Polymer Degradation and Stability 79 (2003) 529–533

www.elsevier.com/locate/polydegstab

Methanolysis of polyethylene terephthalate (PET) in the presence

of aluminium tiisopropoxide catalyst to form dimethyl
terephthalate and ethylene glycol
Hideki Kurokawa*, Masa-aki Ohshima, Kazuo Sugiyama, Hiroshi Miura
Department of Applied Chemistry, Faculty of Engineering, Saitama University, Shimo-Okubo 255, Saitama-shi, Saitama 338-8570, Japan

Received 3 April 2002; received in revised form 18 September 2002; accepted 18 September 2002

Abstract
Aluminium triisopropoxide (AIP) promoted the methanolysis of polyethylene terephthalate (PET) to form monomers, dimethyl
terephthalate (DMT) and ethylene glycol (EG), in an equimolar ratio. The methanolysis at 200  C in methanol with an AIP catalyst
gave DMT and EG in 64% and 63% yields, respectively. The yields were increased by using a toluene/methanol mixed solvent
containing 20–50 vol.% toluene; maximum yields, 88% for DMT and 87% for EG, were obtained at 20 vol.% toluene. These
results indicate that the rate of methanolysis strongly depends on the solubility of PET. The results of GPC analysis suggest that the
methanolysis of PET in the absence of the catalyst includes three steps. In the first step, the depolymerisation occurred at a tie
molecule connecting PET crystals and the chain length was shortened to about 1/3. The shortened chain was depolymerized to
oligomers in the second step. The GPC curve of the oligomers tailed to low molecular weight, clearly indicating that the depoly-
merization took place at random positions on the polymer chain. The third step, the depolymerisation from the oligomers to the
monomers, was promoted only in the presence of the AIP catalyst.
# 2002 Elsevier Science Ltd. All rights reserved.
Keywords: PET; Methanolysis; Aluminium triisopropoxide; Chemical recycling; Catalyst

1. Introduction decomposition of PET in non-aqueous media to form

TPA salt, which is analogous to hydrolysis under basic
Consumption of polyethylene terephthalate (PET) conditions, has also been reported [7–8]. Meanwhile,
has increased considerably in recent years, because it glycolysis of PET with ethylene glycol in the presence
has come into use for beverage containers. Used PET of a catalyst yields bis-2-hydroxyethyl terephthalate
containers are mainly recycled to polyester textiles (BHET) [9–12].
(material recycle), but the supply of PET for recycling Methanolysis of PET affords dimethyl terephthalate
will exceed the demand for recycled textile for the (DMT) and ethylene glycol. These products are recycled
foreseeable future. Chemical recycling of PET to to the transesterification process of PET production and
monomers constitutes effective resource utilization, enjoy some advantages over the products of hydrolysis
because processes for such recycling can be carried out and glycolysis; for example, easy purification steps.
repeatedly without substantial waste. Three methods Methanolysis has been attempted by Lotz et al. [13] and
(hydrolysis, glycolysis and methanolysis) have been Gruschke et al. [14], as described in the references,
proposed for the chemical recycling of PET [1]. In which are patents. In the process of Gruscheke et al.,
hydrolysis, PET is depolymerized to terephthalic acid PET was completely (more than 99%) depolymerized to
(TPA) and ethylene glycol (EG) with water, when DMT and EG by reacting molten PET with methanol at
necessary, under acidic or basic conditions [2–6]. Alkali 210  C in the absence of a catalyst. Some patents [15–
18] teach improved processes for increasing the quality
* Corresponding author. Tel.: +81-48-858-9547; fax: +81-48-858-
of the products and lowering recycling costs.
9547. Sako et al. [19,20] proposed a new process of metha-
E-mail address: [email protected] (H. Kurokawa). nolysis by treating PET with supercritical methanol. In
0141-3910/03/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0141-3910(02)00370-1
530 H. Kurokawa et al. / Polymer Degradation and Stability 79 (2003) 529–533

this process PET can be completely depolymerized to 3. Results and discussion

DMT, EG, and oligomers above 300  C at 11 MPa for
30 min without a catalyst. 3.1. Methanolysis in methanol
As is well known, methanolysis, which is a form of
transesterification, proceeds effectively in the presence Detailed results are summarized in Table 1. The first
of catalysts. The patent literature discloses useful cata- run (Run 1) was a blank test carried out at 200  C in the
lysts for the methanolysis of PET, such as zinc acetate, absence of a catalyst. Small amounts of DMT and EG
manganese salts (acetate and phosphate), calcium were formed, in yields of 2.9% and 3.6%, respectively.
phosphate, oxides and hydroxides of alkali metals, After the products were dissolved in 100ml of methanol,
oxides and hydroxides of alkali-earth metals, sodium most of the PET (92.2%) was recovered as chips by fil-
silicate, and lead oxide [2–5], but the results of reac- tration, and a large number of checkered cracks [Fig. 1
tions using such catalysts are not actually shown. (a) and (b)] were observed on the chip surface by SEM.
Meanwhile, results are shown for use of magnesium These results indicate that most of the PET chips failed
acetate for glycolysis and use of stannous oxalate for to dissolve in methanol under the reaction conditions.
alcoholysis with an alcohol having 6–15 carbon atoms Runs 2–5 show the results of methanolysis at 160–
[12,21]. 200  C in the presence of AIP catalyst. At 160  C, the
Although chemical recycling of PET by methanolysis methanolysis proceeded only slightly (Run 2), and PET
requires catalysts exhibiting high activity, the depoly- was recovered in the same shape as before [Fig. 1(c)]. At
merisation of PET in the presence of catalysts has not 180–190  C, PET was depolymerized to DMT and EG
been widely studied. In the present work, we have (10–20% yield) in equimolar ratio (Run 3–4). The form
investigated the methanolysis of PET to form DMT and of recovered PET was the same as before, but the sur-
EG in the presence of aluminium triisopropoxide (AIP). face of PET chips became rough [Fig. 1(d)]. These
AIP serving as a Lewis acid catalyst is known to pro- results suggest that PET was gradually depolymerized
mote transesterification of dialkylcarbonates with to the monomers from the chip surface. When the tem-
phenol [22]. perature was raised from 190 to 200  C, the yields of
DMT and EG increased dramatically, from 18.6% and
18.7% (Run 4) to 67.3% and 68.3% (Run 5), respec-
2. Experimental tively. The insoluble component was recovered as pow-
dery PET. About 27% of the charged PET could not be
Aluminium triisopropoxide (> 95%) was purchased recovered after methanolysis at 200  C, indicating that
from Wako Pure Chemical Industries, Ltd. Methanol the unidentified products dissolved in methanol at
and toluene were supplied from Kanto Chemical CO., ambient temperature. Methanolysis of PET is generally
Inc. Those chemicals were used without further pur- considered to take place at random positions on the
ification. PET specimens were prepared by cutting a polymer chain, decreasing the chain length and finally
beverage bottle to square chips measuring 3 mm3 mm. producing DMT and EG. In view of both solubility in
The catalyst, the PET chips, and the solvent were methanol and the above reaction path, the unrecovered
charged into an autoclave equipped with a magnetic component is assumed to exist as oligomers. Indeed,
stirrer and a purge valve. After being sealed, the auto- Sako et al. determined the structure of oligomers in
clave was heated, while the solution was stirred, to methanolysis products by field desorption mass spec-
reaction temperature (160–200  C) by use of an oil bath.
Methanolysis was carried out for 2 h, after which the
Table 1
reactor was immediately cooled in an ice bath. The PET methanolysis in methanol in the presence of aluminium triipro-
reaction mixture was dissolved in 100 ml methanol. proxidea
About 1 ml of the suspended solution was filtered, and
Run no. Reaction Catalyst charged Yield (%) Recovered PETb
the filtrate was analysed by a gas chromatograph
temp ( C)
equipped with a capillary column (DB-WAX, 30 m). (mg) (nmol) DMT EG (%) Shape
Finally, the remaining solution was filtered, and the
residue was dried and weighed as a recovered PET. 1 200 0 0 2.9 3.6 92.2 Chip+powder
2 160 50 0.245 0.8 0 100 Chip
The average molecular weight and molecular weight 3 180 50 0.245 10.5 8.8 84.1 Chip
distribution of the PET specimen were determined by a 4 190 50 0.245 18.6 18.7 76.0 Chip
GPC (TOSOH, SEC-PAD) equipped with an UV 5 200 50 0.245 67.3 68.3 4.6 Powder
detector, and a TSKgel GMHhr-M column (length=60 6 200 32 0.157 57.4 60.9 12.3 Powder
cm). The values were normalized by use of polystyrene 7 200 12 0.059 20.2 24.0 60.8 Chip
standards. A scanning electron microscope (SEM, a
Reaction conditions; reaction time=2 h, PET=0.50 g (2.60 mmol
Hitachi S-4100) was used for surface observation of the unit), methanol=10 ml (247 mmol).
b
recovered PET specimen. Recovered from the reaction mixture by filtration.
 H. Kurokawa et al. / Polymer Degradation and Stability 79 (2003) 529–533 531

Fig. 1. SEM images of PET recovered after methanolysis. The samples were obtained by methanolysis (a), (b) at 200. in the absence of catalyst, (c)
at 160  C in the presence of catalyst, and (d) at 190  C in the presence of catalyst.

trometry; the structures were mainly HOOC–C6H4– PET was recovered after the blank test (Run 8) in 20
COOCH2CH2OH and HOOC–C6H4–COOCH2- vol.%-toluene/methanol, whereas original PET was
CH2OOC–C6H4–COOH [19]. recovered after methanolysis in methanol without AIP
Runs 5–7 represent the relationship between yield and catalyst (Run 1). These results indicate that the solubi-
the amount of catalyst charged. The yields of monomers lity of PET increased by addition of toluene to
increased with the amount of catalyst up to 32 mg and methanol.
did not depend on the catalyst loading at yields higher When toluene content was raised from 50 vol.% to 80
than 60% because of equilibrium between the mono- vol.%, the yields of DMT and EG decreased from
mers and oligomers. 82.0% and 83.4% (Run 11) to 25.2% and 31.6% (Run
13), respectively. One possibility is the lower concen-
3.2. Methanolysis in methanol/toluene mixed solvent tration of methanol.

The AIP catalyst promotes the methanolysis of PET Table 2

in methanol at 200  C, however, the yields of DMT and PET methanolysis in methanol/toluene in the presence of aluminium
EG are insufficient for practical use. The moderate rate triisopropoxidea
of depolymerization is attributed to the low solubility of Run no. MeOH:Tol (ml:ml) Yield (%) Recovered PETb
PET in methanol. A modification with mixed solvent
DMT EG (%) Shape
was previously carried out by Hu et al. in the methano-
lysis of polycarbonate to form dimethyl carbonate and 8c (8:2) 2.9 4.6 67.9 Powder
bisphenol A [23]. They reported that methanolysis was 9 (8:2) 88.5 87.2 0 –
10 (6:4) 86.8 85.1 0 –
promoted in methanol/toluene mixed solvent containing
11 (5:5) 82.0 83.4 0 –
20 vol.% toluene. 12 (4:5) 63.6 62.3 12.4 Powder
To increase the solubility of PET, a mixed solvent of 13 (2:8) 25.2 31.6 62.0 Powder
methanol and toluene was employed. The results of a
Reaction conditions: reaction temperature=200  C, reaction
PET methanolysis in methanol/toluene are summarized
time=2 h, PET=0.50 g (2.60 mmol unit), catalyst weight=50 mg
in Table 2. The yield of DMT increased from 67.3% in (0.245 mmol), solvent=10 ml.
methanol (Run 5) to 82–89% in the mixed solvent b
Recovered from the reaction mixture by filtration.
containing 20–50 vol.% toluene (Runs 9–11). Powdery c
In the absence of catalyst.
532 H. Kurokawa et al. / Polymer Degradation and Stability 79 (2003) 529–533

Methanolysis of PET under severe conditions is both powder and chips became sharp in comparison
known to yield diethylene glycol (DEG) [1]. This con- with that of the initial PET. These findings suggest that,
densation of EG is catalyzed by an acidic–COOH group when Mn is above 4000, depolymerization does not take
of partially hydrolysed PET [24]. The formation of place at random positions of the polymer chain. When
DEG was also reported in hydrolysis of PET in the methanolysis was carried out in the presence of the cat-
presence of an acid catalyst [25]. Therefore, we carefully alyst (Run 6, AIP=32 mg), most PET was depolymer-
analyzed the formation of DEG by a GC in Run 8. As a ized to the monomers and 12.3% for the initial charge
result, we confirmed the formation of DEG in small was recovered. Mn of the recovered PET was equal to
quantity (about 0.8%), but did not detect 2-methoxy- 2200 and its GPC curve tailed to low molecular weight
ethanol formed by condensation of ethylene glycol with [Fig. 2 (c), Mw/Mn=2.6]. This means the occurrence of
methanol. These results clearly indicate that the AIP depolymerization at random positions on the polymer
catalyst does not promote the undesirable consecutive chain.
reaction of EG even when EG concentration is rela- As shown in Fig. 2 (d), the molecular weight dis-
tively high. tribution of the recovered PET in Run 8 (methanol/tol-
uene mixed solvent, no catalyst) is broad (Mw/
3.3. GPC analysis Mn=3.8), and peaks derived from oligomers are
observed in the low molecular weight region, clearly
The average molecular weights (Mn and Mw) of indicating that PET was easily depolymerized in the
recovered PET were determined by GPC for investigat- mixed solvent at random positions on the chain. Mn
ing the details of methanolysis, and Fig. 2 shows the (1400) of recovered PET in Run 8 was one-quarter that
GPC curves. Average molecular weight of the starting (5500) in Run 1 and half that in Run 6, but the yield
PET was Mn=16,000, and its distribution was Mw/ (2.9%) of DMT in Run8 is extremely low in comparison
Mn=3.4 [Fig. 2 (a)]. After methanolysis at 200  C in the with that (57.4%) in Run 6.
absence of catalyst (Run 1), PET was recovered as chips The results of GPC analysis suggest that the metha-
(Fig. 2 (b)) with a small amount of powder. Mn of the nolysis in the absence of a catalyst includes at least three
powder portion (4300) was nearly equal to that of the steps represented in Fig. 3. In the first step of methano-
chips (5500) and about one-third lower than that of the lysis, the chain length of PET was shortened to about 1/
starting PET. The molecular weight distributions of 3 [Fig. 3 (i)]. Even when the catalyst is not used, the first
step takes place easily [Fig. 2 (b)]. The details of this
step are not well understood, however the polymer
chain might be cut off at a tie molecule because such
molecules exist in an amorphous part of PET and have
high accessibility. Collins and Zeronian [26] reported
similar results for alkaline hydrolysis of fabric PET;
after the hydrolysis, crystalline material was deposited
on the fibre surface, and its molecular weight was
approximately 2400.
A significant difference in GPC curves was observed
between samples obtained by reacting PET in methanol
and those in toluene/methanol mixed solvent. In tolu-
ene/methanol, most PET was depolymerized to oligo-
mers in the absence of the catalyst [second step, Fig. 3
(ii)] but the peak (log M=4) derived from the crystalline
material was still shown in Fig. 2 (d). This fact denotes
that the oligomers are produced through the first step.
The mixed solvent accelerates the formation of oligo-
mers effectively and this promoting effect is due to
swelling of PET crystals.
The third step [Fig. 3 (iii)], the formation of mono-
mers from oligomers, is promoted only in the presence
of catalyst and the monomers were produced in high
yield. Those results suggest that the catalyst promotes
Fig. 2. GPC curves of PET recovered after methanolysis. Before the formation of monomers from oligomers.
methanolysis. After methanolysis in methanol (b) at 200  C in the
absence of catalyst, and (c) at 200  C in the presence of catalyst. (d)
In conclusion, aluminium triisopropoxide promotes
After methanolysis in methanol/toluene mixed solvent in the absence the methanolysis of PET to form the monomers DMT
of catalyst. and EG. The yields of monomers strongly depend on
 H. Kurokawa et al. / Polymer Degradation and Stability 79 (2003) 529–533 533

[3] Yoshioka Y, Sato T, Okuwaki A. Hydrolysis of waste PET by

sulfuric acid at 150. for a chemical recycling. J Appl Polym Sci
1994;52:1353–5.
[4] Chambret RF. US patent 3,120,561, 1964.
[5] Rosen BI. US patent 5,095,145, 1992.
[6] Benzaria J. US patent 5254,666, 1993.
[7] Hu LC, Oku A, Yamada E, Tomari K. Alkali-decomposition of
poly(ethylene terephthalate) in mixed media of nonaqueous alco-
hol and ether. Study on recycling of poly(ethylene terephthalate).
Polym J 1997;29(9):708–12.
[8] Oku A, Hu LC, Yamada E. Alkali decomposition of poly-
(ethylene terephthalate) with sodium hydroxide in nonaqueous
ethylene glycol: a study on recycling of terephthalic acid and
ethylene glycol. J Appl Polym Sci 1997;63:595–601.
[9] Vaidya UR, Nadkarni VM. Polyester polyols from glycolyzed pet
waste: effect of glycol type on kinetics of polyesterification. J
Appl Polym Sci 1989;38:1179–90.
[10] Chen JY, Ou CF, Hu YC, Lin CC. Depolymerization of poly(-
ethylene terephthalate) resin under pressure. J Appl Polym Sci
1991;42:1501–7.
[11] Dietz HJ, Freese RE. US patent 3,503,904, 1970.
[12] Delattre J, Raynaud R, Thomas C. US patent 4,163,860, 1979.
[13] Lotz R, Wick G, Neuhaus DA. US patent 3,321,510, 1967.
[14] Gruschke H, Hammerschick W, Medem H. US patent 3,403,115,
1968.
[15] Ligorati F, Aglietti G, Nova VE. US patent 3,776,945, 1973.
[16] Everette MR. EP patent 484,963,A2, 1992.
[17] Naujokas AA, Ryan KM. US patent 5,051,528, 1991.
[18] Gamble WJ, Naujokas AA, Debruin BR. US patent 5,298,530,
1994.
[19] Sako T, Sugeta T, Otake K, Nakazawa N, Sato M, Namiki K,
Tsugumi M. Depolymerization of polyethylene terephthalate to
monomers with supercritical methanol. J Chem Eng Japan 1997;
30(2):342–6.
[20] Sako T, Sugeta T, Otake K, Takebayashi Y, Kamizawa C,
Tsugumi M, Hongo M. Kinetic study on depolymerization
of poly(ethylene terephthalate) with methanol at high tem-
perature and pressure. Kobunshi ronbunshu 1998;55(11):685–
Fig. 3. Reaction scheme of PET methanolysis. Solid line: without 90.
catalyst; broken line: with catalyst. [21] Gupta VP, Louis A. US patent 4,929,749, 1990.
[22] Illuminati G, Romano U, Tesei R. Ger. Offen. 2,528,412, 1976
(Chemical Abstract 84:121495h).
the solubility of PET; methanolysis in 20 vol.%-toluene/ [23] Hu LC, Oku A, Yamada E. Alkali-catalyzed methanolysis of
methanol at 200  C in the presence of an aluminium polycarbonate. A study on recycling of bisphenol A and dimethyl
triisopropoxide catalyst produced DMT in 88.5% yield carbonate. Polymer 1998;39:3841–5.
[24] Yang Y, Lu Y, Xiang H, Xu Y, Li Y. Study on methanolytic
and EG in 87.2% yield. depolymerization of PET with supercritical methanol for chemi-
cal recycling. Polym Degrad Stab 2002;75:185–91.
References [25] Kao C-Y, Wan B-Z, Cheng W-H. Kinetics of hydrolytic depoly-
merization of melt poly(ethyleneterephthalate). Ind Eng Chem
[1] Paszun D, Spychaj T. Chemical recycling of poly(ethylene ter- Res 1998;37:1228–34.
ephthalate). Ind Eng Chem Res 1997;36:1373–83. [26] Collins MJ, Zeronian SH. The molecular weight distribu-
[2] Campanelli JR, Kamel MR, Cooper DG. A kinetic study of the tion and oligomers of sodium hydroxide hydrolyzed
hydrolytic degradation of polyethylene terephthalate at high poly(ethylene terephthalate). J Appl Polym Sci 1992;45:
temperatures. J Appl Polym Sci 1993;48:443–51. 797–804.