Trends in metal-biradical exchange interaction for first-row MII(nitronyl nitroxide-semiquinone) complexes

Article abstract of DOI:10.1021/ja020715x

We report molecular structures and temperature-dependent magnetic susceptibility data for several new metal complexes of heterospin triplet ground-state biradical ligands. The ligands are comprised of both nitronyl-nitroxide (NN) and semiquinone (SQ) spin

Full text of DOI:10.1021/ja020715x

Published on Web 01/18/2003
Trends in Metal-Biradical Exchange Interaction for First-Row
MII(Nitronyl Nitroxide-Semiquinone) Complexes
David A. Shultz,*,† Kira E. Vostrikova,† Scot H. Bodnar,† Hyun-Joo Koo,†
Myung-Hwan Whangbo,*,† Martin L. Kirk,*,‡ Ezra C. Depperman,‡ and
Jeff W. Kampf§
Contribution from the Department of Chemistry, North Carolina State UniVersity,
Raleigh, North Carolina 27695-8204, Department of Chemistry, UniVersity of New Mexico,
Albuquerque, New Mexico 87131, and Department of Chemistry, UniVersity of Michigan,
Ann Arbor, Michigan 48109-1055
Received May 21, 2002; Revised Manuscript Received November 1, 2002; E-mail: david_shultz@ncsu.edu;
mike_whangbo@ncsu.edu; mkirk@unm.edu
Abstract: We report molecular structures and temperature-dependent magnetic susceptibility data for several
new metal complexes of heterospin triplet ground-state biradical ligands. The ligands are comprised of
both nitronyl-nitroxide (NN) and semiquinone (SQ) spin carriers. Five compounds are five-coordinate MII
complexes (M ) Mn, Co, Ni, Cu, and Zn), and one is a six-coordinate NiII complex. Five compounds were
structurally characterized. During copper complex formation a reaction with methanol occurs to form a
unique methoxy-substituted SQ ring. Variable-temperature magnetic susceptibility studies are consistent
with strong intraligand (NN-SQ and NN-PhSQ) ferromagnetic exchange coupling. For the five-coordinate
Mn, Co, and Ni complexes, the S ) 1 ligand is antiferromagnetically coupled to the metal. For both the
five-coordinate Cu complex and the six-coordinate Ni complex, the ligand is ferromagnetically coupled to
the metal spins in accordance with orbital symmetry arguments. Despite the low molecular symmetries,
the predicted trend in metal-ligand exchange interactions is supported by spin dimer analysis based on
extended Hu¨ckel calculations. For (NN-SQ)NiTpCum,Me (TpCum,Me ) hydro-tris(3-cumenyl-5-methylpyrazolyl)-
borate), an antisymmetric exchange term was required for the best fit of the magnetic susceptibility data.
Antisymmetric exchange was less important for the other complexes due to inherently smaller ∆g. Finally,
it is shown that intraligand exchange coupling is of paramount importance in stabilizing high-spin states of
mixed metal-biradical complexes.
do not magnetically order at temperatures above ca. 40 K.6,7
Introduction
This is due to the far weaker superexchange interactions that
are attenuated by larger diamagnetic ligands8 and due to low
lattice dimensionality.
Recent efforts in molecular magnetism have produced a
plethora of fascinating molecule-based materials.1 One of the
most productive research areas in modern magnetochemistry
has been preparation and study of Prussian blue analogues. This
class of coordination polymers derives interesting and novel
properties from superexchange interactions mediated by bridging
cyanide. Interesting properties such as high-temperature mag-
netic ordering (T g 300 K),2 photomagnetism,3 and electro-
magnetism4 have been discovered and studied.
Weaker superexchange via large diamagnetic ligands can be
circumvented by using paramagnetic ligands. In this way the
inherently weaker superexchange interactions are dominated by
direct exchange between the metal and the paramagnetic ligand.
The most common paramagnetic ligands include TCNE radical
anion,9 nitroxide-type radicals,10,11 semiquinones,12 and several
interesting new verdazyl-type molecules.13 Application of the
However, coordination polymers comprised of diamagnetic
ligands only slightly larger than cyanide (e.g., oxalate and azide5)
(5) Ribas, J.; Escuer, A.; Monfort, M.; Vicente, R.; Cortes, R.; Lezama, L.;
Rojo, T. Coord. Chem. ReV. 1999, 195, 1027.
* To whom correspondence should be addressed. Telephone: (919) 515-
6972 (D.A.S.); (919) 515-3464 (M.-H.W.); (505) 277-5992 (M.L.K.). Fax:
(919) 515-8920 (D.A.S.); (505) 277-2609 (M.L.K.). Web: http://www.nc-
su.edu/chemistry/das.html (D.A.S.); http://www.ncsu.edu/chemistry/mh-
w.html (M.-H.W.); http://www.unm.edu/∼graduate/research/kirk.html
(M.L.K.).
(6) Decurtins, S.; Schmalle, H. W.; P., S.; Ensling, J.; Gu¨tlich, P. J. Am. Chem.
Soc. 1994, 116, 9521.
(7) Armentano, D.; De Munno, G.; Lloret, F.; Palii, A. V.; Julve, M. Inorg.
Chem. 2002, 41, 2007.
(8) For exceptions to weak exchange coupling through polyatomic, diamagnetic
ligands, see the following and references therein: Ung, V. A.; Thompson,
A.; Bardwell, D. A.; Gatteschi, D.; Jeffery, J. C.; McCleverty, J. A.; Totti,
F.; Ward, M. D. Inorganic Chemistry 1997, 36, 3447.
(9) Manriquez, J. M.; Yee, G. T.; McLean, R. S.; Epstein, A. J.; Miller, J. S.
Science 1991, 252, 1415.
† North Carolina State University.
‡ University of New Mexico.
§ University of Michigan.
(1) Miller, J. S.; Drillon, M. Magnetism: Molecules to Materials II: Molecule-
Based Materials; Wiley-VCH: New York, 2001; p 489.
(2) Ferlay, S.; Mallah, T.; Ouahes, R.; Veillet, P.; Verdaguer, M. Nature 1995,
378, 701.
(10) Caneschi, A.; Gatteschi, D.; Sessoli, R.; Rey, P. Acc. Chem. Res. 1989,
22, 392.
(11) Iwamura, H.; Inoue, K.; Haymizu, T. Pure Appl. Chem. 1996, 68, 243.
(12) Shultz, D. A. Comments Inorg. Chem. 2002, 23, 1.
(13) Hicks, R. G. Aust. J. Chem. 2001, 54, 597.
(3) Sato, O.; Iyoda, T.; Fujishima, A.; Hashimoto, K. Science 1996, 272, 704.
(4) Sato, O.; Iyoda, T.; Fujishima, A.; Hashimoto, K. Science 1996, 271, 49.
9
10.1021/ja020715x CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 1607-1617
1607

A R T I C L E S
Shultz et al.
metal-radical coupling approach has resulted in magnetically
ordered systems, with a high ordering temperature of 46 K.14
Electroactivity is another attractive feature common to many
classes of paramagnetic ligands, and is a key element in the
mechanism of valence tautomerism.15,16
An alternative to the “solid-state approach” to magnetic
materials is the synthesis of single molecules having both a high-
spin ground-state and appreciable magneto-anisotropy. Such
molecules have been termed “single-molecule magnets.”17 If
appropriately designed molecules can be prepared, then indi-
vidual molecules could be magnetized, thus obviating the need
for extended solids and long-range magnetic ordering.
The successful creation of novel materials and magnetizable
solids using either the extended-solid or the single-molecule
approach necessitates the search for new paramagnetic ligands.
Along these lines, we recently reported the preparation and
magnetic properties, of nitronyl nitroxide-semiquinone (NN-
SQ) complexes of ZnII and CuII.18 Analysis of the variable-
temperature magnetic susceptibility data suggested strong
intraligand NN-SQ exchange coupling. Variable-temperature
variable-field magnetic circular dichroism studies of the CuII
derivative were reported subsequently and illustrated the robust-
ness of the exchange couplings in frozen solution and polymer
matrices, where NN-SQ bond torsions are likely to deviate from
values observed in the solid state.19 To continue our study of
(NN-SQ)M complexes, we report molecular structures for five-
coordinate first-row MII (MII ) Mn, Co, Ni, and Cu) complexes
of NN-SQ. A six-coordinate NiII species has also been prepared,
but X-ray quality crystals could not be obtained. Our Mn
complex is the first structurally and magnetically characterized
mono-SQ-MnII species. We also report the preparation, molec-
ular structure and intra-ligand exchange coupling constant in a
methylene chloride, and methanol were distilled from CaH2 under argon.
tert-Butyllithium (1.5 M in pentane) was used as received from Acros
Chemical. Other chemicals were purchased from Aldrich Chemical Co.
Column and radial chromatography were carried out using silica gel
(230-400 mesh for column). X-Band EPR spectroscopy and electro-
chemistry were performed as described previously.22 Fluid solution EPR
spectra were simulated using WinSim.23 NMR spectra were recorded
at 300 MHz for 1H NMR and 75 MHz for 13C NMR in CDCl3 solution
if not otherwise specified. Elemental analyses were performed by
Atlantic Microlab, Inc., Norcross, GA. Mass spectrometry was carried
out at the NC State University Mass Spectrometry Facility. UV-visible
spectra were collected on either an OLIS-converted Cary 14, or on a
Hitachi U-3501 double beam UV-vis-NIR spectrophotometer equipped
to collect data in the wavelength range from 180 to 3200 nm.
Synthesis. Catechol NN-CatH2 and the corresponding metal com-
plexes were prepared as described previously,18 and the synthesis write-
ups for new compounds can be found in the Supporting Information.
Crystallographic Structure Determinations. Experimental infor-
mation for the structure determinations of the metal complexes can be
found in Supporting Information.
Magnetic Susceptibility Measurements. Magnetic susceptibilities
were measured on a Quantum Design MPMS-XL7 SQUID magne-
tometer using an applied field of 0.5 T for Curie plots. Saturation
magnetization values (see Supporting Information) are consistent with
the spin of the ground states as described below. Microcrystalline
samples were loaded into the sample space of a Delrin sample holder
and mounted to the sample rod using string. Data were corrected for
sample holder and molecular diamagnetism using Pascal’s constants.
The decrease in the øT data at low temperatures was accounted for
with a Weiss correction, using the expression øeff ) ø/(1 - ϑø), where
ϑ ) 2zJ′inter /(Ng2â2).24 The origin of J′inter may be zero-field splitting,
intermolecular interaction, saturation effects or some combination of
all three.25 The other terms have their usual meaning.24 The data are
plotted as øT vs T (Figure 2), except for (NN-SQ)NiTpCum,Me which is
also displayed as ø vs T in Figure 4.26
phenyl-spaced derivative of NN-SQ: (NN-PhSQ)ZnTpCum,Me
.
Our experimental results are accompanied by spin dimer analysis
based on extended Hu¨ckel computations. Our results constitute
the first magnetostructural correlations in a series of biradical-
metal ion complexes. This combined experimental/theoretical
approach has shown to be successful in simpler systems, most
notably halide-bridged dinuclear copper(II) complexes.20,21 Our
results confirm the strong ferromagnetic NN-SQ exchange
coupling that is insensitive to NN-SQ bond torsions. Finally,
we illustrate that intraligand exchange coupling is of paramount
importance in stabilizing high-spin states of mixed metal-
biradical complexes.
Experimental Section
General. Unless noted otherwise, reactions were carried out in oven-
dried glassware under nitrogen atmosphere. THF and toluene were
distilled under argon from sodium benzophenone ketyl, and acetonitrile,
(22) Shultz, D. A.; Farmer, G. T. J. Org. Chem. 1998, 63, 6254.
(23) Duling, D. EPR Calculations for MS-Windows NT/95 Software Tools,
Version 0.96 ed.; Public EPR Software Tools, National Institute of
Environmental Health Sciences, National Institutes of Health: Triangle
Park, NC, 1996.
(24) O’Connor, C. J. Prog. Inorg. Chem. 1982, 29, 203.
(25) Caneschi, A.; Dei, A.; Mussari, C. P.; Shultz, D. A.; Sorace, L.; Vostrikova,
K. E. Inorg. Chem. 2002, 41, 1086. Note that these authors used H ) (JS1‚
S2. For comparison of J-values, we have recast their experimentally
determined exchange parameters using a H ) -2JS1‚S2 exchange Hamil-
tonian.
(26) For antiferromagnetically coupled homospin complexes, the ø vs T plot is
often displayed since a maximum is observed in ø, and is therefore
preferable for fitting the data. However, the strongly paramagnetic ground
state dominates the shape of the ø vs T plot for all complexes except (NN-
SQ) NiTpCum,Me (see text), where no maximum in ø vs T is observed. Thus,
the curvature of the øT vs T plots are greater and provide for better analysis
of the data.
(14) Inoue, K.; Hayamizu, T.; Iwamura, H.; Hashizume, D.; Ohashi, Y. J. Am.
Chem. Soc. 1996, 118, 1803.
(15) Shultz, D. A. Valence Tautomerism in Dioxolene Complexes of Cobalt.
In Magnetism: Molecules to Materials II: Molecule-Based Materials;
Miller, J. S., Drillon, M., Eds.; Wiley-VCH: New York, 2001; pp 281-
306.
(16) Pierpont, C. G. Coord. Chem. ReV. 2001, 216, 99.
(17) Christou, G.; Gatteschi, D.; Hendrickson, D. N.; Sessoli, R. MRS Bulletin
2000, 66.
(18) Shultz, D. A.; Bodnar, S. H.; Vostrikova, K. E.; Kampf, J. W. Inorg. Chem.
2000, 39, 6091.
(19) Depperman, E.; Bodnar, S. H.; Vostrikova, K. E.; Shultz, D. A.; Kirk, M.
L. J. Am. Chem. Soc. 2001, 123, 3133.
(20) Crawford, V. H.; Richardson, H. W.; Wasson, J. R.; Hodgson, D. J.;
Hatfield, W. E. Inorg. Chem. 1976, 15, 2107.
(21) Hay, P. J.; Thibeault, J. C.; Hoffmann, R. J. Am. Chem. Soc. 1975, 97,
4884.
9
1608 J. AM. CHEM. SOC. VOL. 125, NO. 6, 2003

Exchange Interaction Studies of (NN-SQ)M Complexes
A R T I C L E S
Scheme 1
The complex (NN-SQ)NiL was prepared from ligand 3,36
nickel(II)perchlorate, and NN-CatH2 in the presence of potas-
sium hydroxide.
A model compound, NN-Ar (Ar ) 5-tert-butyl-3,4-dimethox-
yphenyl) was prepared as shown in Scheme 1. Bromide 431 was
subjected to lithium-halogen exchange, and the resulting aryl-
lithium reagent was quenched with DMF to give the corre-
sponding aldehyde. The aldehyde was subsequently transformed
into NN-Ar using standard protocol.
Extended Hu1ckel Calculations. For a number of different magnetic
solids, it has been shown that trends in spin exchange interactions are
well described by spin dimer analysis based on extended Hu¨ckel (EH)
calculations.27 Thus, we performed EH calculations on (NN-SQ)-
MTpCum,Me to determine the trend in both sign and magnitude of
exchange coupling between metal ion and NN-SQ. The crystal structures
of these compounds were used as structural inputs, and fragment
molecular orbital analysis calculations (see below) were carried out
by employing the CAESAR program package.28 The results are displayed
in Table 3.
X-ray Crystallography. The crystal structure of (NN-SQ)-
ZnTpCum,Me was reported previously, and thermal ellipsoid plots
of the new compounds are shown in Figure 1. Crystallographic
data are collected in Table 1. A tabular collection of bond
lengths and torsion angles can be found in the Supporting
Information.
Results and Analysis
Synthesis. Catechol NN-CatH2 and the corresponding metal
complexes were prepared as described previously.18 Despite
several attempts, we have been unable to prepare the FeII
complex of NN-SQ.29 Because iron(II) offers the possibility for
either an intermediate spin complex, a spin-crossover complex,
or a complex having large zero-field splitting, synthetic efforts
to prepare an FeII complex of NN-SQ continue. The phenyl-
spaced derivative of NN-CatH2, NN-PhCatH2, its ZnTpCum,Me
complex, and (NN-SQ)NiL having a six-coordinate NiII were
prepared according to the Scheme 1.
Boronic ester 130 was reacted with commercially available
4-bromobenzaldehyde under Suzuki coupling conditions,31 fol-
lowed by deprotection to give aldehyde 2 in excellent yield.
Compound 2 was reacted with 2,3-dimethyl-2,3-bis(hydroxyami-
no)butane,32 followed by oxidation to yield NN-PhCatH2.
Complex formation with HOZnTpCum,Me 33 followed standard
procedures.34,35
Magnetic Susceptibility. The magnetic susceptibility data
for all of the compounds are plotted in Figures 2 and 4. The
intraligand exchange coupling has previously been shown to
be strongly ferromagnetic in (NN-SQ)ZnTpCum,Me 18
, with only
the triplet being populated at temperatures as high as 300 K.
Therefore, a lower limit of JNN-SQ was estimated to be +310
cm-1. Because of the strong intraligand exchange coupling, the
exchange Hamiltonian in eq 1 has been used for all of the (NN-
SQ)ML complexes and the data have been fit using the field-
independent van Vleck expression. Here, J is the exchange
interaction between the paramagnetic metal and the strongly
(27) Koo, H.-J.; Whangbo, M.-H. Inorg. Chem. 2001, 40, 2161.
(28) Ren, J.; Liang, W.; Whangbo, M.-H. Crystal and Electronic Structure
Analysis Using CAESAR; PrimeC, 1998.
(29) The crude reaction mixture shows none of the characteristic NN-SQ
electronic absorption bands, rather a broad band centered near 16700 cm-1
(600 nm), consistent with the (NN-Cat)FeIIITpCum,Me form, see: Heistand,
R. H., II; Lauffer, R. B.; Fikrig, E.; Que, L., Jr. J. Am. Chem. Soc. 1982,
104, 2789.
(30) Shultz, D. A.; Hollomon, M. G. Chem. Mater. 2000, 12, 580.
(31) Shultz, D. A.; Boal, A. K.; Driscoll, D. J.; Kitchin, J. R.; Tew, G. N. J.
Org. Chem. 1995, 60, 3578.
(34) Shultz, D. A.; Bodnar, S. H.; Kumar, R. K.; Kampf, J. W. J. Am. Chem.
Soc. 1999, 121, 10664.
(32) Vostrikova, K. E.; Hirel, C.; Pecaut, J.; Rey, P. Chem. Eur. J. 2001, 7,
2007.
(35) Ruf, M.; Noll, B. C.; Groner, M. D.; Yee, G. T.; Pierpont, C. G. Inorg.
Chem. 1997, 36, 4860.
(33) Ruf, M.; Weis, K.; Vahrenkamp, H. J. Chem. Soc., Chem. Commun. 1994,
135.
(36) Uma, R.; Viswanathan, R.; Palaniandavar, M.; Lakshminarayanan, M. J.
Chem. Soc., Dalton Trans. 1994, 1219.
9
J. AM. CHEM. SOC. VOL. 125, NO. 6, 2003 1609

A R T I C L E S
Shultz et al.
Figure 1. Thermal ellipsoid drawings of (NN-SQ)MTpCum,Me and (NN-PhSQ)ZnTpCum,Me. Cumenyl groups and hydrogens have been omitted for clarity.
Table 1. Crystallographic Data for (NN-SQ)MnTpCum,Me, (NN-SQ)CoTpCum,Me, (NN-SQ)NiTpCum,Me, and [NN-SQ(OMe)]CuTpCum,Me and
(NN-PhSQ)ZnTpCum,Me
C H BMnN O
C H BCoN O
C H BN NiO
C H BCuN O
C H BN O Zn
62 73 8 4
56 69
8
4
56 69
8
4
56 69
8
4
57 71
8
5
chemical formula
(NN-SQ)MnTpCum,Me
(NN-SQ)CoTpCum,Me
(NN-SQ)NiTpCum,Me
[NN-SQ(OMe)]CuTpCum,Me
(NN-PhSQ)ZnTpCum,Me
a/Å
b/Å
c/Å
R/deg
â/deg
γ/deg
V/Å3
12.828(5)
17.777(7)
24.075(9)
90
104.274(6)
90
12.582(3)
17.693(4)
24.370(5)
90
98.401(4)
90
12.1536(9)
13.2557(10)
17.3931(12)
96.801(10)
107.518(10)
98.335(10)
2604.6(3)
2
12.2081(8)
13.2504(9)
17.3758(11)
96.958(3)
105.603(3)
99.577(2)
2628.4(3)
2
14.230(5)
17.950(6)
25.171(8)
80.091(5)
75.019(6)
82.382(6)
6092(4)
4
5321(4)
4
5366.9(19)
4
Z
formula weight
space group
T/K
983.94
monoclinic P21/c
158(2)
1.228
987.93
monoclinic P21/c
158(2)
1.223
987.71
triclinic P1h
108(2)
1022.57
triclinic P1h
158(2)
1070.46
triclinic P1h
158(2)
F
calcd/g cm-3
1.259
1.292
1.167
Ra
0.1258d
-
0.1086d
-
0.0459d
0.0993d
-
0.1545d
-
b
Rw
-
0.0818
wR2 (F2)c
0.2668
0.2306
0.2100
0.2512
2
4
a R ) Σ(|Fo| - |Fc|)/ΣFo. b Rw ) [Σ w(Fo - Fc)/(ΣwFo )]1/2
.
c wR2 ) [Σw(Fo - Fc)/(ΣwFo )]1/2; GoF ) S ) [Σw(Fo - Fc)/(n - p)]1/2, where n ) number
of reflections and p ) number of parameters. d For [F2 > 2σ(F2)].
coupled NN-SQ ligand, Sˆ1 is the spin operator for the NN-SQ
ligand (SNN-SQ ) 1), and Sˆ2 is the spin operator for the metal
ion. A similar approach has been used for other high-spin
organic molecules.37 The best-fit parameters for the øT data are
given in Table 2, and qualitative energy level diagrams are given
in Figure 3.
the other symbols have their usual meaning. Inspection of Figure
2 indicates that the exchange interaction is clearly antiferro-
magnetic, yielding an ST ) 3/2 ground state which lies 5J below
5
7
the sextet (ST ) /2) and 12J below the ST ) /2 state. The
exchange coupling constant was determined from the best fit
of eq 2 to the data, yielding J ) -41.3 cm-1
.
The øT vs T data for the five-coordinate nickel complex, (NN-
SQ)NiTpCum,Me (Figure 2) reveal the presence of dominant
antiferromagnetic interactions between SNi ) 1 and SNN-SQ ) 1
spin centers leading to a diamagnetic ST ) 0 ground state.
Additionally, a small paramagnetic impurity manifests itself in
H ) - 2JSˆ1‚Sˆ2
(1)
The øT product for (NN-SQ)MnTpCum,Me is given by eq 2:
Ng2â2 15 + 52.5 e-5J/kT + 126 e-12J/kT
øT )
(2)
4 + 6 e-5J/kT + 8 e-12J/kT
3k
(37) Jacobs, S. J.; Shultz, D. A.; Jain, R.; Novak, J.; Dougherty, D. A. J. Am.
Chem. Soc. 1993, 115, 1744.
where J is the metal-ligand exchange coupling constant and
9
1610 J. AM. CHEM. SOC. VOL. 125, NO. 6, 2003

Exchange Interaction Studies of (NN-SQ)M Complexes
A R T I C L E S
in Table 2 are an average of the values obtained from ø vs T
and øT vs T fits.
The data for six-coordinate nickel complex (NN-SQ)NiL
were fit using eq 5:
Ng2â2 6 e-2J/kT + 30 e-6J/kT
3k
øT )
(5)
1 + 3 e-2J/kT + 5 e-6J/kT
The øT product is consistent with ferromagnetic coupling
between nickel and NN-SQ spins. The best-fit of eq 5 to the
susceptibility data yielded g ) 2.04 and J ) +79.9 cm-1
.
Finally, for (NN-PhSQ)ZnTpCum,Me, the ligand spins are
ferromagnetically coupled and were fit with eq 6 by using J )
JNN-SQ and by fixing g ) 2. Thus, for the phenyl-spaced
derivative of the NN-SQ ligand, the intraligand exchange
interaction is attenuated, and the excited singlet lies 2JNN-SQ
200 cm-1 above the ground-state triplet.
)
Ng2â2
3k
6
øT )
(6)
3 + e-2J/kT
Figure 2.
ø
paraT vs T plots for (NN-SQ)MTpCum,Me, (NN-PhSQ)-
ZnTpCum,Me, and (NN-SQ)NiL.
Interestingly, the øT data for (NN-SQ)CoTpCum,Me cannot
be fit using simple spin-only susceptibility expressions, and this
is likely a manifestation of large spin-orbit coupling effects
within the CoII 4T parent ground state. However, the metal-
ligand interaction is clearly antiferromagnetic, and this is
consistent with Pierpont’s (3,5-DBSQ)CoTpCum,Me complex.35
Trends in Antiferromagnetic Spin Exchange Interactions.
The exchange interactions between the SNN-SQ ) 1 organic
ligand NN-SQ and those of the divalent transition metal ion
are antiferromagnetic for (NN-SQ)NiTpCum,Me (J ) -87.8
cm-1) and (NN-SQ)MnTpCum,Me (J ) -41.3 cm-1), but
ferromagnetic for [NN-SQ(OMe)]CuTpCum,Me (J ) +75.6
cm-1). The exchange parameter J can be written as the sum of
ferro- and antiferromagnetic comtributions (J ) JF + JAF).21,40
Often, the ferromagnetic term JF is small, and J only becomes
ferromagnetic when the antiferromagnetic term JAF is negligible.
Therefore, trends in metal-ligand antiferromagnetic exchange
interactions can be discussed solely on the basis of the
antiferromagnetic terms JAF. The J-values of (NN-SQ)MT-
pCum,Me determined in the present work imply that the magnitude
of antiferromagnetic exchange interaction between the spins of
NN-SQ and MII decreases in the order, (NN-SQ)NiTpCum,Me
> (NN-SQ)MnTpCum,Me > [NN-SQ(OMe)]CuTpCum,Me and
that the antiferromagnetic exchange interaction is very weak in
[NN-SQ(OMe)]CuTpCum,Me. In this section we explore these
implications by analyzing the experimentally determined metal-
ligand exchange interactions present in the (NN-SQ)MTpCum,Me
complexes.
the form of a Curie tail in the ø vs T data at low temperatures
(see Figure 4).
Therefore, the susceptibility data for (NN-SQ)NiTpCum,Me
were initially fit with a model that takes into account a
paramagnetic impurity term. However, best fits to the data
yielded unusually low g-values (g ≈ 1.7) for a five-coordinate
NiII-biradical pair (gNi > 2.0, gNN-SQ ≈ 2.0),38 requiring
reassessment of the model employed. Since the SNi ) 1 and
SNN-SQ ) 1, spin centers in (NN-SQ)NiTpCum,Me are related by
neither an inversion center nor a Cn rotation axis, antisymmetric
exchange interactions39 may be operative. This additional term
in the spin Hamiltonian is given by:39
Hantisymm ) dab[Sa × Sb]
(3)
Antisymmetric exchange describes a mechanism for the
mixing of two states that differ in spin multiplicity. In the case
of (NN-SQ)NiTpCum,Me, the ground state is no longer strictly
diamagnetic due to the triplet admixture, resulting in a finite
value for the low-temperature susceptibility. Thus, the suscep-
tibility expression was modified to include a Boltzmann-
weighted paramagnetic contribution to the low-temperature
susceptibility in addition to the paramagnetic impurity term.
Ng2â2S(S + 1)
3kT
Ng2â2 6 e-2J/kT + 30 e-6J/kT
3kT
ø )
+
+
1 + 3 e-2J/kT + 5 e-6J/kT
[
]
Since we are interested in understanding qualitative trends
in the magnetic properties of closely related systems, it is
sufficient to estimate the relative magnitudes of the exchange
interactions solely on the basis of extended Hu¨ckel calcula-
tions.27 Consider that each spin carrier in an exchange-coupled
dimer contains one unpaired electron and the two spin sites are
represented by magnetic orbitals φa and φb (i.e., singly occupied
molecular orbitals representing the spin sites a and b, respec-
tively). If tab is the hopping integral (i.e., the resonance integral)
between φa and φb, then the JAF term is related to tab as:21,41,42
Cantisymm
1 + 3 e-2J/kT + 5 e-6J/kT
(4)
The best-fit of eq 4 to the susceptibility data yielded g )
1.8, J ) -91 cm-1, and Cantisymm ) 540 × 10-6 cm3/mol, with
<1% of an S ) 1 impurity that presumably results from the
uncoupled spin components. The best-fit parameters presented
(38) Benelli, C.; Dei, A.; Gatteschi, D.; Pardi, L. Inorg. Chem. 1990, 29, 3409.
Note that these authors used H ) (JS1‚S2. For comparison of J-values,
we have recast their experimentally determined exchange parameters using
a H ) -2JS1‚S2 exchange Hamiltonian.
JAF ) -2t2ab/Ueff
(7)
(39) Moriya, T. Phys. ReV. 1960, 117, 635.
9
J. AM. CHEM. SOC. VOL. 125, NO. 6, 2003 1611

A R T I C L E S
Shultz et al.
Table 2. Fit Parameters from Variable-Temperature Magnetic Susceptibility Measurements
-1
compound
g
J/cm
zJ′inter/cm-1a
C
antisymm/emu‚mol-1b
ref
(NN-SQ)MnTpCum,Me
(NN-SQ)CoTpCum,Me
(NN-SQ)NiTpCum,Me
[NN-SQ(OMe)]CuTpCum,Me
(NN-SQ)ZnTpCum,Me
2.03 ( 0.005
-41.3 ( 0.5
<0
-0.01 ( 0.004
-
-
this work
this work
this work
ref 18
-
-
1.78 ( 0.03b
2.02 ( 0.001
-87.8 ( 4.0c
+75.6 ( 1.0
+0.002 ( 0.001
-0.18 ( 0.02
<0
5.0 ( 0.4 × 10-4b
-
-
-
> +300c
-
-
-
-
ref 18
+550d
-
this work
this work
this work
(NN-PhSQ)ZnTpCum,Me
(NN-SQ)NiL
2.0 (fixed)
2.06 ( 0.002
+100.1 ( 3.2a
+78.6 ( 2.3
-0.26 ( 0.02
-0.31 ( 0.02
a Intermolecular correction term, see Experimental Section. b Antisymmetric exchange contribution, see text. c Fit parameters are an average from the øT
c
and ø vs T plots. JNN-SQ intraligand coupling. d From eq 11, see text.
Figure 3. Energy gaps for spin states using H ) -2JS1S2.
two adjacent spin sites have m and n unpaired spins, respec-
tively, the overall spin exchange parameter J is written as:41,43
m
n
1
J )
J
(8)
mn ∑∑
µν
µ)1ν)1
The estimation of qualitative trends in J on the basis of spin-
orbital interaction energies has been discussed elsewhere.44,45
However, this approach is more difficult to employ computa-
tionally for low-symmetry molecules such as (NN-SQ)MT-
pCum,Me (as Figure 6 would imply, vide infra). Thus we consider
the average of the hopping integral squared, t2 , defined by:
m
n
1
t2
)
t2
(9)
∑∑ µν
mn µ)1ν)1
Then the JAF value is approximated by:
JAF ) - 2 t2 /Ueff
(10)
Figure 4. øpara vs T plot for (NN-SQ)NiTpCum,Me
.
Let us now consider how to estimate the t2 values for (NN-
SQ)MTpCum,Me. The NN-SQ ligand contains two unpaired spins
(m ) 2), while the MII ion has five unpaired spins for MnII (n
) 5), two unpaired spins for NiII (n ) 2), and one unpaired
spin for CuII (n ) 1). The primary orbital interactions between
where Ueff is the effective on-site repulsion, which should be
nearly constant for closely related systems. (Note that eq 7 is
valid only when the spin Hamiltonian is expressed as Hˆ )
-2JSˆ1‚Sˆ2 instead of Hˆ ) -JSˆ1‚Sˆ2.) In a magnetic dimer where
superexchange interactions are operative, the two spin sites share
at least one common atom so that their magnetic orbitals are
not well defined for quantitative calculations of tab. In this case,
the value of tab is determined indirectly by performing molecular
orbital calculations for a spin dimer.27,41 For example, when
the two spin sites in an exhange-coupled dimer are equivalent,
the tab integral is related to the spin-orbital interaction energy,
∆e (i.e., the energy separation between the highest two singly
occupied energy levels of a spin dimer) by tab ) ∆e/2. When
the (NN-SQ) ligand and the (MIITpCum,Me +
) fragment in (NN-
SQ)MTpCum,Me involve the out-of-plane oxygen lone-pair
orbitals of the SQ fragment of the NN-SQ ligand. The magnetic
exchange interaction between the two spin sites of (NN-SQ)-
MTpCum,Me involve the two magnetic orbitals φµ ) φNN and
φSQ, which are both π-type molecular orbitals. It is not
(43) Charlot, M. F.; Kahn, O. NouV. J. Chim. 1980, 4, 567.
(44) Koo, H.-J.; Whangbo, M.-H.; Coste, S.; Jobic, S. J. Solid State Chem. 2001,
156, 464.
(45) Dai, D.; Koo, H.-J.; Whangbo, M.-H. Solid State Chemistry of Inorganic
Materials III. In MRS Symposium Proceedings; Geselbracht, M. J., Greedan,
J. E., Johnson, D. C., Subramanian, M. A., Eds.; Materials Research
Society: Warrendale, PA, 2001; Vol. 658, GG5.3.1-5.3.11.
(40) Kahn, O.; Briat, B. J. Chem. Soc., Faraday Trans. 2 1976, 72, 268.
(41) Kahn, O. Molecular Magnetism; VCH: New York, 1993.
(42) Whangbo, M.-H.; Koo, H.-J. Inorg. Chem. 2002, 41, 3570.
9
1612 J. AM. CHEM. SOC. VOL. 125, NO. 6, 2003

Exchange Interaction Studies of (NN-SQ)M Complexes
A R T I C L E S
Table 3. Comparison of the Calculated t2 and JAF Values of
(NN-SQ)MTpCum,Me with Their Experimental J-Values
primary effect this OMe group has on the molecule is to disrupt
the conjugation between NN and SQ units by imposing a torsion
angle of 36.1°. Despite this torsion, the intraligand exchange
remains ferromagnetic, another testimony to the inherently
strong NN-SQ interaction.
a
2
M
m
nb
t2 c/meV
J
AFd/cm-1
Je/cm-1
Mn
Ni
Cu
2
2
2
5
2
1
9700
16200
1300
-41.3
-68.9
-0.7
-41.3
-87.8
+75.6
The metal coordination geometries of the (NN-SQ)MT-
pCum,Me complexes (M ) Mn, Co, Ni) are axially elongated
trigonal bipyramidal, with the NN-SQ ligand occupying posi-
tions on a vertical plane. On the other hand, [NN-SQ(OMe)]-
CuTpCum,Me is best described as a distorted square pyramid with
the NN-SQ ligand occupying positions on the tetragonal plane.35
The metal ligand bond lengths (rM-O and rM-N ≈ 2.1 Å) are
consistent with high-spin electronic configurations for M ) Mn,
Co, and Ni.35,47,51-53 Because X-ray quality crystals of (NN-
SQ)NiL could not be produced, we hypothesize that the metal
geometry is nominally octahedral, and therefore that the NiII is
high-spin.
Ligand Design Elements. Heterospin organic molecules are
those with multiple, different spin-containing moieties.18,22,54-58
To prepare robust, high-spin structures an active atom (one with
positive spin density) of one radical group must be attached to
an inactive atom (one with negative spin density) of another
radical group.59-62 This is the case for our NN-SQ ligand:
nitronyl-nitroxides are known to have large negative spin density
at the middle carbon of the ONCNO chain,63 and SQ has
positive spin density at the C5 position.64
a Number of ligand (NN-SQ) spins. b Number of metal ion spins.
c Average of the hopping integral (resonance integral) squares. d Antifer-
romagnetic contribution to J from eq 10. e Experimental J-values.
straightforward to define the magnetic orbitals describing only
the spins of the MII ion site. If we require that the coordination
sphere of the MII ion be complete in the spin monomer of the
MII ion site, then the spin monomer must include orbitals on
the NN-SQ ligand. However, the NN-SQ ligand also possesses
unpaired spin (SNN-SQ ) 1). To avoid this conceptual difficulty,
we will define the monomer fragment of the MII site as (MII-
TpCum,Me)+ and the magnetic orbitals φν ) φM of this fragment
as those orbitals that are the primary contributors to the singly
filled d-block levels of (NN-SQ)MTpCum,Me. Therefore, there
are five magnetic orbitals for the (MnIITpCum,Me)+ fragment, two
for (NiIITpCum,Me)+, and one for (CuIITpCum,Me)+. Within the
extended Hu¨ckel formalism,46 the tµν values between the orbitals
φSQ and φM are easily evaluated by performing fragment
molecular orbital analysis. Therefore the primary contribution
to JAF arises solely from interactions involving the φSQ orbital.
The results are presented in Table 3.
From a molecular orbital perspective, the SQ SOMO can mix
with both the NN SOMO-1 and the NN LUMO to yield
nondisjoint biradical SOMOs (φNN-SQ and φNN in Figure 5). Such
biradicals have nonzero exchange integrals and therefore possess
triplet ground states.59 The magnetic properties of (NN-SQ)-
ZnTpCum,Me are indeed consistent with a robust triplet ground-
state configuration for NN-SQ.18
Features of the cyclic voltammograms and electronic absorp-
tion spectra (see Supporting Information) corroborate our
suggested interaction between SQ and NN moieties, and a full
spectroscopic study will be presented in detail in a separate
manuscript.65
Discussion
Molecular Structures. The bond lengths for the SQ rings
of the NN-SQ complexes are consistent with the SQ oxidation
state (i.e., MIISQ and not MIIICat, see Supporting Information).47
However, there is a periodic variation in SQ-ring bond lengths
that reflects interaction with the NN group. Previously, we
defined this bond length deviation parameter (Σ|∆i|) for new
quinone-methide-semiquinone (Σ|∆i| ) 0.279 Å) and bis-
(semiquinone) (Σ|∆i| ) 0.116 Å) ligands.48 The Σ|∆i| value
using average bond lengths for NN-SQ is 0.148 Å49 and is
consistent with an interaction between NN and SQ moieties that
causes a moderate redistribution of electron density and therefore
slightly alters bond lengths.
The torsion angles between the ONCNO chains and SQ ring
planes are noteworthy. This torsion will modulate π-interaction
between SQ and NN moieties and therefore controls intraligand
exchange coupling.50 These torsion angles vary from 0.5° to
36.1° (see Supporting Information). Since π-overlap varies as
cos(φ), and the intrinsic intraligand coupling is so strong, we
feel that the effect of the observed torsions on intraligand
exchange is negligible. The fact that we can successfully employ
a one-J Heisenberg Hamiltonian for all complexes lends validity
to our assertion.
Intra-Ligand Exchange Couplings. Variable-temperature
magnetic susceptibility measurements have allowed for the
18
determination of only a lower limit for JNN-SQ
.
As a result,
we wished to evaluate exchange coupling attenuation by addition
of a phenyl ring between the SQ and NN groups. Since the
(51) Pierpont, C. G.; Buchanan, R. M. Coord. Chem. ReV. 1981, 38, 45.
(52) Attia, A. S.; Pierpont, C. G. Inorg. Chem. 1997, 36, 6184.
(53) Benelli, C.; Dei, A.; Gatteschi, D.; Pardi, L. Inorg. Chem. 1988, 27, 2831.
Note that these authors used H ) (JS1‚S2. For comparison of J-values,
we have recast their experimentally determined exchange parameters using
a H ) -2JS1‚S2 exchange Hamiltonian.
(54) Kumai, R.; Matsushita, M. M.; Izuoka, A.; Sugawara, T. J. Am. Chem.
Soc. 1994, 116, 4523.
(55) Inoue, K.; Iwamura, H. Angew. Chem., Int. Ed. Engl. 1995, 34, 927.
(56) Tanaka, M.; Matsuda, K.; Itoh, T.; Iwamura, H. J. Am. Chem. Soc. 1998,
120, 7168.
(57) Liao, Y.; Xie, C.; Lahti, P. M.; Weber, R. T.; Jiang, J.; Barr, D. P. J. Org.
Chem. 1999, 64, 5176.
(58) Ionita, P.; Whitwood, A. C.; Gilbert, B. C. J. Chem. Soc., Perkin Trans. 2
2001, 2, 1453.
(59) Borden, W. T.; Davidson, E. R. J. Am. Chem. Soc. 1977, 99, 4587.
(60) Lahti, P. M.; Rossi, A. R.; Berson, J. A. J. Am. Chem. Soc. 1985, 107,
2273.
(61) Borden, W. T. Diradicals; Wiley: New York, 1982.
(62) Iwamura, H. Pure Appl. Chem. 1993, 65, 57.
The crystal structure of [NN-SQ(OMe)]CuTpCum,Me is
unique in that it shows a methoxy group attached to the SQ
ring. The reaction of methanol with the dioxolene ligand could
have occurred during reaction, or during recrystalization. The
(46) Hoffmann, R. J. Chem. Phys. 1963, 39, 1397.
(47) Pierpont, C. G.; Lange, C. W. Prog. Inorg. Chem. 1994, 41, 331.
(48) Shultz, D. A.; Bodnar, S. H.; Kampf, J. W. Chem. Commun. 2001, 93.
(49) For (NN-PhSQ) ZnTpCum,Me, Σ|∆i| ) 0.164 Å, but this may reflect the
lower quality of structural data.
(63) Pontillon, Y.; Akita, T.; Grand, A.; Kobayashi, K.; Lelievre-Berna, E.;
Pecaut, J.; Ressouche, E.; Schweizer, J. J. Am. Chem. Soc. 1999, 121, 10126.
(64) Wheeler, D. E.; Rodriquez, J. H.; McCusker, J. K. J. Phys. Chem. A 1999,
103, 4101.
(50) Shultz, D. A. Conformational Exchange Modulation in Trimethylen-
emethane (TMM)-Type Biradicals. In Magnetic Properties of Organic
Materials; Lahti, P., Ed.; Marcel Dekker: New York, 1999.
(65) Depperman, E. C.; Kirk, M. L.; Shultz, D. A. Manuscript in preparation.
9
J. AM. CHEM. SOC. VOL. 125, NO. 6, 2003 1613

A R T I C L E S
Shultz et al.
bond torsions) and therefore attenuation of the exchange integral.
Nevertheless, the +100 cm-1 J-value for NN-PhSQ is still quite
large.
Metal-Ligand Exchange Couplings. As mentioned above,
the metal ion coordination geometries for the (NN-SQ)-
MTpCum,Me complexes (M ) Mn, Co, Ni) are distorted trigonal
bipyramidal, except for the copper complex. For the trigonal
bipyramidal complexes, assigning the long O-M-N axis as
the molecular z-axis allows for a tentative description of the
metal d-orbitals in pseudo-C3V symmetry. Figure 6 depicts how
the metal orbitals interact with the NN-SQ SOMO. The five
pathway-dependent metal-ligand exchange interactions are
anticipated to be either ferromagnetic for those metal orbitals
that are orthogonal to the NN-SQ SOMO or antiferromagnetic
for those metal orbitals that are symmetric with the NN-SQ
SOMO.41,73 Thus, the individual Jµν expected to make ferro-
2
2
2
magnetic contributions to JNN-SQ,M are JNN-SQ,d , JNN-SQ,d
,
-
y
z
x
and JNN-SQ,d , while JNN-SQ,d , and JNN-SQ,d are expected to
xy
xz
yz
make antiferromagnetic contributions to JNN-SQ,M. For [NN-SQ-
(OMe)]CuTpCum,Me, the metal coordination geometry is best
described as a distorted square pyramid, and the magnetic orbital
2
2
is expected to be dx -y which is orthogonal to the tetragonal
plane and therefore orthogonal to the ligand magnetic orbitals.
However, due to the low symmetry of the complexes, this
approach may be oversimplified, and a more detailed description
of the orbital nature of the metal-ligand exchange interaction
has been provided using extended Hu¨ckel theory, vide supra.
Complex (NN-SQ)MnTpCum,Me is the first structurally char-
acterized complex with one SQ-type ligand coordinated to hs-
Figure 5. Qualitative MO diagram showing NN and SQ orbital mixing.
exchange interaction is anticipated to scale with the spin
densities of the radical moieties,37,61,66,67 we can estimate JNN-
with the relation:
SQ
MnII,25 and the net antiferromagnetic exchange (J ) -41 cm-1
suggests that the antiferromagnetic Jµν outweigh the ferromag-
netic contributions to JNN-SQ,M
)
FPhSQ
FSQ
aPhSQ
aSQ
JNN-PhSQ ) JNN-SQ
) JNN-SQ
(11)
(
)
(
)
.
The susceptibility data for the complex (NN-SQ)NiTpCum,Me
shows net antiferromagnetic metal-ligand exchange coupling
and is consistent with interstate mixing described by an
antisymmetric exchange term. Documented examples of anti-
symmetric exchange in binuclear centers are quite rare. How-
ever, anomalously low g-values recently observed in copper and
iron-sulfur clusters have been attributed to the presence of
antisymmetric exchange interactions.74,75 The magnitude of the
antisymmetric exchange parameter, |dab|, may be estimated by
considering the following expression (∆g is the deviation of
the average g of the two centers from the free electron ge )
2.0023):39,76
where FPhSQ is the spin density at the phenyl-C4′ position, and
FSQ is the C5 SQ spin density. Since these spin densities are
directly proportional to the associated electron-nuclear hyper-
fine coupling constants,68,69 the spin densities can be replaced
with the corresponding coupling constants (ai), which have been
determined experimentally. Substituting the experimentally
determined JNN-PhSQ ) +100 cm-1, and the ai values70 into eq
11 gives JNN-SQ ≈ +550 cm-1. Thus, the exchange parameter
has been reduced by more than a factor of 5 compared to NN-
SQ, consistent with diminished exchange caused by addition
of the phenylene spacer. Modulation of the exchange interaction
in biradicals by introducing spacers has been observed in
Chichibabin hydrocarbon derivatives,71 and in nitroxide triradi-
cals.11,72 The primary contribution to the reduction of the
pairwise exchange interaction is a decrease in the SOMO-
SOMO overlap density (caused by both additional atoms and
|∆g|
|dab| ≈
JNN-SQ,Ni
(12)
g
Assuming ∆g ≈ 0.1, the antisymmetric exchange parameter
could be as high as |dab| ≈ 4 cm-1. Since the antiferromagnetic
exchange coupling between S ) 1 NiII and S ) 1 NN-SQ gives
rise to a formally diamagnetic ground state, the small positive
deviations from øave ) 0 resulting from antisymmetric exchange
are more readily observed than if the ground state were
paramagnetic with a large increasing øave as T f 0. Therefore,
the observed mixing of a singlet ground state with a low-lying
(66) Rajca, A. Chem. ReV. 1994, 94, 871.
(67) Rajca, A. High-Spin Polyradicals. In Magnetic Properties of Organic
Materials; Lahti, P., Ed.; Marcel Dekker: New York, 1999; p 345.
(68) McConnell, H. M.; Chestnut, D. B. J. Chem. Phys. 1958, 28, 107.
(69) Wertz, J. E.; Bolton, J. R. Electron Spin Resonance; Chapman and Hall:
New York, 1986.
(70) For the hyperfine coupling constant of SQ, we used the value for H4
reported in our previous work (3.33 G), and assumed negligible difference
between the spin densities of C4 and C5: Shultz, D. A.; Boal, A. K.;
Campbell, N. P. Inorg. Chem. 1998, 37, 1540-3; see also: Wheeler, D.
E.; Rodriquez, J. H.; McCusker, J. K. J. Phys. Chem. A 1999, 103, 4101.
For the H4′ hyperfine coupling constant for 3-tert-butyl-5-phenylsemiquione
(0.6 G), we prepared the radical anion and simulated its fluid solution EPR
spectrum, see Supporting Information.
(73) Kahn, O. Angew. Chem., Int. Ed. Engl. 1985, 24, 834.
(74) Clark, P. A.; Solomon, E. I. J. Am. Chem. Soc. 1992, 114, 1108.
(75) Sanakis, Y.; Macedo, A. L.; Moura, I.; Moura, J. J. G.; Papaefthymiou,
V.; Mu¨nck, E. J. Am. Chem. Soc. 2000, 122, 11855.
(76) Gatteschi, D.; Sessoli, R.; Plass, W.; Miller, A.; Krickemeyer, E.; Meyer,
J.; Solter, D.; Adler, P. Inorg. Chem. 1996, 35, 1926.
(71) Waring, R. K.; Sloan, G. J. J. Chem. Phys. 1964, 40, 772.
(72) Kanno, F.; Inoue, K.; Koga, N.; Iwamura, H. J. Phys. Chem. 1993, 97,
13267.
9
1614 J. AM. CHEM. SOC. VOL. 125, NO. 6, 2003

Exchange Interaction Studies of (NN-SQ)M Complexes
A R T I C L E S
triplet excited state is more pronounced than the corresponding
mixing would be had the ground-state been paramagnetic.
Moreover, (NN-SQ)NiTpCum,Me has the lowest-lying excited
state (Figure 3) of all the complexes and will therefore mix more
strongly with the ground state. Antisymmetric exchange con-
tributions undoubtedly occur in the manganese, cobalt, and
copper complexes but are difficult to detect owing to their
paramagnetic ground states and/or the fact that ∆g , 0.1.
The øT data for six-coordinate (NN-SQ)NiL are consistent
with orthogonal metal and ligand orbitals, and therefore also
II
53
2
2
2
with six-coordinate Ni having dz and dx -y magnetic orbitals.
The ferromagnetic metal-ligand exchange coupling in [NN-
SQ(OMe)]CuTpCum,Me was described previously18,19 and is
consistent with strong coupling between orthogonal NN-SQ and
metal spins.
The Nature of the Metal-Ligand Exchange Interaction.
The results of our extended Hu¨ckel calculations, summarized
in Table 3, are in accord with both simple symmetry arguments
and with experimental findings. The experimental J-value of
(NN-SQ)MnTpCum,Me is reproduced in terms of JAF according
to eq 10 (thereby ignoring the JF terms), if Ueff is chosen as
3.76 eV. We employ this Ueff value to estimate the relatiVe JAF
values of (NN-SQ)NiTpCum,Me and [NN-SQ(OMe)]CuT-
pCum,Me. The calculated JAF values quantify the predicted
Figure 6. Qualitative trigonal bipyramidal MO diagram showing NN-SQ-
and d-orbital mixing.
77
antiferromagnetic exchange pathway contributions to JNN-SQ,M
.
In fact, the antiferromagnetic exchange contributions to JNN-
effort, described the magnetic properties of six-coordinate NiII
and five-coordinate MnII complexes of a tris(SQ) ligand.25 The
metal-ligand exchange parameters here were found to be: JSQ,Ni
SQ in the (NN-SQ)MTpCum,Me complexes are calculated to
,M
decrease in magnitude in the order, (NN-SQ)NiTpCum,Me
>
(NN-SQ)MnTpCum,Me > [NN-SQ(OMe)]CuTpCum,Me, and for
the latter the antiferromagnetic exchange interaction is very
weak. Qualitatively, JAF is anticipated to be negligible in [NN-
) +87 cm-1 and JSQ,Mn ) -175 cm-1
.
We can make direct comparisons between the M-SQ
exchange interactions in M-SQ dyads and the (NN-SQ)M triads
presented here by expressing the measured JNN-SQ,M as JSQ,M
via diagonalization of the two-J Heisenberg Hamiltonian at fixed
SQ(OMe)]CuTpCum,Me because the magnetic orbital, φNN-SQ
,
of NN-SQ is nearly orthogonal to the d-block magnetic orbital
of [NN-SQ(OMe)]CuTpCum,Me. The calculated JAF of (NN-
JNN-SQ.
79 This is important, as structural and electronic param-
SQ)MnTpCum,Me is less than half that of (NN-SQ)NiTpCum,Me
.
eters resulting from variations in the ancillary ligands and the
extent of the direct exchange contribution, which scales ac-
cording to the different M-L bond lengths,80 may now be
directly evaluated. If one uses limiting values of JNN-SQ ) +310
cm-1 (the lower limit determined in our previous study)18 and
JNN-SQ ) +550 cm-1 (estimated from eq 11), one can show
that JSQ,Cu ) +187 cm-1 and +166 cm-1 respectively for [NN-
According to eqs 9 and 10, this is understandable because there
are more unpaired spins in the MnII ion site of (NN-SQ)-
MnTpCum,Me than on the NiII ion site of (NN-SQ)NiTpCum,Me
.
Thus, despite the low molecular symmetries, the extended
Hu¨ckel results support predictions based on higher-symmetry
orbital overlap considerations as presented in Figure 6.
Comparison of Metal-Ligand Couplings with Those of
Other SQ Complexes. The experimental metal-ligand ex-
change parameters reported in Table 2 may be compared to those
of other M-SQ complexes78 to obtain greater insight into the
exchange interactions in (NN-SQ)M dyads. For example, Dei
and Gatteschi reported five-coordinate CuII and NiII complexes
of 3,5-di-tert-butylsemiquinone and found JSQ,Cu ) +52 cm-1
for the copper complex and JSQ,M < -300 cm-1 for the nickel
SQ(OMe)]CuTpCum,Me 79
Thus, the ferromagnetic metal-SQ
.
exchange interaction in the NN-SQ-Cu triad is more than tripled
in magnitude compared to the previously mentioned five-
coordinate copper complex38 but is very close to the exchange
interaction in (3,5-DBSQ)CuTpCum,Me 35
.
The spin state energies as a function of JSQ,Cu at constant
JNN-SQ () +310 cm-1 and +550 cm-1) derived from this
analysis are displayed graphically in Figure 7. Note that for
JSQ,Cu > +300 cm-1, the energy gap separating the ground
quartet and the lowest doublet is independent of JSQ,Cu. This is
an extremely important finding if one wishes to enhance the
stability of high-spin states of multispin metal-radical complexes.
38
complex. The same researchers also reported the magnetic
properties of a six-coordinate NiII complex of 3,5-di-tert-
butylsemiquinone for which the exchange is ferromagnetic with
JSQ,Ni > +200 cm-1 53
Recently, one of us, in a collaborative
.
(77) These pathways are described by the Goodenough Kanamori rules, see the
following. Ginsberg, A. P. Inorg. Chim. Acta ReV. 1971, 5, 45. Hatfield,
W. E. In Theory and Applications of Molecular Paramagnetism; Boudreaux,
E. A., Mulay, L. N., Eds.; Chapter 7, Properties of Condensed Compounds
(Compounds with Spin Exchange); Wiley-Interscience: New York, 1976;
pp 381-385.
(78) Pierpont, C. G.; Attia, A. S. Collect. Czech. Chem. Commun. 2001, 66,
33. Note that these authors used H ) (JS1‚S2. For comparison of J-values,
we have recast their experimentally determined exchange parameters using
a H ) -2JS1‚S2 exchange Hamiltonian.
(79) For the copper complex, an 8 × 8 spin Hamiltonian exchange matrix, [H
) -2JCu,SQSCu‚SSQ - 2JNN-SQSSQ‚SNN - 2JCu,NNSCu‚SNN] was derived using
the uncoupled basis set |SCu, MS,Cu, SSQ, MS,SQ, SNN, MS,NN . The matrix
was diagonalized holding constant JCu,NN ) 0 and JNN-SQ ) +310 (+550)
cm-1, and varying JCu,SQ until a value was obtained which gave eigenvalues
matching the quartet-doublet splitting of 227 cm-1. Analogous methods
were used for the nickel complexes.
(80) Rodriguez, J. H.; Wheeler, D. E.; McCusker, J. K. J. Am. Chem. Soc. 1998,
120, 12051.
9
J. AM. CHEM. SOC. VOL. 125, NO. 6, 2003 1615

A R T I C L E S
Shultz et al.
Figure 7. Spin state energies for [NN-SQ(OMe)]CuTpCum,Me as a function
of JSQ,Cu. JNN-SQ ) 310 cm-1 (- - -); JNN-SQ ) 550 cm-1 (s). The lowest-
energy state is the quartet, with the two doublets lying higher in energy.
The arrows in the lower left denote the limits of JSQ,Cu
.
Although strong metal-ligand coupling is obviously a necessary
criterion, the intraligand coupling is of paramount importance
for stabilizing the ground state relatiVe to the first excited spin-
state.
A similar analysis for (NN-SQ)NiTpCum,Me yields JSQ,Ni
)
-244 cm-1 and -525 cm-1 for JNN-SQ ) +310 cm-1 and JNN-
SQ ) +550 cm-1, respectively. For the six-coordinate coordinate
complex, (NN-SQ)NiL, JSQ,Ni ) +193 cm-1 and +320 cm-1
for JNN-SQ ) +550 cm-1 and JNN-SQ ) +310 cm-1, respectively.
These results are displayed graphically in Figure 8. Here it is
clearly apparent that the electronic structure of the metal ion,
which is a function of the coordination environment, can result
in enormous changes in the sign and magnitude of the exchange
interaction. This is a result of different Jµν pathway contributions
to the measured exchange, JSQ,Ni, and is in direct accord with
the Goodenough-Kanimoori rules77 and extended Hu¨ckel
analysis presented above. Nevertheless, JSQ,Ni values for our
NN-SQ complexes are very close to reported values for
complexes with similar coordination geometries, and we
conclude that the NN group does not dramatically affect M-SQ
exchange coupling.
Figure 8. (A) Spin state energies for (NN-SQ)NiTpCum,Me as a function
of JSQ,Ni. JNN-SQ ) +310 cm-1 (- - -); JNN-SQ ) +550 cm-1 (s). The lowest-
energy state is the singlet, with excited triplet, quintet, and triplet states,
respectively. (B) Spin state energies for (NN-SQ)NiL as a function of JSQ,Ni
‚
JNN-SQ ) +310 cm-1 (- - -); JNN-SQ ) +550 cm-1 (s). The lowest-energy
state is the quintet, with excited triplet, singlet, and triplet states, respectively.
The arrows in the lower left denote the limits of JSQ,Ni
.
Conclusions
described with the addition of an antisymmetric exchange term.
The presence of a singlet ground-state, coupled with a moderate
exchange interaction conspire to allow interstate mixing with
the low-lying triplet state. The general manifestations of this
interstate mixing are a reduced g-value, and nonzero paramag-
netic susceptibility in the formally singlet ground spin-state.
However, this model is still insufficient to explanation the data
for the corresponding cobalt complex. This is most likely an
indication of substantial ground-state orbital moment for the d7
cobalt complex. Antisymmetric exchange is less important for
(NN-SQ)MnTpCum,Me, [NN-SQ(OMe)]CuTpCum,Me, and (NN-
SQ)NiL due to the smaller ∆g values and paramagnetic ground
states.
We presented a comprehensive structural and magnetic study
of isostructural first-row, high-spin transition metal complexes
of the high-spin S ) 1 NN-SQ ligand. The intraligand exchange
interactions are strongly ferromagnetic in nature, and generally
greater in magnitude than the metal-ligand exchange. The
metal-ligand exchange interactions are in accord with magnetic
orbital symmetry arguments within the active electron ap-
proximation. Despite the low symmetry (C1) adopted by these
complexes, extended Hu¨ckel calculations support high-symmetry
arguments relating to the sign of the exchange parameter and
have led to highly informative magnetostructural correlations
in these hybrid metal-organic spin triads.
The low-temperature magnetic properties of the five-
coordinate nickel complex, (NN-SQ)NiTpCum,Me, have been
The study also highlights the extreme importance of intrali-
gand coupling for stabilization of the high-spin ground state of
9
1616 J. AM. CHEM. SOC. VOL. 125, NO. 6, 2003

Exchange Interaction Studies of (NN-SQ)M Complexes
A R T I C L E S
multispin metal-radical complexes. This important finding has
a profound impact on the design of high-spin metal-organic
hybrid materials.
M.L.K. acknowledges the National Institutes of Health GM-
057378 for financial support.
Supporting Information Available: Synthetic details, crystal-
lographic data, saturation magnetization plots, electronic spectra,
cyclic voltammograms, and synthesis and EPR spectra for 3-tert-
butyl-5-phenylsemiquinone (PDF/CIF). This material is available
free of charge via the Internet at http://pubs.acs.org.
Acknowledgment. D.A.S. acknowledges the National Science
Foundation (CHE-9910076; CHE-0091247 SQUID Magnetom-
eter) for financial support and thanks the Camille and Henry
Dreyfus Foundation for a Camille Dreyfus Teacher-Scholar
Award. M.-H.W. acknowledges the financial support from the
Office of Basic Energy Sciences, Division of Materials Sciences,
U.S. Department of Energy, under Grant DE-FG02-86ER45259.
JA020715X
9
J. AM. CHEM. SOC. VOL. 125, NO. 6, 2003 1617