9
08
N. Musselwhite et al.
constant 4:1 for 6.5 nm. In a typical synthesis the metal salts
were added to a round bottom flask, triethylene glycol
to 433 K. Quantitative analysis of flow composition was
accomplished with a Hewlett-Packard (5890 Series II) GC
which was equipped with an Aldrich HP-1 capillary col-
umn and a flame ionization detector (FID). A PC based GC
Chemstation software (Hewlett-Packard) was utilized for
automatic GC sampling, data collection and post-run
processing.
(
10 mL) and polyvinylpyrrolidone (28 mg, 55,000 g/mol)
were added, and the solution was sonicated for 30 min to
ensure dissolving of the solids. The flask was then placed
under vacuum for 20 min and then repeatedly purged with
argon and vacuum pumped before finally being filled with
argon, to ensure an oxygen free environment. The flask was
then placed in a molten salt bath held at a constant 503 K for
1
h. During this time the solution changed to dark black in
3 Results and Discussion
color as nanoparticles were formed. Smaller sized bimetallic
nanoparticles were synthesized by addition of NaNO
Nanoparticles were characterized using a Jeol JEM-
2100F scanning transmission electron microscope
(STEM) equipped with an INCA electron dispersive
spectrometer (EDS) was used in scanning mode to obtain
information about the size and composition of the parti-
cles. Briefly, Pt and Rh L lines were simultaneously
probed by EDS using a point-to-point e-beam size of
1.5 nm, and then overlapped on the respective high angle
annular dark field (HAADF) image. STEM–EDS images
are shown in Fig. 1. The larger sized bimetallic particles
were found to be 6.4 ± 0.5 nm with Pt/Rh atom com-
positions of 80/20 and 93/7. Smaller sized Pt Rh were
3
-
5
7 9 10 mol) to a solution containing triethylene glycol,
(
-
PVP and Pt(acac) and Rh(acac) (1.0 9 10 mol total
4
2
3
metal) and following a similar reduction procedure as
reported above. After synthesis, the nanoparticles were
precipitated from solution by addition of acetone and sub-
2
sequent centrifugation (550 9 9.8 m/s , 20 min). Nano-
particles were then supported on MCF-17 type mesoporous
silica, which was synthesized by previously reported
methods [5].
Catalytic measurements were made utilizing a tubular
0
0
fixed catalyst bed reactor at ambient pressure. A 1/4
8
0
20
diameter stainless steel reactor was mounted vertically,
which allowed for a downflow operation. A 0.5 g sample
of catalyst was loaded, which corresponded to a
also characterized and found to have an atom composi-
tion of 80/20 Pt/Rh and size distribution of
a
2.5 ± 0.3 nm. No stray monometallic nanoparticles were
observed. Nanoparticles were also characterized utilizing
synchrotron based ambient pressure X-ray photoelectron
spectroscopy (AP-XPS), as a means to ascertain the
surface composition of the nanocatalysts under relevant
reaction conditions [6]. AP-XPS samples were prepared
by deposition of the particles onto a silica wafer by
Langmuir–Blodgett techniques. The wafer was then loa-
ded into the instrument for characterization. The particle
surface composition was determined by investigating the
Pt 4f and the Rh 3d spectra, utilizing both 380 and
650 eV excitation energies, respectively, both with
probing depths of 0.8 nm. The partial pressures of the
1
0–12 cm bed height. The catalyst was placed in the
center of the reactor tube, capped on each end with
purified thermal silica filter. The remaining space in the
reactor tube was filled with purified fused aluminum
granulate and capped with glass wool. Prior to loading,
the catalyst was pelletized and sieved to obtain
6
0–100 lm size granulates. To avoid misinterpretation
linked with mass and heat transfers, we kept the catalyst
in a kinetic region by holding the total hexane conversion
below 10 %.
The catalysts were first pretreated at 633 K under a gas
mixture of N (Praxair, 5.0 UHP, 10 sccm) and H (Prax-
2
2
air, 5.0 UHP, 10 sccm) for 2 h. The heating rate of the
system was 100 mTorr of H and 20 mTorr of hexane for
2
-
1
pretreatment was limited to 2 K min . After the pre-
treatment the reactor system was cooled to 513 K under the
same gas flow. The gas flow was then changed to 16 sccm
the experiments. The characterization data for the nano-
particles is summarized in Table 1.
By integration of the AP-XPS spectral data, it was
possible to determine the near surface composition of the
bimetallic nanoparticles. The percentage of platinum atoms
is plotted in Fig. 2a as a function of reaction conditions.
Notably, the near surface ratio was found to always be rich
in rhodium with respect to the bulk nanoparticles. Also, it
was determined that the surface composition is dynamic
and varies under different conditions. It was also found that
nanoparticle size had little influence on the surface com-
position of nanoparticles, as shown in Fig. 2b, c. The final
important conclusion is that under reaction conditions (5:1
H and n-hexane (Fluka, C99.0 %) was introduced using a
2
Teledyne ISCO 500D liquid flow pump at a rate of
-
.2 mL h into the preheating reactor head which was
1
1
maintained at 423 K. In the preheating zone, hexane
evaporated and mixed with H , resulting in a two-compo-
2
nent gas flow with a hexane:H ratio of 1:5 entering the
2
reactor at near ambient pressure. A Baratron type (890B,
MKS Instruments) manometer was used to monitor the
reactor inlet pressure. The reaction products were sampled
in the vapor phase at the reactor outlet and analyzed via an
in-line gas chromatograph (GC). All flow lines were heated
H :hex 360 °C) the near surface was found to have an
2
1
23
Musselwhite, Nathan
