urethane science and technology

Editors:

D. Klempner

K.C. Frisch

Advances in Urethane Science and Technology

Editors:

D. Klempner

K.C. Frisch


Advances inUrethane Scienceand Technology

Daniel Klempnerand

Kurt Frisch

Rapra Technology Limited

Shawbury, Shrewsbury, Shropshire, SY4 4NR, United KingdomTelephone: +44 (0)1939 250383 Fax: +44 (0)1939 251118

http://www.rapra.net

First Published in 2001 by

Rapra Technology LimitedShawbury, Shrewsbury, Shropshire, SY4 4NR, UK

©2001, Rapra Technology Limited

All rights reserved. Except as permitted under current legislation no partof this publication may be photocopied, reproduced or distributed in anyform or by any means or stored in a database or retrieval system, without

the prior permission from the copyright holder.

A catalogue record for this book is available from the British Library.

Typeset by Rapra Technology LimitedPrinted and bound by Lightning Source UK

ISBN: 1-85957-275-8

Dedication

In memory of Kurt C FrischOne of the founding Fathers of polyurethanes

January 15th 1919 to October 21st 2000

i

Contents

1 Dimensional Stabilising Additives for Flexible Polyurethane Foams ................ 3

1.1 Introduction ............................................................................................. 3

1.2 Experimental Procedures .......................................................................... 6

1.2.1 Materials ....................................................................................... 6

1.2.2 Handmix Evaluations .................................................................... 8

1.2.3 Machine Evaluation ...................................................................... 9

1.3 TDI - Flexible Moulded Additives .......................................................... 15

1.3.1 Dimensional Stability Additives for TDI ...................................... 16

1.3.2 Low Emission Dimensional Stability Additives ............................ 42

1.4 MDI Flexible Moulded Foam Additives ................................................. 63

1.4.1 Dimensional Stability Additives for MDI .................................... 64

1.4.2 Low Emissions Dimensional Stability Additives in MDI.............. 67

1.5 TDI Flexible Slabstock Low Emission Additives ..................................... 73

1.5.1 Reactivity .................................................................................... 74

1.5.2 Standard Physical Properties ........................................................ 74

1.5.3 TDI Flexible Slabstock Foam Review .......................................... 74

1.6 Foam Model Tool Discussions ................................................................ 75

1.6.1 TDI and MDI Moulded Foam Model .......................................... 75

1.6.2 TDI Flexible Slabstock Foam Model ........................................... 78

1.7 Conclusions ............................................................................................ 81

2 Demands on Surfactants in Polyurethane Foam Production withLiquid Carbon Dioxide Blowing .................................................................... 85

2.1 History of Polyurethane Foams .............................................................. 85

2.1.1 Environmental Concerns in Relation to Flexible Foam Density ... 86


ii

2.2 Current Liquid Carbon Dioxide Technologies for Flexible Slabstock .........Polyether Foam Production .................................................................... 88

2.2.1 Machinery ................................................................................... 88

2.2.2 The Foaming Process ................................................................... 90

2.2.3 Additional Tasks of Silicone Surfactants in Flexible SlabstockFoam Production ......................................................................... 95

2.2.4 Chemistry of a Silicone Surfactant in Flexible SlabstockFoam Production ......................................................................... 99

2.2.5 A Surfactant Development Example .......................................... 101

3 Polyurethane Processing: Recent Developments ........................................... 113

3.1 Industrial Solutions for the Production of Automotive SeatsUsing Polyurethane Multi-Component Formulations ........................... 113

3.1.1 Market Requirements ................................................................ 113

3.1.2 Dedicated Solutions: Metering Equipment ................................ 114

3.1.3 Dedicated Solutions: Mixing Heads........................................... 116

3.1.4 Dedicated Solutions ................................................................... 121

3.2 ‘Foam & Film’ Technology - An Innovative Solution to FullyAutomate the Manufacture of Automotive Sound Deadening Parts ..... 130

3.2.1 The Problem .............................................................................. 131

3.2.2 The Approach to a Solution ...................................................... 131

3.2.3 The Film .................................................................................... 133

3.2.4 Industrial Applications .............................................................. 135

3.2.5 Applications .............................................................................. 137

3.2.6 Advantages ................................................................................ 138

3.3 InterWet - Polyurethane Co-injection ................................................... 138

3.3.1 Glass-Reinforced Polyurethanes, a Well-Known Technology ..... 139

4 Recent Developments in Open Cell Polyurethane-Filled Vacuum InsulatedPanels for Super Insulation Applications ...................................................... 157

4.1 Introduction ......................................................................................... 157

4.2 Some General Properties of Open Cell PU Foams for Vacuum

iii

Contents

Insulated Panels .................................................................................... 158

4.3 Vacuum Issues in the Selection of VIP Components .............................. 163

4.3.1 Vacuum Properties of the Open Cell Foams .............................. 163

4.3.2 Vacuum Properties of the Barrier Film....................................... 167

4.3.3 The Getter Device ...................................................................... 179

4.4 Vacuum Panel Manufacturing Process and Characterisation ................ 188

4.4.1 Some Manufacturing Issues ....................................................... 188

4.4.2 Characterisation of Vacuum Panels ........................................... 191

4.5 Insulation Performances of Open Cell PU-Filled Vacuum Panels .......... 196

4.6 Examples of VIP Applications and Related Issues................................. 199

4.6.1 Household Appliances ............................................................... 199

4.6.2 Laboratory and Biomedical Refrigerators .................................. 203

4.6.3 Vending Machines ..................................................................... 204

4.6.4 Refrigerated/Insulated Transportation ....................................... 205

4.6.5 Other Applications .................................................................... 206

4.7 Near Term Perspectives and Conclusions.............................................. 206

5 Modelling the Stabilising Behaviour of Silicone Surfactants During theProcessing of Polyurethane Foam: The Use of Thin Liquid Films................. 213

5.1 Introduction ......................................................................................... 213

5.2 Film Drainage Rate: Reynold’s Model and Further Modifications ........ 216

5.2.1 Rigid Film Surfaces .................................................................... 216

5.2.2 Mobile Film Surfaces ................................................................. 217

5.2.3 Surface Viscosity ........................................................................ 217

5.2.4 Surface Tension Gradients ......................................................... 218

5.3 Experimental Investigation of Model, Thin Liquid PolyurethaneFilms and the Development of Qualitative and Semi-QuantitativeModels of Film Drainage ...................................................................... 219

5.3.1 Experimental Details ................................................................. 221

5.3.2 Qualitative Description of Polyurethane Films .......................... 223


iv

5.3.3 Quantitative Measurement of Film Drainage Rates:Bulk and Surface Effects ............................................................ 226

5.4 The Development of Theoretical Models of Vertical, DrainingThin Liquid Model PU Films ................................................................ 236

5.4.1 Rigid-Surfaced Collapsing Wedge Model ................................... 236

5.4.2 Deforming Film Models ............................................................ 239

5.4.3 Tangentially-Immobile Films ..................................................... 242

5.4.4 Finite Surface Viscosity .............................................................. 245

5.4.5 Adding Surfactant Transport ..................................................... 249

5.5 Summary .............................................................................................. 254

6 Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea-Acrylic/Styrene Polymer Dispersions ............................................................ 261

6.1 Preface .................................................................................................. 261

6.2 Introduction ......................................................................................... 261

6.2.1 General Considerations ............................................................. 261

6.2.2 Acrylic Dispersions and Polyurethane Dispersions (DPUR) ....... 264

6.2.3 Hybrid Acrylic-Urethane Dispersions ........................................ 266

6.3 Concept of the Study ............................................................................ 268

6.3.1 Selection of Starting Materials ................................................... 268

6.3.2 Assumptions for Synthesis of Hybrid Dispersions ..................... 269

6.4 Methods of Testing ............................................................................... 276

6.4.1 Dispersions ................................................................................ 276

6.4.2 Coatings .................................................................................... 277

6.4.3 Films .......................................................................................... 278

6.5 Experimental results ............................................................................. 279

6.5.1 Characterisation of Starting Dispersions Used for Synthesisof MDPUR ................................................................................ 279

6.5.2 Synthesis of MDPUR and MDPUR-ASD ................................... 288

6.5.3 Investigation of the Effect of Various Factors on theProperties of Hybrid Dispersions ............................................... 290

v

Contents

6.5.4 Additional Experiments ............................................................. 312

6.6 Discussion of results ............................................................................. 320

6.6.1 Estimation of the Effect of Various Factors on the Propertiesof Hybrid Dispersions and Films and Coatings Madefrom Them ................................................................................ 320

6.6.2 Mechanism of Hybrid Particle Formation ................................. 326

6.7 Summary ................................................................................................ 330

7 Adhesion Behaviour of Urethanes ................................................................ 335

7.1 Introduction ......................................................................................... 335

7.2 Surface Characteristics of PU Adhesive Formulations ........................... 335

7.2.1 Experimental ............................................................................. 336

7.2.2 Results and Discussion .............................................................. 338

7.2.3 Conclusions ............................................................................... 347

7.3 Acid/Base Interactions and the Adhesion of PUs to Polymer Substrates 347

7.3.1 Experimental ............................................................................. 348


7.4 The Effectiveness of Silane Adhesion Promoters in the Performanceof PU Adhesives .................................................................................... 355

7.4.1 Experimental ............................................................................. 356


7.4.3 Conclusions ............................................................................... 364

8 HER Materials for Polyurethane Applications ............................................. 369

8.1 Introduction ......................................................................................... 369

8.2 Experimental Conditions ...................................................................... 370

8.2.1 Chain Extenders ........................................................................ 370

8.2.2 Prepolymers ............................................................................... 370

8.2.3 Preparation of Cast Elastomers ................................................. 372

8.2.4 Physical and Mechanical Properties Determination ................... 372

8.3 HER Materials Synthesis and Characterisation .................................... 373


vi

8.4 Cast Poly(Ether Urethanes) ................................................................... 375

8.4.1 Pot Life Determination .............................................................. 375

8.4.2 Polyurethane Castings ............................................................... 376

8.4.3 Calculation of Hard and Soft Segment Contents ....................... 376

8.4.4 Hard Segment Versus Hardness ................................................. 378

8.4.5 Tensile Properties ....................................................................... 378

8.4.6 Tear, Compression Set and Rebound Properties......................... 380

8.4.7 Differential Scanning Calorimetry ............................................. 381

8.4.8 Dynamic Mechanical Analysis ................................................... 383

8.5 Cast Poly(Ester Urethanes) ................................................................... 390

8.5.1 Pot Life Determination .............................................................. 390

8.5.2 Tensile Properties ....................................................................... 390

8.5.3 Tear, Fracture Energy, Compression Set andRebound Properties ................................................................... 390

8.5.4 Differential Scanning Calorimetry Analysis ............................... 393


8.6 Cast Polyurethanes from HER/HQEE Blends ....................................... 397

8.6.1 Freezing Point Determination of HER/HQEE Blends ................ 397

8.6.2 Cast elastomers and Their Properties......................................... 398

8.7 High Hardness Cast Polyurethanes ....................................................... 401

8.7.1 Cast Elastomers and Their Hard and Soft Segment Contents .... 401

8.7.2 Hardness, Tensile, Tear, Compression Set andRebound Properties ................................................................... 401

8.7.3 FT-IR Analysis of Cast Polyurethanes ........................................ 403

8.7.4 Differential Scanning Calorimetric Analysis .............................. 405


8.8 High Thermal Stability Polyurethane with Low Heat Generation ........ 405

8.8.1 Hardness Measurements ............................................................ 408

8.8.2 Tensile Measurements ................................................................ 408

8.8.3 Differential Scanning Calorimetric Analysis .............................. 410

vii


8.9 Conclusions .......................................................................................... 416

9 Ultra-Low Monol PPG: High-Performance Polyether Polyolsfor Polyurethanes ......................................................................................... 421

9.1 Introduction ........................................................................................... 421

9.2 MDI/BDO Cured Elastomers Based on Ultra-Low Monol PPG Polyols . 424

9.2.1 Effect of Monol Content on 4,4´-Methylene DiphenylmethaneDiisocyanate (MDI)/1,4-Butanediol (BDO) Cured Elastomers... 424

9.2.2 Processability and Property Latitude of Elastomers Based onUltra-Low Monol PPG Polyols .................................................. 429

9.2.3 Processing Latitude Improves by IncorporatingOxyethylene Moieties ................................................................ 434

9.3 One-Shot Elastomer System Based on EO-Capped, Ultra-LowMonol PPG Polyols .............................................................................. 436

9.3.1 Effect of Primary Hydroxyl Concentration on One-ShotElastomer Processability ............................................................ 436

9.3.2 Effect of Monol Content on One-Shot Elastomer Processabilityand Properties............................................................................ 438

9.3.3 Processability and Property Latitude of Elastomers Basedon EO-Capped, Ultra-Low Monol Polyols ................................ 440

9.4.1 MDI/BDO Cured Elastomers: Acclaim Polyol 3205 Versus .............PTMEG-2000 ............................................................................ 445

9.4.2 Enhanced Elastomer Properties Utilising Ultra-Low MonolPPG/PTMEG Blends .................................................................. 447

9.5 Polyol Molecular Weight Distribution Effect on Mechanicaland Dynamic Properties of Polyurethanes ............................................ 449

9.5.1 TDI Prepolymers Cured with Methylene Bis-(2-Chloroaniline)[MBOCA] .................................................................................. 450

9.5.2 Moisture-Cured TDI Prepolymers ............................................. 454

9.5.3 Aqueous Polyurethane/Urea Dispersion Coatings ...................... 456

9.5.4 MDI Prepolymers Cured with BDO .......................................... 459

9.6 Conclusions .......................................................................................... 461

Contents


viii

APPENDIX........................................................................................................ 465

Laboratory Preparation of 2,4-TDI and 4,4´-MDI Prepolymers ................... 465

Laboratory Casting of 4,4´-MDI Prepolymers Cured with BDO .................. 465

Laboratory Casting of One-Shot Elastomers Based on Carbodiimide-Modified MDI, Polyol, and BDO ................................................................. 465

Laboratory Casting of 2,4-TDI Prepolymers Cured with MBOCA .............. 466

Laboratory Moisture-Curing of 2,4-TDI Prepolymers ................................. 466

Laboratory Preparation of Aqueous Polyurethane/Urea Dispersionsusing the Prepolymer Mixing Process ........................................................... 466

Abbreviations .................................................................................................... 469

Contributors ...................................................................................................... 473

Author Index ..................................................................................................... 477

Main Index ........................................................................................................ 483

1

Preface

This is a landmark issue of ‘Advances in Urethane Science and Technology’. Notonly is this the first volume of the new millennium, but it is the first to be publishedby Rapra Technology.

On a more solemn note, one of the editors, Kurt C. Frisch, passed away shortlybefore publication. Dr. Frisch, founder of the University of Detroit Mercy’s PolymerInstitute, was one of the pioneers of polyurethanes and was responsible for thesuccessful introduction of polyether polyurethane flexible foams into commerce inthe mid-1950s. Let us not only mourn the loss of, but also celebrate the life of thisgreat scholar by continuing to further the frontiers of urethane science andtechnology. This volume is a good example of this progress.

Polyurethanes continue to be one of the most versatile of all polymers, findingapplications in foams (flexible, rigid, and in-between), elastomers, coatings, sealants,adhesives, paints, textiles, and films. This volume presents some of the majoradvances in polyurethanes, both from the materials and research side of things aswell as processing and applications, and includes studies on foams (additives,vacuum panel applications, blowing and processing), elastomers, adhesionbehaviour and new urethane raw materials.

I would like to take this opportunity to express my gratitude to the authors whocontributed to this book and to the University of Detroit Mercy for itsencouragement of this effort.

I would also like to thank the staff of Rapra, in particular, Frances Powers, ClaireGriffiths and Steve Barnfield.

Daniel Klempner, Ph. D.

Polymer Institute,University of Detroit MercyJuly, 2001


2

3

1.1 Introduction

The issues that an automotive seat manufacturer faces when formulating and producingseats are escalating. Physical properties such as tensile and tear strengths, compressionset and wet set are critical when meeting specific mechanical performance requirementsas defined by the original equipment manufacturer (OEM). As new requirements forcomfort and durability are instituted, tests such as dynamic creep testing, long termvibration characterisation and repeated compression tests under various atmosphericand load conditions have been used to characterise foam performance for comfort.

Comfort properties are best controlled by the polyols used to produce the polyurethanefoam cushion. Significant changes in polyol technology to meet these dynamic comfortproperties have had an impact on the processing of polyurethane foam and on physicalproperties. Increased tightness of the foam article resulting from changes in these rawmaterials has focused more attention by foam producers on crushing methods. Flexiblemoulded polyurethane foam requires some type of mechanical crushing to preventshrinkage and ultimately maintain part stability.

With recent changes made to polyol technology, mechanical methods of crushing do notalways provide the consistency required to produce a part that is dimensionally stable.Additionally, producers of polyurethane articles are continually building more complexityinto their seat designs to meet the aesthetic values required by today’s consumers. Thesecomplex seat designs place more emphasis on crushing capability due to the nature ofthe designs. With all these changes, additives needed to be developed which provide awider processing latitude and increased breathability to the polyurethane article. Widerprocessing latitudes should reduce scrap and repair rates on the foam production lineand improve economics for the polyurethane producer [1].

The formation of moulded foam is a complicated chemical process which involves severalreactions occurring simultaneously. There are rapid volume and temperature increasesand the concurrent development of phase separated polymer networks. To understandhow foam properties can be affected by catalyst and surfactant chemistries severaltechniques are used to identify key performance benefits and issues. A force to crush

1Dimensional Stabilising Additives for FlexiblePolyurethane Foams

Gary D. Andrew, Jane G. Kniss, Mark L. Listemann, Lisa A. Mercando, James D. Tobias and Stephan Wendel


4

(FTC) detection device was used to measure the force required to crush a part to 50% ofits thickness for determination of cell openness. Mass-loss/rate-of-rise was run tounderstand rate of rise and height measurements, weight loss from carbon dioxidegeneration and temperature profiles. A scanning electron microscope (SEM) was used todetermine differences in cell structure and cell distribution caused by changes in thecatalyst and surfactant chemistries. Physical properties were also tested using ASTM testmethods for flexible cellular polyurethane. A novel chemical reaction foam modellingtechnique was also used to determine the selectivity of the catalyst packages, comparedto industrial standard controls [2].

In the past it was thought that the cell structure of polyurethane foam is controlled bythe type and amount of surfactant used. Dabco DC5043 (Air Products and Chemicals,Inc.) was developed to enhance cell wall drainage to better enable cell opening duringcrushing cycles. It was also thought that surfactant technology was the best way toprovide improved crushing techniques; therefore, catalyst technology was ignored [3].As mentioned earlier, with new polyol technology development more emphasis was placedon crushing. New additive technology needed to be developed that would open cellsduring the foam formation and reduce the requirement and criticality of the crushingprocesses. The technology had to go beyond providing easier cell opening at crush toproviding more open cells during the polyurethane formation.

The real challenge in polyurethane foam formation is to control the chemical andphysiochemical processes up to the point where the material finally sets. The sequenceand the rate of the chemical reactions are predominately a function of the catalyst andthe reactivity of the basic raw materials, polyol and isocyanate. The physiochemicalcontribution to the overall stability and processability of a system is provided by thesilicone surfactants. Optimum foaming results will be achieved only if the correctrelationship between chemistry and physics exists [4].

Another rapidly increasing environmental concern is over the emission of volatileorganic compounds (VOC) during and following the production of industrial andconsumer goods. This has stimulated a great deal of effort within the chemical industryto reduce and/or control the ways in which such emissions may occur. In thepolyurethane foam industry, efforts to reduce VOC emissions have greatly impactedthe technologies used in manufacturing processes, especially for the use of organicauxilliary blowing agents such as chlorofluorocarbons. In addition, the ultimate fateof additional foaming additives, including surfactants and catalysts, is now comingunder increased global scrutiny. As a result, foam manufacturers have expressed adesire for polyurethane additives that, among other things, do not exhibit the degreesof fugitivity common to many of the additives that are currently used in polyurethanefoam production today.

5

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Polyurethane foams are prepared from the simultaneous reactions of diisocyanate withwater and with polymeric diols and/or triols to form hydrogen-bonded urea (hard)segments and polyurethane networks (soft segments). The commercial production ofpolyurethanes via isocyanate poly-addition reactions requires the use of one or morecatalysts. Tertiary amines are widely accepted in the industry as versatile polyurethanecatalysts. Amine catalysts are generally stable in the presence of standard polyurethaneformulation components and can have an impact on both the blowing (water-isocyanate)and gelling (polyol-isocyanate) reactions. Although the use of catalysts in themanufacturing of polyurethane foam both speeds the production of the foamed articleand, through the judicious choice of catalyst package, allows control of the physicalproperties of the product, there are some problems associated with the use of theseadditives. A number of commonly used tertiary amine catalysts can volatilise under certainconditions. Release of tertiary amines during foam processing and from consumer productsis generally undesirable. Therefore, identifying alternatives to standard tertiary aminecatalysts which have no or low volatility, yet exhibit the same type of activity in isocyanatepoly-addition reactions, is desirable.

The non-fugitive catalysts reported in this chapter address the problems associated withthe use of polyurethane catalysts by reducing the odour and volatility of these materialsand by eliminating the ability of these additives to escape from finished foam products.One strategy has involved functionalising the catalysts to render the species reactivetoward isocyanates, thereby covalently attaching the catalysts to the polymer network.This strategy not only renders the catalytic material non-fugitive in the final product, butalso reduces the odour and volatility of the catalyst through increases in molecular weightand polarity. These non-fugitive catalysts also provide equivalent or improved physicalproperties when compared to industry standards, whereas conventional reactive aminecatalysts as well as metal catalysts cannot always meet todays ever increasing manufacturerand consumer performance requirements.

These increasingly evolving requirements have led to the development of both novelnon-fugitive catalysts and new cell-opening non-fugitive catalysts for flexible foam. Thesenew low emission additives have been developed to meet the challenge of optimisedfoaming and result in little or no emissions. Several of the non-fugitive catalysts possesscell-opening capability. This new technology allows the manufacturer of polyurethanefoam to optimise their system to achieve the best processing latitude for their foam process.These new additives maintain, or in some cases, improve key physical properties whileproviding a more open foam.


6

1.2 Experimental Procedures

Data presented herein was derived from a combination of handmix and high pressureimpingement-mixing machine produced foam. Foams were prepared using several generaltypes of formulations for toluene diisocyanate (TDI) and two general types of formulationsfor methylenediphenyl diisocyanate (MDI) which are representative of currently utilisedformulations in the automotive interior component industry. In addition, an all waterblown formulation was used to represent the flexible slabstock industry.

1.2.1 Materials

The materials used are shown in Table 1.1.

krowlatnemirepxenidesuslairetaM1elbaT

emanedarT noitalumroF rerutcafunaM

VL33ocbaD GPDniADET%33 ICPA

71-LBocbaD .tsylatacgniwolbenimayraitretnoitcadeyaleD ICPA

35-LBocbaD sedivorphcihw,enimayraitretdepolevedylweN.seitilibapacgninepollecdnagniwolb

ICPA

61-BocbaD .tsylatacerucecafrusenimayraitreT ICPA

)51-CP(51tacyloP .tsylatacenimaevitcaerdecnalaB ICPA

VLBocbaD .dnelbtsylatac1:3ani11-LBocbaD/VL-33ocbaD ICPA

9-TocbaD .tsylatacetaotcosuonnatS ICPA

0601ENocbaD roftsylatacgnillegevitiguf-nondepolevedylweN.snoitacilppadedluomelbixelf

ICPA

5801N-FX gninepollecevitiguf-nondepolevedylweN.tsylatacgniwolb

ICPA

005ENocbaD roftsylatacgnillegevitiguf-nondepolevedylweN.ICPAmorfmaofkcotsbalselbixelf

ICPA

006ENocbaD etaidemretnievitiguf-nondepolevedylweNkcotsbalselbixelfroftsylatacgniwolb

.snoitacilppa

ICPA

002ENocbaD etaidemretnievitiguf-nondepolevedylweNkcotsbalselbixelfroftsylatacgniwolb

.snoitacilppa

ICPA

7

1elbaT deunitnoC

emanedarT noitalumroF rerutcafunaM

60011O-FX gninepollecevitiguf-nondepolevedylweN.tsylatacgnilleg

ICPA

9615CDocbaD .smetsyserucdlocroftnatcafrusremylopocenociliS ICPA

3405CDocbaD erucdlocIDTroftnatcafrusremylopocenociliS.smetsys

ICPA

5852CDocbaD erucdlocIDMroftnatcafrusremylopocenociliS.smetsys

ICPA

7152CDocbaD .tnatcafrusremylopocenociliS ICPA



6851N-FX ,tnatcafrusremylopocenocilisdepolevedylweN.sllecneposetomorphcihw

ICPA

7851N-FX ,tnatcafrusremylopocenocilisdepolevedylweN.sllecneposetomorphcihw

ICPA

FL-AOED)ocbaD(

mroFdiuqiLenimalonahteiD)retaw%51:AOED%58(

ICPA

848ElocrA 5.13fo#HOnahtiwloyloplanoitnevnoC llednoyLlacimehC

158ElocrA 5.81fo#HOnahtiwloylopremylopocsdilos%34 llednoyLlacimehC

036-CN 4.13fo#HOnahtiwloyloP lacimehCwoD

007-CN 0.12fo#HOnahtiwloylopremylopocsdilos%14 lacimehCwoD

2153lonaroV 3.84fo#HOnahtiwloyloprehteyloP lacimehCwoD

AloyloP 5.23fo#HOnahtiwloirtytilanoitcnufhgiH lacimehCwoD

BloyloP 8.32fo#HOnahtiwloirtremylopocsdilos%14 lacimehCwoD

CloyloP 82fo#HOnahtiwloyloprehteyloP lacimehCwoD

DloyloP loylopgninepolleC

897-CRP tnegaesaelerdesab-tnevloS dnerT-mehC

.cnI,slacimehCdnastcudorPriA:ICPAenimalonahteiD:AOED

locylgenelyporpiD:GPDenimaidenelyhteirT:ADET



8

1.2.2 Handmix Evaluations

1.2.2.1 Flexible Moulded Foam Handmix Procedure

Handmix experiments were carried out using the following procedure. Formulationswere blended together for approximately 10 minutes using a mechanical mixerequipped with a 7.6 cm diameter, high shear, mixing blade, rotating at 5000 rpms.Premixed formulations were maintained at 23 ± 1 °C using a low temperatureincubator. Mondur TD-80 (Bayer; a blend of 2,4-TDI and 2,6-TDI isomers in theratio of 4:1) or modified MDI was added to the premix at the correct stoichiometricamount for the reported index for each foam. The mixture was blended togetherwith a Premier Mill Corporation Series 2000, Model 89, dispersator for approximatelyfive seconds. The foaming mixture was transferred to an Imperial Bondware #GDR-170 food container or ‘chicken’ bucket and allowed to free rise in order to obtain theprocessing data.

1.2.2.2 Flexible Slabstock Foam Handmix Procedure

Handmix experiments were carried out using the following procedure. A premixconsisting of polyol, surfactant and water was prepared by blending the componentsin a shaker for approximately 20 minutes. The premix was allowed to stand for 2hours prior to making the foam to allow for degassing of the mixture. A measuredamount of premix was poured into a 1.9 litre paper cup; the required stoichiometricamounts of amine and tin catalysts were added to the contents of the cup and mixedfor 20 seconds using a Premier Mill Corporation dispersator equipped with a 5.5 cmdiameter, high shear, mixing blade, rotating at 6,000 rpm. The corresponding amountof Mondur TD-80 to provide for a 110 index (isocyanate index, which is the amountof isocyanate used relative to the theoretical equivalent amount [5]) was measuredinto a 400 cm3 beaker. Methylene chloride in the correct proportion was added tothe beaker containing the Mondur TD-80; the beaker was carefully swirled for 4 or5 seconds and the contents poured into the paper cup. The mixture was blendedtogether for 6-7 seconds and the foaming mixture poured into a paper bucket for upto 12 seconds and allowed to free rise with the processing data being recorded.Reactivity profiles were determined from hand-mix foams prepared in 5.68 litre paperbuckets. Foams for physical properties were prepared in 35.6 x 35.6 x 25.4 cmcardboard boxes. Identical procedures were followed for both reactivity and physicalproperty experiments.

9

1.2.3 Machine Evaluation

1.2.3.1 TDI Flexible Moulded Foam Procedure

Machine runs for the TDI flexible moulded foam were carried out on a Hi Tech SureShotMHR-50 (Hi-Tech Industries, Inc.), cylinder displacement series, high pressure machine.Fresh premixes, consisting of the appropriate polyols, water, crosslinker, surfactants andcatalysts for each formulation were charged to the machine. Mondur TD-80 was usedthroughout the entire study. All chemical temperatures were held at 23 ± 2 °C via themachine’s internal temperature control units. The foam was poured into an isothermallycontrolled, heated aluminium mould maintained at 71 ± 2 °C. The mould was a typicalphysical property tool designed with internal dimensions of 40.6 cm x 40.6 cm x 10.2cm. The mould has five vents, each approximately 1.5 mm in diameter, centred 10.0 cmfrom each edge and the geometric centre of the lid. The mould was sprayed with asolvent-based release agent, Chem-Trend PRC-798, prior to every pour and allowed todry for one minute before pouring. The foam premix was puddle poured into the centreof the mould with a wet chemical charge weight capable of completely filling the mouldand obtaining the desired core density. Minimum fill requirements were established foreach formulation evaluated. The foam article was demoulded at 240 seconds after theinitial pour. After demoulding, the foam was placed through a mechanical crusher, testedfor FTC measurements, or left uncrushed and set aside for 24 hour shrinkagemeasurements described in Section 1.2.3.2c.

All foams to be tested in each catalyst set were mechanically crushed 1 minute afterdemoulding using a Black Brothers Roller crusher set to a gap of 2.54 cm. Crushing wascarried out three times on each part, rotating the foam 90 degrees after each pass throughthe roller. All parts produced for physical testing were allowed to condition for at leastseven days in a constant temperature and humidity room (23 ± 2 °C, 50 ± 2% relativehumidity).

Three to four specimens were produced for any given set of conditions. Four test specimenswere die-cut from each foam pad and evaluated for each physical property listed insubsequent data tables. All results were included in calculating averages and standarddeviation. Each test was carried out as specified in ASTM D3574 [5].

For each formulation evaluated, duplicate free rise ‘chicken’ buckets were poured at thesame shot size to determine overall reactivities and foam shrinkage. Data recorded werecream time (the time between the discharge of the foam ingredients from the mixinghead and the beginning of the foam rise [5]), top-of-cup (TOC; the time between thedischarge of the foam ingredients from the mixing head and when the centre of the foamreaches the same height as the top of the chicken bucket), string gel (the time between



10

pouring of the mixed liquids and the time that strings of viscous material can be pulledaway from the surface of the foam when it is touched with a tool [5]), full rise time andfinal height. The free rise buckets were again tested for final heights after 24 hours.Measurements of height were made using a Mitutoyo height gauge. In addition to all thestandard tests, several more unique tests were performed where indicated, and aredescribed in Section 1.2.3.2.

1.2.3.2 Tests

1.2.3.2a Maze Flow Mould Test Description

A common type of isothermally heated mould was used to determine the flowability offormulations with each of the catalyst candidates. This maze mould is shown in Figure 1.1.

Machine foam was poured into the mould at the top left corner of the open cavity asindicated by ‘pour spot’ on the figure. The lid was then closed and clamped tightly. Foamwas allowed to free flow consecutively through each of the five gates for the standard 4

Figure 1.1 Diagram of Maze Flow Mould (Top View)

11

minutes prior to demould. Minimum fill was first determined by completely filling thecavity with little or no extrusion through the vent at the end of the fifth gate. Mathematicalreduction of the shot size was performed to obtain the first of three systematically scaleddown foam fill weights. This first foam should have a fifth leg (the foam in gate 5 of themaze flow mould, see Figure 1.1) which barely touches the front cavity wall. The secondreduction in foam fill weight produced a foam that flowed approximately halfway throughthe fifth gate. The third reduction in foam fill weight was equivalent to the step changefrom foam 1 to foam 2. Shot times were held constant for each of the three foam fillweights as compared to the control determined standard shot time in any given solidslevel formulation. These three foams were weighed for total foam pad and fifth leg weight,and measured for fifth leg length to obtain a range of flow values for each of theexperimental catalysts compared to the control additives.

1.2.3.2b Dimensional Stability Test

Foam dimensional stability is essentially the result of a balance between external andinternal forces. The external forces are defined as the ambient pressure along with anyadditional applied loads. The internal forces are the strength of the polymer matrix andthe internal cell pressure [6]. Basically, if the sum of the internal forces is greater than theexternal forces, the foam will expand. Consequently, if the sum of the external forces isgreater than the internal forces the foam will shrink. Any expansion or shrinkage willimpact on the internal and/or external forces until an equilibrium is obtained. It is theinternal forces, i.e., cell pressure and strength of the polymer matrix as defined by ‘greenstrength’ or cure, which will have an impact on the dimensional stability performance ofthe moulded polyurethane.

Dimensional stability can be measured on a freshly demoulded part by determining theamount of force required to open cells, as measured by FTC. FTC measurements weremade thirty seconds after demoulding. The foam pad was removed from the mould,weighed and placed in the FTC apparatus. The force detection device is equipped with a2.2 kg capacity pressure transducer mounted between the 323 cm2 circular plate crosshead and the drive shaft. The actual force is shown on a digital display. This devicemimics the ASTM D3574, Indentation Force Deflection Test [6] and provides a numericalvalue of the freshly demoulded foam’s initial hardness or softness. The foam pad wascompressed to 50 percent of its original thickness at a cross head velocity of 275 mm perminute with the force necessary to achieve the highest compression cycle recorded inwhole Newtons. Several compression cycles were completed. A cycle takes approximately30 seconds to complete. Values are reported as the FTC value for the foam based on theassumption that the lower the FTC values the better the dimensional stability of thefoam. This test requires the foam to be fully cured at demould. A dimensionally stable



12

foam will exhibit little or no tendency to shrink after demoulding. Poor dimensionalstability can result in numerous defects of the polyurethane article, such as lack of fit ofa polyurethane piece to the substrate. These defects will ultimately cause loss of revenueto the polyurethane manufacturers because of increased repair and/or scrap rates.

Additionally the degree of cell openness of polyurethane foam can be measured directlyby the air flow physical properties of the polyurethane part. Higher air flow valuesmeasured for a particular foam would indicate that the foam has less of a tendency toshrink and therefore be more dimensionally stable as compared to a foam with lower airflows. Additionally, higher air flows may also indicate that the foam was much easier tocrush-out thereby breaking many of the cell windows. Dimensionally stable foam shouldreduce scrap and rework by allowing the foam to conform to near its original mouldedshape or at least return to its original shape after being crushed.

1.2.3.2c Shrinkage Test

An additional dimensional stability evaluation was done with a shrinkage templateapparatus (Figure 1.2).

Figure 1.2 Diagram of Shrinkage Template Apparatus

13

It was designed to measure the average foam pad shrinkage of an uncrushed foam. Thisapparatus consists of two 432 mm long x 432 mm wide x 6.35 mm thick, plexiglassplates, mechanically pinned in each corner with threaded bolts to maintain the plates ata constant 102 mm spacing. The single uncrushed foam from each of the catalyst setswas aged 24 hours prior to being placed between the two plates. Both top and bottomplates each contain 5 mm diameter holes evenly spaced, diagonally from corner to corner,25 mm apart in an X-shaped pattern. Nineteen holes are contained in each leg of the X,for a total of 37 holes per plate. Measurements were made with a digital caliper byinserting the end down through each hole to just touch the foam surface with the indicatedvalue being recorded. All measurements were normalised to discount the plexiglass platethickness and subsequently averaged to a single mould cavity and lid value.

1.2.3.2d Time Pressure Release Test

Time Pressure Release (TPR) is the opening of the mould during the curing cycle torelease the internal pressure and then re-closing for the duration of the cure time [8]. Thesudden release of the internally generated pressure bursts the cell windows, therebyimproving the crushability of the foam. The tool is opened only a few millimetres andfor a specific time. TPR can be applied at any time during the curing cycle, however, caremust be taken not to perform the operation too early or too late since surface qualityissues may occur.

A ‘simulated’ TPR process was carried out during this study, whereby the tool lid wasopened approximately 1.5 mm for a three second duration. TPR was applied at varioustime intervals throughout the evaluation. Two mechanical clasps affixed to the top andbottom halves of the tool precisely controlled the gap opening. These clasps were manuallyopened and closed at the desired TPR time interval.

1.2.3.2e Pail Test

Another test was devised to evaluate foam bulk stability in a free-rise mixed foam.Foams prepared from TDI formulations were poured directly from the machine headinto a large open pail (the pail is a common high density polyethylene plastic withapproximate dimensions of 365 mm high and 290 mm in diameter) at a targeted mass.Several pours were carried out to ensure equivalent catalyst activity amongst eachformulation. Foams were allowed to stand for 24 hours prior to removal from the pail.Each foam was weighed to obtain total individual mass. Subsequently, a 25 mm slicewas cut directly through the geometric vertical centre of the foam. Foam slices wereexamined for cell structure.



14

1.2.3.2f Dynamic Fatigue Test

Dynamic Fatigue Constant Pounding testing was carried out using standard testingprocedures outlined in ASTM D3574-95 [6]. A 60 minute recovery time and 80,000cycles were used for each test sample.

1.2.3.2g Fogging Test

Gaseous foam emissions were compared in the standard fogging test procedure outlinedin SAE J1756 [9]. A standard 7.6 cm thick piece of foam was preconditioned for 48hours, then tested in the Hart fogging apparatus at 100 °C for 3 hours. Glossmeterreadings of the foam were taken after 60 minutes at room temperature.

1.2.3.2h Headspace Analysis Test

Foam emissions were also evaluated by cutting equivalent portions, approximately 1 gram,of foam from the geometric centre of the moulded foam articles 60 minutes after demoulding.Each sample was inserted into a 20 cm3 Kimble glass crimp top vial with a Teflon seal.Several vials were sealed without foam to be used as blanks to ensure all emissions hadeluted through the gas chromatograph prior to the injection of a second gas sample. Allvials were loaded into a Tekmar 7000 Headspace Autosampler tray for sequential heatingfor 1 hour at 54 °C. After heating and temperature equilibration, the headspace of the vialwas sampled and directly injected within a closed loop system onto a Hewlett Packard5890 Series II Plus gas chromatograph containing a HP-5 (5%-diphenyl/95%-dimethylsiloxane copolymer) stationary phase column (30 m, 0.25 mm internal diameter,1.00 mm film thickness). A standard oven heating profile was used to separate the gascomponents for detection with a Hewlett Packard 5972 Series Mass Selective Detector.Elution peaks were individually identified by comparison to standard libraries.

1.2.3.3 MDI Flexible Moulded Foam Procedure

Machine runs for the MDI flexible moulded foam were conducted on a Krauss-Maffei,cylinder displacement series, high pressure machine. Fresh premixes, consisting of theappropriate polyols, water, crosslinker, surfactants and catalysts for each formulation werecharged to the machine. Modified MDI was used throughout the entire study. All chemicaltemperatures were held at 25 °C ± 2 °C via the machine’s internal temperature controlunits. Foam pours were made into an isothermally controlled heated aluminium mouldmaintained at 60 °C ± 2 °C. The mould was a typical physical property tool designed with

15

internal dimensions of 40.6 cm x 40.6 cm x 10.2 cm. The mould has two vents eachapproximately 1.0 mm in diameter centred 10.0 cm from each edge and the geometriccentre of the lid. The mould was sprayed with a solvent-based release agent prior to everypour and allowed to dry for one minute before pouring. The foam premix was puddlepoured 15 cm away from the geometric centre of the mould with a wet chemical chargeweight capable of completely filling the mould with the appropriate core density. Minimumfill requirements were established for each formulation evaluated. The foam article wasdemoulded at 300 seconds after the initial pour. Upon demoulding, the foam was placedthrough a mechanical crusher or tested for FTC measurements.

The foams were mechanically crushed 1 minute after demoulding using a roller crusher set toa gap of 3.0 cm. Crushing was carried out three times on each part. All parts produced forphysical testing were allowed to condition for at least seven days in a constant temperatureand humidity room (23 °C ± 2 °C, 50% ± 2% relative humidity). Three to four parts wereproduced for any given set of conditions. Four test specimens were die-cut from each padand evaluated for each physical property listed. All results were included in calculating theaverages and standard deviation. Each test was conducted as specified in ASTM D3574 [5].

For each formulation evaluated, free rise cup foams (see Section 1.2.2.1) were poured todetermine reactivities and foam shrinkage. Data recorded were gel time, full rise time andfinal height. The free rise cup foams were tested for final heights and free rise density after 24hours. Height measurements were carried out using a Mitutoyo height gauge. All experimentalformulations reported in this work were matched by rise profile to each control formulation.

FTC measurements were conducted 90 seconds after demoulding. The foam pad was removedfrom the mould, weighed and placed in the FTC apparatus (Instron 4502). The force detectiondevice is equipped with a 5.0 kN capacity pressure transducer. The actual force is shown ona digital display. This device mimics the ASTM D3574, Indentation Force Deflection Testand provides a numerical value of freshly demoulded foams initial hardness or softness. Thepad was compressed to 70 percent of its original thickness at a cross head velocity of 380 mmper minute with the force necessary to achieve the highest compression cycle recorded inNewtons. Values are reported as the FTC value for the foam based on the assumption thatthe lower the FTC values the better the dimensional stability of the foam.

1.3 TDI - Flexible Moulded Additives

The automotive industry has placed increased pressures on OEM suppliers to improvetheir productivity, quality and cost of the polyurethane articles which they produce.Styling changes, complex designs and OEM productivity demands for automotive seatshave necessitated the need to produce more open foam in all-water blown TDI- and



16

MDI-based systems. Furthermore, changes in polyol technologies designed to improveseat comfort factors can have a negative impact on foam openness and dimensionalstability. Specifically, the comfort of high resilience foam may be related to the degree ofdimensional stability or foam openness.

Flexible moulded polyurethane foam requires mechanical crushing to open foam cells,which in turn prevents shrinkage and improves overall dimensional stability. Currentmechanical methods for cell opening consist mainly of roller crushing, vacuum ruptureand TPR. However, mechanical methods do not always result in complete or consistentcell opening and require a flexible moulded foam producer to invest in additionalmachinery. Additionally, if the polyurethane article is not crushed properly, dimensionalstability suffers which can cause an increase in repair and scrap rates resulting in a negativeimpact on the cost of production. A chemical method for cell opening would be preferred.

1.3.1 Dimensional Stability Additives for TDI

When producing flexible high resilience foam, it is important to provide a wider TPRwindow to expand processing latitude and at the same time maintain or improve physicalproperties. This should result in reduced scrap and/or repair rates, providing improvedeconomics for the polyurethane producer.

The most commonly used catalyst and surfactant package for all water blown TDI-based moulded foam production is a blend of Dabco 33-LV and Dabco BL-11, coupledwith Dabco DC5043 silicone surfactant. Additionally, the acid-blocked counterparts ofthese two catalysts, Dabco 8154 and Dabco BL-17, can also be used for the productionof high resilience moulded foam. A combination of Dabco 33LV and Dabco BL-17 isused to facilitate a short delay in the reactivity of the polyurethane foaming process. Acombination of silicone surfactants, Dabco DC5043 and Dabco DC5169, are utilised toprovide good foam stabilisation, improve cell regulation and cell wall drainage. Thesecombinations of catalysts and surfactants served as the control additives to which theexperimental additives were compared and contrasted.

A newly developed cell opening catalyst, Dabco BL-53, was evaluated to determine itsimpact on dimensional stability and general processability. Dabco BL-53 affords all thebenefits of Dabco BL-11 or Dabco BL-17, with the added advantage of cell opening andslightly delayed initiation times. Dabco BL-53 is not a chemical equivalent for DabcoBL-11 or Dabco BL-17; however, it will provide similar performance. For rapiddemoulding systems, it is recommend that Dabco BL-53 be used at 0.12 to 0.22 pphp,with the optimum level at 0.16 to 0.19 pphp, in combination with a Dabco 33-LV levelat 0.30 to 0.32 pphp.

17

As polyurethane seating design changed from the relatively simple configurations ofthe early 1990s to the more complex designs of today, the need to improve cell openingor dimensional stability has intensified. Accordingly, silicone surfactants providingimproved foam stabilisation, cell regulation and cell wall drainage were needed toenable polyurethane manufacturers to achieve their production goals. Two experimentalsilicone surfactants, X-N1586 and X-N1587, were developed to provide open foamand promote dimensional stability.

A TDI cushion formulation, with a density of 45 kg/m3 and a TDI back formulation,with a density of 35 kg/m3 were used in the TDI automotive study. All of theseformulations were modified accordingly with the appropriate crosslinker, water, andadditive levels for the chosen density range. These formulations are shown inTables 1.2 and 1.3.

m/gk54~2.1elbaT 3 noitalumrofnoihsuc

noitalumroF lortnoC 35LB FSS.pxE FSS.pxE/35LB

noitacifitnedI I II III VI

stnenopmoC phpp phpp phpp phpp

AloyloP 00.86 00.86 00.86 00.86

BloyloP 00.23 00.23 00.23 00.23

retaW 77.2 77.2 77.2 77.2

VL33ocbaD 03.0 03.0 03.0 03.0

71-LBocbaD 81.0 - 81.0 -

35-LBocbaD - 02.0 - 02.0

3405-CDocbaD 57.0 57.0 - -

9615-CDocbaD 52.0 52.0 - -

6851N-X - - 3.0 3.0

7851N-X - - 7.0 7.0

FL-AOEDocbaD 35.1 35.1 35.1 35.1

)xedni001(IDT 3.73 3.73 3.73 3.73

tnatcafrusenocilislatnemirepxe:FSS.pxEloylopderdnuhrepstrap:phpp



18

1.3.1.1 Reactivity

1.3.1.1a Handmix Mass-Loss/Rate-of-Rise

Formulations used in the mass-loss/rate-of-rise are summarised in Tables 1.2 and 1.3.Surfactant and catalyst additives were changed according to the formulation being studied.Foams were run at the optimum index as they were during the machine study. All experimentswere duplicated. Each mixed formulation was poured into a ‘chicken’ bucket equipped witha thermocouple positioned at the centre of the bucket resting on a Mettler PM 30,000 balance.The centre height of the rising foam was recorded every second using a DAPS QA Model#2500 rate-of-rise apparatus. Knowing the foam mass, the rate-of-rise and using the idealgas law, it is possible to calculate the carbon dioxide generated or trapped over time [10, 11].

Figures 1.3 and 1.4 show height versus time achieved for the cushion and back formulationsusing the control formulations I and V, as well as the experimental formulations II, III, IV,VI, VII and VIII. The foam height versus time graphs clearly indicate higher rates forcontrol formulations I and V in both cushion and back formulations. Cushion and back

m/gk53~3.1elbaT 3 noitalumrofkcab

noitalumroF lortnoC 35LB FSS.pxE FSS.pxE/35LB

noitacifitnedI V IV IIV IIIV


AloyloP 00.08 00.08 00.08 00.08

BloyloP 00.02 00.02 00.02 00.02

retaW 72.3 72.3 72.3 72.3

VL33ocbaD 03.0 03.0 03.0 03.0

71-LBocbaD 810 - 81.0 -

35-LBocbaD - 02.0 - 02.0

3405-CDocbaD 57.0 57.0 - 57.0

9615-CDocbaD 52.0 52.0 - -

6851N-X - - 3.0 3.0

7851N-X - - 7.0 7.0

FL-AOEDocbaD 35.1 35.1 35.1 35.1

)xedni001(IDT 2.24 2.24 2.24 2.24

19

formulations, which contained the new additives in formulations II, III, IV, VI, VII andVIII did not achieve the same foam height when compared to the control formulations.Several things could cause the differences observed in the foam heights. First, reactivityrates for control formulations might be faster than the experimental catalyst or surfactants.Second, overall foam stability could be compromised for the experimental catalyst andsurfactants. Lastly, carbon dioxide might be diffusing from the reacting polyurethane foam

Figure 1.3 Foam Height versus Time - Cushion Formulation

Figure 1.4 Foam Height versus Time - Back Formulation



20

at an accelerated rate. All of these possibilities were explored. Data generated andobservations will be reported which demonstrate how these new additives promotedimensional stability through increased carbon dioxide diffusion.

Temperature profiles for control formulations I and V indicate reaction temperatureswhich fall in the middle of the temperature profiles compared to the new additives denotedin Figures 1.5 and 1.6. These temperature profiles clearly demonstrate that carbon dioxide

Figure 1.5 Temperature versus Time - Cushion Formulation

Figure 1.6 Temperature versus Time - Back Formulation

21

conversion is occurring at the same rate in the control as in the experimental formulations.The fact that carbon dioxide conversion is occurring at the same rates would not accountfor the lower foam heights observed in Figures 1.3 and 1.4.

In the control formulations, I and V, the amount of carbon dioxide diffused was less than theamount of carbon dioxide diffused using the new additives. Figures 1.7 and 1.8 were generatedusing the ideal gas law from data generated with the mass-loss/rate-of-rise apparatus. Diffusion

Figure 1.8 Carbon Dioxide Trapped versus Time: Back Formulation

Figure 1.7 Carbon Dioxide Trapped versus Time - Cushion Formulation



22

of carbon dioxide from cells as they form in the free rise reaction apparently keeps thereacting foam from reaching the same foam height when using these experimental additives.

Reactivity profiles for these three new additives are essentially the same. Data discussedlater in Section 9.3.1.1b, Tables 1.4 and 1.5, further supports the fact that there is no

noitalumrofnoihsucgnisunosirapmocytivitcaeresireerfximenihcaM4.1elbaT

lortnoC 35LB FSS.pxE FSS.pxE/35LB

noitalumroF I II III VI

)sdnoceS(maerC 7.4 6.4 9.4 9.4

)sdnoceS(puCfopoT 1.42 6.32 9.32 4.32

)sdnoceS(leGgnirtS 4.55 7.35 3.45 2.45

)sdnoceS(esiRlluF 4.001 8.301 3.501 6.501

oitaRleG:esiR 18.1 10.2 49.1 49.1

)mm(thgieHlaniF 4.872 6.972 8.082 6.272

)mm(thgieHruoH42 4.362 5.562 5.562 6.752

egaknirhS% 4.5 0.5 4.5 5.5

noitalumrofkcabgnisunosirapmocytivitcaeresireerfximenihcaM5.1elbaT

lortnoC 35LB FSS.pxE FSS.pxE/35LB

noitalumroF V IV IIV IIIV

)sdnoceS(maerC 8.4 5.4 8.4 8.4

)sdnoceS(puCfopoT 5.32 9.22 4.52 5.62

)sdnoceS(leGgnirtS 2.35 1.25 6.55 9.55

)sdnoceS(esiRlluF 5.901 7.801 9.111 2.111

oitaRleG/esiR 60.2 80.2 20.2 99.1

)mm(thgieHlaniF 6.372 7.572 3.862 1.072

)mm(thgieHruoH42 6.542 9.642 1.632 9.532

egaknirhS% 2.01 4.01 0.21 7.21

23

significant change in reactivity when comparing the control formulations to theexperimental formulations.

1.3.1.1b Machine Free Rise Reactivity

The machine mix free rise reactivity comparison of all formulations are shown inTables 1.4 and 1.5. This experimental data illustrates that the overall free rise foamreactivity for both the cushion and back formulations remains relatively the same forthe beginning of the reaction. The full rise reactivities in cushion formulations II, IIIand IV and back formulations VII and VIII start to deviate slightly from the controlreference formulations I and V.

Percent shrinkage remains fairly consistent within the cushion formulations, I, II, III andIV. Increased foam shrinkage was observed with back formulations VII and VIII. Thiscould be attributed to better cell wall drainage efficiency, providing more open foamand/or an overall increased carbon dioxide diffusion through the polymer network.

1.3.1.2 Foam Physical Properties

1.3.1.2a TPR Effect on Machine Run Moulded Foam FTC

When producing polyurethane, manufacturers use some type of mechanical crushing to opencells and insure the polyurethane article does not lose dimensional stability. Several techniquescan be used to provide the needed mechanical cell opening. Manufacturers will use TPR,which has been described in Section 1.2.3.2d, along with mechanical roller crushing. Someproducers will rely exclusively on the roller crushing and vacuum crushing techniques toprovide the mechanical cell opening required. In both cases reducing FTC values and improvingfoam openness is important for producing polyurethane articles that are dimensionally stable.

If TPR is carried out too soon during the polyurethane moulding cycle, the article willcollapse (blowout) as indicated in Figure 1.9. This is indicative of the foam beinginsufficiently cured or lacking enough green strength when TPR was applied. If TPR isconducted too late in the manufacturing process scalloping (concave surface areas on thefoam article) and tight foam (insufficient number of open cells within the foam articlethat causes the hot gas to be trapped and upon cooling forces the entire foam part toshrink) may also occur. When scalloping occurs the foam article must be repaired orscrapped. When tight foam occurs dimensional stability will suffer and there will be anegative impact on physical properties as denoted in Figure 1.10. The foam pad in Figure1.10 was produced at a 140 second TPR without crushing using formulation V.



24

Figure 1.9 Example of a Collapse/Blowout Moulded Foam

(Reproduced with permission from APCI)

Figure 1.10 Example of Tight Foam with Shrinkage

(Reproduced with permission from APCI)

25

To understand the benefits of these new additive technologies that provide reduced FTCvalues, a TPR range from 70 to 150 seconds was run for each of the formulations inTables 1.2 and 1.3. At a 70 second TPR all formulations suffered blowout since the foamwas not sufficiently cured and thus lacked green strength. At an 80 second TPR, noformulations evaluated experienced blowouts or collapse; however, slight distortionsand imperfections were evident on the foam surfaces to varying degrees of severity. Tocomplete the TPR window, TPR cycle times were continually ramped up in this fashionto determine the upper limit at which TPR could be applied for each formulation. Theupper limit is reached for a given formulation when the foam displays the obvious signsof scalloping and/or ‘dishing’ (concave surface areas of the foam). When this occurs thefoam is usually very tight and cannot be used as a functional part. Additionally, partswere produced without utilising TPR during the production cycle in order to comparethe difference in foam crushability when TPR is used.

1.3.1.2b Cushion Formulation Machine Evaluation Utilising TPR

Cushion control formulation I, which is listed in Table 1.2, was evaluated at a 90-100second TPR. Initial FTC values of 156 N/323 cm2 for a 90 second TPR and 165 N/323cm2 for an 100 second TPR were observed. These values were acceptable and producedfoam parts of good quality. The new additives in formulations II, III and IV producedmaximum initial FTC values of 160 N/323 cm2 at TPR of 90 to 100 seconds. Foamproduced at an 80 second TPR for the control formulation I and formulations II, III andIV containing the new additives resulted in minor problems with foam quality, i.e.,scalloping. At a 70 second TPR, control and experimental formulations failed because ofsevere blowout. Figures 1.11 and 1.12 show no significant difference of FTC for allformulations evaluated.

When TPR values were increased to 120 seconds for control formulation I, initial valuesincreased to 623 N/323 cm2 and scalloping or foam quality suffered (Figure 1.13).However, increasing the TPR time for cushion formulations II, III and IV to 120 seconds,produced maximum initial FTC values of 205 N/323 cm2 (Figure 1.13). Foam surfacequality was very good. Increasing the TPR to 140 seconds increased initial FTC values toa maximum of 543 N/323 cm2 and 534 N/323 cm2 for formulations II and III, respectively(Figure 1.14). Foam quality was still very good. Formulation IV, which utilises acombination of both surfactant and catalyst technologies, achieved a lower initial FTCvalue at a 140 second TPR of 191 N/323 cm2. This was only slightly higher than the FTCvalues of the control formulation at a 90 second TPR. Figures 1.11, 1.12, 1.13, and 1.14show the results of FTC through the entire TPR range applied in this study. These figuresclearly demonstrate that use of the new additives can reduce FTC values and maintaindimensional stability over the applied TPR range.



26

Figure 1.11 FTC at 90 TPR - Cushion Formulation


27





28

Figures 1.15, 1.16, and 1.17 further demonstrate the reduction of FTC values that canbe achieved when the new catalyst and surfactant technologies are utilised. These figuresclearly show a large difference achieved from plotting initial FTC to the final FTC cycleversus increasing TPR cycle times. Moreover, the lower delta FTC values obtained indicatean improved crush-out capability. The optimum TPR range for control formulation I is90-100 seconds, while the optimum range for formulations II and III is increased to 120

Figure 1.15 Initial FTC versus TPR Time - Cushion Formulation

Figure 1.16 Delta Difference of 1st to 3rd FTC versus TPR Time - Cushion Formulation

29

seconds. Formulation IV provided reduced FTC values up to a 140 second TPR.Formulation I did not provide acceptable foam surface quality above 100 second TPR.Formulations II, III and IV provided good surface quality and acceptable physicalproperties throughout the entire TPR range.

1.3.1.2c Back Formulation Machine Evaluation Utilising TPR

Back formulation evaluations were carried out in the same manner as the cushionformulation study. Because of lower solids and higher water content necessary to obtainspecific densities and physical properties, this system was more sensitive to processingand TPR range than the cushion formulation.

Back formulations V, VI, VII and VIII found in Table 1.3, were all run at TPR times of70-150 seconds and no TPR. Collapsed foam was encountered at TPR times of 70 secondsfor all formulations. FTC values for the back control formulation V had an initial FTCvalue of 138 N/323 cm2 at a 90 TPR and 160 N/323 cm2 at a 100 second TPR time.Formulations VI, VII and VIII produced a maximum initial FTC value at a 90 and 100second TPR of 138 N/323 cm2. Control formulation V and experimental formulationsVI, VII and VIII provided acceptable FTC values at 90 and 100 second TPR times (Figures18 and 19). No significant difference in FTC was realised with these formulations atthese TPR times. The foam quality at this TPR range was also very good. When TPRtimes were increased to 120 seconds for control formulation V, initial FTC increased to645 N/323 cm2. Experimental formulations VI, VII and VIII maintained acceptable initial

Figure 1.17 Delta Difference of Initial to Final FTC versus TPR Time - Cushion Formulation



30

FTC values of 334 N/323 cm2, 245 N/323 cm2 and 156 N/323 cm2, respectively, (Figure1.20). At a 120 second TPR, control formulation V exhibited scalloping while theexperimental formulations VI, VII and VIII continued to produce parts with good foamquality. Initial FTC values for the control formulation V increased to 1,045 N/323 cm2

at a 140 TPR and severe shrinkage was observed. FTC values for formulations VI andVII reached maximum values of 649 N/323 cm2, respectively, and 627 N/323 cm2 at a

Figure 1.18 FTC at 90 TPR - Back Formulation

Figure 1.19 FTC at 100 TPR - Back Formulation

31

140 second TPR (Figure 1.21). Foam produced for these two formulations at a 140second TPR still produced acceptable quality foam. Formulation VIII achieved lowerFTC values than any of the back formulations evaluated at a 140 second TPR range. Theinitial FTC value for formulation VIII was 405 N/323 cm2 with foam surface qualitymaintained (Figure 1.21). Figures 1.18, 1.19, 1.20, and 1.21 show the results of FTCvalues through the entire TPR range utilised in this study.

Figure 1.21 FTC at 140 TPR - Back Formulation.

Figure 1.20 FTC at 120 TPR - Back Formulation.



32

Figures 1.22, 1.23, and 1.24 further demonstrate the reduction of FTC values that canbe achieved in the back formulations when the new catalyst and experimental siliconesurfactant technologies are utilised. At a 90 and 100 second TPR, control formulation Vproduces good quality foam. When the TPR is increased above 100 seconds foam tightnessand surface quality issues arise with the control formulation. Formulations VI, VII andVIII yield good quality foam at all TPR cycle times used in this study. As in the cushionformulation, these lower FTC delta values demonstrate how these newly developedadditives enhance crush-out capabilities. Furthermore, this effectively illustrates that theTPR window can be extended as compared to the control formulation.

Figure 1.22 Initial FTC versus TPR Time - Back Formulation

Figure 1.23 Delta Difference of 1st to 3rd FTC versus TPR Time - Back Formulation

33

1.3.1.2d Cushion and Back Formulation Machine Evaluation Without TPR

Several polyurethane manufactures do not utilise TPR in their production process. Theformulations evaluated in this study were designed to utilise TPR when producingpolyurethane articles. However, indications of reduced force required to crush a padusing roller crushing and/or vacuum rupture can be understood from the data generatedwhen TPR is not applied to the process. The initial FTC values recorded for the controlcushion and back formulations I and V were 1,379 N/323 cm2 and 1,299 N/323 cm2,respectively. Using formulations II, III, VI and VII, FTC values were reduced to 1,152 N/323 cm2 and 1,090 N/323 cm2 for the cushion formulation and 1036 N/323 cm2 and1023 N/323 cm2 for the back formulation. Only minimal shrinkage was observed withthe surface quality being maintained. Using formulations IV and VIII, FTC values couldfurther be decreased to 1032 N/323 cm2 and 943 N/323 cm2, respectively. Again, foamsurface quality was good with only minimal shrinkage being observed. Figures 1.25 and1.26 show the FTC values obtained when no TPR was applied during this study.

1.3.1.2e Machine Physical Property Data Comparisons of Various TPR Times

Physical properties were evaluated for all the formulations found in Tables 1.2 and 1.3with TPR being applied at various times throughout the moulding cycle as previouslydiscussed. Additionally, physical properties were evaluated with no TPR being applied.The 90 second TPR time was chosen to be the minimum time to perform TPR for allformulations. Physical property pads were produced at this TPR time. Extended TPR

Figure 1.24 Delta Difference of Initial to Final FTC versus TPR Time - Back Formulation



34

times were achieved with the experimental formulations. Physical property pads forphysical testing were produced at these extended times.

Various TPR windows were identified for the formulations outlined in Tables 1.2 and 1.3.The control formulations I and V both had an upper TPR range of 100 seconds. At a 110second TPR both formulations I and V had scalloping and surface distortion issues (seeSection 1.3.1.2a). Thus the effective TPR range for formulations I and V is 90 to 100

Figure 1.26 FTC at No TPR - Back Formulation

Figure 1.25 FTC at No TPR - Cushion Formulation

35

seconds. This provides a very narrow process range when utilising TPR with these specificformulations. However, when using the Dabco BL-53 catalyst, or X-N1586/X-N1587experimental silicone surfactants, both cushion and back formulation TPR times can beextended well beyond the 100 second TPR upper limit of the control formulations I and V.

1.3.1.2f Physical Property Comparison at 90 Second TPR

Tables 1.6 and 1.7 provide the physical property comparison for all formulations I-VIIIat a 90 second TPR time. The data clearly demonstrates that physical properties aremaintained, and in several cases improved, compared to the control formulations. Forexample, airflow can be improved by as much as 20% as compared to both cushion andback control formulations, when using Dabco BL-53 and experimental silicone surfactants

noitalumrofnoihsuc:RPT09taseitreporplacisyhP6.1elbaT

ytreporPlacisyhP lortnoC

I

35LB

II

FSS.pxE

III

/35LBFSS.pxE

VI

)N(DLI GVA DS GVA DS GVA DS GVA DS

%52 571 2.2 181 5.1 581 1.3 481 8.2

%56 554 5.3 864 1.2 294 4.2 784 1.4

nruteR%52 451 5.1 951 8.1 061 5.1 361 8.1

)%(dnuobeRllaB 16 )5.1( 36 6.0 16 9.0 26 6.0

)MLS(wolfriA 83 8.5 74 2.4 05 1.5 75 2.6

m/gk(ytisneD 3) 44 8.0 44 4.0 34 8.0 44 7.0

)aPk(elisneT 971 4.8 381 1.3 781 4.5 081 2.3

)m/N(raeT 452 1.21 162 9.8 192 3.31 292 4.9

)%(noitagnolE 671 2.7 281 1.6 281 8.5 971 1.5

)%(teSteW 22 )2.1( 12 9.0 02 9.0 81 7.1

teSnoisserpmoC%05 5 3.0 5 2.0 5 3.0 4 1.0

degAdimuH%05teSnoisserpmoC

21 9.0 01 4.0 9 3.0 8 5.0

noitaiveDdradnatS:DSnoitcelfeddaolnoitatnedni:DLIetunimrepsertildradnats:MLS



36

X-N1586/X-N1587 in combination (formulations IV and VIII). Furthermore, Japanesewet set values [6] can be improved by 5 to 8% with the Dabco BL-53, and experimentalsilicone surfactants X-N1586 and X-N1587. Wet set values are improved nearly 20%when the two additives are utilised in combination as evidenced by Formulations IV andVIII. Equally important, humid aged compression set, tensile and tear physical propertiesdisplay a positive improvement trend as compared to the control formulations.

noitalumrofkcab:RPT09taseitreporplacisyhP7.1elbaT

seitreporPlacisyhP lortnoC

V

35LB

IV

FSS.pxE

IIV

/35LBFSS.pxE

IIIV


%52 521 0.3 431 6.3 441 8.1 551 7.1

%56 453 9.3 073 8.5 693 9.5 224 0.6

nruteR%52 801 7.1 611 9.2 221 7.1 821 1.3

)%(dnuobeRllaB 06 1.2 06 9.0 06 2.1 06 2.1

)MLS(wolfriA 73 5.2 84 2.1 94 1.4 05 8.3

m/gk(ytisneD 3) 53 6.0 63 2.1 63 2.1 63 2.1

)aPk(elisneT 731 3.6 541 1.5 551 7.9 361 1.01

)m/N(raeT 542 6.9 452 5.01 862 5.11 572 8.11

)%(noitagnolE 561 2.8 671 7.4 071 9.8 871 5.01

)%(teSteW 42 1.1 22 1.1 81 2.1 91 2.1

teSnoisserpmoC%05 8 5.0 8 2.0 4 2.0 4 5.0


41 2.1 01 3.0 01 4.0 9 5.0

noitaiveDdradnatS:DSnoitcelfeddaolnoitatnedni:DLIetunimrepsertildradnats:MLS

37

1.3.1.2g Physical Property Comparison at 130 and 150 Second TPR

Tables 1.8 and 1.9 illustrate the physical properties for selected formulations at TPRtimes of 130 and 150 seconds. Control physical properties were not evaluated at theseextended TPR times since the control formulations I and V had visual surface distortionsand severe scalloping. The data in Tables 1.8 and 1.9 demonstrate that the TPR windowcan be extended by using these newly developed additives without negative impact to thephysical properties. In fact, the data indicates that several of the physical properties forthe experimental formulations exceed the control formulation properties at the 90 secondTPR time. For example, airflow measurements are improved when utilising the extendedTPR times. Improvements are greater than 10% with the Dabco BL-53 catalyst andexperimental silicone surfactant combinations. Additional improvements are also observedwith wet set and 50% humid aged compression set values.

noitalumrofnoihsuc:RPT051dnaRPT031taseitreporplacisyhP8.1elbaT

seitreporPlacisyhP RPT031 RPT051

35LBII

FSS.pxEIII

FSS.pxE/35LBVI

)N(DLI GVA DS GVA DS GVA DS

%52 571 8.1 461 5.3 471 1.1

%56 434 5.2 754 0.1 764 5.4

nruteR%52 831 8.1 341 5.1 251 1.1

)%(dnuobeRllaB 16 5.0 26 5.0 16 4.0

)MLS(wolfriA 34 2.3 64 7.4 35 4.7

m/gk(ytisneD 3) 34 8.0 24 8.0 34 7.0

)aPk(elisneT 471 1.4 081 1.5 071 5.3

)m/N(raeT 552 4.7 472 5.01 972 1.21

)%(noitagnolE 081 9.3 081 5.3 371 3.3

)%(teSteW 12 5.1 91 3.1 91 3.0

teSnoisserpmoC%05 5 5.0 5 1.0 4 3.0


01 8.0 01 8.0 8 3.0



38

1.3.1.2h Physical Property Comparison with No TPR

Tables 1.10 and 1.11 illustrate the physical property performance of all formulations Ithrough VIII without utilising TPR. Given that the specific formulations were designedto be used with the TPR process, foam pads produced under these conditions were verytight upon demoulding. Mechanical crushing at 30 seconds after demould was requiredinstead of the one minute time frame; otherwise, irreversible shrinkage may have occurredprior to placing the foam through the roller crusher. The data in Tables 1.10 and 1.11indicate that utilising Dabco BL-53 catalyst and/or the experimental silicone surfactantsX-N1586 and X-N1587, have a positive influence on the physical properties even whenno TPR is applied. Moreover, if the properties of the new additives are compared to thecontrol formulations I and V at the 90 second TPR time (Tables 1.5 and 1.6), physicalproperties are not adversely affected. In some instances the properties are still maintainingtheir improvement over the control formulations. Control formulations run at no TPRexhibited surface quality distortions.

noitalumrofkcab:RPT051dnaRPT031taseitreporplacisyhP9.1elbaT

seitreporPlacisyhP RPT031 RPT051

35LB

IV

FSS.pxE

IIV

/35LBFSS.pxE

IIIV

/35LBFSS.pxE

IIIV


%52 111 1.2 541 2.4 651 7.0 551 2.1

%56 443 1.5 693 5.3 534 0.8 034 0.3

nruteR%52 69 2.2 321 2.2 231 6.0 921 1.2

)%(dnuobeRllaB 95 4.1 95 5.1 26 7.0 06 2.0

)MLS(wolfriA 94 5.5 54 3.2 94 8.1 25 1.2

m/gk(ytisneD 3) 53 8.0 63 0.1 63 8.0 63 2.0

)aPk(elisneT 631 1.5 841 4.5 251 9.7 741 1.6

)m/N(raeT 452 1.9 162 0.8 842 0.8 932 8.5

)%(noitagnolE 071 3.9 661 3.5 361 6.5 571 9.4

)%(teSteW 22 1.1 91 8.0 91 5.0 91 8.0

teSnoisserpmoC%05 9 2.0 5 3.0 4 2.0 4 6.0


01 5.0 01 6.0 9 5.0 9 8.0

39

noitalumrofnoihsuc:RPTontaseitreporplacisyhP01.1elbaT


I

35LB

II

FSS.pxE

III

/35LBFSS.pxE

VI


%52 541 5.0 051 2.1 751 5.0 751 1.2

%56 593 5.0 704 2.2 824 0.3 824 2.4

nruteR%52 721 5.0 131 1.1 631 5.0 831 1.2

)%(dnuobeRllaB 06 7.0 16 4.1 16 0.1 26 7.0

)MLS(wolfriA 43 6.2 24 5.2 74 2.6 94 3.1

m/gk(ytisneD 3) 24 8.0 24 5.0 14 6.0 14 6.0

)aPk(elisneT 951 3.5 061 4.5 571 5.3 961 5.4

)m/N(raeT 142 7.6 252 7.6 962 6.01 862 1.11

)%(noitagnolE 471 1.4 081 2.4 471 5.4 471 3.4

)%(teSteW 42 6.0 22 9.0 22 2.0 02 3.0

teSnoisserpmoC%05 6 2.0 5 1.0 5 2.0 5 2.0


21 2.0 01 1.0 01 7.0 9 1.0

1.3.1.2i Physical Property Review

All of the physical property data generated at the higher TPR cycle times illustrates that thewindow for TPR can be extended when using Dabco BL-53 catalyst and X-N1586, X-N1587 experimental silicone surfactants. Physical properties were not adversely affectedwhen compared to the control samples. By extending the effective TPR window, polyurethanemoulders who utilise this process will enjoy a greater latitude for their processing, providingmore freedom to troubleshoot difficult tooling and moulding operations. Additionally,polyurethane manufacturers that do not practice the TPR process can also benefit withimprovements realised for physical properties and dimensional stability, as illustrated bymany of the physical properties enhancements found in Tables 1.9 and 1.10.

Finally, the question remains whether the stability, especially shear stability, suffers from thesignificantly improved breathability and dimensional stability provided with these newadditives. Examination of cellular structures of the machine mixed free rise ‘chicken’ bucket



40

noitalumrofkcab:RPTontaseitreporplacisyhP11.1elbaT


V

35LB

IV

FSS.pxE

IIV

/35LBFSS.pxE

IIIV


%52 301 5.0 311 1.2 721 7.0 831 0.5

%56 792 0.5 033 7.7 553 6.0 683 4.8

nruteR%52 98 5.0 89 5.2 701 6.0 511 5.3

)%(dnuobeRllaB 85 7.0 16 1.0 16 )1.0 16 4.1

)MLS(wolfriA 93 1.1 64 1.3 74 1.5 64 8.3

m/gk(ytisneD 3) 53 5.0 53 8.0 53 8.0 53 8.0

)aPk(elisneT 321 9.4 531 8.8 231 7.5 041 9.9

)m/N(raeT 532 5.6 542 3.6 252 6.11 932 9.5

)%(noitagnolE 161 2.8 571 0.7 561 1.6 071 2.9

)%(teSteW 32 8.1 22 6.1 91 2.0 12 0.1

teSnoisserpmoC%05 9 0.1 9 4.0 6 4.0 5 3.0


41 8.1 11 5.0 11 6.0 01 6.0

foams demonstrated no evidence of elongated cells or side shear in the outer edges of the cup.Furthermore, when the TPR ramp study was being carried out it was determined that allformulations had similar lower limit TPR times of approximately 90 seconds. Essentially,this early TPR time would mimic a poorly sealed mould so that the shear forces becomeobvious in and around the gap or space created by opening the mould. The foam mixture, ifnot sufficiently stabilised or adequately cured, would be accompanied by a pressure reliefwhich leads to significant mechanical stress, causing collapse. The fact the Dabco BL-53catalyst, X-N1586 and X-N1587 experimental silicone surfactants did not have blowouts orshear collapses at the 90 second TPR illustrates shear stability is not compromised.

1.3.1.3 Scanning Electron Microscopy

A scanning electron microscopy study of moulded foam parts was carried out to understandwhat affect the additives might have on the morphology of the foam cell structure. It isclear from the SEM photomicrographs in Figures 1.27, 1.28 and 1.29 that there is no

41

Figure 1.27Formulation I, Control

(Reproduced with permissionfrom APCI)

Figure 1.28Formulation II, BL-53


Figure 1.29Formulation III, Exp. SSF




42

visual effect on the cell structure of foam produced with the new additives as compared tothe control. The additives did not appear to have an affect on the cell morphology of acrushed foam. However, it was not possible to determine differences in cell morphologyfrom an uncrushed part, since the parts were mechanically crushed to prevent normaldistortion from foam shrinkage. Formulations I, II and III were used for the SEM study.

1.3.2 Low Emission Dimensional Stability Additives

Another emerging need in the automobile market segment is for lower and ultimately noemissions of polyurethane foam additives into the environment. Current efforts revolvearound automotive passenger compartment air quality improvements by reduction ofthe emissions typically from the current migratory additives used to produce interiorfoam components. These emissions can cause a variety of problems in the final applicationof the foam, ranging from odour to vinyl staining and window fogging. The migratorynature of current additives is accelerated by sunlight and heat. Thermally induceddehydrohalogenation of the foam/covering interface is assisted by the migration of fugitivetertiary amine catalysts causing vinyl staining (colour change). Breakdown of the vinylclad covering subsequently releases other chemical compounds, such as plasticisers, whichthen contribute to window fogging. Consumers and automakers are becoming moreconscious of automotive interior air quality, interior vinyl degradation and film formation(windshield fogging) [12].

This section of the chapter reports on the development of non-fugitive gelling and blowingpolyurethane catalysts to address the aforementioned emission issues. These catalystschemically bind to the polyurethane foam matrix (contain active hydrogens) renderingthem incapable of migrating back out of the foam after the reaction is complete, whilestill producing a quality product. Equivalent or improved physical properties such asairflow, compression set and crushability, must be maintained as compared to the industrystandards used in TDI-, MDI-based systems, and TDI/MDI blend systems. Presented inthis section is work using these non-fugitive catalysts in a variety of flexible polyurethanefoam applications, including MDI and TDI automotive and flexible slabstockformulations. The results show that these new additives can give dimensional stabilityand processability to the foam, while providing non-fugitive and non-fogging benefits.Additionally, improvements to the physical properties of the polyurethane article whenusing the new additives, occurs.

Accordingly, these non-fugitive catalysts provide good to excellent foam stabilisation,cell regulation and cell wall drainage and are needed to enable polyurethane manufacturersto maintain their standard production goals. Two experimental non-fugitive catalysts,XF-N1085 and XF-O11006 were developed to provide open foam and promote

43

dimensional stability. During the development of non-fugitive catalysts, it was observedthat TDI formulations, in general, tend to be more closed cell with the use of thesecatalysts compared to MDI systems. Therefore, the technology was advanced toincorporate non-fugitivity and cell opening characteristics within a single catalyst.

XF-N1085 is a novel cell opening blowing, non-fugitive catalyst which affords all thebenefits of the well known Dabco BL-11, with the added advantage of cell opening andno foam emissions. XF-N1085 is not a drop in replacement for Dabco BL-11. For rapiddemoulding of TDI systems, it is recommended that XF-N1085 be used at 0.2 to 0.4php, with the optimum level at 0.25 to 0.35 pphp, in combination with an XF-O11006level at 0.6 to 0.9 pphp. XF-O11006 is a novel cell opening gelling non-fugitive catalystwhich gives all the benefits of the well known Dabco 33-LV, with the added advantage ofcell opening, low volatility and no foam emissions. XF-O11006 is not a drop inreplacement for Dabco 33-LV.

Several representative industry cushion and back formulations, listed in Tables 1.12-1.15, were utilised to compare multiple additives. Dabco NE200 was compared against

m/gk54~21.1elbaT 3 noitalumrofnoihsucllednoyLIDT

noitalumroFnoitacifitnedi

XI X IX IIX


848ElocrA 00.05 00.05 00.05 00.05

158ElocrA 00.05 00.05 00.05 00.05

retaW 43.2 43.2 43.2 43.2

VL33ocbaD 52.0 - - -

11-LBocbaD 01.0 - - -

0601ENocbaD - 80.0 35.0 -

002ENocbaD - 46.0 - -

5801N-FX - - 62.0 62.0

60011O-FX - - - 86.0

3405CDocbaD 57.0 57.0 57.0 57.0

FL-AOEDocbaD 67.1 67.1 67.1 67.1

)xedni001(IDT 47.23 47.23 47.23 47.23



44

m/gk53~31.1elbaT 3 noitalumrofkcabllednoyLIDT

noitacifitnedInoitalumroF IIIX VIX VX IVX


848ElocrA 00.07 00.07 00.07 00.07

158ElocrA 00.03 00.03 00.03 00.03

retaW 42.3 42.3 42.3 42.3

VL33ocbaD 52.0 - - -

11-LBocbaD 01.0 - - -

0601ENocbaD - 80.0 35.0 -

002ENocbaD - 46.0 - -

5801N-FX - - 62.0 62.0

60011O-FX - - - 86.0

3405CDocbaD 00.1 00.1 00.1 00.1

FL-AOEDocbaD 67.1 67.1 67.1 67.1

)xedni001(IDT 78.14 78.14 78.14 78.14

m/gk54~41.1elbaT 3 noitalumrofnoihsucwoDIDT

noitacifitnedinoitalumroF IIVX IIIVX XIX XX


036-CN 00.64 00.64 00.64 00.64

007-CN 00.45 00.45 00.45 00.45

retaW 43.2 43.2 43.2 43.2

VL33ocbaD 52.0 - - -

11-LBocbaD 01.0 - - -

0601ENocbaD - 22.0 07.0 -

002ENocbaD - 46.0 - -

5801N-FX - - 72.0 52.0

60011O-FX - - - 08.0

4615CDocbaD 0.1 0.1 0.1 0.1

9615CDocbaD 5.0 5.0 5.0 5.0

FL-AOEDocbaD 67.1 67.1 67.1 67.1

)xedni001(IDT 07.23 07.23 07.23 07.23

45

m/gk53~51.1elbaT 3 noitalumrofkcabwoDIDT

noitacifitnedinoitalumroF IXX IIXX IIIXX VIXX


036-CN 00.86 00.86 00.86 00.86

007-CN 00.23 00.23 00.23 00.23

retaW 42.3 42.3 42.3 42.3

VL33ocbaD 52.0 - - -

11-LBocbaD 01.0 - - -

0601ENocbaD - 72.0 85.0 -

002ENocbaD - 25.0 - -

5801N-FX - - 23.0 42.0

60011O-FX - - - 77.0

4615CDocbaD 2.0 2.0 2.0 2.0

9615CDocbaD 6.0 6.0 6.0 6.0

FL-AOED 67.1 67.1 67.1 67.1

)xedni001(IDT 29.14 29.14 29.14 29.14

Dabco BL-11 for equivalent blowing efficency. Dabco NE1060 was compared againstDabco 33LV for equivalent gelling efficiency. Required blow:gel ratios are listed in eachtable. Additionally, XF-N1085 and XF-O11006 were specifically developed asdimensional stability/cell opening additives and they are also compared here.

1.3.2.1 Reactivity

Tables 1.16 and 1.17 illustrate that the overall free rise foam reactivity for both the cushionand back formulations remains relatively the same for the entire foaming reaction. Thiswas achieved by balancing the experimental catalyst levels, blow:gel ratios, and matchingthem to the control Dabco BL-11 and Dabco 33LV levels in IX, XIII, XVII and XXI.



46

1.3.2.2 Flexible Moulded Foam Machine Physical Property Data

To understand the benefits of these new additive technologies that provide low volatility,no amine emissions, no fogging and, in some cases, cell opening, all TDI mouldedformulations were run on the Hi Tech machine. Several foams were produced for eachcatalyst combination and formulation to obtain physical property pads, FTC pads,shrinkage pads, and flow evaluations.

IDTgnisunosirapmocytivitcaeresireerfximenihcaM61.1elbaTsnoitalumrofnoihsuc

noitalumroF XI X IX IIX IIVX IIIVX XIX XX

maerC)sdnoceS(

9.4 8.4 8.4 5.4 5.3 8.3 5.4 9.4

puCfopoT)sdnoceS(

73 63 93 53 23 53 63 73

leGgnirtS)sdnoceS(

65 55 65 35 35 35 55 65

esiRlluF)sdnoceS(

38 18 48 87 17 57 18 38

IDTgnisunosirapmocytivitcaeresireerfximenihcaM71.1elbaTsnoitalumrofkcab

noitalumroF IIIX VIX VX IVX IXX IIXX IIIXX VIXX

maerC)sdnoceS(

8.3 8.3 8.3 4.3 2.4 2.3 2.3 2.3

puCfopoT)sdnoceS(

03 33 23 13 23 73 63 73

leGgnirtS)sdnoceS(

85 65 95 75 45 55 75 65

esiRlluF)sdnoceS(

28 58 49 19 57 68 58 88

47

1.3.2.2a Standard Physical Properties

Tables 1.18 to 1.21 provide the physical property comparison for the TDI flexible mouldedformulations IX-XXIV. The data clearly demonstrates that physical properties aremaintained, and in several cases improved, compared to the control formulations,depending on the formulation and the use of experimental cell-opening catalysts. Forexample, ILD properties were comparable or improved over the control in both low andhigh solids formulations, especially in the Dow polyol system. This is attributed to theslightly different polymer matrix resulting from non-fugitive catalysis which can producehigher load properties. In almost all cases, the support factor was comparable to thecontrol. Tensile values in the Dow system were matched to the control using XF-N1085in combination with Dabco NE1060, formulations XIX and XXIII. Dry and humidcompression sets (50% dry, 50% humid aged, and 75% humid aged) were, in manyformulations, slightly elevated, with the Dabco NE1060/XF-N1085 catalyst combination(XIX and XXIII) closest to the control. The compression set data illustrates that this

snoitalumrofnoihsucIDTllednoyLrofseitreporplacisyhP81.1elbaT

ytreporPlacisyhP XI X IX IIX


%52 081 - 071 - 281 - 181 -

%56 435 - 684 - 225 - 715 -

nruteR%52 751 - 051 - 061 - 851 -

)%(dnuobeRllaB 46 0.1 56 8.0 46 2.1 56 7.0

)MLS(wolfriA 84 7.5 54 5.8 26 3.11 04 8.2

m/gk(ytisneD 3) 24 3.1 44 5.0 34 0.1 44 0.1

)aPk(elisneT 341 1.01 631 7.66 361 8.7 671 5.6

)m/N(raeT 391 21 681 7 712 32 391 7

)%(noitagnolE 89 0.7 201 3.6 79 1.6 201 6.01

)%(teSteW 81 0.1 02 2.1 02 9.0 22 4.1

)%(teSnoisserpmoC%05 5 4.0 8 1.1 6 1.0 7 7.0

degAdimuH%05)%(teSnoisserpmoC

9 2.0 61 8.0 6 5.0 71 2.1

)%(siseretsyH 91 1.0 81 1.0 91 1.0 02 4.0

rotcaFtroppuS 0.3 - 9.2 - 9.2 - 9.2 -



48

snoitalumrofkcabIDTllednoyLrofseitreporplacisyhP91.1elbaT

ytreporPlacisyhP IIIX VIX VX IVX


%52 301 - 79 - 99 - 301 -

%56 513 - 782 - 982 - 013 -

nruteR%52 78 - 48 - 68 - 98 -

)%(dnuobeRllaB 26 8.0 46 65.0 46 1.1 26 7.0

)MLS(wolfriA 56 7.5 26 5.8 17 3.11 95 8.2

m/gk(ytisneD 3) 13 0.1 03 9.0 13 3.1 03 9.0

)aPk(elisneT 401 5.4 911 8.6 211 7.8 121 4.4

)m/N(raeT 342 91 422 12 952 61 952 91

)%(noitagnolE 39 7.5 79 6.7 001 2.4 001 4.7

)%(teSteW 72 9.0 52 9.0 62 0.1 62 9.0

)%(teSnoisserpmoC%05 6 7.0 8 4.0 7 35.0 8 4.0


61 6.0 02 2.1 02 1.1 71 0.1

)%(siseretsyH 12 1.0 81 2.0 02 2.0 02 1.0

rotcaFtroppuS 1.3 - 0.3 - 9.2 - 0.3 -

property is very system dependent when using non-fugitive catalysts. For performanceoptimisation, one may want to consider reformulation to produce the desired physicalproperties. The use of XF-N1085 and XF-O11006 will improve or maintain physicalproperties compared to the control for TDI systems. The Japanese wet set data wasdetermined to be comparable or, in the case of the Lyondell back formulation, slightlyimproved when compared to the control foam.

The use of cell opening non-fugitive catalysts was found to improve wet sets in theLyondell high solids formulation. In addition, the cell opening catalysts were also foundto decrease hysteresis and increase tensile strength in the back formulations using Lyondellpolyol compared to the control formulation.

49

snoitalumrofnoihsucIDTwoDrofseitreporplacisyhP02.1elbaT

ytreporPlacisyhP IIVX IIIVX XIX XX


%52 601 - 081 - 081 - 391 -

%56 123 - 315 - 815 - 155 -

nruteR%52 29 - 551 - 651 - 661 -

)%(dnuobeRllaB 65 6.0 55 9.1 65 8.0 75 8.0

)MLS(wolfriA 71 8.2 32 7.5 71 8.2 02 8.2

m/gk(ytisneD 3) 54 4.0 44 6.0 44 8.0 44 7.0

)aPk(elisneT 512 3.41 671 8.6 502 0.4 171 7.31

)m/N(raeT 012 11 012 61 532 81 122 41

)%(noitagnolE 001 0.7 39 6.6 98 9.6 301 8.5

)%(teSteW 61 2.0 12 4.1 02 6.0 71 6.0

)%(teSnoisserpmoC%05 6 6.0 9 1.0 8 3.0 8 3.0


31 5.0 71 8.0 02 6.2 81 7.0

)%(siseretsyH 12 1.0 12 1.0 12 6.0 22 1.0

rotcaFtroppuS 1.3 - 9.2 - 9.2 - 9.2 -

1.3.2.2b Dynamic Fatigue Results

Table 1.22 shows the dynamic fatigue test results for Lyondell and Dow cushion foamstested for 80,000 cycles and 60 minute recovery time. Results indicate that equivalentdynamic fatigue values are obtainable with use of the new non-fugitive and cell openingcatalysts, based on nominal test error of ± 3.

1.3.2.2c FTC Results

In both the Lyondell cushion and back formulations IX through XVI listed in Tables1.11 and 1.12, addition of XF-N1085 and XF-O11006 result in a reduction in FTCvalues (Tables 1.22 and 1.23) throughout the initial several repetitions of the crushingcycle illustrated in Figures 1.30 and 1.31. At the end of the ten complete repetitions, allthree experimental catalyst packages, resulted in the same final FTC value. Without the



50

snoitalumrofkcabIDTwoDrofseitreporplacisyhP12.1elbaT

ytreporPlacisyhP IXX IIXX IIIXX VIXX


%52 801 - 001 - 011 - 511 -

%56 513 - 992 - 313 - 113 -

nruteR%52 49 - 78 - 59 - 001 -

)%(dnuobeRllaB 36 8.0 06 7.1 06 9.1 36 8.0

)MLS(wolfriA 84 8.2 26 5.8 54 7.5 95 7.5

m/gk(ytisneD 3) 33 0.1 33 4.0 33 9.0 23 3.0

)aPk(elisneT 541 4.5 721 4.7 141 8.5 931 8.4

)m/N(raeT 942 9 242 4 742 61 362 32

)%(noitagnolE 601 0.11 89 0.8 901 7.21 501 9.4

)%(teSteW 12 9.0 61 3.1 22 7.1 91 9.1

)%(teSnoisserpmoC%05 5 4.0 7 6.0 8 6.0 7 4.0


51 9.0 13 1.2 12 2.0 62 8.1

)%(siseretsyH 91 1.0 81 2.0 02 1.0 91 2.0

rotcaFtroppuS 9.2 - 0.3 - 9.2 - 7.2 -

)selcyc000,08(]5[59-4753DMTSAeugitafcimanyD22.1elbaT

%04lanigirOnoitcelfeD)N(ecroF

%04laniFnoitcelfeD)N(ecroF

)%(egnahC%

XI 282 442 4.31

X 852 702 9.91

IX 372 222 6.81

IIX 272 832 7.21

IIVX 203 822 4.42

IIIVX 162 602 9.02

XIX 462 002 3.42

XX 082 792 5.92

51

metsysllednoyLninoihsucIDTrofseulavCTF32.1elbaT

)N(sdiloShgiHllednoyL

ycneuqerF eulaVCTF eulaVCTF eulaVCTF eulaVCTF

XI X IX IIX

1 767 329 837 896

2 653 045 543 943

3 242 613 432 942

4 891 632 981 502

5 761 391 261 871

6 941 761 061 171

7 041 151 851 761

8 831 541 651 061

9 831 041 851 061

01 331 721 651 061

Figure 1.30 FTC Graph for TDI Cushion in Lyondell System(Ref. Table 1.22)



52

use of these cell opening catalysts, the Dabco NE1060/Dabco NE200 non-fugitive catalystpackage tightens the foam resulting in slightly elevated values during several initial FTCfootplate depressions.

In the cushion Dow formulations XVII through XX, the XF-N1085 and XF-O11006show similar trends to the Lyondell formulations with reduced FTC (Figure 1.32 andTable 1.24), thereby producing a more dimensionally stable foam. In contrast, for thelow solids back formulations XXI through XXIV, all catalyst combinations exhibitedno significant difference to the control catalyst (Figure 1.33 and Table 1.25). Inherenttightness in this particular back formulation did not allow the same degree of catalystdifferentiation to occur in these foams. Adjustment to the catalyst loading levels wouldbe required to demonstrate the same reduced FTC performance as previously describedin other formulations.

Figures 1.34 and 1.35 further demonstrate the reduction of FTC values that can beachieved with these new catalyst technologies. The graphs indicate the results achievedby plotting the delta of the initial and third FTC value for every formulation. Moreover,lower delta values illustrate an improved crush-out capability.

Figure 1.31 FTC Graph for TDI Back in Lyondell System(Reference Table 1.23)

53

Figure 1.32 FTC Graph of TDI Cushion in Dow System(Reference Table 1.24)

metsysllednoyLnikcabIDTrofseulavCTF42.1elbaT

)N(sdiloSwoLllednoyL


IIIX VIX VX IVX

1 038 858 896 386

2 474 855 724 904

3 492 063 152 542

4 691 632 961 561

5 151 281 921 921

6 811 631 701 501

7 001 611 19 19

8 58 69 08 08

9 67 58 67 17

01 37 08 96 46



54

Figure 1.33 FTC Graphs for TDI Back in Dow System(Reference Table 1.25)

metsyswoDninoihsucIDTrofseulavCTF52.1elbaT

)N(sdiloShgiHwoD


IIVX IIIVX XIX XX

1 4101 4321 959 348

2 456 588 026 615

3 174 146 454 383

4 963 305 063 903

5 413 414 503 762

6 472 743 082 042

7 452 892 652 522

8 632 742 922 112

9 612 942 702 691

01 702 022 781 191

55

Figure 1.34 Delta Difference Graph of First and Third FTC Values in Lyondell System

metsyswoDnikcabIDTrofseulavCTF62.1elbaT

)N(sdiloSwoLwoD


IXX IIXX IIIXX VIXX

1 639 188 109 509

2 005 505 615 005

3 903 113 113 892

4 722 132 112 902

5 171 871 851 761

6 831 041 721 831

7 611 111 701 811

8 001 69 39 501

9 78 28 28 89

01 08 96 67 98



56

1.3.2.2d Pail Test Results

Visual examination of the foam slice’s cellular structure showed no evidence ofelongated cells or side shear collapse near the outer edges. The appearance of theseelongated cells or side shear collapse is a phenomenon created by the frictional dragof the plastic pail (see Section 1.2.3.2e) on the rising foam. Each slice was measuredfor maximum height, normalised by weight and displayed in Figure 1.36 forcomparison of final foam recession to the control. It is evident from the data that nosignificant difference exists in bulk stability when using any of the cell opening ornon-fugitive catalysts.

1.3.2.2e Maze Mould Test Results

Data in the Lyondell system, formulations IX through XVI, generated using thepreviously described maze flow mould is summarised in Tables 1.27 and 1.28 and

Figure 1.35 Delta Difference Graph of First and Third FTC Values in Dow System

57

IIX-XIsnoitalumrofnoihsucIDTllednoyLrofataddluomwolfezaM72.1elbaT

reifitnedInoitalumroF )g(.twdaP )g(etaght5tsapwolF .twdaplatoTfo%

etaGht5MUMINIM

XI 504 03 0.8

X 404 03 0.8

IX 504 53 5.9

IIX 904 03 9.7

etaGht5NAIDEM

XI 474 101 1.72

X 274 401 3.82

IX 874 901 5.92

IIX 974 501 1.82

etaGht5MUMIXAM

XI 425 051 1.04

X 625 061 7.34

IX 135 161 5.34

IIX 825 551 6.14

Figure 1.36 Graph Comparing 24 Hour Pail Height



58

shown graphically in Figure 1.37. The graph utilises a xy scatter plot with linearregression trendlines and R2 values (coefficient of determination) reported for eachformulation group. It compares estimated and actual y-values and ranges in valuefrom zero to one. Since these values are very close to one, there is a near perfectcorrelation in the samples. However, if the coefficient of determination approacheszero, the regression equation is not helpful in predicting a y-value. The data trendssuggest that increasing part weight, regardless of formulation type, shows no significantpenalty in flow compared to the control formulations IX and XIII.

Similar maze flow mould results are also obtained with the Dow cushion and backformulations summarised in Tables 1.28 and 1.29 and shown graphically in Figure 1.38.In these Dow formulations, XVII through XXIV, the data suggests slightly improvedlinear regression trendlines and R2 values. Again, no significant penalty in flow is observedas compared to the control formulations XVII and XXI.

IVX-IIIXsnoitalumrofkcabIDTllednoyLrofataddluomwolfezaM82.1elbaT


etaGht5MUMINIM

IIIX 916 29 5.71

VIX 716 201 8.91

VX 516 801 3.12

IVX 616 701 0.12

etaGht5NAIDEM

IIIX 766 341 3.72

VIX 666 351 8.92

VX 566 061 7.13

IVX 866 651 5.03

etaGht5MUMIXAM

IIIX 817 391 8.63

VIX 517 202 4.93

VX 617 602 4.04

IVX 817 802 8.04

59

Figure 1.37 Maze Flow Mould Graph for Formulations IX-XVI

XX-IIVXsnoitalumrofnoihsucIDTwoDrofataddluomwolfezaM92.1elbaT


etaGht5MUMINIM

IIVX 404 04 0.11

IIIVX 304 63 8.9

XIX 993 43 3.9

XX 404 52 6.6

etaGht5NAIDEM

IIVX 754 39 6.52

IIIVX 754 29 2.52

XIX 854 001 9.72

XX 354 96 0.81

etaGht5MUMIXAM

IIVX 905 941 4.14

IIIVX 805 041 0.83

XIX 805 941 5.14

XX 315 631 1.63



60

VIXX-IXXsnoitalumrofkcabIDTwoDrofataddluomwolfezaM03.1elbaT


etaGht5MUMINIM

IXX 865 56 9.21

IIXX 665 07 1.41

IIIXX 565 46 8.21

VIXX 965 38 1.71

etaGht5NAIDEM

IXX 716 221 7.42

IIXX 616 021 2.42

IIIXX 416 211 3.22

VIXX 816 231 2.72

etaGht5MUMIXAM

IXX 866 861 6.33

IIXX 566 271 9.43

IIIXX 666 461 7.23

VIXX 076 381 6.73

Figure 1.38 Maze Flow Mould Graph for Formulations XVII-XXIV

61

1.3.2.2f Shrinkage Test Results

Shrinkage results displayed in Figures 1.39 and 1.40 were generated with the shrinkageapparatus previously described in the experimental section. Overall, the cavity shrinkage ismuch greater as compared to the lid shrinkage, evidenced in both Lyondell and Dow polyolsystems. The use of Dabco NE1060/Dabco NE200 produced foams with more shrinkagethan the control catalysts. However, utilising non-fugitive cell opening catalysts, XF-N1085and XF-O11006 resulted in shrinkage comparable to the control and in some cases even less.As expected, less shrinkage was observed in both Dow and Lyondell cushion formulationsdue to the increased solids loading as compared to their respective back formulations.

Figure 1.39 Uncrushed Foam Shrinkage for Lyondell System

Figure 1.40 Uncrushed Foam Shrinkage for Dow System



62

9.3.2.2g Fogging Test Results

Several machine foam samples were tested in accordance with fogging test method SAEJ1756-94 [9]. Results are listed in Table 1.31. The results obtained in the Lyondell polyolsystem showed improved percentage reflectance over the control indicating reducedemissions from the foam. Comparable results were obtained in all the Dow polyol systemsamples regardless of catalyst use.

stlusergniggoF13.1elbaT

reifitnedImetsyS ytivitcelfeRfo%

XI 78

X 79

IX 89

IIX 69

IIVX 39

IIIVX 29

XIX 39

XX 39

1.3.2.2h Headspace Analysis and Vinyl Staining Test Results

Headspace analysis in series with gas chromatography and mass spectrometry hasindicated no emissions from the experimental non-fugitive catalysts mentioned in thispaper. Emissions from the fugitive control catalysts were measured and identified. Theseresults support the fact that these experimental catalysts will result in no VOC emissionswhich will ultimately yield no amine fogging and no vinyl staining from the finalpolyurethane foam article. Additional volatility measurements have indicated that theseexperimental catalysts volatilise 3 to 10 times less than the current industry standards(Dabco 33LV and Dabco BL-11) in the temperature range of 0 °C to 150 °C.

Vinyl staining tests using commercial formulations and various grades and types of vinylhave indicated that the experimental catalysts mentioned in this chapter do not contributeto vinyl staining. Compared to industry standard amine catalysts, use of these experimentalnon-fugitive catalysts will result in significant reduction (ΔE of less than 2) in stainingmeasurements [12]. DE is the numerical total colour difference, using lightness andchromaticity factors, between a sample and a known colour standard.

63

1.3.2.2i TDI Flexible Moulded Foam Review

The most commonly used catalyst packages for all water blown TDI flexible foamproduction are blends of Dabco 33LV and Dabco BL-11 at a typical ratio of approximately3:1. These new non-fugitive catalysts, Dabco NE1060, Dabco NE200, XF-N1085, andXF-O11006 are recommended for TDI formulations at various gel:blow ratios dependingupon the selectivity desired. Ratios from 2:1 to 8:1 can be used to tailor the reactionprofile to the preferred selectivity for optimisation of physical properties and adaptationto various processing conditions. The physical properties obtained with these new non-fugitive catalysts are similar to the physical properties generated with standard industrycatalysts. However, slight formulation adjustments may be necessary to improve dry andwet compression set values in certain instances. Improved crush-out capability can alsobe achieved when incorporating the new cell-opening non-fugitive catalysts, XF-N1085and XF-O11006. Complementary methods of emission analysis, headspace and fogging,both suggest reduced emissions from the foam with these new non-fugitive catalysts.

1.4 MDI Flexible Moulded Foam Additives

The current trend toward lower foam densities, faster demould times, and the increaseduse of complex metal and plastic insert frames in the construction of automobile interiorcomponents has increased the difficulty of controlling the numerous process and chemicalvariations a polyurethane manufacturer encounters on a daily basis. To successfullyproduce moulded high resilience polyurethane foam, the manufacturer must maintain acritical balance between foam over stabilisation, resulting in foam shrinkage, and understabilisation, resulting in internal defects, such as basal cell formation and shear collapse.This often difficult to achieve balance is referred to as processing latitude. Variations oneither side of the processing latitude can result in costly increases for the producer inscrap and repair rates. To help minimise production losses, polyurethane manufacturersare essentially forced to over stabilise their foam formulations on a regular basis.Subsequently, the closed cells which normally result from this over stabilisation are openedvia standard mechanical crushing techniques [13].

New surfactant technologies have been developed which promote improved cell walldrainage and contribute to the final level of cell openness of the foam product withoutcausing any of the aforementioned negative attributes typical to a foam process. Due tothis unique cell opening action of these new additive technologies, the foam crushingportion of the process can be greatly simplified or potentially eliminated.

Figure 1.41 shows a variety of flexible moulded Dabco surfactants which can be utilisedin TDI and/or MDI formulations. Using Figure 1.41, proper selection of the surfactantcan be made, enabling polyurethane formulators to expand their processing latitude, cellopenness, and bulk stability of the foam article in their critical formulations.



64

1.4.1 Dimensional Stability Additives for MDI

An increasing North American trend in the moulded polyurethane foam industry todayis the use of MDI. While the advantages and disadvantages of MDI versus TDI are stillbeing debated, the fact remains that most polyurethane producers are using MDI in atleast part of their daily flexible foam production. The surfactant requirements for MDI-based formulations can vary dramatically, dependent upon the type of MDI and polyolsbeing used. Most MDI formulations are inherently quite stable; therefore minimalstabilising contributions are necessary from the surfactant. Conventional TDI highresilience moulded surfactants are far too potent for MDI formulations. For example,utilisation of Dabco DC5043 or Dabco DC5164 in an MDI seating formulation mayresult in over stabilisation and significantly contribute to the shrinkage of the foam article.

To demonstrate the effectiveness of these dimensional stabilising additives, Dabco DC2517and Dabco DC5244, several moulded high resilience foam experiments were carried outusing the formulation illustrated in Table 1.32, to compare with two non-cell openingsurfactants Dabco DC2585 and an additional surfactant (control). Much of the workdescribed next utilises proprietary formulations and only a general description of someof the components can be disclosed.

Figure 1.41 Flexible Moulded Surfactant/Property Relationships

65

1.4.1.1 Reactivity - Handmix Rate of Rise

Data generated in the rate-of-rise comparison tests are summarised in Table 1.33. Foamswere run at an optimum index of 100 and all experiments were duplicated. Each mixedformulation was poured into a ‘chicken’ bucket equipped with a thermocouple positionedat the centre of the bucket resting on a Mettler PM 30 balance. The centred height of therising foam was recorded in millimetres every second using a DAPS (data acquisitionand plotting system) QA Model #2500 rate-of-rise aparatus. Reactivity profiles for thesenew additives are essentially the same. Data discussed next further supports the fact thatthere is no significant impact on reactivity or physical properties when comparing thecontrol to the Dabco DC2517 and Dabco DC5258 formulations.

noitalumrofdedluomelbixelfIDM23.1elbaT

stnenopmoC phpp

dnelbloirt/loid,loyloprehteyloP 001

retaW 5.3

enimalonaporposiiD 0.2

VL33ocbaD 3.0

11-LBocbaD 52.0

61-BocbaD 3.0

stnatcafruS 0.1

)OCN%62,ytilanoitcnuf3.2(remyloperpIDM xednI001

nosirapmocytivitcaeresirfoetarximdnaH33.1elbaT

eliforPnoitalumroF

)sdnoces(1puCfopoT 4.21

)sdnoces(2puCfopoT 1.84

)sdnoces(leGgnirtS 4.35

)sdnoces(esiRlluF 8.101

)mm(thgieHesiRlluF 1.593



66

1.4.1.2 Standard Physical Properties

To understand the benefits of these new additive technologies that provide low volatilityand cell opening, all MDI moulded formulations were run in handmix foam to demonstrateperformance. Several foams were produced for each surfactant formulation to obtainphysical property pads, FTC pads, shrinkage pads, and flow evaluations.

Table 1.34 illustrates moulding foams at a density of 47-48 kg/m3; similar physicalproperties are obtained for ILD, resilience, airflow, 50% dry sets and extensive properties.However, slight nominal improvements in 50% humid aged compression sets (HACS)and wet sets are achieved with Dabco DC2517 and Dabco DC5258.

snoitalumrofdedluomIDMrofseitreporplacisyhP43.1elbaT

ytreporPlacisyhP lortnoC 7152CD 5852CD 8525CD


%52 203 4.3 592 5.4 292 2.5 003 8.1

%56 467 9.9 267 2.8 957 2.7 357 3.1

nruteR%52 322 5.3 312 7.4 212 9.3 912 0.1

)%(dnuobeRllaB 3.44 4.0 7.34 7.0 7.34 7.0 0.44 5.0

)MLS(wolfriA 5.22 1.2 3.42 1.9 3.32 0.5 1.82 4.01

)aPk(elisneT 561 1.5 171 8.9 561 7.11 761 6.8

)M/N(raeT 503 5.11 303 7.51 903 2.11 033 8.91

)%(noitagnolE 59 4.7 49 5.5 39 8.6 79 5.4

)%(teSteW 5.81 3.1 6.61 6.2 1.91 2.0 4.61 6.0

teSnoisserpmoC%05 1.01 1.1 2.01 1.1 4.01 9.0 1.9 8.0

SCAH%05 0.82 2.0 6.62 2.1 7.52 7.0 3.42 6.1

1.4.1.3 FTC Results

The FTC graph shown in Figure 1.41 illustrates an approximate 14% decrease in initialFTC values and continues throughout the remaining FTC cycles. This improvement inFTC represents foam articles that will shrink less immediately after demoulding whencooling of the trapped gas begins, allowing more time to complete the post crushing

67

procedure or potentially eliminating the procedure all together. Several foams wereprepared with each surfactant and allowed to cool for 24 hours in an uncrushed state, toconfirm the relationship between FTC and foam shrinkage. Foams prepared with DabcoDC2517 and Dabco DC5258 showed significantly less shrinkage, approximately 10-15% less, as compared to the control and Dabco DC2585. This is readily comparable tothe differences in FTC values shown in Figure 1.42.

1.4.2 Low Emissions Dimensional Stability Additives in MDI

The control catalyst combination for the MDI seating is a standard commercial catalystpackage. The balanced non-fugitive catalyst package Dabco NE1060/Dabco NE200 andthe low emission catalyst package Dabco NE1060/Dabco BL-11 use levels were set inorder to match performance to the commercial catalyst control package for the MDIformulations. The silicone surfactant control was Dabco 2525. Cell opening non-fugitivecatalysts were not needed for this application based on the reported FTC values discussedlater in this chapter.

An MDI formulation, with a density of 40 kg/m3 and an MDI formulation, with a densityof 55 kg/m3 were used in the MDI automotive study with the formulations shown inTable 1.35. For rapid demoulding of non-fugitive MDI systems, it is recommended thatDabco NE1060 be used at 0.5 to 2.0 pphp, with the optimum level at 1.0 to 1.5 pphpwith 0.3 to 0.7 pphp of Dabco NE200. For low emission systems the authors recommendDabco NE1060 be used at 1.0 to 2.0 pphp, with the optimum level at 1.0 to1.6 pphpwith 0.03 to 0.1 pphp of Dabco BL-11.

Figure 1.42 FTC Graphs for MDI Dimensional Stability Additives



68

1.4.2.1 Low Emissions Moulded Foam Machine Standard Physical Properties

For complete comparison of these new additive technologies that provide low volatility, noamine emissions and no fogging, all moulded formulations were run on a Krauss-Maffeimachine. Several foams were produced for each catalyst combination and formulation toobtain physical property pads and FTC pads, shown in Tables 1.36 to 1.41.

Tables 1.36 and 1.37 provide the physical property comparison for the ~46 kg/m3 MDIflexible moulded formulations XXV-XXXIV. The data clearly demonstrates that physicalproperties are maintained, and in several cases improved, compared to the controlformulations, depending on the formulation and the use of experimental cell-openingcatalysts. For example, the data in Table 1.36 illustrates that at an index of 90, allphysical properties are comparable to the control formulation. When the index is increasedto 105, ball rebound and airflow are improved, with a slight decrease in indentationforce deflection (IFD) properties when the non-fugitive catalyst (XXXIV) is compared tothe control formulation (XXVI). When Dabco NE1060/Dabco NE200 is used (XXIX),the physical properties are equal to the control at an index of 90. Improved Japanese wetsets (see Section 1.3.1.2f) and slightly decreased IFD values are observed at an index of105 compared to the control formulation (XXX).

snoitalumrofIDMdedluomelbixelF53.1elbaT

noitalumroFnoitacifitnedI

-XXV

-XXIV

-XXIIV

-XXIIIV

-XXXI

-XXX

-XXIX

-XXIIX

-XXIIIX

-XXVIX

stnenopmoC phpp phpp phpp phpp phpp phpp phpp phpp phpp phpp

CloyloP 001 001 001 001 001 001 001 001 001 001

DloyloP 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1

retaW 5.3 5.3 5.3 5.3 5.3 5.3 5.3 5.3 5.3 5.3

tsylataClortnoCegakcaP

07.0 - - - - 07.0 - - - -

11-LBocbaD - 80.0 - 50.0 - - 80.0 - 50.0 -

51tacyloP - 03.0 03.0 - - - 03.0 03.0 - -

0601ENocbaD - 00.1 55.0 05.1 02.1 - 00.1 55.0 05.1 02.1

002ENocbaD - - 06.0 - 04.0 - - 06.0 - 04.0

5252CDocbaD 00.1 00.1 00.1 00.1 00.1 00.1 00.1 00.1 00.1 00.1

reknilssorC 06.0 06.0 06.0 06.0 06.0 06.0 06.0 06.0 06.0 06.0

IDMdeifidoMxednI

09 09 09 09 09 501 501 501 501 501

69

m/gk64rofseitreporplacisyhP63.1elbaT 3 09foxedninatasnoitalumrofdedluomIDM

ytreporPlacisyhP VXX IVXX IIVXX IIIVXX XIXX

)N(DLI GVA DS GVA DS GVA DS GVA DS GVA DS

%52 951 - 951 - 941 - 551 - 551 -

%56 954 - 054 - 634 - 054 - 844 -

nruteR%52 721 - 621 - 911 - 321 - 321 -

)%(dnuobeRllaB 15 6.0 15 6.0 15 6.0 15 6.0 15 2.1

)MLS(wolfriA 4.22 6.7 4.42 5.6 92 4.7 43 9.7 71 0.5

m/gk(ytisneD 3) 64 4.3 44 0.1 34 4.2 64 7.3 54 9.0

)aPk(elisneT 921 6.01 892 4.41 521 7.6 331 9.5 721 9.9

)m/N(raeT 212 21 102 61 791 21 712 61 502 02

)%(noitagnolE 411 9.9 811 1.6 211 6.4 021 3.1 811 3.9

)%(teSteW 31 6.0 61 2.1 01 5.0 31 4.0 31 3.1

noisserpmoC%05)%(teS

9 6.1 01 0.1 9 6.1 01 5.0 01 8.0


31 5.0 31 0.2 31 5.1 21 3.1 41 0.1


ytreporPlacisyhP XXX IXXX IIXXX IIIXXX VIXXX


%52 572 - 062 - 462 - 162 - 152 -

%56 867 - 137 - 937 - 837 - 617 -

nruteR%52 902 - 791 - 102 - 891 - 091 -

)%(dnuobeRllaB 34 6.0 54 2.1 54 0.0 54 0.1 74 2.1

)MLS(wolfriA 8.04 5.8 4.13 1.7 83 2.72 33 9.9 64 5.8

m/gk(ytisneD 3) 64 1.1 74 4.4 54 6.0 64 2.3 44 3.1

)aPk(elisneT 671 8.9 904 2.91 371 8.6 861 3.6 961 7.11

m/N(raeT 362 8 252 4 652 21 972 61 442 21

)%(noitagnolE 501 2.4 901 1.4 201 7.6 501 3.4 301 2.6

)%(teSteW 51 8.0 51 4.0 11 2.0 9 7.0 71 2.0


8 6.0 9 3.1 8 3.0 01 9.0 9 0.1


31 0.2 31 8.1 21 0.1 41 7.1 41 0.1



70


ytreporPlacisyhP VXX IVXX IIVXX IIIVXX XIXX


%52 532 - 832 - 812 - 062 - 532 -

%56 736 - 636 - 795 - 137 - 146 -

nruteR%52 881 - 091 - 471 - 791 - 781 -

)%(dnuobeRllaB 64 6.0 44 0.1 64 6.0 54 2.1 24 6.0

)MLS(wolfriA 12 0.4 53 9.11 53 4.5 4.13 1.7 72 5.21

m/gk(ytisneD 3) 35 4.3 05 6.3 05 5.3 74 4.4 25 9.1

)aPk(elisneT 841 7.42 461 0.8 161 2.3 941 2.91 261 0.11

m/N(raeT 702 21 022 21 791 4 252 4 502 21

)%(noitagnolE 711 0.6 611 9.7 021 7.1 901 1.4 811 1.4

)%(teSteW 11 4.1 31 9.0 51 3.1 51 4.0 01 3.0


8 9.0 7 5.1 9 0.1 9 3.1 9 4.0


11 5.1 41 1.2 31 3.1 31 8.1 21 9.0


ytreporPlacisyhP XXX IXXX IIXXX IIIXXX VIXXX


%52 893 - 083 - 883 - 383 - 183 -

%56 6601 - 9201 - 9301 - 2401 - 2301 -

nruteR%52 303 - 882 - 592 - 982 - 982 -

)%(dnuobeRllaB 64 0.1 34 6.0 44 6.0 34 6.0 54 2.1

)MLS(wolfriA 45 3.71 9.26 8.12 84 0.51 76 3.9 84 5.4

m/gk(ytisneD 3) 25 7.3 25 2.1 45 7.0 45 4.0 55 5.0

)aPk(elisneT 712 0.71 812 0.7 702 9.21 712 6.6 322 5.01

)m/N(raeT 992 21 382 42 703 42 992 61 513 61

)%(noitagnolE 501 2.4 801 4.1 201 1.1 111 1.4 901 1.2

)%(teSteW 41 4.1 21 7.0 31 7.0 11 6.0 51 9.0


6 1.1 8 8.0 8 6.0 9 1.1 9 4.0


41 9.1 41 8.0 31 9.0 21 2.1 41 8.0

71

Tables 1.38 and 1.39 provide the physical property comparison for the ~55 kg/m3 MDI flexiblemoulded formulations XXV-XXXIV. The data clearly demonstrates that physical propertiesare maintained, and in several cases improved, compared to the control formulations, dependingon the formulation. For example, the data in Table 1.38 illustrates that all physical propertiesare matched to the control formulation (XXV) at an index of 105 with the exception of aslightly lower ball rebound for the full non-fugitive catalyst package. At an index of 105, allphysical properties for the experimental non-fugitive catalysts formulations were determinedto be similar to the control (XXX). When Dabco NE1060/BL-11 is used as a low emissioncatalyst package (XXXIII), all physical properties at the index of 105 are comparable to thecontrol foam. At an index of 105 (XXXIII), airflow, elongation and Japanese wet sets andHACS are improved over the control (XXX) and 50% dry compression sets are slightly elevated.

1.4.2.2 FTC

Tables 1.40 and 1.41 show numerical values which are shown graphically in Figures1.43 and 1.44. FTC values are dramatically improved versus the control (XXV), shown

09foxedninatasnoitalumrofIDMrofseulavCTF04.1elbaT

noitalumroF mc323/N(CTF 2)m/gk64 3

mc323/N(CTF 2)m/gk55 3

VXX 25 67

IVXX 41 42

IIVXX 53 36

IIIVXX 8 31

XIXX 4 21

501foxedninatasnoitalumrofIDMrofseulavCTF14.1elbaT

noitalumroF mc323/N(CTF 2)m/gk64 3

mc323/N(CTF 2)m/gk55 3

XXX 6.2 5.4

IXXX 3.1 9.1

IIXXX 9.1 2.3

IIIXXX 9.1 2.3

VIXXX 3.1 9.1



72

in Figure 1.43, for the foams with an index of 90. FTC values for all of the formulationsat an index of 105 are very similar to the control (error is ± 6.4 N/323cm2), as illustratedin Figure 1.44.

Figure 1.43 FTC at an Index of 90 for MDI Formulations

Figure 1.44 FTC at an Index of 105 for MDI Formulations

73

1.4.2.3 MDI Flexible Moulded Foam Review

The non-fugitive catalyst package Dabco NE1060/Dabco NE200 and the low emissioncatalyst combination Dabco NE1060/Dabco BL-11 closely match the entire reactivityprofile demonstrated by the standard control catalysts typically used in MDI flexiblefoam. Additionally, the Dabco NE1060/Dabco NE200 combination demonstrates a higherselectivity towards the blowing reaction as compared to a standard 2.5:1 ratio of Dabco33LV to Dabco BL-11. This higher blowing selectivity can help improve cell opening andproduce a more dimensionally stable polyurethane article. These non-fugitive catalystsalso provide a reduction in the force necessary to crush freshly demoulded foam withoutadversely affecting the remaining physical properties of the foam in MDI formulationsover a range of formulation indices.

1.5 TDI Flexible Slabstock Low Emission Additives

Handmix rate of rise comparison tests were performed using the same equipment andprocess as described for TDI and MDI hand mix foam. The formulations used arelisted in Table 1.42 and the rate of rise data reported in Table 1.43. Dabco BLV withDabco T-9 was used as the industry standard catalyst control for the water blownflexible slabstock foam formulations. The silicone surfactant used in the reported flexibleslabstock formulations was DC5160. All water blown formulations were made atdensities of 23 kg/m3.

snoitalumrofkcotsbalselbixelF24.1elbaT

noitalumroFnoitacifitnedI

VXXX IVXXX IIVXXX

stnenopmoC phpp phpp phpp

2153lonaroV 001 001 001

retaW 06.4 06.4 06.4

VLBocbaD 21.0 — —

005ENocbaD — 60.0 61.0

006ENocbaD — 21.0 —

9-TocbaD 52.0 52.0 62.0

0615CDocbaD 09.0 09.0 09.0

)xedni001(IDT 4.75 4.75 4.75



74

1.5.1 Reactivity

Formulations from Table 1.42 were used in rate of rise catalyst comparisons and theresults are summarised in Table 1.43. The data suggests there is no significant change inreactivity when comparing the control formulation (XXXV) to the experimentalformulations.

1.5.2 Standard Physical Properties

Table 1.44 lists the physical property data for the flexible slabstock formulations. Theexperimental non-fugitive catalyst Dabco NE500 can be used with a slight increase inDabco T-9 catalyst level (shown in Table 1.42, formulation XXXVII). FormulationXXXVII compares very well to the control (XXXV). Formulation XXXVI illustrates theuse of Dabco NE500 and Dabco NE600 balanced non-fugitive catalyst combination, tomatch a Dabco BLV control. In this case, all physical properties were also matched, withthe exception of slightly decreased airflow and improved ILD properties. Overall, eitherthe non-fugitive catalyst combination Dabco NE500/Dabco NE600 or the use of DabcoNE500 with increased Dabco T-9 can replace Dabco BLV in all water blown flexibleslabstock formulation.

1.5.3 TDI Flexible Slabstock Foam Review

Dabco NE600 in combination with Dabco NE500 or Dabco NE600 as a sole aminecatalyst used with increased levels of T-9 provides for equal or improved performancecompared to Dabco BLV in all water-blown flexible slabstock foam. In all cases, thesecatalyst combinations can completely replace Dabco BLV and eliminate amine emissionsfrom the flexible slabstock foam.

gnisunosirapmocytivitcaeresireerfximdnaH34.1elbaTsnoitalumrofkcotsbalselbixelf

noitalumroF VXXX IVXXX IIVXXX

)sdnoceS(maerC 31 21 31

)sdnoceS(puCfopoT 52 42 42

)sdnoceS(leGgnirtS 76 76 46

)sdnoceS(esiRlluF 211 111 901

75

1.6 Foam Model Tool Discussions

Details of the TDI and MDI foam model systems have been previously published [2].The models require the use of mono-functional reactants that are quantitatively analysedto correlate structure-activity relationships for various classes of catalysts. A realisticthermal profile is produced through the imposition of an external exotherm. Urethane,urea, allophanate and biuret reaction products are quantified by liquid chromatographicanalysis of quenched reaction samples. The models effectively account for such non-ideal conditions as reactant depletion at variable rates, temperature and concentration-dependent catalyst activity, and catalyst selectivity as a function of isocyanate distribution.

1.6.1 TDI and MDI Moulded Foam Model

The information below highlights the features characteristic of the TDI and the MDIflexible moulded models. Table 1.45 illustrates the conversion of the formulations fromTables 1.12-1.15 into a TDI model system. Table 1.46 illustrates the conversion of theformulations from Table 1.35 into an MDI model system.

Since the TDI and MDI automotive polyols are highly ethylene oxide (EO) tipped thereactive hydroxyl group can be represented by a primary alcohol. The electronic effect of

maofkcotsbalselbixelfrofseitreporplacisyhP44.1elbaT

ytreporPlacisyhP VXXX IVXXX IIVXXX

)N(DLI GVA DS GVA DS GVA DS

%52 891 - 112 - 002 -

%56 104 - 124 - 104 -

nruteR%52 621 - 431 - 821 -

)%(dnuobeRllaB 14 1.0 04 1.0 04 1.0

)MLS(wolfriA 801 3.5 27 0.6 49 8.6

m/gk(ytisneD 3) 22 3.0 22 4.0 22 2.0

)aPk(elisneT 17 6.0 37 5.0 57 5.0

)m/N(raeT 002 - 002 - 002 -

)%(noitagnolE 79 5.7 601 7.6 401 1.6

)%(teSnoisserpmoC%09 6 5.0 7 5.0 6 5.0



76

metsysledommaofdedluomelbixelfIDT54.1elbaT

noitalumroF ssaM)g(

sdnuopmoCledoM ssaM)g(

sloyloPrehteyloP 001 rehtelyhtemlocylgenelyhteiDHC 3 HCO 2 HC 2 HCO 2 HC 2 HO

emylgiDHC 3 HCO 2 HC 2 HCO 2 HC 2 HCO 3

34.01

98.59

enimalonahteiD)sisabeerfretaw( a

47.1 enimalytubiD 41.2

retaW 02.4 H2O 02.4

tnatcafruS ledomrofderiuqertnatcafrusoN

egakcaPtsylataC 53.0 egakcaPtsylataC 53.0

xednI08IDT 501 OCNhP%elom04%elom06 o OCNlyloT-

deiraV

a .ssamrehtelyhtemlocylgenelyhteidhtiwdedulcnislyxordyhenimalonahteiD

metsysledommaofdedluomelbixelfIDM64.1elbaT

noitalumroF ssaM)g(

sdnuopmoCledoM ssaM)g(

loyloPrehteyloP 001 rehtelyhtemlocylgenelyhteiDHC 3 HCO 2 HC 2 HCO 2 HC 2 HO


15.7

45.39

enimalonaporposiiD)eerfretaw( a

58.0 enimalytubiD 18.0

retaW 09.2 H2O 09.2

5252-CDocbaD 00.1 ledomrofderiuqertnatcafrusoN

egakcaPtsylataC 06.1 egakcaPtsylataC 06.1

xednIRMrudnoM 59 OCNhP%elom08%elom02 o OCNlyloT-

deiraV

a .ssamrehtelyhtemlocylgenelyhteidhtiwdedulcnislyxordyhenimalonaporposiiD

77

the ether two carbons away from the alcohol is accounted for by using an ether alcohol.Thus both stearic and electronic effects on the reaction kinetics can be accurately reflectedin the model compound. The unreactive polyether backbone of the polyol can berepresented by the dimethyl ether of diethylene glycol. Dipropylene glycol dimethyl etherwould be more representative of the polyether backbone of a typical EO-tipped propyleneoxide (PO) polyol, but the rate and selectivity measurements are minimally affected bythe choice of model polyol backbone. Note the sum of the masses of the model alcoholand the diglyme slightly exceeds 100 grams, the total polyol mass, since the hydroxylsfrom the alkanolamine crosslinkers are included in the model alcohol mass. Nonetheless,the concentration of reactive groups is comparable to that in the actual foam. Water andcatalysts are the same for both formulation and model. The model system does notrequire a surfactant however, since a foam is not actually produced. TDI 80 and MondurMR are represented by 40:60 mole% and 80:20 mole% mixtures of phenyl and ortho-tolyl isocyanate, respectively. Phenyl isocyanate represents the 4-position in 2,4-TDI or4,4´-MDI, and tolyl isocyanate represents the more sterically hindered 2- and 6-positionsin the 2,4- and 2,6-TDI, or the internal rings of MDI oligomers. It is important to includeboth isocyanate types because using only the more reactive phenyl isocyanate significantlyoverestimates the reactivity of the system.

Detailed procedures for an individual model run can be found in the literature [14]. Forthese runs a masterbatch mixture containing the alcohol, ether, water, catalyst and the3,3´-dimethylbiphenyl internal standard was prepared in advance. The masterbatchmixture was charged to a 50 cm3 roundbottom flask equipped with a glass thermocouplewell, septum, and jacketed mechanical stirrer with gas inlet and a septum. For the TDImodel the flask was placed under a slow argon purge in a sand bath capable of raisingthe internal temperature from 60 °C to 120 °C in 4 minutes in the absence of any reactions.The isocyanate was added via a syringe when the internal temperature reached 60 °C.For the MDI runs the flask was wrapped with insulation but not otherwise heatedexternally, and the isocyanate was added at 25 °C. Samples were then withdrawn via asyringe at 30 second intervals for eight minutes. The samples were quenched withdibutylamine and analysed by liquid chromatography to determine the yields of the ureaand urethane reaction products as well as the unreacted isocyanates, analysed as thedibutyl ureas.

The flexible moulded chemical foam model provides a detailed look at the performanceof the industry standard catalyst package Dabco 33LV/Dabco BL-11 compared to thenon-fugitive catalyst packages Dabco NE1060/Dabco NE200 and Dabco NE1060/XF-N1085. The foam results presented previously can be explained in terms of the overallcatalyst performance. The catalysts were characterised with a model system becauseurea, urethane and isocyanate can be quantified as a function of time, so this approachprovides the highest level of detail. However, a model system is only relevant to the



78

extent that it accurately reproduces the environment in an actual foam. The mixing inthe model systems is less energetic, so this may influence the relative rates of catalysthydrolysis versus reaction with isocyanate. However, if catalysts are compared under aconsistent set of conditions, the relative performance differences should be meaningful.

The most convenient way to compare catalyst performance is to use blow to gel selectivity.Selectivity as a function of time (t), is defined next.

Normalised blowing rate = (% yield of ureas at time, t) / (limiting urea yield)

Normalised gelling rate = (% yield of urethanes at time, t) / (limiting urethane yield)

Blow to gel selectivity = (Normalised blowing rate) / (Normalised gelling rate)

Catalyst selectivity is defined as the ratio of the normalised amount of blowing (ureaformation) to the normalised amount of gelling (urethane formation). A selectivity of1.0 means that the normalised amounts of blowing and gelling are equal at that point inthe reaction. A selectivity substantially below 1.0, for example about 0.3, is indicative ofa strong gelling catalyst. A selectivity greater than 1.0 is indicative of a blowing catalyst.However, it has been shown that a catalyst with a blow-to-gel selectivity greater than 0.8can still serve to balance a strong gelling catalyst such as TEDA, and a good quality foamcan be produced [14].

Note the limiting urea and urethane yields are simply the molar equivalents of water andalcohol, respectively, from Tables 1.45 and 1.46. Selectivities are plotted as a function ofisocyanate conversion in Figures 1.45 and 1.46.

1.6.2 TDI Flexible Slabstock Foam Model

Details on the development and use of foam model systems for flexible moulded foamhave been published [14]. In this section the features characteristic of slabstock foamswill be highlighted. Table 1.47 illustrates the conversion of the formulations from Table1.42 into a model system.

Since the Voranol 3010 polyol is fully PO tipped, the reactive hydroxyl group can berepresented by a secondary alcohol. The unreactive polyether backbone of the polyolcan be represented by the dimethyl ethers of dipropylene glycol and diethylene glycol.The diglyme represents the wt% EO incorporation into the polyol. Note the sum of themasses of the model alcohol and the two dimethyl ethers equals 100 grams, the totalpolyol mass. This ensures that the concentration of reactive groups is comparable to thatin the actual foam. Water and catalysts are the same for both formulation and model.

79

Figure 1.45 TDI Flexible Foam Model Selectivity versus NCO Conversion

Figure 1.46 MDI Flexible Foam Model Selectivity versus NCO Conversion



80

Once again the model system does not require a surfactant, since a foam is not actuallyproduced. As noted previously, TDI 80 is represented by a 40:60 mole% mixture ofphenyl and ortho-tolyl isocyanate.

Detailed procedures for an individual model run can be found in the literature [14]. Forthese runs a masterbatch mixture containing the alcohol, ethers, water, Dabco BLV catalystand the 3,3´-dimethylbiphenyl internal standard was prepared in advance. The masterbatchmixture was charged to a 50 cm3 roundbottom flask equipped with a glass thermocouplewell, septum, and a jacketed mechanical stirrer with gas inlet. The flask was placed undera slow argon purge in a sand bath capable of raising the internal temperature from 50 °Cto 120 °C in 4 minutes in the absence of any reactions. Isocyanate was added via syringewhen the internal temperature reached 50 °C, and the T-9 was injected from a microlitersyringe immediately afterwards. Samples were then withdrawn via the syringe at 30 secondintervals for eight minutes. The samples were quenched with dibutylamine and analysedby liquid chromatography to give yields of the urea and urethane reaction products as wellas the unreacted isocyanates, analysed as the dibutyl ureas.

The chemical flexible slabstock foam model provides a detailed look at the performanceof the industry standard catalyst Dabco BLV compared to the non-fugitive catalysts DabcoNE500 and the Dabco NE500/Dabco NE600 package, keeping the Dabco T-9 levelconstant. Selectivity and conversion can be calculated as described previously. Note that

metsysledommaofkcotsbalS74.1elbaT

noitalumroF )g(ssaM sdnuopmoCledoM )g(ssaM

0103lonaroV 001 lonaporp-2-yxohteM-1HC 3 HCO 2 HC(HC 3 HO)

rehtelyhtemidlocylgenelyporpiDHC( 3 HC(O 3 HCHC) 2)2O


10.9

99.28

0.8

retaW 00.6-00.2 H2O 00.6-00.2

0615CDocbaD 57.0 ledomrofderiuqertnatcafrusoN

VLBocbaD 02.0 VLBocbaD 02.0

9-TocbaD 05.0 01-TocbaD 05.0

xednI08IDT 021,501,09 OCNhP%elom04%elom06 o OCNlyloT-

deiraV

81

the limiting urea and urethane yields are simply the molar equivalents of water andalcohol, respectively, from Table 1.47.

Dabco 33LV is considered to be a strong gelling catalyst, but under slabstock conditionsit is actually more of a blowing catalyst, with an initial selectivity near 2. This reflects thedifficulty of catalysing the secondary alcohol-isocyanate reaction in competition withthe water-isocyanate reaction. Tertiary amine catalysts are highly sensitive to stearichindrance near the reaction site. Secondary alcohols actually become less reactive thanwater, even in the presence of a strong gelling catalyst. This is why tin catalysts arecritical in flexible slabstock formulations because they are less sensitive to stearicallyhindered reactions. Dabco BLV contains some bisdimethylaminoethyl ether (Dabco BL-11), which is a very strong blowing catalyst. Thus the initial selectivity is even higher,near 4. The non-fugitive catalysts give selectivities comparable to that of Dabco BLV.Note that the non-fugitive catalyst packages are more blowing selective than both theTDI and MDI controls. Modern polyol technology tolerates and in some cases evenbenefits from higher initial blowing selectivity because of the reduced levels of monoland diol relative to triol. Thus network formation is more efficient, and correspondinglyless gelling is required to produce a superior urethane network. Higher blowing selectivitycan promote improved cell opening, which is now more critical in the face of improvednetwork formation. In a sense the control formulations are over gelled, but olderformulations could tolerate high gelling due to weaker urethane networks and the frequentuse of Dabco BL-11, which is a potent blowing catalyst.

Although the selectivity curves are different, the catalyst packages are still rate matched.Rate matching in foam is usually accomplished by comparing rise profiles, which arelargely determined by the blowing rate. Thus the non-fugitive catalyst packages provideblowing comparable to the controls, but provide a lower extent of urethane formationearly in the foaming process.

1.7 Conclusions

Incorporation of these newly developed additives in the production of polyurethane foamresults in polyurethane articles that give dimensional stability, low emissions and widerprocessing latitude. These dimensional stability additives provide a significant reductionin the force necessary to crush freshly demoulded foam without adversely affecting thephysical properties of the foam. As increasing demands are placed on the foam producersto meet specific comfort and durability requirements, physical properties andenvironmental concerns, the need for these additives will continue to be a key criterionfor polyurethane manufacturers.



82

Acknowledgments

The authors would like to thank and acknowledge the efforts and support of the followingpeople at Air Products and Chemicals, Inc.: Mark A Eckert, Steven E Robbins for physicalproperty testing of foams generated in this study, and Ilean S Ruhe for preparation andediting of this work. We would also like to thank Air Products and Chemicals, Inc., forsupport of this work and permission to publish it.

References

1. J. D. Tobias, G. D. Andrew, Presented at the SPI Polyurethanes Expo ‘98, Dallas,TX, 1998, p.445.

2. M. L. Listemann, A. C. L. Savoca and A. L. Wressell, Journal of Cellular Plastics,1992, 28, 4, 360.

3. R. G. Petrella and S. A. Kushner, Presented at the SPI Polyurethanes ‘90Conference, Orlando, FL, 1990, p.186.

4. G. Burkhart and M. Klincke, Presented at the SPI Polyurethanes ‘95 Conference,Chicago, IL, 1995, p.297.

5. Flexible Polyurethane Foam, The Dow Chemical Company.

6. ASTM D3574-95Standard Test Methods for Flexible Cellular Materials - Slab, Bonded, andMolded Urethanes Foams.

7. L. J. Gibson and M. F. Ashby, Cellular Solids, Structure and Properties,Pergamon Press, Oxford, 1988, Chapters 5 and 6.

8. K. D. Cavender, Presented at the SPI, Magic of Polyurethane Conference, Reno,NV, 1985, p.314.

9. SAE J1756Test Procedure to Determine the Fogging Characteristics of Interior AutomotiveMaterials, 1994.

10. M. S. Vratsanos, Presented at the SPI, Polyurethanes ’92 Conference, NewOrleans, MS, 1992, p.248.

11. K. F. Mansfield, Air Products & Chemicals, Inc., unpublished results.

83


12. R. G. Petrella and J. D. Tobias, Presented at the SPI Polyurethanes ‘88Conference, Philadelphia, PA, 1988, p.28.

13. D. G Battice and W. I. Lopes, Presented at the SPI Polyurethanes: Exploring NewHorizons Conference, Toronto, Canada, 1986, p.145.

14. M. L. Listemann, K. R. Lassila, K. E. Minnich and A. C. L. Savoca, inventors;Air Products and Chemicals, Inc., assignee; US Patent 5,508,314, 1996.


84

85

2 Demands on Surfactants in Polyurethane FoamProduction with Liquid Carbon Dioxide Blowing

Andreas Weier and Georg Burkhart

2.1 History of Polyurethane Foams

The history of polyurethane foams and of polyurethanes in general is not that long for aplastic material. It has been characterised by one special feature: the consistent changesof the industry.

When Otto Bayer started his research on this new class of plastic material built from theaddition reaction of isocyanates and hydroxyl containing materials in the years 1936/1937he envisioned a plastic material which could be tailored to many different applications bythe broad variety of acidic hydrogen containing compounds already available.

But there were also setbacks to this concept. The first major drawback was the fact thatthe necessary diisocyanates were not available on an industrial scale and large scalesynthesis of these compounds seemed to be a rather high hurdle.

A second major point was an obstacle that Bayer and his group faced again and again.When they tried to make the polyurethanes they envisioned in solid form, like films, etc.,they always ended up with gas bubbles in their plastics. That was due to the inefficiencyof the methods available to synthesise pure water free material. After encountering thiseffect for nearly four years Bayer made a bold move in line with the concept ‘use whatyou cannot prevent’ – in this case the formation of gas bubbles [1, 2]. Thus Bayer’s groupstarted to make polyurethane foams.

Although these new lightweight porous materials were envisioned as supportive materialsas well as insulative materials in the base patents [3, 4] their market development remainedslow. This can be seen by the fact that even in 1952 polyisocyanates, mainly toluenediisocyanate (TDI), were available worldwide in quantities of less than 100 tonnes. Afterthis rather hesitant start of polyurethane history and the first major switch from solidmaterials to porous foamed plastics, the industry has been characterised by significantchanges in concept and the resulting industrial application of these switches.

The first major switch resulted from the basic research enabling the technical productionof soft polyurethane foams in the early 1950s. This technology was mostly focused onpolyester polyols as raw materials. About five years later polyether polyols entered the

86


market as yet another major change within the polyurethane foam industry. Since theearly 1950s the foaming machines for the industrial production of polyurethane foamshad been discontinuous machines. With the use of polyalkyleneglycols as polyols, theimportance of continuous foam production grew significantly, for both performanceand commercial reasons. In the early 1950s, the necessary stabilisers for the productionof polyurethane foams had been silicone oils, with the major suppliers being Bayer, DowCorning, General Electric and ICI. In the late 1950s the picture changed with the broadapplication of the continuous process, first with prepolymers, then the one-shot-process.For this type of polyurethane foam production a different class of stabilisers was needed:the new silicone-polyether copolymers.

The larger variability of the available polyetherpolyols laid the foundation for a significantbroadening of the performance range of polyurethanes – and all this with lower coststhan before. The growth of polyurethanes after this chemical change was fast [5]. Evenin 1960 more than 45,000 tons of flexible polyurethane foam were produced. Besidesits major application as insulation materials and comfort products in furnitureapplications, new usage areas have been continuously identified and tailor-madepolyurethane materials have been supplied. Nowadays the range covers very high densitycontour parts as well as very low density packaging foams, flexible foams as well assemi-rigid and rigid foams, thermoplastic polyurethanes as well as integral skin foams orreaction injection moulded (RIM) materials [6]. In the beginning of the 1990s an estimated500,000 people had jobs in the polyurethane industry, either within production of rawmaterials or in the conversion to final products [7].

2.1.1 Environmental Concerns in Relation to Flexible Foam Density

By definition the production of a foam depends on the formation and stabilisation of gasbubbles in a liquid. That is true even for polyurethane foam when the additional curingof the liquid results in an elastic or rigid solid material. One of the central questions ishow the gas bubbles are generated and how long this generation takes.

The basic chemical gas forming process in polyurethane foaming is the very exothermicreaction of water and isocyanate, resulting in the generation of carbon dioxide, urea andheat. This method as the sole method to generate the gas bubbles of a polyurethane foamis limited to the production of a rather small range of density/hardness combinations.The achievable minimum density is limited by the tolerable heat in a foam bun beforescorching or even self-ignition occurs. Increasing water levels lead to the generation ofincreasing amounts of carbon dioxide and therefore to lower densities. Increasing waterlevels will also result in higher hardness as more urea is generated. During reaction theurea appears in a variety of associated forms before it will ultimately separate as a solidin flexible slabstock polyether foams [8, 9, 10].

87

Demands on Surfactants in Polyurethane Foam Production…

During the history of polyurethane flexible slabstock foam production the accessible density/hardness range was continuously broadened with a wide variety of techniques. The mostobvious one was the use of a physical blowing agent to generate the gas volume necessary toexpand or generate gas bubbles without generating additional heat during the process. Dueto its inert character and its low biological toxicity, CFC-11 (Freon) was the liquid blowingagent of choice for many years in Europe, North America and Asia. That changed afterrecognition of the fact that CFC-11 had both a high global warming potential (GWP) and ahigh ozone depletion potential (ODP). The result was the inception of the Montreal Protocoland subsequently the exploration of a broad variety of alternative blowing agents (ABA),including hydrochlorofluorocarbons (HCFC), acetone, cyclopentane and methylene chloride.

With the use of ABA, both the available densities and the available foam hardnesseswere lowered in comparison to the purely water blown foams, as in these cases gasvolume was produced without generation of urea. Softening of foam was also achievedchemically, either with the use of an additive to change the urea morphology [13] or withthe use of crosslinkers in combination with a low isocyanate index [14].

A few years later machinery modifications enabled foamers to produce low density foamsby the use of increased water levels while reducing traditional blowing agent levels. Thisreaction preferentially utilises the reaction generating carbon dioxide as a blowing agent.When the water in the formulation is increased, the foam bun temperature can becometoo high for safe processing. This problem has been tackled by forced cooling equipmentgenerating an air exchange in the foam bun directly after production and thereby coolingthe bun to prevent scorch or fire problems. There is a variety of technologies for theforced cooling of foams, including Envirocure [13, 14], Rapid Cure [15], Reeves Brothers[16] and lateral cooling [17]. For the production of very soft foam grades some physicalblowing agent is still needed. Also there are always some safety concerns over possibleequipment malfunctions when making these hot foam grades.

A safer way to control flexible foam density and hardness is the variation of productionpressure. Making high as well as low pressures available in the foam productionenvironment can give access to higher density hard grades as well as to low densityfoams without the use of physical blowing agents. The two commercial processesaddressing this approach are Variable Pressure Foaming (VPF) and Foam One. As foamproduction is done in a closed chamber, volatile materials can be trapped rather efficiently.This creates the possibility of a clean and safe, as well as ABA free foam production. Asetback is the high capital investment cost, especially for a VPF unit. Therefore thisequipment only becomes cost effective in very high volume production plants.

Review of the many production techniques reveals that the production of a broad rangeof flexible foam qualities with reduction of environmental hazards (ODP as well as GWP)is possible by numerous options. All of these current technologies offer benefits, butprovide no clear economical solution for total ABA elimination.

88


Within all the blowing agents available today, carbon dioxide probably offers the bestcombination of zero ODP, low GWP and low price. As it is abundant, it doesn’t have tobe generated. For capturing and transporting only a minor amount of energy is needed.Merely the point of emission is changed. On top of all that it is also non-flammable,which minimises safety concerns in the foam plant.

Therefore many attempts have been made to produce polyurethane foams with carbondioxide that is not generated by a chemical reaction during the foaming process. Thedifficulty is that carbon dioxide is a gas at room temperature under atmospheric pressure.

The only way to utilise carbon dioxide not generated chemically is to mix the carbondioxide under pressure or in a liquid state with the chemical components of the foamformulation. Then the pressure is subsequently released causing the frothing or pre-expansion. Carbon dioxide as a non-reactive blowing agent in the frothing techniquehas been suggested for quite some time already [18, 19]. The pressurised reaction mixtureis ejected at atmospheric pressure causing a turbulent vaporisation of the blowingagent. This allows the manufacture of a foam with reduced density, but the cell structureis of very inconsistent quality due to irregular shaped and oversized cells or bubblesbeing present.

2.2 Current Liquid Carbon Dioxide Technologies for Flexible SlabstockPolyether Foam Production

2.2.1 Machinery

The situation changed when Cannon introduced liquid carbon dioxide foaming technologyfor the industrial production of polyurethane flexible foams with its CarDio process. Thisprocess was introduced to the polyurethane foam production industry in 1994. Today foammanufacturers worldwide consider liquid carbon dioxide as a blowing agent and make thisprocess available for the industry. The major point of this technology is the control of thefrothing that occurs as soon as the pressurised liquid carbon dioxide as part of the foamformulation leaves the pressurised equipment and the liquid foam mixture is released intoatmospheric pressure. The controlled release of the mixture into the comparatively lowatmospheric pressure conditions is essential to all the liquid carbon dioxide foaming techniques.

In the CarDio process this is accomplished with the specifically designed lay down device[20, 21] which is called a ‘Gate Bar’ (see Figure 2.1). The gate bar essentially consists ofa metal bar with a pressure drop slot connecting the outside with a feeding tube insidethe bar. Through a liquid mix injection point formulation components with the pressurisedliquid carbon dioxide enter the gate bar.

89


The frothing mixture flows along the frothing cavity and through an outlet apertureacting as a pressure drop zone onto a substrate, e.g., the moving belt of the foam machine.

A variation of the original gate bar design was introduced with the CarDio 2000 featuringa pressure adjustable lay down device [22]. It allows adjustment of the slot width inresponse to the output and the liquid carbon dioxide percentage (see Figure 2.2).

Figure 2.1 Cannon Gate Bar Principle

Figure 2.2 CarDio Froth laydown

90


Another patented technology for liquid carbon dioxide foaming is the NovaFlex processdeveloped by Hennecke and Bayer [23]. The NovaFlex process uses a creamer dispensingthe froth on to the machine. It focuses on the same task of pressure reduction as theCarDio gate bar, except dispensing the froth occurs at one single spot on the belt and isnot distributed over nearly the whole belt width. The same is true with the third versionof liquid carbon dioxide machinery, the Beamech CO-2 equipment. Obviously, both theNovaFlex as well as the Beamech CO-2 process use pressure reduction devices differentfrom the gate bar of the CarDio process as they are said to show a different response tothe use of fillers or generally solid particles within the foam formulations. One modificationof the original NovaFlex process is the MultiStream configuration. It allows the use ofliquid carbon dioxide with a range of polyol types in all different ratios by utilising oneliquid carbon dioxide addition point in one high pressure polyol stream which is thencombined with additional polyols by pass streams [24, 25].

2.2.2 The Foaming Process

2.2.2.1 A Comparison of Rise Profiles

A comparison of the foaming profile found in a standard flexible slabstock foamproduction with the one seen during liquid carbon dioxide foaming clearly shows a largedifference in the rise profiles, especially at the lay down (Figure 2.3).

Figure 2.3 Foaming profiles of standard flexible foam production andliquid carbon dioxide foaming.

91


In standard foam production, a mixture of liquid materials is dispensed and the gas bubbleswithin the foam structure are only generated by a rather slow chemical reaction. In the caseof liquid carbon dioxide foaming the mixture is already dispensed in a two-phase process.

No matter which particular pressure drop system is utilised, the pressure drop is veryfast due to the large flow speeds in the system. The fast pressure drop between the mixinghead and the dispensing compresses this time span for bubble formation to fractions ofa second. As all the gas bubbles resulting in foam cells of the final foam have to begenerated in this short instant, clearly the nucleating powers of the formulation ingredientsare challenged severely. This is especially true when looking at the foam stabilisers usedas they are the main formulation ingredients determining the nucleation.

2.2.2.2 The Nucleation of Gas Bubbles

As early as 1969, Kanner and Decker [26] showed by photomicrography that self-nucleationis essentially absent in a polyurethane foam system. Their results indicate that bubbles areintroduced by the process of mixing and that the presence of a silicone surfactant increasesthe volume of air bubbles introduced during the mixing. Nucleation in this context meansthe formation of gas bubble cores on which a bubble might grow above its critical bubbleradius. The critical bubble radius is the smallest bubble radius for which the addition ofmore gas will result in a net decrease in free energy. A bubble which is smaller than thecritical radius will not spontaneously grow in size because the addition of more gas resultsin a net increase in free energy. Obviously a decrease in the surface tension of a foamingmixture by the silicone surfactant would decrease the energy needed to generate bubbles ofthe necessary critical radius. Even under conditions of a supersaturation pressure of 2.03MPa carbon dioxide in a liquid with a surface tension of 0.025 N/m and a temperature of25 °C, the critical bubble radius is 2.4 x 10-5 cm. It has been proven that silicone surfactantcan aid nucleation by lowering surface tension [27].

However the surface tension lowering of a silicone surfactant might just be one effecthelping to induce new bubble formation by liquid carbon dioxide. The surfactant isentirely soluble in the liquid and therefore cannot act as a heterogeneous nucleatingagent. Besides that other surfactants like fluorocarbons can also lower the surface tensionbut still don’t act like a silicone surfactant does in polyurethane foam production. Apossible explanation of this difference might be related to the exceptional solubility ofoxygen and nitrogen in silicones as demonstrated by Arkles [28, 29]. It is well known inthe polyurethane industry that insufficient air loading of raw materials like the polyolsresults in coarse cell structures even in the conventional foaming process. Obviously asufficient amount of nitrogen and/or oxygen is needed for good nucleation. As siliconesshow a high permeability and solubility for these gases, it is reasonable to assume that

92


the ease of formation of nuclei for gas bubbles is higher in silicone-like surroundingsprovided along the siloxane backbone of a silicone surfactant.

An effect not accessible to Kanner and Decker, and therefore not a subject of theirexperimental work, was the high nucleation demand found in the new liquid carbondioxide foaming technologies. Their finding that bubbles are introduced during the processof mixing by the shear energy of the mixer might reflect the observation that the amountof nuclei generated during mixing has always been sufficient for a standard polyurethanefoaming process. The rather large timescale between mixing and appearance of gas bubbleslarge enough to be visually detectable obviously gave carbon dioxide molecules sufficienttime to diffuse to a nucleation site as they were only generated slowly within the foamformulation. With the new liquid carbon dioxide foaming technologies, a significantlyhigher number of nucleation sites is needed for the carbon dioxide to find a gas bubbleor a gas bubble core during the short time frame of the pressure decrease in the pressurereduction area within the equipment used.

Whereas the former trend in some polyurethane (PU) markets has been to sacrifice moreand more of a surfactant’s nucleating power to gain higher and higher activity, liquidcarbon dioxide foaming has led to a new description of the needed performance profile.This is recognising the fact that quite a number of the very high active surfactants used atthe time of the introduction of the liquid carbon dioxide processes resulted in unacceptablefoam structures. Irrespective of the liquid carbon dioxide system, the high activitysurfactants generated coarse foams. That is in contrast to the findings of Kanner andDecker in standard foam systems. Obviously under these conditions the mixing energyof the machine system has not been high enough to generate sufficient nuclei or gasbubbles to guarantee a regular fine celled foam structure. Obviously under these conditionsthe main quality issue of a surfactant is its nucleation power, and not the stabilisationactivity. Not surprisingly a well established medium active surfactant was used duringthe development of the first commercial liquid carbon dioxide process, the CarDio process.It provided a good balance between nucleation and processing as well as activity. Thereforethis type of surfactant was still the number one choice for processes even four years afterthe introduction of commercial liquid carbon dioxide foaming. Still, due to the highdemands in nucleation, the broad variety of foam grades and the flame retardant (FR)demands of some markets a desire for optimising the different surfactant performanceaspects could clearly be seen. It also provided a focus of interest for the flexible slabpolyurethane community for quite a few years.

2.2.2.3 Emulsification of Raw Materials

In addition to nucleation, there is one further objective to be fulfilled or at least supportedby the silicones used. That is the emulsification of the raw materials used in the foam

93


formulation. An obvious example for this performance aspect is the chemical reaction ofwater and isocyanate, i.e., the formation of carbon dioxide and urea. The different polaritiesof isocyanate and water should normally prevent their reaction with each other in theshort time frame of a PU foam reaction and even the use of polyols as the major formulationcomponent will not always help to overcome the tendency for phase separation.

Even the polyols used in some special foam types can give rise to a number of problems.One of the challenges for the emulsification power of silicone surfactants in flexibleslabstock foam production for example is the combination of a standard polyol with ahypersoft polyol. The consistency of the foaming results and the cell structure obtainedcan be severely influenced by the choice of silicone surfactant used. It can be easilyrationalised that the rise profile and physical foam properties like tensile strength andelongation will be influenced by tendencies for phase separation after the mixing anddispensing of the formulation ingredients.

2.2.2.4 Stabilisation of the Foam During Foam Rise

Probably the most obvious task of any silicone surfactant in flexible polyurethane foamproduction is to stabilise the foam during the foam rise. If the surfactant used iscontaminated or is in too low a concentration, settling of the foam or even foam collapsewill occur. Two possible reasons for the foam instability during the rise time are discussedin the literature, and lead to different theories on the roles of silicone surfactants used.The first explanation of the role of silicone surfactants in flexible polyurethane foam isbased on the fact that as soon as the volume fraction of the gas bubbles exceeds 74%, thespherical bubbles will distort into multi-sided polyhedrals. This means that cell windowswith Gibb’s plateau borders are formed (a polyhedral foam consists of cell windows andstruts. Another term for struts is Gibb’s plateau borders). A pressure difference due tocapillary pressure will cause liquid in the cell window to drain into the struts since thepressure inside the plateau borders is lower than that in the cell windows. Without addingsilicone surfactants this drainage rate will be very fast, so that film rupture and bubblecoalescence occur rapidly. It has been shown that due to the surface tension gradientgenerated by a silicone surfactant the cell window drainage rate is lower [30]. Thereforedifferent surface structures do have different effects on cell window drainage. That hasbeen said to result in a distribution of different cell window thicknesses at the time of cellopening [31]. The critical factor in this picture is the surface elasticity. It is known to becaused by a surface tension gradient along the cell window upon expansion and it willretard the cell window drainage rate. This effect is called Gibbs-Marangoni effect and isreviewed by Scriven and Sternling [32]. Owen and co-workers also studied the dynamicsurface tension of a series of surfactant solutions and stressed the importance of surfaceelasticity [33]. Similar work has been performed by Zhang and co-workers [34].

94


A different view point of the mechanism of flexible polyether polyurethane foam stabilisationby silicone-based surfactants has been published by Rossmy and co-workers [8, 9, 10]. Theauthors investigated the urea precipitation phenomenon in polyurethane foams in greaterdetail. By adding a defoaming agent to a foam formulation they showed that urea precipitationturned a clear foam mix opaque. They also noted that the cell rupture always occurred justafter urea precipitation. In light of this observation the hypothesis is that precipitation ofpolyurea destabilises the foam mix and leads to cell opening. The surfactant then aids instabilising the foam by incorporating the urea precipitate in the foam matrix, adding integrityto the foam. Today it is not doubted any more that there really are urea domain structures inpolyurethane foams and they have been subject to intensive research [35, 36].

2.2.2.5 Control of Blow Off

Another major point of consideration is obviously the processing latitude of siliconesurfactants. This phrase describes the variability of catalyst concentration without runninginto either closed, dead foam or observing first signs of foam instability indicated bysplits in the foam. Not only will a different silicone surfactant result in a differentprocessing window, e.g., amount of stannous octoate variation between the two extremes,but also the position of the processing window might be different.

Figure 2.4 shows the remaining fragments of a cell window after cell opening as found inthe fully cured foam. It is very obvious that the opening of a flexible PU foam is not a

Figure 2.4 Opening of a cell window

95


pressure induced rupture of a foam bubble. Otherwise there would be no explanationwhy the cell opening process clearly starts at a number of locations around the peripheryof the cell window. Obviously weak points in the cell window offer starting points forthe opening of the liquid film. Then drainage occurs, as can be seen by the thin stringsremaining between the opened areas in the cell window. The actual timing of this openingprocess is critical as it determines the physical properties of the foam and the foameconomics. As the viscosity of the liquid material is rising very sharply due to crosslinkingat the end of the foaming profile, there is only a very short time frame in which thematerial is crosslinked enough to support the foam without a high settling once the gasbubbles have opened and no internal gas pressure is supporting the foam anymore. Onthe other hand a further delay in gas bubble opening would lead to an incomplete recessionof the material into the foam struts so that the foam will be showing a rather low airflow or even shrinkage.

2.2.3 Additional Tasks of Silicone Surfactants in Flexible SlabstockFoam Production

2.2.3.1 Influence on burn performance of the foam

As soon as the foam is produced and cured the silicone surfactants used are not neededor desired any more. One exception might be the production of hydrophilic or hydrophobicfoams where a surface active material like a silicone will in some cases have a recognisableinfluence. But with standard foams as well as with liquid carbon dioxide blown foams,additional performance aspects besides the stabilisation during the foam rise might beimportant for surfactant selection. These criteria therefore have to be addressed as welland deserve some consideration as they might also affect the number of key additives afoam operation has to handle. An effect clearly seen in the final foam is the undesirableeffect the silicone surfactant has on the burn performance of the resulting foam. Especiallyin liquid carbon dioxide foaming, suitable conventional (or non-FR) surfactants havebeen available right from the beginning. But the development of optimised universal ornon-FR-surfactants for liquid carbon dioxide foaming took considerable effort and time.To understand the effect of the silicone surfactant on burning it is interesting to have alook at a theory for flame spread development.

The simplified illustration in Figure 2.5 shows that the heat decomposes the organicmaterial at the surface during an endothermic procedure. Pyrolysis products are createdand because of an exothermic reaction with oxygen at the boundary layer of the flame,the formation of even higher reactive decomposition products takes place. These aggressiveradicals are responsible for an accelerated degradation of the polymeric surface. As longas the result of the energy is positive, there is thermal feedback to the endothermic process

96


at the surface and the combustion keeps going. It is believed that the formation of hydrogenradicals and hydroxyl radicals from the organic material are the key factors initiatingand supporting the combustion phenomena. The working mechanisms of FR addressthese phenomena.

The effect of a silicone surfactant on the burning behaviour of the polyurethane foamcan be quite significant even though they normally represent less than 1% of the plasticmaterial. It is due to the fact that in any case the decomposition of a polyurethane foamstarts at the surface. Because of the surface activity of the foam stabilisers it is easy torationalise their enrichment on the surface and it is the surface that is the most influentialpart of the polymer regarding flame spread development.

This point can be further highlighted if we look at the influence of surfactants on theFR-performance of a polyurethane foam with 1 weight percent of surfactant externallyapplied (Figure 2.6). To characterise the burn performance in this case, a horizontalburn test was used and the burn length was measured. Surfactants A, B, C and D arerigid foam surfactants with different structural parameters in the silicone as well as inthe polyether chains. It can be seen in Figure 2.6 that all the different types of siliconesurfactants used do have a negative influence on the burn performance of the foam.The reason for this behaviour can be mainly attributed to the structure of the siloxanebackbone within the siliconepolyethercopolymers [37]. To characterise the effect ofthe polydimethylsiloxane (PDMS) chain used as the backbone of such asiliconepolyethercopolymer, interesting tests have been made by external applicationof pure PDMS in different weights as well as the use of molecules of different averagechain length (Figure 2.7).

Figure 2.5 General assumption about flame spread

97


Not only do increasing amounts of externally applied PDMS lead to increasing burnlength in, but more astonishingly, this effect is obviously correlated to the averagechain length of the applied siloxane. It can be seen that with longer chain length of thePDMS, the detrimental effect on the FR-performance of the foam increases. This effectfar out balances the pure weight effect that would explain why a higher amount ofPDMS leads to higher burn length value due to the larger amount of materials supportingthe decomposition.

To address the question of whether the different volatility of the different PDMSchains is the main reason for this effect, an additional test with silica was carried

Figure 2.6 Influence of surfactants on FR-performance - external applicationof 1% surfactant

Figure 2.7 Effect of 1% PDMS - applied externally

98


out. It is easy to expect that solid materials, especially if inorganic, will have a positiveeffect on the FR-performance of the foam and that they will lead to a decreasingburn length in a burn test. Depending on the chemical nature they could be eitherexpected to be a heat sink, a material to catch free radicals, to char the burn front ofthe foam or to liberate non-gaseous products to dilute the oxygen at the burn front.

Astonishingly, silica, a chemically inert material, under these conditions increasedthe burn length of the foam (Figure 2.8). The assumption here is that on the surfaceof the silica-particles either catalytic effects on the heat initiated oxidation processestake place or the silica particles themselves act like a wick in a candle, providing ameans to transport melted organic material from the surface and increase the diffusionflame front. If the formation of silica-particles on the burn front is indeed the majordetrimental effect of the silicone containing polymer, it should be helpful to decreasethe average length of any unmodified PDMS chain part between any two modifyinggroups. Although in practice this is also an efficient tool to improve the burnperformance of the resulting foam, it negatively affects the nucleation power of thesilicone surfactant. Therefore this way of addressing FR-performance behaviour isdetrimental to the nucleation efficiency of any surfactant for liquid carbon dioxidefoaming [38]. That is the main reason why conventional (non-FR) silicones for theliquid carbon dioxide processes were available before their optimised universal oreven pure FR-counterparts.

Figure 2.8 Addition of solid material

99

2.2.4 Chemistry of a Silicone Surfactant in Flexible Slabstock Foam Production

As outlined previously the main reason for the interest in optimised surfactants for theliquid carbon dioxide processes lies in their role as nucleation promoters.

The silicone surfactants can be viewed as PDMS-polyether-copolymers which are mainlybased on a combination of just three structural units: the methyl substituted siloxanebackbone as well as a sophisticated ratio and arrangement of ethylene oxide and/orpropylene oxide forming the attached polyethers and, in some cases, additionalmodifications.

A typical structure of a silicone surfactant is shown in Figure 2.9.

Figure 2.9 Building blocks of a surfactant

The molecules generally have a siloxane backbone formed by dimethylsiloxane units,substituted methylsiloxane units and some endgroups. Polyether groups and/or additionalmodifications can be attached to the siloxane backbone. Not only are a wide variety ofthose additional modifications possible but also the polyether groups can vary in quite anumber of ways. Although they mostly consist of ethylene oxide and propylene oxide unitsthey obviously can be different in chain length, i.e., their molecular weight. They can alsohave the different monomers distributed along the polyether chain in an either random ora blockwise fashion as well as with alternating blocks. The polarity can be either evenly orunevenly distributed. A more polar side could be either close to the attachment point to thesiloxane or far away from it. As this is true for each of the polyethers used and flexible


100


slabstock surfactants generally have more than one type of polyether attached, this givesrise to virtually millions of different combinations for surfactant molecules. Also, in mostfoaming applications the pendant polyethers have to be unreactive to isocyanates so thatthey do not act like crosslinkers which would lead to tight or shrinking foam.

In addition there are two general ways to link the polyethers to the siloxane chains (seeFigure 2.9). This opens the basis for two separate product lines, each having its specialadvantages over the other.

2.2.4.1 Si-C-Surfactants Versus Si-O-C-Surfactants

The Si-O-C products resulting from the reaction of chlorosiloxanes and hydroxylgroups of polyethers provide an extremely good processing and superior consistencycombined with being hydrolytically stable under the water-amine conditions foundin polyurethane foaming. The Si-C products offer a more beneficial access to highactivity and lead to an easier production of flame retardant foams. These productsare derived from the addition of an Si-H functionality at the siloxanes to a doublebond, a process called hydrosilylation.

Examples of how this variety of structural parameters affects the development of siliconesurfactants especially for liquid carbon dioxide blown foams, are many [39, 40, 41, 42]and this latest drive in surfactant development provides good examples of the importantperformance issues and how they can be addressed.

Figure 2.10 Synthesis of Si-C and Si-O-C linked siloxane-polyether-copolymers

101

2.2.5 A Surfactant Development Example

In surfactant developments for liquid carbon dioxide PU foaming, the describedcomplexity of structural parameters in a surfactant led to the desire to characterisenucleation performance of a surfactant without the need to judge the final, cured foam.One important reason is the fact that a coarse or even irregular cell structure in thefinal foam does not necessarily indicate the reason for such a ‘bad’ cell structure. Itcould arise from a deficiency in nucleation performance so that the number of gasbubbles formed at the beginning of the foaming reaction was not sufficient (lack ofnucleation). Another possible scenario could be that the initial number of bubblesformed was rather high, but many of the gas bubbles were lost during the foam risedue to gas bubble coalescence. That would mean that the stabilising power of thesurfactant would be insufficient and had to be improved.

One of the first published attempts to get an indication of the nucleation performance ofa silicone surfactant was the froth test [39, 40, 41]. In this test polyol and surfactant arestirred in a reproducible fashion and the resulting froth density as well as the time neededfor the foam to collapse is measured. After some experience with industrial surfactantsused in the market these products with known performance in industrial liquid carbondioxide foaming were subjected to the froth test. No direct correlation could be foundbetween the obtained data and the performance of the products on industrial machines(Figure 2.11).

Figure 2.11 Performance of surfactants (Tegostab B 8228 and B 8220) in the froth testComp.: competitive surfactant; EP-H-18: experimental product


102


As can be seen from Figure 2.11, both products Tegostab B 8228 and Tegostab B 8220show nearly identical performance in this mixing test although industrial experience hasshown that Tegostab B 8220 results in significantly finer cell structure than Tegostab B8228, even if both of them yield a very regular cell structure.

Even more confusing is the picture resulting from the comparison of a widely usedsurfactant and an experimental product called EP-H-18 by Burkhart and co-authors[42]. Both of the products again resulted in identical values in the froth stability test,but differed significantly on an industrial production machine. Whereas the competitivematerial resulted in a coarse cell structure, the experimental product yielded a fine andregular foam.

Figure 2.12 Video images of foam formulations with commercial surfactants ofdifferent performance in liquid carbon dioxide processes

A (30 s) B (30 s) C (30 s)

A (150 s) B (150 s) C (150 s)

bad good better

103

Because of these results a video microscopy check on the homogeneity of the gas bubblesin the liquid mixture was performed. For this test the ingredients of a standard foamformulation were mixed and immediately poured onto a polished metal plate. Thecreaming of the foam was then filmed by a video camera with a 10x magnification.Thirty and 150 seconds after the stirring commenced, a video printout was made and thenumber of formed gas bubbles as well as their size homogeneity was characterised (seeFigure 2.12). This procedure is referred to as the video test.

2.2.5.1 Possible Silicone Structures

Not taking cyclic molecules into account, the general structures of industrial siliconesurfactants for flexible slabstock foam production can be seen in Figure 2.13. The mainbuilding blocks of these materials are a PDMS backbone and attached polyethers based onethylene oxide and propylene oxide addition products. The siloxane backbones can eitherbe linear or branched and can have their polyether substituents attached in an either pendantor terminal location. These four general structures are outlined in Figure 2.13).

Figure 2.13 General structures of silicone surfactants (schematic depiction)


104


The simplest type of the possible structures is a linear A-B-A-copolymer with a straightsiloxane backbone and polyether chains at the termini. This type of structure is describedas ST (for straight, terminal). With a straight siloxane backbone there is also the possibilityof attaching polyether groups as pendant chains. This combination is denoted SP (straight,pendant). If the siloxane is branched as well it could have polyether groups attached asterminal or as pendant groups (BT and BP). These four types of structures gave differentresults in the video imaging test (see Figure 2.14).

The ST type of structure resulted in only a small number of gas bubbles being formed in theearly stages of the reaction. Somewhat better was the combination of branched silicone withpendant polyether groups. The best video imaging results were reported with the branchedsiloxane, combined with terminal polyether groups followed closely by the SP type of molecule.

Figure 2.14 Video imaging test, comparison of different silicone structures-, o, +, +(+) are a non numerical quality rating of degree of nucleation. - is poor

and +(+) is very good

ST SP

BT BP

– +

+(+) o

105

These results matched, more or less, with industrial experience. Whereas the ABA typeof copolymer (ST-structure) is known to be a rather poor performing surfactant forpolyurethane production, the SP and the BT type of structures are commonly used. Thestructure that performed best in the video imaging test (the BT type structure) is thestandard structural type that can be found in commercial SiOC-PDMS copolymersobtained by the reaction of a chlorosubstituted PDMS with the hydroxyl group ofpolyether (see previously). Although these SiOC products (named after the silicone-oxygen-carbon bond linking the PDMS and the polyether side chain) provide a very goodcombination of broad processing, good nucleation and consistent quality, their maindisadvantage is their very recognisable negative effect on the burning behaviour of theproduced polyurethane foam.

Therefore these types of products are commonly denoted as conventional or non-FR-silicones.

In contrast, the second best type of structure seemed to be the SP type. These structuresare the fundamental means to build SiC-products which are used as universalsurfactants, for example in North America. Although SP structures seemed to be theobvious choice to build FR or universal type of surfactants (having a less negativeeffect on FR-performance) they did not seem to show as good a nucleation as the BTtype of molecules.

The fourth structural type, BP products, no known experience or industrial applicationwas recognised, and due to the combination of not extremely good nucleation (accordingto the video test) and the complicated synthesis necessary to make them, no reportedwork was undertaken with these types of molecules.

Besides the type of branching in the siloxane backbone and the attachment points ofthe polyether side groups, another factor characterising the siloxanes is the numberof unmodified PDMS groups for each of the modified methylsiloxane groups in themolecule. This ratio is often denoted as the P-value and is especially important inregards to the burn performance of the resulting foams (see Figure 2.10 and discussionabove). As that is another variable that has to be taken into account whencharacterising a silicone surfactant for polyurethane foam production, this factorwas screened for as well.

As can be seen from the video test results, there seems to be a slight improvement in thenucleation efficiency with increasing P-value. However the drawback is that the higherP-values result in surfactants which will lead to a considerably negative effect on theburn performance of the resulting foam [43].

The summary of the data at this point seemed to indicate that the combination of goodnucleation efficiency with the BT-structures and the advantage of high P-values seem to


106


correlate with the fact that, especially based on SiOC type of chemistry, it is easier tobuild a non FR-surfactant structure than to build a universal or true FR-silicone surfactantfor liquid carbon dioxide foaming.

In the case of good FR-surfactants therefore, a compensation for the unavailabilityof high P-values had to be found in regard to nucleation efficiency of the molecule.Because of a suggestion in the literature [44] that the high solubility of carbon dioxidein certain aromatic solvents might be due to interaction between the molecules, theinfluence of aromatic structures in surfactant molecules was screened. Whereas thestandard siloxane backbone in PU surfactants is normally consisting of a PDMSchain between the substituted silicon atoms, high temperature silicone resins areused with phenylmethylsiloxane groups within the chain. Although these types ofmolecules seemed to be viable candidates for testing, they did not seem to be attractiveoptions for industrial use due to the complicated synthesis and the high costs associatedwith them.

Another option to incorporate aromatic structures in organomodified silicones has beenthe use of styrene oxide as a monomer for the synthesis of the polyether side chain withinthe copolymer structure [45] (see Figure 2.16).

Comparison of structural siliconepolyether analogues with and without styrene oxideincorporation in surfactant test molecules showed an indication of a positive trend in thevideo test regarding the nucleation efficiency of those aromatic group substituted siliconepolyether copolymers.

Surprisingly enough in a real foam test these molecules performed rather badly, resultingin a considerably coarser foam than the reference (see Figure 2.17).

Figure 2.15 Video imaging test; nucleation efficiency with increasing P-value

P-value

107

That observation was rationalised with the following assumption. If indeed aromaticrings would increase the solubility of carbon dioxide then these types of structures wouldbe more soluble in a liquid carbon dioxide containing environment. This would result ina decreased tendency to move to the surface of a such a liquid phase. The obvious resultwould be a decrease of surface activity, so the net result could very well be a highernucleation efficiency combined with a lower activity during the rise time of the foam,leading to gas bubble coalescence.

Figure 2.16 Aromatically substituted polyether

Surfactant

without aromatic groups with aromatic groups

Cell structure in foam test:

10 cells/cm 8 cells/cm

Figure 2.17 Comparison of standard and aromatically substituted siliconepolyether


108


If that is true, the polarity and therefore solubility of the molecule would have to beadjusted to lower values. As explained previously the increase of the P-value was not anoption for the generation of a surfactant with good FR-performance.

Therefore to further decrease the solubility of a surfactant in a formulation for flexibleslabstock PU-foam, the polarity of the surfactant molecule was decreased by loweringthe polarity of the polyether side chain with the use of higher alkylene oxides to havemore non-polar groups along the polyether chain (Figure 2.18). This would lead to ahigher carbon:oxygen ratio in the polyether side chain. Since it has been known for morethan 30 years [46] that the polyether polarity affects the cell size distribution of flexibleslabstock PU foams that is a very good method for fine-tuning surfactants. Foaming testsshowed that the carbon:oxygen content had a significant impact on the overall activityof the surfactant, resulting in decreasing foam height with increasing carbon:oxygenratio, especially when a carbon:oxygen ratio of about 2.6 was exceeded.

Figure 2.18 Influence of carbon to oxygen ratio in polyether side chains

The utilisation of the outlined principles and the continuous work with the many possiblevariables by all the additive suppliers again and again has brought new high performancesurfactants, supporting the continuously changing demands in the innovative polyurethaneindustry and most probably will continue to do so. Therefore this industry has all thefascination that comes from the interaction of chemistry and physics, industrial applicationand theoretical interest, commercial importance and environmental awareness. So evenmore than 50 years after its beginning it will remain an area of new challenges andperformance oriented efforts.

109

References

1. No inventor; I.G. Farben AG, assignee; German Patent 728,981, 1937.

2. A. Höchtlen and W. Droste, inventors; Farbenfabriken Bayer AG, assignee;German Patent 913,474, 1941.

3. P. Hoppe, inventor; Farbenfabriken Bayer, assignee; German Patent 851,851, 1948.

4. W. Droste and A. Höchtlen, inventors; Farbenfabriken Bayer, assignee; GermanPatent 860,109, 1952.

5. Chemical and Engineering News, 1961, 39, 11, 62.

6. K. Moser, Kunststoffe, 1983, 73, 12, 764.

7. F. K. Brochhagen in Handbook of Environmental Chemistry, Vol. 3, Part G, Ed.,O. Hutzinger, 1991, Springer Verlag, Berlin, p.73.

8. G. Rossmy, H. J. Kollmeier, W. Lidy, H. Schator and M. Wiemann, Journal ofCellular Plastics, 1981, 17, 6, 319.


10. G. Rossmy, W. Lidy, H. Schator, M. Wiemann and H J. Kollmeier, Journal ofCellular Plastics, 1977, 15, 5, 276.

11. G. Burkhart, H-H. Schlöns and V. Zellmer, inventors; Th. Goldschmidt AG,assignee; US Patent 5 132 333, 1992.

12. R. Ricciardi, Presented at the Polyurethane Foam Association conference inScotsdale, AZ, 1996.

13. M. A. Ricciardi, D. J. Smudin, R. D. Wagner, M. Pcolinsky and J. E. Chaya,inventor; no assignee, US Patent 3,890,414, 1975.

14. M. A. Ricciardi and D. G. Dai, inventors; Crain Industries, Inc., assignee, USPatent 5,171,756, 1992.

15. H. Stone, inventor; PMC, Inc., assignee, US Patent 5,128,379, 1992.

16. A. A. Grizwold, inventor; Reeves Brothers, Inc., assignee; US Patent 4,537,912,1985.


110


17. J. L. Drye and G. C. Cavenaugh, inventors; Trinity American Corporation,assignee, US Patent 5,188,792, 1993.

18. P. Merriman, inventor; Dunlop Rubber Co., Ltd., assignee; US Patent 3,184,419,1965.

19. E. N. Doyle and S. Carson, inventors; no assignee, US Patent 5,120,770,1992.

20. C. Fiorentini, M. Griffith and A. Charles, inventors; Krypton International SA,assignee; European Patent 0645226A2, 1995.

21. C. Fiorentini, M. Taverna and T. Griffiths, Journal of Cellular Polymers, 1994,13, 5, 361.

22. CarDio Newsletter No.4, September 1997, Article no. 97037.

23. R. G. Eiben, Presented at Utech ’96, The Hague, The Netherlands, 1996, PaperNo.31.

24. Innovations, 1998, October, 801, 15.

25. Novaflex Multistream, Hennecke leaflet PI125, October 1998.

26. B. Kanner, T. G. Decker and G. Thomas, Journal of Cellular Plastics, 1969, 5, 1,32.

27. B. Kanner, W. G. Reid and I. H. Petersen, Industrial Engineering Chemistry,Product Research & Design, 1967, 6, 2, 88.

28. B. Arkles, Silanes & Silicones catalogue, Petrach Systems, Bartram Road, Bristol,PA, 19007, USA, 1987, p 87.

29. B. Arkles, Chemtech, 1983, 13, 542.

30. X. D. Zhang, C. W. Macosko, H. T. Davis, A. D. Nikolov and D. T. Wasan,Journal of Colloid and Interface Science, 1999, 215, 2, 270.

31. K. Yasunaga, R. A. Neff, X. D. Zhang and C. W. Macosko, Journal of CellularPlastics, 1996, 32, 5, 427.

32. L. E. Scriven and C. V. Sternling, Nature, 1960, 187, 186.

33. M. J. Owen, T. C. Kendrick, B. M. Kingston and N. C. Lloyd, Journal of ColloidInterface Science, 1967, 24, 2, 141.

111

34. X. D. Zhang, C. W. Macosko and H. T. Davis in Polymeric Foams, Ed., K. C.Khemani, Chapter 9, ACS Symposium Series 669, American Chemical Society,Washington DC, 1997.

35. J. P. Armistead, G. L. Wilkes and R. B. Turner, Journal of Applied PolymerScience, 1988, 35, 3, 601.

36. M. W. Creswick, K. D. Lee, R. B. Turner and L. M. Huber, Presented at the SPI31st Annual Technical/Marketing Conference, Philadelphia, PA, 1988, p.11.

37. A. Weier, G. Burkhart, M. Klincke, Presented at Industrial Applications ofSurfactants IV, Suffolk, UK, 1998, p.260.

38. G. Burkhart, R. Langenhagen and A. Weier, Presented at the Polyurethanes Expo’98, Dallas, TX, 1998, p.129.

39. S. B. McVey, B. L. Hilker and L. F. Lawler, Presented at the Polyurethanes FoamAssociation conference, San Antonio, CA, 1995.

40. S. B. McVey, B. L. Hilker and L. F. Lawler, Presented at Utech 96, The Hague,Paper No.38.

41. G. Burkhart, V. Zellmer and R. Borgogelli, Presented at the SPI PolyurethanesExpo ’96, Las Vegas, NV, 1996, p.144.

42. G. Burkhart, R. Langenhagen and A. Weier, Presented at the Polyurethanes Expo’98, Dallas, TX, 1998, p.129.

43. A. Weier, G. Burkhart and V. Zellmer, Presented at the SPI, Polyurethanes ’94Conference, Boston, MA, 1994, p.202.

44. J. H. Hildebrand, J. M. Prausnitz and R. L. Scott, Regular and Related Solutions:the Solubility of Gases, Liquids and Solids, Van Nostrand Reinhold Co., NewYork, 1970.

45. No inventors; Th. Goldschmidt AG, assignee; German Patent 19, 726, 653, 1999.

46. R. J. Boudreau, Modern Plastics, 1967, 44, 5, 133.


112


113

This chapter highlights three aspects of manufacturing technology, which have recently broughtsignificant benefits to the producers of automotive seats, reinforced parts and generic mouldeditems. These different manufacturing areas have been grouped here under a single title inorder to cover a wider spectrum of reader’s interests and to illustrate the broad area ofimprovement and innovation that can still be found in this steady-growing segment of industry.

3.1 Industrial Solutions for the Production of Automotive Seats UsingPolyurethane Multi-Component Formulations

Cannon has developed complete systems for the manufacture of moulded polyurethaneautomotive seating elements made with varying combinations of several raw materials.The resultant cushions – although produced in a random sequence on the same mouldingline - are characterised by mechanical properties tailored to the specific application ofeach part [1].

This section describes the components of this technology which are:

• dedicated multi-component, high-pressure metering machines with closed-loop controlof both the output and pour pressures

• dedicated mixing heads capable of processing six components (all with high-pressurerecirculation), including low-viscosity toluene diisocyanate (TDI) and water-based additives

• dedicated mould-handling systems for a flexible manufacturing concept based onjust-in-time methods

3.1.1 Market Requirements

A recent analysis (internal Cannon Report) of the automotive seat-manufacturing marketsegment, to identify the future needs and trends of this group highlighted a number ofinteresting considerations for a producer of polyurethane processing equipment.

3 Polyurethane Processing: Recent Developments

Max Taverna

114


First, the number of suppliers is getting smaller and smaller, due to the extensive mergerand acquisition campaign run in recent years by the two largest players (Lear Co., andJohnson Controls). This concentration creates an opportunity for making strategic choicesregarding the ‘make or buy’ decision with respect to manufacturing solutions. Theselarge manufacturers have grown significantly thanks to their own successful developmentof both chemical formulations and moulding plants. Providing them with new machinerysometimes means accepting their request to manufacture a concept that was developedby their own engineering department.

The second aspect concerns the growing demand for simpler, high-efficiency plants, witha reduced number of operators and a high degree of operating flexibility. It must bepossible to produce several different parts on the same moulding line so that, in the caseof a sudden change in production plans, complete projects can be switched from oneproduction line to another within a very short time frame. This increased flexibilityrequires careful design of the metering equipment, mix heads and mould carriers becausethey must be able to perform very different tasks in sequence or by project. A specificrequest involves the potential to process multi-component formulations, where a widerange of foams can be produced on one machine.

The third aspect, potentially in conflict with the previous concept, involves the developmentof families of formulations based on specific chemicals: all diphenylmethane diisocyanate(MDI), MDI/TDI in various percentages, all TDI and special polyols. This means dealingwith very different demould times and moulding conditions, that render a generic ‘seatplant’, that was still so useable just a few years ago, obsolete. In this case, specific packagesmust be available, which conflicts with the concept of high flexibility expressed above.

3.1.2 Dedicated Solutions: Metering Equipment

A high-pressure multi-component metering unit was designed for automotive seatingproducers, with a pump-dosing system capable of precisely dosing TDI and other low-viscosity components. Output adjustment on-the-fly and closed-loop control, easy toobtain with piston-driven metering units, were engineered for pump-driven machines aswell. The number of main components (polyols and isocyanates) can be set, as necessary,since the design is modular: each dosing line includes a dedicated tank, its temperaturecontrol system, high- and low-pressure filters, recirculation valves, a dosing pump andmotor and a portion of the control panel (see Figure 3.1).

The modular design of this unit allows for easy addition of further components to aformulation. Storage, metering, temperature-time control modules and feed lines areassembled in modules and can be added as needed. Computerised process control easilyintegrates the new chemicals in the formulation.

115

Polyurethane Processing: Recent Developments

Two methods were used to enable fast, precise changing of the output and to ensureclosed-loop control of the machine, which is essential to guarantee continuous, accurateoutput of low-viscosity chemicals:

• via an inverter mounted on the pump motor

• via a programmable step-by-step motor that changes the setting of the pump.

The output control on the closed-loop machine operates continuously, comparing the setoutput value of the inverters with the real output as measured by the volumetric flowtransducers. The system enables pouring only when the parameter is within the limits setby the operator via the keyboard and when it is possible to change the pump speed in lessthan half a second.

The unit also includes maintenance and alarm menus. Through the maintenance menu,it is possible to set limitations on the following parameters:

• number of shots

• material consumption

• working hours

Figure 3.1 A high-pressure multi-component metering unit

116


Once these set values are reached, the corresponding warning message informs the operatorto perform the required maintenance operation.

The programmable stepping device is very precise - the pump can be set at 256 discretepositions - and can be easily programmed through the machine programme logic controller(PLC). Since the switching time between formulations can be as low as 0.6 of a second,multiple ratios can be set within the same pour program to produce multi-hardness andmulti-density foams with repetitive results.

3.1.3 Dedicated Solutions: Mixing Heads

Cannon has designed a new mixing head, capable of mixing six components, each withindividual recirculation control. The new mixing head was specifically designed to meet theneeds of the automotive seat producers who wished to mould TDI-based flexible foams witha maximum of flexibility in their formulation. Several pure ingredients are kept separate upto the point of injection and it is possible to operate each stream on demand, provide high-pressure recirculation and avoid the contamination of components (see Figure 3.2) [1, 2].

Figure 3.2 The new Cannon Ax Head, specifically designed for multi-componentpolyurethane formulations. It can mix up to six chemicals all with a high-pressure

recirculation feature.

117


Although this solution is currently available on the market, competitive mix heads arelimited to four streams, their dimension and weight require heavy-duty pour robots whileproviding speed and pour pattern limitations during operation and they have complexpressure set-up and regulation procedures.

3.1.3.1 Multi-Component Operation

The innovative aspect of this mixhead is that TDI - and eventually a TDI-compatiblesixth component - are fed axially into the mixing chamber through the small piston thatcleans the mixing chamber. The other four components, polyols, other catalysts, additives,flame retardants, are fed radially into the mixing chamber. A seal on the small pistonprovides a permanent separation between the polyol and TDI feeding areas so that anycrossover of the low-viscosity components is avoided.

The main advantage of the new mixing head is its reliability. It can operate for millionsof shots because of the remote position of the TDI feeding area. TDI is fed in at a pressureof only 1 MPa, with perfect mixing efficiency (see Figures 3.3, 3.4 and 3.5).

Figure 3.3 Section of the mixing-chamber’s cleaning piston. Four chemicals are fed radiallyin the mixing chamber, while two are fed through a hole drilled in the cleaning piston.

118


Figure 3.5 When the injection signal is received, the mixing chamber’s pistonretracts, the four components fed radially meet the two that are fed axially

through the mixing chamber’s piston. The mixed blend reaches the discharge duct,which is positioned at an angle of 90° to the mixing chamber and leaves the head

through the pouring hole.

Figure 3.4 During the high-pressure recycle phase all six components flowin the grooves carved in the mixing chamber’s piston and return to their

storage tanks.

119


Defining ‘Formulation #1’ as being composed of polyol #1 and isocyanate and ‘Formulation#2’ as polyol #2 and isocyanate, an example of a working sequence is as follows:

1. Pour ‘Formulation #1’: polyol #1 and isocyanate recycle through the mixing chamber’spiston grooves and are poured in the open mould

2. ‘Formulation #2’ remains dormant since polyol #2 continues to recirculate through ahigh-pressure nozzle positioned close to the mixhead

3. The formulation change is implemented by:

• closing the mixing chamber’s piston (the self-cleaning piston is kept open)

• opening the polyol #2 high-pressure nozzle and closing the polyol #1 nozzle

• re-opening the mixing chamber’s piston with the new ‘Formulation #2’

The maximum response time, which occurs between two consecutive shots with differentformulations, is 0.6 seconds.

3.1.3.2 Pour Pressure Control

The pressure control of the component injectors, which is very important for ensuring theproper mixing of the different liquid streams, has been achieved via three different approaches:

• closed-loop control of the injection pressure via hydraulic servo valves and feed-back control from the pressure gauges

• open-loop control via hydraulic servo valves

• fixed-position control via various hydraulic valves, preset at different values

The package supplied with the head to achieve pressure control (see Figure 3.6) includes:

• hydraulically-operated nozzles to be installed on the mixhead for the componentsthat are fed radially; each nozzle has a double function:

- selection of the component to be used in the specified formulation

- control of the re-circulation/injection pressure during the pour

• hydraulic unit for operation of nozzles

• valves for control of the oil flow; each valve has three positions (off, injection pressure#1 and injection pressure #2)

120


• high-performance recycle stream distributors with relevant pressure control tomaintain non-specified components in high-pressure recirculation, ready to be injectedwithin 0.6 seconds

• one set of high-pressure flexible hoses to connect the proportional valves to theinjector nozzles

• appropriate controls

3.1.3.3 Variable Geometry Mixhead

To ensure proper mixing conditions for a wide range of formulations - which can differin chemical composition, viscosity, ratio and output - it is important to provide appropriatebackpressure in the mixing area. This can be obtained with the adjustable geometryregulation of the mixing area: the mix head’s larger piston can be mechanically set topartially block the outlet of the mixing chamber when it is fully retracted. This occlusionincreases the turbulence in the mixing area, causing more efficient mixing to be obtained.

Figure 3.6 Scheme of the circuit used to guarantee a closed-loop control of the pouringpressure of each chemical.

121


3.1.3.4 Advantages

A very compact design: its outside dimensions of 40 x 20 cm and 22 kg weight translateinto a high degree of manoeuvrability.

• The pour pattern can be defined very precisely, according to the design of the mouldand the position of any inserts.

• The pouring operation can be executed in an open mould without fear of collisionwith the upper mould half.

• The robot carrying the head can achieve high acceleration and speeds withoutmechanically stressing the moving elements.

The large number of chemical components that can be handled simultaneously providesa high degree of flexibility in formulations, allowing for optimum use of the mouldingline. Several different parts can be produced on the same moulding line and differenttypes of foam can easily be produced in random sequence without forcing the operatorto use pre-defined sequences of moulds.

3.1.4 Dedicated Solutions

3.1.4.1 Foaming Robots

The basic features of a typical Cannon double-arm pour robot, normally supplied tocarry two heads (see Figure 3.7), are:

• two-axes Cartesian robot with two arms moving independently of each other,

• arm movements (along the ‘X’ and ‘Y’ axes) achieved via a rack-and-pinion systemdriven by electronic variable-speed motors,

• racks fitted with position encoders to detect the relevant ‘X’ and ‘Y’ coordinates.

The Technical Specifications include:

- maximum speed for X axis: 2.4 m/s- maximum speed for Y axis: 2.4 m/s- maximum acceleration: 3.5 m/s2

- maximum load on the arm: 120 kg- maximum ‘X’-axis stroke: 250 cm- maximum ‘Y’-axis stroke: 115 cm- minimum distance between 2 arms: 55 cm- available working area: 307 x 115 cm

122


3.1.4.2 Mould Carriers

The concept of the dedicated flexible foam moulding line relies on a mould carrier havinga simple solid structure, with a minimum number of mechanical parts, which requiremaintenance. It consists of two upper and two lower sections linked by means of ahooking system. The presence of two separate lower platens allows the operator tocompensate for any difference in mould thickness without incurring any problem bymounting two moulds on the same carrier (see Figure 3.8). Insertion, centring and fixationof various moulds within the mould carrier are achieved by means of tensioning screws.

Each mould carrier is usually provided with two pneumatic cylinders for opening andclosing. Upon demand, the closing system can be actuated hydraulically or mechanically.

Mould carrier tilting, for optimum evacuation of air during the filling phase, isaccomplished with a cam system. Each mould carrier can have two tilting positions:either horizontal or frontally inclined by 20 degrees (see Figure 3.9). It is possible toprogramme different angles of opening and tilting to correspond with the type of mouldcurrently being used.

Mould identification – necessary in order to transmit the appropriate mould/formulationinformation to the dosing system - is achieved with a code reader. The electrical controls

Figure 3.7 A double-arm Cartesian robot, designed to perform multi-hardness PUfoaming in open-mould pouring processes for automotive seat moulding.

123


Figure 3.8 Independently-moving lower platens allow for the use of different thicknessof moulds as a double mould carrier.

Figure 3.9 A different degree of tilting can be obtained for each mould carrier, tooptimise the evacuation of air from the moulds during the filling/

polymerisation phase.

124


for the moulding line are interfaced to the metering system controls in order to co-ordinate all the working processes.

The typical working sequence can be defined as follows:

• mould identification: the control system on the moulding line reads the code comingfrom the field

• pour programme selection: inside the mixhead manipulator control unit, a schedule- assigning a specific pour programme to the different moulds - has to be loaded

• pour programme execution: the manipulator control unit sends the input to themachine control unit; when the mould under consideration requires multi-hardnessfoaming, both mix heads are activated, while in the case of a single-hardness seat,only one mixhead is activated

Stroking of the Platens

As stated previously, the lower half of the mould carrier is composed of two independentplatens. They are fitted with a guiding system to ensure maximum parallelism betweenthe platens.

In order to ensure equal clamping pressure and complete closure of the moulds over theentire parting line, pneumatically inflated tubes stroke each platen. They are inflatedafter the mould carrier has been closed and deflated just before opening. The total strokeis a few centimetres. The clamping force afforded by this system is 12,700 kg, using aworking pressure for the air bags of 0.3 MPa.

Mould Temperature Control System

Individual thermoregulator units are used, one for each press. They are mounted on therear of each mould carrier. This solution affords complete autonomy to each press, makingit possible to control the temperature of the moulds prior to being placed on the productionline. Subsequently, when a press is placed on the moulding line, it is ready to commenceproduction.

The thermoregulators work in a closed circuit; the relevant refilling has to be performedoff-line. The control panel for each thermoregulator is mounted on the operator side foreasier access.

125


Mould Carrier Exchange

The mould carrier concept for fixing and centring is designed for quick mould carrierremoval or exchange.

The mould carrier can be removed as follows:

• disconnect the electrical plug (positioned on the operator side)• disconnect the air by removing the relevant joint mounted on the mould carrier• remove the complete mould carrier (with its thermoregulator) using a fork lift

The estimated time to perform this operation is less than five minutes.

Service Station

A dedicated station where service operations on moulds and mould carriers can beperformed, without disturbing the production cycle, is foreseen for each moulding line(see Figure 3.10).

Figure 3.10 All maintenance and setting operations on moulds and mould-carriers canbe executed off-line, in a dedicated service station that performs all the line’s

movements and operative functions.

126


The service station comes with one complete mould carrier on its freestanding frame. Itis equipped with a dedicated thermoregulator for warming of moulds destined to beinserted on the moulding line, along with all the required safety fences/light barriers,controls, and connections for air, water and electrical power.

3.1.4.3 Transport Systems

Cannon provides different mould-carrying systems, each customised to meet the customer’sneeds. The most common solutions are conveyors and turning tables. In the last part ofthis section some innovative concepts, which provide a more compact layout, a minimisedinvestment requirement and a maximum degree of flexibility in a changing manufacturingscenario, are presented.

Oval Conveyor

The conveyor is made of a central steel frame, which is positioned on the floor. It iscomposed of a series of straight-line modules with a curved module at each end andcomes complete with a number of carriages running on an oval steel track (also mountedon the floor). On each carriage, mainly composed of a rigid steel frame positioned onfour pivoting wheels, a single lid mould-carrier is mounted. Four idle wheels, runningalong a central guide plate, maintain the system on its set path (see Figure 3.11).

Continuous movement is achieved by means of one drive system being placed betweenevery two carriages. Each driving system has two entrainment wheels running along theguide plate. The entrainment wheels – operated via an AC motor – are maintained intraction along the guide plate by a spring. An inverter is used to adjust the speed from 4m to a maximum of 9 m/min, while the AC motor is fitted with a brake so that, in caseof an emergency, the plant can stop within a few centimetres of travel.

This system can easily be expanded to incorporate any new business obtained for thatrange of moulded products. Extensions are achieved by fitting an even number of newcarriers to the line and extending the supporting frame accordingly.

Mould Carrying Systems: FlexiDrum

Developed in the early 1980s as a revolutionary tool for the production of foamedrefrigerator doors, the well-known concept of the rotary polymerisation system is nowbeing proposed as a compact and simple mould-carrying system for the production ofautomotive seats (see Figure 3.12).

127

Figure 3.11 A modular oval conveyor for flexible foam moulding. It can be easily extended,adding modules of track and pairs of mould carriers, when higher productivity is required.

Figure 3.12 Flexidrum, a compact moulding line for medium-low volume ofautomotive seat production.

1a Service station: demoulding, cleaning, insert positioning, release agent application;1b Open mould view of 1a; 2 Foaming station, with robot on a platform;

3-6 Polymerisation stations; 7 Six arm rotating structure to support mould carriers; 8 Centralcollector of signals, air and warm water; 9 Supporting structure (and rotor, not shown)


128


Six or more mould carriers can be fixed on the surfaces of a wheel that rotates vertically –similar to a Ferris wheel, rather than horizontally, as with a merry-go-round. The openingand closing movements can be hydraulic or pneumatic. Mould conditioning can beaccomplished using a rotating collector that feeds each mould carrier with water at thedesired temperature. Manual service operations are performed on the lower mould half inthe first station. Foaming is achieved in the subsequent station using a platform-mountedrobot. Polymerisation takes place in the remaining four (or more) elevated stations as theyare passed through a compact suction hood that removes the escaping fumes. Severaladvantages can be highlighted for this new version of mould carrying system:

• simplified mechanical construction that does not incorporate wheels which rollalong the floor picking up and transporting scrap foam and/or dirt

• optimised layout which will work even when space is limited

• reduced power cost thanks to its vertical layout which minimises the suction area,requiring only one fan for extraction of the fumes

• limited number of operators required

Carousel for In Situ Foaming

A number of dedicated moulding lines have been designed for the production of in situmoulded foams. In situ moulding technology adds a delicate operation to the list ofconventional operations (mould cleaning, release agent spraying, insert positioning,foaming and demoulding) that must be executed to mould a standard foamed item: themanual positioning of the textile container into which the foam will be dispensed. This isa delicate operation that requires some time, yet should not penalise the cycle.

A practical solution consists of a carousel line with a row of service positions where theoperators can work on moulds that have been temporarily taken off-line (see Figure 3.13).When the press leaves the curing area, it passes in front of the first free operator and isautomatically disengaged by the dragging system. The moulds can be serviced, taking allthe time required, then, when the textile inserts have been positioned, the carrier can be re-inserted in the first available position in the line.

Multi-Hardness and Multi-Density Foams with Natural Carbon Dioxide

The availability of multi-component mixheads with variable geometry opens the path toanother interesting option: the addition of natural carbon dioxide for expansion of the

129

foam with an environmentally friendly blowing agent and reduction of the foam densityby as much as 20-25%. With this new mixhead, it is possible to use two differentapproaches to add natural carbon dioxide to the formulations:

• CannOxide [3-12] - a technology developed to meter natural carbon dioxide at thepoint of injection, at the desired percentage, into one of the polyol streams

• EasyFroth [1, 2, 4, 11, 12, 13,14] for carbon dioxide – a technology that allows thepremixing of given percentages of blowing agent in one of the two components,usually the isocyanate.

Both methods are currently in industrial production, each having operating andinvestment ‘pro’s and con’s’ that must be evaluated according to the production volumesand flexibility required. For example, a car seat with three different hardnesses can beproduced using two different formulations for the hard and soft parts - (see areas ‘A’and ‘B’ in Figure 3.14) - and incorporate some natural carbon dioxide to decrease theseat density below the thighs (area ‘C’).

Figure 3.13 Oval moulding line with 16 double mould carriers, designed for in situfoaming of textiles.

1 to 4 Service stations, where mould carriers are taken temporarily off-line formould service; 5 Spare station; 6 Foaming station; A Pouring robot; B Metering

equipment, on mezzanine; 7 to 16 Polymerisation stations; C Mould carrierchanging station; D Suction hoods


130


3.2 ‘Foam & Film’ Technology - An Innovative Solution to FullyAutomate the Manufacture of Automotive Sound Deadening Parts

Until now, one of the major limitations in the polyurethane moulding process has beenthe necessity to interrupt the working sequence between each moulding, to remove foamscrap and apply release agent prior to foaming. The introduction of ‘Foam & Film’technology makes manual intervention unnecessary, removing the one factor that hasalways been a major weakness when working with polyurethanes in comparison to otherinjected or extruded plastics [15-19].

The main idea behind this new approach consists of thermoforming a thermoplasticor polyurethane film as part of the moulding sequence. By using a vacuum effect,this film adheres perfectly and smoothly to the mould cavity, without any creases orwrinkle formation. The mould is equipped with a dedicated frame device, specificallydesigned to hold the film. A heating system ensures that the film reaches the desiredtemperature prior to the thermoforming phase and subsequent injection ofpolyurethane into the mould.

Figure 3.14 Triple hardness automotive seat

A – Soft foam; B – Hard foam; C – Soft and lower-density foam, blown with naturalcarbon dioxide

131

Several industrial applications have been found and a number of fully automated plantscurrently incorporate this technology. The manufacture of sound deadening parts forthe automotive industry represents one of the most exciting applications for thisinnovative technology.

3.2.1 The Problem

Mould cleaning is required because of the chemical nature of the polyurethane process.The objective in developing the ‘Foam & Film’ technology was to eliminate the tedious,time-consuming manual operations, which must be performed on the moulds as part ofa discontinuous polyurethane foaming process.

The need to properly vent the mould, to avoid air entrapment, often results in a thinflash of polyurethane being formed around the moulded part, which needs to be manuallyremoved afterwards via a simple trimming operation. These manual operations involveboth cleaning of the mould and spraying of the release agent.

It is necessary to demould parts in the shortest possible time, when the foam is not yetfully polymerised, because the thin cross-section of the flash makes them very fragileduring the first few minutes following the demould. Often small pieces of flash willbreak off when the part is extracted, falling back onto the mould. If they are not removedfrom the mould surface, they could mar the surface of the subsequent parts or they couldallow rising foam to leak from the mould if left along the seal. More extensive cleaningoperations are required every few shifts in order to remove deposits of release agentfrom the mould surface and there require production to be stopped.

Various technologies have been applied to minimise the required downtime, but so far avalid solution has not been found. The adhesive force between polyurethane and metal(either aluminum or steel) requires a release agent to be sprayed onto the moulds everysingle cycle (or every few cycles) to enable easy removal of the part. In addition, variousinternal mould release technologies are available, but they do not apply to all formulations.

As stated previously, these operations take a long time, are expensive and are typicallyexecuted manually and it is impossible to have a fully automated line without getting ridof them. To create a fully automated line, both the need to manually clean the mould andthe need to spray release agents on it had to be eliminated.

3.2.2 The Approach to a Solution

The new approach to polyurethane moulding is to introduce a thermoplastic orpolyurethane film thermoforming process on one or both halves of the mould as part of


132


the overall sequence. By using a vacuum, this film adheres perfectly and smoothly to themould cavity, without any creases or wrinkle formation.

The system’s key concept is a specially designed and patented frame, integrated into themould, which is specifically designed to hold the film in position during this thermoformingprocess. This frame can be either two or three dimensional, to better follow the shapeand the cavities of the mould (see Figure 3.15).

Figure 3.15 ‘Foam and Film’ concept

1. Unrolling and cutting plastic film to required length; 2. Infrared (IR) heating of filmon holding frame; 3. Thermoforming the film in the mould cavity; 4. The same process(1-3) is carried out in the other half of the mould at the same time; 5. Robotised foamdeposition in mould; 6. Mould closed for expansion/polymerisation. Parts are removed

‘wrapped’ in the film.

The frame receives the film after it is unrolled by pinchers and cut dimensionally. Itthen moves it in front of an infrared lamp for several seconds to heat it to the requiredtemperature. A control system ensures that the film reaches the desired temperatureprior to the thermoforming phase. Special infrared heaters, Cannon’s MVL heaters,are used to ensure high efficiency and very low thermal inertia. Once the film is heated,the frame moves into the mould where the film, held in place along its four edges, isvacuum-formed onto it. The use of the frame during the vacuum forming prevents thefilm from folding or wrinkling (see Figure 3.16).

133

3.2.3 The Film

There are two different types of ‘Foam & Film’ processes available, based on the differenttypes of film used: adhesive and releasing.

• In the adhesive type, the film sticks to the part and is unloaded with it, granting anaesthetic finish and a waterproof covering (see Figure 3.17).

• The releasing film, on the contrary, remains vacuum-formed to the mould for severalshots (5-15, dependent upon the process and the materials utilised) and it is thenreplaced when it begins to wear.

Obviously, with the adhesive type, the film has a high adhesion coefficient with polyurethane,while in the releasing version the poorer the adhesion, the better. In both cases, the filmprevents the mould from ever being put in contact with the polyurethane foam. That iswhy both cleaning and spraying are no longer necessary and all operations can be automated.

Figure 3.16 A patented frame holds the releasing film in position so that whenthermoformed over the mould, it is kept under tension and does not form wrinkles.


134


3.2.3.1 Adhesive Film Solution

Basically, three types of film can be used in the adhesive Foam & Film process:

• Thermoplastic polyurethane (TPU)

• Polyethylene (PE)

• Thermoplastic film (TP)

TPU is a good film for both cold and hot processes, providing good adhesion to the foamand excellent mechanical properties but the cost is high (around 0.65 US$ per squaremetre). PE film can only be used in cold processes, it requires a special treatment forperfect adhesion and it gives an overall poorer performance but it has the advantage ofbeing very inexpensive. TP film is only suitable for hot processes, it requires no treatment,and it guarantees good adhesion at a low cost (less than 0.2 US$ per square metre).

This last type of film is what has been developed in depth, obtaining very good tear andimpact resistance, flexibility, elongation and welding ease. This is the solution that hasbeen used for most automated Foam & Film plants supplied so far. Using this film,General Motors makes sound-deadening parts with fully automated equipment that hasa productivity of 8000 parts/day, running 24 hours per day with zero operators [17, 18].

3.2.3.2 Releasing Film Solution

The second type of film, the releasing one, sticks to the mould and not to the polyurethanepart. In the releasing Foam & Film technology, there are two main types of film that can be

Figure 3.17 Two shock absorbing inserts for automotive doors moulded in semi-rigid EA-foam with the ‘Foam & Film’ technology as they appear immediately after demoulding. The

film protects the foaming against moisture and degradation. No release agents or mouldcleaning operations are needed to produce this part in a highly automated moulding plant.

135

used: PE and TP. PE, once again, is very common and inexpensive, but gives poor releasefrom the foam, which minimises the benefits of this technology. TP film provides an easyprocess with good release from the foam for several shots at a very reasonable price.

3.2.4 Industrial Applications

A good example of such a continuous moulding process that incorporates ‘Foam &Film’ technology is a system developed for the production of industrial vehicle carpets.These products are usually made out of a sandwich of two or more layers of differentmaterials. Individually, they provide different features: aesthetics and function (a textilecarpet or a synthetic mat), sound deadening (a layer of polyurethane foam) and protective(a cheap layer to protect the foam from moisture and degradation).

INSOTEC, a technology designed by Cannon for the manufacture of sound-deadeningautomotive components, offers a variety of manufacturing alternatives that providesignificant benefits such as improved quality parts, reduced costs, shorter process times,as well as regular and consistent production cycles (see Figure 3.18).

Figure 3.18 A large full-automated production plant for truck floor-covering matsusing the ‘Foam & Film’ technology. The parts are made of thermoformed PVC (or

thermoplastic elastomer) foam-backed with PU, which is kept separate from themould with a film (adhesive type).


136


These excellent results have recently been improved with the introduction of the new‘Foam & Film’ technology. The technology makes manual intervention unnecessary,removing the one factor, which has always been a major weakness with polyurethaneprocessing and added a significant cost to the parts that are perceived to be an economicalcomponent of a vehicle. These components have a surface area of about 3 m2 and consistof a surface layer of polyvinyl chloride (PVC) or thermoplastic elastomer (referred to asthe ‘heavy layer’), an intermediate sound-deadening layer of flexible, medium-to-lowdensity polyurethane foam and a lower thermoplastic film. The film is designed to preventthe formation of flash during the moulding process, eliminate permanent residue usuallyleft in the mould, and to act as a release agent (once the part is in the vehicle, this filminhibits water absorption, which is a very frequent problem with industrial vehicles). Inthe past, polyurethane films, which were strong mechanically and performed well butwere quite expensive were used. With the new ‘Foam & Film’ technology, it is possibleto replace this polyurethane film with a thinner, less costly thermoplastic one, resultingin high quality production at a lower cost.

The system is composed of two shuttle-bed clamps served by one metering unit, whichdispenses the pre-heated heavy layer in the mould. Pre-heating is carried out using aspecial infra-red heater, which incorporates easily adjustable, special low thermal-inertiaresistances. When a vacuum is applied to the mould, the material adheres to the lowermould half, taking on its shape and embossed design.

The protective film is automatically unrolled from an overhead source using a verticaltraversing frame with pinchers that pulls an appropriate length of film over the frameand cuts it to length. The frame is positioned over the edge of the mould. Another bankof heaters slides laterally in front of the mould and warms the film to the correct formingtemperature; a vacuum is applied at the end of the heating phase to conform the film tothe mould.

The use of two presses, as opposed to one, means that slack periods are eliminated anduse of the cycle time is maximised. While the film and heavy layer are being placed in oneof the moulds and pre-heated, foaming and polymerisation are taking place on the other.

The polyurethane dosing and foaming section of this plant is equipped with a mouldingtechnology that allows chlorofluorocarbon (CFC) blowing agents to be replaced withliquid carbon dioxide. This helps reduce the density of the polyurethane considerablyand thus saves on material costs.

The overall production time is just two and a half minutes per part. Being a two-station plant, this equipment can produce close to 50 finished parts per hour – withouta dedicated operator.

137

3.2.5 Applications

This technology has already been used on turnkey plants for sound insulation parts,carpet back foaming and seat cushions. No limitations for this technology are foreseenand, actually, ‘Foam & Film’ can be implemented in almost every kind of polyurethaneprocess needing either release agent or a cover/surface film.

The system requires the presence of an open mould where the film-holding frame can beinserted to position the film prior to the vacuum-forming phase. Obviously, this technologyis more easily applied to new equipment and new moulds since, most of the time, existingmoulds must be modified to provide the vacuum and hold the frames. When one surfaceof the part is covered by an aesthetic or functional layer (carpet, plastics, etc.), obviouslythe film is only applied to the opposing side where the foam would be in contact with themould (see Figure 3.19).

Figure 3.19 Turnkey installations using the ‘Foam & Film’ technology are producingnumerous parts for the automotive industry.


138


3.2.6 Advantages

‘Foam & Film’ offers significant benefits, such as quality parts, reduced costs, shorterprocess times, as well as regular and consistent production cycles.

• operator intervention is no longer required for mould cleaning and application ofrelease agent; consequently this leads to increased productivity, uniformity and thecycle regularity which comes with a completely automatic line - a fully automatedfoaming process is now possible.

• polyurethane or TP film can be used, the latter being thinner and cheaper but givingthe same performance.

• films which adhere to the product can be used, becoming an integral part of the finishedcomponent; this can be a very useful feature for non-aesthetic parts that are to bemounted in hidden positions and will benefit from this extra protection against humidity,oil, aggressive chemicals and foam-aging agents such as oxygen or other gases.

• films, which adhere to the mould, can be re-used several times as a substitute forrelease agent.

• the availability of a wide range of film sizes means no dimensional limitations on theparts to be moulded.

To summarise, ‘Foam & Film’ Technology:

• automates the production of polyurethane moulded parts,

• eliminates mould cleaning,

• eliminates spraying of release agent, saving its cost plus those of all the relevantdispensing equipment and special fume extraction systems (although regular fumeextraction must be maintained for the polyurethane process),

• offers a low operating cost.

3.3 InterWet - Polyurethane Co-injection

The combined use of polyurethane and reinforcing agents or fillers has been commonpractice for a long time. A recently introduced technology - simultaneously injectingfoam and glass fibre in open moulds - did not give satisfactory results, according to someof the early users. Cannon has developed an industrial solution - named InterWet - that

139

improves both manufacturing performances and mechanical properties of moulded parts[20-27]. In addition, it allows for the use of other reinforcing fibres (natural and synthetic)and opens the way to the easy addition of powder and granulate fillers.

3.3.1 Glass-Reinforced Polyurethanes, a Well-Known Technology

Glass-reinforced polyurethane has been used for many years, utilizing differenttechnologies: reinforced reaction injection moulding (RRIM), open-mould pouring overa flat mat (LD-SRIM), closed-mould injection on preformed glassmat-sandwiches (SRIM).Cannon has for many years dealt with glass-handling polyurethane technologies,developing in the early 1970s a special RRIM mixing head able to mix formulationscontaining high percentages of milled fibre, and launching in the 1980s the HE meteringmachines, closed-loop electronically-controlled piston-metering systems able to cope withthe abrasion deriving from glass and mineral charges. Later in the 1980s the Compotecpreformers were introduced, for the production of preformed glass mats required bySRIM - structural-moulding applications.

A new technology has been introduced by Krauss-Maffei, long fibre injection (LFI) [28],that simultaneously pours polyurethane and chopped glass roving on the surface of openmoulds; after the shot the press is immediately closed to allow for the expansion of thefoam, which surrounds all the fibres and produces a lightweight, resistant compositepanel. Most recently, Hennecke has introduced FipurTec [29, 30], a technology wherethe fibre is chopped outside the head and projected into the flow of reacting chemicals,where this touches the surface of the mould.

In essence these systems carry out a job similar to that performed for very many yearsin the polyester ‘chop-and-spray’ applications, the only differences being thatpolyurethane has a different profile of reactivity and viscosity build-up than polyesterand that, at least in one case, the glass roving is conducted through the mixing deviceinstead of being fed outside of it.

The interest behind this technology lies in the fact that a thin, resistant composite partcan be moulded in only one operation. Using preformers requires more equipment, space,investments, etc. In addition, glass roving costs roughly half that of glass mat, and thiscan represent a considerable saving when producing large, highly reinforced parts.

3.3.1.1 Practical Problems That Needed to be Addressed

A few practical problems have been identified in this technology, mainly deriving fromthe design of the mixing equipment currently available. In one case the chopped fibre is


140


fed through a pipe that is co-axially positioned in a conventional ‘L-shaped’ head andthen the mixed resin is conveyed around the pipe that feeds the chopped fibre. In theother case the fibre is projected separately from the polyurethane; in this case the glassmeets the polyurethane only at the end of the head’s final discharge duct. The separationof the fibre and polyurethane before the exit of the head results in wetting of the fibresnot being optimal; this is visible during moulding, where part of the glass gets stuckvertically into the base of rising foam, visibly dry.

The main problem comes from the dry chopped glass that - leaving the nose of the headstill not wet - flies everywhere near the mould surface. This results in negative effects onworkers, on cleaning and maintenance operations, and on part quality. This problemcan seriously hinder the application of this technology.

3.3.1.2 The Cannon Approach

Cannon started looking into this technology in the summer of 1997, at the urging ofsome car parts producers who were unsatisfied with the performances of the existingones. The main objectives of the project were:

• To supply in the shortest possible time a reliable solution for the injection of glass-reinforced polyurethane foams:

with as little as possible an impact on the working environment,

with the best possible quality,

• To provide the lightest possible mixing/dispensing equipment, in order to reduce theinvestment in head-carrying robots,

• To design a multi-purpose solution, not limiting the choice of charge to glass roving only,but including a wider range of natural and artificial fibres, as well as pulverised fillers,

• To provide maximum flexibility in charge feed, so that products with different contentof charges - in the same moulding or between subsequent moulds - can be produced,

• To provide maximum number of automatic checks on the charge’s feeding line, toavoid the troubles experienced during automatic production cycles.

Due to the short delivery time required, it was not realistic to conceive totally newequipment for this application, and it was decided to concentrate on the use of existing,proven pieces of hardware. Performance of the foaming section had been optimised inprevious years, and the main obstacle was the proper handling of glass and other fillers.

141

3.3.1.3 The Technologies Involved

The project was split into three parallel lines of development, to speed up results and makeuse of the experience of various existing specialists for the specific engineering tasks:

• Charge/Reinforcement (handling, dosing and feeding)

• Polyurethane (dosing, mixing and laydown)

• Robotics.

Handling Reinforcements and Charges

This project required a multi-functional solution, able to accommodate various types offibres, fillers and charges; accordingly, the design of the whole handling section would havehad to respect this future necessity. Since the most urgent need called for the use of glassroving, the first development focused on the design of a simple, reliable device to store, meterand chop conventional type glass-roving rolls. A proper technical specification was draftedto provide optimum handling of the types of glass most suitable for this peculiar application.

Various qualities of roving are available on the market for different end-uses: the diameterof the basic threads (bunches of individual glass fibres) and of the final roving (the ropemade with various threads) define the weight and the field of application, while the typeof coating (a thin layer of special resin applied on the fibres after the glass-extrusionphase) determines the best compatibility of the roving with the polymeric matrix.

A specific type of Owens Corning Fibreglass (OCF) roving was selected as the reference,having a 2400 Tex specific weight, i.e., the roving weighs 2,400 grams per kilometre, withhigh ‘softness’ of the thread. Other types would have been easier to cut (or to be broken,since a fibre of glass is not shear-cut - it is bent beyond its critical radius until it breaks), buttheir individual fibres were more agglomerated in the thread and would have been moredifficult to wet. All parts were designed to cope with a wider range of roving.

The glass-feeding device was conceived to be installed over the polyurethane mixinghead, and its design was made in strict cooperation with the team in charge of designingthe head. It is basically composed of:

• A pulling/cutting device: two opposed rolls that rotate and trap between them one ormore roving, pulling them from their storage. A blade - held in one of the two rolls -forces the fibres to bend beyond their critic radius until they break;

• A conducting system: a flow of compressed air that pushes the chopped fibre fromthe cutting place to the mixing head.


142


A few interesting technical features characterise this roving feeding-chopping section:

• A hydraulic motor, to drive the glass-roving pulling rolls, was chosen to provideproper speed, torque and easy variation of the driving speed. Its high power/weightratio made it preferable to an electric motor;

• A special cutting device was conceived, based on a rotating blade-holder (very easyto access and maintain) capable of accommodating up to four blades. By fixing morethan one blade in the proper holders one can chop the roving in different lengths, thelonger size being that equal to the circumference of the rotating blade-holder. Thepart had to be dimensioned to resist the very high operating speed, up to 5,000 rpm,without suffering from the presence of glass debris around rotating shafts. This designproved useful later, in the further development of a variable-length chopping system;

• A glass-guide system using compressed air was designed to take away all the cutroving from the chopping device and bring it through the mixing head down ontothe mould, with the additional task of providing a final purge of the discharge duct.

The Mixing Head

Most of the development work was concentrated in designing a mixing head able to wetall the solid components before they left, to ensure the projection of a very well wettedmixture rather than a blend of liquid only coupled with a dry, flying filler. The mechanismwhich could ensure a thorough wetting of high percentages of solid fillers with a mixtureof polyurethane is to provoke high turbulence in the mixing head’s area where liquid andsolid meet, and then reduce the turbulence to allow for splash-free open-mould distributionof the blend obtained. It was in fact observed that, with heads where the junction betweenliquid and solid was in a turbulence-free zone, the wetting was not optimal.

By using the Cannon FPL head, an existing, two-decades-proven, piece of hardware itwould have been possible to obtain the best available mixture of the polyurethane componentwith a turbulence-free laydown in the mould. Its internal geometry is L-shaped, with twocylindrical chambers of different diameters connected at a 90º angle. The turbulent flowcreated by impinging of two components in the small mixing chamber is quickly convertedinto a smooth laminar flow as the chemicals are diverted through 90 degrees downwardsinto a large discharge duct and then leave the head. At the end of the pour the smallermixing chamber is sealed and automatically cleaned using a hydraulic piston.

By using the original mixing chamber for the liquid components and boring a hole throughthe main cleaning piston - the plunger - so that a solid component could be fed through it,a wide range of fillers can be used (see Figure 3.20). When the mixing operation begins, the

143

main piston is retracted until it clears the discharge duct at the junction of the two cylindricalchambers. In a controlled sequence, the small piston sealing the mixing chamber retractsand the two chemical components are fired at high pressure against each other throughinjectors in the mixing chamber before leaving it through the discharge duct.

The metering and feeding operation of the solid component from the head’s upper part iscontrolled mechanically, assisted by an intense flow of compressed air. Its feed issynchronised with the arrival of the liquid blend from the mixing chamber. The solidcomponent meets the liquid blend and together they are co-injected into the mould.Following co-injection, the small piston seals the mixing chamber, the large piston sealsand cleans the discharge duct and the feed of solid component is interrupted, althoughthe flow of air is maintained for a short time in order to clean any residual polyurethane- which could interfere with the next pouring operation - from the nose of the head.

Figure 3.20 The Cannon FPL head used for InterWet incorporates in its upper sectionthe glass cutting mechanism and relevant safety checks.


144


The innovative concept in this solution is that the solid component meets the liquidformulation just in front of the mixing chamber, where the turbulent flow of the mixtureis directed through a 90° angle downwards creating a laminar flow. In this way, thekinetic energy from the pressurised liquids is used to wet the stream of solid componentthoroughly and efficiently at a point just 20-25 mm from the point of impingement. Theblend of polyurethane and filler leaves the head already well blended, and there is noevidence of ‘flying glass’ out of the head (see Figure 3.21).

This special design and technique ensures that the solid component is thoroughly internallywetted, hence its name, InterWet.

Figure 3.21 Open-mould co-injection of PU and glass fibre is executed at high speedand in a clean working environment, with an InterWet machine.

145

Robotics

The development of a proper device to carry the mixing head during the pouring operationover open moulds required relatively short time. Basic requirements were multi-axescapacity, speed, precision and easy programming.

The designers concentrated on the use of commercial polar robots, operating up to 6axes at high speed and precision (see Figure 3.22).

The limited weight of the Cannon FPL mixing head allowed for the selection of a medium-sized model able to carry on the wrist 125 kg of payload at up to 2 m/s, with accelerationof 3 m/s2. This robot was fitted with two storage boxes for the roving, mounted on theelbow area (actually this model allows to install there up to 3 glass-roving rolls, for amaximum 60 kg payload). This is a very convenient glass-storage solution for laboratoryand small production applications. When high quantities of glass are required, a solutionable to reduce rolls-replacement times must be provided. A proper device has been designedto accommodate the 1,000 kg of glass demanded for industrial heavy-duty tasks. In thiscase the glass rolls - that could be cascade-joined one another - must be positioned theclosest possible to the head, so that the way the roving must take to reach the chopper isnot restrained by long guides.

Figure 3.22 One of Cannon’s R&D units devoted to the development ofInterWet technology.


146


First Results

The assembly of the pieces of hardware was completed in less than eight weeks from thebeginning of the project, and immediately trials were scheduled in the laboratory using a flattest-plate mould.

A conventional Cannon HE - a piston-driven, electronically closed-loop controlled dosingunit available in the laboratory - was connected to a new mixing head, that had been fitted tothe wrist of the robot by means of proper flanges.

The glass was carried from its storage to the chopper by means of flexible plastic pipes, andthe path was designed so that there would be no obstructions and friction for the roving.Glass roving X900A, a 2400 Tex, very soft roving with average attitude to the cut, wassupplied by Owens Corning Fibreglass.

A chemical formulation supplied by Dow, characterised by cream time of 12 seconds, geltime of 40 seconds, demoulding time of 3 minutes and free rise density of 50-55 kg/m3

was used.

The first tests were run using approximately 30 parts of glass over 100 parts of polyurethane.Surprisingly, the results were extremely satisfactory from the very beginning (see Figure 3.23).

Figure 2.23 Homogenous wetting of all fibres was obtained since the very first trialswith InterWet, thanks to the interval geometry of the mixing head.

147

After fixing minor problems with the cutting control, an optimum distribution of blendwas obtained, and the glass fibres were all wet (this was clearly visible becausepolyurethane was pigmented in black, and absolutely no white spot of dry glass wasshown at the end of an open-mould pouring operation). Being all wet, the fibres did nottend to stick vertically in the rising foam, therefore less possibility remained for the air toremain trapped in the mixture: the moulded parts were, in fact, totally free of air bubbles.

A number of trials were satisfactorily run using thin, transparent polyethylene film asliner for both mould halves, so that when the liner was peeled off it was possible toevaluate the quality of the compound and its surface (see Figure 3.24).

The second series of trials was run on a mould for production of automotive door panelswhere a layer of expanded or compact PVC sheet was unrolled on the lower mould-halfand cut to measure, its edges were blocked with a frame and it was vacuum-formed toperfectly adhere to the mould surface. After this operation the blend of polyurethaneand glass was poured on the rear side of the PVC liner, and the mould carrier closed forpolymerisation. On this occasion - since the first results were very encouraging - the

Figure 3.24 No holes are visible in this InterWet non-presses laboratory sample,photographed against a strong light. The diffuser-mounted on the InterWet provides a

very distribution of material all over the mould.


148


percentage of glass was increased to 35% of the total blend, i.e., approximately 60 partsof glass to 100 parts of polyurethane. Again, very nice parts were obtained at the earlystages (see Figure 3.25).

Figure 3.25 Nice parts, free from air entrapments between reinforced PU and outerskin were obtained at a very early stage of the InterWet development work.

First Improvements

• Improved distribution [25]

One aspect of the laydown - the width of the pouring path - was worth immediateattention; since the reactive blend leaving the head covered only a few centimetres ofmould at each pass, a rather high number of passes was required to properly cover thewhole surface which forced use of the robot at very high speed to complete the wholepattern before creaming started. To avoid the overlapping of two contiguous layers ofmixture, sometimes a narrow gap remained between the layers. This did not createproblems but the distribution of glass in those thin strips was probably different than inthe other areas of the mould.

149

To overcome these problems, a specially-designed pneumatic device was added at theexit of the mixing head, that deviates, under controlled conditions, the flow of mixturefrom its natural, vertical direction and distributes it over a wider pattern and, consequently,in a thinner section. By doing so it is irrelevant if two contiguous layers do overlap (sincethe layers are thinner the effect is negligible) and fewer passes are required to cover thesame surface. The robot can be used at lower speed (or a simpler, less expensive robotcan be used) or more parts can be made in the same unit of time. This could be relevantthinking to carousel-based operations, where the robot would have to perform differentpouring patterns in a sequence, on different moulds.

Another advantage derived from this pneumatic distributor is that the jet of air pusheseven further down the fibres, which results in a very thin layer of mixture applied on themould surface. This means that the male half of the mould - when closing over thefemale part - does not ‘wash down’ the vertical walls from the layer of reacting foam andglass which covers them, since this layer is quite thin and does not constitute an obstaclefor the lowering plug. Parts as thin as 1.5 mm can be moulded with optimum surfaceaspect and homogeneous distribution of glass across the whole surface.

• Variable fibre length, with ‘on-the-fly’ change

The first fibre-cutting device, as stated above, had the possibility of accommodatingmore blades. Therefore, the length of cut fibre obtained was equal to the circumferenceof the cylindrical blade-holder when only one blade was installed, or half of itscircumference with two blades, or one-quarter of it with four blades. Since this wasforeseen in the original wish list, the option of allowing the user to vary the length ofglass fibre at will, even ‘on-the-fly’ on the same moulding, was soon taken intoconsideration. A simple, sturdy mechanical device was designed to allow for this variablecutting. This mechanism which operates via an impulsion given by the robot’s programmeat a given point of the pouring pattern, cuts the roving in shorter sections; this variable-length cutter provides different cuts of fibre: from L to L/12, where L is the longestpossible choice and the combinations depend on the use of the 12 blades installed on theholder.

• Variable glass fibre output

Operated by a hydraulic motor, the pulling device (two opposed rolls that rotate andtrap between them one or more roving, pulling them from their storage) is able to providevariable output of glass. By electronically controlling an inverter or a proportional valveit is quickly possible to vary the motor speed, providing more or less fibre in different


150


parts of the moulding. The motor can be even stopped (thus feeding only polyurethanefoam) or left running after the end of the polyurethane shot (providing dry, unblendedglass fibre over some areas of the mould where an extra reinforcement would be desired).

• Extended safeties

The main concerns in operating this kind of plant derive mostly from the irregular feedof glass to the head. For this reason the latest InterWet machines have been equippedwith detectors able to command an immediate stop to the machine in case of:

• Empty storage of glass roving

• Roving blocked in its roll

• Roving blocked somewhere along the feeding line

• Roving not properly fed into the head after the cutter and forming a ‘birds nest’

• Malfunction of a single blade.

3.3.1.4 Advantages of the InterWet Process

These new developments have resulted in a higher degree of flexibility for the InterWetmachines whose advantages versus the existing and available similar methods can besummarised as follows:

• Optimum overall mechanical properties, due to an optimum glass fibre wetting withPU, because of the mix-head design

• Better distribution of the mechanical properties of the moulded part, i.e., no areaswithout or with little glass, due to a more homogeneous glass fibre distribution inthe mould, because of the special distribution system.

• Programmable variation of mechanical properties on the same part, since thedistribution of the composite’s mixture on the surface of the mould is fullyprogrammable. One can have in different areas of the same part:

• PU and glass in fixed proportion

• PU and a varying proportion of glass

• PU only

• Extra glass over an already filled area.

151

• Easy operation because it uses a standard (modified) FPL mix-head with obviousconsequences in terms of simplicity of maintenance, reliability in use and availabilityof spare parts when required.

• Flexibility in operation because it is designed for more than one type of additivecomponent: it can easily be converted to process fillers or fibres or chopped scrapfoam. Its range of application includes continuous reinforcing fibres, natural fillers(mineral or organic), pulverised plastics, etc.

• All the operational advantages deriving from the use of a light mixing head (robotsize, speed, productivity, etc).

• Economics, because - being based on commercially available components - itsmechanical parts can be less expensive than specially developed ones.

When compared with other glass-reinforcing technologies the following advantages mustbe added:

Versus some RRIM applications:

• Economics, since it is no longer necessary to carry out the expensive dispersion ofsolids in the polyol component - now, by adding them directly in the mixing head, itis possible to eliminate all the traditionally associated problems, i.e., possibleabsorption of polyol into the filler with the resulting slow release of reactivecomponents in the polymer matrix, abrasion, clogging of lines and tanks, fluctuatingpercentages of filler in the liquid, etc.

Versus some mat applications:

• Economics:

in glass (roving costs less than mat)

in equipment (one more machine, storage, handling, floor space, etc.)

in scrap (preformed parts - although optimised - require punch-cutting that producesscrap trim)

in stock (stored preforms can be damaged by sunlight, dust, moisture, and oftenmust be thrown away)

in manpower (to handle, position, punch, etc.)


152


• Quality (foam mixes better when co-injected with glass roving rather than wheninjected into a pressed mat preform)

• Flexibility: it’s easy to mount on a carousel various moulds and produce small seriesexploiting the flexible properties of the robot.

It is evident that there are applications where both RRIM and mat preforms are stillpreferred because of their specific advantages. InterWet has been conceived to fill thegap between the two technologies, and allow for more extended use of reinforcementsand fillers in applications where neither RRIM nor mat preforms could be used [31].

3.3.1.5 Applications

When this technology was first introduced it immediately appealed - for the abovementioned reasons – to automotive and transportation end users. Volumes, cycle timesand final properties meet this market segment’s requirements, but until the currentlyavailable solutions address some of the above-illustrated problems this industry wouldnot select it for mass production.

A market segment where this application would be really interesting is the manufactureof special sanitary-ware parts, for at least three different purposes. The quickest tounderstand for its immediate advantages is the combination of an outer plastic shell -with either aesthetic or functional purposes - back-foamed with reinforced polyurethane;examples include whirlpool tubs, bath tubs, shower trays, and vertical elements for showercabins integrated with accessories. These products are characterised by medium-smallproduction series, wide range of models, and large surfaces. The high operating flexibilityallowed by the programmable robot is very adaptable for the production of this type ofmoulding: the larger the part, the more advantages can be found in this technology.Medium-deep-draw shaped parts can be easily thermoformed from large thermoplasticsheets (even in medium-small series) and immediately reinforced in an integratedproduction line including thermoformer, foaming station and curing area. Dimensions,productivity and investment must be tailored to the project’s requirements.

Another area of interest for this new technology is the re-use of chopped foam scrap,either flexible or rigid; the design of the solid component’s feeding device - as foreseen inthe project’s objectives - has been made in a way that, using an appropriate volumetric-feeding mechanism, powders, granulates and recycled foams can be dosed in the mixingchamber through the same channel bored in the mixing head’s cleaning plunger. It ispossible to add roughly pulverised scrap foam to the virgin formulation directly in themixing head. This would solve a recycling problem, as well as lowering the cost of

153

formulation. Positive trials have been made in a development phase, with filler loadingas high as 10% of the final part’s weight (see Figure 3.26).

Another interesting potential application is the use of continuous fibres, cut in medium-long size, to reinforce thin pads and cushions moulded with flexible foams. The continuousdevelopment of lighter, thinner parts - where weight-reduction and volume-exploitationis the object of intense development - requires the use of formulations characterised byvery high mechanical properties that unreinforced foam cannot meet. The use of longnatural or synthetic fibres to reinforce these flexible mouldings had applied for a longtime, but only under the form of textile inserts manually inserted in the mould prior tothe foaming operation. Being able to co-inject foam and long fibres - applying longer orshorter sizes according to the mechanical properties required - represents an advantagethat comes at a very attractive investment cost.

Figure 3.26 In addition to glass roving a number of other materials can be co-injectedwith PU using the InterWet technology, scrap PU foams, sand, ground cork and other

fillers that would be difficult to add in the component tank.


Note: The Cannon FPL Mixing Head is covered by several international patents.

154


References

1 M. Taverna and B. Pile, Presented at the Polyurethanes Expo ’99 Conference,Orlando, FL, USA, 1999, p.37.

2 Modern Plastics International, 1999, 29, 3, 127.

3 M. Taverna, Presented at the Recycle ’94 Conference, Davos, Switzerland, 1994,Paper No.35.

4 L. White, Urethanes Technology, 1995, 12, 2, 3.

5 Plastics and Rubber Weekly, 1995, No.1612, 10.

6 D. Smock, Plastics World, 1995, 53, 11, 20.

7 British Plastics and Rubber, 1995, November, p.10.


9 M. Taverna, Presented at the Utech Asia ’97 Conference, Suntec City, 1997,Paper No.12.


11 C. Fiorentini, M. Taverna and J. Luca, Presented at the Polyurethanes ’95Conference, Chicago, IL, USA, 1995, p.476.

12 Plastiques Flash, 1998, 304/5, 92.

13 M. Taverna, P. Corradi and B. Biondich, Presented at the Polyurethanes ’95,1995, Chicago, IL, p.484.

14 M. Taverna and P. Corradi, Presented at the Utech Asia ’96 Conference, TheHague, 1996, Paper No.10.

15 P. Mapleston, Modern Plastics International, 1997, 27, 8, 30.

16 Plastics and Rubber Asia, 1997, 12, 72, 34.

17 British Plastics and Rubber, 1997, July/August, 19.

18 Plastics and Rubber Weekly, 1997, No.1687, 17.

155

19 M. Taverna, Presented at the Polyurethanes Expo ’98 Conference, Dallas, TX,USA, 1998, p.685.

20 P. Mapleston, Modern Plastics International, 1997, 27, 11, 41.

21 Plastics Rubber Weekly, 1998, No.1729, 13.

22 Macplas International, 1998, February, 46.

23 G. Graff, Modern Plastics International, 1998, 28, 11, 38.

24 British Plastics and Rubber, 1998, November, 20.

25 M. Taverna and A. Bonansea, Presented at the Polyurethanes Expo ’98, 1998,Dallas, TX, 687.

26 European Plastics News, 1999, 26, 4, 48.

27 Composites – French/English, 1999, 36, 50.

28 J. Stark and F. Peters, Presented at the Utech ’99 Conference, 1999, Singapore,Paper No.6.

29 Plastics and Rubber Asia, 1997, 12, 70, 34.



156


157

4.1 Introduction

Rigid polyurethane (PU) foam is the preferred insulator material in a wide range ofapplications encompassing household, industrial and commercial appliances (refrigerators,freezers, display cases, vending machines), transportation (refrigerated trucks and reefers,shipping containers), insulation in buildings and in industrial plants.

The replacement of chlorofluorocarbons (CFC) with environmentally more benignchemicals, as recently mandated by the Montreal Protocol and subsequent revisions, hascaused a reduction of the insulation efficiency of the PU foams, since the new blowingagents available, like hydrocarbons, hydro-fluorocarbons (HFC) or carbon dioxide haveworse gas thermal conductivity properties.

In applications, the lower insulation efficiency of the CFC-free foams generally leads tohigher energy consumption levels unless specific compensation actions are taken.

This issue is particularly severe for the household appliance industry, which has alsobeen called upon over the past years to progressively reduce the energy consumptionof their products.

With the energy efficiency targets getting more and more demanding and only marginalimprovement obtainable from the optimisation of conventional technologies, the appliancemanufacturers are seriously looking at alternatives which may provide additional benefitin terms of energy savings.

The appliance industry is not the only one involved in the debate on energy consumptionreduction. The growing attention to the global warming issue is forcing policy makers totake actions to cut carbon dioxide emissions and minimise the greenhouse effect. Withthis in mind, the Kyoto Conference has posed ambitious targets for energy saving, theachievement of which calls for improvements in many fields, from transportation, torefrigeration and building insulation.

The ban on CFC and the demand for a more efficient use of energy, both issues coming fromenvironmental concerns, are therefore posing serious challenges to the Industry as a whole.

4 Recent Developments in Open CellPolyurethane-Filled Vacuum Insulated Panelsfor Super Insulation Applications

Paolo Manini

158


Vacuum Insulated Panels (VIP), having a thermal conductivity three to five times lowerthan conventional PU foams, allow the achievement of superior insulation performance.They can be used to partially replace the conventional insulation materials to providea more efficient insulation structure, which allows energy saving without the need toincrease the insulation thickness. Alternatively, VIP can be used in those applicationswhere it is important to reduce the insulation thickness to a minimum value withoutloosing thermal performances.

A VIP is obtained by packaging a microporous low conductivity filler material inside ahighly impermeable gas barrier bag. The filler is then evacuated to a proper vacuum leveland the bag sealed. A gas adsorbent, normally referred to as a getter, inserted in the bagbefore sealing, is also necessary in most panel designs to ensure the proper vacuum levelduring the lifetime of the panel.

Several fillers have been proposed in the past, either in the form of compressed powdersor fibres, however all have some disadvantages in terms of cost, process complexity and/or weight. This has prevented VIP technology from finding widespread acceptance inmost applications. The recent development of open cell rigid foams, first introduced intothe market in the early 1990s has sparked off new interest in this technology. However,in order to fully exploit the unique properties of these insulating materials, the re-evaluation and re-design of the other key elements of a vacuum panel, i.e., the barrierfilm, the gas adsorbent and the manufacturing cycles, has also been necessary. This process,which is still ongoing and has potential for further improvements, has required thecommitment of several sectors of the industry in the last five to six years.

In this chapter, the present status of the open cell PU foam-filled vacuum panel technologyis discussed and some recent developments in film, getter and processing technologiesare reviewed. Special focus is given to the vacuum issues, which are key for the properselection and treatment of the VIP components. Some specific aspects, related to themanufacturing, characterisation and practical use of vacuum panels, as well as theachievable benefits, are also considered. Examples of applications where this technologyis finding its place in the market and areas where further work is necessary to makevacuum panels more cost effective are presented.

4.2 Some General Properties of Open Cell PU Foams for VacuumInsulated Panels

Vacuum panels have been studied as a means to improve thermal insulation for a longtime. Several insulating fillers, such as silica and perlite powders [1, 2, 3], fibre glass [4,5], and aerogels [6, 7, 8] have been proposed as core materials for VIP, each of them

159

having specific advantages and disadvantages in term of weight, thermal conductivity,handling, processing, vacuum properties and cost.

To overcome some of the drawbacks associated with the use of these materials, newfamilies of fillers, based on PU [9, 10, 11, 12] and, more recently, polystyrene (PS) [13]open cell foams, have been developed and proposed to the market. Both open cell PUand polystyrene foams are now being considered as interesting options due to theirmoderate outgassing, good thermal insulation values, low weight, ease of handlingand cost effectiveness.

This chapter will focus on the open cell PU foams, even though many general aspects andrecent developments of the technology, dealing with films, getters and manufacturingprocesses, can be also applied, with minor changes, to other micro porous fillers.

The thermal conductivity, or λ factor (mW/m-k), for some of the most popular fillermaterials is shown in Figure 4.1 as a function of pressure (Pa). The thermal conductivityof cyclopentane (CP) blown closed cell PU foam is also given for comparison.

Figure 4.1 Thermal conductivity as a function of pressure for some filler materials

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…

160


Regardless of the type of filler, all curves have a similar behaviour, with a low pressureregion, where the λ factor is constant, followed by a region where the λ factor increaseswith increasing pressure.

This common trend can be explained considering that the total apparent thermalconductivity in micro porous materials is the sum of four physical contributions, i.e., thethermal conductivity through the solid, λs , the thermal conductivity through the gaseousphase, λg, the thermal conductivity by thermal radiation, λr and by gas convection, λc:

λ = λs + λr + λg + λc (1)

The last term of Equation (1) can be neglected for core material with cell size smaller than 1 mm.

An extensive treatment of the above physical mechanisms and their role in heat transferin low density closed cell foams has been provided by Glicksman [14].

In the case of the open cell PU foams detailed expressions for the first two terms ofEquation (1) have been given by Kodama and co-workers [9].

Both terms are independent of pressure and strictly related to the foam density andmorphology (cell size, structure and degree of anisotropy).

In particular, λs is directly proportional to the thermal conductivity Ks of the foam andinversely proportional to the cellular anisotropy η according to the formula [9]:

λ ηs g sV K= −( )[ ]1 2 3/ / (2)

where Vg is the fraction of foam volume occupied by the gaseous phase.

To reduce the heat transfer through the solid it is therefore necessary to increase thecellular anisotropy and to lower the thermal conductivity of the polymer structure.

Following Kodama and co-workers, the heat transfer by radiation is given by:

λ r g rV H d= 2 3 1 3/ / (3)

where d is the cell size and Hr is the coefficient of thermal conduction by radiation whichdepends on the emissivity of the polymer and the morphology of the cell. To minimisethe λr contribution, the cell size and Hr have to be reduced. The former can be achievedby adjusting the formulation of the foam, the latter by reducing the opening size of cellmembranes to make radiation transmission through the cellular windows less efficient.

161

The heat transfer by gas conduction, λg, becomes progressively more important as thepressure increases, as shown by the following set of equations [15]:

λ λg go L d= ( ) +( )Π / /1 2 (4)

where λgo is the thermal conductivity of air at atmospheric pressure, Π is the porosity ofthe foam and L is the mean free path of air, linked with pressure P by:

L kT P= ( ) ⎛⎝

⎞⎠/ 2 2π σ (5)

where T is the absolute temperature, k the Boltzmann constant and σ the collision diameterof air (about 4 x 10-10 m2).

The λg contribution vanishes for L>>d. For currently available foams, having an averagecell size in the 100-200 μm range, this means that pressure has to be kept at 1.0 Pa orbelow to completely eliminate the heat transfer by gas conduction.

To ensure this condition the foam has to be properly evacuated and all gas sourcesdeteriorating the vacuum during the panel life effectively compensated by a suitableadsorbent, or getter, having sufficient gas capacity and efficiency.

As shown by Equations (1-5), the insulation properties of an open cell foam depend ona complicated interplay of different parameters such as the PU thermal conductivity,foam density, cell size and cell size distribution, cell morphology and anisotropy, whichhave to be optimised during the foam preparation to obtain the best trade-off among thethree thermal conductivity contributions, λs, λr and λg.

In this process, the mechanical properties of the foam cannot be neglected, being essentialto ensure structural stability of the vacuum panel.

In fact, after sealing, the vacuum panel is exposed to the hydrostatic load of theatmospheric pressure and has to withstand it for a long time, which can be 15-20 yearsor even more depending on the application.

Dimensional stability tests carried out on open cell PU foam-filled vacuum panels showedthat creep problems should not occur for properly prepared panels, provided the opencell foam preparation has been optimised [15].

The open cell PU cores are produced according to various preparation technologies,such as lamination and block foaming. In both cases, after the foam has been grown, acutting operation is usually necessary to remove the outer skin of the material, which


162


contains a high percentage of closed cells, and to slice the PU into slabs of the proper sizeand thickness. The removal of the outer skin may be a labour intensive process whichproduces waste, lowers the process yield and increases the overall production costs, thisissue being particularly important in the case of the lamination process. Great attentionis therefore now being paid to design and the open cell foam production process, toreduce waste and increase productivity.

A different approach to making open cell foam slabs has been proposed recently [11],which is based on the use of the PU fluff obtained from the recycling of used refrigerators(Recycled Urethane Fluff, or RUF panel).

During the recycling process, the rigid insulation foam contained in the old refrigeratorsor freezers is mechanically ground into a fine powder, to completely separate and recoverthe CFC present in the foam. The resulting fluff, composed of completely open cells, isdried and introduced into a binding machine where it is sprayed with a certain amountof isocyanate (15-20% by weight) and thoroughly mixed. The blend is then transferredto a mould, heated at 120 °C and compressed at 0.5 MPa for 10 minutes to consolidateit and remove residues of the process. After de-moulding and curing, the RUF panel canbe used as core material in a VIP.

The fluff obtained from recycling one single refrigerator can be used to produce enoughvacuum panels to insulate a new appliance, thus generating a virtual recycling loop.

Due to the use of very fine compressed powder, the density of this open cell foam is threeto four times higher than those produced by lamination or block foaming.

Selected physical properties of some open cell PU foams are listed in Table 4.1.

163

4.3 Vacuum Issues in the Selection of VIP Components

4.3.1 Vacuum Properties of the Open Cell Foams

To fully exploit the insulating performances of the open cell PU foams, pressure in thepanel has to be kept preferably below 1.0 Pa during its life. To achieve this demandingtarget, the foam must be 100% open celled and with a very low outgassing rate.

In spite of the increasing use of open cell PU foam, data from the literature on its vacuumproperties are quite scarce.

To estimate the outgassing contribution, specific tests can be carried out using high vacuumbenches equipped with a quadrupole mass spectrometer.

smaofUPllecnepofoseitreporplacisyhpemoS1.4elbaT

]65[.feR ]65[.feR ]9[.feR ]21[.feR ]11[.feR

noitcudorPssecorp

noitanimaL kcolBmaof

- kcolBmaof

FUR

m/gk(ytisneD 3) 46-06 85-25 25 56-06 002-081

(ezislleC μ )m

egarevA 001 071-031 041 - -

ralucidnepreP - - 57 - -

lellaraP - - 502 - -

lamrehTytivitcudnoc

aP5@)K-m/Wm(

7 5.7-5.5 5-5.4 7 01-8

)aPM(htgnertsevisserpmoC

egarevA 52.0 53.0-32.0 62.0 - -

lellaraP - - 383.0 - -

ralucidnepreP - - 331.0 - -

tnemtaert-erP 021 °Cnim03x

021 °Cnim03x

041 °Cruoh1x

021 °Cnim03x

021 °Cnim03x

tnetnocllecnepO)%(

001 001 001 001 001


164


To quantify the relatively small amounts of gas species released from the sample, it isnecessary to use very clean, bakeable, high or ultra-high vacuum systems made with stainlesssteel and/or glass components. This ensures better leak-tightness and a negligible gas emissionfrom the bench. A sketch of a typical experimental apparatus, having two separate pumpinggroups and base pressure at 1 x 10-6 Pa or lower, is shown in Figure 4.2.

A practical technological implementation of the scheme shown in Figure 4.2 is illustratedin Figure 4.3.

The sample to be analysed is mounted in a glass bulb which is connected to the testbench and evacuated for 10 minutes with a turbo and a rotary pump to a final pressureof 10-4 Pa. The bulb is then isolated from the pumps and the total pressure increase, dueto the sample outgassing, is monitored with a capacitance manometer for some days.The use of the capacitance manometer avoids changes in the gas composition whichmight occur if a hot filament pressure gauge is used. From time to time a small amountof gas is sampled from the test volume, through valves V4 and V6, and passed to themass spectrometer for partial pressure measurement. Before running the outgassing tests

Figure 4.2 Scheme of the stainless steel high vacuum bakeable bench equipped withmass spectrometer for outgassing tests on vacuum components. IG: ionisation gauge.

165

on the PU sample, a blank run has to be carried out. The result of the blank run is thensubtracted from the outgassing test to get the actual gas emission from the PU sample.The use of two distinct pumping groups is essential to minimise the system contamination(mainly water vapour) which occurs when the bulb is opened to the air, to mount thespecimen. In fact, during this operation, valve, V4, is kept closed, reducing the surfacearea of the bench which can absorb water and atmospheric gases.

Before being used in a vacuum panel, the open cell PU foam needs a preliminary heattreatment in air, generally carried out at 120-150 °C for 10-60 minutes to remove waterand other volatile species which otherwise would desorb and rapidly cause the vacuumto deteriorate. The result of a typical outgassing test carried out at 23 °C on a foamsample baked at 120 °C for 30 minutes is shown in Figure 4.4 for all desorbed gases butwater. Water is difficult to quantify since it sticks to the walls of the system and onlypartially reaches the mass spectrometer. Water can be estimated as the difference betweenthe total absolute pressure and the sum of the partial pressures of the other gas species,which can be accurately quantified with the mass spectrometer.

Figure 4.3 A stainless steel outgassing bench equipped with a quadrupole massspectrometer and an external oven for vacuum bake-out(Reproduced by courtesy of SAES Getters SpA, Italy).


166


For each gas species, the amount released by the sample at a given time is obtained bysimply multiplying the test volume (V) by the gas partial pressure (qi(t) = Pi V, qi(t) is theoutgassing rate (Pa-l/h) at time t, Pi being the measured i-th partial pressure).

The experimentally determined qi value for each gas species can be interpolated overtime according to the semi-empirical law quoted in the literature [16]:

q t q ti ii( ) ( )= −

0α (6)

where qi(t) and q0i are the outgassing rate (Pa-l/h) at time t (hour) and the desorbedamount (Pa-l) after 1 hour of the i-th gas species, respectively. The dimensionless parameterαI is related to the desorption mechanism of the i-th gas species, its value usually rangingfrom 0.5 to 1, depending on the gas species and the material considered [16, 17].

Integration over time of Equation (6), provides an estimate of the PU sample outgassingload Qi for each gas species after a given time. As an example, the estimated gas releasedafter 1 to 20 years for the sample of Figure 4.4 and for a 50 x 50 x 2 cm3 panel size isgiven in Table 4.2. The main gases, besides water, are carbon dioxide, nitrogen, carbonmonoxide and hydrogen. It is interesting to note that, given the nature of the desorption

Figure 4.4 Outgassing rate for an open cell PU foam. Water is not shown.

167

process (Equation (1)), most of the gas is given off during the initial period of operationof a VIP.

Results shown in Figure 4.4 and Table 4.2 can vary from sample to sample depending onthe open cell PU foam preparation technique, its microstructure and density and thepre-treatment.

Since desorption is a thermally activated process [18], the outgassing rates increase asthe temperature increases. The outgassing contribution has therefore to be carefullyevaluated in all those applications where the vacuum panels operate continuously attemperatures higher than room temperature, e.g., 60-80 °C, or have to withstand hightemperature peaks, for example 100 °C, even for a relatively short period of time. Examplesof such applications are presented and discussed in Section 4.6.

An additional factor which may have an impact on the vacuum properties is the closedcell content of the PU foam. All the open cell PU foams so far presented in the literaturefor VIP applications are quoted to be 100% open cell based on pycnometric measurement.On the other hand, the accuracy of this test method is generally close to ± 0.2% [15], sothat the possibility of a small fraction of closed cells cannot be completely ruled out andhas to be considered as an additional potential source of pressure build-up in the panel.

Closed cells might be present due to a non optimal foam preparation or to the incompleteskin removal.

4.3.2 Vacuum Properties of the Barrier Film

The barrier film plays an important role in vacuum panel technology since it has the taskof minimising air and moisture penetration into the vacuum core. The barrier must be

5,1retfa)l-aP(iQdaolsagdetamitsednasretemarapgnissagtuO2.4elbaTPIVmc2x05x05arofsraey02dna

saG q i0 )h/l-aP( αi Qi )l-aP()1( Qi )l-aP()5( aP()02(iQ l- )

OC 2 01x8 3- 1 22 52 72

OC 01x2 3- 1 45 46 57

N2 01x9 3- 1 8.01 8.21 51

H2 01x5 5- 5.0 3.2 2.6 31


168


durable and able to resist to puncturing and abrasion. It must be functional over a widetemperature range and retain its physical properties, such as dimensional stability,flexibility and sealability for years.

Thermal conductivity of the skin should also be very low to minimise the heat transferthrough the panel edges which can partially reduce the insulation efficiency of theevacuated panel (‘edge effect’, see Section 3.2.1). These properties have to be coupledwith very low gas permeation and outgassing rates.

Gas permeation through the barrier envelope is one of the most important factorsresponsible for the pressure increase in a panel during its life.

Depending on the structure of the barrier film and the materials used, gas ingress canpreferentially take place through the whole surface of the barrier film (permeationthrough the surface) or through the heat-sealed plastic VIP flanges (permeation at theedges), or both.

4.3.2.1 Gas Permeation Through the Surface

Several types of barrier materials, having different structures and gas transmission rates,are commercially available from the food, packaging and electronic industry. Gas barrierrequirements for vacuum panels are however much more demanding.

Composite plastic films, obtained by laminating several polymeric sheets have beenproposed in the past for precipitated silica-filled panels, where pressures as high as 500-1000 Pa can be tolerated without excessive degradation of the thermal conductivity. Inthe case of open cell PU-filled panels, requiring much lower pressure values, this barrierstructure is generally inadequate, if a lifetime exceeding one to two years is needed.

To improve the permeation properties, barrier materials, such as aluminium and siliconoxide, can be vacuum deposited on the polymer surface in the form of a thin film (0.01-0.05 μm thick) [19].

However, most commercially available metallised samples do not provide enough barrierproperties for long term applications, mainly due to water transmission through the largedensity of pin holes and micro-defects present in the sputtered aluminium layers [20].

To overcome this problem, barrier films obtained by laminating a thicker continuousmetal foil with various polymeric sheets, such as polyethylene terephthalate (PET),Nylon and polyester (PE) have been proposed to the appliance industry some years agoand are now commercially available [15, 21]. Aluminium is the preferred choice for

169

the metal, mainly due to its good workability, which allows production and laminationof virtually pin hole-free foils having thickness of 6 μm or less [21]. The presence ofthe aluminium foil dramatically improves the gas barrier property of the film, wellbeyond the sensitivity limits of the analytical techniques commercially available forthe measurement of gas permeation.

In the case of water, which is one of the most important permeating species, two standardmethodologies are at present widely used to measure its transmission rate through a material,as described by ASTM E96-00 [22]. According to the first procedure (‘the desiccant method’)the barrier sample is sealed to the open mouth of a test plate, which is cylindrical in shape,in which the desiccant is placed, and the assembly kept in a humidity and temperaturecontrolled environment. The permeation of water through the sample is measured by theweight increase of the assembly. In the second procedure (‘water method’) the test platecontains water and the transmission rate through the barrier sample into the controlledenvironment is measured by the weight decrease of the assembly.

A variation of the first method, frequently used in practice, is based on periodicallyweighing vacuum panels, containing of desiccant, and kept in a given temperature andhumidity-controlled environment.

Another very common technique is based on the use of an infrared sensor as describedby ASTM F1249-90 [23]. The barrier sample is sealed between a dry and a wet chamberkept at known temperature and relative humidity. The two chambers make up a diffusioncell which is placed in a test station where the dry chamber and the top of the barrier areflushed with dry air.

Water vapour which penetrates through the barrier film blends with the dry air and istransported to a pressure-modulated infrared sensor. The infrared radiation is absorbed bythe water molecules and the sensor produces an electrical signal which is proportional to theconcentration of water in the gas phase. The intensity of the signal is then compared with thesignal generated by measurement of a sample having a known water transmission rate. Thisallows the determination of the water transmission rate through the test specimen. Thedetection limit for water transmission generally quoted for commercially available instrumentsbased on this principle is in the 0.01 g/m2 day range (at 90% relative humidity and 23 °C).

These techniques are generally more than adequate for many plastic or standard qualitymetallised films but can hardly discriminate between samples containing aluminium foil,which generally have water transmission rates much lower than this value.

To better estimate gas permeation for high quality barriers a novel technique has beendeveloped [24, 25], which is based on the measurement with a quadrupole massspectrometer of the helium transmission rate through the sample.


170


A sketch of the experimental bench is shown in Figure 4.5.

A picture of the bench is shown in Figure 4.6.

The bench is a stainless steel ultra-high vacuum apparatus equipped with rotary, turbo-molecular and non-evaporable getter (NEG) pumps. The sample, a circular coupon ofabout 30 cm2 area, is mounted between two flanges and the helium pressure is applied(generally in the 0.0001-0.1 MPa range) on one side of the sample, the other side beingin view of the quadrupole mass spectrometer. A variable conductance C (l/s) is mountedbetween the mass spectrometer and the evacuation group, so that the helium flow rate F(Pa-l/s) through the conductance can be measured with the mass spectrometer accordingto the equation:

Figure 4.5 Scheme of the experimental bench for the helium permeation test.

171

F = C (P1 - P2) (7)

where P1 and P2 are the pressures before and after the conductance C, respectively.

Under equilibrium conditions, which can be reached after some minutes or some hours,depending on the nature of the sample, the gas flow, F, provides the helium transmissionrate. A typical graph showing the achievement of steady state conditions is given inFigure 4.7.

Figure 4.6 High vacuum bench for helium permeation tests(Reproduced by courtesy of SAES Getters SpA, Italy).


172


The use of the mass spectrometer ensures very high sensitivity for helium, better than10-11 Pa-l/(s m2 Pa), mainly due to the negligible presence of helium as a constituent ofthe gas background. Due to its high sensitivity and the relatively short measuring time,this technique can be used effectively to support the development of improved laminatesand also for quality control in a production environment.

The estimation of the transmission rates for gases other than helium can also be obtained,provided a preliminary calibration is carried out. This is particularly important for water,which is the main gas permeating through the barrier skin. To run the calibrationprocedure, the helium transmission rate is measured for various samples having differentand known permeation rates for water. The linear correlation between helium and watertransmission rates is then established, as shown in Figure 4.8.

The correlation factor for water was found to be close to 500 for most of the PE/PET/Nylon-based films analysed by the authors, either metallised or incorporating analuminium foil [26]. It was then possible to estimate the order of magnitude of the water

Figure 4.7 Achievement of the equilibrium condition in a helium permeation test for alaminated film composed of 15 μm Nylon, 12 μm PET, 6 μm aluminium, 50 μm highdensity polyethylene (HDPE) (or Nylon 15 μm/PET 12 μm/Al 6 μm/HDPE 50 μm).

173

transmission rate for an unknown low permeation rate sample by simply measuring thehelium transmission rate and multiplying it by 500.

Table 4.3 provides a comparison between the helium permeation rates in two films, alaminate incorporating a 6 μm aluminium foil, PET 12 μm/Aluminium 6 μm/HDPE 50 μm(Film B) and a multi-layered barrier composed of four aluminium-sputtered PET sheetslaminated onto a 50 μm PE (Film A). Tests were run at 24 °C. From the measurement of

Figure 4.8 Experimental determination of the correlation factor between helium andwater transmission rates.

smlifhguorhtsetarnoitaemrepretawdnamuileH3.4elbaT

elpmaS (etarnoitaemreP saPM 1- mc 2- aP 1- ) retawdetamitsEPIVanitnuoma

sraey51retfamuileh retaw

AmliF 01x019 21- 01x064 9- g001

BmliF 01x031 51- 01x076 31- g20.0


174


the permeation rates, the quantity of water permeating in 15 years in a 50 x 50 x 2 cm3

vacuum panel has been estimated.

The metallised barrier here considered cannot be used for long term applications, where10-15 years or more are targeted, since the amount of desiccant necessary to compensatefor the water transmission would be too large.

A much lower permeation rate is provided by Film B, which is therefore well suited forlong term applications.

In spite of the extremely good gas barrier properties, VIP prepared with such a filmsuffers from an intrinsic limitation, i.e., the high thermal conductivity of the aluminiumfoil, as shown in Table 4.4.

A fraction of the heat flow is, in fact, transferred from the hot to the cold panel surface by thealuminium foil through the panel flanges, rather than through the core material (so called‘edge effect’). As a result, the average insulation value of the panel is less than the expectedvalue based on the actual insulating properties of the core material (centre of the panelthermal conductivity), this difference being more remarkable the smaller the panel size.

The edge effect can be evaluated, as a function of the panel size, the laminate structure andthe aluminium thickness, by numerical methods, such as the Finite Element Analysis (FEA).

Results of such an analysis are given in one specific example in Figure 4.9, where theaverage thermal conductivity of a vacuum panel using a PET 12 μm/Aluminium 6 μm/HDPE 50 μm laminate is plotted against the panel area.

ro/dnasreirrabsagsadesuslairetamemosfoytivitcudnoclamrehT4.4elbaT.smlifreirrabnisreyallarutcurts

lairetaM ytivitcudnoclamrehT(μ )K-m/

ssenkcihtlacipyT(μ )m

TEP 51.0 51-01

EPDH 50.1 06-04

CDVP 31.0 02-01

nolyN 34.0 02-01

muinimulA 372 7-6

edirolhcenedilynivyloP:CDVP

175

The thermal conductivity and the thickness of the PU core material considered here are6 mW/m-K and 2 cm, respectively.

Figure 4.9 shows that the panel thermal conductivity approaches the value of the corematerial only for a sufficiently large panel area (≈ 1-10 m2). For very small panels, theinsulating properties of the core material are spoiled by the edge effect and the energysavings in the real application may be minimal.

For this reason, efforts are now being made to improve the aluminium–based laminatesand/or to develop products not containing the aluminium foil but still having sufficientlygood gas barrier properties, so as to achieve an acceptable trade-off between energysaving performances and acceptable lifetime. This challenging objective can be addressedby selecting and properly combining together suitable polymers having improved gasbarrier properties.

A typical example of this process, recently reported by Lamb and Zeiler [27], is given inTable 4.5, which shows how significant improvements can be obtained in non aluminium-foil containing skins working on the film structure.

Figure 4.9 Thermal conductivity versus panel area.


176


A comparison of some different types of improved metallised and aluminium foil-containing barriers, also including a preliminary investigation of the effect of temperatureon the gas transmission rates, has been recently reported by Bonekamp [28].

Due to the still developmental nature of most of these barriers, more work is necessaryto assess their usability for long term applications. Extensive characterisation of themechanical, sealability and stress resistance properties are also important issues to consider.

4.3.2.2 Gas Permeation Through the Flanges

The permeation of atmospheric gases through the polymeric sealed flange of a vacuumpanel can be an important contribution to the pressure increase in a VIP for long termapplications. The gas permeation rate depends on the pressure gradient across the vacuumpanel, the type of polymer used as a sealant, the flange geometrical parameters (exposedsurface and width) and temperature. Polyethylene is at present one of the preferred sealingmaterials due to the good trade-off achieved among sealability, mechanical properties,gas permeation, outgassing rates, reliability and cost.

Gas permeability values quoted in the literature for PE are scattered over an order ofmagnitude, depending on the actual density of the sample and the gas considered [29,30, 31]. Typical permeability values corresponding to medium and high density PE areshown in Table 4.6.

Results of the calculation for argon, nitrogen, oxygen and water and the total pressure,are given in Figure 4.10 and show the role of permeation in deteriorating the vacuumlevel in the panel. The width and height of the PE flange here considered are 10 and 0.1mm, respectively and the environment is air at 21 °C and 50% relative humidity.

rofsetarnoissimsnart)RTW(retawdna)RTO(noissimsnartnegyxO5.4elbaT]72[serutcurtsmlifemos

erutcurtsmliF RTO 01x 01- m2 )PTS(yad RTW m/g 2 yad

TEP 8.43 4.34

CDVP+TEP 3.2 3.9

desillatem+CDVP+TEPreyal

40.0 26.0

003LBSR002ralyM 51000.0 51.0<

erusserp,erutarepmet,dradnatS:PTS

177

In the case of panels which are encapsulated, i.e., surrounded by a closed cell PU foam,as happens for panels in household refrigerators, freezers or vending machines (see Section4.6), the picture is more complicated. The gas environment surrounding the vacuumpanel will depend on the type of encapsulating foam and will change with time as aconsequence of the gas out-diffusion from the closed cells and the progressive airpenetration from the outside, thus making predictions more difficult. This ageing process

mc(sesagrofsetarnoitaemreP.6.4elbaT 3 m/ 2 m/g(retawdna)PTS,d 2 )d52ani μ teehsEPkcihtm

saG EPDM 01x 01- EPDH 01x 01- erutarepmeT °CN2 783 57.69 12

O2 8451 7.524 12

rA 6903 477 12

H2O 58.01 56.4 HR%09,83

ytidimuhevitaler:HR;EPytisnedmuidem:EPDM

Figure 4.10 Pressure increase due to gas permeation at the panel PE flanges.


178


will also depend on the temperature and the hermeticity of the case surrounding theencapsulating foam as shown in various papers [32, 33]. Permeation of blowing agentsthrough the seals, even though not a major contribution, has also to be considered.

The use of a different PE grade or a different sealant polymer will provide differentpressure build-up curves in the panel both in terms of total and partial gas pressures.

High barrier polymers, such as polyvinylidene chloride (PVDC), polyacrylonitrile (PAN),polyester, acid copolymers and ionomers have been proposed and are under evaluationas a replacement for PE. However, they present some drawbacks in terms of sealabilityand/or mechanical properties and/or outgassing rates. Cost of these polymers is alsogenerally higher.

4.3.2.3 Outgassing Properties of the Film

In general, very little data have been published on the outgassing properties of the skinsfor vacuum panels, even though they can contribute, in some cases, in a non-negligibleway to the deterioration of the pressure inside the VIP. This can be due to the outgassingproperties of the materials used as barrier layers and/or the lamination process, whichmay introduce volatile substances or trap gases in between the various sheets.

The outgassing properties of a barrier sample can be measured using the same experimentalequipment shown in Figure 4.2 and have to be taken into account to estimate the totalgas load in the panel. For the same film having the barrier property in Table 4.3, theextrapolation of the outgassing experimental data to 20 years in a 50 x 50 x 2 cm3 sizepanel is shown in Table 4.7. Water, which is also given off by the barrier, was not quantifiedin this test.

gnitaroprocnireirrabarofdaolsagdnasretemarapgnissagtuO7.4elbaT6a μ liofmuinimulam

saG q0 )h/l-aP( α )l-aP(Q

H2 01x53 4- 5.0 92

N2 01x34 1- 3.1 331

179

4.3.3 The Getter Device

As discussed in the previous sections, gas desorption from the surfaces exposed to thevacuum (filler and barrier) and gas permeation through the bag contribute to increasethe pressure in a VIP during its life.

These contributions, as well as the residual gases left in the panel after the exhaust andseal-off process, have to be taken into account to provide an accurate estimation of thepressure increase in a VIP as a function of time. Figure 4.11 shows the total pressureincrease, not including water, for some panel sizes, when data from Tables 4.2, 4.3, 4.6and 4.7 are used.

Figure 4.11 Estimated total pressure increase in some open cell foam-filled VIP as afunction of time. The additional pressure increase due to water vapour is not

considered in this calculation.


180


For given temperature and environmental conditions, the pressure build-up depends onthe specific size of the panel, and in particular on the ratio between the perimeter and thethickness, the lower the latter the higher the pressure increase. A barrier film incorporatingan aluminium foil and a panel seal-off pressure of 5 Pa have been here considered. In thisexample, water has been assumed to be completely absorbed by a proper amount ofdesiccant and not to contribute to the pressure increase.

In the case of a typical 50 x 50 x 2 cm3 size VIP, the pressure build-up exceeds 100 Paafter a few years and even not considering water, the total pressure after 20 years exceeds250 Pa.

The deterioration of the thermal conductivity for such a panel, a direct consequence ofthe increase of the internal pressure, is shown in Figure 4.12, for a PU foam followingthe curve of Figure 4.1.

Due to the nature of the thermal conductivity versus pressure curve, the deterioration ofthe λ factor is negligible at the beginning but increases steadily with time.

Figure 4.12 Increase of the thermal conductivity in a vacuum following the pressurebuild-up as per Figure 4.11. Water is supposed to be absorbed by a desiccant and not

contribute to the pressure rise.

181

In some applications, such as in domestic refrigerators and freezers, an additional gasburden may be also generated in the panel during the appliance manufacturing processor by the appliance operating conditions, which may even accelerate the deterioration ofthe thermal conductivity.

These examples show that for medium and long term applications (> 2-3 years) a simpledesiccant is not enough and specific getter materials have to be added to sorb the extraamount of gases generated in the panel.

The wide spectrum of gases present in a panel, including carbon dioxide, carbon monoxide,hydrogen, nitrogen, oxygen and water, as well as traces of blowing agents, also calls foran absorption system having high sorption capacities and adequate pumping speed forall these gases.

Zirconium-based NEG alloys, which are widely used in a variety of vacuum applications[34, 35, 36, 37] cannot be used in plastic evacuated panels due to their limited sorptioncapacity at room temperature and the need to be heat activated at relatively hightemperature (> 350 °C) prior to their use, this process being clearly not compatible withthe panel polymeric components.

Very large area physical adsorbents, like molecular sieves, zeolite or activated charcoal[38, 39] have very good efficiency for water and some organics but present serious limitationsin sorbing carbon monoxide, hydrogen, nitrogen and oxygen at the temperature and pressureconditions typically encountered in VIP applications. Therefore, they have beneficial effectsduring the very initial life of the VIP without being able to ensure long lifetime, as required,for example, by the appliance industry. They are also sensitive to the sorption temperature,i.e., the higher the temperature the lower the sorption performance, which spoils sorptionperformances in high temperature applications, e.g., 60-100 °C.

Physical adsorbents have also a second drawback which is related to their treatment,when mass production quantities have to be handled. To fully take advantage of theirsorption capacity, in fact, a relatively high temperature pre-treatment process which cleansthe surface by promoting gas desorption is required. However, after this treatment, theadsorbents are exposed again to the ambient environmental conditions during thesubsequent handling and mounting operations. Due to their reactivity, they will rapidlysorb some gases from the environment and this will reduce their capacity in a poorlyreproducible way, critically dependent on the environmental conditions, temperatureand exposure time. Fluctuations in their sorption performances as well as in the VIPinsulating efficiency could be the final result.

To eliminate the above problems and address the lifetime issues posed by the applianceindustry, a novel getter device, the COMBOGETTER, has been recently proposed [40].


182


At the heart of this device is a barium-lithium alloy, in a 1 to 4 atomic ratio [41], able toefficiently chemically absorb a large amount of nitrogen at room temperature, up tomore than 2500 Pa-l (N2)/g (alloy) [42].

This very specific feature allows the getter to compensate for the air inlet coming frompermeation, thus ensuring long lifetime requirements.

High efficiency calcium oxide and a metal oxide are also added to barium-lithium toabsorb moisture, hydrogen and some of the most common blowing agents, such as R141b and cyclopentane which could permeate through the VIP during its life. All thesematerials are prepared according to proprietary processes, which confer upon them uniqueproperties in terms of sorption efficiency and capacity. The active powders are compressedin a stainless steel cup according to a configuration which allows optimum sorptionperformances and ease of use (Figure 4.13).

The getter device does not need to be heat activated or pre-treated before being used inVIP and can be handled in air for a reasonable period of time (several minutes), duringthe panel manufacturing process, without affecting its sorption capacity.

Mounting can be accomplished by inserting the getter in a recess cut in the filling material.Alternatively, if the filler is sufficiently soft, as is the case for open cell PU, the getter canbe simply pressed into it.

Figure 4.13 Picture of the COMBOGETTER device for VIP.

183

Gas capacity data for the getter are summarised in Table 4.8.

A typical sorption curve for nitrogen at room temperature is given in Figure 4.14 (sorptionthroughput, Pa-l/s, as a function of the sorbed amount, Pa-l).

ehtfoyticapacnoitprossaG8.4elbaTRETTEGOBMOC

epytsaG 1-aP,yticapacsaG

riA 035

OC/OC 2 008

H2 0039

H2O 00008

stnegAgniwolB 031

Figure 4.14 Sorption curve for nitrogen for the COMBOGETTER (at room temperature)


184


This throughput is adequate to compensate for the relatively slow gas inlet rate occurringduring the VIP lifetime. The effect of the temperature on the sorption performance isshown in Figure 4.15 in the range 20-70 °C.

These data show that the getter is effective also at relatively high temperatures, wherephysical adsorbents become less efficient.

The getter in VIP technology has several roles, as demonstrated by the following examples:

a) Life insurance

Permeating and outgassed species (carbon dioxide, carbon monoxide, hydrogen,nitrogen, oxygen, water) which can significantly increase the pressure are irreversiblyfixed by the getter, as shown by tests of actual panels (Figure 4.16).

Pressure in VIP has been recorded by means of a viscous pressure gauge, as explainedin Section 4.4.2.2. The resulting thermal conductivity of gettered panels is thereforemuch lower in the long term, as shown in Figure 4.17, which compares the λ factorin gettered and ungettered panels after almost 3 years of ageing.

Figure 4.15 Sorption curve for nitrogen for the COMBOGETTER(in the 20-70 °C range)

185

Figure 4.16 Measurement of the pressure in two vacuum panels(with and without getter)

Figure 4.17 Measurement of the thermal conductivity in two panels (with andwithout COMBOGETTER) aged at room temperature


186


b) VIP manufacturing process

The ability of the getter to sorb air can be also used to reduce the evacuation time inpanels, this contribution being particularly interesting for medium and large sizepanels, e.g., > 0.5 m2, where achieving base pressure in the low 1 Pa can be timeconsuming for open cell foams with very small cell openings, which limit the evacuationefficiency. For each application, a trade-off between the getter capacity needed toreduce the evacuation time and that necessary to ensure long lasting serviceperformance has to be found.

A typical example of the getter’s ability to act as a chemical ‘in situ’ pump to sorbresidual gases after the evacuation process is shown in Figure 4.18, which comparesthe pressure trend in panels with and without the getter immediately after theevacuation and seal-off.

Figure 4.18 Pressure decrease in a VIP due to the in situ pumping effectof the COMBOGETTER.

187

c) Process yield and heavy duty conditions

The COMBOGETTER normalises the vacuum level in the production cycle bycompensating for fluctuations in the quality of the VIP components and themanufacturing process.

It can also compensate for the gases generated in the panel during ageing or heattreatment, as happens during storage in the warehouse, the foaming process in arefrigerator cabinet or the high temperature operating conditions required byseveral specific applications. Experimental data on vacuum panel performancesunder these conditions are presented and discussed in Section 4.6.

For some specific VIP sizes and short term applications (for example one to twoyears) a getter for air may not be necessary, since the permeation contribution isless relevant here. In these cases water is usually the most important gas to besorbed and the use of a simple dryer is sufficient to keep the pressure at therequired 10 Pa value. Calcium oxide is one of the preferred desiccants due to itsavailability, low cost, good environmental features and large water sorptioncapacity. The production process influences the calcium oxide physical structure(particle size, porosity and morphology) and thus the water sorption efficiencymay be actually an important parameter to consider in short-term applications(for example in shipping containers, where lifetimes as short as a few months orless, can be of interest). A comparison between a commercially available productand a highly efficient calcium oxide, prepared by SAES according to a proprietaryprocess is given in Figure 4.19.

Also physical absorbents, like molecular sieves and silica gel or activated carbon,can be used for short-term applications.


188


4.4 Vacuum Panel Manufacturing Process and Characterisation

4.4.1 Some Manufacturing Issues

Before being inserted in the barrier bag, the open cell foam needs preliminary pre-treatment in air to remove water and the residue of the foam production process.Volatile materials are in fact used to blow the foam chemicals and remain partiallytrapped in the foam matrix or condense as a liquid on the foam surface after the cellshave been opened.

Even though part of these volatile substances will desorb at room temperature, theremaining quantity will generate an unacceptable pressure build-up in the panel [43].The baking process is also necessary to remove the water absorbed by the PU foamwhen exposed to humid air (about 1.8 % by weight, at 23 °C, 50% relative humidity),this being particularly effective to reduce the evacuation time, as shown by tests ofactual panels [44]. Pre-treatments generally range from 150 °C for 10 minutes to 120°C for up to a few hours [9, 12, 43], depending on the foam type.

Figure 4.19 Water sorption test on two calcium oxide samples. Samples were exposedat 100-200 Pa water partial pressure at 23 °C. Sample A was prepared by SAESaccording to a proprietary process to increase material surface area and porosity.

189

After the foam has been thoroughly baked, it is important to reduce its exposure time tothe working environment to avoid significant moisture re-absorption, which takes placewhen PU foam cools down. As an example, the weight increase due to water re-absorptionduring cooling to room temperature is shown in Figure 4.20 for a small PU sample, driedat 150 °C for 20 minutes and then exposed to 50% relative humidity.

These data indicate the need to design the production process and equipment in such away that the foam is exposed to air for a very short time, not exceeding a few minutes. Alimited exposure is also beneficial to reduce the evacuation time of the panel, whichstrongly depends on the desorption rate of physisorbed water. To further minimise re-absorption of water on the foam it is advisable to process the PU in a dry or humidity-controlled area. The water content picked-up by the foam can be measured in the finishedpanel by means of residual gas analysis (RGA) with the mass spectrometer, as describedin Section 4.4.2.2. In general, the proper pre-treatment of the foam followed by adequatehandling ensures better initial insulation performance and shorter evacuation cycles whichalso mean higher productivity and lower costs per panel.

The open cell foam slabs are generally easier to handle and process than powder-basedcore materials like precipitated silica or perlite. However, precautions have to be adopted

Figure 4.20 Water re-absorption curve of a small PU sample after baking.


190


during handling of the foam to avoid powder generation which could have an impact onthe quality and reliability of the seals. Care also has to be taken during the productionprocess to ensure adequate VIP flatness, which is an important factor for the subsequentproper mounting of the panel in various applications.

Appropriate film handling and processing is also the key to produce reliable and goodquality panels.

Gas barrier properties of films, in fact, mainly depend on the number of defects, such ascracks and pinholes, present in the layered structure. This number increases as aconsequence of the mechanical stresses the film undergoes during its handling. Moreover,the high load applied by the atmospheric pressure on the vacuum panel, this value beingparticularly high in the corners and along the edges, stretches the layers and locallyincreases the defect density.

A preliminary investigation of these effects has been carried out by Sugiyama and co-authors [21] for some laminated plastic films incorporating a 6 μm aluminium foil.

Films have been analysed as such and after having been twisted to simulate handlingconditions. VIP have also been prepared and samples cut from the flat surface of thepanel, the edges and the corners to check for any local increase of the permeation rate.Tests were carried out at 23 °C on samples of 2.5 x 10-3 m2 area applying 0.0505 MPahelium pressure. The structure of the analysed films was Nylon 15 μm/vacuum metallisedPET 12 μm/aluminium 6 μm/HDPE 50 μm. The results, shown in Table 4.9 clearly indicatethat the twisting process is responsible for significantly larger permeation rates. Samplescut at the corners also show higher permeation rates, which confirms that stress conditionsin these areas increase the local pin-hole density in the aluminium foil.

tnereffidotdettimbussmlifreirrabhguorhtsetarnoitaemrepmuileH9.4elbaTstnemtaertlacinahcem

rebmunelpmaS elpmasfoepyT etarnoitaemrepeH01x)s/l-aP( 11-

1 )tnemtaerton(mlifecnerefeR 7<

2 foecafrustalfehtmorftucelpmaSlenapmuucaveht

7

3 ehtforenrocehtmorftucelpmaSlenapmuucav

054

4 detsiwtyllacinahcemelpmaS 032

191

Even though the total permeation rate is still acceptable, data suggests that suitableprecautions have to be adopted to limit as much as possible stress and film deterioration.

Similar behaviour is also expected for metallised films and has to be addressed accordingly.

4.4.2 Characterisation of Vacuum Panels

The rapid development of the open cell foam-filled vacuum panel technology has requiredparallel development and improvements of the analytical techniques necessary to assessVIP performance and reliability. This latter aspect is key for the widespread adoption ofthe technology. The selection of the best components, foam, bag and adsorbent, as wellas the careful control of the manufacturing cycle, minimises the chance of having poorlyperforming insulating panels. However, since the potential risk of defective seals or micro-leaks cannot be completely ruled out, several techniques have been developed to eithersupport and establish the VIP manufacturing cycle or to assess their quality afterproduction.

In the following sections, some techniques to check the insulating performance of vacuumpanels are illustrated, a few of them also being able to be used as a tool for QualityAssurance (QA) and Quality Control (QC) for the manufacturing process.

4.4.2.1 Measurement of the VIP Thermal Conductivity

The direct measurement of the thermal conductivity or λ factor is one of the most commonways to check the quality of a vacuum panel. Several test methods are available, such asthe heat flow meter [45], guarded hot-plate and guarded-calibrated hot box [46, 47]procedures, which measure the heat transferred under steady state conditions throughthe sample whose surfaces are kept at two different given temperatures.

The last two techniques are absolute methods since they do not require the use of referenceor calibration specimen. However they are relatively expensive and time-consuming anddifficult to adapt to rapid testing.

In the heat flow meter technique, the sample is placed between two plates controlled atdifferent temperatures. Thermocouples placed just below the plate surfaces measure thetemperature drop across the sample. One (or more) heat flow transducers mounted onthe plate measures the voltage which is proportional to the heat flow through the sample.Knowing the relationship between the voltage signal of the transducer and the heat flowthrough it, which can be obtained by calibrating the transducer to a known thermalconductivity standard, the thermal conductivity for the specimen can be obtained.


192


Heat flow meter test results generally agree well with those generated by absolute methods.The technique, which is easier to use and is a universally recognised standard test procedure[45], is also usable either for high accuracy measurements to support VIP analyticalmodelling/development or for fast tests for QA/QC programs.

This latter requirement is becoming more and more important as VIP technology movescloser to the market. Recent improvements have been based on the use of a novel dualheat flow transducer design and a tight control of plate temperatures [48].

They allow a pass/fail test on a vacuum panel to be completed within a few minutes, thusproviding the possibility for an extensive quality check during the manufacturing cycle.

This is particularly important not only for VIP manufacturers but also for the final users,to verify continued panel integrity after handling and transportation and before mounting.

Another relatively novel development in thermal conductivity measurement equipmentis based on a transient modified hot wire (MHW) heat reflectance technique [49]. In thistechnique, heat is generated by a constant electrical current in an element that has oneside in contact with the surface of the specimen and the other side in contact with abacking material. The temperature at the interface is measured and the rate of temperaturerise is related to the thermal conductivity of the sample. Quantitative data can be obtainedby calibrating the instrument with known standards in a given environment (pre-settemperature and relative humidity). Preliminary investigations by Dixon and Mathis[50] suggest that the MHW method has the potential to discriminate between panelshaving different vacuum levels, at least in the 100 and 1000 Pa range.

Thanks to the short measuring time (~5 minutes are required between two measurements),this technique also seems promising as a pass/fail test for QA/QC applications.

4.4.2.2 Measurement of the Pressure

Since the thermal performance of the panel is dependent on its internal pressure, analternative approach is to measure the pressure evolution as a function of time.

The spinning rotor gauge (SRG) technique [51], which is based on the measurement ofthe deceleration rate caused by gas friction on a freely spinning spherical metallic ball, isparticularly well suited to monitor the pressure in sealed-off devices and has been proposedin the recent years also for vacuum panel applications [5, 40, 52, 53].

The ball, or rotor, is suspended by magnetic forces between two mutually attractingpermanent magnets and additional coils provide the rotating magnetic field to spin it.

193

After the rotor has been accelerated and has achieved constant speed, the powering coilsare switched off, allowing the rotor to slow down.

A correlation between the deceleration rate and the pressure can be drawn, provided thedensity and the radius of the rotor and the mean velocity of the gas molecules impingingthe rotor surface is known. Positive features of this technique are the wide pressurerange covered, from 10-4 to 105 Pa, the low-cost sensor and the possibility of operatingseveral sensors with a single electronic unit.

The sensor can be connected to the VIP using a high vacuum compatible resin, as shownin Figure 4.21.

In the advanced SRG models, such as the SAES Getters SpiroTorr unit [54], theinstrument can operate basically at any spatial orientation on both the sensor and thegauge head.

Figure 4.21 Diagram of the VIP mounting the spinning rotor gauge. Vacuum tightnessis ensured by a high vacuum epoxy resin.


194


In particular, by mounting the sensor horizontally in the panel it is then possible tocombine the pressure readings with thermal conductivity measurements on the samespecimen.

The measuring equipment is shown in Figure 4.22.

Pressure readings are not very much dependent on the gas type for pressure in the range orbelow 10 Pa, so that the precise knowledge of the gas composition is not strictly necessaryto get a fairly accurate pressure measurement. For higher pressures, the ball (which issuspended by the magnetic field) deceleration rate becomes more dependent on the gasspecies and a precise reading requires the knowledge of the main gases involved [55].

Typical results obtained using the SRG are shown in Figure 4.16 and Figure 4.18.

Figure 4.22 Picture of the SAES Getters SpiroTorr SRG connected to a vacuum panel.

195

Due to the need to mount the sensor, this technique cannot be used as a quality controlmethod. Its role is more in the area of basic investigation and technology assessment.

Another way to measure the pressure is by puncturing the panel and analysing the gasatmosphere with a mass spectrometer. The sample is connected to the analytical benchof Figure 4.3 by a glass tube which is sealed with a high vacuum epoxy resin, as inFigure 4.23.

The tubing is then evacuated, the panel punctured, using for example a metallic hammer,and the internal atmosphere expanded to the mass spectrometer for the residual gasanalysis.

This approach allows measurement of both total and partial pressures in the panel, thusproviding useful information on the gas ratios, this being useful as an R&D tool or tomonitor and improve the manufacturing process. However, since this is a destructiveand relatively time-consuming test it can only be used for one-off samples. A non-destructive technique to measure total pressure in a vacuum panel has been proposedrecently, which is based on the use of a laser beam source coupled to a detection system.

Figure 4.23 Pictorial view of the method used to connect the analytical bench to thevacuum panel for RGA analysis.


196


The panel is placed in a vacuum chamber and the laser beam impinging on the panelsurface is reflected by the laminate and recorded by the detector. The chamber is thenslowly pumped down. When the pressure in the chamber is lower than the internal pressurein the panel, the volume of the bag increases by gas expansion, thus causing the laminatesurface to move and deflect the laser beam in a different direction. This will generate anelectrical feedback signal which is recorded by the equipment as the internal pressure ofthe vacuum panel. Present sensitivity of such types of equipment, although not wellestablished, is around some hundreds of Pa. It is likely that this value can be lowered bycalibrating the laser sensor against an absolute pressure gauge. This measuring techniqueis quite fast (about one panel per minute) and can therefore be used as a QC/QA tool ina manufacturing line.

4.5 Insulation Performances of Open Cell PU-Filled Vacuum Panels

The actual insulation efficiency of PU-based vacuum panels depends on the intrinsic corematerial conductivity, the vacuum level, the type of barrier envelope, the panel size andarrangement in the insulating structure.

All these aspects have to be optimised to get full benefits from VIP technology. Aspreviously discussed, laminates incorporating a continuous aluminium foil provide, atpresent, the best gas barrier properties for VIP. Their main drawback is the loss ofinsulation properties at the panel flanges due to the relatively high thermal conductivityof the aluminium foil which reduces the overall insulating performance, this effect beingdirectly related to the panel size and the aluminium foil thickness.

The influence of the edge effect on the insulating performance of the panel in actualapplications can be better appreciated by introducing the ECR (Energy ConsumptionReduction) factor:

ECR = (Kpu – Kvip )/ Kpu (8)

where Kpu and Kvip (W/m) are the thermal conductivity of the conventional closed cellfoam and of the vacuum panel, respectively.

The ECR factor provides the percentage increase in the insulation efficiency obtained bycompletely replacing the conventional foam with a vacuum panel having given thermalconductivity Kvip. Calculation of Kvip is not easy, since it is a complicated function of thesize of the panel, its geometry, the thermal conductivity of the core material and thethermal conductivity of the laminate (which is basically the aluminium foil, due to itshigh thermal conductivity). The heat flow through the laminate and through the opencell foam, both necessary to get Kvip, can be calculated by FEA.

197

Once Kvip is known, the ECR factor is obtained by Equation (8). The result of the ECRcalculation is shown in Figure 4.24 as a function of the aluminium thickness for somedifferent panel sizes.

It is clear from Figure 4.24 that high ECR values can only be obtained by reducing thealuminium thickness as much as possible and using relatively large vacuum panels, i.e., >0.5 m per side.

If the panel is encapsulated, as actually happens in several applications, the conventionalfoam surrounding it also has to be taken into account in the ECR calculation, since itmay contribute to heat dissipations, its effect being particularly important around theVIP flanges.

The ECR values for encapsulated panels of different sizes prepared with a laminateincorporating a 6 μm aluminium foil (Table 4.3, Film B) are given in Table 4.10. Thecalculation has been made assuming a foam thermal conductivity of 6 mW/m K and athickness of 0.5 cm of additional PU foam (thermal conductivity of 20 mW/m K) for atotal thickness of 0.025 m.

Figure 4.24 ECR values as a function of the aluminium thickness for VIP ofdifferent sizes.


198


The ECR values achievable with the aluminium foil-based laminates are very interestingfor large size panels, where the edge effect is not severe. For small panels, which are alsoimportant, since they can be easily adopted in a variety of applications, from householdappliances to shipping containers, the improvement in the insulation efficiency is not asremarkable.

The energy consumption reduction in two refrigerator cabinets, each using one panel,100 x 50 x 2 cm3 large, covering 40% and 60% of the total surface is shown in Table4.11. The same assumptions used in Table 4.10 have been made for the PU thermalconductivity and thickness of the barrier foil.

sezistnereffidfoPIVdetaluspacneniseulavRCE01.4elbaT

)mc(ezisPIV )%(RCE

2x04x04 5.61

2x05x05 7.62

2x001x05 7.63

2x001x001 1.64

In general, for household appliances, depending on the surface coverage of the cabinetand the panel thickness, energy savings from 10% to 30% have been reported usingopen cell PU foam-filled panels packaged in a 6 μm aluminium foil-based barrier [12,15, 52, 56].

egarevocPIVgniyravhtiwsrotaregirferowtrofseulavRCE11.4elbaT

)mc(ezisPIV rezeerf/rotaregirfeRegarevocecafrus

RCE

2x05x001 %06 %22

2x05x001 %04 %7.41

199

The influence of the vacuum panel thickness to the total insulation thickness ratio aswell as the change in performance, by increasing the foam λ value, has been determinedby Hamilton [54]. In general, results can be optimised by adjusting the VIP size andthickness to provide the most cost effective option.

Further improvements can be obtained by decreasing the aluminium foil thickness orusing a metal foil-free barrier and/or decreasing the thermal conductivity of the foam.The latter can be achieved by careful control of the foam microstructure and in particularthe cell shape, orientation and size, which play an essential role in determining the foamthermal conductivity and its dependence on gas pressure. Conductivity values close to 5mW/m K have been quoted for properly prepared open cell foam samples [9].

4.6 Examples of VIP Applications and Related Issues

4.6.1 Household Appliances

Refrigerators and freezers account for about 20% of the total electricity consumption ofhousehold appliances. For this reason the appliance industry is under pressure to improvethe energy efficiency of their products to cope with the need to reduce carbon dioxideemissions, as recently mandated by the Kyoto Conference. This objective has to be achievedwithout penalising product performance. Several options to decrease the energyconsumption are under evaluation, ranging from high efficiency compressors to theadoption of intelligent electronic devices [57, 58, 59].

Improvement of the insulation through the use of vacuum insulated panels filled withsilica powder or glass wool as core materials has been evaluated, and adopted in limitedamounts, by most refrigerator manufacturers [60, 61, 62], cost being the main obstacleto large sales of these products. The lower cost of the open cell foams, combined with thelow density and good handling properties, has generated renewed interest for thistechnology as an environmentally friendly option to energy consumption reduction. Asecond driving force for the appliance industry is the possibility of increasing the internalvolume of refrigerators and freezers without increasing the outer dimensions. This aspectis particularly important in Europe and Japan, where the built-in appliance market is animportant segment and space constraints play a role.

Adoption of PU-based vacuum panels in refrigerators and freezers requires the properhandling of the foaming process, which may influence the structural and vacuum propertiesof the panel.


200


In fact, during the appliance manufacturing cycle, the panels are glued to the cabinetwalls and then foamed in place with a conventional closed cell PU foam, this processbeing necessary to further improve the overall structural and thermal insulating propertiesof the cabinet. To allow a better flowability of the injected chemicals, the cabinets aregenerally pre-heated at about 45 °C for some minutes. The presence of the panels, whichreduces the free flow and expansion of the foam, requires special care to avoid generationof air-filled void volumes which would deteriorate the insulation performance [56]. Toprevent this, and provide the best integration between VIP and conventional PU foams,specific systems have been recently developed to be used in conjunction with VIP [10].As a combined effect of the cabinet pre-heating and the exothermicity of the foamingreaction, temperatures even higher than 100-120 °C can be reached on the VIP surface,leading to temperature-enhanced gas desorption. A typical result using a 50 x 50 x 2 cm3

panel is shown in Figure 4.25.

Figure 4.25 Measurement of the temperature increase experienced by the two panelsurfaces during the foaming process.

201

This gas contribution has to be evaluated case by case since it depends strictly on theactual foaming conditions, such as the chemicals used, the process variables, therefrigerator design and VIP geometry. However, under some circumstances, the pressureincrease can be a measurable fraction of the maximum acceptable level, thus causing adeterioration of the VIP thermal insulation properties from the very beginning of therefrigerator life.

As an example, Figure 4.26 shows the effect of a thermal treatment carried out at 50, 70and 90 °C on a VIP prepared with an open cell foam and an aluminium foil-containinglaminate. The foam was pre-baked at 150 °C for 20 minutes and no absorbent was used.

After the preparation, the panel was put in a oven and kept at the indicated temperaturefor 15 to 20 minutes. The panel was then removed and the total pressure measured as afunction of time. To allow the continuous monitoring of the pressure, a SpiroTorr spinningrotor gauge was mounted in the panel as shown in Figure 4.21.

Desorbed gases after the thermal treatments are only partially reabsorbed by the foamand generate a measurable pressure build up in the VIP.

Figure 4.26 pressure evolution in a panel submitted to various heat treatments.Pressure readings are taken by means of a SpiroTorr SRG (see Figure 4.3).


202


The effect of the temperature increase resulting from the foaming process, on the vacuumlevel in a panel can also be seen in Table 4.12 which shows the result of an RGA carriedout on two test panels encapsulated by an appliance manufacturer. In the getterless panel,the pressure immediately after the foaming is already exceeding the 10 Pa target valueand close to the maximum acceptable value (generally set at 50 Pa).

The deterioration of the vacuum during the foaming process has been recently addressedby Kücükpinar and co-workers who ran specific tests aimed at quantifying this effect inopen cell foam-filled VIP with and without getters [53].

As shown in this study, the COMBOGETTER was able to compensate for the extra gasload generated during the foaming process, thus confirming the results in Table 4.12.

Further specific outgassing studies are necessary to understand the mechanisms of gasgeneration during the foaming process, even though this investigation presents severaldifficulties because of the large number of parameters involved.

The foaming process may also stress the laminate due to the combined effect of temperatureand applied pressure on critical areas such as the panel edges and the corners. Defectsgenerated here may in turn increase the gas diffusion inside the panel, reducing the designedlifetime.

Another important issue in refrigerator/freezer applications is longevity, since 15 to 20years lifetime is targeted by most manufacturers.

Durability tests on panels are therefore necessary to assess VIP reliability and performanceover time. Figure 4.27 shows the pressure values in some encapsulated panels, as measuredby RGA. Panel A, containing a COMBOGETTER, was aged at room temperature forsome weeks and then kept at 40 °C for more than 30 months. Panels B and C did nothave any adsorbent. The latter was aged at 40 °C for about 3 months, while the formerwas unaged.

dnahtiwslenapdetaluspacnenisisylanAsaGlaudiseR21.4elbaTrettegehttuohtiw

PIVdetaluspacnE erusserPlatoT)aP(

)aP(erusserPlaitraP

riA OC/OC 2 H2O H2

rettegtuohtiW 04 6 92 5 1.0

retteghtiW 7.0 3.0 10.0 3.0 1.0

203

The pressure in panel A is well below the 10 Pa target value as compared to panel Cwhose pressure is one order of magnitude higher. Pressure in panel B, which was justencapsulated, was slightly lower than in panel C. For comparison, the calculated pressureincrease in the panel, as predicted based on the model discussed in Section 4.3, is alsoshown. This result confirms the role of the getter to compensate for the encapsulationprocess and to ensure VIP longevity and also supports the gas load model previouslydescribed.

Still, a better assessment is necessary through longer term tests and, possibly, throughaccelerated tests able to provide experimental data within a shorter time. Activity isongoing in SAES Getters Laboratories, as well as in various other research centres toprovide such an assessment.

4.6.2 Laboratory and Biomedical Refrigerators

This is very special equipment designed to operate at very low temperatures, e.g., from-30 °C to -86 °C, to age samples or to store valuable and perishable goods, like organsand tissues, biological and medical samples or vaccines. Vacuum panels are used mainlyto increase the internal storage volume without increasing the energy consumption.

Figure 4.27 Residual gas analysis in panels encapsulated and aged for different times.Panel A contains one COMBOGETTER, panels B and C are without.


204


Since the conventional insulation must be very thick to ensure the achievement of suchlow temperatures, a partial replacement of the conventional insulator with VIPcontributes significantly to increase the internal volume, from 20% up to 40% ormore in specific models.

As a second advantage, panels provide an overall superior passive insulation, whichmeans that the temperature rise in the case of power failure is less steep and more timeis necessary before a given critical temperature is reached. This provides extra-safetyfor delicate articles which may rapidly deteriorate when their temperature exceeds agiven value.

Due to the generally higher insulation thickness, the foaming process here may be evenmore exothermic than in domestic appliances. Means have therefore to be taken to keepthe VIP temperature as low as possible during the encapsulation.

Ultra-low temperature freezer models using foam filled vacuum panels were successfullyplaced on the market by a leading company some years ago and other companies areexpected to follow soon.

4.6.3 Vending Machines

Vending machines are especially popular in Japan, Korea and the Far East. It is estimatedthat in Japan about 2.5 million of these appliances are in the field with a replacementmarket of about 0.4 million pieces/year.

VIP are finding widespread use in vending machines since most manufacturers areextensively using one or two panels to separate the hot and cold beverage compartments.Also in this case, the main driving force for VIP adoption has been the possibility toincrease the internal volume for the storage of beverages, rather than the improvementin energy efficiency.

Vending machines are quite a demanding application for VIP, since the panel operatingtemperature is cycled between room temperature and 60-70 °C and the lifetime is 5years. As already discussed, the high temperature promotes higher diffusion and outgassingrates and provides additional mechanical stress to the envelope.

Ageing tests have been carried out by vending machines manufacturers to assess VIPusability and reliability in this application.

Results obtained in SAES Getters Laboratories for 50 x 50 x 2 cm3 panels aged forseveral months at 60 °C are shown in Figure 4.28.

205

Again, the deterioration of the thermal conductivity, due to the pressure increase, is quitedramatic in the getterless panels both at 23 and 60 °C. As expected, the thermal conductivityincrease is much more rapid for the panels aged at 60 °C than for the panel kept at roomtemperature. Only minimal differences are measured in the panels containing theCOMBOGETTER, which could compensate for the increased gas load at both temperatures.

4.6.4 Refrigerated/Insulated Transportation

Large vacuum insulated containers, ships and trucks are under evaluation by variouscompanies, mainly in Europe and Japan.

The main advantages offered by using VIP are space saving and better insulation, so asto keep the temperature increase rate as low as possible even in the absence of a powersource. The lifespan is generally from 10 to 15 years.

Since very large VIP, e.g., 2-3 m2 size are considered for this application, the edge effect(see Section 4.3.2.1) is here a marginal issue.

Figure 4.28 Measurement of the thermal conductivity as a function of time for panelsaged at 23 and 60 °C, with and without the COMBOGETTER.


206


Small commercial shipping containers to store and deliver pharmaceuticals, frozen foodand valuable temperature-sensitive products is also another interesting area for the useof VIP.

Long lifespan is not a problem, since most of these shipping containers have a projectedlife of the order of weeks or months. Therefore, it is even possible to use the VIP morethan once, the specifications on the permeation properties of the barrier film are lesstight and the getter not necessary. A specific dryer, to adsorb water, may be used in mostcases. The small size VIP and the limited projected life suggest the possibility of usingmetallised barriers for the skin of the panel in these applications, at least on one side, soas to overcome the edge effect, which, in this case is very critical.

4.6.5 Other Applications

A variety of other potential applications exist for VIP, from cold stores to insulation inbuildings and in industrial plants, e.g., industrial reactors or liquid natural gas tanks.Also super insulation at relatively high temperature applications for water heaters, heatpipes or boiling pots have been considered and are under evaluation. The key issue of theadoption of VIP in this last family of applications is the possibility to bend the open cellPU filled VIP into a round or cylindrical shape. The combination of high temperatureand bending provides additional challenges to VIP technology and further improvementsin the material selection and processing might be required.

4.7 Near Term Perspectives and Conclusions

The open cell PU foam is a very promising filler material for VIP. It shows good vacuumcompatibility, thermal conductivity, light weight, ease of handling and processing andmoderate cost. To allow its use, significant improvements have been made recently in thebarrier and getter technology, as well as in the panel manufacturing and testing procedures.

Laminated barriers, incorporating a 6 μm continuous aluminium foil, are providing areasonable trade-off between permeation properties and energy saving, especially in largesize panels (~1 m2), where the edge effect is less important. These films can therefore beused in applications such as refrigerators, freezers, vending machines and insulatedtransportation, where 10-20 years lifetime and relatively large panels are required.Traditional metallised or multilayered plastic barriers, even though more appealing fromthe energy saving point of view, are not advisable in long term applications, due to theirpoorer gas barrier properties. However, the need to improve performance and VIP costeffectiveness is pushing the film industry to further improve their products either by

207

reducing the aluminium foil thickness to 5 μm and below or developing aluminium-freecomposite barriers. If successful, this latter solution would dramatically increase thevacuum panel insulation efficiency. New families of improved metallised products havebeen developed very recently and proposed for testing and are under evaluation in severallaboratories.

The outgassing and permeation data, on both components and finished panels, and theuse of models quoted in the literature, allow the evaluation of the total gas load and thepressure increase in a vacuum panel as a function of time, size and operating conditions.This in turn is beneficial for designing the most suitable solution for the absorbent.

A high capacity getter system, the COMBOGETTER, capable of adsorbing air, moistureand the other gases of interest without the need for any pre-treatment has been specificallydeveloped for VIP application. Its roles are many since it chemically absorbs residual,outgassed and permeating species and provides the means to shorten the evacuationprocess and increase the manufacturing yield.

Mass production technologies to manufacture high quality open cell foam panels are onthe market. Several techniques are also becoming available for the fast measurement ofthe thermal conductivity. This is a very important issue to build a QA/QC system able toensure the quality and the reliability of the products, not only after production but alsoimmediately before use.

As far as the applications are concerned, open cell PU foams are finding their place in themarket for both low and high temperature appliances, from refrigerators to vendingmachines. Present production volumes are still limited but near term perspectives areencouraging, especially in Asia and Europe. The strong push for energy reduction andthe parallel continuous refinement of the technology should provide further motivationto VIP adoption.

Cost is still the main obstacle to the widespread adoption of this technology. A preliminaryqualitative cost comparison between silica and open cell PU, which includes raw material,panel manufacture and panel installation costs [10] shows that a cost reduction of aboutone-third over silica could be achieved. This has to be improved to bring the cost of thevacuum panel down further, so as to really interest more segments of the insulationindustry.

Very recent breakthroughs in the foam mass production technologies seem to indicatethat this target is reachable.

Even though far from being complete, the material presented in this chapter shows thatsignificant technological achievements have been obtained during the last few years.


208


Thanks to this, vacuum panel technology is becoming a technically viable and cost effectivesolution to the need to reduce energy consumption in household appliances and incommercial and industrial applications. To successfully achieve this target, additionalefforts are necessary to further improve the component quality and reliability (foam,film and adsorbent), to optimise panel production and to reduce costs.

Acknowledgements

The author would like to thank Dr. Paolo della Porta, President and CEO of SAES GettersSpA, for his continuous support and long lasting commitment to VIP technology. Theauthor also acknowledges Dr. Bruno Ferrario, Corporate R&D Director, for fruitfuldiscussions and suggestions.

Special thanks are given to Dr. Roberto Caloi and Dr. Enea Rizzi for their valuabletechnical support in the development of theoretical models and in running experimentalmeasurements on VIP components and finished devices.

References

1. H. A. Fine, Journal of Thermal Insulation, 1989, 12, 183.

2. D. W. Yarbrough, R. S. Graves, F. J. Weaver and D. L. McElroy, presented at ORNL/CON-215, Oak Ridge National Laboratory, Oak Ridge, TN, Sept. 1986, p.1.

3. L. R. Glicksman and N. Solomou, Presented at the SPI Polyurethanes Conference,Boston, MA, 1994, p.182.

4. H. M. Strong, F. P. Bundy and H. P. Bovenkerk, Journal of Applied Physics,1960, 31, 1, 39.

5. B. Veltkamp and P. Hoogendoorn, Vuoto, 1999, 28, 1-2, 27.

6. J. Fricke, Presented at the 13th European Conference on ThermophysicalProperties, Lisboa, Portugal, 1993, p.207.

7. R. L. Tabor, Presented at the SPI Polyurethanes Conference, Boston, MA, 1994,p.288.

8. G. Biesmans, D. Randall, E. Francais and M. Perrut, Presented at thePolyurethanes World Congress ’97, Amsterdam, The Netherlands, 1997, p.294.

209

9. K. Kodama, K. Yuge, Y. Masuda and Y. Tanimoto, Presented at thePolyurethanes World Congress, October 1993, p.140.

10. R. De Vos and I. D. Rosbotham, Cellular Polymers, 1994, 13, 2, 147.

11. K. W. Dietrich and D. W. McCullogh, Presented at UTECH 96, The Hague, TheNetherlands, 1996, Paper No.64.

12. W. H. Tao, W. F. Sung and J. Y. Lin, Journal of Cellular Plastics, 1997, 33, 3,545.

13. P. Pendergast and B. Malone, Vuoto, 1999, 28, 1-2, 27.

14. L. R. Glicksman in Low Density Cellular Plastics: Physical Basis of Behaviour,Ed., N. C. Hilyard and A. Cunningham, Chapman & Hall, London, 1994,Chapter 5, p.104.

15. W. Wacker, A. Christfreund, D. Randall and N. W. Keane, Presented at thePolyurethanes Expo ’96, Las Vegas, NV, 1996, p.35.

16. A. Roth, Vacuum Technology, North Holland Publishing Company, Amsterdam,1982, 186.

17. F. Sciuccati, G. Gasparini and B. Ferrario, Vuoto, 1988, 17, 4, 345.

18. P. A. Redhead, Vacuum, 1962, 12, 203.

19. P. Mercea, L. Muresan and V. Mecea, Journal of Membrane Science, 1985, 24,297.

20. T. A. Beu and P. V. Mercea, Materials Chemistry and Physics, 1990, 26, 309.

21. A. Sugiyama, H. Tada and M. Yoshimoto, Vuoto, 1999, 28, 1-2, 51.

22. ASTM Standard E96-00 Standard Test Methods for Water Vapor Transmission ofMaterials.

23. ASTM Standard F 1249-90 (1995) Standard Test Method for Water VapourTransmission Rate Through Plastic Film and Sheeting Using a ModulatedInfrared Sensor.

24. P. Manini, inventor; Saes Getters SpA, assignee; WO 9803850A1, 1998.

25. P. Manini, Journal of Cellular Plastics, 1999, 35, 5, 403.


210


26. P. Manini and E. Rizzi, Poster presented at the European Workshop on VacuumPanel Technology for Super Insulation in Domestic Appliances and IndustrialApplications, Milano, Italy, 1998.

27. W. Lamb and R. Zeiler, Presented at the Vacuum Insulation Panel Symposium,Baltimore, MD, 1999. Paper available from the SAES Getters web site, http://www.saesgetters.com.

28. J. Bonekamp, Presented at the Vacuum Insulation Panel Symposium, Baltimore,MD, 1999. Paper available from the SAES Getters web site, http://www.saesgetters.com.

29. R. Juran, J. Covington and J. Carley, Modern Plastics Encyclopedia, McGrawHill, NewYork, 1987, 64, 10A.

30. Polymer Handbook, 3rd Edition, Ed., J. Brandrup and E. H. Immergut, JohnWiley and Sons, New York, 1989.

31. S. J. Adam and C. E. David, Presented at the 23rd International SAMPETechnical Conference, New York, NY, 1991.

32. K. E. Wilkes, W. A. Gabbard and F. J. Weaver, ORNL/CP-95971, Oak RidgeNational Laboratory, TN, 1998.

33. G. Biesmans, R. De Vos and I. D. Rosbotham, Presented at the PolyurethanesWorld Congress 1993, Vancouver, BC, 1993, 498.

34. A.Barosi in International Symposium on Residual Gases in Electron Tubes, Ed.,T. A. Giorgi and P. della Porta, Academic Press, London, 1972, 221.

35. C. Boffito, B. Ferrario, P. della Porta and L. Rosai, Journal of Vacuum Scienceand Technology, 1981, 18, 1117.

36. B. Ferrario, A.Figini and M.Borghi, Vacuum, 1984, 35, 13.

37. C. Boffito, B. Ferrario, L. Rosai and F. Doni, Journal of Vacuum Science andTechnology, 1987, A5, 3442.

38. D. W. Breck, Zeolite Molecular Sieves, R. E. Krieger Publishing Company,Malabar, Florida, 1984, Chapter 8.

39. J. S. Mattson and H. B. Mark, Activated Carbon: Surface Chemistry andAdsorption from Solution, Marcel Dekker, New York, 1971.

211

40. P. Manini, Vuoto, 1997, 27, 2, 45.

41. C. Boffito and A. Schiabel, inventors; Getters SpA, assignee; European Patent0514348 A1, 1991.

42. R. M. Caloi, P. Manini, S. Valdrè, E. Magnano, J. Kovac, E. Narducci and M.Sancrotti, Journal of Vacuum Science and Technology, 1999, A17, 2696.

43. R. De Vos, I. D. Rosbotham and J. Deschaght, Presented at the PolyurethanesWorld Congress, 1994, Boston MA, p.194.

44. J. Akita, Vuoto, 1999, 28, 1-2, 59.

45. ASTM Standard C518-98, Standard Test Method for Steady-State ThermalTransmission Properties by Means of the Heat Flow Meter Apparatus.

46. ASTM Standard C236-89 (1993) e1, Standard Test Method for Steady-StateThermal Performance of Building Assemblies by Means of a Guarded Hot Box.

47. ASTM Standard C177-97, Standard Test Method for Steady-State Heat FluxMeasurements and Thermal Transmission Properties by Means of the Guarded-Hot-Plate Apparatus.

48. S. Smith and C. Urso, Vuoto, 1999, 28, 1-2, 64.

49. N. E. Mathis, Measurement of Thermal Conductivity Anisotropy in PolymerMaterials, Chemical Engineering Department, University of New Brunswick,Fredericton, NB, Canada, 1996, Ph.D. Thesis.

50. C. Dixon and N. Mathis, Presented at the Polyurethanes Expo ‘99, 1999, p.607.

51. J. K. Fremerey, Journal of Vacuum Science and Technology, 1985, A3, 3, 1715.

52. E. Kücükpinar, H. Güclü, A. S. Akkurt and F. Özkadi, Vuoto, 1999, 28, 1-2, 31.

53. E. Kücükpinar, F. Özkadi , A. Soyal and H. Güclü, Presented at the PolyurethanesExpo ‘99, 1999, p.339.

54. C. Boffito, M. Moraja and G. Pastore, Journal of Vacuum Science andTechnology, 1997, 15, 4, 2391.

55. B. E. Lindenau and J. K. Fremerey, Journal of Vacuum Science and Technology,1991, A9, 5, 2737.


212


54. A. J. Hamilton, Vuoto, 1999, 28, 1-2, 27.

55. H. A. Fine and J. Lupinacci, Energy-Efficient Refrigerator Prototype Test Results,EPA-43D-R-94-011, June 1994.

56. E. A. Vineyard, J. R. Sand and R. H. Bohman, ASHRAE Transactions, 1995, 1,1422.

57. E. A. Vineyard and J. R. Sand, presented at ORNL/CP-97450, Oak RidgeNational Laboratory, TN, 1998.

58. G. J. Haworth, R. Srikanth and H. A. Fine, Presented at the 44th AnnualInternational Appliance Technical Conference, Ohio, USA, 1993, p.1.

59. E. A. Vineyard, T. Stovall, K. E. Wilkes and K. W. Childs, ASHRAE Transactions,1998, 104, 2.

60. T. F. Potter, D. K. Benson and L. K. Smith, Presented at ACEEE 1988,Proceedings from the Panel on Appliances and Equipment,1998, 4, 4.86.

213

5 Modelling the Stabilising Behaviour of SiliconeSurfactants During the Processing ofPolyurethane Foam: The Use of Thin Liquid Films

Steven A. Snow, Udo C. Pernisz, Benjamin M. Nugent, Robert E. Stevens,Richard J. Braun and Shailesh Naire

5.1 Introduction

The commercial production of polyurethane (PU) foam requires a tremendous diversityof ingredients. Common ingredients include isocyanates, polyether polyols, water, catalystsand stabilisers (surfactants). Most of the surfactants used in the foaming process aresilicone surfactants, the major suppliers being Th. Goldschmidt AG, Witco and AirProducts and Chemicals, Inc. Worldwide volume for silicone surfactants in polyurethanefoam has been recently estimated at 30,000 tonnes/year [1]. In practice, a wide range ofsilicone surfactant structures are necessary to produce PU foam with a sufficiently broadrange of physical properties. This structural diversity also reflects the need for thesurfactant to perform many different tasks in foam formulations. The physical behaviourof these surfactants in the PU foam formulation process [2] has recently been reviewed.These roles of the surfactant include reducing surface (interfacial) tension [3-11], alteringsurface viscoelasticity [5, 7, 11, 12-15], and modifying polymer reaction kinetics andmorphology [16-21]. The focus of this chapter is on the mechanism of PU foamstabilisation and how it is affected by silicone surfactants.

In order to understand a complicated physical mechanism such as PU foam stabilisation,one must first be concerned with the thermodynamics and kinetics of the process. As aPU (or any) foam rises, energy is absorbed by the system and converted into a combinationof excess gravitational and surface energy. For PU foams, this excess energy is created byboth high-shear mixing and energy transfer from the chemical polymerisation reaction(between isocyanates and either polyols or water). One function of the silicone surfactantis to stabilise this foam until it is a self-supporting, gelatinous froth. This froth canfurther harden, yielding elastomeric or rigid foam products. Once hardened, the foamstill contains excess gravitational and surface energy; however, the rate of dissipation ofthis energy is extremely low. Therefore, when discussing foam ‘stability’, one is specificallyconsidering the rate of foam degradation.

Foam degradation involves both physical and energetic changes in the foam. Four physicalprocesses are commonly associated with PU foam degradation:

214


• a decrease in foam volume,

• a decrease in the number of bubbles (cells),

• an increase in average cell size, and

• a decrease in the dispersity of cell sizes.

These physical processes involve the release of excess gravitational and surface energy.

Excess gravitational energy is released during the drainage of liquid from the foam. Theliquid can drain from between the bubbles, thinning the liquid membrane known as thelamella. The liquid can also drain from the plateau borders in the foam. These bordersare the relatively wide channels in the foam created at the intersection point of threelamellae. Mysels, Frankel and Shinoda [22] estimated that, in aqueous foams, the amountof gravitational energy released during degradation was five orders of magnitude lessthan the amount of surface energy released. It is assumed that this estimation roughlyholds for the degradation of PU foams; therefore, the focus here will be on processes thatreduce the surface energy of PU foams.

Two processes, the diffusion of gas from small to larger bubbles (Ostwald Ripening) [23]and bubble coalescence, reduce the surface area of the foam and, therefore, the surfaceenergy. Ostwald ripening involves the diffusive molecular transport of gas from small tolarge(r) bubbles. Ostwald ripening in foams leads to the counterintuitive phenomena oflarge bubbles in a foam expanding as smaller ones shrink. This process results both in areduction in the number of bubbles and in the dispersity of bubble sizes. However,Owen [12] has demonstrated that Ostwald ripening plays only a minor role in PU foamstability; therefore, this chapter will focus on the process of bubble coalescence.

Bubble coalescence involves the merging of two distinct bubbles into one, due to therupture and recession of the lamella. Early in the PU foaming process, the rupture of alamella at the edge of the PU foam will cause bubble collapse as the gas inside is lost tothe surrounding atmosphere. This process will reduce the number of bubbles presentand the foam volume. Lamella rupture within the interior of the foam will cause bubblecoalescence, decreasing the number of bubbles, increasing the average bubble size, butnot necessarily decreasing the foam volume. Later in the foaming process, as the PU gelnetwork sets up, the lamella may rupture but the two bubbles can retain their separateidentities. This process is known as cell opening. If many of the lamellae have rupturedin this fashion, the object is more correctly called a sponge instead of a foam. In thiscase, air can easily be pushed out of the sponge if it is loaded with weight (as for examplewhen one sits on PU foam padding on a car seat). Therefore, the mechanical propertiesof the PU foam depend greatly on the degree of cell opening present. There is generalagreement that a strong stabilising effect from the surfactant is necessary early in the

215

Modelling the Stabilising Behaviour of Silicone Surfactants…

foam process. The stabilising role, if any, of the surfactant late in the foaming processremains a controversial issue [2, 19-21, 23-27]. This chapter will focus on the earlierstages of the foaming process, before the gelling of the liquid PU phase.

The mechanism of rupture depends upon the thickness of the lamella film. Ultra-thinfilms, with thicknesses in the 1-20 nm range, are believed to rupture due to theamplification of thermally-generated surface waves [28-30]. However, studies by Artaviaand Macosko [31] and Akabori and Fujimoto [32] demonstrated that most of the rupturedlamellae in a PU foam are in the 200-1000 nm thickness range. Therefore, it is assumedhere that the mechanism of rupture of these thicker films is most important whenconsidering PU foam.

It has been extensively demonstrated that ‘thick’ liquid films rupture due to theformation (nucleation) of a pinhole in the film which spreads outward radially untilthe film is destroyed. The nucleation of the pinhole could result from either the presenceof fluid instabilities in the film (for example, Marangoni, Rayleigh-Taylor or Benardinstabilities) [33] or else by the presence of immiscible microphases which could serveas nucleation sites.

For PU foams, it has been proposed that film rupture is nucleated by the phase separationof polyurea segments [19-21, 24-26]. It was postulated that the phase separation processcould be retarded by the presence of silicone surfactants containing a high percentage ofpolyoxyethylene. The polyoxyethylene segment of the surfactant could interact stronglywith polyurea via the formation of a hydrogen bonding network. However, in caseswhere the silicone contains lesser amounts of polyoxyethylene, such as in many commercialfoam stabilisers, a direct effect of the concentration and structure of the silicone surfactanton this phase separation process has not been found [2, 26].

Circumstantial evidence of the influence of Marangoni instabilities on PU foam stability,and presumably film stability and the nucleation of pinholes, has been amply demonstrated[34]. Marangoni instability in a fluid film refers to a situation where a gradient of surfacetension exists, stimulating compensating surface flows. For example, the stretching ofthe film reduces the surface concentration and, therefore, increases the surface tension.Furthermore, this stretched and thinned spot would be prone to rupture. However,within a certain range of surfactant concentration, this thin spot can re-thicken due tothe diffusion/convection of surfactant and underlying bulk liquid to the thin spot(Marangoni flow).

Once nucleated, the film hole can spread out radially at a rate inversely proportional tothe bulk viscosity of the film. Spreading can occur until the film is destroyed. Forexample, in typical aqueous films, the rate of spreading is so high that high-speedphotography is necessary to capture the event. However, in a gelling PU foam, the rate

216


of spreading can be very low, and, in fact, films can have multiple holes in them withoutbeing destroyed [24-26, 32].

Detailed investigations on the effect of silicone surfactants on the nucleation and growthof holes in PU films have not been reported. Furthermore, that is not the topic of thischapter which is specifically concerned with the effect of the silicone surfactant on therate of drainage of PU films. However, this topic is highly related to film rupture as it hasbeen shown that the probability of film rupture is an inverse squared function of the filmthickness [35]. The time-dependent film thickness is a function of the drainage rate ofthe film.

5.2 Film Drainage Rate: Reynold’s Model and Further Modifications

5.2.1 Rigid Film Surfaces

Thin film drainage has been investigated in detail both theoretically and experimentally[22, 28-30, 35]. An extensively applied physical model of a draining thin film is the oneof Reynold’s, a form of which is shown in Equation 1:

V Fk RRe = 2 33 4πμ (1)

where VRe is the velocity of film thinning, F is the external force on the film (causingdrainage), k is the film thickness, μ is the dynamic viscosity, and R is the film radius.

The Reynold’s model assumes that:

• the liquid inside the film is Newtonian in its rheological behaviour.

• the surfaces of the film are circular, immovable, parallel plates; surface velocity equalszero, surface viscosity is infinite.

• the flow inside the film is laminar.

• the flow inside the film is driven by gravity and/or pressure gradients and is resistedby the bulk viscosity of the liquid.

The dependencies of drainage (flow) rate on the variables of film thickness, area andbulk viscosity are quantified in Equation 1. For example, with decreasing film thicknessk, the drainage rate decreases in proportion with k3; furthermore the drainage ratedecreases with increasing viscosity.

217


Previous discussions [34] of PU film drainage rate (and ultimately foam stability and cellopening) focused on the dependence of rate on bulk viscosity. Per Reynold’s model, factorswhich increased the bulk viscosity, such as catalyst concentration and polyol reactivity,presumably decreased film drainage rate and yielded a more stable, less porous foam.

5.2.2 Mobile Film Surfaces

The Reynold’s model and equation have also been modified to account for surface effectson film drainage rate [29]. To accomplish this, the assumption (boundary condition) ofa surface velocity of zero must be relaxed. This change also decreases the surface viscosityfrom an infinite to a finite value. This process yields Equation 2. Allowing for a finitesurface velocity increases the film drainage rate from that which would be expectedunder the Reynold’s conditions.

V V ef/ /Re = +1 1 (2)

Where V/VRe is the measured velocity of thinning divided by the velocity of thinningunder the conditions of the Reynold’s equation and ef is the surface mobility factor.

The surface mobility factor is analysed in detail in [36]. This analysis demonstrates thattwo factors control the degree of surface mobility: the surface viscosity and the presenceof surface tension gradients.

5.2.3 Surface Viscosity

The rates of thinning of vertically-supported, thin liquid films of polyol solutions ofvarious silicone surfactants have been measured [37]. It was found that the rate for filmsstabilised by a trimethylsilyl-capped polysilicate (TCP; a highly branched silicone notcontaining polyethers), was much lower than that for the films stabilised by commonsilicone polyether copolymer surfactants. The retardation of drainage rate was correlatedwith an increase in surface viscosity. Furthermore, it was noted that PU foams preparedusing TCP were significantly more stable than those containing the commercial surfactants.

Surface viscosity scales directly with the surfactant surface concentration, theintermolecular cohesion between the surface molecules and the intermolecular adhesionbetween the surfactant molecules and the underlying bulk liquid layer.

Increasing the surface concentration decreases the average distance between molecules,therefore increasing the sum value of the attractive molecular forces between them. In

218


many cases, high surface concentration yields a low film drainage rate and a highlystabilised foam.

The intermolecular cohesion is the result of the balance of attractive (van der Waals,polar bonding) and repulsive (ionic) intermolecular forces. The total cohesive energy is aproduct of the total area of cohesion and the cohesive energy per unit area. For example,the cohesion of the polyether chains of a typical silicone surfactant yields most of thesurface viscosity contribution in PU films and foams [38]. Although it has not beeninvestigated, one could speculate that the surface viscosity would scale with the length ofthe polyether chain. Increasing the chain length increases the potential area of contactbetween the chains.

The unit area cohesive energy is a function of the strength of the intermolecular bondspresent. For example, silicone surfactant molecules could possibly cohere via van derWaals bonding of the methyl groups of the siloxane backbone or else via polar bondingbetween the ethylene oxide units in the polyether chain. The van der Waals bondingbetween methyl groups is known to be weak; consequently, the surface viscosity ofunsubstituted PDMS is close to zero [15, 38]. However, the polar bonding between thepolyether chains would be much stronger and, therefore, the source of surface viscositymeasured in silicone polyether copolymers.

Intermolecular adhesion between the polyether chains of a typical silicone surfactantand the underlying ‘polyurethane’ liquid matrix could increase the surface viscositycontribution in PU films and foams.

In order to maximise surface viscosity, and therefore minimise the drainage rate, thesurfactant concentration, the intermolecular cohesion and adhesion should all be high.Later in the chapter evidence is presented correlating silicone surfactant concentrationand structure (which influences the intermolecular cohesion and adhesion) to surfaceviscosity and film drainage rate.

5.2.4 Surface Tension Gradients

An idealised model of the surface of a thin liquid film is one of a monolayer of evenly-distributed surfactant molecules. However, a more realistic model is one where themolecules are not evenly distributed; therefore, the surface concentration depends onsurface position. The result of this heterogeneous distribution is that gradients of surfaceconcentration, and therefore surface tension, are present. One example of this was pointedout in Section 5.1 on the effect of Marangoni instabilities on film rupture. Regardingfilm drainage, a surface tension gradient exerts a surface stress that can either impede or

219


accelerate the underlying bulk flow (drainage). In thin films, these gradients typicallymake the surface rigid, retarding flow and therefore decreasing, drainage rate. Thegradients can be relieved by the diffusion of surfactant from areas of high to lowconcentration; the drainage rate of the film is proportionately increased.

Diffusive surfactant fluxes are functions of the intensity of the concentration gradientand the diffusion coefficient of the surfactant. The intensity of the concentration gradientdepends upon the overall surfactant concentration. At low or high concentrations, thesegradients are relatively weak. At intermediate concentrations, they are quite strong.

The diffusion coefficient of the surfactant is a function of its size and shape. The mostsimple and common case to analyse is where the surfactant assumes a spherical shape.In this case, the diffusion coefficient of the surfactant scales inversely with the solvatedmolecular volume (Stoke’s law).

Overall, to accelerate film drainage, it is useful to have a high concentration of surfactantwhose surface partition coefficient is also high. The surface partition coefficient is ameasure of the tendency of the surfactant to adsorb at the surface instead of remainingin the bulk. In the case of silicone polyether surfactants, the coefficient scales with theratio of silicone to polyether. Under these conditions (high surfactant concentration,high ratio of silicone to polyether) the flux of surfactant to the surface, relieving surfacetension gradients, will also be high. A large diffusion coefficient (for the surfactantmolecule, a small size and a compact shape) is also helpful. Later in this chaptercorrelations of silicone surfactant concentration, partition coefficient, and diffusioncoefficient to film drainage rate will be discussed.

5.3 Experimental Investigation of Model, Thin Liquid PolyurethaneFilms and the Development of Qualitative and Semi-QuantitativeModels of Film Drainage

This section begins with a qualitative description of thin liquid PU films. This initialinvestigation had five goals in mind: to confirm that stable, vertically-oriented, thin liquidfilms could be prepared using mixtures of ingredients designed to model a PU foam, tostudy the hydrodynamic phenomena in the films, to compare the physical behaviour ofthese films to the behaviour of the more common aqueous soap films, to observe specificsurfactant effects on the properties of these films, and to extrapolate conclusions aboutthe behaviour of these films to operational PU foam.

After this qualitative description, an accurate measurement of the drainage rates of thesefilms was sought in order to study the effect of bulk and surface variables on the rate. In

220


order to make an accurate measurement, a novel interferometric method was developedand implemented. This method was validated by the discovery that film drainage ratescaled with the reciprocal of bulk viscosity as predicted by Reynold’s equation.

Once this method was validated, the effect of silicone surfactant concentration and structureon the drainage rate of the films was investigated. Generally, the drainage rate displayed amaximum as a function of surfactant concentration. This maximum was consistent with aphysical model where the two major influences on drainage rate are surface tension gradientsand surface viscosity. As surfactant concentration increased, the retarding effect of gradientsdecreased, and the film drainage rate increased. However, above a certain concentration, thedrainage retarding effect of surface viscosity overcame the gradient effect.

The dependence of film drainage rate on silicone surfactant molecular structure was alsosystematically investigated. In order to understand this correlation, three physicalparameters of the film affected by surfactant structure must be considered. Theseparameters are the surface partition coefficient, the surfactant molecular diffusioncoefficient and the degree of intermolecular cohesion within the surface layer. Specifically,as the length of the polyether (solvophilic) portion of the surfactant increased, the surfacepartition coefficient decreased, the diffusion coefficient decreased, and the degree ofcohesion increased. This resulted, at constant surfactant concentration, in a complexeffect on the film drainage rate.

A quantitative physical model of a draining vertical thin film was developed from firstprinciples. The starting point was the Navier-Stokes equation. The initial model featureda fixed-surface, wedge-shaped vertical film, with immobile surfaces. This is essentiallythe Reynold’s model modified to this film shape and orientation. Good agreement of thepredictions of this model with experimental data was obtained.

The next model relaxed the condition of fixed film shape (allowed for curvature in thefilm surface) and analysed the effects of the menisci on the film drainage. Analysis of theNavier-Stokes equation was simplified by the application of the lubrication approximation.The results from this analysis agreed extremely well with experimental values both interms of film drainage rate and the changes in film shape with time. The developmentand growth of bulges and waves on the bottom of the film were particularly intriguing.This phenomenon was experimentally observed in these films.

Finally, models were developed where the condition of infinite surface viscosity was relaxed.This allowed the analysis of surfactant effects on film drainage, in particular surface viscosityand surface transport. Specifically, the model predicted the decrease in drainage rate as surfaceviscosity increases, as expected from the qualitative models and measured experimentally.The effect of surface transport was significantly less than that of surface viscosity.

221


5.3.1 Experimental Details

5.3.1.1 Film Formulation

Unless noted otherwise, the experimental work in this chapter involved a model flexibleslabstock PU foam formulation [39] at the instant of mixing (see below):

• 20.00 g VORANOL 3137 (3100 MW polyether copolymer, The Dow ChemicalCompany),

• 5.46 g toluene,

• 0.20 g DABCO DC 198 (surfactant Air Products and Chemicals, Inc.)

For simplicity’s sake, chemical reactions were avoided. This was accomplished bysubstituting toluene for toluene diisocyanate (TDI) in the formulation. This formulationhad a nominal viscosity of about 0.8 Pa-s.

5.3.1.2 Description of the Experimental Interferometer

The following experimental variables were addressed during the design and constructionof the interferometer:

• control of film area,

• control of the withdrawal rate of the film,

• control of spurious vibration potentially leading to film rupture,

• control of solvent evaporation,

• precise vertical alignment of the film,

• optical control of the interferometric measurement (illumination with plane parallellight, use of monochromatic light, uniform illumination of the film surface),

• Imaging of the film including magnification, presentation on a video screen, andvideo recording capabilities, and

• Precise measurement of film lifetimes.

222


Consideration of these points led to the experimental setup sketched in Figure 5.1.

The film is formed between two narrow blades set by 10 mm apart and supported in asurrounding frame; a film of 20 mm height could be formed between the blades. Theframe was vertically clamped on an optical bench. A glass cuvette with rectangularcross section was fabricated with a closely fitting lid through which the clamp rod fitstightly in order to prevent evaporation of the solvent and at the same time to allow thecuvette to be lowered and raised to adjust the bulk liquid level or to draw a film. Thecuvette stands on a small pedestal which was driven by a computer-controlled steppermotor assembly allowing film size control at velocities of the receding or advancingbulk liquid level up to 25 mm/s. The sample holder assembly was mounted on anoptical bench.

The two nearly plane-parallel surfaces of the film constitute a Fizeau interferometerwhich produces interference fringes in the reflected light [40]. The fringes are linesof constant film thickness. Film drainage data were acquired by generatingmonochromatic light by means of an interference filter (λ = 505 nm) or with a HeNelaser (λ = 632.8 nm).

Figure 5.1 Schematic of the thin liquid film interferometer.

223


5.3.1.3 Film Formation

The process of film formation is depicted in Figure 5.2. Films were formed by loweringthe sample liquid from a point where the vertical frame was completely immersed to apoint where it was only partially immersed. In most experiments, the vertical film soformed was left in direct contact with the bulk liquid. Films were withdrawn from thebulk liquid at a rate varying from 0.25 to 50 mm/s.

5.3.2 Qualitative Description of Polyurethane Films

In a typical experiment, stable, vertical, liquid films were formed from the model PUformulation (see Section 5.3.1.1). These films were stable for two to five minutes. Theseries of photographs in Figure 5.3 depicts many of the physical features of these films.Thirty to sixty seconds after film formation, dark horizontal interference fringes wereobserved that initially appeared at the top of the film and steadily moved downward.Over time, the number of fringes decreased. Flow patterns, including eddy currents,fingering patterns and swirls, appeared at the bottom and sides of the films and roseupwards. The rate and amount of these flows decreased as the lifetime of the film increased.

Films were subjected to a rapid raising and lowering of the bulk liquid level. Thesemotions strongly stimulated the surface flows previously mentioned.

Figure 5.2 Depiction of film formation. The film is formed within the boundaries ofthe inner frame by the lowering of the liquid level at a controlled rate. The frame and

liquid are within the transparent walls of the enclosed glass cuvette.

224


The pattern of horizontal fringes observed on the face of the films suggested a verticalwedge shape, with the film thickness increasing as one descends the film. The wedgeshape of the vertical PU films has a number of implications for operational PU foam.Any cell window in PU foam does not have uniform thickness. This non-uniformity hasbeen reported in at least two studies [24-26, 32]. In addition, it would be expected thatthe rupture of the cell window would occur at its thinnest point [24-26, 32].

As seen in Figure 5.3, the interference fringes progress down the face of the film and thedistance between the fringes increases as the film continues to drain. This indicates adecrease in the film thickness gradient during the drainage of the film. The drainage ofthe wedge-shaped film can be visualised in that the sides of the wedge, intersecting at the

Figure 5.3 The visual sequence of a draining film. The first photograph in thesequence is in the top left hand corner and the sequence of photographs proceedsclockwise. Photograph 2: one minute after formation. The dark horizontal linesfaintly present are the interference fringes. Photograph 4: 3-5 minutes after film

formation. Fewer fringes are present and span the height of the film. The reductionin the number of fringes versus Photograph 2 is consistent with a significant amountof film drainage. Photograph 3: 8-12 minutes after film formation. Photograph 1:

approximately 15 minutes after film formation.

225


top of the film, pivot inwards as fluid drains out of the bottom. This is a collapsinginwards of the wedge.

As the film is draining, interference fringes radiate outwards from the top centre of thefilm (photographs 2-4 in Figure 5.3). This radiation suggests that both horizontal andvertical flow processes are present. These flow processes are driven by both gravity (verticalflow) and suction into the Plateau borders of the film (horizontal flow). Once the fluidhas entered the Plateau border, it flows down a relatively wide channel into the bulkliquid. The Plateau borders in these films can be seen in Figure 5.3. They are the darkregions between the film and the frame. The width of the Plateau channel increases asone descends the face of the film, giving it roughly a triangular shape. Plateau bordersare also present in operational polyurethane foams. In fact, horizontally-aligned filmswithin these foams can only drain due to Plateau border suction. The individual Plateauborders within the PU foam form a network which allows for liquid drainage down thefoam. Ultimately, this network becomes the structural ‘struts’ of the foam.

As reported in the literature, interference fringes in many vertical aqueous films arehorizontal with little curvature. These films also displayed ‘mobile’ surfaces, rapid andextensive surface flows and rapid rates of drainage [22, 28-30]. These features were alllinked to the presence of a low surface viscosity. Based on our experimental observations,the PU films stabilised by DABCO DC 198 have low surface viscosities. This conclusionis supported by reports of measurements of low surface viscosities in polyol solutionscontaining silicone polyether surfactants [12, 15, 38].

The phenomena of edge turbulence, fingering and upward flows in aqueous soap filmshave been extensively investigated [22, 28-30, 41-46]. These hydrodynamic phenomenahave been defined as marginal regeneration and gravity convection. Marginal regenerationrefers to a process where, simultaneously, thick films are sucked into the Plateau borderand thinner films are pulled out. This exchange results in a net increase of material inthe Plateau border, essentially draining the film. Within the film, after the materialexchange, the new thin spots then migrate upward (gravity convection) until they reacha height where their thickness equals the film thickness.

Instead of describing these phenomena in PU films as ‘gravity convection ‘, it is proposedthat this flow near the surface of the film is driven by surface tension gradients. Theseare examples of Marangoni flows, and would proceed from an area of high surfacetension to one of low surface tension. Marangoni flows are extremely common and onewell-known example is the phenomenon of ‘tears’ in a glass of port wine or brandy. Thisview is also gaining some acceptance in the field of aqueous thin liquid films. Stein [42-44] recently proposed that the flows observed in mobile-surfaced, aqueous soap filmsare Marangoni flows.

226


The surface tension gradients necessary to drive this type of flow can arise from threedifferent physical phenomena: surface tension gradients are necessary to support theweight of the vertical film, flows in the film or the channel of the Plateau border imposestresses on the surface monolayer which are balanced by a surface tension gradient, andextension or compression of the surface layer can create transient surface tension gradients.

Regarding the cell windows of operational PU foam, it is apparent that surface tensiongradients are constantly present, due to all three mechanisms discussed previously. Inparticular, the constant stretching of the bubbles during the process of foam growth actsas a potent stimuli for these flows.

5.3.3 Quantitative Measurement of Film Drainage Rates: Bulkand Surface Effects

5.3.3.1 A Method to Measure Film Drainage Rates: The ‘CollapsingWedge’ Model

The distribution of the interference fringes in a liquid film can be viewed as a contourmap of the film thickness, with each successive fringe representing a section of constantfilm thickness (isopach). Therefore, the distance between fringes throughout the filmgives the gradient of film thickness. In order to precisely measure PU film drainage rates,a physical model correlating the time change in the fringe density (ds/dt; the quantitythat can be measured in the laboratory) to the time change in volume (dV/dt, the filmdrainage rate) of the film must be applied.

Qualitatively it appears that the drainage of the film collapsed the sides of the film wedgeinwards. During this collapsing process, the two sides of the wedge act as if they werehinged together at the top of the film. With this geometric model of film shape, thedrainage rate of the film, dV/dt, can be expressed mathematically by Equation 3.

dd

dd

Vt

cWL

st

=2

2 (3)

Where V is the film volume, t is time, c is the change in film thickness per interferencefringe, W and L are the film width and height, respectively, and s is the fringe density.

If both sides of Equation 3 are divided by the cross-sectional area (WcsL) of the film atthe bottom Equation 4 is obtained:

227

(1/WScL) dV/dt = L/2s ds/dt (4)

The quantity on the left of Equation 4 is the flux. If it is assumed that the flux is ahyperbolic function of time, namely,

(1/WScL) dV/dt = b/t + t0 (5)

where b is a drainage constant and t0 is the time where hyperbolic drainage begins.

Solving the differential equation for s yields Equation 6.

s k t t mb

L

m= +( ) = −0 0

2 with (6)

where k0 is the fringe density at t = t0

As Equation 6 shows, a log-log plot of fringe density versus time should be linear for t >t0 with a slope of m (<0) and an intercept of logs0 = logk0 + mlogt0. This prediction wasrealised experimentally with high degrees of precision and accuracy.

Therefore, the instantaneous film drainage rate is explicitly given by Equation 7.

dV dtc

/ = ⋅ = −− −

22

01

01WL mk t cWLbk tm m (7)

5.3.3.2 The Effect of Bulk Dynamic Viscosity on Film Drainage Rate

The drainage rate measurement described in Section 5.3.3.1 was novel and requiredexperimental verification of its accuracy and precision. This was accomplished bymeasurement of the dependence of film drainage rate on the reciprocal of bulk viscosity(see Figure 5.4). The Reynold’s equation (Equation 1) predicts that this correlation shouldbe linear. As seen in Figure 5.4, a linear correlation was obtained. It should be noted thatthis form of the Reynold’s equation assumed an infinite surface viscosity. This conditionwas not able to be achieved in practice. However, it was assumed that the surface viscositywas constant throughout the range of experiments and that Equation 1 was valid underthis condition.

The excellent correlation achieved between film drainage rate and reciprocal bulk viscositymotivated the measurement of the correlations between drainage rate and surface variables.


228


5.3.3.3 The Effect of Surface Variables on Film Drainage Rate

5.3.3.3.1 Analysis of Commercial Silicone Surfactants

Initially, the drainage rates of PU films stabilised by various commercial silicone surfactantswere measured. These surfactants are applied in the marketplace to stabilise flexibleslabstock PU foam. These data are given in Table 5.1.

The surfactants listed in Table 5.1 span a range of drainage rates of 1.5 to 2. Thenarrowness of this range is not surprising because the PU foam stability (in terms offoam volume and porosity) achievable with these commercial surfactants does not varyto a great extent. By comparison, as described in Section 5.3.3.3.2, rigid-surfaced filmshave been observed which have drainage rates of less than 1% of those listed in Table5.1. Within this context, the drainage rates in Table 5.1 are rather fast, consistent withthe previously discussed low surface viscosity of the films. This is also consistent with

Figure 5.4 The experimental dependence of film drainage rate on bulk viscosity. Thebulk viscosity was varied by varying the ratio of toluene to polyol in the formulation

229

the consideration that these surfactants have been optimised to provide sufficient drainagerate to produce porous, flexible slabstock PU foam.

Incidentally, one surfactant investigated (FC 171; 3M Company) gave a significantlyhigher drainage rate (271 x 10-3 mm/min) than the surfactants shown in Table 5.1. Thissurfactant, a non-ionic fluorocarbon-based material, is not sold as a stabiliser for PUfoam. Although foam tests were not attempted with this stabiliser, one must presume,based on its high drainage rate, that these foams would be quite unstable.

5.3.3.3.2 The Effect of Surfactant Concentration on Film Drainage Rate

Physical models for the potential effect of surfactant concentration on film drainage ratehave been discussed in detail in Section 5.2. It was predicted that surface tension gradientsand surface viscosity affect film drainage rate. It was expected that the drainage rate wouldincrease with an increase in surfactant concentration due to the lessening of the intensity ofsurface tension gradients. However, It was also expected that the drainage rate woulddecrease as the surfactant concentration increased due to the increase of surface viscosity.

This correlation was examined for two different silicone surfactants as illustrated in Figure5.5. These plots show a maximum in film drainage rate as a function of surfactantconcentration, consistent with the concept that the overall rate is the result of two opposing

UPkcotsbalselbixelflaicremmocfonosirapmocA1.5elbaT)tnatcafrushpp1(smlifdiuqilnihtUPnistnatcafrusmaof

tnatcafruS 01xtd/Vd 4 mm( 3 )nim/rorrE( ± )%01

891CD a 561

0595CD a 091

9305CD a 071

026L b 371

0732FB c 701

ocbaD:CDa .cnI,slacimehCdnastcudorPriAb seitlaicepSoctiWc GAtdimhcsdloG.hT


230


effects. At lower concentrations, the increase in drainage rate observed with increase insurfactant concentration is consistent with a decrease in the intensity of surface tensiongradients in the film due to increased bulk-to-surface surfactant flux. The decrease in drainagerate seen at higher surfactant concentrations is consistent with an increase of surface viscosity.

These conclusions are supported by several published experimental studies.

In their studies of three-phase capillary air slug (bubble) flow, Stebe and Maldarelli [47]correlated surfactant concentration to equivalent flow rate invoking similar reasoning.At a higher surfactant concentration, flow rate increased with surfactant concentration,due to the relief of surface tension gradients at the air bubble surface. When usingprotein-based surfactants, retardation of flow due to surface viscous effects (flow ratedecreased with an increase in protein concentration) was observed.

Figure 5.5 The effect of surfactant concentration on film drainage rate: Comparison ofthe two surfactants DC 198 (solid line) and DC 5950 (dotted line) 10 mm high films

were drawn from the standard PU film formulation at a rate of 0.25 mm/s

231

Following the work of Kanner and co-workers [37], the drainage of films stabilised by aTCP was examined. As previously mentioned, they observed a significantly low filmdrainage rate and attributed it to the high surface viscosity measured for these solutions.

In studies of high TCP concentrations, rigid-surfaced films with very slow drainage (asdetected by the movement of interference fringes) and surface flows were observed. Atlow TCP concentrations, the films were mobile-surfaced and fast draining. Theseobservations are consistent with the concept that the surface viscosity increases withsurface concentration. Further reinforcement of this concept is given by the data inTable 5.2. Assuming that surface viscosity is proportional to surfactant concentration,the inverse dependence of film drainage rate on surface viscosity is quite apparentfrom this data.

smlifnihtUPfoeganiardmliffoetarehT2.5elbaTPCTfonoitartnecnocehtfonoitcnufasa

]PCT[)mpp(

td/Vd01x 4 mm( 3 )nim/

01 596

06 113

021 33

033 7

noillimrepstrap:mpp

Previously, in the literature, the ‘extremes’ of mobile and rigid surfaces were muchdiscussed without significant mention of ‘intermediate’ behaviour. Intermediate behaviourwas studied here by simple titration of the appropriate amount of TCP into the filmsolutions. In this regime, the film drainage rates were intermediate between those ofmobile and rigid-surfaced films, as were the rate and extent of surface flows.

For both Dabco DC 198 and Dabco DC 5950 (Air Products and Chemicals, Inc.), thesurfactant concentration (approximately 1 pph) leading to maximal drainage rate, issimilar to that optimally used in manufactured PU foam. Presuming that the surfactantconcentration needed for effective film stabilisation is similar to that needed for effectivefoam stabilisation, it appears that a maximal drainage rate aids in the optimal productionof flexible slabstock polyurethane foam.


232


5.3.3.3.3 The Effect of Surfactant Molecular Structure on Film Drainage Rate

Once the effect of surfactant concentration on film drainage rate had been measured, theeffect of molecular structure was investigated. It was hypothesised that the surfactantstructure would affect the intensity of surface tension gradients within the film and thesurface viscosity. Within this framework, the effect of molecular structure on the diffusionand surface partition coefficients, and on the degree of intermolecular cohesion in thesurface layer was investigated.

For TCP mentioned in the previous section, an accurate structural characterisation ofthe material was not available. However, it was known that it was a complex three-dimensional siloxane structure, capped by Me3SiO- or -OH units.

By comparing the films prepared from TCP solutions to those from solutions of highlypurified (Me3SiO)8Si8O12, it was deduced that the source of the surface rigidity was theOH (silanol) groups on TCP. These groups can hydrogen bond to each other, thus forminga highly cohesive network at the surface and yielding a high surface viscosity. A similareffect was also noted in aqueous films containing carbohydrate functional organicsurfactants [48]. The source of OH groups in these surfactants are the carbohydratefunctionalities.

These results motivated a more careful look at the specific effects of surfactant structureon film drainage rate using traditional silicone PU foam stabilisers. In order to accomplishthis, the drainage rates of PU films stabilised by a series of surfactants of the generalmolecular structure (Me3SiO)(SiMe20)100(SiMeRO)10SiMe3 (R= -(CH2)3(EO)x

(PO)xOC(O)Me, EO = ethylene oxide, PO= propylene oxide, X= 6, 12, 18, 21, 30) wereinvestigated. These five surfactants differed only in the length of the alkylene oxidechains. Therefore, for the rest of this chapter, they will be referred to only by their Xvalue. Relevant measured physical properties are listed in Table 5.3. Film drainage ratedata is shown in Figure 5.6. Unfortunately, it was not possible to obtain accurate surfaceviscosity data for these materials.

Analysis of the data in Table 5.3 and Figure 5.6 begins with four hypotheses regardingthe physical consequences, to a silicone/polyether copolymer, of an increase in polyetherlength (X) at constant siloxane chain length. An increase in X leads to a copolymer with:

1) A higher ratio of solvophilicity (through the polyether portions) to solvophobicity(through the silicone portion) and therefore, a higher solubility in the liquid (polyether/toluene mixture) phase.

2) A larger solvated volume and therefore a lower diffusion (through the liquid phase)coefficient.

233

stnatcafrusenocilisrofatadnoitasiretcarahC3.5elbaT

X Γ∞ ×01x( 01 mc/lom 2)

AA( 2)

K da

01x 71 )mc(D

01x 11 mc( 2 )s/6 67.1 49 2.33 32721 62.1 231 3.22 47581 89.0 961 8.21 42412 MN MN MN 77303 40.1 061 0.12 66

× sniahcedixoenelyklafohtgnel=Γ∞ noitprosdaecafrusehtsanwonkosla,noitarutasecafrustanoitartnecnocecafrus=

ecafrusehttaelucelomrepaera=AK da tnatsnocnoititrapecafrus=

tneiciffeocnoisuffid=Dderusaemton=MN

Figure 5.6 Film drainage rate as a function of surfactant concentration and structure(X = length of alkylene oxide chain)


234


3) A larger area at the liquid/air interface.

4) A larger area of polyether chain cohesion in the interfacial layer. This increase incohesion area should result in an increase in cohesion energy per molecule andtherefore, at surface saturation, a higher surface viscosity.

Generally, the experimental data in Table 5.3 are consistent with these hypotheses.Firstly, the surface partition coefficient, Kad, is the ratio of the surface saturationconcentration (Γ∞) to the overall amount of surfactant needed to saturate the surface(keeping in mind that a significant amount of the surfactant remains solubilised).Therefore, at relatively constant Γ∞, Kad should decrease as the solubility of thecopolymer increases. This can be seen in Table 5.3. Secondly, the solvated volumeswere not directly measured; however, the diffusion coefficients were measured andthey decreased as X increased. It is generally known that the diffusion coefficient variesas the inverse of the solvated volume [49]. From this relationship, one would concludethat the solvated volumes of the copolymers increased with X. Thirdly, the interfacialareas of the copolymers generally increased as X increased. Finally, we were not ableto measure the surface viscosities of these materials in solution. However, all of themyielded mobile-surfaced films. This means that their surface viscosities were relativelylow, as expected.

It is unclear why the X = 30 copolymer did not give data consistent with the expectedtrends in the data observed for X = 6-18. It would have been helpful to have had moredata on the X = 21 species; perhaps, the deviations observed for the X = 30 specieswould have also been recognisable for the X = 21 data. For X = 30, one possible factorwhich could have yielded non-compliant data would have been an unexpectedconfiguration of the polyether groups. After all, they are quite long (total of 60 monomerunits). Perhaps they assumed an unusual coiled or spiral configuration both in bulksolution and at the interface. This would be consistent with an unexpected solvatedvolume, degree of solvation, interfacial area and area of polyether cohesion. Previouswork with silicone polyethers has demonstrated unexpected results with exceptionallylong polyethers [50].

One goal of this study was to correlate the data in Table 5.3 with the experimentaldrainage rate data for these surfactants, displayed in Figure 5.6. In order to interpretthese correlations, the following hypothesis were developed.

• Low values of X correspond to a high tendency of the surfactant to adsorb at thesurface (high Kad) and a large diffusion coefficient (D). This favours rapid relief ofsurface tension gradients. Furthermore, the surface viscosity values should beminimal under these conditions. Therefore, the films should drain rapidly and berelatively unstable.

235

• As X increases, Kad and D should decrease, which decreases the efficiency of relievingsurface tension gradients. Also, the surface viscosity should increase. Both factorsshould cause a relative decrease in film drainage rate and furthermore, an increase infilm stability.

The following conclusions can be made on analysing Figure 5.6.

The profile for X = 6 is rather complex. Ignoring the points at very low concentrations(which are difficult to interpret), the profile has only the ‘first half’ of the expectedparabolic shape. At moderately low concentrations, the drainage rate increases withconcentration, consistent with the decrease in the intensity of surface tension gradients.Incidentally, this decrease in gradient is also reflected in the fact that the film lifetimesdecrease in this range. In fact, above a certain concentration, films are too unstable toallow for the measurement of the drainage rate. This high degree of (Marangoni)instability, only observed for the X = 6 species, is consistent with an overly-efficient reliefof surface tension gradients in the film. This efficiency is due to the high bulk-to-surfaceflux of this surfactant, which itself is due to its high values of Kad and D. At high surfaceconcentrations, the films could potentially be stabilised by a high surface viscosity.However, for X = 6, the surface viscosity would be expected to be low, and this is borneout by the experimental data, i.e., the films are so unstable that drainage rate measurementscould not be made.

The drainage rate for X = 6 should be the highest of any of the surfactants. This is notconfirmed by experiment. It is possible that the films where this would be demonstrated(at higher surfactant concentrations) were simply too unstable to be investigated.

The expected parabolic shape of the drainage profile is observed for X = 12, and at arelatively low concentration, as might be expected. These films were more stable thanthose exhibited for X = 6, due to a less efficient relief of surface tension gradients. Thiswould be expected, as Kad and D should be lower than for X = 6. The decrease in drainagerate at higher concentrations is due to the build up of surface viscosity. A higher surfaceviscosity than for X = 6 would be expected, as the area of polyether cohesion has increased.

The drainage rate profile for X = 18 is at a lower value than either X = 6 or X = 12,consistent with stronger surface tension gradients and higher surface viscosity. The factthat film stabilities were higher than for either X = 6 or X = 12 is consistent with thisconclusion.

It is intriguing that, for X = 30, the ‘first half’ of the drainage profile is not observed atall. Perhaps it could exist at very low surfactant concentrations, but these films prematurelyrupture. A sufficiently high concentration of surfactant is necessary at the surface toallow for the presence of the surface tension gradients that stabilise films in this regime


236


(in other words, ‘high’ but not ‘too high’). These low gradients would result in a highdrainage rate and, a marginally stable film. As the concentration of this surfactant isincreased, the drainage rate is strongly retarded due to the build up of surface viscosity.Viscous effects would be expected to be particularly strong in this system due to the longpolyether chains, and the resulting large area of cohesion.

Overall, the limitations of analysing these data in terms of simple, qualitative modelsmotivated the development of more refined, quantitative models of the draining film.The development of these models is the topic for the next section.

5.4 The Development of Theoretical Models of Vertical, Draining ThinLiquid Model PU Films

In quite a few of the experiments, there was an extension region in the upper part of thefilm that was essentially two-dimensional. In particular, this occurred in the early partof the experiments (typically several minutes, at least) and in the upper half to two-thirdsof the film for nearly the entire width. Under these circumstances, the interference fringeswere essentially horizontal and parallel. These were also the conditions under which theexponent for the drainage rate was measured. This leads us to consider two-dimensionalmodels for film drainage; they can be considered as a cross-sectional slice down thecentre of the film perpendicular to the surface of the film. All of the theory described inthis section will be for this two-dimensional case.

To develop a quantitative model, the problem was approached in stages. The firstmodel assumed that the surface of the film was wedge-shaped; the narrowing withtime was predicted in the tangentially-immobile case [51]. The film was then modelledwith a deforming surface; the film still thins, but it takes on more complicated shapespredicted by the solutions of non-linear partial differential equations. Thesegeneralisations have resulted in a series of models that gradually incorporate more ofthe experimental behaviour. A few details and some results from these models willnow be discussed.

5.4.1 Rigid-Surfaced Collapsing Wedge Model

The simplest model was a theoretical version of the collapsing film wedge. As discussedpreviously (Section 5.3.3.1), the collapsing wedge geometry was imposed on the analysisof the experimental data to generate instantaneous film drainage rates. Therefore, itseemed reasonable to use this same shape in the development of the first theoreticalmodel [51]. This led to the condition that the shape of the film was invariant with

237

time. The other key assumptions applied in the construction of this model were thatall of the flow in the film was downward (parallel flow) and that the surface of the filmwas tangentially immobile (infinitely large surface viscosity). Therefore, outside ofsurface viscosity, there was no specific surfactant effects on the drainage rate of thefilm. In fact, no surfactant transport or concentration gradient effects were incorporatedin the model.

A diagram of the model construction is shown in Figure 5.7. The film shape is representedby the solid line; the wedge shape that replaces the deforming film shape is sketched withthe dashed line.

Figure 5.7 Schematic representation of the model film studied. The dashed-dotted lineis the centreline of the film; the film is assumed to be symmetric around it. The solidcurve is a sketch of the actual film surface; the dashed line is a wedge approximation

to the film shape. Differences are exaggerated to clarity.


238


Within the wedge, the momentum in the vertical direction is conserved by that componentof the Navier-Stokes equations:

ρ ∂∂

ρ ∂∂

μ ∂∂

wt

gpz

wc

= − +2

2(8)

To simplify matters, it is assumed that the flow is so slow and gradually changing thatthe left hand side of the equation can be neglected. The flat sides of the wedge alongwith the slow flow assumption allow the pressure gradient term to be neglected. Thedifferential equation for the velocity profile in the film is then obtained:

∂∂

ρμ

2

20

wx

g+ = (9)

The solution to this equation is subject to ∂w/∂x = 0 in the centre of the film and ω = 0 onthe surface of the film, x = k (z, t). The velocity profile is given by

w x tg

k x,( ) = −( )ρμ2

2 2(10)

and the average velocity of the film is then found to be

wg k= ρμ3 4

2

(11)

This equation clearly indicates that drainage of the film is accelerated by increasingdensity and inhibited by increasing viscosity.

Further analysis of this model yields a prediction that the slope of the plot of log (fringedensity) against log (time) should be -0.5. These predicted results are in excellentagreement with experimental results where a rigid-surfaced film was investigated;experimental data gave a slope of -0.47. For the mobile-surfaced films investigated, thevalue of the shape ranged from -0.6 to -0.92; the more negative the value, the faster thefilm drainage.

At this point, theory could be developed only for the rigid case; the development oftheory for the mobile cases requires more complicated mathematical models, which aredescribed next.

239

5.4.2 Deforming Film Models

A schematic of the draining film useful for setting up the mathematical theory is given inFigure 5.8. The free surface of the film, k(z, t), can now deform along its length withtime, and partial differential equations for this and other dependent variables of interestwill be developed.

The mathematical problems in this section are specified in the following way, unless statedotherwise. The Navier-Stokes equations govern the motion of the incompressible,Newtonian fluid inside the film [52]. The film is assumed to be symmetric about its verticalcentreline (the z-axis). On the free surface of the fluid, several equations must be satisfied:the kinematic condition, and the normal and tangential stress conditions [53, 54]. Anequation governing the transport of surfactant in the free surface by both advection (fluid

Figure 5.8 Schematic representation of the model studied


240


motion) and diffusion also holds [55, 56]. The surface tension, σ (in mks units: N/m), ofthe film-air boundary is assumed to vary linearly with the surface concentration, Γ:

σ σ ∂σ∂

= + −( )m m mΓΓ Γ Γ (12)

The subscript m denotes the reference values of σ and Γ; these reference values are chosenfor convenience in matching up with experiments. The additional parameters that showup in the boundary conditions on the free surface are the surface shear viscosity andsurface dilatational viscosity, represented by μs and κs, respectively, and the surfacediffusivity of the surfactant denoted by Ds.

No fluid can enter or exit through the wire frame at the top, and the film surface ispinned there. At the bottom, the film drains into the bath (or pool) in the cuvette, fromwhich the film is drawn. The bath is otherwise inert, and it is assumed that the filmshape tends to a static meniscus shape over the bath. This assumption will allow boundaryconditions over the bath to be specified. The details of the mathematical specification ofthe problem appear elsewhere [57-60]; in this section, the focus is on simplified modelsand comparing results from them with the experimental results discussed previously.

The equations mentioned above will be made dimensionless with the length scales

dVg

Dg

d D= = =μρ

σρ

0 1 3 2 3, / /and l (13)

d, D and l are the film thickness, the static meniscus radius, and an intermediate scale,respectively.

The horizontal length scale is d, while the length scale along the film is l . Of particularimportance is that, for the PU solvent of interest here, the ratio of the lengths d/ l is asmall quantity. The ratio

δ2 1≡ <<d / l (14)

is used to reduce the original mathematical fluid dynamics problem to a simpler set ofpartial differential equations; the mathematical method is a multiple scale analysis, andwhen it is applied to thin fluid layers, it is called lubrication theory. For recent reviews ofthis type of approach, see references [29, 61]. We choose

Vgd

0

2

= ρμ (15)

241

as the velocity scale; in mks units, V0 is in m/s. The time scale is l /V0 and the pressure

scale is based on viscous shear, μ δV04/ ( ).l The surface concentration is made

dimensionless with the reference value, Γm.

Several dimensionless groups appear in the non-dimensional equations; these are givenin Table 5.4. These non-dimensional groups serve to highlight the relative importance ofthe myriad parameters in the problem; the interpretation of these groups in terms ofratios of physical forces are also given in Table 5.4. These parameters will be referred toas appropriate in the following sections.

For the typical experiment of interest, the Reynolds number is very small; this will helpmake the simplifications that follow valid. The Péclet number (P) is large; it may be aslarge as 104 at the start of some experiments, according to our estimates. In all thecomputations presented here, P = 102 will be used. The typical value of δ is 0.1 to 0.2;the capillary number is then smaller than 10-4. The other quantities will be varied in theresults to follow.

ehT.snoitaterpretniriehtdnasretemaraplanoisnemid-nonfoelbaT4.5elbaTehtnidetneserpersecroflacisyhpehtfonoitaterpretninasevignmulocdrihtrotanimonedehtnidetneserpersecroflacisyhpehtotderapmocrotaremun

=CsgnilacsruonitahtetoN( σ6)

rebmuN noitinifeD secroffonosirapmoC

sdlonyeR

deifidoMqsenissuoB

yrallipaC

inognaraM

telcéP

ℜ = ρμV0l Inertial

Viscous shear

SV

V d= +( ) /

/κ μ

μ0

2

0

l Surface viscous shearViscous shear

CV

m

= μσ

0 Viscous shearSurface tension

MV

m m

=Γ

ΓΓ∂σ

∂δ

μ

2

0

Shear from concentration gradientsViscous shear

p vVDs

= 0l Advective transportDiffusive transport


242


5.4.3 Tangentially-Immobile Films

For high concentrations of some TCP in PU, the surfactant develops structure on thefilm surface and the surface of the film still deforms but doesn’t move tangentially. This‘rigid’ case [22] is the slow draining limit of the experiment. In that case, the tangentialstress condition may be replaced by the condition that the tangential velocity componentis zero. We have been able to show mathematically that this case is achieved in the limitof large surface viscosity [59, 60] or S → ∞; this will be discussed further in the nextsection.

When lubrication theory is used in this context, a single, nonlinear partial differentialequation for the film shape k(z,t) is obtained. In one approach, which is called the‘whole film’ model, the equation obtained is:

kk

z

t z+ +( )⎡⎣⎢

⎤⎦⎥

=3

31 0κ (16)

which holds over the domain 0 < z < L. The subscripts t and z denote partialdifferentiation; for example κz = δk/δz. Here

κδ

=+( )

k

k

zz

z1 4 2 3 2/ (17)

is the curvature of the film. At the top of the film, z = 0, the film is pinned (k(0,t) = 1)and no fluid leaves or enters, by requiring that

κz = −1 (18)

At the bottom of the film, we specify the slope, kz(L,t) = C, and the second derivative isspecified via

k L,tC

CC szz ( ) = −

+

⎛

⎝⎜⎞

⎠⎟+( ) ≡2

1

11

2 4 2

4 2 3 2

1δ δδ

/

(19)

This approach is based on methods for including the static meniscus in film models thathave been developed by Ruschak [62] and Kheshgi and co-workers [63]; equation (12) isa first integral of the equation for the static meniscus that is the surface of the bath.

243

The initial condition used for all calculations was a linear shape for k. This was aconvenient shape, and for later times in the computations, the results are ratherinsensitive to the initial conditions. This problem is readily solved using numericalmethods described in references [57].

Numerical results are given in Figure 5.9; one can see in the figure that distinct regions ofthe film developed: (i) a meniscus at the top of the film (at the wire frame); (ii) a long flatmiddle region where effects from the average surface tension were negligible; (iii) a regionwhere bumps and dips developed near the bottom of the film and (iv) a static meniscus atthe bottom of the film where it joined the bath.

Figure 5.9 Film shapes k(z,t) versus z with L = 37.5, h = 0.05, kz(L) = 5 and δ = 0.2.The initial slope of the film was 0.23


244


A number of interesting possibilities arise. The bumps and dips at the bottom of thisfilm could be the precursor to complicated three-dimensional behaviour observed at thebottom of films in experiments. Only further work will verify this conjecture, and we arecurrently pursuing that line of inquiry.

The long middle region of the film offers the possibility of still further simplification. Ifone rescales the variables in the problem again, a much simpler equation for the (rescaled)free surface shape K(Z,T) arises, which we shall call the ‘flat film’ model. This model is

Figure 5.10 Computed thickness for models without and with surface tension, solid anddashed lines, respectively, at t = 15 on the right and t = 231.9 on the left;

here δ = 0.25. The case with surface tension is computed with kz(L,t) = 0 and neglectingall terms with d in equations (17) and (using C = 0 in Equation (19)); see [57] for details

245

K K KT z+ =2 0 (20)

with the single boundary condition specifying K(0,T). Here Z =z/δ and T =t/δ are therescaled space and time variables, respectively. Solutions found numerically for the filmshapes from Equation (20) and a more complicated model for the film shape [57] aregiven in Figure 5.10. As time increases, the agreement between the approximationsimproves. In the middle part of the film, no matter which simplification is used, thenumerical results indicate that the film is flattening out with t-1/2 behaviour at long times.

This simple equation also has an exact solution, which has similarity behaviour for longtimes given by K = Z T/ [57, 61]. Similar solutions have been found by Mysels andco-workers [22] and Moriarty and co-workers [64]. In this way, it can be foundanalytically that the rate of flattening of the film is proportional to t-1/2 for long times, ingood agreement with the slowest-draining experimental case of t-0.47.

Based on that success, this kind of simplification will be attempted with more complicatedmodels to see if similar results can be achieved.

5.4.4 Finite Surface Viscosity

When there is finite surface viscosity, the surface of the film is mobile. Lubrication theoryfor our vertical draining film gives a system of two partial differential equations describingthe evolution of the film surface k(z,t) and the velocity of the fluid at the surface w(S)(z,t).These equations are

k kwks

t z

z

+ + +( )⎡

⎣⎢

⎤

⎦⎥ =( )

3

31 0κ (21)

wk

NSs

zz z( ) + +( ) =1 0κ (22)

Here

Nk

≡+

1

1 4 2δ z(23)

The boundary conditions at the top are

k t t w ts( , ) , ( , ) ( , )( )0 1 0 1 0 0= = − =κz and (24)


246


and at the bottom are

k L,t C k L,t s w L,tsz zz and( ) , ( ) ( )( )= = =1 0 (25)

Here C is the specified slope and a consistent s1 can be calculated with the first integralof the equilibrium meniscus shape [62, 63], as in the rigid case discussed above. InEquations (21) and (24), κ is the curvature of the film given in Equation (27).

Previous work [62, 63] has shown that it is possible to keep the full curvature inthe normal and tangential stress conditions and integrate the partial differentialequation for the free surface k through the matching region (where the film meetsthe bath) onto the static meniscus; this model relies on those results. While we donot rigorously apply matched asymptotic expansions in this work, the equationscontain all of the terms necessary to match the film onto a static meniscus (for thebath surface) and it has been shown that the terms neglected are uniformly smallfor δ<<1 [65].

By introducing surface velocity, w(S) (z,t), as a dependent variable in the model, weavoid computing a fifth order derivative of k. This formulation greatly simplifiesour task; a complete description of the above derivation can be found in references[58, 60, 65]. During this work such a formulation was applied in a similar situation[66, 67].

Some computed results are shown in Figures 5.11 and 5.12. For large S, the filmdrains slowly; the downward parabolic shape in the middle region (Figure 5.11; freesurface shapes for varying S) agrees with the typical rigid film profiles given in ChapterIII of Mysels and co-workers [22]. In the limit of S → ∞, w(S) approaches zero alongthe whole film and the free surface shapes tend to that of the tangentially-immobilecase [57]. It is also found that for very large S, the film thickness in the middleregion tends to decrease with the power law t-0.5 as before; this is very close to theexperimental rate of t-0.47 obtained at Dow Corning [68] for their most rigid film andagrees with results from Mysels and co-workers [22]. As S decreases, drainage ismuch faster and the film thins more rapidly (Figure 5.11); free surface shapes forsmaller S, take on the typical shapes for mobile films given in Chapter IV of Myselsand co-workers [22]. In the limit of small S, t-1 thinning in the middle of the film isfound which agrees well with the maximum experimental rate of t-0.92. This model isable to span the film drainage behaviour from rigid to mobile films.

If the simplification for the flat middle part of the film is pursued. In that case, thesimplified equations become

247

K KWK

Ts+ +

⎡

⎣⎢

⎤

⎦⎥ =( )

3

30

z

(26)

WKSzz

s( ) + = 0 (27)

Once again the capitalised variables are rescaled versions of those in Equations (25-29)[60]. These are subject to the boundary conditions:

Figure 5.11 Film shape at t = 16 for several values of S (shown in the upper right).The film is much thinner for small values of S because the free surface is mobile. For

large values of S, the free surface becomes tangentially immobile.


248


K W Zs= = =( ) , 0 0at (28)

and W Z LZs( ) , = =0 at (29)

The initial condition for K is

K(Z Z, ) .0 1 0 23= + (30)

and a consistent W(s)(Z,0) is found from its equation. The consistent value of W(S)(Z,0)is required for the numerical solution of the problem [60].

Figure 5.12 w(s) with t = 16 and kz(L,t) - 10 varying S shown in the upper right. Thesurface velocity decreases as S increases

249

By studying these equations for large and small S, the slow and fast draining limits of theexperiment have been obtained using analytical expressions [60]. For large S, solutionsof the following form are found:

Kf ZT

Wf ZT

s≈ ≈11 2

21 2

( ),

( )/

( )/

and (31)

where f1(Z) and f2(Z) satisfy ordinary differential equations. The solutions approximatecomputed solutions for the rigid film case very well. In this way the rigid case for largesurface viscosity within lubrication theory is analytically recovered. For small S, we canfind solutions of the form

Kf Z

TW

f ZT

s≈ ≈3 4( ),

( )( )and (32)

where f3(Z) and f4(Z) also satisfy ordinary differential equations. In this way, the fastdraining limit of the experiments, where the measured exponents for the fastest thinningcase are -0.92, is analytically recovered.

Surfactant transport will now be included in the model, in an effort to obtain intermediatethinning rates as observed experimentally.

5.4.5 Adding Surfactant Transport

Surfactant transport is also important in the vertical draining film; surfactant concentrationgradients may develop, which may in turn strongly affect the fluid flow via the Marangonieffect. When surfactant transport is considered, lubrication theory then gives threenonlinear partial differential equations for the free surface shape k(z,t), the surface velocityw(S)(z,t) and the surface concentration of surfactant Γ(z,t). The mathematical problemto be solved for these dependent variables is:

k kwks

2

3

13

0+ + +⎡

⎣⎢

⎤

⎦⎥ =( ) ( )κz

z

(33)

wk

NSM

N Ss

zz z z( ) ( )+ + + =1 0

2κ Γ (34)


250


Γ Γ ΓtsN w

PN+ −⎡

⎣⎢⎤⎦⎥

=2 210( )

z

z(35)

inside the film. Boundary conditions are, at the top:

k t w t t ts( , ) , ( , ) ( , ) , ( , )( )0 1 0 0 0 0 1= = = = −Γz zκ (36)

and at the bottom:

k L,t C k L,t s w L,t L,tsz zz zand( ) , ( ) ( ) ( )( )= = = =1 0Γ (37)

where L is the lower end of the computational domain; C is specified and a consistent s1

can be calculated with the first integral of the equilibrium meniscus shape [62, 63], asdiscussed previously.

Some normalisation factors in the surface transport equation for G that ordinarily wouldbecome unity in lubrication theory at leading order have also been retained. While wedo not rigorously apply matched asymptotic expansions in this work, the equationscontain all of the terms necessary to match the film onto a static meniscus (for the bathsurface) and it has been shown that the terms neglected are small for δ<<1 [65].

Marangoni effects have a substantial impact for smaller values of S, retarding surfacedrainage and enhancing film thickness; Figure 5.13 shows free surface shapes for severalM with S = P = 102 and δ = 0.1. For a small M, the Marangoni effect is too weak to retarddrainage and the film thins rapidly. As M increases, the Marangoni effect become moreprominent, retarding drainage and the film is considerably thicker at the same time inthe computation. As M → ∞, w(S) → 0 all along the film and the surface becomes rigid.A rigid fluid surface associated with a large Marangoni number was observed in a levellingproblem [69] and in a similar geometry to this work.

Interesting free surface shape features are shown in Figure 5.14, for M = -10, S = 102

with varying t. There are parts of the free surface shape that appear to be drainingrapidly, as if they are mobile, and other parts that drain slowly, as if they were rigid-surfaced. Concave-out profiles above an upward-propagating wave indicate a mobilesurface while an essentially downward-parabolic profile following the wave indicates arigid surface. These assertions are supported by the thinning rates of t-1 above the waveand t-1/2 below it. The wave that separates the two regions is driven by surface tensiongradients that are shown in Figure 5.15. The localised gradient of the surfactantconcentration on the surface is sufficient to drag fluid up the film; the gradient is preserved

251

because the Péclet number is large, and so it is not smeared out by diffusion. The associatedsurface velocity profiles are shown in Figure 5.16; the negative surface velocity clearlyshows that fluid is moving up the film on its surface, as a result of the Marangoni effect.

The wave seen in Figure 5.14 propagates up the entire length of the film. Similar behaviouris seen in the experiments of Snow and co-workers [70, 71] with PU films, where whatappears to be bumps propagate up the edges of the film. On that basis, it is believed thatthe wave observed may be related to those structures observed in the three-dimensionalfilms. Work is currently being done to verify this conjecture.

Figure 5.13 Film shape at t = 16 with S = 102 fixed for several values of M (shownin the upper right). The film is much thinner for small values of M. For large values

of M, Marangoni effects are stronger, drainage is retarded and the free surfacebecomes tangentially immobile.


252


If S < 102, it is found that the film drainage is typically too rapid to allow for much effectfrom surface tension gradients (and thus Marangoni stresses). For S > 104, we find thatthe films behave, in large part, as if they were rigid [70]. It is clear that the maximumimpact of the Marangoni effect occurs at an intermediate value of the surface viscosity.These results lend support to the assertion that the greatest foam stability occurs whenthe Marangoni and surface viscous effects are at intermediate value, as given in the firstsection of this chapter.

Figure 5.14 Film shape with S = 102 and M = -10 fixed for several values of time t(shown in the upper right). A wave driven by surface tension gradients propagates upthe film. Thinning proportional to t-1 and concave-out profiles above the propagating

wave indicate a mobile surface; below the wave, t1/2 thinning and the downwardparabolic shape indicate a rigid surface.

253

Figure 5.15 Surfactant distribution with S = 102 and M = -10 fixed for several valuesof time t (shown in the upper right). The surface tension gradients driving the wave up

the film are easily seen.


254


5.5 Summary

The stabilising behaviour of silicone surfactants during the formation of PU foam can beinvestigated in detail, both experimentally and theoretically by the use of a foam modelsystem, specifically vertical, thin liquid PU films. These films were investigated by acombination of direct observation and interferometric measurements. The formation andtemporal evolution of structures on the surface of the film, including fingering patterns, wasa result of the Marangoni effect. Experimentally, these films demonstrate a wedge shape andthe gravity-driven drainage of these films causes a collapsing inwards of the wedge.

The evolution of structures and the drainage of these films were functions of theconcentration and molecular structures of the surfactants. The two film physical

Figure 5.16 Surface velocities with S = 102 and M = -10 fixed for several values oftime t (shown in the upper right). Negative values indicate flow up the film, and flow

up the film is caused by the surfactant concentration gradients.

255

parameters correlated to film structures, film drainage and the properties of the surfactantare: (1) the strength and lifetime of surface tension gradients, (2) the surface viscosity.Two extremes of film behaviour were identified. Firstly, a rigid-surfaced film where therate of drainage and structural evolution were low. Secondly, a mobile-surfaced filmwhere the rate of drainage and structural evolution were high. Intermediate cases werealso detected and measured.

The theoretical and numerical model for vertical liquid film drainage that has beendeveloped reproduces a number of features described in these experiments. These featuresinclude film shapes and thinning rates.

Complex, time-dependent, three-dimensional structures were observed in the region wherethe film meets the bath (away from the edges of the film). It is also believed that thelocalised bumps and dips at the bottom of the film in these two-dimensional results, maybe the precursor of those observed motions in the experiments; this notion is currentlybeing investigated.

Future work will include adding intermolecular forces that act across the film andextending our work to three-dimensions in order to make closer comparison withexperiment. Furthermore we endeavor to obtain accurate independent measurements ofsurface viscosity and correlate them to both film properties and surfactant parameters.

Acknowledgements

RJB and SN gratefully acknowledge support from Dow Corning Corporation for this work.

RJB is also grateful for the support of the National Science Foundation via grants DMS-9623092 and DMS-9631287.

SAS and UCP would like to thank Ben Vesper, Mike Owen, Andy Goodwin, John Frey,Randy Hill, Dave Battice, Mike Stanga, Kathy Goudie, Joan Sudbury-Holtshlag, MikeReuter, and Gloria Lyon for many helpful discussions and their support.

References

1. D. Reed, Urethanes Technology, 1995, 11, 6, 22.

2. S. A. Snow and R. E. Stevens in Silicone Surfactants, Surfactant Sciences Series,#86, Ed., R. M. Hill, Marcel Dekker, New York, 1999, Chapter 5.


256


3. B. Kanner, W. G. Reid and I. H. Petersen, Industrial Engineering Chemistry,Product Research & Design, 1967, 6, 2, 88.

4. R. J. Boudreau, Modern Plastics, 1967, 44, 5, 133.

5. E. G. Schwarz in Applied Polymer Symposia #14: Silicone Technology, JohnWiley and Sons, New York, 1970, p.71.

6. B. Kanner, E. D. Goddard and R. D. Kulkarni, Presented at the SPI and FSKConference on Cellular and Non-cellular Polyurethanes, Strasbourg, France,1980, p.647.

7. B. Kanner, and T. G. Decker, Journal of Cellular Plastics 1969, 5, 1, 32.

8. T. E. Lipatova, S. H. G. Vengerovskaya, A. E. Feinerman and L. S. Sheinina,Journal of Polymer Science: Polymer Chemistry Edition, 1983, 21, 7, 2085.

9. T. C. Kendrick, B. M. Kingston, N. C. Lloyd and M. J. Owen, Journal of Colloidand Interface Science, 1967, 24, 2, 135.

10. T. C. Kendrick and M. J. Owen, Presented at the Chimica Physica AppliedPractical Ag. Surfactant C.R. Congress International Detergents, 1968, 2, 1, 571.

11. E. G. Dubyaga, A. B. Komarova and O. G. Tarakanov, Colloid Journal of theUSSR, 1986, 4, 881.

12. M. J. Owen and T. C. Kendrick, Journal of Colloid and Interface Science, 1968,27, 1, 46.

13. M. Dahm, Publication No.1462 of the National Academy of Science, NationalResearch Council, 1966, 52.

14. M. J. Owen, T. C. Kendrick, B. M. Kingston and N. C. Lloyd, Journal of Colloidand Interface Science, 1967, 24, 2, 141.

15. J. Ritter, J. E. Dubjaga and A. B. Komarowa, Plaste und Kautschuk 1979, 26, 11,624.

16. L. I. Kopusov and V. V. Zharkov, International Polymer Science and Technology,1981, 8, 3, T34.

17. G. Rossmy, G. Sanger and H. Seyffert, VDI-Berichte, 1972, 182, 173.

18. H. Hamann and J. Ritter, Plaste und Kautschuk, 1983, 30, 7, 364.

257




22. K. J. Mysels, K. Shinoda and S. P. Frankel in Soap Films-Studies of their Thinningand a Bibliography, Pergamon Press, New York, 1959.

23. D. F. Evans and H. Wennertröm, The Colloidal Domain, VCH Publishers, NewYork, 1994, p.55.

24. R. Neff, Reactive Processing of Flexible Polyurethane Foam, University ofMinnesota, USA, 1995, [Ph.D Thesis].

25. X. Zhang, Role of Silicone Surfactant in Polyurethane Foaming Process,University of Minnesota, USA, 1998, [Ph.D Thesis].

26. K. Yasunaga, R. A. Neff, X. D. Zhang and C. W. Macosko, Journal of CellularPlastics, 1996, 32, 5, 427.

27. S. A. Snow, W. N. Fenton and M. J. Owen, Journal of Cellular Plastics, 1990, 26,2, 172.

28. J. S. Clunie, J. F. Goodman and B. T Ingram, Surface and Colloid Science, 1971,3, 167.

29. I. B. Ivanov, in Thin Liquid Films, Surfactant Science Series No.29, MarcelDekker Inc., New York, 1988.

30. R. J. Pugh, Advances in Colloid and Interface Science, 1996, 64, 0, 67.

31. L. D. Artavia, A Model for Low Density Foams, University of Minnesota, USA,1991, [Ph.D Thesis].

32. K. Akabori, and K. Fujimoto, International Progress in Urethanes, 1980, 2, 41.

33. C. A. Miller and P. Neogi, Interfacial Phenomena, Surfactant Science Series,Volume 17 Marcel Dekker, Inc., New York, 1985, p.196 and p.243-262.


258


34. R. M. Herrington, and R. B. Turner, Advances in Urethane Science andTechnology, 1992, 11, 1.

35. A. J. De Vries, Recueil, 1958, 77, 383.

36. I. B. Ivanov, in Thin Liquid Films, Surfactant Science Series No.29, MarcelDekker Inc., New York, 1988, p.419.

37. B. Kanner, B. Prokai, C. S. Eschbach and G. J. Murphy, Journal of CellularPlastics, 1979, 15, 6, 315.

38. M. J. Owen, Industrial Engineering Chemistry, Product Research andDevelopment, 1980, 19, 67.

39. R. Herrington, K. Hock, Flexible Polyurethane Foams, Dow Plastics, Midland,MI, 1991, 9.

40. Photonics Dictionary, Volume 3, The Photonics Design & ApplicationsHandbook, in the chapter on ‘Interpretation of Interferograms’. LaurinPublishing Co., Oakbrook Terrace, IL, 1991.

41 J. B. M. Hudales and H. N. Stein in The Structure, Dynamics and EquilibriumProperties of Colloidal Systems, Eds., D. M. Bloor and E. Wyn-Jones, KluwerAcademic Publishers, The Netherlands, 1990, 549.

42. J. B. M. Hudales and H. N. Stein, Journal of Colloid and Interface Science, 1991,138, 2, 354.

43. H. N. Stein in Surfactants in Solution, Volume 11, Eds., K. L. Mittal and D. O.Shah, Plenum Press, New York, 1991, 481.

44. H. N. Stein, Advances in Colloid and Interface Science, 1991, 34, 175.

45. K. Koczo, J. K. Koczone and D. T. Wasan, Journal of Colloid and InterfaceScience, 1994, 166, 1, 225.

46. J. B. M. Hudales and H. N. Stein, Journal of Colloid and Interface Science, 1990,137, 2, 512.

47. K. J. Stebe and C. Maldarelli, Journal of Colloid and Interface Science, 1994,163, 1, 177.

48. S. A. Snow, B. M. Nugent and F. L. Baines, Unreported observations.

259

49. Principles of Colloid and Surface Chemistry, 2nd Edition, Ed., P. C. Heimenz,Marcel Dekker, Inc., New York, 1986, p.65 and p.83.

50. T. E. Gentle and S. A. Snow, Langmuir, 1995, 11, 2906.

51. U. C. Pernisz and S. A. Snow, The Drainage of Liquid Polyurethane Films, DowCorning Corporation Research Report, 1999, 1999-I000-46182, Midland, MI.

52. G. K. Bachelor, An Introduction to Fluid Dynamics, Cambridge University Press,Cambridge, 1987.

53. D. A. Edwards, H. Brenner and D. T. Wasan, Interfacial Transport Processes andRheology, Butterworth-Heinemann, Boston, 1991.

54. J. C. Slattery, Interfacial Transport Phenomena, Springer-Verlag, New York,1990.

55. H. A. Stone, Physics of Fluids A, 1990, 2, 1, 111.

56. H. Wong, D. Rumschitzki and C. Maldarelli, Physics of Fluids, 1996, 8, 11,3203.

57. R. J. Braun, S. A. Snow and U. C. Pernisz, Journal of Colloid and InterfaceScience, 1999, 219, 2, 225.

58. S. Naire, R. J. Braun and S. A. Snow, Dow Corning Corporation ResearchReport, 1999, 1999-I0000-47199.

59. S. Naire, R. J. Braun and S. A. Snow, Journal of Colloid and Interface Science,2000, 230, 1, 91 .

60. S. Naire, R. J. Braun and S. A. Snow, SIAM Journal on Applied Mathematics,2000, 61, 3, 889.

61. A. Oron, S. H. Davis and G. B. Bankoff, Reviews of Modern Physics, 1997, 69,931.

62. K. J. Ruschak, AIChE Journal, 1978, 24, 705.

63. H. S. Kheshgi, S. F. Kistler and L. E. Scriven, Chemical Engineering and Science,1992, 47, 683.

64. J. A. Moriarty, L.W. Schwartz and E.O. Tuck, Physics of Fluids A, 1991, 3, 5,733.


260


65. S. Naire, R. J. Braun and S. A. Snow, An Insoluble Surfactant Model for aVertical Draining Free Film II. Film Plus Meniscus in preparation, 2000.

66. L. W. Schwartz and R. V. Roy, Journal of Colloid and Interface Science, 1999,218, 309.

67. O. E. Jensen, Physics of Fluids, 1994, 6, 3, 1084.

68. S. A. Snow, B. M. Nugent and U.C. Pernisz, Dow Corning Corporation ResearchReport, 1996, 1996-I0000-42699.

69. L. W. Schwartz, R. A. Cairncross and D. E. Weidner, Physics of Fluids, 1996, 8,7, 1693.

70. S. A. Snow, U. C. Pernisz and R.E. Stevens, Dow Corning Corporation ResearchReport 1995, 1995-I0000-40859.

71. S. A. Snow, U. C. Pernisz and R. E. Stevens, Presented at the Polyurethanes Expo‘98, Dallas, TX, 1998, p.217.

261

6 Synthesis and Characterisation of AqueousHybrid Polyurethane-Urea-Acrylic/StyrenePolymer Dispersions

Janusz Kozakiewicz, Anita Koncka-Foland, Jan Skarzynskiand Izabella Legocka

6.1 Preface

Because of several limitations in using solvent-based systems worldwide, there is a growingdemand for new types of aqueous polymer dispersions for a wide application range (coatings,adhesives, leather or textile finishing compounds). Especially interesting seems to be combiningthe excellent application properties of vinyl/acrylic polymer and copolymer dispersions suchas adhesion, suitable viscosity, high solids content and surface wettability with excellentproperties of films made of polyurethane dispersions (DPUR) such as mechanical strength,elasticity, solvent resistance and abrasion resistance. Since blending of both types of dispersionsdid not result in truly positive effects, the solution of choice must be synthesis of combined(hybrid) dispersion systems containing both polymers. This chapter will be devoted to adescription of methods of synthesis and a discussion of the structure/properties relationshipsof such systems, especially when applied as coatings.

6.2 Introduction

6.2.1 General Considerations

There are strict ecological limitations relating to allowable volatile organic compounds(VOC) content in coatings introduced in many regions and countries, e.g., the CaliforniaRule of 1995 [1] or special provisions in the EU, and which force the manufacturers tosearch for new, more environment-friendly formulations for particular products whichusually requires replacing of the substantial binder (polymer or resin). Undoubtedly theproducts for which the development has been the most significant in the field of coatingsover the last few decades are aqueous dispersion-based and powder coatings. In theformer group, DPUR, which are presently being widely used as binders for coatingformulations have already taken a leading position [2-5] and compete successfully withacrylic dispersions despite the distinctly higher market prices. Both blends of DPUR andacrylic dispersions [6], and (recently) hybrid acrylic-urethane dispersions [7-12], obtainedby synthesis, and not simple blending are now available. The latter products are much

262


more interesting from both a scientific and practical point of view since the synthesis ofa hybrid system may lead to new polymer architectures which may result in considerableimprovement of properties. Especially attractive is the possibility of synthesisinginterpenetrating polymer network (IPN) systems based on polyurethanes and acrylicpolymers using this principle since it has been shown that in such systems synergisticeffects may occur [12], i.e., the properties of a hybrid may be better than those of a blendof its components. At the same time, hybrid dispersion particles usually have quite complexmorphology (in most cases a ‘core-shell’ type) which affects properties of dispersionsand films and coatings made of them [13].

If a comparison were made between hybrid dispersion systems taking into account thekind of dispersion particles they contain, it would appear that the best properties shouldgenerally be observed for systems obtained by synthesis for which the particles have auniform structure (they may be called ‘true hybrid’ particles) and the worst – for simpleblends of the two dispersions [14] (see Figure 6.1).

Obviously, both types of dispersion particle morphology presented in Figure 6.1 (‘core-shell’ and ‘true hybrid’) should be considered as idealised cases. In practice, a variety ofdifferent particle morphologies may be observed. Three of them are shown in Figure 6.2.

The most important factor that affects the morphology of hybrid dispersion particlesseems to be the method of preparation of such hybrids.

Generally, the methods of preparation of aqueous hybrid polymer dispersions may beclassified into four groups [17]:

1) blending of two dispersions

(a) addition of polymer A dispersion to polymer B dispersion

(b) addition of polymer B dispersion to polymer A dispersion

2) synthesis of polymer A in a dispersion of polymer B

3) synthesis of polymer B in a dispersion of polymer A

4) ‘in situ’ preparation of hybrid dispersions

(a) co-emulsifying of polymers A and B in water

(b) co-emulsifying the mixture of polymers A and B

(c) co-emulsifying the system consisting of polymers A and B:

• IPN composed of polymers A and B (or a precursor of such IPN)

• chemically linked polymers A and B (or a precursor of such system)

263

Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…

a) Blend of dispersions

⇓no synergistic effect

• poor stability• matte and hazy films and coatings

• always two Tgs• poor mechanical properties of films

b) ‘Core-shell’⇓

synergistic effect possible• good stability

• usually matte and hazy films andcoatings

• usually two Tgs• good mechanical properties of films

c) ‘True-hybrid’⇓

synergistic effect probable• excellent stability

• usually glossy and transparent filmsand coatings

• usually one Tg

• excellent mechanical properties of films

Figure 6.1 Comparison of hybrid dispersion systems obtained by (a) blending, (b) and(c) by synthesis, taking into account particle type.

264


In practice, methods (2) and (3) seem to be the most commonly used. Recently, yetanother method of preparation of hybrid polymer dispersions, called an ‘active blending’has been reported [18]. The components of such ‘active blends’ or ‘active mixtures’remain inactive when in the presence of water, and cross-react during film formation.

The other factors that may influence hybrid dispersion particle morphology will bediscussed in detail in Section 6.3.2 of this Chapter.

6.2.2 Acrylic Dispersions and Polyurethane Dispersions (DPUR)

Aqueous acrylic dispersions have been used as binders for coatings for over 60 years.Their main advantages include: low cost and excellent tolerance of additives or pigments;however, they have several drawbacks such as the presence of relatively large amounts ofwater-sensitive additives (protective colloids, emulsifiers, etc.) and the need to usecoalescing agents when a continuous film is desired.

Aqueous DPUR appeared on the market much later (in the 1960s), and after only a couple ofyears achieved the quality that allowed them to be widely applied as coating binders. Substantialadvantages of DPUR include excellent mechanical properties and solvent resistance. Theirdrawbacks are relatively high cost and poor tolerance of additives and pigments.

Figure 6.2 Examples of different particle morphologies of hybrid dispersions obtainedby synthesis (such morphologies may be considered intermediate between ‘core-shell’

and ‘true hybrid’ morphology shown in Figure 6.1).

‘engulfed’ or ‘englued’morphology [15] (the

particles of one polymerare ‘glued’ to the

particles of the other)

‘gradient’ morphology [15](there is a gradient ofconcentration of one

polymer in the other in thehybrid particle)

‘fruit-cake’ morphology[16] (particles of one

polymer are distributedin a matrix of the other)

265


Methods of preparation of aqueous acrylic dispersions are well known [19] and willnot be discussed here. Emulsion polymerisation of acrylic monomers is usually carriedout using a water-soluble initiator or a ‘redox’ system. In this process, relatively highlevels (up to several %) of emulsifiers and protective colloids are used. The dispersedpolymer generally has a linear structure and a high molecular weight (MW). Theproperties of dispersions and films are mostly affected by the kind of monomer(s)used [20].

DPUR can be prepared in many ways [21, 22], which eventually lead to one of the structures(polyurethane or polyurethane-urea) shown in Figure 6.3.

In practice, the structures shown in Figure 6.3 are partly or fully crosslinked. The twomain methods which lead to these structures are the ‘acetone process’ (historically older)[23] and the ‘prepolymer-ionomer process’ [24] that is commonly used to manufactureDPUR at present. A combination of these two methods can also be used.

In the ‘acetone process’, the first step is the synthesis of NCO-terminated urethaneprepolymer which is then dissolved in acetone and reacted with a di-hydroxy or di-amino compound containing groups which are the precursors of ionic groups, e.g., -SO3,-COOH, N≡. The next steps are the emulsification of the resulting product in watercombined with the formation of ionic groups and distillation of acetone.

Figure 6.3 Main structures of polyurethane or polyurethane-urea chain in DPUR.(a) Hydrophilic groups are not incorporated into the polymer chain (are present aspendant groups); (b) Hydrophilic groups are incorporated into the polymer chain

(a)

(b)

- polyol chain

- hydrophilic moiety

- hard segment (urethane or urea)

266


In the ‘prepolymer-ionomer process’, the compound which contains the moieties whichare the precursors of ionic groups is incorporated in the polymer chain already at thestage of urethane prepolymer formation. Then they are neutralised, which results in theformation of a ‘prepolymer-ionomer’ which is emulsified in water and eventuallycrosslinked. In this process, the prepolymer-ionomer is usually dissolved in a small amountof water-miscible solvent of high boiling point, e.g., N-methylpyrolidone, which playsthe role of coalescing agent in the process of film formation. It is then possible to obtainDPUR which contain either ‘cationic DPUR’ with a pH of less than 7 (cationic moietiesare attached to the polyurethane or polyurethane-urea chain) or ‘anionic DPUR’ with apH of greater than 7 (anionic moieties are attached to the polyurethane or polyurethane-urea chain). If non ionic hydrophilic moieties are attached to or incorporated in thepolyurethane or polyurethane-urea chain, then a ‘nonionic DPUR’ may be obtained.

Properties of DPUR depend on several factors. The most important are structure of thepolyurethane or polyurethane-urea chain, degree of crosslinking and the amount ofcoalescing agent [25].

6.2.3 Hybrid Acrylic-Urethane Dispersions

Information on the preparation of hybrid acrylic-urethane dispersions by synthesis,published so far is quite scarce. More detailed data on this process are available only inthe patent literature.

The basic process was first patented by Inmont, USA, in 1982 [9]. They simply claimedpolymerisation of vinyl polymers in DPUR. In 1986, a patent for another method ofpreparation of hybrid acrylic-urethane dispersions (diluting of prepolymer-ionomer withmonomers, emulsifying the resulting solution in water and polymerisation) was grantedto Witco, USA [26].

In subsequent years the number of patents dealing with hybrid acrylic-urethane dispersionsincreased rapidly, and approximately 30 patents have been granted so far worldwide,the majority in Japan.

Despite the large number of patents, at the time acrylic-urethane hybrid dispersions enteredthe market (in about 1990), their properties, and specifically their structure/propertiesrelationships and particle morphology, had not been described in the scientific literatureuntil 1992, when Gruber from Air Products presented a paper at the ACS Meeting in SanFrancisco [27] that dealt with the general characterisation of such dispersions. In 1996,Hegedus and Kloiber (also from Air Products) published an important paper [28] wherethey presented the properties not only of dispersions but also of coatings made on theirbasis. However, there was no information on composition and chemical structure of products

267


referred to in their article. To prepare hybrid dispersions, they diluted a prepolymer-ionomerwith acrylic monomers, emulsified them in water and carried out emulsion polymerisationof the monomers. They found, amongst other things, that unlike films obtained from blendsof acrylic dispersion and DPUR, the film obtained from the hybrid dispersion showed onlyone glass transmission temperature (Tg), but no information on the particle morphologywas given. Some data on particle morphology was presented in 1997 in two publicationsby Zeneca [29, 30] who described the properties of hybrid acrylic-urethane dispersionscrosslinked by formation of azomethine groups from the reaction between carbonyl groupspresent as pendant groups in the polyurethane-urea chain and amino groups from thepolyamine used for crosslinking the dispersion by reaction with free NCO groups of theprepolymer-ionomer.

In one of these articles [30] the properties of this kind of dispersion were compared withthe properties of its analogue crosslinked additionally by oxidation of double bondspresent in polyurethane-urea chain (air-drying) and it was concluded that the formerresulted in higher coating hardness. The authors of both publications presented theirschematic drawings showing the ‘core-shell’ morphology of their dispersion particles,but did not reveal any experimental evidence that would confirm such morphology.

Another interesting paper dealing with hybrid vinyl/acrylic-urethane dispersions was publishedin 1996 by Cytec [31]. They prepared hybrid dispersions by swelling DPUR particles withmonomers at a polyurethane:vinyl component ratio of 3:1 (w/w) and then carrying outemulsion polymerisation. The resulting hybrid dispersions were additionally crosslinked bythe reaction of epoxy groups (when glycydyl methacrylate was used as one of the monomers)with carboxylic groups attached to the polyurethane-urea chain or by using TMI (unsaturatedaliphatic isocyanate; Cytec industries, USA) as one of the monomers. The authors suggestedthat IPN structures were formed in these reactions. The paper contains very good photographsof dispersion particles, but no information on particle morphology is given.

The effect of type of diisocyanate and polyol used in synthesis of hybrid PU/acrylic polymerdispersions was recently reported by authors from Lamberti [32].

A recently published patent [33] from Zeneca described in detail the methods ofpreparation of hybrid polyurethane/vinyl polymer dispersions but neither discussed theeffect of various factors on the properties of the hybrid dispersions nor provided anyinformation on their particle morphology.

Based on this brief literature survey, it seems quite clear that not all features of the hybridacrylic-urethane dispersions have been studied so far. The purpose of the present study isto clarify the effect of various factors on the properties of acrylic/styrene-urethanedispersions. Some results of preliminary investigations of the particle morphology willalso be revealed.

268


6.3 Concept of the Study

6.3.1 Selection of Starting Materials

It was decided that synthesis of DPUR would be carried out according to the ‘prepolymer-ionomer’ method generally described in Section 6.2. The scheme of the reactions leadingto DPUR is presented in Figure 6.4.

For the synthesis of DPUR, a cycloaliphatic diisocyanate, isophorone diisocyanate (IPDI) fromHüls, was used since unlike aromatic isocyanates, it does not result in yellowing of the coatingsand films. The polyols used were polytetramethylene glycol (PTMG) with a MW of 2000

Step I - polyaddition

Step II - neutralisation

Figure 6.4 Schematic representation of the reactions leading to DPUR using the‘prepolymer-ionomer process’

Step III - emulsifying + crosslinking

269


supplied by BASF and commercially available polyesterdiols commonly used in coatings (Bester190 and 195, supplied by Polychimica) with hydroxyl numbers equal to 32 and 122, respectively.Polyesterdiol containing double bonds (Irpurate 240Tz from InterRokita, Poland, with hydroxylnumbers equal to 276) was also used in some experiments. Both polyether- and polyester-diolswere used to look at the effect of structure of the polyurethane-urea chain on the properties ofthe dispersions, films and coatings. N-methylpyrrolidone (NMP) from Sidaco was used as co-solvent (coalescing agent) and 2,2´-dimethylolpropionic acid (DMPA) from Angus was used asthe source of pendant carboxylic groups in the polyurethane-urea chain.

Polymerisation of acrylic monomers was carried out according to the standard procedureusing two different initiating systems:

(a) Water-soluble initiator – potassium persulphate

(b) ‘redox’ system – cumene hydroperoxide (CHP) + Rongalit with the structure:

Rongalit

The choice of initiating systems was made by considering their possible effect on particlemorphology (see section 6.3.2). The other additives (emulsifiers, protective colloids) werestandard materials commonly used in emulsion polymerisation of acrylic monomers.

It was assumed that three monomers would be used in this study: methyl methacrylate,butyl acrylate and styrene. The possiblity of using styrene seemed to be especially interestingbecause of its very low cost compared to the other monomers. The other two monomerswere selected for the significant differences in their polarity and solubility in water thatmight affect the particle morphology and the Tg of poly(methyl methacrylate) and poly(butylacrylate). Difunctional acrylate monomer (1,4-butanediol dimethacrylate) was also usedin some experiments since it encouraged the formation of an IPN structure.

6.3.2 Assumptions for Synthesis of Hybrid Dispersions

6.3.2.1 General Considerations

Based on the literature survey (see Section 6.2) and the author’s preliminary experimentsthe following methods of synthesis of hybrid dispersions (after this point in the chapterthey will be called modified DPUR (MDPUR)) were selected:

270


1) radical polymerisation of monomers in DPUR obtained as presented in Figure 6.4and then diluted with water.

a) with continuous feeding of monomers

b) after swelling of DPUR particles with monomers for 20 hours

2) using the monomers as active diluents (co-solvents) of prepolymer-ionomer followedby their polymerisation after emulsifying the prepolymer-ionomer in water

3) emulsifying the prepolymer-ionomer in diluted acrylic-styrene dispersion (ASD)

It was assumed that the method of MDPUR synthesis would have be the major factorthat would determine particle morphology. In Figures 6.5-6.7, the methods (1) – (3)applied in this study are presented schematically.

Hybrid dispersions obtained by polymerisation, i.e., according to methods 1a, 1b and 2,will be designated as MDPUR-ASD and hybrid dispersions obtained by synthesis of PURin ASD, i.e., according to method 3, will be designated as MDPUR.

Figure 6.5 Method 1 (radical polymerisation of monomers in DPUR

271


Figure 6.6 Method 2 (using the monomers as active diluents for prepolymer-ionomer)

Figure 6.7 Emulsifying the prepolymer-ionomer in diluted ASD

272


6.3.2.2 Analysis of the Process of MDPUR Synthesis Using DispersionParticle Morphology

• Method 1: Polymerisation of monomers in DPUR

Before the polymerisation starts, the system will consist of the following components:

(a) DPUR particles that will be only slightly swelled with monomers (in Method 1a –continuous feeding of the monomers) or fully swelled with monomers (in Method 1b– swelling of DPUR particles before polymerisation)

(b) Monomer droplets which will be obviously more numerous in Method 1a

(c) Monomer dissolved in water

~ ~~~

~~~

~~~~~

~ ~ ~

~~

~ ~

~~~ ~ ~

(d) Micelles of emulsifier

273


When the process is carried out according to Method 1a it may be assumed that the kindof initiating system (water-soluble initiator or redox system) does not have any significanteffect on the morphology of hybrid dispersion particles since the polymerisation proceedsmainly on the surface of DPUR particles and monomer droplets and in the micelles ofemulsifier. In the case of water-soluble initiator (potassium persulphate) the polymerisationstarts in water where the ion-radicals are formed from the monomer molecules dissolvedin water. These ion-radicals can diffuse to DPUR particles, monomer droplets andemulsifier micelles. After this happens the system will consist of:

(a´) DPUR particles partly swelled with monomer, with adsorbed ion-radicals

~ ~~~

~~~

~~~~

~ ~ ~

~~~

~~~ ~

~~

(b´) monomer droplets of the size much smaller than before polymerisation and with theadsorbed ion-radicals

~

~~

(c´) micelles of emulsifier with the adsorbed ion-radicals

274


After polymerisation is completed both acrylic/styrene polymer particles (a) and the hybridparticles (b1 and b2) may be formed:

Hybrid particles of uniform structure (‘true hybrid’) (b1) or hybrid particles of ‘core-shell’ structure where the core is composed of polyurethane-urea (b2):

Obviously, other morphologies of dispersed particles are also possible (see Section 6.2).The factor which probably has quite a significant effect on formation of separate particlesof the acrylic/styrene polymer is concentration of initiator. The higher it is, the higher isthe probability that polymerisation would proceed in monomer droplets. If it is relativelylow, this chance is diminished since monomer droplets diffuse into DPUR particles.Probability of formation of hybrid particles of more uniform structure increases withincrease in solubility of monomer in water since longer-lived ion-radicals may be formedand their adsorption on DPUR particles is easier. The second important factor affectingthe hybrid particle morphology is solubility of monomer in polyurethane-urea whichconstitutes the particles of DPUR. If it is higher, swelling of DPUR particles with monomeris easier and the probability of formation of hybrid particles of more uniform structure ishigher.

When polymerisation is carried out after initial swelling of DPUR particles in monomers(Method 1b), then formation of separate particles of acrylic/styrene polymer is also possible,but much less probable since the amount of monomer present in the system as monomerdroplets is very low in this case (assuming that the amount of monomer used to swellDPUR particles was not higher than the maximum amount resulting from the equilibriumswell curve (see Section 6.5.4)). It can be anticipated that in this method of MDPUR synthesismany more hybrid particles of much more uniform structure will be formed than when themonomers are continuously added to the reaction system (Method 1a), especially when a

(a)

(b1) (b2)

275

‘redox’ type of initiating system is used. Furthermore, formation of hybrid particles ofreversed morphology (core – acrylic/styrene polymer, shell – polyurethane-urea) cannot beexcluded, especially when the polarity of monomer will be low:

If this happens, then gelling of the whole reacting system becomes very probable.

• Method 2: Using monomers as active diluents for prepolymer-ionomer

In this case it can be assumed that polymerisation will proceed in a similar way as describedpreviously for Method 1b. The only difference may be that both structure and size ofDPUR particles formed from the prepolymer-ionomer diluted with monomers will differfrom the ‘standard’ case, i.e., when NMP is used as diluent for prepolymer-ionomer.Moreover, some fraction of the monomer molecules may be so closely ‘trapped’ in thecrosslinked polyurethane-urea structure in DPUR particles that their participation inpolymerisation will be difficult. This may lead to a decrease in average MW of acrylic/styrene polymer which will be formed inside DPUR particles.

• Method 3: Emulsifying prepolymer-ionomer in diluted acrylic/styrene dispersion

In this case, the course of MDPUR synthesis seems to be much simpler than in Methods1 and 2 because no polymerisation of monomers will proceed in the system. Before theprocess starts, only ASD particles are present in the system:

After addition of prepolymer-ionomer and crosslinking it with polyamine some of theseparticles may remain untouched, but both new hybrid ‘core-shell’ particles where acrylic/styrene polymer will constitute the core and polyurethane-urea will form the shell:


276


as well as DPUR particles:

will appear in the system.

It may be assumed that the factors that would decide the course of the process and thefraction of hybrid particles would be:

(a) Difference in surface tension of prepolymer-ionomer and the emulsifier used insynthesis of ASD. If the surface tension of the former is higher, then formation of‘core-shell’ hybrid dispersion particles is also higher since adsorption of prepolymer-ionomer on acrylic/styrene polymer dispersion particles is easier. If the surface tensionof the latter is higher, then adsorption is more difficult and a greater number ofseparate DPUR particles may be formed.

(b) Polarity of acrylic/styrene polymer. If it is high, then formation of ‘core-shell’ hybridparticles is less probable because of difficult adsorption of prepolymer-ionomer. If itis low, then it is expected that formation of hybrid dispersion particles will be easier.

6.4 Methods of Testing

6.4.1 Dispersions

The viscosity of prepolymer-ionomers and dispersions was determined at 25 °C using aHoppler Rheoviscometer. The NCO content of the prepolymer-ionomer was measuredby titration. MW was determined by gel permeation chromatography (GPC) accordingto ISO/TC61/ISC5 using tetrahydrofuran (THF) as solvent and Shimadzu CR 4A

277

Chromatopac apparatus. Free monomer content in the dispersions was measured by gaschromatography (GC) (Giede 18.3 apparatus)

Mechanical stability of dispersions was determined using a laboratory centrifuge withrotational speed of 3700 rpm. Particle size, particle size distribution and zeta potentialof the dispersions was determined using Malvern Zeta Sizer 4 equipment. Minimumfilm formation temperature (MFFT) was measured using a Coesfeld apparatus.

To produce photographs of dispersion particles, transmission electron microscopy(TEM) was used. Photographs were taken at the Lomonosov University, Moscow, (seeFigures 6.12, 6.16 and 6.23-6.25) and the Institute of Chemical Physics of the RussianAcademy of Sciences in Moscow), see Figures 6.31-6.33. A special method developedin this Institute was used for investigating dispersion particle morphology. Accordingto this method, a drop of dispersion (very diluted, 0.5-1.0%) and stained with osmiumtetroxide) was placed between the ‘thermal electrodes’ of a ‘thermodiffusion’ apparatusconstructed at the Institute. The temperature gradient between the electrodes was 35°C and the time taken to settle the dispersion particles was 30-40 minutes. The substratewith the settled particles was then washed with water and dried. Photographs weretaken using a TEM-EM 301 apparatus (Philips). It was shown in previous studies [34]that the substrate does not affect the dispersion particle morphology.

The maximum degree of swelling of DPUR particles with monomers was determinedusing the equilibrium swelling method described in Section 6.4.3.

6.4.2 Coatings

Coatings were made from dispersions by applying them onto degreased glass plates witha 120 μm gap applicator. The resulting coating was dried at room temperature for 72 h(for hardness and adhesion tests) and at 40 °C (for water resistance test).

Water resistance was determined by examination of the coating after 24 hours immersion.It was designated by numbers (1-4) and letters (A-D) where the higher number means ahigher number of bubbles and the ‘higher’ letter means a higher dissipation of bubbleson the plate [35]. Adhesion was determined by a ‘double-cutting knife’ method anddesignated by numbers (1-4). Higher numbers mean lower adhesion. For determinationof hardness the pendulum method (Persoz) was applied [36].

The drying time of the coatings was determined according to the relevant Polish standard(corresponding to DIN 53150 [37]). Times needed to achieve 1° of hardness (touch dry)and 3° of hardness (hard dry) were measured.


278


6.4.3 Films

Films were obtained from dispersions by casting them onto glass Petri dishes and allowingthem to dry for 14 days at room temperature. Mechanical properties, i.e., tensile strength(dr), stress at 100% elongation (d100), elongation at break (er) and Shore A hardness weredetermined according to ISO 527 [38] (strip type specimens were tested).

Percentage swell in water, methylethylketone (MEK) and xylene was determined forsquare specimens (25 x 25 mm) after 2 weeks of immersion at room temperature andcalculated from the following equation:

% swellingL= ⋅ ⎛

⎝⎜⎞⎠⎟

−10025

1003

where L is maximum length.

The Tg was determined by differential scanning calorimetry (DSC) according to ASTMD3418-82 [39].

Surface free energy was calculated from experimental values of dynamic contact angledetermined using a tensiometer constructed in the Institute of Chemical Physics of theRussian Academy of Sciences.

The degree of crosslinking was calculated from the maximum degree of swelling determinedaccording to the following procedure: samples of film weighing approximately 0.1 g eachwere immersed in toluene and the degree of swelling was calculated from the equation:

α τ τ( )

( )= −m mm

0

0

where m(τ) – mass of the sample after time τα(τ) – degree of swelling, and

m0 – initial mass of the sample.

After the maximum degree of swelling was reached (determined graphically), the samplewas weighed and dried to constant solids content (mc). Content of the soluble fraction ofthe sample (S, %) was calculated according to the equation:

Sm m

mc= −0

0

279

Crosslinking density (j) was calculated from the equation:

jS S

=+1

and fraction of active chains (VS), from the equation:

V S jS jss = − − +( ) ( )( )1 1 2 12

6.5 Experimental results

6.5.1 Characterisation of Starting Dispersions Used for Synthesis of MDPUR

Table 6.1 shows the properties of DPUR used for synthesis of MDPUR (MDPUR-ASD)together with the properties of prepolymer-ionomers used as intermediates in theirpreparation.

Figures 6.8 and 6.9 show the MW distribution for standard prepolymer-ionomers usedin DPUR synthesis, obtained from polyetherdiol (MW = 2000) and a mixture ofpolyesterdiols (MW = 920 and 3190), respectively. The weight average MW (Mw) andnumber average MW (MN) of these prepolymer-ionomers were (not counting free IPDI):

- using polystyrene standard

MN MW MW/MN

Polyetherdiol-based 6083 13604 2,24

Polyesterdiol-based 6619 16010 2,41

- using polyethylene glycol (PEG) standard

MN MW MW/MN

Polyetherdiol-based 5215 11153 2,1

Polyesterdiol-based 6999 16811 2,40

For DPUR the MW distribution could not be determined by GPC as the films did notdissolve completely.


280

Advances in Urethane Science and Technologyb1

dnaa1

sdohtem

otgnidrocca

deniatboDS

A-R

UPD

Mrof

slairetam

gnitratssa

desuR

UPD

foseitreporP

1.6elba

Tfo

dna)2.3.6

noitceSees(

ehtfo

seitreporP.noitaraperprieht

nidesu

)enotecaro

PM

Nhti

wdetulid(

sremonoi-re

myloperpDS

A-R

UPD

Mfo

sisehtnysni

setaidemretni

sadesu

)R

UPPsa

detangised(sre

monom

htiw

detulidsre

monoi-remyloperp

.elbatsiht

nidedulcni

neebosla

evah2

dohtem

otgnidrocca

noitangiseD

a

RUP

Dfo

)R

UPP(loyloP

PM

Ntnetno

C%

-remyloperp

foseitreporP

sremonoi

RUP

Dfo

seitreporP

%,O

CN

ytisocsiV

s-aPm

sdiloS%

ytisocsiV

s-aPm

Hplacinahce

M,ytilibatS

setunim

elcitraPezis

noitubirtsid

egarevA

elcitraPezis

ateZ

laitnetopV

mTFF

M°C

.roehT

.tcarPmn

mn

732R

UPD

)233,442,042(0002-

GM

TP0

32.395.2

005,79.03

029.7

09>55-73031-57

3.856.45-

0<

852R

UPD

b0002-

GM

TP3.51

57.251.2

027,41.63

5133.7-0.7

09>024-021

8.1612.12-

0<

882R

UPP0002-

GM

TP2.3

05.286.2

--

--

--

--

-

982R

UPP0002-

GM

TPetaruprI

dna3

66.369.3

--

--

--

--

-

333R

UPP0002-

GM

TP0

57.284.2

--

--

--

--

-

543R

UPD

0002-G

MTP

8.7147.2

23.2083

9.53002

9.7-6.709>

002,1-0111.112

9.7-0<

643R

UPD

0002-G

MTP

342.3

40.3005,7

5.1302

6.706

077-575.512

5.4-0<

453R

UPD

,091retse

BetaruprI

5910

23.427,4

003,60.93

8014.7-1.7

09>041-03

2.656.06-

0<

653R

UPD

0002-G

MTP

etaruprIdna

435.4

32,4008,3

4.825.51

4.7-1.709>

058-031.57

8.89-0<

rebmun

elpmas

yrotarobaleht

otrefer

sre bmun

ehta

desusa

weni

maylopgniknilssorc

tnereffidb

281

Particle size distribution and zeta potential determined for one of the DPUR from Table6.1 are shown in Figures 6.10 and 6.11, respectively. TEM photographs of the particlesof the same dispersion are shown in Figure 6.12.

Properties of films and coatings made from DPUR included in Table 6.1 are shown inTables 6.2 and 6.3, respectively.

DSC thermogram for a typical DPUR from Table 6.1 (DPUR-237) is presented in Figure 6.13.

Properties of ASD used as starting materials in synthesis of MDPUR according to method3 are presented in Table 6.4.

Figure 6.8 MW distribution of prepolymer-ionomer synthesised from polyetherdiol(PTMG 2000, MW = 2000) as determined by GPC

Figure 6.9 MW distribution of prepolymer-ionomer synthesised from a mixture ofpolyesterdiols (Bester 190 and 195, MW 920 and 3190, respectively)


282


Figure 6.10 Particle size distribution for typical DPUR synthesised in this study (DPUR268 = DPUR 237 from Table 6.1)

Cell type AZ104 Cross Beam ModeF(ka) = 1.50 (Smoluchowsky)

Data taken on 20/11/96 at 14:48:58

Cell voltage 138.1 V Current 0.1 mAConductivity 0.03 mS Temperature 24.8Viscosity 8.95 x 10-4 Pa-s Dielectric Constant 79.0Count rate 491761.9 pH 3.91

f(ka) 1.50 Cell position 85.40

Zeta Potential -54.6 mV StDev 5.7

ateZ ytisnetnI ateZ ytisnetnI ateZ ytisnetnI

0.051-5.731-0.521-5.211-0.001-

5.78-0.57-5.26-

0.00.00.00.00.00.01.58.83

0.05-5.73-0.52-5.21-

0.0-5.210.525.73

8.341.21

1.00.00.00.00.00.0

0.055.260.575.780.0015.2110.5215.731

0.00.00.00.00.00.00.00.0

7.5htdiw6.45-naem1kaeP

Figure 6.11 Zeta potential for typical DPUR synthesised in this study(DPUR 268 = DPUR 237 from Table 6.1)

283

Figure 6.12 Particles of typical DPUR synthesised in this study(DPUR 237 from Table 6.1). Photograph was taken using TEM technique.Reproduced with permission from Professor I. A. Grickova at Lomonosov

University, Moscow.

sisehtnysrofslairetamgnitratssadesuRUPDmorfsmliffoseitreporP2.6elbaT.b1dnaa1sdohtemotgnidroccaDSA-RUPDMfo

noitangiseDRUPDfo

smliffoseitreporP

.raeppAδδδδδ 001

aPM

δδδδδr

aPM

εεεεεr

%

llews% Tg

°CH2O KEM enelyX

732RUPD)233,442,042(

T 2.3 0.91 086 ,hWseceiP

seceiP seceiP 1.97-

543RUPD T 50.7 9.7 031 1tNhW

4.471rB

031rB

6.17-

643RUPD PrO,T 1tN 1tN 1tN 1tN 1tN 1tN 1.17-

453RUPD ,TskcarC

- - - - - - 7.51-

653RUPD lE,T - 3.61 72 21 991rB,T

7.901rB,T

0<

ootfoesuacebdetsetton-1tN,leepegnaro-PrO,etihw-hW,elttirb-rB,citsale-lE,tnerapsnart-Tetartsbusotnoisehdahgih

Synthesis and Characterisation of Aqueous Hybrid Polyurethane-urea…

284


rofslairetamgnitratssadesuRUPDmorfsgnitaocfoseitreporP3.6elbaT.b1dnaa1sdohtemotgnidroccaDSA-RUPDMfosisehtnys

noitangiseDRUPDfo

sgnitaocfoseitreporP

nim,emitgniyrD ssendraH)zosreP(

noisehdAretaW

ecnatsiser1(ssendraH °) 3(ssendraH °)

732RUPD)442,042(

01 51 21.0 4 selbbuboN

543RUPD 081 043 461.0 1 D1

643RUPD 52 05 871.0 3 2tN

453RUPD 52 52 52.0 enoN -

653RUPD 04 06 63.0 4 tonseod,C1netihw

ecafrusnevenufoesuacebdetsetton-2tNsedocecnatsiserretawfonoitanalpxenarof2.4.6noitceseeS

Figure 6.13 DSC thermogram for the film obtained from a typical DPUR synthesisedin this study (DPUR 237 from Table 6.1).

285

3dohte

mot

gnidroccaR

UPD

Mfo

sisehtnysrof

slairetam

gnitratssa

desuDS

Afo

seitreporP4.6

elbaT ngise

DDS

Afo

remono

MrotaitinI

DSA

foseitreporP

,sdiloS%

ytisocsiV

s-aPm

Hphce

Mytilibats

nim

eerF,.

monom

%

elcitraPrtsid

ezismn

egarevA

ezistrapmn

ateZ

laitnetopV

m

TFFM

°CT

g

°C

51DS

AS/

MM/

AB

/23.1/1()46.1

8O2S2

K2.42

022.6

54/1.0/40.0

1.0541-04

7.939.21-

05+>0.85+

71DS

AM

M8

O2S2K

9.4241

0.5-0.454

00.158-53

0.454.34-

05+>1.321+

71DS

AA

B8

O2S2K

9.426

0. 5-0.454

35.0041-03

7.953.14-

0.88-

02DS

AS

8O2S2

K3.32

53.5-0.5

5442.0

081-045.97

5.55-05+>

0.89+

462DS

AS/

MM/

AB

/23.1/1()46.1

/PH

Ctilagno

R5.22

55. 6-2.6

09>t

Nt

Nt

Nt

N05+>

tn

662DS

Aevoba

sA

K2S

2O8

4.126

5.6-2.609>

tN

tN

tN

tN

05+>tn

etalyrcalytub:A

B :Seneryts

etalyrcahtemlyhte

m:M

Mdetset

ton:tN


286


In Figures 6.14, 6.15 and 6.16 the particle size distribution, zeta potential and TEMmicrograph of dispersion particles of typical ASD from Table 6.3 (ASD sample number15) are presented.

Figure 6.14 Particle size distribution for typical ASD synthesised in this study (ASD 15from Table 6.3).

287

Figure 6.16 TEM micrograph of dispersion particles of ASD (see Table 6.3).Reproduced with permission from Professor I. A. Grickova at Lomonosov

University, Moscow.

Figure 6.15 Zeta potential for typical ASD synthesised in this study(ASD 15 from Table 6.3).


288


6.5.2 Synthesis of MDPUR and MDPUR-ASD

Hybrid polyurethane-urea-acrylic/styrene polymer dispersions were prepared accordingto methods ‘1a’, ‘1b’, ‘2’ and ‘3’ described in Section 6.3.2. Dispersions designated asMDPUR-ASD were made by polymerisation of monomers in DPU according to themethods 1a, 1b and 2 while dispersions designated as MDPUR were made by synthesisof DPUR in ASD according to method 3. In all syntheses the ratio of polyurethane-ureato acrylic/styrene polymer in the hybrid was 2:1.

Tables 6.5, 6.6, 6.7 and 6.8 show the compositions of MDPUR-ASD prepared accordingto methods 1a, 1b, 2 and 3, respectively.

a1dohtemotgnidroccaderaperpDSA-RUPDMfonoitisopmoC5.6elbaT)sremonomehtfognideefsuounitnochtiwRUPDnisremonomfonoitasiremylop(

fonoitangiseDDSA-RUPDM

fonoitangiseDRUPDgnitrats)1.6elbaTees(

remonoM )PMN(tnecselaoC-RUPDMnitnetnoc

%,DSA

22DSA-RUPDM 042RUPD MM 0

32DSA-RUPDM 442RUPD AB 0

42DSA-RUPDM 442RUPD S,MM,AB 0

41DSA-RUPDM 442RUPD S 0




003DSA-RUPDM 453RUPD S,MM 0

703DSA-RUPDM 653RUPD S,MM,AB 3.3

903DSA-RUPDM 653RUPD S,MM,AB 3.3

b1dohtemotgnidroccaderaperpDSA-RUPDMfonoitisopmoC6.6elbaT)sremonomhtiwselcitrapRUPDgnillewsretfaRUPDnisremonomfonoitasiremylop(

fonoitangiseDDSA-RUPDM

fonoitangiseDRUPDgnitrats)1.6elbaTees(

remonoM)PMN(tnecselaoC-RUPDMnitnetnoc

%,DSA

072DSA-RUPDM 643-RUPD S,MM,AB 2

272DSA-RUPDM 643-RUPD S,MM,AB 2

403DSA-RUPDM 653-RUPD S,MM,AB 2.3

289

remonoi-re

myloperpeht

fognitulid(

2dohte

mot

gnidroccaderaperp

DSA-

RUP

DM

fonoitisop

moC

7.6elba

T-re

myloperpgnitrats

foseitreporP.)re

monoi-remyloperp

fogniknilssorc

dnanoitasire

mylop,sremono

mni

.dedulcniosla

erasre

monoi

fonoitangise

DDS

A-R

UPD

M

fonoitangise

Dgnitrats

-remyloperp

remonoi

)R

UPP(

loyloPre

monoM

tnecselaoC

)PM

N(ni

tnetnocDS

A-R

UPD

M%

remonoi-re

myloperpgnitrats

foseitreporP

%,O

CN

taytisocsi

V 52°

s-aPm,

C.teroeh

T.tcarP

DSA-

RUP

DM 79

882R

UPP0002-

GM

TPS,

MM,

AB

305 .2

86.2067

DSA-

RUP

DM 89

882R

UPP0002-

GM

TPS,

MM,

AB

305.2

86.2067

DSA-

RUP

DM

201982

RUPP

0002-G

MTP

S,M

M,A

B3

66.369.3

037

DSA-

RUP

DM

891333

RUPP

0002-G

MTP

etaruprIS,

MM,

AB

-57.2

84.2t

N

detsetton

–t

N

.)DS

Ani

RUP

Dfo

sisehtnys(3

dohtem

otgnidrocca

deraperpR

UPD

Mfo

snoitisopmo

C8.6

elbaT

dedulcniosla

erasre

monoi-remyloperp

gnitratsfo

seitreporP

fonoitangise

DR

UPD

MloyloP

tnecselaoC

tnetnoC

)PM

N(%,

RUP

DM

ni

fonoitangise

DDS

Agnitrats

remonoi-re

myloperpgnitrats

foseitreporP

%,O

CN

taytisocsi

V 52°

s-aPm,

C.teroeht

.tcarP

342R

UPD

M0002-

GM

TP6.11

51DS

A47.2

46.20004

542R

UPD

M0002-

GM

TP6.11

71DS

A47.2

43.20054

742R

UPD

M0002-

GM

TP6.11

81DS

A47.2

42.20024

052R

UPD

M0002-

GM

TP6.11

02DS

A47.2

15.20004

743R

UPD

M0002-

GM

TP6.3

462DS

A21.3

58.20075

843R

UPD

M0002-

GM

TP6.3

662DS

A21.3

97.20005


290


6.5.3 Investigation of the Effect of Various Factors on the Properties ofHybrid Dispersions

6.5.3.1 Method of Hybrid Dispersion Synthesis and Kind ofInitiating System

The results of investigations of the effect of method of hybrid dispersion synthesis(1a, 1b, 2 or 3 – see Section 3.2) on the properties of dispersions as well as of filmsand coatings made from them are presented in Tables 6.9 to 6.11 (dispersionsprepared using water-soluble initiator) and in Tables 6.12 to 6.14 (dispersionsprepared using ‘redox’ initiating system). In all dispersions the chemical structureof the polyurethane-urea and acrylic/styrene polymer component was the same (seerelevant tables in Section 6.5.2). All the dispersions contained a similar low level(2-3.6%) of NMP.

Particle size distribution and zeta potential of a typical hybrid dispersion obtained usinga water-soluble initiator are presented in Figures 6.17 and 6.18, respectively.

291

tnereffidot

gnidroccaderaperp

noitisopmoclaci

mehce

maseht

fosnoisrepsid

dirbyhfo

seitreporP9.6

elbaT

K(rotaitini

elbulos-retaw

agnisu

sdohtem

2S2O

8)

dohteM fo

sisehtnyS

noitangiseD

dirbyhfo

noisrepsid

snoisrepsidfo

seitreporP

,sdiloS%

ytisocsiV

s-aPM

Hplacinahce

MytilibatS

niM

eerF,re

monom

%,S/M

M/A

B

eziselcitraP

noitubirtsid

egarevA

ezistrapmn

topate

ZV

MTFF

M°C

a1-

RUP

DM

862DS

A7.03

94.7-1.7

51t

N032-03

9.164.81-

0<

b1-

RUP

DM

072DS

A2.43

011.7-8.6

51t

N076-05

4.491.93-

0<

2-

RUP

DM

79DS

A4.13

011.7-8.6

51/20.0/7.0

510.0021-05

4.8017.06-

0<

3R

UPD

M843

6.14002

7.7-4.751

tN

025-554.511

0.32-0

<

detsetton

–t

N

deraperpnoitisop

moclacimehc

emas

ehtfo

snoisrepsiddirbyh

morfeda

ms

mliffo

seitreporP01.6

elbaT

K(rotaitini

elbulos-retaw

gnisusdohte

mtnereffid

otgnidrocca

2S2O

8)

dohteM fo

sisehtnyS

noitangiseD

dirbyH

fonoisrepsi

D

smlif

foseitreporP

.raeppA

δ001

aPM

δ r

aPM

ε r

%

llews

%T

g°C

H2O

KE

Menely

X

a1-

A-R

UPD

M862

DSl

E,T

2.88.01

461l

E,hW

31r

B,T

743r

B,T

8914.37-

b1-

A-R

UPD

M072

DSl

E,T

2.93.31

081l

E,hW

0r

B,T

743r

B,T

4224.37-

2-

A-R

UPD

M79

DSl

E,T

5.60.01

012h

W31

G,T

143G,

T013

6.65-

3R

UPD

M843

lE,

M1.41

5.41711

lE,h

W31

rB,

T891

rB,

T891

7.57-

suonitaleg-

G,elttirb-

rB,etih

w-

hW,ta

M-

M,citsale

-lE,tnerapsnart

-T


292

Advances in Urethane Science and Technologyderaperp

noitisopmoclaci

mehce

maseht

fosnoisrepsid

dirbyhmorf

edam

sgnitaocfo

seitreporP11.6

elbaT

K(rotaitini

elbulos-retaw

agnisu

sdohtem

tnereffidot

gnidrocca2S

2O8)

fodohte

MsisehtnyS

dirbyhfo

noitangiseD

noisrepsiD

sgnitaocfo

seitreporP

nim,e

mitgniyr

Dssendra

H)zosreP(

noisehdA

ecnatsiserreta

W1°

3°a1

862DS

A-R

UPD

M03

0522.0

4h

1retfa

sraeppasidgninetih

wA1

b1072

DSA-

RUP

DM

5254

02.04

h1

retfasraeppasid

gninetihw

C2

279

DSA-

RUP

DM

5306

21.04

h1

retfasraeppasid

gninetihw

B2

3843

-R

UPD

M02

5202.0

1h

1retfa

sraeppasidgninetih

wA2

sedocecnatsiser

retaw

fonoitanalpxe

narof

2.4.6noitces

eeS

tnereffidot

gnidroccaderaperp

noitisopmoclaci

mehce

maseht

fosnoisrepsid

dirbyhfo

seitreporP21.6

elbaT

rotaitini’xoder‘a

gnisusdohte

m

dohteM fo

sisehtnyS

noitangiseD

dirbyhfo

noisrepsid

snoisrepsidfo

seitreporP

sdiloS%

ytisocsiV

s-aPM

Hplacinahce

MytilibatS

niM

eerF,re

monom

S/M

M/A

B%

eziselcitraP

noitubirtsid

egarevA

ezistrapmn

ateZ

laitnetopV

M

TFFM

°C

a1-

RUP

DM

472DS

A7.92

71. 7-8.6

51t

N013-53

0.863.93-

0<

b1-

RUP

DM

272DS

A3.92

014.7-1.7

51t

N056-58

4.096.23-

0<

2-

RUP

DM

89DS

A4.13

81.7-8.6

51/20.0</800.0<

100.0<561-501

6.0310.93-

0<

3R

UPD

M743

5.2384

9.7-6.751

tN

062-067.141

4.52-0

<

detsetton

–t

N

293

deraperpnoitisop

moclacimehc

emas

ehtfo

snoisrepsiddirbyh

morfeda

ms

mliffo

seitreporP31.6

elbaT

rotaitini’xoder‘a

gnisusdohte

mtnereffid

otgnidrocca

fodohte

MsisehtnyS

dirbyH

fonoitangise

Dnoisrepsi

D

smlif

foseitreporP

.raeppA

001δ001δ001δ001δ001δ

aPM

ρδρδ ρδρδρδ

aPM

εε εεεr %

llews

%T

g

°CH

2OK

EM

enelyX

a1472

DSA-

RUP

DM

lE,

T3.6

2.11052

lE

62r

B,T

544r

B,T

4737.86-

b1272

DSA-

RUP

DM

lE,

T6.7

0.21012

lE,h

W,0r

B,T

082r

B,T

0822.27-

289

DSA-

RUP

DM

lE,

T3.5

6.01582

lE,h

W04

G,T

473r

B,T

9046.07-

3743

RUP

DM

le,M

5.210.51

781l

E,hW

0r

B,T

031r

B,T

4712.77-

suonitaleg-

G,elttirb-

rB,etih

w-

hW,ta

M-

M,citsale-l

E,tnerapsnart-

T

deraperpnoitisop

moclacimehc

emas

ehtfo

snoisrepsiddirbyh

morfeda

msgnitaoc

foseitreporP

41.6elba

T.rotaitini’xoder‘

agnisu

sdohtem

tnereffidot

gnidrocca

fodohte

MsisehtnyS

dirbyhfo

noitangiseD

noisrepsid

sgnitaocfo

seitreporP

nim,e

mitgniyr

Dssendra

H)zosreP(

noisehdA

ecnatsiserreta

W1°

3°

a1472

DSA-

RUP

DM

5254

81.03

D2h

1retfa

sraeppasidgninetih

w

b1272

DSA-

RUP

DM

5205

61.04

C3h

1retfa

sraeppasi dgninetih

w

289

DSA-

RUP

DM

04021

80.01

A1h

1retfa

sraeppasidgninetih

w

3743

RUP

DM

0304

91.01

A4h

1retfa

sraeppasidgninetih

w

sedocecnatsiser

retaw

fonoitanalpxe

narof

2.4.6noitces

eeS


294


Figure 6.17 Particle size distribution for a typical polyurethane-urea-acrylic/styrenehybrid dispersion synthesised in this study (MDPUR-ASD 97 from Table 6.7 prepared

according to method 2 using a water-soluble initiator).

Figure 6.18 Zeta potential for a typical polyurethane-urea-acrylic/styrene hybriddispersion synthesised in this study (MDPUR-ASD 97 from Table 6.7 prepared

according to method 2 using a water-soluble initiator)

295

The DSC thermogram of the film obtained from a typical hybrid dispersion obtainedusing a water-soluble initiator is presented in Figure 6.19.

Particle size distribution and zeta potential of a typical hybrid dispersion obtained usinga ‘redox’ initiator are presented in Figures 6.20 and 6.21, respectively.

The DSC thermogram of the film obtained from a typical hybrid dispersion obtainedusing ‘redox’ initiator is presented in Figure 6.22.

Figure 6.19 DSC thermogram for typical hybrid polyurethane-urea-acrylic/styrenedispersion synthesised in this study (MDPUR-ASD 97 prepared according to method 2

using water-soluble initiator)


296


Figure 6.21 Zeta potential for typical polyurethane-urea-acrylic/styrene hybriddispersion synthesised in this study (MDPUR-ASD 98 from Table 6.7 prepared

according to method 2 using a ‘redox’ initiator)

Figure 6.20 Particle size distribution for typical polyurethane-urea-acrylic/styrenehybrid dispersion synthesised in this study (MDPUR-ASD 98 from Table 6.7 prepared

according to method 2 using a ‘redox’ initiator)

297

6.5.3.2 Concentration of Coalescing Agent (NMP)

The properties of hybrid dispersions prepared using different methods (1a, 2 and 3 – seeSection 6.3.2) and different levels of coalescent (NMP) as well as of films and coatingsmade from these dispersions are presented in Tables 6.15 to 6.17. In all dispersions thechemical structure of the polyurethane-urea and acrylic/styrene polymer component wasthe same (see relevant Tables in Section 6.5.2). All dispersions were prepared using awater-soluble initiator.

Figure 6.22 DSC pattern for typical hybrid polyurethane-urea-acrylic/styrenedispersion synthesised in this study (MDPUR-ASD 98 prepared according to method 2

using a ‘redox’ initiator)


298


ehtni

ylnognireffid

noitisopmoclaci

mehce

maseht

fosnoisrepsid

dirbyhfo

seitreporP51.6

elbaT

)PM

N(tnecselaoc

fonoitartnecnoc

dohteM fo

sisehtnys

noitangiseD

dirbyhfo

noisrepsid

tnetnoC

foP

MN

%

snoisrepsidfo

seitreporP

sdiloS%

ytisocsiV

s-aPm

Hplacinahce

MytilibatS

niM

remono

meerF

S/M

M/A

B,%

eziselcitraP

noitubirtsidmn

egarevA

eziselcitrap

mn

ateZ

laitnetopV

m

TFFM

°C

a1-

RUP

DM

142DS

A0

3.4336

8.6-5.654

/100.0/20.0<800.0

886-425.431

2.24-0<

a1-

RUP

DM

862DS

A0.2

7.039

4. 7-1.751

tN

032-039.16

4.81-0<

a1-

RUP

DM

952M

UPa

6.11-

--

--

--

--

2-

RUP

DM

891DS

A0

3.8312

5.6-2.651

/500.0/20.0<800.0

092-033.162

6.36-0<

2-

RUP

DM

79DS

A0.3

4.1301

1.7-8.651

/20.0/7.0510.0

021-054.801

7.06-0<

3R

UPD

M843

6.36.14

0027.7-4.7

51t

N025-55

4.5110.32-

0<

3R

UPD

M342

6.111.53

tN

)gnilttes(6.7

54/71.0/1.0

46.1331-33

4.954.41-

tN

asisehtnys

gniruddelleg

noisrepsiddetset

ton-

tN

299

ylnognireffid

noitisopmoclaci

mehce

maseht

fosnoisrepsid

dirbyhmorf

edam

smlif

foseitreporP

61.6elba

T.)P

MN(

tnecselaocfo

noitartnecnoceht

ni

dohteM fo

sisehtnyS

noitangiseD

dirbyH

fonoisrepsi

D

tnetnoC

PM

Nfo

%

smlif

foseitreporP

ecnaraeppA

δ001

aPM

δ r

aPM

ε r

%

gnillewS

Tg

°CH

2OK

EM

enelyX

a1-

A-R

UPD

M42

DS0

lE,

TyltraPdekcarc

5.98.51

5120 T

G,T

seceiPG,

TseceiP

8.06-

a1-

A-R

UPD

M862

DS0.2

rB,

T2.8

8.01461

31r

B,hW

743r

B,T

891r

B,T

4.37-

a1-

A-R

UPD

Ma952

DS6.11

--

--

--

--

2-

A-R

UPD

M891

DS0

lE,

T6.5

2.11552

0E,h

W082G,

T473G,

T0.95-

2-

A-R

UPD

M79

DS0.3

lE,

T5.6

0.01012

31h

W143G,

T013G,

T6.65-

3R

UPD

M8 43

6.3l

E,M

1.415.41

71131

E,hW

891r

B,T

891r

B,T

2.57-

3R

UPD

M342

6.11l

E,M

tonmliF

nevet

Nt

Nt

N21

rB,h

W003r

B,T

152r

B,T

3.66-

asisehtnys

gniruddelleg

noisrepsidsuonitaleg

-G,elttirb

-r

B,etihw

-h

W,ettaM

-M

,citsale-l

E,tnerapsnart-

Tt

Ndetset

ton-


300


gnireffidnoitisop

moclacimehc

emas

ehtfo

snoisrepsiddirbyh

morfeda

msgnitaoc

foseitreporP

71.6elba

T)P

MN(

tnecselaocfo

noitartnecnoceht

niylno

fodohte

MsisehtnyS

fonoitangise

Dnoisrepsi

Ddirby

H

tnetnoC

PM

Nfo

%

sgnitaocfo

seitreporP

nim,e

mitgniyr

Dssendra

H)zosreP(

noisehdA

ecnatsiserreta

Wssendra

H1(

°)ssendra

H3(

°)

a142

DSA-

RUP

DM

-03

0541.0

4C2

h1

retfasraeppasid

gninetihw

a18 62

DSA-

RUP

DM

0.203

0522.0

4A1

h1

retfasraeppasid

gninetihw

a1952

DSA-

RUP

DM

6.11tn

tntn

tntn

h1

retfasraeppasid

gninetihw

2891

DSA-

RUP

DM

-53

5421.0

3B1

h1

retfasraeppasid

gninetihw

279

DSA-

RUP

DM

0.353

0621.0

4B2

h1

retfasraeppasid

gninetihw

3843

RUP

DM

6.302

5202.0

1A2

3342

RUP

DM

6.11021

56141.0

1tn

neveton

ecafrus

detsetton

-tn

sedocecnatsiser

retaw

fonoitanalpxe

narof

2.4.6noitces

eeS

301

6.5.3.3 Chemical Structure of the Polyurethane-Urea Portion of the Hybrid

(a) Effect of polyol type

In Tables 6.18 to 6.20 the properties of two hybrid dispersions prepared using the samemethod 1a (see Section 3.2) and having the same structure of the acrylic/styrene part ofthe hybrid but based on different polyols, as well as of films and coatings made fromthem are presented. Neither dispersion contained coalescent (NMP)

(b) Effect of introducing double bonds to the polyurethane-urea part of the hybrid

Properties of hybrid dispersions prepared according to the different methods (1a, 1b and2 - see Section 6.3.2) and based on the same polyol (PTMG 2000), but differing in thepresence or absence of double bonds in the polyurethane-urea part of the hybrid, as wellas of films and coatings made of them, are presented in Tables 6.21 and 6.22. Alldispersions have a similar low level (2.0-3.3%) of coalescent and have the same structureof the acrylic/styrene part of the hybrid. ‘Redox’ initiator was used in the synthesis ofdispersions according to the method 2, and in all other dispersions presented in thesetables a water-soluble initiator was applied.


302

Advances in Urethane Science and Technologyaeru-enahteruylop

ehtfo

erutcurtseht

niylno

gnireffidsnoisrepsid

dirbyhfo

seitreporP81.6

elbaT

dirbyheht

fotrap

noitangiseD

dirbyhfo

noisrepsiD

loyloP

snoisrepsidfo

seitreporP

sdiloS%

ytisocsiV

s-aPM

Hplacinahce

MytilibatS

niM

eerF,re

monom

S/M

M/A

B%

eziselcitraP

noitubirtsidmn

egarevA

elcitrapezis

mn

ateZ

topV

m

TFFM

°C

DSA-

RUP

DM 42

loidreh teyloP)0002-

GM

TP(3.43

368.6-5.6

54-0.0/20.0<

800.0/10886-42

5.4312.24-

0<

DSA-

RUP

DM

003retse

B(loidretseyloP)591

retseB

+091

5.9381

8.6-5.651

tN

791-876.571

8.26-81+

tN

detsetton

–

foerutcurts

ehtni

ylnognireffid

snoisrepsiddirbyh

morfeda

ms

mliffo

seitreporP91.6

elbaT

dirbyheht

fotrap

aeru-enahteruylop

fonoitangise

Dnoisrepsi

Ddirbyh

loyloP

smlif

foseitreporP

.raeppA

δ001

aPM

δ r

aPM

ε r

%

gnillewS

Tg

°CH

2OK

EM

enelyX

42DS

A-R

UPD

MloidrehteyloP)0002-

GM

TP(l

E,T

5.98.51

5129.0

seceip8.0

,G,

Tseceip

8.06-

003DS

A-R

UPD

Mretse

B(loi dretseyloP)591

retseB

+091

seceip,T

--

--

--

-

suonitaleg-

G,elttirb-

rB,etih

w-

hW,etta

m-

M,citsale-l

E,tnerapsnart-

T

303

ehtfo

erutcurtseht

niylno


dirbyhmorf

edam

sgnitaocfo

seitreporP02.6

elbaT

dirbyheht

fotrap

aeru-enahteruylop

dirbyhfo

noitangiseD

noisrepsiD

loyloP

sgnitaocfo

seitreporP

nim,e

mitgniyr

Dssendra

H)zosreP(

noisehdA

ecnatsiserreta

Wssendra

H1(

°)ssendra

H3(

°)

42DS

A-R

UPD

M)0002-

GM

TP(loidrehteyloP03

0541.0

4h

1retfa

sraeppasidgninetih

wC2

003DS

A-R

UPD

MloidretseyloP

)591retse

B+

091retse

B(52

523.0

2gninetih

won

B1

sedocecnatsiser

retaw

fonoitanalpxe

narof

2.4.6noitces

eeS

ehtni

sdnobelbuod

foecneserp

ehtni

ylnognireffid

snoisrepsiddirbyh

foseitreporP

12.6elba

Tdirbyh

ehtfo

trapaeru-enahteruylop

dohteM fo

sisehtnys

noitangiseD

dirbyhfo

noisrepsid

sdnobelbuo

Dni

-enahteruylopaeru

snoisrepsidfo

seitreporP

sdiloS%

ytisocsiV

s-aPm

Hplacinahce

MytilibatS

.niM

.mono

meerF

,%

S/M

M/A

B

eziselcitraP

noitubirtsidmn

egarevA

ezisel citrap

mn

ateZ

laitnetopV

m

TFFM

°C

a1-

A-R

UPD

M862

DS-

7.039

4.7 -1.751

-4.471-8.34

9.164.81-

0<

a1-

A-R

UPD

M703

DS+

3.5342

8.6-5.651

-081 -53

7.182.13-

0<

b1-

A-R

UPD

M072

DS-

2.4301

1.7-8.651

-1. 224-8.66

4.491.93-

0<

b1-

A-R

UPD

M403

DS+

4.4381

8.6-5.651

-07-03

5.237.0+

0<

2-

A-R

UPD

M89

DS-

4.138

1.7-8.651

/800.0</20.0<100.0<

561-5016.031

8.93-0<

2-

A-R

UPD

M201

DS+

6.339

2.6-9.509>

/800.0</20.0<100.0<

091-529.66

1.33-0<


304


sdnobelbuod

foecneserp

ehtni

ylnognireffid

snoisrepsiddirbyh

morfeda

ms

mliffo

seitreporP22.6

elbaT

dirbyheht

fotrap

aeru-enahteruylopeht

ni

dohteM fo

sisehtnyS

noitangiseD

dirbyH

fonoisrepsi

D

nisdnob

elbuoD

-enahteruylopaeru

smlif

foseitreporP

ecnaraep pA

δ001

aPM

δ r

aPM

ε r

%

gnillewS

Tg

°CH

2OK

EM

enelyX

a1DS

A-R

UPD

M862

-l

E,T

2.88.01

46131

lE,h

W743r

B,T

891r

B,T

4.37-

a1DS

A-R

UPD

M703

+skcar

C,T

--

--

--

1.45-

b1DS

A-R

UPD

M072

-l

E,T

2.93.31

08 10

lE,h

W743r

B,T

422r

B,T

4.37-

b1DS

A-R

UPD

M403

+l

E,T

-0. 11

3665

rB,h

W471r

B,T

151r

B,T

2.36-

2DS

A-R

UPD

M89

-l

E,T

3.56.01

58241

lE,h

W473G,

T904G,

T6.07-

2DS

A-R

UPD

M201

+l

E,T

-5. 71

5962

lE,h

WG,

T422

471l

E,T

3.15-

suonitaleg-

G,elttirb-

rB,etih

w-

hW,etta

m-

M,citsale

-lE,tnerapsnart

-T

305

elbuodfo

ecneserpeht

niylno


dirbyhmorf

edam

sgnitaocfo

seitreporP32.6

elbaT

dirbyheht

fotrap

aeru-enahteruylopeht

nisdnob

dohteM fo

sisehtnys

fonoitangise

Dnoisrepsi

Ddirby

H

nisdno

Belbuo

Daeruenahteruylop

%

sgnitaocfo

seitreporP

nim,e

mitgniyr

Dssendra

H)zosreP(

noisehdA

ecnatsiserreta

Wssendra

H1(

°)ssendra

H3(

°)

a1862

MUP-

RUP

DM

-03

0522.0

4sraeppasid

gninetihw,

A1h

1retfa

a1703

MUP-

RUP

DM

+51

0424.0

4sraeppasid

gninetihw,

C1h

1retfa

b1072

MUP-

RUP

DM

-52

5491.0

4sraeppasid

gninetihw,

C2rh

1retfa

b1403

MUP-

RUP

DM

+02

0483.0

4sraeppasid

gninetihw,

C1h

1retfa

289

MUP-

RUP

DM

-04

02180.0

1sraeppasid

gninetihw,

A1h

1retfa

2201

MUP-

RUP

DM

+52

0661.0

1sraeppasid

gninetihw,

C2h

1retfa

sedocecnatsiser

retaw

fonoitanalpxe

narof

2.4.6noitces

eeS


306


6.5.3.4 Chemical structure of the acrylic/styrene part of the hybrid

The properties of hybrid dispersion prepared according to two different methods (1aand 3 – see Section 3.2) and having the same chemical composition and differing only inthe chemical structure of the acrylic/styrene polymer component are presented in Tables6.24 to 6.26. All dispersions were synthesised using a water-soluble initiator. Dispersionsprepared according to method 1a did not contain any coalescent while dispersionsprepared according to method 3 contained 11.6% of NMP.

TEM photos of the particles of two hybrid dispersions prepared using less hydrophobicand more hydrophilic monomer prepared according to method 1a are presented in Figures6.23 and 6.24. A TEM micrograph of the particles of a hybrid dispersion preparedaccording to method 3 using more hydrophobic monomer is presented in Figure 6.25.

6.5.3.5 Crosslinking of the Acrylic/styrene Part of the Hybrid

The properties of hybrid dispersions prepared using method 1a (see Section 6.3.2), havingthe same chemical composition and differing only in the presence or absence of partialcrosslinking of the acrylic/styrene polymer component are presented in Tables 6.27 to6.29. Dispersions contained 3.3% of NMP.

Figure 6.23 Particles of hybrid polyurethane-urea-acrylic/styrene dispersion preparedaccording to method 1a using less hydrophobic monomer and water-soluble initiator

(MDPUR-ASD 22). Photograph was taken using TEM.Reproduced with permission from Professor I. A. Grickova, Lomonosov University, Moscow.

307

eneryts/cilyrcaeht

foerutcurtslaci

mehceht

niylno


dirbyhfo

seitreporP42.6

elbaT

dirbyheht

fotrap

dohteM fo

sisehtnys

noitangiseD

dirbyhfo

noisrepsidre

monoM

snoisrepsidfo

seitreporP

sdiloS%

ytisocsiV

s.aPm

Hplacinahce

MytilibatS

.niM

.mono

meerF

,%

S/M

M/A

B

eziselcitraP

noitubirtsidmn

egarevA

eziselcitrap

mn

ateZ

laitnetopV

m

TFFM

°C

a1-

RUP

DM

32DS

AA

B6.23

575.6-2.6

54-/-/3.0

054-517.39

6.72-0<

a1-

RUP

DM

22DS

AM

M5.43

074.7-1.7

54-/51.1/-

063-817.38

3.91-0<

a1-

RUP

DM

41DS

AS

0.4331

5.6-2.654

20.0/-/-525-61

3.696.62-

0<

a1-

RUP

DM

42DS

AS/

MM/

AB

46.1/23.1/13.43

368.6-5.6

54/100.0/20.0<

800.0886-42

5.4312.24-

0<

3R

UPD

M742

AB

1.5304

5.8-2.854

-/-/90.0072-04

8.373.13-

0<

3R

UPD

M542

MM

6.5344

2.854

-/29.0/-031-53

1.166.13-

tn

3R

UPD

M052

S2.5 3

636.7-3.7

0911.0/-/-

052-041.78

9.72-0<

3R

UPD

M342

S/M

M/A

B46.1/23.1/1

1.53tn

)gnilttes(6.7

5446.1/71.0/1.0

331-334.95

4.41-tn

detsetton

-tn


308


foerutcurtslaci

mehceht

niylno


dirbyhmorf

edam

smlif

foseitreporP

52.6elba

Tdirbyh

ehtfo

trapeneryts/cilyrca

eht

dohteM fo

sisehtnyS

noitangiseD

dirbyH

fonoisrepsi

Dre

monoM

smlif

foseitreporP

.raeppA

δ001

aPM

δ r

aPM

ε r

%

gnillewS

gT °C

H2O

KE

Menely

X

a1-

RUP

DM

32DS

AA

Bl

E,T

5.32.4

63131

lE,h

W743G,

T904G,

T6.75-6.97+

a1-

RUP

DM

22DS

AM

MseceiP,

Ttn

tntn

tntn

tn8.95-4.17+

a1-

RUP

DM

41DS

AS

seceiPh

Wtn

tntn

tntn

tn0.45-1.27+

a1-

RUP

DM

42DS

A

/M

M/A

B/23.1/1S

46.1l

E,T

5.98.51

5120 T

G,T

G,T

8.06-

3742

RUP

DM

AB

lE,

T3.4

7.6032

31h

W384G,

T143G,

T73.77-83.74-

3542

RUP

DM

MM

seceiP,T

tntn

tntn

tntn

6.66-0.77+

3052

RUP

DM

Sl

E,hW

4.015.41

10262

lE,h

W422r

B,T

152r

B,T

7.76-

3342

RUP

DM

/M

M/B

A/23.1/1S

46.1

lE,

Mneven

Umlif

tntn

tn21

rB,h

W013r

B,T

152r

B,T

3.66-

suonitaleg-

G,elttirb-

rB,etih

w-

hW,etta

M-

M,citsale

-lE,tnerapsnart

-T

detsetton

-tn

309

foerutcurtslaci

mehceht

niylno


dirbyhmorf

edam

smlif

foseitreporP

62.6elba

Tdirbyh

ehtfo

trapeneryts/cilyrca

eht

dohteM fo

sisehtnys

fonoitangise

Dnoisrepsi

Ddirby

Hre

monoM

sgnitaocfo

seitreporP

nim,e

mitgniyr

Dssendra

H)zosreP(

noisehdA

ecnatsiserreta

Wssendra

H1(

°)ssendra

H3(

°)

a132

DSA-

RUP

DM

AB

5305

90.01

C2

a122

DSA-

RUP

DM

MM

5305

72.04

B1

a141

DSA-

RUP

DM

S53

0521.0

1D1

h1

retfasraeppasid

gninetihw

a142

DSA-

RUP

DM

S/M

M/A

B46.1/23.1/1

0305

41.04

C2h

1retfa

sraeppasidgninetih

w

3R

UPD

M7 42

AB

5306

60.04

B2h

1retfa

sraeppasidgninetih

w

3542

RUP

DM

MM

tnskcar

Ctn

tntn

tn

3052

RUP

DM

S04

0541 ,0

4D1

h1

retfasraeppasid

gninetihw

3342

RUP

DM

S/M

M/A

B46.1/23.1/1

021561

41.01

tn

sedocecnatsiser

retaw

fonoitanalpxe

narof

2.4.6noitces

eeSdetset

ton-

tn


310


Figure 6.24 Particles of hybrid polyurethane-urea-acrylic/styrene dispersion preparedaccording to method 1a using more hydrophobic monomer and water-soluble initiator

(MDPUR-ASD 14). Photograph was taken using TEM.Reproduced with permission from Professor I. A. Grickova, Lomonosov University, Moscow.

Figure 6.25 Particles of hybrid polyurethane-urea-acrylic/styrene dispersion preparedaccording to method 3 using more hydrophobic monomer and water-soluble initiator

(MDPUR 250). Photograph was taken using TEM.Reproduced with permission from Professor I. A. Grickova, Lomonosov University, Moscow.

311

ehtfo

gniknilssorcfo

ecnesbaro

ecneserpeht

niylno


dirbyhfo

seitreporP72.6

elbaT

dirbyheht

fotrap

eneryts/cilyrca

fonoitangise

Dnoisrepsi

Ddirbyh

gniknilssorC

/cilyrcA

foeneryts

snoisrepsidfo

seitreporP

sdiloS%

ytisocsiV

s-aPM

Hplacinahce

MytilibatS

niM

eziselcitraP

noitubirtsidmn

egarevA

eziselcitrap

mn

topate

ZV

mTFF

M°C

703DS

A-R

UPD

M-

3.5342

8.6-5.651

002-537.18

2.13-0<

903DS

A-R

UPD

M+

6.5371

1.7-8.654

051-520.05

5.46-0< fo

ecnesbaro

ecneserpeht

niylno


dirbyhmorf

edam

smlif

foseitreporP

82.6elba

Tdirbyh

ehtfo

trapeneryts/cilyrca

ehtfo

gniknilssorc

noitangiseD

dirbyH

fonoisrepsi

D

fogniknilssor

Ceneryts/cilyrc

Adirbyh

fotrap

smlif

foseitreporP

.raeppA

δ001

aPM

δ r

aPM

ε r

%

gnillewS

Tg

°CH

2OK

EM

enelyX

-R

UPD

M703

DSA

-skcarc,

T-

--

--

-1.45-

-R

UPD

M903

DSA

+l

E,T

-1.11

3565

rB,h

Wt

NseceiP

031r

B,T

5.75-

suonitaleg-

G,elttirb-

rB,etih

w-

hW,tta

m-

M,citsale

-lE,tnerapsnart

-T

tN

detsetton

-


312


6.5.4 Additional Experiments

Additional experiments were performed in order to better clarify the mechanism of the processof formation of hybrid dispersions and the effect of various factors on their properties.

6.5.4.1 Investigations of the Course of Swelling of DPUR Particles with Monomers

This experiment was designed in order to further clarify the mechanism of hybriddispersion particles formation. In this experiment, monomer was continuously added toDPUR (of the composition corresponding to DPUR-237 – see Section 6.3.1) at a rate of1 ml/h. After each hour, samples of the dispersion were taken and the following parameterswere determined: particle size, particle size distribution and zeta potential. Changes ofparticle size distribution in the course of swelling are shown in Figure 6.26, and changesof average particle size and zeta potential in Figure 6.27.

dirbyhfogniknilssorcfoeergedfonoitagitsevnifostluseR92.6elbaTgnitratsehtrofdeniatbostluseR.tsetgnillewsmuirbiliuqeehtgnisusnoisrepsid

nosirapmocrofdedulcnioslaerasnoisrepsidDSAdnaRUPD

fonoitangiseDnoisrepsiddirbyh

noitangiseDgnitratsfonoisrepsid

dohteMdirbyhfosisehtnys

α xam

%

elbuloS

,S,noitcarf

%

gniknilssorC,ytisned

j

fonoitcarFevitca

sV,sniahc

22DSA-RUPDM 042RUPD a1 032 0.14 59.0 11.0

32DSA-RUPDM 442RUPD a1 013 0 ∞ 00.1

42DSA-RUPDM 442RUPD a1 071 0.68 8.2 94.0

41DSA-RUPDM 732RUPD a1 023 5.7 9.2 95.0

342RUPDM 51DSA 3 022 0.4 2.4 17.0

542RUPDM 71DSA 3 0231 5.52 3.1 52.0

742RUPDM 81DSA 3 092 0.52 8.1 52.0

052RUPDM 02DSA 3 57 5.12 5.1 92.0

- 732RUPD - 041 14 59.0 11.0

- 51DSA - 012 03 0.1 91.0

- 71DSA - D D D D

- 81DSA - D D D D

- 02DSA - D D D D

eneulotnidevlossidmlif-D

313

Figure 6.26 Changes of particle size distribution in the course of swelling of DPURparticles with monomer


After 5 h

Initial sample

After 1 h

After 2 h

After 9 h

After 3 h

314


Taking advantage of the capabilities offered by the Malvern Zeta Sizer apparatus thecomputer analysis of total volume of smaller and bigger particles (it may be noticed thaton the curves showing particle size distribution in Figure 6.26 that there are two peaksvisible). Results are:

gnilpmasfoemiT llamsfoemuloV)1kaep(selcitrap

gibfoemuloV)2kaep(selcitrap

elpmaslaitinI %001

h1retfA %89 %2

h2retfA %49 %6

h3retfA %66 %43

h4retfA %04 %06

h5retfA %05 %05

Figure 6.27 Changes of average particle size and zeta potential during addition ofmonomer to DPUR

315

6.5.4.2 Determination of the Maximum Degree of Swelling of DPUR Particles

This experiment was aimed at clarifying what is the maximum amount of monomer thatwould be able to swell polyurethane-urea, i.e., would be able to take part in the formationof hybrid dispersion particles, assuming that enough time is allowed to achieve equilibrium.Films made from DPUR 258 were used in this experiment. The following results wereobtained:

Monomer Maximum degree of swelling αmax, %

MM 88.14Styrene 88.21MM/Styrene = 1/1 89.0

The changes of the mass of polyurethane-urea film in the course of swelling with monomersis shown in Figures 6.28 to 6.30.

Figure 6.28 Change of the mass of polyurethane-urea film (made from DPUR 258) inthe course of swelling with MM


316


Figure 6.29 Change of the mass of polyurethane-urea film (made from DPUR 258) inthe course of swelling with styrene

Figure 6.30 Change of the mass of polyurethane-urea film (made from DPUR 258) inthe course of swelling with MM/S = 1/1 (w/w) mixture

317

6.5.4.4 Investigation of the Effect of the Method of Synthesis of HybridDispersions on their Degree of Crosslinking

The degree of crosslinking of the starting dispersions (DPUR and ASD) and the hybridswas determined using the equilibrium swelling test described in Section 4.3. The followingparameters were calculated based on the results of this experiment:

• Maximum degree of crosslinking, amax (%)• Content of soluble fraction, S (%)• Crosslink density (number of crosslinks per polymer molecule), j• Fraction of active chains (composed of segments linked with crosslinks at both ends), Vs

The results are shown in Table 6.29

6.5.4.5 Investigation of Hybrid Dispersion Particles Morphology

The morphology of dispersion particles was investigated using the method describedin Section 6.4.1. Examples of different morphologies of particles of hybrid dispersionssynthesised in this study according to methods 1a, 1b and 3 are presented in Figure6.31 in comparison with particles of the starting dispersion of BA/MM/S copolymer.The contrast was selected so that in pictures ‘d’ and ‘c’ white colour represents thepolyurethane-urea part of the hybrid and in picture ‘b’ the same colour represents theacrylic/styrene part of the hybrid.

The morphology of particles of hybrid dispersion synthesised according to method 2using water-soluble and ‘redox’ initiators is presented in Figures 6.32 and 6.33,respectively. Both pictures show both the single particles and the coalesced particles todemonstrate what happens to the particle morphology in the process of film formation.White colour represents the polyurethane-urea part of the hybrid.

6.5.4.6 Determination of Surface Free Energy of Films Made from HybridDispersions

The surface free energy of films made from hybrid dispersions was determined in orderto look at the values of its non-polar and polar component which could provide someinformation on whether polyurethane-urea or acrylic/styrene polymer is on the film surface(see Table 6.30).


318


Figure 6.31 Examples of different morphologies of particles of hybrid polyurethane-urea-acrylic/styrene dispersions synthesised in this study. Micrographs were taken

using TEM.

(a) ASD dispersion (BA/MM/S) (shown for comparison); (b) hybrid MDPUR-ASD dispersion (thesame monomers) prepared according to method 1b (see section 3.2); (c) hybrid MDPUR-ASDdispersion (the same monomers) prepared according to method 1a (see Section 3.2); (d) hybrid

MDPUR-ASD dispersion (the same monomers) prepared according to method 3 (see Section 3.2)Reproduced with permission from Professor A. E. Czalych, Institure of Chemical Physics of

the Russian Academy of Sciences, Moscow.

(a) (b) (c) (d)

Figure 6.32 Morphology of particles of hybrid polyurethane-urea-acrylic/styrene hybriddispersion prepared according to method 2 (See Section 6.3.2) using water-soluble initiator(MDPUR-ASD 97). Micrograph was taken using TEM. Both single particle and coalescedparticles are shown. Reproduced with permission from Professor A. E. Czalych, Institure of

Chemical Physics of the Russian Academy of Sciences, Moscow.

319

Figure 6.33 Morphology of particles of hybrid polyurethane-urea-acrylic/styrene hybriddispersion prepared according to method 2 (See Section 6.3.2) using ‘redox’ initiator

(MDPUR-ASD 98). Micrograph was taken using TEM. Both single particle and coalescedparticles are shown. Reproduced with permission from Professor A. E. Czalych, Institure of

Chemical Physics of the Russian Academy of Sciences, Moscow.

(ygreneeerfecafruS03.6elbaT γ (ralopstidna) γp (ralop-nondna) γd)otgnidroccaderaperpsnoisrepsiddirbyhmorfedamsmliffostnenopmoc

gnitratsdnaRUPDmorfedamsmlifrofdeniatbostluseR.3dohtemnosirapmocrofnwohsoslaerasnoisrepsideneryts/cilyrca

fonoitangiseDnoisrepsiddirbyh

fotrapeneryts/cilyrcAdirbyheht

γm/Jm 2

γpm/Jm 2

γdm/Jm 2

342RUPDM 46.1/23.1/1S/MM/AB 4.53 0.4 4.13

542RUPDM MM 7.43 3.3 4.13

742RUPDM AB 3.13 3.6 0.52

052RUPDM S 7.63 7.0 0.63

51DSA 46.1/23.1/1S/MM/AB 5.22 0.81 5.4

81DSA AB 0.32 8.1 2.12

852RUPD - 7.33 3.2 4.13


320


6.6 Discussion

6.6.1 Estimation of the Effect of Various Factors on the Properties of HybridDispersions and Films and Coatings Made from Them

6.6.1.1 Method of Hybrid Dispersion Synthesis and Type of Initiator

• Properties of dispersions

When the properties of hybrid dispersions having the same composition but preparedaccording to different methods using water-soluble and ‘redox’ initiator (Tables 6.9 and6.12) are compared with the properties of starting DPUR and ASD (Tables 6.1 and 6.4)it can be noted that:

a) The type of initiator does not affect the properties of hybrid dispersions significantlyexcept for free monomer content which is distinctly lower when a ‘redox’ initiator isused. This suggests that the mechanism of hybrid particle formation is similar inboth cases. This is strongly supported by the direct evidence that there is no differencein morphologies of particles of hybrid dispersions obtained according to method 2(see Section 6.3.2) using water-soluble or ‘redox’ initiator (compare Figures 6.32and 6.33).

b) Neither method of synthesis nor type of initiator have any influence on mechanicalstability, viscosity or MFFT of hybrid dispersions. However, the following observationscan be made:

• mechanical stability of all hybrid dispersions is slightly lower than that of thestarting dispersions, but nevertheless it is good enough for successful applicationof hybrid dispersions in practice (there was no settling of dispersions duringstorage). Relatively high negative values of zeta potential observed for all hybriddispersions confirm that their stability should be good. For all hybrid dispersionsMFFT is < 0 °C (just as for starting DPUR) which should undoubtedly beconsidered as their great advantage when compared to ADS.

• the viscosity of all hybrid dispersions is quite low and similar to the starting DPUR

a) The average particle size of hybrid dispersions depends strongly on the method ofsynthesis (regardless of the initiator). The lowest values of average particle size areobserved for MDPUR-ADS prepared according to method 1a, i.e., with continuousfeeding of monomers to DPUR, and the highest for MDPUR prepared according tomethod 3, i.e., emulsifying of prepolymer-ionomer in ASD followed by crosslinking

321

with polyamine. This result is in good correlation with the assumptions concerningthe mechanism of hybrid particles formation presented in Section 6.3.2. If thedispersion is prepared according to method 1a, smaller ASD particles can be formedalong with bigger hybrid particles which may have various morphologies. If thedispersion is prepared according to method 3, only bigger ‘core-shell’ particles canbe formed. Detailed comparison of the size of particles of starting DPUR (DPUR346) and hybrid dispersions prepared from it according to method 1a (MDPUR-ASD 268 and MDPUR-ASD 274) confirms this assumption that the average particlesize for starting DPUR is higher than for hybrid dispersions prepared from it.

b) The particle size distribution observed for hybrid dispersions is the lowest for MDPUR-ASD prepared according to method 2, i.e., diluting prepolymer-ionomer with monomers,emulsifying this solution in water, crosslinking with polyamine and polymerisation ofmonomers. This can be explained by the fact that in this method, formation of hybriddispersion particles proceeds ‘in situ’, i.e., both prepolymer-ionomer (precursor ofpolyurethane-urea component of the hybrid) and the monomers (precursors of acrylic/styrene component of the hybrid) are emulsified at the same time, which means thatthe chance of formation of hybrid dispersion particles having similar morphology andsize is very high. Higher uniformity of the whole system in the case when the hybriddispersions are prepared according to method 2 is also confirmed by the higher negativevalues of zeta potential observed for these dispersions.

• Properties of films

From a practical point of view it is very important that, regardless of the type of initiatorand method of synthesis, the films made from hybrid dispersions are transparent (thisproves that their structure is quite uniform), strong, elastic and resistant to water (verygood resistance) and organic solvents (good resistance).

When the properties of films made from hybrid dispersions having the same compositionbut prepared according to different methods using water-soluble and ‘redox’ initiator(Tables 6.10 and 6.13) are compared with the properties of films made from the startingDPUR (Table 6.2) it can be noted that:

a) Type of initiator practically does not have any influence on mechanical properties,water resistance or solvent resistance of films made from hybrid dispersions. Slightlyhigher values of elongation at break observed for films made from hybrid dispersionsprepared using ‘redox’ initiator (regardless of the method of synthesis) can resultfrom the fact that less grafting (or crosslinking) proceeds in the course ofpolymerisation.


322


b) The effect of method of synthesis on the mechanical properties, water resistance andsolvent resistance of the films can be clearly seen when the results obtained for thefilms made from hybrid dispersions prepared according to method 3 and accordingto other methods are compared. In the cases where method 3 has been applied, thefilms are distinctly stronger, less elastic and better resistant to water and organicsolvents. It is also interesting that their Tgs are lower (closer to the one observed forDPUR). The explanation of this phenomenon is not clear since taking into accountthe assumed hybrid particles formation mechanism presented in Section 6.3.2 onecould rather expect that not only Tg, but also the mechanical properties of the filmsmade from hybrid dispersions prepared according to method 3 would resemble thoseof films made from DPUR (high elasticity and water/solvent resistance) more thanthe properties of the films made of dispersions prepared according to the othermethods. Nevertheless, when this problem is more deeply analysed, it can beanticipated that the reason for this phenomenon is the very high (much higher thanin dispersions prepared according to the other methods) number of particles of ‘purecore-shell’ structure in dispersions prepared according to method 3. In the film madeof hybrid dispersion prepared according to this method, a very regular heterogeneousstructure is formed during coalescence of particles. In such a structure, acrylic/styrenepolymer (which constitutes the core in dispersion particles) is the dispersed phasewhile polyurethane-urea (which constitutes the shell in dispersion particles) is thematrix. Acrylic/styrene polymer then plays the role of toughener while polyurethane-urea increases the water and solvent resistance of the film and provides its low Tg.The second Tg corresponding to acrylic/styrene polymer, should also be detected inthis particular system, but DSC does not have the resolution which would be neededhere. However, when the torsion pendulum method was applied we were able tospot this second Tg quite easily. We believe that this particular hybrid system is veryinteresting and requires further studies to explain the mechanisms of both hybridparticles and film formation.

• Properties of coatings

Regardless of the initiator and method of synthesis, all coatings made from hybriddispersions show very short drying time, reasonable hardness and water resistance. Basedon the results presented in Tables 6.11 and 6.14 it can be concluded that the best propertiesof coatings have been achieved with dispersions prepared according to method 3, i.e., byemulsifying the prepolymer-ionomer in ASD, and the worst – for dispersions preparedaccording to method 2, i.e., by diluting the prepolymer-ionomer with monomers. Theproperties of coatings made from hybrid dispersions are generally much better than thoseof coatings made from starting DPUR (starting ASD dispersions did not form coatingssince they did not contain any coalescent).

323

6.6.1.2 Concentration of Coalescent in Hybrid Dispersions


As can be seen in Table 6.15, concentration of coalescent has a significant effect on theproperties of hybrid dispersions, regardless of the method of synthesis. Very high levelsof coalescent lead to gelling in the course of polymerisation (see MDPUR-ASD preparedaccording to method 1a) or partial settling of dispersion during storage (see MDPUR243 prepared according to method 3). It can also be noted that for dispersions preparedwithout coalescent according to methods 1 and 2, average particle size is distinctly higherand particle size distribution broader than when the process is carried out in the presenceof coalescent. This phenomenon can be explained by the positive influence of smallamounts of coalescent on the process of hybrid particle formation in the course ofpolymerisation. It can be assumed that the particles of starting DPUR are partly swelledby coalescent which helps the monomers to diffuse into them. When hybrid dispersionshave been prepared according to method 3, the reverse effect is observed. Here, increasingthe coalescent concentration results in diminishing both the average particle size andparticle size distribution range. The reason for this may be easier formation of separateDPUR particles at lower levels of coalescent.


The effect of coalescent on the properties of films made from hybrid dispersions seems tobe not quite significant (see Table 6.16) which is strange when taking into account theimportant role of coalescent in the process of film formation. The only observation thatcan be made here is that, as has already been stated in Section 6.6.1.2.1, high levels ofcoalescent cannot be recommended in preparation of hybrid dispersions.


Here, the effect of coalescent is clearly visible (see Table 6.17), but only if the coalescentcontent is high (MDPUR 243). Then, drying time increases and coating hardness decreases.However, no differences can be observed between the properties of coatings made fromhybrid dispersions which do not contain any coalescent and those which contain onlysmall amounts of coalescent.


324


6.6.1.3 Chemical Structure of the Polyurethane-Urea Component of the Hybrid

6.6.1.3a Substituting Polyetherdiol for Polyesterdiol


As can be seen from Table 6.18, the chemical structure of the polyurethane-urea part ofthe hybrid does not affect the properties of hybrid dispersions, obviously not countingthe MFFT which is much higher if polyesterdiol has been used as a starting material forthe prepolymer-ionomer synthesis (compare MDPUR-ASD 300 and MDPUR-ASD 24).


Here, the effect of chemical structure of the polyurethane-urea part of the hybrid is, ofcourse, substantial (see Table 6.19). For hybrid dispersion synthesised without coalescentusing polyesterdiol as a starting material for the prepolymer-ionomer, the Tg is so high thatfilms cannot be obtained. On the other hand, if polyetherdiol is applied as a starting materialin the synthesis of the same dispersion, films of very good mechanical properties are obtained.This was the reason for using polyetherdiol rather than polyesterdiol as the starting materialfor synthesis of dispersions in this study.


It can be clearly seen from Table 6.20 that the hybrid dispersion prepared using polyesterdiol(MDPUR-ASD 300) forms faster-drying coatings of better hardness, adhesion and waterresistance. This information may be important from the practical point of view.

6.6.1.3b Introducing Double Bonds to Polyurethane-Urea


As can be seen from Table 6.21, introducing double bonds into the polyurethane-ureacomponent of the hybrid does not affect the properties of hybrid dispersions regardlessof the method used for their synthesis.


Based on the data contained in Table 6.22 it can be concluded that the effect of introducingdouble bonds to polyurethane-urea on the properties of films made from hybrid dispersions

325

is exceptionally strong. Regardless of the method of hybrid dispersion synthesis, the presenceof double bonds in polyurethane-urea results in a distinct increase in Tg and a drasticincrease in film stiffness (decrease in film elasticity), and, an increase in resistance to waterand organic solvents. This may mean that grafting or even crosslinking takes place duringpolymerisation and/or crosslinking via oxidation of double bonds suggested earlier [29],proceeds during air-drying of the films.


The presence of double bonds in polyurethane-urea has very significant effects on theproperties of coatings made from hybrid dispersions (see Table 6.23). It leads to drastic(two-fold) increase in coating hardness and to distinct shortening of drying time. However,there is no effect of double bonds in polyurethane-urea on adhesion or water resistanceof the coatings.

Generally, based on the results shown above, it can be concluded that introducing doublebonds into polyurethane-urea is undoubtedly one of the ways to obtain coatings ofexcellent properties based on hybrid polyurethane-urea-acrylic/styrene dispersions.

6.6.1.4 Chemical Structure of the Acrylic/styrene Part of the Hybrid


Table 6.24 shows that the chemical structure of the acrylic/styrene part of the hybrid hasno effect on the macroscopic properties of hybrid dispersions, although the appearanceof the dispersion particles may be quite different (compare Figures 6.23 and 6.24).Difference in the appearance of particles may result from the degree of hydrophobicityof the monomer (this effect is explained in Section 6.3.2.1).


The chemical structure of the acrylic/styrene part of the hybrid has a substantial effect onthe properties of films made from hybrid dispersions (see Table 6.24). When the monomersthat form polymers of high Tg (styrene or MM) are used and the synthesis is carried outaccording to method 1a, no film is obtained. For the same monomers but a different methodof dispersion synthesis (method 3), a film was obtained only for styrene and only when ahigh level of coalescent was applied. In this case, the mechanical properties, water andsolvent resistance of the film were quite good, but the film was not transparent, which


326


indicated that the uniformity of the system, i.e., the compatibility of polystyrene withpolyurethane-urea was rather poor. Based on the results of investigations of the effect ofstyrene on the morphology of the model polyurethane-urea-acrylic/styrene hybrid systems[41] this assumption can be proved.

When BA or a mixture of monomers (BA/MM/S) was used for synthesis of hybriddispersions, transparent films of very good mechanical properties and water resistancewere obtained (the best properties were obtained when the dispersions were preparedaccording to method 3). Solvent resistance was good only for films made from thedispersion prepared according to method 3.


Using data presented in Table 6.26 it can be concluded that the type of monomer used inthe synthesis of hybrid dispersions has no effect on the drying time of coatings madefrom these dispersions, but has a significant effect on coating hardness. The highesthardness can be achieved with coatings made from dispersions prepared using MM as astarting monomer, if it is possible to make a coating (compare MDPUR-ASD 22 andMDPUR 245 in Table 6.26). It is noteworthy that it was not possible to make anycoating from MDPUR 245 (prepared according to method 3), despite the fact that itcontained a high level of coalescent.

6.6.1.5 Crosslinking of the the Acrylic/styrene Part of the Hybrid

Table 6.27 shows that the properties of hybrid dispersions prepared with additionalcrosslinking of acrylic/styrene polymer do not differ from the properties of dispersionsof the same composition prepared without additional crosslinking. Also the propertiesof films made from both dispersions were very similar (see Table 6.28). However, lack ofpositive effects of additional crosslinking of the acrylic/styrene part of the hybrid mayresult from improper selection of starting materials for hybrid synthesis. Therefore, weintend to investigate this effect further in another study along with further studies on theeffect of introducing double bonds to the polyurethane-urea part of the hybrid.

6.6.2 Mechanism of Hybrid Particle Formation

Based on the results of supplementary investigations (see Section 6.5.4) that were carriedout to better understand the mechanism of hybrid particle formation and to determineparticle morphology, the following observations can be made:

327

(a) Investigations of the kinetics of the process of swelling the DPUR particles withmonomers (see Section 6.5.4.1) have shown that the number of ‘bigger’ particles(particles swelled with monomer and monomer droplets) increases gradually whilethe number of ‘smaller’ particles (DPUR particles not swelled with monomers)decreases. After some time (about 5 hours), equilibrium is reached and then abimodal particle size distribution occurs. This direct observation of the process ofswelling DPUR particles with monomers is very interesting because it does notcorrelate with the course of swelling the film with monomers (compare with Section6.5.4.2). In the process of swelling the film with monomers, equilibrium is reachedafter a much longer time (12 hours) regardless of the monomer used. This happensat a much higher degree of swelling (approximately 90%) as compared to the processof swelling DPUR particles (maximum 40% assuming that all monomer diffusedinto the DPUR particles). In practice, a significant amount of the monomer remainsin the form of monomer droplets suspended in DPUR.

It is also interesting to note that during the feeding of the monomers, the zetapotential becomes less negative which indicates that the uniformity and stability ofthe whole system is diminishing.

Investigations of swelling of DPUR particles with monomers indicate that althoughall monomer being fed into the system may theoretically enter DPUR particles,since the maximum degree of swelling is more than twice that of the monomer/DPUR solids w/w ratio used commercially in synthesis of hybrid dispersions, inpractice there will always be a natural tendency of the system to reach equilibriumbetween monomer droplets, non-swollen DPUR particles and swollen DPURparticles. Therefore, when hybrid dispersions are prepared according to method1a, 1b or 2 (see Section 6.3.2) the result will always be a mixture of particles ofhybrid structure with DPUR and ASD particles, as has been explained in Section6.3.2.1).

(b) The results of investigations of crosslink density of films made from hybriddispersions (see Section 6.5.4.2) show that when the values of the crosslink densityof the films determined by swelling in toluene are compared it can be concludedthat regardless of the method used for synthesis of hybrid dispersions the films areall crosslinked, or at least they behave as if they were crosslinked. This may meaneither that the polymers in the films are actually crosslinked or very strongly branched(by the grafting of monomers on polyurethane-urea) and/or that there is strongdomination of ‘core-shell’ structure with a ‘shell’ made of polyurethane-urea. Thelatter possibility is supported by the properties of films made from hybrid dispersionsprepared according to method 3 (emulsifying of prepolymer-ionomer in ASD andthen crosslinking with polyamine). It may also be additionally supported by:


328


• the experimentally proven existence of an outer polyurethane-urea shell in theparticles of hybrid dispersions prepared not only according to method 3 (whichseems quite obvious), but also according to method 2 (see Figures 6.32and 6.33).

• the values of the polar and non-polar (dispersed) components of surface freeenergy determined for the films made from hybrid dispersions being very similarto the values found for the films made from DPUR, not ASD (see Table 6.30).This may be considered as evidence for ‘hiding’ of the acrylic/styrene part of thehybrid under the film surface, which results presumably from the ‘core-shell’structure of the coalesced dispersion particles that form the film, wherepolyurethane-urea constitutes the shell and acrylic/styrene polymer the core.

(c) The morphology of hybrid dispersion particles shown in Figure 6.31 looks quiteinteresting. It can be clearly seen that regardless of the method of synthesis, resultingin particles of different morphologies, they all have ‘core-shell’-like structures anddiffer substantially from the particles of ASD, which have a uniform structure.Deeper analysis of these structures leads to the conclusion that for MDPUR-ASDprepared according to Method 1b, the particle morphology resembles that describedas ‘engulfed’ [15] for other types of hybrid dispersions (see Figure 6.1). For ourhybrid dispersion, small round-shaped particles of acrylic/styrene polymer are‘engulfed’ or ‘englued’ in the bulky core made of polyurethane-urea.

The most interesting result seems to be, however, the specific morphology of theparticles of hybrid dispersions prepared according to method 2 shown in Figures6.32 and 6.33. It is shown schematically in Figure 6.34.

Here, the acrylic/styrene polymer constitutes an under-surface sphere ‘embedded’ ina particle made of polyurethane-urea. This sphere is approximately 25-40 nm thickand is situated approximately 15-20 nm from the particle surface. After coalescenceof the particles this specific structure is retained (see coalesced particles in Figure6.32) so that the film made from a dispersion of this particular particle morphologyshould have the structure presented schematically in Figure 6.35.

This kind of structure is undoubtedly very interesting since it is a ‘nanostructure’where ‘nanospheres’ of acrylic/styrene polymer are embedded in a polyurethane-urea matrix. Such a structure suggests that very specific properties of the film couldbe obtained if the components of the hybrid were especially selected.

329

Figure 6.34 Schematically presented ‘embedded sphere’ morphology of the particles ofhybrid dispersions prepared according to method 2 (diluting the prepolymer-ionomer with

monomers, emulsifying it in water, crosslinking with polyamine and polymerisation)

Figure 6.35 Schematically presented structure of film made from a hybrid dispersionof the particle morphology shown in Figure 6.34


330


6.7 Summary

In this study, aqueous hybrid polyurethane-urea-acrylic/styrene polymer dispersions wereobtained using four different methods. The dispersions are stable and their viscosity, andpH as well as average particle size and particle size distribution are similar to thoseobserved for DPUR. Most of the hybrid dispersions formed transparent films of goodmechanical properties, water resistance and organic solvent resistance.

Detailed investigations of coating properties of hybrid dispersions proved that they couldform transparent coatings of reasonable hardness, adhesion and water resistance.

The effect of various factors on the properties of hybrid dispersions as well as offilms and coatings made from them was determined. The results are summarised inTable 6.31.

Based on the data in Table 6.31 the following dependencies may be concluded:

• the properties of hybrid dispersions: dependent on method of hybrid synthesis andchemical structure of the polyurethane-urea part of the hybrid

• the properties of films made from hybrid dispersions: dependent on method of hybridsynthesis, chemical structure of the polyurethane-urea part of the hybrid, presence ofdouble bonds in the polyurethane-urea and the chemical structure of the acrylic/styrene part of the hybrid

• the properties of coatings made from hybrid dispersions: dependent on chemical structureof the polyurethane-urea part of the hybrid and presence of double bonds in thepolyurethane-urea

The morphology of dispersed particles revealed by TEM appeared to be veryinteresting. It was found that the method of hybrid synthesis had a substantial influenceon the particle morphology and that usually ‘core-shell’ or ‘englued’ morphologiesdescribed earlier in the literature were observed. However, in one case the unusual‘embedded sphere’ morphology was seen. In these particles, the core made ofpolyurethane-urea is surrounded by a 25-40 nm thick sphere made of acrylic/styrenepolymer and covered by a 15-20 nm thick outer layer of polyurethane-urea. Thestructure of film made from such dispersions is very interesting since it is a two-phase structure where ‘nanospheres’ of acrylic/styrene polymer are suspended in apolyurethane-urea matrix.

The supplementary experiments, and especially investigations of the process of swellingof dispersion particles with monomers and film surface free energy determinations, allowedconfirmation of the assumed mechanism of hybrid particle formation.

331

remylop

eneryts/cilyrca-aeru-enahteruylopdirbyh

foseitreporp

ehtno

srotcafsuoirav

fotceff

E13.6

elbaT

ydutssiht

nidesisehtnys

snoisrepsidsnoisrepsid

foseitreporP

smlif

foseitreporP

sgnitaocfo

seitreporP

ytilibatSytisocsi

VHp

eziselcitraP

ezisdna

noitubirtsidTFF

Mlacinahce

Mseitreporp

retaW

dnatnevlosecnatsiser

Tg

gniyrD

retaW

ecnatsiserssendra

Hnoisehd

A

dirbyhfo

dohteM

sisehtnys-

++-

++

-+

++

+

rotaitinifo

epyT

--

--

--

--

--

tnetnoctnecselao

C-

+-

--

-+

-+

-

erutcurtslacimeh

Caeru-enahteruylop

fodirbyh

fotrap

--

++++

++++

++

+++

elbuodfo

ecneserPni

sdnobaeru-enahteruylop

dirbyhfo

trap

--

-++

++++

++-

++-

erutcurtslacimeh

Ceneryts/cilyrca

fodirbyh

fotrap

--

-++

+++

-+

++

fogniknilssor

Ceneryts/cilyrca

remylop

--

--

--

--

--

devresbosa

wtceffe

emos:+

devresbosa

wtceffe

on:-devresbo

saw

tceffetcnitsid:++


332


Acknowledgements

The authors wish to thank Professor I. A. Grickova (Lomonosov University, Moscow)and Professor A. E. Czalych (Institute of Chemical Physics, Russian Academy of Sciences,Moscow) for valuable consultations. The assistance of our co-workers from the IndustrialChemistry Research Institute in carrying out the experiments and testing is also gratefullyacknowledged. This study was carried out under project 3TO9811410 sponsored by thePolish National Committee for Scientific Research.

References

1. R. R. Roester and R. W. Rimmer, Paint and Coatings Industry, 1991, 7, 3, 22.

2. J. Kozakiewicz in Adhesion 15, Ed., K. W. Allen, Elsevier Applied SciencePublishers, London, 1991, Chapter 6.

3. J. Kozakiewicz, Aqueous Polyurethane Dispersions as Coating Binders, presentedat Ekolak Symposium, Ustron, Poland, 1993.

4. J. W. Rosthauser and K. Nachtkamp, Journal of Coated Fabrics, 1986, 16,July, 39.

5. J. Goldsmith, Journal of Coated Fabrics, 1988, 18, July, 12.

6. Waterborne PU Dispersions, ICI leaflet.

7. H. C. Goos and G. C. Overbeek, inventors; Polyvinyl Chemie Holland BV,assignee; EP 0308115A2, 1989.

8. M. Hartung, M. Grabbe and P. Mayenfels, inventors; BASF Lacke + Farben,German Patent DE 4 010 176 A1, 1991.

9. M. Guagliardo, inventor; Inmont Corporation, assignee; US 4318833, 1982.

10. J. T. Huybrechts, inventor; E.I. DuPont de Nemours and Company, assignee; US4888383, 1989.

11. R. G. Coogan and R. Vartan-Boghossian, inventors; ICI Americas, Inc., assignee;US 4927876, 1990.

12. P. Penczek in Polymer Blends, Eds., E. Maruscelli, R. Palumbo and M.Kryszewski, Plenum Press, New York, 1986, p.333.

333

13. S. Muroi, H. Hashimoto and K. Hosoi, Journal of Polymer Science: PolymerChemistry, 1984, 22, 6, 1365.

14. J. Kozakiewicz, Presented at the 7th European Polymer Federation Symposium onPolymeric Materials, Szczecin, 1998, p.94.

15 D. E. Mestach and F. Loos, Presented at the 23rd FATiPEC Congress, Brussels,1996, Paper No.B-91.

16 N. Subramaniam, M. J. Monteiro, J. R. Taylor, A. Simpson-Gomes and R. G.Gilbert, Presented at the 7th European Polymer Federation Symposium onPolymeric Materials, Szczecin, 1998, p.53.

17 J. Kozakiewicz, Hybrid Dispersion Systems as Coating Binders, Presented at theSeminar of Polifarb Cieszyn-Wroccaw, 9.05.1998, Ustron-Zawodzie.

18 U. Desor, B. Lean and R. Kuropka, Presented at the 25th FATiPEC Congress,Torino Lingotto, 2000, Volume 2, p.117.

19 Z. Florianczyk and S. Penczek, Polymer Chemistry, Warsaw Technical UniversityPress, Warsaw, 1995.

20 M. Schwartz and H. Kossmann, European Coatings Journal, 1998, 3, 134.

21 J. W. Rosthauser and K. Nachthamp, Advances in Urethanes Science andTechnology, 1987, 10, 121.

22 J. Kozakiewicz, Presented at the 2nd Symposium on Aqueous Dispersions andSolutions of Polymers, Ustron, Poland, 1993, p.27.

23 D. Dieterich, Angewandte Makromolekulare Chemie, 1981, 98, 133.

24 R. E. Tirpak and P. H. Markusch, Journal of Coatings Technology, 1986, 58,738, 49.

25 I. Bechara, Journal of Coatings Technology, 1998, 70, 236.

26. P. Loewrigkeit and K. A. Van Dyk, inventors; Witco Corporation, assignee; EP0189945B1, 1990.

27. B. Gruber, Presented at the ACS Meeting, San Francisco, CA, 1992, PaperNo.108.


334


28. C. R. Hegedus and K. A. Kloiber, Journal of Coatings Technology, 1996, 68,860, 39.

29. R. Tennebroek, J. M. Geurts, A. Overbeek and A. Hermsen, European CoatingsJournal, 1997, 11, 1016.

30. J. M. Geurts, R. Tennenbroek, A. Overbeek and Y. W. Smak, Presented at the23rd FATiPEC Congress, Brussels, Belgium, 1996, Paper No.D-259.

31. H. R. Lucas, W. E. Mealmaker and N. Giannopous, Progress in OrganicCoatings, 1996, 27, 1-4, 133.

32. M. Locatelli, V. Alonzo, S. Giordano, M. Piccinini and E. Sironi, Presented at the25th FATiPEC Congress, Torino Lingotto, 2000, Volume 2, p.135.

33. G. C. Cornelius, Y. W. Smak and E. Martin, inventors; Zeneca Resins BV,assignee; WO Patent 9916805A1, 1999.

34. A. E. Czalych, unpublished results

35. ISO 1522 Paints and Varnishes - Pendulum Damping Test, 1998.

36. ISO 2812-1 Paints and Varnishes - Determination of Resistance to Liquids - Part1 - General Methods, 1993.

37. DIN 53150 Testing of Paints, Varnishes and Similar Coating Materials;Determination of the Drying stage of Coatings (Modified Bandow WolffMethod), 1995

38. ISO 527-1, Plastics – Determination of Tensile properties – General Principles,1993.

39. ASTM D3418-99, Standard Test Method for Transition Temperatures ofPolymers by Differential Scanning Calorimetry.

40. G. C. Overbeek, Y. W. Smak and E. Martin, inventors; Avecia BV, assignee;EP1015507A1, 1999.

41. I. Legocka, J. Kozakiewicz and A. Czalych, unpublished results.

335

7 Adhesion Behaviour of Urethanes

Ashok Sengupta and H.P. Schreiber

7.1 Introduction

Two-component polyurethanes (PU) are applied in many areas of advanced technology.For example PU are used as architectural/maintenance coatings or as adhesives for avariety of substrates including glass and polymers such as polyvinyl chloride (PVC),acrylonitrile-butadiene-styrene (ABS) and polymer blends, e.g., XenoyTM. Twocomponent PUs are created by the reaction of a polyisocyanate and a polyol or co-reactant. The strength and permanence of bonds forged by the adhesive is a vital part ofthe system’s performance. A comprehensive understanding of factors influencing thepolymer/PU bond is therefore important.

The adhesion behaviour of urethanes depends not only on their composition but also onthe contribution of physico-chemical factors, in particular on the surface properties ofurethanes. Two specific aspects of surface behaviour are considered in this chapter, whichis arranged in three sections: the first concerns the ability of some polymers, includingPU to adopt a variety of surface structures or orientations, as dictated by preparationconditions and the medium in contact with the urethane polymer. The ability to undergosurface restructuring is a factor in the well documented difference between surface andbulk properties of all polymers. A second aspect of surface behavior, exerting majorinfluence on the adhesion of urethanes to a variety of plastics or glass, is the acid-baseinteraction between the adhesive polymer and the substrate. Acid-base interactions are asubject of discussion in the second section of this chapter. Finally, the role of silanemodifiers in affecting the adhesion behavior of defined urethane adhesive formulationsis considered.

7.2 Surface Characteristics of PU Adhesive Formulations

This study involves surface analyses on formulations typical of PU adhesives. The objectiveis to examine the surface properties of such polymers; to investigate the possibility ofsurface changes occurring when the adhesive, cured to different degrees, is contactedwith orienting fluids notably water; and to indicate the degree to which surfacerestructuring may influence the bonding characteristics of PU adhesives.

336


7.2.1 Experimental

7.2.1.1 Materials

The major portion of this work used as reference a two-part soft-segment PU formulation.A prepolymer of methylene bis-4-cyclohexylisocyanate (MDI) with a large amount of2,4-isomer constituted Part A. Part B was a mixture of two tri-functional polyols withmolecular weights of 900 and 500. These were used in weight ratios of 70/30. Part B alsocontained about 0.08 wt% of a tin catalyst (UL-28, Witco Chemicals).

The formulations were cured to various degrees. In one cure procedure, the polymer wasfirst hardened at 100 oC for 1 hour and then exposed to ambient conditions for fivedays. In a number of cases, this was followed by further curing at 50 °C/50% relativehumidity. In an alternative procedure, the initial 100 °C/1 hour exposure was followedby conditioning for varying lengths of time at room temperature and 50% RH. Thesealternative post-cure processes were followed to produce a range of crosslink levels andto vary also the degree of enolisation occurring during cure.

7.2.1.2 Surface Analyses

In order to produce PU films, parts A and B were mixed to give ratios of NCO:OH of1.03, and in some cases, 1.13. These mixtures were cast onto previously degreasedglass microscope slides and drawn down with doctor blades to give final polymercoating thicknesses of 1.2 mm. Samples were cured by the regimes outlined previously.Surface energies were obtained from experimental determinations of contact anglesfor sessile drops, using a Rame-Hart goniometer. All experiments were conducted at30 °C. Data were obtained for wetting by dispersion force fluids, represented by aseries of n-alkanes, and by formamide, glycerol, ethylene glycol, and water (liquidswith appreciable non-dispersive forces). In all cases, the dead space around the enclosedsample was in equilibrium with the vapour of the contacting liquid, and contact anglesat zero contact time were obtained by extrapolation of data at finite times of contact.This avoided the risk of changes due to any possible liquid-substrate interactions. Theprocedures of Gent and Schultz [1] were used for determinations of dispersive andnon-dispersive contributions to the overall surface energies, γs

d and γsnd. These were

obtained with experimental uncertainties of less than 6%.

Following initial surface energy evaluations, the PU films were immersed in water atcontrolled temperatures and for controlled times. Following immersion, films weredried under vacuum (room temperature, 24 hours) and surface energies wereredetermined.

337

Adhesion Behaviour of Urethanes

Analyses of chemical structure in cured films, before and after immersion, were carriedout by IR spectroscopy, utilising a Bonem-Michelson-100 FTIR instrument.

Another aspect of surface analysis involved the use of inverse gas chromatographic (IGC)techniques [2]. Specifically, IGC was used to evaluate the acid-base interaction potentialof PU surfaces and of substrates against which the adhesive was to be bonded. In theIGC determination, the polymers are the non-volatile phase, while the volatile phaseincluded n-alkane vapours and vapours defined as acids and bases by contemporarytheories. Such as the theory of Gutmann [3], applied in the present case along withdevelopments reported separately by Saint Flour and Papirer [4] and by Deng and Schreiber[5]. In this protocol [5], chloroform represents a pure acid vapour; tetrahydrofuran anddiethylether are strong bases; while acetone may be used as an amphoteric vapour. TheIGC procedure, described in detail elsewhere [2, 4, 5], yields interaction parameters forthe test solids. The parameter acceptor number (AN) is a measure of surface acidity; thedonor number (DN), evaluates the base interaction potential; and an averaged interactionpotential is given by

Ki = DN – AN. (1)

Positive values of Ki designate a surface which is predominantly basic. Predominantsurface acidity will produce a negative Ki. Generally, polymers show finite values of bothAN and DN, with different groups of the surface region acting as electron donors andacceptors. A surface is considered to be amphoteric when AN and DN are finite and ofsimilar values, with Ki ~ 0. When both AN and DN are zero, the surface is considered tobe neutral.

For IGC work with PU, the polymer was cured against Teflon sheet, peeled off, andpowdered by pestle and mortar. About 1.0 g of powder was packed into previouslywashed and dried steel columns, 15 cm long and 1.2 cm in diameter. Polymer substrates,such as PVC and ABS, were deposited onto Chromosorb support (acid-washed, 60/40mesh diatomaceous earth) following well-established procedures [6]. Small, nanoliteramounts of the volatile phase were injected into the chromatograph, with helium as thecarrier gas at a flow rate of 10 cm3/min.

7.2.1.3 Adhesion

Bond strengths of PU adhesives were obtained from single lap-shear determinations.Substrate/adhesive/substrate joints were made with ABS and PVC coupons, 8 x 2.5 cm,as substrates; the overlap area was 4.5 cm2. The reference PU here was compoundedwith glass beads to provide adhesive layers with uniform thicknesses of 1.27 mm. Lap

338


shear data were obtained with an Instron tester at 25 mm/min jaw separation speed. Atleast three, and in most cases five to eight, separate determinations were carried out foreach of the systems, with averaged results presented here. The experimental uncertaintyin these determinations ranged from 8 to about 20%.

7.2.2 Results and Discussion

7.2.2.1.Surface Analyses

Throughout this section, reference will be made to PU specimens cured to varyingdegrees. For convenience, cure processes will be coded as indicated by the entries inTable 7.1.

serudecorperucUPfoyrammuS1.7elbaT

edoCeruC erudecorP

0C 001,h5.0 °C1C 001,h1 °C2C 001,h1 ° TR,syad5+C

3C 001,h1 ° 05,yad1+TR,syad5+C ° HR%05,C

4C 05,syad2+2C ° HR%05,C

5C 05,syad5+2C ° HR%05,C

6C 05,syad01+2C ° HR%05,C

7C 05,syad41+2C ° HR%05,C

8C HR%05,TR,yad1+1C

9C HR%05,TR,syad2+1C




erutarepmetmooR=TRytidimuhevitaler=HR

folanruoJ,rebierhcS.P.HdnaatpugneS.AmorfnoissimrephtiwdecudorpeR.1elbaT,749,11,5,1991,ygolonhceTdnaecneicSnoisehdA © .VBPSV,1991

339


The capability of (partially) cured PU surfaces to restructure under the influence oforienting fluids is illustrated in Figure 7.1. Here, the polymer cross-linked by procedureC2 initially shows a non-dispersion force contribution of about 2mJ/m2 to a measuredtotal surface energy of about 38 mJ/m2. If the surface and bulk of the polymer areidentical in composition, γnd is lower than an expected value near 98.5 mJ/m2. Thereference polymer has a 41 wt% contribution from the polar hard segment, leadingto the conclusion that as initially prepared, the polymer surface has an excessconcentration of the soft-segment moiety. Essentially, this is a hydrocarbon with anexpected surface energy in the 35 mJ/m2 range. This finding confirms those of Chenand Ruckenstein [7] and further substantiates results from earlier studies [5]. Asshown in [5], a marked increase in γnd takes place on contact with water. The non-dispersive component rises toward a plateau value of about 8.5 mJ/m2, again in

Figure 7.1 Time-dependent increase of non-dispersive surface energy for polyurethaneimmersed in water. Polymer cure code C2 (Table 7.1). Immersion at (▲) 38.5 °C,

(●) 60 °C, and (●) 81 °C.

Reproduced with permission from A. Sengupta and H. P. Schreiber, Journal ofAdhesion Science and Technology, 1991, 5, 11, 947, Figure 1. ©1991 VSP BV

340


agreement with data in reference [5], and the rate of change is strongly temperature-dependent. Activation energies for an assumed diffusion-dependent process may becalculated from the results in Figure 7.1. The value obtained, 11.4 kcal/mol, is inexcellent agreement with expected values for diffusion processes of chain moleculesthrough polymer matrices, reported to be in the 10-12 kcal/mol interval [8]. By thisinterpretation, contact with water (γnd = 52 mJ/m2) leads to the diffusion of hardsegments into the surface region of the PU, driven by the thermodynamic need tominimise the interfacial tension. In the new equilibrium state, the polymer has aconsiderably higher total surface energy, since as shown in Figure 7.2, the dispersion-force component is not greatly affected by contact with water. When in theconformation dictated by water, the total γs of the PU surface is 42.5 mJ/m2, morethan 10% greater than when this polymer is in its ‘air-conformation’ state.

Figure 7.2 Response of the polyurethane dispersion-force surface energy to waterimmersion at (●) 38.5 °C, (●) 60 °C, and (■) 81 °C.


341


While the emphasis above is on structural reconformation as the cause of changes in γsnd,

other mechanisms may be considered. They include:

(i) plasticisation of the polymer by water; and

(ii) enolisation, by water, of NCO linkages in the polymer.

The former of these may be rejected, since the data in Figure 7.1 were shown to beindependent of the drying time following removal of samples from water. Vacuum-dryingfor 5 hours, for the standard 24 hours, and 48 hours produces samples in which γs

nd wasunchanged within experimental error. The observed surface energy effects are thereforeunlikely to be caused by the physical incorporation of water in the polymer.

The second mechanism, that of enolisation, may be investigated using IR data, focusing onthe NCO absorption peak at 2300 cm–1 and on the OH signal at 3300 cm–1. Pertinentresults are given in Table 7.2 for specimens prepared under the full range of cure conditions.

selbairaverucfotceffe:evisehdaUPfoartcepSRI2.7elbaTmc(aerAkaeP 2)

edoCeruC a mc0032(OCN- 1- ) mc0033(HO- 1- )

0C 13.1 22.2

1C 60.1 71.2

2C 48.0 88.1

3C 77.0 73.1

4C 46.0 27.1

5C 07.0 85.1

6C 34.0 06.1

7C 25.0 44.1

8C 39.0 40.2

9C 47.0 51.2

01C 16.0 37.1

11C 86.0 78.1

21C 95.0 07.1a 1.7elbaTees,snoitinifedroF

folanruoJ,rebierhcS.P.HdnaatpugneS.AmorfnoissimrephtiwdecudorpeR.2elbaT,749,11,5,1991,ygolonhceTdnaecneicSnoisehdA © .VBPSV,1991

342


Comparison of the NCO and OH absorption peak areas shows that chemical changescontinue to take place as the urethane polymer ages. Specimens made by routes C0 and C1show that an additional 30 minutes at 100oC reduces both signals, and more particularlythat for NCO. Additional reductions are observed for samples cured via routes C2 and C3.The series C3-C7 shows continued decreases in the OH peak and a slight downwardtendency for the NCO peak. The sequence C8-C12, referring to room temperature exposureat controlled humidity, follows similar patterns, though the magnitude of the reductions inpeak areas is somewhat diminished. It is concluded that chemical changes do indeed takeplace in the curing polymer, and that even after 14 days at 50 °C these changes have notgone to completion. The speculation on enolisation effects therefore appears to be supportedanalytically. The relative contribution of the supposed enolisation effects to observed changesin surface energy must now be questioned.

One aspect of the question is considered by the results in Figure 7.3. This shows the responseof γnd of the reference polymer to immersion in a non-polar, dispersion-force fluid

Figure 7.3 Response of the non-dispersion force surface energy of referencepolyurethane to immersion in heptane, 60 °C. Polymer previously immersed in water.


343


(n-heptane). Prior to immersion in heptane at 60 °C, the polymer had been immersed inwater at the same temperature for 48 hours to bring the surface state into equilibrium withthe orienting fluid. Consequently, the ordinate in Figure 7. 3, namely the difference betweenγnd of the polymer in the ‘water’ and the ‘air’ steady states, originates at about 6 mJ/m2. Ifthis increase, shown in Figure 7.1, were due entirely to surface restructuring, then immersionin heptane would restore the ‘air’ steady-state and Δγnd = 0. Instead, however, the non-dispersive surface energy levels out at about 1.8 mJ/m2. Conformational and reversiblerestructuring accounts for the major portion of surface energy changes in the partiallycured PU. A smaller, yet significant irreversible portion, however, arises from changes inthe surface chemistry of the polymer when it is allowed to condition in aqueous media.

Another view of the relationship between surface mobility and the level of cure in the PUis shown in Figure 7.4. This follows the γs

nd on immersion in water at 60 °C of samplescured under procedures C1, C5, and C10 (see Table 7.1). Clearly, in each case there is asignificant rise in γnd, but the magnitude of the increase varies with the cure procedure.

Figure 7.4 Effect of the degree of cure on the surface restructuring of polyurethane onwater immersion, 60 °C. Cure states coded according to Table 7.1.

(●) C1, (■) C5; (●) C10.


344


The most drastic cure in this group was applied to sample C5, which produces the lowestrestructuring effect; C10 and C1 represent progressively milder cure strategies and therestructuring tendency correspondingly increases. Thus, surface chain dynamics of PUsare a function of the degree of cure, but under each of the procedures used in this workthe restructuring effect remains sufficiently important, potentially, to affect adhesion,and related properties of the polymer surface.

7.2.2.2 Adhesion Performance

The lap-shear experiments described previously, clearly show that the adhesive bondbetween PU and a polymer substrate is not uniquely defined by the composition of thematerials in the joint. This is shown by the results in Table 7.3 which gives the adhesivebond strengths of ABS/PU joints, in which the adhesive was cured by the various routes

stniojUP/SBAfohtgnertsevisehdA3.7elbaT03tastsetraehs-paleraatadlla( ° )C

edoCeruC a )aPM(aW

1C 6.3 ± 3.0

2C 1.2 ± 3.0

3C 7.1 ± 3.0

4C 5.1 ± 3.0

5C 2.1 ± 3.0

6C 2.1 ± 3.0

7C 2.1 ± 3.0

8C 6.3 ± 3.0

9C 1.3 ± 3.0

01C 3.2 ± 3.0

11C 4.2 ± 3.0

21C 2.2 ± 3.0a 1.7elbaTees,snoitinifedroF

dnaatpugneS.AmorfnoissimrephtiwdecudorpeRdnaecneicSnoisehdAfolanruoJ,rebierhcS.P.H

.3elbaT,749,11,5,1991,ygolonhceT © .VBPSV,1991

345


described in Table 7.1. A wide variation in bond strengths is noted; the major trend istoward reduced adhesion with increasing duration of cure, and thus also with increasingcontact time between PU and ABS. In this case, of course, the orienting medium for theadhesive is the ABS surface. The results, therefore, suggest that the steady-stateconformation of PU relative to ABS is less favourable to strong bonding than is an initial,transient conformation of the adhesive surface. This somewhat surprising result may berationalised with a knowledge of the acid-base characteristics of the polymer.

IGC determinations of AN, DN, and Ki values for ABS, PVC, and PU are summarised inTable 7.4, all data being at 30 °C.

These results show ABS to be a moderate base and PVC a strong acid. PU, measured herein its surface conformation relative to air, is mildly basic. There is, therefore, no tendencyfor specific non-dispersive forces to act across the ABS/PU interface, the basicity of thematerials being unfavourable to the establishment of strong, short-range forces of thiskind. On prolonged contact, the initial tendency for dispersion-force soft segments topopulate the PU surface preferentially should become more pronounced, accounting forthe data sequence in Table 7.3. Of course, the trends should be reversed with PVC, for inthis case acid-base attraction should promote adhesive bond strength. Only fragmentaryexperimental results are available to test the hypothesis, but as shown in Figure 7.5,these support it. They are also consistent with the recent work of Berger [9], showing theinfluence of non-dispersion forces on the lap shear strength of polymer/metal joints.

A matter of much practical importance is the durability of bond strength when jointsare exposed to aggressive environments. A cursory examination of this point was carriedout as follows: joints of PVC/PU and ABS/PU were prepared by cure procedure C2.Several of these were aged at 50 °C/50% relative humidity for up to 10 days and lap-

KdnaND,NA4.7elbaT i seulav

NA ND Ki

SBA 1.5 5.8 4.3

CVP 4.9 1.2 3.7-

UP 4.2 4.4 0.2

346


shear evaluations were carried out as described previously. The results in Figure 7.5,however, are limited for the most part to single specimen evaluations. A lower degreeof confidence therefore applies to them. The effects seen in the figure, however, areentirely consistent with the preceding analyses. With ABS, as before, a slightdeterioration takes place in bond strength. The initial bond strength with PVC assubstrate is significantly higher than that with ABS, and, moreover, increases withageing, as called for by the acid-base concept. The orienting potential of polymersincluding PU therefore may arise from the nature of acid-base interactions at the pointof contact between materials. These effects may have serious consequences on theperformance parameters of materials such as PU adhesives.

Figure 7.5 Effect of ageing at 50 °C, 50% RH on lap-shear adhesion of polyurethane/PVC (▲) and polyurethane/ABS (●) joints. Polyurethane adhesive cured according to

code C2 (Table 7.1)


347


7.2.3 Conclusions

The results of this study lead to the following conclusions:

(1) The surface energy of a PU formulation has been shown to vary with the nature ofthe medium in contact with the polymer. The variations are attributed to the non-dispersive part of the overall surface energy. The observed trends follow closelyearlier observations on two-part, linear PU surfaces.

(2) Thermodynamic or acid-base interaction forces appear to be the dominant drivingforce for surface restructuring in PU, leading to the surface energy effects.

(3) Chemical changes in the PU surface following prolonged exposure to hightemperatures and/or high humidity contribute to a lesser degree to variations in thesurface energy.

(4) The adhesive bond strength of PU varies with the degree of acid-base interactionwith the substrate and with the conformational state of the adhesive molecule atthe time of testing. Being basic, PU develops stronger bonds with acidic substrates(PVC in this case) than with basic ones, this is exemplified by ABS.

7.3 Acid/Base Interactions and the Adhesion of PUs to PolymerSubstrates

As noted in the preceding section, IGC is an excellent tool for measurements of surfaceproperties and of acid base interaction potentials of macromolecular solids (2, 10, 16).In this section the importance of acid-base interactions relative to the adhesion of PUsis considered in greater detail. IGC has been applied to a series of PU adhesives and toselected polymer substrates, allowing quantitative measurements to be made of theacid/base (electron donor-acceptor) interaction parameters applicable to the surfacesof these materials. Acid base pair-interaction parameters for substrate/PU combinationshave been calculated. The bond characteristics of polymer/PU combinations have beenmeasured, in part by conventional lap-shear procedures and in part, by the more recentconstrained blister detachment method [11, 12]. Possible relationships between bondproperties and acid base interactions have been considered, and a comparison of thetwo adhesion tests has been made.

348


7.3.1 Experimental

7.3.1.1 Materials

Six PU adhesives, modelled on practical formulations, have been studied. They are codedPU1 – PU6. All were two-part compounds, denoted as A and B in Table 7.5. In all cases,part B contained a tin catalyst.

In each of these formulations A and B were thoroughly mixed, applied to the substrateand bonded in less than 3 minutes following mixing.

The substrates were rigid PVC, ABS impact resin and the GE polyblend Xenoy. All werecommercial samples, and all were injection moulded to form 8.0 cm x 2.5 cm rectangles,1.5 mm in thickness.

6UP-1UPsevisehdaUPfonoitalumroF5.7elbaT

1UP .30.1foHO:OCNhtiw)447-PXrudnoM(etanaycosiylopcitamorAA

.loirtdesab-enotcalorpacylopdnaliorotsacfoerutximAB

2UP .1UPnisaetanaycosiylopcitamorAA

loidenatub-4,1dna)0521WM(loidenotcalorpacylopfoerutximAB

3UP )noitaroproCreyaBmorfdeniatbo,447-PX(etanaycosiylopcitamorAA.30.1foHO:OCNhtiw

.loidenatub-4,1dna)0001WM(loylopnirdyhorolhcipedetanimret-HOB

4UP M,enimayloP5229etanibuRA n ;0001=

MenimayloPB n .thgiewyb2.5:02fooitaratadeximowtehT.0001=

5UP 001/g4.1taretomorpnoisehdaenalisfoBotnoitiddahtiwtub1UPsA.Bfostrap

6UP 8.0taretomorpnoisehdaenalisfoBotnoitiddahtiwtub4UPsA.Bniloylopfostrap001/strap

Mn thgiewralomegareva-rebmun=

7.3.1.2 Procedures

For IGC work, PU and substrate polymers were coated onto Chromosorb AW support(60/80 mesh) from solution. From 6.0 – 11 wt% of polymer was supported on the

349

Chromosorb. The coated support was packed in stainless steel columns (0.5 cm id, about0.8 m long) and served as the stationary phase in IGC. The stationary phases werecontacted with selected vapours at very high dilution and described procedures werefollowed [6, 11] to measure the net retention volume, Vn, of the vapours probing thedeposited solids. A Perkin-Elmer Sigma-2 chromatograph with hot wire detector wasused. All determinations were in the range 30 – 60 °C, and values of Vn were measuredin at least triplicate, with a reproducibility better than 4%.

The use of IGC data to obtain acid and base interaction parameters for polymer solids hasbeen described in detail elsewhere [2,13,14]. The procedure calls for a reference relationshipbetween RTln Vn and the normal boiling point of mobile-phase materials able to interactthrough dispersion forces only, represented in this work by n-alkanes (C6-C9) . Vapoursknown to be acids and bases were then injected into the columns, again at extreme dilution.Following the theories of Gutmann [3], diethyl ether (DEE) and chloroform (CHL) wereused as reference base and acid, respectively. As shown in Figure 7.6 for PVC as stationary

Figure 7.6 Retention volume of vapour probes interacting with PVC as function ofnormal boiling point of probe liquids. Data are for 40 °C.

Reproduced with permission from PU Congress.


350


phase, the points for DEE and CHL both deviate from the straight line defined by thealkanes. Thus, PVC interacts with both the acidic and basic probe, but with a pronouncedprevalence of interaction with DEE. The polymer is therefore a net acid. Following theprinciples of reference [1], electron acceptor and donor indices, AN and DN for the polymersare calculated from the behaviour of the DEE and CHL vapours.

The adhesion at substrate/PU interfaces was measured by lap-shear and by constrainedblister test methods [11, 12]. For lap-shear testing, adhesives were mixed with glassbeads, Class 4A from Ferro Microbeads, and deposited on PVC and Xenoy substrates toyield dry films 0.35 mm thick. The adhesives were cured by exposure of 3 days at roomtemperature and 50% relative humidity, followed by 4 days at 50 °C, 0% relative humidity.The lap-shear specimen arrangement si shown in Figure 7.7a. All determinations were

Figure 7.7 Schematic of arrangements for adhesion tests. (a) Specimen construction forlap-shear test. (b) Specimen construction and installation for blister detachment test.

351

made in triplicate with an Instron tester, at a separation speed of 50 mm/min. The blistertest arrangement, used with all substrates, is shown schematically in Figure 7.7b. Thesupport polymer here had a central hole, 6 mm in diameter, which connected to a pumpingsystem able to channel distilled water onto the polymer/adhesive assembly. Duringapplication of the adhesive this opening was blocked by a Teflon plug, removed after theurethane was cured. Adhesive films, again 0.35 mm thick, were deposited on the PVC,ABS and Xenoy surfaces and cured as above. Bond energies, Ea, were obtained by followingprotocols described in recent literature [4, 5]. These require measurement of the pressuregenerated by the liquid pumped against the joint at which the adhesive debonds from thesubstrate. Bond energy data were reproducible to better than 8%; lap-shear determinationscarried a similar uncertainty.


7.3.2.1 IGC Determinations

Acid-base interaction parameters for the materials of this research are given in Table 7.6.As anticipated by the results in Figure 7.7, and in agreement with the results shown inthe preceding section, these show PVC to be predominantly acidic. ABS has a moderatelybasic surface and Xenoy is an amphoteric solid with nearly equal AN and DN indices.All of the PU systems function as both donors and acceptors, with PU5 and PU6 skewedto acidity, PU1 and PU4 to basicity, while PU3 and (to a lesser degree) PU2 are amphoteric.

03talla(setartsbusremylopdnaUPfosexedniesab/dicA6.7elbaT ° .)C

lairetaM NA ND

SBA 2.3 0.7

yoneX 6.6 8.6

CVP 5.8 1.2

1UP 7.2 3.4

2UP 7.3 3.5

3UP 6.6 9.6

4UP 9.3 6.7

5UP 2.8 3.5

6UP 2.7 9.3


352


The acidic adhesives would be expected to interact strongly with ABS, the basic oneswith PVC. The magnitude of acid-base forces is open to question when amphoteric surfacessuch as Xenoy are involved.

Any attempt to relate acid/base interactions between adhesives and substrates to thebond properties of the relevant pair depend on a quantitative evaluation of a pairinteraction parameter. There are no broadly accepted theoretical guidelines to such anevaluation, but a pair interaction parameter, Isp, can be defined by

Isp = ([AN]1 [DN]2)1/2 + ([AN]2 [DN]1)1/2

Here subscripts 1 and 2 designate adherent and adhesive, respectively. Values of Isp forthe polymer/adhesive pairs are given in Table 7.7.

03tasretemarapnoitcaretniriaP7.7elbaT °CI ps CVP SBA yoneX

1UP 4.8 0.8 6.9

2UP 5.9 2.9 9.01

3UP 4.11 5.11 4.01

4UP 9.01 1.01 8.31

5UP 6.21 7.11 4.41

6UP 7.9 6.01 1.21

7.3.2.2 Adhesive Bond Properties

A summary of lap-shear and blister detachment test data is presented in Table 7.8. Thesix adhesives perform differently with each of the substrates, suggesting a link with specificinteractions at the adherent/adhesive interface. This is considered in the next section.

Although lap-shear evaluations are more familiar, they do not necessarily evaluate thestrength of adherent/adhesive interfaces. Cohesive failure of the weaker component ofan assembly can be a complicating factor, tearing of the PU layer being a possibility incertain of the present measurements. In principle, the detachment of the adhesive in theblister test is a pure interfacial event and one that can be quantified provided parameters

353

such as adhesive thickness and the geometry of the detaching blister are known [11, 12].The two test methods may therefore complement each other and the existence ofquantitative correlations between the test data is a point of interest. The theory is testedby means of Figure 7.8. Although the results for PVC and Xenoy substrates are separatedby a significant margin, in both cases the correlations are very satisfactory. The correlationcoefficient for the more strongly bonded Xenoy-PU system is 0.988, that for the weakerPVC-PU bonds is 0.935.

Thus it is concluded that for assemblies of the type used here, lap-shear and blisterdetachment test methods produce sets of internally consistent data which correlate in astatistically acceptable manner. The choice of test procedure, therefore, may rest onconvenience and on the proven relevance of test results to field performance.

7.3.2.3 Adhesion and Acid/Base Interaction

As already noted, the adhesion test data suggests the existence of significantcontributions to bond strengths from non-dispersion forces acting at the adhesive/substrate interface. Indeed, quantitative correlations have been confirmed between Isp

and either of the two sets of adhesion performance results. A demonstration is given inFigure 7.9, using the more complete blister test data. The correlation coefficient forthis linear plot is 0.982.

03taseulavtnemhcatedretsilbdnaraehspaL8.7elbaT °CetartsbuS CVP SBA yoneX

evisehdA SL Ea SL Ea SL Ea

1UP 5.02 5.73 - 8.04 0.72 5.24

2UP 0.62 5.24 - 5.04 5.82 0.44

3UP 2.72 5.74 - 2.84 7.92 5.64

4UP 8.52 0.74 - 7.44 5.13 7.94

5UP 0.82 0.94 - 8.74 2.13 3.05

6UP 5.52 5.24 - 2.74 0.03 0.84

m/Jniraehspal:SL 2-

Ea m/Jniygrenednob: 2-


354


Figure 7.8 Correlation between lap-shear and blister detachment evaluations of bondstrengths in Xenoy/PU and PVC/PU assemblies.

The value is in fact surprisingly large when it is noted that no account has been taken ofcontributions made to bond strengths by dispersion (van der Waals) forces. Acid/baseforces appear to play a determinant role when bonding occurs at interfaces of materialswith pronounced electron donor and acceptor tendencies. This alone seems to justifyconcern for expressing these tendencies quantitatively, and for favouring the IGC techniqueas a convenient route to the objective.

Another noteworthy feature in Figure 7.9 is the inclusion of all polymer-substrate pairs isthe relationship. This encourages speculation that Isp is an objective measure of theinteractions generated by electron donor and acceptor sites in the surfaces of the presentpolymers. It also suggests that IGC determinations of Isp may be useful for screening purposesin designing PU adhesives for optimum performance with specified polymeric substrates.These suggestions form the origin of further research which is being carried out.

355

Figure 7.9 The dependence of equilibrium bond energies for PU/polymer assemblieson acid/base pair interaction. ● - PU1, ● - PU2, ▲ - PU3, ■ - PU4

7.4 The Effectiveness of Silane Adhesion Promoters in thePerformance of PU Adhesives

One of the most frequent procedures in the formulation of adhesives is the addition ofadhesion promoters. Of these, silanes are by far the most frequently used. The silanes forma multi-membered family of chemicals. Numerous publications have described aspects ofsilane chemistry and the applications of silanes [15-18]. Although many silanes are readilyavailable, in many industrial formulations of adhesives the most frequently used of theseare the primary amine and diamine versions because of their relatively low cost. However,these do not necessarily produce optimum benefits in adhesion for all applications.Guidelines for the selection of silanes best suited for specific applications would thereforebe useful. These should be based on a fundamental understanding of the manner in whichsilane compounds affect the interface and interphase in adhesively bonded systems. The


356


last section relates to this objective by describing the effect of various silanes on the adhesionperformance of a model PU adhesive in joints involving selected polymers and glass. Theadhesion and its retention under aggressive ageing conditions is interpreted in terms ofacid-base interactions at the adhesive/substrate interface. The interpretative basis appearssuited for the selection of preferred silanes in the cases under study.

7.4.1 Experimental

7.4.1.1 Materials

The adhesive used was a model semi-structural PU, similar to one described in an earlierwork [7]. The two component formulation consisted of methylene bis-4-cyclohexylisocyanate prepolymer and trifunctional polyols with molecular masses in the range 500-900. A small amount of tin catalyst (UL-28, Witco Chemicals) was also present. The PUwas used as control without additives and was modified by the following silanes:

APS - aminopropyl triethoxy silane

CS - chloropropyl trimethoxy silane

VS - vinyl trimethoxy silane

MS - mercaptobutyl trimethoxy silane

ES - epoxy trimethoxy silane

The silanes were obtained from Dow-Corning Corporation and were added at 0.5 wt%levels. All formulations were cured by a standard procedure of 1 hour at 100 °C, followedby 5 days at 50 °C and 50% relative humidity.

As in the preceding section, the polymers to be bonded were PVC, ABS, and the thermoplasticpolyester blend Xenoy. In addition microscope slide glass was also studied. Surfaces ofthese materials have been shown to be capable of widely differing acid-base interactions.The PVC was an unplasticised polymer (MW = 54,000 g/mol) from Synergistic Chemicals,Inc. It contained 5 phr of Advastab TM-821SP thermal stabiliser. The ABS (Cycolac AR3501), was a moulding grade, commercial product of GE Plastics, Inc. The same sourcesupplied the Xenoy blend; the exact composition of this material was not determined.

7.4.1.2 Procedures

The establishment of silane concentration in modified PU adhesives was based on thetendency of PU to restructure in response to the orienting strength of media in contact

357

with the polymer [5, 19]. The process involves the diffusion-controlled accretion of polarmoieties from the bulk to the surface layer of the polymer. It has been shown earlier inthis chapter that surface restructuring can be followed by monitoring γnd, the non-dispersive contribution to the polymer surface energy, when a film, initially dried in air,is immersed in a polar medium. Static contact angles were measured for n-decane,methylene iodide and water, on a freshly prepared film specimen of pure PU, and also onfilms of PU containing up to 2 wt% APS and ES. The harmonic mean procedure [20]was used to compute γnd. The measurements were repeated after each of the films hadbeen exposed for 7 days to water at 90 °C. Preliminary work showed that this wasadequate to allow the PU surface to equilibrate with respect to the aqueous environment.Whilst the silane-free film restructured in water to increase γnd by about 5 mJ/m2, thepresence of silane sharply reduced the effect. The silane apparently anchors surface-localised moieties and restricts their ability to adopt new conformations. Of the silanes,ES appeared to be more successful in inhibiting the restructuring effect, but both ES andAPS, when added at a concentration of 0.5 wt%, were found to greatly reducerestructuring. This was, therefore, chosen as the concentration level in subsequent work.

For testing bond properties, single lap-shear specimens were assembled from polymerand glass coupons cut following the practice outlined in reference [19]. The polymerswere moulded to a thickness of 1.5 mm and cut to 8 x 2.5 cm size. No mould releaseagents were used in moulding operations. For convenience, the selected overlap area of4.5 cm2 was bonded with adhesives drawn to a thickness of 0.4 mm. Followingconditioning (24 hours, room temperature), initial bond strengths were measured intriplicate with an Instron apparatus at a draw speed of 10 mm/min. To ascertain propertyretention, certain of the systems were exposed in air ovens to 60 °C, 100% relativehumidity for up to 3 weeks under an applied stress. To generate stress, small holeswere drilled near the ends of the lap-shear coupons, one of the holes being used tosuspend the test pieces from a horizontal beam within the oven. Consistent with certainindustrial procedures, stress was generated by a 200 g weight, suspended from the freeend of the joint. Repeat evaluations of lap-shear bond strengths were carried out aftervarious ageing times.

Interfacial properties in assemblies were represented by acid-base interactions, as evaluatedby methods of IGC [2], as described earlier. PU and polymer stationary phases wereprepared as described in preceding sections. Since grinding microscope slides proved tobe ineffective for the production of a solid with sufficiently high surface area, powderedglass beads were used as a representative glass substrate. Following well-establishedprocedures [2, 6], the solids were investigated at infinite dilution with vapours includingn-alkanes (C6 to C10), and with acidic and basic probes chosen on the basis of Gutmann’sclassification [3]: chloroform and benzene as acids, diethyl ether, tetrahydrofuran as basesand acetone as an amphoteric substance. The temperature throughout IGC determinations


358


was in the range 35-65 °C. Net retention volumes, Vn, were obtained from at least triplicatevapour injections, generally with a reproducibility better than 4%.

The protocols leading to the determination of acid-base interaction potentials in thisportion of the work were somewhat different from those documented earlier. Here theinteraction potential of solids is described by the acid and base interaction constants, Ka

and Kb. The derivation of these is given in detail in recent publications [4, 14, 21]. Asbefore, an initial requirement is the establishment of a linear plot of log Vn against thenormal boiling point, Tb, of the alkane vapours [9]; this represents the dispersion-forceinteractions between solid and vapour. The slope of the linear plot may also be used toevaluate the dispersion contribution to the solid’s surface energy, γs

d. The position of Vn

for the acid and base probes, at their respective Tb, defines the contribution of acid-baseinteractions to the free energy of adsorption, ΔGab. When measured over a specifiedtemperature range, the corresponding enthalpy, ΔHab, is also obtained. The enthalpy isthen used to estimate Ka and Kb from the expression [4, 14, 21]:

-ΔHab/AN = KaDN/AN + Kb

Here AN and DN are Gutmann’s values [3] for the acidity and basicity of the vapourprobes. When Ka and Kb values for substrates and adhesives are known, an acid-basepair interaction value Isp is calculated:

Isp = (Ka)1(Kb)2 + (Ka)2(Kb)1

where subscripts 1 and 2 refer to adhesive and polymer (or glass) substrate, respectively.In this work, Isp values were primary tools for the generation of correlations betweencomponent interactions and the performance of the bonded systems. In so doing, therelevance to adhesion of a parameter related to the acid-base contribution to the enthalpyof component interaction is established.


The acid-base interaction potentials of polymer substrates and of the various adhesiveformulations are given in Table 7.9 as are the relevant values of Isp. The IGC experimentsshow the model PU to interact as both mild acid and base, with these tendencies wellbalanced. The presence of APS has no pronounced effect on the electron acceptor capacityof the surface, but shifts the balance to basicity, presumably because of the surfacelocalisation of the silane’s primary amine groups. Each of the CS, VS and MS additivesproduces the opposite trend, the adhesive surface now displaying stronger acidity, butwithout significant change in the basic interaction potential. Addition of ES raises both

359

the acid and base interaction potential with a predominance of basicity. Furthermore,the total interaction potential, given by the sum Ka + Kb, is most affected by the ESadhesion modifier. As expected, the PVC surface is strongly acidic, that of ABS is basic,and Xenoy is amphoteric with excellent balance between donor and acceptor capabilities.The glass surface is a weak acid. As a result of choosing the Gutmann concept of acidity/basicity, all of the surfaces retain both acidic and basic interaction potentials.

enahteruyloprofsretemarapnoitcaretniriapdnastnatsnocnoitcaretnI9.7elbaTsetartsbusdnasevisehda

stnatsnocnoitcaretniesab-dicA:A

lairetaM Ka Kb Ka K+ b

CVP 5.8 1.2 6.01

SBA 2.3 0.7 2.01

yoneX 6.6 8.6 4.31

ssalG 6.2 9.1 5.4

UP 1.3 8.3 9.6

SPA+ 7.3 0.6 7.9

SC+ 9.4 0.4 9.8

SM+ 0.5 2.4 2.9

SV+ 6.4 9.3 5.8

SE+ 0.5 2.6 2.11

I,noitcaretniriaP:B ps

CVP SBA yoneX ssalG

UP 93 43 64 61

SPA+ 95 54 56 32

SC+ 44 74 06 02

SM+ 64 84 26 02

SV+ 14 54 75 91

SE+ 46 55 47 62

sedocenaliSfonoitinifedroftxeteeS

folanruoJ,atpugneS.AdnaniQ.R,rebierhcS.P.HmorfnoissimrephtiwdecudorpeR.1elbaT,13,2-1,86,8991,noisehdA © .srehsilbuPecneicShcaerBdnanodroG,8991


360


The resulting Isp data in Table 7.9, show ES and APS to be the most effective promotersof interaction between adhesives and PVC. APS is not an effective interaction modifierfor ABS substrates; all the other surface modifiers exert a greater effect. Indeed, for eachof the four substrates in this work, ES produces the maximum increase of interaction,although for glass the interaction level with these adhesives remains low when comparedwith the polymeric substrates.

The effects of silane additives on the bond characteristics of assemblies is summarised inTable 7.10. The initial bond strength, prior to accelerated ageing, is reported. The contributionmade by the silanes is variable, ranging from a few percentage points, e.g., PVC/PU + MS, toa doubling of the bond strength (Glass/PU + ES). The largest increment in bond strength isgenerated by ES. As seen in Table 7.9, this confers the largest increase to the total interactionpotential of the adhesive surface, while maintaining a balance between donor and acceptorcharacteristics. For the systems selected for study in this work, ES is closest to a ‘universal’adhesion modifier. In other cases, the effects of acid/base functionality are evident. Thus, theacidic substrate (PVC) benefits from the presence of the basic APS modifier, but does notrespond well to silanes which accentuate the acidity of the adhesive surface. ABS demonstratesthe analogous effect of responding more strongly to the acidic modifiers than it does to APS.In joints based on Xenoy, a strongly amphoteric substrate, the choice of silane is less critical,although ES maintains a slight advantage over the remainder. The strong performance of theepoxy silane notwithstanding, it seems doubtful whether a ‘universal’ silane promoter can beidentified for application to a wide range of substrates.

deifidomevisehdaUPybdeniojseilbmessafohtgnertsdnoblaitinI01.7elbaTsevitiddaenalisyb

m/J(dnobraehs-paL 2 rof)evisehdadnaetartsbus

CVP SBA yoneX ssalG

evitiddaonUP 72 02 82 41

SPA+ 43 42 53 12

SC+ 82 42 43 91

SM+ 03 782 43 91

SV+ 13 92 13 71

SE+ 93 43 93 72

sedocenalisfonoitinifedroftxetees;%tw5.0tasenalisllA


361

The critical question of interdependence between initial lap-shear bond strengths and theIsp parameters is addressed in Figure 7.10. Although each of the substrates tends to definea specific envelope of behaviour (more particularly evident for glass), the first order regressionline drawn through all of the points has a correlation coefficient of 0.915, sufficient tojustify the claim of a true relationship. A more rigorous correlation is probably unrealistic,since no account is taken here of varying chemical interactions between the silanes and thesubstrates, or of variations in the contributions to bonding made by longer range dispersionforces. Acid-base interactions at the adherent/adhesive interface thus make importantcontributions to the initial lap-shear bond strength of these systems. The extrapolation toIsp = 0 suggests that dispersion forces make a net contribution of about 12 J/m2 to the bondproperty. Within experimental error (~10%) this is the initial bond strength for glass joinedby unmodified PU, even though an Isp value of 16 is reported for this pair in Table 7.7.

Figure 7.10 Initial bond strengths as a function of pair interaction parameter

Reproduced with permission from H. P. Schreiber, R. Qin and A. Sengupta, Journal ofAdhesion, 1998, 68, 1-2, 31, Figure 2. ©1998 Gordon and Breach Science Publishers.


362


06(gniegafotceffE11.7elbaT ° htgnertsdnobraehs-palno)ytidimuhevitaler%001,C

)h(emitegA 0 3.1 5.3 9.6 11 4.22 R

m/J(htgnertsdnoB 2)

UP/CVP 72 22 71 41 41 9 33.0

SPA+ 43 33 03 72 42 22 56.0

SV+ 13 82 32 12 71 41 54.0

SE+ 83 73 73 33 03 82 47.0

UP/SBA 02 81 61 21 8 7 53.0

SPA+ 42 12 12 61 71 31 45.0

SV+ 82 52 42 22 02 91 86.0

SE+ 43 43 03 92 72 62 67.0

UP/yoneX 82 32 91 61 61 21 34.0

SPA+ 53 53 23 82 82 52 17.0

SV+ 13 92 72 42 91 71 55.0

SE+ 83 63 53 23 23 92 77.0

UP/ssalG 41 21 7 0 0 0 00.0

SPA+ 12 02 61 11 11 7 33.0

SV+ 81 71 21 9 7 4 22.0

SE+ 72 52 52 02 71 31 84.0

sedocenalisfonoitinifedroftxeteeShtgnertsdnoblaitiniotlaniffooitarehtstneserperR


Evidently at this level of Isp, the strength of specific interactions is insufficient to influencethe lap-shear bond strength as measured by the chosen procedure.

Of equal interest to the initial bond strength is the ability of the assembly to withstandaggressive ageing. Joints using as adhesives the unmodified PU and the PU with 0.5 wt%APS, VS or ES adhesion promoters were subjected to the ageing protocol of this work.

363

Results of the procedure are presented in Table 7.11. Ageing periods are shown as thesquare root of exposure hours, in the expectation that property loss would be a diffusion-controlled process due to the intrusion to the interface of water. In Table 7.11, R isdefined as

R = (Bond Strength)f /(Bond Strength)i

where the subscripts f and i represent the bond strength after three weeks of ageingand the initial value, respectively. Each of the systems is adversely affected, but Rvaries significantly among the systems as well as within them. Qualitatively, arelationship between R and the initial bond strength becomes apparent on inspectionof the data. The ES modified PU, which has the strongest acid-base interaction witheach of the substrates, produces the strongest initial bond, also produces the smallestproperty loss.

A clearer view of the apparent correlation is presented in Figure 7.11. This indicatesthat below initial bond strengths of about 13-14 J/m2, corresponding to Isp < 18-20,the selected ageing regime will totally destroy the bonded assembly. The best fit ofthe data is obtained from the second order regression curve shown, with a coefficientof 0.932. However, a first order linear fit to the data (not shown) with a coefficientof 0.917 is not much inferior and has the attraction of being extrapolatable to R =1.0. If the linear fit is accepted as realistic, then the extrapolation predicts that anassembly with an initial lap-shear bond strength > 45 J/m2 would withstand the chosenageing exposure indefinitely, without loss of bond strength. As the correlation inFigure 7.11 makes evident, the degree of acid/base interaction at the bonded interfacealso should correlate with the residual bond property. The matter is put to the testwith results shown in Figure 7.12. The resulting linear fit to experimental data has acorrelation coefficient of 0.938, with much of the scatter contributed by the glassjoints, where acid-base interactions are feeble and most subject to uncertaindetermination. An extrapolation to R = 1.0, identifies Isp to be greater than or equalto 85, as the strength of acid-base interactions required to ensure long-term resistanceto the selected ageing procedure. The forecasts should be useful in optimising thechoice of silane (and other) adhesion promoters in the formulation of adhesives suchas the present PU. However, the physico-chemical analyses on which these are basedcannot specify the chemical approaches to be followed. Moreover, the generalapplicability of the physico-chemical analyses followed here calls for the study of abroader range of adhesive/additive/substrate systems.


364


Figure 7.11 Bond strength ratio versus initial bond strengths


7.4.3 Conclusions

1. A strong correlation has been shown to exist between initial bond strengths displayedby polymer and glass surfaces bonded by PU adhesives containing a variety of silaneadhesion promoters and the acid-base interaction at substrate/adhesive interfaces.

2. Silane additives change the surface interaction potential of PU in distinct ways; APSaccentuates surface basicity, while CS, MS and VS accentuate surface acidity. Theaddition of ES results in significant increases in both surface acidity and basicity.

3. Within the scope of the present work, ES was found to be the most effective adhesionpromoter for each of the substrates. APS was particularly useful with the acidic PVCsubstrate, but the CS, MS and VS were preferred additives for the basic ABS substrate.

365

Figure 7.12 Dependence of pair interaction parameter on bond strength ratio


4. A correlation has been developed between the residual adhesion of bonded assembliesfollowing accelerated ageing and the magnitude of acid-base interfacial interactions.The strength of interactions needed to avoid property loss under the chosen ageingconditions can be estimated. This capability represents a guideline to the selection ofpreferred silane additives for the adhesion of PU adhesives to substrates with knownacid-base interaction potentials.

5. IGC was shown to be a convenient and effective tool for evaluation of acid-baseinteractions at substrate and adhesive surfaces.


366


References

1. A. Gent and J. Schultz, Journal of Adhesion, 1972, 3, 4, 281.

2. Inverse Gas Chromatography, Eds., D. R. Lloyd, T. C. Ward and H. P. SchreiberACS Symposium Series No. 391, American Chemical Society, Washington, DC1989.

3. V. Gutmann, The Donor-Acceptor Approach to Molecular Interactions, PlenumPress, New York, 1978.

4. C. Saint Flour and E. Papirer, Industrial and Engineering Chemistry ProductResearch and Development, 1982, 21, 4, 666.

5. Z. Deng and H. P. Schreiber in Contemporary Topics in Polymer Science, Vol. 6,Ed., W. M. Culbertson, Plenum Press, New York, 1989, p.385.

6. D. G. Gray, Progress in Polymer Science, 1977, 5, 1.

7. J. H. Chen and E. Ruckenstein, Journal of Colloid and Interface Science, 1990,135, 496.

8. V. Stannett in Diffusion in Polymers, Ed., J. Crank Butterworths, London, 1968,Chapter 2.

9. E. J. Berger, Journal of Adhesion Science and Technology, 1990, 4, 5, 373.

10. J. M. Braun and J. E. Guillet, Advances in Polymer Science, Volume 21, SpringerVerlag, Berlin, 1976, p.108.

11. A. N. Gent, and L. H. Lewandowski, Journal of Applied Polymer Science, 1987,33, 5, 1567.

12. Y-S. Chang, Y-H. Lai and D. A. Dillard, Journal of Adhesion, 1989, 27, 4, 197.

13. C. Saint Flour and E. Papirer, Journal of Colloid and Interface Science, 1983,91, 63.

14. U. Panzer and H. P. Schreiber, Macromolecules, 1992, 25, 14, 3633.

15. E. P. Plueddemann, Interfaces in Polymer Matrix Composites, Academic Press,New York, 1974, p.173.

16. E. P. Plueddemann, Silane Coupling Agents, Plenum Press, New York, 1982.

367

17. D. E. Leyden, Silanes, Surfaces and Interfaces, Gordon and Breach Science, NewYork, 1986.

18. H. Ishida in The Interfacial Interactions in Polymeric Composites, Ed., G.Akovali, NATO ASI Series E, Volume 230, Kluwer Academic Publishers, TheNetherlands, 1993, p.169.

19. A. Sengupta and H. P. Schreiber, Journal of Adhesion Science and Technology,1991, 5, 11, 947.

20. D. T. Sawyer and D. J. Brookman, Journal of Analytical Chemistry, 1968, 40,1847.

21. J. Schultz, L. Lavielle and C. Martin, Journal of Adhesion, 1987, 23, 1, 45.


368


369

8 HER Materials for Polyurethane Applications

Raj B. Durairaj

8.1 Introduction

Polyurethane elastomers have a wide range of industrial applications due to their uniquecombination of valuable physical and mechanical properties. These elastomers, ingeneral, are composed of phase separated flexible soft segments and crystalline hardsegments. The soft segment is based on polyether or polyester type aliphatic diols andthe hard segment is the result of reaction between aromatic diisocyanate and lowmolecular weight diol chain extenders [1]. The soft segments have low polarity and arelonger than the hard segments and provide a soft and flexible matrix. On the otherhand, the hard segments are shorter, highly polar and have a strong tendency toaggregate. The thermodynamic incompatibility of these two segments leads to phaseseparation, which ultimately influences the physical and mechanical properties ofurethane elastomers.

In the development of cast polyurethane elastomers, various chain extenders are usedwith toluene diisocyanate (TDI) and 4,4´-diphenylmethane diisocyanate (MDI)terminated prepolymers based on polyether or polyester polyols. With MDI-terminatedprepolymers, 1,4-butanediol is the most commonly used chain extender. In order tomaintain desired mechanical properties at elevated temperature and to improve hardnessand tear resistance, aromatic diols such as bis-(β-hydroxyethyl) ether of resorcinol(HER) or hydroquinone (HQEE) are often used. One of the difficulties with HQEE isits processing problem associated with a high melting point (~102 °C). On the otherhand, the structurally similar isomer HER offers a lower melting point (89 °C), thusfacilitating more forgiving chemistry and processing ease. Previous studies [2-5] haveshown that HER chain extender offers significant processing advantages over HQEEand the properties of HER extended elastomers are comparable to those of HQEEcured elastomers.

With the development of technical grade HER (HER TG), the processing temperatureof the HER was further lowered without sacrificing the physical and mechanicalproperties of the final cured cast elastomers. The major difference between HER HPand HER TG chain extenders is the presence of some higher molecular weight (MW)reactive aromatic diol monomer in the technical grade material. With this unique

370


product, HER TG, there is no need to add plasticisers, diluents or aliphatic diols tolower the melting points of aromatic diol chain extenders. It is also important to notethat the presence of about 10 weight percent of this higher MW aromatic diol monomerdoes not appear to affect the performance of the cast elastomers. With this discovery,it was decided to investigate different grades of HER materials with varying levels ofhigher MW monomer and to study their performance in cast elastomers based on MDIterminated polyether prepolymer.

The following are some of the objectives of this chapter:

1. To develop low melting and highly processable aromatic diol extenders from resorcinoland ethylene carbonate (HER materials).

2. To determine the nature of products or by-products formed by the resorcinol-ethylenecarbonate reaction.

3. To investigate how the high MW aromatic diol monomer affects the phase separationbehavior of polyurethane elastomers.

4. To examine the effect of various HER chain extenders on the physical, mechanicaland dynamic mechanical properties of cast polyurethanes made with MDI-terminatedprepolymers.

5. To study both polyether- and polyester-based polyurethane systems.

8.2 Experimental Conditions

8.2.1 Chain Extenders

HER materials used as chain extenders in this work were prepared based on INDSPECChemical Corporation’s patented technology and the product details are given inTable 8.1.

Hydroquinone (di-β-hydroxyethyl) ether (HQEE) and 1,4-butanediol (BD) were obtainedfrom commercial sources.

8.2.2 Prepolymers

Polyether and polyester-based MDI prepolymers obtained from commercial sources wereused. The details are given in Table 8.2

371

HER Materials for Polyurethane Applications

1.8elba

Tsnoitacilppa

enahteruyloprof

slairetam

RE

Hfo

tnempoleve

D

epyT

rednetxE

PH-

RE

H-

RE

H012

GT

)012-G

T(

-R

EH

522G

T)522-

GT(

-R

EH

052G

T)052-

GT(

-R

EH

572G

T)572-

GT(

QIL-

RE

H

mroflacisyhPenillatsyr

Cdilos

netloM

dilosnetlo

Mdilos

dilosyxa

Wdilos

ytsaPdiuqi

L

52,s-aP(ytisocsi

V°

)C

dndn

dndn

dn02

(tniop

gnitlem

yrallipaC

°)

C29-88

09-8758-67

87-5656-36

dn

)mrehtodne(

sisylanaCS

D

(erutarep

mettesn

O°

)C

7.686.18

176.56

SK

AEP

3dn

(erutarep

metkaeP

°)

C1.98

586.08

7.1756-53

)C

Gyb,

%.tw(

tnetnocR

EH

99>

3928

8645

31

HC

O(H/

RE

H2

HC

2)n

HC(

O-R-

O2

HC

2)

OnH

)oitar.tw;laciteroeh

T;3dna

2=

n(0/001

7/3971/38

23/8664/45

001/0

HC

O(H/

RE

H2

HC

2)n

HC(

O-R-

O2

HC

2)

OnH

)oitar.tw;

RM

Nyb;2

=n

(99>

4/6951/58

13/9604/06

dn

denimrete

D:rebmunlyxordy

Hdetaluclac,laciteroeh

T:065-555

665055-545

455535-525

535025-005

605084-074

964973283

sutatSM

OC

MO

CV

ED

VE

DV

ED

VE

D

laicrem

moC:

MO

Clatne

mpoleveD:

VE

Ddeni

mretedto

N:dnytirup

hgiH:P

Hedarglacinhce

T:G

Tyhpargota

morhcsa

G:C

G

372


8.2.3 Preparation of Cast Elastomers

All elastomers used in this project were prepared by the following standard procedure.

A known quantity of prepolymer was weighed in a glass jar. Separately, in another glassjar, the chain extender was weighed to achieve 95 or 90% stoichiometry. Both jars werethen placed in a vacuum oven at the desired temperatures and held under vacuum untilthe bubbles disappeared indicating that all the dissolved gases and any moisture, if present,was removed. This process required about 1 mm Hg vacuum and 2.5 to 3.0 hours atabout 80-110 °C depending on the chain extenders (HER materials or HQEE) used forcasting. After degassing, both the prepolymer and extender were mixed thoroughly forone minute using a stirrer to avoid bubbles. The mix was poured into a heated stainlesssteel plate mould pretreated with Teflon or silicone mould release agent. After pouring,the mould was placed in a programmable oven and cured for 16 hours. The samples,after curing, were allowed to cool slowly to room temperature. Test specimens were cutfrom the cured sheet for determination of tensile, tear strength and dynamic mechanicalanalysis (DMA) and then stored for at least 7 days at 22 °C, 50% relative humidityconditions before testing. Compression set samples were prepared separately, followingthe degassing procedure similar to that for plate casting, and aged similarly before testing.

8.2.4 Physical and Mechanical Properties Determination

The following test methods were used to determine the properties of cast polyurethane elastomers:

• Pot life determinations were done using a Brookfield viscometer at controlled temperatures.

• Shore Hardness, Durometer A/D (ASTM D2240) [6].

• Tensile Modulus, strength and % elongation (ASTM D412) [7].

2.8elbaTremyloperP-IDM reilppuS OCN%

retseyloP242-SMcetyaB.1090-SMcetyaB.2

reyaBreyaB

6.60.9

rehteyloP526BENAHTARBIV.1

050-EMcetyaB.2lacimehClayorinU

reyaB4.68.5

373


• Tear Resistance, Die C (ASTM D624) [8].

• Bashore Rebound (ASTM D2632) [9].

• Compression Set % (ASTM D395) [10], Method B.

• DMA was determined using a Rheometrics RMS-800 instrument at 1 Hz frequencywith a heating rate of 2-10 °C/minute.

• Differential scanning calorimetry (DSC) analysis of the cast elastomers was performedon a Perkin Elmer (DSC-7) instrument at a heating rate of 10 °C/minute under anitrogen atmosphere.

• Fourier transform-infra-red (FT-IR) analysis was performed on a Perkin Elmer System-2000 instrument.

8.3 HER Materials Synthesis and Characterisation

High purity (HP) and technical grade HER materials were synthesised by the reaction ofvarious ratios of resorcinol and ethylene carbonate. Previous studies had demonstratedthat the HER TG has a performance equal to that of the HER HP material [5]. It wasalso observed that, in the cast polyurethanes, HER TG could be processed at a lowertemperature than the HER-HP. These interesting experimental facts prompted thedevelopment of various HER-TG materials and an investigation into their effectivenessas chain extenders to enhance the processing of cast and other polyurethanes. By varyingthe molar ratios of resorcinol and ethylene carbonate, the technical grade materials suchas HER TG-210, HER TG-225, HER TG-250, HER TG-275 and HER LIQ weredeveloped. HER TG-210 is the same as HER-TG reported in prior work and in thischapter. The major difference between the HP and TG materials is the presence of higherMW reactive diol monomers present in the TG materials. From the resorcinol and ethylenecarbonate reaction, depending upon the excess of ethylene carbonate, two types of higherMW diols can be expected:

HOH2CH2CO-R-OCH2CH2OCH2CH2OH

and

HOH2CH2C(OH2CH2C)n-O-R-O(CH2CH2O)nCH2CH2OH

where R = meta phenylene, n = 1 and 2

374


On the basis of theoretical calculations and also from the gas chromatography (GC)analysis, hydroxyl number and nuclear magnetic resonance (NMR) analysis results, ithas been confirmed that the reactive high MW diol present in the technical grade materialshas the structure:

HOH2CH2C(OH2CH2C)nO-R-O(CH2CH2O)nCH2CH2OH

where n = 1 and 2

For HER TG-210, TG-225 and TG-250 materials, the reactive high MW diol haspredominantly n = 1 and is present at approximately 7, 17 and 32 weight percent,respectively. With higher molar ratios of ethylene carbonate:resorcinol, high MWdiol with n = 1 and 2, having MW of 286 and 374, respectively, are observed to bepresent. In HER LIQ, the HMW diol monomers were estimated to be as high as 87weight percent.

It was suspected that the presence of these reactive diol monomers in the HERcould have several advantages. Since they have primary hydroxyl groups, theirreactivity towards isocyanate groups is similar to that of other reactive diol chainextenders. The molecular weight of these high MW diols is still low enough thatthey act as chain extenders and not as polyols. In addition, they can act as reactiveplasticisers in various polyurethanes and other polymers. Due to the presence ofthis high MW material, the melting point of HER is lowered, thereby increasing itsprocessability in polyurethane applications. Laboratory observations indicated thatHER and other technical grade HER materials have supercooling tendencies. Infact, HER TG-275 stays liquid at room temperature for a period of 2-3 months.The TG-250 remains liquid at room temperature indicating that processing couldbe easier with this material alone and it would be a good candidate for mixing withother small chain extenders, such as butane diol, and polyols (polyether andpolyester) at lower temperatures.

The melting points determined by the capillary method and DSC analysis are givenin Table 8.1. The varied melting point determined by the capillary is associated withthe high MW diol impurity. To determine the melting point by DSC, measurementswere done at a heating rate of 5 °C/minute under a nitrogen atmosphere. HER-HPshowed one sharp endothermic peak. With HER TG-210, TG-225 and TG-250materials, the major endothermic peak shifted to lower temperatures and also showedone minor peak, indicating one impurity. For HER TG- 275, three endothermic peakswere observed between 35-65 °C indicating three components. For this material, GCanalysis showed two major peaks in addition to the HER peak.

375


8.4 Cast Poly(Ether Urethanes)

8.4.1 Pot Life Determination

Before making the castings, pot life determinations were performed to determine theprocessing window of the prepolymer and chain extender materials. As a standard practice,both the prepolymer and chain extender were preheated to the desired temperatures,degassed thoroughly, mixed at the desired stoichiometry (90 or 95%) and poured into athermostatted vial. The viscosity was measured, using a Brookfield viscometer, as afunction of time. The measurements were terminated when either the time of measurementreached 20 minutes or the viscosity of the mix reached 0.1 Pa-s.

Figure 8.1 shows the viscosity-time plots of VIBRATHANE B 625 extended with HQEEand various HER materials. As can be seen, the pot life temperatures for HER materialsare considerably lower than HQEE. The shorter pot life for HQEE is due to the higherprocessing temperature. If the processing temperature were lower than 110 °C, thenHQEE would crystallise out from the solution. Premature crystallisation is a majorproblem encountered by most users of HQEE. But these problems are eliminated withHER-HP and other HER-TG materials due to their supercooling behaviour.

Figure 8.1 Pot-life curves for Vibrathane B-625-HER materials

376


8.4.2 Polyurethane Castings

The prepolymer % NCO content and the equivalent weight of the chain extender materialswere determined to calculate the amount of each material to achieve 95% stoichiometry.After careful degassing and mixing, the material was poured into the stainless steel mouldand cured for 16 hours continuously at 110 °C. Table 8.3 summarises the experimentaldetails relating to cast elastomer preparation.

In order to compare the properties of cast polyurethanes, conditions relating to castings,ageing and testing were kept constant throughout this work.

8.4.3 Calculation of Hard and Soft Segment Contents

HER-HP has a purity of >99%. The other HER-TG materials contained at least one ortwo high MW reactive diol monomers. The presence of these reactive diol monomerscan play a major role in the formation of hard segments, hard segment content andphase separation behavior of the polyurethanes, which will ultimately dictate the finalcured physical and mechanical properties of the cast elastomer materials. A previousstudy on HER-LIQ showed that the cast elastomers based on this extender producedsofter material compared to HER-HP [5]. The elastomeric materials developed usingHER-LIQ appeared to be more transparent indicating that the high MW diols present inthis chain extender were responsible for this change. The MW of these high MW diolsare still low enough to produce a hard segment in the polyurethanes, but the nature ofthe hard segment can be different from that formed between the isocyanate and HERdue to a longer and more flexible chain.

Based on the extensive study made on HER and HQEE chain extenders, it is well knownthat HER can produce hard segments with MDI and other isocyanate monomers, whichare crystalline in character [4]. Due to their high flexibility, the high MW reactive diolmonomers can produce hard segments which may be more amorphous than crystallinein character. The amorphous hard segment formed can mix with the soft segment of thepolyurethane elastomers and ultimately change the physical and mechanical propertiesof the elastomeric materials. Therefore, it was decided to calculate the amount of hardand soft segment contents of the polyurethanes resulting from the VIBRATHANE B 625prepolymer and HER materials containing various levels of the high MW reactive diols.

VIBRATHANE B 625 is a polyether prepolymer based on polytetramethylene glycol(PTMEG) and MDI raw materials. Based on the 6.5% NCO content, it can be assumedthat this was prepared using 1 mole of PTMEG-1000 polyol and 2.2 moles of MDI. Itcan also be assumed that the HER chain extender produces crystalline hard segments

377


ecnamrofrep

sremotsale

remyloperp

)rehteylop(526-

BlayorinU

nodesab

sremotsale

tsaC

3.8elba

Tslaireta

mG

T-R

EH

suoiravhti

w

slairetaM

RE

HE

EQ

HP

H-R

EH

-R

EH

012G

T-

RE

H522

GT

-R

EH

052G

T-

RE

H572

GT

-R

EH

QIL

(,erutarepmet

remyloperP

°)

C011

00109

0909

0908

(,erutarepmet

rednetxE

°)

C011

00109

0909

0908

)rednetxe/remyloperP(

oitarxi

M6.41/001

5.41/0019.41/001

5.51/0011.61/001

2.71/0017.02/001

)laciteroeht%(,yrte

moihciotS59

5959

5959

5909

011/h(,eruC

°)

C61

6161

6161

6161

)aPM(

seitreporPelisne

T

suludoM

%00153.11

65.0117.01

48.889.7

49.999.3

suludoM

%00240.31

17.3175.41

77.1153.11

44.0136.6

suludoM

%00346.41

33.7180.91

6.5153.61

97.6155.11

htgnertselisne

T9143

61837393

21737304

28631032

noitagnolE

%835

955674

264314

273123

)m/

Nk(,kaerbot

ygrenE

7697

0735

9404

02

)m/

Nk(,C

EID,htgnerts

raeT

721511

20179

6827

44

%,tesnoisserp

moC

124.51

4.7171

5.714.92

1.9

%,dnuobererohsa

B55

8506

0645

6494

)D(

A-erohS,ssendraH

6959

)84(59

)44(29

)34(09

)04(98

)03(87

378


and the high MW diols produce amorphous hard segments. Taking into account theamount of HER and high MW diols present in the chain extender, for the 95%stoichiometry, the crystalline hard segment, amorphous hard segment and polyether softsegment contents were calculated for different cast polyurethanes made using theVIBRATHANE B 625 prepolymer.

These calculations were based on the reaction scheme (8.1) shown opposite and anidealised polyurethane structure from this scheme and the details are given in Table 8.4.

As can be seen from Table 8.4, the crystalline hard segment content is decreased from48.7% to 35.4% when the high MW diol content increases from 0% to 87%.Concurrently, the amorphous hard segment content increases from 0% to 16.6%. At thesame time, the soft segment content resulting from the polyol was affected only slightly(changed from 51.3% to 48%). Since the high MW diol changed the concentrations ofthe hard and soft segments, it is interesting to see how this will affect the physical andmechanical properties of the cast elastomers.

In general, the contributions of both the soft and hard segments in the polyurethanes canbe correlated with the properties observed. The soft rubbery block primarily affectsresiliency, wear, tear, compression set and low temperature properties, while the hardblock affects hardness, modulus and tensile properties [11].

8.4.4 Hard Segment Versus Hardness

The Durometer hardness was measured after the cast elastomers were aged for at least 7days at 23 °C and 50% relative humidity. The details are given in Table 8.3. From theresults, it is evident that as the crystalline hard segment content of the elastomer isdecreased, the elastomer hardness also decreases. By analysing the soft and hard segmentsand their relation to hardness, it appears that the amorphous hard segment behaves likethe polyether soft segment.

8.4.5 Tensile Properties

The tensile data obtained from the cast elastomers are summarised in Table 8.3, plottedand shown in Figure 8.2. From this figure, it can be seen that the 100% modulus and %elongation decreased for the HER-TG materials compared to HQEE and HER-HP chainextenders. On the other hand, the 200% and 300% modulus values were higher forHER-HP and HER TG-210 extenders compared to HQEE and the values were lower forthe other technical grade materials. Except for HER-LIQ extender, all HER based

379


senahteruyloptsac

nistne

mgestfos

dnadrah

fonoitubirtsi

D4.8

elbaT rednetx

Eniahc-re

myloPnoitalucla

ClaciteroehT

sisylanA

slairetaM

RE

H)

%(tne

mgeSdra

HtfoS

tnemgeS

)%(

tnetnoC

RE

H)

%(H

CO(

H2

HC

2)n

HC(

OR

O2

HC

2)

OnH

)%(

enillatsyrC

suohprom

A

)0001-G

EM

TP/ID

M(re

myloperP.1)1/2.2

=oitar

eloM;

%5.6=O

CN(

rednetxeo

N0

3.140

7.85

PH-

RE

H+

526-B

enahtarbiV.2

0010

7. 840

3.15

012G

T+

526-B

enahtarbiV.3

397

849.0

1.15

522G

T+

526-B

enahtarbiV.4

2881

9.642.2

9.05

052G

T+

526-B

enahtarbiV.5

8623

4.441.5

5.05

572G

T+

526-B

enahtarbiV.6

4564

1.347.6

2.05

QIL-

RE

H+

526-B

enahtarbiV.7

3178

4.536.61

84

OC

N-R

-NH

CO

O(P

olye

ther

)OO

CH

N-R

-NC

O +

HO

(H2C

H2C

O) n

-R′-(

OC

H2C

H2)

n-O

H

(Pre

poly

mer

)(n

= 1

-3, c

hain

ext

ende

r)

O-H

2CH

2CO

-R′-O

H2C

H2C

-OO

CH

N-R

-NH

CO

O-(

Poly

ethe

r)-O

OC

HN

-R-N

HC

OO

(CH

2CH

2O) 2

-3-R

′-(O

CH

2CH

2)2-

3-O

|So

ft|

||

|—C

ryst

allin

e H

ard

Segm

ent

—|

Segm

ent

||—

Am

orph

ous

(Har

d) S

egm

ent

—|

whe

re R

- d

iphe

nyl m

etha

ne a

nd R

´-1,

3-ph

enyl

ene

grou

ps.

Rea

ctio

n sc

hem

e 8.

1

380

Advances in Urethane Science and TechnologyHER Materials for Polyurethane Applications

extenders showed either equal or better tensile strength values than the HQEE extender.In general, the HER-TG materials showed good tensile properties.

8.4.6 Tear, Compression Set and Rebound Properties

These properties are summarised in Table 8.3.

Tear strength is generally a very important property in many applications. By comparingthe tear strength values of the technical grade materials against HQEE and HER-HP,they can be seen to decrease continuously as the hardness of the elastomers is decreased.

HER-HP had a lower compression set than the HQEE extended elastomer. This valueincreases when the concentration of high MW aromatic diols content is increased in theHER-TG extenders. In fact, HER TG-275 extender with about 46% high MW aromaticdiol showed a high compression set of 29.4%. On the other hand, when this high MWaromatic diol concentration reached 87% in HER-LIQ, the elastomer based on this chainextender produced a very low compression set material.

Figure 8.2 Tensile properties of cast polyurethanes

381

The most important property to look for in choosing materials for applications such asskate wheels is resilience or rebound [12]. The ether-based elastomers, in general, givehigher rebound resilience. The measured Bashore technique rebound values, show thatHER-HP, HER TG-210 and HER TG-225 extenders have slightly better resilience thanHQEE. The rebound values appeared to be lower for HER TG-250- and HER TG-275-based elastomers. This value was higher for HER-LIQ (49%) when compared to TG-275 (46%) extender. Since the rebound is generally measured at room temperature, itsvalue would relate to tan δ data given in Table 8.5. A comparison of the two indicates acorrelation between tan δ and the rebound values. Obviously, HER-LIQ extender showedlower tan δ (0.044) compared to HER TG-275 (0.076) material.

8.4.7 Differential Scanning Calorimetry

The influence of the high MW diols present in the HER chain extenders on the transitionbehavior of cast elastomers was studied by DSC analysis. Figure 8.3 compares the DSCcurves of HQEE, HER-HP and different HER-TG extended polyurethanes. The transitions,which are observed to occur above room temperature, are associated with the melting of


Figure 8.3 DSC curves of cast polyurethanes from Vibrathane B-625-HER materials

382

Advances in Urethane Science and Technologyecna

mrofrepsre

motsalere

myloperp)rehteylop(

526Blayorin

Uno

desabsre

motsaletsa

C5.8

elbaT

slairetam

GT-

RE

Hsuoirav

htiw

stluseR

sisylanAlacinahce

Mci

manyD

slairetaM

RE

HE

EQ

HP

H-R

EH

012G

T522

GT

052G

T572

GT

QIL-

RE

H

)%(

tnetnocS

Henillatsyr

C7.84

849.64

4.441.34

4.53

)%(

tnetnocS

Hsuohpro

mA

09.0

2.21.5

7.66.61

G,suludoM

egarotS′

)aPM(

52ta

°C74

0382

1241

4193

051ta

°C34

2291

418.9

4.81.3

ytilibatSlamreh

TG

tnatsnocrof

mumixa

merutarep

meT(

′)(

erutarepme

T°

)C

071061

061061

061551

061

G′

01x

aPM,

85.93

9.914.71

6.218.8

7.79.2

G,suludoM

ssoL

′′)aP

M( T,erutarepmet

kaePg

(°)

C9.07-

8.07-8.86-

76-3.56-

5.36-2.95-

suludom

kaeP2.09

8.599.09

6.983.38

9.674.56

52ta

°C9.1

1.11.1

88.096.0

0.171.0

051ta

°C98.0

53.013.0

52.081.0

61.050.0

atleD

naT

T,erutarepmet

kaePg

(°)

C3.36-

4.16-2.95-

4.15-4.34-

3.54-4.53-

atlednat

kaeP891.0

552.0152.0

703.0592.0

642.0174.0

52ta

°C930.0

830.0830.0

340.0740.0

670.0440.0

051ta

°C20.0

610.0610.0

610.0810.0

20.0810.0

32/z

H01

ta°

)ycneuqerf(C

550.0440.0

240.050.0

750.0290.0

270.0

J,ecnailpmo

Csso

L′′

)aPM(

(erutarep

metkaeP

°)

C3.35-

3.35-4.94-

3.93-9.42-

4.58.41-

ecnailpmoc

kaeP007;0011

0031;00710061;0081

00930046

006500623

52ta

°C008

00310041

00120033

006500111

051ta

°C005

007008

00110091

00320065

383


enahtarbiVmorfsenahteruyloptsacnostluserCSD6.8elbaTslairetamREH-526B

erutarepmeTkaePcimrehtodnE

gnitaeHtsriF gnitaeHdnoceS

rednetxEniahC (erutarepmeT ° )C Δ )g/J(H (erutarepmeT ° )C

EEQH.1 502 71 432,502

PH-REH.2 991,681 22 422,481

012GT-REH.3 591 91 291,551

522GT-REH.4 381 31 291,261

052GT-REH.5 181 9 051

572GT-REH.6 371 31 341

QIL-REH.7 dn dn dn

hard segment domains. The melting behaviour of the hard segments can be seen as anendothermic peak in the DSC curve. The hard segment melting temperatures and heats oftransitions observed from the DSC analysis are given in Table 8.6. From the heats oftransition data (ΔH), it is clear that when the hard segment content is decreased, the energyassociated with the melting of the hard segments is also decreased. In addition to this, thepresence of amorphous hard segments appeared to decrease the melting temperatures ofthe crystalline hard segment. Since the temperature at which the first transition takes placeabove room temperature in the DSC curve is considered to indicate the limit of the thermalstability of the polyurethane elastomer, the thermal stability of these elastomers was foundto be decreased when the high MW diol contents were increased. This might be due to thereduction of hard segment content. Higher hard segment transition in the DSC indicateshigher thermal stability of the resulting polyurethane.

8.4.8 Dynamic Mechanical Analysis

DMA is an analysis technique used to determine the dynamic properties of theelastomers [13, 14]. Dynamic properties of the elastomeric materials are importantbecause they influence the performance of certain parts such as wheels and tyres.This method determines the storage modulus G´ (elastic behaviour), loss modulusG´´ (energy dissipation), tan δ, loss compliance J´´ and glass transition temperature(Tg) values. The Tg of the soft segment can determine the low temperature behaviourof polyurethane elastomers. This is not only influenced by the nature of the soft

384


segment (polyether) but also the degree of phase separation between the hard andsoft segments. In this work, the DMA method was utilised to understand the effectof HQEE, HER-HP and various HER-TG materials on the cast elastomers madeusing VIBRATHANE B 625 (PTMEG-Polyether) prepolymer. The results aresummarised in Table 8.5.

8.4.8.1 Storage Modulus G´ Property

The storage modulus, G´, values determined by the DMA for the elastomers extendedwith HER chain extenders are given in Table 8.5, and the storage modulus curves areshown in Figure 8.4. Storage modulus quantitatively measures the material’s elasticproperties and also qualitatively determines the elastomer’s stiffness and hardness.Since the hardness of the elastomer is related to its hard segment content, the DMAstorage modulus was found to decrease as the hardness of the elastomers was decreaseddue to the increasing amounts of high MW diols present in the HER-TG materials.From the DMA curve, it was observed that the G´ was considerably decreased forHER-LIQ extended elastomer which contained about 87% of the high MW diol

Figure 8.4 Storage modulus (G′) curves of cast polyurethanes

385

monomers. It has been observed and reported in the literature [13, 15] that elastomersbased on PTMEG polyol exhibit a secondary transition (additional dispersity) whichappears as a step in the temperature—shear modulus function curves. The same kindof behaviour is observed for elastomers based on HQEE, HER-HP and HER TG-210chain extenders at temperatures below 0 °C. A sharp transition in the G´ curve maybe associated with good phase separation of the elastomer. In this respect, a gradualtransition occurs with HER TG-225, TG-275 and HER-LIQ compared to HQEE,HER-HP and HER-TG-210 extended elastomers indicating that poor phase separationoccurrs in these elastomers.

8.4.8.2 Thermal stability of Elastomers

The thermal stability of urethane elastomers can be determined by the DMA method.The temperature at which the storage modulus (G´) decreases significantly in therubbery region is considered to be the limit of thermal stability of the elastomers[16]. From the G´ values of the elastomers, it was determined that HER extendedmaterials were stable up to 160 °C whereas HQEE extended elastomer showed a10 °C higher stability.

8.4.8.3 Loss Modulus Property (G´´)

The loss modulus, G´´, is a quantitative measure of energy dissipation in theelastomers. Low values of G´´ are indicative of low energy dissipation, low hysteresisand, consequently, low heat build up, an important requirement for applicationssuch as in-line skate wheels [17]. Figure 8.5 shows the plots of loss modulus as afunction of temperature. The loss modulus values are high at low temperature andfound to decrease as the temperature is increased. The curves show maximacorresponding to the Tg of the soft segments of the elastomers which graduallyshift to higher temperatures indicating the effect of high MW diol monomer of theHER materials.

In the temperature range between 0 and 160 °C, the loss modulus decreases into a longplateau and then starts to increase when the temperature reaches beyond 170 °C and180 °C. This might be due to the melting of the hard segments.

Figure 8.6 shows the plots of storage modulus (G´) and loss modulus (G´´) as a functionof various technical grade HER materials along with HQEE and HER HP measured at25 and 150 °C. The curves clearly demonstrate that G´ and G´´ decrease as the hardsegment content is decreased in these elastomers.


386


Figure 8.5 Loss modulus (G′′) curves of cast polyurethanes

Figure 8.6 Effect of HER materials on shear modulus properties

387

8.4.8.4 Tan Delta Property (tan δ)

The measure of tan δ is a qualitative tool to determine the hysteresis or heat build up inan elastomer during dynamic flex conditions. Figure 8.7 shows the plot of tan δ as afunction of temperature and the values obtained from this determination are shown inTable 8.5. As the hard segment content is decreased, the tan δ peak height is increasedindicating higher hysteresis of the elastomers associated with the higher amounts of highMW diols. The peak widths are also observed to increase, corresponding to the decreasein the hard segment content of the elastomers. By comparing the tan δ peak (Tg) to thatof loss modulus peak (Tg) temperatures, the tan δ temperatures appeared to be higher.The Tg associated with the soft segment were found to shift higher as the amorphoushard segment content of the elastomers increased. This temperature shift might beassociated with the restriction of mobility of the soft segments due to phase mixing.Also, the greater the amount of hard segment dissolved in the soft segment phase, thehigher the expected Tg of the soft segment.


Figure 8.7 Tan delta curves of cast polyurethanes

388


The Tg of the elastomers increased gradually when the hard amorphous content ofthese materials was increased. Since this hard amorphous segment is the result ofreaction between the isocyanate and high MW aromatic diol monomers containingether linkages, there is a possibility that this amorphous hard segment can mixeffectively with the soft polyether polyol segments of the elastomers. As the amorphoushard segment content is increased, it is possible that the effect of phase mixing betweenthis phase and the soft polyol segment is enhanced, thereby increasing the Tg of thesoft segment of the resulting polyurethane elastomers. The Tg of the HER-LIQextended elastomer is increased by as much as 26 °C with 16.6% amorphous hardsegment content. These results indicate that the high MW aromatic diol monomerspresent in the HER-TG materials, influence the low temperature properties of thecast polyether elastomer materials.

8.4.8.5 Loss Compliance Property (J´´)

Loss compliance quantifies the heat generated during dynamic loading in applicationslike tyres, wheels and rollers. Polyurethane elastomers with lower loss compliancevalues will experience less heat build up and, consequently, may suffer fewer fieldfailures [14]. Figure 8.8 shows the graph of loss compliance as a function oftemperature for the cast elastomers made with HER extended materials. The resultsare summarised in Table 8.5.

Loss compliance curves show peak maxima at the soft segment Tg. The values appearto be lower for elastomers containing higher hard segment contents. In the case ofHQEE, HER-HP and HER-TG 210 extended elastomers, loss compliance curves showedtwo peaks at the Tg. The two peaks appear to merge into a single peak as the amorphoushard segment of the elastomers is increased. Based on the loss compliance values, itappears that HQEE based elastomers show less heat generation than the HER materialscontaining higher levels of high MW diols. However, HER-HP and HER TG-210 basedelastomers behave similarly and, therefore, comparable performance can be expectedin their applications.

In order to compare the Tg observed from the peak temperatures of loss modulus (G´´),tan δ and loss compliance (J´´) curves, the peak temperatures are plotted as a functionof the amorphous hard segment and shown in Figure 8.9. As can be seen in the figure,the amorphous hard segment produced by the high MW aromatic diol extenders has apronounced effect on the soft segment Tg which, in turn, ultimately dictate the lowtemperature properties of the elastomers made from them.

389


Figure 8.8 Loss compliance (J′′) curves of cast polyurethanes

Figure 8.9 Amorphous hard segment versus peak temperatures

390


8.5 Cast Poly(Ester Urethanes)

Polyurethane elastomers containing polyester soft segments are used in applications suchas wheels, rollers, gaskets, seals, sheet goods, belting, cyclone liners and forklift tyresdue to their high tensile strength, tear strength, abrasion resistance, oil resistance andheat-aging properties.

In this work, MDI-terminated polyester prepolymer based on ethylene adipate polyol(Baytec MS-242) was used to prepare cast poly(ester urethane) materials. The HER-TG materials, HER-HP and HQEE were used as chain extenders. As discussed insection 8.4, technical grade HER materials contain various levels of high MW diolsas reactive impurities. Therefore, as with poly(ether urethanes), the high MW diolsare expected to play a major role in the performance of poly(ester urethanes) madewith these materials.

8.5.1 Pot Life Determination

The experimental details on the pot life determination and elastomer casting conditionsare given in Table 8.7.

Pot lives exceeding 20 minutes can be easily achieved with HER materials due totheir low melting and processing temperatures. Because of their supercoolingbehaviour and tendency to stay liquid after cooling to room temperature, furtherenhancement in the pot life can be achieved by lowering the mixing temperature ofTG-250 and TG-275 materials.

8.5.2 Tensile Properties

The data obtained from the tensile measurements are summarised in Table 8.7 and thegraphs are shown in Figure 8.10. From the results, it can be seen that 100% modulusand %elongation are decreased and the tensile strength increased as the high MW diolscontent of the chain extender is increased.

8.5.3 Tear, Fracture Energy, Compression Set and Rebound Properties

These properties are summarised in Table 8.7 and the values are plotted against elastomerhardness and shown in Figure 8.11.

391


remyloperp

)retseylop(242-S

Mreya

Bno

desabsre

motsaletsa

C7.8

elbaT

seitreporPre

motsalE

nonoitisop

moC

EE

QH

-R

EH

fotceff

Eslaireta

ME

EQ

H/R

EH

EE

QH

PH-

RE

H012

GT-

RE

H522

GT-

RE

H052

GT-

RE

H572

GT-

RE

Hnoitani

mreteD

efil-toP(

erutarepme

T°

)C

011001

5858

5858

)setunim(

efil-toP5.7

0202>

02>02>

02>)s-aP(

ytisocsivefil-toP

1.018.01

7.451.6

5.53.5

snoitidnoC

gnitsaC

(erutarep

metre

myloperP°

)C

011001

5858

5858

(erutarep

metdednetx

E°

)C

011001

5858

5858

)rednetxe/remyloperp(

oitarxi

M7.41/001

51/0011.51/001

61/00171/001

6.71/001laciteroeht

%,yrtemoihciotS

5959

5959

5959

011/h(,eruC

°)

C61

6161

6161

61)detalucla

C(enahteruyloP

tsaC

)%(

tnemges

draH

DN

3.736.63

4.537.33

8.13)

%(tne

mgestfoS

DN

7.265.26

3.268.16

4.16)

%(tne

mgesdrah

suohprom

AD

N0

9.03.2

5.48. 6

)aPM(

ytreporPelisne

Tsuludo

M%001

35.1131.11

55.019.8

90.791.5

suludoM

%0021.41

39.4194.41

95.2141.01

84.7suludo

M%003

93.618.81

31.913.71

26.5137.01

htgnertselisne

T24.82

6.9239.43

94.4358.93

71.43noitagnole

%627

996975

994854

034)

m/Nk(

kaerB

otygren

E)evruc

niarts-ssertsrednu

aera(701

221001

5708

04

)m/

Nk(C

eiD,htgnertS

raeT

841441

731011

40177

)%(

tesnoisserp

moC

2.624.03

4.239.15

5.5515

)%(

dnuobererohsa

B14

9343

3382

51)

D(A-erohS,ssendra

HD

05D

64D

44)

04(

49)

93(

98)

43(48

392


Figure 8.10 Tensile property of Baytec MS-242 prepolymer extended withHER chain extenders

Figure 8.11 Variation of cast polyurethanes properties as a function of hardness

393

The hardness of cast polyurethanes is decreased as the high MW diols content of thechain extender is increased. This eventually affects some of the most important elastomerproperties such as the tear strength, 100% tensile modulus and fracture energy. Thepresence of high MW diols in the TG materials also increases the amount of amorphoushard segment which, in turn, interferes with the phase separation of hard and softsegments of the polyurethane elastomers. The continuous reduction in the reboundproperty is associated with the formation of higher amorphous hard segments fromthese TG materials.

8.5.4 Differential Scanning Calorimetry Analysis

During the DSC analysis, the elastomer sample is first heated to 225 °C (first heating),cooled back to room temperature (cool down) and then heated again to 250 °C (secondheating) at 10 °C/minute heating and cooling rates. The endothermic transitions observedabove room temperature during the heating are associated with the melting of hardsegments and exothermic peaks are due to the crystallisation of the hard segment uponsubsequent cooling. DSC studies made on the segmented polyurethanes revealed threehard segment-associated endotherms at temperatures of 60-80 °C, 120-190 °C and above200 °C [18-21]. These endothermic transitions are due to short-range, long-range andmicrocrystalline ordering of hard segment domains.

The endothermic peak temperature determined from the second heating is taken to bethe thermal stability limit of the elastomer. The hard segment melting and crystallisationtemperatures and their heats of transitions determined from the DSC analysis are givenin Table 8.8. Figure 8.12 shows the DSC curves obtained during the heating and coolingof cast elastomers based on HER-HP and HER TG-250 chain extenders.

From the second heating endothermic peak temperatures, HER-HP extension resultedin a thermal stability at 177 °C and the stability decreased to 135 °C for an elastomermade from TG-275 extender. An interesting observation was also made in the DSCanalysis during cooling. An exothermic peak associated with the crystallisation of hardsegment was seen for the elastomers made from HER-HP, TG-210 and TG-225extenders. Although this exothermic peak is not seen for TG-250 and TG-275 extendersduring the DSC cooling curve, this exothermic behaviour is seen during the secondheating. This strongly suggests that the higher amorphous hard segments formed fromthe isocyanate and high MW diols interfere with the crystallisation of the crystallinehard segment.

From the endothermic transition energy, HER-HP extension produced a higher hardsegment content compared to the other TG materials.


394


slairetam

RE

H-242-SM

morfsenahteruylop

tsacno

stluserCS

D8.8

elbaT

gnitaeH

tsriFn

woD

gnilooC

gnitaeH

dnoceS

mrehtodnE

mrehtoxE

mrehtodnE

mrehtoxE

kaePygren

EkaeP

ygrenE

kaePygren

EkaeP

ygrenE

lairetaM

EE

QH/

RE

HskaeP

°Cg/J

°Cg/J

°Cg/J

°Cg/J

EE

QH.1

rojaM

09182

20101

88101

DN

DN

roniM

102812

9

PH-

RE

H.2roja

M881

3338

41771

31D

ND

N

roniM

112,302

012G

T-R

EH.3

rojaM

98132

3701

3717

DN

DN

r oniM

DN

041

522G

T-R

EH.4

rojaM

58151

666

8616

DN

DN

roniM

402041

052G

T-R

EH.5

rojaM

07141

DN

DN

3414

176

roniM

512

572G

T-R

EH.6

rojaM

DN

DN

DN

5313

475

roniM

7122

=D

NN

detcetedto

395


The storage modulus (G´), thermal stability, Tg and loss compliance (J´´) propertiesobtained from the DMA are given in Table 8.9.

Poly(ester urethanes) based on HER chain extenders behaved in the same way as poly(etherurethanes). When the high MW diol content of the extender is increased, the storagemodulus is decreased due to the reduction of the hard segment content. As was seenfrom thermal analysis, the thermal stability of the HER-HP extended elastomer is slightlyhigher than the technical grade materials.

In addition to storage modulus reduction, a continuous reduction in the Tg, an increasein tan δ and increase in the loss compliance values are also noticed when the amorphoushard segment content of the elastomers is increased.

Figure 8.12 DSC curves of cast polyurethanes from Baytec MS-242


396


242-SMcetyaBmorfsremotsaletsacfostluserAMD9.8elbaT

slairetaMREH EEQH PH-REH 012GT 522GT 052GT 572GT

G,suludoMegarotS ′ )aPM(

52ta °C 5.15 3.03 7.82 8.32 4.21 5.6

001ta °C 7.63 2.81 4.61 1.21 3.6 3

051ta °C 5.33 7.51 6.01 4.7 1.4 2.2

erutarepmeT(ytilibatSlamrehTGtnatsnocrofmumixam ′)

(erutarepmeT ° )C 071 071 061 061 051 051

G′ aPM, 6.02 8.7 7.8 7.4 1.4 2.2

G,suludomssoL ′′ )aPM(

T,erutarepmetkaeP g (° )C 6.92- 7.72- 62- 6.72- 8.91- 9.31-

suludomkaeP 251 661 371 951 401 641

52ta °C 92.3 46.1 97.1 81.2 72.1 45.0

001ta °C 84.1 26.4 73.0 5.0 61.0 50.0

051ta °C 95.1 91.4 73.0 15.0 42.0 1.0

atleDnaT

T,erutarepmetkaeP g (° )C 5.12- 1.02- 8.91- 4.91- 6.7- 5.0

atlednatkaeP 72.0 764.0 254.0 904.0 5.0 947.0

52ta °C 460.0 450.0 260.0 190.0 201.0 580.0

001ta °C 40.0 520.0 220.0 40.0 620.0 910.0

051ta °C 740.0 720.0 530.0 70.0 850.0 340.0

32/zH01ta ° )ycneuqerf(C 580.0 770.0 380.0 421.0 451.0 22.0

J,ecnailpmocssoL ′′ )aPM(

(erutarepmetkaeP ° )C 9.0 9.11- 9.9- 5.01 7.5 1.01

ecnailpmockaeP 0661 0604 0693 0014 0021 0053

52ta °C 0321 0871 0612 0283 0518 0821

001ta °C 0901 0041 0831 0643 0704 0036

051ta °C 0141 0961 0233 0049 0241 0391

397

8.6 Cast Polyurethanes from HER/HQEE Blends

Although the cast polyurethanes made using HQEE chain extender produce highperformance materials, processing problems are always encountered with thismaterial. In order to use HQEE, the processing temperatures must be as high as120-130 °C. If the temperatures are lower than this, then ‘starring’ (after curing a‘snowflake’ type material embedded in the continuous matrix, (cast elastomer) canbe seen) occurs due to the crystallisation of HQEE out of the elastomer system. Toovercome this processing problem, HQEE is often blended with various aliphaticdiols such as ethylene glycol, butanediol, and so on [22-24]. The addition of aliphaticdiols into HQEE, therefore, is expected to affect the high temperature physical andmechanical properties of the resulting elastomers. These problems can be easilyovercome if an aromatic diol chain extender like HER is used in the place of aliphaticdiols. Due to the supercooling behaviour of HER, this extender offers a uniqueopportunity to lower the melting point of HQEE without the customary degradationof properties.

8.6.1 Freezing Point Determination of HER/HQEE Blends

To study the eutectic nature of the HER/HQEE blends, freezing point determinationswere made. HER/HQEE blends were prepared at different weight ratios and meltedat elevated temperatures. Then the melt was allowed to cool slowly under controlledconditions. The temperature at which the first crystallisation was observed wasconsidered to be the freezing point of the melt corresponding to that particular HER/HQEE composition.

The plot of freezing point versus HER/HQEE blend is shown in Figure 8.13. As canbe seen from the plot, HQEE melting can be considerably reduced with the help ofan HER extender.


398


Figure 8.13 Apparent freezing points of HER/HQEE blends

8.6.2 Cast elastomers and Their Properties

The experimental details on the cast elastomers prepared using Baytec MS-242 and BaytecME-050 prepolymers and their physical, mechanical and dynamic mechanical propertiesare given in Tables 8.10 and 8.11.

In the case of poly(ester urethanes), HER and HQEE extended elastomers showedcomparable tensile, tear, glass transition temperature (Tg) and hysteresis (tan δ) properties.Compared to HER, HQEE extension produced elastomer with about 2 Shore-D unitshigher hardness. An improvement in the tensile strength value was observed for theHER/HQEE blend at 25:75 weight ratio. In general, all the physical and mechanicalproperties were slightly affected or lowered with HER/HQEE blends compared to eitherHER or HQEE extended material.

In the case of poly(ether urethanes), HQEE extended elastomer had slightly higher hardness,tensile strength and tear strength values than the HER extended material. HER extensionproduced elastomer with a lower compression set compared to HQEE extension.

DMA properties of HER and HQEE extended elastomers are similar. All the other physicaland mechanical properties of the elastomers from the HER/HQEE blends were lowerthan the HER or HQEE extenders.

399

.remyloperp)retseylop(242-SMcetyaBnodesabsremotsaletsaC01.8elbaTseitreporpremotsalenonoitisopmocEEQH-REHfotceffE

)oitar.tw(EEQH/REH 0/001 52/57 05/05 57/52 001/0

(,erutarepmetremyloperP ° )C 011 301 011 501 011

(,erutarepmetrednetxE ° )C 011 011 011 011 011

oitarxiM)rednetxeniahc/remyloperp(

0.51/001 0.51/001 1.51/001 3.51/001 7.41/001

)laciteroeht%(,yrtemoihciotS 59 59 59 59 59

011/h(,eruC ° )C 61 61 61 61 61

)aPM(ytreporPelisneT

suludoM%001 31.11 56.8 19.8 74.01 35.11

suludoM%002 39.41 6.11 68.11 80.31 1.41

suludoM%003 8.81 6.51 53.51 2.71 4.61

htgnertselisneT 3924 1244 3933 8174 2214

noitagnolE% 996 846 705 436 627

)m/Nk(CeiD,htgnertsraeT 441 621 021 531 841

)%(,tesnoisserpmoC 4.03 2.92 5.82 1.72 2.62

D-erohS,ssendraH 25 05 05 25 45

sisylanAlacinahceMcimanyD

32/zH01taatlednaT °C)ycneuqerf(

370.0 601.0 790.0 290.0 670.0

52taatlednaT °C)peewserutarepmet(

450.0 270.0 960.0 360.0 650.0

Tg ,)kaepatlednat( °C 9.12- 2.61- 5.51- 1.81- 1.42-

Tg G( ′′ ,)kaep °C 4.92- 8.32- 8.52- 9.52- 7.92-


400


.remyloperp)rehteylop(050-EMcetyaBnodesabsremotsaletsaC11.8elbaTseitreporpremotsalenonoitisopmocEEQH-REHfotceffE

)oitar.tw(EEQH/REH 0/001 52:57 05:05 57:52 001:0




5.21/001 1.21/001 9.21/001 6.21/001 7.21/001

)yroeht%(,yrtemoihciotS 09 09 09 09 09

011/h(,eruC ° )C 61 61 61 61 61


suludoM%001 3.9 3.8 8.8 9 2.01

suludoM%002 1.21 9.11 3.11 3.11 21

suludoM%003 3.51 6.51 8.41 5.31 6.31

htgnertselisneT 2143 4343 1393 7873 2373

noitagnolE% 506 905 825 165 236

)m/Nk(CeiD,htgnertsraeT 011 89 101 901 811

)%(,tesnoisserpmoC 6.41 4.61 9.51 3.81 4.61

D-erohS,ssenraH 74 84 05 05 05

sisylanAlacinahceMcimanyD

32/zH01taatlednaT °C)ycneuqerf(

830.0 930.0 830.0 730.0 330.0

52taatlednaT °C)peewserutarepmet(

230.0 30.0 50.0 430.0 920.0

Tg ,)kaepatlednat( °C 9.26- 8.95- 1.75- 5.75- 4.16-

Tg G( ′′ ,)kaep °C 4.07- 5.96- 76- 3.96- 3.96-

401

8.7 High Hardness Cast Polyurethanes

High hardness polyurethanes are used in applications such as rolls and rollers. Whenthe hardness of the elastomer is increased, certain other physical and mechanicalproperties such as the tensile strength, tear strength, compressive strength and tensilemodulus are also increased. To develop materials with the best possible propertiessuch as high hardness, solvent, oil and abrasion resistance, MDI-terminated prepolymerswith NCO contents of greater than 9% are often used.

Hard polyurethane elastomers, produced from the higher NCO containing prepolymers,are generally more hydrolytically stable than the soft ones, since the hard segment ishydrophobic and is less rapidly attacked [16]. With the high NCO containingprepolymers, HER chain extended materials are expected to produce harderpolyurethane elastomers due to longer chain length than the conventional 1,4-butanediolchain extender. On the other hand, the presence of a high MW diol chain extenderimpurity in the HER-TG materials would affect the physical and mechanical propertiesof the resulting polyurethanes.

The main objective of this work was to investigate and study the properties andperformances of hard cast polyurethane elastomers obtained from the HER chainextenders.

8.7.1 Cast Elastomers and Their Hard and Soft Segment Contents

Cast elastomers were prepared using Baytec MS-090 prepolymer and HER chainextenders with 95% stoichiometry. The experimental details of the elastomer castingsand the mix ratios of the prepolymer and chain extender are given in Table 8.12.

Since the cast urethane elastomer properties are greatly influenced by the soft andhard segment contents and their phase separation behaviour, they are theoreticallycalculated from the molar ratios of the prepolymer and chain extenders and arereported in Table 8.12. Table 8.12 shows the hard segment content of the elastomeris 45.5% for HER-HP and is reduced to 38.2% for TG-275. This reduction in thehard segment content is due to the formation of 8.8% amorphous hard segmentfrom TG-275 chain extender.

8.7.2 Hardness, Tensile, Tear, Compression Set and Rebound Properties

These properties are summarised in Table 8.12.


402


sremotsaletsacfoseitreporP21.8elbaT

slairetaMREH PH-REH 012-GT 522-GT 052-GT 572-GT

snoitidnoCgnitsaC




4.02/001 2.02/001 1.12/001 8.12/001 42/001

)laciteroeht%(,yrtemoihciotS 59 59 59 59 59

011/h(,eruC ° )C 011/61 011/61 011/61 011/61 011/61

)detaluclac(senahteruyloptsaC

)%(tnemgesdraH 5.54 5.44 34 6.04 2.83

)%(tnemgestfoS 5.45 3.45 45 5.35 35

)%(tnemgesdrahsuohpromA 0 2.1 3 9.5 8.8


suludoM%001 44.02 42.02 79.71 24.51 89.01

suludoM%002 23.32 66.32 3.22 62.12 20.71

suludoM%003 28.52 95.62 5.72 44.92 10.72

htgnertselisneT 24.62 97.62 41.92 40.23 69.33

noitagnolE% 454 983 304 543 553

)m/Nk(ygreneerutcarF)evrucniarts-ssertsrednuaera(

501 18 28 37 55

)m/Nk(,CeiD,htgnertsraeT 941 241 531 661 19

)%(,tesnoisserpmoC 82 4.63 9.73 4.54 8.74

)%(,dnuobererohsaB 13 13 82 62 42

)D(A-erohS,ssendraH D26 D26 )D65(89 )D65(79 )D84(59

403

HER-HP and TG-210 chain extenders produced elastomer hardnesses in excess of 60Shore D, but the hardness decreased to 48 Shore D for the elastomer extended with TG-275 material.

By correlating the elastomer hardness with that of other physical and mechanicalproperties, a clear trend can be seen showing that as the hardness is decreased, the 100%tensile modulus, fracture energy and tear strength values are also decreased.

On the other hand, a gradual increase in tensile strength and compression set values isobserved for cast elastomers containing more amorphous hard segment contents.

Polyurethane elastomers with higher hard segment content result in more hard segmentsmixed into the soft phase [25-27]. Also, polyester soft segments are more compatiblewith urethane hard segments. The decrease in Bashore rebound values of TG-225, TG-250 and TG-275 extended materials is associated with the lower degrees of phaseseparation in the polyurethanes.

8.7.3 FT-IR Analysis of Cast Polyurethanes

In order to understand the effect of high MW diols on the formation of amorphous hardsegments and their interaction with the crystalline hard segments, FT-IR analysis of thecast polyurethanes was performed. Samples were cast as films on a NaCl plate from ahot dimethyl sulphoxide (DMSO) solution.

The crystalline structure of polyurethanes is controlled by the formation of hydrogenbonds between NH and C=O groups of the urethane linkages [28, 29]. From the IRanalysis, the hydrogen bonded C=O present in the crystalline domain can be identifiedby the peak absorbance at 1699-1706 cm-1 and the non-bonded or free C=O groupshows peaks at 1735 cm-1 and 1748 cm-1.

Figure 8.14 shows the C=O stretching vibration bands of FT-IR spectra of castpolyurethanes obtained from the HER materials. The appearance of peaks at 1695 cm-1

and 1705 cm-1 strongly suggests hydrogen bonded C=O groups exist with all the HERextended materials. The intensities of these two peaks are gradually reduced when thehard segment content of the elastomer is decreased. The complete disappearance of the1695 cm-1 peak might be associated with the strong interaction or interference of theamorphous hard segment in the phase separation of elastomers.

IR analysis also showed a peak at 1736 cm-1 associated with the free C=O groups of theelastomers. From the IR absorbances of bonded and free C=O groups, the hydrogen-bonding index, ‘R,’ which is the ratio of Abonded C=O/A nonbonded C=O, can be calculated [30].


404


Figure 8.14 FT-IR stretching vibration bands of HER materials extendedcast polyurethanes

The greater the value of R indicates an increase in participation of the C=O group in thehydrogen bonding. The values of R calculated from the absorbance of hydrogen bondedC=O (1705 cm-1) and free C=O (1736 cm-1) peak values are given in Table 8.13.

The values clearly suggest an increased hydrogen bonding with HER HP extended materialdue to high hard segment content. The reason for the higher R value for TG-225 extenderis not known at the present time.

senahteruyloptsacrofxednignidnobnegordynfoeulaV31.8elbaTrednetxEniahC xednIgnidnoBnegordyH

6371A/5071A=R

PHREH 08.0

012-GTREH 76.0

522-GTREH 38.0

052-GTREH 06.0

572-GTREH 05.0

405

8.7.4 Differential Scanning Calorimetric Analysis

The crystalline hard segment melting, crystallisation and heats of transition observedduring the DSC analysis are given in Table 8.14.

From the endothermic melting temperatures, HER-HP and TG-210 extended elastomersshow a thermal stability of 180 °C. Increase in hard segment content in the elastomerresults in an increase in hydrogen bonding between adjacent hard segments which mayresult in higher transition energy. By comparing the energies of endothermic peaks observedduring the first and second heating DSC experiments, HER-HP extended elastomer hasthe highest hard segment content. The presence of amorphous hard segment appeared toaffect the crystallisation behaviour, during cooling, and also the thermal stability of theurethane elastomers.


The DMA analysis results are summarised in Table 8.15

Due to higher hard segment content, HER-HP and TG-210 extended elastomers hadhigh storage modulus (G´) compared to the other HER chain extenders. By comparingthe tan δ, Tg and loss compliance properties, it is seen that HER-HP and TG-210 extendedelastomers behaved similarly.

8.8 High Thermal Stability Polyurethane with Low Heat Generation

It is well known that conventional polyester-based urethane elastomers extended withbutanediol can withstand continuous use temperatures of about 80 °C. At highertemperatures, a reduction in the physical and mechanical properties is seen due todegradation of the material. The thermal stability of the polyurethanes is related to thenature of the starting materials such as the aromatic diisocyanate and diol chain extender.The hard segment of the urethane elastomer is primarily responsible for temperatureresistance, and the soft segment determines the material’s performance at low temperature.

Butanediol chain extender has four aliphatic carbon atoms in the chain. On the otherhand, HER is an aromatic diol chain extender with the following structure:

HOH2CH2CO OCH2CH2OH


406


slairetam

RE

H-090-SM

morfsenahteruylop

tsacno

stluserCS

D41.8

elbaT

gnitaeH

tsriFn

woD

gnilooC

gnitaeH

dnoceS

mrehtodnE

mrehtoxE

mrehtodnE

mrehtoxE

kaePygren

EkaeP

ygrenE

kaePygren

EkaeP

ygrenE

lairetaM

EE

QH/

RE

HskaeP

°Cg/J

°Cg/J

°Cg/J

°Cg/J

PH-

RE

Hroja

M681

1348

81081

51dn

dn

roniM

dn441

1

012G

T-R

EH

rojaM

081,39152

2751

18102

dndn

roniM

461251

522G

T-R

EH

rojaM

88191

859

5718

755.1

roniM

361,371251

052G

T-R

EH

rojaM

17102

dndn

8416

1701

roniM

dn

572G

T-R

EH

rojaM

66101

dndn

2415

1811

roniM

6121

=dn

Ndetceted-no

407

Based on the structures of butanediol and HER, it can be expected that the HER-basedhard segment would have a longer chain length than the butanediol-containing hardsegment. The domains containing these longer hard segments are also expected to showa higher melting temperature than the shorter ones. Therefore, HER based elastomersare expected to be more thermally stable than the butanediol extended material.

remyloperp090-SMnodesabsremotsaletsacfostluserAMD51.8elbaT

slairetaMREH PH-REH 012GT 522GT 052GT 572GT

G,suludoMegarotS ′ )aPM(

52ta °C 6.89 6.39 4.57 7.84 5.92

051ta °C 7.33 8.52 2.22 9.21 5.8

erutarepmeT(ytilibatSlamrehTGtnatsnoCrofmumixaM ′)

(erutarepmeT ° )C 561 051 051 051 051

G′ )aPM(, 33 62 22 31 9

G,suludomssoL ′′ )aPM(

T,erutarepmetkaeP g (° )C 7.21- 4.41- 35.8- 5.2- 4.5

suludomkaeP 831 341 741 741 641

52ta °C 5.01 3.01 01 9.8 9.11

051ta °C 97.0 76.0 74.0 22.0 41.0

atleDnaT

T,erutarepmetkaeP g (° )C 4.6- 2.6- 4.0- 1.7 4.41

atlednatkaeP 952.0 642.0 3.0 4.0 75.0

52ta °C 601.0 11.0 331.0 381.0 693.0

051ta °C 320.0 620.0 120.0 710.0 610.0

32/zH01ta ° )ycneuqerf(C 41.0 41.0 81.0 3.0 16.0

J,ecnailpmocssoL ′′ )aPM(

(,erutarepmetkaeP ° )C 5.92 9.92 5.92 8.91 7.42

ecnailpmockaeP 0701 0911 0571 0373 00711

52ta °C 0601 0611 0371 0363 00711

051ta °C 007 0001 069 0431 0091


408


In order to demonstrate that HER compared to butanediol-based elastomer has betterphysical and mechanical properties at elevated temperatures, cast elastomers with 95%stoichiometry have been made from an MDI terminated polyester prepolymer (BaytecMS-242) extended with butanediol and HER-HP (from INDSPEC Chemical Corporation).Test specimens were cut from a sheet post cured at 110 °C for 16 hours, for the tensile,tear, DSC and DMA determinations. In the preparation of material for the dry heataging evaluation, tensile, DSC and DMA samples were placed in an air oven with thetemperature controlled within ± 2 °C of the set point and aged for 28 days at 100 °C, 21days at 120 °C and 14 days at 135 °C. Then, all the samples were stored for at leastseven days at room temperature and 50% relative humidity before testing.

8.8.1 Hardness Measurements

The Durometer hardness measured on the elastomers before and after the heat aging isgiven in Table 8.16. Since hardness is the measure of hard segment content of the elastomer,the unaged HER elastomer showed a 96A (Shore) hardness compared to 88A for thebutanediol elastomer clearly suggesting a higher hard segment content with HERextension. On heat ageing, the HER elastomer maintained its hardness even after 21 and14 days of aging at 120 °C and 135 °C, respectively. The butanediol extended elastomer’shardness was reduced by 4 Shore A units at all heat-ageing temperatures suggestingmaterial deterioration during the heat treatment.

8.8.2 Tensile Measurements

Polyurethane elastomers are known for their high elongation, tensile strength and modulusproperties. The combination of these properties provides toughness and durability infabricated parts. Cast elastomers extended with butanediol can maintain these tensileproperties when the use temperature is about 80 °C. When these elastomers are subjectedto higher temperatures, reduction in the tensile properties are observed due to theweakening of physical bonds in the elastomer.

The tensile properties measured on the unaged and heat-aged elastomer materials madefrom HER and butanediol are given in Table 8.16.

The 100, 200 and 300% tensile modulus values of unaged elastomers are higher forHER than for the butanediol elastomer. Analysing the 100, 200 and 300% modulusvalues of 100 °C heat-aged samples showed that HER elastomer retained about 95-97%of its original unaged values compared to 75-76% for butanediol. At 120 °C, there was93-94% modulus retention for HER whilst 85-88% was observed for butanediol. When

409

242-SM

cetyaB

morfsre

motsaletsac

foseitreporplacinahce

mdna

gniegatae

H61.8

elbaT

PH-

RE

Hloidenatu

B-4,1

snoitidnoC

gniegA

taeH

(erutarep

meT

°)

CT

R001

021531

TR

001021

531

gniegata

syaD

enoN

8212

41eno

N82

1241

A-erohS,ssendraH

6949

6969

8848

4848

)aPM(

ytreporpelisne

T

suludoM

%00169.01)001(

83.01)59(

53.01)49(

68.9)09(

85.6)001(

39.4)57(

43.5)18(

0.5)67(

suludoM

%00235.41)001(

57.31)59(

66.31)49(

48.21)88(

15.9)001(

53.7)77(

40.8)58(

73.7)77(

suludoM

%00364.81)001(

53.71)49(

71.71)39(

79.51)88(

40.51)001(

04.11)67(

42.31)88(

83.11)67(

htgnertselisne

T91.82)001(

82.72)79(

51.62)39(

59.12)87(

27.64)001(

08.83)38(

33.24)09(

16.53)67(

noitagnolE

%916

196536

646435

246006

916

)m/

Nk(kaerb

otygren

E)evruc

niarts-ssertsrednu

aera(011

421111

40129

00189

19

)m/

Nk(C

eiD,htgnerts

raeT

241dn

dndn

501dn

dndn

.gniegataeh

ehtretfa

deniaterseitreporp

fosegatnecrep

ehtera

sisehtnerapeht

niseulav

ehT:eto

Ndeni

mretedto

N:dn


410


the aging temperature is increased to 135 °C, HER elastomer showed better modulusretention property than the butanediol (87% versus 77%).

Although the butadiene elastomer has an unaged tensile strength of 46.72 MPa, thisvalue drops to 38.80 MPa (83% retention) when this material is heated for 28 days at100 °C. On the other hand, the HER extended elastomer retains 97% of its original(unaged) tensile strength under the same temperature heat-aging conditions, indicatingits excellent high temperature stability.

In spite of the severe aging conditions, HER extended elastomer maintained its excellent(greater than 600%), tensile elongation property. The combination of higher modulusand good elongation of the HER-based elastomer clearly demonstrates its ability towithstand higher tensile loads without the risk of failure.

Tensile measurements can also be used to determine the fracture energy of the material.The area under the stress-strain curve is the measure of fracture energy which is recognisedas a measure of toughness. As can be seen in Table 8.16, HER extended elastomer showedhigher fracture energy before and after heat aging suggesting that it has a higher toughnessthan the butanediol elastomer.

Tear strength is a property commonly used to determine the cut growth propagation inan elastomeric material. The unaged tear strength is much higher for HER than thebutanediol elastomer. The combination of fracture energy and tear strength propertiesclearly predicts that the HER elastomer will have better wear resistance than the butanediolextended elastomer.

8.8.3 Differential Scanning Calorimetric Analysis

From the DSC analysis, the thermal stability of polyurethanes can be determined by theendothermic transition temperatures associated with the melting of hard segment domains.The results are summarised in Table 8.17 and also shown in Figure 8.15. The first heatingDSC endothermic peaks for HER extended elastomer are broader and appeared at highertemperatures than the butanediol elastomer. Higher endothermic peak energy suggeststhat larger size crystalline hard segment domains exist within the HER elastomer.Considering the molecular structures of HER and butanediol chain extenders, there isno doubt that the higher thermal stability of the HER elastomer is associated with theformation of large size hard segments from the aromatic isocyanate and extender reaction.

On cooling the heated samples, exothermic peaks, due to the crystallisation of hardsegment domains, were seen in the DSC curves. Again, the higher exothermic peaktemperature suggested the existence of larger size hard segments for HER elastomer.

411

morfsremotsaletsacdegataehfostluserCSD71.8elbaTremyloperp242-SMcetyaB

gnitaeHtsriF nwoDgnilooC gnitaeHdnoceS

mrehtodnE mrehtoxE mrehtodnE

kaeP ygrenE kaeP ygrenE kaeP ygrenE

°C g/J °C g/J °C g/J

loidenatuB4,1:rednetxEniahC

(erutarepmeTgniegA ° )C

TR 371 01 07 5 061 4

022 1

001 671 11 07 5 861 2

021 571 9 27 2 661 3

531 081 7 DN DN

712 5.0< 071 2

PH-REH:rednetxEniahC

(erutarepmeTgniegA ° )C

TR 671 71 67 21 971 11

402 5.0<

001 691,971 42 28 21 781 11

512 5.0< 512 5.0<

021 391,871 02 18 11 781 11

912 5.0<

531 591,971 12 08 8 681 11

:tnemirepxECSD522ot52morfelpmasehttaeH.1 ° 01taC ° )gnitaehtsrif(nim/C

52ot522morflooC.2 ° 01taC ° )nwodgnilooc(nim/C522ot52morfelpmasehttaeH.3 ° 01taC ° )gnitaehdnoces(nim/C

During the second heating, DSC curves showed that a well-defined endothermic meltingpeak appeared at higher temperature confirming the higher HER elastomer thermalstability. Based on the DSC results, HER extension produced about 20 °C higher hardsegment melting temperature indicating the superior high temperature properties of HERcompared to butanediol elastomer.


412



DMA measures the ability of the material to store and dissipate mechanical energy. DMAproperties of the elastomeric materials are important because they often correlate toactual field performance. This method determines the storage modulus (G´, elasticbehaviour), loss modulus (G´´, energy dissipation), tan δ, loss compliance (J´´) and Tg

values. In this work, the DMA method was effectively utilised to understand theadvantages of HER over the butanediol extended elastomer. The results are summarisedin Table 8.18.

8.8.4.1 Storage Modulus Property (G´)

Storage modulus quantitatively measures the material’s elastic properties and alsoqualitatively determines the elastomer’s stiffness and hardness. The unaged elastomerstorage modulus values are higher for HER at 25, 100 and 150 °C temperatures comparedto butanediol. On heat ageing at 100 °C, the HER elastomer retained about 85-86% ofits original G´ value compared to 60-85% for butanediol. After ageing for 28 days at

Figure 8.15 DSC curves of 135 °C heat aged urethane elastomers from Baytec MS-242

413

remyloperp

242-SM

cetyaB

morfenahteruylop

tsacdega

taehfo

AM

D81.8

elbaT

PH-

RE

Hloidenatu

B-4,1

(erutarep

meT

gniegA

°)

CT

R001

021531

TR

001021

531

G,suludoM

egarotS′

)aPM(

52ta

°C1.33

9.8203

4.929.01

3.97.9

1.9

001ta

°C2.12

8181

8.212.7

1.53.5

9.3

051ta

°C4.41

4.217.21

9.81.6

7.39.3

8.2

G,suludoM

ssoL

′′)aP

M( T,erutarepmet

kaePg

(°)

C8.92-

6.92-03-

6.92-4.72-

6.52-82-

62-

suludom

kaeP171

861761

561761

711461

951

atleD

naT

T,erutarepmet

kaePg

(°)

C9.32-

7.32-8.12-

6.32-3.91-

61-8.71-

7.71-

atlednat

kaeP293.0

404.0104.0

683.0756.0

826.056.0

326.0

52ta

°C270.0

80.0380.0

190.0150.0

780.080.0

301.0

001ta

°C910.0

230.0330.0

460.0420.0

180.0460.0

790.0

051ta

°C920.0

250.0150.0

80.0720.0

570.0660.0

190.0

J,ecnailpmoc

ssoL

′′)aP

M(

(,erutarepmet

kaeP°

)C

6.11-1.21-

4.9-31-

3.9-6-

7.7-6.3-

ecnailpmoc

kaeP0852

00030182

00620641

04610351

00151

52ta

°C0612

07720572

09030274

02290128

00211

001ta

°C009

08710581

00050533

08510911

00152

051ta

°C0302

01240404

04980544

000020961

00023

051>

—0:aera

kaeP°C

2243

5376

75091

061292

051>

—05-:aera

kaeP°C

8214

1437

28612

581613


414


100 °C temperature, a very sharp decline in G´ values (60%) for butanediol and a highretention for HER were observed at 150 °C.

For 120 °C heat-aged samples, 85-90% retention of G´ values for HER compared to 63-89% for butanediol were seen. The same kind of trend is seen with 135 °C heat-agedelastomer storage modulus values. Thus the HER extended elastomer has better retentionof storage modulus at higher temperatures than the butanediol elastomer.

8.8.4.2 Loss Modulus Property (G´´)

Loss modulus measures the energy dissipation in the elastomer. The Tg measured fromthe loss modulus curves (Table 8.18) showed lower values for the HER than butanediolelastomer suggesting good phase separation.

8.8.4.3 Tan Delta Property (tan δ)

The soft segment Tg is often defined as the maximum in the tan δ versus temperaturecurve which determines the low temperature behavior of urethane elastomers. The resultsobtained are summarised in Table 8.18. The Tg obtained from the tan δ curves confirmsagain that a good phase separation existed with HER elastomer. As a consequence, theHER-based polyurethane has a higher flexibility at low temperature.

Although the unaged 25 °C tan δ value appeared to be higher for HER, the effect of heatageing showed an increase of 26% compared to 101% for BD. Similarly, the 100 °C and 135°C heat-aged samples showed very high percentages of elevated temperature (100 °C and150 °C) tan δ value increases in butanediol-based elastomer as compared to the HER material.

These experimentally determined tan δ values from the DMA analysis strongly suggestthat more heat generation problems can be anticipated with butanediol-based elastomerwhen they are exposed to temperatures above 80 °C than with HER elastomer.

8.8.4.4 Loss Compliance Property (J´´)

Figures 8.16 and 8.17 show the plots of loss compliance for butanediol and HERelastomers, respectively and the results are summarised in Table 8.18.

The loss compliance values of unaged and heat-aged butanediol extended elastomer areabout 3-5 times greater than the HER elastomer.

415

Figure 8.17 Heat ageing study - loss compliance of MS-242-HER elastomer

Figure 8.16 Heat ageing study - loss compliance of MS-242-butanediol elastomer


416


In addition to the loss compliance data at various temperatures, the area under the losscompliance curve can also be used to predict the tendency of the elastomeric materials toconvert mechanical energy to heat. A smaller area means less mechanical energy isconverted to heat, indicating better dynamic performance of the elastomer over thetemperature range of interest.

By comparing the loss compliance peak areas of heat-aged samples, there is no doubt theHER extended elastomer is far superior in performance to butadiene elastomer.

Based on the current study, HER extended elastomers has the following room temperatureand high temperature physical and mechanical properties advantages over butanediol-based elastomers:

• Excellent retention of hardness.

• Retention of high tensile modulus and elongation.

• Higher fracture energy.

• High tear strength.

• Higher thermal stability due to high crystalline hard segment melting.

• High flexibility due to low Tg.

• High storage modulus.

• Lower hysteresis and compliance values.

• Broader service temperature.

8.9 Conclusions

The HER materials developed for polyurethane and other applications have a widerange of melting characteristics. Lower melting temperatures of these materials canprovide improved processability in the development of cast and thermoplasticpolyurethane elastomers. The high MW diols present in the HER materials not onlyreduce the melting point of the HER extender, but also exhibit a profound influence onthe concentration of crystalline hard segments and, as a result, ultimately affect thephysical and mechanical properties of the elastomers. Polyurethanes with varyingproperties can be obtained by using HER materials containing different concentrationsof high MW diols. Since the high MW diols can act as plasticisers, the addition ofexternal plasticisers can be avoided by use of these materials. Cast polyurethanes showgood tensile, tear, compression set and hysteresis properties with HER-based chainextenders. This results in a wide choice in the selection of a suitable aromatic diolchain extender from the HER materials family.

417

Acknowledgements

The author would like to thank Jason Burchianti and Stephen J. Ondrey for preparingelastomer samples and for conducting pot life measurements, Richard F. George for theDMA, physical properties and DSC work and Vaughn J. Romell for the FT-IR analysiswork. My special thanks to Fred M. Covelli, V.P. Technology, and Joseph de Almeida,Marketing Manager of Specialty Products, for their continuous encouragement in thiswork.

References

1. K. C. Frisch and S. L. Reegen in Advances in Urethane Science and Technology,Volume 2, Eds., K. C. Frisch and D. Klempner, Technomic Publishing Co., Inc.,Lancaster, PA, 1973, 29.

2. D. Klempner and K. C. Frisch in Advances in Urethane Science and Technology,Volume 8, Eds., K. C. Frisch and D. Klempner, Technomic Publishing Co., Inc.,Lancaster, PA, 1981, 93.

3. J. E. Tiedemann, Presented at the Fall PMA Meeting, Miami, FL, USA, 1993.

4. M. A. Mendelsohn, F. W. Navish and D. Kim, Rubber Chemistry andTechnology, 1986, 58, 5, 997.

5. R. B. Durairaj, J. Burchianti, S. J. Ondrey, R. F. George and J. de Almeida, PaperPresented at the Fall PMA Meeting, Baltimore, MD, USA, 1998.

6. ASTM D2240-00Standard Test Method for Rubber Property-Durometer Hardness.

7. ASTM D412-98aStandard Test Methods for Vulcanized Rubber and Thermoplastic Rubbers andThermoplastic Elastomers-Tension.

8. ASTM D624-00e1Standard Test Method for Tear Strength of Conventional Vulcanized Rubber andThermoplastic Elastomers.

9. ASTM D2632-96Standard Test Method for Rubber Property-Resilience by Vertical Rebound.


418


10. ASTM D395-98Standard Test Methods for Rubber Property-Compression Set.

11. Polyurethane Handbook, 2nd Edition, Ed., G. Oertel, Hanser Publications,Munich, 1994.

12. R. Palinkas, Paper Presented at the Fall PMA Meeting, Cincinnati, OH, USA, 1990.

13. J. E. Doyle, Paper Presented at the PMA Meeting, Milwaukee, WI, USA, 1994.

14. S. M. Clift, Paper Presented at the Fall PMA Meeting, Charlotte, NC, USA, 1991.

15. J. R. Lin and L. W. Chen, Journal of Applied Science, 1998, 69, 8, 1575.

16. C. Hepburn, Polyurethane Elastomers, 2nd Edition, Elsevier Applied Science,London, 1992.

17. S. Madan, S. Franyutti, K. Recker and R. S. Pantone, Presented at thePolyurethanes World Congress ‘97, Amsterdam, The Netherlands, 1997, 381.

18. R. W. Seymour and S. L. Cooper, Journal of Polymer Science, Part B, PolymerLetters, 1971, 9, 7, 689.

19. T. R. Hesketh, J. W. C. van Bogart and S. L. Cooper, Polymer Engineering andScience, 1980, 20, 3, 190.

20. L. M. Leung and J. T. Koberstein, Macromolecules, 1986, 19, 3, 706.

21. G. Pompe, A. Pohlers, P. Potschke and J. Pionteck, Polymer, 1998, 39, 21, 5147.

22. Eastman HQEE-80, As a Chain Extender for Liquid Castable UrethaneElastomers, Industrial Chemical Division, Eastman Chemical Products, Inc.,Kingsport, TN-37662, USA, Publication No. D-161, December 1982.

23. M. Palmer, Paper Presented at the Spring PMA Meeting, Kansas City, KY, 1983.

24. M. Palmer, J. H. Davis, T. L. Douglas and M. R. Wilhelm, Paper Presented at theFall PMA Meeting, Chicago, IL, USA, 1977.

25. J. W. C. van Bogart, A. Lilaonitkul, L. E. Lerner and S. L. Cooper, Journal ofMacromolecular Science B, 1980, 17, 2, 267.

26. J. W. C. van Bogart, P. E. Gibson and S. L. Cooper, Journal of Polymer Science:Polymer Physics, 1983, 21, 1, 65.

419

27. C. B. Hu, R. S. Ward, Jr., and N. S. Schneider, Journal of Applied PolymerScience, 1982, 27, 6, 2167.

28. R. Bonart, L. Morbitzer and E. H. Muller, Journal of Macromolecular Science,1974, B9, 3, 447.

29. J. Blackwell, J. R. Quay, M. R. Nagarajan, L. Born and H. Hespe, Journal ofPolymer Science: Polymer Physics, 1984, 22, 7, 1247.

30. R. W. Seymour, G. M. Estes and S. L. Cooper, Macromolecules, 1970, 3, 5, 579.


420


421

9 Ultra-Low Monol PPG: High-PerformancePolyether Polyols for Polyurethanes

Stephen D. Seneker, Nigel Barksby and Bruce D. Lawrey

9.1 Introduction

Polyoxypropylene glycols (PPG) have been available since the early 1960s.Polyurethanes based on these polyols are used in a diverse range of applicationsincluding flexible foams (moulded and slabstock), sealants and adhesives. Untilrecently, their use in high-performance applications such as cast and thermoplasticurethane elastomers has been limited due to their inferior physical properties andprocessing characteristics. These problems can be largely attributed to the considerableamount of monol (monofunctional hydroxyl-containing species) present in thesepolyols. Monol acts as a chain terminator, which severely reduces both the molecularweight (MW) build during polymer formation (processability) and the ultimatepolymer MW (physical properties).

Conventional polyoxypropylene diols are produced commercially through the base-catalysed (KOH) propoxylation of glycol starters such as propylene glycol. However,the base catalyses not only the addition of propylene oxide to the growing polyolmolecule, but also a side reaction in which propylene oxide isomerises to allyl alcohol[1, 2, 3]. Allyl alcohol acts as a monofunctional starter resulting in propoxylatedallyl alcohol (‘monol’) as shown in Figure 9.1. Since each monol chain contains aterminal double bond, the amount can be quantified by measuring the unsaturationlevel. The level of unsaturation is typically reported in milliequivalents of unsaturationper gram of polyol (meq/g).

Table 9.1 shows the relationship between conventional polyoxypropylene diol MWand unsaturation level, functionality, and monol content (mole%). As the MW ofconventional PPG increases, there is a dramatic increase in the monol content. Forexample, a PPG with a MW of 2000 typically has an unsaturation or monol level ofabout 0.03 meq/g, corresponding to a functionality of 1.94, while a diol with a MWof 4000 has a monol content of 0.09 meq/g and a functionality of only 1.69. Aconventional diol with a MW of 4000 is considered to be a mixture of about 70%diol and 30% monol (molar basis). This is why the practical upper limit for commercialPPG is a 2000 equivalent weight (4000-MW diol or 6000-MW triol).

422


Figure 9.1 Reaction of propylene oxide to form polyoxypropylene diol and monol

folonom%elomdnaytilanoitcnuf,)noitarutasnu(lonoM1.9elbaTsloylopGPPlanoitnevnoc

loiDGPP #HO )g/qem(.tasnU ytilanoitcnuF lonoM%eloM

WM0001 111 10.0 99.1 1

WM0002 65 30.0 49.1 6

WM0003 73 50.0 68.1 41

WM0004 82 90.0 96.1 13

*WM0008 41 *— *— *—

yllaicremmocgnisutsylatacHOKlanoitnevnochtiwGPPWM-0008naecudorptonnaC*ehtfoerusaemasihcihw,rebmunlyxordyhehtsi#HO.snoitidnocnoitcaerelbisaefehtybthgiewtnelaviuqeloylopehtotdetalersitI.spuorglyxordyhfonoitartnecnoc

.thgiewtnelaviuqe=#HO/001,65:noitauqe

423

Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes

Advances in catalyst technology demonstrated that PPG could be prepared withsignificantly lower monol contents using double metal cyanide (DMC) catalysts such aszinc hexacyanocobaltate [4, 5, 6, 7, 8]. This technology was originally developed byGeneral Tire Inc., [9, 10, 11]. These DMC catalysts produce PPG with monol contentstypically in the range of 0.015 to 0.020 meq/g, which corresponds to a diol with a MWof 4000 with a functionality of 1.94. The synthesis and utility of these low monol PPGpolyols in polyurethanes have been reported extensively [6, 7, 8, 12, 13, 14]. Althoughthe physical properties and processability of polyurethanes improved significantly withlow monol PPG polyols, the cost to performance ratio was still considered unacceptable;therefore, these polyols achieved limited commercial success.

Further improvements in catalyst technology resulted in Lyondell (formerly ARCOChemical Company) commercialising a new family of PPG with ultra-low monol contentsdesignated Acclaim Polyols. Table 9.2 summarises the unsaturation, functionality, andmonol contents of these polyols. Comparing Tables 9.1 and 9.2, one can clearly see thedramatic improvements in functionality. For example, a 4000-MW, ultra-low monolPPG (Acclaim Polyol 4200) has a functionality of 1.98 versus 1.69 for a conventional4000-MW PPG diol.

In this chapter, the dramatic effect that monol content or polyol functionality has onprocessability and properties of polyurethane cast elastomers is discussed. This effect isshown for elastomers prepared by both the prepolymer and one-shot processes. Furtherimprovements in elastomer processability and formulating latitude can be achieved byincorporating oxyethylene moieties into the polyol backbone. For one-shot elastomerprocesses, ultra-low monol PPG polyols capped with ethylene oxide have beencommercialised.

mialccAfolonom%elomdnaytilanoitcnuf,)noitarutasnu(lonoM2.9elbaT)sloylopGPPlonomwol-artlu(sloylop

)WM(loyloP #HO )g/qem(.tasnU ytilanoitcnuF lonoM%eloM

*GPPWM-0001 211 500.0 599.1 5.0

0002 65 500.0 99.1 1

**0003 53 400.0 99.1 1

0004 82 500.0 89.1 2

0008 41 500.0 69.1 4

elpmaslatnemirepxE*.seiteiomenelyhteyxotnecrepthgiew02gniniatnocloidenelyhteyxo-enelyporpyxoyloP**

rebmunlyxordyh:#HO

424


This chapter also focuses on the effect of the polyol molecular weight distribution (MWD)or polydispersity on mechanical and dynamic properties of polyurethanes andpolyurethane/ureas. Ultra-low monol PPG polyols are unique in that they have a verynarrow MWD in comparison to other high-performance polyols such aspolytetramethylene ether glycol (PTMEG) or polyester polyols. Narrow polydispersityresults in lower viscosity polyols as well as lower viscosity isocyanate-terminatedprepolymers based on those polyols. However, a not so obvious effect is that polyolswith a narrow MWD result in polyurethanes with significantly different mechanical anddynamic property profiles. These differences will be highlighted in various polyurethaneand polyurethane/urea systems and explained using dynamic mechanical thermal analysis(DMTA). By understanding the effect of polyol MWD, we will illustrate how thepolydispersity can be adjusted in order to maximise the mechanical and dynamic propertiesobtained from ultra-low monol PPG polyols.

The experimental procedures utilised to prepare the various polyurethane andpolyurethane/urea polymers described in this chapter are discussed in the Appendix.

9.2 MDI/BDO Cured Elastomers Based on Ultra-Low Monol PPG Polyols

9.2.1 Effect of Monol Content on 4,4´-Methylene DiphenylmethaneDiisocyanate (MDI)/1,4-Butanediol (BDO) Cured Elastomers

The importance of monol content in 4,4´-MDI/BDO cured elastomer systems was determinedby comparing 4000-MW PPG diols prepared via ultra-low monol technology, DMC andpotassium hydroxide. They are designated as ultra-low monol, low monol and conventionaland have monol contents of 0.005, 0.016 and 0.085 meq/g, respectively. This correspondsto functionalities of 1.98, 1.94 and 1.71. We prepared 6% NCO 4,4´-MDI prepolymersand chain extended with BDO at an isocyanate to hydroxyl ratio (NCO:OH) of 1.03 [15].Table 9.3 summarises the monol effect on elastomer processing characteristics (pot lifeand demould time), and physical properties.

The monol content has a major effect on elastomer processability. Lower monol contentresults in a faster molecular weight build, which reduces demould time. Elastomers basedon the 4000-MW, ultra-low monol PPG had demould times as short as 20 minutes,whereas, the low monol PPG gave demould times of over one hour. The elastomer basedon conventional polyol had a demould time greater than three hours.

Lower monol content clearly has a positive effect on all elastomer physical properties.Dramatic property improvements are even seen when comparing elastomers based onultra-low monol and low monol polyol. This is particularly surprising considering the

425

,lonomwol-artlunodesabsremotsaleenahteruylopfonosirapmoC3.9elbaTsloidWM-0004lanoitnevnocdnalonomwol

)derucODB;sremyloperpIDMOCN%6(

)g/qem(tnetnoClonoM 500.0 610.0 580.0

ytilanoitcnuFloyloP 89.1 49.1 17.1

s-aP,ytisocsiVremyloperP

02 °C 5.8 5.7 1.8

04 °C 2.2 9.1 0.2

06 °C 87.0 96.0 37.0

08 °C 83.0 23.0 83.0

SCITSIRETCARAHCGNISSECORP

)setunim(efiltoP 3-1 3-1 3-1

)setunim(emiTdluomeD 54-02 021-06 081>

SEITREPORPREMOTSALE

AerohS,ssendraH A27 *)A66(A86 *)A83(A15

%,dnuobeRmuludneP 07 26 84

%,noitagnolEetamitlU 019 0201 031

aPM,htgnertSelisneT 6.02 4.21 4.1

aPM,suludoM%001 4.3 0.3 6.1

aPM,suludoM%002 9.4 3.4 —



m/Nk,raeTCeiD 76 94 01

07@h22(%,teSnoisserpmoC ° )C 52 04 35

aPM,noitcelfeDnoisserpmoC

%5 8.0 4.0 1.0

%01 2.1 8.0 2.0

%51 7.1 2.1 3.0

%52 8.2 9.1 8.0* sisehtnerapnidetacidnieulavehtotelpmasehtotnidetartenepylwolseldeenssendraHerohS


426


small difference in functionality (1.98 versus 1.94). For example, Shore A hardnessincreased from 68 to 72; rebound from 62 to 70%; tensile strength from 12 to 21MPa; Die C tear strength from 49 to 67 kN/m; and compression set decreased from 40to 25%. The load bearing capabilities improved significantly as shown by thecompression deflection results. The stress/strain curves in Figure 9.2 clearly show thatthe elastomer based on ultra-low monol PPG exhibits much greater modulus buildwith increasing elongation.

Figure 9.2 Elastomer stress/strain curves(6% NCO MDI/4000-MW diol prepolymers; BDO cured)

427

Reducing the monol content also improves the dynamic properties of cast elastomers.Figure 9.3 shows the DMTA response for the two polymers described above. The polymerderived from the ultra-low monol polyol has a flatter rubbery plateau region. The highermodulus in the rubbery plateau is consistent with the polymer’s higher tensile modulus.The substantial reduction in tan delta (δ) across the entire temperature range should alsobe noted. Lower tan delta translates into improved performance in dynamic applicationsdue to lower heat buildup and improved rebound as noted above.

Figure 9.3 Elastomer DMTA curves(6% NCO MDI/4000-MW diol prepolymers; BDO cured)


428


All these improvements indicate that significant changes are occurring in the elastomeras the monol content is reduced and the polyol functionality increased from 1.94 to1.98. We believe the benefits are directly attributable to an increase in the polymermolecular weight. Figure 9.4 shows the relationship between functionality and theoreticalelastomer molecular weight (calculated from the Carothers equation). Raising thefunctionality from 1.94 to 1.98 increases the ultimate polymer molecular weight by afactor of three from 82,000 to 270,000. This increase in MW could easily account forthe observed improvements in processability and properties.

Figure 9.4 Effect of polyol functionality on theoretical elastomer molecular weight

429

9.2.2 Processability and Property Latitude of Elastomers Based on Ultra-LowMonol PPG Polyols

Ultra-low monol PPG diols are commercially available in molecular weights from 2000to 12,000. A product summary is given in Table 9.4. To determine the formulating rangeof the various diols in MDI/BDO cured elastomers, we prepared MDI prepolymers basedon the 2000, 4000 and 8000 MW diols at a wide range of isocyanate contents. Theseprepolymers were chain extended with BDO to produce elastomers with a broad rangeof properties. Tables 9.5 to 9.7 summarise the prepolymer compositions and viscosities,processing characteristics and physical properties of these elastomers.

eniltcudorploylopmialccA4.9elbaT

mialccAloyloP

raluceloMthgieW

noitisopmoC ytilanoitcnuF epyTlyxordyH

0022 0002 loiDOP 99.1 yradnoceS

5023 0023 loiDOP/OE 99.1 yradnoceS




0036 0006 loirTOP 79.2 yradnoceS

0222 0522 loiDdeppaC-OE 99.1 )%58>(yramirP

0224 0004 loiDdeppaC-OE 79.1 )%58>(yramirP

0236 0006 loirTdeppaC-OE 49.2 )%58>(yramirP


430


GPPlonomwol-artluWM-0002nodesabsremotsaleenahteruyloP5.9elbaT)derucODB;sremyloperpIDM(

NOITISOPMOCREMYLOPERP

remyloperPOCN% 0.4 0.6 0.8 0.01 0.21

HO/OCN 2.2 0.3 8.3 8.4 0.6

)wbp(IDM-´4,4 7.72 9.63 7.74 2.06 0.57

)wbp(GPPWM-0002 001 001 001 001 001


02 °C 52 51 5.7 0.4 3.3

04 °C 6.4 7.2 54.1 39.0 85.0

06 °C 54.1 58.0 94.0 33.0 22.0

08 °C 47.0 04.0 52.0 51.0 1.0

GNISSECORPSCITSIRETCARAHC

56@remyloperP ° nitlytubid;30.1=HO/OCN;TR@ODB;C001@sdluoM;tsylatacetarualid ° 001@sruoh61deruc;C °C

)wbp(remyloperP 001 001 001 001 001

)wbp(ODB,loidenatub-4,1 2.4 2.6 3.8 4.01 5.21

)setunim(efiltoP 3-1 3-1 3-1 3-1 3-1

)setunim(emiTdluomeD 54-02 54-02 54-02 06-03 021-06

SEITREPORPREMOTSALE 32@skeew4denoitidnocselpmaS ° ytidimuhevitaler%05/C

AerohS,ssendraH A75 A86 A08 A98 )D54(A79

%,dnuobeRmuludneP 06 16 16 65 45

%,noitagnolEetamitlU 018 017 007 046 085

aPM,htgnertSelisneT 9.51 1.22 2.62 5.52 4.32

aPM,suludoM%001 4.1 5.3 7.5 5.8 7.11

aPM,suludoM%002 8.1 3.5 1.8 9.01 7.31

aPM,suludoM%003 5.2 4.7 3.01 0.31 4.51

aPM,suludoM%004 5.3 9.9 8.21 4.51 5.71

m/Nk,raeTCeiD 62 85 47 19 701

%,teSnoisserpmoC07@h22( ° )C

92 22 81 61 02


%5 3.0 5.0 9.0 4.1 1.2

%01 6.0 0.1 9.1 8.2 1.4

%51 8.0 6.1 6.2 1.4 1.6

%52 5.1 8.2 7.4 8.6 4.01

erutarepmetmoor:TR;thgiewybstrap:wbp

431

GPPlonomwol-artlu,WM-0004nodesabsremotsaleenahteruyloP6.9elbaT)derucODB;sremyloperpIDM(


remyloperPOCN% 0.4 0.6 0.8 0.01

HO/OCN 3.3 7.4 3.6 2.8

)wbp(IDM-´4,4 6.02 4.92 5.93 3.15

)wbp(GPPWM-0004 001 001 001 001


02 °C 61 5.8 5.4 1.3

04 °C 8.3 2.2 52.1 98.0

06 °C 23.1 87.0 5.0 63.0

08 °C 36.0 83.0 52.0 81.0


56@remyloperP ° nitlytubid;30.1=HO/OCN;TR@ODB;C001@sdluoM;tsylatacetarualid ° 001@sruoh61deruC;C °C

)wbp(remyloperP 001 001 001 001

)wbp(ODB,loidenatub-4,1 2.4 3.6 3.8 4.01

)setunim(efiltoP 3-1 3-1 3-1 3-1

)setunim(emiTdluomeD 54-02 54-02 06-03 021-06

SEITREPORPREMOTSALE 32@skeew4rofdenoitidnocselpmaS ° ytidimuh%05/C

AerohS,ssendraH A06 A27 A28 A19

%,dnuobeRmuludneP 07 07 56 06

%,noitagnolEetamitlU 098 019 079 096

aPM,htgnertSelisneT 9.51 7.02 2.71 2.51

aPM,suludoM%001 9.1 4.3 5.4 6.6

aPM,suludoM%002 9.2 9.4 2.6 7.8

aPM,suludoM%003 0.4 3.6 6.7 1.01

aPM,suludoM%004 3.5 8.7 9.8 3.11

m/Nk,raeTCeiD 74 76 57 88

m/Nk,raeTtilpS 7.4 8.8 6.9 2.41


91 52 91 12


%5 3.0 8.0 0.1 5.1

%01 6.0 2.1 8.1 0.3

%51 9.0 7.1 5.2 3.4

%52 7.1 8.2 0.4 3.7

erutarepmetmoor:TR;thgiewybstrap:wbp


432


-0006fotceffE:GPPlonomwol-artluWM-0008nodesabsremotsalE7.9elbaT)derucODB;sremyloperpIDM(noitiddaloirtlonomwol-artluWM

NOITISOPMOCREMYLOPERPremyloperPOCN% 0.2 0.3 0.4

HO/OCN 9.2 1.3 0.4 2.4 1.5 5.5)wbp(IDM-´4,4 9.9 6.9 5.31 2.31 4.71 1.71

)wbp(loirtGPPWM-0006 6.2 — 6.2 — 6.2 —)wbp(loidGPPWM-0008 4.79 001 4.79 001 4.79 001s-aP,ytisocsiVremyloperP

02 °C 27 05 04 53 42 22

04 °C 81 21 01 9 4.6 9.506 °C 7.6 5.4 7.3 3.3 4.2 3.208 °C 2.3 7.1 6.1 5.1 1.1 38.0


56@remyloperP ° nitlytubid;30.1=HO/OCN;TR@ODB;C001@sdluoM;tsylatacetarualid ° 001@sruoh61deruC;C °C

)wbp(remyloperP 001 001 001 001 001 001)wbp(ODB,loidenatub-4,1 1.2 1.2 1.3 1.3 2.4 2.4

)setunim(efiltoP 3-1 3-1 3-1 3-1 3-1 3-1)setunim(emiTdluomeD 06-03 081> 54-02 54-02 54-02 54-02

SEITREPORPREMOTSALE 32@skeew4rofdenoitidnocselpmaS ° ytidimuh%05/CAerohS,ssendraH A43

*)A23(A72

*)A61(A94 A54

*)A14(A85 A65

*)A35(%,dnuobeRmuludneP 95 55 86 56 17 96%,noitagnolEetamitlU 0301 **0031> 0311 0911 029 069aPM,htgnertSelisneT 0.4 9.2 7.21 1.01 2.31 8.11

aPM,suludoM%001 5.0 5.0 2.1 2.1 7.1 6.1

aPM,suludoM%002 7.0 5.0 8.1 7.1 5.2 3.2aPM,suludoM%003 8.0 6.0 4.2 2.2 4.3 1.3aPM,suludoM%004 0.1 7.0 1.3 9.2 4.4 0.4

m/Nk,raeTCeiD 61 31 24 24 94 74


43 35 22 04 91 12

aPM,noitcelfeDnoisserpmoC%5 1.0 1.0 2.0 2.0 3.0 3.0

%01 3.0 3.0 4.0 4.0 6.0 6.0

%51 4.0 4.0 6.0 6.0 0.1 8.0%52 7.0 7.0 0.1 0.1 7.1 5.1

* sisehtnerapnidetacidnieulavehtotelpmasehtotnidetartenepylwolseldeenssendraHerohS** gnitsetgnirudkaerbtondidselpmaS

433

A review of the elastomer processability and physical property data indicates that thereis an optimum hardness range for each polyol MW. Based on the demould time and thepercent compression set, the following elastomer Shore hardness ranges are recommended.The lower hardness limits of these ranges can be extended by addition of low levels (lessthan 10%) of 6000-MW, ultra-low monol triol (Acclaim Polyol 6300).

Acclaim Polyol 2200 (2000-MW diol) 70 to 90 Shore AAcclaim Polyol 4200 (4000-MW diol) 55 to 75 Shore AAcclaim Polyol 8200 (8000-MW diol) 50 to 65 Shore A

The rebound of the elastomers based on 2000-MW, ultra-low monol PPG ranges from55 to 60%. For high-performance dynamic applications, use of higher MW diols arerecommended, which give rebound values up to 10 units higher, ranging from 65 to70%. Figure 9.5 shows the DMTA response for the 70 Shore A elastomers based on the2000, 4000 and 8000 MW diols. The glass transition temperature (Tg) for the elastomersbased on the higher MW diols is about 15 °C lower, which probably accounts for thehigher rebound of these materials.

Figure 9.5 DMTA curves of 70 Shore A elastomers: Effect of polyol molecular weight(6% NCO MDI prepolymers; BDO cured)


434


The 8000-MW, ultra-low monol PPG is used particularly in the development of soft,plasticiser-free elastomers [16, 17]. The use of this high MW, low polydispersity polyolallows for the preparation of low viscosity prepolymers with low isocyanate contents.Initial evaluations of these very soft elastomers showed lower than expected physicalproperties. The lower properties can be attributed to the very low hard-segment contentof these polymers. The lack of hard segment (physical crosslinks) can be compensatedfor by the incorporation of low levels of chemical crosslinks (triol) [18]. This isaccomplished by the addition of a 6000-MW, ultra-low monol triol (Acclaim Polyol6300) into the polymer matrix. Table 9.7 shows the triol effect on prepolymer viscosities,elastomer processability and physical properties. It should be noted that very low levelsof crosslinking are needed to improve the elastomer properties.

9.2.3 Processing Latitude Improves by Incorporating Oxyethylene Moieties

The ultimate goal for MDI/BDO cured elastomers was to design a polyol that wouldhave a combination of good processability and high rebound in the mainline hardnessrange of 80 to 95 Shore A. This was an elusive goal as one can see from the elastomerdata in Tables 9.5 and 9.6. A 90 Shore A elastomer based on the 4000-MW diol hadpoor processability but a high rebound, whereas, the 90 Shore A elastomer based on the2000-MW diol was processable but had a lower rebound.

We attributed the poor processability of the 90 Shore A elastomer based on the 4000-MW diol to the excessive phase-separation of the MDI/BDO hard segment [19]. Toalleviate the hard segment phase-separation problem, the compatibility of the MDI/BDOhard segment in the soft-segment phase was increased by incorporating oxyethylenemoieties into the polyol backbone. This technique dramatically enhanced the elastomerprocessability. The polyol MW and oxyethylene content were varied to obtain the bestpossible combination of processability, high hardness and high rebound. This optimisationprocess led to the development of Acclaim Polyol 3205, a 3200-MW diol containingapproximately 20 weight percent oxyethylene moieties. Table 9.8 shows the enhancedprocessing latitude of elastomers based on this polyol. In addition to enhancedprocessability, elastomer physical properties were improved. Acclaim Polyol 3205 is aunique material in that it allows the formulation of MDI/BDO cured elastomers from 75to 95 Shore A with excellent processability and high rebound.

435

5023loylopmialccAnodesabsremotsaleenahteruyloP8.9elbaT)derucODB;sremyloperpIDM(


remyloperPOCN% 0.4 0.6 0.8 0.01 0.21

HO/OCN 9.2 0.4 3.5 9.6 7.8

)wbp(IDM 4.22 2.13 5.14 5.35 7.76

)wbp(5023loyloPmialccA 001 001 001 001 001


02 °C 41 5.8 6.4 0.3 4.1

04 °C 2.3 9.1 0.1 8.0 56.0

06 °C 0.1 7.0 14.0 23.0 62.0

08 °C 5.0 3.0 91.0 61.0 41.0


56@remyloperP ° nitlytubid;30.1=HO/OCN;TR@ODB;C001@sdluoM;tsylatacetarualid ° 001@sruoh61rofderuc;C °C

)wbp(remyloperP 001 001 001 001 001

)wbp(ODB,loidenatub-4,1 2.4 2.6 3.8 4.01 5.21

)setunim(efiltoP 3-1 3-1 3-1 3-1 3-1

)setunim(emiTdluomeD 04-51 04-51 04-51 05-02 06-03

setunim03@htgnertSneerG tnellecxe tnellecxe tnellecxe doog egareva

SEITREPORPREMOTSALE 32@skeew2rofdenoitidnocselpmaS ° ytidimuhevitaler%05/C

AerohS,ssendraH )A63(A34 A86 A28 A09 A59




aPM,suludoM%001 5.1 0.3 9.4 1.7 8.9

aPM,suludoM%002 1.2 7.4 7.6 9.8 4.11

aPM,suludoM%003 2.3 6.6 4.8 8.01 7.21

aPM,suludoM%004 7.4 9.8 3.01 6.21 1.41

m/Nk,raeTCeiD 03 45 07 98 39

%,teSnoisserpmoC07@h22( ° )C 06 24 91 12 42


%5 2.0 5.0 0.1 6.1 2.3

%01 4.0 0.1 1.2 3.3 8.5

%51 6.0 4.1 0.3 7.4 8.7

%52 1.1 5.2 0.5 3.7 5.11


436


9.3 One-Shot Elastomer System Based on EO-Capped, Ultra-LowMonol PPG Polyols

The lower reactivity of secondary hydroxyl groups on the initial ultra-low monol PPGpolyol products limited their use to mainly prepolymer systems. PTMEG and polyesterpolyols have primary hydroxyl groups, allowing them to be easily formulated withcuratives such as 1,4-butanediol and reacted with polyisocyanates using a ‘one-shot’ or‘quasi-prepolymer’ approach. For this reason, high-reactivity versions of ultra-low monolPPG were developed. These polyols give the performance enhancements expected fromultra-low monol PPG with the added benefit of being one-shot processable [20, 21]. Thehigher reactivity is achieved by capping the propylene-oxide based polyols with ethyleneoxide (EO) resulting in products with a high percentage of primary-hydroxyl groups. Asignificant amount of work has been reported on polyurethanes based on EO-cappedlow monol PPG [20, 21, 22, 23, 24, 25].

9.3.1 Effect of Primary Hydroxyl Concentration on One-Shot ElastomerProcessability

A key factor in determining whether or not a polyol can be used in a one-shot process isreactivity. As a general rule, the polyol reactivity needs to be nearly equal to or greaterthan the reactivity of the curative. For example, most one-shot elastomers based on MDIuse BDO as a curative which contains primary hydroxyl groups. A typical PPG polyolcontains about 95% secondary hydroxyl groups. If one attempts to process a blend ofMDI, BDO and PPG, the MDI will react preferentially with the BDO to form an MDI-BDO hard segment. The MDI-BDO hard segment is incompatible in the system andphase separates before the PPG polyol has a chance to react. The end result is a low MW,two-phase system in which the MDI-BDO hard segment is dispersed in partially reactedpolyol, which results in very poor physical properties. Therefore, successful one-shotprocessing requires that the polyol and curative have similar reactivities.

The reactivity of PPG polyols can be increased by reacting with EO, which ring opens toform a primary hydroxyl group. A 100% primary hydroxyl content cannot be achievedusing this approach since EO has a tendency to form chains instead of distributing evenlyto each secondary hydroxyl group; however, it is well known that a 100% primarycontent is not required to achieve one-shot processability. The effect of the primaryhydroxyl content on the processability of a one-shot elastomer system is discussed next.

To determine the effect of the polyol primary content on the processability of a one-shotsystem, two 4000-MW, ultra-low monol PPG diols, which were EO-capped to obtainprimary contents of 80 and 87 percent were prepared. These were compared with a

437

2000-MW PTMEG-2000, which has 100 percent primary hydroxyl content. These polyolswere evaluated in a one-shot elastomer system consisting of carbodiimide-modified MDI(29.3% NCO), polyol and BDO formulated to an 80 Shore A hardness. Processabilitywas assessed by measuring the hardness build up with time. The reaction rate was heldconstant by adjusting the dibutyltin dilaurate catalyst concentration to achieve a constantpot life of 90 seconds. Pot life is defined as the work life of the solution, which is the timeperiod until the solution becomes too viscous to be poured into a mould. The elastomersolution was poured into moulds preheated to 100 °C and then placed into a 100 °Cvented oven until the elastomer had developed enough integrity to be demoulded. Afterdemoulding, the elastomer hardness was measured at regular intervals and placed backin the oven. The final hardness measurement was made after 16 hours cure at 100 °C.All the data were normalised to 100% using the final hardness measurement. Thenormalised hardness versus time were then plotted.

The hardness build of 80 Shore A, one-shot elastomers based on polyols with primarycontents of 80, 87 and 100 percent are shown in Figure 9.6. As expected, the elastomer

Figure 9.6 Hardness buildup of 80 Shore A one-shot elastomers(carbodiimide-modified MDI/polyol/BDO)


438


based on PTMEG-2000 has the fastest hardness build, achieving 90% of its finalhardness after only 30 minutes. The 87% primary content, 4000-MW, EO-capped diolgave the next fastest hardness build as it took 60 minutes to achieve 90% of its finalhardness. Surprisingly, there was a large difference between the 80 and 87% primarycontent 4000-MW, EO-capped diols as the time to achieve 90% of its final hardnesswent from 60 minutes to over 4 hours. Based on this information, it was decided thatthe primary-hydroxyl content needs to be at least 85%.

The previous studies led to the development of three products, a 2250-MW diol (AcclaimPolyol 2220), 4000-MW diol (Acclaim Polyol 4220) and a 6000-MW triol (AcclaimPolyol 6320). Table 9.4 shows the typical properties of these new higher-reactivity,ultra-low monol PPG polyols. The behaviour of these products in one-shot systems isdiscussed below.

9.3.2 Effect of Monol Content on One-Shot Elastomer Processabilityand Properties

The effect of monol content on the processability and properties of one-shot elastomersis shown in Table 9.9. We prepared 80 Shore A elastomers based on carbodiimide-modified MDI and BDO using a 4000-MW, EO-capped, ultra-low monol diolcontaining 3% monol content and a conventional EO-capped diol containing 15%monol. These polyols had comparable primary hydroxyl concentrations of 87 and89%, respectively.

As expected, the monol content of the EO-capped PPG diols has a significant effecton the elastomer processability. The Shore hardness of the conventional elastomerwas only 57A after 16 hours cure at 100 °C, whereas the elastomer based on ultra-low monol PPG had a hardness of 78A. It took three weeks for the conventionalelastomer to achieve its final hardness of 82A.

Virtually all elastomer properties improved with lower monol contents (see Table 9.9). Ofparticular note, the rebound increased from 59 to 68 percent, the elongation from 470 to550%, tensile strength from 10 to 19 MPa, tear strength from 49 to 68 kN/m, compressionset decreased from 64 to 23%, and the Taber abrasion loss (ASTM D4060-95 [26]) decreasedfrom 220 to 80 mg loss/1000 cycles. The improvement in the stress/strain curves is shownin Figure 9.7. Lower monol content results in significantly higher ultimate polymer MW,which results in improved mechanical properties.

439

lanoitnevnocsusrevlonomwol-artlU:nosirapmocremotsaletohs-enO9.9elbaT)ODB/loylop/IDMdeifidom-edimiidodrac(GPPdeppac-OE

NOITISOPMOCLOYLOP

loyloP lanoitnevnoC lonoMwoL-artlU

%ralom,tnetnoClonoM 51 3

%,tnetnoClyxordyHyramirP 98 78

NOITISOPMOCREMOTSALE

)wbp(IDM-DC 9.24 0.34

)wbp(loyloP 001 001

)wbp(ODB,loidenatuB-4,1 9.01 9.01


)setunim(efiltoP 5.1 5.1

)setunim(emiTdluomeD 02 02

setunim03@htgnertSneerG dooG tnellecxE

001,eruch61(ssendraHerohSlaitinI ° )C A75 A87

SEITREPORPREMOTSALE

AerohSssendraH A28 A28

%,dnuobeRmuludneP 95 86

%,noitagnolEetamitlU 074 055

aPM,htgnertSelisneT 4.01 4.91

aPM,suludoM%001 3.4 3.5




m/Nk,raeTCeiD 94 86

07@h22(%,teSnoisserpmoC ° )C 46 32

)ver0001/ssolgm(,*noisarbArebaT 022 08


%5 2.1 3.1

%01 1.2 3.2

%51 0.3 3.3

%52 4.4 2.5* )ygieG-abiC(6701xonagrI%tw1deniatnocremotsalE


440


9.3.3 Processability and Property Latitude of Elastomers Based onEO-Capped, Ultra-Low Monol Polyols

The processability and property latitude of elastomers based on EO-capped, ultra-lowmonol polyols were determined by preparing a series of elastomers at varying hard-segment contents. The percent hard-segment content is defined as:

(weight MDI + weight BDO)/total elastomer weight x 100.

The elastomers were prepared using the one-shot process by the hand-casting techniqueat a pot life of 1.5 minutes.

The compositions, processability and properties of elastomers prepared using the one-shot process are given in Tables 9.10 to 9.12. For the 4000-MW, EO-capped diol, theupper hardness processability limit via hand casting is about 85 Shore A as indicated bythe green strength in Table 9.10. Machine casting using a shorter pot life (30 to 45

Figure 9.7 Stress/strain curves of 80 Shore A one-shot elastomers based on 4000-MW,EO-capped PPG: Ultra-low monol versus conventional

441

wol-artlu,deppac-OE,WM0004nodesabsremotsaletohs-enO01.9selbaT)ODB/loylop/IDMdeifidom-edimiidobrac(GPPlonom


tnetnoCtnemgeS-draH 51 02 52 53 04 54

)wbp(IDM-DC 2.51 8.02 2.72 0.34 2.35 4.46

loyloP0224mialccA 001 001 001 001 001 001

)wbp(ODB,loidenatuB-4,1 5.2 2.4 1.6 9.01 9.31 5.71


)setunim(efiltoP 5.1 5.1 5.1 5.1 5.1 5.1

)setunim(emiTdluomeD 02 02 02 02 02 02

htgnertSneerG tnellecxE tnellecxE tnellecxE tnellecxE dooG rooP

SEITREPORPREMOTSALE

AerohS,ssendraH A05 A06 A07 A08 A09 A59)D54(

%,dnuobeRmuludneP 86 96 07 86 06 65

%,noitagnolEetamitlU 046 016 085 055 005 054

aPM,htgnertSelisneT 9.4 6.7 7.01 4.91 4.02 4.12

aPM,suludoM%001 1.1 8.1 4.2 3.5 6.7 0.01

aPM,suludoM%002 5.1 5.2 6.3 8.7 7.01 1.31

aPM,suludoM%003 9.1 4.3 0.5 0.01 1.31 9.51

aPM,suludoM%004 4.2 6.4 6.6 5.21 9.51 7.81

m/Nk,raeTCeiD 12 53 74 96 07 58

%,teSnoisserpmoC07@h22( ° )C 04 33 03 52 02 22


%5 2.0 4.0 6.0 3.1 2.1 0.3

%01 4.0 7.0 0.1 3.2 8.2 3.5

%51 6.0 0.1 5.1 3.3 2.4 9.6

%52 0.1 8.1 6.2 2.5 3.7 4.01

seconds) could extend the hardness limit to the low to mid-90s. The lower hardness limitas indicated by the increase in compression set is about 60 Shore A. The lower hardnesslimit can be extended down to at least 40 Shore A by replacing some of the 4000-MW


442


,deppac-OE,lonomwol-artlunodesabsremotsaletohs-enO11.9elbaT)ODB/loylop/IDM-DC(sdnelbloirtWM-0006/loidWM-0004


tnetnoCtnemgeS-draH 01 01 01 01 01

)wbp(IDM-DC 2.01 2.01 2.01 2.01 2.01



)wbp(ODB,loidenatuB-4,1 9.0 9.0 9.0 9.0 9.0


)setunim(efiltoP 5.1 5.1 5.1 5.1 5.1

)setunim(emiTdluomeD 02 02 02 02 02

htgnertSneerG tnellecxE tnellecxE tnellecxE tnellecxE tnellecxE

SEITREPORPREMOTSALE

AerohS,ssendraH A04 A24 A34 A54 A54




aPM,suludoM%001 8.0 9.0 9.0 9.0 0.1

aPM,suludoM%002 9.0 2.1 2.1 2.1 3.1

aPM,suludoM%003 1.1 4.1 5.1 6.1 —

aPM,suludoM%004 1.1 6.1 7.1 — —

m/Nk,raeTCeiD 51 41 21 11 9

07@h22(%,teSnoisserpmoC ° )C 54 83 02 51 21


%5 1.0 1.0 1.0 1.0 1.0

%01 3.0 3.0 3.0 3.0 3.0

%51 4.0 4.0 5.0 5.0 5.0

%52 8.0 8.0 9.0 9.0 9.0

diol with the 6000-MW, EO-capped, ultra-low monol triol as shown by the reduction incompression set in Table 9.11. Thus, using combinations of 4000-MW diol and 6000-MW triol, a broad hardness range from about 40 to 85 Shore A can be achieved.

443

The 2250-MW (50 OH#) EO-capped, ultra-low monol diol was designed to enhanceprocessability at higher Shore hardnesses. The processability and properties of one-shotelastomers based on this polyol are shown in Table 9.12. One-shot elastomers withexcellent processability and properties can be formulated from an 80 Shore A up toabout a 55 Shore D. Thus, a broad range of Shore hardnesses (40A to 55D) can beachieved using EO-capped, ultra-low monol polyols.

wol-artlu,deppac-OE,WM0522nodesabsremotsaletohs-enO21.9elbaT)ODB/loylop/IDM-DC(GPPlonom


tnetnoCtnemgeS-draH 52 53 54 55 56

)wbp(IDM-DC 5.82 2.44 6.56 6.69 2.541


)wbp(ODB,loidenatuB-4,1 8.4 6.9 2.61 6.52 5.04


)setunim(efiltoP 5.1 5.1 5.1 5.1 5.1

)setunim(emiTdluomeD 02 02 02 02 02

htgnertSneerG tnellecxE tnellecxE tnellecxE tnellecxE egarevA

SEITREPORPREMOTSALE

AerohS,ssendraH A76 A28 A49)D44(

D25 D06




aPM,suludoM%001 6.2 1.5 7.9 9.51 4.91

aPM,suludoM%002 7.3 6.7 1.31 4.91 6.52

aPM,suludoM%003 4.5 4.01 6.61 4.12 —

aPM,suludoM%004 3.8 5.41 0.02 — —

m/Nk,raeTCeiD 44 76 88 79 721

07@h22(%,teSnoisserpmoC ° )C 83 82 42 dn dn

denimretedton:dn


444


9.4 Comparison of Ultra-Low Monol PPG Polyols with PTMEG

When comparing ultra-low monol PPG polyols with PTMEG it is essential to keep inmind the fundamental differences between these two high-performance polyether polyolsin order to obtain the maximum benefit in urethane applications. Table 9.13 summarisesthese differences in terms of their chemical structure, hydroxyl type (reactivity), MWDand crystallisability.

GPPlonomwol-artlufoscitsiretcarahclatnemadnuffonosirapmoC31.9elbaTGEMTPdnasloylop

GPPlonoMwoL-artlU GEMTP

erutcurtSlacimehC HC-(O- 2 )O-HC- x HC- 3

Tg ≈ 57- °CHC-(O- 2 HC 2 HC 2 HC 2 )O- x-

Tg ≈ 48- °C

epyTlyxordyH lyxordyHyradnoceS)ytivitcaeRrewoL(

lyxordyHyramirP)ytivitcaeRrehgiH(

thgieWraluceloMnoitubirtsiD

M:worraN w M/ n [email protected]:0022mialccA( ° )C

M:daorB w M/ n 4.2-7.1=04@s-aP21:0002-GEMTP( ° )C

ytilibasillatsyrC elbasillatsyrc-noN33-:tnioPruoP;diuqiL( ° )C

elbasillatsyrC04-82:pm;diloSyxaW( ° )C

tniopgnitlem:pm

Both PPG and PTMEG have polyether backbones, which inherently impart excellenthydrolytic stability and microbial resistance to polyurethanes. The Tg of PPGapproaches that of PTMEG resulting in polyurethanes that can be used in cold, harshenvironments. PTMEG has primary hydroxyl groups, whereas, ultra-low monol PPGare available with secondary hydroxyl groups or EO-capped versions for primaryhydroxyl contents in excess of 85%. Ultra-low monol PPG polyols have narrow MWD,which gives polyols that are an order of magnitude lower in viscosity than PTMEG.The storage and handling requirements for PPG polyols are minimal since they arenon-crystallising liquids with pour points of about –30 °C. Additionally, polyurethanesbased on non-crystallising polyols will not ‘cold harden’ like some of those based oncrystallisable polyols. A comparison of ultra-low monol PPG versus PTMEG in MDI/BDO cured elastomers is given below.

445

9.4.1 MDI/BDO Cured Elastomers: Acclaim Polyol 3205 Versus PTMEG-2000

For a comparison of PTMEG in MDI/BDO cured elastomers the 3200-MW ultra-lowmonol PPG diol was chosen which contains about 20 wt% EO (Acclaim Polyol 3205)since it gives the best overall properties in this type of system. A 2000-MW PTMEG wasused since it gives the best performance at an 80 Shore A hardness. Elastomers of 80Shore A hardness were prepared using the prepolymer process. Table 9.14 summarisestheir prepolymer viscosities, elastomer processing characteristics and physical properties.

The initial surprise was to find that it takes an 8% NCO MDI/Acclaim 3205 prepolymer tomake a 80 Shore A elastomer, whereas, PTMEG-2000 requires a 6% NCO prepolymer. Thisfinding will be explained in the section on the effect of polyol MWD (see section 9.5.4).

Note the dramatic difference in prepolymer viscosities in Table 9.14. The prepolymerbased on ultra-low monol polyol has a viscosity almost ten times lower than that ofPTMEG-2000. This is due to a combination of the narrower MWD and the higher %NCO of the Acclaim 3205 prepolymer.

Lower viscosity prepolymers offer significant advantages in processability. This gives thepolyurethane manufacturer much more latitude in terms of processing conditions suchas temperature and catalyst combinations. Elastomers based on Acclaim Polyol 3205have demould times and green strengths comparable to those based on PTMEG-2000.When hand casting elastomers using a 1 to 3 minute pot life, this can give demould timesof significantly less than 30 minutes.

As expected, the fundamental differences in the polyols also have a significant impact onthe elastomer physical properties. Although elastomers based on ultra-low monol PPGhave pendulum rebounds significantly higher than conventional or low monol polyols,they are about 10 units lower that PTMEG at an equal Shore hardness. It is believed thisis due to fundamental differences in the polyol backbone structure. PPG polyols have apendant methyl group which may inherently absorb more energy than the linear methylenechain of PTMEG.

Elastomers based on Acclaim Polyol 3205 have ultimate elongations that are twice thatof PTMEG. It is believed that this is due to the fact that PTMEG stress crystallises as it isstretched. Crystallites behave like crosslinks and lower the elongation. To test this theory,a series of MDI/BDO cured elastomers were prepared using 2000-MW crystallisablepolyesters (polybutylene adipate and polycaprolactone) and non-crystallisable polyesters(polyethylenebutylene adipate and poly-2-methyl-1,3-propylene adipate). Figures 9.8 and9.9 compare the stress/strain curves of the crystallisable versus non-crystallisablepolyethers and polyesters, respectively. From these figures the importance of the polyolstress crystallisability in determining the elastomer stress/strain profile can be seen.


446


erohS08ta0002-GEMTPhtiwderapmocsremotsale5023mialccA41.9elbaT)derucODB:sremyloperpIDM(ssendrah


epyTloyloP 5023mialccA 0002-GEMTP

remyloperPOCN% 0.8 0.6

HO/OCN 3.5 0.3


02 °C 6.4 24

04 °C 0.1 11

06 °C 14.0 7.3

08 °C 91.0 6.1

SCITSIRETCARAHCGNISSECORP 56@sremyloperPmialccA ° @sremyloperPGEMTP;C08 ° etarualidnitlytubid;30.1=HO/OCN;TR@ODB;C

001@sdluoM;tsylatac ° 001@sruoh61rofderuc;C °C

)setunim(efiltoP 3-1 3-1

)setunim(emiTdluomeD 04-51 04-51

SEITREPORPREMOTSALE

AerohS,ssendraH A28 A38

%,dnuobeRmuludneP 56 57

%,noitagnolEetamitlU 069 015

aPM,htgnertSelisneT 42 13





m/Nk,raeTCeiD 07 07

07@h22(%,teSnoisserpmoC ° )C 91 51


%5 0.1 8.0

%01 1.2 7.1

%51 0.3 6.2

%52 0.5 6.4

447

Figure 9.8 Stress/strain curves of 80 Shore A elastomers: Crystallisable versusnon-crystallisable polyethers

The tensile strengths of ultra-low monol based elastomers are lower than that of PTMEG.However, if one considers the total energy to break as measured by the area under thestress/strain curve, the elastomers based on Acclaim Polyol 3205 are significantly tougherthan those based on PTMEG. The tear strengths of the elastomers are comparable.

9.4.2 Enhanced Elastomer Properties Utilising Ultra-Low MonolPPG/PTMEG Blends

Previously, it was shown that elastomers based on non-crystallising polyols such as ultra-low monol PPG have twice the ultimate elongation of crystallising polyols such as PTMEG.The next experiments were aimed at determining how much ultra-low monol PPG it wouldtake to eliminate the stress crystallisability of PTMEG [19]. A series of 80 Shore A elastomers


448


Figure 9.9 Stress/strain curves of 80 Shore A elastomers: Crystallisable versusnon-crystallisable polyesters

were prepared using MDI prepolymers based on 4000-MW, ultra-low monol PPG/PTMEG-2000 blends chain extended with BDO. The results of the addition of 0, 10, 40and 100 weight percent 4000-MW PPG based on the total polyol weight are given inTable 9.15.

It was found that only about 10 weight percent of 4000-MW PPG is required to breakup the stress crystallisability of PTMEG-2000 as shown in Table 9.16 with the elongationincreasing from 510 to over 1000%. Surprisingly, it was found that 4000-MW PPGaddition gave a synergistic response in terms of elongation, tensile and tear strength.Even more surprising was that the synergy doesn’t reach a maximum until up to 40weight percent of 4000-MW PPG. This elastomer has a 1270% elongation, 35 MPatensile strength and a 77 kN/m tear strength. Additionally, the rebound was only threeunits lower than the PTMEG elastomer system and the prepolymer viscosity wassignificantly lower which enhances processability.

449

desabsremotsaleAerohS08otnoitiddaGPPWM-0004fotceffE51.9elbaT)derucODB:sremyloperpIDM(0002-GEMTPno

SCITSIRETCARAHCREMYLOPERP

GPPWM-0004%tW 0 01 04 001

remyloperPOCN% 6 6 6 8

@s-aP,ytisocsiVremyloperP

02 °C 24 03 22 5.4

04 °C 11 7.7 9.5 3.1

06 °C 7.3 7.2 1.2 5.0

08 °C 6.1 2.1 9.0 3.0

SEITREPORPREMOTSALE




aPM,htgnertSelisneT 13 33 53 71

aPM,suludoM%001 3.5 0.5 5.4 5.4

aPM,suludoM%003 9.9 4.8 9.7 6.7

m/Nk,raeTCeiD 07 18 77 57

So one can maximise the benefit from ultra-low monol PPG polyols in urethanesystems such as MDI/BDO cured elastomers by understanding their fundamentaldifferences in comparison to other high-performance polyols like PTMEG. A detailedlook at the effect of the polyol MWD on the mechanical and dynamic properties ofpolyurethanes follows.

9.5 Polyol Molecular Weight Distribution Effect on Mechanical andDynamic Properties of Polyurethanes

A less well-understood factor, which affects the properties of polyurethanes, is thepolyol MWD. Ultra-low monol PPG polyols have very narrow MWD (polydispersity< 1.1) compared to other high-performance polyols such as PTMEG or polyester polyols(polydispersity: 1.7 to 2.4) as illustrated by the gel permeation chromatography (GPC)curve in Figure 9.10. As expected, a narrower MWD results in an order-of-magnitude


450


reduction in polyol and prepolymer viscosities [7]. Additionally, it was found thatpolyol MWD has a dramatic effect on the polymer mechanical and dynamic properties,particularly in polyurethane/urea systems [19, 27].

9.5.1 TDI Prepolymers Cured with Methylene Bis-(2-Chloroaniline) [MBOCA]

In formulating polyurethane/urea systems using ultra-low monol PPG polyols, it wasfound that polyol MWD has a major effect on polymer properties. A broadened MWDis required to achieve a property profile similar to a PTMEG-based system. The MW-distribution effect is illustrated using TDI prepolymers cured with MBOCA, TDI moisture-cured prepolymers and aqueous polyurethane/urea dispersion coatings.

Traditionally, high-performance TDI/MBOCA-cured elastomers are based on PTMEG. Theyare prepared by reacting TDI with PTMEG at a 1.7:1 to 2:1 NCO to OH ratio to form aprepolymer and then extend the polymer chain with MBOCA. For these studies, prepolymerswere prepared using pure 2,4-TDI reacted with polyol at a 2:1 NCO to OH ratio.

The properties of a MBOCA-extended prepolymer based on a 1000-MW PPG comparedto a 1000-MW PTMEG are shown in columns 1 and 2 of Table 9.16. Surprisingly, the

Figure 9.10 GPC comparison of 4000-MW ultra-low monol PPG to 2000-MW PTMEG

451

elastomer properties were much lower than expected as shown by the hardness (83Aversus 96A); rebound (15 versus 45%); tensile strength (9.2 versus 45 MPa) and tearstrength (39 versus 110 kN/m). After several attempts to improve the properties viaprocessing conditions, catalyst, etc., it was decided to artificially broaden the MWD ofan ultra-low monol PPG polyol to approximate that of PTMEG-1000. This was achievedby blending a 4000-MW PPG with a low-MW glycol diethylene glycol (DEG), to anaverage MW of 1000. The property profile of this elastomer is shown in the last columnof Table 9.16. Interestingly, the property profile is now closer to the elastomer based onPTMEG-1000 and is very similar to a commercial PTMEG-based 93A product, AirthanePET-93A, which is a product of Air Products and Chemicals, Inc.

semyloperpIDTderuc-ACOBMfoseitreporplacinahceM61.9elbaTepyTloyloP 0001-GPP 0001-GEMTP enahtriA

*A39-TEPWM-0004GED/GPP

dnelB




aPM,htgnertSelisneT 3.9 8.44 9.73 7.12

aPM,suludoM%001 8.2 2.41 0.11 0.11

m/Nk,raeTCeiD 93 011 27 88* ,sremotsalEenahteruyloPelbatsaCfoseitreporPgnireenignEmorfnekatataD

.4991,.cnI,slacimehCdnastcudorPriA.cnI,slacimehCdnastcudorPriAfokramedartderetsigerasienahtriA

It is thought that PTMEG gives a higher hardness/modulus polymer because a significantamount of the low-MW tail (see Figure 9.10) ends up in the hard-segment phase. Also,the PTMEG-based elastomer gives a higher rebound because the soft-segment phaseconsists of PTMEG with MW significantly higher than 1000.

The above conclusions are supported by the DMTA curves shown in Figure 9.11. The storagemodulus in the rubbery plateau region is significantly higher for the elastomer based on thePPG/DEG blend, supporting the theory that DEG is contributing to the hard segment. Also,note the transition at 130-150 °C, which is likely to be the softening point of the DEG/TDI/MBOCA hard segment. The dramatic lowering of the soft-segment Tg from 0 °C to -42 °Calso supports the view that the PPG polyol and DEG are going into separate phases.


452


Figure 9.11 DMTA curves of TDI prepolymers cured with MBOCA-PPG-1000 versus4000-MW PPG/DEG blended to 1000 MW

In Figure 9.12, the effect of varying the glycol chain length on the DMTA profiles isshown by using propylene glycol (PG), dipropylene glycol (DPG), tripropylene glycol(TPG) and a 400-MW PPG. The short-chain glycols (PG, DPG, TPG) all gave similarstorage-modulus profiles, while the 400-MW PPG gave a dramatically different profile.This indicates that if the MW of the glycol is too high it becomes miscible in the soft-segment phase and increases the Tg, thus reducing the elastomer rebound and dynamicperformance.

In some cases it may not be practical or cost effective to perform DMTA analysis on allsamples. It was found that materials can be screened by measuring hardness and reboundat a series of temperatures (-10 °C, 23 °C and 105 °C) and these results are shown inTable 9.17. Small changes in properties imply a broad temperature use range for thepolymer. These results are consistent with the DMTA findings.

In Table 9.18 the effect of a higher average MW polyol is shown by comparing 2000-MW ultra-low monol PPG with a blended product (4000-MW PPG/DEG). Again, thedifferences in hardness, rebound and modulus are dramatic and are consistent with theabove trends.

453

Figure 9.12 Effect of glycol chain length on DMTA curves of TDI prepolymerscured with MBOCA

fotceffe(sremyloperpIDTderuc-ACOBMfoseitreporplacinahceM71.9elbaT)htgnertsniahclocylg

–epyTloyloPWM0001

ssendraHerohS@

501/32/01-°C

erohsaB@dnuobeR501/32/01-

°C

noitagnolE%

elisneThtgnertS

aPM

%001suludoM

aPM

CeiDraeT

htgnertSm/Nk

0001-GPP A96/A38/D47 24/41/15 043 3.9 8.2 93

GP/0004 A88/A29/A49 26/24/72 014 0.92 9.01 39

GPD/0004 A19/A49/A59 26/84/82 053 8.42 0.21 39

GPT/0004 A98/A29/A59 85/24/82 063 0.92 0.31 79

004GPP/0004 A68/D85/D96 53/83/64 063 3.82 3.21 68


454


sremyloperpIDTderuc-ACOBMfoseitreporplacinahceM81.9elbaTepyTloyloP GPPWM-0002 A08-TEPenahtriA GPPWM-0004

dnelBGED

AerohS,ssendraH A86 A08 A08




aPM,suludoM%001 1.2 1.4 3.4

m/Nk,raeTCeiD 33 16 76* stcudorPriA,sremotsalEenahteruyloPelbatsaCfoseitreporPgnireenignEmorfnekatataD

.4991,.cnI,slacimehCdna.cnI,slacimehCdnastcudorPriAfokramedartderetsigerasienahtriA

9.5.2 Moisture-Cured TDI Prepolymers

To determine if the polyol MWD effect can be seen in other polyurethane/urea systems,a moisture-cured film was prepared using a 2:1 prepolymer based on 2,4-TDI and PPG-1000 and this was compared with one based on 4000-MW PPG blended with DEG to aMW of 1000. Table 9.19 shows the mechanical properties of the films in bold type. Thedramatic increases in 100% modulus (1.3 to 3.6 MPa), tensile strength (14 to 35 MPa)and tear strength (18 to 58 kN/m) obtained by broadening the MWD of the polyol can

seitreporplacinahcemnoWMdnelbGED/GPPWM-0004fotceffE91.9elbaTsremyloperpIDTderuc-erutsiomfo

0001-GPP sdnelBlocylGenelyhteiD/GPPWM-0004

WMdnelBloyloP 0001 0003 0061 0001 077

)yroeht(OCN%remyloperP 32.6 45.2 42.4 32.6 35.7

%,noitagnolE 037 0821 058 027 045


aPM,suludoM%001 3.1 4.0 0.2 6.3 3.8

aPM,suludoM%003 1.2 6.0 0.4 6.5 6.41

m/Nk,raeTCeiD 81 21 04 85 39

455

clearly be seen. The DMTA curves of these materials show the same trends as with theMBOCA-cured TDI prepolymers as seen in Figure 9.13. A formulation with 4000-MWPPG decreases the soft-segment Tg from –16 °C to –58 °C greatly enhancing the lowtemperature performance of this system.

The versatility of using this approach can be clearly seen by blending 4000-MW PPGand DEG to a wide range of MW (770 to 3000) as shown in Table 9.19. By simplyblending these two components, one can vary the 100% modulus from 0.4 to 8.2 MPa.Traditionally, hardness would be adjusted by using various polyol MWs; however, whenusing this approach there is a reduction in the low-temperature flexibility as shown bythe DMTA curves in Figure 9.13.

Figure 9.13 DMTA curves of TDI moisture-cured prepolymer films - PPG-1000 versus4000-MW PPG/DEG blended to 1000 MW

Using the blend approach, the hardness can be varied over a wide range while retainingthe low-temperature flexibility as shown by the DMTA curves in Figure 9.14. Notethat with increasing amounts of DEG, there is a transition beginning at about 100 °Cthat becomes more pronounced and shifts to about 80 °C. This transition is likely


456


associated with the Tg of the TDI/DEG/urea hard segment. Similar property profilesare observed in isophorone diisocyanate-based (IPDI), moisture-cured systems utilisedin deck-coating materials.

9.5.3 Aqueous Polyurethane/Urea Dispersion Coatings

We also determined the effect of polyol MWD in aqueous polyurethane/ureadispersion coatings by first comparing a 2000-MW, ultra-low monol PPG withPTMEG-2000. The dispersions were prepared by the prepolymer-mixing processusing the formulations given in Table 9.20 [15]. The prepolymer was based onIPDI, polyol and dimethylol propionic acid (DMPA). It also contained 10 weightpercent N-methyl pyrrolidone (NMP) as a coalescing aid. Triethylamine was usedas a neutraliser and ethylene diamine as chain extender. All aqueous dispersionswere prepared to 40% solids content. A detailed experimental procedure is givenin a previous article [28].

Figure 9.14 Effect of 4000-MW PPG/DEG blend MW on DMTA curves of TDImoisture-cured prepolymer films

457

snoitalumrofnoisrepsidaeru/enahteruylopsuoeuqA02.9elbaT

loyloP GPPWM0002 0002-GEMTPWM0004GPP/GPP

WM,dnelBloyloP 0002 0002 0002

NOITISOPMOCREMYLOPERP ).viuqe(g

etanaycosiiDenorohposI )785.0(2.56 )785.0(2.56 )685.0(1.56

loyloP )691.0(6.591 )691.0(6.591 )590.0(9.881

locylGenelyporpiD — — )101.0(8.6

dicacinoiporplolyhtemiD )731.0(2.9 )731.0(2.9 )731.0(2.9

enodilorryplyhteM-N 03 03 03

thgieWlatoT 003 003 003

SCITSIRETCARAHCREMYLOPERP

HO/OCN 67.1 67.1 67.1

)laciteroeht(OCN% 6.3 5.3 6.3

)lautca(OCN% 52.3 22.3 03.3

etalyxobraC% 21.1 01.1 11.1

08@ytisocsiV ° s-aP,C dn 7.3 7.0

NOISNETXENIAHCDNANOISREPSID

resilartueN enimalyhteirT enimalyhteirT enimalyhteirT

rednetxEniahC enimaidenelyhtE enimaidenelyhtE enimaidenelyhtE

SCITSIRETCARAHCNOISREPSID

%tw,tnetnoCsdiloS 04 04 04

%tw,tnetnoCtnevlosoC 0.4 0.4 0.4

The stress/strain curves in Figure 9.15 show that the film based on the 2000-MW PPGhas a significantly lower hardness than that based on PTMEG-2000. To take into accountthe differences in polyol MWD, we then prepared an aqueous dispersion using 4000-MW PPG and DPG blended to a 2000 MW. Figure 9.15 shows that these stress/straincurves are now very similar up to 200% elongation. At higher elongations, the filmbased on PTMEG-2000 begins to stress crystallise resulting in increased modulus andreduced ultimate elongation. The DMTA curves in Figure 9.16 show that the film basedon the blend of 4000-MW PPG and DPG has a profile more similar to PTMEG-2000than to that based on 2000-MW PPG.


458


Figure 9.15 Stress/strain curves of aqueous polyurethane/urea dispersion films

Figure 9.16 DMTA curves of aqueous polyurethane/urea dispersion films

459

9.5.4 MDI Prepolymers Cured with BDO

Table 9.21 demonstrates the polyol MWD effect by comparing BDO cured 6% NCO4,4´-MDI prepolymers based on 2000-MW PPG and PTMEG-2000. The elastomer basedon the narrow-MWD polyol has a significantly lower hardness (68A versus 83A) andrebound (60 versus 75%). This is consistent with PTMEG-2000 containing some short-chain diol, which phase separates into the hard segment promoting higher hardness.Also, the soft-segment Tg of the PTMEG-2000 elastomer would be lower since the softsegment actually consists of PTMEG with MW significantly higher than 2000. Based onour previous discussion, to best approximate the property profile of a broad-MWD polyolsuch as PTMEG, one needs to use a higher MW ultra-low monol PPG polyol with someadditional chain extender (BDO).

Table 9.21 illustrates this effect by increasing the PPG MW from 2000 to 4000 and theprepolymer % NCO from 6 to 8. This increases the hardness to the same level as thePTMEG-based elastomer and increases the rebound to 65%. As an aside, increasing theprepolymer % NCO from 6 to 8 represents a polyol/BDO blend of 2160 MW.

We conclude with some remarks regarding recommendations as to which short chaindiols to use with polyurethanes versus polyurethane/ureas. In elastomeric materials,optimal dynamic performance is obtained in a well phase-separated system (a low soft-segment Tg and a high softening or melting-temperature hard-segment). For polyurethanesystems, optimal dynamic performance is achieved by maintaining symmetry in the hardsegment. This is accomplished by using a symmetrical diisocyanate (such as pure 4,4´-MDI) with a symmetrical chain extender (such as ethylene glycol, BDO, DEG or 1,6-hexanediol). This symmetry results in a high-melting-temperature crystalline hard-segmentphase, enhancing phase separation in the polymer. Complete phase separation results ina low soft-segment Tg and excellent rebound or dynamic performance. Disrupting thesymmetry of the hard segment potentially reduces the rebound or dynamic performance.In formulating with ultra-low monol PPG in polyurethane systems, optimal propertiesare attained using a high MW polyol with additional symmetrical chain extender. Use ofmixed chain extenders will disrupt the hard-segment symmetry and lead to a reductionin dynamic properties.

In polyurethane/urea systems, phase separation and optimal dynamic performance isdriven by the hydrogen bonding of the urea linkage and not by the hard-segment symmetry.So when formulating ultra-low monol PPG polyols in polyurethane/urea systems, thereis more latitude in the choice of short-chain glycols. However, it is undesirable to choosean extender that is so short or symmetrical that the isocyanate adduct is incompatible inthe isocyanate-terminated prepolymer. Additionally, the glycol MW must be limited toprevent it from phase mixing with the soft segment.


460


ODBhtiwderucsremyloperpIDMfoseitreporpremotsalE12.9elbaTNOITISOPMOCREMYLOPERP

epyTloyloP GPPWM-0002 0002-GEMTP GPPWM-0004

remyloperPOCN% 0.6 0.6 0.8

HO/OCN 0.3 0.3 3.6

)wbp(IDM-´4,4 9.63 9.63 5.93

)wbp(loyloP 001 001 001


02 °C 51 24 5.4

04 °C 7.2 11 52.1

06 °C 58.0 7.3 5.0

08 °C 4.0 6.1 52.0


)wbp(remyloperP 001 001 001

)wbp(ODB,loidenatub-4,1 2.6 2.6 3.8

)setunim(efiltoP 3-1 3-1 3-1

setunim(emiTdluomeD 54-02 54-02 06-03

SEITREPORPREMOTSALE 32@skeew4rofdenoitidnocselpmaS ° %05/Cytidimuhevitaler

AerohS,ssendraH A86 A38 A28




aPM,suludoM%001 5.3 3.5 5.4

aPM,suludoM%002 3.5 1.7 2.6

aPM,suludoM%003 4.7 9.9 6.7

aPM,suludoM%004 9.9 7.41 9.8

m/Nk,raeTCeiD 85 07 57

07@h22(%,teSnoisserpmoC ° )C 22 51 91


%5 5.0 8.0 0.1

%01 0.1 7.1 8.1

%51 6.1 6.2 5.2

%52 8.2 6.4 0.4

461

The benefits from using ultra-low monol PPG polyols in polyurethane or polyurethane/urea systems can be realised by understanding the fundamental differences between themand other high-performance polyols such as PTMEG. In particular, when consideringthe polyol MWD, PTMEG should be regarded as a high MW polyol blended with a lowMW chain extender (or pseudo chain extender). In contrast, ultra-low monol PPG containno low-MW fraction, therefore, to mimic the behavior of PTMEG it is necessary to addadditional low MW glycol into the prepolymer or resin components. Because of thisfeature, ultra-low monol PPG have a broader formulating latitude, since chain extendercan be added to tailor the property profile to the specific application.

9.6 Conclusions

Ultra-low monol PPG polyols have excellent properties and ease of processing. Theelimination of monol enables polymers to be produced having much higher final MW,i.e., improved properties, as well as significantly faster MW buildup during polymerformation, i.e., shorter demould times and excellent green strength.

Another important feature of these polyols is their narrow MWD. Low polydispersityresults in lower viscosities in both the polyol and isocyanate-terminated prepolymers.Polyol MWD has a significant effect on the mechanical and dynamic properties ofpolyurethane and polyurethane/urea systems. Broad-MWD polyols such as PTMEG orpolyester polyols contain significant amounts of hard-segment diol, which limits theirsoftness. Acclaim polyols contain no hard-segment diol and thus have inherent superiorproperties in soft systems. They can be made to approximate broad-MWD polyols suchas PTMEG by incorporation of a low MW glycol.

Acknowledgements

We would like to thank Steve Bailey, Daphne Hale, Beth Lambert, Tony Loveday andMike Robinson, who carefully and conscientiously prepared the samples and carried outthe physical property testing.

References

1. D. M. Simons and J. J. Verbanc, Journal of Polymer Science, 1960, 44, 144, 303.

2. G. Gee, W. C. E. Higginson, K. J. Tailor and H. W. Trenholme, Journal of theChemical Society, 1961, 4298.


462


3. W. H. Snyder, K. J. Tailor, N. S. Chu and C. C. Price, Transactions of the NewYork Academy of Science, Series II, 1962, 24, 341.

4. R. J. Herold in Macromolecular Syntheses, Volume 5, Ed., E. L. Wittbecker, JohnWiley and Sons, Inc., New York, 1974, 9.

5. H. van der Hulst, G. A. Pogany and J. Kuyper, inventors; Shell Oil Company,assignee; US Patent 4,477,589, 1984.

6. J. L. Schuchardt and S. D. Harper, Presented at the 32nd Annual PolyurethaneTechnical/Marketing Conference, San Francisco, CA, USA, 1989, p.360.

7. R. L. Mascioli, Presented at the 32nd Annual Polyurethane Technical/MarketingConference, 1989, San Francisco, CA, USA, p.139.

8. J. W. Reisch and D. M. Capone, Presented at the 33rd Annual PolyurethaneTechnical/Marketing Conference, Orlando, FL, USA, 1990.

9. J. Milgram, inventor; General Tire Corporation, assignee; US Patent 3,278,457,1966.

10. R. J. Belner, inventor; General Tire Corporation, assignee, US Patent 3,278,458,1966.

11. R. J. Herold, inventor; General Tire Corporation, assignee, US Patent 3,278,459,1966.

12. R. J. Herold and R. A. Livigini, Polymerization Kinetics and Technology, Ed., N.A. J. Platzer, Advances in Chemistry Series No.128, American Chemical Society,Washington, DC, 1973, 208.

13. C. P. Smith, J. W. Reisch and J. M. O’Connor, Presented at the PolyurethaneWorld Congress, 1991, Nice, France, p.313.

14. N. Barksby, G. L. Allen, Presented at the Polyurethane World Congress, 1993,Vancouver, BC, Canada, p.445.

15. N. Barksby, S. D. Seneker and G. L. Allen, Presented at the PolyurethaneManufacturing Association Conference, Pittsburgh, PA, November 1995.

16. S. D. Seneker and N. Barksby, Presented at the Utech 96 Conference, The Hague,The Netherlands, March 1996, Paper No.46.

17. N. Barksby, S. D. Seneker and G. L. Allen, Urethanes Technology, 1996, 13, 1, 36.

463

18. R. J. Tuinman, T. L. Fishback and C. J. Reichel, Presented at the PolyurethanesExpo ’98 Conference, Dallas, TX, 1998, p.535.

19. S. D. Seneker, N. Barksby and B. D. Lawrey, Presented at the Polyurethanes Expo’96 Conference, Las Vegas, NV, 1996, p.305.

20. S. D. Seneker, C. C. Shen and N. Barksby, Presented at the Polyurethanes WorldCongress ‘97, Amsterdam, The Netherlands, 1997, p.568.

21. S. D. Seneker, C. C. Shen and N. Barksby, Presented at the PolyurethaneManufacturing Association Conference, San Diego, CA, March 1998.

22. R. J. Tuinman, R. L. Fishback and C. J. Reichel, Presented at the PolyurethanesWorld Congress ‘97, Amsterdam, The Netherlands, 1997, p.560.

23. T. L. Fishback and C. J. Reichel, Presented at the Polyurethanes Expo ‘96International Conference, Las Vegas, NV, 1996, p.282.

24. A. T. Chen, R. R. Wells, C. P. Smith, J. W. Reisch, M. M. Emmet and J. M.O’Connor, Presented at the Polyurethanes World Congress 1993, Vancouver, BC,Canada, 1993, p.388.

25. C. P. Smith, J. W. Reisch and J. M. O’Connor, Journal of Elastomers and Plastics,1992, 24, 4, 306.

26. ASTM D4060-95, Standard Test Method for Abrasion Resistance of OrganicCoatings by the Taber Abraser.

27. S. D. Seneker, N. Barksby and B. D. Lawrey, Presented at the Polyurethanes Expo’98 Conference, Dallas, TX, 1998, p.195.

28. S. D. Seneker, N. Barksby and G. B. Ellerbe, Presented at the Utech Asia ‘97Conference, Suntec City, Singapore, 1997, Paper No.41.


464


465

APPENDIX

Laboratory Preparation of 2,4-TDI and 4,4´-MDI Prepolymers

Isocyanate prepolymers were prepared in glass reaction vessels in a laboratory fumehood. The reaction vessel was equipped with a stirrer, temperature controller and nitrogeninlet. The reaction was carried out under a dry-nitrogen atmosphere to minimise exposureto atmospheric moisture and polyol oxidation.

Isocyanate is added first to the reaction vessel and heated to approximately 50 to 60 °C.Polyol is then added to the isocyanate, carefully controlling the addition rate to ensurethe reaction temperature does not exceed 85 °C. Higher temperatures can cause harmfulside reactions that can affect prepolymer performance. After reacting at 80 °C forapproximately 4 hours, the extent of reaction was determined by measuring the percentisocyanate content (% NCO) of the prepolymer. Heating was continued until theprepolymer was slightly below the theoretical % NCO value.

Laboratory Casting of 4,4´-MDI Prepolymers Cured with BDO

MDI/BDO cast-elastomer systems were made by mixing and reacting two components,the MDI-terminated prepolymers (described previously) and BDO. The elastomers inthis chapter were prepared using the hand-mixing procedure. The MDI prepolymer waspreheated to ~65 °C and degassed under vacuum (130 to 665 Pa) until foaming stopped.Catalyst (25-50 ppm dibutyltin dilaurate) was added to the prepolymer and mixedthoroughly to give the desired pot life of 1 to 3 minutes.

The BDO was added (isocyanate to hydroxyl ratio (NCO:OH) of 1.03) and mixed thoroughlyusing a Jiffy mixer until the blend was homogeneous. This solution was immediately pouredinto preheated moulds at 100 °C treated with mould release. After demoulding, the sampleswere post-cured at 100 °C for 16 hours. The polymer samples were conditioned at 23 °C and50% relative humidity for at least four weeks prior to testing.

Laboratory Casting of One-Shot Elastomers Based on Carbodiimide-Modified MDI, Polyol, and BDO

One-shot elastomers were prepared by mixing and reacting three components:carbodiimide-modified MDI (CD-MDI), high primary-hydroxyl-content polyol and BDO.The polyol was preheated to 60 °C and degassed under vacuum (130 to 665 Pa). The


466


CD-MDI, polyol and BDO were charged to a 250 ml glass jar. Dibutyltin dilaurate catalyst(25-50 ppm) was added to adjust the pot life to approximately 90 seconds. Thesecomponents were mixed thoroughly and poured into preheated moulds at 100 °C, treatedwith mould release. After demoulding, the samples were post-cured at 100 °C for 16hours. The polymer samples were conditioned at 23 °C and 50% relative humidity for atleast four weeks prior to testing.

Laboratory Casting of 2,4-TDI Prepolymers Cured with MBOCA

The TDI/MBOCA cured elastomers were prepared by mixing and reacting twocomponents: 2,4-TDI prepolymers (described previously) and MBOCA. The elastomersin this chapter were prepared using the hand-mixing procedure. The 2,4-TDI prepolymerwas preheated to 90 °C and degassed under vacuum (130 to 665 Pa) until foamingstopped. The MBOCA was melted at 120 °C. The 2,4-TDI prepolymer and MBOCAwere stirred thoroughly using a Jiffy mixer at an NCO:NH ratio of 1.05 without catalyst.The solution was immediately poured into preheated moulds at 100 °C treated withmould release. After demoulding, the samples were post-cured at 100 °C for 16 hours.The polymer samples were conditioned at 23 °C and 50% relative humidity for at leastfour weeks prior to testing.

Laboratory Moisture-Curing of 2,4-TDI Prepolymers

The 2,4-TDI prepolymers were cast as films on glass plates using a 0.762 mm draw-down bar (a draw-down bar is a metal bar with a gap, which is pulled down a glass plateleaving a thin film of prepolymer. The prepolymer is then moisture-cured to form thepolymer film). The films were cured/conditioned at 23 °C and 50% relative humidity forat least four weeks prior to testing.

Laboratory Preparation of Aqueous Polyurethane/Urea Dispersionsusing the Prepolymer Mixing Process

Isocyanate prepolymers were prepared in glass reaction vessels in a laboratory fumehood. The reaction vessel was equipped with a stirrer, temperature controller, and nitrogeninlet. The reaction was carried out under a dry-nitrogen atmosphere to minimise exposureto atmospheric moisture and polyol oxidation. The stepwise procedure is:

Charge polyol, DMPA, (stabiliser), NMP, (coalescing aid) in a reaction vessel equippedwith a thermocouple (thermometer) and agitator. Care should be taken to keep

467

moisture away from the reaction mixture by using a drying tube or a positive pressureof dry nitrogen.

Heat the mixture to 95 °C and apply vacuum to remove any residual water (DMPA andpolyol stored in the laboratory may contain high water levels).

Cool the mixture to 50-60 °C, then weigh in isocyanate (IPDI, etc.)

Heat the mixture to 80 to 100 °C for aliphatic diisocyanates or 60 to 80 °C for aromaticdiisocyanates. If necessary, a small amount of a tin catalyst, (e.g., 25 to 100 ppm ofstannous octoate) can be added.

React this mixture until the % NCO content is 0.2-0.5 units lower than the theoretical% NCO. Reaction time is usually 6 to 12 hours depending on the % NCO of theprepolymer.

Cool the mixture to 50-60 °C, and then add neutraliser, i.e., triethylamine, to theprepolymer. Mixing is critical during this step. The neutralisation forms the water-dispersible carboxylate ammonium salt.

Heat the prepolymer amine salt to 60 to 80 °C and slowly pour into a resin kettle chargedwith a known amount deionised water at 25 to 50 °C with strong agitation. Add theprepolymer to give the desired solids content.

Immediately add the chain extender (ethylene diamine), which has been diluted in waterslowly to the aqueous polyurethane/urea dispersion. Typically chain extend only 80-90% of the actual free NCO groups because some will be lost due to reaction with water.As an alternative, the amine extender may be added to the water prior to dispersing theneutralised prepolymer.

Cook out the residual excess isocyanate by heating the aqueous dispersion at 60-80 °Cfor two hours. Excess unreacted isocyanate in an aqueous polyurethae/urea dispersionwill lead to carbon dioxide gas and pressure buildup in the container.


468


469

Abbreviations and Acronyms

ABA Alternative blowing agentsABS Acrylonitrile butadiene styreneAN Acceptor numberAPCI Air Products and Chemicals, Inc.APS Aminopropyl triethoxy silaneASD Acrylic-styrene dispersionASTM American Society for Testing and MaterialsAVG AverageBA Butyl acrylateBD 1,4-butadieneBDO 1,4-butanediolCD-MDI Carbodiimide-modified MDICFC ChloroflurorocarbonsCHL ChloroformCHP Cumene hydroperoxideCOM CommercialCP CyclopentaneCS Chloropropyl triethoxy silaneDAPS Data acquisition and plotting systemDC DabcoDEE Diethyl etherDEG Diethylene glycolDEOA DiethanolamineDEOA-LF Diethanolamine liquid formDMA Dynamic mechanical analysisDMC Double metal cyanideDMPA Dimethyl propionic acidDMSO Dimethyl sulphoxideDMTA Dynamic mechanical analysisDN Donor numberDPG Dipropylene glycolDPUR Polyurethane dispersions

470


DSC Differential scanning calorimetryECR Energy consumption reductionEO Ethylene oxideES Epoxybutyl trimethoxy silaneFEA Finite element analysisFR Fire retardantFTC Force-to crushFT-IR Fourier transform infra-red spectroscopyGC Gas chromatographyGPC Gel permeation chromatographyGWP Global warming potentialHACS Humid aged compression setsHCFC HydrochloroflurocarbonHDPE High density PEHER Bis-(beta-hydroxyethyl) ether of resorcinolHER-TG HER Technical gradeHFC HydrofluorocarbonsHP High purityHQEE Hydroquinone (di-beta-hydroxyethyl) etherid Internal diameterIFD Indentation force deflectionIGC Inverse gas chromatographyILD Indentation load deflectionIPDI Isophorone diisocyanateIPN Interpenetrating polymer networkIR InfraredKOH Potassium hydroxideLD-SRIM Low density structural RIMLFI-PU Long fibre injection PUMBOCA Methylene bis- (2-chloroaniline)MDI Methylene diphenyl isocyanateMDPE Medium density PEMDPUR Modified polyurethaneMEK Methyl ether ketoneMFFT Minimum film formation temperatureMHW Modified hot wireMM Methyl methacrylateMn Number average molecular weight

471

MOCA Methylene bis – (2-chloroaniline)mp Melting pointMS Mass spectrometerMS Mercaptobutyl trimethoxy silaneMW Molecular weightMWD Molecular weight distributionNCO IsocyanateND Not detectednd Not determinedNEG Non-evaporable getterNM Not measuredNMP N-methyl pyrrolidoneNMR Nuclear magnetic resonanceOCF Owens Corning fibre glassODP Ozone depletion potentialOEM Original equipment manufacturerOH# Hydroxyl numberOTR Oxygen permeation ratePAN PolyacrylonitrilePbw Parts by weightPDMS PolydimethylsiloxanePE PolyethylenePEG Polyethylene glycolPET Polyethylene terephthalatePG Propylene glycolPLC Programme logic controllerPO Propylene oxidePPG Polyoxypropylene glycolPPH Parts per hundredpphp Parts per hundred polyolppm Parts per millionPPUR Polyurethane prepolymerPS PolystyrenePTMEG Polytetramethylene ether glycolPTMG Polytetramethylene glycolPU, PUR PolyurethanePVC Polyvinyl chloridePVDC Polyvinylidene chloride

Abbreviations and Acronyms

472


QA Quality assuranceQC Quality controlR&D Research and DevelopmentRGA Residual gas analysisRH Relative humidityRIM Reaction Injection MouldingRRIM Reinforced reaction injection mouldingRT Room temperatureRUF Recycled urethane fluffS StyreneSAE Society of Automotive EngineersSD Standard deviationSEM Scanning electron microscopeSLM Standard litres per minuteSRG Spinning rotor guageSRIM Structural RIMSSF Silica surfactantSTP Standard temperature and pressureTCP Trimethylsilyl-capped polysilicateTDI Toluene diisocyanateTEDA Triethylene diamineTEM Transmission electron microscopyTg Glass transition temperatureTG Technical gradeTHF TetrahydrofuranTOC Top-of-the-cupTP ThermoplasticTPG Tri propylene glycolTPR Time pressure releaseTPU Thermoplastic PUVIP Vacuum insulated plateVOC Volatile organic compoundsVPF Variable pressure foamingVS Vinyl trimethyl silaneWTR Water permeation rate

473

Gary D. AndrewAir Products and Chemicals, Inc.7201 Hamilton Blvd.Allentown, PA 18195USA

Nigel BarksbyBayer CorporationSouth Charleston Technical Center3200 Kanawha TurnpikeBuilding 727South CharlestonWV 25303USA

Richard J. BraunDepartment of Mathematical ScienceUniversity of DelawareNewarkDE 19716USA

Georg BurkhartGoldschmidt AGGoldschmidtstrasse 100D-45127 EssenGermany

Raj B. DurairajDirector of ResearchIndspec Chemical CorporationPittsburghPA 15238USA

Contributors

Daniel KlempnerPolymer InstituteUniversity of Detroit Mercy4001 W. McNichols RoadDetroitMI 48219-0900USA

Jane G. KnissAir Products and Chemicals, Inc.7201 Hamilton Blvd.Allentown, PA 18195USA

Anita Koncka-FolandIndustrial Chemistry Research Institute8, Rydygiera Street01-315 WarsawPoland

Janusz KozakiewiczIndustrial Chemistry Research Institute8, Rydygiera Street01-315 WarsawPoland

Bruce D. LawreyBayer CorporationSouth Charleston Technical Center3200 Kanawha TurnpikeBuilding 727South CharlestonWV 25303USA

474


Izabella LegockaIndustrial Chemistry Research Institute8, Rydygiera Street01-315 WarsawPoland

Mark L. ListemannAir Products and Chemicals, Inc.7201 Hamilton Blvd.Allentown, PA 18195USA

Paolo ManiniSAES Getters SpAViale Italia 7720020 Lainate (MI)Italy

Lisa A. MercandoAir Products and Chemicals, Inc.7201 Hamilton Blvd.Allentown, PA 18195USA

Shailesh NaireDepartment of Mathematical ScienceUniversity of DelawareNewarkDE 19716USA

Benjamin M. NugentCentral Research and DevelopmentMail # CO43C1Dow Corning CorporationMidlandMI 48686-0994USA

Udo C. PerniszCentral Research and DevelopmentMail # CO43C1

Dow Corning CorporationMidlandMI 48686-0994USA

Arturo PlanaAir Products and Chemicals, Inc.7201 Hamilton Blvd.AllentownPA 18195USA

H.P. SchreiberChemical Engineering DepartmentEcole PolytechniqueP.O. Box 6079Stn. Centre-VilleMontrealQuebecCanadaH3C 3A7

Stephen D. SenekerAnderson Development Company1415 E Michigan streetAdrianMI 49221USA

Ashok SenguptaCorporate Research & Development3M Canada CompanyLondonOntarioCanadaN6A 4T1

Jan SkarzynskiIndustrial Chemistry Research Institute8, Rydygiera Street01-315 WarsawPoland

475

Steven A. SnowCentral Research and DevelopmentMail # CO43C1Dow Corning CorporationMidlandMI 48686-0994USA

Robert E. StevensAir Products and Chemicals, Inc.7201 Hamilton Blvd.AllentownPA 18195-1501USA

Max TavernaCannon GroupVia C Colombo 4920090 Trezzano s/NaviglioMilanoItaly

James D. TobiasAir Products and Chemicals, Inc.7201 Hamilton Blvd.Allentown, PA 18195USA

Andreas WeierDegussa AGKarl-Arnold Plazt 1aD-40474 DusseldorfGermany

Stephan WendelAir Products and Chemicals, Inc.Robert Koch Strasse 2722851 NorderstedtGermany

Contributors

476


477

Author Index

Adam, S.J. 210Akabori, K. 215, 257Akita, J. 211Akkurt, A.S. 211Allen, G.L. 462Allen, K.W. 332Alonzo, V. 334Andrew, G.D. 82Arkles, B. 91, 110Armistead, J.P. 111Artavia, L.D. 215, 257Ashby, M.F. 82Bachelor, G.K. 259Baines, F.L. 258Bankoff, G.B. 259Barksby, N. 462, 463Barosi, A. 210Battice, D.G. 83Bayer, O. 85Bechara, I. 333Belner, R.J. 462Benson, D.K. 212Berger, E.J. 366Beu, T.A. 209Biesmans, G. 208, 210Biondich, B. 154Blackwell, J. 419Boffito, B. 210Boffito, C. 210, 211Bohman, R.H. 212Bonansea, A. 155Bonart, R. 419Bonekamp, J. 176, 210Borghi, M. 210Borgogelli, R. 111

Born, L. 419Boudreau, R.J. 111, 256Bovenkerk, H.P. 208Brandrup, J. 210Braun, J.M. 366Braun, R.J. 259, 260Breck, D.W. 210Brenner, H. 259Brochhagen, F.K. 109Brookman, D.J. 367Bundy, F.P. 208Burchianti, J. 417Burkhart, G. 82, 102, 109, 111Cairncross, R.A. 260Caloi, R.M. 211Capone, D.M. 462Carley, J. 210Carson, S. 110Cavenaugh, G.C. 110Cavender, K.D. 82Chang, Y-S. 366Charles, A. 110Chaya, J.E. 109Chen, A.T. 463Chen, J.H. 339, 366Chen, L.W. 418Childs, K.W. 212Christfreund, A. 209Chu, N.S. 462Clift, S.M. 418Clunie, J.S. 257Coogan, R.G. 332Cooper, S.L. 418, 419Cornelius, G.C. 334Corradi, P. 154


478

Covington, J. 210Creswick, M.W. 111Cunningham, A. 209Czalych, A.E. 319, 334Dahm, M. 256Dai, D.G. 109David, C.E. 210Davis, H.T. 110, 111Davis, J.H. 418Davis, S.H. 259de Almeida, J. 417De Vos, R. 209, 210, 211De Vries, A.J. 258Decker, T.G. 91, 92, 110, 256della Porta, P. 210Deng, Z. 337, 366Deschaght, J. 211Desor, U. 333Dieterich, D. 333Dietrich, K.W. 209Dillard, D.A. 366Dixon, C. 192, 211Doni, F. 210Douglas, T.L. 418Doyle, E.N. 110Doyle, J.E. 418Droste, W. 109Drye, J.L. 110Dubjaga, J.E. 256Dubyaga, E.G. 256Durairaj, R.B. 417Edwards, D.A. 259Eiben, R.G. 110Ellerbe, G.B. 463Emmet, M.M. 463Eschbach, C.S. 258Estes, G.M. 419Evans, D.F. 257Feinerman, A.E. 256Fenton, W.N. 257Ferrario, B. 209, 210

Figini, A. 210Fine, H.A. 208, 212Fiorentini, C. 110, 154Fishback, T.L. 463Florianczyk, Z. 333Francais, E. 208Frankel, S.P. 214, 257Franyutti, S. 418Fremerey, J.K. 211Fricke, J. 208Frisch, K.C. 417Fujimoto, K. 215, 257Gabbard, W.A. 210Gasparini, G. 209Gee, G. 461Gent, A. 336, 366Gentle, T.E. 259George, R.F. 417Geurts, J.M. 334Giannopous, N. 334Gibson, L.J. 82Gibson, P.E. 418Gilbert, R.G. 333Giordano, S. 334Glicksman, L.R. 160, 208, 209Goddard, E.D. 256Goldsmith, J. 332Goodman, J.F. 257Goos, H.C. 332Grabbe, M. 332Graff, G. 155Graves, R.S. 208Gray, D.G. 366Grickova, I.A. 283, 287, 306, 311Griffith, M. 110Griffiths, T. 110Grizwold, A.A. 109Gruber, B. 266, 333Guagliardo, M. 332Güclü, H. 211Guillet, J.E. 366

479

Author Index

Gutmann, V. 337, 349, 357, 359, 366Hamann, H. 256Hamilton, A.J. 199, 212Harper, S.D. 462Hartung, M. 332Hashimoto, H. 333Haworth, G.J. 212Hegedus, C.R. 266, 334Heimenz, P.C. 259Hepburn, C. 418Hermsen, A. 334Herold, R.J. 462Herrington, R.M. 258Hesketh, T.R. 418Hespe, H. 419Higginson, W.C.E. 461Hildebrand, J.H. 111Hilker, B.L. 111Hill, R.M. 255Hilyard, N.C. 209Höchtlen, A. 109Hock, K. 258Hoogendoorn, P. 208Hoppe, P. 109Hosoi, K. 333Hu, C.B. 419Huber, L.M. 111Hudales, J.B.M. 258Hutzinger, O. 109Huybrechts, J.T. 332Immergut, E.H. 210Ingram, B.T. 257Ishida, H. 367Ivanov, I.B. 257, 258Jensen, O.E. 260Juran, R. 210Kanner, B. 91, 92, 231, 256, 258Kanner, T.G. 110Keane, N.W. 209Kendrick, T.C. 110, 256Khemani, K.C. 111

Kheshgi, H.S. 242, 259Kim, D. 417Kingston, B.M. 110, 256Kistler, S.F. 259Klempner, D. 417Klincke, M. 82, 111Kloiber, K.A. 266, 334Koberstein, J.T. 418Koczo, K. 258Koczone, J.K. 258Kodama, K. 160, 209Kollmeier, H.J. 109, 257Komarova, A.B. 256Kopusov, L.I. 256Kossmann, H. 333Kovac, J. 211Kozakiewicz, J. 332, 333, 334Kryszewski, M. 332Kücükpinar, E. 202, 211Kulkarni, R.D. 256Kuropka, R. 333Kushner, S.A. 82Kuyper, J. 462Lai, Y-H. 366Lamb, W. 175, 210Langenhagen, R. 111Lassila, K.R. 83Lavielle, L. 367Lawler, L.F. 111Lawrey, B.D. 463Lean, B. 333Lee, K.D. 111Legocka, I. 334Lerner, L.E. 418Leung, L.M. 418Lewandowski, L.H. 366Leyden, D.E. 367Lidy, W. 109, 257Lilaonitkul, A. 418Lin, J.R. 418Lin, J.Y. 209


480

Lindenau, B.E. 211Lipatova, T.E. 256Listemann, M.L. 82, 83Livigini, R.A. 462Lloyd, D.R. 366Lloyd, N.C. 110, 256Locatelli, M. 334Loewrigkeit, P. 333Loos, F. 333Lopes, W.I. 83Luca, J. 154Lucas, H.R. 334Lupinacci, J. 212Macosko, C.W. 110, 111, 215, 257Madan, S. 418Magnano, E. 211Maldarelli, C. 230, 258, 259Malone, B. 209Manini, P. 209, 210, 211Mansfield, K.F. 82Mapleston, P. 154, 155Mark, H.B. 210Markusch, P.H. 333Martin, C. 367Martin, E. 334Maruscelli, E. 332Mascioli, R.L. 462Masuda, Y. 209Mathis, N.E. 192, 211Mattson, J.S. 210Mayenfels, P. 332McCullogh, D.W. 209McElroy, D.L. 208McVey, S.B. 111Mealmaker, W.E. 334Mecea, V. 209Mendelsohn, M.A. 417Mercea, P.V. 209Merriman, P. 110Mestach, D.E. 333Milgram, J. 462

Miller, C.A. 257Minnich, K.E. 83Monteiro, M.J. 333Moraja, M. 211Morbitzer, L. 419Moriarty, J.A. 245, 259Moser, K. 109Muller, E.H. 419Muresan, L. 209Muroi, S. 333Murphy, G.J. 258Mysels, K.J. 214, 245, 246, 257Nachtkamp, K. 332, 333Nagarajan, M.R. 419Naire, S. 259, 260Narducci, E. 211Navish, F.W. 417Neff, R.A. 110, 257Neogi, P. 257Nikolov, A.D. 110Nugent, B.M. 258, 260O’Connor, J.M. 462, 463Oertel, G. 418Ondrey, S.J. 417Oron, A. 259Overbeek, A. 334Overbeek, G.C. 332Owen, M.J. 93, 110, 214, 256, 257, 258Özkadi, F. 211Palinkas, R. 418Palmer, M. 418Palumbo, R. 332Pantone, R.S. 418Panzer, U. 366Papirer, E. 337, 366Pastore, G. 211Pcolinsky, M. 109Penczek, P. 332Penczek, S. 333Pendergast, P. 209Pernisz, U.C. 259, 260

481

Perrut, M. 208Peters, F. 155Petersen, I.H. 110, 256Petrella, R.G. 82Piccinini, M. 334Pile, B. 154Pionteck, J. 418Platzer, N.A.J. 462Plueddemann, E.P. 366Pogany, G.A. 462Pohlers, A. 418Pompe, G. 418Potschke, P. 418Potter, T.F. 212Prausnitz, J.M. 111Price, C.C. 462Prokai, B. 258Pugh, R.J. 257Qin, R. 360, 361, 362, 365Quay, J.R. 419Randall, D. 208, 209Recker, K. 418Redhead, P.A. 209Reed, D. 255Reegen, S.L. 417Reichel, C.J. 463Reid, W.G. 110, 256Reisch, J.W. 462, 463Ricciardi, R. 109Rimmer, R.W. 332Ritter, J. 256Rizzi, E. 210Roester, R.R. 332Rosai, L. 210Rosbotham, I.D. 209, 210, 211Rossmy, G. 94, 109, 256, 257Rosthauser, J.W. 332, 333Roth, A. 209Roy, R.V. 260Ruckenstein, E. 339, 366Rumschitzki, D. 259

Ruschak, K.J. 242, 259Saint Flour, C. 337, 366Sancrotti, M. 211Sand, J.R. 212Sanger, G. 256Savoca, A.C.L. 82Sawyer, D.T. 367Schator, H. 109, 257Schiabel, A. 211Schlöns, H-H. 109Schneider, N.S. 419Schreiber, H.P. 337, 338, 341, 342, 344,347, 359, 360, 361, 362, 366, 367Schuchardt, J.L. 462Schultz, J. 336, 366, 367Schwartz, L.W. 259, 260Schwartz, M. 333Schwarz, E.G. 256Sciuccati, F. 209Scott, R.L. 111Scriven, L.E. 93, 110, 259Seneker, S.D. 462, 463Sengupta, A. 338, 341, 342, 344, 347,359, 360, 361, 362, 365, 367Seyffert, H. 256Seymour, R.W. 418, 419Sheinina, L.S. 256Shen, C.C. 463Shinoda, K. 214, 257Simons, D.M. 461Simpson-Gomes, A. 333Sironi, E. 334Slattery, J.C. 259Smak, Y.W. 334Smith, C.P. 462, 463Smith, L.K. 212Smith, S. 211Smock, D. 154, 155Smudin, D.J. 109Snow, S.A. 251, 255, 257, 258, 259, 260Snyder, W.H. 462

Author Index


482

Solomou, N. 208Soyal, A. 211Srikanth, R. 212Stannett, V. 366Stark, J. 155Stebe, K.J. 230, 258Stein, H.N. 225, 258Sternling, C.V. 93, 110Stevens, R.E. 255, 260Stone, H. 109Stone, H.A. 259Stovall, T. 212Strong, H.M. 208Subramaniam, N. 333Sugiyama, A. 190, 209Sung, W.F. 209Tabor, R.L. 208Tada, H. 209Tailor, K.J. 461, 462Tanimoto, Y. 209Tao, W.H. 209Tarakanov, O.G. 256Taverna, M. 110, 154, 155Taylor, J.R. 333Tennebroek, R. 334Thomas, G. 110Tiedemann, J.E. 417Tirpak, R.E. 333Tobias, J.D. 82Trenholme, H.W. 461Tuck, E.O. 259Tuinman, R.J. 463Turner, R.B. 111, 258Urso, C. 211Valdrè, S. 211

van Bogart, J.W.C. 418van der Hulst, H. 462Van Dyk, K.A. 333Vartan-Boghossian, R. 332Veltkamp, B. 208Vengerovskaya, S.H.G. 256Verbanc, J.J. 461Vineyard, E.A. 212Vratsanos, M.S. 82Wacker, W. 209Wagner, R.D. 109Ward Jr., R.S. 419Ward, T.C. 366Wasan, D.T. 110, 258, 259Weaver, F.J. 208, 210Weidner, D.E. 260Weier, A. 111Wells, R.R. 463Wennertröm, H. 257White, L. 154Wiemann, M. 109, 257Wilhelm, M.R. 418Wilkes, G.L. 111Wilkes, K.E. 210, 212Wittbecker, E.L. 462Wong, H. 259Wressell, A.L. 82Yarbrough, D.W. 208Yasunaga, K. 110, 257Yoshimoto, M. 209Yuge, K. 209Zeiler, R. 175, 210Zellmer, V. 109, 111Zhang, X.D. 93, 110, 111, 257Zharkov, V.V. 256

483

A

Acclaim elastomers 446Acclaim polyol 423, 429, 435, 438, 445,Acclaim polyolsAcetone process 265Acid copolymers 178Acid/base interactions 347Acrylic dispersions 261Acrylic polymer 262Activated charcoal 181Additives 3, 15, 16, 42, 63, 64, 67, 73

Dimensional stability 3, 16, 64Flexible slabstock low emission 73For TDI 16Low emission dimensional stability 42

in MDI 64, 67MDI flexible moulded foam 63TDI flexible moulded 15

Adhesion 218, 325, 337, 344, 353Acid/base interaction 353Bond properties 352Determination 277Performance 344Polyurethane 355Strength 344Tests 350

Adhesion promoter 355, 364Silane 355

Adhesive Ageing 362Air flow values 12Alloy 182

Barium-lithium 182Aluminium 168

Sputtered 168

Analysis 174, 189, 195, 336, 373, 395,403, 405, 410, 412

Differential scanning calorimetric373, 381, 383, 393-395, 405, 406,410Dynamic mechanical 373, 383, 395,396, 405, 407, 412Finite element 174, 196Fourier transform-infra red 373, 403Gel permeation chromatography 450Head space 14, 62Inverse gas chromatography 345,

347-349, 354, 357-358, 365Residual gas 189, 195Surface 336, 338

Anisotropy 160Cellular 160

Appliance 198Household 198

Applications 135, 187Industrial 135Shipping containers 187

Arcol 7ASTM 15

D3574, Indentation force deflectiontest 15Test methods 4

B

Bag 158Gas barrier 158

Barrier 168, 174, 198, 206Aluminium foil-based 198

Main Index

Main Index


484

Gas 168Metallised 174, 206Multilayered plastic 206

Film 167Vacuum properties 167Properties 168

Bashore rebound 373Beamech CO-2 90Blister detachment 350, 352, 353, 354Block foaming 161Blowing agents 4, 87, 88, 136, 178

Alternative 87Carbon dioxide 87, 88CFC-11 87Chloroflurorcarbon 187Physical 87

Bond strength 337, 360, 361, 364Boundary 240

Conditions 240, 245Film-air 240

Bubble coalescence 214Bubble formation. See NucleationBulk stability 13Bulk viscosity 217Butyl acrylate 269

C

Calcium oxide 182CannOxide 129Capillary number 241Carbon dioxide 88, 128, 166

Liquid 88Carbon monoxide 166CarDio 88, 89, 90, 92Carrier 122

Mould 122, 123Cast elastomers 372, 377, 382, 391, 398-402, 409, 411

Hard segment content 401Heat ageing 409, 411Mechanical properties 409

Preparation 372Properties 398, 402Soft segment content 401

Cast poly(ester urethanes) 390Cast poly(ether urethanes) 375Cast polyurethanes 379, 380, 392, 393,397, 401, 404, 413

Analysis 403Hard segments 379Hardness 392, 393Heat aged 413High hardness 401HER/HQEE blends 397Properties 392Soft segments 379Tensile properties 380

Catalyst 16, 43, 52, 75, 81Activity 75Cell opening 16, 52Cell opening blowing 43Non-fugitive 43, 52, 81Selectivity 75

Cell opening 4, 93-94, 214, 217Blow off 94

Cell structure 101, 107Cell wall drainage 4Chain extenders 370Chamber 118

Mixing 118Chem-Trend 7Chem-Trend PRC-798 9Chemical foam model 75

Flexible cellular polyurethane 4Chicken buckets 9Chromatography 276, 337

Gel permeation 276Inverse gas 337, 345. See also IGC

Closed cells 162Coalescing agent 297-300

NMP 297-300Coating 277

Binder 264

485

DPUR 284Drying time 277Properties 292-293, 300, 303, 305,320, 322-326

Coating Coatings 456Aqueous polyurethane/urea dispersion456

Coefficient 232Diffusion 232Surface partition 232

Cohesion 218, 220Intermolecular 218Failure 352

Collapse 23Collapsing wedge model 226COMBOGETTER 181-187, 202, 203,207

for VIP 182Comfort properties 3Compression set 373, 380, 390, 401

Tests 3Container 206

Shipping 206Contamination 165

System 165Control 119

Pour pressure 119Conveyor 126

Oval 126Cooling 87

Lateral 87Creep 161Crosslinking 278

Degree of 278Density 279

Crusher 9Black Brothers Roller 9

Cumene hydroperoxide 269

D

Dabco 6, 80

Deforming surface 236Dehydrohalogenation 42

Thermally induced 42Density 17Differential scanning calorimetry 373,381, 393Diffusion 214, 215, 219, 220, 240Diffusion coefficient 232, 234Dimensional stability 11-12Dispersion 264, 266, 267, 276, 279

Acrylic 264, 267Characterisation 279Hybrid 267Hybrid acrylic-urethane 266Polyurethane 264Properties 323, 324, 325Systems 263

Hybrid 263DMA see AnalysisDMTA 427, 433, 452, 453, 454, 458DPUR particle 312, 315

Swelling of 312, 315Drainage rate 218, 219, 220, 229, 236DSC see AnalysisDynamic creep testing 3Dynamic fatigue 14, 49Dynamic mechanical analysis see Analysis

E

EasyFroth 129Edge effect 174Effect 226

Bulk 226Surface 226

Elastomer processability 424, 438One-shot 438

Elastomers 385, 424, 433MDI/BDO cured 424rebound 433Thermal stability 385

Electron microscopy 277

Main Index


486

Transmission 277Embedded sphere 329Emissions 4Emulsion polymerisation 265Energy 218

Cohesive 218Efficiency 157

Envirocure 87Environmental concernS 86Equipment 114

Metering 114, 115

F

FEA See AnalysisFiller 158

Aerogels 158Compressed powder 158Fibre 158Fibre glass 158Perlite 158, 189Precipitated silica 189Silica 158

Film 134, 178, 213, 215-217, 219, 221,226, 228, 231, 278, 317

Adhesive 134Bulk viscosity 215Crosslink density 327Drainage rate 216, 217, 226, 227,228Formation 223Formulation 221Free surface 239Mechanical properties 278Mobile surface 217Models 239

Deforming 239Outgassing 178Properties 291, 293, 299, 302, 304,308-309, 311, 320-321, 323-325Releasing 134Rigid surface 216, 228, 231

Film rupture 215, 216, 218, 221Film stability 235Surface free energy 317Tangentially-immobile 242Thin liquid 213Thin liquid polyurethane 219Wedge 236

Collapsing 236Film drainage 219, 255

Rate 227-229, 232, 235Surface variables 228

Vertical liquid 255Fingering pattern 223, 254FipurTec 139Fire-retardants 96, 97, 98, 108Flexible moulded foam 14, 46, 63, 73

MDI 14, 68, 73TDI 63

Flexible slabstock 73-74, 88, 90, 95, 99Polyether 88Polyurethane 229

FlexiDrum 126Flux 219, 227, 230Foam 74, 76, 78-80, 85, 88, 107, 128, 157-159, 161, 163, 177, 179, 201, 213, 254

CFC-free 157Cooling 87Density 86

Hardness 86Minimum 86

Encapsulating 177Evacuated 161Flexible slabstock 75Formation 86Hardness 87Soft 87MDI flexible 79

Moulded 76Mechanical properties 161Multi-density 128Multi-hardness 128Open cell 163, 179, 201

487

Open cell PU 158, 163Physical properties 163Polyether 88Polystyrene 159Polyurethane 85, 96, 213Porosity 161Rise 93

Stabilisation 93Slabstock 80TDI flexible 79Moulded 76Slabstock 74, 78

Foam & film 133Foam One 87Foaming 87, 90, 128

Carousel 128In situ 128Liquid carbon dioxide 90Variable pressure 87

Fogging 14, 62Force to crush 3, 23, 66, 71, 72Formulations 72

MDI 72Fracture energy 390Freezer 177, 204

Ultra-low temperature 204Froth 101, 213

Density 101Gelatinous 213

FTC see Force to crushFull rise time 15

G

Gas permeation 168, 176, 177Flanges 176, 177

Gas transmission 168Gate Bar 88Gauge 192

Spinning rotor 192Gel time 15Getter 158, 170, 181, 183, 185, 193

Device 179Material 181Non-evaporable 170SpiroTorr 193, 194, 201Technology 206

Gibbs-Marangoni 93Global warming 157Goniometer 336

Rame-Hart 336GPC See AnalysisGradient 230Gravitational energy 214Gravity convection 225Greenhouse effect 157Guarded hot-plate 191

H

Handmix procedure 8Mass-loss 18Flexible moulded foam 8Flexible slabstock foam 8

Hard segment 378Content 376, 378

Hardness 86-87, 378, 401, 408Determination 277

Heat flow meter 191Heat reflectance 192

Modified hot wire 192HER materials 369, 370, 371, 373

Characterisation 373Polyurethane applications 369, 371Synthesis 373

HER/HQEE blends 397, 398Freezing point determination 397

High vacuum benches 163Hot box 191

Guarded-calibrated 191Hybrid 301, 306

Chemical structure 306, 325Crosslinking 306, 326Polyurethane-urea 301

Main Index


488

Hybrid acrylic-urethane dispersions 261Hybrid dispersion 269, 320

Chemical structure 324Coalescent 323Crosslinking 312, 317Film 328Initiator 320Particle size 320, 321Properties 290, 291, 292, 298, 302,303, 307, 311, 320, 330Surface free energy 319Synthesis 269, 290, 320, 322

Hybrid dispersion particlesMorphology 317, 328

Hybrid dispersion system 262Preparation 262

Hybrid polymer dispersion 262Hydrogen 166

I

IGC See AnalysisINSOTEC 135Instabilities 215, 218, 235

Fluid 215Benard 215Marangoni 215, 235Rayleigh-Taylor 215

Marangoni 218Insulation 196Interference fringe 222, 223, 224, 225,226, 231, 236Interferometer 221Interpenetrating polymer network 262InterWet 138, 143, 144, 145, 146, 147,150, 153

Advantages 150Applications 152Improvements 148

Ionomers 178Isopach 226Isophorone diisocyanate 268

K

Krauss-Maffei 14Kyoto conference 157

L

Laboratory moisture-curing 4662,4-TDI prepolymers 466

Lamella 214, 215Laminate 201

Aluminium foil 201Lamination 161, 162Lap-shear 337, 344, 347, 352, 353, 354,362Liquid carbon dioxide blowing 85Long fibre injection 139Long-term vibration 3Loss compliance 388, 414, 415, 416Loss modulus 385, 386, 414Lubrication theory 240, 242, 245, 249

M

Machine 9Hi Tech SureShot MHR-50 9

Machine evaluation 9, 25, 29, 33Back formulation utilising TPR 29Back formulation without TPR 33Cushion formulation utilising TPR 25Cushion formulation without TPR 33TDI flexible moulded foam 9

MarangoniEffect 249, 250, 251, 252, 254Flows 215, 225

Marginal regeneration 225Mass spectrometer 163, 164, 189

Quadrupole 163Mass-loss/rate-of-rise 4Maze flow mould 10, 56MDI Prepolymers 459-460, 465

Cured with BDO 459, 460

489

Laboratory casting 465Laboratory preparation 465

Mechanical crushing 3Mechanical properties 372Melting points 374Metal foil 168Methyl methacrylate 269Micelles 273Micro-leaks 191Microscopy 4, 40

Scanning electron 4, 40Mill

Premier Mill Corporation dispersator8

Mixing head 116, 120, 142Variable geometry 120

Molecular sieves 181Mondur TD-80 8Monolayer 218, 226Monomer 272, 275

as active diluents 275Polymerisation 272

Montreal Protocol 157Morphology 262, 264, 267, 272

Dispersion particle 262, 272Engulfed 264Fruit-cake 264Gradient 264Particle 267

Mould 56, 60, 124Carrier 129Carrier exchange 125Carrying system 126FlexiDrum 126Maze flow 56, 57, 58, 59, 60Temperature control system 124

Moulded foam 3, 68Low emission 68

Moulded formulation 69, 70MDI 69, 70, 71

Moulding 127, 130Flexible foam 127

Polyurethane 130Multi-component operation 117

N

N-methylpyrrolidone 269Navier-Stokes equation 220, 238, 239NEG alloy 181

Zirconium-based 181Newtonian fluid 239Nitrogen 166NovaFlex 90Nucleation 91, 101, 105, 215Nylon 168

O

OEMSpecifications 3

One-shot elastomer 436-437, 439-443,465

Hardness buildup 437Laboratory casting 465Processability 436Properties 438

Open cell 158, 159PU 159PU foam 158

Ostwald ripening 214Outgassing 163, 165, 202

Load 166Oxyethylene moieties 434Oxygen transmission 176

P

Pail test 13, 56Panel 158, 177, 207

Encapsulated 177Lifetime of 158Open cell foam 207Partial pressure 164

Main Index


490

Particle 327Morphology 330Swelling 327

Particle formation 326Particle size 277, 313, 314

Distribution 277, 313PDMS 218Péclet number 241, 251Permeation 170, 171, 172

Helium 170, 171, 172Physical properties 3, 23, 33, 35, 37-39,47, 66, 68-70, 74, 372

Flexible slabstock foam 75Foam 23MDI 69, 70Machine data 33Review 39TPR 35, 37

Plateau borders 93, 214, 225, 226Gibb’s 93

Polarity 5Polyacrylonitrile 178Polycat 6Polyester 168, 178Polyether chains 218Polyethylene 176Polyethylene terephthalate 168Polyisocyanate 85Polymer dispersions 261Polymerisation 270

Radical 270Polyol 75, 77, 78, 449

EO-tipped propylene oxide 77Ethylene oxide tipped 75Molecular weight distribution 449Polyolefin tipped 78

Polyol functionality 428Polyoxyethylene 215Polyoxypropylene glycol 421

Ultra-low monol 421Polysilicate 217

Trimethylsilyl-capped 217

Polytetramethylene glycol 268Polyurea 215Polyurethanes 85, 92, 113, 138, 139,157, 262, 335, 347, 405, 449

Adhesion 347Adhesive 335, 360

Formulations 335IR spectra 341Silane 356Silane additives 360

Co-injection 138Cure 338Dynamic properties 449Filled 196

Open cell 196Film 216, 236

Draining thin liquid 236Rate of drainage 216Vertical 236

Foam 166, 200Closed cell 200Open cell 166Insulation 157

Glass-reinforced 139High thermal stability 405History 85Markets 92, 113Mechanical properties 449Multi-component formulations 113Two-component 335

Polyurethane castings 376Polyurethane catalysts 42

Blowing 42Non-fugitive gelling 42

Polyurethane elastomer 430, 431Polyurethane Foams 3, 5, 93

Catalysts 5Flexible 3, 93Non-fugitive catalysts 5Tertiary amine catalysts 5

491

Polyurethane-urea 324Double bonds 324

Polyurethane-urea-acrylic dispersions261Polyurethane/urea dispersions 466

Formulations 457Laboratory preparation 466

Polyvinylidene chloride 178Pot life 372, 375

Determination 375, 390PPG polyols 424, 429, 436

One-shot elastomer system 436Processability 429Processing latitude 434Property latitude 429Ultra-low monol 424, 444

PPG/PTMEG blends 447Elastomer properties 447

Prepolymer-ionomer 265, 275, 276Processing 113, 167, 179

Desorption 166Exhaust 179Latitude 3Polyurethane 113Seal-off 179Thermally activated 167Yield 187

Propionic acid, dimethylol 269PU see PolyurethanePVC 136

Q

Quality assurance 191Quality control 172, 191

R

Rapid Cure 87Rate 169

Helium transmission 169Water transmission 169

Rate-of-rise 18Raw materials 92

Emulsification 92Reactive amine catalysts 5Reactivity 18, 23, 45, 65, 74

Handmix rate of rise 65Machine free rise 23

Rebound properties 380, 390, 401Recycling 162

Refrigerators 162Reeves Brothers 87Refrigerator 177, 203

Biomedical 203Laboratory 203

Reinforcement 141Handling 141

Reynold’sEquation 220, 227Model 216Number 241

Rigid foam 158Open cell 158

Robots 121, 145Cartesian 122Foaming 121

Robotics 145

S

Scalloping 23, 25Seating 17

Automotive 113, 122, 130Back formulation 19-22, 30-34, 36,

38, 40, 44-46, 48, 60, 113Cushion formulation 19-22, 26-29,34-35, 37, 39, 43-44, 46, 49, 57, 59Polyurethane 17PU back formulation 18PU cushion formulation 17

Service station 125Shear modulus 386Shore hardness 372

Main Index


492

Shrinkage 12, 61Foam 61Template 12

Silanes 356Additive 364

Silica 98Silicone surfactant 4, 95, 99, 213, 215-217, 219-220, 228-229, 254

Burning 95Characterisation data 233Chemistry 99, 103Dabco 4Stabilisation 93

Soft segment content 376Solvated volume 232, 234Sound deadening 130Sponge 214Stoke’s law 219Storage modulus 384, 385, 412Stress conditions 239, 246

Normal 239, 246Tangential 239, 246

Stress/strain curves 426, 440, 447, 458Structure 174

Laminate 174Styrene 269

Polymer dispersions 261Surface characteristics 335Surface energy 213, 214Surface free energy 278, 317Surface partition coefficient 219, 220Surface tension 91, 230, 276

Gradient 218-220, 225-226,229, 232, 234-235, 250, 252-253Surface viscosity 217Surfactants 64, 91-92, 101, 108, 213,229, 232

Activity 92Concentration 229Development 101Emulsification 92Flexible moulded 64

Molecular structure 232Non-ionic fluorocarbon 229Silicone 91, 92, 100, 213

Synthesis 288MDPUR 288MDPUR-ASD 288

T

Tan delta 387, 414TDI 80, 465

Laboratory preparation 465MBOCA 450, 453Mechanical properties 454Moisture-cured 454Prepolymers 450, 453, 454, 466

Laboratory casting 466Tear 380, 390, 401

Resistance 373Strength 410

Technology 130Foam & film 130-132, 134-138

Tegostab 102Tensile 401, 408Tensile

Measurements 410Modulus 372Properties 378, 390, 392

Thermal conductivity 159, 174, 175, 180Thermoregulation 124Time pressure release test 13Toluene diisocyanate 85TPR 23TPR times 33Transmission rate 169

Helium 173Water 169, 173

Transport 126Systems 126

Treatment 165Heat 165

493

U

Urethane elastomers 413Heat aged 413

Urethanes 335Adhesion behaviour 335

V

Vacuum panel 158, 161, 188, 191Characterisation 188Insulating performance 191Manufacturing process 188

Vacuum panel technology 158Van der Waals bonding 218Vending machines 177, 204Vibrathane 375-376, 378, 383Video imaging 106Video microscopy 103, 105Vinyl staining 42, 62VIP 190, 193, 197, 202, 203

Longevity 203Performance 202Reliability 202

Flange 168Viscosity 217-218, 220, 227-228, 245,276

Bulk 220Bulk dynamic 227Finite surface 245Surface 217-218, 220, 225, 227-232,235-237, 242, 255Surface dilatational 240Surface shear 240

Volatile organic compounds 4Voranol 7, 80

W

Water resistance 277Water transmission 176Window fogging 42

Z

Zeolite 181Zeta potential 314

Main Index


494

Rapra Technology Limited

Rapra Technology is the leading independent international organisation with over 80 years of experience providing technology, information and consultancy on all aspects of rubbers and plastics.

The company has extensive processing, analytical and testing laboratory facilities and expertise, and produces a range of engineering and data management software products, and computerised knowledge-based systems.

Rapra also publishes books, technical journals, reports, technological and business surveys, conference proceedings and trade directories. These publishing activities are supported by an Information Centre which maintains and develops the world’s most comprehensive database of commercial and technical information on rubbers and plastics.

Shawbury, Shrewsbury, Shropshire SY4 4NR, UKTelephone: +44 (0)1939 250383 Fax: +44 (0)1939 251118

http://www.rapra.net

ISBN: 1-85957-275-8

urethane science and technology

Documents

owens corning

xi xii ix

jacketed mechanical

ideal gas

slow argon

dimensional

90physical

105physical