Carbon-carbon bond activation of 2,2,6,6-tetramethyl-piperidine-1-oxyl by a RhII metalloradical: A combined experimental and theoretical study

Article abstract of DOI:10.1021/ja078157f

Competitive major carbon-carbon bond activation (CCA) and minor carbon-hydrogen bond activation (CHA) channels are identified in the reaction between rhodium(II) meso-tetramesitylporphyrin [Rh^II(tmp)] (1) and 2,2,6,6-tetramethyl-piperidine-1-oxyl (TEMPO) (2). The CCA and CHA pathways lead to formation of [Rh^III(tmp)Me] (3) and [Rh^III(tmp)H] (5), respectively. In the presence of excess TEMPO, [Rh^II(tmp)] is regenerated from [Rh^III(tmp)H] with formation of 2,2,6,6-tetramethyl- piperidine-1-ol (TEMPOH) (4) via a subsequent hydrogen atom abstraction pathway. The yield of the CCA product [Rh^III(tmp)Me] increased with higher temperature at the cost of the CHA product TEMPOH in the temperature range 50-80°C. Both the CCA and CHA pathways follow second-order kinetics. The mechanism of the TEMPO carbon-carbon bond activation was studied by means of kinetic investigations and DFT calculations. Broken symmetry, unrestricted b3-lyp calculations along the open-shell singlet surface reveal a low-energy transition state (TS1) for direct TEMPO methyl radical abstraction by the Rh^II radical (S_H2 type mechanism). An alternative ionic pathway, with a somewhat higher barrier, was identified along the closed-shell singlet surface. This ionic pathway proceeds in two sequential steps: Electron transfer from TEMPO to [Rh^II(por)] producing the [TEMPO] ⁺[Rh^I(por)]^- cation-anion pair, followed by net CH₃⁺ transfer from TEMPO⁺ to Rh^I with formation of [Rh^III(por)Me] and (DMPO-like) 2,2,6-trimethyl-2,3, 4,5-tetrahydro-1-pyridiniumolate. The transition state for this process (TS2) is best described as an S_N2-like nucleophilic substitution involving attack of the d_z2 orbital of [Rh^I(por)]^- at one of the C_Me-C_ring σ* orbitale of [TEMPO] ⁺. Although the calculated barrier of the open-shell radical pathway is somewhat lower than the barrier for the ionic pathway, R-DFT and U-DFT are not likely comparatively accurate enough to reliably distinguish between these possible pathways. Both the radical (S_H2) and the ionic (S _N2) pathway have barriers which are low enough to explain the experimental kinetic data.

Full text of DOI:10.1021/ja078157f

Published on Web 01/19/2008
Carbon-Carbon Bond Activation of
2,2,6,6-Tetramethyl-piperidine-1-oxyl by a RhII Metalloradical:
A Combined Experimental and Theoretical Study
Kin Shing Chan,*,† Xin Zhu Li,† Wojciech I. Dzik,‡ and Bas de Bruin*,‡
Department of Chemistry, The Chinese UniVersity of Hong Kong, Shatin, New Territories,
Hong Kong, People’s Republic of China, and Department of Homogeneous and Supramolecular
Catalysis, Van’t Hoff Institute for Molecular Sciences (HIMS), UniVersity of Amsterdam,
Nieuwe Achtergracht 166, 1018 WV, Amsterdam
Received October 24, 2007; E-mail: ksc@cuhk.edu.hk; bdebruin@science.uva.nl
Abstract: Competitive major carbon-carbon bond activation (CCA) and minor carbon-hydrogen bond
activation (CHA) channels are identified in the reaction between rhodium(II) meso-tetramesitylporphyrin
[RhII(tmp)] (1) and 2,2,6,6-tetramethyl-piperidine-1-oxyl (TEMPO) (2). The CCA and CHA pathways lead
to formation of [RhIII(tmp)Me] (3) and [RhIII(tmp)H] (5), respectively. In the presence of excess TEMPO,
[RhII(tmp)] is regenerated from [RhIII(tmp)H] with formation of 2,2,6,6-tetramethyl-piperidine-1-ol (TEMPOH)
(4) via a subsequent hydrogen atom abstraction pathway. The yield of the CCA product [RhIII(tmp)Me]
increased with higher temperature at the cost of the CHA product TEMPOH in the temperature range
50-80 °C. Both the CCA and CHA pathways follow second-order kinetics. The mechanism of the TEMPO
carbon-carbon bond activation was studied by means of kinetic investigations and DFT calculations. Broken
symmetry, unrestricted b3-lyp calculations along the open-shell singlet surface reveal a low-energy transition
state (TS1) for direct TEMPO methyl radical abstraction by the RhII radical (SH2 type mechanism). An
alternative ionic pathway, with a somewhat higher barrier, was identified along the closed-shell singlet
surface. This ionic pathway proceeds in two sequential steps: Electron transfer from TEMPO to [RhII(por)]
producing the [TEMPO]+[RhI(por)]- cation-anion pair, followed by net CH3+ transfer from TEMPO+ to RhI
with formation of [RhIII(por)Me] and (DMPO-like) 2,2,6-trimethyl-2,3,4,5-tetrahydro-1-pyridiniumolate. The
transition state for this process (TS2) is best described as an SN2-like nucleophilic substitution involving
attack of the dz2 orbital of [RhI(por)]- at one of the CMe-Cring σ* orbitals of [TEMPO]+. Although the calculated
barrier of the open-shell radical pathway is somewhat lower than the barrier for the ionic pathway, R-DFT
and U-DFT are not likely comparatively accurate enough to reliably distinguish between these possible
pathways. Both the radical (SH2) and the ionic (SN2) pathway have barriers which are low enough to explain
the experimental kinetic data.
or comparable in strength compared to M-H bonds.2 Examples
Introduction
of unstrained, non-aliphatic CCA and CHA have been reported
Carbon-carbon bond activation (CCA) by a transition metal
complex is much less frequently observed than C-H bond
activation (CHA).1 It has been commonly accepted that C-C
bond activation is sterically more demanding than C-H bond
activation.1 Furthermore, C-H bonds are statistically more
abundant than C-C bonds in most substrates, thereby favoring
CHA. Among the C-C bonds, aliphatic CCA is sterically most
demanding and difficult. Examples are rarely encountered in
unstrained systems.1
in the reactions of Mn2(CO)10 with pentamethylcyclopentadi-
ene,3 in Rh-PCP and Ir-PCP systems,4 and for a Ni-allyl
cyanide complex.5
The rich chemistry of RhII and IrII has been extensively
studied.6,7 In our groups we have a strong focus on gaining a
(2) Martinho Simo˜es, J. A.; Beauchamp, J. L. Chem. ReV. 1990, 90, 629-
688.
(3) Hemond, R. C.; Hughes, R. P.; Locker, H. B. Organometallics 1986, 5,
2391-2392.
(4) (a) Rybtchinski, B.; Vigalok, A.; Ben-David, Y.; Milstein, D. J. Am. Chem.
Soc. 1996, 118, 12406-12415. (b) Sundermann, A.; Uzan, O.; Milstein,
D.; Martin, J. M. L. J. Am. Chem. Soc. 2000, 122, 7095-7104.
(5) Brunkan, N. M.; Brestensky, D. M.; Jones, W. D. J. Am. Chem. Soc. 2004,
126, 3627-3641.
Although CCA is kinetically less favorable than CHA, it is
thermodynamically possible and competitive especially for late
transition metal complexes in which M-C bonds are stronger
(6) (a) Cui, W. H.; Wayland, B. B. J. Am. Chem. Soc. 2006, 128, 10350-
10351. (b) Cui, W. H.; Wayland, B. B. J. Am. Chem. Soc. 2004, 126, 8266-
8274. (c) Cui, W.; Zhang, X. P.; Wayland, B. B. J. Am. Chem. Soc. 2003,
125, 4994-4995. (d) Zhang, X.-X.; Wayland, B. B. J. Am. Chem. Soc.
1994, 116, 7897-7898. (e) Wayland, B. B.; Ba, S.; Sherry, A. E. J. Am.
Chem. Soc. 1991, 113, 5305-5311. (f) Sherry, A. E.; Wayland, B. B. J.
Am. Chem. Soc. 1990, 112, 1259-1261. (g) Farnos, M. D.; Woods, B. A.;
Wayland, B. B. J. Am. Chem. Soc. 1986, 108, 3659-3663.
† Chinese University of Hong Kong.
‡ University of Amsterdam.
(1) (a) Crabtree, R. H. Chem. ReV. 1985, 85, 245-269. (b) Murakami, M.;
Ito, Y. In Topics in Organometallic Chemistry; Murai, S., Ed.; Springer:
Berlin 1999; Vol. 3, pp 97-129. (c) Rybtchinski, B.; Milstein, D. Angew.
Chem. Int. Ed. 1999, 38, 870-883. (d) Jun, C.-H. Chem. Soc. Rev. 2004,
33, 610-618.
9
10.1021/ja078157f CCC: $40.75 © 2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 2051-2061
2051

A R T I C L E S
Chan et al.
Scheme 1. Competing CCA and CHA TEMPO Activation
Pathways upon Reaction with 1a
deeper understanding of the metalloradical reactivity of several
RhII and IrII radicals.8-10 A particular unique chemistry concerns
the activation of aliphatic carbon-carbon bonds in a variety of
organic substrates including nitroxyl radicals,10a,b ketones,10c
amides,10d esters,10d and nitriles10e,f by rhodium(II) meso-
tetramesitylporphyrin, [RhII(tmp)] (1).6e
TEMPO (2,2,6,6-tetramethyl-piperidine-1-oxyl) is often used
as a radical inhibitor in various (catalytic) reactions and can
mediate transition metal-catalyzed oxidation reactions. In these
reactions (the oxygen atom of) its nitroxyl radical moiety is
generally believed to be responsible for its actions, and (radical
mediated) decomposition reactions of TEMPO involving its
aliphatic moieties are generally not considered relevant. Con-
sidering the results presented in this paper this assumption is
not always justified.
In this report we focus on providing a deeper mechanistic
understanding of our previously communicated CCA of un-
strained CR3-Cring bonds of cyclic nitroxyl radicals by RhII.10b
We studied in detail the kinetics of the CCA reaction of [Rh-
(tmp)] (1) with TEMPO 2 leading to [RhIII(tmp)Me] (3). In this
process we identified a competing CHA pathway leading to
formation of TEMPOH via [RhIII(tmp)H]. The selectivity of
CCA over CHA increases at higher temperatures. Herein we
disclose a detailed mechanistic picture supported by DFT
calculations.
a Reactions in benzene under a N2 atmosphere at 70 °C for 4 h.
Table 1. Yields of Rh(tmp)Me and TEMPOH
temp
TEMPO
equiv
%
%
%
Rh(tmp)Me:
TEMPOH
entry
°
C
Rh(tmp)Me
TEMPOH
total yield
1
2
3
4
5
6
7
70
70
70
70
50
60
80
1
2
5
20
20
20
20
60
76
80
82
73
76
85
5.7
8.0
9.0
65.7
83.0
89.0
91.3
89.8
88.3
88.9
10.5:1
9.5:1
8.9:1
8.8:1
4.4:1
6.1:1
21.8:1
9.3
16.8
12.3
3.9
Reaction of Rh(tmp) 1 with TEMPO 2 for 4 h gives rise to
formation of [RhIII(tmp)Me] 3 in high yield (Scheme 1, Table
1).10b Careful analysis of the crude reaction mixture by GC-
MS revealed the presence of 2,2,6,6-tetramethyl-piperidin-1-ol
TEMPOH 4, suggestive for a competing reaction. Other organic
coproducts which can be expected to result from these reactions,
such as nitrone 6 and azaoxetane 7, proved too unstable to be
detected by GC-MS or isolated (Scheme 1).11 However, a
closely related analogue of 6 was detected in the reaction of
[RhII(tmp)] with 2,2- dimethyl-5,5-diphenylpyrrolidin-1-oxyl,
yielding 2-methyl-2,5-diphenyl-3,4-dehydronitrone and [RhIII-
(tmp)Me].10b
Increasing the TEMPO/Rh ratio, while keeping all other
reaction conditions identical (70 °C), gives rise to increased
overall yields of both Rh(tmp)Me 3 and TEMPOH 4, but also
decreases the 3/4 ratio. (Table 1, entries 1-4). Increasing the
TEMPO/Rh ratio from 1 to 20 gradually increases the [RhIII-
(tmp)Me] yield from 60 to 82% and the TEMPOH yield from
5.7 to 9.3%. The highest total product yields (92%) were reached
using 20 equiv of TEMPO, although the yields already leveled
off to 89% when using 5 equiv of TEMPO. Using more than 5
equiv TEMPO gives rise to a nearly constant 3/4 ratio of about
8.8.
We observed a clear influence of the temperature on the
relative rates of the two competing pathways (50-80 °C).
Keeping the TEMPO and [RhII(tmp)] concentrations constant
while increasing the temperature results in higher yields of [RhIII-
(tmp)Me] but lower yields of TEMPOH, thus resulting in a
higher 3/4 ratio (Table 1, entries 5, 6, 4, and 7). The total product
yields remained high and ascertained that other side reactions
were minor.
(7) (a) Paonessa, R. S.; Thomas, N. C.; Halpern, J. J. Am. Chem. Soc. 1985,
107, 43333-4335. (b) Chen, M. J.; Rathke, J. W. Organometallics 1994,
13, 4875-4880.
(8) For recent overviews, see: (a) Hetterscheid, D. G. H.; Gru¨tzmacher, H.;
Koekoek, A. J. J.; de Bruin, B. In The Organometallic Chemistry of Rh,
Ir, Pd and Pt Based Radicals; Higher Valent Species; Karlin, K. D., Ed.;
Vol 55; Wiley: 2007, pp 247-354. (b) de Bruin, B.; Hetterscheid, D. G.
H. Eur. J. Inorg. Chem. 2007, 2, 211-230.
(9) (a) Hetterscheid, D. G. H.; Klop, M.; Kicken, R. J. N. A. M.; Smits, J. M.
M.; Reijerse, E. J.; de Bruin, B. Chem. Eur. J. 2007, 13, 3386-3405. (b)
de Bruin, B.; Russcher, J. C.; Gru¨zmacher, H. J. Organomet. Chem. 2007,
692, 3167-3173. (c) Hetterscheid, D. G. H.; de Bruin, B. J. Mol. Catal.
A. Chem., 2006, 251, 291-296. (d) Hetterscheid, D. G. H.; Kaiser, J.;
Reijerse, E. J.; Peters, T. P. J.; Thewissen, S.; Blok, A. N. J.; Smits, J. M.
M.; de Gelder, R.; de Bruin, B. J. Am. Chem. Soc. 2005, 127, 1895-1905.
(e) Hetterscheid, D. G. H.; Bens, M.; de Bruin, B.; Dalton Trans. 2005,
979-984. (f) Knijnenburg, Q.; Horton, A. D.; van der Heijden, H.; Kooistra,
T. M.; Hetterscheid, D. G. H.; Smits, J. M. M.; de Bruin, B.; Budzelaar, P.
H. M.; Gal, A. W. J. Mol. Catal. A. Chem. 2005, 232, 151-159. (g)
Hetterscheid, D. G. H.; Smits, J. M. M.; de Bruin, B. Organometallics.
2004, 23, 4236-4246. (h) de Bruin, B.; Budzelaar, P. H. M.; Gal, A. W.;
Angew. Chem., Int. Ed. 2004, 43, 4142-4157. (i) Hetterscheid, D. G. H.;
de Bruin, B.; Smits, J. M. M.; Gal, A. W. Organometallics 2003, 22, 3022-
3024. (j) de Bruin, B.; Thewissen, S.; Yuen, T.-W.; Peters, T. P. J.; Smits,
J. M. M.; Gal, A. W.; Organometallics 2002, 21, 4312-4314. (k) de Bruin,
B.; Peters, T. P. J.; Thewissen, S.; Blok, A. N. J.; Wilting, J. B. M.; de
Gelder, R.; Smits, J. M. M.; Gal, A. W. Angew. Chem., Int. Ed. 2002, 41,
2135-2138. (l) Willems, S. T. H.; Russcher, J. C.; Budzelaar, P. H. M.;
de Bruin, B.; de Gelder, R.; Smits, J. M. M.; Gal, A. W. Chem. Commun.
2002, 148-149.
The formation of TEMPOH and the origin of the hydroxyl
hydrogen in TEMPOH atom is puzzling. It seems most likely
that the reaction proceeds via the hydride intermediate [RhIII-
(tmp)H].12 Rh-H bonds of [RhIII(por)] species are weak (∼60
kcal/mol)6d and can donate their hydrogen atoms easily,12b and
TEMPO is known to react rapidly with [MIII(oep)H] hydride
(11) (a) A more stable diphenyl substituted nitrone co-product analogue of 6
has been isolated; see ref 7. (b) Breuer, E.; Aurich, H. G.; Nielsen, A.
Nitrones, Nitronates and Nitroxides: Wiley & Sons: New York, 1989. (c)
A derivative of azaoxetane 7 has been a proposed, unstable intermediate:
Baldwin, J. E.; Bhatnagar, A. K.; Choi, S. C.; Shortridge, T. J. J. Am.
Chem. Soc. 1971, 93, 4082-4084.
(10) (a) Mak, K. W.; Yeung, S. K.; Chan, K. S. Organometallics 2002, 21,
2362-2364. (b) Tse, M. K.; Chan, K. S. J. Chem. Soc. Dalton Trans. 2001,
510-511. (c) Zhang, L.; Chan, K. S. J. Organomet. Chem. 2006, 691,
3782-3787. (d) Zhang, L.; Chan, K. S. J. Organomet. Chem. 2007, 692,
2021-2027. (e) Chan, K. S.; Li, X. Z.; Fung, C. W.; Zhang, L.
Organometallics 2007, 26, 20-21. (f) Chan, K. S.; Li, X. Z.; Fung, C.
W.; Zhang, L. Organometallics 2007, 26, 2679-2687.
(12) (a) Chan, K. S.; Leung, Y. B. Inorg. Chem. 1994, 33, 3187. (b) Zhang, L.;
Chan, K. S. Organometalllics 2007, 26, 679-684.
9
2052 J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008

Carbon−Carbon Bond Activation of TEMPO
A R T I C L E S
species (M ) Rh, Ir) at room temperature to give TEMPOH
and the corresponding metal dimers.12a Also in our hands, [RhII-
(tmp)H] was indeed found to react with TEMPO in benzene to
give [RhII(tmp)] and TEMPOH (eq 4) in >96% yield, so
formation of TEMPOH via an intermediate [RhIII(tmp)H] species
is a very likely pathway. However, this still leaves the question
to what is source of the hydrogen atom. We considered four
possible hydrogen sources: (1) the substrate TEMPO, (2) the
solvent benzene, (3) the starting material [RhII(tmp)], and (4)
the product [RhIII(tmp)Me]. Each of these sources can have its
unique hydrogen transfer mechanistic pathway. We performed
some additional experiments aiming at revealing the most likely
hydrogen source.
Since TEMPO proved stable in benzene at 70 °C for 24 h
without any decomposition (determined by GC-MS analysis
using an internal standard), we rule out a direct intermolecular
hydrogen atom transfer from benzene (or another TEMPO
moiety) to TEMPO.13 Also [RhII(tmp)] is stable in benzene at
130 °C for at least 2 days.6e,12b Even though [RhIII(tmp)Me]
reacts with [RhII(tmp)] at 130 °C to give [RhIII(tmp)Et] via the
proposed intermediate [RhIII(tmp)H],10f no reaction is observed
at all at 70 °C. Therefore, [RhIII(tmp)Me] cannot be the hydrogen
source for formation of TEMPOH. Hydrogen atom transfer from
benzene, [RhIII(tmp)Me] or another [RhII(tmp)] moiety to [RhII-
(tmp)] is clearly difficult, so we can safely exclude these
reagents serving as a hydrogen source for TEMPOH (sources
3 and 4). This leaves a hydrogen atom transfer pathway
involving CHA of TEMPO by [RhII(tmp)] to give the [RhIII-
(tmp)H] intermediate, which in turn transfers its hydrogen atom
to another TEMPO moiety leading to TEMPOH with reforma-
tion of RhII(tmp)] (source 1 via [RhIII(tmp)H]) as the most likely
pathway (Scheme 1).
Figure 1. UV-vis titration spectra of Rh(tmp) with TEMPO at 10.0 °C.
Table 2. Summary of Spectrophotometric Titration of Rh(tmp) with
TEMPO
T
1/T
TEMPO:
Rh(tmp)
1
entry
(
°C)
(K-
)
log K1
R
1
2
3
4
5
6
10.0
15.0
20.0
25.0
30.0
70.0
0.00353
0.00347
0.00341
0.00336
0.00330
0.98 ( 0.02
1.01 ( 0.01
0.98 ( 0.02
0.98 ( 0.01
0.99 ( 0.01
4.62 ( 0.09
4.53 ( 0.06
4.35 ( 0.07
4.06 ( 0.04
3.93 ( 0.05
2.69 (extrapolated)
0.996
0.999
0.998
0.997
0.995
The formation of [RhIII(tmp)H] through a chelation-assisted
CHA by [RhII(tmp)] is kinetically favorable.12b While this C-H
bond activation is apparently uphill (since a weaker Rh-H bond
is formed by breaking a strong C-H bond (Rh(tmp)-H ) 60
kcal/mol,6e Me3CCH2-H ) 100 kcal/mol),13 the formation of
the organic coproduct azaoxetane 7 could render the process
energetically favorable. Furthermore, the CHA reaction could
be simply driven by the fast and favorable follow-up reaction
of [RhIII(tmp)H] with TEMPO to give TEMPOH and [RhII-
(tmp)]. This would imply that the TEMPOH yield should
increase at higher TEMPO concentrations. Indeed, increasing
the [TEMPO] from 1 to 20 equiv increases the TEMPOH yield.
Not only in absolute sense, but the CHA pathway is also
relatively more favored at higher [TEMPO] compared to the
CCA, as can be derived from the decreasing 3/4 product ratio
(Table 1, entries 1 to 4). This implies that the overall CHA
activation process likely involves more equiv of TEMPO (likely
2) than the CCA process (likely 1).
Figure 2. Measured absorbance against concentration of TEMPO added
at 10 °C.
Since the yields of the CCA and CHA reactions are both
dependent on the TEMPO concentration we decided to evaluate
the binding constant K1 () k1/k-1) for TEMPO binding to [RhII-
(tmp)]. The binding stoichiometry and binding constant were
evaluated spectrophotometrically from changes in absorbance
at 521 nm in the temperature range from 10 to 30 °C (Figure 1
and Table 2). Analyses of the data confirmed a 1:1 adduct (See
Figure 2, for details see Experimental Section) and yielded a
binding constant K1 () k1/k-1) ) 2.24 × 104 ( 1.2 M-1 at
20.0 °C. A van’t Hoff plot of the binding constants at different
temperatures provided the enthalpy and entropy values: ∆H1
) -14.7 ( 1.4 kcal mol-1, and ∆S1 ) -30.6 ( 4.6 cal mol-1
K-1 (Figure 3). The extrapolated value of K1 at 70 °C (the
reaction temperature at which the competing CCA and CHA
reactions were studied) is also listed in Table 2 for a subsequent
comparison with the kinetic data below.
Binding Studies and Kinetic Investigations. To shed some
more light on the reaction mechanism, we investigated the
binding affinity of TEMPO to [RhII(tmp)], the kinetics of the
CCA and CHA reactions as well as the kinetics associated with
the hydrogen atom transfer from [RhIII(tmp)H] to TEMPO.
(13) This is understandable, because the energy of the TEMPOH O-H bond is
only 70 kcal/mol and thus not strong enough to allow hydrogen atom
transfer from the much stronger alkyl C-H bonds of TEMPO or benzene
(Me3CCH2-H ) 100 kcal/mol, Ph-H ) 113 kcal/mol). See: Luo, Y.-R.
Handbook of Bond Dissociation Energies in Organic Compounds; CRC
Press: Boca Raton, FL, 2003.
The kinetics of the reaction between [RhII(tmp)] 1 and
TEMPO 2 were investigated spectophotometrically by following
9
J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008 2053

A R T I C L E S
Chan et al.
rate ) kobs[[RhII(tmp)]]n[TEMPO]m
) k′obs[[RhII(tmp)]]
(1)
(2)
(3)
) kobs[[RhII(tmp)]][TEMPO]
The temperature dependent observed rate constants k′obs were
measured in the temperature range from 50-80 °C and are
tabulated in Table 4. An Erying plot (Figure S1) yielded the
q
corresponding enthalpies and entropies of activation: ∆Hobs
q
) 15.02 ( 0.71 kcal mol-1 and ∆Sobs ) -20.93 ( 0.88 cal
K-1 mol-1. Given the product distributions in the applied
temperature range (Table 1), the obtained activation parameters
should mainly reflect those of the major CCA process.
To ascertain that the rate determining step of the CHA
pathway is the CHA step and not the follow-up reaction, we
investigated the kinetics of the reaction between [RhIII(tmp)H]
5 and TEMPO 2. TEMPO reacted with [RhIII(tmp)H] at room
temperature in 10 min to give TEMPOH in 96% yield as
determined by GC-MS analysis (eq 4). Since [RhII(tmp)] is
air-sensitive and difficult to purify, its yield was measured
indirectly as [RhIII(tmp)I], which is formed from [RhII(tmp)]
by a subsequent reaction with added excess iodine. [RhIII(tmp)I]
was obtained in 93% yield (average of three runs) after column
chromatography. With the assumption of quantitative yield of
iodination (2 [RhII(tmp)] + I2 f 2 [RhIII(tmp)I]), the yield of
[RhII(tmp)] from the reaction with TEMPO and [RhIII(tmp)H]
was estimated at 93% (eq 4).
Figure 3. van’t Hoff plot of binding of Rh(tmp) with TEMPO.
C6H6, N2
Rh(tmp)H + TEMPO
8 Rh(tmp) +
(4)
TEMPOH
96%
rt, 10 min
93%
Figure 4. Typical time scan of the reaction between [RhII(tmp)] and
TEMPO.
With the follow-up hydrogen atom transfer (HAT) reaction of
[RhIII(tmp)H] with TEMPO being established as a clean reaction,
the kinetics of the reaction were followed spectrophotometrically
at 522 nm under the conditions 15.0 °C, initial concentrations
(2.96-8.87) × 10-5 M [RhIII(tmp)H] and (9.11-5.47) × 10-3
M TEMPO with TEMPO always in at least 10-fold excess
(Table 5). The observed rate constants k′HAT,obs follow a first-
order decay in [RhIII(tmp)H] and the observed rate constants
did not change with the initial concentration of [RhIII(tmp)H]
(Table 5, entries 1-3, Figure 5). Therefore, the kinetic order
in [[RhIII(tmp)H]] was evaluated to be one. A linear plot of
k′HAT,obs () kHAT[Rh(tmp)H]) against [TEMPO] revealed a first-
order reaction in [TEMPO] (Figure 6). The rate law conformed
to: rate ) kHAT [Rh(tmp)H] [TEMPO] where kHAT ) 0.21 L
mol-1 s-1. The rates were also measured in the temperature
range from 15 to 30 °C and the data are tabulated in Table 6.
An Eyring plot (Figure 7) gave the corresponding enthalpies
Table 3. k′obs under Different TEMPO and Rh(tmp)
Concentrations at 70 °C
1
entry
[Rh(tmp)] (10-5 M)
[TEMPO] (10-4 M)
k
′
obs (10-5 s-
)
1
2
3
4
5
6
2.775
5.549
5.549
5.549
5.549
5.549
8.282
8.282
16.56
24.97
33.13
49.70
8.79 ( 0.01
8.74 ( 0.02
16.99 ( 0.05
24.45 ( 0.05
34.97 ( 0.08
49.47 ( 0.08
changes in the absorbance at 523 nm under the reaction
conditions 70.0 °C, initial concentrations (2.775 - 8.324) ×
10-5 M Rh(tmp) and (0.828 - 4.970) × 10-3 M TEMPO with
TEMPO always in at least 10-fold excess. The reactions were
monitored for at least four half-lives.
The kinetics conformed to first-order reactions in both [1]
and [2]. The rate laws are defined in eqs 1-3. With excess
TEMPO, the observed rate constants k′obs (eq 3, Figure 4) were
evaluated by first-order fitting of the absorbance changes and
did not change with the initial concentrations of [RhII(tmp)]
(Table 3, entries 1-2). The reaction order in [1] (n) is clearly
equal to one. Variation of the concentration TEMPO at fixed
concentration of [RhII(tmp)] yielded a linear plot (Table 3,
entries 3-6, Figure 3), in good agreement with a reaction order
in [2] (m) equal to one. The resulting overall rate law is shown
and entropies of activation: ∆HHATq ) 21.90 ( 0.33 kcal mol-1
and ∆SHAT ) 14.43 ( 0.26 cal K-1 mol-1
.
q
The second-order kinetics suggest a direct hydrogen atom
transfer from [RhIII(tmp)H] to TEMPO in the transition state.
The positive value of ∆Sq
is rather intriguing, and may
HAT
indicate a dissociative activation process, perhaps related to
breaking of the Rh-H bond. A bimolecular associative mech-
anism is not consistent with these data. However, the obtained
rate constant kHAT might not be an elementary one, and we have
no further mechanistic information. Besides a direct HAT
in eq 3 where kobs ) 0.11 L mol-1 s-1
.
9
2054 J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008

Carbon−Carbon Bond Activation of TEMPO
A R T I C L E S
Table 4. Observed Rate Constants kobs at Different Temperatures
Table 6. Temperature Dependent Observed Rate Constant k′HAT,
obs and kHAT
1
1
1
entry
T (°
C)
1/T (K-
)
k′
obs (10-4 s-
)
k
obs (L mol-1 s-
)
ln(kobs/T)
k
′
kHAT
HAT, obs
1
2
3
4
5
6
50.0
55.0
60.0
70.0
75.0
80.0
0.00310
0.00305
0.00300
0.00292
0.00287
0.00283
0.772 ( 0.005
1.107 ( 0.009
1.698 ( 0.002
3.497 ( 0.075
5.217 ( 0.014
6.969 ( 0.022
0.0233
0.0334
0.0512
0.1056
0.1575
0.2104
-9.537
-9.192
-8.779
-8.086
-7.701
-7.425
1
1
1
temp (
°
C)
1/T (K-
)
(10-4 s-
)
(L mol-1 s-
)
ln(kHAT/T)
15.0
20.0
25.0
30.0
0.00347
0.00341
0.00336
0.00330
1.94 ( 0.04
3.68 ( 0.07
6.70 ( 0.12
13.2 ( 0.18
0.2134
0.4045
0.7358
1.4493
-7.2074
-6.5853
-6.0038
-5.3427
Table 5. k′HAT,obs under Different Rh(tmp)H Concentrations at
15 °C
[Rh(tmp)H]
10-5 M)
[TEMPO]
k′
HAT,obs
1
entry
(
×
(
×
10-4 M)
(× )
10-4 s-
1
2
3
4
5
6
7
2.958
5.916
8.874
5.916
5.916
5.916
5.916
9.108
9.108
9.108
18.22
27.32
36.43
54.65
1.965 ( 0.042
1.944 ( 0.041
1.973 ( 0.032
4.022 ( 0.073
6.267 ( 0.018
8.256 ( 0.026
13.21 ( 0.028
Figure 7. Eyring plot of reaction between Rh(tmp)H and TEMPO.
a separate paper. In this paper we further focus on the
mechanistic details of the CCA (and necessarily the competing
CHA).
It is clear that the measured hydrogen atom transfer rate
constant kHAT (0.21 L mol-1 s-1 at 15 °C) is much larger than
the overall rate constant for the competing CCA and CHA
reactions kobs (0.11 L mol-1 s-1 at 70 °C). So clearly the
preceding CHA and CCA steps are rate-limiting in the two
competing pathways.
Figure 5. Linear relationship between k′obs and [TEMPO].
The extrapolated binding constant K1 for TEMPO binding
to [RhII(tmp)] (488 M-1) is still relatively large even at 70 °C.
This binding constant implies that up to 70% of [RhII(tmp)] is
associated with TEMPO as the adduct [Rh(tmp)(TEMPO)] 8
under the applied reaction conditions. The rate equation: rate
) {(k2 + k3)K1[TEMPO][Rh(tmp)]0}/{1 + K1[TEMPO]} would
apply for competitive CHA and CCA reactions each proceeding
via 8 (without subsequent isomerization of 8 to 8′, see
Supporting Information for a derivation). This equation cannot
be further simplified (with K1[T] ) 488 M-1 × 3.313 × 10-3
M ) 1.61 being of the same order as one) and such the binding
constant K1 ) 488 M-1 would without doubt lead to saturation
kinetics if the rate-limiting steps would follow directly from 8.
This is not compatible with the clear experimental first-order
kinetic behavior in both [TEMPO] and [[RhII(tmp)]]0. This
implies that formation of 8 might not be productive, or its
formation is followed by subsequent reaction with a much
smaller equilibrium constant K2. We considered the possibility
that the Rh(tmp)-η1-O-TEMPO complex 8 could react with
another molecule of TEMPO to undergo the CCA and CHA
reactions. However, kinetic modeling would yield the rate
equation: rate ) {(k2 + k3)K1[TEMPO][Rh(tmp)]0} [TEMPO]/
{1 + K1[TEMPO]} (Supporting Information (SI), eq 16 and
Figure 6. Linear relationship between k′HAT obs and [TEMPO].
reaction, a proton coupled electron transfer (PCET)14 might be
operative. We are currently working on obtaining kinetic isotope
data and DFT calculations for this reaction, to be published in
(14) For a leading reference of PCET, see: Rhile, I. H.; Mayer, J. M. Angew.
Chem., Int. Ed. 2005, 44, 1598-1599.
9
J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008 2055

A R T I C L E S
Chan et al.
Scheme 2 Kinetic Pathways for CCA and CHA of TEMPO by [RhII(tmp)] Consistent with the Experimental Data
17), which also cannot be further simplified and, more impor-
tantly, is not consistent with the observed first-order kinetics
in [TEMPO].
mode, for which a couple of precedents have been reported.16
The hindered tetramesityl porphyrin is however expected to
impose a rather unfavorable steric constrain for this binding
mode. The influence of steric hindrance is illustrated by reported
change in the TEMPO coordination modes: while TEMPO is
η2-Ν,Ã coordinated in TiCl3(TEMPO),16i its is η1-O coordinated
in the more hindered complex CpTiCl2(TEMPO).16e η2-N.O
TEMPO coordination to Rh(tmp) might thus not be favorable.
Heterolytic Rh-O bond splitting and electron transfer are
possible alternatives for these equilibria.
DFT Calculations. The reaction mechanism of the experi-
mentally observed methyl transfer from TEMPO to [RhII(por)]
was further investigated with DFT. We used nonfunctionalized
rhodium-porhyrinato complexes [Rh(por)] as model compounds
for the experimental [Rh(tmp)] systems.
We first considered a two-step mechanism, in which TEMPO
first splits into a methyl radical and the DMPO-like product
2,2,6-trimethyl-2,3,4,5-tetrahydro-1-pyridiniumolate via ho-
molytic C-C bond cleavage, which could then be followed by
capture of the methyl radical by [RhII(por)]. The first step of
this mechanism is prohibitedly endergonic according to the DFT
calculations (Scheme 3).
However, the calculations do reveal that the respective
TEMPO C-CMe bond is rather weak, with an overall thermo-
dynamic bond dissociation enthalpy (BDE) of only +32 kcal/
mol (compared to BDE values >60 kcal/mol for regular alkane
This leaves two possibilities each for both the CCA and CHA
pathways: (I) formation of 8 is not productive, and the reaction
proceeds without any pre-coordination of TEMPO to Rh, or
(II) formation of 8 is followed by a subsequent equilibration
step 8 T 8′ with a much smaller equilibrium constant K2
(Scheme 2).15 If we leave all remaining possibilities open we
arrive at the generalized rate equation: rate ) (k3K1K2 + k′3 +
k4K1K2 + k′4)[TEMPO][Rh(tmp)]0, which conforms to the
experimentally derived rate eq 3 with kobs ) (k3K1K2 + k′3 +
k4K1K2 + k4). Although the k3 and k′3 pathways are not likely
to have similar rate constants, and k4 and k′4 are neither likely
to be of the same order, the data do not allow us to discriminate
between the possible pathways (k′n . knK1K2 vs knK1K2 . k′n).
In any case, given the product distributions (Table 1), the
kinetics must be dominated by the CCA pathway (kobs ≈ k3K1K2
+ k′3).
Although the carbon-carbon activation of TEMPO by [Rh-
(tmp)] may well proceed without pre-coordination, it is note-
worthy that all substrates for which CCA by [Rh(tmp)] has been
observed are coordinating ones, such as ketones,10c amides,10d
esters,10d and nitriles.10e,f Metalloradical aliphatic carbon-carbon
activation of non-coordinating organic compound has not been
reported yet. Several reports of CHA by [RhII(por)] species have
been reported (for an overview, see ref 8a), and we cannot rule
out the direct hydrogen abstraction k4′ pathway for CHA of
TEMPO, as even the C-H bonds of the “non-coordinating”
CH4 can be cleaved between two [RhII(tmp)] radicals.6
(16) For references of TEMPO-transition metal complexes: (a) η1-O-(neutral)
Rh(II) complex: Felthouse, T. R.; Dong, T.-Y.; Hendrickson, D. N.; Shieh,
H.-S.; Thompson, M. R. J. Am. Chem. Soc. 1986, 108, 8201-8214. (b)
η1-O(neutral)-Cu(II) complex: Baskett, M.; Lathi, P. M.; Palacio, F.
Polyhedron, 2003, 22, 2363-2374. (c) η1-O(neutral)-Ru(II) porphyrin:
Selyer, J. W.; Fanwick, P. E.; Leidner, C. R. Inorg. Chem. 1992, 31, 3699-
3700. (d) η1-O(anionic)-Si(IV) porphryin: Zheng, J.-Y.; Konishi, K.; Aida,
T. J. Am. Chem. Soc. 1998, 120, 9838-9843. (e) η1-O(anionic)-Ti(IV)
complex: Mahanthappa, M. K.; Huang, K.-W.; Cole, A. P.; Waymouth,
R. M. Chem. Commun. 2002, 502-503. (f) η1-O(anionic)-Ir(III) com-
plex: Hetterscheid, D. G. H.; Kaiser, J.; Reijerse, E.; Peters, T. P. J.;
Thewissen, S.; Blok, A. N. J.; Smits, J. M. M.; de Gelder, R.; de Bruin, B.
J. Am. Chem. Soc. 2005, 127, 1895-1905. (g) η1-O(anionic)-µ-η1:η2-ON
Sm(III) complex: Evans, W. J.; Perotti, J. M.; Doedens, R. J.; Ziller, J.
W. Chem. Commun. 2001, 2326-2327. (h) η2-O(anionic)N-Ni(II) com-
plex: Mindiola, D. J.; Waterman, R.; Jenkins, D. M.; Hillhouse, G. L.
Inorg. Chim. Acta 2003, 345, 299-308. (i) η2-O (anionic)N-Ti(IV)
complex: ref 16e.
The above 8 T 8′ and 1 + 2 T 8′ equilibria, possibly relevant
for the unusual CCA pathway, could involve TEMPO η2-O,N-
side-on binding instead of the expected η1-O end-on binding
(15) In case of rate limiting steps following formation of species 8′, K2 must be
relatively small (<1) so to make K1K2[TEMPO] , 1, thus simplifying the
individual rate equations rate ) {knK1K2[TEMPO][Rh(tmp)]0}/{1 + K1K2-
[TEMPO]} to rate ) kobs[T][Rh(tmp)]0 (with kobs ) knK1K2 and n ) 3 or
4).
9
2056 J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008

Carbon−Carbon Bond Activation of TEMPO
A R T I C L E S
The binding of TEMPO to rhodium could in principle involve
formation of the expected η1-O-bound TEMPO adduct II
(Scheme 4), or alternatively its η1-N-bound or η2-N,O-bound
isomer. Attempts to optimize (singlet) η1-N-bound or η2-N,O-
bound TEMPO adducts all converged to the η1-O-bound
TEMPO adduct II. The optimized geometry of II corresponds
well with our previously reported diamagnetic η1-O-bound
TEMPO-IrIII adduct formed from TEMPO and an IrII species.16f
Optimizing complex II in a triplet spin state led to TEMPO
dissociation from Rh, and (broken symmetry) attempts to
optimize II as singlet biradical did not lead to a lower energy.
The energy values associated with TEMPO capture by the [RhII-
(por)] radical to form the diamagnetic η1-O-bound TEMPO
adduct [RhIII(por)(TEMPO)] II are not reliable, since they
involve comparing systems of different spin states.18-20 This
problem is evident from the calculated exothermicity of forma-
tion of II at the bp86-SV(P) level (in reasonable agreement with
the experimentally derived ∆G0 ) -5.3 kcal/mol value associ-
ated with TEMPO binding), while at the b3-lyp-TZVP level of
theory formation of II from the TEMPO and [RhII(por)] radicals
is endothermic. Apparently, the hybrid b3-lyp functional
overestimates the relative stability of the separated radicals.20
Relative energies of the (closed-shell) singlet species formed
after the TEMPO capture to form II should be more reliable,
and for these subsequent steps we found no large differences
between the bp86 and b3-lyp levels of theory. We thus decided
to take the TEMPO adduct II as a reference point for comparing
the relative energies (corresponding with the experimental
data).
Figure 8. SOMO (left) and spin density (right) plots of TEMPO, revealing
significant unpaired electron density at the “axial” methyl fragments (b3-
lyp, TZVP).
Scheme 3. Calculated Standard Free Energies (kcal/mol),
Standard Enthalpies (kcal/mol), Standard Entropies
(cal‚mol-1‚deg-1) and Bond Distances Associated with Methyl
Radical Dissociation from TEMPO (Homolytic C-C Bond Splitting,
b3-lyp, TZVP)
Table 7. Energies and Free Energies Associated with the
Calculated Reaction Paths for Me Transfer from TEMPO to
[RhII(por)]
Ic
IId TS1e
IIId
TS2d
IVd
(17) Electron transfer from TEMPO to [RhII(por)] (formation of III, Scheme
4) is noteworthy. [RhII(por)] species and nitroxyl radicals are both redox
active and both have the ability to act as reducing and as oxidizing
agents. The reverse SET direction (from Rh to TEMPO) seems applicable
in the formation of the TEMPO adduct [RhIII(por)(TEMPO-)] II. Hypo-
thetical outer-sphere ET from [RhII(tmp)] to TEMPO would yield a
[RhIII(por)]+[TEMPO]- cation-anion pair. This is perhaps possible in polar
solvents, but in the gas-phase or in benzene it would easily collapse to
form the [Rh(por)(TEMPO)] adduct II. In previous reports, we have already
shown that the methyl transfer reaction cannot proceed via reaction of
neutral nitroxyl radicals with [RhIII(por)]+ cations or [RhI(por)]- anions.10b
(18) Gosh, A. J. Biol. Inorg. Chem. 2006, 11, 712-724.
∆E
bp86, SVPa
∆E
+10.7 ) 0
-7.2 ) 0
-18.2 ) 0
-16.6 ) 0
-
+13.9
+20.4
-5.0
+14.3 +15.7
+11.7 +11.5
+11.7 +11.6
+22.7 -13.8
+17.0 -23.6
+17.4 -22.6
b3-lyp, TZVPa
∆G0
298K
b3-lypb
∆G353K
b3-lypb
a Energies (kcal/mol). b Free energies (kcal/mol) corrected for dielectric
solvent effects (COSMO, ꢀ ) 2.28). c Two gas phase separated doublets.
d Singlet. e Open-shell singlet (singlet biradical): broken symmetry, unre-
stricted b3-lyp solution.
(19) Harvey, J. N.; Poli, R.; Smith, K. M. Coord. Chem. ReV. 2003, 238-239,
347-361.
(20) This is not the first time that hybrid HF/DFT functionals (such as b3-lyp)
are found to overestimate the relative stability of radicals compared to
closed-shell systems: (a) Saeys, M.; Reyniers, M.-F.; Marin, G. B.; Van
Speybroeck, V.; Waroquir, J. Phys. Chem. A 2003, 107, 9147-9159. (b)
Sustmann, R.; Sicking, W.; Huisgen, R.; J. Am. Chem. Soc. 2003, 125,
14425-14434. (c) Jensen, K. P.; Ryde, U.; J. Phys. Chem. A 2003, 107,
7539-7545. (d) Gosh, A.; J. Biol. Inorg. Chem. 2006, 11, 712-724.
(21) (a) Ozinskas, A. J.; Bobst, A. M. HelV. Chim. Acta 1980, 63, 1407-1411.
(b) Fields, J. D.; Kropp, P. J. J. Org. Chem. 2000, 65, 5937-5941.
(22) Wayland, B. B.; Van Voorhees, S. L.; Wilker, C. Inorg. Chem. 1986, 25,
4039-4042.
C-C bonds). This is clearly the result of concerted formation
of a CdN π-bond associated with the homolytic C-C splitting.
The calculations did reveal a significant spin density located
at the two “axial” methyl fragments of TEMPO, caused by some
(antibonding) mixing-in of the respective C-C bonding orbitals
with the half-filled N-O π* orbital (see Figure 8). This insight
led us to investigate a direct methyl radical transfer pathway
from TEMPO to [RhII(por)] along the “broken symmetry”
(unrestricted b3-lyp, open-shell singlet) singlet biradical potential
surface (i.e., a bimolecular homolytic substitution; SH2). The
experimental results clearly reveal that TEMPO binds to [RhII-
(por)], with the kinetics not being conclusive whether this is a
productive step in the formation of the final [RhIII(por)Me]
product, or not. Therefore we also considered an alternative
pathway, proceeding via the [TEMPO]+ [RhI(por)]- cation-
anion pair, obtainable via electron transfer from the TEMPO
radical to [RhII(por)] (I) or via heterolytic bond splitting of the
Rh-O bond of the TEMPO adduct (II), respectively (Scheme
4).17 The energies associated with the species depicted in
Scheme 4 are shown in Table 7.
(23) Perkampus H. H. UV-VIS Spectroscopy and Its Applications; Springer-
Verlag: Berlin, Heidelberg, 1992.
(24) Atkins, P. Physical Chemistry, 6th ed.; Oxford University Press: Oxford,
U.K., 1998.
(25) (a) Ahlrichs, R; et al. Turbomole, version 5; Theoretical Chemistry Group,
University of Karlsruhe: Karlsruhe, Germany, 2002. (b) Treutler, O.;
Ahlrichs, R. J. Chem. Phys. 1995, 102, 346-354. (c) Turbomole, version
5, Turbomole basisset library (see ref 25a). (d) Scha¨fer, A.; Horn, H.;
Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571-2577. (e) Andrae, D.;
Haeussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. Theor. Chim. Acta 1990,
77, 123-141. (f) Scha¨fer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994,
100, 5829-5835. (g) Ahlrichs, R.; May, K. Chem. Phys. 2000, 2, 943.
(26) (a) PQS, version 2.4; Parallel Quantum Solutions: Fayetteville, AR, 2001
(the Baker optimizer is available separately from PQS upon request). (b)
Baker, J. J. Comput. Chem. 1986, 7, 385-395.
(27) (a) Becke, A. D. Phys. ReV. A 1988, 38, 3098-3100. (b) Perdew, J. P.
Phys. ReV. B 1986, 33, 8822-8824.
(28) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789. (b)
Becke, A. D. J. Chem. Phys. 1993, 98, 1372-1377. (c) Becke, A. D. J.
Chem. Phys. 1993, 98, 5648-5652. (d) Calculations were performed using
the Turbomole functional “b3-lyp”, which is not identical to the Gaussian
“B3LYP” functional.
9
J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008 2057

A R T I C L E S
Chan et al.
Figure 9. DFT optimized geometries (b3-lyp, TZVP) of II, TS1, III, TS2, and IV.
Scheme 4. Computational Reaction Mechanisms for Methyl Transfer from TEMPO to [RhII(por)], in Good Agreement with the Experimental
Kinetic Data
Formation of the cation-anion pair III from II is about +12
kcal/mol uphill. From this cation-anion pair we readily obtained
an SN2-like transition state (TS2) for methyl transfer from
TEMPO to Rh. This transition state involves a nucleophilic
attack of the electron rich RhI atom to one of the methyl groups
of the TEMPO cation, with (DMPO-like) 2,2,6-trimethyl-
2,3,4,5-tetrahydro-1-pyridiniumolate acting as an efficient leav-
ing group (SN2 type mechanism, see Scheme 4). TS2 lays only
about +17 kcal/mol uphill with respect to II. Formation of the
final products IV from II is exergonic by nearly -24 kcal/mol.
The transition state for direct Me radical transfer from the
TEMPO radical to the [RhII(por)] radical (SH2-type mechanism),
along the open-shell singlet (singlet biradical) surface, has an
even lower relative energy (∆Gq ) +12 kcal/mol). On this basis,
the actual reaction might well take place via direct methyl radical
transfer (TS1) instead of the alternative ionic pathway (TS2).
However, the hybrid b3-lyp functional seems to overestimate
the relative stability of radicals.20 It is not so clear if, and to
what extend, this also holds for the singlet biradicaloid transition
state TS1 (which is “on the way” to become closed-shell; S2
9
2058 J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008

Carbon−Carbon Bond Activation of TEMPO
A R T I C L E S
from 1.5 to 1. The “oxidized” TEMPO moiety in III has no
electrons in the N-O π* antibonding orbital, leading to a net
bond order of 2. The nucleophilic attack of RhI at the oxidized
TEMPO of III leads to a net “reduction” with formation of the
(DMPO-like) nitrone 2,2,6-trimethyl-2,3,4,5-tetrahydro-1-pyri-
diniumolate of IV via TS2. As expected, the N-O bond
distances increase and the C-N bond distances decrease along
the (SN2-type) ionic pathway from III via TS2 to IV. However,
the changes of bond distances along the (SH2-type) radical
pathway I f TS1 f IV are less clear. On going from I to TS1
the N-O bond distance first decreases to eventually become
even longer in IV (compared to I). It is clear that TS1 is
optimized as a mixture of the “real” singlet biradical and its
triplet state. This is one possible explanation for the observed
discrepancy in the behavior of the N-O bond lengths. Another
possible explanation is that some partial electron transfer from
TEMPO to Rh also occurs in TS1, giving rise to a shortening
of the N-O distance in the transition state.
The obtained barriers TS1 and TS2 for the ionic (SN2) and
the direct radical transfer (SH2) pathways are rather low,
suggesting that the reaction might even proceed at room
temperature. However, considering the fact that the experimental
[RhII(tmp)] system is sterically much more shielded, the
calculated mechanistic pathways seem both qualitatively and
quantitatively in good agreement with the experimental results.
Since both pathways are consistent with the experimental kinetic
data, we are not able to make a distinction (neither experimen-
tally nor computationally) between these two possibilities.
Conclusions
Major carbon-carbon bond activation (CCA) and minor
carbon-hydrogen bond activation (CHA) in reaction between
rhodium(II) meso-tetramesitylporphyrin Rh(tmp) 1 and 2,2,6,6-
tetramethyl-piperidine-1-oxyl (TEMPO) 2 are both associated
with first-order kinetics in TEMPO and [RhII(tmp)]. The
selectivity for CCA increases at higher temperatures. U-DFT
reveals a low-energy transition state (TS1) for direct abstraction
of a methyl radical from TEMPO by the RhII radical (SH2-type
mechanism). R-DFT reveals an alternative ionic pathway along
the closed-shell singlet surface. This ionic pathway proceeds
in two sequential steps: electron transfer from TEMPO to [RhII-
(por)] producing the [TEMPO]+[RhI(por)]- cation-anion pair,
followed by CH3+ transfer from TEMPO+ to RhI with formation
of [RhIII(por)Me] and a (DMPO-like) 2,2,6-trimethyl-2,3,4,5-
tetrahydro-1-pyridiniumolate leaving group. The transition state
for this process (TS2) is best described as an SN2-like nucleo-
philic substitution involving attack of the dz2 orbital of
[RhI(por)]- at a C-C σ* orbital of [TEMPO]+. The R-DFT
and U-DFT calculations are comparatively not accurate enough
to reliably distinguish between these possible pathways. Both
the radical (SH2) and the ionic (SN2) pathway have barriers
which are low enough to explain the experimental data, and
both pathways are consistent with the experimental kinetic data.
These remarkable and unprecedented reactivity patterns for
TEMPO suggest that TEMPO can act as methyl radical donor,
a reducing agent, and (in its oxidized form) as a CH3+ donor to
(late) transition metals. Since TEMPO is often used as a radical
inhibitor in various (catalytic) reactions, the possible implications
of these new insights need to be further explored.
Figure 10. Selected bond-length changes associated with the methyl group
transfer pathways.
) 0.47). It might well be that the stability of TS1 relative to
TS2 is also overestimated. This seems to be supported by the
fact that at the (pure DFT) bp86 level of theory, TS1 does not
exist (all attemps to find TS1 at the (u)bp86 level led to TS2).
On the other hand, the computational restrictions associated with
single determinant (u)b3-lyp optimization of TS1 unavoidably
led us to optimize its geometry as a mixture of its real
(multideterminant singlet biradical) electronic structure and its
triplet state.29 This is likely to result in a too high energy. To
what extent these effects counterbalance is not clear, and
therefore we are unable to reliably distinguish between the ionic
(TS2) and direct radical abstraction (TS1) pathways on the basis
of these DFT calculations.
The optimized geometries of II, III, IV, TS1, and TS2 are
depicted in Figure 9. The bond length changes are represented
in Figure 10, which require some additional comments. It is
clear that the N-O bond distance reflects the “oxidation state”
of the TEMPO moiety. Going from I to II, the TEMPO moiety
is reduced by one electron, which decreases the N-O bond order
(29) Knijnenburg, Q.; Hetterscheid, D.; Kooistra, T. M.; Budzelaar, P. H. M.;
Eur. J. Inorg. Chem. 2004, 1204-1211 and references cited therein.
9
J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008 2059

A R T I C L E S
Chan et al.
color. The reaction mixture was then cooled down to 0 °C under N2,
and HCl (40 mL, 0.1 M) was added via syringe. An orange suspension
formed immediately and was stirred for 15 min at 0 °C. The workup
was carried out under N2. The reaction mixture was worked up by the
addition with degassed benzene/H2O. The crude product was extracted
with degassed benzene (50 mL), washed with H2O (10 mL × 3), dried
over MgSO4, filtered under N2 using a cannular with the tip wrapped
with a filter paper. Then the solvent was removed under high vacuum.
A red solid (41.6 mg, 0.047 mmol, 96%) was obtained. 1H NMR (C6D6,
300 MHz): δ -40.07 (d, 1 H, 1JRhH ) 45 Hz), 1.80 (s, 12 H), 2.14 (s,
12 H), 2.44 (s, 12 H), 7.10 (s, 4 H), 7.19 (s, 4 H), 8.77 (s, 8 H).
Reaction of TEMPO with [RhIII(tmp)H]. TEMPO (38 mg, 0.24
mmol) was added to Rh(tmp)H (42 mg, 0.048 mmol) in anhydrous
C6H6 (30 mL) under N2. The mixture was stirred at room temperature
(rt) for 30 min. Then, the reaction mixture was analyzed by GC-MS
with anthracene (30 µL, 0.0749 M) as the internal standard, and
TEMPOH was obtained in 96% yield. The addition of excess iodine
to the reaction mixture yielded Rh(tmp)I (45 mg, 0.045 mmol, 93%)
after column chromatography. Assuming the yield to be 100% in the
reaction between Rh(tmp) and I2, the yield of Rh(tmp) was estimated
to be 93%.
Kinetic Studies of the Reaction between TEMPO (2) and [RhII-
(tmp)] (1). Benzene was freshly distilled over sodium, degassed by
the freeze-pump-thaw method (three cycles) and refilled with N2.
To a Teflon screw-head stoppered flask, [RhIII(tmp)CH3] 3 (10.0 mg,
0.011 mmol) was added carefully. Then dried and distilled C6H6 (4.000
mL) was added with a gastight syringe. The reaction mixture was then
degassed by the freeze-pump-thaw method (three cycles) and refilled
with N2. The reaction mixture was irradiated under a 400 W Hg-lamp
at 6-10 °C until complete consumption of Rh(tmp)CH3 was confirmed
by TLC analysis (10-11 h). Stock benzene solution of [RhII(tmp)] 1
(2.225 × 10-3 M, 4.000 mL) was prepared. Sublimed 2,2,6,6-
tetramethyl-pyperidin-1-oxyl (TEMPO) 2 (0.2595 g) was added to a
5.00 mL volumetric flask and was made to the volume with dried
benzene. The solution was transferred to a Teflon screw-head stoppered
flask. It was degassed by the freeze-pump-thaw method for three
cycles. Then it was refilled with N2. Stock benzene solution of TEMPO
(0.3321 M) was prepared.
UV-visible time scans for monitoring the reaction of [RhII(tmp)] with
TEMPO were carried out with varied initial concentrations of Rh(tmp).
The stock benzene solution of Rh(tmp) (2.225 × 10-3 M) and dried
benzene were transferred separately to a Teflon-stopped Schlenk UV
cell with gastight syringe under N2. The mixture was stirred 10 min at
70.0 ( 0.2 °C inside the sample compartment. Then the stock benzene
TEMPO (0.3321 M) was transferred to the UV cell with a gastight
syringe under N2. The reaction was monitored at 523 nm for 4-5 half-
lives.
Experimental Section
All chemicals were purchased from commercial suppliers and used
without further purification unless otherwise specified. Dichloromethane
was distilled from calcium hydride under nitrogen. Chloroform was
distilled from calcium chloride under N2. Hexane was distilled from
calcium chloride. Benzene was distilled from sodium under N2. 1,2-
Dichloroethane was distilled from calcium hydride under N2. 2,2,6,6-
Tetramethyl-pyperidin-1-oxyl (TEMPO) was purified by vacuum
sublimation.
Thin layer chromatography was performed on Merck precoated silica
gel 60 F254 plates. Column chromatography was performed on silica
gel (70-230) or neutral aluminum oxide (activity I, 70-230 mesh).
1H NMR spectra were recorded on a Bruker DPX-300 (300 MHz).
Chemical shifts were referenced with the residual solvent protons in
C6D6 (δ 7.15 ppm), CDCl3 (δ 7.24 ppm), or with tetramethylsilane (δ
0.00 ppm) as the internal standard. Chemical shifts (δ) were reported
as part per million (ppm) in δ scale.
Gas chromatography was performed on a HP G1800 GCD system
using a HP5MS column (30 m × 0.25 mm × 0.25 µm).
UV-vis spectra were performed on a Hitachi U-3300 spectropho-
tometer equipped with a Neslab temperature circulator RTE-110 for
temperature control with a mixture of ethylene glycol and water (v/v
) 1/1) used as the circulating liquid. The temperature was measured
by a Fluke thermometer connected to a K-type thermal couple wire
placed in an adjacent dummy UV cell filled with benzene. Spectral
and rate data were analyzed with OriginPro 7.5 software.
Reaction of [RhII(tmp)] with TEMPO. To a Teflon screw-head
stoppered flask, [RhIII(tmp)CH3] 3 (10.0 mg, 0.011 mmol) was dissolved
in C6H6 (4.0 mL) to form a clear orange solution. The reaction mixture
was then degassed by the freeze-pump-thaw method (3 cycles) and
refilled with N2. The reaction mixture was irradiated under a 400 W
Hg-lamp at 6-10 °C until complete [RhIII(tmp)CH3] consumption was
confirmed by TLC analysis (10-11 h). Benzene solution of [RhII(tmp)]
1 (0.0088 mmol, 4.0 mL, 80%) was prepared.6e,10b Then TEMPO 2
(0.055 mmol) was added to the benzene solution of [RhII(tmp)] and
the reaction mixture was heated under N2 in the absence of light. After
cooling down to room temperature, 1 µL of degassed anthracene
solution (0.0749 M) was added as internal standard. The yield of
TEMPOH 4 was measured by GC-MS.
The reaction mixture was purified by chromatography on silica gel
eluting with a solvent mixture of hexane/CH2Cl2 (10:1) to hexane/CH2-
Cl2 (7:1) to give the orange solid of [RhIII(tmp)CH3] with H NMR
1
spectrum identical to that of an authentic sample.
Preparation of TEMPOH 4.21 To TEMPO 2 (500 mg, 3.2 mmol),
a 1:1 mixture of acetone/water (60 mL) and 85% Na2S2O4 (717 mg,
3.5 mmol) were added. The orange solution decolorized instantly upon
swirling. After stirring for 0.5 h at room temperature, acetone was
removed under reduced pressure. The remaining aqueous solution was
extracted with degassed diethylether under N2. Then the ether extract
was dried over K2CO3 under N2. The solution was filtered under N2
using a cannular with the tip wrapped with a filter paper. Then the
solvent was removed under high vacuum. White solid (392 mg, 2.5
UV Titration of Coordination between [RhII(tmp)] (1) and
TEMPO (2). Stock benzene solution of TEMPO (25 µL, 0.3321 M)
was transferred to a Teflon screw-head stoppered flask with a gastight
syringe. Then dried benzene (2.500 mL) was added with a gastight
syringe. It was degassed by the freeze-pump-thaw method for three
cycles. Then it was refilled with N2. Stock benzene solution of TEMPO
(3.288 × 10-3 M) was prepared. The UV-visible wavelength scan
was carried out at 10.0 ( 0.2 °C. The stock benzene solution of [RhII-
(tmp)] (100 µL, 2.225 × 10-3 M) was transferred to a Teflon-stopped
Schlenck UV cell with a gastight syringe under N2. Then dried benzene
(3.900 mL) was added with a gastight syringe under N2. The stock
benzene solution of TEMPO (3.288 × 10-3 M) was then titrated into
the [RhII(tmp)] solution via a gastight microsyringe in 2.0 µL steps up
to a total of 20.0 µL and then in 4.0 µL steps up to a total of 60.0 µL
TEMPO solution added under N2. Finally 50 µL of higher concentration
TEMPO solution (0.3321 M) was added to the sample solution to obtain
the estimated absorbance for the [Rh(tmp)TEMPO)] complex. The UV
spectra were scanned at 600 nm/min between 280 and 700 nm. The
1
mmol, 78%) was obtained. H NMR (C6D6, 300 MHz): δ 1.16 (s, 6
H), 1.31 (m, 2 H), 1.38 (m, 4 H), 3.86 (s, 1 H). GCD (column
program: initial temperature 70 °C, duration 1 min; increment rate 40
°C/min, duration 3 min; increment rate 45 °C/min, duration 2 min;
increment rate 10 °C/min, duration 1 min; final temperature 290 °C,
duration 3 min) tR ) 3.87 min; EIMS: m/z 157.
Preparation of (5,10,15,20-Tetramesitylporphyrinato)hydrido
Rhodium(III) [RhIII(tmp)H] (5).10f,22 A red suspension of Rh(tmp)I
(50 mg, 0.049 mmol) in EtOH (40 mL) and a solution of NaBH4 (7.5
mg, 0.20 mmol) in aq NaOH (0.5 M, 2 mL) were purged with N2
separately for about 15 min. The solution of NaBH4 was added slowly
to the suspension of [RhIII(tmp)I] via a cannular. The reaction mixture
was heated at 55 °C for 3 h and the color changed to deep brown in
9
2060 J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008

Carbon−Carbon Bond Activation of TEMPO
A R T I C L E S
absorbance data at 521 nm for the titration at 15.0, 20.0, 25.0, 30.0 (
0.2 °C, respectively, were measured for evaluation of binding constants.
The binding constants and the number of TEMPO ligand coordinated
to each Rh(tmp) complex were calculated by the equation:23
sample compartment under N2. Then stock benzene solution of TEMPO
(10 µL, 0.3321 M) was transferred to the UV cell with a gastight syringe
under N2. The experimental concentrations were [[RhIII(tmp)H]] ) 5.916
× 10-5 M, [TEMPO] ) 9.108 × 10-4 (15 equiv).
DFT Calculations. The geometry optimizations were carried out
with the Turbomole program25a,b coupled to the PQS Baker optimizer.26
Geometries were fully optimized as minima or transition states at the
bp8627 level using the SV(P) basis set25d on all atoms (small-core
pseudopotential25c,e on rhodium and iridium) and at the b3-lyp level28
using the polarized triple-ú TZVP basis25c,f (small-core pseudopotential25c,e
on Rh). The “real” energy ꢀs of the (multideterminant) open-shell singlet
TS1 (singlet biradical) was estimated from the energy ꢀ0 of the
optimized single-determinant broken symmetry unrestricted b3-lyp
Ae - A
An - Ae
log Kn ) log
m - n log[L]
where n ) no. of TEMPO ligand coordinated to each Rh(tmp), Kn )
binding constant, [L] ) conc. of TEMPO in the UV sample, Ae is the
experimental absorbance measured (volume correction made), Am is
the absorbance of the Rh(tmp) without TEMPO, An is the theoretical
absorbance obtained for the Rh(tmp)-TEMPO complex (obtained
from the absorbance measured for the infinite point with volume
correction).
2
(TZVP) solution (S0 ) 0.454) and the energy ꢀ1 from a separate
2
unrestricted ms ) 1 calculation at the same geometry (S1 ) 2.007),
The measured Ae values were corrected for volume change due to
the addition of ligand solution and the thermal expansion/contraction
by Vt ) V0(1 + ât) and â ) 0.00372 (for benzene),24 where V0 is the
volume before correction, t is the changed temperature and Vt is the
volume after correction.
using the formula29
2
2
S1 ꢀ0 - S0 ꢀ1
ꢀs ≈
2
S12 - S0
Kinetic Study of the Reaction between TEMPO (2) and [RhIII
-
Solvent corrections from single-point cosmo calculations (ꢀ ) 2.28)
were applied for all species.30 The stationary points optimized at the
b3-lyp level (RDFT and UDFT) were characterized by vibrational
analysis (numerical frequencies); ZPE and thermal corrections (entropy
and enthalpy, 298 and 353 K, 1 bar) from these analyses are included.
The thus obtained (free) energies (kcal/mol) are reported in Table 7.
Optimized geometries of II, III, IV, TS1, and TS2, visualized with
the PLATON31 program (rendered with POVRAY), are shown in Figure
9. The orbital and spin density plots shown in Figure 8 were generated
with Molden.32
(tmp)H] (5). The procedure was the same as UV-visible wavelength
scan of the reaction between TEMPO 2 and [RhII(tmp)] 1. Stock
benzene solution of [RhIII(tmp)H] 5 (2.373 × 10-3 M) and TEMPO 2
(0.3321 M) were prepared. First, UV-visible wavelength scan was
carried out. The stock benzene solution of [RhIII(tmp)H] (100 µL, 2.373
× 10-3 M) was transferred to a Teflon-stopped UV cell with a gastight
syringe under N2. Then dried benzene (3.900 mL) was added with a
gastight syringe under N2. The mixture was stirred 10 min at 25.0 (
0.2 °C inside the sample compartment under N2. Then stock benzene
solution of TEMPO (10 µL, 0.3321 M) was transferred to the UV cell
with a gastight syringe under N2. The experimental concentrations were
[[RhIII(tmp)H]] ) 5.916 × 10-5 M, [TEMPO] ) 9.108 × 10-4 M (15
equiv), temperature ) 25.0 ( 0.2 °C. The UV spectra were scanned at
600 nm/min between 280 and 700 nm. The spectra were recorded at 0,
2, 5, 10, 30, 60, 90, and 150 min. From the results of wavelength scan,
the rates of reactions were monitored at 522 nm, where maximum
absorbance changes occurred.
Acknowledgment. We thank the Research Grants Council
of Hong Kong (CUHK 400104) and The Netherlands Organiza-
tion for Scientific Research (NWO-CW) for financial support.
We thank Prof. P. H. M. Budzelaar (Manitoba) and Prof. P.-O.
Norrby (Gothenburg) for helpful discussions about the DFT
calculations.
Second, the UV-visible time scan was carried out. The concentration
of [RhIII(tmp)H] was varied. The stock benzene solution of [RhIII(tmp)H]
(2.373 × 10-3 M) and dried benzene were transferred separately to a
Teflon-stopped UV cell with gastight syringe under N2. The mixture
was stirred 10 min at 15.0 ( 0.2 °C inside the sample compartment.
Then the stock benzene TEMPO (0.3321 M) was transferred to the
UV cell with a gastight syringe under N2. The reaction was monitored
at 522 nm for 4-5 half-lives.
Finally, the reactions were monitored at 522 nm for 4-5 half-lives
at temperatures from 15.0 to 30.0 °C. The stock benzene solution of
[RhIII(tmp)H] (100 µL, 2.373 × 10-3 M) was transferred to a
Teflon-stopped Schlenck UV cell with a gastight syringe under N2.
Then dried benzene (3.900 mL) was added with a gastight syringe under
N2. The mixture was stirred 10 min at a certain temperature inside the
Supporting Information Available: Experimental details,
kinetic data and derivation of rate laws and activation param-
eters, complete ref 25a, (u)b3-lyp optimized geometries of I,
II, III, IV, TS1, TS2, and the species depicted in Scheme 3.
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA078157F
(30) Klamt, A.; Schu¨u¨rmann, G.; J. Chem. Soc., Perkin Trans. 2 1993, 5, 799-
805.
(31) Spek, A. L. PLATON, A Multipurpose Crystallographic Tool; Utrecht
University: Utrecht, The Netherlands, 2003.
(32) Schaftenaar, G.; Noordik, J. H. J. Comput.-Aided Mol. Des. 2000, 14 (2),
123-134.
9
J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008 2061