Hydrotreatment and Hydrocracking of Oil Fractions B. Delmon, G.F. Froment and P. Grange (Editors) �9 1999 Elsevier Science B.V. All rights reserved. 161

Effect of Chelating Agents on HDS and Aromatic Hydrogenation over CoMo- and NiW/AI20 3

Yukie Ohta, Takehiro Shimizu, Takehide Honma, and Muneyoshi Yamada

Department of Applied Chemistry, Graduate School of Engineering, Tohoku University

Aoba, Aramaki, Aoba-ku, Sendai 980-8579, JAPAN

Abstract

Hydrotreatment catalysts (CoMo-, NiMo- and NiW/A1203) were prepared by an

impregnation method with solutions containing a chelating agent (nitrilotriacetic acid (NTA),

ethylenediaminetetraacetic acid (EDTA) or cyclohexanediaminetetraacetic acid (CyDTA)),

and were subjected to some test reactions: hydrodesulfurization (HDS) ofbenzothiophene and

dibenzothiophene, and hydrogenation (HGN) of o-xylene and 1-methylnaphthalene.

Chelating agent modified CoMo- and NiW/A1203 showed higher activities in both HDS

and HGN than the catalysts without the modification. The chelating agents had little effect on

the activity of NiMo/A1203. CyDTA was the most effective for HDS activity of CoMo/A1203

and HGN activity of NiW/A1203. The chelating agents had no effect on the activity of each

component catalysts (i.e., Co-, Ni-, Mo-, and W/A1203). The activity of the CyDTA-modified

catalysts depended strongly on sulfiding temperature. CyDTA exhibited the improving effect

at higher sulfiding temperatures, while working as an inhibitor at lower temperatures. Our

previous study had indicated that the pre-formation of the MoS2-1ike structure was necessary

to induce the intrinsic promoting effect of Co on the activity of Mo/A1203. The chelating

agent, forming Co complex which decomposes at a rate depending on sulfiding temperature,

was considered to adjust timing when Co ions interact with the MoS2-1ike structure.

1. I N T R O D U C T I O N

From the recent environmental point of view, hydrotreatment of petroleum fractions

to produce clean and high quality transportation fuels is becoming more and more important.

In particular, HDS and aromatic HGN of diesel fuel are very important, and there is a

growing need for improvements in catalyst performance of CoMo/A1203 and NiW/A1203,

which are widely used for HDS and HGN treatments of petroleum fractions.

162

The catalysts mentioned above are conventionally prepared by impregnation

followed by drying, calcination, and sulfiding for activation. In order to obtain a high

performance catalyst, every preparation step should be optimized. Since one of the most

important characteristics of these catalysts is the activity-promoting effect of Co or Ni, the

preparation steps should be optimized to induce the intrinsic promoting effect of Co or Ni.

Van Veen group reported that CoMo catalysts prepared by impregnating the supports (carbon,

silica, or A1203 ) with a solution containing Mo, Co and NTA showed higher activity for HDS

of thiophene at atmospheric pressure. They have assigned the higher HDS activity to the

formation of a complex of NTA with Mo and Co on the supports. Being stimulated by their

work (HDS of thiophene at atmospheric pressure), we have started to investigate the

applicability of other chelating agents in other reactions.

In the present work, impregnating solutions containing NTA, EDTA or CyDTA were

used to prepare modified CoMo- and NiW/A1203. EDTA forms complexes with both Mo and

Co as well as NTA does, while CyDTA only with Co.

2. E X P E R I M E N T A L

Every catalyst examined here was prepared by an incipient wetness method as

follows [2, 3] : -A1203 was impregnated with an aqueous solution containing a chelating

agent, cobalt nitrate (or nickel nitrate) and ammonium paramolybdate (or ammonium

metatungstate), then dried at 393 K in air. The molar ratio of a chelating agent to Mo (or W)

was 1.2 for NTA, and 0.6 for EDTA or CyDTA. These catalysts are abbreviated hereinafter

as "chelating agent"-"combination of metals", e.g., NTA-CoMo. The catalysts thus obtained

were used without calcination.

The catalysts were packed in a conventional fixed bed flow reactor and sulfided in-

situ in the stream of 5% HzS/H 2 under 1.1 MPa at 673 K. Immediately after the sulfiding,

activity tests were started by feeding the reactant into the reactor. The reaction conditions of

HDS were as follows: 5 tool% benzothiophene (BT) in n-dodecane (or 2.5 mol%

dibenzothiophene (DBT) in decalin), 543 (or 573) K, 5.1 MPa, LHSV 300 h -1, in H2 stream

(300 ml/min), Hz/reactant feed 1,000 vol/vol. The reaction conditions of HGN were as

follows: o-xylene (or 1-methylnaphthalene (1-MN)), 573 K (603 K), 5.1 MPa, LHSV 7 (or

75) h -1, in H2 stream (300 ml/min), HJreactant feed 1,000 vol/vol. Products were analyzed

with GC (FID) and/or GC-MS. Details of the apparatus and the procedure were described in

our preceding papers [4, 5].

163

3. R E S U L T S A N D D I S C U S S I O N

3.1. H D S R e a c t i o n s

In HDS reaction of BT, ethylbenzene (EB) and a little amount of

dihydrobenzothiophene (DHBT) were produced. In order to compare the catalyst activities

easier, the conversion level of BT was controlled below 30-40% by adjusting the catalyst

loading

EDTA-CoMo NTA-CoMo

CyDTA-CoMo CoMo

Commercial CoMo

(a) EDTA-CoMo NTA-CoMo

CyDTA-CoMo CoMo

Commercial CoMo

biph~yl cyclohexylbenzene,,,1

~ ] , 6 (a)

0 0.3 0.6 0.9 1.2 EB Yield/metal/%/gmol

0 0.2 0.4 0.6 0.8 1 BP and CHB Yields/metal/%/gmol

EDTA-NiW NTA-NiW

CyDTA-NiW NiW

Commercial NiW

(b)

0 0.3 0.6 0.9 1.2 EB Yield/metal/%/gmol

EDTA-NiW

NTA-NiW

CyDTA-NiW

NiW

(b)

0 0.2 0.4 0.6 0.8 1 BP and CHB Yields/metal/%/gmol

EDTA-NiMo

NTA-NiMo

CyDTA-NiMo

NiMo

0 0.3 0.6 0.9 1.2 EB Yield/metal/%/gmol

CyDTA-NiMo ~ (c)

NiMo I I I I

0 0.2 0.4 0.6 0.8 1 BP and CHB Yields/metal/%/gmol

Figure 1. Effects of chelating agents on Figure 2. Effects of chelating agents on benzothiophene HDS activity, dibenzothiophene HDS activity. Reaction conditions : 543 K, 5.1 MPa. Reaction conditions : 573 K, 5.1 MPa.

in the reactor. The EB yield was used as an index of HDS activity of the catalysts. The

catalyst activities thus obtained, were compared based on the %EB yield/mol-metal. The

activities of both CoMo and NiW were improved by the addition of the chelating agents in the

following order: C y D T A > EDTA > NTA > none (Figures 1 (a) and 1 (b)).

C y D T A - C o M o and C y D T A - N i W attained about 70% and 65% higher HDS

activities than the corresponding unmodif ied catalysts, respectively. On the other hand,

164

activity of NiMo was not affected by the chelating agents (Figure 1(c)).

Effects of these chelating agents on the catalyst activity were further examined in

HDS of DBT. In this reaction, biphenyl (BP) was mainly produced and cyclohexylbenzene

(CHB) was also produced. The sum of BP and CHB yields was regarded as HDS activity of

each catalyst. The activities of CoMo and NiW were also improved by the addition of the

chelating agents in the following order: CyDTA > EDTA > NTA, none (Figures 2(a) and

2(b)). The activity of CoMo was increased ca. 25% by the addition of CyDTA, but remained

unchanged by NTA. Van Veen et al. have reported negative results for HDS of DBT on

NTA-modified CoMo/A1203 [6]. Reasons for this contradiction are not yet clear. HDS

activity for DBT was much more promoted in NiW catalyst than in CoMo catalyst by the

addition of the chelating agent. HDS activity of the NiW catalyst might be more susceptible

to the preparation method than that of the CoMo catalyst.

3.2. HGN Reactions

For HDS of DBT, it has often been said that HGN activity of the catalyst is

important. So we have examined HGN activities of the catalysts prepared with the chelating

agents in HGN of o-xylene as a test reaction. Reaction products were 1,2-

dimethylcyclohexane, some dimethylcyclohexanes, and m-, p-xylene. HGN activity of NiW

for o-xylene (the sum of yields of 1,2-dimethylcyclohexane and other cycloaliphatic

compounds) was improved three-fold by the addition of CyDTA (Figure 3).

CyDTA-CoMo

CoMo

CyDTA-NiW

NiW

0 5 10 15 20 HGN Yield/metal/%/mmol

Figure 3. HGN activity of CyDTA-modified CoMo- and NiW/A1203 for o-xylene. Reaction conditions �9 573 K, 5.1 MPa.

5-methyltetralin 1-methyltetralin

EDTA-NiW

NTA-NiW

CyDTA-NiW

NiW

0 0.1 0.2 0.3 0.4 0.5 HGN Yield/metal/%/gmol

Figure 4. HGN activities of chelating agent- modified NiW/A120 3 for 1-methylnaphthalene. Reaction conditions" 603 K, 5.1 MPa.

1-MN was also hydrogenated to examine the effect of the chelating agents on the

HGN activity of the catalysts. 1-methyltetralin and 5-methyltetralin were mainly produced.

The sum of the yields of 1-methyltetralin and 5-methyltetralin was used as an index of HGN

activity. HGN activity of NiW was improved about 40% by the addition of CyDTA (Figure

4). The activity of each component system, i.e., Co-, Mo-, Ni- or W/A1203 was not

165

improved by the addition of the chelating agents. The chelating agents are considered to

improve synergy between Co and Mo or between Ni and W, leading to induce the intrinsic

promoting effects of Co for CoMo and Ni for NiW. The chelating agents may have a role to

improve the formation of specific active phase.

3.3. Complex Formation Constants An important development has recently been made in research of the Co-Mo-S

structure by van Veen et al. [ 1 ]. In the study expecting to understand "a real support effect",

they found that Co-Mo-S phase was selectively formed by using NTA in impregnating

solution by means of M6ssbauer emission spectroscopy. The preparation method, originally

invented by a researcher of Shell for SiO 2 supported hydrotreatment catalysts [7], has been

applied to the study using the extended X-ray absorption fine structure (EXAFS)

measurement, and it was suggested that Co was located at the edge site o f M o S 2 structure and

was coordinated with five or six sulfur atoms [8, 9].

Table 1

Complex formation constants of literature 10)

Co Ni Mo W

EDTA 16.31 18.62 18.76 19.67

NTA 10.38 11.54 18.60 19.03

CyDTA 18.92 19.40 - * - *

* We confirmed with NMR that no complexes were formed.

Van Veen reported that the improving effect of NTA on the activity of C o M o / S i O 2

(or CoMo/active-C) is due to its ability to form complex with Mo and Co at the same time. In

the present work, we have examined van Veen's proposition under different conditions. Table

1 shows literature values of complex formation constant of the chelating agents with the

related metal ions. NTA forms complexes with all the metal ions. CyDTA, the most

effective chelating agent, however, forms complexes with Co or Ni ions, but not with Mo or

W ions. Considering CyDTA was more effective than NTA, the ability to make a complex of

chelating agent and promoter (e.g., Co, Ni) is rather important to improve the synergy

between promoter and Mo or W ions.

3.4. Sulfiding Temperature Dependence In order to investigate the mechanism in which chelating agents improve synergy

between Mo and Co or between W and Ni, effects of sulfiding temperature on the catalyst

166

activity were examined.

Figure 5(a) shows the sulfiding temperature dependence of HDS activities of

CyDTA-CoMo and CoMo. Figure 5(b) shows the dependence of HGN activities of CyDTA-

NiW and NiW. In these Figures, the following two points are noticed with respect to the

improving effect of CyDTA. First, the activity of the CyDTA-modified catalysts depends

more strongly on sulfiding temperature than that of the unmodified catalysts. Secondly, the

order of the catalytic activities of the CyDTA-modified and unmodified catalysts are inverted

at a lower sulfiding temperature. At higher sulfiding temperatures, the CyDTA modified

catalysts show higher activities in HDS and HGN reactions than the unmodified catalysts. At

a lower sulfiding temperature, however, the activities of CyDTA-modified catalysts are lower

than those of the unmodified catalysts. That is, CyDTA exhibits the improving effect at

higher sulfiding temperatures, while working as an inhibitor at lower temperatures.

1 . 2 �9

~0.8 . , . a

~0.6 E

0.4 . , . .~

~0.2

m

-8 _$

CyDTA-CoMo O

�9 0 o

(a) CyDTA-NiWO

~20

-~15 -

~10 -

N5 Z

:=0 450

o 0

(b)

CoMo -i~ NiW

I I Q I I 0 450 550 650 750 550 650 750

Sulfiding temperature/K Sulfiding temperature/K

Figure 5. The sulfiding temperature dependence of catalytic activities of modified and unmodified catalysts. (a) Benzothiophene HDS Reaction : 543 K, 5.1 MPa. (b) o-Xylene HGN Reaction : 573 K, 5.1 MPa.

3.5. Role of Chelat ing Agents

The sulfiding temperature dependency of the effect of CyDTA is considered to be

caused by the strong interaction between CyDTA and Co (or Ni) ion as shown in Table 1. At

relatively lower sulfiding temperatures, CyDTA interacts strongly with Co (or Ni) ion,

resulting in inhibiting Co (or Ni) ion from interacting with Mo (or W) or A1203. At higher

sulfiding temperatures, however, the complex between CyDTA and Co (or Ni) ion

decomposes, resulting in the interaction between Co and Mo (or Ni and W). Our concept is

depicted in Figure 6.

In the preceding paper [11], we reported the effects of various pretreatments

(including sulfiding and reducing) on the activity and structure of CoMo/A1203. In the report,

the intrinsic high activity of CoMo was found to be induced by proper sulfiding pretreatment.

From the results of activity test and Mo K-edge EXAFS analysis, it was concluded that the

167

appearance of the intrinsic promoting effect of Co was closely connected with the formation

of MoSz-like structure by proper sulfiding pretreatment. The pre-formation of the MoSz-like

structure was necessary to induce the intrinsic promoting effect of Co on the activity of

Mo/AI203. That is, the intrinsic promoting effect of Co was induced on the surface of the

MoS2-1ike structure.

Figure 6. Scheme of the fornation of active sites in CyDTA-CoMo/A1203.

The present results support our previous proposition.

As shown in Figure 6, without chelating agents, Co ions can react with A1203 or be

sulfided to form less active COA1204 or Co9S8, respectively. Co ions also interact freely with

Mo to interfere with the formation of the MoSz-like structure.

In the presence of a chelating agent such as CyDTA, Co ions are so strongly complexed

with the chelating agent that sulfiding of the Co ions or interaction of the Co ions with Mo or

A1203 will be inhibited, while Mo ions are sulfided to form the MoSz-like structure.

At lower sulfiding temperatures, the chelated Co ions remain undecomposed.

Accordingly, highly active sites resulted from interaction between Co and the MoS~-like

168

structure will not be formed. A chelating agent works as an inhibitor.

At higher sulfiding temperatures, the chelated Co ions decompose with time. In other

words, the chelated Co ions decompose after Mo ions are sulfided to some extent to form the

MoSz-like structure. The Co ions thus formed, eventually interacts with the MoS2-1ike

structure to form highly active sites.

Thus, the role of chelating agents is considered to adjust timing when Co ions interact

with the MoS2-1ike structure, leading to induce the intrinsic synergy between Mo and Co.

4. A C K N O W L E D G E M E N T

A part of this work has been carried out as a research project of the Japan Petroleum

Institute commissioned by the Petroleum Energy Center with the subsidy of the Ministry of

International Trade and Industry.

5. R E F E R E N C E S

7

8

R.

9

J. A. R. van Veen, E. Gerkema, A. M. van der Kraan, and A. Knoester, J. Chem.

Soc.,Chem. Commun., (1987) 1684.

T. Shimizu, S. Kasahara, T. Kiyohara, K. Kawahara, and M. Yamada, Sekiyu

Gakkaishi, 38 (1995) 384.

K. Hiroshima, T. Mochizuki, T. Honma, T. Shimizu, and M. Yamada, Appl. Surf.

Sci., 121/122 (1997) 433.

M. Yamada, A. Saito, T. Wakatsuki, T. Obara, J.-W. Yan, and A. Amano, Sekiyu

Gakkaishi, 30 (1987) 412.

M. Yamada, Y.-L. Shi, T. Obara, and K. Sakaguchi, Sekiyu Gakkaishi, 33 (1990)

227.

J. A. R. van Veen, H. A. Colijn, P. A. J. M. Hendriks, and A.J. van Welsenes, Fuel

Processing Technology, 35 (1993) 137.

M. S. Thompson, Eur. Pat. Appl., EP 181035 (1986).

S. M. A. M. Bouwens, J. A. R. van Veen, D. C. Koningsberger, V. H. J. de Beer, and

Prins, J. Phys. Chem., 95 (1991) 123.

S. M. A. M. Bouwens, F. B. M. van Zon, M. P. van Dijk, A. M. van der Kraan, V.

H. J. de Beer, J. A. R. van Veen, and D. C. Koningsberger, J. Catal., 146 (1994) 375

10

11

L.G.Silen and A.E.Martell, Stability constants of metal-complexes vol.2, Chemical

Society, London, 1964.

S. Kasahara, Y. Udagawa, and M. Yamada, Appl.Catal.B. Environmental, 12, (1997)

225.