Functional aspects of Gif-type oxidation of hydrocarbons mediated by iron picolinate H2O2-dependent systems: Evidence for the generation of carbon- and oxygen-centered radicals

Article abstract of DOI:10.1021/ja000063h

The present investigation explores the functional features of several novel and other previously ill-defined ferrous and ferric complexes of the picolinic acid anion (Pic), which are used to mediate Gif-type oxidation of hydrocarbons by H₂O₂. Complexes [Fe(Pic)₂(py)₂], [Fe(Pic)₃]·0.5py, [Fe₂O(Pic)₄(py)₂], [Fe₂(μ-OH)₂(Pic)₄], and FeCl₃ have been employed in oxygenations of adamantane by H₂O₂ mostly in py/AcOH to reveal that tert- and sec-adamantyl radicals are generated in Gif solutions. The alleged absence of sec-adamantyl radicals from Gif product profiles has been previously interpreted as compelling evidence in support of a non-radical mechanism for the activation of secondary C-H sites in Gif chemistry. The product profile is entirely dictated by trapping of the diffusively free adamantyl radicals, since authentic tert- and sec-adamantyl radicals are shown to partition between dioxygen and protonated pyridine at 4% O₂ (in N₂), or between dioxygen and TEMPO at 100% O₂, in a manner analogous to that observed in Gif oxygenations of adamantane. The low tert/sec selectivity (2.2-4.5) obtained, increasing with increasing dioxygen partial pressure, and the small intramolecular kinetic isotope effect values revealed by employing adamantane-1,3-d₂ (1.06(6) (Ar); 1.73(2) (4% O₂ in N₂)), indicate the presence of an indiscriminate oxidant under inert atmosphere, coupled to a more selective oxidant at higher partial pressures of dioxygen. Gif oxygenation of DMSO by H₂O₂ mediated by [Fe(Pic)₂(py)₂] provides pyridine-trapped methyl radicals under argon, as expected for the addition reaction of hydroxyl radicals to DMSO. The reaction is progressively inhibited by increasing amounts of EtOH, generating pyridine-captured CH₃·CHOH and ·CH₂CH₂OH radicals. Quantification of the DMSO- versus EtOH-derived alkyl radicals affords an estimate of k(EtOH)/k(DMSO) equal to 0.34(3), in reasonable agreement with the kinetics of radiolytically produced hydroxyl radicals (k(EtOH)/k(DMSO) = 0.29). The formation of methyl radicals in Gif oxygenation of DMSO is also supported by the quantitative generation of tert-adamantyl radicals in the presence of 1-iodoadamantane. These results are consistent with the action of hydroxyl radicals in Gif oxygenations by Fe(II/III)/H₂O₂ (Ar), most likely coupled to substrate-centered alkoxyl radicals under O₂. The oxygen-centered radicals perform H-atom abstractions from Gif substrates to generate diffusively free carbon-centered radicals, in accord with previously reported findings.

Full text of DOI:10.1021/ja000063h

J. Am. Chem. Soc. 2000, 122, 7503-7517
7503
Functional Aspects of Gif-Type Oxidation of Hydrocarbons Mediated
by Iron Picolinate H2O2-Dependent Systems: Evidence for the
Generation of Carbon- and Oxygen-Centered Radicals
Salma Kiani,† Amy Tapper,† Richard J. Staples,‡ and Pericles Stavropoulos*,†
Contribution from the Departments of Chemistry, Boston UniVersity, Boston, Massachusetts 02215, and
HarVard UniVersity, Cambridge, Massachusetts 02138
ReceiVed January 4, 2000
Abstract: The present investigation explores the functional features of several novel and other previously
ill-defined ferrous and ferric complexes of the picolinic acid anion (Pic), which are used to mediate Gif-type
oxidation of hydrocarbons by H2O2. Complexes [Fe(Pic)2(py)2], [Fe(Pic)3]‚0.5py, [Fe2O(Pic)4(py)2], [Fe2(µ-
OH)2(Pic)4], and FeCl3 have been employed in oxygenations of adamantane by H2O2 mostly in py/AcOH to
reveal that tert- and sec-adamantyl radicals are generated in Gif solutions. The alleged absence of sec-adamantyl
radicals from Gif product profiles has been previously interpreted as compelling evidence in support of a
non-radical mechanism for the activation of secondary C-H sites in Gif chemistry. The product profile is
entirely dictated by trapping of the diffusively free adamantyl radicals, since authentic tert- and sec-adamantyl
radicals are shown to partition between dioxygen and protonated pyridine at 4% O2 (in N2), or between dioxygen
and TEMPO at 100% O2, in a manner analogous to that observed in Gif oxygenations of adamantane. The low
tert/sec selectivity (2.2-4.5) obtained, increasing with increasing dioxygen partial pressure, and the small
intramolecular kinetic isotope effect values revealed by employing adamantane-1,3-d2 (1.06(6) (Ar); 1.73(2)
(4% O2 in N2)), indicate the presence of an indiscriminate oxidant under inert atmosphere, coupled to a more
selective oxidant at higher partial pressures of dioxygen. Gif oxygenation of DMSO by H2O2 mediated by
[Fe(Pic)2(py)2] provides pyridine-trapped methyl radicals under argon, as expected for the addition reaction of
hydroxyl radicals to DMSO. The reaction is progressively inhibited by increasing amounts of EtOH, generating
•
•
pyridine-captured CH3 CHOH and CH2CH2OH radicals. Quantification of the DMSO- versus EtOH-derived
alkyl radicals affords an estimate of kEtOH/kDMSO equal to 0.34(3), in reasonable agreement with the kinetics
of radiolytically produced hydroxyl radicals (kEtOH/kDMSO ) 0.29). The formation of methyl radicals in Gif
oxygenation of DMSO is also supported by the quantitative generation of tert-adamantyl radicals in the presence
of 1-iodoadamantane. These results are consistent with the action of hydroxyl radicals in Gif oxygenations by
FeII/III/H2O2 (Ar), most likely coupled to substrate-centered alkoxyl radicals under O2. The oxygen-centered
radicals perform H-atom abstractions from Gif substrates to generate diffusively free carbon-centered radicals,
in accord with previously reported findings.
Introduction
t-BuOOH) (1:20) in py/AcOH (2:1),4b has been proposed by
Sawyer5 to mediate “oxygenated Fenton” chemistry. Although
the presumed active oxidants are differently formulated (a high-
valent FeVdO unit in Gif1 and an unusual [Fe(OOR)(O2)(Pic)2-
(pyH+)] species in oxygenated Fenton chemistry5), both systems
have been interpreted to perform oxygenation of substrates via
pathways that, with a few exceptions, do not involve oxygen-
or carbon-centered radicals.
One argument in support of the non-radical mechanism in
Gif chemistry is the alleged selectivity for the oxygenation of
secondary positions (leading primarily to ketones) over tertiary
or primary sites (sec > tert > prim).6 This unusual selectivity
(which in our hands has been confirmed for isopentane,7 albeit
Early versions of Gif-type hydrocarbon-oxidizing systems1
were composed of an FeII/III complex, a reducing agent (Fe, Zn,
electrochemical cathode) and dioxygen (or air), operating
primarily in a pyridine/carboxylic acid matrix (10:1 v/v).2 More
recently, Gif reagents have been represented by FeIII/H2O2 or
FeIII/t-BuOOH (60 °C) combinations,3 frequently assisted by
the presence of 2-picolinic acid (PicH) to enhance the rate of
catalytic oxygenations.4 A typical system employed by Barton
to accumulate most of the mechanistic data known to date
involves the combination of FeCl3/PicH/H2O2 (1:4:4) in pyridine
or py/AcOH.3 A closely related reagent, Fe(Pic)2/H2O2 (or
† Boston University.
‡ Harvard University.
(5) Sawyer, D. T.; Sobkowiak, A.; Matsushita, T. Acc. Chem. Res. 1996,
29, 409-416.
(1) (a) Barton, D. H. R.; Doller, D. Acc. Chem. Res. 1992, 25, 504-
512. (b) Barton, D. H. R. Aldrichim. Acta 1990, 23, 3-11.
(2) Barton, D. H. R.; Be´vie`re, S. D.; Chavasiri, W.; Csuhai, E.; Doller,
D.; Liu, W.-G. J. Am. Chem. Soc. 1992, 114, 2147-2156.
(3) Barton, D. H. R.; Hu, B.; Taylor, D. K.; Rojas Wahl, R. U. J. Chem.
Soc., Perkin Trans. 2 1996, 1031-1041.
(4) (a) Balavoine, G.; Barton, D. H. R.; Boivin, J.; Gref, A. Tetrahedron
Lett. 1990, 3, 659-662. (b) Tung, H.-C.; Kang, C.; Sawyer, D. T. J. Am.
Chem. Soc. 1992, 114, 3445-3455.
(6) (a) Barton, D. H. R.; Boivin, J.; Gastiger, M.; Morzycki, J.; Hay-
Motherwell, R. S.; Motherwell, W. B.; Ozbalik, N.; Schwartzentruber, K.
M. J. Chem. Soc., Perkin Trans. 1 1986, 947-955. (b) Barton, D. H. R.;
Doller, D.; Ozbalik, N.; Balavoine, G.; Gref, A. Tetrahedron Lett. 1990,
31, 353-356. (c) Barton, D. H. R.; Chavasiri, W. Tetrahedron 1997, 53,
2997-3004.
(7) Singh, B.; Long, J. R.; Fabrizi de Biani, F.; Gatteschi, D.; Stavro-
poulos, P. J. Am. Chem. Soc. 1997, 119, 7030-7047.
10.1021/ja000063h CCC: $19.00 © 2000 American Chemical Society
Published on Web 07/22/2000

7504 J. Am. Chem. Soc., Vol. 122, No. 31, 2000
Kiani et al.
at such low yields as to warrant caution) has been accounted
for by a general mechanism1 (eqs 1-6; illustrated for cyclo-
hexane) that involves concerted [2 + 2] addition of C-H bonds
across high-valent FeVdO units (eqs 1, 2). The proposed four-
atom centered intermediate would thus favor sec-C-H oxy-
genation as a compromise between steric encumbrance and C-H
bond strength. Another important fact taken into account in the
formulation of the Barton mechanism is the detection of
cyclohexylhydroperoxide2 (eq 4), eventually decomposing via
metal-dependent steps to yield cyclohexanone (major product)
and cyclohexanol (minor product) (eqs 5, 6). The important
observation by M. J. Perkins8 that the oxygen incorporated in
cyclohexylhydroperoxide is derived from dioxygen, raised the
question whether the immediate precursor to the alkylhydro-
peroxide (ROOH) is actually free alkylperoxyl radical (ROO•).
The latter will be generated by the reaction of diffusively free
alkyl radicals (R•) with internally produced or externally
provided dioxygen. A non-radical version of the sequence of
events leading to the formation of the alkylhydroperoxide has
been accommodated in Barton’s mechanism by assuming metal-
bound alkyl (eq 3) and alkylperoxo moieties (eq 4).
and Ingold9b,10,11 lends credence to a typical Haber-Weiss-
Walling13 radical mechanism (eqs 9-17), featuring involvement
of t-BuO•/t-BuOO• radicals and substrate-centered alkyl/alkyl-
peroxyl radicals. This pathway is usually suspected by account
of a one/2-ol ratio of g 1 (typical values tend to be close to 1)
owing to Russell14 decomposition of alkylperoxyl radicals (eqs
15, 16). In addition, mixed alkyl(tert-butyl)peroxides are
observed11c in good yields (eq 17). Other diagnostic elements
of alkoxyl radical (RO•) involvement as the hydrogen abstracting
agent include observation of byproducts of â-scission and
intramolecular hydrogen abstraction,13a as well as trapping of
RO•/ROO• radicals by analogues of vitamin E (R-tocopherol)15
or diphenylamine.10 Following Barton’s reappraisal16 of the
chemistry of t-BuOOH-based Gif systems, there is now
consensus that the main role of these Fe(II)/Fe(III) reagents is
to generate t-BuO•/t-BuOO• radicals (eqs 9, 10).
FeII + t-BuOOH f FeIII + OH- + t-BuO•
FeIII + t-BuOOH f FeII + H+ + t-BuOO•
(9)
(10)
(11)
t-BuO• + c-C6H12 f t-BuOH + c-C6H11
•
FeIII + H2O2 f FeVdO + H2O
(1)
(2)
t-BuO• + t-BuOOH f t-BuOH + t-BuOO•
2t-BuOO• f 2t-BuO• + O2
(12)
(13)
FeVdO + c-C6H12 f FeV(OH)(c-C6H11)
FeV(OH)(c-C6H11) + H2O2 f
FeIII(OH)(c-C6H11) + O2 + 2H+ (3)
c-C6H11• + O2 f c-C6H11OO•
(14)
2c-C6H11OO• f c-C6H11OOOOc-C6H11 f
FeIII(OH)(c-C6H11) + O2 f
FeIII(OH)(OO-c-C6H11) (+ H2O) h
c-C6H10(O) + c-C6H11OH + O2 (15)
FeIII + 2OH- + c-C6H11OOH (4)
FeIII(OH)(OO-c-C6H11) f FeIII + 2OH- + c-C6H10(O) (5)
FeIII(OH)(OO-c-C6H11) + reducing agents f c-C6H11OH (6)
c-C6H11OO• + t-BuOO• f c-C6H11OOOOt-Bu f
c-C6H10(O) + t-BuOH + O2 (16)
c-C6H11• + t-BuOO• f c-C6H11OOt-Bu
(17)
The unusually high one/2-ol ratio (∼3-10) obtained in Gif
oxygenations of sec-C-H positions (cyclohexane, adamantane)
indicates that, in addition to eqs 15 and 16, decomposition of
sec-alkylperoxyl radicals may take place via metal-dependent
pathways. In Barton’s mechanism,1 the enhancement of ketone
production is attributed to formation and decay of Fe(III)-
alkylperoxy intermediates (eqs 4, 5) in a non-radical mechanism.
More recently, a general argument concerning the non-appli-
cability of the Haber-Weiss-Walling mechanism for the Fe(II)-
dependent decomposition of secondary and primary (as opposed
to tertiary) alkylhydroperoxides has been proposed by Barton17
to explain that not much alcohol is obtained in pyridine
solutions. In other solvents (CH3CN/AcOH, AcOH/H2O), the
In addition to Fe(III)/H2O2 based Gif-type reagents that give
rise to the non-radical “FeIII/FeV manifold” noted above, Barton
has also recognized3 the existence of an independent “FeII/FeIV
manifold”, generated by circumstantial Fe(II)/H2O2 interactions,
that yields alkyl radicals via eqs 7 and 8.
FeII + H2O2 f FeIVdO + H2O
(7)
FeIVdO + R-H f HO-FeIV-R f HO-FeIII + R• (8)
The applicability of a non-radical mechanism has been
experimentally challenged, primarily for those branches of
Gif9 and “oxygenated Fenton”10 systems (as well as other
biomimetically designed hydrocarbon-oxidizing reagents11)
employing t-BuOOH and other diagnostic alkylhydroperoxides.
Compelling evidence provided for these systems by Minisci9a,12
(12) (a) Minisci, F.; Fontana, F.; Araneo, S.; Recupero, F.; Banfi, S.;
Quici, S. J. Am. Chem. Soc. 1995, 117, 226-232. (b) Minisci, F.; Fontana,
F.; Araneo, S.; Recupero, F. J. Chem. Soc., Chem. Commun. 1994, 1823-
1824. (c) Minisci, F.; Fontana, F. Tetrahedron Lett. 1994, 35, 1427-1430.
(13) (a) Walling, C. Acc. Chem. Res. 1998, 31, 155-157. (b) Walling,
C. Acc. Chem. Res. 1975, 8, 125-131.
(8) Knight, C.; Perkins, M. J. J. Chem. Soc., Chem. Commun. 1991, 925-
927.
(9) (a) Minisci, F.; Fontana, F.; Araneo, S.; Recupero, F.; Zhao, L. Synlett
1996, 119-125. (b) Snelgrove, D. W.; MacFaul, P. A.; Ingold, K. U.;
Wayner, D. D. M. Tetrahedron Lett. 1996, 37, 823-826.
(10) MacFaul, P. A.; Wayner, D. D. M.; Ingold, K. U. Acc. Chem. Res.
1998, 31, 159-162.
(11) Arends, I. W. C. E.; Ingold, K. U.; Wayner, D. D. M. J. Am. Chem.
Soc. 1995, 117, 4710-4711. (b) MacFaul, P. A.; Arends, I. W. C. E.; Ingold,
K. U.; Wayner, D. D. J. Chem. Soc., Perkin Trans. 2 1997, 135-145. (c)
MacFaul, P. A.; Ingold, K. U.; Wayner, D. D. M.; Que, L., Jr. J. Am. Chem.
Soc. 1997, 119, 10594-10598.
(14) Russell, G. A. J. Am. Chem. Soc. 1957, 79, 3871-3877.
(15) (a) Evans, C.; Scaiano, J. C.; Ingold, K. U. J. Am. Chem. Soc. 1992,
114, 4589-4593. (b) Valgimigli, L.; Banks, J. T.; Ingold, K. U.; Lusztyk,
J. J. Am. Chem. Soc. 1995, 117, 9966-9971. (c) Burton, G. W.; Ingold, K.
U. Acc. Chem. Res. 1986, 19, 194-201.
(16) (a) Barton, D. H. R. Synlett 1997, 229-230. (b) Barton, D. H. R.;
Le Gloahec, V. N.; Patin, H.; Launay, F. New J. Chem. 1998, 22, 559-
563. (c) Barton, D. H. R.; Le Gloahec, V. N.; Patin, H. New J. Chem. 1998,
22, 565-568.
(17) Barton, D. H. R.; Launay, F. Tetrahedron 1997, 53, 14565-14578.

Generation of Radicals in Gif-Type Oxidations
J. Am. Chem. Soc., Vol. 122, No. 31, 2000 7505
Scheme 1
involved. Furthermore, by using suitable substrates (dimethyl
sulfoxide, ethanol) and competition kinetics, we provide com-
pelling evidence that the hydrogen-abstracting agent is hydroxyl
radicals under inert atmosphere, coupled to a more selective
oxidant (most likely adamantyloxyl radicals) in the presence
of dioxygen.
amount of alcohol generated is comparable to that of ketone.
However, the alternative radical proposition18 of pyridine-
facilitated 1,2-hydrogen shift19 (Scheme 1) merits further
consideration.
Among other novel features, an important aspect of the
present systematic exploration is the employment of mainstream
Gif reagents and substrates to unravel product profiles which
frequently are in disagreement with experimental data on the
basis of which mechanistic interpretations have been put forward
by Barton and co-workers.2,3 Thus, this work substantiates
Perkins’ “radical reappraisal”22 of Gif chemistry by means of
expanding and amending the experimental record, as well as
by employing a handful of new diagnostic substrates to directly
interrogate the nature of the active oxidant(s) involved in H2O2-
dependent Gif chemistry. The notion that specialized substrates
such as Newcomb’s radical clocks,23 should be treated as
exceptional instances, can no longer be maintained, as generation
of carbon-centered radicals is shown here to pervade the most
common reagents and substrates used in Gif chemistry.
The propensity of t-BuOOH supported oxygenation systems20
to generate t-BuO•/t-BuOO• radicals even in cases in which high-
valent iron-oxo units are otherwise regarded to play a prominent
role (for instance, in shunt pathways of P-45021), raises doubt
whether these systems can be used as representative indicators
of mechanism for other mainstream Gif reagents. Mechanistic
insights regarding Gif systems employing H2O2 or O2/Zn have
been put forward by M. J. Perkins22 in a lucid reinterpretation
(vide infra) of Barton’s results, hinting toward an oxygen- and
carbon-centered radical mechanism for the totality of Gif
chemistry. However, unequivocal detection and quantification
of the aforementioned radicals in these systems is currently
resting on insufficient experimental evidence. A notable excep-
tion is an elegant study by M. Newcomb and co-workers,23 in
which an ultrafast radical-clock probe has been used to
distinguish between pathways adopted by biological monooxy-
genases (P-450,24 sMMO25) and a Gif-type reagent (FeCl3 in
py/AcOH), as indicated by the generation of short-lived (<100
fs) and diffusively free alkyl radicals, respectively. These results
have been interpreted by Barton as exceptional cases of alkyl
radical formation via homolytic FeV-R cleavage, similar to
those obtained for the tertiary positions of the strained hydro-
carbons adamantane2 and decalin.6a
In the present article, we report that two well-characterized
iron picolinate precursor species, ferrous [Fe(Pic)2(py)2] and
ferric [Fe(Pic)3]‚0.5py, frequently used as mediators of “oxy-
genated Fenton”4b and Gif chemistry3 respectively, enable
oxygenation of adamantane by H2O2 that demonstrates formation
of not only tertiary adamantyl radicals, as previously recog-
nized,2 but also of secondary adamantyl radicals. The alleged
absence of the latter has been central to Barton’s argument1 in
support of a non-radical mechanism. Other Gif-type reagents
([Fe2O(Pic)4(py)2], FeCl3, [Fe2(µ-OH)2(Pic)4]) also perform
adamantane oxidations in which sec-adamantyl radicals are
Experimental Section
Preparation of Compounds. All operations were performed under
a pure dinitrogen or argon atmosphere, using Schlenk techniques on
an inert gas/vacuum manifold or in a drybox (O2, H2O <1 ppm).
Hexane, petroleum ether, and toluene were distilled over Na, and THF
and diethyl ether, over Na/Ph2CO. Acetonitrile and methylene chloride
were distilled over CaH2. Ethanol and methanol were distilled over
the corresponding magnesium alkoxide, and acetone, over drierite.
Anhydrous pyridine, dimethyl sulfoxide and dimethyl formamide (water
< 0.005%) and double-distilled acetic acid (metallic impurities in parts-
per-trillion) were purchased from Aldrich. Deuterated solvents for NMR
experiments were purchased from Cambridge Isotope Laboratory. All
solvents, with the exception of methanol and water were degassed by
three freeze-pump-thaw cycles. Methanol and water were degassed
by bubbling nitrogen or argon for 0.5 h. Starting materials were
purchased from Aldrich and are of the highest available purities.
[Fe(Pic)2]n (1). A mixture of Fe powder (99.99%, 160 mg, 2.86
mmol) and picolinic acid (665 mg, 5.40 mmol) was stirred in CH2Cl2
(50 mL) for 4 days under nitrogen in an inert-atmosphere drybox. The
red-orange solution was then filtered from any remaining Fe, and excess
diethyl ether was added to the filtrate to precipitate the red powder of
1 (540 mg, 63%). 1H NMR (CD2Cl2): δ 35.62, 47.08, 58.93. IR
(KBr): νOCO 1653, 1594, 1576 cm-1. Anal. Calcd for C12H8N2Fe1O4:
C, 48.04; H, 2.69; N, 9.34. Found: C, 48.24; H, 2.71; N, 9.42.
[Fe(Pic)2(py)2] (2). A mixture of Fe powder (99.99%, 160 mg, 2.86
mmol) and picolinic acid (665 mg, 5.40 mmol) was suspended in
pyridine (50 mL). The reaction mixture was stirred for 4 days in a
nitrogen-atmosphere glovebox. Unreacted Fe powder was filtered off
and diethyl ether was slowly diffused into the filtrate to afford a brick
red microcrystalline solid (650 mg, 50%). Efforts to obtain X-ray quality
(18) We thank a reviewer for proposing this alternative radical pathway.
(19) 1,2-Hydrogen migrations are more common in aqueous solutions,
see: Beckwith, A. L. J.; Ingold, K. U. In Rearrangements in Ground
and Excited States; de Mayo, P., Ed.; Academic Press: New York, 1980;
Vol. 1, pp 161-310.
(20) (a) Nguyen, C.; Guajardo, R. J.; Mascharak, P. K. Inorg. Chem.
1996, 35, 6273-6281. (b) Rabion, A.; Chen, S.; Wang, J.; Buchanan, R.
M.; Seris, J.-L.; Fish, R. H. J. Am. Chem. Soc. 1995, 117, 12356-12357.
(21) (a) Meunier, B. Chem. ReV. 1992, 92, 1411-1456. (b) Mansuy,
D.; Bartoli, J.-F.; Momenteau, M. Tetrahedron Lett. 1982, 23, 2781-2784.
(c) Mansuy, D.; Bartoli, J.-F.; Chottard, J.-C.; Lange, M. Angew. Chem.,
Int. Ed. Engl. 1980, 19, 909-910. (d) Ledon, H. C. R. Hebd. Seances
Acad. Sci., Ser. C 1979, 288, 29-31. (e) Baccouche, M.; Ernst, J.; Fuhrhop,
J.-H.; Schlo¨zer, R.; Arzoumanian, H. J. Chem. Soc., Chem. Commun.
1977, 821-822.
(22) Perkins, M. J. Chem. Soc. ReV. 1996, 229-236.
(23) Newcomb, M.; Simakov, P. A.; Park, S.-U. Tetrahedron Lett. 1996,
37, 819-822.
(24) (a) Newcomb, M.; Le Tadic-Biadatti, M.-H.; Chestney, D. L.;
Roberts, E. S.; Hollenberg, P. F. J. Am. Chem. Soc. 1995, 117, 12085-
12091. (b) Atkinson, J. K.; Hollenberg, P. F.; Ingold, K. U.; Johnson, C.
C.; Le Tadic, M.-H.; Newcomb, M.; Putt, D. A. Biochemistry 1994, 33,
10630-10637.
(25) (a) Choi, S.-Y., Eaton, P. E.; Hollenberg, P. F.; Liu, K. E.; Lippard,
S. J.; Newcomb, M.; Putt, D. A.; Upadhyaya, S. P.; Xiong, Y. J. Am. Chem.
Soc. 1996, 118, 6547-6555. (b) Liu, K. E.; Johnson, C. C.; Newcomb,
M.; Lippard, S. J. J. Am. Chem. Soc. 1993, 115, 939-947.
1
crystals of this compound were unsuccessful. H NMR (CD2Cl2): δ
7.58-8.69 (m), 18.77 (br, 2H, 2H-py), 35.37 (br, 1H, 4H-Pic), 46.69
(br, 1H, 3H-Pic), 58.67 (br, 1H, 5H-Pic), 138.53 (br, 6H-Pic). IR
(KBr): νOCO 1645, 1594, 1348 cm-1. UV-vis (py): λmax ) 428 (ꢀM
)
2800) nm. Anal. Calcd for C22H18N4Fe1O4: C, 57.56; H, 3.96; N, 12.23.
Found: C, 57.60; H, 3.98; N, 12.39.
[Fe(Pic)(py)3Cl] (3). Recrystallization of 2 from a mixture of
pyridine and methylene chloride (1:1 v/v) affords red crystals of 3 and
white needle-shaped crystals of 1,1′-methylene-bispyridinium dichloride
([C5H5N-CH2-NC5H5]Cl2). The latter compound is also obtained from
a mixture of py/CH2Cl2 (1:1 v/v) in the absence of metal. IR(KBr):
νOCO 1634, 1601, 1530, 1483, 1442. Anal. Calcd for C21H19N4Cl1-
Fe1O2: C, 55.96; H, 4.25; N, 12.43. Found: C, 55.60; H, 4.32; N,
12.52.
[Fe2(Pic)4(4-tert-Bupy)2] (4). Compound 1 (100 mg, 0.33 mmol)
was dissolved in 4-tert-butylpyridine (10.0 mL) to afford a deep red

7506 J. Am. Chem. Soc., Vol. 122, No. 31, 2000
Kiani et al.
solution. Diethyl ether was slowly diffused to the filtrate of the solution
to yield carmine red microcrystals of 4 (123 mg, 86%). 1H NMR
(CD2Cl2): δ 1.08 (s, 9H, C(CH3)3), 6.2-8.4 (m), 11.82 (br, 2H, 2H-
py) 25.60 (br, 2H, 4H-Pic), 45.56 (br, 2H, 3H-Pic), 60.41 (br, 2H, 5H-
using a syringe-pump over a period of 6 h upon vigorous stirring. The
oxidations were performed under inert atmosphere (N2 or Ar) and/or
under various partial pressures of dioxygen (O2 (4%) in N2, air, O2
(100%)). Other variables included addition of picolinic acid (0.2 or
16.0 mmol) and/or Zn (20.0 mmol) to the reaction mixture. Occasion-
ally, acetic acid was omitted from the solvent matrix. When TEMPO
was used in experiments directed toward trapping of adamantyl radicals,
the exact procedure noted above was followed with the addition of
TEMPO (2.0 or 5.0 mmol, as specified).
Catalytic oxidations of DMSO and EtOH were conducted under the
general conditions noted in oxidations of adamantane, employing
amounts specified in Table 4. In the case of oxidations in the presence
of 1-iodoadamantane (1.31 g, 5.0 mmol), the reaction and follow-up
procedure were protected from light exposure but were otherwise
performed as indicated for the other substrates, using amounts of
materials as specified in the text.
Pic), 118.48 (br, 6H-Pic). IR (KBr): νOCO 1640, 1610, 1550, 1490 cm-1
.
Anal. Calcd. for C42H42N6Fe2O8: C, 57.94; H, 4.86; N, 9.65. Found:
C, 57.48; H, 4.79; N, 9.95.
[Fe2(Pic)4(DMF)2] (5). Quantitative amounts of this compound were
obtained by dissolving 1 in dimethylformamide (DMF), or better 2 (200
mg, 0.44 mmol) in 30.0 mL of DMF. Red crystals of 5 (137 mg, 83%)
suitable for X-ray analysis were grown by diffusion of diethyl ether
into DMF at 0 °C. 1H NMR ((CD3)2NCDO): δ 36.2 (br, 1H, 4H-Pic),
46.5 (br, 1H, 3H-Pic), 59.9 (br, 1H, 5H-Pic), 137.5 (br, 6H-Pic). IR
(KBr): νOCO 1642, 1575, 1320, 1270 cm-1. UV-vis (DMF): λmax
)
462 (ꢀM ) 1120) nm. Anal. Calcd for C30H30N6Fe2O10: C, 48.28; H,
4.05; N, 11.26. Found: C, 48.36; H, 4.12; N, 11.38.
[Fe(Pic)2(MeOH)2] (6). Solid [Fe(Pic)2]n (1, 200 mg, 0.67 mmol)
was dissolved in freshly degassed methanol (15.0 mL). The solution
was filtered and allowed to stand at room temperature over a period of
one week, upon which large red-plate crystals (180 mg, 74%) suitable
for X-ray analysis were obtained. 1H NMR (CD2Cl2): δ 3.39 (br, 1H,
MeOH), 3.75 (br, 3H, CH3OH), 23.94 (br, 1H, 4H-Pic), 45.57 (br, 1H,
3H-Pic), 62.20 (br, 1H, 5H-Pic). IR (KBr): νOCO 1624, 1589, 1564,
1444 cm-1. Anal. Calcd for C14H16N2Fe1O6: C, 46.18; H, 4.43; N, 7.69.
Found: C, 45.96; H, 4.56; N, 7.70.
[Fe2O(Pic)4(py)2] (8). This compound was generated in solution by
dissolving 9 (80 mg, 0.11 mmol) in pyridine (20.0 mL). Careful layering
of diethyl ether onto pyridine affords light brown solid of analytically
pure 8 (72 mg, 84%) upon standing at ambient temperature.1H
NMR(CD2Cl2): δ 7.71 (br), 9.07 (br), 111.30 (br), 125.22 (br). IR
(KBr): νOCO 1654, 1569, 1351, 1475, νFeOFe 856 cm-1. UV-vis (py):
λmax ) 350 (ꢀM ) 8000) nm. Anal. Calcd for C34H26N6Fe2O9: C, 52.75;
H, 3.38; N, 10.85. Found: C, 52.95; H, 3.38; N, 10.84.
The workup procedure for identification and quantification of
products of adamantane oxidation was performed as follows. Excess
oxalic acid (5 equiv over iron) and triphenylphosphine (2 equiv over
H2O2) were occasionally added at the end of the reaction to ensure
complexation of iron to inert iron-oxalato species, and reduction of
any remaining H2O2 and adamantylhydroperoxides to water and
corresponding alcohols, respectively. The internal standard (hexa-
methylbenzene or 1,3,5-trisisopropylbenzene) was added to the reaction
mixture, and an aliquot (2 mL) was withdrawn for GC or GC-MS
analysis. After addition of water (3 mL), the aliquot was extracted with
diethyl ether (3 × 5 mL). The combined diethyl ether layers were dried
over magnesium sulfate. The samples were analyzed by a Hewlett-
Packard 5890 Series II Capillary GC, employing Supelco SPB-1 or
SPB-50 capillary columns (30 m (length), 0.32 mm (i.d.), 0.25 µm
(df)), a flame ionization detector, and a Hewlett-Packard 3395 integrator.
The following was a typical temperature program for the SPB-1
column: initial temperature ) 50 °C; hold temperature for 9 min;
increase temperature by 3 °C/min up to 135 °C, thereafter by 5 °C/
min up to the final temperature 260 °C. Products were identified
by their respective retention times versus authentic samples and by
mass determination on a Finnigan MAT-90 MS coupled to a Varian
3400 GC. Adamantane and authentic oxygenated adamantane products
(1-Ad-ol, 2-Ad-ol, 2-Ad-one) as well as 1-chloroadamantane were
purchased from Aldrich. 2-Chloroadamantane was a gift from Barton’s
group. Authentic tert-adamantylpyridines (2-(1-Ad)-py, 4-(1-Ad)-py),26
sec-adamantylpyridines (2-(2-Ad)-py, 4-(2-Ad)-py),26 1-Ad-TEMPO,27
and 2-Ad-TEMPO27 were prepared by photolysis of suitable adamantyl-
radical generating precursors in the presence of protonated pyridine or
TEMPO according to literature procedures.
The follow-up analytical procedure for quantification of the pyridine-
trapped alkyl radicals, generated in oxidations of DMSO and EtOH, is
similar to the one noted above, with the exception that the reaction
aliquot was rendered alkaline (NaOH 20% w/v) prior to extraction
by diethyl ether. Furthermore, a Hewlett-Packard Carbowax column
(30 m (length), 0.50 mm (i.d.), 0.25 µm (df)) was employed under a
temperature program identical to that noted above, with the exception
that the final temperature was set at 220 °C. Authentic 2-, 3-,
4-picolines, 2-(2-CH2CH2OH)py, and 4-(2-CH2CH2OH)py were pur-
chased from Aldrich. (()-R-Methyl-4-pyridinemethanol was purchased
from Fluka. A specimen of the corresponding (()-R-methyl-2-
pyridinemethanol was unavailable, and this important compound was
quantified using the isomer 2-(2-CH2CH2OH)py as equivalent authentic
sample.
[Fe2O(Pic)4(DMF)2] (9). Solid 2 (100 mg, 0.22 mmol) was dissolved
in 20.0 mL of DMF. This solution was then exposed to dry oxygen for
30 min at room temperature. The color of the solution changed from
dark red to brown. X-ray quality crystals of 9 (68 mg, 82%) were
1
obtained by diffusing diethyl ether into the DMF solution at 0 °C. H
NMR (CD2Cl2): δ 2.75 (br), 2.89 (br), 7.95 (br), 111.01 (br), 123.30
(br). IR (KBr): νOCO 1653, 1344, 1288, νFeOFe 867 cm-1. UV-vis
(DMF): λmax ) 350 (ꢀM ) 7500) nm. Anal. Calcd for C30H30N6-
Fe2O11: C, 47.27; H, 3.97; N, 11.02. Found: C, 47.39; H, 3.97; N,
11.08.
[Fe(Pic)3]‚0.5py (10). A mixture of picolinic acid (6.15 g, 0.05 mol)
and sodium hydroxide (2.0 g, 0.05 mol) in 10.0 mL of water was slowly
added to a slurry of Fe2(SO4)3 (3.3 g, 0.008 mol) in 100 mL of
acetonitrile. The reaction mixture was stirred for 2 h, followed by
filtration and extraction of the precipitate with 3 × 60 mL of acetonitrile.
The combined organic layers were dried over MgSO4. Removal of
solvent under vacuum afforded 2.8 g (40%) of crude [Fe(Pic)3]. Yellow
crystals of the pyridine-solvated adduct 10, suitable for X-ray analysis,
were grown by diffusing diethyl ether into a pyridine solution of
the crude product (see text for other preparation methods).1H NMR
(CD2Cl2): δ 7.4 (br), 7.7 (br), 8.7 (br), 109.72 (br), 123.67 (br). IR
(KBr): νOCO 1676, 1604, 1576, 1326, 1287 cm-1. UV-vis (py): λmax
) 341 (2580) nm. Anal. Calcd for C20.5H14.5N3.5Fe1O6: C, 53.27; H,
3.16; N, 10.60. Found: C, 53.11; H, 3.09; N, 10.47.
[Fe2(µ-OMe)2(Pic)4] (11). Compound [Fe2(µ-OH)2(Pic)4] (100 mg,
0.16 mmol) was dissolved in methanol (20.0 mL) upon mild heating.
Yellow-green plate-shaped crystals of 11 (85 mg, 81%) were obtained
by slow crystallization from methanol at room temperature. IR (KBr):
Control Experiments Involving Authentic Adamantyl Radicals.
Barton’s esters of N-hydroxypyridine-2-thione26 with 1-adamantane-
carboxylic acid and 2-adamantanecarboxylic acid28 (0.115 mmol of
each) were dissolved in a mixture of pyridine (10.0 mL) and acetic
acid (1.0 mL). The iron reagents [Fe(Pic)2(py)2] (2) (4.2 mg, 9.2 µmol)
or [Fe(Pic)3]‚0.5py (10) (4.3 mg, 9.2 µmol) were added, and the mixture
was subjected to photolysis for 2 h by means of an ACE-Hanovia high-
νOCO 1672, 1604, 1473, 1331, 1286 cm-1. UV-vis (MeOH): λmax
)
251 (ꢀM ) 1400) nm. Anal. Calcd for C26H22N4Fe2O10: C, 47.16; H,
3.35; N, 8.46. Found: C, 47.43; H, 3.28; N, 8.57.
Catalytic Oxidations. A typical oxidation of substrate (adamantane)
mediated by various iron-containing species (2, 8, 10, FeCl3, [Fe2(µ-
OH)2(Pic)4]) was conducted as follows. In a 50.0 mL round-bottom
flask, the iron reagent (0.20 mmol) was dissolved under anaerobic
conditions in 30.0 mL of pyridine and 3.0 mL of acetic acid followed
by addition of adamantane (681 mg, 5.0 mmol). Degassed aqueous
solution (30%) of hydrogen peroxide (0.24 mL, 2.0 mmol) was added
(26) Barton, D. H. R.; Halley, F.; Ozbalik, N.; Schmitt, M.; Young, E.;
Balavoine, G. J. Am. Chem. Soc. 1989, 111, 7144-7149.
(27) Barton, D. H. R.; Be´vie`re, S. D.; Chavasiri, W.; Doller, D.; Liu,
W.-G.; Reibenspies, J. H. New J. Chem. 1992, 16, 1019-1029.
(28) Farcasius, D. Synthesis 1972, 615-616.

Generation of Radicals in Gif-Type Oxidations
J. Am. Chem. Soc., Vol. 122, No. 31, 2000 7507
pressure mercury-vapor lamp, while the solution was continuously
purged by O2 (4%) in N2. The same procedure was employed in those
experiments in which TEMPO (0.500 mmol) was added, with the
exception that 0.250 mmol of each ester was used and pure O2 was
bubbled through the solution under photolysis. In all cases, the product
profile was analyzed by GC and GC-MS as indicated above.
All non-hydrogen atoms are refined anisotropically. Hydrogens were
located by difference Fourier maps and refined isotropically. For 5, 7,
and 11 hydrogens were calculated by geometrical methods and refined
as a riding model. Water-molecule hydrogens in 7 were placed to
accommodate hydrogen bonding. The solvent molecule of pyridine in
10 was modeled for disorder over the 2-fold axis. The crystals used
for the diffraction studies showed no decomposition during data
collection. All ORTEP drawings deposited as Supporting Information
are done at 50% probability ellipsoids.
Kinetic Isotope Effect (KIE) Measurements. The intermolecular
deuterium kinetic isotope effect was measured for adamantane by con-
ducting catalytic oxidations on a mixture of adamantane/adamantane-
d16 (1:1). The following was a typical procedure. The iron reagent,
[Fe(Pic)2(py)2] (2) (9.2 mg, 0.02 mmol) or [Fe(Pic)3]‚0.5py (10) (9.2
mg, 0.02 mmol), and adamantane (0.25 mmol)/adamantane-d16 (0.25
mmol) were dissolved in a mixture of pyridine (3.0 mL) and acetic
acid (0.3 mL). Aqueous H2O2 (30%, 30 µL, 0.2 mmol) was added to
this stirred solution via a syringe pump over a period of 6 h under
argon (delivered to the headspace of the flask under static conditions).
Throughout the experiment, the mixture was maintained at 25.0(1) °C
with the aid of a Brinkman RM6 thermostat. The workup procedure
was performed in a manner similar to that described for adamantane
oxidations. KIE values were obtained for each product from the ratio
of the corresponding protio/deuterio product yield (average of three
trials), evaluated by GC separation of all products and GC-MS
(selective-ion mode) detection and calculation of the relative intensities
of the respective molecular ions selected over the fully resolved GC
peaks. Corrections for the exact adamantane/adamantane-d16 ratio used,
as determined by GC-MS, were applied.
The intramolecular deuterium KIE was determined on adamantane-
1,3-d2 (deuterium enrichment: 97.7(1) %), synthesized from 1,3-
dibromoadamantane according to a literature procedure.29 The reaction
(three trials) was performed under Ar and O2 (4%) in N2 in a manner
similar to that described for the evaluation of the intermolecular KIE.
KIE values for 1-adamantanol were calculated from the ratio of
1-adamantanol-d2 over 1-adamantanol-d1 determined by GC-MS
evaluation of the relative areas corresponding to the respective molecular
ions selected over the entire GC peak (unresolved for the two
1-adamantanols). Corrections were applied to the mass peak intensity
of 1-adamantanol-d2 (154) due to contribution from the masses of
naturally abundant 13C and 2H isotopes in 1-adamantanol-d1 (153). An
isotopic purity correction for the incompletely deuterated substrate (that
is, due to contamination of adamantane-1,3-d2 by adamantane-1-d1) was
implemented following a published analysis.29
1
Other Physical Measurements. H NMR spectra were recorded
on JEOL GSX-270 and Varian XL-400 NMR spectrometers. The
isotropically shifted peaks for the iron-containing compounds were
assigned when possible based on chemical shift, integration, and
selective deuteration. FT-IR spectra were obtained on a Perkin-Elmer
1800 spectrometer. UV-vis spectra were obtained on a Hewlett-Packard
8452A diode array spectrometer, equipped with an Oxford DN1704
cryostat and ITC4 temperature controller for low-temperature measure-
ments. Electrospray ionization (ESI) mass spectra were recorded using
a Platform II MS (Micromass Instruments, Danvers, MA). Samples
were introduced from solutions specified in the text at a flow rate of
5 µL/min from a syringe pump (Harvard Apparatus). The electrospray
probe capillary was maintained at a potential of 3.0 kV, and the orifice
to skimmer potential (“cone voltage”) was varied from 15 to 30 V.
Spectra were collected in the multichannel acquisition mode. EI and
CI mass spectra were obtained on a Finnigan MAT-90 mass spectrom-
eter. Fast atom bombardment (FAB) mass spectra were measured on a
Kratos MS instrument. X-band ESR spectra were recorded on a Varian
E9 spectrometer located at the Chemistry Department of Harvard
University, employing an Oxford ESR-10 liquid-helium cryostat.
Microanalyses were done by H. Kolbe, Mikroanalytisches Laborato-
rium, Mu¨lheim an der Ruhr, Germany, and by Quantitative Technolo-
gies Inc., Whitehouse, NJ.
Results and Discussion
Synthesis and Characterization of Ferrous Picolinate
Species. We initiate our investigation by summarizing the
pertinent features and interconversion chemistry of ferrous and
ferric (vide infra) picolinate species employed as oxidation
reagents in the present study. Some of these precursors have
been previously used in Gif and “oxygenated Fenton” chemistry,
but have remained ill-defined. Stoichiometric transformations
of the appropriate ferrous picolinate sites in pyridine suggest
that these precursors are not suitable to mediate Gif oxygenations
supported by O2/Zn. On the other hand, ferric picolinate species
are readily transformed to their ferrous congeners upon exposure
to a small excess of H2O2. Thus, the oxygenation chemistry
detailed in subsequent sections is most likely associated with
oxidants produced by FeII/H2O2 combinations, irrespective of
whether ferrous or ferric species are used as catalyst precursors.
Scheme 2 summarizes the structural features and intercon-
version chemistry of key iron picolinate species under a variety
of conditions. Detailed structural evidence has been deposited
as Supporting Information. [FeII(Pic)2]n (1) precipitates as a red
powder upon stirring metallic Fe and PicH (2 equiv) in CH2Cl2
under a dinitrogen atmosphere. Polymeric 1 dissociates to
discrete mononuclear and/or dinuclear species in a variety of
solvents. In pyridine, Gif’s special solvent, 1 dissolves to afford
orange-red crystals of analytically pure [FeII(Pic)2(py)2] (2).
Compound 2, which has not been amenable to X-ray analysis,
is assigned as a mononuclear structure on the basis of micro-
X-ray Structure Determinations. Crystallographic data for com-
pounds 3, 5-7, and 9-11 for which structures were determined are
included only as Supporting Information. Single crystals were picked
from the crystallization vessel (coated with Paratone-N oil if necessary
due to air-sensitivity or desolvation), mounted on a glass fiber using
grease, and transferred to a Siemens (Bruker) SMART CCD (charge
coupled device) based diffractometer equipped with an LT-2 low-
temperature apparatus operating at 213 K. Data were measured using
omega scans of 0.3° per frame for 30 s (10 s for 5; 60 s for 6 and 11),
such that a hemisphere was collected. A total of 1271 frames (1400
for 9) were collected with a maximum resolution of 0.75 Å (0.85 Å
for 5, 9 and 10). The first 50 frames were recollected at the end of
data collection to monitor for decay. Cell parameters were retrieved
using SMART software30 and refined using the SAINT software31 which
corrects for Lp and decay. Absorption corrections were applied using
SADABS32 supplied by George Sheldrick. The structures are solved
by the direct method using the SHELXS-9733 program and refined by
least-squares method on F2, SHELXL-97,34 incorporated in SHELXTL-
PC V 5.10.35
The structures were solved in the space groups specified in tables
deposited as Supporting Information by analysis of systematic absences.
(29) Sorokin, A.; Robert, A.; Meunier, B. J. Am. Chem. Soc. 1993, 115,
7293-7299.
(30) SMART V 5.050 (NT) Software for the CCD Detector System;
Bruker Analytical X-ray Systems: Madison, WI, 1998.
(31) SAINT V 5.01 (NT) Software for the CCD Detector System; Bruker
Analytical X-ray Systems: Madison, WI, 1998.
(32) SADABS Program for Absorption Corrections Using Siemens CCD
Based on the Method of Robert Blessing; Blessing, R. H. Acta Crystallogr.
1995, A51, 33-38.
(33) Sheldrick, G. M. SHELXS-97 Program for the Solution of Crystal
Structure; University of Go¨ttingen: Go¨ttingen, Germany, 1997.
(34) Sheldrick, G. M. SHELXL-97 Program for the Refinement of Crystal
Structure; University of Go¨ttingen: Go¨ttingen, Germany, 1997.
(35) SHELXTL 5.10 (PC-Version) Program Library for Structure
Solution and Molecular Graphics; Bruker Analytical X-ray Systems:
Madison, WI, 1998.

7508 J. Am. Chem. Soc., Vol. 122, No. 31, 2000
Kiani et al.
Scheme 2
analytical and positive-ion electrospray MS data obtained from
that the mononuclear structure persists in solution. A similar
structure has been reported36 for the water adduct [FeII(Pic)2-
(H2O)2]‚2H2O, which in the present study is obtained by
dissolving polymeric 1 in water, but also via all other Fe(II)
complexes noted above. Interestingly, a light-blue compound
of stoichiometry [FeII(Pic)2]‚2H2O (7) has been isolated in low
yields among the decomposition products of the reaction of
[Fe(Pic)3]‚0.5py (10) with H2O2 (vide infra), and recrystallized
from wet methanolic solutions. The structure of 7 (see
Supporting Information) reveals layers composed of discrete
planar centrosymmetric [Fe(Pic)2] units featuring weak inter-
molecular interactions (Fe-O ) 2.734 Å) via adjacent carbonyl
moieties along the axial coordination vector. Surprisingly,
compound 7, dissolved in methanol or water, retains its light
blue color and does not convert to 6 or [FeII(Pic)2(H2O)2]. It is
also worthwhile noting that the solvated ferrous species 2-6
interconvert by virtue of facile solvent exchange.
1
pyridine solutions of 2 (m/z ) 459 [M + H]+). H NMR data
in CD2Cl2 and pyridine-d5 suggest that the solid-state structure
is retained in solutions. In an attempt to crystallize 2 from a
mixture of pyridine and CH2Cl2 (1:1 v/v) red crystals of
[FeII(Pic)(py)3Cl] (3) were deposited (see Supporting Informa-
tion for solid-state structure) along with colorless crystals of
[C5H5N-CH2-NC5H5]Cl2. The latter compound, which re-
markably precipitates even from mixtures of routinely purified
pyridine and CH2Cl2 (1:1 v/v) alone, is presumably the source
of chloride found in 3.
Polymeric 1 dissociates in 4-tert-butylpyridine to afford
carmine-red [FeII (Pic)4(4-t-Bupy)2] (4), albeit in microcrystals
2
not suitable for X-ray analysis. Based on microanalytical data,
the compound is formulated as a dimeric structure similar to
that encountered in the case of the DMF adduct 5 (see below).
1H NMR spectra in CD2Cl2 support this assignment and extend
it to solution structures by providing an integrated ratio of Pic/
4-t-Bupy equal to 2:1. In DMF, compound 1 dissolves to yield
Synthesis and Characterization of Ferric Picolinate Spe-
cies. Compound [FeII(Pic)2(py)2] (2) is moderately stable upon
exposure to dioxygen. This behavior may explain why catalytic
oxygenation of hydrocarbons mediated by Fe(II)-picolinate
species in pyridine or py/AcOH would not operate under O2/
Zn but require H2O2 instead. Compound [FeIII2O(Pic)4(py)2] (8)
can be synthesized indirectly upon dissolving [FeIII2O(Pic)4-
(DMF)2] (9) in pyridine. Unfortunately, we have been unable
to obtain crystals of light-brown 8 suitable for X-ray analysis.
Microanalytical, IR (νFe-O-Fe(asym) ) 856 cm-1), 1H NMR, and
ESI-MS data (m/z ) 696 [M + H - py]+) strongly support the
assignment of 8 as a diferric µ-oxo structure similar to that
encountered in the solid-state structure of brown-red 9 (see
Supporting Information). Generation of 937 is possible via
compound 2, which, when dissolved in DMF, transforms to 5
and becomes sensitive upon exposure to dioxygen, most likely
due to facile replacement of DMF. However, stoichiometric
red crystals of [FeII (Pic)4(DMF)2] (5). The same compound can
2
also be obtained in quantitative yields by dissolving 2 in DMF.
The solid-state structure of 5 (see Supporting Information)
consists of a dimeric molecule featuring a pair of ferrous sites
bridged via two carboxylato oxygens derived from two Pic
moieties. The dinuclear structure of 5 is apparently retained
in dimethylformamide, as indicated by ESI-MS data (m/z )
746 [M]+).
Hints to the structure of the pyridine adduct 2 are offered by
the geometric features of [FeII(Pic)2(MeOH)2] (6) and the known
compound [FeII(Pic)2(H2O)2]‚2H2O.36 The methanol adduct 6,
which has been crystallographically investigated (see Supporting
Information), is obtained as orange crystals from solutions of 1
or 2 in MeOH. The 1H NMR spectrum of 6 in CD2Cl2 suggests
(36) Shova, S. G.; Brashovyanu, N. V.; Turte, K. I.; Mazus, M. D. Russ.
J. Coord. Chem. 1996, 22, 438-443.

Generation of Radicals in Gif-Type Oxidations
J. Am. Chem. Soc., Vol. 122, No. 31, 2000 7509
Table 1. Product Profile of Adamantane Oxidation by H2O2
Mediated by [Fe(Pic)2(py)2] (2)
amounts of pyridine in solutions of 5 in DMF are necessary to
achieve formation of 9, otherwise the oxo bridge of 9 is not
retained and only [Fe(Pic)3]‚DMF is obtained. Further verifica-
tion of the structure of 9 is provided by the detection of the
νFe-O-Fe(asym) stretching band in the solid state (867 cm-1) in
accord with values reported38 for similar µ-oxo diferric com-
pounds. 1H NMR spectra of 9 in CD2Cl2 and ESI-MS data from
DMF solutions (m/z ) 690 [M + H - DMF]+) indicate that
this structure is also retained in solution.
In the presence of protic solvents such as excess water or
CH3COOH, the oxo bridge of 8 has also proven to be fragile,
leading to generation of [Fe(Pic)3]‚0.5py (10) in pyridine.
Importantly, exposure of 10 to 2 equiv of H2O2 in pyridine
generates purple colored solutions at low temperatures (<-20
°C) which decompose rapidly at higher temperatures (>5 °C)
to reform, upon evolution of O2, the reduced [FeII(Pic)2(py)2]
(2) and trace amounts of [FeII(Pic)2]‚2H2O (7). The nature of
the intermediate purple solution is currently under consideration.
Finally, the ferrous species [FeII(Pic)2(MeOH)2] (6) and
[FeII(Pic)2(H2O)2]‚2H2O36 are oxidized by dioxygen to yield
[FeIII2(µ-OMe)2(Pic)4] (11) and [FeIII2(µ-OH)2(Pic)4]39 in metha-
nol and water, respectively. The pale-yellow hydroxo-bridged
ferric compound has been previously reported39 and proven
difficult to crystallize, as it precipitates rapidly from water and
transforms to other species in a variety of solvents. In pyridine,
[FeIII2(µ-OH)2(Pic)4] reverts to [FeIII2O(Pic)4(py)2] (8), while in
py/AcOH affords [Fe(Pic)3]‚0.5py (10). The slightly more acidic
methanol converts the hydroxy-bridged [FeIII2(µ-OH)2(Pic)4] to
the corresponding methoxy-bridged species 11 (see Supporting
Information for solid-state structure).
Catalytic Oxidation of Adamantane Mediated by Fe(II)
Reagents. Table 1 summarizes product profiles for the oxy-
genation of adamantane (5 mmol) by H2O2 (2 mmol; added by
syringe pump in the course of the reaction) catalyzed by the
ferrous compound [FeII(Pic)2(py)2] (2) (0.2 mmol). The oxy-
genation reactions are performed in a solvent matrix composed
of pyridine (30.0 mL) and carboxylic acid(s), namely AcOH
(3.0 mL) and/or PicH (50 mg, 0.40 mmol, or 2 g, 16.3 mmol).
Occasionally, excess Zn (1.3 g, 20.0 mmol) is added to the
reaction mixture as a sacrificial reducing agent. The catalytic
reactions are conducted under inert atmosphere (N2, Ar) or
variable partial pressures of dioxygen (O2 (4%) in N2, air, O2
(100%)).
a General conditions, unless otherwise noted: 0.2 mmol 2; 5 mmol
AdH; 2 mmol H2O2 (added by syringe pump over 6 hours at RT);
py/AcOH (30/3 mL). nd ) not detected. tr ) trace amount. na ) not
applicable. b 20 mmol of Zn dust added. c 0.4 mmol of PicH added.
d 16.3 mmol of PicH added; AcOH omitted. e 0.4 mmol of PicH added;
AcOH omitted. f Py omitted; 30 mL of CH3CN added.
also sec-adamantylpyridines (2-(2-Ad)-py, 4-(2-Ad)-py) are
detected in comparable yields. These coupling products domi-
nate the product profile especially with regards to the activation
of the tert-adamantyl positions (the ratio of pyridine-coupled
versus oxo products is 34.6) and to a lesser extent of the sec-
adamantyl sites (the corresponding ratio is 0.8). Other important
mechanistic indicators are (i) the prominence of 2-one formation
over 2-ol (2-one/2-ol ≈ 3), which is not due to overoxidation
of 2-ol to 2-one, and thus precludes simple “oxenoid” insertion
into C-H bonds; and (ii) the overall selectivity for the activation
of tertiary versus secondary positions on a per hydrogen basis
(3°/2° ) 2.88), which suggests that the active oxidant involved
displays modest selectivity. Similarly low tert/sec selectivity
values are reported for H2O2-40 or O2/Zn-dependent41 oxidations
of adamantane mediated by other oxygenation systems. The
t-BuOOH supported systems42 provide a value 3°/2° ≈ 10 in
accord with radical hydrogen-abstraction from adamantane by
t-BuO•.3 These values need to be contrasted with values of 11-
48 for oxidations of adamantane by P-450/PhIO biomimetic
sytems,43 although a surprisingly low value (3°/2° ) 3) has been
reported44 for adamantane oxidation by sMMO.
Inspection of the product profile obtained from the oxygen-
ation of adamantane reveals activation of both tertiary and
secondary positions. The products detected are divided into two
categories: oxo products (1-adamantanol (1-ol), 2-adamantanol
(2-ol), adamantanone (2-one)) and adamantyl-pyridines (2-(1-
Ad)-py, 4-(1-Ad)-py, 2-(2-Ad)-py, 4-(2-Ad)-py). In addition,
products of pyridine oxidation, namely hydroxypyridines and
bipyridines, are obtained. These products have been previously
quantified in GifIV-type oxygenation chemistry (Fecat/O2/Zn);7
no attempts to account for products of pyridine oxidation have
been made with the present systems.
The detection of adamantylpyridines is prima facie indication
of the involvement of adamantyl radicals, since alkyl radicals
are known to attack protonated pyridine at positions 2- and 4-,
(40) (a) Fish, R. H.; Konings, M. S.; Oberhausen, K. J.; Fong, R. H.;
Yu, W. M.; Christou, G.; Vincent, J. B.; Coggin, D. K.; Buchanan, R. M.
Inorg. Chem. 1991, 30, 3002-3006. (b) Vincent, J. B.; Huffman, J. C.;
Christou, G.; Li, Q.; Nanny, M. A.; Hendrickson, D. N.; Fong, R. H.; Fish,
R. H. J. Am. Chem. Soc. 1988, 110, 6898-6900.
(41) (a) Kitajima, N.; Ito, M.; Fukui, H.; Moro-oka, Y. J. Chem. Soc.,
Chem. Commun. 1991, 102-104. (b) Kitajima, N.; Fukui, H.; Moro-oka,
Y. J. Chem. Soc., Chem. Commun. 1988, 485-486.
(42) (a) Kojima, T.; Leising, R. A.; Yan, S.; Que, L., Jr. J. Am. Chem.
Soc. 1993, 115, 11328-11335. (b) Me´nage, S.; Vincent, J. M.; Lambeaux,
C.; Chottard, G.; Grand, A.; Fontecave, M. Inorg. Chem. 1993, 32, 4766-
4773.
Entry 1 (Table 1) reveals the most important finding in these
series of adamantane oxidations. Under dinitrogen, not only tert-
adamantylpyridines (2-(1-Ad)-py, 4-(1-Ad)-py) are obtained but
(37) A compound assigned to [Fe2O(Pic)4] has been previously synthe-
sized from DMF solutions. The reported compound is most likely identical
to 9: (a) Sheu, C.; Richert, S. A.; Cofre´, P.; Ross, B., Jr.; Sobkowiak, A.;
Sawyer, D. T.; Kanofsky, J. R. J. Am. Chem. Soc. 1990, 112, 1936-1942.
(b) Cofre´, P.; Richert, S. A.; Sobkowiak, A.; Sawyer, D. T. Inorg. Chem.
1990, 29, 2645-2651.
(38) Kurtz, D. M., Jr. Chem. ReV. 1990, 90, 585-606.
(39) Schugar, H. J.; Rossman, G. R.; Gray, H. B. J. Am. Chem. Soc.
1969, 91, 4564-4566.
(43) Groves, J. T.; Nemo, T. E. J. Am. Chem. Soc. 1983, 105, 6243-
6248.
(44) Green, J.; Dalton, H. J. Biol. Chem. 1989, 264, 17698-17703.

7510 J. Am. Chem. Soc., Vol. 122, No. 31, 2000
Kiani et al.
Table 2. Product Profile of Adamantane Oxidation by H2O2
via pathways extensively studied by Minisci.45 This is further
reinforced by the results of oxidation experiments under
dioxygen (entry 2), which mark a substantial deviation from
the product distribution observed under N2 (entry 1). Indeed,
sec-adamantylpyridines have now disappeared in favor of 2-one
and to a lesser extent 2-ol, and tert-adamantylpyridines have
been diminished to low yields in favor of the major oxygenation
product 1-ol. Interestingly, the tert/sec selectivity is enhanced
(3°/2° ) 3.78), suggesting partial interference of a more selective
active oxidant. At low partial pressure of dioxygen (4% of O2
in N2 (entry 3)), an intermediate product profile, between those
observed in pure N2 and O2, is obtained. The conditions under
4% O2 provide good yields of all products for control experi-
ments purported to scrutinize whether the entire product
distribution is due to partitioning of tert- and sec-adamantyl
radicals between dioxygen and protonated pyridine (vide infra).
In the presence of excess zinc under N2, added with the intention
of rapidly reducing any Fe(III) formed to Fe(II) and thus
avoiding byproducts of Fe(III)/H2O2 dependent cycles (including
dioxygen), the product profile (entry 4) shifts almost exclusively
toward formation of tert- and sec-adamantylpyridines, as
expected, but surprisingly the tert/sec selectivity (3°/2° ) 4.57)
is one of the highest observed in the present study. Addition of
small amounts of picolinic acid (2 equiv over 2) to the reaction
mixture (py/AcOH) shows no ligand-enhancing effect on the
product profile of oxygenations conducted under argon (entry
5) or dioxygen (entry 6), with the exception that adamantyl-
pyridines are prominently represented under argon, probably
due to the effectiveness of argon in removing adventitious
dioxygen. Replacement of acetic acid by equimolar amounts
of picolinic acid diminishes the overall yield of products (entries
7 and 8) to almost half. In contrast, catalytic amounts of picolinic
acid (2 equiv over 2) in the absence of AcOH, are as effective
as AcOH (especially under O2 (entry 10)) in bringing about
oxidation of adamantane, although under argon (entry 9) the
amounts of adamantylpyridines are relatively suppressed in favor
of the oxo products (tert Ad-py/Ad(O) ) 6.4, sec Ad-py/
Ad(O) ) 0.16), most likely owing to the limited protonation of
pyridine.45 Finally, attempts to replace py/AcOH by CH3CN/
AcOH (10:1 v/v) yielded low amounts of oxo products (entries
11 and 12). The alcohol 1-ol is the major product followed by
2-ol, which exceeds the amount of 2-one observed. The
preponderance of 2-ol over 2-one may not necessarily signify
change in mechanism, as it could possibly reflect different
processing of sec-adamantyloxyl radicals in CH3CN/AcOH
(usual H-atom abstraction)17 versus py/AcOH (C- to O-atom
1,2-H migration (Scheme 1)).
Mediated by [Fe(Pic)3]‚0.5py (10), [Fe2O(Pic)4-(py)2] (8),
[Fe2(µ-OH)2(Pic)4], and FeCl3
a General conditions, unless otherwise noted: 0.2 mmol 10; 5 mmol
AdH; 2 mmol H2O2 (added by syringe pump over 6 hours at RT);
py/AcOH (30/3 mL). nd ) not detected. b 0.2 mmol of PicH added
c AcOH omitted. d 10 replaced by 8. e 10 replaced by [Fe2(µ-OH)2(Pic)4].
f 10 replaced by FeCl3. g 1-Cl-Ad. h 2-Cl-Ad.
observed in oxygenations mediated by [FeII(Pic)2(py)2] (2),
chief among which is the detection of both tert- and sec-
adamantylpyridines. However, by comparison to the corre-
sponding oxo products, the amount of adamantylpyridines is
suppressed (tert Ad-py/Ad(O) ) 6.55, sec Ad-py/Ad(O) )
0.48), probably due to internally produced dioxygen (and/or
superoxide). Under pure O2 or 4% O2 (in N2) (entries 2, 3), the
product profile shifts gradually in favor of the oxo products, in
a manner paralleling that observed with the Fe(II) reagent,
including the tendency for the tert/sec selectivity to increase
with increasing amounts of dioxygen (3°/2° ) 2.64 (N2), 3.61
(O2)). Addition of one more equivalent of PicH over iron in
the reaction mixture has no effect on product distribution (entries
4, 5). However, if this small excess of PicH is present in solution
in the absence of any AcOH, then, in analogy to results obtained
with [FeII(Pic)2(py)2] (2), the amount of adamantylpyridines is
relatively diminished in favor of oxo products (entries 6, 7).
Adamantane oxygenations mediated by [FeIII2O(Pic)4(py)2] (8)
show similar product profiles (entries 8, 9) to those obtained
with 10, but [FeIII2(µ-OH)2(Pic)4] (entries 10, 11) affords only
traces of sec-adamantylpyridines under Ar. It is unlikely that
the near absence of these products signifies any change in
mechanism. The low tert/sec selectivity (3°/2° ) 2.22) observed
in these experiments minimizes the possibility that internally
produced fluxes of O2 may have affected the product profile.
However, the unusually high amounts of 2-ol obtained relative
to 2-one suggest that FeIII-OH units may efficiently scavenge
sec-adamantyl radicals to generate 2-ol. A similar situation is
encountered with a reagent commonly used by Barton in
mechanistic studies,3 namely FeCl3 in py/AcOH. Indeed,
oxygenation of adamantane by this reagent does not afford
any detectable amounts of sec-adamantylpyridines under inert
atmosphere (entry 12). However, the presence of 2-chloro-
adamantane, formed at much higher quantities than 1-chloro-
adamantane, fully compensates for the lack of sec-adamantyl-
pyridine products. Chlorine-atom abstraction by adamantyl
Catalytic Oxidation of Adamantane Mediated by Fe(III)
Reagents. Collected in Table 2 are product profiles obtained
from H2O2-dependent oxidations of adamantane mediated by a
number of Fe(III) reagents ([Fe(Pic)3]‚0.5py (10), [FeIII2O(Pic)4-
(py)2] (8), [FeIII2(µ-OH)2(Pic)4], as well as FeCl3). Most data
derive from oxygenations involving [Fe(Pic)3]‚0.5py (10), since
this reagent is readily generated by other Fe(III) picolinate
species (such as 8 or [FeIII2(µ-OH)2(Pic)4]) in py/AcOH.
Conditions employed in catalytic oxygenations by Fe(III)/H2O2
reagents are similar to those used with Fe(II)/H2O2 combinations,
thus permitting direct comparison of the two systems.
The product distribution obtained from the oxidation of
adamantane (5 mmol) by H2O2 (2 mmol) in py/AcOH (10:1
v/v) in the presence of [Fe(Pic)3]‚0.5py (10, 0.2 mmol) under
N2 (Table 2, entry 1) exhibits similar features to those previously
(45) Minisci, F.; Vismara, E.; Fontana, F. Heterocycles 1989, 28, 489-
519.

Generation of Radicals in Gif-Type Oxidations
J. Am. Chem. Soc., Vol. 122, No. 31, 2000 7511
Scheme 3
Scheme 4
numerical discrepancy with respect to the control experiment
was explained as preferential addition of tert-adamantyl radicals
to pyridine coordinated to iron), the near absence of sec-
adamantylpyridines was taken as evidence for the lack of
diffusively free sec-adamantyl radicals in Gif chemistry. Minisci
has recently reinterpreted46 the latter result by hypothesizing
that the Gif conditions do not favor the irreversible rearoma-
tization process, thus, by virtue of the more reversible sec-
adamantyl-radical addition to pyridinium, dioxygen is permitted
to compete most effectively by trapping sec-adamantyl radicals.
However, the same argument should also predict that by
comparison to the control experiment, tert-adamantyl oxo
products ought to be somewhat more enhanced under Gif
conditions; the reverse is actually realized in Barton’s experi-
ments26 by a large margin (see below for further discussion).
In the present study, authentic tert- and sec-adamantyl
radicals are generated via photolysis of the corresponding
N-adamantylcarboxylpyridine-2-thione derivatives (0.115 mmol
of each) in py/AcOH (10:1 mL) under a stream of 4% O2 in N2
(Scheme 4). The so formed tert- and sec-adamantyl radicals
are competitively captured by three traps present in solution:
dioxygen, protonated pyridine, and the 2-pyridylthiyl moiety
of Barton’s esters. In contrast to Barton’s control experiments
which are conducted in the absence of metal, reagents [FeII(Pic)2-
(py)2] (2) or [Fe(Pic)3]‚0.5py (10) are added to py/AcOH at
concentrations comparable to those employed in Gif oxygen-
ations. This is necessary in order to better mimic the conditions
of the Gif experiment, especially since metal ions can be readily
involved in redox reactions with alkyl radicals and thus influence
the result of the competition kinetics. Importantly, neither
reagent generates Gif-type oxygenation systems in the presence
of O2 and a reducing agent (they are strictly H2O2-dependent
in their oxygenation activity), so potential complications due
to interference of oxygen-centered radicals such as HO•/HOO•
are minimized.
Collected in Table 3 are numerical results of the quantification
(by GC) of oxygenated and pyridine-coupled adamantyl products
generated in the control experiments noted above, in the
presence of [FeII(Pic)2(py)2] (2) and [Fe(Pic)3]‚0.5py (10). Mass
balances are provided by AdS-py and AdCOOH (derived by
hydrolysis of the precursor esters), both of which were not
quantified as they do not alter the results of adamantyl radical
partitioning between dioxygen and protonated pyridine. In the
presence of 2 (entry 1), the calculated ratio of oxygen-trapped
versus pyridine coupled tert-adamantyl radicals (tert Ad(O)/
Ad-py) equals 0.37, whereas the corresponding ratio for the
secondary position (sec Ad(O)/Ad-py) is 12.0. Similar results
are obtained in the presence of 10 (entry 2), for which tert
Ad(O)/Ad-py is 0.44 and sec Ad(O)/Ad-py equals 12.8. By
comparison to Barton’s results26 derived from experiments in
the absence of metals, the corresponding numerical values (tert
Ad(O)/Ad-py ) 0.74; sec Ad(O)/Ad-py ) 4.3) are within the
same order of magnitude, although the presence of the iron
reagents tends to relatively enhance pyridine-trapped products
in the case of tert-adamantyl radicals, and dioxygen-captured
derivatives in the case of sec-adamantyl radicals. Another
interesting observation is that 2-one is favored over 2-ol both
radicals from FeCl3 derivatives (eq 18) is most likely the source
of chloroadamantanes.
FeIII-Cl + Ad• f FeII + Ad-Cl
(18)
Control Experiments with Authentic Adamantyl Radicals.
To secure that the product profiles of adamantane oxidations
mediated by [FeII(Pic)2(py)2] (2) and [Fe(Pic)3]‚0.5py (10) are
uniquely determined by competition between dioxygen and
protonated pyridine for capturing tert- and sec-adamantyl
radicals, we revisited an early control experiment26 that has been
central to Barton’s argument1 in support of a non-radical
mechanism for the activation of sec-adamantyl sites in Gif
chemistry. Assuming that authentic tert- and sec-adamantyl
radicals (Scheme 3) are generated in py/AcOH (10/1 v/v) under
4% O2 in N2 (for instance, via photolysis of Barton’s PTOC
esters26), their fate will be determined by competitive partition-
ing between dioxygen trapping and addition to pyridinium
cation. Dioxygen trapping is expected22 to be diffusion con-
trolled for both types of radicals to yield the corresponding
adamantylperoxyl radicals, which are the precursors of all oxo
products. On the other hand, the reVersible addition of tert- and
sec-adamantyl radicals at positions 2- and 4- of protonated
pyridine to generate the corresponding adamantylpyridinium
radicals has been anticipated22 to proceed at different rates for
the two radicals. Indeed, recent measurements by Minisci46 have
provided rate constants of 2.2 × 106 M-1 s-1 and 1.3 × 104
M-1 s-1 for the addition of tert- and sec-adamantyl radicals to
pyridinium cation, respectively. The origin of the difference,
by 2 orders of magnitude, between the two rate constants is
traced46 to the substantial charge-transfer character of the
transition state, which favors the more nucleophilic tert-
adamantyl radical. The so obtained adamantylpyridinium radi-
cals are highly reducing and readily rearomatize to the final
adamantylpyridine products. Barton’s control experiments26 with
authentic adamantyl radicals showed that at 4% O2 in N2 the
ratio of oxygen-trapped over pyridine-coupled adamantyl prod-
ucts is 0.74 for the tertiary positions and 4.3 for the secondary
sites, reflecting the different rates for tert- and sec-adamantyl
radical addition to pyridinium. In contrast, a typical oxidation
of adamantane by a GifIV-type reagent (FeCl2/Zn/O2 (4%))
afforded26 a corresponding ratio of 0.03 for the tertiary positions
and 94 for the secondary sites. While the preponderance of tert-
adamantylpyridines in Gif chemistry was interpreted26 as solid
evidence for the presence of tert-adamantyl radicals (and the
(46) Recupero, F.; Bravo, A.; Bjørsvik, H.-R.; Fontana, F.; Minisci, F.;
Piredda, M. J. Chem. Soc., Perkin Trans. 2 1997, 2399-2405.

7512 J. Am. Chem. Soc., Vol. 122, No. 31, 2000
Kiani et al.
Table 3. Partition of Authentic and Gif Produced tert- and sec-Adamantyl Radicals between Dioxygen, Protonated Pyridine and TEMPO
a 9.2 µmol 2 or 10 as specified; 0.115 mmol of N-adamantyl-1-carboxylpyridine-2-thione and N-adamantyl-2-carboxyl-pyridine-2-thione each;
hν (ACE-Hanovia mercury lamp); py/AcOH (10/1 mL); 2 h. nd ) not detected. b 0.20 mmol 2 or 10 as specified; 5.0 mmol AdH; 2.0 mmol H2O2
(added by syringe pump over 6 hours at RT); py/AcOH (30/3.0 mL). c 5.0 mmol of TEMPO added to solutions specified in (b). d 2.0 mmol of
TEMPO added to (b). e 0.02 mmol 2 or 10 as specified; 0.250 mmol of N-adamantyl-1-carboxylpyridine-2-thione and N-adamantyl-2-carboxylpyridine-
2-thione each; 0.50 mmol of TEMPO; hν (ACE-Hanovia mercury lamp); py/AcOH (3.0/0.30 mL); 2 h.
in the absence of metal in Barton’s experiments (although the
presence of 2-Ad-OOH obscures the picture) and most promi-
nently in the presence of iron reagents in the present study.
The product profiles provided by the corresponding Gif
oxidations under 4% O2 in N2 afford calculated values of tert
Ad(O)/Ad-py and sec Ad(O)/Ad-py equal to 0.43 and 10.0 in
the presence of [FeII(Pic)2(py)2] (2) (entry 3) and 0.43 and 11.5
in the presence of [Fe(Pic)3]‚0.5py (10) (entry 4), respectively.
These values are in much better agreement to those obtained
via control experiments featuring genuine adamantyl radicals
(entries 1, 2), by comparison to values (tert Ad(O)/Ad-py )
0.08; sec Ad(O)/Ad-py ) 94) reported by Barton,26 derived from
GifIV-type (Fe(II)/O2 (4%)/Zn) oxygenations of adamantane. The
agreement is better between experiments featuring authentic
adamantyl radicals and Gif oxygenations involving [Fe(Pic)3]‚
0.5py (10), and less so for those dependent on [FeII(Pic)2(py)2]
(2). However, in the latter case values of tert Ad(O)/Ad-py and
sec Ad(O)/Ad-py obtained from several Gif oxygenations were
varying within a wider range, namely from 0.3 to 0.5 for the
tertiary sites and from 8 to 12 for the secondary positions.
TEMPO-trapped Adamantyl Radicals. To ascertain that
diffusively free tert-adamantyl radicals and most importantly
sec-adamantyl radicals are generated in H2O2-dependent Gif
oxygenations mediated by [FeII(Pic)2(py)2] (2) and [Fe(Pic)3]‚
0.5py (10) under high partial pressures of O2, a radical trap
superior to protonated pyridine was sought to compete with
dioxygen. This is particularly important for those oxygenations
in which adamantylpyridines (especially sec-adamantyl-
pyridines) are formed in trace amounts under the reaction
conditions employed (high partial pressures of oxygen and/or
limited protonation of pyridine) and would thus obscure
interpretation of competition experiments.
reactions of alkyl radicals with TEMPO have been measured47a
to be ∼1 × 109 M-1 s-1 in hydrocarbons (somewhat lower
values are obtained for resonance-stabilized radicals) and about
half this value in water. The rate constants are inversely related
to the thermodynamic stability of the carbon-centered radicals,
and are also affected by steric factors. TEMPO was used in the
present study to assess the formation of tert-Ad-TEMPO and
sec-Ad-TEMPO in the course of Gif-type oxygenation of
adamantane. Fortunately, both Ad-TEMPO products are volatile
and stable under GC conditions, so that they can be quantified
along with the other products of adamantane oxidation.
Table 3 (entries 5-10) summarizes product profiles of
adamantane oxidation in the presence of TEMPO under variable
conditions. Adamantane (5 mmol, entry 5) is oxidized by H2O2
(2 mmol) in reactions mediated by [FeII(Pic)2(py)2] (2) (0.2
mmol) under N2 in the presence of TEMPO (5 mmol, 0.15 M)
to afford almost exclusively 1-Ad-TEMPO and 2-Ad-TEMPO
in amounts exceeding by 5-fold those of the corresponding oxo
products (present in low yields). Minor amounts of only tert-
adamantylpyridines are also present, confirming the superiority
of TEMPO as a radical trap. The tert/sec selectivity (3°/2° )
3.10) is typical of similar Gif oxidations noted above, thus
TEMPO does not seem to have an effect on the generation of
the active oxidant involved. However, parallel oxidation pro-
cesses do occur, as 4-oxo-TEMPO is detected among the
reaction products in low yields. Furthermore, a control experi-
ment under identical conditions indicated that adamantan-2-ol
is readily converted to adamantanone in the presence of TEMPO
(but only to a limited extent in the absence of TEMPO). The
transformation is probably due to metal-dependent one-electron
oxidation of TEMPO to the corresponding oxoammonium
cation, which is known48 to mediate oxidation of alcohol to
aldehydes or ketones, albeit slowly for secondary alcohols.
The competition between dioxygen and TEMPO becomes
evident at lower concentrations of TEMPO (0.06 M, entry 6),
as the preponderance of Ad-TEMPOs versus oxo products is
TEMPO (2,2,6,6-tetramethyl-1-piperidinyloxy) is a widely
used trap of carbon-centered radicals.47 Rate constants for the
(47) (a) Chateauneuf, J.; Lusztyk, J.; Ingold, K. U. J. Org. Chem. 1988,
53, 1629-1632. (b) Beckwith, A. L. J.; Bowry, V. W., Moad, G. J. Org.
Chem. 1988, 53, 1632-1641. (c) Bowry, V. W.; Lusztyk. J.; Ingold, K. U.
J. Am. Chem. Soc. 1989, 111, 1927-1928. (d) Bowry, V. W.; Lusztyk, J.;
Ingold, K. U. Pure Appl. Chem. 1990, 62, 213-216.
(48) (a) Semmelhack, M. F.; Schmid, C. R.; Corte´s, D. A.; Chou, C. S.
J. Am. Chem. Soc. 1984, 106, 3374-3376. (b) Semmelhack, M. F.; Chou,
C. S.; Corte´s, D. A.; J. Am. Chem. Soc. 1983, 105, 4492-4494.

Generation of Radicals in Gif-Type Oxidations
J. Am. Chem. Soc., Vol. 122, No. 31, 2000 7513
reduced to approximately 2-fold. The product profile under 4%
O2 in N2 (not shown) is also very similar to that observed under
N2. It is only under pure dioxygen (entry 7) that the amounts
of oxo products supersede those of the corresponding Ad-
TEMPOs (tert Ad-TEMPO/Ad(O) ) 0.52; sec Ad-TEMPO/
Ad(O) ) 0.44) in the presence of 0.15 M of TEMPO. Authentic
tert- and sec-adamantyl radicals, generated by photolysis of the
corresponding Barton’s PTOC esters as previously indicated,49
partition between dioxygen (100%) and TEMPO (0.15 M) in
the presence of [FeII(Pic)2(py)2] (2) in a closely analogous
manner (entry 8: tert Ad-TEMPO/Ad(O) ) 0.56; sec Ad-
TEMPO/Ad(O) ) 0.46). Similar results have been obtained in
oxidation experiments mediated by [Fe(Pic)3]‚0.5py (10). For
example, under pure dioxygen (entry 9), the ratios of TEMPO-
versus oxygen-trapped adamantyls (tert Ad-TEMPO/Ad(O) )
0.50; sec Ad-TEMPO/Ad(O) ) 0.64) are analogous to those
obtained in Gif oxygenations mediated by 2. Once again,
authentic adamantyl radicals generated via photolysis in the
presence of 10, partition between dioxygen (100%) and TEMPO
(0.15 M) in an analogous fashion (entry 10: tert Ad-TEMPO/
Ad(O) ) 0.55; sec Ad-TEMPO/Ad(O) ) 0.69).
Scheme 5
KIE and may also be influenced by the kinetics of activation
of the second hydrogen atom.53
To circumvent some of the limitations associated with the
intermolecular KIE evaluation, we determined an intramolecular
primary KIE for the tertiary position using adamantane-1,3-d2
as substrate (Scheme 5).29 This substrate, which is readily
prepared29 by deuteration (LiAlD4) of commercially available
1,3-dibromoadamantane, has been successfully employed in KIE
determinations29 for oxygenations by P-450 model systems.
Observing all critical corrections and precautions recommended
in that elegant study29 (especially with respect to the deuterium
enrichment of the synthesized substrate (97.7(1) %), and the
necessity to integrate over the entire GC-eluted peak of the
unresolved products, composed of 1-adamantanol-d2 and -d1,
in order to accurately evaluate (MS) the relative molecular ion
content), we obtain a primary KIE value of 1.06(6) under N2
and 1.73(2) under O2 (4%) in N2 for the oxygenation of
adamantane to 1-adamantanol mediated by 2 (average of three
trials). The KIE value obtained in the presence of N2 is
consistent with that expected for hydroxyl radical activity, while
the higher value under low partial pressure of O2 signifies that
a more selective dioxygen-dependent oxidant (most likely,
adamantyloxyl radicals) is responsible for hydrogen-atom
abstraction in conjunction with hydroxyl radicals. It is also
reasonable to suggest that the intermolecular KIE value associ-
ated with the formation of 2-adamantanol (KIE ) 1.46(3))
reflects mostly the contribution of the secondary KIE, which
in the case of hydrogen-atom abstraction leading to carbon
rehybridization can be as high as 1.4, as recently observed in
FeCl3-catalyzed oxidation of ethylbenzene by t-BuOOH.54
(b) Gif Oxygenation of DMSO and DMSO/EtOH Com-
petition Experiments. The generation of hydroxyl radicals in
the course of Gif oxygenations was further investigated by
employing a highly specific reaction of hydroxyl radicals,
namely addition of HO• to DMSO (k ) 6.6 × 109 M-1 s-1) to
generate methanesulfinic acid and methyl radicals (eq 19).55 The
reaction favors addition of HO• to sulfur by a large margin
(92%)55a versus the alternative hydrogen-atom abstraction from
the methyl group (8%).56 The addition reaction is also strongly
indicative of the presence of hydroxyl radicals, inasmuch as
metal-oxo moieties are known to mediate oxygen atom transfer
to sulfides57 in the order R2S f R2SdO f R2S(O)2. However,
to exclude the possibility that some other oxidant (especially
iron (hydro)peroxo units58) may be responsible for the same
addition reaction, we chose to investigate the competition
The close correspondence between the aforementioned control
experiments involving genuine adamantyl radicals and Gif
oxygenations of adamantane mediated by [FeII(Pic)2(py)2] (2)
or [Fe(Pic)3]‚0.5py (10) leads to the conclusion that the product
profile of the oxygenation reactions is entirely dictated by the
generation of tert- and sec-adamantyl radicals.
In Search of the Active Oxidant. (a) Kinetic Isotope Effect.
An intermolecular deuterium kinetic isotope effect (KIE) has
been determined for the competitive Gif oxygenation of
adamantane and adamantane-d16 by 2 and 10 under dinitrogen.
The KIE values obtained via GC-MS analysis of the product
profile from three independent trials are as follows: 1-ol )
0.94(6) (2), 1.29(9) (10); 2-ol ) 1.46(3) (2), 1.40(7) (10); 2-one
) 1.88(4) (2), 1.99(14) (10); 2-(1-Ad)py ) 0.81(3) (2), 0.86(5)
(10); 4-(1-Ad)py ) 1.41(4) (2), 1.39(2) (10); 2-(2-Ad)py )
1.64(15) (2), 1.51(8) (10); 4-(2-Ad)py ) 1.38(5) (2), 1.36(3)
(10). Intermolecular KIE values50 between 1 and 2 are difficult
to interpret, as they may signify the influence of (i) a
non-selective hydrogen-abstracting agent, (ii) a more selective
oxidant acting via a bent-transition state, or (iii) a rate-limiting
step other than the C-H activation process. The KIE value of
approximately 250 for the formation of the major product,
adamantanone, has been frequently used to distinguish the active
oxidant of Gif systems from the much more selective oxidant
of the biological monooxygenases P-450 (KIE ) 10-14;51 more
modest values have also been reported29) or sMMO25b (KIE g
50 for methane, 4 for ethane (Methylosinus trichosporium)52),
but also from the action of hydroxyl radicals, for which a value
close to 1 is expected. Barton’s interpretation1 of this value
invokes a side-on [2 + 2] approach of the C-H bond to the
FedO unit in the transition state. However, the KIE value for
adamantanone can be a composite of primary and secondary
(53) Hoffmann, P.; Robert, A.; Meunier, B. Bull. Soc. Chim. Fr. 1992,
129, 85-97.
(49) Barton, D. H. R.; Chabot, B. M.; Hu, B. Tetrahedron 1996, 52,
10301-10312.
(50) Barton, D. H. R.; Doller, D.; Geletii, Y. V. Tetrahedron Lett. 1991,
32, 3811-3814.
(51) (a) Sono, M.; Roach, M. P.; Coulter, E. D.; Dawson, J. H. Chem.
ReV. 1996, 96, 2841-2887. (b) Ortiz de Montellano, P. R. In Cytochrome
P-450: Structure, Mechanism, and Biochemistry, 2nd ed.; Ortiz de
Montellano, P. R., Ed.; Plenum Press: New York, 1995; pp 245-303. (c)
Groves, J. T.; Han, Y.-Z. In Cytochrome P-450: Structure, Mechanism,
and Biochemistry, 2nd ed.; Ortiz de Montellano, P. R., Ed.; Plenum Press:
New York, 1995; pp 3-48.
(52) (a) Nesheim, J. C.; Lipscomb, J. D. Biochemistry 1996, 35, 5,
10240-10247. (b) Nesheim, J. C.; Lipscomb, J. D. J. Inorg. Biochem. 1995,
59, 369.
(54) (a) Merrigan, S. R.; Le Gloahec, V. N.; Smith, J. A.; Barton, D. H.
R.; Singleton, D. A. Tetrahedron Lett. 1999, 40, 3847-3850. (b) Singleton,
D. A.; Thomas, A. A. J. Am. Chem. Soc. 1995, 117, 9357-9358.
(55) (a) Veltwisch, D.; Janata, E.; Asmus, K.-D. J. Chem. Soc., Perkin
Trans. 2 1980, 146-153. (b) Urbanski, S. P.; Stickel, R. E.; Wine, P. H. J.
Phys. Chem. A 1998, 102, 10522-10529.
(56) Clement, J.-L.; Gilbert, B. C.; Ho, W. F.; Jackson, N. D.; Newton,
M. S.; Silvester, S.; Timmins, G. S.; Tordo, P.; Whitwood, A. C. J. Chem.
Soc., Perkin Trans. 2 1998, 1715-1717.
(57) Mukerjee, S.; Stassinopoulos, A.; Caradonna, J. P. J. Am. Chem.
Soc. 1997, 119, 8097-8098.
(58) (a) Ho, R. Y. N.; Roelfes, G.; Feringa, B. L.; Que, L., Jr. J. Am.
Chem. Soc. 1999, 121, 264-265. (b) Kim, C.; Chen, K.; Kim, J.; Que, L.,
Jr. J. Am. Chem. Soc. 1997, 119, 5964-5965.

7514 J. Am. Chem. Soc., Vol. 122, No. 31, 2000
Kiani et al.
Table 4. Product Profile of DMSO and DMSO/EtOH Oxidation
by H2O2 Mediated by [Fe(Pic)2(py)2] (2)
methyl radical (the least selective and least nucleophilic of all
alkyl radicals63) to all aromatic positions of pyridine and/or
protonated pyridine. Traces of methylated bipyridines are also
detected by GC-MS. Addition of EtOH to the reaction mixture
at various increasing concentrations versus DMSO (entries 2-4)
causes progressive reduction of the amounts of all picolines and
concomitant augmentation of the amounts of pyridine-trapped
hydroxyethyl radicals. As observed in the case of adamantyl
radicals, these latter ethanol-derived radicals add exclusively
at positions 2- and 4- of pyridine, owing to the high rate and
regioselectivity of addition of carbon-centered radicals to
protonated (rather than non-protonated) pyridine, both increasing
with the nucleophilicity of the alkyl radical.45 From the ratio
of pyridine-captured methyl over hydroxyethyl radicals and
knowledge of the relative concentrations of DMSO and EtOH,
a ratio of rate constants kEtOH/kDMSO of 0.34(3) is calculated,
which is in reasonable agreement with that reported from pulse
radiolysis experiments (kEtOH/kDMSO ) 0.29).59 Furthermore,
the internal competition for hydrogen-atom abstraction from the
R- and â-carbon atoms of EtOH, as judged by the ratio py-
CH(CH3)OH/py-CH2CH2OH, yields a selectivity of 7.3(4) which
is not unlike that documented by radiation chemistry65 (eq 20),
considering the uncertainty in determining the low-yield py-
CH2CH2OH products. The qualitative results and quantitative
estimates of the experiments presented in this section strongly
support the involvement of hydroxyl radicals as major oxidant
in Gif-type oxygenations.
a General conditions unless otherwise noted; 0.2 mmol 2; 2 mmol
H2O2 (added by syringe pump over 6 hours at RT); DMSO (8.0 mL);
py/AcOH (15.0/1.5 mL). b DMSO (8.8020 g, 112.6 mmol); EtOH
(5.1149 g, 111.0 mmol). c DMSO (8.8070 g, 112.7 mmol); EtOH
(6.6289 g, 143.9 mmol). d DMSO (8.8180 g, 112.8 mmol); EtOH
(8.0220 g, 174.1 mmol).
kinetics of the underlying oxidant by coupling the DMSO
addition reaction to hydrogen-atom abstraction from ethanol
(k ) 1.9 × 109 M-1 s-1 59
)
(eq 20).
Me2SdO + HO• f MeS(dO)OH + Me•
(19)
CH3CH2OH + HO• f H2O + •CH2CH2OH (13.2%) +
•
CH3 CHOH (84.3%) + CH3CH2O• (2.5%) (20)
(c) Gif Oxygenations in the Presence of 1-Iodoadaman-
tane. To ascertain that the formation of picolines observed in
the oxygenation of DMSO by Gif reagents is due to generation
of methyl radicals via eq 19, the reaction of HO• with DMSO
was coupled to the well documented methyl radical abstraction
of iodine atoms from alkyl iodides,66 conveniently applied to
1-iodoadamantane67 (eq 22). Similar reaction sequences have
been introduced by Minisci68 for the preparative scale synthesis
of alkylpyridines in the presence of protonated N-heterocycles.
Hydroxyl radicals were provided in those experiments by Fenton
reagents.
DMSO has been frequently used as an efficient trap of
hydroxyl radicals generated from aqueous oxygenation sys-
tems.60 Hydroxyl radicals are quantified by usually monitoring
the accumulation of methanesulfinic acid by HPLC,60b UV-
vis61 or ion chromatography.60a Quantification of methyl radicals
by trapping with fluorescamine-derivatized nitroxyl radicals has
also been reported.62 In the present study, the characteristic
addition of alkyl radicals to protonated pyridine63 permits facile
quantification (by GC) of all carbon-centered radicals generated
in eqs 19 and 20, provided that the reaction is performed under
strictly inert atmosphere (argon) to avoid capturing of these
radicals by dioxygen to the greatest possible extent. Furthermore,
reactions with Fe(II) precursor reagents are solely entertained
to minimize possible interference from the diffusion controlled
oxidation of R-hydroxyethyl radicals to acetaldehyde by Fe(III)
sites (eq 21).45,64 Fortunately, as mentioned above, Fe(III) Pic
species reductively decompose to their Fe(II) congeners in the
presence of H2O2.
Me• + 1-Ad-I f Me-I + 1-Ad•
(22)
A typical Gif oxygenation of DMSO (8 mL) by H2O2
(2 mmol) mediated by 2 (0.2 mmol) in the presence of
1-iodoadamantane (5 mmol) in py/AcOH (15.0/1.5 mL) under
argon affords cleanly 2-(1-Ad)py (0.55 mmol) and 4-(1-Ad)py
(0.44 mmol), as well as trace amounts of 1-Ad-OH. No picolines
are detected, suggesting that the methyl radicals, generated via
reaction 19 in amounts commensurating with those observed
in the absence of 1-Ad-I, are predominantly diverted toward
iodine abstraction and 1-Ad• formation (eq 22).
Under the same experimental conditions, but in the absence
of DMSO, the two tert-adamantylpyridines are still observed
in lower yields (2-(1-Ad)py: 0.16 mmol; 4-(1-Ad)py: 0.13
mmol), accompanied by a host of other products. Those have
been identified, but not quantified, by GC-MS and include
iodopyridines, four diiodoadamantanes (1,2-I2-Ad (racemic),
1,3-I2-Ad, 1,4-I2-Ad (two diastereomers)), and eight 1-iodo-
FeIII + CH3 CHOH f FeII + H+ + CH3CHO (21)
•
Table 4 summarizes the results of several Gif oxygenation
experiments mediated by 2 (0.2 mmol) in py/AcOH (15.0/1.5
mL) in the presence of DMSO or DMSO/EtOH (various relative
concentrations). In the presence of DMSO (entry 1), 2-, 3-, and
4-picoline are obtained, as expected for the addition of the
(59) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. J.
Phys. Chem. Ref. Data 1988, 17, 513-886.
(60) (a) Yurkova, I. L.; Schuchmann, H.-P.; von Sonntag, C. J. Chem.
Soc., Perkin Trans. 2 1999, 2049-2052. (b) Daniels, J. S.; Gates, K. S. J.
Am. Chem. Soc. 1996, 118, 3380-3385.
(65) Asmus, K.-D.; Mo¨ckel, H.; Henglein, A. J. Phys. Chem. 1973, 77,
1218-1221.
(61) Steiner, M. G.; Babbs, C. F. Arch. Biochem. Biophys. 1990, 278,
478-481.
(66) Hawari, J. A.; Kanabus-Kamiuska, J. M.; Wayner, D. D. M.; Griller,
D. In Substituent Effects in Radical Chemistry; Viehe, H. G., Ed.; Reidel
Publishing Company: Dordrecht, 1986; pp 91-107.
(67) Ryu, I.; Yamazaki, H.; Ogawa, A.; Kambe, N.; Sonoda, N. J. Am.
Chem. Soc. 1993, 115, 1187-1189.
(62) Li, B.; Gutierrez, P. L.; Blough, N. V. Anal. Chem. 1997, 69, 4295-
4302.
(63) Minisci, F.; Vismara, E.; Fontana, F.; Morini, G.; Serravalle, M.;
Giordano, C. J. Org. Chem. 1987, 52, 730-736.
(64) Minisci, F.; Citterio, A.; Vismara, E. Tetrahedron 1985, 41, 4157-
4170.
(68) Fontana, F.; Minisci, F.; Vismara, E. Tetrahedron Lett. 1988, 29,
1975-1978.

Generation of Radicals in Gif-Type Oxidations
J. Am. Chem. Soc., Vol. 122, No. 31, 2000 7515
adamantyl-pyridine coupled products (most likely due to addi-
tion of the three possible 1-I-Ad• radicals to positions 2- and 4-
of pyridine, including two pairs of diastereomers). All products
are well understood and arise from initial hydroxyl-radical
addition to pyridine59,69 and hydrogen-atom abstraction from
1-iodoadamantane, generating carbon-centered radicals. These
in turn perform iodine-atom abstraction from 1-iodoadamantane
and/or coupling with protonated pyridine. With the exception
of the tert-adamantylpyridines, all other products are reduced
to trace amounts in the presence of DMSO (8 mL, 113 mmol)
as previously noted. Similarly, formation of these products
is inhibited to a lesser extent in the presence of EtOH (8 mL,
136 mmol). These results further confirm that the active
oxidant involved initiates genuine radical chemistry under Gif
conditions.
comparisons were made1 with systems in which much more
selective oxygen-centered radicals were operating.76 These and
many other inconsistencies and fallacies in the interpretation
of Gif chemistry have been summarized in an insightful review
by M. J. Perkins22 to which the reader is referred.
Nevertheless, adamantane was perceived by Barton2 as an
exception to the application of an all pervasive non-radical
mechanism, since the formation of tert-adamantyl radicals could
not be denied in light of the presence of tert-adamantylpyridines.
The argument was thus reduced to the applicability of the non-
radical activation path to the secondary C-H positions, for
which no sec-adamantylpyridines were detected. These conclu-
sions were further supported by the competition experiments26
noted above, indicating that the degree of partitioning of genuine
tertiary and secondary adamantyl radicals between dioxygen
(4%) and protonated pyridine differed from what was obtained
in GifIV-type chemistry. Two observations which should have
caused concern about the validity of these control experiments
are (i) the discrepancy of the numerical values even for the
tertiary adamantyl radicals, and (ii) that the amounts of sec-
adamantylpyridines produced in the Gif experiment were small,
but not zero. A major finding of the present work is that, under
inert atmosphere, sec-adamantylpyridines can be readily detected
in good yields in Gif oxygenations of adamantane by H2O2.
Moreover, the analogous competition experiments involving
authentic adamantyl radicals, when conducted under conditions
permitting genuine comparison with the Gif experiment, provide
partitioning values which are very similar to those obtained in
Gif chemistry. It is possible that the high sec Ad(O)/Ad-py value
(∼94) obtained by Barton26 via a GifIV-type reagent (FeCl2/
Zn/O2 (4%)), if not erroneous, may be due to the reductive
environment of this system. As suggested by Minisci,46 these
conditions may not be as effective as the Barton PTOC esters
in inducing oxidative rearomatization to generate sec-adamantyl-
pyridines. We have also noted that static delivery of O2 (4%)/
N2 in the headspace of the Gif reaction tends to increase the
amount of tert-adamantylpyridines at the expense of 1-ol,
probably due to O2 depletion. The presence of Zn, which is
known77 to reduce O2 to superoxide in py/AcOH, may contribute
to this effect. The present results provide strong evidence that
the product profile of Gif oxygenation of adamantane is dictated
by the generation of diffusively free tert- and sec-adamantyl
radicals.
(d) Miscellaneous Control Experiments. As an additional
test of the viability of the FeIVdO hypothesis we attempted to
oxygenate adamantane (5 mmol) by the oxo donor reagent 2,6-
dimethyl-1-iodosylbenzene (2 mmol) in the presence of 2 in
py/AcOH (30.0/3.0 mL) under argon. Not even traces of oxo
products or adamantylpyridines were observed. Although the
experiment does not conclusively negate the action of FeIVdO
units, as it is not known whether coordinated pyridine can be
easily replaced by the oxo donor moiety,70 it is nevertheless
supportive of the main conclusion of the present study.
Finally, we note that attempts to generate authentic hydroxyl
radicals in py/AcOH in the presence of adamantane, by
71
continuous UV photolysis of H2O2 (by means of a high-
pressure xenon lamp, 75 W), afforded product profiles which
demonstrate the usual partioning of adamantyl radicals between
dioxygen and protonated pyridine. However, the high tert/sec
selectivities obtained even under a constant stream of argon (3°/
2° ≈ 10), and the yellowing of the reaction solutions, suggested
that other less indiscriminate radical species generated during
photolysis (most likely derivatives of pyridinyl radicals72)
interfere with the outcome of this control experiment.
Further Discussion and Conclusions
Early reports of Gif chemistry by D. H. R. Barton and co-
workers1 were centered on the oxygenation of adamantane as a
mechanistic probe of what was perceived to be a non-radical
pathway for the activation of hydrocarbons. Initial product
profiles largely favored (sec/tert ) 22)73 the formation of
secondary adamantyl oxo products (particularly 2-one) over
1-adamantanol. The subsequent addition of tert-adamantyl-
pyridines (but not sec-adamantylpyridines) to the product profile
reduced the sec/tert selectivity to approximately 1.1.74 These
values had not been corrected on a per-hydrogen basis at the
time, thus concealing the actual preponderance of tertiary over
secondary C-H activation (tert/sec ) 2.7) in adamantane
oxidations. This selectivity is only slightly higher than what
would be expected from the action of the indiscriminate
hydroxyl radicals.75 This point was not taken into account, as
Newcomb’s results23 on the oxidation of the radical-clock
substrate 1-methyl-2-phenylcyclopropane by the Gif reagent
FeCl3/H2O2 provided the first strong indication that diffusively
free alkyl radicals are generated in Gif solutions. Indeed, the
entire product profile was derived from the rapidly rearranging
(k ) 3 × 1011 s-1 (ring opening)), free (2-phenyl-cyclopropyl)-
carbinyl radical, and no products were observed from insertion
into the methyl C-H bonds of the unrearranged substrate. These
results were interpreted by Barton78 as an exceptional case,
similar to that observed for the tertiary position of adamantane,
in which a putative FeV-R bond supposedly collapses to FeIV
and R•. The present work shows that there is nothing special
(69) (a) Cercek, B.; Ebert, M. Trans. Faraday Soc. 1967, 63, 1687-
1698. (b) Steenken, S.; O’Neill, P. J. Phys. Chem. 1978, 82, 372-374.
(70) For reservations concerning the intermediacy of metal-oxo units
in oxygenations mediated by iodosylbenzene derivatives see: Nam, W.;
Valentine, J. S. J. Am. Chem. Soc. 1993, 115, 1772-1778.
(71) (a) Nadezhdin, A.; Dunford, H. B. J. Phys. Chem. 1979, 83, 1957-
1961. (b) Livingston, R.; Zeldes, H. J. Chem. Phys. 1966, 44, 1245-1259.
(72) Kosower, E. M. In Free Radicals in Biology; Pryor, W. A., Ed.;
Academic Press: New York, 1976; pp 1-53.
(75) (a) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B.
J. Phys. Chem. Ref. Data 1988, 17, 513-886. (b) Baker, R. R.; Baldwin,
R. R.; Walker, R. W. Trans. Faraday Soc. 1970, 66, 2812-2826. (c) Russell,
G. A. In The Chemistry of Alkanes and Cycloalkanes; Patai, S., Rappoport,
Z., Eds.; Wiley: New York, 1992; pp 966-983. (d) Kochi, J. K. In Free
Radicals; Kochi, J. K., Ed.; Wiley: New York, 1973; Vol. II, pp 689-
690.
(76) Fossey, J.; Lefort, D.; Massoudi, M.; Nedelec, J.-Y.; Sorba, J. Can.
J. Chem. 1985, 63, 678-680.
(73) Barton, D. H. R.; Gastiger, M. J.; Motherwell, W. B. J. Chem. Soc.,
Chem. Commun. 1983, 41-43.
(77) Sawyer, D. T. Oxygen Chemistry; Oxford University Press: New
York, 1991; pp 120-187.
(74) Balavoine, G.; Barton, D. H. R.; Boivin, J.; Lecoupanec, P.;
Lelandais, P. New J. Chem. 1989, 13, 691-700.
(78) Barton, D. H. R. Personal communication.

7516 J. Am. Chem. Soc., Vol. 122, No. 31, 2000
Kiani et al.
about the tertiary position of adamantane, and the same should
also apply to the radical-clock substrates. Furthermore, there is
no evidence to support that high-valent FedO units are involved
as active oxidants in Gif chemistry (vide infra). An even earlier
report by Perkins,8 in which a Gif oxidation of cyclohexane
led to detection of cyclohexyl radicals by spin-trapping, did not
receive its due recognition at the time, and was dismissed as
an overly qualitative result.
circuitous interpretation is not supported by the results of the
present study, since adamantyl radicals are clearly present even
in the absence of TEMPO.
FeVdO + TEMPO f FeIVdO + [TEMPO]+ (24)
The most compelling evidence against the double radical
manifold is the realization that HO• radicals play a major role
in H-abstraction and addition reactions of Gif chemistry. The
high-valent FeIV/VdO units proposed by Barton3 have always
been suspected on account of the widely differing product
profiles obtained by Gif reagents (C-H and CdC ketonization)
versus those realized with biological monooxygenases51,81 (C-H
insertion to form alcohols; CdC epoxidation). The involvement
of hydroxyl radicals has been inferred in the present study by
virtue of diagnostic reactions, chief among which is their highly
specific addition of HO• to DMSO to afford pyridine-trapped
methyl radicals under Gif conditions. Competitive interception
of hydroxyl radicals by DMSO and EtOH confirms that the
competition kinetics are consistent with the presence of this
oxidant rather than some other active species generated in Gif
solutions. The low tert/sec selectivity values and the intra-
molecular KIE value close to unity, obtained under argon, are
also typical of HO• kinetics.75 Under increasing amounts of
dioxygen a more selective oxidant, most likely substrate-centered
alkoxyl radicals,22 may contribute to the hydrogen-abstracting
process, as judged by the increased tert/sec selectivity and
magnitude of the intramolecular KIE value. These results also
reverse scepticism expressed in our previous publication7 in
regards to the role of hydroxyl radicals in Gif chemistry,
although a handful of product profile discrepancies noted in
that study need to be revisited. In retrospect, we recognize that
the GifIV-type systems previously employed are unsuitable for
detailed mechanistic work due to complications introduced by
the obligatory presence of O2/Zn. As noted above, O2 increases
the selectivity of Gif reagents, most likely by increasing
the contribution of substrate-centered alkoxyl radicals versus
hydroxyl radicals. Thus, typical features of hydroxyl radical
chemistry, such as hydroxylation of aromatics, are suppressed.
In addition, O2 conceals the presence of sec-alkyl radicals. As
indicated in the present study, Zn powder also increases the
selectivity of the Gif reaction, probably via selective reduction
of the oxygen-centered radicals. The heterogeneity of the Gif
solutions introduced by Zn is another source of concern; notably,
mass balances are inferior in the presence of Zn. These
limitations are all circumvented with the present homogeneous
Fe/H2O2 combinations which permit addition of O2 at will,
therefore assisting in unravelling the underlying nature of Gif
chemistry.
Preliminary results of the present work have also been
criticized by Barton78 as indicating operation of Fe(II)/H2O2
dependent cycles (“FeII/FeIV radical manifold”; eqs 7, 8), even
in cases in which ferric reagents are employed as catalyst
precursors under scrupulous argon purging. This argument is
associated with the late introduction3 of two manifolds, the
radical noted above, and the original FeIII/FeV non-radical
manifold. The distinction was most likely introduced once it
was realized3 that Fe(II)/H2O2 based Gif oxygenations of
cyclohexane produce cyclohexyl chloride (in the presence of
Cl-) or cyclohexylpyridines (in the absence of Cl-), both
indicating the presence of cyclohexyl radicals, while Fe(III)/
H2O2 Gif reagents mostly produce cyclohexanone and cyclo-
hexanol, assumed to indicate non-radical pathways. An impor-
tant addendum to this observation was that, the Fe(II)/H2O2
system starts producing more oxo products under an oxygen
stream, while the Fe(III)/H2O2 system starts generating cyclo-
hexyl chloride in the presence of Cl- under an argon stream.
While these important results can be explained in a straight-
forward manner as consistent with the generation and partition-
ing of cyclohexyl radicals among O2 > FeIII-Cl > [pyNH]+
under all conditions, they were perceived3 instead as indicating
interconversion of the two manifolds via reversible loss of
dioxygen79 from a putative diferric peroxo unit (eq 23).
FeIII-O-O-FeIII h 2 FeII + O2
(23)
The present results refute the double manifold theory at
various levels of consideration, without disputing the fact that
all reagents examined most likely produce their active oxidant
via Fe(II)/H2O2 interactions. First, the product profiles of
adamantane oxygenations clearly illustrate that tert- and sec-
adamantyl radicals are produced by both Fe(II)/H2O2 and Fe(III)/
H2O2 systems, even for Fe(III) reagents such as Fe(Pic)3 which
have been extensively used in Barton’s experiments.3,80 The
amounts of sec-adamantylpyridines are indeed reduced when
Fe(III)/H2O2 systems are applied, but this is simply due to the
increased in situ production of O2 by these systems. Notable is
also the generation of 2-chloroadamantane at the expense of
sec-adamantylpyridines in the presence by FeCl3/H2O2 in py/
AcOH, a ferric reagent which has also been routinely used by
Barton.3 Most importantly, the results from Gif oxygenations
of adamantane in the presence of TEMPO and the related
competition experiments, show clearly that generation of tert-
and sec-adamantyl radicals dictates the product profile eVen
under pure O2. The formation of TEMPO-trapped sec-alkyl
radicals had been noted by Barton,49 but because sec-alkyl-
pyridines had not been observed in related experiments, it
was hypothesized49 that the generation of alkyl radicals was
induced by TEMPO via one electron-reduction of the putative
FeVdO oxidant of the non-radical manifold to the corresponding
FeIVdO oxidant of the radical manifold (eqs 24 and 7-8). This
It should also be noted that M. J. Perkins was the first to
pinpoint in an early report8 that the p-hydroxylation of phenyl-
alanine by an Fe(III)/H2O2 reagent is consistent with the action
of HO• radicals. More recently, M. J. Perkins22 has also drawn
attention to the well-documented Gif oxidation of pyridine
to hydroxypyridines and bipyridines, as another instance of
hydroxyl radical attack. Some controversy with respect to the
regioselectivity of pyridine hydroxylation in Gif chemistry
(which also gives some 3-HOpy, as expected for the electrophilic
HO•) and by photolytically generated HO• radicals (reportedly27
producing 2-HOpy and 4-HOpy in a ratio of 2:1) need to be
eventually reconciled. Furthermore, hydroxylation of aromatic
moieties by Gif reagents has been detected by M. Newcomb,23
who has also noted that the absence of molecular selectivity in
(79) Kitajima, N.; Tamura, N.; Amagai, H.; Fukui, H.; Moro-oka, Y.;
Mizutani, Y.; Kitagawa, T.; Mathur, R.; Heerwegh, K.; Reed, C. A.; Randall,
C. R.; Que, L., Jr.; Tatsumi, K. J. Am. Chem. Soc. 1994, 116, 9071-9085.
(80) Barton, D. H. R.; Beck, A. H.; Delanghe, N. C. Tetrahedron Lett.
1996, 37, 1555-1558.
(81) (a) Wallar, B. J.; Lipscomb, J. D. Chem. ReV. 1996, 96, 2625-
2657. (b) Feig, A. L.; Lippard, S. J. Chem. ReV. 1994, 94, 759-805.

Generation of Radicals in Gif-Type Oxidations
J. Am. Chem. Soc., Vol. 122, No. 31, 2000 7517
competitive Gif oxygenations of cycloalkanes82 can be consistent
with H-atom abstraction by hydroxyl radicals.83 Notably, Barton1
frequently used to refer to the beneficial role of pyridine in Gif
chemistry in trapping any residual hydroxyl radicals.
needs to be better defined. The stoichiometry of the Fenton
reaction is better served by assuming an FeII-OOH precursor
of hydroxyl radicals and an FeIII-OOH precursor of hydro-
peroxyl radicals. In contrast, recent studies11 of iron-based
ROOH-supported (R ) H, t-Bu) oxygenation of substrates favor
homolytic cleavage of the O-O bond in FeIII-OOR precursor
species as the source of HO•/RO• radicals. The latter mode of
decomposition, which may be influenced by the strength of the
ligand field,58a has been documented with the assistance of
CoIII-OOR model species in a thorough study.91 Future
investigations are directed toward unraveling those metal-
centered (hydro)peroxo intermediates which may be the im-
mediate precursors of hydroxyl radicals and possibly active oxo
donors to olefinic substrates.
Can there be an additional active oxidant responsible for
the chemistry observed in the current investigation? For
unfunctionalized alkanes, the answer is largely negative.
Preliminary oxygenation experiments of cis-stilbene by H2O2
mediated by 2 or 10 in py/AcOH or CH3CN/AcOH indicate
that the product profile is mostly composed of benzaldehyde,
deoxybenzoin, trans-stilbene oxide, and cis-stilbene oxide
(especially in CH3CN/AcOH). While the former two products
can be readily generated via hydroxyl-radical addition to the
double bond84 (although other mechanisms involving electron
transfer from alkenes to metal oxo reagents have also been
discussed85), stilbene epoxides may instead originate from oxo
transfer processes.86 High-valent iron-oxo units87 need not be
invoked, since iron (hydro)peroxo species (FeIII-OO(H))88 may
be efficient oxo donor moieties,89 as also indicated in recent
reconsiderations of the role of the active oxidants in P-450
oxygenations.90 These FeII/III-OOH species may also be the
immediate precursors of hydroxyl radicals, although the im-
portance of the different oxidation states of iron in these units
Acknowledgment. We thank Corina E. Rogge and Professor
John P. Caradonna for a sample of 2,6-dimethyl-1-iodosyl-
benzene and fruitful discussions. Patricia C. Ong and Remle
C¸ elenligil-C¸ etin are also acknowledged for offering valuable
experimental assistance. We also thank Dr. Andrew Tyler for
assisting in obtaining electrospray ionization mass spectra and
Michael Creech for providing GC-MS data. The present work
was generously supported by grants from the U.S. Environ-
mental Protection Agency (R823377-01-1) and the donors of
the Petroleum Research Fund administered by the ACS (ACS-
PRF-29383-G3), and in part by a grant from the Division of
Chemical Sciences, Office of Basic Energy Sciences, Office of
Science, U.S. Department of Energy (DE-FG02-99ER14978).
(82) Barton, D. H. R.; Launay, F.; Le Gloahec, V. N.; Li, F.; Smith, F.
Tetrahedron Lett. 1997, 38, 8491-8494.
(83) Newcomb, M.; Simakov, P. A. Tetrahedron Lett. 1998, 39, 965-
966.
(84) (a) von Sonntag, C. The Chemical Basis of Radiation Biology; Taylor
& Francis: London, 1987; pp 37-38. (b) von Sonntag, C.; Schuchmann,
H.-P. Angew. Chem., Int. Ed. Engl. 1991, 30, 1229-1253.
(85) (a) Gross, Z.; Nimri, S. J. Am. Chem. Soc. 1995, 117, 8021-8022.
(b) Traylor, T. G.; Miksztal, A. R. J. Am. Chem. Soc. 1987, 109, 2770-
2774.
(86) Ostovic, D.; Bruice, T. C. Acc. Chem. Res. 1992, 25, 314-320.
(87) (a) Groves, J. T.; Nemo, T. E. J. Am. Chem. Soc. 1983, 105, 5786-
5791. (b) Castellino, A. J.; Bruice, T. C. J. Am. Chem. Soc. 1988, 110,
158-162. (c) Castellino, A. J.; Bruice, T. C. J. Am. Chem. Soc. 1988, 110,
7512-7519.
(88) Sam, J. W.; Tang, X.-J.; Peisach, J. J. Am. Chem. Soc. 1994, 116,
5250-5256.
(89) Murugesan, N.; Hecht, S. M. J. Am. Chem. Soc. 1985, 107, 493-
500.
Supporting Information Available: ORTEP diagrams and
Tables 1-35 containing listings of crystal data and structure
refinement, atomic coordinates and equivalent isotropic dis-
placement parameters, interatomic distances and bond angles,
anisotropic displacement parameters, and hydrogen coordinates
and isotropic displacement parameters for compounds 3, 5-7,
and 9-11 (PDF). X-ray crystallographic file in CIF format. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(90) (a) Newcomb, M.; Shen, R.; Choi, S.-Y.; Toy, P. H.; Hollenberg,
P. F.; Vaz, A. D. N.; Coon, M. J. J. Am. Chem. Soc. 2000, 122, 2677-
2686. (b) Toy, P. H.; Newcomb, M.; Coon, M. J.; Vaz, A. D. N. J. Am.
Chem. Soc. 1998, 120, 9718-9719.
JA000063H
(91) Chavez, F. A.; Rowland, J. M.; Olmstead, M. M.; Mascharak, P.
K. J. Am. Chem. Soc. 1998, 120, 9015-9027.