A Novel Catalyst for the Glycolysis of Poly(ethylene

terephthalate)

K. Troev,1 G. Grancharov,1 R. Tsevi,1 I. Gitsov2

1
Institute of Polymers, Bulgarian Academy of Sciences, Sofia 1113, Bulgaria
2
The Michael M. Szwarc Polymer Research Institute and Department of Chemistry, College of Environmental Science
and Forestry, State University of New York, Syracuse, New York 13210

Received 24 September 2002; accepted 6 January 2003

ABSTRACT: Poly(ethylene terephthalate) waste materials pared with compounds traditionally used in this process like
were depolymerized by ethylene glycol (EG), diethylene Zn(OOCCH3)2. The oligomer distribution in the glycolysis
glycol (DEG), and propylene glycol (PG) in the presence of products was studied by size-exclusion chromatography.
a novel catalytic system: titanium (IV)-phosphate. The new © 2003 Wiley Periodicals, Inc. J Appl Polym Sci 90: 1148 –1152, 2003
catalyst was synthesized through a reaction of TiCl4 with
triethyl phosphate (C2H5O)3P(O). It was found that the de-
polymerization of poly(ethylene terephthalate) fiber pro- Key words: catalysis; glycolysis; organometallic catalysts;
ceeds faster in the presence of titanium (IV)-phosphate com- poly(ethylene terephthalate); recycling; thermoplastics

INTRODUCTION titanium alkoxides compared with the degradation

catalyzed by metal acetates. Regrettably, the former
Increased environmental awareness, legislative mea-
compounds also catalyze side reactions that yield
sures, and public demand for environmental sustain-
products with an undesirable yellowish color.14 On
ability are leading to an increased interest in plastics
the other side, it was found that during the PET syn-
recycling.1 Poly(ethylene terephthalate) (PET), a high
thesis similar reactions could be avoided in the pres-
volume thermoplastic polyester2 that is extensively
ence of thermal stabilizers like trialkyl- or aryl phos-
used in diverse applications—textiles, high strength
phates affording PET with good whiteness and trans-
fibers, photographic films, soft-drink bottles (annually
parency.17–20
1.5 million ton of PET are used in the production of The primary goals of this study are the design,
bottles and jars alone),3 and others—is also one of the synthesis, and evaluation of a new catalyst for PET
largest components of the postconsumer plastics glycolysis that will combine the advantages of the
waste in landfills4 (1,700,000 tons in the U.S. alone1). previously applied compounds while avoiding their
Therefore, the chemical PET decomposition and con- shortcomings. The leading idea is to incorporate tita-
version into reusable chemical products of added nium and alkyl phosphate moieties in a single mole-
value is one of the important recycling strategies for cule that could act simultaneously as a catalyst and as
this material.5,6 Glycolysis with ethylene glycol (EG)7 a thermal stabilizer during the process. Here we report
propylene glycol, and their derivatives is one of the the results from the synthesis of titanium (IV)-phos-
most widely studied processes for PET degradation.8 phate compound and its application in the glycolysis
The reaction is a very sluggish one when performed of PET wastes.
without a catalyst. It was found that the activation
energy for the uncatalyzed glycolysis was 32 kcal/mol
whereas the catalyzed process requires only 19 kcal/ EXPERIMENTAL
mol.9 Therefore various metal-containing catalysts,
such as zinc acetate,8,10 –12 lead-, cobalt-, or manganese Materials
acetate,10 –12 and titanium alkoxides14 –16 were used to PET pellets used for fiber production (Mw ⫽ 30,000)
facilitate the process. The reaction proceeds faster with were supplied by Yambolen Ltd., Bulgaria. PET pellets
for bottle manufacturing (Mw ⫽ 51,000) were supplied
by Bulgarian private company. Ethylene glycol (EG),
Correspondence to: K. Troev ([email protected]) and diethylene glycol (DEG), and 1,2-propylene glycol
I. Gitsov ([email protected]).
(PrG), TiCl4, triethyl phosphate, and Zn (OOCCH3)2
Journal of Applied Polymer Science, Vol. 90, 1148 –1152 (2003) were purchased from Aldrich (Milwaukee, WI) and
© 2003 Wiley Periodicals, Inc. used without further purification.
GLYCOLYSIS OF POLY(ETHYLENE TEREPHTHALATE) 1149

out a catalyst were also performed. The reaction was

monitored by size-exclusion chromatography.

Measurements
The molecular weight characteristics and the oligomer
distribution of the glycolysis products were deter-
mined by size-exclusion chromatography (SEC). The
analyses were performed at 40°C on a Waters 244
instrument, equipped with differential refractive in-
dex detector R410, 450 UV-Vis detector (both from
Waters), and 5 mm American Polymer Standards col-
umns (Linear, 104 Å, and 500 Å) in tetrahydrofuran
Scheme 1 (THF) eluting at 1 mL/min. The molecular weights
were calculated by a conventional calibration with
Synthesis of titanium(IV)-phosphate catalyst, 1 polystyrene standards.
31
P NMR spectra were recorded on a Bruker 500
Triethyl phosphate 18.2 g (0.1 mol) and TiCl4 3.78 g MHz spectrometer in solid state. Thermal studies
(0.02 mol) were placed into a three-necked flask, were performed on a Perkin Elmer DSC-7 differential
charged with 30 mL of high boiling gasoline, and scanning calorimeter in nitrogen atmosphere and
equipped with a capillary for argon purging, ther- heating rate of 5°C/min. The instrument was cali-
mometer, and reflux condenser. The reaction was per- brated by indium and lead standards. FTIR spectra
formed at 120°C for 6 h. The contents in the flask were were recorded on a Bruker Vector 22 instrument in
allowed to cool down to room temperature after the KBr.
completion of the reaction (controlled by the amount
of C2H5Cl evolved). The precipitate formed was sep-
arated by filtration, washed several times with dry RESULTS AND DISCUSSION
ethanol in order to separate the unreacted triethyl
Synthesis of titanium (IV)-phosphate catalyst
phosphate, and then dried at 50°C. Yield 7.9 g (60%).
Titanium (IV)-phosphate ester is chosen as the catalyst
in the PET glycolysis because of the known efficiency
General procedure of PET glycolysis
of phosphate compounds as thermal stabilizers.11 The
A typical experiment is performed as described: 25 g heterochain catalyst 1, containing Ti-O-P sequences, is
(0.13 mol/repeating unit) of PET pellets (fiber grade or synthesized by reacting TiCl4 with triethyl phosphate
bottle grade) were added to 0.36 mol of glycols (EG, (Scheme 1).
DEG, or PrG). The molar ratio of PET repeating unit to Suggestions that polymeric compounds like 1 might
glycol was 1:2.77. The mixtures, together with differ- be formed during this reaction were briefly mentioned
ent amounts of titanium (IV)-phosphate catalyst 1, (in in some early studies21–24 without any experimental
wt% based on the weight of PET), were charged into a description and analytical evidence. The structure of
three-necked flask, equipped with a stirrer, thermom- the individual building block in 1 is confirmed by 31P
eter, and reflux condenser. Comparative experiments NMR and IR spectroscopy. The 31P{H} NMR spectrum
in the presence of Zn(OOCCH3)2 (0.3 wt%) and with- shows only one quintet signal at with d ⫽ ⫺0.24 ppm

TABLE I
Reaction Conditions and Main Products During PET (Fiber Grade) Glycolysis with EG
PET (mol/ Reaction SEC analysis of glycolyzed products
repeating EG Cat. Temperature time
unit) (mol) (wt %) (°C) (min) BHET (%) Dimer (%) Trimer (%) Tetramer (%)

0.13 0.36 — 190–200 300 96.4 2.4 0.9 0.3

0.13 0.36 0.05 190–200 195 95.3 3.5 1.0 0.2
0.13 0.36 0.1 190–200 180 96.5 2.4 0.8 0.2
0.13 0.36 0.2 190–200 160 95.5 3.2 1.0 0.3
0.13 0.36 0.3 190–200 150 97.5 1.6 0.7 0.2
0.13 0.36 0.5 190–200 155 95.3 3.5 1.0 0.3
0.13 0.36 0.3a 190–200 150 62.8 26.8 8.6 1.8
a
Catalyst ⫽ Zn(OOCCH3)2.
1150 TROEV ET AL.

TABLE II
Characterization Data for PET (Fiber Grade) Glycolysis with DEG
PET (mol/ Reaction SEC analysis of glycolyzed products
repeating DEG Cat. Temperature time
unit) (mol) (wt %) (°C) (min) BHEET (%) Dimer (%) Trimer (%) Tetramer (%)

0.13 0.36 — 220 50 12.8 2.8 81.8 2.0

0.13 0.36 0.3 220 12 91.2 4.5 2.8 1.5
0.13 0.36 0.5 220 13 90.4 4.6 3.1 1.9

and 3J(P,H) ⫽ 7.7 Hz. The phosphorus chemical shift is reaction time is more complex (Table I, run 7). It
shifted upfield by 0.50 ppm with respect to triethyl contains 62.8% BHET, 26.8% dimer, 8.6% trimer, and
phosphate (d ⫽ 0.26 ppm) due to the presence of an 1.8% tetramer. It is obvious that the full decomposition
ionic bond in this structure.25 In the IR spectrum of 1 of PET is significantly facilitated in the presence of 1
the P ⫽ O stretching frequency is lowered by approx- compared with process catalyzed by Zn(OOCCH3)2
imately 235 cm⫺1 when compared with the analogous under identical reaction conditions.
band in triethyl phosphate. These data confirm that In the case of DEG the depolymerization is per-
the phosphorus atom is surrounded by two ethoxy formed at 220°C. The experimental results show that
groups. 0.3 wt % of 1 is again the optimal catalyst concentra-
tion because the decomposition is complete within 20
min, compared to the same process run for 65 min
Glycolysis of fiber grade PET
without a catalyst (Table II). Bis[2(2-hydroxyethoxy)-
The glycolysis is carried out with EG, DEG, or PrG ethyl] terephthalate (BHEET) is the main product of
with the molar ratio PET to glycol kept constant at the glycolysis.
1:2.77. The depolymerization with PrG is carried out at the
During the glycolysis with EG the temperature of boiling temperature of the glycol (180°C). Similar to
the oil bath is kept at 200°C, close to the boiling the EG case, the reaction temperature increases during
temperature of the glycol. It is also observed that the the decomposition process. The optimal concentration
temperature of the system increases from 190 to 200°C of the catalyst 1 is also 0.3 wt %. The depolymerization
during the depolymerization process. Different con- proceeds at this concentration for 480 min, whereas
centrations of the catalyst 1 were investigated (Table I, depolymerization without catalyst runs for 885 min.
runs 2–5). It is found that the reaction proceeds faster SEC analysis shows that the main product is bis(2-
in the presence of the titanium (IV)-phosphate ester, hydroxypropyl) terephthalate (BHPT) (Table III).
(Table I). At 0.3 wt % of 1, the optimal concentration of
the catalyst, the reaction time is reduced from 300 min
Glycolysis of bottle grade PET
(noncatalyzed glycolysis, Table I, run 1) to 150 min.
SEC analysis shows (Table I) that the decomposition From the experimental data presented in Table IV it is
mixtures consist mainly of bis(2-hydroxyethyl) tereph- clear that the reaction time with EG is reduced from
thalate (BHET). DSC thermograms of the products 480 min for the noncatalyzed depolymerization (run 1)
contain two endothermic peaks. The first peak at to 105 min for catalyzed process (0.3 wt % of 1, run 2).
107.8°C, can be attributed to the melting transition of It should be mentioned that the content of BHET at the
BHET. The second peak at 124.8°C is small and well catalyzed depolymerization (66.7%) is notably higher
below the reported melting temperature of the BHET compared with the noncatalyzed process (37.4%). Sur-
dimer (162°C).13 The glycolysis mixture produced in prisingly, the results obtained show that in this case of
the presence of 0.3 wt % of Zn(OOCCH3)2 and 150 min the glycolysis there is no substantial difference in the

TABLE III
Reaction Conditions and Main Products During PET (Fiber Grade) Glycolysis with PrG
PET (mol/ Reaction SEC analysis of glycolyzed products
repeating PEG Cat. Temperature time
unit) (mol) (wt %) (°C) (min) BHPT (%) Dimer (%) Trimer (%) Tetramer (%)

0.13 0.36 — 180–194 885 94.1 3.3 1.7 0.9

0.13 0.36 0.3 180–188 480 93.6 3.6 1.8 1.0
0.13 0.36 0.5 180–188 530 93.3 3.7 1.9 1.1
GLYCOLYSIS OF POLY(ETHYLENE TEREPHTHALATE) 1151

TABLE IV
Characterization Data for PET (Bottle Grade) Glycolysis with EG
PET (mol/ Reaction SEC analysis of glycolyzed products
repeating EG Cat. Temperature time
unit) (mol) (wt %) (°C) (min) BHET (%) Dimer (%) Trimer (%) Tetramer (%)

0.13 0.36 — 190–200 480 37.4 23.7 24.4 13.9

0.13 0.36 0.3 190–200 105 66.7 28.9 4.4 —
0.13 0.36 0.5 190–200 110 62.9 29.1 7.9 —
0.13 0.36 0.3a 190–200 105 61.1 28.5 8.7 1.7
a
Catalyst ⫽ Zn(OOCCH3)2.

TABLE V
Characterization Data for PET (Bottle Grade) Glycolysis with DEG
PET (mol/ Reaction SEC analysis of glycolyzed products
repeating DEG Cat. Temperature time
unit) (mol) (wt %) (°C) (min) BHEET (%) Dimer (%) Trimer (%) Tetramer (%)

0.13 0.36 — 220 16 9.7 21.0 32.0 37.3

0.13 0.36 0.3 220 8 64.2 3.3 2.1 30.4
0.13 0.36 0.5 220 8 58.0 8.6 6.5 26.9

TABLE VI
Reaction Conditions and Main Products During PET (Fiber Grade) Glycolysis with PrG
PET (mol/ Reaction SEC analysis of glycolyzed products
repeating PEG Cat. Temperature time
unit) (mol) (wt %) (°C) (min) BHPT (%) Dimer (%) Trimer (%) Tetramer (%)

0.13 0.36 — 180–194 1210 38.8 29.8 19.6 11.8

0.13 0.36 0.3 180–190 285 52.8 27.8 11.6 5.0
0.13 0.36 0.5 180–190 400 47.7 27.9 13.2 11.2

activity of catalyst 1 and Zn(OOCCH3)2. The glycoly- chloride and triethyl phosphate can be used as a cat-
sis of bottle grade PET in the presence of 0.3 wt % of alyst for PET glycolysis with EG, DEG, and PrG. It is
Zn(OOCCH3)2 (Table IV, run 4) generates the follow- found that the optimal concentration of 1 is 0.3 wt %.
ing decomposition products: BHET 61.1%, dimer The depolymerization rate of fiber grade PET substan-
28.5%, trimer 8.7%, and tetramer 1.6%. tially increases in the presence of this catalyst. At
The depolymerization with DEG was carried out at comparable reaction times (195 min, Table I; 480 min,
220°C. The glycolysis is faster in the presence of 1 Table III) the depolymerization achieved by the un-
(Table V, runs 2 and 3). The SEC analysis shows that catalyzed process is less than 60%. At these times the
the content of the monomer, dimer, trimer, and tet- reaction produces mostly oligomers. The mechanism
ramer for the noncatalyzed depolymerization process of the catalyzed reaction most probably involves com-
is several times lower, compared with the process plexes of the polymer, catalyst and the corresponding
where the glycolysis is carried out in the presence of a
glycol through hydrogen bond formation.26,27 It
catalyst 1 (Table 5).
should be mentioned that 1 shows markedly higher
The same trend is observed for the PET glycolysis
catalytic activity in this system compared with Zn(O-
by PrG. At the optimal concentration of 1 (0.3 wt %)
OCCH3)2. In the case of the depolymerization of bottle
the glycolysis proceeds for 285 min. In the absence of
a catalyst the reaction time is significantly longer, at grade PET no substantial difference in the activity of 1
1210 min. The content of BHPT is 52.8% for catalyzed and Zn(OOCCH3)2 is observed. However, the color of
depolymerization of PET whereas for the noncata- the degradation products is notably better. The good
lyzed process it is 38.8% as revealed by SEC analysis product whiteness observed in all depolymerizations
(Table VI). studied is an additional advantage of the new catalyst
system.

CONCLUSIONS
The authors thank the Bulgarian National Foundation for
The results obtained show that the titanium (IV)-phos- Scientific Investigations (grant X-811) and Allied Signal for
phate (1) synthesized by reaction of titanium tetra- the financial support of this research.
1152 TROEV ET AL.

References 15. Park, L. S.; Yoon, L. H. Pollimo 1994, 18, 693.

16. Billiau-Loreau, M.; Durand, G.; Tersac, G. Polymer 2002, 43, 21.
1. Subramanian, P. M. Res Cons Recycl 2000, 28, 253. 17. Chimura, K.; Itoh, K.; Takashima, Sh.; Shindoh, M.; Kawashima,
2. Fakirov, S. Handbook of Thermoplastic Polyesters: Homopoly- M. (Mitsubishi Rayon Co. Ltd) Jpn Kokai 1973, 73, 35954; C A
mers, Copolymers, Blends and Composites; John Wiley & Sons; 1975, 83, 59874e.
New York, 2002. 18. Tsuji, T.; Takeuchi, N. (Kuraray Co. Ltd) Jpn Kokai 1977, 77, 63,
3. Paszun D.; Spychaj T. Ind Eng Chem Res 1997, 36, 1375. 289; C A 1977, 87, 119161q.
4. Yoda, N. Polym Int 1999, 48, 944. 19. Okumura, N.; Kuno, M.; Norimatsu, Y. (Teijin Ltd) Jpn Kokai
5. Lin, C. C. Macromol Symp 1998, 135, 129. 1979, 79, 45937; C A 1979, 91, 57841h.
6. Spychaj T.; Paszun D. Macromol Symp 1998, 135, 137. 20. Shindo, M.; Matsuda, M. (Mitsubishi Rayon Co. Ltd) Jpn Kokai
7. Korshak, V. V.; Bekasova, N. I.; Zamyatina, V.A. Izv Akad Nauk Tokyo Koho 1980, 79, 101897; C A 1980, 92, 43211m.
USSR, Ser Khim 1958, 614. 21. Andrianov, K. A.; Kuznetsova, I. K. Izv Akad Nauk SSSR Ser
8. Chen, J. W.; Chen, L. W. J Appl Polym Sci 1999, 73, 35. Khim 1964, 4, 651.
9. Simoto, M.; Kito, A.; Inoue, R. Kobunshi Kagaku 1958, 15, 664. 22. Andrianov, K. A.; Kuznetsova, I. K. Izv Akad Nauk SSSR Ser
10. Vaidya, U. R.; Nadkarni, V. M. J Appl Polym Sci 1987, 34, 235. Khim 1965, 6, 945.
11. Vaidya, U. R.; Nadkarni V. M. J Appl Polym Sci 1989, 38, 1179. 23. Andrianov, K. A.; Kuznetsova, I. K.; Smirnow, Yu. N. Izv Akad
12. Campanelli, J. R.; Kamal, M. R.; Cooper, D. G. J Appl Polym Sci Nauk SSSR Neorg Mat 1965, 1, 301.
1994, 54, 1731. 24. Pandit, S. K.; Nerlekar, P. G.; Gupta, J. Indian J Chem 1967, 5,
13. Baliga, S.; Wong, W. T. J Polym Sci Part A Polym Chem 1989, 27, 121.
2071. 25. Troev, K. Rev Heteroatom Chem 1995, 13, 99
14. Nawata, K. (Teijin Ltd.) Jpn Kokai 1977, 77, 57, 291; C A 1977, 87, 26. Gramstad, T. Acta Chem Scand 1961, 15, 1337.
69240s. 27. Aksnes, G.; Gramstad, T. Acta Chem Scand 1960, 14, 1485.