Work Function Calculation for Charged Systems:

[ling chen]

Dear Experts,

I am currently working on calculating the work function of a system and have encountered a few challenges that I would appreciate your insights on.

For the zero-charge system, my current method involves calculating the Hartree energy, taking the value in the vacuum region (usually a horizontal line, as shown in the attached figure), and subtracting the Fermi energy obtained from the .out file. This approach yields results that agree well with those obtained using VASP. 

 However, when dealing with charged systems, such as charge = -2, the Hartree energy in the vacuum region becomes a diagonal line (as shown in the attached figure). Using the average value and subtracting the Fermi energy results in a significant discrepancy compared to the VASP calculations. It has been suggested that I use the SCCS implicit water model to stabilize the vacuum energy level, which does indeed restore the horizontal line in the Hartree energy. However, the calculated work function deviates even further from the VASP results.

I am unsure about the cause of this discrepancy and would like to inquire if you have any recommendations for calculating the work function in charged systems. Additionally, I have noticed that the work function calculation is significantly influenced by the CUTOFF and REL_CUTOFF parameters in MGRID. What values would you suggest to be appropriate for these parameters? 

 I would be grateful for your expertise in addressing these concerns. Thank you in advance for your time and assistance. 

Kind Regards,

Ling Chen

[Lucas Lodeiro]

Hi,

About the vacuum level, when a charged system is calculated, the charges in the slab system interacts with the replica in the "non-periodic" direction, creating a electric field itself and making the Hartree potential non plane in the vacuum zone... this also happens sometimes for neutral slab systems with dipoles into the slab or on the surfaces which do not cancel each other. In VASP there are dipoles and electric field corrections to accomplish this task without this effect maintaining the 3D periodicity... I guess you are using those in your VASP calculations. In CP2K AFAIK there is dipole corrections, but do not remember if there is electric field correction, but here there is the possibility to turn off the periodicity of the cell in the "non-periodical" direction of the slab, maybe this is the way you need to go.

Regards - Lucas

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/035d2574-d796-4112-bdd7-f8e4cad2ab01n%40googlegroups.com.

[ling chen]

Dear [Recipient],

Thank you for your prompt and insightful response to my previous inquiry. Your expertise has been truly enlightening! I have three more questions that I would like to ask for your guidance on:

1.       I did not explicitly use the dipole and electric field corrections in my VASP calculations, but I did enable VASPSol. Does VASPSol come with built-in dipole and electric field correction capabilities?

2.       Similarly, when calculating a charged slab using the SCCS implicit solvent model in CP2K, the Hartree potential is indeed flattened in the vacuum region. However, the work function calculated using this method differs significantly from the results obtained using VASP. What could be the reason for this discrepancy?

3.       Could you please provide me with the dipole correction statement for CP2K?

Thank you in advance for your help, and I look forward to your response.

Best regards,

Ling Chen