A. Senol et al., Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions, Chem. Biochem. Eng. Q.

, 30 (3) 317–330 (2016) 317

Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions

Using Tri-n-propyl amine/Diluent and Dibenzyl amine/Diluent Systems
A. Senol,* M. Bilgin, B. Baslioglu doi: 10.15255/CABEQ.2015.2223
Department of Chemical Engineering, Original scientific paper
Faculty of Engineering, Istanbul University, Received: April 29, 2015
34320 Avcilar, Istanbul, Turkey Accepted: August 31, 2016

Reactive extraction of valeric acid from water by tri-n-propyl amine (TPA) and
dibenzyl amine (DBA) dissolved in polar oxygenated aliphatic diluents (diethyl sebacate,
diethyl succinate, diethyl malonate, ethyl caprylate, ethyl valerate and isoamyl alcohol)
has been studied at= T 298 ± 0.2 K and= Pp 101.3 ± 0.7 kPa . Distribution data have been
subjected to formulation of an optimization structure for effective acid separation. The
optimization approach uses separation ratio R and synergistic enhancement SE ­factors to
efficiently identify optimum extraction ranges. Among the examined alipha­­tic ester and
alcohol diluents, monoesters exhibit higher solvation efficiency comprising ­acid1–amine1
complex formation, while isoamyl alcohol yields larger loading factors. The uptake ca-
pacity of the amine/diluent system is ranging in the order TPA > DBA.
Modeling efforts based on the mass-action law principles have shown considerable
success. The mass action law chemodel and modified Langmuir approach are quite accu-
rate yielding mean errors of 0.9 % and 0.7 %, respectively.
Key words:
extraction, valeric acid, amine, modeling, optimization

Introduction herent to this uncoupling, the complementary solva-

tion effect of an oxygen-containing, substituted
Long-chain saturated aliphatic tertiary amines aromatic or aliphatic diluent is considered to be an
(e.g., Alamine 336; 308) dissolved in protic, important complexation factor.1–14 However, the ex-
non-protic, and polar diluents are effective extract- perimental findings of Yang et al.4 and Senol et
ants for carboxylic acids.1–14 The amine extractants al.7–14 give evidence for the reversible complexation
have been widely used for the extractive recovery between the tertiary amine and the non-dissociated
of carboxylic acids from aqueous solutions, such as part of the acid in the organic phase being overly
fermentation broth and wastewater including lower sensitive to the solvation efficiency of diluent. The
than 10 % (w/w) acid content.1–8 The experimental extraction power of an amine/diluent system has
findings of King and co-workers1–3 and Senol and been found to decrease in the order: valeric acid >
co-workers7–14 have revealed that the characteriza- butyric acid > propionic acid > levulinic acid > ace-
tion of acid–amine complexation is intimately con- tic acid > formic acid.4,7–14
nected to the polarity and hydrogen bonding affinity Comprehensive studies of reactive extraction
of diluent. Three major factors have been found to systems covering diluents from protic, non-protic,
influence the equilibrium characteristics of amine po­­lar, and inert classes reveal that the stoichiometry
extraction of carboxylic acids from aqueous solu- of the acid–amine complexes is intimately connected
tions, i.e., the nature of acid, concentrations of the to the strength of the complex solvation by the dilu-
acid and extractant, and the type of diluent.1–14 Si- ent increasing in the order: aliphatic hydrocarbon <
multaneously, the influence of additional controlling alkyl aromatic < halogenated aromatic < ketone <
factors, such as temperature, pH, swing effect of a proton-donating halogenated aliphatic hydrocarbon
mixed diluent and a third phase formation can also < nitrobenzene < alcohols.1–18 The effect of diluent
modify the reversible complexation stage.15 The im- is mainly focused on its ability to solvate polar ion-
plementation of amine extraction method argues an pair organic species through dipole-dipole interac-
uncoupling of the behaviors relative to the diluent/ tion or hydrogen bonding, favoring the formation of
complex interaction from the physical solubility of one or simultaneously at least two acid–carrier
the acid to establish a sentence structure distin- complexes.2–14 The experimental findings manifest
guishing the dominating factors of extraction. In- the fact that a polar diluent is capable of increasing
*
Corresponding author (A. Senol), E mail: [email protected], the extracting power of non-polar amines by pro-
Fax: 90 212 4737180 viding additional solvating power that allows higher

A. Senol et al., Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions
317–330
318 A. Senol et al., Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions, Chem. Biochem. Eng. Q., 30 (3) 317–330 (2016)

levels of polar acid–amine complexes to stay in the TPA/diluent and DBA/diluent solvent systems can
organic phase.2–5,7–14 improve the extraction efficiency of the considered
Distribution of valeric acid between water and hydrophobic acid. The distribution data have been
basic amine extractants tri-n-propyl amine (TPA) used to develop an equilibrium model for reactive
and dibenzyl amine (DBA) dissolved in polar oxy- extraction of valeric acid. However, this study also
genated aliphatic diluents, as well as the extraction deals with a new conceptual definition for optimum
capacity of pure diluent alone have been studied at extraction as the locus of the proposed separation
isothermal conditions = ( T 298 ± 0.2 K ). This arti- factors being used as the optimization criteria for
cle will also discuss the effect of the solvent struc- the considered reactive extraction system. The opti-
ture on the extraction power of amine/diluent sys- mization problem is solved both graphically and
tem, as well as the competition between physical ­analytically along with considering a non-homoge-
extraction and chemical interaction regarding the neous differential equation to represent conformably
behavior of both protic and non-protic polar alipha­ the non-linear variation profile of the optimized
tic solvents, diethyl sebacate, diethyl succinate, di- quantity. In order to accomplish this task, the deriva­
ethyl malonate, ethyl caprylate, ethyl valerate, and ti­­ve variation method has been applied to identifying
isoamyl alcohol. Experimental data for the present the optimization range.
systems of TPA/diluent/valeric acid and DBA/dilu- The possibility of achieving a synergism by a
ent/valeric acid are not available in the open litera- solvent mixture composed mainly of a commercial
ture. amine and a conventional diluent has led to exten-
It may be desirable to use a high-boiling sive research over optimum recovery of valeric acid
amine-reactive extractant that does not have to be from aqueous solutions depending on various fac-
distilled so long as no azeotropes appear. Regarding tors like the types and concentrations of amine, di-
the technical and economic merits of high-boiling luent and acid, temperature, etc.19,20 It can be ob-
amine extractants during the regeneration by distil- served from the experimental results, reported by
lation, the selection of TPA (Tb = 429 K) and DBA Luque et al.19 and Senol20 for amine extraction of
(Tb = 573 K) of higher boiling temperatures than valeric acid by commercial extractants Amberlite
water (Tb = 373 K) was made. In addition, four se- LA-2/toluene and Alamine 336/diluent, that the ef-
lected non-protic oxygen-containing diluents – di- fect of temperature and aqueous-phase acid concen-
ethyl sebacate (Tb = 585 K), diethyl succinate (Tb = tration on the phase equilibrium should be accus-
491 K), diethyl malonate (Tb = 472 K) and ethyl tomed to a regularly small change in separation
caprylate (Tb = 480 K) have higher boiling tempera- efficiency of amine because of a limited acid solu-
tures than valeric acid (Tb = 461.5 K) and water, bility in water, therefore, the extraction at different
whereas the polar ethyl valerate (Tb = 418 K) and temperatures and aqueous acid concentrations has
isoamyl alcohol (Tb = 403 K) diluents have higher not been studied here.
vapor pressures than valeric acid. In particular, sol- The properties of amine/diluent/valeric acid
vents used in the present extraction systems should system of hydrogen-bond formation can be estimat-
be of low cost, low toxicity, and rather high boiling ed through theoretically-based models depending
temperature properties, while their viscosities and on the mass action law methodology14,21 and linear
densities should be close to those of water. Howev- solvation energy relation (LSER) principles.22,23 Ex-
er, they should give proper liquid-liquid equilibrium tensions to these models for predicting the phase
(LLE) data for the excellent design and productive behavior of reactive extraction systems containing
operation of the related extraction equipment. especially amine/polar or non-polar diluent/mono-
carboxylic acid have expanded the model’s versatil-
As a continuation of the previous study7–14, the
ity.8,11,12,19,20 In this study, the results were correlated
present work aims to generate new LLE data for the
in terms of a chemodel and a modified Langmuir
reactive extraction of valeric acid from water at
model, and checked for consistency in reproducing
T 298 ± 0.2 K and=
= P 101.3 ± 0.7 kPa using
the observed optimization quantity.
TPA/diluent and DBA/diluent solvent systems of
lower vapor pressure (higher boiling temperature)
than water. No dependable results were found in the Theoretical
literature for the studied extractants and diluents ap-
plied to the valeric acid extraction. It is, therefore, Criteria of extraction efficiency
of interest to extend the previous works to accom-
modate the additional data on the amine extraction The results were interpreted in terms of the dis-
of valeric acid from the aqueous solution and to tribution ratio D, the degree of extraction E, the
model analytically the properties of relevant reac- overall loading factor of the amine Zt, the stoichio-
tive extraction systems. Due to the synergistic effect metric loading factor Zs and the chemical separation
of physical extraction and chemical interaction, factor s fchem representing the acid separation due to a
A. Senol et al., Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions, Chem. Biochem. Eng. Q., 30 (3) 317–330 (2016) 319

chemical interaction.7–14 We also introduce here the alone, respectively. C AM

0
and C TA
0
stand for the initial
physical separation factor s fphys to characterize the concentrations of the extractant and acid, respec-
degree of physical extraction. Additionally, two new tively. CHA, CTA and CTA represent the concentrations
factors including both chemical (s fchem) and physical of the complexed acid, the overall extracted acid by
(s fphys) interaction terms are used to account for the the amine/diluent system and the aqueous phase
optimum amine extraction of valeric acid, namely, acid content, respectively. Concentrations are given
the separation ratio optimization factor R and the in (mol dm–3) unit. The species in the organic phase
synergistic enhancement factor SE. are presented by the overbar.
D and E factors, defined by eqs. (1a) and (1b), Traditionally, D and E account for the distribu-
respectively, are widely used for representing the tion effect of acid, whereas Zt is a measure of the
effectiveness of an extraction process. The overall effectiveness with which the amine loading is con-
loading factor Zt is the ratio of total amount of acid tributed to the acid extraction. s fchem and s fphys factors
extracted C TA to the initial organic phase concentra- are fair measures of relative contribution of chemi-
tion of the amine (AM) C AM 0
, eq. (1c). The stoichio-
cal and physical interactions to the overall ex-
metric loading factor Zs is the ratio of the overall
traction process, respectively. The competition be-
complexed acid in the organic phase to the initial
tween chemical interaction and physical extraction
amount of the amine, eq. (1d). This factor includes
a correction term ( vCTA d
) for the amount of acid ex- is quantitatively measured by R, while SE fairly
tracted by the diluent in the solvent mixture. The accounts for the effectiveness with which valeric
chemical separation factor s fchem stands for the ratio acid can be synergistically extracted by amine/dilu-
of the complexed acid with the extractant CHA to the ent system. Firmly, D, E and Zt are essential in eval-
overall extracted acid in the organic phase CTA, uating the phase behavior of a reactive extraction
eq. (1e). As reciprocal of s Fchem, the physical separa- system, but they are not exactly sufficient for the
tion factor s fphys accounts for the degree of physical- description of the physical event. Here, Zs some-
ly extracted acid portion by the diluent, eq. (1f). A what gives evidence for a probable stoichiometry
relative proportion between chemical and physical (or degree) of chemical aggregation between acid
interactions is evaluated in terms of the separation and amine, while s fchem characterizes quantitatively
ratio optimization factor R, eq. (1g). For an equiva- whether the chemical association is formally domi-
lent contribution of both physical and chemical in- nated over physical interaction or not. Whereas R
teractions, R = 1. The synergistic enhancement fac- and SE are originally intended for the description of
tor SE, eq. (1h), is a measure of the synergistic the optimum extraction field of relevant systems. In
extraction power of amine/diluent system. general, the examined optimization structure using
R and SE gives a realistic picture of whether an op-
D = CTA CTA (1a) timum point exists and how its values can be deter-
mined, and this process brings a new perspective to
 C 
=E 100 CTA C
= 0
TA(100 1 − TA
0 )(%) (1b) literature. Although R and SE are about equally
 CTA  strongly dependent on s fchem, they exhibit opposite
0
variation profiles and different curve slopes with an
Z t = CTA CAM (1c) increase in amine concentration, so that the relative
proportion between chemical and physical interac-
d
CTA − vCTA CHA tions and the degree of synergism can be quantified
=Zs = (1d) and thus independently applied to the estimation of
0 0
CAM CAM optimum conditions. The present study differs from
the literature works, as it investigates and compares
sfchem = CHA CTA (1e) special amine extractants and ester diluents simulta-
neously on an efficiency-basis using eight extraction
sfphys = 1 − sfchem (1f) factors, which imperatively take a role in the de-
sfchem scription of phase behavior of a complex reactive
R= (1g) extraction system composed of (tertiary or second-
sfphys ary amine/monobasic ester or dibasic ester or alco-
1 1 hol/valeric acid). This is a stringent test of the suit-
SE
= chem
+ phys (1h) ability of TPA/diluent and DBA/diluent systems for
sf sf
reactive extraction of valeric acid, and the applica-
where v and CTA d
designate the volume fraction of bility of the proposed optimization structure to the
diluent in the solvent mixture and the amount of description of optimum extraction field in terms of
acid extracted by the pure (amine-free) diluent the above-mentioned factors.
320 A. Senol et al., Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions, Chem. Biochem. Eng. Q., 30 (3) 317–330 (2016)

Equilibrium models of mass action law tions for p = 1 – k and q = 1 – l should not require
to be explicitly evaluated.7–14 In the prediction
Using the chemical modeling concepts of mass of equilibrium, different sets of an appropriate
action law7–14, the overall extraction equilibrium of acid-amine aggregation have been selected for
valeric acid/TPA/diluent and valeric acid/DBA/dilu- valeric acid, regarding the overall loading region
ent systems can be described by a complex forma- and the maximum loading values, i.e. the plateau of
tion through the interfacial reaction, eq. (2). The the loading curve. Accordingly, aggregation of sim-
conditioned extraction constant β pq including the ple complexes into larger adducts has been assumed.
activity coefficients of species is defined in the mo-
The equilibrium data for the reactive extraction
larity scale (mol dm–3)1–p–q by eq. (3):
of valeric acid can be interpreted in terms of the
pHA + q NR 3 =
(HA) p (NR 3 ) q mass action law chemodel given by eq. (7) associat-
, (2a) ed with eqs. (4) and (5) to achieve a model structure
=p 1,=
k ; q 1, l involving Z t , CHA and CTA 0
quantities.

pHA + q NR 2 H =
(HA) p (NR 2 H) q
, (2b)
=p 1,=
k ; q 1, l
(7)

β pq = C pq (C HA
p q
CAM = )
, p 1,=
k ; q 1, l (3)

where HA, NR 3 and NR 2 H represent the non-dis-

sociated acid in the aqueous phase, tri-n-propyl where Cd vD 0 CTA
= 0
/ (1 + D0 ) represents the con-
amine (TPA) and dibenzyl amine (DBA), respec- centration of the physically extracted acid part by
tively. (HA) p (NR 3 ) q and (HA) p (NR 2 H) q stand the diluent. D0 is the distribution ratio of the acid
for the acid-amine complexes. The overbar denotes referred to the diluent alone. The adjustable ex-
species in the organic phase. CHA , C AM and C pq des- traction constant β pq for the relevant system has
ignate the equilibrium concentrations of non-disso- been regressed due to eq. (7) supposing one or two
ciated acid in the aqueous phase, non-complexed (p, q) acidp–amineq complex formation.
amine and acid–amine (p, q) complex, respectively. In a similar way, the Langmuir equilibrium
The total equilibrium content of complexed acid model of Bauer et al.25 has been applied to relevant
(C HA), is the sum of contributions of the individual extraction systems along with considering an over-
complexes defined by eq. (4). all acid-extractant complexation with an associated
k l
q
number (z) relative to the maximum loading of the
CHA = ∑ ∑ pβ pq CHA p CAM (4) extractant, z = Z s,max . By incorporating CHA from
=p 1=q 1 the Bauer model to the Z t factor, a unified model
By incorporating eq. (4) into the overall bal- structure eq. (8) is derived involving both physical
ance equation for acid, the equilibrium model is de- and chemical interaction terms.
rived, eq. (5), along with considering equal volumes
of organic and aqueous phases. Cd + CHA 0
vD0CTA z β L CHA z
=Zt = + (8)
0
CTA =Cd + CHA + CTA (5)
0
CAM (1 + D0 ) CAM
0 1 + β L CHA z

0 where Cd , D 0 , CAM0
, CTA
0
and CHA stand for the
where CTA , CTA and Cd stand for the initial and
same quantities as defined by eq. (7). The Langmuir
total aqueous phase acid concentrations, and con-
extraction constant β L in (mol dm–3)–z is attributed
centration related to the acid portion physically ex-
to the overall reaction eq. (2a) or (2b), supposing
tracted by the diluent in the solvent mixture, respec-
the formation of only one type ( p q = z ) of aggre-
tively. The non-dissociated aqueous phase acid
gated structure. The assumption inherent in this
concentration CHA is to be calculated from CTA, pH
­approach is attributed to an associated number of
and the dissociation constant of valeric acid (Ka) in
acid-amine complexation related to the maximum
the aqueous phase due to eq. (6) using pK a = 4.842
for valeric acid.24 CH+ is the molar aqueous-phase loading
= z Z= s,max (CHA ) 0
CAM
max
.
concentration of proton.
=CHA CTA CH+ (C H+
+ Ka ) (6) Experimental
Interpretation of the equilibrium results relative Valeric acid (99 %, GC), as well as the reactive
to amine-based reactive extraction systems has re- extractants tri-n-propyl amine TPA (98 %) and
vealed that all possible acid-carrier (p, q) combina- dibenzyl amine DBA (98 %), and six organic sol-
 A. Senol et al., Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions, Chem. Biochem. Eng. Q., 30 (3) 317–330 (2016) 321

vents – diethyl sebacate (98 %), diethyl succinate Ta b l e 1 – Mutual solubility of binaries (w1 water + w2 sol-
(>98 %), diethyl malonate (99 %), ethyl caprylate vent) in terms of mass fraction (w) at T = 298 K
(98 %), ethyl valerate (>98 %), and isoamyl alcohol and P = 101.3 kPaa
(99 %) of analytical grade purity were furnished by Solvent (2) Water (1)
Merck and Aldrich. All the chemicals were used as in water (1) in solvent (2)
Binary system
received without further purification. Deionized and
redistilled water was used in all experiments. w1 w2 w1 w2

The extraction experiments were performed us- Water + diethyl sebacate 0.9990 0.0010 0.0019 0.9981
ing an equilibrium glass cell equipped with a mag-
Water + diethyl succinate 0.9981 0.0019 0.0104 0.9896
netic stirrer and thermostatted at= T 298 ± 0.2 K.26
Equal volumes (10 cm ) of initial aqueous valeric
3 Water + diethyl malonate 0.9806 0.0194 0.0200 0.9800
acid and organic (TPA/diluent or DBA/diluent) Water + ethyl caprylate 0.9979 0.0021 0.0036 0.9964
phases were agitated for 1 h and then left for 2 h to
Water + ethyl valerate 0.9942 0.0058 0.0082 0.9918
settle down into aqueous and solvent layers at a
fixed temperature = ( T 298 ± 0.2 K ) and pressure Water + isoamyl alcohol 0.9784 0.0216 0.1030 0.8970
( P 101.3 ± 0.7 kPa ).26 The contact time enough to
= Water + tri-n-propyl amine 0.99927 0.00073 0.00061 0.99939
reach equilibrium, and the waiting time required to
Water + dibenzyl amine 0.99990 0.00010 0.00014 0.99986
separate the conjugate phases were determined in
preliminary analysis, which were found to be suffi-
a
Standard uncertainties u are u (T ) = 0.2 K , u ( P ) = 0.7 kPa,
cient for a complete extraction.26 The effective sep- u ( w) = 0.002.
aration of the phases was ensured by centrifugation.
Aqueous-phase pH was measured using an Orion ter and alcohol diluents for the identical experimen-
601A pH-meter. Aqueous-phase acid concentration tal conditions at A:O = 1 : 1 (v/v),= T 298 ± 0.2 K,
was determined by titration with aqueous NaOH =P 101.3 ± 0.7 kPa and the initial aqueous acid
(Titrosol A, Merck) and phenolphthalein indicator, solution of CTA0
= 0.3329 mol dm–3-3
used as a simu-
as well as using an UV-spectrophotometer (Perkin lated synthetic fermentation sample. The solubilities
Elmer Lambda 35 Model). The organic phase acid of the extractant, diluent and organic complex in the
concentration was analyzed by Hewlett-Packard GC aqueous phase are negligible in the range of the
Analyzer, Model 5890A, equipped with FID and a variables investigated. Similarly, the change in the
capillary column, HP1-type 50 m × 0.2 mm × 0.5 phase volume was neglected. This is confirmed by
µm. Nitrogen was used as a carrier gas at a flow the experimental results in Table 1 for the mutual
rate of 5 mL min–1. The initial amine/diluent content solubility of (water + solvent) binaries obtained by
in the organic phase was determined gravimetrically the cloud point method.26,27
by weighing with a Sartorius scale accurate to with-
in ±10–4 g, in addition to chromatographical analy- As reported by Luque et al.19 and Senol20, the
sis using Hewlett-Packard GC Analyzer. variation of temperature or aqueous-phase acid con-
centration can produce a slight effect on removal
The acid analysis was checked by a mass bal-
efficiency of valeric acid by commercial amine ex-
ance. It was confirmed from two independent repli-
tractants (Amberlite LA-2, Alamine 336), therefore,
cates that the valeric acid extraction experiments
the extraction equilibrium at different temperatures
were reproducible within at most 3 % standard devi-
and aqueous acid concentrations will be according-
ation.26 The initial acid content in the aqueous phase
was kept at CTA ly considered redundant and not studied here.
0
= 0.3329 mol dm -3–3. To eliminate a
third phase formation, the initial amine concentration
was restricted in the range of 0.25–1.05 mol dm–3.
Results and discussion
Tests covering the influence of the solvent structure
and concentration on the extraction degree of valer-
Factors affecting the extraction power
ic acid were performed using protic (isoamyl alco- of TPA and DBA
hol), and polar and proton-accepting (diethyl seba-
cate, diethyl succinate, diethyl malonate, ethyl The equilibrium results for the extraction of
caprylate, ethyl valerate) oxygenated diluents for valeric acid by pure diluent alone and amine/diluent
two basic amine extractants TPA and DBA. The mixture are provided in Table 2 and Figs. 1–4.
physical extraction of the acid by the diluent alone Study of the extraction systems given in Table 2
was also studied. However, the relative dependence containing CTA0
= 0.3329 mol dm -3–3 initial aqueous-
of the extraction efficiency on the structural proper- phase acid solution and TPA or DBA dissolved in
ties of the carrier and diluent has been elucidated by diethyl sebacate, diethyl succinate, diethyl malonate,
comparing the extraction capabilities pertaining to ethyl caprylate, ethyl valerate, and isoamyl alcohol
TPA and DBA dissolved in the above-mentioned es- diluents reveals that the physical extraction of valer-
322 A. Senol et al., Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions, Chem. Biochem. Eng. Q., 30 (3) 317–330 (2016)

Ta b l e 2 – Variation of extractability factors (E, D, Zt) with concentration of components for extraction of valeric acid by tri-n-pro-
0
pyl amine (TPA)/diluent, dibenzyl amine (DBA)/diluent, and pure diluent alone at T = 298  ±  0.2 K ( CTA = 0.3329 mol dm –3 )a

0
CAM b CTA d E 0
CAM b CTA d E
pH c D Zt pH c D Zt
(mol dm ) –3 (mol dm ) –3 (%) (mol dm ) –3 (mol dm ) –3 (%)

TPA + Diethyl sebacate DBA + Diethyl sebacate

0.0000e 3.41 0.0282 91.53 10.805 0.0000 e

3.41 0.0282 91.53 10.805

0.2605 3.49 0.0241 92.76 12.813 1.185 0.2576 3.52 0.0209 93.72 14.928 1.211
0.5197 3.56 0.0175 94.74 18.023 0.607
0.5244 3.54 0.0198 94.05 15.813 0.597
0.7884 3.63 0.0134 95.97 23.843 0.405
0.7891 3.57 0.0157 95.28 20.204 0.402
1.0409 3.68 0.0106 96.82 30.406 0.310
1.0553d 3.68 0.0113 96.61 28.460 0.305
DBA + Diethyl succinate
TPA + Diethyl succinate
0.0000 e
3.45 0.0263 92.10 11.658
0.0000e 3.45 0.0263 92.10 11.638 0.2581 3.54 0.0195 94.14 16.072 1.214
0.2591 3.52 0.0207 93.78 15.082 1.205 0.5214 3.58 0.0153 95.40 20.758 0.609
0.5299 3.57 0.0142 95.73 22.444 0.601 0.7823 3.67 0.0118 96.46 27.212 0.410
0.7894 3.71 0.0087 97.39 37.264 0.411 1.0410 3.70 0.0109 96.73 29.541 0.309

1.0522 3.89 0.0050 98.50 65.580 0.312 DBA + Diethyl malonate

0.0000 e
3.44 0.0271 91.86 11.284
TPA + Diethyl malonate
0.2597 3.52 0.0191 94.26 16.429 1.208
0.0000e 3.44 0.0271 91.86 11.284
0.5217 3.54 0.0173 94.80 18.243 0.605
0.2640 3.58 0.0154 95.37 20.617 1.203
0.7869 3.56 0.0165 95.04 19.176 0.402
0.5246 3.67 0.0110 96.70 29.264 0.614
1.0415 3.59 0.0146 95.61 21.801 0.306
0.7886 3.70 0.0083 97.51 39.108 0.412 DBA + Ethyl caprylate
1.0571 3.82 0.0063 98.11 51.841 0.309 0.0000 e
3.32 0.0383 88.50 7.692
TPA + Ethyl caprylate 0.2568 3.48 0.0265 92.04 11.562 1.193
0.0000 e
3.32 0.0383 88.50 7.692 0.5198 3.55 0.0181 94.56 17.392 0.606

0.2653 3.51 0.0242 92.73 12.756 1.164 0.7828 3.57 0.0147 95.58 21.646 0.406
1.0415 3.64 0.0116 96.52 27.698 0.308
0.5269 3.63 0.0128 96.15 25.008 0.608
DBA + Ethyl valerate
0.7916 3.78 0.0075 97.75 43.387 0.411
0.0000e 3.45 0.0262 92.13 11.706
1.0558 3.86 0.0055 98.35 59.527 0.310
0.2603 3.56 0.0178 94.65 17.702 1.211
TPA + Ethyl valerate
0.5211 3.58 0.0144 95.67 22.118 0.611
0.0000e 3.45 0.0262 92.13 11.706 0.7864 3.61 0.0123 96.31 26.065 0.408
0.2668 3.65 0.0125 96.25 25.632 1.201 1.0411 3.65 0.0108 96.76 29.824 0.309
0.5255 3.73 0.0074 97.78 45.986 0.619 DBA + Isoamyl alcohol
0.7905 3.94 0.0046 98.62 71.370 0.415 0.0000 e
3.64 0.0127 96.19 25.213

1.0513 4.11 0.0031 99.07 106.387 0.314 0.2602 3.66 0.0108 96.76 29.824 1.238
0.5203 3.70 0.0094 97.18 34.415 0.622
TPA + Isoamyl alcohol
0.7812 3.73 0.0084 97.48 38.631 0.415
0.0000e 3.64 0.0127 96.19 25.213
1.0408 3.79 0.0072 97.84 45.236 0.313
0.2577 3.68 0.0098 97.06 32.969 1.254
a
Initial concentration of valeric acid. Initial concentration of
b

0.5264 3.71 0.0086 97.42 37.709 0.616 amine dissolved in oxygen-containing diluent. c Aqueous phase
acidity. d Aqueous phase acid concentration; organic phase acid
0.7972 3.75 0.0073 97.81 44.603 0.408 concentration C= 0
CTA − CTA . e Properties referred to pure
TA

1.0550 3.83 0.0060 98.20 54.483 0.310 diluent alone (i.e., diethyl sebacate, diethyl succinate, diethyl
malonate, ethyl caprylate, ethyl valerate, isoamyl alcohol).
A. Senol et al., Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions, Chem. Biochem. Eng. Q., 30 (3) 317–330 (2016) 323

ic acid in pure diluent alone is reasonably high with being indicative of the formation of non-overload-
a distribution ratio ( D0 ) of about 25 for isoamyl al- ed (one acid per multiple amines) valeric acid-carri-
cohol = ( D0 25.1;
= E 96.2%   %), and less than 12 for er complexes in the working range. In fact, as
others ranging from 7.7 for ethyl caprylate to 11.7 shown in Table 2 and Figs. 1 and 2, it is a foregone
for ethyl valerate. In general, a high physical ex- conclusion that D, E and sfchem proportionally in-
tractability of valeric acid by conventional solvents crease as the amine content increases, while Z t and
could be attributable to the strong hydrophobic na- Z s gradually decrease when increasing the amine
ture of the acid due to a long R-chain structure and concentration in the organic phase. In general, this
its relatively low ionizing strength ( pK a = 4.842). type of phase behavior is prevalent for amine/dilu-
This behavior is also related to the simultaneous ef- ent/acid systems, but the range of increasing (or de-
fect of several specific solvent characteristics, such creasing) of D, E, sfchem , Z t and Z s is intimately
as polarity and hydrophobicity, probably varying connected to the types and concentration levels of
dependently with the functional group configuration amine, diluent, and acid.7–14,20 But unexpectedly,
in the solvent structure. It turns out from the results here it is observed that, in the amine load working
given in Tables 1 and 2 that the solubilities of water interval 0.25 − 1.05 mol dm–3 -3
, the overall loading
and valeric acid in the organic phase vary with the factors almost invariably range between
structural properties and polarity of conventional 0.3 〈 Z t 〈 1.3 for all the examined amine/diluent
solvents, following approximately the order: (i) for mixtures. As seen in Fig. 1, the stoichiometric load-
water, alcohol > dibasic ester > monobasic ester; ing factors Z s for the studied amine concentration
(ii) for valeric acid, alcohol > monobasic ester > level are markedly low, ranging Z s 〈 0.12 . Regard-
dibasic ester. This is perhaps not surprising consi­
ing Fig. 2, the same remarks hold for the chemical
dering the more polar structure of water as com-
separation factors s fchem varying in the range s fchem ≈
pared to that of valeric acid and organic solvent. As
evident from Table 1, except for isoamyl alcohol, 0.07 – 0.30, which in turn manifests the fact that the
there is less tendency of the ester and amine coex- contribution of chemical interaction to the overall
traction in the aqueous phase, but inevitably water acid separation is about 3–12 times smaller than
should carry a small amount of acid in the organic that of physical extraction. It is concluded from
phase containing especially isoamyl alcohol, which Figs. 1–3 that all the studied amine/diluent solvent
can affect the phase behavior in an unexpected way. mixtures generally exhibit low chemical interaction-
dependent factors 0.06 〈 Z s 〈 0.12, 0.07 〈 sfchem 〈 0.3
Inspection of the experimental results in Table
2 reveals that, for both TPA/diluent and DBA/dilu-
ent systems, isoamyl alcohol and ethyl valerate di- 0.12
luents allow for achieving larger separation factors0.12 A
D and E as compared to those of dibasic ester and 0.11 A
0.11
ethyl caprylate. Accordingly, the smallest extraction
0.10
efficiency displayed by amine/diethyl sebacate and0.10
amine/ethyl caprylate would likely come from the
Zs

0.09
Zs

long R-chain structure of these diluents that is even-0.09

0.08
tually responsible for a steric hindrance in the or-0.08
ganic phase. As shown in Table 2, a regular decrease 0.07
in the overall loading factor Z t with increasing the0.07
amine concentration CAM 0
is also observed, but the0.06 0.06
0.11
range of decreasing is about equally large for all the0.11 B
B
examined TPA/diluent and DBA/diluent systems. 0.10
0.10
Regarding the magnitude of extraction factors
0.09
in Table 2, it turns out that the largest extraction0.09
Zs
Zs

power is exhibited by TPA/isoamyl alcohol or DBA/ 0.08

isoamyl alcohol yielding D 〉 29 , Z t 〉 0.3 and0.08
E 〉 95 95% % due to a complementary interaction effect0.07 0.07
of hydroxyl (OH), carboxyl (COOH), and amine
0.06
(NR3, NHR2) functional groups, promoting simulta-0.06
neously physical extraction and chemical interac- 0.2 0.4 0.6 ______0.8 1.0 1.2
0.2 0.4 0.6 ______0.8 C 0 1.0 -3 )
(mol dm1.2
tion.19,20 On the other hand, considering here Z s and C AM
0 dm-3 )
(molAM
sfchem  to be quantitative criteria for the degree of acid-
amine aggregation, it is observed from Figs. 1 and 2 F i g . 1 – Variation of stoichiometric loading factor Z s with
0
organic phase amine concentration
Figure 1 C AM (mol dm–3); A. TPA/
that isoamyl alcohol diluent exhibits unimpressive Figure 1
diluent, B. DBA/diluent; diethyl sebacate, diethyl succi-
solvation efficiency towards the acid-base complex- nate, diethyl malonate, ethyl caprylate, ethyl valerate,
es related to small Z s ≈ 0.07 and sfchem ≈ 0.06 − 0.21, isoamyl alcohol; (CTA 0
= 0.3329 mol dm –3 )
324 A. Senol et al., Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions, Chem. Biochem. Eng. Q., 30 (3) 317–330 (2016)

and 0.06 〈 R 〈 0.39 indicating that moderately strong 0.30

interactive forces should dominate during valeric0.12 A
A
acid-amine complexation. Specifically, it is pre-0.11 0.24
sumed that the formation of non-overloaded (1:1

s chem
and 1:2) or (1:1 and 1:3) ion-pair acid-amine com-0.10 0.18

f
plexes with a lowered solvation degree by the dilu-0.09

Zs
0.12
ent would likely proceed at the complaxation stage
of extraction. Together with this, the differences0.08 0.06
among D and SE factors provided in Table 2 and0.07
Fig. 4 indicate that the complementary effect of 0.00
chemical interaction and physical extraction is a0.06 0.30
B
critical factor for TPA- or DBA-based reactive ex- 0.11
0.24
B
traction of valeric acid. 0.10

s fchem
However, it is essential that this phenomenon 0.18
should have a significant impact on the implemen- 0.09
Zs
tation of a selected extraction method. As seen in0.08 0.12

Table 2 and Figs. 1 and 2, valeric acid is physically 0.06

easier to extract by the protic isoamyl alcohol alone0.07
as compared to others, whereas the magnitude of 0.00
the acid-amine complexation is relatively larger for0.06 0.2 0.4 0.6 ______0.8 1.0 1.2
amine/monobasic ester (ethyl caprylate and ethyl 0.2 0.4 0.6 ______0.8 C 0 1.0(mol dm -3
1.2 )
valerate) regarding the magnitude of Z s and sfchem C AM
0 (molAMdm-3 )
factors. It turns out from Figs. 1 and 2 that the con- F i g . 2 – Variation of chemical separation factor sfchem with
trolling factor for acid-amine chemical association organic phase amine Figure 2 C 0 (mol dm–3); A. TPA/
Figureconcentration
1 AM
is the solvation efficiency of diluent ranging as fol- diluent, B. DBA/diluent; diethyl sebacate, diethyl succi-
lows: ethyl caprylate > ethyl valerate > diethyl suc- nate, diethyl malonate, ethyl caprylate, ethyl valerate,
0
cinate ≈ diethyl malonate > diethyl sebacate > iso- isoamyl alcohol; (CTA = 0.3329 mol dm –3 )
amyl alcohol. It is worth mentioning here again that
SE is quantitatively assessed as a criterion for a syn-
ergism in extraction since it includes both chemical
and physical interaction terms. In view of SE fac- 0.40
tors in Fig. 4, the synergistic extraction power of 0.12 A
the solvent mixture increases in the order: isoamyl 0.32 A
0.11
alcohol > dibasic ester > monobasic ester, and TPA
> DBA. Typically, TPA is a more effective separa- 0.10
0.24

tion agent than DBA for the identical diluents

R
Zs

0.16
checked according to sfchem , D and SE factors. This
0.09

could be attributable to the steric hindrance and the 0.08 0.08

resonance π electron effect in the DBA structure
and the more structured (or equivalent less polar) 0.07
0.00 36
formula of TPA due to the larger number of organic 0.06 0.40
B
35

radicals in the latter. Unfortunately, there is a lack 0.11

B
of experimental data reflecting the range of the ste- 0.32
0.10
ric hindrance and resonance effect provided by an 0.24
extractant in solution; however, here the geometric
R

0.09
Zs

structure of the solvent molecule is thought to be an 0.16

appropriate reference of the expected magnitude of 0.08
the steric effect along the entire composition range. 0.08
0.07
It is apparent from Fig. 1 that the maximum 0.00
stoichiometric loading ( Z s,max ) corresponding to the0.06
0.2 0.4 0.6 ______0.8 1.0 1.2
plateau in the loading curve appears at Z s,max 〈 1 , 0.2 0.4 0.6 ______0.8 C AM 1.2-3 )
0 1.0 (mol dm
signifying a tendency toward the formation of two C AM
0 (mol dm ) -3

types of non-overloaded valeric acidp–amineq struc-

tures, i.e. an aggregation related to one acid per F i g . 3 – Variation of separation ratio optimization factor R
Figure 3 0
with organic Figure
phase 1amine concentration C AM (mol dm–3);
multiple amines ( p : q = 1: 2 or p : q = 1: 3), and an A. TPA/diluent, B. DBA/diluent; diethyl sebacate, diethyl
equimolar acid-amine interaction ( p : q = 1:1). This succinate, diethyl malonate, ethyl caprylate, ethyl valer-
fact seems to be a common strategy for designing ate, isoamyl alcohol; (CTA 0
= 0.3329 mol dm –3 )
A. Senol et al., Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions, Chem. Biochem. Eng. Q., 30 (3) 317–330 (2016) 325

the reactive extraction of valeric acid. For the sake recovery against practically permissible optimum
of simplicity, we will consider here a maximum concentration range of the solvent mixture. A feasi-
stoichiometric loading Zs,max to be 0.03 (about 22– ble way to achieve these purposes lies in processing
30 %) larger than the upper limit value of Zs ob- an effective optimization method depending on R
tained from Fig. 1 as follows: 0.104, 0.111, 0.132, and SE factors. To simplify the complexity of the
0.139, 0.139 and 0.103 for TPA/diluent and 0.117, optimization problem, only the ranges of R and SE
0.116, 0.120, 0.133, 0.121 and 0.099 for DBA/dilu- factors should be subjected to formulation of an op-
ent for diethyl sebacate, diethyl succinate, diethyl timization structure along with applying the deriva-
malonate, ethyl caprylate, ethyl valerate and isoam- tive variation method described earlier by Senol.27–29
yl alcohol, respectively. Regarding the above Zs,max This method implies that: (1) the contribution of the
values, one may conclude that the examined dilu- derivatives to the optimized property is validated by
ents affect almost invariably the diluent-complex the slope analysis, and (2) the identification of the
aggregation through hydrogen bonding or dipole-di- optimum conditions is governed by the range of
pole interaction. Weak interactive forces producing changes in the derivative value.
low solvation degree of acid-amine complexes have Development of a new conceptual definition
been obtained for all the tested solvents yielding of optimum extraction conditions for the acid re-
Z s,max 〈 0.15. In comparison with isoamyl alcohol covery requires an interpretation of R = f ( xiv ) and
and dibasic ester, monobasic ester allows for higher SE = f (1 xiv ) curves both graphically and analyti-
complex solvation efficiency in the organic phase. cally, where xiv stands for the independent variable
It ought to be pointed out that larger distribution of 0
defined as xiv = CAM CTA . To reduce the complexi-
the acid in pure diluent alone mostly is not a guar- ty of the optimization problem, an uncoupling and
antee for achieving better complexation efficiency an independent dealing with the slopes of the ob-
of the amine.20 served and modeled curves will be processed, where
Consequently, the used amines do not function the variation profile of the modeled quantity (R or
well in extracting valeric acid as proven by low Zs SE) can be expressed by a non-homogeneous differ-
values. However, isoamyl alcohol, as a less hydro- ential equation, eq. (9).
phobic diluent, can extract a significant amount of
valeric acid by itself, thus eventually improving the =R Rmax 1 − exp ( rxiv ) 
; (9)
overall extraction. The observed Zt and SE factors
from Table 2 and Fig. 4 show that TPA is a slightly = SE SEmax 1 − exp ( sxiv−1 ) 

more effective carrier than DBA for which the aro- Rmax and SEmax are the maximum values of ex-
matic ring is likely responsible for a steric effect.
For both amine samples, there is a tendency toward traction factors. The substance-dependent adjust-
an acid-carrier complexation including predomi- able coefficients r and s have been estimated by
nantly one mole acid per multiple moles extractant means of linear regression. The derivative variation
related to Z s 〈 0.12 and sfchem 〈 0.3 . test (slope analysis) of the considered variables has
been performed using linear programming algo-
Optimization criteria for amine extraction rithm.30,31 The maximum extraction factors Rmax and
of valeric acid SEmax , as well as the adjustable coefficients r and s
obtained from eq. (9) are as follows, (a) for TPA:
An important aspect of liquid-liquid equilibri- Rmax = 0.42, SEmax = 17.44 (r = –0.015357, s =
um systems containing a reactive amine extractant –33.418 diethyl sebacate); Rmax = 0.44, SEmax =
is conducted on optimizing the extraction process. 16.40 (r = –0.008187, s = –43.348 diethyl succi-
While various models, based on hydrogen-bond nate); Rmax = 0.44, SEmax = 13.26 (r = –0.009135, s
theory, group-contribution method, and dipole-di- = –58.279 diethyl malonate); Rmax = 0.49, SEmax =
pole interaction concepts exist for reactive ex- 12.21 (r = –0.009456, s = –41.618 ethyl caprylate);
traction systems, only a few works have focused on Rmax = 0.44, SEmax = 12.53 (r = –0.005289, s =
optimizing analytically the extraction efficiency of –80.832 ethyl valerate); Rmax = 0.38, SEmax = 18.55
a LLE system. However, the literature revealed very (r = –0.006974, s = –79.487 isoamyl alcohol), (b)
little insight relating to the validity of a generalized for DBA: Rmax = 0.42, SEmax = 15.32 (r = –0.014654,
method for the prediction of optimum extraction s = –38.735 diethyl sebacate); Rmax = 0.41, SEmax =
limits pertaining to reactive extraction systems.27–29 15.64 (r = –0.014060, s = –41.852 diethyl succi-
The study deals with a new conceptual defini- nate); Rmax = 0.40, SEmax = 14.95 (r = –0.017567, s
tion for optimum extraction as the locus of the new- = –39.709 diethyl malonate); Rmax = 0.46, SEmax =
ly proposed separation ratio optimization factor R 13.04 (r = –0.017632, s = –32.539 ethyl caprylate);
and synergistic enhancement factor SE being used Rmax = 0.41, SEmax = 14.76 (r = –0.013999, s =
as the optimization criteria. The goal is to determine –45.411 ethyl valerate); Rmax = 0.41, SEmax = 19.44
the most suitable extract composition for the acid (r = –0.008290, s = –71.060 isoamyl alcohol).
326 A. Senol et al., Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions, Chem. Biochem. Eng. Q., 30 (3) 317–330 (2016)

The present amine/diluent/valeric acid reactive 20

extraction system may be categorized as a system0.12 A
exhibiting large physical solubility of the distribut- 16
A
0.11
ed acid in pure diluent alone with D0 〉 7 . An inter-
pretation of the optimum conditions of relevant sys-0.10

SE
tems through analyzing the derivative variation 12

Zs
0.09
profile (slopes) of the observed (Figs. 3 and 4) and
modeled (eq. (9)) properties results in the following0.08 8
quantitative ranges of the R factor attributed to the
0.07
amine extraction of valeric acid. 4
0.06 20
0 〈 R 〈 0.5 (extractant is a poor separating agent) 0.11 B
B
0.5 〈 R 〈 1 (extractant is moderately effective) 16
0.10
1〈R〈3 (extractant is an excellent separating

SE
agent) Zs
0.09 12

R〉3 (very large extractant load is used) 0.08

8
Since the physical extraction of valeric acid by0.07
the diluent is prevalent in the examined extraction 4
systems, the ranges of the synergistic (overall) ex-0.06
0.2 0.4 0.6 ______0.8 1.0 1.2
traction factors Zt and SE decrease with increasing 0.2 0.4 0.6 ______0.8
C 0 1.0 -3 )
(mol dm1.2
the volume fraction of the amine. In this case, the C AM
0 dm-3 )
(molAM
most appropriate synergistic extraction power of
F i g . 4 – Variation of synergistic enhancement factor SE with
relevant amine/diluent system can be identified due organic phase amine concentration
Figure 4 C AM0
(mol dm–3); A. TPA/
to the SE ranges regarding the slope variation pro- Figure 1
diluent, B. DBA/diluent; diethyl sebacate, diethyl succi-
file (slope changes) of the observed (Fig. 4) and nate, diethyl malonate, ethyl caprylate, ethyl valerate,
0
modeled (eq. (9)) curves defined as follows: isoamyl alcohol; (CTA = 0.3329 mol dm –3 )

SE 〈 4 (very large extractant load is used)

4 〈 SE 〈 7 (solvent system is moderately effective)
7 〈 SE 〈 20 (solvent system is an excellent
separating agent) A - Eq. (7)
1.2
SE 〉 20 (solvent system works in favor of A - Eq. (7)
Z t (modeled)

physical extraction). 1.2

0.9

By analyzing the variation profiles of the quan-

Z t (modeled)

0.9
tities in question, it is recognized that the most ap- 0.6

propriate ranges are 0.5 RR 〈〈3R3 〈and

0.5 〈 0.5 3 4 〈 SE 〈 20 0.6
where the curve slope is changed considerably. 0.3
However, both R and SE factors are varying with 0.3 38
the extractant content in the solvent mixture, and 0.0 35
their optimum values are intimately connected to 0.0
1.2
B - Eq. (8)
the physical solubility of the acid in pure diluent B - Eq. (8)
1.2
alone. For a practically insoluble acid in the select-
t (modeled)

0.9
ed diluent alone and carrier, the optimization factors
Z t Z(modeled)

0.9
R and SE are devoid of the physical meaning. This
leaves us with the conclusion that the proposed op- 0.6
0.6
timization structure based on R and SE factors is
capable of representing reliably the behavior of a 0.3
0.3
reactive extraction system involving a physically
very soluble valeric acid in pure diluent alone with 0.0
D0 〉 7 .
0.0
0.0 0.3 0.6 0.9 1.2
0.0 0.3 0.6 0.9 1.2
Depending on the above R and SE conditions, Z t(observed)
(observed)
Zt
an analysis was conducted of the observed (Figs. 3
F i g . 5 – Reliability analysis of mass-action law models as a
and 4) and modeled (eq. (9)) performance results in plot of the model performance (Z5t) 5against the observed proper-
Figure
optimum extraction conditions dictating preferably Figure
ties: TPA/diluent, DBA/diluent. (A) Chemodel, eq. (7); (B)
the use of monobasic esters and isoamyl alcohol as Langmuir model, eq. (8)
A. Senol et al., Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions, Chem. Biochem. Eng. Q., 30 (3) 317–330 (2016) 327

Ta b l e 3 – Extraction constants βpq and bL of eqs. (7) and (8), root-mean-square deviation (σ ), and mean relative error ( e )a of ­model
estimates for valeric acid–amine complexation
Modified Langmuir model, eq. (8) Mass-action law chemodel, eq. (7)

System bL; ( z = Z s,max )b e ( Zt ) β pq1 ; ( p, q)c β pq 2 ; ( p, q)c e ( Zt )

σ ( Zt ) σ ( Zt )
( mol dm )
–3 − z
( mol dm )
–3 1− p − q
( mol dm –3 )
1− p − q
(%) (%)

TPAe + Diethyl sebacate Sd 0.3779 · 101; (0.104) 0.0005 0.08 0.4416 · 101; (1, 1) 0.019 3.85

Td 0.2254 · 101; (1, 1) 0.5198 · 101; (1, 2) 0.005 1.08

TPA + Diethyl succinate S 0.4826 · 10 ; (0.111)
1
0.0007 0.12 0.6127 · 10 ; (1, 1)
1
0.035 7.22

T 0.6105 · 100; (1, 1) 0.1597 · 102; (1, 2) 0.014 2.31

TPA + Diethyl malonate S 0.3541 · 101; (0.132) 0.0081 1.19 0.9777 · 101; (1, 1) 0.019 3.49

T 0.6017 · 101; (1, 1) 0.9271 · 101; (1, 2) 0.003 0.52

TPA + Ethyl caprylate S 0.4571 · 10 ; (0.139)
1
0.0062 1.18 0.7651 · 10 ; (1, 1)
1
0.044 8.99

T 0.5432 · 101; (1, 1) 0.2367 · 102; (1, 3) 0.010 2.09

TPA + Ethyl valerate S 0.3943 · 10 ; (0.139)
1
0.0085 1.25 0.1436 · 10 ; (1, 1)
2
0.035 6.62

T 0.3803 · 101; (1, 1) 0.3021 · 102; (1, 2) 0.006 1.18

TPA + Isoamyl alcohol S 0.3358 · 101; (0.103) 0.0022 0.30 0.9902 · 101; (1, 1) 0.011 2.04

T 0.7264 · 10 ; (1, 1)
1
0.5875 · 10 ; (1, 2)
1
0.003 0.68
DBA + Diethyl sebacate S
e
0.3470 · 10 ; (0.117)
1
0.0040 0.54 0.5694 · 10 ; (1, 1)
1
0.016 3.18

T 0.4928 · 101; (1, 1) 0.5029 · 101; (1, 3) 0.002 0.27

DBA + Diethyl succinate S 0.3512 · 10 ; (0.116)
1
0.0037 0.57 0.6195 · 10 ; (1, 1)
1
0.014 2.62

T 0.5615 · 101; (1, 1) 0.4273 · 101; (1, 3) 0.005 0.92

DBA + Diethyl malonate S 0.2860 · 10 ; (0.120)
1
0.0073 0.99 0.5212 · 10 ; (1, 1)
1
0.003 0.70

T 0.5678 · 10 ; (1, 1)
1
0.2434 · 10 ; (1, 3) 0.004
–1
0.71
DBA + Ethyl caprylate S 0.3752 · 101; (0.133) 0.0070 1.23 0.5843 · 101; (1, 1) 0.020 3.48

T 0.5228 · 101; (1, 1) 0.5339 · 101; (1, 3) 0.006 0.75

DBA + Ethyl valerate S 0.3285 · 10 ; (0.121)
1
0.0061 0.91 0.6698 · 10 ; (1, 1)
1
0.008 1.48

T 0.6553 · 101; (1, 1) 0.2563 · 101; (1, 3) 0.001 0.13

DBA + Isoamyl alcohol S 0.3467 · 10 ; (0.099)
1
0.0005 0.08 0.8551 · 10 ; (1, 1)
1
0.009 1.73

T 0.7755 · 101; (1, 1) 0.4847 · 101; (1, 3) 0.001 0.13

0.5
e =(100/N )∑ N =1 ( Z t,obs − Z t,mod ) Z t,obs , σ =  ∑ i =1 ( Z t,obs − Z t,mod ) N  .
N N 2
a
 
Langmuir extraction constant (bL) in ( mol dm –3 )
−z
b
due to eq. (8). Maximum loading values z = Z s ,max are given in parenthesis.
c
Extraction constant (βpq) in ( mol dm )
–3 1− p − q
for a given acid-amine (p, q) aggregation due to eq. (7).
d
One (S) or two (T) complex formation considered.
e
TPA, tri-n-propyl amine; DBA, dibenzyl amine.

the most appropriate diluents in combination with tested amine concentration range, indicating that
TPA or DBA. The latter diluents provide a relative- generally weaker interactive forces between the di-
ly appropriate medium for valeric acid and ac- luent and the formed complexes appear during the
id-amine complexes to stay in the organic phase as complexation stage. These concepts are supported
compared to less effective dibasic esters. Disap- by the results for SE factor from Fig. 4, signifying
provingly, all the studied diluents consistently yield- preferably an effective synergistic separation of
ed very small R factors (R R 〈 0.4 ) with regard to the valeric acid by amine/ester and amine/isoamyl alco-
328 A. Senol et al., Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions, Chem. Biochem. Eng. Q., 30 (3) 317–330 (2016)

hol systems at lower concentration levels of the car-

rier. Consequently, monobasic esters in combination The estimated β L values in (mol dm–3)–z due to
with the amines appear to be the most effective sol- eq. (8) and statistical deviation results in terms of
vating agents for the formed acid-amine complexes, Zt are given in Table 3. This table also presents the
while isoamyl alcohol yields the largest synergistic maximum loading values ( z = Z s,max ) of relevant
extraction efficiency among the considered amine/ amine/diluent systems used in eq. (8). Statistical
diluent systems. analysis of eq. (8) results in a quite precise repre-
sentation of the model performance, reproducing
Statistical analysis of equilibrium models data with an average error of e (Z t ) = 0.7 %
( σ (Z t ) = 0.005 ). The reliability analysis of eqs. (7)
The equilibrium results in Table 2 were inter- and (8) has also been performed graphically through
preted in terms of the mass action law chemodel a plot of the model performance against the ob-
and modified Langmuir approach using the re- served property with respect to the selected two (T)
gressed overall apparent extraction constants β pq complex formations for eq. (7) and one complex
and β L due to eqs. (7) and (8), respectively. Fig. 5 relative to z = Z s,max for eq. (8), as depicted in Fig. 5.
illustrates the graphical confidence tests of the mod- Both eqs. (7) and (8) are expected to be reliable
el performance against the observed property per- in data fit for the associated reactive extraction sys-
taining to eqs. (7) and (8). However, the chemodel, tems because the distribution along the diagonal
eq. (7), presumes the formation of either one or at line remained in an acceptable narrow band (Fig.
least two complexes. Estimates were performed us- 5). Further, the random pattern of comparison points
ing the multivariable procedures of linpack algo- at each side of the diagonal line implies that the ex-
rithm31 for one, two or three selected appropriate isting mass-action law models are almost free of
complex combinations regarding Zs. The best fits systematical errors. Referring to Fig. 5, one may
display the approach comprising the formation of conclude that the models yielded a relatively fair
one ( p:q = 1:1 for TPA and DBA) or simultaneous distribution verifying the goodness-of-fit.
two associated acidp–extractantq (p, q) structures of
different stoichiometry depending on the diluent
used, i.e. (a) TPA: (1, 1) and (1, 3) for ethyl capry- Conclusions
late and (1, 1) and (1, 2) for other diluents; (b)
DBA: (1, 1) and (1, 3) for all the examined diluents. A detailed study on the reactive extraction of
The Langmuir model, eq. (8), has been derived con- valeric acid from aqueous solutions by TPA/diluent
sidering the formation of only one associated struc- and DBA/diluent has been performed. The work
ture p= :q z= :1 Z s ,max . The regressed equilibrium leads to the following conclusions:
constants ( β pq ) for one (S) and two (T) selected –  The extraction efficiency of valeric acid by
individual complexes and Langmuir extraction con- amine/diluent mixture is dependent almost equally
stant ( β L ) in terms of the mean relative error strongly on the structural properties of the carrier
(100 / N ) ∑ i =1 ( Z t,obs − Z t,mod ) / Z t,obs
N
e (%) and root- and the solvation capability of the examined dilu-
0.5 ent. The physical extraction favors over the chemi-
mean-square deviation σ =  ∑ i =1 ( Z t,obs − Z t,mod ) N 
N 2

  cal interaction. The synergistic extraction power of

of the Zt factor are provided in Table 3. Referring to amine/diluent system decreases in the order: isoam-
the β pq definition and considered complex forma- yl alcohol > ethyl valerate > diethyl malonate > di-
tion from Table 3, the chemodel (eq. (7)) reproduc- ethyl succinate ≈ diethyl sebacate > ethyl caprylate.
es the overall Zt data quite accurately, yielding the The more structured TPA is slightly more effective
mean deviations of e (Z t ) = 3.8 % ( σ ( Z t ) = 0.019), than the less structured DBA. The resonance effect
and e (Z t ) = 0.9 % ( σ ( Z t ) = 0.005) for one and of aromatic π electron system and the steric hin-
two complex formation, respectively. The chemod- drance are other controlling factors for DBA.
el, eq. (7), matches the experimental data for two –  Characterization of acid-amine complexation
considered complex formations slightly more pre- is intimately connected to the polarity and hydrogen
cisely in comparison with one complex formation bonding affinity of the diluent. A favorable valeric
due to an increased number of degrees of freedom. acidp–amineq complexation pertains to the forma-
Additionally, eq. (7) reproduces the observed Z t tion of (1:1) and (1:2) complexes for TPA, and (1:1)
properties for the DBA/diluent system more reli- and (1:3) complexes for DBA.
ably, yielding e (Z t ) = 0.5 % (σ ( Z t ) = 0.003) as –  The proposed R and SE factors provide an
compared to e (Z t ) = 1.3 % (σ ( Z t ) = 0.007) for the analytical structure for prediction of optimum ex-
TPA/diluent system considering two complex for- traction. The evaluated optimization structure de-
mations. pending on R and SE factors is of particular interest
A. Senol et al., Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions, Chem. Biochem. Eng. Q., 30 (3) 317–330 (2016) 329

in dealing with complex non-linear phenomena, P –  Pressure, kPa

such as that in a reactive extraction system. p, q –  Number of acid and extractant molecules in-
–  The mass action law models, eqs. (7) and volved in the complex
(8), were able to reproduce the experimental data R –  Separation ratio optimization factor
satisfactorily yielding mean errors of 0.9 % and 0.7 r, s –  Coefficients
%, respectively. chem phys
sf , sf –  Chemical and physical separation factors of
solvent mixture
Acknowledgments SE –  Synergistic enhancement factor
T –  Temperature, K
This work was supported by the Research Fund
v –  Volume fraction of diluent in solvent mixture
of Istanbul University; Project number 4402.
xiv –  Independent variable
Y –  Independent variable
Nomenclature Zs –  Stoichiometric loading factor
Zt –  Overall loading factor of extractant
Symbols z –  Associated number
CAM –  Concentration of non-complexed amine, ( )
overbar –  Species in organic phase
mol dm–3
0
CAM –  Initial concentration of amine in solvent mix-
ture, mol dm–3 Greek letters
Cd –  Concentration of acid extracted by diluent,
βL –  Langmuir extraction constant, (mol dm–3)–z
mol dm–3
βpq –  Apparent equilibrium extraction constant,
CH + –  Proton concentration of acid in aqueous
(mol dm–3)1–p–q
phase, mol dm–3
σ –  Root-mean-square deviation,
CHA –  Concentration of undissociated acid in aque- 0.5
σ =  ∑ i =1 (Yi ,obs − Yi ,mod ) N 
N 2
ous phase, mol dm–3  
CHA –  Overall concentration of complexed acid,
mol dm–3
C pq –  Concentration of acidp–amineq complex, mol Subscripts
dm–3
CTA –  Overall concentration of acid in aqueous mod –  Modeled
phase, mol dm–3 obs –  Observed
CTA –  Overall concentration of acid in organic
phase, mol dm–3
0
CTA –  Initial concentration of acid, mol dm–3 References
d
C TA –  Concentration of acid extracted by diluental- 1. Kertes, A. S., King, C. J., Extraction chemistry of fermenta-
one, mol dm–3 tion product carboxylic acids, Biotechnol. Bioeng. 28
(1986) 269.
D –  Distribution ratio of acid relative to solvent
doi: http://dx.doi.org/10.1002/bit.260280217
mixture
2. Tamada, J. A., Kertes, A. S., King, C. J., Extraction of car-
D0 –  Distribution ratio of acid relative to pure di- boxylic acids with amine extractants. 1. Equilibria and law
luent alone of mass action modeling, Ind. Eng. Chem. Res. 29 (1990)
E –  Extraction degree of acid relative to solvent 1319.
doi: http://dx.doi.org/10.1021/ie00103a035
mixture, %
3. Tamada, J. A., King, C. J., Extraction of carboxylic acids
e –  Mean relative error, with amine extractants. 2. Chemical interaction and inter-
e = (100 N ) ∑ i =1 (Yi ,obs − Yi ,mod ) Yi ,obs , %
N
pretation of data, Ind. Eng. Chem. Res. 29 (1990) 1327.
http://dx.doi.org/10.1021/ie00103a036
HA –  Monocarboxylic acid 4. Yang, S. T., White, S. A., Hsu, S. T., Extraction of carboxyl-
(HA) p (NR 3 ) q –  Acid-amine complex ic acids with tertiary and quaternary amines: Effect of pH,
Ind. Eng. Chem. Res. 30 (1991) 1335.
(HA) p (NR 2 H) q –  Acid-amine complex doi: http://dx.doi.org/10.1021/ie00054a040
Ka –  Dissociation constant of acid 5. Bízek, V., Horácek, J., Koušová, M., Amine extraction of
citric acid: Effect of diluent, Chem. Eng. Sci. 48 (1993)
N –  Number of observation 1447.
NR3, NR2H –  Tertiary and secondary amines doi: http://dx.doi.org/10.1016/0009-2509(93)80051-Q
330 A. Senol et al., Optimal Reactive Extraction of Valeric Acid from Aqueous Solutions, Chem. Biochem. Eng. Q., 30 (3) 317–330 (2016)

6. Poposka, F. A., Nikolovski, K., Tomovska, R., Equilibrium 19. Luque, S., Alvarez, J. R., Pazos, C., Coca, J., Recovery of
and mathematical models of extraction of citric acid with valeric acid from aqueous solutions by solvent extraction,
isodecanol/n-paraffins solutions of trioctylamine, J. Chem. Solv. Extr. Ion Exch. 13 (1995) 923.
Eng. Japan 30 (1997) 777. doi: http://dx.doi.org/10.1080/07366299508918310
doi: http://dx.doi.org/10.1252/jcej.30.777 20. Senol, A., Extraction equilibria of valeric acid using (Ala-
7. Senol, A., Extraction equilibria of formic, levulinic and ace- mine 336/diluent) and conventional solvent systems. Mod-
tic acids using (Alamine 336/diluent) and conventional sol- eling considerations, Chem. Eng. Process. 41 (2002) 681.
vent systems: Modeling considerations, J. Chem. Eng. Ja- doi: http://dx.doi.org/10.1016/S0255-2701(01)00188-X
pan 32 (1999) 717.
21. Kirsch, T., Maurer, G., Distribution of oxalic acid between
doi: http://dx.doi.org/10.1252/jcej.32.717
water and organic solutions of tri-n-octylamine, Ind. Eng.
8. Senol, A., Effect of diluent on amine extraction of acetic Chem. Res. 35 (1996) 1722.
acid: Modeling considerations, Ind. Eng. Chem. Res. 43 doi: http://dx.doi.org/10.1021/ie9505827
(2004) 6496.
doi: http://dx.doi.org/10.1021/ie0400954 22. Kamlet, M. J., Doherty, R. M., Abraham, M. H., Marcus, Y.,
Taft, R. W., Linear solvation energy relationships. 46. An
9. Senol, A., Extraction equilibria of nicotinic acid using Ala-
improved equation for correlation and prediction of octa-
mine 336 and conventional solvents: effect of diluent,
Chem. Eng. J. 83 (2001) 155. nol/water partition coefficients of organic nonelectrolytes
doi: http://dx.doi.org/10.1016/S1385-8947(00)00236-9 (including strong hydrogen bond donor solutes), J. Phys.
Chem. 92 (1988) 5244.
10. Senol, A., Influence of conventional diluents on amine ex-
doi: http://dx.doi.org/10.1021/j100329a035
traction of picolinic acid, Sep. Purif. Technol. 43 (2005) 49.
doi: http://dx.doi.org/10.1016/j.seppur.2004.09.013 23. Marcus, Y., Linear solvation energy relationships. Correla-
11. Senol, A., Dramur, U., Predicting liquid–liquid equilibria of tion and prediction of the distribution of organic solutes
amine extraction of carboxylic acid through solvation ener- between water and immiscible organic solvents, J. Phys.
gy relation, Solv. Extr. Ion Exch. 22 (2004) 865. Chem. 95 (1991) 8886.
doi: http://dx.doi.org/10.1081/SEI-200030286 doi: http://dx.doi.org/10.1021/j100175a086
12. Senol, A., Extraction equilibria of formic and levulinic ac- 24. Dean, J. A., Lange’s Handbook of Chemistry, 30th ed., Mc-
ids using Alamine 308/diluent and conventional solvent Graw-Hill, New York, 1985.
systems, Sep. Purif. Technol. 21 (2000) 165. 25. Bauer, U., Marr, R., Ruckl, W., Siebenhofer, M., Reactive
doi: http://dx.doi.org/10.1016/S1383-5866(00)00200-8 extraction of citric acid from aqueous fermentation broth,
13. Senol, A., Influence of diluent on amine extraction of pyru- Ber. Bunsen-Ges. Phys. Chem. 93 (1989) 980.
vic acid using Alamine system, Chem. Eng. Process. 45 doi: http://dx.doi.org/10.1002/bbpc.19890930911
(2006) 755. 26. Baslioglu, B., Investigation of the Separation of Carboxylic
doi: http://dx.doi.org/10.1016/j.cep.2006.03.002 Acids from the Aqueous Solutions by Using Liquid Mem-
14. Senol, A., Lalikoglu, M., Bilgin, M., Modeling extraction brane Technique, Ph.D. Thesis (in Turkish), Istanbul Uni-
equilibria of butyric acid distributed between water and tri- versity, Institute of Science and Technology, 2012.
n-butyl amine/diluent or tri-n-butyl phosphate/diluent sys- 27. Senol, A., Optimum extraction equilibria of the systems
tem: Extension of the LSER approach, Fluid Phase Equilib. (water + carboxylic acid + 1-hexanol/Alamine): Thermody-
385 (2015) 153. namic modeling, Fluid Phase Equilib. 360 (2013) 77.
doi: http://dx.doi.org/10.1016/j.fluid.2014.10.043 doi: http://dx.doi.org/10.1016/j.fluid.2013.09.012
15. Heyberger, A., Procházka, J., Volaufova, E., Extraction of
28. Senol, A., Liquid–liquid equilibria for mixtures of (water +
citric acid with tertiary amines – third phase formation,
pyruvic acid + alcohol/Alamine): Modeling and optimiza-
Chem. Eng. Sci. 53 (1998) 515.
tion of extraction, J. Chem. Eng. Data 58 (2013) 528.
doi: http://dx.doi.org/10.1016/S0009-2509(97)00321-7
doi: http://dx.doi.org/10.1021/je3012265
16. Malmary, G., Faizal, M., Albet, J., Molinier, J., Liquid-liq-
uid equilibria of acetic, formic, and oxalic acids between 29. Senol, A., Optimal extractive separation of chromium(VI)
water and tributyl phosphate + dodecane, J. Chem. Eng. from acidic chloride and nitrate media by commercial
Data 42 (1997) 985. amines: Equilibrium modeling through linear solvation en-
doi: http://dx.doi.org/10.1021/je9700939 ergy relation, Ind. Eng. Chem. Res. 52 (2013) 16321.
17. Ricker, N. L., Michaels, J. N., King, C. J., Solvent proper- doi: http://dx.doi.org/10.1021/ie4014309
ties of organic bases for extraction of acetic acid from wa- 30. Chang, Y.–L., Sullivan, R. S., Quantitative Systems for
ter, J. Sep. Proc. Technol. 1 (1979) 36. Business Plus (QSB+) V 2.0, Prentice–Hall Inc., Engle-
18. Juang, R. S., Huang, W. T., Equilibrium studies on the ex- wood Cliffs, New Jersey, 1991.
traction of citric acid from aqueous solutions with tri-n-oc- 31. Himmelblau, D. M., Riggs, J., Basic Principles and Calcu-
tylamine, J. Chem. Eng. Japan 27 (1994) 498. lations in Chemical Engineering, eighth ed., Prentice–Hall
doi: http://dx.doi.org/10.1252/jcej.27.498 Inc., Englewood Cliffs, New Jersey, 2012.