| 335 |
|
|
| 336 |
|
|
| 337 |
|
\subsection{Forcefield validation} |
| 338 |
< |
The CO-metal cross interactions were compared directly to DFT results |
| 338 |
> |
The CO-Pt cross interactions were compared directly to DFT results |
| 339 |
|
found in the supporting information of Tao {\it et al.} |
| 340 |
< |
\cite{Tao:2010} These calculations are estimates of the stabilization |
| 341 |
< |
energy provided to double-layer reconstructions of the perfect 557 |
| 342 |
< |
surface by an overlayer of CO molecules in a $c (2 \times 4)$ pattern. |
| 343 |
< |
To make the comparison, metal slabs that were five atoms thick and |
| 344 |
< |
which displayed a 557 facet were constructed. Double-layer |
| 340 |
> |
\cite{Tao:2010}, while the CO-Au results are interpreted on their own. |
| 341 |
> |
These calculations are estimates of the stabilization |
| 342 |
> |
energy provided to double-layer reconstructions of the perfect (557) |
| 343 |
> |
surface by an overlayer of CO molecules in a $c (2 \times 4)$ pattern. |
| 344 |
> |
To make the comparison, metal slabs of both Pt and Au that were five atoms thick and |
| 345 |
> |
which displayed a (557) facet were constructed. Double-layer |
| 346 |
|
(reconstructed) systems were created using six atomic layers where |
| 347 |
< |
enough of a layer was removed from both exposed 557 facets to create |
| 347 |
> |
enough of a layer was removed from both exposed (557) facets to create |
| 348 |
|
the double step. In all cases, the metal slabs contained 480 atoms |
| 349 |
|
and were minimized using steepest descent under the EAM force |
| 350 |
|
field. Both the bare metal slabs and slabs with 50\% carbon monoxide |
| 351 |
|
coverage (arranged in the $c (2 \times 4)$ pattern) were used. The |
| 352 |
|
systems are periodic along and perpendicular to the step-edge axes |
| 353 |
< |
with a large vacuum above the displayed 557 facet. |
| 353 |
> |
with a large vacuum above the displayed (557) facet. |
| 354 |
|
|
| 355 |
< |
Energies using our force field for the various systems are displayed |
| 356 |
< |
in Table ~\ref{tab:steps}. The relative energies are calculated as |
| 357 |
< |
$E_{relative} = E_{system} - E_{M-557-S} - N_{CO} E_{CO-M}$, |
| 358 |
< |
where $E_{CO-M}$ is -1.84 eV for CO-Pt and -0.39 eV for CO-Au. For |
| 359 |
< |
platinum, the bare double layer is slightly less stable then the |
| 360 |
< |
original single (557) step. However, addition of carbon monoxide |
| 361 |
< |
stabilizes the reconstructed double layer relative to the perfect 557. |
| 362 |
< |
This result is in qualitative agreement with DFT calculations in Tao |
| 363 |
< |
{\it et al.}\cite{Tao:2010}, who also showed that the addition of CO |
| 364 |
< |
leads to a reversal in stability. |
| 365 |
< |
|
| 365 |
< |
The DFT calculations suggest an increased stability of 0.08 kcal/mol |
| 366 |
< |
(0.7128 eV) per Pt atom for going from the single to double step |
| 367 |
< |
structure in the presence of carbon monoxide. |
| 355 |
> |
Energies calculated using our force field for the various systems are |
| 356 |
> |
displayed in Table ~\ref{tab:steps}. The relative energies are calculated |
| 357 |
> |
as $E_{relative} = E_{system} - E_{M-557-S} - N_{CO}*E_{M-CO}$, |
| 358 |
> |
where $E_{M-CO}$ is -1.8 eV for CO-Pt and -0.39 eV for CO-Au. Our |
| 359 |
> |
calculated CO-Pt minimum is actually at -1.83 eV at a distance of 1.53~\AA, |
| 360 |
> |
which was obtained from single-atom liftoffs from a Pt(111) surface. The |
| 361 |
> |
arrangement of CO on the single and double steps however, leads to a |
| 362 |
> |
slight displacement from the minimum. For a 1 ps run at 3 K, the single |
| 363 |
> |
step Pt-CO average bond length was 1.60~\AA, and for the double step, |
| 364 |
> |
the bond length was 1.58~\AA. This slight increase is likely due to small |
| 365 |
> |
electrostatic interactions among the CO and the non-ideality of the surface. |
| 366 |
|
|
| 367 |
< |
The gold systems show much smaller energy differences between the |
| 368 |
< |
single and double layers. The weaker binding of CO to Au is evidenced |
| 369 |
< |
by the much smaller change in relative energy between the structures |
| 370 |
< |
when carbon monoxide is present. Additionally, as CO-Au binding is |
| 371 |
< |
much weaker than CO-Pt, it would be unlikely that CO would approach |
| 372 |
< |
the 50\% coverage levels operating temperatures for the gold surfaces. |
| 367 |
> |
For platinum, the bare double layer is less stable then the original single |
| 368 |
> |
(557) step by about 0.25 kcal/mole per Pt atom. However, addition of carbon |
| 369 |
> |
monoxide to the double step system provides a greater amount of stabilization |
| 370 |
> |
when compared to single step system with CO on the order of 230 kcal/mole |
| 371 |
> |
for this system size. The absolute difference is minimal, but this result is in |
| 372 |
> |
qualitative agreement with DFT calculations in Tao {\it et al.}\cite{Tao:2010}, |
| 373 |
> |
who also showed that the addition of CO leads to a reversal in stability. |
| 374 |
> |
|
| 375 |
> |
The gold systems show a smaller energy difference between the clean |
| 376 |
> |
single and double layers when compared to platinum. Upon addition of |
| 377 |
> |
CO however, the single step surface becomes much more stable. These |
| 378 |
> |
results, while helpful, need to be tempered by the weaker binding energy |
| 379 |
> |
of CO to Au. From our simulations we see that at the elevated temperatures |
| 380 |
> |
we are running at, it is difficult for the gold systems to maintain > than 25\% |
| 381 |
> |
coverage, despite their being enough CO in the system. |
| 382 |
|
|
| 383 |
|
%Table of single step double step calculations |
| 384 |
|
\begin{table}[H] |
