Journal of Water Process Engineering 7 (2015) 237–243

Contents lists available at ScienceDirect

Journal of Water Process Engineering

journal homepage: www.elsevier.com/locate/jwpe

Adsorption characteristics of gadolinium and dysprosium with

microcapsules containing an extractant
Kazuo Kondo ∗ , Masashi Umetsu, Michiaki Matsumoto
Department of Chemical Engineering and Materials Science, Doshisha University, Kyotanabe, Kyoto 610-0321, Japan

a r t i c l e i n f o a b s t r a c t

Article history: The adsorption and separation of gadolinium and dysprosium from aqueous solution were conducted
Received 15 January 2015 using microcapsules containing a metal extractant, dibenzoylmethane as an extractant. The spherical
Received in revised form 2 June 2015 microcapsules with an average diameter of 196 ␮m were prepared through encapsulation of the extrac-
Accepted 11 June 2015
tant by in situ radical polymerization of divinylbenzene. The results for adsorption equilibrium of each
Available online 20 July 2015
metal were applied to the Langmuir adsorption isotherm model to explain the experimental results. The
extent of metal adsorbed on the microcapsules increased as the pH of the metal solution increased. In
Keywords:
a binary system of gadolinium and dysprosium, only dysprosium was adsorbed on the microcapsules at
Adsorption
Separation
pH 6.0; this shows that the separation of these metals can be achieved at pH 6.0. The adsorption rate of
Microcapsule the metals was also measured, which indicated that the metal complex formation reaction and the diffu-
Rare earth metal sion of the metal complex through the pores of the microcapsules contributed to the overall adsorption
Extractant process of the metals.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction amount of reagent [2], the formation of secondary phases, and the
difficulty of back extraction.
In recent years, several metals have been in great demand for For the last several decades, alternative solvent extraction tech-
various fundamental materials. In particular, rare earth metals play niques for the recovery of metals from dilute solutions have been
very important roles as components in light emissive material, per- developed. Supported liquid membranes [3], emulsion liquid mem-
manent magnets, and catalysts. Most demands for gadolinium and branes [4], solvent impregnated resins [5], and microcapsules
dysprosium are as contrast media for examinations of MRI and opti- (hereafter, MC) containing a metal extractant [6] are typical exam-
cal magnetic disks, respectively. However, rare earth metals and ples. Among these techniques, we have investigated microcapsules
some other valuable metals can not be obtained as raw materials for the recovery and separation of rare metals and lanthanides [7,8].
in Japan. Therefore, it has become more important to recover such The metals are extracted by the microcapsules to achieve quite
valuable metals from waste materials. high loadings and they can be easily back-extracted with an acidic
The increase in the industrial demand for rare earth metals will solution.
result in an increased need for their recycling. Dilute aqueous solu- We now propose the use of microcapsules as a new alter-
tions containing such metals can be obtained from industrial waste native technological method for the recovery and separation of
treatment processes. Therefore, it is important to develop technolo- rare earth metals. The use of polymeric materials that contain a
gies for the recovery and separation of rare earth metals from such selective extractant offers some advantage over the use of sol-
dilute solutions. vent extraction techniques [9–12]. In the work reported so far,
Solvent extraction methods have generally been used for the the extractants impregnated in microcapsules are limited to acidic
recovery of metals from dilute solutions [1]. However, the separa- organophosphorus compounds or long-chain alkyl amines. The
tion of aqueous and organic phases after extraction is often difficult author previously carried out the extraction of copper with diben-
with this method. Moreover, the conventional solvent extraction zoylmethane as an extractant [13]. In this study, the adsorption
technique has some disadvantages, such as the use of a large equilibrium of gadolinium and dysprosium was measured using
the microcapsules containing dibenzoylmethane and the adsorp-
tion rate of the metals was also measured to clarify the extraction
mechanism as well as study this industrial application.
∗ Corresponding author.
E-mail address: kkondo@mail.doshisha.ac.jp (K. Kondo).

http://dx.doi.org/10.1016/j.jwpe.2015.06.006
2214-7144/© 2015 Elsevier Ltd. All rights reserved.
238 K. Kondo et al. / Journal of Water Process Engineering 7 (2015) 237–243

The DBM content in the MC was measured as follows. Micro-

Nomenclature capsules of mass 0.1 g and 50 cm3 of toluene were injected in a
multiple-dose container and shaken for 48 h at 100 rpm and 303 K.
AMC effective surface area After that, the mixture was filtered and the DBM concentration
Ceq metal concentration at equilibrium in the toluene solution was measured by ultra-violet and visible
D distribution ratio of metal spectrophotometry (SHIMADZU UV-2500PC) from which the DBM
Ea apparent activation energy content in the MC was calculated.
K extraction equilibrium constant or adsorption equi-
librium constant
2.3. Solvent extraction
k reaction rate constant
m,n reaction order
The aqueous solution was prepared by dissolving GdCl3 ·6H2 O
qeq amount of metal adsorbed at equilibrium
and/or DyCl3 ·6H2 O in a hydrochloric acid and sodium acetate mix-
qmax maximum amount of metal adsorbed
ture solution to produce a concentration of 1.0 mmol/dm3 . The pH
R0,MC initial adsorption rate of metal
of the solution was adjusted using 0.1 mol/dm3 sodium chloride
T temperature
and sodium acetate mixture solution. The organic solution was
t time
prepared by dissolving DBM in heptane of analytical grade such
w weight of microcapsule
that the DBM concentration was 30 mmol/dm3 . Aqueous solution
of 20 cm3 and organic solution of 20 cm3 were injected in an Erlen-
meyer flask and shaken for 48 h at 100 rpm at 303 K. After that,
2. Experimental the metal ion concentration in the aqueous phase was determined
by inductively coupled plasma spectrometry (Shimadzu ICPS-8000,
2.1. Reagents ICP-OES).

Dibenzoylmethane (DBM) was purchased from Wako Pure

2.4. Metal adsorption equilibrium
Chemical Industry Co. Ltd. in Japan and used without further
purification. For microencapsulation, divinylbenzene (DVB) was
2.4.1. Effect of pH
employed as a monomer to form the shell of the microcapsules.
The adsorption behavior of Gd and Dy onto the MC containing
Toluene was used as the diluent for the extractant. 2,2-Azobis(2,4-
DBM was examined batchwise. In a single metal system, the metal
dimethylvaleronitrile) (ADVN) was used as the polymerization
ion concentration was adjusted to 1.0 mmol/dm3 for each metal
initiator. ADVN is a special grade chemical and was used without
ion. In a binary metal system, the metal ion concentration was
further purification.
adjusted to 0.5 mmol/dm3 for each metal ion. Thus, the total metal
Aqueous solutions of the metals were prepared by dissolving
concentration was 1.0 mmol/dm3 . The pH of the aqueous solution
GdCl3 ·6H2 O and/or DyCl3 ·6H2 O in a hydrochloric acid and sodium
was adjusted in a similar way to that described above. 10 mL of a
acetate mixture solution to produce a desired concentration. These
metal solution and 0.1 g of MC were shaken together in a test tube
reagents were analytical grade.
at 303 K over 60 h at 100 rpm to attain adsorption equilibrium.
The metal concentrations in the aqueous solutions before and
2.2. Preparation of the microcapsules after adsorption were measured using ICP-OES. The metal adsorbed
onto the MC was calculated from the mass balance difference before
The reactor used for preparation of the microcapsules was a and after equilibrium.
mixing vessel with four baffle plates and a capacity of 1 dm3 . The
continuous phase was 900 g of distilled water containing 2% w/w
2.4.2. Effect of metal concentration
Arabic gum to act as a dispersion stabilizer. The continuous phase
In this experiment, the metal concentration was adjusted to
was stirred in the vessel using a flat paddle. After the temperature of
0.5–5.0 mmol/dm3 for each metal ion. The pH of the aqueous solu-
the continuous phase reached 343 K, the dispersed phase consisting
tion was adjusted to 7.0 at equilibrium using a 0.1 mol/dm3 sodium
of DVB, ADVN, toluene, and the extractant was added to the con-
chloride – sodium acetate mixture solution. The other experimental
tinuous phase. The reactor was purged with nitrogen gas and the
procedure was the same as that mentioned above.
contents were agitated at 300 rpm for 3 h at 343 K under a nitrogen
atmosphere to produce the MC through the in situ radical poly-
merization of DVB. The conditions used for the preparation of MC 2.5. Metal adsorption rate
containing DBM are shown in Table 1. The obtained MC were col-
lected by suction filtration, washed with distilled water, and then 2.5.1. Effect of temperature
dried at room temperature. The MC was then screened using sieves The adsorption rates of Gd and Dy onto the MC were measured
with mesh sizes of 150–210 ␮m. The sizes of the resultant particles varying the solution temperature (303–323 K). The metal solution
were then measured with a laser diffraction particle size analyzer was prepared in a similar way to that described in the metal equi-
(SALD-2000A, Shimadzu Co. Ltd., Japan). librium experiment.
The pH of the aqueous solution was adjusted to 7.0 using
0.1 mol/dm3 sodium chloride – sodium acetate mixture solution.
Table 1 Ten milliliters of the aqueous solution and 0.1 g of MC were shaken
Conditions for the preparation of MC.
together in a test tube at 150 rpm. After a periodical adsorption
Materials Weight or volume time, the aqueous solution was filtered and the metal concentra-
Continuous Distilled water 882 g tion was determined. The same experiment as above was carried
phase Arabic gum 18.0 g out under the condition of infinite agitation speed.
Dispersed DVB 15.0 g
phase ADVN 3.6 g 2.5.2. Effect of metal concentration
DBM 12.0 g In this experiment, the metal adsorption rate was measured
Toluene 50.0 cm3
while varying the initial metal concentration (1.0–5.0 mmol/dm3 ).
 K. Kondo et al. / Journal of Water Process Engineering 7 (2015) 237–243 239

Fig. 1. SEM image of the microcapsule (×350).

The pH of the solution was adjusted to 7.0. The other experimental

procedure was the same as that mentioned above.
Fig. 3. Relationship between log D and equilibrium pH (Gd: logD = 3.00pH-18.8,
R2 = 0.975 ;Dy: logD = 3.05pH-18.6, R2 = 0.993).
3. Results and discussion

3.1. Preparation of microcapsules Now, the distribution ratio D is defined as follows.

[MRn ]
The microcapsules obtained were yellowish spherical particles D=  
Mn+
and their average diameter was 196 ␮m. The amount of DBM con-
tained in the microcapsules was 1.76 mmol/g-MC. Any leak of DBM By using Eq. (3), Eq. (2) can be rearranged as follows.
from the MC was not seen in a pH range of 2–8. Therefore, it is
anticipated that the adsorbed metals do not leak out of the MC. The logD = npH + logK + nlog[HR] (4)
scanning electron microscopy image of the microcapsule is shown The experimental data were arranged in Fig. 3 according to Eq.
in Fig. 1. (4). Two straight lines with slope of 3 were obtained, from which
it was found that K is 1.44 × 10−19 for gadolinium extraction, and
3.2. Solvent extraction 2.43 × 10−19 for dysprosium extraction, respectively.
From the result, the metal extraction by DBM is expressed as
Fig. 2 shows the structural formula of DBM. It can be considered follows.
that the extraction of metal ion by DBM is caused by a chelating
M3+ + 3HR  MR 3 + 3H+ (5)
complex formation reaction of the metal with enol form of DBM.
This can be expressed as follows [13].
3.3. Metal adsorption equilibrium
Mn+ + nHR  MR n + nH+ (1)
3.3.1. Effect of pH on the metal adsorption equilibrium
where M is Gd or Dy, and HR is DBM. Fig. 4 shows the relationship between the extent of metal
In Eq. (1), the extraction equilibrium constant K is expressed as adsorbed and the equilibrium pH in a single metal system, and
follows. Fig. 5 shows the same relationship in a binary metal system. From
 n Figs. 4 and 5, the extent of metal adsorbed increased as the equilib-
[MRn ] H+ rium pH increased. This is because the metal adsorption is caused
K= (2) by exchanging hydrogen ion and metal ion which is expressed in
[M n+ ] [HR]n
Eq. (5). At a higher pH of the aqueous solution, in other words, a
low hydrogen ion concentration in the solution, DBM contained
in microcapsules tends to release a lot of hydrogen ions. As the
result, DBM molecules were easily ionized to extract more metal
ions. From Fig. 5, it was found that only dysprosium was adsorbed
at pH 6.0. So, the two metal ions can be separated at this pH in a
batch system.
As the results of a repeated experiment, the MC was stable in at
least three times experiments.

3.3.2. Effect of metal concentration on the metal adsorption

equilibrium
The adsorption equilibrium for the metals was also examined
using the MC containing DBM. Fig. 6 shows the time course of
Fig. 2. Structural formula of DBM. the extent of metal adsorbed for Gd and Dy. It was found that the
240 K. Kondo et al. / Journal of Water Process Engineering 7 (2015) 237–243

Fig. 6. Relationship between extent of metal adsorbed on MC and time.

Fig. 4. Relationship between extent of metal adsorbed and equilibrium pH (single

metal system).

adsorption equilibrium was attained over 60 h. This result is rea-

sonable because the reaction rate between the rare earth metals
and DBM is slow. Fig. 7 shows the adsorption isotherms of Gd and
Dy based on the Langmuir isotherm model. This model assumes
that the adsorption occurs at specific homogeneous adsorption
sites within the adsorbent and so intermolecular forces decreases
rapidly with the distance from the adsorption surface. The widely
used Langmuir isotherm has found successful application in many
real adsorption processes. This equation is described as follows.

qmax K and C eq
qeq = (6)
1 + K and C eq

Fig. 7. Relationship between qeq and Ceq .

Table 2
Langmuir parameters determined for the metal.

K [dm3 /mmol] qmax [mmol/g]

Gd 1.06 0.441
Dy 11.40 0.436

Eq. (6) can be rearranged as follows.

1 1 1 1
= + (7)
qeq qmax K Ceq qmax

where K [dm3 /mmol] and qmax [mmol/g] are the apparent adsorp-
tion equilibrium constant and the maximum amount of metal
adsorbed, respectively, and Ceq [mmol/dm3 ] is the metal concen-
tration at equilibrium.
A linearized plot of 1/qeq versus 1/Ceq is shown in Fig. 8 for Gd
and Dy. The parameters qmax and K were calculated from the slopes
and the intercepts of each straight line. The values of the Langmuir
parameters are listed in Table 2.
As it can be seen from Table 2 and Fig. 8, the experimental
Fig. 5. Relationship between extent of metal adsorbed and equilibrium pH (binary data were well correlated by the Langmuir adsorption equilibrium
metal system). model. Moreover, the Langmuir constant K of Dy adsorption is
 K. Kondo et al. / Journal of Water Process Engineering 7 (2015) 237–243 241

Fig. 8. Langmuir plots for Gd (1/qeq = 2.14/Ceq + 2.27, R2 = 0.983) and Dy

(1/qeq = 0.201/Ceq + 2.29, R2 = 0.990). Fig. 10. Adsorption rate of Dy.

larger than that of Gd; this tendency was seen also in the solvent The following relation holds between the initial metal adsorp-
extraction. These results show that Dy has higher affinity to DBM tion rate, R0,MC , and temperature T, using Arrhenius equation:
than Gd. dq (t) Ea 1
InR0 ∝ In |t = 0 ∝ − (9)
dt R T
3.4. Metal adsorption rate
where Ea [J/(K mol)] is the apparent activate energy and R is the gas
3.4.1. Effect of temperature on the metal adsorption rate constant.
The relationship between the amount of metal adsorbed and From Figs. 9 and 10, ln (dq/dt) was calculated and the relation-
adsorption time is shown in Figs. 9 and 10, for Gd and Dy, respec- ship between ln (dq/dt) and (1/T) is shown in Fig. 11. From the figure,
tively. The initial rate of metal adsorption onto the MC, R0,MC , was the apparent activation energy, Ea, was calculated and the values
calculated by Eq. (8) using the experimental results [14,15]: are listed in Table 3.
Metal ion adsorption onto the MC is considered to consist of
w dq (t) three steps as follows.
R0 = |t = 0 (8)
AMC dt
1. Diffusion of metal ion to the MC through the boundary film sur-
where AMC [m2 ] is the effective surface area of the MC, w [g] is
rounding the MC.
the weight of MC, and q(t) is the amount of metal adsorbed onto
2. Complex formation reaction between DBM and metal ion.
the MC. Here AMC was calculated using the average diameter of a
3. Diffusion of the metal complex through the fine pores of the MC.
microcapsule (196 ␮m).

Fig. 11. Apparent Arrehenius plot for Gd (ln(dq/dt) = −1.05/T × 104 + 27.8,
Fig. 9. Adsorption rate of Gd. R2 = 0.937) and Dy (ln(dq/dt) = −5.00/T × 103 + 10.8, R2 = 0.967).
242 K. Kondo et al. / Journal of Water Process Engineering 7 (2015) 237–243

Table 3
Apparent activation energy and reaction order for Gd and Dy.

Agitation speed [rpm] Ea [kJ/mol] m

Gd 150 87.3 0.90

∞ 88.6 0.97

Dy 150 41.6 0.82

∞ 68.2 1.02

Generally, it is known that the complex formation reaction is

the rate-determining step in the overall adsorption process if Ea
is higher than 40 kJ/mol, and the diffusion of metal ion is the rate-
determining step if Ea is lower than 20 kJ/mol [12,16]. From Table 3,
it is found that Ea is higher than 40 kJ/mol. So, it is considered that
the complex formation reaction would be the rate-determining
step in the overall adsorption process of the metal ion.
In order to confirm this idea, the same experiment was carried
out under the condition where the agitation speed is infinite. Prac-
tically speaking, the test tubes were shaken so intensely by hand
that the diffusion resistance of metal ion to the MC through the
boundary film surrounding the MC could be neglected. The result
is also shown in Table 3, from which it is found that the complex for-
Fig. 13. Adsorption rate of Dy.
mation reaction is completely the rate-determining step at infinite
agitation speed.

3.4.2. Effect of the temperature on the metal adsorption rate

The relationship between the amount of metal adsorbed and
adsorption time is shown in Figs. 12 and 13, for Gd and Dy, respec-
tively.
The initial adsorption rate of the metals is assumed to be as
follows.
 m
R0 = k M3+ [DBM]n0 (10)
0

where k is the reaction rate constant, [M3+ ]0 is the initial metal

ion concentration, and [DBM]0 is the initial DBM concentration, and
m and n are the reaction order for metal ion and DBM, respectively.
Eq. (10) can be rewritten as follows.
 
logR0 = mlog M3+ + nlog[DMB]0 + logk (11)
0

dq (t)   AMC
log | = mlog M3+ + nlog[DBM]0 + logk + log (12)
dt t=0 0 w
From Figs. 12 and 13, log (dq/dt) was calculated and the rela-
tionship between log (dq/dt) and log [M3+ ]0 is shown in Fig. 14
Fig. 14. Relationship between log(dq/dt) and log[M3+ ] for Gd
(log(dq/dt) = 0.900log[Gd3+ ]-2.68, R2 = 0.944) and Dy (log(dq/dt) = 0.816
3+ 2
log[Dy ]−2.62, R = 0.954).

for Gd and Dy from which the value of m was calculated and also
shown in Table 3. Generally, it is known that the complex forma-
tion reaction is the rate-determining step in the overall adsorption
process if m is 1.0. Table 3 shows that m is a little lower than 1.0
at 150 rpm. This is because that the effect of diffusion of the metal
ion to the MC through the boundary film surrounding the MC can
not be absolutely neglected. Concluding this, at 150 rpm the overall
metal adsorption process is considered to be affected by both the
complex formation reaction and the diffusion of the metal com-
plex through the pores in the MC. However, the complex formation
reaction controlled the overall metal adsorption process at infinite
agitation speed.

4. Conclusions

In this study, the microcapsules containing DBM as a metal

extractant were successfully prepared. The microcapsules obtained
Fig. 12. Adsorption rate of Gd. were yellowish spherical particles and their average diameter was
 K. Kondo et al. / Journal of Water Process Engineering 7 (2015) 237–243 243

196 ␮m. The amount of DMB contained in the MC was 1.76 mmol/g- [5] J.L. Cortina, A. Warshawsky, Developments in Solid–Liquid Extraction by
MC. Solvent-Impregnated Resins, in Ion Exchange and Solvent Extraction, in: J.A.
Marinsky, Y. Marcus (Eds.), Marcel Dekker Publishers, New York, 1997.
Batch extraction of Gd and Dy with the MC containing DBM was [6] H. Yoshizawa, K. Fujikubo, Y. Uemura, Y. Kawano, K. Kondo, Y. Hatate,
dependent on the hydrogen ion concentration and could be applied Preparation of divinylbenzene homopolymeric microcapsules with highly
to the Langmuir adsorption isotherm model, from which the Lang- porous membranes by in situ polymerization with solvent evaporation, J.
Chem. Eng. Japan 28 (1995) 78–84.
muir constants, K and qmax , were determined. The results confirm [7] E. Kamio, K. Kondo, Separation of rare metal ions by a column packed with
that the adsorption equilibrium of each metal is well explained by microcapsules containing an extractant, Ind. Eng. Chem. Res. 41 (2002)
the Langmuir adsorption isotherm model. In a binary system of 3669–3675.
[8] K. Kondo, M. Ishihara, M. Matsumoto, E. Kamio, Extraction of precious and
Gd and Dy, only Dy was adsorbed on microcapsules at pH 6.0, this
base metals with microcapsules containing an extractant, J. Chem. Eng. Japan
showing that these metals can be separated at this pH. 40 (2007) 31–35.
The adsorption rates of Gd and Dy onto the MC were also mea- [9] H. Mimura, Development of advanced nuclide separation and recovery
methods using ion-exchange techniques in nuclear backend, J. Nucl. Sci.
sured. The effects of the metal concentration and temperature on
Technol. 38 (2001) 342–348.
the metal adsorption rates were examined. From these results, both [10] E. Kamio, K. Kondo, Separation and concentration of lanthanides using
the complex formation reaction and the diffusion of the metal com- microcapsules containing acidic organophosphorus compounds as an
plex through the pores in the MC were found to contribute to the extractant, J. Chem. Eng. Japan 35 (2002) 574–581.
[11] S. Kiyoyama, S. Yonezawa, Y. Yoshida, Extraction rate of palladium using
overall metal adsorption process at 150 rpm. However, the complex divinylbenzene microcapsules containing tri-n-octylamine as the extractant,
formation reaction controlled the overall metal adsorption process React. Funct. Polym. 67 (2007) 522–528.
when the agitation speed is infinite. [12] K. Kondo, M. Sawada, M. Matsumoto, Adsorption and separation of palladium
and platinum with microcapsules containing tri-n-octylamine hydrochloride,
J. Water Proc. Eng. 1 (2014) 115–120.
References [13] K. Kondo, T. Tsuneyuki, K. Hiramatsu, F. Nakashio, Kinetic studies of
extraction and stripping of copper with dibenzoylmethane, Sep. Sci. Technol.
[1] H. Narita, M. Tanaka, K. Morisaku, T. Abe, Extraction of gold (III) in 25 (1990) 1213–1224.
hydrochloric acid solution using monoamide compounds, Hydrometallurgy [14] E. Kamio, M. Matsumoto, K. Kondo, Sorption behavior of Ga(III) and In(III) into
81 (2006) 153–158. a microcapsule containing long-chain alkylphosphoric acid monoester, Ind.
[2] L. Ghezzi, B.H. Robinson, F. Seco, M.R. Tine, M. Venturini, Binding of Pd(II) to Eng. Chem. Res. 44 (2005) 2265–2272.
pada in water/micellar system: complex formation kinetics in water and [15] E. Kamio, Y. Fujiwara, M. Matsumoto, K. Kondo, F. Valenzuela, Investigation
DTAC solution, Colloids Surf. A 292 (2007) 139–147. on extraction rate of lanthanides with extractant-impregnated microcapsules,
[3] S.S. Fu, M. Teramoto, H. Matsuyama, Uphill transport of Ce(III) by supported Chem. Eng. J. 139 (2008) 93–105.
liquid membranes containing [16] K. Kondo, S. Hirao, M. Matsumoto, Column separation of nickel and cobalt
octyl(phenyl)-N,N-diisobutylcarbamoylmethylphosphine oxide in using a microencapsulated extractant, Hydrometallurgy 149 (2014) 87–96.
2-nitrophenyl octyl ether, Sep Sci. Technol. 39 (2004) 517–538.
[4] K. Kondo, M. Matsumoto, Separation and concentration of indium (III) by an
emulsion liquid membrane containing diisostearylphosphoric acid as a
mobile carrier, Sep. Purif. Technol. 13 (1998) 109–115.