Synthesis, structures and mass spectrometry of lanthanide nitrate complexes with tricyclohexylphosphine oxide

Article abstract of DOI:10.1016/j.poly.2007.06.023

The complexes Ln(NO₃)₃(OPCy₃)₃(EtOH)_x (Cy = cyclohexyl, C₆H₁₁ x = 0 for Ln = Eu, Er, x = 0.5 for Ln = La, Nd and x = 1 for Ln = Tm, Yb) have been prepared by reaction of lanthanide nitrates with Cy₃PO in ethanol. The single crystal X-ray structures for Ln = La, Nd, Eu, Er, Tm and Yb are reported. The structures for Ln = La-Eu have two molecules in the unit cell in which the nitrates are all bound as bidentate ligands. The unit cell for Ln = Er contains two distinct molecules; one with three bidentate nitrates and one with two bidentate and one monodentate nitrate. The Tm and Yb complexes have one molecule in the unit cell with two bidentate and one monodentate nitrate ligands. The monodentate nitrates are hydrogen bonded to ethanol in the Tm and Yb structures but not in the Er complex. The infrared spectra of the three classes of complex do not readily permit identification of the monodentate nitrate groups. Electrospray mass spectrometry indicates that redistribution/ionisation reactions occur in solution. Ions formed by solvolysis reactions are attributed to gas phase processes associated with the electrospray technique. Tandem mass spectrometry for the La, Ho and Yb complexes shows that in the gas phase loss of Cy₃PO is the sole fragmentation pathway for all but the Yb complex where the higher energy required for initial fragmentation leads to a more complex fragmentation pattern.

Full text of DOI:10.1016/j.poly.2007.06.023

Polyhedron 26 (2007) 4865–4876
www.elsevier.com/locate/poly
Synthesis, structures and mass spectrometry of lanthanide
nitrate complexes with tricyclohexylphosphine oxide
a
b
b,
*
Ann P. Hunter , Anthony M.J. Lees , Andrew W.G. Platt
a
EPSRC National Mass Spectrometry Service Centre, Chemistry Department, University of Wales, Swansea, Singleton Park, Swansea SA2 8PP, UK
b
Faculty of Sciences, Staffordshire University, College Road, Stoke-on-Trent, ST4 2DE, UK
Received 23 May 2007; accepted 19 June 2007
Available online 4 July 2007
Abstract
The complexes Ln(NO3)3(OPCy3)3(EtOH)x (Cy = cyclohexyl, C6H11 x = 0 for Ln = Eu, Er, x = 0.5 for Ln = La, Nd and x = 1 for
Ln = Tm, Yb) have been prepared by reaction of lanthanide nitrates with Cy3PO in ethanol. The single crystal X-ray structures for
Ln = La, Nd, Eu, Er, Tm and Yb are reported. The structures for Ln = La–Eu have two molecules in the unit cell in which the nitrates
are all bound as bidentate ligands. The unit cell for Ln = Er contains two distinct molecules; one with three bidentate nitrates and one
with two bidentate and one monodentate nitrate. The Tm and Yb complexes have one molecule in the unit cell with two bidentate and
one monodentate nitrate ligands. The monodentate nitrates are hydrogen bonded to ethanol in the Tm and Yb structures but not in the
Er complex. The infrared spectra of the three classes of complex do not readily permit identification of the monodentate nitrate groups.
Electrospray mass spectrometry indicates that redistribution/ionisation reactions occur in solution. Ions formed by solvolysis reactions
are attributed to gas phase processes associated with the electrospray technique. Tandem mass spectrometry for the La, Ho and Yb com-
plexes shows that in the gas phase loss of Cy3PO is the sole fragmentation pathway for all but the Yb complex where the higher energy
required for initial fragmentation leads to a more complex fragmentation pattern.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Keywords: Crystal structures; Tandem mass spectrometry; Lanthanide nitrate; Complexes; Tricyclohexylphosphine oxide
1. Introduction
stoichiometry can be isolated depending on the metal to
ligand ratio. Complexes such as Ce(NO3)3(Ph3PO)2(CH3-
CH2OH) can be formed with low Ln: phosphine oxide
ratios [4]. With Ln:Ph3PO ratios of approximately 1:3
nine-coordinate Ln(NO3)3(Ph3PO)3 are formed and with
more ligand rich reaction media Ln(NO3)3(Ph3PO)4 can
be isolated [5]. Structurally characterised complexes in
the tetrakis(triphenylphosphine oxide) series show that
for the lighter lanthanides two nitrates are chelating whilst
one is monodentate and that as the ionic radii of the lan-
thanide decreases the monodentate nitrate is expelled from
the primary coordination sꢀphere giving cationic complexes,
[Ln(NO3)2(Ph3PO)4]+NO3 [5]. The replacement of the
phenyl by the more sterically demanding cyclohexyl group
in ligands has been shown to give significant changes in
structures. For instance 1:1 and 1:2 complexes are formed
between Ln(NO3)3 and 2,6(CH2P(O)R2)2C5H3NO when
Complexes of lanthanide nitrates with phosphine oxide
donors have been extensively studied [1]. One of the main
reasons for this interest lies in their use in nuclear fuels
reprocessing [2] where the high affinity of phosphine oxides
for lanthanide and actinide metals, together with the ready
adaptation of their peripheral structure to enhance solvent
extraction properties and their chemically robust nature
facilitates application. Many ligand systems with multiple
donors have been synthesised to increase selectivity in the
extraction process [3]. The chemically simpler triphenyl-
phosphine oxide systems have been structurally investi-
gated [4–6]. Here it is found that complexes of varying
*
Corresponding author.
E-mail address: A.Platt@staffs.ac.uk (A.W.G. Platt).
0277-5387/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.poly.2007.06.023

4866
A.P. Hunter et al. / Polyhedron 26 (2007) 4865–4876
R = Ph whilst with the bulkier ligand (R = Cy) only a 1:1
complex was isolated [3c] and [Li(R2PC(NiPr)2]n forms a
dimer with R = Ph, but a less sterically crowded hexamer
for R = Cy [7].
Yb complexes have one molecule in the unit cell in which
there are two bidentate and one monodentate nitrate
ligands. The structure of the Tm complex is shown in
Fig. 2.
We examined a series of lanthanide nitrate complexes
with tricyclohexylphosphine oxide to investigate the effect
of increasing the steric bulk of the ligand on the structural
properties of the resulting complexes. The increased lipo-
philicity of Cy3PO compared to Ph3PO may make its com-
plexes more suited to solvent extraction. In this respect any
transition from a neutral to cationic complex may also pro-
vide a mechanism for increased selectivity in extraction,
with ionic complexes having lower affinity to a non-aque-
ous phase.
In these complexes the monodentate nitrate is hydrogen
bonded to the lattice ethanol and this causes a distinct
asymmetry in the bonding of the monodentate nitrate.
˚
The H-bonded oxygen–Ln distance is over 4.4 A whilst
˚
the non-H-bonded distances are in the region of 3.5 A for
the Tm and Yb complexes.
The Ln–O distances have a good linear correlation with
the nine-coordinate (La to Er) and eight-coordinate (Er to
Yb) ionic radii of the respective lanthanide ions [9]. A plot
of the Ln–O(Nbidentate) and Ln–O(P) distances versus ionic
radius is shown in Fig. 3.
2. Results and discussion
In the Er structure the monodentate nitrate has an Er–
O(N) distance of 2.287(5) A compared with longer dis-
˚
˚
˚
2.1. Synthesis and structures
tances for the Tm (2.310(2) A) and Yb (2.304(2) A), respec-
tively. This discrepancy can be explained by the hydrogen
bonding between the nitrate and ethanol. In the Tm and
Yb complexes, the Oꢁ ꢁ ꢁO contacts between the nitrate
The reaction of lanthanide nitrates with tricyclohexyl-
phosphine oxide in metal to ligand ratios between 1:3
and 1:4 led to the isolation of Ln(NO3)3(OPCy3)3 as etha-
nol solvates except for the Eu and Er complexes. The 1:4
complexes which can be formed with Ph3PO [5] and the
smaller Ph2MePO [8] cannot be obtained with the more
bulky Cy3PO under these conditions. The complexes
formed in high yield giving crystals suitable for X-ray
diffraction.
The structures comprise two molecules in the unit cell
for the La, Nd, Eu and Er complexes and one for Tm
and Yb. Details of the data collection and refinement are
given in Table 1 and selected bond lengths and angles in
Table 2. The geometries about the metals are all distorted
mer-octahedra if the nitrate ligands are considered to act
as pseudo-monodentate ligands bonded via the nitrogen
atoms. The ‘‘cis’’ angles range from 104.5ꢁ to 74.0ꢁ with
an average of 90.9ꢁ over all the complexes, whilst the
‘‘trans’’ angles range from 153.4ꢁ to 176.9ꢁ with an average
of 164.7ꢁ.
The coordination mode of the nitrate ligands changes
with the decrease in the ionic radius of the lanthanide
ion. Thus for the larger ions La–Eu the nitrates all act as
bidentate ligands and the complexes are 9-coordinate.
Where the unit cells contain ethanol molecules, they are
not hydrogen bonded to any of the nitrate ligands. The
structure of one of the molecules in the unit cell of the lan-
thanum complex is shown in Fig. 1. The Er structure has
two molecules in the unit cell which differ in the binding
of the nitrates. One molecule has three bidentate nitrates
and a similar overall structure to those of the La, Nd
and Eu complexes, whilst the other has two bidentate
and one monodentate nitrate and is hence eight-coordi-
nate. It is interesting to note here that the monodentate
nitrate is not hydrogen bonded to the lattice ethanol, and
this appears to lead to relatively symmetrical bonding to
the metal with non-bonded distances between the O–Er
˚
˚
and ethanol of 2.932(5) A (Tm) and 2.943(5) A (Yb) are
shorter than the sum of the Van der Waals radius for oxy-
gen (3.04 A) and are thus indicative of a hydrogen bond.
˚
2.2. Infrared spectroscopy
Infrared spectroscopy has routinely been used as a
means of identifying the coordination mode of nitrate
ligands to metals [10]. Thus although monodentate and
bidentate coordinated nitrates have the same local symme-
try the splitting of the N–O stretching modes has been
taken as a means of assigning structures. Further, ionic
nitrate, with a higher local symmetry should be readily dis-
tinguishable by the number of bands observed alone.
We have recently encountered complexes in which for-
mally ionic nitrates are hydrogen bonded to lanthanide
coordinated ligands or lattice solvent molecules [11]. The
spectra in these instances are typical of bidentate nitrates.
It was thus apparent that the use of spectroscopy alone
in determining the coordination mode of nitrate ligands
was prone to uncertainties when H-bonding ligands were
present.
The infrared spectra of the Cy3PO complexes are as
expected for complexes of phosphine oxides with lantha-
nide nitrates. The P@O stretch decreases on coordination
from 1146 cmꢀ1 to about 1100 cmꢀ1 on coordination to
the metal. The spectra of the N–O region in these com-
plexes differ only subtly despite the presence of monoden-
tate nitrates in the Er, Tm and Yb complexes. All
complexes show absorptions typical of coordinated nitrate
with strꢀo1ng bands in the 1480–1450 cmꢀ1 and 1300–
1290 cm
region and a weaker absorption around
1040 cmꢀ1. Complexes which contain both monodentate
and bidentate nitrates do not display features which would
allow unambiguous identification of the presence of
monodentate nitrates.
˚
of 3.940(6) and 3.979(5) A. The structures of the Tm and

Table 1
Crystal data, data collection and refinement parameters
Compound
[La(Cy3PO)3(NO3)3] Æ
0.25EtOH
[Nd(Cy3PO)3(NO3)3] Æ
0.25EtOH
[Eu(Cy3PO)3(NO3)3]
C108H198Eu2N6O24P6
[Er(Cy3PO)3(NO3)3]
C108H198Er2N6O24P6
[Tm(Cy3PO)3(NO3)3] Æ
EtOH
[Yb(Cy3PO)3(NO3)3] Æ
EtOH
Crystal data
Molecular formula
C108H198La2N6O24P6
C108H198N6Nd2O24P6
+ CH3O0.5
2462.06
monoclinic
P21
C54H99N3O12P3Tm
+ C2H6O
1290.27
orthorhombic
Pc21n
C54H99N3O12P3Yb
+ C2H6O
1294.38
orthorhombic
Pc21n
+ CH3O0.5
2451.40
monoclinic
P21
Molecular weight, Mr
Crystal system
Space group
Unit cell dimensions
2454.46
monoclinic
P21
2485.06
monoclinic
P21
˚
a (A)
11.7015(7)
18.7997(6)
28.6979(16)
90
100.953(2)
90
11.6360(4)
18.7227(12)
28.4733(17)
90
101.063(3)
90
11.6216(2)
18.7278(4)
28.4435(6)
90
101.5490(10)
90
11.62110(10)
18.6661(2)
27.9691(4)
90
99.1778(4)
90
14.2639(2)
22.2990(4)
19.1100(3)
90
90
90
14.25470(10)
22.3444(2)
19.13000(10)
90
90
90
˚
b (A)
˚
c (A)
a (ꢁ)
b (ꢁ)
c (ꢁ)
3
˚
Cell volume, V (A )
Formula units/unit cell,
6198.3(5)
2
6087.8(6)
2
6065.3(2)
2
5989.40(12)
2
6078.33(17)
4
6093.15(8)
4
Z
Dcalc (Mg mꢀ3
F(000)
)
1.313
2594
1.343
2606
1.344
2592
1.378
2612
1.410
2720
1.411
2724
Reflections to obtain cell
parameters
98420
55177
135855
102414
28935
75486
h (ꢁ)
2.91–26.73
0.71073
0.825
2.92–26.01
0.71073
0.991
2.91–27.48
0.71073
1.171
2.95–27.46
0.71073
1.540
2.91–27.48
0.71073
1.600
2.91–27.49
0.71073
1.675
˚
k (Mo Ka) (A)
l (mmꢀ1
T (K)
)
120(2)
120(2)
120(2)
120(2)
120(2)
120(2)
Crystal description
Colour
Dimensions (mm)
block
plate
block
block
prism
shard
colourless
0.3 · 0.16 · 0.14
colourless
0.1 · 0.1 · 0.02
colourless
0.1 · 0.08 · 0.04
colourless
0.2 · 0.1 · 0.1
colourless
0.3 · 0.2 · 0.2
colourless
0.38 · 0.18 · 0.12
Data collection and SORTAV absorption correction [19]
Tmax
Tmin
0.97779
0.92571
34864
1.01367
0.95229
37736
1.35378
0.68948
73696
1.09172
0.86566
80937
1.00179
0.90410
13116
0.80646
0.72406
68372
Measured reflections
Independent reflections
20342
19181
24697
26588
13116
13420
Observed reflections
14127 (n = 4)
14798 (n = 4)
20285 (n = 4)
20777 (n = 4)
11568 (n = 4)
11804 (n = 4)
(Fo > nrFo))
Rint
0.0703
0.0706
0.1465
0.1293
0.0000
0.0710
hmax (ꢁ)
26.73
26.01
27.45
27.46
27.48
27.49
h
k
l
ꢀ14 ! 13
ꢀ23 ! 20
ꢀ28 ! 36
ꢀ12 ! 14
ꢀ23 ! 21
ꢀ34 ! 35
ꢀ15 ! 14
ꢀ22 ! 24
ꢀ36 ! 36
ꢀ15 ! 15
ꢀ24 ! 24
ꢀ36 ! 36
ꢀ18 ! 18
ꢀ28 ! 28
ꢀ24 ! 24
ꢀ18 ! 18
ꢀ27 ! 29
ꢀ24 ! 24
Refinement on F2
Scattering factors from International Tables for Crystallography (vol. C)
Reflections
Parameters/
restraints
20342
1342/1
19181
1342/1
24697
1317/37
26588
1307/1
13116
691/1
13420
690/1
(continued on next page)

Table 1 (continued)
Compound
[La(Cy3PO)3(NO3)3] Æ
0.25EtOH
[Nd(Cy3PO)3(NO3)3] Æ
0.25EtOH
[Eu(Cy3PO)3(NO3)3]
[Er(Cy3PO)3(NO3)3]
[Tm(Cy3PO)3(NO3)3] Æ
EtOH
[Yb(Cy3PO)3(NO3)3] Æ
EtOH
Weighting
scheme, wa
Final R
indicesb
(Fo > nrFo)
R (F)
[r2(Fo2) + (0.0354P)2]ꢀ1 [r2(Fo2) + (0.0331P)2]ꢀ1 [r2(Fo2) + (0.0901P)2 + 7.72P]ꢀ1 [r2(Fo2) + (0.0599P)2 + 0.81P]ꢀ1 [r2(Fo2) + (0.0426P)2 + 14.02P]ꢀ1 [r2(Fo2) + (0.0365P)2]ꢀ1
n = 4
n = 4
n = 4
n = 4
n = 4
n = 4
0.0585
0.0981
0.987
0.0546
0.0889
1.007
0.0734
0.1756
1.091
0.0634
0.1339
1.031
0.0307
0.0630
1.046
0.0305
0.0646
1.014
wR2(F2)
S
Final R indicesb (all data)
R(F)
0.1093
0.1086
0.001
1.24
0.0862
0.0960
0.002
0.51
0.0924
0.1867
0.001
2.77
0.0887
0.1456
0.001
1.07
0.0416
0.0663
0.002
0.0395
0.0679
0.002
1.08
wR2(F2)
(D/r)max
Dqmax
0.999
ꢀ3
˚
(e A
Dqmin (e A
Extinction
)
˚
ꢀ3
)
ꢀ1.38
ꢀ0.77
ꢀ3.17
ꢀ1.85
ꢀ0.623
ꢀ0.99
none
none
none
none
none
none
correction
Absolute
structure
ꢀ0.011(10)
ꢀ0.014(8)
ꢀ0.017(10)
ꢀ0.020(7)
0.520(5)c
ꢀ0.030(4)
parameter
a
b
c
P = [max(Fo2, 0) + 2Fc2]/3.
P
P
P
P
R1 = iFoi ꢀ Fci/ jFo; wR2 = [ [w(Fo ꢀ Fc2)2]/ [w(Fo2)2]]0.5
.
2
Batch scale factor for TWIN refinement.

Table 2
Selected bond lengths (A), bond angles (ꢁ) and geometry between non-bonded atoms of Ln(NO3)3(Cy3PO)3
˚
[Ln(Cy3PO)3(NO3)3] Æ 0.25C2H5OH
La Nd
2.596(5) 2.534(5)
[Ln(Cy3PO)3(NO3)3]
Eu
[Ln(Cy3PO)3(NO3)3] Æ C2H5OH
Tm Yb
2.449(3) 2.450(2)
Er
2.444(5)
2.431(5)
2.441(5)
2.443(6)
2.287(5)
Ln(1)–O(1a)
Ln(1)–O(3a)
Ln(1)–O(4a)
Ln(1)–O(6a)
Ln(1)–O(7a)
Ln(1)–O(9a)
Ln(1)–O(10a)
Ln(1)–O(11a)
Ln(1)–O(12a)
Ln(1)ꢁ ꢁ ꢁN(1a)
Ln(1)ꢁ ꢁ ꢁN(2a)
Ln(1)ꢁ ꢁ ꢁN(3a)
Ln(1)ꢁ ꢁ ꢁP(1a)
Ln(1)ꢁ ꢁ ꢁP(2a)
Ln(1)ꢁ ꢁ ꢁP(3a)
Ln(2)–O(1b)
Ln(2)–O(3b)
Ln(2)–O(4b)
Ln(2)–O(6b)
Ln(2)–O(7b)
Ln(2)–O(9b)
Ln(2)–O(10b)
Ln(2)–O(11b)
Ln(2)–O(12b)
Ln(2)ꢁ ꢁ ꢁN(1b)
Ln(2)ꢁ ꢁ ꢁN(2b)
Ln(2)ꢁ ꢁ ꢁN(3b)
Ln(2)ꢁ ꢁ ꢁP(1b)
Ln(2)ꢁ ꢁ ꢁP(2b)
Ln(2)ꢁ ꢁ ꢁP(3b)
2.493(6)
2.515(6)
2.526(6)
2.509(7)
2.499(7)
2.546(7)
2.313(6)
2.314(6)
2.333(6)
2.944(8)
2.950(7)
2.961(9)
3.8029(22)
3.8409(23)
3.7959(21)
2.487(7)
2.531(6)
2.529(7)
2.511(6)
2.545(6)
2.584(7)
2.303(5)
2.336(5)
2.312(6)
2.912(7)
2.925(7)
2.967(9)
3.7875(19)
3.8375(20)
3.7677(25)
Ln(1)–O(1)
Ln(1)–O(3)
Ln(1)–O(4)
Ln(1)–O(6)
Ln(1)–O(7)
Ln(1)–O(9)
Ln(1)–O(10)
Ln(1)–O(11)
Ln(1)–O(12)
Ln(1)ꢁ ꢁ ꢁN(1)
Ln(1)ꢁ ꢁ ꢁN(2)
Ln(1)ꢁ ꢁ ꢁN(3)
Ln(1)ꢁ ꢁ ꢁP(1)
Ln(1)ꢁ ꢁ ꢁP(2)
Ln(1)ꢁ ꢁ ꢁP(3)
2.609(5)
2.616(5)
2.610(5)
2.635(5)
2.618(5)
2.420(5)
2.419(5)
2.424(4)
3.014(7)
3.049(7)
3.030(7)
3.8906(20)
3.9284(21)
3.8809(19)
2.610(5)
2.623(4)
2.640(5)
2.605(4)
2.653(5)
2.663(5)
2.408(4)
2.432(4)
2.400(5)
3.031(6)
3.041(7)
3.066(9)
3.8787(20)
3.9323(18)
3.8818(22)
2.552(5)
2.549(5)
2.533(5)
2.558(5)
2.575(5)
2.360(5)
2.348(4)
2.363(4)
2.954(7)
2.977(7)
2.959(6)
3.8263(20)
3.8604(17)
3.8175(19)
2.534(5)
2.541(5)
2.567(5)
2.526(5)
2.581(5)
2.596(4)
2.340(4)
2.366(3)
2.338(5)
2.958(7)
2.979(7)
3.001(6)
3.8079(17)
3.8644(14)
3.8062(22)
2.411(2)
2.442(2)
2.440(2)
2.310(2)
2.405(2)
2.439(2)
2.428(2)
2.304(2)
2.279(6)
2.225(4)
2.254(5)
2.857(7)
2.880(7)
3.275(6)
3.7423(22)
3.7453(18)
3.7195(20)
2.422(5)
2.459(5)
2.461(6)
2.410(6)
2.497(6)
2.545(7)
2.262(4)
2.273(4)
2.255(6)
2.859(6)
2.880(7)
2.955(8)
3.7329(18)
3.7668(16)
3.7242(24)
2.219(2)
2.241(2)
2.220(2)
2.852(3)
2.860(3)
3.322(3)
3.6995(9)
3.7531(10)
3.7129(9)
2.202(2)
2.228(2)
2.203(2)
2.848(3)
2.854(3)
3.321(3)
3.6858(8)
3.7452(9)
3.6971(8)
N(3a)ꢁ ꢁ ꢁLn(1)ꢁ ꢁ ꢁP(2a)
N(1a)ꢁ ꢁ ꢁLn(1)ꢁ ꢁ ꢁN(2a)
N(1a)ꢁ ꢁ ꢁLn(1)ꢁ ꢁ ꢁN(3a)
N(2a)ꢁ ꢁ ꢁLn(1)ꢁ ꢁ ꢁN(3a)
N(3a)ꢁ ꢁ ꢁLn(1)–O(11a)
P(1a)ꢁ ꢁ ꢁLn(1)ꢁ ꢁ ꢁP(3a)
O(10a)–Ln(1)–O(11a)
O(10a)–Ln(1)–O(12a)
O(11a)–Ln(1)–O(12a)
168.70(15)
174.68(19)
91.43(21)
87.01(19)
168.72(19)
162.64(4)
86.61(16)
156.51(15)
85.42(16)
167.40(13)
175.52(19)
92.21(20)
86.29(19)
167.31(16)
162.30(4)
87.06(16)
156.58(15)
85.83(16)
167.07(19)
176.91(21)
92.61(24)
86.99(24)
166.64(24)
161.57(5)
86.95(22)
155.10(21)
85.83(22)
172.35(13)
173.39(20)
83.19(19)
93.08(19)
174.18(17)
158.16(5)
97.50(18)
157.67(20)
90.37(18)
167.55(6)
167.30(6)
N(3)ꢁ ꢁ ꢁLn(1)ꢁ ꢁ ꢁP(2)
N(1)ꢁ ꢁ ꢁLn(1)ꢁ ꢁ ꢁN(2)
N(1)ꢁ ꢁ ꢁLn(1)ꢁ ꢁ ꢁN(3)
N(2)ꢁ ꢁ ꢁLn(1)ꢁ ꢁ ꢁN(3)
N(3)ꢁ ꢁ ꢁLn(1)–O(11)
P(1)ꢁ ꢁ ꢁLn(1)ꢁ ꢁ ꢁP(3)
O(10)–Ln(1)–O(11)
O(10)–Ln(1)–O(12)
O(11)–Ln(1)–O(12)
(continued on next page)
176.24(9)
90.79(9)
90.02(9)
166.75(9)
156.19(2)
92.91(11)
160.15(9)
95.61(9)
176.44(8)
91.00(9)
89.52(8)
166.55(9)
156.02(2)
92.90(10)
159.80(8)
95.94(8)

4870
A.P. Hunter et al. / Polyhedron 26 (2007) 4865–4876
Fig. 1. The structure of one of the molecules of La(Cy3PO)3(j2-NO3)3.
Thermal ellipsoids are drawn at the 50% probability level and cyclohexyl
groups have been omitted for clarity.
Fig. 2. The structure of Tm(Cy3PO)3(j1-NO3)(j2-NO3)2. Thermal ellip-
soids are drawn at the 50% probability level and cyclohexyl groups have
been omitted for clarity.
2.3. Electrospray mass spectrometry
The solution properties of the complexes were examined
by electrospray mass spectrometry in methanol solution,
and the data and assignments for the complexes studied
are given in Table 3. Assignments are made on the basis
of calculated m/z ratios, all of which are within 0.5 Da of
the theoretical values and the observation of analogous

A.P. Hunter et al. / Polyhedron 26 (2007) 4865–4876
4871
La
2.650
2.600
2.550
2.500
2.450
2.400
2.350
2.300
2.250
2.200
2.150
Nd
Eu
Tm
Yb
Er (9-coordinate)
Er (8-coordinate)
Ln-O(N bidentate)
Ln-O(P)
0.95
1.00
1.05
1.10
1.15
1.20
1.25
Ionic Radius / Å
Fig. 3. The variation in Ln–O distances with ionic radius of the lanthanide ion.
Table 3
The electrospray mass spectraa of Ln(NO3)3L3 in methanol/water
[Ln(NO3)2L3]+
[Ln(NO3)L4]2+
[Ln(NO3)L3]2+
[Ln(OH)L4]2+
[Ln(OMe)L4]2+
679.8(15)
[Ln(OMe)L3]2+
La
1151.6(<5)
1156.6(15)
1165.7(<5)
1170.6(5)
1171.6(20)
1177.6(20)
1181.6(20)
1186.7(25)
692.9(10)
695.5(50)
700.0(10)
702.3(15)
703.2(75)
705.9(50)
709.9(10)
710.4(50)
544.8(<5)
547.3(10)
501.0(<5)
554.0(5)
554.5(15)
557.5(10)
559.5(5)
670.4(10)
672.8(30)
677.5(<5)
680.0(15)
680.7(65)
683.4(45)
685.4(10)
687.9(45)
529.3(20)
531.7(20)
536.3(10)
538.8(10)
539.5(40)
542.3(25)
544.4(20)
546.8(20)
Nd
Eu
Gd
Tb
Ho
Tm
Yb
686.5(5)
687.0(10)
690.0(5)
562.2(15)
694.8(<5)
Tandem mass spectra
Precursor ion
Fragment ion
[Ln(NO3)2L3]+
[Ln(NO3)2L2]+
855.1
881.2
890.0
[Ln(NO3)L4]2+
[Ln(NO3)L3]2+
544.6
557.5
562.1
[Ln(OH)L4]2+
[Ln(OH)L3]2+
522.1
534.9
539.6
[Ln(OMe)L3]2+
[Ln(OMe)L2]2+
380.8
394.0
398.4b
La
Ho
Yb
a
m/z (relative intensity %).
Significant fragmentation of this ion was observed as discussed in the text.
b
species for the series of complexes studied. In addition the
theoretical isotope patterns give excellent agreement with
those observed. An example for the [Gd(NO3)2L3]+ ion is
typical and shown in Fig. 4. The spectra all show strong
signals from [L+H]+, [L+Na]+, [2L+H]+ whilst those from
lanthanide containing ions indicate that the complexes
undergo redistribution reactions in solution. Such redistri-
bution reactions are common for lanthanide complexes
and have previously been observed for b-diketonate [12]
and a variety of phosphine oxide [13] complexes. Peaks
from [Ln(NO3)2L3]+ are observed for all complexes, often
as low intensity clusters of signals. The main lanthanide
containing signals are derived from loss of nitrate to give
[Ln(NO3)L3]2+ and from loss of nitrate and attachment
of further ligand to give [Ln(NO3)L4]2+. There appears to
be no correlation between the abundance of +2 cations
and the presence of the monodentate nitrates observed in
the solid state structures. In addition to these processes sol-
volysis reactions are also apparent giving rise to intense
signals assigned to [Ln(OH)L4]2+, [Ln(OMe)L3]2+ and
[Ln(OMe)L4]2+. Such solvolysis processes have been
observed in the spectra of lanthanide salts and complexes
before [14] and are thought to be due to gas phase pro-
cesses rather than reflecting the nature of solution species.
The mass spectra of several of the complexes were recorded
again after the solutions had aged for about three weeks

4872
A.P. Hunter et al. / Polyhedron 26 (2007) 4865–4876
1170.6
100
Theoretical for
[Gd(NO3)2L3]2+
1169.6
1168.6
1172.6
1171.6
%
1167.6
1173.6
1174.6
1166.6
0
1170.7
100
Observed
1168.71169.7
1172.7
%
1171.7
1167.7
1173.7
1174.7
1166.7
m/z
0
1172
1158
1160
1162
1164
1166
1168
1170
1174
1176
1178
1180
1182
1184
Fig. 4. Theoretical isotope pattern for [Gd(NO3)2L3]+ (L = Cy3PO) and that observed in the positive ion electrospray mass spectrum.
and showed no significant differences with spectra obtained
from fresh solutions. The positive ion ESMS of
Ho(NO3)3L3 is representative of the spectra obtained and
is shown in Fig. 5.
Selected ions were also examined by tandem mass spec-
trometry to investigate their behaviour in the gas phase.
The ions were chosen to represent species which might be
expected initially to be present in solution namely
[2L +H]+
100
[L+H]+
[Ho(OH)L4]2+
[Ho(NO3)L4]2+
[Ho(NO3)L3]2+
%
[Ho(OMe)L ]2+
3
+
[Ho(NO3)2L2]
[3L + NH4]+
0
m/z
1200
1400
100
300
600
700
800
900
1000 1100
1300
200
400
500
Fig. 5. The positive ion Electrospray mass spectrum of Ho(NO3)3L3 (L = Cy3PO) in methanol/water with principal assignments.

A.P. Hunter et al. / Polyhedron 26 (2007) 4865–4876
4873
[Ln(NO3)2L3]+ and [Ln(NO3)L4]2+ and also ions which are
more likely to have been formed as part of the electrospray
process such as [Ln(OH)L4]2+ and [Ln(OMe)L3]2+. The
lanthanum, holmium and ytterbium complexes were stud-
ied to represent the range of lanthanides. The tandem mass
spectra of most of the ions are simple, showing loss of only
one tricyclohexylphosphine oxide ligand for all ions stud-
ied. The exception to this was [Yb(OMe)L3]2+ which
Fig. 6. The tandem mass spectra of the [La(NO3)2L3]+ (upper) and [Yb(OMe)L3]2+ (L = Cy3PO) (lower) ions.

4874
A.P. Hunter et al. / Polyhedron 26 (2007) 4865–4876
- L
Cy
-MeO.
- H2
- Cy.
Cy
[Yb(OMe)L3]2+
Cy
[Yb(OMe)L2]2+
398.4
[YbL2]2+
2+
[LYb
381.9
30%
O
2+
P
[LYb
]
O
P
]
546.7
100%
382.9
55%
340.8
50%
55%
-
-
Cy
H
2+
O
[
Yb
P
]
L
2+
[
O
LYb
P
]
H
299.7
30%
300.8
25%
Fig. 7. A possible fragmentation pathway for [Yb(OMe)L3]2+ (L = Cy3PO).
required a higher collision offset (COFF) to induce frag-
mentation similar to that of the La and Ho complex ions.
Typical spectra for the La and Yb complexes are shown
in Fig. 6. The Yb complex ion required COFF 50 eV,
whilst the corresponding La and Ho ions yielded simple
and meaningful fragment ions with COFF 30–40 eV. The
stability of the Yb complex to loss of ligand is presumably
due to the higher charge density on the metal ion. Under
the harsher conditions required for its fragmentation a
more complex spectrum was produced. This showed loss
of Cy3PO with further fragmentation due to loss of MeOÅ,
cyclohexyl radicals and possibly cyclohexadiene. Loss of
cyclohexene and cyclohexyl radicals under electron impact
has been observed previously in the mass spectra of aro-
matic cyclohexyl ethers [15] and in Cy2P(O)C2H4P(O)Cy2
[16] and whilst electrospray ionisation is generally less ener-
getic, the application of higher COFF in the tandem mass
spectra of the Yb complex presumably supplies sufficient
energy for radical formation.
97) [21]. Refinement was by full-matrix least-squares on
F2 data using SHELXL-97 [21]. Solutions and all refinements
were conducted using SHELX-97 from within the WinGX
[22] suite of software.
The Nd, Eu and Er structures were each refined in space
group P21 after isomorphous replacement into the La
structure. The Tm structure was solved by similar replace-
ment into the Yb structure (space group Pc21n) after
re-indexing the data and transforming the unit cell axes
from the alternative space group setting, Pbn21 for which
the data were collected. The asymmetric unit in each of
the La–Er structures consisted of two discrete lanthanide-
containing molecules. The asymmetric units of the Tm
and Yb structures each consisted of a single lanthanide-
containing molecule.
Attempts to refine the Er structure after initial substitu-
tion into the La structure showed the positions of the
atoms of one of the bidentate nitrate groups in one of
the two molecules to be unstable. After deletion of nitrate
atoms, peaks were re-located in subsequent Fourier synthe-
ses to build a stable model containing a monodentate
nitrate group – consistent with the molecular structure
observed in the Yb and Tm complexes.
In each of the structures, non-H atoms were refined
anisotropically with the exception of the C atoms of
one of the cyclohexyl groups in each of the La and Nd
structures and two cyclohexyl groups in the Er structure.
The initial, fully anisotropic, models of the La–Er struc-
tures displayed marked librational elongation of the ther-
mal ellipsoids of the C-atoms in these groups, and their
geometry deviated from that expected for the cyclohexyl
moiety. A disordered, isotropic, model consisting of two
alternative sites with complementary site occupation fac-
tors was adopted for the La, Nd, and Er structures.
SIMU restraints were applied to C-atoms of the corre-
sponding group in the Eu structure for which data were
poorer.
The loss of H2 from a 2+ charged ion was also apparent
by the presence of peaks separated by one m/z unit. Fig. 7
illustrates
a possible fragmentation pathway which
accounts for the presence of the major ions observed. It
must be noted that the mechanism is tentative. Cleavage
of the Yb–OMe bond leading to loss of MeOÅ generates
an Yb(II) fragment which may explain why whilst further
decomposition of [Yb(OMe)L2]2+ was observed the corre-
sponding ion for La and Ho, for which the divalent states
are less stable, does not undergo this decomposition.
3. Experimental
3.1. Crystal structure determinations
Data were collected by the EPSRC National Crystallog-
raphy Service at the University of Southampton using pre-
viously described procedures [17–19].
The positions of the metal atoms in the La and Yb struc-
tures were estimated by Patterson methods using SHELXS-97
[20], and all remaining non-H atom positions were
obtained through subsequent Fourier syntheses (SHELXL-
An ethanol molecule was located near to one of the
nitrate groups in each of the La, Nd, Tm and Yb struc-
tures. The site occupation factors of the solvent molecules
in the La and Nd lattices were initially refined, and finally

A.P. Hunter et al. / Polyhedron 26 (2007) 4865–4876
4875
fixed at half occupancy. A similar solvent molecule was
3.4. Synthesis and characterisation
located in the Er structure apparently H-bonded to one
of the bidentate nitrate groups. This site occupation factor
of this solvent refined to <0.6%, and made little improve-
ment to the figures of merit. It was excluded from the final
structure. No similar solvent could be located in the differ-
ence Fourier maps of the Eu structure.
Tricyclohexylphosphine oxide was prepared by oxida-
tion of an acetone solution of tricyclohexylphosphine
with 30% H2O2. In a typical preparation the phosphine
(20.04 g) was suspended in 100 ml of acetone and hydro-
gen peroxide (9.00 g 30% aqueous solution) was added in
portions allowing the vigorous reaction to subside before
the next addition. The resulting clear solution was
allowed to cool to room temperature and the resulting
crystalline solid filtered, washed with a little cold acetone
and dried at the pump to give 19.18 g of a waxy colour-
less solid. NMR (acetone) d31P 51.6, literature 50.0 [25]
All H atoms were included in the refinements at idea-
lised positions riding on the atoms to which they were
˚
C–
bonded
(methane
C–H = 1.00 A,
methylene
˚
˚
˚
H = 0.99 A, methyl C–H = 0.98 A, O–H = 0.84 A) The
H-atoms of each functional group were assigned a single,
refined isotropic displacement parameter, except those of
solvent molecules which were assigned a common, refined
isotropic displacement parameter.
mPO 1146 cmꢀ1
.
The nitrate complexes were prepared by mixing hot
solutions of tricyclohexylphosphine oxide with the appro-
priate lanthanide nitrate in ethanol. On cooling crystals
formed in 80–100% yield. Representative syntheses and
characterisation are described below.
The non-centrosymmetric structures refined to give
Flack absolute structure parameters of ꢀ0.011(10),
ꢀ0.014(8), ꢀ0.017(10), ꢀ0.020(7), ꢀ0.030(4) for the La,
Nd, Eu, Er and Yb structures, respectively. The absolute
structure of the Tm complex could not be unambiguously
determined. ROTAX analysis [23] of the Tm data indicated
merohedral or pseudo-merohedral twinning. The Tm struc-
ture refinement was repeated assuming racemic twinning by
using the default TWIN setting in SHELXL-97 [21], and
refinement converged to yield a batch scale factor of
0.520(5).
La(NO3)3(Cy3PO)3 Æ 0.5EtOH. La(NO3)3 Æ 5H2O (0.23
g 0.55 mmol) in 10 ml hot ethanol was mixed with Cy3PO
(0.71 g 2.4 mmol) in 10 ml hot ethanol. The resulting
solution was warmed for 15 min during which time crys-
tals began to form. The mixture was cooled to room tem-
perature and allowed to stand overnight. The colourless
crystals were filtered, washed with a small quantity of
cold ethanol and dried at the pump to give 0.65 g
Molecular graphics were generated using SNOOPI [24]
(99%), mPO 1097 cmꢀ1
.
3.2. Mass spectrometry
Anal. Calc. for La(NO3)3(Cy3PO)3 Æ 0.5EtOH: C, 53.39;
H, 8.31; N, 3.40. Found: C, 52.88; H, 8.20; N, 3.28%.
Electrospray mass spectra were recorded on a either a
Waters ZQ-4000 or a Micromass Quattro II for the MS/
MS measurements. Samples in methanol solution were
loop injected into a flow of 1:1 methanol: water running
at 40lL/min into the electrospray capillary. Nebulisation
was pneumatically assisted by a flow of N2(g) at 15 L/min
and a drying gas of N2(g) at 100 L/min warmed by a source
temperature of 80 ꢁC. The spray voltage was 3.5 kV and
the cone voltage varied between 20 V and 90 V depending
on the mass and charge of the ions of interest. For tandem
mass spectra, Ar(g) at a pressure of 2 · 10ꢀ3 mbar was
introduced into the collision chamber (termed quadrupole
2 or Q2); ions formed during the ES process and mass
selected for individual analysis by the first quadrupole
(Q1), were accelerated through this region of relatively high
pressure by a voltage termed the collision offset (COFF).
The COFF voltage was adjusted to obtain optimal frag-
mentation for each precursor ion, and the resulting frag-
ment ions analysed by Q3.
Nd(NO3)3(Cy3PO)3 Æ 0.5EtOH.
Nd(NO3)3 Æ 6H2O
(0.26 g 0.59 mmol) in 5 ml hot ethanol was mixed with
Cy3PO (0.66 g 2.2 mmol) in 10 ml hot ethanol. The mixture
was cooled to room temperature and allowed to stand over-
night. The lilac crystals which formed were filtered, washed
with a small quantity of cold ethanol and dried at the pump
to give 0.65 g (99%), mPO 1099 cmꢀ1
.
Anal. Calc. for Nd(NO3)3(Cy3PO)3 Æ 0.5EtOH: C, 53.16;
H, 8.27; N, 3.38. Found: C, 52.63; H, 8.19; N, 3.32%.
Eu(NO3)3(Cy3PO)3. Eu(NO3)3 Æ 6H2O (0.26 g 0.58
mmol) in 10 ml hot ethanol was mixed with Cy3PO
(0.72 g 2.4 mmol) in 10 ml hot ethanol. The mixture was
cooled to room temperature and allowed to stand over-
night. The colourless crystals which formed were filtered,
washed with a small quantity of cold ethanol and dried
at the pump to give 0.63 g (88%) mPO 1102 cmꢀ1
Anal. Calc. for Eu(NO3)3(Cy3PO)3: C, 52.85; H, 8.13; N,
3.42. Found: C, 52.38; H, 8.15; N, 3.23%.
.
Tm(NO3)3(Cy3PO)3. Tm(NO3)3 Æ 6H2O (0.24 g 0.52
mmol) in 5 ml hot ethanol was mixed with Cy3PO (0.62 g
2.0 mmol) in 10 ml hot ethanol. The mixture was cooled
to room temperature and allowed to stand overnight.
The colourless crystals which formed were filtered, washed
with a small quantity of cold ethanol and dried at the pump
3.3. Infrared spectroscopy
Infrared spectra were recorded in the range
400–4000 cmꢀ1 on a Thermo Nicolet Avatar 370 instru-
ment operating in ATR mode. Samples were analysed
without pre-treatment other than crushing onto the optical
surface.
to give 0.55 g (84%), mPO 1104 cmꢀ1
Anal. Calc. for M(NO3)3(Cy3PO)3: C, 52.13; H, 8.02; N,
3.38. Found: C, 52.05; H, 8.20; N, 3.15%.
.

4876
A.P. Hunter et al. / Polyhedron 26 (2007) 4865–4876
[8] M. Bosson, W. Levason, T. Patel, M.C. Popham, M. Webster,
Acknowledgements
Polyhedron 20 (2001) 2055.
[9] R.D. Shannon, Acta Crystallogr., Sect. A 32 (1976) 751.
[10] K. Nakamoto, Infrared and Raman Spectra of Inorganic and
Coordination Compounds, 5th ed., Wiley, New York, 1997.
[11] A.M.J. Lees, A.W.G. Platt, Inorg. Chem. 42 (2003) 4673.
[12] J.M. Curtis, P.J. Derrick, A. Schnell, E. Constantin, R.T. Gallagher,
J.R. Chapman, Inorg. Chim. Acta 201 (1992) 197.
We are grateful to the Royal Society of Chemistry for
support through the research fund and EPSRC for access
to the X-ray data collection service at Southampton
University.
[13] (a) J. Fawcett, A.W.G. Platt, Polyhedron 22 (2002) 967;
(b) J. Fawcett, A.W.G. Platt, S. Vickers, M.D. Ward, Polyhedron 23
(2004) 2561;
Appendix A. Supplementary material
CCDC 602138 to 602143 contain the supplementary
crystallographic data for La to Yb. These data can be
obtained free of charge via http://www.ccdc.cam.ac.uk/
conts/retrieving.html, or from the Cambridge Crystallo-
graphic Data Centre, 12 Union Road, Cambridge CB2
1EZ, UK; fax: (+44) 1223-336-033; or e-mail: deposit@
ccdc.cam.ac.uk. Supplementary data associated with this
article can be found, in the online version, at
doi:10.1016/j.poly.2007.06.023.
(c) A.M.J. Lees, A.W.G. Platt, Polyhedron 24 (2005) 427.
[14] (a) I.I. Stewart, G. Horlick, Anal. Chem. 66 (1994) 3983;
(b) R. Colton, A. D’Agostino, J.C. Traeger, Mass Spectrom. Rev. 14
(1995) 79;
(c) S. Colette, B. Amekraz, C. Madic, L. Berthon, G. Cote, C.
Moulin, Inorg. Chem. 42 (2003) 2215.
[15] T.A. Molenaar-Langeveld, L.M. Van der Burg, J.W.H. Peeters, C.G.
De Koster, S. Ingemann, J. Mass Spectrom. 39 (2004) 1075.
[16] L.C. Baldwin, M.J. Fink, J. Organomet. Chem. 646 (2002) 230.
[17] DENZO – Data collection and processing software Z. Otwinowski, W.
Minor, in: C.W. Carter Jr., R.M. Sweet (Eds.), Methods in
Enzymology, Macromolecular Crystallography, Part A, vol. 276,
Academic Press, London, 1997, pp. 307–326.
References
[18] R. Hooft, COLLECT – Data collection software, Nonius B.V., 1998.
[19] (a) SORTAV – absorption correction software package: R.H. Blessing,
Acta Crystallogr., Sect. A 51 (1995) 33;
(b) R.H. Blessing, J. Appl. Crystallogr. 30 (1997) 421.
[20] (a) G.M. Sheldrick, SHELXS-97, Release 97-2: A Program for Auto-
matic Solution of Crystal Structures, University of Go¨ettingen,
Go¨ettingen, Germany, 1997;
(b) G.M. Sheldrick, Z. Dauter, K.S. Wilson, H. Hope, L.C. Sieker,
Acta Crystallogr., Sect. D 49 (1993) 18.
[21] G.M. Sheldrick, SHELXL-97, Release 97-2: A Program for Crystal
Structure Refinement, University of Go¨ettingen, Go¨ettingen, Ger-
many, 1997.
[1] (a) For reviews, see: S.A. Cotton, in: J.A. McCleverty, T.J. Meyer
(Eds.), Comprehensive Coordination Chemistry II, vol. 3, Elsevier,
Amsterdam, 2004, p. 93;
(b) T.S. Lobana, in: F.R. Hartley (Ed.), The Chemistry of Organo-
phosphorus Compounds, vol. 2, Wiley, New York, 1992, p. 409.
[2] (a) K.L. Nash, C. Lavallette, M. Borkowski, R.T. Paine, X. Gan,
Inorg. Chem. 41 (2002) 5849;
(b) A.N. Turanov, V.K. Karandashev, A.N. Yarkevich, Z.V. Safro-
nova, Solvent Extr. Ion Exc. 22 (2004) 391.
[3] (a) X. Gan, R.T. Paine, E.N. Duesler, H. No¨th, J. Chem. Soc., Dalton
Trans. (2003) 153;
[22] (a) L.J. Farrugia, WinGX: A Windows Program for Crystal Structure
Analysis, University of Glasgow, Glasgow, 1998;
(b) X. Gan, E.N. Duesler, R.T. Paine, Inorg. Chem. 40 (2001)
4420;
(b) L.J. Farrugia, J. Appl. Crystallogr. 32 (1999) 837.
[23] ROTAX: Simon Parsons & Bob Gould, University of Edinburgh with
additions by Richard Cooper (Oxford) and Louis Farrugia (Glas-
gow), November, 2001.
(c) X. Gan, B.M. Rapko, E.N. Duesler, I. Binyamin, R.T.
Paine, B.P. Hay, Polyhderon 24 (2005) 469.
[4] W. Levason, E.H. Newman, M. Webster, Acta Crystallogr., Sect. C
56 (2000) 1308.
[24] K. Davies, SNOOPI – Molecular Plotting Program, Chemical Crystal-
lography Laboratory, University of Oxford, 1983.
[25] M.M. Crutchfield, C.H. Dungan, J.H. Letcher, V. Mark, J.R. Van
Wazer, Topic in Phosphorus Chemistry, vol. V, Interscience, New
York, 1967.
[5] W. Levason, E.H. Newman, M. Webster, Polyhedron 19 (2000) 2697.
[6] M.J. Glazier, W. Levason, M.L. Matthews, P.L. Thornton, M.
Webster, Inorg. Chim. Acta 357 (2004) 1083.
[7] N.E. Mansfield, M.P. Coles, P.B. Hitchcock, J. Chem. Soc., Dalton
Trans. (2006) 2052.