B. H. Lipshutz et al.
Table 3. Influence of copper iodide as co-catalyst.[a]
trasts with related samples of
Pd/C, which, likewise using
TEM analyses, tend to form
much larger, less reactive, and
Conditions A[b]
no CuI
Conditions B[c]
no CuI
non-recyclable
aggregates
Entry
Catalyst
5 mol% CuI
5 mol% CuI
during the reaction.[5l,17] These
observations correlate well with
the X-ray diffraction (XRD)
pattern that indicates Pd0 crys-
tallites of 60 nm average diame-
ter (Figure 2). By contrast, all
commercially available catalysts
1
2
3
UC Pd
Aldrich
Degussa E101
95%
90%
45%
55%
40%
95%
90%
–
26%
traces
–
68%
[a] Isolated yield on 2 mmol scale. [b] 5 mol% Pd/C, 10 mol% PPh3, 95% ethanol, 1208C (microwaves),
15 min. [c] 2 mol% Pd/C, 2 mol% XPhos, 95% ethanol, 508C, 5 h.
Suzuki–Miyaura[17] couplings, as well as in Sonogashira reac-
tions.[7e]
examined by XRD consist of smaller nanoparticles (6–
15 nm on average).
Carbon monoxide temperature programmed desorption
(CO-TPD) was used to quantify the Pd surface sites due to
the specificity of the Pd-CO bonding ratio of 1:1. It is ob-
served (Figure 3) that the relative accessible palladium sur-
face areas (as integrated area under each desorption curve)
are dependent on the source. Additionally, there are small
differences in the effective binding/desorption energies
(peak positions) suggesting slightly different chemical envi-
ronments for Pd in each catalyst. Both samples of Pd/C
either sold by Alfa Aesar, or prepared as UC Pd, show the
greatest surface area of Pd (Table S1, entries 5, 6), even
though the size of the nanoparticles for UC Pd is larger than
that of all other commercial samples (Figure 2).
Table 4. Recycling of catalysts (isolated yields).
Run UC Pd
[%]
Alfa Aesar
[%]
Degussa E101
[%]
Aldrich
[%]
Selcat Q6
[%]
1st
95
92
89
87
94
45
10
–
75
21
–
84
49
15
–
90
54
20
–
2nd
3rd
4th
–
Since the data as to particle size alone cannot be used to
explain both the greater reactivity and unique recyclability
of UC Pd compared to other sources of Pd/C, additional sur-
face analyses were conducted. X-ray photoelectron spectros-
copy (XPS) was employed to determine the oxidation state
of palladium at the surface. Given that the “escape depth”
for photoelectrons generated within any given substance is
quite shallow, XPS measures a signal from only the top 2–
4 nm of a material. This surface specificity is particularly in-
formative, in contrast to the “bulk” signal observed in XRD.
All Pd/C catalysts examined by this technique contain
roughly 30% Pd0 at their surface, with the exception of De-
gussa E101 that shows 15% (Table S2). The amounts of PdII
and PdIV present vary widely between samples, and hence,
no trend with respect to particle size or activity is observed.
However, after use, UC Pd retains its distribution of oxida-
tion states (ca. 30% Pd0; Figure 4), while the Degussa E101
and Alfa Aesar catalysts drop from 15 to 5%, and from 27
to 9% Pd0, respectively. The substantially better stability of
UC Pd may derive from its higher levels of reduced metal at
the surface relative to other catalysts due to its larger metal
nanoparticles. These are less likely to stabilize oxides on
their surface relative to smaller nanoparticles.[18] The impli-
cation is that there exists an optimum particle size which
balances surface area with surface stability. Thus, while to
some extent counterintuitive, using smaller size particles
does not translate into greater reactivity. This balance im-
pacts both catalyst activity and recyclability; only UC Pd
emerges with minimal modification to its surface following
participation in Sonogashira cross-couplings.
Several examples that lead to products 4–17 and attest to
the generality of this mild, heterogeneous cross-coupling are
illustrated in Table 5. Aryl bromides bearing either activat-
ing or deactivating substituents gave the corresponding dis-
ubstituted acetylenes in high isolated yields. Successful
cross-couplings involving an assortment of terminal acety-
lenes containing a variety of functional groups, such as ni-
trile, chloride, acetal, and alcohol groups, lend further cre-
dence to the attractiveness of this protocol. Featured cases
include aryl bromides bearing o-CF3 (entry 2), m-amino (en-
tries 3, 4), p-cyano (entries 5, 6), p-nitro (entry 8), keto (en-
tries 9, 10), and 3-pyridyl groups (entry 11), and more highly
functionalized substrates (entries 12 and 13). Acetylenes
containing an w-cyano (entries 2, 3), a-quaternary carbon
(entries 9, 11, 13), and a cholesteryl residue (entry 14) all re-
acted smoothly. Double arylation of a diyne (entry 4) is also
worth mentioning. Tests were conducted on UC Pd to assess
the extent to which it provides mainly heterogeneous cataly-
sis. Analysis of a reaction mixture (as in Table 4) premature-
ly stopped (at 60% conversion) and filtered hot to remove
Pd/C showed no further consumption of educts upon further
treatment at 508C. ICP-AES analysis on the filtrate further
confirmed that only traces of Pd are detected in solution (2–
4 ppm).
Analyses of Pd/C: Bright field transmission electron micros-
copy (TEM) imaging of UC Pd shows palladium nanoparti-
cles with a fairly broad size distribution, in the 10–300 nm
range, both before and after catalysis (Figure 1). This con-
3368
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 3366 – 3371