JOURNAL OF CATALYSIS 75, 233-242(1982)
Benzene Hydrogenation over Nickel Catalysts at Low and High
Temperatures: Structure-Sensitivity and Copper Alloying Effects
G.A. MARTIN AND J.A. DALMON
Institut de Recherches SW lu Catalyse (CNRS), 2 Atvenue Albert Einstein, 69626 Villeurbanne CPdrx, Frrrr~ce
Received February 9, 1981
The rate of benzene hydrogenation has been measured over various silica-supported Ni, unsup-
ported Ni, and Ni-Cu/SiOz catalysts from 300 to 600 K, and has been referred to unit area of metal.
It is shown that benzene hydrogenation at low temperature is to be considered as a structure-
sensitive reaction. As particle size increases from 2.5 to 6 nm, the activity increases. Reduction of
catalysts at high temperature results in a marked decrease of activity. This behavior is very similar
to that observed for ethane or propane hydrogenolysis. Copper addition slowly decreases the
activity. The active site is an ensemble composed of 3 7 2 adjacent nickel atoms. The influence of
particle size and of copper alloying on the rates of benzene conversion into cyclohexane or methane
at high temperatures is quite different from that observed at room temperature. Finally, a correla-
tion between alloying effects and particle size-sensitivity for particles smaller than 6 nm is pre-
sented.
INTRODUCTION metallic atoms involved in the active site
(6). In this respect, of special interest is the
Although benzene hydrogenation into cy- influence of alloying nickel with copper on
clohexane has received much attention (I), the rate of benzene hydrogenation. Some
no consensus of opinion exists in the litera- data are already available in the literature
ture about the structure-sensitivity of the (7, 8). The activity of nickel toward ben-
reaction. The rate of reaction per unit me- zene hydrogenation at low temperatures
tallic area over nickel catalysts is generally (below 373 K) decreases slowly as copper
considered as being little influenced by sur- concentration increases (7, 8), while it in-
face structure (2, 3). There are, however, creases at high temperatures (8). In both
some papers dealing with benzene hydroge- cases, however, the reaction was carried
nation over nickel catalysts which report a out on films (7) or powders (8) which are
more or less marked structure-sensitivity. known to exhibit a strong surface enrich-
On the one hand, Coenen et al. (4) found ment in copper, thus making difficult a
small nickel particles (size smaller than 1.2 quantitative study of the alloying effect.
nm) to be less active than large particles by This situation has prompted us to reex-
a factor of 5. On the other hand, Selwood et amine the problem of benzene hydrogena-
al. (5) reported a drop in activity which tion on nickel, and this paper presents new
greatly exceeds the fall of specific surface data on the structure-sensitivity of the reac-
area when nickel catalysts are thermally tion and on the influence of copper addition
sintered. This shows that the options on on the activity of well-defined catalysts.
structure-sensitivity of benzene hydrogena-
tion are far from being completely unequiv- EXPERIMENTAL
ocal. The morphological characteristics of the
It has been suggested that the structure- catalysts, which were partially described in
sensitivity of catalytic reactions could be previous papers (9, II), are summarized in
related to alloying effects and that both phe- Table 1. Unsupported nickel powder sam-
nomena could be ascribed to the number of ples were obtained by decomposition of
233
0021-95 17/82/060233-10$02.00/O
Copyright 0 1982 by Academic Press. Inc.
AU rights of reproducflon in any form reserved.
234 MARTIN AND DALMON
TABLE 1
Sample Morphology of Nickel Catalysts
Number Precursor Ni Reducing Degree of DS S
loading temperature reduction @ml (m”/g Ni)
1% wt) f”K)
1 Ni(OHh 62 520 80” 12
2 Ni(OHh 62 570 1.01 67” 10
3 Ni(OH)* 62 670 96” 7
4 Ni(OH), 62 770 167O 4
5 Ni(OH)&SiOz
(hexamine) 4.5 800 0.95 2.5
6 Ni(OH)JSiOZ
(hexamine) 4.5 920 1 3.2
7 Ni(OH)JSiOz
(hexamine) 11 970 1 5.7
8 Ni(OH).JSiO,
(hexamine) 23 920 0.98 6.3
9 Ni(OH),/SiO,
(hexamine) 23 lOoa 1 7.5
10 Ni(OH)JSiOz
(hexamine) 23 1200 1 14
11 Ni*+/SiO,
(ex nitrate) 6 770 1 12
a As deduced from the BET surface area assuming spheres.
nickel hydroxide (prepared by the ammonia with diameters of the other samples calcu-
method) in flowing helium for 1 hr, followed lated from magnetic data. The Ni-Cu/SiOz
by reduction for 2 hr in flowing hydrogen catalysts were prepared by reduction at 920
(gas flow, at a rate of 4 liters/hr), at temper- K in flowing hydrogen of precursors ob-
atures listed in the table. Surface areas were tained by adding silica to solutions of nickel
calculated from nitrogen adsorption data at and copper nitrate hexamine (12). Ad-
liquid nitrogen temperature. The Ni/SiOz sorption and magnetic studies have shown
catalysts were prepared by reduction for 15 that a homogeneous alloy was formed and
hr of precursors obtained by reacting the that the surface composition of the metallic
support (SiOs, Aerosil Degussa, 200 m*/g) particles (diameter, 6 nm) was very similar
with a solution of nickel nitrate hexamine to the bulk composition (12), in contrast
(samples 5 to 10) or with a solution of nickel with the surface enrichment in copper gen-
nitrate (sample 11). Reduction temperatures erally observed on Ni-Cu films and unsup-
and metallic loadings were varied to cover a ported powders.
wide range of particle sizes. Prior to reduc- Kinetic experiments were carried out in a
tion, samples 5 and 11 were treated in va- flow system with a fixed-bed reactor at at-
cuo for 1 hr at the reduction temperature. mospheric pressure. The microreactor con-
Magnetic measurements showed that re- sisted of a quartz tube with a porous disk
duction of the samples was almost complete and samples were held by quartz wool. The
and also allowed us to calculate the surface total flow at room temperature was 120 ml/
average diameters, D, listed in Table 1 min. Hydrogen and helium (used as a dil-
(9, II). The volumes of adsorbed hydrogen uent) had an initial purity better than
at saturation were found to be in agreement 99.99%, and were further purified by a De-
with the BET surface area of sample 2, and 0x0 catalyst followed by a zeolite trap. Par-
BENZENE HYDROGENATION OVER Ni 235
tial pressures of benzene were varied by crease of activity with time obeys a simple
bubbling a flow of helium in liquid benzene. logarithmic law which allows extrapolation
Pure thiophene-free benzene from Merck to initial time. The variations of the initial
was further purified by adding metallic so- activity per unit area with metallic particle
dium and by a clean Pt/SiO, catalyst to diameters are shown in Fig. 2. Experimen-
remove minute amounts of sulfur com- tal points have been arranged more or less
pounds (no important difference of behav- arbitrarily into two groups (curve A and B
ior however, was detected between initial in Fig. 2) corresponding to data obtained
and purified benzene). Gas analyses were with samples reduced at temperatures
performed by gas chromatography with a higher and lower than 770 K, respectively.
flame-ionization detector. Conversion was Results obtained by Coenen et al. (13) and
always less than 2% except for aging exper- by Nikolajenko et al. (2) can be compared
iments. The rates reported here refer to the to ours in Fig. 3. Figure 4 shows the activi-
rate of conversion of the parent hydrocar- ties of samples 5-10 toward various reac-
bon (benzene). tions, namely, benzene hydrogenation, and
hydrogenolysis of ethane and propane (14,
RESULTS
1.5). From these data, it can be seen that the
Benzene Hydrogenation at Low* various activities vary more or less in a par-
Temperatures allel way, particularly for particle diameters
The variations of catalytic activity with larger than 6 nm. Below this critical diame-
time were first examined. As can be seen in ter, the activities decrease with nickel parti-
Fig. 1, the phenomenon of catalyst aging is cle size and the corresponding slopes vary
observed for samples prepared below 520 according to the reaction considered. Varia-
and 770 K for unsupported and supported tions of the rate of benzene hydrogenation
catalysts, respectively. The observed de- at low temperature with the copper concen-
trationx are shown in Fig. 5, from which it
can be deduced that the variations of the
LOG conversion
activity A, with x can be approximately
represented by
-1
A, = A,(1 - x)”
with N = 3 T 2.
Apparent activation energies, E,, and
partial reaction orders with respect to hy-
-2
c.m3Tr~- 3
drogen and benzene partial pressure, nH and
nB, of the samples were also measured. The
results thus obtained can be compared in
Table 2. To a first approximation, they can
be considered as constant, showing that the
-3 mechanism does not change with particle
size and alloying.
O\ 1 Reaction of Benzene with Hydrogen at
TIME]min )
High Temperatures
30 60 The reaction was also studied over a wide
FIG.1. Catalyst aging: Log conversion as a function range of temperature. A typical Arrhenius
of time for samples 1, 11, 3, and 8 (curves 1-4, respec- plot showing the influence of temperature
tively). on conversion is represented in Fig. 6. As
236 MARTIN AND DALMON
LOG r (molecules/sea/cm2 Ni )
13
12
11
LOG Ds(nm)
1 1
.5 1 1.5 2
FIG. 2. Log rate hydrogenation of benzene into cyclohexane at T = 303 K, Pa = 600 Torr, Pa = 4
Torr as a function of the average particle size D,. Curves A and B correspond to samples reduced at
temperatures higher and lower than 770 K, respectively.
previously observed (16 -Is), the curve served at high temperatures, corresponding
goes through a maximum at 420 K and in probably to benzene hydrogenolysis ac-
the high temperature region the apparent cording to:
activation energy is -20 kcal/mole; the or- CGH, + 9Hz ---) 6CH4
ders with respect t0 partial pressures are na
= 3 and nB = 0.4, in accord with the data of The apparent kinetic parameters of this
van Meerten and Coenen (16). Beside cy- reaction are E, = 28 kcal/mole, nH = 0.3,
clohexane, formation of methane is also ob- and nB = -0.4 at 520 K. Small amounts of
13 LOG r (molecules/ sec/cm2 Ni )
FIG. 3. Structure-sensitivity of benzene hydrogenation: comparison of our data (curves 2A and 2B)
with those of Coenen et al. (4) and Nikolajenko et al. (2), curves 1 and 3, respectively.
BENZENE HYDROGENATION OVER Ni 237
1 LOG I TABLE 2
Apparent Kinetic Parameters of Benzene
Hydrogenation”
Samples Number Cu E, %i “a
content (kcalimole)
Unsupported
Ni I 13 0.45 -
2 11.4 0.54 0.06
3 I2 --
Ni/SiOz 5 10.4 0.55 0.14
8 11.6 0.51 0.12
10 11.7 0.57 0.05
II 13 0.45 -
Ni-Cu/SiO, 12 14 12.7 - -
13 36 12.1 0.48 0.08
D&l”‘) Average
I 5 10 15
values 11.9 0.5 0.09
FIG. 4. Log rate as a function of the average nickel
a Measured at room temperature, P, = 300 Torr, P,
particle size: Curve 1: Benzene hydrogenation at T = = 40 Torr.
303 K, Pa = 600 Torr, Pa = 4 Torr. Curve 2,3, and 4: zene hydrogenation into cyclohexane, cal-
hydrogenolysis of ethane, hydrogenolysis of propane
into methane + ethane, and hydrogenolysis of propane culated on the basis of data given in Ref.
into 3 molecules of methane, respectively, at T = 507 (I 6). It shows that the approach to the ther-
K, P, = 160 Torr, and Phydmcarbon= 25 Torr. modynamic equilibrium conversion does
not seem to limit the reaction rate, as al-
ready concluded by van Meerten and
cyclohexene are also detected, as illus- Coenen (161. This has been experimentally
trated in Fig. 6, particularly on sample 10. verified by varying the gas flow. The per-
The apparent kinetic parameters of cyclo- centage of benzene converted into cyclo-
hexene formation on sample 10 at 500 K are hexane is proportional to the contact time,
E, = -19 kcaVmole, II~ = 1.7, and nB = showing that thermodynamic equilibrium is
1.1. not attained.
In Fig. 6 is also shown the thermody- The variations with particle size of the
namic equilibrium curve of reaction of ben- rate of benzene conversion into cyclohex-
Log Ax I Ao
2< -
0 -----J-w I __
-- --
-z
I --
FIG. 5. Log activity in benzene hydrogenation measured at T = 303 K, P, = 30 Torr, Pn = 730 Torr,
as a function of x, the copper concentration of the alloy. The dashed curves are calculated assuming
that the rate varies as (1 - x)“.
238 MARTIN AND DALMON
LOG ,’
:onvers.f
I
If
I
I
/
I
2 25 3
FIG. 6. Log conversion as a function of the reciprocal temperature measured at PH = 300 Torr, Pa =
40 Torr for sample 10. Curve 1, benzene hydrogenation into cyclohexene. Curve 2, benzene hydrogena-
tion into cyclohexene. Curve 3, benzene hydrogenolysis into methane. Curve 4, thermodynamic
equilibrium benzene-hydrogen-cyclohexane.
ane and methane at high temperatures are The effects of alloying nickel with copper
shown in Fig. 7. The curve corresponding are shown in Fig. 8. As already reported by
to the former reaction is very different from Van Barneveld and Ponec (8), copper addi-
that observed for benzene hydrogenation at tion depresses the rate of benzene hydro-
low temperatures (Fig. 2). As particle size genolysis, and increases the rate of benzene
increases from 2.5 to 6 nm the high temper- hydrogenation at high temperatures (8). It
ature rate decreases slowly (curve 1, Fig. can be noticed that alloying produces oppo-
7). Beyond 6 nm, the activity remains con- site effects on benzene hydrogenation at
stant. The rate of benzene hydrogenolysis low and high temperatures.
into methane (curve 2, Fig. 7) first increases
as particle size increases, then remains DISCUSSION
more or less constant. It can be noticed that
the structure-sensitivity of benzene hydro- Structure-Sensitivity
genolysis is different from that of ethane Results reported in Fig. 2 show unambig-
and propane hydrogenolysis and is very uously that benzene hydrogenation into cy-
similar to that of CO hydrogenation (19). clohexane at low temperatures is a struc-
BENZENE HYDROGENATION OVER Ni 239
trast, when large diameters are obtained by
Log r ( molecules/sect cm2Ni)
reduction at high temperatures, there is a
sharp decrease in activity as diameters in-
crease (curve A in Fig. 2). This finding
confirms earlier results obtained by Sel-
wood et al. (5) and reconciles the appar-
ently conllicting results of the literature.
What is the origin of the low activity to-
ward benzene hydrogenation at room tem-
perature and ethane or propane hydrogen-
olysis exhibited by Ni/SiOP reduced at high
Wnm)
I I temperatures? This point has been already
5 10 15
partly discussed in previous work (I I >. Let
FIG. 7. Log rate conversion of benzene into cyclo- us recall that the hypothesis of the forma-
hexane (curve 1) and methane (curve 2) measured at T tion of bulk Ni-Si alloys with a concentra-
= 572 K, PH = 600 Torr, Pa = 40 Torr as a function of
tion of silicon larger than 1% is to be ruled
nickel particle size D,.
out and that the gas accessibility is not dra-
matically affected. In another paper (22),
ture-sensitive reaction. If we compare our we have suggested that Ni/SiOo reduced at
results with those of Coenen et al. (4) and of high temperatures could be in a state simi-
Nikolajenko et al. (2), the same general lar to that observed on Pt/TiO, reduced at
trend is noticed (Fig. 3). As particle size ca. 800 K (the so-called SMSI state), since
increases the activity first increases, O2 treatments followed by a low-tempera-
reaches a maximum, then remains constant ture reduction are capable of restoring the
or decreases. The diameter corresponding activity of the Ni/SiOz catalyst. It can be
to the maximum is 6 nm in our case. The noticed that the reduction of Ni/SiOz at
value is between those observed by Nikola-
jenko et al. (7 nm) and Coenen et al. (1.2
nm). It is hard to decide whether these dis-
crepancies are significant-differences in
particle size distributions, degrees of reduc-
tion, or in experimental conditions may be
invoked to account for observed differ-
ences-or are to be attributed to artifacts
resulting from the different techniques and
assumptions used to estimate nickel parti-
cle sizes. These points probably require
further investigations.
Our results show also that activity does
not depend unequivocally on particle size
but rather on conditions prevailing as parti-
cles are prepared, particularly for large par-
ticle sizes. When samples are prepared at
moderate temperatures, say below 750 K,
variation of the activity with diameter is not
very pronounced (curve B in Fig. 2). This FIG. 8. Log rate conversion of benzene into cyclo-
observation is in agreement with data in hexane (curve 1) and methane (curve 2) measured at T
papers reporting on nickel catalysts pre- = 572 K, P, = 600 Torr, Pa = 40 Torr as a function of
pared below 750 K (2, 3, 4, 20, 21). In con- the copper concentration x of the alloy.
240 MARTIN AND DALMON
high temperatures does not seem to have however, that this value is underestimated,
any influence on benzene hydrogenolysis since these authors have assumed that satu-
into methane, and benzene hydrogenation ration in hydrogen coverage is obtained at
at high temperatures (Fig. 7). This means room temperature under 10 Tort-. We know,
that the occurrence of the SMSI (strong now, that this is to be considered as a first
metal-support interaction) state affects approximation, and that in these condi-
some catalytic reactions but does not influ- tions, dH lies between 0.6 and 0.8 (IO), 0.7
ence others. Further work is desirable to being a plausible average value. If one cor-
elucidate this intriguing behavior. rects the n value obtained by Candy et al.,
Results reported in this study show one obtains IZ = 1.6/0.7 = 2.3. We can con-
clearly that the structure-sensitivity of a sider that this value is in agreement with N
given reaction may depend on conditions as deduced from Ni-Cu experiments.
prevailing as the rate is measured. When When the reaction is performed at high
measured at high temperatures, benzene temperatures, the effect of copper addition
hydrogenation into cyclohexane appears to on benzene hydrogenation is reversed (Fig.
be a reaction which is less structure-sensi- 8). Our data confirm those of Van Barneveld
tive than at low temperatures. The small and Ponec (8). Let us recall that these au-
structure-sensitivity of benzene hydrogena- thors have attributed this positive alloy ef-
tion at high temperatures reported here fect to some inhibition of side-reactions,
agrees well with the conclusions of Dalmai- like coke formation and hydrogenolysis,
Imelik and Massardier (23) who have re- which would result in an increase of the
ported no significant differences of activity activity toward benzene hydrogenation.
toward benzene hydrogenation on (11 l), Our results confirm that copper addition de-
(IOO), and (110) planes of nickel, when the creases sharply the rate of benzene hydro-
reaction is performed at 443 K under PH = genolysis into methane, in accordance with
50 Tori-, i.e., in high-temperature conditions this hypothesis (8).
(16). Finally, it is interesting to notice that the
hypothesis of Van Barneveld and Ponec (8)
Alloying Effects accounts for the results of Fig. 7. When
The variations of the rate of benzene hy- nickel particle sizes increase from 2.5 to 15
drogenation at low temperature on Ni-Cu nm, the activity toward benzene hydroge-
with X, the copper concentration, can be nation at high temperatures first decreases
represented by a law of the type (1 - x)~ then remains constant, when the activity
with N = 3 + 2. It can be recalled that (1 - toward benzene hydrogenolysis, one of the
x)~ is the probability of finding an ensemble side-reactions invoked by these authors,
of N adjacent Ni atoms assuming the ran- first increases then remains constant.
dom distribution of metallic atoms (24).
This means that the active site of benzene
hydrogenation at low temperature is an en- Reaction between Alloying EfSect and
semble of N = 3 7 2 adjacent nickel atoms. Structure-Sensitivity
The number of nickel atoms of the active Let us define the particle size-sensitivity
site may be compared to that deduced from (Yand the alloying-sensitivity /3 as,
volumetric measurements by Candy et al.
dA dA
(25). The consumption of hydrogen in con- “=a. P=,,
tact with a nickel surface precovered with
benzene shows that the number of nickel where A is the activity per unit area, D the
atoms bonded to one molecule of benzene particle diameter, and x the copper concen-
which is hydrogenated is n = 1.6 (average tration. In Table 3 are displayed the signs of
value of five determinations). We think, (Y and /3 for the reactions studied here and
BENZENE HYDROGENATION OVER Ni 241
TABLE 3
Alloying Effects and Particle Size Sensitivity
Reactions Alloying effectsa Particle size sensitivity”
dA
*=;T; D< 6nm D>6nm
dA dA
P=a P=;ils
C,H, + 3H2 = C6H,2
low temperature
C,H, + 3H, = C,H,,
high temperature + - 0
C6H6 + 9HZ = 6CH4 - + 0
CO + 3H2 = CH, + H,O - + 0
’ +, 0, and - mean that (Yor p are positive, nearly equal to zero, or negative, respectively.
for methanation (19) on samples 5-l 1. which requires an ensemble of N adjacent
There is no correlation between alloying metallic atoms is a complex function of N,
and particle size effects for particles larger D, diameter of particles, and also of 8, the
than 6 nm. Below this critical size, how- degree of coverage with adsorbed species.
ever, a clear parallelism is observed: cop- This function is capable of accounting for
per addition leads qualitatively to the same the observed particle size sensitivities in a
effect as decreasing particle size. This cor- more quantitative way.
relation results probably from the fact that
the active site involves a certain number of CONCLUSION
adjacent metallic atoms varying from one From this work it can be therefore con-
reaction to another (15, 19, 24), accounting cluded that: (i) benzene hydrogenation at
for the observed decrease of activity and room temperature exhibits a structure-sen-
change of selectivity with copper addition. sitivity similar to that of ethane or propane
As particle size decreases, the number of hydrogenolysis, particularly for particle
available metallic surface atoms per parti- sizes larger than 6 nm; (ii) the influence of
cle decreases, then the activity is expected particle size and of copper alloying on the
to vanish below some critical diameter. rate of benzene reaction with hydrogen at
However satisfying from the qualitative low temperature is quite different from that
viewpoint this explanation may be, it can- observed when the reaction is performed at
not be overlooked that it raises some high temperatures; (iii) a correlation does
difficulties from the quantitative viewpoint exist between alloying effects and particle
since critical diameters thus calculated are size sensitivity for particles smaller than
very small (ca. 0.5 nm for a reaction involv- 6 nm.
ing 12 adjacent nickel atoms). Moreover, it
is not easy to understand why the particle REFERENCES
size effects of benzene hydrogenation at 1. Moyes, R. B., and Wells, P. B., in “Advances in
low temperatures and of ethane hydrogen- Catalysis” (D. D. Eley, H. Pines, and P. B. Weisz,
olysis which require ca. 3 and 12 or more Eds.), Vol. 23, p. 121. Academic Press, New York,
adjacent nickel atoms, respectively, are of 1973.
2. Nikolajenko, V., Bosafek, V., and Danes, V. L.,
the same order of magnitude. As a matter of J. Caral. 2, 127 (1963).
fact, it will be shown in another paper (26) 3. Aben, P. C., Platteuw, J. C., and Stouthamer, B.,
that the size sensitivity, (Y, of a reaction Rec. Trav. Chim. 89, 449 (1970).
242 MARTIN AND DALMON
4. Coenen, J. W. E., van Meerten, R. 2. C., and 14. Martin, G. A., J. C&a/. 60, 345 (1979).
Rijnten, H. T., “Proc. Int. Congr. Catalysis, 5th 15. Guilleux, M. F., Dalmon, J. A., and Martin, G. A.,
(Florida 1972)” (J. W. Hightower, Ed.), p. 671. J. Catal. 62, 235 (1980).
North-Holland, Amsterdam, 1973. 16. van Meerten, R. 2. C., and Coenen, J. W. E., J.
5. Selwood, P. W., Adler, S., and Phillips, T. R., J. Catal. 37, 37 (1975); 46, 13 (1977).
Amer. Chem. Sot. 76,2281(1954); 77, 1462(1955). 17. Herbo, C., Bull Sot. Chem. Be/g. 50, 257 (1941).
6. Boudart, M., “Proc. Int. Congr. Catalysis, 6th 18. Germain, J. E., Maurel, R., Bourgeois, Y., and
(London 1976)” (G. C. Bond, P. B. Wells, and F. Sinn, R., J. Chim. Phys. 60, 1219 (1963).
C. Tompkins, Eds.), p. 1. Chemical Society, Lon- 19. Dalmon, J. A., and Martin, G. A., “Proc. Int.
don, 1977. Congr. Catalysis, 7th (Tokyo 1980),” p. 402. Else-
7. Van der Plank, P., and Sachtler, W. M. H., J. vier, Amsterdam, 1981.
Catal. 12, 35 (1968). 20. Dixon, G. M., and Singh, K., Trans. Faraday Sot.
8. Van Bameveld, W. A. A., and Ponec, V., Rec. 65, 1128 (1969).
Trav. Chim. 93, 243 (1974). 21. Taylor, W. F., and Staffin, H. K., Trans. Faraday
9. Prime& M., Dalmon, J. A., and Martin, G. A., J. Sm. 63, 2309 (1967).
Catal. 46, 25 (1977). 22. Martin, G. A., and Dalmon, J. A., React. Kinet.
10. Martin, G. A., J. Catal. 60, 452 (1979). Catal. Lett., in press.
II. Martin, G. A., Dutartre, R., and Dalmon, J. A., in 23. Dalmai-Imelik, G., and Massardier, J., “Proc. Int.
“Growth and Properties of Metal Clusters” (J. Congr. Catalysis, 6th” (G. C. Bond, P. B. Wells,
Bourdon, Ed.), p. 467. Elsevier, Amsterdam, and F. C. Tompkins, Eds.), p. 90. Chemical Soci-
1980. ety, London, 1977.
12. Dalmon, J. A., J. Catal. 60, 325 (1979). 24. Dalmon, J. A., and Martin, G. A., J. Catal. 66,
13. Coenen, J. W. E., van Meerten, R. Z. C., and 214 (1980).
R&ten, H. T., “Proc. Int. Congr. Catalysis, 5th 25. Candy, J. P., Fouilloux, P., and Imelik, B., Nou-
(Florida 1972)” (J. W. Hightower, Ed.), p. 671. veau J. Chim. 2, 45 (1978).
North-Holland, Amsterdam, 1973. 26. Martin, G. A., and Dalmon, J. A., to be published.
