# Fermi level change in surfaces with adsorbed molecules

In a DFT standard calculation of a solid 3D material (i.e. bulk metal), one of the properties that can be obtained is the Fermi level $$E_F$$. This feature is related with the required work to add an electron to this material.

In organic electronics, one possible utility of that value is to extract information about the charge-transfer properties of such material when it is contacted by a single molecule or even a molecular layer (organic semiconductor), thus obtaining the interfacial properties of the complete junction. As the orbitals of both entities tend to hybridize, the properties of the modified surface can be radically different from those of the initial isolated materials. This is particularly critical when either or both of the materials are magnetic, in which case it is called Spinterface.

In this context, some question arises. Which is the most adequate way of calculating a molecular surface on top of a metal? How to account on rearrangement of atomic positions, image-charge properties and orbital hybridization? and last but not least, How to model a descriptive molecular surface in which different molecules can be arranged in different positions?

• Salva, it would be nice if this question gets an answer, but one thing that might be making it hard is that there's at least 3 questions in one (there's 3 questions in the last paragraph). Might it be appropriate to focus on one question here, and then maybe ask a part 2 separately if necessary? This way if someone knows the answer to one of the 3 parts of the question, but not the other two, they might not feel discouraged to answer! – Nike Dattani Jun 15 '20 at 0:55
• The work function is surface-dependent, and sensitive to surface reconstructions and passivation, amongst other things. The Fermi level of the infinite bulk material would be a rather crude proxy for the workfunction. Also remember that the Fermi level as computed in most periodic DFT codes is calculated with respect to an arbitrary pseudopotential-dependent reference energy, not a "zero vacuum level", so you would need to correct for this. – Phil Hasnip Jul 5 '20 at 23:11
• @PhilHasnip This seems like it would the start of a great answer. – Tyberius Jul 14 '20 at 20:13
• Hi @PhilHasnip, we've dipped below 90% answered, I wonder if you could write that comment as an answer? – Nike Dattani Jul 28 '20 at 0:32

## 1 Answer

The work function is surface-dependent, and sensitive to surface reconstructions and passivation, amongst other things. The Fermi level of the infinite bulk material would be a rather crude proxy for the workfunction.

You also need to be aware that the Fermi level as computed in most periodic DFT codes is calculated with respect to an arbitrary pseudopotential-dependent reference energy, not a "zero vacuum level", so you would need to correct for this. Essentially, the DFT eigenenergies (and hence the Fermi energy) are computed with respect to the mean electrostatic potential in the simulation cell, and this includes a contribution from the local part of the pseudopotential. Adding a molecule to the surface in a periodic simulation is actually adding an infinite, periodic array of molecules and this changes the mean electrostatic potential (in addition to any actual physical or chemical phenomena).