Batch Liquid-Liquid Extraction 39

Mauri S. A. Palma1 Research Article

Cintia Shibata1
José L. Paiva2 Batch Liquid-Liquid Extraction of Phenol from
Mario Zilli3
Attilio Converti3 Aqueous Solutions
1
Department of Biochemical The aim of this work is the study of batch liquid-liquid extraction of phenol from
Pharmaceutical Technology, aqueous solutions in a bench-scale well-mixed reactor. The influence of the ratio
Faculty of Pharmaceutical of phase volumes, temperature, and rotational speed on phenol removal (0.72–
Sciences, University of São 1.1 % w/w) was investigated using methyl isobutyl ketone as an extracting
Paulo, São Paulo–SP, Brazil. solvent. For this purpose, the ratio of phase volumes were set at 0.1 and 0.2, the
2
Department of Chemical temperature at 10, 20, and 30 °C, and the rotational speed at 300, 400, and
Engineering, Polytechnical 500 rpm. A physical model based on the material balance of the phases as well as
School, University of São the equation of mass flux between the phases allowed the estimation of the overall
Paulo, São Paulo–SP, Brazil. coefficient of mass transfer coupled with the superficial area. Moreover, it proved
3
Department of Chemical and to fit, satisfactorily well, the experimental data of residual phenol concentration
Process Engineering, in the organic phase versus time under all the conditions investigated.
University of Genoa, Genoa,
Italy. Keywords: Batch process, Liquid-liquid extraction, MIBK, Phenol
Received: May 29, 2009; revised: July 8, 2009; accepted: July 27, 2009
DOI: 10.1002/ceat.200900279

1 Introduction The various techniques available for the treatment of pheno-

lic effluents can be subdivided into two main categories, the
Phenol and other phenolic compounds are common constitu- destruction and the recovery methods [5].
ents of wastewater from various kinds of industries and pro- Among the destruction methods, there are biological treat-
cess operations, such as chemical (polymeric resin, bisphenol ments [6–8], incineration, ozonization in the presence of UV
A, alkyl phenols, caprolactams, adipic acid, etc.), petrochem- radiation, oxidation with wet air [5], and electrochemical oxi-
ical (oil refining), metallurgical (smelting, iron, steel, and dation [9]. On the other hand, the recovery methods include
coke), pharmaceutical, textile, plastic, explosive, coffee, ceram- liquid-liquid extraction [10–12], adsorption and electro-ad-
ic, paint and varnish, pesticide production industries, and elec- sorption with activated charcoal [13, 14], ionic exchange with
trolytic strip tin-coating plants [1]. resins [15], and membrane processes, such as pervaporation
Phenols released into the environment may directly or indi- [16] and extraction with membrane [17], supported liquid
rectly cause serious health and odor problems. They can, in membrane [18], and liquid membrane in emulsion [19].
fact, inhibit growth of or exert lethal effects on aquatic organ- Phenol removal and recovery by liquid-liquid extraction,
isms even at relatively low concentrations (5 to 25 mg/L, which started to be used during the Second World War [10], is
depending on the temperature and the organism state of economically feasible with respect to other techniques when
maturity), and impart off flavors in drinking water and food phenolic effluents are highly concentrated (> 1000 ppm) and/
processing water [2]. or released at high flow rates [19]. Such a process is basically
Eleven of these compounds are among the 129 major pollu- made up of three unit operations: (a) extraction of phenol
tants present in the list of the Environmental Protection from the contaminated stream by a solvent, which gives an
Agency [3]. Most of the overall world production of phenol, extract and a raffinate phase, (b) phenol separation from the
which was 7.78·106 tons in 2001, is related to the production solvent (extract), and (c) separation of the solvent present in
of bisphenol A (39 %), phenolic resins (27 %), caprolactam the treated effluent (raffinate) [20]. The operation of phenol
(16 %), alkylphenols (5 %), 2,6-xylenol (3 %), and anilines recovery from the solvent can be performed in different ways,
(2 %) amongst others (8 %) [4]. among which is a second extraction by solvent, distillation,
evaporation or chemical reaction [21].
The main physicochemical characteristics, which have to be
taken into consideration to select the solvent suited to the
– extraction, are the coefficient of solute distribution between
Correspondence: Prof. A. Converti (converti@unige.it), Department the solvent and the aqueous phase, solubility in water, inter-
of Chemical and Process Engineering, University of Genoa, via Opera facial tension, viscosity, boiling point, latent heat of vaporiza-
Pia 15, 16145, Genoa, Italy. tion, and relative specific mass with respect to water.

Chem. Eng. Technol. 2010, 33, No. 1, 39–43 © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.cet-journal.com
40 M. S. A. Palma et al.

In this work, three sets of tests were carried out in a bench-    

dCAI KaV II I qI V I qI V I I
scale, mechanically-mixed reactor to check the influence of the ˆ C A 1 ‡ m m C mCAII
dt VI qII V II qII V II A
ratio of phase volumes, temperature, and rotational speed on
(6)
phenol removal from aqueous solutions with methylisobutyl-
ketone, using the phenol concentration in the organic phase to where CAI° and CAII° are the initial concentrations (% w/w) of
compare the results. The aim of this study was to investigate phenol in phases I and II, respectively.
the kinetics of phenol removal and the time to reach steady Defining a as the ratio between the volumes of phases II and I:
state conditions so as to apply this process to the treatment of
polluted wastewater.
a = VII / VI (7)

and substituting into Eq. (6), we obtain the equation:

2 Theory
   
dCAI qI qI I0
Mass transfer in a heterogeneous system in a mixed vessel is a ˆ Kaa CAI 1 ‡ m II m C mCAII0 (8)
dt q a qII a A
complex phenomenon that depends on the composition of
the system (solute/solvent/aqueous phase), properties of
fluids, and mechanical characteristics (geometry, mixing, etc.)
[22]. From the mass balance of Eq. (1), the concentration of phe-
For batch extraction in a closed vessel, the mass balance of nol in the organic phase can be obtained:
solute A (phenol) is described by the equation:
CAI0 CAI qI


CAII ˆ ‡ CAII0 (9)
d qI V I CAI d qII V II CAII a qII
ˆ (1)
dt dt
Numerical integration of Eq. (8), taking into account Eq. (9)
where VI and VII are the volumes (m3) of the aqueous phase and neglecting the term CAII° = 0, allowed us to obtain the
(I) and the solvent (II), CAI and CAII are the concentrations time behavior of CAII and the related value of Ka.
(% w/w) of solute A in phases I and II, respectively, and qI In particular, for short times (t → 0), combination of
and qII are the densities of phases I and II (kg/m3); t is the time Eqs. (8) and (9) provides:
(s).
The flux of mass transfer can be expressed as: dCAII qII Io
ˆ Ka C (10)

dt qI A
d qI V I CAI

ˆ KAqI CAI mCAII (2)
dt
Then, the initial slope of any CAII curve versus time has to
where: be proportional to the initial concentration in the aqueous
phase and to Ka, which emphasizes the importance of short
m = CAIeq/CAIIeq (3)
time data for accurate determination of this parameter.
is the coefficient (dimensionless) of phenol distribution be-
tween phases I and II, which depends on phenol concentra-
tions at equilibrium, CAIeq and CAIIeq (% w/w). K is the overall
3 Materials and Methods
coefficient of mass transfer (m/s) and A is the interfacial area
of mass transfer (m2). 3.1 Chemicals and Equipments
The specific interfacial area of mass transfer, a, is defined as
The chemical compounds employed in this work were methy-
the ratio of A to VII:
lisobutylketone (MIBK) of commercial grade, pure phenol for
A analysis (Labsynth Ltda., Diadema, Brazil), and distilled water.
aˆ (4) The experimental setup of the equipment utilized to carry
V II
out the liquid-liquid extraction tests is depicted in Fig. 1. Ac-
and is expressed in m–1. cording to the literature [22, 23], a standard Rushton mixed
Therefore, the coefficients of mass transfer coupled with the vessel was used. The mixed vessel (A), having a 5-L working
total area, KA, and the superficial area, Ka, which are expressed volume, was made of glass and was equipped with an external
in m3/s and s–1, respectively, can be correlated to each other by jacket for recirculation of thermostatted water. It was equipped
the equation: with an upper stainless steel plate with 5 holes to host a
mechanical mixer (B), a temperature sensor (C), two tubes for
KA the feed of aqueous phenol solution (about 1 % w/w) and
Ka ˆ (5)
V II solvent, and a syringe sampling device (D). At the bottom of
the vessel, there was a valve (E) for sampling both phases.
Making reference to a dilute system, substitution of Eqs. (4) Temperature was controlled by means of a thermostatic bath
and (5) into Eq. (2) gives: (F) that allowed one to perform tests at variable temperatures

www.cet-journal.com © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Chem. Eng. Technol. 2010, 33, No. 1, 39–43
 Batch Liquid-Liquid Extraction 41

(10–30 ± 1 °C). The mechanical mixer was moved by an engine analysis of results was conducted using standard deviations to
with rotational speed variable from 100 to 1500 rpm. A pH evaluate their reproducibility. Standard deviations with respect
sensor (G) and a conductometer (H) were positioned in the to the mean values varied from 2.6 to 11.3 %.
water phase since the solute distribution coefficient is a func-
tion of both pH and electrical conductivity of the medium.
3.3 Analytical Methods

Phenol concentrations in the organic phase was determined by

means of a gas chromatograph, model CG 3400 CX (Varian,
Walnut Creek, CA), equipped with a melted silica capillary
column, model CP-Sil 8 CB Low Bleed/MS (CP 5862) (Var-
ian), having a 30-m length and a 0.25-mm inner diameter, and
containing a 1 lm-thick film of polyethylene glycol.
The concentration of phenol in the starting aqueous phases
was checked by a Total Organic Carbon (TOC) analyzer, mod-
el 5000A (Shimadzu, Kyoto, Japan), taking into account the
stoichiometric carbon content of phenol.

3.4 Calculations

The curves of phenol concentration in the solvent versus time

were obtained by numerical integration of Eq. (8). The coeffi-
cient of mass transfer coupled with superficial area, Ka, was
obtained by fitting the experimental data of phenol concentra-
tion in the organic phase by means of the above mathematical
Figure 1. Experimental setup of the equipment used for liquid- model under the conditions of the dynamic regime.
liquid extraction of phenol from aqueous solutions. A = extrac-
tion vessel, B = mechanical mixer, C = temperature sensor, D =
sampling device, E = valve, F = thermostatic bath, G = pH sen- 4 Results and Discussion
sor, H = conductometer.
To check the ability of Eqs. (8) and (9) to describe the extrac-
tion process, a preliminary batch run was conducted in tripli-
3.2 Liquid-Liquid Extraction Tests cate at a = 0.1, N = 400 rpm, and T = 20 °C, whose results were
expressed in terms of phenol concentration in the organic
The aqueous solution of phenol was fed into the mixed ther- phase versus time (see Fig. 2). Using for the distribution coeffi-
mostatted vessel. The volume of aqueous phenol solution was cient, m, the values suggested by Won and Prausnitz [24],
set at 5.0 L and that of the solvent at 1.0 or 0.5 L, according to which varied according to concentration between 0.011 to
circumstances, corresponding to ratios of phase volumes, a, of 0.024, such a mathematical model allowed us to estimate
0.2 and 0.1, respectively. Ka = 0.040 s–1 by the least square method and then KA =
Once the selected temperature was achieved in the medium, 2.0·10–5 m3/s. It is noteworthy that there was a reasonable
the solvent was added in the shortest time possible, through a
separation funnel and then the extraction time was
measured by a chronometer. After given time in- Table 1. Experimental conditions of phenol liquid-liquid extraction tests and re-
tervals, we interrupted mixing and waited for the lated values of mass transfer coefficients coupled either with total area (KA) or
separation of phases (about 1 min). The organic with superficial area (Ka).
(superior) phase was then sampled for determina-
Test N a T CAI° KA Ka
tion of the concentration of phenol by gas chroma- [rpm] [°C] [% w/w] [× 10–5 m3s–1] [× 10–2 s–1]
tography. Samples (5 mL) were transferred to glass
tubes sealed with caps, containing about 1 g of an- 1 400 0.1 20 1.1 2.0 4.0
hydrous sodium sulfate needed to remove any 2 500 0.2 30 0.84 11 11
trace of water. Samples were stored in dark bottles
at about 5 °C and used for gas chromatographic 3 300 0.1 30 0.89 1.4 2.8
analysis within 5 days. After sampling, mixing and 4 400 0.1 30 0.80 2.7 5.4
time measurements re-started, and the tests pro-
5 500 0.1 30 0.76 4.0 8.0
ceeded up to the next sampling.
The experimental conditions under which the 6 500 0.2 10 0.73 2.5 2.5
tests were conducted are summarized in Tab. 1. All
7 500 0.2 20 0.72 8.0 8.0
tests were performed in triplicate and the statistical

Chem. Eng. Technol. 2010, 33, No. 1, 39–43 © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.cet-journal.com
42 M. S. A. Palma et al.

14 from 4.0·10–5 to 10·10–5 m3/s and of Ka from

8.0·10–2 to 10·10–2 s–1, respectively, likely because
12 the increased solvent volume with respect to that
of the water phase could have enlarged the inter-
10 facial area available for mass transfer.
Tab. 2 shows the results obtained varying the
CA II (% w/w)

8 rotational speed (N) in the range 300–500 rpm, at

0.76 ≤ CAI° ≤ 0.89 % (w/w), and keeping constant
6 the other conditions (a = 0.1, T = 30 °C). As ex-
pected from better mixing of the reaction medium,
4 an increase in N progressively shortened the time
needed to reach equilibrium conditions from
2 about 25 min at N = 300 rpm to 10 min at
N ≥ 400 rpm. This result is in accordance with the
0 increase in Ka with N (see Tab. 1), as the likely
0 5 10 15 20 25 30 35 40 consequence of a size reduction of the organic
Time (min) phase drops dispersed within the aqueous phase
associated with the increased tangential speed of
Figure 2. Time behavior of phenol concentration in the solvent phase, CAII, the paddles, i.e., increased Reynolds number.
during liquid-liquid extraction of phenol with MIBK at N = 400 rpm, T = 20 °C, On the other hand, the different values of final
CAI° = 1.1 %, a = 0.1. (—) model fitting. equilibrium concentration of phenol in the solvent

reproducibility of the experimental data of this test and that Table 2. Effect of rotational speed (N) on the performance of
equilibrium conditions were achieved after about 15 min. phenol liquid-liquid extraction carried out at 30 °C and a = 0.1,
The influence of the ratio of phase volumes, a, on the using MIBK as an extracting solvent.
extraction process is evident in Fig. 3, which illustrates the
Test N CAI° CAII t
experimental data of tests carried out at T = 30 °C and [rpm] [% w/w] [% w/w] [min]
N = 500 rpm. It should be noted that equilibrium occurred for
both runs after only ∼10 min or even sooner, because of the 3 300 0.89 9.7 25
relatively high temperature and rotational speed employed, 4 400 0.80 8.4 10
and that the mathematical modeling gave satisfactory fitting.
These results demonstrate that the system exhibited, under 5 500 0.76 7.1 10
these two conditions (a = 0.1 and 0.2), similar hydrodynamic I
CA ° = Initial concentrations of phenol in the aqueous phase.
behavior and practically the same dispersion characteristics of CAII = Phenol concentration in the solvent phase.
the organic phase. It should be noted that an increase in a t = Time to reach almost equilibrium conditions.
from 0.1 to 0.2 resulted in a corresponding increase in KA

12 phase were due to the slightly different values of

CAI°.
Finally, Tab. 3 shows the results obtained at a =
10
α = 0.1 0.2 and N = 500 rpm, varying temperature from 10
α = 0.092
to 30 °C. It is evident that the temperature did not
8 have any clear effect because it notoriously influ-
CA (% w/w)

ences, at the same time, a large number of factors,

6 αα = 0.2
0.19 namely the equilibrium position, solute solubility,
II

transport properties, and dispersion in the organic

4 phase. These results agree with those of Yang et al.
[20], who did not observe any significant improve-
2 ment of phenol extraction with MIBK, consequent
to a temperature variation in the range 25–70 °C.
0
Moreover, they clearly show that a temperature in-
crease from 10 to 30 °C reduced the time needed to
0 5 10 15 20 25 30 35 40
reach equilibrium conditions from 15 to 4 min.
Time (min) The values of the determination coefficient were
acceptable under most conditions (0.89 ≤ r2 ≤
Figure 3. Influence of the ratio of the phase volumes, a, on phenol concentration 0.98), which means that the model was suitable to
in the solvent phase, CAII, during liquid-liquid extraction of phenol with MIBK describe the experimental data of phenol extrac-
performed at N = 500 rpm and T = 30 °C. (~) CAI° = 0.84 %, a = 0.2; (䊏) CAI° =
tion with MIBK from water.
0.76 %, a = 0.1; (—) model fitting.

www.cet-journal.com © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Chem. Eng. Technol. 2010, 33, No. 1, 39–43
 Batch Liquid-Liquid Extraction 43

Table 3. Effect of temperature (T) on the performance of phenol References

liquid-liquid extraction carried out at N = 500 rpm and a = 0.2,
using MIBK as an extracting solvent.
[1] S. R. Rosatto, R. S. Freire, N. Duran, L. T. Kobuta, Química
Test T CAI° CAII t Nova 2001, 24 (1), 77.
[°C] [% w/w] [% w/w] [min] [2] L. Theodore et al., in Perry’s Chemical Engineer’s Handbook
(Eds: R. H. Perry, J. O. Maloney), 7th ed., Mc Graw-Hill,
6 10 0.73 4.14 15 New York 1997.
7 20 0.72 4.09 5 [3] S. H. Lin, C. L. Pan, H. G. Leu, J. Hazard. Mater. 1999, B65
(3), 289. DOI: 10.1016/S0304-3894 (98)00273-8
5 30 0.76 4.76 4
[4] Anonymous, Chemical Week 2002, 64 (2), 31.
CAI° = Initial concentrations of phenol in the aqueous phase. [5] S. C. Saxena, C. K. Jotshi, Prog. Energy Combust. Sci. 1996, 22
CAII = Phenol concentration in the solvent phase. (5), 401. DOI: 10.1016/S0360-1285 (96)00007-X
t = Time to reach almost equilibrium conditions. [6] M. Zilli, B. Fabiano, A. Ferraiolo, A. Converti, Biotechnol.
Bioeng. 1996, 49 (4), 391. DOI: 10.1002/(SICI)1097-0290
(19960220)49:4<391::AID-BIT5>3.0.CO;2-K
[7] M. Zilli, A. Converti, in The Encyclopedia of Bioprocess Tech-
5 Conclusions nology: Fermentation, Biocatalysis, and Bioseparation (Eds:
M. C. Flickinger, S. W. Drew), Wiley, New York 1999.
In conclusion, the experimental data of batch tests of phenol
[8] R. S. Freire, R. Pelegrini, L. T. Kubota, N. Duran, Química
liquid-liquid extraction from aqueous solutions with MIBK
Nova 2000, 23, 504.
proved to be satisfactorily reproducible, thus demonstrating
[9] Z. Mojovic, A. Milutinovic-Nikolic, S. Mentus, D. Jovanovic,
the feasibility of the proposed methodology. However, to make
Chem. Eng. Technol. 2009, 32 (5), 738. DOI: 10.1002/
it economically and environmentally feasible at an industrial
ceat.200800546
scale, solvent has to be removed from the aqueous raffinate
[10] D. Mackay, M. Medir, in Handbook of Solvent Extraction
and phenol from the solvent, for example, through a second
(Eds: T. C. Lo, M. H. I. Baird, C. Hanson), 1st ed., Wiley,
extraction with NaOH.
New York 1983.
The physical model utilized in this study, which was based
[11] H. Jiang, Y. Fang, Y. Fu, Q. X. Guo, J. Hazard. Mater. 2003,
on the material balance of the phases as well as the equation of
B101 (2), 179. DOI: 10.1016/S0304-3894 (03)00176-6
mass flux between the phases, allowed estimation of the coeffi-
[12] H. Jiang, Y. Tang, Q. X. Guo, Sep. Sci. Technol. 2003, 38 (11),
cient of mass transfer coupled with superficial area, Ka, by fit-
2579.
ting the experimental data of residual phenol concentration in
[13] E. Ayranci, B. E. Conway, J. Electrochem. Chem. 2001, 513
the organic phase under the conditions of the dynamic regime.
(2), 100. DOI: 10.1016/S0022-0728 (01)00529-0
Moreover, it proved to describe, sufficiently well, phenol liq-
[14] A. K. Jain, S. Bhatnagar, A. Bhatnagar, J. Colloid Interface Sci.
uid-liquid extraction with MIBK under different conditions.
2002, 251 (1), 39. DOI: 10.1006/jcis.2002.8395
As expected, Ka increased with the ratio of the phase vol-
[15] D. T. Palepu, S. P. Chauhan, K. P. Amanth, in Industrial Pol-
umes and the rotational speed, likely because of the enlarged
lution Prevention Handbook (Ed: H. M. Freeman), 1st ed.,
interfacial area available for mass transfer and the reduced size
McGraw-Hill, New York 1995.
of solvent drops, respectively, whereas the overall effect of tem-
[16] N. C. Pradhan, C. S. Sarkar, S. Niyogi, B. Adhikari, J. Appl.
perature was difficult to interpret due to its simultaneous
Polym. Sci. 2002, 83 (4), 822. DOI: 10.1002/app.10055
influence on a large number of parameters.
[17] A. Boam, S. Zhang, The Chemical Engineer 2001, 720, 40.
[18] W. Kujawski et al., Desalination 2004, 163 (1–3), 287. DOI:
10.1016/S0011-9164 (04)90202-0
Acknowledgements [19] S. E. Kentish, G. W. Stevens, Chem. Eng. J. 2001, 84 (2), 149.
DOI: 10.1016/S1385-8947 (01)00199-1
The authors thank the Fundação de Amparo à Pesquisa do Es-
[20] C. Yang, Y. Qian, L. Zhang, J. Feng, Chem. Eng. J. 2006, 117
tado de São Paulo (FAPESP) for the financial support.
(2), 179. DOI: 10.1016/j.cej.2005.12.011
[21] M. S. A. Palma, J. L. Paiva, M. Zilli, A. Converti, Chem. Eng.
The authors have declared no conflict of interest.
Proc. 2007, 46 (8), 764. DOI: 10.1016/j.cep.2006.10.003
[22] P. A. Quadros, C. M. S. G. Baptista, Chem. Eng. Sci. 2003, 58
(17), 3935. DOI: 10.1016/S0009-2509 (03)00302-6
[23] E. Sada, H. Kumazawa, T. Amano, Chem. Eng. J. 1977, 13
(1), 41. DOI: 10.1016/0300-9467 (77)80007-5
[24] K. W. Won, J. M. Prausnitz, J. Chem. Thermodyn. 1975, 7,
661.

Chem. Eng. Technol. 2010, 33, No. 1, 39–43 © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.cet-journal.com