Copyright © 2011 American Scientific Publishers Journal of

All rights reserved Nanoscience and Nanotechnology

Printed in the United States of America Vol. 11, 824–828, 2011

Metal-Oxide-Doped Silica Nanoparticles for the

Catalytic Glycolysis of Polyethylene Terephthalate
Muhammad Imran1 , Kyoung G. Lee1 , Qasim Imtiaz1 , Bo-kyung Kim2 ,
Myungwan Han2 , Bong Gyoo Cho3 , and Do Hyun Kim1
∗
1
Department of Chemical and Biomolecular Engineering and Center for Ultramicrochemical Process Systems,
Korea Advanced Institute of Science and Technology (KAIST), 373-1 Gusong-dong, Yuseong-gu, Daejeon 305-701, Korea
2
Department of Chemical Engineering, Chungnam National University, 220 Gung-dong,
Yuseong-gu, Daejeon 305-764, Korea
3
Mineral Processing Department, Korea Institute of Geoscience and Mineral Resources, Gwahang-no 92,
Yuseong-gu, Daejeon 305-350, Korea

Polyethylene terephthalate (PET) was depolymerized to monomer bis(2-hydroxyethyl) terephthalate

(BHET) using excess ethylene glycol (EG) in the presence of metal oxides that were impregnated
on different forms of silica support [silica nanoparticles (SNPs) or silica microparticles (SMPs)] as
glycolysis catalysts. The reactions were carried out at 300  C and 1.1 MPa at an EG-to-PET molar
ratio of 11:1 and a catalyst-to-PET-weight ratio of 1.0% for 40–80 min. Among the four prepared cat-
alysts (Mn3 O4 /SNPs, ZnO/SNPs, Mn3 O4 /SMPs, and ZnO/SMPs), the Mn3 O4 /SNPs nanocomposite
had the highest monomer yield (>90%). This high yield may be explained by the high surface area,
amorphous and porous structure, and existence of numerous active sites on the nanocomposite
Delivered by Publishing Technology to: Adelaide Theological Library
catalyst. The BHET yield increased with time and reached the highest level where equilibrium was
IP: 218.23.112.86 On: Sun, 24 Jan 2016 07:14:01
established between BHET and its dimer.
Copyright: The catalysts
American were
Scientific characterized by their SEM, TEM,
Publishers
RESEARCH ARTICLE

and BET surface areas, and via XRD, whereas the monomer BHET was characterized by HPLC
and FT-IR. The glycolysis with the Mn3 O4 /SNPs nanocomposite as the glycolysis catalyst produced
a maximum BHET in a short reaction time.
Keywords: PET Recycling, BHET, Depolymerization, Glycolysis, Nanocomposite.

1. INTRODUCTION glycolysis using EG as a depolymerization solvent is

Polyethylene terephthalate (PET) is a widely used resin
for synthetic fibers, audio and video taps, X-ray films,
and softdrink bottles due to its excellent thermal, mechan-
ical, and chemical properties. Industrial and societal con-
cern with the recovery and disposal of huge volumes
of post-consumer plastic waste is growing. Severe legal
restrictions are being imposed on landfills and incineration,
which are the most practiced waste management meth-
ods so far, but the trend is changing rapidly towards the
evolving recycling technologies.1 PET is one of the most
extensively recycled polymeric materials. There are many
processes by which post-consumer plastics can be recy-
cled, but the one that is acceptable for sustainable devel-
opment is the chemical recycling process because it leads
to the formation of monomers from which the polymer
is made.2 Among various chemical recycling processes,
∗
Author to whom correspondence should be addressed.
the oldest, simplest, and least capital intensive process.3
Moreover, the recovered monomer (BHET) from glycoly-
sis can be blended with fresh BHET, and the mixture can
be used again for virgin PET production.
There have been numerous studies on glycolysis that
used different catalysts and different reaction conditions.
Baliga and Wong4 reported the use of metal acetates (zinc,
manganese, cobalt, and lead) as glycolysis catalysts, and
they found that the zinc acetate had the best monomer
yield. Shukla et al.5–7 studied several catalysts (sodium
carbonate, sodium bicarbonate, glacial acetic acid, lithium
hydroxide, sodium sulfate, potassium sulfate, and zeolites)
and showed that their monomer yield is comparable with
that of conventionally used catalysts such as zinc acetate
and lead acetate. Wang et al.8 recently reported the use
of ionic liquids for the depolymerization of PET. Most
glycolysis processes are conducted at a wide range of tem-
peratures, from 160 to 220  C, and with a reaction time
of 0.5–8 h. Because of the long reaction time and the low

824 J. Nanosci. Nanotechnol. 2011, Vol. 11, No. 1 1533-4880/2011/11/824/005 doi:10.1166/jnn.2011.3201

Imran et al. Metal-Oxide-Doped Silica Nanoparticles for the Catalytic Glycolysis of Polyethylene Terephthalate
monomer yield of existing catalysts, a new process with
nano-sized silica supporting metal oxide nanocomposites
as glycolysis catalysts is proposed and compared with
microcomposites of conventional micro-sized silica sup-
porting metal oxide catalysts.
Interest in nanomaterials for various applications is
growing due to their unique optical, electrical, mechan-
ical, and catalytic properties. Over the past decade, the
exceptional and novel activity of nanomaterials, in combi-
nation with recent advances in materials science, has led
to an explosive development in nanocatalysis. 9–11
Since all
the elementary reactions occur at the atomic or molec-
ular scale, catalysis also takes place at the nanoscale
or sub-nanoscale. Nanometer-sized catalysts have specific
physical properties that differ from those of their bulk
counterpart and lead towards higher catalytic activity,
12
selectivity, and stability.
Due to the unique features of nanocomposites at a
reduced length scale, such as a high surface area and uni-
form dispersion, silica nanoparticles (SNPs) were synthe-
sized and used for nanocatalyst support in this study. Their
catalytic performance significantly depends on their prepa-
ration method, the dispersion of active catalytic species,
and the type of the catalyst support. Oxides of manganese
and zinc (Mn3 O4 , ZnO) were impregnated on the silica
nanoparticles (SNPs) through ultrasound irradiation, and
on the silica microparticles (SMPs), to compare their cat-
Delivered by Publishing Technology 2.3.to: Adelaide
alytic activities. The catalysts that were prepared using
IP: 218.23.112.86 On: Sun, 24 Jan 2016 07:14:01
sonochemical reactions showed higher catalytic American
Copyright: activity Scientific Publishers
than those that were prepared using conventional impreg-
nation methods.13 PET was glycolized with EG at 300  C
and 1.1 MPa in the presence of the catalysts that were
prepared as previously described. The effect of the reac-
tion time on the monomer yield was investigated to study
the catalytic performance (in terms of the monomer yield)
of the nanocomposites (Mn3 O4 /SNPs and ZnO/SNPs) and
the microcomposites (Mn3 O4 /SMPs and ZnO/SMPs).
2. EXPERIMENTAL DETAILS
2.1. Materials
Virgin PET pellets (3-mm × 3-mm) were obtained from
SK Chemicals from Korea. These pellets were mixed
with dry ice and ground to fine powder with a particle
size of 202
m in a common grinder. The number aver-
age (MWn ) and weight average (MWw ) molecular weights
of the virgin PET were 11,053 and 38,378, respectively.
Tetraethyl orthosilicate (TEOS), ammonium hydroxide,
manganese (II) nitrate hydrate (Mn(NO3 2 · xH2 O), zinc
nitrate hexahydrate (Zn(NO3 2 · 6H2 O), anhydrous EG,
and standard BHET were purchased from Sigma-Aldrich.
Tetrahydrofuran (THF), ethanol, and SMPs (1–20
m)
were purchased from J. T. Baker, Merck, and Junsei Chem-
icals, respectively. All the reagents were used as pur-
chased, without further purification.
J. Nanosci. Nanotechnol. 11, 824–828, 2011
2.2. Catalyst Synthesis
The silica nanoparticles were synthesized via the Stöber
method,14 with a few modifications. 8.0 ml of ammonium
hydroxide (28 wt%) and 6.0 ml of deionized water were
added to 100 ml of ethanol in a sealed round bottom flask
and mixed for 15 min. 4.7 ml of TEOS was added to this
solution and stirred with a magnetic stirrer at room tem-
perature for 3 h. The resulting precipitates were separated
via centrifugation and washed several times with water and
ethanol, after which they were dried in an oven at 70  C
for 8 h. The dried powder was finally calcined at 500  C
for 12 h.
The silica (SNPs or SMPs)-supported manganese
oxide or zinc oxide was prepared by adding a pre-
determined amount of silica support to the 1.0 M pre-
cursor [Mn(NO 3 2 · xH2 O or Zn(NO3 2 · 6H2 O] solution.
The metal oxide loading was kept at 15 wt%. The solu-
tion was sonicated for 45 min using a horn-type sonicator
(Branson Sonifier). At the beginning of the sonication, a
0.1 M ammonia solution was added to maintain the pH
of the solution at around 9.5. The mixture was separated
via centrifugation and washed with water and ethanol. The
catalyst samples were then dried at 100  C for 8 h and

calcined at 350 C for 3 h.
Glycolysis Theological
of PET Library
RESEARCH ARTICLE
The glycolysis reactions were carried out in a stainless
steel (SS 316) batch-type reactor (10 ml) at 300  C
and 1.1 MPa. A mixture of 0.3 g of PET, 1.1 g of
EG, and 0.003 g of a catalyst (Mn3 O4 /SNPs, ZnO/SNPs,
Mn3 O4 /SMPs, or ZnO/SMPs) were loaded into the reactor,
and the air inside the reactor was purged with argon
to avoid oxidation of the PET and the glycolysis prod-
ucts. The reactor was placed in a furnace at 300  C for
40–80 min. After the given reaction time, the reactor was
quenched in cold water to immediately stop the reaction.
To quantitatively determine the monomer BHET, the reac-
tion products were dissolved in THF and analyzed via
HPLC. The molar yield of BHET was calculated based on
the following relationship:
moles of BHET produced
Yield of BHET (mol%) = ×100
moles of PET units
To determine the functional groups of the monomer
BHET qualitatively, the glycolysis reaction products were
dissolved in boiling water, and the resulting hot solution
was immediately filtered. The BHET was considerably sol-
uble in boiling water, and the filter cake was washed sev-
eral times to extract almost all the BHET. The filtrate was
stored in a refrigerator at 4  C for 24 h. White BHET crys-
tals were formed, which were filtered and dried at 70  C
for 12 h. The dried crystals were then used for the FT-IR
analysis.
825
 Metal-Oxide-Doped Silica Nanoparticles for the Catalytic Glycolysis of Polyethylene Terephthalate
2.4. Characterization
The catalyst particles were characterized via FE-SEM
(S-4800, Hitachi), FE-TEM (JEM-2100F, Jeol Ltd.), and
XRD [D/Max-RB (12 KW), Rigaku], and with a BET sur-
face area analyzer (Tristar-3000, Micromeritics).
The monomer BHET was analyzed via HPLC (PU-980,
UV-975, Jasco) and FT-IR (IFS 66/S and Hyperion 3000,
Bruker Optiks). For the HPLC, a reverse-phase Zorbax-C8
column and an ultraviolet (UV) detector that was set at
254 nm were used. A 50:50 (v/v) THF/H2 O solution was
used for the mobile phase at a flow rate of 1.0 ml/min.
3. RESULTS AND DISCUSSION
3.1. Catalyst Morphology and Characterization
Metal oxide is an important catalyst in a wide range of
applications. However, pure metal oxides easily agglom-
erate during their synthesis, which reduces their active
surface area for the catalysis. For better dispersion with
less agglomeration, manganese and zinc oxides were
doped separately on two different forms of support: SNPs
(150±10 nm) and SMPs (1–20
m), with the use of ultra-
sound irradiation. The principle of sonochemical reactions
is based on cavity explosions that produce local high tem-
perature and pressure.Delivered
This localbyhigh
Publishing Technology
temperature and
IP: 218.23.112.86 On: Sun,products.
pressure assist the growth of the metal oxide on the surface
Copyright: American Scientific
RESEARCH ARTICLE
(a) (b)
(c) (d)
(e) (f)
Fig. 1. Morphologies of the nanocomposites and the microcomposites:
SEM images of (a) Mn3 O4 /SNPs; (b) ZnO/SNPs; (c) Mn3 O4 /SMPs; and
(d) ZnO/SMPs, and TEM images of (e) Mn3 O4 /SNPs and (f) ZnO/SNPs.
826
Imran et al.
of the support material. The morphologies of the nanocom-
posites (Mn3 O4 /SNPs and ZnO/SNPs) and microcompos-
ites (Mn3 O4 /SMPs and ZnO/SMPs) are shown in Figure 1.
The SEM (Figs. 1(a and b)) and TEM (Figs. 1(e and f))
images of the nanocomposites showed uniform disper-
sion of the metal oxides around the silica nanoparticles.
In the case of the microcomposites, their SEM images
(Figs. 1(c and d)) showed aggregates of the metal oxide
nanoparticles around the micro-sized silica. This difference
in the metal oxide distribution was due to the size and shape
of the support material, as well as the interaction between
the active metal oxide and the support material.
The SNPs were amorphous and the SMPs were polycrys-
talline, as shown by the XRD patterns in Figure 2. Crys-
talline ZnO was observed on the polycrystalline SMPs and
the amorphous SNPs, as shown in Figures 2(a) and (b),
respectively. Crystalline Mn3 O4 was observed on the poly-
crystalline SMPs, and the crystalline Mn3 O4 had no dif-
feraction peaks on the amorphous SNPs, as shown in
Figures 2(c) and (d), respectively. These show that the
sonochemically impregnated Mn3 O4 was homogenously
dispersed on the amorphous SNP support and exhibited
an amorphous state. Sonochemical reactions could produce
amorphous materials, depending on the temperature in the
ring region of the collapsing bubbles.15 After the collapse
of the bubbles, the high cooling rate of several hundred K/s
hinders the organization
to: Adelaide Theological andLibrary
crystallization of the reaction
24 Jan 2016 07:14:01
For catalytic
applications, high-surface-area materials,
Publishers
whether polycrystalline or amorphous, are generally
: Zincite-ZnO Intensity (a.u.)
: Hausmannite-Mn3O4 : ZnO
(a) (b)
20 30 40 50 60 70 80
2 Theta
Intensity (a.u.)
Intensity (a.u.)
(d)
(c)
20 30 40 50 60 70 80
2 Theta
20 30 40 50 60 70 80
2 Theta
Fig. 2. XRD patterns of the synthesized catalysts: (a) ZnO/SMPs; (b)
ZnO/SNPs; (c) Mn3 O4 /SMPs; and (d) Mn3 O4 /SNPs. (All the diffraction
peaks in (a) and (c), other than the ZnO and Mn3 O4 peaks, corresponded
to that of the polycrystalline silica).
J. Nanosci. Nanotechnol. 11, 824–828, 2011
Imran et al. Metal-Oxide-Doped Silica Nanoparticles for the Catalytic Glycolysis of Polyethylene Terephthalate

needed. The use of supported oxides, which consist of 100

Mn3O4/SNPs
active metal oxides dispersed on an inert support with a 90 ZnO/SNPs
high surface area, is one of the ways of obtaining high- Mn3O4/SMPs
80 ZnO/SMPs
surface-area catalysts. The surface area, pore volume, and No catalyst

Monomer (BHET) yield

average pore diameter of the supported catalysts that were 70
measured using the BET method are shown in Table I. High
60
surface areas were obtained in the case of the nanocompos-
ites, because of the small support size and the high pore 50
volume. 40

30
3.2. Glycolytic Depolymerization of PET
20
PET glycolysis involves the incorporation of diols into
10
PET chains to produce BHET. A mechanism with three
steps has been proposed: (i) glycol diffusion into the 0
polymer, (ii) polymer swelling, which increases the dif- 40 50 60 70 80
Reaction time (min)
fusion rate, and (iii) the reaction of glycol with an ester
bond in the PET chain.16 Polymer degrades to oligomers, Fig. 3. Effect of the type of catalyst on the BHET yield at 300  C and
dimers, and ultimately, to monomer BHET (PET + EG → 1.1 MPa.
oligomers → dimer ↔ BHET).
A mixture of 0.3 g of PET, 1.1 g of EG (EG/PET: nanocomposite.17 18 PET glycolysis is a heterogeneous
11/1, mole ratio), and 0.003 g of a catalyst (catalyst/PET = reaction, and the interface between the solid catalyst and
1 wt%) were used for the glycolysis reaction (40–80 min) the liquid reaction mixture controls the reaction. The large
 number of active sites and the high surface area of the
at 300 C and 1.1 MPa. A relatively high pressure of
1.1 MPa was needed to keep the EG in the liquid phase. catalyst could weaken the polymer network and help in
The effect of the type of catalyst on the BHET yield is the depolymerization of PET in a short reaction time.7
shown in Figure 3. The Delivered
BHET yieldbysignificantly Technology Figure
Publishingincreased 3 illustrates
to: Adelaide the orderLibrary
Theological of occurrence of the catalytic
IP: 218.23.112.86 On:
with time in the case of the catalytic glycolysis, contrary Sun,activity
24 Jan in terms
2016 of
07:14:01the monomer yield (Mn3 O4 /SNPs >
Copyright: American Scientific >
Publishers >

RESEARCH ARTICLE
to the slow increase without a catalyst. This shows that the ZnO/SNPs Mn 3 O 4 /SMPs ZnO/SMPs), which shows
glycolysis without a catalyst is a slow process under these reasonable agreement with the order of the catalyst sur-
operating conditions, and that a suitable catalyst is indis- face area and the pore volume in Table I. It is thus con-
pensable to the enhancement of the reaction rate. Figure 3 cluded that the Mn 3 O 4 /SNPs catalyst could provide the
shows that the BHET yields are higher with nanocom- most active surfaces for catalysis among the four pre-
posites than with microcomposites. This is because of the pared catalysts, which enhances the reaction rate and max-
uniform dispersion of the active metal oxide on the SNP imizes the monomer yield. From these observations, it is
support with a high surface area, as shown in the BET
surface areas in Table I. The BHET yield increased with
time and reached the highest level (>90%) after an 80-min
Aldrich BHET
reaction time with the Mn3 O4 /SNPs catalyst, during which
an equilibrium was established between the monomer and
its dimer. A longer reaction time may increase the dimer
% Transmittance

amount at the expense of the BHET yield.8 The high Our BHET
monomer yield in the short reaction time can be explained
by the amorphous structure, high surface area, and high
pore volume of the catalyst, as well as by the existence of
(b) (d)
numerous active sites on the surface of the Mn3 O4 /SNPs
(a)
(c) (d)
Table I. BET surface area, pore volume, and average pore diameter of (a) (b)
the supported catalysts. O
HOCH2CH2O C C OCH2CH2OH (a)
BET surface Pore volume Average pore (c)
Catalyst area (m2 /g) (cm3 /g) diameter (nm) 4000 3500 3000 2500 2000 1500 1000 500

Mn3 O4 /SNPs 45.09 0.214 18.9 Wavenumber (cm–1)

ZnO/SNPs 22.44 0.154 30.2
Fig. 4. Comparison of the FT-IR spectra of the standard BHET and
Mn3 O4 /SMPs 3.38 0.020 24.4
the BHET that was produced in this study with Mn3 O4 /SNPs as the
ZnO/SMPs 2.49 0.014 21.4
glycolysis catalyst.

J. Nanosci. Nanotechnol. 11, 824–828, 2011 827

 Metal-Oxide-Doped Silica Nanoparticles for the Catalytic Glycolysis of Polyethylene Terephthalate
concluded that the SNPs support is more beneficial to the
preparation of active metal oxide nanocomposites than the
SMPs support.
The FT-IR spectra of the BHET crystals and the stan-
dard BHET are shown in Figure 4. The agreement of
the FT-IR spectra proved that the monomer produced by
the nanocomposite (Mn3 O4 /SNPs) as the glycolysis cat-
alyst was almost pure BHET. The spectrum showed an
–OH band at 3,424 and 1,128 cm−1 , an aromatic C–H at
1,456–1,502 cm−1 , C O at 1,712 cm−1 , and alkyl C–H
at 2,879 and 2,964 cm−1 .
4. CONCLUSION
The glycolysis of PET with anhydrous EG was carried
out at 300  C and 1.1 MPa in the presence of sono-
chemically produced supported metal oxides as glycoly-
sis catalysts. The nanocomposites performed better as as
catalysts than the microcomposites. Among the nanocom-
posites, the highest monomer yield >90%) was obtained
with Mn3 O4 /SNPs. This high yield may be attributed to
the high surface area, amorphous and porous structure, and
existence of numerous active sites on the surface of the
nanocomposite catalyst, which weakened the polymer net-
work and helped depolymerize the PET in a short time.
The monomer BHET was analyzed via HPLC and FT-IR,
and the FT-IR spectrumDelivered by Publishing
of the produced BHET wasTechnology
com-
IP: 218.23.112.86
pared with that of the standard BHET, which verifiedOn:
theSun,15.
Copyright: American Scientific
RESEARCH ARTICLE
high level of purity of the produced BHET.
Acknowledgments: The authors acknowledge the
financial support of the Resource Recycling R&D Center
sponsored by the 21C Frontier R&D Program and of the
828
Imran et al.
Center for Ultramicrochemical Process Systems sponsored
by KOSEF.
References and Notes
1. L. Lundquist, Y. Leterrier, P. Sunderland, and J. A. E. Manson, Life
Cycle Engineering of Plastics: Technology, Economy, and the Envi-
ronment, Elsevier Science Ltd., UK (2000), pp. 39–40.
2. D. S. Achilias and G. P. Karayannidis, Water, Air, & Soil Pollution
4, 385 (2004).
3. J. Scheirs and T. E. Long, Modern Polyesters: Chemistry and Tech-
nology of Polyesters and Copolyesters, John Wiley & Sons Ltd.,
England (2003), p. 579.
4. S. Baliga and W. T. Wong, J. Polym. Sci., Part A: Polym. Chem.
27, 2071 (1989).
5. S. R. Shukla and K. S. Kulkarni, J. Appl. Polym. Sci. 85, 1765
(2002).
6. S. R. Shukla and A. J. Harad, J. Appl. Polym. Sci. 97, 513 (2005).
7. S. R. Shukla, V. Palekar, and N. Pingale, J. Appl. Polym. Sci.
110, 501 (2008).
8. H. Wang, Y. Liu, Z. Li, X. Zhang, S. Zhang, and Y. Zhang,
Eur. Polym. J. 45, 1535 (2009).
9. T. Sanders, P. Papas, and G. Veser, Chem. Eng. J. 142, 122 (2008).
10. J. J. Bong, K. Y. Jung, K. Wooyoung, K. Pil, and Y. Jongheop,
J. Nanosci. Nanotechnol. 8, 5130 (2008).
11. R. Ji, X. Lu, and J. Zhang, J. Nanosci. Nanotechnol. 9, 5134
(2009).
12. U. Heiz and U. Landman, Nanocatalysis, Springer-Verlag, New York
(2007), p. 245.
13. W. R. Moser, Advanced Catalysts and Nanostructured Materials:
Modern Synthetic Methods, Academic Press, USA (1996), p. 202.
to: Adelaide Theological Library
14. W. Stober and A. Fink, J. Colloid Interface Sci. 26, 62 (1968).
24S.Jan 2016 07:14:01
D. Jackson and S. J. Hargreaves, Metal Oxide Catalysis, Wiley-
Publishers
VCH, Germany (2009), p. 631.
16. J. Aguado and D.P. Serrano, Feedstock Recycling of Plastic Wastes,
The Royal Society of Chemistry, UK (1999), p. 33.
17. H. Einaga and A. Ogata, J. Hazard. Mater. 164, 1236 (2009).
18. C. Reed, Y. K. Lee, and S. T. Oyama, J. Phys. Chem. B 110, 4207
(2006).
Received: 20 July 2009. Accepted: 8 December 2009.
J. Nanosci. Nanotechnol. 11, 824–828, 2011