4
A.M. Bauer et al. / Tetrahedron Letters 60 (2019) 151193
but the two oxidants gave comparable results with triisopropyl
phosphite. The phosphine substrates gave higher percent conver-
sions than phosphites. Pyridine-N-oxide gave better results than
x-PVP-N-oxide for tricyclohexylphosphine, but x-PVP-N-oxide led
to slightly better results for tri-n-butylphosphine. Despite the fact
that pyridine-N-oxide resulted in a higher percent conversion for
three of the four substrates, the overall results for both oxidants
were quite similar in direct oxidation reactions. This suggests that
while there will likely be minor differences in reactivity between a
molecular OAT reagent and a heterogeneous reagent with similar
structures, it is reasonable to expect comparable results between
the two in direct substrate oxidation reactions.
Recycling the polymer backbone
Fig. 5. Percent conversion of phosphorus trichloride to phosphorus(V) oxychloride
with recycled x-PVP-N-oxide.
One advantage of introducing a heterogeneous reagent is the
opportunity to recycle the material and use it for multiple reac-
tions. In the case of x-PVP-N-oxide, the polymer backbone would
need to be re-oxidized in a separate step after each OAT reaction
with a substrate (Scheme 2). Phosphorus trichloride was selected
as a substrate because high percent conversions were achieved at
room temperature. It was desirable to keep x-PVP-N-oxide in the
same reaction vessel for all steps of the recycling experiments to
minimize inevitable loss of solid when transferring vessels. Re-oxi-
dizing x-PVP to x-PVP-N-oxide at room temperature using meta-
chloroperoxybenzoic acid (mCPBA), according to the method of
Sadun [30], was a desirable strategy because it would allow both
OAT to phosphorus trichloride and re-oxidation to x-PVP-N-oxide
to be accomplished at room temperature in the same reaction
vessel. All filtration and washing steps were also designed to keep
x-PVP-N-oxide in the same reaction vessel by drawing solutions
out of the vessel with a thin 21.5 gauge needle attached to a syr-
inge. In this way, loss of the heterogeneous x-PVP-N-oxide would
be minimized over the course of multiple cycles. Despite these
precautions, some loss of x-PVP-N-oxide was still observed in each
Catalytic reactions
N-oxides have been used as the stoichiometric oxidant in metal-
catalyzed oxidation reactions [31,32]. It was therefore logical to
test if x-PVP-N-oxide could interact with metal catalysts in oxida-
tion reactions. In order to test this, the oxidation of a-diazo-benze-
neacetic acid methyl ester to methyl benzoylformate using copper
triflate catalyst, Cu(OTf)2, was attempted with x-PVP-N-oxide as
the stoichiometric oxidant (Table 2) [32].
The reactions presented here included slight modifications of
the literature procedure, which employed 1,2-dichloroethane as
the solvent, 4 Å molecular sieves as an additive, and 12 h of stirring
[32]. In this present study, molecular sieves were not used, reac-
tions were stirred for 18 h, and acetonitrile was chosen as the sol-
vent, as acetonitrile has been shown to be the most effective
solvent when using x-PVP-N-oxide. As expected, similar reactivity
was observed when using one equivalent of pyridine-N-oxide as
the OAT reagent (74% conversion to methyl benzoylformate, entry
2, average of two runs) relative to the reported reaction that used
4-phenylpyridine-N-oxide as the OAT reagent (71% isolated yield
of 4-phenylpyridine, entry 1). Once comparable results to the liter-
ature were achieved, 4.6 equivalents of x-PVP-N-oxide were sub-
stituted for pyridine-N-oxide while keeping all of the other
conditions the same as entry 2. With x-PVP-N-oxide in the catalytic
reaction, only 3.9% conversion to methyl benzoylformate was
achieved (entry 3, average of three runs). This dramatic change
in reactivity was accompanied by conversion of the starting mate-
rial to multiple, unidentified products besides methyl benzoylfor-
mate (Supporting Information). That uncontrolled, off-cycle
catalytic reactions were occurring suggested that the heteroge-
neous x-PVP-N-oxide was not interacting efficiently with the
catalyst.
To evaluate the effect of temperature on the interaction of
x-PVP-N-oxide with the copper catalyst, the reaction was run at
110 °C (entry 4). The results of the reaction at 110 °C were essen-
tially identical to the reaction at 60 °C, with multiple unidentified
products and only a 4.8% conversion to methyl benzoylformate
(average of three runs). These results demonstrated that
substituting the heterogeneous x-PVP-N-oxide for the molecular
pyridine-N-oxide fundamentally altered the catalytic reaction.
To determine if x-PVP-N-oxide had any interaction with the
copper catalyst in the reactions represented by entries 3 and 4
(3.9% conversion and 4.8% conversion, respectively), two control
reactions were run, one in the absence of the copper triflate cata-
lyst (entry 5) and one in the absence of x-PVP-N-oxide (entry 6).
Only a trace of methyl benzoylformate product was observed in
the absence of catalyst (0.5%, entry 5) and no product was detected
in the absence of oxidant (entry 6), confirming that x-PVP-N-oxide
was the source of the oxygen atoms and that x-PVP-N-oxide
step. A higher ratio of x-PVP-N-oxide (54 mg, 450
lmol, 4.5 equiv-
alents) to phosphorus trichloride (100 mol) was used in compar-
l
ison to the standard conditions (2.2 equivalents oxidant) to help
compensate for the loss of solid during recycling steps. The results
of the recycling experiments are shown in Fig. 5.
As expected, the first phosphorus trichloride oxidation
resulted in full conversion to phosphorus(V) oxychloride. Subse-
quent re-oxidation of the x-PVP and x-PVP-N-oxide mixture was
facile using mCPBA (100 lmol). Washing the regenerated x-PVP-
N-oxide (3 ꢀ 3 mL acetonitrile, 3 ꢀ 3 mL 0.14 M triethylamine
solution in acetonitrile, 3 ꢀ 3 mL acetonitrile) and drawing sol-
vent out of the reaction vial, however, did result in some loss of
the solid in each step. This led to lower percent conversions in
the second oxidation of phosphorus trichloride (88%) and also
the third oxidation (57%). Despite decreasing percent conversions,
it was clear that x-PVP-N-oxide could be recycled. Characteriza-
tion of the recycled x-PVP-N-oxide was not attempted at the
end of the recycling reactions because too much of the material
was lost in the washing steps.
Scheme 2. Basic concept for recycling x-PVP-N-oxide.