Allylic oxidation of cyclohexene and indene by cis-[Ru(IV)(bpy)2 (py)(O)]2+

Article abstract of DOI:10.1021/ja991722x

The kinetics of oxidation of cyclohexene, cyclohexen-1-ol, and indene by cis-[Ru(IV)(bpy)₂(py)(O)]²⁺ (bpy = 2,2'-bipyridine and py = pyridine) have been studied in CH₃CN. The reactions are first-order in both Ru(IV)=O²⁺ and substrate in an initial, rapid stage in which Ru(IV) is reduced to Ru(III). The rate constants are 0.16 ± 0.01, 1.10 ± 0.02, and 5.74 ± 0.74 M^-1 s^-1 for cyclohexene, cyclohexen-l-ol, and indene, respectively. A k(α,α'-H₄)/k(α,α'-D₄) kinetic isotope effect of 21 ± 1 is observed for the oxidation of cyclohexene. At a 2:1 ratio of Ru(IV)=O²⁺ to olefin, the reactions of Ru(IV)=O²⁺ with either cyclohexene or indene give Ru(II)- NCCH₃²⁺ and the 4-electron ketone products, 2-cyclohexen-l-one and indenone, respectively, as identified by GC-MS. As the ratio of cyclohexene to Ru(IV)=O²⁺ is increased, cyclohexen-l-ol becomes an increasingly competitive product. The mechanisms of these reactions are highly complex. They involve two distinct stages and the formation and subsequent reactions of Ru(III)-substrate bound intermediates.

Full text of DOI:10.1021/ja991722x

5984
J. Am. Chem. Soc. 2000, 122, 5984-5996
Allylic Oxidation of Cyclohexene and Indene by
cis-[RuIV(bpy)2(py)(O)]2+
Laura K. Stultz,† My Hang V. Huynh,| Robert A. Binstead,‡ Maria Curry, and
Thomas J. Meyer*,§
Contribution from the Department of Chemistry, Venable and Kenan Laboratories, The UniVersity of North
Carolina at Chapel Hill,Chapel Hill, North Carolina 27599-3290
ReceiVed May 24, 1999
Abstract: The kinetics of oxidation of cyclohexene, cyclohexen-1-ol, and indene by cis-[RuIV(bpy)2(py)(O)]2+
(bpy ) 2,2′-bipyridine and py ) pyridine) have been studied in CH3CN. The reactions are first-order in both
RuIV)O2+ and substrate in an initial, rapid stage in which Ru(IV) is reduced to Ru(III). The rate constants are
0.16 ( 0.01, 1.10 ( 0.02, and 5.74 ( 0.74 M-1 s-1 for cyclohexene, cyclohexen-1-ol, and indene, respectively.
A k(R,R′-H4)/k(R,R′-D4) kinetic isotope effect of 21 ( 1 is observed for the oxidation of cyclohexene. At a
2:1 ratio of RuIVdO2+ to olefin, the reactions of RuIVdO2+ with either cyclohexene or indene give RuII-
NCCH32+ and the 4-electron ketone products, 2-cyclohexen-1-one and indenone, respectively, as identified by
GC-MS. As the ratio of cyclohexene to RuIVdO2+ is increased, cyclohexen-1-ol becomes an increasingly
competitive product. The mechanisms of these reactions are highly complex. They involve two distinct stages
and the formation and subsequent reactions of RuIII-substrate bound intermediates.
Introduction
loporphyrins have been used as catalysts for the radical
autoxidation of cyclohexene by O2 giving 2-cyclohexen-1-one
as the primary product, but typically after long induction
periods.9a-c Tabushi and co-workers showed that the autoxi-
dation of cyclohexene catalyzed by manganese tetraphenylpor-
phyrin chloride (MnIIITPPCl) can be significantly promoted by
the presence of reducing agents.9d,e MnIIITPPCl efficiently
catalyzes the autoxidation of cyclohexene to 2-cyclohexen-1-
one but in the presence of NaBH4 or H2 and colloidal platinum,
the primary products are cyclohexanol and cyclohexene oxide,
respectively. The cyclohexanol formed in the reaction with
NaBH4 is most likely due to the reduction of cyclohexene oxide
formed as the initial product.
The oxidation of cyclohexene by Cr(V)-oxo heteropolytung-
state anions gives both epoxidation and allylic oxidation
products. The product ratios can be varied by changing the
solvent, the polytungstate ligand, and the countercation.10 The
reaction has been proposed to proceed by radical intermediates.
A magnesium oxide supported polysilazane ruthenium com-
plex is capable of catalyzing the oxygenation of indene to
indenone without byproducts under mild conditions. The sup-
ported ruthenium complex is also effective without the presence
of a cocatalyst.11
The results of a variety of studies on the oxidation of
cyclohexene by transition metal complexes have been reported
in the literature. Oxidation by cytochrome P-450 gives cyclo-
hexene oxide and 2-cyclohexen-1-ol as products.1 Similar results
have been found for model systems based on metalloporphyrins
of Fe,2 Mn,3 and Cr4 with iodosylbenzene as the source of the
oxidative equivalents. Cyclohexene is also oxidized by high
oxidation state complexes of Cu,5 Mo,6 V,7 and Ti.8 Metal-
† Current address: Birmingham-Southern College, Box 549022, Arka-
delphia Road, Birmingham, AL 35254.
| Current address: Los Alamos National Laboratory, Chemistry Division
J514, Los Alamos, NM 87545.
‡ Current address: Spectrum Software Associates, P.O. Box 4494, Chapel
Hill, NC 27515-4494. Telephone/Fax: 1-919-942-4061. E-mail: specsoft@
compuserve.com.
§ Current address: Los Alamos National Laboratory, ALDSSR MS A127,
Los Alamos, NM 87545. Telephone: 1-505-667-8597. Fax: 1-505-667-
5450. E-mail: tjmeyer@lanl.gov.
(1) Groves, J. T.; Krishnan, S.; Avaria, G. E.; Nemo, T. E. In Biomimetic
Chemistry; Dolphin, D., McKenna, C., Murakami, Y., Tabushi, I., Eds;
Advances in Chemistry Series 191; American Chemical Society: Wash-
ington, DC, 1980; Chapter 15.
(2) Groves, J. T.; Kruper, W. J.; Nemo, T. E.; Myers, R. S. J. Mol. Catal.
1980, 7, 169. (b) Medina, J. C.; Gabriunas, N.; PaezMozo, E. J. Mol. Catal.
A-Chem. 1997, 115(2), 233. (c) Lei, Z. Q. React. Funct. Polym. 1999, 39(3),
239.
(3) Nappa, M. J.; McKinney, R. J. Inorg. Chem. 1988, 27, 3740. (b)
KnopsGerrits, P. P.; Vankelecom, I. J. F.; Beatse, E.; Jacobs, P. A. Catal.
Today 1996, 32(1-4), 63. (c) Matsushita, T.; Sawyer, D. T.; Sobkowiak,
A. J. Mol. Catal. A: Chem. 1999, 137(1-3), 127.
(4) Garrison, J. M.; Bruice, T. C. J. Am. Chem. Soc. 1989, 111, 191. (b)
Agarwal, D. D.; Goswami, H. S. (c) Agarwal, D. D.; Goswami, H. S.; Raina,
D.; Agarwal, S. C. Ind. J. Chem. B 1995, 34(4), 357.
(5) Zakharov, A. N.; Zefirov, N. S. Kinet. Catal. 1998, 39(4), 493. (b)
Kim, Y.; Kim, H.; Lee, J.; Sim, K.; Han, Y.; Paik, H. Appl. Catal., A 1997,
155(1), 15.
(6) Olason, G.; Sherrington, D. C. Macromol. Symp. 1998, 131, 127.
(7) Bahranowski, K.; Kooli, F.; Poltowicz, J.; Serwicka, E. M. React.
Kinet. Catal. 1998, L64(1), 3. (b) Schuchardt, U.; Guerreiro, M. C.; Shul’pin,
G. B. Russ. Chem. Bull. 1998, 47(2), 247.
(8) Chen, L. Y.; Chuah, G. K.; Jaenicke, S. Catal. Lett. 1998, 50(1-2),
107. (b) Talsi, E. P.; Shalyaev, K. V. J. Mol. Catal. A: Chem. 1996, 105(3),
131. (c) Jappar, N.; Xia, Q. H.; Tatsumi, T. J. Catal. 1998, 180(2), 132.
In earlier work, we showed that cis-[RuIV(bpy)2(py)(O)]2+
oxidizes trans-stilbene, cis-stilbene, styrene, and norbornene to
the corresponding epoxides in CH3CN.12 For olefins containing
R-C-H bonds, allylic oxidation to the R-ketone occurs in ace-
(9) Modnicka, T. In Metalloporphyrins in Catalytic Oxidations; Sheldon,
R. A., Ed.; Marcel Dekker: New York, 1994, Chapter 9. (b) Paulson, D.
R.; Ullman, R.; Sloanne, R. B.; Closs, G. L. J. Chem. Soc., Chem. Commun.
1974, 186. (c) Peterson, M. W.; Rivers, D. S.; Richman, R. M. J. Am. Chem.
Soc. 1985, 107, 2907. (d) Tabushi, I.; Koga, N. J. Am. Chem. Soc. 1979,
101, 6456. (e) Tabushi, I.; Yazaki, A. J. Am. Chem. Soc. 1981, 103, 7371.
(10) Khenkin, A. M.; Hill, C. L. J. Am. Chem. Soc. 1993, 115, 8179.
(11) Wang, T. J.; Sun, F. Y.; Huang, M. Y.; Jiang, Y. Y. React. Kinet.
Catal. Lett. 1996, 57(1), 147. (b) Hatsuit, T. H. Bull. Chem. Soc. Jpn. 1980,
53(9), 2655.
(12) Stultz, L. K.; Binstead, R. A.; Reynolds, M. S.; Meyer, T. J. J. Am.
Chem. Soc. 1995, 117, 2520.
10.1021/ja991722x CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/10/2000

Allylic Oxidation of Cyclohexene and Indene
J. Am. Chem. Soc., Vol. 122, No. 25, 2000 5985
tonitrile, as reported for cyclohexene in an earlier work (eq 1).13
Leising and Takeuchi also reported the catalytic, aerobic
oxidation of cyclohexene by cis-[RuII(bpy)2(PPh3)(H2O)]2+
which yields 2-cyclohexen-1-one, 2-cyclohexen-1-ol, and small
amounts of cyclohexene oxide.14
Table 1. Pseudo-First-Order Rate Constants for the Oxidation of
Cyclohexene by 1.7 × 10-4 M cis-[Ru(bpy)2(py)(O)]2+ in CH3CN at
T ) 25.0 ( 0.1 °C
a
[cyclohexene] M
kobs (s-1
)
2.04 × 10-2
4.09 × 10-2
6.13 × 10-2
8.17 × 10-2
1.02 × 10-1
1.23 × 10-1
1.43 × 10-1
1.64 × 10-1
1.84 × 10-1
2.04 × 10-1
(1.35 ( 0.02) × 10-2
(2.53 ( 0.03) × 10-2
(3.71 ( 0.05) × 10-2
(5.14 ( 0.03) × 10-2
(6.50 ( 0.04) × 10-2
(7.71 ( 0.06) × 10-2
(8.76 ( 0.04) × 10-2
(1.02 ( 0.02) × 10-1
(1.16 ( 0.03) × 10-1
(1.28 ( 0.03) × 10-1
We report here the results of a detailed mechanistic study on
the oxidations of cyclohexene and indene by cis-[RuIV(bpy)2-
(py)(O)]2+. By the use of a combination of UV-visible and
FTIR measurements with the application of global analysis, we
have been able to identify key intermediates and to demonstrate
that these reactions occur by mechanisms which involve a
complicated series of kinetically coupled reactions.
a Calculated as kobs ) 4 × k1[olefin], k1 ) 0.16 ( 0.01 M-1 s-1
.
mixing apparatus consisting of Hamilton gastight syringes and a T-valve
mixer connected to the solution IR cell with PTFE tubing.
UV-visible spectra were collected as a function of time by the use
of a Hewlett-Packard 8452A diode array spectrophotometer interfaced
with an IBM-compatible PC. The measurements were made in standard
quartz cuvettes or in a solution IR cell (see above) when short path
lengths were required. The temperature of solutions during kinetic
studies was maintained to within ( 0.2 °C with use of a Lauda RM6
circulating water bath and monitored with an Omega HH-51 thermo-
couple probe. For rapid reactions, UV-visible spectral changes were
followed by use of an OLIS (Bogart, GA) RSM/1000 rapid-scanning,
dual beam spectrophotometer linked by liquid light guide to a Hi-Tech
Scientific (Salisbury, England) SF-61MX stopped-flow mixing ap-
paratus. Some experiments were performed by use of a Hi-Tech
Scientific SF-51 stopped-flow apparatus linked by fiber-optic coupling
to either a Beckman DU or a Harrick rapid scan monochromator
interfaced with a Zenith 158 computer system employing OLIS data
acquisition hardware and software. The temperature of the reactant
solutions was controlled to within (0.2 °C by use of a Neslab RTE-
110 water bath and monitored via the Pt resistance thermometers of
the SF-51 or SF-61MX mixing units.
UV-visible Kinetic Studies. All kinetic studies were performed in
acetonitrile solution at T ) 25.0 ( 0.2 °C, generally with a pseudo-
first-order excess of the organic substrate. The concentrations of RuIVd
O2+ were varied from 4.0 × 10-6 M to 9.4 × 10-3 M with use of
0.020 to 10 cm path length cells in order to maintain absorbance
readings within an appropriate range. The concentrations of substrate
were varied from 2.04 × 10-2 to 2.04 × 10-1 M for cyclohexene (Table
1), 5.0 × 10-3 to 1.0 M for 2-cyclohexen-1-ol, and 0.010 to 0.200 M
for indene.
For convenience, the following abbreviations will be used
throughout the manuscript:
cis-[RuII(bpy)2(py)(H2O)]2+ ) RuII-OH2
2+
2+
cis-[RuII(bpy)2(py)(NCCH3)]2+ ) RuII-NCCH3
cis-[RuIII(bpy)2(py)(OH)]2+ ) RuIII-OH2+
cis-[RuIV(bpy)2(py)(O)]2+ ) RuIVdO2+
cis-[RuIV(bpy)2(py)(18O)]2+ ) RuIVd18O2+
Experimental Section
Materials. High purity acetonitrile was used as received from
Burdick & Jackson. House-distilled water was purified by the use of a
Barnstead E-Pure deionization system. High purity NMR solvents and
18O-enriched water were purchased from Cambridge Isotope Labora-
tories and used as received. Cyclohexene was purified by passage
through a column of alumina to remove the BHT stabilizer followed
by fractional distillation. 3,3,6,6-tetradeuterocyclohexene was purchased
from MSD Isotopes and passed through a plug of powdered alumina
immediately prior to use. Common laboratory chemicals employed in
the preparation of compounds were reagent grade and used without
further purification.
The presence of intermediates in the reactions of RuIVdO2+ with
the olefins was investigated by monitoring the UV-visible spectral
changes after quenching reaction mixtures at various times by the
addition of a small excess of either hydroquinone or ascorbic acid.
Both reductants rapidly reduce all Ru(IV) or Ru(III) species to their
Ru(II) forms, which are more easily detected by their characteristic
Preparations. The salts cis-[Ru(bpy)2(py)(O)](PF6)215 and cis-[Ru-
(bpy)2(py)(18O)](ClO4)212 were prepared by literature methods. Caution!
Although the preparation of the perchlorate salt has been repeated
numerous times without incident, particular care should be exercised
in its preparation and handling. Perchlorate salts of organometallic
cations, metal complexes, and organic cations have been known to
explode spontaneously. Hexafluorophosphate was used as the coun-
teranion except when solubility required the use of the perchlorate salt.
Instrumentation. Organic products from the oxidation reactions
were analyzed by the use of a Hewlett-Packard 5890 series II gas
chromatograph with a 12 m × 0.2 mm × 0.33 µm HP-1 column (cross-
linked methyl silicone gum) and a Hewlett-Packard 5971A mass
selective detector, both interfaced with an HP Vectra PC computer
system. IR spectra were recorded by use of a Mattson Galaxy 5020
series FT-IR spectrophotometer interfaced with an IBM-compatible PC.
IR measurements were made either in KBr pellets or in CD3CN solution
by the use of a demountable cell with NaCl plates and Teflon spacers.
Kinetic studies with this cell were performed with a manually operated
MLCT absorption bands16 and their different rates of solvolysis to RuII-
2+
NCCH3
.
Multi-wavelength kinetic data were processed by the use of the
program SPECFIT.17 When necessary, the program was used with molar
absorptivity spectra of known species and/or known rate constants to
constrain the global least-squares fit. A more detailed description of
the global kinetic analysis is given in an earlier publication.16 Single
wavelength data were analyzed with the program KINFIT.18
Results
Organic Product Analysis and Stoichiometry. Organic
products from the reactions between RuIVdO2+ and cyclohexene
(13) Seok, W. K.; Dobson, J. C.; Meyer, T. J. Inorg. Chem. 1988, 27,
5. (b) Kanmani, A. S.; Vancheesan, S. J. Mol. Catal. A: Chem. 1997,
125(2-3), 127. (c) Cheng, W. C.; Yu, W. Y.; Li, C. K.; Che, C. M. J. Org.
Chem. 1995, 60(21), 6840.
(14) Leising, R. A.; Takeuchi, K. J. Inorg. Chem. 1987, 26, 4391.
(15) Moyer, B. A.; Meyer, T. J. Inorg. Chem. 1981, 20, 436.
(16) Binstead, R. A.; Stultz, L. K.; Meyer, T. J. Inorg. Chem. 1995, 34,
546.
(17) SPECFIT, Spectrum Software Associates: Chapel Hill, NC, 1995.
(18) Binstead, R. A. KINIFT, Nonlinear Curve Fitting for Kinetic
Systems; Chemistry Department: University of North Carolina, Chapel Hill,
NC, 1990.

5986 J. Am. Chem. Soc., Vol. 122, No. 25, 2000
Stultz et al.
Table 3. Results of Isotope Labeling for Olefin Oxidations by
Table 2. Product Analyses by GC-MS for Cyclohexene and
Indene Oxidation by cis-[RuIV(bpy)2(py)(O)]2+ in CH3CN at T ) 25
( 1 °C
cis-[Ru(bpy)2(py)(O)]2 in CH3CN at T ) 25 ( 1 °C by GC-MS
stoichio-
metry
fractional
yield
fractional
reactiona
product
reactiona
RuIVdO2+
cyclohexene
RuIVdO2+:olefin
products
yieldb
RuIVdO2+
d4-cyclohexene
+
2:1
1:1
1:2
1:5
d2-cyclohexenone
d3-cyclohexenone
d2-cyclohexenone
d3-cyclohexenone
d2-cyclohexenone
d3-cyclohexenone
16O-cyclohexenol
18O-cyclohexenol
16O-cyclohexenone
18O-cyclohexenone
16O-cyclohexenol
18O-cyclohexenol
16O-cyclohexenone
18O-cyclohexenone
16O-indenone
98
2
95
5
80
20
14
0
25
61
7
0
89
4
20
80
97
3
+
2:1
2-cyclohexen-1-ol
2-cyclohexen-1-one
2-cyclohexen-1-ol
2-cyclohexen-1-one
2-cyclohexen-1-ol
2-cyclohexen-1-one
2-cyclohexen-1-ol
2-cyclohexen-1-one
2-cyclohexen-1-ol
2-cyclohexen-1-one
indene
8
92
14
86
37
63
51
49
53
47
9c
91
55c
45
80c
20
2:1.5
1:5
RuIVd18O2+
cyclohexene
+
1:25
1:50
2:1
RuIVdO2+
+
1:1
cyclohexene + H218Ob
RuIVdO2+
+
indene
indenone
1:1
indene
RuIVd18O2+ + indene
1:1
1:1
indenone
18O-indenone
1:2
indene
RuIVdO2+ + indene +
16O-indenone
indenone
H218Ob
18O-indenone
a cis-[Ru(bpy)2(py)(O)]2+ ) 6.5 mM for cyclohexene oxidation and
15 mM for indene oxidation. b The total product yield was 93-96%
a [RuIVdO2+] ) 15 mM. b CH3CN solvent 2% v/v in H218O.
based on added RuIVdO2+ c The ratio of unreacted indene to indenone.
.
Scheme 1
or indene in CH3CN under argon were determined by GC-MS
analysis of mixtures obtained after reaction times of 2-24 h at
room temperature. Calibration curves and product analyses for
the oxidation of cyclohexene are provided in Supporting
Information Figures S1 and S2. The product distributions
obtained at RuIVdO2+ to olefin ratios of 2:1, 1:1.5, 1:5, 1:25,
and 1:50 are listed in Table 2. For cyclohexene oxidation, the
total yields for organic products were 93-96% based on RuIVd
16O2+ added.
The oxidation of indene yielded only the 4-electron allylic
product, indenone. On the basis of oxidation by 2 equiv of
RuIVdO2+, indenone was formed in ∼90% yield. The corre-
sponding allylic alcohol and epoxide products were not detected
by GC-MS analysis, even with a 50:1 ratio of indene to RuIVd
O2+
.
In the oxidation of cyclohexene by RuIVdO2+, both the allylic
alcohol, 2-cyclohexen-1-ol, and allylic ketone, 2-cyclohexen-
1-one, were detected. At a 2:1 ratio of RuIVdO2+ to cyclohex-
ene, only 8% of the alcohol was formed, but as the ratio of
substrate to oxidant was increased, the proportion of alcohol
increased. When the reaction was performed with a 50-fold
excess of cyclohexene, the percentage yield of the alcohol
product rose to 53%. Epoxide products were not observed in
any of the GC-MS product analyses. Further experimental
details are given in Supporting Information.
18O-Labeling Studies. 18O-labeling experiments were con-
ducted in CH3CN. In these experiments, cyclohexene and indene
were oxidized by RuIVd18O2+ and the products determined by
both FTIR spectroscopy and GC-MS analysis. The results of
the GC-MS analyses are summarized in Table 3.
The 18O-labeling study for the oxidation of cyclohexene was
performed at a 1:5 ratio of RuIVd18O2+ to cyclohexene. Under
these conditions, both 2-cyclohexen-1-one and 2-cyclohexen-
1-ol are products. None of the alcohol product was 18O-labeled,
while 2-cyclohexen-1-one was found to be 61 ( 15% labeled,
as shown by FTIR band intensity measurements. With the 18O
label, ν(CdO) in the cyclohexenone product shifted to 1685
cm-1 from 1648 cm-1 in the 18O product. Baseline inconsisten-
cies and overlap with nearby bands limited the accuracy of the
analysis. Oxidation of indene by RuIVd18O2+ at a 1:1 ratio gave
80 ( 5% 18O incorporation into the indenone product by GC-
MS analysis. When the same reaction was studied by FTIR,
ν(CdO) in the product shifted from 1712 cm-1 (16O) to 1684
cm-1 (18O).
The oxidations of cyclohexene and indene by RuIVd16O2+
were also performed in CH3CN containing 2% v/v H218O, and
the products were analyzed by GC-MS. The 2-cyclohexen-1-
ol product contained no 18O label, while 2-cyclohexen-1-one
was 4% 18O-labeled. Similarly in the oxidation of indene, only
3% of the indenone produced contained the 18O label.
Oxidation of 3,3,6,6-Tetradeuterocyclohexene. Oxidation
of 3,3,6,6-tetradeuterocyclohexene (d4-cyclohexene) was inves-
tigated as a probe for possible radical intermediates. As
illustrated in Scheme 1, radical formation followed by allylic
rearrangement19 would lead to inequivalent ketone products, one
containing two deuteriums having molecular weight 98 amu and
the other containing three deuteriums having molecular weight
99 amu. Allylic rearrangement can be distinguished by the
appearance of parent ions at m/e ratios of 98 and 99, and parent
minus ethylene at 69 and 70 with correction for 13C. Results at
ratios of RuIVdO2+ to cyclohexene of 2:1, 1:1, and 1:2 are listed
in Table 3. From these data, it is evident that at a 2:1 ratio of
RuIVdO2+ to cyclohexene, only 2% of the ketone product
underwent allylic rearrangement, while at a 1:2 ratio, the amount
of rearranged ketone increased to 20%.
(19) Groves, J. T.; Subramanian, D. V. J. Am. Chem. Soc. 1984, 106,
2177.

Allylic Oxidation of Cyclohexene and Indene
J. Am. Chem. Soc., Vol. 122, No. 25, 2000 5987
Figure 1. Rapid-scan stopped-flow, UV-visible spectral changes for the reaction between 1.0 × 10-4 M cis-[RuIV(bpy)2(py)(O)]2+ and 1.0 M
cyclohexene in CH3CN at T ) 25.0 ( 0.2 °C, path length ) 1.0 cm, 0 < t < 5 s.
Figure 2. Diode array, UV-visible spectral changes with manual mixing for the reaction between 1.0 × 10-4 M cis-[RuIV(bpy)2(py)(O)]2+ and 1.0
M cyclohexene in CH3CN at T ) 25.0 ( 0.2 °C, path length ) 1.0 cm, 0 < t < 5 s.
Spectral Changes during the Oxidation of Cyclohexene.
In Figure 1 are shown UV-visible spectral changes that occur
during the initial stage of the oxidation of cyclohexene (1.0 M)
by RuIVdO2+ (1.0 × 10-4 M). In the first 5 s after stopped-
flow mixing, the initial spectrum of RuIVdO2+ was replaced
by a spectrum characteristic of RuIII-OH2+ (λmax ) 380 nm, ꢀ
468 nm, ꢀ ) 8750 M-1 cm-1).16 Spectral changes collected on
a diode array spectrophotometer after manual mixing are shown
in Figure 2. In the time required to mix the reactants in the cell
(5-10 s), the first stage of the reaction was nearly complete.
In the second stage of the reaction, shown in Figure 2, the
absorbance at 380 nm decreased, while the absorbance at 440
nm increased. The final trace corresponds to the spectrum of
) 5520 M-1 cm-1).16 A small shoulder also appeared at ∼470
nm consistent with the MLCT band for RuII-OH2 (λmax
)
RuII-NCCH3 (λmax ) 440 nm, ꢀ ) 8160 M-1 cm-1).16
2+
2+

5988 J. Am. Chem. Soc., Vol. 122, No. 25, 2000
Stultz et al.
Figure 3. Predicted concentration profiles determined by kinetic simulation of the disproportionation of RuIII-OH2+ followed by RuIVdO2+ oxidation
in the presence of (a) 1 M cyclohexene and (b) 0.01 M cyclohexene, RuIVdO2+ (‚‚‚), RuIII-OH2+ (- - -), RuII-OH22+ (s), RuII-NCCH32+ (- ‚ -).
The nature of the products of the initial stage (t ≈ 10 s) was
investigated further by a series of quenching experiments in
which an excess of hydroquinone was added to reduce Ru(IV)
and Ru(III) to Ru(II) (Supporting Information Figure S3).
Singular value decomposition (SVD) of the spectral data after
quenching indicated the presence of three predominant colored
components and two distinct kinetic processes both leading to
RuII-NCCH32+. A global fit of the data to the biexponential
model, A f C, B f C (kA and kB are the rate constants for the
two steps) gave kA ) (1.28 ( 0.04) × 10-2 s-1 and kB ) (1.46
( 0.02) × 10-3 s-1. The latter is consistent with k for solvolysis
the predicted spectrum of A, a Ru(II) intermediate that
undergoes solvolysis more rapidly than RuII-OH22+ (Supporting
Information Figure S4). On the basis of this analysis, [A]:[ RuII-
OH22+] ≈ 1:5 although this ratio includes the small amount of
RuII-OH2 present before the addition of hydroquinone.
The nature of the bound intermediate was investigated by
GC-MS analysis of reaction mixtures after quenching. With
cyclohexene at 1.1 × 10-2 M and RuIVdO2+ at 2.3 × 10-3 M,
hydroquinone 2.7 × 10-2 M was added after 1 min, and the
initial products allowed to solvolyze for an additional 30 min.
2-cyclohexen-1-one, benzoquinone, and hydroquinone were
detected by GC-MS. No 2-cyclohexen-1-ol was detected
initially, but it did appear as a product over time. Quenching
after 5 min gave a [ketone]:[alcohol] ratio of 6:1.
2+
of RuII-OH2 in 1 M cyclohexene/CH3CN, ksolv ) (1.48 (
0.02) × 10-3 s-1, T ) 25.0 ( 0.2 °C. The total concentration
of Ru was adjusted from the predicted spectrum of the final
product C to obtain the correct molar absorptivity for RuII-
NCCH32+. With B identified as RuII-OH22+, the initial con-
centrations were adjusted so that the predicted spectrum of B
matched that of RuII-OH22+. The resulting global fit returned
The kinetics of the initial stage in the oxidation of cyclo-
hexene by RuIVdO2+ (1.7 × 10-4 M) were investigated under
pseudo first-order conditions with a large excess of olefin (2.04
× 10-2 - 2.04 × 10-1 M). Single wavelength measurements

Allylic Oxidation of Cyclohexene and Indene
J. Am. Chem. Soc., Vol. 122, No. 25, 2000 5989
Figure 4. IR spectral changes for the reaction between 7.6 × 10-3 M cis-RuIV(bpy)2(py)(O)]2+ and 2.78 × 10-2 M cyclohexene in CD3CN, path
length ) 0.2 mm (NaCl).
at 388-400 nm were fit to pseudo-first-order kinetics. The
experimental rate constant, kobs, was found to vary linearly with
olefin concentration, and matched well with studies performed
at 1.0 M added olefin.
Although this mechanism fits the absorbance-time traces
reasonably well, it is incomplete. It does not account for the
fact that 2-cyclohexen-1-ol becomes an increasingly important
product as the RuIVdO2+:olefin ratio is increased. It also does
not explain how the free ketone product appears in the solution.
The spectral changes in Figure 1 show that Ru(IV) is reduced
to Ru(III) in the first step. The presence of ketone in the reaction
mixture after quenching shows that more than two oxidative
equivalents per olefin are consumed. This demonstrates that the
mechanism is complex and does not involve a single kinetic
process. This conclusion is supported by the appearance of both
The oxidation of 3,3,6,6-tetradeuterocyclohexene by RuIVd
O2+ is considerably slower than the oxidation of cyclohexene.
UV-visible spectral changes with time are shown in Supporting
Information Figures S5 and S6. The reactions were performed
with only a 10-fold excess of reductant to guard against possible
interference by small amounts of partially deuteriated impurities.
For d4-cyclohexene, the reaction was performed with 1.0 × 10-2
M d4-cyclohexene and 1.0 × 10-3 M RuIVdO2+ (path length
) 0.10 cm) which was compared to the reaction between 1.0
× 10-3 M cyclohexene and 1.0 × 10-4 M RuIVdO2+ (path
length ) 1.00 cm). In both cases, the UV-visible spectral
changes were analyzed with SPECFIT. The SVD analysis of
the spectral-time traces indicated the presence of five colored
components, but two were present at very low levels. This
precluded the use of a complete kinetic model; however, the
data could be fit to biexponential kinetics, A f B f C which
was adequate to isolate the apparent rate constant for the initial
stage. This procedure yielded k ′ ) (9.75 ( 0.03) × 10-3 M-1
s-1 for d4-cyclohexene, and k ′ ) 0.20 ( 0.01 M-1 s-1 for
cyclohexene. On the basis of these values, the k(R,R′-H4)/
k(R,R′-D4) kinetic isotope effect is 21 ( 1 for this reaction.
The reaction between cyclohexene and RuIVdO2+ was
also monitored by FTIR at ratios of between 2:1 and 1:7 of
RuIVdO2+ to cyclohexene. The results of a typical experiment
are shown in Figure 4. A series of ν(bpy) ring stretching modes
appeared in the region 1300-1600 cm-1 that varied in energy
and molar absorptivity with oxidation state of the complex.
Kinetic analysis of the IR data showed that the initial, rapid
stage of the reaction was first-order in both cyclohexene and
RuIVdO2+. The latter stages were relatively independent of
both. The SVD of the spectral data showed the presence of
four contributors. The kinetics could be modeled by the rate
2+
RuII-OH2 and a second Ru(II) intermediate after hydro-
quinone quenching.
On the basis of this analysis, the interpretation of kobs and its
olefin dependence must include both rate-determining oxidation
of cyclohexene and following steps in which further oxidation
by RuIVdO2+ occurs. The rate constant for the rate-determining
step, k′, can be calculated from eq 2
kobs
k ′ )
(2)
n[olefin]
in which n is the stoichiometry of RuIVdO2+ consumed in the
initial stage. For cyclohexene, the value n ) 4 was used to
calculate k ′ based on the analysis in Discussion.
The quenching studies demonstrate that there are two forms
of Ru(III) after the initial stage but predominantly RuIII-OH2+
.
The slower spectral changes that occur past 10 s (Figure 2) can
be accounted for largely by a mechanism involving: (1) initial
disproportionation of RuIII-OH2+
,
-1
81 M-1 s
2 RuIII-OH2+ y
z RuIVdO2+ + RuII-OH2
(3)
2+
4070 M-1 s-1
(K ) 0.02)
(2) further oxidation of cyclohexene by RuIVdO2+; and, (3)
solvolysis of RuII-OH22+. Concentration profiles calculated by
kinetic simulation of this mechanism in the presence of 1 and
0.01 M cyclohexene are shown in Figure 3.

5990 J. Am. Chem. Soc., Vol. 122, No. 25, 2000
Stultz et al.
Figure 5. Rapid-scan, UV-visible spectral changes for the reaction between 1.0 × 10-4 M cis-[RuIV(bpy)2(py)(O)]2+ and 1.0 M 2-cyclohexen-
1-ol in CH3CN solution at T ) 25.0 ( 0.2 °C, path length ) 1.0 cm.
Table 4. Rate Constants for the Oxidation of Indene by
cis-[RuIV(bpy)2(py)(O)]2+ in CH3CN at T ) 25.0 ( 0.2 °Ca
Scheme 2
indene [RuIV)O2+
(M)
]
M
k1 (s-1
)
k2 (s-1
)
k3 (s-1
)
0.010 2.0 × 10-5 0.142 ( 0.009 0.025 ( 0.003 0.012 ( 0.003
0.010 1.0 × 10-4 0.144 ( 0.002 0.018 ( 0.005 0.004 ( 0.001
0.020 2.0 × 10-5 0.290 ( 0.020 0.030 ( 0.003 0.011 ( 0.002
0.020 1.0 × 10-4 0.282 ( 0.008 0.076 ( 0.009 0.015 ( 0.001
0.050 2.0 × 10-5 0.680 ( 0.020 0.032 ( 0.005 0.010 ( 0.001
0.050 1.0 × 10-4 0.655 ( 0.006 0.046 ( 0.009 0.008 ( 0.002
0.100 2.0 × 10-5 1.390 ( 0.060 0.040 ( 0.007 0.010 ( 0.001
0.100 1.0 × 10-4 1.270 ( 0.010 0.042 ( 0.004 0.008 ( 0.002
0.200 2.0 × 10-5 2.300 ( 0.100 0.040 ( 0.005 0.009 ( 0.001
0.200 1.0 × 10-4 2.410 ( 0.050 0.044 ( 0.006 0.008 ( 0.003
law (A f B f C f D). While simplified, this model allowed
the rate constant for the initial oxidation of cyclohexene by
RuIVdO2+ to be separated from the following steps. On the
basis of this analysis, for the reaction between 2.78 × 10-2
M
cyclohexene and 7.6 × 10-3 M RuIVdO2+, kobs ) (1.53 ( 0.08)
× 10-2 s-1 and k′ ) 0.138 ( 0.008 M-1 s-1. These values
are consistent with values obtained from UV-visible measure-
ments.
a The data were fit to the model A f B f C f D f E, with k1, k2,
k3, and k4 ) the rate constants for each step from A to E, with SPECFIT.
k4 was fixed at 1.66 × 10-3 s-1, the rate constant for solvolysis of
RuII-OH2 in CH3CN.
2+
The appearance of the ketone product was monitored by the
growth in ν(CdO) at 1685 cm-1. It occurred in the latter stages
of the reaction and deviates slightly from single-exponential
kinetics. The results of the global fit returned almost identical
spectra for species C and D. These observations suggest that
the free ketone is released in two separate steps with k2 ) (1.4
As shown in Supporting Information Figure S7, kinetic
simulation with k1 ) 2 M-1 s-1 demonstrated that the growth
of RuIII-OH2+ is biexponential under conditions of high RuIVd
O2+ (1.0 × 10-4 M) and cyclohexenol (1 M). At lower
concentrations, the kinetic behavior is close to exponential.
Results are summarized in Table S1. A plot of kobs (s-1) versus
the cyclohexenol concentration was linear, and a linear least-
squares fit gave kobs ) 2.20 ( 0.02 M-1 s-1 with an intercept
of 0.016 ( 0.009 s-1. On the basis of eq 2 with n ) 2, k ′ )
( 0.2) × 10-3 and k3 ) (1.7 ( 0.1) × 10-4 s-1
.
Spectral Changes for the Oxidation of 2-Cyclohexen-1-
ol. The oxidation of 2-cyclohexen-1-ol by RuIVdO2+ was
studied independently. The spectral changes at early time are
shown in Figure 5. The kinetics in this case were highly
wavelength dependent. Absorbance changes from of 389-400
nm were used to follow the growth of Ru(III). The data in this
region exhibited biexponential kinetics with an approximate ratio
of 2:1 for the observed rate constants. This behavior was more
pronounced at high concentrations of RuIVdO2+ and cyclohex-
enol. Our observations are consistent with initial 2-electron
oxidation of the alcohol followed by comproportionation, as
shown in Scheme 2.
1.10 ( 0.02 M-1 s-1
.
Spectral Changes for the Oxidation of Indene. In Figure
6A are shown UV-visible spectral traces with time for the
reaction between RuIVdO2+ (9.1 × 10-5 M) and indene (5.04
× 10-2 M) in CH3CN at 25 °C. At early times (t < 5 s), RuIVd
O
2+ disappeared with the concomitant growth of an intermediate
with absorption maxima near 380 nm and ∼550 nm. The
ultimate product was RuII-NCCH32+ (λmax ) 440 nm, ꢀ ) 8160
M-1 s-1).16

Allylic Oxidation of Cyclohexene and Indene
J. Am. Chem. Soc., Vol. 122, No. 25, 2000 5991
Figure 6. (A) Rapid-scan, UV-vis spectral changes for the reaction between 9.1 × 10-5 M cis-[RuIV(bpy)2(py)(O)]2+ and 0.0504 M indene in
CH3CN at T ) 25.0 ( 0.2 °C. Each trace represents the average of 160 × 1 ms scans. (B) The first six temporal eigenvectors from U × S. (C) First
six spectral eigenvectors, V, from the singular value decomposition analysis of the data matrix, Y. By definition, the scan data can be reconstructed
as Y ) U × S × V.
Singular value decomposition (SVD) of this data set gave
evidence for the presence of five distinct contributing compo-
nents and four-coupled kinetic processes (Figure 6B and C).
The kinetics could be modeled by a multi-exponential series
consistent with the mechanism A f B f C f D f E, with k1,
k2, k3, and k4 ) the rate constants for each step from A to E, A
) RuIVdO2+, and E ) RuII-NCCH32+. Predicted spectra and
concentration profiles from a typical global fit are shown in
Figure 7A and B. Observed first-order rate constants for the
first three stages were listed in Table 4 as a function of the
concentrations of indene and RuIVdO2+. Since some of the
data sets did not extend far enough in time to determine the
fourth rate constant uniquely, its value was fixed at 1.66 × 10-3
s-1. This is the rate constant for solvolysis of RuII-OH22+. The
rate constants derived from the kinetic fits were the same for
data obtained in aerated solutions and under a nitrogen
atmosphere.
k ′ ) 5.74 ( 0.4 M-1 s-1 from the slope by using eq 2 with
n ) 2 (see Discussion). The intercept was near zero (0.07 (
0.03 s-1).
The predicted spectrum of the product of the first stage
(Figure 7A) appeared to be a superposition of RuIII-OH2+ and
an intermediate with λmax ≈ 550 nm. To determine the nature
of the intermediate, the reaction was quenched at this point by
addition of either hydroquinone or ascorbic acid as reducing
agent (Supporting Information Figure S8). SVD of the spectra
after the addition of the reductant indicated the presence of a
minimum of four colored components. The most rapid spectral
change was dependent on the concentration of added reductant
and led to further formation of Ru(II). This arises from reduction
of an intermediate since RuIII-OH2+ is reduced rapidly by
hydroquinone (k ≈ 1 × 106 M-1 s-1) and would not be
detectable on this time scale.20 There must be at least two
Ru(II) intermediates, and both undergo solvolysis to give RuII-
NCCH32+. The first (λmax ) 468 nm, ksolv ) (1.56 ( 0.02) ×
The observed rate constant (kobs, s-1) for the first step varied
linearly with the concentration of indene. A linear least-
squares fit to the data yielded the second-order rate constant
10-3 s-1) is RuII-OH2 formed by reduction of RuIII-OH2+
2+
or unreacted RuIVdO2+ 18
The second contains either coordi-
.

5992 J. Am. Chem. Soc., Vol. 122, No. 25, 2000
Stultz et al.
Figure 7. (A) Predicted spectra obtained from a global fit of the data in Figure 6A to the multi-exponential kinetic model (A f B f C f D f E).
A(s), B(- - -), C(‚‚‚), D(_ _ _), E(- ‚ -) with A ) RuIVdO2+ and E ) RuII)NCCH32+. (B) As in Figure 7A showing predicted concentration
profiles.
nated ketone or alcohol, solvolyzes with ksolv ) (2.6 ( 0.2) ×
10-2 s-1, and has an MLCT absorption feature at λmax ) 550-
In the second stage of the reaction, both Ru(III) interme-
diates were reduced to Ru(II). The rate constants obtained from
the multiexponential fits appeared to be independent of the
concentrations of both indene and initially added RuIVdO2+ with
k2 ) 0.042 s-1 (Table 4). k2 was kinetically well-defined only
at relatively high olefin concentrations because of kinetic overlap
with the solvolysis steps. The rate constants for the third and
650 nm. From the known molar absorptivities of RuII-OH2
2+
and RuII-NCCH32+, the ratio of RuII-OH2 to Ru(II) inter-
2+
mediate was ∼1.4:1.
(20) Binstead, R. A.; McGuire, M. E.; Dovletoglou, A.; Seok, W. K.;
Roecker, L. E.; Meyer, T. J. J. Am. Chem. Soc. 1992, 114, 173.

Allylic Oxidation of Cyclohexene and Indene
J. Am. Chem. Soc., Vol. 122, No. 25, 2000 5993
fourth steps were also independent of both indene and RuIVd
the O-atom in the ketone product comes from the oxo group of
RuIVdO2+ and not from H2O in the solvent. (6) The organic
product of the Ru(IV) f Ru(III) stage remains bound to
Ru(III) and is the ketone as shown by FTIR. It is released
following hydroquinone reduction of Ru(III) to Ru(II) as shown
by GC-MS analysis. 2-Cyclohexen-1-ol only appears as a
product at the Ru(III) f Ru(II) stage (see below).
O
2+. The latter appeared to involve solvolysis of a Ru(II)
intermediate with k3 ) 0.010 ( 0.002 s-1 and solvolysis of
2+
2+
RuII-OH2 to give RuII-NCCH3
.
The oxidation of indene by RuIVdO2+ was also monitored
by FTIR spectroscopy at ratios of between 2:1 and 1:6 RuIVd
O
2+ to indene at room temperature. ν(RudO) at 800 cm-1 was
absent in even the earliest scans, showing that RuIVdO2+
had been consumed by 5-10 s after mixing, consistent with
the results of the stopped-flow experiments. The growth of
ν(CdO) for the ketone at 1712 cm-1 was independent of indene
and initial RuIVdO2+ with kobs ) (1.2 ( 0.4) × 10-2 s-1 which
coincided with k3 in the multiexponential fit to the UV-visible
data. There was no evidence for ν(CdO) in the intermediate
before solvolysis (Supporting Information Figure S9).
On the basis of these results, the first step appears to be
insertion of RuIVdO2+ into an R-C-H bond by a nonradical
pathway with O-atom transfer. This would give a bound alcohol,
Discussion
but the initial product formed is the ketone bound to Ru(III)
rather than the alcohol bound to Ru(II). To explain these
observations, it is necessary to invoke rapid oxidation of the
Unlike cytochrome P-450 and some model porphyrin com-
plexes, epoxides are not observed as products in the oxidations
of indene or cyclohexene by RuIVdO2+. At least at 25 °C,
RuIVdO2+ prefers to insert into an allylic C-H bond rather
than to add to the double bond of the olefin. Epoxidation is a
known reaction for the stilbenes and other olefins where allylic
oxidation cannot occur.12
initial RuII-alcohol2+ intermediate by RuIVdO2+
.
There is an additional complication in that the expected
ν(CdO) stretch for a ketone bound to Ru(III) does not appear
in the FTIR spectrum. Given this observation, the Ru(III)
intermediate is probably better formulated as the enol
For the two olefins, the products are similar as are the spectral
changes that accompany the reactions. They appear to occur
by somewhat related mechanisms. In the product distribution
for cyclohexene, both 2-electron (alcohol) and 4-electron
(ketone) products appear whose relative amounts vary with the
RuIVdO2+ to olefin ratio. The ketone dominates when this ratio
is high with the alcohol increasing from <10% at 2:1 to ∼50%
at 1:50. In the oxidation of indene, the allylic ketone is the only
product even at RuIVdO2+:indene ratios as high as 1:50.
The results of the various kinetics and product studies allow
detailed mechanisms to be proposed for these reactions. They
are complex, multistep processes similar in complexity to the
epoxidation of olefins by RuIVdO2+ reported earlier.12 There
are three discernible stages in both reactions: (1) reduction of
Ru(IV) to Ru(III), (2) reduction of Ru(III) to Ru(II), and (3)
solvolysis of Ru(II). The degree to which they overlap kineti-
cally depends on the olefin concentration with the three stages
well separated at high olefin concentrations.
or enolate,
The bound alcohol should be acidic. The pKa for loss of a proton
from cis-[RuIII(bpy)2(py)(H2O)]3+ is 0.85.15
Assuming the enolate, further oxidation of the RuII-alcohol2+
intermediate occurs by
Oxidation of Cyclohexene. The Ru(IV) f Ru(III) Stage.
The results of a number of experiments provide direct evidence
about the details of the first stage of cyclohexene oxidation in
which Ru(IV) is reduced to Ru(III). (1) The reaction is first-
order in both RuIVdO2+ and olefin. The spectral changes are
consistent with reduction of Ru(IV) to Ru(III) (predominantly
as RuIII-OH2+). The same pattern of spectral changes is
observed over a wide range of RuIVdO2+ to olefin ratios
with concentrations of cyclohexene as high as 1 M. (2) Addition
of hydroquinone after the first stage causes rapid reduction
The net reaction is
of Ru(III) to Ru(II). Two products form, RuII-OH2 and
This reaction is consistent with the evidence for two different
forms of Ru(III) at the end of the Ru(IV) f Ru(III) stage.
Ru(III) f Ru(II) Stage. In the second stage, Ru(III) is
reduced to Ru(II), and the free ketone appears as a product in
solution. The alcohol also begins to appear as a product at this
stage. Absorbance-time traces can be simulated by a mechanism
in which there is initial disproportionation, eq 3, followed by
RuIVdO2+ oxidation of cyclohexene, eqs 4 and 5.
To explain the data in Tables 2 and 3, it is necessary to invoke
a competing parallel pathway. At RuIVdO2+:olefin ) 2:1, the
product is largely ketone with nearly complete 18O transfer and
no allylic rearrangement. As this ratio is increased, the alcohol
2+
a Ru(II) intermediate (λmax ≈ 442 nm), Ru(II)′. The latter
undergoes solvolysis with ksolv ) (1.28 ( 0.04) × 10-2 s-1
.
The ratio of RuII-OH2
to Ru(II)′ is 4-5:1. (3) At
2+
RuIVdO2+:cyclohexene ) 2:1, the product is principally 2-
cyclohexen-1-one (92%, Table 2). Oxidation of 3,3,6,6-tetra-
deuterocyclohexene under these conditions gives largely unre-
arranged ketone (Table 3) which is consistent with a mechanism
in which cyclohexenyl radical is not formed. (4) From the rate
constant data for the R,R′-H4 and R,R′-D4 olefins, there is a
H/D kinetic isotope effect of 21 ( 1 for the reaction between
RuIVdO2+ and cyclohexene. (5) From the 18O-labeling results,

5994 J. Am. Chem. Soc., Vol. 122, No. 25, 2000
Stultz et al.
product appears. There is incomplete 18O transfer, and some
allylic rearrangement occurs. The yield of alcohol increases from
8% at RuIVdO2+:olefin ) 2:1 to 53% at 1:50.
Given these observations, the parallel pathway may occur
by 1-electron oxidation by RuIII-OH2+ to give the cyclohexyl
radical
∼80% 18O-labeled with little incorporation from added H218O.
Addition of excess hydroquinone or ascorbic acid at the end of
2+
this stage gives RuII-OH2
and a second component
2+
which solvolyzes to give RuII-NCCH3 with ksolv ) (2.6 (
0.2) × 10-2 s-1. The ratio of RuII-OH2 to intermediate
2+
is ∼1.4:1.
Oxidation of indene also appears to occur by insertion
From known autoxidation chemistry, this would be followed
by O2 oxidation in a radical chain mechanism,9,10 possibly with
followed by further oxidation to an RuIII-alkoxide2+ intermediate
termination by
Given the 18O-labeling results in Table 3, the O-atom in
the alcohol product is not derived from the oxo group of
RuIVdO2+ or from H2O and must come from O2.
Once formed under close to stoichiometric conditions, the
alcohol is more reactive toward oxidation by RuIVdO2+ than
the olefin by a factor of 14 (k ) 1.10 versus 0.16 M-1 s-1) and
would not build up in the solution.
The net reaction is
A radical pathway at higher olefin to RuIVdO2+ ratios would
explain both the appearance of allylic rearrangement and the
less than complete 18O transfer in the ketone product. The
alcohol is favored as the olefin concentration is increased
because direct oxidation by RuIII-OH2+ (eq 7) becomes
increasingly competitive with disproportionation and oxidation
by RuIVdO2+. Direct oxidation increases in importance as the
reaction proceeds, and [RuIII-OH2+] decreases since the rate
of disproportionation depends on the square of the RuIII-OH2+
concentration.
With this interpretation, the low-energy absorption feature at
∼550 nm could be an alkoxide f Ru(III) charge-transfer band
in RuIII-alkoxide2+ which, upon reduction, gives RuII-alcohol2+
as the observed intermediate. An analogous N3- f Ru(III) band
appears at 498 nm (ꢀ ) 7370 M-1 cm-1) for [RuIII(bpy)2(py)-
(N3)]2+ in CH3CN.21
The oxidative equivalent in the putative RuIII-enolate2+
intermediate may contribute both to disproportionation and direct
oxidation:
The principal difference in behavior between the oxidations
of cyclohexene and indene at this stage is in the extent of
oxidation with indene oxidized by two electrons and cyclohex-
ene by four. The difference can be rationalized if the key to
4-electron oxidation is proton loss and formation of a bound
enolate. A RuIII-enolate2+ form is inaccessible for indene
because there is no dissociable C-H proton.
A product of both is RuII-ketone2+ which, from the quenching
experiments with added hydroquinone, undergoes solvolysis
The mechanism in eqs 14 and 15 predicts a RuII-OH22+:RuII-
alcohol2+ ratio of 1:1 following hydroquinone or ascorbic acid
reduction of Ru(III). The experimental value is ∼1.4:1. Also,
when RuIVd18O2+ is the oxidant, only ∼80% of the indenone
is 18O-labeled, and the O-atom in the unlabeled portion does
not come from H2O (Table 3).
with k ) (1.28 ( 0.04) × 10-2 s-1
.
This reaction and the solvolysis of RuII-OH22+ comprise the
slow, third stage of the reaction.
These two observations suggest that a competing reaction may
exist at the Ru(IV) f Ru(III) stage, perhaps 1-electron transfer
to give the indenyl radical,
Oxidation of Indene. Ru(IV) f Ru(III) Stage. The oxida-
tion of indene by RuIVdO2+ is also first-order in olefin and
RuIVdO2+ but is more rapid than cyclohexene by a factor of
36. RuIII-OH2+ is a product along with a second component
having λmax ≈ 550 nm. There is no FTIR evidence for ν(CdO)
of a coordinated ketone complex although the only product
at the end of the reaction is indenone. With RuIVd18O2+ as the
oxidant and RuIVdO2+:indene ) 1:1, the indenone product is
(21) Brown, G. M.; Callahan, R. W.; Meyer, T. J. Inorg. Chem. 1975,
14, 1915.

Allylic Oxidation of Cyclohexene and Indene
J. Am. Chem. Soc., Vol. 122, No. 25, 2000 5995
Scheme 3
One-Electron
Solvolysis
Radical chain autoxidation would give the alcohol,
was obtained by FTIR and UV-visible monitoring. In a re-
lated reaction, Keene et al.22 investigated the intramolecular
oxidation of a chelated alkoxide to the aldehyde. The kinetics
were consistent with initial disproportionation to Ru(II) and Ru-
(IV) followed by rate-limiting proton release and internal
oxidation.
This is not the case here where there is a first-order
dependence on Ru(III), and yet RuII-OH22+ and RuII-ketone2+
appear simultaneously. This suggests a mechanism in which a
reactive form of the alkoxide complex undergoes oxidation by
RuIII-OH2+ as in eq 19.
and further oxidation by RuIVdO2+, indenone.
Ru(III) f Ru(II) Stage. In the reduction of Ru(III) to
Ru(II), RuIII-OH2+ and RuIII-alkoxide2+ are simultaneously
reduced to Ru(II) to give RuII-OH22+ (ksolvolysis ) k4 ) 1.66 ×
10-3 s-1) and RuII-ketone2+. The ketone intermediate undergoes
solvolysis with k ) 1.2 × 10-2 s-1 to give free ketone,
ν(CdO) at 1712 cm-1. The same rate constant for solvolysis
(22) Ridd, M. J.; Gakowski, D. J.; Sneddon, G. E.; Keene, F. R. J. Chem.
Soc., Dalton Trans. 1992, 1949.

5996 J. Am. Chem. Soc., Vol. 122, No. 25, 2000
Stultz et al.
Table 5. Summary of Oxidations Involving cis-[RuIV(bpy)2(py)(O)]2+ in CH3CN or H2O
k
∆Hq
∆Sq
reductant
cyclohexene
solvent
product
(25 °C M-1 s-1
)
kH/kD
(kcal mol-1
)
(cal deg-1 mol-1
)
ref
CH3CN 2-cyclohexen-1-one
CH3CN indenone
CH3CN trans-stilbene oxide
H2O
H2O
CH3CN benzoquinone
CH3CN OPPh3
CH3CN OS(CH3)2
CH3CN O2S(CH3)2
0.16 ( 0.01
5.74 ( 0.74
0.278 ( 0.002
21 ( 1
7.4 ( 0.1b
-34 ( 3.4b
this work
this work
12
indene
trans-stilbene
hydroquinone
H2O2
phenol
PPh3
S(CH3)2
OS(CH3)2
PhCH2OH
4.4 ( 0.1
0.64 ( 0.04
6.0 ( 0.3
-46 ( 0.4
-29.0 ( 0.1
-37 ( 3
-14 ( 2
-19 ( 3
-26 ( 3
-39 ( 3
-38 ( 1
benzoquinone
O2
(9.6 ( 0.3) × 105
1.74 ( 0.18
28.7 ( 1.0
22.0 ( 1.2
20
29
13
26
25
25
21
27
(1.9 ( 0.4) × 102
1.60 ( 0.02a
1.71 × 101
5.5 ( 0.2 10.3 ( 0.6
4.7 ( 0.5
8.0 ( 0.8
1.34 × 10-1
6.8 ( 0.2
H2O
PhCHO
2.43 ( 0.03
50 ( 3
5.8 ( 0.4
[RuII(bpy)2(py)(OH2)]2+ CH3CN [RuIII(bpy)2(py)(OH)]2+ (4.07 ( 0.13) × 103 14.6 ( 0.7
a Measured at 25.8 ( 0.2 °C. b See ref 28.
According to this mechanism, kobs ) ka in the limit, kb . k-a
with kobs ) ka. In this limit, the rate determining step is
conversion of RuIII-alkoxide2+ to the activated form. One
possibility for the activated form is proton migration,
bond breaking. In this reaction, the electron flow is from the
σ(C-H) molecular orbital
to give a carbanion intermediate stabilized by electronic
delocalization to Ru(III).
Mechanistic Summary. A global summary of the mecha-
nisms for oxidation of cyclohexene and indene is given in
Scheme 3.
to the lowest acceptor orbital or orbitals at RuIVdO2+. These
are a pair of orbitals, dπ*Ru, largely dπ in character but
antibonding with regard to the Ru-O interaction through mixing
with oxygen 2p orbitals. There is a double vacancy in this set
of orbitals which allows RuIVdO2+ to function as a 2-electron
acceptor.
Most allylic oxidations by transition metal complexes occur
by radical autoxidation mechanisms. RuIVdO2+ is capable of
allylic oxidation but without the intervention of radical path-
ways. The radical chemistry, that does occur, has its origin in
This mechanism can be added to others whose details have
been elucidated in earlier studies and are summarized in Table
5. They include hydride transfer from alcohols,23 electrophilic
oxo attack on phenol with net C-H insertion,24 oxygen atom
transfer to sulfides25 and phosphines,26 oxygen atom transfer
RuIII-OH2+ and not in RuIVdO2+
.
There are important implications for synthesis in the mech-
anism in Scheme 3. The initial reaction appears to be a concerted
process that gives an alcohol product. Further oxidation gives
the ketone and Ru(III) intermediates. They complicate the
mixture of products by initiating 1-electron, radical chemistry.
A related complication has been observed from Ru(III) in the
epoxidations of cis- and trans-stilbene and styrene by cis-[Ru-
in the epoxidation of olefins,12 and proton-coupled electron
2+ 27
transfer in the oxidations of hydroquinones20 and RuII-OH2
.
Acknowledgment is made to the National Science Founda-
tion under Grant No. CHE-9503738 and to the Department of
Energy under Grant No. LM 19X-SX 092C for support of this
research.
(bpy)2(py)(O)]2+
.
These results point to an advantage in immobilizing the
oxidant on a surface or in a membrane as a way of limting
oxidation to the alcohol stage. We are currently investigating
this possibility on high surface area oxide supports.
Supporting Information Available: Procedure of generat-
ing calibration curves for analysis, procedure of product analysis,
and Figures S1-9 (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
Mechanism of Oxidative Insertion. Oxidative insertion into
an allylic C-H bond occurs in the first step in Scheme 3. An
argument could be made for consecutive 1-electron steps and a
Ru(III) alkoxy radical intermediate with a lifetime too short for
allylic rearrangement to occur. However, it should be noted that
there is a significant kinetic advantage for RuIVdO2+ compared
to RuIII-OH2+ even though they have comparable redox
potentials. With reductants that undergo 1-electron transfer by
proton-coupled electron-transfer such as hydroquinone,20 rate
constants are comparable. Compared to RuIVdO2+, oxidation
of olefins by RuIII-OH2+ occurs by 1-electron transfer and far
more slowly.
JA991722X
(23) Thompson, M. S.; Meyer, T. J. J. Am. Chem. Soc. 1982, 104, 4106.
(b) Roecker, L. E.; Meyer, T. J. J. Am. Chem. Soc. 1987, 109, 746.
(24) Seok, W. K.; Meyer, T. J. J. Am. Chem. Soc. 1988, 110, 7358. (b)
Seok, W. K.; Dobson, J. C.; Meyer, T. J. Inorg. Chem. 1988, 27, 3.
(25) Roecker, L. E.; Dobson, J. C.; Vining, W. J.; Meyer, T. J. Inorg.
Chem. 1987, 26, 779.
(26) Moyer, B. A.; Sipe, B. K.; Meyer, T. J. Inorg. Chem. 1981, 20,
1475.
(27) Binstead, R. A.; Moyer, B. A.; Samuels, G. J.; Meyer, T. J. J. Am.
Chem. Soc. 1981, 103, 2897.
(28) Curry, M. E. Ph.D. Dissertation, University of North Carolina at
Chapel Hill, NC, 1990.
(29) Gilbert, J. A.; Gersten, S. W.; Meyer, T. J. J. Am. Chem. Soc. 1982,
104, 6872.
Given the magnitude of the kinetic isotope effect for
cyclohexene, oxidative insertion probably occurs through an
asymmetrical transition state at which there is extensive C-H