Ligand exchange upon oxidation of a dinuclear Mn complex-detection of structural changes by FT-IR spectroscopy and ESI-MS

Article abstract of DOI:10.1039/b415148h

The structural rearrangements triggered by oxidation of the dinuclear Mn complex [Mn₂(bpmp)(μ-OAc)₂]⁺ (bpmp = 2,6-bis[bis(2-pyridylmethyl)amino]methyl-4-methylphenol anion) in the presence of water have been studied by combinations of electrochemistry with IR spectroscopy and with electrospray ionization mass spectrometry (ESI-MS). The exchange of acetate bridges for water (D₂O) derived ligands in different oxidation states could be monitored by mid-IR spectroscopy in CD ₃CN-D₂O mixtures following the v_as(c-o) bands of bound acetate at 1594.4 cm^-1 (II,II), 1592.0 cm^-1 (II,III) and 1586.5 cm^-1 (III,III). Substantial loss of bound acetate occurs at much lower water content (< 0.5% v/v) in the III,III state than in the II,II and II,III states (≥10%). The ligand-exchange reactions do not initially reduce the overall charge of the complex but facilitate further oxidation by proton-coupled electron transfer as the water-derived ligands are increasingly deprotonated in higher oxidation states. In the IR spectra deprotonation could be followed by the formation of acetic acid (DOAc, ～1725 cm^-1, v_(c-o)) from the released acetate (1573.6 cm ^-1, v_as(c-o)). By the on-line combination of an electrochemical flow cell with ESI-MS several product complexes could be identified. A di-μ-oxo bridged III,IV dimer [Mn₂(bpmp)(μ-O) ₂]²⁺ (m/z 335.8) can be generated at potentials below the III,III/II,III couple of the di-μ-acetato complex (0.61 V vs. ferrocene). The ligand-exchange reactions allow for three metal-centered oxidation steps to occur from II,II to III,IV in a potential range of only 0.5 V, explaining the formation of a spin-coupled III,IV dimer by photo-oxidation with [Ru[bpy) ₃]³⁺ in previous EPR studies. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b415148h

View Article Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
Ligand exchange upon oxidation of a dinuclear Mn complex–
detection of structural changes by FT-IR spectroscopy and ESI-MS†
Gerriet Eilers,a Camilla Zettersten,b Leif Nyholm,‡b Leif Hammarstro¨ma and Reiner Lomoth*a
a Department of Physical Chemistry, Uppsala University, BOX 579, S-75123, Uppsala, Sweden.
E-mail: Reiner.Lomoth@fki.uu.se
b Department of Analytical Chemistry, Uppsala University, BOX 599, S-75124, Uppsala,
Sweden
Received 30th September 2004, Accepted 31st January 2005
First published as an Advance Article on the web 15th February 2005
The structural rearrangements triggered by oxidation of the dinuclear Mn complex [Mn2(bpmp)(l-OAc)2]+ (bpmp =
2,6-bis[bis(2-pyridylmethyl)amino]methyl-4-methylphenol anion) in the presence of water have been studied by
combinations of electrochemistry with IR spectroscopy and with electrospray ionization mass spectrometry
(ESI-MS). The exchange of acetate bridges for water (D2O) derived ligands in different oxidation states could be
monitored by mid-IR spectroscopy in CD3CN–D2O mixtures following the vas(C bands of bound acetate at
–O)
1594.4 cm−1 (II,II), 1592.0 cm−1 (II,III) and 1586.5 cm−1 (III,III). Substantial loss of bound acetate occurs at much lower
water content (< 0.5% v/v) in the III,III state than in the II,II and II,III states (≥10%). The ligand-exchange reactions
do not initially reduce the overall charge of the complex but facilitate further oxidation by proton-coupled electron
transfer as the water-derived ligands are increasingly deprotonated in higher oxidation states. In the IR spectra
deprotonation could be followed by the formation of acetic acid (DOAc, ∼1725 cm−1, v(C O)) from the released acetate
–
(1573.6 cm−1, vas(C O)). By the on-line combination of an electrochemical flow cell with ESI-MS several product
–
complexes could be identified. A di-l-oxo bridged III,IV dimer [Mn2(bpmp)(l-O)2]2+ (m/z 335.8) can be generated at
potentials below the III,III/II,III couple of the di-l-acetato complex (0.61 V vs. ferrocene). The ligand-exchange
reactions allow for three metal-centered oxidation steps to occur from II,II to III,IV in a potential range of only 0.5 V,
explaining the formation of a spin-coupled III,IV dimer by photo-oxidation with [Ru[bpy)3]3+ in previous EPR studies.
upon oxidation of a Mn2(II,II) complex has also been reported
Introduction
representing a more special case of charge compensation by
Synthetic Mn complexes have been extensively studied as
deprotonation.16 Charge compensating exchange of ligands
structural models for the active site of metalloenzymes like
is also observed in the catalytic disproportionation of H2O2
catalase or the oxygen evolving center of photosystem II
by di(l-carboxylato)Mn2(II,II) complexes where the exchange
(PSII).1–5 The active site metal center of PSII goes through
of acetate bridges for oxo ligands leads to the active di(l-
several oxidation states as it couples single electron transfer steps
to multiple electron catalytic reactions. Subsequent oxidations
of the synthetic complexes, however, are typically separated by
several hundreds of mV. This imposes severe problems for appli-
oxo)Mn2(III,IV) species.17
For the dimeric Mn2(II,II) complex [Mn2(bpmp)(l-OAc)2]+
(1a) {bpmp− is the anion of 2,6-bis[bis(2-pyridylmethyl)-
amino]methyl-4-methylphenol} oxidation to a Mn2(III,IV) com-
plex has been accomplished by photo generated [Ru[bpy)3]3+.18,19
Stepwise oxidation of [Mn2(bpmp)(l-OAc)2]+ by electron trans-
fer to the oxidized sensitizer could account for the formation of
[Mn2(bpmp)(l-OAc)2]2+ (1b) and [Mn2(bpmp)(l-OAc)2]3+ (1c)
with the potentials of the II,III/II,II and III,III/II,III couples well
below the oxidant. Further oxidation of the same complex
beyond the III,III state is however impossible and the third
oxidation step resulting in a III,IV complex must rely on precedent
rearrangements of the ligand set. Since the formation of the
Mn2(III,IV) dimer was observed only in partly aqueous solutions
it was proposed that in higher oxidation states the exchange of
acetate bridges (OAc−) for more negative oxo (O2−) ligands might
provide the charge compensation needed for a third oxidation
step to occur below the oxidation potential of the [Ru[bpy)3]3+
oxidant.18,19 A mono- or di-l-oxo bridging motif was suggested
but the structure of the Mn2(III,IV) dimer and the mechanism of
its formation could not be revealed from the EPR data.
cations as functional mimics in artificial photosynthesis,6–11 since
photo-oxidation of the Mn complex, even with strong oxidants
as [Ru(bpy)3]3+ (E0 ∼ 1.3 V vs. SCE), is difficult to achieve
more than once or twice before further oxidation becomes
thermodynamically infeasible or results in oxidative degradation
of the ligand. In the natural reaction center of PSII12,13 all four
oxidation steps of the Mn cluster needed for water oxidation
are compressed in about 0.3 V only. This narrow potential
range is a result of deprotonation reactions that provide charge
compensation in response to the oxidation steps. Therefore,
reactions that compensate for the increasing charge on the metal
centers are most probably indispensable for a functional mimic.
For some synthetic Mn complexes proton-coupled electron
transfer has been reported and the potentials for oxidation
of the oxo-bridged Mn dimers have been shown to drop by
several hundreds of mV in response to deprotonation.14,15 The
transformation of a terminal aquo ligand into an oxo bridge
In this paper we describe an investigation of the structural
rearrangements of [Mn2(bpmp)(l-OAc)2]n+ (n = 1, 2, 3) in
the presence of water by combinations of electrochemistry
with IR spectroscopy and with electrospray ionization mass
spectrometry (ESI-MS). By mid-IR spectroelectrochemistry we
were able to monitor the loss of acetate ligands in the II,II; II,III
and III,III state as well as the release of protons from bound
water in response to oxidation. Further information on the
product structures was obtained by the on-line combination
† Electronic supplementary information (ESI) available: Fig. S1:
Spectroelectrochemical changes upon oxidation of [MnII,II2(bpmp)-
(l-OAc)2]+ in CD3CN. Fig. S2: IR spectra of [MnII,II2(bpmp)(l-
OAc)2][ClO4] and [MnII,III2(bpmp)(l-OAc)2][ClO4]2 in the solid state
and in CD3CN. Fig. S3: Selected full range EC–ESI-MS spec-
tra of [MnII,II2(bpmp)(l-OAc)2]+. See http://www.rsc.org/suppdata/
dt/b4/b415148h/
‡ Current address: Department of Materials Chemistry, Uppsala Uni-
versity, BOX 538, S-75121, Uppsala, Sweden.
T h i s j o u r n a l i s
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5
D a l t o n T r a n s . , 2 0 0 5 , 1 0 3 3 – 1 0 4 1
1 0 3 3
©

View Article Online
of an electrochemical flow cell with ESI-MS.20–22 This method
was now employed for the first time for the in situ detection
of ligand-exchange products formed upon oxidation of a metal
complex. By this approach a number of ligand-exchanged dimers
formed in response to oxidation of [Mn2(bpmp)(l-OAc)2]+ could
be identified including a III,IV dimer with a m/z value consistent
with the complex [Mn2(bpmp)(l-O)2]2+.
electrode was a Pt grid (1024 meshes cm−2, 65% open area). The
counter electrode was a Pt cylinder concentric with the working
electrode and sufficiently separated from the thin layer compart-
ment to ensure that no electrolysis product from the counter
electrode can diffuse into the analysed volume on the timescale
of the experiments. The reference electrode was of the same type
as described for electrochemistry and connected to the cell via a
salt bridge (0.5 M KPF6 in CD3CN).
CD3CN, D2O and KPF6 were purchased from Sigma–Aldrich.
CD3CN was dried over molecular sieve (3 A) prior to use and
Experimental
˚
KPF6 was dried overnight at 353 K in vacuum.
Materials
[Mn2(bpmp)(l-OAc)2][ClO4] and [Mn2(bpmp)(l-OAc)2][ClO4]2
were provided by Magnus Anderlund and Prof Licheng Sun
(Stockholm University, Sweden) and had been prepared accord-
ing to reported methods.23
EC–ESI-MS
A thin layer electrochemical flow cell (Bioanalytical Systems
Inc., West Lafayette, IN) was coupled to a PE-Sciex API III+
triple quadrupole mass spectrometer (Concord, ON, Canada)
with an IonSpray interface. An isolation transformer (type PVM
440 0019, Tufvassons, Sigtuna, Sweden) was used to decouple
the electrochemical cell from the ESI high voltage so the EC
cell and the potentiostat (Polarographic Analyzer/Stripping
Voltammeter, Model 264, EG&G Princeton Applied Research,
Princeton, NJ) were floating on the high voltage. Details of
the coupling of the electrochemical thin layer flow cell to the
mass spectrometer are described elsewhere.22 Sample solution
was introduced to the floating system (isolated from ground by
pieces of Styrofoam) by a syringe pump (Syringe Infusion Pump
22, Harvard Apparatus, Inc., Cambridge, MA).
In the electrochemical equipment, the working electrode was
a series dual glassy carbon (GC) disc electrode (3 mm diameter)
embedded in a PEEK block (Bioanalytical Systems Inc.). The
spacer gasket (Bioanalytical Systems Inc.) that was used was
16 lm thin. The stainless steel block of the flow cell was
used as a pseudo-reference electrode and a laboratory made
stainless steel electrode, which was made to fit the reference
electrode compartment of the flow cell, was used as the counter
electrode. This modified set-up of the flow cell was made
to minimize interferences as a result of the electrochemical
reactions occurring at the counter electrode.
Electrochemistry
Cyclic voltammetry, differential pulse voltammetry and con-
trolled potential electrolysis were carried out using an Autolab
potentiostat with an GPES electrochemical interface (Eco
Chemie). The working electrode was a glassy carbon disc
(diameter 3 mm, freshly polished) for voltammetry or a platinum
grid for bulk electrolysis, respectively. A platinum spiral in a
compartment separated from the bulk solution by a fritted
disk was used as counter electrode. The reference electrode was
a non-aqueous Ag+/Ag electrode (CH Instruments, 0.010 M
AgNO3 in acetonitrile) with a potential of −0.08 V vs. the
ferrocenium–ferrocene (Fc+/Fc) couple in acetonitrile as an
external standard. All potentials reported here are vs. the Fc+/Fc
couple by adding −0.08 V to the potentials measured vs. the
Ag+/Ag electrode.
All solutions were prepared from dry acetonitrile (Merck,
˚
spectroscopy grade, dried with MS 3 A) with 0.1 M tetra-
butylammonium perchlorate (Fluka, electrochemical grade)
as supporting electrolyte that has been dried in vacuum at
383 K. Before all measurements oxygen was removed by
bubbling the stirred solutions with solvent-saturated argon
and the samples were kept under argon atmosphere during
measurements.
Experiments were made in positive electrospray mode. The
fused silica spray capillary from the flow cell (250 mm long,
50 lm inner diameter and 184 lm outer diameter, total volume
of 0.49 ll, Polymicro Technologies, Phoenix, AZ) to the mass
spectrometer was centered in a stainless steel capillary counter
assembly delivering 0.7 L min−1 of air (zero grade, generated in
house) for pneumatically assisted ESI-MS. The flow rate of dry
nitrogen counter-current curtain gas (99.999% purity, 5.0, AGA,
Stockholm, Sweden), which was heated to 63 ◦C, was 1.2 L min−1
over the sampling orifice. Mass spectrometric parameters: ion
spray voltage (ISV) 3400 V, interface plate voltage (IN) 650 V,
orifice lens voltage (OR) 35 V, and AC entrance rod (R0)
30 V. About 30 scans between m/z 150–900 or 1000 (step size
0.2), with a dwell time of 1.0 ms and pause time of 0.052 ms,
were recorded in multichannel acquisition (MCA) mode during
1.97 min. Data were collected on a Macintosh computer by the
Tune2.5-FPU software and analyzed by Multiview 68 K 1.3.
The use of supporting electrolyte was limited by the sup-
pression effect24–27 from the TBA+ ions on the analyte ions in
the electrospray process. This required relatively high concen-
trations of the analyte to replace most or the entire support-
ing electrolyte in the EC-MS experiments. The measurements
were performed with 45 lM [Mn2II,II(bpmp)(l-OAc)2][ClO4] in
acetonitrile and 97 lM [Mn2II,II(bpmp)(l-OAc)2][ClO4], 25 lM
TBAP in acetonitrile–water solution. Under these conditions
conductivity was sufficient for the EC cell while suppression
effects were avoided.
IR Spectroscopy
IR absorption spectra were recorded between 1000 and
5000 cm−1 at a resolution of 1 cm−1 on a Bruker FT-IR
spectrometer (IFS 66 v/S) with the sample as a KBr pellet or
as solution in a liquid-sample-cell (Bruker A140) between CaF2
windows and a pathlength of 125 lm.
The solutions were prepared in CD3CN (Aldrich, 99.8 atom%
D) or mixtures of CD3CN with D2O (Aldrich, 99.9 atom% D).
With the deuterated solvents and KPF6 (0.1 M) as supporting
electrolyte the background absorption in the range of the
carboxylate bands (1800–1200 cm−1) was minimized and could
be subtracted for D2O contents of ≤75% (v/v). Minor differ-
ences in residual H2O content between sample and background
measurement gave rise to incomplete background correction in
some spectra. Where necessary the remaining absorption bands
of H2O (1632 cm−1, d(H H)) and HOD (1432 cm−1, d(H
were subtracted from the presented spectra.
)
–O
–
–O–D)
Higher oxidation states of the complex were prepared by
exhaustive electrolysis of [Mn2(bpmp)(l-OAc)2][ClO4] (6 mM)
in CD3CN. After electrolysis the samples were diluted 1 : 1 with
CD3CN or CD3CN–D2O mixtures to allow for the variation of
the D2O-concentration while keeping the concentration of the
complex constant.
IR spectroelectrochemical measurements were performed on
solutions of [Mn2(bpmp)(l-OAc)2][ClO4] (6 mM) in CD3CN or
CD3CN–D2O with KPF6 (0.1 M). The spectra were recorded
in absorption mode in a thin layer cell equipped with CaF2
windows and the optical path length set to 125 lm. The working
Samples were pumped through the thin layer electrochemical
flow cell at a flow rate of 10 ll min−1 in acetonitrile–water and
5 ll min−1 in the less conductive acetonitrile solution where the
flow rate was reduced to obtain a stable electrospray.
1 0 3 4
D a l t o n T r a n s . , 2 0 0 5 , 1 0 3 3 – 1 0 4 1

View Article Online
the absence of additional voltammetric peaks on the second
oxidation scan indicates that also the II,III state has little
reactivity towards water.
Results and discussion
Cyclic voltammetry
The irreversibility of the second wave and the new peak on
the reverse scan clearly demonstrate the effect of water on the
III,III state. Assigning the irreversibility tentatively to a charge
compensating ligand-exchange reaction (see below) the new
cathodic peak can be attributed to the reduction of one of the
products. The new cathodic peak does not have any counterpeak
in the second oxidation scan, which indicates that the ligand-
exchange product rapidly and quantitatively re-establishes the
acetate bridges upon reduction. Another indication for this
behavior is that the first oxidation wave reaches the same
amplitude in the second scan as in the first scan.
It could be anticipated that ligand-exchange reactions involve
the relatively weakly bound acetate bridges. The C–O stretching
frequencies of carboxylate ions are known to change upon
coordination to a metal center. This provides a probe for the
degree of acetate loss from the complex in different oxidation
states.
In dry acetonitrile, cyclic voltammograms of [Mn2(bpmp)(l-
OAc)2]+ show two quasi-reversible waves that have been at-
1
tributed earlier to the Mn2(II,III)/Mn2(II,II) (E (1) = 0.06 V
2
vs. Fc+/0) and Mn2(III,III)/Mn2(II,III) (E (2) = 0.61 V) couples,
1
2
respectively18,23,28 (see Fig. 1(b)). The potential and chemical
reversibility of the first process is nearly unaffected by the
presence of water (v/v ≤10%)18 (Fig. 1(a)). The peak current is
only somewhat diminished and the peak split slightly increased
in acetonitrile/water (90 : 10) as compared to neat acetonitrile.
The second oxidation wave, however, becomes broader and
chemically irreversible when changing from acetonitrile to
acetonitrile–water mixtures. On the reverse CV scans extending
to sufficiently negative potential an additional cathodic peak is
observed at −0.25 V provided that water is present and that the
second oxidation step was accomplished in the forward scan
(Fig. 1(a), bold line). This peak occurs at a potential more
negative than the re-reduction of [Mn2II,III(bpmp)(l-OAc)2]2+ to
[Mn2II,II(bpmp)(l-OAc)2]+ but in subsequent scans no counter
peak is observed on the forward scans and peak heights of the
original oxidation peaks are unaffected by repeated scanning.
IR Spectroscopy
Band assignments and different redox states in anhydrous
acetonitrile. Fig. 2 shows the spectra of the complex in the
II,II (a), the II,III (b) and III,III states (c) in anhydrous acetonitrile
(CD3CN) solution together with spectra of the Hbpmp ligand
(d), free acetate (as sodium acetate, e) and acetic acid (f).
All spectral changes observed upon oxidation of the complex
could be fully reversed upon re-reduction in carefully dried
acetonitrile and several isosbestic points were preserved in the
interconversions (see ESI†) The spectra in Fig. 2 show absolute
spectra for the different redox-states rather than difference
spectra, because maxima of absorption changes do not represent
the maxima in the absolute spectra in the case of overlapping
absorption bands.
In the II,II state (Fig. 2(a)) the most intense bands observed at
1594 and 1423 cm−1 can be assigned to the asymmetric (vas) and
symmetric stretching (vs) modes of the acetate ligands though
they are superimposed to absorptions of the bpmp− ligand
(compare with Fig. 2(d)) that arise mainly from ring stretching
modes (1571, 1595 cm−1) of the pyridyl and phenolic groups. The
shoulder at 1601 cm−1 in the complex corresponds to the band
at 1595 cm−1 of the Hbpmp ligand that shifts by 6 cm−1 upon
coordination while the peak at 1594 cm−1 can be attributed to the
mas mode of the acetate ligands. This assignment is supported by
the spectrum of the complex without acetate ligands (see below).
Oxidation to the II,III state (Fig. 2(b)) shifts both acetate
bands to lower frequencies (see Table 1) and decreases the peak
intensity of the mas band. Upon further oxidation to the III,III state
both bands shift to even lower frequencies and the intensity of
the mas band is further reduced. The decreasing intensity of the
Table 1 IR band assignments
m˜a/cm−1 (e/M−1 cm−1
)
Acetate
Pyridine, phenol
Compoundb
vas(C–O)
vs(C–O)
v(C–C)c
1a
1594.4 (1870)
1592.0 (840)
1586.5 (700)
1573.6 (840)
1422.0 (670)
1390.0 (370)
1384.4 (270)
1410.0 (470)
1601d (350)
1605d (1000)
1609d (470)
—
Fig. 1 Cyclic voltammograms (0.100 V s−1) of 1a (2 mM) in CH3CN
with 10% (v/v) of water (a) and neat CH3CN (b) with 0.1 M
[(n-C4H9)N][ClO4] as supporting electrolyte. Inset: Second scan.
1b
1c
CH3COO−e
CH3COODf
Hbpmp
1723.5 (730)
—
The effects of water on the electrochemical behavior of
[Mn2(bpmp)(l-OAc)2]+ can be rationalized in terms of ligand-
exchange reactions. The rather small effect on the first voltam-
metric wave demonstrates that with a water content up to 10%
only a minor fraction of the initial II,II state undergoes ligand-
exchange reactions. The reversibility of the first oxidation and
—
1595.6 (560)
0.3 cm−1 b In CD3CN solution. c Several ring stretching modes of
.
a
the aromatic moieties occur in this range: pyridine (1581, 1598 cm−1),
phenol (1598, 1606 cm−1). d Shoulder. m˜ and e from fit with lorentzian
bands. e As sodium acetate. f With 10% D2O.
D a l t o n T r a n s . , 2 0 0 5 , 1 0 3 3 – 1 0 4 1
1 0 3 5

View Article Online
(ESI†) are identical to the spectra in solution with Dm˜ ≥ 200 cm−1
also for the solid II,III material. Thus, the spectra in Fig. 2(a)–
(c) are all assigned to bis-l-acetato bridged complexes and the
different frequencies, intensities, and values of Dm˜ are attributed
to the direct effect of metal oxidation state rather than changes in
coordination mode. In all three oxidation states the frequencies
of these bands (see Table 1), in particular of the vas band, are
distinctly different from those of the free acetate ion (and acetic
acid, Fig. 2(f)), allowing quantitative monitoring of the loss of
acetate ligands in the presence of water.
Effect of water on the II,II state. Fig. 3 shows the spectrum of
[Mn2(bpmp)(l-OAc)2]+ in dry acetonitrile and spectra obtained
in mixtures with 10, 30, 50, 70 and 90% D2O. The 90% D2O
spectrum does not extend below 1500 cm−1 because of excessive
solvent absorption impeding proper background correction. The
peaks at 1594 and 1423 cm−1 assigned to the vas and vs modes
of acetate bound to the II,II complex decrease with increasing
water content, while the absorption of free acetate increases
concomitantly as indicated by the arrows. The vas band of free
acetate accounts for the peak around 1560 cm−1 while the vs
band gives rise to the increase in absorption around 1400 cm−1.
With 90% water the vas band of bound acetate disappears
almost completely and the intensity of the band at 1560 cm−1
corresponds to almost two equivalents of free acetate. From this
data it can be concluded that basically all acetate bound to the
II,II complex is released as free ions at a water content of 90%.
The remaining peak at 1607 cm−1 in the spectra with 70 and 90%
Fig. 2 IR spectra of 1a (a), 1b (b) and 1c (c) obtained by spectroelec-
trochemistry of [Mn2(bpmp)(l-OAc)2][ClO4] (3 mM) in CD3CN with
0.5 M KPF6. IR spectra of Hbpmp (3 mM in CD3CN) (d), NaOAc
(6 mM in CD3CN) (e) and DOAc (6 mM HOAc in CD3CN with 10%
(v/v) D2O) (f).
vas band in the higher oxidation states can not be attributed to
loss of bound acetate as the vs band is about equally intense in
all oxidation states and no major spectral contributions of free
acetate or acetic acid (compare with Fig. 2(e)) are observed.
The separation Dm˜ = m˜as − m˜s is usually correlated to the
coordination mode of carboxylate ligands29 and the value of
171 cm−1 of the II,II state would be in agreement with bridging
coordination. In the II,III and III,III states Dm˜ increases to 203 and
201 cm−1, respectively, and values of Dm˜ ≥ 200 cm−1 are generally
considered to indicate monodentate coordination. However, it
has been shown that Dm˜ for bridging acetate can largely depend
on metal oxidation state of isostructural Mn complexes30–32 and
the simple Dm˜ criterion has been questioned.33 In the solid state
bridging coordination is evident from X-ray crystallography of
the II,II34 and II,III23 state and IR spectra of the solid materials
Fig. 3 (a) IR spectra of [Mn2II,II(bpmp)(l-OAc)2]+ (3 mM) in CD3CN
and in CD3CN–D2O mixtures with 10, 30, 50, 70 and 90% (v/v) D2O
(0.1 v/v% D2O is equivalent to 55 mM D2O). Arrows indicate increasing
D2O content. (b) Absorbance of bound acetate (vas) with increasing water
content in the II,II (᭺), II,III (ꢀ) and III,III (ꢀ) state.
1 0 3 6
D a l t o n T r a n s . , 2 0 0 5 , 1 0 3 3 – 1 0 4 1

View Article Online
D2O corresponds to the shoulder on the vas bands in anhydrous
solution. It can be attributed to a vibration of the bpmp ligand
that occurs at 1595 cm−1 in Hbpmp and is shifted to 1607 cm−1
upon coordination to the manganese. Based on this assignment
the IR spectra provide evidence that even so the acetate bridges
are lost the Mn ions are still bound to the bpmp ligand.
In agreement with the electrochemistry results most of the
complex is found intact at 10% water with less than 15% of the
bound acetate being lost. If the partial loss of absorption from
bound acetate at lower water contents is arising from the loss of
one or both acetate ligands per complex can not be elucidated
from the IR data. Fig. 3(b) shows a plot of the absorption from
bound acetate vs. water concentration. The sharp vas band of
bound acetate is the most reliable indicator for the loss of acetate
as the broader band of free acetate is more overlapping with
other absorptions and the shape of the acetic acid band depends
on the solvent composition (see Figs. 4(a) and 5(a)). For the II,II
state the decrease of bound acetate vs. water concentration is
nearly linear (Fig. 3(b)). Such a behavior can not be rationalized
in terms of a simple ligand-exchange equilibrium. This is not
surprising, since the equilibrium constant itself would change
with such a large change in solvent composition. From the IR
data we can not decide whether water coordinates to manganese
as acetate ligands are lost. Since no acetic acid is observed, as in
higher oxidation states (see below), water would have to bind to
the II,II complex as aquo rather than hydroxo or oxo ligand.
Fig. 5 (a) IR spectrum of [Mn2III,III(bpmp)(l-OAc)2]2+ (3 mM) gener-
ated by bulk electrolysis (0.82 V) in CD3CN and spectra with 0.1, 0.5,
2, 5, 20 and 50% (v/v) D2O added after electrolysis. Arrows indicate
increasing D2O content. (b) Spectroelectrochemical changes during
oxidation (0.82 V) of [Mn2II,III(bpmp)(l-OAc)2]+ (3 mM) in CD3CN
with 1% (v/v) D2O. Arrows indicate increasing time of electrolysis (20,
100, 600, 1080 s).
For the higher oxidation states the effect of water was
studied in two different ways: in one approach water was added
after exhaustive bulk electrolysis to the II,III or III,III state in
dry acetonitrile or, alternatively, to the chemically oxidized
preparation of the II,III complex. In the other approach water
was added to the solution of the II,II complex prior to electrolysis
in the spectroelectrochemical cell. With the first approach a
simpler situation is created where potential ligand-exchange
reactions are restricted to the oxidation state prepared by the
initial electrolysis. The latter approach, however, includes the
possibility of further oxidation of ligand-exchange products if
the potential for this process comes close to or drops below the
electrode potential that is continuously applied.
Note that after complete electrolysis no spectral changes were
observed for at least 20 min, in any of the redox states, with any
of the solvent compositions.
Effect of water on the II,III state. Fig. 4(a) shows the
spectrum of [Mn2(bpmp)(l-OAc)2]2+ in dry acetonitrile and
spectra obtained in mixtures with 10, 30, 50 and 75% D2O
added after oxidation. The increasing loss of acetate ligands with
increasing water concentration (see also Fig. 3(b)) is evidenced
by the depletion of the 1592 cm−1 band that arises from the vas
mode of the bound acetate. Concomitantly rising absorptions
around 1560 and 1720 cm−1 are observed indicating that part of
the released acetate is protonated. At the highest water content
basically all bound acetate is released and the absorption of free
acetate amounts to about two thirds of two equivalents of acetate
Fig. 4 (a) IR spectrum of [Mn2II,III(bpmp)(l-OAc)2]2+ (3 mM) gen-
erated by bulk electrolysis (0.32 V) in CD3CN and spectra with 10,
30, 50 and 75% (v/v) D2O added after electrolysis. Arrows indicate
increasing D2O content. (b) Spectroelectrochemical changes during
oxidation (0.32 V) of [Mn2II,III(bpmp)(l-OAc)2] (3 mM) in CD3CN with
10% (v/v) D2O. Arrows indicate increasing time of electrolysis (35, 80,
1200 s).
D a l t o n T r a n s . , 2 0 0 5 , 1 0 3 3 – 1 0 4 1
1 0 3 7

View Article Online
while the acetic acid approximately accounts for the remaining
third. Exact quantification of acetic acid is difficult however due
to the variation of the spectrum with D2O content. These results
show that the exchange of acetate for water in the II,III complex is
accompanied by deprotonation of the binding water as a charge
compensation for the higher metal oxidation state. Assuming
that all protons released in this process result in the formation
of acetic acid we can estimate that on average not more than one
proton is released per complex.
In a complementary experiment (see Fig. 4(b)) 10% D2O
was added to a solution of the II,II complex before electrolysis
at 0.32 V that oxidized the intact complex to the II,III state.
With this water content only about 15% of the II,II complex
undergoes ligand exchange (see above). Thus, electrolysis under
these conditions oxidizes predominantly the intact complex.
In the spectrolectrochemical thin-layer cell, exhaustive elec-
trolysis occurs in less than 30 s, and the first spectrum shown
in Fig. 4(b) was obtained after 35 s. It largely resembles the
spectrum observed after addition of 10% D2O to the II,III
complex (Fig. 4(a)) showing the loss of bound acetate and
the release of both free acetate and acetic acid. Keeping the
applied potential the acetic acid peak continues to rise while the
absorptions of bound (1592 cm−1) and free acetate (∼1570 cm−1)
decrease. In the final spectrum all free acetate has been converted
into acetic acid and most of the bound acetate is lost. The overall
process, i.e. immediate and delayed changes, turned out to be
reversible. Re-reduction at −0.28 V quantitatively regenerated
the spectrum of the intact II,II complex.
Fig. 5(b) shows spectra that were obtained during electrolysis
at 0.82 V of a solution containing 1% D2O added prior to elec-
trolysis. The solution of the II,II complex had been electrolyzed
before at 0.32 V to prepare the II,III state. At this water content,
reaction of the II,II and II,III states can be neglected, allowing
for the clean preparation of the intact II,III state at 0.32 V (see
Figs. 3 and 4(a)). On the other hand the reactivity of the III,III
state towards water is sufficiently high to ensure almost complete
release of one acetate per complex already with 1% of D2O (see
Fig. 5(a)). The first spectrum was taken after 20 s when the
primary electrolysis is essentially completed and the spectrum is
similar to the one obtained after subsequent addition of water
to the III,III state (see above). With the potential of 0.82 V
(Fig. 5(b)) applied to the cell further transformation of bound
acetate into acetic acid is observed. This indicates that at this
potential the products of the initial ligand-exchange reactions
undergo further oxidation that eventually triggers the release
of the residual acetate ligands. All acetate is found as acetic
acid but it can not be decided whether additional protons are
released in the course of this oxidation since already the initial
ligand-exchange reaction presumably produces two eqivalents of
protons. The position of the 1606 cm−1 band of the bpmp− ligand
is maintained indicating that the Mn ions are still coordinated.
The changes shown in Fig. 5(b) are not reversible, however. This
indicates that the oxidation products are not stable under the
conditions of the strongly oxidizing potential and/or the long
time scale of the experiment.
In summary, IR spectroscopy can monitor the loss of acetate
ligands and the deprotonation of the water derived ligands in
higher oxidation states. However, the IR spectra report average
effects of samples that consist most likely of product mixtures.
This required a complementary technique that could identify
individual product complexes.
Since no spectral changes in time were observed when D2O
was added to the II,III state after electrolysis the delayed spectral
changes can be attributed to the electrolysis of ligand-exchange
products initially formed from the II,III state. Further oxidation
could account for both the release of protons from existing aquo
or hydroxo ligands as well as further loss of bound acetate. The
sluggishness of these processes might arise from slow secondary
ligand-exchange reactions as well as slow electrolysis if the
electrode potential is not sufficiently positive to drive oxidation
of products with diffusion controlled rate.
EC–ESI-MS
To identify products of the ligand-exchange reactions an on-
line combination of electrochemistry and electrospray ionization
mass spectrometry (ESI-MS) was used. The sample solutions
were passed through a thin-layer electrochemical (EC) flow cell
(total dead volume of 1.0–1.2 ll) in 6–12 s and transferred to
the mass spectrometer via a silica transfer line (dead volume of
0.5 ll) in 3–6 s using a modification of a recently described EC–
ESI-MS setup.22 Starting from potentials where no oxidation of
[Mn2II,II(bpmp)(l-OAc)2]+ occurs the potential applied to the EC
flow cell was raised in increments of 0.10 V to a final potential
where no further change of the mass spectra was observed. Fig. 6
shows exemplary mass spectra in the m/z-range 320–390, for
open circuit condition, and for different potentials applied to
the working electrode (see ESI†for full range spectra). The figure
compares spectra obtained in both dry acetonitrile (Fig. 6(a)–
(d)) and acetonitrile–water mixture (90/10 (v/v) (Fig. 6(e)–(h)).
Based on their m/z value and the charge of the ion the peaks
were assigned according to Table 2. The first group of peaks in
the top part of the table comprises the different oxidation states
of the intact [Mn2(bpmp)(l-OAc)2]n+ while the second group of
peaks lists the ligand-exchanged complexes with the suggested
structure shown in Scheme 1. For all structures a preference of
the water-derived ligands for bridging coordination has been
assumed. Alternative structures with terminal ligands can not
be excluded from the MS data but seem less likely with the open
coordination sites.
Effect of water on the III,III state. Fig. 5(a) compares
the spectrum of [Mn2(bpmp)(l-OAc)2]3+ in dry acetonitrile to
spectra obtained in solutions with 0.1, 0.5, 2, 5, 20 and 50%
D2O added after oxidation at 0.82 V.
In the III,III state already small quantities of less than 1% D2O
result in the loss of about half of the bound acetate (see also
Fig. 3(b)) that is not found as free acetate but exclusively as
acetic acid. Further increase of the water content (1 to 50%)
does not reduce the absorption of the residual acetate ligands,
however. Assuming that the spectra can be essentially assigned
to a specific product rather than a mixture this would indicate
that the III,III state exchanges one acetate bridge much more
easily but binds the remaining acetate ligand more strongly than
the lower oxidation states.
Concomitantly with the loss of acetate ligands, a shift of the
1609 cm−1 band of the bpmp ligand to 1606 cm−1 is observed,
which is not found for the lower oxidation states. The frequency
is closer to but distinctively different from the value of the
free ligand indicating that bpmp is still ligated to the Mn ions.
The absence of free acetate indicates an at least stoichiometric
release of protons. This means, that the acetate ligands in the
III,III state are replaced by hydroxo or oxo ligands to compensate
for the higher charge of the metal centers.
In the open circuit spectrum (Fig. 6(a)) in anhydrous acetoni-
trile the dominating peaks at m/z 757.6 (see full range spectra in
ESI†) and m/z 349.3 arise from [Mn2II,II(bpmp)(l-OAc)2]+ (1a)
and [Mn2II,II(bpmp)(l-OAc)]2+ (2). The minor peak observed at
m/z 378.8 can be assigned to [Mn2II,III(bpmp)(l-OAc)2]2+ (1b).
In solution complex 2 has probably aquo ligands36 bound to the
open coordination sites that are lost in the mass spectrometer.
The fact that ligand-exchanged products were found also in
With the low D2O content required for structural changes
to occur in the III,III state the spectra could be analyzed also
in the v(O
range (2500–2700 cm−1, not shown) that becomes
–D)
obscured with the higher D2O content required in the lower
oxidation states. No absorption attributable to the v(O mode
–D)
of OD− ligands35 could be observed providing evidence for the
formation of O2− rather than OD− ligands.
1 0 3 8
D a l t o n T r a n s . , 2 0 0 5 , 1 0 3 3 – 1 0 4 1

View Article Online
Table 2 Assignments of mass spectrometry peaks
m/z
Compound
Ox. state
Structure
757.5
378.8
856.5
953.5
349.3
357.8
337.3
327.8
335.8
241.9
186.0
[Mn2(bpmp)(l-OAc)2]+
[Mn2(bpmp)(l-OAc)2]2+
[Mn2(bpmp)(l-OAc)2]2+·[ClO4]−
[Mn2(bpmp)(l-OAc)2]3+·2[ClO4]−
[Mn2(bpmp)(l-OAc)]2+
[Mn2(bpmp)(l-OAc)(l-OH)]2+
[Mn2(bpmp)(l-OH2)(l-OH)]2+
[Mn2(bpmp)(l-O)]2+
II,II
II,III
II,III
III,III
II,II
II,III
II,II
II,III
III,IV
—
1a
1b
1b
1c
2
3
4
5
[Mn2(bpmp)(l-O)2]2+
6
[N(n-C4H9)4]+ a
—
—
[Fe(Cp)2]+ b
III
a From tetrabutylammonium perchlorate supporting electrolyte. b From
ferrocene standard.
Scheme 1
At oxidizing potentials a group of peaks was arising in the
range of m/z 300–400 as shown in Fig. 6(b)–(d) and (f)–(h).
These peaks can be ascribed to dimeric Mn complexes with
the bpmp ligand and different bridging ligands (Table 2). In
Scheme 1 the suggested structures are arranged according to
increasing oxidation states (from left to right) and increasing
degree of acetate loss (top to bottom). Beside the doubly
charged complexes, minor peaks of singly charged clusters of Mn
complexes and perchlorate counter ions were found at m/z >
800. Thus, essentially all peaks arising upon oxidation in both
acetonitrile and acetonitrile–water mixtures can be ascribed to
the Mn complexes given in Table 2.
Fig. 7 summarizes the results of the EC–ESI-MS measure-
ments. The plots show the intensity of the MS peaks of the Mn
complexes as a function of potential applied to the EC cell. To
guide the eye the data points were fitted by simple functions that
follow the trend of the data. The potential axis was calibrated
with ferrocene as an external standard by defining the inflection
point for the intensity of the ferrocenium peak as 0.00 V vs.
Fc+/0. The bar in the middle indicates the potential ranges for
the oxidation states of the unaltered [Mn2(bpmp)(l-OAc)2]n+
complex, i.e. 1a, 1b and 1c as determined by voltammetry in
anhydrous acetonitrile.
The m/z 757.5 peak of the starting complex 1a is saturated
at all potentials. Instead, the isotope peak with m/z 759.4 was
plotted in Fig. 7. In this way it was ensured that all signals
shown in Fig. 7 are within the dynamic range ((9–12) × 107
counts s−1 for 30 scans) and only the most intense signals exceed
the linear range ((3–4) × 107 counts s−1 for 30 scans) of the mass
spectrometer. As the mass spectrometer could not be calibrated
for the product complexes strictly quantitative comparisons
between signals from different complexes are not possible. The
conversion efficiency of the flow cell that can be estimated
from the depletion of 1a at oxidizing potentials (Fig. 7(b)) is
comparable to the value of 50% that has been found earlier for
the same setup.22
Fig.
6 EC–ESI-MS spectra of doubly charged products from
[Mn2II,II(bpmp)(l-OAc)2]+ (45 lM) in CH3CN without applied potential
(a) and with 0.2 V (b), 0.7 V (c) and 1.2 V (d) applied to the EC cell. Mass
spectra of doubly charged products from [Mn2II,II(bpmp)(l-OAc)2]+
(97 lM) in CH3CN with 10% (v/v) H2O without applied potential (e)
and with 0.2 V (f), 0.4 V (g) and 0.6 V (h) applied to the EC cell.
neat acetonitrile is possibly due to exposure to atmospheric
humidity in the electrospray process or small quantities of water
in the acetonitrile solution. The small signal from the oxidized
complex 1b that is found in the open circuit spectra could be
due to oxidation in the electrospray process or due to oxidation
by oxygen. IR spectroelectrochemistry did not indicate major
redox impurities in the preparation of 1a as a reducing potential
did not result in any spectral change.
Oxidation in acetonitrile. The signal of the oxidation prod-
1
uct 1b (m/z 378.8) rises at potentials above E (1) and reaches
2
a plateau at ca. 0.8 V. No depletion of the signal from 1b was
1
observed even at potentials above E (2) possibly due to the
2
incomplete electrolysis resulting in comproportionation of 1c
and 1a to give 1b. This could also explain that only a very weak
signal from 1c (m/z 953, not shown) is observed at potentials
above 0.9 V.
D a l t o n T r a n s . , 2 0 0 5 , 1 0 3 3 – 1 0 4 1
1 0 3 9

View Article Online
occupying the open coordination sites in solution is in agreement
with the IR results in the II,II state (Fig. 3(a)) that demonstrate
the loss of acetate ligands without deprotonation of water-
derived ligands
The signal of the first oxidation product 1b does not rise above
the open circuit level when the potential of the first oxidation
is passed. Instead, the predominant II,III product signals arise
from 3 and 5 that might be formed by reaction of 1b with water,
consistent with the results in neat acetonitrile. Complexes 3 and
5 could be also formed to some extent from 2 which shows
some depletion above 0.1 V. The structures of 3 and 5 with
deprotonated water-derived ligands are in agreement with the
IR spectra (Fig. 4(a)) that show a mixture of free acetate and
acetic acid when 1b is exposed to water.
The III,IV complex 6 is formed clearly below the potential of
the second oxidation of the mother compound indicating that 6
is not formed from 1c in the acetonitrile–water (90/10) mixture.
Depletion of 3 and 5 around the peak of 6 suggests that their
further oxidation leads to the formation of 6. In response to
oxidation to their III,III states, 3 and 5 most probably deprotonate
(in solution also 5 has water ligands) and a product such as 7
might be formed, as suggested by the IR data. The deprotonation
could allow for further oxidation at the same or only slightly
higher potential that eventually results in the III,IV complex
6. This would explain the absence of ligand-exchanged III,III
complexes in the mass spectra even if these would undergo no
or less complete comproportionation with the starting material
due to the lower oxidation potentials compared to the unaltered
complex 1c.
From Scheme 1 it can be seen that in a given oxidation state
ligand exchange does not necessarily result in a lower overall
charge compared to the intact complex (compare 3 with 1b or 1a
with 2 and 4). The ligand-exchange reactions, however, facilitate
further oxidation that is coupled to deprotonation of the water-
derived ligands. Thus the III,IV-complex 6 has only one more
positive charge than the starting II,II-complex 1a. Similarly the
III,III-complex 7 in which the oxo bridge is deprotonated has a
lower charge than the diacetato complex 1c.
With respect to the formation of the most oxidized complex,
6, the EC–ESI-MS measurements suggest that already at a water
content of 10% the dominating pathway is the oxidation of
products such as 3 and 5, which were formed by ligand exchange
in the II,III state 1b. Only at lower water content 6 should be
formed predominantly via the intact III,III complex (1c).
Fig. 7 Intensities of EC–ESI-MS peaks assigned to manganese com-
plexes 1–6 as a function of potential applied to the EC cell in CH3CN
(a) and CH3CN with 10% (v/v) H2O (b).
The II,II complexes 2 and 4 are the only ligand-exchange
products that can be found in solutions with no applied potential
(see open circuit signal) indicating that they are formed directly
from 1a. That they are not depleted over the whole potential
range might indicate that they are formed from 1a after the EC
cell.
The formation of the II,III complexes 3 and 5 is triggered by
the first oxidation step as their signals are rising concomitantly
with 1b. As the potential for the second oxidation (1b → 1c)
is passed a III,IV complex, 6, is formed. This indicates that 1c
is a precursor of 6 but no sizable signal from the intact III,III
complex is observed and no ligand-exchanged III,III complex
is found. It is conceivable that the intact III,III complex (1c)
is sufficiently oxidizing to oxidize not only unelectrolyzed II,II
complex but possibly also ligand-exchanged III,III complexes
thereby generating the III,IV complex 6. Thus, 1c would be
consumed by a homogeneous reaction before it reaches the mass
spectrometer.
The occurrence of ligand-exchange products in the open
circuit spectrum in neat acetonitrile can be explained by either
exposure to atmospheric humidity in the electrospray process or
by small quantities of water in the acetonitrile solution. The first
possibility implies that ligand-exchange reactions occur only
after the sample passed the electrochemical cell, while the latter
includes the option of secondary electrolysis steps after ligand
exchange.
Conclusions
By the combination of electrochemistry with IR spectroscopy
and ESI-MS we were able to monitor the loss of acetate ligands
from [Mn2(bpmp)(l-OAc)2]n+ in different oxidation states and
identify products of the ligand-exchange reactions. Our results
show that the II,II and II,III states are less prone to acetate loss
in solutions of low water content while the III,III state releases
acetate ligands already with much less water. This finding is in
agreement with our earlier conclusion based on voltammetric
data that the ligand-exchange reactions with a water content
≤10% are most efficiently triggered by the second oxidation step
to the III,III state. In a given oxidation state the ligand-exchange
reactions do not lower the overall charge compared to the intact
complex but, as expected, the water-derived ligands that replace
the acetate bridges are increasingly deprotonated in the higher
oxidation states thereby facilitating further oxidation.
The di-l-oxo bridged III,IV dimer, [Mn2(bpmp)(l-
O)2]2+generated by electro-oxidation of [Mn2(bpmp)(l-OAc)2]+
can be assigned to the strongly spin coupled III,IV dimer
18,19
detected in the EPR studies
in the course of photo-
oxidation of [Mn2(bpmp)(l-OAc)2]+. The mechanism of the
formation of the III,IV dimer seems to depend on the water
content of the solution. [Mn2(bpmp)(l-O)2]2+ can be formed
via ligand-exchange reactions starting from the intact III,III
Oxidation in acetonitrile–water. In solutions with no applied
potential the II,II complex 2 is the dominating ligand-exchange
product. The structure of 2 with aquo ligands most likely
1 0 4 0
D a l t o n T r a n s . , 2 0 0 5 , 1 0 3 3 – 1 0 4 1

View Article Online
complex [Mn2(bpmp)(l-OAc)2]3+ if little water is present but
also via ligand-exchange products of the lower oxidation states
if high water content favors ligand exchange in these states.
In the latter case the potential required for the formation
of the di-l-oxo bridged III,IV dimer drops even below the
potential of the III,III/II,III couple (0.61 V vs. Fc+/0) of the
di-l-acetato complex. The formation of the III,IV dimer by
photo-oxidation with [Ru(bpy)3]2+ (0.90 V vs. Fc+/0) could be
accomplished by either pathway but under the conditions of
the EPR experiments (10–90% water) ligand exchange in the
II,II and II,III state of [Mn2(bpmp)(l-OAc)2]n+ is most likely the
dominating mechanism.
10 L. Sun, M. K. Raymond, A. Magnuson, D. LeGourrierec, M. Tamm,
M. Abrahamsson, P. H. Kenez, J. Ma˚rtensson, G. Stenhagen, L.
˚
Hammarstro¨m, S. Styring and B. Akermark, J. Inorg. Biochem., 2000,
78, 15–22.
11 D. Burdinski, E. Bothe and K. Wieghardt, Inorg. Chem., 2000, 39,
105–116.
12 M. P. Klein, K. Sauer and V. K. Yachandra, Photosynth. Res., 1993,
38, 265–277.
13 R. D. Britt, in Oxygenic Photosynthesis: The Light Reactions, ed.
D. R. Ort and C. F. Yocum, Kluwer Academic, Dordrecht, 1996, vol.
4, pp. 137–164.
14 M. J. Baldwin and V. L. Pecoraro, J. Am. Chem. Soc., 1996, 118,
11325–11326.
15 H. H. Thorp, J. E. Sarneski, G. W. Brudvig and R. H. Crabtree,
J. Am. Chem. Soc., 1989, 111, 9249–9250.
16 T. Tanase and S. J. Lippard, Inorg. Chem., 1995, 34, 4682–4690.
17 L. Dubois, R. Caspar, L. Jacquamet, P.-E. Petit, M.-F. Charlot,
C. Baffert, M.-N. Collomb, A. Deronzier and J.-M. Latour, Inorg.
Chem., 2003, 42, 4817–4827.
Acknowledgements
Stenbjo¨rn Styring, Ann Magnuson, Ping Huang (Uppsala),
Joakim Ho¨gblom (Lund), Magnus Anderlund, Bjo¨rn Akermark
and Licheng Sun (Stockholm) are gratefully acknowledged
for discussions and continuous collaboration, and M. A.
and L. S. for providing samples of [Mn2(bpmp)(l-OAc)2][ClO4]
and [Mn2(bpmp)(l-OAc)2][ClO4]2.
˚
18 P. Huang, A. Magnuson, R. Lomoth, M. Abrahamsson, M. Tamm,
˚
L. Sun, B. van Rotterdam, J. Park, L. Hammarstro¨m, B. Akermark
and S. Styring, J. Inorg. Biochem., 2002, 91, 159–172.
19 P. Huang, J. Ho¨gblom, M. F. Anderlund, L. Sun, A. Magnuson and
S. Styring, J. Inorg. Biochem., 2004, 98, 733–745.
20 V. Kertesz and G. J. Van Berkel, Electroanalysis, 2001, 13, 1425–1430.
21 F. Zhou and G. J. Van Berkel, Anal. Chem., 1995, 67, 3643–3649.
22 C. F. Bo¨kman, C. Zettersten, P. J. R. Sjo¨berg and L. Nyholm, Anal.
Chem., 2004, 76, 2017–2024.
23 H. Diril, H. R. Chang, M. J. Nilges, X. Zhang, J. A. Potenza, H. J.
Schugar, S. S. Isied and D. N. Hendrickson, J. Am. Chem. Soc., 1989,
111, 5102–5114.
Financial support from the following organisations is grate-
fully acknowledged: the Swedish Energy Agency, the Swedish
Research Council and the Swedish Foundation for Strategic
Research. L. H. gratefully acknowledges a research fellow
position from The Royal Swedish Academy of Sciences.
24 L. Tang and P. Kebarle, Anal. Chem., 1993, 65, 3654–3668.
25 P. Kebarle and L. Tang, Anal. Chem., 1993, 65, 972A–986A.
26 A. M. Bond, R. Colton, A. D’Agostino, A. J. Downard and J. C.
Traeger, Anal. Chem., 1995, 67, 1691–1695.
References
1 V. L. Pecoraro and W.-Y. Hsieh, Met. Ions Biol. Syst., 2000, 37,
429–504.
27 X. Xu, W. Lu and R. B. Cole, Anal. Chem., 1996, 68, 4244–4253.
28 M. Suzuki, M. Mikuriya, S. Murata, A. Uehara, H. Oshio, S. Kida
and K. Saito, Bull. Chem. Soc. Jpn., 1987, 60, 4305–4312.
29 G. B. Deacon and R. J. Phillips, Coord. Chem. Rev., 1980, 33, 227–
250.
30 K. Wieghardt, U. Bossek, B. Nuber, J. Weiss, J. Bonvoisin, M.
Corbella, S. E. Vitols and J. J. Girerd, J. Am. Chem. Soc., 1988,
110, 7398–7411.
2 S. Mukhopadhyay, S. K. Mandal, S. Bhaduri and W. H. Armstrong,
Chem. Rev., 2004, 104, 3981–4026.
3 J. Limburg, V. A. Szalai and G. W. Brudvig, J. Chem. Soc., Dalton
Trans., 1999, 1353–1362.
4 K. Wieghardt, Angew. Chem., 1989, 101, 1179–1198.
5 W. Ruettinger and G. C. Dismukes, Chem. Rev., 1997, 97, 1–24.
˚
6 L. Sun, L. Hammarstro¨m, B. Akermark and S. Styring, Chem. Soc.
Rev., 2001, 30, 36–49.
31 K. Wieghardt, U. Bossek, J. Bonvoisin, P. Beauvillain, J. J. Girerd, B.
Nuber, J. Weiss and J. Heinze, Angew. Chem., 1986, 98, 1026–1027.
32 K. Wieghardt, U. Bossek, D. Ventur and J. Weiss, J. Chem. Soc.,
Chem. Commun., 1985, 347–349.
33 J. C. Smith, E. Gonzalez-Vergara and J. B. Vincent, Inorg. Chim.
Acta, 1997, 255, 99–103.
7 L. Sun, H. Berglund, R. Davydov, T. Norrby, L. Hammarstro¨m, P.
Korall, A. Bo¨rje, C. Philouze, K. Berg, A. Tran, M. Andersson, G.
Stenhagen, J. Ma˚rtensson, M. Almgren, S. Styring and B. Akermark,
˚
J. Am. Chem. Soc., 1997, 119, 6996–7004.
8 M. L. A. Abrahamsson, H. B. Baudin, A. Tran, C. Philouze,
˚
K. E. Berg, M. K. Raymond-Johansson, L. Sun, B. Akermark,
34 S. Blanchard, G. Blondin, E. Riviere, M. Nierlich and J.-J. Girerd,
Inorg. Chem., 2003, 42, 4568–4578.
S. Styring and L. Hammarstro¨m, Inorg. Chem., 2002, 41, 1534–
1544.
35 Z. Shirin, V. G. Young, Jr. and A. S. Borovik, Chem. Commun., 1997,
1967–1968.
9 K. E. Berg, A. Tran, M. K. Raymond, M. Abrahamsson, J. Wolny,
S. Redon, M. Andersson, L. Sun, S. Styring, L. Hammarstro¨m,
H. Toftlund and B. Akermark, Eur. J. Inorg. Chem., 2001, 1019–
1029.
˚
36 A. Boisen, A. Hazell and C. J. McKenzie, Chem. Commun., 2001,
2136–2137.
D a l t o n T r a n s . , 2 0 0 5 , 1 0 3 3 – 1 0 4 1
1 0 4 1