Materials Transactions, Vol. 47, No. 1 (2006) pp.

136 to 142
#2006 The Mining and Materials Processing Institute of Japan

Modeling of the Equilibria of Yttrium(III) and Europium(III) Solvent Extraction

from Nitric Acid with PC-88A
Nianxin Fu1;2 and Mikiya Tanaka1; *
1
Research Institute for Environmental Management Technology, National Institute of Advanced Industrial Science and Technology (AIST),
Tsukuba, Ibaraki 305-8569, Japan
2
College of Materials and Metallurgy, Northeastern University, Shenyang, Liaoning 110006, P. R. China

The equilibria of yttrium(III) and europium(III) (RE) extraction with PC-88A dissolved in Shellsol D70 from nitric acid solutions were
studied under nonideal conditions in order to develop a chemically-based model enabling the engineering prediction of the equilibrium
distribution ratios. The distribution of the extractant between the nitrate solution and the diluent was also examined. As a result, it was found that
(i) the majority of PC-88A is dimerized in the organic phase, (ii) the cation exchange extraction is predominant in the low-acidity region, and (iii)
solvation extraction becomes appreciable in the high-acidity region. By using a model that considers these two extraction equilibria together
with the complexation of the RE with NO3
, the nonideality of the PC-88A dimer, and the competing extraction of HNO3 , the extraction
equilibrium constants and the coefficient for the effective concentration of the extractant were determined by the nonlinear least-squares method.
The proposed model can reproduce the experimental data with good accuracy for single-metal systems over wide feed concentration ranges of
RE up to 0.1 kmolm
3 , nitric acid (0.05 to 3 kmolm
3 ), and the extractant (0.25 to 1 kmolm
3 ). Its applicability to the prediction of the
extraction from a binary metal system was also confirmed.

(Received May 10, 2005; Accepted October 28, 2005; Published January 15, 2006)
Keywords: solvent extraction, yttrium, europium, PC-88A, modeling, nitric acid

1. Introduction commercial name is PC-88A, abbreviated HR) were ana-

Rare earth (RE) elements are widely used in various
advanced functional materials due to their superior character-
istics such as fluorescent, magnetic, optical, and catalytic
properties; thus, the recovery of RE metals from spent
materials is of significant interest. It is difficult, however, to
mutually separate adjacent RE metals due to the similarity in
their chemical properties. Solvent extraction using acidic
organophosphorus reagents is often employed for this
purpose on an industrial scale. Since many stages of
extraction are required to achieve sufficient separation, a
computer simulation code should be developed to find the
optimum operating conditions. Thus, it is desired to study the
extraction equilibria of RE for seeking an appropriate
chemically-based mathematical model which can be used
for the multistage extraction simulation.
Fundamental extraction studies of RE metals with acidic
organophosphorus reagents have been conducted by a
number of investigators under ideal conditions.1–7) Namely,
very low metal concentrations were adopted, while the ionic
strength of the aqueous solution was kept constant to avoid
any changes in the activity coefficients of the chemical
species participating in the reaction. Thus, the experimental
results were mathematically analyzed in order to obtain
relevant conditional equilibrium constants. On the other
hand, few studies have been carried out on the extraction
equlibrium of RE metals in macro concentrations under
practical and nonideal conditions,8–10) where the nonideality
of the aqueous phase was corrected by considering the
appropriate complexations.
In the present study, the extraction equilibria of yttri-
um(III) and europium(III) from a nitric acid solution with 2-
ethylhexyl phosphonic acid mono-2-ethylhexyl ester (its
*Corresponding author, E-mail: mky-tanaka@aist.go.jp
lyzed under the practical operating conditions for the
engineering simulation of the separation of both metals from
used fluorescent lights. The aqueous solubility of the
extractant was first investigated by measuring its distribution
ratio between the aqueous phase at various pHs and the
organic solvent. The distribution of each metal between the
aqueous and organic phases was then measured by varying
the feed concentrations of the hydrogen ion and PC-88A over
a wide range. The observed data were analyzed by the
chemically-based model to estimate the stoichiometries of
the extraction and the relevant equilibrium constants. The
proposed model was validated by comparing the calculated
distribution ratio with the experimental data. Its applicability
was also checked against different feed concentrations of the
metals in the single metal systems and in the binary metal
system.
2. Experimental Procedures
2.1 Reagents and solutions
PC-88A (Daihachi Chemicals) was used as the extractant
without further purification for the extraction experiment. For
investigating the aqueous solubility of PC-88A, however, it
was purified by washing five times with the following
solutions in sequence: the mixture of 2 kmolm
3 NaOH and
0.5 kmolm
3 NaCl, 6 kmolm
3 HCl, and then distilled
water. Shellsol D70 (Shell Chemicals) was used as a nonpolar
diluent. This diluent consists predominantly of paraffins and
naphthenes distilling around 471–523 K, and has a density of
0.792 Mgm
3 . Organic working solutions containing from
0.01 to 1.25 kmolm
3 of the extractant were prepared by
dissolving PC-88A in the diluent. The concentration of the
extractant was determined by the nonaqueous neutralization
titration.
An aqueous stock solution of RE(NO3 )3 was prepared by
 Modeling of the Equilibria of Yttrium(III) and Europium(III) Solvent Extraction from Nitric Acid with PC-88A 137

slowly dissolving the desired amount of RE2 O3 (Wako,
99.99% purity, RE ¼ Y or Eu) into excess concentrated nitric
acid heated at 423 K, followed by the removal of the excess
acid by evaporation and then by dilution with 0.1 kmolm
3
HR ð2Þ
nitric acid. Aqueous working solutions were prepared by
appropriately diluting the stock solution with nitric acid of
different concentrations for the feed concentrations of
Hþ þ R
ð3Þ
RE(NO3 )3 from 0.005 to 0.5 kmolm
3 and HNO3 from
0.05 to 3 kmolm
3 . All other chemicals were of reagent
grade.
2.2 Distribution of PC-88A
Equal volumes (15 cm3 ) of the organic working solution of
a known PC-88A concentration and the aqueous solution of
0.4 kmolm
3 (Na,H)NO3 were poured into a 50 cm3 -stop-
pered conical flask and shaken overnight in a thermostatted
waterbath at 298  0:1 K to attain equilibrium. The two
phases were then centrifuged and carefully separated by
pipetting. The pH values of the aqueous phases were
measured using an HM-60G pH meter (TOA) attached with
a glass combination electrode (TOA, GS-5017C) by immers-
ing the electrode in the solution for 5 min to obtain the stable
reading. The pH meter was calibrated beforehand by the pH
standard solutions of 1.7, 4.0, and 6.9. The concentrations of
PC-88A in the aqueous phases were determined by measur-
ing the phosphorus contents with an SPS4000 plasma
spectrometer (Seiko, ICP-AES). Since the amounts of PC-
88A dissolved in the aqueous phase were low, the method of
standard addition was employed to improve the measurement
accuracy.11) The concentrations of PC-88A in the organic
phases were calculated by mass balance.
2.3 Extraction equilibria of RE
The procedure in this experiment was similar to that in the
PC-88A distribution experiment. Equal volumes (15 cm3 ) of
the aqueous and organic working solutions were continu-
ously shaken for 3 h at 298  0:1 K to attain equilibrium.
After phase separation, the pH values of the aqueous phases
were measured in the same manner as previously described,
and the concentrations of RE in the phases were determined
by EDTA titration and by ICP-AES when they were
relatively high and low, respectively. The concentrations of
RE in the organic phases were also measured by this method
after completely stripping the metal from the organic solution
into 6 kmolm
3 HCl.
3. Results and Discussion
(2) Distribution of monomeric PC-88A into the aqueous
phase
Kd
HR
(3) Dissociation of PC-88A in the aqueous phase
Ka
HR
The total concentrations of PC-88A in the aqueous and
organic phases can be expressed using eqs. (4) and (5),
respectively, balanced with the feed concentration of PC-88A
using eq. (6)
½HRt ¼ ½HR þ ½R
 ¼ Kd ½HRð1 þ Ka =½Hþ Þ ð4Þ
2
½HRt ¼ ½HR þ 2½ðHRÞ2  ¼ ½HR þ 2Km ½HR ð5Þ
½HRf ¼ ½HRt þ ½HRt ð6Þ
where [ ] represents concentration, and the subscripts f and t
denote the feed and total amounts, respectively.
Figure 1 shows the equilibrium data as a plot of
logf½HRt =ðkmolm
3 Þg versus the equilibrium pH for vari-
ous ½HRf values. It can be seen that [HR]t depends on ½HRf
but is independent of pH, which shows according to eq. (4),
that the dissociation of PC-88A in the aqueous phase can be
ignored. Thus, [HR]t is approximated as follows:
½HRt 
¼ ½HR ¼ Kd ½HR ð7Þ
The distribution ratio of PC-88A can be given by combining
eqs. (5) and (7):
DHR ¼ ½HRt =½HRt ¼ 1=Kd þ 2ðKm =Kd 2 Þ½HRt ð8Þ
With respect to this equation, the plot of DHR vs. [HR]t (the
average value of the data given in Fig. 1) was made for each
½HRf in Fig. 2. The plotted points lie on a straight line as
expected by eq. (8). From the slope and intercept of this line,
the values of Km and Kd were evaluated as shown in Table 1
together with the constants published in the literature12,14–17)
for comparison. Unlike di(2-ethylhexyl)phosphoric acid
(henceforth D2EHPA), the constants reported for PC-88A
are very few. The scattering of these values is not small due
to the use of different aqueous phases and diluents. The
values obtained in the present work are within this range of
-3.5
0.01
[ HR ] / kmol m-3
f
0.1
-4.0 0.25
Log ([HR] / kmol m-3)

0.5
0.99
3.1 Distribution of PC-88A -4.5

The following equilibria were considered for the distribu-

t

tion of PC-88A between the aqueous and organic phases, -5.0

assuming that the extractant aggregates into the dimer in
the diluent and only its monomer can distribute into the
-5.5
aqueous phase, since it has been shown by a vapor-phase
osmometric study that PC-88A is dimerized in an aliphatic
solvent:12,13) -6.0
0.2 0.4 0.6 0.8 1.0 1.2 1.4
(1) Dimerization of PC-88A in the organic phase
pH
Km
2HR
ðHRÞ2 ð1Þ
Fig. 1 Aqueous solubility of PC-88A in 0.4 kmolm
3 (Na,H)NO3 in the
where the over bar denotes the species in the organic phase. lower pH region. ½Hþ f ¼ 0:07, 0.2, 0.4 kmolm
3 .
138 N. Fu and M. Tanaka

4.5 5.0
-3 Slope=3
4.0 [ HR ]f / kmol m 3
10-4DHR = 1.4 105 [HR]t / kmol m-3 + 0.0178 4.0

Log D - 3log {[(HR)2]/ kmol m-3}

3.5 0.25 1
0.5
3.0 3.0 1.0
10-4

0.25
2.5 0.5
2.0 1.0
HR

2.0
D

1.5 1.0
1.0
0.0
0.5
0.0 -1.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0

[HR] 105 / kmol m-3 -2.0

t
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
pH
Fig. 2 Relation between the distribution ratio of PC-88A and its total a

concentration in 0.4 kmolm
3 (Na,H)NO3 in the lower pH region.

½Hþ f ¼ 0:07, 0.2, 0.4 kmolm
3 .
scatter. High Km and low Kd values, respectively, suggest that
the majority of the PC-88A molecules are dimerized in
Shellsol D70 and that the aqueous solubility is very low. In
the following analysis, therefore, it is assumed for conven-
ience that all the PC-88A is present as dimer in the organic
phase and no PC-88A is present in the aqueous phase.
3.2 Extraction equilibria of RE
The equilibrium distribution ratios of RE were measured
using the feed concentration of 0.05 kmolm
3 Y3þ and
0.005 kmolm
3 Eu3þ for the Y(NO3 )3 and Eu(NO3 )3
systems, respectively. It is accepted that the trivalent RE is
extracted by PC-88A through the following cation exchange
reaction at the low loading ratio and low acid concentra-
tion:5,6,10)
M(NO3 )3 2nHR  ð12Þ
M3þ þ 3ðHRÞ2
MR3 (HR)3 þ 3Hþ  ð9Þ
where M denotes the RE. The apparent extraction equili-
brium constant of eq. (9), K1 , is written as

K1 ¼ ½MR3 (HR)3 ½Hþ 3 =f½ðHRÞ2 3 ½M3þ g

Suppose that M3þ and MR3 (HR)3 are the only species of RE
in the aqueous and organic phases, respectively; therefore,
the distribution ratio of RE, D, is described as
log D ¼ log K1 þ 3pHa þ 3 logf½ðHRÞ2 =ðkmolm
3 Þg ð11Þ
Here the pH value measured by the pH meter, which would
deviate from the true pH value especially in the higher-
Fig. 3 LogD
3 log½ðHRÞ2  vs. aqueous equilibrium pHa for the extrac-
tion of RE with PC-88A. Open keys: Y(NO3 )3 , ½Y3þ f ¼ 0:05 kmolm
3 .
Solid keys: Eu(NO3 )3 , ½Eu3þ f ¼ 0:005 kmolm
3 . ½Hþ f ¼ 0:07{3
kmolm
3 .
acidity region, is written as pHa (apparent pH) for conven-
ience.
Based on the experimental data, the values of log D

3 logf½ðHRÞ2 =ðkmolm
3 Þg are plotted versus the aqueous
phase pHa values as shown in Fig. 3 according to eq. (11),
where 2½ðHRÞ2  ¼ ½HRf
6½MR3 (HR)3 . The experimen- 
tal data lie on a straight line of slope = 3 in the low-acidity
region for each ½HRf but deviate upwards from these linear
relationships with an increase in the acidity. This indicates
that solvation extraction becomes appreciable as expressed
by eq. (12) in the high-acidity region.18,19)
K2
M3þ þ 3NO3
þ nðHRÞ2
where n is solvation number.
The dependence of log D
3pHa on log{[ðHRÞ2 ]/(kmol
m
3 )} in the lower-acidity region (½Hþ f < 1 kmolm
3 ) is
shown in Fig. 4. When [Hþ ]f is 0.05 kmolm
3 for the
ð10Þ
Y(NO3 )3 system, the experimental plots lie on a straight line
of slope = 3 when log½ðHRÞ2  is less than
2 and deviate
downward with increasing logf½ðHRÞ2 = ðkmolm
3 Þg. This
negative deviation from the third-order dependence, seen
also in the other [Hþ ]f levels for the Y(NO3 )3 system and in
the Eu(NO3 )3 system, can be attributed to the nonideality of
ðHRÞ2 caused by weak solute–solute interactions,20) where
the interacting solute species is the PC-88A dimer.

Table 1 Equilibrium constants of D2EHPA and PC-88A.

Aqueous phase Km /
Extractants Diluent Kd Sources
I  /kmolm
3 Electrolyte (kmolm
3 )
1
D2EHPA 0.1 (Na,H)NO3 n-heptane 3:16  104 6:25  10
4 14)
D2EHPA 0.3 (Na,H)Cl Shellsol-K 285 5:9  10
2 15)
D2EHPA 0.1 (Na,H)ClO4 toluene 1:26  105 1:62  10
5 16)
D2EHPA 0.1 (Na,H)ClO4 hexane 3:39  104 2:95  10
5 17)
PC-88A None n-heptane 3:0  103 — 12)
PC-88A 0.4 (Na,H)NO3 Shellsol D70 2:2  104 5:6  10
3 Present work

: Ionic strength

: Vapor-phase osmometer
 Modeling of the Equilibria of Yttrium(III) and Europium(III) Solvent Extraction from Nitric Acid with PC-88A 139

1.0 where Dcalc. and Dexpt. denote the calculated and observed
[ H+ ] / kmol m-3 Slope = 3 distribution ratios of RE, respectively, and N represents the
0.0 f
number of the experimental points.
0.05
-1.0 Three hydrogen ions are discharged into the aqueous phase
0.2
-2.0 0.4
for every RE ion extracted according to eq. (9). However, the
LogD -3pHa

0.07 hydrogen ion concentration decreased as the extraction

-3.0 0.2 progressed in the higher-acidity region for both systems.
0.4
-4.0 This can be explained by the competing extraction of HNO3
by PC-88A, which is assumed to follow eq. (18). The
-5.0
extraction of HNO3 with D2EHPA and PC-88A has also been
-6.0 reported in the literature.22,23)
-7.0 KH
-3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 Hþ þ NO3
þ ðHRÞ2
HNO3 2HR  ð18Þ
Log {[(HR) ] / kmol m-3}
2
On the basis of all the above analysis, the distribution ratio
Fig. 4 LogD
3pHa vs. log½ðHRÞ2  for the extraction of RE with PC-88A can be calculated as follows:
in the lower-acidity region. Open keys: Y(NO3 )3 , ½Y3þ f ¼ 0:05 kmol
m
3 , ½HRf ¼ 0:01{1:25 kmolm
3 . Solid keys: Eu(NO3 )3 , ½Eu3þ f ¼ D ¼ ½Mt =½Mt ¼ fK1 ½ðHRÞ2 3 y3 ðHRÞ2 ½Hþ a
3
0:005 kmolm
3 , ½HRf ¼ 0:25{1 kmolm
3 .
þ K2 ½NO3
3 ½ðHRÞ2 n yn ðHRÞ2 g=ð1 þ
1 ½NO3
Þ ð19Þ

The total concentration of the extractant, that is, the feed

Also, the line connecting the experimental plots at each
[Hþ ]f level is located in slightly different positions from one
another for the same metal system; that is, the log D
3pHa
values decrease with increasing [Hþ ]f . This is expected to be
due to the complexation of RE with NO3
. The formation of
the first-order complex was assumed as defined by eq. (13)
under the present conditions.

1
M3þ þ NO3

MNO3 2þ ð13Þ charge balance to the experimental data, the most suitable
The values of the stability constants,
1 , were taken from the
literature21) as 0.631 and 1.995 (kmolm
3 )
1 for the
Y(NO3 )3 and Eu(NO3 )3 systems, respectively. The non-
ideality of ðHRÞ2 can be corrected by the effective concen-
tration of the PC-88A dimer, aðHRÞ2 , expressed as
aðHRÞ2 ¼ yðHRÞ2 ½ðHRÞ2 
where yðHRÞ2 denotes the coefficient for the effective concen-
tration of the extractant. Based on eqs. (9), (13), and (14), the
distribution ratio is expressed as
D ¼ ½Mt =½Mt ¼ K1 ½ðHRÞ2 3 y3 ðHRÞ2 ½Hþ a
3
 ð1 þ
1 ½NO3
Þ
1
By examining the experimental data at [Hþ ]f of 0.05 kmol
m
3 for the Y(NO3 )3 system shown in Fig. 4, it was found
that yðHRÞ2 is approximately linear with respect to ½ðHRÞ2  as
yðHRÞ2 ¼ 1
A½ðHRÞ2 
where A is a coefficient. By fitting eqs. (15) and (16) to the
experimental data at [Hþ ]f of 0.05 kmolm
3 for the Y(NO3 )3
system, the most likely values of K1 and A were determined
as 42.1 and 0.735 (kmolm
3 )
1 (½ðHRÞ2   0:625 kmol
m
3 ), respectively, by the nonlinear least-squares method to
minimize the standard deviations in terms of D, , expressed
by eq. (17).
( )1=2
XN
¼ N
1
½ðDcalc.
Dexpt. Þ=Dexpt.  2
ð17Þ
i¼1
concentration of the monomeric extractant can be expressed
by eq. (20).
½HRf ¼ 2½ðHRÞ2  þ 6K1 ½M3þ ½ðHRÞ2 3 y3 ðHRÞ2 ½Hþ a
3
þ 2nK2 ½M3þ ½NO3
3 ½ðHRÞ2 n yn ðHRÞ2
þ 2KH ½Hþ a ½NO3
½ðHRÞ2 yðHRÞ2 ð20Þ
By fitting these equations together with the equation of
combinations of the values of K1 , K2 , and KH were obtained
by the nonlinear least-squares method to minimize the
standard deviations for each value of n (¼ 1; 2; 3; . . .). The
value of  was smaller when n ¼ 2 than the others. Thus, the
most likely values of these constants were determined at
n ¼ 2, which are summarized in Table 2. Hirashima et al.
ð14Þ
reported that the solvation number was 2 when the
lanthanoids were extracted by D2EHPA in benzene from
highly acidic solutions.18)
Using these constants and the feed concentrations of RE,
Hþ , and HR, the distribution ratios and the pH values of the
aqueous phases were calculated by the iterative method and
ð15Þ were compared with the observed data as shown in Figs. 5
and 6 for the Y(NO3 )3 and Eu(NO3 )3 systems, respectively.
The calculated values shown by the solid curves agree with
the experimental data for the two systems. This indicates that
the proposed model can reproduce the experimental data with
ð16Þ good accuracy over a wide range of acidity (½Hþ f ¼ 0:05 to
3 kmolm
3 ).
In this modeling, the changes in the activity coefficients of
the aqueous RE ions are not fully considered, although their
complexation with the nitrate ion is considered. Moreover,
the activity coefficients of the hydrogen ion are not fully
considered, because the true pH values will deviate from the
measured pH values particularly under the high-acidity
region. One method to overcome these points is to employ
the estimation method of the activity coefficients; however,
the parameters required for the estimation are often unavail-
able, and the model tends to be very complicated. In
the present study, therefore, we have avoided such an
140 N. Fu and M. Tanaka

3.0 4.0
[ M3+ ]f / kmol m-3
-3
[ HR ]f / kmol m
3.0 0.01
2.0 0.25
0.1
0.5
1.0 2.0 0.01
1.0 0.1

1.0
Log D

Log D
0.0
0.0
-1.0
-1.0
-2.0
-2.0
-3.0
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 -3.0
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2
pH
a pH
a

Fig. 5 Effect of the equilibrium pHa on the distribution ratio of Y. The

Fig. 7 Distribution ratio of RE vs. pHa for different [M3þ ]f . ½HRf ¼
curves denote the calculated values. ½Y3þ f ¼ 0:05 kmolm
3 , ½Hþ f ¼
0:5 kmolm
3 . Curves denote the calculated values. Open keys: Y(NO3 )3 .
0:07{3 kmolm
3 .
Solid keys: Eu(NO3 )3 . ½Hþ f ¼ 0:07{3 kmolm
3 .

3.0 -0.4
[ HR ] / kmol m -3 -0.6
2.0 f
0.25
-0.8
0.5
1.0
1.0 -1.0
Log D

0.0 -1.2
Log D

[ Y3+ ]f / kmol m-3

-1.4
-1.0 0.5
-1.6 0.25
-2.0 0.5-Cal.
-1.8 0.25-Cal.

-3.0 -2.0
-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8
-4.0
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 pH
a

pH
a
Fig. 8 Distribution ratio of Y vs. pHa in the Y(NO3 )3 system for different
[Y3þ ]f . ½HRf ¼ 0:5 kmolm
3 . ½Hþ f ¼ 0:07{3 kmolm
3 . Curves de-
Fig. 6 Effect of the equilibrium pHa on the distribution ratio of Eu. The note the calculated values.
curves denote the calculated values. ½Eu3þ f ¼ 0:005 kmolm
3 , ½Hþ f ¼
0:07{3 kmolm
3 .

Table 2 Determined equilibrium constants and standard deviations.

K2 / KH /
1 /
Systems K1 n ðDÞ
(kmolm
3 )
ð3þnÞ (kmolm
3 )
2 (kmolm
3 )
1
Y(NO3 )3 (42.1) 0.36 2 0.25 (0.631) 0.178
Eu(NO3 )3 0.49 0.20 2 0.25 (1.995) 0.235

The constants in parentheses were obtained before this analysis.

approach for the engineering purpose. In this viewpoint,
the constants shown in Table 2 are specific and should be
used in caution.
3.3 Applicalility of model for different metal concen-
trations
The applicability of the proposed model was examined for
different feed concentrations of RE extending from 0.01 to
0.5 kmolm
3 as a function of the initial aqueous acidity at
½HRf ¼ 0:5 kmolm
3 . The observed distribution ratios
cluster on the respective calculated curves up to 0.1
kmolm
3 [M3þ ]f as shown in Fig. 7, but are greater than
the calculated values in the case of ½Y3þ f ¼ 0:25 to
0.5 kmolm
3 as shown in Fig. 8 for the Y(NO3 )3 system
as an example. These results indicate that the model is
applicable for the feed RE concentrations at least up to
0.1 kmolm
3 and the constants in Table 2 are valid for the
aqueous phase at the ionic strength up to 3.5 kmolm
3 . The
deviation from the observed data at high [M3þ ]f , correspond-
ing to a high loading ratio, is probably due to the change in
the activity coefficient of the RE ion and the formation of
polymeric metal-extracted complexes:7,24)
 Modeling of the Equilibria of Yttrium(III) and Europium(III) Solvent Extraction from Nitric Acid with PC-88A
3.0
[ HR ]f / kmol m-3
2.0 0.25
0.5
1.0
1.0
Log D
0.0
-1.0
-2.0
-3.0
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
pH
a
Fig. 9 Effect of the equilibrium pHa on the distribution ratio of Y in the
binary system. Solid curves were calculated by the model. ½Y3þ f ¼
0:05 kmolm
3 , ½Eu3þ f ¼ 0:005 kmolm
3 and ½Hþ f ¼ 0:07{3 kmol
m
3 .
mM3þ þ ð3m=2ÞðHRÞ2
(MR3 )m þ 3mHþ ð21Þ
where m denotes the degree of polymerization.
3.4 Applicability of the model to the binary metal
system
The applicability of the proposed model for the single-
metal systems to binary metal systems was also examined
under the condition that ½Y3þ f ¼ 0:05 kmolm
3 , ½Eu3þ f ¼
0:005 kmolm
3 , and ½HRf ¼ 0:25 to 1.0 kmolm
3 as a
function of [Hþ ]f (0.07 to 3 kmolm
3 ), assuming that the
interaction between the two metals is negligible.10) Combin-
ing the equilibrium constants of the various reactions for the
two metals and knowing the feed concentrations of all
solutes, the distribution ratios of RE as a function of the
equilibrium aqueous pHa values were calculated by the
iterative method. The experimental data are plotted together
with the calculated values in Figs. 9 and 10 for the
distribution of yttrium and europium, respectively. The
experimental data are in good agreement with the calculated
values, indicating that the model obtained for the single-
metal systems can be also used for the binary metal system
over a wide range of acidity.
4. Conclusions
For the engineering simulation, the equilibria of yttrium
and europium extraction with 2-ethylhexyl phosphonic acid
mono-2-ethylhexyl ester (PC-88A) dissolved in Shellsol D70
from nitric acid solutions were studied under nonideal
conditions. The following results were obtained:
(1) Based on the distribution data of PC-88A between the
aqueous and organic phases, the dimerization and
distribution constants of the extractant were obtained,
which suggests that PC-88A dimerizes in Shellsol D70,
and its aqueous solubility is quite low.
(2) A chemical equilibrium model was established by
analyzing the distribution data for the extraction of
141
2.0
[ HR ] / kmol m-3
1.0 f
0.25
0.5
0.0 1.0
Log D
-1.0
-2.0
-3.0
-4.0
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
pH
a
Fig. 10 Effect of the equilibrium pHa on the distribution ratio of Eu in
the binary system. Solid curves were calculated by the model.
½Y3þ f ¼ 0:05 kmolm
3 , ½Eu3þ f ¼ 0:005 kmolm
3 and ½Hþ f ¼ 0:07{
3 kmolm
3 .
RE (½Y3þ f ¼ 0:05 kmolm
3 or ½Eu3þ f ¼ 0:005 kmol
m
3 ) over wide feed concentration ranges of nitric acid
(0.05 to 3 kmolm
3 ) and the extractant (0.01 to
1.25 kmolm
3 ). The extraction equilibrium constants
were determined by fitting the model to the exper-
imental data using the nonlinear least-squares method.

The species MR3 (HR)3 is mainly formed by the cation
exchange reaction in the low-acidity region, and the
solvation extraction becomes important in the high-
acidity region. Meanwhile, the complexation of RE
with NO3
and the competing extraction of HNO3 with
the extractant were considered, and the nonideality of
ðHRÞ2 was corrected by replacing ½ðHRÞ2  with its
effective concentration.
(3) The model can reproduce the experimental data with
good accuracy for the single-metal systems for feed
concentrations of RE up to 0.1 kmolm
3 and is also
applicable to the prediction of the extraction from the
binary metal system.
Acknowledgment
The authors are grateful to the Japan Society for the
Promotion of Science for supporting this fellowship work.
REFERENCES
1) O. Navrátil and I. Dubinin: J. Inorg. Nucl. Chem. 31 (1969) 2927–2932.
2) P. W. Cassidy and A. R. Burkin: Proceedings of the International
Solvent Extraction Conference, ISEC 1971, (Hague, the Netherlands,
April 19–23, 1971) pp. 476–482.
3) T. Harada, R. G. Bautista and M. Smutz: Metall. Trans. 2 (1971) 195–
197.
4) G. W. Mason, D. N. Metta and D. F. Peppard: J. Inorg. Nucl. Chem. 38
(1976) 2077–2079.
5) Y. Mori, H. Ohya, H. Ono and W. Eguchi: J. Chem. Eng. Jpn. 21 (1988)
86–91.
6) F. Kubota, M. Goto and F. Nakashio: Solvent Extr. Ion Exch. 11 (1993)
437–453.
7) M. Iglesias, E. Anticó, V. Salvadó, A. Masana and M. Valiente: Solvent
142 N. Fu and M. Tanaka

Extr. Ion Exch. 17 (1999) 277–300.
8) T. G. Lenz and M. Smutz: J. Inorg. Nucl. Chem. 30 (1968) 621–637.
9) Y. Hirashima, M. Mugita and J. Shiokawa: J. Inorg. Nucl. Chem. 37
(1975) 1249–1253.
10) A. Hino, S. Nishihama, T. Hirai and I. Komasawa: J. Chem. Eng. Jpn.
30 (1997) 1040–1046.
11) I. M. Kolthoff: Quantitative Chem. Analysis, (Macmillan Co.: London,
1969) pp. 32–35.
12) F. Nakashio, K. Kondo, A. Murakami and Y. Akiyoshi: J. Chem. Eng.
Jpn. 15 (1982) 274–279.
13) E. Ma, X. Yan, S. Wang, H. Long and C. Yuan: Scientia Sinica XXIV
(1981) 1237–1247.
14) I. Komasawa, T. Otake and Y. Higaki: J. Inorg. Nucl. Chem. 43 (1981)
3351–3356.
15) M. Imai and S. Furusaki: Kagaku Kogaku Ronbunshu 11 (1985) 515–
519.
16) D. H. Liem: Acta Chem. Scand. 26 (1972) 191–204.
17) V. S. Ul’yanov and R. A. Sviridova: Radiokhimiya 12 (1970) 47–64.
18) Y. Hirashima, Y. Yamamoto, S. Takagi, T. Amano and J. Shiokawa:
Bull. Chem. Soc. Jpn. 51 (1978) 2890–2893.
19) T. Sato and M. Ueda: J. Inorg. Nucl. Chem. 35 (1973) 1003–1010.
20) P. R. Danesi and G. F. Vandegrift: J. Inorg. Nucl. Chem. 17 (1981)
109–115.
21) S. Kotrly and L. Sucha: Handbook Chem. Equi. Anal. Chem, (Ellis
Horwood Limited: New York, 1985) 116, 131.
22) D. F. Peppard and J. R. Ferraro: J. Inorg. Nucl. Chem. 15 (1960) 365–
370.
23) D. Li, X. Wan, D. Lin, Y. Xie, S. Lin, Z. Wang, H. Li and E. Ji:
Proceedings of the International Solvent Extraction Conference, ISEC
1980, (Liege, Belgium, Sept. 6–12, 1980), Vol. 3, pp. 1–10.
24) Y. Hirashima, H. Kodera and J. Shiokawa: Nippon Kagakukaishi
(1982) 627–631.