poly(ethylene isophthalate)s: effect of the tert-butyl substituent on structure and properties

8
Poly(ethylene isophthalate)s: effect of the tert-butyl substituent on structure and properties Robert Quintana a , Antxon Martı ´nez de Ilarduya a , Elisabet Rude ´ b , Darwin P.R. Kint a , Abdelilah Alla a , Juan A. Galbis c , Sebastia ´n Mun ˜oz-Guerra a, * a Departament d’Enginyeria Quı ´mica, Universitat Polite `cnica de Catalunya, ETSEIB, Diagonal 647, 08028 Barcelona, Spain b Departament d’Enginyeria Quı ´mica i Metal·lu ´rgia, Facultat de Quı ´mica, Universitat de Barcelona, Martı ´ i Franque `s 1, 08028 Barcelona, Spain c Departamento de Quı ´mica Orga ´nica y Farmace ´utica, Facultad de Farmacia, Universidad de Sevilla, 41071 Sevilla, Spain Received 6 January 2004; received in revised form 12 May 2004; accepted 13 May 2004 Available online 9 June 2004 Abstract A comparative study of the two isophthalic acid deriving homopolyesters poly(ethylene isophthalate) (PEI) and poly(ethylene 5-tert-butyl isophthalate) (PE t BI), including synthesis, crystal structure, and thermal and permeability properties, was carried out. The two polyesters were prepared by condensation polymerization in the melt. In both cases, minor amounts of cyclic dimers were observed to form, which were characterized by nuclear magnetic resonance and mass spectroscopy. PEI and PE t BI were thermally stable up to 400 8C and they appeared to be semicrystalline polyesters, having their melting temperatures between 130 and 135 8C. Their glass-transition temperatures were 62 and 94 8C, respectively. The crystal structure adopted by the two polyesters seemed to consist of a regularly folded conformation, clearly different from the almost extended conformation characteristic of poly(ethylene terephthalate). Gas permeability measurements for N 2 ,O 2 , and CO 2 revealed that PE t BI is more permeable to these gases than PEI, in spite of having a higher T g : Furthermore, water vapor diffusion was found to be increased by the insertion of the tert-butyl group, whereas water absorption diminished. The differences in gas and water vapor transport properties observed for these two polyesters were discussed on the basis of their respective molecular structures. q 2004 Elsevier Ltd. All rights reserved. Keywords: Polyesters; Poly(ethylene isophthalate)s; Poly(ethylene terephthalate) 1. Introduction It is widely known that the insertion of isophthalate units in poly(ethylene terephthalate) (PET) introduces significant structural changes and subsequent property modifications that are of academic and industrial interest [1,2]. In fact, isophthalic acid has become the most widely accepted PET modifier for packaging applications due to considerable reduction in crystallization without affecting other general properties. As a result of this, the homopolymer poly (ethylene isophthalate) (PEI) has been specifically investi- gated with the aim of elucidating its structure and to evaluate their basic properties when compared to PET [3,4]. PEI can be obtained by using the same polymerization procedure established for PET; the polymer has superior barrier properties than PET, but it exhibits a lower T g and its mechanical properties are insufficient for most of common applications. PEI is a semicrystalline polymer whose melting temperature is not well established and its structure in the solid state is not well known. Several authors have reported on such issues, however, structural data are scarce and there are some contradictions in their interpretation [3]. The content of isophthalate units in PET as a crystal- lization repressor has to be limited to less than , 5% to avoid deterioration in other polymer properties of prime importance. The utilization of isophthalic acid bearing a bulky side group has been considered as a valuable alternative [5]. Recently, the preparation and properties of random poly(ethylene terephthalate-co-5-tert-butyl iso- phthalate) copolymers, abbreviated PET t BI, has been reported for a wide range of compositions [6]. This work evidenced the strong repressing influence of the 5-tert-butyl isophthalate units on the ‘crystallizability’ of PET, and showed that the T g of the copolymer increased with the 0032-3861/$ - see front matter q 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.polymer.2004.05.028 Polymer 45 (2004) 5005–5012 www.elsevier.com/locate/polymer * Corresponding author. Tel.: þ 34-93-4016681; fax: þ 34-93-4016600. E-mail addresses: [email protected] (S. Mun ˜oz-Guerra), rude@ angel.qui.ub.es (E. Rude ´), [email protected] (J.A. Galbis).

Upload: robert-quintana

Post on 14-Jul-2016

214 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Poly(ethylene isophthalate)s: effect of the tert-butyl substituent on structure and properties

Poly(ethylene isophthalate)s: effect of the tert-butyl substituent on

structure and properties

Robert Quintanaa, Antxon Martınez de Ilarduyaa, Elisabet Rudeb, Darwin P.R. Kinta,Abdelilah Allaa, Juan A. Galbisc, Sebastian Munoz-Guerraa,*

aDepartament d’Enginyeria Quımica, Universitat Politecnica de Catalunya, ETSEIB, Diagonal 647, 08028 Barcelona, SpainbDepartament d’Enginyeria Quımica i Metal·lurgia, Facultat de Quımica, Universitat de Barcelona, Martı i Franques 1, 08028 Barcelona, Spain

cDepartamento de Quımica Organica y Farmaceutica, Facultad de Farmacia, Universidad de Sevilla, 41071 Sevilla, Spain

Received 6 January 2004; received in revised form 12 May 2004; accepted 13 May 2004

Available online 9 June 2004

Abstract

A comparative study of the two isophthalic acid deriving homopolyesters poly(ethylene isophthalate) (PEI) and poly(ethylene 5-tert-butyl

isophthalate) (PEtBI), including synthesis, crystal structure, and thermal and permeability properties, was carried out. The two polyesters

were prepared by condensation polymerization in the melt. In both cases, minor amounts of cyclic dimers were observed to form, which were

characterized by nuclear magnetic resonance and mass spectroscopy. PEI and PEtBI were thermally stable up to 400 8C and they appeared to

be semicrystalline polyesters, having their melting temperatures between 130 and 135 8C. Their glass-transition temperatures were 62 and

94 8C, respectively. The crystal structure adopted by the two polyesters seemed to consist of a regularly folded conformation, clearly different

from the almost extended conformation characteristic of poly(ethylene terephthalate). Gas permeability measurements for N2, O2, and CO2

revealed that PEtBI is more permeable to these gases than PEI, in spite of having a higher Tg: Furthermore, water vapor diffusion was found to

be increased by the insertion of the tert-butyl group, whereas water absorption diminished. The differences in gas and water vapor transport

properties observed for these two polyesters were discussed on the basis of their respective molecular structures.

q 2004 Elsevier Ltd. All rights reserved.

Keywords: Polyesters; Poly(ethylene isophthalate)s; Poly(ethylene terephthalate)

1. Introduction

It is widely known that the insertion of isophthalate units

in poly(ethylene terephthalate) (PET) introduces significant

structural changes and subsequent property modifications

that are of academic and industrial interest [1,2]. In fact,

isophthalic acid has become the most widely accepted PET

modifier for packaging applications due to considerable

reduction in crystallization without affecting other general

properties. As a result of this, the homopolymer poly

(ethylene isophthalate) (PEI) has been specifically investi-

gated with the aim of elucidating its structure and to

evaluate their basic properties when compared to PET [3,4].

PEI can be obtained by using the same polymerization

procedure established for PET; the polymer has superior

barrier properties than PET, but it exhibits a lower Tg and its

mechanical properties are insufficient for most of common

applications. PEI is a semicrystalline polymer whose

melting temperature is not well established and its structure

in the solid state is not well known. Several authors have

reported on such issues, however, structural data are scarce

and there are some contradictions in their interpretation [3].

The content of isophthalate units in PET as a crystal-

lization repressor has to be limited to less than ,5% to

avoid deterioration in other polymer properties of prime

importance. The utilization of isophthalic acid bearing a

bulky side group has been considered as a valuable

alternative [5]. Recently, the preparation and properties of

random poly(ethylene terephthalate-co-5-tert-butyl iso-

phthalate) copolymers, abbreviated PETtBI, has been

reported for a wide range of compositions [6]. This work

evidenced the strong repressing influence of the 5-tert-butyl

isophthalate units on the ‘crystallizability’ of PET, and

showed that the Tg of the copolymer increased with the

0032-3861/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.

doi:10.1016/j.polymer.2004.05.028

Polymer 45 (2004) 5005–5012

www.elsevier.com/locate/polymer

* Corresponding author. Tel.: þ34-93-4016681; fax: þ34-93-4016600.

E-mail addresses: [email protected] (S. Munoz-Guerra), rude@

angel.qui.ub.es (E. Rude), [email protected] (J.A. Galbis).

Page 2: Poly(ethylene isophthalate)s: effect of the tert-butyl substituent on structure and properties

content of these units in the copolymer. PET terpolymers

containing both isophthalate and 5-tert-butyl isophthalate

units were also investigated and found to have intermediate

properties, which could be tuned by adjusting the terpoly-

mer composition [7].

In this work we have carried out a comparative study of

the homopolyesters PEI and PEtBI, including synthesis,

structure, and properties. The objective is to estimate the

influence of the tert-butyl group on the polymerization

reaction with ethylene glycol, on the molecular confor-

mation, and on some structure-related properties, such as

melting and permeability.

2. Experimental

2.1. Materials and measurements

Dimethyl isophthalate (DMI) (99 þ %) and 5-tert-butyl

isophthalic acid (tBIA) (99 þ %) were purchased from

Sigma-Aldrich, Inc. Both comonomers were used without

further purification. Ethylene glycol (EG) (99 þ %, Sigma-

Aldrich) was reagent grade and used as received. Tetrabutyl

titanate (TBT) and dibutyl tin oxide (DBTO) were obtained

from Merck–Schuchardt and Sigma-Aldrich, respectively,

and they were used without further purification. Solvents

were all of either technical or high-purity grade and used as

received.1H and 13C NMR spectra were recorded on a Bruker

AMX-300 spectrometer at 25.0 8C operating at 300.1 and

75.5 MHz, respectively. Polyesters were dissolved in a

mixture of deuterated chloroform (CDCl3)/trifluoroacetic

acid (TFA) (8/1 v/v), and spectra were internally referenced

to tetramethylsilane. About 10 and 50 mg of sample

dissolved in 1 ml of deuterated solvent were used for 1H

and 13C, respectively. Sixty-four scans were registered for1H and 1000–10,000 for 13C NMR with 32 and 64 K data

points and relaxation delays of 1 and 2 s, respectively.

CP-MAS 13C NMR spectra were recorded in the same

spectrometer operating at 75.5 MHz with a 32-kHz spectral

width and 6K data points at a MAS rate of 4 kHz. Spectra

were acquired at 25 8C under the following conditions:

pulse width of 5 ms (908), contact time of 1.5 ms, and

repetition time of 3 s. 13C chemical shifts were externally

calibrated with adamantane (with a higher field peak

appearing at 29.5 ppm). High resolution mass spectra

(HRMS) were recorded on a Micromass AUTOSPEC Q

mass spectrometer, using the technique of Chemical

Ionization (CI) at 45 eV, and methane as reactive gas.

Samples were introduced by a solid probe at 30 8C and were

progressively heated up to 280 8C.

Intrinsic viscosities of the polymers dissolved in

dichloroacetic acid were measured using an Ubbelohde

viscosimeter thermostated at 25 ^ 0.1 8C. Gel permeation

chromatography (GPC) was carried out at 35 8C using

an o-chlorophenol/chloroform mixture (1/9 v/v) and

tetrahydrofuran as mobile phase for PEI and PEtBI,

respectively. GPC experiments were performed on a Waters

GPC system equipped with a refractive-index detector.

Molecular weights and molecular weights distributions

were calculated against polystyrene and poly(methylmeth-

acrylate) standards for PEI and PEtBI, respectively, using

the Maxima 820 software.

The thermal behavior of the polyesters was examined by

DSC using a Perkin Elmer DSC Pyris 1 instrument

calibrated with indium. DSC data were obtained from

4–6-mg samples at heating/cooling rates of 10 8C min21

under nitrogen circulation. Thermogravimetric analysis

(TGA) was carried out with a Perkin Elmer TGA-6

thermobalance at a heating rate of 10 8C min21 under a

nitrogen atmosphere. Wide angle X-ray scattering (WAXS)

experiments were accomplished on a Statton-type camera

using nickel-filtered Cu Ka radiation with a wavelength of

0.1542 nm.

Gas transport properties of PEI and PEtBI were evaluated

using amorphous films having a thickness between 20 and

40 mm and an effective membrane area of 9.1 cm2.

Permeability to nitrogen, oxygen, and carbon dioxide was

tested at 25 8C at atmospheric pressure. The films were

prepared from a 5% (w/v) polymer solution in chloroform

by evaporating the solvent under an infrared lamp. The

thickness of the film was taken as the average of a number of

measurements made with a micrometer all over the whole

area of the sheet. Permeability experiments were carried out

in a pressure-rise constant volume system, which was

described in a previous publication [6b]. Initially the

membrane, which was placed in the diffusion cell, was

evacuated from any residual gas by applying high vacuum

to both sides for at least 24 h. At time t ¼ 0 the upstream

side of the membrane was exposed to the desired gas and the

pressure on the downstream chamber was measured as a

function of time.

Water vapor diffusion coefficients and the amount of

absorbed water were obtained by dynamic vapor sorption

(DVS). Measurements were conducted at 30 8C in three

consecutive steps: 0 to 30, 30 to 60, and 60 to 85% relative

humidity, respectively, on completely dried samples, using

a surface measurement systems DVS 2085. The water vapor

diffusion coefficient of the aromatic polyesters appeared to

be independent on the relative humidity, and the reported

coefficient was taken as the average value of the three steps.

The amount of absorbed water was taken at 85% relative

humidity. The sample thickness was in the range of

200–250 mm.

2.2. Synthesis

2.2.1. Poly(ethylene isophthalate) and poly(ethylene

tert-butyl isophthalate)

DMI (20 g; 120 mmol) and EG (16.5 g; 265 mmol) in a

feed molar ratio of 1:2.2 were introduced into a three-

necked 100-ml round-bottom flask equipped with a

R. Quintana et al. / Polymer 45 (2004) 5005–50125006

Page 3: Poly(ethylene isophthalate)s: effect of the tert-butyl substituent on structure and properties

mechanical stirrer, a nitrogen inlet, and a vacuum distilla-

tion outlet. The temperature was raised to 185 8C and, after

complete homogenization of the monomer mixture,

approximately 0.6 mmol of TBT catalyst per mol of diester

was added. Transesterification was carried out under a low

nitrogen flow for a period of 5 h with the formation of

methanol. The temperature was then raised to 260 8C and

the pressure lowered to 0.5–1 mbar, and the condensation

polymerization reaction was allowed to proceed isothermally

under these conditions for 120 min. The reaction mixture

was then cooled down to room temperature, and atmos-

pheric pressure was recovered with a nitrogen flow to

prevent degradation. The solid mass was dissolved in

chloroform, and the polymer precipitated with cold diethyl

ether, collected by filtration and extensively washed with

cold methanol and diethyl ether. All samples were dried

at 60 8C under reduced pressure for a minimum of 24 h

before use.1H NMR (CDCl3/TFA (8/1 v/v), d (ppm)): 4.78 (s, 4H,

CH2), 7.58 (t, 1H, ArH-5), 8.27 (dd, 2H, ArH-4,6) and 8.71

(t, 1H, ArH-2). 13C NMR (CDCl3/TFA (8/1 v/v), d (ppm)):

63.76 (CH2), 129.31 (ArC-5), 129.54 (ArC-1,3),

131.31(ArC-2), 134.93 (ArC-4,6), and 167.35 (CO).

The same reaction system was used to prepare the PEtBI

polymer from tBIA (20 g; 90 mmol) and EG (27.9 g;

450 mmol) in a feed molar ratio of 1:5. Esterification

reaction was carried out under a low nitrogen flow for a

period of 3 h at 190 8C followed by second period of 2 h at

220 8C, with the formation of water. No catalyst was added

in this step. Polycondensation was carried out in the

presence of TBT (0.6 mmol per mol of diacid) at 250 8C

and 0.5–1 mbar for a period of 120 min. The polymer was

isolated and purified as described above for PEI polyester.1H NMR (CDCl3/TFA (8/1 v/v) d (ppm)): 1.30 (s, 9H,

CH3), 4.79 (s, 4H, CH2), 8.28 (d, 2H, ArH-4,6) and 8.54 (t,

1H, ArH-2). 13C NMR (CDCl3/TFA (8/1 v/v) d (ppm)):

30.97 (CH3), 35.22 (C(CH3)3), 63.97 (CH2), 128.96 (ArC-2),

129.65 (ArC-1,3), 132.41 (ArC-4,6), 153.50 (ArC-5), and

168.11 (CO).

2.2.2. Cyclo di(ethylene isophthalate)(cDEI) and cyclo

di(ethylene 5-tert-butyl isophthalate) (cDEtBI)

In a 250-ml round-bottom flask equipped with a

condenser were introduced 10 mmol of PEI and 125 ml of

dry 1,2-dichlorobenzene (DCB) in the presence of DBTO

(3 mol%). The cyclodepolymerization reaction was allowed

to proceed under reflux temperature (180 8C) for 10 days,

and a continuously increasing amount of precipitated

material was observed. The mixture was then cooled

down to room temperature and the DCB was evaporated

off under vacuum (20 mmHg) to dryness. A mixture of the

cyclic dimer (.90%) and higher cyclic oligomers were

deposited on a coldfinger upon sublimation at 275 8C and

20 mmHg. The cyclo(ethylene isophthalate) dimer (cDEI)

was purified by repeated washing with chloroform and dried

at 60 8C under reduced pressure for 2 days before

characterization.1H NMR (CDCl3/TFA (8/1 v/v) d (ppm)): 4.79 (s, 4H,

CH2), 7.70 (t, 1H, ArH-5), 8.35 (dd, 2H, ArH-4,6) and 8.70

(t,1H, ArH-2). 13C NMR (CDCl3/TFA (8/1 v/v) d (ppm)):

63.51 (CH2), 129.68 (ArC-5), 129.78 (ArC-1,3), 130.98

(ArC-2), 135.27 (ArC-4,6) and 166.51 (CO).

HRMS (CI): [M þ 1]þ found: 385.0923. Calcd for

[C20H17O8]þ: 385.0923.

The synthesis of cyclo di(ethylene 5-tert-butyl isophthal-

ate) was carried out from PEtBI using the same procedure as

for cDEI. In this case sublimation provided the pure cyclic

dimer making unnecessary any further purification

treatment.1H NMR (CDCl3/TFA (8/1 v/v) d (ppm)): 1.32 (s, 9H,

CH3), 4.61 (s, 4H, CH2), 8.27 (d,2H, ArH-4,6) and 8.65

(t,1H, ArH-5). 13C NMR (CDCl3/TFA (8/1 v/v) d (ppm)):

31.09 (CH3), 35.03 (C(CH3)3), 62.63 (CH2), 127.82 (ArC-2),

130.12 (ArC-1,3), 131.59 (ArC-4,6), 152.65 (ArC-5), and

165.56 (CO).

HRMS (CI): [M þ 1]þ found: 497.2171. Calcd for

[C28H33O8]þ: 497.2175.

3. Results and discussion

Ethylene glycol deriving aromatic polyesters containing

exclusively isophthalic units, either unsubstituted (PEI) or

bearing a tert-butyl group attached to the meta position

(PEtBI), were prepared by melt condensation polymeriz-

ation. The typical two-step procedure consisting of

transesterification, followed by polymerization was applied

in the two cases (Fig. 1). The methyl diester and the free

diacid monomers were used for PEI and PEtBI, respectively.

Since similar procedure was used for the two cases, a close

comparison of synthesis results regarding the chemical

constitution of the polymers could be made (Table 1). The

conclusion drawn from this comparative study is that no

significant differences in reaction results were observed for

the two systems. The higher molecular weight and narrower

polydispersity found for PEtBI is a consequence of the

partial removal of the low molecular fraction that takes

place upon precipitation with ether. Such a fractionation has

a more pronounced effect in the case of PEtBI due to the

higher solubility of this polymer in chloroform.

Fig. 1. Polymerization scheme leading to PEI and PEtBI.

R. Quintana et al. / Polymer 45 (2004) 5005–5012 5007

Page 4: Poly(ethylene isophthalate)s: effect of the tert-butyl substituent on structure and properties

The chemical structure of the polyesters was ascertained

by NMR spectroscopy. The two polymers contained the

same amount of diethylene glycol in spite that the dimethyl

ester and the free acid were the monomers used for PEI and

PEtBI, respectively. According to both viscosity and GPC

measurements, PEI and PEtBI were obtained with molecular

sizes of similar order of magnitude. However, the values

provided by GPC should not be taken as absolute because

the calibration of a polyester with polystyrene standards is

certainly not an accurate method. In fact, it is known from

studies of several authors that PS calibrated GPC measure-

ments largely overestimate the Mn’s of aliphatic polyesters

[9]. Furthermore the slight discrepancies observed by GPC,

which point out a higher molecular weight for PEtBI, should

be taken with caution, as a different hydrodynamic behavior

should be expected for the two polymers since different

solvents were used in each case. The evolution of the

polycondensation reaction was followed by viscosimetry,

which revealed that polymer size steadily increased along

the first two hours of reaction, after which an abrupt decay in

viscosity occurred. According to previously reported work

on the synthesis of PEI [10], thermal depolymerization with

the formation of cyclic oligomers will be responsible for

such change in the polymerization trend. In fact, a certain

amount of oligomeric material could be recovered from the

vessel wall when the polymerization reaction was left to

proceed for long times, after which low molecular weight

polymers were recovered ðh ¼ 0:35–0:45Þ:

The formation of cyclic oligolactones in the poly-

condensation reaction of ethylene glycol and isophthalic

acid has been reported by several authors [11,12]. However,

there is no agreement upon what is the size of the main

product present in this oligomeric mixture. In order to help

to identify and characterize the oligomers produced in the

polymerization reactions leading to PEI and PEtBI, we

performed the thermal depolymerization of the two

polyesters dissolved in o-dichlorobenzene catalyzed by

DBTO (Fig. 2). A combined analysis of NMR and mass

spectroscopy revealed unequivocally that the cyclic dilac-

tone was the oligomer predominantly formed in both cases.

This is in accordance with the results obtained by Hodge

et al. [12], but in disagreement with the work of Lim et al.

[11], who reported the trimer to be the main product present

in the oligomeric material generated in the polymerization

of PEI.

The thermal properties of PEI and PEtBI as well as their

powder X-ray diffraction spacings obtained by WAXS, are

given in Table 2, where similar data for PET have been

included for comparison. Moreover, thermal and diffraction

data for their corresponding cyclic dimers were also

measured because of their significance in the interpretation

of the structure of the polymers.

The most striking result is that both PEI and PEtBI have

almost the same melting temperature ðTmÞ;which turns to be

about 120 8C lower than that of PET, and that their melting

enthalpies are also significantly lower than that of PET.

Furthermore, the melting of the cyclic dimers happened

very sharply at a very high temperature (between 320 and

330 8C), which is in accordance with expectations for highly

crystalline low molecular weight compounds (Fig. 3). In a

previously reported work on PEI, we reported a melting

temperature around 240 8C for this compound [6a]. Other

authors however interpreted this temperature as correspond-

ing to the melting of cyclic compounds depleted by the

presence of the polymer [3b]. We have studied the melting

behavior of mixtures of PEI and cDEI containing from 1 up

to 10% of cyclic material and observed that deviations in Tm

for both the polymer and the dimer were less than 10 8C.

According to our present results, the PEI polymer melting at

,130 8C should be due to a crystal form that has not been

previously reported for PEI. Along our present work we

have learned that the crystal form reported to melt above

200 8C is very elusive, the melting peak appearing at 240 8C

being occasionally observed with very small enthalpy in

samples of uncertain purity. The possibility that such peak

could arise from melting of contaminating oligomers cannot

be discarded.

Table 1

Synthesis results and solubility

Polymer tpa DEGb [h]c GPC Solubilityd

Mne Mw

e PDe Acetone THF Chloroform Toluene

PEI 120 2.3 0.61 38,700 75,900 2.0 2 2 þþ 2

PEtBI 120 2.3 0.65 58,900 97,900 1.7 þ þþ þþ þ

a Polycondensation time (min).b Diethylene glycol content (mol%) calculated from 1H NMR spectra.c Intrinsic viscosity (dl g21) measured in dichloroacetic acid at 25 8C.d Estimated according to the method of Braun [8]: (2) insoluble, (þ ) soluble on warming at 100 8C or at boiling point, (þþ ) soluble at room temperature.

THF, tetrahydrofuran.e Number- and weight-average molecular weights and polydispersity determined by GPC.

Fig. 2. Cyclodepolymerization of PEI and PEtBI.

R. Quintana et al. / Polymer 45 (2004) 5005–50125008

Page 5: Poly(ethylene isophthalate)s: effect of the tert-butyl substituent on structure and properties

Powder X-ray diffraction patterns of PET, PEI, and

PEtBI are compared in Fig. 4, and the spacings observed

therein are listed in Table 2. Both PEI and PEtBI patterns

display diffraction rings greater in number and sharper than

those observed in the case of PET, revealing the occurrence

of larger and better formed crystals in the isophthalate

polyesters. What is most outstanding in these patterns

however is the presence in the inner region of highly intense

rings corresponding to long interplanar spacings without

parallelism in the diffraction pattern of PET. Such spacings

strongly suggest the existence of a regularly folded

conformation with a large molecular diameter, which

would give rise to a significant enlargement of the interchain

packing distances. Although indexing is not feasible at this

stage, it can be reasonably concluded that the type of crystal

structure adopted by the isophthalate polymers must be

different from that of PET, which is known to consist of a

triclinic lattice with chains in an almost extended confor-

mation [13]. It should be remarked here that PET

copolymers containing either isophthalic or 5-tert-butyl

isophthalic units retain the PET crystalline structure with

crystallites consisting exclusively of terephthalic units [6]. It

seems therefore that a fully homogeneous constitution is

required by these polyesters to take up a regular

conformation.

The interpretation of the PEI and PEtBI melting at

,130 8C as consisting of a crystal form with the polymer

chain arranged in a regularly folded conformation was

further supported by CP-MAS 13C NMR analysis. Such

helical structure would imply a more complex atomic

topology than the extended conformation with certain

chemical groups being arranged in more than one chemical

environment and therefore being discernible by NMR. In

Fig. 5, the 50–170 ppm 13C NMR spectral window is

compared for the three polyesters in the solid state. These

spectra of the poly(isophthalate)s are the difference spectra

that result by subtracting the spectrum arising from a totally

amorphous sample to that obtained from the semicrystalline

sample. They should be taken therefore as due exclusively

to the crystalline phase of the polymer. By this means,

possible differences arising from amorphous and crystal

phases of the same polymer are suppressed. The splitting in

the signal of the oxyethylenic unit observed in the spectra of

PEI and PEtBI is taken as indicative therefore of the

existence of two magnetically distinguishable methylene

groups in the crystal phase of the polyesters made of

isophthalate units. On the contrary, a single type of

methylene seems to be present in the chain of PET

according to what should be expected from its confor-

mation. The conformational complexity of the isophthalate

chains is also evidenced by comparing the signal appearing

around 165 ppm, which arises from the carbonyl carbon. In

this case however the splitting of the signal is not clearly

resolved for the case of PEtBI. Note that Kobayashi and

Hachibosni [3c] studied the crystal structure of a stretched

fiber of PEI having a melting point of 240 8C and proposed

for this polymer a triclinic crystal structure with a geometry

closely resembling the crystal structure of PET. Although

we have not been able to characterize this form, the presence

of a crystal dimorphism in the poly(ethylene isophthalate)s

is perfectly assumable.

Permeability measurements of the PEI and PEtBI

Table 2

Thermal properties and X-ray spacings of polyesters and cyclic dimers

Polymer DSC TGA X-ray spacings

Tga (8C) T

m

b (8C) DHm (J g21) Tdc (8C) Tds

d (8C) RWe (%) dhkl (nm)

PEI 62 132 30.6 409 433 12 1.29, 0.65 0.49, 0.43, 0.38, 0.34

PEtBI 94 134 20.8 410 441 12 1.82, 0.92, 0.61 0.53, 0.46, 0.43, 0.34

cDEI – 332 76.4 378 432 18 – –

cDEtBI – 323 86.7 393 432 17 – –

PET 82 256 41.2 406 436 20 0.52, 0.39, 0.34, 0.28

a Second scan recorded from the molten sample at 10 8C min21.b First scan from powder precipitated sample.c Temperature at which a 10% weight loss was observed in the TGA traces recorded at 10 8C min21.d Temperature of maximum degradation rate.e Remaining weight at 550 8C.

Fig. 3. DSC from powder samples (first scan) of PEI and PEtBI polyesters

and their corresponding cyclic dimers cDEI and cDEtBI.

R. Quintana et al. / Polymer 45 (2004) 5005–5012 5009

Page 6: Poly(ethylene isophthalate)s: effect of the tert-butyl substituent on structure and properties

homopolyesters were carried out in a pressure-rise constant

volume system by the conventional time-lag method.

Downstream pressures were plotted against time, and from

the extrapolation of the steady-state part of the curve, the

intercept with the time axis, the time-lag u was obtained.

The diffusion coefficient D was calculated as D ¼ L2=6u;L

being the thickness of the film. The permeability constant P

is calculated from the slope of the steady-state part of

the p– t plot using the following equation,

P ¼TSTP

pSTP

VL

ATpup

dpðtÞ

dt

where V is the volume of the system, p the pressure recorded

versus time t; pup is the applied upstream pressure, T is the

absolute temperature, A is the effective area of the film, and

TSTP and pSTP are the temperature and pressure under

standard conditions. The solubility coefficient S is then

calculated from the diffusion and the permeability coeffi-

cients as S ¼ P=D: The plot of upstream pressure versus

time is depicted in Fig. 6 for the three gases and the two

polymers studied in this work.

Diffusion, solubility, and permeability coefficients deter-

mined for the transport of N2, O2, and CO2 gases through

PEI and PEtBI films are compared in Table 3, where data for

PET have been included for reference. As expected from the

disturbing packing effect caused by the presence of the tert-

butyl side group, all three parameters were observed to be

substantially larger in PEtBI when compared to PEI, for the

three gases that were investigated. Such effect has been

previously reported for other families of polymers, such as

polysulfones, polyamides, and other polyesters [14]. We

have recently studied the influence of the incorporation of

5-tert-butyl isophthalate units in PET [6b]. In that case, the

diffusion and solubility coefficients appeared to be affected

in opposite sense so that the resulting influence of

composition on permeability was not too severe. In the

case of PEI, all the three transport parameters changed in the

same sense when the tert-butyl isophthalate units were

incorporated in the polyester chain. It was striking the fact

that the barrier properties of the polymer decreased in spite

that the Tg increased, a behavior that could be explained by

the difference in the amount of free volume that is available

for the chain segment and for the gas molecule. Never-

theless, the selectivity for the mixtures O2/N2 and CO2/O2

(a*) were essentially the same as in the unsubstituted PEI.

The water vapor diffusion coefficient and the amount of

absorbed water were determined for both polyesters by

dynamic vapor sorption measurements. The diffusion

Fig. 4. Powder X-ray diffraction patterns of PET (a), PEI (b) and PEtBI (c).

Fig. 5. CP-MAS 13C NMR spectra of the indicated polyesters. (*spinning

side bands).

Fig. 6. Pressure vs. time curves for PEI and PEtBI for the three indicated

gases.

R. Quintana et al. / Polymer 45 (2004) 5005–50125010

Page 7: Poly(ethylene isophthalate)s: effect of the tert-butyl substituent on structure and properties

coefficient of PEtBI increased by one order of magnitude,

when compared to PEI. This is in agreement with the

diffusion parameters obtained for the transport of N2, O2,

and CO2 gases. However, the amount of absorbed water was

significantly reduced when replacing the isophthalate units

for the tert-butyl substituted isophthalate ones. This result

faithfully reflects the high hydrophobic character of the

PEtBI polymer chain, which exerts a pronounced negative

influence on water absorption. It is apparent that such

hydrophobic effect outweighs the influence that could be

expected from the higher amount of free volume provided

by tert-butyl group when compared to PEI.

4. Conclusions

A comparative study of poly(ethylene isophthalate)

(PEI) and poly(ethylene 5-tert-butyl isophthalate) (PEtBI),

including synthesis, crystal structure, and thermal and gas

transport properties, was carried out. No significant

differences in synthesis results were observed when

dimethyl isophthalate was replaced by 5-tert-butyl iso-

phthalic as the diacidic monomer. Tg increased with the

presence of the tert-butyl substituted units but Tm was

essentially the same for the two poly(isophthalate)s and for

both cases, much lower than for PET. Both PEI and PEtBI

appear to adopt a crystal form with chains presumably

arranged in a helical conformation with topological features

clearly different from those described for the triclinic crystal

form of PET. Barrier properties were found to decay in

PEtBI with respect to PEI, but selectivity appeared to be

essentially unchanged. Water vapor absorption greatly

diminished with the presence of the tert-butyl group. It

was concluded that the free volume in PEtBI accessible to

small particles is higher than in PEI, in spite of having

higher Tg: On the other hand, the higher hydrophobic

character of the PEtBI polyester induced by the tert-butyl

group is responsible for the lower water uptake that PEtBI

has with respect to PEI.

Acknowledgements

Part of this work has been supported by MCYT with

projects MAT2002-04600-CO2, MAT2003-06955-CO2

and QUI99-0533. Financial support received from Catalana

de Polımers S.A. (Barcelona, Spain) is also gratefully

acknowledged. The authors are indebted to Dr J.J Bou for

his assistance with GPC experiments.

References

[1] (a) Ha WS, Chun YK, Jang SS, Rhee DM, Park CR. J Polym Sci Part

B: Polym Phys 1997;35:309. (b) Lee SW, Ree M, Park CE, Jung YK,Table

3

Permeabilityofgases

andwater

absorptionin

PET,PEI,andPE

t BI

Polymer

Tg(8C)

P·1015

(cm

3(STP)·cm

·cm

22·s21·Pa2

1)

ap O

2=N

2ap C

O2=O

2D·109

(cm

2·s21)

S·106

(cm

3(STP)·cm

23·Pa2

1)

D·109

(cm

2·s21)

A (%)

N2

O2

CO2

N2

O2

CO2

N2

O2

CO2

H2Oa

H2Oa

PET

78

0.68

4.5

21.0

6.6

4.7

1.1

4.6

1.5

0.62

0.98

14.0

19.4

0.89

PEI

62

0.13

0.90

2.2

6.9

2.4

0.25

1.1

0.24

0.52

0.82

9.2

5.4

0.80

PE

t BI

94

2.8

19.5

64.7

7.0

3.3

4.4

6.7

3.6

0.64

2.9

18.0

47.0

0.47

Gas

transportcoefficientswereobtained

atroom

temperature

(258C

)andatmospheric

pressure.

aWater

vapordiffusionandabsorptioncoefficientsobtained

at308C

and85%

ofrelativehumidity.

R. Quintana et al. / Polymer 45 (2004) 5005–5012 5011

Page 8: Poly(ethylene isophthalate)s: effect of the tert-butyl substituent on structure and properties

Park CS, Jin YS, Bae DC. Polymer 1999;40:7137. (c) Li B, Yu J, Lee

S, Ree M. Eur Polym J 1999;35:1607.

[2] Kint DPR, Martınez de Ilarduya A, Sansalvado A, Ferrer J, Munoz-

Guerra S. J Appl Polym Sci 2003;90:3076.

[3] (a) Conix A, Van Kerpel R. J Polym Sci 1959;40:521. (b) Yamadera

R, Sonoda T. J Polym Sci Polym Lett 1965;3:411. (c) Kobayashi S,

Hachiboshi M. Rep Prog Polym Phys Jpn 1970;13:157.

[4] (a) Light RR, Seymour RW. Polym Engng Sci 1982;22:857.

(b) Polyakova A, Liu RY, Schiraldi DA, Hiltner A, Baer E. J Polym

Phys 2001;39:1889.

[5] (a) Shaltter MJ, Butler JC. US Patent 2,794,794; 1957. (b) Fenoglio

DJ, Foster JJ. J Polym Sci Part A: Polym Chem 1990;28:2753.

[6] (a) Kint DPR, Martınez de Ilarduya A, Munoz-Guerra S. J Polym Sci

Part A: Polym Chem 2001;39:1994. (b) Kint DPR, Rude E, Llorens J,

Munoz-Guerra S. Polymer 2002;43:4749.

[7] Kint DPR, Martınez de Ilarduya A, Alla A, Munoz-Guerra S. J Polym

Sci Part A: Polym Chem 2003;41:124.

[8] Braun D, Chedron H, Kern W. Praktikum der Makromolekularen

Organischen Chemie Alfred. Heidelberg, Gemany: Alfred Huthig

Verlag; 1966.

[9] (a) Pasch H, Rode K. J Chrom A 1995;699:21. (b) Kricheldorf HR,

Eggerstedt S. Macromol Chem Phys 1998;199:283. (c) Kowalski A,

Duda A, Penczek S. Macromolecules 2000;33:7359.

[10] Wick G, Zeitler H. Angew Makromol Chem 1983;112:59.

[11] Lim BH, Kwon SH, Kang EC, Park H, Lee HW, Kim WG. J Polym

Sci Part A: Polym Chem 2003;41:881.

[12] Hall AJ, Hodge P, McGrail CS, Rickerby J. Polymer 2000;41:1239.

[13] Daubeny R, Bunn C, Brown C. Proc R Soc Lond 1954;A226:531.

[14] Pitxon MR, Paul DR, Polymeric gas separation membranes, vol. 83.

Boca Raton, FL: CRC Press; 1994. p. 153.

R. Quintana et al. / Polymer 45 (2004) 5005–50125012