l.W. Hightower, W.N. Delgass, E. Iglesia and A.T. Bell (Eds.

11th International Congress on Catalysis - 40th Anniversary

Studies in Surface Science and Catalysis, Vol. 101
1996 Elsevier Science B.V. All rights reserved.

751

The montmorillonite catalyzed production of phenol and acetone from

cumene hydroperoxide.
W.A. de Groot, E.LJ. Coenen, B.F.M. Kuster and G.B. Marin
Laboratorium voor Chemische Technologie, Schuit Institute of Catalysis, Eindhoven
University of Technology, PO Box 513, 5600 MB Eindhoven, The Netherlands.
The kinetics of the montmorillonite F-20 catalyzed production of phenol and acetone from
cumene hydroperoxide have been investigated. Batch experiments were carried out at initial
cumene hydroperoxide concentrations of 50 - 500 mol m'3 at 303 K in cumene with dried
catalyst in the absence and presence of 5 - 30 mol m'3 of typical feed impurities such as 2phenyl-2-propanol, dicumyl peroxide, acetophenone and a-methyl styrene. A reaction
network is presented. Regression analysis of the experimental data, using a mUlti-response
Marquardt-algorithm, allowed the data to be adequately described by a model based on a
Langmuir-Hinshelwood mechanism. The large inhibiting effect of water could be best
incorporated by a negative order dependence of 2,13 0.08 on the water concentration in the
acid catalyzed steps.

1. INTRODUCTION
The industrial production of phenol and acetone mainly takes place by the acid catalyzed
decomposition of cumene hydroperoxide, which is formed by oxidation of cumene. Hock and
Lang [1 J invented this process in 1944 and in 1952 the first production plant was put in
operation, Nowadays 90 % of the phenol and 70 % of the acetone production proceeds via this
route [2]. The decomposition is typically carried out at atmospheric pressure and temperatures
from 348 to 368 K with acetone as refluxing solvent with 0.1 to 2 wt% sulphuric acid.
The oxidation of cumene is performed with molecular oxygen and is autocatalytic. As byproducts 2-phenyl-2-propanol, acetophenone and minor amounts of dicumyl peroxide are
formed. These by-products are also fed to the plant in which cumene hydroperoxide is
decomposed to phenol and acetone. Next to sulphuric acid other strong mineral acids like
sulphur dioxide and perchloric acid are employed as catalyst, although in Russia also acidic
ion exchangers are used. The selectivity obtained is usually high, but due to the presence of 2phenyl-2-propanol, some dicumyl peroxide, a-methyl styrene and dimers are formed together
with minor by-products as mesityl oxide, cumyl phenols and cumyl phenyl ether.
Although the decomposition step with mineral acids is fast and selective and only little
corrosion occurs, heterogeneous catalysts can achieve the same or better activity and
To whom correspondence should be addressed

752

selectivity without any corrosion. Heterogeneous catalysts have the advantage that less
process steps are necessary, because the separation of the catalyst is much easier. Many solid
acids have been claimed in the patent literature as potential catalysts for the production of
phenol and acetone, including zeolites [3,4], heteropoly acids [5], acidic ion exchangers [5],
clays [6], mineral acids deposited on inert carriers [7] and combinations of these solid acids
[5,8]. A kinetic investigation however is still missing in the open literature. This paper reports
on such a study with a montmorillonite clay as catalyst.

2. EXPERIMENTAL
2.1. Materials
Cumene hydroperoxide (CHP) of technical quality was obtained (Sigma, 80 %) and was
purified by procedures described by Hock and Lang [1] and Fil'makova et. al. [9]. The resulting purity was 94 %, the other 6 % being cumene. The latter was employed as solvent during
kinetic experiments and therefore did not constitute a problem. Phenol (PH), dicumyl
peroxide (DCP), acetophenone (ACP), 2-phenyl-2-propanol (PP), <X-methyl styrene (AMS)
(all from Aldrich), cumene (Janssen, 99%) and acetone (AC) (Merck, p.a.) were obtained in
the highest purity available and used as received.
The highly activated montmorillonite clay Filtrol Grade F-20, kindly made available by
Engelhard was chosen as catalyst. To avoid potential transport limitations the catalyst was
sieved and the fraction particles with a diameter smaller than 38 ~m was used for the
experiments. The volume averaged catalyst diameter was 20 ~m as measured by the Coulter
LS 130 apparatus. The catalyst was dried at 1 bar and 363 K during 16 hours before use.

2.2. Procedures
Table 1. Range of experimental conditions
Reaction procedure and conditions.
Temperature
303 K
Reactions were performed in a thermoSolvent
Cumene
stated batch reactor of 100 cm3 under
Catalyst
Filtrol F-20
nitrogen atmosphere with cumene as
Batchtime
0 - 1oo s
solvent. In a typical experiment solvent
C cat
0.25-0.50 kg m-3
was pipeted in the reactor and the reactor
CCHP,O
0 - 500 mol m-3
was brought at the designated temperature.
CAc,o, CPH,o
0 - 120 mol m-3
The reactor was brought under nitrogen by
Cpp,o, CAMS,O, Cocp,o, CACP,O 0 - 30 mol m- 3
repeatedly switching between vacuum and
CH20 ,O
2 - 8 mol m- 3
nitrogen atmosphere. Reactant(s) and
toluene, which was used as internal
standard, were added in the required
quantity, while a small overpressure of nitrogen was maintained to avoid introduction of
oxygen. A sample was taken by a syringe punctured through a septum and subsequently
catalyst was added, marking the start of the experiment. At suitable intervals samples were
taken from which catalyst was removed by syringe filters. The range of experimental
conditions applied for this investigation is shown in Table 1. Reactions were performed with
single components as well as with combinations of cumene hydroperoxide with one of the

753

other components. Blank experiments showed that a mixture of cumene, acetone and phenol
did not undergo any reactions. No reactions were observed in the absence of catalyst.
Experiments in which cumene hydroperoxide was repeatedly added to the reactor, using the
same batch of catalyst, showed that the disappearance rate decreased a little on every new
batch of cumene hydroperoxide, but it was concluded that for a single batch deactivation of
the catalyst could be neglected. Theoretical calculations and experiments with catalysts with
varying average particle diameter showed that experiments were in the intrinsic kinetic
regime, i.e. were not influenced by transport phenomena
HPLC-UV analysis. For the quantitative analysis of the components in the reaction mixture
a HPLC method has been developed. Samples diluted to appropriate concentrations were
separated over a reversed phase column (Zorbax ODS-2, 150x4.6 mm, 5 !J.1ll particles) with
gradient elution of 1 ml min-I acetonitrile/water mixtures. This gradient started with 15 %
acetonitrile for 1 minute and was subsequently raised to 100 % during the next 20 minutes,
where it was kept for another 4 minutes. Detection and quantification was performed by UV
absorption at 254 nm. The components eluted in the following order (retention times in
minutes): acetone (2.2), phenol (7.7), 2-phenyl-2-propanol (9.6), acetophenone (10.5), cumene
hydroperoxide (11.0), toluene (15.3), a-methyl styrene (17.1), cumene (17.8), dicumyl
peroxide (20.7), phenoxy-isopropoxy-isopropyl cumyl peroxide (PllCP) (21.7) and the 0.methyl styrene dimers (DIM) (21.4, 21.9 and 22.1). Calibration factors were obtained easily
for all components except PllCP, which presence will be discussed later and the a-methyl
styrene dimers, which are not commercially available. Quantification of the total amount of
these dimers was done by assuming 100 % selectivity of the reaction of pure a-methyl styrene
to its dimers, which is reasonable, for the HPLC-chromatogram was free of other components.
For the quantification ofPllCP the calibration factor of dicumyl peroxide was used.
HPLC-MS analysis. Identification of PllCP was performed by HPLC with on-line MS
detection (Fisons Trio2000 apparatus, MassLynx software). The HPLC conditions were
identical to the conditions used for HPLC-UV. Two ionization techniques were used. On
ElectroSpray (ES) ionization the post column flow of 1 mlImin was split in a ratio 1:50 and to
the resulting 20 J.1l1min was added 10 J.11Imin of 30 mM ammonium acetate in
acetonitrile/water (1/1) to enable the formation of (M + NH4t adducts. A capillary voltage of
4.2 kV and a cone voltage of 24 V were applied at a temperature of 60 0c. Particle beam
Electron Impact (EI) ionization with an electron energy of 70 eV was employed to obtain
fragmentation patterns.
Karl-Fischer analysis. Water concentrations were determined by automatic Karl-Fischer
coulometric titrations performed on a Metrohm 684 KF Coulometer with a 703 Ti stand. Of
all determinations the water concentration could be measured the least accurately due to the
low water concentrations, causing larger relative errors. Furthermore it appeared to be very
difficult to set the initial water concentration, due to differing humidity of the reactants.
2.3. Parameter estimation
The regression of the experimental data was based on the maximum-likelihood criterion, as
outlined by Froment and Hosten [10]. Parameter estimates were obtained by applying the least
squares criterion to the observed product concentration, [P), and the calculated product
concentrations, [P), i.e. by minimizing the residual sum of squares:

754

S(~) =

f f cr i ([PIe]j - [i\ l; )([P,l; - [P\]i) ~ Min

ld

1e=1 1=1

(1)

i=\

v being the number of responses, i.e. CHP, AC, PH, PP, AMS, DCP, DIM, PIICP and H20,
see Figure 1, m being the number of experimental time sampling points and dd being the (k,l)
elements of the inverse of the covariance matrix of the experimental errors on the responses
[Pl. The latter was estimated from a preliminary parameter estimation based upon a minimization of Eq. 1 with dd being the .elements of the unity matrix, i.e. assuming equal variances
for the errors of the different responses and neglecting any correlation. Minimization was
achieved with a multi-response Marquardt algorithm [11]. The calculated products
concentrations, [P], were obtained by integration of the corresponding continuity equations:
d[PIe ] - R
dt - v.PIe

k= I,... ,v

(2)

with R v Plc the net rate of production of the corresponding component.

The parameter estimates were tested for statistical significance by means of their
approximate individual t values. The significance of the global regression was expressed by
means of the ratio of the mean regression sum of squares to the mean residual sum of squares,
which is distributed according to F [12]. A high value of the F ratio corresponds to a high
significance of the global regression over the whole range of investigated conditions. The
adequacy of the mathematical models used for the regression was tested by analysis of
residuals. Model discrimination was based on this adequacy, on the physical meaning of the
kinetic parameter estimates and on the statistical testing of the significance of the kinetic
parameters and of the global regression.

3. REACTION NETWORK
3.1. Decomposition of pure cumene hydroperoxide
Cumene hydroperoxide decomposed to acetone and phenol with a final selectivity
approximating 100 %, see Figure 2a. The activity of the dried montmorillonite F-20 clay is
comparable to sulphuric acid when used in the same solvent with the same water
concentration and related to the number of protons available for catalysis. The latter amounts
to 0.65 mmol g.l for F-20, based on the value reported for acid activated montmorillonite
clays by Benesi [13].
During decomposition of cumene hydroperoxide a component was observed, which sofar
had not been reported in literature. Water was formed together with this component. Due to
this increase in the water concentration the reaction rate strongly decreased, as is generally the
case with clay catalyzed reactions [14]. Furthermore the homogeneous decomposition is also
known to be sensitive to the water concentration [15-17]. The yield of this component and
thus of the water formed, increased at higher initial cumene hydroperoxide concentrations, see
Figure 2b, leading to a strong decrease in the decomposition rate. At higher initial cumene
hydroperoxide concentrations also Significant differences between the acetone and phenol
concentrations were observed, indicating incorporation of acetone in this component. The

755
PbeDoxy-isopropoxy-isopropyi ClDDyi peroDcle

latter is formed as
an
intermediate
o-~o-~o-o-~
and
as
no
CH, CH,
b:;=I
influence on the
Pbenol
C _ bJdroperoxide
CHP
CHI'~ __
final
selectivity
(PH)
(CUP)
/~AC
.AC~;(
was observed it
-lip
H+
+~
+
decomposes
to
CH,
Dicumyl peroDde
phenol
and
acetone.
The
inter~-61, ~ ~
mediate compoO~o-H
~ Dimen
nent eluted at 21.7
- CH,
-11,0
CH,
(DIM)
2-Pbeuyl-2-propmol
a-methyl styrene
minutes in the
(PP)
I, Inte..-diates
I,
(AMS)
HPLC chromatogram.
This
0_g.H,
~
relatively
large
61,
- CH,
retention
time,
Figure l. Network of observed global reactions and intermediates via
comparable
to
which reactions are suggested to proceed.
dicumyl peroxide
and the a-methyl
styrene dimers, suggests an apolar structure. HPLC-MS with ElectroSpray ionization revealed
a molecular mass of 344 a.u. Electron Impact ionization showed a fragmentation pattern, with
a m1z ratio of 135 being dominant, which corresponds to the structure of intermediate II in
Figure I, suggesting that this structure is part of the intermediate.
Based on the above the structure and the formation mechanism depicted in Figure 3 are
proposed for this component, tentatively named phenoxy-isopropoxy-isopropyl cumyl
peroxide (PIICP). PIICP is supposed to be formed by addition of the intermediate II to water
(PDCP)

0,

/+

OOH
0 , E'o-o-H
~O't'!:::'~d
H')

0,

Or

0'

6r-----------,

150r-----------,10

6
4
2
2000
Time [5)

3000

_ ___ _ _ _

o .. CQIP.
mM
.. .. ..= 50
.. .. .. _.

Left Y
o CHP
0

AC

PH

~-~~.~-~~~~-

RightY

<)

COI.= 225 mM

t.

COI... = 500 mM

P!ICP

1000

2000

3000

Time [5)

Figure 2. a. Concentration vs. batchtime during the decomposition of cumene hydroperoxide

in cumene at 303 K with 0.5 kg.m-3 F-20. b. Yield of PIICP vs. batchtime for different initial
cumene hydroperoxide concentrations. Symbols: experimental, curves: calculated with the
parameter estimates reported in Table 3.

756

followed by an acid catalyzed

nucleophilic addition of acetone and
finally a condensation with cumene
hydroperoxide. The reaction network
presented in the upper half of Figure
1 now

summarizes

the

reactions

ir->. CH,

'(I ~-()-()-H

+ Ii'

~'1

Oo-r-'

+8,0

-- O
-

1'=1 OI,

CH,

----+ f

CH,

o-~+

CH,

CH,

'\ o-e-OH

ell,

+H,O

+ Ii'

observed starting from pure cumene

Oo-~OH
hydroperoxide.
The quantification of PllCP based
CH,
PllCP
on the calibration factor of dicumyl
peroxide appeared to be reasonable.
The calculated amount of PllCP
+H,O
based on this factor was comparable
Figure
3.
Structure
and
formation
mechanism
to the increase in the water
concentration as well as to the proposed for phenoxy-isopropoxy-isopropyl cumyl
difference between the acetone and peroxide (PllCP).
phenol concentration and allowed to
close the mass balance. The increase
in the yield from almost zero at an initial cumene hydroperoxide concentration of 50 molm 3
to more than 5 % at 500 molm 3 as shown in Figure 2b can be explained by a higher order in
cumene hydroperoxide c.q. a first order in water, which are both probable, based on the
mechanism postulated.
3.2. Influence of the presence of feedstock components and reaction products on the
decomposition of cumene hydroperoxide
The presence of the most abundant by-products of the oxidation of cumene influence the
decomposition of cumene hydroperoxide to different degrees as will be outlined below.
Acetophenone. Acetophenone does not react nor influence the reactions that occur and
therefore does not appear in the network in Figure 1.
2-phenyl-2-propanol. While 2-phenyl-2-propanol dehydrates to a-methyl styrene in the
absence of cumene hydroperoxide, in the presence of cumene hydroperoxide, dicumyl
peroxide is formed and hardly any a-methyl styrene was observed, see Figure 4. During both
reactions water is formed, which decreases acid activity. So an increase in the 2-phenyl-2propanol concentration leads to a decrease in the decomposition rate of cumene hydroperoxide, compare the decomposition curves of cumene hydroperoxide in Figures 2a and 4a.
Dicumyl peroxide. Decomposition of dicumyl peroxide gives phenol, acetone, a-methyl
styrene and dimers as products, see Figure 5b. During this decomposition no cumene
hydroperoxide and 2-phenyl-2-propanol were observed, suggesting that these components are
not formed as intermediates, or that the consecutive reactions to resp. phenol and acetone and
a-methyl styrene and dimers are much faster than the reactions to the expected intermediate
components. When dicumyl peroxide and cumene hydroperoxide were treated simultaneously,
cumene hydroperoxide decomposed to phenol and acetone as if no dicumyl peroxide was
present. Dicumyl peroxide on the contrary did not react at all, see Figure 5a. An explanation is
that cumene hydroperoxide and its reaction products adsorb so strongly on the catalyst that
dicumyl peroxide is prevented from adsorption and thus from reaction.

757
25,-------------,

150 ,r - - - - - - - - - - - , 2 0 Left Y

faD
o

------------- :.-- 15

ow
AC

.------- ..

b
~20

'0

S15~
c

..,

0' 0

,_

'

pp

AMS

DIM

~1O

AMS

OCP

PIICP

Time [sI

t.>

'A

5~
00

Time [sI

8
10
(Thousands)

Figure 4. Concentration vs. batchtime for a: cumene hydroperoxide (Co=125 mol m- 3) mixed
with 2-phenyl-2-propanol (Co=13 mol m- 3) in cumene at 303 with 0.5 kg m-3 F-20, b: 2phenyl-2-propanol (Co=22 mol m- 3) under same conditions. Symbols: experimental, curves:
calculated with the parameter estimates reported in Table 3.
a-methyl styrene. The results of the experiments with a-methyl styrene are comparable to
the results with dicumy1 peroxide. Alone a-methyl styrene dimerizes, while in the presence of
cumene hydroperoxide no reaction of a-methyl styrene takes place and the decomposition of
cumene hydroperoxide is not influenced by the presence of a-methyl styrene.
Conclusion. The presence of PP, DCP and AMS results in the occurrence of reactions
depicted in the lower half of Figure 1, which in total represents the network of all reactions
observed during the production of phenol and acetone from industrial feed with F-20 as
catalyst in cumene as solvent. Typical other impurities sometimes found during the sulphuric
acid catalyzed decomposition, like cumyl phenyl ether, cumyl phenols and mesityl oxide
[2,15] were not observed.

20r-------------,
o

ow

CHP

AC

PH

758

4. MODELLING OF THE REACTION NETWORK

Several models based on elementary reaction steps have been proposed in order to obtain
more insight in the reaction mechanism. Discrimination between the tested models with the
criteria outlined in paragraph 2.3
resulted in the mechanism depicted in
Table 2. Elementary steps and reaction paths Table 2. The corresponding parameter
corresponding to the considered global reactions
estimates are shown in Table 3. They
allow an adequate description of the
O"AO"BO"CO"DO"EO"F
experimental data, viz. Figures 2, 4 and
110000
1 CHP+*
~CHP*
5 and can be used for industrial reactor
1 1
2 CHP*
- - 11* + H 20
design purposes, when taking into
3 1,* + H 20
_ _ AC* + PH* 1 - 1 account their temperature dependency.
-I 0 o 0 0 0
4 AC + *
~ AC*
The final model is generally a
-I 0 o 0 0 0
5 PH + *
~ PH*
Langmuir-Hinshelwood mechanism and
of adsorption
of the
consists
6 I,*+H 20+AC+CHP
~ PllCP*+ HzO - 1 -I components on the acid sites, denoted
7 PllCP + *
~ PllCP*
0 -I 1 0 0 0
by an asterisk, preceding the acid
8PP+*
~PP*
000010
catalyzed steps. The adsorption is
9 pp*
~ Iz* + HzO
- 1 1 assumed to be in quasi-equilibrium. The
10 Iz* + CHP* ~ DCP* + *
- - 1 catalyst is assumed to consist of uniform
II DCP + *
~ DCP*
0 0 0 0 -I 0
acid sites. Upon acid catalyzed
12 Iz*
~ AMS*
- 1 - -I
rearrangements the two intermediates II
13 AMS + *
~ AMS*
0 0 0 -I 0 2
or 12, which are depicted in Figure 1, are
14 Iz* + AMS* _ _ DIM + 2*
- - - 1
formed. For these intermediates the
_ _ AC + PH
A CHP
pseudo-steady state approximation was
B 2CHP + AC - - PllCP + HzO
applied. Therefore no absolute values,
but ratios of the parameters k3 and ~
C PllCP + ~O - - CHP+2AC+PH
D PP
~ AMS + HzO
and k..9, k lO , k l2 and k l4 were obtained.
The most critical element in the
~ DCP + HzO
E PP + CHP
F 2AMS
_ _ DIM
tested models was the incorporation of
the inhibition by water. A simple model
in which acid sites on which water had
been adsorbed were assumed to be inactive could not adequately describe all experimental
results. This was already expected from what is known about clay catalysis, because by
adsorption of water the acid sites do not lose their activity completely, but the proton activity
is decreased by dissipation of the cation charge [18]. The adsorption capacity of one acid site
furthermore is not limited to one water molecule. Upon adsorption of more water the activity
of the acid sites continues to decrease [19]. By this process an unknown distribution of sites
on which 0 or 1 or 2 or more water molecules have been adsorbed all with differing activities
originates:

(3)
in which MZ+ represents a proton or strongly polarizing cations like Al 3+. The dependence of
the activity on the number of adsorbed water molecules is not known exactly, but the activity

759
is known to decrease strongly with
Table 3. Parameter estimates with their
increasing amount of adsorbed water.
individual 95% confidence intervals for the
Solomon et. al. [19] for example found
model with the highest F ratio. (F=7220)
for a kaolinite clay that the Hammett
Parameter
Regression analysis
acidity of the strongest sites decreased
estimate.
kz K I I m3 kg-I S-I
from Ho<-8.2 on the dried clay, via Ho=-3
4.4.10- 1 0.71O- 1
on a clay with I % (w/w) water to Ho=+3
~ K7 I mb kg- I mor S-I 5.4.10- 1 2.4.10- 1
z
z
k91 mol kg-I SI)
with 10 % (w/w) water adsorbed.
1.810 1.5lO
Because of these unknown relations
k..JO I mol kg' S-I
1.2.10- 1 0.2.10- 1
k..l2 K13 I m~ kg- I S-I
the inhibiting effect of water was
2.3.10-3 0.3.10-3
l
approximated by introducing a partial
4.2.10-3 1.2.10-3
K.t + Ks I m> mor
l
reaction order n for water of the acid
Ksl m> mor
2.3IO- z 0.8-IO- z
catalyzed steps, i.e. steps 2, -6, 9, -10 and
KII 1m> mor l
5.6-10-2 1.4IO- z
n 1-12 in Table 2. For simplicity the partial
-2.13 0.08
reaction order for water was assumed to
3.310-6 0.810-6
be equal for all these steps. This
approximation worked out very well as
k..~lzI m> mor l
4.84.3
can be seen from the 95% confidence
1.9-10- 1 0.3.10- 1
kJO KI/klzl m> mor l
intervals, which reveal good statistical
l
kl 4 KI3/k 12 I m> mor
2.1-10- 1 0.2.10- 1
significance for the parameter estimates.
The order in water turned out to be rather
negative, Le. -2.13 but its small 95% confidence interval shows that this parameter estimate is
highly significant. On basis of the F ratios for the different models tested it could furthermore
be concluded that this model was the model which could by far the most accurately describe
the experimental results.
The rate of reactions 6 and -6 was best described by a first order dependence in all
components involved, suggesting that all steps outlined in Figure 3, with the exception of the
first are in quasi-equilibrium. Due to the involvement of water in step -6 the overall order of
water in this step is one less, i.e. n-l. Because in all experiments equimolar mixtures of
acetone and phenol were used only the sum of their adsorption constants K.t+Ks could be
determined. The adsorption constants of cumene hydroperoxide, PIICP and a.-methyl styrene,
i.e. Kt, K7 and K 13 , could not significantly be estimated, which means that they are low
compared to K.t+Ks, K9 and Klz. Consequently only values for kzKI, k..6K7 and kIZK 13 were
obtained.

5. CONCLUSIONS.
The kinetics of the montmorillonite catalyzed production of phenol and acetone from
cumene hydroperoxide in the absence and presence of typical feed impurities have been investigated. The activity of dried Filtrol Grade F-20 is comparable to sulphuric acid and 100 %
selectivity towards phenol and acetone was achieved. A sofar unknown side product has been
identified. Water strongly inhibited all observed reactions. The presence of 2-phenyl-2-propanol lead to a decreased reaction rate due to an increase in the water concentration which was
caused by reaction of 2-phenyl-2-propanol with cumene hydroperoxide to dicumyl peroxide

760

and water. No dehydration of 2-phenyl-2-propanol to a-methyl styrene was observed in the

presence of cumene hydroperoxide. Acetophenone, dicumyl peroxide and a-methyl styrene
did not influence the decomposition of cumene hydroperoxide. The reactions of dicumyl
peroxide and a-methyl-styrene: decomposition to phenol, acetone, methyl styrene and dimers
resp. dimerization were inhibited when carried out in the presence of cumene hydroperoxide.
A kinetic model is presented which could adequately describe the experimental data over a
broad range of concentrations with a single set of physically reasonable parameter values.
Kinetic parameters were obtained by regression of the data with the multi-response Marquardt
algorithm. In this model the water inhibition was approximated by a negative order in water
for the protonation steps.
ACKNOWLEDGEMENT.

The authors are grateful to T. de Joode and P.J.W. Schuyl of Unilever Research Laboratory
Vlaardingen for their assistance in carrying out the HPLC-MS analyses. This research was
financial supported by Unilever Research Laboratory Vlaardingen.
REFERENCES.
I.
2.
3.
4.
5.
6.
7.
8.
9.
10.
II.
12.
13.
14.
15.
16.
17.
18.
19.

H. Hock and S. Lang, Ber. 77B (1944) 257.

W. Gerhatz et. aI. (eds.), Ullmann's encyclopaedia of industrial chemistry, 5th edition, Al (1985) 79 and
A19 (1991) 299.
U. Romana. M.G. Clerici, G. Belussi and F. Buonomo, Catalyst for the selective decomposition of cumene
hydroperoxide and process using it, European Patent No. 203632 (1986).
J.F. Knifton and P.-S.E. Dai, Method for the production of phenol/acetone from cumene hydroperoxide,
European Patent No. 429807A2 (1991).
J.F. Knifton and J.R. Sanderson, Method for the production of phenol/acetone from cumene
hydroperoxide, US Patent No. 4898995 (1990).
J.F. Knifton, Method for the production of phenol/acetone from cumene hydroperoxide, US Patent No.
4870217 (1989).
J.F. Knifton and N.J. Grice, Method for the production of phenol/acetone from cumene hydroperoxide, US
Patent No. 4876397 (1989).
J.F. Knifton, Method for the production of phenol/acetone from cumene hydroperoxide, US Patent No.
4898987 (1989).
.
L.A. FiI'makova, V.A. Galegov, I. Yeo Pokrovskaya and I.I. Kubasova, SOy. Chern. Ind., (1974) 679.
G.F. Froment and L.H. Hosten in J.R. Anderson and M. Boudart (eds.), Catalytic Science and
Technology, Springer Verlag, Berlin, 1981, Chap. 3.
DW. Marquardt, J. Soc. Ind. Appl. Math., II (1963) 431.
N.R. Draper and H. Smith, Applied Regression Analysis, Wiley, New York, 1966.
H.A. Benesi, J. Phys. Chern., 61 (1957) 587.
M.M. Monland and K,V. Raman, Clays Clay Min., 16 (1968) 393.
P. Cavalieri d'Oro, C. Perego and L. Raimondi, 3rd World Congress on Chemical Engineering, Sept. 2125, Tokyo, 1986.
G. Burtzlaff, U. Felber, H. Hiibner, W. Pritzkow and W. Rolle, J. Prakt. Chern., 28 (1965) 305.
G. Jung and G. Just, J. Prakt. Chern., 313 (1971) 377.
J.M. Adams, T.V. Clapp and D.E. Clement, Clay Min., 18 (1983) 411.
D.H. Solomon, J.D. Swift and A.J. Murphy, J. Macromol. Sci., Chern., 61 (1971) 587.