Dendrimer-Encapsulated Nanoparticle Precursors to Supported Platinum Catalysts

Article abstract of DOI:10.1021/ja0364120

In this contribution, we report the successful preparation of supported metal catalysts using dendrimer-encapsulated Pt nanoparticles as metal precursors. Polyamidoamine (PAMAM) dendrimers were first used to template and stabilize Pt nanoparticles prepare

Full text of DOI:10.1021/ja0364120

Published on Web 11/07/2003
Dendrimer-Encapsulated Nanoparticle Precursors to
Supported Platinum Catalysts
Huifang Lang, R. Alan May, Brianna L. Iversen, and Bert D. Chandler*
Contribution from the Department of Chemistry, Trinity UniVersity, One Trinity Place,
San Antonio, Texas 78212-7200
Received May 29, 2003; E-mail: bert.chandler@trinity.edu
Abstract: In this contribution, we report the successful preparation of supported metal catalysts using
dendrimer-encapsulated Pt nanoparticles as metal precursors. Polyamidoamine (PAMAM) dendrimers were
first used to template and stabilize Pt nanoparticles prepared in solution. These dendrimer-encapsulated
nanoparticles were then deposited onto a commercial high surface area silica support and thermally activated
to remove the organic dendrimer. The resulting materials are active oxidation and hydrogenation catalysts.
The effects of catalyst preparation and activation on activity for toluene hydrogenation and CO oxidation
catalysis are discussed.
Introduction
plated iron oxide nanoparticles have been used as precursors
for the synthesis of carbon nanotubes.18 In this manuscript, we
Polyamidoamine (PAMAM) dendrimers have drawn consid-
erable interest in recent years due to their potential applications
in medicine, nanotechnology, and catalysis.1-5 These macro-
molecules can also be functionalized to incorporate transition
metal complexes into the dendrimer backbone; such function-
alized dendrimers have shown a variety of novel properties when
employed as homogeneous catalysts.4,5 Functionalized or modi-
fied PAMAM dendrimers can also be immobilized on solid sup-
ports and employed as “heterogenized” homogeneous catalysts.6-8
The ability to control dendrimer interior/exterior functional-
ities and the macromolecular architecture of PAMAM den-
drimers (open spaces within the interior) also create an ideal
environment for trapping guest species.9,10 Specifically, PAM-
AM dendrimers can bind a defined number of transition metal
cations and thus template and stabilize metal oxide or metal
nanoparticles.3,11,12 Templated Pt, Pd, and Pt-Pd nanoparticles
have been employed as homogeneous catalysts,13-17 and tem-
report a new application of dendrimer-encapsulated nanopar-
ticles: as precursors to supported metal particle catalysts
(Scheme 1).
Finely dispersed metal nanoparticles supported on inorganic
oxide carriers are a mainstay of commercial heterogeneous
catalysts.19 These supported-metal catalysts are generally pre-
pared via impregnation of metal salts onto an oxide support,
followed by high-temperature oxidation and/or reduction.19
Because such methods are relatively inexpensive, they are
widely applied with numerous metals and supports; however,
they provide limited control over particle size and distribution
in the ultimate catalyst. This is a considerable drawback to
studying and understanding the catalytic mechanisms at work
on these catalysts.
Dendrimer-encapsulated metal nanoparticles (DENs), on the
other hand, can be prepared in solution with reproducible and
variable synthetic schemes. Utilizing DENs as catalyst precur-
sors offers the opportunity to exert a degree of control over
metal particle size and composition, while varying the carrier
or substrate (see Scheme 1). The potential to ultimately control
nanoparticle size and composition11,17,20 makes DENs extremely
attractive as potential precursors for studying supported-metal
catalysts. Because the particles are prepared ex situ and can be
deposited onto a substrate or support, DENs offer the op-
portunity to bridge the gap between surface science studies of
model systems and real world catalysts on high surface area
(1) Zeng, F.; Zimmerman, S. C. Chem. ReV. 1997, 97, 1681.
(2) Bosman, A. W.; Janssen, H. M.; Meijer, E. W. Chem. ReV. 1999, 99, 1665.
(3) Crooks, R. M.; Lemon, B. I.; Sun, L.; Yeung, L. K.; Zhao, M. Top. Curr.
Chem 2001, 212, 82-135.
(4) Twyman, L. J.; King, A. S. H.; Martin, I. K. Chem. Soc. ReV. 2002, 31,
69-82.
(5) Kreiter, R.; Kleij, A. W.; Gebbink, R. J. M.; van Koten, G. Top. Curr.
Chem 2001, 217, 163-199.
(6) Chung, Y.-M.; Rhee, H.-K. Chem. Commun. 2002, 238-239.
(7) Antebi, S.; Arya, P.; Manzer, L. E.; Alper, H. J. Org. Chem. 2002, 67,
6623-6631.
(8) Alper, H.; Arya, P.; Borque, S. C.; Jefferson, G. R.; Manzer, L. E. Can. J.
Chem. 2000, 78, 920-924.
(9) Fisher, M.; Vogtle, F. Angew. Chem., Int. Ed. Engl. 1999, 38, 884.
(10) Cooper, A. I.; Londono, J. D.; Wignall, G.; McClain, J. B.; Samulski, E.
T.; Lin, J. S.; Dobrynin, A.; Rubinstein, M.; Burke, A. L. C.; Frechet, J.
M. J.; DeSimone, J. M. Nature 1997, 389, 368-371.
(11) Crooks, R. M.; Zhao, M.; Sun, L.; Chechik, V.; Yeung, L. K. Acc. Chem.
Res. 2001, 34, 181-190.
(15) Ooe, M.; Murata, M.; Mizugaki, T.; Ebitani, K.; Kaneda, K. Nano Lett.
2002, 2, 999-1002.
(16) Niu, Y. H.; Yeung, L. K.; Crooks, R. M. J. Am. Chem. Soc 2001, 123,
6840-6846.
(17) Chung, Y.-M.; Rhee, H.-K. Catal. Lett. 2003, 85, 159-164.
(18) Choi, H. C.; Kim, W.; Wang, D.; Dai, H. J. Phys. Chem. B 2002, 106,
12361-12365.
(19) Ponec, V.; Bond, G. C. Catalysis by Metals and Alloys; Elsevier:
Amsterdam, 1995; Vol. 95.
(20) Scott, R. W. J.; Datye, A. K.; Crooks, R. M. J. Am. Chem. Soc. 2003, 123,
3708-3709.
(12) Ottaviani, M. F.; Montalti, F.; Turro, N. J.; Tomalia, D. A. J. Phys. Chem.
B 1997, 101, 158-166.
(13) Yeung, L. K.; Lee, C. T.; Johnston, K. P.; Crooks, R. M. Chem. Commun.
2001, 21, 2290-2291.
(14) Rahim, E. H.; Kamounah, F. S.; Frederiksen, J.; Christensen, J. B. Nano
Lett. 2001, 1, 499-501.
9
14832
J. AM. CHEM. SOC. 2003, 125, 14832-14836
10.1021/ja0364120 CCC: $25.00 © 2003 American Chemical Society

Dendrimer Route to Supported Pt Catalysts
A R T I C L E S
Scheme 1. Schematic Route to Dendrimer-Derived Supported
Nanoparticle Catalysts
supports. The possibility of varying particle size and composition
through nanoparticle preparation schemes makes DENs uniquely
suited to exploring and understanding the relative importance
of these effects on catalytic reactions using materials comparable
to those employed as industrial catalysts. In this manuscript,
we report the initial proof of this concept using dendrimer-
encapsulated Pt nanoparticles.
Figure 1. Infrared spectra showing dendrimer decomposition under flowing
O2/He.
Results
Generation 5 PAMAM dendrimers (G5-OH) were used to
prepare dendrimer-encapsulated Pt nanoparticles (DENs) via
literature procedures.11 Two samples were prepared in aqueous
solution using Pt:dendrimer ratios of 100:1 (Pt100) and 50:1
(Pt50). After purification by dialysis, the Pt content of the DEN
solutions was determined via AA spectroscopy. DENs were then
deposited onto DAVICAT SI- 1403 silica via wetness impreg-
nation to make 0.3 wt % Pt materials. Because the DEN
preparation procedures utilized very dilute G5-OH solutions,
impregnation of DEN solutions onto the silica required condens-
ing the solutions prior to deposition. For comparison purposes,
a standard supported Pt catalyst was prepared via insipient
wetness impregnation of H2PtCl6 onto the same silica support.
The supported, intact DENs do not bind CO and are not active
catalysts. Presumably, in the absence of solvent, the dendrimer
collapses onto the nanoparticle, preventing even small substrates
from accessing the metal surface.21 Activation conditions for
the supported DENs were chosen on the basis of thermal
decomposition experiments in a controlled atmosphere transmis-
sion IR cell. Destruction of the dendrimer amide linkages was
followed while heating under flowing O2/He by monitoring
spectral changes in the 1750-1500 cm-1 region. Figure 1 shows
the onset of dendrimer decomposition began near 75 °C;
however, the spectrum continued to change until the sample
had been treated at 300 °C for several hours.
Figure 2. CO oxidation catalysis after oxidation at 300 °C for 4 h.
Table 1. Characterization Data
Pt/SiO
Pt100/SiO
Pt50/SiO
2
2
2
Elemental Analysis
0.25
0
% Pt
% C
% N
0.30
<0.5
<0.2
0.32
<0.5
<0.2
CO Adsorption Data
ν
CO (cm-1
peak width at
)
2074
57
2082
23
2078
30
half-height (cm-1
dispersion (%)
)
43
66
49
dcalc (nm)
2.5
1.7
2.3
TEM
supported DENsa
activated catals
1.9 ( 0.3
2.2 ( 0.5
2.1 ( 0.6
2.2 ( 0.5
2.9 ( 1.5
a Average diameter with standard deviation.
On the basis of these results, samples were oxidized with
flowing O2/He at 300 °C for 4 h. As shown in Figure 2, Pt50/
SiO2 and Pt100/SiO2 subjected to this oxidation treatment showed
very similar activity for CO oxidation catalysis but were less
active than the traditionally prepared catalyst. Further treatment
with flowing H2 yielded materials that bind CO and are much
more active catalysts (vida infra). On the basis of these results,
we used an activation protocol involving an oxidation step (O2,
300 °C, 4 h) and a reduction step (H2, 300 °C, 2 h) to yield the
dendrimer-derived nanoparticle catalysts (DDNCs) Pt50/SiO2 and
Pt100/SiO2. Elemental analysis of these catalysts (Table 1) shows
the C and N content to be below the detection limits. The
traditionally prepared Pt/SiO2 catalyst was also activated under
these conditions and contained less than 500 ppm of Cl after
activation.
Infrared spectra of CO adsorbed to the activated catalysts are
presented in Figure 3; the peak shapes are evaluated by the peak
widths at half-height in Table 1. Particle size distribution
histograms from TEM data on supported DENs and activated
DDNCs are compared with the traditionally prepared catalyst
in Figure 4. Further characterization data, including elemental
(21) Deutsch, S. D.; Lafaye, G.; Lang, H.; Chandler, B. D.; Liu, D.; Williams,
C. T.; Gao, J.; Murphy, C. J.; Amiridis, M. D. Langmuir 2003, submitted.
9
J. AM. CHEM. SOC. VOL. 125, NO. 48, 2003 14833

A R T I C L E S
Lang et al.
Figure 3. Infrared spectra of CO adsorbed on (a) Pt/SiO2, (b) Pt100/SiO2, and (c) Pt50/SiO2.
Figure 4. Particle size distribution histograms for (a) intact Pt100/SiO2, (b) activated Pt100/SiO2, (c) intact Pt50/SiO2, (d) activated Pt50/SiO2, and (e) Pt50/
SiO2.
analysis, CO adsorption, and average particle sizes measured
from TEM, are compiled in Table 1. CO chemisorption and
TEM data for the DDNCs indicate that the metal particles are
small and well dispersed, and the infrared data are consistent
with CO adsorbed on supported-Pt catalysts.
The activated catalysts were tested using the toluene hydro-
genation and CO oxidation test reactions. Figure 5 presents CO
oxidation light-off curves (plots of conversion vs temperature
at constant space velocity), and Table 2 contains data extracted
from the catalysis experiments. Toluene hydrogenation Arrhe-
nius plots yielded nearly identical apparent activation energies
(Eapp) for all three catalysts. Rates at 60 °C (Table 2) show some
differences between catalysts in total activity. However, toluene
hydrogenation turnover frequencies (TOFs), which correct the
overall rate for the available surface Pt, were essentially the
same for all three catalysts.
CO oxidation light off curves (Figure 5a) for the activated
catalysts show that the DDNCs are more active than the tra-
ditionally prepared Pt/SiO2 catalyst. The TOF vs temperature
plots (Figure 5b) are derived from the conversion data, cor-
recting for the slight differences in platinum loading and dis-
persion. After these corrections, the catalytic activity of Pt100
/
SiO2 is indistinguishable from Pt/SiO2. The CO oxidation tests
also show Pt50/SiO2 to be distinct from the other two catalysts.
The onset of activity by this catalyst is 25-30 °C lower than
with Pt/SiO2, TOFs for Pt50/SiO2 are 2.5-3 times higher than
the other catalysts, and the apparent activation energy extracted
from the TOF data (Table 2) is 3 kcal/mol lower.
9
14834 J. AM. CHEM. SOC. VOL. 125, NO. 48, 2003

Dendrimer Route to Supported Pt Catalysts
A R T I C L E S
ticles. One other attempt to activate supported DENs with high-
temperature oxidation has been reported.22 In this study, Sun
and Crooks deposited dendrimer encapsulated Pd nanoparticles
onto a flat mica support and evaluated particle mobility with
AFM. They clearly showed Pd particle mobility and sintering
are pronounced at very high temperatures (600 °C) but did not
address activation under more moderate conditions. In this study,
we monitored the decomposition of the PAMAM dendrimer
using in-situ infrared spectroscopy to explore the possibility of
activating the supported DENs at lower temperature. Particle
agglomeration processes on oxide supports are expected to be
much faster at higher temperatures, so identifying low-temper-
ature activation processes is critical to eventually realizing the
possible advantages of the dendrimer route to model, study, and
develop industrially relevant supported catalysts.
The decomposition experiments illustrated in Figure 1 indicate
that the amide bonds of the PAMAM dendrimers are largely
destroyed after oxidation at 300 °C. However, CO oxidation
experiments on oxidized supported DENs (Figure 2) show that
the nanoparticles are not very active catalysts. Further treatment
with H2 at 300 °C yielded catalysts active for oxidation and
hydrogenation reactions. Elemental analysis confirms that most
of the C and N have been removed, but the relatively high
detection limits do not exclude the persistence of some organic
residues on the catalyst surface.
Figure 5. CO oxidation catalysis for after oxidation at 300 °C (20% O2/
He, 4 h) and reduction at 300 °C (20% H2/He, 2 h).
Table 2. Catalysis Data for Traditional and DEN Catalysts
Pt/SiO
Pt100/SiO
Pt50/SiO
2
2
2
Toluene Hydrogenation
rate at 60 °C (s-1
TOF (s-1
app (kcal/mol)
)
0.057
0.13
0.085
0.13
0.069
0.14
13.8
On the basis of a 100 atom spherical Pt particle, the expected
)
E
13.7
13.6
dispersion (% surface Pt) and average particle sizes for Pt100
/
CO Oxidation
150
17.5
SiO2 are 78% and 1.4 nm, respectively. The chemisorption data
(66%, dcalc ) 1.7 nm), which inherently samples a much greater
number of particles than is possible with TEM, is in reasonably
good agreement with the expected value and indicates that there
has been little particle agglomeration during the activation of
Pt100/SiO2. The particles imaged with TEM also have a slightly
larger average diameter than the expected value and the
chemisorption value, which is consistent with the literature re-
ports of Pt and Pd DENs3,11 and with the presence of the support
(vida infra). After the activation protocol, the average diameter
and size distribution change only slightly.
The three-dimensional nature of the support complicates
interpretation of the TEM data in two ways: (1) The three-
dimensional landscape of the support makes it difficult to find
areas of relatively consistent thickness necessary to obtain good
contrast between the nanoparticles and the support. (2) Since
the nanoparticles are dispersed throughout this three-dimensional
support, they are often in different focal planes, making many
of the particles in a given micrograph slightly out of focus.
Given these factors, the differences in TEM particle size before
and after activation may not be significant. At most, they rep-
resent only a small increase in average particle size. The high
chemisorption value also suggests that the TEM measurements
may slightly overestimate the actual average particle size, either
through this inherent imprecision in measuring the observed
particles or through undercounting of the smaller particles.
TEM data from Pt50/SiO2 also show little to no change in
average particle size and distrubution resulting from the acti-
vation protocol. However, the particles are larger than would
be expected for a 50 atom particle, suggesting that metal aggre-
gation has occurred during the deposition process. In the
preparation of the catalyst via wetness impregnation, the volume
T
E
25 (°C)a
app (kcal/mol)
141
17.5
125
14.5
a Temperature required for 25% conversion.
Discussion
PAMAM dendrimers have been used to template and stabilize
dendrimer encapsulated nanoparticles (DENs) of Pt, Pd, Au,
Cu, and Ru.3,11 DENs, particularly those of noble metals, offer
a number of opportunities for the study of heterogeneous cat-
alysts. Because DENs can be prepared and purified in aqueous
solution, they can then be delivered to any number of solid
carriers. This allows for the delivery of nanoparticles with
known and reproducible synthetic histories (or even from the
same synthetic batch) to a variety of different substrates. For
example, the deposition of Pt DENs onto high surface area
amorphous silica, mica, and Si/SiO2 TEM substrates would
allow for atomic force microscopy (AFM) and TEM studies to
be performed on the appropriate samples. At the same time,
direct comparisons could be made with bulk catalyst activity
measurements using high surface area catalysts prepared from
the same batch of nanoparticles. Alternately, delivery of nano-
particles of consistent size and distribution to different supports
(e.g. silica, alumina, titania, ceria) may eventually allow for the
evaluation of support effects and metal-support interactions.
Because the nanoparticles of conventionally prepared supported
catalysts are generally prepared on the surface of the support,
particle sizes and distributions are often dissimilar on different
carriers, thus making the separate evaluation of support and
particle size effects difficult.
The results from this study serve as the initial proof of this
concept and outline the important factors for developing het-
erogeneous catalysts from dendrimer encapsulated nanopar-
(22) Sun, L.; Crooks, R. M. Langmuir 2002, 18, 8231-8236.
9
J. AM. CHEM. SOC. VOL. 125, NO. 48, 2003 14835

A R T I C L E S
Lang et al.
of the Pt50 solution had to be concentrated to a greater extent
than the Pt100 solution to maintain consistent Pt loading.
Concentrating this solution prior to deposition would be
expected to foster some degree of particle agglomeration during
deposition.
After activation, the TEM particle size distributions for the
DDNCs are very similar. CO uptake by Pt50/SiO2, however, is
somewhat lower than Pt100/SiO2. This may be due to the
persistence of organic residues on the catalyst surface. If the
metal loadings are to remain consistent, depositing the Pt50
DENs necessitates doubling the dendrimer loading relative to
the Pt100 DENs. It is perhaps not surprising that a single
activation protocol is not optimal for catalysts with such different
dendrimer loadings as the Pt50/SiO2 catalyst has more organic
material to remove. Our continuing work is aimed at addressing
the differences in activation conditions and optimizing condi-
tions for individual catalysts.
Infrared spectra of CO adsorbed to the dendrimer-derived
nanoparticle catalysts (DDNCs) are consistent with CO bound
to supported Pt catalysts.19 The CO stretching frequencies (Table
1) are similar for all three catalysts, with the DDNCs slightly
blue shifted relative to Pt/SiO2. Figure 3 also shows that the
DDNC peak shapes are more narrow and symmetrical than the
Pt/SiO2 catalyst, suggesting that the surface platinum sites on
the DDNCs are more uniform than on the conventionally
prepared catalyst. This result is consistent with the TEM data,
which show more narrow particle size distributions for the
DDNCs.
Toluene hydrogenation and CO oxidation catalysis were
chosen as test reactions that are representative of the broad range
of reactions for which supported Pt catalysts are active. Toluene
hydrogenation is particularly attractive because it is well-known
as a structure insensitive reaction on monometallic Pt and Pd
catalysts.23 By the term “structure insensitive”, we mean that
the reaction rate depends only on the number of exposed metal
atoms; i.e., the turnover frequency for this reaction is indepen-
dent of particle size and does not measurably change on various
surface planes.24 Thus, although rates for toluene hydrogenation
at 60 °C are slightly different for the three catalysts, the turnover
frequencies (TOFs) are the same when correcting for available
surface Pt measured with CO chemisorption (Table 2). Similarly,
the apparent activation energies (Eapp) for toluene hydrogenation
(Table 2) are the same for all the catalysts, consistent with the
reaction’s structure insensitivity.23 This structure insensitivity
makes toluene hydrogenation a useful test reaction because it
directly tests the activity of the available surface platinum,
regardless of particle size or morphology. This reaction also
provides an additional indirect estimate of exposed metal atoms
that is consistent with the CO chemisorption data for all three
catalysts.
(Figure 5a) than the conventional catalyst, which is internally
consistent with the higher dispersion of Pt100/SiO2. Once the
total activity is corrected for the number of available Pt atoms
(Figure 5b), there are no significant differences in CO oxidation
TOFs. This result shows that DDNCs are appropriate models
for traditionally prepared catalysts and indicates that the small
amount of chlorine left on the Pt/SiO2 catalyst does not
substantially affect CO oxidation catalysis.
Pt50/SiO2 is clearly more active for CO oxidation than the
other two catalysts. At all of the temperatures studied, TOFs
for Pt50/SiO2 are 2-3 times greater than TOFs for Pt100/SiO2
and Pt/SiO2. We are currently investigating the origination of
this enhanced activity and are considering several possibilities.
The TEM and CO chemisorption data suggest that small amount
of organic residues may persist on this catalyst. These might
affect CO oxidation catalysis by breaking up islands of CO that
may serve effectively as poisons for some active sites25 or by
affecting the availability of weakly bound O atoms.26 We cannot
rule out that there may be differences in the surface oxidation
state of these catalysts or perhaps the presence of a different
particle morphology.
Conclusions
In this study, we show that dendrimer-encapsulated noble
metal nanoparticles can be used as precursors to supported metal
particle catalysts. Activation conditions were chosen on the basis
of in-situ infrared spectroscopy experiments, which showed that
the decomposition of PAMAM dendrimer amide bonds was
largely completed after treatment at 300 °C for several hours.
Further treatment with hydrogen at 300 °C was also required
to prepare active catalysts, and this activation protocol caused
little sintering of the metal nanoparticles. The resulting catalysts
are active for both oxidation and hydrogenation reactions and
prove to be appropriate models for traditionally prepared
supported catalysts.
Acknowledgment. We are very grateful to Stephen Mal-
donado and Prof. Keith Stevenson at the University of Texas
at Austin for their assistance in obtaining TEM data. This
research was supported by an award from Research Corp.
Acknowledgment is also made to the donors of the Petroleum
Research Fund, administered by the American Chemical Society,
and the National Science Foundation (Grant CTS-030135) for
partial support of this research. R.A.M. gratefully acknowledges
support from the Welch Foundation (Departmental Grant
W-0031). We also thank the National Science Foundation (Grant
CHE-0116731) for support of the chemisorption instrumentation.
Supporting Information Available: A complete experimental
section. This material is available free of charge via the Internet
at http://pubs.acs.org.
CO oxidation catalysis results highlight some of the potential
of the dendrimer route. Pt100/SiO2 shows slightly higher activity
JA0364120
(23) Lin, S. D.; Vannice, M. A. J. Catal. 1993, 143, 554-562.
(24) Bennet, C. O.; Che, M. AdV. Catal. 1989, 36, 55-172.
(25) Fanson, P. T.; Delgass, W. N.; Lauterbach, J. J. Catal. 2001, 204, 35-52.
(26) Bourane, A.; Bianchi, D. J. Catal. 2002, 209, 126-134.
9
14836 J. AM. CHEM. SOC. VOL. 125, NO. 48, 2003