Multinuclear Paramagnetic NMR Spectra and Solid State X-ray Crystallographic Characterization of Manganese(III) Schiff-Base Complexes

Article abstract of DOI:10.1021/ic970747z

The ¹H NMR spectra of symmetrically substituted paramagnetic [(R,R'-SALOPHEN)Mn^III]⁺ and [(R,R'-SALOPHEN)Mn^III]₂(μ-O) complexes (SALOPHEN is 1,3-bis(salicylideneamino)benzene) were obtained. All of the monomers and dimers yielded well-resolved but isotropically shifted ¹H NMR spectra. A diamagnetic suppression routine was applied to the ¹H NMR spectra that allowed observation of all rapidly relaxing resonances. The binding of axial ligands, acetate, methanol, water, and pyridine was also observable by ¹H NMR. Two X-ray crystal structures were obtained: [(3,3′-17-crown-6-SALOPHEN)Mn^III]·(CH ₃C(O)O)·BaTf₂·2H₂O, [9b]OAc·BaTf₂·2H₂O, crystallized in the triclinic space group P1?; with a = 10.747(2) A?, b = 12.271(8) A?, c = 16.206(6) A?, α = 106.10(4)°, β = 103.07(2)°, γ = 106.06(3), and Z = 2, while [(3,3′,6,6′-(CH₃)₄-SALOPHEN)Mn ^III]PF₆·H₂O, [6b]PF₆·H₂O, crystallized in the monoclinic space group P2₁/n, with a = 12.465(1) A?, b = 13.366(2) A?, c = 15.240(1) A?, β = 103.051(7)°, and Z = 2. [9b]OAc·BaTf₂·2H₂O has the Mn(III) center ligated in the axial positions by one oxygen of the acetate group and an oxygen from one SO₃CF^3-. The second oxygen of the acetate is bound to the Ba²⁺ residing in the crown ether. The water molecule in [6b]PF₆·H₂O is O-bound to the Mn(III) center and also hydrogen bonded to one of the fluorine atoms of the PF₆ counterion, O-F 3.01 A?. ¹H and ¹⁹F NMR spectra revealed that the hydrogen bond in [6b]PF₆·H₂O was preserved in CD₃OD.

Full text of DOI:10.1021/ic970747z

4968
Inorg. Chem. 1997, 36, 4968-4982
Multinuclear Paramagnetic NMR Spectra and Solid State X-ray Crystallographic
Characterization of Manganese(III) Schiff-Base Complexes
Yangzhen Ciringh,† Scott W. Gordon-Wylie,† Richard E. Norman,‡ George R. Clark,§
Susan T. Weintraub,| and Colin P. Horwitz*,†
Department of Chemistry, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213, Department of
Chemistry and Biochemistry, Duquesne University, Pittsburgh, Pennsylvania 15282, Department of
Chemistry, The University of Auckland, Auckland, New Zealand, and Department of Biochemistry,
The University of Texas Health Science Center, San Antonio, Texas 78284
ReceiVed June 17, 1997X
The 1H NMR spectra of symmetrically substituted paramagnetic [(R,R′-SALOPHEN)MnIII]+ and [(R,R′-
SALOPHEN)MnIII]2(µ-O) complexes (SALOPHEN is 1,3-bis(salicylideneamino)benzene) were obtained. All of
1
the monomers and dimers yielded well-resolved but isotropically shifted H NMR spectra. A diamagnetic
suppression routine was applied to the 1H NMR spectra that allowed observation of all rapidly relaxing resonances.
1
The binding of axial ligands, acetate, methanol, water, and pyridine was also observable by H NMR. Two
X-ray crystal structures were obtained: [(3,3′-17-crown-6-SALOPHEN)MnIII]‚(CH3C(O)O)‚BaTf2‚2H2O, [9b]-
OAc‚BaTf2‚2H2O, crystallized in the triclinic space group P1h; with a ) 10.747(2) Å, b ) 12.271(8) Å, c )
16.206(6) Å, R ) 106.10(4)°, â ) 103.07(2)°, γ ) 106.06(3), and Z ) 2, while [(3,3′,6,6′-(CH3)4-SALOPHEN)-
MnIII]PF6‚H2O, [6b]PF6‚H2O, crystallized in the monoclinic space group P21/n, with a ) 12.465(1) Å, b ) 13.366-
(2) Å, c ) 15.240(1) Å, â ) 103.051(7)°, and Z ) 2. [9b]OAc‚BaTf2‚2H2O has the Mn(III) center ligated in the
axial positions by one oxygen of the acetate group and an oxygen from one SO3CF3-. The second oxygen of the
acetate is bound to the Ba2+ residing in the crown ether. The water molecule in [6b]PF6‚H2O is O-bound to the
Mn(III) center and also hydrogen bonded to one of the fluorine atoms of the PF6 counterion, O-F 3.01 Å. 1H
and 19F NMR spectra revealed that the hydrogen bond in [6b]PF6‚H2O was preserved in CD3OD.
Introduction
base complexes possessing alkane backbones.10-13 Chemical
shift assignments were made for ligand hydrogens lying outside
what is typically considered as the diamagnetic region (0-14
ppm) through systematic substitution of the ligand hydrogens.
Two important observations from those studies are that (a) ligand
geometry (e.g., planar Vs twisted) markedly affected the
chemical shift pattern for the ligand hydrogens10 and (b)
chemical reactions of paramagnetic Mn(III) complexes13,14 can
be followed by NMR spectroscopy.
Polymetallic coordination compounds of Mn(III) and Mn-
(IV) containing bridging oxo ligands have been studied as
structural and chemical models for the tetranuclear manganese
site in the oxygen-evolving complex (OEC) of photosystem II
(PSII) and the dinuclear site in manganese catalase1-3 and also
for their magnetic properties.2,4 This has led to a tremendous
amount of insight into the chemical and physical properties of
oxo-bridged manganese cluster compounds. NMR spectroscopy
is one technique that has not been applied extensively to Mn-
(III) coordination complexes because they are almost always
paramagnetic (S ) 2), which complicates acquisition and
interpretation of their NMR spectra. However, well-resolved,
isotropically shifted 1H NMR spectra of high-spin Mn(III)
We have been examining the reaction of O2 with Mn(II)
Schiff-base complexes Ia-c (R is an alkyl substituent) in
nonaqueous solution.15-18 The products isolated from these
complexes can be obtained.5-9 For example, high-quality H
1
NMR spectra were recently reported for some Mn(III) Schiff-
† Carnegie Mellon University.
‡ Duquesne University.
§ The University of Auckland.
| The University of Texas Health Science Center.
X Abstract published in AdVance ACS Abstracts, September 15, 1997.
(1) Wieghardt, K. Angew. Chem., Int. Ed. Engl. 1989, 28, 1153.
(2) Manchanda, R.; Brudvig, G. W.; Crabtree, R. H. Coord. Chem. ReV.
1995, 144, 1.
reactions include Mn(III) monomers, MnIV2 bis(µ-oxo) dimers,
(3) Larson, E. J.; Pecoraro, V. L. In Introduction to Manganese Enzymes;
Larson, E. J., Pecoraro, V. L., Eds.; VCH Publishers: New York,
1992; p 1.
(4) Wemple, M. W.; Tsai, H.-L.; Wang, S.; Claude, J. P.; Streib, W. E.;
Huffman, J. C.; Hendrickson, D. N.; Christou, G. Inorg. Chem. 1996,
35, 6437.
(5) Bertini, I.; Turano, P.; Vila, A. J. Chem. ReV. 1993, 93, 2833.
(6) La Mar, G. N. NMR of Paramagnetic Molecules; Principles and
Applications; Academic Press: New York, 1973.
(7) La Mar, G. C.; Walker, F. A. J. Am. Chem. Soc. 1973, 95, 6950.
(8) La Mar, G. C.; Walker, F. A. J. Am. Chem. Soc. 1975, 97, 5103.
(9) LaMar, G. N.; Horrocks, W. D., Jr.; Holm, R. H. NMR of Paramagnetic
Molecules; Academic Press: New York, 1973.
(10) Bonadies, J. A.; Kirk, M. L.; Lah, M. S.; Kessissoglou, D. P.; Hatfield,
W. E.; Pecoraro, V. L. Inorg. Chem. 1989, 28, 2037.
(11) Gohdes, J. W.; Armstrong, W. H. Inorg. Chem. 1992, 31, 368.
(12) Li, X.; Pecoraro, V. L. Inorg. Chem. 1989, 28, 3403.
(13) Caudle, M. T.; Riggs-Gelasco, P.; Gelasco, A. K.; Penner-Hahn, J.
E.; Pecoraro, V. L. Inorg. Chem. 1996, 35, 3577.
(14) Rogers, K. R.; Goff, H. M. J. Am. Chem. Soc. 1988, 110, 7049.
(15) Dailey, G. C.; Horwitz, C. P.; Lisek, C. A. Inorg. Chem. 1992, 31,
5325.
(16) Horwitz, C. P.; Ciringh, Y.; Liu, C.; Park, S. Inorg. Chem. 1993, 32,
5951.
(17) Horwitz, C. P.; Dailey, G. C. Comments Inorg. Chem. 1993, 14, 283.
S0020-1669(97)00747-7 CCC: $14.00 © 1997 American Chemical Society

Manganese(III) Schiff-Base Complexes
Inorganic Chemistry, Vol. 36, No. 22, 1997 4969
Table 1. Ligands Used in This Study
substituent
5,5′ 6,6′ 7,7′ 8,8′ 9.9′
Scheme 1
complexa
3,3′
4,4′
1
2
3
4
5
6
7
8
9b
H
CH3
H
H
(CH3)3C
CH3
H
(CH3)3C
C3H6O2
H
H
OCH3
H
H
H
H
H
H
H
H
H
OCH3
CH3
H
H
CH3
H
H
H
H
H
H
CH3
H
H
H
H
H
H
H
H
H
H
D
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
CH3
H
H
and MnIII µ-oxo dimers. The nature of the oxidized product
2
a The designators a, b, and c in the text refer to Mn(II) complexes,
Mn(III) monomer complexes, and µ-oxo-bridged Mn(III) dimers,
respectively. b This substituent is a polyether chain linking the 3 and
3′ positions like the one shown in Scheme 1.
depends both on the solvent used during the oxygenation19 and
on ligand backbone flexibility. Oxygenation in methanol yields
cationic Mn(III) monomers while oxygenation in CH3CN or
CH2Cl2 using the same Mn(II) complex produces a neutral oxo-
bridged dimer. Backbone flexibility also is important in
determining which dimer forms because the ligand twists into
ligand geometry. This was of interest to us because we felt it
might be useful to employ this information to distinguish and
characterize those Mn(III) compounds for which we have been
unable to obtain single-crystal X-ray crystallographic data.
Furthermore, it was anticipated that the results of the NMR study
would provide a means of determining if the oxo-bridged dimers
have significantly different ligand geometries than their corre-
sponding monomers, which might account for their unusual
electron transfer properties.
a cis-â geometry in the bis(µ-oxo) dimer.11,20 Thus, the MnIV
2
bis(µ-oxo) dimer is isolated when Mn(II) complexes with
flexible ligand backbones, Ia or Ib, are oxidized.
We recently reported that oxygenation of 3,3′-17-crown-6-
SAL-4-CH3-(OPHEN)MnII, II, in CH3CN produced a MnIII
2
µ-oxo dimer, III (Scheme 1).21 The crown ether groups on III
incorporate alkali and alkaline earth cations. The cation-binding
property of III is interesting because it allows one to begin
probing the role that Ca2+, an inorganic cofactor for the function
of the OEC,22 might play in the water oxidation process
catalyzed by PSII.23 Analogous chemistry has now been
The principal focus here is the characterization by paramag-
netic 1H NMR spectroscopy of the Mn(III) monomers and MnIII
2
µ-oxo dimers with SALOPHEN type ligands, IV (Table 1).
extended to the MnIV bis(µ-oxo) dimers.24,25 We found for
2
III that large anodic potential shifts occurred for the MnIII/II
2
couple upon cation incorporation and that the shifts were
consistent with an electrostatic effect. For example, Ba2+
incorporation into III shifted the MnIII/II2 couple anodically by
300 mV when compared to the cation-free complex.
Electrochemical experiments on III revealed that the two
metal centers were extremely weakly coupled despite being
connected by the bridging oxo ligand. This weak coupling was
manifested by the two Mn(III) centers having two reversible
one electron reduction processes at nearly the same potential.
Some compounds in this study were also examined by para-
magnetic 19F and 23Na spectroscopy. An application of the
standard inversion recovery technique26 was used to gather the
1H NMR data. The method described in the Experimental
Section provided a means of enhancing the paramagnetic signals
while decreasing those arising from residual hydrogens of the
deuterated solvents and any hydrogen-containing diamagnetic
impurities, primarily water, in the solvents. This allowed us to
observe all of the ligand hydrogens and also to probe axial
ligation properties of Mn(III). More importantly, the data
acquisition method should be applicable to other paramagnetic
transition metal complexes having electronic properties similar
to those found for Mn(III) and thus could provide a new method
for examining chemical and physical properties of these
molecules.
We show in this report that NMR spectroscopy is extremely
useful for characterizing and distinguishing Mn(III) monomers
from oxo-bridged dimers and that the planar ligand geometry
of the monomer likely is conserved in the dimer. In addition,
we found that multinuclear NMR spectroscopy permits one to
readily observe the binding of acetate, methanol, water, and
pyridine to Mn(III) centers in the monomers. Coordination of
water and pyridine to Mn(III) in the dimers also is observable.
Finally, paramagnetic 19F NMR data reveal that the hydrogen
bond between an axially bound water molecule and the PF6
counteranion observed in the solid state for one of the monomer
complexes is preserved in CD3OD.
In contrast, MnIII µ-oxo dimers having bridging groups such
2
as acetate connecting the Mn(III) centers in addition to the oxo
bridge, exhibit strongly coupled metal centers, and therefore
have distinct mixed-valence redox states.1 We have now
prepared and characterized several MnIII µ-oxo dimers with
2
SALOPHEN derivatives and have found that all show the same
electron transfer characteristics as III. Unfortunately, single
crystals suitable for X-ray crystallographic analysis have not
yet been obtained for any of the derivatives, so a structurally
based rationalization of the electron transfer properties is not
yet possible.
1
As noted above, the chemical shift pattern in the H NMR
spectrum of the hydrogens on Schiff-base ligands depends on
(18) Horwitz, C. P.; Winslow, P. J.; Warden, J. T.; Lisek, C. A. Inorg.
Chem. 1993, 32, 82.
(19) Gallo, E.; Solari, E.; Re, N.; Floriani, C.; Chiesi-Villa, A.; Rizzoli, C.
Angew. Chem., Int. Ed. Engl. 1996, 35, 1981.
(20) Larson, E. J.; Las, M. S.; Li, X.; A., B. J.; Pecoraro, V. L. Inorg.
Chem. 1992, 31, 373.
(21) Horwitz, C. P.; Ciringh, Y. Inorg. Chim. Acta 1994, 225, 191.
(22) Yocum, C. F. In The Calcium and Chloride Requirements for
Photosynthetic Water Oxidation; Yocum, C. F., Ed.; VCH Publishers,
Inc.: New York, 1992; p 71.
(23) Ghanotakis, D. F.; Yocum, C. F. Annu. ReV. Plant Physiol. Plant Mol.
Biol. 1990, 255, 41.
(24) Horwitz, C. P.; Warden, J. T.; Weintraub, J. T. Inorg. Chim. Acta
1996, 246, 311.
(25) Ciringh, Y.; Horwitz, C. P. Unpublished results.
(26) Harris, R. K. Nuclear Magnetic Resonance Spectroscopy; Longman
Scientific & Technical: Essex, 1986.

4970 Inorganic Chemistry, Vol. 36, No. 22, 1997
Ciringh et al.
Aldehyde Syntheses. Salicylaldehyde derivatives not commercially
available were prepared by the method of Casiraghi et al.,30 but the
purification procedures were modified as follows: 2-hydroxy-3-
methylbenzaldehyde was purified by vacuum distillation; 2-hydroxy-
3-tert-butyl-5-methylbenzaldehyde was purified by sublimation under
vacuum; 2-hydroxy-3,6-dimethylbenzaldehyde was purified by column
chromatography on silica gel using 80% hexanes/20% ether as eluant,
and the first band is the desired product. The 2-hydroxy-3-tert-butyl-
5-methylbenzaldehyde-d was prepared using CD2O and purified by
Experimental Section
Physical Measurements. 1H NMR spectra were measured at 300
MHz on an IBM NR/300 FT-NMR spectrometer. FTIR spectra (KBr
disks) were taken on a Mattson 5000 FTIR spectrometer while UV/vis
spectra were obtained using a Hewlett-Packard HP8452A spectropho-
tometer. Elemental analyses were performed by Midwest Microlab,
Indianapolis, IN. Electrochemical measurements were performed using
a Princeton Applied Research Model 273 potentiostat/galvanostat with
current voltage curves recorded on a Graphtec Model WX1200 X-Y
recorder. The supporting electrolyte was 0.1 M [n-Bu4N][PF6]
(TBAPF6, Fluka puriss). Cyclic voltammograms were performed under
N2 in a three-compartment cell using a glassy carbon disk working
electrode, a Pt wire counter electrode, and a sodium chloride saturated
calomel electrode (SSCE) reference electrode. Electrospray ionization
mass spectrometry (ESI-MS) was performed at University of Texas
Health Science Center at San Antonio. ESI mass spectra were acquired
on a Finnigan MAT SSQ700 mass spectrometer fitted with an Analytica
of Branford electrospray ionization interface. Electrospray voltages
of 2400-3400 V were utilized. Samples were dissolved in either CH3-
CN or CH2Cl2 at concentrations of approximately 10 pmol/µL and were
introduced into the ESI interface by direct infusion at a flow rate of 1
µL/min. Spectral averaging for 1-2 min was employed prior to data
acquisition.
1
vacuum sublimation. Product purities were determined by H NMR
spectroscopy.
Metal Complex Syntheses. (R,R′-SALOPHEN)MnII (R ) R′ )
H, 1a; 3-CH3, 2a; 4-CH3O, 3a; 5-CH3O, 4a; 3-t-Bu-5-CH3, 5a; 3,6-
(CH3)2, 6a). The same general procedures were followed for the
preparation of all of the Mn(II) complexes, so only the synthesis of 2a
is described. The solids 2-hydroxy-3-methylbenzaldehyde (556 mg, 2
mmol) and 1,2-diaminobenzene (108 mg, 1 mmol) were combined in
a Schlenk flask and then dissolved in methanol (20 mL), resulting in
the formation of a yellow solution, which was refluxed under N2 for
about 30 min. The solution was cooled to 20 °C, KOH (112 mg, 2
mmol) in deoxygenated methanol (10 mL) was added, and the solution
was stirred for 10 min. Mn(CH3C(O)O)2‚4H2O (245 mg, 1 mmol)
(MnCl2 can be used in place of the acetate) dissolved in deoxygenated
methanol was added to the solution of the deprotonated ligand, which
caused the solution to immediately become deep red; in most cases a
yellowish brown precipitate formed . The mixture was refluxed for 1
h to ensure complete reaction. After cooling to 20 °C, the solid was
recovered by filtration, washed with 20 mL of deoxygenated water,
and dried under vacuum.
Diamagnetic Suppression Routine for 1H NMR Data. An
inversion recovery pulse sequence, π-τ-π/2-acquire, was applied.
A series of spectra was obtained where the variable delay time, τ, was
chosen to be in the range of approximately 0.1-5 times the T1 value
for the paramagnetically shifted hydrogens (T1p). At short τ (τ1
∼
0.1T1p), the signals arising from the paramagnetic and the diamagnetic
(solvent and water) portions were inverted. Α longer τ, τ2 ) (3-5)-
T1p, was then chosen such that the paramagnetic signals were fully
relaxed, i.e., close to the maximum positive intensity, while the
diamagnetic signals remained inverted. It was necessary to choose τ2
- τ1 to be small compared to the T1 values for the diamagnetic protons
(T1d) so that the net change in the diamagnetic signal intensity was
minimal. Subtraction of the spectra at the chosen τ2 and τ1 values
then capitalized on the differential relaxation effects between the
paramagnetic and diamagnetic signals and resulted in an enhancement
of the paramagnetic signals and suppression of the diamagnetic ones.
Typical operating parameters were as follows: spectrometer frequency
(SF), 300.135 MHz; data block size (SI), 16K; resolution (RES), 3.59
Hz/pt; acqusition time (AQ), 0.279 s; and relaxation delay (RD), 0 s.
19F and 23Na NMR Data. 19F and 23Na NMR data were obtained
using parameters standard for the instrument (19F, SF 282.386 MHz,
SI 32K, RES 7.629 Hz/pt, and AQ 0.131 s; 23Na, SF 79.391 MHz, SI
8K, RES 20.345 Hz/pt, and AQ 0.030s), and in both cases the RD was
set to 0. Wide sweep widths were used in survey spectra so that signals
of reasonable intensity did not go unobserved.
Materials. The reagent grade chemicals 2-tert-butyl-4-methylphe-
nol, SnCl4, (H2CO)n, NaH, triethylene glycol ditosylate, 1,2-diami-
nobenzene, 2,3-dihydroxybenzaldehyde, 3,4-diaminotoluene, 2,6-
lutidine, 2-methylphenol, 2,4-dimethylphenol, 2-hydroxy-4-methoxy-
benzaldehyde, 2-hydroxy-5-methoxybenzaldehyde, Mn(CH3C(O)O)2‚
4H2O, and Ba(SO3CF3)2 (BaTf2) were used as received from Aldrich.
The [SO3CF3]- salts of Na+ (NaTf), K+ (KTf), and Ca2+ (CaTf2), as
well as [Cp2Fe][PF6], were prepared by methods described in the
literature.27,28 (D2CO)n (99% d2) CD2Cl2, CD3CN, C5D5N, and CD3-
OD were obtained from Cambridge Isotope Laboratories and used
without further purification. Cp2Co (Strem) was purified by sublimation
prior to use. The ligand 3,3′-17-crown-6-SALOPHEN-H2 was prepared
by the method of van Staveren et al.27 and its purity determined by 1H
NMR spectroscopy. The solvents CH3CN, CH2Cl2, DMF, and CH3-
OH were purified by standard methods29 while dry ether (Aldrich) was
used as received. Air sensitive compounds were prepared under
nitrogen using Schlenk techniques and stored in an Ar-filled glovebox
(Vacuum Atmospheres).
[(R,R′-SALOPHEN)MnIII]PF6 (R ) R′ ) H, [1b]PF6; 3-CH3, [2b]-
PF6; 4-CH3O, [3b]PF6; 5-CH3O, [4b]PF6; 3-t-Bu-5-CH3, [5b]PF6;
3,6-(CH3)2, [6b]PF6). The same procedures were used to prepare all
of these Mn(III) complexes, so only the synthesis of [2b]PF6 is
described. Compound 2a (50 mg, 0.13 mmol) was slurried in 10 mL
of O2-free CH3OH, and then O2 was bubbled through the solution for
about 15 min, producing a dark brown solution. A 4-fold molar excess
of NH4PF6 was added as a solid to the solution, and this mixture was
stirred until all of the solid was consumed and a dark brown
homogeneous solution remained. The methanol was removed in Vacuo,
and the solid product was washed with copious quantities of H2O. After
the solid was redissolved in dry CH3CN and the solution was dried
over Na2SO4, the solvent volume was reduced to about 5 mL, and a
solid precipitated on addition of dry ether. The solid was recovered
by filtration and dried under vacuum at approximately 100 °C for 10
h. Yield: 80-85%. Anal. Calcd (found): for C20H14N2O2MnPF6,
[1b]PF6, C 46.69 (46.97), H 2.75 (2.87), N 5.45 (5.77); C22H18N2 ‚O2-
MnPF6‚1/3CH3CN, [2b]PF6; C 48.17 (48.09), H 3.43 (3.60), N 5.42
(5.11); C22H18N2O2MnPF6, [3b]PF6, C 45.99 (45.86), H 3.16 (3.3), N
4.88 (5.05); C22H18N2O2MnPF6‚1/2H2O, [4b]PF6, C 45.28 (45.09), H
3.28 (3.31), N 4.8 (5.09); C30H34N2O2MnPF6‚1/2CH3CN ‚1/2H2O, [5b]-
PF6, C 54.41 (54.51), H 5.38 (5.66), N 5.12(5.42); C22H18N2 O2MnPF6‚1/
3H2O, [6b]PF6, C 49.99 (50.04), H 3.97 (4.22), N 4.86 (4.77). UV/vis
λ/nm (ꢀ × 103 M-1 cm-1): for [2b]PF6 (CH3OH) 246 (43.2), 306 (23.2),
350 (31.4), 444 (9.4); [3b]PF6 (CH3CN) 256 (32.3), 314(sh) (22), 352
(29.5), 394 (sh) (17.2), 432 (sh) (14); [5b]PF6 (CH3CN) 254 (31.3),
308 (20.5), 342 (21.4), 358 (22.5), 474 (6.1).
[(SAL-3,4-(CH3)2-OPHEN)MnIII]PF6 ([7b]PF6). The procedure
used to prepare this compound was the same as the one for the [(R,R′-
SALOPHEN)MnIII]PF6 complexes, except that 4,5-dimethyl-1,2-phen-
ylenediamine was used in place of 1,2-diaminobenzene. Anal. Calcd
(found) for C22H18N2O2MnPF6‚1/2CH3CN, [7b]PF6: C 49.06 (48.93),
H 3.49 (3.56), N 6.22 (6.08). UV/vis in CH3OH, λ/nm (ꢀ × 103 M-1
cm-1): 250 (40.5), 298 (21.6), 344 (26.5), 432 (10.9).
[(3,3′-(t-Bu)2-5,5′-(CH3)2-SALOPHEN-d2)MnIII]PF6 ([8b]PF6).
2-Hydroxy-3-tert-butyl-5-methylbenzaldehyde-d (192 mg, 1.0 mmol)
and 1,2-diaminobenzene (54 mg, 0.5 mmol) were dissolved in CH3-
OD (5 mL, 99% D), and the resulting yellow solution was refluxed
under N2 for about 30 min. After cooling to room temperature, KOD
in 1 mL of CH3OD (prepared by dissolving KOH (56.1 mg, 1.0 mmol)
(27) van Staveren, C. J.; van Eerden, J.; M van Veggel, F. C. J.; Harkema,
S.; Reinhoudt, D. N. J. Am. Chem. Soc. 1988, 110, 4994.
(28) Yang, E. S.; Chan, M. S.; Wahl, A. C. J. Phys. Chem. 1975, 79, 2049.
(29) Perrin, D. D.; Armarego, W. L. F. Purification of Laboratory
Chemicals, 3rd ed.; Pergamon Press: Oxford, England, 1988.
(30) Casiragi, G.; Casnati, G.; Puglia, G.; Startori, G.; Terenghi, G. J. Chem.
Soc., Perkin Trans 1 1980, 1862.

Manganese(III) Schiff-Base Complexes
Inorganic Chemistry, Vol. 36, No. 22, 1997 4971
in CH3OD) was added to the solution of the ligand, followed by stirring
for 10 min. Exchangeable hydrogens on Mn(CH3C(O)O)2‚4H2O were
exchanged with deuterium by dissolving the Mn(II) salt in CH3OD and
then removing the CH3OD in Vacuo. A CH3OD (1 mL) solution of
the deuterium-exchanged Mn(II) salt (122.5 mg, 0.5 mmol) was then
added to the solution containing the deprotonated ligand. The mixture
was refluxed for 1 h to ensure complete reaction and then cooled to
room temperature, and then O2 was bubbled through the solution. After
the reaction was judged to be complete as evidenced by a homogeneous
solution (30-60 min), a 4-fold molar excess of solid NH4PF6 was
added. Purification was accomplished by the procedure described above
for 5b. ESI-MS: [8b]+ m/z 511.4.
Table 2. Crystallographic Data for [9b]OAc‚BaTf2‚2H2O and
[6b]PF6‚H2O
[9b]OAc‚BaTf2‚2H2O
[6b]PF6‚H2O
formula
formula weight
C30H31BaF6MnN2O16S2 C20H24O3N2MnPF6
1045.97
588.37
cryst dimens (mm) 0.35 × 0.30 × 0.11
0.40 × 0.15 × 0.10
space group
a (Å)
P1h
P21/n
10.747(2)
12.271(8)
16.206(6)
106.10(4)
103.07(2)
106.06(3)
1864.0(14)
1.864
12.465(1)
13.366(2)
15.240(1)
b (Å)
c (Å)
R (deg)
â (deg)
γ (deg)
V (Å3)
103.051(7)
[(R,R′-SALOPHEN)MnIII]2(µ-oxo) (R ) R′ ) H, 1c; 3-CH3, 2c;
4-CH3O, 3c; 5-CH3O, 4c; 3-t-Bu-5-CH3, 5c; 3,6-(CH3)2, 6c). The
2473.5(4)
1.580
d
calc (g cm-3
)
same basic procedures were used to prepare all of the MnIII µ-oxo
2
Z
2
4
dimers, so only the one for 1c is described. Compound 1a (50 mg,
0.14 mmol) was loaded in a flask in the glovebox and dissolved in
CH2Cl2 (25 mL), and the flask sealed with a rubber septum. After the
flask was removed from the glovebox, O2 was allowed to flow over
the top of the solution slowly, causing the golden yellow slurry to
become a homogeneous red brown solution. The solvent volume was
reduced to 5 mL, and pentane was added to precipitate a brown powder.
The solid was recovered by filtration through a medium-porosity glass
frit and dried under vacuum. Anal. Calcd (found) for C40H28N4O5-
Mn2‚2CH2Cl2, 1c: C 58.71 (58.61), H 3.61 (3.39), N 6.68 (6.44). ESI-
MS: [1cH]+ m/z 755. Anal. Calcd (found) for C60H68N4O5Mn2‚3H2O,
5c: C 66.15 (66.26), H 6.85 (6.48), N 5.15 (4.85). ESI-MS: [5cH]+
m/z 1035.4. UV/vis (CH2Cl2) λ/nm (ꢀ × 103 M-1 cm-1): 306 (29.9),
344 (30.5), 360 (30.0), 486 (13.7).
radiation, λ (Å)
Mo KR, 0.710 69
-80
0.0352
Mo KR, 0.710 69
21
0.065
T (°C)
Ra
b
wR2
Rw
0.0878
c
0.052
2
a R1 ) Σ||Fo| - |Fc||/Σ|Fo|. b wR2 ) [Σ[w(Fo2 - Fc2)2/Σ[w(Fo )2]]1/2
.
2
c Rw )[(Σw(|Fo| - |Fc|)2/ΣwFo )]1/2
.
filtration and dried under vacuum. This crystalline material was used
for a single-crystal X-ray structural determination. Anal. Calcd (found)
for C28H27N2O8Mn‚Ba(CF3SO3)2: C 35.65 (35.11), H 2.69 (2.80), N
2.77 (2.66). ESI-MS: m/z 861 [[9b]OAc‚Ba(CF3SO3)]+, m/z 356 [[9b]-
OAc‚Ba]2+, m/z 951 [[9b]‚Ba(CF3SO3)2]+.
[(3,3′-(t-Bu)2-5,5′-(CH3)2-SALOPHEN-d2)MnIII]2(µ-O) (8c). The
Mn(II) starting material 8a was synthesized as described for compounds
1a-6a except that CH3OD was used during the Schiff-base condensa-
tion and metalation reactions. After isolation, 8a was oxidized and
isolated by a procedure analogous to the one described for the synthesis
of 5c. ESI-MS: [8cH]+ m/z 1039.4.
(3,3′-17-Crown-6-SALOPHEN)MnII‚CH3C(O)OH (9a‚HOAc). The
ligand H2-3,3′-17-crown-6-SALOPHEN (150 mg, 0.32 mmol) was
dissolved in dried, deaerated CH3OH (15 mL) and refluxed under N2
for 20 min to remove traces of O2 from the solution. After cooling to
20 °C, Mn(CH3C(O)O)2‚4H2O (80 mg, 0.32 mmol) dissolved in 10
mL of CH3OH was added to the ligand, producing a golden yellow
solution. The solution was refluxed under N2 for 30-45 min to assure
complete reaction. No precipitate appeared at 20 °C, so the solvent
was removed in Vacuo leaving a golden yellow solid, which was washed
with copious quantities of H2O. After drying, the solid was redissolved
in CH3CN the solution filtered, and then the solvent removed in Vacuo.
The golden yellow solid was dried at 50 °C for 2 h. Yield: 80-85%.
Anal. Calcd (found) for C26H24O6N2Mn‚CH3C(O)OH: C 58.44 (58.48),
H 4.9 (5.11), N 4.87 (5.34).
[(3,3′-17-Crown-6-SALOPHEN)MnIII]‚PF6 ([9b]PF6). This com-
pound was prepared by a method described for the analog with a
4-methyl-1,2-diaminobenzene backbone.21 Yield: 85-90%. Anal.
Calcd (found) for C26H24N2O6MnPF6‚3H2O: C 45.41(45.22), H 3.96
(3.73), N 4.08 (4.13). UV/vis (CH3CN) λ/nm (ꢀ x 103 M-1 cm-1):
246 (73.5), 314 (39.6), 370 (41.8), 474 (8.0).
[(3,3′-17-Crown-6-SALOPHEN)MnIII][CH3C(O)O] ([9b]OAc). A
sample of 9a‚HOAc (50 mg, 0.097 mmol) was loaded in a Schlenk
flask in an inert atmosphere glovebox and dissolved in deoxygenated
CH3CN (20 mL). The flask was removed from the glovebox, and then
O2 was slowly passed over the solution by means of a T-tube fitted
onto the Schlenk flask. Oxygenation was carried out for approximately
16 h, during which time the solution changed from golden yellow to
brown and a brown precipitate formed. The solid was recovered by
filtration through a medium-porosity glass frit, washed with CH3CN,
and dried under vacuum. Yield 80-85%.
Crystallographic Data Collection and Refinement of the Struc-
tures. (a) [9b]OAc‚BaTf2‚2H2O. Crystals of [9b]OAc‚BaTf2‚2H2O
were obtained by the slow diffusion of Et2O into an CH3CN solution
of [9b]OAc‚BaTf2 in air. X-ray diffraction measurements were made
on an Enraf-Nonius CAD-4 diffractometer with graphite-monochro-
mated Mo KR radiation. Cell constants, Table 2, and an orientation
matrix were obtained from a least-squares refinement by using the
setting angles of 25 carefully centered reflections in the range 19.5° <
2θ < 25.4°. The unit cell is triclinic (P1h; No. 2) with cell constants as
given in Table 2. Intensity data were collected at 193 ( 2 K using the
ω-2θ scan technique to a maximum 2θmax ) 54° for (h,+k,(l. Of
the 7434 reflections collected, 7120 were unique (Rint ) 0.021). Data
were corrected for Lorentz and polarization effects, and for absorption
by the empirical ψ scan method.31
The structure was solved by direct methods using SHELXS.32 The
non-hydrogen atoms were refined anisotropically. All hydrogen atoms
could be located in the difference electron density maps and were
refined satisfactorily with individual isotropic temperature factors. The
final cycles of full-matrix least-squares refinement on F2 were based
on 7119 reflections and 656 parameters and converged (largest
parameter shift was 0.021 times its esd) with R1 ) 0.0352 and wR2 )
0.0878 [I > 2σ(I)], Table 2. The refinement program employed was
SHELXL93.33 Atomic scattering factors for neutral atoms were taken
from International Tables for X-ray Crystallography.34
(b) [6b]PF6‚H2O. Crystals of [6b]PF6‚H2O were obtained by the
slow diffusion of Et2O into an CH3CN solution of [6b]PF6 in air. All
measurements were made on a Rigaku AFC7R diffractometer with
graphite-monochromated Mo KR radiation. Cell constants, Table 2,
and an orientation matrix for data collection, which was obtained from
a least-squares refinement using the setting angles of 25 carefully
centered reflections in the range 16.81° < 2θ < 22.37°, corresponded
to the monoclinic space group P21/n (No. 14). The data were collected
at 21 °C using the ω-2θ scan technique to 2θmax ) 45.0°. ω scans of
several intense reflections that were made prior to data collection had
an average width at half-height of 0.25°, with a takeoff angle of 6.0°.
Scans of (1.21 + 0.35 tanθ)° were made at 16.0°/min (in omega). Weak
[(3,3′-17-Crown-6-SALOPHEN)MnIII]‚OAc‚BaTf2 ([9b]‚OAc‚
BaTf2). A sample of [9b]OAc (50 mg, 0.048 mmol) was combined
with BaTf2 (42.6 mg, 0.096 mmol) in 10 mL of CH3CN, and the mixture
was stirred until all of the solid dissolved. A deep red solution formed.
The solvent volume was reduced to about 5 mL, and then Et2O was
allowed to slowly diffuse into the CH3CN solution, resulting in the
precipitation of a red brown crystalline solid, which was recovered by
(31) North, A. C.; Phillips, D. C.; Matthews, F. S. Acta Crystsllogr. 1968,
A24, 351.
(32) Sheldrick, G. M. Crystallographic Computing 3; Sheldrick, G. M.,
Ed.; Oxford University Press: Oxford, England, 1985, p 175.
(33) Sheldrick, G. M., SHELXL93.
(34) Ibers, J. A.; Hamilton, W. C. International Tables for X-ray Crystal-
lography; Kynoch Press: Birmingham, England, 1974; Vol. IV.

4972 Inorganic Chemistry, Vol. 36, No. 22, 1997
Ciringh et al.
reflections (I < 10.0σ(I)) were rescanned (four scans maximum), and
the counts were accumulated to ensure good counting statistics.
Stationary background counts were recorded on each side of the
reflection. The ratio of peak counting time to background counting
time was 2:1. Of the 3591 reflections collected, 3413 were unique
(Rint ) 0.054). The intensities of three representative reflections
measured after every 150 reflections. Over the course of data collection,
the standards increased by 0.6%. A linear correction factor was applied
to the data to account for this phenomenon. The data were corrected
for Lorentz and polarization effects.
The structure was solved by direct methods35 and expanded using
Fourier techniques.36 Non-hydrogen atoms were refined anisotropically.
Hydrogen atoms were included but not refined. The final cycle of
full-matrix least-squares refinement was based on 1624 observed
reflections (I > 2.00σ(I)) and 334 variable parameters and converged
(largest parameter shift was 0.01 times its esd) with R ) 0.065 and Rw
) 0.052, Table 2. The standard deviation of an observation of unit
weight was 2.76. The weighting scheme was based on counting
statistics and included a factor (F ) 0.004) to downweight the intense
reflections. The maximum and minimum peaks on the final difference
Fourier map corresponded to 0.54 and -0.38 e-/Å3, respectively.
Neutral atom scattering factors37 and anomalous dispersion terms were
included.38 All calculations were performed using teXsan39 from
Molecular Structure Corporation.
conditions used to prepare the corresponding Mn(III) monomer
from 9a‚HOAc, the acetate ion from the acetic acid solvate binds
to the Mn(III) center.
Syntheses of Mn(III) Monomers. Two principal methods
were used to prepare the Mn(III) monomers. In the first, a Mn-
(II) complex was oxidized in CH3OH by bubbling O2 through
the solution (this method is referred to as the CH3OH/O2
method; eq 1). The resulting Mn(III) product, which was brown
for [1b]+, [2b]+, [4b]+, [5b]+, and [6b]+ and dark blue for [3b]+
and [7b]+, was isolated as a PF6- salt. Paramagnetic 1H NMR
results, presented below, revealed that complexes prepared in
this manner likely have CH3OH as an axial ligand. The second
method involved oxidation of a Mn(II) complex under N2 or
Ar with [Cp2Fe]+ in CH3CN (CH3CN/[Cp2Fe]+ method; eq 2)
followed by isolation as the PF6- salt. These compounds appear
to have CH3CN as the axial ligand. Compounds prepared by
either synthetic method and “dried” in Vacuo at 60-80 °C
exhibited identical combustion analytical data (<0.4% error for
each element analyzed), indicating loss of the axial ligand during
the drying process. We refer to compounds dried at 60-80 °C
as “dried monomers” while those only briefly exposed to
vacuum at 20 °C are designated as “solvated monomers”. A
dried monomer is less soluble in CH3CN than a solvated one,
but both are freely soluble in CH3OH. The 1H NMR spectrum
of a dried monomer in CD3CN (presented below) is suggestive
of loss of the axial CH3OH or CH3CN with an accompanying
dimerization or oligomerization of the monomers for a portion
of the sample.
Results and Discussion
Syntheses of Mn(II) Monomers. The ligands used in this
study (Table 1) were prepared by condensation of 1,2-phen-
ylenediamine or 4,5-dimethyl-1,2-phenylenediamine with the
appropriate aldehyde in CH3OH under an inert atmosphere. In
general, the ligands were not isolated but were metalated in
situ using either Mn(CH3C(O)O)2‚4H2O or MnCl2. KOH in
CH3OH was added prior to the Mn(II) treatment to facilitate
the metalation process. Most of the Mn(II) complexes, which
are yellow-brown to brown except for 9a, which is orange,
precipitated from the CH3OH solution. The solids were
recovered by filtration, washed with copious quantities of
deoxygenated water to remove residual salts, and then dried in
Vacuo (80-100 °C) for 3-5 h. For those Mn(II) compounds
that were soluble in CH3OH, the solvent was stripped off in
Vacuo and the solid washed with deoxygenated water and then
dried in Vacuo. Acceptable combustion analytical data were
obtained for all of the Mn(II) complexes.
Deprotonation of the ligand by KOH was not employed
during the synthesis of 9a to avoid incorporation of K+ in the
crown ether. As a consequence, the product precipitated with
one acetic acid molecule per Mn(II) even though it was washed
with water and dried under vacuum for an extended period of
time. This compound is referred to as 9a‚HOAc. The presence
of the acetate ion was confirmed by synthesizing the deuterated
analog of 9a‚HOAc using Mn(CD3C(O)O)2, which was prepared
by refluxing Mn(CH3C(O)O)2 in CD3C(O)OD overnight fol-
lowed by removal of the acid under vacuum, and then drying
the solid for 3 h at 80 °C under vacuum. The IR spectrum of
9a‚CD3C(O)OD displayed absorption bands in the 2265 and
2105 cm-1 regions from C-D stretching vibrations. The
corresponding bands in 9a‚HOAc are masked by other ligand
C-H vibrations. As discussed below, under some synthetic
(CH3OH)(L)MnII + O2 CH OH8
3
∆
[(CH3OH)x(L)MnIII]+ vacuum8 [(L)MnIII]+ (1)
(CH3CN)(L)MnII + Cp2Fe+ CH CN8
3
∆
[(CH3OH)x(L)MnIII]+ vacuum8 [(L)MnIII]+ (2)
L ) R,R′-SALOPHEN; x ) 1 or 2
Oxygenation of 9a‚HOAc in CH3OH followed by the addition
of NH4PF6 produced the Mn(III) monomer, [9b]PF6. In
contrast, oxygenation of 9a‚HOAc in CH3CN produced [9b]-
OAc as a red-brown precipitate. The low solubility of [9b]-
OAc in CH3CN may be due to formation of a polymeric material
in which acetate groups bridge the Mn(III) centers.10,40-42 As
discussed below, oxygenation of all of the other Mn(II)
complexes in CH3CN resulted in the formation of MnIII2 µ-oxo
dimers. Apparently the acetic acid present in 9a‚HOAc either
intercepted intermediates during µ-oxo dimer formation or
cleaved the dimer after it formed.
The addition of Na+, K+, Ca2+, or Ba2+ ([SO3CF3]- salts)
to a slurry of [9b]OAc in CH3CN resulted in the formation of
a homogeneous red-brown solution within approximately 5-10
min. Single crystals of [9b]OAc containing Ba2+ were isolated
from the slow diffusion of Et2O into a CH3CN solution of the
compound in air. The result of the X-ray analysis is shown in
Figure 1a. Crystallographic data for [9b]OAc‚BaTf2‚2H2O are
given in Table 2, and selected bond distances and angles are
(35) Altomare, A.; Cascarano, M.; Giacovazzo, C.; Guagliardi, A.; Burla,
M. C.; Polidori, G.; Camalli, M. J. Appl. Crystallogr. 1994, 27, 435.
(36) Beurskens, P. T.; Admiraal, G.; Beurskens, G.; Bosman, W. P.; Garcia-
Granda, S.; Gould, R. O.; Smits, J. M. M.; Smykalla, C. The DIRDIF
program system. Technical Report of the Crystallography Laboratory,
University of Nijmegen: Nijmegen, The Netherlands, 1992.
(37) Cromer, D. T.; Waber, J. T. International Tables for X-ray Crystal-
lography; Kynoch Press: Birmingham, England, 1974; Vol. IV, Table
2.2 A.
(40) Aurangazeb, N.; Hulme, C. E.; McAuliffe, A. A.; Pritchard, R. G.;
Watkinson, M.; Bermejo, M. R.; Sousa, A. J. Chem. Soc., Chem.
Commun. 1994, 2193.
(41) Aurangazeb, N.; Hulme, C. E.; McAuliffe, A. A.; Pritchard, R. G.;
Watkinson, M.; Garcia-Deibe, A.; Bermejo, M. R.; Sousa, A. J. Chem.
Soc., Chem. Commun. 1992, 1524.
(38) Ibers, J. A.; Hamilton, W. C. Acta Crystallogr. 1964, 17, 781.
(39) teXsan, Crystal Structure Analysis Package.
(42) Davies, J. E.; Gatehouse, B. M.; Murray, K. S. J. Chem. Soc., Dalton
Trans. 1973, 2523.

Manganese(III) Schiff-Base Complexes
Inorganic Chemistry, Vol. 36, No. 22, 1997 4973
Table 3. Selected Bond Lengths (Å) and Angles (deg) for
[9b]OAc‚BaTf2‚2H2O
a
Ba1-O16
Ba1-O8
Ba1-O15
Ba1-O1
Ba1-O5
Ba1-O3
Ba1-O2
Ba1-O4
Ba1-O12
Ba1-O6
Ba1-Mn1
2.691(4)
2.732(3)
2.767(4)
2.786(3)
2.791(3)
2.817(3)
2.863(3)
2.863(3)
2.884(3)
2.917(3)
3.741(2)
Mn1-O1
Mn1-O2
Mn1-N1
Mn1-N2
Mn1-O7
Mn1-O9
S1-O9
S2-O12
O7-C27
O8-C27
1.867(3)
1.877(3)
1.969(3)
1.978(3)
2.107(3)
2.489(4)
1.435(3)
1.445(3)
1.268(5)
1.234(5)
O1-Mn1-O2
O1-Mn1-N1
O2-Mn1-N1
O1-Mn1-N2
O2-Mn1-N2
N1-Mn1-N2
O1-Mn1-O7
O2-Mn1-O7
N1-Mn1-O7
N2-Mn1-O7
90.71(12)
92.59(13)
175.21(13)
168.58(13)
93.18(13)
82.94(13)
100.27(12)
93.85(13)
88.99(13)
90.19(12)
O1-Mn1-O9
O2-Mn1-O9
N1-Mn1-O9
N2-Mn1-O9
O7-Mn1-O9
C27-O7-Mn1
C27-O8-Ba1
S1-O9-Mn1
S2-O12-Ba1
87.42(12)
94.98(13)
81.70(13)
81.54(12)
168.21(11)
128.2(3)
140.9(3)
150.9(2)
159.2(2)
b
which is described below, reveals that the Mn(III) center is
ligated by acetate in CD3OD. The Mn-N and Mn-O distances
to the Schiff-base ligand are similar to those found in other
Schiff-base complexes.42-47 Not unexpectedly, the Mn-O
acetate distance, Mn1-O7 2.107(3) Å, is significantly shorter
than the Mn-O triflate length, Mn1-O9 2.489(4) Å. The O7-
Mn1-O9 angle is 168.21(11)°.
The coordination environment around the Ba2+ is novel
(Figure 1b,c). The cation is coordinated by 10 oxygens: four
from the crown ether, two from the H2O molecules, two from
the Schiff-base phenolate groups, one from the triflate group
not bound to manganese, and one from the acetate. Except for
the Ba1-O16 distance (2.691(4) Å), the Ba2+-O distances
(2.767(4)-2.917(3) Å) are within the range predicted48 and
found49 for barium in a crown ether type environment. There
is no obvious reason for the Ba1-O16 distance to be signifi-
cantly shorter than the Ba1-O15 distance (2.767(4) Å) since
both oxygens belong to H2O molecules bound to Ba2+ and
neither is involved in an H-bonding interaction with a neighbor-
ing complex. The arrangement of the ligating oxygens is best
described as possessing C3V local symmetry, with one oxygen
(O15) (and the Ba2+) on the C3 axis, six oxygens in a hexagon
above the Ba2+, and three oxygens below the Ba2+ in a staggered
arrangement with respect to the hexagon. The coordination
polyhedron, thus, consists of a hexagonal pyramid with its six
triangular faces at the “top”, and a 3-fold alternation of triangular
and trapezoidal faces downward to a triangle at the bottom. A
more graphical description is a hexagonal pyramid on a
triangular base. This 1:6:3 coordination geometry has not been
seen previously. Other workers have categorized 10-coordina-
tion environments and, by energy minimization calculations,
have deduced that six structures are possible.50-52 They have
c
(43) Ashmawy, F. M.; Beagly, B.; McAuliffe, C. A.; Parish, R. V.;
Pritchard, R. G. J. Chem. Soc., Chem. Commun. 1990, 936.
(44) Gohdes, J. W.; Armstrong, W. H. Inorg. Chem. 1988, 27, 1841.
(45) Matsumoto, N.; Takemoto, N.; Ohyosi, A.; Okawa, H. Bull. Chem.
Soc. Jpn. 1988, 61, 2984.
(46) Pecoraro, V. L.; Butler, W. M. Acta Crystallogr. 1986, C42, 1151.
(47) Horwitz, C. P.; Dailey, G. C.; Tham, F. S. Acta Crystallogr. 1995,
C51, 815.
Figure 1. (a) ORTEP diagram and atom-labeling scheme of [9b]OAc‚
BaTf2‚2H2O with hydrogen atoms omitted and thermal ellipsoids at
the 50% probability level (in this view O11 obscures Mn1). (b) View
of the 1:6:3 coordination at Ba2+. (c) View down the local C3 axis at
Ba2+
.
found in Table 3. Mn1 adopts a pseudo-octahedral geometry
with the equatorial plane consisting of two imine nitrogens and
two phenolate oxygens of the Schiff-base ligand. An oxygen
from one triflate group and an oxygen from the acetate are the
axial ligands. The 1H NMR spectrum of [9b]OAc‚BaTf2‚2H2O,
(48) Hay, B. P.; Rustad, J. R. J. Am. Chem. Soc. 1994, 116, 6316.
(49) Costes, J.-P.; Laurent, J.-P.; Chabert, P.; Commenges, G.; Dahan, F.
Inorg. Chem. 1997, 36, 656.
(50) Favas, M. C.; Kepert, D. L. Prog. Inorg. Chem. 1981, 28, 309.
(51) Robertson, B. E. Inorg. Chem. 1977, 16, 2735.
(52) Lin, Y. C.; Williams, D. E. Can. J. Chem. 1972, 51, 312.

4974 Inorganic Chemistry, Vol. 36, No. 22, 1997
Ciringh et al.
found that the bicapped square antiprism, the tetracapped
trigonal prism, and the tetradecahedron are more stable than
the pentagonal antiprism, the pentagonal prism, and the bicapped
square prism. Thus, the coordination geometry observed here
is not one of the predicted ones.
Synthesis of MnIII2 µ-Oxo Dimers. The MnIII2 µ-oxo dimers
were prepared by oxygenation of the corresponding Mn(II)
complexes in CH3CN or CH2Cl2 (eq 3). We found that it was
important to carry out the oxygenation slowly (24-72 h)
because, otherwise, the sample could be contaminated with other
oxo-bridged dimers. This slow oxidation was achieved by
passing O2 over an unstirred solution of the monomer. The 1H
NMR data for these compounds, described below, indicate that
water is an axial ligand for the dimers; the water likely comes
from trace water present in the solvents used during oxygenation.
CH3CN or CH2Cl2
(L)(L′)MnII + O2
8 [(H2O)(L)MnIII]2(µ-O) (3)
trace H2O
Figure 2. 1H NMR spectra of a solvated sample of compound [1b]+
in CD3OD: (a) before and (b) after application of the diamagnetic
suppression routine.
L′ is CH3CN or CH3OH; L is a Schiff-base ligand
Paramagnetic NMR Spectroscopy. The observation of
signals from paramagnetic molecules by 1H NMR spectroscopy
is not as straightforward as for diamagnetic species, but
paramagnetic NMR spectroscopy is becoming an important tool
for characterizing a number of transition metal complexes.5,10,53
We found that, in order to obtain high-quality spectra of the
Mn(III) Schiff-base complexes studied here, it was key that the
intensity of the resonances arising from residual protons in the
pared to the remainder of the spectrum. The signal intensity
of the upfield resonance has been enhanced 10-fold relative to
the other resonances in Figure 2b and smoothed for purposes
of clarity. The six resonances lying outside the diamagnetic
region (0-14 ppm) are observed in both spectra, but the
resonance at 1.8 ppm is clearly visible only after application of
the diamagnetic suppression routine. The resonance at 1.8 ppm
is masked in Figure 2a by the resonance arising from CH3CN
that was introduced during the synthesis of [1b]+ (see below).
When pyridine-d5 (py-d) was used as the solvent, the resonance
at 1.8 ppm was still observed, so this signal is not an artifact of
the spectral subtraction routine.
1
deuterated solvents as well as any H NMR active impurities
in the solvents, typically water, be diminished. These kinds of
resonances are referred to here as diamagnetic signals. An
application of the inversion recovery pulse sequence was used
to perform the diamagnetic suppression. Specific experimental
parameters used for collecting the data are given in the
Experimental Section. Typical suppression factors for the
diamagnetic signals were 5:1 to 10:1 while paramagnetic signals
were enhanced by approximately 2:1. Although the overall
effect of the suppression routine may appear to be fairly modest
compared to more sophisticated pulse sequences, the experiment
is easily performed and instrument setup parameters are
straightforward. We have found that even paramagnetic peaks
that are overlapped by diamagnetic peaks can be resolved by
this approach because of the inherently different relaxation rates
of the paramagnetic and diamagnetic resonances.
It has been shown for Mn(III) Schiff-base complexes that
the contact term which occurs through the phenol oxygens and
imine groups accounts for most of the chemical shift.54 The
alternation in upfield/downfield resonances that one might
expect as the aromatic rings of the ligand arms are traversed is
not observed in the R,R′-SALOPHEN complexes because of
the competing chemical shift determining pathways. However,
since there is only one delocalization mechanism for the
aromatic backbone, the resonance for either 8H or 9H should
be in the upfield region and the other in the downfield region.
It was possible to determine the chemical shift assignments
for all of the resonances observed in [1b]+ Via a systematic
substitution of the ligand hydrogens (Table 1). The matrix
generated by the substitutions (Figure S1, Supporting Informa-
tion) was used in providing the assignments indicated on Figure
2b. Given in Table 4 are the measured chemical shifts for the
hydrogens on all of the complexes; T1p values for some
complexes are also included. The imine hydrogens (7H) of
[8b]+ were selectively deuterated in order to show that 7H gave
rise to the most upfield resonance. Synthesis of the deuterated
ligand was straightforward as (CD2O)n (99% d) could be used
in place of (CH2O)n in the aldehyde synthesis (see Experimental
Section). It was necessary to perform the Schiff-base condensa-
tion and metalation reactions in CH3OD to ensure that no loss
of the deuterium label occurred. The compound having a CH3
group on the 7-position also was prepared, but that substitution
appeared to distort the ligand from planarity, thus affecting all
of the resonance positions in this molecule.
Sample preparation was important for achieving high-quality
spectra. High concentrations of complex were used where
possible (approximately 20 mg/mL of solvent) in order to
decrease the total data acquisition time. Typically, monomers
were dissolved in CH3OD or CD3OD prior to spectral acquisi-
tion to exchange CH3OH, CH3CN, or H2O, the CH3OD or CD3-
OD was removed in Vacuo, and then the sample was dissolved
in the solvent of choice. This procedure could not be applied
to the MnIII µ-oxo dimers because they decomposed in
2
methanol. The dimers were dried at approximately 50 °C in
Vacuo prior to dissolution.
1
[LMnIII]+ Monomers. The H NMR spectra obtained for
[1b]+ as a solvated monomer in CD3OD, before and after
application of the diamagnetic suppression routine, are shown
in Figure 2 traces a and b, respectively. A total of seven
resonances are expected for [1b]+ assuming C2V symmetry, and
seven are observed (Figure 2b). The resonance at -387 ppm
is extremely broad and weak. Clear observation of the upfield
resonance required a 10-fold longer acquisition time as com-
We attempted to determine the T1p value for 7H in [1b]+,
but its extremely rapid relaxation rate only allowed us to
(53) Bertini, I.; Luchinat, C. NMR of Paramagnetic Molecules in Biological
Systems; B. Cummings: Menlo Park, CA, 1986.
(54) Bonadies, J. A.; Maroney, M. J.; Pecoraro, V. L. Inorg. Chem. 1989,
28, 2044.

Manganese(III) Schiff-Base Complexes
Inorganic Chemistry, Vol. 36, No. 22, 1997 4975
1
Table 4. Paramagnetic H NMR Chemical Shifts, T1 Values, and Electrochemical Data
δ/ppm (T1/ms)
substituent:
δ/ppm
CH3 of
bound NMR
complex
3H
4H
5H
6H
7H
8H
9H
bound OAc- H2O solvent E°/mVa
CD3OD -20
[1b]+
[2b]+
-4.2
-28.9
(2.6)
-34.5
(2.58)
-23.0
(3.1)
-22.3
(3.73)
-17.5
1.8 -398 25.0
-17.6
(3.1)
(0.43)
(0.8) (<0.4) (0.65)
1.6
b
24.7
-17.6
(3.37)
-16.4
-18.0
-17.7
3-CH3: 19.5
CD3OD -36
[3b]+
[4b]+
[5b]+
-10.6
-1.9
2.3
1.3
b
b
26.1
24.0
4-CH3O: 7.6
5-CH3O: 5.3
3-t-Bu: 25.2
5-CH3: 6.3
3-CH3: 18.4
6-CH3: 5.9
3-CH3: 21.8
6-CH3: 6.8
3-CH3: 22.9
6-CH3: 7.1
3-CH3: 18.4
6-CH3: 5.9
9-CH3: 6.3
3-t-Bu: 6.3
5-CH3: 25.2
7-D
CD3OD -65
CD3OD -78
CD3OD -165
-21.8
-29.6
2.8 -387 26.6
[6b]+
-38.9
-38.9
-39.0
-39.4
-20.2
-23.1
-24.2
-20.3
b
b
b
b
b
25.2
26.0
25.9
25.9
-16.6
-17.0
-17.1
-16.6
CD3OD -105
12.7c CD3CN
CD3CN
[6b]+‚CH3OH
[6b]+‚CH3CN
[6b]PF6‚H2O
2.1 CD3OD
[7b]+
[8b]+
-4.7
-27.4
-38.9
-23.6
-20.2
2.8
25.0
25.2
CD3OD -74
CD3OD
-16.6
[9b]PF6
-34.7
(0.8)
-33.2
-33.7
(2.5)
-29.1
-30.0
(1.6)
-25.2
-26.9
(2.1)
-25.7
-27.0
(2.3)
-28.5
-35.1
-24.3
(2.3)
-23.6
-24.6
(0.7)
-19.7
-21.8
(2.0)
-17.7
-21.7
(2.5)
-15.1
-20.1
(2.3)
-22.9
-26.1
-16.2
-1.5
(0.9)
-3.0
b
b
23.9
(0.5)
25.2
25.2
(2.2)
25.9
26.1
(0.4)
26.6
25.6
(0.5)
26.0
25.5
(0.6)
-19.4
(1.9)
-17.9
-19.1
CD3CN -17
[9b]+ ‚Na+ d
[9b]+‚K+ d
18
15
CD3CN
CD3CN
-1.83 b
19
(2.1)
(0.5)
b
b
[9b]+‚Ca2+ d
[9b]+‚ Ba2+ d
-7.4
-3.3
(0.9)
-2.0
7.2
(0.7)
-5.7
-1.4
(0.5)
-13.9 -354 23.6
-11.1 -348 25.2
2.5
16.7
14
-17.5
-19.3
(1.7)
-13.8
-15.0
(2.2)
-15.1
-16.9
(2.1)
28
17
CD3CN
CD3CN
215
[9b]OAc‚Na+ e
[9b]OAc‚ K+ e
b
b
39.1
41.3
CD3CN
CD3CN
25
0
[9b]OAc‚Ca2+ e
b
b
33.5
36.1
(0.7)
CD3CN
CD3CN
240
200
[9b]OAc‚ BaTf2‚2H2Of
1c
1c
2c
3c
5c
31.6
40.5
23.9
11.7
-21.1
-23.5
-14.3
-7.2
44.7 C5D5Ng
51.4 CD2Cl2 -130h
b
b
24.7
30
3-CH3
43.8 C5D5Ng
-16
-21.2
4-CH3O: 7.5
3-t-Bu: 34
5-CH3: 8.6
3-CH3; 6-CH3
48.4 CD2Cl2 -215
45.9 CD2Cl2 -270h
-345 29.2
-9.2
6c
-25
-20.3
b
24
-15.4
44
C5D5Ng
a 0.1 M (n-Bu4N)PF6/CH3CN; E Vs SSCE. b Not measured. c CH3 of bound CH3OH. d Prepared in situ by the reaction of [9b]PF6 with Mn+
-
-
(SO3CF3 salt). e Prepared in situ by the reaction of [9b]OAc with Mn+ (SO3CF3 salt). f Sample used for crystallographic analysis. g Spectrum
obtained within 1 h of dissolution in py. h 0.1 M (n-Bu4N)PF6/CH2Cl2; E Vs SSCE.
estimate that T1p was less than 0.4 ms. This value corresponds
to the most rapid T1p measured for the Mn(III) monomer
compounds in this study. Since 7H is in close proximity to the
Mn(III) center and a conjugation pathway exists through the
sp2-hybridized imine group, both dipolar and scalar coupling
mechanisms likely serve to very efficiently relax 7H. Owing
to the long acquisition times required to observe 7H, no attempt
was made to observe this resonance in all of the other
compounds prepared in this study.
the new resonances. Anion coordination to Mn(III) does not
occur, as evidenced by the fact that vibrational bands of the
anions in the FT-IR spectra of solvated and dried monomers
are the same. As noted above, combustion analytical data reveal
that a dried monomer loses its axial ligands. In order for Mn-
The spectrum of [1b]+ as a dried monomer dissolved in CD3-
CN is shown in Figure 3. In CD3OD, the spectrum is identical
to the one in Figure 2b. It is evident that, even though there
are more resonances for the dried monomer than for the solvated
monomer, the intense resonances in the spectrum have the same
chemical shifts as those found for the solvated monomer.
Similar increases in the number of resonances were recorded
for all of the other dried monomers. As discussed below, the
chemical shifts for the ligand hydrogens on monomers having
CH3OH, CH3CN, or H2O as axial ligands are not strongly
influenced by the type of axial ligand. Therefore, the “new”
resonances are not due to simply changing axial ligation.
It is likely that the new resonances arise from formation of
dimers, V, of the Mn(III) monomers: trimers, tetramers, and
other oligomers also are viable species that could give rise to
(III) to satisfy its requirement for axial ligands, a likely scenario
is that a phenol oxygen from a nearby Mn(III) complex occupies
an axial site. This type of interaction is observed in Cu, V,
and Fe55-57 Schiff-base chemistry, and we report here an
example of such a dimer pair in the solid state (see Figure 8b).
It appears that CD3CN is not as strongly coordinating as CD3-
(55) Hall, D.; Waters, T. N. J. Chem. Soc. 1960, 2644.
(56) Gerloch, M.; Lewis, J.; Mabbs, F. E.; Richards, A. Nature 1966, 809.
(57) Yamamoto, K.; Oyaizu, K.; Tsudhida, E. J. Am. Chem. Soc. 1996,
118, 12665.

4976 Inorganic Chemistry, Vol. 36, No. 22, 1997
Ciringh et al.
these are assigned to the -CH2- groups of the crown ether.
Dried samples of [9b]PF6 showed an increase in the number of
resonances like the dried form of [1b]+, but no obvious changes
occurred for the crown ether portion of the molecule.
Cation-containing complexes of [9b]PF6 were prepared in situ
for 1H NMR measurements by adding 1 equiv of a salt such as
KTf to [9b]PF6 in CD3CN (Figure 4b and Table 4). Cation
incorporation into [9b]+ caused the resonance at 8 ppm to
sharpen and shift upfield to 1.6 ppm. Similar effects occurred
for the other crown ether resonance upon cation incorporation,
but the upfield shift was from 2.1 to 1.9 ppm. All other signals
from the ligand showed little change in their chemical shifts.
The T1p values for most of the hydrogens of the salt-containing
complexes, [9b]+‚Mn+, are close to those measured for [9b]-
PF6 itself with the notable exception being 4H; because 4H is
closest to the crown ether, its environment appears to be the
most affected by the cation and any solvent molecules that might
be bound to it. A new resonance from a water molecule or
hydroxide ion bound to Mn(III) also appeared at 15 ppm
following cation incorporation (see below). Since only the two
resonances associated with the crown ether change significantly
upon cation incorporation, this is strong evidence for cation
insertion into the crown ether cavity. This same conclusion
was deduced from electrochemical measurements on these
(Table 4) and related compounds.21
Figure 3. 1H NMR spectrum of a dried sample of compound [1b]+ in
CD3CN with diamagnetic suppression applied. Arrows indicate tentative
resonance assignments for the dried monomer as derived from the
solvated monomer spectrum. The resonance marked 3H is a composite
of 3H and 9H in the dried monomer sample. An asterisk (*) indicates
an impurity in the sample or solvent.
We believe that the broad resonance at 15 ppm in Figure 4b
arises from H2O or OH- bound to Mn(III) for the following
reasons: (i) the resonance was present only when the compound
was prepared using KTf that had not been rigorously dried;
similar behavior was observed for the other salt-containing
compounds (Table 4); (ii) this signal was lost upon the addition
of D2O; (iii) the chemical shift of the resonance was sensitive
to the amount of water in solution; it shifted upfield, i.e., in the
direction of free water, as the water concentration increased;
and (iv) the very rapid T1p value (∼1 ms) supports binding to
the Mn(III) center rather than to the cation in the crown ether.
It is not possible to distinguish H2O from OH- with our current
experiments; however, since a bound OH- would reduce the
overall charge on the complex following cation incorporation,
at this time we favor hydroxide binding to the Mn(III) center.
The propensity of the [9b]+‚Mn+ type complexes to bind
water to Mn(III) suggests that the presence of the Mn+ cation
increases the Lewis acidity of the Mn(III) center and enhances
its water/hydroxide binding properties. While monomers like
[9b]+‚Mn+ are structurally unrelated to PSII, the enhanced
affinity of Mn(III) for water that results from having a cation
within 3.6 Å of the metal center27 leads us to suggest that the
Ca2+ cation, which might be as close as 3.3 Å to the Mn4 core
of the OEC, may fill a similar role in the functioning enzyme.
The 23Na NMR spectrum of [9b]+‚Na+ provided interesting
information regarding cation insertion into the crown ether. As
a reference we assigned NaTf in CD3CN to 0 ppm. The
chemical shift for NaTf in 18-crown-6/CD3CN is -8.3 ppm,
shifting slightly upfield upon the addition of water. The
spectrum of a dried monomer of [9b]+‚Na+ in CH3CN has two
relatively broad resonances: an intense signal at 12 ppm, which
is assigned to Na+ in the crown ether of the monomer, and a
smaller one at 421 ppm, which is assigned to Na+ in the crown
ether portion of the dimer/oligomer (Figure 5a). The 421 ppm
resonance is lost upon the addition of water or methanol to the
CD3CN solution (Figure 5b); recall that methanol can be used
to dissociate other dried monomers. Since both resonances are
in different spectral regions than “free” Na+ or Na+ in
18-crown-6 in CD3CN, this supports the binding of sodium by
the crown ether in [9b]+. Furthermore, the broadness of the
Figure 4. 1H NMR spectra of a solvated sample of compound [9b]PF6
in CD3CN with diamagnetic suppression applied (a) before and (b) after
the addition of 1 molar equiv of hydrated KSO3CF3. An asterisk (*)
indicates an impurity in the sample or solvent.
OD and does not break up all of the dimers/oligomers formed
during the drying process. The dimerization/oligomerization
may occur for all or just a fraction of the complexes in a given
sample; we do observe that the relative ratio of dried monomer
to solvated monomer resonances depends on the length of the
drying process and the ligand substitution. If the dimer/oligomer
hypothesis is correct, then the close proximity of one or more
S ) 2 Mn(III) centers to the ligand hydrogens on a nearby Mn-
(III) complex could account for the shifts observed in the
resonance positions. Tentative assignments for the new reso-
nances are given in Figure 3. However, as some of the
intensities were extremely weak, particularly those for 3H and
6H in [1b]+, it is difficult to unambiguously assign all of the
resonances.
Crown Ether Complex. The chemical shift assignments in
the spectrum of [(3,3′-17-crown-6-SALOPHEN)MnIII]PF6,
[9b]PF6, taken in CD3CN are based on the results from the
solvated [(R,R′-SALOPHEN)MnIII]PF6 monomers discussed
above as the pattern of three far-upfield resonances and one
downfield resonance observed for [1b]+ is preserved in [9b]-
PF6 (Figure 4a and Table 4) (we did not attempt to locate 7H
for [1b]+ or its salt-containing derivatives). Two signals which
are broad compared to solvent peaks but sharp compared to
the paramagnetic signals appear in the diamagnetic region, and

Manganese(III) Schiff-Base Complexes
Inorganic Chemistry, Vol. 36, No. 22, 1997 4977
binding/bridging is, in turn, likely to influence both the contact
and dipolar shifts arising from the Mn(III) center and thus the
chemical shift of the methyl group. We are in the process of
preparing crystalline forms of some of the cation-containing
compounds to determine crystallographically how acetate bind-
ing depends on the cation in the crown ether.
Axial Ligation to [LMn(III)]+ Monomers. Since we could
observe differences in the type of axial ligands bound to the
1
Mn(III) in [9b]OAc‚BaTf2‚2H2O and [9b]+‚Ba2+ by H NMR
spectroscopy, we examined axial ligation to other monomers
prepared in this study. It is well-known that Mn(III) porphyrin
and Schiff-base complexes bind one or two axial ligands.59-61
The binding constants and other thermochemical parameters for
some of these binding events have been determined, but these
parameters do not assess either how the Mn(III) center perturbs
the axial ligand or how the axial ligand might affect properties
of the chelate ligand. With regard to photosystem II, it is
significant to ascertain how water molecules in particular interact
with the various sites of the tetramanganese cluster. In a recent
publication,62 ENDOR and ESEEM data have been used to
observe the binding of water and methanol to the MnIII center
of a MnIII,MnIV dimer.63 The very interesting result from that
study is that the hydrogens of the bound water molecule are at
different distances from the MnIII center. 1H NMR spectroscopy
can be an important tool for providing complimentary data
regarding water binding to Mn(III) centers.
Figure 5. 23Na spectra of a dried sample of [9b]PF6 in CD3CN
containing 1 molar equiv of NaSO3CF3 (a) before and (b) after the
addition of 200 molar equiv of H2O.
We have found that [6b]+ is an especially valuable complex
for observing axial ligand binding to Mn(III) by multinuclear
NMR spectroscopy. The discussion that follows pertains to
[6b]+ unless otherwise noted. Solvated monomers of [6b]+
prepared by the CH3OH/O2 and CH3CN/[Cp2Fe]+ oxidation
methods produced compounds with different axial ligands. The
differences are readily apparent when the compounds are
examined by 1H NMR spectroscopy in CD3CN. Other analyti-
cal methods (electrochemistry, FT-IR, and UV/Vis spectroscopy)
that we applied to the characterization of [6b]+ that had been
prepared by the two oxidation methods erroneously led us to
believe that the two oxidation methods produced identical
compounds. Shown in Figure 7a is the spectrum of [6b]+
prepared by the CH3OH/O2 oxidation method while Figure 7b
is the spectrum of [6b]+ synthesized by the alternative oxidation
method. The broad, rapidly relaxing (T1p ) 0.7 ms) resonance
at 16 ppm in Figure 7a arises from the CH3 group of a CH3OH
molecule bound to Mn(III) as this resonance is lost upon the
addition of CD3OD; this compound is designated as [6b]+‚
CH3OH. For the compound used to obtain the spectrum in
Figure 7b, we assume that CH3CN is the axial ligand prior to
dissolution in CD3CN, but that it is rapidly replaced by CD3-
CN; this complex is referred to as [6b]+‚CH3CN. [6b]+‚CH3-
CN is not sufficiently soluble in CH2Cl2 to permit determining
the chemical shift of the bound CH3CN. Comparison of the
resonance positions for the Schiff-base ligand hydrogens on
[6b]+‚CH3OH and [6b]+‚CH3CN reveals that they are only
slightly influenced by the axial ligand. A similar observation
has been made for Mn(III) Schiff-base compounds with alkane
backbones.11
Figure 6. 1H NMR spectrum of [9b]OAc‚BaTf2‚2H2O in CD3CN with
diamagnetic suppression applied. The resonance at 38 ppm is from the
CH3 group of the [CH3CO(O)]- bound to Mn(III).
resonances for [9b]+‚Na+ suggests that the Na+ feels the
magnetic field of the Mn(III) possibly in both a dipolar sense
and a contact term; the contact term arises from the simultaneous
binding of Na+ and Mn(III) to the phenolic oxygens.
1
The H NMR spectrum of [9b]OAc‚BaTf2‚2H2O (Figure 6
and Table 4) in CD3CN is different from the spectrum of [9b]+‚
Ba2+ in CD3CN. [9b]OAc‚BaTf2‚2H2O has a broad resonance
at 36 ppm that is absent in [9b]+‚Ba2+, and no resonance for a
H2O or OH- bound to the Mn(III) center is observed at 17 ppm
(Table 4). A resonance from a bound water molecule could be
broadened beyond our detection limits or masked by crown ether
resonances. The resonance at 36 ppm for [9b]OAc‚BaTf2‚
2H2O arises from the CH3 group of the bound acetate. This
was shown by adding CD3CO(O)D to a CD3CN solution of this
complex, which caused the loss of the 36 ppm resonance and
concurrent appearance of a signal at 2.1 ppm for free
[CH3C(O)O]-. Furthermore, the resonance at 36 ppm was
absent when [9b]OAc was prepared using Mn(CD3C(O)O)2. In
the related compound [MnIII(µ-O)(O2CCH3)2(HB(pz)3)2]
-
(HB(pz)3 ) hydrotris(1-pyrazolyl)borate), the methyl group
of the bridging acetate ligand has a chemical shift of +65.6
ppm.58 In addition to the large downfield shift of the acetate
CH3 group in [9b]OAc‚BaTf2‚2H2O, its rapid T1p value, 0.7
ms, indicates that the paramagnetic nature of the Mn(III) center
is projected strongly through the carboxylate group.
X-ray quality single crystals of [6b]+ were obtained by
dissolving a sample of [6b]+‚CH3OH in CH3CN without
All of the salt complexes prepared using [9b]OAc as a starting
material exhibit spectra consistent with acetate binding to Mn-
(III) (Table 4). The chemical shift of the acetate CH3 group
depends on which cation is in the crown ether. This effect may
arise from different degrees of acetate binding in a bridging
mode to the cation in the crown ether. The variability of acetate
(59) Collman, J. P.; Brauman, J. I.; Hampton, P. D.; Tanaka, H.; Bohle,
D. S.; Hembre, R. T. J. Am. Chem. Soc. 1990, 112, 7980.
(60) Hill, C. L.; Williamson, M. M. Inorg. Chem. 1985, 24, 2836.
(61) Collman, J. P.; Lee, V. J.; Keller-Yuen, C. J.; Zhang, X.; Ibers, J. A.;
Brauman, J. I. J. Am. Chem. Soc. 1995, 117, 692.
(62) Randall, D. W.; Gelasco, A.; Caudle, M. T.; Pecoraro, V. L.; Britt, R.
D. J. Am. Chem. Soc. 1997, 119, 4481.
(63) Larson, E.; Haddy, A.; Kirk, M. L.; Sands, R. H.; Hatfield, W. E.;
Pecoraro, V. L. J. Am. Chem. Soc. 1992, 114, 6263.
(58) Sheats, J. E.; Czernuszewicz, R. S.; Dismukes, G. C.; Rheingold, A.
L.; Petrouleas, V.; Stubbe, J.; Armstrong, W. H.; Beer, R. H.; Lippard,
S. J. J. Am. Chem. Soc. 1987, 109, 1435.

4978 Inorganic Chemistry, Vol. 36, No. 22, 1997
Ciringh et al.
Figure 7. 1H NMR spectra with diamagnetic suppression applied, of
(a) [6b]+‚CH3OH in CD3CN, (b) [6b]+‚CH3CN in CD3CN, and (c) [6b]-
PF6‚H2O in CD3OD. The inset in part c highlights the spectral region
containing the resonance tentatively assigned as a H2O molecule
strongly bound to the Mn(III) center. Peaks for CH3CN and (CH3-
CH2)2O respectively, are from the crystallization process.
Table 5. Selected Bond Lengths (Å) and Angles (deg) for
[6b]PF6‚H2O
Figure 8. (a, top) ORTEP diagram and atom-labeling scheme for [6b]-
PF6‚H2O. Ellipsoids correspond to 50% probability. (b, bottom) A
schematic view of [6b]PF6‚H2O that emphasizes the Mn1-O3-F1
interaction and the association of O2 with Mn1 on a neighboring
compound.
Mn1-O1
Mn1-O3
Mn1-N2
Mn1-O2
Mn1-N1
O3-F1
1.863(7)
2.217(8)
1.959(8)
1.836(7)
1.997(8)
3.01(1)
P1-F2
P1-F4
P1-F6
P1-F1
P1-F3
P1-F5
1.52(1)
1.463(9)
1.54(1)
1.52(1)
1.51(1)
1.46(1)
inhibitor of the catalase enzyme66 and it is proposed that a water
molecule bound to the manganese dimer is hydrogen bonded
to the fluoride anion. Despite the hydrogen bond to F1 in [6b]-
O1-Mn1-O3
O2-Mn1-O3
O3-Mn1-N1
O3-Mn1-N-2
100.8(3)
F1-P1-F2
F3-P1-F5
F4-P1-F6
174.4(7)
177(1)
177.2(8)
-
PF6‚H2O, the bond distances and angles in the PF6 are not
88.4(4)
92.2(4)
94.7(4)
significantly distorted (Table 5). There is a weak contact
between the second axial site on the Mn1 and a phenol oxygen
from a neighboring complex, Mn1-O2′ 2.780(8) Å. This
contact slightly distorts the square plane of the ligand; this is
shown schematically in Figure 8b. We believe that this is the
type of interaction that occurs in the dried monomers.
excluding air and then allowing the diffusion of Et2O vapor
into the CH3CN solution. The molecular structure is shown in
Figure 8a. Data pertinent to the structure can be found in Table
2 while selected bond distances and angles are given in Table
5. The Mn(III) center resides in an essentially square pyramidal
environment with the equatorial plane defined by the Schiff-
base ligand and the axial site occupied by a water molecule,
Mn1-O3 2.217(8) Å. We refer to this molecule as [6b]-
PF6‚H2O. MnIII-O and MnIII-N distances to the Schiff-base
ligand are similar to those found in other five- and six-coordinate
Schiff-base complexes. The bound water molecule in [6b]-
PF6‚H2O is particularly interesting because of the hydrogen-
bonding interaction with F1, O3-F1 3.01 Å. Hydrogen bonding
to PF6- is rare, and we have found only a few recent examples
Shown in Figures 9, traces a and b, are FT-IR spectra of [6b]-
PF6‚H2O and [6b]+‚CH3OH, respectively. It is evident in the
spectrum of [6b]PF6‚H2O that the bands at 842 and 558 cm-1
-
characteristic of free PF6 are absent. If [6b]+‚CH3OH was
recrystallized from CH3CN/Et2O in an inert atmosphere glove-
box using dried solvents (<5 ppm H2O), the precipitate was a
finely divided powder that exhibited absorption bands charac-
-
teristic of free PF6
. The addition of two drops of H2O
(approximately 0.2 mL) to another sample of [6b]PF6 in CH3-
CN under the same inert atmosphere conditions produced
-
crystalline [6b]PF6‚H2O. Clearly, hydrogen bonding to the PF6
is essential for the formation of X-ray quality single crystals of
this molecule, and there must be a strong driving force for
binding water to the Mn(III) center.
-
of transition metal aqua complexes with hydrogen bonds to PF6
in the solid state.64,65 As discussed below, the hydrogen bond
in [6b]PF6‚H2O persists in CD3OD. Identification of the
hydrogen bond in [6b]PF6‚H2O is significant because F- is an
Compound [6b]PF6‚H2O is insoluble in CH3CN while [6b]+‚
CH3OH is freely soluble in that solvent. Both compounds are
(64) Colquhoun, H. M.; Doughty, S. M.; Maud, J. M.; Stoddart, J. F.;
Williams, D. J.; Wolstenholme, J. B. Isr. J. Chem. 1985, 25, 15.
(65) Sletten, J. Acta Chem. Scand. Ser. A 1983, 37, 569.
(66) Penner-Hahn, J. E. In Structural Properties of the Manganese Site in
the Manganese Catalases; Penner-Hahn, J. E., Ed.; VCH Publishers:
New York, 1992; p 29.

Manganese(III) Schiff-Base Complexes
Inorganic Chemistry, Vol. 36, No. 22, 1997 4979
Figure 9. FTIR spectra of (a) [6b]PF6‚H2O and (b) [6b]+‚CH3OH as
KBr disks.
soluble in CH3OH. The 1H NMR spectrum of [6b]PF6‚H2O in
CD3OD exhibits a broad, rapidly relaxing signal at 2.1 ppm
which we tentatively assign to the axially bound H2O molecule
in addition to the resonances expected for the ligand (Figure
7c). We expected that either CD3OD would exchange with the
bound water molecule or an H/D exchange of the protons on
this H2O molecule would occur, resulting in a spectrum
resembling the one for [6b]+‚CH3OH in CD3OD. However,
the continued observation of the bound water signal in the
spectrum of [6b]PF6‚H2O suggested that CD3OD did not
displace the water. In addition, the resonance position of the
bound water also was at abnormally high fields compared to
the other Mn(III) compounds we have investigated. The lack
of H/D exchange is especially interesting as the water is bound
to a cationic metal center which should render it rather acidic.67
We speculated that the unusual properties of the bound water
Figure 10. 19F NMR spectra in CD3OD of (a) [6b]+‚CH3OH, (b)
[6b]+‚CH3OH containing 0.5 molar equiv of (n-Bu4N)BF4, (c) [6b]-
PF6‚H2O, and (d) [6b]PF6‚H2O containing 0.5 molar equiv of (n-Bu4N)-
BF4.
the water hydrogens. This may be pictured as an arrangement
in which the PF6- tumbles on the bound water molecule. The
second possibility is that only one of the axial fluorines, probably
the one trans to the hydrogen bond, or all of the equatorial
fluorines give rise to the doublet, and there is weak or no F-F
coupling. The third explanation is that the signal arises from a
species other than [6b]PF6‚H2O but one that also has the PF6
associated with the Mn(III) center. We believe that the second
explanation is the most correct one. If the trans fluorine is the
origin of this resonance, then one would expect that the signal
intensity should be, at best, 1/6 that of an unbound PF6-. From
signal to noise considerations one would expect that a 36-fold
increase in the acquisition time would be required, which is in
good agreement with our observations. When the influence of
the magnetic field from the Mn(III) is included, this would
further diminish the predictions for signal intensity. In view
of the long acquisition times required to obtain spectra of [6b]-
PF6‚H2O (20 h), variable temperature studies have not been
initiated which could be a means to distinguish the three possible
explanations or suggest others.
1
molecule observed in the H NMR spectrum arose from the
-
hydrogen bond to the PF6 remaining intact in CD3OD. This
hypothesis was tested by acquiring the 19F NMR spectra of [6b]-
PF6‚H2O and [6b]+‚CH3OH in CD3OD.
The 19F NMR spectrum of [6b]+‚CH3OH in CD3OD has a
doublet for the PF6- counterion centered at 40.3 ppm (JP-F 710
Hz) as shown in Figure 10a; a spectrum of KPF6 in CD3OD
has a doublet at 41 ppm (JP-F 706 Hz). In contrast, [6b]PF6‚H2O
in CD3OD has a broader upfield-shifted doublet centered at 32.5
ppm (JP-F 980 Hz) and only a trace of free PF6- (Figure 10c).
We estimate that the concentration of free PF6- in solution for
[6b]PF6‚H2O is on the order of 500 times lower than the
-
concentration of the hydrogen-bonded PF6 in this sample. It
was substantially more difficult to obtain the 19F NMR spectrum
of [6b]PF6‚H2O than to obtain that of [6b]+‚CH3OH. The
spectrum shown in Figure 10c is the result of approximately
20 h of data collection whereas only 20 min was needed for
the one in Figure 10a; the concentrations of [6b]PF6‚H2O and
[6b]+‚CH3OH were 0.033 and 0.036 M, respectively, in these
experiments. The difficulty in observing the 19F signal for [6b]-
-
It was of interest to determine if the hydrogen-bonded PF6
could undergo exchange reactions with BF4-. For this experi-
ment, 0.5 molar equiv of (n-Bu4N)BF4 was added to a solution
of [6b]+‚CH3OH in CD3OD and the 19F NMR spectrum
obtained, Figure 10b; the 19F NMR spectrum of (n-Bu4N)BF4
alone appears as a singlet at -39.2 ppm in CD3OD. There are
no differences in the spectrum when compared to the one in
Figure 10a. In contrast, clear changes occurred when 0.5 molar
equiv of (n-Bu4N)BF4 was added to [6b]PF6‚H2O in CD3OD.
It can be seen in Figure 10d that the signal associated with the
bound PF6- decreased in intensity and broadened compared to
-
PF6‚H2O indicated that the PF6 strongly experienced the
1
magnetic field of the Mn(III) center. The H and 19F NMR
data taken together are consistent with preservation of the
-
hydrogen bond between the Mn(III)-bound water and its PF6
counterion as observed in the solid state data for [6b]PF6‚H2O.
Instead of one doublet for the hydrogen-bonded PF6- coun-
terion in the 19F NMR spectrum of [6b]PF6‚H2O, we expected
to observe three resonances, all with 31P and 19F coupling,
-
the one in Figure 10c, while the doublet for the free PF6
-
because the PF6 has its symmetry lowered to C4V by the
increased in intensity by a factor of about 10. The concentration
hydrogen bond. There are at least three explanations which
might account for the observed spectral doublet. First, the
doublet may arise from an average structure for the anion in
solution where each fluorine resides for only a short time on
-
of “free” PF6 was still about 50 times lower than the amount
of bound PF6- at 8 h after the addition of BF4-, as determined
-
by integration of the “free” PF6 signals. The peak width at
-
half-height for the signal from the BF4 was approximately
-
twice as large as it is in Figure 10b, indicating that BF4 can
(67) Cotton, F. A.; Wilkinson, G. AdVanced Inorganic Chemistry, 5th ed.;
J. Wiley & Sons: New York, 1988.
hydrogen bond to the Mn(III)-bound water molecule. In view

4980 Inorganic Chemistry, Vol. 36, No. 22, 1997
Ciringh et al.
of at least two resonances. Py-d in the absence of [1b]+ has
resonances at 8.74 (R-H; vertical line in the figure), 7.58 (γ-
H), and 7.22 (â-H) ppm. Clearly there are no resonances for
“free” py-d in the spectrum (a capillary containing acetone was
placed in the NMR tube to serve as the chemical shift standard).
On the basis of a proximity argument, we presume that the most
downfield resonance arises from the R-hydrogen and the other
resonance is from the γ- and â-hydrogens.
1H NMR results for paramagnetic Ni(II) compounds that
strongly bind pyridine and Fe(III) dimers69,70 that irreversibly
bind ligands composed of pyridyl groups indicate that the
following three features should be observed in the spectrum of
[1b]+ if the py-d is bound strongly to the Mn(III) center: (i)
chemical shifts for the py-d should be 10-100 ppm from those
of free py-d; (ii) three well-separated py resonances should be
present; and (iii) the chemical shift for the â-hydrogen should
be downfield of the γ-hydrogen. Since none of these three
conditions were observed here, we concluded that the py-d was
undergoing an exchange process at the Mn(III) center that was
rapid on the NMR time scale and that the â- and γ-hydrogens
gave rise to the upfield resonance. The rapid exchange process
is not surprising because axial ligands on Mn(III) centers in
square planar complexes are generally labile and the concentra-
tion of [1b]+ (38.9 mM) is low relative to that of pyridine (12.5
M).
As noted above, dry conditions had to be used to prepare the
samples in py-d because water competes effectively with py-d
for binding to the Mn(III) center. Shown in Figure 11b is the
spectrum taken after 170 equiv of D2O had been added to the
solution used to obtain the spectrum in Figure 11a. There are
now three resonances for the py-d (9.19, 7.79, and 7.64 ppm),
and a broad peak for water appears at 6.37 ppm; free water in
py-d is at ∼5 ppm. The insert in Figure 11 is a plot of the
chemical shift for the py-d resonances Vs quantity of D2O added.
The significant points regarding the plot are (i) the change in
chemical shift for the py-d resonances is not a linear function
of D2O added; (ii) all of the py-d resonances shift upfield as a
consequence of the D2O addition, but even after 170 equiv (6.67
M) of D2O have been added, the chemical shifts are not those
of free py-d; and (iii) each resonance exhibits a different rate
of upfield shift. The first two points along with the chemical
shift value for the water in Figure 11b lead us to conclude that
there is an equilibrium between water and py binding to the
Mn(III) center. Equilibria such as those in eq 4 are likely, and
we are in the process of evaluating the relative binding affinities
of [1b]+ for water and py. Regarding the third point, it can be
Figure 11. 1H NMR spectra of a dried sample of [1b]+ dissolved in
py-d containing (a) 0 and (b) approximately 170 equiv of added D2O;
the vertical line at 8.74 ppm indicates the chemical shift for the
R-hydrogens of free pyridine. The graphical insert depicts the change
in chemical shift of the pyridine hydrogens as a function of the amount
of D2O added to the sample.
of the extremely long acquisition times required for examining
the reaction of [6b]PF6‚H2O with BF4-, we have not attempted
-
to measure the equilibrium constant for the PF6 /BF4- exchange.
The results of the NMR experiments described above illustrate
some rather interesting properties of [6b]+. First, [6b]+ can be
prepared with three different axial ligands, each of which has
characteristic NMR properties. In this context, the 19F NMR
data were important for correcting our initial interpretation of
-
the IR data for [6b]PF6‚H2O where we thought that PF6 was
absent (Figure 9). Second, the precipitation of [6b]PF6‚H2O
from solution must be the driving force for formation of this
compound because the deliberate addition of water (0.5-4
equiv) to [6b]+‚CH3OH in CD3CN does not produce [6b]-
PF6‚H2O within 24 h, as determined by 1H NMR measurements.
Third, the water molecule appears to remain coordinated to the
Mn(III) center during the BF4-/PF6- exchange. Finally, it was
-
anticipated that BF4- would displace the hydrogen-bonded PF6
as it has a less diffuse charge. However, the results suggest
-
that PF6 binding is favored under the current reaction condi-
tions.
Pyridine Coordination. The final axial ligand examined was
pyridine-d5, py-d. Pyridine is often used as an axial ligand for
planar Mn(II) and Mn(III) compounds, and equilibrium binding
constants have been obtained for Mn(II) and Mn(III) porphyrin
complexes.59-61,68 In the case of the Mn(III) porphyrins, both
five- and six-coordinate complexes are observed. Electrochemi-
cal data for some Mn(III) Schiff-base compounds examined in
our laboratory16 indicated that pyridine and N-methylimidazole
were axial ligands for Mn(III) centers; in particular, MnIII/II
couples shifted cathodically and O2 binding rates for Mn(II)
compounds changed upon inclusion of these axial ligands. We
wished to ascertain if pyridine binding was observable by NMR
spectroscopy. A similar NMR study has previously been
performed on Mn(III) porphyrins,60 but the results here are
somewhat different because water competes with pyridine for
binding to the Mn(III) center in the Schiff-base complexes.
The NMR samples for these experiments were prepared by
dissolving [1b]+ (dried form) in py-d from a freshly opened
ampule in an inert atmosphere glovebox to ensure the driest
possible conditions. The 1H NMR spectrum of [1b]+ prepared
under these conditions exhibited two resonances at 9.64 and
7.96 ppm in addition to those arising from the ligand (Figure
11a; for purposes of clarity, ligand resonances are not included,
but they correspond closely to those shown in Figure 2). The
asymmetry of the upfield peak suggested that it was composed
(4)
seen on the plot that the chemical shifts for the two peaks arising
from the single upfield resonance in Figure 11a are diverging
as the quantity of water increases. This requires that the more
downfield member of this pair be due to the γ-hydrogen.
Furthermore, if one makes the extrapolation to irreversible
binding of the py to Mn(III), then the â-hydrogen would be
downfield of the γ-hydrogen as observed for Ni(II) complexes.
It is quite interesting that [1b]+ is so effective at scavenging
even small quantities of water in pyridine; even at 15 equiv of
(69) Norman, R. E.; Holz, R. C.; Menage, S.; O’Connor, C. J.; Zhang, J.
H.; Que, L., Jr. Inorg. Chem. 1990, 29, 4629.
(70) Norman, R. E.; Yan, S.; Que, L., Jr.; Backes, G.; Ling, J.; Sanders-
Loehr, J.; Zhang, J. H.; O’Connor, C. J. J. Am. Chem. Soc. 1990,
112, 1554.
(68) Collman, J. P.; Brauman, J. I.; Fitzgerald, J. P.; Hampton, P. D.; Naruta,
Y.; Michida, T. Bull. Chem. Soc. Jpn. 1988, 61, 47.

Manganese(III) Schiff-Base Complexes
Inorganic Chemistry, Vol. 36, No. 22, 1997 4981
resonance and the pattern of four upfield and three downfield
resonances, plus the one extra, is preserved, we assume that
the ligand is essentially square planar in the µ-oxo dimer. From
an approach involving the systematic substitution of ligand
hydrogens, the most downfield resonance, which occurs at 51.4
ppm in Figure 12a, was concluded to be the additional one; of
particular importance was the observation that this resonance
did not arise from a hydrogen on the Schiff-base ligand.
Indicated on the figure are the resonance assignments based on
a substitution matrix similar to the one used in determining the
resonance assignments given in Figure 2b. Since the additional
resonance in the spectrum of the µ-oxo dimer was not from the
Schiff-base ligand and it has a rapid relaxation rate, it must be
an axial ligand on Mn(III). In view of the synthetic protocol
used to produce the µ-oxo dimers, two likely candidates for
the putative axial ligand were methanol and water.
Figure 12. 1H NMR spectra of 1c in (a) CD2Cl2 and (b) py. The
resonances marked with an asterisk (*) in part b arise from 1c having
py coordinated to the Mn(III) center(s).
Our initial thought was that a water or methanol molecule
was introduced during the synthesis of the Mn(II) starting
material (see above) and that, following the oxygenation
reaction, the methanol or water remained bound to the Mn(III)
center. In order to test this hypothesis, (SALOPHEN)MnII was
synthesized in CD3OD; the resulting solid was washed with
D2O, dried under vacuum, and then oxygenated in CD2Cl2 to
D2O in the presence of 320 equiv of py-d, some displacement
of pyridine occurs. This behavior most likely reflects the
oxophilic nature of manganese. Furthermore, in almost all of
our 1H NMR spectra, the traces of water that are always present
even in high-quality NMR solvents are not detected under
routine data collection conditions (no diamagnetic suppression
applied). Presumably, when the trace water in a given solvent
binds to Mn(III), the paramagnetic nature of the metal center
broadens the resonance and shifts it far from the chemical shift
region being probed. Thus, it appears that all of the Mn(III)
monomer compounds used in this study bind water strongly.
1
produce 1c. The H NMR spectrum of this material showed
the resonance at 51.4 ppm to be of the same intensity relative
to 8H as a sample of 1c prepared using 1a synthesized in CH3-
OH and washed with H2O. This experiment eliminated
methanol or water as an axial ligand being introduced at the
Mn(II) stage. We surmised then that the axial ligand was water
that becomes bound during the oxygenation process.
MnIII µ-Oxo Dimers. Some general comments about
2
obtaining NMR spectra of the dimers are warranted before
presentation of the data. First, the µ-oxo dimers are more
difficult to examine than the monomers because of their lower
solubility. Only non-hydroxylic solvents can be used because
the µ-oxo dimers are cleaved in alcohol solvents. Even some
non-hydroxylic solvents, DMF and DMSO, appear to cleave
selected µ-oxo dimers, indicating that the Mn-O-Mn func-
tionality has limited stability. Second, from systematic substitu-
tion of the ligand hydrogens, we found that the resonance
positions for a few of the ligand hydrogens are more variable
in the µ-oxo dimers than they are in the monomers. This
behavior may arise from both steric interactions and enhanced
dipolar effects from an adjacent Mn(III) center. Third, axial
ligation to the Mn(III) centers occurs in the µ-oxo dimers, but
in the dimers, unlike the monomers, chemical shifts of the ligand
hydrogens are strongly affected by the axial ligand. Finally,
the solvent can cause shifts of up to 10 ppm upfield or downfield
for some resonances. Thus, there are some occasional difficul-
ties in making definitive chemical shift assignments for the
µ-oxo dimers. However, owing to the sensitivity of the µ-oxo
dimers to the conditions under which the spectrum is obtained
and the presence of some signature resonances, one can
immediately identify a µ-oxo dimer in the presence of a
monomer. The ability to differentiate between these two species
was a primary goal of the NMR study.
The spectrum of [(SALOPHEN)MnIII]2(µ-O), 1c, in CD2Cl2
displays eight resonances (Figure 12a and Table 4), which is
one more resonance than for [1b]+ in CD3OD. The spectrum
of each of the dimers synthesized in either CH2Cl2 or CH3CN
showed one extra resonance as compared to the spectrum of its
corresponding monomer. Seven resonances are expected for a
µ-oxo dimer composed of square planar SALOPHEN ligands
adopting a cofacial or bent arrangement and maintaining roughly
C2V or Cs symmetry. Fourteen resonances might be observed
for a distorted ligand structure, but the 1H NMR spectra of other
distorted Mn(III) Schiff-base complexes10,12,13 show fewer
resonances. Since the spectrum of 1c has only one additional
That the downfield resonance was due to bound water was
confirmed by obtaining the spectrum of 1c in py-d (δ ) 44.7
ppm for the downfield resonance in py-d). We assumed that
py-d would displace some or all of the bound water resulting
in loss of the downfield resonance; the released water ligand
then could be conclusively identified from its chemical shift.
The spectrum of a sample of 1c prepared under Ar and taken
within 20 min of being dissolved in py-d is shown in Figure
12b. We have interpreted the peaks marked with an asterisk
(*) as arising from the small portion of µ-oxo dimer molecules
that become coordinated by py-d. We were somewhat surprised
to find that the concentration of py-coordinated dimer increased
only very slowly. However, our studies using the Mn(III)
monomers described above indicated that water binds strongly
to Mn(III) centers.
Slow loss of the bound H2O was confirmed by allowing the
sample to stand in the py-d and acquiring spectra periodically
over 3 weeks. The NMR tube was maintained under Ar
between runs. Selected spectra are shown in Figure 13. It is
clear that the resonance at 44.7 ppm decreased in intensity during
the experiment while one at 5 ppm for free H2O in py-d
increased. A plot of the normalized increase in signal intensity
for water released and the normalized decrease in the signal at
44.7 ppm Vs time is shown as an inset in Figure 13. This result
indicates that the free water comes from the species responsible
for the resonance at 44.7 ppm. As there is little change after
14 days, the reaction appears to reach equilibrium by that time.
The complicated nature of the spectrum taken after 14 days
is consistent with multiple species. First note that there is no
[1b]+(py)(H2O) (compare with Figure 11) so decomposition of
the dimer did not appear to occur to any significant extent. We
believe that there are three species in solution: MnIII2(µ-O)-
(H2O)2, MnIII2(µ-O)(H2O)(py-d), and MnIII2(µ-O)(py-d)2. No
attempt has been made to assign the resonances arising from
the latter two compounds.

4982 Inorganic Chemistry, Vol. 36, No. 22, 1997
Ciringh et al.
compounds. Its utility is greatly enhanced when it is coupled
with a simple diamagnetic suppression routine that reduces the
intensity of solvent peaks. Through the application of the
diamagnetic suppression routine, it was possible to identify and
assign all of the peaks in most of the spectra of the Schiff-base
compounds examined in this study. Diamagnetic suppression
was particularly useful for those compounds in which a ligand
resonance would be masked by a more intense solvent resonance
if suppression was not used.
The suppression routine also was a valuable component in
revealing the subtleties of axial ligation in these compounds.
While many previous workers have shown that axial ligation
to Mn(III) coordination complexes is an integral part of the
chemistry of these species, the variety of molecules that bind
strongly to the Mn(III) centers in our Schiff-base complexes
was somewhat surprising. The probing of axial ligation proved
to be an effective means for detecting subtle differences in what
appeared to be, by other spectroscopic methods, identical
compounds. It also revealed the strong, persistent hydrogen
bond of a Mn(III)-bound water molecule in [6b]PF6‚H2O to
PF6-, an anion generally considered to be noncoordinating. As
noted above, the ability to observe this hydrogen-bonded species
by 19F NMR supports the hypothesis that inhibition of the
catalase reaction by F- may be due to a hydrogen bond between
a Mn-bound water molecule and fluoride.
This NMR study also suggested several reasons why we have
had difficulty in obtaining single crystals of µ-oxo dimers. In
coordinating solvents like py, the existence of multiple species
undoubtedly compromises the ability to grow single crystals of
a pure compound. While only a single species appears to exist
in a noncoordinating solvent like CH2Cl2, the very low solubility
of the µ-oxo dimer in this solvent generally results in too rapid
precipitation and consequently formation of powdery materials.
Finally, the NMR results for the [9b]+‚Mn+ complexes are
intriguing because a Ca2+ cation resides near the tetramanganese
core of the OEC, but the function of the Ca2+ in the water
oxidation process is still a matter of debate. We have shown
that cations within approximately 3.5 Å of the MnIII2 µ-oxo core
in the crown ether complex cause substantial anodic shifts in
the redox potentials of the dimers.21 Similar effects are also
observed for MnIV2 bis(µ-oxo) dimers that incorporate cations.
Our NMR and electrochemical results taken together suggest
that Ca2+ may serve at least two functions in the OEC: (i)
enhancing water binding to manganese and (ii) increasing the
oxidizing power of the manganese cluster. Even in the absence
of Ca2+ the affinity of the Mn(III) Schiff-base monomers and
dimers for water is much larger than one might have anticipated.
Thus, when examining the reaction chemistry of Mn(III)
compounds, one must be prepared to at least consider that water
will be involved in the reaction of interest.
Figure 13. 1H NMR spectra of 1c in py-d obtained at the times
indicated on the figure. The resonances denoted with the symbol (
arise from H2O bound to Mn(III) while those marked with the symbol
b are for free H2O in py. The inset shows the changes in normalized
signal intensities for bound and free H2O as a function of time.
The NMR data suggest that the µ-oxo dimers have essentially
the same square planar ligand geometry observed for the
monomer and that the overall symmetry of the dimer is Cs or
C2V. At least two options are available for the orientation of
the ligands: trans (VI) or cis (VII) orientation of aromatic rings.
In view of the steric demands of the tert-butyl groups on 8c,
we suggest that VI is a better description than VII. The
proposed structure is similar to ones observed in solid state
structures of oxo-bridged Ti(IV)71 and Fe(III)72-76 Schiff-base
complexes although some VVdO Schiff-base dimers do exhibit
the cis-ligand orientation.57,77 The available data do not allow
us to determine if the Mn(III)-O-Mn(III) unit is linear or bent
or if the manganese centers reside within the plane defined by
the Schiff-base ligand. However, six-coordinate Mn(III) com-
pounds tend to have the metal center in the plane of the ligand.
Conclusions. 1H NMR spectroscopy is clearly a powerful
method to characterize paramagnetic Mn(III) coordination
Acknowledgment. This work was supported by the NSF
(CHE 9496236 to C.P.H.).
(71) Sasaki, C.; Nakajima, K.; Kojima, M.; Fujita, J. Bull. Chem. Soc. Jpn.
1991, 64, 1318.
(72) Marini, K. S.; Murray, K. S.; West, B. O. J. Chem. Soc., Dalton Trans.
1983, 143.
(73) Davies, J. E.; Gatehouse, B. M. Acta Crystallogr. 1973, B29, 1934.
(74) Coggon, P.; McPhail, A. T.; Mabbs, F. E.; McLachlan, V. N. J. Chem.
Soc. A 1971, 1014.
(75) Gerloch, M.; McKenzie, E. D.; Towl, A. D. C. J. Chem. Soc. A 1969,
2850.
Supporting Information Available: Tables listing detailed crystal-
lographic data, atomic positional parameters, and bond lengths and
angles for [9b]OAc‚BaTf2‚2H2O and [6b]PF6‚H2O and Figure S1,
depicting a matrix of ligand substitutions and assignments of hydrogen
resonances (8 pages). Ordering information is given on any current
masthead page.
(76) Lewis, J.; Mabbs, F. E.; Richards, A. R. J. Chem. Soc. A 1967, 1699.
(77) Oyaizu, K.; Yamamoto, K.; Yoneda, K.; Tsuchida, E. Inorg. Chem.
1996, 35, 6634.
IC970747Z