Visual observation of redistribution and dissolution of palladium during the Suzuki-Miyaura reaction

Article abstract of DOI:10.1002/anie.200800153

(Chemical Presented) Now you see it, now you don't: A specially designed reactor that heats only a small area of Pd foil during a Suzuki-Miyaura coupling permits observation of the surface changes during the reaction. Dissolution of Pd occurs only in the heated zone, and only in the presence of aryl iodide, whereas deposition of Pd occurs preferentially on the unheated zones adjacent to the reactive zone. SEM and XPS are employed to probe the surface before and after reaction.

Full text of DOI:10.1002/anie.200800153

Angewandte
Chemie
DOI: 10.1002/anie.200800153
Palladium Catalysis
Visual Observation of Redistribution and Dissolution of Palladium
during the Suzuki–Miyaura Reaction**
Stephanie MacQuarrie, J. Hugh Horton, Jack Barnes, Kevin McEleney, Hans-Peter Loock, and
Cathleen M. Crudden*
The use of Pd catalysts in cross-coupling reactions has
revolutionized the way organic molecules are made.
Miyaura reaction itself and treatment with the aryl iodide
alone cause changes in surface chemistry and morphology
only where the temperature is sufficient to cause the coupling
reaction to take place. In addition, we found that re-
deposition of Pd occurs preferentially on the periphery of
the reactive zone in cases when only a small portion of the
surface is heated.
[1,2]
Suzuki–Miyaura, Mizoroki–Heck, Kumada–Corriu, Stille,
and Negishi couplings are but a few of the named reactions
that provide versatile methods for the construction of organic
frameworks. Although advances have been made in the use of
alternative metals such as nickel or iron in cross-coupling
reactions, palladium remains by far the catalyst of choice for
these important reactions.
[
7–11]
As the catalyst, we employed 250 mm thickPd foil.
Its
surface was examined by scanning electron microscopy
(SEM), optical microscopy, and X-ray photoelectron spec-
troscopy (XPS) prior to reaction. The foil is characterized by a
relatively smooth, pit-free surface that has ridges spaced at
non-periodic distances presumably as a result of the rolling
process used to generate the foil. Rather than anneal the foil,
we employed these surface features as a frame of reference
during the various treatments.
With efficient ways of manufacturing molecules in hand,
the main difficulty then becomes removing the catalyst at the
[
3,4]
completion of the reaction.
The strict controls regulatory
agencies place on the levels of heavy metals in pharmaceutical
and related products makes removal of metal catalysts after a
reaction an even more serious issue in the pharmaceutical
industry.
An obvious approach to solving this problem is to use
heterogeneous catalysts which should be removable by
filtration. However, the groundbreaking studies of Arai and
Kohler showed independently that even traditional hetero-
geneous catalysts such as Pd/C act by releasing small amounts
of soluble Pd, which then redeposit at the completion of the
For the Suzuki–Miyaura reaction, we employed the
pinacol ester of phenyl boronic acid, and p-nitrophenyl
iodide. An electron-deficient aryl iodide was chosen to
facilitate the coupling reaction. The reaction was first carried
out with the entire piece of Pd foil immersed in DMFat 1008C
in a conventional reactor, without stirring, to prevent damage
to the Pd surface.
[
5,6]
reaction.
Since this re-deposition removes Pd from solu-
tion, understanding how this takes place is critical to
developing effective catalysts that are removed after reaction.
With this in mind, we embarked on a study of the Suzuki–
Miyaura reaction using Pd foil as the catalyst, to permit the
direct visualization of the changes that take place on the
metal surface as a result of the Suzuki–Miyaura reaction,
including the re-deposition phenomenon. In addition to bulk
studies with Pd foil, we have employed a specially designed
reactor that allows us to heat only a small portion of the
surface to a temperature where reaction can take place, while
exposing the entire surface of the foil to the reaction mixture.
Using this technique, we demonstrate that both the Suzuki–
Figure 1A shows the Pd foil after treatment with DMF/
[
12]
water at 1008C, which results in no visible change in the
surface morphology. However under the reaction conditions,
which gave the desired product in 45% yield, considerable
pitting of the surface of the Pd was observed (Figure 1B).
Exposure to the aryl iodide alone also led to considerable
[*] Dr. S. MacQuarrie, Prof. J. H. Horton, Dr. J. Barnes, K. McEleney,
Prof. H.-P. Loock, Prof. C. M. Crudden
Department of Chemistry, Queen’s University
90 Bader Lane, Kingston, ON, K7L 3N6 (Canada)
Fax: (+1)613-533-6669
E-mail: cruddenc@chem.queensu.ca
Homepage:
http://www.chem.queensu.ca/people/faculty/Crudden/crudden.html
[
**] The Natural Sciences and Engineering Research Council of Canada
is acknowledged for support of this research in terms of operating
grants to C.M.C., H.P.L., and J.H.H., and for scholarships to K.M.
Queen’s University is acknowledged for support to K.M.
Figure 1. A) Pd surface after exposure to DMF/H O at 1008C for 24 h;
2
B) Pd surface upon completion of successful Suzuki–Miyaura reaction;
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
C) Pd surface after exposure to the aryl iodide in DMF/H O at 1008C
for 24 h. Scale bar 20 mm.
2
Angew. Chem. Int. Ed. 2008, 47, 3279 –3282
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3279

Communications
restructuring of the Pd surface (Figure 1C), consistent with
Although iodine is present, assignment of the unknown
[25]
the commonly held view that it is the oxidative addition of the
aryl iodide that is responsible for removing Pd from the
surface in much the same way as magnesium metal is
signal based on the I3d5/2 spectra alone is difficult. The N 1s
region of sample C shows one signal attributable to an amine-
like environment (likely trace quantities of Hünigꢀs base) and
one in the correct range for a nitro group. XPS data indicate
[
13–15]
dissolved upon formation of a Grignard reagent.
Con-
sistent with this, we observed about 100 ppb of Pd in solution
by ICPMS after the coupling reaction and after treatment
with aryl iodide.
Analysis of the surfaces of the different samples using
XPS was carried out to determine what if any chemical
changes accompanied the obvious morphological changes
observed in the Pd foil, and to rule out simple contamination
of the surface with reactants or products. In Figure 2, the XPS
that the NO /I area ratio is (1.6 Æ 0.4):1, which is close to the
2
1:1 value that would be expected for the oxidative addition
product. Although contamination of the surface with a trace
of unreacted aryl iodide cannot be ruled out, the observation
of these signals in addition to a new signal in the Pd spectra is
suggestive of a species resulting from oxidative addition into
the C(aryl)ÀI bond.
This type of species has been postulated independently by
[
13]
[24,26]
Reetz and Westermann and Trzeciaket al.
in studies of
Pd nanoparticles in solution and on support. Although a
separate signal for the proposed oxidatively added complex
2
À
(
[Pd(Ph) Br ] ) was not observed definitively in the Pd3d
x
4
À
x
5
/
2
spectra, Trzeciaket al. suggested that the increased breadth of
the signal observed at 336.9 eV may be attributed to this
oxidative addition product. The signal we observe at 335.7 eV
II
would also be consistent with Pd , since a lower binding
energy would arise from stabilization of the core hole by
extra-atomic relaxation of the polarizable iodo and aryl
fragments.
Interestingly, treatment with the aryl iodide alone (sam-
ple D) showed no changes in the Pd3d5/2 or N1s spectra.
However, the I3d5/2,3/2 regions of the XP spectrum do show
weaksignals similar to those observed in sample B. The small
size of these signals may be attributed to removal by washing
with DMF. The low inherent intensity of N1s spectra and the
Figure 2. X-ray photoelectron spectra: A) Pd foil washed with DMF;
B) Pd foil after the Suzuki–Miyaura coupling reaction; C) sample B
after prolonged washing with DMF; D) Pd foil exposed to a solution of
p-iodonitrobenzene in DMF, E) PdI2.
0
significant width of the Pd peakma ke the observation of
small quantities of this species difficult.
spectra (Pd3d , N1s, and I3d ) obtained for variously
3/2,5/2
reacted Pd foils are collected in addition to that for PdI2.
Confident that the observed changes in surface structure
are not due to adsorption of organic impurities but rather
correspond to changes in the surface of the metal, we carried
out the same reaction in a cell that permitted us to heat only a
small portion of the foil to a temperature at which coupling
will take place, while exposing the entire surface to the
reaction conditions.
Pd foil was employed as the base of the reactor, onto
which a 10 mm diameter Teflon cylinder was attached. Heat
was applied to the backof the foil using a metal tip about
1 mm in diameter. A heat sinkwas attached to the foil to
establish a well-defined temperature profile.
5
/2
[16]
Sample A, which is Pd foil that has been exposed to DMF,
displays a signal in the Pd3d region at a binding energy of
5
/2
[
17]
3
35.0 eV. Based on literature data, this peakis reasonably
[
18–20]
assigned to bulkPd metal.
As expected, no signal was
[21]
observed in the I3d5/2,3/2 spectra. Following initiation of the
Suzuki coupling reaction on the Pd foil surface, substantial
changes in the Pd3d5/2 region of the spectrum can be seen (see
sample B). First, the intensity of the spectra is greater than in
the case of the starting Pd foil. This is likely due to the
increase in surface area that accompanies the observed pitting
of the surface (Figure 1B).
The actual temperature profile was examined from the
[
27]
Three peaks appear in the Pd3d5 region at binding
top with an infrared thermal imaging camera. With the heat
sinkset at 37 8C, we were able to determine that the area of
the foil that reached temperatures greater than 608C was no
more than 3 mm in diameter, and temperatures greater than
1008C are limited to 1 mm diameter (Figure 3). The bulk
temperature in the reaction solvent never increased above
ambient as expected from the much higher thermal con-
ductivity of Pd metal compared to the aqueous solution.
We then proceeded to carry out the Suzuki–Miyaura
reaction under the same conditions employed in the “bulk”
reactor, using the microreactor described above. Although
the reaction was considerably slower (ca. 15% yield after
4 days), coupled product was observed in solution. The
decrease in the yield is likely due to the smaller surface
/2
[
22]
energies of 334.8, 335.7, and 336.8 eV.
attributed to bulkmetallic Pd, and the signal at 336.8 eV is
The first peakis
[23]
consistent with either Pd oxide or PdI2.
After more
vigorous washing, the peakat 336.8 eV remains, but no
iodine is present (sample C, left). Thus, this peakis li ke ly
[
19,24]
attributable to Pd oxide.
The peakat 335.7 eV, which is removed by extensive
washing, appears to be unique to sample B, and contains Pd,
N, and I. The spectrum of PdI (sample E) has a signal at
2
3
37.7, 2 eV higher than the unknown signal. Another
possibility is that this signal is attributable to the oxidative
addition product from the reaction between Pd and the aryl
iodide.
3
280 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 3279 –3282

Angewandte
Chemie
Figure 4. SEM images of the surface of the Pd foil after treatment with
spot heater. A) Pd surface outside the heated zone but still exposed to
the reactants; B) the transition between heated and not-heated areas
of the Pd surface and resultant pitting of the surface; C) Pd surface
inside the heated zone where reaction takes place. Scale bar 20 mm.
the heated zone are characterized by significant amounts of
what appears to be re-deposited Pd (Figure 5). This is
consistent with the re-deposition of Pd occurring in the cool
areas of the Pd surface, not far from the heated zone.
Figure 5. SEM image of the Pd foil in spot-heated experiment showing
a section adjacent to the heated area where significant Pd re-
deposition has taken place. Scale bar 5 mm.
Figure 3. Image of the Pd foil (250 mm) recorded with an IR camera
FLIR ThermaCAM SC1000). The tip of a soldering iron heated to
008C was held approximately at the center of the foil underneath the
(
4
reactor. To record the heat distribution, the Pd foil was coated with
soot to decrease reflectance, while mounted on a heat sink held at
This leads to the conclusions that 1) the mobility of the Pd
is not sufficiently large under these conditions to cover the
entire surface and that 2) deposition appears to take place
preferentially on the cooler parts of the surface. This latter
point is important in terms of the results obtained by Arai and
3
78C.
2
area that is heated (0.8 mm ), and therefore the smaller
volume of reaction.
[
5,6]
Local heating of the Pd foil in the absence of reactants led
to no change in the surface morphology; however, in the
presence of the reactants, we observed pitting of the Pd foil
only in the area roughly defined by the high-temperature
probe. Figure 4A shows the area of the foil that is outside the
heated zone where reaction takes place. It is indistinguishable
from the foil treated with only DMF. However, the area
contained within the heated zone (Figure 4C) is dramatically
different. The surface exposed to the reactants and heated to
an appropriate temperature shows considerable pitting.
Figure 4B shows a region that is at the boundary of the
heated and non-heated zones. As previously observed, the
aryl iodide also led to significant changes in the surface
morphology but other reagents did not.
Kohler,
who showed that deposition of Pd occurs upon
cooling of the surface of supported catalysts. It is for this
reason that during the commonly employed “hot-filtration-
test” for heterogeneity, reactions must be filtered while still
hot to prevent deposition of solubilized Pd. To the best of our
knowledge, this is the first time the re-deposition phenom-
enon has been visually observed.
In conclusion, the redistribution of Pd during the Suzuki–
Miyaura reaction was observed visually by SEM examination
of the surface of Pd foil used as a catalyst for this important
reaction. By designing a reactor with which it is possible to
heat only a small area of the surface to reactive temperatures,
we have been able to demonstrate that pitting of the surface
takes place only in reactive zones. Furthermore, re-deposition
of Pd appears to take place at the cool edges of these zones
rather than uniformly across the surface. XPS examination of
the surfaces of these Pd samples demonstrates that the change
in surface morphology is not a result of deposition of by-
One significant difference between this experiment and
those conducted by heating the entire surface of the Pd foil is
the inhomogeneity of the Pd re-deposition. In the case of the
point-heated surface, some of the areas immediately beside
Angew. Chem. Int. Ed. 2008, 47, 3279 –3282
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3281

Communications
products and furthermore that new Pd species are formed as a
[7] To the best of our knowledge, Pd foil has not been employed in
coupling reactions, although it has been used in oxidation and
reduction chemistry (see references [8–11]).
result of the reaction. Treatment with aryl iodide also results
in pitting of the surface. These results show that at least with
Pd foil, reaction likely occurs by dissolution of Pd and re-
deposition, both of which are thermally controlled.
[
8] D. Teschner, A. Pestryakov, E. Kleimenov, M. Havecker, H.
Bluhm, H. Sauer, A. Knop-Gericke, R. Schlögl, J. Catal. 2005,
2
30, 195.
9] C. Descorme, P. W. Jacobs, G. A. Somorjai, J. Catal. 1998, 178,
68.
[
6
Experimental Section
[10] R. S. Monteiro, D. Zemlyanov, J. M. Storey, F. H. Ribeiro, J.
Catal. 2001, 199, 291.
General procedure for the Suzuki–Miyaura coupling reaction: p-
Nitrophenyl iodide (62 mg, 0.25 mmol) (Aldrich), N,N-diisopropyle-
thylamine (64 mg, 0.50 mmol), PhBpin (pinacol ester of phenyl
boronic acid, 75 mg, 0.375 mmol) were weighed into the reaction
vesicle and dry DMF (2.5 mL), degassed by bubbling argon, and
[
[
[
11] E. H. Voogt, A. J. M. Mens, O. L. J. Gijzeman, J. W. Geus, Surf.
Sci. 1997, 373, 210.
12] The morphology of the surface is identical to that which has not
been treated with any solvents or reagents.
13] M. T. Reetz, E. Westermann, Angew. Chem. 2000, 112, 170;
Angew. Chem. Int. Ed. 2000, 39, 165; .
distilled H O (0.125 mL) were added by syringe. The solution was
2
heated with the palladium foil (Aldrich) without stirring and the
reaction progress was monitored by GC-FID.
Bulkpalladium foil samples A, B, and D were extracted from the
reaction vessel and washed with DMF at 808C overnight. After this,
the samples were air-dried at room temperature for 24 h. Sample C
[
14] J. G. de Vries, Dalton Trans. 2006, 421.
[
15] No change in the surface was observed if the Pd was exposed to
the boronic ester component alone (see Figure S-1 in the
Supporting Information).
[
16] XPS data were also obtained in the B1s region, but in no case
was any signal detected. C1s and O1s spectra were also obtained
but were mostly unremarkable; for the C1s region, they showed
a single peakat a binding energy of 285.0 eV. The O1s spectra
consisted of a single peakat binding energy of 532.6 eV in all
cases. Both are likely due to trace contaminants on the surface as
they are present in all samples.
(Figure 2) was prepared from sample B after it was re-washed with
DMF at 808C for 8 h, refluxed in THF for 8 h, and then rinsed with
EtOAc and CH Cl .
2
2
Samples of palladium foil (shown in Figure 4 and 5) that were
exposed to the tip of a soldering iron were washed with DMF at 808C
for 4 h to remove any organic material, prior to examination of the
surface.
[
[
[
17] Note that the second signal observed at about 340 eV is the
Pd3d3/2 signal.
18] M. L. Kantam, M. Roy, S. Roy, B. Sreedhar, S. S. Madhavendra,
B. M. Choudary, R. L. De, Tetrahedron 2007, 63, 8002.
19] E. H. Voogt, A. J. M. Mens, O. L. J. Gijzeman, J. W. Gues, Surf.
Sci. 1996, 350, 21.
XPS measurements were performed by using a Thermo Instru-
ments Microlab 310F surface analysis system (Hastings, UK) under
ultrahigh vacuum conditions with a MgKa radiation source at
1
253.6 eV. Scans were acquired at fixed analyzer transmission
(FAT) mode at a pass energy of 20 eV. All spectra were calibrated
to the C1s line at a binding energy of 284.5 eV. Spectra were
background-subtracted using a Shirley fit algorithm and using a
Powell peak-fitting algorithm provided in the spectrometer soft-
[
[
20] P. Weightman, P. T. Andrews, J. Phys. Chem. C 1980, 13, L815.
21] The as-received foil before extensive washing with DMF showed
strong signals in the C1s and O1s spectra, indicative of the
presence of a considerable amount of oxide and carbonaceous
contaminants on the surface, in addition to a peakassignable to
Pd oxide. Except for some small residual C1s and O1s signals,
these all disappeared after the samples had been washed with
DMF.
[
28]
ware. Peakareas from different samples were normalized by using
the ratio of the background signal 5 eV beyond the Fermi level cut off
from the valence band spectra to account for small variations in X-ray
source intensity between analyses. XP spectra of PdI2 (Aldrich
7
790387) were obtained by spreading the powdered Pd salt onto a
piece of double-sided Cu tape (SPI Supplies Toronto, Ontario), which
was then mounted onto the sample holder and transferred through an
air lockinto the spectrometer.
[
22] It should be noted that poorly resolved XP spectra can be easily
“overfit” to more peaks than the data warrant. However, if
spectrum B is fit with only two peaks, the best fit is obtained with
peaks of full width half maximum (FWHM) of 1.2 and 2.1 eV.
This does not seem physically reasonable, as the FWHM for all
other Pd spectra studied here was (1.5 Æ 0.2) eV. By fitting three
peaks, the FWHM values become consistent with those for other
spectra observed here and also by Takasu et al., who report
values of 1.6 eV: Y. Takasu, R. Unwin, B. Tesche, A. M.
Bradshaw, Surf. Sci. 1978, 77, 219.
SEM images were recorded by mounting the palladium foil
samples on pins, no conductive coating was required for these
samples. A JEOL JSM-840 scanning electron microscope was used to
investigate the surfaces.
Received: January 11, 2008
Keywords: cross-coupling · heterogeneous catalysis · leaching ·
.
[23] L. Hilaire, P. Lꢁgarꢁ, Y. Holle, G. Maire, Solid State Commun.
979, 32, 157.
palladium · surface chemistry
1
[
24] A. Gniewek, A. Trzeciak, J. J. Ziolkowski, L. Kepinski, J.
Wrzyszcz, W. Tylus, J. Catal. 2005, 229, 332.
[
25] The range of binding energies observed for iodide compounds is
very small, for example, the I3d5/2 binding energies of CuI, NiI2,
AgI, ZnI , and CdI spans a range of only 0.8 eV: S. W.
[
1] Handbook of Organopalladium Chemistry for Organic Synthesis
Eds.: E.-I. Negishi, A. de Meijere), Wiley-VCH, Weinheim,
002.
2
2
(
2
Gaarenstroom, N. Winograd, J. Chem. Phys. 1977, 67, 3500.
26] I. Pryjomska-Ray, A. Gniewek, A. M. Trzeciak, J. J. Ziolkowshi,
W. Tylus, Top. Catal. 2006, 40, 173.
27] To decrease reflection, the Pd surface needed to be coated with
carbon black.
[
[
[
[
2] Metal-Catalyzed Cross-Coupling Reactions (Eds.: F. Diederich,
P. J. Stang), Wiley-VCH, Weinheim, 1998.
[
3] K. Königsberger, G.-P. Chen, R. R. Wu, M. J. Girgis, K. Prasad,
O. Repic, T. J. Blacklock, Org. Process Res. Dev. 2003, 7, 733.
4] C. E. Garrett, K. Prasad, Adv. Synth. Catal. 2004, 346, 889.
5] K. Kohler, W. Kleist, S. S. Prockl, Inorg. Chem. 2007, 46, 1876.
6] F. Y. Zhao, M. Shirai, Y. Ikushima, M. Arai, J. Mol. Catal. A
28] Practical Surface Analysis by Auger and X-ray Photoelectron
Spectroscopy (Hrsg.: D. Briggs, M. P. Seah), Wiley, New York,
[
[
[
1983.
2002, 180, 211.
3
282
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 3279 –3282