Single Stage Oxychlorination of Ethanol To Ethylene Dichloride Using A Dual Catalyst Bed
Single Stage Oxychlorination of Ethanol To Ethylene Dichloride Using A Dual Catalyst Bed
Single Stage Oxychlorination of Ethanol To Ethylene Dichloride Using A Dual Catalyst Bed
a r t i c l e i n f o a b s t r a c t
Article history: The direct production of ethylene dichloride, EDC, from ethanol over a dual-catalyst bed consisting of a
Received 25 May 2010 zeolite pre-bed and a CuCl2 –NaCl/␥-Al2 O3 oxychlorination catalyst has been shown to give greater than
Received in revised form 31 July 2010 80% yield of EDC with ethyl chloride as the main by-product. The effects of temperature, HCl/air ratio
Accepted 18 August 2010
and space velocity were investigated. The most effective zeolite was ZSM-5. It is concluded that the main
Available online 26 August 2010
route to EDC was via ethyl chloride and its oxychlorination and disproportionation.
© 2010 Elsevier B.V. All rights reserved.
Keywords:
Ethylene dichloride
Ethanol
Oxychlorination
Zeolite
0926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.08.036
R. Al-Hajri, D. Chadwick / Applied Catalysis A: General 388 (2010) 96–101 97
Nomenclature
AcA acetaldehyde
DEE diethyl ether
EtCl ethyl chloride
EDC 1,2-dichloroethane
EtOH ethanol
GHSV gas hourly space velocity, h−1
TCA trichloroethane
VCM vinyl chloride monomer
X conversion
and structure, and which have been studied widely for the dehy-
dration of ethanol [11–14]. Mordenite has been shown to be highly (counter current configuration at 150 ◦ C) where it mixed with an
active for dehydration of ethanol to ethylene at temperatures as Ar flow. The ethanol/argon mixture was then fed into a mixer/pre-
low as 180 ◦ C [13]. It was expected, therefore, that ethanol would heater made of glass lined s.s. tubing where HCl and air were added
be dehydrated over the zeolite in the first bed producing ethylene, to adjust the reactor feed composition. For experiments with ethyl
and that the product ethylene would be oxychlorinated to EDC in chloride, ethyl chloride/N2 premixed gas was supplied to the feed
the presence of HCl and air over the CuCl2 catalyst in the second gas mixer/pre-heater via a mass flow controller. To avoid corro-
bed. However, we show that the main route to EDC is via formation sion and product reactions in the reactor exit lines, PFA tubing was
ethyl chloride and its subsequent oxychlorination and dispropor- used for all lines prior and after the reactor containing HCl. The
tionation. The effects of reaction temperature, feed composition, lines were trace heated to avoid any condensation. NaOH solution
and space velocity were studied. In addition, the performance of and Norit activated carbon were used as scrubber for any unre-
the dual-bed configuration is compared to a physical mixture. We acted HCl and halogenated products. The product stream from the
demonstrate that greater than 80% yield of EDC can be achieved reactor was analyzed continuously by online FID GC with a Car-
directly from ethanol when ZSM-5 is used in the pre-bed. bopack B column. Using a temperature program (isothermal at
45 ◦ C for 4 min, increase of temperature up to 220 ◦ C with a heat-
2. Experimental ing rate of 8 ◦ C/min and then hold at 220 ◦ C for 5 min) ethylene,
EtCl, ethanol and EDC were easily separated at retention times of
2.1. Catalyst preparation and pretreatment 1.9, 6.2, 6.8 and 15.1 min, respectively. The main by-products were
also readily identified. The response factors for chlorinated com-
Na-mordenite(13), Y-zeolite(60), ZSM-5(280) and ZSM-5(50) pounds were determined using a calibration mixture obtained from
were obtained from Zeolyst and checked by XRD. The number in Restek. Data was taken over 1 h time-on-stream at each condition.
parentheses is the SiO2 /Al2 O3 ratio. NH4 -mordenite was prepared There was no noticeable deactivation. The carbon balances were
by ion-exchange method with NH4 NO3 . The H-forms of the zeolites ±2% at all the conditions studied demonstrating that the produc-
were prepared by calcination at 550 ◦ C in air for 10 h ex situ imme- tion of carbon oxides was very small. Values of selectivity are given
diately before use. The oxychlorination catalyst for the second bed, in mol%.
14 wt%CuCl2 + 5 wt%NaCl on ␥-Al2 O3 , was prepared according to
the procedure described in US patent 2,442,285 (1948) [8]. The Cu 3. Results and discussion
content of the catalyst was determined by ICP. BET surface area was
determined to be 150 m2 /g. The catalyst and pre-bed was heated 3.1. Reactions over CuCl2 –NaCl/-Al2 O3
overnight in situ in the reactor at 300 ◦ C with argon flow. This treat-
ment was carried out in order to activate the zeolite pre-bed in situ 3.1.1. Catalyst characterisation
and also to remove water of hydration from the CuCl2 catalyst. The XRD of the CuCl2 –NaCl/␥-Al2 O3 catalyst as prepared and after
temperature was then decreased to the reaction temperature and in situ pretreatment at 300 ◦ C in Ar is given in Fig. 1. In the as
a flow of HCl and air was passed over the catalyst for 1 h before prepared catalyst, Fig. 1a, the diffraction peaks of CuCl2 ·2H2 O are
introducing the ethanol flow. clearly seen. From XRD the average particle size of the supported
CuCl2 ·2H2 O was estimated to be 15 nm. After the pretreatment no
2.2. Catalyst activities diffraction peaks from Cu species were detected. Redistribution of
Cu in oxychlorination catalysts is well known [15], and could be a
A 10 mm I.D. tubular flow reactor was used to study the gas contributory factor to the loss of diffraction peaks. TEM analysis of
phase reaction. The reactor was electrically heated and the temper- the spent catalysts, however, revealed a reasonably even covering
ature was controlled by a thermocouple placed in the centre of the of particles of about 15 nm diameter on the alumina support. From
catalyst bed. 0.7 g of zeolite and 0.7 g of CuCl2 catalyst (0.5–0.85 mm this we conclude that the CuCl2 ·2H2 O particles lose crystallinity
particle diameter) were each mixed with SiC at 4:1 dilution and when the water of hydration is removed by heating in Ar and may
placed as separate beds in the middle of the reactor supported by also partially oxidise on exposure to air [15], but the majority of the
quartz wool from both ends. The same dilution was used for stud- particles remain about 15 nm in diameter.
ies of single catalysts. Values of GHSV are based on the undiluted
catalyst volume (either zeolite or CuCl2 catalyst). Gas feeds (ethy- 3.1.2. Ethylene oxychlorination
lene, Ar, air, N2 , HCl/N2 ) to the reactor were controlled by mass flow The C2 H4 oxychlorination activity of catalyst was tested at
controllers. Liquid ethanol was pumped into an evaporator/mixer 250–300 ◦ C with a close to stoichiometric feed composition of 9,
98 R. Al-Hajri, D. Chadwick / Applied Catalysis A: General 388 (2010) 96–101
Table 2
EtCl reactiona with HCl/air over 14 wt%CuCl2 –5 wt%NaCl/␥-Al2 O3 .
3.1.4. Ethanol oxychlorination
The EDC yield from ethanol oxychlorination over CuCl2 –NaCl/␥-
T EtCl Product selectivity, % Al2 O3 was low at 250–300 ◦ C; the main product was EtCl with
◦
C X, % C2 H4 EDC TCA selectivities of over 80%, Table 3. This contrasts sharply with US
250 1.1 30 70 0.00
patent 2,442,285 (1948) which claimed a good yield of EDC from
275 9.0 13 87 0.00 ethanol in the same temperature range [8]. The activity for ethanol
300 35 4.5 95 0.65 conversion to EtCl was much higher than for ethylene or ethyl chlo-
TCA: trichloroethane. ride oxychlorination to EDC at the same conditions, which is a clear
a
Feed (vol%): 2%EtCl, 18%HCl, 31%air; GHSV = 11550 h−1 ; 1 barg. indication that EtCl is formed by the rapid direct hydrochlorination
of ethanol (7), and not by dehydration to ethylene (8) followed by
hydrochlorination of ethylene by reverse of (4). In fact as pointed
18, and 31 vol% for C2 H4 , HCl, and air respectively at a GHSV of out above, the equilibrium of (4) lies towards ethylene at the con-
11,550 h−1 . This feed composition and reaction temperature is typ- ditions used and it does not take place at a significant rate over the
ical of industrial oxychlorination processes [16]. The results are catalyst.
given in Table 1. The catalyst showed good activity for C2 H4 oxy-
chlorination with high selectivity to EDC. Small amounts of EtCl C2 H5 OH + HCl → CH3 CH2 Cl + H2 O (7)
and VCM were detected with a total selectivity less than 2%. VCM
increased with increasing temperature due to dehydrochlorination C2 H5 OH → C2 H4 + H2 O (8)
of the product EDC over the ␥-Al2 O3 according to reaction (3). The
ethylene conversion reached 25% at 300 ◦ C with corresponding to EDC increases as temperature and selectivity to EtCl falls
a reaction rate of 3.2 mmol l−1 s−1 , and apparent activation energy suggesting a sequential reaction of EtCl to EDC, mainly by oxychlo-
of 70 kJ/mol. rination (6) as noted above (there is a loss of selectivity to EtCl at the
lower temperature due to the formation of oxygenated by-products
3.1.3. Ethyl chloride oxychlorination such as acetaldehyde), Table 3. The EDC cracking to VCM was very
Ethyl chloride oxychlorination was studied from 250 to 300 ◦ C low at all conditions due to the low acidity of the NaCl promoted
with a feed composition of 2, 18, and 31 vol% for EtCl, HCl, and air, ␥-Al2 O3 catalyst and agrees well with the findings of Carmello et
respectively. The results are given in Table 2. The reaction selec- al. [17]. Diethyl ether (DEE) is observed at the higher temperatures
tively produced EDC and ethylene with the selectivity to ethylene due to a small amount of dehydration of ethanol on the ␥-Al2 O3
falling as temperature increased. Trichloroethane, TCA, was the support. The low selectivity to ethylene is also due to the low acid-
only significant by-product, which is produced by the hydrochlo- ity of the catalyst which does not dehydrate ethanol to ethylene
rination reaction of VCM which was not detected. The conversions significantly below 300 ◦ C [14].
at 250 and 275 ◦ C are lower and similar in magnitude to ethylene The effect of temperature on the extent of these reactions was
oxychlorination, Table 1, which might be taken as evidence that opposite to that reported in US patents 4,814,527 (1989) and
the route to EDC may be via dehydrochlorination as in (4) followed 4,910,176 (1990) [9,10], although it was noted that the ratio of
by oxychlorination of ethylene according to (2). Thermodynamic EDC/ethylene could be varied over a wide range by balancing the
calculation by Aspen shows that the equilibrium of (4) lies toward CuCl2 loading and support acidity. Our results show that higher
ethylene at the conditions used. However, reaction (4) is known reaction temperature favours the production of EDC rather C2 H4 .
to be catalysed by acidic sites on the alumina support which are This is a consequence of the dominance of the direct oxychlori-
blocked by CuCl2 and NaCl in typical oxychlorination catalysts [17], nation of EtCl (6) due to the high loading of CuCl2 compared to
suggesting that this is not the main route to EDC. The dispropor- the zeolite supported catalysts of US patents 4,814,527 (1989) and
tionation of EtCl to EDC and ethylene (5) followed by ethylene 4,910,176 (1990), and the relatively low catalyst acidity. In any
oxychlorination to EDC by (2) and direct oxychlorination of ethyl case, any produced ethylene from disproportionation (5) and/or
chloride to EDC (6) can also take place. At 300 ◦ C, the conversion dehydrochlorination of EtCl (4), would be oxychlorinated to pro-
of ethyl chloride to EDC, Table 2, is greater than that of ethylene duce more EDC as the temperature increased further increasing
to EDC, Table 1. The much lower concentration of ethyl chloride the selectivity to EDC.
Table 3
EtOH oxychlorinationa over 14 wt%CuCl2 –5 wt%NaCl/␥-Al2 O3 .
Table 4
Effect of space velocity at different reaction temperaturesa .
3.2. Ethanol oxychlorination over the dual-catalyst bed 3.2.2. Effects of temperature and space velocity
configuration A pre-bed of ZSM-5(50) was used to explore the effects of
temperature, 250–300 ◦ C and space velocity on the product dis-
3.2.1. Effect of zeolite tribution, Table 4. Higher space velocity over the oxychlorination
The direct production of EDC from EtOH was carried out at catalyst in the dual-catalyst bed system was achieved by reduc-
250–300 ◦ C using a dual-catalyst bed configuration consisting of a ing the catalyst volume of the second bed (and a corresponding
zeolite catalyst in the pre-bed and the 14 wt%CuCl2 –5 wt%NaCl/␥- increase in the dilution), so that the flow rate and hence residence
Al2 O3 catalyst in the second bed. A range of zeolites (ZSM-5(50), time in the pre-bed remained unchanged. The ethanol conversion
ZSM-5(280), mordenite(13), Y-zeolite(80)) were investigated ini- was always high and decreased only slightly at the higher space
tially for the pre-bed spanning variation in acidity and structure. In velocity showing that almost all the ethanol was converted over
general, while EtCl was the main product at 250 ◦ C EDC increased the ZSM-5 pre-bed. The selectivity to EDC decreased sharply for
with temperature becoming the main product for all except the the higher space velocity indicating that it is formed almost exclu-
ZSM-5(280) pre-bed at 300 ◦ C. Fig. 2 shows the results obtained sively on the second bed. This was confirmed by a study of ethanol
at 300 ◦ C, 1 barg operating pressure, 11,550 h−1 GHSV (based on reaction with HCl and air over the ZSM-5(50) alone, Table 6, which
the volume of the CuCl2 catalyst bed or zeolite catalyst bed) with a shows almost the same ethanol conversion, but low selectivity to
feed gas composition of 8 vol%EtOH, 18 vol%HCl, and 31 vol%air. The EDC and higher selectivities to ethylene and ethyl chloride at the
selectivity to VCM was very low reaching 0.3% maximum at 300 ◦ C same reaction temperature. At both space velocities, the selectivity
indicating a low EDC cracking rate over the dual-bed catalyst con- to EDC increased with reaction temperature at the expense of EtCl,
figuration. By far the best performance in terms of both ethanol further underlining the sequential nature of EDC formation.
Table 5
Effect of feed gas compositiona .
Table 6
EtOH oxychlorinationa over ZSM-5.
3.2.3. Effects of HCl and air composition version of less than 10%. On balance therefore, since much less EDC
Table 5 gives the effects of HCl and air compositions in the feed than ethylene is produced, it is likely that the ethylene is produced
gas stream on the product distribution over ZSM-5(50) pre-bed and mainly by direct dehydration (8).
14 wt%CuCl2 –5 wt%NaCl/␥-Al2 O3 at 275 ◦ C and 11550 h−1 GHSV. Comparing the results obtained from ZSM-5(280) and ZSM-
The main products were EDC, EtCl, and ethylene. The minor by- 5(50), Table 6, we see the effect of zeolite acidity. Ethanol
products were methyl chloride, CH3 Cl, acetaldehyde (AcA), vinyl conversion is higher for the more acid zeolite as expected. How-
chloride, VCM, and DEE. In both cases, a maximum in EDC selec- ever, at the same time the selectivity to EtCl is lower and that
tivity (minimum EtCl) was observed indicating that an optimum to C2 H4 is higher. Again, this probably reflects the greater direct
feed composition was required for efficient EDC formation. This ethanol dehydration activity of the more acidic zeolite, although as
probably relates to the formation of the catalytically active state of noted above dehydrochlorination of the main product ethyl chlo-
the copper in the catalyst, as evidenced by XPS studies [15]. These ride via (4) would also be expected to be greater over ZSM-5(50).
results may also explain the lower EDC yield over CuCl2 /zeolite It is also worth noting that despite the low level of EDC produc-
reported in US patents 4,814,527 (1989) and 4,910,176 (1990) tion, the higher zeolite acidity favours the disproportionation of
[9,10]. We believe that the high excess of air and low HCl/ethanol EtCl producing EDC as expected.
ratio used was the reason why only 27% EDC selectivity was
achieved.
3.4. Dual-catalyst bed vs. physical mixture
3.3. Reactions over ZSM-5 The efficiency of the 2-bed configuration for ethanol oxy-
chlorination was compared to a physical mixture of ZSM-5(50)
In order to understand the role of zeolite in the direct pro- and 14 wt%CuCl2 –5 wt%NaCl/␥-Al2 O3 catalyst. The results obtained
duction of EDC from ethanol in the dual-catalytic bed system, the from the physical mixture are given in Table 7. The physical mixture
ethanol reaction with HCl and air was studied over ZSM-5(280) and showed similar dependency on temperature to the dual-bed: high
ZSM-5(50), Table 6. For both zeolites, the main product was EtCl selectivity to ethyl chloride at 250 ◦ C and high selectivity to EDC
with some C2 H4 and EDC. The other by-products were methane, at 300 ◦ C. However, ethanol conversion and the selectivity to EDC
CH3 Cl (methanol hydrochlorination), acetaldehyde, AcA (ethanol were always lower than achieved over the dual-catalyst bed con-
oxidation), VCM (EDC cracking), CHCl3 (CH3 Cl chlorination), and figuration for the same temperature, gas composition and GHSV. A
DEE (ethanol dehydration). Two main parallel reactions took place significant difference was the higher selectivity to VCM at 300 ◦ C
over the zeolites: hydrochlorination of ethanol producing EtCl and for the physical mixture. There could be several reasons for the
water (7), and ethanol dehydration producing C2 H4 and water (8). difference between the dual-catalyst bed and the equivalent phys-
Thermodynamic calculations showed that reaction (7) can proceed ical mixture. The ZSM-5(50) in the physical mixture is exposed to
to completion at the conditions used. The relative contribution of a significantly different gas composition compared to when it is a
these reactions depends on the temperature and zeolite properties. pre-bed, and this may give rise to more cracking of EDC. For exam-
At lower temperature the selectivity is overwhelmingly towards ple, in the separate bed configuration, the main EDC production was
ECl. Higher the reaction temperature leads to increased produc- over the second bed and the only available sites for cracking would
tion of C2 H4 . EtCl can disproportionate producing C2 H4 and EDC be the exposed surface of the ␥-Al2 O3 which has much lower acidity
according to reaction (5) and dehydrochlorinate as in (4), which is than ZSM-5(50) [17]. In the physical mixture, part of the EDC would
thermodynamically favoured especially at the higher temperature. be produced in the early zone of the reactor, leaving it exposed to
Dehydrochlorination is known to be catalysed by acid sites [17]. further contact with the acid catalyst as it passed down the reactor
However, measurement of ethyl chloride dehydrochlorination to leading to increased cracking to VCM. This effect could be mitigated
ethylene over ZSM-5(50) at 300 ◦ C in the presence of air gave a con- by axial grading of the physical mixture.
Table 7
EtOH oxychlorinationa over a physical mixture of ZSM-5(50) and 14 wt%CuCl2 –5 wt%NaCl/␥-Al2 O3 .
4. Conclusions References
We have demonstrated that a dual-catalyst bed of zeo- [1] B.K. Hodnett, Heterogeneous Catalytic Oxidation: Fundamental and Technolog-
ical Aspects of the Selective and Total Oxidation of Organic Compounds, Wiley,
lite + oxychlorination catalyst can be used for the selective 2000.
production of ethylene dichloride from ethanol with minimum [2] A. Demirbas, Energy Sources, Part A: Recovery, Utilization, and Environmental
cracking to VCM. Greater than 80% EDC yield was achieved using Effects 28 (2006) 955–964.
[3] C. Cardona, J. Quintero, I. Paz, Bioresource Technology 101 (2010) 4754–
a ZSM-5 pre-bed with ethyl chloride and ethylene as the main 4766.
by-products. An optimum feed composition of HCl and air was [4] K. Fang, D. Li, M. Lin, M. Xiang, W. Wei, Y. Sun, Catalysis Today 147 (2009)
required to activate the oxychlorination catalyst and maximize 133–138.
[5] X. Zhang, Z. Li, Q. Guo, H. Zheng, K. Xie, Fuel Processing Technology 91 (2010)
the EDC yield. The dual-bed configuration gave better performance 379–382.
than the equivalent physical mixture. It was shown that EDC was [6] V. Subramani, S. Gangwal, Energy & Fuels 22 (2008) 814–839.
mainly produced by direct hydrochlorination of ethanol to ethyl [7] M. Bruscino, Hydrocarbon Processing (August) (2010) 65.
[8] H.A. Cheney, US patent 2,442,285 (1948).
chloride followed by the oxychlorination and disproportionation of
[9] R.W. Diesen, US patent 4,814,527 (1989).
ethyl chloride. Ethanol dehydration followed by ethylene oxychlo- [10] R.W. Diesen, US patent 4,910,176 (1990).
rination playing the minor role in the temperature range studied. [11] T.M. Nguyen, R.L.V. Mao, Applied Catalysis A 61 (1990) 119–
The results open the possibility of a single-stage process from bio- 129.
[12] C. Phillips, R. Datta, Industrial and Engineering Chemistry Research 36 (1997)
derived ethanol to EDC. 4466–4475.
[13] I. Takahara, M. Saito, M. Inaba, K. Murata, Catalysis Letters 105 (2005) 249–
Acknowledgments 252.
[14] X. Zhang, R. Wang, X. Fang, F. Zhang, Microporous & Mesoporous Materials 116
(2008) 210–215.
We thank DTI(UK) and Ineos Technology for financial support [15] P. Prasad, K. Prasad, P. Rao, V. Kaushik, Journal of Materials Science 32 (1997)
under contract TP/16307, and in particular Dr. Andrea Marsella of 1479–1482.
[16] J.W.S. Charles, E. Wilkes, C.A. Daniels (Eds.), PVC Handbook, Hanser Gardner
Ineos Vinyls for general advice on oxychlorination. We are grateful Publications, 2005.
to Dr. Sami A.I. Barri for advice and help with XRD, and Dr. Laura [17] D. Carmello, E. Finocchio, A. Marsella, B. Cremaschi, G. Leofanti, M. Padovan,
Torrente Murciano for help with the equilibrium calculations. Journal of Catalysis 191 (2000) 354–363.