Mechanistic studies of (porphinato)iron-catalyzed isobutane oxidation. Comparative studies of three classes of electron-deficient porphyrin catalysts

Article abstract of DOI:10.1021/ic000284o

We report herein a comprehensive study of (porphinato)iron [PFe]-catalyzed isobutane oxidation in which molecular oxygen is utilized as the sole oxidant; these catalytic reactions were carried out and monitored in both autoclave reactors and sapphire NMR tubes. In situ ¹⁹F and ¹³C NMR experiments, coupled with GC analyses and optical spectra obtained from the autoclave reactions have enabled the identification of the predominant porphyrinic species present during PFe-catalyzed oxidation of isobutane. Electron-deficient PFe catalysts based on 5,10,15,20-tetrakis(pentafluorophenyl)porphyrin [(C₆F₅)₄PH₂], 2,3,7,8,12,13,17,18-octabromo-5,10,15,20-tetrakis(pentafluorophenyl)porphyrin [Br₈(C₆F₅)₄PH₂], and 5,10,15,20-tetrakis(heptafluoropropyl)porphyrin [(C₃F₇)₄PH₂] macrocycles were examined. The nature and distribution of hydrocarbon oxidation products show that an autoxidation reaction pathway dominates the reaction kinetics, consistent with a radical chain process. For each catalytic system examined, PFe(II) species were shown not to be stable under moderate O₂ pressure at 80 °C; in every case, the PFe(II) catalyst precursor was converted quantitatively to high-spin PFe(III) complexes prior to the observation of any hydrocarbon oxidation products. Once catalytic isobutane oxidation is initiated, all reactions are marked by concomitant decomposition of the porphyrin-based catalyst. In situ ¹⁷O NMR spectroscopic studies confirm the incorporation of ¹⁷O from labeled water into the oxidation products, implicating the involvement of PFe-OH in the catalytic cycle. Importantly, Br₈(C₆F₅)₄PFe-based catalysts, which lack macrocycle C-H bonds, do not exhibit augmented stability with respect to analogous catalysts based on (C₆F₅)₄PFe and (C₃F₇)₄PFe species. The data presented are consistent with a hydrocarbon oxidation process in which PFe complexes play dual roles of radical chain initiator, and the species responsible for the catalytic decomposition of organic peroxides. This modified Haber-Weiss reaction scheme provides for the decomposition of tert-butyl hydroperoxide intermediates via reaction with PFe-OH complexes; the PFe(III) species responsible for hydroperoxide decomposition are regenerated by reaction of PFe(II) with dioxygen under these experimental conditions.

Full text of DOI:10.1021/ic000284o

Inorg. Chem. 2000, 39, 3125-3139
3125
Mechanistic Studies of (Porphinato)Iron-Catalyzed Isobutane Oxidation. Comparative
Studies of Three Classes of Electron-Deficient Porphyrin Catalysts
Kevin T. Moore,† Istva´n T. Horva´th,*,‡ and Michael J. Therien*,†
Department of Chemistry, University of Pennsylvania, Philadelphia, Pennsylvania 19104-6323, and
Corporate Research Laboratories, Exxon Research and Engineering Company,
Annandale, New Jersey 08801
ReceiVed March 13, 2000
We report herein a comprehensive study of (porphinato)iron [PFe]-catalyzed isobutane oxidation in which molecular
oxygen is utilized as the sole oxidant; these catalytic reactions were carried out and monitored in both autoclave
reactors and sapphire NMR tubes. In situ 19F and 13C NMR experiments, coupled with GC analyses and optical
spectra obtained from the autoclave reactions have enabled the identification of the predominant porphyrinic
species present during PFe-catalyzed oxidation of isobutane. Electron-deficient PFe catalysts based on 5,10,15,-
20-tetrakis(pentafluorophenyl)porphyrin [(C6F5)4PH2], 2,3,7,8,12,13,17,18-octabromo-5,10,15,20-tetrakis(pen-
tafluorophenyl)porphyrin [Br8(C6F5)4PH2], and 5,10,15,20-tetrakis(heptafluoropropyl)porphyrin [(C3F7)4PH2]
macrocycles were examined. The nature and distribution of hydrocarbon oxidation products show that an
autoxidation reaction pathway dominates the reaction kinetics, consistent with a radical chain process. For each
catalytic system examined, PFeII species were shown not to be stable under moderate O2 pressure at 80 °C; in
every case, the PFeII catalyst precursor was converted quantitatively to high-spin PFeIII complexes prior to the
observation of any hydrocarbon oxidation products. Once catalytic isobutane oxidation is initiated, all reactions
are marked by concomitant decomposition of the porphyrin-based catalyst. In situ 17O NMR spectroscopic studies
confirm the incorporation of 17O from labeled water into the oxidation products, implicating the involvement of
PFe-OH in the catalytic cycle. Importantly, Br8(C6F5)4PFe-based catalysts, which lack macrocycle C-H bonds,
do not exhibit augmented stability with respect to analogous catalysts based on (C6F5)4PFe and (C3F7)4PFe species.
The data presented are consistent with a hydrocarbon oxidation process in which PFe complexes play dual roles
of radical chain initiator, and the species responsible for the catalytic decomposition of organic peroxides. This
modified Haber-Weiss reaction scheme provides for the decomposition of tert-butyl hydroperoxide intermediates
via reaction with PFe-OH complexes; the PFeIII species responsible for hydroperoxide decomposition are
regenerated by reaction of PFeII with dioxygen under these experimental conditions.
Introduction and Background
state, commonly referred to as compound I), yet must be
available to facilitate the 2 f 3 and 4 f 6 reductive conversions
outlined in Scheme 1. Such formidable requirements for an
abiological catalytic system were recognized by Groves,2-7
Traylor,8-11 Bruice,12-19 Meunier,20-22 Mansuy,23-29 Collman,30-34
The development of selective catalytic hydrocarbon oxidation
reactions is of primary importance for the functionalization of
natural products, drug design, development of new chemical
feedstocks, and small molecule activation. The catalytic para-
digm for such selective substrate oxygenations at ambient
temperature has been the enzyme cytochrome P450 (Scheme 1).1
While an attractive target for biomimetic catalyst development,
two major problems circumvent technological application of
such yet-to-be realized catalysts that derive their oxidizing
equivalents from dioxygen, the world’s most abundant and least-
expensive oxidant: (i) the need for coreductants and (ii) the
fact that one oxygen atom of O2 is wasted in the evolution of
water. These two facts highlight the quandaries for (porphinato)-
iron [PFe] catalysts that are modeled after cytochrome P450. For
instance, in addition to adding cost to any hydroxylated or
epoxidized hydrocarbons obtained by such a catalytic cycle,
coreductants must be sheltered from high-valent iron oxo species
responsible for selective hydrocarbon oxidation (a P•+FeIVdO
(2) Groves, J. T., Adhyam, D. V. J. Am. Chem. Soc. 1984, 106, 2177-
2181.
(3) Groves, J. T.; Quinn, R. Inorg. Chem. 1984, 23, 3844-3846.
(4) Groves, J. T. J. Chem. Educ. 1985, 62, 928-931.
(5) Groves, J. T.; Watanabe, Y. J. Am. Chem. Soc. 1986, 108, 507-508.
(6) Groves, J. T.; Watanabe, Y. J. Am. Chem. Soc. 1986, 108, 7834-
7836.
(7) Groves, J. T.; Watanabe, Y. J. Am. Chem. Soc. 1988, 110, 8443-
8452.
(8) Traylor, T. G.; Marsters, J. C., Jr.; Nakano, T.; Dunlap, B. E. J. Am.
Chem. Soc. 1985, 107, 5537-5539.
(9) Traylor, T. G.; Tsuchiya, S. Inorg. Chem. 1987, 26, 1338-1339.
(10) Traylor, T. G.; Miksztal, A. R. J. Am. Chem. Soc. 1987, 109, 2770-
2774.
(11) Traylor, T. G.; Hill, K. W.; Fann, W. P.; Tsuchiya, S.; Dunlap, B. E.
J. Am. Chem. Soc. 1992, 114, 1308-1312.
(12) Nee, M. W.; Bruice, T. C. J. Am. Chem. Soc. 1982, 104, 6123-6125.
(13) Calderwood, T. S.; Bruice, T. C. J. Am. Chem. Soc. 1985, 107, 8272-
8273.
(14) Dicken, C. M.; Lu, F. L.; Nee, M. W.; Bruice, T. C. J. Am. Chem.
Soc. 1985, 107, 5776-5789.
(15) Lee, W. A.; Bruice, T. C. J. Am. Chem. Soc. 1985, 107, 513-514.
(16) Dicken, C. M.; Woon, T. C.; Bruice, T. C. J. Am. Chem. Soc. 1986,
108, 1636-1643.
* To whom correspondence should be addressed.
† University of Pennsylvania.
‡ Exxon Research and Engineering Company. Present address: Depart-
ment of Organic Chemistry, Eotvos Lorand University, Pazmany Peter
Setany 1/A, H-1117 Budapest, Hungary.
(1) Dawson, J. H. Science 1988, 240, 433-439.
10.1021/ic000284o CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/22/2000

3126 Inorganic Chemistry, Vol. 39, No. 15, 2000
Scheme 1. Cytochrome P450 Catalytic Cycles1
Moore et al.
Scheme 2. Proposed Mechanism for Alkane Hydroxylation
Reactions Catalyzed by Electron-Deficient Iron Porphyrin
Catalysts That Exploit Dioxygen as the Stoichiometric
Oxidant
that utilize an abbreviated catalytic cycle (2 f 7 f 1 f 2) in
which PFeIII complexes are reacted directly with powerful,
sacrificial O atom donors, to produce P•+FeIVdO species that
possess reactivity similar to that of compound I of cytochrome
P450
.
For the specialized functions of cytochrome P450 (detoxifi-
cation, enantioselective biosynthesis of oxygenated organic
substrates), Nature can afford the energetic cost to split dioxygen
into water and a metal-bound O atom with the input of two
protons and two electrons; most potential applications of
biomimetic P450 catalysts cannot. For example, in a potential
large-scale selective oxidation of a hydrocarbon feedstock
chemical, not only would a P450-type biomimetic catalyst in
which one molecule of water was generated through each turn
of the catalytic cycle be energetically wasteful, but the associated
engineering and disposal problems would assuredly impact the
economic feasibility of the process. The tremendous need for
liquefied oxygenated fuels from lightweight hydrocarbons has
led to a number of notions to circumvent this problem; perhaps
the most intriguing hypothesis, first proposed by Ellis and
Lyons,39-47 involved the efficacy of potential electron-deficient
porphyrin-based catalysts in which an FeII-derived PFeIVdO
species mirrors the reactivity of compound I’s P•+FeIVdO
oxidation level. The primary advantage of such a scheme is
that both O atoms of O2 could be incorporated into hydrocarbon
substrates (Scheme 2).
Previous work showed that PFe catalysts that featured large
numbers of electron-withdrawing groups fused to the macrocycle
periphery facilitated the oxidation of branched hydrocarbons
(particularly isobutane) to oxidized alkanes when dioxygen was
utilized as the stoichiometric oxidant, suggesting that an
oxidation pathway similar to that outlined in Scheme 2 might
be operative.39-47 Not only did key questions arise related to
mechanism, but this work gave rise to speculation that perhaps
even more electron-deficient porphyrin macrocycles could lead
to the development of catalysts capable of hydroxylating primary
alkanes. Work by Labinger and Gray concluded that the likely
mechanism of these isobutane oxygenations was radical chain
autoxidation,48-50 where ferric and ferrous porphyrin complexes
played key roles in alkyl hydroperoxide decomposition; such a
and a number of other investigators,35-38 and have led to the
development of a family of porphyrin-based P450-type catalysts
(17) Woon, T. C.; Shirazi, A.; Bruice, T. C. Inorg. Chem. 1986, 25, 5,
3845-3846.
(18) Woon, T. C.; Dicken, C. M.; Bruice, T. C. J. Am. Chem. Soc. 1986,
108, 7990-7995.
(19) Castellino, A. J.; Bruice, T. C. J. Am. Chem. Soc. 1988, 110, 158-
162.
(20) Hoffmann, P.; Labat, G.; Robert, A.; Meunier, B. Tetrahedron Lett
1990, 31, 1991-1994.
(21) Hoffmann, P.; Robert, A.; Meunier, B. Bull. Soc. Chim. Fr. 1992,
129, 85-97.
(22) Hoffmann, P.; Meunier, B. New J. Chem. 1992, 16, 559-561.
(23) Mansuy, D.; Devocelle, L.; Artaud, I.; Battioni, J. P. NouV. J. Chim.
1985, 9, 711-716.
(24) Mansuy, D. Pure Appl. Chem. 1987, 59, 759-770.
(25) Artaud, I.; Gregoire, N.; Battioni, J. P.; Dupre, D.; Mansuy, D. J.
Am. Chem. Soc. 1988, 110, 8714-8716.
(26) Bartoli, J. F.; Brigaud, O.; Battioni, P.; Mansuy, D. J. Chem. Soc.,
Chem. Commun. 1991, 440-442.
(27) Mandon, D.; Ochsenbein, P.; Fischer, J.; Weiss, R.; Jayaraj, K.; Austin,
R. N.; Gold, A.; White, P. S.; Brigaud, O.; Battioni, P.; Mansuy, D.
Inorg. Chem. 1992, 31, 2044-2049.
(28) Mansuy, D. Coord. Chem. ReV. 1993, 125, 129-142.
(29) Mansuy, D.; Battioni, P.; Sheldon, R. A. Metalloporphyrins Catal.
Oxid. 1994, 99-132.
(30) Collman, J. P.; Sorrell, T. N.; Hoffman, B. M. J. Am. Chem. Soc.
1975, 97, 913-914.
(39) Bhinde, M. V.; Lyons, J. E.; Ellis, P. E., Jr. U.S. Patent 5550301 A,
1996.
(31) Kodadek, T.; Raybuck, S. A.; Collman, J. P.; Brauman, J. I., Papazian,
L. M. J. Am. Chem. Soc. 1985, 107, 4343-4345.
(32) Collman, J. P.; Hampton, P. D.; Brauman, J. T. J. Am. Chem. Soc.
1990, 112, 2977-2986.
(40) Ellis, P. E., Jr.; Lyons, J. E. J. Chem. Soc., Chem. Commun. 1989,
1189-1190.
(41) Ellis, P. E., Jr.; Lyons, J. E. Catal. Lett. 1989, 3, 389-397.
(42) Ellis, P. E., Jr.; Lyons, J. E. J. Chem. Soc., Chem. Commun. 1989,
1315-1316.
(33) Collman, J. P.; Zhang, X.; Lee, V. J.; Uffelman, E. S.; Brauman, J. I.
Science 1993, 261, 1404-1411.
(43) Ellis, P. E., Jr.; Lyons, J. E. J. Chem. Soc., Chem. Commun. 1989,
1187-1188.
(34) Collman, J. P. Inorg. Chem. 1997, 36, 5145-5155.
(35) Nappa, M. J.; Tolman, C. A. Inorg. Chem. 1985, 24, 4711-4719.
(36) Naruta, Y.; Tani, F.; Ishihara, N.; Maruyama, K. J. Am. Chem. Soc.
1991, 113, 6865-6872.
(44) Lyons, J. E.; Ellis, P. E., Jr. Catal. Lett. 1991, 8, 45-51.
(45) Ellis, P. E.; Jr., Lyons, J. E. Coord. Chem. ReV. 1990, 105, 181-193.
(46) Lyons, J. E.; Ellis, P. E., Jr.; Myers, H. K., Jr.; Wagner, R. W. J.
Catal. 1993, 141, 311-315.
(37) Lee, K. A.; Nam, W. J. Am. Chem. Soc. 1997, 119, 1916-1922.
(38) Urano, Y.; Higuchi, T.; Hirobe, M.; Nagano, T. J. Am. Chem. Soc.
1997, 119, 12008-12009.
(47) Wijesekera, T. P.; Lyons, J. E.; Ellis, P. E., Jr. Catal. Lett. 1995, 36,
69-73.

(Porphinato) iron-Catalyzed Isobutane Oxidation
Inorganic Chemistry, Vol. 39, No. 15, 2000 3127
were performed by Robertson Microlit Laboratories, Inc. (Madison,
NJ). Mass spectra were performed at the Mass Spectrometry Center at
the University of Pennsylvania.
mechanism accounted for the observation that primary C-H
bonds were inert to functionalization under these reaction
conditions.
General Procedure for the Preparation of (Porphinato)iron(II)-
(3-Fluoropyridine)2 Derivatives. The appropriate PFe-(py)2 species
(∼0.1 mmol) was dissolved in 3-fluoropyridine (3-F-py) (3 mL, 35.0
mmol) and brought to reflux under a N2 atmosphere in a 50 mL Schlenk
flask equipped with a condenser for 3 h. The solvent was evaporated
in vacuo, and the crude PFe-(3-F-py)2 complex was purified by
recrystallization from hot hexanes.
[5,10,15,20-Tetrakis(heptafluoropropyl)porphinato]iron(II)-(3-
F-py)2. Isolated yield: 95 mg (95% based on 100 mg of (C3F7)4PFe-
(py)2). 19F NMR (400 MHz, C6D6, 20 °C): δ -79.61 (s, 12F), -84.69
(br, 8F), -118.66 (br s, 8F), -125.28 (s, 2F). UV-vis (C6H6) [λmax
(nm) (ꢀ (M-1 cm-1))]: 340 (4.47), 418 (4.95), 542 (3.76), 575 (4.28).
[5,10,15,20-Tetrakis(pentafluorophenyl)porphinato]iron(II)-(3-
F-py)2. Isolated yield: 95 mg (95% based on 100 mg of (C6F5)4PFe-
(py)2). 19F NMR (400 MHz, C6D6, 20 °C): δ -125.95 (s, 2F), -138.14
(s, 8F), -152.15 (t, 4F), -162.15 (d, 8F). UV-vis (C6H6) [λmax (nm)
(ꢀ (M-1 cm-1))]: 312 (4.39), 416 (5.26), 524 (4.07), 553 (3.75).
[2,3,7,8,12,13,17,18-Octabromo[5,10,15,20-tetrakis(pentafluo-
rophenyl)porphinato]]iron(II)-(3-F-py)2. Isolated yield: 95 mg (95%
based on 100 mg of Br8(C6F5)4PFe-(py)2). 19F NMR (400 MHz, C6D6,
25 °C): δ -123.97 (s, 2F), -138.77 (s, 8F), -151.27 (t, 8F), -162.78
(d, 2F). UV-vis (C6H6) [λmax (nm) (ꢀ (M-1 cm-1))]: 345 (4.40), 452
(5.12), 556 (3.70), 588 (3.74).
To provide new insight into this mechanistic controversy, we
brought to bear two types of in situ spectroscopy to probe the
kinetics of these reactions and to examine the nature of products,
reactants, and PFe species present throughout the oxidation
reaction. This work couples in situ NMR spectroscopic experi-
ments51 with electronic absorption spectroscopic analysis of
oxidation reactions carried out in autoclave reactors; such studies
provide information distinct from that obtained previously from
product analysis of the oxygenated hydrocarbons derived from
reactions of hydrocarbons with dioxygen in the presence of PFe
catalysts.
In this paper, we examine the extent to which catalyst
electronic structure impacts catalyst stability and reactivity, as
well as product distribution, in the catalytic oxygenation of
isobutane, probing three different electron-deficient porphyrin
ligand environments: 5,10,15,20-tetrakis(pentafluorophenyl)-
porphyrin [(C6F5)4PH2], 2,3,7,8,12,13,17,18-octabromo-5,10,-
15,20-tetrakis(pentafluorophenyl)porphyrin [Br8(C6F5)4PH2], and
5,10,15,20-tetrakis(heptafluoropropyl)porphyrin [(C3F7)4PH2].
We present evidence for a radical chain autoxidation mechanism,
in which (porphinato)iron(III)-OH (PFe-OH) species not only
are responsible for the breakdown of the tert-butyl hydroper-
oxides generated in situ during the catalytic reaction, but also
play the role of radical chain initiator in the autoxidation process.
Synthesis of Four-Coordinate (C3F7)4PFe. ZnHg amalgam was
prepared using a procedure described previously by Scheidt.57 Activated
Zn turnings (600 mg, 9.23 mmol) were added to ∼ 2 mL of freshly
distilled Hg, and the mixture was stirred under a N2 atmosphere for 1
h. (C3F7)4PFe-Cl (15 mg, 0.014 mmol) was dissolved in 2.0 mL of
dry, degassed C6D6 and transferred to the ZnHg suspension. The mixture
was stirred for 30 min, at which time the color of the solution changed
from deep purple to dark green. The solution was filtered under N2
Experimental Section
Materials. Inert atmosphere manipulations and solvent purification
were carried out as described previously.52 Methods detailing the
preparation of 2,3,7,8,12,13,17,18-octabromo-5,10,15,20-tetrakis(pen-
tafluorophenyl)porphyrin,27 5,10,15,20-tetrakis(pentafluorophenyl)por-
phyrin,53 and 5,10,15,20-tetrakis(heptafluoropropyl)porphyrin54 have
been described previously. The syntheses of 2,3,7,8,12,13,17,18-
octabromo[5,10,15,20-tetrakis(pentafluorophenyl)porphinato]iron(III)
chloride and [5,10,15,20-tetrakis(pentafluorophenyl)porphinato]iron(III)
chloride followed those described in the literature.27,53,55 The syntheses
of [5,10,15,20-tetrakis(heptafluoropropyl)porphinato]iron(II)-(pyri-
dine)2, and {[5,10,15,20-tetrakis(heptafluoropropyl)porphinato]iron-
(III)}2O have been reported recently;56 [2,3,7,8,12,13,17,18-octabromo-
5,10,15,20-tetrakis(pentafluorophenyl)porphinato]iron(II)-(pyridine)2
and [5,10,15,20-tetrakis(pentafluorophenyl)porphinato]iron(II)-(pyri-
dine)2 were synthesized using an experimental procedure similar to that
described for [5,10,15,20-tetrakis(heptafluoropropyl)porphinato]iron-
(II)-(pyridine)2.56 Chemical shifts for 1H NMR spectra are relative to
residual protium in the deuterated solvents (C6D6, δ ) 7.15 ppm), while
19 F NMR chemical shifts are reported relative to fluorotrichloromethane
(CFCl3, δ ) 0.00 ppm). Chromatographic purification of these
compounds was accomplished on the benchtop using neutral alumina
(Fisher Scientific, Brockmann Activity I, 200 mesh). Elemental analyses
1
and transferred to an NMR tube fitted with a Teflon valve. H NMR
(200 MHz, C6D6, 23 °C): δ -2.80 (s, 8 H). 19F NMR (200 MHz, C6D6,
23 °C): δ -76.84 (s, 12 F); -87.57 (s, 8 F); -106.98 (br s, 8 F).
Instrumentation. Electronic spectra were recorded on an OLIS UV-
vis-near-IR spectrophotometry system that is based on the optics of a
Carey 14 spectrophotometer. Variable-temperature and in situ high-
pressure NMR spectra were recorded on a Varian 400 MHz spectro-
photometer. High-pressure NMR experiments were performed using
single-crystal sapphire NMR tubes; their design has been described
elsewhere.58 GC analyses of our catalytic oxidation reactions were
accomplished using a Perkin-Elmer gas chromatographic autosystem
equipped with a 30 mm × 0.53 mm × 2.65 µm thickness Hewlett-
Packard HP1 cross-linked methyl silicone column. A 125 mL Parr
reactor equipped with a magnetic stirrer was used for the autoclave
oxidations.
Autoclave Reactor Oxidation Studies. A 125 mL autoclave reaction
vessel, lined with a glass sleeve, was charged with 20 mL of a 0.25
mM solution of the PFe catalyst in the appropriate solvent and a known
amount of isobutane. The vessel was then heated to 80 °C, and once
the temperature had equilibrated, oxygen (120 psi) was introduced into
the system. An oxygen gas feed vessel was utilized to maintain constant
pressure in the reactor throughout the course of the catalytic oxidation.
Aliquots (2 mL) were removed periodically from the autoclave reactor
for analysis; electronic absorption spectroscopy (25 °C) determined the
nature of the PFe species present as a function of time, while GC
experiments enabled identification and quantification of the organic
products generated in the reaction. The reaction was terminated when
dioxygen ceased to be consumed. After depressurization and thorough
cleaning of the autoclave, the vessel was charged with solvent,
isobutane, and oxygen and heated to 80 °C for 24 h to confirm that no
substrate oxidation took place in the absence of catalyst. If trace catalytic
activity due to reactor impurities was observed, the reactor was
repetitively recleaned and retested with the catalyst-free (blank) system
(48) Grinstaff, M. W.; Hill, M. G.; Labinger, J. A.; Gray, H. B. Science
1994, 264, 1311-1313.
(49) Labinger, J. A. Catal. Lett. 1994, 26, 95-99.
(50) Boettcher, A.; Birnbaum, E. R.; Day, M. W.; Gray, H. B.; Grinstaff,
M. W.; Labinger, J. A. J. Mol. Catal., A 1997, 117, 229-242.
(51) Moore, K. T.; Horva´th, I. T.; Therien, M. J. J. Am. Chem. Soc. 1997,
119, 1791-1792.
(52) DiMagno, S. G.; Lin, V. S.-Y.; Therien, M. J. J. Org. Chem. 1993,
58, 5983-5993.
(53) Adler, A. D.; Longo, F. R.; Finarelli, J. D.; Goldmacher, J.; Assour,
J.; Korsakoff, J. J. Org. Chem. 1967, 32, 476.
(54) DiMagno, S. G.; Williams, R. A.; Therien, M. J. J. Org. Chem. 1994,
59, 6943-6948.
(55) Birnbaum, E. R.; Hodge, J. A.; Grinstaff, M. W.; Schaefer, W. P.;
Henling, L.; Labinger, J. A.; Bercaw, J. E.; Gray, H. B. Inorg. Chem.
1995, 34, 3625-3632.
(57) Safo, M. K.; Nesset, M. J. M.; Walker, F. A.; Debrunner, P. G.;
Scheidt, W. R. J. Am. Chem. Soc. 1997, 119, 9438-9448.
(58) Horva´th, I. T.; Ponce, E. C. ReV. Sci. Instrum. 1991, 62, 1104-1105.
(56) Moore, K. T.; Fletcher, J. T.; Therien, M. J. J. Am. Chem. Soc. 1999,
121, 5196-5209.

3128 Inorganic Chemistry, Vol. 39, No. 15, 2000
Moore et al.
Figure 1. Electron-deficient (porphinato)iron structures.
until no hydrocarbon oxidation products could be observed for the
control reaction.
High-Pressure NMR Studies of Catalytic Hydrocarbon Oxida-
tion. Sapphire NMR tubes were charged with a 0.5 mM solution of
the PFe catalyst in benzene-d6 under argon in a glovebox. The tube
was cooled in a liquid N2 bath; isobutane was then condensed in the
chilled vessel to give a hydrocarbon concentration of ∼7 M. After the
tube was warmed to room temperature, oxygen gas was added at a
pressure of ∼160 psi, providing an isobutane-rich system possessing
an 8:1 molar ratio of isobutane/O2. The NMR tubes were heated to 80
°C, and the fate of the PFe species was monitored by 19F or 1H NMR;
Figure 2. Variable-temperature 19F NMR spectra of (C3F7)4PFe-(3-
F-py)2 in toluene-d8 under Ar. L represents the 19F signal for the 3-F-
py axial ligand, while R, â, and γ refer to the 19F signals that derive
from the meso-perfluoropropyl groups.
the hydrocarbon oxidation products were conveniently analyzed by 13
C
NMR. While these NMR tube reactions were run until catalysis ceased,
all such reactions exhibited isobutane conversions that ranged from
5% to 10%, corresponding to incorporation of 40-80% of the O2
present in hydrocarbon oxidation products.
complexes in noncoordinating solvents that lack a sufficient
quantity of an appropriate Lewis base.
Following isolation of the PFe-(py)2 complexes, facile
replacement of the pyridine ligands with 3-fluoropyridine (3-
F-py) was accomplished by dissolving the PFe-(py)2 compound
in neat 3-F-py and heating at reflux for 3 h. This axial ligand
was chosen since it provides a convenient 19F NMR handle for
our spectroscopic studies of the catalytic isobutane oxidation
reaction (vide infra). While binding constants for PFeIII com-
plexes vary with the electronic properties of the axial ligands,
it has been shown for conventional porphyrin macrocycles that
the absolute magnitude of the analogous ligand binding constants
for the corresponding PFeII derivatives does not depend ap-
preciably on the acidity constant for the BH+/B couple of the
axial Lewis base; PFe-L2 formation constants are in fact
relatively uniform for axial ligands having pKa values that range
from 0.7 to 10.1.59
NMR Spectroscopy. Figure 2 displays temperature-depend-
ent 19F NMR spectral data obtained for (C3F7)4PFe-(3-F-py)2
in toluene-d8 over a -60 to +80 °C temperature domain;
detailed, tabulated variable-temperature (VT) 19F NMR data for
(C3F7)4PFe-(3-F-py)2, (C6F5)4PFe-(3-F-py)2, and Br8(C6F5)4-
PFe-(3-F-py)2 complexes are presented in the Supporting
Information. For the (C3F7)4PFe-(3-F-py)2 complex, four
distinct 19F NMR resonances are observed; the three meso-
heptafluoropropyl signals are denoted as R, â, and γ, while the
single 3-F-py axial ligand signal is labeled as L. The temper-
ature-dependent NMR behavior for this compound is complex.
Note that the R-CF2 and â-CF2 resonances broaden as the sample
is warmed from -60 to 0 °C, but progressively sharpen and
shift to higher field upon further increases in temperature. These
temperature-dependent shifts and changes in breadth of the C3F7
19F NMR signals derive from two factors: the magnitude of the
barrier to rotation of the CF2-Cmeso bond as a function of
coordination number, and the fact that diamagnetic [(C3F7)4-
tert-Butyl Hydroperoxide Decomposition Studies. (A) Electronic
Absorption Spectroscopy. Benzene solutions of the PFe catalysts (10-5
M) were transferred to a Schlenk-style optical cell sealed under N2; a
sufficient quantity of a 5 M decane solution of tert-butyl hydroperoxide
was added at ambient temperature to provide a 1000- 3000-fold molar
excess of this reagent with respect to the PFe complex. The fate of the
porphyrin catalyst was monitored by electronic absorption spectroscopy.
(B) NMR Spectroscopy. PFe catalysts were dissolved in toluene-
d8 (concentration 10-2 M) and transferred to an NMR tube under an
atmosphere of Ar; 1000 equiv of tert-butyl hydroperoxide was then
added to the tube at -198 °C. The hydroperoxide decomposition
reaction was monitored by 19F NMR spectroscopy at -70 °C. GC data
were analyzed for tert-butyl alcohol, acetone, and tert-butyl peroxide
as well as other possible oxidation products.
Isotopic Labeling Studies. Benzene solutions of the PFe catalyst
were treated with 5 µL of 17OH2 (∼20% enrichment) prior to
pressurization with isobutane and unlabeled oxygen in sapphire NMR
tubes. The NMR tubes were heated to 80 °C; 19F and 17O NMR spectra
were taken alternately to monitor both the state of the PFe catalyst and
degree of 17O incorporation into the oxidation products as a function
of time.
Results and Discussion
Synthesis of (C3F7)4PFe, (C6F5)4PFe, and Br8(C6F5)4PFe
Complexes. The three electron-deficient PFe systems utilized
in this study are displayed in Figure 1. All three macrocycles
are appended with electronegative substituents which stabilize
the FeII redox state in the presence of Lewis-basic axial ligands.
The corresponding low-spin PFe-(py)2 complexes can be
synthesized by the reaction of the respective free base porphyrin
(PH2) with an FeII salt in pyridine at reflux. The resulting ferrous
PFe-(py)2 complexes can be subsequently purified via chro-
matography on neutral alumina (eluant 9:1 CH2Cl2/pyridine)
in air and do not evince any observable oxidation to PFeIII
species. These ferrous complexes are robust in the solid state
and stable in aerated solutions at ambient temperature containing
excess pyridine, but they are irreversibly oxidized to PFeIII
(59) Walker, F. A.; Nasri, H.; Turowska-Tyrk, I.; Mohanrao, K.; Watson,
C. T.; Shokhirev, N. V.; Debrunner, P. G.; Scheidt, W. R. J. Am.
Chem. Soc. 1996, 118, 12109-12118.

(Porphinato) iron-Catalyzed Isobutane Oxidation
Inorganic Chemistry, Vol. 39, No. 15, 2000 3129
PFe-(3-F-py)2] is in equilibrium in solution with paramagnetic
56
five- and four-coordinate species.
Because coordinatively unsaturated PFeII-L or PFeII species
derived from PFeII-L2 catalyst precursors will be reactive with
dioxygen, we examined the temperature-dependent axial ligand
1
dynamics using VT H NMR ((C3F7)4PFe-(L)2, (C6F5)4PFe-
(L)2) and VT 19F NMR ((C3F7)4PFe-(L)2, (C6F5)4PFe-(3-F-
py)2, Br8(C6F5)4PFe-(3-F-py)2) spectroscopies (L ) 3-F-py,
pyridine). While these data (presented as Supporting Informa-
tion) will be analyzed in detail elsewhere, it is important to note
that all of these species show spectral behavior consistent with
the presence of significant concentrations of PFe species with
reduced coordination number (PFe and PFe-L complexes) at
80 °C.
One interesting aspect of these data is the apparent disparate
temperature-dependent axial ligand dynamical behavior that is
manifest as a function of macrocycle electronic structure;
(C3F7)4PFe binds ligands more avidly than (C6F5)4PFe and Br8-
(C6F5)4PFe, consistent with axial ligand binding constant data
determined from electrochemical experiments.56 Moreover, in
contrast to what has been observed for more conventional
porphyrin macrocycles,59 axial ligand binding affinity increases
markedly with increasing Lewis basicity of the axial donor for
(C3F7)4PFe-(L)2 species, consistent with the strongly σ-electron-
withdrawing nature of the macrocycle’s meso-perfluoroalkyl
groups.56
Figure 3. In situ 19F NMR spectra of a PFe-catalyzed isobutane
oxidation reaction in benzene-d6 that employed (C3F7)4PFe-(3-F-py)2
as the catalyst precursor. (A) (C3F7)4PFe-(3-F-py)2 at 80 °C under
Ar. (B) (C3F7)4PFe species present 2 days after pressurization with O2
in the absence of isobutane at 80 °C. (C) (C3F7)4PFe species observable
at 80 °C 1 h following the addition of isobutane/O2. (D) (C3F7)4PFe-
derived species that are observed while oxidation products are being
produced (reaction time 96 h at 80 °C, spectrum taken at 26 °C). Prime
notation represents the 19F NMR signals that coincide with those
reported for [(C3F7)4PFe]2O.56 L represents the 19F NMR resonance for
the 3-F-py axial ligand (δ ) -127.1 ppm), and X denotes the 19F NMR
signal corresponding to pyridine N-oxide; the signal denoted with an
asterisk is tentatively assigned to (C3F7)4PFe-OH.
To quantify the magnitude of the axial pyridyl binding
constant, four-coordinate (C3F7)4PFe was synthesized; 1H NMR
spectroscopy (Supporting Information) shows that the â-H atoms
resonate at -2.80 ppm at 23 °C and exhibit Curie behavior,
shifting to lower field with increasing temperature. Because K2
. K1 (eq 1) for nitrogenous PFeII axial ligands,60-62 and the
K1
K2
PFe-L2 y
\
z PFe-L + L y
\
z PFe + 2L
(1)
σ-electron- withdrawing nature of the (C3F7)PH2 macrocycle
engenders exceptional stability to the ferrous oxidation state
when the metal resides in the macrocycle plane (thus destabiliz-
ing (C3F7)4PFe(py) relative to both (C3F7)4PFe(py)2 and (C3F7)4-
PFe),56 the equilibrium constant for axial pyridyl binding can
be determined from this temperature-dependent NMR data,
giving Keq ) (3.0 ( 0.6) × 109 for the overall formation
constant (â2) of the six-coordinate complex at 23 °C; note that
an axial pyridyl binding constant of this magnitude exceeds by
over a 1000-fold that determined for ferrous centers of
conventional porphyrin macrocycles.63,64
High-Pressure NMR Spectroscopic Studies. NMR experi-
ments that examined the spectroscopic behavior of (C3F7)4PFe-
(3-F-py)2, (C6F5)4PFe-(3-F-py)2, and Br8(C6F5)4PFe-(3-F-py)2
catalyst precursors under high dioxygen pressure in the presence
and absence of isobutane substrate were carried out. Figures 3
and 4 detail these studies for systems utilizing (C3F7)4PFe-(3-
F-py)2 and (C6F5)4PFe-(3-F-py)2 catalyst precursors. In contrast
to earlier work that has emphasized the stability of Br8(C6F5)4-
PFe-based oxidation catalysts,40-45,65 under the experimental
conditions explicated in this paper, these catalysts show the least
Figure 4. In situ 19F NMR spectra of a (C6F5)4PFe-(3-F-py)2-catalyzed
isobutane oxidation reaction in benzene-d6. (A) (C6F5)4PFe-(3-F-py)2
at 80 °C under Ar. (B) (C6F5)4PFe species present 2 days after
pressurization with O2 in the absence of isobutane at 80 °C. (C) (C6F5)4-
PFe species observable at 80 °C 1 h following addition of O2/isobutane.
(D) 19F NMR spectrum taken after oxidation products are first observed
(reaction time 48 h at 80 °C; spectrum taken at 26 °C). Prime notation
represents the 19F NMR resonances corresponding to [(C6F5)4PFe]2O.
The double prime notation denotes the 19F NMR signals assignable to
(C6F5)4PFe-OH. The triple prime notation denotes resonances that are
tentatively ascribed to a (C6F5)4PFe-(O-tert-butyl) complex. L signifies
the 3-F-py axial ligand resonance.
stability, and evince (in contrast to analogous data obtained for
reactions that employed (C3F7)4PFe-(3-F-py)2 and (C6F5)4PFe-
(3-F-py)2 catalyst precursors) NMR signals that could not readily
be ascribed to well-characterized PFe complexes. As such,
detailed analyses of NMR data obtained for Br8(C6F5)4PFe-
catalyzed reactions are not possible; these spectra are presented
for comparative purposes in the Supporting Information.
Reactivity of (Porphinato)iron Species with Dioxygen.
Sapphire NMR tubes containing the (C3F7)4PFe-(3-F-py)2 and
(C6F5)4PFe-(3-F-py)2 catalyst precursors were prepared under
argon, and their 19F NMR spectra were recorded at 80 °C
(Figures 3A and 4A). The tubes were then pressurized with
dioxygen (∼160 psi); the 19F NMR spectra shown in Figures
(60) Wayland, B. B.; Mehne, L. F.; Swartz, J. J. Am. Chem. Soc. 1978,
100, 2379-2383.
(61) Rougee, M., Brault, D. Biochem. Biophys. Res. Commun. 1973, 55,
1364-1369.
(62) Rougee, M.; Brault, D. Biochemistry 1975, 14, 4100-4106.
(63) Kadish, K. M.; Bottomley, L. A. Inorg. Chem. 1980, 19, 832-836.
(64) Nesset, M. J. M.; Shokhirev, N. V.; Enemark, P. D.; Jacobson, S. E.;
Walker, F. A. Inorg. Chem. 1996, 35, 5188-5200.
(65) Lyons, J. E.; Ellis, P. E., Jr.; Durante, V. A. Stud. Surf. Sci. Catal.
1991, 67, 99-116.

3130 Inorganic Chemistry, Vol. 39, No. 15, 2000
Moore et al.
3B and 4B were manifest immediately at 80 °C. No further
spectral changes were observed over a one-week period at 80
°C. Figure 3B reveals two new distinct resonances for the R-CF2
fluorines, one of which is assigned to [(C3F7)4PFe]2O (prime
notation).56 Figure 4B shows that a mixture of PFe-OH (double
prime label) and [PFe]2O (single prime label) exists for the
(C6F5)4PFe system under dioxygen; note that multiple species
are present under these conditions for the Br8(C6F5)4PFe system
(Supporting Information). In these NMR experiments, it is
important to emphasize that, at 80 °C under the dioxygen
pressure utilized, ferrous porphyrin species are not observed
for any of these electron-deficient ligand environments. All three
PFe-(3-F-py)2 systems evince (i) complete loss of the axial
ligand following O2 pressurization, (ii) an NMR signal that
integrates as 2 equiv of the free 3-F-py ligand, and (iii) the
quantitative conversion to exclusively PFeIII species after O2 is
added, suggesting that the oxidizing agent involved in the
ferrous-to-ferric conversion derives solely from dioxygen.
Analogous experiments utilizing (C3F7)4PFe-(3-F-py)2 catalyst
precursor concentrations as high as 0.1 M (data not shown)
reveal stoichiometric conversion to the PFeIII species highlighted
in Figure 3B under similar O2 pressures, effectively ruling out
solvent impurities (e.g., peroxides) playing a significant role in
PFeII oxidation. Hence, these experiments show that while Br8-
(C6F5)4PH2 and (C3F7)4PH2 similarly stabilize the ferrous oxida-
tion state,56 neither class of these exceptionally electron-poor
PFeII complexes is thermodynamically stable under moderate
O2 pressure at 80 °C.
Differences between the19F NMR spectra of Figures 3B and
4B lie in the identities and distribution of the predominant PFeIII
species present under dioxygen; the nature of these PFeIII species
clearly depends on the steric and electronic properties of the
macrocyclic ligands. In Figure 3, we observe conversion of the
ferrous porphyrin complex to [(C3F7)4PFe]2O µ-oxo dimer upon
O2 pressurization, the spectral properties of which have been
previously reported.56 The resonance labeled with an asterisk
in Figure 3B has not been conclusively identified; it is likely,
however, that it corresponds to (C3F7)4PFe-OH, though such
a species has yet to be independently isolated and characterized.
Similarly, when solutions of (C6F5)4PFe-(3-F-py)2 are pres-
surized with dioxygen at 80 °C, a mixture of [(C6F5)4PFe]2O
and (C6F5)4PFe-OH species55 exist in nearly equal proportion
(Figure 4B). Due to steric factors that derive from the Br8(C6F5)4-
PH2 macrocycle’s highly saddled structure, the formation of a
[Br8(C6F5)4PFe]2O µ-oxo dimer species66 is precluded; hence,
all ferric porphyrin species present when solutions of Br8(C6F5)4-
PFe-(3-F-py)2 are pressurized with dioxygen at 80 °C (Sup-
porting Information) must correspond to mononuclear ferric
complexes.
infra). Nevertheless, these NMR experiments provide the first
direct analysis of the predominant PFe species that exist when
the catalytic conversion of isobutane to isobutanol commences.51
Furthermore, they illuminate the time-dependent spectral changes
that take place for these three classes of electron-deficient PFe
species while isobutane is being oxidized. Concomitantly
obtained 13C and/or 17O NMR data allow us to chronicle the
time-dependent distribution of organic products, as well as
determine the approximate times that the catalytic oxidation
reaction initiates and ceases.
Figures 3C and 4C show the 19F NMR spectral data obtained
for the systems that, respectively, employ (C3F7)4PFe-(3-F-
py)2 and (C6F5)4PFe-(3-F-py)2 catalyst precursors, once pres-
surization with isobutane and dioxygen has taken place. Note
that Figure 4C indicates that, upon isobutane addition, the
concentration of (C6F5)4PFe-OH has diminished, and at least
one new PFeIII species is present. On the basis of the similarity
of the chemical shifts of its resonances to those of (C6F5)4PFe-
OH, we tentatively assign this complex as a (C6F5)4PFe-(O-
tert-butyl) species. Whatever the nature of this compound, we
have to date only observed it under catalytic conditions. Figures
3D and 4D reveal the PFe species observable in solution at the
onset of catalysis, which we define as the point at which the
first evidence of oxidation products is observable in the 13C
NMR spectrum. Comparison of the 19F NMR data presented in
panels C and D of Figures 3 and 4 with the corresponding 13
C
NMR spectral data utilized to monitor the isobutane oxidation
products (data not shown) make apparent several general trends:
(i) No oxidation products are observed immediately upon tube
pressurization with isobutane and dioxygen (Figures 3C and 4C);
note also that only PFeIII species are observable at this stage in
each of these three systems. (ii) In each experiment, an induction
period is observed during which no further PFe spectral changes
are evident in the 19F NMR spectrum and no oxidation products
are detectable by 13C NMR spectroscopy; for Figures 3 and 4,
the induction period spans the time between the acquisition of
the NMR spectra C and D. (iii) For the (C3F7)4PFe- and (C6F5)4-
PFe-based catalysts, the length of the induction period ranges
from several hours to several days; for catalytic oxidations
employing Br8(C6F5)4PFe-(3-F-py)2 catalyst precursors, the
observed induction periods are typically less than 1 h (Sup-
porting Information). When (C3F7)4PFe- and (C6F5)4PFe-based
catalysts are employed, the duration of the induction period
increases at higher PFe concentrations; in contrast, the induction
periods for Br8(C6F5)4PFe-catalyzed isobutane oxidations are
insensitive to the concentration of the Br8(C6F5)4PFe-(3-F-py)2
catalyst precursor (data not shown). These experiments dem-
onstrate that the duration of the induction period is a function
of the nature of the catalyst precursor; for a given PFe-L2
concentration, the length of the induction period varies with
the steric properties of the PFe species [(C3F7)4PFe > (C6F5)4-
PFe . Br8(C6F5)4PFe]. Given the apparent dependence of the
induction period upon the extent of steric encumbrance at the
macrocycle periphery, it is thus plausible that greater concentra-
tions of [PFe]2O relative to PFe-OH at the onset of catalysis
are responsible for longer induction periods, congruent with the
supposition that the [PFe]2O dimer is catalytically inert.51
High-Pressure NMR Spectroscopic Studies of PFe-
Catalyzed Isobutane Oxidation Reactions that Utilize Di-
oxygen. Sapphire NMR tubes were charged with benzene-d6, a
PFe-(3-F-py)2 complex, isobutane, and oxygen; these tubes
were then heated at 80 °C (see the Experimental Section);
substrate oxidation products were monitored and quantitated by
13C or 17O NMR spectroscopy, while 19F NMR was used to
characterize the porphyrinic species present in solution. Since
substrate-to-product conversions could be rate limited by O2
transfer from the gas to solution phase across the small gas-
liquid interface of the NMR tube, a more detailed analysis of
the product distribution with respect to time was obtained for
isobutane oxidation reactions carried out in autoclaves (vide
Following the induction period, tert-butyl peroxide, acetone,
and tert-butyl alcohol were observed as the dominant oxidation
products for each catalyst. Once oxidation is initiated, new
porphyrinic resonances are observed in the 19F NMR spectra
(Figures 3D and 4D; Supporting Information); these signals
diminish slowly in intensity during the course of the reaction,
indicating that the benzene-soluble PFe species are decomposing
(66) Grinstaff, M. W.; Hill, M. G.; Birnbaum, E. R.; Schaefer, W. P.;
Labinger, J. A.; Gray, H. B. Inorg. Chem. 1995, 34, 4896-4902.

(Porphinato) iron-Catalyzed Isobutane Oxidation
Inorganic Chemistry, Vol. 39, No. 15, 2000 3131
during isobutane oxidation. For all three catalyst systems, the
oxidation ceases at the time when no porphyrin species could
be observed by NMR; at this point in time, typically 40-80%
of the O2 present originally in the NMR tube had been
consumed. Importantly, these results show clearly that even the
perhalogenated dodecasubstituted porphyrin macrocycle, Br8-
(C6F5)4PH2, is not robust under the conditions of catalytic
isobutane oxidation; this is contrary to the supposition that a
macrocycle void of C-H bonds would necessarily be rendered
less susceptible to oxidative degradation, as well as to a number
of other previous reports.40-45,65
and electronic properties; instead the concentration of the PFe
species utilized and the nature of the solvent appear to be much
more important factors. Entry 3 of Table 1B, which describes
a catalytic system with a PFe concentration 20-fold greater than
that of the other experiments listed therein, shows a much higher
relative percentage of the tert-butyl peroxide oxidation product;
it is also noteworthy that, at this catalyst concentration,
drastically reduced turnover numbers are observed. In further
agreement with a hydrocarbon autoxidation mechanism49 is the
observed dependence of the overall conversion of isobutane to
products upon reaction time, which also appears independent
of macrocycle electronic properties.
Batch Reactor Studies on PFe-Catalyzed Isobutane Oxi-
dation Reactions That Utilize Dioxygen. To compliment the
in situ NMR experiments, an analogous set of isobutane
oxidation reactions were performed in a glass-lined autoclave.
Unlike the NMR experiments, the batch reactor studies enable
detailed kinetic analyses of the catalytic isobutane oxidation
reactions and, importantly, direct monitoring of dioxygen
consumption throughout an autoclave reactor oxidation, which
is experimentally untenable in the NMR tube reactions. Control
experiments in which the autoclave reactors are charged with
benzene, isobutane, and oxygen were performed to verify that
no oxidation took place in the absence of a PFe complex; if
any activity was observed in the control experiments, the
autoclave was repetitively cleaned and retested until no detect-
able hydrocarbon oxidation products were observable in the
catalyst-free oxidation experiment.
As would be expected from a radical process dominating PFe-
catalyzed isobutane oxidation, irreproducibility from experiment
to experiment is observed with respect to the length of the
induction periods, turnover frequency, and product distribution.
For example, Figure 5 shows a plot of turnover frequency vs
reaction time for four independent Br8(C6F5)4PFe-catalyzed
isobutane oxidation reactions carried out under identical condi-
tions. Note that the initial turnover frequency, the rate at which
this turnover frequency diminishes, and the final turnover
frequency differ for each catalytic run, despite the extreme care
taken to hold all controllable experimental parameters constant.
Because such discrepancies exist between catalytic oxidations
carried out under presumably identical conditions, little can be
concluded regarding any potential roles that catalyst electronic
properties (Table 1) play in these reactions.
In addition to the major oxidation products described above,
there are two oxidation products characterized in gas-phase GC
analyses of these reactions that are produced in measurable
quantity: CO2 and CO. It is plausible that these products derive
from the methyl radical generated in the conversion of isobu-
tyloxy radical to acetone (vide infra).
Typically, the reactors were charged with benzene, PFe-(3-
F-py)2, and isobutane and heated to 80 °C with stirring. The
only reaction conditions that differed between the catalytic
oxidation experiments carried out in NMR tubes and the
autoclave reactors were catalyst concentration (NMR experi-
ments, [PFe] ) 0.5 mM; autoclave oxidations, [PFe] ) 0.25
mM) and the fact that the reactor utilized a continuous O2 feed
that maintained constant pressure. After oxygen was introduced
to the reactor, an immediate reduction in pressure in the oxygen
gas feed vessel was observed, indicating the dissolution of O2
in the reaction medium. Samples were drawn from the reactor
at various time intervals and analyzed by GC and electronic
absorption spectroscopy. Parts A-C of Table 1A-C show the
batch reactor data for reactions charged, respectively, with
(C3F7)4PFe, (C6F5)4PFe, and Br8(C6F5)4PFe catalyst precursors.
Product distribution, turnover frequency, and conversion (%)
were consistent with those observed in the NMR tube experi-
ments; the observed induction periods, however, were signifi-
cantly shorter in the autoclave experiments relative to those
observed for the NMR tube reactions. The shorter induction
periods for the autoclave reactions may derive from better
mixing of O2 with solvent in an autoclave reactor equipped with
a magnetic stirrer relative to an NMR tube which was neither
stirred nor spun. It should be noted that the larger concentration
of the PFe catalyst in the NMR tube experiments was neces-
sitated by the practical requirement of minimizing the acquisition
time required to obtain an adequate signal-to-noise ratio.
(Porphinato)iron optical absorption properties are sensitive
to macrocycle electronic structure, the oxidation state of the
iron center, and the number and nature of axial ligands. Figure
6A shows the time-dependent spectral changes that occur
throughout an oxidation reaction carried out in an autoclave
that was initially charged with (C3F7)4PFe-(py)2, benzene, and
isobutane, under conditions described previously; at t ) 0, three
intense transitions are observed at 418, 542, and 578 nm (log ꢀ
) 4.95, 3.76, and 4.28, respectively).56 Consistent with NMR
tube experiments, at 80 °C following pressurization with
dioxygen, quantitative conversion of the PFeII complex to PFeIII
species is evident in the t ) 1 h electronic spectrum of Figure
6A, which displays broad Soret features at 378 and 420 nm
and Q-band absorbances at 560 and 602 nm; this spectrum
corresponds to that previously reported for [(C3F7)4PFe]2O.56
Similar to the NMR experiments discussed above, the t ) 1 h
spectrum of Figure 6A is observed as soon as an optical
spectrum can be obtained after the autoclave is charged with
isobutane and dioxygen (t ≈ 5 min), indicating that the PFeII-
to-PFeIII conversion is rapid. Furthermore, as was shown in
Figures 3 and 4, catalyst degradation occurs after the onset of
catalytic activity; this is marked by a steady decrease in
absorption oscillator strength throughout the 350 f 650 nm
spectral regime during the time domain over which isobutane
oxidation is observed (Figure 6A). The 2 and 20 h spectra of
Figure 6A show that a (C3F7)4PFe-derived absorbing species is
present while catalytic isobutane oxidation is taking place; the
fact that these spectra cannot be fit as an appropriately weighted
sum of the optical spectra of independently characterized PFeII,
As seen in Table 1, three primary oxidation products were
obtained in the autoclave reactor studies: tert-butyl alcohol,
acetone, and tert-butyl peroxide. The distribution of these
oxidation products is consistent with that observed in the NMR
tube experiments; importantly the observed tert-butyl alcohol/
acetone/tert-butyl peroxide distribution is indicative of an
autoxidation process for isobutane oxidation as discussed by
Labinger and Gray.48,50,67 Congruent with such a mechanism,
the quantity and nature of isobutane oxidation products do not
appear to significantly depend on porphyrin macrocycle steric
(67) Birnbaum, E. R.; Grinstaff, M. W.; Labinger, J. A.; Bercaw, J. E.;
Gray, H. B. J. Mol. Catal., A 1995, 104, 119-122.

3132 Inorganic Chemistry, Vol. 39, No. 15, 2000
Moore et al.
Table 1. Time-Correlated Product Distribution of (C3F7)4PFe-, (C6F5)4PFe-, and Br8(C6F5)4PFe-Catalyzed Isobutane Oxidations at 80 °Ca
selectivity (%)e
reaction
no. of
conversiond tert-butyl
tert-butyl
alcohol acetone peroxide
entry
1
catalyst (mmol)
(C3F7)4PFe-(py)2 (0.005)
solvent
A. (C3F7)4PFe-Catalyzed Oxidation
time (h)b turnoversc
(%)
benzene
1
2
3
5
21
5
23.5
2
19
43
91
22
2034
4120
5422
9292
9974
460
7800
3010
4340
5720
22440
8340
3.38
6.84
9.15
16.51
16.98
0.77
13.99
4.99
7.40
9.87
35.54
12.33
84.2
83.9
82.3
76.4
80.5
78.3
69.5
81.0
78.3
76.9
84.0
93.3
7.0
7.4
7.2
7.2
8.0
4.3
5.4
6.0
5.5
5.6
8.1
3.6
8.8
8.7
10.5
16.4
11.5
17.4
25.1
13.0
16.1
17.5
7.8
2
3
(C3F7)4PFe-(py)2 (0.005)
benzene (dried over Na)
benzene
(C3F7)4PFe-(3-F-py)2 (0.005)
4
5
6
(C3F7)4PFe-(3-F-py)2 (0.005)
(C3F7)4PFe-(3-F-py)2 (0.005)
(C3F7)4PFe-Cl (0.005)
C7F14
3.1
9:1 C7F14/benzene
benzene
22
9320
13.94
90.8
5.6
3.6
18
90
114
1.5
17
91
2540
3440
8080
1400
5200
6560
3.65
4.92
11.47
2.27
8.52
11.00
91.3
89.5
90.6
90.0
88.5
85.4
7.1
9.3
8.9
2.9
3.5
4.0
1.6
1.2
0.5
7.1
8.1
7
1
[(C3F7)4PFe]2O (0.005)
benzene
10.7
B. (C6F5)4PFe-Catalyzed Isobutane Oxidation
(C6F5)4PFe-Cl (0.005)
benzene
1
2
4
21
28
44
121
2
1418
1870
2058
2438
3582
3914
6500
6166
13920
21774
0
2.77
3.74
4.18
5.01
7.39
8.09
12.74
9.56
21.48
34.08
0.00
64.0
61.1
59.3
57.9
57.4
56.9
64.4
87.1
86.3
83.0
4.5
4.4
4.3
4.0
4.0
4.0
3.8
7.7
8.9
10.7
31.5
34.5
36.4
38.1
38.6
39.0
31.8
5.3
2
3
(C6F5)4PFe-Cl (0.005)
(C6F5)4PFe-Cl (0.100)
20:1 benzene/water
benzene
5
71
2
4.8
6.3
20
72
5
47
19
43
70
84
74
290
2.92
55.4
60.7
81.4
78.1
82.6
80.1
79.0
79.9
4.1
4.5
7.4
7.4
4.9
4.9
5.1
5.3
40.5
34.8
11.2
14.5
12.5
15.0
15.9
14.8
10.97
13.93
18.40
13.15
15.42
18.84
24.90
4
5
(C6F5)4PFe-OH/[(C6F5)4PFe]2O (0.005)f
(C6F5)4PFe-OH/[(C6F5)4PFe]2O (0.005)
benzene
7840
10060
8140
9340
11320
15100
20:1 benzene/1.0 M aq NaOH
C. Br8(C6F5)4PFe-Catalyzed Isobutane Oxidation
1
Br8(C6F5)4Fe-Cl (0.005)
benzene
1
3
6
42
187
1894
2174
2472
1830
10640
3.57
4.12
4.74
3.29
15.02
22
77.0
76.4
75.0
71.0
73.6
85.0
5.9
5.8
6.0
1.7
5.0
8.5
17.1
17.8
19.0
27.2
21.4
7.0
2
3
Br8(C6F5)4PFe-(py)2 (0.005)
benzene
theoretical result for the radical chain mechanismg
a All catalytic oxidations except entry 1 of part C were performed to the exhaustive limit, which we define as the point in time where PFe species
are no longer observed by electronic spectroscopy and O2 consumption is negligible for a period of 8 h. Reaction conditions are elaborated in the
Experimental Section. b For reactions carried out to the exhaustive limit, the last reaction data point recorded per entry corresponds to the time at
which the production of oxidized hydrocarbons had completely ceased. c Turnovers are defined as total millimoles of product per millimole catalyst.
d Conversion (%) is defined as the total millimoles of product per millimole of isobutane times 100%. The amount of isobutane utilized in these
experiments ranged from 0.46 to 0.48 mol. e Selectivity is based on the relative intensities of the three oxidation products tert-butyl alcohol, acetone,
and tert-butyl peroxide in the GC analyses. Other oxidation products comprised <1% of the total integrated peak intensities for tert-butyl alcohol,
acetone, and tert-butyl peroxide in the GC experiments. f PFe-OH and (PFe)2O species exist as an equilibrium mixture; at ambient temperature,
in H2O-saturated benzene, the molar ratio of these PFe complexes is approximately 1:1. g See ref 49.
PFeII-L2, PFeIII-X, (PFeIII)2O, and PFeIII-L2 complexes
strongly suggests that porphyrin macrocycle destruction occurs
over the time course of this experiment; consistent with this,
considerable precipitate is observed at the reaction end point,
further supporting this hypothesis. Likewise, systems utilizing
(C6F5)4PFe-Cl and Br8(C6F5)4PFe-Cl catalyst precursors also
evince prominent changes in their electronic spectral features
once catalysis has initiated, along with similar diminishment
of absorption intensity in the visible and low-energy UV regions
of the electromagnetic spectrum (Figure 6B,C). Note that, under
these autoclave reactor conditions, complete bleaching of Br8-
(C6F5)4PFe-derived electronic absorption spectral features is
observed at t ) 20 h; consistent with the results from NMR
studies of the catalytic isobutane oxidation reaction, the Br8-
(C6F5)4PH2 macrocycle is seen to be the least robust under
catalytic conditions, contrasting previous reports of its superior
stability with respect to oxidative degradation.40-45,65 For the
PFe-catalyzed isobutane oxidations carried out in batch reactors,
it is important to note that regardless of the PFe catalyst
precursor employed, cessation of catalysis coincided with the

(Porphinato) iron-Catalyzed Isobutane Oxidation
Inorganic Chemistry, Vol. 39, No. 15, 2000 3133
visible region, showing that the ultimate porphyrin-derived
decomposition products are catalytically inactive.
While NMR and autoclave experiments show that the relative
stabilities of the ferrous oxidation state in these macrocycle
environments are surprisingly similar, there is no direct cor-
relation between macrocycle stability under catalytic conditions
and the thermodynamic stability of the ferrous oxidation state
[(E1/2[FeII/III] (benzonitrile, 1.0 M pyridine) vs SCE: Br8(C6F5)4-
PFe-(py)2, 780 mV; (C3F7)4PFe-(py)2, 540 mV; (C6F5)4PFe-
(py)2), 370 mV].56 This suggests that a common catalyst
deactivation mechanism may be operative, and implicates a role
for radicals formed during the catalytic oxidation reaction (vide
infra) in Fe-N, C-N, and/or C-C bond cleavage events.
It is likely that the enhanced reactivity reported previously
for isobutane oxidation conditions utilizing Br8(C6F5)4PFe-based
catalysts arises from two factors: (i) the high concentration of
PFe-OH initially present in the reactions, which correlates with
the observation that reactions utilizing this catalyst exhibit a
burst of substrate turnovers in the first few minutes of reaction
(Figure 5), and (ii) the fact that multiple 19F signals are observed
just after initiation of a Br8(C6F5)4Fe-catalyzed reaction (Sup-
porting Information). This latter point indicates that significant
degradation of the highly halogenated macrocycle occurs just
after catalysis commences, which raises the possibility that these
(porphinato)iron decomposition products may also serve as
initiators of isobutane autoxidation. Again, it is important to
note that the overall activity of this catalyst with respect to the
total turnovers of isobutane determined at the exhaustive limit
of the oxidation reaction are similar to those observed for
(C3F7)4PFe- and (C6F5)4PFe-catalyzed processes (Table 1),
indicating that the concentration of active Br8(C6F5)4PFe-based
catalysts must diminish rapidly with respect to time.
Figure 5. Turnover frequency vs reaction time for four independent
Br8(C6F5)4PFe-catalyzed isobutane oxidations that employed identical
reaction conditions ([Br8(C6F5)4PFe] ) 0.25 mM; [isobutane] ) 22.8
mM; PO ) 8 bar). Aliquots (2 mL) were taken out of the reaction
2
mixture at 5, 15, 30, 60, 120, 180, and 300 min intervals. Turnovers
were calculated from GC chromatographic analyses of the oxidation
products.
Axial Ligand and Solvent Effects on Catalyst Activity and
Stability. Axial ligand electronic properties play a key role in
the activation of dioxygen in heme-containing enzymes.1 It has
been demonstrated for synthetic metalloporphyrin catalysts,
particularly for those systems which utilize O atom donors to
generate reactive compound I analogues P•+FedOs(L), that
the nature of the metal axial ligand L influences both catalyst
reactivity and stability; for instance, the addition of imidazole
to a solution containing a PFe-Cl complex and excess organic
peroxyacid, generates a PFe catalyst capable of epoxidizing
olefins, while the high-valent PFe species generated in the
imidazole-free solution are unreactive with respect to unsaturated
hydrocarbons.7 Likewise, for metalloporphyrin-catalyzed oxida-
tions that consume oxidizing equivalents derived from dioxygen,
Ellis and Lyons have reported that the introduction of azides
as axial ligands for metalloporphyrin-catalyzed alkane oxidation
results in augmented catalytic activity.43,68
The temperature-dependent 19F and 1H NMR studies (Figure
2; Supporting Information) evince reversible axial ligand
dissociation for these PFe-L2 systems, and demonstrate that
the 3-F-py axial ligand resonance broadens due to exchange
with increasing temperature. The effect of exogenous pyridine
upon the catalytic isobutane oxidation reaction was probed in
both autoclave and NMR reaction vessels. As seen in entries 1,
6, and 7 of Table 1A, no significant differences in catalyst
reactivity or stability are observed for oxidation reactions in
which autoclaves were charged with (C3F7)4PFe-Cl and
[(C3F7)4PFe]2O species relative to those in which (C3F7)4PFe-
(py)2 served as the catalyst precursor. While the rate at which
PFe-(3-F-py)2 complexes are converted to PFeIII species
Figure 6. Time-dependent electronic absorption spectra of PFe-
catalyzed isobutane oxidation reactions carried out in autoclave reactors
when (A) (C3F7)4PFe-(py)2, (B) (C6F5)4PFe-Cl, and (C) Br8(C6F5)4-
PFe-Cl are utilized as catalyst precursors. Note the steady disappear-
ance of benzene-soluble PFe complexes throughout the course of
catalysis. Reactor conditions: T ) 80 °C, [PFe] ) 0.25 mM, PO2 ) 8
bar, [isobutane] ) 22.8 mM (see the Experimental Section for further
details). Optical spectra displayed were recorded at ambient temperature
and pressure.
point in time at which electronic spectral analysis of the reaction
medium indicated complete bleaching of all absorbances in the
(68) Lyons, J. E.; Ellis P. E., Jr.; Shaikh, S. N. Inorg. Chim. Acta 1998,
270, 162-168.

3134 Inorganic Chemistry, Vol. 39, No. 15, 2000
Moore et al.
decreases with increasing exogenous 3-F-py concentrations (19F
NMR and autoclave reactor data not shown), once complete
conversion of the PFe-L2 precursor to the ferric porphyrin
complex has taken place, a normal induction period is observed;
both time-dependent electronic absorption and 19F NMR reaction
profiles are essentially identical to those obtained for the
analogous reaction carried out in the absence of added axial
ligand. These results reveal that enhanced stability for the ferrous
oxidation state may be temporarily achieved by the addition of
exogenous axial ligand, since hexacoordination precludes
reactivity with dioxygen; augmented concentrations of 3-F-py,
however, do not play a significant role in catalytic activity,
product distribution, or ultimate catalyst lifetime.
[PFeIII]2O + 17OH2 h {FeIII-OH + PFeIII-17OH (2)
carried out in aqueous systems,69 17O NMR experiments
executed with an independently synthesized sample of (C6F5)4-
PFe-17OH verify that the exchange process of eq 2 is facile in
benzene. Interestingly, incorporation of the 17O label into the
oxidation products is observed when catalytic isobutane oxida-
tion reactions are carried out in the presence of 17OH2 (Table
2).
An alternating sequence of 17O NMR and 19F NMR spectra
were repetitively recorded in which 5 µL of 20% enriched 17OH2
was added to a 5 µM benzene solution of the PFeII catalyst
precursor prior to the sapphire NMR tube being charged with
isobutane and unlabeled dioxygen under conditions identical
to those described previously (Figure 7). As described above,
following NMR tube pressurization with isobutane and dioxy-
gen, 19F NMR spectroscopy reveals the complete and irreversible
formation of PFeIII species from the PFeII-L2 precursor, while
17O NMR initially shows only one signal corresponding to the
labeled water present in benzene-d6 (Figure 7, spectrum B). At
the point where new 17O signals corresponding to hydrocarbon
oxidation products are produced (spectrum F), a change in the
nature of porphyrinic species present in solution is concomitantly
observed (Figure 7, spectra E-G). The three major 17O NMR
signals in spectra F and H of Figure 7 correspond to acetone,
tert-butyl peroxide, and tert-butyl alcohol; the integrated, relative
intensities of the 17O NMR peaks are consistent with those
observed previously for analogous experiments in which 13C
NMR spectroscopy was used to monitor the relative amounts
of isobutane oxidation products. The fact that 17O signals are
observed in the oxidation products highlights the direct involve-
ment of water in the PFe-catalyzed oxidation of isobutane.
Two control experiments were performed to determine
whether any 17O oxygen incorporation into the hydrocarbon
oxidation products occurs following the PFe-dependent catalytic
reaction. Sapphire NMR tubes were charged with a benzene-d6
solution of (C6F5)4PFe-(3-F-py)2, 17O-labeled water, and either
tert-butyl alcohol or tert-butyl peroxide (Table 2). The tubes
were heated for a prolonged period (t ) 2 days at 80 °C) and
the organic products monitored by GC and 17O NMR spectro-
scopy. As seen in Table 2, the experiment involving tert-butyl
alcohol produced no other oxidized hydrocarbons; moreover,
no 17O incorporation into this alkanol was observed. However,
when the tube was charged with tert-butyl peroxide under these
conditions, acetone was generated with the incorporated 17O
label.70 Thus, on the basis of these control experiments, we
conclude that both tert-butyl alcohol and tert-butyl peroxide do
not undergo 17O exchange after their production in the isobutane
oxidation, and this exchange must take place prior to their
formation. Acetone, however, does exchange with water after
its formation, and is produced in significant quantity through
the decomposition of tert-butyl peroxide under the conditions
of catalytic isobutane oxidation.
Five different solvent systems were utilized for the PFe-
catalyzed isobutane oxidation reactions: benzene, perfluoro-
methylcyclohexane (C7F14), 9:1 C7F14/benzene, 20:1 benzene/
H2O, and 20:1 benzene/(1 M aqueous NaOH). The PFe catalyst
precursors and isobutane are highly soluble in both benzene
and C7F14; while the major oxidation products are very soluble
in benzene, their solubility is limited in C7F14. In general, little
difference can be noted in these reactions as a function of
solvent. Entries 3, 4, and 5 in Table 1A, however, relate the
relative turnover numbers, product distribution, and selectivity
for isobutane oxidation using a (C3F7)4PFe-(3-F-py)2 catalyst
precursor in benzene, C7F14, and a 9:1 C7F4/benzene mixture;
these results show that although the overall turnover numbers
and conversion (%) do not vary significantly as a function of
time in these reactions, the relative selectivity for tert-butyl
alcohol over tert-butyl peroxide is enhanced in the oxidation
reactions carried out with C7F14 solvent. Similarly, entries 2
and 3 of Table 1B seem to show that added H2O enhances the
selectivity of tert-butyl alcohol over tert-butyl peroxide. While
the data set is too limited to make broad conclusions regarding
the role played by solvent in determining the distribution of
oxidized hydrocarbons, the increased activity observed in
catalytic hydrocarbon oxidations carried out with added water
may derive from shifting the equilibrium described in Scheme
4 (equilibrium v, vide infra) toward PFe-OH. The lower level
of tert-butyl peroxide (and the corresponding higher level of
acetone) obtained for catalytic oxidations carried out in the
presence of water likely correlates with enhanced rates of tert-
butyl peroxide decomposition in this medium.
Role of Water in PFe-Catalyzed Isobutane Oxidation.
High-Pressure 17O NMR Spectroscopic Studies. Spectroscopic
evidence (vide supra) shows that PFe-OH is present under
catalytic conditions; production of this species would not be
possible without a proton or hydroxide source. Because water
would need only be present at levels approaching that of the
catalyst concentration to be implicated in PFe-OH production,
it is likely that sufficient moisture to do so would be present in
the solvents and apparatus used for these reactions. (Except for
entry 2 of Table 1A, no special care was taken to rigorously
dry the solvents or reaction vessels used in this study.) To better
understand the role played by water in PFe-catalyzed isobutane
oxidations, small amounts of 20% enriched 17OH2 were added
to either benzene-d6 solutions of [(C3F7)4PFe]2O or a [(C6F5)4-
PFe]2O/(C6F5)4PFe-OH equilibrium mixture under argon. Im-
mediately, 17O NMR reveals broadening of the free water signal
relative to that observed for an 17OH2 standard in benzene-d6
(data not shown), suggesting either 17OH2 association/dissocia-
tion is occurring with the PFe species present in solution on
the time scale of the NMR experiment, or a chemical exchange
process (eq 2) is taking place. Consistent with experiments
Role of Hydroperoxides in PFe-Catalyzed Isobutane
Oxidation. tert-Butyl Hydroperoxide Decomposition Experi-
ments. The catalytic decomposition of hydroperoxides by
halogenated PFe complexes has been previously studied.41-43
The major oxidation products reported in this reaction, namely,
(69) Ostrich, I. J.; Liu, G.; Dodgen, H. W.; Hunt, J. P. Inorg. Chem. 1980,
19, 619-621.
(70) Interestingly, no isobutanol was produced in this reaction, implying
that isobutane or tert-butyl hydroperoxide must be present to generate
this species, since the acetone generated in this experiment assuredly
derives from tert-butoxy radical.

(Porphinato) iron-Catalyzed Isobutane Oxidation
Inorganic Chemistry, Vol. 39, No. 15, 2000 3135
Table 2. Role of Water in PFe-Catalyzed Isobutane Oxidation:
Evidence for Oxygen Exchange from Water Prior to Product
Formationa
absorption spectroscopy, which contrasts the results reported
above for the isobutane oxidation experiments (Figure 4).51
The potential role played by tert-butyl hydroperoxide in a
PFe-catalyzed isobutane oxidation reaction was examined using
in situ 19F and 13C NMR methods for a catalytic reaction
employing a (C3F7)4PFe-(3-F-py)2 catalyst precursor; for these
reactions, the NMR tube was charged with the PFe-(3-F-py)2
complex, benzene-d6, and 1000 equiv of tert-butyl hydroper-
oxide prior to pressurization with isobutane and dioxygen
(Figure 10). Consistent with the data of Figure 3C, [(C3F7)4-
PFe]2O is the dominant PFe species present 30 min after the
isobutane oxidation reaction has initiated (Figure 10B). It is
noteworthy that if an induction period is taking place for this
reaction, it must be less than the 2 h time period required for
the acquisition of the spectrum of Figure 10B, since significant
quantities of organic products were detected in a subsequently
acquired 13C NMR spectrum (data not shown). The induction
periods for (C3F7)4PFe-catalyzed reactions carried out under
identical conditions without added hydroperoxide are minimally
8 h in duration. These data suggest that tert-butyl hydroperoxide
serves to initiate the isobutane oxidation reaction. After 2 days
at 80 °C, the spectrum in Figure 10C is evident. No further
time-dependent spectral changes were observed due to the fact
that all of the oxygen within the sapphire NMR tube was
completely consumed at this point in time.
major
oxidation products oxidation products
(GC or 13C NMR) (17O NMR)
17O-labeled
substrate
isobutane (400 mg)
tert-butyl alcohol tert-butyl alcohol
acetone acetone
tert-butyl peroxide tert-butyl peroxide
tert-butyl alcohol (control) (40 µL) tert-butyl alcohol none
tert-butyl peroxide (control)
(40 µL)
acetone
tert-butyl peroxide
acetone
a Experimental conditions: solvent ) benzene-d6; [(C3F7)4PFe-(3-
F-py)2] ) 1.0 mM; [17OH2] ) 0.14 M (20% 17O enrichment);
[isobutane] ≈7 M; O2 pressure ) ∼160 psi; T ) 80 °C; reaction time
) 2 days.
tert-butyl alcohol, acetone, and tert-butyl peroxide, are produced
in proportions similar to those observed in PFe-catalyzed
isobutane oxidations in which the oxidizing equivalents are
derived from dioxygen; thus, the decomposition of such
peroxides has been surmised to be important in the alkane
oxidation reaction mechanism.46,47,71 We have augmented this
earlier work by utilizing in situ electronic absorption and 19F
NMR spectroscopic methods to probe the importance of this
process in the PFe-catalyzed isobutane oxidation reaction and
monitor the spectral changes that PFe complexes undergo during
hydroperoxide decomposition.
Summary and Conclusions
Figure 8 shows the time-dependent electronic absorption
spectral changes that occur for the (C3F7)4P-, (C6F5)4P-, and
Br8(C6F5)4P-based iron catalyst systems in the presence of a
3000-fold molar excess of tert-butyl hydroperoxide. Isosbestic
behavior was observed for the Br8(C6F5)4PFe-catalyzed reactions
over a 1 h time period, highlighting the clean conversion of the
ferrous porphyrin species to a single PFeIII complex (Figure 8A).
In contrast, nonisosbestic behavior was observed in the analo-
gous experiment utilizing the (C6F5)4PFe catalyst (Figure 8B);
interestingly, subsequent absorption spectra taken after the initial
oxidation of (C3F7)4PFe (Figure 8C) show little change with
respect to the electronic absorption spectrum recorded at t )
10 min, indicating that these species may be a preferable
peroxide decomposition catalyst due to its superior stability
under these conditions. Note that, at 25 °C, no catalyst
destruction is observed in the electronic absorption spectra,
suggesting that the processes involved in PFe decomposition
at 80 °C under oxidative conditions are slow at this temperature.
A variety of mechanisms have been proposed for PFe-
catalyzed alkane oxidations that utilize dioxygen as the sto-
ichiometric oxidant. Initially, it was hypothesized that the
electron-deficient PFe species entered into a catalytic cycle
which involves a high-valent PFeIVdO intermediate.40-45,65
Labinger and Gray proposed recently that an autoxidation
reaction pathway likely dominated the kinetics; the primary role
of the metalloporphyrin involved decomposing the hydroper-
oxides produced during the autoxidation pathway.48-50 While
the observed ratios of organic products produced in catalytic
isobutane oxidation reactions did not coincide with that predicted
by a theoretical model for such an autoxidation,49 it was
emphasized that if a P450-type mechanism involving reactive
PFeIVdO species were to dominate, the selectivity toward tert-
butyl alcohol would have been much higher than what is
typically observed for these reactions.
Clearly, the mechanism by which the stoichiometric conver-
sion of low-spin PFe-L2 precursors to high-spin PFeIII species
is accomplished at 80 °C in the presence of O2 and isobutane
provides additional information regarding the PFe oxidation
states potentially important for catalysis; Scheme 3 shows three
established pathways by which PFeIII species can be produced
stoichiometrically from a PFeII precursor. Routes that generate
PFeIII requiring the production of superoxide by either a SN1
or SN2 process (pathway A) are not consistent with the
observation of induction periods in both the NMR tube and
autoclave experiments, as both these processes would produce
peroxide; addition of even trace amounts of peroxide to either
an autoclave or sapphire NMR tube isobutane oxidation reaction
results in instantaneous termination of the induction period and
the formation of organic products. Likewise, a water-assisted
metal-superoxo decomposition (pathway B) produces the
hydroperoxyl radical, an established radical chain autoxidation
initiator. Control experiments rule out the generation of PFeIII
species from trace peroxide impurities present in either the
solvent or isobutane gas (pathway C) since (i) PFeIII species
are stoichiometrically produced from the PFe-L2 precursors
Significant macrocycle destruction, however, is observed in
the electronic absorption spectra of these complexes recorded
over longer time periods (∼48 h) when large excesses of tert-
butyl hydroperoxide are present; Figure 9 highlights such a set
of data obtained when (C3F7)4PFe-(3-F-py)2 is used as the
catalyst precursor. Time-dependent 19F NMR spectra obtained
for analogous tert-butyl hydroperoxide decomposition reactions
in which (C3F7)4PFe-(3-F-py)2 and (C6F5)4PFe-(3-F-py)2 were
utilized as catalyst precursors corroborate these results (Sup-
porting Information); furthermore, these experiments (T ) -70
to -20 °C) show the rapid conversion of these complexes to
intermediates which were not observed during the in situ
oxidation experiments and that, prior to the onset of the catalyst
degradation, a single high-spin PFeIII species is present. It is
interesting to note that, for (C6F5)4PFe-catalyzed tert-butyl
hydroperoxide decomposition reactions, there is no indication
of the presence of [(C6F5)4PFe]2O by 19F NMR or optical
(71) Lyons, J. E.; Ellis, P. E., Jr. U.S. Patent 5345008 A, 1994.

3136 Inorganic Chemistry, Vol. 39, No. 15, 2000
Moore et al.
Figure 7. An alternating sequence of consecutively acquired 19F and 17O NMR spectra for a PFe-catalyzed isobutane oxidation reaction carried out
at 80 °C in benzene-d6 in which (C6F5)4PFe-(py)2 is used as the catalyst precursor. (A) 19F NMR spectrum of (C6F5)4PFe-L2 at 80 °C under Ar.
Spectra B-H that follow were taken during the time interval specified after the addition of labeled water, isobutane, and oxygen to the NMR tube.
Prime notation represents the 19F signals corresponding to [(C6F5)4PFe]2O, double prime notation represents resonances observed for (C6F5)4PFe-
OH, and triple prime notation denotes resonances that are tentatively ascribed to a (C6F5)4PFe-(O-tert-butyl) complex. Resonances labeled with X
are presently unassigned. All 17O NMR spectra recorded at time intervals after spectrum F show the same relative intensities of the 17O products;
19F NMR spectra taken after spectrum F evince a steady decrease in concentration of benzene-soluble PFe species.
Figure 9. In situ time-dependent electronic absorption spectra recorded
over long time intervals that chronicle a tert-butyl hydroperoxide
decomposition reaction in which (C3F7)4PFe-(py)2 was utilized as a
catalyst precursor. (A) (C3F7)4PFe-(py)2 under Ar. (B) (C3F7)4PFe-
(py)2-derived species observed at t ) 5 min following the addition of
a 1000-fold molar excess of tert-butyl hydroperoxide. (C) (C3F7)4PFe-
(py)2-derived species observed 12 h following the initial addition of
tert-butyl peroxide. No further changes in the optical spectrum were
apparent over an additional 1 h time period; at this point, an additional
2000 equiv of tert-butyl peroxide was added. (D) Optical spectra
recorded 36 h after spectrum C.
at such a level would have been detectable by 1H or 13C NMR
spectroscopy in the absence of catalyst, which was not the case),
and (iii) reducing the concentration of the PFe-L2 precursor to
1 µM shows only the production of PFeIII species in optically
monitored oxidation experiments (oxidation products are ob-
served only after the system is charged with isobutane and
dioxygen, following the intermediacy of a normal induction
period). If peroxide impurities exceeded this concentration level,
Figure 8. In situ time-dependent electronic absorption spectra of tert-
one would expect to see reaction initiation and immediate
butyl hydroperoxide decomposition reactions in benzene solvent in
observation of isobutane oxidation products upon addition of
isobutane and dioxygen, which is not the case.
which (A) Br8(C6F5)4PFe-(py)2, (B) (C6F5)4PFe-(py)2, and (C) (C3F7)4-
PFe-(py)2 were utilized as catalyst precursors. The trace labeled “initial
spectrum” corresponds to the isolated ferrous porphyrin species in
benzene under Ar; subsequent electronic absorption spectra taken at
the labeled time intervals record the time-dependent optical changes
that occur once a 3000-fold molar excess of tert-butyl hydroperoxide
has been added. All spectra were taken at 26 °C. Experimental
conditions: [PFe] ) 10-5 M; [tert-butyl hydroperoxide] ) 3 × 10-2
M.
Given these facts, and the previously presented experimental
evidence, we propose Scheme 4 as the most plausible mecha-
nism by which PFeIII species are produced from these PFeII-
L2 precursors. This scheme is based on a fourth, and extremely
well-precedented, mechanism72-74 that involves a transient
PFeIVdO species in the O2-mediated ferrous-to-ferric porphyrin
(72) Chin, D.-H.; La Mar, G. N.; Balch, A. L. J. Am. Chem. Soc. 1980,
102, 4344-4350.
(73) Balch, A. L.; Renner, M. W. Inorg. Chem. 1986, 25, 303-307.
(74) Arasasingham, R. D.; Balch, A. L.; Cornman, C. R.; Latos-Grazynski,
L. J. Am. Chem. Soc. 1989, 111, 4357-4363.
in the absence of isobutane under O2, (ii) ferrous PFe-L2
complexes are completely converted to PFeIII species under
dioxygen pressure even when the PFe-L2 concentration is as
high as 0.1 M (any peroxide or hydroperoxide impurity present

(Porphinato) iron-Catalyzed Isobutane Oxidation
Inorganic Chemistry, Vol. 39, No. 15, 2000 3137
in benzene.79 (iii) Any HO• that escapes the [PFeII-OH] cage
would initiate an autoxidation mechanism. (iv) The lengths of
the induction periods that are observed for these reactions
correlate with the magnitudes of previously reported PFe-OH
bond dissociation energies.80
Once H atom abstraction from isobutane occurs, the organic
radical will react with dioxygen and initiate an autoxidation
radical chain process; this modified Haber-Weiss cycle is
shown in Scheme 5. While some of the details of Scheme 5 are
speculative in nature, note that the propagation step described
in Scheme 5A produces the tert-butyl hydroperoxyl radical.
Oxidized organic products are created through the PFe-catalyzed
decomposition of the hydroperoxide (Scheme 5B,C); extensive
low-temperature spectroscopic data obtained by Balch establish
precedent for the mechanistic role ascribed to PFeIII-bound tert-
butyl hydroperoxo compounds.74,81-84 The 17O NMR experi-
ments provide insight into the nature of the PFe-catalyzed
hydroperoxide decomposition process; because the 17O label is
observed in all the major organic products produced in the
reaction (tert-butyl alcohol, acetone, and tert-butyl peroxide),
PFe-OH must play a major role in tert-butyl hydroperoxide
decomposition as well as reaction initiation.
A number of reactions known to occur under Haber-Weiss
conditions have been omitted from Scheme 5 for the sake of
clarity. With respect to the 17O-labeling experiments, a few of
these such reactions are worth noting in brief. It is well-known
that (porphinato)iron(III)-(alkyl peroxide) species (Scheme 5B),
in addition to undergoing bond homolysis reactions that produce
PFe(II) species and peroxyl radicals, decompose via both
homolytic and heterolytic O-O bond cleavage processes.7,37,85-87
Either homolytic or heterolytic cleavage enables facile incor-
poration of the 17O label into the â-oxygen atom of the peroxyl
group. Once this occurs, two pathways allow incorporation of
the 17O label into tert-butyl alcohol. As noted previously, the
17O-labeling experiments were carried out in sapphire NMR
tubes that utilized 8:1 isobutane/dioxygen (see the Experimental
Section); these reactions were run until catalysis ceased, and
exhibited isobutane conversions that ranged from 5% to 10%.
Because this corresponds to incorporation of 40-80% of the
O2 present in the NMR tubes into products, the fact that cycle
Figure 10. In situ 19F NMR spectra that examine a PFe-catalyzed
isobutane oxidation reaction in which a (C3F7)4PFe-(3-F-py)2 catalyst
precursor was utilized in the presence of 1000 equiv of tert-butyl
hydroperoxide. (A) The PFe-L2 complex at 26 °C under Ar in benzene-
d6. (B) 19F NMR spectrum taken at 80 °C 30 min after O2/isobutane/
tert-butyl hydroperoxide addition. Alcohol production is observed in
the 13C NMR at this time (spectrum not shown). (C) 19F NMR spectrum
obtained at 80 °C after data were acquired over the 48-50 h time
following O2/isobutane/tert-butyl hydroperoxide addition; at this point
in time, the dioxygen contained in the NMR tube was completely
consumed. The resonance labeling scheme is identical to that described
in Figure 3.
conversion. Although PFeIVdO species have been isolated and
characterized,75 because the species in brackets are not spec-
troscopically detected under our experimental conditions, this
route for the PFeII-to-PFeIII conversion can only be considered
a postulate. We emphasize, however, that the most attractive
feature of this proposal is the fact that it produces no species
(HOOH, O2-) that either are known to directly initiate or can
be converted to species that directly initiate isobutane autoxi-
dation in the presence of PFe complexes.49 It is important to
note that all other commonly invoked mechanisms for an O2-
mediated PFeII-to-PFeIII conversion fail this crucial test. Pro-
cesses i-v of Scheme 4 must be rapid, since PFe-OH and/or
(PFe)2O complexes are observed immediately upon pressuriza-
tion of either a sapphire NMR tube or an autoclave reactor
containing a PFeII-L2 solution with dioxygen (t < 15 min. If
the mechanism of Scheme 4 is operative, the transiently
generated PFeIVdO species must either react with hydrocarbon
substrates on a time scale that is long with respect to that for
reaction iv (Scheme 4) or not be thermodynamically competent
to promote alkane hydroxylation. This latter hypothesis is
consistent with thermodynamic data compiled by Sawyer which
indicate that the FeIVdO + H• f FesOH bond formation
energy (-∆GBF) compares unfavorably with the dissociative
bond energy (-∆HBDE) required for isobutane CsH bond
homolysis.76-78 We thus propose that isobutane oxidation is
initiated by a mechanism that involves the thermodynamically
unfavorable homolysis of the PFe-OH bond. This proposal is
attractive for four reasons: (i) It is consistent with the long and
often widely variable induction periods. (ii) It accounts for the
presence of the small quantities of the primary alkane oxidation
product, 2-methyl-1-propanol, observed in these reactions,
because H-atom abstraction processes involving hydroxyl
radicals and primary alkanes are thermodynamically favored
(79) H3CCH2CH3 f H3CCH2CH2• + H• (-∆HBDE ) 100 kcal/mol); H2O
f HO• + H• (-∆HBDE ) 111 kcal/mol).
(80) Assuming that a typical induction period of 6 h correlates directly
with •OH escape from a [PFeII•OH] cage, and that the cage recombina-
tion reaction to produce PFe-OH occurs at diffusion-controlled rates
(109 M-1 s-1), gives an equilibrium constant K ) 4.63 × 10-14 M-1
at 80 °C. Provided that ∆S° for this reaction does not deviate
appreciably from 23 eu, a standard value for bond homolysis reactions
(see for example: (a) Handbook of Chemistry and Physics, 71st ed.;
CRC Press: Boca Raton, FL, 1990; pp D51-D95. (b) Woska, D. C.;
Xie, Z. D.; Gridnev, A. A.; Ittel, S. D.; Fryd, M.; Wayland, B. B. J.
Am. Chem. Soc. 1996, 118, 9102-9109) allows an estimation of
∆HBDE ) 30 kcal mol-1. This value lies intermediate in the range of
∆HBDE values for PFe-OH complexes that have been measured to
date (see: Richert, S. A.; Tsang, P. K. S.; Sawyer, D. T. Inorg. Chem.
1989, 28, 2471-2475. Tung, H. C.; Chooto, P.; Sawyer, D. T.
Langmuir 1991, 7, 1635-1641).
(81) Arasasingham, R. D.; Balch, A. L.; Latos-Grazynski, L. J. Am. Chem.
Soc. 1987, 109, 5846-5847.
(82) Arasasingham, R. D.; Balch, A. L.; Har, R. L.; Latos-Grazynski, L.
J. Am. Chem. Soc. 1990, 112, 7566-7571.
(83) Balch, A. L. Inorg. Chim. Acta 1992, 198, 297-307.
(84) Balch, A. L.; Olmstead, M. M.; Safari, N.; St. Claire, T. N. Inorg.
Chem. 1994, 33, 2815-2822.
(75) Groves, J. T.; Gross, Z.; Stern, M. K. Inorg. Chem. 1994, 33, 5065-
5072.
(85) He, G.-X.; Bruice, T. C. J. Am. Chem. Soc. 1991, 113, 2747-2753.
(86) Traylor, T. G.; Tsuchiya, S.; Byun, Y.-S.; Kim, C. J. Am. Chem. Soc.
1993, 115, 2775-2781.
(76) FeIVdO + H• f Fe(III)-OH (-∆GBF ) 78 kcal/mol); (CH3)3CH f
(CH3)3C• + H• (-∆HBDE ) 93 kcal/mol).
(77) Sawyer, D. T. J. Phys. Chem. 1989, 93, 7977-7978.
(78) The ActiVation of Dioxygen and Homogeneous Catalytic Oxidation;
Barton, D. H. R., Martel, A. E., Sawyer, D. T., Eds.; Plenum Press:
New York, 1993; pp 1-7.
(87) Traylor, T. G.; Kim, C.; Richards, J. L.; Xu, F.; Perrin, C. L. J. Am.
Chem. Soc. 1995, 117, 3468-3474.
(88) Lyons, J. E.; Ellis, P. E., Jr.; Myers, H. K., Jr. J. Catal. 1995, 155,
59-73.

3138 Inorganic Chemistry, Vol. 39, No. 15, 2000
Moore et al.
Scheme 3. Mechanisms by Which PFeIII Species Can Be Generated from PFeII Precursors under Oxidative Conditions: (A)
Generation of Superoxide by SN1 or SN2 Pathways; (B) H2O-Assisted Metal Superoxo Decomposition; (C) Direct Oxidation via
Hydroperoxide Impurities Present in Isobutane and/or Solvent
Scheme 4. Proposed Initiation Mechanism for PFe-Catalyzed Isobutane Oxidation
C of Scheme 5 produces labeled dioxygen enables 17O to be
introduced into the R-peroxyl position and hence into the tert-
butyl alcohol, since labeled dioxygen would be free to reenter
the catalytic cycle of Scheme 4A. Alternatively, the common
Haber-Weiss reaction that produces tert-butyl peroxide via the
reaction of tert-butyl radical and tert-butyl peroxyl radical
provides a second route to 17O-labeled tert-butyl alcohol
(Scheme 5C). Because the majority of the 17O label at the end
of the catalytic reaction is incorporated into products, the PFe-
OH-dependent process of Scheme 5B must be the focal point
for peroxide decomposition.
complexes are competent to catalyze isobutane oxidation at
moderate temperatures without the input of stoichiometric
reductants, such simple catalysts are not stable under the
autoxidation reaction conditions. While porphyrin ring haloge-
nation leads to more robust epoxidation catalysts that utilize
PFeIII complexes and an O atom donor to produce reactive
compound I analogues (P•+FeIVdO),9,91,92 PFe alkane hydroxy-
lation catalysts that derive their oxidizing equivalents from O2
(89) Ellis, P. E., Jr.; Lyons, J. E.; Shaikh, S. N. Catal. Lett. 1994, 24, 79-
83.
(90) Lyons, J. E.; Ellis, P. E., Jr. Appl. Catal., A 1992, 84, L1-L6.
(91) Traylor, T. G.; Dunlap, B. E.; Miksztal, A. R. Basic Life Sci. 1988,
49, 509-512.
(92) Traylor, P. S.; Dolphin, D.; Traylor, T. G. J. Chem. Soc., Chem.
Commun. 1984, 279-280.
A large body of work has been dedicated to the study of PFe-
catalyzed alkane oxidation using dioxygen as the stoichiometric
oxidant.41-43,46-48,50,67,88-90 Although electron-deficient PFe

(Porphinato) iron-Catalyzed Isobutane Oxidation
Inorganic Chemistry, Vol. 39, No. 15, 2000 3139
Scheme 5. Modified Haber-Weiss Radical Chain
Mechanism for PFe-Catalyzed Alkane Oxidation and
Hydroperoxide Decomposition
This study highlights two roles for the PFe complexes in
catalytic isobutane oxidation: that of the radical chain initiator
and, consistent with work by Labinger and Gray,48,50,67 the
species responsible for the catalytic decomposition of organic
peroxides. It is thus appropriate to consider the dual requirements
of radical chain initiation and facile hydroperoxide decomposi-
tion, and whether a set of experimental conditions and catalysts
•
can be elucidated that allows for control of OH concentration
and facile hydroperoxide decomposition, while protecting the
catalyst from decomposition. Furthermore, with respect to
catalyst design and the original proposal by Ellis and Lyons,40-45,65
the development of a ligand framework that will allow a
PFeIVdO species to replicate the chemistry of the biological
P•+FeIVdO paradigm, and direct the incorporation of both O
atoms of O2 into organic substrates while suppressing autoxi-
dation radical chain processes, clearly remains a challenging
problem in biomimetic chemistry.
Acknowledgment. M.J.T. thanks the Office of Naval
Research (Grant N00014-98-1-0187) for their generous support
of this work, and gratefully acknowledges the Searle Scholars
Program (Chicago Community Trust), the Arnold and Mabel
Beckman Foundation, and the National Science Foundation for
Young Investigator Awards, as well as the Camille and Henry
Dreyfus and Alfred P. Sloan Foundations for research fellow-
ships. I.T.H. acknowledges the support of the Exxon Research
and Engineering Co. for his work in the area of hydrocarbon
activation. We thank James T. Fletcher for preparing several
figures in this manuscript, acknowledge Raymond A. Cook,
Kenneth A. Eriksen, and Jeffrey E. Bond for their invaluable
experimental assistance, and express gratitude to Drs. Paul A.
Stevens, Walter Weismann, Istva´n Pelcer, and David Woska
for many stimulating discussions.
that are based on highly halogenated Br8(C6F5)4PH2 ligand
systems possess stabilities that are diminished relative to those
of analogous catalysts that feature (C6F5)4PH2 and (C3F7)4PH2
ligand frameworks.
Supporting Information Available: Spectroscopic data that detail
the temperature and axial ligand concentration dependences of NMR
chemical shifts of these (porphinato)iron complexes, along with
additional catalytic isobutane oxidation and tert-butyl hydroperoxide
decomposition data. This material is available free of charge via the
Internet at http://pubs.acs.org.
Last, with respect to the conversion of PFeII to PFeIII within
the context of an autoxidation mechanism, it has been proposed
that peroxides play a key role in this process and that alkoxyl
radicals and hydroxide anions are produced stoichiometrically
each time a ferrous center is oxidized.48 This study demonstrates
that such a role for ROOH is unlikely due to its extremely low
concentration under these experimental conditions and that O2
is responsible for this conversion in the catalytic cycle (Scheme
5). Since hydroperoxides are not detected at any point in the
catalytic reaction, their concentration must be lower than the
detection limit of the NMR spectrometer. Assuming that
hydroperoxide present at concentrations higher than 1/25 of that
for the catalyst would be detectable, the upper limit for the
hydroperoxide concentration is ∼10-3 mM. This value is at least
4 orders of magnitude smaller than the concentration of O2 in
benzene under these conditions, which is ∼10 mM at 353 K
and 1 atm of O2 pressure.93,94
IC000284O
(93) Byrne, J. E.; Battino, R.; Danforth, W. F. J. Chem. Thermodyn. 1974,
6, 245-250.
(94) Bimolecular rate constants for dioxygen binding at five-coordinate
PFe-(L) complexes typically range from 1 × 106 to 5 × 108 M-1
s-1 at ambient temperature (See: Momenteau, M.; Reed, C. A. Chem.
ReV. 1994, 94, 659-698), with sterically unencumbered ferrous
porphyrin complexes displaying an average bimolecular rate constant
of ∼5 × 107 M-1 s-1. Because the dioxygen concentration exceeds
that of tert-butyl hydroperoxide by at least a factor of 104, for
reoxidation of PFe(II) by tert-butyl hydroperoxide to compete with
the dioxygen-mediated reaction, the bimolecular rate constant for this
process would have to exceed that for a diffusion-controlled process
by (at least) a factor of 1000.