Hydrogen abstraction from H-donor substrates by the 6-CF 3-benzotriazol-N-oxyl radical (TFNO)

Article abstract of DOI:10.1002/poc.1783

The aminoxyl radical 6-trifluoromethyl-benzotriazol-N-oxyl (TFNO) has been generated from the parent hydroxylamine 6-CF₃-1-hydroxy-benzotriazole (TFBT) by one-electron oxidation with a Ce^IV salt and characterized by spectrophotometry and cyclic voltammetry (CV). Rate constants of H-abstraction (k_H) by TFNO from a number of H-donor benzylic substrates have been determined spectrophotometrically in MeCN solution at 25°C. A radical H-atom transfer (HAT) route of oxidation is substantiated for TFNO by several pieces of evidence. The kinetic data also testify the relevance of stereoelectronic effects upon the HAT reactivity of TFNO.

Full text of DOI:10.1002/poc.1783

Research Article
Received: 4 May 2010,
Revised: 8 July 2010,
Accepted: 15 July 2010,
Published online in Wiley Online Library: 28 September 2010
(wileyonlinelibrary.com) DOI 10.1002/poc.1783
Hydrogen abstraction from H-donor
substrates by the 6-CF3-benzotriazol-N-oxyl
radical (TFNO)
a
a
a
*
Mahelet Aweke Tadesse , Carlo Galli and Patrizia Gentili
The aminoxyl radical 6-trifluoromethyl-benzotriazol-N-oxyl (TFNO) has been generated from the parent hydroxyla-
mine 6-CF3-1-hydroxy-benzotriazole (TFBT) by one-electron oxidation with a CeIV salt and characterized by spectro-
photometry and cyclic voltammetry (CV). Rate constants of H-abstraction (kH) by TFNO from a number of H-donor
benzylic substrates have been determined spectrophotometrically in MeCN solution at 25 -C. A radical
H-atom transfer (HAT) route of oxidation is substantiated for TFNO by several pieces of evidence. The kinetic data
also testify the relevance of stereoelectronic effects upon the HAT reactivity of TFNO. Copyright ß 2010 John Wiley &
Sons, Ltd.
Keywords: aminoxyl radicals; bond energy; hydrogen atom transfer; kinetics; oxidation reactions; stereoelectronic effects
INTRODUCTION
RESULTS AND DISCUSSION
Characterization of the radical TFNO
Removal of hydrogen atom from a substrate is a fundamental
event that allows the subsequent interaction with dioxygen and
therefore leads to an oxidation outcome.[1] Transfer of hydrogen
atom from C—H bonds to free-radical species, and to oxygen-
centred radicals in particular, takes place in many instances,[2]
including not only the degradation of chemicals in the tropo-
sphere,[3] biochemical reactions,[4] synthetic organic transform-
ations,[5] but also the pathogenesis of some diseases.[6] Oxygen-
centred radicals, such as the aminoxyl radicals (R2NOꢀ) from
appropriate hydroxylamine (R2NO—H) precursors, participate in
many industrial procedures,[5,7,8] and promote the environmental
The generation of an aminoxyl radical from the parent hydro-
xylamine (>NO—H) is in principle possible by either H-atom
abstraction or else by one-electron abstraction followed by
deprotonation.[9,10] The resulting >N—Oꢀ species can be either
persistent or short-living depending on the structure, and
characterized by appropriate spectroscopic methods.[9] For example,
PINO was generated from precursor N-hydroxyphthalimide (HPI)
by oxidation with Pb(OAc)4 in AcOH solution, and showed an
electronic spectrum with maximum of absorbance at 380 nm.[14]
It underwent spontaneous decay with a half-life of 7900 s.
Analogously, BTNO was generated from precursor HBT by
oxidation with cerium(IV) ammonium nitrate (CAN) in MeCN
solution; it showed lmax at 474 nm and a spontaneous decay
with half-life of 110 s.[19] Additional characterization of these
aminoxyl radicals was possible by means of laser flash
photolysis, cyclic voltammetry (CV) and electron paramagnetic
resonance (EPR) spectroscopy.[9,10] The latter technique was the
one formerly employed in the kinetic investigation with BTFN
(Fig. 1) in Freon.[20]
benign aerobic oxidation of
a number of substrates by
H-abstraction.[9–11] Phthalimide N-oxyl radical (PINO) provides a
prominent case (Fig. 1), and the synthetic value of PINO-induced
oxidations is recognized and has prompted studies upon the
reactivity features of the aminoxyl radical intermediates.[9,10,12,13]
Kinetic surveys have been already performed with PINO or else
with its X-aryl-substituted cognates,[13–17] and have been recently
complemented by a kinetic investigation with another aminoxyl
radical, that is benzotriazole N-oxyl (BTNO),[18] originating from
precursor 1-hydroxy-benzotriazole (HBT).[19] Another kinetic
study had been reported with bis(trifluoromethyl)aminoxyl
radical,[20] (CF3)2N—Oꢀ (dubbed here BTFN). We now compare
the previous kinetic data with newly acquired data concerning
the H-abstraction reactivity of 6-trifluoromethyl-benzotriazol-
N-oxyl radical (dubbed TFNO) (Fig. 1) towards H-donors.
Mechanistic evidence has been collected with this particular
aminoxyl radical, and a deeper insight into the H-abstraction
reactivity of this class of compounds now becomes possible. In
addition, and by resorting to suitable substrates, we gain a
quantitative appreciation of the stereoelectronic effects that
drive the H-abstraction step by the aminoxyl radicals.
We report here on the first generation of TFNO from the
commercially available 6-CF3-1-hydroxy-benzotriazole (TFBT)
precursor. Following the addition of a stoichiometric amount
`
Correspondence to: C. Galli, Department of Chemistry, Universita ‘La Sapienza’,
*
P.le A. Moro 5, Roma 00185, Italy.
E-mail: carlo.galli@uniroma1.it
a
M. A. Tadesse, C. Galli, P. Gentili
Department of Chemistry and IMC-CNR Sezione Meccanismi di Reazione,
`
Universita ‘La Sapienza’, Roma, Italy
J. Phys. Org. Chem. 2011, 24 529–538
Copyright ß 2010 John Wiley & Sons, Ltd.

M. A. TADESSE, C. GALLI AND P. GENTILI
Figure 1. Aminoxyl radicals (>N—Oꢀ) of interest in this study (full names
are given in the text)
Figure 3. Cyclic voltammetry in MeCN solution (V vs. Ag/AgCl):
[TFBT] ¼ 2 mM, [Py] ¼ 4 mM
of the monoelectronic oxidant CAN (E8 1.3 V vs. NHE)[18] to a
0.8 mM solution of TFBT (curve ꢁ) in MeCN at 25 8C, a new and
broad absorption band develops in the range of 350–600 nm
(curve !), with lmax at 450 nm and e 1516 Mꢂ1 cmꢂ1 (Fig. 2). The
generated species is not stable but decays according to an
exponential curve having half-life of 70 s.
In agreement with the quantitative oxidation of TFBT by the
one-electron oxidant CAN, investigation by CV shows a reversible
scan for a 2 mM solution of TFBT in MeCN in the presence of
pyridine (4 mM) and 0.1 M LiClO4 (Fig. 3), from which the redox
potential (E8) of the >NOꢀ/>NOꢂ couple is obtained as 1.06 V
versus normal hydrogen electrode (NHE). In buffered (pH 6.9)
water solution, the E8 is 1.16 V versus NHE. These data are in
keeping with a reversible one-electron oxidation of the conjugate
base of TFBT to TFNO (Scheme 1).[21,22]
bond of TFBT by EPR equilibration methods,[23] but nevertheless
have attained an evaluation of that BDENO—H as 86 ꢃ 1 kcal/mol
from resorting to a thermochemical cycle.[24,25] By and large, all
the evidence collected on TFNO is in keeping with the formerly
reported characterization of the parent BTNO species from HBT
(Table 1).[18] In particular, the BDENO—H values are almost equal
within the errors (refer to Experimental section).
Hydrogen-abstraction reactivity by TFNO
The spontaneous decay of TFNO is strongly accelerated in the
presence of H-donor substrates bearing C—H bonds of suitable
energy. Accordingly, it becomes possible to investigate the
H-atom transfer (HAT) reactivity (kH) of TFNO from substrates,
such as benzyl alcohols, alkylarenes or allylalkanes (Eqn (1)), by
using UV–Vis spectroscopy, and monitoring the disappearance of
the OD450 of TFNO under the above described conditions in
MeCN solution at 25 8C.
Because of the limited stability of TFNO, we have not been able
to carry out the determination of the energy value of the NO—H
kH
TFNO þ ArCH2OH ꢂ! ArCHðꢀÞOH
-TFBT
O2
ꢂ! productðe:g: ArCHOÞ
(1)
TFNO is generated in the spectrophotometric cuvette by
quickly mixing a 0.5 mM CAN solution with a 1 mM TFBT solution,
both in MeCN at 25 8C. The reading of the initial absorption at
450 nm confirms quantitative generation of TFNO. Fast addition
of a solution of the C—H bearing substrate marks the beginning
of the kinetic experiment; 3–5 initial concentrations of the
substrate are typically investigated in concentration exceeding
Figure 2. UV–Vis spectrum of TFBT (0.8 mM) in MeCN in the absence of
CAN (ꢁ); in the presence of CAN (0.8 mM) (!) 15 ms after mixing
Scheme 1. Electrochemical generation of TFNO
wileyonlinelibrary.com/journal/poc
Copyright ß 2010 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2011, 24 529–538

HYDROGEN ABSTRACTION FROM H-DONOR SUBSTRATES
Table 1. Comparison of the properties of the >NO—H and >NO. species in the benzotriazole family
HBT
TFBT
pKa (NO—H)a
4.6
0.967
1.06
85c
4.0
1.06
1.16
86c
E8, V vs. NHE, in MeCNb
E8, V vs. NHE, in buffered water (pH 6.9)
BDENO—H, in kcal/molc
lmax of >NO. (in MeCN)
Half-life of >NOꢀ(in MeCN)d
474 nm
110 s
450 nm
70 s
[21,22]
a From Ref.
.
b In the presence of pyridine; refer also Ref.
.
[23]
c Calculated from a thermochemical cycle (ꢃ1 kcal/mol), refer Experimental section and Ref.
.
[23]
d Spontaneous decay, k(decay)
.
that of TFNO by 10–100 fold. The time-dependent depletion of
TFNO at OD450 is monitored by stopped-flow spectrophotometry
because the reaction times are in general in the few seconds
range. Under these pseudo first-order conditions, the H-abstraction
rate from the substrate is higher than the spontaneous decay of
TFNO in all cases reported in Table 2. The pseudo first-order rate
constants k0, determined from at least duplicated experiments,
are converted into second-order H-abstraction rate constants
(kH) by determining the slope of a k0 versus [subst]8 linear plot
(Fig. 4).[18] No investigation is possible with either H-donor
substrates, too reactive for our experimental equipment, or
slower-reacting substrates whose ‘apparent’ first-order rate con-
stant matches the k(decay) of TFNO. Typical uncertainty in the
kinetic determinations ranges from 2 to 5%.
It is appropriate to state that, although the kH data in Table 2
cover a range exceeding two powers of ten, the corresponding
range of bond dissociation energy (BDE) values is unfortunately
not wide enough, nor the latter data are known with sufficient
precision, to warrant a significant correlation of reactivity with
bond energy, as we were able to do in the previous investigation
with BTNO.[18] Nevertheless, the kH values of Ph2CH2 and
Ph2CHOH (5.4 and 25 Mꢂ1 sꢂ1) obtained here with TFNO do
scale with the available BDEC—H values (84 and 78 kcal/mol,
Table 2. Second-order rate constants of H-abstraction (kH)
from H-donors (RH) by TFNO at 25 8C in MeCN. Available bond
dissociation energy values (BDEC—H) of RH are given.
BDEC—H
(kcal/mol)b
Substrate, RH
kH (Mꢂ1 sꢂ1 a
)
4-MeO-C6H4CH2OH
3-MeO-C6H4CH2OH
4-Me-C6H4CH2OH
PhCH2OH
4-Cl-C6H4CH2OH
4-CF3-C6H4CH2OH
4-NO2-C6H4CH2OH
(PhCH2)2O
4-MeO-C6H4CH3
4-MeO-C6H4CH2CH3
4-MeO-C6H4CHMe2
PhCH(Me)OH
cyclohexene
54
26
40
23
15
7.6
7.2
22
1.8
6.2
27
23
27
5.4
25
24
0.4
78–82
78–82
78–82
80
83–85
83–85
83–85
n.a.c
89–90
85–87
83–85
n.a.c
82
84
Ph2CH2
Ph2CHOH
fluorene
78
82
n.a.
c
PhCH2CN
a Experimental conditions: [TFNO] 0.5 mM, [RH] 5–25 mM in
Figure 4. Pseudo first-order (k0) rate constants for H-abstraction from
4-Cl-benzyl alcohol by TFNO (0.5 mM) in MeCN at 25 8C. The slope of the
linear plot gives the second-order rate constant kH; the ratio of the initial
concentrations of TFNO and substrate is indicated
MeCN solution. Determinations in triplicate; typical error ꢃ3%.
[18]
[26]
b From Refs
and
.
c Not available.
J. Phys. Org. Chem. 2011, 24 529–538
Copyright ß 2010 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/poc

M. A. TADESSE, C. GALLI AND P. GENTILI
respectively).[26] Likewise, the rate constants of abstraction of a
primary versus secondary versus tertiary benzylic hydrogen atom
from, respectively, MeO-C6H4CH3 versus MeO-C6H4CH2CH3 versus
MeO-C6H4CHMe2 are compatible with the corresponding values
of BDEC—H (ca. 89, 86 and 84 kcal/mol, respectively).[26] Therefore,
a higher HAT reactivity for weaker bonds can be inferred, a result
consistent with analogous findings for PINO and BTNO.[14,16,17]
Summing it up, the reactivity features of TFNO in the HAT
process (Table 2) are akin to those of other aminoxyl radicals,[10]
including the finding of an accelerating effect from electro-
n-donor and the retarding effect from electron-withdrawing
substituents on the benzylic substrate. Significant figures for this
comparison of reactivity among aminoxyl radicals are selected
and given in Table 3, proper allowance being made to the
different solvents employed (no major effects upon the rates are
however expected in these radical processes[14,15,27]).
It is worth to point out that the reactivity of abstraction of
benzylic hydrogens by the present aminoxyl radicals is sub-
stantially lower than that documented for other oxygen-centred
radicals,[10,20,28] such as HOꢀ or t-BuOꢀ, the second-order rate
constants of which being in the range of 106–109 for similar
substrates under similar conditions. Because HOꢀ and t-BuOꢀ form
much stronger O—H bonds upon HAT (119 and 109 kcal/mol,
respectively)[26] than the aminoxyl radicals (>NO—H ca.
80–89kcal/mol),[9,17] a fair convergence of thermodynamic and
reactivity features emerges: the stronger the O—H bond formed by
the abstracting oxygen-radical (ROꢀ) the higher the H-abstraction
reactivity. More on this point, and although the difference in
BDENO—H values among the hydroxylamine precursors of the
aminoxyl radicals of Table 3 is admittedly quite small, some
comments about their different reactivity are anyhow possible. In
fact, there is no doubt that PINO is more reactive than BTFN, in
keeping with the higher energy of the NO—H bond it forms on
H-abstraction (88 vs. 84 kcal/mol).[10] The BDENO—H of TFBT and
HBT are instead intermediate in value between the two above
Scheme 2. Polar effects upon the transition state of a radical HAT
process by an electrophilic aminoxyl radical
Table 3. Comparison of kinetic data (kH, Mꢂ1 sꢂ1, at 25 8C) of H-abstraction among aminoxyl radicals (structures and names in Fig. 1;
n.d. stands for not determined)
PINO,b in AcOH
BTFN,c in Freon
BTNO,d in MeCN
TFNO in MeCN
¼
Substrate, RH
(BDEC—H, in kcal/mol)a
BDENO—H
¼
BDENO—H
¼
BDENO—H
¼
BDENO—H
88.3 kcal/mole
84.0 kcal/mole
85 kcal/molf
86 kcal/molf
PhCH2OH (80)
p-MeO-C6H4CH2OH (ca. 79)
PhCH3 (89)
PhCH2CH3 (85)
PhCHMe2 (84)
Ph2CHOH (78)
Ph2CH2 (84)
11
45
0.62
5.4
27
58
13
59
40
n.d.
n.d.
1.8
7.8
0.26g
0.70g
0.55g
3.2
0.72
2.3
3.8
23
54
1.8g
6.2g
54g
25
5.4
n.d.
24
8.9 10ꢂ3
0.3
0.3
n.d.
0.48
8.8
Ph3CH (81)
Fluorene (82)
n.d.
[26]
a From Ref
.
[14,27]
b From oxidation of HPI with Pb(OAc)4 at 25 8C, Refs
.
c From Ref. [20], at 25 8C.
d From oxidation of HBT with CAN at 25 8C, Ref.
.
[10]
[18]
e Experimental value (ꢃ0.5 kcal/mol), Refs and
[9]
.
f A reasonable confidence limit for our calculated BDENO—H values is ꢃ1 kcal/mol, Ref.
g As p-MeO-derivatives.
.
[23]
wileyonlinelibrary.com/journal/poc
Copyright ß 2010 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2011, 24 529–538

HYDROGEN ABSTRACTION FROM H-DONOR SUBSTRATES
Table 4. Kinetic data of either H- or D-abstraction by TFNO
from a series of benzylic alcohols, in MeCN at 25 8C
Substrate
kH or kD (Mꢂ1 sꢂ1
)
kH/kD
Scheme 3. Competition of ET versus HAT routes with an aminoxyl radical
4-MeOC6H4CH2OH
4-MeOC6H4CD2OH
PhCH2OH
PhCD2OH
4-CF3C6H4CH2OH
4-CF3C6H4CD2OH
54
8.6
23
1.4
7.6
0.44
6.3
16
17
cases and indeed very close together,[10,23] so that it is a little
puzzling that TFNO comes out 3–10 fold more reactive than
BTNO. However, because the >N—Oꢀ radicals are known to be
electrophilic,[9,10,17,29] the presence of the electron-withdrawing
CF3-substituent in TFNO may confer additional polar stabilization
to the transition state of the HAT (cf. Scheme 2). It will assist the
development of a partial negative charge on the abstracting
aminoxyl radical, as mirrored by the favourable effect from the
electron-donor substituents (X) on the benzylic alcohol counter-
part, where a partial positive charge develops (vide also infra).
Therefore, the H-abstraction reactivity of the aminoxyl radicals
appears to be not only dominated by the enthalpic requirements
of the NO—H bond energy in formation, but also by the polar
effects sketched in Scheme 2 that may provide electrostatic
stabilization to the transition state of the radical HAT route.
Hammett correlation
The H-abstraction reactivity of TFNO towards X-substituted
benzyl alcohols (from kH data in Table 2; cf. Eqn (1)) correlates
with the sþ parameter of the X-substituents according to an
Hammett-like equation (log kX/kH ¼ rsþ),[32] and gives a linear
plot with r ¼ ꢂ0.62 (ꢃ0.06) (Fig. 5). This is indeed a small slope,
as expected for a radical process,[18] and compares well with a
r value of ꢂ0.55 for the reaction of BTNO versus the same
substrates in MeCN,[19] as well as with a r of ꢂ0.41 for PINO
whenever generated from HPI by Pb(OAc)4 in AcOH solution.[14]
Mechanistic insight
Recent studies have supported the capability of the aminoxyl
radicals to remove electrons from electron-rich substrates, giving
an electron-transfer (ET) route as opposed to the expected HAT
pathway (Scheme 3).[10,30,31] Inspired by the finding that polar
effects (Scheme 2) are relevant for TFNO and make it somewhat
more reactive than BTNO, i.e. the parent aminoxyl radical for
which a clean HAT route had been substantiated,[18,19] we have
deemed that a full mechanistic characterization of the reaction
of TFNO towards H-donors was required in the event that the
competing ET route could take over.
Scheme 4. Products study for reaction of TFNO with probe substrate 1:
Ca—Cb (above) versus Ca—H (below) bond cleavage
Figure 6. Stereoelectronic interaction between the lone-pair of an
amine and an adjacent C—H bond. Reprinted with permission from
Griller et al. [39]. Copyright (1981) American Chemical Society
Figure 5. Hammett correlation for reaction of TFNO with X-substituted
benzyl alcohols in MeCN solution at 25 8C
J. Phys. Org. Chem. 2011, 24 529–538
Copyright ß 2010 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/poc

M. A. TADESSE, C. GALLI AND P. GENTILI
Table 5. Rate constants of H-abstraction by TFNO in MeCN solution at 25 8Ca
Entry
1
SubstrateA
SubstrateB
krel (subA/subB)b
Ph2CHOH
kH ¼ 25 Mꢂ1 sꢂ1
Ph2CH2
kH ¼ 5.4 Mꢂ1 sꢂ1
9
BDEC—H ¼ 78 kcal/mol
BDEC—H ¼ 84 kcal/mol
2
16
kH ¼ 190 Mꢂ1 sꢂ1
BDEC—H ¼ 73 kcal/mol
(PhCH2)2O
kH ¼ 24 Mꢂ1 sꢂ1
BDEC—H ¼ 82 kcal/mol
PhCH2CH2Ph
3
4
5
25
kH ¼ 22 Mꢂ1 sꢂ1
kH ¼ 0.88 Mꢂ1 sꢂ1
BDEC—H ¼ 87 kcal/mol
PhCH2OC(O)Ph
BDEC—H ¼ ca. 82 kcal/mol
PhCH2OH
>80
>100
kH ¼ 23 Mꢂ1 sꢂ1
kH ¼ unreactivec
BDEC—H ¼ 80 kcal/mol
BDEC—H ¼ n.a.d
kH ¼ 62 Mꢂ1 sꢂ1
BDEC—H ¼ n.a.d
kH ¼ unreactivec
BDEC—H ¼ n.a.d
6
>50
kH ¼ 6.8 Mꢂ1 sꢂ1
BDEC—H (allylic) ¼ ca. 82 kcal/mol
PhCH2NHC(O)Ph
kH ¼ unreactivec
BDEC—H (allylic) ¼ ca. 90 kcal/mol
PhCH2OC(O)Ph
7
8
>15
kH ¼ 2.2 Mꢂ1 sꢂ1
kH ¼ unreactivec
BDEC—H ¼ n.a.d
BDEC—H ¼ n.a.d
Ph2CH2
4.4
kH ¼ 24 Mꢂ1 sꢂ1
kH ¼ 5.4 Mꢂ1 sꢂ1
BDEC—H ¼ 84 kcal/mol
PhCH2CH2Ph
BDEC—H ¼ 82 kcal/mol
9
140
kH ¼ 124 Mꢂ1 sꢂ1
kH ¼ 0.88 Mꢂ1 sꢂ1
BDEC—H ¼ 84 kcal/mol
BDEC—H ¼ 87 kcal/mol
(Continues)
wileyonlinelibrary.com/journal/poc
Copyright ß 2010 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2011, 24 529–538

HYDROGEN ABSTRACTION FROM H-DONOR SUBSTRATES
Table 5. (Continued)
Entry
10
SubstrateA
SubstrateB
Ph2CHOH
krel (subA/subB)b
7.6
kH ¼ 190 Mꢂ1 sꢂ1
kH ¼ 25 Mꢂ1 sꢂ1
BDEC—H ¼ 73 kcal/mol
BDEC—H ¼ 78 kcal/mol
[26]
a BDE data from Ref.
.
b The ratio has been normalized for the number of equivalent H-atoms (i.e. ‘intrinsic reactivity’ per H-atom).
c Rate value comparable to the spontaneous decay of TFNO (i.e. <0.3 Mꢂ1 sꢂ1).
d Not available. A rate value too high for our spectrophotometric mixing device (8 ms of dead-time; i.e. kH > 80 Mꢂ1 sꢂ1).
Additional r values (in the range of ꢂ0.5 to ꢂ0.7) are reported for
X-substituted PINOs generated from X-substituted HPIs in MeCN
solution.[17] In all these cases better Hammett correlations are
obtained versus the sþ parameter of the X-substituents than with
the simple s parameter, in line with the electrophilicity of the
aminoxyl radicals.[17,29] Thus, an uniform pattern of selectivity
among these structurally similar >N—Oꢀ species does emerge.
Small r values such as these are more compatible with a
single-step HAT route rather than with the formal alternative of a
concerted proton–electron transfer route (CPET; cf. last structure
in Scheme 2).[30,33] For example, in the recently reported oxidation
of 4-X-substituted N,N-dimethylanilines by the very BTNO, a much
larger r value of ꢂ3.8 has been obtained, and judged compatible
with a two step electron transfer–proton transfer mechanism.[34]
Upon oxidation of 1 with TFNO in MeCN (cf. Experimental
section), a 4% yield of ketone 3 is obtained with no traces of the
aldehyde 2; the rest of mass balance being the unreacted 1.
Under similar conditions, BTNO gave a strictly comparable
result.[19,35] Thus, the HAT pathway is utterly confirmed for both
these aminoxyl radicals, and any competition by an ET route (cf.
Scheme 3) can be discarded.
Stereoelectronic effects
In line with the concept of the anomeric effect,[37,38] Griller
et al.[39] have convincingly demonstrated that conjugation from
the lone-pair of a heteroatom weakens an adjacent C—H bond
and stabilizes the intervening C-radical generated by H-abstraction.
The contribution from this weakening effect upon the energy of
the adjacent C—H bond is the greater the more co-linear are
C—H bond and lone-pair.[39,40] Figure 6 shows the case of an
amine where the effect reportedly accounts for a 7–9 kcal/mol
weakening of the adjacent C—H bond. A computational study
similarly attains a 5 kcal/mol weakening of the a C—H bond for
conjugation from an oxygen lone-pair.[41]
For torsion angles (u) wider than 308, the relevance of the
effect fades away drastically.[39] Our finding of an appreciable
rate increase when passing from Ph2CH2 to Ph2CHOH (Table 2) is
likely due to the profitable stereoelectronic interaction from
the oxygen lone-pair of the alcohol with the scissile a C—H bond,
an interaction that clearly lacks in the hydrocarbon counterpart,
thereby confirming the analogous finding in the kinetic study
with BTNO.[18]
Kinetic isotope effect determination
The kinetic survey with TFNO has been extended towards a few
deuteriated substrates, pertinent data being collected in Table 4.
The large values of the intermolecular kH/kD ratio, ranging from
6 to 17, support a single-step rate-determining H-abstraction
step, and confirm analogous results (in the range of 6–27)
obtained for PINO and BTNO.[9,14,15,17,18,27] The contribution from
tunnelling is clear, as also documented elsewhere.[14,17]
Products study for reaction of TFNO with a probe substrate
As described previously,[35] the test substrate 1 presents the
distinct advantage of giving rise to two different end-products
depending on the operating mechanism of oxidation. In fact,
under genuine ET conditions, 1 gives a transient 1ꢀþ inter-
mediate that cleaves the Ca—Cb bond of the side-chain, and
affords the corresponding aldehyde (2) plus tert-butyl radical
(Scheme 4).[36]
The cleavage is driven by steric and stereoelectronic factors
that make the loss of a tertiary radical easy.[36] Conversely, under
bona fide radical HAT conditions, 1 undergoes cleavage of the
Ca—H benzylic bond, and produces the corresponding ketone
3 exclusively.[35,36] This clear-cut behaviour makes 1 a significant
probe, capable of differentiating the mechanism of an oxidation
reaction from product analysis.
By exploiting the higher reactivity of TFNO than BTNO in HAT
reactions, we have been able to extend the previous kinetic
investigation on the stereoelectronic effects[42,43] to additional
and significant substrates. Table 5 gives the results for H-abstraction
by TFNO, the investigated substrates being organized pairwise
with their experimental kH and BDEC—H values (whenever
available).[26]
A few comments are in order:
(1) Abstraction of hydrogen from a C—H bond adjacent to an
oxygen atom is always faster than from the parent hydro-
carbon (entries 1–3),[44] as mirrored by the corresponding
BDEC—H data,[26] owing to the favourable stereoelectronic
J. Phys. Org. Chem. 2011, 24 529–538
Copyright ß 2010 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/poc

M. A. TADESSE, C. GALLI AND P. GENTILI
alcohols affords a small value for r (i.e. ꢂ0.62), compatible with a
single-step radical H-atom abstraction from the benzylic position,
rather than with a CPET route (cf. Scheme 3),[31] that it is expected
to provide larger r values (as ꢂ3.8,[33,34] or even larger[46]). A
product study for reaction of TFNO with a probe substrate
confirms a HAT route of oxidation. The primary kinetic isotope
effect supports the H-abstraction as rate determining. The HAT
reactivity has been compared within a series of four aminoxyl
radicals and an uniform pattern of selectivity (Table 3) does
emerge. In particular, the H-abstraction from H-donor substrates
appears to be the faster the stronger the O—H bond (i.e.
BDENO—H) that the >N—Oꢀ species gives rise to in the HAT step.
In this respect, the slightly higher reactivity of TFNO than the
parent BTNO species is at odds with the comparable values of
their BDENO—H and ascribed to the electron-withdrawing
CF3-substituent in TFNO that confers additional polar stabilization
to the transition state of the radical HAT (Scheme 2). Finally, and
by exploiting the higher reactivity of TFNO than BTNO in HAT
reactions, we have been able to substantiate a wider number of
exemplary cases of stereoelectronic effects on the HAT reactivity.
In the inspected cases, the conjugation of a C—H bond with
either the lone-pair of an adjacent heteroatom or a p-system
weakens the C—H bond and fosters its cleavage by the aminoxyl
radical. The efficiency of the interaction is enhanced whenever
a proper co-linear conformation is attained (Fig. 6), and this is
particularly well met in rigid cyclic substrates (Fig. 8) where the
loss of conformational freedom is not severe.
Figure 7. Reduced availability of the oxygen lone-pair owing to conju-
gation with an adjacent group
Figure 8. Enforced co-linearity of
a benzydrylic radical with the
p-orbitals of the adjacent aromatic rings, as due to the planar
5-membered ring interconnecting the two p-systems (the case of entry
8 in Table 5 is shown)
interaction between oxygen lone-pair and C—H bond (cf.
Fig. 6);
(2) whenever the availability of the adjacent oxygen lone-pair is
lessened by resonance with an electron withdrawing group,
such as a carbonyl (Fig. 7), the stereoelectronic assistance in
Fig. 6 is reduced (entries 4–6);[45]
EXPERIMENTAL
(3) the stereoelectronic effect from the lone-pair of nitrogen is
somewhat stronger than from oxygen (entry 7).[39–41]
(4) clearly, the stereoelectronic effect is particularly appreciable
here because the >N—Oꢀ species is an electrophilic
radical,[9,10,12,13,17,29] and, therefore, susceptible to both bond-
polarization effects (cf. Fig. 6) and polar effects (Scheme 2)
upon reactivity.
General
Commercial TFBT (Aldrich) was used as received; all the sub-
strates were purchases and used without further purification, if
not already available in the laboratory.[18] In particular, substrate 1
and products 2 and 3 were available from a previous study.[35]
Cerium(IV) ammonium nitrate (CAN; Erba RPE) was oven-dried
before use.[18] Reagent grade acetonitrile (Erba RPE) was used as
the solvent. Bis a-deuteriated benzyl alcohols (cf. Table 4) were
already available in the laboratory or synthesized as described
previously[35,47] by reduction of the corresponding carboxylic
acids with LiAlD4 in dry THF. Full 1H-NMR (samples concentration:
65 mM) and MS characterization of the bis a-deuteriated benzyl
alcohols were as follows.
Besides the stereoelectronic effect pertaining to the lone-pair
of an adjacent heteroatom, H-abstraction from a benzylic (or
allylic) C—H bond is faster whenever cyclic systems are
considered, as opposed to open-chain counterparts.[42] This is
confirmed when comparing the values of BDEC—H for the cyclic/
open pairs in entries 8–10, and supports previous kinetic findings
in HAT processes by BTNO or PINO.[10,42] Clearly, the homolytic
cleavage of a benzylic (or allylic) C—H bond is assisted by
p-conjugation with an aromatic (or vinylic) moiety that stabilizes
the intervening carbon-radical. Such assistance is fostered in a
cyclic compound where co-linearity between the C—H
bond being cleaved and the adjacent p-system is enforced by
steric restrictions from the ring system (Fig. 8), whereas in the
corresponding open-chain compound the appropriate co-linearity
would be somewhat disrupted owing to conformational free-
dom.[42,43]
(1) [1,1-2H1]-benzyl alcohol
(a) dH (200 MHz; CDCl3) 1.9 (bs, 1H, OH), 7.3 (m, 5H, ArH);
(b) MS (m/z): 110 (Mþ), 93, 77, 51.
(2)
[1,1-2H1]-4-(trifluoromethyl)benzyl alcohol
(a) dH (200 MHz; CDCl3) 1.9 (bs, 1H, OH), 7.63–7.45 (AA‘XX’, 4H,
ArH);
(b) MS (m/z): 178 (Mþ), 160, 109, 81, 77, 51.
Instrumentation
The kinetic study was carried out with a Hi-Tech SFA-12 stopped
flow instrument, interfaced to a HP 8453 diode array spectro-
photometers, and having a thermostated cell holder. A thermo-
couple was employed for reading the temperature in the quartz
cuvette. A VARIAN 3400/Star gas chromatograph, fitted with a
CONCLUSIONS
Rate constants of H-abstraction (kH) by TFNO from H-donor
substrates have been determined spectroscopically. Hammett
correlation of the reactivity of TFNO versus X-substituted benzyl
wileyonlinelibrary.com/journal/poc
Copyright ß 2010 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2011, 24 529–538

HYDROGEN ABSTRACTION FROM H-DONOR SUBSTRATES
20 m ꢄ 0.25 mm methyl silicone gum capillary column, was
employed in the GC analyses. GC-MS analyses were performed on
a HP 5892 GC, equipped with a 12 m ꢄ 0.2 mm methyl silicone
gum capillary column and coupled to a HP 5972 MSD instrument
operating at 70 eV.
Kinetic procedure
The TFNO species was generated in the cuvette of the stopped-
flow by adding a 0.5 mM solution of CAN to a 1.0 mM solution of
TFBT, both in MeCN. A broad absorption band developed almost
immediately (8 ms) in the 350–600 nm region (lmax at 450 nm, e
1516 dm3 molꢂ1 cmꢂ1). The OD450 reading was not stable, but
decayed with a half-life of 70 s. The rate of decay was unaffected
by the use of de-aerated MeCN. Rate constants of H-abstraction
from H-donor substrates were determined at 25 8C in MeCN, by
following the decrease of OD450 of TFNO. The initial concen-
tration of the substrate was in excess with respect to the TFNO
species, and in the range of 5–25 mM, so to enable a pseudo
and equation:[9,24,24]
ꢀ
ꢁ
ꢀ
ꢁ
R2NOꢀ
Hþ
BDENOꢂH ¼ pKaðR2NOHÞ þ Eꢅ
þ Eꢅ2
(2)
1
-
R2NO
Hꢀ
This calculation has been calibrated in the case of HPI, obtaining
a BDENO—H value of 88 kcal/mol versus the experimental value of
88.3 kcal/mol.[9] This suggested us that a reasonable confidence
limit for the calculated values is ꢃ1 kcal/mol (in Tables 1 and 3).
first-order treatment of the kinetic data. Plots of (ODt-OD ) versus
1
time were well fitted by a first-order exponential over more than
three half-lives, and the rate constant (k0) reckoned. From a plot of
three-to five k0 versus [subst]8 data pairs, the second-order rate
constant of H-abstraction (kH) was obtained. The intermolecular
kinetic isotope effect for bis a-deuteriated-X-substituted benzyl
alcohols (Table 4) was calculated from the ratio of the kH and kD
rate constants with TFNO.
Acknowledgements
The Authors thank the Italian MIUR for financial support (COFIN
and FIRB projects).
Products study with probe substrate (1)
REFERENCES
As described previously,[18,23,35] 60 mmol of substrate 1, 20 mmol
of TFBT and 20 mmol of CAN were let to react at room
temperature in 3 ml MeCN. After a 5 h reaction time and
conventional workup with ethyl acetate, GC-MS analysis enabled
us to detect the formation of ketone 3 and the absence of
aldehyde 2; no other products besides unreacted 1 were
detected. Quantitative determination of the yield of 3 (i.e. 4%)
was carried out by GC with the internal standard method.
[1] J. A. Labinger, J. E. Bercaw, Nature 2002, 417, 507–514.
´
[2] S. Olivella, J. M. Anglada, A. Sole, J. M. Bofill, Chem. Eur. J. 2004, 10,
3404–3410 (and references therein).
[3] J. Warnatz, in Combustion Chemistry (Ed.: W. C. Gardiner), Springer,
New York, 1984.
[4] R. B. Silverman, in The Organic Chemistry of Enzyme-Catalyzed Reac-
tions, Academic Press, San Diego, 2000.
[5] I. W. C. E. Arends, R. A. Sheldon, in Modern Oxidation Methods, Chap. 4
(Ed.: J.-E. Ba¨ckvall), Wiley-VCH, Weinheim, 2004.
[6] E. J. Cadenas, K. J. A. Davies, Free Radic. Biol. Med. 2000, 29, 222–230.
[7] Y. Ishii, S. Sakaguchi, T. Iwahama, Adv. Synth. Catal. 2001, 343,
393–427.
[8] S. Riva, Trends Biotechnol. 2006, 24, 219–226.
[9] C. Galli, Nitroxyl radicals. in The Chemistry of Hydroxylamines, Oximes
and Hydroxamic Acids – Part 2, Chap. 15 (Eds.: Z. Rappoport, J. F.
Liebman), Wiley, Chichester, 2009.
[10] C. Galli, P. Gentili, O. Lanzalunga, Angew. Chem. Int. Ed. 2008, 47,
4790–4796.
[11] Y. Ishii, S. Sakaguchi, in Modern Oxidation Methods, Chap. 5 (Ed.: J.-E.
Ba¨ckvall), Wiley-VCH, Weinheim, 2004.
[12] R. Amorati, M. Lucarini, V. Mugnaini, G. F. Pedulli, F. Minisci, F.
Recupero, F. Fontana, P. Astolfi, L. Greci, J. Org. Chem. 2003, 68,
1747–1754.
[13] F. Recupero, C. Punta, Chem. Rev. 2007, 107, 3800–3842.
[14] N. Koshino, B. Saha, J. H. Espenson, J. Org. Chem. 2003, 68,
9364–9370.
Electrochemical determinations
CV at the steady disc electrode (glassy-carbon disc, 1.5 mm in
diameter)[48] was carried out with TFBT (2 mM) at 25 8C in MeCN in
the presence of pyridine (4 mM) and 0.1 M LiClO4 (Fig. 3). A
three-electrode circuit was used with a potentiostat with a
positive feedback ohmic-drop compensation and a MPLI Hard-
ware (Vernier Software & Technology) controlled by a program
written in Cþþ language for Windows 95/98.[49] A platinum wire
(1 cm2) and a saturated calomel electrode (SCE) as reference were
used; the E8 value is obtained as the mean of the cathodic and
anodic peak values of the reversible scan (Fig. 3),[50] and referred
to the normal hydrogen electrode (NHE), NHE ¼ SCE þ 0.241 V.[51]
Scan rates in the range of 10 mV/s to 10 V/s were investigated.
[15] Y. Cai, N. Koshino, B. Saha, J. H. Espenson, J. Org. Chem. 2005, 70,
238–243.
[16] F. Minisci, F. Recupero, A. Cecchetto, C. Gambarotti, C. Punta, R. Faletti,
R. Paganelli, G. F. Pedulli, Eur. J. Org. Chem. 2004, 109–119.
[17] C. Annunziatini, M. F. Gerini, O. Lanzalunga, M. Lucarini, J. Org. Chem.
2004, 69, 3431–3438.
BDE determination
[18] P. Brandi, C. Galli, P. Gentili, J. Org. Chem. 2005, 70, 9521–9528.
[19] C. Galli, P. Gentili, O. Lanzalunga, M. Lucarini, G. F. Pedulli, Chem.
Commun. 2004, 2356–2357.
The BDENO—H of TFBT has been obtained from the following
thermochemical cycle (pertinent data in Table 1, to be converted
into kcal/mol), the value of E28 being 54 kcal/mol:[23–25]
[20] T. Doba, K. U. Ingold, J. Am. Chem. Soc. 1984, 106, 3958–3963.
J. Phys. Org. Chem. 2011, 24 529–538
Copyright ß 2010 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/poc

M. A. TADESSE, C. GALLI AND P. GENTILI
[21] A pKa value of 4.0 is given for TFBT: I. Koppel, J. Koppel, I. Leito, V. Pihl,
L. Grehn, U. Ragnarsson, J. Chem. Res. (S) 1993, 446–447.
[22] See also: R. Subiro`s-Funosas, R. Prohens, R. Barbas, A. El-Faham,
F. Albericio, Chem. Eur. J. 2009, 15, 9394–9403.
[23] P. Astolfi, P. Brandi, C. Galli, P. Gentili, M. F. Gerini, L. Greci,
O. Lanzalunga, New J. Chem. 2005, 29, 1308–1317.
[36] E. Baciocchi, S. Belvedere, M. Bietti, O. Lanzalunga, Eur. J. Org. Chem.
1998, 299–302.
[37] P. Deslongchamps, Tetrahedron 1975, 31, 2463–2490.
[38] A. J. Kirby, in The Anomeric Effect and Related Stereoelectronic Effects at
Oxygen, Springer, New York, 1983.
[39] D. Griller, J. A. Howard, P. R. Marriott, J. C. Scaiano, J. Am. Chem. Soc.
1981, 103, 619–623.
´
[24] J. Lind, G. Merenyi, J. Phys. Chem. A 2006, 110, 192–197.
[25] and based on the seminal paper by: D. D. M. Wayner, V. D. Parker, Acc.
Chem. Res. 1993, 26, 287–294.
[26] Y.-R. Luo, in Comprehensive Handbook of Chemical Bond Energies, CRC
Press, Boca Raton, Florida, USA, 2007.
[27] N. Koshino, Y. Cai, J. H. Espenson, J. Phys. Chem. A 2003, 107,
4262–4267.
[28] Y. Sawaki, in: The Chemistry of Hydroxyl, Ether and Peroxide Groups,
Wiley-Interscience, New York, 1993.
[29] F. Minisci, C. Punta, F. Recupero, F. Fontana, G. F. Pedulli, J. Org. Chem.
2002, 67, 2671–2676.
[30] E. Baciocchi, O. Lanzalunga, M. F. Gerini, J. Org. Chem., 2004, 69,
8963–8966.
[31] E. Baciocchi, M. Bietti, M. Di Fusco, O. Lanzalunga, D. Raponi, J. Org.
Chem. 2009, 74, 5576–5583.
[32] T. H. Lowry, K. S. Richardson, in Mechanism and Theory in Organic
Chemistry, 2nd edn, Harper & Row, New York, 1981.
[33] J. M. Mayer, D. A. Hrovat, J. L. Thomas, W. T. Borden, J. Am. Chem. Soc.
2002, 124, 11142–11147.
[40] V. Malatesta, K. U. Ingold, J. Am. Chem. Soc. 1981, 103, 609–614.
[41] F. Bernardi, H. B. Schlegel, A. J. Wolfe, Mol. Struct. 1976, 35,
149–153.
[42] P. Brandi, C. Galli, P. Gentili, J. Phys. Org. Chem. 2006, 19,
552–554.
`
[43] A. Coniglio, C. Galli, P. Gentili, R. Vadala, Org. Biomol. Chem. 2009, 7,
155–160.
[44] D. P. Curran, W. Shen, J. Am. Chem. Soc. 1993, 115, 6051–6059.
[45] See also: A. Gross, L. Fensterbank, S. Bogen, R. Thouvenot, M.
Malacria, Tetrahedron 1997, 53, 13797–13810. (and references
therein).
[46] F. d’Acunzo, P. Baiocco, M. Fabbrini, C. Galli, P. Gentili, New J. Chem.
2002, 26, 1791–1794. (and references therein).
[47] E. Baciocchi, M. Fabbrini, O. Lanzalunga, L. Manduchi, G. Pochetti, Eur.
J. Biochem. 2001, 268, 665–672.
[48] F. d’Acunzo, C. Galli, Eur. J. Biochem. 2003, 270, 3634–3640.
[49] Thanks are due to Dr L. Mottier for writing and implementing the
electrochemical software for us.
[34] E. Baciocchi, M. Bietti, O. Lanzalunga, A. Lapi, D. Raponi, J. Org. Chem.
2010, 75, 1378–1385.
[50] H. Lund, M. M. Baizer, in Organic Electrochemistry, 3rd edn, Marcel
Dekker, New York, 1991.
[35] P. Baiocco, A. M. Barreca, M. Fabbrini, C. Galli, P. Gentili, Org. Biomol.
Chem. 2003, 1, 191–197.
[51] D. J. G. Ives, G. J. Janz, in Reference Electrodes, Academic Press, New
York, 1961.
wileyonlinelibrary.com/journal/poc
Copyright ß 2010 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2011, 24 529–538