Analytical instrumentation
The innate complexity of catalytic systems makes the rationalization of structure-function relationships challenging. Ultra-high vacuum (UHV) provides a method of precisely controlling the environment around a catalyst system, and when paired with surface science, affords in-depth understanding of catalyst structures and compositions. Typically, these approaches have been limited to idealized planar model systems (e.g., single crystals or nanoparticles supported on metal oxide thin films) which has resulted in a so-called “materials gap” between surface science and “real-world” industrial catalysis, which is performed over powdered catalysts. Rarely have powdered catalysts been studied in ultra-high vacuum conditions.
Here we demonstrate how a Harrick Low Temperature Reaction Chamber from Specac can be utilized to perform ultra-high vacuum diffuse reflectance infrared Fourier transform spectroscopy (UHV-DRIFTS) to directly bridge this “materials gap,” affording in-depth understanding of powdered catalysts. As an example, we present UHV-DRIFTS measurements that reveal how adsorption features evolve over a standard Pd/Al2O3 catalyst as a function of CO pressure and consequently, CO coverage.
The experiments herein were performed using a Harrick Low Temperature Reaction Chamber (CHC-CHA-4) which was attached to an ultra-high vacuum system containing a residual gas analyzer (RGA, Hiden 3F PIC). A variable leak valve (Duniway, VLVE-1000) (Figure 1A) was attached to the ¼” VCO gas inlet port on the CHC-CHA-4 (Figure 1B) via a DN35CF to Swagelok adaptor, which allows gas to be introduced to the vacuum system so that the pressure can be precisely varied from < 10-8 to 10-3 Torr.
A constant flow of gas was supplied behind the leak valve from a gas handling system via mass flow controllers (Alicat Scientific). The gas outlet port on the chamber was connected to a VCO to Swagelok connector and attached to a ¼” stainless steel flex hose (Figure 1C). The hose was then connected to a three-port manual switching valve (Swagelok, SS-43GXS4) with two outputs: one which was connected to a scroll pump (Edwards nXDS15iC), and the other which was connected to the UHV system containing the RGA for detection of gas phase species. T
he reaction chamber was coupled with a Harrick Praying Mantis(tm) diffuse reflection accessory from Specac, which was then placed in the sample compartment of a Bruker INVENIO Fourier-Transform Infrared Spectrometer (FT-IR) for detection of gas phase and catalyst surface-bound species. An annotated picture of this setup is provided in Figure 1.
To prepare the sample, a 250-mesh stainless steel disc was placed in the sample cup and approximately 150 mg of 46 grit SiC was added to fill the cup to 1 mm from the top. Next, 11.21 mg of a standard Pd catalyst (5% Pd/Al2O3, ESCAT™ 1241, pressed and sieved to 50-70 mesh) was added to the sample cup, ensuring the catalyst was flat and flush with the sides of the sample cup. The temperature of the sample cup was monitored and controlled by the thermocouple inside the reactor bed.
The Low Temperature Reaction Chamber was then inserted into the Praying Mantis and pumped down to UHV. Great care should be taken when pumping down the cell, as any rapid change in pressure can blow the catalyst powder out of the sample cup due to the airflow from the bottom of the sample cup up into the dome. To pump down the system, first the scroll pump was used to achieve rough vacuum, then the three-port manual switching valve was used to change the flex hose from the scroll pump to the UHV system.
The typical base pressure for this setup was < 1 x 10-8 Torr, which was read via a wide-range gauge (Edwards WRG-S-DN40CF) at the inlet of the UHV system, after the three-port valve.
To remove any residual carbonaceous species, the fresh catalyst was first heated under vacuum to 200 ℃, then a 20% O2 (balanced in Ar) gas mixture was leaked into the system until the pressure in the vacuum chamber read 1 × 10-4 Torr. The sample was oxidized for approximately 2 hours, until no more CO2 production could be detected in the mass spectrometer. When checking for CO2 production the leak rate was temporarily reduced so that the pressure was 1 × 10-7 Torr to avoid damaging the mass spectrometer.
After initial oxidation treatment, the sample was reduced using a H2 pretreatment to generate a pristine metallic Pd surface. The sample was first heated to 350 ℃ under vacuum and a 10% H2 (balanced in Ar) gas mixture was leaked into the system until the pressure gauge read 3 × 10-5 Torr. The sample was reduced in this environment for 30 minutes. After the reductive treatment, the leak valve was closed, the cell was pumped down to the base pressure of the system and cooled to 50 ℃.
For the CO adsorption experiments, a 20% CO (balanced in Ar) gas mixture was leaked into the system containing the pristine Pd catalyst so that the system pressure varied from 5 × 10-8 to 5 × 10-4 Torr. DRIFT spectra were taken as the average of 500 scans with a resolution of 4 cm-1 using a liquid nitrogen cooled MCT detector. The background measurement was taken over the pristine Pd catalyst at 50 ℃. During the CO adsorption experiments, FTIR spectra were taken every ~30 seconds.
During the experiment, the CO pressure was held until the observed spectra did not change over time, after which the pressure was then increased. Figure 2 shows a representative spectrum of CO saturation at each given pressure.
The time resolution of the repeated DRIFT spectra allows us to observe the pressure-dependent population of binding sites. The three features indicating surface binding of CO to metallic Pd observed are well-defined in literature.1–3 The feature around 2100-2080 cm-1 corresponds to linearly bound CO atop a Pd atom, the feature around 1950-1900 cm-1 corresponds to bridge-bound CO, and the feature around 1850-1900 cm-1 corresponds to CO in a 3-fold hollow site (Figure 2).
At low pressures of CO, the 3-fold hollow sites saturate first, and as pressure increases the bridge and atop sites continue to populate. As pressure increases, the features blueshift to higher wavenumbers. This could be indicative of a coverage effect: as CO surface coverage increases, lateral interactions between adsorbed CO molecules increase and the back donation of electron density decreases, both of which are known to induce shifts at high coverages.
Alternatively, the shift could indicate a change in populated sites such as the increasing population of CO adsorbed to bridge and atop sites compared to 3-fold hollow sites, and/or a change in what subtypes of sites are populated. For instance, the populated atop site may shift from undercoordinated Pd atoms at edges and kinks to well-coordinated Pd atoms on terraces as pressure increases, which would be expected to induce a blueshift. The ability to precisely control the environment around a catalyst to observe the pressure-dependent preferential population of catalyst sites yields invaluable insight into the evolution of catalyst surfaces under dynamic conditions.
The work was supported by the Rowland Fellowship through the Rowland Institute at Harvard.
1. Zorn, K. et al. CO oxidation on technological Pd-Al2O3 catalysts: Oxidation state and activity. J. Phys. Chem. C 115, 1103–1111 (2011).
2. Perich, M. P. et al. In situ analysis of gas dependent redistribution kinetics in bimetallic Au-Pd nanoparticles. J. Mater. Chem. A 12, 32760–32774 (2024).
3. Marx, S., Krumeich, F. & Baiker, A. Surface Properties of Supported, Colloid-Derived Gold/Palladium Mono- and Bimetallic Nanoparticles. J. Phys. Chem. C 115, 8195–8205 (2011).
PIN 27.2 Apr/May 2026
.jpg)
Product directory
