Journal of the Japan Petroleum Institute, 50, (1), 44-52 (2007)

44

[Regular Paper]

Effects of Small Amounts of Nitrogen Compounds in Feedstock on

Performance of Hydrocracking Catalyst
Manabu Kobayashi, Seiji Togawa, and Katsuaki Ishida
Petroleum Refining Research & Technology Center, Japan Energy Corp., 3-17-35 Niizo-Minami, Toda, Saitama 335-8502, JAPAN
(Received February 22, 2006)

Poisoning by very small amounts of nitrogen compounds was investigated for the catalysts for two types of hydrocracking processes.Nitrogen-free vacuum gas oil was used as feedstock with carbazole and tributyl amine as the
model nitrogen compounds.Carbazole is abundant in hydrotreated oil as carbazole is difficult to remove.
Tributyl amine is easily converted to ammonia in the reactor.
Vacuum gas oils containing various amounts of carbazole were passed over the main hydrocracking catalyst of
the two-stage hydrocracking process.Strong effects on both hydrocracking activity and middle distillate selectivity were observed, especially if nitrogen content was less than 2 wtppm.The nitrogen poisoning effect was
stronger on the catalyst with higher activity than on the catalyst with lower activity.
Poisoning of the main hydrocracking catalyst of the single-stage hydrocracking process, which is affected by
both organic nitrogen compounds and ammonia, was evaluated with several ratios of combinations of carbazole
and tributyl amine, maintaining total nitrogen concentration at 300 wtppm.Increased carbazole ratio lowered
the hydrocracking activity and increased the middle distillate selectivity especially if the concentration of nitrogen
as carbazole was less than 20 wtppm.These experimental results indicate that there is an optimum range of nitrogen concentration in the effluent to the main hydrocracking reactor to maximize middle distillate selectivity
without severely affecting the hydrocracking activity.
Keywords
Vacuum gas oil, Hydrocracking, Poisoning, Ammonia, Carbazole

classified into two main types6),7) (Fig. 1).The twostage hydrocracking process consists of two reactors
The hydrocracking (HC) process for heavy oils has
with the gas/liquid separator between the reactors.
become more important recently because of the inThe first reactor operates as a hydrotreater for hydrocreased demand for cleaner middle distillates.The
genation of unsaturated molecules, hydrodesulfurmost common catalysts for the HC process are biization, and hydrodenitrogenation, as well as partial
functional with acidic and hydrogenation functions1). hydrocracking, and the second reactor completes the
The acidic function usually depends on crystalline or
hydrocracking process.Most organic nitrogen comuncrystalline alumino_silicates, but these materials have
pounds are converted to ammonia in the first reactor,
strong affinity with nitrogen compounds in the feedand generated ammonia is removed at the gas/liquid
stock.Therefore, the nitrogen compounds have a
separator.Therefore, the main HC catalyst in the seclarge effect on the performance of the hydrocracking
ond reactor is exposed only to the remaining organic
catalysts2), and the processing of nitrogen compounds is
nitrogen compounds7), typically at levels of less than
3)5)
a key technology for the HC process
.However,
100 wtppm.
nitrogen compounds also have a positive effect that inThe single-stage hydrocracking process converts
creases the selectivity for middle distillate.Both posmost of the organic nitrogen compounds in the feeditive and negative effects are caused by adsorption of
stock to ammonia in the first reactor, but the generated
the nitrogen compounds on the strong acid sites of HC
ammonia is not removed and remains in the effluent to
catalysts, which have hydrocracking activity even for
the second reactor, as no gas separator is present beless reactive middle distillate.
tween the first and the second reactors.The total niThe process flow of hydrocracking processes can be
trogen content in the effluent from the first reactor is
determined by the fresh/recycle feed ratio regardless of

the denitrogenation level in the first reactor.

To whom correspondence should be addressed.

Many studies to improve the efficiency of HC proE-mail: [email protected]

1.Introduction

J. Jpn. Petrol. Inst.,

Vol. 50,

No. 1, 2007

45

Fig. 1Simplified Flow Diagram of Hydrocracking Units

cess have investigated the nitrogen poisoning effects on

the hydrocracking catalyst3)5) as well as catalytic
cracking (Fluid Catalytic Cracking)8),9), hydrodesulfurization10),11), and hydrogenation12).However, very little quantitative information on poisoning by very small
amounts of nitrogen compounds is available, especially
for the single-stage HC process.
Investigation of the hydrotreated coker gas oil,
derived from Athabasca bitumen, concluded that the
nitrogen compounds were dominated by a series of
alkyl carbazoles13).Analysis of hydrotreated coker
gas oils from a delayed coker obtained similar results14).
These findings suggest that carbazole and its derivatives
are abundant in the effluent from the first hydrotreatJ. Jpn. Petrol. Inst.,

ment reactor of HC units.

We performed two series of experiments to investigate the poisoning effect of very small amounts of nitrogen compounds on the main HC catalysts of the twostage type and single-stage type HC processes with
actual feedstock combined with carbazole and tributyl
amine (TBA) as model nitrogen compounds.The
evaluation test for the two-stage hydrocracking process
was performed with the feedstock containing various
amounts of carbazole.In the single-stage hydrocracking process, the ammonia generated in the first hydrotreatment reactor still remains in the effluent to the
main reactor.TBA, which is immediately converted
to ammonia in the reactor and is more convenient to

Vol. 50,

No. 1, 2007

46

NF-VGO are shown in Table 2.

Metal loadings and other catalyst properties that affect the HC performance are shown in Table 3.
2. 2.Hydrocracking Reaction
The catalyst (100 ml) was filled in a fixed bed flow
reactor with -alumina (100 ml) as diluent, and used
for the hydrocracking reaction.The hydrocracking
conditions were as follows.
Evaluation test for two-stage HC: Catalyst; Cat-A, B,
C, H2 pressure; 14.7 MPa, H2/oil ratio; 800 nl/l, operation temperature; 360, 370, 380, 390 C (NF-VGO),
370, 380, 390 C (N-VGOs), liquid hourly space velocity
(LHSV); 1.36 h1.
Evaluation test for single-stage HC:
HC Catalyst; Cat-D,
l l, operation
l/
H2 pressure; 14.7 MPa, H2/oil ratio; 800 nl/
temperature; 330, 340, 350, 360 C (NF-VGO), 390, 400,
410 C (N-VGO-5), 380, 390, 400 C (N-VGO-6 to 8),
390, 400 C (N-VGO-9), LHSV; 1.60 h1.
Before providing the feedstocks, catalysts were sulfided with nitrogen-free gas oil containing 1 vol% carbon disulfide.Liquid samples were obtained after oil/
gas separation.Concentrations of gas products were
analyzed by GC (gas chromatography).
2. 3.Analysis of Product Oils
The evaluation tests for the two-stage HC measured
HC activity and the yield of each fraction at 60 wt%
conversion of the 293 C+ fraction.The evaluation test
for the single-stage HC measured HC activity and the
yield of each fraction at 70 wt% conversion of the
360 C+ fraction.These performance parameters were
based on the actual operation conditions of the specific
HC units.The targeted fraction of the two-stage HC

handle than ammonia, was added to the feedstock with

carbazole to evaluate the effect of hydrodenitrogenation
level in the first hydrotreatment reactor on HC in the
second reactor.
2.Experimental
2. 1.Feedstock and Model Nitrogen Compounds
Two series of experiments were performed to quantitatively evaluate the nitrogen poisoning phenomena on
the main HC catalysts in both the two-stage and the
single-stage HC processes.
For the two-stage HC, poisoning behaviors on three
amorphous silica_alumina catalysts (Cat-A, Cat-B, and
Cat-C) with different HC activities were evaluated with
vacuum gas oils containing various amounts of carbazole (N-VGO-1 to 4) and a nitrogen-free vacuum gas
oil (NF-VGO), which was a recycle oil from an actual
hydrocracking unit.The properties of the NF-VGO
are shown in Table 1.
For the single-stage HC, the poisoning effect on a
catalyst (Cat-D) consisting of silica _ alumina and a
small amount of USY zeolite was evaluated with vacuum gas oils containing carbazole and TBA at several
ratios (N-VGO-5 to 9), but maintaining the total nitrogen content at 300 wtppm, which approximately corresponds to the expected nitrogen content in the effluent
to the second reactor, which can be calculated from the
fresh and recycle feedstock flow rate.N-VGO-5 to 9
were prepared by the addition of carbazole and TBA to
NF-VGO.The nitrogen contents in the N-VGOs and
Table 1Properties of Nitrogen-free Vacuum Gas Oil (NF-VGO)

Table 2Nitrogen Contents in Feedstocks

NF-VGO
Density (15 C)
Sulfur
Nitrogen
Distillation
IBP
T10
T30
T50
T70
T90
T95
FBP

[g/ml]
[wt%]
[wtppm]

0.8249
0.001
1

(ASTM D 2887)
[ C]
[ C]
[ C]
[ C]
[ C]
[ C]
[ C]
[ C]

263
335
375
398
424
470
501
554

Nitrogen as carbazole
[wtppm]

Nitrogen as TBA
[wtppm]

0
0.5
2
8
16
0
10
20
50
75

0
0
0
0
0
300
290
280
250
225

NF-VGO
N-VGO-1
N-VGO-2
N-VGO-3
N-VGO-4
N-VGO-5
N-VGO-6
N-VGO-7
N-VGO-8
N-VGO-9

Table 3Properties of Hydrocracking Catalysts

W
[wt%]

Ni
[wt%]

Silica_alumina weight
per volume
[g/cm3]

USY weight
per volume
[g/cm3]

Metal (WNi) / Silica_alumina

ratio
[wt%/wt%]

Cat-A
Cat-B
Cat-C

11
22
10

1
2
1

0.485
0.480
0.607

0.177
0.430
0.128

Cat-D

22

0.468

0.02

0.450

J. Jpn. Petrol. Inst.,

Vol. 50,

No. 1, 2007

47
Table 4293 C+ Conversion of NF-VGO and N-VGOs
293 C+ conversion

Sample number
[]

Time
[h]

Feed
[]

Nitrogen in
feedstock
[wtppm]

Operation
temperature
[ C]

Cat-A
[wt%]

Cat-B
[wt%]

Cat-C
[wt%]

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22

54
67
78
91
102
115
138
151
164
188
201
214
237
250
263
287
300
313
337
349
361
373

NF-VGO
NF-VGO
NF-VGO
NF-VGO
NF-VGO
NF-VGO
N-VGO-1
N-VGO-1
N-VGO-1
N-VGO-2
N-VGO-2
N-VGO-2
N-VGO-3
N-VGO-3
N-VGO-3
N-VGO-4
N-VGO-4
N-VGO-4
NF-VGO
NF-VGO
NF-VGO
NF-VGO

0
0
0
0
0
0
0.5
0.5
0.5
2.0
2.0
2.0
8.0
8.0
8.0
16.0
16.0
16.0
0
0
0
0

390
380
370
360
390
380
390
380
370
390
380
370
390
380
370
390
380
370
390
390
390
390

82.4
66.8
46.0
27.1
79.0
62.7
76.4
54.2
34.9
70.8
46.8
28.0
56.4
30.8
15.8
30.9
13.7
6.7
50.2
57.8
60.9

71.8
49.8
25.4
14.6
62.3
38.2
55.2
34.4
20.4
49.8

16.6
39.8
22.8
11.4
23.9
12.0
6.0
27.4
33.4
34.4
38.7

93.8
82.8
61.7
38.5
94.7
78.3
91.8
73.5
46.9
86.3
62.5
36.8
72.1
40.8
21.4
43.1
20.2
10.2
72.5

H2 pressure; 14.7 MPa, H2/oil ratio; 800 nl/

l l, liquid hourly space velocity (LHSV); 1.36 h1.
l/

unit discussed in the present paper is jet fuel (kerosene) fraction and the cut point was set at 293 C whereas
the single-stage HC is targeted at middle distillates (kerosene and gas oil) and the cut point was set at 360 C.
Distillation properties of liquid samples obtained
from hydrocracking reaction were evaluated with the
ASTM D 2887 method to determine the 293 C+ conversion defined as Eq. (1).
C0C1) / C0100 (1)
293 C+ conversion (wt%)(C
where C0 (wt%)293 C+ fraction of feedstock (2)
C1 (wt%)293 C+ fraction of product oil (3)
The 360 C+ conversion was calculated with a similar
equation.
Yields of each fraction were also derived from the
NF-VGO, N-VGO-1 (0.5 wtppm), N-VGO-2 (2 wtppm),
result of ASTM D 2887 analysis.Boiling ranges were
N-VGO-3 (8 wtppm), N-VGO-4 (16 wtppm).
as follows: naphtha (boiling point: C5-127 C), kerosene
Note: In Fig. 2, 293 C+ conversions at 390 C operation with Cat-A
(127-293 C), gas oil (293-360 C), bottom oil (360 C+). were extracted from Table 4.
The gas fraction consisted of methane, ethane, propane,
Fig. 2Change in Hydrocracking Activity for Vacuum Gas Oils
butane, and isobutane.
with and without Nitrogen Compounds
Required operation temperatures for 60 wt% and
70 wt% conversion (ROT60 and ROT70) were calculated from Arrhenius plots assuming that the HC reac3.Results and Discussion
tion is second order.Approximate yields of each fraction at 60 wt% or 70 wt% conversion were calculated
3. 1.Effects of Nitrogen Compounds on Two-stage
from the correlation curve of conversion and yield of
Hydrocracking Catalysts
each fraction.
The results of the evaluation tests for poisoning of
two-stage hydrocracking catalysts and 293 C+ conversion are summarized in Table 4.Figure 2 compares
the 293 C + conversion at 390 C operation on Cat-A,
and shows that HC activity was almost stable with NFJ. Jpn. Petrol. Inst.,

Vol. 50,

No. 1, 2007

48

Cat-A, Cat-B, Cat-C.

Cat-A, Cat-B, Cat-C.

Fig. 3Hydrocracking Activity of the Present Catalysts for

Nitrogen-free Vacuum Gas Oil (NF-VGO)

Fig. 4Required Operation Temperature for 60 wt% Conversion as

a Function of Nitrogen Content in VGO Feedstock

VGO feed, declined according to the increase of nitrogen content, and after 315 h recovery of HC activity
with NF-VGO was observed.NF-VGO was supplied
for almost 50 h and HC activity recovered gradually,
but 50 h was apparently still on the path of recovery.
This result demonstrated the reversibility of nitrogen
poisoning of HC activity and that the changes in HC activity observed during the present study were caused by
nitrogen poisoning.
The HC activities of Cat-A, Cat-B and Cat-C for
nitrogen free vacuum gas oil (NF-VGO) are compared
in Fig. 3.The highest HC activity was observed for
Cat-C and the lowest for Cat-B.The highest HC activity of Cat-C can be interpreted as the highest concentration of silica _ alumina and the lowest metal/
silica_alumina ratio.The lowest HC activity of Cat-B
can be explained by the very high metal/silica_alumina
ratio.
Figure 4 shows the effect of nitrogen in the feedstock (N-VGO-1 to 4) on the activity of catalysts
(Cat-A, Cat-B and Cat-C).Even very small amounts
of nitrogen compounds at less than 10 wtppm greatly
lowered the HC activity of the catalysts.The decrease
of HC activity of all catalysts was rapid at nitrogen content between 0 and 2 wtppm, and ROT60 increased
gradually with nitrogen content of more than 2 wtppm.
Figure 5 shows the relative rate constant at 370 C as
expressed by the division of the rate constants for
N-VGOs (kkN-VGO) by that of NF-VGO (kkNF-VGO).The
effect of nitrogen poisoning is stronger for the hydrocracking catalyst with higher activity.Figure 5 also
indicates the logarithmic relationship between the rate
constant of the HC reaction and nitrogen content.
Figures 6 to 10 show the effect of nitrogen on the
yields of gas, naphtha, kerosene, gas oil, and vacuum
gas oil (VGO) at 60 wt% conversion, respectively.
The gas yields were almost constant regardless of the
J. Jpn. Petrol. Inst.,

Cat-A, Cat-B, Cat-C.

Fig. 5Effects of Nitrogen Content in VGO Feedstock on Relative
Rate Constant of the Hydrocracking Reaction at 370 C (second order)

employed catalyst or the nitrogen content, but the yields

of liquid fractions were affected by nitrogen content of
less than 2 wtppm.
The strong acid sites that crack small or medium size
molecules decrease the middle distillate yield by secondary cracking.Therefore, the increases in kerosene
and gas oil yields with nitrogen contents of less than
2 wtppm can be attributed to poisoning of the strong
acid sites.On the other hand, increase of nitrogen
content to more than 2 wtppm did not improve the yield
balance, whereas the HC activity was strongly affected
by nitrogen content.This result indicates that most of
the strong acid sites are deactivated by nitrogen of
2 wtppm.The acid sites that are deactivated by nitrogen of more than 2 wtppm can be regarded as weak
acid sites that hardly convert the middle distillates (ker-

Vol. 50,

No. 1, 2007

49

Cat-A, Cat-B, Cat-C.

Cat-A, Cat-B, Cat-C.

Fig. 6Relationship between Nitrogen Content in VGO Feedstock

and Yield of Gas Fraction (C1-C4) at 60 wt% Conversion

Fig. 8Relationship between Nitrogen Content in VGO Feedstock

and Yield of Kerosene Fraction at 60 wt% Conversion

Cat-A, Cat-B, Cat-C.

Cat-A, Cat-B, Cat-C.

Fig. 7Relationship between Nitrogen Content in VGO Feedstock

and Yield of Naphtha Fraction at 60 wt% Conversion

Fig. 9Relationship between Nitrogen Content in VGO Feedstock

and Yield of Gas Oil Fraction at 60 wt% Conversion

osene and gas oil) to the lighter fraction.

3. 2.Effects of Nitrogen Compounds on Singlestage Hydrocracking Catalysts
The correlation between the nitrogen content of carbazole in the feedstock and ROT70 is shown in Fig. 11.
ROT70 increased with increased carbazole ratio in the
feedstock, indicating decreased HC activity.With the
nitrogen content of carbazole at less than 20 wtppm, the
carbazole ratio affects HC activity more than with the
nitrogen content of carbazole at more than 20 wtppm.
ROT70 increased by 0.27 C/wtppm and 0.04 C/wtppm
with nitrogen contents of carbazole of less than
20 wtppm and more than 20 wtppm, respectively.
The contents of organic nitrogen compounds in product oils, carbazole or TBA in the present study, were
less than the detectable limit of 1 wtppm with all samples, which indicates that the stronger poisoning effect

of carbazole than that of TBA on HC activity only persists until the carbazole is converted to ammonia.The
difference in the strength of poisoning between carbazole and TBA suggests that the denitrogenation reaction of carbazole does not proceed immediately as for
TBA, so carbazole has a significantly stronger poisoning effect than TBA.
The activation energies of the HC reaction, assuming
that the HC reaction is apparent second order, are also
strongly affected by the carbazole/TBA ratio (Fig. 12).
These trends suggest that the affinity of carbazole with
the active sites on the catalyst is stronger than that of
ammonia.Various nitrogen compounds were tested in
the poisoning of thiophene hydrodesulfurization on
NiMo/Al2O3 catalyst10).The calculated adsorption
equilibrium constants of ammonia and carbazole were
0.048 kPa 1 and 0.51 kPa 1, respectively.The size/

J. Jpn. Petrol. Inst.,

Vol. 50,

No. 1, 2007

50

Cat-A, Cat-B, Cat-C.

Fig. 10Relationship between Nitrogen Content in VGO Feedstock
and Yield of VGO Fraction at 60 wt% Conversion

Note: In Figs. 11-17, nitrogen contents originated from carbazole

were varied maintaining total nitrogen contents constant at
300 wtppm balanced by tributyl amine.

Fig. 12Relationship between Content of Nitrogen as Carbazole in

VGO Feedstock and Activation Energy (second order)

Fig. 13Relationship between Content of Nitrogen as Carbazole

in VGO Feedstock and Gas Yield (C1-C4) at 70 wt%
Conversion

Fig. 11Required Operation Temperature for 70 wt% Conversion

as a Function of Content of Nitrogen as Carbazole in the
VGO Feedstock

mass of the poison molecule affects the strength of poisoning as well as basicity8).These previous findings
also support the stronger effects of carbazole than ammonia on the HC reaction.
The correlation between the carbazole ratio and the
yields of each fraction at 70 wt% conversion are shown
in Figs. 13-16.If the nitrogen content of carbazole
was less than 20 wtppm, with increased carbazole ratio,
gas yield increased, naphtha yield decreased, gas oil
yield increased, and kerosene yield was almost constant.The increase of gas oil yield continued up to approximately 3 wt%, which corresponds to the decrease
of naphtha yield.On the other hand, if the nitrogen
content as carbazole was more than 20 wtppm, the yield
balance of each fraction was almost constant regardless
J. Jpn. Petrol. Inst.,

Fig. 14Relationship between Content of Nitrogen as Carbazole in

VGO Feedstock and Naphtha Yield at 70 wt% Conversion

Vol. 50,

No. 1, 2007

51

Fig. 15Relationship between Content of Nitrogen as Carbazole in

VGO Feedstock and Kerosene Yield at 70 wt% Conversion

ROT70, C, Middle distillate (MD) yield, wt%.

Fig. 17Optimum Nitrogen Content for High Middle Distillate
Yield

Figs. 8 and 9.The difference can be attributed to the

number of strong acid sites on the catalysts, as the catalyst employed for the evaluation of single-stage HC catalyst in the present study contains zeolite whereas those
for two-stage HC are amorphous catalysts.
4.Conclusion

Fig. 16Relationship between Content of Nitrogen as Carbazole in

VGO Feedstock and Gas Oil Yield at 70 wt% Conversion

of the carbazole ratio.The constant yield of kerosene

is attributed to the low reactivity of the molecules in the
kerosene fraction, and the increase of gas fraction with
increased carbazole ratio is due to the increased
ROT70.
These results suggest that there is a suitable range of
the denitrogenation target for the hydrotreatment step in
the first reactor.Figure 17 indicates that 10 wtppm
of nitrogen as carbazole significantly increases the yield
of middle distillate, whereas more than 10 wtppm nitrogen as carbazole results in almost constant middle distillate yield regardless of the nitrogen content of carbazole.On the other hand, HC activity declines
according to the increase of the nitrogen content as
carbazole until 80 wtppm.From these results, we conclude that 10 to 15 wtppm of nitrogen could significantly
increase the middle distillate selectivity without severely
affecting the HC activity.On the other hand, in the
two-stage HC, 2 wtppm of nitrogen is adequate for improvement of the middle distillate yield, as shown in
J. Jpn. Petrol. Inst.,

Poisoning by very low concentrations of nitrogen

compounds was investigated for the catalysts used for
the two types of hydrocracking processes.Nitrogenfree vacuum gas oil was used as the feedstock with
model nitrogen compounds, carbazole and tributyl
amine.
Strong effects both on hydrocracking activity and
middle distillate selectivity were observed for two-stage
hydrocracking, especially if the nitrogen content was
less than 2 wtppm.Increased carbazole ratio lowered
the hydrocracking activity and increased the middle distillate yield for single-stage hydrocracking.Those effects were especially strong if the nitrogen concentration derived from carbazole was less than 20 wtppm.
These experimental results indicate the suitable range
of nitrogen content in the effluent to the main hydrocracking reactor to maximize the middle distillate selectivity without unacceptable decrease of HC activity.
The present study suggests that around 2 wtppm and
10-15 wtppm are suitable for two-stage and single-stage
hydrocracking, respectively.
References
1) Ward, J. W., Fuel Process. Technol., 35, 55 (1993).
2) Minderhoud, J. K., van Veen, J. A. R., Fuel Process. Technol.,
35, 87 (1993).
3) Jacobs, P. A., Heylen, C. F., J. Catal., 34, 267 (1974).
4) Dufresne, P., Quesada, A., Mignard, S., Catalysts in
Petroleum Refining 1989, eds. by Trimm, L. et al., Elsevier,
Vol. 50,

No. 1, 2007

52
Amsterdam (1990), p. 301-315.
5) Nat, P. J., Erdoel Kohle, Erdgas, Petrochem. Brennst.-Chem.,
42, (11), 447 (1989).
6) Scherzere, J., Gruia, A. J., Hydrocracking Science and
Technology,ed. by Heinmann, H., Marcel Dekker, New York
(1996), p.174-199.
7) Scott, J. W., Bridge, A. G., Adv. Chem. Ser., 103, 113 (1971).
8) Ho, T. C., Katritzky, A. R., Cato, S. J., Ind. Eng. Chem. Res.,
31, 1589 (1992).

9) Fu, C., Schaffer, A. M., Ind. Eng. Chem., Prod. Res. Dev., 24,
68 (1985).
10) Vopa, V. L., Satterfield, C. N., J. Catal., 110, 375 (1998).
11) Nagai, M., Sato, T., Aiba, A., J. Catal., 97, 52 (1986).
12) Lee, C. M., Satterfield, C. N., Energy & Fuels, 6, 315 (1992).
13) Jokuty, P. L., Gray, M. R., Energy & Fuels, 5, 791 (1991).
14) Ignatiadis, I., Kuroki, M., Arpino, P. J., J. Chromatogr., 366,
251 (1986).

335-8502 3-17-35

300 wtppm

20 wtppm

2 wtppm

J. Jpn. Petrol. Inst.,

Vol. 50,

No. 1, 2007