Acta Chini. Slov. 2001, 48, 585-596.
585
SEPARATION PROCESSES IN H202 PRODUCTION
Matjaž Krajne*, Janvit Golob, Aleš Golja, Roman Burja
University of Ljubljana, Faculty of Chemistry and Chemical Technology, Ljubljana
Matevž Ložar, Slavko Flajs, Marko Grčar
Beiinka Perkemija, Ljubljana, Slovenia
This paper is dedicated to Prof. Roman Modic at his 90th Birthday. Received 28-05-2001
Abstract
The article presents basic physico-chemical and process parameters, which are important in capacity enhancement design of the hydrogen peroxide production The limiting steps in the greater production, taking into account the existing production plant, are oxidation and extraction. The oxidation is a moderately slow reaction occurring mainly in bulk liquid, whose rate is affected by the mass transfer across the liquid film. Therefore, both interfacial area and liquid hold-up should be maximized in the industrial reactor. In the extraction process, the idea of pseudo equilibrium line is introduced. It was found out that knowing transport coefficients and interfacial area is not enough to predict the process in the extraction column.
Introduction
The production of hydrogen peroxide (H2O2) via anthraquinone occurs in three steps (Figure 1), i.e.: (1) hydrogénation of 2-ethyltetrahydroanthraquinone (ETQ); (2) oxidation of the product obtained in a gas-liquid reactor; (3) extraction of hydrogen peroxide with demineralized water. The industrial process is called all-tetra, because the principal active anthraquinone compounds are in tetrahydro forms. A working solution with the active agents consisting of 70% of ETQ and 30% of 2-ethylanthraquinone (EQ) dissolved in a proprietary mixture of organic solvents is used. " In the hydrogénation step,   anthraquinones   are   reduced   to   2-ethyltetrahydroanthraquinol   (ETQH2)   and
2-ethylanthraquinol (EQH2). However, the latter readily reacts to give'
EQH2                           ETQ                               EQ                             ETQH2
M. Krajne, J. Golob, A. Golja, R. Burja, M. Ložar, S. Flajs, M. Grčar: Separation Processes in H202...
586
Acta Chini. Slov. 2001, 48, 585-596.
This equilibrium is almost completely shifted to the right. In order to avoid secondary reactions, only 60-70% of the anthraquinone mixture is reduced in the hydrogénation step. Therefore, only ETQH2 is involved in the successive oxidation step.
extraction
Figure 1.           Schematic representation of H2O2 production.
The oxidation step, a gas-liquid reaction, is considered as slow to moderately fast. The absorption rate of oxygen is essential for the overall transfer rate. The oxidation reaction obeys the first-order kinetic law for both oxygen and anthraquinone.
The aim of this work is to define basic physico-chemical and process parameters, which are important in capacity enhancement design of the H2O2 production. The limiting steps in the greater production, taking into account the existing production plant, are oxidation and extraction. Therefore, the research was oriented toward (i) the study of oxygen absorption followed by chemical reaction in an oxidizer, (ii) analyzing process parameters which exert an influence on global reaction rate of oxidation, (iii) in the extraction process the determination of equilibrium diagram EbCh-water-working solution at 50 °C, and (iv) the study of effect of process parameters and physico-chemical properties on the efficiency of the column with perforated plates.
Experimental
Oxidation experiments were performed in a batch reactor with a volume of 4.6 L (d = 0.14 m; h = 0.30 m). The mixing equipment consisted of a six-bladed pitch-blade turbine with a diameter of 0.047 m. The reactor was placed in a water-bath thermostat to
M. Krajne, J. Golob, A. Golja, R. Burja, M. Ložar, S. Flajs, M. Grčar: Separation Processes in H202...
Acta Chini. Slov. 2001, 48, 585-596.
587
maintain a constant reaction temperature and it was fed with a 2.2 L of working solution containing hydrogenated anthraquinone. After reaching a desired reaction temperature (40 and 60 °C) air was purged at the bottom through a sparger. The flow of air varied between 200 and 500 L/h. Formed H2O2 has a greater specific weight as the working solution and tends to deposit at the reactor bottom. To eliminate the separation effect the range of the stirring rate was chosen between 1750 and 2250 min" . The amount of hydrogen peroxide formed and anthraquinone converted were determined by hydrogen peroxide titration.
Extraction experimental work comprised determination of chemical and physical characteristics of a system working solution - hydrogen peroxide - water at 50 °C, engineering analysis of extraction process and prediction of industrial experiment. The last phase of the research was realization of an industrial experiment followed by the analysis of the results.
Results and discussion
Oxidation. Oxidation of ETQH2 by oxygen occurs through the following steps: (1) diffusion of oxygen from the gas phase to the gas-liquid interphase, (2) diffusion of oxygen from the liquid interphase into the bulk of the solution and (3) chemical reaction in the solution. It was assumed that the main resistance is at the liquid side. Therefore, the first step can be ignored. The other two steps can occur simultaneously. Therefore, in order to evaluate the extent of the reaction occurring in the liquid film, the enhancing factor or Hatta number, Ha, must be determined.
To define qualitatively the effect of process parameters on chemical reaction and mass transfer the oxidation experiments at different stirring speeds, temperatures and gas flow rates were performed and compared. All experiments were conducted by changing only one parameter at time leaving the other two constant. The effect of mixing is presented at Figure 2. It can be concluded that mass transfer affects the global reaction rate. An interfacial area is higher at higher stirring speed resulting in higher volumetric mass-transfer coefficient, kiA
M. Krajne, J. Golob, A. Golja, R. Burja, M. Ložar, S. Flajs, M. Grčar: Separation Processes in H202...
588
Acta Chini. Slov. 2001, 48, 585-596.
^    0,4
g
A
D 1750 min"1 O 2000 min"1 A     2250 mill"1
D O
D O
O     ?
0              5             10            15            20            25            30            35            40
t, min
Figure 2. The effect of stirrer speed on the ETQH2 disappearance (T = 60 °C; (J>g = 500 L/h).
The global reaction rate also depends on gas flow rate. Higher flow rate results in higher interfacial area and higher amount of oxygen present in the reaction mixture. Therefore, H2O2 production rate is higher. However, there is an upper limit. At flow rate higher then 700 L/h, a contact time between gas and liquid is too short. Consequently, oxidation is not efficient. To eliminate this side effect the experiments were performed at flow rates between 200 and 500 L/h. Figure 3 shows the effect of gas flow rate.
^ 0/
2n
D O
A    O    ?
O
_1_
_l_
_1_
_l_
o      D 0
D 200 L/h O 300 L/h A     500 L/h
_1_
A   a iW.rPn   H  .   n
5   10  15  20  25  30  35  40  45  50  55  60
t, min
Figure 3. The effect of gas flow rate on the ETQH2 disappearance (T = 60 °C; N 2000 min"1).
M. Krajne, J. Golob, A. Golja, R. Burja, M. Ložar, S. Flajs, M. Grčar: Separation Processes in H202...
Acta Chini. Slov. 2001, 48, 585-596.
589
Temperature dependence of global reaction rate was determined at constant stirrer speed and gas flow rate at two different temperatures (40 and 60 °C). The effect is presented in Figure 4.
?    40 C exp.
-----40°C cal.
O    60°C exp. -----60°C cal.
ULi___1^^
10       15       20       25       30       35       40       45       50       55       60
t, min
Figure 4. Experimental and calculated concentration profiles ((J>g = 500 L/h; N 2000 min"1).
To quantify the experimental results, it was assumed that the reaction rate obeys a second-order kinetic law. Employing the relation
- kLa-(CB    CB) -kr -CA -CB
dt          dt
that is, eliminating Cb the rate of ETQH2 disappearance becomes
dCA dt
c:
1     1
—+—
r     A        a^Ć/
(2)
(3)
where A and B represent ETQH2 and oxygen, respectively. By integrating Eq. 3 the following expression describing consumption of ETQH2 may be written
1
kLa
(c°A-cA)+^-\n^ = c;-t
(4)
The values of kr and Henry constant, He, were used as found in literature '
kr  = (1,14 ± 0,08) • IO" • eK14213±1036)/*7-)   [cm3/mol s]                                            (5)
M. Krajne, J. Golob, A. Golja, R. Burja, M. Ložar, S. Flajs, M. Grčar: Separation Processes in H202...
590
Acta Chini. Slov. 2001, 48, 585-596.
He (40°C)= 110 bar L/mol
He (60°C)= 108 bar L/mol
By fitting experimental data to Eq. 4 the values of kLa may be estimated at different temperatures and gas flow rates. To compare experimentally obtained values of kLa with known correlations the correlation developed by Gaddis was used
kLa = (kLay/(v/ë2)m                                                                            (6)
where is
(kLa)* -(9.8xlO"5 0P/rr4)/(#     +0.81xl0-°65ii)
and
(p/Fr=(/>/F)/(p-(vgTj)
B = y,/D2,)-(yg)
-1/4
(7)
(8) (9)
The comparison of correlated and calculated volumetric mass transfer coefficients is presented in Table 1. The good agreement between calculated results from Eq. 4 and experimental data is presented in Figure 4.
Table 1. Comparison of ^a.
(^ =200 L/h
kLa, s l
cal. (Eq.4)
kLa, s l (Eq. 6)
(k =300 L/h
kLa, s l
cal. (Eq.4)
kLa, s (Eq. 6)
(k =500 L/h
kLa, s l
cal. (Eq.4)
T = 40°C
0.0402     0.0358     0.0507     0.0444     0.0542      0.0588
T = 60°C
0.0471  0.0478  0.0617  0.0594  0.0863  0.0789
kLa, s (Eq. 6)
To determine the influence of mass transfer on the global reaction rate the Hatta number was introduced. The Hatta number, Ha, compares the maximum chemical conversion in the mass transfer film to the maximum diffusion flux through this liquid film, thus is
Ha
K
(10)
M. Krajne, J. Golob, A. Golja, R. Burja, M. Ložar, S. Flajs, M. Grčar: Separation Processes in H202...
Acta Chini. Slov. 2001, 48, 585-596.
591
Since the parameter kL appears independently in the expression of the Hatta number, its value was calculated from the experimentally obtained values of k^a by
o
estimating the value of a from the correlation developed by Calderbank
\pjvr-p02'
a = 1.44x
<j
.0.6
r    a05
\UrJ
(H)
where the terminal velocity, ut, was taken constant of 0.265 m/s.
Table 2. Hatta numbers.
T,°C	40	60
(|)g, L/h	Ha,/	Ha,/
200	0.15	0.24
300	0.16	0.23
500	0.14	0.21
Table 2 presents Ha at different gas flow rates and temperatures. It can be seen that the Hatta number is in the range between 0.14-0.24, that is, an intermediate region , where 0.05 < Ha< 1, a range to which corresponds an enhancement factor of 1. The conclusion is that oxidation of ETQH2 is a moderately slow reaction occurring mainly in the bulk liquid, whose rate is affected by the mass transfer across the liquid film. Therefore, both interfacial area and liquid hold-up should be maximized in the industrial reactor.
Extraction. A standard method was used to determine the equilibrium concentrations of H2O2 in working solution and in water. From the data obtained the equilibrium diagram (Figure 5) was constructed for the system Y [kg ^C^/kg H2O]-C [kg ^C^/m WS]. It can be described by linear relation
Y = KC                                                                                                  (12a)
and
y
Y--
i-y
(12b)
M. Krajne, J. Golob, A. Golja, R. Burja, M. Ložar, S. Flajs, M. Grčar: Separation Processes in H202...
592
Acta Chini. Slov. 2001, 48, 585-596.
where the equilibrium constant K at 50 °C was found to be 0.16 m3 WS/kg H20 at Y < 0.7 (y = 41.2%).
Figure 5.
C [kgH,0,/m3WS]
Equilibrium diagram.
It is known from previous work that the resistance in organic phase is a limiting step in a mass transfer of hydrogen peroxide from organic to water phase, where the mass transfer coefficient may be in the range of 10" - 10" m/s depending on hydrodynamics (mass transfer in jet regime volume or dispersion layer). Furthermore,
7       "i
the interfacial area is also dependent on process conditions (~100 m /m of dispersion). Due to diversity of both regimes on the tray the study of the kLa and mass transfer of H2O2 dependence on organic phase flow is very complicated. To define the region and dependence trends an industrial experiment was employed. However, to successfully perform an industrial experiment some key limiting factors of the process should be known, i.e., the coalescence and hydraulic limitations on maximum allowable concentration of organic phase leaving the extraction column which should not be higher than 0.2 g/L.
Hydraulic operation of column can be predicted from semiempirical relationship of
II   17
dispersion phase height, Ah, and organic phase flow rate. (J>, (Eq. 13) ' and energetic balance, which can be described as a relation between the height of organic phase below
1 ^
the perforated plate, ho, and organic flow rate (Eq. 14).
M. Krajne, J. Golob, A. Golja, R. Burja, M. Ložar, S. Flajs, M. Grčar: Separation Processes in H202...
Acta Chini. Slov. 2001, 48, 585-596.
593
Ah=Kt -<f>3[m3 /h] where are
(13)
K^tsep[s\
1
3600 C A, h
=58.5x10"
m
time of the dispersion separation tsep = 36 s with hold-up of the dispersed phase h = 0.5,
'j l'i
specific characteristic constant C = 0.15 (m/s)   , and cross-sectional area of Ak =
3.14 m2
and
h0[m] = K2-<f>2[m3/h]                                                                                        (14)
where is
~h2
K,
Pd
A2f2a2gAp36002
=15.5x10"
m
where a = 0.67, pd = 950 kg/m , area of perforations Af = 0.05 m , Ap = 200 kg/m  and hc - water phase layer above the plate, where 0.9 m is the distance between two plates hc = 0.9-h0-Ah                                                                                     (15)
110              120               130               140               150               160
[m/h]
Figure 6.          Heights of layers on a tray at different organic flow rates.
M. Krajne, J. Golob, A. Golja, R. Burja, M. Ložar, S. Flajs, M. Grčar: Separation Processes in H202...
594
Acta Chini. Slov. 2001, 48, 585-596.
It is reasonable to introduce an idea of pseudo equilibrium line for the extraction columns that are known to operate at efficiency of 0.2-0.25. Pseudo equilibrium line represents actual compositions at tray assuming that the efficiency is equal at all trays. The relation between pseudo equilibrium line and efficiency may be described as
Y/
/ =
K
(16)
c(\-e)+ey/k
where are fxK slope of the pseudo equilibrium line, K slope of the equilibrium line, E
efficiency and C,Y compositions in organic and water phase at tray, respectively. For E =
0.23, C = 6.52 kgH202/m3, Y = 0.54 kgH202/kgH20 and K = 0.16 m3/kgH20 fis
equal to 0.58 and corresponding slope of pseudo equilibrium line is 0.093 m /kg H20. A
slope of operating line can be calculated from mass balance
L = AY =   0.54-0 V    AC    6.52-0.19
where L and V are organic and water phase, respectively, and their ratio is equal to
0.085 m /kg H20. Lines are presented in Figure 7 where one may find 26 real contact
stages. Analysis of the samples of organic and aqueous phase at chosen trays did not
correspond to compositions found at pseudo equilibrium line, what does not confirm the
assumption of equal efficiency at all trays.
equilibrium line pseudo equilibrium line - operating line
C [kgH,0,/rnWS]
Figure 7.          Equilibrium, pseudo equilibrium and operating lines.
M. Krajne, J. Golob, A. Golja, R. Burja, M. Ložar, S. Flajs, M. Grčar: Separation Processes in H202...
Acta Chini. Slov. 2001, 48, 585-596.
595
The study of the flow rate effect on the H2O2 concentration of organic phase in the effluent gave the results in the form of mass transfer presented in Table 3, where is
Wc=t{e.52-Corg)[kglh]                                                                        (19)
Table 3. Mass transfer and relevant process criteria at different flow rates.
4> [m3/h ]	ho [m]	Ah [m]	L-org? exp. [mol/L]	WCMlo2 [kg/h] overall	WkhA [kg/h] average per tray
130	0.26	0.13	0.18	822	31.6
150	0.35	0.23	0.18	950	36.5
160	0.4	0.2	0.19	1000	38
From the data in Table 3 may be concluded that the column is operating at the upper hydraulic (ho = 0.4 m), coalescence (Ah = 0.23 m) and outlet concentration (Corg. = 0.19 g/L) conditions.
Conclusions
The value of the Hatta number is often used as a criterion for reactor selection. In the case of fast film reactions a reactor with a high interfacial area should be selected. For a slow bulk reaction the liquid hold-up is to be maximized. For the intermediate region, as the one presented in this article, a reactor with both the interfacial area and liquid hold-up being large should be chosen. In this region, the enhancing effect changes significantly with the Hatta number. Therefore, reactors within this region should not be selected on the basis of Hatta number only.
The raise in the flow of organic phase gives the higher mass transfer arising from one or all the factors in kLaxtcontact product. Consequently, knowing transport coefficients and interfacial area is not enough to predict the process in the extraction column. It is necessary to know the factors that affect the product kLaxtcontact.
Acknowledgement
M. Krajne, J. Golob, A. Golja, R. Burja, M. Ložar, S. Flajs, M. Grčar: Separation Processes in H202...
596
Acta Chini. Slov. 2001, 48, 585-596.
This work was supported by Belinka Perkemija, d.o.o. under the contract 2809/2000.
References
1.    U.S. Patent 3752885, 1973; Liebert, M., Delle H., Kabisch, G.
2.    U.S. Patent 3880596, 1975; Liebert, M., Delle H., Kabisch, G.
3.    U.S. Patent 5399333, 1995; Kato, T, Ishiuchi, Y., Takamiya, H, Ogasawara, K, Minato, K.
4.    U.S. Patent 5902559, 1999; Ledon, H.
5.     Oinas, P., Wild, G, Midoux, N., Haario, H., Chem. Eng Proc. 1995, 34, 503.
6.     Santacesaria, E., Ferro, R., Ricci, S., Carra, S. Ind. Eng. Chem. Res. 1987, 26, 155.
7.     Gaddis, E.S., Chem. Eng. Proc. 1999, 38, 503.
8.     Calderbank, P. H., Trans. Inst. Chem. Engrs. 1958, 36, 443.
9.     Charpentier, J.C., laAdvances in Chemical Engineering, Vol. II; Academic: New York, 1981.
10.   Golob, J., Grilc, V, Modic, R., Int. Solv. Extr. Conf. ISEC '83, Denver, Colorado, Book of Abstracts, 1983, 4a.
11.   Golob, J., R. Modic, Trans. IChemE 1977, 55, 207
12.  Zuiderweg, F. J., Fysische scheidingsmethoden; Technische Hogesschool Delft, Delft, 1980.
13.  Perry, R. H., Green, D, Perry's Chemical Engineers' Handbook, 6th Edition, McGraw-Hill, New York, 1984.
Povzetek
V članicu so prestavljeni osnovni fizikalno-kemijski in procesni parametri, ki so pomembni za povečevanje kapacitete v proizvodnji kapacitete vodikovega peroksida. Limitna procesa sta oksidacija in ekstrakcija. Oksidacija poteka predvsem v glavni masi tekoče faze, njeno hitrost pa omejuje absorpcija kisika v tekočo fazo. Za povečevanje industrijskega reaktorja je zato potrebno zagotoviti čim večjo medfazno površino in sloj tekoče faze. Pri študiju ekstrakcije je bil uveden pojem psevdoravnotežne črte. Ugotovljeno je bilo, da poznavanje transportnih koeficientov in medfazne površine ne zadošča za napoved dogajanja v ekstrakcijski koloni.
M. Krajne, J. Golob, A. Golja, R. Burja, M. Ložar, S. Flajs, M. Grčar: Separation Processes in H202...