Experimental and computational study of a direct O2-coupled wacker oxidation: Water dependence in the absence of Cu salts

Article abstract of DOI:10.1021/ja1057218

The kinetics of the Pd[(-)-sparteine]Cl₂ catalyzed oxidation of decene using oxygen as the sole oxidant have been studied in the absence of copper salts and high [Cl^-]. Saturation kinetics are observed for [decene] as well as a third order dependence on [water]. A mechanism is proposed involving the dissociation of two chlorides and rate-limiting formation of a three-water hydrogen bridged network and subsequent oxypalladation as supported by computational studies.

Full text of DOI:10.1021/ja1057218

Published on Web 08/05/2010
Experimental and Computational Study of a Direct O2-Coupled Wacker
Oxidation: Water Dependence in the Absence of Cu Salts
Brian J. Anderson,† John A. Keith,*,‡ and Matthew S. Sigman*,†
Department of Chemistry, UniVersity of Utah, 315 S 1400 E, Salt Lake City, Utah 84112-085, and Institut fu¨r
Elektrochemie, UniVersita¨t Ulm, Albert-Einstein-Allee 47, D-89069 Ulm, Germany
Received May 25, 2010; E-mail: sigman@chem.utah.edu; john.keith@uni-ulm.de
Abstract: The kinetics of the Pd[(-)-sparteine]Cl2 catalyzed
oxidation of decene using oxygen as the sole oxidant have been
studied in the absence of copper salts and high [Cl-]. Saturation
kinetics are observed for [decene] as well as a third order
dependence on [water]. A mechanism is proposed involving the
dissociation of two chlorides and rate-limiting formation of a three-
water hydrogen bridged network and subsequent oxypalladation
as supported by computational studies.
Over 50 years ago Smidt and co-workers at Consortium fu¨r
elektrochemische Industrie, the research organization for Wacker
Chemie, reported the Pd-catalyzed oxidation of olefins to give
aldehydes and ketones using molecular oxygen as the terminal
oxidant.1 The so-called Wacker Process has been used commercially
since then to convert ethylene to acetaldehyde.2 A key to developing
the Wacker process into an industrial success was using CuCl2,
which efficiently links the redox cycle of Pd0 and PdII to one
involving O2. Initial mechanistic details determined by Smidt and
Figure 2. Plots of initial rate vs [reactant] (B, C, and D at 0.3 M [decene]).
co-workers showed that the oxygen atom in the product’s carbonyl
arises from water and not O2, while the protons are conserved from
the starting olefin. Additionally, R-olefins give the Markovnikov
designed by Steinhoff and Stahl.12,13 O2 consumption was monitored
addition product, and exogenous Cl- inhibits the reaction. This
simultaneously in up to six reaction wells, and the substrate conversion
seminal work has inspired chemists through the decades to
was independently confirmed by gas chromatographic analysis. Decene
understand the details of the mechanism and develop improved
was utilized as the standard substrate in DMA/H2O solvent mixtures.
synthetic methodology.3-9 Herein, we report experimental and
Importantly, the reaction was found to be independent of O2 pressure
computational mechanistic studies on olefin oxidation with a unique
above 20 psi. Subsequent reactions were performed at 25 psi O2 and
variant of the Wacker oxidation in the absence of copper salts.
stirred at 800 rpm.14 Initial rates were measured up to 2% conversion.
The oxidation reaction exhibits saturation kinetics in [decene]
(Figure 2A), so subsequent studies were performed under two
limiting conditions: first order in [decene] (0.1 M) and zero order
in [decene] (0.3 M). Substrate saturation kinetics have not been
observed in previous studies of ethylene oxidation.4 A first order
dependence in Pd[(-)-sparteine]Cl2 is observed (Figure 2C), and
the oxidation reaction is inhibited by addition of NaCl although a
noninteger dependence is observed (Figure 2D, vide infra).
Unfortunately, we were unable to measure the dependence on [HCl]
because catalyst decomposition occurs presumably due to proton-
ation of the ligand, (-)-sparteine, under the reaction conditions.
Furthermore, an approximate third order dependence on [H2O] was
observed (Figure 2B), consistent with organization of three waters
prior to or at the rate-limiting step.15 As further evidence of the
role of water, a k(H2O)/k(D2O) of 2.6 ( 0.1 at high [decene] and
1.6 ( 0.1 at 0.1 M [decene] is observed.16 The empirical rate laws
at low and high [decene] are depicted in eqs 1 and 2 respectively.
Figure 1. Copper-free Pd-catalyzed aerobic oxidation of decene to
2-decanone.
Traditional Wacker chemistry and most of its related reactions
use a Cu cocatalyst whose role in the chemistry is usually assumed
to be only the regeneration of the PdII catalyst. This assumption
has been called into question.10 Unlike classical Wacker studies,3a,b,e
we have recently reported a copper-free oxidation of olefins to
methyl ketones using O2 as the sole oxidant and well-defined
catalyst precursor Pd[(-)-sparteine]Cl2 (Figure 1).11 Although
copper-free systems are less studied, they remove many complica-
tions due to Cu salts or high [Cl-] that may affect the mechanism.
Reaction rates of the Pd-catalyzed oxidation of decene in the absence
of Cu salts were measured using a gas uptake system originally
[Pd(-) - (sparteine)Cl2][decene][H2O]3
νobs
)
[Cl-]x
† University of Utah.
‡ Universita¨t Ulm.
at low [decene] (1)
9
11872 J. AM. CHEM. SOC. 2010, 132, 11872–11874
10.1021/ja1057218  2010 American Chemical Society

C O M M U N I C A T I O N S
[Pd(-) - (sparteine)Cl2][H2O]3
νobs
)
[Cl-]x
at high [decene] (2)
Chloride dissociation was measured in DMA/water electrochemi-
cally for Pd[(-)-sparteine]Cl2 revealing dissociation of one chloride.17
Density functional theory (DFT) calculations18 indicate that binding
enthalpy of water and DMA to Pd(II) is stronger than that of Cl- and
H2O by ∼4 and ∼8 kcal/mol, respectively. Since the free energies of
these species in such mixed solvents cannot straightforwardly be
obtained, an analysis of reaction free energies is not currently possible.
Nevertheless, the calculated enthalpies alone indicate a particularly
strong affinity between palladium and DMA, implicating that DMA
will bind to the catalyst in the resting state.
Based on these results, we propose the following mechanism (Scheme
1). From the dissociation constant for Pd[(-)-sparteine]Cl2, A, the catalytic
cycle is initiated with the cationic Pd-species B, which loses another
chloride to form C prior to reversibly binding the olefin. Intermediate D,
the expected precursor to the rate-determining step, then will incorporate
three water molecules to account for the [water] dependence and the water
isotope effect. No kinetic isotope effect was observed when using
2-deutero-undec-1-ene as the substrate, suggesting the rate-determining
step occurs before ꢀ-hydride elimination.
Figure 3. Quadratic fit to the inverse rate vs [NaCl]: (A) at 0.1 M [decene]
and (B) at 0.3 M [decene].
because a simple inverse first order dependence on [Cl-] is not revealed.
Indeed, the derived rate law can be rearranged to solve for [Cl-] resulting
in a quadratic relationship to 1/rate (general expression is eq 5).16 Fitting
this equation to the observed [Cl-] dependence results in an excellent fit
(R2 ) 0.97) and is consistent with both chlorides dissociating prior to the
rate-limiting step (Figure 3).
1
V
) a[Cl-]2 + b[Cl-] + c
(5)
An intriguing aspect of the current catalytic process is that one can
experimentally determine its rate dependence on [H2O]. The observation
that multiple water molecules play a role in Wacker catalysis is consistent
with previous computational reports9a,d,f that have all suggested the lowest-
energy oxypalladation process in the Wacker process occurs via anti-attack
with a three- or four-water hydrogen-bonded chain analogous to our
mechanistic proposal (Scheme 1). Note that none of these simulations
included Cu salts, and it may not be a coincidence that our experimental
results find a related dependence on water for this nucleophilic attack.
Indeed, low-energy enthalpic barriers (5-10 kcal/mol) for external
nucleophilic attack were first believed to show nucleophilic attack was an
equilibrium process in the classic Wacker reaction.9a Inclusion of free
energy contributions from water in solution raise these barriers consider-
ably, however.9c Furthermore, recent state-of-the-art CPMD simulations
on the classic Wacker system by Stirling and Ujaque show anti-
nucleophilic attack is potentially rate-determining in the absence of Cu.9f
With the Pd[(-)-sparteine]Cl2 catalyst, the enthalpic barriers for this
nucleophilic attack are low: +5.2 and +11.9 kcal/mol in continuum
solvation models for H2O and DMA, respectively. Although free energy
contributions cannot readily be calculated, it is possible these will play a
significant role in raising the barrier for nucleophilic attack to be a rate-
determining process.
Some comparisons to the classical Wacker process studies deserve
mention here. The decades-long controversy surrounding the mode
of attack for the nucleophilic addition is challenging to unequivocally
settle due to sensitivity of the mechanism with respect to reaction
conditions.3e In this report, we use a Pd catalyst with a bidentate amine
ligand in a DMA/water mixture while the classical Wacker reaction
uses a PdCl42- catalyst in an aqueous environment with low concentra-
tions of [CuCl2]. Kinetics studies indicate that both complexes release
two Cl- ions into solution in order for the reaction to proceed. Given
the presence of relatively strongly coordinating DMA ligands, an
outersphere reaction mechanism would be expected.3e
Scheme 1. Proposed Mechanism for the Wacker Reaction in the
Absence of Cu Salts
Using the above mechanistic proposal, a rate law was derived with the
assumption that the formation of the three-water chain and subsequent
oxypalladation involving water hydrogen bonding are rate determining.
The rate of the reaction can then be expressed as shown in eq 3.
V ) krds[D][H2O]3
(3)
Due to the observed saturation kinetics, the Pd mass balance must
be solved for [D] using the equilibrium expressions for the two chloride
dissociations and the olefin association (KD1, KD2, KA respectively).16
The derived rate law (eq 4) is consistent with our experimental
observations and shows third order dependence on [water], first order
in [Pd], saturation in [decene], and inhibition by Cl-.
Figure 4 compares the structures and calculated NBO charges
of our anti-oxypalladation processes with those of Pd[ethene]Cl2,
the complex used in previous theoretical studies.9 Despite the
fact the current process involves a dication while theoretical
studies on the classic Wacker process involve a neutral complex,
only very minor differences are seen between the two transition
states. Calculated geometries depict a slightly earlier transition
state, but surprisingly, partial atomic charges (except for Pd)
are nearly identical.
krdsk1KD2KA[Alk][Pd]T[H2O]3
V )
(4)
(k-1 + krds)[Cl-](X)
k1KAKD2[H2O]3[Alk]
(k-1 + krds)[Cl-]
[Cl-]
KD1[L]
KD2[L]
[Cl-]
KAKD2[Alk]
X ) 1 +
+
+
+
[Cl-]
(
)
A key question is whether one or two chlorides dissociate during
catalysis. We derived the rate law assuming that both chlorides dissociate
9
J. AM. CHEM. SOC. VOL. 132, NO. 34, 2010 11873

C O M M U N I C A T I O N S
calculations are consistent with other calculations on the Wacker
process without copper that suggest oxypalladation proceeds through
an anti attack from a three-water hydrogen-bond bridged network. Of
particular interest, the presence of CuCl2 in the reaction mixture has a
dramatic effect on the mechanism. Current studies are ongoing to better
understand the oxypalladation step and the precise role of copper salts
in the Wacker oxidation and to apply this information to the design of
new alkene oxidation catalysts.
Acknowledgment. This work was supported by the National
Institutes of Health (NIGMS RO1 GM3540). The authors thank
Amanda King and Shannon Stahl (University of Wisconsin) as well
as Dale Heisler (University of Utah) for their help in the construc-
tion and programming of the O2-uptake instrument. J.A.K. thanks
the Alexander von Humboldt (AvH) Foundation for funding and
the bwGRiD project20 for computational resources.
Supporting Information Available: Kinetic data, rate law deriva-
tion, and computational data. This material is available free of charge
via the Internet at http://pubs.acs.org.
Figure 4. PdCl2 and Pd[(-)-sparteine] anti-oxypalladation transition states.
Table inset lists partial natural bond order charges (δ) in atomic units and
interatomic distances (r) in Å.
References
(1) Smidt, J.; Hafner, W.; Jira, R.; Sedlmeier, J.; Sieber, R.; Ruttinger, R.;
Kojer, H. Angew. Chem. 1959, 71, 176.
(2) Tsuji, J. New J. Chem. 2000, 24, 127.
Last, we sought to investigate the effect of copper salts on our
Wacker oxidation system. Previous studies have assumed these to
be uninfluential in the olefin oxidation. Interestingly, the addition
of CuCl2 has a dramatic effect on the reaction mechanism (Figure
5). At low concentration of CuCl2 (0.001 M, 1:1 Cu/Pd), the rate
dependence on [decene] is first order, but at higher concentrations
of CuCl2 (10:1 Cu/Pd) a greater than second order dependence on
[decene] is observed.19 If NaCl is substituted for CuCl2, saturation
kinetics are observed similar to results in the absence of CuCl2
demonstrating that the affect of CuCl2 does not arise from [Cl-]
inhibition. As proposed by Hosokawa,10 these results strongly
suggest that Cu salts play a complex role in Wacker chemistry and
are influential in the alkene oxidation process.
(3) For reviews of the Wacker oxidation, see: (a) Negishi, E.-i., Ed. Handbook
of Organopalladium Chemistry for Organic Synthesis; Wiley: New York,
2002; Vol. 2. (b) Henry, P. M. Palladium Catalyzed Oxidation of
Hydrocarbons; Catalysis by Metal Complexes; D. Reidel Publishing Co.:
Dordrecht, 1980; Vol. 2. (c) Takacs, J. M.; Jiang, X.-t. Curr. Org. Chem.
2003, 7, 369. (d) Cornell, C. N.; Sigman, M. S. Inorg. Chem. 2007, 46,
1903. (e) Keith, J. A.; Henry, P. M. Angew. Chem., Int. Ed. 2009, 48,
2–14.
(4) (a) Henry, P. M. J. Am. Chem. Soc. 1964, 86, 3246. (b) Henry, P. M. J. Am.
Chem. Soc. 1966, 88, 1595.
(5) Henry and coworkers evaluated the oxidation of chiral allylic alcohols and
propose syn addition; see:Wan, W. K.; Zaw, K.; Henry, P. M. Organo-
metallics 1988, 7, 1677. It should be noted that allylic alcohol substrates
give a wide range of products. For a review of this topic, see: Muzart, J.
Tetrahedron 2007, 63, 7505.
(6) Stille, J. K.; Divakaruni, R. J. Am. Chem. Soc. 1978, 100, 1303–4.
(7) Ba¨ckvall, J. E.; Akermark, B.; Ljunggren, S. O. J. Chem. Soc., Chem.
Commun. 1977, 264.
(8) (a) Ba¨ckvall, J. E.; Akermark, B.; Ljunggren, S. O. J. Am. Chem. Soc.
1979, 101, 2411. (b) Ba¨ckvall, J. E.; Akermark, B.; Ljunggren, S. O.
J. Chem. Soc., Chem. Commun. 1977, 264.
(9) (a) Siegbahn, P. E. M. J. Phys. Chem. 1996, 100, 14672. (b) Keith, J. A.;
Oxgaard, J.; Goddard, W. A., III. J. Am. Chem. Soc. 2006, 128, 3132. (c)
Keith, J. A.; Nielsen, R. J.; Oxgaard, J.; Goddard, W. A., III. J. Am. Chem.
Soc. 2007, 129, 12342. (d) Beyramabadi, S. A.; Eshtiagh-Hosseini, H.;
Housaindokht, M. R.; Morsali, A. Organometallics 2008, 27, 72. (e) Keith,
J. A.; Nielsen, R. J.; Oxgaard, J.; Goddard, W. A., III; Henry, P. M.
Organometallics 2009, 28, 1618. (f) Comas-Vives, A.; Stirling, A.; Lledo´s,
A.; Ujaque, G. Chem. Eur. J. 2010, DOI: 10.1002/chem.200903522.
(10) (a) Hosokawa, T.; Nomura, T.; Murahashi, S.-I. J. Organomet. Chem. 1998,
551, 387. (b) Hosokawa, T.; Nomura, T.; Murahashi, S.-I. J. Organomet.
Chem. 1998, 566, 293.
(11) Cornell, C. N.; Sigman, M. S. Org. Lett. 2006, 8, 4117.
(12) Steinhoff, B. A.; Stahl, S. S. J. Am. Chem. Soc. 2006, 128, 4348.
(13) Steinhoff, B. A. PhD thesis, University of Wisconsin (USA), 2005.
(14) The reaction stir rate was kept above 800 rpm to avoid mass transfer
problems.
Figure 5. Plots of initial rate vs [decene] in the presence of 0.001 CuCl2
(left) and 0.01 M CuCl2 or NaCl (right).
(15) Note that this experiment requires an ∼10% change in [DMA].
In summary, the kinetics of the Pd[(-)-sparteine]Cl2 catalyzed
oxidation of decene using oxygen as the sole oxidant have been
investigated in the absence of copper salts and high [Cl-]. Our reaction
conditions differ from the original Wacker conditions in that our Pd-
catalyst contains a bidentate amine ligand and the reactions were
performed in a DMA/water mixture. From the kinetic analysis, a
mechanism is proposed involving dissociation of both chlorides
followed by rate-limiting organization of a three-water hydrogen-bond
bridged chain and subsequent oxypalladation. While our studies do
not directly address the long-standing mechanistic question of syn or
anti oxypalladation, the observed experimental kinetics and theoretical
(16) See Supporting Information for more details.
(17) No change of the dissociation constant for Pd[(-)-sparteine]Cl2 was
observed when the ratio of H2O/DMA was altered.
(18) Calculations were carried out with Jaguar 7.5 using the B3LYP/
LACV3P**++ level with implicit solvation parameters for DMA. See
Supporting Information for more details.
(19) Similar kinetic results were obtained if (-)-sparteine was added to chelate
CuCl2 ruling out ligand exchange processes affecting catalysis.
(20) Calculations were run on the bwGRiD (http://www.bw-grid.de), member
of the German D-Grid initiative funded by the Ministry for Education and
Research (Bundesministerium fu¨r Bildung und Forschung) and the Ministry
for Science, Research and Arts Baden-Wu¨rttemberg (Ministerium fu¨r
Wissenschaft, Forschung und Kunst Baden-Wu¨rttemberg).
JA1057218
9
11874 J. AM. CHEM. SOC. VOL. 132, NO. 34, 2010