Polymer Degradation and Stability 87 (2005) 347e354 www.elsevier.com/locate/polydegstab

Application of pyrolysis/gas chromatography/Fourier transform infrared spectroscopy and TGA techniques in the study of thermal degradation of poly (3-hydroxybutyrate) A. Gonzalez1, L. Irusta, M.J. Ferna´ndez-Berridi*, M. Iriarte, J.J. Iruin Polymer Science & Technology Department and Institute for Polymer Materials (POLYMAT), University of the Basque Country, P.O. Box 1072, 20080 San Sebastian, Spain Received 29 April 2004; received in revised form 6 September 2004; accepted 17 September 2004 This is dedicated to Prof. Cecilia Sarasola.

Abstract The thermal degradation behaviour of bacterial poly (3-hydroxybutyrate) has been studied by PyrolysiseGCeFTIR using a semi continuous furnace in the temperature range of 200e600  C. At temperatures lower than 400  C, 2-butenoic acid and higher degradation products have been obtained. However, at higher temperatures propene and carbon dioxide are the major degradation products. The relative composition of the detected compounds changes as a function of pyrolysis temperature. The application of Hi-Res TGA technique permits the minimum temperature of PHB decomposition to be determined with increased precision. Ó 2004 Elsevier Ltd. All rights reserved. Keywords: Poly (3-hydroxybutyrate) thermal degradation; Poly (3-hydroxybutyrate) degradation mechanism; PyrolysiseGas chromatographye Fourier transform infrared spectroscopy; Hi-Res TGA

1. Introduction Poly (3-hydroxybutyrate) (PHB) belongs to the family of biodegradable aliphatic polyesters that can be produced from renewable resources [1,2]. PHB is an optically active polyester, which is produced as an intracellular polymer by a large variety of micro-organisms. The polymer is fully biodegradable by hydrolytic processes or microbial activity [3e5].

* Corresponding author. Tel.: C34 943 018194; fax: C34 943 212236. E-mail address: [email protected] (M.J. Ferna´ndez-Berridi). 1 On leave from the Chemical Engineering Faculty, Universidad de Oriente, Santiago de Cuba, Cuba. 0141-3910/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.polymdegradstab.2004.09.005

It is well known that PHB is rather unstable at temperatures near its melting temperature (180  C) [6,7], and that the polymer can suffer some molecular weight reduction when kept at temperatures even 10  C below its melting point, which limits its processability. It is widely believed that this degradation occurs almost exclusively via a random chain scission mechanism involving a six-membered ring transition state [8,9]. Different techniques have been used to study PHB pyrolysis, and many authors have devoted work to the determination of the composition and yields of PHB degradation products. Thus, Lehrle et al. [10,11] studied the pyrolysis mechanism of PHB by Gas chromatography/Mass spectrometry. Nishida and Tokiwa [12] observed degraded PHB films by SEM. Abate et al. [13] described a method for the structural analysis of hydroxybutyrate/hydroxyvalerate copolymers, based on

348

A. Gonzalez et al. / Polymer Degradation and Stability 87 (2005) 347e354

the direct FAB-MS analysis of the partial pyrolysis products. Other workers have studied the degradation of PHB by DSC and GC and FAB-MS and NMR [14,15]. In all cases, they arrived at the same conclusion about the degradation mechanism. The objective of the present paper is to test the sensitivity of the Pyrolysis/GCeFTIR technique to characterise the volatile products of PHB thermal degradation under a range of temperatures. Additionally, the application of high resolution TGA (Hi-Res TGA) [16] has permitted us to establish precisely the minimum degradation temperature of PHB as a function of the resolution level, and the results have been compared with previous work reported in literature [17].

2.2. Pyrolysis/GC/FTIR equipment Pyrolysis was performed by depositing a range of microgram quantities of the polymer onto a semicontinuous furnace pyrolysis unit. This unit is a Pyrojector SGE (Konik) coupled to the Gas chromatograph (GC) injection unit. The apparatus used was a Shimadzu GC-14A. An on-line Magna 560 (Nicolet) infrared spectrometer was used to characterise gas chromatographic peaks. The pyrolysis chamber was set at the head of the injector, and different pyrolysis temperatures in the range of 220e600  C were used. The pyrolysis chamber pressure was 180 kPa. The chromatographic column was a Supelco 30 m SPBÔ-1 capillary column of internal diameter 0.25 mm. The temperatures of the injector, detector and transfer line units were set at 200  C and the following temperature program was used for the GC oven: isothermal at 50  C for 15 min, temperature ramped at 10  C/min to 200  C, where it was maintained for a further 15 min. Nitrogen was used as the carrier gas at a pressure of 127 kPa. GC separated products were identified in a FTIR spectrometer equipped with a MCT detector cooled with liquid nitrogen. The chromatograms were built from the interferograms registered at different times.

2. Experimental section 2.1. Materials Bacterial iPHB used in this study was supplied by Biomer Germany as Biopol. The average molecular weights determined by gel permeation chromatography were Mn Z 220 000 and Mw Z 374 000 (referred to polystyrene calibration standards). Its melting temperature (Tm) is 451 K and glass transition temperature (Tg) is 273 K.

H CH

C

H

O C O

CH3 H

C

C

H

O

O

H

O

CH C CH

n

O

CH3

H

O C

n

O

H

H

C

C

CH3 H

O +

C x

CH

C

O

CH3

OH

H

C O

CH

O

C C

CH3 H

O y

O H3C

C

C

H

H

C OH

"monomer"

Scheme 1. PHB degradation mechanism.

349

A. Gonzalez et al. / Polymer Degradation and Stability 87 (2005) 347e354

first scan from room temperature to 140  C at a heating rate of 100  C/min, was carried out, where the sample was kept under isothermal conditions for 30 min. The following scans were performed at the same heating rate from the previous temperature at steps of 10  C.

O H3C

C

C

H

H

CO2 +

C OH

HC

CH2

CH3

Scheme 2. Secondary decomposition of trans-2-butenoic acid.

Data acquisition was performed using a compatible PC fitted with Omnic software (Nicolet).

3. Results and discussion

2.3. Thermogravimetric analysis (TGA)

3.1. Pyrolysis/GC/FTIR analysis

Thermogravimetric analysis (TGA) was performed in a TGA Q500 thermobalance, with standard furnace coupling and nitrogen flow rate of 50 cc/min. Two independent experiments were set. In the first instance, dynamic scans from room temperature to 500  C at a heating rate of 40  C/min were carried out at different resolutions (4, 5, 6, 7 and 8). The Hi-Res operates similar to the traditional constant heating rate ramp segment, except that the heating rate is varied dynamically during the ramp in response to the derivative of weight change (%/min); as %/min increases, heating rate is decreased. As derivative decreases, heating rate is increased. The heating rate is constrained to the range 0.001  C/min (minimum) to the maximum specified in the range segment. The resolution setting is a dimensionless number used to select the most useful band of %/min values for proportional heating rate control. The second experiment was devoted to calculate the weight loss of PHB in isothermal conditions during 30 min, in the temperature range from 140 to 210  C. A

It is generally assumed that the thermal degradation of PHB proceeds via a b-elimination reaction [8], according to Scheme 1. The composition and yields of the degradation products, however, depend on the range of pyrolysis temperature employed. Thus, in the low temperature range (170e200  C), water is produced as a consequence of the condensation of hydroxyl and acidic groups originally present as the end groups of the polymer molecules. When PHB is pyrolysed between 200 and 300  C, monomeric (crotonic acid), dimeric, trimeric and tetrameric volatile products are assumed to evolve. Oligomers higher than tetramer are not volatile enough, and at these temperatures would remain within the polymer. Finally, when PHB is pyrolysed at 500  C, propene and carbon dioxide are now the major products, possibly formed as a consequence of further decomposition of crotonic acid, according to Scheme 2. In accordance with these results, PHB was pyrolysed at four different temperatures and the volatile degradation products were identified by GC/FTIR. Fig. 1 shows

600°C

Intensity

400°C

300°C

2

3 4

220°C

0

5

10

15

20

25

30

35

40

Times (minutes) Fig. 1. Gram-Schmidt Chromatograms obtained at 220, 300, 400 and 600  C pyrolysis temperatures.

350

A. Gonzalez et al. / Polymer Degradation and Stability 87 (2005) 347e354

Peak 3

Absorbance

Absorbance

Peak 2

4000

3600

3200

2800

2400

2000

1600

1200

800

Wavenumber (cm-1) 1800

1750

1700

1650

1600

Fig. 3. Infrared spectrum of the degradation compound termed ‘‘dimer’’ (peak 3 in the chromatograms). Insert: Scale expanded spectrum in the C]O (1850e1600 cmÿ1) stretching region.

Wavenumber (cm-1) Fig. 2. Infrared spectrum of trans-2-butenoic acid (peak 2 in the chromatograms). Insert: Scale expanded spectrum in the C]O (1850e1600 cmÿ1) stretching region.

double carbon/carbon bond stretching vibrations, two bands at 1170 and 1100 cmÿ1, due to eCeCeOe stretching vibrations and a band at 970 cmÿ1, attributable to the o.o.p. bending vibrations of olefinic CeH [18,19]. It must be pointed out, however, that the carbonyl band is in fact constituted by two well-separated contributions at 1770 and 1759 cmÿ1, respectively, that can be attributed to the S-trans and S-cis conformations. This result is consistent with that obtained by Shibano et al. [20] in the analysis of microwave spectra of trans-2-butenoic acid, that revealed the existence of these forms in the gas phase. The infrared spectrum (Fig. 3) of peak 3 (retention time 28 min) is very similar to that of 2-butenoic acid. However, two carbonyl stretching vibration bands can be observed at 1783 and 1744 cmÿ1, respectively. The position of these bands, together with the appearance of a new band at 1179 cmÿ1 reveals the existence of acid

the corresponding chromatograms obtained at 220, 300, 400 and 600  C pyrolysis temperatures. As can be observed, the number of peaks and their retention times are clearly dependent on the pyrolysis temperature. Thus, the chromatogram obtained at the lowest pyrolysis temperature (220  C) displays three peaks, which have been numbered as peaks 2, 3 and 4, respectively. Peak 2 has been identified from its infrared spectrum (Fig. 2) as trans-2-butenoic acid (crotonic acid). The spectrum of this compound is mainly characterised by the following bands: a sharp, low intensity band centred at 3600 cmÿ1, of the monomeric acidic OH group stretching vibration, a high intensity band due to the stretching vibration of carbonyl groups at about 1760 cmÿ1, a band centred at 1664 cmÿ1, assignable to H C

C

H

O

O

H CH C

CH

O

CH3

O C

O

H

H

C

C

CH3H

O CH

CH2

CH3

O

O C

COOH

+

CH3

CH=CH

OH x

C

O

CH CH3

"dimer" Scheme 3. Dimer formation in PHB degradation.

CH2

COOH

351

A. Gonzalez et al. / Polymer Degradation and Stability 87 (2005) 347e354

Peak 4

Absorbance

A

CO2

Peak 1: Contribution A

4000

3600

3200

2800

2400

2000

1600

1200

800

Wavenumber (cm-1) 4000

3600

3200

2800

2400

2000

1600

1200

800

Fig. 5. Infrared spectra of CO2 and that corresponding to the contribution A of peak 1.

Wavenumber (cm-1) Fig. 4. Infrared spectrum of the degradation compound named ‘‘trimer’’ (peak 4 in the chromatograms). Insert: Scale expanded spectrum in the C]O (1850e1600 cmÿ1) stretching region.

degradation products is the same, although their relative concentrations differ from one experiment to the other. In the chromatogram at 400  C, the degradation products corresponding to ‘‘monomer’’ and ‘‘dimer’’ are also observed. However, there is no evidence of the evolution of ‘‘trimer’’ and, on the contrary, a new peak at about 2 min (peak 1) is observed. Nonetheless, from the study of its infrared spectrum it can be deduced that this peak has in fact two contributions, which are identified as carbon dioxide and propene, respectively. Individual spectra of these compounds have been obtained by subtracting one from another. Figs. 5 and 6 show the corresponding spectra and their comparative library spectra. Finally, when the pyrolysis is carried out at 600  C, carbon dioxide and propene are now the only degradation products detected. These compounds are formed by the secondary decarboxylation of crotonic acid (monomer). Therefore, it is evident that as pyrolysis

and a, b unsaturated ester groups in the molecule. According to this, peak 3 has been identified as the structure shown in Scheme 3, and it will be termed ‘‘dimer’’. The spectrum of peak 4 (Fig. 4) at a retention time of 34 min is also very similar to that of peak 3. However in this case, a new carbonyl band at 1757 cmÿ1 can be observed, indicating the existence of saturated ester groups in addition to the acid and a, b unsaturated ester groups, detected in the dimer. This fact responds to the structure of the ‘‘trimer’’, where the three types of groups coexist (Scheme 4). Infrared spectra of the three peaks observed in the chromatogram at 300  C are identical to those obtained at 220  C, indicating that in the range of moderate temperature degradation the nature of the volatile

H C

C

H

O

O

H

O

CH C CH

O O

CH3

CH

CH2

C O

CH3

CH

CH2

COOH

CH3

O C

O

H

H

C

C

CH3 H

O

O C

+ CH3 OH

x

CH=CH

O

C O

CH

CH2

CH3

C O

CH CH3

"trimer" Scheme 4. Trimer formation in PHB degradation.

CH2

COOH

352

A. Gonzalez et al. / Polymer Degradation and Stability 87 (2005) 347e354

A

Propene

Peak 1: Contribution B

4000

3600

3200

2800

2400

2000

1600

1200

800

Wavenumber (cm-1)

discrepancy can be the way the chromatograms have been obtained. Thus, the Gram-Schmidt chromatogram is built upon the basis of the infrared responses of the different functional groups. In the case of chromatograms of compounds with equal functional groups, the appearance of FID and Gram-Schmidt chromatograms can be quite similar as is the case of our system at low pyrolysis temperatures. However, at higher pyrolysis temperatures, once the evolution of carbon dioxide starts, the intensities of the different peaks in the chromatograms will be markedly different and therefore, impossible to quantify.

Fig. 6. Infrared spectra of propene and that corresponding to the contribution B of peak 1.

3.2. Thermogravimetric analysis

Table 1 Relative yields of the different volatile degradation products Temperature (  C)

CO2 C propene

220 300 400 600

0 0 5G2 100

Monomer

Dimer

As stated in Section 2, two experimental procedures have been used. Fig. 7 shows the TG curves at five different resolution levels. As can be seen, the TG curve shifts toward lower temperatures as the resolution increases. Thus at the highest resolution (Hi-Res 8), the weight loss starts at 193  C whereas the degradation begins near 260  C at the lowest Hi-Res (4). These results are fairly different to those reported in literature [21] where TG curves of PHB, measured in a conventional TGA, shift to higher temperatures. This discrepancy can be due to either the different origin of PHB or/and the poorer resolution of the conventional technique. It must be pointed out that as the resolution is increased, the initial degradation temperature is calculated with higher precision. This effect is normal because with progressively higher resolution setting, transitions are constrained to lower decomposition rates, which can only be maintained at lower temperatures. Fig. 8 shows the results of the isothermal experiment at different temperatures. As can be seen, the beginning of the degradation does not start until 180  C. Under these conditions PHB loses 3% of the original weight in

100

% Remanent weight

temperature increases the volatile degradation products go from oligomers to monomer and finally to CO2 and propene. These results are basically in accordance with those reported on literature about the chemical nature of the main degradation products of PHB. However, there are some minor points that it is worthwhile to consider. First of all, under our experimental conditions, ‘‘tetramer’’ moiety has not been detected at any of the pyrolysis temperatures employed. Taking into account the results reported by Lehrle et al. [11], who suggested that tetramer is totally formed via secondary reactions and its detection depends on polymer mass, one possible reason for which this oligomer has not been detected is that under our experimental conditions, either secondary reactions are minimized and/or the polymer mass employed has not exceed the mass limit. In the second place, the yields of degradation products, measured as the pyrogram peak areas, change with pyrolysis temperature (Table 1). Thus, at the lowest temperature the ratio monomer/dimer is almost unity and the trimer quantity accounts for only 10%. However, at 300  C, there is a dramatic increase in the amount of the monomer at the expense of the dimer, and at 400  C dimer content increases while trimer yield goes to zero. Although it seems evident that monomer yield must increase in detriment of oligomer content as pyrolysis temperature increases, there is no straightforward reason for the observed increment of dimer content from 300 to 400  C. One reason to explain this

HR8

80

HR7

HR6

HR5 HR4

60 40 20

Trimer 0

50 G 2 87 G 2 78 G 2 0

40 G 2 10 G 2 17 G 2 0

10 G 2 3G2 0 0

180

200

220

240

260

Temperature (°C) Fig. 7. Thermogravimetric analysis of PHB as a function of the instrument resolution (Hi-Res: 4, 5, 6, 7 and 8).

353

A. Gonzalez et al. / Polymer Degradation and Stability 87 (2005) 347e354 120

220 3.0 %

100 200

Weight (%)

36.6 % 180 60

160

40

Temperature (°C)

80

42.0 % 20 140 0

40

80

120

160

200

240

280

Time (minutes) Fig. 8. Thermogravimetric isothermal analysis of PHB between 140 and 210  C.

30 min. As can be expected, the weight loss increases as temperature is raised, reaching 100% at 210  C. It is important to point out that the initial weight loss temperature is very close to the melting point of PHB (193  C in Hi-Res 8). Moreover keeping PHB at lower temperatures for 30 min, the degradation temperature decreases to 180  C. 4. Conclusions These results confirm the sensitivity of the GCeFTIR technique in the study of PHB degradation. Thus, it has been possible to identify the primary volatile products as 2-butenoic acid experiences decarboxylation, giving rise to carbon dioxide and propene. However, this technique cannot be applied for quantitative analysis, especially when the nature of the degradation products is clearly different. Hi-Res TGA has proven to be a powerful analysis tool in the determination of the initial degradation temperature of PHB, showing a significant resolution improvement in comparison with the conventional one. The minimum temperature for PHB degradation is a function of the heating time. Therefore, the two parameters (temperature and time) must be considered in order to establish the optimal processing conditions. Acknowledgments We express our thanks to the University of the Basque Country for its continuous support through the Consolidated Groups Program (9/UPV 00203.21513519/2001).

References [1] Murase T, Iwata T, Doi Y. Direct observation of enzymatic degradation behaviour of poly((R)-3-hydroxybutyrate) lamellar single crystals by atomic force microscopy. Macromolecules 2001; 34:5848e53. [2] Doi Y, Sudesh K, Abe H. Synthesis, structure and properties of polyhydroxyalkanoates: biological polyesters. Prog Polym Sci 2000;25:1503e55. [3] Doi Y, Kunioka M, Nakamura Y, Soga K. Nuclear magnetic resonance studies on poly(b-hydroxybutyrate) and a copolyester of b-hydroxybutyrate and b-hydroxyvalerate isolated from Alcaligenes eutrophus H16. Macromolecules 1986;19:2860e4. [4] Abe H, Matsubara I, Doi Y. Physical properties and enzymatic degradability of polymer blends of bacterial poly[(R)-3-hydroxybutyrate] and poly[(R, S )-3-hydroxybutyrate] stereoisomers. Macromolecules 1995;28:844e53. [5] Focarete ML, Ceccorulli G, Scandola M, Kowalczuk M. Further evidence of crystallinity-induced biodegradation of synthetic atactic poly(3-hydroxybutyrate) by PHB-depolymerase A from Pseudomonas lemoignei. Blends of atactic poly(3-hydroxybutyrate) with crystalline polyesters. Macromolecules 1998;31: 8485e92. [6] Yu P, Chua H, Huang A, Lo W, Cheng G. Conversion of food industrial wastes into bioplastics. Appl Biochem Biotechnol 1998; 70:603e14. [7] King PP. Biotechnology. An industrial view. J Chem Technol Biotechnol 1982;32:2e8. [8] Grassie N, Murray EJ, Holmes PA. The thermal degradation of poly(-(D)-b-hydroxybutyric acid). Part I. Identification and quantitative analysis of products. Polym Degrad Stab 1984;6: 47e61. [9] Nguyen S, Yu G, Marchessault R. Thermal degradation of poly(3-hydroxyalkanoates): preparation of well-defined oligomers. Biomacromolecules 2002;3(1):219e24. [10] Lehrle RS, Williams RJ. Thermal degradation of bacterial poly(hydroxybutyric acid): mechanisms from the dependence of pyrolysis yields on sample thickness. Macromolecules 1994;27: 3782e9. [11] Lehrle RS, Williams RJ, French C, Hammond T. Thermolysis and methanolysis of poly(b-hydroxybutyrate): random scission

354

[12]

[13]

[14]

[15]

A. Gonzalez et al. / Polymer Degradation and Stability 87 (2005) 347e354 assessed by statistical analysis of molecular weight distributions. Macromolecules 1995;28:4408e14. Nishida H, Tokiwa Y. Effects of higher-order structure of poly(3hydroxybutyrate) on its biodegradation. I. Effects of heat treatment on microbial degradation. J Appl Polym Sci 1992;46:1467e76. Abate R, Ballistreri A, Montaudo G, Impallomeni G. Thermal degradation of microbial poly(4-hydroxybutyrate). Macromolecules 1994;27:332e6. Kunioka M, Doi Y. Thermal degradation of microbial copolyesters: poly(3-hydroxybutyrate-co-3-hydroxyvalerate) and poly(3-hydroxybutyrate-co-4-hydroxybutyrate). Macromolecules 1990;23:1933e6. Ballistreri A, Montaudo G, Garozzo D, Giuffrida M, Montaudo MS. Microstructure of bacterial poly(b-hydroxybutyrate-co-b-hydroxyvalerate) by fast atom bombardment mass spectrometry analysis of the partial pyrolysis products. Macromolecules 1991;24:1231e6.

[16] Gill PS, Sauerbrunn SR, Crowe BS. High resolution thermogravimetry. J Therm Anal 1992;38:255e66. [17] Li S, Yu P, Cheung M. Thermogravimetric analysis of poly(3hydroxybutyrate) and poly(3-hydroxybutyrate-co-3-hydroxyvalerate). J Appl Polym Sci 2001;80:2237e44. [18] Socrates G. Infrared characteristic group frequencies. Chichester: John Wiley & Sons; 1994. [19] Fei B, Chen C, Wu H, Peng S, Wang X, Dong L. Quantitative FTIR study of PHBV/bisphenol A blends. Eur Polym J 2003;39: 1939e46. [20] Shibano J, Matsumoto T, Ishida T, Onda M, Sakaizumi T, Ohashi O, et al. Microwave spectra and conformation of trans-2butenoic acid. J Mol Struct 1988;190:377e85. [21] Li S, He J, Yu P, Cheng M. Thermal degradation of poly(3hydroxybutyrate) and poly(3-hydroxybutyrate-co-3-hydroxyvalerate) as studied by TG, TGeFTIR, and PyeGC/MS. J Appl Polym Sci 2003;89:1530e6.