Dendrimer-based synthesis of Pt catalysts for hydrocarbon conversion

Article abstract of DOI:10.1016/j.apcata.2010.05.045

Well-defined platinum-based catalysts were prepared using fourth-generation hydroxyl-terminated poly(amidoamine) (PAMAM) dendrimers as templates. Nanoparticles with an average of 40 Pt atoms and an average diameter of 1.5 nm were produced this way, and dispersed on a homemade sol-gel silica support. Different treatments were tested for the activation of the resulting catalyst, including heating in vacuum or in oxygen or hydrogen atmospheres. A combination of electron microscopy, infrared absorption spectroscopy, X-ray photoelectron spectroscopy, and catalytic kinetic measurements were used to investigate the effect of those treatments on the decomposition and elimination of the material associated with the dendrimer and on the activity and selectivity of the resulting catalysts. Neither heating in vacuum nor burning with oxygen was deemed acceptable, because much graphitic material is deposited with the former treatment and metal oxidation is induced by the latter. Hydrogen pretreatments, on the other hand, lead to fairly active catalysts with high selectivity for the conversion of trans olefins to their cis counterparts.

Full text of DOI:10.1016/j.apcata.2010.05.045

Applied Catalysis A: General 391 (2011) 386–393
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Dendrimer-based synthesis of Pt catalysts for hydrocarbon conversion
Manuel A. Albiter, Ricardo Morales, Francisco Zaera∗
Department of Chemistry, University of California, Riverside, CA 92521, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Well-defined platinum-based catalysts were prepared using fourth-generation hydroxyl-terminated
poly(amidoamine) (PAMAM) dendrimers as templates. Nanoparticles with an average of 40 Pt atoms
and an average diameter of 1.5 nm were produced this way, and dispersed on a homemade sol–gel silica
support. Different treatments were tested for the activation of the resulting catalyst, including heating in
vacuum or in oxygen or hydrogen atmospheres. A combination of electron microscopy, infrared absorp-
tion spectroscopy, X-ray photoelectron spectroscopy, and catalytic kinetic measurements were used to
investigate the effect of those treatments on the decomposition and elimination of the material associ-
ated with the dendrimer and on the activity and selectivity of the resulting catalysts. Neither heating in
vacuum nor burning with oxygen was deemed acceptable, because much graphitic material is deposited
with the former treatment and metal oxidation is induced by the latter. Hydrogen pretreatments, on the
other hand, lead to fairly active catalysts with high selectivity for the conversion of trans olefins to their
cis counterparts.
Received 14 January 2010
Received in revised form 27 April 2010
Accepted 28 May 2010
Available online 4 June 2010
Keywords:
Dendrimer-encapsulated nanoparticles
Platinum catalysts
Olefin conversion
Cis–trans isomerization
Selectivity
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
[15,16]. The resulting dendrimer-encapsulated metal nanoparti-
cles (DENs) have already been successfully tested in a few catalytic
Thanks to recent advances in materials-science synthetic
methodology, there has been a renewed interest in the design of
well-defined heterogeneous catalysts [1]. Arguably the key chal-
lenge in heterogeneous catalysis for the 21st century is the control
of selectivity in highly complex processes [2–4], and that requires
exquisite control over the design of the catalytic active site. Typi-
cal heterogeneous catalysts are comprised of an active phase finely
dispersed onto a high-surface-area support, and the process of dis-
persion often leads to the deposition of ill-defined phases with
a multitude of local structures. Recently, though, novel colloidal
and self-assembly approaches have been used to synthesize cata-
lyst with metal nanoparticles displaying excellent size and shape
distributions [5–7]. Those catalysts have also been shown to dis-
play selective behavior toward specific catalytic conversions [8,9],
in many instances for reactions where no structure sensitivity had
been identified before [10,11].
Colloidal chemistry in particular has proven quite adept to the
synthesis of metal nanoparticles with different well-defined shapes
[12–14]. Unfortunately, it is quite difficult to control the growth of
colloidal particles to sizes below ∼5 nm. This can be a limitation
in catalysis, where smaller particles are often required. An alterna-
tive way to grow small nanoparticles, with diameter of the order of
1–2 nm and narrow size distributions, is using dendrimer templates
applications [17–20], and have even been shown useful when used
directly in liquid phase [21–23]. The use of such small particle size
is expected to display unique catalytic behavior in many reactions.
We in our laboratory have focused on the study of the pro-
motion of hydrocarbon conversions by metal catalysts [4,24–30],
and have recently turned our attention to the isomeriza-
tion of carbon–carbon double bonds in olefins using platinum
[7,10,11,31–35] or palladium [36,37] surfaces. It has been tradi-
tionally believed that this mild and mechanistically subtle reaction
is not affected by the structure of the surface of the catalyst. How-
ever, we have recently established that that is in fact not the case:
it was shown in our work that the (1 1 1) facets of the Pt metal favor
the preferential rearrangement of trans-olefins to their cis counter-
parts [10], whereas its (1 0 0) facets (and rougher surfaces) promote
the reverse cis-to-trans conversion [11]. As an extension of that
work, we are now turning our attention to the effect of particle size,
since small platinum particles display unique rough surfaces and
also particular electronic properties bordering on those expected
from discrete molecular structures.
In this report we provide a summary of our initial results on
the preparation and catalytic activity of Pt-DEN-based catalysts
for the promotion of olefin conversion reactions. It is shown that
well-defined catalysts with Pt nanoparticles ∼1.5 nm in diame-
ter can be made using a fourth-generation hydroxyl-terminated
poly(amidoamine) (PAMAM) dendrimer and by dispersing those
on a high-surface-area silica support. Activation of the catalyst may
potentially be accomplished by burning the organic matter using
∗
Corresponding author. Tel.: +1 951 827 5498; fax: +1 951 827 3962.
E-mail address: zaera@ucr.edu (F. Zaera).
0926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.05.045

M.A. Albiter et al. / Applied Catalysis A: General 391 (2011) 386–393
387
Fig. 1. Transmission electron microscopy (TEM) image and energy-dispersive X-ray spectroscopy (EDS) spectra for the Pt-based dendrimer-encapsulated nanoparticles
(Pt-DEN) used in the research reported here. The EDS corresponds to samples before and after oxidation in an O2 atmosphere at 675 K.
oxygen, but that sometimes leads to the oxidation of the metal and
to a diminished activity. Instead, we show that it is best to pretreat
these catalysts with hydrogen exclusively. No significant sintering
of the metal was seen until temperatures above ∼625 K. In terms of
selectivity, a trade off between butane and 1-butene formation was
seen during the conversion of trans-2-butene, with the yields for
the latter increasing with temperature at the expense of the former.
Perhaps more interestingly, a high selectivity was seen towards the
production of cis-2-butene, its yield amounting to more than 50%
of all products formed in the temperature range studied.
chromatograph (GC) equipped with a 23% SP-1700 on 80/100 Chro-
mosorb PAW column (30 ft × 1/8 in. o.d., stainless steel; Supelco)
and a flame ionization detector [46–48].
All liquid and solid reagents used in the synthesis of the Pt-DENs,
sol–gel silica, and regular Pt/SiO2 catalysts, including the PAMAM
dendrimer, were purchased from Aldrich and used as supplied. The
trans-2-butene gas (>95% purity) was purchased from Matheson
and the hydrogen (>99.995% purity) and argon (99.998% purity)
from Liquid Carbonic, and also used as received.
3. Results
2. Experimental
The first step in our research was the synthesis in solution phase
of the dendrimer-encapsulated Pt nanoparticles (Pt-DENs) that
were used in our catalyst preparation. The size and shape of those
nanoparticles were checked by transmission electron microscopy
(TEM). As shown in Fig. 1, most nanoparticles show a nearly spher-
ical shape, and approximately the same size. Statistical analysis of
TEM images such as this led to an estimate for the particle diameter
of ꢀdꢁ = 1.5 0.3 nm. In addition, the energy-dispersive X-ray spec-
troscopy (EDS) data shown at the top of Fig. 1 indicate that those
nanoparticles contain substantial amounts of carbon, as expected
since the original particles are encapsulated in the dendrimer tem-
plate. The bottom EDS trace in Fig. 1 indicates that all this carbon
could be burned in an oxygen atmosphere at temperatures around
675 K. Not only all EDS carbon signal was eliminated that way, but
the Pt signal was also increased, presumably because the Pt-DENs
become exposed after the removal of the dendrimer matter.
The thermal stability of our Pt-DENs was investigated further
using X-ray photoelectron spectroscopy (XPS). Fig. 2 summarizes
the C 1s (left) and Pt 4f (right) XPS data obtained after anneal-
ing the Pt-DENs under vacuum at different temperatures. The
Pt-DENs were first deposited from an ethanol suspension onto a
niobium foil (the XPS data of which are provided at the bottom
of Fig. 2 for reference), after which the solvent was left to evap-
orate. Only carbon was detected in the analysis of the original
deposited Pt-DENs (Fig. 2, second traces from bottom), presumably
because the Pt particles are masked by the organic matter from
the dendrimers. Deconvolution of the C 1s XPS trace indicated the
presence of one main peak at 285.3 eV binding energy amounting
to approximately 80% of the total signal intensity, and a broader
feature around 287.0 eV accounting for the remaining 20% (decon-
volution not shown). Although the original PAMAM dendrimers
contain several types of carbon atoms, their chemical environ-
The experiments were carried out using nanoparticles con-
taining an average of 40 platinum atoms encapsulated in a
fourth-generation
(PAMAM)
hydroxyl-terminated
[G4-OH(Pt40)].
poly(amidoamine)
These dendrimer-
dendrimer
encapsulated nanoparticles (DENs) were synthesized by mixing
together appropriate aliquots of aqueous solutions of K2PtCl4
and G4-OH followed by reduction of the metal with a tenfold
excess of NaBH4 [38,39]. The supported catalysts were prepared by
dispersing either 0.25 or 0.50 wt% Pt-DENs on a homemade sol–gel
silica via wetness impregnation [40]. Regular Pt/SiO2 catalysts
were also prepared for reference, by impregnation of the support
using hexachloroplatinic acid as reported in the literature.
Initial characterization of both the original Pt-DEN made by
the colloidal synthesis and the catalysts that result after their
dispersion on the silica support was carried out by transmission
electron microscopy (TEM) using a FEI-Philips CM300 instrument
equipped with an EDAX energy-dispersive X-ray spectrometer
(EDS) with a Si/Li detector window [41]. X-ray photoelectron
spectroscopy (XPS), used to determine atomic compositions and
oxidation states, was performed using a Leybold EA11-MCD system
equipped with an Al-K␣ X-ray source (1486.6 eV) and a 100-mm
concentric hemispherical electron energy analyzer [42,43]. Survey
and high-resolution spectra were acquired using band-pass ener-
gies of 100.8 and 31.5 eV, respectively, which correspond to spectral
resolutions of ∼2.0 and ∼0.9 eV once the natural line width of the X-
ray source is taken into consideration. Infrared absorption spectra
were acquired in transmission mode using a Bruker FT-IR spec-
trometer and a cell designed for in situ treatments [44,45]. The
kinetics of conversion of trans-2-butene with hydrogen were fol-
lowed in a 150 ml close-loop recirculation reactor by periodically
taking 0.1-ml aliquots of the gas sample and analyzing them in a gas

388
M.A. Albiter et al. / Applied Catalysis A: General 391 (2011) 386–393
Fig. 3. Infrared absorption spectra for catalysts made out of 0.25 wt% Pt-DEN
dispersed on a sol–gel silica support after heating in vacuum to the indicated tem-
peratures. The spectra are reported in differential mode, ratioed against the previous
trace, in order to highlight the changes that occur after each step.
Fig. 2. C 1s (left) and Pt 4f (right) X-ray photoelectron spectroscopy (XPS) data from
our Pt-DEN samples as a function of the treatment used for their cleaning. Data are
shown for the nanoparticles as prepared, after annealing to 475, 675 and 875 K, after
heating at 1000 K in O2 for 1 min, and after 10 min of ion sputtering.
with PtO2 [49]. Finally, ion sputtering was used to further remove
the carbon-containing surface species. That also led to the removal
of some of the platinum (Fig. 2, top traces).
ments are for the most part not sufficiently different to display
measurable differences in binding energy. Accordingly, we assign
the main C 1s XPS feature to all carbon atoms bound to nitrogen
(aminic carbons) and to hydroxyl groups, and the broad high-
binding-energy peak to the carboxylic carbons. This assignment is
consistent with what has been reported for other hydrocarbons
[49].
Additional experiments on the thermal stability of the Pt-DENs
after dispersion onto a high-surface-area sol–gel silica support
were performed using transmission infrared (IR) absorption spec-
troscopy. Fig. 3 reports representative traces for the catalysts after
heating under vacuum to the indicated temperatures, in differen-
tial mode (ratioed against the trace from the previous treatment)
in order to highlight the changes induced by each temperature.
A vibrational mode assignment, based on typical group vibrations
[50] and previous reports [51,52], is provided in Table 1. The first
thing to notice from the data in Fig. 3 is that all peaks but the one
at 3741 cm−1 point upwards, indicating the disappearance of the
corresponding molecular moieties. The 3741 cm−1 feature corre-
sponds to the O–H stretching frequency of isolated hydroxo groups
in silica [53], and the fact that it appears as a downward peak indi-
cates either that new OH species form on the silica surface or, more
likely, that the surface OH are freed upon thermal conversion of the
adsorbed Pt-DENs. It is quite possible that the dendrimers attach
to the silica surface via H-bonds between the terminal OH groups
Annealing of the Pt-DENs at 475 K under vacuum leads to the
loss of approximately 40% of the C 1s XPS signal. This indicates
a significant loss of the dendritic structure. On the other hand,
the remaining C 1s XPS peak is still centered around 285.3 eV
after this treatment, which implies that the remaining carbon
remains in chemical environments similar to those encountered
in the original organic framework. One possible explanation for
this observation is that the dendrimer may initially decompose
via the breaking away of its outer layers. In any case, sufficient
carbon-containing material remains on the surface to still encap-
sulate the platinum nanoparticles, since no Pt 4f XPS signal is yet
detected at this stage. That is no longer the case after anneal-
ing at temperatures above 675 K, at which point a small amount
of additional carbon is lost, somewhere between 10 and 20% of
the original signal, and the remaining C 1s XPS signal is redis-
tributed into two new peaks centered at 284.7 eV (major peak) and
282.8 eV (minor feature). The first feature in this spectrum corre-
sponds to a graphitic deposit, presumably the result of complete
dehydrogenation of the organic surface material, whereas the sec-
ond is easily identified with a metal carbide, a surface film that
forms by reaction with the niobium substrate [49]. At this stage
a small Pt 4f XPS signal becomes visible, with the Pt 4f7/2 peak
centered around 74.8 eV, indicative of non-metallic, oxidized, plat-
inum.
Table 1
Vibrational assignment for the infrared absorption peaks in the spectra of the silica-
supported Pt-DENs reported in Fig. 3. Relative intensities are given in parenthesis.
All modes are associated with the PAMAM dendrimer unless otherwise indicated.
Wavenumber (cm−1
)
Vibrational mode
3741 (downward)
3450–3520 (broad)
3450–3520 (broad)
2965(strong)
O–H stretching, isolated OH in silica
O–H stretching, H-bonded OH in silica
O–H and N–H stretching
CH2 asymmetric stretch
2934(strong)
CH2 asymmetric stretch
2880(medium)
2865(medium)
1708 (strong)
CH2 symmetric stretch
CH2 symmetric stretch
C
O stretching (amide I)
An attempt was made to remove the carbon deposits result-
ing from the high-temperature annealing of the Pt-DEN by burning
them in an oxygen atmosphere at 1000 K. Some carbon removal
was indeed possible even after a 1 min O2 treatment (Fig. 2, second
from top traces), mostly the graphitic layer likely to be covering
the platinum nanoparticles. The Pt 4f7/2 peak is somewhat more
intense at this point, and centered at 75.0 eV, a value consistent
1599 (very weak)
1580 (very weak)
1490(weak)
N–H deformation (amide II)
N–H deformation (amide II)
CH2 scissoring
1466 (medium)
1450 (shoulder)
1388 (medium)
1301 (strong)
CH2 scissoring
N–H bending
CH2 wagging/rocking
C–N stretching (amide III), O–H deformation

M.A. Albiter et al. / Applied Catalysis A: General 391 (2011) 386–393
389
of the PAMAM and the surface hydroxo species, and that as the
organic matter decomposes those bonds break and free the surface
species.
All remaining features in the IR absorption spectra in Fig. 3
show up as positive bands, the result of a loss in signal with
increasing temperature, and evolve in approximately the same
way, implying extensive decomposition of the dendrimers. The
decomposition starts at temperatures as low as 350 K, but is the
most significant at 400 K and is almost complete by 475 K. There
are some subtle changes in the relative intensities of the peaks as
a function of temperature, notably an initial prominence of the
peaks at 2934 and 2865 cm−1 due to the asymmetric and sym-
metric C–H stretching of some of the methylene moieties. This
suggests an early dehydrogenation of some of the hydrocarbon
chains. Also to note is the early disappearance of the strong fea-
ture at 1301 cm−1, which is gone by 450 K. This can be interpreted
as the result of the elimination of the terminal OH groups in PAMAM
by that temperature. Finally, the broad feature between 3450 and
3520 cm−1 shifts to higher frequencies with temperature, per-
haps also a reflection of the loss of the terminal PAMAM hydroxo
moieties.
The possibility of removing all organic matter and of activating
the silica-supported Pt-DENs for catalysis using either hydrogen or
oxygen was tested using XPS. Fig. 4 shows typical C 1s (left) and N 1s
(right) XPS data obtained after such treatments. The as-deposited
Pt-DEN/SiO2 display two features in the C 1s XPS, at approximately
285.1 and 287.4 eV. The first peak is easily assigned to the den-
drimer by comparison with Fig. 2, while the second is typical of
metal carbonyls [54], and suggests some adsorption of CO on the Pt
nanoparticles. A peak is also seen in the N 1s XPS data at 401.9 eV,
a surprisingly high value only seen in multiply bonded nitrogen or
in amines in highly ionic environments [54]. This again suggests
Fig. 4. C 1s (left) and N 1s (right) XPS for a 0.25 wt% Pt-DEN/SiO2 catalyst after
treatment in H2 (100 Torr, 2 h, 425 K) and after sequential treatments in O2 and H2
(100 Torr each, 4 and 3 h, respectively, 575 K).
hydrogen bonding to the OH groups on the surface, perhaps with
the amine nitrogen atoms of the dendrimer.
Hydrogen treatments may help remove some of the dendrimer
material in the Pt-DEN/SiO2, but more severe oxidation using O2
may be needed to remove all that organic matter. Indeed, the C 1s
XPS data in the left panel of Fig. 4 shows a shift to 285.5 eV after
Fig. 5. TEM of a 0.25 wt% Pt-DEN/SiO2 catalyst after treatment in H2 at different temperatures, from room temperature to 675 K.

390
M.A. Albiter et al. / Applied Catalysis A: General 391 (2011) 386–393
exposure of the catalyst to H2 at 425 for 2 h, indicating a change in
the nature of the carbon-containing species. Significant amounts
of carbon are seen even after sequential O2 and H2 treatments at
575 K, but that most likely reflects the presence of carbon from con-
tamination during the handling of the catalyst, during its insertion
into the XPS instrument. A better indication of the fate of the den-
drimers is provided by the N 1s XPS signal (Fig. 4, right), which is
reduced by approximately 40% upon the H2 exposure at 425 K, and
completely after the more severe O2 + H2 treatment at 575 K.
The degree of Pt-DEN dispersion in the Pt-DEN/SiO2 catalysts
and the possibility of sintering upon thermal treatment were inves-
tigated by TEM. Fig. 5 displays typical TEM images obtained after
annealing at different temperatures between 300 and 675 K under
a hydrogen atmosphere. Good dispersion is seen in the original cat-
alysts, with Pt nanoparticles displaying similar size distribution as
in the colloidal samples, before deposition on the silica support.
Indeed, the average diameter of the Pt nanoparticles in the sup-
ported catalyst is still ꢀdꢁ ∼ 1.5 nm. The nanoparticles retain their
size, without sintering, until temperatures above 625 K, at which
point some Pt nanoparticles with diameters of d ∼ 2–3 nm can be
seen. Nevertheless, even after H2 treatments at 675 K, the major-
ity of the Pt nanoparticles still retain their original size; those that
sinter are still a minority.
Fig. 6. Catalytic activity of our 0.25 wt% Pt-DEN/SiO2 catalyst toward trans-2-
butene conversion as a function of pretreatment. Shown are the percentages of
trans-2-butene converted as a function of time in our batch reactor for a reaction
temperature of 425 K and starting with a 10 Torr C4H8 + 0.5 Torr H2 + 500 Torr Ar
reaction mixture. Left: comparison of results from pretreating the catalyst at 575 K
with O2 only, H2 only, or O2 followed by H2 (all at 400 Torr, 2 h each). Right: kinetic
catalytic data after H2 pretreatments at different temperatures and for different
periods of time.
Next, the Pt-DEN/SiO2 catalysts were tested for their ability to
promote the conversion of trans-2-butene in the presence of small
amounts of hydrogen. A reaction mixture composed of 10 Torr of
trans-2-butene, 0.5 Torr of hydrogen, and 500 Torr of argon was
used in all the kinetic experiments reported here, which were
carried out in a close-loop batch reactor. Fig. 6 displays typical
conversion curves as a function of reaction time at 425 K for cat-
alysts treated under different conditions. The left panel compares
the total activity of the catalyst after treatments with either H2 or
ments at 425 K, but full promotion is attained only after hydrogen
exposures at temperatures of 475 K or above.
Fig. 7 displays the kinetic data obtained at three different reac-
tion temperatures for a catalyst activated with H2 at 625 K. Similar
total activity was observed in the temperature range studied, from
375 to 475 K, but changes in selectivity were clearly detected: more
1-butene is produced at higher temperatures at the expense of
lower yields for hydrogenation to butane. To note in particular is
the fact that the production of cis-2-butene from isomerization of
the central carbon–carbon double bond accounts for more than half
O2 alone or after a sequential O2 + H2 treatment (all at 575 K and for
2 h per gas). Clearly, high activities are obtained only after hydro-
gen treatments; treatments with oxygen lead to the oxidation of
the platinum nanoparticles (Fig. 2), and that appears to passivate
the surface. The right panel of Fig. 6 reports the activity of the cata-
lyst after hydrogen treatments at different temperatures and after
different lengths of time. Some catalyst activity is seen after treat-
Fig. 7. Kinetic data for the conversion of trans-2-butene on a 0.5 wt% Pt-DEN/SiO2 catalyst as a function of reaction temperature. Shown are the accumulated yields for butane
(left), 1-butene (center), and cis-2-butene (right) as a function of time. The catalyst was precondition by treatment with 400 Torr of H2 at 625 K, and a 10 Torr C4H8 + 0.5 Torr
H2 + 500 Torr Ar reaction mixture was used.

M.A. Albiter et al. / Applied Catalysis A: General 391 (2011) 386–393
391
Fig. 8. Kinetic data for the conversion of trans-2-butene on regular Pt/SiO2 catalysts. Data are shown as a function of reaction time (left), platinum loading (center), and
temperature (right). The reaction conditions are similar to those reported in Figs. 6 and 7.
of the products, and that double-bond migration can account for as
much as another 30–40%. This makes the Pt-DEN based catalysts
quite selective toward olefin rearrangements.
ing their structure, in particular their narrow size distribution. In
this project we explored different treatments, and ultimately tested
their effectiveness by evaluating the performance of the resulting
catalysts towards the conversion of trans-2-butene.
Finally, Fig. 8 reports similar kinetic data for a set of regular
Pt/SiO2 catalysts, prepared without the aid of dendrimers as the
other catalysts reported above. The left panel reports the product
accumulation results for a run with trans-2-butene and hydrogen
carried out at 475 K, under identical conditions to those used in the
comparable experiment in Fig. 7. In general, the reaction on this
catalyst seems to occur at a faster rate, since the final yields are
reached by 100 min of reaction. Another, perhaps more important,
noticeable difference is that in this case the production of 1-butene
is almost the same as that of cis-2-butene: the regular catalysts
does not display the same high selectivity toward carbon–carbon
double-bond cis–trans isomerization seen with the Pt-DEN/SiO2
catalysts. Additional tests were carried out with catalysts of dif-
ferent platinum loadings, between 0.1 and 1.0 wt% (Fig. 8, center),
and at different reaction temperatures, from 375 to 475 K (Fig. 8,
right), but in no case the high cis–trans isomerization reported with
Pt-DEN/SiO2 could be achieved. It should be noted, however, that
the differences are not particularly large. This is because the aver-
age size of the Pt nanoparticles in the regular catalysts, according to
measurements from TEM images (not shown), are also small, on the
order of 1.5–2.0 nm for all catalysts but the one with 0.1 wt% Pt (for
which the average size is ∼1.4 nm). The main difference between
the regular and the dendrimer-based catalysts is the narrow size
distribution obtained for the latter.
Supported platinum catalysts based on dendrimer-
encapsulated nanoparticles have been studied by a few groups
already in the past [20,55]. The work of Chandler and co-workers
[18,56–58] and that of the groups of Williams and Amiridis [59–62]
merit particular mention, since they have offered the most detailed
characterization of the chemistry involved in the treatment of
metal-DEN-based supported catalysts. Given that intuitively the
activation of these catalysts is expected to require the burning of
the organic matter associated with the dendrimer, most catalysts
pretreatments tried so far have involved exposures to an oxygen
atmosphere at different temperatures [56]. For instance, catalysts
prepared by dispersion of Pt-DENs onto DAVICAT SI-1403 silica
were shown to start loosing the dendritic material at temperatures
as low as 350 K, but to require higher temperatures, up to 575 K or
more, to remove all the amide groups [18,56,58]. Similar results
have also been reported using other supports such as alumina or
zirconia [62]. However, it has also been shown that oxygen treat-
ments lead to the oxidation of the platinum nanoparticles [39],
and consequently to a diminished catalytic activity. Subsequent
reduction using either hydrogen [18] or carbon monoxide [58] was
proven in some instances to increase the capacity of the platinum
nanoparticles for gas adsorption, and also for improved catalytic
activity.
The results reported here are in general qualitative agree-
ment with those reported before. In particular, we show that the
organic material from the Pt-DENs used to prepare Pt heteroge-
neous catalysts thermally decomposes in vacuum (Figs. 2 and 3)
but can be more effectively removed using an oxygen atmo-
sphere (Figs. 1, 2 and 4). This is true for the Pt-DENs either alone
(Figs. 1 and 2) or once dispersed on the silica support (Figs. 3 and 4).
It was also seen that the decomposition of the dendrimer occurs
over a wide range of temperatures (Figs. 2 and 3), and may involve
several steps, possibly including the partial elimination of the outer
layers of the dendrimer and/or partial dehydrogenation steps in the
early, low-temperature, stages of these reactions (Fig. 3). A third
observation in agreement with previous research is that oxidation
4. Discussion
In this report we summarize the data obtained in our attempt
to prepare well-defined catalysts consisting of platinum nanopar-
ticles finely dispersed on a high-surface-area silica sol–gel support.
The platinum nanoparticles were first prepared in solution using
dendrimers as templates in order to obtain small particles of nearly
uniform size, to test the possibility to manipulate selectivity in
hydrocarbon conversion reactions. One challenge in this approach
is to develop a procedure for activating the catalyst once the
nanoparticles are dispersed onto the porous support while preserv-

392
M.A. Albiter et al. / Applied Catalysis A: General 391 (2011) 386–393
of the Pt-DENs occurs upon thermal treatments in O2 atmospheres
(Figs. 2 and 6), and that their reduction to a metallic state and the
restitution of their catalytic activity can be achieved upon sub-
sequent treatments with H2 (Fig. 6). Finally, we also report the
preservation of the dispersion and size distribution of the Pt-DENs
on the silica support upon treatments in hydrogen atmospheres to
temperatures of up to 625 K.
375 and 475 K: any potential increase in the reaction rate constant
for 2-butyl formation (the common intermediate for all butane,
1-butene, and cis-2-butene formation) on the surface may be com-
pensated by a lowering in the steady-state coverage of the surface
hydrogen needed to make such intermediate.
5. Conclusions
One key observation from the data reported here is the fact that
the Pt-DEN/SiO2 catalysts are in general more active after pretreat-
ments using hydrogen, ether by itself or after previous oxidation
using O2 (Fig. 6). This is the case even though hydrogen treatments
alone are not always sufficient to remove the organic matter (Fig. 4).
The effect that the dendritic material exerts on the performance of
these catalysts is still not fully understood. On the one hand, dry
Pt-DEN based supported catalysts have been shown to not be capa-
ble of adsorbing any significant amounts of gas-phase molecules
such carbon monoxide [23,56]; only in solutions the dendrimer
structure appears to be able to open up and let those adsorbates dif-
fuse to the surface of the metal nanoparticles [23,59]. On the other,
significant catalytic activity has been reported for Pt-DEN-based
catalysts toward hydrocarbon conversion reactions even without
any pretreatment, and high performance has been seen with cata-
lysts treated with H2 only [20]. The same conclusion can be derived
from our catalytic kinetic data (Fig. 6). The reasons for this are not
fully known, but it is possible for the hydrocarbon reactants to at
least partially displace the dendrimer ligands closer to the metal
nanoparticles. It is also worth mentioning that our complementary
studies on the uptake of probe molecules such as carbon monoxide,
nitrous oxide, and acetylene on these catalysts show roughly the
same trends in reactivity reported here [63].
Next, we briefly highlight the implications of the selectiv-
ity results illustrated by the kinetic data in Fig. 7. Of particular
relevance is the high selectivity detected toward carbon–carbon
double-bond isomerizations in olefins, either toward cis–trans
interconversion, which here accounts for more than half of all
the products, or toward double-bond migration to the end carbon
atoms. The high extent of cis–trans isomerization seen here is par-
ticularly noteworthy, since the cis isomer is thermodynamically
less stable than the original trans-2-butene reactant. Preference
for cis formation was previously reported by us on catalysts with
exposed (1 1 1) platinum facets [7,10,11], but none of the catalysts
previously studied match the high cis-2-butene yields seen here.
We speculate that trans-to-cis conversion requires significant sur-
face reconstruction of the metal nanoparticles [10,34], and that
such Pt atomic rearrangements may be easier in the small nanopar-
ticles used here. This is nevertheless a subject that merits further
research.
Well-defined platinum-based catalysts were prepared via an
initial synthesis of dendrimer-encapsulated Pt nanoparticles (Pt-
DENs) and their subsequent dispersion onto a high-surface-area
homemade sol–gel silica support. The original Pt-DENs display
almost uniform size and shape, with most particles being close to
spherical and containing approximately 40 Pt atoms. The average
particle size was measured by TEM to be ꢀdꢁ ∼ 1.5 nm. Both size and
shape were preserved upon deposition of the DENs onto the silica
support.
The resulting Pt-DEN/SiO2 catalysts contained not only plat-
inum but also most of the organic matter associated with the
dendrimer template. In order to activate such catalysts, different
treatments were explored. It was shown that heating in vacuum
leads to extensive dehydrogenation over a wide range of tempera-
tures, with the most chemistry occurring around 400 K, and to the
formation of graphitic deposits, which block access to the metal
surface. Exposures to oxygen atmospheres at high temperatures
burn the organic matter, but also oxidize the platinum particles
and therefore render them almost inactive for catalysis. Hydrogen
treatments, on the other hand, may not fully remove the dendritic
material, but still make the catalyst quite active toward hydro-
carbon conversion reactions. The platinum nanoparticles are quite
stable under the conditions of these treatments, and do not start to
sinter until treatments with H2 at temperatures above 625 K.
The hydrogen-activated Pt-DEN/SiO2 catalysts were tested for
the conversion of trans-2-butene in the presence of small amounts
of hydrogen gas. Hydrogen pretreatments to temperatures as low
as 425 K resulted in catalysts with some activity, but H2 exposures
at 475 K or higher temperatures were required to attain maxi-
mum catalytic conversion. The Pt-DEN/SiO2 catalysts proved to be
particularly selective for the promotion of trans-to-cis isomeriza-
tion: cis-2-butene amounted to more than half of all the products
obtained under the conditions used in our catalytic kinetic exper-
iments. Also, the same as with other Pt-based catalysts, a switch
in selectivity from butane formation to 1-butene production was
identified as the reaction temperature was increased.
Acknowledgments
The other clear trend identified in Fig. 7 in terms of catalytic
selectivity is the trade off between the production of butane and 1-
butene. A significant decrease in butene hydrogenation to butane
is seen with increasing reaction temperature, compensated by an
almost mirror increase in the yield of 1-butene. Similar trends were
in fact seen before with supported Pt catalysts with both tetrahe-
dral and cubic nanoparticles [11]. Since the two reactions share a
common 2-butyl surface intermediate, this trend reflects the rel-
ative rates of the two competing subsequent steps, namely, the
reductive elimination with surface hydrogen to produce butane
and the beta-hydride elimination at the terminal carbon atom to
produce 1-butene [31,64–66]. The results in Fig. 7 clearly show a
lower effective activation barrier for the former reaction. This could
reflect an intrinsic lower barrier for the bimolecular butyl + H reac-
tion (compared to the unimolecular hydride elimination step), but
may also be influenced by a lowering in the steady-state concen-
tration of surface atomic hydrogen with increasing temperature.
The latter could also explain why there is an almost negligible tem-
perature effect in the total trans-2-butene conversion rate between
Funding for this project was provided by the U. S. National Sci-
ence Foundation. We also thank Prof. Richard M. Crooks for his
initial help with the preparation of the Pt-DENs and for his helpful
scientific input.
References
[1] F. Zaera, J. Phys. Chem. Lett. 1 (2010) 621–627.
[2] F. Zaera, J. Phys. Chem. B 106 (2002) 4043–4052.
[3] G.A. Somorjai, J.Y. Park, Angew. Chem., Int. Ed. 47 (2008) 9212–9228.
[4] F. Zaera, Acc. Chem. Res. 42 (2009) 1152–1160.
[5] A. Miyazaki, I. Balint, Y. Nakano, J. Nanopart. Res. 5 (2003) 69–80.
[6] R. Narayanan, M.A. El-Sayed, J. Phys. Chem. B 109 (2005) 12663–12676.
[7] I. Lee, R. Morales, M.A. Albiter, F. Zaera, Proc. Natl. Acad. Sci. U.S.A. 105 (2008)
15241–15246.
[8] J.C. Serrano-Ruiz, A. López-Cudero, J. Solla-Gullón, A. Sepúlveda-Escribano, A.
Aldaz, F. Rodríguez-Reinoso, J. Catal. 253 (2008) 159–166.
[9] G. Somorjai, C. Kliewer, React. Kinet. Catal. Lett. 96 (2009) 191–208.
[10] I. Lee, F. Delbecq, R. Morales, M.A. Albiter, F. Zaera, Nat. Mater. 8 (2009) 132–138.
[11] I. Lee, F. Zaera, J. Catal. 269 (2010) 359–366.
[12] T.S. Ahmadi, Z.L. Wang, T.C. Green, A. Henglein, M.A. El-Sayed, Science 272
(1996) 1924–1926.

M.A. Albiter et al. / Applied Catalysis A: General 391 (2011) 386–393
393
[13] H. Lee, S.E. Habas, S. Kweskin, D. Butcher, G.A. Somorjai, P. Yang, Angew. Chem.,
[43] B.C. Kan, J.H. Boo, I. Lee, F. Zaera, J. Phys. Chem. A 113 (2009) 3946–3954.
[44] F. Zaera, Int. Rev. Phys. Chem. 21 (2002) 433–471.
Int. Ed. 45 (2006) 7824–7828.
[14] Y. Xiong, J.M. McLellan, Y. Yin, Y. Xia, Angew. Chem., Int. Ed. 46 (2007) 790–794,
S790/791–S790/798.
[45] H. Tiznado, S. Fuentes, F. Zaera, Langmuir 20 (2004) 10490–10497.
[46] A. Loaiza, M. Xu, F. Zaera, J. Catal. 159 (1996) 127–139.
[47] M. Aryafar, F. Zaera, Catal. Lett. 48 (1997) 173–183.
[48] J. Wilson, H. Guo, R. Morales, E. Podgornov, I. Lee, F. Zaera, Phys. Chem. Chem.
Phys. 9 (2007) 3830–3852.
[49] C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder, G.E. Muilenberg (Eds.), Hand-
book of X-Ray Photoelectron Spectroscopy, Perkin-Elmer Corporation, Eden
Prairie, MN, 1978.
[50] G. Socrates, Infrared Characteristic Group Frequencies: Tables and Charts, 2nd
ed., Wiley, Chichester, 1994.
[15] D. Astruc, F. Chardac, Chem. Rev. 101 (2001) 2991–3023.
[16] R.W.J. Scott, O.M. Wilson, R.M. Crooks, J. Phys. Chem. B 109 (2005) 692–704.
[17] L.K. Yeung, R.M. Crooks, Nano Lett. 1 (2001) 14–17.
[18] H. Lang, R.A. May, B.L. Iversen, B.D. Chandler, J. Am. Chem. Soc. 125 (2003)
14832–14836.
[19] E. De Jesús, J.C. Flores, Ind. Eng. Chem. Res. 47 (2008) 7968–7981.
[20] W. Huang, J.N. Kuhn, C.K. Tsung, Y. Zhang, S.E. Habas, P. Yang, G.A. Somorjai,
Nano Lett. 8 (2008) 2027–2034.
[21] Y. Niu, R.M. Crooks, Chem. Mater. 15 (2003) 3463–3467.
[22] D. Méry, D. Astruc, Coord. Chem. Rev. 250 (2006) 1965–1979.
[23] M.A. Albiter, R.M. Crooks, F. Zaera, J. Phys. Chem. Lett. 1 (2010) 38–40.
[24] F. Zaera, Prog. Surf. Sci. 69 (2001) 1–98.
[25] F. Zaera, Appl. Catal. A 229 (2002) 75–91.
[26] F. Zaera, Mol. Phys. 100 (2002) 3065–3073.
[27] F. Zaera, Catal. Lett. 91 (2003) 1–9.
[28] F. Zaera, Top. Catal. 34 (2005) 129–141.
[29] F. Zaera, J. Mol. Catal. A 228 (2005) 21–26.
[30] F. Zaera, Chem. Rec. 5 (2005) 133–144.
[51] A. Manna, T. Imae, K. Aoi, M. Okada, T. Yogo, Chem. Mater. 13 (2001) 1674–
1681.
[52] F. Tarazona-Vasquez, P.B. Balbuena, J. Phys. Chem. B 108 (2004) 15982–15991.
[53] R.S. McDonald, J. Phys. Chem. 62 (2002) 1168–1178.
[54] C.D. Wagner, A.V. Naumkin, A. Kraut-Vass, J.W. Allison, C.J. Powell, J.R. Rumble
Jr., NIST X-ray Photoelectron Spectroscopy Database, NIST Standard Reference
Database 20, Version 3.5, NIST, 2007, http://srdata.nist.gov/xps/.
[55] Y. Niu, R.M. Crooks, Comp. Rend. Chim. 6 (2003) 1049–1059.
[56] D.S. Deutsch, G. Lafaye, D. Liu, B. Chandler, C.T. Williams, M.D. Amiridis, Catal.
Lett. 97 (2004) 139–143.
[31] I. Lee, F. Zaera, J. Phys. Chem. B 109 (2005) 2745–2753.
[32] I. Lee, F. Zaera, J. Am. Chem. Soc. 127 (2005) 12174–12175.
[33] I. Lee, F. Zaera, J. Phys. Chem. C 111 (2007) 10062–10072.
[34] F. Delbecq, F. Zaera, J. Am. Chem. Soc. 130 (2008) 14924–14925.
[57] L.W. Beakley, S.E. Yost, R. Cheng, B.D. Chandler, Appl. Catal. A 292 (2005)
124–129.
[58] A. Singh, B.D. Chandler, Langmuir 21 (2005) 10776–10782.
[59] D. Liu, J. Gao, C.J. Murphy, C.T. Williams, J. Phys. Chem.
12911–12916.
B 108 (2004)
[35] I. Lee, M.K. Nguyen, T.H. Morton, F. Zaera, J. Phys. Chem.
14117–14123.
C 112 (2008)
[60] O. Ozturk, T.J. Black, K. Perrine, K. Pizzolato, C.T. Williams, F.W. Parsons, J.S.
Ratliff, J. Gao, C.J. Murphy, H. Xie, H.J. Ploehn, D.A. Chen, Langmuir 21 (2005)
3998–4006.
[36] B. Brandt, J.-H. Fischer, W. Ludwig, S. Schauermann, J. Libuda, F. Zaera, H.-J.
Freund, J. Phys. Chem. C 112 (2008) 11408–11420.
[37] B. Brandt, W. Ludwig, J.H. Fischer, J. Libuda, F. Zaera, S. Schauermann, J. Catal.
265 (2009) 191–198.
[61] O.S. Alexeev, A. Siani, G. Lafaye, C.T. Williams, H.J. Ploehn, M.D. Amiridis, J. Phys.
Chem. B 110 (2006) 24903–24914.
[38] M. Zhao, R.M. Crooks, Angew. Chem., Int. Ed. 38 (1999) 364–366.
[39] M. Zhao, R.M. Crooks, Adv. Mater. 11 (1999) 217–220.
[40] R.D. Gonzalez, T. Lopez, R. Gomez, Catal. Today 35 (1997) 293–317.
[41] University of California Central Facility for Advanced Microscopy and Micro-
analysis, 2009, http://micron.ucr.edu/.
[62] D.S. Deutsch, A. Siani, P.T. Fanson, H. Hirata, S. Matsumoto, C.T. Williams, M.D.
Amiridis, J. Phys. Chem. C 111 (2007) 4246–4255.
[63] M.A. Albiter, F. Zaera, Langmuir, ASAP, doi:10.1021/la100753g.
[64] F. Zaera, Langmuir 12 (1996) 88–94.
[65] F. Zaera, Chem. Rev. 95 (1995) 2651–2693.
[42] H. Tiznado, F. Zaera, J. Phys. Chem. B 110 (2006) 13491–13498.
[66] Z. Ma, F. Zaera, Surf. Sci. Rep. 61 (2006) 229–281.