You are on page 1of 12

Kinetics Modelling of 2Butene Metathesis Over Tungsten Oxide

Containing Mesoporous Silica Catalyst

Tazul I. Bhuiyan,1,2 Palani Arudra,1 Mohammad M. Hossain,1,2 Mohammad N. Akhtar,1 Abdullah M. Aitani,1
Raed H. Abudawoud3 and Sulaiman S. AlKhattaf1,2*
1. Center of Research Excellence in Petroleum Rening and Petrochemicals, King Fahd University of Petroleum & Minerals, Dhahran 31261,
Saudi Arabia
2. Department of Chemical Engineering, King Fahd University of Petroleum & Minerals, Dhahran 31261, Saudi Arabia
3. Research & Development Center, Saudi Aramco, Dhahran 31311, Saudi Arabia

This article reports the activity, reaction mechanism and reaction kinetics of 2butene metathesis over tungsten oxide containing mesoporous silica
(SBA15) catalyst. Physicochemical characterisation of the catalyst indicates that the dispersion and nature of tungsten oxide species on the catalyst
are mainly responsible for the metathesis activity. The ample availability of Brnsted acid sites created by tetrahedrally coordinated tungsten species
also enhances the metathesis reaction. The product analysis indicates that lower temperature favours the butene isomerisation, while higher
temperature is required for metathesis. The high temperature is also responsible for cracking reactions. A kinetics model is developed based on the
experimental observations and the possible reactions including isomerisation, metathesis and cracking. The model parameters are estimated by
tting the experimental data implemented in MATLAB. The estimated apparent activation energy of 2butene isomerisation reaction was found to
be the lowest (39.4 kJ/mol) among the competing reactions. The activation energy of cross metathesis of 2butene and 1butene, selfmetathesis of
1butene and 2butene cracking are 71.3, 176.9 and 73.1 kJ/mol, respectively. These results are consistent to the product selectivity and the proposed
reaction scheme, which suggests that the isomerisation of 2butene gives 1butene and it further reacts (cross metathesis) with 2butene to produce

Keywords: metathesis, 2butene, propylene, tungsten on mesoporous material, kinetic modelling

INTRODUCTION be used in metathesis, the heterogeneous catalysts are preferred

due to their higher thermal stability, easy handling and possibility

he demand for propylene is increasing for the downstream of reuse after regeneration.[7,8] In the open literature, supported
applications such as polypropylene, acrylonitrile, cumene, rhenium, molybdenum and tungsten oxides are considered as the
oxoalchohols, propylene oxide, acrylic acid and isopropy- most promising candidates as metathesis catalysts.[9,10] Sang[11]
lalcohol and polygas chemicals.[1] Current global propylene and Balcar[12] reported high activity and selectivity of alumina,
demand is 73 million tons which is expected to riseup to silicaalumina and mesoporous organised alumina supported
110 million tons by 2020. On the other hand, the projected global rhenium (Re) oxide catalysts under mild reaction conditions.
production of propylene by 2020 stands at about 100 million tons However, the activity of these catalyst are poor with low Re loading
per year. Therefore, one can foresee a signicant decit in the (<10 wt %).[13] The deactivations rate of Re catalysts are also
annual global propylene production in just after 10 years from now fast.[14] The other major issue with Re based catalyst is the high
if the current production remains similar. Conventionally, propyl- price of rhenium.[6] The molybdenum oxide based catalysts have
ene is produced mostly by naphtha steam cracking and uid received much attention because of their relatively low price, and
catalytic cracking (FCC) units.[2] These conventional processes comparable activity under mild reaction conditions.[1517] The
cannot push further to increase the propylene production given main limitation of molybdenum catalysis is that it requires a
these processes have the priorities to produce ethylene and comparatively high pure feed stream.[18] The tungsten based
transportation fuels.[3] Considering these facts, in recent years catalysts shows good activity and selectivity for a longer period of
several onpurpose propylene production technologies including time. They are also resilient to frequent poisoning.[19] In addition,
metathesis of butene have been investigated to meet the high continuous regeneration without negative effects on the catalyst
propylene demand.[2] structure makes the tungsten based catalysts more attractive for
The metathesis reaction is an important organic transformation commercial use.[19] Usually the tungsten oxide catalysts require
that transforms less desired olens to higher value products a comparatively higher temperature (2505008C) to achieve
through C C bonds rearrangement. For instance, in the presence

of a tungsten catalyst, metathesis of the C4 olens converts C4 into
more valuable propylene.[4,5] At present, the olens conversion
technology (OCT) process licensed by Lummus Technology is the *Author to whom correspondence may be addressed.
only commercial olen metathesis technology producing propyl- Email address:
ene for the petrochemical industry.[6] The key step in olen Can. J. Chem. Eng. 92:12711282, 2014
metathesis is the intermediacy of metal carbene active species and 2014 Canadian Society for Chemical Engineering
DOI 10.1002/cjce.21976
consequently the formation of unstable metallocyclobutane. Published online 25 February 2014 in Wiley Online Library
Although, both homogeneous and heterogeneous catalysts can (

appreciable activity. The catalytic activity of tungsten based reaction model is dened by observing the reaction rates at a
catalysts widely depend on the oxidation state of tungsten species, different set of experimental conditions. The mechanism of
the content of tungsten oxide, the properties of support and the metathesis reaction was derived with the catalyst investigat-
pretreatment condition.[2,5,19,20] ed along with the effect of temperature and spacetime.
Regarding the support materials, galumina has been widely Kinetic parameters are assessed using their physical signi-
used as catalyst supports because of their stability and reasonable cances and various statistical indicators.
surface area. On the other hand, periodic mesoporous silicas (i.e.
MCM41, SBA15) having large pores (520 nm) with hexagonal EXPERIMENTAL
structure are more attractive due to their large channel diameter
and its thermal as well as hydrothermal stability.[21,22] The high Materials
surface (6001200 m2 g1) of these materials also makes them Pluronic123 (EO20PO70EO20), tetraethylorthosilicate (TEOS,
interesting support material.[23,24] Direct tungsten incorporation 98%), sodium tungstate (NaWO4.2H2O), HCl were purchased
using a hydrothermal method has advantages over a conventional from SigmaAldrich, Chemie Gmbh Munich, Germany and used
impregnation method such as high tungsten species dispersibility without any further purication. 2butene with 99.9% purity (50%
and the presence of more active tungsten species into the cis2butene and 50% trans2butene) was purchased from Saudi
support.[25,26] Industrial Gas Company, Dammam, Saudi Arabia.
Considering the great promise of tungstenbased catalysts, the
present research is focussed on the highly active and selective Catalyst Synthesis
tungsten oxide containing mesoporous silica (WO3SBA15) as Hexagonally ordered mesoporous SBA15 was synthesised using
the catalyst for metathesis of 2butene to produce propylene. There triblock copolymer P123 (EO20 PO70EO20) as the structure
are two important aspects of this present research, which have not directing agent and tetraethylorthosilicate (TEOS) as the silica
been reported in the open literature: precursor. Synthesis of SBA15 was performed according to the
(i) Metathesis of 2butene using only 2butene as a feed: the procedure reported in literature.[25,26] In a typical batch, 6 g of
advantage of the present approach is the use of much cheaper Pluronic 123 was dissolved to 210 mL of 2 M HCl and stirred at 408C
butene ($775/ton) as a raw material than the conventional to achieve a transparent solution. After 34 h, 12.8 g TEOS was
metathesis of 2butene with higher value ethylene ($1200/ quickly added into the solution. Then, a calculated amount of
ton) to produce propylene ($1375/ton).[27,28] There are sodium tungstate solution was added to the solution and stirred at
several studies describing the production of propylene by 1 958C for 3 days under hydrothermal conditions. The resultant
butene metathesis.[26,29,30] However, 1butene is expensive solution was stirred at 958C for 3 days. The solid products obtained
as compared to 2butene because 1butene has a huge were recovered by ltration and dried at 808C overnight. The
demand as a monomer for the production of liner low density template was removed by calcination in air at 5508C for 5 h.The
polyethylene (LLDPE) and polybutene.[31] To the best of catalysts obtained in this way were identied as WO3SBA15(x),
our knowledge, there is no communication dealing with where x represented the molar ratio of silica to tungsten (Si/W).
the application of tungsten oxide containing mesoporous The molar ratio of the gel composition is SiO2: 0.033 WO3: 0.0167
silica (WO3SBA15) as a catalyst for 2butene metathesis P123: 5.82 HCl: 190 H20.
reaction. In this study, we also investigate the reaction
mechanism of 2butene metathesis considering a wide range Catalyst Characterisation
of temperature. The chemical composition of synthesised tungsten containing SBA
(ii) Investigate the kinetics of 2butene metathesis over tungsten 15 was determined by inductively coupled plasma atomic emission
oxide containing mesoporous silica (WO3SBA15): In the spectroscopy (ICPAES), using Ultima 2 (Horiba scientic)
open literature, there are only a very limited number of equipment.
contributions considered for the kinetics of propylene The prepared catalysts were characterised by the Xray powder
metathesis. For instance, Kapteijn et al.[32] and Spinicci diffraction (XRD) method, using a Rigaku Miniex II system using
and Tofanari[33] investigated the kinetics of propylene nickel ltered CuKa radiation (l 0.154 nm) at a scanning rate of
metathesis over rhenium oxide based catalyst support on 18/min in the 2u range of 1.3508. The Xray tube was operated at
alumina. The apparent activation energy for metathesis of 40 kV and 30 mA.
propylene was 46.8 kJ/mol.[33] Cheng and Lo[7] developed a The textural properties of the samples were measured by
density functional theory only considering ethylene and 2 nitrogen adsorption measurements at 77 K using Quantachrome
butene metathesis reaction over WO3 catalysts. Li et al.[34] AUTOSORB 1C adsorption analyser. Before adsorption measure-
investigated the inuence of Mo oxidation state for metathe- ments, the solid samples (ca. 0.1 g) were degassed at 2208C under
sis of ethylene and 2butene over Mo/HBeta catalyst by vacuum nitrogen ow for 3 h. The nitrogen adsorption isotherms of
developing a DFT theory. Surprisingly, there is no work calcined materials were measured at liquid nitrogen temperature
reported in the open literature studying the kinetics (77 K). The surface areas and pore sizes were calculated by the
modelling of 2butene metathesis to produce propylene. On BrunauerEmmettTeller (BET) method and the BarrettJoyner
the other hand, the kinetics study of this reaction is very Halenda (BJH) method, respectively.
critical in determining the reactor/process design and Raman spectra were recorded at room temperature on Horiba
performance analysis. Keeping the above factors in mind, Jobin Yvon in situ Raman, iHR 320 with synapse CCD detector(-
the present work is focused on the development of kinetic excitation line, 532.5 nm of an Ar ion laser; source power, 40
model(s) for the metathesis of 2butene over tungsten based 60 mW; resolution, 4 cm1).
SBA15 catalyst. This is an attempt starting from an extensive UV vis DR spectra were collected on a JASCO V670
set of kinetic experiments using a laboratory scale xed bed spectrophotometer at 200 nm/min scan rate using BaSO4 as
reactor. The ratecontrolling step of the heterogeneous reference.

The acidity of the synthesised material was analysed by pyridine Conversion of 2butene
adsorption followed by FTIR study. The catalyst sample was nely
ground and pressed into a selfsupporting wafer. The wafers were moles of 2butene convertedW2C
4 =4
X 2C4  100% 1
calcined under a vacuum at 4008C for 1 h, followed by exposure to moles 2butene fed
pyridine vapour. The adsorbed pyridine was desorbed at various
temperatures (1504508C). The thin wafer was placed in the FTIR
cell and the spectrum was recorded on a Nicolet FTIR spectrometer Yield
(Magna 500 model). The difference between the spectra of pyridine
adsorbed on the samples and that of the reference was obtained by X2C4  WC n =n
Yield; YC
n  100%
subtraction. The amount of Brnsted acid sites (B) and Lewis acid WC2 =2 WC

3 =3 WC4 =4
sites (L) were calculated using the following equation as reported
WC5 =5 WC6 =6
in literature[35]:
C pyridine on B sites 1:88 IAB R =W;
2 Selectivity

Yield; YC
C pyridine on L sites 1:42 IAL R2 =W SC
n  100% 3
Conversion of 2butene; X 2C4
where C is the concentration (mmol/g catalyst); IA(B, L) is the
integrated absorbance of B or L band (cm1); R is the radius of In the formula, YC
n and SCn is the yield and selectivity of
catalyst disk (cm) and W is the weight of disk (mg). alkene with carbon number n, and WC n is the mass percent of
alkene with different carbon number n.
Kinetics Experiments
For this kinetics investigation, the metathesis reaction of 2butene RESULTS AND DISCUSSION
(1:1 mixture of trans2butene and cis2butene) was performed in
a xedbed tubular reactor (stainless steel tube grade 316 material, Physicochemical Properties of the Catalyst
0.312 inch ID  0.562 inch OD  8 inch length). A schematic The Xray diffraction patterns of the calcined SBA15, WO3SBA15
diagram of the reactor is shown in Figure 1. In a typical experiment, (30) samples are shown in Figure 2. Two diffraction peaks in
the reactor was charged with 2 mL of catalyst previously sieved to a the region of 2u 1.528, which can be indexed to the [1 1 0] and
particle size of 0.51.0 mm diameter. The catalyst sample was rst [2 0 0] diffractions characteristic of the p6mm hexagonal sym-
activated in a nitrogen stream (99.9% purity) at 5508C for 1 h. The metry, were observed for the SBA15 samples, which is also
metathesis reaction was carried out at different temperatures, reported by Herrera et al.[37] In both cases, 100 reections were not
ranging from 3505508C, GHSV (gas hourly space velocity) of 900 observed due to the limitation of the XRD instrument. There is no
2700 h1 and atmospheric pressure. The quantitative analysis of signicant difference in the XRD patterns for the SBA15 samples
the reaction products was carried out on line using Varian GC with showing the retention of hexagonal symmetry after tungsten
FID (Varian 450GC), equipped with a CPAl2O3/Na2SO4 column incorporation in the framework. There were small peaks assigned
(50 m length  0.32 mm ID  df 5 mm). To check the reproduc- to the crystalline phase of WO3 as observed for WO3SBA15(30)
ibility of the experimental results, a number of runs were sample at high angle XRD. Table 1 summarises the textural
performed and typical errors were in the range of 1%. properties of the supports and catalysts calculated from nitrogen
The GC analysed data were further processed to calculate the adsorption studies.
conversion and selectivity using the following equations according Raman spectra of the support SBA15 and catalyst WO3SBA15
to Liu et al.[36]: (30) are shown in Figure 3. For the catalyst WO3SBA15(30), three
peaks were observed at 267, 712 and 807 cm1. These crystalline
WO3 peaks were assigned to the symmetric stretching mode of
W--O, bending mode of W--O and the deformation mode of
W--O--W, respectively.[29] The band at 970 cm1 was associated to
the symmetric stretching mode of the terminal W O bond of the

tetrahedrally coordinated tungsten oxide species, which suggested
as the active site of the WO3SBA15(30) catalyst.[26,38] It is obvious
that pure SBA15 do not have any peaks in these regions.
Diffuse reectance spectra of tungsten supported samples in the
region of 200600 nm are shown in Figure 4. The spectrum of
sodium tungstate presents a spinal structure at 230 nm which
mainly implies the presence of ligand to metal charge transfer
involving isolated transition metal sites and considered direct proof
of the framework tungsten oxide species in tetrahedral co
ordination.[39] For pure SBA15, there is no evidence of the band
(not in the gure). The sample tungsten incorporated into the
framework, WO3SBA15(30), shows a similar band at 230 nm can
be attributed to the presence of isolated [WO4]2 tetrahedral
species. The second band at 290 nm may be assigned to isolated
tungsten species or lowcondensed oligomeric tungsten oxide
Figure 1. Schematic diagram of the xed bed tubular reactor. species. For the sample bulk WO3, a broad band at 350400 nm is

Figure 2. XRD of support SBA15 and WO3SBA15(30) catalyst at (A) low angle and (B) wide angle.

observed, implying that mainly extraframework bulk WO3 reaction temperature during the metathesis of 2butene over WO3
particles or clusters in the sample.[5,26] SBA15(30) catalyst.
Pyridine adsorption measured by IR spectroscopy was carried
out to evaluate the strength and types of acid sites present in the Effects of Temperature
tungsten impregnated and framework sample. Figure 5 shows In order to study the inuence of reaction temperature the
the FTIR spectra of support and the catalyst recorded after the experiments were conducted at different temperatures which
adsorption of pyridine and subsequent evacuation at 1508C and varied between 3508C and 5508C with a constant GHSV (900 h1)
2508C. The PyFTIR spectrum recorded after adsorption of at atmospheric pressure. Figure 6 plots the 2butene conversion
pyridine at room temperature on the pure SBA15 shows two and selectivity of different products as a different reaction
major peaks at 1595 and 1445 cm1, ascribed to pyridine temperature. It is clear from the gure that the temperature has
coordinately bonded to weak surface Lewis acid sites. These peaks signicant effects both on 2butene conversion and product
disappeared after out gassing at 3508C. In the case of the WO3SBA selectivity. This observation indicates that several reactions are
15(30) sample, two additional peaks at 1545 and 1635 cm1 were involved during this wide range temperature including isomer-
also observed due to protonated pyridine bonded to the surface of isation, primary cross metathesis, selfmetathesis, secondary cross
Brnsted acid sites, indicating the presence of tungstenincorpo- metathesis and cracking. As expected, 2butene conversion was
rated species.[38,40] The peak at 1485 cm1 has been identied as increasing with the increase of the reaction temperature. It is
the combination of both Lewis and Brsted acid sites.[35] The instructive to concentrate rst on the temperature range below
intensities of all of these bands were decreased after out gassing at 4008C where the 2butene conversion was low (35.8%). In this
2508C. It can be concluded that the presence of Brnsted acid sites temperature range 1butene is the major product while the
in the tungsten containing SBA15 samples are related to the formation of ethylene, propylene, 2pentene and 3hexene are
presence of tetrahedrally coordinated tungsten oxide species. minimal. When the reaction temperature was raised from 4008C to
4508C, 2butene conversion was drastically increased from 35.8 to
Metathesis of 2Butene Over WO3SBA15(30) Catalyst
As mentioned in the experimental section, the metatheses of 2
butene experiments were conducted in a laboratory scale xed bed
reactor. In order to identify the possible products, the experiments
were conducted in a wide range of experimental conditions (T:
3505508C; space time: 824 s). The product analysis results
during the metathesis of 2butene over WO3SBA15(30) catalyst
are presented in Table 2. One can see that the main products of 2
butene metathesis are propylene, ethylene, pentene, hexene and 1
butene. However, remarkable differences observed both in 2
butene conversion and product yields at different temperature
levels. Therefore, it is very important to discuss the effects of

Table 1. Physicochemical properties of synthesised catalysts

Surface Pore Brnsted Lewis

% of WO3 area diameter acid site acid site
Samples (ICPAES) (m2/g) (nm) (1508C) (1508C)

SBA15 833 4.4 0.0265 0.3818

WO3SBA15(30) 9.19 501 6.9 0.1086 0.3521
Lewis and Brnsted acid sites are calculated by FTIR pyridine sorption
(only at 1508C). Figure 3. Raman spectra of support SBA and catalyst WO3SBA15(30).

Figure 6. 2Butene conversion and selectivity of product over WO3SBA15
(30) catalyst from 3508C to 5508C temperature at 24 s spacetime.

1butene selfmetathesis:
Figure 4. UVvis diffuse reectance spectra of Na2WO4.2H2O, WO3SBA15
(30) and bulk WO3. 1butene $ ethylene 3hexene starts at  4508C

According to the above reactions, the molar ratio of propylene

72.4%, 1butene selectivity also dropped from 58 to 13.5 mol% and and pentene should be 1:1. However, the product analysis shows
the selectivity of the propylene and 2pentene increased to 41.3 and that propylene selectivity was 15% higher than that of 2pentene at
25.9 mol%, respectively. At 4508C the other products such as 4508C. This shows the possibility of some secondary metathesis
ethylene and 3hexene started to appear. One can speculate the reactions. In the temperature range between 4508C and 5008C both
following reactions steps based on the observed shift of conversion the 2butene conversion and propylene and ethylene selectivity
and product selectivity due to the change in the reaction proportionately increased while the selectivity of pentene and
temperature from 4008C to 4508C: hexene decreased with the increase of the reaction temperature
2butene isomerisation: (Figure 6). When the reaction temperature was further increased
from 5008C to 5508C, 2butene conversions and selectivity of
2butene $ 1butene starts at  3508C propylene and ethylene were increased, but the selectivity of 2
pentene and 3hexene started to decrease. This is a clear indication
2 and 1butene cross metathesis: of 2pentene and 3hexene cracking, which are favourable at a
higher temperature in the presence of an acidic catalyst. Due to the
2butene 1butene $ propylene 2pentene starts at  4508C cracking of 2butene more propylene and ethylene were produced.
In order to determine the effects of cracking, the C3 /C5 and
C2/C6 ratios are plotted at a different temperature (Figure 7). It
was noticed that, at 4508C C2 
  ratio was 1.02, which

increased to 1.69 and 2.74 at 5008C and 5508C, respectively. In the
same way, C3 
 ratio was 1.62 at 4508C, which increased to
1.89 and 2.36 at 5008C and 5508C. These increments of ratios with
temperature are the clear evidence of 2butene and 2pentene
cracking for the formation of ethylene and propylene at a higher

Figure 5. FTIR spectra of pyridine adsorbed on the sample recorded at

different temperature (a) SBA15 at 1508C, (b) SBA15 at 2508C, (c) WO3
SBA15(30) at 1508C, (d) WO3SBA15(30) at 2508C. B, Bronsted acidity; L, Figure 7. 2Butene conversion and selectivity of product over WO3SBA15
Lewis acidity. (30) catalyst from 3508C to 5508C temperature at 24 s spacetime.

Table 2. Product distribution (mol %) for metathesis reaction of 2butene at 3505508C and 8 to 24 s spacetime over WO3SBA15(30) catalyst

Temp (8C) C. time (s) Conv. 2C4



550 24 81.2 18.8 8.2 36.4 14.7 11.2 4.2

18 78.8 21.2 9.2 36.2 15.3 10.2 3.5
14 74.3 25.7 10.0 34.0 16.7 7.6 3.9
12 66.2 31.8 9.0 31.9 16.7 5.1 3.5
10 61.6 38.4 8.4 29.7 16.2 4.0 3.4
9 56.1 43.9 7.3 26.6 16.4 2.9 3.0
8.5 51.8 48.2 6.9 24.8 15.0 2.4 2.7
8.25 49.1 50.9 6.2 23.6 14.7 2.2 2.5
8 41.6 58.4 5.1 20.1 13.0 1.5 2.0
500 24 80.6 19.38 7.77 33.83 17.83 9.22 5.48
18 76.6 23.38 9.24 33.21 18.35 7.51 5.18
14 72.4 27.61 10.02 31.37 18.36 5.77 4.67
12 64.0 35.96 8.30 30.02 18.00 4.03 3.69
10 57.8 42.16 7.77 26.53 17.50 2.75 3.29
9 44.4 55.60 7.65 19.23 14.15 1.38 1.99
8.5 42.2 57.85 7.51 18.80 13.99 1.18 1.90
8.25 38.8 61.22 7.10 17.23 12.95 1.08 1.04
8 30.6 69.39 4.94 14.16 10.62 0.00 0.88
450 24 72.7 27.29 9.62 29.97 18.77 5.15 5.00
18 65.0 34.99 12.45 26.14 16.98 3.79 3.82
14 56.3 43.73 14.67 21.50 15.15 2.43 2.52
12 53.2 46.79 11.19 21.91 15.84 1.97 2.31
10 42.2 57.83 13.36 14.83 12.04 0.95 0.99
9 29.8 70.22 16.29 5.93 4.56 0.00 0.00
8.5 26.6 73.37 13.08 9.20 7.35 0.00 0.00
8.25 26.5 73.50 10.13 9.17 7.19 0.00 0.00
8 18.7 81.34 11.20 4.11 3.35 0.00 0.00
400 24 35.4 64.60 20.90 6.87 4.98 0.39 0.20
18 32.1 67.92 21.95 5.11 3.74 0.00 0.00
14 26.6 73.39 23.97 1.99 0.65 0.00 0.00
12 24.0 76.05 22.84 1.12 0.00 0.00 0.00
10 21.3 78.72 21.28 0.00 0.00 0.00 0.00
9 17.9 82.07 17.93 0.00 0.00 0.00 0.00
8.5 16.4 83.60 16.90 0.96 0.00 0.00 0.00
8.25 15.5 84.52 15.48 0.00 0.00 0.00 0.00
8 12.0 88.03 11.97 0.00 0.00 0.00 0.00
350 24 27.5 72.51 20.48 3.26 2.03 0.00 0.00
18 24.7 75.34 21.23 1.72 0.63 0.00 0.00
14 22.6 77.45 21.73 0.82 0.00 0.00 0.00
12 20.0 80.03 19.97 0.00 0.00 0.00 0.00
10 17.7 82.30 17.70 0.00 0.00 0.00 0.00
9 15.6 84.36 15.64 0.00 0.00 0.00 0.00
8.5 14.9 85.12 14.88 0.00 0.00 0.00 0.00
8.25 13.1 86.92 13.08 0.00 0.00 0.00 0.00
8 9.5 90.48 9.52 0.00 0.00 0.00 0.00

temperature. In summary, the isomerisation of 2butene to 1 spacetime. The same trend was followed to other products. The
butene takes place at a relatively lower temperature (3508C) involvement of secondary metathesis and cracking of 2butene
while the metathesis and cracking reactions occur at a higher could be explained with the comparison of C3 
 ratios at

(>4008C) temperature. different temperature and different spacetime, which are pre-

sented in Figure 9A. The ratio of C3
 decreased more at 5508C

than 4508C with the decreasing of spacetime. From this trend, it is
Effect of SpaceTime assumed that spacetime inuenced cracking more than the
In order to study the effects of spacetime, the experiments were secondary metathesis reactions for the production of propylene.
conducted at various feed ow rates using the same amount of The ratio of C2/C6 at different spacetime and temperature is

catalyst. The spacetime were varied between 8 and 24 s. Figure 8 presented in Figure 9B. The ratio of C2 
/C6 decreased with the

shows the products yield as a function of space time at 5508C and decreasing of spacetime and there was no formation of ethylene
atmospheric pressure. With the increasing of the space time (8 and 3hexene at 4508C at 9 s spacetime. From this result, it is
24 s), 2butene yield was decreased from 58.4 to 18.8 mol%. conrmed that selfmetathesis of 1butene for the production of
However, propylene yield was increasing with the increase of ethylene and 3hexene is highly inuenced by the spacetime.

Figure 8. Product distribution from 8 to 24 s spacetime over WO3SBA15
(30) catalyst at 5508C.

Scheme 1. Possible reaction scheme of 2butene metathesis reaction along

Catalytic Stability with isomerisation and cracking reactions over WO3SBA15(30) catalyst.
The catalytic stability of WO3SBA15(30) catalyst was investigat-
ed at 5508C for 16 h of stream. The results are presented in
lumped as C5 
, while hexane isomers combined as C6

Figure 10. The conversion of 2butene, yield of propylene and
ethylene were initially 81.1%, 37.5 and 12.11 mol%, respectively. According to Scheme 1, the desired product, propylene is produced
After 16 h, the catalyst has almost the same metathesis activity via three alternative routes: (i) cross metathesis of 1butene and 2
with 2butene conversion 80.2%, propylene yield 35.7 mol% and butene, (ii) secondary metathesis of ethylene and 2butene and (iii)
ethylene yield 10.5 mol%, respectively. There was very little coke by cracking of 2butene. From the carbene metal complex
formation after 16 h of operation. Furthermore, the opportunity of formation of 2butene, it is predicted that 2butene does not
pore blocking is low due to the large pore diameter of WO3SBA15 undergo selfmetathesis.[7,34,41] Due to the low concentration of
(30) catalyst. ethylene (by product) the secondary metathesis as a source of
propylene is considered to be negligible. Although the contribu-
Reaction Scheme tions of each of the pathways widely depend on the characteristics
Based on the above discussion, during the metathesis of 2butene of the catalysts material, it is important to recognise that the 1
over WO3SBA15(30) catalyst one can consider Scheme 1 for the butene required in cross metathesis via the isomerisation of 2
possible reaction pathways. In this scheme all the pentene isomers butene. On the other hand, 1butene also undergoes selfmetathesis
to form ethylene and 3hexene. An excessive isomerisation of 2
butene will promote the 1butene selfmetathesis. Therefore, a
balance of 2butene isomerisation is crucial to maximise propylene
production by cross metathesis and to minimise side products.
Aromatic compounds are not considered in the reaction scheme
due to its negligible amount.

Mass Transfer Limitations

The overall reaction rate for metathesis of 2butene over WO3SBA
15(30) catalyst involves diffusion, adsorption and intrinsic
reaction steps. In the context of the present study, the possibilities
of both the external and internal diffusions were examined by
evaluating the mass transfer limiting criterions using the
experimental data. The possibility of external mass transfer
(bulk uid to catalyst surface) limitations of the reactants and
products is assessed using the Sherwood number with Froesslings
dimensionless correlations evaluated at the experimental con-

essling correlation
Sh 2 0:6 Sc0:33 Re0:66 Fro 4

kc 5
r 00 ABobs
C AB;b  C AB;s film DC film 6
Figure 9. Ratio of (A) propylene to 2pentene (B) ethylene to 3hexene over 1  103 T 1:75 1=M A 1=M B 1=2
WO3SBA15(30) catalyst from 8 to 24 s spacetime at 4508C and 5508C DAB 1=3 1=3
Patm SvA SvB 2

bulk 2butene concentration. The small values of DClm suggest
that in the context of the present study the external mass transfer
limitation is not signicant.
The inuence of intraparticle mass transfer (diffusion) is
examined by evaluating the WeiszPrater criterion.[42] According
to the WeiszPrater criterion the diffusion mass transfer can be
considered negligible if the following inequality satises:

r 00 ABobsrp R2p
C WP 1 9
Deff C AB;s

where, r 0 ABobs is the observed reaction rate (mol/kg s), rP is the

particle density (kg/m3), Rp is the particle radius (m), Deff is the
effective diffusivity (m2/s) and CAB,s is the concentration of gaseous
Figure 10. Results of 2butene metathesis reaction over WO3SBA15(30) reactant at the surface of solid particles (mol/m3). The effective
catalyst at 16 h operation. diffusivity (Deff) is estimated as[43]:
Deff 0:1DAB 10
where, Sh, Sc and Re are Sherwood, Schmidt and Reynolds
numbers, respectively. kc is the external mass transfer coefcient The initial concentration of 2butene is considered as the
(m/s), subscript A is using for 2butene and B is using for nitrogen, concentration of 2butene at the surface of the catalyst. This
DAB molecular diffusivity of 2butene in carrier gas nitrogen (m2/ ensures the evaluation of WeiszPrater criterion at conditions,
s), dp is the particle diameter (m),r 00 ABobs is the observed where the possibility of diffusion limitations is the highest. Table 4
reaction rate (mol/m2 s), CAB,b is the bulk concentration of gaseous reports the calculated WeiszPrater parameters and the other
reactant and CAB,s is the surface concentration of gaseous reactant variables used for this evaluation. It is clear for Table 4 that even in
(mol/m3). The diffusivity for 2butene in nitrogen (DAB) is the worst case scenarios the values of CWP was much smaller than 1
estimated using the Fuller, Schettler and Giddings correlation (one). Thus, the concentration gradient within the catalyst carrier
(Equation 7).[44] Frossling correlation is evaluated by considering pellet is expected to be negligible. Thus, it can be concluded that in
that the uid surrounding the particles is almost stagnant (highest the studied reaction conditions, the chemical reaction is the
possible gas lm resistances). The observed reaction rate (mol/ limiting resistance for all the runs developed in the xed bed
m2 s) in the xed bed reactor can be expressed as: reactor.

F A0 X A Kinetic Modelling
r 00 ABobs  8
Sex W
Model development
where, FA0 is the molar ow rate of reactant 2butene, XA is the The mathematical models representing the rates of chemical
conversion of 2butene and Sex is the external surface area per gram reactions are developed based on Scheme 1 which considers (i)
of sample and W is the weight of the catalyst. isomerisation of 2butene to 1butene (ii) cross metathesis of 2
The calculated DClm (using Equation 6) values are listed in butene and 1butene, (iii) selfmetathesis of 1butene and (iv)
Table 3. One can easily notice that at calculated values of DClm, at cracking of 2butene. As previously discussed, all these reactions
various reaction conditions, are very small when compared to the are greatly inuenced by temperature and spacetime. In formula-

Table 3. Parameter used to calculate the effects of external mass transfer limitations

Temp. (8C) DAB (m2/s) kc (m/s) Spacetime (t), s r 00 AB obs (mmol/m2 s) CAB,b (mol/m3) DCfilm (mmol/m3)

350 3.7  105 0.12 24 98.3 43.0 0.80

400 4.2  105 0.14 24 122.8 43.0 0.87
450 4.8  105 0.16 24 256.2 43.0 1.60
500 5.4  105 0.18 24 284.3 43.0 1.58
550 6.1  105 0.20 24 287.8 43.0 1.43

Table 4. Parameter used to calculate the effects of internal mass transfer limitations

Temp. (8C) Deff (m2/s) Spacetime (t), s r 00 AB obs(mmol/kg s) CAB,s (mol/m3) CWP
350 3.7  10 24 3.34 43.0 5.7  104
400 4.2  106 24 4.18 43.0 6.2  104
450 4.8  106 24 8.71 43.0 1.1  103
500 5.4  106 24 9.67 43.0 1.1  103
550 6.1  106 24 9.79 43.0 1.0  103

tion of the kinetic model the following assumptions have been (average of all the temperatures) is introduced to reduce parameter
considered: interaction.
i. Isomerisation, cross metathesis and selfmetathesis reactions For the thermodynamic consistency of the reaction, the
are reversible while 2butene cracking is irreversible. temperature dependent equilibrium constants were formulated
ii. The pentene isomers are considered as 2pentene and the using Gibbs energy of formation of pure species:
hexene isomers as 3hexene. Isomerisation of 2butene to 1butene
iii. All parafn, isobutene and aromatic products are negligible.   
iv. Contribution of secondary metathesis in propylene produc- 9600 1 1
k1e 21:45 exp   19
tion is insignicant. R T T0
Based on these assumptions and using power law model, the rate
Cross metathesis of 2butene and 1butene
of disappearance/formation of each of the species involved in 2
butene metathesis can be expressed as follows:   
1200 1 1
Disappearance of 2butene: k2e 196:5 exp   20
R T T0
dCA 1 1
k1 CA  CB  k2 CA CB  CC CD  2k4 C24 Selfmetathesis of 1butene
dt0 k 1e k2e
26100 1 1
k3e 181:8 exp   21
R T T0
Formation of 1butene:
dCB 1 1 Parameter estimation and kinetics evaluation
k 1 CA  C B  k 2 C A CB  CC C D
dt0 k1e k2e To estimate the parameters and evaluate the developed models, the
1 temperature dependent specic reaction rate constants (Equa-
 2k3 CB 2
k3e tion 18) and temperature dependent equilibrium constants
(Equations 1921) were rst combined with the kinetics model
Formation of propylene: equations (Equations 1116). The concentrations of various
species were expressed in terms of mass fraction according
  Equation 17. The resulting differential equations were then
dCC 1
k 2 C C
A B  CC D 2k 4 CA
C 2
13 numerically solved in conjunction with a least square tting of
dt0 k2e
the experimental 2butene metathesis data to be obtained from the
xed bed reactor. In solving the differential equation MATLAB
Formation of 2pentene: ODE 45 subroutine (RungeKuttaGill method) was employed
  while the parameters were estimated using LSQCURVEFIT
dCD 1 subroutine (Modied Marquad method).
k2 CA CB  CC CD 14
dt0 k 2e In order to ensure reliable model parameter estimation, the
catalytic 2butene metathesis experiments were conducted at a
Formation of ethylene: wide range of reaction temperatures (3505508C) and various
  space times (824 s). In this fashion a total of 270 data points were
dCE 1 obtained and used in parameter estimation. Therefore, eight
k 3 CB  CE F k 4 CA
C 2
dt 0 k3e kinetics parameters were determined with a degree of freedom of
262 (degree of freedom No. of experimental data points  No. of
Formation of 3hexene: model parameters). The sufcient numbers of experimental data
  points obtained/used in parameter estimation provides reliability
dCF 1 of the model parameter.
k 3 C2
 CE CF 16
dt0 B

The concentration of various species in Equation (11)

through (16) can be written in terms of mass fraction of (yi)
according to the following equation:

yi F TM
Ci 17
MWi q

The rate constants ki can be represented with the temperature

dependent form of the Arrhenius equation:
Ei 1 1
ki ki0 exp   18
R T T0

Figure 11. Comparison between experimental results and model

where, ki0 and Ei are the preexponential factor and activation predictions for 2butene conversion at the reaction temperature at 350 8C,
energy of the reaction i, respectively. T0, the centering temperature 4508C and 5508C.

(iii) Lower SSR (sum of the squares of the residuals).
(iv) Lower crosscorrelation coefcient (g).
(v) Smaller individual condence intervals for the model

Figures 11 and 12 show the comparison between the experi-

mental data and model predictions for the reaction scheme. We
observed in these plots that model predictions compared favour-
ably with the experimental data for the reactions at 3508C, 4508C
and 5508C. Some deviations due to high conversion are predicted at
5508C and 24 s space time. Several reactions are involved in the
reaction scheme to produce various products. During the long
range of space time the model predicted a higher conversion at a
Figure 12. Comparison between experimental results and model
higher temperature. Figure 13 shows the Arrhenius plots of various
predictions for propylene concentration at the reaction temperature at
specic reaction rate constants. The straight line with negative
3508C, 4508C and 5508C.
slopes of all the Arrhenius plots (as shown in Figure 13) of the rate
constants conrms the thermodynamic consistency of the present
The following model evaluation criteria are considered: Power Law model.
Table 5 lists the estimated kinetics parameters for different
(i) All the kinetics parameters (specic reaction rates and reaction paths displayed in Scheme 1. From Table 5, it is noticed
activation energies) should be consistent with physical that the estimated activation energy of 2butene isomerisation
principles. forming 1butene requires the least activation energy (39.4 kJ/
(ii) Coefcient of determination (R2). mol) as compared to the other competing reactions. The estimated

Figure 13. Arrhenius plots of different reactions rate constants obtained from power law model (A) isomerisation reaction, (B) cross metathesis of 1butene
an 2butene, (C) selfmetathesis of 1butene and (D) cracking of 2butene.

Table 5. Estimated values of the parameters, 95% condence intervals However, the probability of cross metathesis of 2butene and 1
and R2 butene is much higher than the selfmetathesis of 1butene.
4. The kinetics investigation shows that among the competing
Estimated 95% confidence reactions, isomerisation of 2butene to 1butene requires the
Parameter value interval R2
least amount of activation energy (39.4 kJ/mol) while the
k01  102 (m3/kg) 4.72 0.17 0.9727 activation energy for selfmetathesis of 1butene is the highest
E1 (kJ/mol) 39.4 1.92 (176.9 kJ/mol). The activation energy of cross metathesis of 2
k02  102 (m6/kg) 5.7 0.4 butene and 1butene, selfmetathesis of 1butene and 2butene
E2 (kJ/mol) 71.3 6.1 cracking are 71.3, 176.9 and 73.1 kJ/mol, respectively. These
k03  102 (m6/kg) 1.64 0.48 results are consistent to the proposed reaction mechanisms
E3 (kJ/mol) 176.9 36.8 and the product selectivity data, which suggests that the
k04  102 (m3/kg) 1.8 0.031 isomerisation of 2butene gives 1butene which further reacts
E4 (kJ/mol) 73.1 15.3 (cross metathesis) with 2butene to produce propylene.

activation energies for cross metathesis of 2butene and 1butene is
71.3 kJ/mol, cracking of 2butene is 73.1 kJ/mol. Unfortunately, in Ci concentration of species i in the fixed bed reactor (mmol/
the open literature the activation energies for butene metathesis cm3)
(either 1 and 2butene cross metathesis or 1butene self CI confidence limit
metathesis) are not available for comparison. The value of Ei apparent activation energy of the ith reaction (kJ/mol)
activation energy for selfmetathesis of 1butene found to the ki apparent rate constant for the ith reaction (1/s)
highest (176.5 kJ/mol). The high activation energy for 1butene ki0 preexponential factor for the ith reaction after reaction
selfmetathesis is consistent to the experimental low selectivity 3 parameterisation
hexene as shown in Figure 6 and Table 2. On the other hand, the MWi molecular weight of the species
low value of activation energy of 2butene isomerisation is also ri rate constant of the species
reasonable given that the product 1butene further reacts with 2 vo volumetric flow rate of the feed
butene (cross metathesis) to give propylene. Therefore, isomer- t the spacetime
isation of 2butene to 1butene is a prerequirement of cross yi mass fraction of species i in the reactor
metathesis. The high selectivity of propylene as seen in Figure 6 FTM total mass flow rate
and Table 2 conrms this explanation. One can argue that the Symbols
origin of propylene might be the cracking of 2butene. However, at
a low temperature (until 4008C) the selectivity of the other cracking A 2butene
products ethylene is negligible, while the selectivity of propylene is B 1butene
still high. This conrms the formation of propylene mainly via C propylene
cross metathesis of 2butene and 1butene. D 2pentene
E ethylene
F 3hexene
The present investigation is to develop a kinetic model on 2butene
metathesis reaction over WO3SBA15 catalyst which leads to the The authors would like to acknowledge the support provided by
following conclusions: Saudi Aramco for funding project #CRP02252. The support from
the Ministry of Higher Education, Saudi Arabia, in establishment of
1. The synthesised WO3SBA15(30) catalyst has tetrahedral the Center of Research Excellence in Petroleum Rening &
tungsten oxide species as conrmed by Raman spectroscopy Petrochemicals at King Fahd University of Petroleum & Minerals
and UVvis DRS. The tetrahedral tungsten oxide species act as (KFUPM) is appreciated. The support of KFUPM is also highly
the active site for the metathesis reaction. Tungsten oxide appreciated.
effectively incorporated into the SBA15 framework and
highly dispersed into the lattice of the support as revealed
by XRD analysis. The FTIR pyridine adsorption experiment REFERENCES
conrms the availability of both Lewis and Brnsted acid sites [1] H. J. Liu, L. Zhang, X. J. Li, S. J. Huang, S. L. Liu, W. J. Xin,
on the catalyst surface. S. J. Xie, L. Y. Xu, J. Nat. Gas Chem. 2009, 18, 331.
2. The product selectivity is signicantly inuenced by the [2] Q. F. Zhao, S. L. Chen, J. S. Gao, C. M. Xu, Transit. Met. Chem.
reaction temperature. Lower temperature favours the 2butene 2009, 34, 621.
isomerisation to 1butene. On the other hand, a higher [3] X. X. Zhu, X. J. Li, S. J. Xie, S. L. Liu, G. L. Xu, W. J. Xin, S. J.
temperature is required for metathesis. However, a high
Huang, L. Y. Xu, Catal. Surv. Asia 2009, 13, 1.
temperature is also responsible for undesired cracking
reactions. Selectivity of propylene and ethylene are increased [4] J. C. Mol, P. W. M. Leeuwen, Metathesis of Alkenes
with the increasing of temperature, while pentene and hexene Handbook of Heterogeneous Catalyst, Wiley Online Library,
selectivities are decreased after 4508C. 2008, 14, 3240.
3. The reaction mechanisms suggest that metathesis reactions [5] D. R. Hua, S. L. Chen, G. M. Yuan, Y. L. Wang, Q. F. Zhao,
are initiated by the Wcarbene complex formation. Under the X. L. Wang, B. Fu, Micropor. Mesopor. Mater. 2011, 143, 320.
studied reaction conditions both the cross metathesis of 2butene [6] P. Topka, H. Balcar, J. Rathousky, N. Zillkova, F. Verpoort, J.
and 1butene and selfmetathesis of 1butene are possible. Cejka, Micropor. Mesopor. Mater. 2006, 96, 44.

[7] Z. Cheng, C. S. Lo, ACS Catal. 2012, 2, 341. [37] J. E. Herrera, J. H. Kwak, J. Z. Hu, Y. Wang, C. H. F. Peden, J.
[8] T. I. Bhuiyan, P. Arudra, M. N. Akhtar, A. M. Aitani, R. H. Macht, E. Iglesia, J. Catal. 2006, 239, 200.
Abudawoud, M. A. AlYami, S. S. AlKhattaf, Appl. Catal. A [38] X. L. Yang, W. L. Dai, R. H. Gao, K. N. Fan, J. Catal. 2007, 249,
2013, 467, 224. 278.
[9] S. J. Huang, S. L. Liu, W. J. Xin, J. Bai, S. J. Xie, Q. X. Wang, [39] E. Briot, J. Y. Piquemal, M. Vennat, J. M. Bregeault, G.
L. Y. Xu, J. Mol. Catal. A 2005, 226, 61. Chottard, J. M. Manoli, J. Mater. Chem. 2000, 10, 953.
[10] H. Balcar, J. Cejka, Coord. Chem. Rev. 2013, 257, 3107. [40] C. Martin, P. Malet, G. Solana, V. Rives, J. Phys. Chem. B 1998,
[11] L. Sang, S. L. Chen, G. M. Yuan, M. Zheng, J. You, A. E. Chen, 102, 2759.
R. Li, L. J. Chen, J. Nat. Gas Chem. 2012, 21, 105. [41] US Patent 6777582 B2, 2004, R. J. Gartside, M. I. Greene, Q. J.
[12] H. Balcar, R. Hamtil, N. Zilkova, J. Cejka, Catal. Lett. 2004, Jones.
97, 25. [42] H. S. Fogler, Elements of Chemical Reaction Engineering, 4th
[13] A. Andreni, X. D. Xu, J. C. Mol, Appl. Catal. 1986, 27, 31. edition, Princeton Hall International, New Jersey 2006, p.
[14] A. Behr, U. Schller, K. Bauer, D. Maschmeyer, K. D. Wise, F. 839.
Nierlich, Appl. Catal. A 2009, 357, 34. [43] M. M. Hossain, H. I. de Lasa, AIChE J. 2007, 53, 1817.
[15] J. Handzlik, J. Ogonowski, J. Stoch, M. Mikolajczyk, P. [44] J. R. Welty, C. E. Wicks, R. E. Wilson, G. Rorrer,
Michorczyk, Appl. Catal. A 2006, 312, 213. Fundamentals of Momentum, Heat and Mass Transfer. 5th
[16] H. Balcar, D. Mishra, E. Marceau, X. Carrier, N. Zilkova, Z. edition, John Wiley & Sons, Inc, New York 2006, p. 412.
Bastl, Appl. Catal. A 2009, 359, 129.
[17] D. P. Debecker, M. Stoyanova, U. Rodemerck, A. Lonard,
B. L. Su, E. M. Gaigneaux, Catal. Today 2011, 169, 60. Manuscript received July 30, 2013; revised manuscript received
[18] US Patent Application 20120289617 A1. 2012. Y. Wang, C. October 7, 2013; accepted for publication October 8, 2013.
[19] A. Spamer, T. I. Dube, D. J. Moodley, C. van Schalkwyk, J. M.
Botha, Appl. Catal. A 2003, 255, 133.
[20] S. J. Huang, F. C. Chen, S. L. Liu, Q. J. Zhu, X. X. Zhu, W. J.
Xin, Z. C. Feng, C. Li, Q. X. Wang, L. Y. Xu, J. Mol. Catal. A
Chem. 2007, 267, 224.
[21] D. Y. Zhao, Q. E. Huo, J. L. Feng, B. F. Chmelk, G. D. Stucky, J.
Am. Chem. Soc. 1998, 120, 6024.
[22] D. Bek, H. Balcar, N. Zilkova, A. Zukal, M. Horacek, J. Cejka,
ACS Catal. 2011, 1, 709.
[23] A. Taguchi, F. Schuth, Micropor. Mesopor. Mater. 2005, 77, 1.
[24] R. M. MartinAranda, J. Cejka, Top. Catal. 2010, 53, 141.
[25] L. F. Chen, J. U. Hu, Y. D. Wang, K. Zhu, R. Richards, W. M.
Yang, Z. C. Liu, W. Xu, Mater. Lett. 2006, 60, 3059.
[26] J. C. Hu, Y. D. Wang, L. F. Chen, R. Richards, W. M.
Yang, Z. C. Liu, W. Xu, Micropor. Mesopor. Mater. 2006, 93,
[27] V. Y. Wan, Propylene production via metathesis of ethylene
and butenes, Process economics program, Review no. 2011
04, 2011.
[28] Technology economic program. Light olen seriesPropyl-
ene production via metathesis, 2011.
[29] Y. D. Wang, Q. L. Chen, W. Yang, Z. Xie, W. Xu, D. Huang,
Appl. Catal. A 2003, 250, 25.
[30] L. Harmse, C. van Schalkwyk, E. van Steen, Catal. Lett. 2010,
137, 123.
[31] K. G. Lee, J. K. Jeon, E. H. Hwang, Y. S. Ko, Y. K. Park, S. J.
Lee, J. H. Lee, J. Ind. Eng. Chem. 2007, 13, 1062.
[32] F. Kapteijn, H. L. G. Bredt, E. Homburg, J. C. Mol, Ind. Eng.
Chem. Prod. Res. Dev. 1981, 20, 457.
[33] R. Spinicci, A. Tofanari, Appl. Catal. 1985, 14, 261.
[34] X. Li, J. Guan, D. H. Zhou, G. H. Li, X. W. Han, W. P. Zhang,
X. H. Bao, Comp. Theor. Chem. 2009, 913, 167.
[35] C. A. Emeis, J. Catal. 1993, 141, 347.
[36] N. Liu, S. Ding, Y. I. Cui, N. H. Xue, L. I. Peng, X. F. Guo, W. P.
Ding, Chem. Eng. Res. Des. 2013, 91, 573.