Synthesis, crystal structure, and high-precision high-frequency and -field electron paramagnetic resonance investigation of a manganese(III) complex: [Mn(dbm)2(py)2](ClO4)

Article abstract of DOI:10.1021/ic049180u

The complex [Mn(dbm)₂(py)₂](ClO₄) (dbm = anion of 1,3-diphenyl-1,3-propanedione (dibenzoylmethane), py = pyridine) was synthesized and characterized by X-ray crystallography. It has tetragonally distorted geometry with the axial positions occupied by the py ligands and the equatorial positions by the dbm ligands. This mononuclear complex of high-spin Mn(III) (3d⁴, S = 2) was studied by high-frequency and -field electron paramagnetic resonance (HFEPR) both as a solid powder and in frozen dichloromethane solution. Very high quality HFEPR spectra were recorded over a wide range of frequencies. The complete dataset of resonant magnetic fields versus transition energies was analyzed using automated fitting software. This analysis yielded the following spin Hamiltonian parameters (energies in cm ^-1): D = -4.504(2), E = -0.425(1), B₄⁰ = -1.8(4) × 10^-4, B₄² = 7(3) × 10^-4, B₄⁴ = 48(4) × 10^-4, g_x = 1.993(1), g_y = 1.994(1), and g_z = 1.983(1), where the B₄ⁿ values represent fourth-order zero-field splitting terms that are generally very difficult to extract, even from single-crystal measurements. The results here demonstrate the applicability of HFEPR at high-precision measurements, even for powder samples. The zero-field splitting parameters determined here for [Mn(dbm)₂(py) ₂]⁺ are placed into the context of those determined for other mononuclear complexes of Mn(III).

Full text of DOI:10.1021/ic049180u

Inorg. Chem. 2005, 44, 187−196
Synthesis, Crystal Structure, and High-Precision High-Frequency and
-Field Electron Paramagnetic Resonance Investigation of a
Manganese(III) Complex: [Mn(dbm)2(py)2](ClO4)
Guillem Arom´ı,† Joshua Telser,‡ Andrew Ozarowski,§ Louis-Claude Brunel,§
Helen-Margaret Stoeckli-Evans,| and J. Krzystek*,§
Departament de Qu´ımica Inorga`nica, Facultat de Qu´ımica, UniVersitat de Barcelona,
E-08028 Barcelona, Spain, Chemistry Program, RooseVelt UniVersity, Chicago, Illinois 60605,
National High Magnetic Field Laboratory, Florida State UniVersity, Tallahassee, Florida 32310,
and Laboratoire de Cristallographie, Institut de Chimie, UniVersite´ de Neuchaˆtel,
CH-2007 Neuchaˆtel, Switzerland
Received June 23, 2004
The complex [Mn(dbm)2(py)2](ClO4) (dbm
) anion of 1,3-diphenyl-1,3-propanedione (dibenzoylmethane), py )
pyridine) was synthesized and characterized by X-ray crystallography. It has tetragonally distorted geometry with
the axial positions occupied by the py ligands and the equatorial positions by the dbm ligands. This mononuclear
complex of high-spin Mn(III) (3d4, S
) 2) was studied by high-frequency and -field electron paramagnetic resonance
(HFEPR) both as a solid powder and in frozen dichloromethane solution. Very high quality HFEPR spectra were
recorded over a wide range of frequencies. The complete dataset of resonant magnetic fields versus transition
energies was analyzed using automated fitting software. This analysis yielded the following spin Hamiltonian
2
4
parameters (energies in cm-1): D ) −4.504(2), E ) −0.425(1), B40 ) −1.8(4)
48(4) 1.993(1), gy 1.994(1), and gz
10-4, gx
×
10-4, B4
)
7(3)
×
10-4, B4
)
n
×
)
)
)
1.983(1), where the B4 values represent fourth-order
zero-field splitting terms that are generally very difficult to extract, even from single-crystal measurements. The
results here demonstrate the applicability of HFEPR at high-precision measurements, even for powder samples.
The zero-field splitting parameters determined here for [Mn(dbm)2(py)2]+ are placed into the context of those
determined for other mononuclear complexes of Mn(III).
Introduction
representing different coordination numbers and geometries.
The original HFEPR detection involved, in one case, five-
coordinated porphyrinic complexes, Mn(TPP)Cl and its
porphyrazine analogues,2 representing a square-pyramidal
coordination sphere of the type N4Cl, and, in the other, a
six-coordinate complex, Mn(dbm)3 (where Hdbm is diben-
zoylmethane, formally, 1,3-diphenyl-1,3-propanedione), rep-
resenting a distorted octahedral geometry of exclusively
oxygen donors (O6).3 Mn(III) with O6 coordination had also
been studied early on by EPR4 and quite recently by INS,5
as well as by HFEPR.6 Related, five-coordinate complexes
with O4Cl and O4Br donors have been studied by HFEPR
Ever since the first successful application of high-
frequency and -field electron paramagnetic resonance
(HFEPR)1 in 1997 to detect paramagnetic resonances in high-
spin Mn(III) (3d4, S ) 2),2,3 this formerly “EPR-silent” ion
has been studied in a variety of molecular complexes
* Author to whom correspondence should be addressed. E-mail:
krzystek@magnet.fsu.edu. Phone: +1 850 644 6077. Fax: +1 850 644
9462.
† Universitat de Barcelona.
‡ Roosevelt University.
§ Florida State University.
| Universite´ de Neuchaˆtel.
(1) Abbreviations used: acac, anion of 2,4-pentanedione (acetylacetone);
dbm, anion of 1,3-diphenyl-1,3-propanedione (dibenzoylmethane);
EPR, electron paramagnetic resonance; HFEPR, high-frequency and
-field EPR; INS, inelastic neutron scattering; NIR, near infrared; OEP,
dianion of octaethylporphine; py, pyridine; rms, root-mean-square
error; TPP, dianion of meso-tetraphenylporphine; zfs, zero-field
splitting.
(2) Goldberg, D. P.; Telser, J.; Krzystek, J.; Montalban, A. G.; Brunel,
L.-C.; Barrett, A. G. M.; Hoffman, B. M. J. Am. Chem. Soc. 1997,
119, 8722-8723.
(3) Barra, A.-L.; Gatteschi, D.; Sessoli, R.; Abbati, G. L.; Cornia, A.;
Fabretti, A. C.; Uytterhoeven, M. G. Angew. Chem., Intl. Ed. Engl.
1997, 36, 2329-2331.
(4) Gerritsen, H. J.; Sabisky, E. S. Phys. ReV. 1963, 132, 1507-1512.
10.1021/ic049180u CCC: $30.25
Published on Web 12/16/2004
© 2005 American Chemical Society
Inorganic Chemistry, Vol. 44, No. 2, 2005 187

Arom´ı et al.
as well.7 Further work has been reported on the properties
of complexes with nitrogen donor ligands. These include
porphyrinic complexes of Mn(III) with varying axial liga-
tion: N4;8 N4O;9 N4X (X ) Cl,10 Br7). Six-coordinate
complexes with all nitrogen donors (N6)11 and various mixed
nitrogen and halide donors (N3F3)12 and N4X2 (X ) Br, I)13,14
have also been studied. Last, a five-coordinate complex with
the mixed donor set N2O2Cl has also been investigated by
low-frequency EPR15 and HFEPR.16 Given the number of
complexes studied, qualitative and quantitative conclusions
can already be drawn about their magnetic properties as well
as electronic vs geometric structure (see Discussion). The
principal aim of this study is thus to characterize structurally
and spectroscopically, chiefly by HFEPR, a mononuclear
Mn(III) complex with an N2O4 coordination sphere, which
is a type hitherto not studied, to our best knowledge. This
was realized by synthesizing and investigating the complex
[Mn(dbm)2(py)2](ClO4), subsequently termed 1.
the first time that fourth-order zfs terms have been accurately
determined for a mononuclear high-spin Mn(III) complex
from powder spectra.
Experimental Section
Synthesis. All manipulations were performed in air using
reagents as received. Bun NMnO4 was prepared18 as reported in
4
the literature.19 To a stirred light yellow solution of Mn(ClO4)2‚
6H2O (450 mg, 1.24 mmol) and pyridine (3 mL, 37.09 mmol) in
MeCN (15 mL) was added dropwise a purple solution of freshly
prepared Bun NMnO4 (111 mg, 0.31 mmol) in MeCN (2 mL). Solid
4
Hdbm (700 mg, 3.13 mmol) was added immediately to the resulting
dark brown solution, and the mixture was stirred for a few hours,
after which a brown precipitate of [Mn(dbm)2(py)2](ClO4) (1) had
formed. The solid was collected by filtration, washed with Et2O,
and dried under vacuum. The yield was ∼50%.
IR (neat): ν (cm-1) ) 1599.3 m, 1585.6 w, 1511.4 vs, 1484.8
vs, 1445.9 m, 1433.9 m 1341.1 s, 1321.7 s, 1230.4 m, 1185.9 w,
1159.6 m, 1094.2 vs, 1067.3 m, 1022.7 m, 1007.9 w, 1000.0 w,
940.2 w, 762.9 w, 748.2 w, 726.4 w, 705.5 m, 685.9 w, 622.2 m,
589.6 w, 552.0 w, 422.1 w.
The coordination environment of Mn(III) in 1 is of
particular interest with respect not only to other six-
coordinate complexes of Mn(III) described above but also
because so many of the Mn(III) complexes thus far charac-
terized contain equatorial N4 donor ligands (i.e., tetrapyr-
roles8-10,17 and saturated tetraazamacrocycles13,14) with vari-
ous weaker field, axial ligands. In this study, 1 has an
equatorial O4 donor set and axial nitrogen donors. The two
dbm ligands do not comprise a macrocycle, but each is
π-conjugated and, taken together, the result is equatorial
ligation that could be described as an O4 “porphyrinoid”.
Anal. Calcd for 1 (Mr ) 759.10): C, 63.29; H, 4.25; N, 3.69;
Mn, 7.24. Found: C, 63.31; H, 4.10; N, 3.67; Mn, 7.27. Large
crystals suitable for X-ray crystallography could be obtained
overnight if the stirring was interrupted just a few minutes after
the addition of Hdbm and the system was left unperturbed.
X-ray Crystallography. Suitable crystals of 1 were obtained
as dark brown rods. The intensity data were collected at 153 K
(-120 °C) on a Stoe Mark II image plate diffraction system
equipped with a two-circle goniometer and using Mo KR graphite-
monochromated radiation. The image plate distance was 100 mm,
and ω rotation scans were conducted with angles of 0-180° at φ
0° and 0-64° at φ 90°, with a step ∆ω ) 1.0°. The 2θ range was
2.29-59.53°, and dmax-dmin was 17.779-0.716 Å. The structure
was solved by direct methods using the program SHELXS-97.20
The refinement and all further calculations were carried out using
SHELXL-97.21 The H atoms were located from Fourier difference
maps and refined isotropically. The non-H atoms were refined
anisotropically, using weighted full-matrix least squares on F2. The
molecular structure and crystallographic numbering scheme are
illustrated in Figure 1.
HFEPR. The experiments were performed using a single-pass
transmission-type spectrometer, in which the sub-THz waves are
propagated in cylindrical lightpipes, as described previously.22 Sub-
THz frequencies were generated by either of two Gunn oscillators,
operating at 95 ( 3 and 110 ( 3 GHz, respectively. In addition to
the fundamentals, the frequencies were also multiplied by factors
of 2-4 using Schottky diode-based multipliers. The power emitted
varied strongly with the harmonic generated this way. A 17 T
superconducting magnet from Oxford Instruments was used. The
magnetic field values were read from the power supply current
through proper calibration. Errors associated with this procedure
Concerning the spectroscopic characterization, of particular
attention was the possibility to increase the accuracy in
determining the spin Hamiltonian parameters describing the
S ) 2 spin state of Mn(III), in particular the fourth-order
zfs terms. This advance proved possible thanks to the very
high quality of the HFEPR spectra obtained from polycrys-
talline solid samples. To the best of our knowledge, this is
(5) Basler, R.; Tregenna-Piggott, P. L. W.; Andres, H.; Dobe, C.; Gu¨del,
H.-U.; Janssen, S.; McIntyre, G. J. J. Am. Chem. Soc. 2001, 123,
3377-3378.
(6) Krzystek, J.; Yeagle, G.; Park, J.-H.; Meisel, M. W.; Britt, R. D.;
Brunel, L.-C.; Telser, J. Inorg. Chem. 2003, 42, 4610-4618.
(7) Krzystek, J.; Telser, J.; Knapp, M. J.; Hendrickson, D. N.; Aromi,
G.; Christou, G.; Angerhofer, A.; Brunel, L.-C. Appl. Magn. Reson.
2001, 23, 571-585.
(8) Krzystek, J.; Telser, J.; Hoffman, B. M.; Brunel, L.-C.; Licoccia, S.
J. Am. Chem. Soc. 2001, 123, 7890-7897.
(9) Bendix, J.; Gray, H. B.; Golubkhov, G.; Gross, Z. J. Chem. Soc., Chem.
Commun. 2000, 1957-1958.
(10) Krzystek, J.; Telser, J.; Pardi, L. A.; Goldberg, D. P.; Hoffman, B.
M.; Brunel, L.-C. Inorg. Chem. 1999, 38, 6121-6129.
(11) Limburg, J.; Vrettos, J. S.; Crabtree, R. H.; Brudvig, G. W.; de Paula,
J. C.; Hassan, A.; Barra, A.-L.; Duboc-Toia, C.; Collomb, M.-N. Inorg.
Chem. 2001, 40, 1698-1703.
(12) Mantel, C.; Hassan, A. K.; Pe´caut, J.; Deronzier, A.; Collomb, M.-
N.; Duboc-Toia, C. J. Am. Chem. Soc. 2003, 125, 12337-12344.
(13) Mossin, S.; Weihe, H.; Barra, A.-L. J. Am. Chem. Soc. 2002, 124,
8764-8765.
(14) Mossin, S.; Stefan, M.; ter Heerdt, P.; Bouwen, A.; Goovaerts, E.;
Weihe, H. Appl. Magn. Reson. 2002, 21, 586.
(15) Campbell, K. A.; Lashley, M. R.; Wyatt, J. K.; Nantz, M. H.; Britt,
R. D. J. Am. Chem. Soc. 2001, 123, 5710-5719.
(16) Krzystek, J.; Telser, J. J. Magn. Reson. 2003, 162, 454-465.
(17) Krzystek, J.; Pardi, L. A.; Brunel, L.-C.; Goldberg, D. P.; Hoffman,
B. M.; Licoccia, S.; Telser, J. Spectrochim. Acta, Part A 2002, 58,
1113-1127.
(18) Warning! There haVe been reports of a detonation of quaternary
ammonium permanganates during drying at eleVated temperature. We
recommend appropriate care be taken in the use of organic perman-
ganates. The Bun N+ salt should be dried in Vacuo at room temper-
4
ature.
(19) Vincent, J. B.; Folting, K.; Huffman, J. C.; Christou, G. Inorg. Chem.
1996, 25, 996-999.
(20) Sheldrick, G. M. Acta Crystallogr. 1990, A46, 467-473.
(21) Sheldrick, G. SHELXL; University of Go¨ttingen: Go¨ttingen, Germany,
1999.
(22) Hassan, A. K.; Pardi, L. A.; Krzystek, J.; Sienkiewicz, A.; Goy, P.;
Rohrer, M.; Brunel, L.-C. J. Magn. Reson. 2000, 142, 300-312.
188 Inorganic Chemistry, Vol. 44, No. 2, 2005

Synthesis and Characterization of [Mn(dbm)2(py)2](ClO4)
Figure 2. Effect of particular zero-field splitting terms of the spin
Hamiltonian (eq 1) acting on quintet state wave functions. ∆ ) 3E2/D.
Figure 1. ORTEP plot of 1 at 50% probability level. Hydrogen atoms
have been omitted for clarity.
state wave functions are shown in Figure 2. A spin Hamiltonian
for S ) 2 can also include higher order field-dependent (Zeeman)
terms, such as B.S3,26 which can complicate determination of g
values; however, it was not necessary to consider such terms here.
are discussed elsewhere.22 Temperature control was achieved with
an Oxford Instruments CF1200 continuous-flow liquid-helium
cryostat and an IC503 controller. Field modulation and the ensuing
phase-sensitive detection using a Stanford SR830 lock-in amplifier
was applied. Experiments were performed on either polycrystalline
solid or a solution in CH2Cl2 (Aldrich-Sigma, analytical grade). In
the second case, the solvent was purged with nitrogen to reduce
the concentration of dissolved oxygen, which often produces strong
HFEPR signals at low temperatures.23 An attempt to apply toluene
as a glassing agent to improve the frozen solvent quality16 resulted
in precipitating the complex and was therefore abandoned. The CH2-
Cl2 solution was of sufficiently high quality, judging by the obtained
spectra. Typically, 20-30 mg of the sample was used in either the
solid-state or solution measurement. In addition, a conventional
X-band EPR experiment was performed on both solid and solution
sample.
Least-Squares Fitting Procedure. Single-frequency powder
spectra were simulated using a locally written program (available
from A.O.). The transition probabilities were calculated from the
eigenvectors in a standard way.27 The calculated transition intensities
were corrected for the Boltzmann populations of the levels involved
because of the relatively large values of the zfs parameters and the
high magnetic fields at which certain transitions occurred. After
finding by the trial-and-error method the initial parameters g, D,
and E that gave reasonable powder simulations, all transitions
observed in the experimental powder spectra were identified as
corresponding to the x, y, z or to the off-axial turning points. A
2-dimensional array whose rows contained frequency (in the range
of ∼90-440 GHz), polar angles Θ and Φ, and resonance field
was used as our input data. This array was simultaneously fitted
by use of a nonlinear least-squares (Simplex) method to minimize
the function
Other Physical Measurements. FTIR spectra were collected
on a Perkin-Elmer Spectrum RXI spectrometer. H NMR spectra
1
were recorded on a 250 MHz Bruker DRX 250 spectrometer.
Elemental analyses were performed with a Carlo Erba Instruments
CHNS-O EA-1108 elemental analyzer. UV/vis/NIR spectra were
recorded on a Jasco V-570 spectrophotometer. Compound 1 in CH2-
Cl2 solution exhibits a very strong charge-transfer band below 450
nm, a band at 480 nm (ꢀ ) 1550 M-1 cm-1), and shoulders at 505
and 520 nm. No absorption was observed above 700 nm.
Theory for EPR Spectra. To analyze the EPR spectra, we
applied the most general spin Hamiltonian for an S ) 2 spin state
including both the second-order and the fourth-order terms of the
zero-field splitting:24
N
ø2 )
(fi(calc) - fi(exp) 2
)
(2)
∑
i)1
where fi are the calculated and experimental resonance fields.
Because we assumed (and confirmed in separate experiments
described in ref 22) that all experimental resonance fields bear
approximately the same uncertainty, there was no need to include
uncertainties σ2 in the denominator. The resonance fields fi(calc) were
i
evaluated using the same Householder method28 to diagonalize the
spin Hamiltonian in eq 1 that the powder simulation had employed,
but no transition probabilities nor Boltzmann factors were calculated
in this case. The least-squares procedure was carried out in
conjunction with human judgment, which was used to eliminate
mathematically possible, but unphysical, results to obtain best-fit
parameters for the entire field vs energy 2-D array of EPR
transitions. After convergence had been achieved, the Hessian
0
2
0
2
4
4
H ) âB‚g‚S + B2 O20 + B2 O22 + B4 O40 + B4 O42 + B4 O4
(1a)
In the following discussion we will apply the commonly used
second-order zfs parameters D ≡ 3B20 and E ≡ B22 while retaining
the form of the fourth-order terms as shown above:
H ) âB‚g‚S + D(Sz2 - S(S + 1)/3) + E(Sx2 - Sy ) +
2
0
2
4
4
B4 O40 + B4 O42 + B4 O4 (1b)
(25) Sometimes, a different set of fourth-order zfs terms is used, comprising
0
4
parameters a and F. Parameter a is equal to 120B4 or 24B4 in a
cubic environment. Parameter F reflects deviation from the perfect
m
4
0
The operators O4 as well as their matrix elements may be found
in many texts including Abragam and Bleaney.24,25 The zero-field
energy levels resulting from this spin Hamiltonian acting on quintet
cubic symmetry and is equal to 36B4 - 180B4 .
(26) McGavin, D. G.; Tennant, W. C.; Weil, J. A. J. Magn. Reson. 1990,
87, 92-109.
(27) Bencini, A.; Gatteschi, D. In Transition Metal Chemistry; Melson, G.
A., Figgis, B. N., Eds.; Marcel Dekker: New York, 1982; Vol. 8, p
1.
(23) Pardi, L. A.; Krzystek, J.; Telser, J.; Brunel, L.-C. J. Magn. Reson.
2000, 146, 375-378.
(24) Abragam, A.; Bleaney, B. Electron Paramagnetic Resonance of
Transition Ions; Dover Publications: New York, 1986.
(28) Press: W. H.; Flannery, B. P.; Teukolsky, A. A.; Vetterling, W. T. In
Numerical Recipes in Pascal; Cambridge University Press: Cam-
bridge, U.K., 1989; p 572.
Inorganic Chemistry, Vol. 44, No. 2, 2005 189

Arom´ı et al.
matrix was calculated. The Hessian matrix29 is formally a matrix
containing the second derivatives of ø2 with respect to parameters
p,
4Mn(ClO4)2 + Bun NMnO4 + 10Hdbm + 10py f
4
5[Mn(dbm)2(py)2](ClO4) + Bun NClO4 + 2HClO4 + 4H2O
4
(7)
δ2ø2
δpiδpj
Hij )
(3)
Description of the Structure. Crystallographic data for
1 are collected in Table 1. Selected interatomic distances
and angles are listed in Table 2. An ORTEP representation
of the complex is shown in Figure 1.
but, as is almost always done, it was calculated from
calc
calc
Compound 1 is a complex of Mn(III) in an ax-N2-eq-O4
coordination environment. This coordination sphere is well-
known for complexes formally of Mn(II), indeed with axial
py and equatorial π-conjugated O2 donor ligands (e.g.,
catecholates/quinones).33,34 The ax-N2-eq-O4 coordination
environment is less common for Mn(III), although there are
several known examples of this donor set.35-37 Complex 1,
however, represents the first of this coordination environment
for mononuclear Mn(III) in which the axial ligands are py
(or similar heterocycles) and the equatorial ligands are
π-conjugated O2 donors. A very similar coordination sphere
has been reported for a dinuclear complex of Mn(III).38
As is often found for pseudooctahedral high-spin d4
systems, complex 1 displays a static Jahn-Teller distortion
in the form of a tetragonal elongation. The long axis is
formed by the N-Mn vectors arising from pyridine mol-
ecules, with bond distances of 2.273(1) and 2.291(1) Å,
respectively. The equatorial positions are occupied by the
O-donors of two dbm- chelating ligands, featuring Mn-O
bond distances in the range 1.902(1)-1.918(1) Å. The
idealized planes of the pyridine groups are inclined to one
another by 34.52(4)°. This is in contrast with the reported
dinuclear complex [Mn2(L)2(py)4] (H3L ) 3-(3-oxo-3phe-
nylpropionyl)-5-methylsalicylic acid), where each Mn(III)
is in the same coordination environment as in 1 but with the
trans pyridine ligands almost perpendicular to each other.38
Three of the four phenyl groups per molecule of 1 are
participating in intermolecular π-π stacking interactions
(range of intercentroid distances 3.570-3.781 Å), thereby,
each molecule being connected to two other complexes via
either one or two such interactions, respectively. This
supramolecular arrangement leads to infinite chains of
molecules within the crystal structure, which is reminiscent
of a similar organization observed in the solid state of the
molecules [MnCl(dbm)2] or [MnCl(Me2dbm)2], which are
related to 1.31 Between the [Mn(dbm)2(py)2] chains lie the
N
δfk
δfk
Hij )
(4)
∑
(
)(
)
δpi
δpj
k)1
The derivatives of the resonance fields, fkcalc, with respect to the
parameters p had to be evaluated numerically. Errors in the best-
fit parameters were finally estimated as
ø2
σi )
(H-1
)
(5)
ii
xN - P
where N is the number of experimental resonance fields and P is
the number of fitted parameters.
Results
Synthesis. The chelating ligand Hdbm has been used in
the preparation of a considerable number of high oxidation
state Mn complexes with a wide range of nuclearities.30-32
In the synthesis of such compounds, an often encountered
challenge is that of preventing the formation of the very
stable mononuclear tris-chelate complex [Mn(dbm)3].3 How-
ever, it is possible to prepare in pure form the bis-chelate
monomers [MnX(dbm)2] (X ) Cl, Br)31 or [Mn(dbm)2(py)2]-
(ClO4) if the stoichiometry is well controlled and the
appropriate terminal ligand (X or py) is present. Complex 1
was prepared in a manner very similar to that previously
described.32 The preparation consists of the generation in situ
of Mn(III) in the presence of pyridine and a counterion,
followed by the addition of the chelating ligand Hdbm. Given
the absence of a suitable source of Mn(III) with labile ligands
that is soluble in organic solvents, this ion can be formed
by comproportionation of Mn(II) and Mn(VII), prior to the
addition of Hdbm, according to the redox scheme
4Mn(II) + Mn(VII) f 5Mn(III)
(6)
The product forms in MeCN solution and precipitates pure,
with no need for recrystallization. A balanced chemical
equation can be written for this transformation (eq 7). Large
crystals of the compound 1 can be obtained directly from
the reaction mixture if it is left unperturbed after mixing all
reagents.
-
ClO4 counterions.
HFEPR. The most important aspect of HFEPR spectra
of the polycrystalline complex is the complete absence of
(33) Lynch, M. W.; Hendrickson, D. N.; Fitzgerald, B. J.; Pierpont, C. G.
J. Am. Chem. Soc. 1984, 106, 2041-2049.
(34) Attia, A. S.; Pierpont, C. G. Inorg. Chem. 1995, 34, 1172-1179.
(35) Shongwe, M. S.; Vandayar, S. K.; Fernandes, M. A.; Marques, H.
M.; Morris, M. J.; Heath, S. L. Polyhedron 2001, 20, 2195-2201.
(36) Nakamura, T.; Niwa, K.; Usugi, S.; Asada, H.; Fujiwara, M.;
Matsushita, T. Polyhedron 2001, 20, 191-201.
(37) Bermejo, M. R.; Deibe, A. G.; Rey, M.; Sanmartin, J.; Sousa, A.;
Aurangzeb, N.; Hulme, C. E.; McAuliffe, C. A.; Pritchard, R. G.;
Watkinson, M.; Helliwell, M. J. Chem. Soc., Dalton Trans. 1994,
1265-1269.
(38) Arom´ı, G.; Gamez, P.; Roubeau, O.; Berzal, P. C.; Kooijman, H.;
Spek, A. L.; Driessen, W. L.; Reedijk, J. Inorg. Chem. 2002, 41, 3673-
3683.
(29) Wilkinson, J. H. The Algebraic EigenValue Problem; Clarendon
Press: London, 1970.
(30) Arom´ı, G.; Aubin, S. M. J.; Bolcar, M. A.; Christou, G.; Eppley, H.
J.; Folting, K.; Hendrickson, D. N.; Huffman, J. C.; Squire, R. C.;
Tsai, H. L.; Wang, S.; Wemple, M. W. Polyhedron 1998, 17, 3005-
3020.
(31) Arom´ı, G.; Knapp, M. J.; Claude, J. P.; Huffman, J. C.; Hendrickson,
D. N.; Christou, G. J. Am. Chem. Soc. 1999, 121, 5489-5499.
(32) Wang, S.; Wemple, M. S.; Yoo, J.; Folting, K.; Huffman, J. C.; Hagen,
K. S.; Hendrickson, D. N.; Christou, G. Inorg. Chem. 2000, 39, 1501-
1513.
190 Inorganic Chemistry, Vol. 44, No. 2, 2005

Synthesis and Characterization of [Mn(dbm)2(py)2](ClO4)
Table 1. Crystallographic Data for [Mn(dbm)2(py)2](ClO4) (1)
cryst shape/color
cryst size, mm
formula
fw
block/dark brown
0.40 × 0.40 × 0.40
C40H32ClMnN2O8
759.07
cryst system
space group
a, Å
monoclinic
P21n
8.2851(4)
b, Å
24.0826(9)
c, Å
17.9576(9)
R, deg
90
â, deg
100.144(4)
90
3527.0(3)
γ, deg
V, Å3
Z
4
T, K
F
153.2
1.429
0.710 73
0.508
calc, g/cm3
wavelength (Mo KR), Å
µ, mm-1
unique data
9792
R,a wR2b (I > 2σ(I))
0.0367, 0.0935 (8317 reflns)
2
2
2
a R ) Σ||Fo| - |Fc||/Σ|Fo|. b wR2 ) [Σ[w(Fo - Fc )2]/Σ[w(Fo )2]]1/2
.
Table 2. Selected Bond Distances (Å) and Angles (deg) for
[Mn(dbm)2(py)2](ClO4) (1)
Mn(1)-O(4)
Mn(1)-O(3)
Mn(1)-O(1)
1.902(1)
1.910(1)
1.910(1)
Mn(1)-O(2)
Mn(1)-N(2)
Mn(1)-N(1)
1.918(1)
2.273(1)
2.291(1)
O(4)-Mn(1)-O(3)
O(4)-Mn(1)-O(1)
92.67(4)
87.04(4)
O(3)-Mn(1)-O(1) 174.98(5)
O(4)-Mn(1)-O(2) 174.87(5)
O(1)-Mn(1)-N(2)
O(2)-Mn(1)-N(2)
O(4)-Mn(1)-N(1)
O(3)-Mn(1)-N(1)
O(1)-Mn(1)-N(1)
O(2)-Mn(1)-N(1)
94.17(4)
87.36(5)
90.73(4)
88.96(4)
86.03(4)
94.10(4)
Figure 3. HFEPR spectra of solid polycrystalline 1 at two selected
frequencies and their simulations: (A) experiment at 187.51 GHz and 40
K; (B) simulation at the same conditions as A; (C) experiment at 438.15
GHz and 40 K; (D) simulation at the same conditions as C. Spectra B and
D were simulated using spin Hamiltonian parameters as given in Table 3.
The line marked with an asterisk originates from a Mn(II) impurity and is
not reproduced in the simulations.
O(3)-Mn(1)-O(2)
O(1)-Mn(1)-O(2)
O(4)-Mn(1)-N(2)
O(3)-Mn(1)-N(2)
89.13(4)
91.58(4)
87.82(5)
90.83(4)
N(2)-Mn(1)-N(1) 178.53(5)
any field-induced torquing phenomena, which occur very
frequently in other solid Mn(III) complexes.10 This results
in perfect powder spectra, free of artifacts, and allows the
experimentalist to work with the solid “as is” without a need
to implement crystallite-constraining methods. Spectra could
be recorded in the 4.2-77 K temperature range, with signal-
to-noise ratio very good at low temperatures, but becoming
progressively worse at the upper limit of that range. No
temperature-dependent changes in the spectra other than
intensity were observed. Figure 3 shows HFEPR spectra at
40 K of polycrystalline complex 1 recorded at two selected
frequencies accompanied by simulations assuming a perfectly
random orientation of the crystallites with regard to the
magnetic field. (The simulation parameters were obtained
from a multifrequency dataset in a method described before
and not from individual spectra.) A striking feature of Figure
3 is the nearly prefect agreement between the simulations
and the experiments not only in terms of the position of the
resonances but also their shape and intensity, which confirms
the absence of field-induced artifacts. This agreement persists
to the highest frequencies and fields used in this work (440
GHz, and 15 T, respectively). Indeed, the best spectra were
obtained at the highest limit of the available frequency range.
As some of us argued elsewhere,6 the magnetic properties
of a given complex in a solid not only depend on its intrinsic
characteristics but are often influenced by the properties of
the crystal lattice. It is therefore desirable to carry out an
experiment on a complex isolated in an “inert” (noncoordi-
nating) solvent, necessarily here as a frozen solution, to
eliminate the unwanted crystal effects. We have thus
performed HFEPR measurements on 1 dissolved in CH2Cl2
in a concentration of about 100 mM. The resulting spectra
are shown in Figure 4, at two selected frequencies, ac-
companied again by simulations assuming a perfectly random
orientation of the molecules with regard to the field. The
very good agreement between the simulations and the
experiment indicates a high quality of the frozen solution;
however, the spectra also show a line-broadening effect
compared to the solid state. Most importantly, there is very
little difference in spin Hamiltonian parameters between the
solid and the frozen solution, which proves that CH2Cl2 is
an “innocent” solvent for 1 and also that the properties of
the complex in its solid form are not determined by the
crystal packing constraints or similar factors.39
Rather than extracting spin Hamiltonian parameters by
fitting them to individual single-frequency spectra, as is
usually done in EPR, we collected all the observed reso-
nances at multiple frequencies in a single 2-D array
(resonance fields as function of sub-THz quantum energy),
which is indicated by squares in Figure 5 for the case of
polycrystalline 1. We have previously used this procedure
for triplet (S ) 1) spin systems, such as molecular oxygen,23
(39) HFEPR spectra for 1, either as a solid or diluted in frozen solution,
do not exhibit resolved hyperfine splitting from 55Mn, in concurrence
with all other HFEPR studies of Mn(III) molecular complexes known
to us.
Inorganic Chemistry, Vol. 44, No. 2, 2005 191

Arom´ı et al.
Figure 5. Dependence of the resonance fields on sub-THz quantum energy
in polycrystalline 1. The squares represent the experimental resonances
including the single “non-Kramers” line observed at X-band, while the lines
were drawn using the spin Hamiltonian parameters as in Table 3, including
the fourth-order terms. The red lines are the x canonical turning points in
the powder pattern, the blue lines are the y turning points, and the black
lines are the z turning points, while the green lines represent the off-axis
turning points.
Figure 4. HFEPR spectra of a solution of 1 in methylene chloride at two
selected frequencies and their simulations: (A) experiment at 92.818 GHz
and 30 K; (B) simulation at the same conditions as A; (C) experiment at
273.45 GHz; (D) simulation at the same conditions as C. Spectra B and D
were simulated using spin Hamiltonian parameters as given in Table 3.
The line marked with an asterisk originates from a Mn(II) impurity while
the lines marked with crosses are generated by the fourth harmonic of the
fundamental frequency. Neither of these lines are reproduced in the
simulations.
strongly correlated with E (correlation coefficient 0.604; see
Supporting Material), which is not surprising as their
2
2
respective operators, O22 and O4 , both contain terms in [S+
2
2
+ S- ], and is probably the reason B4 is rarely reported.
Parameter B4 , on the other hand, is only weakly correlated
4
with the other Hamiltonian terms and is significantly different
from zero, taking a value 0.0048(4) cm-1. The actual
correlation matrix (see Supporting Material) indicates no
excessive dependencies among other parameters in the solid
sample.
high-spin Ni(II),40 and V(III),41 as well as for a quartet (S )
3/2), Co(II).42 A quintet spin state (S ) 2), particularly with
rhombic symmetry, as studied here, however, displays many
more possible transitions than does a spin triplet or quartet.
Consequently, we employed the software as described in the
Experimental Section to extract spin Hamiltonian parameters
from the complete 2-D dataset. This approach ensures that
an equally optimal fit is achieved at any frequency and allows
one to rigorously calculate the experimental error from the
statistical analysis.43
The final fit quality can be judged visually by comparing
the experimental resonances (squares in Figure 5) with the
simulated curves (lines in Figure 5). The effect of fourth-
order parameters on improving the fits was also observable
in single-frequency spectra. Particular transitions are sensitive
to fourth-order zfs parameters to a varying degree. We found
the strongest dependence in spectra obtained at 377 GHz,
as shown in Figure 6, which presents the high-field part of
an EPR spectrum of polycrystalline 1 and compares it to
spectra simulated by using the best-fit parameters with fourth-
order zfs terms either present or absent. Clearly, a better
agreement is achieved when the fourth-order zfs terms are
included in the simulations. The final set of spin Hamiltonian
parameters for solid 1 is shown in Table 3. For the frozen-
solution spectra, an analogous analysis of a 2-D dataset was
performed, and the spin Hamiltonian parameters obtained
from it are also given in Table 3. Extraction of fourth-order
zfs parameters, however, was not successful in the frozen
solution, since the statistical error was on the order of, or
even larger than, the parameters themselves (Table 3). This
was primarily caused by the increased line width of the
solution resonances compared to the solid-state spectra,
For the polycrystalline sample of 1, a very satisfactory
agreement was found between the experimental and simula-
tion resonant fields using the set of spin Hamiltonian
parameters including only the Zeeman term and second-order
zero-field splitting terms D and E (rms ) 48 mT). However,
if fourth-order terms of the Hamiltonian (eq 1) are included,
then the agreement becomes significantly better (rms ) 16
2
4
mT). In particular, both the parameters B4 and B4 are
nonnegligible. Parameter B42 of 0.0007(3) cm-1 is relatively
(40) Krzystek, J.; Park, J.-H.; Meisel, M. W.; Hitchman, M. A.; Stratemeier,
H.; Brunel, L.-C.; Telser, J. Inorg. Chem. 2002, 41, 4478-4487.
(41) Krzystek, J.; Fiedler, A. T.; Sokol, J. J.; Ozarowski, A.; Zvyagin, S.
A.; Brunold, T. C.; Long, J. R.; Brunel, L.-C.; Telser, J. Inorg. Chem.
2004, 43, 5645-5658.
(42) Krzystek, J.; Zvyagin, S. A.; Ozarowski, A.; Fiedler, A. T.; Brunold,
T. C.; Telser, J. J. Am. Chem. Soc. 2004, 126, 2148-2155.
(43) Tregenna-Piggott, P. L. W.; Weihe, H.; Barra, A.-L. Inorg. Chem.
2003, 42, 8504-8508.
192 Inorganic Chemistry, Vol. 44, No. 2, 2005

Synthesis and Characterization of [Mn(dbm)2(py)2](ClO4)
be significant in specific cases, as in high-spin molecular
clusters such as Mn12, where they were proposed to influence
quantum tunneling probability between the particular (MS
states through breaking the symmetry of the zfs tensor.44 It
is thus reassuring that the fourth-order parameters extracted
by us for the mononuclear Mn(III) complex, 1, from its
powder HFEPR spectra fall within the same range as
determined in the two cases for Mn(III) studied by single-
crystal EPR. In the case of the [Mn(cyclam)Br2]Br complex,
4
B4 was also established as the largest of the three fourth-
order parameters and found positive14 as in the case of 1
(17 × 10-4 vs 48 × 10-4 cm-1). In the case of Mn(III) in
rutile, only the cubic parameter a was reported and was equal
m
to 0.13 cm-1,4 which when recalculated into the B4
0
parameters, assuming cubic symmetry, yields B4 ) 11 ×
10-4, equivalent to B44 ) 55 × 10-4 cm-1, which is in good
4
agreement with B4 ) 48 × 10-4 cm-1 determined for
complex 1. Also, although we were unable to accurately
determine fourth-order zfs parameters in the complex Mn-
(acac)3, whose second-order zfs values are very close to those
of complex 1,6 the estimate of the cubic parameter a ≈ 0.15
Figure 6. High-field part of a HFEPR spectrum of polycrystalline 1 at
377.3 GHz. For transition identification, one should draw a vertical line in
Figure 5 corresponding to the experimental frequency. The middle (black)
trace is the experiment. The uppermost (blue) trace was simulated with the
following best-fit spin Hamiltonian parameters without fourth-order terms:
D ) -4.509 cm-1; E ) -0.422 cm-1; gx ) 1.989; gy ) 1.994; gz ) 1.983.
The lowest (red) trace was simulated with the best-fit parameters that did
include fourth-order terms as given in Table 3.
4
cm-1 yields B4 ≈ 63 × 10-4 cm-1, again in reasonable
agreement with the value of 48 × 10-4 cm-1 found in
complex 1.
The accurate determination by HFEPR of not only second-
order but also fourth-order zfs parameters of the spin
Hamiltonian for a quintet (S ) 2) state is quite instructive
as to the perils of determining these parameters from single-
frequency spectra, particularly those obtained at low frequen-
cies, such as 9.5 GHz typical for X-band EPR. We have
shown previously that such a procedure when applied to a
Mn(III) complex leads to underdetermination;6 that is,
different sets of parameters may equally well simulate a given
“non-Kramers” X-band signal. In the case of complex 1, the
non-Kramers signal at X-band is only weakly dependent on
the fourth-order zfs parameters, so that their inclusion in the
simulation leads to a shift of the resonance amounting to
ca. 4 mT (from 87 to 83 mT at 9.45 GHz) compared to the
line width of ca. 25 mT. As shown in Figure 6, certain
resonances appearing at much higher frequencies are more
sensitive to fourth-order zfs parameters, as their inclusion
in the simulations leads to spectral shifts as large as 150
mT compared to experimental line widths on the order of
50 mT. HFEPR is thus inherently more accurate in determin-
ing spin Hamiltonian parameters in high-spin rhombic Mn-
(III) complexes (and other high-spin systems characterized
by large zfs anisotropy) than conventional EPR spectroscopy,
provided the line widths and line shapes are sufficiently
narrow and well-defined to allow a precise reading of the
resonance field position.
which made the accurate determination of particular turning
points more difficult than in the solid sample. Also, fewer
data points were collected in the case of the frozen solution
than for the solid sample.
Discussion
The HFEPR spectra of the solid polycrystalline 1 are
remarkable for their high quality, which is at least partly
due to the absence of any field-induced torquing effects. This
results in a system that can be considered as a model rhombic
S ) 2 spin complex and can be measured as a solid without
any preparation of the sample to yield repeatable and highly
accurate results.
The high quality of the HFEPR data allowed us for the
first timesto our best knowledgesto determine the fourth-
order zfs parameters of an S ) 2 spin system from powder
2
4
spectra. In particular, both parameters B4 and B4 were found
to be significantly different from zero, 7(3) × 10-4 and 48-
(4) × 10-4 cm-1, respectively (the other fourth-order
0
2
parameter, B4 , is zero within statistical error). B4 appears
to be correlated with E, however, which is probably the
reason it is rarely considered. Previously, fourth-order zfs
parameters in a S ) 2 system could be extracted exclusively
from single-crystal spectra.4,14 Although the determination
of the fourth-order zfs parameters could be considered an
academic enterprise, these terms of the spin Hamiltonian may
Table 3. Spin Hamiltonian Parameters for 1 Obtained from a Simultaneous Least-Squares Fit to the Complete Dataset for the Polycrystalline Solid
(Represented by Squares in Figure 5) and an Analogous Dataset Collected for the Solution (Not Shown)
D (cm-1
)
E (cm-1
)
B40 (10-4 cm-1
)
B42 (10-4 cm-1
)
B44 (10-4 cm-1
)
gx
gy
gz
polycryst
soln
-4.504(2)
-4.500(8)
-0.425(1)a
-0.430(5)
-1.8(4)
0(1)
7(3)
-3(12)
48(4)
18(17)
1.993(1)
1.990(4)
1.994(1)
1.978(5)
1.983(1)
1.972(5)
a Because the choice of the sign of E is arbitrary, we followed a convention that attributes to E the same sign as of the parameter D, i.e., in our case
negative. Changing the sign of E would require a corresponding permutation of the gx and gy values.
Inorganic Chemistry, Vol. 44, No. 2, 2005 193

Arom´ı et al.
Table 4. Summary of Spin Hamiltonian Parameters for Mn(III) Complexes
0
2
4
coord
sphere
B4
B4
(10-4 cm-1
B4
complex
D (cm-1
)
E (cm-1
)
(10-4 cm-1
)
)
(10-4 cm-1
)
gx
gy
gz
a
N6
Mn(terpy)(N3)3
Mn(bpea)(N3)3
-3.29(1)
+3.50(1)
0.51(1)
0.82(1)
-0.0135(6)
0.034
2.000(5)
2.02(1)
2.005(4 )
2.00
2.005(3 )
2.005(5 )
1.980(5)
1.98(1)
2.036(2 )
2.00
2.005(3 )
2.005(5 )
2.010(5)
1.95(1)
2.015(2)
1.99
1.98(2)
2.00(2)
1.984
2.00(1)
2.00(2)
1.98(1)
1.980(4)
2.00(1)
1.96(1)
1.98(1)
2.00(2)
1.983(1)
1.97
b
N4Br2
N4I2
N4Cl
[Mn(cyclam)Br2]Brc
-1.1677(7)
+0.604
9(3)
6(2)
17(3)
d
Mn(cyclam)I2
Mn(TPP)Cle
Mn(Pc)Cle
-2.290(5)
-2.31(1)
-2.33(1)
-2.40(1)
-3.12(2)
-1.07(1)
-2.69(2)
-2.64(1)
-3.82(2)
-3.67(2)
-2.47(2)
-4.504(2)
-4.35
-4.52(2)
-4.524(1)
-4.491(7)
-4.431(9)
-3.4(1)
-2.45(3)
-1.40(2)
0.00(1)
0.00(1)
0.00(1)
e0.02(1)
0.00(1)
0.00(1)
0.030(3)
0.015(5)
0.75(2)
0.70(2)
0.17(1)
-0.425(1)
0.26
0.25(2)
0.276(1)
0.248(5)
0.258(8)
0.116(1)
0.00(1)
0.00(1)
Mn(ODMAPz)Clf
Mn(OEP)Clg
Mn(TSP)h
2.00(1)
2.00(2)
2.01(1)
1.994(4 )
2.02(1)
1.97(2)
1.96(1)
2.00(2)
1.993(1 )
1.99
2.00(1)
2.00(2)
2.01(1)
1.994(4)
2.02(1)
2.04(1)
1.98(1)
2.00(2)
1.994(1)
1.99
N4Br
N4O
N4
Mn(OEP)Brg
Mn(tpfc)(OPPh3)i
Mn(cor)j
b
N3F3
Mn(terpy)F3
b
Mn(bpea)F3
N2O2C l
N2O4
O6
Mn(salen)h
k
[Mn(dbm)2(py)2](ClO)4
-1.8(4)
7(3)
48(4)
55
l
Mn(dbm)3
Mn(acac)3
m
1.99(1)
1.99(1)
1.99(1)
CsMn(SO4)2‚12D2On
CsMn(SO4)2‚12D2Oo
CsMn(SO4)2‚12H2Oo
1.981(5 )
2.001(5)
2.00(2)
2.03(2)
1.98(2)
1.993(5)
1.997(7)
2.00(2)
2.03(2)
1.98(2)
1.988(5)
1.966(12)
1.99(1)
2.02(2)
1.98(2)
p
Mn3+ in TiO2
11
O4Cl
O4Br
Mn(Me2dbm)Clg
Mn(Me2dbm)Brg
a Reference 11. b Reference 12. c Reference 14. d Reference 13. e Reference 10. f Reference 2. A complex with another axial ligand, DTC, is also reported.
g Reference 7. An analogous complex with iodine is also described. h Reference 16. i Reference 9. j Reference 8. k This work. l Reference 3. m Reference 6.
n Reference 5. These workers employed zero-field inelastic neutron scattering (INS), which directly provides the zfs parameters but gives no information on
g values. o Reference 43. p Reference 4. These authors reported the cubic fourth-order zfs parameter a equal to 0.13 cm-1, which we recalculated along the
convention we use with an assumption of a cubic symmetry. See note 25.
The successful determination of the fourth-order zfs terms
and the high precision obtained for the other spin Hamilto-
nian parameters (2 × 10-3 cm-1 for D, 1 × 10-3 cm-1 for
E, and 1 × 10-3 for the g matrix values) was possible not
only because of the high spectral quality but also through
the use of a novel software. Traditionally, spin Hamiltonian
parameters were fitted to single-frequency spectra in EPR
because most spectrometers operated at a single (low)
frequency. With the advent of HFEPR it was realized that
truly optimal parameters were those that fit several frequen-
cies, particularly for high-spin systems. A multifrequency
approach was thus adopted,44 with the spectra recorded at
several frequencies and the Hamiltonian parameters fitted
to them for the most part using human judgment. Since
human judgment has its limitations, it was not unusual to
report parameters that were not entirely frequency-indepen-
dent but allowed for some variation depending on the
frequency. The methodology that we report here was
previously used by some of us in determining Hamiltonian
parameters in S ) 1 systems such as Ni(II), V(III), and
dataset, as opposed to individual spectra. This approach
ascertains that the equally optimal fit is achieved at any
frequency and allows one to rigorously calculate the experi-
mental error from the statistical analysis. This is particularly
important in the quintet spin state (S ) 2) as studied here,
since many more possible transitions between the spin
sublevels are observed than in a triplet or a quartet, making
the field vs energy diagrams such as the one shown in Figure
3
5 much more complicated than for S ) 1 or S ) /2. In
addition to a vastly increased number of canonical reso-
nances, the quintet state powder pattern contains multiple
off-axis turning points (green lines in Figure 5), which were
used in the fits along with the canonical orientations. The
resulting agreement between the experiment and simulation
can be judged from both single-frequency spectra (Figures
3 and 6 for the solid, Figure 4 for the frozen solution) and
the multifrequency dataset (Figure 5, solid only; the frozen
solution produced an analogous dataset, and the agreement
was almost as good).
The large amount of data obtained by HFEPR and other
techniques, such as INS,5 allows us to survey qualitatively
the relation between structure and spin Hamiltonian param-
eters for a variety of Mn(III) complexes, which are sum-
marized in Table 4. The zfs parameters are most informative
in this regard, as described below. In principle, the g values
could also be correlated with molecular and electronic
structure; however, this is not apparent. In all of these
complexes of Mn(III), both the isotropic deviation from ge
(2.00) and anisotropy in g are minimal: the g values range
overall between 1.95 and 2.03 but are normally 1.99 ( 0.02.
One major structural type of Mn(III) complexes is square-
pyramidal, with either bidentate or macrocyclic equatorial
3
molecular oxygen23,40,41and subsequently also in a S ) /2
system, Co(II).42 Very recently it was employed in a S ) 2
system (high-spin Fe(II)).45 A similar approach has been used
by Tregenna-Piggott et al.43,46 and by Mossin et al.14 The
method depends on collecting all the observed resonances
in a 2-D array (resonant field vs transition energy) and fitting
the Hamiltonian parameters simultaneously to the complete
(44) Barra, A.-L.; Brunel, L.-C.; Gatteschi, D.; Pardi, L.; Sessoli, R. Acc.
Chem. Res. 1998, 31, 460-466.
(45) Ozarowski, A.; Zvyagin, S. A.; Reiff, W. M.; Telser, J.; Brunel, L.
C.; Krzystek, J. J. Am. Chem. Soc. 2004, 126, 6574-6575.
(46) Tregenna-Piggott, P. L. W.; Weihe, H.; Bendix, J.; Barra, A.-L.; Gu¨del,
H.-U. Inorg. Chem. 1999, 38, 5928-5929.
194 Inorganic Chemistry, Vol. 44, No. 2, 2005

Synthesis and Characterization of [Mn(dbm)2(py)2](ClO4)
5
ligands and a variety of axial ligation. Complexes with axial
chloride are all characterized by strictly axial zfs tensors (E
) 0 within experimental error) with the zfs parameter D
ranging from -2.30 to -2.50 cm-1, regardless of whether
the equatorial coordination is N4 (porphyrinic) or O4 (bis-
(diketonate)).47 Replacing the porphyrinic ligands with
corroles increases the magnitude of D to -2.60 to -2.70
cm-1 8,9 and introduces a small nonzero |E| of 0.015-0.030
cm-1 due to reduced symmetry of the corrole macrocycle.
A combination of N and O donors in the equatorial
coordination sphere, as in Mn(salen) (N2O2Cl),16 preserves
the D value of about -2.50 cm-1 but introduces a moderate
rhombicity of E ) -0.17 cm-1.
The other major structural type of Mn(III) complexes is
distorted octahedral. The most common among these is the
O6 coordination sphere, corresponding to a tetragonally
elongated octahedron. Regardless of whether the coordination
is by tris(diketonates) or hexaaqua ligands, these complexes
are characterized by a D between -4.35 and -4.52 cm-1
and a moderate E very close to -0.25 cm-1.48 The complex
of interest here, 1, despite having a mixed donor set (axial
N2, equatorial O4) fits within this range of D as well. The
class of complexes characterized by the coordination sphere
of N6 and N3F311,12 displays D on the order of -3.30 to -3.80
cm-1. The very low symmetry of these complexes causes a
strong rhombicity of the zfs tensor, with |E| varying from
ca. 0.50 to 0.80 cm-1. In one particular case, this results in
the change of sign of D, which for Mn(bpea)(N3)3 becomes
+3.50 cm-1.12 Finally, introducing a heavy atom into the
coordination sphere of Mn(III) results in pronounced changes
of D, due to the significant spin-orbit coupling attendant
with heavy atoms, even when they are present only as
ligands.13 Thus, in the three known instances where Br was
coordinated to Mn(III), N4Br and O4Br7 and N4Br2,14 the
magnitude of D is significantly reduced to between -1.07
and -1.40 cm-1. The presence of iodine in the coordination
sphere influences the zfs parameters even further, with D
close to zero in Mn(OEP)I7 and achieving a positive value
(+0.604 cm-1) in Mn(cyclam)I2.13
within the Eg (in Oh or S6 symmetry) undistorted ground
state, and the visible band was assigned to 5Ag f 5T2g (Oh),
which corresponds to the octahedral splitting (in strong-field
3
2
notation: t2 e1 f t2 e2), wherein the deviations from octa-
hedral symmetry were not resolved.43 These assignments,
combined with the extensive data available for aqua ligands,
allowed these workers to estimate the bonding parameters
(using the AOM apparatus50) of the axial and equatorial aqua
ligands in [Mn([H,D]2O)6]3+, which could then be used in
combination with estimates for the Racah and spin-orbit
coupling parameters to calculate values for spin Hamiltonian
parameters in excellent agreement with experiment.43
An analogous analysis is exceedingly difficult here. Unlike
aqua complexes, the electronic absorption spectrum of metal
diketonates is dominated by charge-transfer and/or π-π*
transitions involving the π-conjugated ligands. Nevertheless,
for Mn(acac)3 and various other tris(â-diketonates) and
related complexes (e.g., K3Mn(ox)3, ox ) C2O42-), it is
possible to observe a ligand-field (d-d) transition in the NIR
5
5
region, which is assigned in D4h symmetry to B1g f A1g
(i.e., the tetragonal splitting of Eg).51 Two features in the
visible region are also observed that can be assigned to
ligand-field transitions: B1g f 5B2g; 5B1g f 5Eg. These three
bands allow a unique determination of the three Ballhausen52
tetragonal crystal-field parameters, Dq, Ds, and Dt.51 Use
of these parameters for Mn(acac)3 in combination with the
Racah and spin-orbit parameters employed for [Mn-
(H2O)6]3+ (∼80% of the free-ion values), yields an axial zfs
parameter, D ≈ -3.7 cm-1, with a small cubic component
(a ≈ 0.06 cm-1), reasonably close to that determined here
for 1.
5
5
Thus, the electronic absorption features observed for 1 in
the range 480-520 nm (∼20 000 cm-1) could likewise be
5
5
assigned to Eg f T2g transitions in Oh symmetry so that
the octahedral splitting in [Mn(dbm)2(py)2] + is indeed similar
to that in [Mn(H2O)6]3+ and Mn(III) tris(â-ketonate) com-
plexes. However, in contrast to these complexes, no bands
are visible in the NIR region that would shed light onto the
+
magnitude of tetragonal splitting in [Mn(dbm)2(py)2] to
enable a calculation using a crystal-field model, which could
refine the calculation described above.
Ideally, the spin Hamiltonian parameters derived from
analysis of the HFEPR data should also be correlated with
electronic energy levels obtained from optical spectroscopy.
For example, in their study of the cesium manganese alum
(CsMn(SO4)2‚12[H,D]2O), Tregenna-Piggott et al. also re-
corded variable-temperature, single-crystal electronic absorp-
tion spectra that clearly showed two ligand-field bands, one
in the NIR at 10 400 cm-1 (960 nm) and one in the visible
region at 21 140 cm-1 (470 nm).43 The hexaaqua Mn(III)
cation has octahedral symmetry that is reduced to S6 due to
the aqua ligands, and Jahn-Teller distortion further lowers
the symmetry to Ci.49 The NIR band was thus assigned to
5Ag f 5Ag (Ci), which corresponds to the tetragonal splitting
A better method, in any case, is to use the AOM, as was
done for the aqua complexes.43 Unfortunately, complex 1 is
heteroleptic, which increases the number of AOM bonding
parameters; moreover, the pyridine and dbm ligands are
inherently much more complex than the aqua ligand, due to
their π-conjugation (the Orgel effect).53 Therefore, we can
only note the remarkable similarity in D values between
[Mn(dbm)2(py)2] + and [Mn([H,D]2O)6]3+,5,43 as well as the
â-diketonate complexes with O6 coordination: Mn(acac)3,6
and Mn(dbm)3.3 Indeed, as seen from Table 3, the difference
in D values between [Mn(dbm)2(py)2] + and [Mn(D2O)6]3+,
both as determined by HFEPR, is less than the difference in
(47) There is a sole exception to this rule: the complex tetraphenylsul-
fonateporphyrinatoMn(III) has a much larger magnitude of D equal
to -3.12 cm-1 16 but still a rigorously axial zfs tensor.
(50) Scha¨ffer, C. E. Struct. Bonding 1968, 5, 68-95.
(51) Davis, T. S.; Fackler, J. P., Jr.; Weeks, M. J. Inorg. Chem. 1968, 7,
1994-2002.
(48) An exception is found for Mn(III) as a dopant in rutile, but this is not
a molecular complex.
(52) Ballhausen, C. J. In Introduction to Ligand Field Theory; McGraw-
Hill: New York, 1962; pp 99-103.
(49) Tregenna-Piggott, P. L. W.; Andres, H.-P.; McIntyre, G. J.; Best, S.
P.; Wilson, C. C.; Cowan, J. A. Inorg. Chem. 2003, 42, 1350-1365.
(53) Scha¨ffer, C. E.; Yamatera, H. Inorg. Chem. 1991, 30, 2840-2853.
Inorganic Chemistry, Vol. 44, No. 2, 2005 195

Arom´ı et al.
D values for [Mn(D2O)6]3+, as determined by HFEPR43
versus INS.5 Likewise, the difference in D values between
the protio and deuterio hexaaqua Mn(III) cations is greater
than the difference in D values between [Mn(dbm)2(py)2] +
HFEPR spectra were recorded for 1 as both a polycrystalline
solid and in CH2Cl2 frozen solution over a wide range of
frequencies in the sub-THz regime. The complete dataset of
resonant magnetic fields versus sub-THz quantum energy
was analyzed using an automated fitting software. This
and [Mn(D2O)6]3+. Thus the AOM parameters determined
ax, eq 43
for [Mn(H2O)6]3+, e(σ,π)
,
would produce the D value
analysis yielded the following spin Hamiltonian parameters
0
for [Mn(dbm)2(py)2]+, simply by relabeling the parameters
as e(σ,π) and e(σ,π)dbm, respectively.
(energies in cm-1): D ) -4.504(2), E ) -0.425(1), B4
)
py
-1.8(4) × 10-4, B42 ) 7(3) × 10-4, B44 ) 48(4) × 10-4, gx
n
The sole difference between [Mn(dbm)2(py)2]+ and the
homoleptic diketonate or aqua complexes is the larger value
for the rhombic zfs, E, in 1. The origin of E was discussed
by Tregenna-Piggott and co-workers in their study of [Mn-
([H,D]2O)6]3+, who quantitatively related its magnitude to
the aqua ligand twist angle, æ.43 Given that we have no
independent basis for estimation of the AOM parameters in
[Mn(dbm)2(py)2]+, we simply note that the axial pyridine
ligands exhibit a significant twist with respect to the
equatorial plane, as described above in the description of
the molecular structure, and this effect could be responsible
for the marked rhombicity of the zfs tensor. The phase-
coupling of the equatorial dbm ligands54 could also play a
role in enhancing the rhombic splitting.
) 1.993(1), gy ) 1.994(1), and gz ) 1.983(1), where the B4
values represent fourth-order zero-field splitting terms that
are generally difficult to extract from experiment and usually
available only from sophisticated single-crystal measure-
ments. The present results demonstrate the applicability of
HFEPR to high-precision measurements, even for polycrys-
talline samples in propitious cases such as complex 1.
Comparison with other six-coordinate complexes of Mn-
(III), in particular hexaaqua and tris(â-diketonate) complexes
suggests that the electronic structure of all of these shows a
truly remarkable commonality, despite distinct variation in
ligation.
Acknowledgment. We acknowledge the Spanish Minis-
try of Science and Technology for the contract “Ramo´n y
Cajal” (G.A.) and the Swiss National Science Foundation
(H.-M.S.-E.). The NHMFL is funded by the NSF through
the Cooperative Agreement No. DMR-0084173 and the State
of Florida.
Conclusions
A mononuclear complex of Mn(III), [Mn(dbm)2(py)2]-
(ClO4), 1, was prepared and structurally characterized. The
complex has axial pyridine ligands and equatorial O-donor
ligands within a π conjugated system, which represents a
structural type that has been reported in many instances for
Mn(II) but not for mononuclear Mn(III). Very high quality
Supporting Information Available: An X-ray CIF file, the
details of powder spectra simulations, and the Hessian and
correlation matrices resulting from the statistical analysis of HFEPR
dataset. This material is available free of charge via the Internet at
http://pubs.acs.org.
(54) Ceulemans, A.; Dendooven, M.; Vanquickenborne, L. G. Inorg. Chem.
1985, 24, 1153-1158.
IC049180U
196 Inorganic Chemistry, Vol. 44, No. 2, 2005