How to saturate slab bonds with partial charge H in Dmol3?

Hello,

I am attempting to create a slab of the (001) GaSb surface for an adsorption study. To properly simulate the bonding of the Ga atoms in the bottom of the slab requires capping them with partially charged (1.25e) Hydrogen (see yellow atoms in the image below) I have not been able to find any documentation on specifying partial charge H atoms within DMol, but I did find a paper which does just that: Carter et al. Phys. Rev. B. 77, 115349, 2008. Is this still possible in DMol? If so, how would I go about it?

Thank you.

slab_frac_H.bmp

Answers 2
FH
FH 2014-02-24

Hi,

I had a brief look into how difficult it is to work with those partially charged hydrogens you asked about. We might be able to do something involving customized basis set files which you would have to set up by hand... However. This is not something we ever validated at Accelrys, nor is it something we would support for the time being. I am not sure if I can manage, but maybe I could help you out if you promise to do your own validation and make sure this works for you. Please let me know if you are interested.

In the meantime, you should be able to work with integer-charged H atoms in your GaSb surface - I would expect that an additional surface layer or two should screen the effects of the integer H's quite nicely. This might be the better short-term option for you.

Also, I'd like to query the sample image for the GaSb(111) surface you sent. As you show it, the top surface looks quite unstable to me (I have not done any calculation to check though). With all top atoms Sb and all bottom atoms Ga (or vice versa?), you are probably creating a macroscopic dipole across the entire slab, which is likely to destabilize the entire crystal. I think this holds true even if you manage to saturate the bottom slab with H-atoms, partially charged or not. Are you sure that your model is a reasonable and relevant description for the surface structure?

Best,

Felix

BA
BA 2014-02-24

Hi Felix,

Thank you for your reply. I guess the authors of the paper I linked must have done some customizing of the DMol basis set files on their own. For now, I will refrain from working with custom basis set files unless I see no other option. I have been working with this surface using other DFT software packages, so I will try comparing the effect of the different bond termination on my adsorption energies. Your suggestion to add more layers to screen out any charge is a good point, though I hope to eventually run MD calculations on this surface which would be hampered by the extra atoms.

Regarding the sample image, this was just a test structure I was setting up and does not reflect the actual surface I am studying. You are correct that the shown surface is unstable. Experimental and theory work shows that the alpha-(4x3) reconstruction (with 4 Ga-Sb dimers and one Sb-Sb dimer formed on the surface) minimizes the surface energy. Either way, I dipole does form in this slab, though I expect dipole corrections to properly smooth out the potential on the reactive surface.

Thank you for your assistance,

-Alejandro