How large is the difference between SZ and DZP(or DZ) for gold basis?

[bxu4]

Dear everyone,

      Recently, I perform the calculations of a large system, the electrods and the surface layers are chosen to be gold. In this system, the number of the gold atoms is about 200.  When I use the DZP basis for gold atoms, I find that the convergence can not be reached.
Therefore, I replace the DZP basis with SZ, and the convergence can be reached. But I think it is very different between the SZ and DZP results. So, I would like to know how large the difference is between SZ and DZP(or DZ) for gold basis, whether the qualitative results remain or not. Could someone provide some test results about the different basis for gold? I will appreciate any reply.

[Anders Blom]

Using SZ is generally not advisable, for some elements (not gold, though) it doesn't even work. The minimum basis set to consider, except for quick-and-dirty testing, should be SingleZetaPolarized (SZP).

The differences for gold between SZP and DZP are likely to be small. You can test this by e.g. running a simple band structure calculation for bulk Au, or a slab (to mimic the surface).

If SZ converges and DZP doesn't it can most likely be fixed by increasing the temperature to say 1000 K. Otherwise, make sure you have enough k-points and electrode layers. For more detailed assistance, post the script you use.