(734c) Carbon Decoration in Supported Palladium Catalysts: Discrepancy in STEM-Chemisorption Particle Sizes

Introduction

       
Metal-support interactions have been known to stabilize metal nanoparticles in
harsh reaction environments or cause hindrances in the chemisorption surface
area during chemical reactions [1, 2]. Although it is necessary for the active
sites to remain exposed without any poisoning to allow the maximum adsorption
of reactants in heterogeneous reactions, it is also vital to maintain the
structural stability in chemical reactions at elevated temperatures. Carbon-supported
palladium catalysts find a wide variety of applications in the fine chemicals
industry, as electrodes in fuel cells and in a large number of
hydrogen-mediated aqueous phase catalytic reactions.   While it has been
frequently seen in the literature that active sites in the metal are blocked by
carbonaceous deposits [3], there is also ample evidence that carbon coatings
are frequently used to coat oxide supports and metal nanoparticles for
enhancing their hydrothermal stability in aqueous phase hydrogenation reactions
at elevated temperatures [4] due to their inherent hydrophobicity and
inertness.

During the preparation of palladium catalysts on
carbon black in our laboratory, it has been found that well dispersed catalysts
are formed as confirmed with Scanning Transmission Electron Microscopy (STEM)
and X-Ray Diffraction (XRD) but the chemisorption of hydrogen is substantially
suppressed on the catalysts leading to a disagreement between particle sizes
obtained by STEM and chemisorption [5]. This work focusses on understanding the
STEM-chemisorption discrepancy based on the hypothesis that during the
preparation of carbon supported palladium nanoparticles, the metal surface
becomes coated from the surface/sub-surface/interstitial carbon that migrates
from the support to the metal nanoparticles or with a partial surface blockage
by residual chloride from the chloride precursor used. This suppresses the
hydrogen chemisorption and leads to larger particle sizes as compared to
transmission electron microscopy TEM/X-ray diffraction (XRD). The discrepancy
is also believed to be dependant on the type of carbon used, the functional
groups on the surface, the precursor chosen and pre- treatment conditions of
the support.

Materials and Methods

       
Strong Electrostatic Adsorption (SEA) was used to deposit Pd-nanoparticles onto
two sets of different carbons: VXC72 (carbon black) and Darco G60 (activated
carbon) that were oxidized and heat-treated to different temperatures. The
pre-treated catalysts were then dried at 120°C and reduced in
hydrogen at 180°. The catalysts were characterized for particle sizes
using STEM, XRD and chemisorption. Systematic Temperature Programmed Oxidation
(TPO) studies were conducted to burn off the surface carbon and pre and post
TPO chemisorption sizes were compared. Chloride and non-chloride precursors
were also compared to determine the effect of chloride ions on the discrepancy.

       To understand the discrepancy better, the
ratio of chemisorption and STEM surface average sizes were plotted against the
pretreatment temperatures. The discrepancy for both DarcoG60 and VXC72 is the
greatest when they are oxidized and decreases with the increase in pretreatment
temperature which confirms that there is a relation between the surface oxygen
groups and carbon decoration. It is also seen that the unoxidized Darco series
did not show any discrepancy. Incidentally, this was the only series of samples
that had very large Pd particles (>10 nm) with wide standard deviations (Fig
1). After the TPO burn off experiment, the discrepancy is reduced by 50-60% for
the highest discrepancy (oxidized carbon samples) (Fig 2). For the unoxidized
Darco sample, there was no change in the ratio since the discrepancy did not
exist in the first place. However, it is seen that burning off the surface
carbon does not recover the chemisorption surface completely. This led to the
comparison of results using chloride and non-chloride precursors which showed
that chemisorption gives a 30-50% reduction in the discrepancy for the nitrate
precursor as compared to chloride precursor. It may, thus, be concluded that
carbon decoration coupled with chloride poisoning is responsible for the
discrepancy in the STEM-chemisorption particle sizes.

References

[1] D. Nerhing, H. Dreyer. Chem.Tech. 1960, 12, 343.

[2]
G.M. Schwab. Adv. Catalysis, 1978, 27, 1-22.

[3]
P. Forzatti , L. Lietti, Catalysis Today, 1999, 52,165-181.

[4]
H. N. Pham, A. E. Anderson, R. L. Johnson, T. J. Schwartz,B. J. O’Neill, P.
Duan, K. Schmidt-Rohr, J.
A. Dumesic and A. K. Datye, ACS Catalysis, DOI: 10.1021/acscatal.5b00329 ACS
Catal. 2015, 5, 4546−4555.

[5]
J.M.M.Tengco, Y.K.Lugo-Jose, J.R.Monnier, J.R.Regalbuto, Catalysis Today, 2015,
246, 9.