Propane activation by palladium complexes with chelating bis(NHC) ligands and aerobic cooxidation

Article abstract of DOI:10.1002/anie.201309568

The development of efficient aerobic oxidation methods remains a challenge for the selective functionalization of C-H bonds in alkanes. Herein we report the development of a C-H functionalization procedure for propane by using a palladium catalyst with chelating bis(N-heterocyclic carbene) ligands in trifluoroacetic acid together with a vanadium co-catalyst. Halides play a decisive role in the reaction. The experimental results are presented together with supporting kinetic data and an isotope effect. The reaction can be run with dioxygen as the oxidant if vanadium salts and halides are present in the reaction mixture. Experimental as well as computational results favor a mechanism involving C-H activation by palladium(II), followed by oxidation to palladium(IV) by bromine. Reoxidation by dioxygen: The combination of C-H activation by a homogeneous catalytic palladium complex with a vanadiumoxo co-catalyst allows the selective aerobic oxidation of propane with dioxygen (see scheme). Copyright

Full text of DOI:10.1002/anie.201309568

Angewandte
Chemie
DOI: 10.1002/anie.201309568
À
C H Activation
Propane Activation by Palladium Complexes with Chelating Bis(NHC)
Ligands and Aerobic Cooxidation**
Dominik Munz and Thomas Strassner*
Abstract: The development of efficient aerobic oxidation
methods remains a challenge for the selective functionalization
À
of C H bonds in alkanes. Herein we report the development of
À
a C H functionalization procedure for propane by using
a palladium catalyst with chelating bis(N-heterocyclic carbene)
ligands in trifluoroacetic acid together with a vanadium co-
catalyst. Halides play a decisive role in the reaction. The
experimental results are presented together with supporting
kinetic data and an isotope effect. The reaction can be run with
Scheme 1. Trifluoroacetoxylation of propane catalyzed by chelated
palladium-bis(NHC) complexes.
dioxygen as the oxidant if vanadium salts and halides are
present in the reaction mixture. Experimental as well as
À
computational results favor a mechanism involving C H
activation by palladium(II), followed by oxidation to palla-
dium(IV) by bromine.
moderate differences for the chemo- and regioselectivity
(isopropyl trifluoroacetate: ca. 90%, Scheme 1).
T
he selective functionalization of hydrocarbons and espe-
Other research groups explored the reactivity with alkyl
cially methane has proven to be a formidable challenge and
has been deemed to be the “holy grail” of transition-metal
catalysis.[1] Methane and lower hydrocarbons constitute an
enormous reservoir of comparably clean fossil energy, and
a potential feedstock for the chemical industry. From an
esters and evaluated the industrial relevance of closely related
systems.[6] C H functionalization usually relies on econom-
À
ically unfavorable oxidants, such as K2S2O8 in our case.[2c]
Therefore, further development has to focus on the realiza-
tion of aerobic oxidation procedures.
À
economic perspective, oxidative C H functionalization
Although there has been substantial progress in the field
approaches offer an attractive alternative to conventional
gas-to-liquids processes, as they do not rely on the energy-
intensive formation of syngas.[2] We reported earlier that cis-
chelated palladium(II)-bis(NHC) (NHC = N-heterocyclic
carbene) complexes catalyze the conversion of methane into
methyl trifluoroacetate in trifluoroacetic acid (HOTFA).[3]
The catalytic activity was found to be comparable to the
platinum(II) system with bispyrimidine ligands in oleum
reported by Catalytica.[4] In subsequent studies we investi-
gated the influence of the transition metal (Pt versus Pd), the
halogenido counterions, the chelating bis(NHC) ligand, and
extended our investigations from methane to propane
(Scheme 1).[5] Whereas the catalytic activity turned out to
be highly dependent on the ligand design and the transition
metal employed (TON = 0–30), we only observed very
of homogeneous, transition-metal-catalyzed activation of sp2-
[7]
À
hybridized C H bonds, the selective conversion of alkane
[8]
À
C H bonds using sustainable oxidants still remains a chal-
lenge 40 years after the seminal work by Shilov and Shulpin.[9]
À
The related palladium-catalyzed oxidation of C H bonds in
the presence of a directing group usually relies on strong
oxidants such as peroxides, F+ reagents, halogens, or hyper-
valent iodine species to generate palladium(III) or palladiu-
m(IV) intermediates.[10] However, whereas the reactivity of
palladium(0)[8a,11] with molecular oxygen is comparably well
understood, examples of the oxidation of palladium(II)
compounds by dioxygen remain extremely rare.[8b,12] Thus
we reasoned that a strategy based on a redox co-catalyst for
the activation of dioxygen should be a versatile strategy.[2c,9]
DFT calculations have proven to be a valuable tool for the
[13]
À
investigation of C H activation mechanisms.
Thus, we
decided to study by DFT calculations the functionalization of
methane by a palladium-bis(NHC) catalyst.[14] We proposed,
on the basis of previous work on the synthesis and reactivity
of potential intermediates,[15] that the reaction involves
a palladium catalyst in the oxidation states of + II and + IV,
and that the oxidation step is mediated by bromine
(Scheme 2).
[*] Dr. D. Munz, Prof. T. Strassner
Physikalische Organische Chemie, Technische Universitꢀt Dresden
01069 Dresden (Germany)
E-mail: thomas.strassner@chemie.tu-dresden.de
[**] We thank the Center for Information Services and High Performance
Computing (ZIH) for computation time and the Deutsche For-
schungsgesellschaft (DFG) (STR 526/7-1 and STR 526/7-2) for
financial support. D.M. thanks the Studienstiftung des Deutschen
Volkes for support. Donations of trifluoroacetic acid by Solvay Fluor
are gratefully acknowledged. NHC=N-heterocyclic carbene.
Therefore, we envisioned a combination of the catalytic
À
cycle for the C H activation with a second redox cycle
involving oxygen activation.
Supporting information for this article (experimental details) is
available on the WWW under http://dx.doi.org/10.1002/anie.
201309568.
Based on the experimental observation of bromine fumes
at a high catalyst loading as well as, in the case of methane, the
Angew. Chem. Int. Ed. 2014, 53, 2485 –2488
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2485

.
Angewandte
Communications
importance of generating free coordination sites at the
palladium catalyst. Catalysis by a heterogeneous palladium
catalyst appears unlikely considering the large differences in
the catalytic activity of the different catalysts investigated
herein and those reported before with different bis(NHC)
ligands.[5c]
Next, we examined a series of oxidants which are well
known to oxidize bromide rapidly under acidic conditions. As
anticipated, we found catalytic turnover with the oxidants
HAuCl4, KMnO4, H2O2, Pb(OAc)4, and F+ reagents. Select-
fluor[10b] gave the same results as K2S2O8. Oxidants with
a lower redox potential than the bromine/bromide redox
couple, but which have been reported to oxidize palladium(0),
such as, for example Cu(OAc)2 + O2, NaVO3, Ag(OTFA), or
tetrachloro-1,4-benzoquinone, turned out to be completely
inactive!
Scheme 2. Proposed mechanism.
inhibition of the reaction on addition of KBr[5b] we decided to
study the influence of halides in the reaction with K2S2O8 as
the oxidant. Consequently, the catalytic activity of a series of
bis(NHC) catalysts with different halogenido counterligands
(Figure 1) was investigated in the reaction with propane.
Peroxodisulfate and polyhalides can react with alkanes
and trifluoroacetic acid by radical pathways.[16,17] To exclude
that carbon-centered radicals participate in the reaction and
À
to check whether polyhalomethanes mediate the C H
functionalization we added the radical-transfer reagents
CBrCl3 and CCl4 in large excess to the reaction mixture
(4.2 mmol catalyst, 210 mmol CX4, 210 mmol K2S2O8, 1 bar
propane, 2.5 mL HOTFA, 608C, 17 h). We did not observe the
formation of considerable amounts of chloro- and bromopro-
pane (CCl4: TON = 16; CCl3Br: TON = 9), as has been
reported for radical mechanisms under comparable reaction
conditions.[18] Furthermore, as the reaction seems not to be
sensitive to the presence of dioxygen, we conclude that
radicals do not play a role during the course of the reaction.
Inspired by the work of Hirao and co-workers,[19] who
showed that vanadium salts catalyze the aerobic oxidation of
bromide under acidic conditions, we then explored the
oxidation of propane in the presence of NaVO3 and dioxygen.
As expected, isopropyl trifluoroacetate formed at 908C in
trifluoroacetic acid (25 mL, HOTFA, 84 mmol 1_Br2;
840 mmol NaVO3, 17 h, 6 bar propane, 4 bar dioxygen:
TON = 3). The addition of related vanadium precatalysts
such as NH4VO3 or VO(OAc)2 led to a comparable level of
catalytic activity. The product distribution obtained matched
the regio- and chemoselectivity when K2S2O8 and Selectfluor
were used as oxidants (Figure 1), although we also identified
the formation of traces of methyl and ethyl trifluoroacetate as
well as isopropyl bromide. During the course of the reox-
idation process, one equivalent of water forms per molecule
Figure 1. Dependence of catalytic activity on counterligands. Reaction
conditions: 4.2 mmol catalyst, 210 mmol K2S2O8, 1 bar propane, 2.5 mL
HOTFA, 608C, 17 h.
We found a comparable catalytic activity for the catalysts
1_Br2 (TON = 18) and 1_Cl2 (TON = 17) but a significantly
lower activity for 1_I2 (TON = 6), while all the catalysts led to
a comparable regio- and chemoselectivity. This is expected, as
bromine and chlorine are known to oxidize 1_Br2 to the
corresponding palladium(IV) complex, whereas the oxidation
strength of iodine is not sufficient for this reaction. The
catalytic activity of 1_(OTFA)2 and Pd(OAc)2 (TON = 5 and
TON = 6) was found to be substantially lower. In those
catalyst systems, and likewise for 1_I2, a different mechanism
for the reoxidation of the palladium centers is clearly
operative. The tetrafluoroborate complex 1_(MeCN)2(BF4)2
with coordinating acetonitrile ligands (TON = 1) as well as
the corresponding platinum complex Pt1_Br2 (TON = 0)
turned out to be inactive. The last result in particular
underlines the significance of the nature of the transition-
metal center.
À
of ester. As described above, water inhibits the C H
activation cycle. We therefore examined the effect of the
addition of trifluoroacetic anhydride (21 mL HOTFA + 4 mL
TFA2O) and found, as expected, a remarkable increase in
activity (908C: TON = 10, 1008C: TON = 12). Only traces of
CO2 were detected at the end of the reaction, which points
toward negligible decomposition of HOTFA by radical
pathways. The catalytic activity is highly preserved at lower
catalyst loadings and TONs of over 60 for the palladium
catalyst could be obtained (Figure 2). The reaction is also
clearly catalytic in NaVO3 (TON = 8). The catalytic activity
remained high after 65 h as long as not too much water had
accumulated in the reaction mixture and not all the dioxygen
had been consumed.
As in the case of methane,[5b] the reaction with propane is
strongly inhibited by the addition of two or five equivalents of
potassium bromide to the reaction mixture (4.2 mmol catalyst,
8.4 mmol or 21 mmol KBr, 210 mmol K2S2O8, 1 bar propane,
2.5 mL HOTFA, 608C, 17 h: TON = 4 and TON = 2) or by the
addition of water. The addition of trifluoroacetic anhydride
only led to a slight inhibition (for details see the Supporting
Information). These observations taken together highlight the
2486 www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 2485 –2488

Angewandte
Chemie
after a reaction time of 23 h. Remarkably, the reaction
proceeded with a comparable initial reaction rate if we added
new oxidant after 23 h. The reaction rate shows a first order
dependence on the catalyst and the substrate[20] concentra-
tion. We were able to determine the Arrhenius relationship to
get an estimate of the reaction barrier and important
information on the rate-determining step of the catalytic
cycle. From the experimental data we calculated a reaction
barrier of (29.0 Æ 1.6) kcalmolÀ1, which is in good agreement
with reported values for related reactions.[13a] To gain more
information on the rate-determining state of the reaction, we
then measured the isotope effect for the model substrate
cyclohexane.[20] We obtained an isotope effect (IE) of 3.4 Æ
0.8, which is indicative of a primary IE. This value does not
Figure 2. Variation of the concentration of the palladium catalyst 1_Br2
and the reaction time. 21 mL HOTFA+4 mL TFA2O, 168 mmol NaVO3,
6 bar propane, 4 bar dioxygen, 908C.
In agreement with reports in the literature,[18a] we also
observed a blue-colored, paramagnetic reaction mixture
when all the dioxygen had been consumed. This blue color,
which rapidly vanished upon exposure to air, has been
attributed to a reduced vanadium(III) or vanadium(IV)
species.[18a]
À
only confirm that the C H bond is broken in the rate-
determining step of the reaction, but is also in good agree-
ment with reports in the literature[21] on the carbonylation of
cyclohexane by palladium catalysts (IE = 3.2) and our DFT
calculations (IE = 3.9).
À
To verify that the same C H activation mechanism is
In conclusion, the mechanism of the palladium-bis(NHC)-
catalyzed functionalization of propane has been studied in
detail. There is a very strong indication that the catalytic cycle
proceeds via palladium in the oxidation state + IV and that
neither radical nor heterogeneous reactivity play an impor-
tant role. Furthermore, there is evidence that halides are
involved in the reaction. We demonstrated that the palla-
operative with the vanadium co-catalyst as in the case of the
oxidants K2S2O8 or Selectfluor, we examined the catalytic
activity of different catalysts and KBr in the presence of the
vanadium co-catalyst. No significant catalytic activity (TON
ꢀ 0) was obtained for the platinum complex Pt1_Br2, the
palladium trifluoroacetate complex 1_(OTFA)2, palladium
acetate, palladium bromide, or in the absence of any
palladium catalyst. In the presence of KBr (84 mmol KBr:
TON = 1, 84 mmol Pd(OAc)2 + 0.17 mmol KBr: TON = 1) we
determined minor background reactions, which we attribute
to vanadium catalysis and/or radical reactivity. Evidently, the
concomitant presence of a palladium catalyst with a bis(NHC)
ligand, a halide source, and the vanadium salt are necessary
for an efficient oxidation reaction.
À
dium-catalyzed C H activation can be coupled through
a bromide/bromine redox pair to an aerobic cooxidation
procedure based on NaVO3. We therefore present a very
promising and new cooxidation concept for the development
of catalytic reactions with palladium in high oxidation states
(+ III, + IV) and dioxygen used as the terminal oxidant.
Received: November 4, 2013
Revised: December 2, 2013
Published online: January 31, 2014
On the basis of all of our results we therefore propose that
the palladium catalyst is responsible for the C H function-
alization, while the vanadium catalyst mediates the dioxygen
activation (Scheme 3). We expect that both catalytic cycles
are connected by a bromide/bromine redox pair, which
mediates the oxidation step of the alkane functionalization.
À
À
Keywords: C H activation · oxygen · palladium ·
reaction mechanisms · vanadium
.
[1] B. A. Arndtsen, R. G. Bergman, T. A. Mobley, T. H. Peterson,
Acc. Chem. Res. 1995, 28, 154 – 162.
[2] a) R. H. Crabtree, Chem. Rev. 1995, 95, 987 – 1007; b) J. R.
Webb, T. BolaÇo, T. B. Gunnoe, ChemSusChem 2011, 4, 37 – 49;
c) E. G. Chepaikin, Russ. Chem. Rev. 2011, 80, 363 – 396; d) J. A.
Labinger, J. E. Bercaw, Nature 2002, 417, 507 – 514; e) H.
Schwarz, Angew. Chem. 2011, 123, 10276 – 10297; Angew.
Chem. Int. Ed. 2011, 50, 10096 – 10115.
[3] M. Muehlhofer, T. Strassner, W. A. Herrmann, Angew. Chem.
2002, 114, 1817 – 1819; Angew. Chem. Int. Ed. 2002, 41, 1745 –
1747.
[4] B. G. Hashiguchi, S. M. Bischof, M. M. Konnick, R. A. Periana,
Acc. Chem. Res. 2012, 45, 885 – 898.
Scheme 3. Proposed redox cycles.
[5] a) S. Ahrens, A. Zeller, M. Taige, T. Strassner, Organometallics
2006, 25, 5409 – 5415; b) T. Strassner, M. Muehlhofer, A. Zeller,
E. Herdtweck, W. A. Herrmann, J. Organomet. Chem. 2004, 689,
1418 – 1424; c) D. Munz, A. Poethig, A. Tronnier, T. Strassner,
Dalton Trans. 2013, 42, 7297 – 7304; d) D. Meyer, M. A. Taige, A.
Zeller, K. Hohlfeld, S. Ahrens, T. Strassner, Organometallics
2009, 28, 2142 – 2149; e) S. Ahrens, T. Strassner, Inorg. Chim.
Acta 2006, 359, 4789 – 4796.
Finally, we looked at various aspects of the reaction with
the oxidant K2S2O8 to assess the plausibility of the proposed
mechanism: We found that the reaction follows a pseudolinear
kinetic model, at least for the first 15 h (ca. 50% yield based
on the oxidant). We observed neither a significant induction
period nor signs of decomposition of the palladium catalyst
Angew. Chem. Int. Ed. 2014, 53, 2485 –2488
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 2487

.
Angewandte
Communications
[6] a) X.-F. Hou, Y.-N. Wang, I. Gçttker-Schnetmann, Organome-
tallics 2011, 30, 6053 – 6056; b) M. Golombok, T. Nijbacker, M.
Raimondi, Ind. Eng. Chem. Res. 2004, 43, 6001 – 6005.
[7] a) C. S. Yeung, V. M. Dong, Chem. Rev. 2011, 111, 1215 – 1292;
b) L. Ackermann, Chem. Rev. 2011, 111, 1315 – 1345; c) F.
Glorius, J. Wencel-Delord, T. Drçge, F. Liu, Chem. Soc. Rev.
2011, 40, 4740 – 4761; d) T. W. Lyons, M. S. Sanford, Chem. Rev.
2010, 110, 1147 – 1169; e) J. Choi, A. H. R. MacArthur, M.
Brookhart, A. S. Goldman, Chem. Rev. 2011, 111, 1761 – 1779.
[8] a) A. N. Campbell, S. S. Stahl, Acc. Chem. Res. 2012, 45, 851 –
863; b) L. Boisvert, K. I. Goldberg, Acc. Chem. Res. 2011, 45,
899 – 910.
[13] a) D. Balcells, E. Clot, O. Eisenstein, Chem. Rev. 2010, 110, 749 –
823; b) Y. Boutadla, D. L. Davies, S. A. Macgregor, A. I.
Poblador-Bahamonde, Dalton Trans. 2009, 5820 – 5831.
[14] D. Munz, D. Meyer, T. Strassner, Organometallics 2013, 32,
3469 – 3480.
[15] a) D. Meyer, Dissertation Thesis, TU Dresden (Germany), 2011;
b) D. Meyer, S. Ahrens, T. Strassner, Organometallics 2010, 29,
3392 – 3396; c) A. S. McCall, H. Wang, J. M. Desper, S. Kraft, J.
Am. Chem. Soc. 2011, 133, 1832 – 1848; d) S. S. Subramanium,
L. M. Slaughter, Dalton Trans. 2009, 6930 – 6933; e) S. N.
Sluijter, S. Warsink, M. Lutz, C. J. Elsevier, Dalton Trans. 2013,
42, 7365 – 7372.
[9] A. E. Shilov, G. B. Shulpin, Chem. Rev. 1997, 97, 2879 – 2932.
[10] a) L. M. Xu, B. J. Li, Z. Yang, Z. J. Shi, Chem. Soc. Rev. 2010, 39,
712 – 733; b) K. M. Engle, T. S. Mei, X. Wang, J. Q. Yu, Angew.
Chem. 2011, 123, 1514 – 1528; Angew. Chem. Int. Ed. 2011, 50,
1478 – 1491; c) D. C. Powers, T. Ritter, Acc. Chem. Res. 2012, 45,
840 – 850.
[11] a) Z. Z. Shi, C. Zhang, C. H. Tang, N. Jiao, Chem. Soc. Rev. 2012,
41, 3381 – 3430; b) B. V. Popp, S. S. Stahl, Top. Organomet.
Chem. 2007, 22, 149 – 189; c) M. Zhou, R. H. Crabtree, Chem.
Soc. Rev. 2011, 40, 1875 – 1884; d) K. J. Stowers, A. Kubota, M. S.
Sanford, Chem. Sci. 2012, 3, 3192 – 3195.
[12] a) A. N. Vedernikov, Acc. Chem. Res. 2012, 45, 803 – 813; b) J. R.
Khusnutdinova, N. P. Rath, L. M. Mirica, J. Am. Chem. Soc.
2010, 132, 7303 – 7305; c) F. Z. Tang, Y. Zhang, N. P. Rath, L. M.
Mirica, Organometallics 2012, 31, 6690 – 6696; d) J. R. Khusnut-
dinova, F. R. Qu, Y. Zhang, N. P. Rath, L. M. Mirica, Organo-
metallics 2012, 31, 4627 – 4630; e) J. R. Khusnutdinova, N. P.
Rath, L. M. Mirica, J. Am. Chem. Soc. 2012, 134, 2414 – 2422;
f) G. J. Chuang, W. K. Wang, E. Lee, T. Ritter, J. Am. Chem. Soc.
2011, 133, 1760 – 1762; g) P. L. Alsters, H. T. Teunissen, J.
Boersma, A. L. Spek, G. van Koten, Organometallics 1993, 12,
4691 – 4696.
[16] I. Akhrem, A. Orlinkov, Chem. Rev. 2007, 107, 2037 – 2079.
[17] a) M. V. Kirillova, A. M. Kirillov, A. J. L. Pombeiro, Chem. Eur.
J. 2010, 16, 9485 – 9493; b) M. Zerella, S. Mukhopadhyay, A. T.
Bell, Org. Lett. 2003, 5, 3193 – 3196; c) A. A. Fokin, P. R.
Schreiner, Chem. Rev. 2002, 102, 1551 – 1594.
[18] M. V. Kirillova, M. L. Kuznetsov, P. M. Reis, J. A. da Silva, J. J.
da Silva, A. J. Pombeiro, J. Am. Chem. Soc. 2007, 129, 10531 –
10545.
[19] a) K. Kikushima, T. Moriuchi, T. Hirao, Tetrahedron 2010, 66,
6906 – 6911; b) K. Kikushima, T. Moriuchi, T. Hirao, Tetrahedron
Lett. 2010, 51, 340 – 342; c) T. Moriuchi, K. Kikushima, T.
Kajikawa, T. Hirao, Tetrahedron Lett. 2009, 50, 7385 – 7387; d) K.
Kikushima, T. Moriuchi, T. Hirao, Chem. Asian J. 2009, 4, 1213 –
1216.
[20] As a consequence of the high solubility of propane in HOTFA,
the experiments on the rate law for propane and the isotope
effect have been conducted instead with the model substrate
cyclohexane.
[21] K. Nakata, T. Miyata, T. Jintoku, A. Kitani, Y. Taniguchi, K.
Takaki, Y. Fujiwara, Bull. Chem. Soc. Jpn. 1993, 66, 3755 – 3759.
2488 www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 2485 –2488