Relationship between surface acidity and liquefaction yield of hydrotreating catalysts

Relationship between surface acidity and liquefaction yield of hydrotreating catalysts

Relationship liquefaction Wang L. Yoon, between surface acidity and yield of hydrotreating catalysts In C. Lee and Won K. Lee* Fossil Fuel// Labo...

668KB Sizes 2 Downloads 25 Views

Relationship liquefaction Wang

L. Yoon,

between surface acidity and yield of hydrotreating catalysts

In C. Lee and Won

K. Lee*

Fossil Fuel// Laboratory, Korea Institute of Energy and Resources, Taejeon, 30 7 -343, Korea *Department of Chemical Engineering, Korea Advanced Institute of Science and Technology, Seoul, 136-797, Korea (Received 12 March 1990; revised 21 August 1990)

Alaskan Usibelli coal was liquefied in naphthalene or tetralin under 13.78 MPa hydrogen pressure using six commercial hydrotreating catalysts of varying physical properties: Amocat lA, HDN-30, HDN-60, Shell 317, Shell 324 and Amocat 1C. Also, stepwise thermal desorption (STD) of pyridine was carried out under chromatographic conditions to measure the amount of acid and the strength distribution on each catalyst, to relate the catalyst activity with the solid surface acidity. Comparing the results ofcoal liquefaction

efficiency with the acidity measurements of six hydrotreating catalysts, it appears that weak acidity, which can provide reversible adsorption sites for reactants and products of catalytic reactions, is very important to improve liquefaction yield. The results showed that with the commercial hydrotreating catalyst, and having the same metals (Ni/Mo) on the surface, the catalytic activity for coal liquefaction with both the weak catalyst acidity and the pore volume in 3&200 A diameter pores. (Keywords:

liquefaction;

catalyst;

surface

Ltd.

well

properties)

To efficiently produce distillate products through direct coal liquefaction, catalysis is necessary to effect hydrogenation, reduction in molecular weight, and removal of heterocyclic compounds. Bifunctional catalysts with combined hydrogenation and cracking activities are the most suitable for coal liquefaction reactions’. Hydrogenation is necessary to stabilize the radical fragments liberated from the cleavage of connecting linkages in coal structure by thermal or catalytic action. It is thought that recent developmental trends in coal dissolution catalysts are directed towards cheap dispersed catalysts able to attain intimate contact with the coal. The principal research objective as regards supported catalysts now used in upgrading coal liquids should be development of a catalyst with high resistance to deactivation by coke and metal deposition. The recognition that different conditions are required for dissolution and upgrading has led to the development of a number of two-stage processes. Surface acidity which can provide cracking activity is a very important aspect of the characterization of catalysts for coal conversion and particularly for coal liquefaction and upgrading of coal liquids. If a catalyst has acid sites that are too weak it would not be active, but on the other hand, acid sites that are too strong can cause excessive cracking and condensation reactions resulting in carbon deposition. Therefore, the catalyst should be designed to have controlled acid site strength distribution as well as hydrogenation. Actually, determination of acid type, strength distribution and number is very important to understand the catalytic phenomena of solid catalysts and to select the best catalysts for given reactions. The acid strength and number of a solid surface can be determined experimentally by several methods’: 001~2361/91/0101074I6 c 1991 Butterworth~Heinemann

correlated

1. non-aqueous n-butylamine titration using indicators of varying pKa3 2. calorimetric determination by measurement of heat generated from the adsorption of a base on acid sites4 3. i.r. spectroscopic technique3 4. adsorption and desorption of basic gases such as ammonia, quinoline, pyridine or trimethylamine5. Infra monitoring of pyridine on solid surfaces has been shown to distinguish acidity type (Lewis or Brbnsted), which is otherwise difficult to do6-8. But this technique has a limitation in that it is difficult to quantify the accurate strength distribution and amount of acidity by measuring the absorption band and its area. Adsorption and desorption of base molecules by gas chromatography is valuable because the experiments can be carried out under conditions similar to reaction conditions’. This technique can provide precise data on the irreversible adsorption of a base at different temperatures, and also on the distribution of acid sites of different strength in terms of the amount of base desorbed in the different temperature steps. The present study investigated a simple and rapid gas chromatographic method, based on saturated adsorption and stepwise thermal desorption (STD) of pyridine under chromatographic conditions, used for measuring the acid strength distribution on solid catalyst at temperatures close to those employed in catalytic reactions. Also, batch liquefaction tests were carried out using Usibelli subbituminous coal from Alaska to relate liquefaction activity with catalyst acidity and catalyst physical properties, and to select a catalyst that gives the highest activity among the commercial catalysts to develop a new generation of effective well balanced catalysts.

FUEL,

1991,

Vol 70, January

107

Surface

acidity

and liquefaction

yield of hydrotreating

catalysts: W. L. Yoon et al. Liquefaction procedure Coal liquefaction for initial activity testing of each catalyst was carried out as described previously”. Namely, 3 g of coal was liquefied at 400°C for 30 min under reaction pressure of 13.78 MPa hydrogen using naphthalene or tetralin (6 g) as solvent, both with and without catalyst (0.1 g). After a given reaction time, and after venting the gaseous products, the residual products remaining in the minibomb were collected. Their solubilities in n-pentane benzene, and pyridine were then measured. The direct extraction technique described by Roberto and Cronauer12 was used. The gas yield was determined from the difference between the initial and residual sample weights. The products of liquefaction are characterized by the following equations:

EXPERIMENTAL Materials Coal from Usibelli in Alaska was used as a sample. Its analytical data are shown in Table 1. The sample was ground to pass 150 mesh (0.105 mm) screen and stored in a nitrogen atmosphere. The slurry vehicle solvents were tetralin and naphthalene and were used without further purification. The physical and chemical properties of six commercial catalysts are listed in Table 2. As shown in the data, one of the catalysts (Amocat 1A) is cobalt-molybdenum supported on alumina, whereas the others are nickel-molybdenum on alumina. Each catalyst was crushed to a mesh size of -60 (0.246 mm) to + 100 (0.147 mm), and then sulphided before use in coal liquefaction, following the procedure described elsewhere”.

gas = initial Table 1 Analysis

of Alaskan

Usibelh

subbituminous

oil = recovered insoluble

coal ~.___.

Result as recd. Proximate analysis Moisture Volatile matter Fixed carbon Ash

(%)

Elemental analysis Hydrogen Carbon Nitrogen Sulphur Oxygen (ind.) Ash

(X)

Heating

value (Btu lb- ‘)

16.50 40.65 34.26 8.59

N/A 48.68 41.03 10.29

N/A 54.26 45.14 N/A

5.86 52.87 0.72 0.19 31.77 8.59

4.82 63.31 0.86 0.23 20.49 10.29

5.31 70.57 0.96 0.26 22.84 N/A

10840

12083

GO.01 0.01 0.19

GO.01 0.01 0.22

GO.01 0.01 0.25

Properties

of commercial

catalysts

=

sample

(mf) - n-pentane

insoluble-benzene - pyridine

as pyridine

insoluble insoluble

insoluble

daf coal -residue daf coal

Catalyst acidity measurement The acidity of each catalyst was measured by stepwise thermal desorption (STD) of pyridine as a base. The apparatus consisted of a gas chromatograph with flame ionization detector and temperature programmed oven. The catalyst (0.30.57 g, particle size 0.2-0.3 mm), was placed in a 15 cm long stainless steel tube (o.d. 3 mm, i.d. 2 mm) and then connected in the oven of the g.c. as a normal column. To minimize the dead volume, one end of the column was directly connected to the detector and the other end to the injector through a 50 cm stainless steel tube (o.d. 1.5 mm, i.d. 0.7 mm), which acted as a preheater. Pyridine (ACS grade reagent) was used as an adsorbate without further purification. After connecting the adsorbent column to the g.c. oven the catalyst was calcined in situ to remove traces of moisture. Calcination was performed at 723 K for 24 h in a flow of helium gas at 25 ml min-’ in all experiments. The oven was heated

43.83 17.03 6.20 0.94 22.41 2.55 1.21 0.96

R,O

residual

= benzene insoluble

coal conversion

in ash (%)

MgO, Na,O

sample

= n-pentane

preasphaltene

9052

A&O, Fe,& TiO, CaO,

Table 2

maf

residual

residue = determined

Sulphur forms Sulphate Pyritic Organic Major elements SiO,

asphaltene

mf

sample-recovered

used in this study Catalyst

Properties

Amocat

Shape

Cylind.

Size (inch) Ni (wt%)

l/16” _

co (wt%)

2.5 9.8

MO (wt%) BET surface

area (m’ g- I)

Total pore volume

(ml g-l)

216

1A

HDN-30

HDN-60

Shell 317

Shell 324

Cylind.

Trilobe

Trilobe

Trilobe

Cylind.

l/16”

0.05”

0.05”

l/20”

l/32”

Amocat

2.3

1C

3.0

5.0

2.1

2.7

_ 10.4 182

20.5 146

22.0 145

11.6 195

0.70

0.71

0.34

0.29

0.63

13.2 147 0.38

Pore volume

(d c 30 A)

0.70

0.71

0.34

0.29

0.63

0.38

Pore volume

(30
0.47

0.41

0.29

0.27

0.44

0.35

Bimodal

Bimodal

Unimodal

Unimodal

Bimodal

Unimodal

Pore size distribution Compacted

108

bulk density

FUEL,

1991,

(lb ftt3)

Vol 70, January

41

42

52

54

36

54

Surface acidity and liquefaction at a rate of 10K min-’ until this temperature was reached. After calcination, the oven temperature was decreased to 473 K, the initial temperature of STD, and the pyridine was injected via a syringe until the adsorbent surface was fully saturated. The STD of pyridine irreversibly adsorbed at 473 K on a catalyst was carried out by raising the temperature of the catalyst from 473 to 673 K in a number of temperature steps, each of 50 K, and quantitatively measuring the pyridine desorbed in each step with a detector. After the maximum temperature of each step was attained, sufficient time (120 min) was allowed for the desorption of reversibly adsorbed pyridine. The amount of pyridine desorbed in a particular step gave the number of sites having an acid strength designated by the temperature limits of the desorption step. The amount of pyridine adsorbed irreversibly at the maximum temperature (673 K) chosen for STD was determined using the pulse technique described elsewhere13. RESULTS

AND DISCUSSION

Initial catalyst activity testing A series of catalyst activity screening tests was carried out with six commercial hydrotreating catalysts at the prescribed liquefaction reaction conditions. Reaction conditions and product distributions for each run are given in Table 3. Liquefaction efficiency was evaluated on the ability to convert coal to pyridine soluble (coal conversion) and pentane soluble (oil) matter. When naphthalene was used as a solvent, the catalytic activities, in terms of coal conversion and oil yield, were: Shell 1A > HDN1C > Shell 324 > Amocat 3 17 > Amocat 30 = HDN-60. With tetralin, the catalyst activities were: Shell 3 17 > Amocat 1C > Shell 324 > Amocat 1A = HDN30=HDN-60. In both cases, Shell 317 was found to give the highest coal liquefaction yields. Comparison of the physical properties of Amocat 1A (CO-MO) and Shell 317 (Ni-Mo) catalysts indicated that Ni-Mo catalyst showed higher activity for coal liquefaction than the CO-MO catalyst. Both catalysts have similar physical properties, apart from a difference in the

Table 3

Summary

of reaction

conditions

and product

metal component acting as a chemical promoter. It is likely that the higher activity of Shell 317 is caused by the use of nickel instead of cobalt. This implies that nickel promoter is more effective for the hydrogenative stabilization of free radical fragments than cobalt14. Of the live Ni-Mo catalysts used in this study, Shell 317 gave the highest activity. Although HDN-30 and HDN-60 had higher metal loadings than other catalysts they gave the lowest liquefaction yields. Catalyst performance thus did not appear to depend on metal content. The relatively poor activities of both catalysts could be ascribed to the lower surface area of these catalysts. As shown in Table 2, it is thought that the only physical property that is significant is surface area, as this can provide uniform metal dispersion and can thus increase the specific activity of the catalyst. The catalysts used in this study were all hydrotreating type (Co(or Ni)-Mo supported on Al,O,), to promote hydrogenolysis, hydrogenation and hydrocracking. If a hydrocracking catalyst such as Ni-Mo supported on silica-alumina were tested, it is expected that amounts of gas formed would be much higher for the hydrotreating catalyst. This because a co-gelled mixture of silica and alumina containing mainly silica (about l&12% alumina) has high acidity and is therefore highly active in cracking3. However, the data given in Table 3 show little difference in gas yield among the catalysts tested. This fact is interpreted as a result of the presence of few strong sites that can cause excessive cracking resulting in gas formation.

Acidity measurement The acidity is proportional to the amount of base molecules chemisorbed on the acid sites. The acid site strength distributions for the six commercially available catalysts, as determined by stepwise thermal desorption of pyridine, are presented graphically in Figures 1-6. In these figures, the first shaded column of the distribution represents the total acidity in terms of the total number of moles adsorbed at 473 K, and the other columns show the strength distribution of the total acid sites. The adsorption capacities at 673 K as a measure of strong

for coal runs (reaction

conditions:

4OO”C, 30 min, 2000 psi of Hz)

5

6

1

naph

naph

naph

naph

naph

HDN-30

HDN-60

Shell 3 I7

Shell 324

Amocat

0.120 0.128 0.572

0.086 0.213 0.073 0.124 0.495

0.086 0.211 0.077 0.126 0.499

0.086 0.298 0.121 0.245 0.750

0.107 0.234 0.115 0.257 0.713

0.094 0.251 0.086 0.293 0.724

8

9

10

11

12

13

14

tet _

tet

tet

tet

tet

tet

Amocat

HDN-30

HDN-60

Shell 3 17

Shell 324

Amocat

0.098 0.345 0.103 0.23 1 0.777

0.064 0.354 0.175 0.201 0.794

0.073 0.354 0.133 0.244 0.803

0.086 0.354 0.120 0.240 0.799

0.077 0.392 0.175 0.222 0.867

0.068 0.361 0.138 0.248 0.816

0.086 0.365 0.145 0.227 0.823

2

Solvent

naph _

naph

0.107 0.183

0.111 0.213

0.120 0.056 0.466

Run no. Solvent

Catalyst

Amocat

1A

1C _

(wt frac.)

Asphaltene Preasphaltene Coal conversion

Catalyst Product distributions Gas Oil Asphaltene Preasphaltene Coal conversion

W, L. Yoon et al.

4

1

distributions

catalysts:

3

Run no.

Product Gas Oil

distributions

yield of hydrotreating

tet 1A

1C

(wt frac.)

FUEL,

1991,

Vol 70, January

IO!3

Surface

acidity

and liquefaction

yield of hydrotreating

W. L. Yoon et al.

catalysts:

0.2

0.2

qi

versus

T curve

\ 7 01

7

m 0

0

E

:

7) =

0.1

Lz

b

b

6

0.1

a

0 0

300

200 Temperature

0

400

0

(OC)

(‘Cl

Temperature

Figure 1

Acid strength

distribution

of Amocat

1A Figure 4

versus

2 5

400

300

200

Acid strength

distribution

of Shell 317

T curve

0.1

0

300

200 Temperature

Figure 2

Acid strength

distribution

0

400

200

[“C)

Figure 5

of HDN-30

0.2

‘;

300

Acid strength

distribution

400

[OC)

of Shell 324

r

lx

E E

-I

-0 D

; P

.

Temperature

0.2

b .-

-

0

n_ -!! I

0.1

6 .cs

r

0

_I

200

Figure 3

110

Acid strength

FUEL,

1991,

300 Temperature

distribution

Vol

70,



300

200

[“C)

of HDN-60

January

0

400

Temperature

Figure 6

Acid strength

distribution

400

(OC)

of Amocat

1C

Surface Table 4 Weak and strong acid sites as determined (mm01 per g catalyst) Catalyst

Total acidity

Strong

Amocat 1A HDN-30 HDN-60 Shell 317 Shell 324 Amocat 1C

0.217 0.114 0.127 0.235 0.186 0.209

0.048 0.024 0.027 0.068 0.070 0.064

The values in parentheses

represent

acidity

by the STD method

sites

Weak sites

(22.1) (21.1) (21.3) (28.9) (37.6) (30.6)

the percentages

and liquefaction

0.169 0.090 0.100 0.167 0.116 0.145

(77.9) (78.9) (78.7) (71.1) (62.4) (69.4)

of each site

yield of hydrotreating

W. L. Yoon et al.

catalysts:

had the second highest acidity, it appears to be less active than other Ni-Mo catalysts, such as Amocat lC, which have almost the same acidity. The lower activity of this catalyst may be due to the use of Co instead of Ni as a promoter. Actually, a nickel promoter is more active for hydrogenation than cobalt, and acid catalysts can promote the cleavage of linkages that connect structural units, and thus crack the structures which comprise these units, but do not promote hydrogenation to stabilize the cracked products. As a consequence of their inadequate hydrogenative stabilization of the cracked products, cracking reactions are accompanied by condensation reactions leading to the formation of coke. Therefore, a proper coal liquefaction catalyst should possess a suitably

A

8

I

I

I

I

10

12

14

16

Weak acidity

(mmol per g cat. X 102)

Figure 7 Weak acidity versus oil yields with Ni-Mo naphthalene as solvent; A, tetralin as solvent

catalysts:

A,

8

I

I

I

I

10

12

14

16

Weak acidity

sites, which are shown in the final shaded column, were determined by the pulse techniquer3. The amount of pyridine reversibly desorbed by raising the temperature of the catalyst from 473 to 673 K is regarded as a measure of weak acid sites. The plot of amount of pyridine irreversibly adsorbed (qi) versus temperature (T) shows the acid strength distribution, where the number of acid sites is expressed by the amount of pyridine adsorbed irreversibly as a function of the adsorption temperature. The results in these figures show that the commercial hydrotreating catalysts used in this study had a broad acid strength distribution. The percentages and amounts of weak and strong acid sites, for qualitative comparison between catalysts, are also given in Table 4. Weak acid sites that give reversible adsorption sites can play a very important role in the catalytic process, and temperature programmed desorption of the reversibly adsorbed species would give a better understanding of the catalyst surface. The weak acidity decreased in the following order: Shell 3 17 > Amocat 1A > Amocat 1C > Shell 324> HDN-60 > HDN-30. The weak acidity results correlate well with the catalytic activity in coal liquefaction, except for Amocat 1A (CO-MO), as shown in Figures 7 and 8. The Shell 317 catalyst, which gave the highest coal liquefaction activity, had the greatest number of acid sites. Although the Amocat 1A catalyst

(mmol per g cat. x 102)

Figure 8 Weak acidity versus coal conversion with Ni-Mo 0, naphthalene as solvent; 0, tetrahn as solvent

25

8

11 11 11 8 11 11 II1I Weak acidity

Figure 9

II1I

12

Correlation

between

FUEL,

catalysts:

18 11 1 16

(mmol per g cat. x 102)

weak acidity

1991,

and pore volume

Vol 70, January

111

Surface

acidity

and liquefaction

yield of hydrotreating

combined dual function of both cracking and hydrogenation activity. It is interesting to note that a physical property of the catalyst, pore volume in 30-200 A pore diameter range, correlates with catalyst surface acidity, as shown in Figure 9. This might be partially explained by the fact that pyridine molecules do not diffuse into acid sites in small pores, resulting in poor utilization of the interior surface area with pore size less than 30 A. It has been reported previouslyi5,i6 that the micropore volume in a critical pore size range and the surface area are very important aspects in the performance of a liquefaction catalyst. It was recommended that to maintain good activity, the pore size distributions of the micropores should be rather narrow, with pores having 50-100 A radii to prevent plugging of the high surface area regions of the catalyst by coking. It has also been pointed out I7 that the catalyst activities for both hydrogen uptake and hydrodenitrogenation correlate with pore volume in 60-200 A diameter pores. The above results indicate that the specific activity of the catalyst increases with increasing average pore diameter in the preferred range. Consequently, it can be inferred that the catalyst activity correlates with either pore volume in 30-200 A diameter pores or with surface weak acidity. However, it is impossible to elucidate which factor has the more important effect on catalyst activity. To quantitatively relate the coal liquefaction catalyst activity and selectivity with catalyst surface properties, determination of the adsorption and desorption characteristics of hydrogen (hydrogenation ability) and of the surface acidity (cracking ability) are likely to be very important. Also, development of support material having high pore volume in the preferred range (30-200 A pore diameter) can have a significant influence upon the catalyst activity.

catalysts:

provide reversible adsorption sites for reactants and products of catalyst reaction is very important in improving liquefaction yield. With commercial hydrotreating catalysts having the same metals (Ni/Mo) on the surface, it was experimentally demonstrated that the catalytic activity for coal liquefaction correlated well with both the weak acidity of solid surface and the pore volume in 30-200 A pore diameters. However, it is impossible to determine which factor has the more important effect on activity. REFERENCES 1 2

3

4 5

6 7 8 9 10

11 12 13 14 15

CONCLUSIONS

16

Comparing the results of coal liquefaction efficiency with the acidity measurements of six commercial hydrotreating catalysts, it appears that the weak acidity which can

17

112

FUEL,

1991,

Vol 70, January

W. L. Yoon et al.

Derbyshire, F. Am. Chem. Sot. Div. Fuel Chem. Prepr. 1988, 33, 188 Anderson, R. B. and Dawson, P. T. in ‘Experimental methods in catalytic research’, Academic Press, New York, USA, 1968, 1st edition, Vol. 1, p. 362 Gates, B. C., Katzer, J. R. and Schuit, G. C. A. in ‘Chemistry of Catalvtic Processes’. McGraw-Hill Book Co.. 1979./ 1st edition, p. 19 Chessick, J. J. and Zettlenmoyer, A. C. Ado. Catul. 1959,11,263 Parviczak, T. in ‘Gas Chromatoeranhv in Adsorotion and Cataiysis’, Ellis Horwood Limited: Poland, 1986, 1st edition, p. 206 Ward, J. W. J. Caral. 1961, 9, 225 Ward, J. W. J. Catal. 1968, 10, 34 Ward, J. W. J. Catal. 1969, 14, 365 Choudhary, V. R. J. Chromatography 1983, 268, 207 Shabtai, J. S. and Oblad, A. G. ‘Chemistry and Catalysis of Coal Liquefaction: Catalytic and Thermal Upgrading of Coal liquids: and Hydrogenation of CO to produce Fuels’, DOE Final Report DOE/ET/14700-Tl, June 1985, Vol. 3, p. 229 Yoon, W. L., Jin, G. T., Kim, Y. I. et (11.Fuel 1989, 68, 614 Roberto, R. G. and Cronauer, D. C. Fuel Proc. Technol. 1979. 2, 215 Choudhary, V. R. and Nayak, V. S. Applied Catal. 1982,4,31 Rao, S. N., Schindler, H. D. and McGurl, G. V. Am. Chem. Sot. Diu. Fuel Chem. Prepr. 1988, 33, 145 Bertolacini, R. J., Gutberlet, L. C., Kim, D. K. and Robinson, K. K. ‘Catalyst development for coal liquefaction’, Final report AF-1084, EPRI, Palo Alto, CA, USA, June 1979 Kim. D. K.. Bertolacini. R. J.. Foreac. J. M. et al. ‘Catalvst development for coal liquefaction’, Fmal’report AF-1233, EPRI, Palo Alto, CA, USA, November 1979 McCormick, R. L., King, J. A., King, T. R. and Haynes Jr., W. Ind. Eng. Chem. Res. 1989, 28, 940