Single-molecule magnets: Tetranuclear vanadium(III) complexes with a butterfly structure and an S = 3 ground state

Article abstract of DOI:10.1021/ja9732439

Reactions of VCl₃(THF)₃, bpy, and NaO₂CR (R = Et, Ph; bpy = 2,2'-bipyridine) in a 1:1:3 ratio in Me₂CO give [V₄O₂(O₂CR)₇(bpy)₂](ClO₄) (R = Et, 1; R = Ph, 4) following addition of NBu(n)₄ClO₄. Use of 4,4'-dimethyl- or 5,5'-dimethylbipyridine (4,4'-Me₂bpy and 5,5'-Me₂bpy, respectively) and R = Et leads similarly to [V₄O₂(O₂CEt)₇(L-L)₂](ClO₄) (L-L = 4,4'-Me₂bpy, 2; L-L = 5,5'-Me₂bpy, 3). Yields are in the 38-90% range. The cation of 1 is isostructural with previously prepared [M₄O₂(O₂CR)₇(bpy)₂]⁺ (M = Cr(III), Mn(III), Fe(III)) species and possesses a [V₄O₂] butterfly core. 1D and 2D COSY ¹H NMR spectra of 1 show the solid-state structure is retained on dissolution. The effective magnetic moment (μ(eff)) per V₄ for 1 gradually rises from 5.79 μ(B) at 300 K to a maximum of 6.80 μ(B) at 25.0 K and then decreases rapidly to 4.72 μ(B) at 2.00 K. The data in the 7.00-300 K range were fit to the appropriate theoretical expression (based on H = 2JS(i)·S(j)) to give J(bb) = -31.2 cm^-1, J(wb) = +27.5 cm^-1, and g = 1.82, (b = body, w = wingtip). These values indicate a S(T) = 3 ground state, confirmed by magnetization vs field studies. Similar results were obtained for the 2-picolinate (pic) analogue of 1 (complex 5). The S(T) = 3, 1, 3, and 0 ground states for the M = V(III), Cr(III), Mn(III), and Fe(III), respectively, are rationalized using spin frustration arguments based on competition between J(bb) and J(wb) interactions. AC magnetic susceptibility studies down to 1.7 K on 1 and 5 show weak out-of-phase signals (χ''(M)) below 4.0 K and corresponding small decreases in the in-phase signals (χ'(M)), indicating that the relaxation of magnetization is unusually slow and comparable with the oscillating AC field (250-1000 Hz). This is a characteristic signature of a single-molecule magnet. Simultaneous application of AC and DC fields has the effect of increasing the barrier to magnetization relaxation, causing the χ''(M) signal to move to higher temperature and consequently leading to a much stronger χ''(M) signal and, for 5, the observation of a peak at ~2.0 K. A dependence of the χ''(M) peak position of 5 on the DC field intensity and AC field oscillation frequency is found.

Full text of DOI:10.1021/ja9732439

J. Am. Chem. Soc. 1998, 120, 2365-2375
2365
Single-Molecule Magnets: Tetranuclear Vanadium(III) Complexes
with a Butterfly Structure and an S ) 3 Ground State
Stephanie L. Castro,1a Ziming Sun,1b Craig M. Grant,1a John C. Bollinger,1a
David N. Hendrickson,*,1b and George Christou*,1a
Contribution from the Department of Chemistry and Molecular Structure Center, Indiana UniVersity,
Bloomington, Indiana 47405-4001, and Department of Chemistry-0358, UniVersity of California at
San Diego, La Jolla, California 92093-0358
ReceiVed September 15, 1997
Abstract: Reactions of VCl3(THF)3, bpy, and NaO2CR (R ) Et, Ph; bpy ) 2,2′-bipyridine) in a 1:1:3 ratio
in Me2CO give [V4O2(O2CR)7(bpy)2](ClO4) (R ) Et, 1; R ) Ph, 4) following addition of NBun ClO4. Use
4
of 4,4′-dimethyl- or 5,5′-dimethylbipyridine (4,4′-Me2bpy and 5,5′-Me2bpy, respectively) and R ) Et leads
similarly to [V4O2(O2CEt)7(L-L)2](ClO4) (L-L ) 4,4′-Me2bpy, 2; L-L ) 5,5′-Me2bpy, 3). Yields are in
the 38-90% range. The cation of 1 is isostructural with previously prepared [M4O2(O2CR)7(bpy)2]+ (M )
1
CrIII, MnIII, FeIII) species and possesses a [V4O2] butterfly core. 1D and 2D COSY H NMR spectra of 1
show the solid-state structure is retained on dissolution. The effective magnetic moment (µeff) per V4 for 1
gradually rises from 5.79 µB at 300 K to a maximum of 6.80 µB at 25.0 K and then decreases rapidly to 4.72
µB at 2.00 K. The data in the 7.00-300 K range were fit to the appropriate theoretical expression (based on
Hˆ ) -2JSi‚Sj) to give Jbb ) -31.2 cm-1, Jwb ) +27.5 cm-1, and g ) 1.82, (b ) body, w ) wingtip). These
values indicate a ST ) 3 ground state, confirmed by magnetization vs field studies. Similar results were
obtained for the 2-picolinate (pic) analogue of 1 (complex 5). The ST ) 3, 1, 3, and 0 ground states for the
M ) VIII, CrIII, MnIII, and FeIII, respectively, are rationalized using spin frustration arguments based on
competition between Jbb and Jwb interactions. AC magnetic susceptibility studies down to 1.7 K on 1 and 5
show weak out-of-phase signals (ø′′M) below 4.0 K and corresponding small decreases in the in-phase signals
(ø′M), indicating that the relaxation of magnetization is unusually slow and comparable with the oscillating
AC field (250-1000 Hz). This is a characteristic signature of a single-molecule magnet. Simultaneous
application of AC and DC fields has the effect of increasing the barrier to magnetization relaxation, causing
the ø′′M signal to move to higher temperature and consequently leading to a much stronger ø′′M signal and, for
5, the observation of a peak at ∼2.0 K. A dependence of the ø′′M peak position of 5 on the DC field intensity
and AC field oscillation frequency is found.
Introduction
been discovered that certain molecules can function as
nanomagnets,4-27 and such a molecule has been termed a
“single-molecule magnet” (SMM).25 They are prepared by
Magnetic particles of size < 100 nm are expected to exhibit
unusual properties and are therefore the subject of much current
interest.2 Such nanomagnets have been most frequently pre-
pared by fragmenting bulk ferro- or ferrimagnetic materials, but
this approach unfortunately suffers from the production of
particles with a distribution of sizes, complicating the study of
size vs property relationships and precluding a uniform response
to an external influence such as an applied magnetic field.
Attempts to overcome such problems have included the use of
the protein ferritin as a vessel for the synthesis and study of
nanoscale magnetic particles.3
(5) Villain, J.; Hartman-Boutron, F.; Sessoli, R.; Rettori, A. Europhys.
Lett. 1994, 27, 159.
(6) Eppley, H. J.; Tsai, H.-L.; De Vries, N.; Folting, K.; Christou, G.;
Hendrickson, D. N. J. Am. Chem. Soc. 1995, 117, 301.
(7) Novak, M. A.; Sessoli, R.; Caneschi, A.; Gatteschi, D. J. Magn. Magn.
Mater. 1995, 146, 211.
(8) Barbara, B.; Wernsdorfer, W.; Sampaio, L. C.; Park, J. G.; Paulsen,
C.; Novak, M. A.; Ferre´, R.; Mailly, D.; Sessoli, R.; Caneschi, A.;
Hasselbach, K.; Benoit, A.; Thomas, L. J. Magn. Magn. Mater. 1995, 140-
144, 1825.
(9) Tsai, H.-L.; Eppley, H. J.; De Vries, N.; Folting, K.; Christou, G.;
Hendrickson, D. N. Mol. Cryst. Liq. Cryst. 1995, 274, 167.
(10) Novak, M. A.; Sessoli, R. In Quantum Tunneling of Magnetization
- QTM ’94; Kluwer Academic Publishers: Dordrecht, 1995; pp 171-
188.
Exciting developments in the past few years have demon-
strated an alternative approach to nanoscale magnets: it has
(1) Indiana University. (b) University of California.
(2) Awschalom, D. D.; Di Vincenzo, D. P. Physics Today 1995, 48, 43.
(b) Leslie-Pelecky, D. L.; Rieke, R. D. Chem. Mater. 1996, 8, 1770. (c)
Gunther, L. Physics World 1990, December, 28. (d) Awschalom, D. D.; Di
Vincenzo, D. P.; Smyth, J. F. Science 1992, 258, 414. (e) Stamp, P. C. E.;
Chudnosvsky, E. M.; Barbara, B. Int. J. Mod. Phys. 1992, B6, 1355.
(3) Gider, S.; Awschalom, D. D.; Douglas, T.; Mann, S.; Chaparala, M.
Science 1995, 268, 77.
(4) Sessoli, R.; Tsai, H.-L.; Schake, A. R.; Wang, S.; Vincent, J. B.;
Folting, K.; Gatteschi, D.; Christou, G.; Hendrickson, D. N. J. Am. Chem.
Soc. 1993, 115, 1804. (b) Sessoli, R.; Gatteschi, D.; Caneschi, A.; Novak,
M. A. Nature 1993, 365, 141.
(11) Paulsen, C.; Park, J.-G. In Quantum Tunneling of Magnetization -
QTM ’94; Gunther, L., Barbara, B., Eds.; Kluwer Academic Publishers:
Dordrecht, 1995; pp 189-207.
(12) Paulsen, C.; Park, J.-G.; Barbara, B.; Sessoli, R.; Caneschi, A. J.
Magn. Magn. Mater. 1995, 140-144, 379 and 181.
(13) Friedman, J. R.; Sarachik, M. P.; Tejada, J.; Maciejewski, J.; Ziolo,
R. J. Appl. Phys. 1996, 79, 6031. (b) Friedman, J. R.; Sarachik, M. P.;
Tejada, J.; Ziolo, R. Phys. ReV. Lett. 1996, 76, 3830.
(14) Thomas, L.; Lionti, F.; Ballou, R.; Gatteschi, D.; Sessoli, R.; Barbara,
B. Nature 1996, 383, 145.
(15) Tejada, J.; Ziolo, R. F.; Zhang, X. X. Chem. Mater. 1996, 8, 1784.
S0002-7863(97)03243-5 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/27/1998

2366 J. Am. Chem. Soc., Vol. 120, No. 10, 1998
Castro et al.
solution methods from simple reagents and, being readily
purified, are composed of discrete molecular units of a single,
sharply defined size. Unlike normal magnets, whose properties
are the result of interactions between large numbers of individual
spin carriers within domains in crystals, the behavior of a SMM
arises from the intrinsic properties of individual molecules, i.e.,
no intermolecular interactions are required. Single-molecule
magnets are magnetizable magnets: in an external magnetic
field, their magnetic moments can all be oriented either “up”
or “down”, and when the external field is removed, the moments
(spins) of the molecules will reorient only very slowly if the
temperature is low enough, i.e., the magnetization is retained
in zero field. A SMM is thus magnetizable because there is a
big enough potential energy barrier between its spin “up” and
“down” states. There are two requirements for this: (i) the
SMM must have a relatively large spin (S) ground state, and
(ii) there has to be a large and negative magnetic anisotropy,
macroscopic scale underlies classical behavior at the macro-
scopic scale, given that for nanomagnets it is possible to observe
macroscopic quantum tunneling of magnetization, as demon-
strated for both the [Mn12O12(O2CR)16(H2O)4]0,- 13-15,30 and
[Mn4O3Cl(O2CMe)3(dbm)3]26 species. In the present work, we
report that a new class of VIII4 complexes has been synthesized
and fully characterized and that representative examples have
been identified as the first examples of a SMM outside Mn and
Fe chemistry. The complexes [V4O2(O2CR)7(bpy)2](ClO4) (bpy
) 2,2′-bipyridine) and (NEt4)[V4O2(O2CEt)7(pic)2] (pic )
2-picolinate) have been found to exhibit the out-of-phase AC
magnetic susceptibility signals diagnostic of a SMM. Some
portions of this work have been previously communicated.31
Experimental Section
All manipulations were performed under anaerobic conditions using
an inert-atmosphere glovebox and a vacuum line in conjunction with
standard Schlenk glassware. All solvents were distilled from appropri-
ate drying agents unless otherwise indicated. Solvents were degassed
by vacuum/nitrogen cycles. VCl3(THF)3 was prepared as described in
the literature.32 “[Cr3(OH)2(O2CMe)7]”, 2,2′-bipyridine, and 4,4′-
dimethyl-2,2′-bipyridine (4,4′-Me2bpy) were used as received (Aldrich).
5,5′-Dimethyl-2,2′-bipyridine (5,5′-Me2bpy) was prepared by a literature
2
arising from zero-field splitting (H ) DSz ) in the ground state
with D < 0. Both these requirements are necessary because
the potential energy barrier is S2|D| for integer S and (S2-1/4)|D|
for half-integer spin.
The most thoroughly studied example of a SMM is
[Mn12O12(O2CMe)16(H2O)4]‚2MeCO2H‚4H2O with S ) 10.4-23,28
The related complexes [Mn12O12(O2CR)16(H2O)x] (R ) Et, x
) 3;6 R ) Ph, x ) 46) with S ) 9 or 10 also show SMM
properties, as does (PPh4)[Mn12O12(O2CEt)16(H2O)4], the first
ionic SMM.6,29,30 More recently, a second class of SMM has
been discovered, the family of Mn molecules of formulation
[Mn4O3X(O2CMe)3(dbm)3] (X ) various; dbm- ) the anion
of dibenzoylmethane) with a [Mn4(µ3-O)3(µ3-X)]6+ highly
distorted cubane core and an S ) 9/2 ground state.24-26
Additionally, [Fe8O2(OH)12(tacn)6]8+ (tacn ) 1,4,7-triazacy-
clononane) has been reported to exhibit SMM behavior.27
The SMM field is now established and among the many
challenges for the future is the synthesis of new members of
this class to help expand our knowledge of this phenomenon.
This is particularly important given (i) the potential of nano-
magnets comprising single molecules to provide the ultimate
high-density memory device and (ii) the ability to use such
species to study how quantum mechanical behavior at the
procedure.33 NBun ClO4 was prepared by neutralization of a 40% w/w
4
solution of NBun OH in H2O (Aldrich) with HClO4 (conc). The
4
resulting white precipitate was recrystallized from CH2Cl2/Et2O and
dried in vacuo at room temperature. Sodium propionate and benzoate
were prepared by adding elemental sodium to a slight excess of the
carboxylic acid in THF; the resulting white solids were collected by
filtration, washed copiously with Et2O, and dried in vacuo.
[V4O2(O2CEt)7(bpy)2](ClO4) (1). A reaction slurry consisting of
VCl3(THF)3 (1.12 g, 3.00 mmol), bpy (0.468 g, 3.00 mmol), and NaO2-
CEt (0.829 g, 9.00 mmol) in acetone (50 mL; degassed but not distilled)
was stirred overnight at room temperature to give a red-brown solution
and an off-white precipitate of NaCl. The latter was removed by
filtration, and NBun ClO4 (0.26 g, 0.75 mmol) was added to the filtrate.
4
The solvent was then removed in vacuo, and the resulting red-brown
oil was washed with Et2O (3 × 50 mL) and redissolved in acetone (10
mL; distilled from 4A molecular sieves). Addition of a large excess
of Et2O precipitated a brown microcrystalline precipitate which was
collected by filtration, was washed with Et2O, and dried in vacuo; the
yield was 70-90%. Anal. Calcd (found) for 1‚Et2O‚H2O (C45H63-
N4O22ClV4): C, 43.20 (43.29); H, 5.08 (5.12); N, 4.48 (4.29).
Electronic spectrum in CH2Cl2: λmax, nm (ꢀM, L mol-1 cm-1); 214
(38 200), 246 (29 600), 300 (29 100), 488 (2450), 678 (920). Selected
IR data (cm-1, Nujol): 1583 (s), 1495 (m), 1300 (s), 1219 (m), 1174
(w), 1161 (w), 1080 (s), 1028 (s), 962 (w), 812 (m), 779 (s), 736 (s),
682 (s), 652 (s), 638 (s), 623 (m), 576 (m), 422 (s).
[V4O2(O2CEt)7(4,4′-Me2bpy)2](ClO4) (2). This complex was made
by substituting 4,4′-dimethyl-2,2′-bipyridine (4,4′-Me2bpy) for bpy in
the procedure for complex 1. The product was obtained in 46% yield.
The same procedure but with 5,5′-Me2bpy gave [V4O2(O2CEt)7(5,5′-
Me2bpy)2](ClO4) (3) in 38% yield. Complexes 2 and 3 were pure by
1H NMR spectroscopy.
[V4O2(O2CPh)7(bpy)2](ClO4) (4). This complex was made by
substituting NaO2CPh for NaO2CEt in the procedure for complex 1.
The product was obtained in 47% yield. Anal. Calcd (found) for 4
(C69H56N4O20ClV4): C, 55.4 (55.32); H, 3.4 (3.50); N, 3.7 (3.73).
[NEt4][V4O2(O2CEt)7(pic)2] (5). To a stirred mixture of VCl3(THF)3
(0.747 g, 2.00 mmol) and NaO2CEt (0.552 g, 6.00 mmol) in MeCN
(20 mL) was added Na(pic) (0.145 g, 1.00 mmol) dissolved in H2O
(250 µL). The mixture was stirred overnight to give a dark green
solution and an off-white precipitate (NaCl). The solution was filtered,
(16) Hernandez, J. M.; Zhang, X. X.; Luis, F.; Bartolome´, J.; Tejada, J.;
Ziolo, R. Europhys. Lett. 1996, 35, 301.
(17) Chudnovsky, E. M. Science 1996, 274, 938.
(18) Lionti, F.; Thomas, L.; Ballou, R.; Barbara, B.; Sulpice, A.; Sessoli,
R.; Gatteschi, D. J. Appl. Phys. 1997, 81, 4608.
(19) Friedman, J. R.; Sarachik, M. P.; Hernandez, J. M.; Zhang, X. X.;
Tejada, J.; Molins, E.; Ziolo, R. J. Appl. Phys. 1997, 81, 3978.
(20) Aubin, S. M. J.; Spagna, S.; Eppley, H. J.; Sager, R. E.; Folting,
K.; Christou, G.; Hendrickson, D. N. Mol. Cryst. Liq. Cryst., in press.
(21) Eppley, H. J.; Aubin, S. M. J.; Wemple, M. W.; Adams, D. M.;
Tsai, H.-L.; Grillo, V. A.; Castro, S. L.; Sun, Z.; Folting, K.; Huffman, J.
C.; Hendrickson, D. N.; Christou, G. Mol. Cryst. Liq. Cryst., in press.
(22) Luis, F.; Bartolome´, J.; Fernandez, J. F. Phys. ReV. B 1997, 55,
11448.
(23) Hernandez, J. M.; Zhang, X. X.; Luis, F.; Tejada, J.; Friedman, J.
R.; Sarachik, M. P.; Ziolo, R. Phys. ReV. B 1997, 55, 5858.
(24) Wemple, M. W.; Adams, D. M.; Hagen, K. S.; Folting, K.;
Hendrickson, D. N.; Christou, G. J. Chem. Soc., Chem. Commun. 1995,
1591.
(25) Aubin, S. M. J.; Wemple, M. W.; Adams, D. M.; Tsai, H.-L.;
Christou, G.; Hendrickson, D. N. J. Am. Chem. Soc. 1996, 118, 7746.
(26) Aubin, S. M. J.; Dilley, N. R.; Wemple, M. W.; Maple, M. B.;
Christou, G.; Hendrickson, D. N., submitted for publication.
(27) Barra, A. L.; Debrunner, P.; Gatteschi, D.; Schultz, C. E.; Sessoli,
R. Europhys. Lett. 1996, 35, 133.
(28) Lis, T. Acta Crystallogr. 1980, B36, 2042.
(31) Castro, S. L.; Sun, Z.; Bollinger, J. D.; Hendrickson, D. N.; Christou,
G. J. Chem. Soc., Chem. Commun. 1995, 2517.
(32) Manzer, L. E. Inorg. Synth. 1982, 21, 138.
(33) Ebmeyer, F.; Vo¨gtle, F. Chem. Ber. 1989, 122, 1725. (b) Badger,
G. M.; Sasse, W. H. F. J. Chem. Soc. 1956, 616. (c) Sasse, W. H. F.; Whittle,
C. P. J. Chem. Soc. 1961, 1347.
(29) Schake, A. R.; Tsai, H.-L.; De Vries, N.; Webb, R. J.; Folting, K.;
Hendrickson, D. N.; Christou, G. J. Chem. Soc., Chem. Commun. 1992,
181.
(30) Aubin, S. M. J.; Spagna, S.; Eppley, H. J.; Sager, R. E.; Christou,
G.; Hendrickson, D. N. Chem. Commun. In press.

Single-Molecule Magnets: Tetranuclear Vanadium(III) Complexes
J. Am. Chem. Soc., Vol. 120, No. 10, 1998 2367
Table 1. Crystallographic Data for Complex 1‚2CH2Cl2
Packard Model 8452A spectrophotometers, respectively. DC magnetic
susceptibility data were collected on powdered, microcrystalline samples
(restrained in Vaseline to prevent torquing) on a Quantum Design
MPMS5 SQUID magnetometer equipped with a 5.5 T (55 kG) magnet.
A diamagnetic correction to the observed susceptibilities was applied
using Pascal’s constants. AC magnetic susceptibility measurements
were made on a Quantum Design MPMS2 SQUID magnetometer
equipped with a 1 T (10 kG) magnet and capable of achieving
temperatures of 1.7-400 K. The AC field range is 1 × 10-4-5 G
oscillating at a frequency in the range 5 × 10-4-1512 Hz. AC
susceptibility data were collected on powdered, microcrystalline samples
in an AC field of 1.0 G oscillating in the 250-1000 Hz and in an
applied DC field of 0-8.0 kG.
formula
a )
C43H55Cl5N4O20V4
13.302(3) Å
13.265(3) Å
15.905(4) Å
87.51(1)°
fwa
1328.95 g‚mol-1
P1h
-161 °C
0.710 69 Å
1.588 g cm-3
9.459 cm-1
0.0716
space group
T )
b )
c )
λ )
R )
â )
Dcalc )
96.70(1)°
µ )
γ )
86.66(1)°
R(Fo)c )
Rw(Fo)d
V )
Z )
2778.8 Å3
2
0.1649
a Including solvate molecules. b Graphite monochromator. c R )
2
∑||Fo| - |Fc||/∑|Fo|. d Rw ) [∑w(|Fo| - |Fc|)2/∑w|Fo| ]1/2 where w )
1/σ2(|Fo|).
Results
and NEt4Cl (0.083 g, 0.50 mmol) was added to the filtrate. The MeCN
was removed in vacuo, and the remaining green oil redissolved in
anhydrous acetone (10 mL). THF (10 mL) was added, and the green
solution was left undisturbed at room temperature. After a few days,
a green-black microcrystalline solid had precipitated, and this was
collected by filtration, washed with THF (10 mL), and dried in vacuo.
The yield was 53%. Anal. Calcd (found) for C41H63N3O20V4: C, 43.90
Syntheses. The possibility for the preparation of [V4O2(O2-
CR)7(bpy)2]+ salts was originally suggested by the fact that
[M4O2(O2CR)7(bpy)2](ClO4) complexes were already known for
M ) CrIII (6),34,36 MnIII (7),37 and FeIII (8),38 and there seemed
no reason to believe that the M ) VIII analogues should not
also be accessible if the right procedure could be developed.
After some preliminary experimentation, a successful procedure
was indeed devised. Treatment of VCl3(THF)3, bpy, and NaO2-
CEt in a 1:1:3 molar ratio in Me2CO led to isolation of
[V4O2(O2CEt)7(bpy)2](ClO4) (1) after addition of ClO4-. The
oxide ion most likely arises from solvent H2O impurities (thus,
to optimize yield, the solvent was degassed but not dried), and
the preparation may be summarized as in eq 1. The 2-fold
(43.20); H, 5.66 (5.49); N, 3.75 (3.64)%. Selected IR data (cm-1
,
Nujol): 1664 (s), 1593 (s, br), 1342 (s), 1294 (s), 1242 (m), 1184 (2),
1167 (m), 1082 (m), 1049 (m), 1026 (w), 1010 (m), 885 (w), 854 (m),
810 (m), 773 (m), 707 (m), 698 (m), 540 (m, br), 416 (s) cm-1
.
[Cr4O2(O2CMe)7(bpy)2](ClO4) (6). This complex was prepared by
a modification to the procedure of Vincent and co-workers.34 A green
slurry of “Cr3(OH)2(O2CMe)7” (1.81 g, 3.00 mmol), bpy (1.41 g, 9.00
mmol), and NBun ClO4 (1.03 g, 3.00 mmol) in PhCN (60 mL) was
4
maintained at reflux for 3.5 h. The purple solution was cooled, and
Et2O (∼100 mL) was added to precipitate a solid. The solid was
collected by filtration, washed with Et2O, and recrystallized from a
CH2Cl2/hexanes layering. The product was rather oily, and it was
separated by decantation, washed with hexanes, and dried in vacuo to
give a free-flowing microcrystalline purple solid. The yield was 33%.
The solid appears hygroscopic and was stored in a glovebox. Anal.
Calcd (found) for 6‚1/2Et2O (C36H42N4O20.5ClCr4): C, 39.23 (39.67);
H, 3.84 (3.65); N, 5.08 (4.94).
4VCl3 + 2bpy + 11NaO2CEt + 2H2O f
[V4O2(O2CEt)7(bpy)2]+ + 4EtCO2H + Cl- + 11NaCl (1)
excess of bpy over that required by eq 1 was found to be
beneficial in providing high yields of product, although the
reason for this is not clear. A similar bpy effect has been noted
previously.39 The reaction can readily be altered with respect
to the identities of the carboxylate and the chelate, as confirmed
by the preparation of the 4,4′-Me2bpy (2) and 5,5′-Me2bpy (3)
derivatives, and the PhCO2- derivative (4). No doubt the same
procedure could also be extended to a number of other
carboxylate and substituted-bpy (and phen) groups. In addition,
anionic chelates may be incorporated. Thus, use of Na(pic)
(pic- ) picolinate) leads to the formation of (NEt4)[V4O2(O2-
X-ray Crystallography. Crystals suitable for crystallography were
grown from a CH2Cl2/hexanes layering, and these were subsequently
identified crystallographically as 1‚2CH2Cl2. A suitable single crystal
was chosen from the bulk sample and affixed to the tip of a glass fiber
with silicone grease. The mounted sample was transferred to the
goniostat and cooled to -161 °C for characterization and data collection.
The equipment employed was a Picker four-circle diffractometer; details
of the diffractometry, low-temperature facilities, and computational
procedures employed by the Molecular Structure Center are available
elsewhere.35 A systematic search of a limited hemisphere of reciprocal
space located a set of intensities with no Laue symmetry or systematic
absences, indicting a triclinic space group. Subsequent solution and
refinement of the structure confirmed the choice of the centrosymmetric
space group P1h. After data collection (6°e 2θ e 45°; +h,(k,(l), the
intensities were corrected for Lorentz and polarization effects, and
equivalent data were averaged (R for averaging ) 0.040). The structure
was solved by direct methods (SHELXTL) and Fourier techniques. All
non-hydrogen atoms, including those of two well-behaved CH2Cl2
solvate molecules, were refined with anisotropic thermal parameters.
Hydrogen atoms were located in a difference Fourier map phased on
the non-hydrogen atoms and included in the final least-squares
refinement cycles with isotropic thermal parameters. A final difference
Fourier map was essentially featureless, with the largest peak having
an intensity of 1.46 e/Å.3 Crystallographic data are summarized in
Table 1.
CEt)7(pic)2] (5), the VIII analogue of (NEt4)[Mn4O2(O2CR)7-
4
(pic)2] reported several years ago.40
Description of Structure. A labeled ORTEP plot of the
anion of 1 is shown in Figure 1, and interatomic distances and
angles are collected in Table 2. Complex 1‚2CH2Cl2 crystallizes
in the triclinic space group P1h with the asymmetric unit
containing the entire cation, one anion, and two lattice CH2Cl2
molecules. The cation contains a [V4(µ3-O)2]8+ core comprising
four VIII ions with a “butterfly” disposition and a µ3-O2- ion
bridging each V3 “wing”; both the oxide ions, O(5) and O(6),
lie slightly out of their respective V3 planes, by 0.328 and 0.317
Å, respectively. Peripheral ligation is provided by seven
EtCO2- and two bpy chelate groups, the former in their common
syn,syn-bridging modes and the latter attached to the wingtip
(36) Bino, A.; Chayat, R.; Pedersen, E.; Schneider, A. Inorg. Chem. 1991,
30, 856.
(37) Vincent, J. B.; Christmas, C.; Chang, H.-R.; Li, Q.; Boyd, P. D.
W.; Huffman, J. C.; Hendrickson, D. N.; Christou, G. J. Am. Chem. Soc.
1989, 111, 2086.
(38) McCusker, J. K.; Vincent, J. B.; Schmitt, E. A.; Mino, M. L.; Shin,
K.; Coggin, D. K.; Hagen, P. M.; Huffman, J. C.; Christou, G.; Hendrickson,
D. N. J. Am. Chem. Soc. 1991, 113, 3012.
Physical Measurements. 1H NMR spectra were collected on a 300
MHz Varian Gemini 2000 spectrometer with the protio-solvent signal
used as reference. IR and electronic spectra were recorded on Nujol
mulls and solutions, respectively, on Nicolet Model 510P and Hewlett
(34) Ellis, T.; Glass, M.; Harton, A.; Folting, K.; Huffman, J. C.; Vincent,
J. B. Inorg. Chem. 1994, 33, 5522.
(35) Chisholm, M. H.; Folting, K.; Huffman, J. C.; Kirkpatrick, C. C.
Inorg. Chem. 1984, 23, 1021.
(39) Castro, S. L.; Martin, J. D.; Christou, G. Inorg. Chem. 1993, 32,
2978.

2368 J. Am. Chem. Soc., Vol. 120, No. 10, 1998
Castro et al.
Table 3. Comparison of Selected Structural Parameters (Å, deg)a,b
for the [M4O2(O2CR)7(bpy)2]+ Cations of Complexes 1 and 6-8
parameterc
V(1)
Cr(6)
Mn(7)
Fe(8)
Mb‚‚Mb
2.909(2)
3.318(2)
3.473(2)
1.930(4)
1.977(5)
1.842(4)
96.3(2)
123.6(2)
131.2(2)
133.1
2.784(1)
3.3163(9)
3.4274(8)
1.891(3)
1.914(2)
1.852(3)
94.0(1)
124.8(1)
131.0(1)
143.0
2.848(5)
3.306(5)
3.378(5)
1.903(15)
1.920(15)
1.824(16)
96.3(7)
123.7(8)
130.2(7)
135.4
2.855(4)
3.306(4)
3.439(4)
1.926(5)
1.947(5)
1.819(5)
95.0(2)
123.9(3)
131.9(3)
139.6
b,d
b,e
Mb‚‚Mw
Mb‚‚Mw
Mb-Otb,d
Mb-Otb,e
Mw-Ot
b
Mb-Ot-Mb
b,d
b,e
Mb-Ot-Mw
Mb-Ot-Mw
θ (dihedral) f
g
∆
0.323
0.338
0.334
0.323
a Averaged using C2 virtual symmetry. b Quoted esd’s for averaged
values are the esd’s of the individual values. c Subscripts: b ) body,
w ) wingtip, t ) triply-bridging. d For M atoms bridged by two RCO2
-
groups. e For M atoms bridged by one RCO2- group. f Dihedral angle
between two M3 least-squares planes. g Distance of Ot to M3 least-
squares plane.
Figure 1. ORTEP representation of the [V4O2(O2CEt)7(bpy)2]+ cation
of 1 at the 50% probability level.
Table 2. Selected Interatomic Distances (Å) and Angles (deg) for
Complex 1‚2CH2Cl2
3.465(2) Å). The dihedral angle θ between the two V3 least-
squares planes is 133.1°.
V(1)-V(2) 2.909(2) V(2)-O(5) 1.981(5) V(3)-O(13) 2.053(5)
V(1)-V(3) 3.315(2) V(2)-O(32) 1.995(5) V(3)-N(42) 2.126(5)
V(1)-V(4) 3.480(2) V(2)-O(27) 2.019(5) V(3)-N(53) 2.173(6)
V(1)-O(5) 1.928(4) V(2)-O(8) 2.024(4) V(4)-O(6) 1.841(4)
V(1)-O(6) 1.972(5) V(2)-O(37) 2.055(4) V(4)-O(38) 1.993(5)
V(1)-O(12) 1.988(5) V(2)-V(4) 3.321(2) V(4)-O(33) 2.042(5)
V(1)-O(22) 2.004(4) V(2)-V(3) 3.465(2) V(4)-O(23) 2.050(5)
V(1)-O(7) 2.053(4) V(3)-O(5) 1.830(4) V(4)-N(65) 2.116(6)
V(1)-O(17) 2.085(4) V(3)-O(18) 1.981(4) V(4)-N(54) 2.159(6)
V(2)-O(6) 1.931(4) V(3)-O(28) 2.049(5)
As stated earlier, the cation of 1 is the fourth member of the
isostructural [M4O2(O2CR)7(bpy)2]+ (M ) FeIII, MnIII, CrIII)
family of complexes, and it is thus pertinent to ask whether
there are any significant structural trends or differences across
this series. In Table 3 are compared selected structural
parameters for the four complexes, i.e., 1 and 6-8. It is readily
apparent that the complexes are essentially isometric, the main
variations being between some of the values for the CrIII
complex compared with the others; for example, the Crb‚‚Crb
distance appears short, but this is undoubtedly due to the
variation in M3+ size as gauged by ionic radii41 values of 0.640
(V3+), 0.615 (Cr3+), 0.645 (Mn3+), and 0.645 (Fe3+) Å,
reflecting population of σ-antibonding d-orbitals for high-spin
Mn3+ and Fe3+. The decreased Mb‚‚‚Mb separation for Cr3+
then also affects the Mb‚‚Ot and Mb-Ot-Mb values. Note also
that the Jahn-Teller distortion for high-spin Mn3+ complicates
close comparisons involving complex 7. In general, the four
complexes are very similar and may be quite satisfactorily
described as isostructural.
O(5)-V(1)-O(6)
O(5)-V(1)-O(12)
O(6)-V(1)-O(12)
O(5)-V(1)-O(22)
O(6)-V(1)-O(22)
O(12)-V(1)-O(22)
O(5)-V(1)-O(7)
O(6)-V(1)-O(7)
O(12)-V(1)-O(7)
O(22)-V(1)-O(7)
O(5)-V(1)-O(17)
O(6)-V(1)-O(17)
O(12)-V(1)-O(17)
O(22)-V(1)-O(17)
O(7)-V(1)-O(17)
O(6)-V(2)-O(5)
O(6)-V(2)-O(32)
O(5)-V(2)-O(32)
O(6)-V(2)-O(27)
O(5)-V(2)-O(27)
O(32)-V(2)-O(27)
O(6)-V(2)-O(8)
O(5)-V(2)-O(8)
O(32)-V(2)-O(8)
O(27)-V(2)-O(8)
O(6)-V(2)-O(37)
O(5)-V(2)-O(37)
O(32)-V(2)-O(37)
O(27)-V(2)-O(37)
O(8)-V(2)-O(37)
O(5)-V(3)-O(18)
O(5)-V(3)-O(28)
O(18)-V(3)-O(28)
82.6(2)
95.9(2)
176.0(2)
167.1(2)
87.6(2)
94.4(2)
91.1(2)
87.1(2)
89.2(2)
96.8(2)
89.1(2)
96.0(2)
87.7(2)
83.6(2)
176.8(2)
82.3(2)
96.3(2)
174.8(2)
167.4(2)
87.8(2)
94.2(2)
91.3(2)
86.6(2)
88.5(2)
95.7(2)
89.6(2)
97.0(2)
87.9(2)
84.0(2)
176.4(2)
96.8(2)
93.5(2)
93.8(2)
O(5)-V(3)-O(13)
O(18)-V(3)-O(13)
O(28)-V(3)-O(13)
O(5)-V(3)-N(42)
O(18)-V(3)-N(42)
O(28)-V(3)-N(42)
O(13)-V(3)-N(42)
O(5)-V(3)-N(53)
O(18)-V(3)-N(53)
O(28)-V(3)-N(53)
O(13)-V(3)-N(53)
N(42)-V(3)-N(53)
O(6)-V(4)-O(38)
O(6)-V(4)-O(33)
O(38)-V(4)-O(33)
O(6)-V(4)-O(23)
O(38)-V(4)-O(23)
O(33)-V(4)-O(23)
O(6)-V(4)-N(65)
O(38)-V(4)-N(65)
O(33)-V(4)-N(65)
O(23)-V(4)-N(65)
O(6)-V(4)-N(54)
O(38)-V(4)-N(54)
O(33)-V(4)-N(54)
O(23)-V(4)-N(54)
N(65)-V(4)-N(54)
V(3)-O(5)-V(1)
V(3)-O(5)-V(2)
V(1)-O(5)-V(2)
V(4)-O(6)-V(2)
V(4)-O(6)-V(1)
V(2)-O(6)-V(1)
95.1(2)
89.5(2)
170.5(2)
99.2(2)
163.9(2)
83.4(2)
91.0(2)
174.4(2)
88.7(2)
86.7(2)
84.3(2)
75.3(2)
95.4(2)
95.0(2)
89.2(2)
93.0(2)
93.2(2)
171.4(2)
98.0(2)
166.4(2)
91.6(2)
84.1(2)
173.7(2)
90.8(2)
83.9(2)
87.8(2)
75.8(2)
123.8(2)
130.8(2)
96.2(2)
123.4(2)
131.6(2)
96.4(2)
2
NMR Spectroscopy. 1H and H NMR spectroscopy has
proven useful in the past for studying complexes 6-8 in
solution,37-39 including assessing whether the solid-state struc-
tures are retained on dissolution, and complementary studies
have therefore been performed with the present complexes. This
knowledge will also be of use in future reactivity studies of
these molecules. The 1H NMR spectrum of complex 1 in CD2-
Cl2 is shown in Figure 2, and pertinent data for complexes 1-3
are collected in Table 4. Peak assignments were made by
consideration of peak integrations, T1 values, the effect of Me
substitution at the bpy 4 and 5 positions, and comparisons with
the spectra of 7 and 8.
Excluding signals from the solvent CD2Cl2 and its impurities,
there are 20 peaks in the δ ) +90 to -50 ppm range that can
be assigned to the [V4O2(O2CEt)7(bpy)2]+ cation; this is exactly
the number expected for this cation with effective C2 solution
symmetry: 8 bpy resonances, 4 CH3 resonances in a 2:2:2:1
relative integration ratio, and 8 CH2 resonances in a 2:2:2:2:2:
2:1:1 relative ratio resulting from the diastereotopic nature of
each CH2 pair of hydrogen atoms. These eight resonances occur
in the δ ) 30-90 ppm range as four pairs with comparable
chemical shifts and T1 times; the less intense pair at δ ) 33.9
VIII centers, V(3) and V(4). The metals are all octahedrally
coordinated, and the entire cation has virtual C2 symmetry but
C1 crystallographic symmetry. The V‚‚‚V separations fall into
three types: the body/body separation, V(1)‚‚‚V(2), is short
(2.909(2) Å) relative to the wingtip/body separations, which
themselves separate into two types, those bridged by two
EtCO2- groups (V(1)‚‚V(3), 3.315(2); V(2)‚‚V(4), 3.321(2) Å),
and those bridged by one (V(1)‚‚V(4), 3.480(2); V(2)‚‚V(3),
(40) Libby, E.; McCusker, J. K.; Schmitt, E. A.; Folting, K.; Hendrickson,
D. N.; Christou, G. Inorg. Chem. 1991, 30, 3486.
(41) Shannon, R. D. Acta Crystallogr. Sect. A 1976, A32, 751.

Single-Molecule Magnets: Tetranuclear Vanadium(III) Complexes
J. Am. Chem. Soc., Vol. 120, No. 10, 1998 2369
Table 4. 1H NMR Spectral Datae for [V4O2(O2CEt)7(L-L)2]+,-
(L-L ) bpy, pic-) Complexes
1
2
3
5 f
CH2
CH3
3/3′
86.7, 70.7a
84.0, 50.7b
75.2, 64.9c
38.2, 33.9d
2.9d
85.6, 69.6
83.1, 49.3
74.0, 63.8
37.3, 33.3
3.2
1.4
1.0 (br)
87.3, 69.1
82.1, 47.9
74.7, 62.3
51.9, 33.6
3.3
1.5
1.1
74.4, 64.6
69.7, 42.3
66.3, 49.4
30.7
2.3
1.3
0.5
0.4
16.3
1.6a
0.6c
0.5b
1.0
10.5, 8.0
11.3, 8.8
10.4, 8.0
4/4′ -32.7, -22.9 (47.2, 34.1)g -21.7, -31.5
-27.5
18.6
-47.5
5/5′
15.1, 14.7
15.3, 14.8
(-5.8, -10.9)g
6/6′ -48.7, -35.2 -34.1, -47.7 -36.1, -49.1
a-d Superscripts identify the CH2/CH3 pairings deduced from the 2D
e
spectra. ∼23 °C; in CD2Cl2 (1-3) or (CD3)2CO (5); values are
chemical shift from TMS using the δ scale (downfield shifts are
positive). f Excluding small peaks assigned to the anti isomer. g Values
in parentheses are for Me groups at these positions.
π system of the bpy rings. This mechanism is also supported
by the change in sign of the isotropic shifts of the 4/4′ or 5/5′
Me groups compared with the shifts of hydrogens at these
positions (Table 4).
The spectrum of 5 in (CD3)2CO comprises 17 peaks (Table
4). Seven peaks of equivalent intensity were found for the CH2
groups between 30 and 75 ppm, similar to the 30-90 ppm
region in Figure 2. The absence of an eighth resonance in the
methylene region can be explained by the equivalent integration
of the seven peaks; coincidental overlap of the two resonances
for the unique bridging carboxylate (which would otherwise be
half the intensity of the wingtip-body propionate methylene
resonances) would account for the eighth methylene proton. Four
peaks were found significantly shifted from the diamagnetic
region; two upfield and two downfield; these resonances were
assigned to the methyl protons of the propionate; there are also
two resonances in a 2:3 ratio in the expected positions for the
tetraethylammonium cation. These data are consistent with the
anion being structurally similar to the cation of 1 with pic-
replacing the bpy groups. However, an additional 18-20
signals, all approximately 5% intensity relative to the major
peaks of 5, were also observed in the spectrum. It was first
assumed that these were due to an impurity, but recrystallized
material gave the same spectrum, with little change in the
relative intensities, and the elemental analysis was consistent
with the formula of 5. We suspect that 5 occurs as a mixture
of isomers. The two picolinate ligands can bind either syn or
anti to each other (with respect to the orientation of the six-
membered ring of picolinate). The syn configuration is observed
in the crystal structures of [M4O2(O2CR)7(pic)2]- (M ) Cr,43
Mn40) and is the most likely configuration for the major
component of 5. The anti isomer is observed in the neutral
complex [Mn4O2(O2CPh)6(pic)2(MeCN)2].44 The large number
of small additional peaks is consistent with an anti isomer (of
lower symmetry than the C2 syn isomer), and it is reasonable
that the cause of the extra peaks should somehow involve the
pic- groups, since there is no evidence of an additional species
in the spectra of 1, 2, or 3.
Figure 2. 1H NMR spectrum of [V4O2(O2CEt)7(bpy)2](ClO4) (1) in
CD2Cl2 at ∼23 °C. Primed and unprimed refer to the two halves of the
bpy group. The peak marked with an asterisk is from solvent protio-
impurities.
and 38.2 ppm is clearly from the unique EtCO2- group bridging
the body V atoms V(1) and V(2). The CH3 resonances are less
paramagneticallly shifted, as expected from their greater distance
from the paramagnetic centers (paramagnetic shift R r-3), and
occur in the δ ) 0-3 ppm range, two only partially resolved;
the integration value of the δ ) 2.9 resonance indicates it to be
-
from the unique EtCO2 group.
The assignment of the bpy resonances was facilitated by Me-
substitution at the 4,4′ and 5,5′ positions (complexes 2 and 3,
respectively), which allowed identification of the resonances
from the corresponding 4,4′ and 5,5′ hydrogen atoms as labeled
in Figure 2; the other resonance positions did not differ much
between 1, 2, and 3. The 6,6′ hydrogen atoms are the closest
to the metal centers and consequently give the most paramag-
netically broadened resonances at δ ) -35.2 and -48.7 ppm.
The remaining resonances at δ ) 11.3 and 8.8 ppm are then
assigned by elimination to the 3,3′ hydrogen atoms. The bpy
resonances display the alternating-shift pattern characteristic of
contact shifts resulting from spin delocalization onto the bpy
ligand via a π-delocalization pathway.42 Thus, the 3/3′ and 5/5′
hydrogen resonances are paramagnetically shifted downfield,
while the 4/4′ and 6/6′ pairs are shifted upfield. A π spin
delocalization mechanism is consistent with unpaired spin only
in dπ orbitals of near octahedral VIII and their overlap with the
T1 and 2D COSY Experiments. With the exception of the
resonances for the unique body-body bridging carboxylate
(42) NMR of Paramagnetic Molecules; La Mar, G. N., Horrocks, W.
DeW., Jr., Holm, R. H., Eds.; Academic Press: New York, 1973.
(43) Donald, S.; Terrell, K.; Vincent, J. B.; Robinson, K. D. Polyhedron
1995, 14, 971.
(44) Libby, E.; Folting, K.; Huffman, C. J.; Huffman, J. C.; Christou,
G. Inorg. Chem. 1993, 32, 2549.

2370 J. Am. Chem. Soc., Vol. 120, No. 10, 1998
Castro et al.
Table 5. T1 Relaxation Times for the Propionate Hydrogen Nuclei
of Complex 1
chemical shift
(ppm)
chemical shift
(ppm)
T1 (ms)
T1 (ms)
CH2
CH3
86.7
84.0
75.2
70.7
10.4(3)
10.4(3)
12.4(3)
10.6(2)
64.9
50.7
38.2
33.9
12.0(2)
11.0(2)
9.9(6)
10.7(3)
2.9
1.6
19.1(4)
15.6(3)
0.6
0.5
21.4(9)
18.5(3)
1
group, it was not possible to determine from the 1D H NMR
spectra which pairs of CH2 resonances arise from protons bound
to the same carbon or which CH3 group is associated with which
CH2 protons. Because the longitudinal relaxation of a hydrogen
nuclei is strongly influenced by its environment, even small
changes in the hydrogen environment can have a substantial
effect on the relaxation time (T1). With this in mind, the T1
times of each of the CH2 and CH3 resonances in 1 were
measured. The measurements were made using a standard
inversion-recovery pulse sequence45 with a 90° pulse width
of 10.4 µs. Calculation of the T1 time was made using the
exponential decay equation (eq 2), where Mz is the magnetization
Mz ) M0[1 - 2 exp(-τ/T1)]
(2)
in the z direction and τ is the delay between the 180° and 90°
pulses. The results of the experiment are summarized in Table
5. It was hoped that the T1 times would be different enough to
allow pairs of diastereotopic protons to be assigned on the basis
of similar T1 times. This did not turn out to be the case, since
the T1 times for the CH2 resonances were in a narrow range of
9.9(6)-12.4(3) ms and the CH3 T1 times in the range 15.6(3)-
21.4(9) ms, precluding confident assignments. However, the
desired assignments were achieved from a 2D COSY experi-
ment. The broadness of the resonances and the large spectral
window involved (160 ppm) caused some difficulties, but a
satisfactory 2D spectrum was nevertheless obtained (Figure 3)
with all the cross peaks visible. All cross peaks resulting from
hydrogen nuclei in the same CH2 group were clearly evident,
allowing the eight CH2 resonances to be assigned as the four
pairings given in Table 4. In addition, cross peaks for the three-
bond CH2/CH3 couplings were observed, as is better seen in
the expanded view of Figure 4, leading to the CH2CH3 pairings
of Table 4. Cross peaks for the protons of the bpy ligands were
evident as well, and this allowed the resonances pertaining to
each half of the bpy group to be grouped together: 11.3, -32.7,
15.1, -48.7 ppm (3, 4, 5, 6) and 8.8, -22.9, 14.7, -35.2 ppm
(3′, 4′, 5′, 6′).
DC Magnetic Susceptibility Studies. Variable-temperature
magnetic susceptibility studies were performed on powdered
samples of complexes 1, 5, and 6 in the temperature range 2.00-
320 K in a 10.0 kG (1T) applied magnetic field. The samples
were restrained in Vaseline to prevent torquing. For complex
1, the effective magnetic moment (µeff) per V4 gradually rises
from 5.79 µB at 300 K to a maximum of 6.80 µB at 25.0 K and
then decreases rapidly to 4.72 µB at 2.00 K (Figure 5). The
300 K value is slightly higher than the spin-only (g ) 2) value
of 5.66 µB for a cation comprising four noninteracting VIII
centers, suggesting the presence of some ferromagnetic exchange
Figure 3. 2D COSY 1H NMR spectrum of [V4O2(O2CEt)7(bpy)2](ClO4)
(1) in CD2Cl2. The vertical line of points at 20 ppm is an artifact that
appears at the center of the spectral window.
interactions in 1; the low-temperature values may be compared
with the 4.90, 6.93, and 8.94 µB spin-only values expected for
an exchange-coupled cation with a S ) 2, 3, or 4 ground-state
spin value, respectively.
The experimental µeff vs T data were fit to the theoretical
expression derived previously37,38,40 for a tetranuclear butterfly
complex with C2V symmetry, adjusted for the present case of
4VIII (S ) 1) centers. This derivation using C2V symmetry
assumes all four body/wingtip interactions of the [V4O2]8+ core
are equivalent, and this simplifying and reasonable approxima-
tion allows application of the equivalent operator approach based
on the Kambe vector coupling method.46 The resulting spin
Hamiltonian is given in eq 3, where Jwb is the wingtip/body
2
2
Hˆ ) -Jwb(SˆT2 - SˆA2 - SˆB ) - Jbb(SˆA2 - Sˆ12 - Sˆ3 ) (3)
interaction, Jbb is the body/body interaction, SˆA ) Sˆ1 + Sˆ2, SˆB
) Sˆ3 + Sˆ4, and SˆT ) SˆA + SˆB, where SˆT is the resultant spin of
the complete V4 cation. The interaction between the two
wingtip atoms V(3) and V(4) is assumed to be negligible, given
their large separation (5.634(2) Å). The corresponding eigen-
value expression is given in eq 4 where constant terms have
been ignored. In the present case of four VIII centers, S1 ) S2
) S3 ) S4 ) 1, and there are a total of 19 spin states (ST) with
values ranging from 0-4.
E(ST) ) -Jwb[ST(ST + 1) - SA(SA + 1) - SB(SB + 1)] -
Jbb[SA(SA + 1)] (4)
(45) Derome, A. Modern NMR Techniques for Chemistry Research;
Pergamon Press: Oxford, 1987.
(46) Kambe, K. J. Phys. Soc. Jpn. 1950, 5, 48.
A theoretical expression for the µeff vs T behavior of 1 was
derived from the use of eq 4 and the Van Vleck equation; this

Single-Molecule Magnets: Tetranuclear Vanadium(III) Complexes
J. Am. Chem. Soc., Vol. 120, No. 10, 1998 2371
Figure 6. Plots of reduced magnetization (M/NµB) vs H/T for complex
1 at 10 (b), 20 (9), 30 (2), 40 ([), and 50 (>) kG. The solid lines are
fits to the appropriate expression; see the text for the fit parameters.
a triply degenerate second excited state, comprising the (2,0,2),
(1,0,1), and (0,0,0) states, at energies above the ground state of
14.8 and 47.6 cm-1, respectively. The low-lying ST ) 4 first
excited-state rationalizes why the µeff vs T plot in Figure 5 rises
to a maximum of 6.80 µB at 25.0 K, before decreasing to 6.36
µB at 7.00 K as the ST ) 4 first excited state is progressively
depopulated and more of the molecules enter the ST ) 3 ground
state. The µeff vs T data for complex 5 look similar to those
shown in Figure 5 for complex 1. Fitting of the data gave Jbb
) -31.2 cm-1, Jwb ) +27.0 cm-1, and g ) 1.85. Complex 5
also has a ST ) 3 ground state.
1
Figure 4. Expanded region of the 2D COSY H NMR spectrum of
complex 1 in CD2Cl2 showing CH2/CH3 cross peaks: CH2/CH3 pairings
are indicated by the letters a-d.
For complex 6, the µeff (ømT) value per Cr4 slowly decreases
from 6.43 µB (5.17 cm3 K mol-1) at 320 K to 2.91 µB (1.06
cm3 K mol-1) at 2.00 K (Figure 5). The data were fit to the
theoretical µeff vs T expression derived for a C2V butterfly model
(eqs 3 and 4) except that S1 ) S2 ) S3 ) S4 ) 3/2. The fitting
parameters were found to be Jbb ) -19.4 cm-1, Jwb ) -11.2
cm-1, and g ) 1.99, with no paramagnetic impurity term being
required, and the TIP held constant at 400 × 10-6 cm3 mol-1
.
These values are very similar to those found for [Cr4O2(O2-
CMe)7(bpy)2](PF6) by Bino, Pedersen, and co-workers using
the same “two-J” model:36 Jbb ) -25.5 cm-1 and Jwb ) -15.7
cm-1, with g fixed at 1.98. The results for 6 indicate the ground
state to be the (1,2,3) state with a low-lying (2,1,3) first excited
state and a (0,3,3) second excited state at 12.2 and 26.7 cm-1
respectively, above the ground state.
,
DC Magnetization vs Magnetic Field Studies. Confirma-
tion of the ground-state spin value for complexes 1, 5, and 6
was sought by DC magnetization vs field studies employing
applied magnetic fields (H) of 500 G to 50.0 kG in the 2.00-
30.0 K temperature range. The data for complex 1 are shown
in Figure 6 as reduced magnetization M/NµB vs H/T, where M
is magnetization, N is Avogadro’s number, and µB is the Bohr
magneton. For a spin system populating only the ground state
and exhibiting no ZFS effects, the various isofield lines should
be superimposed and the M/NµB plots should saturate at gST,
or approximately 6 for g ≈ 2 and ST ) 3. Figure 6 shows that
(i) the isofield lines are not superimposed, and (ii) the data at
the highest field and lowest temperature saturate at M/NµB ≈
4. Clearly, there is appreciable ZFS. The M/NµB vs H/T data
were fit to a model that assumes that only the ground state is
populated at the lowest temperatures and that includes Zeeman
Figure 5. Plots of effective magnetic moment (µeff) per M4 vs
temperature for complexes 1 (]) and 6 (O). The solid lines are fits to
the appropriate expressions; see the text for the fitting parameters.
expression was then used to least-squares fit the experimental
data measured in the 7.00-300 K range, and the fit is shown
as a solid line on the µeff vs T plot in Figure 5. Data for
temperatures below 7 K were omitted because the very rapid
decrease in µeff is likely due primarily to zero field splitting
(ZFS) effects within the ground-state spin manifold, and these
effects were not incorporated into our fitting model. The fitting
parameters were found to be Jbb ) -31.2 cm-1, Jwb ) +27.5
cm-1, and g ) 1.82, with no correction term for paramagnetic
impurities and a temperature-independent paramagnetism (TIP)
term held constant at 400 × 10-6 cm3 mol-1. These values
indicate that the ground state of complex 1 is, in the format
(ST,SA,SB), the (3,1,2) state, with a (4,2,2) first excited state and
2
and axial ZFS (DSz ) interactions. A matrix diagonalization

2372 J. Am. Chem. Soc., Vol. 120, No. 10, 1998
Castro et al.
Table 6. Comparison of Magnetic Parameters for
[M4O2(O2CR)7(bpy)2]+ Complexes
V(1)
Cr(6)
Mn(7)
Fe(8)
J
wb, cm-1
+27.5
-11.2
-7.8
-45.5
-8.9
Jbb, cm-1
ST
-31.2
-19.4
-23.5
3
1
3
0
Mb-O-Mb, dega
96.3(2)
94.0(1)
124.8(1)
131.0(1)
96.3(7)
123.7(8)
130.2(9)
95.0(2)
123.9(3)
131.9(3)
Mw-O-Mb, dega 123.6(2)
131.2(2)
Mb‚‚Mb, Å
2.909(2)
2.784(1)
2.848(5)
2.855(4)
a Averaged values.
body V atoms to be concomitantly aligned antiparallel to each
other and parallel to the spins of the wingtip V atoms. The
preferred spin alignments are consequently frustrated because
neither Jwb nor Jbb is insignificant (relative to the other). The
ground-state thus becomes sensitive to the relative magnitude
of the competing exchange interactions, i.e., the Jwb/Jbb ratio.
The (ST, SA, SB) values of the ground states of 1 and 6 have
been identified above as (3,1,2) and (1,2,3), respectively, and
this allows the spin alignments in the ground state to be
depicted: for comparison, the previously determined alignments
for the M ) MnIII (7)37,40 and M ) FeIII (8)38 cases are also
Figure 7. Plots of reduced magnetization (M/NµB) vs H/T for complex
6 at 10 (b), 20 (2), 30 (9), 40 ([), and 50 (>) kG. The solid lines are
fits to the appropriate expression; see the text for the fit parameters.
procedure described elsewhere was employed.6 The solid lines
in Figure 6 are the least-squares fits of the five isofield data
sets plotted. Two different fits of the data were found, one with
D < 0 and the other with D > 0; two minima are generally
seen in the fitting of magnetization data. The fitting parameters
for the fit shown as solid lines in Figure 6 are ST ) 3, g )
1.93, and D ) -1.52 cm-1. The other fit gave ST ) 3, g )
1.96, and D ) +3.55 cm-1. It is physically reasonable that D
< 0 for a complex comprised of VIII ions. The D set of
parameters is further supported by the AC susceptibility data
(vide infra). The magnetization vs field studies thus confirm
the conclusion from the øM vs T studies that complex 1 has a
ST ) 3 ground state. The data for complex 5 (essentially
identical with those for 1) were similarly fit to give ST ) 3, g
) 2.00, and D ) -1.50 cm-1
.
For complex 6, the M/NµB vs H/T data are shown in Figure
7, and comparison with Figure 6 shows two main differences:
(i) the various isofield lines for 6 are more nearly superimposed
than those for 1, suggesting smaller ZFS in 6 vs 1; and (ii)
saturation of the M/NµB plot is not quite achieved, even at the
highest H/T values, although extrapolation suggests saturation
near ∼2, the gST value expected for a ST system with g ≈ 2.
Fitting of the data as for 1 and 5 gave the solid lines in Figure
7 with ST ) 1, D ) -0.17 cm-1, and g ) 2.04, confirming the
ST ) 1 ground-state deduced from the øM vs T studies.
Origin of the Ground States of 1 and 6. We have in
previous reports described how spin frustration within the
butterfly like [M4O2]8+ cores lead to S ) 3 and S ) 0 ground
states for the M ) MnIII and FeIII cases, respectively. By spin
frustration is here meant the presence within appropriate metal
topologies of competing exchange interactions of comparable
magnitude that lead to prevention (frustration) of the preferred
spin alignments.47,48 As a result, spin frustration can often lead
to ground-state spin values larger than expected from the nature
of the exchange interactions (e.g., even when all are antiferro-
magnetic). For the present cases of 1 and 6, it is clear that
spin frustration effects are operative. For 1, Jwb is positive and
Jbb is negative, but it is clearly impossible for the spins of the
shown, and the same atom numbering scheme as Figure 1 is
employed. The Jwb and Jbb values for 1, 6, 7, and 8 are collected
in Table 6. The body metal atoms are, of course, equivalent
by symmetry and thus, for the VIII4 complex, for example, this
simple depiction is to be interpreted as indicating that the spins
of V(1) and V(2) align neither perfectly parallel (SA ) 2) nor
antiparallel (SA ) 0) but in an intermediate manner to give a
resultant SA ) 1 which then aligns parallel to the SB ) 2
resultant to give the overall ST ) 3 ground state. The tendency
of the body spins to align antiparallel (Jbb is negative) is opposed
by the body-wingtip interaction (Jwb) which favors parallel
alignment of spins. Thus, an intermediate situation is obtained,
one that depends on the Jwb/Jbb ratio. Figure 8 shows the spin
manifold as a function of Jwb/Jbb: For 0.5 < |Jwb/Jbb| < 1.0,
which includes the experimentally determined value for |Jwb
/
Jbb| of 0.88, the ground state has ST ) 3 as depicted above.
Note that even though Jbb (-31.2 cm-1) is stronger than Jwb
(+27.5 cm-1), there is only one of the former and four of the
latter; thus, the body spins are not aligned antiparallel, and Jwb
primarily determines the ground state. For |Jwb/Jbb| > 1.0,
however, the ability of the antiferromagnetic Jbb interaction to
prevent perfectly parallel alignment of the body spins (as
dictated by the ferromagnetic Jwb) is totally overcome, and all
spins are parallel to give the ST ) 4 ground state. Similarly,
for |Jwb/Jbb| < 0.5, the body spins are aligned antiparallel giving
a triply degenerate ground state with ST ) 0, 1, and 2.
(47) In accord with the suggestion by Kahn,48 we shall employ terms
such as “spin frustration degeneracy” for the special cases when competing
exchange interactions lead to a degenerate ground state. Thus, we shall
continue to employ the term “spin frustration” for the more general case
where competing exchange interactions of comparable magnitude prevent
(frustrate) the preferred spin alignments. This is in keeping with our practice
to date, and with the most common meaning of the verb “to frustrate”.
(48) Kahn, O. Chem. Phys. Lett. 1997, 265, 109.
Similar considerations rationalize the ground state of the CrIII
4
complex 6. Jwb and Jbb are both antiferromagnetic, and the

Single-Molecule Magnets: Tetranuclear Vanadium(III) Complexes
J. Am. Chem. Soc., Vol. 120, No. 10, 1998 2373
spin down by passing through the intermediate quantized levels.
2
2
The potential-energy barrier is of height |D|[Sˆz |(3 - Sˆz |0 ]
) 9|D|. The barrier for complex 1 is calculated to be 9(1.52
cm-1) ) 13.68 cm-1, if D ) -1.52 cm-1. Note that the other
fit of magnetization data for complex 1 gave D ) 3.55 cm-1
;
a positive D value changes the diagram in Figure 9 such that
the highest energy states are the Ms ) (3 states. When D is
positive, there is no potential-energy barrier to the reversal of
the magnetization.
If a molecule has an appreciable barrier for reversal of the
direction of its magnetic moment, then it can function as a SMM.
Negative magnetic anisotropy (D < 0) and a relatively large
spin ground state are required. As stated in the Introduction,
in response to an external magnetic field, the magnetic moment
of a SMM can be magnetized with its spin either “up” or “down”
along the magnetic anisotropy axis; if the external field is now
removed, the moment (spin) of the molecule will reorient only
very slowly if the temperature is low. After saturation in a large
field, the magnetization of [Mn12O12(O2CMe)16(H2O)4]‚
2MeCO2H‚4H2O relaxes with a half-life of greater than 2
months at 2K. There are several ways to determine if a
molecule is a SMM. The most classic indication is a hysteresis
loop in the plot of magnetization vs magnetic field. Several
Mn12 complexes have been shown to exhibit magnetization
hysteresis loops,4-27 but none of the [V4O2]8+ butterfly com-
plexes show hysteresis effects in magnetization vs magnetic field
data measured down to 1.70 K, the lowest limit of our SQUID
magnetometer. However, a frequency-dependent out-of-phase
AC magnetic susceptibility (ø′′M) signal is an equally good
indicator that a molecule is a SMM. In the AC susceptibility
experiment, the AC magnetic field is oscillated at a particular
frequency, and a ø′′M signal is observed when the rate at which
the magnetic moment of a molecule flips is close to this value.
Thus, if a collection of SMMs are kept at a certain temperature
and the frequency of the AC magnetic field is varied, a
maximum in the ø′′M signal will occur when the frequency of
the field equals the rate at which a molecule can interconvert
between the two halves of the potential-energy double well
shown in Figure 9. Frequency-dependent ø′′M signals have been
reported for all of the known Mn12 SMMs.4-27
Figure 8. Plot of |Jwb/Jbb| vs relative energy for the idealized C2V
symmetry core of [V4O2(O2CEt)7(bpy)2](ClO4) (1) showing the change
in ground state, indicated as (ST, SA, SB).
Figure 9. Double-well potential energy vs magnetization direction
diagram for a ST ) 3 species with a negative ZFS parameter D. The
thermal barrier for reversal of magnetization direction from ms ) 3 to
ms ) -3 is |9D|.
AC magnetic susceptibility data were collected for complexes
1 and 5. Some of the data for [V4O2(O2CEt)7(bpy)2](ClO4) (1)
are shown in Figure 10. A polycrystalline sample was studied
in the 1.7-30 K range with a 1.0 G AC field oscillating at
either 250, 499, or 997 Hz. The DC magnetic field was held
at zero. In the 10-30 K range, the value of ø′MT, where ø′M is
the in-phase AC susceptibility, remains constant at a value of
4.81 cm3 K mol-1 (i.e., µeff ) 6.20 µB), consistent with a ST )
3 ground state with g ) 1.82. As can be seen in Figure 5, this
AC value for ø′MT is smaller than the µeff ≈ 6.8 µB value in the
20-30 K range in a 10.0 kG DC field. The difference is likely
due to the first excited state with ST ) 4 being partially
populated in this temperature range in the presence of a 10.0
kG field, since this ST ) 4 state is only 14.8 cm-1 above the
ground state.
system is frustrated. Jbb (-25.5 cm-1) is stronger than Jwb
(-15.7 cm-1) but again, the greater number of the latter type
causes the ground-state alignments to be dominated by Jwb
,
enforcing almost parallel alignment of the body spins to give
the ST ) 1 ground state. An increase in the Jwb/Jbb ratio would
give a ST ) 0 ground state, as seen for the FeIII4 complex, while
a decrease in Jwb/Jbb would lead to a ground state with a higher
ST, as in the MnIII complex.
4
AC Magnetic Susceptibility Studies. Conclusive evidence
that these [V4O2]8+ complexes function as single-molecule
magnets was obtained with AC magnetic susceptibility mea-
surements. If, as indicated by one of two fits of the variable-
field magnetization data, the ST ) 3 ground state of 1 and 5
experiences ZFS with D ) -1.5 cm-1, then the potential energy
diagram in Figure 9 is applicable. This diagram is a plot of
the potential energy of a single molecule vs the magnetization
direction. The ST ) 3 ground state is split into ms ) (3, (2,
(1, and 0 levels. In zero magnetic field and when D < 0,
there are two states with the same lowest energy value, the ms
) 3 and ms ) -3 states. The ms ) 3 state can be viewed as
the case where the magnetic moment of an individual molecule
is “up” and the ms ) -3 state where the moment is “down”.
The double well shown in Figure 9 represents the change in
potential energy as a single molecule converts from spin up to
At temperatures below 4.0 K, there are signs of slow magnetic
relaxation for complex 1. The value of ø′MT decreases from
4.13 cm3 K mol-1 at 4.0 K to 3.47 cm3 K mol-1 at 1.7 K. This
decrease could be due to an out-of-phase AC susceptibility signal
(ø′′M), i.e., the temperature has been reduced to such a low value
that the molecule cannot easily reverse the direction of its
magnetization and therefore cannot keep in phase with the
oscillating magnetic field. If so, a ø′′M signal should be seen.
Indeed, complex 1 does exhibit a ø′′M signal at low temperatures

2374 J. Am. Chem. Soc., Vol. 120, No. 10, 1998
Castro et al.
Figure 10. Plots of ø′MT vs T (top) and ø′′M vs T (bottom) for [V4O2(O2-
CEt)7(bpy)2](ClO4) (1) in a 1.0 G AC field oscillating at the indicated
frequencies (and with no applied DC field), where ø′M and ø′′M are the
in-phase and out-of-phase magnetic susceptibilities, respectively.
(Figure 10), and, furthermore, it is frequency-dependent.
However, the ø′′M signal is quite weak: at its maximum at the
lowest accessible temperature of 1.70 K, it is only ∼1-2% of
the in-phase ø′M signal.
Figure 11. Plots of the out-of-phase AC susceptibility (ø′′M) vs T for
(NEt4)[V4O2(O2CEt)7(pic)2] (5) in a 1.0 G AC field oscillating at 1000
Hz and applied DC fields 0-2.0 kG (top) and 3.0-8.0 kG (bottom).
Additional experiments were carried out to firmly establish
that these [V4O2]8+ complexes are SMMs. If they do indeed
have the double-well potential shown in Figure 9, then the
application of an external DC field would affect the temperature
at which the ø′′M signal is seen.16 The application of a DC
field changes the double-well potential. An increase in the DC
field leads to one-half of the double well becoming stabilized
relative to the other half, i.e., the ms ) -3 level will be at lower
energy than the ms ) 3 level. Furthermore, the barrier becomes
larger as the DC field is increased. Thus, a DC field leads to
the ø′′M AC peak moving to higher temperatures. The effect
of an external DC field was examined for complexes 1 and 5.
It was found that the ø′′M signals for both complexes were
shifted to higher temperature with an increasing DC field. The
ø′′M signal for complex 1 became progressively larger with DC
fields up to 10.0 kG, but it did not shift enough to show a peak
in the ø′′M vs T curve. On the other hand, the ø′′M signal for
the picolinate complex 5 did shift to high enough temperatures
to show a peak in the ø′′M vs T plot (Figure 11). With the 1.0
G AC field oscillating at 1000 Hz, there is only a small
frequency-dependent ø′′M signal seen when the DC field is zero,
but an increase in the DC field to 1.0 kG leads to an increase
in the magnitude of the ø′′M signal as the peak in the ø′′M vs T
curve is presumably shifting to higher temperatures. Finally,
when the DC field is 2.0 kG, there is a peak in the ø′′M vs T
curve. Further increases in the DC field beyond 2.0 kG lead to
a further shift to higher temperatures, until DC fields of ∼5.0
kG, above which the temperature of the peak remains constant
at ∼2.2 K (Figure 12).
Figure 12. Variation of the temperature of the ø′′M peak maximum
(in a 1.0 G AC field oscillating at 1000 Hz) as a function of applied
DC field.
K. Further reduction to 50 Hz causes the ø′′M peak to shift to
below 1.70 K. All of the above AC magnetic susceptibility
data conclusively establish that complexes 5 and 1 function as
SMMs. When magnetization vs magnetic field data are
collected for these complexes at temperatures below 1.7 K, it
is possible that a magnetization hysteresis loop may be observed.
It was noted that the peak in the ø′′M vs T data moved to
temperatures > 1.7 K for complex 5 when the external DC field
was increased, whereas the ø′′M vs T peak did not move above
1.7 K for complex 1. This can be rationalized by reference to
Figure 9. The barrier (U) for reversal of the magnetization
direction is U ) |D|S2 for integer S, which is equal to 9|D| for
The frequency dependence of the ø′′M vs T peak for complex
5 was also studied. With the DC field held constant at 5.5 kG,
an appreciable frequency dependence is seen (Figure 13). As
the frequency of the AC field is changed from 1000 to 500 to
250 Hz, the peak in the ø′′M signal shifts from 2.2 to 2.0 to 1.7

Single-Molecule Magnets: Tetranuclear Vanadium(III) Complexes
J. Am. Chem. Soc., Vol. 120, No. 10, 1998 2375
Concluding Comments
The described synthesis procedures provide convenient access
to the new family of [V4O2(O2CR)7(bpy)2]+ (R ) Et, Ph) species
and certain Me-substituted bpy derivatives. Yields and purity
are good, and the procedures involve readily available starting
materials. X-ray crystallographic studies have confirmed that
the representative complex 1 contains a butterfly [V4O2]8+ core,
making this the fourth member of the [M4O2(O2CR)7(bpy)2]+
series, which currently encompasses M ) VIII, CrIII, MnIII, and
FeIII; it is possible that the series could be extended to include
the, e.g., TiIII and CoIII analogues, but this has yet to be explored.
The structures of the four complexes are essentially superim-
posable.
DC magnetic susceptibility studies show that [V4O2]8+
complex 1 has a ST ) 3 ground state and that, as might be
expected from the differing electron counts involved, the four
members of the [M4O2] family have differing ground-state spins,
i.e., ST ) 3 (V and Mn), 1 (Cr), or 0 (Fe). Both the ST ) 3
value for 1 and the variation in the M ) V-Fe series can be
readily rationalized using straightforward ideas of spin frustra-
tion arising from the particular combinations of magnitudes and
signs of the exchange interactions present in the [M4O2]8+ cores.
This is particularly satisfying given the relatively large size and
relatively low symmetry of these species.
Figure 13. Plot of the out-of-phase AC susceptibility (ø′′M) vs T for
(NEt4)[V4O2(O2CEt)7(pic)2] (5) in a 1.0 G AC field oscillating at the
indicated frequencies and with a 5.5 kG DC field.
a SMM with a S ) 3 ground state. Thus, the barrier height
depends on the magnitude of D. If 1 has a smaller value of
|D| than 5, then it would have a smaller barrier, and it would
consequently require a lower temperature for complex 1 to show
slow magnetization relaxation. Thus, its ø′′M vs T peak would
occur at a lower temperature than that for 5. It is known that
fitting variable-field magnetization data does not give a very
precise determination of D; in the analysis of our data, we
obtained comparable D values for complexes 1 and 5. High-
field EPR data are needed to get better determinations of D for
these complexes.
AC magnetic susceptibility experiments, both with and
without an external DC magnetic field, conclusively establish
that these [V4O2]8+ complexes are new examples of single-
molecule magnets. The combination of a ST ) 3 ground state
2
and a negatiVe magnetoanisotropy associated with ZFS (DSz ),
where D ≈ -1.5 cm-1, gives a large enough barrier (U ≈ 13.5
cm-1) to magnetization reversal that the complexes exhibit slow
relaxation of magnetization. For comparison, the Mn12 SMMs
have ST ) 9 or 10 ground states with D ≈ -0.5 to -0.6 cm-1
and a barrier U ≈ 50 cm-1, giving out-of-phase AC susceptibil-
ity signals (ø′′M) in the 4-7 K (ST ) 10) or 2-4 K (ST ) 9)
regions. The [Mn4O3X]6+ distorted cubane complexes have ST
) 9/2, D ≈ -0.3 to -0.4 cm-1, and U ≈ 6-8 cm-1 and exhibit
ø′′M signals in the 2-3 K range.
Additional comments about the DC field dependence of the
ø′′M AC peak temperature can be made. Since the barrier for
magnetization reversal is ∼13.5 cm-1 at temperatures below 3
K where ø′′M vs T peaks are seen for complex 5, there is not
nearly enough thermal energy to go over the barrier. Each
complex must reverse its magnetization direction by quantum
mechanically tunneling from one side of the double-well
potential curve to the other side. The tunneling is due to the
presence of a transverse magnetic field or zero-field interaction
terms. The transverse magnetic field arises from a combination
of the external magnetic field or an internal magnetic field from
either the nuclear spins of the atoms or dipolar interactions
between [V4O2]8+ molecules. The rate of tunneling will vary
as the magnitude of the external DC field is changed.16 At zero
DC field, the energy levels (ms ) -3, -2, -1) on the right-
hand of the double well are aligned with those on the left side
(ms ) 3, 2, 1), and the rate of tunneling is relatively large. This
is termed “resonant magnetization tunneling”. When a DC field
is introduced, the energy levels on the two sides of the double
well are not aligned, and the rate of tunneling will be less than
when the DC field is zero. If the DC field is increased to a
value that the ms ) 3 level lines up with the ms ) -2 level,
then there would be another increase in the rate of tunneling.
At this large field, the ø′′M vs T peak temperature would also
decrease. Thus, the DC field dependence of the ø′′M vs T peak
temperature shown in Figure 11 for complex 5 is understandable.
There is a decrease at zero DC field, and the peak temperature
is constant in the 5-10 kG range. If the DC field was increased
to a value given by ∆H ) |D|/gµB, then the ms ) 3 level would
be aligned with the ms ) -2 level, and the ø′′M peak temperature
would again show a decrease. However, it would require a field
of ∼17.0 kG to see this second dip in the ø′′M peak temperature,
and unfortunately our AC SQUID used has a DC field limitation
of 10.0 kG.
The identification of new SMMs outside Mn and Fe chemistry
is a significant development and augurs well for the future
breadth and scope of this new phenomenon. The negative
anisotropy of VIII and the [V4O2]8+ complexes suggest that
additional SMM examples may be found in this area of early
transition metal chemistry as long as ground state spin values
are also sufficiently large, as might result from spin frustration
effects within appropriate Vx topologies. Also significant are
1
the results of the 1D and 2D H COSY NMR studies that
establish that the complexes retain their structure on dissolution
in CH2Cl2. Among the many advantages of SMMs over normal
nanomagnets are single, sharply defined size and solubility, the
latter having obvious benefits in future technological applications
that might require, e.g., thin films. Confirming the retention
of structural and therefore magnetic integrity on dissolution
makes these potential applications feasible.
Acknowledgment. This work was supported by the National
Science Foundation.
Supporting Information Available: Tables of X-ray data
collection and structure solution details, fractional coordinates,
thermal parameters, and interatomic distances and angles (14
pages). See any current masthead page for ordering information
and Web access instructions.
JA9732439