Multiwfn forum

FanMo

Member

Registered: 2025-08-14

Posts: 2

Composite DFT Methods: (1) ECPs & Deep Contraction ; (2) 2025 Pople?

Hello! Here is my attempt at understanding.

Composite DFT methods involve many empirical corrections, like geometric counterpoise (gCP) and functional reparameterization. A recent champion is wB97X-3c, which unlike its predecessors avoids using gCP and "SRB" (short-range basis) corrections, opting merely for the D4 dispersion correction and a specialized basis set (vDZP) [1]. Considering the RSH-mGGA form of wB97X-V, the cost associated with the functional is higher compared to GGA-based composite methods (e.g., B97-3c, PBEh-3c, HSE-3c) [2]. So, Grimme's 2023 work focused heavily on developing their vDZP basis set, using many effective core potentials (ECPs) and relatively deep contraction (more Gaussian primitives) to minimize BSSE and cost [1].

1. It feels contradictory to use ECPs and deep contraction simultaneously?

Shortly after, someone compares vDZP against the target 6-311+G(3df,2p), using several different combinations of diffuse and polarization functions [3].

2. I feel there are much better comparisons to make? Though one can't simply modify "balanced" basis sets like Ahlrichs family (and indeed the original Grimme article already evaluates them as is), I do wonder.

Thank you!

[1] https://pubs.aip.org/aip/jcp/article/15 … hybrid-DFT
[2] https://corinwagen.github.io/public/blo … osite.html
[3] https://pubs.acs.org/doi/pdf/10.1021/ac … le_openPDF

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 2,045

Website

Re: Composite DFT Methods: (1) ECPs & Deep Contraction ; (2) 2025 Pople?

There is no necessary contradiction. For example, MOLOPT basis set in CP2K is fully generally contracted, while it is used in combination with GTH pseudopotential. I suggest checking original paper of MOLOPT: THE JOURNAL OF CHEMICAL PHYSICS, 127, 114105 (2007).

FanMo

Member

Registered: 2025-08-14

Posts: 2

Re: Composite DFT Methods: (1) ECPs & Deep Contraction ; (2) 2025 Pople?

I see, thank you!

I found combining these concepts unintuitive because an advantage of pseudopotentials is needing fewer primitive gaussians per atom—but there are other benefits! The question becomes what is the relative importance and difficulty of modulating the other factors influenced by the pseudopotential, but I leave it there.

Last edited by FanMo (2025-08-15 18:35:16)