Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Magnetic susceptibility measurements
Electrochemical studies
Magnetic measurements for polycrystalline samples of
complexes 1–6 were collected in the temperature range, 2–290 K
in an applied field of 1 T. At 290 K the µeff values of 6.09µB
(χMT = 4.64 cm3 K molϪ1) for 1 and 6.42µB (χMT = 5.16 cm3 K
molϪ1) for 2 decrease monotonically with decreasing tem-
perature until the µeff value of 0.52µB (χMT = 0.034 cm3 K
molϪ1) for 1 and 0.99µB (χMT = 0.12 cm3 K molϪ1) for 2 at 2 K
is attained; this is a clear indication of antiparallel exchange
coupling between two paramagnetic Mn() centers (SMn = 2)
with a resulting St = 0 ground state for both complexes 1 and 2.
Using the isotropic Heisenberg spin-Hamiltonian in the form
H = Ϫ2J S1.S2, the experimental magnetic data were simulated
to evaluate the following best fit parameters: J = Ϫ11.8 cmϪ1,
gMn = 1.998 for 1 and J = Ϫ8.2 cmϪ1, gMn = 1.998 for 2.
The experimental data with the simulations are provided as
Fig. S1 and S2 (supplementary material†). The evaluated
antiferromagnetic coupling constants between the high spin
Mn() ions fall in the range observed for other µ-alkoxo
dimanganese() complexes with comparable Mn ؒ ؒ ؒ Mn
distances.9
Above 15 K complexes 3 and 4 exhibit essentially temper-
ature independent magnetic moments of µeff = 4.74 0.03 for
3 and 4.86 0.10 µB for 4, confirming the high spin d4 elec-
tronic configuration for the manganese centers. Simulations of
the experimental magnetic data using SMn = 2.0, the zero-
field splitting parameter D and gMn, yield gMn = 1.987, D = ϩ2.9
cmϪ1 for 3 and gMn = 1.986, D = ϩ3.4 cmϪ1 for 4 (Fig. S1 and
S2, supplementary material†). We want to emphasize that
the quality of fit decreases dramatically on changing the sign
of D, thus indicating a positive zero-field splitting parameter
for both 3 and 4. We recall that too much significance should
not be put on the sign and values of D evaluated through the
fitting procedure of powder susceptibility data. We have
confirmed the positive sign of the zero-field splitting para-
meters by performing temperature dependent magnetization
measurements at different applied magnetic fields of 1, 4 and 7
T. Variable temperature-variable field magnetization measure-
ments for 4 and their simulations are depicted in Fig. 4. The
simulated parameters are S = 2.0 (fixed), g = 1.98 and D = ϩ3.4
cmϪ1.
The electrochemical properties of the monomeric complexes 3,
4, 5 and 6 were investigated in CH2Cl2 solutions containing 0.1
M [(n-Bu)4N]PF6 as supporting electrolyte. A glassy carbon
working electrode (2 mm disk; for higher scan rates a 0.25 mm
Pt electrode was employed) and a Ag/AgNO3 reference elec-
trode was used. Ferrocene (Fc) was used as an internal standard
and potentials are referenced vs. the ferrocenium/ ferrocene
(Fcϩ/Fc) couple.
The MnIII and MnIV forms of the complexes (3, 4 and 5, 6
respectively) can be electrochemically interconverted. The cyclic
voltammograms of 3 and 5 and of 4 and 6 are identical and
controlled potential coulometry of a solution of 3 at ϩ0.3 V
yields the UV-vis spectrum of 5 after a one-electron oxidation.
The cyclic voltammograms belonging to such interconversion
are shown for a slow scan rate (0.05 V sϪ1) in Fig. 5a). Oxidative
and re-reductive peaks are seen, but the separation of the peaks
is about 0.3 V, i.e. much higher than that for an uncomplicated
1eϪ redox process (0.06 V). This high peak separation was
found to depend strongly on the material of the working elec-
trode (Pt and glassy carbon was used) and on the treatment of
its surface (polishing procedure) prior to the measurements.
This provides evidence that the high peak separation originates
from a slow heterogeneous electron exchange rate rather than
from interfering homogeneous reactions.
ˆ
ˆ ˆ
At higher scan rates (measured up to 30 V sϪ1) (Fig. 5a) the
cyclic voltammograms are more complex. When (with all com-
plexes) the scan is started at positive potentials (i.e. the sur-
rounding of the electrode consists after an equilibration time
(15 s) only of the MnIV forms 5 or 6), the reductive waves of 5 or
6 show at higher scan rates only the characteristics of slow
electron transfer: peak broadening and a shift of the wave’s
maximum towards negative potentials. When, however, the scan
is started at negative potentials, the reduction is the second pro-
cess in the CV which takes place, depending on scan rate, only a
short time after the preceding oxidation. In this case an addi-
tional reduction wave at more positive potentials grows in with
increasing scan rate, at the expense of that for the “direct”
reduction of MnIV to MnIII seen at low scan rates. Clearly this
extra peak must arise from a short-lived intermediate formed
during the oxidation of 3 or 4. At 30 V sϪ1 scan rate the wave for
“direct” reduction has almost disappeared and the reduction
takes place predominantly via this intermediate. It is noted that
the oxidation appears at similar potentials as a single peak at all
scan rates. This intermediate, which becomes visible at high
scan rates because the “direct” metal-centered MnIII/MnIV oxid-
ation is kinetically hindered, must be formed via an alternative
oxidation pathway of 3 or 4. The only feasible alternative is a
ligand-centered oxidation which generates a MnIII–phenoxyl
III
ؒ
radical (Mn –O ) species. In the course of a slow scan it trans-
forms into the thermodynamically slightly more stable MnIV
which is re-reduced at potentials ≤ Ϫ0.3 V. At high scan rates,
when the time required for a scan is shorter than the life time of
the phenoxyl radical, only electrochemical reversible ligand-
centered reduction at 0 V takes place.
The reaction scheme for the electrochemical 3/5 inter-
conversion can be written as:
Fig. 4 The field-dependent magnetization of 4 as a function of
temperature.
MnIII Ϫ eϪ
MnIV, E 0Ј = Ϫ0.23 V,
ks = 5 × 10Ϫ4 cm sϪ1, α = 0.7 (1)
Above 10 K complexes 5a and 5 show essentially temperature
independent µeff values of 4.30
0.10µB and 3.65
0.03µB,
respectively, thus corroborating with the d3 electronic configur-
ation for the manganese centers in these complexes. The higher
µeff value for the Mn() center in 5a can be explained by the
contribution of high TIP from ferrocene (low spin d6) present in
the crystals of 5a. The magnetic moment of 6 is 3.83 0.02µB,
which is identical with the spin-only value for a d3 system. Thus
the experimental µeff values clearly indicate that we are dealing
with genuine Mn() species in 5, 5a and 6.
III
MnIII Ϫ eϪ
Mn –O , E 0Ј = Ϫ0.05 V
(2)
(3)
ؒ
MnIV, k ≈ 1 sϪ1
III
ؒ
Mn –O
where ks is the heterogeneous rate constant, α is the transfer
coefficient and k is the unimolecular rate constant of reaction
(3).
D a l t o n T r a n s . , 2 0 0 3 , 3 1 3 6 – 3 1 4 4
3139
Paine, Tapan Kanti
