Characterization, via ESR Spectroscopy, of Radical Intermediates in the Photooxidation of Arylcarbinols by Ceric Ammonium Nitrate

Article abstract of DOI:10.1021/jo991856t

The photooxidation by ceric ammonium nitrate (CAN) of several aryl and naphthylcarbinols has been studied by means of ESR spectroscopy. For all the investigated arylcarbinols, but not for the naphthyl derivatives, it has been possible to detect radical intermediates deriving from the parent alkoxyl radicals. In particular, in the photooxidation of 1,1-diphenylethanol, a bridged-radical intermediate has been detected. The assignment has been validated through experiments with two different labeled compounds: the 1,1-[2′, 3′, 4′, 5′, 6′, 2″, 3″, 4″, 5″, 6″-²H₁₀]diphenylethanol and the 1,1-diphenyl[2, 2, 2-²H₃]ethanol. A similar bridged radical has been found to be formed in the photooxidation of triphenylmethanol, while, for the 1,1-diphenylpropanol, the only detectable species has been the ethyl radical deriving from a competitive β-scission process. Finally, for the 2-phenylpropan-2-ol (cumyl alcohol), two radical species have been identified: the methyl, deriving from the β-scission process, and the cyanomethylene, deriving from H-abstraction of the cumyloxyl radical from the solvent. A kinetic study on the competition of the two processes has also been conducted and the parameters of the Arrhenius equation for the latter process have been estimated.

Full text of DOI:10.1021/jo991856t

2748
J . Org. Chem. 2000, 65, 2748-2754
Ch a r a cter iza tion , via ESR Sp ectr oscop y, of Ra d ica l In ter m ed ia tes
in th e P h otooxid a tion of Ar ylca r bin ols by Cer ic Am m on iu m
Nitr a te
Loris Grossi* and Samantha Strazzari1
Dipartimento di Chimica Organica “A. Mangini”, Universita` di Bologna, Viale Risorgimento 4,
I-40136 Bologna, Italy
Received December 2, 1999
The photooxidation by ceric ammonium nitrate (CAN) of several aryl and naphthylcarbinols has
been studied by means of ESR spectroscopy. For all the investigated arylcarbinols, but not for the
naphthyl derivatives, it has been possible to detect radical intermediates deriving from the parent
alkoxyl radicals. In particular, in the photooxidation of 1,1-diphenylethanol, a bridged-radical
intermediate has been detected. The assignment has been validated through experiments with
two different labeled compounds: the 1,1-[2′, 3′, 4′, 5′, 6′, 2′′, 3′′, 4′′, 5′′, 6′′-2H10]diphenylethanol
and the 1,1-diphenyl[2, 2, 2-2H3]ethanol. A similar bridged radical has been found to be formed in
the photooxidation of triphenylmethanol, while, for the 1,1-diphenylpropanol, the only detectable
species has been the ethyl radical deriving from a competitive â-scission process. Finally, for the
2-phenylpropan-2-ol (cumyl alcohol), two radical species have been identified: the methyl, deriving
from the â-scission process, and the cyanomethylene, deriving from H-abstraction of the cumyloxyl
radical from the solvent. A kinetic study on the competition of the two processes has also been
conducted and the parameters of the Arrhenius equation for the latter process have been estimated.
Sch em e 1
In tr od u ction
The 1,2 migration of an aryl group in free radicals,
known as the neophyl rearrangement,2 has been exten-
sively studied for 2-arylethyl radicals, and it is thought
to occur via a spiro[2.5]octadienyl radical intermediate
or transition state,3 although direct detection for this
type of intermediate has often failed. However, in at least
one case4 the intermediacy of a more stable spiro[cyclo-
propaneanthracen]-9-yl radical has been proved by ESR
spectroscopy.
Similarly, the neophyl-like rearrangement of alkoxyl
radicals, carrying at least two aryl substituents, is
believed to proceed through a bridged-radical interme-
diate,5,6 as shown in Scheme 1.
The direct detection, by time-resolved absorption spec-
troscopy, of the bridged intermediate for the 1,2 aryl
migration of the 1,1-diphenylethoxyl radical was reported
by Schuster,5 but five years later Scaiano6 demonstrated
that the absorption band which Schuster attributed to
the spiro[oxiranecyclohexadien]yl intermediate could
instead be assigned to the starting alkoxyl radical.7
Nevertheless, a series of AM1-UHF calculations which
were also conducted6 indicated that the bridged-radical
is expected to be a “true” intermediate of this process and
not only a transition state.
We decided to study the mechanism of the neophyl-
like rearrangement of the 1,1-diphenylethoxyl radical by
ESR spectroscopy. Since the formation of the bridged
intermediate could be thought to occur via the 1,3-exo
cyclization of the alkoxyl moiety onto one double bond of
the aromatic system, we chose to apply the same spin
trapping technique which we had used for the detection
of the oxiranylcarbinyl radicals formed by cyclization of
the corresponding allyloxyl radicals.8 But, when a solu-
tion of 1,1-diphenylethanol (1) and t-BuONO was photo-
lyzed directly in the cavity of the ESR spectrometer, no
signals were obtained, except for a weak one which was
assigned to the adduct to t-BuONO of the cyanomethyl
radical, deriving from H-abstraction of the t-BuO. from
the solvent. The lack of detection of any radical species
related to 1,1-diphenylethoxyl radical suggested that this
method could not be exploited to achieve successfully the
precursor alkoxyl radical, most likely due to the scarce
efficiency of the exchange reaction between t-BuONO and
aryl alcohols.8,9
(1) In partial fulfillment of the requirements for a Ph.D. degree in
Chemical Sciences, University of Bologna.
(2) Urry, W. H.; Kharasch, M. S. J . Am. Chem. Soc. 1944, 66, 1438.
(3) Beckwith, A. L. J .; Ingold, K. U. In Rearrangements in Ground
and Excited States; de Mayo, P., Ed.; Academic Press: New York, 1980;
pp 161-310.
Thus, a different procedure was considered in order to
generate the alkoxyl radical.
Previous studies10 conducted on the photoinduced
oxidation of alcohols by CAN had demonstrated that,
(4) Leardini, R.; Nanni, D.; Pedulli, G. F.; Tundo, A.; Zanardi, G.;
Foresti, E.; Palmieri, P. J . Am. Chem. Soc. 1989, 111, 7723.
(5) Falvey, D. E.; Khambatta, B. S.; Schuster, G. B. J . Phys. Chem.
1990, 94, 1056.
(6) Banks, J . T.; Scaiano, J . C. J . Phys. Chem. 1995, 99, 3527.
(7) (a) Avila, D. V.; Lusztyk, J .; Ingold, K. U. J . Am. Chem. Soc.
1992, 114, 6576. (b) Avila, D. V.; Ingold, K. U.; Di Nardo, A. A.;
Zerbetto, F.; Zgierski, M. Z.; Lusztyk, J . J . Am. Chem. Soc. 1995, 117,
2711.
(8) The allyloxyl radicals were obtained by photolysis of the corre-
sponding nitrites, generated directly via an exchange reaction between
the appropriate allylic alcohol and tert-butyl nitrite, which also acted
as the spin trap: (a) Grossi, L.; Strazzari, S. J . Chem. Soc., Chem.
Commun. 1997, 917. (b) Grossi, L.; Strazzari, S.; Gilbert, B. C.;
Whitwood, A. C. J . Org. Chem. 1998, 63, 8366.
(9) Doyle, M. P.; Terpstra, J . W.; Pickering, R. A.; LePoire, D. M. J .
Org. Chem. 1983, 48, 3379.
10.1021/jo991856t CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/12/2000

Photooxidation of Arylcarbinols
J . Org. Chem., Vol. 65, No. 9, 2000 2749
Ta ble 1. ESR Hyp er fin e Sp littin g Con sta n ts (a /m T) of
th e Detected Ra d ica lsa
F igu r e 1. (a) ESR spectrum obtained in the photolysis of an
acetonitrile solution of 1, 20 mM, and CAN, 0.5 wt %, at 230
K, showing both radicals 2 and 3*. Spectrometer setting:
microwave power, 7 mW; modulation amplitude, 0.1 mT; time
constant, 2 s; gain, 2.5 × 104; scan range, 10 mT; scan time,
30 min. (b) Simulation of the spectrum of radical 2.
during this reaction, the corresponding alkoxyl radicals
are formed. As a result, we considered it convenient to
produce the alkoxyl radical of interest by photolysis of
an acetonitrile solution of 1 and CAN directly in the ESR
cavity.
Using the same procedure, experiments were also
conducted on other aryl and naphthylcarbinols. The
present study sets out the results obtained in the
interaction of the analyzed substrates with this single
electron oxidant.
Resu lts a n d Discu ssion
(a ) P h otooxid a tion of 1,1-Dip h en yleth a n ol (1)
w ith CAN. (i) Resu lts. Acetonitrile solutions of 1, 20
mM, and CAN, 0.5 wt %, were continuously flowed and
irradiated (λ > 250 nm) directly in the cavity of an ESR
spectrometer at low temperature (the best conditions
were found at ca. 230 K). The procedure led to the
detection of two different radical species (see Figure 1).
The first radical, 2, was characterized by hyperfine
splitting constants conforming to those reported in the
literature11 for cyclohexadienyl-type radicals (Table 1);
the second, 3, showed an ESR pattern typical of a peroxyl
radical (a sole broad line with a relatively high g-value,
see Table 1). The mechanism of formation of these two
species will be described later (vide infra).
a
b
Typically at 230K. ( 0.0002; the g factors have been deter-
mined by comparison with the g factor of the DPPH (2.0037 (
0.0001). c At 283 K.
cating that there is no involvement of the methyl sub-
stituent in the structure of 2. On the other hand, when
5 was analyzed in the same experimental conditions, it
These findings led to assign to radical 2 a structure
like that of a spiro[oxiranecyclohexadien]yl radical:
enabled the detection (see Figure 2a) of the peroxyl
radical (7) (Table 1), as well as a second species assigned
the structure 6 (Figure 2b). The latter was characterized
by deuterium splittings each of ca. 1/6.5 that of the
corresponding protons (Table 1).
To obtain further evidence for the assignment of 2,
we performed two different experiments on two labeled
compounds: the 1,1-diphenyl[2, 2, 2-2H3]ethanol (4) and
the 1,1-[2′, 3′, 4′, 5′, 6′, 2′′, 3′′, 4′′, 5′′, 6′′-2H10]diphenyl-
ethanol (5).
The photolysis of an acetonitrile solution of 4, 20 mM,
and CAN, 0.5 wt %, led to the same ESR spectrum
(Figure 1) obtained in the photooxidation of 1, thus indi-
These experiments strongly supported the structure
(10) Grossi, L. Res. Chem. Intermed. 1996, 22, 315.
(11) Eiben, K.; Schuler, R. H. J . Chem. Phys. 1975, 62, 3093.
assigned to 2, i.e., the same as the bridged intermediate

2750 J . Org. Chem., Vol. 65, No. 9, 2000
Grossi and Strazzari
Sch em e 2
F igu r e 2. (a) ESR spectrum obtained in the photolysis of an
acetonitrile solution of 5, 20 mM, and CAN, 0.5 wt %, at 230
K, showing both radicals 6 and 7*. Spectrometer setting:
microwave power, 7 mW; modulation amplitude, 0.1 mT; time
constant, 2 s; gain, 2.5 × 104; scan range, 10 mT; scan time,
30 min. (b) Simulation of the spectrum of radical 6.
With regard to the mechanism of formation of the
second detected species, 3, since cyclohexadienyl radicals
are known7b,15 to react rapidly with oxygen to form the
corresponding peroxyl radicals, it can be assumed that
the traces of oxygen dissolved in the solution could
quench the spiro[oxiranecyclohexadien]yl radical 2 to
yield 3, as shown below:
in the neophyl-like rearrangement of the 1,1-diphenyl-
ethoxyl radical. However, since many unsuccessful at-
tempts had been made to detect this intermediate by
ESR,12 further support for these spectroscopic results was
required.
The 1,1-diphenylethoxyl radical was hence generated
through an independent route, i.e., by photolysis of the
corresponding 1,1-diphenylethyl peroxide (8).
An acetonitrile solution of 8, ca. 25 mM, was flowed
(flow rate: 3.5 mL/min) and directly photolyzed in the
ESR cavity at 230 K: in these conditions, only the peroxyl
radical 3 was detected.
In the experiment on the diphenylethyl peroxide (8),
however, two different peroxyl radicals could in principle
be formed. The peroxyl species could in fact arise from
quenching by oxygen of either the bridged intermediate
2 or the final product of the neophyl-like rearrangement,
the 1-phenoxy-1-phenylethyl radical (see Scheme 1),
which is also expected15a to react rapidly with oxygen.
If the latter would be the case, it could explain the
nondetection, by ESR, of this benzyl radical in such
experiment.
It is noteworthy that the photooxidation of 1 by CAN
did not lead to the detection of any radical species
deriving from competitive â-scission processes. It is
established that the 1,1-diphenylethoxyl radical does not
undergo fragmentation,5,7b as a result of a faster carbon-
to-oxygen migration of a phenyl group; similarly the
CAN-induced cyclization seems to occur faster than the
competitive â-scission processes.
(ii) Mech a n istic In ter p r eta tion . ESR detection of
the 2-methyl-2-phenyl-1-oxaspiro[2.5]octa-4,7-dienyl radi-
cal (2) has never been reported before. We believe that,
in the present study, an important role is played by CAN
in the formation and in the stabilization of this interme-
diate. In fact, it is known10 that in solution this nitrate
salt can coordinate a tertiary alcohol molecule, leading
to a 1:1 complex.13,14 The electron-transfer process can
thus be hyphotesized10 to occur directly between the
lanthanide in the excited state and the alcohol. The
coordination of the oxygen of the hydroxyl group to
the transition metal is expected to increase the electro-
philicity of this moiety and, as a result, to favor the
cyclization process (Scheme 2).
Most likely, the resulting bridged intermediate is
stabilized inside the coordination sphere by additional
interactions, for instance, by H-bonding between the
hydrogen atom of the hydroxyl function and one oxygen
atom of a nitrate ligand (see Scheme 2). In the absence
of CAN, this extra stabilization is not present and the
neophyl-like rearrangement proceeds so fast that it is not
possible to detect radical intermediate 2, for which a
lifetime of less than 100 ns has been estimated.6 This
could account for the nondetection of 2 in the ESR
experiments with the diphenylethyl peroxide (8).
(b) P h otooxidation of Tr iph en ylm eth an ol (9) with
CAN. ESR experiments on acetonitrile solutions of 9, 20
mM, and CAN, 0.5 wt %, at 230 K led to analogous
results with those found for the 1,1-diphenylethanol (1),
i.e., the detection of the 2,2-diphenyl-1-oxaspiro[2.5]octa-
4,7-dienyl radical (10) and the corresponding peroxyl
radical 11, Table 1.
It is known5 that the neophyl-like rearrangement for
the triphenylmethoxyl radical is extremely fast compared
to other reactions of alkoxyl radicals, thus, also in this
case, a direct involvement of CAN in the mechanism of
formation and stabilization of the bridged intermediate
has to be invoked (Scheme 2).
(c) P h otooxid a tion of 1,1-Dip h en ylp r op a n ol (12)
w ith CAN. The photolysis of a solution of 12, 20 mM,
and CAN, 0.5 wt %, in the same experimental conditions
(12) (a) Kochi, J . K.; Krusic, P. J . J . Am. Chem. Soc. 1969, 91, 3940.
(b) Hamilton, E. J ., J r.; Fischer, H. Helv. Chim. Acta 1973, 56, 795. (c)
Maillard, B.; Ingold, K. U. J . Am. Chem. Soc. 1976, 98, 1224.
(13) Greatorex, D.; Kemp, T. J . Trans. Faraday Soc. 1971, 56.
(14) The use of cerium(IV) as a qualitative colorimetric reagent for
the detection of alcohols has been known for a long time and even
quantitative methods for the analysis of alcohols based on the red color
of the Ce(IV)-alcohol complexes have been developed: (a) Young, L.
B.; Trahanovsky, W. S. J . Am. Chem. Soc. 1969, 91, 5060. (b)
Trahanovsky, W. S.; Macaulay, D. B. J . Org. Chem. 1973, 38, 1497.
(15) (a) Maillard, B.; Ingold, K. U.; Scaiano, J . C.J . Am. Chem. Soc.
1983, 105, 5095. (b) Pan, X.-M.; Schuchmann, M. N.; von Sonntag, C.
J . Chem. Soc., Perkin Trans. 2 1993, 289.

Photooxidation of Arylcarbinols
J . Org. Chem., Vol. 65, No. 9, 2000 2751
Sch em e 3
Sch em e 4
F igu r e 3. ESR spectrum obtained in the photolysis of an
acetonitrile solution of 14, 50 mM, and CAN, 0.5 wt %, at 283
K, showing both radicals 15 and 16*. Spectrometer setting:
microwave power, 7 mW; modulation amplitude, 0.1 mT; time
constant, 1 s; gain, 6.3 × 103; scan range, 10 mT; scan time,
16 min.
used to analyze 1 and 9, was performed directly in the
ESR cavity. Only one radical species was detected: the
ethyl radical (13), Table 1, deriving from a competitive
â-scission process (Scheme 3).
Ta ble 2. Va r ia tion of th e Ra tio of Ra d ica ls [15]/[16] in
Exp er im en ts on Aceton itr ile Solu tion s of 14, 50 MM, a n d
CAN, 0.5 w t %, a s a F u n ction of Tem p er a tu r ea
T (K)
[15]/[16]
No evidence for the formation of a spiro[oxiranecyclo-
hexadien]yl radical could be obtained. This finding shows
that, for the 1,1-diphenylpropyloxyl radical, even in the
presence of CAN, the main process is the â-scission,
which is evidently faster than the CAN-induced cycliza-
tion.
In the presence of traces of molecular oxygen in
solution, we should expect the ethyl radical to react
rapidly with O2 to yield the corresponding peroxyl radical.
But, in fact, we did not detect any peroxyl species, as it
could be predicted, since it is known10,16 that such type
of radicals cannot be detected when the precursor alkyl
radicals have low molecular weights, i.e., they have a
short carbon atom chain (n < 4).
(d) P h otooxidation of 2-P h en ylpr opan -2-ol, Cu m yl
Alcoh ol, (14) w ith CAN. (i) Resu lts. When an aceto-
nitrile solution of 14, 50 mM, and CAN, 0.5 wt %, was
continuously flowed (optimum conditions were found at
3.5 mL/min) and photolyzed inside the ESR cavity, two
different radical species were detected (see Figure 3).
Both species were generated in the competitive decay
processes,17,18 which the cumyloxyl radical can undergo
(Scheme 4): the â-scission, yielding to the methyl radical
(15), and the hydrogen abstraction from the solvent,
which in this case leads to the cyanomethyl radical (16),
Table 1.
268
273
276
283
291
293
76/24
73/27
70/30
69/31
65/35
60/40
a
At fixed flow rate (3.5 mL/min).
Kin etic Stu d ies. The ESR spectra obtained by photo-
oxidation of 14 with CAN were recorded at variable tem-
perature and, correspondingly, a variation of the relative
concentrations of the two detected radical species as a
function of the temperature was observed (Table 2). In
particular, raising the temperature led to an increase in
the proportion of H-abstraction (see Table 2), suggesting
that this process is characterized by higher activation
energy. This enabled us to conduct a study on the kinetic
competition between the â-scission reaction and the
H-abstraction from acetonitrile for the cumyloxyl radical.
Scheme 5 sets out19 the reactions, which can be
assumed to be involved in the whole process.
In the steady-state approximation:20
d[15]/dt ) k2[14a ] - 2k4[15]2 - 2k5[15] [16] ) 0 (1)
d[16]/dt ) k3′[14a ] - 2k6[16]2 - 2k5[15] [16] ) 0 (2)
The lack of detection of the bridged intermediate in
this experiment indicated that its formation, in this case,
is disfavored, even if mediated by CAN. This is consistent
with the fact that the cumyloxyl radical is not expected
to undergo the neophyl-like rearrangement,5 as the
radical species, which would arise from phenyl migration,
could not be stabilized by conjugation with phenyl groups.
Since it is established21,22 that the bimolecular self-
reactions of all simple alkyl radicals proceed at the
(19) In this scheme, radical-radical termination of two alkoxyl
radicals and of an alkoxyl and a carbon centered radical, yielding
respectively a peroxide and an ether, have not be included. These
reaction can in fact be assumed to be relatively unimportant both for
the very low steady-state concentration of the alkoxyl radical, which
rapidly undergoes â-scission and H-abstraction, and for the low
(16) Fessenden, R. W.; Schuler, R. H. J . Chem. Phys. 1963, 39, 2147.
(17) Walling, C. Pure Appl. Chem. 1967, 15, 69.
energies involved in the formed bonds. Also, reaction
2 can be
(18) (a) Baigne´e, A.; Howard, J . A.; Scaiano, J . C.; Stewart, L. C. J .
Am. Chem. Soc. 1983, 105, 6120. (b) Avila, D. V.; Brown, C. E.; Ingold,
K. U.; Lusztyk, J . J . Am. Chem. Soc. 1993, 115, 466.
considered to follow a pseudo-first-order kinetic equation, since aceto-
nitrile, present as the solvent in large excess, can be assumed not to
change its concentration significantly.

2752 J . Org. Chem., Vol. 65, No. 9, 2000
Grossi and Strazzari
Sch em e 5
F igu r e 4. Log k3′ values vs 1/T (10-3 K-1) in experiments on
acetonitrile solutions of 14, 50 mM, and CAN, 0.5wt %, at fixed
flow rate (3.5 mL/min).
observed (see Figure 4) and the following equation is
obtained:
Ta ble 3. Va r ia tion of th e Ra te Con sta n ts of th e
Com p etitive P r ocesses of â-scission (k2) a n d
H-Abstr a ction fr om CH3CN (k3′) for th e Cu m yloxy
Ra d ica l a s a F u n ction of Tem p er a tu r e
log k3′ ) (15.56 ( 0.35) - [(13.72 ( 0.45)]/θ (4)
where θ ) 2.303RT kcal/mol.
Thus, the activation energy for this process can be
estimated as
T (K)
1/T (10-3 K-1
)
k2 (104 s-1
)
k3′ (104 s-1
)
log k3′
293
291
283
276
273
268
3.41
3.44
3.53
3.62
3.66
3.73
36.3
32.4
20.3
13.2
10.9
7.83
24.2
17.5
9.12
5.65
4.02
2.47
5.38
5.24
4.96
4.75
4.60
4.39
Ea ) (13.72 ( 0.45) kcal/mol
and the value of k3′ at 303 K can be calculated from eq 4
k3′ ) 4.81 × 105 s-1
diffusion-controlled limit (which means k4 ) k5 ) k6),
steady-state analysis leads to the simple relation:
which sounds reasonable.
k3′/k2 ) [16]/[15]
(3)
(e) P h ot ooxid a t ion of Na p h t h ylca r b in ols w it h
CAN. The 1,2 migration of a 2-naphthylethyl radical is
2 orders of magnitude faster21 than the analogous migra-
tion of an aryl group in 2-arylethyl radicals. For the
former, in fact, the loss of stabilization for resonance, due
to the formation of the bridged intermediate, is lower
than that for the 2-arylethyl system. Thus, for a 2-naph-
thylethyl radical, the cyclization process, leading to a
bridged-radical, is more favored, i.e., the activation
energy for the formation of this intermediate is lower.
Similarly, it is expected23 that the neophyl-like rear-
rangement of an alkoxyl radical with a naphthyl sub-
stituent is even more favored than the analogous process
for a phenyl-substituted alkoxyl radical.
Thus, experiments on the 2-(1-naphthyl)propan-2-ol
(17) and the 1-naphthyldiphenylmethanol (18) were also
performed. However, the photolysis of an acetonitrile
solution of 17, 30 mM, and CAN, 1 wt %, as well as that
of a solution of 18, 30 mM, and CAN, 1 wt %, did not
allow the detection of any radical species.
Hence, the ratio of the concentrations of the cyano-
methyl radical (16) and the methyl radical (15) parallels
the ratio of the kinetic constants of the H-abstraction and
the â-scission processes.
The rate constant for the â-scission process of the
cumyloxyl radicals, in acetonitrile, has been determined
18b
at 303 K [k2 ) (6.33 ( 0.43) × 105 s-1
]
and a value18a
for the preexponential factor of log A2 ) 12.4 can be
assumed for this reaction. Thus, it is possible to estimate
the values of k2 at different temperatures and, through
eq 3, the corresponding values of k3′ (see Table 3).
A plot of log k3′ versus 1/T (see Table 3) should give
a straight line with an intercept of logA3′ and a slope
of -(Ea3′/2.303R), from which it should be possible
to evaluate the parameters of the Arrhenius equation
for the pseudo-first-order reaction of H-abstraction
from acetonitrile for the cumyloxyl radical. When these
data are plotted in this way, excellent agreement is
Since no radical species were detected in these experi-
ments, neither the bridged intermediates nor radicals
deriving from competitive processes, such as â-scission
or H-abstraction, we believe that only a weak interaction
occurs between the R-naphthylcarbinols and CAN. Most
likely, the coordination to the ceric ion of the hydroxyl
(20) The factor 2 in these equations has two different meanings (see
ref 21). First, rate constants are generally defined in terms of the rate
of formation of products, but in this system we are interested in the
loss of reactants; actually, in reactions 4 and 6 two radicals 15 and
two radicals 16, respectively, are destroyed and so k4 and k6 must be
multiplied by two. Second, the cross-termination reaction 5 is statisti-
cally favored by a factor of 2 (i.e., we have the same probability that
radical 15 reacts with 16 as that 16 react with 15) and hence k5 must
be multiplied by two.
(21) Griller, D.; Ingold, K. U. Acc. Chem. Res. 1980, 13, 317.
(22) Griller, D.; Ingold, K. U. Acc. Chem. Res. 1980, 13, 193.
(23) Neckers, D. C.; Linden, S.-M.; Taliano, R. J .; Williams, B. L.;
Zakrzewski, D. Tetrahedron Lett. 1988, 29, 3029.

Photooxidation of Arylcarbinols
J . Org. Chem., Vol. 65, No. 9, 2000 2753
Sch em e 6
than the rates of competitive processes, which the corre-
sponding alkoxyl radicals can undergo, such as â-scission
and H-abstraction. Radical species deriving from the oc-
currence of these processes have been identified by ESR.
Finally, a kinetic study on the competition between the
â-scission and H-abstraction decay processes for the
cumyloxyl radical has been made. This study has led to
estimate the parameters of the Arrhenius equation for
the pseudo-first-order reaction of the H-abstraction from
acetonitrile for the cumyloxyl radical [log A ) (15.56 (
0.35); Ea ) [(13.72 ( 0.45) kcal/mol].
function of these substrates cannot be accomplished due
to steric hindrance. As a result, the formation of the
complex, which seems to play an important role in the
mechanism of production and stabilization of the bridged
intermediate, cannot take place.
Exp er im en ta l Section
Mater ials. 1,1-Diphenylethanol, triphenylmethanol, 2-phen-
yl-2-propanol, and ceric ammonium nitrate (99.99% pure) were
commercial products used as received. Acetonitrile was 99.9+%
HPLC grade.
Con clu sion s
The ESR detection of the bridged-radicals 2 and 10,
which are believed5,6 to be involved as intermediates in
the neophyl-like rearrangements respectively of the 1,1-
diphenylethoxyl and the triphenylmethoxyl radicals, has
been obtained, by generating the precursor alkoxyl
radicals by photooxidation of the corresponding alcohols,
1 and 9, with CAN. Most likely, this nitrate salt favors
the cyclization process, which leads to the bridged
intermediates, by increasing the electrophilicity of the
alkoxyl moiety by coordination. This accounts for the
nondetection of such intermediates in the experiments
with R-naphthyl derivatives, where the CAN-alcohol
complex cannot be formed, probably due to steric hin-
drance.
In the absence of CAN, as in the experiment with
1,1-diphenylethyl peroxide (8), no evidence for the forma-
tion of 2 is obtained. This result leads to hypothesize that
the bridged intermediate can be stabilized by interaction
with CAN. If this stabilization lacks, the neophyl-like
rearrangement proceeds fast and the bridged-radical
cannot be detected due to its short lifetime.
1,1-Dip h en yl[2,2,2-2H3]eth a n ol (4). A suspension of turn-
ings of Mg (22 mmol in 15 mL of anhydrous THF under an N2
atmosphere) was gently refluxed in the presence of crystals
of I2 and ca. one-third of a solution of CD3I (20 mmol in 5 mL
of anhydrous THF) was added dropwise. The mixture was then
warmed until the violet coloring, due to the iodine, disap-
peared. The remaining CD3I was added and the solution
refluxed for additional 30 min. Subsequently, the solution of
the freshly prepared CD3MgI was transferred in a three-
necked flask by means of a long double-tipped deflecting needle
so that the Grignard reagent could be easily separated from
the excess of Mg. After dilution with 10 mL of anhydrous Et2O,
the mixture, kept under N2, was cooled at -20 °C and then a
solution of benzophenone (20 mmol in 10 mL of anhydrous
Et2O) was added dropwise. After stirring for 30 min at room
temperature, the reaction was quenched with sat. NH4Cl/H2O
and the organic layer extracted with Et2O, dried over Na2SO4
and concentrated. The crude product was purified by chroma-
tography on silica gel (petroleum ether/dichloromethane )
1:1): 2.44 g; 61%; mp 78-80 °C; 1H NMR (CDCl3, 200 MHz) δ
2.44 (1H, s, OH), 7.16-7.53 (10H, m); 13C NMR (CDCl3, 50.3
MHz) δ 30.5 (m, CD3), 76.8 (C), 126.1 (CH), 127.5 (CH), 128.8
(CH), 148.1 (C); HRMS calcd for C14H11D3O 201.1233, found
201.1234.
A different mechanism for the formation of an alkoxyl
radical from the interaction of the corresponding alcohol
with CAN could also be taken into account. In fact, it is
claimed24 that, in the photolysis of CAN, a nitrate radical
is produced:
1,1-[2′,3′,4′,5′,6′,2′′,3′′,4′′,5′′,6′′-2H10]Dip h en yleth a n ol (5).
To a stirred solution of bromobenzene-d5 (11 mmol in 40 mL
of anhydrous THF), kept at -50 °C and under an N2 atmo-
sphere, first n-butyllithium (12 mmol, 1.6 M in hexane) and
then [2′, 3′, 4′, 5′, 6′-2H5]acetophenone (11 mmol in ca. 5 mL
of anhydrous THF) were added dropwise. The mixture was
allowed to react at room temperature (30-40 min), and
subsequently the reaction was quenched with a saturated
aqueous solution of NH4Cl. After being extracted with Et2O,
the organic layer was dried over Na2SO4 and concentrated in
vacuo. The purified product was obtained by chromatography
on silica gel with CH2Cl2 as the eluent: 1.16 g; 52.1%; mp
73-76 °C; 1H NMR (CDCl3, 200 MHz) δ 1.91 (3H, s), 2.30 (1H,
s, OH); 13C NMR (CDCl3, 50.3 MHz) δ 30.7 (CH3), 76.1 (C),
125.4 (t, CD), 126.3 (t, CD), 127.6 (t, CD), 147.8 (C); HRMS
calcd for C14H4D10O 208.1672, found 208.1674.
Ce(IV)NO3- 9hν8 Ce(III) + NO3
•
This radical could then attack the hydroxyl function
either by direct hydrogen abstraction or by electron
transfer, leading in both cases to the corresponding
alkoxyl radical (Scheme 6).
But, if that would be the case, we could not explained
the fact that the naphthyl derivatives are not oxidized
by CAN, since an intimate interaction should not be
required. Moreover, this oxidation pathway does not lead
to any extra-stabilization for the bridged intermediate;
thus, we would not expect that the use of CAN could favor
the formation of this radical, as it actually is.
The same procedure, i.e., the photooxidation by CAN,
has been applied to the study of the 1,1-diphenylpropanol
(12) and the cumyl alcohol (14). For these substrates, it
has not been possible to detect the spiro[oxiranecyclo-
hexadien]yl-type intermediates, probably because the
rate of the cyclization process, mediated by CAN, is lower
[2′,3′,4′,5′,6′-2H5]Acetop h en on e. 1-[2′, 3′, 4′, 5′, 6′-2H5]-
Phenylethanol (17 mmol in 40 mL of CH2Cl2) was oxidized to
the corresponding ketone with pyridinium chlorochromate 98%
(3.75 g). The oxidation course was followed by thin-layer
chromatography, and after 1 h, the reaction was quenched with
Et2O anhydrous. The organic phase was washed with Et2O
and separated from chromium salts by filtration. The solvent
was removed under reduced pressure and the pure product
recovered by chromatography on silica gel (petroleum ether/
diethyl ether ) 2:1): 1.34 g; 61.8%; 1H NMR (CDCl3, 200 MHz)
δ 2.58 (3H, s); 13C NMR (CDCl3, 50.3 MHz) δ 27.3 (CH3), 128.9
(t, CD), 129.1 (t, CD), 133.7 (t, CD), 137.7 (C), 198.8 (C); HRMS
calcd for C8H3D5O 125.0889, found 125.0890.
(24) (a) Baciocchi, E.; Del Giacco, T.; Rol, C.; Sebastiani, G. V.
Tetrahedron Lett. 1985, 26, 541. (b) ibidem, 3353. (c) Del Giacco, T.;
Baciocchi, E.; Steenken, S. J . Phys. Chem. 1993, 97, 5451.
1-[2′,3′,4′,5′,6′-2H5]P h en yleth a n ol. To a cooled solution
(-50 °C) of phenyllithium-d5 (30 mmol in 50 mL of anhydrous

2754 J . Org. Chem., Vol. 65, No. 9, 2000
Grossi and Strazzari
THF under N2), obtained by halogen-metal exchange from
bromobenzene-d5 (30 mmol) and n-butyllithium (32 mmol, 1.6
M in hexane), was added dropwise 3.4 mL of acetaldehyde (60
mmol in 5 mL of anhydrous THF). After additional stirring
for 30-40 min at room temperature, the reaction was quenched
with saturated NH4Cl/H2O and the mixture extracted with
Et2O. The organic layer was then dried over Na2SO4 and
concentrated in vacuo. After chromatography on silica gel
(petroleum ether/diethyl ether ) 1:1) the desired product was
obtained: 3.10 g; 81.4%; 1H NMR (CDCl3, 200 MHz) δ 1.43
(3H,d), 3.11 (1H, s, OH), 4.80 (1H, q); 13C NMR (CDCl3, 50.3
MHz) δ 24.9 (CH3), 69.9 (CH), 124.8 (t, CD), 126.6 (t, CD),
127.7 (t, CD), 145.6 (C); HRMS calcd for C8H5D5O 127.1045,
found 127.1043.
isolated by chromatography on silica gel (petroleum ether/
diethyl ether ) 3:1): 2.10 g; 67.7%; mp 95-96 °C (lit.26 mp 95
°C); 1H NMR (CDCl3, 200 MHz) δ 0.87 (3H, t), 2.08 (1H, s,
OH), 2.31 (2H, q), 7.15-7.46 (10H, m); 13C NMR (CDCl3, 50.3
MHz) δ 8.7 (CH3), 34.9 (CH2), 79.0 (C), 126.6 (CH), 127.3 (CH),
128.6 (CH), 147.4(C).27
1-Na p h th yld ia lk ylm eth a n ols. The 2-(1-naphthyl)propan-
2-ol (17) was prepared according to a literature method.28 The
1-naphthyldiphenylmethanol (18) was synthesized adopting
the same procedure:28 2.99 g; 69%; mp 135-136 °C (lit.29 mp
1
134-135 °C); H NMR (CDCl3, 200 MHz) δ 3.32 (1H, s, OH),
6.85 (1H, d), 7.20-7.42 (13 H, m), 7.82 (2H, t), 8.08 (1H, d);
13C NMR (CDCl3, 50.3 MHz) δ 84.0 (C), 124.9 (CH), 126.0 (CH),
126.2 (CH), 127.8 (CH), 128.4 (CH), 128.7 (CH), 128.9 (CH),
129.5 (CH), 130.0 (CH), 131.9 (C), 135.6 (C), 142.7 (C), 147.6
(C); HRMS calcd for C23H18O 310.1358, found 310.1362.
ESR Exp er im en ts. The acetonitrile solutions of the aryl-
carbinols (10-50 mM) and CAN (0.5-1wt %) were purged with
gaseous N2 for at least 45 min before use. The mixtures were
then continuously flowed (flow rate 2.0-3.5 mL/min) through
a flat quartz cell (0.3 mm width) inside an ESR cavity and
directly irradiated with an OSRAM HBO 500 W/2 high-
pressure mercury lamp. A Varian E-104 spectrometer,
equipped with a variable temperature control system, was
used. The temperatures inside the spectrometer cavity, in the
kinetic experiments, were calibrated by means of a Ni/Cu
thermocouple inserted into the flat quartz cell after the
spectral measurements. Relative radical concentrations and
hyperfine splitting assignments were obtained by means of
computer simulations.
1,1-Dip h en yleth yl P er oxid e (8).25 To a mechanically
stirred solution of 1,1-diphenylethanol (20 mmol) dissolved
in 60 mL of glacial acetic acid and 20 mL of Et2O, kept under
N2 at 0 °C, was added dropwise hydrogen peroxide 30 wt %
(10 mL), together with four drops of concentrated sulfuric acid.
The reaction course was followed by thin-layer chromatogra-
phy, and after ca. 8 h, the mixture was worked up by addition
of an aqueous solution of NaOH, taking care that the temper-
ature would not exceed 2/3 °C. The solution was then kept at
-4 °C for 12 h, after which time the organic phase was
extracted with Et2O, washed with saturated NaHCO3/H2O,
and dried over Na2SO4. The solvent was evaporated carefully
under reduced pressure and the crude, consisting of a mixture
of 1,1-diphenylethyl hydroperoxide and 1,1-diphenylethanol
(70:30), used without further purification. In the second step
of the synthesis, the mixture of hydroperoxide and alcohol was
suspended in 100 mL of formic acid 90% and stirred for ca. 4
h. The formation of the peroxide was checked by thin-layer
chromatography. The solution was kept at -4 °C overnight
so that the peroxide could slowly precipitate. The solid phase
was then filtered, washed with distilled water and dissolved
in Et2O. The obtained solution was dried over Na2SO4 and
concentrated in vacuo. The residue was crystallized twice from
petroleum ether and benzene, yielding the pure product: 1.65
Ack n ow led gm en t. The authors thank F. Montanari
for skillful assistance. They also thank Prof. R. Leardini
for helpful suggestions for the synthesis of the 1,1-
diphenylethyl peroxide and Prof. J . C. Scaiano for stimu-
lating discussions. This work was carried out with the
financial support of the Ministry of the University and
Scientific and Technological Research (MURST), Rome.
1
g; 41.9%; mp 148-150 °C; H NMR (CDCl3, 200 MHz) δ 1.98
(6H, s), 7.25 (20H, s); 13C NMR (CDCl3, 50.3 MHz) δ 27.0 (CH3),
86.7 (C), 127.7 (CH), 127.8 (CH), 128.4 (CH), 145.7 (C).
1,1-Dip h en ylp r op a n ol (12). To a solution of benzophenone
(14 mmol in 30 mL of anhydrous THF), stirred under N2 and
cooled at -50 °C, was added dropwise ethylmagnesium
bromide (5 mL, 15 mmol, 1.0 M in THF). The reaction mixture
was allowed to warm to room temperature and, after 50 min,
was quenched with saturated NH4Cl/H2O. The organic phase
was then extracted with Et2O and dried over Na2SO4. The
solvent was removed under reduced pressure and the product
J O991856T
(25) The synthesis of this substrate was performed following a
procedure suggested by Prof. R. Leardini.
(26) Bergmann, E. D.; Meyer, A. M. J . Org. Chem. 1969, 30, 2840.
(27) The spectroscopic assignments were conforming to those
reported in the literature: Guijano, D.; Guillena, G.; Mancheno, B.;
Yus, M. Tetrahedron 1994, 50, 3427.
(28) Casarini, D.; Lunazzi, L.; Mazzanti, A. J . Org. Chem. 1997, 62,
3315.
(29) Russel, B. J . Am. Chem. Soc. 1966, 88, 5491.