.
Angewandte
Communications
sion NMR experiments were performed to determine the
Stokes radius of 1 and to confirm its structure in solution.[29]
The diffusion coefficient for 1 in [D3]MeCN at À108C was
(0.541 Æ 0.087) ꢁ 10À5 cm2 sÀ1. Assuming that the shape of
1 can be approximated as a sphere, this yields a radius of
(7.0 Æ 1.1) ꢀ, a value in good agreement with the crystal
structure of 1, in which the outermost hydrogen atoms lie
between 5 and 8 ꢀ from the center of the molecule.
A series of mass spectrometry and catalysis experiments
were performed to assess the role of 1 in catalytic alcohol
oxidations with the dimeric precursor 22(OTf)2. The speci-
ation of the Pd compounds was monitored during the
oxidation of 0.14m 1,2-propanediol catalyzed by 22(OTf)2
(7 mm, 10 mol% Pd) in MeCN by measuring the ESI-MS of
diluted aliquots at various time points (species that showed
significant variation during catalysis are plotted in Figure 4,
observed are present in the reaction mixture rather than being
generated in the course of microdroplet evaporation.
To assess the chemical and kinetic competence of 1 as an
intermediate in the catalytic oxidation reaction, the isolated
complex 1(BF4)2 was investigated for its activity in the
catalytic oxidation of 1,2-propanediol at room temperature.
Oxidation of 1,2-propanediol (0.3m, MeCN) occurred readily
under 1 atm of air in the presence of 0.83 mm of 1(BF4)2
(2.5 mm Pd, 0.83% loading). The initial rate of alcohol
oxidation and TON (TON = mmol diol consumed/mmol Pd,
see
Table S2)
with
1(BF4)2
[ratei = (0.77 Æ 0.08) ꢁ
10À1 mm minÀ1] are lower than those with 22(BF4)2 [ratei =
(6.06 Æ 0.26) ꢁ 10À1 mm minÀ1] under similar conditions.[40]
Furthermore, a significant amount of a black precipitate,
likely Pd black, formed during catalysis with 1(BF4)2, thus
suggesting that LPd0 oxidation was not efficient under these
conditions. By contrast, no Pd black is observed for the
oxidation of 1,2-propanediol with 22(BF4)2 under these
conditions. The addition of 3 equiv of HOAc relative to
1 led to an increase in the initial rate and TON [rate = (0.98 Æ
0.11) ꢁ 10À1 mm minÀ1] of diol oxidation with 1, however the
activity was lower than that of 22(BF4)2 and Pd black still
formed. By contrast, when a 2:1 cocatalytic mixture of 1(BF4)2
and [LPd(MeCN)2](BF4)2 (7(BF4)2) was used (0.55 mm 1-
(BF4)2, 0.83 mm 7(BF4)2, 2.5 mm total Pd,), the initial rate of
1,2- propanediol conversion and TON per Pd atom [rate =
(5.84 Æ 0.23) ꢁ 10À1 mm minÀ1] were indistinguishable to those
for 22(BF4)2 at the same Pd loading.[41] Under these conditions,
the formation of Pd black was significantly reduced (see the
Supporting Information). This result suggests that some
amount of LPdII with exchangeable ligands is required for
efficient oxidation of LPd0 when 1 is used as a catalyst
precursor.
Figure 4. ESI-MS reaction monitoring. [Pd] for each sample=0.2 mm,
ESI potential=2.0 kV. Intensities for each time point are normalized to
the total ion current at that time point. Peaks areas for related species
are summed and plotted against the left axis. The rate of 1,2-propane-
diol conversion (black circles), measured between two time points, is
plotted against the right axis. Lines connecting data points are
provided merely to illustrate the trends in the data.
The catalytic activity of 1(BF4)2 in combination with
7(BF4)2 indicates that the trinuclear Pd3O2 complex 1(BF4)2 is
both a chemically and kinetically competent intermediate in
the catalytic oxidation of 1,2-propanediol. That both 1(BF4)2
and 7(BF4)2 are needed to achieve a similar rate to that
achieved with the cationic dimer 22(BF4)2 implies that the
efficient activation and reduction of O2 to H2O is facilitated
by the cooperative behavior of multiple Pd centers.[11,34,42]
In Figure 5, we present a mechanistic hypothesis in which
others are described and plotted in Figure S4). An independ-
ent reaction was carried out under identical conditions except
that the aliquots were passed through silica plugs and
analyzed by gas chromatography with flame ionization
detection (GC–FID) to measure the rate of 1,2-propanediol
oxidation. As shown in Figure 4, the appearance and decay of
the trinuclear compound 1 correlate with the rate of 1,2-
propanediol consumption. While this correlation is only
qualitative, given the challenges in quantifying complex
mixtures by ESI-MS, the data nevertheless imply that 1 is
associated with the catalytic oxidation reaction in MeCN. To
assess whether 1 might have been formed during the ESI
ionization process, a similar monitoring experiment was
carried out with nanoelectrospray ionization (nanospray)
MS[37] and provided similar results to those from the ESI-MS
experiments (see the Supporting Information). Because
ionization by nanospray occurs ca. 103 times faster than by
ESI,[38,39] the observation of similar species in both the ESI-
MS and nanospray-MS implies that 1 and the other Pd species
the Pd3O2 complex 1 liberates the catalytically active m-OH
[15,20]
compound 92
and [LPdII]2+ (7) in the presence of acid.
Oxidation of the alcohol would generate [LPd-H]+ and
subsequently the h2-peroxo Pd complex LPd(O2) (8).[30]
Regeneration of the Pd3O2 complex 1 could occur in
a manner similar to that utilized in its synthesis (Scheme 1),
where coupling of the peroxo complex 8 with [LPdII]2+ (7)
generates the proposed dicationic m-peroxo 10, which reacts
with another unit of 8, thereby displacing O2 in order to form
1 (see the Supporting Information for DFT calculations).[43]
Alternatively, 10 might also be generated from hydrogen
peroxide generated from protonolysis of peroxo 8.[30] That
complex 1 can be formed in aprotic solvents indicates that
hydrogen peroxide is not necessary to generate 10 or 1, but
these intermediates may be relevant to the proposed dis-
proportionation of hydrogen peroxide.[45,46]
5650
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 5648 –5652