ISSN 0965-5441, Petroleum Chemistry, 2007, Vol. 47, No. 4, pp. 273–284. © Pleiades Publishing, Ltd., 2007.

Original Russian Text © V.M. Zakoshansky, 2007, published in Neftekhimiya, 2007, Vol. 47, No. 4, pp. 301–313.

The Cumene Process for Phenol–Acetone Production

V. M. Zakoshansky
ILLA International, LLC 4756 Doncaster Ct., Long Grove, IL 60047, USA
e-mail: vlzak@illallc.com
Received January 11, 2007

Devoted to the memory of the inventors of this chemical process,

Russian scientists R.Yu. Udris, M.S. Nemtsov, and B.D. Kruzhalov

Abstract—The history of designing and the evolution of the process for the joint production of phenol and ace-
tone, which is now the basic industrial method for the manufacture of these products, the state of the art and
parameters of its key steps in industrial facilities among world manufacturers of phenol are presented. Special
attention is given to the consideration of ways of improving the characteristics of the process and its environ-
mental safety on the basis of the recent advances in science and to methods for the solution of scientific
and practical questions of the reconstruction and renovation of out-of-date industrial phenol manufacturing facil-
ities.
DOI: 10.1134/S096554410704007X

HISTORICAL INFORMATION The cumene process for phenol–acetone (PA) man-

ufacture is one of the unique high-volume petrochemi-
The purpose of this historical paper is to correct cal processes in which simultaneously two products
unintentionally introduced errors concerning the dis- (phenol and acetone) are obtained from one reactant
covery of the reaction of acid-catalyzed decomposition (cumene), each of the products finding useful applica-
of cumene hydroperoxide (CHP), whose authorship is tion; however, which is most important, their reaction
attributed to respected scientist Hoch, who indepen- with one another in the joint synthesis results in bisphe-
dently invented this process in 1944, i.e., two years later nol, a chemical that forms the basis of the large-scale
than R.Yu. Udris. production of polycarbonate plastics, which are known
In 1942, Rudolf Yur’evich Udris (Russia) was the for their unique consumer properties.
first in the world to isolate cumene hydroperoxide The chemical beauty of the reactions proceeding in
(CPH) from cumene oxidation products. the cumene process and their tremendous practical
In 1942, Rudolf Yur’evich Udris (Russia) discov- importance have drawn the attention from researchers
ers the reaction of acid-catalyzed CHP conversion into for decades, although the complexity and danger of the
process for a long time did not allow any significant
phenol and acetone.1 improvement in its characteristics over those reached
In 1943, Mark Semyonovich Nemtsov, Rudolf by Th pioneers: the yield of the unutilizable byproduct,
Yur’evich Udris, and Boris Dmitrievich Kruzhalov the so-called phenol tar, remained at the level of
with co-authors develop an industrial process for 150−200 kg/t of phenol, and the yield of useful α-meth-
cumene oxidation into CHP (patent no. 106906, Rus- ylstyrene (AMS) was 40–50% of theoretical values.
sia, 1947) and acid CHP degradation into phenol and Giving the authors and the first founders of the
acetone (patent no. 106992, Russia, 1947). cumene process for PA manufacture their due, it should
In 1944, Hoch independently discovers the reaction be noted that the design of the modern PA process has
of acid-catalyzed CHP decomposition. naturally required practically the complete abandon-
ment of the strategy and philosophy of the process
In 1949, Two industrial units for phenol–acetone developed by the pioneers in all its parts without exclu-
manufacturing plants for the production are com- sion.
missioned for the first time (Dzerzhinsk, Russia, and Even the chemistry, let alone the production engi-
Canada, British Petroleum.2 neering of the key steps, underwent a radical change, a
1 The
development which allowed the modern cumene pro-
authorship of the discovery of the reaction is wrongly attrib- cess for phenol manufacturing to dominate over alter-
uted to Hoch, who independently discovered, in 1944, the reac-
tion of acid CHP decomposition. native methods; to attain a level of selectivity very close
2 The process was developed and implemented independently by to the theoretical value, 97 mol %; and, accordingly, to
these authors. reduce the yield of phenol tar to 25–35 kg/t of phenol,

273
274 ZAKOSHANSKY

OFF-GAS CLEANUP C
WITH ZEOLITES ISOLATION U
OF COMMERCIAL ACETONE å
E
ALDEHYDE COMMERCIAL N
BINARY CUMENE
ALKALINE H2O ACETONE AMS HYDROGENATION E
AGENT
OXIDATION
NaOH
INCLUDING CUMENE
FRESH
OBTAINED FROM CUMENE
COAL-TAR NH4OH
BENZENE H2O H2
NaOH
air
ISOLATION
OF COMMERCIAL AMS
DISTILLATION STEP OF NEUTRALIZATION
OF UNREACTED AND REMOVAL OF HYDROXYACETONE COMMERCIAL
CUMENE AMS
crude CHP
CHP DECOMPOSITION
RECYCLING ACETONE
ALKALINE
AND
OXIDATIVE
H2O ∆– AGENTS

H2SO4 TO COMMERCIAL
DEPHENOLIZATION PHENOL
H2O &
NH4OH
PHENOL
TAR TWO-STEP
CATALYTIC
PURIFICATION OF PHENOL

Fig. 1. Simplified flow chart of the phenol process. Radical key engineering changes in the ILLA process relative to the conventional
processes and those of other licensiates.

and to increase the yield of AMS to ~90 mol %. This and its parameters fundamentally. Designers of new
breakdown was attained within the last 15–20 years phenol manufacturing processes are few in number glo-
owing to the close combination of profound experimen- bally. Note that only one research center (ILLA Inter-
tal and theoretical studies of all steps of the phenol pro- national) is engaged in extensive scientific research
cess in practice. aimed at the radical improvement of the cumene pro-
cess, the solution of the “painful” points of the tradi-
tional process, the design of systems for management,
PHENOL–ACETONE MARKET and the control, design, and licensing of their own phe-
AND KEY PLAYERS ON THIS MARKET nol processes and their commissioning.
The rate of phenol production will remain at the
same level (~5 rel. %/y) in the coming decade. Up to
2010, it is planned to build new facilities with a general TECHNOLOGY
total capacity of ~1.5 million t/year for phenol, a value A simplified scheme of the ILLA phenol process used
that makes up ~20 % of the current capacity to discuss the current process is represented in Fig. 1.
(7.5 million t/year). Practically all steps and components have experienced
Some licensors of phenol manufacturing processes radical changes during the last 10–15 years, thus mak-
in the world are listed below: ing it possible to develop an essentially new PA manu-
(1) Callog (United States), the base process of Brit- facturing process.
ish Petroleum (BP); The technical characteristics of the process are
(2) Lammus (United States), ILLA International determined by its three key steps: cumene oxidation,
(United States) and the GE process (based on the BP CHP decomposition, and phenol cleaning of impurities
process); to ensure the quality of phenol. The parameters of aux-
(3) UOP (United States), the Allied process; iliary steps of the process are practically identical in all
types of licensed processes.
(4) Mitsui (Japan) basically develops processes for
its own plants;
(5) ILLA International (United States–Russia) Step of Cumene Oxidation to Cumene Hydroperoxide
licenses its own the processes. The cumene oxidation reaction follows the radical
It is easy to conclude that most of these companies chain mechanism. Therefore, the general problem of
do not license their own developments; thus, they this process is the absence of effective methods to over-
upgrade the old technologies mainly at the engineering come the inhibiting effect of impurities that are present
level, which does not change the nature of the process in both initial cumene (sulfur compounds, such as mer-

PETROLEUM CHEMISTRY Vol. 47 No. 4 2007

 THE CUMENE PROCESS FOR PHENOL–ACETONE PRODUCTION 275

Selectivity, mol %

96 (a) 97 (b)

with NH4OH with NH4OH

95 96

94 95
without NH4OH without NH4OH
93 94

92 93

5 10 15 20 25 30 5 10 15 20 25 30
CHP, wt %

Fig. 2. Change in total selectivity of processes: (a) “wet” oxidation with and without ammonia; (b) “dry” oxidation with and without
ammonia.

captans, thiophene, etc.) and are formed during its oxi- reduction in pH to 2–3, and, consequently, increased
dation. the danger of the cumene oxidation process and the
Despite almost 65 years of existence of the process, CHP concentration step.
the problem of inhibition remained rather pressing. (d) Alkali supply to dry (water-free) cumene oxida-
Researches at ILLA proved that the traditional tion reactors partly solved the problem of the neutral-
approach to controlling the inhibition by means of ization of organic acids (safety issue), but has led to a
NaOH—via its feeding to reactors or the pretreatment marked loss of the selectivity both at the cumene oxida-
of the feedstock with alkali—was a mistake. On the one tion and at the CHP concentration steps.
hand, this approach solved the problem of attaining a Escape from this circle of problems that seem unre-
reasonable conversion of cumene, but, on the other solvable simultaneously was found by ILLA [1]
hand, it simultaneously led to a significant decrease in (Ammoxidation process, Figs. 2a, 2b) via the arrange-
the selectivity of the process. Other methods used to ment of the so-called chemical trap, the scavenging of
control inhibition (see below) also turned out to be formaldehyde by ammonia introduced for the purpose
practically ineffective: and its transformation into hexamethylenetetramine
(a) Application of the so-called wet (emulsion) oxi- (HMTA). Owing to this arrangement, the route for the
dation of cumene (Fig. 2a), in which the inhibiting role transformation of ëç2é into formic acid, which medi-
of formed phenol (BP process) is supposed to be ates the acid decomposition of CHP and the appearance
reduced with the aid of water and Na2CO3 supplied to of an inhibitor (phenol) in the cumene oxidation prod-
reactors, has not completely solved the problem and ucts, is barred:
resulted (due to Na2CO3 recycling) in the formation of CH3 CH2O HCOOH
a new inhibitor (NaHCO3) whose inhibiting activity is
equivalent to that of phenol. Moreover, the production Ar–CH(CH3)2 HCOOH
ArOH + CO(CH3)2
engineering has become more complicated, leading to
a marked increase in the reactor volume, i.e., an 6CH2O + 4NH3 C6H12N4.
increase in capital costs. In this case, it is quite important that NH3 does not
(b) Application of a physical method (significant form salts with CHP, unlike NaOH, which gives the Na
decrease in the temperature of the process, UOP pro- salt of CHP. The decomposition of the sodium salt of
cess) made it possible to increase the selectivity of the CHP when the alkali is used to control the inhibition in
process, but simultaneously led to the necessity of the conventional process is a source of the formation of
using larger reactors; a larger volume will make the dimethyl phenyl carbinol (DMPC) and acetophenone
process extremely inflexible in response to a change in (ACP) as principal byproducts and leads to a decrease
the loading and maintenance of the regime, as well as in the selectivity of the cumene oxidation process.
potentially dangerous owing to a huge amount of CHP The nonstandard “chemical trap” method made it
present in reactors. possible to increase the selectivity of cumene oxidation
(c) Abandonment of alkali supply to the reactors of to 94.5 mol %, to increase the productivity of reactors
the dry (water-free) oxidation of cumene simplified the by 15–20 rel. % without a loss in selectivity, to reduce
process somewhat, but, did not solve the problem of the threefold the size of the reactors in comparison with the
enhancement of selectivity, simultaneously leading to a physical method of inhibition control, and to solve the

PETROLEUM CHEMISTRY Vol. 47 No. 4 2007

276 ZAKOSHANSKY

[O2]/[O2–]av S, rel. %
0.9 95
0.8
94
0.7
0.6
93
0.5
0.4
92
0.3
0.2 91
90
0.1
10 15 20 25 30 35 40 45 50 55 60
Specific CHP yield (kg/m3)/h

Fig. 3. Dependence of the ratio of the concentration of dissolved oxygen to its equilibrium value in the liquid phase on the specific
yield of CHP.

safety problem by running the reaction in the pH sharply reduces the formation of phenol, the inhibitor
range 7–8. of the process [1].
The natural abandonment of the section reactors In addition to the aforementioned, the combination
designed and realized by the pioneers, as well as the of the “chemical trap” technique with a number of other
shift of the cumene oxidation process toward the cas- chemical and engineering methods made it possible to
cade of in-series connected reactors, is no more than the solve the problem of the use of cumene contaminated
requirement of the kinetics of consecutive and parallel with various sulfur-containing impurities. This combi-
side reactions. Accordingly, new plants that are under nation allowed a selectivity of 94–94.5 mol % to be
construction use a cascade of reactors and the old attained, a value that is very close to the selectivity of
cumene oxidation facilities with a battery of parallel the aforementioned process, and the use of much
low-selectivity single plate reactors, as a rule, are mod- cheaper coal-tar benzene.
ernized into the cascade. A further increase in the selectivity of the process is
possible, in principle, but, is uneconomical because of
As applied to a cascade of reactors, the result of the both inadequately high power expenditures for the
optimally organized cumene oxidation process, consid- organization for the recycling of the returnable
ering the CHP yield from a unit volume, its selectivity, streams–while cumene conversion decreases—and a
and minimal power consumption at the CHP concentra- significant growth in capital costs for the construction
tion step, is represented in Fig. 3. As a result of the con- of very large reactors.
ducted studies and calculations, it was found that the
ratio of the concentration of dissolved oxygen to its The industrial process of cumene oxidation to CHP
equilibrium value for various industrial conditions for has undergone significant changes from the time of its
the realization of an oxidation reaction is within the creation and practically attained a selectivity of
range 0.2–0.9. Industrial experience shows that the 94.5−95 mol %, which is determined by the theory.
higher the ratio of the concentration of dissolved oxy-
gen to its equilibrium concentration, the higher the Off-Gas Purification Step
selectivity of cumene oxidation. However, this is
attained by a decrease in CHP removal from a unit vol- The sorption method for off-gas cleaning with car-
ume of a reactor, i.e., by the reduction of the process bon adsorbents has become morally obsolete. Serious
temperature and, accordingly, by a strong increase in defects of this method are the absence of an opportunity
the reactor volume, leading to greater capital costs. For to remove methanol from offgases and a short lifetime
this reason, most industrial reactors operate at values of of in adsorbent (1–1.5 years).
CHP yield from 30 to 50 kg/m3 per hour, obviously In our opinion, the method of purification on zeolite
directing the reaction toward a decrease in selectivity. adsorbents (ILLA process) is preferable; it has success-
The only exception is the ammoxidation process for dry fully solved the main problems:
cumene oxidation, in which a selectivity of about
95 mol % is attained at a CHP yield of 35–40 kg/m3 per (a) total sorption capacity is comparable to that of
hour. This selectivity is not due to an increase in the carbons;
oxygen concentration in the reaction medium and is (b) the degree of removal of cumene and acids is
owing to the application of the chemical trap, which equal to that in carbons;

PETROLEUM CHEMISTRY Vol. 47 No. 4 2007

 THE CUMENE PROCESS FOR PHENOL–ACETONE PRODUCTION 277

∆S, abs. %
1.2
1.0
Conventional process
0.8
0.6
0.4
ILLA process
0.2
0
20.0 20.5 21.0 21.5 22.0 22.5 23.0 23.5 24.0
CHP yield, (kg h)/m3

Fig. 4. Losses in CHP concentration depending on the CHP yield in both conventional and ILLA processes.

(c) the required sorption capacity for methanol, overall process parameters in general become unsatis-
which is absent in carbons; factory. That is why numerous investigators of the phe-
(d) the lifetime is at least 3 years (presumably, 7); nol process and phenol manufacturers since the time of
(e) the exclusion of yearly shutdowns and expenses the design of the phenol process and the commissioning
for adsorbent recharging. of the first facilities in 1949 in Russia (Dzerzhinsk) and
Canada (by British Petroleum) have deliberately
Thus, the problem of environmentally hazardous searched for ways of increasing the selectivity of the
gas emissions during the cumene process is success- given process; the yield of the valuable byproduct
fully resolved. (AMS), which is one of the criteria of selectivity, was
35 to 50% of the theoretical value. Correspondingly, the
Cumene Hydroperoxide Concentration Step yield of nonutilizable byproducts, so-called phenol tar,
varied from 300 to 150 kg/t of phenol.
Loss in selectivity at this step is 0.1 to 1.2 abs. % and The chemical transformations proceeding during
is predetermined by the chemical idea of the cumene the decomposition of crude CHP (CCHP) are extremely
oxidation process, rather than the engineering design of diverse and complex.
the concentration step. At low pH values in cumene oxi-
dation reactors, the CHP concentration step becomes It is rather difficult to choose a way that will be use-
dangerous. ful for maximizing the yields of the three desired prod-
ucts (phenol, acetone and AMS).
When NaOH is used for the treatment of recycling
cumene streams or for feeding cumene oxidation reac- Since the CHP degradation reaction is accompanied
tors, the loss of CHP (and, hence, cumene) increases to by the evolution of a huge amount of heat (486 kcal/kg
~1.2 abs. %. However, the problems of chemical loss of of CHP, which is equivalent to the adiabatic increase in
CHP and safety problems have been successfully temperature by ~700°C), the degradation reaction of
solved in the last decade owing to the use of NH3 at the CHP itself takes a few seconds and it is difficult to
cumene oxidation step. It is seen from Fig. 4 that losses equilibrate heat release and heat withdrawal within this
significantly decrease in the CHP concentration step period of time. Therefore, the inventors of the homoge-
conducted according to the traditional and the neous process of CCHP transformation (Fig. 5) were
improved (ILLA) process; i.e., in our opinion, the envi- forced to increase the residence time of products in the
ronmental protection problems pertaining to the CHP reactor and in several series arranged water-cooled heat
concentration step of the process are optimally solved. exchangers.
As a result, the requirements on the conditions and
regimes of the realization of the basic reaction have con-
Crude CHP Degradation Step flicted with conditions and requirements for the realiza-
It is this step, not the cumene oxidation step, that tion of side reactions. Moreover, the approach of devel-
determines the identity of the phenol process as a whole opment engineers to decompose CHP by 100% and very
from the viewpoint of both its selectivity and safety, fast (3–5 s) in combination with a huge heat effect of this
and the formation of trace impurities that complicate reaction have made the created process extremely dan-
the manufacture of high-quality commercial products. gerous in operation.
The good parameters of the CHP decomposition step In the process that suggests the withdrawal of the
reasonably well compensate for the poor characteristics heat of the reaction by evaporating acetone (BP pro-
of the oxidation step and vice versa. As a result, the cess, Fig. 6), heat release and heat withdrawal are

PETROLEUM CHEMISTRY Vol. 47 No. 4 2007

278 ZAKOSHANSKY

Degradation
products CHP
cw
500
400 Heat
P-1 released
300 in the reactor Heat removed
in the heat exchanger
200

H2SO4 100

0 1 2 3 4 5 6 7 8
Time, Ò

Fig. 5. Heat release and heat withdrawal in the one-step process for homogeneous CHP decomposition.

Reflux acetone
2–2.5 t/t of phenol

T = 79 °C = const
CHP
H2O
H2SO4
Products
Recycling acetone
from the step of rectification of CHP
of 10–12 rel. %/t of CHP decomposition

Fig. 6. Schematic of the one-step CHP decomposition process using heat withdrawal by vaporized acetone.

almost automatically balanced by the vaporized ace- so-called ebullating-acetone process, manages to
tone. To keep this balance, i.e., for the sake of the safety achieve these goals. It seemed that the first designers of
of the process, a very high concentration of the catalyst these two technologies for CHP decomposition have
(H2SO4) was maintained, thus ruling out the possibility initially found optimum conditions of these two basi-
of attaining a high selectivity of the process. Unfortu- cally different processes and it is, in principle, impossi-
nately, this innovation [2, 3] not only solved the afore- ble to improve the selectivity.
mentioned problems of low selectivity, but also sharply To resolve these problems, ILLA abandoned the
reduced safety in comparison with the single-step pro- concept of the processes designed by the trailblazers.
cess [4]. Within 20 years, since 1980, a fundamentally new pro-
Over almost 50 years, the search for ways of cess of CHP decomposition has been created and real-
increasing the selectivity of the process (lowering the ized at many plants in the world:
catalyst (H2SO4) concentration, variation of the type of
acid catalyst, change in the residence time of products (1) On the basis on the detailed kinetic study, the
in the reactor, the lowering or elevation of temperature, one-step process was transformed (1980) to a two-step
variation in the composition of the reaction medium by version.
various additives, etc.) was actually unsuccessful: it (2) The idea of the superfast and complete (100%)
was impossible in practice to increase the yield of AMS decomposition of CHP was rejected as erroneous and
and, consequently, to decrease that of phenol tar. Nei- opposite approaches [5−7] were offered:
ther the homogeneous CHP degradation technology
(the version of the process first developed by Udris, —slow CHP decomposition;
Nemtsov, and Kruzhalov) nor another, radically differ- —incomplete CHP decomposition—the CHP con-
ent technology designed by British Petroleum (BP), the version per pass of at most 65%;

PETROLEUM CHEMISTRY Vol. 47 No. 4 2007

 THE CUMENE PROCESS FOR PHENOL–ACETONE PRODUCTION 279

—distribution of CHP conversion over three succes- despite a substantial growth in the acetone concentra-
sive first-step reactors; tion in the reaction medium, and an amount of MOX
—conversion of the residual amount of CHP in a decreases 10–20-fold in comparison with conventional
second-step reactor. processin.
(3) A radical change in the composition of the reac- (6) The reactions of the irreversible formation of
tion medium for CHP decomposition, i.e., the transfer polycondensed phenols, which were ~20 rel. % of total
of the process from the equimolar acetone/phenol ratio phenol “tars” (waste products) in conventional process-
to the acetone : phenol : cumene = 1.2 : 1 : 0.2 reaction ing, do not occur.
medium [6]. The indicated changes made on the basis of the
(4) Control of CHP conversion in the first step of the detailed study of the influence of the medium composi-
process and control of the safety and selectivity of the tion on the kinetics and mechanism of the principal [5]
process with the aid of an additional reactor, a calorim- and side [9] reactions revealed the reasons behind the
eter [6, 7]. attempts of various researchers throughout the world to
(5) The organization of the second steps of the pro- improve the parameters of the CHP decomposition pro-
cess as an adiabatic reactor of replacement for the trans- cess. As a result of these studies, it was found that:
formation of both: CHP nondecomposed in the first (1) The reaction medium (PA containing H2SO4) has
step; DMPC to AMS, and also dicumyl peroxide syn- a high acidity (Fig. 7), which is determined by the
thesized in the first step to the main products of the pro- S
cess [6, 7]. Hammett acidity function H0 (more exactly, H 0 ), and
(6) The realization of transformations of products in this predetermines the extremely high rate of CHP
a reactor of the second step: decomposition;
—in the reaction medium different from that of the (2) The CHP decomposition reaction and most side
first step [7]; reactions pertain to those of specific, not general, acid
—application of a “soft” acid catalyst (NH4HSO4) catalysis, i.e., they are determined by the acidity func-
S
instead of “hard” H2SO4 [6, 7]. tion H 0 , not by H2SO4 concentration.
(7) The transfer of the process on computer start-up S
[7, 8]. (3) The acidity value h 0 of the reaction medium,
(8) The creation and realization of a three-leveled when the acetone/phenol ratio is varied [10], as well as
protection of the process instead of a single-leveled when very small amounts of additives of protic (for
one. example, ç2é [10], alcohols, etc.) and aprotic solvents
(9) Computer (on-line) control and protection of the (for example, cumene) can vary by dozen of times,
process against factors disturbing a regime and the opti- leading to quite significant changes in the rate constants
mizing the selectivity of the process [8]. of the CHP decomposition reaction and other reactions
of the process.
Regarding the current process for CHP decomposi-
tion, it is interesting that a change in processing is (4) At a variable composition of the reaction medium
caused by a change in the chemistry of the main and (acetone/phenol/cumene/H2SO4concentration/ç2é con-
side reactions: S
H SO
centration) based on the h 0 data and temperature effect
(1) The side reaction CHP 2 4 DMPC + AMS, S
which leads to a loss of the target products, is ruled out. on the h 0 data, the optimal composition and the regime
(2) To protect DMPC from conversion to bypro- have been selected to realize the first and second steps
ducts, practically quantitative synthesis of dicumyl per- of a commercial run:
oxide (DCP) is realized in the first step, and the revers- —various (in weight) additives of acetone and water
ible reaction is shifted to DCP : CHP + DMPC in the first and second steps;
DCP + H2O. —the increased cumene content;
(3) Reversible reaction DMPC AMS + H2O in —optimal H2SO4 concentration in the first step and
the first step is shifted to DMPC. replacement of H2SO4 by NH4HSO4 in the second step;
(4) Irreversible conversion of DMPC to byproducts —the decreased temperature in the first step provid-
of the process (AMS dimers and cumylphenols) ing a strictly certain and safety value of CHP conver-
2DMPC DI,II,III + H2O sion to DCP and PA, and increased temperature in the
DMPC + phenol o-, p-CP second step, providing together with “soft” acid cata-
is practically completely stopped in the first step and lyst NH4HSO4 increase in acetone and cumene contents
minimized in the second step to the degree when the and a high selectivity of DCP and DMPC conversion to
yields of the byproducts decrease by a factor of 7–10 the desired products.
compared to conventional processing. All this has allowed ILLA to implement a process
(5) The reaction of the formation of mesityl oxide having a high level of safety for today and the highest
(MOX) from acetone practically does not proceed, selectivity, 97.1 mol %.

PETROLEUM CHEMISTRY Vol. 47 No. 4 2007

280 ZAKOSHANSKY

h0
(‡) 21 (b)
9 18
7 15
5 12
9
3 6
1 3
0
0 0.5 1.0 1.5 2.0 2.5 0 100 200 300 400 500 600
H2O, wt. % H2SO4, wt. %

Fig. 7. Change in acidity h0 as a function of: (a) H2O in an equimolar phenol/acetone mixture ([H2SO4] = 100 ppm, T=25°C);
[H2SO4] in phenol/acetone mixture at (1) H2O concentration 0.52 wt % and (2) 0.31 wt %.

The problem of designing a safe, computer-con- phenol purification from HA even at extremely high (by
trolled, and highly selective process for CHP decompo- ~5–10 times higher than usual) values of its concentra-
sition (table) has been effectively solved in the tion in raw phenol from the real commercial run. The
FAN-2000 technology successfully used in industrial term of the continuous operation of a catalyst without
practice. It is noteworthy that it is the first process of regeneration is not less than 1.5 years. The level of phe-
CHP decomposition designed in which the realized nol purification from impurities on the catalyst of the
logic of the safety and protection of the process does second steps (acid zeolite catalyst) considerably
not interfere with the purpose of achieving its maximal exceeds the parameters characteristic of all known sul-
selectivity. Using the ILLA process, ~30 rel. % of the fonated cation-exchangers. The industrial term of the
global amount of phenol–acetone—2.5 million t/year of continuous operation of the catalyst of the second step
phenol and 1.5 million t/year of acetone—is produced, without regeneration is 7 years. Parameters of the
and phenol production will increase by 1.5 million t/year purity of the obtained phenol after the two-step purifi-
in 2010. cation correspond to a level of phenol used in the pro-
duction of carbonate plastics.
Phenol Quality Problem
This problem has been the Achilles’ heel of the phe- Feedstock and Steam Consumption
nol process for a long time, because only top-grade for the Process as a Whole
phenol having a total impurity level of no more than It is seen from the practical data on the selectivity of
0.01% and meeting the stringent color standards can be the basic steps of the phenol process and from the values
used for the manufacture of carbonate plastics. The of the physical and chemical losses at the auxiliary steps
color characteristics are directly related to the concen- of the process, the overall selectivity (for phenol) of the
tration of hydroxyacetone (HA) and 2-methylbenzofu- phenol process on the fed cumene basis is 93.9 mol %
ran (2-MBF) that is formed from HA in the presence of and 97.1 mol % (1360 kg/t and 1315 kg/t, respectively).
an acid catalyst. This problem has not been completely In our opinion, advertised values of the coefficient of
solved by many phenol manufactureses. materials consumption below 1315 kg/t are unrealistic.
When performing detailed studies of all chemical Depending on the type of the applied CHP decom-
reactions proceeding in the traditional single-step puri- position technology, the level of steam consumption in
fication of phenol with the use of sulfonated cation- optimal processes varies from ∼4.2 to 2.2 t/t of phenol.
exchangers and having revealed its principal draw- The optimal value was found to be ∼4,2 t/t, whereas the
backs, ILLA completely rejected this method and value of 2.2 t/t was attained owing to the practically
developed a new two-step method for phenol purifica- complete integration of heat, leading to rather great dif-
tion [12]. ficulties for those who directly operate the process. The
Cardinal distinction of this method is that, at the first consumption of cooling water is ∼300–350 m3/t of phe-
step, complete HA removal on special heterogeneous nol and is comparable for all types of licensed pro-
catalyst under conditions excluding the formation of cesses.
2-MBF is carried out; at the second step, purification is In our opinion, there are approaches to improve the
carried out with the use of traditional heterogeneous selectivity of the step of cumene oxidation of isopropyl
acid catalysts (preferably zeolite type). Results of the toluene and the step of CHP decomposition, and, it is
operation of a two-step purification ILLA unit on heter- the most important factor is that they allow one to trans-
ogeneous catalysts are presented in Figs. 8a and 8b. form the process to practically wasteless and extremely
Based on these data, it can be concluded that the cat- economic, or the so-called “wasteless process for phe-
alyst of the first step successfully solves the problem of nol–acetone production,” WEPP. The chemical and

PETROLEUM CHEMISTRY Vol. 47 No. 4 2007

 THE CUMENE PROCESS FOR PHENOL–ACETONE PRODUCTION 281

Principal characteristics and parameters of the CHP decomposition process

Processes

“boiling” (Cellog, Mitsui, homogeneous

homogeneous, ILLA
Shell, Sunoco/UOP) (Sunoco/UOP)

First step, H2SO4, ppm 450–700 100–200 150–180

Second step, H2SO4, ppm 450–700 100–200 10–30

Reaction medium, acetone/phenol 1.05 : 1 1:1 1.5 : 1

Computer start-up Cannot be realized in prin- None Is realized in all plants

ciple

Computer control Cannot be realized in prin- None, realization is diffi- Is realized in all plants (**)
ciple cult (*)

Number of protection levels of the 1 1 3

process

Theoretically possible yield of 74 80 93

AMS, mol %*

Actual yield of AMS, mol %* ∼65 67–74** 89–92.5**

Yield of phenol tar, kg/t of phenol* 90 70 45

[HA], ppm 1800–2000 1500–1700 1300

[MOX], ppm 500 300–200 <50

Notes: * As borrowed by ILLA workers from published data, the inspection data obtained in active plants, the data of kinetic studies and
modeling of the process on pilot units.
** The AMS yield of ~85 [11] has been reported in the advertising materials of Sunoco/UOP; however, it has not been reported in a
concrete unit and its capacity where this result has been attained; all plants and their capacities are known (GE, ~350000 t/year,
INEOS, two plants for 200000 t/year, FCFC, 200000 t/year, FQ, 42000 t of phenol/year), where the AMS yield has been ~89–
92.5% under industrial conditions according to the ILLA process of CHP decomposition.

technological essence of this process is presented in from the CHP decomposition step) to WEPP is shown
[13, 14]. in Fig. 9.
If pilot tests will be successful, the expected cumene However, until the pilot tests of a sorption–desorp-
consumption will be 1283–1285 kg/t of phenol, i.e., tion step will be finished, it is premature to declare the
selectivity for cumene will achieve 99.5 mol %. The creation of a new production technique. The repre-
second important difference and advantage of WEPP in sented data on this ILLA process have an especially
comparison with traditional cumene processing is the informational character.
safety and simplicity of the step of CHP decomposi-
tion, low power inputs to separate products of CHP
decomposition, and the high quality of the products RECONSTRUCTION OF OUT-OF-DATE
produced. The third important difference and advan- PRODUCTION ENGINEERING AND PLANTS
tage is the simple transfer from traditional cumene pro- Economic expediency of the radical reconstruction
cessing to WEPP. of imperfect and obsolete production engineering does
The simplified scheme (developed at ILLA) to not require demonstrations—numbers of annual losses
transform the traditional cumene process (beginning of such plants speak for themselves. In particular,

PETROLEUM CHEMISTRY Vol. 47 No. 4 2007

282 ZAKOSHANSKY

HA concentration at the inlet and outlet

of a pilot unit for phenol purification
[HA], ppm
1000
(‡)
800
600
400
200
0
0 50 100 150 200 250 300 350 400
Test time, day
(b)
[Total carbonyls], ppm [Mesityl oxide],
Feedstock: 400 ppm Feedstock: 100 ppm
150 30
100 20
1 year 1 year
50 7
10 new 7 years new
new year new
0 0
Zeolite Cation exchanger Zeolite Cation exchanger
Phenol purification of total carbonyls Phenol cleaning of mesityl oxide

[Total impurities], ppm

Feedstock: 400 ppm
150
100
7 years 1 year
50 1 year
new year 7 new
0
Zeolite Cation exchanger Zeolite Cation exchanger
Total impurities after purification lifetime

Fig. 8. Treatment of crude phenol for hydroxyaxetone removal (a) on the ILLA heterogeneous catalyst (first step); (b) on the zeolite-
containing catalyst (second step).

plants with a low capacity (40000–60000 t/year of phe- (2) So-called “reanimation” is realized, owing to
nol), which continue to use old production engineering, minimal possible changes in production engineering,
lose from US $5 to $15 million annually. Plants of which are carried out on the operative (or available)
medium (70000–120000 t/year of phenol) and high equipment, to vary regimes of the key steps of the pro-
capacity (150000–300000 t/year of phenol) lose, cess to improve it. But, this step does not exclude
accordingly, much more, especially in proportion to implementation of a part of the production engineering
capacity. However, note that many plants of medium stipulated by the next, third stage of modernization.
and high capacity once every 10–20 years conduct (3) Radical reconstruction, i. e., a total transfer of
modernization that leads to some decrease in their operations to modern production engineering.
losses. Manufacturing applications of modern produc-
tion engineering developed by the best licensors result The first stage is in full measure a prerogative for the
in radical improvements of all production parameters, manufacturer of products and his responsibility; the
economy, safety, and ecology. second and third steps are based on the purchase of
licenses, accordingly, the programs “reanimation” and
As experience shows at ILLA, which has success- “radical reconstruction” are developed by the licensor.
fully conducted numerous reconstructions at various
plants in the world, to achieve success in the modern- The out-of-date plants really need the realization of
ization of obsolete plants and production engineering, a the first step, owing to the depreciation of equipment,
comprehensive approach is usually needed, including the use of obsolete measuring devices, and their insuf-
three steps: ficient amount, etc. Expenses at this step can be both
greater and smaller and are predetermined by what
(1) Expenses to maintain a process under operable measure fixed capital was renovated during long main-
conditions. tenance processes. But, in any case, these expenses are

PETROLEUM CHEMISTRY Vol. 47 No. 4 2007

 THE CUMENE PROCESS FOR PHENOL–ACETONE PRODUCTION 283

Commercial acetone
2
Products
of CHP Commercial Recycled
decomposition 1 phenol cumene
3
Fraction
Recycled of ACP
cumene and cumene

Condensation
products
CATALYTIC Acetophenone
PHENOL
PURIFICATION

Fig. 9. Simplified flow chart after the reconstruction of the conventional cumene production engineering and its transfer onto the
WEPP technology (cumene oxidation and sorption-desorption steps are absent).

forcedly necessary if the producer intends to make of the out-of-date technology or the absence of invest-
products instead of closing the plant by virtue of its ments in production, but also a consequence of the
moral and physical obsolescence. technological level of their top managers and engineer-
There is one more objective reason why reconstruc- ing staff. The evaluation is based on the facts that (1) a
tion of obsolete plants is the difficult problem. That is, number of plants have the byproducts yields, which are
only some of all licensors know one another’s produc- lower by a factor of 1.5–2 even if the out-of-date tech-
tion engineering. And, the use of local engineering and nology is used and (2) modeling of the out-of-date tech-
licensing companies, even if they have developed and nology based on the kinetics of the key steps of the pro-
designed in the past the plants to be reconstructed, cess (cumene oxidation and CHP decomposition which
appears to be completely senseless, because these com- mainly determine parameters of the process) unambig-
panies do not possess either the rights to modern pro- uously confirms that the yield of byproducts should not
duction engineering, or the knowledge of these produc-
tion engineering. exceed 140–145 kg/t of phenol.
If the aforementioned difficulties are overcome and Practice has convincingly proven that:
the customer begins reconstruction, the realization of
the second stage of modernization is economically (a) It is impossible to attain a high selectivity of
expedient because the volume of borrowed funds for cumene oxidation in reactors of old design.
the implementation of main stage—radical reconstruc- (b) The conversion of CHP decomposition reactors
tion—decreases owing to the partial increase in the
selectivity of the key steps of the process. Moreover, the of old designs into modern production engineering is
implementation of the second stage allows one to impossible unless the safety of the process and the loss
reduce markedly the payback time of expenses for in selectivity, higher than that in the original process of
overall modernization. the licensor, are disregarded.
In our opinion, it is possible in practice to choose In conclusion of the consideration of the stages of
three types of obsolete plants that are easily distin- the improvement of the performance characteristics of
guished in terms of yields of byproducts, so-called phe- the cumene method for phenol production, it is neces-
nol tar (without taking cracking into account): sary to mention alternative methods for direct phenol
a) very obsolete, 170–250 kg/t of phenol; preparation advertised recently. Analysis of the litera-
b) an average level of obsoleteness, 120–170 kg/t of ture, patent, and promotional materials, nevertheless,
phenol; shows that so-called “single-step” methods for benzene
c) obsolete, 100–120 kg/t of phenol. oxidation to phenol, even explicitly elegant from the
As shown by the experience of ILLA, all three types chemical point of view, rank considerably below the
of obsolete plants can be successfully transformed by ILLA cumene process in their technical-and-economic
means of modern technology into those having a yield parameters. Accordingly, the prospects for industrial
of phenol tar of 50–70 kg/t of phenol, a value that is implementation of “single-step” processes in the near
equivalent to feedstock (cumene), saving from 50 to future seem doubtful. Most likely, the newly developed
200 kg/t of phenol. cumene process and its modifications will dominate the
Note that the high yield of phenol tar (170–250 kg/t market of phenol production even in the remote per-
of phenol) at very old plants is not only a consequence spective.

PETROLEUM CHEMISTRY Vol. 47 No. 4 2007

284 ZAKOSHANSKY

REFERENCES 8. V. M. Zakoshansky, AICHE Second International Aro-

matics Processing Conference (New Orleans, 2003),
1. RU Patent No. 1455596 (1987); US Patent p. 794.
No. 5,908,962 (1998); RU Patent No. 2146670 (1998).
9. V. M. Zakoshansky, AICHE Third International Aromat-
2. V. M. Zakoshansky, AICHE First International Aromat- ics Processing Conference (New Orleans, 2004), p. 504.
ics Processing Conference (New Orleans, 2001), p. 166.
10. V. M. Zakoshansky, Yu. N. Yur’ev, G. S. Idlis, and A. S. Alek-
3. V. M. Zakoshansky, Katal. Khim. Neftekhim. Prom., sandrov, Zh. Fiz. Khim. 58 (5), 1265 (1984).
No. 1, 29 (2002).
11. UOP / R.J. Schmidt // 3rd Asia Phenol/Acetone & Deriv-
4. V. M. Zakoshansky, Second Asia Phenol/Acetone & atives Market. 2006. Peking. Febrary 23−24.
Derivatives Market (2005). 12. RU Patent No. 2111203 (1993); US Patent
5. V. M. Zakoshansky, Katal. Khim. Neftekhim. Prom., No. 5,502,259 (1996); RU Patent No. 2266275 (2004).
No. 5, 2 (2004). 13. RU Patent No. 2125038 (1996); US Patent No. 6,252,124
6. SU Patent No. 1361937 (1985); SU Patent No. 1563181 (2000).
(1990); US Patent No. 5,254,751 (1993). 14. V. M. Zakoshansky et al., AICHE Second International
7. RU Patent No. 2141938 (1999); RU Patent Aromatics Processing Conference (New Orleans, 2003),
No. 2142 932 (1999); US Patent No. 6,057,483 (2000). p. 75.

PETROLEUM CHEMISTRY Vol. 47 No. 4 2007