work with hypberbranched grafts on gold surfaces or thiophene-
containing hyperbranched grafts on PE films had indicated that
these functional grafts were ultrathin (10–100 nm thick) and
2-dimensionally uniform. To determine that the chemistry
above results in uniform incorporation of phosphine ligands and
Pd catalysts on these surfaces, we prepared a Pd(0)-DPPA/
3-PAA sample using a hyperbranched 3-PAA/PE film sample.8
Multipoint XPS analysis over 1 mm region (Fig. 2) of the
resulting film shows uniform surface functionalization. Nota-
bly, the average P/Pd ratio of ca. 3/1 determined with this
experiment suggests that these surface-bound catalysts are
similar in coordination environment to the active catalysts in a
typical reaction using (PPh3)4Pd. Earlier work with thiophenes
polymerization has demonstrated that these lightly crosslinked
grafts are flexible enough to oligomerize so formation of bis-
and tris-phosphine ligated Pd(0) catalysts seems plausible.
However, the known reactivity of Pd(0) crystallites themselves
in allylic substitution chemistry and the known feasibility of
using added phosphine ligands to promote this reactivity led us
to briefly examine the possibility that our Pd(0) catalysts were
actually just Pd(0) crystallites activated by phosphine ligands in
the matrix. Several experiments were performed. First, XPS
analysis of Pd crystallites in a hyperbranched graft (Pd(0)-
4PAA/PE, prepared by reduction of a Pd carboxylate salt)
showed a Pd 3d3/2 peak at 333.25 eV. This peak was discernibly
different than the peak for the molecular catalyst—a phosphine-
ligated Pd (Pd(0)-DPPA/4-PAA). Next, we examined the
reactivity of the crystallites with and without external triphenyl-
phosphine in allylic substitution of allyl acetate by piperidine on
a 20 mmol scale. While low activity was seen without external
ligand, complete conversion to the allylic amine was seen in 16
h using 0.05 mol% Pd catalyst with 0.6 mol% added
triphenylphosphine. However, unlike the results with a Pd(0)-
DPPA/4-PAA catalyst, ca. 2–3% Pd leaching was seen. We
earlier noted that molecular Pd(0)-DPPA/4-PAA catalysts
could be converted to phosphine oxides and Pd(0) crystallites
by refluxing the yellow Pd(0)-DPPA/4-PAA powder in metha-
nol in air to form a grey powder. This grey powder too had low
activity in allylic amination of allyl acetate. As was true with
other films containing Pd(0) crystallites, activity was restored
on addition of external triphenylphosphine but at the cost of
probable leaching of Pd. A final experiment used DPPA to
modify unreacted –CO2H groups in a hyperbranched graft
containing Pd(0) crystallites (a PE/4-PAA/Pd(0) powder).10
This powder containing both Pd crystallites and a covalently
bound phosphine did have some activity in allylic amination but
the activities were ca. 10-fold lower than those of the molecular
Pd(0)-DPPA/4-PAA catalysts.
Fig. 2 Multipoint XPS analysis of a Pd(0)-DPPA/3-PAA/PE film: -,
atom% values for Pd (± 0.1); :, atom% values for N (± 0.1); 1, atom%
values for P (± 0.4). Atom% values for C and O across this 1 cm section of
film averaged 78.9 (± 1.6) and 14.4% (± 1.4), respectively.
This phosphine-ligated Pd(0) catalyst/PE powder was active
for allylic substitution chemistry. These reactions (Scheme 2)
were carried out at room temperature in the absence of solvent.†
The regioselectivity of the reaction between cinnamyl acetate
and piperidine is similar to earlier work we reported using a
soluble diphenylphosphine-terminated polyethylene ligand for
Pd(0).9 The catalyst was recycled up to 5 times without a
noticeable decrease in activity. If the catalysts were exposed to
air, oxidation and catalyst deactivation did occur. This oxidation
was verified by 31P MAS NMR spectroscopy. Indeed, in all
cases when catalytic activity was significantly slower and where
the polymer was recovered and analyzed, 31P-MAS NMR
spectroscopy showed that phosphine ligand oxidation had
occurred presumably due to adventitious oxidation. Such
catalyst deactivation was also detectable visually—the yellow
active catalyst becomes darker when the activity has noticeably
decreased. This oxygen sensitivity makes recovery and reuse of
these catalysts experimentally more difficult and tedious.
Support by the R.W. Welch Foundation, the National Science
Foundation (DMR grant # 9977911) and ArQule is gratefully
acknowledged.
Notes and references
† Cinnamyl acetate (1 g, 5.7 mmol) and Et2NH (1.67 g, 22.8 mmol) were
combined in a 15 mL flask and this mixture was degassed 33 by freeze/
pump/thaw. A 40 mL centrifuge tube with Pd(0)-DPPA/4-PAA/PE powder
catalyst (200 mg, 0.0081 mmol of Pd) was evacuated and backfilled with N2
33. The reaction mixture was transferred by forced syphon through cannula
into the catalyst-containing tube which was degassed 33 more. This
mixture was mixed for 2 h by magnetic stir bar, during which time
diethylamine-HOAc salt precipitated out forming a thick mixture. The
product was isolated from three 10 mL portions of distilled degassed THF
that were first transferred into the centrifuge tube and then separated by
centrifugation and forced siphon from the more dense powder. The THF
was then removed under reduced pressure. The remaining oil was dissolved
in ethyl acetate and washed with water (33), brine, and then dried with
MgSO4. The ethyl acetate was removed under reduced pressure and product
dried under vacuum yielding a yellow oil (89–92% isolated yield).
1 D. C. Sherrington, J. Polym. Sci., Polym. Chem., 2001, 39, 2364.
2 C. A. Parrish and S. L. Buchwald, J. Org. Chem., 2001, 66, 3820.
3 H. Danjo, D. Tanaka, T. Hayashi and Y. Uozumi, Tetrahedron, 1999,
55, 14341.
4 I. Fenger and C. Le Drian, Tetrahedron Lett., 1998, 39, 4287.
5 S.-B. Jang, Tetrahedron Lett., 1997, 38, 1793.
6 D. E. Bergbreiter, G. Tao and A. M. Kippenberger, Org. Lett., 2000, 2,
2853.
7 J. R. Rasmussen, E. R. Stedronsky and G. M. Whitesides, J. Am. Chem.
Soc., 1977, 99, 4736; J. R. Rasmussen, D. E. Bergbreiter and G. M.
Whitesides, J. Am. Chem. Soc., 1977, 99, 4746.
8 D. E. Bergbreiter, J. G. Franchina and K. Kabza, Macromolecules, 1999,
32, 4993.
Scheme 2
Leaching is an important issue in any catalyst immobilization
scheme. Digestion of a successful reaction mixture and DCP
analysis showed that < 0.1% Pd had leached, based on loadings
determined by combustion of the powder and DCP analysis of
the digested residue. This catalyst was also screened as a
possible catalyst for Heck coupling reactions between iodoar-
enes and activated olefins but was unsuccessful in our hands.
Further studies will include the use of the hyperbranched grafts
on polyethylene powder as supports for other ligands.
9 D. E. Bergbreiter and D. A. Weatherford, J. Org. Chem., 1989, 54,
2726.
10 D. E. Bergbreiter and M. Liu, J. Polym. Sci., Polym. Chem., 2001, 39,
4119.
The observation of oxidation sensitivity of these catalysts is
not remarkable. It is a general problem faced in recycling or
using these and other Pd(0) phosphine ligated catalysts.
CHEM. COMMUN., 2002, 2158–2159
2159
Bergbreiter, David E.
