Kinetic study of 2-propanol and benzyl alcohol oxidation by alkaline hexacyanoferrate (III) catalyzed by a terpyridyl ruthenium complex

Article abstract of DOI:10.1002/1097-4601(2000)32:12<760::AID-KIN3>3.0.CO;2-7

The complex (Trpy)RuCl₃ (Trpy = 2,2′:6′,2″-terpyridine) reacts with alkaline hexacyanoferrate(III) to form a terpyridyl ruthenium(IV)-oxo complex that catalyzes the oxidation of 2-propanol and benzyl alcohol by alkaline hexacyanoferrate(III). The reaction kinetics of this catalytic oxidation have been studied photometrically. The reaction rate shows a first-order dependence on [Ru(IV)], a zero-order dependence on [hexacyanoferrate(III)], a fractional order in [substrate], and a fractional inverse order in [HO^-]. The kinetic data suggest a reaction mechanism in which the catalytic species and its protonated form oxidize the uncoordinated alcohol in parallel slow steps. Isotope effects, substituent effects, and product studies suggest that both species oxidize alcohol through similar pericyclic processes. The reduced catalytic intermediates react rapidly with hexacyanoferrate(III) and hydroxide to reform the unprotonated catalytic species.

Full text of DOI:10.1002/1097-4601(2000)32:12<760::AID-KIN3>3.0.CO;2-7

Kinetic Study of 2-Propanol
and Benzyl Alcohol
Oxidation by Alkaline
Hexacyanoferrate(III)
Catalyzed by a Terpyridyl
Ruthenium Complex
ERIC P. KELSON, PROMA P. PHENGSY
Department of Chemistry, California State University at Northridge, Northridge, CA 91330, USA
Received 8 February 2000; accepted 20 June 2000
ABSTRACT: The complex (Trpy)RuCl3 (Trpy ϭ 2,2Ј:6Ј,2Љ-terpyridine) reacts with alkaline hexa-
cyanoferrate(III) to form a terpyridyl ruthenium(IV)-oxo complex that catalyzes the oxidation
of 2-propanol and benzyl alcohol by alkaline hexacyanoferrate(III). The reaction kinetics of this
catalytic oxidation have been studied photometrically. The reaction rate shows a first-order
dependence on [Ru(IV)], a zero-order dependence on [hexacyanoferrate(III)], a fractional order
in [substrate], and a fractional inverse order in [HOϪ]. The kinetic data suggest a reaction
mechanism in which the catalytic species and its protonated form oxidize the uncoordinated
alcohol in parallel slow steps. Isotope effects, substituent effects, and product studies suggest
that both species oxidize alcohol through similar pericyclic processes. The reduced catalytic
intermediates react rapidly with hexacyanoferrate(III) and hydroxide to reform the unproton-
ated catalytic species. ᭧ 2000 John Wiley & Sons, Inc. Int J Chem Kinet 32: 760–770, 2000
INTRODUCTION
gest that the surprising mild and selective activity of
the catalyst comes from the coordination of the alcohol
to ruthenium followed by an inner-sphere oxidation of
the alcohol [2–5].
Simple oxo-complexes of ruthenium including ruth-
enate, perruthenate, and ruthenium tetroxide have been
extensively studied for their ability to catalyze the ox-
idation of alcohols by inexpensive co-oxidants [1]. A
particularly interesting aspect of this reactivity is the
capability to utilize relatively mild co-oxidants such
as hexacyanoferrate(III) under alkaline conditions [2–
5]. The reported kinetics of ruthenate-catalyzed alco-
hol oxidation by alkaline hexacyanoferrate(III) sug-
Polypyridine donors offer a means of gaining lig-
and control over high-oxidation-state ruthenium
chemistry. Although polypyridine ruthenium com-
plexes are well known for the stoichiometric, catalytic,
and electrocatalytic oxidation of alcohols in aqueous
and nonaqueous media [6–8], this chemistry has yet
to be investigated under alkaline conditions. Polypyr-
idine ligand systems have several advantages. The
synthetic chemistry of these systems is well estab-
lished, which allows ligand variations to be made eas-
Correspondence to: E. P. Kelson (kelson@csun.edu)
᭧ 2000 John Wiley & Sons, Inc.

2-PROPANOL AND BENZYL ALCOHOL OXIDATION BY ALKALINE HEXACYANOFERRATE(III)
761
ily. Ligand variations have been shown to influence
the electrochemical potential of ruthenium-oxo com-
plexes and the specifics of their activity. Polypyridine
ruthenium complexes are also notable for their inher-
ent stability, which makes them well suited for kinetic
investigations.
SPECFIT/32 (Spectrum Software Associates) running
in Windows 95. Further kinetic analyses were carried
out in Kaleidagraph 3.0 (Abelbeck Software) in Mac-
intosh OS 8.5.
We report here the formation of a high-oxidation-
state ruthenium-oxo species from (Trpy)RuCl3 (Trpy
is 2,2Ј:6Ј,2Љ-terpyridine) and describe the kinetics of
its catalytic activity toward alcohol oxidation by hex-
acyanoferrate(III) in aqueous hydroxide.
Products and Stoichiometry
Products and yields from the oxidation of benzyl, 4-
methoxybenzyl, and 4-nitrobenzyl alcohols were mea-
sured as follows. The precatalyst (Trpy)RuCl3 (2.6 mg,
5.9 ␮mol) was added to 10.0 mL of 0.30 M NaOH
and dissolved by stirring in a 45ЊC water bath for 5
min. Then 10.0 mL of 12.0 mM K3Fe(CN)6 was
added, and the solution was stirred in a 35ЊC water
bath for 30 minutes to complete formation of the cat-
alytic species. The resulting solution was then diluted
with 30.0 mL of water, and additional K3Fe(CN)6
(0.388 g, 1.18 mmol) was added and dissolved. Then
10.0 mL of the catalyst/oxidant solution and 10 mL of
the substrate solution (see Table I) were thermally
equilibrated at 25ЊC and mixed. After 30 min, the re-
action was carefully acidified with 0.30 M HCl and
extracted three times with 10 mL portions of CH2Cl2.
The extracts were dried with Na2SO4 and analyzed by
GC-FID. The benzyl alcohols and the corresponding
benzaldehyde and benzoic acid products (Table I)
were identified by their retention times.
Products from the oxidation of cyclopropyl meth-
anol and cyclobutanol were determined as follows.
The precatalyst (Trpy)RuCl3 (2.6 mg, 5.9 ␮mol) was
added to 1.00 mL of 0.30 M NaOH and dissolved by
stirring in a 45ЊC water bath for 10 minutes. Then 3.00
mL of water and 1.00 mL of 0.156 M K3Fe(CN)6 were
sequentially added to the catalyst solution, and the cat-
alyst/oxidant solution was stirred in a 35ЊC water bath
for 30 min. Then 30 ␮L of the substrate (cyclopropyl
methanol or cyclobutanol) was added, and the reaction
was stirred in a 35ЊC bath. On completion of the re-
action (within 5 min), the solution was carefully acid-
ified with 0.30 M HClO4, and the resulting solution
EXPERIMENTAL
Materials and Instrumentation
House deionized water was repurified with a Barn-
stead Super-Q system. Deuterium oxide, 99.8% (Cam-
bridge Isotope Laboratories), potassium hexacyano-
ferrate(III) (Aldrich), and sodium hydroxide (Fisher
Scientific), sodium perchlorate (Alfa-Aesar) were
used without additional purification. The complex
(Trpy)RuCl3 was prepared as previously reported from
the reaction of RuCl3 · x HO (Alfa-Aesar) and Trpy
2
(Aldrich) in ethanol [9]. Dideuteriophenylcarbinol
(benzyl alcohol-␣,␣-d2), cyclopropyl methanol, and
cyclobutanol were obtained from Aldrich and used
without further purification. All other substrates were
obtained from the Aldrich Chemical Co. and purified
by vacuum distillation or recrystallization. Purities
were verified by GC-FID. All other materials were re-
agent grade and were used without additional purifi-
cation.
UV-visible spectra were recorded on a Hewlett
Packard 8452 diode array spectrophotometer with a
Peltier thermostatted cell holder. GC-FID analyses
were carried out with a Hewlett Packard 5890 Series
II gas chromatograph with a HP-Waxcolumn (30 m
length, 0.32-mm diameter, 0.25-␮m film thickness).
Global Factor Analysis of spectra was performed with
Table I Oxidation of Benzyl Alcohols by the (Trpy)RuCl3/Fe(CN)6 Ϫ/HOϪ System after 30 Min at 25ЊCa
3
Substrate
[ROH]0/[FeIII]0
Product
Turnovers
% Yield
Benzyl alcohol
Benzyl alcohol
1.0
0.5
Benzaldehyde
Benzaldehyde
Benzoic acid
101
92
6
93
84
6
4-Methoxybenzyl alcohol
4-Nitrobenzyl alcohol
1.0
1.0
4-Methoxybenzaldehyde
4-Nitrobenzaldehyde
4-Nitrobenzoic acid
103
37
62
95
34
57
5
3
2
2
a Reactions carried out in water with [Ru]T ϭ 5.9 · 10Ϫ M; [K3Fe(CN)6 Ϫ] ϭ 1.28 · 10Ϫ M; [NaOH] ϭ 3.0 · 10Ϫ M. Yields based on
amount of oxidant consumed. Turnovers based on stoichiometries in Eqs. (1) and (2) in which benzoic acids are formed sequentially through
benzaldehyde through two turnovers of catalyst.

762
KELSON AND PHENGSY
was analyzed by GC-FID directly. The carbonyl prod-
ucts (cyclopropanecarbaldehyde and cyclobutanone)
were confirmed by their retention times.
of benzyl alcohols. The catalyst/oxidant/hydroxide so-
lution was prepared by the combination of stock so-
lutions to give typical reaction concentrations of 0.100
mM catalyst, 1.7 mM Fe(CN)6 Ϫ, and 50 mM OHϪ.
3
The oxidation of triphenylphosphine was con-
firmed and quantified as follows. The precatalyst
(Trpy)RuCl3 (2.0 mg, 4.5 ␮mol) was added to 2.0 mL
of 0.172 M NaOH and dissolved by stirring in a 45ЊC
water bath for 5 min. Then 50 mL of 0.172 M NaOH
and 50 mL of 4.3 mM K3Fe(CN)6 were sequentially
added to the catalyst solution, and the catalyst/oxidant
solution was stirred in a 35ЊC water bath for 30 min.
Then freshly ground PPh3 (0.137 g, 0.52 mmol) was
added, and the reaction was stirred vigorously in a
The ionic strength of each reaction was adjusted to
␮ ϭ 0.50 M by the appropriate amount of NaClO4. To
ensure the reproducible formation of the catalytic spe-
cies from (Trpy)RuCl3, the catalyst solution was pre-
pared as follows. The precatalyst (Trpy)RuCl3 (2.6
mg, 5.9 ␮mol) was added to 10.0 mL of 0.30 M NaOH
and stirred in a 45ЊC water bath for 5 min. Then 10.0
mL of 12 mM K3Fe(CN)6 was added, and the solution
was stirred in a 35ЊC water bath for 30 min. During
45ЊC oil bath until the yellow color of Fe(CN)6 Ϫ dis-
this time, three equivalents of Fe(CN)6 are con-
sumed by the ruthenium complexto form the catalytic
intermediate. This procedure gives a solution that is
3
3Ϫ
appeared. Then a 3.0-mL portion of the resulting sus-
pension was extracted with 1.0 mL of CH2Cl2, and the
extract was analyzed by GC-FID. Triphenylphosphine
oxide product and unreacted triphenylphosphine were
identified by their retention times and quantified
through FID integrations.
The dissociation of chloride ligands from
(Trpy)RuCl3 in NaOH was quantified as follows. The
complex(Trpy)RuCl 3 (0.100 g, 0.23 mmol) was added
to 38 mL of 0.30 M NaOH and dissolved by stirring
in a 45ЊC water bath for 10 min. The reaction was
quickly acidified with 1.0 M HClO4 and treated with
2.0 mL of 0.57 M AgNO3 (1.14 mmol). The resulting
AgCl was immediately isolated by centrifugation,
washed, and dried in vacuo. This afforded 0.0923 g of
AgCl, which corresponds to a 95% yield, assuming all
three chlorides of (Trpy)RuCl3 dissociate.
0.30 mM in catalyst, 5.1 mM in Fe(CN)6 Ϫ, and 0.15
3
M in OHϪ.
Purging reaction solutions with argon had no ap-
parent effect on reaction kinetics, so reactions were
not deoxygenated.
RESULTS
Stoichiometry and Products
As shown in Table I, the precatalyst (Trpy)RuCl3 in
alkaline hexacyanoferrate(III) effectively oxidizes
benzyl alcohols to benzaldehyde and acid products.
Oxidation of excess benzyl and 4-methoxybenzyl al-
cohols exclusively produces the corresponding alde-
hydes according to the stoichiometry in Eq. (1).
Kinetic Measurements
3
C6H5CH2OH ϩ 2Fe(CN)6 Ϫ ϩ 2OHϪ !:
Rate data for the consumption of the oxidant
3
C6H5CHO ϩ 2Fe(CN)6 Ϫ ϩ 2H2O (1)
4
Fe(CN)6 Ϫ were collected by following absorbance de-
creases in its ␭max at 420 nm. Besides the decay of this
Stoichiometric oxidation of benzyl alcohol affords a
small yield of benzoic acid that is likely formed from
oxidation of product benzaldehyde according to the
stoichiometry in Eq. (2).
peak, no other changes occur in the reaction spectra
3Ϫ
until the Fe(CN)6 is almost completely consumed.
Further, the decay of 420 nm vs. time is linear for most
of the reaction. For these reasons, initial reaction rates
were calculated through the relation
␯
ϭ
obs
3
3
C6H5CHO ϩ 2Fe(CN)6 Ϫ ϩ 2OHϪ !:
Ϫ␦[Fe(CN)6 Ϫ]/␦t ϭ Ϫ(␦A420 /␦t)/(␧III Ϫ ␧II)b in which
␦A420 /␦t was the least-squares slope for the absorbance
decay at 420 nm, ␧III and ␧II were the absorptivities of
C6H5CO2H ϩ 2Fe(CN)6 Ϫ ϩ 2H2O (2)
4
3
Ϫ
4Ϫ
Fe(CN)6 and Fe(CN)6 at 420 nm (measured as
The oxidation of twice the stoichiometric amount of
4-nitrobenzyl alcohol affords a significant amount of
the benzoic acid, indicating the electron-withdrawing
nitro group encourages benzaldehyde oxidation. Elec-
tron-withdrawing groups have also been reported to
accelerate the oxidation of substituted benzaldehydes
[10] and furfurals [11] by alkaline permanganate. The
electron-withdrawing substituent may be encouraging
the reaction of 4-nitrobenzaldehyde with hydroxide to
1.04 · 103 MϪ cmϪ , and 1.41 MϪ cmϪ , respectively),
1
1
1
1
and b was the cell path length (1.00 cm).
A typical kinetic run was initiated by the addition
of 0.50 mL of alcohol solution to 2.5 mL of solution
containing the catalyst, K3Fe(CN)6, NaOH, and
NaClO4. The alcohol solutions were prepared volu-
metrically to provide typical reaction concentrations
2
4
of 2.0 · 10Ϫ M of aliphatic alcohols or 6.7 · 10Ϫ
M

2-PROPANOL AND BENZYL ALCOHOL OXIDATION BY ALKALINE HEXACYANOFERRATE(III)
763
of hexacyanoferrate(III) and reappears upon the ad-
dition of more oxidant. This peak at 472 nm likely
represents the oxidized form of the catalyst that pre-
dominates in the presence of oxidant. Since the initial
spectral changes during catalytic reactions appear to
be only the consumption of hexacyanoferrate(III), the
early progress of the reaction was easily monitored
through the decay in the absorption at 420 nm.
Kinetics
3
؊
Effect of Fe(CN)6 Concentration. When an excess
of hexacyanoferrate(III) is present with respect to ru-
thenium, the decay of the 420 nm absorbance of hex-
acyanoferrate(III) is linear, indicating a zero-order de-
pendence on the oxidant (Fig. 1). This linearity was
unaffected by reaction conditions used in this study.
This made measurement of initial reaction rates
straightforward from the initial slope of absorption
versus time data. Typical initial reaction rates are listed
in Table II.
Figure 1 Absorbance trace (420 nm) vs. time plot for the
oxidation of 2-propanol catalyzed by the (Trpy)RuCl3/
3
Ϫ
Ϫ
Fe(CN)6 /HO system in water, [2-propanol] ϭ 5.2 ·
Ϫ
2
Ϫ4
10 M, [Ru]T ϭ 1.00 · 10 M, [K3Fe(CN)6] ϭ 1.7 ·
Ϫ3
Ϫ2
10 M, [NaOH] ϭ 5.0 · 10 M, [NaClO4] ϭ 0.45 M at
35ЊC.
form the 4-NO2C6H4CH(OH)(O)Ϫ anion. Rapid oxi-
dation of this electron- rich intermediate would pro-
duce the observed benzoic acid. The mechanism of
aldehyde oxidation by this catalytic system is under
investigation.
Effect of Catalyst Concentration. Plots of ␯obs vs. cat-
alyst concentration are linear with zero-intercepts(Fig.
2) over the concentrations used in this study. This con-
firms that the reaction is first order to catalyst concen-
tration and the rate of the uncatalyzed reaction is neg-
ligible compared to the catalyzed oxidation.
Spectral Changes
The precatalyst 1 dissolves into alkaline hexacyano-
ferrate(III) to form a complexwith a ␭ at 472 nm
Effect of Substrate Concentration. At low concentra-
tions of 2-propanol and benzyl alcohol, the initial rates
are proportional to [alcohol]; at higher concentrations
of both alcohols saturation-type kinetics are observed.
max
(␧ ϭ 1.8 · 103 MϪ cmϪ ). The addition of alcohol to a
solution of (Trpy)RuCl3 in alkaline hexacyanofer-
rate(III) results in the decay of the ␭max of hexacyano-
ferrate(III) at 420 nm. In the early stages of the cata-
lytic reaction, no other changes occur in the spectrum.
The peak at 472 nm disappears shortly after depletion
1
1
Plots of 1/␯ vs. 1/[alcohol] for 2-propanol (Fig. 3)
obs
and benzyl alcohol (Fig. 4) are linear with positive
slopes and small intercepts, consistent with Eq. (3).
Table II Representative Zero-Order Rate Constants for Alcohol Oxidation by the
(Trpy)RuCl3/Fe(CN)6 Ϫ/HOϪ Systema
3
Ϫ6
Ϫ1
Substrate
[Substrate] (M)
␯
obs (10 M · s )
Ϫ2
Ϫ2
Ϫ4
Ϫ4
Ϫ4
Ϫ4
Ϫ4
Ϫ4
Ϫ4
Ϫ4
2-Propyl alcohol
2-Propyl alcohol-d8
Benzyl alcohol
Benzyl-␣,␣-d2 alcohol
4-Methoxybenzyl alcohol
4-Methylbenzyl alcohol
4-Fluorobenzyl alcohol
4-Chlorobenzyl alcohol
4-Trifluoromethylbenzyl alcohol
4-Nitrobenzyl alcohol
2.0 · 10
2.0 · 10
6.7 · 10
6.7 · 10
6.7 · 10
6.7 · 10
6.7 · 10
6.7 · 10
6.7 · 10
6.7 · 10
4.4
0.89
3.38
0.80
3.52
3.45
3.06
3.69
4.02
6.29
a Reactions carried out in water with [Ru]T ϭ 1.0 · 10Ϫ4 M; [K3Fe(CN)6] ϭ 1.7 · 10Ϫ3 M; [NaOH] ϭ
5.0 · 10Ϫ2 M; [NaClO4] ϭ 0.45 M at 25ЊC for benzyl alcohols and 35ЊC for 2-propyl alcohols.

764
KELSON AND PHENGSY
Figure 4 Plot of 1/␯ vs. 1/[BnzOH] for benzyl alcohol
obs
Figure 2 Plot of ␯ vs. [Ru]T for 2-propanol oxidation
obs
3Ϫ
Ϫ
3Ϫ
Ϫ
oxidation catalyzed by the (Trpy)RuCl3/Fe(CN)6 /HO
catalyzed by the (Trpy)RuCl3/Fe(CN)6 /HO system in
Ϫ
5
Ϫ
3
system in water, [Ru]T ϭ 1.6 · 10 M, [K3Fe(CN)6] ϭ
water, [K3Fe(CN)6] ϭ 1.7 · 10 M, [2-propanol] ϭ 2.2 ·
Ϫ3
Ϫ2
Ϫ2
Ϫ2
2.0 · 10 M, [NaOH] ϭ 5.0 · 10 M, [NaClO4] ϭ 0.45 M
at 25ЊC.
10 M, [NaOH] ϭ 5.0 · 10 M, [NaClO4] ϭ 0.45 M at
35ЊC.
1
a2
linear with a positive slope and a small intercept con-
sistent with Eq. (4).
ϭ a1 ϩ
(3)
␯
[alcohol]
obs
b2
[HOϪ]
b1[HOϪ] ϩ b2
Although this suggests alcohol coordination is in-
volved in the reaction mechanism, the initial rates for
2-propanol oxidation are not affected by large concen-
trations of added tert-butanol (up to 0.2 M). High con-
centrations of 2-propanol also do not change the 472
nm peak in the reaction spectra in any noticeable way.
␯
ϭ b1 ϩ
ϭ
(4)
obs
[HOϪ]
The plot of ␯ vs. 1/[HOϪ] for benzyl alcohol (Fig.
6) was not linear, but a plot of ␯ vs. 1/[HOϪ] was
obs
obs
easily fit to Eq. (5).
c1(1/[HOϪ]) ϩ c2 c1 ϩ c2[HOϪ]
Effect of Hydroxide Concentration. The initial rates
for the oxidation of 2-propanol and benzyl alcohol in-
crease with decreasing concentrations of hydroxide.
The plot of ␯obs vs. 1/[HOϪ] for 2-propanol (Fig. 5) is
␯
ϭ
ϭ
(5)
obs
c3(1/[HOϪ]) ϩ 1
c3 ϩ [HOϪ]
Although reaction rates were very fast at low hydrox-
Ϫ
Figure 3 Plot of 1/␯ vs. 1/[2-PrOH] for 2-propanol ox-
Figure 5 Plot of ␯ vs. 1/[OH ] for the oxidation of 2-
obs
obs
3Ϫ
Ϫ
3Ϫ
Ϫ
idation catalyzed by the (Trpy)RuCl3/Fe(CN)6 /HO sys-
propanol catalyzed by the (Trpy)RuCl3/Fe(CN)6 /HO sys-
tem in water, [Ru]T ϭ 1.0 · 10 M, [K3Fe(CN)6] ϭ 1.7 ·
10 3 M, [2-propanol] ϭ 2.2 · 10 2 M, ␮ ϭ 0.50 M at 35ЊC.
Ϫ
4
Ϫ4
tem in water, [Ru]T ϭ 1.0 · 10 M, [K3Fe(CN)6] ϭ 1.7 ·
Ϫ3
Ϫ
2
Ϫ
Ϫ
10 M, [NaOH] ϭ 5.0 · 10 M, [NaClO4] ϭ 0.45 M at
35ЊC.
(NaClO4 added to adjust ionic strength to 0.50 M.)

2-PROPANOL AND BENZYL ALCOHOL OXIDATION BY ALKALINE HEXACYANOFERRATE(III)
765
Figure 7 Plot of log(␯obs) vs. Hammett sigma for the oxi-
dation of 4-substituted benzyl alcohols by the (Trpy)RuCl3/
Ϫ
Figure 6 Plot of ␯ vs. 1/[OH ] for the oxidation of (a)
obs
4-NO2C6H4CH2OH, (b) 4-CH3OC6H4CH2OH, and (c)
C6H5CH2OH catalyzed by the (Trpy)RuCl3/Fe(CN)6 /HO
system in water, [Ru]T ϭ 1.00 · 10 M, [K3Fe(CN)6] ϭ
1.7 · 10 M, [Alc.] ϭ 6.7 · 10 M, ␮ ϭ 0.50 M at 25ЊC.
3
Ϫ
Ϫ
Ϫ4
Fe(CN)6 /HO system in water, [Ru]T ϭ 1.00 · 10 M,
3Ϫ
Ϫ
Ϫ
3
Ϫ
Ϫ2
[K3Fe(CN)6] ϭ 1.7 · 10
[NaClO4] ϭ 0.45 M at 25ЊC.
M, [HO ] ϭ 5.0 · 10
M,
Ϫ
4
Ϫ3
Ϫ4
(NaClO4 added to adjust ionic strength to 0.50 M.)
dation of 4-nitrobenzyl and 4-methoxybenzyl alcohols
exhibited the same dependencies on hexacyanofer-
rate(III), alcohol, and hydroxide concentrations, as ob-
served for benzyl alcohol.
ide concentrations, catalyst turnover numbers rapidly
fell below 10 at pH Ͻ 11.7. For this reason, kinetic
measurements were typically made at pH ϭ 12.7,
where turnover numbers were sufficient to fully con-
sume the alcohol many times.
Free Radical Detection. Free radicals are probable
intermediates from the one-electron oxidation of al-
cohols. The oxidation of radical-clock substrates cy-
clopropyl methanol and cyclobutanol exclusively
formed cyclopropanecarbaldehyde and cyclobutan-
one, respectively. The lack of ring-opened products
indicated that free-radical intermediates were not
formed in the oxidation of these substrates [12,13].
Further, deoxygenation of 2-propanol and benzyl al-
cohol oxidations had no effect on their initial rates.
These observations indicate that free radicals are not
formed in these reactions.
Kinetic Isotope Effects. Primary kinetic isotope ef-
fects were observed in the oxidation of 2-propanol
(␯obs,H/␯obs,D ϭ 4.9) at 35ЊC and benzyl alcohol (␯
/
obs,H
␯
ϭ 4.2) at 25ЊC. Significant solvent isotope ef-
obs,D
fects were also observed when the oxidation rates
of 4-nitrobenzyl alcohol and 4-methoxybenzyl alco-
hol in D2O were compared to those in H2O (␯
/
obs, H2O
␯
obs, D O ϭ 1.88 and 1.63, respectively; Table III).
2
Substituent Effects. An attempted Hammett plot of
the initial oxidation rates for 4-substituted benzyl al-
cohols exhibited a surprising concave curve with its
minimum near ␴ ϭ 0 (Fig. 7). Both electron-with-
drawing and electron-releasing substituents acceler-
ated the reaction, although the effect was stronger for
electron-withdrawing groups. Initial rates for the oxi-
DISCUSSION
Although attempts at direct isolation of the catalytic
complexresulted only in decomposition products, in-
3
Ϫ
Table III Solvent Kinetic Isotope Effects for Alcohol Oxidation by (Trpy)RuCl3/Fe(CN)6
OHϪ at 25ЊCa
/
Ϫ6
Ϫ1
Ϫ1
Substrate
Media
␯
obs (10
M
s
)
␯H/␯
D
4-Methoxybenzyl alcohol
4-Methoxybenzyl alcohol
4-Nitrobenzyl alcohol
4-Nitrobenzyl alcohol
H2O
D2O
H2O
D2O
3.77
2.31
8.08
4.29
1.63
1.88
3
a Reactions carried out with [Ru]T ϭ 1.40 · 10Ϫ4 M; [Alc.] ϭ 6.7 · 10Ϫ4M; [K3Fe(CN)6] ϭ 2.5 · 10Ϫ M;
[NaOH] ϭ 7.1 · 10Ϫ2 M; [NaClO4] ϭ 0.64 M.

766
KELSON AND PHENGSY
direct evidence suggests that a ruthenium-oxo species
is responsible for this reactivity [14]. When a solution
of (Trpy)RuCl3 in NaOH is acidified and treated with
an excess of silver nitrate, AgCl precipitates immedi-
ately in sufficient quantity to account for all three chlo-
ride ligands of the precatalyst. This suggests that
(Trpy)RuCl3 likely exchanges one or more chloride
ligands in NaOH for hydroxide or water. In the spec-
trophotometric titration of (Trpy)RuCl3 dissolved in
NaOH with hexacyanoferrate(III), three equivalents of
the oxidant are required to produce the catalytic spe-
complexfor alcohol. However, terpyridine ruthenium
systems are six-coordinate and exchange ligands rel-
atively slowly by dissociative processes. Hydroxide
dissociation will not likely be fast enough to account
for the observed reaction rates. Further, the presence
of tert-butanol has no measurable effect on the oxi-
dation rate of 2-propanol even up to a molar ratio of
10 : 1. This is not consistent with an alcohol coordi-
nation mechanism, although the steric bulk of tert-
butanol may attenuate its impact on such a reaction.
The observed inverse hydroxide dependence is
more likely a result of proton transfer from water to
the predominant form of the catalytic complex, as
shown in Eq. (6). An oxo or hydroxyl group likely
acts as the proton acceptor.
cies (␭
ϭ 472 nm). An excess of hexacyanofer-
max
rate(III) results in no further changes. Although the
reactions leading to the catalytic species appear com-
plex, the addition of hexacyanoferrate(III) results in
an immediate change in the initial spectrum of
(Trpy)RuCl3 in NaOH followed by a slow formation
k1
(RuIV) ϩ H2O I
R
J H(RuIV)ϩ ϩ HOϪ
(6)
of the catalytic species (␭
ϭ 472 nm) over min-
max
kϪ
1
utes. Since polypyridyl RuIV complexes such as
[(Bpy)2(Py)RuIV(O)]2ϩ (where Bpy is 2,2Ј-bipyridine)
and [(Trpy)(Phen)RuIV(O)]2 are prone to ligand hy-
where (RuIV)
ϭ
(TrpyO)RuIV(O)(LЈ)(LЉ) and
ϩ
H(RuIV)ϩ ϭ (TrpyO)RuIV(O)(HLЈ)(LЉ)ϩ. Terpyridine
ruthenium-oxo/hydroxyl complexes are noted to have
multiple protonation states [16,17], and protonation
would enhance the electrophilicity and reactivity of
the ruthenium center. Interestingly, proton transfer to
and from ruthenium-oxo/hydroxyl complexes can be
relatively slow due to accompanying isomerization of
the complex. Slow proton transfer and isomerization
has been reported in the oxidation of diphosphines by
droxylation in basic solutions [15], the slower reaction
is likely the hydroxylation of the terpyridine ligand
in a RuIV complex. A rapid two-electron oxidation
of the resulting RuII complexwould afford a
Ϫ
(TrpyO)RuIV(L)(LЈ)(LЉ) complexwhere (TrpyO) is
the hydroxylated (and deprotonated) terpyridine lig-
and and L, LЈ, LЉ are oxo, hydroxyl, aquo, and/
or chloro ligands. The oxidation chemistry of
(Bpy)2RuII(OH2)2 and (Trpy)RuII(OH2)3 suggests
that the high oxidation state of the catalytic species
will likely encourage one or more of its ligands to be
oxo groups [16,17]. Since reported dimeric and tri-
meric polypyridine ruthenium complexes all exhibit
strong absorption peaks at 600 nm and longer [18], the
472 nm absorption of the catalytic species seems char-
acteristic of a monomer. In addition to catalyzing al-
cohol oxidation, the catalytic species mediates the ox-
idation of triphenylphosphine by hexacyanoferrate(III)
to form triphenylphosphine oxide. Such oxygen trans-
fer is familiar in metal-oxo chemistry and is consistent
with the catalyst being a ruthenium-oxo complex [19–
22]. The kinetic isotope effects for both 2-propanol
and benzyl alcohol are also typical for alcohol oxida-
tion reactions by ruthenium-oxo systems [23,24].
Overall, a possible formulation for the catalytic spe-
cies is (TrpyO)RuIV(O)(LЈ)(LЉ) where LЈ and LЉ are
oxo, hydroxyl, aquo, and/or chloro ligands. Attempts
to directly prepare and characterize such complexes
are underway.
2
ϩ
3ϩ
ϩ
trans-[(Trpy)RuVI(O)2(H2O)]2 in water/acetonitrile
[25].
The saturation-like kinetics for alcohol concentra-
tions could be due to a slow alcohol oxidation follow-
ing a relatively sluggish preequilibrium. The simplest
mechanism would have this oxidation following the
slow protonation step mentioned earlier. This is con-
sistent with the expectation that protonation would
produce a more reactive oxidizer for alcohol. The ab-
sence of ring-opened products from the oxidation of
radical clock substrates and the insensitivity of the re-
action to oxygen indicate that the oxidation step does
not proceed by hydrogen transfer from the alcohol to
form radical intermediates. Alcohol oxidation by the
protonated intermediate is more likely a two-electron
process to produce a RuII intermediate, as shown in
Eq. (7). The specifics of this reaction will be addressed
later in this discussion.
k2
9:
H(RuIV )ϩ ϩ RRЈCHOH 9
Traditionally, the inverse hydroxide dependence
and saturation-like kinetics for alcohols would suggest
the exchange of a hydroxyl ligand on the catalytic
H3(RuII )ϩ ϩ RRЈC"O (7)
where H3(RuII)ϩ ϭ (TrpyO)RuII(OH)(H2LЈ)(LЉ)ϩ.

2-PROPANOL AND BENZYL ALCOHOL OXIDATION BY ALKALINE HEXACYANOFERRATE(III)
767
The small intercept in the ␯obs vs. 1/[OHϪ] plot for the
oxidation of 2-propanol (Fig. 5) suggests a parallel
alcohol oxidation step independent of the protonation
equilibrium. The simplest explanation would have the
unprotonated ruthenium species oxidize alcohol as
shown in Eq. (8) at a slower rate than the protonated
intermediate.
[H(RuIV)ϩ] from Eq. (12) into Eq. (13) and then solv-
ing for [H(RuIV)ϩ] gives Eq. (14):
[H(RuIV )ϩ ] ϭ
k1
[Ru]T
(14)
k1 ϩ kϪ1[HOϪ] ϩ k2[ROH]
Substitution of Eq. (14) into Eq. (13) and then solving
for [(RuIV)] leads to Eq. (15):
k3
9:
(RuIV ) ϩ RRЈCHOH 9
H2(RuII ) ϩ RRЈC"O (8)
kϪ1[HOϪ] ϩ k2[ROH]
k1 ϩ kϪ1[HOϪ] ϩ k2[ROH]
[(RuIV )] ϭ
[Ru]T (15)
where H2(RuII) ϭ (TrpyO)RuII(OH)(HLЈ)(LЉ).
The zero-order dependence on hexacyanofer-
rate(III) is consistent with the reduced ruthenium spe-
cies from alcohol oxidation rapidly reacting with two
equivalents of hexacyanoferrate(III) and an appropri-
ate amount of hydroxide to reform the predominant
catalytic complexas shown in Eqs. (9) and (10). This
would close the catalytic cycle.
Substitution of the expressions for [(RuIV)] and
[H(RuIV)ϩ], Eqs. (14) and (15), into Eq. (11) leads to
Eq. (16):
3Ϫ
d[Fe(CN)6
]
␯
ϭ Ϫ
obs
dt
2k1k2[ROH] ϩ 2kϪ1k3[HOϪ][ROH]
fast
ϩ 2k2k3[ROH]2
3
Ϫ
H2(RuII ) ϩ 2 Fe(CN)6 ϩ 2 HOϪ 9
9:
ϭ
[Ru]T
(16)
k1 ϩ kϪ1[HOϪ] ϩ k2[ROH]
4
Ϫ
(RuIV ) ϩ 2 Fe(CN)6 ϩ 2 H2O (9)
fast
3
Ϫ
H3(RuII )ϩ ϩ 2 Fe(CN)6 ϩ 3 HOϪ 9
9:
Since k2[ROH] and k3[ROH] represent the slow steps
of the catalytic reaction, their product, k2k3[ROH]2,
is likely very small compared to k1k2[ROH] and
4
Ϫ
(RuIV ) ϩ 2 Fe(CN)6 ϩ 3 H2O (10)
kϪ k3[HOϪ][ROH], so Eq. (16) can be approximated
1
Given the hypothetical mechanism outlined in Eqs.
(6)–(10), the rate for hexacyanoferrate(III) consump-
tion would be given by Eq. (11):
as Eq. (17):
3Ϫ
d[Fe(CN)6
]
␯
ϭ Ϫ
obs
3
Ϫ
dt
d[Fe(CN)6
]
␯
ϭ Ϫ
2k1k2[ROH] ϩ 2kϪ1k3[HOϪ][ROH]
k1 ϩ kϪ1[HOϪ] ϩ k2[ROH]
obs
dt
Ϸ
[Ru]T
(17)
ϭ 2k2[H(RuIV )ϩ ][ROH] ϩ 2k3 [(RuIV )][ROH]
(11)
This rate law is in good agreement with the exper-
By applying the steady-state hypothesis for
[H(RuIV)ϩ], we obtain Eq. (12)
imental data. The alcohol dependence in this equation
fits the observed relationship in Eq. (3) with the fol-
lowing expressions for a1 and a2:
k1[(RuIV )]
kϪ1[HOϪ] ϩ k2[ROH]
[H(RuIV )ϩ ] ϭ
(12)
k2
a1 ϭ
a2 ϭ
(18)
(19)
2(kϪ1k3[HOϪ] ϩ k1k2)[Ru]T
Since both H2(RuII) and H3(RuII)ϩ are consumed in fast
steps, the amount of both should be negligible as long
as hexacyanoferrate(III) is present. The total ruthe-
nium concentration may be given as Eq. (13):
k1 ϩ kϪ1[HOϪ]
2(kϪ1k3[HOϪ] ϩ k1k2)[Ru]T
[Ru]T ϭ [(RuIV)] ϩ [H(RuIV)ϩ]
(13)
The rate law in Eq. (17) is also consistent with the
hydroxide dependencies for 2-propanol and benzyl al-
cohol. In the case of 2-propanol, the following ex-
Replacing the expression for [(RuIV)] in terms of

768
KELSON AND PHENGSY
pressions for b1 and b2 and the assumption in Eq. (22)
make Eq. (17) fit the observed behavior in Eq. (4):
kϪ1[HOϪ] ϩ k2[ROH]
k1 ϩ kϪ1[HOϪ] ϩ k2[ROH]
␯ ϭ 2k3[ROH]
[Ru]T
(25)
3
ͩ ͪ
2kϪ1k3[ROH][Ru]T
b1 ϭ
b2 ϭ
ϭ 2k3[ROH][Ru]T (20)
kϪ
1
This expression can be simplified using k1 ϽϽ
kϪ [HOϪ] (which applies to 2-propanol and benzyl al-
cohol) to give Eq. (26):
1
2k1k2[ROH][Ru]T
(21)
kϪ
1
kϪ1[HOϪ] ϩ k2[ROH]
kϪ1[HOϪ] ϩ k2[ROH]
␯ ϭ 2k3[ROH]
[Ru]T
and
3
ͩ ͪ
ϭ 2k3[ROH][Ru]T
(26)
k1 ϩ k2[ROH] ϽϽ kϪ [HOϪ]
(22)
1
This expression is equivalent to Eq. (20) for b1. Under
the conditions in Figure 5 with 0.050 M NaOH (a typ-
ical value), ␯3 represents 18% of the observed rate. As
expected, alcohol oxidation by the protonated form of
the catalyst, H(RuIV)ϩ, is largely responsible for the
observed rate. The moderate kinetic deuterium isotope
A relatively small k2[ROH] with respect to kϪ [HOϪ]
1
is reasonable considering the relatively slow oxidation
rates for 2-propanol compared to benzyl alcohol. A
small k1 with respect to kϪ [HOϪ] would make the un-
1
protonated species the strongly preferred form of the
catalyst over the range of hydroxide concentrations
studied. This is consistent with the ␭max 472 nm peak
(that likely represents the unprotonated catalyst) ap-
pearing unchanged by the hydroxide concentration.
The relatively small size of k1 allows Eq. (17) to be
simplified to Eq. (23):
effect (␯obs,H/␯
ϭ 4.9) would largely represent k2
obs,D
and indicate that this step likely involves the cleavage
of an alcohol C–H bond.
For benzyl alcohol oxidation, the following ex-
pressions for c1, c2, and c3 make Eq. (23) fit the ob-
served behavior represented by Eq. (5):
3
Ϫ
d[Fe(CN)6
]
␯
ϭ Ϫ
2k1k2[ROH][Ru]T
obs
dt
c1 ϭ
c2 ϭ
c3 ϭ
(27)
ϭ 2k3[ROH][Ru]T (28)
(29)
kϪ
1
2k1k2[ROH] ϩ 2kϪ1k3[HOϪ][ROH]
kϪ1[HOϪ] ϩ kϪ2[ROH]
Ϸ
[Ru]T
(23)
2kϪ1k3[ROH][Ru]T
kϪ
1
k2[ROH]
The intercept and slope from the plot of ␯ vs. 1/
obs
[HOϪ] for 2-propanol (Fig. 5) give values of k3 ϭ
kϪ
1
1.91 · 10Ϫ MϪ sϪ and k1k2/kϪ ϭ 4.25 · 10Ϫ sϪ
1
1
1
2
1
1
through Eqs. (20) and (21). The contribution of alco-
hol oxidation by unprotonated catalyst, (RuIV) to the
observed rate, ␯obs, can be calculated through Eq. (24):
Table IV lists values of k1, k3, k2/kϪ , and ␯3/␯ ob-
1
obs
tained from Eqs. (27), (28), (29), and (26) for benzyl
alcohol, 4-nitrobenzyl alcohol, and 4-methoxybenzyl
alcohol. The values for k1 are within experimental er-
ror of each other as would be expected for this mech-
anism. The results for ␯3/␯obs confirm that alcohol ox-
idation by the protonated catalyst, H(RuIV)ϩ, is the
major contributor to the observed oxidation chemistry.
␯3 ϭ 2k3[(RuIV)][ROH]
(24)
Using Eq. (15) to express [(RuIV)] in terms of [Ru]T
gives Eq. (25):
Table IV Rate Constants from Eq. (23) for the Oxidation of Benzyl Alcoholsa
Ϫ1
Ϫ
Ϫ1
Ϫ1
Substrate
k1 (s
)
k3 (M 1 s
)
k2/kϪ (M
)
␯3/␯
obs
1
Ϫ2
Ϫ2
Ϫ2
4-Methoxybenzyl alcohol
Benzyl alcohol
4-Nitrobenzyl alcohol
(4.5 Ϯ 0.8) · 10
(4.4 Ϯ 2.4) · 10
(5.7 Ϯ 1.7) · 10
5.3 Ϯ 0.4
5.1 Ϯ 0.9
13.4 Ϯ 1.7
44 Ϯ 8
40 Ϯ 19.8
82 Ϯ 21
0.19 Ϯ 0.03
0.20 Ϯ 0.06
0.26 Ϯ 0.03
4
3
4
a Reactions carried out with [Ru]T ϭ 1.00 · 10Ϫ M; [K3Fe(CN)6] ϭ 1.7 · 10Ϫ M; [Alc.] ϭ 6.7 · 10Ϫ M, ␮ ϭ 0.50 M at 25ЊC. (NaClO4
added to adjust ionic strength to 0.50 M.)

2-PROPANOL AND BENZYL ALCOHOL OXIDATION BY ALKALINE HEXACYANOFERRATE(III)
769
The moderate isotope effect for the oxidation of ben-
zyl alcohol (␯obs,H/␯obs,D ϭ 4.2) clearly represents a pri-
mary kinetic deuterium isotope effect in k2 indicating
cleavage of an alcohol C9H bond in this step.
amount to a two-electron–two-proton reduction of the
catalyst, which is not surprising in light of the familiar
proton-coupled electrochemistry of ruthenium-aquo/
hydroxyl/oxo complexes [16,17].
Values for k2/kϪ1 and k3 show electron-withdrawing
and -releasing substituents accelerating both k2 and k3
over their values for unsubstituted benzyl alcohol.
This is consistent with the unusual concave shape of
the attempted Hammett plot for 4-substituted benzyl
alcohols (Fig. 7). Since the kinetic behavior of both 4-
methoxybenzyl and 4-nitrobenzyl alcohols mirror that
of benzyl alcohol, these substituent effects may rep-
resent shifts in the transition states of k2 and k3 [26].
Increases in k2 and k3 with electron-withdrawing and
-releasing substituents suggest the buildup of modest
negative and positive charges, respectively, at the al-
cohol carbon in the reaction with both protonated and
unprotonated forms of the catalyst. Although the for-
mation of radical intermediates could account for such
behavior, this system exhibits no ring opening in the
oxidation of cyclopropyl methanol and cyclobutanol
radical-clock substrates and exhibits no oxygen de-
pendence. A more consistent explanation for the sub-
stituent effects is a concerted mechanism in which al-
cohol deprotonation accompanies hydride transfer
from an alcohol C9H bond to an oxo ligand. The
partial charge that accumulates on the alcohol carbon
would depend on the relative timing of these two
events; electron-withdrawing substituents would en-
courage more alcohol deprotonation and more nega-
tive charge buildup in the transition state. The solvent
isotope effects for 4-nitrobenzyl alcohol [␯obs(H2O)/
Although pericyclic mechanisms have been pro-
posed for the oxidation of coordinated alcohols in
chromium(VI)-, rhenium(V)-, and ruthenium(VI)-oxo
complexes, this system provides evidence for a pos-
sible pericyclic oxidation of an uncoordinated alcohol
[27–33]. In nonaqueous media, terpyridine ruthe-
nium-oxo complexes oxidize alcohols by simple hy-
dride-transfer. The strong alkaline environment in this
system may support a low protonation state of the ru-
thenium catalyst that can act as a base as well as a
hydride acceptor. Simultaneous alcohol deprotonation
and hydride transfer would avoid energetic carbocat-
ion and radical intermediates and may exhibit activa-
tion enthalpies significantly below that of hydride
transfer alone. This may make catalytic alcohol oxi-
dation possible at low oxidation potentials. Although
this issue is not so important for synthesis, it is partic-
ularly important for direct alcohol fuel cells in which
the anode needs to collect electrons from alcohols near
their thermodynamic oxidation potentials. A detailed
survey of the substituent effects on k2 and k3 and their
activation parameters is underway.
The authors acknowledge the financial assistance of the
Camille and Henry Dreyfus Foundation and the donors of
the Petroleum Research Fund administered by the American
Chemical Society. We would also like to thank the Office
of Graduate Programs and the College of Science and Math-
ematics at California State University, Northridge.
␯
obs(D2O) ϭ 1.88] and 4-methoxybenzyl alcohol
[␯obs(H2O)/␯obs(D2O) ϭ 1.63] appear consistent with
the expectation for greater O–(H,D) cleavage in the
oxidation of 4-nitrobenzyl alcohol. Since the hydrox-
ide dependence in the kinetics precludes attack of so-
lution hydroxide in k2 or k3, the catalyst may be re-
sponsible for both C9H and O9H cleavage as part
of a pericyclic process (Scheme I). Overall, this would
BIBLIOGRAPHY
1. Griffith, W. P. Chem Soc Rev 1992, 179.
2. Mucientes, A. E.; Poblete, F. J.; Rodriguez, D.; Rodri-
guez, M. A.; Santiago, F. Int J Chem Kinet 1999, 31(1),
1.
3. Behari, K.; Narayan, H.; Shukla, R. S.; Gupta, K. C. Int
J Chem Kinet 1984, 16, 195.
4. Dwivedi, R. K.; Verma, K. M.; Kumar, P.; Behari, K.
Tetrahedron 1983, 39(5), 815.
5. Lalitha, T. V.; Rao, M. P.; Sethuram, B. Indian J Chem
1992, 31A, 309.
6. Navarro, M.; De Giovani, W. F.; Romero, J. R. Synth
Commun 1990, 20(3), 399.
7. Navarro, M.; Galembeck, S. E.; Romero, J. R.; De Giov-
ani, W. F. Polyhedron 1996, 15(9), 1531.
8. Meyer, T. J. J Electrochem Soc 1984, 131, 221C and
references therein.
Scheme I

770
KELSON AND PHENGSY
9. Sullivan, B. P.; Calvert, J. M.; Meyer, T. J. Inorg Chem
1980, 19, 1404.
21. Tokita, Y.; Yamguchi, K.; Watanabe, Y.; Morichima, I.
Inorg Chem 1993, 32, 329.
10. Wilberg, K. B.; Stewert, R. J Am Chem Soc 1955, 77,
1786.
22. Moyer, B. A.; Sipe, B. K.; Meyer, T. J. Inorg Chem
1981, 20, 1475.
11. Freeman, F.; Brant, J. B.; Hester, N. B.; Kamego,
A. A.; Kasner, M. L.; McLaughlin, T. G.; Paull, E. W.
J Org Chem 1970, 35(4), 982.
12. Maillard, B.; Forest, D.; Ingold, K. U. J Am Chem Soc
1976, 98, 7024.
23. Thompson, M. S.; Meyer, T. J. J Am Chem Soc 1982,
104, 4106.
24. Lee, D. G.; van den Engh, M. Can J Chem 1972, 50,
200.
25. Dovletoglou, A.; Meyer, T. J. J Am Chem Soc 1994,
116(1), 215.
26. Swain, C. G.; Langsdorf (Jr), W. P. J Am Chem Soc 73,
1951, 2813.
27. Wang, Z.; Chandler, W. D.; Lee, D. G. Can J Chem
1998, 76, 919.
28. Deng, L.; Ziegler, T. Organometallics 1997, 16, 716.
29. Dumez, D. D.; Mayer, J. M. Inorg Chem 1995, 34, 6396.
30. Scott, S. L.; Bakac, A.; Espenson, J. H. J Am Chem Soc
1992, 114, 4205.
31. Rocek, J.; Westheimer, F. H.; Eschenmoser, A.; Mol-
dovanyi, L.; Schreiber, J. Helv Chim Acta 1962, 45,
2554.
32. Wilberg, K. B.; Schafer, H. J Am Chem Soc 1967, 89,
455.
13. Effio, A.; Griller, D.; Ingold, K. U.; Beckwith, A. L. J.;
Serelis, A. K. J Am Chem Soc 1976, 102, 1734.
14. Attempts to isolate the catalytic intermediate including
concentration and counterion precipitation techniques
resulted in obvious spectroscopic changes and poorly
defined products.
15. Roecker, L.; Kutner, W.; Gilbert, J. A.; Simmons, M.;
Murrya, R. W.; Meyer, T. J. Inorg Chem 1985, 24, 3784.
16. Dobson, J. C.; Meyer, T. J. Inorg Chem 1988, 27, 3283.
17. Adeyemi, S. A.; Dovletglou, A.; Guadalupe, A. R.;
Meyer, T. J. Inorg Chem 1992, 31, 1375.
18. Lebeau, E. L.; Adeyemi, S. A.; Meyer, T. J. Inorg Chem
1998, 37, 6476.
19. Dovletglou, A.; Meyer, T. J. J Am Chem Soc 1994, 116,
215.
33. Lee, D. G.; Stewart, R. J Org Chem 1967, 32, 2868.
20. Cheng, S. Y. S.; James, B. R. J Mol Catal 1997, 117,
91.