MOLSTR 11215

Journal of Molecular Structure 521 (2000) 19–23 www.elsevier.nl/locate/molstruc

Studies of bisphenol-A–polycarbonate aging by Raman difference spectroscopy 夽 S.-N. Lee a, V. Stolarski b, A. Letton b, J. Laane a,* a Department of Chemistry, Texas A&M University, College Station, TX 77843, USA Polymer Technology Consortium and Department of Mechanical Engineering, Texas A&M University, College Station, TX 77843, USA

b

Received 9 July 1999; accepted 1 September 1999

Abstract Raman difference spectroscopy (RDS) has been used to measure small frequency shifts (up to ^3 cm ⫺1) in the Raman frequencies of the bisphenol-A–polycarbonate polymer when subjected to elongation or thermal aging. While only minor shifts were observed upon thermal treatment, the elongation treatment resulted in shifts of ⫺2:51 ^ 0:23 cm⫺1 for the C–O stretching band at 1235 cm ⫺1 of ⫺0:78 ^ 0:15 cm⫺1 for the CH out-of-plane wag at 885 cm ⫺1. Smaller shifts were also observed for the 1606 and 1111 cm ⫺1 bands. This work demonstrates that conformational processes can be monitored using RDS. 䉷 2000 Elsevier Science B.V. All rights reserved. Keywords: Raman difference spectroscopy; Bisphenol-A–polycarbonate, polymer; Physical aging

1. Introduction Bisphenol-A–polycarbonate (BPAPC) is known for its unusual ability to withstand a sudden impact of energy. For weaker polymers impact is often sufficient to break chemical bonds leading to macroscopic failure.

夽

In honour of Professor Giuseppe Zerbi on the occasion of his 65th birthday. * Corresponding author. Tel.: ⫹1-409-854-3352; fax: ⫹1-409845-3154. E-mail address: [email protected] (J. Laane).

In the polycarbonate system, however, the energy is dissipated through the movement or reorientation of molecular segments leaving the chemical bonds intact [1]. Processes such as injection molding, which employ extreme conditions (high temperature and pressure), are routinely used when making components out of BPAPC [2]. These extreme conditions force the polymer chains to reside in high energy, high volume conformations. Once molded, the component is usually quenched to below its glass transition temperature. This sudden removal of energy locks in the high-energy conformations, even at ambient conditions. It is in this state that the polymer is at its toughest, with sufficient residual volume for the molecular segments to rotate and dissipate energy as needed. Over time, however, thermodynamics pushes the chains to reach a lower energy equilibrium state. Through this process, a densification of the material is observed indicating that the packing efficiency of the polymer chains has increased. This

0022-2860/00/$ - see front matter 䉷 2000 Elsevier Science B.V. All rights reserved. PII: S0022-286 0(99)00422-6

20

S.-N. Lee et al. / Journal of Molecular Structure 521 (2000) 19–23

packing does not lead to crystallization, making this behavior unique. It is believed that the increased steric repulsions between the chains prohibit rotations of molecular segments. Thus, should an external force be applied to the polymer in this condition, the added energy cannot be dissipated and the material becomes brittle and suffers macroscopic failure. This phenomenon is referred to as “physical aging” [3–5]. Thus, it is the chemical structure of BPAPC that contributes not only to the observed toughness of this polymer, but also to the physical aging process. Spectroscopic methods are generally not sensitive enough to detect subtle conformational changes in polymer structure. Vibrational frequencies from infrared or Raman spectroscopy depend primarily on bond stretching and bond bending force constants in molecules. Since these force constants are only little changed by conformational changes, the vibrational frequencies typically change by at most one or two cm ⫺1, which are too small to be detected by conventional methods. However, Raman spectroscopy has been effectively used to determine the conformations of polymers such as polyethylene [6] and polybutadiene [7]. Raman difference spectroscopy (RDS) is a specialized method, which allows frequency differences as low as 0.1 cm ⫺1 between similar samples to be measured. This technique is thus well adapted to measuring conformational changes during a polymer aging process, which result in only small frequency shifts. By comparing the vibrational frequencies of aged and unaged polymers and by identifying the individual vibrational bonds, which show the frequency shifts, more detailed information on the aging process can be ascertained. The RDS technique is based on simultaneously or nearly simultaneously recording the spectra of two similar samples (e.g. aged and unaged BPAPC). This ensures that there is no frequency error in the difference between the two samples since the instrument monochromator is not moved. Furthermore, although the band maxima are difficult to measure with great accuracy, the difference spectrum can be computed and used to readily get ^0.1 cm ⫺1 accuracy. The theoretical principles have been outlined [8] and the RDS method has been reviewed [9]. In the present study, we first undertook a study of the Raman and infrared spectra of BPAPC and then investigated the frequency shifts of the conformation-

ally sensitive bands by RDS after the polymer has been subjected to heating for many hours (induced aging) or to deformation by elongation. No detailed vibrational studies of BPAPC have been previously reported. 2. Experimental 2.1. Raman spectra and RDS The Raman spectra of the unaged and aged BPAPC samples were recorded using an ISA U-1000 double monochromator. The 532 nm line of a Coherent Radiation DPSS-200 or -400 frequency doubled Nd:YAG laser was used as the excitation source (200 or 400 mW power) utilizing a 90⬚ scattering geometry. A liquid nitrogen cooled CCD was employed as the detector and a Fe–Ne lamp was used to precisely calibrate the frequency scale of each recorded spectrum. Galactic LabCalc software was used to manipulate the spectra and to determine the frequency shifts in the RDS spectra. 2.2. Infrared spectra The infrared spectra of films of BPAPC were recorded using a BioRad FTS-60 instrument. The films were prepared by dissolving about 100 mg of polymer in 3 ml of methylene chloride (Aldrich). Three to five drops of solution were placed on a KBr disc and the solvent was allowed to evaporate. Spectra (256–1024) scans were recorded at resolutions of 1–4 cm ⫺1. 2.3. Sample preparation BPAPC samples supplied as Lexan 141-111 extrusion grade pellets from General Electric Company (Mw ˆ 54 000–57 000 and Mn ˆ 22 700–23 900) were compression molded in our laboratories. These pellets were dried under vacuum at 100⬚C for 48 h prior to molding. Compression molding was performed using a heating press. Approximately 75 mg of BPAPC pellets were put into an aluminum mold which was placed between Kapton film to enhance their release after molding. The pellets were heated from room temperature until softening began (approx. 190⬚C) when contact pressure was

S.-N. Lee et al. / Journal of Molecular Structure 521 (2000) 19–23

21

5000 lb load displaced at the rate of 0.1 in./min until the breaking point. The sample elongation before failure reached 45%. Immediately following the mechanical testing, the Raman spectra of the samples were recorded. Scans were taken of the break edge, the necking area, and the undeformed portion of the sample in order to examine the changing morphology due to orientation.

Fig. 1. The fingerprint region of Raman spectrum of the BPAPC polymer.

applied. Heating continued until the temperature reached 230⬚C. At this time isothermal heating ensued until the pellets were completely melted (after about 6 min). The temperature was then decreased to 170⬚C where the samples were isothermally heated for an additional 10 min. This allowed for the relaxation of mold enhanced stresses and for the samples to reach conformational uniformity. The samples were then quenched into a cold water bath (18⬚C) and popped out of the mold. 2.3.1. Deformation by elongation For the RDS studies, which required an oriented sample, a BPAPC rectangular piece was pulled on an Instron 4206. The sample was subjected to a

Fig. 2. The fingerprint region of the infrared spectrum of the BPAPC polymer.

2.3.2. Thermal aging To assure the maximum amount of homogeneity possible, all samples used henceforth were cut from a compression molded plaque. From the plaque, two 2 cm square samples were cut with a small saw. One sample was placed in a desiccator as the unaged. Once the oven had stabilized at 130⬚C (^3⬚), the other sample were placed inside. The oven was evacuated and filled with argon. After 500 h, the sample was removed and quenched in a water bath and it was used as the thermally aged sample. 3. Results and discussion Figs. 1 and 2 show the 500–1900 cm ⫺1 regions of the Raman spectra and infrared spectra, respectively, for an unoriented 55 000 amu Lexan polycarbonate plaque. The observed frequencies and assignments are presented in Table 1. The assignments were aided by group frequency considerations [10–12]. Broad peaks in both Raman and infrared spectra occur in the 1200–1600 cm ⫺1 region. These peaks were attributed to the C–O stretching and phenyl ring breathing, respectively. Their broadness is an indication that there are a variety of polymer conformations and that these motions play a central role in contributing to the morphological changes giving rise to the aging mechanism. Seven of the Raman bands were selected for the RDS studies. These were generally the more intense bands, but they include two band shoulders on these strong bands. For the bands of weaker intensity, the accuracy of determining the frequency shifts for the oriented samples are considerably lower. Fig. 3 shows the Raman spectra of an oriented (elongated) sample and an unoriented sample for the 1111 cm ⫺1 band along with the difference spectrum resulting from the RDS measurement. The results show that the

22

S.-N. Lee et al. / Journal of Molecular Structure 521 (2000) 19–23

Table 1 Assignments for BPAPC from infrared and Raman spectra (i.p., in-plane; o.p., out of plane) Vibration

IR (solution)

Intensity a

Raman (polymer)

Intensity a

CyO stretch Ring stretch Ring stretch Ring stretch CH def CH3 bend CH3 bend CO def C–O stretch CH3 bend CH wag (i.p) CH wag (i.p) C–O stretch Ring stretch CH wag (o.p) C–CH3 stretch C–CH3 stretch CH wag (o.p) CH wag (o.p) Ring def (o.p) Ring def (i.p)

1772

m

1595 1504 1462 1389 1366 1238

w s vw vw vw s

1777 1606 1592

w s sh

1468

vw

1235 (broad)

s

1194 (broad) 1161

s s

1016

w

884

w

827 720

m w

1180 1111 1025 1012 936 923 889 885 821 710 643

s s vw w vw w s sh w m m

a

w, weak; m, medium; s, strong; v, very; and sh, shoulder.

Fig. 3. Raman difference spectrum (lower solid curve) of the 1111 cm ⫺1 band of BPAPC polymer between the oriented and unaged polymer sample: (A) oriented polycarbonate; (B) unaged (unoriented); and (C) difference between A and B representing a ⫺0.14 cm ⫺1 shift in oriented sample.

oriented sample has shifted by ⫺0.14 cm ⫺1. Ten of these measurements were made using different samples and the shift upon 45% elongation was determined to be ⫺0:14 ^ 0:03 cm⫺1 : Similar measurements were made on six other bands and these results are shown in Table 2. As can be seen, five of the Raman bands show statistically significant frequency shifts upon elongation, while the 1592 and 1180 cm ⫺1 bands do not. The largest shifts were observed for the C–O stretching at 1235 cm ⫺1, the out-of-plane CH wag 885 cm ⫺1 and the C–CH3 stretching at 889 cm ⫺1. Table 2 also shows the frequency shift measurements for samples which had been “aged” by heating to 130⬚C for 500 h. What is evident is that only minor frequency shifts can be detected for three or four of the bands. The broad C–O stretching band at 1235 cm ⫺1 is shifted by ⫺0:49 ^ 0:20 cm⫺1 while bands at 1111, 889, and 885 cm ⫺1 appear to have statistically significantly shifts although each individual measurement does not necessarily show a shift. Moreover, the band at 1111 cm ⫺1 appears to shift up whereas upon elongation the shift is down. If both elongation and heating help to transform the polymer morphology closer to the lowest

S.-N. Lee et al. / Journal of Molecular Structure 521 (2000) 19–23

23

Table 2 Frequencies (cm ⫺1) and shifts of Raman bands of polycarbonate (i.p., in-plane; o.p., out of plane) Vibration

Raman frequency

Frequency shift (elongation by 45%)

Frequency shift (thermal aging a)

Ring stretch Ring stretch C–O stretch CH wag (i.p.) CH wag (i.p.) C–CH3 stretch CH wag (o.p.)

1605:64 ^ 0:03 1591:58 ^ 0:09 (overlapped) 1234:83 ^ 0:10 (broad) 1180 ^ 0:04 1111:33 ^ 0:02 889:36 ^ 0:06 885:01 ^ 0:11

⫺0:16 ^ 0:05 ⫺0:06 ^ 0:13 ⫺2:51 ^ 0:23 ⫺0:06 ^ 0:06 ⫺0:14 ^ 0:03 ⫺0:33 ^ 0:09 ⫺0:78 ^ 0:15

⫹0:04 ^ 0:08 ⫹0:04 ^ 0:08 ⫺0:49 ^ 0:20 ⫺0:05 ^ 0:07 …⫹0:23 ^ 0:05† b ⫺0:22 ^ 0:14 …⫺0:14 ^ 0:05† b

a b

Heating at 130⬚C for 500 h. Based on only five measurements.

conformational energy, the shifts lowest conformation energy, the shifts should be in the same direction. In any case, what is clear is that the 500 h of heating at 130⬚ has very little effect on the Raman frequencies and thus apparently no major conformational processes have occurred. This is in contrast to the elongation studies where several significant shifts can be observed. Some insight into the conformational processes can be inferred. The largest shifts involve the broad C–O stretching at 1235 cm ⫺1 and the C–CH3 stretching and C–H out-of-plane wagging at 889 and 885 cm ⫺1, respectively (the assignments could be reserved). A ring stretching at 1605 cm ⫺1 and a C–H in-plane wag shift by smaller amounts. Elongation and aging are both expected to change the conformation through internal rotations about the various single bonds in the polymer so the 1235 and 889 cm ⫺1 shifts for stretching bonds of C–O and C–CH3 single bonds is not unexpected. Smaller shifts for C–H bending and phenyl ring stretching models suggest changes in polymer packing resulting in shifts due to non-bonded interactions.

4. Conclusions The results demonstrate that RDS can readily be used to measure Raman frequency shifts resulting from sample elongation. Thermal aging, however, produced minimal shifts. The observed shifts are assumed to result from conformational changes, which result in minor effects on the stretching or bending force constants or on the non-bonded interactions. The broad C–O stretching band at 1235 cm ⫺1 is the

most sensitive to these effects, and it would appear that this can be used as a means of monitoring internal rearrangement within the BPAPC polymer. It is anticipated that similar “aging diagnostics” would be available for other polymers as well. Acknowledgements J. L. and A. L. thank the Texas A&M Research Enhancement Program for financial support. This paper is dedicated to Professor Gus Zerbi for his contributions to science. References [1] L.C.E. Struik, Physical Aging in Amorphous Polymers and other Materials, Elsevier, Amsterdam, 1987. [2] K. Neki, P.H. Geil, J. Macromol. Sci. Phys. B 8 (1973) 295. [3] A.F. Yee, S.A. Smith, Macromolecules 54 (1981) 14. [4] B. Erman, D.C. Marvin, P.A. Irvine, P.J. Flory, Macromolecules 15 (1982) 6670. [5] T.S. Chow, Polymer 34 (1993) 341. [6] Y.J. Sung, W. Hagedorn, J. Polym. Sci. 16 (1978) 1181. [7] S.W. Cornell, J.L. Koenig, Rubber Chem. Technol. 43 (1970) 322. [8] J. Laane, W. Kiefer, J. Chem. Phys. 72 (1980) 5305. [9] J. Laane, in: J.R. Durig (Ed.), Vibrational Spectra and Structure, 11, Elsevier, Amsterdam, 1983 chap. 6. [10] D.O. Hummel (Ed.), Infrared Spectra of Polymers, 207, Interscience, New York, 1966. [11] R.T. Conley, Infrared Spectroscopy, Allyn & Bacon, Boston, MA, 1972, p. 92. [12] D. Lin-Vien, N.B. Colthup, W.G. Fateley, J.G. Grasselli, The Handbook of Infrared and Raman Characteristic Frequencies of Organic Molecules, Academic Press, Boston, MA, 1991.