COVER SHEET

This is the author version of article published as:

Nagle, Dylan J. and Celina, Mathew and Rintoul, Llewellyn and Fredericks, Peter M. (2007) Infrared microspectroscopic study of the thermo-oxidative degradation of hydroxy-terminated polybutadiene/isophorone diisocyanate polyurethane rubber. Polymer Degradation and Stability 92(8):pp. 1446-1454.

Copyright 2007 Elsevier Accessed from http://eprints.qut.edu.au

1

Infrared Microspectroscopic study of the thermo-oxidative degradation of hydroxy-terminated polybutadiene/isophorone diisocyanate polyurethane

rubber

Dylan J. Nagle1, Mathew Celina2, Llewellyn Rintoul1 and Peter M. Fredericks1*

1School of Physical & Chemical Sciences, Queensland University of Technology, GPO Box 2434, Brisbane, Qld 4001, Australia

2Sandia National Laboratory, Albuquerque, NM 87185-1411, USA Abstract Hydroxy-terminated polybutadiene/isophorone diisocyanate (HTPB-IPDI) polyurethane rubber that was aged in air at elevated temperatures has been studied by infrared microspectroscopy. Spectra were collected in transmission mode on microtomed samples. Analysis of sets of spectra taken across the sectioned material showed that most of the degradation was occurring in the polybutadiene part of the polymer and that the urethane linkage was essentially unchanged. The trans isomer of the polybutadiene appears to be preferentially degraded compared with the vinyl isomer. The IR technique does not provide significant information about the cis isomer. The IR spectra indicated that likely degradation products included acids, esters, alcohols, and small amounts of other products containing a carbonyl functional group. Band area ratios, supported by a principal components analysis, were used to derive degradation profiles for the material. These profiles were steep-sided indicating an oxygen diffusion limited process.

*Corresponding author. Tel.: +61 7 3138 2341; fax: +61 7 3138 1804. E-mail address: p.fredericks@qut.edu.au (P.M. Fredericks)

2

1. Introduction Polyurethanes refer to a broad group of elastomers that may contain just one common

element; a urethane linkage [1]. One of the distinguishing features of polyurethanes is

the hard and soft segments that can be found in the macromolecule. These differences

in hardness are due to crystalline and amorphous regions in the elastomer structure. In

crystalline regions, the polymer is tightly packed in a regular repeating structure,

while the amorphous regions consist of random, intertwined coils [2]. These

crystalline and amorphous regions are held together by hydrogen bonding and van der

Waals forces. Normally, diisocyanates are used to create urethane linkages in the

manufacture of polyurethanes [3].

The addition of isocyanate linkages to the polyurethane elastomer results in an

increase in hardness and modulus of the elastomer [1]. This is coupled with a gradual

decrease in tensile strength and elongation at break with increasing diisocyanate

concentration.

Polyurethane-based adhesives and binders can demonstrate a wide range of physical

properties depending upon the type of chain extender and diisocyanate. Polyurethane

adhesives bond to substrates such as glass, wood, plastics and ceramics [2]. The high

polarity of polyurethane elastomers encourages bonding to substrates through

intermolecular forces (Van der Waals, dipole forces and hydrogen bonding). Strongest

bonding can be obtained using aromatic diisocyanates such as toluene diisocyanate

(TDI), while aliphatic isocyanates such as hexamethylene diisocyanate (HDI) provide

light and colour stability. Furthermore, the regions of varying hardness in

polyurethane adhesives provide flexibility, adding to the usable lifetime of the

adhesive[3]. Binders such as cured hydroxy-terminated polybutadiene/isophorone

diisocyanate (HTPB/IPDI) used in solid propellants also fall into this category.

Fourier transform IR techniques such as transmission, reflection and ATR (attenuated

total reflection) are commonly used in the investigation of reaction [4] and

degradation [5] kinetics, polyurethane characterisation [6-9], and hydrogen bonding

of the urethane carbonyl [10]. Resonance Raman scattering has been employed to

determine information regarding radiation absorption of chromophores in some

3

polyurethanes [11], and the reduced fluorescence problems associated with FT-Raman

have allowed researchers to study polyurethanes occurring in natural systems [12].

Vibrational spectroscopy is also commonly used to follow polyurethane degradation

experiments [13-15]. A series of papers published by Wilhelm and Gardette et al. [16-

18] in the late 1990s investigated photo-oxidative degradation of aliphatic and

aromatic polyurethanes using IR spectroscopy. Although the study was performed

primarily on the effects of photo-oxidation, the temperature at which degradation was

performed was slightly elevated (60°C). Thermo- and photo-oxidation processes often

follow similar pathways [19], and thus chemical changes that are pertinent to one

form of oxidation are commonly relevant to the other oxidation process.

The first paper published by Wilhelm and Gardette [16] concerned the photo-

oxidation of MDI (4,4’-diphenylmethane) diisocyanate-cured aliphatic poly(ester

urethane). Degradation was found to occur at both the polyester and the urethane

linkage components. The second paper [17] examined the degradation of the aliphatic

polyurethane poly(ether urethane). An observation made in the first paper [16] that

soft segments in polyurethanes (comprised of the chain extenders) are the most

susceptible to photo-oxidation was experimentally confirmed in the poly(ether

urethane) study. Hydroxy-terminated chain extender polyurethanes form hydroperoxy

radicals with relative ease, which then react further with the polymer. Ether groups

are catastrophically destroyed during the photodegradation process, reacting to

produce formates as the main photoproducts, which react further to produce low-

molecular weight by-products. Degradation of the urethane linkage did occur, albeit

not to the same extent as degradation of the polyether components. The third paper in

the series [18] concerned the photodegradation of aromatic polyurethanes. The

aromatic urethane linkages were found to undergo oxidation to form chromophoric

degradation products.

Elastomeric binders based on functionalized polybutadienes can be cured with various

crosslinkers, using hydroxyl-isocyanate (PU linkage), carboxylate-epoxy or

carboxylate-aziridine reactions. Only small amounts of crosslinkers (on the order of

10%) are required to produce suitable rubbers and hence the mechanical and chemical

properties are largely dominated by the soft polybutadiene segments. The mechanical

properties of such polyurethanes depend largely on the degree of crosslinking, which

4

in turn is dependant on the NCO/OH ratio [20]. Due to the well-known sensitivity of

polybutadiene to oxidative degradation and the utilization of these materials in

defence related applications, considerable efforts have been devoted to better

understand the long term aging and degradation mechanisms of these materials [21-

25]. Thermal aging of HTPB/IPDI materials has been studied in great detail with an

emphasis on oxidation sensitivity and chemical-mechanical property correlations [26] ,

degradation chemistry [27], NMR polymer chain mobility relaxation experiments

[28,29] chemical derivatization approaches [30] [and the development of more

suitable techniques for lifetime prediction purposes [31,32], As a brief summary,

these materials are surprisingly sensitive to oxidation despite the addition of

significant stabilization, and undergo thermal crosslinking, densification, hardening

and display a general loss of their required elastomeric properties during aging.

Detailed knowledge of chemical degradation processes are required, an area where IR

spectroscopic analysis holds significant promise and may have been underutilized.

To better understand the degradation chemistry Harris et al. [27] performed studies on

the aging of HTPB/IPDI polyurethane rubber using 17O isotopic labelling of

degradation products. Polyurethane samples (containing the antioxidant Vanox

MPBC) were aged in an oven at 80°C for up to 239 days. The aged samples and

degradation products were characterised using 17O NMR, 13C NMR, 1H NMR and

spin diffusion experiments. Before aging, approximately one half of the carbons of the

polymer were unsaturated; however, following thermo-oxidation, the number of

unsaturated carbons reduced dramatically. Hydroxylated by-products were detected

via 17O NMR, which was in agreement with previous studies [33]. Although these

alcohol by-products were dominant, making 60% of products formed, other by-

products are also observed, including acids, esters and gases (CO2 and CO). The

gaseous products were created by conversion of other by-products at high

temperatures. Harris et al. found that the majority of degradation occurred in the

butadiene segment of the polymer. In 1989 Adam et al. [34] investigated the

photooxidation of polybutadiene via infrared spectroscopy. Degradation of different

isomeric compositions (trans, cis and vinyl) were found to be similar. Bands in the

infrared spectrum at 1696 cm-1 and 1710 cm-1 were considered to derive from ketones

and acids respectively. In 1999 Ahlblad et al. [23,24] reported the results of a

chemiluminescence study on the oxidative degradation of polybutadiene.

5

Embrittlement of the rubber occurred as a result of accelerated aging tests, which was

attributed to crosslinking and the formation of a secondary network. A significant

difference was found in the level of oxidation at the surface compared to the bulk.

Most solids experience heterogeneous oxidation as ambient oxygen cannot permeate

into and react with the sample bulk as readily as with the surface layers. The reasons

for the development of diffusion limited oxidation profiles, or DLO, due to the

balance of oxygen permeation versus oxidation rate of the material has been discussed

several times in the literature [35-38]. Adam et al. [34] also found that oxidation

decreased with increased distance from the surface of a polybutadiene sample due to

DLO effects. Oxidation profiles of bulk rubber samples can be constructed to

demonstrate the influence of DLO [24]. These profiles typically plot an oxidation

indicator, such as modulus or carbonyl band intensity in the infrared spectrum, as a

function of sample depth.

This work presents an IR spectroscopic analysis of HTPB/IPDI polyurethane that has

undergone accelerated aging at elevated temperatures in the presence of oxygen. The

spectroscopic measurements are evaluated for the presence of oxidation profiles to

investigate any DLO effects, and to determine the nature of the oxidation products

involved during the degradation process. Chemometrics in the form of principal

component analysis (PCA) has also been employed to aid in spectral analysis.

6

2. Experimental

Materials

The polyurethane was composed of HTPB cured with IPDI, for which the HTPB

component was known to consist of approximately 25% cis, 25% vinyl and 50% trans

polybutadiene isomers as reported by the manufacturer. Thermal aging of

polyurethane strips approximately 6mm wide cut from rubber sheets (~2mm

thickness) was carried out in temperature controlled (±1 °C stability), commercial air-

circulating aging ovens at various temperatures under partial oxygen pressure

equivalent to atmospheric conditions in Albuquerque.. The aging details of the

samples are given in Table 1. The aged elastomers were cross-sectioned and then

cooled with liquid nitrogen, after which a thin strip was microtomed to approximately

20 μm thickness and placed on a BaF2 disc for transmission IR spectroscopy

measurements.

Transmission infrared spectroscopy

IR spectra were collected using a Nicolet 870 Nexus Fourier transform infrared (FT-

IR) system including a Continuμm™ IR microscope equipped with a liquid-nitrogen-

cooled MCT detector (Nicolet Instrument Corp., Madison, WI). Control of the

instrument, data collection and mapping were conducted (or performed) with the

OMNIC 5.2 and Atlμs 1.1 software packages (Nicolet Instrument Corp., Madison,

WI). Spectra were manipulated with the GRAMS/32 6.00 software package (Galactic

Corp., Salem, NH.). Spectra were plotted with the Origin 6.1 software package

(OrginLab Corp., Northampton, MA, USA).

To measure line maps, spectra were collected contiguously across the sectioned face

of the sample. The aperture was set at 40 μm (in the direction of travel) x 120 μm to

allow an optimum balance between spatial resolution and signal strength. A step size

of 40 μm was used, so that typically 50 spectra were taken over an approximately

2mm cross-sectioned face (Fig 2). Spectral collection was under computer control so

that after taking one spectrum the microscope stage was driven in the XY plane to the

next measurement position. Spectra were collected in the spectral range 4000 to 650

cm-1, using 128 scans and 4 cm-1 resolution. The background was rerun every 5th

7

spectrum. The measurement time per spectrum was about 1 minute, and a complete

line map took approximately 1 hour.

Some edge effects were encountered when taking transmission spectra of the

cryogenically microtomed polyurethane samples. The very soft and elastic

consistency of the relatively undegraded samples caused many samples to have a

poorly defined edge, and in some cases the polyurethane tended to bunch up at the

edges, despite carrying out the microtoming at low temperature. In these instances

optical distortions and the increased thickness due to bunching created artefacts and

over-absorption in the infrared spectra measured at the sample margins.

Chemometrics

Chemometric analyses were performed using the GRAMS/32 6.00 software package

(Galactic Corp., Salem, NH.). The data was mean centred, and all spectra were

normalised to equal amide II integrated absorbance (1550 cm-1 to 1500 cm-1) prior to

PCA analyses. The region between 2650 cm-1 and 2000 cm-1 does not contain any

relevant chemical information, and was omitted from the analysis to prevent

atmospheric CO2 absorption bands from affecting the result.

8

3. Results and discussion Spectra

Visual inspection of the spectra measured at the exterior (A) and interior (B) of aged

HTPB/IPDI polyurethane in Figs 3 and 4 show only subtle differences. Therefore

spectral manipulation must be performed to highlight any degradation-related changes

occurring in the sample.

Spectral subtraction is a useful method for enhancing small differences between

spectra. Use of a reference band when subtracting one spectrum from another

emphasises relative variations. An absorption band suitable as a reference should

display minimal change between the spectra – it is a constant that other absorptions in

the spectra can be measured relative to. HTPB/IPDI polyurethane has two distinct

segments – the urethane linkage, and the polybutadiene chain. Based on relevant

literature [16-18] it is expected that the polybutadiene segment should initially

degrade more significantly than the urethane linkage, and the constancy of the amide

II and III absorption bands associated with the urethane linkage in our polyurethane

sample reflect this. On this basis spectral subtraction has been performed relative to

the intensity of the amide II (1517 cm-1) and III (1236 cm-1) absorption bands.

Infrared band assignments based on previous literature [16-18,34,39] for the

polybutadiene segment and the urethane linkage are given in Tables 2 and 3,

respectively.

OH region

The subtraction spectrum in Fig 3 shows an increase in absorption in the OH

stretching region. This is a broad absorption between 3600 and 3200 cm-1. The

increase of absorption intensity in this region with extended aging conditions can be

seen in Fig 5.

Other researchers have pointed to hydroxyl groups contributing significantly in

degradation product formation [1,16,27]. The broad nature of this band and the

9

increasing absorption intensity with further aging can be confidently attributed to

degradation products containing alcohol and carboxylic acid functional groups.

CH absorption

The CH stretching region in Fig 3 and CH wag in Fig 4 show only negative bands in

the difference spectrum. This points to an overall loss of CH absorption relative to the

urethane linkage bands.

The CH stretching bands of the vinyl isomer at 3074 cm-1 and 2977 cm-1 display some

loss of absorption in the difference spectrum (Fig 3C). The trans isomer bands

however at 2918 cm-1 and 2845 cm-1 experience a much greater loss of absorption. It

is clear that there is initially significantly less vinyl C=C than trans present in the

polyurethane, based on the intensities of the relative absorptions (approximately twice

as much trans isomer assuming similar absorptivities). This is supported by the

material specifications stated above in the Materials section. The decrease in

absorption of the CH stretching vibration is expected to be complemented by a

decrease in CH deformation bands. The appearance of negative bands in the

difference spectrum (Fig 4C) at around 1440 (CH2 def.), 993, 967 (CH wag) and 912

cm-1 (CH2 wag) confirms that this is indeed the case.

The CH bands below 1000 cm-1 provide more specific information regarding the loss

of the different types of CH. The three isomers of polybutadiene are shown in Fig 6.

The CH and CH2 wag bands of the vinyl moiety at 993 cm-1 and 912 cm-1 respectively

show a significant reduction in intensity with degradation. However, the trans isomer

shows a more significant change than the vinyl isomer. The C=C stretching vibration

at 1640 cm-1 of the vinyl isomer shows only a relatively small degree of change with

degradation in the difference spectrum.

It is a little more difficult to give a definitive statement regarding the relative loss of

cis C=C. Any absorption would be expected to appear in the region of 730-650 cm-1

[39]. The relatively weak absorption below 750 cm-1 does not allow an unambiguous

insight into the changes in CH wag of the cis isomer, as mentioned earlier. Instead,

information must be obtained from the CH stretching and deformation bands. The cis

C-H stretching absorption at 3004 cm-1 shows a significant loss of intensity with

10

degradation. This loss is comparable to that noted for the trans isomer band, which

suggests that the trans and cis isomers degrade to a similar degree. Overlapping of

absorption bands in the CH deformation region make it hard to distinguish between

cis and trans isomers.

Curve fitting

Curve fitting of the CH wag region between 1020 cm-1 and 890 cm-1 was performed

to further establish the effect of degradation on the trans and vinyl isomers. It was

hoped that a comparison of the area under the CH wag bands in degraded and

undegraded samples would give a quantitative result, although over-absorption at the

low wavenumber end of the spectrum made it difficult to estimate some band ratios.

A comparison of band areas relative to the trans 967 cm-1 band for some samples is

shown in Table 4. A trend of increasing vinyl-CH wag area (993 cm-1 and 912 cm-1) to

trans-CH wag area (967 cm-1) ratio with oxidation could be observed, implying that

the trans-CH wag absorption decreases relative to the vinyl isomer absorption with

degradation. This trend of increasing vinyl-CH wag absorption to trans-CH wag

absorption supports preferential trans isomer degradation observed in the spectral

subtraction results.

The carbonyl stretching region

The presence of the C=O stretching vibration of the urethane functional group

significantly masks carbonyl signatures arising from degradation products in the C=O

stretching region. There are, however, a number of studies that have been published

detailing possible degradation products of polyurethanes, similar to that being

investigated here, which include a carbonyl group. Relevant information will be

briefly presented, followed by an analysis of the data collected in this study.

The carbonyl stretching vibration often appears as a complex absorption band

consisting chiefly of two overlapping bands [40]. These bands represent free C=O and

hydrogen-bonded C=O. Queiroz et al. [41] assigned the free C=O of poly(propylene

oxide)/polybutadiene polymers at a wavenumber 20 cm-1 higher than H-bonded C=O

(1730 cm-1 compared with 1710 cm-1). The appearance of an H-bonded C=O

11

stretching absorption at slightly lower wavenumbers than a free C=O absorption is a

common trend in the literature [16-18,42-44].

It is well known that polyurethanes generally degrade to produce oxygenated products

such as ketones, carboxylic acids, esters and small molecules [23]. Accelerated aging

experiments have been performed on HTPB/IPDI polyurethane [27] using 17O NMR

and GC analyses. Acid and ester degradation products combined to total nearly 20%

of the products formed, with alcohols being the most common degradation product.

The overlapping twin bands of the free and hydrogen-bonded C=O stretching

vibrations can be observed in Fig 4 at approximately 1725 cm-1 and 1712 cm-1

respectively. Spectral subtraction in the carbonyl stretching region of a spectrum

taken from the exterior of the sample from an interior spectrum yields the difference

spectrum, shown in Fig 7.

A small increase in absorption can be seen in the subtraction spectrum (Fig 7C) at

higher wavenumbers than the urethane C=O stretching bands, indicating the

formation of possible new carbonyl C=O stretching vibrations. One band has a

maximum at 1742 cm-1 in the difference spectrum, while a second absorption of lower

intensity is also visible in this region at 1775 cm-1. This indicates the formation of

some degradation-related products in small concentrations. The absorption with a

maximum at 1742 cm-1 is assigned to an ester [39], while the absorption at higher

wavenumbers could possibly be indicative of degradation products such as anhydrides,

lactones, cyclic esters or ketones.

Additional absorption can be seen in the difference spectrum in Fig 7 with a

maximum at around 1660 cm-1. This absorption extends from around 1720 cm-1 to

1550 cm-1. A previous study [27] found that NMR evidence pointed to the formation

of acids as a result of degradation of HTPB/IPDI polyurethane. This degradation-

related absorption is attributed to the presence of acid salt degradation products.

Absorption in the carbonyl region continues to increase with more extensive aging of

the polyurethane. Fig 8 compares different samples under various aging conditions.

The appearance of absorption bands in this region following accelerated aging can be

12

observed in all samples. Extensive aging (Fig 8A) results in a significant broadening

of the urethane C=O absorption as shoulders appear due to degradation product

formation. This broad absorption extending from 1850 cm-1 to 1550 cm-1 is probably

due to numerous degradation products absorbing in this region.

Degradation Profiles

Comparison of the area under bands that change with degradation over the

polyurethane cross-section provides degradation information in a spatial context. The

need for a reference band when comparing band areas in order to normalise the data

has been discussed earlier, and for this analysis the amide II band was again chosen as

an internal reference. Plots of the area profiles of the CH wag band at 963 cm-1 (trans-

polybutadiene) compared with that of the CH wag band of the vinyl isomer at 912 cm-

1, normalised to the amide II band, are shown in Fig 9. The plots are to scale, and two

outliers have been removed. These outliers are due to the superimposition of a sine

wave in the infrared spectrum caused by interference fringes.

It is clear from this comparison that the curvature at the edges of the ‘U’ shape of the

vinyl CH2 wag profile is not as pronounced as the trans isomer profile. It can be

inferred from this that the degradation is not as severe in the vinyl isomer; thus there is

a less dramatic difference in the area under the band at the edges compared with the

trans CH wag. This is in agreement with the spectral subtraction analysis performed

earlier, and supports the argument for preferential degradation of the trans-

polybutadiene isomer.

Due to the weakness of the cis isomer band in the infrared spectrum, a reliable

degradation profile for this isomer could not be obtained.

Application of Chemometrics

Chemometrics refers to multivariate data analysis techniques that can assist in

identifying relationships within a data set. This statistical method of data manipulation

is regularly employed to aid in the investigation of spectral data [45]. Although

chemometrics can be a powerful tool, care must be taken to ensure that one interprets

the data in the context of the data treatment that has been applied. In this case,

chemometrics is utilised to provide two main functions: firstly, to determine the major

13

sources of change in the spectra, and secondly to plot these changes across the cross-

section of the polymer in order to understand degradation-related changes as a

function of sample depth. An advantage of using chemometrics over other spectral

manipulation techniques such as band area ratios is that chemometrics uses the whole

spectrum, whereas band area ratio methods only use two relatively small regions.

To achieve these objectives a principal component analysis (PCA) was performed on

the data. Through a series of mathematical treatments this isolates the main factors, or

components, of change in the spectra. The regions of the polyurethane spectrum that

change the most relative to all the spectra in the data set are incorporated in principal

component 1, or PC1, the second orthogonal major component of change is termed

PC2, and so forth. Typically, only the first few principal components can be

considered significant; further PCs have little weight, and often describe factors such

as spectral noise.

Figs 10 and 12 show the PC1 loadings plots for some polyurethane samples. A

loadings plot is an indication of the extent of variation in the spectral data set as a

function of wavenumber. Inspection of Figs 10 and 12 reveals that PC1 is remarkably

similar to the difference spectrum in Figs 3 and 4. The loading plot of PC1 shows a

broad positive band in the OH stretching region and negative bands in CH stretching,

deformation and wagging vibrations, and can be interpreted in terms of oxidation of

the polymer.

The PC1 loadings plots further support the information obtained earlier regarding the

specific degradation of trans-polybutadiene isomers. The vinyl absorptions are not as

strongly affected by degradation, although they show negative bands in PC1.

Conversely the trans isomer bands show drastic negative bands.

Since PC1 loadings clearly represent changes in the spectra as a consequence of

degradation, it follows that the score of PC1 is a convenient measure of the extent of

degradation. It can be inferred that plotting of PC1 scores against location of the

spectral collection along the cross-section provides a depth-profile of degradation-

related changes in the polyurethane rubber. Some examples of such plots are provided

14

in Figs 11 and 13. These profiles obtained by the application of PCA are very similar

in appearance to those obtained by comparing band area ratios.

4. Conclusions Transmission infrared spectra of the HTPB/IPDI polyurethane samples gave strong

evidence for the degradation of the polybutadiene segment, whilst the polyurethane

linkage appeared relatively unaffected. Spectral subtraction showed a loss of CH2

stretching bands, as well as losses of the vinyl and trans CH wag bands of the

polybutadiene isomers. A relatively greater loss of the trans-polybutadiene CH wag

band compared with the vinyl isomer indicated preferential degradation of the trans

isomer. The subtraction spectrum also showed a degradation-related absorption

increase in the region between 1640 cm-1 and 1550 cm-1, which may be due to the

presence of acid salts. An increase in OH stretching at the edges of the degraded

rubber pointed to hydroxyl functional groups.

The chemometric technique PCA was employed to aid in the spectral investigation.

The analysis supported the proposal of preferential degradation of the trans isomer,

and also provided a degradation profile utilising the entire spectrum as a function of

sample depth. The degradation profiles calculated from band area ratios showed a

similar oxidation profile shape to that obtained via chemometrics. Furthermore, the

oxidation profile demonstrated the higher rate of degradation of the trans isomer

compared to the vinyl isomer. Increasing the aging time and temperature of the

polyurethane resulted in a higher concentration of degradation products, although

there did not appear to be any differences in the nature of the products formed. The IR

results presented here are consistent with the analysis of chemical degradation

products previously obtained via NMR studies for this material (Harris et al. [27]).

15

References

[1] Wright P, Cumming APC. Solid Polyurethane Elastomers. London, UK: Maclaren and Sons LTD, 1969.

[2] Szycher M. Szycher's Handbook of Polyurethanes. Fl, USA: CRC Press, 1999. [3] Hepburn C. Polyurethane Elastomers. NY, USA: Elsevier Science Publishing

Co., 1982. [4] Burel F, Feldman A, Bunel C. Hydrogenated hydroxy-functionalized

polyisoprene (H-HTPI) and isocyanurate of isophorone diisocyanates (I-IPDI): reaction kinetics study using FTIR spectroscopy. Polymer 2005;46:15.

[5] Dannoux A, Esnouf S, Begue J, Amekraz B, Moulin C. Degradation kinetics of poly(ether-urethane) Estane[trademark] induced by electron irradiation. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms;In Press, Corrected Proof.

[6] Ri Zhang D, Wha Oh S, Pyo Hong Y, Kang YS. Synthesis and characterization of a new adhesion-activator for polymer surface. International Journal of Adhesion and Adhesives 2005;25:371.

[7] Chen Y, Zhou S, Yang H, Gu G, Wu L. Preparation and characterization of nanocomposite polyurethane. Journal of Colloid and Interface Science 2004;279:370.

[8] Sari B, Talu M, Yildirim F, Balci EK. Synthesis and characterization of polyurethane/polythiophene conducting copolymer by electrochemical method. Applied Surface Science 2003;205:27.

[9] Subramanian K. Hydroxyl-terminated poly (azidomethyl ethylene oxide-b-butadiene-b-azidomethyl ethylene oxide)--synthesis, characterization and its potential as a propellant binder. European Polymer Journal 1999;35:1403.

[10] Huang S-L, Chao M-S, Ruaan R-C, Lai J-Y. Microphase separated structure and protein adsorption of polyurethanes with butadiene soft segment. European Polymer Journal 2000;36:285.

[11] Myer YP, Chen ZJ, Frisch HL. Resonance Raman spectroscopic study of an iodine-doped pseudo-interpenetrating polymer network of natural rubber-poly(carbonate urethane). Polymer 1997;38:729.

[12] Edwards HGM, Farwell DW. Fourier transform-Raman spectroscopy of amber. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 1996;52:1119.

[13] Hatchett DW, Kodippili G, Kinyanjui JM, Benincasa F, Sapochak L. FTIR analysis of thermally processed PU foam. Polymer Degradation and Stability 2005;87:555.

[14] van der Ven LGJ, Leuverink R, Henderiks H, van Overbeek R. Durability prediction of p-urethane clearcoats. Progress in Organic Coatings 2003;48:214.

[15] Lattimer RP, Williams RC. Low-temperature pyrolysis products from a polyether-based urethane. Journal of Analytical and Applied Pyrolysis 2002;63:85.

[16] Wilhelm C, Gardette J-L. Infrared analysis of the photochemical behaviour of segmented polyurethanes: 1. Aliphatic poly(ester-urethane). Polymer 1997;38:4019.

[17] Wilhelm C, Gardette J-L. Infrared analysis of the photochemical behaviour of segmented polyurethanes: aliphatic poly(ether-urethane)s. Polymer 1998;39:5973.

16

[18] Wilhelm C, Rivaton A, Gardette J-L. Infrared analysis of the photochemical behaviour of segmented polyurethanes : 3. Aromatic diisocyanate based polymers. Polymer 1998;39:1223.

[19] Allen NS, Barcelona A, Edge M, Wilkinson A, Merchan CG, Ruiz Santa Quiteria V. Thermal and photooxidation of high styrene-butadiene copolymer (SBC). Polymer Degradation and Stability 2004;86:11.

[20] Haska SB, Bayramli E, Pekel F, Ozkar S. Mechanical Properties of HTPB-IPDI-Based Elastomers. Journal of Polymer Science 1997;64:2347.

[21] Celina M, Graham AC, Gillen KT, Assink RA, Minier LM. Thermal degradation studies of a polyurethane propellant binder. Rubber Chemical Technology 2000;73:678.

[22] Judge MD. An investigation of composite propellant accelerated ageing mechanisms and kinetics. Propellants, Explosives, Pyrotechnics 2003;28:114.

[23] Ahlblad G, Reitberger T, Terselius B, Stenberg B. Thermal oxidation of hydroxyl-terminated polybutadiene rubber I. Chemiluminescence studies. Polymer Degradation and Stability 1999;65:179.

[24] Ahlblad G, Reitberger T, Terselius B, Stenberg B. Thermal oxidation of hydroxyl-terminated polybutadiene rubber II. Oxidation depth profiles studied by imaging chemiluminescence. Polymer Degradation and Stability 1999;65:185.

[25] Sinturel C, Billingham NC. A theoretical model for diffusion-limited oxidation applied to oxidation profiles monitored by chemiluminescence in hydroxy-terminated polybutadiene. Polymer International 2000;49:937.

[26] Celina M, Wise J, Ottesen DK, Gillen KT, Clough RL. Correlation of chemical and mechanical property changes during oxidative degradation of neoprene. Polymer Degradation and Stability 2000;68:171.

[27] Harris DJ, Assink RA, Celina M. NMR Analysis of Oxidatively Aged HTPB/IPDI Polyurethane Rubber: Degradation Products, Dynamics and Heterogeneity. Macromolecules 2001;34:6695.

[28] Assink RA, Celina M, Gillen KT. Monitoring the condition of thermally aged polymers by NMR relaxation measurements. Polymer News 2003;28:102.

[29] Mowery DM, Assink RA, Celina M. Sensitivity of proton NMR relaxation times in a HTPB based polyurethane elastomer to thermo-oxidative aging. Polymer 2005;46:10919.

[30] Skutnik JM, Assink RA, Celina M. High-sensitivity chemical derivatization NMR analysis for condition monitoring of aged elastomers. Polymer 2004;45:7463.

[31] Celina M, Skutnik Elliott JM, Winters ST, Assink RA, Minier LM. Correlation of antioxidant depletion and mechanical performance during thermal degradation of an HTPB elastomer. Polymer Degradation and Stability 2006;91:1870.

[32] Celina M, Trujillo AB, Gillen KT, Minier LM. Chemiluminescence as a condition monitoring method for thermal aging and lifetime prediction of an HTPB elastomer. Polymer Degradation and Stability 2006;91:2365.

[33] Foldes E, Lohmeijer J. Relationship between chemical structure and performance of primary antioxidants in PBD. Polymer Degradation and Stability 1999;66:31.

[34] Adam C, Lacoste J, Lemaire J. Photo-oxidation of Elastomeric materials. Part 1 - Photo-oxidation of Polybutadienes. Polymer Degradation and Stability 1989;24:185.

17

[35] Gillen KT, Clough RL. Rigorous experimental confirmation of a theoretical model for diffusion-limited oxidation. Polymer 1992;33:4358.

[36] Cunliffe AV, Davis A. Photo-oxidation of thick polymer samples--Part II: The influence of oxygen diffusion on the natural and artificial weathering of polyolefins. Polymer Degradation and Stability 1982;4:17.

[37] Audouin L, Langlois V, Verdu J, de Bruijn J. Role of oxgyen diffusion in polymer aging: kinetic and mechanical aspects. Journal of Materials Science 1994;29:569.

[38] Wise J, Gillen KT, Clough RL. Quantitative model for the time development of diffusion-limited oxidation profiles. Polymer 1997;38:1929.

[39] Lin-Vien D, Colthup NB, Fateley WG, Grasselli JG. The Handbook of Infrared and Raman Characteristic Frequencies of Organic Molecules. USA: Academic Press, 1991.

[40] Romanova V, Begishev V, Karmanov V, Kondyurin A, Maitz MF. Fourier transform Raman and Fourier transform infrared spectra of cross-linked polyurethaneurea films synthesized from solutions. Journal of Raman Spectroscopy 2002;33:769.

[41] Queiroz DP, Norberta de Pinho M. Structural characteristics and gas permeation properties of polydimethylsiloxane/poly(propylene oxide) urethane/urea bi-soft segment membranes. Polymer 2005;46:2346.

[42] Schoonover JR, Dattelbaum DM, Osborn JC, Bridgewater JS, Kenney I, John W. Pressure-dependent Fourier transform infrared spectroscopy of a poly (ester urethane). Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2003;59:309.

[43] Han Q, Urban MW. Surface/Interfacial Changes During Polyurethane Crosslinking: A Spectroscopic Study. V. Journal of Applied Polymer Science 2001;81:2045.

[44] Auten KL, Petrovic ZS. Synthesis, Structural Characterization, and Properties of Polyurethane Elastomers Containing Various Degrees of Unsaturation in the Chain Extenders. Journal of Polymer Science: Part B: Polymer Chemistry 2002;40:1316.

[45] Guilment J, Bokobza L. Determination of polybutadiene microstructures and styrene-butadiene copolymers composition by vibrational techniques combined with chemometric treatment. Vibrational Spectroscopy 2001;26:133.

18

NH

CO

O

OHn

NH

CO

O

OH

m

Figure 1 HTPB/IPDI polyurethane structure

Figure 2 Diagram showing the parameters for the IR microspectroscopic linemaps of the sectioned aged polyurethane samples.

120 μm

40 μm

1st step

2nd step

3rd step

4th step etc

Degraded elastomer edge

Direction of linemap

19

3600 3400 3200 3000 2800 2600

C

B

A

2845

2918

3004

3074

Abso

rban

ce

Wavenumbers (cm-1)

Figure 3 Infrared microspectra in the region 3700-2500 cm-1 of polyurethane aged 10 days at 110 °C taken from (A) exterior; and (B) interior of the same sample. Spectrum (C) is the

difference spectrum (A-B) x2

1800 1600 1400 1200 1000 800

1712

1640

Amid

e III

Amid

e II

CB

A

912

967

993

1308

1418

1448

, 143

6

Abso

rban

ce

Wavenumbers (cm-1)

Figure 4 Infrared microspectra in the region 1900-650 cm-1 of polyurethane aged 10 days at 110 °C taken from (A) exterior; and (B) interior of the same sample. Spectrum (C) is the

difference spectrum (A-B) x2

20

4000 3800 3600 3400 3200 3000 2800

B

D

C

A

Wavenumbers (cm-1)

Abso

rban

ce

Figure 5 IR microspectra in the OH stretching region showing the formation of hydroxyl groups with degradation. (A) was aged 235 days at 80 °C; (B) 8 days at 125 °C; (C) 10 days at 110 °C;

(D) unaged. All spectra were obtained at the edges of their respective samples

OH

OHy

x

z

cis vinyl trans

Figure 6 Possible structural isomers of polybutadiene.

21

1800 1600

C

B

AAb

sorb

ance

Wavenumbers (cm-1)

Figure 7 The C=O stretching region of the IR microspectra of (A) the exterior of polyurethane aged 10 days at 110 °C and (B) the interior of the sample. Also shown is the difference spectrum

(C) which is (A) – (B).

1900 1800 1700 1600 1500 1400

B

D

C

A

Wavenumbers (cm-1)

Abso

rban

ce

Figure 8 IR microspectra in the carbonyl stretching region showing the degradation-related changes occurring in the carbonyl region. (A) was aged 235 days at 80 °C; (B) 8 days at 125 °C;

(C) 10 days at 110 °C; (D) unaged. All spectra were obtained at the edges of their respective samples, and have been offset.

22

0 500 1000 1500 2000

1.5

1.0

0.5

0.0

963 (trans) 912 (vinyl)

A amid

e II /

AC

H w

ag

Distance (μm)

Figure 9 Comparison of degradation profiles, derived from IR microspectra, of trans and vinyl polybutadiene isomers. The plots have been offset for clarity.

4000 3500 3000 2500 2000 1500 1000-0.16-0.14-0.12-0.10-0.08-0.06-0.04-0.020.000.020.040.060.08

PC1

load

ings

Wavenumbers (cm-1)

Figure 10 PC1 loadings for polyurethane derived from an IR line map of cross-sectioned polyurethane aged at 110 oC for 10 days.

23

0 500 1000 1500 2000

0.2

0.0

-0.2

-0.4

-0.6

Scor

e

Distance (μm)

Figure 11 Scores plot from Fig 8 used to give a degradation profile.

4000 3500 3000 2500 2000 1500 10000.10

0.05

0.00

-0.05

-0.10

-0.15

PC1

load

ings

Wavenumbers (cm-1)

Figure 12 PC1 loadings for polyurethane derived from an IR line map of cross-sectioned polyurethane aged at 125 oC for 8 days.

24

0 200 400 600 800 1000 1200 14000.2

0.0

-0.2

-0.4

-0.6

-0.8

Scor

e

Distance (μm)

Figure 13 Scores plot from Fig 10 used to give a degradation profile.

25

Table 1 Details of aging conditions

Temperature (°C) Days Aged

80 235

110 2, 10, 18

125 5, 8

Table 2 Assignment of infrared absorption bands for the polybutadiene segment.

Wavenumber (cm-1) IntensityAssignment

[16-18,34,39]

Cis Trans Vinyl

3660-3100 3660-3100 3660-3100 w νOH 3074 m νasCH2=

3004 m,sh ν(CH=CH)

2977 m,sh νsymCH2=

2918 2918 2918 vs νasCH2

2845 s νsymCH2

1671 m,sh νC=C

1658 vw,sh νC=C

1640 m νC=C

1448, 1436 s δsymCH2

1436 s δsymCH2

1418 m δCH2=

1404 w δ(=CH)

1364-1344 1364-1344 1364-1344 w CH2 wag

1308 m δ(=CH2)

993 m CH wag

967 vs CH wag

912 CH2 wag

ν = stretch, δ = deformation, s = strong, m = medium, w = weak, vw = very weak, sh = shoulder

26

Table 3 Assignment of infrared absorption bands for the urethane segment.

Frequency (cm-1) Intensity Assignment [16-18,34,39]

3430-3340 m νNH

1725 s νC=O (free)

1712 s νC=O (hydrogen bonded)

1517 m Amide II

1236 m Amide III

ν = stretch, δ = deformation, s = strong, m = medium, w = weak, vw = very weak, sh = shoulder Table 4 Curve fitting band area ratios.

Band Area Area relative to 967 cm-1

unaged 993 cm-1 6.7 0.25 967 cm-1 26 1 912 cm-1 11 0.42 125 °C 8 days 993 cm-1 4.1 1.0 967 cm-1 3.9 1 912 cm-1 5.9 0.66 80 °C 235 days 993 cm-1 7.5 0.44 967 cm-1 19 1 912 cm-1 14 0.73