Instability of the oxidation catalysts [OsIII(tpy)(bpy)(py)] 3+ and [OsIII(bpy)3]3+ in alkaline solution

Article abstract of DOI:10.1016/j.molcata.2003.11.001

The Os(III) complexes [Os^III(tpy)(bpy)(py)]³⁺ and [Os^III(bpy)₃]³⁺ (bpy: 2,2′-bipyridine, py: pyridine, tpy: 2,2′,2″-terpyridine) are unstable toward self-reduction in alkaline solution. Product analysis performed by using UV-Vis spectroscopy, cyclic voltammetry, HPLC and GC show that ca. 90% of the reduced product that appears is the unmodified polypyridyl complex of Os(II), [Os ^II(tpy)(bpy)(py)]²⁺ or [Os^II(bpy) ₃]²⁺. The remaining 10% of the products are accounted for by other Os(II) complexes, including several which appear to have undergone varying degrees of ligand oxidation or substitution. Another product of the reduction of [Os^III(tpy)(bpy)(py)]³⁺ in basic solution has been identified as [Os^II(tpy)(bpy)(OH₂)]²⁺ which results from metal reduction and concomitant pyridine ligand loss. No O₂ was detected as a product of the base-catalyzed self-reduction reaction. The Os(III) complexes are stable towards self-reduction and ligand loss in the dark in acetonitrile or aqueous solutions buffered below pH = 6. The rates observed for the self-reduction processes are sufficiently rapid at pH values above 7 that useful catalytic rate data cannot be obtained in basic solution unless the substrate of interest is sufficiently reactive and is present in large excess. Triple mixing, pH jump techniques have successfully been employed to observe reaction rates under strongly alkaline conditions.

Full text of DOI:10.1016/j.molcata.2003.11.001

Journal of Molecular Catalysis A: Chemical 212 (2004) 83–89
Instability of the oxidation catalysts [OsIII(tpy)(bpy)(py)]3+ and
[OsIII(bpy)3]3+ in alkaline solution
Estelle L. Lebeau∗
Department of Chemistry, Central Michigan University, 268 Dow Science Complex, Mount Pleasant, MI 48859, USA
Received 20 August 2003; received in revised form 31 October 2003; accepted 4 November 2003
Abstract
The Os(III) complexes [OsIII(tpy)(bpy)(py)]3+ and [OsIII(bpy)3]3+ (bpy: 2,2ꢀ-bipyridine, py: pyridine, tpy: 2,2ꢀ,2ꢀꢀ-terpyridine) are unstable
toward self-reduction in alkaline solution. Product analysis performed by using UV-Vis spectroscopy, cyclic voltammetry, HPLC and GC show
that ca. 90% of the reduced product that appears is the unmodified polypyridyl complex of Os(II), [OsII(tpy)(bpy)(py)]2+ or [OsII(bpy)3]2+
.
The remaining 10% of the products are accounted for by other Os(II) complexes, including several which appear to have undergone varying
degrees of ligand oxidation or substitution. Another product of the reduction of [OsIII(tpy)(bpy)(py)]3+ in basic solution has been identified
as [OsII(tpy)(bpy)(OH2)]2+ which results from metal reduction and concomitant pyridine ligand loss. No O2 was detected as a product of the
base-catalyzed self-reduction reaction. The Os(III) complexes are stable towards self-reduction and ligand loss in the dark in acetonitrile or
aqueous solutions buffered below pH = 6. The rates observed for the self-reduction processes are sufficiently rapid at pH values above 7 that
useful catalytic rate data cannot be obtained in basic solution unless the substrate of interest is sufficiently reactive and is present in large
excess. Triple mixing, pH jump techniques have successfully been employed to observe reaction rates under strongly alkaline conditions.
© 2003 Elsevier B.V. All rights reserved.
Keywords: Oxidation; Catalysis; Osmium; Polypyridyl; Self-reduction
step upon addition of [FeIII(bpy)3]3+ to basic solution is
hydroxide attack on the metal center. Creutz and Sutin have
observed only trace levels of O2 production, whereas Nord
and co-workers observe significant O2 production with
1. Introduction
Polypyridyl complexes containing high valent group 8
metals have been used as stoichiometric or catalytic oxi-
dants for a wide variety of organic and inorganic substrates.
Efforts have been directed toward the design of stable, yet
strongly oxidizing, homogeneous catalysts. A potentially
important variable in using these reagents as oxidants is
pH because ligand loss and ligand oxidative degradation
have been noted as primary sources of catalyst deactiva-
tion, especially in basic media. The instability arises from
self-reduction of the high valent metal center by ligand
oxidation as the pH is increased as has been reported
for many pyridyl and polypyridyl complexes, including
[RuIII(bpy)3]3+ and [FeIII(bpy)3]3+ [1–3]. Creutz and Sutin
describe the kinetics of decomposition of [RuIII(bpy)3]3+ in
basic media as dominated by rate-determining nucleophilic
attack by hydroxide on the bound bypyridine ligand. How-
ever, Nord and co-workers suggest that the rate-determining
concomitant formation of >85% [FeII(bpy)3]2+
.
Oxidation of water to dioxygen is the terminal reac-
tion of Photosystem II. Although much has been learned
about this system, little is known about the how O2 is
evolved. There are a few well defined molecules which
effect this conversion [4–6] one being the “blue dimer”,
cis,cis-[(bpy)2(H2O)RuIIIORuIII(OH2)(bpy)2]4+, and sev-
eral of its derivatives. These molecules are important be-
cause they have provided insights into possible mechanisms
for water oxidation [7–10]. Development of a water oxida-
tion catalyst that can cycle through Eqs. (1) and (2) is of
considerable interest for the development of alternative fuel
sources.
4(catalyst) + 2H2O → 4(catalyst) + O2+4H+
2(catalyst) + 2H+ → 2(catalyst)++H2
(1)
(2)
∗
For typical polypyridyl group 8 complexes, dioxygen
formation is rarely reported, however. Base-catalyzed
Tel.: +989-774-3116; fax: +989-774-3883.
E-mail address: lebea1el@cmich.edu (E.L. Lebeau).
1381-1169/$ – see front matter © 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2003.11.001

84
E.L. Lebeau / Journal of Molecular Catalysis A: Chemical 212 (2004) 83–89
self-reduction can occur without dioxygen formation, and
the nature of the oxidized products and the mechanism of
their formation is, therefore, of some interest. We have in-
vestigated and report here the results of a systematic series
of studies on the instability of [OsIII(tpy)(bpy)(py)]3+ and
[OsIII(bpy)3]3+ toward self-reduction and ligand loss in ba-
sic solution. Although a similar study with [OsIII(bpy)3]3+
has been published [11], we chose to reinvestigate the
base-catalyzed self-reduction of this substrate under iden-
tical conditions as [OsIII(tpy)(bpy)(py)]3+. The compound
[OsIII(tpy)(bpy)(py)]3+ was selected to allow for the evalu-
ation of the stability of monodentate, bidentate, or tridentate
pyridine moieties towards ligand loss and pH induced in-
tramolecular electron transfer. The goal of this work was
to evaluate the availability of [OsIII(tpy)(bpy)(py)]3+ and
other high valent polypyridyl osmium complexes to act
as oxidation catalysts in basic solution. It is an important
study because it is necessary to define pH regions of cata-
lyst stability if these compounds are to regularly be used as
one-electron, outer-sphere oxidants.
combined with 0.54 g tpy in 50 ml ethylene glycol and was
heated at reflux under Ar for 1 h. After cooling to room tem-
perature, 50 ml water and 5 ml Na2S2O4 were added to the
stirring solution. A few milliliters of saturated ammonium
hexafluorophosphate, NH4PF6, were added and the precipi-
tate was collected on a fritted funnel. The solid was purified
by elution from an alumina column using 1:1 (v:v) acetoni-
trile:toluene as eluant. The first purple/brown band eluted is
the desired product. The solvent was removed under vacuum
by rotary evaporation. The solid was further purified by dis-
solving the complex in a small volume of acetonitrile and
adding this solution dropwise to a large excess of rapidly
stirring ether. The precipitated complex was collected by fil-
tration and washed (3 × 30 ml) with fresh ether.
2.2.2. [OsIII(tpy)(bpy)(Otf)](Otf)2 (Otf is
trifluoromethanesulfonate)
A 0.20 g sample of [OsII(tpy)(bpy)(Cl)](PF6) was com-
bined with 5 g neat triflic acid in a 25 ml roundbottom flask.
The solution was stirred at reflux under Ar for 1 h, then
cooled to RT. The solid was precipitated by slowly dripping
the solution into ca. 300 ml of rapidly stirring ether. The
pale brown precipitated was collected on a glass frit and was
washed 3 × 10 ml with fresh ether. The product was dried
in a vacuum desiccator overnight.
2. Experimental
2.1. Materials
High purity acetonitrile (Burdick & Jackson) was used
as the solvent for the kinetic studies after distillation from
P2O5 with use of a Vigereux column. High quality wa-
ter was prepared by distilling house-deionized water from
alkaline potassium permanganate. (NH4)2OsCl6 was ob-
tained from Englehard Ind. 2,2ꢀ,2ꢀꢀ-Terpyridine was ob-
tained from G.F. Smith and 2,2ꢀ-bipyridine was obtained
from Aldrich. Sodium phosphate monobasic monohydrate,
NaH2PO4·H2O, sodium phosphate dibasic heptahydrate,
Na2HPO4·7H2O, sodium phosphatetribasic dodecahydrate,
Na3PO4·12H2O, and phosphoric acid, H3PO4 were ob-
tained from Fisher Scientific and were used without further
purification in the preparation of buffer solutions. All other
materials were reagent grade and used without additional
purification.
2.2.3. [OsIII(tpy)(bpy)(py)](PF6)2
It was prepared through a modification of a literature
procedure [13]. A 1.4 ml sample of pyridine was added to
5 ml of an Ar sparged solution of ethylene glycol. To this
was added 0.2 g [OsIII(tpy)(bpy)(Otf)](Otf)2. The solution
was heated at reflux under Ar for 5 h. Under an Ar at-
mosphere, the solution was cooled to room temperature.
Twenty milliliters of H2O and several milliliters of saturated
NH4PF6 were added. The solution was stirred at 0◦ for 2 h.
The precipitate was collected on a fritted funnel and was
washed with a minimum amount of water followed by ether.
The solid was purified by elution from an alumina column
using 1:1 (v:v) acetonitrile:toluene as eluant. The first brown
band eluted is the desired product. The solvent was removed
under vacuum by rotary evaporation. The solid was further
purified by dissolving the complex in a small volume of ace-
tonitrile and adding this solution dropwise to a large excess
of rapidly stirring ether. The precipitated complex was col-
lected by filtration and washed (3 × 30 ml) with fresh ether
and was air dried.
2.2. Preparations
2.2.1. [Os(bpy)Cl4] and [OsII(tpy)(bpy)(Cl)](PF6)
These compounds were prepared by slight modification
of the procedure of Buckingham et al. [12]. Typically, a 1 g
sample of (NH4)2OsCl6 was dissolved in 50 ml of 3 M HCl
and was stirred at 70 ◦C. To this was added 0.42 g bpy which
was dissolved in a minimum of 3 M HCl. The solution was
cooled to 0 ◦C for 2 h. The precipitate was collected by vac-
uum filtration using a glass frit, washed (3×10 ml each) with
3 M HCl, water, and diethyl ether. The (bpy-H2) [OsCl6]
was dried in a vacuum oven at 70 ◦C overnight, and then
was converted into [Os(bpy)Cl4] by pyrolysis in a molten
salt bath (290 ◦C for 6 h). A 1 g sample of [Os(bpy)Cl4] was
2.2.4. [OsII(tpy)(bpy)(OH2)](PF6)2
It was prepared by modification of a literature preparation
[14]. To a 0.5 g sample of [OsIII(tpy)(bpy)(Otf)](Otf)2 was
added 10 ml H2O and a zinc amalgam. The solution was
stirred under Ar for 3 h. After the zinc amalgam was sep-
arated from the solution through decantation, the addition
of several milliliters of saturated NH4PF6 solution induced
precipitation of the desired complex. The compound was pu-
rified by column chromatography using Sephadex SP-C25

E.L. Lebeau / Journal of Molecular Catalysis A: Chemical 212 (2004) 83–89
85
cation exchange as the column support and Na2SO4 as the
eluant. By use of a 0.2 M ionic strength sodium sulfate elu-
ant, a brown band was recovered from the Sephadex column.
The solution was reduced in volume to ca. 25 and 2 ml of
saturated NH4PF6 were added. The solution was cooled at
0 ◦C overnight and a dark brown product was collected by
filtration.
comparable to the GC–FID. Initial oven temperature of
40 ◦C, 250 ◦C final oven temperature (10 ◦C/min tempera-
ture ramp, 15 min hold time), 250 ◦C injection port temper-
ature, and 300 ◦C detector temperature. The injection size
was 2 ␮l.
A Bioanalytical Systems (BAS) CV-50W voltammetric
analyzer and BAS CV-50W version 1.0 software was used
for all cyclic voltammogram measurements and bulk ox-
idation experiments. In a typical electrolysis experiment
(bulk electrochemical oxidation of [OsII(tpy)(bpy)(py)]2+ or
[OsII(bpy)3]2+), a reticulated vitreous carbon electrode in a
three-arm electrolysis cell was used. After a given amount of
charge was passed, electrolysis was interrupted and the solu-
tion subjected to CV analysis using a glassy-carbon working
electrode. Electrolysis and cyclic voltammeric experiments
were conducted using a platinum-wire auxiliary electrode
and a saturated sodium calomel reference electrode (SSCE)
for aqueous solutions or Ag/Ag+ in acetonitrile. Typical CV
scan rates are 10 mV/s or 100 mV/s. An ionic strength of
0.1 M was maintained using phosphate buffers in aqueous
solutions and tetrabutyl ammonium hexafluoro phosphate
(TBAH) in acetonitrile. Buffer solutions for electrochemi-
cal measurements were prepared from aqueous perchloric
acid, HClO4, with LiClO4 (pH = 1–2), and HClO4 with
NaH2PO4·H2O, Na2HPO4·7H2O, Na3PO4·12H2O (pH =
2–12), and NaOH with Na3PO4·12H2O (pH = 12–14). The
E0.5 values reported in this work were calculated from cyclic
voltammeric waveforms as an average of the oxidative and
reductive peak potentials, (Ep,a+Ep,c)/2. All cyclic voltam-
mograms were obtained after purging with argon, and the
solutions were kept in the dark to avoid photolabilization of
the pyridine ligands.
3. Instrumentation
Routine UV-Vis spectra were recorded on a Hewlett-
Packard 8453 diode array spectrophotometer by using
standard quartz cells. The temperature of the reactant so-
lutions was controlled to 25 0.1 ◦C by using a Peltier
temperature controller. Manual mixing of solutions from
conventional syringes into standard cuvettes was employed.
The pH of solutions was determined by using a Corning
320 pH Meter with a general purpose combination electrode
(476530) after calibration with standard buffer solutions.
In the kinetic experiments the initial concentrations of
[OsII(bpy)3]2+, [OsIII(bpy)3]3+, [OsIII(tpy)(bpy)(py)]3+
,
and [OsII(tpy)(bpy)(py)]2+ were varied from 8 × 10−7 to
2 × 10−4 M. Rate data in water were collected by following
visible spectral changes at a series of pH values. Wave-
lengths were chosen where large spectral changes were ob-
served or where component absorbances could be isolated.
Triple mixing was employed at basic pHs (pH > 6) because
of the instability of [OsIII(bpy)3]3+ for extended periods
under these conditions. The pH jump experiments minimize
complications from decomposition of [OsIII(bpy)3]3+
.
Dioxygen detection was achieved by use of an Orion
Model 97-08-00 O2 electrode interfaced to a Standard
pH Meter. HPLC isolation and purity determination of
[OsII(tpy)(bpy)(py)]2+ was achieved using a HEMA cation
exchange column with either 450 nm or 290 nm detection
using a Varian Prostar 210 Model pump system with a
Varian Prostar 340 Model UV-Vis detector and a flow rate
4. Results and discussion
[OsIII(tpy)(bpy)(py)]3+ and [OsIII(bpy)3]3+ are stable
for extended periods, ca. days, in acetonitrile or aque-
ous acidic solutions below pH 6. Cyclic voltammetry and
UV-Vis measurements before and after bulk electrolysis of
[OsII(tpy)(bpy)(py)]2+ oxidation to [OsIII(tpy)(bpy)(py)]3+
or from [OsII(bpy)3]2+ to [OsIII(bpy)3]3+ are identical in
acetonitrile and under acidic conditions.
The major products (>90%) of the spontaneous re-
ductions of [OsIII(tpy)(bpy)(py)]3+ or [OsIII(bpy)3]3+ in
basic solution are the corresponding Os(II) complexes,
[OsII(tpy)(bpy)(py)]2+ or [OsII(bpy)3]2+. However, under
basic conditions waves appeared after electrolysis that were
not present in the initial solution. After acidification with
1 M HClO4 or 1 M trifluoromethanesulfonic acid (HOtf),
precipitates can be separated from each product solution
which are identified as the unmodified Os(II) complexes
by UV-Vis spectroscopy, cyclic voltammetry, 1H NMR and
HPLC. HPLC enabled the separation of the unmodified
Os(II) complexes from at least two other Os(II)-containing
side products.
1
of 0.5 ml/min. H NMR spectra were obtained in CD3CN,
D2O or CD3CN using a General Electric QE-300 MHz
FT-NMR spectrometer. Chemical shifts are reported as
ppm versus TMS at 20 ◦C. IR spectra were recorded us-
ing a Nicolet Magna-IR 560 spectrometer and were made
in KBr pellets or made in CD3CN solution by use of a
demountable cell with NaCl plates and teflon spacers or a
fused BaF2 cell with a 1 mm path length. GC measurements
were made using a Hewlett-Packard 5890 Series II Model
chromatograph or an Agilent (Hewlett-Packard) Model
6890 equipped with flame ionization detection (GC–FID,
J&W DB-5 columns (30 M × 0.32 mm with a 0.25 ␮m
film thickness). He carrier gas, 2.0 ml/min flow rate. APEX
ProSep 800XT pre-column injection system (held at 20 ◦C
for injection and then programmed to 300 ◦C at 300 ◦C/min,
10:1 split ratio). Gas chromatography–mass spectrometry
(GC–MS) was performed using a Hewlett-Packard Model
5973 GC/MS/DS or a Finnigan 700 with column conditions

86
E.L. Lebeau / Journal of Molecular Catalysis A: Chemical 212 (2004) 83–89
Fig. 1. Cyclic voltammograms at glassy-carbon electrode of deoxygenated ( ) 10−5 mol [Os(tpy)(bpy)(py)]2+ at pH = 4.5, before electrolysis; (
)
[Os(tpy)(bpy)(py)]3+ at pH = 4.5, 1 h after electrolysis at 650 mV (80 min, n = 0.97e−); and (—) [Os(tpy)(bpy)(py)]3+ at pH = 7.71, 1 h after electrolysis
and base addition. The potential scan rate was 0.1 V/s.
In a typical electrolysis experiment 10−5 to 10−6 mol
of [OsII(tpy)(bpy)(py)]2+ was dissolved in ca. 6 ml of so-
lution and electrolyzed at ≥650 mV using a reticulated
vitreous carbon electrode in a three-arm electrolysis cell.
After a given amount of charge was passed, electrolysis
was interrupted and the solution subjected to analysis by
CV using a glassy-carbon electrode placed in the working
electrode compartment. In Fig. 1 is shown a cyclic voltam-
mogram of [OsII(tpy)(bpy)(py)]2+ which was dissolved at
pH = 4.5. The originally greenish solution was oxidized
to a pinkish/purple [OsIII(tpy)(bpy)(py)]3+ by exhaustive
electrolysis at 650 mV versus SSCE (n = 0.97e−). A CV of
the [OsIII(tpy)(bpy)(py)]3+ solution at pH = 4.5 is nearly
identical to the CV before electrolysis and shows only the
waves for the OsII/III couple. This solution is stable for
several hours in the dark as shown by CV and UV-Vis anal-
ysis. However, after 1 h at pH = 7.71 (generated by base
addition to the original solution), the wave for the OsII/III
couple for [OsII(tpy)(bpy)(py)]2+ is still present, but there
is also a new wave attributed to the OsII/III couple for
[OsII(tpy)(bpy)(OH2)]2+. At slower scan rates (≤20 mV/s),
a wave for the OsIII/IV couple is also present (not
shown).
at 650 mV versus SSCE. Electrolysis was followed by base
addition and the total scan time = 10 h after base addition.
These data reveal an increased absorbance with time in the
region from 400 to 600 nm and a decreased absorbance at
260 nm (as shown in the inset). The resulting spectrum is
consistent with polypyridyl Os(II) species, however, the mo-
lar extinction coefficients (calculated from A = εbc) are
not equal to that of pure [OsII(tpy)(bpy)(py)]2+. These data
combined with the CV data of the same solution (Fig. 1)
reveal that other Os(II) species are present in the alkaline
solution.
Fig. 2. UV-Vis spectral changes observed for the base-catalyzed
self-reduction of [OsIII(tpy)(bpy)(py)]3+ at pH
=
7.71 at 25.0 ◦C.
Fig. 2 shows UV-Vis spectral changes observed for the
base-catalyzed self-reduction of [OsIII(tpy)(bpy)(py)]3+ at
pH = 7.71 at 25.0 ◦C. [OsIII(tpy)(bpy)(py)]3+ was gener-
ated from [OsII(tpy)(bpy)(py)]2+ at pH = 4.5 by electrolysis
[OsIII(tpy)(bpy)(py)]3+ was generated from [OsII(tpy)(bpy)(py)]2+ at
pH = 4.5 by electrolysis at 650 mV vs. SSCE. Electrolysis was followed
by base addition and the total scan time = 10 h after base addition. Inset
shows decrease in absorbance at λ = 260 nm.

E.L. Lebeau / Journal of Molecular Catalysis A: Chemical 212 (2004) 83–89
87
Fig. 3. 1H NMR of [OsII(tpy)(bpy)(py)]2+ in CD3CN after isolation following self-reduction reaction.
After acidification with 1 M HClO4, HPLC enabled the
separation of the unmodified Os(II) complex from at least
two other Os(II)-containing side products in this case. The
conditions, the number of electron equivalents passed never
exceeded one. However, a series of experiments were per-
formed in which the [OsII(tpy)(bpy)(py)]2+ was electrolyzed
to [OsIII(tpy)(bpy)(py)]3+ under basic conditions and each
resulted in greater than one equivalents of electrons. In one
such experiment at pH = 7.82, 5 ml of a 5.85 × 10−6 M so-
lution of [OsII(tpy)(bpy)(py)]2+ was electrolyzed at 750 mV
for 2 h using a reticulated vitreous carbon electrode in a
three-arm electrolysis cell. A total of 1.28 C, or 2.25 equiv-
alents of electrons, were passed in this time period. Fig. 4
shows the cyclic voltammograms of the solution before and
after electrolysis. In addition, a CV of an independently
prepared sample of [OsII(tpy)(bpy)(OH2)]2+ at the same
pH is shown. Before electrolysis, only the OsII/III wave of
1
aromatic region of a H NMR in CD3CN of the isolated,
unmodified [OsII(tpy)(bpy)(py)]2+ is shown in Fig. 3. One
of the side products was isolated using a Sephadex SP-C25
cation exchange column. This compound was identified
as [OsII(tpy)(bpy)(OH2)]2+ by UV-Vis measurements,
consistent with the CV identification, and results from
pyridine labilization, solvent ligation, and metal center re-
duction. No O2 production was detected as a product of the
base-catalyzed self-reduction.
In all cases where the electrolysis of [OsII(tpy)(bpy)(py)]2+
to [OsIII(tpy)(bpy)(py)]3+ was performed under acidic
Fig. 4. Cyclic voltammograms at glassy-carbon electrode of deoxygenated 10−5 mol [OsII(tpy)(bpy)(py)]2+ at pH = 7.82 before electrolysis ( ) and after
electrolysis (—) at 750 mV with 2.25 equivalents of e− passed. Also shown is the CV of an authentically prepared sample of [OsII(tpy)(bpy)(OH2)]2+
at pH = 7.82 ( ). The potential scan rate was 0.1 V/s.

88
E.L. Lebeau / Journal of Molecular Catalysis A: Chemical 212 (2004) 83–89
the [OsII(tpy)(bpy)(py)]2+ complex is observed. After elec-
trolysis, that wave remains, but there are additional waves
clearly present that correspond to the OsII/III and OsIII/IV
waves of [OsII(tpy)(bpy)(OH2)]2+. Literature reports [15]
have suggested that the complex [OsIV(tpy)(bpy)(O)]2+
will lose a bpy ligand once oxidized to the OsV form,
[OsV(tpy)(bpy)(O)]2+. This accounts for the irreversibility
of the OsIV/V wave at ca. 0.95 V. The waves at E0.5 = −173
and −315 mV are the OsII/III and OsIII/IV couples of the
complex [OsII(tpy)(OH2)3]2+. Aside from these voltammet-
ric, UV-Vis, and HPLC observations, no further attempts
were made to isolate or characterize the various products.
Experiments were conducted at a series of more basic pH
values producing similar results when [OsII(tpy)(bpy)(py)]2+
was electrolyzed under basic conditions. A greater number
of electron equivalents were passed in a given time period
as the solutions were buffered at more basic pH values.
One explanation is that the rate of pyridine substitution by
hydroxide is increased as the pH is increased. At pH = 7,
there is a driving force of approximately 200 mV for the dis-
proportionation reaction shown in Eq. (3). The reaction be-
tween [OsII(tpy)(bpy)(OH2)]2+ and [OsIII(tpy)(bpy)(py)]3+
has a driving force of approximately 500 mV and results
in the electron transfer reaction shown in equation 4. It is
likely that much of the unmodified [OsII(tpy)(bpy)(py)]2+
results from a combination of these two reactions.
are susceptible to attack by hydroxide, particularly when
the metal is in an oxidized form. Once an initial oxidation
has been performed on a polypyridyl ligand, it is activated
towards further oxidation. The mechanism is reported to
involve nucleophilic addition of water or hydroxide to a
carbon atom of the activated aromatic amine ligand. High
valent metals activate the aromatic amine moiety and OH−
can act as a nucleophile. Although it would require extreme
conditions, a fully oxidized bpy ligand could account for up
to 60 electron equivalents, Eq. (5), or 58 e− per bpy, Eq. (6).
C10H10N2+15O2 → 10CO2+2HNO3+4H2O
2C10H10N2+29O2 → 20CO2+4NO2+10H2O
(5)
(6)
We could not identify individual ligand oxidation prod-
ucts, but believe (anecdotally) that it initiates the reactiv-
ity observed. [OsIV(tpy)(bpy)(O)]2+ and [OsIII(tpy)(bpy)
(OH)]2+ would be reduced to the observed [OsII(tpy)(bpy)
(OH2)]2+ species by oxidizing a ligand moiety.
5. Conclusions
Spontaneous reduction of [OsIII(tpy)(bpy)(py)]3+ and
[OsIII(bpy)3]3+ in water is promoted by basic conditions.
At high pH, a fairly simple, first-order dependence of
the reduction rate on [OH−] is found. Given the absence
of O2 formation in the base-catalyzed self-reduction of
[OsIII(tpy)(bpy)(py)]3+, the concurrent formation of sev-
eral Os-containing compounds, and the >90% yield of
[OsII(tpy)(bpy)(py)]2+, we conclude that the initial redox
chemistry observed here is ligand-based. The apparent lig-
and decomposition reactions observed in this work (and
elsewhere) suggest that polypyridine ligands are not always
merely “spectator” ligands, especially under alkaline con-
ditions, however useful catalytic rate data can be obtained
in basic solution if the substrate of interest is sufficiently
reactive and is present in large excess.
[OsIII(tpy)(bpy)(OH)]2++[OsIII(tpy)(bpy)(py)]3+
→ [OsIV(tpy)(bpy)(O)]2++[OsII(tpy)(bpy)(py)]2+ (3)
[OsII(tpy)(bpy)(OH2)]2++[OsIII(tpy)(bpy)(py)]3+
→ [OsIII(tpy)(bpy)(OH)]2++[OsII(tpy)(bpy)(py)]2+
(4)
Dioxygen production was not observed as a prod-
uct of exhaustive, prolonged oxidative electrolysis of
[OsII(tpy)(bpy)(py)]2+ in 0.1 M NaOH.
It is clear, however, that the pyridine ligand must be
replaced by OH− to initiate the reactions in 3 and 4. A
first-order dependence of the rate of substitution on hydrox-
ide was determined to ca. pH = 11 by a series of triple
mixing pH jump experiments (using the initial rate method).
Beyond pH = 11, the rate of substitution and self-reduction
became to fast to be observed by conventional diode array
UV-Vis techniques.
Acknowledgements
Acknowledgment is made to the Central Michigan Uni-
versity Faculty Research and Creative Endeavors Committee
for support of this research.
Proton loss from an aqua or hydroxy ligand upon ox-
idation prevents build-up of charge on the complex from
multiple electron loss at the metal and also allows the ba-
sic coordination environment for the metal in the different
oxidation states to remain the same throughout a catalytic
cycle. Because of the increased capacity for ␲-donation
from the lone pairs of the hydroxy or oxo ligands, the higher
oxidation states of the metal are stabilized as compared to
with the pyridine ligand.
References
[1] C. Creutz, N. Sutin, Proc. Nat. Acad. Sci. U.S.A. 72 (1975) 2858–
2862.
[2] L.R. Roecker, W. Kutner, J.A. Gilbert, M. Simmons, R.W. Murray,
T. Meyer, J. Inorg. Chem. 24 (1985) 3284.
[3] G. Nord, B. Pedersen, E. Bjergbakke, J. Am. Chem. Soc. 105 (1983)
1913–1919.
Gillard and Irving [16] and Sutin and co-workers [17]
have presented evidence that bpy and phen type ligands
[4] R. Manchanda, G. Brudvig, R.H. Crabtree, Coord. Chem. Rev. 144
(1995) 1–38, and references therein.

E.L. Lebeau / Journal of Molecular Catalysis A: Chemical 212 (2004) 83–89
89
[5] V.L. Pecoraro, M.J. Baldwin, A. Gelasco, Chem. Rev. 94 (1994)
807–826, and references therein.
[12] D.A. Buckingham, F.P. Dwyer, A.M. Sargeson, Aust. J. Chem. 17
(1964) 622–631.
[6] W. Rüttinger, G.C. Dismukes, Chem. Rev. 97 (1997) 1–24.
[7] C.W. Chronister, R.A. Binstead, J. Ni, T. Meyer, J. Inorg. Chem. 36
(1997) 3814–3815.
[8] D. Geselowitz, T. Meyer, J. Inorg. Chem. 29 (1990) 3894–3896.
[9] Y. Lei, J.K. Hurst, Inorg. Chim. Acta. 226 (1994) 179–185.
[10] P. Comte, M.K. Nazeeruddin, F.P. Rotzinger, A.J. Frank, M. Grätzel,
J. Mol. Cat. 52 (1989) 63–84.
[13] E.M. Kober, J.V. Caspar, B.P. Sullivan, T. Meyer, J. Inorg. Chem.
27 (1988) 4587–4598.
[14] K.J. Takeuchi, M.S. Thompson, D.W. Pipes, T. Meyer, J. Inorg.
Chem. 23 (1984) 1845–1851.
[15] K.J. Takeuchi, M.S. Thompson, D.W. Pipes, T. Meyer, J. Inorg.
Chem. 23 (1984) 1845–1851.
[16] R.D. Gillard, H.M. Irving, Chem. Rev. 65 (5) (1965) 603–616.
[17] P.K. Gosh, B.S. Brunschwig, M. Chou, C. Creutz, N. Sutin, J. Am.
Chem. Soc. 106 (1984) 4772–4783.
[11] G. Nord, O. Wernberg, J. Chem. Soc., Dalton Trans. (1975) 845–
849.