Ion-size recognition of Group 13 metals (Al3+, In3+) with modified β-diketones

Article abstract of DOI:10.1039/a702377d

Ion-size recognition of Group 13 metals (Al³⁺ and In³⁺) with modified β-diketones, 3-phenylpentane-2,4-dione (α-phenylacetylacetone, HL²) and 1,2-diphenylbutane-1,3-dione (α-phenylbenzoylacetone, HL⁴), has been

Full text of DOI:10.1039/a702377d

View Article Online / Journal Homepage / Table of Contents for this issue
3؉
3؉
Ion-size recognition of Group 13 metals (Al , In ) with modified
â-diketones
Quyen T. H. Le, Shigeo Umetani* and Masakazu Matsui
Institute for Chemical Research, Kyoto University, Uji, Kyoto 611, Japan
3ϩ
3ϩ
Ion-size recognition of Group 13 metals (Al and In ) with modified β-diketones, 3-phenylpentane-2,4-dione
2
4
(
α-phenylacetylacetone, HL ) and 1,2-diphenylbutane-1,3-dione (α-phenylbenzoylacetone, HL ), has been studied
1
by the liquid–liquid extraction method and compared with that of pentane-2,4-dione (acetylacetone, HL ) and
3
1
3
2
1
-phenylbutane-1,3-dione (benzoylacetone, HL ). While HL and HL quantitatively extracted both ions, HL and
4 3ϩ 3ϩ 3ϩ 3ϩ
HL greatly discriminated Al from In : Al was readily extracted into benzene at a pH below 5, while In
3ϩ
3ϩ
was entirely unextractable, leading to an effective extraction separation of Al from In . This could be ascribed
to the large ‘bite size’ (O ؒ ؒ ؒ O distance in the chelate ring) in the indium β-diketonates. The steric repulsions
2
4
between the α-phenyl group and the terminal methyl or phenyl groups prevent HL and HL from widening their
3ϩ
bite size. In addition, the extraction of Al was greatly influenced by the interligand contact due to its remarkably
1
3
3ϩ
3ϩ
small ionic radius. The extraction order with HL and HL is Al > In obeying the common tendency that the
metal ion with a smaller ionic radius is more readily extracted, while that with ligands having bulky terminal
5
groups, such as 1,3-diphenylpropane-1,3-dione (dibenzoylmethane, HL ), is the opposite. These steric factors have
3ϩ
3ϩ
been evaluated by the investigation of the crystal structures of the complexes of Al and In and by molecular
mechanics calculations. The steric factors have been confirmed through extraction behaviors with various kinds of
β-diketones.
β-Diketones have been proven to be versatile ligands for various
metal ions. A large number of studies have been done on these
chelating reagents, especially in the field of solvent extraction
of metal ions for the purpose of separation and concen-
R3
R3
R1
R2
R1
R2
O
O
O
O
1–3
H
tration. Recently, β-diketones have been examined in the
4–7
R1
R2
R3
disposal of radioactive waste using supercritical fluids, in
β-Diketone
8,9
1
superconducting thin-film manufacturing1 and as metal–
Me
Me
Me
Me
Ph
C4H3S
Me
Ph
Me
Me
Ph
Ph
Ph
CF3
CF3
CF3
CF3
H
Ph
H
Ph
H
H
H
H
Me
HL
0,11
2
organic chemical vapor deposition sources.
Much attention
HL
3
HL
has also been paid to the structural chemistry of β-diketones
4
HL
involving keto–enol tautomerism and the intramolecular
5
12,13
HL
hydrogen bond.
However, there seems to be a lack of sys-
6
HL
tematic studies on the correlation between the structure of
β-diketones and their selectivity for metal ions.
7
HL
8
HL
9
Studies on the solvent extraction of metal ions with
HL
Ph
14,15
16
acylpyrazolones
and acylisoxazolones, which are structur-
ally analogous to β-diketones derived from five-membered
heterocyclic compounds, revealed that the improved extraction
with strongly acidic extractants is usually accompanied by poor
selectivity. These compounds were found to have longer dis-
tances between the two donating oxygens as compared to the
conventional β-diketones, such as acetylacetone and 4,4,4-tri-
fluoro-1-(2-thienyl)butane-1,3-dione (thenoyltrifluoroacetone),
according to molecular orbital calculations. These facts
prompted us to consider that the O ؒ ؒ ؒ O distance plays an
important role in complexing metal ions. Recently, on studying
the extraction of lanthanides using 2-trifluoroacetylcyclo-
gate how the O ؒ ؒ ؒ O distance and other factors in the structure
19
of β-diketones affect their complexation reactions.
In the present paper the selectivity of modified β-diketones
3
ϩ
3ϩ
for Al (r = 0.53 Å for a co-ordination number of six) and In
20
(r = 0.80 Å) has been examined via the solvent-extraction
3ϩ
3ϩ
method, and the stability of β-diketones of Al and In dis-
cussed by means of X-ray crystallographic studies and molecu-
lar mechanics calculations.
Experimental
Syntheses
17
alkanones and acylpyrazolones having bulky substituents
18
4
at the 4 position, it was found that the separability of
lanthanides was clearly improved with those ligands having the
shorter O ؒ ؒ ؒ O distance. It could be concluded that the O ؒ ؒ ؒ O
distance is one of the most significant factors that govern select-
ivity in the complexation of β-diketones with lanthanides. If the
structure of the β-diketone could be suitably modified by intro-
ducing bulky groups at suitable positions to create a steric effect
the O ؒ ؒ ؒ O distance could be intentionally controlled, and,
consequently, the extractability and/or the separability could be
improved. Besides lanthanide ions with very similar ionic radii,
we now expand our research to the metal ions of Group 13
having large differences in ionic radii in an attempt to investi-
á-Phenylbenzoylacetone (HL ). This compound was syn-
21
thesized according to the literature through the catalysed
acylation of deoxybenzoin with acetic anhydride in the pres-
ence of boron trifluoride, followed by treatment with aqueous
sodium acetate. To a rapidly stirred dichloromethane solution
3
(200 cm ) of deoxybenzoin (2-phenylacetophenone) (0.3 mol)
and acetic anhydride (0.6 mol) kept cold in an ice-bath, BF (0.6
3
mol) was slowly added and the mixture rapidly stirred for
3
around 20 h. An aqueous solution (500 cm ) of sodium acetate
(1.2 mol) was then poured in, and the mixture was refluxed for
4
3 h. The crude HL was twice extracted with diethyl ether and
accumulated by evaporation. The pale yellow solid obtained
J. Chem. Soc., Dalton Trans., 1997, Pages 3835–3840
3835

View Article Online
22
was purified by the copper salt method, twice recrystallized
centrifugation, the pH of the aqueous phase was measured and
with hexane and dried in vacuo at 50 ЊC (22.4% yield), m.p.
taken as the equilibrium value. The metal concentration in the
aqueous phase was determined by inductively coupled argon-
plasma atomic emission spectrometry. That in the organic
phase was measured in the same way after back-extraction by
stripping with hydrochloric acid solution. The sum of the metal
concentrations in the two phases agreed well with the initial
concentration.
1
8
1
7.6 ЊC. H NMR (CDCl ): δ 2.06 (s, 3 H, CH ), 7.08–7.32 (m,
3
3
0 H, C H ) and 17.39 (s, 1 H, OH) (Found: C, 80.59; H, 5.90;
6
5
O, 13.47. Calc. for C H O : C, 80.65; H, 5.92; O, 13.43%); m/z
16
14
2
ϩ
2
38 (M ).
5
3
Aluminium dibenzoylmethanate [AlL ] 1. This compound was
synthesized according to a previously reported general pro-
23,24
cedure.
An ethanol solution of dibenzoylmethane was
Calculations
added while stirring to a 5% aqueous solution of a stoichio-
metric quantity of AlCl3 buffered by sodium acetate. The
chelate immediately appeared, and the mixture was stored
overnight in a refrigerator for complete precipitation. The pre-
cipitate was filtered off, twice recrystallized in a chloroform–
heptane mixture and dried at 37 ЊC in vacuo. A single crystal
2
7
Semiempirical MNDO/H calculation was performed on fully
optimized molecular geometries on a Cray Y-MP2E/264 com-
puter using MNDO 93 (Cray Research Inc.).
Molecular mechanics calculations were carried out using the
CACHE system on a Macintosh Quadra 800 computer. The
program is based on Allinger’s MM2 force field
28
29,30
and com-
suitable for X-ray analysis was grown in acetone, m.p. 298.5 ЊC.
1
putes the net force acting on a molecule as the sum of the
energy terms of bond stretch, bond angle, dihedral angle,
improper torsion, van der Waals, electrostatics, and hydrogen
bond. The augmented force-field parameters obtained by
CACHE were used without modification. A charge of 3ϩ was
assigned to Al and In and M᎐O bonds were described as
coordinate bonds. The crystal structure was used as an initial
conformation for the calculation of the ligands and complexes.
H NMR (CDCl ): δ 6.94 (s, 1 H, CH) and 8.05–7.32 (m, 10 H,
3
C H ) (Found: C, 77.81; H, 4.76. Calc. for C H AlO : C,
6
5
45 33
6
7
7.58; H, 4.77%).
5
3
Indium dibenzoylmethanate [InL ]2. This compound was syn-
thesized and a single crystal suitable for X-ray analysis obtained
5
3
in a manner similar to that of [AlL ]; the solvent for recrystal-
1
lization was acetone. M.p. 255.6 ЊC. H NMR (CDCl ): δ 6.81
3
(
s, 1 H, CH) and 7.31–8.02 (m, 10 H, C H ) (Found: C, 68.88;
6 5
X-Ray crystallography
H, 4.15. Calc. for C H InO : C, 68.89; H, 4.24%).
45
33
6
Crystallographic data for compounds 1 and 2 are summarized
in Table 3. Crystals were mounted on a fine glass fiber with
epoxy cement. The lattice parameters and intensity data were
measured on a Rigaku AFC7R four-circle diffractometer
equipped with Ni-filtered Cu-Kα radiation (λ = 1.54178 Å) at
Chemicals
,4,4-Trifluoro-2-methyl-1-phenylbutane-1,3-dione (α-methyl-
4
9
benzoyltrifluoroacetylacetone, HL ) was synthesized according
25
to the method of Barkley and Levine through the conden-
sation of methyl trifluoroacetate and propiophenone in the
2
0 ± 1 ЊC. The ω–2θ scan technique was used to a maximum 2θ
2
6
2
3
value of 120Њ. An empirical absorption correction using the
presence of sodium methoxide. The compounds HL , HL ,
,4,4-trifluoro-1-phenylbutane-1,3-dione (benzoyltrifluoro-
31
program DIFABS was applied to the data sets. The data were
corrected for Lorentz and polarisation effects. A correction for
secondary extinction was applied. The structures were solved by
4
8
acetone, HL ) and 1,1,1-trifluoropentane-2,4-dione (trifluoro-
acetylacetone, HL ) were from Dojindo and used without
further purification. Aqueous solutions of Al and In were
prepared from standard solutions of Wako Chemicals and
stocked in 0.02 mol dm hydrochloric acid solution. Other
7
3
2
33
3ϩ
3ϩ
direct or heavy-atom Patterson methods, expanded using
33
the Fourier technique and refined by full-matrix least squares
for 3758 and 4497 reflections, respectively, with |F |> 3σ(F ).
Ϫ3
o
o
The non-hydrogen atoms were anisotropically refined. Hydro-
gens were fixed at calculated positions. All calculations were
performed using the TEXSAN crystallographic software pack-
chemicals were of analytical reagent grade. Water was
demineralized and distilled.
34
age. The complexes have a distorted octahedral geometry.
Their mean dimensions are displayed together for comparison
in Table 4.
Apparatus
Metal-ion concentrations were measured with a Japan Jarrell
Ash Model ICAP-500 inductively coupled argon-plasma
atomic emission spectrophotometer. pH Measurements were
made with a Hitachi-Horiba F-8L pH meter equipped with a
Horiba 6028 glass combination pH electrode. Proton NMR
spectra were recorded with a Varian VXR-200 spectrometer
CCDC reference number 186/682.
Results and Discussion
Molecular orbital calculations
(
200 MHz) at 25 ЊC in CDCl3.
The distances between the two donating oxygens were estimated
2
7
Measurements of acid dissociation constants of the ligands (pKa)
by semiempirical MNDO/H calculation which takes hydrogen
bonding into consideration. Our examination of the semi-
1
2
3
4
The pK values of compounds HL , HL , HL and HL were
a
35
empirical molecular orbital calculations including AM1 and
PM3 shows that MNDO/H is most suitable to evaluate the
structure of β-diketones. The O ؒ ؒ ؒ O distances found by
MNDO/H for various β-diketones are shown in Table 1.
The distances between the two donating oxygens of HL and
HL calculated by MNDO/H are 2.50 and 2.51 Å, respectively,
whereas those for HL and HL are reduced to 2.44 Å. A short-
ening of this distance was made by introducing a phenyl group
at the α position of HL and HL . The steric repulsion between
the phenyl group and the two terminal groups in the structure
determined by a potentiometric titration method: a solution
36
3
Ϫ3
(
(
20 cm ) of 0.02 mol dm diketone in water–1,4 dioxane
26
Ϫ3
25:75 v/v) containing 0.1 mol dm tetramethylammonium
perchlorate kept at 25 ЊC was rapidly stirred and titrated by 0.1
1
Ϫ3
mol dm tetramethylammonium hydroxide.
3
2
4
Solvent-extraction procedure
3
Ϫ4
1
3
A 10 cm aliquot of an aqueous phase containing 1 × 10 mol
Ϫ3
3ϩ
3ϩ
Ϫ3
dm Al or In , 0.1 mol dm sodium perchlorate and 0.01
Ϫ3
1
3
mol dm sodium acetate as buffering component was adjusted
of HL and HL results in a narrowing of this distance. This
structural change is also supported by H NMR data. The very
broad peaks assigned to the enolic proton (OH) for HL and
HL appear at δ 15.44 and 16.15, respectively, while those for
1
to the desired pH with hydrochloric acid or sodium hydroxide
solution. The aqueous phase was shaken with an equal volume
of benzene containing the required amount of diketone in a
1
3
3
2
4
centrifuge tube (30 cm ) at 25 ЊC for the appropriate time. After
HL and HL are very sharp and shift downfield to δ 16.67 and
3
836
J. Chem. Soc., Dalton Trans., 1997, Pages 3835–3840

View Article Online
1
2
3
4
Table 1 Comparison of properties of HL , HL , HL and HL
1
2
3
4
HL
HL
HL
HL
a
O ؒ ؒ ؒ O distance (Å)
HA
2.50
3.09
65.56
15.44
12.65
2.44
2.83
77.53
16.67
12.85
2.51
3.10
63.51
16.15
12.66
2.44
2.84
75.69
17.39
12.90
Ϫb
Aa
OH
Ϫ1
E
/kJ mol
1
H NMR, δ(OH)
c
pKa
a
b
c
From MNDO/H. Fixed at planar. Measured in 1,4-dioxane–water
Ϫ3
(
75:25 v/v), [NMe ClO ] = 0.1 mol dm .
4 4
1
7.39, respectively. It was reported that the hydrogen-bonded
37
enolic proton signal appears at that low magnetic field. These
3
ϩ
3ϩ
3
3
Ϫ3
Fig. 1 Extraction of Al and In with HL . [HL ] = 0.1 mol dm in
o
signals did not move on changing the concentration (0.01–0.1
Ϫ3
benzene, [NaClO ] = 0.1 mol dm
Ϫ3
4
mol dm ) but disappeared after adding D O. The sharp and
2
downfield signal of the enolic proton is caused by the strong
intramolecular hydrogen bonding due to the narrow O ؒ ؒ ؒ O
separation. The intramolecular hydrogen-bond energy (EOH
)
can be estimated as the energy difference between the hydrogen-
bonded structure and the open structure in which the O᎐H is
rotated by 180Њ in order to minimize the hydrogen-bond energy;
these values are shown in Table 1.
In addition, the facts that the enolic proton signals corre-
spond to one proton and no methyne protons were observed for
2
4
HL and HL indicate that both compounds exist quantitatively
in the enol form in CDCl . It has been reported that the pres-
3
1
2
ence of an α-substituent usually favors the keto tautomer, and
our previous results are in agreement with this tendency. The
results obtained here agree with reports
26
38,39
demonstrating that
an α-aryl group markedly increases the enol content unlike
other α-substituents, such as the methyl or bromo groups. The
increase in the keto content diminishes the applicability as an
organic ligand.
Acid dissociation constants (pKa)
All the pH measurements for determining pK were made in
a
aqueous 1,4-dioxane, so the reading of the pH meter was cali-
brated. In this work the pH measurements were carried out at
Ϫ3
2
5 ЊC with an ionic strength of 0.1 mol dm (tetramethylam-
monium perchlorate) in water–1,4-dioxane (25:75 v/v), and the
40
pH reading correction proposed by Irving and Mannot has
been adopted; equation (1), where B is the actual pH-meter
reading.
ϩ
Ϫlog[H ] = B ϩ 0.28
(1)
3
ϩ
3ϩ
2
Fig. 2 Extraction of Al and In into benzene with (a) HL and
4
2
4
Ϫ3
1
2
3
4
(b) HL . [HL ]
o
= [HL ]
o
= 0.1 mol dm in benzene, [NaClO
4
] = 0.1 mol
The pK values of HL , HL , HL and HL were found to be
Ϫ3
a
dm
1
2.65, 12.85, 12.66 and 12.90, respectively. The difference
between the pK values of a compound and its α-phenyl deriv-
a
3ϩ
The extractions are enhanced with increase in pH and Al is
ative could be attributed to the fact that the shorter O ؒ ؒ ؒ O
distances of the latter strengthen the intramolecular hydrogen
bond and thus restrain the release of the proton, resulting in a
decrease in acidity.
3ϩ
extracted better than In ; however, their extraction behaviors
are fairly similar. The extraction behavior with HL was analo-
gous to that with HL . The extraction behaviors significantly
changed through the introduction of a phenyl group at the α
position. The solvent extractions of Al and In from a 0.1
mol dm sodium perchlorate aqueous phase to a 0.1 mol dm
1
3
3ϩ
3ϩ
Equilibrium analysis
Ϫ3
Ϫ3
3ϩ
2
4
For the extraction of In a shaking time of 3 h was found to be
long enough to reach equilibrium with all the β-diketones. It is
reported that the rate of exchange of co-ordinated water for
HL or HL benzene phase are shown in Fig. 2. Almost 100%
3ϩ 2 3ϩ
of Al is extracted around pH 4.6 with HL , while In is
totally unextracted below pH 5. A similar result was obtained
3
ϩ
41
4
3ϩ
Al is very slow. Consequently, quite a long shaking time was
for HL . The extraction of In over pH 5 was also null, while
the concentration of In in the aqueous phase decreased,
3ϩ
3ϩ
required for the extraction of Al . For example, the extraction
with HL reached equilibrium at 26 h of shaking.
The complexation of the extractants HL and HL with Al
and In , compared to that of HL and HL , has been investi-
2
probably due to hydrolysis.
2
4
3ϩ
To clarify the composition of the extracted species, the effects
of pH and log[HA]o on log D (D = [M ] /[M ]) was also
3ϩ
1
3
3ϩ
3ϩ
o
gated via the solvent-extraction method. Fig. 1 shows plots of
examined. Here, HA stands for the diketone and the subscript o
denotes the species in the organic phase. Plots of log D vs. pH
3ϩ
3ϩ
the extraction percentages of Al and In vs. the pH of the
Ϫ3
3
aqueous phase in the extraction with 0.1 mol dm HL in ben-
and vs. log [HA] are straight lines with a slope of 3, indicating
o
Ϫ3
zene from 0.1 mol dm sodium perchlorate aqueous media.
that the extracted species is AlA and no self-adducts, such as
3
J. Chem. Soc., Dalton Trans., 1997, Pages 3835–3840
3837

View Article Online
5
5
Table 2 Extraction data for Al and In
Table 3 Crystallographic data* for AlL 1 and InL
2
3
3
a
pH1/2
log Kex
1
2
Formula
M
a/Å
b/Å
c/Å
α/Њ
C H AlO
696.73
C H InO
6
784.57
4
5
33
6
45 33
Diketone
Al
In
Al
In
1
HL2
3.30
4.01
3.20
4.24
2.12
3.00
3.47
3.95
—
3.39
—
2.63
2.51
4.98
Ϫ6.90
Ϫ8.85
—
Ϫ7.17
—
Ϫ4.89
Ϫ4.53
Ϫ11.97
10.7308(8)
17.939(2)
9.8365(9)
94.784(9)
107.184(6)
93.213(9)
1796.0(3)
11.497(2)
16.339(1)
10.5361(4)
97.319(8)
102.079(9)
107.474(8)
1087.3(4)
HL3
HL4
HL
Ϫ9.03
Ϫ6.60
Ϫ9.72
Ϫ3.36
Ϫ6.00
Ϫ7.41
7
HL8
β/Њ
HL9
HL
γ/Њ
U/Å
3
a
Ϫ3
Ϫ3
b
Crystal size/mm
0.20 × 0.15 × 0.60
1.288
9.05
0.40 × 0.20 × 0.40
1.442
56.32
[
HA] = 0.1 mol dm in benzene, [NaClO ] = 0.1 mol dm
.
Data
Ϫ3
o
4
c
D
c
/g cm
taken from ref. 42. Data taken from ref. 19.
Ϫ1
µ/cm
Reflections collected
Independent reflections
Observed reflections
R
5082
4793
3758
0.044
0.058
1.75
5111
4835
4497
0.034
0.047
1.66
AlA (HA) (s = 1,2 … ) are observed. Based on these slope
3
s
analyses, the extraction equilibrium can be expressed as in
equation (2). The extraction constant, K , is defined in
ex
RЈ
Goodness of fit
3ϩ
ϩ
Al ϩ 3HAo
AlA3,o ϩ 3H
(2)
¯
Details in common: triclinic, space group P1 (no. 2); Z = 2; R =
2
2 ¹
*
Σ( F | Ϫ |F )/Σ|F |; RЈ = [(Σw(|F | Ϫ |F |) /Σw|F | ] �² .
o
c
o
o
c
o
equations (3) and (4). The log K values are obtained by substi-
ex
ϩ 3
3ϩ
3ϩ
ϩ 3
Al and In , the opposite extraction order is frequently seen
depending on the bulkiness of the ligands. Aluminium is
[
[
AlA ] [H ]
[H ]
3
o
Kex
=
= D
(3)
(4)
3ϩ
Al ][HA]o3
[HA]o3
3ϩ
1
3
extracted better than In with HL and HL . In the case of
bulky β-diketones, such as HL and HL , the steric crowding
of the ligands around Al greatly destabilizes complexation
5
6
l
o
g
K
=
l
o
g
D
Ϫ
3
p
H
Ϫ
3
l
o
g
[
H
A
]
e
x
o
3ϩ
because of its very small ionic radius, while this is not so serious
tuting the pH1/2 values, i.e. the pH read from plots of log D vs.
pH at which half of the metal ion is extracted (log D = 0), in
equation (4) and are summarized in Table 2. Since the concen-
tration of the metal ions is much smaller than that of the
extractant, extractant concentration in the organic phase
3ϩ
3ϩ
3ϩ
in the case of In . As a result, In is better extracted than Al
with these bulky β-diketones. It should be noted here that the
1
3
6
5
O ؒ ؒ ؒ O distances of HL , HL , HL and HL are similar
according to the molecular orbital calculations.
It is seen from the extraction constants (Table 2) that
{
[HA] in equation (4)} can be considered equal to the initial
o
3ϩ
3ϩ
2
4
although Al is completely isolated from In by HL and HL ,
concentration.
As previously mentioned, Al and In are extracted with
1
3ϩ
3ϩ
the extractions were disturbed compared to those for HL and
3
3ϩ
2
1
3
HL , i.e., the distribution curves log D vs. pH of Al for HL1
HL and HL with similar extraction constants. On the con-
4
3
ϩ
2
4
3ϩ
and HL shifted more to the alkaline region than those of HL
trary, In is not extracted at all with HL and HL , while Al
is extracted with smaller extraction constants compared to
3
3ϩ
2
4
and HL . The poor extractions of Al with HL and HL are
partly ascribed to the lower acidity, though this is not so
1
3
those for HL and HL . Considering that the substituents at the
α position of the β-diketones should essentially not affect their
complexation with the metal ions, the present results are
surprising.
important because the differences in the pK values are very
a
small. It could be also attributed to the interligand contact in
the aluminium complex. The steric crowding of the ligands is
enhanced by the α-phenyl group which disturbs the free
rotation of the terminal methyl or phenyl groups. The free
rotation is expected to diminish the interligand contact.
An X-ray crystallographic consideration of β-diketones of
Al and In was made to confirm the role of the substituent at the
1
α position. The O ؒ ؒ ؒ O distances of the β-diketones in AlL
3
4
3
1
44
6 45
3
,
InL 3
4
and InL 3 were reported to be 2.726, 2.905 and
2
.905 Å, respectively. A survey of X-ray data for β-diketones of
Crystal structures of â-diketonates of Al and In
Al and In shows that the O ؒ ؒ ؒ O distances in aluminium
β-diketones are among the shortest ones, while those in indium
β-diketones are among the longest. According to MNDO/H
calculations (Table 1), the O ؒ ؒ ؒ O distances in the anionic form
The steric factors were evaluated by examining the crystal
5
5
1
structures of AlL 1 and InL 3 2 in addition to those of AlL 3
3
43
1
44
3
and InL 3 4.
The M᎐O distances for compounds 1 and 3 are much shorter
than those for 2 and 4 due to the remarkably small ionic radius
2
4
1
of HL and HL are 2.83 and 2.84 Å, while those for HL and
3
3ϩ
HL are 3.09 and 3.10 Å. The non-extractability of In with
3ϩ
1
of Al . The M᎐O distance in AlL is one of the shortest
among the tervalent metal acetylacetonates, while that in InL
is one of the longest. Consequently, the mean distance of
2
4
3
HL and HL could be ascribed to the phenyl group at the α
position which prevents the O ؒ ؒ ؒ O separation of the anionic
form from widening to fit to the desired configuration in the
indium complex owing to the steric repulsion between the
α-phenyl group and the terminal groups.
1
3
46
diagonal O ؒ ؒ ؒ O for aluminium complexes is significantly short,
meaning that three ligands are forced to be packed into a small
space. As seen in Table 4, the M᎐O distances for 1 and 2 are
shorter than those for 3 and 4, respectively. The shorter M᎐O is
brought about by the terminal phenyl group. This corresponds
with the fact that metal ions are usually better extracted in the
Interligand contact in â-diketonates of Al and In
The intracomplex interligand contacts should have great influ-
5
3
1 45
3ϩ
ence on the stability of aluminium β-diketonates, which could
order HL > HL > HL , except for Al where the effect of
the interligand contact is significant. The shorter M᎐O in 1 than
in 3 would enhance the interligand contact.
3ϩ
3ϩ
be deduced from the extraction order of Al and In with β-
1,3,42
diketones.
In general, organic ligands including β-diketones
form more stable complexes with metal ions having smaller
ionic radii when the valency is the same. Consequently, the ex-
traction constants for the smaller metal ions are usually higher
than those for the larger ones. However, in the extraction of
Indium β-diketonates are characterized by their long O ؒ ؒ ؒ O
1
distances. The O ؒ ؒ ؒ O distance in InL 3 is the longest among
1
3
the tervalent metal acetylacetonates, while that in AlL is one
46
of the shortest. Such long O ؒ ؒ ؒ O distances would be intrinsic
3
838
J. Chem. Soc., Dalton Trans., 1997, Pages 3835–3840

View Article Online
5
5
1
Table 4 Mean dimensions (Å or Њ) in AlL 1, InL 2, AlL 3 and
3
3
3
1
InL 3
4
a
b
1
2
3
4
M᎐O
Bite size (O ؒ ؒ ؒ O)
Bite angle
1.874(1)
2.658(2)
90.30(4)
3.747
2.122(1)
2.872(2)
85.2(1)
4.234
1.892(6)
2.726(4)
90.30(2)
3.783
2.132(2)
2.904(1)
85.53(1)
4.255
Diagonal O ؒ ؒ ؒ O
a
b
Data taken from ref. 43. Data taken from ref. 44.
Ϫ1
Table 5 Steric energy change (kJ mol ) for complex formation
5
1
3ϩ
3ϩ
7
8
9
∆
U(ML )
∆U(ML )
∆(∆U)
Fig. 3 Extraction of Al and In with HL , HL and HL . [HA]
o
=
3
3
Ϫ3
Ϫ3
3
ϩ
0.1 mol dm in benzene, [NaClO ] = 0.1 mol dm
4
Al
In
Ϫ372.0
Ϫ368.4
Ϫ277.0
Ϫ214.1
Ϫ95.0
Ϫ154.3
3ϩ
8
3ϩ
with HL shifted to the alkaline region, while that of In
scarcely changed. It can be concluded that the reduced extrac-
tion was brought about by the interligand contact which does
for indium complexes. The enlargement of the O ؒ ؒ ؒ O distance
could be disturbed by the α-substituent.
3ϩ
not work for In . When a methyl group was introduced at the α
8
9
position of HL to get HL the O ؒ ؒ ؒ O distance and pK are
a
Molecular mechanics calculations
26
reduced. Owing to the lower acidity and the enhanced inter-
ligand contact caused by the α-methyl group as in the case of
HL and HL , the extraction constant for Al decreased; how-
The interligand contact is quantitatively elucidated by molecu-
lar mechanics calculations estimating the change in steric
energy (strain energy) that arises from the complexation of one
metal trivalent atom with three ligands to yield a tris complex.
2
4
3ϩ
ever, the influence of the narrowed O ؒ ؒ ؒ O distance is marked
3ϩ
3ϩ
for the extraction of In . The extraction of In drastically
decreased; the shift of the pH value being 2.47; consequently,
The change in steric energy on MA complex formation (∆U)
1/2
3
3ϩ
3ϩ
Al was extracted much better then In resulting in a facili-
tated separation.
was assessed on the basis of equations (5) and (6) where UMA3,
3
ϩ
Ϫ
M
ϩ 3A
MA3
(5)
(6)
Conclusion
∆
U = UMA3 Ϫ UM3ϩ Ϫ 3UA
Ϫ
From our results so far there are two factors governing the
3ϩ
3ϩ
complexation of Al and In with β-diketones: the distance
between the two donating oxygens and the interligand inter-
action, and their balance should decide the stability of each
complex, that is, the extraction order, as well as the separation
UM3ϩ and UA
Ϫ
represent the steric energy of the complex, metal
ion and ligand, respectively. For simplicity, the value of UM3ϩ
was assumed to be zero; UMA3 and UA were obtained by opti-
mizing the crystal structures in the MM2 force field.
Table 5 summarizes the increase in the steric energy of MA3
complex formation, ∆U(MA ). Since U 3ϩ is assumed to be 0
for all metals, the comparison of the absolute values of
U(MA ) is meaningless. Therefore, we noted the difference
Ϫ
3ϩ
3ϩ
of Al and In . This observation may contribute to the basic
knowledge of organic ligands, especially on the concepts of
their complexation with metals, and confirms our suggestion
for a perspective strategy for designing novel ligands of high
3
M
∆
3
26
1
5
3
selectivity from well known typical ones.
between ∆U(ML ) and ∆U(ML ) for each metal, equation (7).
3
5
3
1
3
∆
(∆U) = ∆U(ML ) Ϫ ∆U(ML )
(7)
Acknowledgements
This research is partly supported by a Grant-in-Aid for
Scientific Research (07640803) from the Ministry of Education,
Science, and Culture of Japan. Computation time was provided
by the Supercomputer Laboratory, Institute for Chemical
Research, Kyoto University.
The term UM3ϩ is cancelled in ∆(∆U), so that it can be com-
3ϩ
pared between metals. As can be seen in Table 5, ∆(∆U) of Al
Ϫ1
3ϩ
is 59.3 kJ mol larger than that for In ; the steric energy
1
5
change in complexation from HL to HL is more drastic for
3ϩ
3ϩ
5
Al than that for In due to the bulkiness of HL and the
3ϩ
small size of Al . This confirms our postulation that the com-
3ϩ
3ϩ
1
plexation of Al and In with HL follows the general concep-
References
5
tion, while those with HL and other bulky β-diketones do not.
1
T. Sekine and Y. Hasegawa, Solvent Extraction Chemistry, Marcel
Dekker, New York, 1977.
In the present calculations solvation was not taken into
account and the electron distribution was simplified. Never-
theless, it is very probable that the above steric energy change
dominates the stability in the real system.
2 D. N. Solokov, Russ. Chem. Rev., 1988, 57, 956.
3 Principles and Practices of Solvent Extraction, eds. J. Rydberg,
C. Musikas and G. R. Choppin, Marcel Dekker, New York, 1992.
4
K. E. Laintz, C. M. Wai, C. R. Yonker and R. D. Smith, J. Supercrit.
Fluids, 1991, 4, 194.
K. E. Laintz, C. M. Wai, C. R. Yonker and R. D. Smith, Anal.
Chem., 1992, 64, 2875.
Simultaneous effects of the bite size and the interligand contact
Finally, the effects of the O ؒ ؒ ؒ O distance and the interligand
5
6
7
8
Y. Lin and C. M. Wai, Anal. Chem., 1994, 66, 1971.
J. Wang and W. D. Marshall, Anal. Chem., 1994, 66, 1658.
R. M’Hamdi, J. R. Bocquet, K. Chhor and C. Pommier, J. Supercrit.
Fluids, 1991, 4, 55.
3ϩ
3ϩ
contact on the solvent extraction of Al and In have been
7
8
9
confirmed again using HL , HL and HL . The results are
shown in Fig. 3 and Table 2. Aluminium was better extracted
3ϩ
7
7
than In with HL ; the O ؒ ؒ ؒ O distance of HL is 2.50 Å
according to MNDO/H, and no significant interligand contact
seems to take place among the terminal methyl and trifluoro-
methyl groups. When the methyl group is replaced by the bulky
9 B. N. Hansen, B. M. Hybertson, R. M. Barkley and R. E. Sievers,
Chem. Mater., 1992, 4, 748.
0 W. Lin, T. H. Warren, R. G. Nuzzo and G. S. Girolami, J. Am.
Chem. Soc., 1993, 115, 1164.
1 I. Baxter, S. R. Drake, M. B. Hursthouse, K. M. A. Malik,
J. McAleese, D. J. Otway and J. C. Plakatouras, Inorg. Chem., 1995,
34, 1384.
1
1
8
phenyl group to obtain HL , its O ؒ ؒ ؒ O distance (2.52 Å) and
7
3ϩ
pK are still similar to those of HL . The extraction of Al
a
J. Chem. Soc., Dalton Trans., 1997, Pages 3835–3840
3839

View Article Online
1
1
2 J. Emsley, Struct. Bonding (Berlin), 1984, 57, 147.
Sheldrick, C. Kruger and R. Goddard, Oxford University Press,
Oxford, 1985, p. 175.
3 G. Gilli, F. Bellucci, V. Ferretti and V. Bertolasi, J. Am. Chem. Soc.,
1
989, 111, 1023.
33 P. T. Beurskens, G. Admiraal, G. Beurskens, W. P. Bosman,
S. Garcia-Granda, R. O. Gould, J. M. M. Smits and C. Smykalla,
The DIRDIF program system, Technical Report of the Crystallo-
graphy Laboratory, University of Nijmegen, 1992.
34 TEXSAN, Molecular Structure Corporation, Houston, TX, 1985.
35 M. J. S. Dewar, E. G. Zeobisch, E. F. Healy and J. J. P. Stewart,
J. Am. Chem. Soc., 1985, 107, 3902.
36 J. J. P. Stewart, J. Comput. Chem., 1989, 10, 209.
37 R. M. Silverstein, G. C. Bassler and T. C. Morrill, Spectrometric
Identification of Organic Compounds, Wiley, New York, 1981.
38 J. Emsley and Y. Y. L. Ma, J. Mol. Struct., 1988, 178, 297.
39 J. Emsley, J. N. Freeman, A. P. Bates and B. M. Hursthouse,
J. Chem. Soc., Perkin Trans. 1, 1988, 297.
40 H. M. N. H. Irving and U. S. Mannot, J. Inorg. Nucl. Chem., 1968,
30, 1215.
41 M. Eigen, Pure Appl. Chem., 1963, 6, 105.
42 J. Stary, The Solvent Extraction of Metal Chelates, Pergamon,
London, 1964.
43 P. K. Hon and C. E. Pfluger, J. Coord. Chem., 1973, 3, 67.
44 G. J. Palenik and K. R. Dymock, Acta. Crystallogr., Sect. B, 1980,
36, 2059.
1
1
1
4 A. Roy and K. Nag, J. Inorg. Nucl. Chem., 1978, 40, 331.
5 S. Umetani and H. Freiser, Inorg. Chem., 1987, 26, 3179.
6 Q. T. H. Le, S. Umetani, H. Takahara and M. Matsui, Anal. Chim.
Acta, 1993, 272, 293.
1
1
1
7 S. Umetani, Y. Kawase, H. Takahara, Q. T. H. Le and M. Matsui,
J. Chem. Soc., Chem. Commun., 1993, 78.
8 S. Umetani, Y. Kawase, T. H. Q. Le and M. Matsui, unpublished
work.
9 T. H. Q. Le, S. Umetani, M. Suzuki and M. Matsui, J. Chem. Soc.,
Chem. Commun., 1995, 2271.
2
2
2
2
2
0 R. D. Shannon, Acta. Crystallogr., Sect. A, 1976, 32, 751.
1 C. R. Hauser, F. W. Swamer and J. T. Adams, Org. React., 1954, 59.
2 J. C. Reid and M. Calvin, J. Am. Chem. Soc., 1950, 72, 2948.
3 E. W. Berg and J. T. Truemper, J. Am. Chem. Soc., 1960, 64, 487.
4 E. G. Zaitseva, I. A. Baidina, P. A. Stabnikov, S. V. Borisov and
I. K. Igumenov, Zh. Strukt. Khim., 1990, 31, 184.
2
5 L. B. Barkley and R. Levine, J. Am. Chem. Soc., 1953, 75, 2059.
6 T. H. Q. Le, S. Umetani, M. Suzuki and M. Matsui, J. Chem. Soc.,
Dalton Trans., 1997, 643.
2
27 K. Y. Burstein and A. N. Isaev, Theor. Chim. Acta, 1984, 64, 397.
28 CACHE, Version 3.8, CAChe Scientific, Oxford, 1995.
29 N. L. Allinger, J. Am. Chem. Soc., 1977, 99, 8127.
30 U. Burkert and N. L. Allinger, Molecular Mechanics, The American
Chemical Society, Washington, DC, 1982.
45 H. Soling, Acta Chem. Scand., Ser. A, 1976, 30, 163.
46 E. C. Lingafelter and R. L. Braun, J. Am. Chem. Soc., 1966, 88,
2951.
3
3
1 N. Walker and D. Stuart, Acta Crystallogr., Sect. A, 1983, 39, 158.
2 G. M. Sheldrick, in Crystallographic Computing 3, eds. G. M.
Received 7th April 1997; Paper 7/02377D
3
840
J. Chem. Soc., Dalton Trans., 1997, Pages 3835–3840