A phosphine-mediated through-space exchange coupling pathway for unpaired electrons in a heterobimetallic lanthanide-transition metal complex

Article abstract of DOI:10.1002/chem.200700355

The reaction of P(CH₂OH)₃ with methyl anthranilate NH₂C₆H₄-2CO₂Me produced the ligand precursor P(CH₂NHC₆H₄-2-CO₂Me) ₃ (1). The reaction of 1 with [Y(N(SiMe₃)) ₂]₃] produced hexadentate yttrium complex [Y(P(CH ₂NC₆H₄-2-CO₂Me)₃}] (2), in which the metal centre is coordinated by three amido donors and the three carbonyl oxygen atoms of the ester groups. The ³¹P(¹H) NMR spectrum features ¹J_Y,P= 15 Hz, and DFT calculations demonstrate that through-space interaction between the minor lobe of the phosphine lone pair and the yttrium centre allows a large Fermi contactcontribution to this spin coupling constant. The EPR spectrum of the analogous paramagnetic Gd complex [Gd[P(CH₂NC₆H ₄-2-CO₂Me)₃)] (3) can be modelled by using a B₂₀ crystal field parameter of ±0.19 cm^-1. Heterodinuclear complexes were prepared by the reactions of 1 and 3 with [5,10,15,20tetrakis(4-methoxyphenyl)porphinato]cobalt(II), by binding of the phosphinelone pair to the d⁷ cobalt centre. The solid-state EPR spectrum of the heterodinuclear yttrium complex 4 exhibits large superhyperfine coupling to the phosphorus nucleus, indicative of an S=1/2 complex in which the unpaired electron resides in the cobalt d_z2 orbital directed at the phosphine donor. The magnetic susceptibility of the heterodinuclear Gd-Co complex S demonstrates that through-space antiferromagnetic coupling occurs between unpaired electrons on the gadolinium and cobalt centres.

Full text of DOI:10.1002/chem.200700355

FULL PAPER
DOI: 10.1002/chem.200700355
A Phosphine-Mediated Through-Space Exchange Coupling Pathway for
Unpaired Electrons in a Heterobimetallic Lanthanide–Transition Metal
Complex
Ritu Raturi,[a] Julie Lefebvre,[b] Daniel B. Leznoff,[b] Bruce R. McGarvey,[a] and
Samuel A. Johnson*[a]
Abstract: The reaction of P
(CH2OH)3
contribution to this spin coupling con-
stant. The EPR spectrum of the analo-
gous paramagnetic Gd complex
[Gd{P(CH2NC6H4-2-CO2Me)3}] (3) can
be modelled by using a B20 crystal field
parameter of ꢀ0.19 cmꢁ1. Heterodinu-
clear complexes were prepared by the
reactions of 1 and 3 with [5,10,15,20-
tetrakis(4-methoxyphenyl)porphinato]-
cobalt(II), by binding of the phosphine
lone pair to the d7 cobalt centre. The
solid-state EPR spectrum of the heter-
odinuclear yttrium complex 4 exhibits
large superhyperfine coupling to the
phosphorus nucleus, indicative of an
with methyl anthranilate NH2C6H4-2-
CO2Me produced the ligand precursor
P(CH2NHC6H4-2-CO2Me)3 (1). The re-
action of 1 with [Y{N(SiMe3)2}3] pro-
A
duced hexadentate yttrium complex
[Y{P(CH2NC6H4-2-CO2Me)3}] (2), in
which the metal centre is coordinated
by three amido donors and the three
carbonyl oxygen atoms of the ester
groups. The 31P{1H} NMR spectrum
S=1= complex in which the unpaired
2
electron resides in the cobalt dz2 orbital
directed at the phosphine donor. The
magnetic susceptibility of the heterodi-
nuclear Gd–Co complex
5 demon-
strates that through-space antiferro-
magnetic coupling occurs between un-
paired electrons on the gadolinium and
cobalt centres.
1
features JY,P =15 Hz, and DFT calcula-
Keywords:
pounds
heterometallic com-
lanthanides magnetic
tions demonstrate that through-space
interaction between the minor lobe of
the phosphine lone pair and the yttri-
um centre allows a large Fermi contact
·
·
properties · N,O ligands · through-
space interactions
Introduction
are also responsible for transmission of exchange coupling,
frequently by a superexchange pathway. Single atoms bridg-
ing metal centres often perform this role, but larger p-conju-
gated bridging ligands are also capable of propagating ex-
change coupling, and the cyano donor in particular has re-
cently revolutionised the design of magnetic network struc-
tures[5] and single-molecule magnets.[6] An increase in the
variety of ligands that could facilitate a building-block ap-
proach[7] to magnetic materials would be a boon to this
field, and in particular building blocks with at least axial
symmetry would aid in the modelling of magnetically aniso-
tropic systems.
Essential to the design of molecule-based magnets,[1,2] in-
cluding single-molecule magnets, is the ordering of spins via
exchange coupling.[3] Progress in this area demands synthetic
approaches to the design of intricate architectures, which
can include macromolecular polynuclear complexes, 1D
wires and 2D or 3D networks.[2,4] Typically, the same bond-
ing interactions that assemble these complexes and networks
[a] R. Raturi, Prof. B. R. McGarvey, Prof. S. A. Johnson
Department of Chemistry & Biochemistry
University of Windsor
We have shown how formally trianionic tripodal triamido-
401 Sunset Ave, Windsor ON, N9B3P4 (Canada)
Fax : (+1)519-973-7098
E-mail: sjohnson@uwindsor.ca
phosphine ligands such as P(CH2NPh)3 can be used to gen-
C
[b] J. Lefebvre, Prof. D. B. Leznoff
Department of Chemistry, Simon Fraser University
8888 University Drive, Burnaby BC, V5A1S6 (Canada)
obimetallic analogues, these P(CH2NPh)3M moieties could
G
act as valuable building blocks. Due to the ubiquity of phos-
phine donors in coordination chemistry, it should prove
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2008, 14, 721 – 730
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
721

disiloxane solution. An ORTEP depiction of the solid-state
molecular structure is shown in Figure 2. The structure fea-
tures intramolecular hydrogen bonding between the carbon-
Figure 1. Through-space and through-bond interactions in tripodal com-
plexes of P(CH2NPh)3 with metal centres M and M’ via the minor and
ACHTREUNG
major lobes, respectively, of the lone-pair orbital on phosphorus.
facile to utilize these complexes in the assembly of magnetic
polynuclear complexes and modify them to generate extend-
ed networks. The four s bonds separating M and M’ would
seem to preclude any significant exchange coupling via this
pathway, but the minor lobe of the phosphine lone pair ex-
tends towards metal centre M, which in theory could pro-
vide a unique combination of through-bond and through-
space pathways for exchange coupling. The study of d–f het-
erodinuclear complexes of lanthanides and transition
metals[10–13] is of current interest due to reports that the
large magnetic anisotropies of the heavier lanthanides[14] can
be utilised in the design of single-molecule magnets.[12,15]
Herein we show how this through-space interaction can be
utilised with paramagnetic metal centres to facilitate ex-
change coupling between M and M’, even in the difficult
case where M is a lanthanide bearing core-like f electrons.
Figure 2. Solid-state structure of 1 as determined by X-ray crystallogra-
phy. Hydrogen atoms are omitted for clarity, except those associated with
N1, N2 and N3. Hydrogen-bonding interactions are shown as dashed
lines. Selected bond angles [8]: C(1)-P(1)-C(2) 102.6213, C(1)-P(1)-C(3)
98.4414, C(1)-P(1)-C(3) 97.25(13).
yl and amino groups. The positions of the three amino hy-
drogen atoms were refined by using isotropic thermal pa-
rameters. The sum of C-P-C angles is 298.31(2)8, which is
typical for phosphine donors.
Synthesis of a mononuclear yttrium complex: The similar
ionic radius of YIII to those of heavier lanthanides such as
GdIII prompted us to start our studies by preparing an yttri-
um complex of 1 as a diamagnetic model complex. Reaction
Results and Discussion
Synthesis of the ligand precursor: To achieve the goal of
stable well-defined lanthanide building blocks, we sought to
modify the tripodal triamido donors we had previously uti-
lised to incorporate additional donors, due to the propensity
of the trivalent lanthanides to assume high coordination
numbers. The target ligand precursor P(CH2NHC6H4-2-
of 1 with [Y{N(SiMe3)2}3] in toluene at room temperature
[Eq. (2)] resulted in precipitation of rhombohedron-shaped
A
yellowcrystals of [Y{P(CH NC6H4-2-CO2Me)3}] (2). The ar-
2
1
omatic region of the H NMR spectrum of 2 features four
multiplets, which were assigned by a combination of 2D
COSY and 2D NOESY spectroscopy. Analysis of the reac-
1
CO2Me)3 (1) was synthesised by reaction of P
A
tion mixture by H and 31P{1H} NMR spectroscopy revealed
with methyl anthranilate H2NC6H4-2-CO2Me in toluene
[Eq. (1)]. The water produced in this reaction was removed
by azeotropic distillation with a
Dean–Stark apparatus.[16] Solu-
tions of 1 oxidize rapidly in air,
and thus the solid product was
stored in an inert-atmosphere
glove box.
that no side products were formed, other than the by-prod-
uct HN(SiMe3)2.
A
X-ray quality crystals of 1
were obtained by slow evapora-
tion of a benzene/hexamethyl-
The solid-state molecular structure of 2 was determined
by X-ray crystallography, and an ORTEP depiction is shown
in Figure 3. Despite the collection of X-ray diffraction data
on several crystals of 2 that appeared suitable for solid-state
722
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 721 – 730

Through-Space Coupling in a Lanthanide–Transition Metal Complex
FULL PAPER
orbit, dipolar and diamagnetic spin–orbit contributions to
the coupling constant are all calculated to be much smaller
than the Fermi contact term, with values of ꢁ0.8, ꢁ0.14 and
ꢁ0.026 Hz, respectively. An analysis of the molecular orbi-
tals predicted[20] from this calculation determined that only
HOMOꢁ3 has any significant overlap of density between P
and Y, as required for a Fermi contact term. This orbital,
which is primarily associated with the lone pair on phospho-
rus, is depicted in Figure 4. The minor lobe of the lone pair
Figure 3. Solid-state molecular structure of 2 as determined by X-ray
crystallography. Hydrogen atoms are omitted for clarity.
structure determination, twinning complicated solution of
the structure. All atoms in the model structure were treated
isotropically, and some restraints had to be used, which lim-
ited the accuracy of bond lengths and angles. Complex 2 has
crystallographic C3 symmetry, and features a six-coordinate
Y centre chelated by the ligand amido donors and the car-
bonyl oxygen atoms of the ester groups. The phosphine lone
pair is directed away from the metal centre and thus is avail-
able for donation to a second metal.
Figure 4. Depiction of the HOMOꢁ3 orbital associated with the phos-
phorus lone pair, obtained from
0.01 eŠ3 isosurface.
a DFT calculation, shown as the
Through-space 31P–89Y coupling: The 31P{1H} chemical shift
of 1 is d=ꢁ33.6 ppm, whereas for 2 the 31P signal is ob-
served at d=ꢁ57.0 ppm. This shift to higher field is unusual,
because the increase in C-P-C angles upon coordination of a
large metal centre such as Y should cause the opposite
effect;[17] we have previously ascribed this unusual shift in
related complexes to interactions of the minor lobe of the
lone pair with the adjacent metal centre.[9]
orbital clearly extends back towards the Y centre, and this
allows for a through-space interaction.
Synthesis of a mononuclear gadolinium complex: The reac-
tion of ligand precursor 1 with Gd[N(SiMe3)2]3 produced the
A
paramagnetic complex [Gd{P(CH2NC6H4-2-CO2Me)3}] (3),
as shown in Equation (2). Similar to the synthesis of 2,
yellowrhombohedron-shaped crystals precipitated from tol-
uene. The crystal structures of 2 and 3 are isomorphous, and
only small metric differences in the unit-cell parameters,
due to the slightly larger size of GdIII, were observed for this
complex. As with complex 2, this C3-symmetric complex
crystallised in a trigonal space group, difficulties associated
with twinning complicated solution of this structure and con-
nectivity could not be accurately determined.
The X-band EPR spectrum of a powdered sample of com-
plex 3 was obtained at 77 K (Figure 5). The spectrum dis-
plays resonances from nearly zero field to 12500 G. The
spectrum can be adequately modelled by using only a B20
value of ꢀ0.194 cmꢁ1 (2080 G) and a g value of 1.994.[21] As
would be expected for GdIII, the B40 B43, B60, B63 and B66
crystal-field parameters were found to be much smaller than
the B20 term, and attempts to fit these parameters did not
produce a significantly better model of the experimental
data; in fact more significant improvements in fit were ob-
tained by modelling the line widths anisotropically. The
zero-field splitting in 3 is almost an order of magnitude
larger than has been reported for a related anionic bis-
phthalocyaninato gadolinium complex,[22] whose Dy, Tb[23]
and Ho[24] analogues have been shown to behave as mono-
Also notable in the 31P{1H} NMR of complex 2 is the ob-
servation of a 15 Hz coupling between yttrium (89Y, I=1= )
2
3
and phosphorus. This JP, Y value is large considering that
1JP, Y values are typically in the range of 50–80 Hz,[18] and
2JP, Y values are typically 4–6 Hz, though values as large as
11 Hz have been reported in conjugated systems.[19] The
3
largely ionic nature of bonding to Y suggests that JP,Y cou-
pling constants should be smaller, due to smaller Fermi con-
3
tact terms, but we could find no JP, Y values in the literature
for comparison. This suggests that coupling between Y and
P could be mediated by a through-space interaction. The X-
ray data suggest an approximate Y···P distance of 3.36 Š,
and the proximity of the P and Y atoms could allowa for-
mally bonding interaction to occur between the minor lobe
of the phosphorus lone pair and the metal centre.
Ab initio calculation with the DFT B3YLP method and
DGDZVP basis sets was used to test this theory. This calcu-
lation predicts a JP, Y value of ꢁ11.2 Hz whose absolute value
is close to the 15 Hz observed experimentally, and the Fermi
contact term of ꢁ10.2 Hz predominates. The sign of this
value is consistent with direct interaction between nuclei
mediated by a pair of electrons and the opposite signs of the
magnetogyric ratios for 31P and 89Y. The paramagnetic spin–
Chem. Eur. J. 2008, 14, 721 – 730
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
723

S. A. Johnson et al.
tance of this anisotropy in any observed exchange coupling
is not readily ascertained. Regardless, the nearly tempera-
ture independent value of cmT in complex 3 facilitates the
observation of spin–spin exchange coupling in heterobime-
tallic complexes prepared using this precursor.
DFT-calculated spin density: A DFT geometry optimization
was performed on 3, and the resultant calculated spin densi-
ty is shown in Figure 6. The unpaired spins are primarily lo-
Figure 5. X-band EPR spectrum of a powdered sample of 3 at 77 K
(lower trace) and a simulated spectrum (upper trace) with S=7/2, B20
=
ꢀ0.194 cmꢁ1 (ꢀ2080 G), g=1.994 and anisotropic line widths (k=240,
?=360 G).
nuclear single-molecule magnets due to the large ligand-in-
duced zero-field splittings of their ground states. The ap-
proximate energy difference between the lowest and highest
sublevels at zero field for complex 3 is 2.328 cmꢁ1; however,
the relative sign of B20 cannot be determined from these
EPR data, and thus it is not clear whether the S=ꢀ1/2 or
S=ꢀ7/2 substates are lowest in energy.
3
Figure 6. The 0.0001 eŠ isosurface of the calculated spin density for 3.
cated on the GdIII centre, which is expected due to the con-
tracted nature of the f orbitals; however, some spin density
of the opposite sign is found on the amido and carbonyl
donor atoms. Only a very small contribution from unpaired
spin density is found on the phosphorus atom or the carbon
atoms. To use complex 3 as a magnetic building block, it is
necessary to observe exchange coupling between the gadoli-
nium centre and a transition metal bound to the available
phosphine donor. The localised nature of the unpaired elec-
trons on gadolinium necessitates a careful choice of transi-
tion metal to attach to the phosphine; it must be capable of
delocalizing its unpaired electron density towards the gadoli-
nium centre.
Magnetic susceptibility of 3: The molar magnetic susceptibil-
ity cm of a powdered sample of complex 3 immobilised in ei-
cosane was studied over the temperature range of 300.0–
2.0 K. A plot of cmT versus T for 3 is provided in the Sup-
porting Information. The cmT value of 3 at room tempera-
ture of 7.82 cm3 Kmolꢁ1 corresponds to the expected value
for a Gd3+ ion at room temperature (7.88 cm3 Kmolꢁ1), and
is also in agreement with the value of 7.8 cm3 Kmolꢁ1 in tol-
uene solution at room temperature determined by Evansꢁs
method. The magnitude of cmT for the powdered sample of
3 decreases slightly below15 K, and reaches a value of
7.43 cm3 Kmolꢁ1 at 2 K. The modelled temperature depend-
ence of cmT obtained using the negative value of the B20
crystal field value (i.e., ꢁ0.194 cmꢁ1) obtained from the sim-
ulation of the EPR data is also provided in the Supporting
Information. This model predicts a slight drop in cmT at low
temperatures, as is observed, but the fit is not sufficient to
determine the sign of B20, because a similar decrease in cmT
is predicted when a positive B20 value is used.
Synthesis of heterobimetallic complexes: To demonstrate
that through-space exchange coupling is viable in heterobi-
metallic complexes of these tripodal amidophosphine li-
gands, a transition metal complex must be attached to the
phosphine donor that satisfies two requirements: Firstly, it
must direct unpaired electron density towards the com-
plexed gadolinium centre; secondly, it must have magnetic
properties that are easy to model. We chose [Co
(TPP=5,10,15,20-tetrakis(4-methoxyphenyl)porphine)
ACHTREUNG
The calculated magnetic anisotropy at lowtemperatures
for this magnitude of B20 is significant at temperatures
below20 K. At 1 K ckmT and c?mT are predicted to be 15.74
and 3.31 cm3 Kmolꢁ1, respectively, when B20 is ꢁ0.194 cmꢁ1.
the transition metal moiety, because it has a single unpaired
electron in a nondegenerate orbital and thus should exhibit
straightforward magnetism. Additionally, the SOMO in ad-
ducts of [Co
ACHTREUNG
For the corresponding positive value of B20, the ckmT and c
pendicular to the plane of the [Co
ACHTREUNG
?
mT values at 1 K are 2.60 and 10.07 cm3 Kmolꢁ1, respectively.
The common assumption that Gd3+ in a crystal field can be
treated as magnetically isotropic is clearly not valid in this
system. Unfortunately, this complicates the modelling of this
building block in heterobimetallic complexes, and the impor-
should directly overlap with the phosphine donor orbital.
Related five-coordinate phosphine adducts of CoII porphyr-
ins with diamagnetic phosphine donors are known.[25]
Complexes 2 and 3 react with [Co
A
provide brownish-purple crystalline solids which were iden-
724
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 721 – 730

Through-Space Coupling in a Lanthanide–Transition Metal Complex
FULL PAPER
[26]
tified
as
the
heterobimetallic
complexes
large range, from as short as 2.2127(8) Š
to as long as
[(TPP)Co{P(CH2NC6H4-2-CO2Me)3}Ln] [Ln=Y (4) or Gd
(5); Eq. (3)].
2.479(5) Š.[27] The relatively elongated Co P bond length in
5 can be rationalised from a crystal-field theory argument
that the lone unpaired electron in this d7 species occupies
the dz2 orbital, which is directed towards the phosphine
donor. The ionic radius of GdIII is approximately 0.04 Š
larger than that of YIII,[28] and this results in slightly different
ꢁ
ꢁ
bond lengths between 4 and 5. In 5, the average Gd N and
ꢁ
Gd O distances are 2.330(2) and 2.315(2) Š, respectively,
ꢁ
ꢁ
whereas in 4 the average Y N and Y O bond lengths are
2.299(1) and 2.2771(9) Š, respectively. The P(1)···Y(1) dis-
tance of 3.2575(6) Š in 4 is less than 7% longer than the
[29]
ꢁ
longest reported Y P bond length of 3.045(2) Š, although
[18,30]
ꢁ
shorter Y P bond lengths are more typical.
In compari-
son, the P(1)···Gd(1) distance of 3.2440(9) Š in 5 is actually
shorter than the P(1)···Y(1) distance in 4, which can be ra-
tionalised by the slightly larger size of the GdIII ion requir-
ing a greater strain in the ancillary ligand fro its accommo-
dation. This assertion is confirmed by the slightly larger sum
of C-P-C angles in 5 than in 4, which are 326.70(26) and
324.69(17)8, respectively. These values are approximately
288 larger than the sum of C-P-C angles for ligand precursor
1, which can be attributed primarily to the coordination of
the large lanthanide ions.[9]
Single crystals of 4 and 5 were obtained by slow evapora-
tion of saturated toluene solutions. The initial solubility of
these complexes in toluene appears kinetic, as the resultant
product has only modest solubility. The solid-state structures
of 4 and 5 were determined by X-ray crystallography, and
an ORTEP depiction of the solid-state molecular structure
of 5 is shown in Figure 7. As anticipated, the triamido
EPR spectra of 4 and 5: The X-band EPR spectrum of a
frozen toluene solution of 4 was obtained at 77 K (Figure 8).
It confirms that the complex remains a 1:1 adduct of 2 and
Figure 8. X-band EPR spectrum of a powdered sample of 4 at 77 K (solid
line) and a simulated spectrum (dotted line, offset above) obtained by
using gk =1.98, g? =2.21, APk =176 G, AP ? =144 G, ACok =50 G, ACo ?
=
63 G.
Figure 7. Solid-state molecular structure of 5 as determined by X-ray
crystallography. Hydrogen atoms are omitted for clarity. Selected distan-
ꢁ
ꢁ
ꢁ
ces [Š]: P(1) Co(1) 2.4128(9), Gd(1) N(1) 2.325(3), Gd(1) N(2)
[Co
(TPP)] in solution, as well as the low spin S=1= nature
2
A
ꢁ
2.332(3), Gd(1) N(3) 2.336(3), Gd(1)···P(1) 3.2440(9).
of the complex. A simulated spectrum was obtained by con-
sidering the cobalt centre in 4 to have approximate axial
symmetry and was fitted by using anisotropic gk and g?
values of 1.98 and 2.21, respectively. The simulation re-
vealed that the phosphorus superhyperfine coupling con-
stants APk and AP ? of 176 and 144 G, respectively, are
larger than the cobalt (59Co, I=7/2) hyperfine coupling con-
stants ACok and ACo ? of 50 and 63 G, respectively. The simu-
lated spectrum is also shown in Figure 8. The EPR signal in
solution at room temperature is a doublet due to superhy-
donors chelate the lanthanide centre, and the phosphine
lone pair is bound to the Co(TPP) moiety. Although 4 and 5
A
display slightly different metrical parameters, they are iso-
structural. These structures confirm the connectivity deter-
mined for the mononuclear complexes 2 and 3; the tripodal
ligand chelates via the amido donors and the carbonyl
ꢁ
oxygen atoms. The Co P bond length in complex 5 is
2.4128(9) Š. Reported CoII–phosphine distances cover a
Chem. Eur. J. 2008, 14, 721 – 730
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
725

S. A. Johnson et al.
perfine coupling to 31P, and ACo cannot be resolved, which
further confirms that in these adducts considerable unpaired
electron density resides on the phosphine donor.
and 4.9% residing in the phosphorus 3p orbital.
ACok ¼ K þ P½4=7ꢁð4=7Þc3ꢁð6=7Þc1 þ ð2=63Þ c32
ꢁð64=63Þ c25 þ ð30=14Þ c21 þ ð8=21Þ c5c3ꢃ
It has previously been shown[31] in related complexes that,
by using third-order perturbation theory, the g values can be
ð7Þ
ð8Þ
ð9Þ
2
used to determine the energy separation of the A1 ground
ACo? ¼ K þ P½ꢁ2=7 þ ð2=7Þ c3 þ ð45=7Þ c1
2
4
4
state and the E, A2, and E excited states in phosphine ad-
ducts of cobalt porphyrins. In this complex the lowest
energy d orbitals should be the degenerate dxz and dyz orbi-
tals, followed by the dxy orbital and the singly occupied dz2
orbital. The dx2ꢁy2 orbital is highest in energy and is unoccu-
pied in the ground state. The c1, c3 and c5 parameters are re-
lated to the gk and g? values as shown in Equations (4) and
þð12=63Þ c23 þ ð40=63Þ c52ꢁð57=14Þ c21ꢁð34=21Þ c5c3ꢃ
P ꢂ ðACokꢁACo?Þ=½6=7ꢁð6=7Þ c3ꢁð51=7Þ c1
þð12=63Þ c23 þ ð87=14Þ c21ꢃ
A DFT calculation was performed on 4 using the solid-
state structural data obtained by X-ray crystallography. The
spin density predicted from this calculation is shown in
Figure 9. Consistent with the EPR spectrum of 4, this analy-
(5). Here c1 =x/(E[2E]ꢁE
C
G
2
orbit coupling constant. The E excited configuration results
from promotion of an electron from doubly degenerate dxy
and dyz to the dz2 orbital. Similarly, parameters c3 =x/
(E[4E]ꢁE
C
[4A2]ꢁE
[2A1]) correspond to
4
4
as phosphines, the A2 configuration, which results from pro-
motion of an electron from the dxy orbital to the dx2ꢁy2 orbi-
4
tal, is similar in energy to the E state, which results from
promotion of an electron from the doubly degenerate dxy
and dyz orbitals to the dx2ꢁy2 orbital. This is because the dxz
and dyz orbitals are relatively close in energy to the dxy orbi-
tal, whereas the dx2ꢁy2 orbital is significantly higher in
energy. The approximation that c3 ꢂc5 is therefore reasona-
ble,[32,33] and a simplified expression for g? is obtained
[Eq. (6)]. For complex 4 this analysis results in the values
c1 =0.039 MHz and c3 =0.11 MHz, which is consistent with
previously reported data for related complexes.[33]
3
Figure 9. A depiction of the 0.0004 eŠ isosurface of the calculated spin
density for heterodinuclear yttrium cobalt complex 4 (left) and a simplifi-
cation of the antibonding interaction associated with the SOMO (right).
gk ¼ 2:0023 þ 2 c23ꢁ3 c12
ð4Þ
ð5Þ
ð6Þ
sis reveals that considerable unpaired electron density re-
sides on the phosphorus atom. Notably, this unpaired elec-
tron density extends back towards the Y centre, which
should allowfor through-space exchange coupling in com-
plexes where Y is replaced by a paramagnetic lanthanide.
The extensive delocalisation of the unpaired electron densi-
ty onto the phosphorus atom and towards the yttrium centre
can be rationalised by considering the nature of the SOMO,
which can be approximated as an antibonding interaction
between the phosphine lone pair, which has contributions
from both the 3s and 3p orbitals of the phosphorus atom,
and the dz2 orbital of the cobalt centre. A simplified depic-
tion of this interaction is also shown in Figure 9.
The X-band EPR spectrum of a powdered sample of 5 at
77 K is provided in the Supporting Information. The spec-
trum cannot be modelled to a reasonable fit by assuming no
interaction between the CoII and GdIII centres for any com-
bination of the B20 and B22 crystal-field parameters. The
complete spin Hamiltonian for 5 would have to take into ac-
count the crystal-field splitting of the Stark substates and ex-
change coupling between the two metal centres. This is com-
plicated by the presence of only pseudo-axial symmetry in 5,
as well as the significant exchange coupling interaction be-
tween the unpaired electrons associated with CoII and GdIII.
g? ¼ 2:0023 þ 6 c1ꢁ6c21 þ ð2=3Þ c32 þ ð8=3Þ c25ꢁð4=3Þ c3c5
g? ꢂ 2:0023 þ 6 c1ꢁ6 c21 þ 2 c32 ðc3 ꢂ c5Þ
This analysis can be extended to the evaluation of the hy-
perfine and superhyperfine coupling constants.[31] The ex-
pressions for the hyperfine coupling constants ACok and
ACo ? are given in Equations (7) and (8) and can be used to
estimate values of P, the dipolar term for the hyperfine cou-
pling constant and K, the Fermi contact term. An expression
for P under the assumption c3 ꢂc5 is shown in Equation (9).
The P value for complex 4 was determined to be 683 MHz,
which is very close to the value of 689 MHz for the free
ion[34] and is indicative of primarily dz2 spin density on the
CoII centre, as was anticipated. The value of
K is
ꢁ187 MHz. This analysis predicts that the sign of ACo ? is
negative.[31] Analysis of the superhyperfine coupling con-
stants to 31P allows evaluation of the contribution of the
phosphorus 3s and 3p orbitals to the unpaired spin density
on the phosphine donor.[33] This breakdown estimates that
12% of the unpaired spin density resides on the phosphine
donor, with 4.5% associated with the phosphorus 3s orbital
726
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 721 – 730

Through-Space Coupling in a Lanthanide–Transition Metal Complex
FULL PAPER
The exchange interaction is more easily estimated from the
magnetic susceptibility data.
These data can be modelled by using the isotropic spin
Hamiltonian given in Equation (10), where SGd and SCo are
the spin operators associated with GdIII and CoII, and J is
the magnetic exchange constant. The simplified Hamiltonian
results in an expression that can be used to fit the tempera-
ture dependence of cmT by using the coupling constant J, as
shown in Equation (11), where N represents Avogadroꢁs
number, b the Bohr Magneton, and k Boltzmannꢁs constant.
A J value of ꢁ2.1 cmꢁ1 was determined from this analysis,
and the modelled fit is provided in the Supporting Informa-
tion. This J value corresponds to an separation of the S=3
ground state and the S=4 state of 8.4 cmꢁ1.
Magnetic properties of 4 and 5: The value of cmT for a tolu-
ene solution of complex 4 at 298 K was difficult to measure
by Evansꢁs method, due to the lowsolubility of 4 and the
significant diamagnetic contribution cancelling the majority
of the paramagnetic contribution to the susceptibility. The
uncorrected value was determined to be 0.36 cm3 Kmolꢁ1,
which is close to the value of 0.375 cm3 Kmolꢁ1 expected for
a species bearing a single unpaired electron in a nondegen-
erate orbital, but after subtraction of the diamagnetic contri-
bution a corrected value of cmT of 0.59 cm3 Kmolꢁ1 was ob-
tained. The molar magnetic susceptibility of 4 was measured
over the temperature range of 1.8–300 K in an applied mag-
netic field of 1 T, and cmT (corrected for a slight tempera-
ture-independent paramagnetism of 1.38”10ꢁ3 cm3 molꢁ1)
was found to be independent of temperature, with a value
of 0.39 cm3 Kmolꢁ1 at 1.8 K. The simplicity of the magnetism
of this species renders the observation of exchange coupling
in 5 facile, because in the absence of coupling negligible de-
pendence of cmT on temperature is predicted.
At room temperature, the anticipated value of cmT for
complex 5 is 8.26 cm3 Kmolꢁ1. This corresponds to the value
for two uncoupled isolated ions GdIII (S=7/2, cmT=
7.88 cm3 Kmolꢁ1) and CoII (S=1/2, cmT=0.375 cm3 Kmolꢁ1).
The value of cmT for a solution of 5 in toluene at 298 K de-
termined by Evansꢁs method was 8.40 cm3 Kmolꢁ1, which is
approximately that anticipated in the absence of coupling.
The molar magnetic susceptibility of a powdered sample of
complex 5 was measured over the temperature range of 1.8–
300 K in an applied magnetic field of 0.01 T. The plot of cmT
versus temperature for 5 is shown in Figure 10. At room
^
^
^
ð10Þ
ð11Þ
Hex ¼ ꢁJSGd ꢄ SCo
ꢀ
ꢁ
4Nb2g2 15 þ 7 eꢁ4J=kT
cmT ¼
9 þ 7 eꢁ4J=kT
k
^
^
^
^
Hex ¼ gGdb½BzSzGd þ ð1=2ÞðBþSꢁGd þ BꢁSþGdÞꢃ
^
^
^
þb½gkCoBzSzCo þ ð1=2Þg?CoðBþSꢁCo þ BꢁSþCoÞꢃ
þB20ðSzGdꢁ21=4ÞꢃꢁJkSzGd
ꢄSzCoꢁð1=2ÞJ?½SþGd ꢄ SꢁCo þ SꢁGd ꢄ SþCo
ð12Þ
2
^
^
^
^
^
^
^
ꢃ
A more complete anisotropic Hamiltonian which accounts
for the anisotropy at the Co and Gd centres is given in
Equation (12), where the field in the z direction is repre-
sented by Bz, and the + and ꢁ subscripts represent raising
and lowering operators, respectively (Bꢀ =Bx ꢀiBy). The
magnetic susceptibility modelled with this Hamiltonian is
drawn as a solid line in Figure 10. The assumption that both
Jk and J? are equal to ꢁ2.1 cmꢁ1 provides a good fit; how-
ever, although the fit is strongly influenced by Jk, small
changes in J? do not result in large changes in the predicted
temperature dependence of cmT. The gkCo, g?Co, gGd and B20
values used were those obtained from the analysis of the
EPR spectra of complexes 3 and 4, though the fit is also not
particularly sensitive to small changes in B20. The calculated
c
kmT and c?mT contributions obtained from the simulation
are shown as dotted lines. This analysis predicts that the
lower energy S=3 state and higher energy S=4 state pre-
dicted from the isotropic model are both split by zero-field
splitting, into ꢀ3, ꢀ2, ꢀ1, 0 and ꢀ4, ꢀ3, ꢀ2, ꢀ1, 0 sub-
states. The zero-field splitting of the S=4 state is approxi-
mately (3/4)B20, and that of the S=3 state is about (5/4)B20,
in accordance with theoretically predicted values.[35] These
energy levels and their field dependences are included in
the Supporting Information.
^
Figure 10. Plots of cmT versus T for 5 ( ) and the simulated fit obtained
using gCok =1.98, gCo ? =2.21, gGd =1.994, B20 =ꢁ0.194 cmꢁ1 and Jk =J?
=
ꢁ2.1 cmꢁ1, shown as a solid line. The calculated ckmT and ckmT contribu-
tions obtained from the simulation are shown as dotted lines.
The antiferromagnetic nature of the exchange coupling
between Co and Gd is opposite to that most typically ob-
served,[10,11] and provides support for the suggestion that the
exchange mechanism involves direct overlap of the magnetic
orbitals which contain the unpaired electrons associated
with the CoII and GdIII centres, primarily due the extensive
delocalisation of the magnetic orbitals associated with CoII
over the phosphorus donor. The magnitude of J found here
temperature, the experimental value of cmT is close to that
of 8.26 cm3 Kmolꢁ1 predicted in the absence of significant
coupling. This value decreases at lowtemperatures and
reaches a value of 6.20 cm3 Kmolꢁ1 at 1.8 K, which suggests
weak antiferromagnetic coupling between the S=1/2 and
S=7/2 centres that results in an S=3 ground state, which
should have a cmT value of 6.0 cm3 Kmolꢁ1.
Chem. Eur. J. 2008, 14, 721 – 730
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
727

S. A. Johnson et al.
constant cross-sectional area. Elemental analyses were performed by the
Centre for Catalysis and Materials Research (CCMR) at the University
of Windsor. The compounds tris(hydroxymethyl)phosphine, methyl an-
thranilate, [5,10,15,20-tetrakis(4-methoxyphenyl)-21H,23H-porphine]co-
balt(II) and anhydrous yttrium trichloride were purchased from Aldrich.
Anhydrous GdCl3 was purchased from Strem. All the reagents were used
is comparable to values determined for other d–f heterobi-
metallics,[10] which demonstrates that through-space ex-
change coupling between transition metal and lanthanides
mediated by this tripodal ligand is equally as effective as the
superexchange mechanism which commonly operates in d–f
complexes in which bridging donor atoms are shared by the
transition metal and lanthanide. This result suggests that
paramagnetic tripodal complexes of the transition metals or
lanthanides with analogous supporting ligands should be ef-
fective as magnetic building blocks.
without further purification. The compounds Y[N
(SiMe3)2]3 were synthesised by literature methods.[40]
Synthesis of P(CH2NHC6H4-2-CO2Me)3 (1): A mixture of P(CH2OH)3
A
AHCTREUNG
A
(5 g, 0.040 mol), methyl anthranilate (30.45 g, 0.20 mol) and toluene
(70 mL) were mixed in a 250-mL three-neck flask equipped with Dean–
Stark trap and a condenser. The solution was heated to reflux for 1 h and
the water produced was removed azeotropically. After cooling to room
temperature the solvent was evaporated to dryness under vacuum, and
the creamy white residue was rinsed with diethyl ether 2–3 times to
remove excess methyl anthranilate. The product was then collected by fil-
tration and dried under vacuum. Yield: 20 g, 95%. X-ray quality crystals
Conclusion
Even though calculations suggest that the GdIII complex
[Gd{P(CH2NC6H4-2-CO2Me)3}] has negligible spin density
on the phosphorus atom, with the appropriate choice of
transition metal complex it proved possible to observe mag-
netic exchange coupling in the heterobimetallic complex
[(TPP)Co{P(CH2NC6H4-2-CO2Me}3Gd]. This coupling is
mediated by delocalisation of the spin density of the cobalt
centre onto the phosphine donor, which allows direct over-
lap of the magnetic orbital associated with CoII with the f
electrons on the GdIII centre. Contrary to what is typically
observed in d–f complexes, where ligand superexchange
pathways usually operate, this through-space interaction re-
sults in antiferromagnetic coupling. The magnitude of this
exchange coupling is equal to those of other d–f complexes,
which bodes well for potential use of the mononuclear lan-
thanide complexes of this ligand as building blocks for
larger polymetallic complexes in which through-space inter-
actions yield magnetically ordered systems or single-mole-
cule magnet behaviour.
were obtained by slow evaporation of a solution in benzene and hexame-
2
thyldisiloxane. 1H NMR (C6D6, 300 MHz, 298 K): d=3.36 (d, JP, H
=
5.1 Hz, 6H, PCH2), 3.46 (s, 9H, CH3), 6.49 (dd, 3H, ArH), 6.71(d, 3H,
ArH), 7.15 (ddd, 3H, ArH), 7.96 (dd, 3H, ArH), 8.29 ppm (br, 3H, NH);
13C{1H} NMR (C6D6, 75.5 MHz, 298 K): d=38.1 (d, JP, C =15.4 Hz, PCH2),
51.1 (s, CH3), 110.9, 112.1, 115.3, 131.9 and 134.8 (s, Ar-C), 151.6 (d, J=
2.7 Hz ipso-C), 169.1 ppm (s, CO2); 31P{1H} NMR (C6D6, 121.5 MHz,
298 K): d=ꢁ33.6 ppm (s); elemental analysis calcd (%) for
C27H30N3O6P: C 61.94, H 5.78, N 8.03; found: C 61.90, H 5.68, N 8.13.
Synthesis of [Y{P(CH2NC6H4-2-CO2Me)3}] (2):
P(CH2NHC6H4CO2Me)3 (1.00 g, 1.91 mmol) and [Y{N
A
mixture of
(SiMe3)2}3]
ACHTREUNG
(1.633 g, 2.86 mmol) was stirred in toluene (70 mL) for 5 h. The resultant
yellow crystalline precipitate was isolated by filtration, rinsed with 50 mL
pentane and dried for 4 h (77%, 1.45 g). 1H NMR (C6D6, 300 MHz,
2
298 K): d=3.25 (s, 9H, CH3), 3.92 (d, JP, H =7.1 Hz, 6H, PCH2), 6.46 (dd,
3JH,H =8.1, 6.6 Hz, 3H, C6H4 5-H), 6.78 (d, 3JH,H =8.8 Hz, 3H, C6H4 3-H),
7.27 (ddd, 3JH,H =8.8, 6.6 Hz, 4JH,H =1.8 Hz, 3H, C6H4 4-H), 8.1070 ppm
(dd, 3JH,H =8.1, 4JH,H =1.8 Hz, 3H, C6H4 6-H); 13C{1H} NMR (C6D6,
75.5 MHz, 298 K): d=38.1 (d, JP,C =15.4 Hz, PCH2), 51.5 (s, CH3), 172.1
(s, CO2), 108.7, 112.1, 114.4, 132.9 and 136.6 (s, Ar-C), 153.6 ppm (ipso-
C); 31P{1H} NMR (C6D6, 121.5 MHz, 298 K): d=ꢁ57.0 ppm (d, JP, Y
=
15.1 Hz); elemental analysis calcd (%) for C27H27N3O6PY: C 53.21, H
4.47, N 6.90; found: C 53.10, H 4.45, N 6.96.
Synthesis of [Gd{P(CH2NC6H4-2-CO2Me)3}] (3):
P(CH2NHC6H4CO2Me)3 (500 mg, 0.938 mmol) and [Gd{N
A
mixture of
(SiMe3)2}3]
ACHTREUNG
Experimental Section
(600 mg, 0.938 mmol) was stirred in toluene (20 mL) for 30 min. The so-
lution was filtered and remaining yellow crystalline solid was rinsed with
pentane (50 mL) and dried for 4 h (67%, 425 mg). Elemental analysis
calcd (%) for C27H27N3O6PGd: C 47.04, H 3.95, N 6.09; found: C 47.28,
H 4.02, N 6.24.
General procedures: Unless otherwise stated, all experiments were per-
formed under an inert atmosphere of nitrogen using either Schlenk tech-
niques or an MBraun glove box. Dry oxygen-free solvents were used
throughout. Anhydrous pentane and toluene were purchased from Al-
drich, sparged with nitrogen and passed through activated alumina under
a positive pressure of nitrogen gas; toluene and hexanes were further de-
oxygenated on Ridox catalyst columns.[36] Deuterated benzene was dried
by heating at reflux over potassium in a sealed vessel under partial pres-
sure, then trap-to-trap distilled and freeze–pump–thawdegassed three
times. 1H, 13C{1H} and 31P{1H} NMR spectra were recorded on a Bruker
AMX (300 MHz) or Bruker AMX (500 MHz) spectrometer. 1H NMR
spectra were referenced to residual protons (C6D5H, d=7.15) with re-
spect to tetramethylsilane at d=0.0. 13C{1H} spectra were referenced to
solvent resonances (C6D6, d=128.0). 31P{1H} NMR spectra referenced to
external 85% H3PO4 at d=0.0. EPR spectra of all solid samples were
collected on an X-band Bruker ESR 300E spectrometer. The program
Simpip[37] was used to model the CoII spectra of complex 4. The program
Spin[38] was used to simulate the Gd3+ spectrum of 3 using only the B20
crystal field parameter, and the program Sim[39] was used to generate
spectra with the B20, B40 B43, B60, B63 and B66 crystal-field parameters.
Unless otherwise noted, magnetizations were measured at 100 G with a
Quantum Design Evercool MPMS-XL system. Corrections for the dia-
magnetic contributions of compounds were made by using Pascalꢁs con-
stants. Samples were run in a PVC holder specially designed to have a
Synthesis of [(TPP)Co{P(CH2NC6H4-2-CO2Me)3}Y] (4): A mixture of
[Y{P(CH2NC6H4-2-CO2Me)3}] (450 mg, 0.738 mmol) and [5,10,15,20-tet-
rakis(4-methoxyphenyl)porphinato]cobalt(II) (584.67 mg, 0.737 mmol)
was stirred in toluene (25 mL) for 30 min. The solution was filtered and
the resultant reddish-purple crystalline solid was washed with pentane
(50 mL) and dried for 4 h (65.2%, 675 mg). X-ray quality crystals were
obtained by performing the reaction without stirring, and the structure
contained 2 equivalents of co-crystallised toluene. The complex is spar-
ingly soluble in toluene and benzene. 1H NMR (C6D6, 300 MHz, 298 K):
d=3.2 (br, 9H, CO2CH3), 4.4 (br, 18H total, OCH3 and PCH2), 5.8 (br,
3H, C6H4), 6.9 (br, 3H, C6H4), 7.3 (br, 3H, C6H4), 8.2 (br, 3H, C6H4), 8.8
(br, 8H, TPP m-H), 11.5 (vbr, 8H, TPP o-H), 15.1 ppm (v br, 8H, pyrrole
H); elemental analysis calcd (%) for C75H63N7O10PYCo: C 64.29, H 4.53,
N 7.00; found: C 64.23, H 4.82, N 6.81.
Synthesis of [(TPP)Co{P(CH2NC6H4-2-CO2Me)3}Gd] (5): A mixture of
[Gd{P(CH2NC6H4-2-CO2Me)3}] (500 mg, 0.737 mmol) and [5,10,15,20-tet-
rakis(4-methoxyphenyl)porphinato]cobalt(II),
[CoII
A
(584 mg,
0.737 mmol) was stirred in toluene (25 mL) for 30 min. The solution was
filtered and the resultant reddish purple crystalline solid was washed with
pentane (50 mL) and dried for 3–4 h (47.5%, 515 mg); elemental analysis
728
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 721 – 730

Through-Space Coupling in a Lanthanide–Transition Metal Complex
FULL PAPER
Table 1. X-ray crystallographic data for 1–5.
1
2
3
4
5
empirical formula
formula weight
crystal system
a [Š]
b [Š]
c [Š]
C30H33N3O6P
562.56
triclinic
C27H27N3O6PY
609.40
trigonal
14.6660(3)
14.6660(3)
20.9460(9)
90
90
120
3905.06(20)
R3c
6
1.56
2.349
C27H27N3O6PGd
708.7
trigonal
14.6765(11)
14.6660(11)
21.2161(3)
90
90
120
3957.7(7)
R3c
6
1.70
2.684
C89H79N7O10PCoGd
1653.74
triclinic
14.442(2)
15.495(2)
18.737(3)
90.648(2)
112.571(2)
89.894(2)
C89H79N7O10PCoY
1585.46
triclinic
14.4404(19)
15.509(2)
18.736(2)
10.8108(17)
11.3382(18)
12.513(2)
104.725(2)
104.725(2)
90.718(2)
1429.3(4)
a [8]
b [8]
90.2840(10)
112.6130(10)
90.1050(10)
3873.5(10)
g [8]
V [Š3]
3871.5(10)
¯
¯
¯
space group
Z
P1
P1
P1
2
1.31
0.144
2
1.36
1.150
2
1.42
1.046
1calcd [gcmꢁ1
]
m
A
]
T [K]
173
173
173
173
173
2qmax [8]
min./max. transmission
total reflns
50.0
0.9473/0.9420
16454
50.0
0.791/0.584
13993
50.0
55.0
0.708/0.811
43148
52.5
0.818/0.977
44292
not fully solved[a]
13580
unique reflns (residue)
parameters
R1; wR2 (all data)
6348 (Rint =0.0439)
356
0.0624; 0.1800
1530 (Rint =0.0867)
39
1999 (Rint =0.0821)
not fully solved[a]
not fully solved[a]
17077 (Rint =0.0323)
991
0.0397; 0.1115
7230 (Rint =0.0258)
991
0.0391; 0.1041
0.1372; 0.4167[a]
[a] Twinning complicated solutions of structures for 2 and 3.
calcd (%) for C75H63N7O10PGdCo: C 61.30, H 4.32, N 6.67; found: C
61.47, H 4.12, N 6.54.
Acknowledgements
Acknowledgement is made to the National Sciences and Engineering Re-
search Council (NSERC) of Canada and the Ontario Research and De-
velopment Challenge Fund (ORDCF) for their financial support.
X-ray crystallography: Each crystal was covered with Paratone, mounted
on a glass fibre and rapidly placed into the cold N2 stream of the Kryo-
Flex low-temperature device. The data were collected using SMART[41]
software on a Bruker APEX CCD diffractometer with graphite-mono-
chromated MoKa radiation (l=0.71073 Š). Details of crystal data, data
collection and structure refinement are listed in Table 1. Data reduction
was performed using SAINT[42] software, and the data were corrected for
absorption using SADABS.[43] The structures were solved by direct meth-
ods using SIR97[44] and refined by full-matrix least-squares on F2 using
SHELXL-97[45] and the WinGX[46] software package, and the thermal el-
lipsoid plots were produced using ORTEP32.[47] In general, thermal pa-
rameters for non-hydrogen atoms were treated anisotropically, and all hy-
drogen atoms were placed in idealised locations. The co-crystallised ben-
zene solvent in 1 was modelled as an idealised hexagon, with equal iso-
tropic thermal parameters on the six carbon atoms, and the hydrogen
atoms associated with this disordered moiety were omitted. The hydro-
gen atoms associated with the three amino groups that were involved in
hydrogen bonding in 1 were located in an electron-density difference
map and their positions and isotropic thermal parameters were refined.
Multiple data sets were acquired for complexes 2 and 3, but all suffered
from twinning. All atoms in the solution for 2 were treated isotropically
and the arene ring was modelled as a perfect hexagon with each carbon
bearing identical thermal parameters. Only the P and Gd centres were
readily located in the attempted solution of the structure of 3.
[1] a) K. Itoh, M. Kinoshita, Molecular Magnetism: New Magnetic Ma-
terials, Taylor & Francis, London, 2002; b) O. Kahn, Molecular Mag-
netism, VCH, Weinheim, 1993; c) R. Jain, K. Kabir, J. B. Gilroy,
K. A. R. Mitchell, K.-C. Wong, R. G. Hicks, Nature 2007, 445, 291–
294.
[2] J. S. Miller, Dalton Trans. 2006, 2742–2749.
[3] J. S. Miller, A. J. Epstein, Angew. Chem. 1994, 106, 399–432; Angew.
Chem. Int. Ed. Engl. 1994, 33, 385–415.
[4] J. Mrozinski, Coord. Chem. Rev. 2005, 249, 2534–2548.
[5] L. M. C. Beltran, J. R. Long, Acc. Chem. Res. 2005, 38, 325–334.
[6] a) D. Li, S. Parkin, G. Wang, G. T. Yee, R. Clerac, W. Wernsdorfer,
S. M. Holmes, J. Am. Chem. Soc. 2006, 128, 4214–4215; b) C. P. Ber-
linguette, D. Vaughn, C. Canada-Vilalta, J. R. Galan-Mascaros, K. R.
Dunbar, Angew. Chem. 2003, 115, 1561–1564; Angew. Chem. Int.
Ed. 2003, 42, 1523–1526.
[7] a) O. Kahn, NATO Adv. Study Inst. Ser. Ser. C 1997, 499, 287–302;
b) L. M. Toma, L. D. Toma, F. S. Delgado, C. Ruiz-Perez, J. Sletten,
J. Cano, J. M. Clemente-Juan, F. Lloret, M. Julve, Coord. Chem.
Rev. 2006, 250, 2176–2193; c) T. E. Vos, J. S. Miller, Angew. Chem.
2005, 117, 2468–2471; Angew. Chem. Int. Ed. 2005, 44, 2416–2419;
d) S. Tanase, J. Reedijk, Coord. Chem. Rev. 2006, 250, 2501–2510;
e) B. D. Koivisto, R. G. Hicks, Coord. Chem. Rev. 2005, 249, 2612–
2630; f) H. Miyasaka, R. Clerac, Bull. Chem. Soc. Jpn. 2005, 78,
1725–1748; g) M. Pilkington, S. Decurtins, Persp. Supramol. Chem.
2003, 7, 275–323.
CCDC-638601 (1), CCDC-638602 (2), CCDC-638603 (4) and CCDC-
638604 (5) contain the supplementary crystallographic data for this
paper. These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
Calculations: Ab initio DFT calculations were performed by using the
hybrid-functional B3LYP or UB3LYP[48] method with the Gaussian03
package.[49] The basis functions used were the DGDZVP set for complex
2, and CEP-31G for complex 3. Both were optimised with C3 symmetry.
For the model complex of 4 the CEP-121G* basis set was used on all
atoms except for Y and Co, for which the CEP-121G basis set was used.
All these basis sets are provided in the Gaussian03 program suite.
[8] a) J. A. Hatnean, R. Raturi, J. Lefebvre, D. B. Leznoff, G. Lawes,
S. A. Johnson, J. Am. Chem. Soc. 2006, 128, 14992–14999; b) A. L.
Keen, M. Doster, H. Han, S. A. Johnson, Chem. Commun. 2006,
1221–1223; c) H. Han, S. A. Johnson, Organometallics 2006, 25,
5594–5602.
[9] H. Han, M. Elsmaili, S. A. Johnson, Inorg. Chem. 2006, 45, 7435–
7445.
Chem. Eur. J. 2008, 14, 721 – 730
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
729

S. A. Johnson et al.
[10] a) M. Sakamoto, K. Manseki, H. Okawa, Coord. Chem. Rev. 2001,
219–221, 379–414; b) C. Benelli, D. Gatteschi, Chem. Rev. 2002,
102, 2369–2387.
[25] a) B. B. Wayland, M. E. Abd-Elmageed, J. Am. Chem. Soc. 1974, 96,
4809–4814; b) B. B. Wayland, J. V. Minkiewixz, M. E. Abd-Elma-
geed, J. Am. Chem. Soc. 1974, 96, 2795–2801; c) B. De Castro, M.
Rangel, J. B. Raynor, J. Chem. Soc. Dalton Trans. 1990, 3311–3318.
[26] M. D. Fryzuk, D. B. Leznoff, R. C. Thompson, S. J. Rettig, J. Am.
Chem. Soc. 1998, 120, 10126–10135.
[27] D. C. Bradley, M. B. Hursthouse, R. J. Smallwood, A. J. Welch, J.
Chem. Soc. Chem. Commun. 1972, 872.
[28] R. D. Shannon, Acta. Crystallogr. Sect. A 1976, 32, 751–767.
[29] P. B. Hitchcock, M. F. Lappert, I. A. MacKinnon, J. Chem. Soc.
Chem. Commun. 1988, 1557–1558.
[30] P. W. Roesky, M. T. Gamer, N. Marinos, Chem. Eur. J. 2004, 10,
3537–3542.
[31] B. R. McGarvey, Can. J. Chem. 1975, 53, 2498–2511.
[32] W. C. Lin, Inorg. Chem. 1976, 15, 1114–1118.
[33] B. B. Wayland, A. E. Sherry, A. G. Bunn, J. Am. Chem. Soc. 1993,
115, 7675–7684.
[34] B. R. McGarvey, J. Phys. Chem. 1967, 71, 51–66.
[35] J. Owen, J. Appl. Phys. 1961, 32, S213–217.
[36] A. B. Pangborn, M. A. Giardello, R. H. Grubbs, R. K. Rosen, F. J.
Timmers, Organometallics 1996, 15, 1518–1520.
[37] Simpip, An EPR simulation program, M. Nilges, Illinois EPR Re-
search Centre.
[38] Spin, EPR simulation program, A. Ozarowski, National High Mag-
netic Field Laboratory, Florida.
[39] Sim, version 2004-5, An EPR Simulation Program, H. Weihe, De-
partment of Chemistry, University of Copehagen.
[40] a) D. C. Bradley, J. S. Ghotra, F. A. Hart, J. Chem. Soc. Chem.
Commun. 1972, 349–350; b) D. C. Bradley, J. S. Ghotra, F. A. Hart,
J. Chem. Soc. Dalton Trans. 1973, 1021–1023.
[41] SMART, Molecular analysis research tool, Bruker AXS Inc., Madi-
son, WI, 2001.
[42] SAINTPlus, Data reduction and correction program’; Bruker AXS
Inc.: Madison, WI, 2001.
[43] SADABS, An empirical absorption correction program, Bruker
AXS Inc., Madison, WI, 2001.
[11] a) A. Bencini, C. Benelli, A. Caneschi, R. L. Carlin, A. Dei, D. Gat-
teschi, J. Am. Chem. Soc. 1985, 107, 8128–8136; b) C. Benelli, A.
Caneschi, D. Gatteschi, O. Guillou, L. Pardi, Inorg. Chem. 1990, 29,
1750–1755; c) M. Sasaki, K. Manseki, H. Horiuchi, M. Kumagai, M.
Sakamoto, H. Sakiyama, Y. Nishida, M. Sakai, Y. Sadaoka, M.
Ohba, H. Okawa, J. Chem. Soc. Dalton Trans. 2000, 259–263; d) J.-
P. Costes, F. Dahan, B. Donnadieu, J. Garcia-Tojal, J.-P. Laurent,
Eur. J. Inorg. Chem. 2001, 363–365; e) N. Matsumoto, M. Sakamoto,
H. Tamaki, H. Okawa, S. Kida, Chem. Lett. 1990, 853–854; f) M.
Andruh, I. Ramade, E. Codjovi, O. Guillou, O. Kahn, J. C. Trombe,
J. Am. Chem. Soc. 1993, 115, 1822–1829; g) A. J. Blake, R. O.
Gould, C. M. Grant, P. E. Y. Milne, S. Parsons, R. E. P. Winpenny, J.
Chem. Soc. Dalton Trans. 1997, 485–495; h) J.-P. Costes, M. Auchel,
F. Dahan, V. Peyrou, S. Shova, W. Wernsdorfer, Inorg. Chem. 2006,
45, 1924–1934; i) J.-P. Costes, F. Dahan, A. Dupuis, J.-P. Laurent,
Inorg. Chem. 1996, 35, 2400–2402; j) C. Benelli, A. J. Blake, P. E. Y.
Milne, J. M. Rawson, R. E. P. Winpenny, Chem. Eur. J. 1996, 1, 614–
618; k) X.-M. Chen, Y.-L. Wu, Y.-X. Tong, X.-Y. Huang, J. Chem.
Soc. Dalton Trans. 1996, 2443–2448; l) J. F. Petit, J. C. Trombe, A.
Gleizes, J. Galy, C. R. Acad. Sci. Ser. II 1987, 304, 1117–1120; m) R.
Kempe, H. Noss, T. Irrgang, J. Organomet. Chem. 2002, 647, 12–20;
n) R. Koner, H.-H. Lin, H.-H. Wei, S. Mohanta, Inorg. Chem. 2005,
44, 3524–3536; o) J.-P. Sutter, M. L. Kahn, Magnetism: Molecules to
Materials V, Wiley-VCH, Weinheim, 2005, pp. 161–187.
[12] C. M. Zaleski, E. C. Depperman, J. W. Kampf, M. L. Kirk, V. L. Pe-
coraro, Angew. Chem. 2004, 116, 4002–4004; Angew. Chem. Int. Ed.
2004, 43, 3912–3914.
[13] F. Pointillart, K. Bernot, R. Sessoli, D. Gatteschi, Chem. Eur. J.
2007, 13, 1602–1609.
[14] V. S. Mironov, Y. G. Galyametdinov, A. Ceulemans, C. Gorller-Wal-
rand, K. Binnemans, J. Chem. Phys. 2002, 116, 4673–4685.
[15] a) A. Mishra, W. Wernsdorfer, S. Parsons, G. Christou, E. K. Bre-
chin, Chem. Commun. 2005, 2086–2088; b) A. Mishra, W. Werns-
dorfer, K. A. Abboud, G. Christou, J. Am. Chem. Soc. 2004, 126,
15648–15649; c) S. Osa, T. Kido, N. Matsumoto, N. Re, A. Pochaba,
J. Mrozinski, J. Am. Chem. Soc. 2004, 126, 420–421; d) M. Ferbin-
teanu, T. Kajiwara, K.-Y. Choi, H. Nojiri, A. Nakamoto, N. Kojima,
F. Cimpoesu, Y. Fujimura, S. Takaishi, M. Yamashita, J. Am. Chem.
Soc. 2006, 128, 9008–9009; e) J. Tang, I. Hewitt, N. T. Madhu, G.
Chastanet, W. Wernsdorfer, C. E. Anson, C. Benelli, R. Sessoli,
A. K. Powell, Angew. Chem. 2006, 118, 1761–1765; Angew. Chem.
Int. Ed. 2006, 45, 1729–1733; f) G. Christou, D. Gatteschi, D. N.
Hendrickson, R. Sessoli, MRS Bull. 2000, 25, 66–71; g) J.-P. Costes,
F. Dahan, W. Wernsdorfer, Inorg. Chem. 2006, 45, 5–7.
[16] A. W. Frank, G. L. Drake, Jr., J. Org. Chem. 1972, 37, 2752–2755.
[17] K. W. Feindel, R. E. Wasylishen, Can. J. Chem. 2004, 82, 27–44.
[18] M. D. Fryzuk, T. S. Haddad, D. J. Berg, Coord. Chem. Rev. 1990, 99,
137–212.
[19] a) S. O. Grim, S. A. Sangokoya, J. Chem. Soc. Chem. Commun.
1984, 1599–1600; b) A. A. Pinkerton, W. L. Earl, J. Chem. Soc.
Dalton Trans. 1979, 1347–1349.
[20] R. Stowasser, R. Hoffmann, J. Am. Chem. Soc. 1999, 121, 3414–
3420.
[21] C. A. Hutchison, Jr., B. R. Judd, D. F. D. Pope, Proc. Phys. Soc. B
1957, 70, 514–520.
[22] N. Ishikawa, N. Tanaka, Y. Kaizu, Inorg. Chim. Acta 2004, 357,
2181–2184.
[23] N. Ishikawa, M. Sugita, W. Wernsdorfer, Angew. Chem. 2005, 117,
2991–2995; Angew. Chem. Int. Ed. 2005, 44, 2931–2935.
[24] N. Ishikawa, M. Sugita, W. Wernsdorfer, J. Am. Chem. Soc. 2005,
127, 3650–3651.
[44] A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano, C. Giaco-
vazzo, A. Guagliardi, A. G. G. Moliterni, G. Polidori, R. Spagna, J.
Appl. Crystallogr. 1999, 32, 115–119.
[45] Sheldrick, G. M. SHELXL-97, Universitat Gottingen: Gottingen,
1997.
[46] L. J. Farrugia, J. Appl. Crystallogr. 1999, 32, 837–838.
[47] L. J. Farrugia, J. Appl. Crystallogr. 1997, 30, 565.
[48] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[49] Gaussian03, RevisionC.01, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomery,
Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar,
J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega,
G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R.
Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross,
V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann,
O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y.
Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg,
V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O.
Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Fores-
man, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski,
B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L.
Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Na-
nayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen,
M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wallingford
CT, 2004..
Received: March 2, 2007
Revised: October 3, 2007
Published online: November 14, 2007
730
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 721 – 730