Fuel Processing Technology 81 (2003) 201 – 209

www.elsevier.com/locate/fuproc

Synthesis, characterization, and application of

hydotalcites in hydrodesulfurization of FCC gasoline
Ruiyu Zhao*, Changlong Yin, Huiji Zhao, Chenguang Liu
College of Chemistry and Chemical Engineering, Key Laboratory of Catalysis, CNPC, University of Petroleum,
Dongying, Shandong 257061, China
Accepted 1 November 2002

Abstract

Magnesium – aluminum, copper – aluminum, zinc – aluminum hydrotalcite(HT) compounds were

synthesized using co-precipitation method. The effects of stirring rate, feeding rate of reactants, pH,
calcination temperature on the properties of Mg – Al mixed oxides were investigated by using XRD,
TG-DTA and BET techniques. The catalytic activity and selectivity of CoMo/g-Al2O3, CoMo/g-
Al2O3+HT catalysts for hydrodesulfurization of FCC gasoline were examined in a high pressure
microreactor. The results showed that the catalysts with mixed oxide obtained from hydrotalcite as
support give lower levels of olefin hydrogenation than the catalyst with g-Al2O3 as support, and
lower octane number reduction and hydrodesulfurization yield, too. The hydrodesulfurization
activity order was as follows: CoMo/g-Al2O3>CoMo/g-Al2O3+Mg – Al(HT)>CoMo/g-Al2O3+Cu –
Al(HT)>CoMo/g-Al2O3+Zn – Al(HT), and the olefin hydrogenation activity and octane number
reduction order were as follows: CoMo/g-Al2O3+Zn – Al(HT)<CoMo/g-Al2O3+Cu – Al(HT)<CoMo/
g-Al2O3+Mg – Al(HT)<CoMo/g-Al2O3.
D 2003 Elsevier Science B.V. All rights reserved.

Keywords: Hydrotalcites; Synthesis; Hydrodesulfurization; Catalyst; Activity

1. Introduction

It is well known that a major source of air pollution is the exhausts from fuel combusted
in hundreds of millions of motor vehicles. Regulations have been enacted to reduce
harmful motor vehicle emissions through more stringent fuel standards. Fuels containing

* Corresponding author. Tel.: +86-546-839-1710; fax: +86-546-839-2284.

E-mail address: [email protected] (R. Zhao).

0378-3820/03/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0378-3820(03)00012-2
202 R. Zhao et al. / Fuel Processing Technology 81 (2003) 201–209

sulfur give sulfur oxide, which is the largest contributor to NOX emissions from fuel
combustion [1]. On the other hand, sulfur compounds in the fuel are poisonous to catalytic
converter fixed in the automobile. To reduce or resolve these environmental problems, the
sulfur content in the fuel has been, and will continue to be reduced to a lower and lower
level.
All cuts go to the gasoline pool, nearly all of the sulfur (85 –95%), in a typical refining
factory, comes from FCC gasoline containing a high olefin content and thus a high octane
number [2,3]. As a result, FCC gasoline is the focus for sulfur reduction. Elimination of
the sulfur from cracked naphtha is easy, but during the HDS processing using traditional
HDS catalysts, saturation of the olefins occurs significantly, resulting in the great reduction
of the octane number. Therefore, the key technical problem for HDS of FCC gasoline is to
do deep sulfur removal, at same time, to reduce the loss of octane number occurring in the
HDS process. One feasible method for the aim is to design a catalyst with favorable
activity that can finish needed HDS but with lower loss of octane number.
In order to reach the target, some researches have been carried out. Yu and Myers [4]
and Sudhakar [5] disclose selective desulfurization of cracked naphtha by using special
catalysts with hydrotalcite as support. Kukes et al. [6] reported the use of novel catalysts
comprising a hydrogenation component and a support component which comprises a
magnesium component and an alkali metal component. Hydrotalcite compounds have
been used in recent years as precursors of oxide catalysts and catalyst support [7 –9]. It is
the key for HDS of FCC gasoline to prepare an efficient catalyst for removing sulfur from
olefin containing feedstocks with minimum loss of octane number. Now, it has been found
that the selectivity for hydrodesulfurization was improved significantly by using a catalyst
support comprising essentially a Mg – Al hydrotalcite composition, but HDS activity was
reduced unavoidably [10].
In this paper, Mg – Al, Cu – Al, Zn –Al hydrotalcite compounds were synthesized using
co-precipitation method. The activity and selectivity of CoMo sulfide catalysts supported
on Mg – Al, Cu – Al, Zn – Al mixed oxide supports obtained from hydrotalcite compounds
were evaluated by hydrodesulfurization of FCC gasoline. TG-DTA, XRD, BET techniques
were used to characterize the structure and properties of supports. The relationship
between the hydrodesulfurization activity and catalysts properties were discussed. The
discussion is focus on the affection of support on the HDS, hydrogenation(HYD) activity
and research octane number (RON) loss of FCC gasoline.

2. Experimental

2.1. Preparation of hydrotalcite

Hydrotalcite samples were prepared by co-precipitating an aqueous solution of

magnesium and aluminum salts with a highly basic carbonate solution [11]. The first
solution (A solution) containing Mg(NO3)2
6H2O and Al(NO3)3
9H2O, Cu(NO3)2
6H2O
and Al(NO3)3
9H2O, Zn(NO3)2
6H2O and Al(NO3)3
9H2O, respectively, dissolved in
deionized water at Mg/Al, Cu/Al, Zn/Al molar ratio of 3:1. The second solution (B
solution) containing appropriate amounts of sodium hydroxide – sodium carbonate was
 R. Zhao et al. / Fuel Processing Technology 81 (2003) 201–209 203

added on a certain ratio so that the final Al/CO32 molar ratio equals to 2 and the pH of the
final solution was kept at 8.5F0.3. Two solutions were added dropwise at different rate to
keep the pH of the slurry in the range of 8.5F0.3. Then the mother liquid was poured into
a closed vessel with inner liner of teflon(polytetrafluoroethylene) and aged at 110 jC for
12 h. The precipitate was separated by filtering and washed with deionized water, and then
the precipitate was dried at 393 K overnight.

2.2. Preparation of catalysts

The g-Al2O3 was mixed with Mg – Al, Zn – Al, Cu – Al hydrotalcites in mass ratio of
1:3, respectively, to obtain the intermixtures. The g-Al2O3 and the intermixtures were
mixed with 1% nitric acid, then extruded into extrude bar of 1.6 mm diameter. The
extrudes were dried at 383 K overnight, then calcined at 773 K for 4 h to obtain supports.
CoMo/g-Al2O3+Mg – Al, CoMo/g-Al2O3+Zn – Al, CoMo/g-Al2O3+Cu – Al and CoMo/g-
Al2O3 were prepared by a standard incipient wetness technique. The calcined supports
were impregnated with an aqueous solution of ammonium heptamolybdate and cobalt
nitrate according to the required Mo and Co loading. After each impregnation, the supports
were dried at 383 K overnight, then calcined at 723 K for 4 h.

2.3. Characterization of the catalysts

The specific surface area and average pore size of supports were determined by
nitrogen adsorption using an ASAP 2010 micrometer instrument.

Fig. 1. Effect of stirring rate on the physical properties of Mg – Al hydrotalcite.

204 R. Zhao et al. / Fuel Processing Technology 81 (2003) 201–209

Table 1
Effect of feeding rate of reactants on the physical properties of Mg – Al hydrotalcite
Feeding ratea, ml/min BET surface area, m2/g Average pore size, Å
4.0 237.9 129.9
8.0 259.2 121.9
12.0 242.0 116.0
a
Volume flow rate of A solution.

X-ray powder diffraction analysis was carried out with a Rigaku D/max-II diffrac-
tometer using a graphite-filter CuKá radiation at a scan rate of 2j/min.
TG-DTA analysis was carried out with WCT-2 computer controlled TG-DTA meter
using temperature program of 5 jC/min.
Catalytic activity measurements were carried out in a high pressure microreactor. The
catalyst was ground to 20 –40 mesh and 8 ml catalyst was loaded into the reactor.
Presulfiding of catalysts was carried out at 260 jC for 3 h, and 280 jC for 3 h with a liquid
stream containing 3 wt.% CS2 in cyclohexane. FCC gasoline was then pumped at test
temperature. Reaction conditions are temperature 280 jC, pressure 2.1 MPa, H2/oil
volume ratio 150, space velocity 3.53 h 1. The average hydrodesulfurization activity
was calculated as the percentage conversion of sulfur compounds and in units of HDS%.
The average hydrogenation activity was calculated as the conversion of olefin and in units
of HYD%. The PIONA (paraffin, isoalkane, olefin, naphthene, and aromatics) components
and RON (research octane number) of feedstock and effluents were analyzed by GC with
FID detector.

3. Results and discussion

3.1. Effects of preparation conditions on the physical properties of hydrotalcites

The relationship between the stirring rate and the BET surface area and average pore
size of Mg – Al mixed oxide is showed in Fig. 1. The experimental results in the Fig. 1

Fig. 2. The effects of aging time on relative crystallinity of Mg – Al hydrotalcite at 65 jC.

 R. Zhao et al. / Fuel Processing Technology 81 (2003) 201–209 205

Fig. 3. The effects of aging time on relative crystallinity of Mg – Al hydrotalcite at 110 jC.

reveal that the BET surface area varies slightly with the increase of stirring rate, but the
average pore size increases rapidly with the rising of stirring rate.
The effects of reactant feeding rate on the properties of Mg – Al hydrotalcite are listed in
the Table 1. The experimental results showed that the feeding rate of reactants affects the
physical properties of Mg – Al oxides. The samples have lower BET surface area at both
lower and faster feed rate and bigger BET surface area when the feed rate of A solution is 8
ml/min. However, the average pore size was reduced gradually with increase of the feed
rate of A solution.
The effects of aging temperature and time on relative degree of crystallinity are showed
in Figs. 2 and 3. The experiment results in Figs. 2 and 3 show that the relative crystallinity
increases with rising of aging temperature and prolonging of crystallization time, however,
the crystallization time decreases with rising of aging temperature. When aging at 110 jC,
the relative crystallinity reaches about 100% after 4 h, whereas aging at 65 jC the relative
crystallinity just reaches about 85% after 36 h. The results in Fig. 4 reveal that the crystal
size of Mg –Al hydrotalcite increases with the aging time at 110 jC, so the sample

Fig. 4. The effects of aging time on crystal size of Mg – Al hydrotalcite at 110 jC.
206 R. Zhao et al. / Fuel Processing Technology 81 (2003) 201–209

Fig. 5. XRD patterns of hydrotalcite samples obtained at different aging temperature; (a) hydrothermal aging at
110 jC, (b) aging at 65 jC.

obtained from aging at 110 jC has bigger crystal size. In a word, aging at 110 jC can give
bigger crystal size and need shorter time to finish crystallization.
The XRD patterns of sample obtained by hydrothermal aging at 110 and 65 jC were
illustrated in Fig. 5. The XRD peaks of the samples obtained from hydrothermal aging at
110 jC are narrower in width and higher in intensity, compared with the sample obtained
from aging at 65 jC. It is well known that the crystal size is inverse proportional to the half
band width according to Scherrer’s equation. Therefore, above results indicate that the
hydrotalcite aged at 110 jC has more uniform crystal and bigger crystal size.
The X-ray diffraction patterns of Zn –Al, Mg –Al, Cu – Al hydrotalcites prepared in our
laboratory were showed in Fig. 6. The patterns reveal that all of the hydrotalcites samples
of Zn – Al, Mg –Al, Cu – Al have similar XRD pattern and typical hydrotalcite’s structure.
It is also found that the XRD peaks, particularly 003 peak, of the Zn –Al hydrotalcite are
narrower in width and higher in intensity, compared with the Mg – Al hydrotalcite, while
the Cu –Al hydrotalcite has widest band and lowest intensity. These indicate that the Zn –

Fig. 6. The X-ray diffraction patterns of hydrotalcites; (a) Zn – Al HT, (b) Mg – Al HT, (c) Cu – Al HT.
 R. Zhao et al. / Fuel Processing Technology 81 (2003) 201–209 207

Fig. 7. TG-DTA profiles of Mg – Al hydrotalcites.

Al hydrotalcite has more uniform crystal and bigger crystal size than the Mg – Al and Cu –
Al hydrotalcites.
Figs. 7 – 9 provides the TG-DTA profiles of Mg –Al, Cu – Al and Zn – Al hydrotalcites.
The results of TG-DTA analysis of hydrotalcites revealed that Mg – Al hydrotacite has two
obvious weight loss processes. The Mg – Al hydrotacite loses interbedded water at about
285 jC, while loses OH 1 and anion in the layer and between the layers, respectively, at
about 448 jC, and completed decomposition over about 550 jC. As to Cu – Al hydro-
talcite, it loses interbedded water at about 218 jC, and loses OH 1 and anion in the layer
and between the layers, respectively, at about 440 jC, and completed decomposition over
about 520 jC. As to Zn –Al hydrotalcite, it loses interbedded water at 322 jC, and loses
OH 1 and anion in the layer and between the layers at about 483 jC. These indicate that
500 –550 jC is the appropriate temperature for calcination of supports containing hydro-
talcites.

3.2. Catalytic activities of Co –Mo catalysts supported on hydrotalcites

The experimental results of various catalyst evaluated for the hydrodesulfurization of

FCC gasoline are listed in Fig. 10 (RONR: octane number retention percentage of
effluents).

Fig. 8. TG-DTA profiles of Cu – Al hydrotalcites.

208 R. Zhao et al. / Fuel Processing Technology 81 (2003) 201–209

Fig. 9. TG-DTA profiles of Zn – Al hydrotalcites.

The results of Fig. 10 display that the catalyst with g-Al2O3 as support and Co –Mo as
active components has higher HDS and hydogenation (HYD) activity, but lower octane
number retention percentage (RONR). As to the catalysts with hydrotalcites as support, the
results show that they have lower HDS and HYD activity, but have higher RONR
compared with Co –Mo/g-Al2O3 catalyst. This indicated that hydrotalcites increase the
selectivity of catalysts for hydrodesulfurization of sulfur-containing compounds over
hydrogenation of olefins. Of the hydrotalcite-containing catalysts, Zn –Al catalyst shows
lowest HDS and HYD activity, and highest RONR, while Cu – Al and Mg –Al catalyst
show lower HDS and HYD activity, and higher RONR than CoMo/g-Al2O3 catalyst. The
reason for CoMo/g-Al2O3+Mg –Al catalyst having the lower HDS and HYD, as well as
higher RONR, can attribute in part to that the Mg – Al mixed oxide has greater number of
basic sites per unit of surface area compared with g-Al2O3. As to Cu – Al and Zn – Al
hydrotalcite catalysts, there are still mechanism problems need to make clear.

Fig. 10. Activity and selectivity of Co – Mo catalysts for hydrodesulfurization of FCC gasoline (conditions:
pressure 2.0 MPa, temperature 280 jC, H2/oil ratio 150, LHSV 3.5 h 1).
 R. Zhao et al. / Fuel Processing Technology 81 (2003) 201–209 209

4. Conclusions

The properties of Mg-Al mixed oxides obtained from hydrotalcite were affected by the
preparation conditions of hydrotalcites. BET surface areas increase with the raise of
stirring rate. The hydrotalcite has bigger surface area and the greater average pore size
when the feeding rate of reactants was 4 ml/min of Mg – Al salt solution. The Mg – Al
mixed oxides show the biggest surface area when the hydrotalcite was calcined at 550 jC,
while it has the greater average pore size at 500 jC. It can also be concluded from the
results of desulfurization test of FCC gasoline that the Co –Mo catalysts supported
on hydrotalcite are superior to the Co – Mo catalyst supported on alumina for the
octane number retention. The hydrodesulfurization activity order was as follows:
CoMo/g-Al2O3>CoMo/g-Al2O3+Mg – Al(HT)>CoMo/g-Al2O3+Cu – Al(HT)>CoMo/g-
Al2O3+Zn –Al(HT), and the olefin hydrogenation activity and octane number reduc-
tion orders were as follows: CoMo/g-Al2O3+Zn – Al(HT)<CoMo/g-Al2O3+Cu –
Al(HT)<CoMo/g-Al2O3+Mg – Al(HT)<CoMo/g-Al2O3.

Acknowledgements

This work was supported by the CNPC and State Key Laboratory of Heavy Oil
Processing.

References

[1] K.L. Rock, H.M. Putman, NPRA Annual Meeting, San Francisco, AM-98-37 (1998).
[2] L.L. Upson, M.W. Schnaith, Low-sulfur specifications cause refiner to look at hydrotreating options, Oil
Gas J. 8 (1997) 47 – 51.
[3] J.A. Anabtarri, S.A. Ali, M.A. Ali, Impact of gasoline and diesel specifications on the refining industry,
Energy Sources 18 (1996) 203 – 214.
[4] A.P. Yu, E.C. Myers, U.S. Patent 4,132,632 (1979).
[5] C. Sudhakar, U.S. Patent 5,851,382 (1998).
[6] S.G. Kukes, H.P. Donald, R.L.A. Olleandorff, et al., U.S. Patent 5,348,928 (1994).
[7] R.J. Davis, E.G. Derouane, Nature 349 (1991) 313.
[8] H. Schaper, J.J. Berg-Slot, W.H. Stork, Appl. Catal. 54 (1989) 79.
[9] O. Clause, B. Rebours, E. Merlen, F. Trifiro, A. Vaccari, J. Catal. 133 (1992) 231.
[10] P.-S.E. Dai, J. Sherwood, E. David, et al., U.S. Patent 5,340,466 (1994).
[11] J. Shen, J.M. Kobe, Y. Chen, J.A. Dumesic, Langmuir 10 (1994) 3902.