Table of Contents
Fetching ...

CO$_2$ Dissociative Sticking on Cu(110)

Federico J. Gonzalez, Carmen A. Tachino, H. Fabio Busnengo

TL;DR

This study addresses CO2 activation on Cu(110), a system central to copper-catalyzed CO2 conversion, by combining density functional theory with an vdW-DF2 functional and an active-learning ANN potential energy surface to perform quasi-classical trajectory simulations across a broad range of impact energies and surface temperatures. The authors map dissociative sticking probabilities $P_{diss}$, molecular adsorption $P_{molads}$, and the final-state distributions of dissociation products, while characterizing the PES with minima (physisorbed CO2, bent CO2*, and dissociated CO_ads+O_ads) and low-energy transition states TS1 and TS2. Vibrational excitation increases reactivity, and dissociation can drive pronounced surface distortions, including Cu adatom formation, particularly at high $E_i$ and room temperature, offering a potential explanation for experimentally observed higher oxygen coverage at high energies. Overall, the work demonstrates that a rigorously trained ANN-PES based on DFT enables detailed, large-scale dynamics that reconcile several experimental trends and highlight surface reconstruction as a key factor in CO2 activation on Cu(110).

Abstract

In this work we investigate the dissociation of CO$_2$ on Cu(110) by performing density functional theory calculations using the vdW-DF2 exchange-correlation functional, with a potential energy surface parameterized using artificial neural networks. We computed quasi-classical trajectory calculations of molecular and dissociative adsorption probabilities as a function of the initial impact energy of the molecules and surface temperature, by comparing our results with available supersonic molecular beam experimental data for normal incidence. Concerning the general dependence of the molecular and dissociative adsorption probabilities on the initial translational energy of the molecules, our theoretical results agree with experiments. Also in agreement with experiments, we have found that dissociative adsorption is not affected by surface temperature between 50 and 400 K, for impact energies for which the dissociation probability is larger than $\sim 10^{-3}$. We have investigated the influence of impact energy and surface temperature on the final state of the dissociation products by extending the time integration of the reactive trajectories up to 10 ps. We have found that above $\sim 2.5$ eV and close to or above room temperature, CO$_2$ dissociation induces strong surface distortions including final structures involving Cu adatoms. The creation of Cu vacancy-adatom pairs is stimulated by the presence of both CO$_{ads}$ and O$_{ads}$ which interact strongly with the Cu adatoms and even give rise to unexpected (O-Cu-CO)$_{ads}$ linear moieties anchored to the surface by the dissociated O atom and involving a Cu adatom almost detached from the surface. These surface distortions produced by dissociation products of high energy CO$_2$ molecules at and above room temperature might explain recent experiments that have found a greater saturation oxygen coverage for high energy molecules.

CO$_2$ Dissociative Sticking on Cu(110)

TL;DR

This study addresses CO2 activation on Cu(110), a system central to copper-catalyzed CO2 conversion, by combining density functional theory with an vdW-DF2 functional and an active-learning ANN potential energy surface to perform quasi-classical trajectory simulations across a broad range of impact energies and surface temperatures. The authors map dissociative sticking probabilities , molecular adsorption , and the final-state distributions of dissociation products, while characterizing the PES with minima (physisorbed CO2, bent CO2*, and dissociated CO_ads+O_ads) and low-energy transition states TS1 and TS2. Vibrational excitation increases reactivity, and dissociation can drive pronounced surface distortions, including Cu adatom formation, particularly at high and room temperature, offering a potential explanation for experimentally observed higher oxygen coverage at high energies. Overall, the work demonstrates that a rigorously trained ANN-PES based on DFT enables detailed, large-scale dynamics that reconcile several experimental trends and highlight surface reconstruction as a key factor in CO2 activation on Cu(110).

Abstract

In this work we investigate the dissociation of CO on Cu(110) by performing density functional theory calculations using the vdW-DF2 exchange-correlation functional, with a potential energy surface parameterized using artificial neural networks. We computed quasi-classical trajectory calculations of molecular and dissociative adsorption probabilities as a function of the initial impact energy of the molecules and surface temperature, by comparing our results with available supersonic molecular beam experimental data for normal incidence. Concerning the general dependence of the molecular and dissociative adsorption probabilities on the initial translational energy of the molecules, our theoretical results agree with experiments. Also in agreement with experiments, we have found that dissociative adsorption is not affected by surface temperature between 50 and 400 K, for impact energies for which the dissociation probability is larger than . We have investigated the influence of impact energy and surface temperature on the final state of the dissociation products by extending the time integration of the reactive trajectories up to 10 ps. We have found that above eV and close to or above room temperature, CO dissociation induces strong surface distortions including final structures involving Cu adatoms. The creation of Cu vacancy-adatom pairs is stimulated by the presence of both CO and O which interact strongly with the Cu adatoms and even give rise to unexpected (O-Cu-CO) linear moieties anchored to the surface by the dissociated O atom and involving a Cu adatom almost detached from the surface. These surface distortions produced by dissociation products of high energy CO molecules at and above room temperature might explain recent experiments that have found a greater saturation oxygen coverage for high energy molecules.
Paper Structure (11 sections, 4 equations, 13 figures, 2 tables)

This paper contains 11 sections, 4 equations, 13 figures, 2 tables.

Figures (13)

  • Figure 1: Schematic representation of the CO$_2$ molecule in its equilibrium configuration far from the surface, and coordinate system used throughout this work. $\mathrm{Z}$=0 corresponds to the plane containing the outermost-layer Cu atoms, in the lowest-energy clean surface structure. High symmetry sites: top ($\mathrm{X}$=0,$\mathrm{Y}$=0); short-bridge ($\mathrm{X}$=a$_{Cu}/(2\sqrt{2})$,$\mathrm{Y}$=0); long-bridge ($\mathrm{X}$=0, $\mathrm{Y}$=a$_{Cu}/2$); hollow ($\mathrm{X}$=a$_{Cu}/(2\sqrt{2})$,$\mathrm{Y}$=a$_{Cu}/2$).
  • Figure 2: DFT-vdW-DF2 total energy of CO$_2$ in its equilibrium geometry in vacuum, as a function of the molecule surface distance measured by the $\mathrm{Z}$ coordinate of the C atom above the outermost-layer of surface atoms which are kept fixed in their equilibrium positions. Each line corresponds to a different surface high symmetry site: top, short-bridge (sb), long-bridge (lb), and hollow (hol) and/or a different orientation as indicated in the inset by the values of the polar and azimuthal angles of the molecule. For instance, hol 90,0 (lb 0) indicates that the center of mass of the molecule is on a hollow (long-bridge) site and the polar and azimuthal angles of (the polar angle of) the molecular axis containing the two C--O bonds are 90 deg and 0 deg respectively (is 0 deg).
  • Figure 3: (a) Histogram of energy differences (E$_\text{DFT}$ - E$_\text{NN}$) for the training, test, and validation sets. (b) Force prediction errors for the validation set.
  • Figure 4: Energetics of CO$_2$ physisorption, molecular chemisorption and dissociation on Cu(110), including the involved local minima and transition states found along minimum energy pathways connecting them.
  • Figure 5: (a) Top view, and (b) side view of the transition state TS2 along the minimum energy pathway for CO$_2$ dissociation on Cu(110), connecting the physisorbed and the most stable dissociated state. Dark and light orange distinguish topmost-layer and second-layer Cu atoms respectively. Relevant distances/coordinates: d(C-O$_1$)=1.39 Å, d(C-O$_2$)=1.21 Å, $\mathrm{Z_{CM}}$=1.58 Å, d(C-Cu$_1$)=2.05 Å, d(O$_2$-Cu$_2$)=2.02 Å, d(O$_2$-Cu$_3$)=2.13 Å, d(O$_2$-Cu$_4$)=2.45 Å. In (a) the red arrows represent the displacements of Cu atoms (larger than 0.1 Å) with respect their positions in the reference configuration (see the text) of lengths: 0.23 Å, 0.24 Å, 0.33 Å, and 0.15 Å for the Cu atoms labeled as 1, 2, 3, and 4 respectively.
  • ...and 8 more figures