As described previously,# TPD experiments utilize an in-house-modified chamber that is connected to the photoelectron spectrometer analysis chamber for direct sample transfer between techniques.  Custom-fabricated sample pucks integrate resistive heating in a Phi-compatible puck shape (HeatWave Labs, Inc., Watsonville, CA) while a K-type thermocouple in mechanical contact with the top face of the sample puck acquires temperatures.  A stage raises the sample puck to within 5 mm of the entrance grid of a 1–200 amu quadrupole, 70 eV electron-impact mass spectrometer with a channel-electron multiplier (channeltron) equipped detector (RGA200, Stanford Research Systems, Sunnyvale CA). An in-house designed, LabVIEW-based program acquires mass spectra as a function of time while concurrently controlling a power supply (PGM-2010, GW Instek, New Taipei City, Taiwan) to yield a linear heating rate of 0.33 K s–1. During sample heating, the mass spectrometer quantifies the intensity of species desorbing at x, y, z m/z.  Following Redhead, eq ?? models relates the desorption profile to an interaction energy between the adsorbate and the surface.*

dN/dT =A Nn β–1 exp (–Ea/RT)

Equation ?? utilizes the surface density of adsorbed species N, desorption order n, sample temperature T, heating rate β, and gas constant R. Assuming first-order desorption, model parameters include an Arrhenius-like preexponential factor, A, that we estimate as 1×1013 s–1 due to with first-order desorption;* and an Arrhenius-like activation energy for desorption, Ea. As an experimentally determined mass intensity vs temperature plot should be proportional to –dN/dT, simulated desorption traces with parameterized Ea values estimate the activation energy for desorption rather than strict fits as employed in photoelectron spectroscopy.

# Cite the first manuscript from the group that described our TPD instrument, 10.1021/acs.jpcc.8b05352.

* Cite Redhead, 10.1016/0042-207X(62)90978-8.