Selective oxidation of methane to methanol catalyzed, with C-H activation, by homogeneous, cationic gold

Article abstract of DOI:10.1002/anie.200461055

Cationic gold catalyzes the title reaction in strong acid solvents with Se^VI as the stoichiometric oxidant. At 180°C and 27 bar of CH₄ turnover numbers of about 30 and methanol concentrations of 0.6 M have been observed with selectivities >90%. DFT calculations indicate that cationic Au^I and Au^III are both viable catalysts that operate by electrophilic C-H activation and oxidative functionalization (see picture).

Full text of DOI:10.1002/anie.200461055

Communications
ꢀ
Homogeneous Catalysis
active for heterolytic C H activation as well as oxidative
functionalization; this is likely because AuI (d10, two-coor-
dinate) and AuIII ions (d8, square planar) are both “soft”
electrophiles that are isoelectronic and isostructural with the
neighboring cations (HgII and PtII, respectively) that have
Selective Oxidation of Methane to Methanol
ꢀ
Catalyzed, with C H Activation, by
Homogeneous, Cationic Gold**
ꢀ
been shown to be active for electrophilic C H activation of
methane.[2,4] This situation is not common, and in most
catalytic systems based on “soft”, redox-active electrophiles
CJ Jones, Doug Taube, Vadim R. Ziatdinov,
Roy A. Periana,* Robert J. Nielsen, Jonas Oxgaard, and
William A. Goddard III
ꢀ
only one oxidation state of the redox couple is active for C H
activation.Thus, we sought to explore the catalytic chemistry
of gold cations for the oxidation of methane.To our knowl-
edge, while gold complexes have been reported to facilitate
free-radical reactions of alkanes with peroxides in low
yields,[5] no homogeneous gold catalysts that operate by
Some of the most efficient homogeneous catalysts for the low-
temperature, selective oxidation of methane to functionalized
products employ a mechanism involving C H activation
with an electrophilic substitution mechanism.Several such
systems have been reported based on the cations HgII,[2] PdII,[3]
and PtII.[4] These catalyst systems typically operate by two
[1]
ꢀ
ꢀ
heterolytic C H activation and oxidative functionalization
have been reported for the selective functionalization of
alkanes.This is possibly because of the strong propensity for
irreversible formation of gold metal, and any attempts to
develop redox catalysis based on homogeneous Au cations
must address this issue.
ꢀ
general steps that involve: A) C H activation by coordina-
tion of the methane to the inner sphere of the catalyst (En+
)
ꢀ
followed by cleavage of the C H bond by overall electrophilic
substitution to generate En+ CH3 intermediates, and B) oxi-
In strong acid solvents such as triflic or sulfuric acid,
AuIII cations (generated by dissolution[6] of Au2O3) react with
methane at 1808C to selectively generate methanol (as a
mixture of the ester and methanol) in high yield (Table 1,
entries 1 and 2).As expected, the irreversible formation of
metallic gold is very evident after these reactions and, unlike
reactions with HgII,[2] PtII,[4d] and PdII [3a] that are catalytic in
96% H2SO4, only stoichiometric reactions (turnover numbers
(TONs) < 1) are observed with AuIII [Eq.(1)]. Soluble
cationic gold is essential for these reactions as no methanol
is observed under identical conditions without added
AuIII ions (entry 3), or in the presence of metallic gold
(entry 4) which is not dissolved in hot H2SO4.
ꢀ
dative functionalization involving redox reactions of En+
ꢀ
CH3 to generate the desired oxidized product CH3X.[4a]
Consequently, efficient catalysts that follow this pathway
would be expected to be “soft”, highly electrophilic species
that form relatively strong covalent bonds to carbon atoms
and that are also good oxidants.
We considered that gold cations could be uniquely
efficient electrophilic catalysts for methane conversion
because, as shown in the conceptual catalytic cycle
(Scheme 1), both AuI and AuIII oxidation states could be
HX
180 oC, 2 h
3 CH ð27 barÞ þ 2 AuX þ 3 H O
3 CH OH þ 6 HX þ 2 Au0
ƒƒƒƒƒ!
4
3
2
3
X ¼ HSO4
ð1Þ
Consistent with the known nobility of gold metal, no
methanol formed when SO3 or persulfate (K2S2O8) were
added as possible oxidants of metallic gold (entries 5 and 6).
We considered the use of SeVI ions as a more suitable oxidant.
SeVI ions are a more powerful oxidizing agent than SVI ions
(Eo = 1.5 V SeO42ꢀ/H2SeO3, Eo = 0.17 V SO42ꢀ/H2SO3, respec-
tively) and, critically, the hydrated form, selenic acid
(H2SeO4), is known to oxidize gold metal.[7] Equally impor-
tant, selenic acid is also almost as acidic as sulfuric acid,[7,8]
and it was anticipated that neither the acid nor the conjugate
ꢀ
Scheme 1. Proposed functionalization catalytic cycles involving C H
activation by AuI and AuIII species.
[*] C. Jones, D. Taube, V. R. Ziatdinov, Prof. Dr. R. A. Periana
University of Southern California
Department of Chemistry, Loker Hydrocarbon Institute
Los Angeles, CA 90089 (USA)
Fax: (+)213-821-2656
ꢀ
base (HSeO4 ) would inhibit catalysis with electrophilic,
cationic gold by blocking the required coordination of
E-mail: rperiana@usc.edu
methane to the metal center.[4a]
R. J. Nielsen, J. Oxgaard, Prof. Dr. W. A. Goddard III
Materials and Process Simulation Center
Beckman Institute (139-74)
Division of Chemistry and Chemical Engineering
California Institute of Technology, Pasadena, CA 91125 (USA)
We now report that a 3m solution of H2SeO4 in 96%
sulfuric acid solvent containing 27 mm metallic gold (added as
a 20 mesh powder) leads to catalytic oxidation of methane
(27 bar) to methanol at 1808C [Eq.(2)]. [9] Turnover numbers
of up to 30 and turnover frequencies of about 10ꢀ3 sꢀ1 have
[**] This work was supported by NSF Grant CHE-0328121. The
computer facilities of the MSC were provided through support from
DURIP and MURI. We thank Prof. R. J. Puddephatt for helpful
discussion.
CH4 þ H2SeO4 ! CH3OH þ H2SeO3
ð2Þ
4626
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200461055
Angew. Chem. Int. Ed. 2004, 43, 4626 –4629

Angewandte
Chemie
Table 1: Yield of CH3OH under different reaction conditions.[a] [9]
Entry Au source [mm][b] Additives t [h] TON TOF [sꢀ1
]
CH3OH (% sel) [CH3OH] [mm] CH4 (% conv) CO2 [mol%]
1
2
3
4
5
6
7
8
Au2O3 [50]
Au2O3 [50]
–
–
1
1
6
1
3
2
2
9
10
2
0.8
0.9
0
0
0
0
9
32
4
8
10ꢀ3
10ꢀ3
0
0
0
>90
>90
0
60
70
0
0
0
<5
240
178
350
215
627
69
<2
<2
<1
<1
<1
<1
11
8
15
10
28
<1
<1
<1
<1
<1
<2
3
2
6
2
2
100% CF3SO3H
–
–
Au0 [27][c]
Au0 [27][c]
Au0 [27][c]
Au0 [27][c]
Au0 [5][c]
Au2O3 [81]
Au0 [27][c]
Au0 [81][c]
–
0
0
0
81
77
74
86
94
68
2 wt% SO3
0.1m K2S2O8
3m H2SeO4
3m H2SeO4
3m H2SeO4
3m SeO3 +6 bar O2
3m SeO3 +2% SO3
3m H2SeO4
0
10ꢀ3
10ꢀ3
10ꢀ3
10ꢀ3
10ꢀ3
–
9
10
11
12
3
6
15
0
3
<2
[a] Reaction temperature 1808C, 27 bar CH4 with 96% D2SO4 as solvent. [b] Based on in situ generated AuIII species assuming complete dissolution of
added gold. [c] Added as metallic (20 mesh) gold powder.
been observed in methanol concentrations of up to 0.6m in
sulfuric acid with > 90% selectivity, (Table 1, entries 7–11).
Both gold and selenic acid are required for the reaction; no
methanol is observed in the absence of SeVI ions with or
without Au0 (entries 3 and 4) and only small amounts of
methanol are formed in the presence of selenic acid without
Au0 (entry 12).Use of 100% enriched [ 13C]methane as well as
1H, 13C NMR spectroscopic, and HPLC analyses of the crude
reaction mixtures confirmed that methanol is the predom-
Control experiments confirm that selenic acid can dissolve
gold metal.Thus, addition of metallic gold powder or gold
metal (separately obtained from stoichiometric reactions
between Au2O3 and methane) to a 3m solution of selenic
acid in 96% sulfuric acid leads to rapid dissolution of the gold
metal upon warming to 1008C to generate a clear yellow
solution of cationic gold.While we have not obtained kinetic
data on the various reaction steps proposed in Figure 1, this
visual observation of rapid dissolution at low temperature
suggests that the rate-determining step for methane oxidation
catalyzed by cationic gold is not Au0 reoxidation.
We have not been able to determine the exact oxidation
state of the dissolved gold, but on the basis of known redox
potentials we expect that AuIII ions are the major species.
However, it is possible that AuI ions (potentially stabilized to
disproportionation by Se anions) could be present.The
possibility that both AuIII and AuI ions may be present in
solution leads to the challenging question of whether one or
I
III
ꢀ
both oxidation states (Au and Au ) are active for the C H
activation of methane.This is an important point because, as
discussed above, both AuI and AuIII ions are potential
catalysts.The observation that reactions starting from
AuIII ions (with Au2O3 as the source) lead to stoichiometric
conversion of methane into methanol [Eq.(1)] does not
ꢀ
require that AuIII ions are the active species.The C
H
Figure 1. 13C NMR spectrum of the crude reaction mixture resulting
activation reaction could occur with catalytic amounts of
AuI ions, formed by reduction of AuIII ions with adventitious
reductants, and in both cases the stoichiometry shown in
Equation (1) would be expected.As observed with the related
HgII,[2] PtII,[4] and PdII [2] systems, some evidence for the
from the reaction of 13CH4.[10]
inant liquid-phase product generated from methane.As can
be seen from an NMR spectrum of the crude reaction mixture
(Figure 1), only a small amount of methane selenoic acid
(CH3SeO3H) is observed in the liquid phase.Very low levels
(< 5% based on added CH4) of CO2 is observed in the gas
phase along with recovered CH4.The observation that
addition of O2 to the reaction mixture (entry 10) has no
significant effect on the reaction (compare entry 7) and the
high reaction selectivity and reproducibility suggests that free
radicals are not involved.Carrying out the stoichiometric
reaction with Au2O3 at several temperatures between 100 and
2008C leads to an estimated activation barrierof about
30 kcalmolꢀ1.
ꢀ
formation of an Au CH3 intermediate is the observation that
low levels of CH3D (< 2%, which is not produced in the
absence of added AuIII) are produced when the reactions are
carried out in D2SO4.So far, attempts at synthesizing likely
I
III
ꢀ
ꢀ
Au CH3 or Au CH3 intermediates (without “soft” ligands
such as phosphine) have been unsuccessful.
Given the synthetic challenges, we turned to DFT
calculations to determine the feasibility of the steps proposed
++ [11]
in Scheme 1.B3LYP/LACVP**
,
as implemented by the
Jaguar 5.5 package,[12] was applied with the basis set for sulfur
augmented with an extra, compact d-like polarization shell to
Angew. Chem. Int. Ed. 2004, 43, 4626 –4629
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4627

Communications
better describe high oxidation states.[13,14] A dielectric con-
tinuum surrounding the models at a probe radius of 2.205 Š
represented the sulfuric acid solvent with e = 50.[15] Enthalpies
describe solutes in ideal solution at 1758C.Entropy changes,
which are not accounted for, must be kept in mind as species
enter and exit the models.A survey of complexes of gold ions
with HnSO4 (n = 0,1,2) showed that the square planar
AuIII(SO4)2ꢀ ion (1) and linear AuI(HSO4)2ꢀ ion (7) are the
most stable AuIII and AuI species in solution.However,
bidentate sulfate ions, bisulfate ions, and water made similarly
strong bonds to gold, so other complexes are most likely
present in equilibrium quantities.
Figure 2 illustrates the AuIII pathway, by which 1 reacts
with methane to produce the [AuIII(CH3)(HSO4)3]ꢀ complex
5.Calculations show that a key role of the acid solvent is the
generation of various protonated states of the gold complexes,
such as 2, that are sufficiently labile to allow methane to
ꢀ
Figure 3. Calculated enthalpies (kcalmolꢀ1) at 1758C relative to 7,
including solvation by sulfuric acid (e=50). For clarity, spectator
H2SO4 groups have been omitted from some structures. The top curve
illustrates electrophilic substitution, while the bottom curve illustrates
oxidative addition.
substitute into the coordination sphere of the metal.C
H
activation can proceed through a number of transition states,
the most accessible of which is an electrophilic substitution
pathway 3°, which features an intramolecular abstraction of a
proton from methane by the neighboring bisulfate, with an
activation energy of 28.1 kcalmolꢀ1.Alternative mechanisms
with slightly higher enthalpies include intermolecular reac-
I
ꢀ
While C H activation with Au ions is feasible, direct
functionalization from this species is not.Both reductive
elimination and nucleophilic substitution yields an
AuꢀI·2H2SO4 ion that is calculated to be 37.6 kcalmolꢀ1
higher in energy than 10.However, 10 can most likely be
oxidized to the AuIII species 5, either through selenic acid,
sulfuric acid, or 1.Indeed, preliminary calculations show that
the redox exchange 10 + 1!5 + 7 is exothermic by about
32 kcalmolꢀ1, and would thus be feasible even for the
stoichiometric process.
III
ꢀ
tions with bisulfate ions in solution.The resulting Au
CH3
complex 4 most likely converts into the lowest energy AuIII
CH3 complex 5 before functionalization.
ꢀ
Oxidative functionalization of 5 proceeds with a low
energy barrier (DH = 10.8 kcalmolꢀ1) through an SN2 type
attack on the methyl group by a free bisulfate ion 6°.This
process is facilitated by the dissociation of the ligand trans to
the CH3 group.A reductive elimination mechanism was also
found, but with a DH° value of 32.9 kcalmolꢀ1.
Based on these results, it appears that oxidative addition
to AuI ions is the favored pathway, which is 7.0 kcalmolꢀ1
lower in energy than the electrophilic substitution pathway
for AuIII ions.However, these energetics do not take the
relative concentrations of the AuIII and AuI cations into
account, and the concentration of AuI ions should be signifi-
AuI species can follow a similar electrophilic substitution
ꢀ
C H activation mechanism (Figure 3) to produce
[AuI(CH3)(H2SO4)] (10) with an activation energy of
26.2 kcalmolꢀ1 from 7.Interestingly, an oxidative addition
mechanism was also found with an activation energy of
21.1 kcalmolꢀ1, somewhat lower than the electrophilic sub-
stitution.The resulting Au III species 12 is a stable intermedi-
cantly lower than AuIII ions in the presence of excess oxidant
ꢀ1
SeVI.A difference of 70. kcalmol
corresponds to a ratio of
ate, but the compound is acidic enough that deprotonation
approximately 2500:1 at 1758C, and thus if the ratio of AuI to
I
would rapidly lead to the same Au CH3 complex 10 for the
electrophilic substitution pathway.
AuIII ions is less than 1:2500, the AuIII pathway would be
ꢀ
favored.Since this is a feasible ratio, both Au and AuIII
I
Figure 2. Calculated enthalpies (kcalmolꢀ1) relative to 1, including solvation by sulfuric acid (e=50). Spectator sulfuric acid molecules have been
omitted from some structures for clarity.
4628
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 4626 –4629

Angewandte
Chemie
implied by this finite basis cause average errors of 4 kcalmolꢀ1
which the implicit solvent model will increase.
[12] L.L.C.Schrꢂdinger, Jaguar 5.5, Portland, OR, 1991–2003.
[13] J.M.L.Martin, J. Chem. Phys. 1998, 108, 2791.
[14] The d-like polarization shell with exponent a = 0.65 was
replaced with two shells: a single Gaussian with exponent a =
0.55 and a two-Gaussian contraction with exponents a1 = 7.5 and
a2 = 1.8 with contraction coefficients c1 = 0.2 and c2 = 1.0.
[15] A high-temperature (1758C) value was projected assuming the
dielectric constant of sulfuric acid decays with temperature at a
rate similar to that of water; ꢀ(1/e) @e/@T= 0.0045.
,
remain possible active species, which is in contrast to findings
for PtII [4a] and HgII ions.[2]
In summary, we report that cationic gold catalyzes the
selective, low-temperature, oxidation of methane to methanol
in strong acid solvent using SeVI ions as the stoichiometric
oxidant.The reaction does not appear to proceed through
free radicals and DFT calculations indicate that AuI or
AuIII species are both viable catalysts that operate by
ꢀ
mechanisms involving overall electrophilic C H activation
and oxidative functionalization.
Received: June 23, 2004
Published Online: August 12, 2004
ꢀ
Keywords: C H activation · gold · homogeneous catalysis ·
methane · oxidation
.
[1] a) B.A.Arndtsen, R.G.Bergman, T.A.Mobley, T.H.Peterson,
Acc. Chem. Res. 1995, 28, 154; b) A.E.Shilov, G.B. Shulꢀpin,
Chem. Rev. 1997, 97, 2879; c) R.H. Crabtree, J. Chem. Soc.
Dalton Trans. 2001, 17, 2437; d) W.D.Jones, Science 2000, 287,
1942.
[2] R.A. Periana, D.J. Taube, E.R. Evitt, D.G. Loffler, P.R.
Wentrcek, G.Voss, T.Masuda, Science 1993, 259, 340.
[3] a) R.A.Periana, O.Mironov, D.Taube, G.Bhalla, C.J.Jones,
Science 2003, 301, 814; b) A.Sen, Acc. Chem. Res. 1998, 31, 550.
[4] a) K.X.J.Xu, R.A.Periana, W.A.Goddard III,
Organometal-
lics 2003, 22, 2057; b) J.A.Labinger, J.E.Bercaw, Nature 2002,
507-514, 507; c) M.Lin, C.Shen, E.A.Garcia-Zayas, A.Sen,
Am. Chem. Soc. 2001, 123, 1000; d) R.A.Periana, D.J.Taube, S.
J.
Gamble, H.Taube, T.Satoh, H.Fuji, Science 1998, 280, 560; e) S.
Stahl, J.A. Labinger, J.E. Bercaw, Angew. Chem. 1998, 110,
2298; Angew. Chem. Int. Ed. 1998, 37, 2180; f) A.E.Shilov, G.B.
Shulꢀpin, Chem.Rev. 1997, 97, 2879.
[5] a) G.B.Shulꢀpin, A.E.Shilov, G.Sꢁss-Fink,
Tetrahedron Lett.
2001, 42, 7253; b) G.B.Shulꢀpin, J. Mol. Catal. A 2002, 189, 39.
[6] a) M.S.Wickleder, K.Esser, Z. Anorg. Allg. Chem. 2002, 628,
911; b) M.S.Wickleder, O.Buchner, Z. Naturforsch. B 2001, 56,
1340.
[7] N.N. Greenwood, A. Earnshaw, Chemistry of the Elements,
2nd ed., Butterworth-Heinemann, Oxford, 1997, p.782.
[8] a) D.H.McDaniel, L.H.Steinert,
J. Am. Chem. Soc. 1966, 88,
4826; b) S.Wasif, J. Chem. Soc. A 1967, 1, 142; c) M.A.Hussein,
G.M.Iskander, M.M.Nour, S.Wasif, H.M.Zeidan,
J. Chem.
Soc. Dalton Trans 1982, 9, 1645.
[9] In a typical reaction, a 50-mL glass-lined, stirred, high-pressure
reactor containing 27 mm of 20 mesh, gold powder and 3m
H2SeO4 (3 mL) in 96% D2SO4 was pressurized with methane
to a final pressure of 27 bar.This reaction mixture was vigorously
stirred and maintained at 1808C for 2 h.After cooling the
mixture, the reaction gas phase was vented into an evacuated
cylinder (to obtained both dissolved and undissolved gases) and
analyzed by GC-MS.The liquid phase was analyzed by 1H and
13C NMR spectroscopy and HPLC.[4d]
[10] This NMR spectrum cannot be used to obtain quantitative
information nor is it representative of the 13CH4 dissolved under
reaction conditions.See reference [9].
[11] Relativistic effects are incorporated into the Los Alamos Core-
Valence Pseudopotential (LACVP) used for Au (P.J.Hay, W.R.
Wadt, J. Chem. Phys. 1985, 82, 299).The 6-31G** + + basis was
employed for other atoms (M.M. Francl, W.J. Pietro, W.J.
Hehre, J.S.Binkley, M.S.Gordon, D.J.DeFrees, J.A.Pople,
J.
Chem. Phys. 1982, 77, 3654).In our experience, approximations
Angew. Chem. Int. Ed. 2004, 43, 4626 –4629
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4629