20.1999.kinetic and Catalytic Aspects in The Hydrogen Peroxide Production Via Anthraquinone
20.1999.kinetic and Catalytic Aspects in The Hydrogen Peroxide Production Via Anthraquinone
20.1999.kinetic and Catalytic Aspects in The Hydrogen Peroxide Production Via Anthraquinone
Abstract
The kinetic and catalytic aspects involved in the hydrogen peroxide production via anthraquinone (method all-tetra) have been
examined considering three of the four main production steps, i.e., hydrogenation of 2-ethyltetrahydroanthraquinone catalyzed by
palladium-supported catalysts, oxidation of the product obtained in the "rst step, and reconversion of the deteriorated molecules for
the regeneration of active quinones. For each reaction considered, a mechanism and a kinetic law are suggested with the numerical
values of the related parameters. 1999 Elsevier Science Ltd. All rights reserved.
1. Introduction 2. Results
The production of hydrogen peroxide via anthra- 2.1. Examination of the problems involved in the xrst step:
quinone with the method named all-tetra occurs in four hydrogenation of THEAQ
steps, i.e.: (1) hydrogenation of 2-ethyltetrahydroanthra-
quinone (THEAQ) in a slurry reactor (Santacesaria et al., In the mentioned process all-tetra (Ulmann, 1994),
1988, 1994); (2) oxidation of the product obtained (THE- hydrogen peroxide is produced subjecting a mixture
AQH ) in a gas}liquid reactor (Santacesaria et al., 1987); of 30% 2-ethylanthraquinone (EAQ) and 70%
(3) extraction of hydrogen peroxide with an acid aqueous THEAQ dissolved in an appropriate mixture of solvents
solution; and (4) regeneration of active quinones from to cyclic reduction and oxidation as in the following
chemically deteriorated molecules and puri"cation of the scheme:
solution on a packed bed tubular reactor (Ulmann, 1994;
Sandelin et al.,1994). In the present work, we will brie#y
summarize the results previously obtained in the kinetic
studies of step (1) and (2) and already published (Santa-
cesaria et al., 1988,1987, 1994). More details on this subject
can be found in the mentioned literature. Step (4) will be
examined in more detail, because very few papers have
been published on this topic (Sandelin et al.,1994), despite (1)
its importance for the management of industrial plants.
0009-2509/99/$ } see front matter 1999 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 0 9 - 2 5 0 9 ( 9 8 ) 0 0 3 7 7 - 7
2800 E. Santacesaria et al. /Chemical Engineering Science 54 (1999) 2799}2806
(2)
poisoning occurs through the following reaction: activity simulation when catalyst poisoning is neglected
as well as the agreement with our model.
The necessity to add fresh catalyst to keep the catalytic
. (6) activity in the industrial reactor constant is the most
relevant consequence of catalyst poisoning. Similarly, we
need to add EAQ to keep the optimal ratio EAQ/THE-
Reversible poisoning is responsible for an about 30% AQ constant, which is altered by the slow hydrogenation
catalyst activity loss in 1 or 2 days of operation. Irrevers- process for the aromatic rings of EAQ and THEAQ.
ible poisoning is much slower; to study it, we interpreted These reactions do not a!ect the kinetic studies per-
the activity data of the industrial plant. By de"ning h(t) as formed in laboratory reactors for the hydrogenation of
the fraction of active sites at time t, h (t) as the fraction of anthraquinones to oxygen, but cannot be neglected in the
simulation of big-sized continuous industrial plants for
available sites, i.e., of not permanently poisoned sites, and
as the fraction of the sites occupied by water, we can put long-time operations. Ring hydrogenation reactions are
the actual kinetic constant of hydrogenation propor-
tional to h, that is k "k h, k being the kinetic constant
& ! !
of the fresh catalyst. The evolution with time of h should
be
dh 1 h*p
"!k'C h 1! & . (7)
dt B 5 k C h
CB 5
For the evolution of h we assumed: . (9)
dh
"!k''h . (8)
dt B
others
&&
mk b [EAQ]p
r " & & , (11)
& (1#b [EAQ]#b [THEAQ])
mk b [THEAQ]p
r " & & , (12)
& (1#b [EAQ]#b [THEAQ])
Fig. 3. Simulation of an industrial reactor. The dashed line corresponds
to the simulation of anthraquinone conversions with a model neglect-
r "mk [EAQ]? . (13)
ing deactivation. The continuous line interpolating experimental points & &
is the simulation achieved by considering the occurrence of catalyst
deactivation. The heavy continuous line corresponds to the percent of Reaction (3) contributes to the loss of EAQ for less than
increasing catalyst hold up in the plant. 5%, a results about 1.
2802 E. Santacesaria et al. /Chemical Engineering Science 54 (1999) 2799}2806
The kinetic parameters for the same palladium-sup- We have seen that during the hydrogenation step,
ported catalyst used in the previously mentioned kinetic secondary products are formed owing to the hydrogen-
runs are collected in Table 1. ation of aromatic rings. Normally, THEAQH is ob-
tained from EAQ and THEAQH forms H EAQH . By
2.2. Examination of the problems involved in the second analysis we have also seen small amounts of partially
step: THEAQH oxidation hydrogenated intermediates such as 2-ethylhexahydro-
anthraquinone H EAQH and 2-ethyldihydroanthra-
quinone H EAQH . EAQH and THEAQH can also
THEAQH oxidation is normally performed at indus-
give tautomerization to oxanthrone followed by the
trial level in bubble column continuous reactors with air
and with the working solution #owing countercurrent. formation of anthrone and dianthrone (Ulmann 1994)
Air #ow rates are high enough to keep the liquid phase
well mixed. The reaction occurs with a second-order
kinetic law of the type (Santacesaria et al., 1987)
k "(1.14$0.08);10 exp(!(14213$1036)/RT)
(cm/mol s). (15)
All these secondary products are formed as a conse-
The mass transfer and oxygen solubility parameters were quence of very slow reactions not a!ecting the laboratory
determined independently. kinetic studies but they accumulate in the big-sized in-
The kinetic law and related parameters have success- dustrial plants and their reaction gives place to less
fully been tested simulating the bubble column of both soluble condensation products probably responsible for
a pilot and di!erent industrial plants, the values of the the permanent poisoning of the palladium-supported
gas}liquid interphase area in the examined bubble col- catalyst. Besides adding fresh EAQ to compensate the
umns were somewhat uncertain as the working solution loss of active quinones, we therefore need to regenerate
tended to give copious foam at the top of the reactor. deteriorated molecules as much as possible and to purify
From the experimental viewpoint, it is interesting to the working solution from the formed tars and the acidity
observe that in the runs performed in the laboratory developed during the autoxidation step. Alkali supported
semibatch reactor no induction time has been observed, on alumina or on silica}alumina with high surface area
that is, steady-state conditions for the involved radicals are normally used as catalysts and adsorbents in this
are readily achieved. Moreover, a second-order kinetic reconversion step (Ulmann, 1994). The operation can be
E. Santacesaria et al. /Chemical Engineering Science 54 (1999) 2799}2806 2803
1. THEAQ dehydrogenation, mainly occurring in the However, the most adsorbed active anthraquinone was
oxidized solution with the following stoichiometry: THEAQH owing to the possibility of giving hydrogen
bonds with the oxygens and hydroxyls of the solid sur-
face. Exhausted alumina catalysts contain about 12% of
organic compounds but only 2.4% of them are active
quinones. Therefore, alkaline alumina catalysts have the
e!ect of purging and regenerating the working solution
developing three fundamental actions:
Table 2
Kinetic parameters obtained by mathematical regression on
experimental data
(22)
because H EAQ is always present in small amounts in
the working solution. As the slow reactions occur on the
surface between adsorbed molecules:
r "k h h (23)
2&#/ &#/
From the experimental evidence, it results that THE-
AQH is the most strongly adsorbed molecule, therefore,
b [THEAQ] Fig. 4. Kinetic runs of THEAQ dehydrogenation performed in batch
h " 2&#/ (24) conditions. ¹"1003C, catalyst concentration"100 g/dm, < "100 cm,
2&#/ 1#b [THEAQH ] *
2&#/& [THEAQ]"8.3;10\ mol/dm. Symbols correspond to experi-
mental data: 䊏 EAQ, 䊉 THEAQ, 䉱 THEAQH ; lines are obtained by
and
calculations.
b [H EAQ]
h " &#/ . (25)
&#/ 1#b [THEAQH ]
2&#/&
By considering the "rst step in equilibrium, it results in
k [THEAQ]
[H EAQ]" . (26)
[THEAQH ]
By introducing all the terms in Eq. (23), we obtain
k [THEAQ]
r " , (27)
(1#b [THEAQH ])[THEAQH ]
2&#/&
where k "k b b k .
2&#/ &#/
However, we observed that two di!erent basic sites are
operative, one related to the supported alkali and the
other existing on the surface of the support. The latter
sites give place to adsorption phenomena which are neg-
ligible if compared with those shown by supported alkali, Fig. 5. Kinetic runs of THEAQ dehydrogenation performed in batch
therefore, for the activity of these sites, we can give conditions. ¹"1403C, catalyst concentration"100 g/dm, < "
*
a simpler kinetic equation of the type 100 cm, [THEAQ]"8.3;10\ mol/dm. Symbols correspond to ex-
perimental data: 䊏 EAQ, 䊉 THEAQ, 䉱 THEAQH ; lines are obtained
[THEAQ] by calculations.
r "k [THEAQ][H EAQ]"k , (28)
[THEAQH ]
where k "k k .
d[THEAQH ]
"2(r #r )
Experimental kinetic runs can now easily be simulated by
dt
integrating the equations:
d[THEAQ] d[EAQ]
"!3(r #r ) "r #r . (29)
dt dt
E. Santacesaria et al. /Chemical Engineering Science 54 (1999) 2799}2806 2805
b [THEAQH ]
h " 2&#/& (31)
2&#/& 1#b [THEAQH ]#b [THEAQ]#b [H O]#b [Sol]
2&#/& 2&#/ &- 1
b and [Sol] are di!erent according to the solvent
1
applied.
Hence,
k [THEAQH ] [EPOX]
r " . (32)
1#b [THEAQH ]#b [THEAQ]#b [H O]#b [Sol]
2&#/& 2&#/ &- 1
2806 E. Santacesaria et al. /Chemical Engineering Science 54 (1999) 2799}2806
In this case, too, two types of sites are operative, the same k THEAQH aromatic ring hydrogenation rate
&
already seen for THEAQ dehydrogenation. The basic sites constant, mol g\ atm\ h\
of the support give negligible adsorption e!ects, therefore, k rate constant of EAQH degradation, cm
&
g\ h\
r "k [THEAQH ][EPOX] . (33)
k gas-liquid mass-transfer coe$cient, cm/s
*
The evolution with time of the di!erent involved molecu- k kinetic constant of ethyl-anthraquinone oxida-
lar species can easily be followed by integrating the tion, cm mol\ s\
following di!erential equations: k liquid-solid mass-transfer coe$cient, cm/s
1
K equilibrium constant for the reversible catalyst
d[EPOX] CB
"!(r #r ) poisoning, atm mol\ cm\
dt m catalyst hold-up, g cm\
d[THEAQH ] P partial pressure of H , atm
"!(r #r ) &
r reaction rate of EAQH aromatic ring hydro-
dt &
genation, mol dm\ h\
d[THEAQ] r reaction rate of THEAQH aromatic ring hy-
"2(r #r ) &
dt drogenation, mol dm\ h\
r reaction rate of EAQH2 degradation,
d[H O] &
"r #r . (34) mol dm\ h\
dt r rate of i reaction, (i"1,4)
G
The best "tting of the experimental data have been ob- r reaction rate of ethyl-anthraquinone oxidation,
MV
tained with the parameters collected in Table 3, where mol cm\ s\
b is related to 2,5-dimethyl-4-eptanol. t time, s
1 ¹ temperature, K
The most relevant kinetic and catalytic phenomena a reaction order for the gaseous reagent
occurring in the hydrogen peroxide production via g overall e!ectiveness factor
anthraquinone have been described. For the hydrogen- h fraction of active sites at time t
ation and oxidation steps, it has been shown that h* fraction of sites reversibly poisoned by water
h fraction of catalyst sites occupied by i
laboratory results can usefully be extrapolated to inter- G
h fraction of available sites, i.e., of not permanently
pret the performance of industrial plants. The same work
is in progress to describe the reconversion section. poisoned sites, initially equal to 1
References
Acknowledgements
Chaudari, R.V., & Ramanchandran, P.A. (1980). Ing. Engng Chem.
Ausimont SpA is acknowledged for the "nancial sup- Fundam., 19, 201.
port. Chaudari, R.V., & Ramanchandran, P.A. (1983). ¹hree phase catalytic
reactors. London: Gordon and Breach.
Berglin, T., & ShoK oK n, N.T. (1983) Ind. Engng. Chem. Process Des. Dev.,
Notation 22, 150.
Etienne, A., & Fellion, Y.C.R. (1954) Hebd. Seances. Acad. Sci., Ser. C,
a speci"c gas}liquid interface area, cm\ 238, 1429.
* Levenspiel, O., & Godfrey, J.M., J. H. (1974) Chem. Engng Sci. 29, 1723.
a speci"c liquid}solid interface area, cm\
1 Sandelin, F., Oinas, P., Salmi, T., & Haario, H (1994). Catalyst deactiva-
b adsorption equilibrium constant of i species, cm
G tion. In: B. Delmon, & G.F. Froment (Eds.), studies in surface and
or dm/mol catalysis (Vol. 88).
C water concentration, mol/cm Santacesaria, E., Di Serio, M., Velotti, R., & Leone, U. (1984). J. Mol.
Catal. 94, 37}46.
D e!ective di!usion coe$cient, cm/s
C Santacesaria, E., Wilkinson, P., Babini, P., & Carrà S. (1988) Ind. Engng.
H Henrys parameter of the gaseous reagent solu-
Chem. Res. 5, 27.
bility, cm mol\ atm\ Santacesaria, E., Di Serio, M., Velotti, R., & Leone, U. (1994) Kinetics,
k' rate constant of reversible catalyst poisoning,
B mass transfer, and palladium catalyst deactivation in the hydrogen-
cm mol\ day\ ation step of peroxide synthesis via anthraquinones. Ind. Engng
k'' rate constant of irreversible catalyst poison- Chem. Res. 2, 33.
B Santacesaria, E., Ferro, R., Ricci, S., & Carrà, S. (1987) Kinetic aspects
ing/day
in the oxidation of hydrogenated 2-ethyl-tetrahydroanthroquinone.
k rate constant of i reaction (i"1,4)
G Ind. Engng Chem. Res. 26, 155}159.
k EAQH aromatic ring hydrogenation rate con-
& Ulmann's Encyclopedia of industrial chemistry; (1994) (Vol. A13, 443)
stant, mol g\ atm\ h\ VCH: Weinheim.