Synergistic extraction and solution structures of ternary complexes of lanthanoids with 2-thenoyltrifluoroacetone and linear poly(oxyethylene) in 1,2-dichloroethane

Article abstract of DOI:10.1246/bcsj.80.2357

Synergistic extraction of thenoyltrifluoroacetone (tta) complexes of 14 trivalent lanthanoids (Ln³⁺) into 1,2-dichloroethane with linear poly(oxyethylene) compounds (POE) was investigated at 25.0°C, where the linear POE [HO-(CH₂CH₂O-)_nR] are monodispersed DEOn (R = C₁₂H₂₅; n = 4, 6, and 8) and polydispersed TX-100 (R = octyl-phenyl; n_ave = 9.6). Equilibrium studies showed a 1:1 adduct complex with POE compound, Ln(tta)₃ · (POE), formed in the organic phase. The adduct formation constant, β_add, increased as the ionic radius of lanthanoid increased for any kind of POE. The very high stability of the 18-crown-6 adducts is explained by the incorporation of Ln³⁺ ion into the cavity of the crown ether. Adduct formation constants of the POE having short ethylene oxide (EO) chain, such as DE04 and 12-crown-4, were fairly small; thus, it was estimated that those coordinate to the metal ion as a bidentate or tridentate ligand. Relatively large values of β_add of long chain POE (n ≥ 6) indicate the indirect outer-sphere interaction of the uncoordinated residual EO units with the metal ion. Adduct formation with linear POEs significantly reduced the difference in extraction constants among the lanthanoid ions. Therefore, this synergistic extraction system is advantageous for the separation of lanthanoid ions as a group from other metal ions.

Full text of DOI:10.1246/bcsj.80.2357

Ó 2007 The Chemical Society of Japan
Bull. Chem. Soc. Jpn. Vol. 80, No. 12, 2357–2364 (2007) 2357
Synergistic Extraction and Solution Structures of Ternary Complexes
of Lanthanoids with 2-Thenoyltrifluoroacetone and Linear
Poly(oxyethylene) in 1,2-Dichloroethane
ꢀ1
Gene Frederick Gagabe,1 Yuta Satoh,1 Keiichi Satoh,2 and Kiyoshi Sawada
1Graduate School of Science and Technology, Niigata University, Ikarashi 2-Nocho, Niigata 950-2181
2Department of Chemistry, Faculty of Science, Niigata University, Ikarashi 2-Nocho, Niigata 950-2181
Received June 12, 2007; E-mail: sawada@sc.niigata-u.ac.jp
Synergistic extraction of thenoyltrifluoroacetone (tta) complexes of 14 trivalent lanthanoids (Ln3þ) into 1,2-
dichloroethane with linear poly(oxyethylene) compounds (POE) was investigated at 25.0 ꢁC, where the linear POE
[HO–(CH2CH2O–)nR] are monodispersed DEOn (R ¼ C12H25; n ¼ 4, 6, and 8) and polydispersed TX-100 (R = octyl-
phenyl; nave ¼ 9:6). Equilibrium studies showed a 1:1 adduct complex with POE compound, Ln(tta)3 (POE), formed in
ꢂ
the organic phase. The adduct formation constant, ꢀadd, increased as the ionic radius of lanthanoid increased for any kind
of POE. The very high stability of the 18-crown-6 adducts is explained by the incorporation of Ln3þ ion into the cavity
of the crown ether. Adduct formation constants of the POE having short ethylene oxide (EO) chain, such as DEO4 and
12-crown-4, were fairly small; thus, it was estimated that those coordinate to the metal ion as a bidentate or tridentate
ligand. Relatively large values of ꢀadd of long chain POE (n ꢃ 6) indicate the indirect outer-sphere interaction of the
uncoordinated residual EO units with the metal ion. Adduct formation with linear POEs significantly reduced the differ-
ence in extraction constants among the lanthanoid ions. Therefore, this synergistic extraction system is advantageous
for the separation of lanthanoid ions as a group from other metal ions.
There has been a growing interest in lanthanoid (Ln) coor-
dination chemistry, because of the increasing industrial, chem-
ical, medical, and sensor applications of rare-earth ions and
their compounds.1,2 ꢀ-Diketones, represented by thenoyltri-
fluoroacetone (Htta), have been widely used as an effective ex-
tractant for the separation and determination of rare-earth ele-
ments.3–7 Addition of an electrically neutral lipophilic ligand
or Lewis base enhances the extractability of the metal ions and
the mutual separation of the lanthanoid ions. The synergistic
effects on Htta extraction of Ln3þ ions have been extensively
studied for monodentate and polydentate ligands, such as
phosphoric esters, phosphine oxides, 1,10-phenanthroline, and
polyethers.8–11 This effect is explained by the formation of a
hydrophobic adduct through the displacement of residual
water molecules coordinated to the ꢀ-diketonato–lanthanoid
complex with the neutral ligand.8
Poly(oxyethylene) derivatives (POE compounds), which
consist of repeating units of ethylene oxide (EO), form cationic
complexes with alkali, alkaline-earth and rare-earth metal
ions.12–22 Because of this peculiar nature, POEs have been
used for the separation, concentration, and determination of
metal ions.23–25 Particularly, cyclic POE compounds (crown
ethers) incorporate metal ions within their cavity. In other
words, they have very high selectivity for metal ions by the
size-fitting effect.26–29
reported the synergistic extraction of a series of lanthanoid
ions with Htta and various crown ethers and have estimated
the structure of the adduct complexes formed in the organic
phase.34 Non-cyclic poly(oxyethylene)s (linear POEs) react
with metal ions in the same manner as crown ethers. Linear
POEs have been widely used for analytical and industrial
purpose, because they are nontoxic and inexpensive compared
to crown ethers.35 Although the selectivity of metal ion com-
plexation of linear POEs is inferior to the crown ethers, linear
POEs have flexible structures, and long-chain derivatives
are also available.35–39 These properties are advantageous for
efficient simultaneous extraction of metal ions as a group.
In the present paper, the synergistic extraction of 14 lantha-
noid complexes of Htta with linear POEs having various eth-
ylene oxide chain lengths is reported. The formation constants
and structures of the adduct complexes of the linear POEs in
organic phase (1,2-dichloroethane) are compared to those of
the crown ether adducts.
Experimental
Reagents. Stock solutions of lanthanoid nitrates were pre-
pared by dissolving appropriate amounts of their metal oxides,
Ln2O3 (Ln ¼ La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm,
Yb, Lu; Shinetsu), in 6 M HNO3, followed by evaporation of the
solvent and dissolution of the residue in distilled water. Thenoyl-
trifluoroacetone (Htta, Dojindo) was used as received. Monodis-
persed monododecyl ethers (DEOn; HO–(CH2CH2O–)nR, where
R ¼ C12H25 and number of EO units n ¼ 4, 6, and 8) and polydis-
persed octylphenyl ether (TX-100; R = octylphenyl and average
number of EO units39 ðnaveÞ ¼ 9:6) were obtained from Nikko
Synergistic extraction of the lanthanoid ions with crown
ethers has been studied for various extraction systems and or-
ganic solvents;30–33 however, systematic data have rarely been
reported on the adduct formation of Htta complexes of Ln3þ
ions with crown ethers. Thus, in our previous paper, we have
Published on the web December 13, 2007; doi:10.1246/bcsj.80.2357

2358 Bull. Chem. Soc. Jpn. Vol. 80, No. 12 (2007)
Lanthanoid Complexes with Polyether
Chemicals and Wako Pure Chemicals, respectively, and were used
without further purification. 1,2-Dichloroethane (Nacalai Tesque,
AR grade) was washed twice with distilled water prior to use.
All other reagents used were of analytical reagent grade.
adduct complex with linear POE in the organic phase. By
assuming the formation of [Ln(tta)3 m(POE)] in the organic
phase, the extraction equilibrium is given by
ꢂ
Ln3þaq þ 3Httaorg þ mPOEorg
Procedure. The number of thenoyltrifluoroacetonate (Htta)
involved in the extracted adduct complexes was determined by
the following procedure. The pH of an aqueous solution contain-
ing 14 lanthanoid ions was adjusted to the desired value (3.5 to
5.5) with 0.01 M acetic acid–potassium acetate buffer. It has been
shown that the acetate ion has no effect on the extraction of the
lanthanoid ions. The ionic strength was maintained at 0.1 M with
potassium nitrate. The initial concentration of each lanthanoid was
5:0 ꢄ 10ꢅ7 M (M = mol dmꢅ3). The aqueous solution (20 cm3)
thus prepared was mixed with an equivolume of an 1,2-dichloro-
ethane solution containing 1:0 ꢄ 10ꢅ2 M Htta and 1:0 ꢄ 10ꢅ2 M
POE in a centrifuge tube (50 cm3). The mixture was shaken for
30 min at 25:0 ꢆ 0:1 ꢁC in a thermostated mechanical shaker. This
shaking time was sufficient to attain equilibrium. To determine the
chemical form of the adduct and its formation constant, a portion
of the aqueous phase (20 cm3, pH 4.9) was shaken with an
equivolume of organic phase containing 1:0 ꢄ 10ꢅ2 M Htta and
various concentrations of POE (3:0 ꢄ 10ꢅ7–1:0 ꢄ 10ꢅ2 M). After
phase separation through centrifugation, the pH of the aqueous
phase was measured by using a pH meter (DKK PHL-40). The
concentrations of lanthanoid ions remaining in the aqueous phases
were determined with an inductively coupled plasma-mass
spectrometer (HP4500 ICP-MS, Yokogawa Agilent). The Ln3þ
concentrations in the organic phase were determined in the same
manner as that in the aqueous phase after back-extraction of Ln3þ
into 0.1 M HNO3.
The distribution coefficient, Kd, of 12-crown-4 (12C4, Aldrich)
between 1,2-dichloroethane and water was determined by the
following procedure. A 1,2-dichloroethane (15 cm3) solution con-
taining an appropriate amount of 12C4 was mixed with an equi-
volume of aqueous phase in a centrifuge tube. The mixture was
shaken for 30 min. After phase separation, the solvents were
evaporated under low pressure. The amount of 12C4 in each phase
was then determined by weighing. The distribution coefficient of
12C4 (Kd = [12C4]org/[12C4]aq) was determined to be 0.83. It
has been confirmed that the potassium ion is not extracted by
the nitrate ion under the present experimental conditions. MM2
calculations of the molecular structures were performed by using
SPARTAN software.40
Kꢀex,add [Ln(tta)3 m(POE)]org þ 3Hþ
The extraction constant of adduct complex is
ð4Þ
ð5Þ
aq
ꢂ
[Ln(tta)3 m(POE)]org[Hþ]3
ꢂ
aq
Kex,add
¼
:
[Ln3þ]aq[Htta]3org[POE]omrg
As the total concentration of each Ln3þ ion in the organic
phase is the sum of the concentrations of [Ln(tta)3] and
[Ln(tta)3 m(POE)], the distribution ratio of Ln3þ in the pres-
ence of POE is given as
ꢂ
[Ln(tta)3]org þ [Ln(tta)3 m(POE)]org
ꢂ
According to extraction equilibria (2) and (4), the formation of
D ¼
:
ð6Þ
[Ln3þ
]
aq
the adduct [Ln(tta)3 m(POE)] in the organic phase is given by
ꢂ
ꢀadd
[Ln(tta)3]org þ mPOEorg ꢀ [Ln(tta)3 m(POE)]org
ð7Þ
ð8Þ
ꢂ
and the adduct formation constant, ꢀadd, is defined as
[Ln(tta)3 m(POE)]org
ꢂ
ꢀadd
¼
:
[Ln(tta)3]org[POE]omrg
Consequently, the distribution ratio in Eq. 6 becomes
[Ln(tta)3]org þ [Ln(tta)3]orgꢀadd[POE]omrg
D ¼
ð9Þ
[Ln3þ
]
aq
[Ln(tta)3]org(1 þ ꢀadd[POE]omrg
)
¼
:
ð10Þ
[Ln3þ
]
aq
Substitution of Eq. 3 into Eq. 10 results in the following:
Kex[Htta]3org(1 þ ꢀadd[POE]morg
)
D ¼
:
ð11Þ
[Hþ]3aq
Taking the logarithm of Eq. 11 yields
log D ¼ log Kex þ 3 log [Htta]
þ 3 pH þ logð1 þ ꢀadd[POE]omrg):
ð12Þ
Results
To make sure that number of Htta molecules involved in the
adduct is three, the distribution ratio, D, was determined as a
function of pH with the concentrations of Htta and POE kept
constant. The plots of log D vs. pH thus obtained are shown
in Figs. 1 and 2, where the results of extraction with DEO6
and TX-100 are depicted as examples. The plots were straight
lines with a slope of three for all Ln3þ, and the same results
were obtained for the other POEs. Thus, it is confirmed that
the adduct complex also involves three tta molecules, that is,
The distribution ratio (D) of a lanthanoid ion is defined as
CLn,org
D ¼
;
ð1Þ
CLn,aq
where CLn,org and CLn,aq are the total concentrations of the Ln
ions in the organic and aqueous phases, respectively. The ex-
traction equilibrium and extraction constant, Kex, of trivalent
lanthanoid ions (Ln3þ) with Htta have been determined30,34 as
Kex
the adduct [Ln(tta)3 m(POE)].
Equation 12 can be simplified as
Ln3þaq þ 3Httaorg ꢀ [Ln(tta)3]org þ 3Hþ
;
ð2Þ
ꢂ
aq
[Ln(tta)3]org[Hþ]a3q
[Ln3þ]aq[Htta]o3rg
log D ¼ log D0 þ logð1 þ ꢀadd[POE]morgÞ;
ð13Þ
Kex
¼
:
ð3Þ
where log D0 ¼ log Kex þ 3 log [Htta]org þ 3 pH, and D0 is the
distribution ratio of Ln3þ in the absence of POE. The value of
D0 is constant when the pH and concentration of Htta are fixed.
Thus, the number of POE in the extracted adduct, m, can be
The extraction of Ln3þ with Htta is significantly enhanced by
the addition of linear POE in the same manner as the crown
ether system.34 This result indicates the formation of the

G. F. Gagabe et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 12 (2007) 2359
3
2
3
2
(a)
(b)
1
1
0
0
-1
-2
-3
-1
-2
-3
3
4
5
6
3
4
5
6
pH
pH
Fig. 1. Plots of log D vs. pH in the extraction of Ln3þ with 1:0 ꢄ 10ꢅ2 M Htta and 1:0 ꢄ 10ꢅ2 M DEO6. (a) Odd-atomic number
Ln: = La, = Pr, = Eu, = Tb, = Ho, ꢄ = Tm, = Lu; (b) Even-atomic number Ln: = Ce, = Nd, = Sm,
= Gd, = Dy, ꢄ = Er,
= Yb. The slopes of solid lines are 3.
3
3
(a)
(b)
2
1
2
1
D
D
0
0
-1
-2
-3
-1
-2
-3
3
4
5
6
3
4
5
6
pH
pH
Fig. 2. Plots of log D vs. pH in the extraction of Ln3þ with 1:0 ꢄ 10ꢅ2 M Htta and 1:0 ꢄ 10ꢅ2 M TX-100. (a) Odd-atomic number
Ln, and (b) Even-atomic number Ln. Symbols are the same as those in Fig. 1. The slopes of solid lines are 3.
determined from the slope of the plot of log D vs. log [POE]org
.
regression to the plots of Fig. 3. The solid lines in Fig. 3 depict
the calculated regression curves, and these are in good agree-
ment with the experimental results. The logarithmic values of
adduct formation constant, log ꢀadd, thus obtained are listed in
Table 1. The values of log ꢀadd for some crown ether adducts34
are also listed in this table for comparison. The values of the
adduct formation constants of the crown ethers listed in this
table were corrected for the distribution of the crown ether into
the aqueous phase by using Kd values of 1.05 and 0.83 (pres-
ent work) for 18-crown-6 (18C6)41 and 12-crown-4 (12C4),
respectively.
The distribution coefficient of DEOn and TX-100 in 1,2-di-
chloroethane is larger than 103,37,38 and the concentration of
POE is much higher than that of metal complexes. Thus, the
concentration of POE in the organic phase, [POE]org, can be
assumed to be equal to its total concentration, CPOE. Figure 3
shows the plots of log D as a function of log [POE]org at pH 4.9
and 1:0 ꢄ 10ꢅ2 M Htta. The slopes of the plots approached
unity in higher [POE]org region irrespective of the kind of
POE. The other lanthanoids, not depicted in Fig. 3, also
showed the same results. These results indicate that the
[Ln(tta)3] complex forms an adduct with one molecule of
POE, that is, m ¼ 1. Thus, the chemical form of the adduct
formed in the synergistic extraction with the linear POEs
Discussion
Trend in Adduct Formation Constants Across the Lan-
thanoid Series. Lanthanoid complexes of Htta are extracted
was determined to be [Ln(tta)3 (POE)].
The values of the adduct formation constant, ꢀadd, were cal-
culated based on Eq. 13 by applying nonlinear least-square
ꢂ
into the organic phase as the hydrated complex [Ln(tta)3
nH2O], where n ¼ 2 or 3.42 In synergistic systems, the extract-
ꢂ

2360 Bull. Chem. Soc. Jpn. Vol. 80, No. 12 (2007)
Lanthanoid Complexes with Polyether
4
4
3
(b)
(a)
3
2
2
1
1
0
0
-1
-2
-1
-2
-7
-6
-5
-4
-3
-2
-7
-6
-5
-4
-3
-2
log [DEO4]
log [DEO6]
org
org
4
3
4
3
(c)
(d)
2
2
1
1
0
0
-1
-2
-1
-2
-7
-6
-5
-4
-3
-2
-7
-6
-5
-4
-3
-2
log [DEO8]
log [TX-100]
org
org
Fig. 3. Plots of log D vs. log [POE]org in the extraction of Ln3þ with 1:0 ꢄ 10ꢅ2 M Htta at pH 4.9. POE: (a) DEO4, (b) DEO6,
(c) DEO8, (d) TX-100; Ln: ꢄ = La,
= Nd,
= Gd, = Er, = Lu. Solid lines are calculated curves; see text.
Table 1. Formation Constant of the POE Adduct [Ln(tta)3 (POE)], ꢀadd, in 1,2-Dichloroethane
ꢂ
Ln3þ Ionic radii log Kex
log ꢀadd
DEO8
a)
DEO4
DEO6
TX-100
12C4a) 18C6a)
˚
/A
La3þ
Ce3þ
Pr3þ
1.216
1.196
1.179
1.163
1.132
1.120
1.107
1.095
1.083
1.072
1.062
1.052
1.042
1.032
ꢅ10:09 3:90 ꢆ 0:03 5:11 ꢆ 0:06 5:39 ꢆ 0:02 5:49 ꢆ 0:04
ꢅ9:42 3:82 ꢆ 0:03 5:09 ꢆ 0:04 5:37 ꢆ 0:02 5:48 ꢆ 0:03
ꢅ9:12 3:65 ꢆ 0:03 4:95 ꢆ 0:04 5:44 ꢆ 0:02 5:52 ꢆ 0:03
ꢅ8:92 3:52 ꢆ 0:03 4:91 ꢆ 0:03 5:40 ꢆ 0:02 5:50 ꢆ 0:03
ꢅ8:23 3:27 ꢆ 0:03 4:89 ꢆ 0:02 5:39 ꢆ 0:02 5:45 ꢆ 0:03
ꢅ8:08 3:20 ꢆ 0:03 4:81 ꢆ 0:02 5:29 ꢆ 0:02 5:38 ꢆ 0:03
ꢅ8:08 3:10 ꢆ 0:03 4:70 ꢆ 0:02 5:22 ꢆ 0:02 5:29 ꢆ 0:04
ꢅ7:63 3:00 ꢆ 0:03 4:60 ꢆ 0:02 5:15 ꢆ 0:03 5:27 ꢆ 0:04
ꢅ7:48 2:93 ꢆ 0:04 4:50 ꢆ 0:03 5:02 ꢆ 0:03 5:21 ꢆ 0:04
ꢅ7:45 2:91 ꢆ 0:04 4:48 ꢆ 0:03 5:14 ꢆ 0:03 5:15 ꢆ 0:04
ꢅ7:35 2:81 ꢆ 0:04 4:39 ꢆ 0:03 4:88 ꢆ 0:04 5:17 ꢆ 0:05
ꢅ7:16 2:76 ꢆ 0:05 4:25 ꢆ 0:04 4:85 ꢆ 0:04 4:97 ꢆ 0:05
ꢅ7:03 2:59 ꢆ 0:05 4:07 ꢆ 0:04 4:50 ꢆ 0:04 4:79 ꢆ 0:05
ꢅ7:02 2:51 ꢆ 0:05 3:91 ꢆ 0:04 4:36 ꢆ 0:04 4:64 ꢆ 0:05
3.87
3.84
3.86
3.84
3.81
3.73
3.69
3.61
3.56
3.51
3.49
3.35
3.31
3.18
6.35
6.49
6.54
6.51
6.42
6.42
6.33
6.25
6.15
6.15
6.07
5.96
5.75
5.68
Nd3þ
Sm3þ
Eu3þ
Gd3þ
Tb3þ
Dy3þ
Ho3þ
Er3þ
Tm3þ
Yb3þ
Lu3þ
a) Cited from Ref. 34. The values of ꢀadd for 12C4 and 18C6 were corrected by taking into consideration the
distribution coefficient of crown ethers.

G. F. Gagabe et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 12 (2007) 2361
ionic size.45–48 The increase in the number of oxygen atoms
of POE coordinating to the lanthanoid ion due to the increase
in the ionic radii also causes the increase in the stability of the
adduct.
7
Comparison of Adducts Formed among Polyethers. The
value of log ꢀadd for 18C6 is about 2.5 orders of magnitude
higher than that for the 12C4 adducts of the corresponding
metal ion. These large values of ꢀadd for 18C6 complexes are
attributable to the stable structures of these complexes. The
6
5
β
˚
cavity size of 18C6 has been estimated as 1.45 A and is larger
than the ionic radius of any Ln3þ ion.49 Thus, 18C6 forms sta-
ble complexes with Ln3þ by incorporating the metal ion inside
of its cavity. The metal ion coordinated by 18C6 does not have
enough space for the direct coordination of three bidentate tta
ligands. Hence, it is thought that one tta molecule is expelled
from the inner coordination sphere of the complex, and the cat-
4
3
2
ionic complex [Ln(tta)2 (18C6)] is formed. The formation of
this kind of cationic complex has been reported for the extrac-
ꢂ
ꢅ
35
tion of Ln3þ ions in the presence of ClO4
.
Moreover, for-
mation of [Ln(tta)2 (18C6)] in crystal and in organic solution
ꢂ
has been confirmed by X-ray crystallographic and 1H NMR
spectroscopic analyses.50 Equilibrium studies indicate the
participation of three tta ions for the adduct formation. Accord-
ingly, the third tta may form an ion pair with the cationic com-
1.00 1.05 1.10 1.15 1.20 1.25
ionic radius / Å
Fig. 4. The plots of logarithmic formation constants of ad-
ducts as a function of ionic radii of Ln. POE:
plex as [Ln(tta)2 (18C6)]þ(ttaꢅ). On the basis of the preceding
ꢂ
= DEO4,
= 12C4,
information, the structure of [Ln(tta)2 (18C6)]þ(ttaꢅ) was es-
ꢂ
= DEO6,
= 18C6.
= DEO8,
= TX-100,
timated by MM2 calculations. The resulting structure is shown
in Fig. 5a. Generally, the complex formation constant of the
third step is much smaller than those of the first and second
steps. Thus, the third ttaꢅ coordinating the metal ion must
be substituted by the 18C6, which is a multidentate ligand
and can incorporate the lanthanoid ion in its cavity.
ability of the metal complex considerably increases by the
replacement of these water molecules with lipophilic POE
molecules. The logarithmic formation constants of the adducts
[Ln(tta)3 (POE)] of the linear POEs, log ꢀadd, are plotted as a
ꢂ
function of ionic radii of lanthanoids43 in Fig. 4, where the re-
sults for crown ethers are also shown. Similar to the results for
crown ethers, the value of ꢀadd for linear POEs generally in-
creased as the radius of Ln3þ increased. The hydration energy
decreases as the radius of Ln3þ ions increases. Because of this,
the displacement of coordinated water molecules by the POE
would be much easier for larger Ln3þ than for smaller ones.
Hence, lighter Ln3þ ions more favorably form the adduct
with the POE. In case of the adduct formation with a mono-
dentate ligand such as tributylphosphate (TBP) or trioctylphos-
phine oxide (TOPO), the formation constants of the adduct
[Ln(tta)3 (TBP)] or [Ln(tta)3 (TOPO)] show an opposite trend
In case of 12C4 adduct, the metal ion cannot be incorporat-
ed inside of the crown ether ring, because the cavity size of
12C4 (0.72 A)49 is smaller than the ionic radii of the lanthanoid
˚
ions. The ꢀadd values for 12C4 adducts are significantly lower
than those of 18C6. Thus, the 12C4 may coordinate to the
metal ion of the complex [Ln(tta)3] as a bidentate or tridentate
ligand. The structure of [Ln(tta)3 (12C4)] estimated by MM2
calculations, which is shown in Fig. 5b, reasonably explained
the above results.
ꢂ
The adduct formation constants of DEO4 are on the same
orders of magnitude as those of 12C4 adducts. The number
of ethylene oxide groups in DEO4 is not enough to surround
the Ln3þ ions; hence, the stability and structure of the DEO4
adduct must be similar to those of the 12C4 adducts. The struc-
ture of the adduct of La–hexafluoroacetylacetonate complex
with the linear polyether tetraglyme has been reported for
the crystalline complex. The La3þ ion is encapsulated by the
six oxygen atoms of three ꢀ-diketones, and is coordinated
by four oxygen donor atoms of the tetraglyme to form a 10-
ꢂ
ꢂ
with the present results for POEs, that is, log ꢀadd for TBP or
TOPO decreases as the radius of Ln3þ ion increases.7,9,11
Lewis basicity of TBP and TOPO are stronger than water, thus
the stability of the adduct increases by a decrease in ion size
of the metal ion. On the other hand, POEs are much weaker
bases; thus, the hydration energy of the lanthanoid ions has a
greater effect on adduct formation.
The coordination number of lanthanoid ions increases from
eight to nine by the increase in the ionic radius of Ln3þ. The
coordinate complex.46 The structure of the adduct [Ln(tta)3
ꢂ
(DEO4)] in solution could be similar to this crystal structure.
However, because of the bulkiness of the ligands, i.e., tta
and DEO4, it is thought that a 9-coordinate adduct is formed,
that is, three oxygen atoms of DEO4 is coordinating to the
metal ion. The terminal –OH may coordinate to the metal
ion, because the basicity of the hydroxy oxygen is higher than
that of the ethereal oxygen.51 An optimized structure of
usual coordination number is eight for the heavier Ln3þ
,
Tb3þ to Lu3þ, nine for the lighter Ln3þ, La3þ to Sm3þ, and
both for Gd3þ and Eu3þ
.
Crystal structures of the adducts
44
of Ln3þ ions with hexafluoroacetylacetone and linear oligo-
glymes have shown that the number of oxygen atoms of the
polyether coordinating to the Ln3þ ion center increases with

2362 Bull. Chem. Soc. Jpn. Vol. 80, No. 12 (2007)
Lanthanoid Complexes with Polyether
(a)
(b)
(c)
(d)
Fig. 5. Proposed structures of the adduct complexes of [La(tta)3 (POE)], POE: (a) 18C6, (b) 12C4, (c) DEO4, and (d) DEO6.
ꢂ
Optimized by MM2 calculations; see text. La = green, O = red, C = dark gray, H is omitted for simplicity, tta is given by
stick frame.
[Ln(tta)3 (DEO4)] adduct obtained by MM2 calculations is
shown in Fig. 5c.
may be attributed to some additional interactions of the DEO6
with the metal ion center. The uncoordinated oxygen atoms of
the residual EO chain of DEO6 may electrostatically interact
with the Ln3þ ion center at the outer sphere of the complex.
ꢂ
The adduct formation constant for DEO6 is much smaller
than 18C6, although the number of ethylene oxide units is
the same between them. This may indicate that the number
of oxygen atoms coordinating to the metal ion in DEO6 is
smaller than that of 18C6. When the counter anion of the metal
complex of linear POE is a non-coordinating ion, such as pic-
rate (picꢅ) or nitrate, the metal ion center is surrounded by the
oxygen atoms of the ethylene oxide chain of DEOn. For exam-
ple, in the solid-state adduct of linear pentaethylene glycol
Molecular structure of the [Ln(tta)3 (DEO6)] adduct opti-
ꢂ
mized by MM2 calculations is shown in Fig. 5d. The outer
sphere interaction of residual oxygen atoms of DEO6 is
supported by the structure generated by MM2 calculation.
The differences in log ꢀadd values between DEO6 and
DEO8, and between DEO8 and TX-100 (nave ¼ 9:6) are not
so significant compared to that between DEO4 and DEO6.
As can be seen from the structure of DEO6 complex (Fig. 5d),
six ethylene oxide units are enough for the outer sphere
interaction with the [Ln(tta)3] complex.
(PEG5) with neodymium nitrate, [Nd(NO3)3 (PEG5)], all of
ꢂ
the six oxygen atoms of PEG5 coordinate to the Nd3þ ion
by forming a ring-like configuration around the metal ion cen-
ter similar to that of the crown ether complex.17 NMR studies
revealed that the POE complex has similar structure even in
solution. That is, in the [K(DEO6)](pic) complex, the EO moi-
ety of DEO6 is coordinating to the Kþ ion by wrapping around
the metal ion center.37,51 On the other hand, DEO6 does not
substitute the coordinated tta ligands in the present extraction
system, since the coordinating ability of tta is higher than that
of POE. This may imply that the structure of the DEO6 adduct
might be similar to that of the DEO4 complex, that is, two or
three oxygen atoms of DEO6 coordinate to the metal ion in
[Ln(tta)3] complex. However, the ꢀadd values of DEO6 adduct
are 30-fold higher than that of the DEO4 adduct. These results
Extraction Constant of POE Adducts.
synergistic extraction constant, Kex,add, in Eq. 5 is given by
The over-all
combining Eqs. 3 and 8:
log Kex,add ¼ log Kex þ log ꢀadd
:
ð14Þ
The values of log Kex,add thus obtained are listed in Table 2.
The log Kex,add is plotted as a function of ionic radius of
Ln3þ in Fig. 6 together with log Kex and log ꢀadd, where the
results for DEO6 system are shown as the example. Because
of its negative charge and chelating effect, the coordination
ability of the tta ligand is higher than that of the water mole-
cule. Thus, the log Kex value decreases as the size of Ln3þ

G. F. Gagabe et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 12 (2007) 2363
Table 2. Extraction Constants of the Adduct Complex [Ln(tta)3 (POE)],
Kex,add, into 1,2-Dichloroethane
ꢂ
Ln3þ
log Kex,add
DEO4
DEO6
DEO8
TX-100
12C4
18C6
La3þ
Ce3þ
Pr3þ
ꢅ6:19
ꢅ5:60
ꢅ5:47
ꢅ5:40
ꢅ4:96
ꢅ4:88
ꢅ4:98
ꢅ4:63
ꢅ4:55
ꢅ4:54
ꢅ4:54
ꢅ4:40
ꢅ4:44
ꢅ4:51
ꢅ4:98
ꢅ4:35
ꢅ4:17
ꢅ4:01
ꢅ3:34
ꢅ3:27
ꢅ3:38
ꢅ3:03
ꢅ2:98
ꢅ2:97
ꢅ2:96
ꢅ2:91
ꢅ2:96
ꢅ3:11
ꢅ4:70
ꢅ4:05
ꢅ3:68
ꢅ3:52
ꢅ2:84
ꢅ2:79
ꢅ2:86
ꢅ2:48
ꢅ2:46
ꢅ2:31
ꢅ2:47
ꢅ2:31
ꢅ2:53
ꢅ2:66
ꢅ4:60
ꢅ3:94
ꢅ3:60
ꢅ3:42
ꢅ2:78
ꢅ2:70
ꢅ2:79
ꢅ2:36
ꢅ2:27
ꢅ2:30
ꢅ2:18
ꢅ2:19
ꢅ2:24
ꢅ2:38
ꢅ6:22
ꢅ5:58
ꢅ5:26
ꢅ5:08
ꢅ4:42
ꢅ4:35
ꢅ4:39
ꢅ4:02
ꢅ3:92
ꢅ3:94
ꢅ3:86
ꢅ3:81
ꢅ3:72
ꢅ3:84
ꢅ3:74
ꢅ2:93
ꢅ2:58
ꢅ2:41
ꢅ1:81
ꢅ1:66
ꢅ1:75
ꢅ1:38
ꢅ1:33
ꢅ1:30
ꢅ1:28
ꢅ1:20
ꢅ1:28
ꢅ1:34
Nd3þ
Sm3þ
Eu3þ
Gd3þ
Tb3þ
Dy3þ
Ho3þ
Er3þ
Tm3þ
Yb3þ
Lu3þ
small difference in the extraction constants of the lanthanoid
ions is preferable for these purposes.
10
8
-2
Conclusion
-4
The synergistic extraction of thenoyltrifluoroacetone com-
plexes of the whole series of lanthanoid ions (except promethi-
um) with linear poly(oxyethylene) was studied and compared
to that of crown ether systems. (1) The extraction of lantha-
noids is significantly enhanced by the formation of the 1:1 ad-
6
-6
β
duct [Ln(tta)3 (POE)] in organic phase. (2) The adduct forma-
ꢂ
tion constant increased with an increase in the ionic radius of
lanthanoid ion. (3) The order of the adduct formation constants
among the POE compounds is as follows: 18C6 ꢇ TX-100 ꢈ
DEO8 > DEO6 ꢇ 12C4 ꢃ DEO4. (5) The structures of ad-
duct complexes were estimated from the adduct formation
constants with the aid of MM2 calculation. (6) The synergistic
extraction with linear POE is advantageous for the group
separation of lanthanoid ions.
4
2
0
-8
-10
-12
1.00
1.05
1.10
1.15
1.20
1.25
References
ionic radius / Å
1
A. A. Ansari, Z. Ahmed, A. Iftukhar, Spectrochim. Acta,
Part A 2007, 68, 176.
D. Chapon, C. Husson, P. Delangle, C. Lebrun, P. Vottero,
J. Alloys Compd. 2001, 323–324, 128.
Fig. 6. Plots of extraction constant of [Ln(tta)3], and
extraction and adduct formation constants of [Ln(tta)3
ꢂ
´
2
(DEO6)] as a function of ionic radius.
= log Kex,add = log ꢀadd
= log Kex;
;
.
3
4
J. C. Reid, M. Calvin, J. Am. Chem. Soc. 1950, 72, 2948.
J. Alstad, J. H. Augustson, L. Farbu, J. Inorg. Nucl. Chem.
ion increases. On the other hand, the log ꢀadd increases as the
ionic radius of lanthanoid increases. These opposite trends be-
tween the log Kex and log ꢀadd resulted in the leveling of the
range of equilibrium constants of the synergistic extraction,
log Kex,add. In particular, the values of Kex,add of heavy lantha-
noids (Sm3þ to Lu3þ) barely change.
Usually, the synergistic extraction of lanthanoid ions has
been investigated for the purpose of mutual separation among
them. Recently, ICP method has become popular for the simul-
taneous determination of elements. Thus, the group separation
is of great importance for the simultaneous analysis of metal
ions.52 Also, group separation is quite essential since the ex-
traction of rare-earth ions is the primary step in the partitioning
of high-level liquid wastes containing fission products.53 Thus,
1974, 36, 899.
5
34, 3221.
6
1961, 16, 323.
7
1481.
8
E. F. Kassierer, A. S. Kertes, J. Inorg. Nucl. Chem. 1972,
A. M. Poskanzer, B. M. Foreman, Jr., J. Inorg. Nucl. Chem.
T. Sekine, D. Dyrssen, J. Inorg. Nucl. Chem. 1967, 29,
N. Kameta, H. Imura, Bull. Chem. Soc. Jpn. 2001, 74,
D. D. Ensor, A. H. Shah, Solvent Extr. Ion Exch. 1984, 2,
1641.
9
591.
10 S. Nakamura, N. Suzuki, Bull. Chem. Soc. Jpn. 1993,
66, 98.
11 T. Sekine, Y. Koike, Y. Hasegawa, Bull. Chem. Soc. Jpn.

2364 Bull. Chem. Soc. Jpn. Vol. 80, No. 12 (2007)
Lanthanoid Complexes with Polyether
1969, 42, 432.
Chem. 1992, 159, 353.
34 G. F. Gagabe, Y. Satoh, K. Satoh, K. Sawada, Monatsh.
Chem. 2006, 137, 1015.
35 R. Rogers, R. Etzenhouser, J. Murdoch, E. Reyes, Inorg.
Chem. 1991, 30, 1445.
12 R. Izatt, J. S. Bradshaw, S. A. Neilsen, J. D. Lamb, J. J.
Christensen, Chem. Rev. 1985, 85, 271.
13 S. Yanagida, K. Takahashi, M. Okahara, Bull. Chem. Soc.
Jpn. 1978, 51, 1294.
14 F. Vogtle, E. Weber, Angew. Chem., Int. Ed. Engl. 1979,
18, 753.
15 T. Okada, Analyst 1993, 118, 959.
16 B. Tummler, G. Mass, E. Weber, W. Wehner, F. Vogtle,
¨
¨
J. Am. Chem. Soc. 1977, 99, 4683.
17 Y. Hirashima, K. Kanetsuki, J. Shiokawa, N. Tanaka, Bull.
Chem. Soc. Jpn. 1981, 54, 1567.
`
18 G. Malandrino, I. Fragala, Coord. Chem. Rev. 2006, 250,
36 K. Sawada, K. Satoh, C. Haruta, Y. Kikuchi, Phys. Chem.
Chem. Phys. 1999, 1, 2737.
37 Y. Kikuchi, Y. Nojima, H. Kita, T. Suzuki, K. Sawada,
Bull. Chem. Soc. Jpn. 1992, 65, 1506.
38 Y. Kikuchi, T. Suzuki, K. Sawada, Anal. Chim. Acta 1992,
264, 65.
39 Y. Kikuchi, N. Takahashi, T. Suzuki, K. Sawada, Anal.
Chim. Acta 1992, 256, 311.
¨
1605.
19 S. Yanagida, K. Takahashi, M. Okahara, Bull. Chem. Soc.
Jpn. 1977, 50, 1386.
40 Spartan ’02, Wavefunction, Inc., Irvine CA, USA.
41 Y. Kikuchi, Y. Sakamoto, Anal. Chim. Acta 2000, 403,
325.
20 A. M. Y. Jaber, G. J. Moody, J. D. R. Thomas, Analyst
1981, 106, 439.
42 I. Matsubayashi, E. Ishiwata, T. Shionoya, Y. Hasegawa,
Talanta 2004, 63, 625.
21 Y. Sakai, H. Nakamura, M. Takagi, K. Ueno, Bull. Chem.
Soc. Jpn. 1986, 59, 381.
22 I. Yoshida, R. Takeshita, K. Ueno, M. Takagi, Anal. Sci.
1986, 2, 53.
43 R. D. Shannon, Acta Crystallogr., Sect. A 1976, 32, 751.
44 T. Toraishi, S. Nagasaki, S. Tanaka, Inorg. Chim. Acta
2007, 360, 1575.
`
45 G. Malandrino, R. Licata, F. Castelli, I. Fragala, C. Benelli,
23 B. S. Mohite, S. M. Khopkar, Anal. Chem. 1987, 59, 1200.
24 L. Favretto, F. Tunis, Analyst 1976, 101, 198.
Inorg. Chem. 1995, 34, 6233.
46 G. Malandrino, C. Benelli, F. Castelli, I. Fragala, Chem.
`
ˆ
25 K. Toei, S. Motomizu, T. Umato, Talanta 1982, 29, 103.
Mater. 1998, 10, 3434.
47 K. D. Pollard, J. J. Vital, G. P. A. Yap, R. J. Puddephatt,
J. Chem. Soc., Dalton Trans. 1998, 1265.
26 B. P. Czech, D. A. Babb, B. Son, R. A. Bartsch, J. Org.
Chem. 1984, 49, 4805.
`
48 G. Malandrino, R. Lo Nigro, I. Fragala, C. Benelli, Eur. J.
27 S. Kitazawa, K. Kimura, H. Yano, T. Shono, J. Am. Chem.
Soc. 1984, 106, 6978.
28 P. Seiler, M. Dobler, J. D. Dunitz, Acta Crystallogr., Sect.
B 1974, 30, 2744.
Inorg. Chem. 2004, 500.
49 R. Colton, S. Mitchell, J. C. Traeger, Inorg. Chim. Acta
1995, 87, 231.
29 K. Sawada, K. Sato, Bunseki 2001, 11, 599.
30 Y. Kitatsuji, Y. Meguro, Z. Yoshida, T. Yamamoto, K.
Nishizawa, Solvent Extr. Ion Exch. 1995, 13, 289.
31 Y. Meguro, Y. Kitatsuji, T. Kimura, Z. Yoshida, J. Alloys
Compd. 1998, 271–273, 790.
32 G. M. Nair, D. R. Prabhu, J. Radioanal. Nucl. Chem. 1988,
121, 83.
33 F. A. Shehata, S. M. Khalifa, H. F. Aly, J. Radioanal. Nucl.
50 Y. Yano, M. Yatabe, T. Ohsawa, A. Aruga, Anal. Sci.
2001, 17, i1647.
51 K. Sawada, A. Imai, K. Satoh, Y. Kikuchi, J. Phys. Chem.
B 2007, 111, 4361.
52 X.-Y. Zhang, K. Satoh, A. Satoh, K. Sawada, T. Suzuki,
Anal. Chem. 1997, 13, 891.
53 A. Fujiwara, Y. Nakano, T. Yaita, K. Okuno, J. Alloys
Compd. 2007. doi:10.1016/j.jallcom.2007.02.056.