Catalytic Hydrogen Treatment of Aromatic Alcohols

Article abstract of DOI:10.1006/jcat.1997.1920

The gas phase hydrogenation/hydrogenolysis of phenol over the temperature range 423 K ≤ T ≤ 573 K has been studied using a 1.5 % w/w Ni/SiO₂ catalyst. The effects on reaction rate and product selectivity of varying such process variables as reaction time and temperature, contact time, phenol and hydrogen partial pressures, and the phenol solvent were considered. Hydrogenation was observed to occur in a stepwise fashion where the selective formation of the intermediate, cyclohexanone, was promoted at low hydrogen partial pressures and high temperatures, while hydrodeoxygenation to benzene proceeded at temperatures in excess of 523 K. The catalytic hydrogen treatment of cyclohexanone and cyclohexanol, under identical reaction conditions, was examined in order to establish the reaction mechanism. The observed conversions and selectivities are compared to those calculated from equilibrium constant data and it is shown that reaction equilibria are prevalent on the catalyst surface where the equilibrium is shifted to complete hydrogenation to cyclohexanol. The overall reaction is first order in phenol and the apparent dependency on hydrogen concentration increased with increasing temperature. Process selectivity is interpreted in terms of reactant/catalyst interactions where phenol adsorption is viewed as occurring via the aromatic π-electron system and/or the hydroxyl substituent where the latter interaction promotes hydrogenolysis as a direct loss of selectivity with respect to cyclohexanone formation. The hydrogen treatments of the three isomers of cresol were also considered for comparative purposes where the turnover frequency decreased in the order phenol ≥ m-phenol ≥ p-cresol ≥ o-cresol; reactivity and selectivity trends are accounted for in terms of electronic and steric effects.

Full text of DOI:10.1006/jcat.1997.1920

JOURNAL OF CATALYSIS 173, 450–459 (1998)
ARTICLE NO. CA971920
Catalytic Hydrogen Treatment of Aromatic Alcohols
Eun-Jae Shin and Mark A. Keane1
Department of Chemical Engineering, The University of Leeds, Leeds LS2 9JT, United Kingdom
Received July 4, 1997; revised September 15, 1997; accepted September 15, 1997
combustion process is necessary which has resulted in com-
plex, oversized, and costly combustion units; products of
incomplete combustion (PIC) can include organic acids,
aldehydes, CO, amines, etc., depending on the nature of
the incinerator feed (4, 5). In order to minimise environ-
mental effects and conserve resources, the first abatement
option that should be considered is the recovery and reuse
of raw materials. Catalytic hydrogen treatment is proposed
here as a viable alternative to thermal/catalytic inciner-
ations for transforming hazardous substances into useful
products. Catalytic hydroprocessing has long been used in
industry for hydrorefining petroleum but the application of
this methodology to the detoxification of organic hazardous
wastes has not yet been fully explored (6). In a pilot plant
study, Kalnes and James (7) have successfully hydropro-
cessed some hazardous liquid organic wastes and demon-
strated the economic and environmental advantages of
hydrotreatment compared with thermal incineration. Aro-
matic alcohols such as phenol and cresol originate from
natural as well as industrial sources and are part of the
many hazardous organic compounds present in coal, tar,
and petroleum (8). Effluents from coke ovens, oil refineries,
petrochemical units, polymeric resin manufacturing, and
plastic units contain such aromatic alcohols in large quan-
tities (9). These emissions are of major concern because of
their persistence in air and toxicity to all life forms (10)
and phenol entry into the environment is now regulated
in terms of mpc (maximum permissible concentrations) by
the World Health Organisation (WHO) and a number of
environmental agencies (11).
The gas phase hydrogenation/hydrogenolysis of phenol over the
temperature range 423 K ꢀ T ꢀ 573 K has been studied using a 1.5%
w/w Ni/SiO2 catalyst. The effects on reaction rate and product se-
lectivity of varying such process variables as reaction time and tem-
perature, contact time, phenol and hydrogen partial pressures, and
the phenol solvent were considered. Hydrogenation was observed
to occur in a stepwise fashion where the selective formation of the
intermediate, cyclohexanone, was promoted at low hydrogen par-
tial pressures and high temperatures, while hydrodeoxygenation to
benzene proceeded at temperatures in excess of 523 K. The cata-
lytic hydrogen treatment of cyclohexanone and cyclohexanol, under
identical reaction conditions, was examined in order to establish the
reaction mechanism. The observed conversions and selectivities are
compared to those calculated from equilibrium constant data and
it is shown that reaction equilibria are prevalent on the catalyst sur-
face where the equilibrium is shifted to complete hydrogenation to
cyclohexanol. The overall reaction is first order in phenol and the
apparent dependency on hydrogen concentration increased with in-
creasing temperature. Process selectivity is interpreted in terms of
reactant/catalyst interactions where phenol adsorption is viewed as
occurring via the aromatic ꢀ-electron system and/or the hydroxyl
substituent where the latter interaction promotes hydrogenolysis as
a direct loss of selectivity with respect to cyclohexanone formation.
The hydrogen treatments of the three isomers of cresol were also
considered for comparative purposes where the turnover frequency
decreased in the order phenol ꢁ m-phenol > p-cresol > o-cresol; re-
activity and selectivity trends are accounted for in terms of elec-
tronic and steric effects.
ꢂc 1998 Academic Press
INTRODUCTION
The hydrogenation of phenol in the gas (12–18) and liq-
uid (19, 20) phase has been reported, in the main, for a range
of palladium (12, 13, 15–17, 20, 21) and platinum (12, 14,
18) based catalyst systems where the emphasis was placed
on the selective production of cyclohexanone, a key ma-
terial in the manufacture of nylon 6 and polyamide resins
(21). The use of nickel catalysts has been considered to a
far lesser extent (22, 23). A variety of activity/selectivity
responses to changes in such process variables as reaction
time (13, 15–17), temperature (12, 13, 15, 16, 24), and reac-
tant(s) concentration (12, 13, 16, 17, 22) for the diverse cata-
lyst systems are documented in the literature. The role of
The stringent statutory regulations that now govern the
handling of hazardous organic substances present in many
industrial waste streams have accelerated the need to estab-
lish effective treatment technologies (1). Thermal and/or
catalytic oxidations are the widely favoured air pollution
abatement techniques, offering high destructive efficiency
(2, 3). To ensure a continuous and complete destruction
of the hazardous chemicals, a conservative design of the
1 To whom correspondence should be addressed. Fax:+44-113-2332405.
E-mail: chemaak@leeds.ac.uk.
450
0021-9517/98 $25.00
c
Copyright ꢂ 1998 by Academic Press
All rights of reproduction in any form reserved.

HYDROGEN TREATMENT OF AROMATIC ALCOHOLS
451
metal loading has also been considered (13, 14, 16–18, 22), organic feed at a fixed rate which has been carefully cali-
and limited kinetic information is available in the form of brated, and the vapour was carried through the catalyst bed
reaction orders (13, 16) and activation energies (24). It is, in a stream of purified hydrogen. Alcoholic solutions (nor-
however, fair to state that comprehensive kinetic and mech- mally methanolic) of phenol and, for comparative purposes,
anistic studies are still required to fully assess the feasibility methanolic solutions of cyclohexanol, cyclohexanone and
of catalytic hydrogenation/hydrogenolysis as a waste mini- o-, m-, and p-cresol were used as feedstock where the
misation/pollution abatement methodology. The gas phase aromatic concentration was in the overall range 2 ꢅ 10ꢃ4
–
hydrogen treatment of phenol over a well-characterised 0.3 mol hꢃ1. The catalytic measurements were made at an
Ni/SiO2 catalyst (25, 26) is reported in this paper. The ef- overall space velocity of 560–2250 hꢃ1 (STP) and at W/F
fects of reaction time and temperature, reactant(s) partial values in the range 17–2.2 ꢅ 103 g molꢃ1 h where W is the
pressure(s), and contact time on product distribution have weight of activated catalyst and F is the flow rate of the
been examined, and reproducible turnover frequencies are aromatic. Under these conditions, the catalyst system has
quoted as part of a kinetic and mechanistic study. The varia- been shown to operate with negligible internal or external
tion of overall reactivity and selectivity due to the presence diffusion retardation of reaction rate (26). The reaction or-
of a methyl substituent (o-, m-, and p-cresols) has been con- der with respect to hydrogen in the conversion of phenol
sidered and the role of the reaction solvent is addressed. was determined by dilution in helium to vary the hydrogen
This paper represents a fundamental study of the catalytic partial pressure in the range 0.4–0.8 atm; the reaction order
hydrogen treatment of phenols and will form the basis for with respect to phenolwaslikewise evaluated byvaryingthe
a future feasibility study of hydrodechlorination as a vi- aromatic partial pressure in the range 0.01–0.15 atm. The
able methodology in the treatment of the highly toxic (27) reactor effluent was frozen in a liquid nitrogen trap for sub-
chlorophenols.
sequent analysis, which was made using an AI Cambridge
GC94 chromatograph equipped with a flame ionization de-
tector and employing a DB-1 50 m ꢅ 0.20 i.d., 0.33 ꢁm
capillary column (J&W Scientific). Data acquisition and
analysis were performed using the JCL 6000 (for Win-
dows) chromatography data system and the overall level
of reactant hydrogenation/hydrogenolysis was converted to
mol% conversion using calibration plots for each feedstock.
Mol% conversion is defined as (mi ꢃ mo)/(mi) ꢅ 100 where
mi is the initial concentration or number of moles of reac-
tant entering the reactor per unit time and mo is the number
of moles of product exiting the reactor per unit time. Molar
selectivity in terms of product x is defined by mx/mtot ꢅ 100
where mtot is the total number of moles of product. All the
reactants were AnalaR grade and were used without fur-
ther purification.
EXPERIMENTAL
Catalyst Preparation and Activation
The catalyst wasprepared bythe homogeneousprecipita-
tion/deposition of nickel onto a nonporous microspheroidal
Cab–O–Sil 5M silica of surface area 194 m2 gꢃ1 as described
in detail elsewhere (25, 28). The nickel loading was deter-
mined by atomic absorption (Perkin–Elmer 360 AA spec-
trophotometer) and the water content by thermogravime-
try as outlined previously (29). The hydrated sample, sieved
in the 150–200 ꢁm mesh range, was reducedꢃ,1without a pre-
calcination step, by heating in a 100 cm3 min stream of dry
hydrogen (99.9% ) at a fixed rate of 5 K minꢃ1 (controlled
using a Eurotherm 91e temperature programmer) to a final
temperature of 673 ꢄ 1 K which was maintained for 18 h.
The catalyst activation procedure and characterisation by
carbon monoxide chemisorption and transmission electron
microscopy are given a full treatment elsewhere (25, 28).
RESULTS AND DISCUSSION
Effect of Varying Reaction Conditions
The catalyst precursor has a nickel loading of 1.5% w/w
and low water content (<1% w/w). Under the stated activa-
Catalytic Procedure
All of the catalytic reactions were carried out under tion conditions, the chemisorption data yield a nickel metal
atmospheric pressure in a fixed bed glass reactor (i.d. = dispersion equal to 73% which represents an estimated av-
15 mm) over the temperature range 423–573 K. The cata- erage nickel particle size of 1.4 nm and a metal surface
lyst was supported on a glass frit and a layer of glass beads area of 74 m2 gꢃ1. Transmission electron microscopy anal-
above the catalyst bed ensured that the reactants reached ysis of the activated catalyst (25) supports the chemisorp-
the reaction temperature before contacting the catalyst. tion data and revealed a narrow nickel metal particle size
The reactor temperature was monitored by a thermocou- distribution. One batch of this catalyst was used to gen-
ple inserted in a thermowell within the catalyst bed; reactor erate all the catalytic data presented in this paper which
temperature was constant to within ꢄ1 K. The catalytic re- represents time-on-stream in excess of 500 h over the tem-
actor has been described previously in some detail (26, 30). perature interval 423 K ꢀ T ꢀ 573 K, processing a range of
A Merck–Hitachi LC-6000A pump was used to deliver the hydrogen/organic combinations with no appreciable loss of

452
SHIN AND KEANE
catalytic activity. The hydrogen treatment of phenol in the
presence of the nickel catalyst generated cyclohexanone
and cyclohexanol as the products of aromatic ring reduction
while benzene was generated at temperatures in excess of
523 K via hydrogenolysis of the hydroxyl substituent. Trace
amounts (<1 mol% ) of cyclohexane were also isolated in
the product mixture. Passage of phenol in a stream of hy-
drogen through a fixed bed of glass beads (in the absence of
Ni/SiO2) over the same temperature range did not result in
any measurable transformation of phenol. Previous reports
have revealed a dependence of reaction rate (24) and pro-
cess selectivity (22) on contact time. In this study, the phe-
nol turnover frequency (number of molecules converted
per second per active site) and the product selectivity were
found to be constant with variations in contact time (within
FIG. 2. The temperature dependence of the observed (solid lines) and
calculated equilibrium (dashed lines) phenol hydrogenation selectivities
with respect to cyclohexanone and cyclohexanol formation: the dotted line
represents the combined cyclohexanone + benzene selectivity generated
via catalysis.
experimental error) over the range 4 ꢅ 10ꢃ4–18 ꢅ 10ꢃ4
h
and all subsequent catalytic data were obtained at the inter-
mediate value of 9 ꢅ 10ꢃ4 h. Moreover, the short-term deac-
tivation reported for palladium systems (13, 15–17) was not
observed in thisstudywhere catalyticactivityand selectivity
were both essentially time invariant and steady state con-
version was readily achieved and maintained. In the only di-
rectly comparable catalytic study (22), the quoted rate (per
gram of catalyst) of phenol conversion over a higher loaded
(9% w/w) nickel-on-silica was less than half that recorded
in this laboratory. It has, however, been recognised (22) that
phenol hydrogenation activity increases with increasing dis-
persion where the smaller supported nickel particles are the
more active. The catalyst used in this study was prepared by
homogeneous precipitation/deposition and has been shown
to exhibit stronger metal/support interactions than is the
case with impregnated catalysts, which results in the forma-
tion of smaller nickel metal crystallites with a narrow size
distribution (25). The generation of such a well-dispersed
nickel metal phase can be considered to be the source of
the higher reaction rates generated in this study.
erated cyclohexanol as by far the predominant product
(S > 98% ) but with an increase in temperature cyclohex-
anone formation was increasingly promoted and was the
preferred product at the highest temperatures (>540 K)
considered. Similar temperature effects have been reported
for palladium systems (12, 15) but the data generated in this
study are in conflict with a previous report wherein temper-
ature had a negligible effect on selectivity for nickel cata-
–
lysts (22). Hydrogenolysis of the substituent C OH bond
to form benzene is energetically more demanding and was
only initiated and promoted at temperatures greater than
523 K, as has been observed in the hydrogen treatment
of other oxygenated aromatics (31, 32). The overall hy-
drogenation of phenol to cyclohexanol has been viewed
as occurring in a stepwise fashion with cyclohexanone as
a reactive intermediate (12, 13). The equilibrium constants
for the hydrogenation of phenol to cyclohexanone and cy-
clohexanone to cyclohexanol are available in the literature
(15, 33, 34) and are given below in the form of temperature
dependencies:
Variations in reaction temperature, however, had a con-
siderable effect on reaction selectivity, as illustrated in
Fig. 1. At 423 K, the hydrogen treatment of phenol gen-
K1
C6H5OH + 2H2
C6H10O,
ꢀ
ꢁ
PC H
20131
10O
6
K1 =
= 2.7604ꢅ10ꢃ16 exp
,
[1]
(PC H OH )(PH )2
T
6
5
2
K2
C6H10O + H2
C6H11OH,
ꢀ
ꢁ
PC H
7897
11OH
6
K2 =
= 3.6844 ꢅ 10ꢃ7 exp
.
[2]
(PC H O )(PH )
T
6
10
2
Using these equilibrium constants the equilibrium prod-
uct compositions were calculated at representative tem-
peratures, and the resultant reaction selectivities in the
hydrogenation of phenol are compared in Fig. 2 with
those generated in this study. The selectivities generated
from equilibrium constant calculations and experimental
FIG. 1. Selectivity with respect to cyclohexanol (᭿), cyclohexanone
(᭡), and benzene (᭹) formation as a function of temperature.

HYDROGEN TREATMENT OF AROMATIC ALCOHOLS
453
catalysis exhibit the same pattern in that cyclohexanone
formation is promoted with increasing temperature in both
cases. At the lower reaction temperatures (<453 K), the
equilibrium product distribution generated via heteroge-
neous catalysis is essentially the same as that which is calcu-
lated for the homogeneous gas phase system. At higher re-
action temperatures, the two systems diverge and the nickel
catalyst clearly shifts the equilibria shown in Eqs. [1] and
[2] fully to the right, i.e., complete hydrogenation to cyclo-
hexanol; a constant temperature differential of 20 K is ob-
served for cyclohexanol selectivities in the range 40–75% .
The selective formation of cyclohexanone is consistently
lower for the catalytic system, with a marked divergence at
temperatures greater than 523 K. The latter effect can be
accounted for when the additional conversion of phenol to
benzene is incorporated into the selectivity plot. The dot-
ted line in Fig. 2 represents the combined selectivity with
respect to cyclohexanone and benzene formation and this
profile is characterised by a corresponding displacement of
20 K from the calculated line where T ꢁ 490 K. In other
words, in what appears to be an equilibrium hydrogenation
of phenol on the catalyst surface, the selective formation
of the partially hydrogenated cyclohexanone is sacrificed
as the hydrogenolysis step is promoted. The equilibrium
formation of the fully hydrogenated cyclohexanol is, how-
ever, not noticeably disrupted due to this secondary reac-
tion when compared to the theoretical equilibrium. The de-
gree of conversion of the phenol feed is compared to that
which is expected under noncatalysed equilibrium condi-
tions in Fig. 3. The conversion in both cases declines at ele-
vated temperatures but the presence of the catalyst ensures
FIG. 4. The temperature dependence of the observed (solid lines) and
calculated equilibrium (dashed lines) cyclohexanone hydrogenation/dehy-
drogenation selectivities with respect to cyclohexanol and phenol forma-
tion.
a consistently higher degree of conversion (by up to a factor
of 2.1) as a direct result of a greater shift in the equilibrium
to full hydrogenation. The drop in activity with increasing
temperature has been reported previously (12, 13, 15, 16)
and has been attributed to a temperature-induced decrease
of the fraction of the catalyst surface that is covered by
reactants (13). In this study, however, the observed tem-
perature activity dependency mirrors that predicted from
the thermodynamic equilibrium and may be attributed to
the thermodynamic constraints where the catalyst serves
to extend conversion at elevated (>523 K) temperatures
by promoting full hydrogenation to cyclohexanol and hy-
drogenolysis to benzene.
Kinetic and Mechanistic Studies
Cyclohexanone and cyclohexanol were used as feed un-
der reaction conditions identical to those of the phenol hy-
drogenation procedure in order to probe the overall re-
action network. Reaction selectivity and conversions are
compared to the equilibrium calculations in Figs. 3–5. In
FIG. 3. Effect of reaction temperature on the degree of (a) phenol,
(b) cyclohexanone, and (c) cyclohexanol conversion generated in the cata-
lytic step (solid lines) and calculated from equilibrium constants (dashed
lines): x, mole fraction.
FIG. 5. The temperature dependence of the observed (solid lines)
and calculated equilibrium (dashed lines) cyclohexanol dehydrogenation
selectivities with respect to cyclohexanone and phenol formation.

454
SHIN AND KEANE
the case of the hydrogen treatment of cyclohexanone, the
observed selectivities again show a pattern similar to that
predicted using the thermodynamic data in that cyclohex-
anol is the preferred product at lower temperatures with a
shift in selectivity in favour of the dehydrogenated product
(phenol) at higher temperatures. At temperatures in ex-
cess of 500 K, the catalytic system exhibits markedly higher
(differentials up to 50% ) selectivities with respect to cyclo-
hexanol where the profiles are consistently shifted by 30 K
to higher temperatures. It can be stated that the presence of
the nickel catalyst significantly shifts the equilibrium to the
hydrogenated (cyclohexanol) product. The predicted equi-
librium conversion of cyclohexanone falls with increasing
temperature to pass through a minimum at 523 K as a di-
rect consequence of the different temperature dependen-
cies of K1 and K2 where the dehydrogenation step begins
to predominate at higher temperatures and phenol is the
favoured product. It is evident from Fig. 3b that the cata-
lytic system shows the same response, which again suggests
that the nickel promoted hydrogenation/dehydrogenation
process is in equilibrium, but the temperature-related ac-
tivity minimum appears at a higher temperature (ca. 540 K)
due to the overall shift in favour of hydrogenation. Selec-
tivity in terms of the catalytic conversion of cyclohexanol
also mimicked that predicted, as shown in Fig. 5. The dehy-
drogenation reaction occurs stepwise where cyclohexanone
formation is first preferred and complete dehydrogenation
to phenol is increasingly favoured at the higher temper-
atures. Whereas there is a clear thermodynamically pre-
dicted switch in selectivities at ca. 560 K, the nickel cata-
lyst exhibited a consistently higher selectivity with respect
to cyclohexanone formation at every temperature that was
studied. The greater shift in the equilibrium to full hydro-
genation of phenol, revealed in Fig. 2, is accompanied by an
inhibition of the full dehydrogenation of cyclohexanol and
the degree of dehydrogenation of the alcohol is lower than
the predicted values at every temperature that was consid-
ered, Fig. 3c. In addition, the yield of benzene, generated
in the catalytic study, from the three reactants increased in
the order cyclohexanol < cyclohexanone ꢆ phenol, which
is a direct consequence of the increasing concentration of
surface phenol that is available for the hydrogenolysis step.
The kinetics for the hydrogen treatment of phenol may
be represented by the power equation
FIG. 6. Turnover frequency of phenol as a function of phenol partial
pressure at 423 K (᭿) and 573 K (᭹).
and representative logarithmic plots depicting the variation
of TOF with the partial pressure of phenol at 423 and 573 K
are shown in Fig. 6. The computed reaction order with re-
spect to phenol is one for both temperatures. Values of 1 and
ꢃ1 (at 433 K) have been quoted previously for Pd/Al2O3
(13) and Pd/MgO (16), respectively. By comparison, in the
hydrogenation of benzene and methylbenzenes over the
same nickel/silica (26, 30) and nickel/zeolite (35, 36) cata-
lysts, the reaction order with respect to the aromatic was
observed to increase from 0 to 0.4 in the temperature range
393 K ꢀ T ꢀ 523 K. Benzene is known to be adsorbed on
nickel catalyst via ꢀ-bond interactions in which the aro-
matic ring lies parallel to the active surface (37, 38). It has
been shown that the presence of methyl group(s) on the
benzene ring stabilises the adsorbed ꢀ-complex with the
resultant introduction of a higher energy barrier for aro-
matic ring hydrogenation (26, 30, 36, 39). The increase in
the reaction order with respect to the aromatic at higher
temperatures has been attributed to a decrease in the sur-
face coverage by the reactant(s) (26, 30). A temperature
invariant reaction order of one with respect to phenol can
be considered to be diagnostic of a much weaker interac-
tion than has been observed for benzene or methylben-
zene(s). It has been proposed that the stability of the ad-
sorbed ꢀ-complex increases (and reactivity consequently
decreases) with decreasing ionisation potential (40). The
ionisation potential (41) of phenol (8.51 eV) is appreciably
lower than that of benzene (9.24 eV) or toluene (8.82 eV)
and is within the range of values (8.44–8.58 eV) quoted for
the xylenes. In terms of the ꢀ-complex stabilities, the ex-
pected order of decreasing ease of ring hydrogenation is
then benzene > toluene > phenol ꢇ xylenes. The observed
(under identical reaction conditions) TOFs, recorded in
Table 1, reveal a markedly higher rate (by a factor of 2.5)
of phenol hydrogenation compared with benzene. It is in-
structive to note that the hydrogen treatment of benzyl al-
cohol over the same catalyst only generated products of hy-
[4] drogenolysis (toluene and benzene) while the aromatic ring
TOF = kPCm H OH PHn ,
[3]
6
5
2
where k represents the rate constant and m and n represent
the orders of the reaction with respect to the phenol and
hydrogen partial pressures, respectively. The reaction or-
ders were determined by means of logarithmic plots where,
for instance, at constant reaction temperature and partial
pressure of hydrogen,
ꢂ
ꢃ
log TOF = log kPHn + mlog PC H OH
,
6
5
2

HYDROGEN TREATMENT OF AROMATIC ALCOHOLS
455
crease in the contribution of phenoxide/catalyst interac-
tions to the overall reactivity of the system where the effect
is inductive. The additional loss of activity in the conver-
sion of the o-isomer may be rationalised in terms of steric
hindrance and increased ring resonance due to the electron
donation to the ring. The latter effect serves to lower the
reactivity in a coplanar arrangement of the reactant on the
surface where a two-point interaction of the reactant (via
aromatic ring and substituent) is envisaged. Alcoholic so-
lutions of phenol were used as feed where there was no
detectable product resulting from a reaction between phe-
nol and the solvent. The nature of the alcoholic solvent
can, however, have an effect on the ultimate hydrogena-
tion rate as shown in Table 2. While methanol-, ethanol-,
and 1-butanol-based feed generated equivalent turnover
frequencies, there is a definite solvent dependency for the
higher straight chain alcohols. A solvent effect has been
reported previously (20) for a liquid phase batch hydro-
TABLE 1
Comparison of the Turnover Frequencies of a Number of Aromatic
Reactants under Identical Reaction Conditions: T = 473 K
Aromatic reactant
103 TOF sꢃ1
Benzene
Toluene
o-Xylene
m-Xylene
p-Xylene
Phenol
o-Cresol
m-Cresol
p-Cresol
8.9
5.0
3.5
4.2
4.7
22.2
13.9
21.9
18.1
remained intact (32). This observation supports the conclu-
sions of Moreau et al. (42) that hydrogenolysis activity is
higher for slightly electron-donating groups (such as ben-
zyl alcohol, IP = 9.14 (41)) whereas hydrogenation activ-
ity is enhanced by strongly electron-donating substituents
as is the case with phenol. The higher activity and higher
reaction order in the case of phenol suggests a mode of
adsorption different from that of the nonoxygenated aro-
matic reactants. In addition to aromatic ring interactions,
phenol can also adsorb on the active surface via the hy-
droxyl substituent (43, 44) where the aromatic ring is copla-
nar with (43) or perpendicular to (44) the catalysts sur-
face. The sequence of decreasing hydrogenation rate for
the methyl substituted benzene systems, i.e., toluene > p-
xylene > m-xylene > o-xylene, has been attributed to an in-
creasing strength of interaction with the surface due to a
combination of electronic and steric effects. The methyl-
substituted phenols (cresols) exhibit lower turnover fre-
quencies than phenol where the observed order of decreas-
ing reactivity, phenol ꢁ m-cresol > p-cresol > o-cresol, is in
agreement with that reported previously for the liquid
phase batch hydrogenation using palladium catalysts (20).
It should be noted that the rate sequence, in the case of
the aromatic alcohols, follows the order of increasing pKa
values (41) for phenol and the m- and p-isomer, i.e., phenol
(9.99)ꢈm-cresol (10.00) < p-cresol (10.26). The exception
to this trend is the o-isomer (pKa = 10.26) which exhibits
lower TOFs. Such a relationship between rate and the acid
ionisation constant suggests that in the case of phenol, p-,
and m-cresol, the ease of hydrogenation can be related to
the stability of the associated phenoxide ion in solution. If
this represents a valid interdependence, then the gas phase
hydrogenation is promoted by phenoxide/catalyst interac-
tions. Ring substituents can have marked effects on the re-
activity of phenols by means of inductive and/or resonance
effects where a methyl substituent is “electron-releasing”
towards the aromatic ring and serves to destabilise phe-
noxide ion-contributing structures (45). The lower rate of
p-cresol hydrogenation may then be attributed to a de-
–
genation using Pd C catalysts where the rate was lower in
polar solvents, and this effect was attributed to a solvation
around the phenoxide ion which weakened adsorption. The
opposite trend is evident in this study of a gas phase con-
tinuous flow system. If the equilibrium formation of the
phenoxide ion in solution is greater in the more polar alco-
holic solvents (methanol, ethanol, and 1-butanol) and the
dissociation equilibrium is less favoured as the polarity of
the alcohol decreases (in the order: 1-pentanol; 1-hexanol;
1-heptanol) and this effect is prevalent and is maintained
in the gas phase, then the contribution of phenoxide ion
interactions is greater in the former case. It would appear
from the activity sequence in Table 2 that the interaction
of phenol with the catalyst via the substituent oxygen is
the principal source of phenol activation. The same must
be true for p- and m-cresol, and the role of direct aromatic
ring activation is more significant for the o-isomer due to
steric effects.
The variation of process selectivity with phenol partial
pressure, illustrated in Fig. 7, sheds some light on the possi-
ble mode(s) ofphenoladsorption. The equilibrium constant
calculations predict a selectivity invariance with phenol
concentration, and this is indeed the case at lower reaction
temperatures where the only products formed are the result
TABLE 2
Effect of Varying the Nature of the Alcoholic Solvent on
the Overall Turnover Frequency of Phenol: T = 473 K
Solvent
102 TOF sꢃ1
Methanol
Ethanol
1-Butanol
1-Pentanol
1-Hexanol
1-Heptanol
2.2
2.1
2.3
1.5
1.2
1.0

456
SHIN AND KEANE
in Table 3 where the data for phenol processing is also
included for comparative purposes. Both p- and m-cresol
show similar selectivity trends to phenol in that selective
partial hydrogenation and hydrogenolysis are preferred at
higher temperatures; in contrast, 1,2-methylcyclohexanone
formation from o-cresol is less favoured. Hydrogenolysis or
hydrodeoxygenation, in the case of p- and m-cresol, was ini-
tiated at the lower temperature of 473 K and was promoted
with a greater degree of selectivity than was observed for
the phenol feed. Such an increase in hydrogenolysis sele-
ctivity can be attributed to an electronic effect where elec-
tron donation from the methyl substituent to the ring serves
–
to weaken the C O bond. The order of reactivity of cresols
with respect to hydrodeoxygenation has been reported in
the literature to be m- > p- > o-cresol (46, 47). In this study,
m- and p-cresol exhibited essentially the same selectivity
in terms of toluene formation and the electronic effect can
be considered to be equivalent, under the stated reaction
conditions, for CH3 positioned p- and m-cresol to the hy-
droxyl function. The o-cresol reactant exhibited lower hy-
drogenolysis reactivity than phenol, an observation that is
both borne out (48) and in direct conflict (49) with earlier
reports. Steric hindrance to adsorption due to the proxim-
ity of the methyl group in the o-isomer has been invoked
in a number of hydrogenolysis systems (46, 47, 50) to ex-
plain the lower reactivity of this isomer. As the selectivity
pattern observed in this study for the processing of o-cresol
reactant displayed such a deviation from that recorded (in
Table 3) for phenol and the remaining isomers, it is pro-
posed that the nature of the reactant/catalyst interaction is
different in the former case. The presence of o-CH3 must
impede the nonplanar arrangement or rather cant the aro-
matic ring towards a planar adsorption mode, the result
FIG. 7. Selectivity with respect to cyclohexanol (᭿), cyclohexanone
(᭡), and benzene (᭹) formation as a function of phenol partial pressure
at (a) 423 K and (b) 573 K.
of hydrogenation. At higher reaction temperatures cyclo-
hexanol formation was again largely independent of phenol
partial pressure but the formation of cyclohexanone was
certainly enhanced at the direct expense of benzene where
the partialpressure ofphenolexceeded 0.05atm. Thisseem-
ing competition between partial hydrogenation and hy-
drogenolysis augments the data presented in Fig. 2 and may
be accounted for in terms of aromatic/catalyst interactions.
Neri et al. (13) have proposed that the coplanar orienta-
tion of phenol on the catalyst favours complete hydrogena-
tion to cyclohexanol whereas the nonplanar arrangement
favours a stepwise hydrogenation. It is not possible from
the catalytic data alone to conclusively identify the predom-
inant surface arrangement. The data generated in this study
support surface hydrogenation/dehydrogenation equilibria
where the catalyst promotes complete hydrogenation. The
nonplanar mode of adsorption involving a single point of in-
teraction between the hydroxyl substituent and active sur-
face should certainly promote hydrogenolysis. It is tenta-
tively proposed that cyclohexanol formation is promoted
by hydrogen addition to phenol adsorbed in a coplanar ori-
entation with probable surface interactions involving both
the hydroxyl substituent and the aromatic ring. Incomplete
hydrogenation to cyclohexanone also results from such a
TABLE 3
Effect of Reaction Temperature on Process Selectivity in the Hydro-
gen Treatment of Phenol and the Three Isomeric Forms of Cresol
S%
ketonea
S%
alcoholb
S%
aromaticc
Reactant
T = 473 K
T = 573 K
Phenol
o-Cresol
m-Cresol
9
35
20
12
91
65
80
88
0
0
<1
<1
surface arrangement or alternatively/additionally via the p-Cresol
stepwise addition of hydrogen to phenol adsorbed in a
nonplanar orientation where at higher temperatures hy-
drogenolysis is promoted in direct competition with par-
tial hydrogenation. If this is the case then the selectivity
dependence of phenol partial pressure suggests that hy-
drogenolysis, and by inference a nonplanar adsorption of
phenol, is favoured at low concentrations of phenol. Reac-
tion selectivity at two representative temperatures for the
hydrogen treatment of the three cresol isomers is presented
Phenol
o-Cresol
m-Cresol
p-Cresol
53
12
46
40
31
80
33
38
16
8
22
22
a
Cyclohexanone from phenol and 1,2-, 1,3-, or 1,4-methylcyclohexa-
none from o-, m-, and p-cresol, respectively.
b
Cyclohexanol from phenol and 1,2-, 1,3-, or 1,4-methylcyclohexanol
from o-, m-, and p-cresol, respectively.
c Benzene from phenol and toluene from o-, m-, and p-cresol.

HYDROGEN TREATMENT OF AROMATIC ALCOHOLS
457
FIG. 9. Effect of hydrogen partial pressure on the observed (sym-
bols) and calculated equilibrium (lines) selective formation of cyclohex-
anol from phenol at 423 K (᭿, solid line), 523 K (᭡, dashed line), and
573 K (᭹, dotted line).
FIG. 8. Turnover frequency of phenol as a function of hydrogen par-
tial pressure at 423 K (᭿), 523 K (᭡), and 573 K (᭹).
being that o-cresol exhibits a greater deal of interaction be-
tween the catalyst and the aromatic nucleus. Such a surface
arrangement will limit the extent of hydrogenolysis and the
increase in ring resonance due to electron donation from
the methyl substituent means higher energy requirements
for complete hydrogenation, which results in the observed
shift in selective alcohol formation to higher temperatures.
The reaction order with respect to hydrogen was deter-
mined by varying the partial pressure of hydrogen at a
constant partial pressure of phenol, employing helium as
a diluent and representative logarithmic plots are shown in
Fig. 8. The computed reaction order increased with a rise
in reaction temperature from 423 K (0.5) to 523 K (1.1)
and 573 K (1.7). A reaction order of 2 and 1.3 at 433 K
has been reported for Pd/Al2O3 (13) and Pd/MgO (13, 16),
respectively, where the latter catalyst system exhibited an
increase in hydrogen dependency to 1.8 at 473 K (16). Such
a temperature-induced increase in hydrogen reaction or-
der has been observed for a range of nickel-promoted aro-
matic hydrogenation reactions (26, 30, 35, 36, 51). It has
been demonstrated elsewhere (26) that such variations in
the experimentally determined reaction order are due to
the temperature dependence of the surface concentration
of hydrogen. A recent chemisorption and TPD study of hy-
drogen on Ni/Al2O3 (52) revealed different hydrogen ad-
sorption states which are strongly dependent on tempera-
ture. The greater dependence of the phenol hydrogenation
rate on hydrogen partial pressure may then be attributed
to an increased mobility and temperature induced loss of
surface reactive hydrogen. The effect of hydrogen partial
pressure on the selective formation of cyclohexanol and cy-
clohexanone is illustrated in Figs. 9 and 10, respectively. The
observed selectivity with which cyclohexanol was produced
decreased with decreasing hydrogen partial pressure and
this effect was more pronounced at higher reaction temper-
atures (Fig. 9). The loss of cyclohexanol selectivity coincides
with an increasingly favoured production of cyclohexanone
mation was essentially independent of hydrogen concentra-
tion. These observations are in qualitative agreement with
the earlier finding that cyclohexanone is promoted at lower
hydrogen/phenol molar ratios (22). However, at temper-
atures greater than 423 K, the observed selectivities with
respect to cyclohexanone and cyclohexanol formation are
consistently lower and higher, respectively, than those pre-
dicted thermodynamically. These relationships again illus-
trate the role of the catalyst in shifting the overall equilib-
rium towards complete hydrogenation.
A reaction mechanism that accounts for the observed
rate data is summarised as
KA
–
C6H5OH + A
2H2 + 4S
C6H5OH A
[5]
[6]
[7]
[8]
K4H
K10
–
4H S
–
–
–
C6H5OH A + 4H S
C6H10O A + 4S
C6H10O(g) + A
K1D
–
C6H10O A
FIG. 10. Effect of hydrogen partial pressure on the observed (sym-
bols) and calculated equilibrium (lines) selective formation of cyclohex-
anone from phenol at 423 K (᭿, solid line), 523 K (᭡, dashed line), and
at lower hydrogen partial pressures (Fig. 10); benzene for- 573 K (᭹, dotted line).

458
SHIN AND KEANE
K2H
K20
In the hydrogenation of cyclohexanone to cyclohexanol,
the temperature dependence of the equilibrium constants
generated by catalysis can likewise be represented by
–
H2(g) + 2S
2H S
[9]
[10]
[11]
–
–
–
C6H10O A + 2H S
C6H11OH A + 2S
C6H11OH(g) + A,
K2D
ꢀ
ꢁ
–
PC H
6644
C6H11OH A
11OH
6
K2ꢉ =
= 1.0 ꢅ 10ꢃ5 exp
,
[20]
PC H O PH
T
6
10
2
where A and S denote adsorption sites for the aromatic
and hydrogen (adsorbed dissociatively) reactants, respec-
tively. The model, in keeping with the earlier study (26)
of benzene and methylbenzene(s) hydrogenation, assumes
that both reactants adsorb at different sites and the sur-
face reactants and products exist in quasi-equilibrium with
the gas phase. The surface hydrogenation reaction pro-
ceeds stepwise where partially hydrogenated cyclic alco-
hols are not isolated in the product stream, tautomerisation
of cyclohexene-1-ol to cyclohexanone must occur readily,
and the formation of cyclohexanone and cyclohexanol is in
equilibrium. The generation of benzene is taken to be an
irreversible hydrogenolysis step which lowers the overall
selective formation of cyclohexanone:
where
K20 K2H K2D
K2ꢉ =
.
[21]
K1D
The deviation in equilibria induced by the nickel catalyst
in a continuous flow system can be seen from a direct com-
parison of Eq. [18] with Eq. [1] and Eq. [20] with Eq. [2].
CONCLUSIONS
The data presented in this paper support the following
conclusions:
k1
(i) The hydrogenation of phenol to cyclohexanol pro-
ceeds stepwise via the formation of cyclohexanone as a re-
active intermediate. The observed rate data are explained
by a mechanism involving equilibrated adsorption and hy-
drogenation steps where the catalyst serves to shift the equi-
librium in favour of complete hydrogenation.
–
–
–
C6H5OH A + 2H S → C6H6 A + H2O(g) + 2S [12]
K3D
–
C6H6
A
C6H6(g) + A
[13]
Quasi-equilibria for the adsorption and desorption steps
yield, for instance,
(ii) Phenol can adsorb on the catalyst at the aromatic
ring and/or the hydroxyl substituent where the latter inter-
action promotes hydrogenolysis to benzene at the direct
expense of cyclohexanone formation.
(iii) Steady state conversion, which is independent of
contact time, is readily achieved with no appreciable de-
activation after extended (>500 h) use.
ꢂ
C6H5OH
KA
=
=
,
[14]
[15]
PC H OH ꢂA
6
5
ꢂH4
PH2 ꢂS4
K4H
,
2
and
(iv) The hydrogenation rate exhibits a temperature in-
variant first order dependence on phenol concentration and
a lower turnover frequency in less polar alcoholic solvents,
which is taken to be diagnostic of a predominantly weak
interaction with the catalyst that involves phenoxide ion
formation.
(v) The reaction order with respect to hydrogen in-
creases with increasing temperature due to an induced loss
of catalytically active hydrogen from the catalyst surface
where lower hydrogen partial pressure promotes the selec-
tive formation of cyclohexanone.
PC H O ꢂA
6
10
K1D
=
,
[16]
ꢂ
C6H10
O
where ꢂ denotessurface coverage. The equilibrium constant
for the surface partial hydrogenation of phenol to cyclohex-
anone is given by
ꢂC H O ꢂ4
6
10
S
K10 =
[17]
ꢂC H OH ꢂ4
6
5
H
(vi) The sequence of decreasing hydrogenation rate,
phenol ꢇ m-cresol > p-cresol > o-cresol, is attributed to in-
ductive and steric effects where the selectivity pattern gen-
erated by the o-isomer deviates significantly from the other
aromatic alcohols and suggests a different mode of adsorp-
tion/activation.
and the temperature dependence (where 423 K ꢀ T ꢀ
573 K) of this surface-controlled equilibrium is related to
the gas phase reactant partial pressures by
ꢀ
ꢁ
PC H
17960
10O
6
K1ꢉ =
= 4.9 ꢅ 10ꢃ14 exp
,
[18]
[19]
PC H OH P2
T
6
5
H2
ACKNOWLEDGMENT
where
K1ꢉ = K10 KA K4H K1D
.
E.J.S. acknowledges partial financial support from the British Council.

HYDROGEN TREATMENT OF AROMATIC ALCOHOLS
459
REFERENCES
25. Keane, M. A., Can. J. Chem. 72, 372 (1994).
26. Keane, M. A., and Patterson, P. M., J. Chem. Soc., Faraday Trans. 92,
1. Joglekar, H. S., Samant, S. D., and Joshi, J. B., Water Res. 25, 135 (1991).
1413 (1996).
2. van der Vaart, D. R., Marchand, E. G., and Bagely-Pride, A., Critic. 27. Hoke, J. B., Gramiccioni, G. A., and Balko, E. N., Appl. Catal. B:
Rev. Environ. Sci. Tech. 24, 203 (1994). Environmental 1, 285 (1992).
3. Noordally, E., Richmond, J. R., and Tahir, S. F., Catal. Today 17, 359 28. Keane, M. A., and Webb, G., J. Catal. 136, 1 (1992).
(1993). 29. Coughlan, B., and Keane, M. A., J. Catal. 123, 364 (1990).
4. Cooper, C. D., Clausen, C. A., Tomlin, D., Hewett, M., and Martinez, 30. Keane, M. A., J. Catal. 166, 347 (1997).
A., J. Hazard. Mater. 27, 273 (1991).
31. Keane, M. A., Bull. Soc. Chim. Belg. 104, 63 (1995).
32. Keane, M. A., J. Mol. Catal. 118, 261 (1997).
33. Chitwood, H. C., Fitzpatrick, J. T., Fowlerand, G. W., and Freure,
B. T., Ind. Eng. Chem. 44, 1696 (1952).
5. Dial, C. J., Chem. Eng. Prog. 82, 16 (1986).
6. Gioia, F., J. Hazard. Mater. 26, 243 (1991).
7. Kalnes, R. C., and James, R. B., Environ. Prog. 7, 185 (1988).
8. Bhargara, D. S., Bhatt, D. J., and Panesar, P. S., J. Environ. Eng. 109, 34. Cubberley, A. H., and Mueller, M. B., J. Am. Chem. Soc. 69, 1535
369 (1983).
(1947).
9. Ong, S. K., and Bowers, A. R., J. Environ. Eng. 116, 1013 (1990).
35. Coughlan, B., and Keane, M. A., Zeolites 11, 12 (1991).
10. Shieh, W. K., Puhakka, J. A., Melin, E., and Tuhkanen, T., J. Environ. 36. Coughlan, B., and Keane, M. A., Catal. Lett. 5, 101 (1990).
Eng. 116, 683 (1990). 37. Minot, C., and Gallezot, P., J. Catal. 123, 34 (1990).
11. Bezuglaya, E. Y., Shchutskaya, A. B., and Smirnova, I. V., Atmos. 38. Coughlan, B., and Keane, M. A., J. Chem. Soc., Faraday Trans. 86, 3961
Environ. 27, 773 (1993). (1990).
12. Talukdar, A. K., Bhattacharyya, K. G., and Sivasanker, S., Appl. Catal. 39. Stanislaus, A., and Cooper, B., Catal. Rev. Sci. Eng. 36, 75 (1994).
A: General 96, 229 (1993).
13. Neri, G., Visco, A. M., Donato, A., Milone, C., Malentacchi, M., and
Gubitosa, G., Appl. Catal. A: General 110, 49 (1994).
40. Vo¨lter, J., Hermann, M., and Heise, K., J. Catal. 12, 307 (1968).
41. Dean, J. A., “Handbook of Organic Chemistry.” McGraw–Hill, New
York, 1987.
14. Srinivas, S. T., Lakshmi, L. J., and Rao, P. K., Appl. Catal. A: General 42. Moreau, C., Joffre, J., Saenz, C., and Geneste, P., J. Catal. 122, 448
110, 167 (1994).
(1990).
15. Itoh, N., and Xu, W.-C., Appl. Catal. A: General 107, 83 (1993).
43. Taylor, R., and Lundlum, K. H., J. Phys. Chem. 76, 2882 (1972).
16. Galvagno, S., Donato, A., Neri, G., and Pietropaolo, R., J. Chem. Tech- 44. Tanabe, K., and Nishizaki, T., in “Proceedings, 6th International
nol. Biotechnol. 51, 145 (1991).
17. Narayanan, S., and Krishna, K., Appl. Catal. A: General 147, L253
Congress on Catalysis” (G. C. Bond, P. B. Wells, and F. C. Tompkins,
Eds.), p. 863. Catalysis Society, Letchworth, 1977.
(1996).
45. Brown, W. H., “Organic Chemistry.” Saunders, New York, 1995.
18. Srinivas, S. T., and Rao, P. K., J. Chem. Soc. Chem. Commun., 33 (1993). 46. Odebunmi, E. O., and Ollis, D. F., J. Catal. 80, 56 (1983).
19. Tang, U. M., Hung, M. Y., and Jiang, Y. Y., Macromol. Rapid Commun. 47. Rollmann, L. D., J. Catal. 46, 243 (1977).
15, 527 (1994).
20. Higashijima, M., and Nishimura, S., Bull. Chem. Soc. Jpn. 65, 2955 49. Bredenberg, J. B., Huuska, M., Ra¨ty, J., and Korpio, M., J. Catal. 77,
(1992). 242 (1982).
21. Dodgson, I., Griffen, K., Barberis, G., Pignatora, F., and Tauszik, G., 50. Gevert, S. B., Eriksson, M., Eriksson, P., and Massoth, F. E., Appl.
48. Weigold, H., Fuel 61, 1021 (1982).
Chem. Ind., 830 (1989).
Catal. A: General 117, 151 (1994).
22. Narayanan, S., and Sreekanth, G., Ind. J. Technol. 31, 507 (1993).
23. Narayanan, S., and Sreekanth, G., React. Catal. Lett. 51, 449 (1993).
51. Smeds, S., Murzin, D., and Salmi, T., Appl. Catal. A: General 125, 271
(1995).
24. Gonzalez-Velasco, J. R., Gutierrez-Ortiz, J. I., Gonzalez-Marcos, J. A., 52. Smeds, S., Salmi, T., Lindfors, L. P., and Krause, O., Appl. Catal. A:
and Romero, A., Kinet. Catal. Lett. 32, 505 (1986). General 144, 177 (1996).