rights / license: research collection in copyright - non ......it is concerned with the improvement...
TRANSCRIPT
Research Collection
Doctoral Thesis
Improvement of One-Component Polyurethane Bonded WoodenJoints under Wet Conditions
Author(s): Kläusler, Oliver F.
Publication Date: 2014
Permanent Link: https://doi.org/10.3929/ethz-a-010280586
Rights / License: In Copyright - Non-Commercial Use Permitted
This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.
ETH Library
DISS. ETH NO. 22157
IMPROVEMENT OF
ONE-COMPONENT POLYURETHANE BONDED
WOODEN JOINTS UNDER WET CONDITIONS
A thesis submitted to attain the degree of
DOCTOR OF SCIENCE of ETH ZURICH
(Dr. sc. ETH Zurich)
presented by
OLIVER FREDERIK KLÄUSLER
Dipl. Holzwirt, University of Hamburg
born May 29th 1972 in Solothurn
citizen of Zürich (ZH), Herznach (AG) and Germany
accepted on the recommendation of
Prof. Dr.-Ing. habil. Dr. h.c. Peter Niemz, ETH Zürich, examiner
Prof. Dr. Ingo Burgert, ETH Zürich, co-examiner
Prof. Dr. Rupert Wimmer, BOKU, Vienna, co-examiner
2014
© by Oliver F. Kläusler
All rights reserved.
Dedicated to
my wife Alessandra Cristina
and our three sons
Jonas Frederik, Benjamin Robert and David Luca
Preface
Nowadays architects, civil engineers and their customers are becoming more and more aware
of the limitations regarding the availability of building materials. Therefore renewable natural
resources like wood are increasingly gaining attention. The overwhelming majority of today’s
wooden load bearing construction materials, such as glulam, cross laminated timber (CLT) or
laminated veneer lumber (LVL), is produced by using adhesive bonding technology. As a
basis for the required safety, such bondings have to provide long term reliability under various
ambient climate conditions.
This thesis is focused on the improvement of the performance of 1C PUR bonded wooden
joints for load bearing structures under moisture load. Such joints show good and reliable
performance in the dry state, but do not meet the requirements of the Canadian standard CSA
O112.9-04 regarding wood failure percentage (WFP) in the wet state. It is the overriding goal
of this thesis to find new approaches to enhance the WFP of such joints under high moisture
conditions. Such an approach could contribute to the basic understanding of 1C PUR bonded
joints and help 1C PUR manufacturers to open new sales markets in North America. 1C PUR
is the prefered kind of adhesive in this thesis, since it does not introduce formaldehyde into
the gluing procedure and is comparatively easy to handle. Furthermore, the cured glued joints
do not emit VOC.
This thesis is part of a research project, which is mainly concerned with the optimization of
1C PURs used for the structural bonding of wood in highly humid environments. Project
partners are the Commission for Technology and Innovation (CTI) of the Federal Department
of Economic Affairs, Education and Research (Bern, Switzerland), the adhesive manufacturer
Purbond AG (Sempach-Station, Switzerland), and the ETH Zurich (Swiss Federal Institute of
Technology, Switzerland).
Acknowledgements
During my work on this thesis I received substantial support from a number of people, whom
I would like to thank most sincerely. First of all I would like to thank my supervisor
Prof. Dr.-Ing. habil. Dr. h.c. Peter Niemz for the opportunity to work on this CTI- project and
for his strong and continuous support, particularly during quite challenging project phases.
Also I would like to thank my co-examiners Prof. Dr. Ingo Burgert and Prof. Dr. Rupert
Wimmer for their experienced advice and the very friendly and positive working cooperation.
Furthermore I would like to thank the industrial partners of this project: The Purbond AG,
represented by Walter Stampfli, Dr. Joseph Gabriel and Dr. Carlos Amen, not only
contributed a substantial part of the project funding, but also many fruitful discussions and
various additional inputs, such as modified adhesives. Also I would like to express my
gratitude to the Currenta GmbH, represented by Dr. Alexander Karbach and Wilhelm
Bergmeier, for their very professional contributions in terms of analytical methods and
procedures.
Moreover, I am very thankful for the personal and professional support I received from the
members of the ETH Wood Physics Group over many years, since I came to the ETH for the
first time in 2003. In particular, Dr. Sebastian Clauss, who was my HIF office colleague and
one of the project initiators, was a great help for me. Dr. Philipp Hass, who was Sebastian’s
successor in the office, took on the job of proofreading parts of my papers and of this thesis.
From Dr. Walter Sonderegger and Dr. Michaela Zauner I received valuable hints regarding
the “dos and don’ts” in the final stage of the thesis.
I am especially grateful for the immense and continuous support given by my wife Alessandra
and for the liveliness of our three youngsters at home, offering me the sometimes needed
distraction from project issues. Thank you!
Table of contents
Summary .................................................................................................................................... 1
Zusammenfassung ...................................................................................................................... 3
1. Introduction ........................................................................................................................... 5
1.1 Relevance and motivation ............................................................................................... 5
1.2 Research gap, objectives and project partners ................................................................ 8
1.3 Work outline and methods .............................................................................................. 9
2. State of the art ..................................................................................................................... 13
2.1 Literature overview on wood failure percentage of 1C PUR bonded joints ................. 13
2.2 Relevant technical standards ......................................................................................... 13
2.3 Wood ............................................................................................................................. 16
2.3.1 Macroscopic anatomy ............................................................................................ 16
2.3.2 Microscopic anatomy ............................................................................................. 17
2.3.3 Ultrascopic anatomy ............................................................................................... 21
2.3.4 Wood species used in this thesis ............................................................................ 23
2.3.5 Wood as an adherend ............................................................................................. 27
2.4 One-component polyurethane adhesives for the structural bonding of wood ............... 28
2.4.1 Basic chemistry of 1C PUR wood adhesives ......................................................... 28
2.4.2 Benefits and drawbacks of 1C PUR for structural bonding of wood ..................... 30
3. Preliminary investigations ................................................................................................... 32
3.1 Round robin test ............................................................................................................ 32
3.2 Tests on wood failure percentage in the wet state ......................................................... 33
3.3 Tests with modified 1C PUR ........................................................................................ 36
3.3.1 1C PUR with modified filler content ..................................................................... 37
3.3.2 1C PUR with modified molecular weight .............................................................. 38
3.3.3 1C PUR with more hydrophilic pMDI ................................................................... 38
3.4 Tests with modified polyurethane prepolymers ............................................................ 38
4. Main investigations ............................................................................................................. 39
4.1 Paper I ........................................................................................................................... 40
4.2 Paper II .......................................................................................................................... 63
4.3 Paper III ......................................................................................................................... 82
4.4 Paper IV ....................................................................................................................... 110
5. Additional investigations ................................................................................................... 133
5.1 Sorption behaviour of adhesive films .......................................................................... 133
5.2 Diffusion behaviour of adhesive films ........................................................................ 134
6. Synthesis ............................................................................................................................ 138
6.1 Main findings .............................................................................................................. 138
6.2 Conclusive discussion ................................................................................................. 140
6.3 Potential for future research ........................................................................................ 142
7. References ......................................................................................................................... 144
List of abbreviations
1C PUR One- component moisture- curing polyurethane wood adhesive
ASTM American society for testing and materials
CLT Cross laminated timber
CSA Canadian standards association
DMAC Dimethylacetamid
DMF N,N-dimethylformamide
EMC Equilibrium wood moisture content
EN European standard
EO Ethylene oxide
EPI Emulsion polymerized isocyanate
ESEM Environmental scanning electron microscope
EU Europe / European
Glulam Glued laminated timber
HMR Hydroxymethylated resorcinol
HS Hard segments of the cured polyurethane
LVL Laminated veneer lumber
MFA Cellulose micro fibril angle
MOE Modulus of elasticity [MPa], [GPa]
MUF Melamine urea formaldehyde adhesive
NA North America / North American
NCO Isocyanate
PO Propylene oxide
Prepo Polyurethane prepolymer
PRF Phenol resorcinol formaldehyde wood adhesive
RH Relative humidity of the ambient air [%]
SS Soft segments of the cured polyurethane
TSS Tensile shear strength [MPa], [N/mm2]
UF Urea formaldehyde adhesive
VOC Volatile organic compounds
WFP Wood failure percentage [%]
WMC Wood moisture content [%]
1
Summary
This thesis has been carried out in the framework of a research project supported by the Swiss
Federal Commission for Technology and Innovation (CTI, Bern). It was initiated by a
research cooperation, involving the adhesive manufacturer Purbond AG (Sempach- Station)
and the Wood Physics Group of ETH Zürich. It is concerned with the improvement of wood
failure percentage (WFP) of one-component moisture-curing polyurethane (1C PUR) bonded
wooden joints in the wet state. The project was spurred by the fact that 1C PUR bonded
wooden joints have difficulties fulfilling the demands of the Canadian standard CSA O112.9-
04 regarding WFP on the fracture surface after testing in the wet state. A substantial
improvement in this field was necessary, since this could contribute to the basic
understanding of 1C PUR bonding of wood and also open up Canada as a market for
structural 1C PUR wood adhesives.
The investigations presented in this thesis are mainly concerned with the effect of moisture on
the mechanical properties of adhesive polymers and interactions between wood and 1C PUR
under high moisture load. The findings illustrate strengths and weaknesses of wooden 1C
PUR bonded joints and will hereby help to make the behaviour of the accordant wooden
constructions under moisture load even more predictable.
Finally the solvent DMF was found to be helpful when used as a primer for 1C PUR bonded
wooden joints. The adhesion between the adhesive polymer and the wood improved
substantially, making it basically possible to meet the demands of the Canadian standard
mentioned above with Douglas fir. For a variety of reasons the use of a toxic adhesion
promoter like DMF is clearly not the optimal solution. However, the findings described in this
work partly served as a basis for the industrial development of more suitable primers.
2
Furthermore, during the preparation of this thesis the DMF- priming was tested in a pilot
project on an industrial scale. The results reveal that by means of this method and under
appropriate conditions the tendency of 1C PUR bonded glulams to delaminate can also be
reduced significantly. Hence, more research in this field would be reasonable, because a more
essential understanding of the detected effects could contribute to an even higher reliability of
polyurethane bonded wooden constructions in the future.
The potential for further improvements of 1C PUR bonded wooden joints has not been
exhausted yet.
3
Zusammenfassung
Die vorliegende Arbeit ist im Rahmen eines Projektes der Kommission für Technologie und
Innovation des Bundes (KTI, Bern) entstanden. Das Projekt wurde von einer
Forschungskooperation initiiert, bestehend aus dem Klebstoffhersteller Purbond AG
(Sempach-Station) und der Gruppe Holzphysik der ETH Zürich. Das Projektziel bestand
darin, die Holzbruchanteile von nass geprüften Holzverklebungen zu verbessern. Diese
wurden unter Verwendung eines Einkomponenten-Polyurethanklebstoffes (1K PUR)
hergestellt. Des Weiteren war es das Bestreben dieser Arbeit, den Holzbruchanteil, der bei
nassgeprüften 1K PUR–Holzverklebungen im Regelfall bei etwa 20% liegt, auf ca. 80% zu
erhöhen. Diesen Grenzwert setzt z.B. die kanadische Norm CSA O 112.9-04. Sollte dieses
Ziel erreicht werden, so würde dies sowohl zu neuen wissenschaftlichen Erkenntnissen
führen, als auch der Firma Purbond den Zugang zum kanadischen Markt für die Herstellung
von formaldehydfrei-verklebten tragenden Holzbauteilen ermöglichen. Bis heute wird dieser
Markt von formaldehyd-basierten Polykondensationsklebstoffen (in NA v.a. PRF) beherrscht.
Die Untersuchungen konzentrierten sich somit zunächst auf den Einfluss von
Umgebungsfeuchte auf die mechanischen Eigenschaften von Klebstoffpolymeren sowie auf
die Wechselwirkung zwischen den hölzernen Fügeteilen und 1K PUR unter starker
Feuchtebelastung. Die Ergebnisse zeigten, dass unter nassen Bedingungen ein
Adhäsionsverlust zwischen Klebstoff und Fügeteilen auftritt, der die Kraftübertragung
erheblich reduziert.
Schliesslich gelang es, mittels eines DMF-Priming-Verfahrens die erforderliche Adhäsion
herzustellen. In der Folge konnten die Normansprüche an Douglasienholz erfüllt werden.
Dennoch ist die Verwendung von toxischem DMF als Haftvermittler für 1K PUR verklebte
Bauteile gewiss keine praktikable Lösung für eine industrielle Anwendung.
Die Untersuchungsergebnisse konnten einen Beitrag zur industriellen Entwicklung von
weniger toxischen und verfahrenstechnisch geeigneteren Primern leisten. Daher bleibt zu
hoffen, dass der Effekt von DMF auf 1K PUR-Verklebungen in zukünftigen Projekten weiter
untersucht wird. Ein tiefergehendes Verständnis der entsprechenden Wirkmechanismen
könnte zu einer weiteren Verbesserung der Holzverklebung führen.
4
Dennoch ist das Priming lediglich ein Hilfsverfahren, das in der Zukunft möglichst durch
Einkomponenten- oder Zweikomponenten-Klebstoffe abgelöst werden sollte.
Während des Verfassens dieser Arbeit wurde das DMF-Priming in einem industriellen
Pilotversuch eingesetzt. Es zeigte sich, dass dieses Verfahren unter geeigneten Bedingungen
auch die Delaminierungsbeständigkeit von 1K PUR-Verklebungen deutlich verbessern kann.
Die Möglichkeiten zur Optimierung von 1K-PUR-verklebten Holzverbindungen sind noch
nicht ausgereizt.
5
1. Introduction
1.1 Relevance and motivation
At present, the application of wood in construction is experiencing a noteworthy upswing.
This is due to the increasing interest of investors, architects and society in issues like the
development of sustainable infrastructures. Aspects like the renewability of building
materials, the life cycle assessment of structures (incl. their carbon footprint) or the energy
efficiency of buildings (incl. their grey energy) are more and more gaining attention.
Therefore the building with wood and also the accordant wood research are financially
supported by various public funding programs, such as the Wood First Act in Canada [1] or
action plan wood [2] and NRP66 [3] in Switzerland. Furthermore, planners and engineers are
becoming increasingly aware of the technological advantages that wooden constructions have
to offer, such as good weight to strength ratio, low deformation under load in case of fire
(compared to load bearing steel elements), advantageous elastic behaviour during earthquakes
and the high degree of industrial prefabrication of wooden building components. Furthermore
design aspects, like the high value appearance of filigree hardwood beams with their more
homogeneous surface are becoming more and more important.
In the recent past these developments have resulted into several quite innovative building
projects, revealing an open-minded interdisciplinary cooperation between investors,
engineers, architects and the wood industry. Following the natural human strive for
development, the involved partners are attempting to extend the boundaries of the state of the
art design and technology of adhesively bonded wooden constructions. This concerns the
dimensions and the design, as well as the fields of application of the wooden structures.
Recent examples of extended dimensions are the Neumatt bridge, at 60m it is the longest free
spanned wooden bridge in Switzerland (inaugurated 2013) [4, 5] and the Saldome 2. With a
span of 120m it is the biggest self- supported wooden domed structure (space framework) in
Europe (inaugurated in 2012) [6]. Remarkable representatives of new design approaches are
for instance the Golfclub building Haesley Nine Bridges in Yeoju, South Korea, with its tree
like curved beams, manufactured by means of 3D CAD technology (realized in 2009) [7], or
the new administration building complex for Swatch and Omega in Biel (Switzerland), which
mimics snake like shapes (construction starting in 2015) [8]. Both projects are examples of
the current stream of bio-inspired wood architecture [9]. Instead, the design of the new
6
Tamedia building in Zürich (seven stories high, inaugurated in 2013) was driven by a demand
for the consistent use of wood (glulam) as the only material for the whole load bearing
skeleton construction. Hence, engineers and planners had to avoid any metal inserts or metal
fittings. Finally, the whole supporting structure was assembled as a wooden connector system.
Specific structural reinforcements at the node connectors were achieved by gluing hardwood
inserts into the softwood elements [10]. Such a combination of hardwood (European ash) and
softwood (spruce / fir) within structural elements (so called “hybrid elements”) was also
employed in the construction of a new general purpose building in Arosa, Switzerland (car
parking, ski school centre, railway service rooms). The building was finished in 2010. Here,
the craft was supplemented by pure hardwood beams (European ash), making quite filigree
elements with comparatively long spans possible [11]. Softwood - hardwood hybrid elements
are also used for the ETH- House of Natural Resources in Zürich (construction started in
2014). In this case a special post-tensioned timber frame construction (held together by a
strong and adjustable steel wire cable) provides the advantages of high stiffness and high
strength in combination with particular economic efficiency due to a high degree of
prefabrication and a comparatively easy deconstruction process at the end of the building’s
service life [12]. The concept of prefabrication is not limited to the beams, as it can also be
applied to room elements (wall or floor elements) or even to modular room units (e.g.
comprising walls, ceiling, floor and parts of the interior installations). This enlarges the fields
of applications for wooden constructions, which therefore can be built much easier and faster
at places far away from access roads, power lines or other convenient infrastructure. A recent
example for such constructions is the high alpine new Monte Rosa hut (canton Valais,
Switzerland). Helicopters transported the CLT- elements to the construction site at an altitude
of 2900 m above sea level, where they were assembled to the six story high mountain hut
(building’s volume 3700 m3, topping out ceremony in 2009) [13] in a couple of weeks.
Further examples of enlarged applications from the recent past are the Timber Tower (a 1C
PUR CLT- construction), which was built near Hannover (Germany) in 2012, carrying a 1.5
MW wind power station with a hub height of about 100 m [14, 15]. Also wooden residential
buildings are reaching new heights: In Bergen (Norway) the completion of the first 14 story
high glulam building is scheduled for the end of 2014. In Vancouver (Canada) feasibility
studies for a 40 story high wooden “skyscraper” are being discussed (with special focus on
static, earthquake safety, fire safety, energy efficiency and life cycle assessment) [16].
7
In summary, the endeavour to extend the application of adhesively bonded wooden
constructions is enthusiastically moving forward, which in turn necessitates reliable wood
bondings, since they provide an important contribution to the serviceability of the different
structures under varying service conditions. Besides the quality of the wood itself and a
proper execution of an appropriate gluing procedure, the quality of adhesively bonded joints
heavily depends on the used adhesive polymer system. The serious accident, which happened
in Bad Reichenhall (Germany) in January 2006, illustrates what can happen if bonded
structural joints fail: The wooden roof construction of the Ice-Sport Centre collapsed; killing
15 people. The reasons for this disaster were multifactorial, but the use of the wrong adhesive
was identified as one of the main factors. For bonding the wooden box girders of the roof
construction, a urea-formaldehyde resin (UF) was used, which does not withstand high
moisture loads in the long term [17, 18]. For making a good adhesive choice, it is essential to
know about the potentials and the limits of the various adhesive systems on the one hand and
about the service conditions of the construction on the other.
In the structural timber industry, as well as in the wood panel industry, the choice for a
specific adhesive is not only dependent on the mechanical performance and the costs, but also
on availability and environmental aspects (also affecting the costs) [19]. Furthermore, product
liability and health concerns gain in importance (customer protection and protection of
employees in production facilities). Therefore, a glulam producer would best avoid the
storage or use of substances that may cause legal or health concerns, as much as possible.
Since the WHO declared formaldehyde as carcinogenic to humans [20], the use of 1C PURs
for the production of bonded structural elements has been increasing. In contrast to more
traditional adhesive systems for structural bonding of wood (like MUF or PRF), no additional
VOC or formaldehyde is emitted when using 1C PUR; neither in the glulam production
facility nor from the installed glued product. However, there are still obstacles that impede a
more extended use of 1C PUR. Two of them are the performance of 1C PUR bonded wood
under high thermal loads and under high moisture load. However, remarkable progress has
recently been achieved regarding the heat resistance of 1C PUR bonded wooden joints [21-
24]. Yet, resistance against moisture impact remains as one of the major issues preventing 1C
PUR from being introduced on to the NA market for structural bonding of wood at a large
scale. This issue is the main topic of this thesis.
8
1.2 Research gap, objectives and project partners
Despite the fact that 1C PURs are being increasingly used in the field of load bearing timber
constructions, there are still some drawbacks associated with such adhesives (see 2.4.2). One
of these drawbacks is the reduction of WFP when the joints are tested in high moisture
conditions. As a consequence, 1C PUR bonded specimens have difficulties in meeting the
demands of NA technical standards, since they set high WFP thresholds at about 80%
(depending on wood species and testing procedure) for glued joints tested under shear load in
the wet state. The accordant 1C PUR bonded joints generate only about 20% WFP, whilst the
more traditional formaldehyde based glulam adhesives (PRF, MUF) are capable of generating
at least about 80% WFP under said conditions. This fact has not proven to be an obstacle for
the use of 1C PUR in Europe, since EU standards do not set thresholds for WFP (regarding
shear tests of wet samples). Concerning shear strength, 1C PUR bonded wooden joints meet
the demands of all the standards (NA & EU) in dry and wet states. However, fundamental
approaches to improve the WFP in the wet state are still lacking, since the specific reasons for
this phenomenon are not yet clear.
Therefore the main objective of this thesis is to investigate the reasons for the low WFP and,
consecutively, countermeasures are proposed. The overriding goal is to find a practicable
method for meeting the accordant WFP thresholds of the NA standards.
The research project was funded by a cooperation, consisting of the partners CTI (Federal
Commission for Technology and Innovation in Bern, Switzerland), the adhesive producer
Purbond AG (Sempach-Station, Switzerland) and the Wood Physics Group of the Federal
Institute of Technology (ETH Zürich, Switzerland).
9
1.3 Work outline and methods
In order to find out why 1C PURs generate less WFP in the wet state than other adhesives
(such as PRF), in a first step some preliminary investigations were carried out (see 3.). These
pre-tests comprised a round robin test, a set of initial shear tests (including the wet state) and
finally tensile shear tests using modified 1C PURs.
By means of the round robin test (see 3.1) it was assured that the laboratory procedures are
carried out in an appropriate manner, using calibrated devices in good working order. The
results confirmed the comparability of the testing results achieved in both laboratories
(Purbond AG and ETH).
Subsequently, preliminary tensile shear tests according to prEN 302-1 (2011) were carried out
(see 3.2). Thus, the results regarding TSS and WFP tested in the dry and wet states mentioned
in the literature were reliably reproduced.
This was followed by tensile shear tests using specifically modified 1C PURs (see 3.3). This
was carried out to investigate, whether a specific variation of the filler content, of the pMDI’s
molecular weight or of the hydrophilic properties of the PUR open a pathway to higher WFP
under humid conditions.
The tests with modified 1C PUR revealed that the goal of attaining a higher WFP in the wet
state is hardly achievable by means of a slight modification of a current adhesive. Therefore,
for the main investigations the research focus was aimed at achieving a more basic
understanding of the behaviour of 1C PUR adhesive and the tested wood adhesive composite
in humid environments. For sure wood properties like the wood density or also testing
conditions, such as the fibre load angle, affect the WFP to a certain degree. But it seemed
important to find bonding techniques that are to a high extend independent of such
parameters, since they can hardly be precisely controlled in industrial practice. Furthermore
competitive adhesives such as PRF reliably produce high WFP in the dry and wet states,
independent of variations of such parameters.
10
The main investigations followed four different approaches. Each of them was investigated in
the form of a publication (papers I to IV). Table 1.3 displays an overview of the laboratory
techniques carried out within the main investigations and the additional investigations. For
detailed descriptions of each technique please refer to the relevant chapter noted there.
The main topic of the first approach (paper I) was the investigation of the mechanical
behaviour of the 1C PUR adhesive polymer under varied ambient RH. It was assumed that
high ambient moisture (water or water vapour) significantly reduces the stiffness and the
strength of 1C PUR polymers, which could finally cause low WFP. A verification of this
basic idea could finally lead to improved WFP under high moisture load by varying the
adhesive polymer’s mechanical properties, since the stress-strain behaviour of 1C PURs can
be adjusted within a certain range [25, 26]. The outcomes of the accordant tensile tests under
varied environmental humidity indicate that 1C PUR, as well as MUF, reveal clear reductions
in strength and MOE under moisture load (compared to the dry state). Nonetheless the MUF
still produces more than 70% WFP in the wet state, whereas the WFP of the equivalently
tested PUR batches decline substantially (see 3.2 and 4.1). Hence, the difference in WFP does
not mainly derive from the reduction of cohesive strength or MOE (this was later confirmed
by findings described in paper IV).
The properties of the boundary layer between adherend and adhesives have a confirmed
influence on the bonding quality [27-30]. Therefore it was reasonable to assume that an
optimized mechanical preparation of the bonding surface might improve the interplay
between adhesive and adherend, thus improving the WFP. This led to the second approach.
The resulting question, whether the WFP in the dry and wet states are significantly influenced
by the applied wood machining procedure prior to bonding, is addressed within paper II.
One of the main findings of said paper is that WFP in the wet state cannot be improved by
performing a specific machining technique. Instead the investigation revealed that a loss of
adhesion in the wet state hinders sufficient load transfer between the adherends. Therefore a
third approach was established, based on the idea that WFP in the wet state can be increased
by improving the adhesion between adhesive and adherend by means of DMF priming. This
topic was tackled in paper III.
11
After it was shown that DMF priming substantially improves, TSS and WFP of 1C PUR
bonded wooden joints in the wet state, the search for the accordant mechanisms of action
began. A detailed understanding of the relevant interactions between the wood, the toxic DMF
and the 1C PUR is needed to develop a priming procedure that is less hazardous in terms of
health concerns. In addition, at a later stage of development an additive could probably be
capable of replacing the priming procedure. Since it was known from literature that the
solvent DMF affects the wood [31-33], it was hypothesized that DMF has a much greater
effect on the adherends and their boundary layers (thus improving adhesion) than on the
adhesive polymer itself. This search for the basic modes of actions of DMF was the line of
study for the investigations presented in paper IV.
12
Table 1.3 Laboratory methods, their application and where to find further details
Tested material
Laboratory method Application Details in chapter
Wood
Surface roughness measurement
Tactile measurement of the bonding surfaces’ roughness
4.2
Wood machining Mechanical preparation of bonding
surfaces (planing, sanding, face milling)
4.2
Contact angle measurement Determination of the wooden
surfaces’ wettability 4.2
Wood priming Chemical preparation of the bonding
surfaces (DMF, pMDI) 4.3
UV-VIS-NIR Spectroscopy Investigation of wood extraction by
means of DMF 4.4
Adhesive
Tensile test Testing of adhesive polymer films
after different kinds of climatic treatments
4.1
IR-ATR Spectroscopy Monitoring of the NCO- conversion in
1C PUR droplets 4.4
Sorption tests Measurement of moisture and time
dependent weight changes of adhesive polymer films
5.1
Diffusion tests Measurement of water vapour
diffusion through adhesive polymer films
5.2
FT-IR Microscopy (chemical imaging)
Investigation of the penetration of glue and solvent into the wood
4.4
Composite
Tensile shear test Testing of adhesively bonded wooden specimens after different kinds of pre
-treatments 3., 4.2, 4.3
ESEM Investigation of fracture surfaces 4.2, 4.3
Reflected light microscopy Detection of adhesive polymer on
fracture surfaces 4.3
UV Fluorescence Microscopy Investigation of 1C PUR penetration
into the wood 4.4
AFM Optical representation of the localisation of nano-indents
4.4
Nanoindentation Determination of MOE and Hardness
of bondlines and cell walls 4.4
13
2. State of the art
2.1 Literature overview on wood failure percentage of 1C PUR bonded joints
The 1C PUR wood adhesives are known for creating high WFP (ca. 80%) under dry
conditions and much lower WFP (ca. 20%) in the wet state [34-38]. A detailed investigation
on the causes of the reduced WFP under high moisture conditions is still leaking. Frihart [39]
proposed a shift of swelling strain as influencing factor for WFP. Accordingly, he stated the
swelling strain model, dividing wood adhesives into the two groups of in situ- polymerised
adhesives and pre- polymerised adhesives. Since 1C PURs belong to the second group, the
accordant swelling strain occurs at the interface, thus preferably leading to low WFP.
However, Vick et al. [40] tackled the low WFP via HMR primer. This primer significantly
improves the WFP of 1C PUR and MUF glued joints in the wet state and also helps to reduce
delamination [37, 41]. But it did not gain acceptance on the market, since it’s application is
too impractical for industrial use and it introduces formaldehyde into the production process.
2.2 Relevant technical standards
One of the most important instruments to avoid failing glued joints in structural wooden
elements in practice is the establishment of technical standards, which define the minimum
requirements that such adhesively bonded joints have to meet. These requirements vary quite
a bit from country to country (exemplarily see Table 2.1).
In Europe for instance, Eurocode 5 [42] is a set of rules and regulations concerning design,
performance and technical execution of timber structures. This framework, developed by the
European Committee for Standardisation, was published under the European Standards’
number EN 1995 and refers to several sub-standards: Wooden glued joints for structural
elements in utility class 3 (wet environment, EMC higher than 20%) have to meet the
demands of EN 1995-1-1 [42] (Part 1-1: General – Common rules and rules for buildings).
Amongst others, this refers to EN 14080 [43] which explains that adhesives for full
weathering (adhesive type I) have to be tested according to EN 301 [44] in combination with
EN 302-1 [45] (longitudinal tensile shear strength tests) and other standards (delamination
tests, creep tests, etc.). If the glued joints are produced by means of a 1C PUR, they have to
meet the requirements of EN 15425 [46]. These standards set thresholds for the joint shear
14
strength tested in the dry and wet states, but not for the WFP. In Europe strength, creep and
delamination are regarded as the most important parameters of wooden glued joints providing
structural safety.
In contrast, in North America WFP is regarded as an important criterion when it comes to the
assessment of wooden glued joints. In addition to parameters like strength, creep and
delamination behaviour, standards like ASTM 2559-04 (Specification for Adhesives for
Structural Laminated Wood Products for Use under Exterior Exposure Conditions) [47] and
CSA O112.9-04 (Evaluation of Adhesives for Exterior Structural Wood Products) [48] define
high minimum limits for WFP of joints tested in the dry and in the wet states (see Table 2.1).
The scope of CSA 0112.9-04 defines requirements for evaluating adhesives intended to bond
solid wood for load bearing exterior applications. For hardwood the median threshold for
WFP in the wet state (80%) is even higher than that for the dry test conditions (60%). This is
most likely because the loss of wood strength when wetting above the fibre saturation point is
taken into account. Furthermore, in the annex of this CSA standard (Commentary C.4.7.1) it
is mentioned that these thresholds are derived from inter -laboratory tests, using PRF as a
basis for comparison. It also notes that the standard refers to medians (instead of average
values), because the obtained datasets are frequently not normally distributed.
To some extent the standard ASTM 2559-04 is the US counterpart of the Canadian standard
mentioned above. In addition to delamination and creep, it is also concerned with resistance to
shear by compression loading under dry conditions. The shear strength of the laminated
product has to meet (or exceed) 90% of the shear strength of the accordant solid wood, based
on the average values published in the Wood Handbook of the Forest Product Society.
In addition, this ASTM standard requires a general average WFP of 75% for all wood species
noted in the Wood handbook. Hence this standard follows the traditional concept that, in
addition to high strength values, a high WFP of the composite is a strong indication of
strength that exceeds the strength of the solid wood itself. Even though ASTM D 2559 – 04
refers to laminated wood for structural purposes under exterior conditions, shear tests after
water contact are beyond its scope.
15
Table 2.2 Exemplary overview on technical standards regarding wood failure percentage of
1C PUR bonded wooden joints after shear tests
Technical standard
Abbreviation for treatment of specimens before testing
Description of treatment
Threshold values
Average tensile shear
strength [MPa] b
Median Block shear
strength [MPa]
Wood failure percentage
(WFP)
Eurocode 5: EN 15425 (1C PUR)
A1 conditioning at
20°C / 65% RH a 10
none A4
specimens placed in boiling water for 6h + submerged at room temperature for 2h, test in the
wet state
6
A5 A4 + A1 8
CSA O112.9 Compression
shear test (inter alia)
dry conditioning
conditioning at 20°C / 65% RH a
19 (HW) 10 (SW)
60 (HW) 85 (SW)
(both medians)
vacuum-pressure
submerged whilst vacuum and
pressure treatment, test in the wet state
11 (HW) 5.6 (SW)
80 (HW) 85 (SW)
(both medians)
boil-dry-freeze boiling, hot drying, freezing (8 cycles), test in the wet state
6.9 (HW) 3.5 (SW)
ASTM D 2559 – 04 Compression
shear test (inter alia)
dry conditioning
conditioned at 23°C / 65% RH a
12.4 (White oak)
7.0 (Douglas fir)
75 (average value)
a RH: Relative humidity of ambient air [%], b Adhesive type I and 0.1 mm thickness of adhesive layer, c HW: Hardwood, SW: Softwood
16
2.3 Wood
In the following the main structures of wood are described on the macroscopic, microscopic
and ultrascopic level.
2.3.1 Macroscopic anatomy
When looking at a piece of dry wood without any optical aids, different macroscopic
structures are visible (see Fig. 2.3.1). These different tissues are crucial for the living tree’s
organism since they fulfil important functions like protection, mechanical strengthening,
conduction of water and assimilates, and storage of various substances.
The basic geometry of a tree stem can be roughly approximated by a cylinder, leading to the
definition of three anatomical main directions: The longitudinal (L) axis runs parallel to the
fibre (grain), the radial axis (R) is normal to the growing increments and runs through the
cylinder jacket and pith, whereas the tangential (T) axis runs perpendicular to the grain and
tangential to the growth increments without touching the longitudinal axis (see Figure 2.3.1).
Hence, a radial section, for instance, forms a LR-plane or a tangential section leads to a LT-
plane.
Fig 2.3.1 Macroscopic view of a tree’s stem section. Schematic drawing [49]
17
Related to figure 2.3.1 the relevant macroscopic structures of a stem are discussed from the
outside to the inside. During growing season, the vascular cambium (layer between bark and
wood) produces the wood (secondary xylem) and the inner bark (secondary phloem). The
tissue of the inner bark provides the transport of assimilates (like sugars) from the leaves to
the roots and to the growing parts of the organism [50]. The outer bark helps to protect the
tree organism from mechanical damage or excessive evaporation. The parts of the wood,
which are located next to the cambium, are called sapwood. The sapwood comprises the
living part of the wood. It conducts the sap (water) from the roots to the leaves and stores
photosynthates (such as starch and lipids) and synthesizes other biochemicals. After a certain
amount of growth periods, most species transform the oldest parts of their sapwood into
heartwood. The heartwood however, located between pith and sapwood, is no longer involved
in water transport. As a result of this (species dependent) transformation, the heartwood
finally exhibits a different chemical composition and a reduced permeability (compared to the
accordant sapwood). This leads to a lower WMC and EMC as well as to darker colour and / or
enhanced durability. The wood grown during spring is called earlywood, which exhibit big
lumen diameters and thin cell walls. The wood grown during summer and autumn is called
latewood, which exhibit smaller lumina and thicker cell walls (see Figure 2.3.2.1, EW and
LW). This sequence of growth characteristics leads to the annual growth ring patterns of
wood grown in temperate zones. In the centre of the tree the stem axis is represented by the
pith, a remnant of the very early (primary) growth period of the plant before wood formation
(secondary growth).
2.3.2 Microscopic anatomy
As an interim result of evolution, two main classes of trees can be differentiated nowadays:
The coniferous softwood trees (gymnosperms) and the deciduous hardwood trees
(angiosperms). As described in the following, their microscopic structures and cellular
compositions differ considerably from each other (Table 2.3.2).
18
Table 2.3.2 Major cell types found in softwoods and hardwoods and their main functions [51]
Function Softwood Hardwood
Mechanical strengthening Tracheids (Latewood) Libriform fibres (thick walled) Fibre tracheids (thin walled)
Transport of water Tracheids (Earlywood) Tracheae (vessels) Vessel tracheids
Storage and transport of assimilates
Ray parenchyma Longitudinal parenchyma
Secretion Epithelial cells (surrounding the resin channels)
For the conduction and distribution of aqueous solutions between the cells, several openings
between the cell walls are needed. Such openings are called pits. The three major types of pits
are simple pits, bordered pits and half-bordered pits [51]. Their basic structure is displayed in
figure 2.3.4. While simple pits reveal a straight-walled pit chamber, bordered pits reveal
secondary walls overarching the pit chamber on each of the two cell walls, thus enlarging the
pit membrane in the middle. Bordered pits typically occur between two conducting cells,
whereas simple pits appear between parenchyma cells. Also half-bordered pits are of
importance, because they provide transport processes between conducting cells and
biochemically active parenchyma cells.
19
Figure 2.3.2 The three varieties of pit formation: a) schematic drawings of simple pits,
bordered pits with Torus (Pinaceae) and half- bordered pits; and micrographs of bordered pits
in spruce (RT- plane) (b), in tracheid cell wall (c) and in a cross- section (d). The scales
correspond to 10 m [89, adapted after 44]
20
2.3.2.1 Softwood
The softwood trees date from a much earlier evolutionary period. Therefore, their wood
tissues present less different kinds of cell types (see table 2.3.2), which fulfil several functions
on a comparatively low performance level. The arrangement of these cell types leads to the
anatomical softwood structure displayed in figure 2.3.2.1. These comparatively long and slim
tracheids fulfil mechanical and conducting functions. The storing function is covered by ray
parenchyma and longitudinal parenchyma, supplemented by epithelial cells (around resin
channels) for secreting functions.
Fig. 2.3.2.1 Cellular structure of spruce softwood [49]. Image generated by means of
synchrotron based tomographic microscopy
2.3.2.2 Hardwood
Hardwood trees present more different cell types, fulfilling more specialized functions (see
table 2.3.2) on a higher performance level.
The various kinds of vessel arrangements found in different hardwood species lead to
different hardwood groups. In particular, diffuse-porous hardwoods (e.g. Acer, Betula or
Fagus spp.), ring-porous hardwoods (e.g. Quercus or Fraxinus spp.), and semi-ring-porous
hardwoods (e.g. Juglans or Prunus spp.) are distinguished by arrangement and size of their
vessels. The basic microscopic structure of European beech (Fagus sylvatica L.) hardwood is
21
exemplarily depicted in Figure 2.3.2.2.
Fig. 2.3.2.2 Cellular (microscopic) structure of European beech hardwood [49]. Image
generated by means of synchrotron based tomographic microscopy
2.3.3 Ultrascopic anatomy
The microscopic structures mentioned above are basically represented by the cell walls of the
different kinds of tissues. Therefore this section exemplarily describes the composition of a
wood fibre cell wall.
The function of the devital wood cells is solely borne by the cell wall structures and the cell
lumina [42]. Whilst the lumen just encloses a void space, the cell wall assembly (e.g. of a
tracheid) consists of three main layers: The middle lamella (M), the primary wall (P) and the
secondary walls (S1 + S2 + S3). This basic structure of the cell wall composition of a wood
fibre (strengthening tissue) is schematically depicted in figure 2.3.3.
22
Figure 2.3.3 Model of the cell wall layers of a wooden fibre after [52], modified by [49].
L: Longitudinal direction, M: middle lamella, P: Primary wall, S1-S3: Secondary walls. Dark
lines represent the texture of the cellulose microfibrils in each cell wall layer
Regardless of the species, the density of the pure cell wall material is about 1500 kg/m3 [53].
Each of the cell wall layers consists of the basic components cellulose and the matrix
materials hemicellulose, pectin (mainly in primary walls) and lignin [54]. From a more
general perspective, cellulose (highest content in S2 layer) represents string-like molecules
with high tensile strength and stiffness (MOE approx. 134 GPa in the axial direction) [55].
Cellulose microfibrils are strictly aligned agglomerates of such molecules, forming very
strong macromolecules. They are embedded in hemicellulose and lignin (MOE about 2.0 GPa,
highest concentration found in M) matrix polymers [51], holding the cellulose fibrils in place
when under compressive load. This combination of fibrils and embedding matrix provides the
cell wall with the needed compressive strength. Compared to cellulose, the hemicelluloses are
shorter, but more branched molecules with a MOE of about 0.02 – 2.0 GPa [55]. It is likely
that hemicelluloses mediate recovery mechanisms in the cell walls [56]. Thus, from this
perspective wood can be considered as a kind of fibre reinforced bio composite material [57].
Not only do the strength and the stiffness of the semicrystalline microfibrils have a major
influence on the mechanical properties of the cell wall, but also their orientation related to the
23
fibres longitudinal axis [58]. In each cell wall layer the microfibrils follow a specific texture
(see figure 2.3.3). Whilst the multidirectional and almost transverse orientation of cellulose
fibrils in primary walls stabilizes the cell in the circumferential direction during longitudinal
growth [59], the three layers of the secondary walls (formed after completion of the cell
growth) each reveal a specific inclination of the aligned fibrils towards the cell axis [58]. This
varying microfibril angle (MFA) makes the whole wood fibre stiff, gives it tensile strength
and contributes to the anisotropic behaviour of wood. The MFA between the layer varies from
0° - 50° in S2 to 60° - 90° in S1 and S3 [59-61]. The MFA in the S2 layer particularly affects
the mechanical properties of the cell wall due to the comparatively high thickness of this
layer. An increase in MFA decreases the cell wall stiffness and increases the strain to fracture.
Due to the hierarchical organization of the plant, this interaction between the MFA of stiff
cellulose and the pliant matrix polymers is crucial. It does not only affect the mechanical
behaviour of the single fibre, but also of the whole tree and helps the tree to specifically adapt
to environmental influences (such as wind) [58].
2.3.4 Wood species used in this thesis
Two different wood species were used to perform the investigations described in this thesis.
The main species was European beech hardwood. For complementary investigations Douglas
fir softwood was used. This section gives a concise overview on the composition and the
properties of these species.
2.3.4.1 Douglas fir
Douglas fir (Pseudotsuga menziesii Mirb. Franco) was used for comparative tests (see 4.3)
since this softwood species is frequently used for timber structures in NA. Compared to
European beech wood, it represents quite different histological, chemical, physical and
mechanical characteristics (see table 2.3.4.1 compared to table 2.3.4.2).
24
Table 2.3.4.1 Composition and the properties of Douglas fir wood [62]
Douglas fir wood (Pseudotsuga menziesii Mirb. Franco)
Tracheids Length [mm]
Lumen diameter [m]
Thickness per two cell walls [m] Proportion [%]
Wood rays
Proportion [%]
Resin channels Proportion [%]
Extractive contents [%]
Benzene- alcohol extraction Cold water extraction Hot water extraction
Structural substances’ contents
Lignin content [%] Total sugar content [%] Cellulose content [%]
Ash content [%]
Tannin content [%] pH
Density (g/cm3) at 12% EMC
Differential volumetric shrinkage [%]
Compressive strength (ǁ to the grain) [N/mm2] a Bending strength [N/mm2] a
Tensile strength (ǁ to the grain) [N/mm2] a Tensile strength (┴ to the grain) [N/mm2] a
Shear strength (ǁ to the grain) [N/mm2] a E- modulus (ǁ to the grain) [N/mm2] a
2.5…4.5...5.6
20.0…30.0…35.0 4.0…8.0…16.0
93%
7.0
0.2
1.6 …4.4 2.2…3.5 2.8...6.5
25.4…34.8 67.0…71.4 43.7…55.6
0.6…1.1
6…10 3.3…4.2 (Heartwood)
0.35…0.51…0.75
0.38…0.42
43…68 68…89
…105… …2.4…
7.8…10.2 11200…13500
a: measured at about 12% EMC
25
2.3.4.2 European beech wood
The majority of the tensile shear tests presented in this thesis was carried out using the
diffuse-porous hardwood species European beech (Fagus sylvatica L.). A concise overview
on the properties of this wood species is given in table 2.3.4.2. This temperate zone species
features ladder shaped interconnections between the vessel elements (scalariform
perforations) as well as circular shaped ones [51, 62]. A remarkable property of European
beech wood is the presence of comparatively thick wood rays, visible by the naked eye. They
are composed of parenchyma tissue of single cell width and multicellular width (for average
percentages please refer to table 2.3.4.2). In addition to storage functions, such ray
parenchyma cells also provide a substantial contribution to the radial strength and elasticity of
the stem [63].
This species was chosen for the tests, since it does not contain high percentages of extractives
(see table 2.3.4.2), which might affect the bonding quality. Furthermore, European beech
wood is known for its comparatively high differential swelling and shrinking. Hence effects
caused by stress due to dimensional changes of the specimens after different pre-treatments
might become more pronounced and therefore better detectable. This hardwood also reveals a
comparatively high strength (compared to most softwood species). Therefore a new technical
approach, revealing an improvement in WFP on beech wood, will most likely also be helpful
when applied to many other species with the same or lower strength. In addition, beech wood
is proposed in EN 302-1 for tensile shear tests, which is actually the prevalent bonding test
method within this thesis. Presently, structural wooden elements are primarily produced from
softwoods, but the producers of these elements are progressively starting to explore the
possibilities of using hardwoods, since their higher strength and stiffness allow for larger
bearing spans and smaller diameters. In addition, architects are progressively coming to
appreciate the more homogeneous look of hardwoods (compared to softwoods). Therefore the
utilization of hardwoods for structural bondings of wood is a current topic of interest.
26
Table 2.3.4.2 Composition and the properties of European beech wood [62]
European beech wood (Fagus sylvatica L.)
Libriform fibres Length [mm]
Lumen diameter [m]
Thickness per two cell walls [m] Proportion [%]
Vessels
Lumen diameter [m] Proportion of pores (vessels) [%]
Average percentages of parenchyma cells
Longitudinal parenchyma [%] Ray parenchyma [%]
Density (g/cm3) at 12% EMC
Density (g/cm3) green
Extractive contents [%] Benzene - alcohol extraction
Water extraction
Structural substances’ contents Lignin content [%]
Total sugar content [%] Cellulose content [%]
Ash content [%]
pH
Density (g/cm3) at 12% EMC Density (g/cm3) green
Differential volumetric shrinkage [%]
Compressive strength (ǁ to the grain) [N/mm2] a Bending strength [N/mm2] a
Tensile strength (ǁ to the grain) [N/mm2] a Tensile strength (┴ to the grain) [N/mm2] a
Shear strength ( ǁ to the grain) [N/mm2] a E- modulus (ǁ to the grain) [N/mm2] a
0.6…1.3
3.5…7.1…11.2 3.6…7.5…10.3
25.2…39.6…57.2
8…45…85 24.6…39.5…52.5
3.5…5.2…7.0 11.2…15.7…21.2
0.54…0.72…0.91 0.82…1.07…1.27
1.5 … 1.9 1.9
11.6…22.7 75.7…85.0 33.7…46.4
0.3…1.2 5.1…5.4
0.54…0.72…0.91 0.82…1.07…1.27
0.46…0.6
41…62…99 74…123…210 57…135…180
7.0…10.7 6.5…8.0…19.0
10000…16000…18000
a: measured at about 12% EMC
27
2.3.5 Wood as an adherend
After application of the adhesive onto the bonding surface, the composite material is then
produced during the pressing procedure. This leads to penetration of adhesive into the porous
adherend, whereby an interphase region is formed between the adhesive layer and the two
adherends. According to the model described by Habenicht [64] the interphase region is
bounded by the interface on one side and the pure adherend on the other (Figure 2.3.1). The
interface however, is the boundary layer between pure adhesive and the adherend. Figure
2.3.2 displays the situation on bonded spruce wood (left) and beech wood (right).
Figure 2.3.1 Model concept of a joined bondline after Habenicht [64]
Figure 2.3.2 SEM images of polyurethane prepolymer bondlines in spruce (left) and beech
wood (right). The scales represent a length of 200 μm [65]
Inside the single material, cohesive electric forces are quasi equally distributed between the
molecules, thus providing the material’s strength [28]. However, at the material surface, the
molecules do not have “equal neighbours” to interact with via dipole forces. Therefore they
interact with other surrounding substances (such as dust particles or adhesives), binding to
28
their surface by means of dipole forces. Hence, in addition to cohesive forces, the adhesive
forces at the interface are also crucial for sufficient load transmission between the adherends
of an adhesively bonded joint [28]. Secondary bonds especially, like hydrogen bonds,
contribute to adhesive forces and play an important role when it comes to the strength testing
of 1C PUR bonded wooden specimens (see 4.1, 4.2, 4.3).
2.4 One-component polyurethane adhesives for the structural bonding of wood
In the late 1990s, established adhesive systems for glulam production, like MUF or PRF, were
supplemented by the rapidly emerging 1C PURs. The latter offers several advantages to the
glulam producers as well as to users of the finished constructions. The main advantage is the
absence of additional formaldehyde in the gluing process and in the glued product,
particularly since the WHO declared formaldehyde carcinogenic to humans [20]. However, it
also has advantages like ready to use delivery and the absence of contaminated wastewater in
its use, which are remarkable. Therefore within this thesis the focus lies on the use of 1C
PURs for the production of adhesively bonded structural elements.
2.4.1 Basic chemistry of 1C PUR wood adhesives
Otto Bayer and his research co-workers discovered polyurethane in 1937 in Leverkusen,
Germany. In the 1950s, the production of a large variety of PUR products started on an
industrial scale. Starting with foams and sealers, adhesives were also later made from
polyurethanes. The basic structure of moisture curing 1C PUR wood adhesives is represented
by prepolymers containing soft segments (SS) and hard segments (HS). The SS derive from a
polyaddition reaction between the hydroxyl groups of (more or less branched) polyols and a
stoichiometric surplus of aromatic methylene diphenyl diisocyanate (MDI) isomers (see
Figure 2.4.1 left), forming urethane groups (see Figure 2.4.1 right).
29
Figure 2.4.1 4, 4’ –MDI (left) and urethane group (right)
In 1C PUR wood adhesives, polyether polyols are the most common type of polyols because
of their availability, low costs and chemical performance [66]. Different kinds of polyether
polyols are available with various molecular weights and functionalities. Their properties
derive from the different types of precursors used for their production, such as initiators
(starting alcohols) and different types of epoxides [66]. For the production of 1C PURs the
preferred epoxides are ethylene oxide (EO) and propylene oxide (PO). By varying the
proportion and functionality of polyether polyols and MDI isomers, the branching within the
fluid prepolymer, and thus also the network density of the cured polyurethane, can be adjusted
within a certain range [67]. Furthermore, the prepolymer is complemented by feeding it
additives like catalysts, fillers, defoamers, etc. This results in a ready-to-use 1C PUR wood
adhesive, tailored according to the customer’s needs. The adhesive’s curing starts by applying
the 1C PUR onto the adherend and is driven by the reaction between excess MDI isomers and
moisture. The latter derives from the adherend and from the ambient humidity. As an
intermediate step to this reaction, carbamic acid is built, which then decomposes to a primary
amine and carbon dioxide gas. Consequently, the amine reacts with the residual isocyanate,
forming urea linkages, taking effect as strong crosslinking points in the final polymer
network. The areas of the polymer, which present polyurea-polyurethane groups, are referred
to as HS of the cured polyurethane polymer [25, 66]. The mechanical characteristics of the
cured polymer, like strength, stiffness or plastic workability, derive from the characteristics of
the SS, the HS and the interaction between those two by primary (covalent) bonds at the
cross-linking points and additional secondary (hydrogen) bonds [25, 68]. The strong hydrogen
bonds (H bonds) between the amino and carbonyl groups of urea and urethane considerably
support the cohesion of the polymer matrix, whereas the H bonds between HS and polyether
are much weaker.
30
2.4.2 Benefits and drawbacks of 1C PUR for structural bonding of wood
Compared to the more traditional adhesives used for structural bonding of wood (MUF, PRF),
1C PUR offers several advantages to the user, but it has also revealed some drawbacks.
Benefits:
No additional formaldehyde emission in the production facility (occupational safety)
No additional formaldehyde emission from the bonded products (consumer safety)
After curing the adhesive polymer does not emit NCO or VOC
Ready to use delivery (No mishandling of adhesive or raw materials regarding mixing
procedure, storage management, etc.)
Water free handling. No need for expensive wastewater treatment (e.g. resulting from
maintenance and cleaning of the gluing line)
Lower adhesive application quantities (g/m2)
No dark glued joints (the colour is even adjustable to a certain degree)
Capability of bonding wood with comparatively high moisture content close to fibre
saturation [69-71]
Adjustable curing times (between several minutes and several hours)
Long storage time of delivered adhesive (min. 1 year)
Curing at room temperature (no heat required)
Uniform and consistent curing process of the adhesive over the whole crosscut of the
pressing, even for adherends with large dimensions, such as very wide multi-stack
beams (no moisture needs to be removed for setting)
Drawbacks:
Creep under moisture load, if the composition of the polyurethane adhesive is not
optimized [69, 72]
Sensitivity to high temperatures, if the chemical composition of the adhesive system is
not optimized for thermal loads [25, 73-77]
Comparatively high procurement price per ton of adhesive
Low resistance against delamination (e.g. compared to certain MUF) [78]
31
Reduced strength and MOE under moisture load [26, 79]
Sensitivity against WMC below 8% [80-82] during gluing procedure
Sensitivity against WMC above fibre saturation during gluing procedure [83]
Low WFP in the wet state hinders 1C PUR bonded joints from meeting the demands
of CSA O112.9-04 (see 4.2, 4.3)
Clad [49] was among the first concerned with the mechanical properties of wood adhesives in
general. In the recent past various surveys have investigated the influence of temperature on
the mechanical properties of 1C PUR adhesives or the various wooden glued [21, 77, 79]
joints and noticeable progress has been achieved in this field [21]. Studies regarding the
influence of RH on the mechanics of 1C PUR wood adhesives for the structural bonding of
wood are comparatively rare [79, 84] and such bondings still reveal difficulties in respect to
generating high WFP in the wet state [69].
32
3. Preliminary investigations
3.1 Round robin test
To a considerable extent, the laboratory work for this thesis consisted of testing adhesively
bonded wooden specimens after different kinds of pre-treatments. Besides the quality of the
selected wood and the used adhesive, the manufacturing technique for the production of the
specimens and the proper execution of the pre-treatments (water storage, re-drying, etc.) also
influence the test results. Therefore a round robin test was carried out prior to any specific
testing to assure state of the art laboratory procedures and devices. This round robin test
comprised the laboratory of the industrial partner Purbond AG and the wood physics
laboratory at the ETH Zürich. Special attention was paid to the selection of the wood, the
handling and the application of the adhesives and the production of adherends, pressings and
resulting test specimens. One of the main outcomes of this test was a much more even
distribution of the pressing forces in the calibrated press by paying more attention to the
machining procedures, supplemented by using a pressing jig (see Figs. 3.1.1 & 3.1.2). In
addition more attention was paid to the testing parameters. Finally, more clarity regarding the
most crucial processing factors and testing parameters was achieved (wood selection, planing
knife quality, planing feed rate, thickness tolerances, pressing jig with 3mm neoprene mats,
testing speed) together with reliable comparability of the results deriving from the two
laboratories.
Figure 3.1.1 Starting point of round robin test: left: pressings lying “loosely” in the press;
right: pressing forces unevenly distributed on the joining surfaces (pressure 1 MPa, visualized
(red colour) using pressure sensitive FUJI Prescale films
33
Figure 3.1.2 Result of various improvement steps at the end of the round robin test. Left:
pressing jig with neoprene mats between the pressings; right: Prescale films reveal more
evenly distributed pressing forces (below: 0.7 MPa, middle: 1 MPa, above: 1.5 MPa)
3.2 Tests on wood failure percentage in the wet state
Subsequent to the round robin test, a preliminary tensile shear test according to prEN 302-1
[45] was carried out, in order to reproduce and confirm the low WFP of 1C PUR joints in the
wet state described in the literature [34, 70, 85, 86].
Material and methods
European beech wood (Fagus sylvatica L.) was conditioned at 20°C / 65% RH until EMC of
about 12.4% (average) was reached. Consecutively the wood was processed according to the
standard mentioned above. The slats used for adherends were mixed in order to scatter
influences resulting from the wood over the whole sampling. A 1C PUR (HBS 709, Purbond),
a PRF (Aerodux 185, Dynea) and a MUF (Kauramin, BASF) were used as adhesives and
processed in accordance with the producers’ recommendations. All the adhesives are
approved for the structural bonding of wood in Europe. MUF and PRF served as reference
adhesives. The two components of the PRF (Resin 185 and hardener HRP 155) and of the
MUF were mixed as depicted in table 3.2.
34
Table 3.2 Adhesive processing parameters
After production, the pressings were stored again in climate 20°C / 65% RH for 5 days to
ensure sufficient curing of the polymers. Hereinafter they were cut to tensile shear specimens
under the terms of EN 302-1 (Fig. 3.2.1) and underwent the pre-treatments A1 (conditioning
in climate 20°C / 65% RH until EMC before testing) and A2 (4d water storage at 20°C,
consecutive testing in the wet state) in accordance with said standard. In addition to the
adhesively bonded specimens also solid wood specimens were produced and pre-treated.
Figure 3.2.1 Specimen according to EN 302-1 [45]
Mixing ratio
(resin : hardener)
Applied quantity per joint [g/m2]
Pressing time [h] Pressure [Mpa]
1C PUR - 180 3 1.2
MUF 100 : 60 400 4.5 1.2
PRF 100 : 20 450 4 1.2
35
Results
As Fig. 3.2.2 displays, all the batches pass the minimum requirements regarding TSS after
treatment A1 (10 N/mm2) and after treatment A2 (6 N/mm2) based on EN 15425 [46] and EN
301 [44], respectively. They reach the TSS of the solid wood itself. In this respect it should be
noted, that the testing results deriving from the solid wood samples are not exactly
comparable to the results deriving from the bonded specimens (see 4.3). For none of the
adhesives a significant difference was detected between the two moisture- treated batches
(95% and water). It might be reasonable to take this into consideration, since structural
elements situated in coastal areas or in tropical regions are exposed to such climate conditions
for the most part of the year. All adhesives revealed a loss in TSS from dry (65%) to wet
(95%, water) stage, which lies in the range of the accordant loss in TSS of the solid wood
itself. After such treatments the 1C PUR samples revealed the lowest WFP by far, compared
to the accordant MUF and PRF lots. Hence, the outcomes of this preliminary tensile-shear test
confirm the findings described in the literature.
Figure 3.2.2 Tensile shear strength (TSS) and wood failure percentage (WFP) tested on
European beech wood (Fagus sylvatica L.) according to EN 302-1. Samples were conditioned
at 65% RH (A1) or 95% RH at 20°C or stored in water at 20°C for 4 days (A2) before testing.
0
10
20
30
40
50
60
70
80
90
100
0
2
4
6
8
10
12
14
16
65%
95%
wat
er
redr
ied
65%
95%
wat
er
redr
ied
65%
95%
wat
er
redr
ied
65%
95%
wat
er
redr
ied
1C PUR MUF PRF Solid wood
TSS [MPa] WFP [%]
36
1C PUR: HB S 709 (Purbond); MUF: Kauramin (resin 683, hardener 688, BASF); PRF:
Aerodux (resin 185, hardener HRP 155, Dynea)
3.3 Tests with modified 1C PUR
Developing a new adhesive and introducing it to the market requires several efforts like R&D
resources (personnel, laboratory, materials, time), certification procedures in several
countries, and additional marketing efforts. Therefore a primary wish of the industrial partner
was to limit any changes of the current commercial products to a minimum. Therefore in a
first step the focus was set on investigating parameter changes in the adhesive system, which
can be implemented comparatively easy. Hence, four variations of HB S 309 (Purbond), a
commercial 1C PUR for the structural bonding of wood, were tested (Table 3.3). The filler
content, the molecular weight of pMDI and the hydrophilicity of the pMDI were varied. Their
influence on TSS and WFP under wet conditions was tested by means of tensile shear strength
on the basis of EN 302-1 using European beech wood (Fagus sylvatica L.). The laboratory
procedures were carried as described under to 3.2.
Table 3.3 Tested variations of HB S 309
Description of variation
VN 3105 increased filler content: 5% chalk
VN 3106 increased filler content: 10% chalk
VN 3107 30% of the isocyanate content replaced by a pMDI with lower molecular weight
VN 3108 VN 3107 with more hydrophilic pMDI
In summary, none of the modified 1C PURs opened a pathway to enhanced WFP under wet
conditions (see 3.3.1 to 3.3.3).
37
3.3.1 1C PUR with modified filler content
As Figure 3.3 depicts, the filler content heavily affects the TSS and WFP after treatment A1
(dry state). This goes in line with Clauss [21], who tested the influenced of various fillers on
the performance of 1C PUR bonded wooden joints in the dry state and under elevated
temperatures. However, within the current investigation no significant influence of the filler
content on TSS in the wet state was detected. An increased content on chalk even reduced the
WFP after A4 treatment (VN 3105 and VN 3106 compared to HB S 309).
Figure 3.3 Tensile shear strength and wood failure percentage (▲) tested acc. to EN 302-1 on
1C PUR (modified HB S 309) and Fagus sylvatica L.; Sample treatments before testing: A1:
storage in 20°C / 65% RH; A4: 6h boiling water + 2h water at 20°C. Error bars: Confidence
intervals ( = 0.05). VN 3105: 5% chalk content. VN 3106: 10% chalk content. VN 3107:
30% of the isocyanate replaced by lower molecular weight pMDI. VN 3108: 30% of the
isocyanate replaced by 30% more hydrophilic pMDI
0
20
40
60
80
100
0
2
4
6
8
10
12
14
16
A1 A4 A1 A4 A1 A4 A1 A4 A1 A4 A1 A4
Solid wood HB S 309 VN 3105 VN 3106 VN 3107 VN 3108
Woo
d fa
ilur
e [%
]
Ave
rage
tens
ile
shea
r st
reng
th [
MP
a]
38
3.3.2 1C PUR with modified molecular weight
A general thought that is often discussed when it comes to bonding of wood, is the idea of
enhanced penetration of single components of the adhesive polymer into the wood cell wall in
the interphase (see Fig. 2.3.1), supported by low molecular weight (see 4.3). Such a
penetration into the cell wall could improve the adhesion and therefore enhance TSS and
WFP in the wet state. This approach was not confirmed by the test results (VN 3107).
Compared to HB S 309 this variation neither significantly improved the TSS nor did it
enhance the WFP (A1 and A4 batches).
3.3.3 1C PUR with more hydrophilic pMDI
The concept described under 3.3.2 is supplemented by the idea to introduce greater polarity
into the low molecular pMDI. This could basically support the reaction between the 1C PUR
adhesive and the (polar) moisture bound in the cell wall. The findings depicted in Fig. 3.3.
reveal, that this variation in polarity enhances TSS in the dry state (A1), but neither improves
WFP in the dry state (VN 3108 compared to VN 3107 and HB S 309) nor creates higher
WFP.
3.4 Tests with modified polyurethane prepolymers
By 2007 the author had already performed tensile shear tests according to prEN 302-1 [45] on
12 specifically modified 1C PUR prepolymers (Prepos) [82] in the dry and wet states. The
varied parameters were content of NCO, content of PO, content of EO, network density,
content of urethane groups and content of urea groups. As a result of this diploma thesis, no
clear evidence for a significant influence of the Prepo composition on TSS or WFP under wet
conditions was found. None of the 12 Prepos reached the TSS –threshold of 6 MPa in the wet
state and none created more than 5% WFP (all average values between 0% and 5%). Hence
the author concluded, that for an improvement of the WFP of 1C PUR bonded wood under
moisture load, approaches other than variation of the Prepo composition are needed (see 1.3).
39
4. Main investigations
List of peer-reviewed papers:
Paper I
Influence of moisture on stress-strain behaviour of adhesives used for structural bonding of
wood
Oliver Kläusler, Sebastian Clauß, Luise Lübke, Jürg Trachsel, Peter Niemz
International Journal of Adhesion & Adhesives (2013) 44: 57 - 65
Paper II
Influence of wood machining on tensile shear strength and wood failure percentage of one-
component polyurethane bonded wooden joints after wetting
Oliver Kläusler, Klaus Rehm, Falko Elstermann, Peter Niemz
International Wood Products Journal (2014) 5, 1: 18 – 26
Paper III
Improvement of tensile shear strength and wood failure percentage of 1C PUR bonded
wooden joints in the wet state by means of DMF priming
Oliver Kläusler, Philipp Hass, Carlos Amen, Sven Schlegel, Peter Niemz
European Journal of Wood and Wood Products (2014) 72: 343–354,
Paper IV
Influence of dimethylformamide on one-component moisture-curing polyurethane wood
adhesives
Oliver Kläusler, Wilhelm Bergmeier, Alexander Karbach, Walter Meckel, Eduard Mayer,
Sebastian Clauß, Peter Niemz
International Journal of Adhesion & Adhesives (2014) 55: 69-76
40
4.1 Paper I
International Journal of Adhesion & Adhesives (2013), 44: 57 - 65
Influence of moisture on stress-strain behaviour of adhesives used for
structural bonding of wood
Oliver Kläusler a, Sebastian Clauß a, Luise Lübke b, Jürg Trachsel a, Peter Niemz a
a Institute for Building Materials, ETH Zurich, Schafmattstrasse 6, 8093 Zurich, Switzerland
b Eberswalde University of Applied Sciences, Alfred-Möller-Straße 1, 16225 Eberswalde,
Germany
a corresponding author’s phone: +41 (44) 632 32 32, fax: +41 (44) 632 11 74,
e-mail: [email protected], url: www.ifb.ethz.ch/wood
Abstract Tensile strength, Young’s modulus and stress-strain behaviour of adhesive films
were investigated by means of tensile tests at 23°C under various ambient moisture conditions
(5% to 95% RH, water exposure and redrying). The adhesive films were produced from two
one-component moisture-curing polyurethane adhesives (1C PUR), one phenol-resorcinol-
formaldehyde resin (PRF) and one melamine-urea-formaldehyde resin (MUF). These four
adhesives are commonly used for structural bonding of wood. In addition, films were made
from three non-commercial 1C PUR prepolymers, all of which had their ethylene oxide (EO)
proportions specifically modified. For all the tested adhesives other than PRF, the findings of
the tensile tests revealed a linear dependency of tensile strength and Young’s modulus on the
relative humidity (RH). Both parameters decreased significantly with increasing RH. The
redried samples illustrate the reversibility of this effect. These observations are mainly
attributed to physical bonds like hydrogen bonds, which are disrupted by water molecules
entering the polymer film and re-established whilst re-drying. No evidence was found for an
influence of the EO content of the prepolymers on their tensile strength or Young’s modulus
at high RH. Regarding the 1C PURs at high RH the findings revealed an influence of
hydrophilic catalyst on tensile strength, but not on Young’s modulus. Under all tested ambient
41
conditions, the fracture strain of PRF and MUF specimens remained below 5%, whereas that
of the 1C PURs and the prepolymers reached at least 20%. This illustrates the ductility of the
tested 1C PUR polymers on all tested climate stages in contrast to the brittleness of MUF and
PRF polymers.
Keywords: Wood Adhesives; polyurethane; phenolic resin; melamine resin, tensile test,
Young’s modulus, tensile strength, moisture
1. Introduction
Over their whole life span, adhesively bonded exterior load bearing timber constructions are
exposed to a range of ambient moisture conditions. Changing moisture influences the
mechanical properties of the wood-adhesive composite material, because it affects the wood
[1], the adhesive layer [2] and their interplay [3]. This study focuses on the effect of increased
ambient moisture on tensile strength, Young’s modulus of elasticity and ductility of the
adhesive layer. Thin adhesive films were made from two moisture-curing one-component
polyurethanes (1C PUR), one phenol-resorcinol-formaldehyde resin (PRF), and one
melamine-urea-formaldehyde resin (MUF). In Europe these commercial cold curing
adhesives are approved and used for the production of structural wood-based composites,
such as glued-laminated timber (glulam) [1, 2].
Generally, urea-formaldehyde resins (UF) are the most common amino plastic adhesives used
for engineered wood products like particleboards [2], however, access of water causes
hydrolysis, subsequent formaldehyde emission and a loss of strength in UF adhesive joints [4-
6]. Therefore, fortified resins were developed by adding proportions of melamine to the UF
adhesive in a co-condensation process, leading to MUF. Herewith, resistance of these
thermosetting, strong, rigid and brittle polycondensation resins against hydrolysis has been
significantly increased [2, 5, 7, 8]. This has allowed for the wide use of MUFs for the
structural bonding of wood [9]. In such amino resins, which require a suitable acid hardener,
the formaldehyde reacts with urea and melamine monomers forming quasi-linear polymer
chains. The substantial crosslinking between the chains is achieved by the excess of
formaldehyde [9, 10]. Therefore the behaviour of MUF adhesives under wet conditions is
heavily influenced by the formaldehyde/urea mole ratio and the melamine content [4].
42
Whereas MUF is generally considered suitable for bonds requiring limited resistance to
weather and water exposure, structural adhesive joints with outstanding resistance against full
weathering are rather made from PRF resins [9]. These are not subjected to substantial
hydrolysis [2] and are known for their high stiffness, brittleness and strength [2, 7]. In
principal, such cold setting PRF adhesives are produced by a two- step procedure. In the first
step, a quite linear low-condensation phenol-formaldehyde (PF) resol polymer with reactive
methylol groups is generated by the reaction of phenol with an excess of formaldehyde under
alkaline conditions. In a second step, a surplus of highly reactive resorcinol is added to the PF
resol under specific conditions, leading to a reaction between the resorcinol and the methylol
groups of the PF. The result is a dark, brownish thermosetting PRF resin with resorcinol
“grafted” onto the PF polymer. The curing is driven by the cross-linking polycondensation
reaction, which starts between the surplus resorcinol and the finally added formaldehyde
hardener [9].
In addition to the basic resins and hardeners, resins like MUF or PRF present additives such
as fillers (strengtheners or extenders) or stabilizers, which often amount 50% of the final
polymer. These also strongly influence polymer properties like strength, water uptake or
swelling and shrinking [11].
In the early 1990s, 1C PURs were introduced to the market of structural wood bonding since
formaldehyde-free bonding became more and more distinct [2]. However, in 2006 the
WHO/IARC evaluated formaldehyde as carcinogenic to humans [12]. Besides the absence of
formaldehyde, 1C PURs are characterized by shorter curing times, comparatively small
application quantities [13] ready-to-use delivery and water-free handling. Prepolymers
represent the basic structure of moisture curing 1C PUR wood adhesives, containing soft
segments (SS) and hard segments (HS). The SS derive from the polyaddition reaction
between the hydroxyl groups of more or less branched polyols and a stoichiometric surplus of
methylene diphenyl diisocyanate (MDI) isomers, forming urethane groups. In 1C PUR wood
adhesives, polyether polyols are the most common type of polyols because of their
availability, low costs and chemical performance [9]. Different kinds of polyether polyols are
available with various molecular weights, functionalities and different hydrophilic properties.
Their properties derive from the different types of precursors used for their production, such
as initiators (starting alcohols) and different types and amounts of epoxides [9]. For this
purpose, the preferred epoxides used are ethylene oxide (EO) and propylene oxide (PO). By
43
varying the proportion and functionality of the various polyether polyols and MDI isomers,
branching within the fluid prepolymer, and thus also the network density of the finally cured
polyurethane polymer, can be adjusted within a certain range [11]. According to customers’
needs, the prepolymer is subsequently tailored by feeding additives like catalysts, fillers,
defoamers, etc., resulting in a ready-to-use 1C PUR wood adhesive. For example, hydrophilic
tertiary amines or organometallics are used for catalysts in 1C PURs [14, 15]. The curing
starts after application of the 1C PUR onto the adherend and is driven by the reactivity of
excess MDI isomers and moisture. The latter derives from the adherend and/or from ambient
humidity. Carbamic acid is built as an intermediate step of this reaction, which then
decomposes to a primary amin and carbon dioxide gas. Consequently, the amin (or an
optional chain extender) reacts with the residual isocyanate, forming urea linkages, which
take effect as strong crosslinking points in the final polymer network. The areas of the
polymer, which present polyurea-polyurethane groups, are referred to as HS of the cured
polyurethane polymer [9, 16]. The mechanical characteristics of the cured polymer, like
strength, stiffness or plastic workability, derive from the characteristics of the SS, the HS and
the interaction between those two by primary (covalent) bonds at the cross-linking points and
additional secondary (hydrogen) bonds [16, 17]. The strong hydrogen bonds (H bonds)
between the amino and carbonyl groups of urea and urethane considerably support the
cohesion of the polymer matrix, whereas the H bonds between HS and polyether are much
weaker. Therefore the HS tend to form agglomerations, leading to HS and SS rich phases in
the polymer [16]. This process may be considered as a microphase separation [18, 19]. Based
on the large variety of precursors, 1C PURs are quite versatile and, within a certain range,
adjustable in terms of their elasto-plastic behaviour. They can represent the behaviour of
duromers, elastomers or thermoplastic polymers [10, 20]. However, 1C PURs for wood
adhesives basically exhibit characteristics of semi-crystalline thermoplastics with a glass
transition temperature below room temperature [11, 16]. Besides tensile strength and Young’s
modulus, the ductility of the polymer films under changing ambient humidity is also a subject
of this investigation. In general, the ability of a material to plastically deform under load
without rupture is defined as ductility [10]. Ductility helps to redistribute internal and external
stresses, leading to appreciable deformation at limit states of serviceability before structural
collapse occurs [21]. Brittleness is the antonym of ductility and, as a rule of thumb, materials
having a fracture strain of less than about 5% are considered to be brittle [22]. In the case of
an extraordinary overload of a load bearing construction, ductility appears to be
44
advantageous, since an imminent collapse is announced by a preliminary deformation [7, 21],
providing precious time for safety intervention. In contrast, brittle materials collapse abruptly
without prior indication via deformation [22].
In the recent past various surveys investigated the influence of temperature on the mechanical
properties of adhesives or wooden glued joints [6, 16, 23, 24]. Studies regarding the influence
of relative humidity (RH) on the mechanics of adhesive polymers regarding the structural
bonding of wood are comparatively rare. Clad [25] was among the first concerned with the
mechanical properties of wood adhesives in general. Konnerth et al. [24] measured Young’s
modulus and hardness of four types of adhesive films (UF, MUF, PRF, 1C PUR) by means of
nanoindentation after treatment with fluid water and after redrying. The results showed a
distinct influence of water on the mechanical properties of all the tested adhesive films.
2. Materials and methods
2.1 Adhesives
In this survey six different adhesives were examined, which all belong to the cold-setting
network polymers [22]. Four of them are commercially used for structural wood bonding
(MUF, PRF, 1C PUR 1 and 1C PUR 2), two additional ones (prepolymer 1 and 2) were
produced for laboratory purposes only.
2.1.1 1C PUR adhesives
The tested 1C PUR 1 (HB S 309) and 1C PUR 2 (HB S 709) were both provided by the
adhesive manufacturer Purbond AG, Sempach-Station, Switzerland. In principle, the
compositions of these two adhesives are about 95% identical, but 1C PUR 1 contains about
three times more aminic catalysts than 1C PUR 2. Both were produced by means of the
prepolymer method. Further details regarding the polymer composition are a trade secret of
the manufacturer. The basic idea behind the modification is to make adhesive 1C PUR
systems available with almost equivalent mechanical properties, but with different processing
parameters (1C PUR 1: 30 min assembly time, 75 min curing time; 1C PUR 2: 70 min
assembly time, 175 min curing time). This was done with respect to customers’ needs, whilst
avoiding the full certification process for every single adhesive.
45
2.1.2 Prepolymers
Bayer MaterialScience, Leverkusen, Germany, provided the prepolymers Prepo 1 and Prepo
2. Their thermal stability has already been subject to recent investigations [16]. They were
produced within the same laboratory lot and on the same raw material basis. Their
specifications are listed in Table 1. The investigated prepolymers consist of a mixture of
methylene diphenyl diisocyanate (MDI) isomers (comprising 4,4’- MDI with a functionality >
2) and polyetherpolyols, but they do not present any additives. The polyetherpolyols were
specifically varied (Table 1) in order to study their influence on the mechanical properties of
the accordant polymer film under changing ambient moisture conditions.
Table 1: Chemical specifications of the prepolymers
Prepolymer 1 Prepolymer 2
Designation of the Prepolymer in [5] M N
Isocyanate content [%] 16
Propylene oxide content [%] 32.78 27.94
Ethylene oxide content [%] 4.83 9.66
NCO/PO/EO ratio [-] 1 / 2 / 0.3 1 / 1.7 / 0.6
Urethane groups [mol/kg] 0.77
Urea group content [mol/kg] 2
Cross-linking density [mol/kg] 0.21
Prepolymer properties, like bulk hydrophilicity, are strongly influenced by the SS [26]. EO
enforces hydrophilic properties, water uptake and water vapour permeance of the 1C PUR
polymer [15, 16, 27, 28]. Therefore, it is reasonable to assume, that a controlled variation of
the PO/EO ratio in a prepolymer affects its mechanical properties under changing ambient
humidity. The more hydrophilic Prepo 2 was expected to show more moisture sensitive
behaviour than Prepo 1 because it has double the content of EO. In 1C PUR wood adhesives
the use of PO is more common because of its higher hydrophobicity [16].
46
2.1.3 MUF and PRF
In this study the MUF Kauramin (BASF, Ludwigshafen, Germany), resin No. 683, hardener
No. 688, was used in a mix ratio of 100:60. A mix ratio between 100:20 and 100:100 is
recommended, based on the technical data sheet of the supplier. The resin represents a
melamine content of 20% to 30%. Further details are a trade secret. The investigated PRF was
Aerodux 185 (Dynea AS, Lillestrøm, Norway), resin No. 185, hardener HRP No. 155, mix
ratio 100 : 20.
2.2 Manufacturing of polymer films and specimens
From all the six polymers studied, cast films with a thickness of approximately 150 μm were
manufactured by means of a film applicator frame under controlled ambient climate
conditions. After sufficient hardening of the films shouldered test bar specimens according to
[29], specimens of type five (dimensions consistent with specimen type IV in [30]), were
punched out of the adhesive films using a moulding tool. This type of specimen is regarded as
suited for direct comparison between plastics of different rigidity [30]. Specimens with
significant voids, cracks, inconsistent thicknesses or other flaws were sorted out, as such
defects could influence the tensile test results [31].
2.3 Tensile tests
The specimens were randomly divided into batches and conditioned for at least two weeks in
controlled climate chambers in order to assure thorough acclimatization and the time-
dependent development of the polymer morphology [19]. The conditioning climates were
applied at 23°C as shown in Table 2.
Table 2: Conditioning batches for the specimens (23°C)
Air humidity RH (%) Water treatment Redrying
35 65 95 4 days immersed
10 days in 65% RH + 4 days immersed
+ redried in 65% RH for min. 2 weeks
Upon completion of the conditioning, the tensile tests were performed according to [32],
which is technically equivalent to [30]. For the testing procedure, a Zwick/Roell Z100
47
universal testing machine (500 N load cell) was used. The specimens were inserted into the
testing machine immediately after being taken out of the conditioning climate. After applying
an initial load of 0.4 N at a speed of 0.25 mm/min for accurate orientation of the sample and
its extensometer marks, the grips were moved at constant crosshead rates in order to perform
a displacement-controlled test. The testing speed for determination of Young’s modulus was
0.75 mm/min. for all tested polymers to achieve a strain rate of approx. 1% of measurement
length per minute in accordance with [32]. For determination of the tensile strength, a testing
speed of 5 mm/min was chosen for the 1C PURs and the prepolymers, because this is the
slowest common testing speed for polymer foils according to [29]. As this rate was still far
too fast for the very brittle PRF and MUF films (samples failed immediately after start of
testing), these polymers were tested at a speed of 2 mm/min based on [32].
During the test procedure, the longitudinal expansion of the specimens was measured by a
video extensometer system Mintron MTV-1362CA. The tensile tests were stopped at a max.
strain of 30%, because of limitations regarding proper video extensometer measurements on
wet specimens. Engineering tensile stress σ [MPa] and tensile strain Ɛ [%] values were
determined. Hence, tensile strength σm and Young’s modulus Et were computed according to
[32], where tensile strength is defined as the stress measured at the first stress-maximum on
the stress-strain curve. Most of the measured stress-strain-curves showed this maximum in the
form of an upper yield point. However, under high ambient moisture conditions some
polyurethane specimens no longer showed a clear yield point. In this particular case the
tensile strength was determined on the basis of proof strength Rp 3.3 (offset 3.3 %) for non-
proportional extension according to [33, 34]. In those stress-strain diagrams showing upper
yield strength, a parallel translation of the Hook straight line by 3.3% runs through the upper
yield point of the stress-strain curve. Therefore, Rp 3.3 provided the best possible
approximation of yield points in measured stress-strain diagrams without yield point. The
basis for the calculation of Young’s modulus was the elongation interval 0.05% ≤ Ɛ ≤ 0.25%
according to [32].
48
3. Results and discussion
The data gained from the tensile tests was examined for outliers according to [35] (Dixon
test). Outliers were excluded from further data examination. In accordance with [32], the
sample size of n = 5 was regarded as suitable since coefficients of variation are comparatively
low (Table 3). A 95% confidence interval of the mean values was used to ensure sufficient
reliability of the mean values. According to the Shapiro-Wilk Normality test at the 0.05 level,
the observations were significantly drawn from a normally distributed population. Table 3
provides the data gained from the tensile tests (overview).
Table 3a: Average values of tensile strengths σm and tensile moduli Et (v [%] = Coefficient of variation
Tensile strength Young’s modulus
Adhesive RH [%] σm [MPa] (v [%]) Et [GPa] (v [%])
1C PUR 1
5 34.3 (6.1) 1.5 (8.0)
35 28.2 (4.3) 1.2 (3.6)
65 23.7 (1.3) 1.1 (6.5)
95 19.9 (2.5) 0.8 (6.2)
water 19.7 (2.5) 0.8 (6.2)
redried 29.1 (6.2) 1.3 (3.9)
1C PUR 2
5 33.0 (0.9) 1.3 (5.9)
35 27.8 (2.5) 1.1 (6.5)
65 26.6 (4.5) 1.0 (6.6)
95 22.6 (1.3) 0.9 (2.5)
water 19.0 (3.7) 0.7 (4.0)
redried 32.6 (4.0) 1.3 (7.6)
Prepolymer 1
5 48.2 (1.7) 2.3 (4.1)
35 43.4 (1.6) 2.0 (5.7)
65 36.9 (1.9) 1.7 (3.9)
95 25.6 (3.5) 1.1 (4.4)
water 20.3 (2.0) 0.8 (1.0)
redried 42.2 (3.6) 1.9 (4.8)
Prepolymer 2
5 39.8 (5.5) 1.8 (5.8)
35 46.2 (1.9) 2.2 (7.1)
65 37.6 (4.8) 1.8 (8.5)
95 24.0 (3.3) 1.0 (5.8)
water 18.8 (4.8) 0.8 (12.7)
redried 36.8 (6.8) 1.8 (6.8)
49
Table 3b: Average values of tensile strengths σm and tensile moduli Et (v [%] = Coefficient of variation
Tensile strength Young’s modulus
Adhesive RH [%] σm [MPa] (v [%]) Et [GPa] (v [%])
MUF
5 66.6 (2.0) 5.1 (4.2)
35 46.6 (8.6) 3.5 (4.5)
65 39.8 (9.5) 2.5 (1.6)
95 31.5 (3.8) 1.2 (4.1)
water 45.5 (4.6) 2.2 (6.3)
redried 52.8 (1.7) 3.1 (8.2)
PRF
5 30.5 (19.3) 4.0 (7.0)
35 33.0 (9.1) 3.6 (7.6)
65 33.8 (13.3) 3.4 (8.9)
95 33.0 (12.7) 2.4 (10.6)
water 33.0 (26.7) 2.2 (18.9)
redried 28.2 (21.6) 2.8 (4.8)
3.1 1C PUR adhesives
Tensile strength and Young’s modulus of the tested 1C PURs and prepolymers reveal a linear
decrease with increasing ambient humidity (Fig. 1, 2). In principal, these findings go along
with Hakala et al. [36], who describe the reduction of Young’s modulus in polyurethanes
under wet conditions compared to dry conditions. They outline, that the presence of moisture
induces increased local mobility within the polymer film, polymer plasticization, and a
reduced glass-transition temperature, going along with a reduced Young’s modulus.
50
Figure 1: Tensile strength of polymer films as a function of ambient relative humidity. Lines: Linear regressions. Data points: Average values. Error bars: 95% -Confidence intervals of average values. R2: Coefficient of determination
Figure 2: Young’s moduli of polymer films as a function of ambient relative humidity. Lines: Linear regressions. Data points: Average values. Error bars: 95% -Confidence intervals of average values. R2: Coefficient of determination
51
The redried films of the 1C PURs (and of the prepolymers) regained their tensile strength and
their Young’s modulus measured at 65% RH with some even exceeding these values (Fig. 3,
4). This corresponds to Konnerth et al. [24], who also measured a decline of Young’s
modulus of elasticity of thin 1C PUR films after immersion in water and the recovery of
Young’s modulus after redrying. This reversibility of the loss of tensile strength and Young’s
modulus is a strong indicator of the importance influence of H bonds inside the polymer,
which are disrupted by water molecules entering the polymer film and that are re-established
whilst redrying. Yilgor et al. [18, 19] discussed the importance of H bonds for the mechanical
performance of segmented polyurethanes by influencing the morphology (microphase
separation) of the polymer. Thus, the high moisture content of the polymer before redrying
might have supported an advantageous re-arrangement of HS and SS in the polymer,
supported by the loss and re-establishment of the H bonds, the polymer polarity, and the
diisocyanate symmetry. Such a change in morphology might be the reason for the high tensile
strength and Young’s modulus measured after re-drying. However, the effect of re-drying on
the mechanical properties of polymer films is intended to be the subject of continuative
investigations, which will also consider the idea of post-crosslinking reactions caused by
residual isocyanate groups inside the polymer film, reacting with entering moisture at room
temperature [37].
52
Figure 3: Boxplots of tensile strengths of adhesive films as a function of ambient moisture,
including water treated and redried specimens. Whiskers show minimum and maximum
values.
0
10
20
30
40
50
60
70
1C PUR 1
05% RH
35% RH
65% RH
95% RH
water
redried
0
10
20
30
40
50
60
70
1C PUR 2
Prepolymer 1
Prepolymer 2
05% RH
35% RH
65% RH
95% RH
water
redried
Te
nsi
le s
tre
ng
th [M
Pa]
05% RH
35% RH
65% RH
95% RH
water
redried
PRF
MUF
53
Figure 4: Boxplots of Young’s moduli of adhesive films as a function of ambient moisture,
including water treated and redried specimens. Whiskers show minimum and maximum
values.
Over the whole range of tested climate stages, the moduli of elasticity of 1C PUR 1 and 1C
PUR 2 do not significantly differ from each other (Fig. 2, 4). In contrast, 1C PUR 2, which is
expected to be less hydrophilic due to the lower content of hydrophilic catalyst, shows higher
tensile strength values at 65% RH and 95% RH than 1C PUR 1 (Fig. 1, 3). From 35% RH to
95% RH, the 1C PUR 2 displays a tensile strength reduction of about 19%. Under the same
conditions, the tensile strength of 1C PUR 1 reveals a reduction of 30%. Hence, it seems fair
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
You
ng'
s m
odu
lus
[G
Pa
]
1C PUR 1
05% RH
35% RH
65% RH
95% RH
water
redried
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
1C PUR 2
Prepolymer 1
Prepolymer 2
05% RH
35% RH
65% RH
95% RH
water
redried
MUF
05% RH
35% RH
65% RH
95% RHwater
redried
PRF
54
to conclude that, under increased RH, a higher content of hydrophilic catalyst supports a
decrease in tensile strength of the polymer film. Additional research, in terms of the moisture
sorption behaviour of these polymer films, was carried out and will be the subject of a
following paper. Nonetheless, the results indicate that the manufacturer has complied with the
basic requirement to provide two 1C PUR polymers with comparable mechanical properties
and substantially different processing parameters.
3.2 Prepolymers
The results do not reveal a significant influence of the EO content on tensile strength or
Young’s modulus at humidity levels from 35% up to water storage (Fig. 1 - 4). Under dry
conditions (5% RH), Prepo 2 shows significantly lower tensile strength and Young’s modulus
compared to Prepo 1. The testing of this batch was repeated, revealing the same results. A
direct comparison between the prepolymers and the 1C PURs should be considered with
caution, because these polymers originate from different laboratories with different raw
material sources and manufacturing conditions. In addition, the detailed specifications of the
prepolymer used for the production of 1C PUR 1 and 1C PUR 2 is trade secret of the adhesive
manufacturer. As a general observation, it was found that under dry conditions (5% and 35%
RH) the tensile moduli of the 1C PUR adhesives are up to about 40% lower than the
corresponding moduli of Prepo 1. Corresponding results were found for the tensile strength
values of these polymers. This exemplarily illustrates the diversity of polyurethanes regarding
mechanical properties. Nonetheless, the findings reveal that the tensile strengths of all the
tested polyurethanes (1C PURs and prepolymers) do not significantly differ from each other
when tested in the wet state (water treatment). Accordant results were observed for the tensile
moduli of these polymers.
55
3.3 MUF and PRF
Among all the tested polymers MUF, achieved the highest values of tensile strength and
Young’s modulus (5% RH), accompanied by the steepest declines from lowest to highest RH
(Fig. 1). It lost about 50% of its tensile strength from 5% to 95% RH (Fig. 1, Table 3). The
corresponding reduction of Young’s modulus 75% (Fig. 2, Table 3). After redrying, the
polymer regained tensile strength and Young’s modulus (Fig. 3, 4) and even exceeded the
accordant values at 65% RH. Concerning MUF bonded particle boards, Ringena et al. [6]
report a correlation between the loss of shear strength and the mass loss of the polymer due to
hydrolysis. The results of the current study can partly be explained based on hydrolysis.
However, additional mechanisms should be taken into account when trying to find
explanations for the enhanced strength after redrying. For example such mechanisms could be
the re-arrangement of polymer structures by enhanced mobility during the wet stage, causing
additional crosslinking [38]. In combination with the influence of secondary bonds (based on
3.1) and re-condensation of the MUF polymer whilst re-drying these mechanisms could lead
to enforced post-crosslinking. However, the results indicate that a high RH (95%) causes a
greater reduction of strength and elasticity of MUF films than fluid water (Fig. 3, 4, 5). It is to
be expected that different results would be found with other ratio mixes (resin : hardener) and
other melamine contents of the resin, because the resistance of the MUF-polymer against
hydrolysis is also highly dependent on the molar melamine-formaldehyde ratio [2, 39].
However, on all tested climate stages the MUF and the PRF polymers show significantly
higher stiffnesses than the 1C PURs.
The tensile strength measured for PRF did not reveal a decline with increasing ambient
moisture, however, its Young’s modulus decreased significantly by about 50% when
changing from 5% RH to 95% RH (Fig. 1, 2). In general, the PRF specimens exhibited the
biggest variance (Fig. 3, 4) among all the tested polymers. Nonetheless, at any stage of RH,
the tensile strength and the modulus of elasticity of MUF and PRF were at least as high as the
accordant values of the best performing polyurethanes (Fig. 1, 2). Regarding PRF films, a
comparison of the values measured at 95% RH and after water treatment does not reveal
significant differences in tensile strength or Young’s modulus (Fig. 3, 4).
56
3.4 Stress-strain behaviour
Figure 5 exemplarily shows the tensile stress-strain charts of 1C PUR 2, MUF and PRF
measured at different ambient climate stages. The 1C PUR specimens revealed a ductile
behaviour under all of the tested moisture conditions. They were extended by a min. of 25%
before failure took place, which is substantially beyond the proportional limit. The findings
also show, that ambient moisture clearly affects the tensile stress level at which the creeping
of the 1C PUR starts (Fig. 5). At 95% RH it is about 30% lower than at 5% RH. The MUF
and PRF samples revealed a brittle behaviour on all tested climate stages, showing
comparatively little or no plastic deformability. The MUF samples ruptured within a range of
about 1.5% to 4.5% strain. The PRF films failed between 0.8% and 1.4% strain (5% RH up to
95% RH). After water treatment, the PRF specimens resisted strain up to about 2.5%.
57
Figure 5: Stress-strain charts of adhesive films (MUF, PRF, 1C PUR 2). A one curve
exemplar for each treatment is illustrated. Tensile tests were terminated at 30% max. strain.
58
4. Conclusions
The findings reveal that the behaviour of the investigated polymers changes significantly
under different RH, which typically occur during the service life of a load bearing
construction. On the one hand, these results should not be directly transferred to construction
elements like adhesively bonded structural beams in situ, since important influential
parameters, such as construction details or interactions between adhesive and adherend, are
not within the scope of this investigation. On the other hand, it was shown that just the
presence of water vapour is sufficient to greatly affect the properties of thin adhesive
polymers. The recovery of the polymers‘ Young’s moduli and tensile strengths after redrying
shows, in principle, that all the investigated adhesive polymers are capable of re-gaining
strength and elasticity, even after 4 days of immersion in water. Some findings of this paper
will be the basis for further investigations, which will focus on the moisture sorption
behaviour of the presented polymers.
Acknowledgements
We gratefully acknowledge the financial support of this research by the Commission for
Technology and Innovation CTI (Bern, Switzerland) and the Purbond AG (Sempach-Station,
Switzerland). Bayer MaterialScience (Leverkusen, Germany) kindly provided the
prepolymers. Türmerleim AG (Basel, Switzerland) kindly made Kauramin available.
References
[1] Wood Handbook, Wood as an Engineering Material, Forest Products Laboratory,
Madison, Wisconsin, USA, 2010.
[2] M. Dunky, P. Niemz, Holzwerkstoffe und Leime, Springer Verlag Berlin 2002.
[3] D. Ren, C.E. Frazier, Wood/adhesive interactions and the phase morphology of moisture-
cure polyurethane wood adhesives, Int. J. Adhes. Adhes., 34 (2012) 55-61.
[4] B.-D. Park, S.-M. Lee, J.-K. Roh, Effects of formaldehyde/urea mole ratio and melamine
content on the hydrolytic stability of cured urea-melamine-formaldehyde resin, Eur. J. Wood
Wood Prod., 67 (2008) 121-123.
59
[5] N.T. Paiva, A. Henriques, P. Cruz, J.M. Ferra, L.H. Carvalho, F.D. Magalhães, Production
of melamine fortified urea-formaldehyde resins with low formaldehyde emission, J. Appl.
Polym. Sci., 124 (2012) 2311-2317.
[6] O. Ringena, R. Janzon, G. Pfizenmayer, M. Schulte, R. Lehnen, Estimating the hydrolytic
durability of cured wood adhesives by measuring formaldehyde liberation and structural
stability, Eur. J. Wood Wood Prod., 64 (2006) 321-326.
[7] M. Brunner, M. Lehmann, S. Kraft, U. Fankhauser, K. Richter, J. Conzett, A Flexible
Adhesive Layer to Strengthen Glulam Beams, J. Adhes. Sci. Technol. , 24 (2010) 1665-
1701.
[8] E. Kehr, G. Riehl, E. Hoferichter, E. Roffael, B. Dix, Feuchtebeständigkeit und
Hydrolyseresistenz von Holz-zu-Holz-Bindungen in Spanplatten, hergestellt mit
formaldehydarmen modifizierten Harnstoff-Formaldehydharzen unter Einsatz verschiedener
Härtungsbeschleunigersysteme, Eur. J. Wood Wood Prod., 52 (1994) 253-260.
[9] A. Pizzi, K.L. Mittal, Handbook of Adhesive Technology, 2 ed., Marcel Dekker Inc.,
Cimarron Road, Monticello, New York 12701, USA, 2003.
[10] D.R. Askeland, P.P. Fulay, W.J. Wright, The Science and Engineering of Materials, 6
ed., Cengage Learning, 200 First Stamford Place, Suite 400, Stamford, CT 06902, USA,
2011.
[11] W. Kaiser, Kunststoffchemie für Ingenieure. Von der Synthese bis zur Anwendung, Carl
Hanser Verlag, Germany, 2007.
[12] IARC Monograph, VOLUME 88: Formaldehyde, 2-Butoxyethanol and 1-tert-
Butoxypropan-2-ol, World Health Organization (WHO), Intl. Agency for Research on
Cancer, Lyon, France, 2006.
[13] I. Hartung, S. Boehm, Civil Constructions, in: Da Silva L.F.M., Öchsner A., Adams R.D.
(Eds.), Handbook of Adhesion Technology, Springer, Heidelberg, 2011, pp. 1264 - 1288.
[14] J.W. Britain, P.G. Gemeinhardt, Catalysis of the Isocyanate-Hydroxyl Reaction, J. Appl.
Polym. Sci., 4 (1960) 207-211.
[15] B. Müller, Additive kompakt, Farbe und Lack Edition ed., Vincentz Network GmbH &
Co. KG, 30062 Hannover, Germany, 30062 Hannover, Germany, 2009.
60
[16] S. Clauβ, D.J. Dijkstra, J. Gabriel, O. Kläusler, M. Matner, W. Meckel, et al., Influence
of the chemical structure of PUR prepolymers on thermal stability, Int. J. Adhes. Adhes., 31
(2011) 513-523.
[17] U. Sebenik, M. Krajnc, Influence of the soft segment length and content on the synthesis
and properties of isocyanate-terminated urethane prepolymers, Int. J. Adhes. Adhes., 27
(2007) 527-535.
[18] I. Yilgor, E. Yilgor, I.G. Guler, T.C. Ward, G.L. Wilkes, FTIR investigation of the
influence of diisocyanate symmetry on the morphology development in model segmented
polyurethanes, Polymer, 47 (2006) 4105-4114.
[19] I. Yilgor, E. Yilgor, Structure-morphology-property behavior of segmented thermoplastic
polyurethanes and polyureas prepared without chain extenders, Polym. Rev., 47 (2007) 487-
510.
[20] G. Habenicht, Kleben - erfolgreich und fehlerfrei, 6 ed., Vieweg + Teuber, Wiesbaden,
Germany, 2012.
[21] J. De Castro San Romàn, System Ductility and Redundancy of FRP Structures with
Ductile Adhesively-Bonded Joints, Faculté Environnement Naturel, Architectural et
Construit, Section d' Architecture, École Polytechnique Fédérale de Lausanne (EPFL),
Switzerland, 2005.
[22] W.D. Callister, D.G. Rethwisch, Materials Science and Engineering, 8 ed., John Wiley &
Sons, Inc., 222 Rosewood Drive, Danvers, MA 01923, 2011.
[23] S. Clauß, M. Joscak, P. Niemz, Thermal stability of glued wood joints measured by shear
tests, Eur. J. Wood Wood Prod., 69 (2010) 101-111.
[24] J. Konnerth, F. Stöckel, U. Müller, W. Gindl, Elastic properties of adhesive polymers.
III. Adhesive polymer films under dry and wet conditions characterized by means of
nanoindentation, J. Appl. Polym. Sci., (2010) 1331-1334.
[25] W. Clad, Ueber das Wesen einer Verklebung und die Fugenelastizität ausgehärteter
Leimfugen bei Holzverleimungen, Technische Hochschule Stuttgart, Germany, 1964.
[26] G.A. Skarja, K.A. Woodhouse, Structure-property relationships of degradable
polyurethane elastomers containing an amino acid-based chain extender, J. Appl. Polym. Sci.,
75 (2000) 1522-1534.
61
[27] G.A. Skarja, K.A. Woodhouse, Synthesis and characterization of degradable
polyurethane elastomers containing and amino acid-based chain extender, J. Biomater. Sci.
Polym. Ed., 9 (1998) 271-295.
[28] Y. Gnanou, Hydrophilic Polyurethane Networks Based on Poly(ethylene oxide):
Synthesis, Characterization, and Properties. Potential Applications as Biomaterialst,
Macromolecules, 17 (1984) 945-952.
[29] DIN EN ISO 527-3:2003 Plastics – Determination of tensile properties – Part 3: Test
conditions for films and sheets, German Institute for Standardization, Beuth Verlag, Berlin,
2003.
[30] ASTM D 638-10 Standard Test Method for Tensile Properties of Plastics, ASTM
international, West Conshohocken.
[31] L.F.M. Da Silva, A. Öchsner, R.D. Adams, Handbook of Adhesion Technology, Springer
Verlag, Berlin Heidelberg, Germany, 2011.
[32] DIN EN ISO 527-1:2010 Plastics – Determination of tensile properties – Part 1: General
principles, German Institute for Standardisation. Beuth Verlag, Berlin, 2010.
[33] ISO/TR 25679:2005 Mechanical testing of metals – Symbols and definitions in published
standards, ISO, Geneva, 2005.
[34] B. Suryanarayana, Experimental Techniques in Materials and Mechanics, CRC Press,
Taylor & Francis Group, LLC, 6000 Broken Sound Parkway NW, Suite 300, Boca Raton,
Florida, USA, 2011.
[35] DIN 53804-1:2002 Statistical evaluation - Part 1: Continuous characteristics, German
Institute for Standardization, Beuth Verlag, Berlin, 2002.
[36] K. Hakala, R. Vatanparast, E. Vuorimaa, H. Lemmetyinen, Monitoring water uptake of
polyurethanes by in situ fluorescence technique, J. Appl. Polym. Sci., 82 (2001) 1593-1599.
[37] G. Tillet, B. Boutevin, B. Ameduri, Chemical reactions of polymer crosslinking and post-
crosslinking at room and medium temperature, Prog. Polym. Sci., 36 (2011) 191-217.
[38] A. Root, P. Soriano, The Curing of UF Resins Studied by Low-Resolution 1H-NMR, J.
appl. Polym. Sci., 75 (2000) 754-765.
62
[39] T.W. Lee, E. Roffael, B. Dix, H. Miertzsch, Influence of Different Catalyst Systems on
the Hydrolytic Stability of Particleboards Bonded with Unmodified and Modified UF-Resins,
Holzforschung, 48 (1994) 101-106.
63
4.2 Paper II
International Wood Products Journal (2014), 5, 1: 18 - 26
Influence of wood machining on tensile shear strength and wood
failure percentage of one-component polyurethane bonded wooden
joints after wetting
Oliver Kläusler *1, Klaus Rehm 2, Falko Elstermann 3, Peter Niemz 1
1 ETH- Zürich, Institute for Building Materials, Schafmattstrasse 6, 8093 Zürich, Switzerland
2 Bern University of Applied Sciences, Architecture, Wood and Civil Engineering,
Solothurnstrasse 102, 2500 Biel 6, Switzerland
3 BA Sachsen, Department of Wood Technology, University of Cooperative Education, Hans-
Grundig- Straβe 25, 01307 Dresden, Germany
*1 Corresponding author’s email: [email protected]
Abstract
This study determines the influence of mechanical surfacing on tensile shear strength (TSS)
and wood failure percentage (WFP) of beech wood (Fagus sylvatica L.) in the wet state acc.
to prEN 302-1:2011 (tensile shear tests). The wood was planed, sanded and face milled, using
different qualities of cutting edges and sanding grits. Roughness and wettability of the
adherends were characterised, supplemented by ESEM images. The specimens were bonded
by means of a one-component polyurethane adhesive and tested at the dry stage (pre-
treatment A1), in the wet state (A4) and after re-drying (mA5). Results determined that the
wettability and the wood structure’s integrity are considered more decisive for the TSS in the
wet state than surface roughness. Planing with very dull knives caused the lowest TSS and
64
WFP after A4 and the most subsurface damage. Face milled and sanded batches revealed the
best performance. TSS and WFP decreased significantly from A1 to A4, but recovered after
mA5.
Keywords
Wood machining, bonding, 1C PUR, contact angle, tensile shear strength, wood failure
percentage, roughness, sanding, planing, face milling
Introduction
The performance of wooden load bearing elements, such as glue laminated timber, is
influenced by various factors such as dimensions, wood properties, adhesive properties or the
manufacturing process. This study is concerned with the effect of wood machining on the
performance of wooden glued joints for such structural elements. Since machining
parameters, like tool geometry, change significantly due to wear (Pahlitzsch and Schulz 1957,
Davim 2008), various studies were concerned with the effect of changes in machining on the
quality of adherend surfaces and the accordant wooden glued joints in the dry state. The
present study enlarges this focus. It deals with the influence of mechanical surfacing and tool
wear on tensile shear strength (TSS) and wood failure percentage (WFP) of beech wood when
tested as described in the methods.
Some common techniques for the mechanical preparation of adherends include planing, belt-
sanding and face milling. The effect of face milling on adherend surfaces during the dry stage
was, for instance, investigated by Riegel (1997) and Heisel and Tröger (1993) who concluded
that this machining principle generates surfaces of higher quality than orthogonal planing,
because a properly maintained face milling machine causes less cutting forces and
consequently less subsurface damage of the wood structure. Such an intact structure of the
adherend helps to avoid the formation of mechanically weak boundary layers (de Moura et al.
2010). Singh et al. (2002) were concerned with orthogonal planing. Their results revealed that
wear of the cutting edge leads to progressive formation of micro-clearance angles, which
cause increased wood compression due to higher cutting forces. Therefore planing with dull
tools causes severe damage in the wood structure, leading to reduced substrate integrity and
bonding strength during the dry stage (River et al. 1991). As Cool and Hernandez (2011b)
report, the belt sanding process can be performed without seriously damaging the subsurface
wood tissue. However, this heavily depends on the applied machining parameters, like e.g.
65
grit size or feed speed. In addition, the wood itself also contributes to the surfacing result,
since factors like wood density, porosity and fibre angle influence the surface profile
(Murmanis et al. 1986).
Little is known about the specific effect of different surfacing techniques and knife wear on
the bonding performance tested in the wet state, although various technical standards set
accordant thresholds for structural elements under exterior exposure. It must also be taken
into consideration that even structures that are not supposed to be exposed to water might be
subjected to substantial wetting during service life (Greiner-Mai 2006). In Europe, wooden
glued joints for load bearing structures have to meet the demands of EN 1995-1-1 (Eurocode
5), whereas standards like CSA O112.9-04 or ASTM D 2559-04 set benchmarks in North
America (Table 1).
Table 1 Cited technical standards in brief
Technical standard
Abbreviation for treatment of
specimens before testing
Description of treatment
Threshold values
Allowed kinds of
machining
Tensile shear
strength [MPa]c
Wood failure percentage
(WFP)
EN 15425 EN 302-1
Tensile shear tests
A1 conditioning at 20°C / 65% RH
b 10
none
planing or sanding (grain size
P100)
A4 specimens for 6h in boiling
water + 2h submerged at room temperature
6
mA5 a A4 + kiln drying at 103°C + A1 8
CSA O112.9 Compression
shear test (inter alia)
dry conditioning conditioning at 20°C / 65% RH
b 60
(hardwood)
planing (sanding
prohibited) vacuum-pressure
submerged whilst vacuum and pressure treatment, test in the
wet state
80 (hardwood)
boil-dry-freeze boiling, hot drying, freezing (8
cycles), test in the wet state
ASTM D 2559 – 04 Compression
shear test (inter alia)
dry conditioning conditioned at 23°C / 65% RH b
75 no
limitations
a mA5 is a modified version of the original A5. The latter does not comprise the kiln drying step. b RH: Relative humidity of ambient air [%] c Adhesive type I and 0.1 mm thickness of adhesive layer.
66
Amongst others, Eurocode 5 refers to prEN 301:2011, which explains that adhesives for full
weathering (adhesive type I) have to be tested according to prEN 302-1:2011 (tensile shear
test) and other standards (delamination tests, creep tests, etc.). If the glued joints are produced
by means of a one-component polyurethane adhesive, they have to meet the requirements of
DIN EN 15425. In this study, glued joints were subjected to the tensile shear tests. Prior to
glue application, the quality of the machined mating surfaces was determined by examining
wettability, roughness and ESEM micrographs. In general, sufficient wetting of the adherend
by the adhesive is important for durable bonding (Wellons 1980, River et al. 1991).
Hernandez and Cool (2008b) stress the importance of good wettability for good adhesion as it
is an essential precondition for mechanical and chemical interlocking and for secondary force
interactions between wood and adhesive. To define a parameter for wettability, the contact
angle of a fluid drop on the solid wood surface is measured. However, due to its porosity,
anatomical and chemical inhomogeneity and hygroscopicity, wood is a non-ideal surface for
contact angle measurements in principle (Gindl et al. 2004; Santoni and Pizzo 2011).
Therefore the measured values should be interpreted with caution.
The influence of the adherends’ surface roughness on the resulting bonding quality was a
matter of consideration in several studies. Hernandez and Cool (2008b) investigated the effect
of surfacing methods on roughness and wettability of wood surfaces. Their findings exhibited
better wettability of the rough surfaces compared to the smooth ones. Moreover de Moura et
al. (2010) concluded that an excessive reduction of roughness could even be disadvantageous
for bonding due to the reduced participating adhering surface. Follrich et al. (2010) studied
the dependence of the bond strength on the surface roughness of the adherends. They
observed increased tensile strength with increased surface roughness. This corresponds with
River et al. (1991) who describe that, up to a certain point, roughness supports bonding
strength, immaterial of whether the roughness is caused by machining or by the wood’s
natural porosity itself. However, the various findings regarding influence of roughness on
bonding performance are not fully consistent. Cool and Hernandez (2011a) examined the TSS
of wooden glued joints after different machining techniques. Surface roughness was lowest
after sanding (grit 80), whereas the surfaces after face milling or planing were about twice as
rough. In this study Cool and Hernandez did not find a significant influence of the roughness
on TSS or WFP. The objective of the current research is to study the influence of different
mechanical surfacing techniques on TSS and WFP of wooden glued joints. In particular the
67
focus is placed on the performance of the joints in the wet state, since various technical
standards set accordant thresholds and studies regarding this topic are very rare.
Materials and methods
Wood
Based upon prEN 302-1: 2011 for tensile shear tests, boards of European beech wood (Fagus
sylvatica L.) were conditioned (20°C / 65% RH) until an equilibrium moisture content of
about 12% was reached. Boards with “defects” like wavy direction of grain, discolorations,
etc. were disqualified for this study. The conditioned boards were cut to size conforming to
the said standard and mixed in order to scatter wood influences over the whole random
sampling.
Surfacing methods
Subsequent to conditioning, the boards were subjected to different mechanical surfacing
methods as listed below in order to generate suitable joining surfaces. These procedures were
repeatedly applied to each board, thus eliminating influences of previous work steps, like
coarse thickness calibration (Table 2: “Min. number of work steps per workpiece”). Various
machining parameters affect the surface characteristics (Pahlitzsch and Schulz 1957; River et
al. 1991; Hernandez and Cool 2008; Cool and Hernandez 2011a & b). Therefore the applied
machine settings are listed in Table 2.
68
Table 2 Applied machining parameters. Terms according to ISO 3002/1-1982 and ISO
3002/3-1984
Planing
Planing was carried out on a thickness planing machine equipped with a two-knife cutter head
holding the knives parallel to its rotation axis (peripheral orthogonal planing). The first pair of
knives used for planing was freshly sharpened and remained unjointed (“over-sharp”). The
other pair was identical in design, but at the end of its tool life, having very blunt cutting
edges. In default of a suitable standard test method for proper evaluation of the cutting edge
quality, pictures of the knifes’ cutting edges were taken by means of a reflected light
microscope. The experimental setup is schematically displayed in Figure 1.
Planing Face milling Sanding
Surfacing machine Thickness planer CNC machine Wide-belt sander
Cutting and sanding materials
Sharp cutting edges (un-
jointed), High speed steel
(HSS)
Blunt cutting edges, HSS
cutting edges in middle of
tool life P80 P120
Feed speed (νf) [m min -1]
8 5 6
Number of cutting edges (z)
2 3
Feed (fz) [mm] 0,5 0,5
Cutting speed (νc) [m s-1]
~ 50
Cutter Head speed (n) [rpm]
8000
Working engagement of the cutting edge (ae) [mm]
per work step 1 2 0.3
Min. number of work steps per workpiece
2 2 3
69
Figure 1 (drawing, left): Scheme of the experimental set-up setup for the assessment of the cutting edge qualities. Left: Oversharp knife. Right: Dull knife Figure 2 (micro photographs, centre & right): Images of the planing knives’ cutting edges. The light shines from the left margin of the photograph onto the flank of the wedge. Left: Approx. 8 μm wide cutting edge of the over-sharp knife. Right: Approx. 58 μm wide micro-flank resulting from blunting of the edge
As Figure 2 shows, the cutting edge of the over-sharp knife was about 8 μm wide. By
contrast, the blunt and rounded cutting edge revealed a micro-flank of approximately 58 μm
width. Such micro-flanks are formed due to progressive knife wear or by jointing (Riegel
1997, Maier 2000; Davim 2008). Knife wear leads to progressively changing tool geometry
(clearance angle, wedge angle, rake angle, cutting edge) due to increased cutting forces
(Pahlitzsch and Schulz 1957, Fritz and Schulze 2008).
Sanding
In addition to planing, the standard prEN 302-1:2011 permits sanding of the adherends using
grit P100, whereas CSA O112.9-04 does not allow sanding of adherend surfaces (Table 1). In
the present study, the sanding treatments were performed by means of a commercial wide-belt
sander provided with sanding belts of grit P80 and P120. The dust extraction system of the
sander removed material lying loosely on the machined surfaces and cleaned the sanding belts
to a certain degree.
Face milling
A CNC-machining centre was used for the face milling. Milling of the bonding surfaces was
accomplished by the three active minor cutting edges (corner radius rƐ = 1 mm) of the face
milling cutter head. For further machining parameters see Table 2.
70
Roughness measurement and contact angle evaluation
Within 24 h after machining, the bonding surfaces were examined by measuring roughness
and contact angle. This complies with the requirements of prEN 302-1:2011, regarding the
maximum time span permitted between joining surface preparation and glue application. For
the comparative evaluation of the surface roughness, a tactile measuring method was applied
by means of a Hommel Tester T8000 device. Surface examination was performed on the basis
of DIN EN ISO 4287:2010 and de Moura (2010) on a measuring length of 4 mm with a cut-
off length of 0.8 mm at a speed of 0.5 mm/s using a diamond tip. The surface roughness
parameter Ra (arithmetic mean of the absolute ordinate values) was determined. According to
Volk (2005) this parameter is preferred for the documentation of the gradual changes of
surfaces due to tool wear. For each surface machining method, the assessment of the surface
profile was performed five times parallel to the fibre direction (Ra ǁ) and five more times
perpendicular to the grain (Ra ┴). The arithmetic average value (Ra x) was calculated from
these ten values for each mechanical surface treatment
On the basis of prEN 828:2010 and Santoni and Pizzo (2011), the static contact angle θ was
measured as a wettability parameter of the machined surfaces (sessile drop technique). The
smaller θ is, the better is the wetting ability, because a small contact angle indicates good
spreading of the fluid over the solid surface. For each batch, 10 droplets (20 μl each) of aqua
destillata were applied randomly onto different spots of the machined surfaces. Contact angles
were measured at a frequency of two pictures per second over 30 s testing time. Data
processing was done using the software ConAngle v.9 and a Gaussian filter. For better
comparability, the θ after 20 s was chosen as the crucial wettability value (θ20). A θ20
corresponds with the onset of the final phase of contact angle development, which is
characterized by a slow and approximately constant decrease of the contact angle over time
(dθ/dt) compared to the much faster and inconsistent change in contact angle measured
between 0 s and 20 s of wetting time (Santoni and Pizzo 2011). After 20 s a constant wetting
rate angle (Nussbaum 1999) was reliably reached.
Bonding and sample preparation
The adherends underwent the gluing process within 24 h after surface preparation. For all the
bondings, the one-component moisture-curing polyurethane (1C PUR) adhesive HB S309
(PURBOND®, Sempach-Station, Switzerland) was used, which is approved for load bearing
71
timber constructions in Europe.150 g/m2 of HB S309 was applied to one side by means of a
toothed spatula. Pressing of the adherends was performed for 75 min at 0.8 MPa in a
calibrated press using a pressing jig. Consequently, the pressed parts were conditioned at 20
°C and 65 % RH for at least one week. The climatised pressings were cut to tensile shear test
samples.
Sample treatment before testing and testing procedure
Prior to testing, the specimens of each surfacing method were mixed and afterwards divided
into three different batches for the treatments displayed in Table 1. In the follow-up to the
treatments, tensile shear tests according to the aforementioned standard were performed on a
Zwick/Roell Z100 universal testing machine. At the constant testing speed of 1 mm/min the
specimens failed between 30 s and 90 s in accordance with the said standard. For each batch,
15 samples were prepared for testing. The evaluation of the WFP on the fracture surfaces of
the dry specimens was performed with the naked eye.
Results and discussion
The results of the TSS tests revealed that all the tested surfacing methods generated wooden
glued joints that comply with the DIN EN 15425:2008 requirements for testing at the dry
stage (A1). After the treatments A4 and mA5 all batches other than planing with dull knives
matched the aforementioned standard. In general the measured TSS were lowest after A4,
thus confirming the results of Richter and Schirle (2002). Within this treatment group the face
milled and the sanded (P120) batches performed best. Also within groups A1 and mA5, the
face milled batches showed strongest performance, followed by the sanded (P120) specimens.
Findings within each group of machining techniques reveal that the TSS diminishes in the wet
state (A4 compared to A1) but recovers by re-drying (mA5 compared to A4 and A1). Hence,
the loss of strength during the wet stage is reversible (Fig. 3).
72
Figure 3 Tensile shear strength. Whiskers show minimum and maximum values, square in
box displays arithmetic mean value, horizontal line shows median.
Figure 4 wood failure percentages. Whiskers show minimum and maximum values, square in
box displays arithmetic mean value, horizontal line shows median.
73
Only batches planed with the blunt knives did not reveal such a recovery (mA5 compared to
A1). In this study a substantial influence of knife edge sharpness on TSS was notable only in
batches that were treated with water (A4, mA5). This result corresponds with the findings of
Bustos et al. (2010). The ESEM images (Fig. 7) of the adherends planed with dull knives
exhibit substantial damage of the subsurface wood tissue. Here, cells are considerably
distorted and partly ruptured: a finding that corresponds with Singh et al. (2002). It is likely
that swelling and shrinking enforce these damages, leading to an even weaker wood structure.
After A4 the specimens were tested during the wet stage in a condition of maximum swelling,
whereas during mA5, the samples had to undergo maximum swelling followed by maximum
shrinking and final reconditioning to 12 % EMC. Despite this, the TSS after mA5 is about as
high as after A1 (except for samples planed with dull knives). It can thus be concluded that
swelling and shrinking do not seriously harm 1C PUR-bonded glued joints – provided that the
wood tissue is not substantially damaged. In contrast, the moisture content of the specimens at
the moment of testing appears to be crucial for the TSS (Fig. 3) and WFP (Fig. 4). However,
it should be noted that repeated swelling and shrinking (alternating climate tests) or varied
sample geometries (e.g. bigger cross sectional areas of the adherends) might produce different
outcomes.
The inspection of the fracture surfaces revealed significantly less WFP after A4 compared to
A1 (Fig. 4). After A4 none of the determined medians achieved 20 % WFP. This applies to all
the examined batches regardless of the surfacing method. In this respect it should be taken
into account that wood loses strength with increasing moisture content up to fibre saturation
(Kollmann 1951, Niemz 1993). Obviously the TSS of the 1C-PUR bonded joints in the wet
state is not high enough to surpass the reduced cohesive strength of beech wood in the wet
state. In general the WFP after mA5 were much higher than those after A4, which correlates
with the results observed for TSS. The sanded mA5 batches performed best, whilst those
planed with dull knives showed poorest WFP performance. The sanded batches (P80 and
P120) even reveal a complete recovery of the WFP after mA5 compared to A1.
Unfortunately, certain standards do not permit sanding (Table 1). In general, the boxplots for
WFP display a large scatter of results and skewed distributions (Fig. 4). This goes along with
the Canadian standard CSA O112.9-04, which mentions that sets of WFP data that are
normally distributed are very rare.
74
The roughness measurements reveal that face milling, comparatively, generates the roughest
surfaces by far (Fig. 5).
Figure 5 Roughness Ra (arithmetic mean of ordinate values measured within the reference length). Whiskers show minimum and maximum values, square in box displays arithmetic mean value, horizontal line shows median.
Within the current study, no significant differences were found regarding roughness after
sharp planing, dull planing or sanding with two different grits. A more detailed differentiation
requires further investigations.
75
Figure 6 Contact angles. Data points show mean out of 10 measurements. Box plots: Whiskers show minimum and maximum values, square in box displays arithmetic mean value, horizontal line shows median. Figure corrected after [87]
The results of the contact angle measurements indicate better wettability of the face milled
and sanded surfaces compared to the planed ones (Fig. 6), going along with Santoni and Pizzo
(2011). Cool and Hernandez (2011b) also measured better wettability on rough surfaces
compared to smoother ones. The change in the sharpness of the cutting edge did not
significantly change the wettability of the wood surface.
0
10
20
30
40
50
60
70
80
90
0 10 20
Con
tact
ang
le θ
[°]
Wetting time (s)
Planing sharp Planing dull Sanding P80 Sanding P120 Face milling
76
Figure 7 ESEM-images of adherend surfaces after machining: a) Planing sharp; b) Planing dull; c) Sanding P120; d) Sanding P80; e) Face milling
77
Conclusions
Tensile shear strength
The results of the TSS tests show that wood machining not only influences wooden glued
joints at the dry stage, but has an even greater influence on the bonding performance in the
wet state or after re-drying. Whilst face milled and sanded (P120) samples showed the best
over-all TSS (A1, A4, mA5), the batch produced with dull knives showed the poorest results
in the wet state (A4). These samples did not comply with the requirements of Eurocode 5 and
DIN EN 15425:2008, respectively. Also the redried batches (mA5) revealed much better TSS
after planing with sharp knifes compared to planing with dull knives. Therefore planing with
very dull knives should be avoided in the production of glued joints for structural wood
products. Regarding the moisture dependent performance of the 1C PUR bonded joints, the
moisture content of the sample at the moment of testing revealed to be much more important
than the sample’s wetting history. The re-establishment of the bonding strength after re-drying
is a strong indication of the major importance of secondary bonds like hydrogen bonds. Such
bonds are present within the adhesive polymer as well as in the wood adhesive interphase.
They are disrupted when water enters the glued joint and they can re-establish as soon as re-
drying takes place.
Wood failure percentage
Generally the outcomes of the WFP assessment should be considered with caution, since the
data show quite skewed distributions. Nonetheless the results reveal that the WFP of 1C PUR
bonded wooden joints in the wet state cannot reliably be influenced by choosing a specific
mechanical surfacing method.
Roughness and contact angle
The face milled surfaces were the roughest, showing comparatively good wettability. The
sanded batches also revealed good wettability, despite having about half the roughness of the
face milled ones. Since these two surfacing methods generated the best TSS, roughness
obviously was not a crucial factor for the strength of the joints. Indeed sufficient wettability of
the adherend and integrity of the wood structure are considered as important preconditions for
good bonding quality.
78
Acknowledgements
We gratefully acknowledge the financial support of this research by the Commission for
Technology and Innovation KTI/CTI (Bern, Switzerland) and the Purbond AG (Sempach-
Station, Switzerland).
References
ASTM D 2559-04 2004. Standard Specification for Adhesives for Structural Laminated
Wood Products for Use Under Exterior (Wet Use) Exposure Conditions. ASTM
International.
Bustos, C. A., Moya, C. L., Lisperguer, J. M. and Viveros, E. M. 2010. Effect of knife wear
on the gluability of planed surfaces of radiata pine. Wood Fiber Sci 42: 185-191.
Cool, J. and Hernandez, R. E. 2011a. Evaluation of four Surfacing Methods on Black Spruce
Wood in Relation to Poly(Vinylacetate) Gluing Performance. Wood Fiber Sci
43: 194-205.
Cool, J. and Hernandez, R. E. 2011b. Improving the Sanding Process of Black Spruce Wood
for Surface Quality and Water-based Coating Adhesion. Forest Prod. J. 61: 372-380.
CSA O112.9-04 2004. Evaluation of adhesives for structural wood products (exterior
exposure). Canadian Standards Association.
Davim, J. P. 2008. Machining - Fundamentals and Recent Advances, London, Springer.
De Moura, L. F., Cool, J. and Hernandez, R. E. 2010. Anatomical Evaluation of Wood
Surfaces Produced by Oblique Cutting and Face Milling. IAWA J. 31: 77-88.
DIN EN 1995-1-1:2010. Eurocode 5: Design of timber structures- Part 1-1: General-
Common rules and rules for buildings; German version EN 1995-1-1:2004 + AC:2006
+ A1:2008. Deutsches Institut für Normung.
DIN EN ISO 4287:2010. Geometrical Product Specifications (GPS) - Surface texture: Profile
method - Terms, definitions and surface texture parameters. Deutsches Institut für
Normung.
79
DIN EN 15425:2008. Adhesives – One component polyurethane for load bearing timber
structures – Classification and performance requirements. Deutsches Institut für
Normung.
Follrich, J., Vay, O., Veigel, S. and Müller, U. 2010. Bond strength of end-grain joints and its
dependence on surface roughness and adhesive spread. J Wood Sci 56: 429-434.
Fritz, A. H. and Schulze, G. 2008. Fertigungstechnik, Berlin, Springer.
Gindl, M., Reiterer, A., Sinn, G. and Stanzl-Tschegg, S. E. 2004. Effects of surface ageing on
wettability, surface chemistry, and adhesion of wood. Holz Roh Werkst 62: 273-280.
Greiner-Mai, D. 2006. Gutachten zu den Ursachen des Einsturzes der Eissporthalle in Bad
Reichenhall vorgelegt. Bautechnik 83: 667-668.
Heisel, U. and Tröger, J. 1993. Qualitativ hochwertige Oberflächen durch Stirnplanfräsen.
HOB 40: 80-88.
Hernandez, R. E. and Cool, J. 2008a. Effects of cutting parameters on surface quality of paper
birchwood machined across the grain with two planing techniques. Holz Roh Werkst
66: 147-154.
Hernandez, R. E. and Cool, J. 2008b. Evaluation of three surfacing methods on paper birch
wood in relation to water- and solvent-borne coating performance. Wood Fiber Sci
40: 459-469.
ISO 3002/1 1982. Basic quantities in cutting and grinding - Part 1: Geometry of the active
part of cutting tools - General terms, reference systems, tool and working angles, chip
breakers. International Organization for Standardization.
ISO 3002/3 1984. Basic quantities in cutting and grinding - Part 3: Geometric and kinematic
quantities in cutting. International Organization for Standardization.
Kollmann, F. 1951. Technologie des Holzes und der Holzwerkstoffe, Berlin, Springer.
Maier, G. 2000. Holzspanungslehre und werkzeugtechnische Grundlagen, Würzburg, Vogel,
Kamprath-Reihe.
80
Murmanis, L., River, B. H. and Stewart, H. A. 1986. Surface and subsurface characteristics
related to abrasive-planing conditions. Wood Fiber Sci 18: 107-117.
Niemz, P. 1993. Physik des Holzes und der Holzwerkstoffe, Leinfelden-Echterdingen, DRW.
Nussbaum, R. M. 1999. Natural surface inactivation of Scots pine and Norway spruce
evaluated by contact angle measurements. Holz Roh Werkst 57: 419-424.
Pahlitzsch, G. and Schulz, K. 1957. Schnittkraftmessung und Schneidenabstumpfung beim
Hobeln von Holz mit kreisender Schnittbewegung. Holz Roh Werkst 15: 159-170.
prEN 301:2011. Adhesives, phenolic and aminoplastic, for load-bearing timber structures-
Classification and performance requirements. European Committee for
Standardization.
prEN 302-1:2011. Adhesives for load-bearing timber structures- Test methods- Part 1:
Determination of longitudinal tensile shear strength. European Committee for
Standardization.
prEN 828:2010. Adhesives - Wettability - Determination by measurement of contact angle
and surface free energy of solid surface. European Committee for Standardization.
Richter, K. and Schirle, M. A. 2002. Behaviour of 1K PUR adhesives under increased
moisture and temperature conditions. In: A. Teischinger and R. Stingl (eds),
Lignovisionen, Proceedings of the International Symposium on Wood Based
Materials, University of Natural Resources and Applied Life Sciences (BOKU),
Vienna 153 - 158.
Riegel, A. 1997. Gefügeschädigungen durch spanende Bearbeitung - Stand der Erkenntnisse
und mögliche Messmethoden. Holz Roh Werkst 55: 111-117.
River, B. H., Vick, C. B. and Gillespie, R. H. 1991. Wood as an Adherend, New York, Marcel
Dekker.
Santoni, I. and Pizzo, B. 2011. Effect of surface conditions related to machining and air
exposure on wettability of different Mediterranean wood species. Int J Adhes Adhes
31: 743-753.
81
Singh, A. P., Anderson, C. R., Warnes, J. M. and Matsumura, J. 2002. The effect of planing
on the microscopic structure of Pinus radiata wood cells in relation to penetration of
PVA glue. Holz Roh Werkst 60: 333-341.
Volk, R. 2005. Rauheitsmessung: Theorie und Praxis, DIN e.V., Berlin, Beuth
Wellons, J. D. 1980. Wettability and Gluability of Douglas-fir Veneer. Forest Prod J
30: 53-55.
82
4.3 Paper III
European Journal of Wood and Wood Products, 2014, Volume 72, Issue 3: 343 - 354
Improvement of tensile shear strength and wood failure percentage of
1C PUR bonded wooden joints at wet stage by means of DMF
priming
Oliver Kläusler a, Philipp Hass a, Carlos Amen b, Sven Schlegel c, Peter Niemz a
a Institute for Building Materials, ETH Zürich, Zürich, Switzerland
b R&D Chemist, Purbond AG, Sempach-Station, Switzerland
c BA Dresden, University of Cooperative Education, Dresden, Germany
Corresponding author’s phone: +41 (44) 632 32 32, fax: +41 (44) 632 11 74,
e-Mail: [email protected], URL: www.ifb.ethz.ch/wood
Abstract
Tensile shear tests according to EN 302-1 for load-bearing timber structures were performed
on European beech wood (Fagus sylvatica L.) and Douglas fir (Pseudotsuga menziesii (Mirb.)
Franco) bonded by means of a one-component polyurethane adhesive (1C PUR). Results
reveal a substantial loss of tensile shear strength (TSS) and wood failure percentage (WFP) in
the wet state compared to the dry stage. As microscopic images display, this is accompanied
by a loss of adhesion at the boundary layer. Therefore the aim of this work was to find basic
approaches for a priming fluid that improves the load transmission between adhesive and
adherend in the wet state without introducing formaldehyde into the gluing process. A
substantial improvement of TSS and WFP was achieved by means of the hygroscopic organic
solvent N,N-Dimethylformamide (DMF). To contribute an explanatory approach contact
83
angle measurement were carried out, revealing that DMF heavily enhances the wettability of
the joining surface. Furthermore it was tried to integrate the outcomes into the swelling strain
model, stated by Frihart in 2009. By way of comparison also the hydroxymethylated
resorcinol (HMR) coupling agent, a mixture of diphenylmethane-4,4'-diisocyanates isomers
(pMDI) and water were tested as priming fluids. The data confirm that TSS and WFP of 1C
PUR bonded wooden joints do not correlate, whilst WFP is mostly not normally (at wet stage
often bimodally) distributed.
Keywords
Wood, Bonding, Adhesive, Polyurethane, Priming, Shear strength, Wood failure, DMF
Verbesserung der Zugscherfestigkeit und des Holzbruchanteils von
1K-PUR verklebtem Holz nach Wasserlagerung durch DMF Priming
Zusammenfassung
Buchenholz (Fagus sylvatica L.) und Douglasienholz (Pseudotsuga menziesii (Mirb.) Franco)
wurden mit einem Einkomponenten-Polyurethanklebstoff (1K PUR) verklebt und
Zugscherprüfungen nach prEN 302-1 (2011) für tragende Holzbauteile unterzogen. Die
Ergebnisse bestätigen einen erheblichen Verlust an Zugscherfestigkeit und Holzbruchanteil an
nassen Prüfkörpern (im Vergleich zu trockenen Prüfkörpern). Wie Mikroskopbilder zeigen,
geht dies mit Adhäsionsversagen einher. Im trockenen Zustand dominieren hingegen
Kohäsionsversagen in der Klebfuge und Holzbruch. Daher wurde in der vorliegenden Arbeit
nach Primern für 1K PUR-Verklebungen gesucht, die die Nass-Adhäsion verbessern. Diese
Primer sollen möglichst kein Formaldehyd in den Verklebungsprozess einbringen. Unter
Verwendung von N,N-Dimethylformamid (DMF) als Primer konnten Zugscherfestigkeit und
Holzbruch erheblich verbessert werden. Die Ergebnisse wurden im Kontext mit dem
„swelling strain model“ von Frihart (2009) diskutiert. Ergänzend zeigen
Kontaktwinkelmessungen eine verbesserte Benetzbarkeit der Fügeflächen nach DMF-
Behandlung. Zusätzlich wurden vergleichend die Primer HMR (hydroxymethyliertes
84
Resorcinol), pMDI (polymeres Diphenylmethandiisocyanat) und Wasser geprüft. Die
Versuchsergebnisse bestätigen, dass Zugscherfestigkeit und Holzbruchanteil nicht
miteinander korrelieren, wobei die Holzbruchanteile mehrheitlich nicht normalverteilt sind
und nach Wasserlagerung oft bimodale Verteilungen aufweisen.
Introduction
Wooden glued joints for load bearing elements, such as glued laminated timber, have to
sufficiently comply with technical standards’ requirements. For one-component polyurethane
(1C PUR) bonded wooden joints the European standard EN 15425 (2008) sets thresholds for
tensile shear strength (TSS) at the dry and wet stages, but not for wood failure percentage
(WFP). Standards like CSA O112.9 (2004) or ASTM D 2559 (2004) are decisive for North
America (NA). They comprise compression shear tests and set various thresholds for shear
strength and for WFP at the dry and wet stages. As a rule of thumb, they demand a WFP of at
least 80 % (median) for hardwoods and 85 % for softwoods, depending on conditioning and
testing conditions. So far 1C PUR bonded joints have passed all the dry stage requirements,
but they have problems overcoming the thresholds for WFP in the wet state (Brandmair et al.
2012). Inter alia Uysal et al. (2006), Lopez-Suevos et al. (2009) and Kläusler et al. (2013)
confirmed a significant reduction of the performance of 1C PUR wood bondings in the wet
state compared to dry stage. Since the use of hardwoods for adhesively bonded structural
elements is an issue of current interest (Schmidt et al. 2010a & b; Strahm 2011; Flüshöh 2012;
Schmidt et al. 2012) also the delamination behavior of 1C PUR bonded beech wood elements
has been investigated. Schmidt et al. (2010b) concluded that the accordant demands of prEN
302-2 for types I and II adhesives can be met using a melamine-urea-formaldehyde polymer
(MUF) with specifically prolonged closed assembly time. The accordant 1C PUR bonded
specimens however did not fulfill the delamination requirements of said standard. In the
1990’s, Vick et al. (1998) developed a hydroxymethylated resorcinol (HMR) coupling agent.
This primer significantly improves the WFP of 1C PUR and MUF glued joints in the wet state
(Vick et al. 2000) and also helps to reduce delamination (Lopez-Suevos et al. 2009;
Ohnesorge et al. 2010). However, it introduces formaldehyde into the gluing process and
requires some laborious process steps (Eisenheld et al. 2005). The detailed mechanism on
which HMR takes effect is not yet completely understood, but several notable efforts have
85
been made on this subject. Gardner et al. (2001) investigated HMR treated wood by means of
contact angle measurements. Their results indicate that the enhanced strength of HMR treated
bonds is not caused by improved wetting of the adherends by the adhesive. The reaction
mechanism between HMR and 1C PUR was investigated by Szczurek et al. (2010). They
proposed that a formation of urethane linkages takes place between methylol groups of the
HMR on the one side and isocyanate groups of the 1C PUR on the other. Son et al. (2004) and
Christiansen (2005) studied the effect of HMR on the wood itself. Their findings indicate that
HMR improves the bonding quality by dimensional stabilization of the wooden substrate,
leading to reduced stress between substrate and adhesive during climatic changes. Son et al.
(2005) investigated the influence of HMR on maple veneer and postulated that this coupling
agent also acts as a lignin plasticizer, generating an interphase which helps to reduce stresses
caused by moisture changes. This finding contrasts with Sun et al. (2005) who reported that
the highly reactive HMR rather stiffens the cell wall, which may be based on a crosslinking
reaction between HMR and lignin. The current work aimed at finding a new basic approach
for the formaldehyde free priming of 1C PUR bonded wood. Therefore the solvent N,N-
dimethylformamide (DMF) was tested in comparison with three other priming fluids. The
latter were water, a mixture of diphenylmethane-4,4'-diisocyanates isomers (pMDI) and the
HMR primer. Water might be one of the simplest “primers” one can think of. The beneficial
effect of water spray on the 1C PUR-gluing results is frequently mentioned in experience
reports from industrial practice. According to Beaud et al. (2006) and Kägi et al. (2006) water
spray is helpful when the ambient conditions are very dry, leading to fast superficial drying of
the wooden adherends. Ashton (1973) proposed improving the adhesion of organic coatings
on wood by means of a physico-chemical wood treatment. His basic approach was to swell
wood in order to make more functional (OH-) groups available to the reagents. A polar fluid
capable of swelling wood to an even higher degree than water is DMF. As Ashton (1973) and
Mantanis et al. (1994 a, b) summarize, it swells wood comparatively fast just by soaking at
room temperature. This hygroscopic and high boiling solvent does not evaporate too quickly,
thus providing some time for interactions with the wood and possibly also with the adhesive
polymer. A quite different basic approach is priming by means of a mixture of
diphenylmethane-4,4'-diisocyanates isomers (pMDI). Such highly functional isocyanates
promote bonding by reacting with polar groups of the wood (Lay and Cranley 2003). Under
ideal conditions pMDI is even capable of bonding covalently to the wood via formation of
urethane linkages (Zhou and Frazier 2001). The hereby modified boundary layer would then
86
represent additional linking points for the adhesive polymer. Gindl W. et al. (2004)
investigated the diffusion of pMDI into cell walls of spruce wood. They concluded that no
pMDI diffuses into the cell walls on a microscopic level and added that this does not exclude
a potential diffusion of pMDI compounds at nanometer scale.
Materials and methods
Wood
Based upon prEN 302-1 (2011) for tensile shear tests, boards of European beech wood (Fagus
sylvatica L.) and Douglas fir (Pseudotsuga menziesii (Mirb.) Franco) were conditioned in
climate 20°C / 65% relative humidity (RH) until equilibrium moisture content (EMC) was
reached. Subsequently the average raw densities of 679 kg/m3 (beech) and 498 kg/m3
(Douglas fir) were determined. All the wood of each species derived from one section of the
same log. Boards with flaws such as a very wavy direction of grain, knots or discolorations
were sorted out. The material was then cut to size and planed conforming to the standard
mentioned above. Prior to any testing the boards were mixed in order to randomly scatter
influences caused by the wood’s inhomogeneity over the whole sampling.
Adhesive
All the bonding procedures were performed using the 1C PUR adhesive HB S 309 (Purbond®
AG, Switzerland), approved for structural bonding of wood in Europe.
Priming liquids
For a concise overview of the used liquids see Table 1. The used DMF (C3H7NO, 73.09 g /
mol) is a polar, high boiling, toxic and hygroscopic solvent, produced by Sigma-Aldrich
(puriss. p.a., ACS reagent, reag. Ph. Eur., ≥ 99.8 % (GC), vapor pressure 2.7 mmHg (20 °C)).
87
Table 1 Priming fluids, applied amounts and waiting times
* Average amount of primer applied per joining surface
The use of highly concentrated toxic DMF in practice might require further safety measures.
Therefore a solution of 5 % DMF was also tested. The pMDI Desmodur® VKS 20 (Bayer
MaterialScience) is a solvent free mixture of diphenylmethane-4,4'-diisocyanates with
isomers and homologues of high functionality (2.9). It contains about 31 % isocyanate and is
preferably used as a hardener component in adhesive systems. The HMR priming fluid was
prepared as described by Lopez-Suevos et al. (2009).
Priming procedure
The adherends underwent the priming process within 30 minutes after planing. For the
applied amounts per joining surface and the accordant waiting times please refer to Table 1.
The HMR priming fluid was applied onto the adherends by means of a paintbrush. For
spraying the deionized water, a standard hand held water-spray bottle was used. Basically
additional moisture accelerates the 1C PUR reaction, but presence of a water film on the
substrate leads to a very sudden reaction, which impedes the proper formation of adhesion
between adhesive and adherend. Therefore, three minutes were allowed for the water to
penetrate and partly evaporate before 1C PUR application. DMF and pMDI were applied onto
a metal sheet using a paintbrush, avoiding a spray mist. Subsequently, the bonding surfaces
were covered with the sheets. This technique provides a more homogeneous liquid spread
than direct brushing onto the wood.
Abbreviated designation
Applied amount [g / m2]
* Waiting time
No primer application Control 0 0
HMR HMR 195 18 h
Water spray Water 20 3 Min
DMF, concentration 5% DMF 5 40 30 Min
DMF, concentration 100% DMF 100 40 30 Min
VKS 20, 30 min. waiting time pMDI 30 30 30 Min
VKS 20, 1 day waiting time pMDI 1d 30 1 day
88
Bonding process and sample manufacturing
After priming, 180 g / m2 1C PUR were applied one-sided using a toothed spatula. The
pressing was performed over 75 minutes at a specific pressure of 0.8 MPa in a calibrated press
by means of a pressing jig. Consequently the pressed parts were again stored in climate 20°C /
65% RH for at least three days in order to assure sufficient hardening of the adhesive before
further processing. The climatized pressings were then cut to tensile shear test samples
according to prEN 302-1 (2011).
Sample treatment, lots and testing procedure
Prior to testing the specimens of each batch were mixed and afterwards divided into different
lots (n = 12) for the treatments depicted in Table 2.
Table 2 Sample treatments and threshold values
In the follow-up to the treatments, tensile shear tests according to the aforementioned standard
were performed on a calibrated universal testing machine. Specimens were subjected to a
constant testing speed of 0.9 mm/min and failed after 30 s to 90 s, in accordance with said
Technical standard
Sample treatments before testing Threshold values
Abbreviated designation Description Tensile shear strength
[MPa] b
EN 15425
EN 302-1 Tensile shear
test
A1 7 days storage at 20°C /
65% RH a 10
A2 4 days water storage at
20°C 6
A4 6h storage in boiling water + 2h submerged at 20°C
6
A5 A4 + reconditioning in
20°C / 65% RH 8
a RH: Relative humidity of ambient air [%] b Adhesive type I with 0.1 mm thickness of adhesive layer
89
standard. At the moment of testing the average moisture contents of the beech specimens were
13.4% (A1), 119.8% (A2), 121.4% (A4), and 14.7% (A5). The evaluation of the WFP on the
fracture surfaces was performed visually on the basis of ASTM D5266 (1999). Since moisture
and temperature affect the performance of the glued joints (Schrödter et al. 2006; Clauß et al.
2010) a run of pre-tests was performed with treatments A1, A2 and A4 in order to be sure
about the more decisive parameter. Whilst A1 functioned as the control batch, A2 and A4
mainly differed from each other regarding temperature sequence. Subsequently the main test-
runs followed, comprising new control batches and primed specimens, supplemented by solid
wood samples. The latter are suitable for an approximate assessment of the wood itself, but
the measured values should be interpreted with caution. Such samples do not have a bondline
and consequently present a different stress distribution during testing. In addition also the
divers arrangement of annual growth rings, wood rays, grain angles, etc. influences the
mechanical properties of the wood (Kollmann 1951; Niemz 1993; Burgert et al. 2001) and
accordingly of the bonded or un-bonded test specimen. Nonetheless, the testing of bonded
samples and solid wood samples appears to be the best feasible way for the comparative and
approximate evaluation of the wooden adherend.
UV-light images
Frequently it is difficult to separate shallow wood failure, adhesion failure and cohesion
failure in the bondline from each other (definitions acc. to ASTM D907-12), especially when
adhesive and wood have almost the same colour. E.g. fracture surfaces with just a few fiber
layers on top of the adhesive layer can falsely look like adhesion failure instead of shallow
wood failure. Particularly in such cases the noted WFP of one and the same sample can vary
quite a bit, depending on the person evaluating it. Advantageously the used adhesive
represents UV-fluorescent markers. Therefore the combination of daylight-images and UV-
light images of the fracture surfaces was used for the assessment of WFP.
ESEM-Images
In order to support the findings obtained by UV-light and daylight photographs further images
of fractured specimens were prepared by means of an Environmental Scanning Electron
Microscope (ESEM). To this end the fractured adherends were reassembled using a reflected
90
light microscope and hereinafter embedded in epoxy resin. A series of about 15 adjacent
ESEM pictures were taken from each of four representative specimens (two samples out of
batch Control A4 and two out of lot DMF 100 A4). The consecutive images (4 x 15) were
joined, thus displaying the complete fracture path (length 20 mm) of each sample. Additional
pictures were taken using energy dispersive X-ray spectroscopy (EDX) to ensure a proper
differentiation between epoxy resin and 1C PUR (images not depicted).
Contact angle measurements
As previous works demonstrated (Wellons 1980; River et al. 1991a; Dunky 2002; Hernandez
et al. 2008 a & b; Kläusler et al. 2013) sufficient wettability of the joining surface is an
important precondition for a good performance of the resulting joints. Therefore the contact
angle sequences of ten water droplets on four freshly planed beech wood joining surfaces
(half rift cut) were measured by means of the static sessile drop method. Two of the surfaces
were DMF 100 primed, the other two remained unprimed. Beforehand the wood was
climatized in standard climate 20°C / 65% RH until EMC was reached. Due to technical
limitations distilled water had to be used instead of the highly viscous 1C PUR. The
measurements were carried out on a dataphysics contact angle system OCA supported by
SCA 20 software. The droplet volume of 12 l was applied 30 minutes after application of the
DMF (in accordance with the waiting time depicted in table 1) and the camera took pictures
of the droplets’ shapes with a frequency of 0.5 Hz (unprimed samples) and 25 Hz (DMF
treated samples) respectively. The change in frequency was necessary due to the very high
wettability of the primed surfaces compared to the non-primed ones. However, due to its
porosity and inhomogeneity wood is a non-ideal surface for contact angle measurements in
principle (Gindl M. et al. 2004; Santoni et al. 2011). Therefore, the measured values should
primarily be interpreted comparatively within the current study.
Results and discussion
The results of the pre-tests reveal that the impact of moisture is much more decisive for the
performance of the 1C PUR bonded joints than the temperature of the water treatment (Fig.
1). Regarding TSS and WFP, no significant differences were detected between treatments A2
and A4. They both show substantial reduction of their values compared to the A1 treatment.
91
Fig. 1 Pre-tests (beech wood): Tensile shear strength and Wood failure percentage of beech
wood specimens. Whiskers: minimum and maximum values; horizontal line: median, square
in box: arithmetic mean
Subsequently for the main test runs with primed samples the A2 batches became obsolete and
were replaced by A5 lots to investigate the bondings after re-drying.
Tensile shear strength of beech wood specimens
The results of TSS tests were firstly evaluated by Boxplots, giving an overview of data
distribution, arithmetic means and medians (Fig. 2).
A1 A2 A4
0
20
40
60
80
100
Woo
d fa
ilure
[%
]
A1 A2 A40
2
4
6
8
10
12
14
16
Ten
sile
she
ar s
tren
gth
[MPa]
92
Fig. 2 Tensile shear strength of beech wood specimens. Whiskers: minimum and maximum
values; horizontal line: median, square in box: arithmetic mean
A Shapiro-Wilk test on normal distribution (suitable for sample sizes 8 ≤ n ≤ 50) was carried
out for the beech wood data (Fig. 2) based on DIN ISO 5479 (2004) at the α = 0.05 level.
Result indicates that 88% of these TSS batches comprise normally distributed data (Table 3).
Solid
woo
d
Contr
olHM
RW
ater
DMF 5
DMF 10
0
pMDI 30
pMDI 1d
Solid
woo
d
Contr
olHM
RW
ater
DMF 5
DMF 10
0
pMDI
30
pMDI
1d
Solid
woo
d
Contr
olHM
RW
ater
DMF 5
DMF 10
0
pMDI 30
pMDI 1d
0
2
4
6
8
10
12
14
16A5A4
Ten
sile
she
ar s
tren
gth
[MP
a]
A1
93
Table 3 Shapiro-Wilk Normality Test
Therefore confidence intervals were taken into consideration for the assessment of average
mean value differences (Fig. 3). The TSS average values of the control samples (A1, A4, and
A5) do not significantly differ from the accordant values of the solid wood samples. Hence,
the 1C PUR bonded samples do not appear to be stronger or weaker than the solid wood itself.
Fig. 3 Average values of tensile shear strength with Confidence intervals at a 95% confidence
level. Beech wood: _B, Douglas fir: _D
0
2
4
6
8
10
12
14
16
Ten
sile
she
ar s
tren
gth
[MP
a]
A1 A4 A5
Solid wood
Control HMR Water DMF
5 DMF 100
pMDI 30
pMDI 1d
Tensile shear
strength [MPa]
A1 x x x x x x x
A4 x x x x x x
A5 x x x x x x x x
Wood failure
[%]
A1
) (
x x x
A4 x
A5 x x x x x
x : At the 0.05 level this data was significantly drawn from a normally distributed population.
)( : Excluded from Normality test
94
But all the batches (incl. solid wood) significantly lose strength from A1 to A4 and regain
strength after re-drying (Fig. 2 and 3). In this respect it should be taken into account that also
the wood itself loses strength when its moisture content rises up to fiber saturation (Kollmann
1951; Niemz 1993). According to River et al. (1991b) the loss of shear strength parallel to the
grain between oven-dry and fiber saturation amounts to about 50%, inter alia depending on
the wood species. The fact that the bonded specimens lose strength at wet stage (A4 compared
to A1) and regain strength due to re-drying (A5) points at the great importance of secondary
bonds between the 1C PUR polymer on the one side and hydroxyl groups of the wood on the
other. Such bonds, in particular hydrogen bonds, are going to be ruptured due to the polar
water molecules entering the interface (at the boundary layer) and the interphase (between
bulk adhesive and bulk wood). The majority of these bonds is going to be re-established as
soon as the water evaporates. In addition swelling and shrinking of the composite material
obviously did neither irreversibly damage the wood nor the adhesive polymer in the bondline.
Otherwise such a regain of TSS (A5 compared to A1) would not be possible. The water and
DMF 5 batches are the only ones revealing a significantly lower TSS after re-drying
compared to A1. The HMR lots show increased scatter of individual values (Fig. 2), whilst
their mean values do not significantly differ from the accordant values of the control batches
(A1, A4, A5). When tested without any previous water contact (A1) the results of the water
spray-batches reveal that this kind of treatment is capable of enhancing the TSS, thus
confirming the industrial experience. However, this treatment is not helpful when the
specimens are tested at the wet or re-dried stage (A4, A5). The A4 batch does not even meet
the requirements of EN 15425 (2008) (Table 2). This also applies for the DMF 5 lots A4 and
A5. On the contrary the DMF 100 batches reveal significant improvements of TSS after A1
and A4 compared to the controls (Fig. 2 and 3). Obviously a sufficient concentration of the
fluid is needed for causing such an effect. Regarding priming with pMDI neither a significant
influence of waiting time nor of the primer itself on TSS was detected (A1, A4, A5 batches of
pMDI 30 and pMDI 1d compared to accordant control batches).
Wood failure percentage of beech wood specimens
Some boxplots for WFP (Fig. 4) exhibit skewed distributions, like e.g. DMF 100 (A4),
showing extreme range (R = Xmax – Xmin = 100%) and a large inter-quartile range (IQR =
Q.75 – Q.25 = 90%). Therefore the Shapiro-Wilk test mentioned above was also carried out
95
for the WFP data and demonstrated that only 43% of the WFP batches reveal normal
distribution (Table 3). The scatter plots (Fig. 5) support these observations and histograms
(not depicted) reveal bimodal distributions for the A4 batches Control, HMR and DMF 100.
The accordant specimens reveal either very high or very low WFP on the same strength level.
Not a single sample with medium WFP (30 – 70%) was found within these A4 batches. These
findings basically go in line with CSA O112.9 (2004), which explains that WFP is rarely
found to be normally distributed. Hence regarding WFP results medians should be given
preference over average values.
Fig. 4 Wood failure percentage of beech wood specimens. Whiskers: minimum and
maximum values; Horizontal line in box: median; square in box: arithmetic means value
The control batches show high WFP after A1, low WFP when tested at wet stage (A4) and
regained WFP after re-drying (Fig. 4). This applies to all the tested batches and goes in line
with Kläusler et al. (2013). The HMR batches reveal high WFP after all the three treatments,
notably the highest median values within the treatment groups A4 and A5 and basically in
agreement with Vick et al. (2000). In contrast, no improving effect of water spray or DMF 5
ControlHM
R
Water
DMF 5
DMF 10
0
pMDI 30
pMDI 1d
ControlHM
R
Water
DMF 5
DMF 10
0
pMDI 30
pMDI 1d
ControlHM
R
Water
DMF 5
DMF 10
0
pMDI 30
pMDI 1d
0
20
40
60
80
100
W
ood failu
re [%
]
A5A4A1
96
on WFP can be found after A4 and A5. But DMF 100 and pMDI 1d reveal a considerable
improvement of WFP (medians compared to control median) after water contact (A4, A5).
Both pMDI batches clearly show reduced WFP after A1 (compared to Control), but higher
medians after A4 and A5. For sure the given wood moisture content after A1 was sufficient
for the pMDI to react (He and Yan 2005). However, the results of the accordant A4 and A5
batches point at a strong influence of the waiting time on WFP after water treatment, which
does not seem to be crucial after A1. In this regard further experiments (e.g. with more
graduations of waiting times) would help to get a clearer picture before further conclusion can
be drawn regarding the influence of pMDI priming on WFP. In summary HMR and DMF 100
turned out to be the only priming liquids capable of enhancing WFP after all three treatments
(A1, A4, A5), whereas HMR (median) is the only one reaching 80% WFP after A4 and even
85% after A5.
Correlation coefficients and Scatter plots regarding beech wood specimens
Calculation of Spearman’s rank correlation coefficient (rSP) was used for evaluation of the
correlation between TSS and WFP of the beech wood batches. This method is suitable for
non-normally distributed lots and was exemplarily performed for the control batches. In
summary, none of these lots revealed a significant correlation between the two parameters at
α = 0.05 level. It is fair to conclude that high (or low) WFP of 1C PUR bonded beech wood
joints does not indicate high (or low) TSS of the 1C PUR bonded composite material (see also
Fig. 5). But nonetheless the assessment of WFP of 1C PUR bonded joints is reasonable.
Especially in case of extremely low strength a high WFP may point at a low wood quality.
97
Fig. 5 Scatter plots. Gray rectangle in diagram: Interquartile range of the TSS measured for
the solid wood samples (beech wood). Values rounded acc. to EN 302-1
Tensile shear strength and WFP of the Douglas fir samples
The finding for the Douglas fir samples (Fig. 6 & 7) basically go in line with those for the
beech wood samples. Confidence intervals (Fig. 3) reveal a significant improvement of TSS
after DMF 100 treatment compared to the accordant controls (A1, A4). After water contact
(A4, A5) also WFP (Fig. 7) benefits from the DMF 100 treatment.
0
20
40
60
80
100
0 5 10 15
WF
P [%
]
TSS [MPa]
Control A1
0
20
40
60
80
100
0 5 10 15
WF
P [%
]
TSS [MPa]
Control A4
0
20
40
60
80
100
0 5 10 15
WF
P [%
]
TSS [MPa]
Control A5
0
20
40
60
80
100
0 5 10 15
WF
P [%
]
TSS [MPa]
HMR A1
0
20
40
60
80
100
0 5 10 15
WF
P [%
]
TSS [MPa]
HMR A4
0
20
40
60
80
100
0 5 10 15
WF
P [%
]
TSS [MPa]
HMR A5
98
Fig. 6 Tensile shear strength of Douglas fir samples. Whiskers: minimum and maximum values; horizontal line: median, square in box: arithmetic mean
Fig. 7 Wood failure percentage of Douglas fir samples
Contr
ol
DMF 5
DMF 10
0
Contr
ol
DMF 5
DMF 10
0
Contr
ol
DMF 5
DMF 10
0
0
2
4
6
8
10
12
14
16A5A4A1
Ten
sile
she
ar s
tren
gth
[MP
a]
Contr
ol
DMF 5
DMF 10
0
Contr
ol
DMF 5
DMF 10
0
Contr
ol
DMF 5
DMF 10
0
0
20
40
60
80
100A5A4A1
Woo
d fa
ilure
[%
]
99
UV light photographs and ESEM Images of beech wood specimens
For several specimens the result of the WFP assessment varied quite a bit, dependent on the
used light source. For example, the specimen depicted in Fig. 8 was rated as 100% WFP at a
first glance by means of daylight. Using UV-light in combination with a reflected light
microscope the value was corrected down to 70% (66% measured and rounded up acc. to
prEN 302-1).
Fig. 8 Fracture surfaces of the beech wood specimen with highest TSS (Control, 13.0 MPa)
after A5. Left: artificial daylight. Right: UV light; dotted frames encircle wood failure
Fig. 9 Fracture surfaces of a beech wood specimen after A4 (Control). Left: artificial
daylight; 10% WFP estimated due to “fibers” on the fracture surfaces. Right: identical
specimen under UV light: 0% WFP detected. Matching shapes (oval - oval, circuit - circuit):
Correspondent spots on the two surfaces, exemplarily disclosing selective loss of adhesion of
the adhesive polymer
In summary, the combination of both light sources plus microscope proved effective. A closer
look at the fracture surfaces of the A4 specimens (Fig. 9) revealed selective detachments of
100
glue from the surfaces of the adherends. In total nine samples of Control batch A4 revealed
0% WFP. Four of them were searched for selective loss of adhesion and it was found on each
of them. Two depicted loss of adhesion on about 50% of the fracture surface. On the contrary,
after A1 or A5 loss of cohesion within the adhesive layer and wood failure were predominant.
Two more specimens of the Control A4 batch were investigated by means of ESEM images
(Fig. 10). Results confirm the finding that fracture surfaces without wood failure of specimens
tested at wet stage (A4) show loss of adhesion at the interface. In principal it is difficult to
draw general conclusions out of a series of 15 ESEM pictures per specimen, representing one
and the same plane within the sample. But nonetheless, the images taken of the fracture path
after A4 (Fig. 10 exemplarily) depict a clear loss of adhesion over the whole width of the
sample (20 mm).
Fig. 10 ESEM image of beech wood specimen (Control) tested on TSS after A4. FP: Fracture
Path (filled with Epoxy resin for preparation); GJ: Former Glued Joint (with residual 1C
PUR) showing no wood failure but fracture between adhesive and adherend; WR: Wood ray
Another two specimens of batch DMF 100 were inspected after treatment A4 (Fig. 11
exemplarily). As the micrographs reveal, the DMF treated bondings are basically capable of
creating deep wood failure (about 200 – 300 m distance between fracture path and glued
joint).
101
Fig. 11 ESEM image of beech wood specimen (DMF 100, A4) showing deep wood failure
after TSS testing. FP: Fracture Path (filled with Epoxy resin for preparation); GJ: Glued Joint
(filled with 1C PUR); WR: Wood rays of the two adherends
Contact angle measurements
Compared to the control samples the DMF treatment heavily reduces the contact angle of the
water on the adherends’ surfaces (Fig. 12), measured 30 minutes after application of the
DMF. It is more than likely that DMF changes the chemical composition of the boundary
layer by influencing the wood extractives, thus affecting the wood’s surface energy and
glueability (Nussbaum 1999; Stehr et al. 2000; Gindl M. et al. 2004).
Fig. 12 Contact angle measurements on DMF 100 treated beech wood
102
Swelling strain model
In search of fundamental explanations for the behavior of adhesively bonded wood under
changing moisture conditions Frihart (2009) stated the swelling strain model. It focusses on
the effect of swelling strain distribution on the failure behavior of wooden glued joints and
recommends establishing the two groups of in situ polymerized adhesives (e.g. phenolic
resins) and pre-polymerized adhesives (e.g. 1C PUR). According to the model the penetration
of in situ polymerized adhesives into the adherends’ cell walls has a stabilizing effect by
reducing the cell walls’ swelling capacity. This promotes higher WFP in the wet state,
because the swelling strain occurs some cell rows away from the joint (Fig. 13, I.) where less
adhesive is present. Also the HMR treatment would basically fit into this group.
Fig. 13 Schematic drawing of glued joints. I. Phenolic resin; II. 1C PUR; III. DMF + 1C PUR; a: glued sample in the dry state; b: glued sample at wet stage with fracture path (wavy black line) after tensile shear test; cuboid in the center: bondline; dotted areas: presence of adhesive polymer; dashed areas: Presence of DMF
On the contrary pre-polymerized adhesives do not penetrate the cell walls (Fig. 13, II.).
Therefore the swelling strain occurs at the interface, hereby advantaging a fracture path with
low WFP. Based on the current results the model could be extended by a third variation. By
using DMF as adhesion promoter for 1C PUR bonded wooden joints (Fig. 13, III.) a highly
pre-swollen state of the adherends is created. It is likely that after pressing some of the high
boiling solvent slowly evaporates (especially at the edges of the specimens) and some of it
remains in the wood, thus establishing a prolonged pre-swollen state. During A4 treatment
II.a II.b III.aI.b I.a III.b
103
this pre-swollen state might cause a shift of swelling strain away from the interface deeper
into the bulk adherends where less DMF is present. According to the swelling strain model
such a shift promotes higher WFP at wet stage. In the current research a period of about two
weeks “evaporation time” elapsed between manufacturing of the pressings and A4 treatment.
Obviously the prolonged high-grade pre-swelling by means of highly concentrated DMF has a
different effect on WFP than the short-term pre-swelling by means of water spray or DMF 5
(very high water content). The water does not swell the wood to such a high extend and
vapors off until EMC of climate 20°C / 65% RH (storage climate for re-drying) is reached.
These observations basically go in line with the swelling strain model. However, further
investigations and a more detailed knowledge about the penetration of pMDI into wooden cell
walls are needed, before this kind of priming can clearly be classified in the model.
Conclusions
Within the present work DMF 100 is the only adhesion promoter which significantly
enhances the TSS of 1C PUR bonded beech wood joints after A1 and after A4. Regarding
WFP the highly concentrated DMF improves the measured medians after all three treatments
(A1, A4 and A5). The additional tests on Douglas fir confirm said results for TSS and WFP.
No significant effect of pMDI priming on TSS was observed and further experiments are
needed to get a clear picture regarding the influence of this priming on WFP. Water spray
improves TSS in the dry state (A1) but not after water contact (A4, A5). Furthermore it does
not improve the WFP of 1C PUR bonded joints. The HMR primer does not reveal a
significant effect on TSS (A1, A4, and A5) but it substantially enhances WFP after all three
treatments.
As UV-light micrographs and ESEM images depict, the loss of performance in the wet state
(Control A4 compared to A1) is accompanied by a loss of adhesion between adhesive
polymer and the wooden adherends. The results after A5 for TSS indicate that this is a
reversible effect, demonstrating the high importance of secondary bonds like hydrogen bonds
for 1C PUR glued joints. Actually the adhesion at wet stage was clearly improved by means
of different priming liquids (HMR, DMF 100), resulting in a higher performance at wet stage
(compared to Control). The UV tracers present in the quasi transparent adhesive turned out to
be quite helpful for the assessment of WFP, which was carried out using a combination of
104
UV-light and artificial daylight. Furthermore the findings confirm that TSS and WFP of 1C
PUR bonded joints do not correlate. The two parameters reveal very different distributions.
Low WFP of 1C PUR bonded beech wood joints does not indicate low TSS of the bonded
composite material and vice versa. It is reasonable to consider this aspect when it comes to the
discussion, whether high WFP of 1C PUR bonded joints can serve as indicator for very high
strength. The present work does not intend to recommend a specific fluid for priming in
practice. Instead it aims to contribute to the fundamental understanding of the mechanisms of
action of such adhesion promoters. Priority attention is being paid to the accordant effect of
the solvent DMF. Sure enough the results presented cannot exhaustively explain the measured
effects of DMF on 1C PUR bonded wooden joints. It is likely that the reasons for the changes
in TSS and WFP are a combination of different influencing factors, such as enhanced
wettability of the bonding surface (likely contributing to the improved adhesion at wet stage),
translocated swelling strain (see swelling strain model) and others. Therefore a subsequent
paper is in preparation, dealing with the influence of DMF on the 1C PUR adhesive polymer
on the one hand and the beech wood on the other. Further investigations could be carried out
regarding possible alternative substances like the less toxic Dimethylacetamid (DMAC), but
also regarding the influence of suitable solvents on the delamination behavior of 1C PUR
bonded joints or the influence of such solvents on the bonding performance of wood species
which so far are difficult to be bonded by means of 1C PUR (e.g. Larix spp.). In addition the
translocation of swelling strain mentioned above could be investigated using Digital Image
Correlation or Speckle Interferometry (Valla et al. 2011; Keunecke et al. 2012).
105
Acknowledgements
The authors would like to thank Dr. Christian Wamprecht (Bayer MaterialScience, Germany)
for his valuable contribution and gratefully acknowledge the financial support of this research
by the Commission for Technology and Innovation KTI/CTI (Bern, Switzerland) and the
Purbond AG (Sempach-Station, Switzerland). Mrs. Gabriele Peschke (IfB, ETH Zürich) is
cordially thanked for preparing the ESEM images.
References
ASTM D2559-04 Standard Specification for Adhesives for Structural Laminated Wood
Products for Use Under Exterior (Wet Use) Exposure Conditions. ASTM International, West
Conshohocken, PA, USA
ASTM D907-12 Standard Terminology of Adhesives. ASTM International, West
Conshohocken, PA, USA
ASTM D5266-99 Standard Practice for Estimating the Percentage of Wood Failure in Bonded
Joints. ASTM International, West Conshohocken, PA, USA
Ashton HE (1973) The Swelling of Wood in Polar Organic Solvents. National Research
Council Canada. Division of Building Research
Beaud F, Niemz P, Pizzi A (2006) Structure-property relationships in one-component
polyurethane adhesives for wood: Sensitivity to low moisture content. J Appl Polym Sci. 101
(6):4181-4192.
Brandmair A, Jans N, Clauss S, Hass P, Niemz P (2012) Bonding of hardwoods with 1C PUR
adhesives for timber construction. Bauphysik 34 (5):210-216. doi:10.1002/bapi.201200025
Burgert I, Eckstein D (2001) The tensile strength of isolated wood rays of beech (Fagus
sylvatica L.) and its significance for the biomechanics of living trees. Trees-Struct Funct 15
(3):168-170. doi:10.1007/s004680000086
Christiansen AW (2005) Chemical and mechanical aspects of HMR primer in relationship to
wood bonding. Forest Prod J 55 (11):73-78
106
Clauß S, Joscak M, Niemz P (2010) Thermal stability of glued wood joints measured by shear
tests. Eur J Wood Wood Prod 69 (1):101-111. doi:10.1007/s00107-010-0411-4
CSA O112.9-04 Evaluation of adhesives for structural wood products (exterior exposure).
CSA Group, Toronto, ON, Canada
DIN ISO 5479 Statistical interpretation of data - Test for departure from the normal
distribution. Deutsches Institut für Normung, Beuth, Berlin, Germany
Dunky M (2002) Teil III Einflussgrössen, Verleimungsrelevante Eigenschaften von
Holzoberflächen. In: Dunky M, Niemz P (eds) Holzwerkstoffe und Leime. Springer, Berlin,
Germany
EN 15425 (2008) Adhesives - One component polyurethane for load bearing timber structures
- Classification and performance requirements. European Committee for Standardization,
Brussels, Belgium
Eisenheld L, Gardner DJ (2005) Developing an industrial friendly process for hydroxymethyl
resorcinol (HMR) priming of wood using a novolak-based HMR. Forest Prod J 55 (9):61-66
Flüshöh F (2012) Von der Nadel zum Laub. Eiche-BSH. Bauen mit Holz (12):48-51
Frihart CR (2009) Adhesive Groups and How They Relate to the Durability of Bonded Wood.
J Adhes Sci Technol 23 (4):601-617. doi:10.1163/156856108x379137
Gardner D, Tze W, Shi S (2001) Adhesive Wettability of hydroxymethyl resorcinol (HMR)
treated wood. In: International Symposium on Wood Adhesives, S Lake Tahoe, 2000. Forest
Products Research SOC Madison, WI, USA 312-327
Gindl W, Schöberl T, Jeronimidis G (2004) The interphase in phenol-formaldehyde and
polymeric methylene di-phenyl-di-isocyanate glue lines in wood. Int J Adh Adh (24): 279-286
Gindl M, Reiterer A, Sinn G, Stanzl-Tschegg SE (2004) Effects of surface ageing on
wettability, surface chemistry, and adhesion of wood. Holz Roh Werkst 62 (4):273-280.
doi:10.1007/s00107-004-0471-4
He G, Yan N (2005) Effect of moisture content on curing kinetics of pMDI resin and wood
mixtures. Int J Adh Adh (25): 450-455
107
Hernandez RE, Cool J (2008a) Effects of cutting parameters on surface quality of paper
birchwood machined across the grain with two planing techniques. Holz Roh Werkst 66
(2):147-154. doi:10.1007/s00107-007-0222-4
Hernandez RE, Cool J (2008b) Evaluation of three surfacing methods on paper birch wood in
relation to water- and solvent-borne coating performance. Wood Fiber Sci 40 (3):459-469
Kägi A, Niemz P, Mandallaz D (2006) Einfluss der Holzfeuchte und ausgewählter
technologischer Parameter auf die Verklebung mit 1K-PUR Klebstoffen unter extremen
klimatischen Bedingungen. Eur J Wood Prod 64 (4):261-268.
doi:10.1007/s00107-005-0088-2
Keunecke D, Novosseletz K, Lanvermann C, Mannes D, Niemz P (2012) Combination of X-
ray and digital image correlation for the analysis of moisture-induced strain in wood:
opportunities and challenges. Eur J Wood Prod 70 (407 413)
Kläusler O, Rehm K, Elstermann F, Niemz P (2013) Influence of wood machining on tensile
shear strength and percentage wood failure of 1C PUR bonded wooden glued joints after
wetting. IWP. doi:10.1179/2042645313y.0000000039
Kollmann F (1951) Elastizität und Festigkeit des Holzes. In: Technologie des Holzes und der
Holzwerkstoffe. Erster Band. 2 edn, Springer, Berlin, Germany
Lay DG, Cranley P (2003) Polyurethane Adhesives. In: Pizzi A, Mittal KL (eds) Handbook of
Adhesive Technology. 2 edn, Marcel Dekker, New York, USA
Lopez-Suevos F, Richter K (2009) Hydroxymethylated Resorcinol (HMR) and Novolak-
Based HMR (n-HMR) Primers to Enhance Bond Durability of Eucalyptus globulus Glulams.
J Adhes Sci Technol 23 (15):1925-1937. doi:10.1163/016942409x12508517390914
Mantanis GI, Young RA, M. RR (1994a) Swelling of wood, Part II. Swelling in Organic
Liquids. Holzforschung 48 (6):480-490
Mantanis GI, Young RA, Rowell RM (1994b) Swelling of wood Part 1. Swelling in water.
Wood Sci Technol 28:119-134
Niemz P (1993) Wirkung wesentlicher Einflussfaktoren der elastomechanischen und
rheologischen Eigenschaften. In: Physik des Holzes und der Holzwerkstoffe. DRW,
Leinfelden-Echterdingen:139-146
108
Nussbaum RM (1999) Natural surface inactivation of Scots pine and Norway spruce
evaluated by contact angle measurements. Holz Roh Werkst 57 (6):419-424. doi:DOI
10.1007/s001070050067
Ohnesorge D, Richter K, Becker G (2010) Influence of wood properties and bonding
parameters on bond durability of European Beech (Fagus sylvatica L.) glulams. Ann For Sci
67 (6):601p601-601p610. doi:10.1051/forest/2010002
prEN 302-2 (2011) Adhesives for load-bearing timber structures – Test methods – Part 2:
Determination of resistance to delamination. Comité Européen de Normalisation CEN,
Brüssel, Belgium
prEN 302-1 (2011) Adhesives for load-bearing timber structures - Test methods - Part 1:
Determination of longitudinal tensile shear strength. Comité Européen de Normalisation
CEN, Brüssel, Belgium
River BH, Vick CB, Gillespie RH (1991a) Chapter III Wood and fibre surfaces. In: Minford
JD (ed) Wood as an Adherend. Treatise on Adhesion and Adhesives, vol 7. Marcel Dekker,
New York, USA
River BH, Vick CB, Gillespie RH (1991b) Chapter II Wood Characteristics Influencing the
Bonding Process and Bond Quality. In: Minford JD (ed) Wood as an Adherend. Treatise on
Adhesion and Adhesives, Vol 7. Marcel Dekker, Inc., New York, USA
Santoni I, Pizzo B (2011) Effect of surface conditions related to machining and air exposure
on wettability of different Mediterranean wood species. Int J Adhes Adhes 31 (7):743-753.
doi:10.1016/j.ijadhadh.2011.07.002
Schmidt M, Frese M, Blass HJ, Glos P (2010a) Bauaufsichtlich zugelassen. Hochfestes
Brettschichtholz aus Buche. Holz-Zentralblatt 136 (4):98-99
Schmidt M, Glos P, Wegener G (2010b) Verklebung von Buchenholz für tragende
Holzbauteile. Eur J Wood Prod 68 (1):43-57. doi:10.1007/s00107-009-0382-5
Schmidt M, Torno S, Knorz M (2012) Laubholz im Bauwesen - ein aktueller Überblick.
Bauen mit Holz (12):34-38
Schrödter A, Niemz P (2006) Untersuchungen zum Versagensverhalten von Klebfugen bei
erhöhter Temperatur und Luftfeuchte. Holztechnologie 47 (1):24-32
109
Son J, Gardner DJ (2004) Dimensional stability measurements of thin wood veneers using the
Wilhelmy plate technique. Wood Fiber Sci 36 (1):98-106
Son J, Tze WTY, Gardner DJ (2005) Thermal behavior of hydroxymethylated resorcinol
(HMR)-treated maple veneer. Wood Fiber Sci 37 (2):220-231
Stehr M, Johansson I (2000) Weak boundary layers on wood surfaces. J Adhes Sci Technol
14 (10):1211-1224. doi:10.1163/156856100742168
Strahm T (2011) Hochfestes Brettschichtholz aus Laubholz und mit hybridem Trägeraufbau:
BSH GL 60 Träger. Bulletin Holzforschung Schweiz (1):7
Sun NJ, Frazier CE (2005) Probing the hydroxymethylated resorcinol coupling mechanism
with stress relaxation. Wood Fiber Sci 37 (4):673-681
Szczurek A, Pizzi A, Delmotte L, Celzard A (2010) Reaction Mechanism of
Hydroxymethylated Resorcinol Adhesion Promoter in Polyurethane Adhesives for Wood
Bonding. J ADHES SCI TECHNOL 24 (8-10):1577-1582. doi:10.1163/016942410x500954
Uysal B, Özçifçi A (2006) Bond strength and durabilty behavior of polyurethane-based
Desmodur-VTKA adhesives used for building materials after being exposed to water-
resistance test. J APPL POLYM SCI 100 (5):3943-3947. doi:10.1002/app.22899
Valla A, Konnerth J, Keunecke D, Niemz P, Müller U, Gindl W (2011) Comparison of two
optical methods for contactless, full field and highly sensitive in-plane deformation
measurements using the example of plywood. Wood Sci Technol 45: 755-765
Vick CB, Okkonen EA (1998) Strength and durability of one-part polyurethane adhesive
bonds to wood. Forest Prod J 48 (11-12):71-76
Vick CB, Okkonen EA (2000) Durability of one-part polyurethane bonds to wood improved
by HMR coupling agent. Forest Prod J 50 (10):69-75
Wellons JD (1980) Wettability and Gluability of Douglas-fir Veneer. Forest Prod J 30 (7):53-
55
Zouh X, Frazier EF (2001) Double labeled isocyanate resins for the solid-state NMR
detection of urethane linkages to wood. Int J Adh Adh (21): 259-264
110
4.4 Paper IV
International Journal of Adhesives and Adhesion 55 (2014) 69 - 76
Influence of N,N- dimethylformamide on one-component moisture-
curing polyurethane wood adhesives
Oliver Kläusler a, Wilhelm Bergmeier b, Alexander Karbach b, Walter Meckel c,
Eduard Mayer c, Sebastian Clauß d, Peter Niemz a
a Institute for Building Materials, ETH Zurich, Schafmattstrasse 6, 8093 Zürich, Switzerland
b Currenta GmbH & Co. Ohg, Kaiser-Wilhelm-Allee 1, 51373 Leverkusen, Germany
c Bayer MaterialScience, 51373 Leverkusen, Germany
d Daimler AG, 71059 Sindelfingen, Germany
a corresponding author’s phone: +41 (44) 632 32 32, fax: +41 (44) 632 11 74,
e-mail: [email protected]
Abstract
The influence of the solvent N,N- dimethylformamide (DMF) on one-component moisture-
curing polyurethane (1C-PUR) bonded wooden specimens was investigated. The applied
methods were ATR-IR spectroscopy, VIS-NIR spectroscopy, FT-IR microscopy, UV
Fluorescence microscopy, Nanoindentation and AFM imaging. Findings reveal that DMF
influences the curing kinetics of the 1C-PUR, supporting its thorough conversion by affecting
the morphology and mechanical properties of the cured adhesive. Furthermore, DMF partially
dissolves the wood’s lignin and seems to make bound wood moisture more available for
reaction with the adhesive. The outcomes confirm that 1C-PUR enters neither the wood cell
walls nor the wood rays. The vessel system of the early wood represents the major pathway
for the penetration of 1C-PUR into the beech wood.
111
KEYWORDS
A: adhesives for wood; A: polyurethane; A: primers and coupling agents; B: wood and wood
composites; not listed: DMF
1. Introduction
N,N-dimethylformamide (DMF) is an organic, polar, toxic and hygroscopic solvent. It is used
in polyurethane synthesis, because it promotes sufficient solubility of reactants such as
diisocyanate and catalysts [1]. This contributes to a homogeneous reaction and to
comparatively high yields in polyurethane production. When used as a primer, DMF is
capable of improving wood failure percentage (WFP) and tensile shear strength (TSS) of
moisture-curing one-component polyurethane (1C-PUR) bonded wooden joints in the dry and
wet states [2]. Within this work, it was also found that DMF greatly enhances the wettability
of the wooden adherend, thus supporting the joints’ performance. Another explanatory
approach for the effect of DMF on the performance of such joints was given by implementing
the outcomes into the Swelling Strain Model described by Frihart [3]. However, various
questions are still open. Therefore the current work intends to contribute to the basic
understanding on how DMF affects the 1C-PUR adhesive (fluid or hardened) and the wooden
substrate.
The most reactive compounds of 1C-PUR wood adhesives are isocyanates, such as polymeric
diphenyl methane diisocyanate (pMDI). The reaction of DMF with arylisocyanates at boiling
temperature was investigated by Weiner [4] and Jovtscheff [5]. They found that aromatic
isocyanates like phenylisocyanate or naphthylisocyanate react with DMF, emitting carbon
dioxide and forming dimethylformamidine. However, DMF did not react with aliphatic
isocyanates under these conditions, demonstrating that reactions between DMF and
isocyanates are highly dependent on the isocyanates’ molecular structure. Joel et al. [6] also
dealt with reactions between phenylisocyanate and DMF. They proclaimed that at 20 °C and
in the presence of water the isocyanate is initially hydrolysed during the first 24 h, leading to
the formation of diphenylurea. However, no conversion of the aryl isocyanate with the solvent
could be found up to 120 h reaction time at 20 °C, but after about 48 h reaction time the
isocyanate and the urea slowly started to convert into triphenylbiuret. In the presence of
112
polyurethane catalysts, substantial amounts of these biuret structures were already found after
24 h at 20 °C. A temperature rise to above 60 °C was required before the phenylisocyanate
and the DMF reacted with each other without catalysts, thereby liberating CO2 gas and
forming phenylformamidine. This is consistent with reactions described by [4] and [5]. As
Ulrich [7] summarized, the interactions between polar solvents (or their impurities) and
diisocyanate at elevated temperatures can also cause undesired side reactions in PUR
production, leading to chain termination or undesired additional crosslinking. Both reactions
disturb the homogeneous molecular weight distribution in the final polyurethane. Ulrich
assumes that the lower the temperature is during polymerization, the less undesired side
reactions occur. Summing up, no evidence in the literature was found for any substantial
direct reactions between DMF and the isocyanates present in 1C-PUR adhesives at 20 °C in
the presence of (wood-) moisture. Nonetheless, DMF is basically capable of influencing the
morphology of PUR polymers after curing, which is characterized by content and distribution
of dispersed hard segments in the soft segment matrix. As Oprea [8] concluded, such a
variation in morphology is accompanied by an alteration of hydrogen bond interactions within
the hard segments and between hard and soft segments. He found accordant changes in tensile
strength and Young’s modulus by means of tensile tests. In addition to its influence on the
polyurethane, the solvent DMF also interacts with wood. Ashton [9] described the fast
swelling of hardwood after immersion in DMF at room temperature and concluded that DMF
has even more swelling potential than water. This corresponds to Mantanis et al. [10], who
added that the comparatively low molecular weight and the high basicity of this amide
support its wood swelling potential. Based on results by Nayer [11], Mantanis et al. [12] also
stress the DMF’s high capability to establish hydrogen bonds to the OH groups of the
cellulose, thus enhancing the wood’s swelling. However, DMF does not dissolve the cellulose
itself, not even after prolonged heating [13]. Therefore, DMF was used for the Soxhlet
extraction of wood chips, leading to polar extracts [14]. The present work investigates the
influence of DMF on the curing process at 20 °C and the resulting properties of 1C-PUR
bonded wooden joints.
113
2. Materials and methods
2.1. Adhesive
The commercial 1C-PUR wood adhesive HB S 309 (Purbond AG, Switzerland) was used for
all the experiments (Viscosity Brookfield ca. 24’000 mPa s). It is approved for glulam
production in Europe.
2.2. DMF and acetone
The DMF (C3H7NO, MW 73.09 g/mol) used for the current work is a high boiling, polar
aprotic, hygroscopic and toxic organic solvent with a comparatively high dipole moment of
3.24 Debye [9, 15, 16] produced by Sigma-Aldrich, puriss. p.a., ACS reagent, reag. Ph. Eur.,
≥ 99.8 % (GC), vapor pressure at 2.7 mmHg (20 °C). The acetone used (puriss. ≥ 99.9 %) was
produced by Azelis GmbH, Germany.
2.3. Wood
The diffuse porous European beech wood (Fagus sylvatica L.) with a measured average raw
density of 0.67 g / cm3 was used as the adherend for the bonding processes. Before further
experimental processing the wood was cut into slats of 8 mm thickness and stored in climate
20 °C / 65 % relative humidity (RH) for at least 2 weeks, resulting in a 12.5 % average wood
moisture content. Slats with discolorations, wavy direction of grain or other flaws were sorted
out. Consequently the slats were mixed to randomly scatter influences of the wood over the
whole sampling.
2.4. Manufacturing of bonded samples
Within one hour after planing the conditioned slats underwent the gluing procedure according
to [17]. An amount of 180 g / m2 1C-PUR adhesive was applied (to one side) by means of a
toothed spatula. Consecutively the pressing was performed over 75 minutes at a specific
pressure of 0.8 MPa using a jig in a calibrated press. Afterwards the pressed parts were stored
at 20 °C / 65 % RH for at least three days in order to assure sufficient hardening of the
adhesive. DMF was used as a primer on some specimens before application of the 1C-PUR.
114
The accordant adherends underwent the priming process within 30 minutes after planing. 40 g
/ m2 DMF was applied onto a metal sheet using a paintbrush, avoiding a spray mist.
Subsequently, the surfaces to be bonded were covered with the sheets. This technique
provided a more homogeneous liquid spread than direct brushing onto the wood. The
adhesive was applied 30 minutes after the DMF treatment.
2.5. IR-ATR investigation on fluid 1C-PUR droplets
Recent studies used FT-IR and IR-ATR for analysis of adhesive droplets and reaction kinetics
[18-20], since this method is accurate, swift and comparatively easy to perform. In the current
study, spectra of the 1C PUR adhesive droplets (Fig. 1 – 3, 5) were collected whilst hardening
by means of a Nicolet Avatar 320 FT-IR-ATR spectrometer with a resolution of about 4 µm�
in combination with Omnic software. The device was equipped with a diamond crystal,
leading to an average light penetration depth of ca. 5 µm. The measured bands of the highly
reactive isocyanate (NCO) groups served as a reference for the progress of the hardening
reaction in accordance with [19]. In addition to the 1C-PUR adhesive, three mixtures were
prepared in total (1C-PUR : DMF = 10 : 1 and 10 : 5; 1C-PUR : acetone = 10 : 5). A droplet
of each fluid was placed on top of the ATR crystal before measurements commenced for a
duration of 24 h. During the measurements, the specimens were exposed to the ambient
laboratory climate of 22 °C and 35 % RH (average values). Each spectrum displayed in Figs.
1 to 4 represents each average set of scans. Each comprises of 10 minutes of scanning,
totalling 394 scans. The extinction was normalized with the absorption peak of the NCO.
2.6. UV-vis-NIR spectroscopy and ATR investigation on wood-DMF extract
214 g of beech wood chips were extracted in 10 mL DMF over a time span of 24 h at 22 °C in
a closed glass jar. Hereafter, the extract was put into a glass cuvette and spectroscopic
investigations on the extract were performed using a Jasco V-670 UV-VIS-NIR
Spectrophotometer (Fig. 7). Consequently the extract was placed onto a heated potassium
bromide (KBr) disk (transparent to IR-radiation) and evaporated to dryness over 2 h in order
to remove the DMF. Afterwards the dark-brown solid extract was investigated by means of a
Nicolet 5700 IR ATR spectrometer with a germanium crystal, providing about 1 µm
penetration depth (spectra displayed in Fig. 8 represents the average of 300 scans).
115
2.7. FT-IR microscopy with chemical imaging
The penetration of 1C-PUR and DMF into the wood substrate was also a matter of interest in
this investigation (Figs. 9 -11). Therefore cross-sections of the adhesively bonded slats were
prepared using a diamond knife microtome. Afterwards the spectroscopic characterization of
the material was performed on a PerkinElmer Spotlight 400 FT-IR Microscope, equipped with
a germanium ATR crystal. This experimental setup permits chemical imaging with a
resolution between one and two micrometres.
2.8. UV-Fluorescence-Microscopy
The 1C-PUR incorporates a UV-marker additive as a standard to support the quality control
of the industrial gluing process. Therefore a reflected light microscope with a UV-light source
was used to investigate the penetration of the adhesive into the adherends (Fig. 12).
2.9. Nanoindentation and AFM images
Wimmer et al. [20] were among the first to successfully use nanoindentation (NI) for the
study of plant cell wall mechanics. In the present paper NI of wood cell walls and bondline
was performed using a DI Atomic Force Microscope (AFM) Nanoscope Dimension 3100,
equipped with a Hysitron nanoindenter head (conical diamond tip). Elastic modulus and
hardness were evaluated from the NI load-depth curves.
3. Results and discussion
3.1. IR-ATR investigation on fluid 1C-PUR droplets
The blue spectra (Figs. 1 – 3, 5) represent average extinction of the first scanning loop
(droplet still in liquid phase). The red spectra were collected after 24 h from the cured
polymer. The green spectra represent the scanning loops performed in the meantime.
116
FIGURE 1: ATR spectrum of a fluid 1C-PUR droplet during hardening. Blue spectrum collected after 10 minutes, red spectrum collected after 24h
At the beginning of the measurements the IR-ATR spectra display a clear extinction near
wave number 2275 cm-1, which is ascribed to the NCO groups of the adhesive [21]. This
peak declines over time, indicating the hardening process of the polymer due to the reaction
of NCO with ambient moisture. However, other NCO contents remain in the droplets after 24
h. In the 1C-PUR droplet (Fig. 1) a residual NCO content of about 4 % remains in the cured
polymer. This is in agreement with [19]. In this context it should be considered that
unconverted NCO might be capable of reducing the cohesive strength of polyurethanes [22].
After curing, the urea groups in the polyurethane polymer appear as a combination of
different bands at about 3335 cm-1 (N-H stretching, hydrogen bonded), 1609 cm-1 (C=C
stretching in aromatic ring), 1600 cm-1 (N-H bending in urethane), 1530 cm-1 (C-N
stretching in urethane group), 1510 cm-1 (N-H bending in urethane group) and 1100 cm-1 (C-
O-C stretching in aliphatic polyether) [19, 23]. The bands around 2800 – 3000 cm-1 display
aliphatic hydrocarbons [21]. Urea groups reveal carbonyl stretching between 1690 cm-1 and
1700 cm-1. However, all peaks shift within a range of about 10 – 50 cm-1, depending on
hydrogen bonds and molecular structure, inter alia [24].
Figure 2 depicts the spectra of a 1C-PUR droplet mixed with DMF at a ratio of 10 : 1. After
completion of the hardening (red spectrum) there is still some unconverted NCO detectable,
117
but compared to Fig. 1 less NCO (about 3 %) remains unconverted after 24 h. The DMF
bands appear around wave numbers 1700 cm-1, 1400 cm-1, 1100 cm-1 and 660 cm-1 [21].
FIGURE 2: ATR spectrum of a fluid droplet during hardening. Droplet mixture 1C PUR : DMF = 10 : 1. Blue spectrum collected after 10 minutes, red spectrum collected after 24h
The occurrence of side reactions, which may eventually be promoted due to the presence of
DMF (e.g. leading to allophanate linkages), is unlikely due to the presence of moisture and
the comparatively low temperature of about 20 °C [7, 25]. However, the bands around 3335
cm-1 reveal the formation of urea.
Figure 3 displays the accordant spectra for a droplet mixture PUR : DMF = 10 : 5. As the
bands for NCO reveal, the conversion speed increased substantially (compared to Fig. 2) by
addition of a higher ratio of DMF to the 1C-PUR. The NCO peak starts at extinction 0.8 after
10 minutes (blue graph) and disappears during hardening. Urea formation is also detectable.
118
FIGURE 3: ATR spectra of a fluid droplet whilst hardening. Droplet mixture 1C PUR : DMF = 10 : 5. Blue spectrum collected after 10 minutes, red spectrum collected after 24h
Thus, it can be concluded, that the speed and the final degree of conversion are both
influenced by the addition of DMF (Figs. 1 - 3): The more DMF that is added to the droplet,
the faster and the more thoroughly the NCO reaction takes place.
Figure 4 displays the normalized NCO extinction of the droplets over time, based on the
spectra displayed in Figs. 1 – 3, 5 (100% = average spectrum after first 10 minutes reaction
time, followed by average reading points every 10 minutes).
119
FIGURE 4: Normalized extinctions of solvents (DMF, acetone) and of NCO groups in fluid
droplets over time (derived from the NCO peak sequences in Figs. 1 to 4)
After 14 hours the 1C-PUR droplet still displayed about 20% of its initial NCO extinction.
The droplet containing DMF (1C PUR : DMF = 10 : 5) passed this value after about 30
minutes reaction time and its NCO extinction disappeared within the first hour. In addition, a
droplet of DMF and one of acetone were placed onto the ATR and their extinctions were
monitored during evaporation. As expected, the evaporation of the high boiling DMF takes
place much slower than that of the low boiling acetone. This affects the conversion of a 1C-
PUR droplet after mixing with the solvents, especially within the first hour of curing.
Therefore the effect of acceleration of the 1C-PUR conversion is much more pronounced in a
droplet with DMF, compared to a droplet with acetone (Fig. 5). After a period of time,
however, the graph of the 1C-PUR : acetone droplet runs approximately parallel to the graph
of the 1C-PUR.
0
20
40
60
80
100
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Norm
alized
extinction [%]
Time [hours]
NCO (1C PUR)
NCO (1C PUR : DMF = 10 : 5)
NCO (1C PUR : Acetone = 10 : 5)
DMF
Acetone
120
FIGURE 5: ATR spectra of a fluid polymer droplet during hardening. Droplet mixture 1C-
PUR : acetone = 10 : 5. Blue spectrum collected after 10 minutes, red spectrum collected after
24h
According to [19], the kinetics of the very initial part of the polyurethane reaction is
chemically controlled, but the higher the conversion degree rises, the more the reaction
becomes diffusion controlled, generating greater crosslink density and consequently leading
to decreased molecular mobility. It is reasonable to assume that the hygroscopic DMF
promotes the diffusion of moisture into the centre of the droplet, thus leading to a higher final
degree of NCO conversion at the end of the curing process. This is supported by a reduction
of viscosity caused by the solvent (temporarily increased molecular mobility). The findings
displayed in Fig. 6 support this hypothesis.
121
FIGURE 6: Subtraction of ATR-spectra. Green: spectrum of cured droplet of 1C PUR : DMF = 10 : 5; black: green spectrum minus blue spectrum; red: spectrum of cured droplet of 1C-PUR; blue: spectrum of DMF
After subtraction of the DMF spectra (blue) from the spectra of the cured mixed droplet
(green), the resulting spectrum (black) does not reveal substantial deviations from the
spectrum of the cured original 1C-PUR (red), with the exception of the NCO extinction
discussed above. These findings indicate that the influence of DMF on the curing kinetics of
1C-PUR is mainly based on physical effects, such as promoted diffusion, rather than on
fundamental changes of the chemical reaction itself. No indication of substantial changes in
the chemical composition of the curing product by DMF was found based on the spectra
displayed in Fig. 6.
The curing process of 1C-PURs starts at the surface of the droplet due to the presence of
ambient moisture, leading to the formation of a urea-coating on the polymer surface. Such a
cured 1C-PUR layer represents a diffusion barrier [26], slowing down the diffusion of
moisture into the uncured polymer in the centre of the droplet. This top layer becomes thicker
and thicker over time, thus enlarging the diffusion barrier and possibly even prohibiting the
1C-PUR in the centre from complete conversion. Daniel-Da-Silva et al. [19] studied this
effect by means of FTIR. Their results confirm that unreacted NCO remains in the
polyurethane polymer after curing. The addition of amine catalysts was needed for the
accordant NCO peaks to disappear. It is reasonable to assume that the presence of DMF in the
122
urea skin reduces its barrier potential and enhances the moisture diffusion.
3.2. UV-VIS-NIR spectroscopy and ATR investigation of DMF wood extract
The UV-VIS Spectroscopy (Fig. 7, vis spectrum 400 – 800 nm) depicts the colour difference
between the extract (amber colour) and the DMF (colourless to slight yellow).
FIGURE 7: UV-vis-NIR spectroscopy. Blue: DMF; red: beech wood-extracted with DMF;
black arrows: overtones of water
Within the range of the NIR spectrum (800 – 3000 nm), the two overtones of water (ca. 1400
nm, 1950 nm) merit consideration, because they reveal that the extract contains more water
than DMF at the beginning of extraction (in a closed jar). Hence the additional water in the
extract derives from the wood, which contained about 12.5 % moisture content when
extraction started. This is indicative of mobilization of the moisture bound in the wood cell
walls by the hygroscopic DMF. The FTIR spectrum of the extract (Fig. 8) indicates that lignin
was extracted from the wood. This is in agreement with [27] who supported lignin extraction
by means of DMF.
123
FIGURE 8: FTIR Spectra of beech wood (red), DMF beech wood extract (blue) and (for comparison ex library) lignin sulphonate (green)
Also [28] used DMF for selective pulping of wheat, consequently resolving lignin at low
cellulose degradation. After re-drying, the extracted wood chips showed a whitish surface
colour, which confirmed an enhanced proportion of cellulose and hemicellulose after
superficial extraction of the brownish lignin. Even though the reference spectra of lignin in
the literature show typical patterns (bands at 1600, 1515, 1459 and 1423 cm-1), a certain
degree of variability persists. This is not only because of the different spectroscopic
techniques applied or the different kinds of lignin [29], but also due to the specific method of
lignin isolation that subsequently influences the nature of the resulting lignin itself [27, 30].
However, the spectra depicted in Figure 8 are in agreement with those published by [29-31].
3.3. FTIR microscopy with chemical imaging
In addition to the droplets mentioned above, the influence of DMF on the curing process of
the 1C-PUR wood adhesive was also investigated on bonded beech wood specimens, since
the in situ behaviour of 1C-PUR in a bondline might be quite different to that of an individual
1C-PUR droplet [32]. Figure 9 (right) and Figure 10, which were both taken from the same
location on the same specimen, depict that the tracheae vessel system with its wide lumina
represents the only major pathway for 1C-PUR to enter from the bonding surface into the
beech wood. Fig. 9 (left) confirms this observation. There is no evidence of substantial
124
penetration of 1C-PUR into the walls or lumina of other cell types, such as libriform fibres or
wood rays. This observation is in agreement with [33-36].
FIGURE 9: FTIR microscopy images with chemical imaging of beech wood, bonded with 1C-PUR. Left: Bonded without priming. Right: Bonded after DMF priming. Blue: Wood. Red: 1C-PUR. Green: Spots where a mixture of wood and 1C-PUR were detected. Black: ambient air
Regarding the penetration of DMF into the adherends, Figure 10 reveals that the DMF
penetrates the wood by at least 250 μm and there is no specific cell type that was not
permeated by DMF.
FIGURE 10: FTIR image with chemical imaging of beech wood, bonded with 1C-PUR after DMF priming. Blue: No DMF detected (presence of wood, 1C-PUR or ambient air). Colours other than blue: Spots where DMF absorbance was detected
The influence of DMF on the degree of NCO conversion, which was observed in a droplet
(see 3.1), was also observed in the bondline (Fig. 11). The images displayed in Fig. 9 (left)
and Fig. 11 (left) were taken from the same location on the same specimen.
125
FIGURE 11: Chemical imaging of NCO absorbance: FTIR images of beech wood with 1C-PUR bond line. Left: Bonded without priming. Right: Bonded after DMF priming. Violet or blue: No NCO detected. Colours other than violet or blue: Spots where NCO absorbance was detected
The chemical imaging of the NCO absorbance (Fig. 11, left) confirms that, after curing,
residual NCO persists inside the cured 1C-PUR polymer (in the bondline and in tracheae
filled with 1C-PUR). This corresponds to results published by [37]. However, after DMF
priming no residual NCO is detectable (Fig. 11, right, taken at the same location as Fig. 10,
and Fig. 9, right). This validates the hypothesis that DMF promotes the diffusion of moisture
through the urea coating diffusion barrier. It is worth noting that all the bonded specimens
were glued together 7 days before the investigations took place. In between the bonding and
chemical imaging periods, they were stored in a climatised box at 20 – 23 °C and 65 % - 70 %
RH, thus preserving a wood moisture content of about 12.5 %, which provides enough
moisture for the 1C-PUR curing reaction.
3.4. UV fluorescence microscopy
As Figure 12 reveals (left and right images), the late wood represents a barrier for the
penetration of 1C-PUR adhesive, from the bondline into the beech wood tissue, since it does
not comprise large-pore tracheae. Furthermore, there is no evidence of the adhesive entering
the wood rays. This is in accordance with observations described under 3.3.
126
FIGURE 12: UV fluorescence microscopy. Left: 1C-PUR bonded beech wood without DMF priming. Right: 1C-PUR bonded beech wood after DMF priming
The combination of UV microscopy and chemical imaging reveals that the UV tracers and the
1C-PUR itself were detected within the same wood structures. Obviously, such tracers clearly
indicate where the adhesive is placed and are not “filtered out” of the adhesive during the
pressing procedure by the wood’s microstructures. A comparison of the two images (Fig. 12)
shows that the bondline of the left sample (no DMF) appears to be more compact, especially
where the late wood meets the bondline. Although, the DMF treated sample (right image)
shows more cavities in the bondline, indicating increased CO2 development due to a more
intense reaction of the 1C-PUR (in agreement with 3.1). However, the bondline images
should be interpreted with caution, since each image displays only one single site on each
specimen.
3.5. Nanoindentation and AFM images
The outcomes of the NI tests (Table 1) indicate that the DMF priming has a greater impact on
the mechanics of the adhesive polymer than on the mechanics of the wood. After priming, the
modulus of elasticity (MOE) and the hardness of the bondlines were significantly reduced by
about 17% and 20%, respectively (compared to the unprimed bondlines). Obviously DMF has
a “softening effect” on the 1C-PUR polymer. No significant changes in the mechanics of the
cell walls were detected. The MOE values measured for the unprimed bondlines are in
accordance with [36] and [38].
TABLE 1: Nanoindentation on 1C-PUR bonded beech wood.
500 m
127
Place of indentation Priming
Elastic Modulus Hardness Number
of
indentsAverage
[GPa]
Conf.
interv.
Average
[GPa]
Conf.
interv.
Bondline
none 2.61 0.081 0.2 0.004 26
DMF 2.16 0.086 0.16 0.007 30
Cell walls next to
bondline
none 11.67 0.828 0.44 0.036 26
DMF 11.9 1.602 0.41 0.043 10
Conf. Interv. = Confidence Interval at = 0.05 level of significance
Changes in MOE and hardness of such segmented 1C-PUR polymers can be caused by
changes in the polymer’s morphology, which is characterized by content and distribution of
hard segments and soft segments [8, 39-42]. The urea linkages of the hard segments in
particular take effect as strong crosslinking points in the hardened polymer network. As the
AFM micrographs reveal, the DMF treatment actually changes the morphology of the cured
1C-PUR polymer in the bondlines. A considerable reduction in the proportion of hard
segments was detected after DMF priming (Fig. 13, granular texture on left image compared
to right image).
FIGURE 13: AFM micrographs of hardened 1C-PUR polymer in the middle of the bondline. Left: Bonding performed without priming. Right: Bonding performed after DMF priming
5 m
128
4. Conclusions
The results of the present study indicate that DMF accelerates the kinetics of the 1C-PUR
hardening at 20 °C, influences the morphology of the hardened polymer and reduces its MOE
and hardness. Since no evidence was found for a chemical reaction between DMF and 1C-
PUR the DMF is regarded as a “diffusion promoter”. It is likely that the combination of the
softening effect and the hygroscopic character of the high boiling DMF both support the
diffusion of moisture through the urea skin, which begins to form on the surface of the
polymer. This supports the thorough conversion of the 1C-PUR in such a way that no residual
NCO persists in the hardened polymer.
Regarding the wooden adherends, it was found that DMF partially dissolves lignin and
mobilizes moisture bound in the wood cell walls, which then also becomes available for the
1C-PUR conversion. The beech wood’s vessel system turned out to be the only major
pathway for the 1C-PUR’s penetration into the wood. No 1C-PUR polymer was found in the
cell walls or lumina of wood rays, libriform fibres or late wood in general. Although, it is
unlikely that priming with toxic DMF is suitable for an implementation in practice, the
outcomes may help to find new approaches for formaldehyde-free priming of 1C-PUR wood
bondings. All results confirm that a primer can influence both, the adherend and the adhesive
polymer. However, in order to suitably change their characteristics, a chemical reaction with
the primer itself is not necessarily needed.
Acknowledgements
We gratefully acknowledge the financial support of this research by the Commission for
Technology and Innovation CTI (Bern, Switzerland) and the Purbond AG (Sempach-Station,
Switzerland).
129
References
[1] Y. Gartia, J. Moore, C. Felton, S. Pulla, B. Berry, P. Munshi, et al., Total replacement of
solvent in polyurethane synthesis using carbon dioxide soluble
1,3-dichlorodistannoxane catalysts, J Appl Polym Sci, 128, 6 (2013) 3522-3527
[2] O. Kläusler, P. Hass, C. Amen, S. Schlegel, P. Niemz, Improvement of tensile shear
strength and wood failure percentage of 1C PUR bonded wooden joints at wet stage
by means of DMF priming, Eur. J. Wood Prod., 72, 3 (2014) 343-354
[3] C.R. Frihart, Adhesive Groups and How They Relate to the Durability of Bonded Wood,
J. Adh. Sci. Techn., 23, 4 (2009) 601-617
[4] M.L. Weiner, Reaction of Phenyl Isocyanate with N,N-Dimethylformamide, J Org Chem,
25, 12 (1960) 2245-2247
[5] A. Jovtscheff, F. Falk, Umsetzung von N-Dimethylformamid mit Phenyl- und
-Naphthylisocyanat, J. Prakt. Chem., 4, 13 (1961) 265-271
[6] D. Joel, P. Muller, R. Ahl, Isocyanate Reactions in and with N,N-Dimethylformamide,
Angew Makromol Chem, 197, (1992) 131-139
[7] H. Ulrich, Synthesis of Polymers from Isocyanates in Polar Solvents, J. Polymer Sci.,
11, (1976) 93-133
[8] S. Oprea, Effect of Solvent Interactions on the Properties of Polyurethane Films,
17, 2 (2005) 163-173
[9] H.E. Ashton, The Swelling of Wood in Polar Organic Solvents, National Research
Council Canada. Division of Building Research, 1973
[10] G.I. Mantanis, R.A. Young, R.R. M., Swelling of wood, Part II. Swelling in Organic
Liquids, Holzforschung, 48, 6 (1994) 480-490
[11] A.N. Nayer, R.L. Hossfeld, Hydrogen Bonding and the Swelling of Wood in Various
Organic Liquids, 71, 8 (1949) 2852-2855
[12] G.I. Mantanis, R.A. Young, R.M. Rowell, Swelling of Wood Part III. Effect of
Temperature and Extractives on Rate and Maximum Swelling, Holzforschung,
49, 3 (1995) 239-248
[13] R.G. Schweiger, Waterless Solvent Systems for Cellulose, Papier, 28,
12 (1974) 534-536
[14] B.J. Fuhr, R.G. Jonasson, F. Tosto, Tappi, Fingerprinting of extractives in woods and
pulps, Pulping Conference, Montreal, CA, 1998, 1389-1402
130
[15] C.S. Brossia, S. D.A., Organic Liquids, in: Baboian R. (Ed.), Corrosion Tests and
Standards: Application and Interpretation, ASTM International, Clearway Logistics,
PA, USA, 2005, p. 867
[16] Sigma-Aldrich, Safety Data Sheet of N,N-Dimethylformamide, Version 6.1,
Sigma-Aldrich International GmbH, St. Gallen, 2013
[17] P. 302-1, prEN 302-1:2011 Adhesives for load-bearing timber structures- Test methods-
Part 1: Determination of longitudinal tensile shear strength, European Committee for
Standardization, Brussels, 2011
[18] K. Groestad, P.O. Kristiansen, Quantitative Determination of Melamine in MUF Resins
Using IR-Analysis, in: Society F.P. (Ed.), Wood Adhesives 2000, Forest Products
Society, South Lake Tahoe, Nevada, 2001, 374-375
[19] A.L. Daniel-Da-Silva, J.C.M. Bordado, J.M. Martín-Martínez, Moisture curing kinetics
of isocyanate ended urethane quasi-prepolymers monitored by IR spectroscopy and
DSC, 107, 2 (2008) 700-709
[20] R. Wimmer, B.N. Lucas, T.Y. Tsui, W.C. Oliver, Longitudinal hardness and Young's
modulus of spruce tracheid secondary walls using nanoindentation technique, Wood
Sci Technol, 31, 2 (1997) 131-141
[21] Sigma-Aldrich, The Aldrich Library of FT-IR Spectra, 2 ed., Sigma-Aldrich Co.,
USA, 1997
[22] A.L. Daniel Da Silva, J.M. Martín-Martínez, J.C.M. Bordado, Influence of the free
isocyanate content in the adhesive properties of reactive trifunctional polyether
urethane quasi-prepolymers, 26, 5 (2006) 355-362
[23] J.G. Dillon, Infrared spectroscopic atlas of polyurethanes, 1989
[24] E. Yilgor, I. Yilgor, E. Yurtsever, Hydrogen bonding and polyurethane morphology. I.
Quantum mechanical calculations of hydrogen bond energies and vibrational
spectroscopy of model compounds, 43, 24 (2002) 6551-6559
[25] A.M. Heintz, D.J. Duffy, S.L. Hsu, W. Suen, W. Chu, C.W. Paul, Effects of reaction
temperature on the formation of polyurethane prepolymer structures, 36, 8 (2003)
2695-2704
[26] W. Sonderegger, S. Hering, D. Mannes, P. Vontobel, E. Lehmann, P. Niemz,
Quantitative determination of bound water diffusion in multilayer boards by means
of neutron imaging, 68, 3 (2010) 341-350
[27] S.N. Sun, M.F. Li, T.Q. Yuan, F. Xu, R.C. Sun, Effect of ionic liquid/organic solvent
131
pretreatment on the enzymatic hydrolysis of corncob for bioethanol production.
Part 1: Structural characterization of the lignins, Ind Crop Prod, 43, (2013) 570-577
[28] A. Ghozatloo, J. Mohammadi-Rovshandeh, Modelling of Pulp Properties During
Delignification on Non-Wood Materials by DMF, Cellulose Chemistry and
Technology, 45, 5-6 (2011) 361-370
[29] O. Faix, Classification of Lignins from Different Botanical Origins by FT-IR
Spectroscopy, 45, s1 (1991) 21-28
[30] L.M. Kline, D.G. Hayes, A.R. Womac, N. Labbe, Simplified Determination of Lignin
Content in Hard and Soft Woods Via Uv-Spectrophotometric Analysis of Biomass
Dissolved in Ionic Liquids, Bioresources, 5, 3 (2010) 1366-1383
[31] K.K. Pandey, A.J. Pitman, FTIR studies of the changes in wood chemistry following
decay by brown-rot and white-rot fungi, Int Biodeter Biodegr, 52, 3 (2003) 151-160
[32] D. Ren, C.E. Frazier, Wood/adhesive interactions and the phase morphology of
moisture-cure polyurethane wood adhesives, Int. J. Adhes. Adhes., 34, (2012) 55-61
[33] P. Hass, F.K. Wittel, M. Mendoza, H.J. Herrmann, P. Niemz, Adhesive penetration in
beech wood: experiments, Wood Sci Technol, 46, 1-3 (2012) 243-256
[34] F.A. Kamke, J.N. Lee, Adhesive penetration in wood - a review, Wood Fiber Sci, 39, 2
(2007) 205-220
[35] J. Konnerth, D. Harper, S.-H. Lee, T.G. Rials, W. Gindl, Adhesive penetration of wood
cell walls investigated by scanning thermal microscopy (SThM), Holzforschung,
62, 1 (2008), 91-98
[36] J. Konnerth, W. Gindl, Mechanical characterisation of wood-adhesive interphase cell
walls by nanoindentation, Holzforschung, 60, 4 (2006), 429-433
[37] D. Ren, C.E. Frazier, Structure/durability relationships in polyurethane wood adhesives:
Neat films or wood/polyurethane composite specimens?, Int J Adh Adh, 45,
(2013) 77-83
[38] F. Stoeckel, J. Konnerth, W. Gindl-Altmutter, Mechanical properties of adhesives for
bonding wood - A review, Int J Adh Adh, 45, (2013) 32-41
[39] S. Clauβ, D.J. Dijkstra, J. Gabriel, O. Kläusler, M. Matner, W. Meckel, et al., Influence
of the chemical structure of PUR prepolymers on thermal stability,
Int. J. Adhes. Adhes., 31, 6 (2011) 513-523
[40] E. Yilgor, E. Yurtsever, I. Yilgor, Hydrogen bonding and polyurethane morphology. II.
Spectroscopic, thermal and crystallization behavior of polyether blends with
132
1,3-dimethylurea and a model urethane compound, Polymer, 43, 24 (2002)
6561-6568
[41] I. Yilgor, E. Yilgor, Structure-morphology-property behavior of segmented thermoplastic
polyurethanes and polyureas prepared without chain extenders, Polym. Rev., 47, 4
(2007) 487-510
[42] O. Kläusler, S. Clauß, L. Lübke, J. Trachsel, P. Niemz, Influence of moisture on
stress–strain behaviour of adhesives used for structural bonding of wood,
Int J Adhes Adhes, 44, 44 (2013) 57-65
133
5. Additional investigations
5.1 Sorption behaviour of adhesive films
As described under 4.1, moisture clearly influences the mechanical behaviour of adhesive
polymer films (except for PRF). Hence the question arises, whether this mechanical behaviour
correlates with the moisture uptake of such polymers. Since secondary bonds like hydrogen
bonds contribute to the polymer strength, it is reasonable to assume that polar water
molecules, accumulated in the bondline polymer, are capable of breaking hydrogen bonds,
thus reducing the polymer’s performance. Therefore vapour sorption isotherms were
measured, depicting the equilibrium amount of water vapour absorbed as a function of
relative vapour pressure (from 0% to 98% RH) at 20°C. The tests were performed by Prof.
Dr. R. Wimmer in cooperation with the Göttingen University; using the identical sample
material previously used for paper I (see 4.1). In addition, protein-based fish glue and
polyvinyl acetate were tested.
The results described by Wimmer et al. [88] revealed a comparatively slow moisture uptake
of 18% for PRF, and 22% for MUF. The 1C PUR samples revealed a medium fast weight
gain of just 3.5%. Hence, there is no evidence that the polymer that reveals the most stable
mechanical performance under changing RH (PRF) is the one showing the least moisture
sorption. On the contrary, the two 1C PURs, which both reveal a clear reduction in
mechanical performance due to increased ambient humidity (see 4.1), absorbed only about
3.5% moisture (see Figure 5.1). The formaldehyde- reacting adhesives (MUF, PRF)
accumulated considerably higher amounts of moisture in their polymer network than the 1C
PUR sample.
134
Figure 5.1 Sorption isotherms of adhesive polymer films (Wimmer et al. 2013)
5.2 Diffusion behaviour of adhesive films
When water vapour gets in contact with diffusion-open gluelines (PRF, MUF), moisture
diffusion through the adhesive layer can take place. On the other hand bondlines made from
PUR or EPI rather act as a kind of diffusion barrier [89-91]. Over time such a diffusion barrier
can lead to a quite uneven distribution of moisture across a building element. In some cases
this outcome is desired (e.g. in the case of wall elements equipped with a vapour barrier foil),
and sometimes it isn’t. Inside a thick load bearing beam, for instance, a very uneven moisture
distribution could result in higher swelling strain and higher stress, which could weaken the
joints by promoting delamination. Therefore it was interesting to investigate, whether the
moisture diffusion through a 1C PUR film can be influenced by adding resolvable “filler” like
salt (NaCl) powder. This investigation was carried out as part of a bachelor thesis [92].
135
After adding 3% salt powder to a commercial 1C PUR (see Table 5.2.1), thin films (about 0.2
mm average thickness) were cast and tested on the basis of EN ISO 12572:2012 [93] by
means of wet cups and dry cups (see Figure 5.2.1).
Table 5.2.1 Overview on batches for diffusion tests
Batch Adhesive Salt addition [%] Particle diameter [µm] n
(dry / wet)
VN3112
1C PUR 309
0 5 / 5
VN3115 3 ≤ 50 5 / 5
VN3116 3 75 – 100 5 / 5
Figure 5.2.1 Wet cup and dry cup with 1C PUR films for diffusion experiments (diameter of
diffusive surface: 5.1 cm)
The water vapour diffusion resistance value (µ) is used to represent diffusivity (see Table
5.2.2). The higher this µ-value, the less permeable the material is to water vapour.
136
Table 5.2.2 Calculation of the water vapour diffusion resistance factor
Water vapour diffusion resistance factor [ - ]
δa . ∗
∗ ∗
. water vapour diffusion coefficient of air [kg/(m*s*Pa)]
p0 Standard air pressure (= 101‘325) [Pa]
p Air pressure [Pa]
Rv Gas constant for water vapour (= 426) [N*m/(kg*K)]
T Temperature in the testing room [K]
δ Water vapour diffusion coefficient [kg/(m*s*Pa)], whereas δ = W *d
W Water vapour transmission coefficient [kg/(m2*s*Pa)]
d Average sample thickness [m]
Results suggest that salt particles influence the water vapour diffusion resistance factor ,
when the film is exposed to high RH (about 95% in the wet cups, which most likely partially
dissolves the salt). The differences in average µ -values are considerable (about – 30% for
batches with salt), but not statistically significant ( = 0.05 level). Evidence for influence of
particle size on the diffusability was not detected in this work (see Figure 5.2.2). Additional
tests, including tensile shear tests of accordantly bonded specimens, would be helpful to gain
a clearer picture on this topic.
137
Figure 5.2.2 Water vapour diffusion resistance factor of modified 1C PUR films measured
by means of dry cup and wet cup tests. Error bars: Confidence intervals ( = 0.05)
VN 3112 VN 3115 VN 3116Dry Cups 11387 11143 11139
Wet Cups 8996 6109 6402
0
2'000
4'000
6'000
8'000
10'000
12'000
14'000
16'000
18'000
µ [
-]
138
6. Synthesis
6.1 Main findings
Under dry conditions 1C PUR bonded wooden joints show good performance (TSS and WFP)
and meet the demands of various international technical standards for loadbearing timber
structures. Under high moisture conditions however, such joints reveal a reduction of TSS
(yet still comply with the demands of the accordant technical standards) and a heavy loss in
WFP. Under such conditions the joints no longer meet the demands of CSA O 112 for WFP.
The loss of adhesion in the wet state turned out to be an important aspect in this context, since
it is a limiting factor regarding load transfer between the two adherends. Therefore an attempt
was made to find a formaldehyde- free coupling agent, that affects the wood and the adhesive
as well. The solvent DMF (see 4.3 and 4.4) proved capable of interacting with the wood
(improved wettability, swelling, partly resolving lignin) and with the 1C PUR (more thorough
conversion, reduction of MOE, and reduction of hardness). As tensile shear tests revealed,
DMF substantially enhances TSS and WFP of 1C PUR bonded wooden joints at dry and wet
stage when used as a wood primer before application of the adhesive. After such priming the
adhesion of the adhesive on the boundary layer of the adherends in the wet state improved,
making higher WFP possible. This was tested on European beech wood (Fagus sylvatica L.)
and Douglas fir (Pseudotsuga menziesii (Mirb.) Franco).
The fact that TSS is reduced in the wet state (compared to the dry state) and reaches the initial
dry-state values again after re-drying, indicates the high importance of hydrogen bonds for the
performance of 1C PUR bonded wooden joints. They are disrupted when water enters the
specimen and are largely re-established during re-drying. These results also confirm that the
wood, the adhesive polymer and the boundary layer of the specimens are not irreversibly
damaged by swelling and shrinking (see 4.2 and 4.3), otherwise this re-gaining of strength
would not occur.
As described under 4.2 and 4.3 the WFP of 1C PUR bonded joints in the wet state reveals
skewed, bimodal distribution. The standard CSA O 112 confirms that this is a non normal-
distribution for PRF- bonded wooden joints. Accordingly, within one testing batch (using
both polymers independently) samples with very high and samples with very low WFP were
139
the most likely outcome, even though the same materials (wood from the same slat, identical
adhesive) and the same procedures (wood preparation, gluing, water treatment, testing) were
used. A consistent explanation for this finding is missing so far. Obviously, the interplay of
the various parameters influencing WFP (wood strength, WMC, swelling- strain, fibre –load –
angle, mechanics of the adhesive polymer, adhesive penetration, etc.) has not yet been
sufficiently investigated. This should be a focus for future research.
As described under 4.1 (paper I), the MOE of the investigated 1C PUR adhesive polymer film
is about half of that of the MUF or PRF film. Therefore it was initially assumed, that the low
WFP correlates with this low MOE, but the investigation described under 4.4 revealed that
DMF treatment reduces the MOE of the glued joint and increases WFP (see 4.3). Hence high
WFP in the wet state can be produced by means of an adhesive polymer with high MOE
(PRF, MUF) or with low MOE (DMF –treated 1C PUR). Hence the MOE cannot be regarded
as the most crucial single influencing factor for WFP. The same accounts for hardness (see
4.4) and cohesive strength (see 4.1 and 4.3).
The concept of chemically and mechanically weak boundary layers and their influence on the
bonding performance is frequently discussed in literature [29, 94, 95]. However, the different
mechanical preparation techniques performed in this work affected surface parameters like
surface roughness or wettability. Although the test results also revealed, that the WFP under
wet conditions cannot be substantially influenced by performing a specific machining
technique (planing, face milling, etc.).
In contrast, chemical surface modifications (e.g. DMF) heavily changed the surface
wettability and WFP in the wet state. Even “water priming” improved the 1C PUR joint
performance in the dry state (see 4.2). Hence, future investigations should focus on the
influence of wood extractives at the boundary layer on the performance of 1C PUR bonded
wooden joints, since they affect the wettability of the joining surfaces (see 4.2).
Various coupling agents, which are capable of reacting directly with the bonding partners,
thus creating covalent bonds, were considered. However, the example of DMF, influencing
the 1C PUR, reveals that such direct reactions are not necessarily required in order to affect
140
the performance of the bonded composite material. Just “promoting” the adhesive’s reaction
or improving the interplay of the bonding partners can increase the performance significantly.
6.2 Conclusive discussion
This thesis is focused on increasing the WFP of 1C PUR bonded structural wooden joints in
wet environments. The need for high WFP basically derives from technical standards,
claiming that high WFP in combination with high bond joint strength is indicative for a joint
strength that clearly exceeds the strength of the wooden substrate itself. Initially, this
“traditional approach” appears to be quite convincing. Obviously the joint strength must be
very high if high grade wood fractures occur outside the bondline. Consequently joint strength
seems to be the most important factor influencing WFP. However, additional influencing
factors such as fiber-load angle or adhesive penetration into the cell wall should be taken into
account (see 4.3, swelling-strain model). Under wet conditions a pre-swollen state could
transfer swelling strain into the adherend, promoting a fracture path within the wood. Here a
correlation between joint strength and WFP is not necessarily expected and was also not
detectable within this thesis (see 4.3). A consistent model explaining the formation of WFP
and the interplay of its polyfactorial causes in the dry state compared to wet state is still
pending. Micro- mechanical tests could help to gain progress in this field (see 6.2). The
general discussion about whether a fracture path running through the wooden substrate is
always superior to a fracture path running through the bondline of a joint of the same strength
is not yet conclusive [38, 96]. No evidence was found for the traditional opinion, that high
WFP on sound wood can serve as an indicator for high TSS in the wet state.
However, a general agreement has been reached regarding the importance of thresholds for
strength, on account of different technical standards committees (EN, ISO, CSA, ASTM)
having all defined various thresholds. The situation regarding WFP is quite different. Even
experts in standards committees give WFP quite different degrees of importance. Some NA
standards set accordant limits under wet conditions (CSA O112.9) and some do for tests in the
dry state (ASTM D 2559 – 04). Some EU standards don’t set any such limits at all (EN 301,
EN 302-1, EN 15425). Others, like EN 14080 [43], define thresholds under dry conditions,
but not in the wet state. The latter even defines variable threshold values for WFP, which are
141
dependent on the joint’s shear strength. With increasing joint strength the accordant
thresholds for WFP decrease. For a specimen with at least 10 MPa shear strength (hence
including those stronger than the wood) a WFP of 20% is declared as sufficient. A WFP of
100% is only required for specimens with less than 6 MPa shear strength, pointing at poor
wood quality.
As Milner [97] described, the discussion regarding the development of more performance-
based standards for structural timber bondlines is on-going. This might probably help to
develop a more consistent and more common view on the WFP topic. One of the few really
common aspects is that the combination of low strength and high WFP is regarded as
indicative of low quality wood.
142
6.3 Potential for future research
Science driven topics
In view of all the investigations performed during the thesis, it was quite surprising that the
application of a solvent proved to be such a powerful way to enhance the performance of the
bonded composite. Obviously the potential of the 1C PUR- wood- system had not formerly
been driven to its full extent. Hence, there are still possibilities to “exploit” this potential
beyond the classical approaches, such as changing raw materials for the prepolymers or
adhesive tailoring with additives. Further tests with other solvents like the less toxic DMAC
(less hygroscopic) could help to achieve a more distinct differentiation between influencing
factors such as hygroscopicity, swelling potential and others. The overriding goal should be to
detect the decisive modes of actions. This could open up pathways for the development of
other primers or even more efficient adhesives. In addition, such investigations could give
valuable hints on the crucial factors influencing WFP.
Furthermore, other unconventional paths of priming should also be taken into consideration.
Compared to the quite polar wood surface, the 1C PUR polymer is of relatively low polarity.
Therefore another kind of “mediator” between the polar wood surface and the less polar fluid
polymer could probably be a kind of emulsifier, which is basically capable of making unpolar
fluids dispersible in polar fluids.
Since the adhesion between wood and adhesive turned out to be a limiting factor for the
performance of 1C PUR bonded joints under moisture load, the boundary layer between the
two should be the focus of further investigations under varied ambient humidity. Regarding
the physical and chemical binding mechanisms between the bonding partners, special
attention should be paid to the influence of secondary bonds such as hydrogen bonds. The
investigations revealed that under wet conditions one and the same 1C PUR polymer
sometimes creates 90% WFP and sometimes 0%, just within the same testing run. Hence ,
since the polymer properties and the testing conditions remained unchanged, the chanced
influencing parameter must be something else. After all it is reasonable to conclude that the
wood extractives play an important role in this respect (see contact angle measurements 4.2 &
4.3).
143
Additionally, micromechanical tensile tests on bonded veneer specimens could be carried out
(not pre-swollen, partly pre-swollen, completely pre-swollen, using different adhesives) under
varied ambient climate conditions. The aim of such tests is to investigate the translocation of
swelling strain (see 4.3). This could be performed using Digital Image Correlation or Speckle
Interferometry [98, 99].
Industry driven topics
Since 1C PURs were introduced on a large scale in the 1990s, long term experiences
regarding the durability of the joints have not yet materialised. Such investigations require
considerable time (natural weathering even several years). Hence, it would be advantageous
to presently start the monitoring of weathering tests (natural and artificial) or creep tests under
constant load and varied ambient climates with an intention to observe the behaviours over
many years.
Furthermore the detected effects of hygroscopic DMF on the bonding performance could be
helpful regarding the bonding of very dry wood (mobilization of moisture by DMF) or in
terms of gluing wood species which are known for their low performance when bonded with
1C PUR (e.g. Larix spp.).
During preparation of this thesis the DMF- priming procedure was tested on an industrial
scale and a significant reduction of the glued joint delamination was achieved. However, since
DMF is a toxic solvent this protocol will certainly not be used on a large scale in the future.
Nonetheless this demonstrates again that further improvement steps are also possible
regarding parameters like climate dependent delamination, which are relevant to safety.
In order to gain industrial acceptance in general, DMF should be replaced by a less toxic
substance, which could probably (in the long term) be implemented into a 2C or even 1C
adhesive system.
144
7. References
[1] Market Development Agency of the Government of British Columbia, FII Ltd. (2014)
Wood First Program. Download April 1st 2014 from
http://www.bcfii.ca/bc-forest-sector/wood-first
[2] Von Büren, C.; Weibel, M. (2014) Aktionsplan Holz - Magazin Phase 1. Bundesamt für
Umwelt (BAFU), Schweizerische Eidgenossenschaft, Bern
[3] Bernhard, T.; Beer-Tóth, K.; Duda, R. (2012) Portrait of the National Research
Programme (NRP 66) - Resource Wood, in: SNF (Ed.), Swiss National Science
Foundation SNF, Bern
[4] Meuter, M. (2014) Herkunftszeichen Schweizer Holz für die Neumattbrücke.
Download April 2nd 2014 from http://www.holz-portal.ch
[5] Egli, U. (2014) Neumattbrücke ist montiert und bereit. Download April 1st 2014 from
http://www.bernerzeitung.ch/region/emmental/Neumattbruecke-ist-montiert-
und-bereit/14587910/print.html
[6] Häring, C.; Schneider, R. (2012) Riesenkuppel aus Holz. Tec21 31: 21-25
[7] Hunziker M. (2011) Eine Lösung für den koreanischen Korb. Baublatt 42: 24-28
[8] Wissmann R. (2014) 150 Millionen für neue Swatch-Hauptsitze. Download April 2nd 2014
from http://www.derbund.ch/bern/kanton/So-soll-der-neue-SwatchHauptsitz
-aussehen/30081304/print.html
[9] Blumer, H. (2009) Bionische Architektur mit Holz – unsere Chance. Schweizer
Holzbau 10: 11-18
[10] Pfäffinger, J. (2013) Bürohaus der Tamedia in Zürich. Mikado 1: 14-21
[11] Kübler, W.; Jacob-Freitag, S. (2011) Pionierleistung mit Ausblick. Mikado 9: 30-34
[12] Frangi, A. (2013) ETH House of Natural Resources. Presentation at the
Montagskolloquium für die Praxis, ETH Zürich
[13] Keller, E.-M.; Von Fischer, S. (2014) Neue Monte-Rosa-Hütte. Fakten zum Projekt.
Fact Sheet. Download April 3rd 2014 from http://www.neuemonterosahuette.ch
[14] Gregor, P. (2013) Der Baustoff Holz in innovativen Geschäftsmodellen – Holztürme für
Windkraftanlagen. 16. Proceedings of the Internationales Holzbau-Forum,
Garmisch Partenkirchen, pp. 1-13
[15] Aue, S. (2013) TimberTower im Aufwind. Press Release, Download April 10th 2014
from http://www.timbertower.de/wpcontent/uploads
145
/2013/11/Pressemitteilung_140mTurm.pdf
[16] Klein, S. (2014) Holzbau kommt dem Himmel immer näher. Holz-Zentralblatt 7: 160
[17] Bogusch, B. (2006) Zum Einsturz der Eishalle in Bad Reichenhall: Ursachen und
Folgerungen. Schweizer Holzbau 72(9): 4-9
[18] Steiger, R. (2013) Einstürze von Dachtragwerken aus Holz: Entstehung, Ursachen,
Folgerungen. Presentation at the FBH Herbsttagung, ETH Zürich
[19] Roll, W. (2000) Wood Adhesives and the Wood Adhesive Industry in Europe.
Proceedings of the Conference on Wood Adhesives, Forest Products Society,
South Lake Tahoe, pp. 51-56
[20] World Health Organization (WHO) (2006) Formaldehyde, 2-Butoxyethanol and 1-tert-
Butoxypropan-2-ol. IARC Monograph Volume 88, Intl. Agency for Research on
Cancer, Lyon
[21] Clauss, S. (2011) Structure-Property Relationship Of One-Component Moisture-Curing
Polyurethane Adhesives Under Thermal Load. PhD thesis No. 20060, ETH Zürich
[22] Klippel, M.; Frangi, A. (2012) Einfluss des Klebstoffes auf das Brandverhalten von
Brettschichtholz. Bauphysik 34(4): 142-152
[23] Klippel, M.; Frangi, A.; Fontana, M. (2011) Influence of the Adhesive on the
Load-Carrying Capacity of Glued Laminated Timber Members in Fire. Presentation
at the 10th International Symposium on Fire Safety Science (IAFSS 2011),
International Association for Fire Safety Science, College Park, Maryland
[24] Klippel, M.; Frangi, A.; Hugi, E. (2014) Experimental Analysis on the Fire Behavior of
Finger-Jointed Timber Members. Journal of Structural Engineering 140(3): 1-12
[25] Clauβ, S.; Dijkstra, D.J.; Gabriel, J.; Kläusler, O.; Matner, M.; Meckel, W.; Niemz, P.
(2011) Influence of the chemical structure of PUR prepolymers on thermal stability,
Int. J. Adhes. Adhes. 31(6): 513-523
[26] O. Kläusler, S. Clauß, L. Lübke, J. Trachsel, P. Niemz, Influence of moisture on
stress–strain behaviour of adhesives used for structural bonding of wood,
Int. J. Adhes. Adhes., 44: 57-65
[27] Dunky, M. (2002) Teil III Einflussgrössen, 1.4.2 Verleimungsrelevante Eigenschaften
von Holzoberflächen. In: Dunky, M.; Niemz, P. (Eds.), Holzwerkstoffe und Leime,
Springer, Berlin
[28] Habenicht, G. (2009) Applied adhesive bonding. Springer, Berlin
[29] Nussbaum, R.M. (1999) Natural surface inactivation of Scots pine and Norway spruce
146
evaluated by contact angle measurements. Holz Roh. Werkst. 57(6): 419-424
[30] Frihart, R.F.; Hunt, C.G. (2010) Adhesives with Wood Materials. Bond Formation and
Performance. In: Ross, R.J. (Ed.), Wood Handbook, Chapter 10, USDA, Forest
Products Laboratory, Madison, Wisconsin
[31] Mantanis, G.I.; Young, R.A.; Rowell, R.M. (1994) Swelling of wood Part 1. Swelling in
water. Wood Sci. Technol. 28: 119-134
[32] Mantanis, G.I.; Young, R.A.; Rowell, R. M. (1994) Swelling of wood, Part II. Swelling
in Organic Liquids. Holzforschung 48(6): 480-490
[33] Mantanis, G.I.; Young, R.A.; Rowell, R. M. (1995) Swelling of Wood Part III. Effect of
Temperature and Extractives on Rate and Maximum Swelling.
Holzforschung 49(3): 239-248
[34] Vick, C.B.; Okkonen, E.A. (1998) Strength and durability of one-part polyurethane
adhesive bonds to wood, Forest Prod J, 48, 11-12: 71-76
[35] Brandmair, A.; Jans, N.; Clauss, S.; Hass, P.; Niemz, P. (2012) Bonding of hardwoods
with 1C PUR adhesives for timber construction., Bauphysik, 34, 5: 210-216
[36] Uysal, B.; Özçifçi, A. (2006) Bond strength and durability behavior of
polyurethane-based Desmodur-VTKA adhesives used for building materials after
being exposed to water-resistance test, J Appl Polym Sci, 100, 5: 3943-3947
[37] Lopez-Suevos, F.; Richter, K. (2009) Hydroxymethylated Resorcinol (HMR) and
Novolak-Based HMR (n-HMR) Primers to Enhance Bond Durability of Eucalyptus
globulus Glulams, J Adhes Sci Technol, 23, 15: 1925-1937
[38] Hass, P.; Kläusler, O.; Schlegel, S.; Niemz, P. (2014) Effects of mechanical and
chemical surface preparation on adhesively bonded wooden joints,
Int J Adhes Adhes, 51: 95-102
[39] Frihart, C.R. (2009) Adhesive Groups and How They Relate to the Durability of Bonded
Wood, 23, 4: 601-617
[40] Vick, C.B.; Okkonen, E.A. (2000) Durability of one-part polyurethane bonds to wood
improved by HMR coupling agent, Forest Prod J, 50, 10: 69-75
[41] Ohnesorge, D.; Richter, K.; Becker, G. (2010) Influence of wood properties and bonding
parameters on bond durability of European Beech (Fagus sylvatica L.) glulams.
Ann. For. Sci., 67, 6: 601-610
[42] EN 1995-1-1 (2010) Eurocode 5. Design of timber structures- Part 1-1: General -
Common rules and rules for buildings. European Committee for Standardization,
147
Brussels
[43] EN 14080 (2013) Timber structures - Glued laminated timber and glued solid timber -
Requirements. European Committee for Standardization, Brussels
[44] prEN 301 (2011) Adhesives, phenolic and aminoplastic, for load-bearing timber
structures- Classification and performance requirements. European Committee for
Standardization, Brussels
[45] prEN 302-1 (2011) Adhesives for load-bearing timber structures- Test methods- Part 1:
Determination of longitudinal tensile shear strength. European Committee for
Standardization, Brussels
[46] EN 15425 (2008) Adhesives - One component polyurethane for load bearing timber
structures - Classification and performance requirements. European Committee for
Standardization, Brussels
[47] ASTM D 2559 -04 (2004) Standard Specification for Adhesives for Structural
Laminated Wood Products for Use Under Exterior (Wet Use) Exposure Conditions.
ASTM International, West Conshohocken, PA
[48] CSA O112.9-04 (2004) Evaluation of adhesives for structural wood products
(exterior exposure). Canadian Standards Association, Missiauga, Ontario
[49] Zauner, M. (2014) In-situ synchrotron based tomographic microscopy of uniaxially
loaded wood: in-situ testing device, procedures and experimental investigations.
PhD thesis No. 21620, ETH Zürich
[50] Wiedenhoeft, A. (2010) Structure and Function of Wood. In: Ross, R.J. (Ed.),
Wood Handbook. Chapter 3, Forest Products Laboratory, Madison, Wisconsin
[51] Fengel, D.; Wegener, G. (2003) Structure and Ultrastructure. In: Wood, Chapter 2,
Kessel, Remagen
[52] Bodig, J.; Jayne, B.A. (1993) Mechanics of Wood and Wood Composites.
Krieger Publishing, Malabar, Florida
[53] Niemz, P. (1993) Physik des Holzes und der Holzwerkstoffe. DRW, Leinfelden
[54] Panshin, A.J.; De Zeeuw, C. (1980) Textbook of technology. 4th Ed., Mac Graw-Hill,
New York
[55] Salmen, L. (2004) Micromechanical understanding of the cell-wall structure.
Comptes rendus biologies 327(9-10): 873-880
[56] Burgert, I.; Keckes, J.; Peter, F. (2006) Mechanics of the Wood Cell Wall. Chapter 3 in:
Stokke, D.D.; Groom, L.H. (Eds.), Characterization of the Cellulosic Cell Wall,
148
Blackwell Publishing
[57] Kerstens, S.; Decraemer, W.F.; Verbelen, J.-P. (2001) Cell Walls at the Plant Surface
Behave Mechanically Like Fiber-Reinforced Composite Materials. Plant Physiology
127: 381-385
[58] Burgert, I. (2006) Exploring the micromechanical design of plant cell walls. American
Journal of Botany 93(10): 1391-1401
[59] Baskin, T.I. (2005) Anisotropic expansion of the plant cell wall. Annual Review of
Cell and Developmental Biology 21: 203-222
[60] Keunecke, D.; Niemz, P. (2008) Axial stiffness and selected structural properties of yew
and spruce microtensile specimens. Wood Research 53(1): 1-14
[61] Lichtenegger, H.; Reiterer, A.; Stanzl-Tschegg, S.E.; Fratzl, P. (1999) Variation of
cellulose microfibril angles in softwoods and hardwoods - a possible strategy of
mechanical optimization. Journal of Structural Biology 128(3): 257-269
[62] Wagenführ, R. (2000) Holzatlas, 5 ed., Fachbuchverlag, Leipzig
[63] Burgert, I.; Eckstein, D. (2001) The tensile strength of isolated wood rays of beech
(Fagus sylvatica L.) and its significance for the biomechanics of living trees,
Trees-Struct Funct, 15, 3: 168-170
[64] Habenicht, G. (2009) Kleben - Grundlagen, Technologien, Anwendungen.
Springer, Berlin
[65] Hass, P.; Wittel, F.K.; Mendoza, M.; Herrmann, H.J.; Niemz, P. (2012) Adhesive
penetration in beech wood: experiments. Wood Sci. Technol. 46: 243-256
[66] Pizzi, A.; Mittal, K.L. (2003) Handbook of Adhesive Technology, 2ed, Marcel Dekker
New York
[67] Kaiser, W. (2007) Kunststoffchemie für Ingenieure. Von der Synthese bis zur
Anwendung. Carl Hanser, Munich
[68] Sebenik, U.; Krajnc, M. (2007) Influence of the soft segment length and content
on the synthesis and properties of isocyanate-terminated urethane prepolymers.
Int. J. Adhes. Adhes. 27(7): 527-535
[69] Properzi, M.; Pizzi, A.; Uzielli, L. (2003) Comparative wet wood glueing performance of
different types of Glulam wood adhesives. Holz Roh. Werkst 61(1): 77-78
[70] Richter, K.; Schirle, M.A. (2002) Behaviour of 1K PUR adhesives under increased
moisture and temperature conditions. In: Teischinger, A.; Stingl, R. (Eds.),
Lignovisionen, Proceedings of the International Symposium on Wood Based
149
Materials, Institute of Wood Science and Technology, University of Natural
Resources and Applied Life Sciences (BOKU), Vienna
[71] Schirle, M.A.; Künniger, T.; Fischer, A.; Richter, K. (2002) Charakterisierung und
Optimierung der Holzverklebung mit 1K-PUR Klebstoffen. KTI- Abschlussbericht,
Projekt 4126.1, EMPA, Dübendorf
[72] Na, B.; Pizzi, A.; Delmotte, L.; Lu, X. (2005) One-component polyurethane adhesives
for green wood gluing: Structure and temperature-dependent creep. J. Appl. Polym.
Sci. 96(4): 1231-1243
[73] Richter, K.; Pizzi, A.; Despres, A. (2006) Thermal stability of structural one-component
polyurethane adhesives for wood - structure-property relationship. Journal of
Applied Polymer Science 102(6): 5698-5707
[74] George, B.; Simon, C.; Properzi, M.; Pizzi, A.; Elbez, G. (2003) Comparative creep
characteristics of structural glulam wood adhesives. Holz Roh. Werkst. 61 (1): 79-80
[75] Frangi, A.; Fontana, M.; Mischler, A. (2004) Shear behaviour of bond lines in glued
laminated timber beams at high temperatures. Wood Sci. Technol. 38(2): 119-126
[76] Clauß, S.; Joscak, M.; Niemz, P. (2010) Thermal stability of glued wood joints measured
by shear tests. Eur. J. Wood Wood Prod. 69(1): 101-111
[77] Konnerth, J.; Gindl, W. (2008) Observation of the influence of temperature on the
mechanical properties of wood adhesives by Nanoindentation.
Holzforschung 62(6): 714-717
[78] Schmidt, M.; Glos, P.; Wegener, G. (2009) Verklebung von Buchenholz für tragende
Holzbauteile. Eur. J. Wood Prod. 68(1): 43-57
[79] Stoeckel, F.; Konnerth, J.; Gindl-Altmutter, W. (2013) Mechanical properties of
adhesives for bonding wood - A review. International Journal of Adhesion and
Adhesives 45: 32-41
[80] Beaud, F.; Niemz, P.; Pizzi, A. (2006) Structure-property relationships in one-component
polyurethane adhesives for wood: Sensitivity to low moisture content. J. Appl.
Polym. Sci. 101(6): 4181-4192
[81] Kägi, A.; Niemz, P.; Mandallaz, D. (2006) Einfluss der Holzfeuchte und ausgewählter
technologischer Parameter auf die Verklebung mit 1K-PUR Klebstoffen unter
extremen klimatischen Bedingungen. Eur. J. Wood Prod. 64(4): 261-268
[82] Kläusler, O. (2007) Untersuchung zur Auswirkung der Zusammensetzung von
Polyurethan-Prepolymeren auf die Verklebungsgüte von Buchenholz. Diploma
150
Thesis, Hamburg University
[83] Na, B.; Pizzi, A. (2011) One component polyurethane adhesives for green wood gluing:
Creep and temperature- dependent creep. Proceedings of the International
Conference on Mechanical Engineering and Technology, London, pp. 125-129
[84] Konnerth, J.; Stöckel, F.; Müller, U.; Gindl, W. (2010) Elastic properties of adhesive
polymers. III. Adhesive polymer films under dry and wet conditions characterized
by means of Nanoindentation. J. Appl. Polym. Sci. 118(3): 1331-1334
[85] Christiansen, A.W.; Vick, C.B.; Okkonen, E.A. (2001) Enhanced Durability of
One-Part Polyurethane Bonds to Wood Due to the Use of HMR Primer.
Proceedings of the International Conference on Wood Adhesives 2000,
Forest Products Society, South Lake Tahoe, pp. 489-494
[86] Brandmair, A.; Jans, N.; Clauss, S.; Hass, P.; Niemz, P. (2012) Verklebung von
Laubhölzern mit 1K-PUR-Klebstoffen für den Holzbau. Bauphysik 34(5): 210-216
[87] Kläusler, O.; Rehm, K.; Elstermann, F.; Niemz, P. (2014) Erratum: Influence of wood
machining on tensile shear strength and wood failure percentage of one component
polyurethane bonded wooden joints after wetting. Int. Wood Prod. Journal
5(2): 127-127
[88] Wimmer, R.; Kläusler, O.; Niemz, P. (2013) Water sorption mechanisms of
commercial wood adhesive films. Wood Sci. Technol. 47(4): 763-775
[89] Bowditch, M.R. (1996) The durability of adhesive joints in the presence of water.
Int. J. Adhes. Adhes. 16(2): 73-79
[90] Volkmer, T.; Schmidt, J.A.; Kranitz, K.; Niemz, P. (2012) Untersuchungen zum
Einfluss der Klebstoffart auf den Diffusionswiderstand von Holzverklebungen.
Bauphysik 34(2): 341-350
[91] Sonderegger, W.; Hering, S.; Mannes, D.; Vontobel, P.; Lehmann, E.; Niemz, P. (2010)
Quantitative determination of bound water diffusion in multilayer boards by means
of neutron imaging. Eur. J. Wood Wood Prod. 68(3): 341-350
[92] Uttinger, S. (2012) Einfluss des Füllstoffgehalts von 1K-PUR-Klebstofffilmen auf
deren Diffusionsverhalten. Bachelor thesis, ETH Zürich
[93] EN ISO 12572 Hygrothermal performance of building materials and products -
Determination of water vapour transmission properties. European Committee for
Standardization, Brussels
[94] Nussbaum, R.M. (1996) The critical time limit to avoid natural inactivation of spruce
151
surfaces (Picea abies) intended for painting or gluing. Holz Roh. Werkst.
54(1): 26-26
[95] Stehr, M.; Johansson, I. (2000) Weak boundary layers on wood surfaces. Journal of
Adhesion Science and Technology 14(10): 1211-1224
[96] Hass, P. (2012) Penetration Behavior of Adhesives into solid Wood and Micromechanics
of the Bondline. PhD thesis, ETH Zürich
[97] Milner, H.R. (2006) Towards a performance-based standard for structural timber
bondlines. Journal of Adhesion Science and Technology 20(8): 897-916
[98] Valla, A.; Konnerth, J.; Keunecke, D.; Niemz, P.; Muller, U.; Gindl, W. (2011)
Comparison of two optical methods for contactless, full field and highly sensitive
in-plane deformation measurements using the example of plywood. Wood Sci.
Technol. 45(4): 755-765
[99] Keunecke, D.; Novosseletz, K.; Lanvermann, C.; Mannes, D.; Niemz, P. (2012)
Combination of X-ray and digital image correlation for the analysis of
moisture-induced strain in wood: opportunities and challenges.
Eur. J. Wood Wood Prod. 70(4): 407-413
152
Publications
Peer reviewed papers
Clauß S., Dijkstra D., Gabriel J., Kläusler O., Matner M., Meckel W., Niemz P. (2011)
Influence of the chemical structure of PUR prepolymers on thermal stability.
International Journal of Adhesion and Adhesives 31, 6: 513 – 523
Hass P., Kläusler O., Schlegel S., Niemz P. (2014) Effects of mechanical and chemical
surface preparation on adhesively bonded wooden joints. International Journal of
Adhesion & Adhesives 51: 95-102
Kläusler O., Bergmeier W., Karbach A., Meckel W., Mayer E., Clauß S., Niemz P. (2014)
Influence of dimethylformamide on one-component moisture-curing polyurethane
wood adhesives. International Journal of Adhesion & Adhesives 55: 69-76
Kläusler O., Hass P., Amen C., Schlegel S., Niemz P. (2014) Improvement of tensile shear
strength and wood failure percentage of 1C PUR bonded wooden joints in the wet
state by means of DMF priming. European Journal of Wood and Wood Products
72:343–354
Kläusler O., Rehm K., Elstermann F., Niemz P. (2014) Influence of wood machining on
tensile shear strength and wood failure percentage of one- component polyurethane
bonded wooden joints after wetting. International wood products journal 5, 1: 18–26
Kläusler O., Clauß S., Lübke L., Trachsel J., Niemz P. (2013) Influence of moisture on
stress-strain behavior of adhesives used for structural bonding of wood. International
Journal of Adhesion and Adhesives 44: 57-65
Wimmer R., Kläusler O., Niemz P. (2013) Water sorption mechanisms of commercial wood
adhesive films. Wood Science and Technology 47, 4: 763-775