Synthesis, properties and solution speciation of lanthanide chloride complexes of triphenylphosphine oxide

Article abstract of DOI:10.1016/j.ica.2003.09.015

The reaction of Ph₃PO with LnCl₃·nH ₂O (Ln = La-Lu ≠ Pm) in a 3.5:1 ratio in acetone produces [LnCl₃(Ph₃PO)₃], whilst from a 6:1 ratio in ethanol the products are [LnCl₂(Ph₃PO)₄] Cl·n(solvate). In the presence of [NH₄][PF₆] in ethanol solution, [LnCl₂(Ph₃PO)₄]PF ₆ can be isolated. The last complexes are stable in solution but the [LnCl₃(Ph₃PO)₃] and [LnCl₂(Ph ₃PO)₄]Cl partially interconvert in non-coordinating solvents, the neutral species being preferred by the lighter lanthanides, the cationic tetrakis complexes becoming more favoured towards the end of the series. The complexes have been characterised in the solid state by analysis and IR spectroscopy and in solution by ³¹P{¹H} NMR spectroscopy and conductance measurements. The crystal structures of trans-[LnCl₂(Ph₃PO)₄]Cl·nEtOH (Ln = Tb or Yb) and mer-[LnCl₃(Ph₃PO)₃]·0. 5Me₂CO (Ln = La or Ce) are reported and discussed.

Full text of DOI:10.1016/j.ica.2003.09.015

Inorganica Chimica Acta 357 (2004) 1083–1091
www.elsevier.com/locate/ica
Synthesis, properties and solution speciation of lanthanide
chloride complexes of triphenylphosphine oxide
*
Mikael J. Glazier, William Levason , Melissa L. Matthews,
Pamela L. Thornton, Michael Webster
Department of Chemistry, University of Southampton, Southampton SO17 1BJ, UK
Received 11 June 2003; accepted 17 September 2003
Abstract
The reaction of Ph3PO with LnCl3 ꢀ nH2O (Ln ¼ La–Lu ¼ Pm) in a 3.5:1 ratio in acetone produces [LnCl3(Ph3PO)3], whilst from
a 6:1 ratio in ethanol the products are [LnCl2(Ph3PO)4]Cl ꢀ n(solvate). In the presence of [NH4][PF6] in ethanol solution,
[LnCl2(Ph3PO)4]PF6 can be isolated. The last complexes are stable in solution but the [LnCl3(Ph3PO)3] and [LnCl2(Ph3PO)4]Cl
partially interconvert in non-coordinating solvents, the neutral species being preferred by the lighter lanthanides, the cationic tet-
rakis complexes becoming more favoured towards the end of the series. The complexes have been characterised in the solid state by
analysis and IR spectroscopy and in solution by 31P{1H} NMR spectroscopy and conductance measurements. The crystal structures
of trans-[LnCl2(Ph3PO)4]Cl ꢀ nEtOH (Ln ¼ Tb or Yb) and mer-[LnCl3(Ph3PO)3] ꢀ 0.5Me2CO (Ln ¼ La or Ce) are reported and
discussed.
Ó 2003 Elsevier B.V. All rights reserved.
Keywords: Lanthanides; Phosphine oxides; Crystal structures
1. Introduction
whilst for the later elements mixtures of [Ln(NO3)3
(R3PO)3] and [Ln(NO3)2(R3PO)4]þ are observed.
Phosphine oxides have often been chosen as suitable
ligands for complexing the oxophilic lanthanide(3+) ions
and have been used in solvent extraction separation pro-
cesses [1]. The most thoroughly studied systems are the
lanthanide nitrates with Ph3PO, and Ph2MePO where
complexes of types [Ln(NO3)3(R3PO)3] (Ln ¼ La–Lu,
¼ Pm), [Ln(NO3)3(Ph3PO)4] (Ln ¼ La, Ce, Pr, Nd) and
[Ln(NO3)2(R3PO)4]þ, (Ln ¼ Tb–Lu) have been charac-
terised [2–6]. X-Ray crystallographic studies have shown
that a range of 8-, 9-, and 10-coordinate structures are
present [3–6]. Solution speciation in organic solvents (as
established by 31P{1H} NMR and conductance studies)
also varies along this series of complexes with
[Ln(NO3)3(R3PO)3] usually present for Ln ¼ La–Gd,
For lanthanide chloride systems, the literature is based
upon a few examples, is inconsistent and contains little
data [1,7,8]. These reports indicate [Ln(R3PO)xCl3] (x ¼ 3
or 4) often as solvates, can be isolated and are non-elec-
trolytes in solution implying 6- and 7-coordinate lantha-
nide, respectively. In the presence of Lewis acids such as
FeCl3 or CuCl2, [LnCl(Ph3PO)5]2þ form [8]. Recent
studies of the HMPA (OP(NMe2)3) complexes have es-
tablished 6-coordinate [LnCl3(HMPA)3] and crystal
structures have been reported for fac-[SmCl3(HMPA)3]
and mer-[LnCl3(HMPA)3] (Ln ¼ Yb or Dy) [9–11]. Re-
cent detailed studies of the reactions of yttrium halides
(X ¼ Cl, Br or I) with Ph3PO, Ph2MePO and Me3PO have
revealed a range of complexes can be isolated depending
upon the R3PO and the reaction conditions used includ-
ing [YX2(Ph3PO)4]þ, [YX3(Ph2MePO)3], [YCl(Ph3
PO)5]2þ, [YCl2(Ph2MePO)4]þ, and [Y(Me3PO)6]3þ [12].
We have also reported that Me3PO, LnCl3 ꢀ nH2O and
* Corresponding author. Tel.: +44-23-80-593792; fax: +44-23-
80593781.
E-mail address: wxl@soton.ac.uk (W. Levason).
0020-1693/$ - see front matter Ó 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.ica.2003.09.015

1084
M.J. Glazier et al. / Inorganica Chimica Acta 357 (2004) 1083–1091
NH4PF6 react in methanol solution to produce octahe-
dral [Ln(Me3PO)6][PF6]3 [13].
are observed at 844 and 560 cmꢂ1 [15]. In 10ꢂ3 mol
dmꢂ3 CH2Cl2 solutions they have conductivities in the
range 15–17 Xꢂ1 cm2 molꢂ1 which are interpreted as 1:1
1
electrolytes (the usual range for 1:1 electrolytes in this
2. Results and discussion
solvent is 20–25 Xꢂ1 cm2 molꢂ1 [4,5,12], the lower values
here being attributable to the very large ions). Addition
of Ph3PO to these solutions does not increase the con-
ductance, showing that there is no tendency to displace
Based upon analogies with the corresponding nitrates
[3–6], the limited literature data on chloride systems [1,7]
and preliminary experiments, we synthesised two series of
complexes [LnCl3(Ph3PO)3] and [LnCl2(Ph3PO)4]PF6.
The former were obtained in modest yield from reaction
of LnCl3 ꢀ nH2O with 3.5 equivalents of Ph3PO in acetone
and the latter from reaction of LnCl3 ꢀ nH2O with Ph3PO
and NH4PF6 in a 1:4.5:1 molar ratio in absolute ethanol.
Analytical and spectroscopic data showed that both series
of complexes were prone to retain solvent and prolonged
drying in high vacuum is necessary, although in some
cases (vide infra) even then lattice solvent is retained. The
[LnCl3(Ph3PO)3] complexes cannot be recrystallised from
chlorocarbons or acetone unless some Ph3PO is added, in
its absence the products have markedly lower (and vari-
able) C, H content and appear to be mixtures. Acetone or
ethanol solutions of [LnCl3(Ph3PO)3] left to evaporate in
the open laboratory hydrolyse with displacement of the
Ph3PO, and in the extreme, the products are Ph3PO and
hydrated lanthanide chlorides, one example of which
[Ce2Cl2(H2O)14]Cl4 was characterised by an X-ray study
[14]. The reactions of LnCl3 ꢀ nH2O with Ph3PO in abso-
lute ethanol are less clear cut, it being possible to isolate
[LnCl3(Ph3PO)3] and/or [LnCl2(Ph3PO)4]Cl ꢀ n(solv) de-
pending on the Ln involved and the conditions. The
[LnCl2(Ph3PO)4]Cl ꢀ n(solv) separate from ethanol solu-
tion containing a 6:1 molar ratio of Ph3PO:LnCl3 ꢀ nH2O.
Although the analytical data (Table 1) are in reasonable
agreement with formulae such as [LnCl2(Ph3PO)4]
Cl ꢀ 2H2O or [LnCl2(Ph3PO)4]Cl ꢀ 3EtOH, 1H NMR
spectroscopy suggests that a mixture of water and ethanol
is present which probably varies from sample to sample.
The solvate water/ethanol is not removed in vacuo at
room temperature and although it is lost on heating, this
was generally avoided since it may be accompanied by
some decomposition of the cations.
1
further chlorides in this way. The H NMR spectra are
uninformative showing the expected phenyl protons,
broadened and shifted in the paramagnetic complexes.
The 31P{1H} NMR spectra (Table 2) show single reso-
nances for the cations, in the cases of the paramagnetic
LnÕs broadened and shifted in generally similar ways to
those reported for the nitrate systems [4,5], and quali-
tatively consistent with theoretical predictions [16]. Only
[GdCl2(Ph3PO)4]PF6 fails to exhibit a resonance pre-
sumably due to fast relaxation by the gadolinium ion.
For the elements Eu–Lu addition of Ph3PO to the so-
lutions results in a new resonance at d 28 due to Ph3PO
and leaves the resonance of the cation unchanged
showing exchange between free and coordinated phos-
phine oxides are slow on the NMR timescale. For Ce–
Sm separate resonances for the cation and Ph3PO are
also observed but these are broad at 295 K indicating
some slow exchange, and for [LaCl2(Ph3PO)4]PF6 only
a single resonance is observed at 295 K which on cooling
below 230 K splits showing exchange is fast at ambient
temperatures. The observation of a single sharp 31P{1H}
resonance for the diamagnetic La and Lu complexes
indicates that all these cations are likely to be trans
isomers, although the broadening caused by the lan-
thanide paramagnetism in most other cases is sufficient
to prevent resolution of closely spaced resonances (in the
diamagnetic Y complexes for example, phosphine oxides
in geometrically different environments or in different
complexes had 31P{1H} resonances which covered only
a 1–2 ppm range [12]). Addition of anhydrous [Bu4N]Cl
to solutions of [LnCl2(Ph3PO)4]þ in CH2Cl2 resulted in
new resonances identified as free Ph3PO and [LnCl3(Ph3
PO)3] showing the expected equilibria are established,
although for the later lanthanides, even with excess Clꢂ
the tetrakis cation is still the major species present. If
‘‘wet’’ [Bu4N]Cl was added to these solutions, some
Ph3PO was displaced but new 31P{1H} resonances were
also generated which presumably result from water en-
tering the coordination sphere of the lanthanide.
2.1. [LnCl2(Ph3PO)4]PF6
These were obtained as powders directly from the
preparations, with analytical and IR data (weak features
at ꢁ3300 and 1650 cmꢂ1) suggesting that the first
members of the series are hydrates (Table 1). The
complexes are readily soluble in chlorocarbons, and
poorly soluble in alcohols or acetone. The IR spectra
show strong rather broad m(PO) modes in the range
1139–1150 cmꢂ1 which increase along the series La–Lu
(Table 2) and which compare to 1195 cmꢂ1 in Ph3PO
[12]. Weak features in the range 225–240 cmꢂ1 are as-
signed as m(Ln–Cl), and the m3 and m4 modes of [PF6]ꢂ
1
Although dichloromethane is not commonly used for conductance
measurements, we have used it successfully for other series of
phosphine oxide complexes. In the present case it was chosen to
permit direct comparisons of the conductances with the NMR data
obtained in this solvent. The lability and ease with which phosphine
oxides are lost from these complexes makes the choice of solvent
important.

M.J. Glazier et al. / Inorganica Chimica Acta 357 (2004) 1083–1091
1085
Table 1
Analytical and physical data
a
Complex
%C
%H
K (Xꢂ1 cm2 molꢂ1b
)
c
d
[LaCl3(Ph3PO)3]
[CeCl3(Ph3PO)3]
[PrCl3(Ph3PO)3]
[NdCl3(Ph3PO)3]
[SmCl3(Ph3PO)3]
[EuCl3(Ph3PO)3]
[GdCl3(Ph3PO)3]
[TbCl3(Ph3PO)3]
[DyCl3(Ph3PO)3]
[HoCl3(Ph3PO)3]
[ErCl3(Ph3PO)3]
[TmCl3(Ph3PO)3]
[YbCl3(Ph3PO)3]
[LuCl3(Ph3PO)3]
59.9(60.1)
59.5(60.0)
60.0(59.9)
60.1(59.8)
58.8(59.4)
59.7(59.3)
59.6(59.0)
58.5(59.0)
58.5(58.8)
58.8(58.6)
58.7(58.5)
58.7(58.4)
58.1(58.2)
57.7(58.1)
4.4(4.5)
4.4(4.2)
4.2(4.2)
4.0(4.2)
4.2(4.2)
4.3(4.2)
4.0(4.1)
4.3(4.1)
4.2(4.1)
4.3(4.1)
3.9(4.1)
4.0(4.1)
4.1(4.1)
3.8(4.1)
1
1
2
2
2
2
4
6
6
6
7
6
8
8
3
4
6
7
7
8
11
14
16
16
18
20
20
21
[LaCl2(Ph3PO)4]PF6 ꢀ H2O
[CeCl2(Ph3PO)4]PF6 ꢀ H2O
[PrCl2(Ph3PO)4]PF6 ꢀ H2O
[NdCl2(Ph3PO)4]PF6 ꢀ H2O
[SmCl2(Ph3PO)4]PF6
[EuCl2(Ph3PO)4]PF6
[GdCl2(Ph3PO)4]PF6
[TbCl2(Ph3PO)4]PF6
[DyCl2(Ph3PO)4]PF6
[HoCl2(Ph3PO)4]PF6
[ErCl2(Ph3PO)4]PF6
57.8(58.2)
57.5(58.1)
58.5(58.1)
57.1(58.0)
57.8(58.5)
57.8(58.4)
57.6(58.2)
57.6(58.1)
58.3(58.0)
57.7(57.9)
57.8(57.8)
57.2(57.7)
57.2(57.6)
57.2(57.5)
3.9(4.2)
3.8(4.2)
3.9(4.2)
3.8(4.2)
3.3(4.0)
3.9(4.1)
3.8(4.1)
3.8(4.0)
3.9(4.1)
3.7(4.0)
3.5(4.0)
3.7(4.0)
3.9(4.0)
3.9(4.0)
15
16
16
16
17
17
16
17
17
17
17
17
16
18
[TmCl2(Ph3PO)4]PF6
[YbCl2(Ph3PO)4]PF6
[LuCl2(Ph3PO)4]PF6
[LaCl2(Ph3PO)4]Cl ꢀ 2H2Oe
[CeCl2(Ph3PO)4]Cl ꢀ 2H2O
[PrCl2(Ph3PO)4]Cl ꢀ 2H2O
[NdCl2(Ph3PO)4]Cl ꢀ 2H2O
[SmCl2(Ph3PO)4]Cl ꢀ 2H2O
[EuCl2(Ph3PO)4]Cl ꢀ 2H2O
[GdCl2(Ph3PO)4]Cl ꢀ 2H2O
[TbCl2(Ph3PO)4]Cl ꢀ 2H2O
[DyCl2(Ph3PO)4]Cl ꢀ 2H2O
[HoCl2(Ph3PO)4]Cl ꢀ 2H2O
[ErCl2(Ph3PO)4]Cl ꢀ 2H2O
[TmCl2(Ph3PO)4]Cl ꢀ 2H2O
[YbCl2(Ph3PO)4]Cl ꢀ 2H2O
[LuCl2(Ph3PO)4]Cl ꢀ 2H2O
61.6(62.0)
62.0(62.0)
62.2(61.9)
61.9(61.8)
61.3(61.5)
61.8(61.4)
61.5(61.2)
61.3(61.1)
61.1(61.0)
61.2(60.8)
60.8(60.8)
60.4(60.6)
60.3(60.5)
60.2(60.5)
4.5(4.6)
4.5(4.6)
4.6(4.6)
4.5(4.6)
4.1(4.6)
4.7(4.5)
4.1(4.6)
4.5(4.5)
4.2(4.5)
4.4(4.5)
4.6(4.6)
4.4(4.5)
4.4(4.5)
4.5(4.5)
3
3
6
6
3
8
3
8
4
7
5
9
7
15
16
17
19
20
20
22
23
7
8
8
8
10
14
15
a Calculated values in parenthesis.
b 10ꢂ3 mol dmꢂ3 CH2Cl2 solutions.
c Freshly prepared 10ꢂ3 mol dmꢂ3 solution.
d Solution in footnote c containing approximately 3 equivalents of Ph3PO.
e Calculated data refer to dihydrates but in most samples 1H NMR spectra show a mixture of water and ethanol present (see discussion in text). The
effect on the analyses is small due to the large molecular weights of the cations.
2.2. [LnCl3(Ph3PO)3]
more soluble in chlorocarbons. In 10ꢂ3 mol dmꢂ3
CH2Cl2 solutions they have small conductivities which
increase along the series (Table 1) but at 1–8 Xꢂ1 cm2
molꢂ1, are interpreted as non-conductors, the NMR
studies (below) providing an explanation for the ob-
served values. The conductances rise on addition of
excess Ph3PO, those of complexes in the latter half of the
series progressively approaching 1:1 electrolytes. The IR
These complexes are best obtained by reaction of
LnCl3 ꢀ nH2O with Ph3PO in a 1:3.5 molar ratio in hot
acetone, and providing the solutions are not too con-
centrated on mixing, materials of adequate analytical
purity separate directly from the solutions. The com-
plexes are poorly soluble in acetone or alcohols, and

1086
M.J. Glazier et al. / Inorganica Chimica Acta 357 (2004) 1083–1091
Table 2
Spectroscopic data
Complex
m(PO) (cmꢂ1a
)
m(Ln–Cl) (cmꢂ1a
)
d(31P)b
[LaCl3(Ph3PO)3]
[CeCl3(Ph3PO)3]
[PrCl3(Ph3PO)3]
[NdCl3(Ph3PO)3]
[SmCl3(Ph3PO)3]
[EuCl3(Ph3PO)3]
[GdCl3(Ph3PO)3]
[TbCl3(Ph3PO)3]
[DyCl3(Ph3PO)3]
[HoCl3(Ph3PO)3]
[ErCl3(Ph3PO)3]
[TmCl3(Ph3PO)3]
[YbCl3(Ph3PO)3]
[LuCl3(Ph3PO)3]
1147, 1171
1146, 1172
1145, 1175
1147, 1176
1150, 1176
1150, 1184
1152, 1186
1153, 1184
1154, 1188,
1153, 1188
1154, 1188
1155, 1189
1155, 1187
1156, 1190
220
225
35.8(10)
78(1650)
225, 237sh
228, 235sh
229
130(700)
136(500)
29(150)
231
230
)70(600)
n.o.
228
228
)95(800), )129*
n.o., 17*
229
239
49(1800), )6*
)91(650), )87*
)38(530), )20*
21(650), 33*
36.0(10), 37.2*
236
240
240
[LaCl2(Ph3PO)4]PF6 ꢀ H2O
[CeCl2(Ph3PO)4]PF6 ꢀ H2O
[PrCl2(Ph3PO)4]PF6 ꢀ H2O
[NdCl2(Ph3PO)4]PF6 ꢀ H2O
[SmCl2(Ph3PO)4]PF6
[EuCl2(Ph3PO)4]PF6
[GdCl2(Ph3PO)4]PF6
[TbCl2(Ph3PO)4]PF6
[DyCl2(Ph3PO)4]PF6
[HoCl2(Ph3PO)4]PF6
[ErCl2(Ph3PO)4]PF6
1139
1141
1140
1142
1142
1143
1145
1146
1146
1148
1148
1149
1150
1150
226
235
237
237
236
235
237
238
238
241
240
240
241
240
37.8(8)
80(100)
131(130)
138(90)
30 (100)
)75(400)
n.o.
)130(450)
17(860)
)6 (870)
)87(340)
)20(100)
33(80)
[TmCl2(Ph3PO)4]PF6
[YbCl2(Ph3PO)4]PF6
[LuCl2(Ph3PO)4]PF6
37.3(10)
[LaCl2(Ph3PO)4]Cl ꢀ 2H2Oc
[CeCl2(Ph3PO)4]Cl ꢀ 2H2O
[PrCl2(Ph3PO)4]Cl ꢀ 2H2O
[NdCl2(Ph3PO)4]Cl ꢀ 2H2O
[SmCl2(Ph3PO)4]Cl ꢀ 2H2O
[EuCl2(Ph3PO)4]Cl ꢀ 2H2O
[GdCl2(Ph3PO)4]Cl ꢀ 2H2O
[TbCl2(Ph3PO)4]Cl ꢀ 2H2O
[DyCl2(Ph3PO)4]Cl ꢀ 2H2O
[HoCl2(Ph3PO)4]Cl ꢀ 2H2O
[ErCl2(Ph3PO)4]Cl ꢀ 2H2O
[TmCl2(Ph3PO)4]Cl ꢀ 2H2O
[YbCl2(Ph3PO)4]Cl ꢀ 2H2O
[LuCl2(Ph3PO)4]Cl ꢀ 2H2O
1140
1141
1141
1141
1143
1144
1145
1146
1146
1147
1147
1150
1150
1150
224
237
237
240
229
236
238
240
239
238
239
240
240
242
33.5(br), see text
80(120), 78**
131(sh), 129**
138(100), 136sh**
30(110), 29**
)74(450), )70**
n.o.
)130(500), )94**
17(900), n.o.**
)6(550), 49**
-87(360), )91**
-20(110), )38**
33(90), 21**
38.0(15), 36.5**
a Nujol mulls.
b CH2Cl2 solution at 295 K relative to external H3PO4. The values in parenthesis are approximate line widths at half height (W 1=2 (Hz)). n.o. – not
observed. Resonances marked * are due to [LnCl2(Ph3PO)4]þ formed in solution by rearrangement (see text), resonances marked ** are similarly due
to [LnCl3(Ph3PO)3] and all the latter are accompanied by free Ph3PO (not listed).
c See footnote e in Table 1.
spectra show broad and sometimes asymmetric t(PO)
modes in the range 1145–1156 cmꢂ1 with a shoulder at
1170–1190 cmꢂ1, the major feature(s) slightly higher for
each specific Ln than observed in the corresponding
[LnCl2(Ph3PO)4]þ, and broad weak bands in the range
220–240 cmꢂ1 are assigned as m(Ln–Cl). In neither case
is there sufficient resolved splitting of m(PO) or m(Ln–Cl)
to clearly distinguish mer (three IR active modes) from
fac (two IR active modes) isomers. In a number of cases
crystals of the complexes separated on standing from the
synthesis solutions, but as observed in related systems
[3–6,12] despite good visual appearance these often gave
rather mediocre quality X-ray structural data, possible
causes including fractional occupancies by solvent
molecules and some disorder in the ligand phenyl rings.
However, crystal structures were obtained for
[LnCl3(Ph3PO)3] ꢀ 0.5Me2CO (Ln ¼ La and Ce). Both
were found to be 6-coordinate and meridional geomet-
rical isomers, and the structures are isomorphous, al-
though of low quality. Both the Ln residues in the cell
had the same stereochemistry but during the refinement
it became clear that there were disorder problems with a

M.J. Glazier et al. / Inorganica Chimica Acta 357 (2004) 1083–1091
1087
few of the phenyl rings (out of the 18) and the final R1
[La(NO3)3(Ph3PO)x] systems we did not observe any
lanthanum resonances, clearly due to the line-broaden-
ing caused by fast quadrupolar relaxation in the low
symmetry environments, and similarly resonances are
not expected for [LaCl3(Ph3PO)x]. However both
CH2Cl2 and MeCN solutions of [LaCl3(Ph3PO)3]
showed a broad resonance at ca. d 600 (W1=2 ¼ 3000 Hz),
which is in poor agreement with that reported for
[LaCl6]3ꢂ in MeCN d 851 [18]. The resonance is largely
lost on adding excess Ph3PO to the solution, suggesting it
is a lanthanum chlorospecies which is consumed as the
equilibrium shifts.
2
3
values were 0.12(Ce) and 0.13(La). The key points
are the Cl–Ln–Cl (Ce: 91.9(1), 92.7(1), 174.7(1)°. La:
91.9(2), 92.9(2), 174.2(2)°) and the O–Ln–O (Ce: 86.1(3),
91.1(3), 176.9(3)°. La: 85.3(5), 91.8(4), 177.0(4)°) which
establishes the mer isomer. As in the tetrakis species (see
later) the Ln–O–P angles are large (Ce: 164.5(6)–
178.5(6)°. La: 165.1(8)–177.7(9)°).
The 1H NMR spectra are uninformative but the
31P{1H} NMR (Table 2) are more useful, except for
[GdCl3(Ph3PO)3] for which no resonance was observed,
and [DyCl3(Ph3PO)3] which showed only a weak reso-
nance assigned to [DyCl2(Ph3PO)4]þ the absence of the
expected resonances being attributable to fast relaxation.
For the complexes of Tb–Lu, two resonances were present
(Table 2) the narrower one of which could be identified as
due to [LnCl2(Ph3PO)4]þ by comparison with the spectra
obtained from [LnCl2(Ph3PO)4]PF6 and the second
broader signal is assigned to [LnCl3(Ph3PO)3]. The mass
balance requires at least one other lanthanide species to be
formed, and since no other 31P NMR resonances are seen,
this is presumably phosphine oxide free. Possible equi-
libria would include:
Addition of Ph3PO to these [LnCl3(Ph3PO)3] solu-
tions changed the relative intensities of the resonances in
favour of that of the tetrakis complex, confirming their
assignments and the presence of a free Ph3PO resonance
in the majority of examples showed exchange was slow
on the NMR time scale. The relative amount of the tet-
rakis complex present increases along the series which
explains the increasing conductances of their solutions.
For the early members of the series Ce–Nd only a single
31P{1H} NMR resonance of coordinated phosphine
oxide was observed. For these elements the very similar
chemical shifts of the tris and tetrakis species (Table 2)
makes it very difficult to identify small amounts of one in
the presence of the other, but from the very small con-
ductances it seems probably that rearrangement into the
tetrakis complexes is minimal here. The diamagnetic
[LaCl3(Ph3PO)3] shows only a single sharp (W1=2 ¼ 10
Hz) resonance which was unexpected given the structure
of the solid (above) reveals a mer arrangement of Ph3PO
groups. It is not possible from the available data to dis-
tinguish between complete rearrangement into the fac
form, fluxionality or accidental coincidence of the reso-
nances. The last seems most likely, since although the
nitrate systems are usually fluxional [4–6] these are
mostly 9-coordinate, and fluxionality is less likely in 6-
coordination. In contrast to the other complexes, addi-
tion of Ph3PO to a solution of [LaCl3(Ph3PO)3] in
CH2Cl2 showed only a broad singlet indicative of fast
exchange, and cooling the solution below 243 K was
necessary to resolve separate resonances. At 195 K res-
onances attributable to [LaCl3(Ph3PO)3], [LaCl2(Ph3
PO)4]þ, Ph3PO and a new unidentified species with d 33
were present. Since lanthanum is the largest of the lan-
thanides, it is not possible to rule out that this unassigned
species may be the 7-coordinate [LaCl3(Ph3PO)4], al-
though it could also be the second geometric isomer of
[LaCl3(Ph3PO)3].
4½LnCl3ðPh3POÞ ꢃ $ 3½LnCl2ðPh3POÞ ꢃþ þ ½LnClxꢃðxꢂ3Þꢂ
3
4
þ ð6 ꢂ xÞClꢂ:
In an attempt to probe this equilibrium, 139La NMR
spectra were recorded from solutions of [LaCl3(Ph3PO)3]
in CH2Cl2 and MeCN. The 139La nucleus (I ¼ 7=2,
99.9%, N ¼ 14.1 MHz, quadrupole moment 0.20 ꢄ 10ꢂ28
m2) is not particularly suited to such studies, the com-
bination of the substantial quadrupole and the lability of
many complexes often resulting in lines too broad to
observe. Most of the species for which a resonance has
been observed are of high symmetry usually LaL6 type
[17]. We reported the chemical shift of [La(Me3PO)6]3þ
as +116 [13] and those of other LaO6 moieties lie in the
region ca. 0–200 ppm [17]. In previous studies [5] of the
2
[CeCl3(Ph3PO)3] ꢀ 0.5Me2CO. Crystals were isolated from the
preparation. Crystal data: monoclinic; a ¼ 19.529(5); b ¼ 15.5490(10);
3
ꢀ
ꢀ
c ¼ 34.844(4) A; b ¼ 103.738(3)°, V ¼ 10277.9(17) A . Z ¼ 8, RMM ¼
1110.32, T ¼ 120 K. Data collected on
a Nonius Kappa CCD
diffractometer using Mo Ka radiation. Space group P21/n (no. 14)
from systematic absences. The data were rather weak (hI=rðIÞi ¼ 3:5)
with Rint ¼ 0:11, and although the two mer octahedral species in the
ꢀ
asymmetric unit were clearly established (Ce–Cl 2.696(4)–2.755(4) A;
ꢀ
Ce–O 2.344(9)–2.421(11) A) there was disorder in the phenyl rings and
refinement failed to reduce R1 (I > 2rðIÞ) below 0.12.
3
[LaCl3(Ph3PO)3] ꢀ 0.5Me2CO. Crystals were isolated from the
preparation. Crystal data: monoclinic; a ¼ 19.574(5); b ¼ 15.608(4);
3
ꢀ
ꢀ
c ¼ 34.837(8) A; b ¼ 103.890(8)°, V ¼ 10332(4) A . Z ¼ 8, RMM ¼
2.3. [LnCl2(Ph3PO)4]Cl ꢀ n(solv)(solv ¼ H2OorEtOH)
1109.11, T ¼ 120 K. Data collected on
a Nonius Kappa CCD
diffractometer using Mo Ka radiation. Space group P21/n (no. 14)
from systematic absences. The data were rather weak (hI=rðIÞi ¼ 5:0)
with Rint ¼ 0:08, and although the two mer octahedral species in the
These complexes were obtained from hot ethanol
solutions of LnCl3 ꢀ nH2O and 6 mol. equivalents of
Ph3PO, and were always obtained solvated (see above).
As expected the IR spectra were very similar to those of
ꢀ
asymmetric unit were clearly established (La–Cl 2.712(7)–2.776(5) A;
ꢀ
La–O 2.351(14)–2.451(14) A) there was disorder in the phenyl rings
and refinement failed to reduce R1 (I > 2rðIÞ) below 0.13.

1088
M.J. Glazier et al. / Inorganica Chimica Acta 357 (2004) 1083–1091
[LnCl2(Ph3PO)4]PF6, except for the absence of the
[PF6]ꢂ and the presence of solvate molecule vibrations.
The presence of trans pseudooctahedral cations was
confirmed in two cases by X-ray studies on the Tb and
Yb compounds. The crystal structures (see Tables 3 and
4 and Fig. 1) are essentially isomorphous with only
small differences in the anion disorder and extent of
solvation being observed. The cations have no crystal-
lographic symmetry but in both cases the two Ln–Cl
distances are similar as are the Ln–O distances. As ex-
pected the Yb distances are on average shorter than the
ꢀ
Tb distances by approximately 0.05 A. All the bond
angles at Ln are within 6° of the ideallised value of 90°.
The geometry of the triphenylphosphine oxide ligand is
unexceptional and the Ln–O–P angle which is very
variable in phosphine oxide complexes is large, spanning
the range 167–177°. Disordered EtOH molecules were
clearly identified in both structures but it proved difficult
to convincingly associate residual peaks in the electron
density maps with O(H2) molecules.
In contrast to the data on the [LnCl2(Ph3PO)4]PF6
salts, the 31P{1H} NMR spectra from CH2Cl2 solutions
(Table 2) mostly show three species identified as
[LnCl2(Ph3PO)4]þ, [LnCl3(Ph3PO)3], and Ph3PO pres-
ent. For [LaCl2(Ph3PO)4]Cl only a broad resonance at d
34 was present, but on cooling to 200 K this resolved
into resonances at 36 ([LaCl3(Ph3PO)3]), a very weak
Fig. 1. The structure of the cation in [TbCl2(Ph3PO)4]Cl ꢀ 1.55EtOH
showing the atom labelling scheme. Displacement ellipsoids are drawn
at the 50% probability level and the H atoms are omitted for clarity.
The structure of the Yb cation is very similar and the same numbering
scheme has been adopted
shoulder at 37.5 ([LaCl2(Ph3PO)4]þ), Ph3PO, and an
unassigned species with d 33. This data and the very
small conductance of the solution (Table 1) shows that
Table 3
ꢀ
Selected bond lengths (A) and angles (°) for [TbCl2(Ph3PO)4]Cl ꢀ 1.55EtOH
Tb(1)–Cl(1)
Tb(1)–O(1)
Tb(1)–O(2)
P–O
2.633(3)
2.260(6)
Tb(1)–Cl(2)
Tb(1)–O(3)
Tb(1)–O(4)
P–C
2.647(2)
2.252(6)
2.269(6)
1.496(7)–1.519(7)
2.266(7)
1.788(10)–1.820(10)
Cl(1)–Tb(1)–Cl(2)
O(1)–Tb(1)–O(2)
O(1)–Tb(1)–O(3)
O(1)–Tb(1)–O(4)
Cl(1)–Tb(1)–O
Tb(1)–O(1)–P(1)
Tb(1)–O(2)–P(2)
O–P–C
176.23(8)
88.8(2)
176.8(2)
O(2)–Tb(1)–O(3)
O(2)–Tb(1)–O(4)
O(3)–Tb(1)–O(4)
88.2(2)
173.3(2)
92.1(2)
91.0(2)
86.2(2)–91.8(2)
177.5(4)
167.2(4)
Cl(2)–Tb(1)–O
Tb(1)–O(3)–P(3)
Tb(1)–O(4)–P(4)
C–P–C
89.3(2)–96.3(2)
171.3(4)
170.6(4)
108.5(4)–112.7(4)
105.1(5)–109.9(4)
Table 4
Selected bond lengths (A) and angles (°) for [YbCl2(Ph3PO)4]Cl ꢀ 2.25EtOH
ꢀ
Yb(1)–Cl(1)
Yb(1)–O(1)
Yb(1)–O(2)
P–O
2.590(3)
2.205(8)
Yb(1)–Cl(2)
Yb(1)–O(3)
Yb(1)–O(4)
P–C
2.585(2)
2.188(8)
2.216(8)
1.488(8)–1.512(8)
2.217(8)
1.777(13)–1.826(14)
Cl(1)–Yb(1)–Cl(2)
O(1)–Yb(1)–O(2)
O(1)–Yb(1)–O(3)
O(1)–Yb(1)–O(4)
Cl(1)–Yb(1)–O
Yb(1)–O(1)–P(1)
Yb(1)–O(2)–P(2)
O–P–C
177.15(10)
91.9(3)
177.6(3)
O(2)–Yb(1)–O(3)
O(2)–Yb(1)–O(4)
O(3)–Yb(1)–O(4)
90.2(3)
173.8(3)
89.8(3)
88.2(3)
86.6(2)–91.7(2)
171.0(5)
170.4(5)
Cl(2)–Yb(1)–O
Yb(1)–O(3)–P(3)
Yb(1)–O(4)–P(4)
C–P–C
89.3(2)–95.5(2)
175.1(5)
167.4(5)
108.5(5)–113.7(6)
105.9(6)–109.5(7)

M.J. Glazier et al. / Inorganica Chimica Acta 357 (2004) 1083–1091
1089
decomposition to form [LaCl3(Ph3PO)3] is close to
complete. For the succeeding elements Ce–Sm, the so-
lution conductances are also low and the 31P{1H} NMR
spectra show a strong Ph3PO resonance and broad
features due to the coordinated phosphine oxide, con-
sistent with extensive dissociation into the neutral tris
complex. For Eu–Tm the conductances rise steadily and
for these complexes the 31P{1H} NMR spectra clearly
show both tris and tetrakis complexes (and free Ph3PO)
present, whilst for Yb and Lu the data supports the
tetrakis species as the major solution form, although
even here the spectra still show some [LnCl3(Ph3PO)3]
present. In earlier studies these low conductances were
interpreted as evidence for neutral 7-coordinate
[LnCl3(Ph3PO)4] complexes in solution, but with the
NMR data available, it is now clear that the mixed so-
lution speciation is responsible for the low values and
the complexes are based upon 6-coordinate cations.
[LnCl2(Ph3PO)4]Cl are only the major solution species
for the last few elements of the series, and their isolation
in the solid state for the others depends upon slow
crystallisation from solutions containing excess phos-
phine oxide.
4. Experimental
Hydrated lanthanide chlorides were obtained com-
mercially (BDH, Aldrich or Alfa) or made by reaction of
the corresponding oxides with aqueous hydrochloric acid.
Absoluteethanol (Aldrich)andothersolventswereusedas
received. 31P{1H} NMR spectra were obtained on a Bru-
ker DPX400 at 161.9 MHz and are referenced to external
85% H3PO4, and 139La spectra similarly at 56.6 MHz and
referenced to [La(H2O)x]3þ at pH 1 in water. Other
physical measurements were made as described [5,12,13].
The complexes were made by a number of general
methods, representative examples of which are described.
Whilst the [PF6]ꢂ salts are readily obtained, isolation of
pure samples of the tris or tetrakis chloro-complexes was
best achieved under conditions where they separated
slowly from solution. Use of concentrated solutions from
which immediate precipitation occurred on mixing rarely
yielded pure samples, and since recrystallisation often
fails to improve purity due to the equilibria present as
discussed above, isolation of pure samples from the re-
action mixtures is preferable. Analytical data and selected
spectroscopic data are given in Tables 1 and 2.
3. Summary and conclusions
The studies have shown that [LnCl3(Ph3PO)3] and
[LnCl2(Ph3PO)4]þ co-exist in solutions of lanthanide
chlorides – Ph3PO in weakly coordinating solvents. Pure
samples of either can be isolated by choice of appropriate
conditions, and in solution the equilibria are readily
shifted by addition of Clꢂ or Ph3PO, but there was no
evidence for the formation of species with fewer than
three or more than four-coordinated phosphine oxides,
although complexes of the latter type, [LnCl(Ph3PO)5]2þ
can be isolated by chloride abstraction using strong Le-
wis acids [8]. There is a clear trend along the lanthanide
series with [LnCl3(Ph3PO)3] strongly favoured at the
beginning of the series and [LnCl2(Ph3PO)4]þ towards
the end. A question arises about how the solvents affect
the speciation – does the isolation of [LnCl3(Ph3PO)3]
from acetone mean these are the major solution species
in this solvent? It appears not. Taking Tm as an example,
since it shows sharp well separated 31P{1H} NMR res-
onances for each species making for easy identification,
acetone solutions of either [TmCl3(Ph3PO)3] or
[TmCl2(Ph3PO)4]þ show both species present. However,
[TmCl3(Ph3PO)3] is only slightly soluble in cold acetone,
and this poor solubility rather than essentially different
solution speciation probably accounts for its ready iso-
lation from acetone solution, the fast solution equilibria
permitting crystallisation of the least soluble form.
Similarly both species are present in MeCN solution, and
probably in EtOH, although in ethanol there are mod-
erate solvent shifts (<3 ppm) on the major resonances
and it may be that some ethanol coordination occurs. We
have not however isolated any complexes of the type
[LnCl3(Ph3PO)2(EtOH)] in this work, although [Ln
(NO3)3(Ph3PO)2(EtOH)] are well established in the
nitrate systems [2,4]. The data suggest that
4.1. [La(Ph3PO)3Cl3]
A solution of LaCl3 ꢀ 7H2O (0.425 g, 1.1 mmol) in
acetone (5 cm3) was added to a solution of Ph3PO (1.11
g, 4.0 mmol) in acetone (10 cm3) and the mixture ref-
luxed for 20 min. The solution was then cooled and
refrigerated for 24 h. The white powder was filtered off
and dried in vacuo. Yield: 0.45 g (38%). 1H NMR
(CDCl3) 300 K. 7.35–7.70(m).
4.2. [PrCl2(Ph3PO)4]Cl ꢀ n(solv)
A solution of PrCl3 ꢀ 7H2O (0.38 g, 1.0 mmol) in
boiling absolute ethanol (10 cm3) was added to a solution
of Ph3PO (1.67 g, 6.0 mmol) in hot ethanol (15 cm3) and
the mixture refluxed for 20 min. It was allowed to cool to
room temperature, filtered to remove any solid, and the
filtrate refrigerated for 24 h. Very pale greenish crystals
separated, which were filtered off, rinsed with diethyl
ether (10 cm3) and dried in vacuo. Yield: 0.60 g (43%).
4.3. [CeCl2(Ph3PO)4]PF6 ꢀ H2O
Solutions of CeCl3 ꢀ 7H2O (0.18 g, 0.50 mmol) in
ethanol (10 cm3) and [NH4][PF6] (0.082 g, 0.50 mmol)

1090
M.J. Glazier et al. / Inorganica Chimica Acta 357 (2004) 1083–1091
in ethanol (10 cm3) were mixed and stirred for 5 min. A
solution of Ph3PO (0.56 g, 2.0 mmol) in ethanol (10
cm3) was added resulting in an immediate precipitate.
The mixture was stirred for 1 h and left to stand
overnight. The precipitate was then filtered off, rinsed
with cold ethanol (5 cm3) and dried in vacuo. Yield:
0.46 g (62%).
structure was well behaved both the structures had
disordered atoms associated primarily with the ethanol
solvate molecules. For the Tb structure rather surpris-
ingly there was disorder in the Clꢂ anion with one
position being shared by an EtOH molecule. The site
occupation factor (sof) was reduced for the remaining
EtOH molecules in both the Tb and Yb structures to
reduce the large isotropic atomic displacement param-
eters values of the C and O atoms to more normal
values. The rather large residual electron-density peaks
were close to EtOH molecules and represent inade-
quately modelled disorder and may even be O(H2)
molecules as noted in the NMR spectra.
4.4. Crystal structure determinations
Suitable crystals were isolated from the syntheses in
ethanol solution and attempts to grow crystals using
the solid products were not successful. Data were re-
corded at 120 K using a Nonius CCD diffractometer
fitted with graphite-monochromator and Mo Ka radi-
ation. An empirical absorption correction using the
SORTAV procedure [19] was applied and the crystal-
lographic data are given in Table 5. The structures were
solved in space group P21/c using direct methods [20] to
locate the heavy atoms (Ln, Cl, P and O) and sub-
sequent structure-factor and electron-density calcula-
tions located the remaining non-hydrogen atoms. H
atoms on the phenyl groups were added in calculated
positions and full-matrix least-squares refinement on F2
[21] was carried out to convergence. While the cation
5. Supplementary material
Atomic coordinates, atomic displacement parameters
and bond lengths and angles have been deposited at the
Cambridge Crystallographic Data Centre (CCDC) as
CCDC Nos. 211998 (Tb) and 211999 (Yb). Copies of
this information may be obtained free of charge from
The Director, CCDC, 12 Union Road, Cambridge CB2
1EZ, UK (fax: +44-1223-336-033; e-mail: deposit@ccdc.
cam.ac.uk or www: http://www.ccdc.cam.ac.uk).
Table 5
Crystallographic data for [MCl2(Ph3PO)4]Cl ꢀ nEtOH (M ¼ Tb and Yb)a
Compound
[TbCl2(Ph3PO)4]Cl ꢀ 1.55EtOH
[YbCl2(Ph3PO)4]Cl ꢀ 2.25EtOH
C76:50H73:50Cl3O6:25P4Yb
1496.12
Empirical formula
Formula weight
Crystal system
Space group
C75:10H69:30Cl3O5:55P4Tb
1449.76
monoclinic
P21/c (no. 14)
14.563(2)
17.364(2)
29.571(4)
97.238(5)
7417.9(13)
1.298
monoclinic
P21/c (no. 14)
14.431(3)
ꢀ
a (A)
ꢀ
b (A)
17.456(4)
29.261(7)
ꢀ
c (A)
b (°)
97.563(10)
7307(3)
1.360
3
ꢀ
U (A )
Dcalc (g cmꢂ3
Z
)
4
2961
4
3054
F ð000Þ
Crystal size (mm)
Reflections measured
Unique reflections (Rint
hkl range
0.16 ꢄ 0.16 ꢄ 0.12
57950
12885 (0.11)
0.16 ꢄ 0.12 ꢄ 0.10
36126
11939 (0.13)
)
)13 to 17, )20 to 20, )35 to 35
1.005, 0.948
)15 to 17, )20 to 19, )33 to 34
0.970, 0.942
Maximum, minimum transmission
Data in refinement
Parameters
12885
786
11939
793
l (cmꢂ1
S
Maximum shift/esd
)
11.95
1.025
0.004
15.27
1.040
0.002
ꢂ3
ꢀ
Residual electron density (eA
Rb(I > 2rðIÞ)
)
)1.16 to +3.04
0. 082 (8496 reflections)
0.131
)1.16 to +3.10
0.089 (6581 reflections)
0.173
R (all data)
wR2c(all data)
0.248
0.241
a
ꢀ
In common: k(Mo KaÞ ¼ 0.71073 A; T ¼ 120 K; max. 2h ¼ 50.0°; SORTAV absorption correction [19].
P
P
b R ¼ jjFoj ꢂ jFcj= jFoj.
P
P
2
1=2
c wR2 ¼ ½ wðFo2 ꢂ Fc2Þ = wFo4ꢃ
.

M.J. Glazier et al. / Inorganica Chimica Acta 357 (2004) 1083–1091
1091
[9] Z. Xing-Wang, L. Xing-Fu, F. Benetollo, G. Bombieri, Inorg.
Chim. Acta 139 (1987) 103.
Acknowledgements
[10] Z. Hou, K. Kobayashi, H. Yamazaki, Chem. Lett. (1991) 265.
[11] S. Petricek, A. Demsar, L. Golic, J. Kosmrlj, Polyhedron 19
(2000) 199.
We thank EPSRC for support and Professor M.B.
Hursthouse for access to the EPSRC Nonius Kappa
CCD diffractometer.
[12] N.J. Hill, W. Levason, M.C. Popham, G. Reid, M. Webster,
Polyhedron 21 (2002) 445.
[13] N.J. Hill, L.-S. Leung, W. Levason, M. Webster, Inorg. Chim.
Acta 343 (2003) 169.
References
[14] W. Levason, M. Webster, Acta Crystallogr. E 58 (2002) 176.
[15] G.M. Begun, A.C. Rutenberg, Inorg. Chem. 6 (1967) 2212.
[16] B. Bleaney, J. Magn. Res. 8 (1972) 91;
[1] T.S. Lobana, in: F.R. Hartley (Ed.), The Chemistry of Organo-
phosphorus Compounds, Wiley, New York, 1992.
[2] D.R. Cousins, F.A. Hart, J. Inorg. Nucl. Chem. 29 (1967) 1745.
[3] G. Valle, G. Casotto, P.L. Zanonato, B. Zarli, Polyhedron 5
(1986) 2093.
A.A. Pinkerton, M. Rossier, S. Spiliadis, J. Magn. Res. 64 (1985)
420.
[17] D. Rehder, in: P.S. Pregosin (Ed.), Transition Metal NMR,
Elsevier, New York, 1991.
[4] W. Levason, E.H. Newman, M. Webster, Polyhedron 19 (2000)
2697.
[18] D.F. Evans, P.H. Missen, J. Chem. Soc. Dalton Trans. (1982)
1929.
[5] M. Bosson, W. Levason, T. Patel, M.C. Popham, M. Webster,
Polyhedron 20 (2001) 2055.
[19] R.H. Blessing, J. Appl. Crystallogr. 30 (1997) 421.
[20] G.M. Sheldrick, SHELXS 97, Program for the solution of
[6] W. Levason, E.H. Newman, M. Webster, Acta Crystallogr. C 56
(2000) 1308.
€
crystal structures, University of Gottingen, Gottingen, Ger-
€
[7] D.R. Cousins, F.A. Hart, J. Inorg. Nucl. Chem. 30 (1968) 3009.
[8] H.-K. Wang, M.-J. Zhang, X.-Y. Jing, J.-T. Wang, R.-J. Wang,
H.-G. Wang, Inorg. Chim. Acta 163 (1989) 19.
many, 1997.
[21] G.M. Sheldrick, SHELXL 97, Program for the refinement of crystal
€
structures, University of Gottingen, Gottingen, Germany, 1997.
€