CCL: Double slash notation in quantum chemistry



A single point energy calculation was done with method mA and basis bA, on geometry that was optimized with method mB and basis bB:

mA/bA//mB/bB


Marcel Swart FRSC FYAE MAE, Prof. Dr.
ICREA Research Professor at University of Girona
Director of Institut de Química Computacional i Catàlisi

Univ. Girona, Campus Montilivi (Ciències)
c/ Maria Aurèlia Capmany i Farnés, 69


vCard


On 23 Feb 2022, at 21:57, Andrew DeYoung andrewdaviddeyoung()gmail.com <owner-chemistry%%ccl.net> wrote:


Hi,

Could you please help me to understand the "double slash" notation that I see in quantum chemical literature?  The notation is: 

method/basis set//method/basis set

What is the convention -- if any -- for the order of the calculations, (a) or (b)?  

(a) Is the calculation before the "//" performed first, while the calculation after the "//" is performed second?  In other words, "do this [left], then that [right]"?

(b) Or is it the other way around, meaning that the left-hand calculation is performed ON the output of the right-hand calculation?  In other words, somewhat like an operator acting on a wave function; "do this [left] on the output of that [right]."

Below are two examples from the literature.  I try to reason through the notation based on these examples:

(1) Maxwell, Tirado-Rives, and Jorgensen. J. Comput. Chem. 1995, 16, 984-1010 ( https://doi.org/10.1002/jcc.540160807 ).  In this paper, ab initio calculations were performed to determine torsional parameters for organic molecules and ions.  From the abstract, "The rotational energy profiles were obtained at the HF/6-31G*//HF/6-31G*."  In certain cases, they also used MP2/6-31G*//RHF/6-31G* or MP2/6-31G*//MP2/6-31G*.  I think this means an initial, unconstrained geometry optimization of the entire molecule/ion was performed first, followed by a geometry optimization at each selected, constrained value of the torsional angle.  In the case of MP2//RHF, then, is MP2 used for the initial optimization and RHF for the scan?  Because RHF is less expensive, I think, that would probably make sense.  Based on this, I would say that MP2//RHF means that MP2 is performed first and RHF is performed second; in other words, the "double slash" notation is read from left to right.  So interpretation (a) above is correct.

(2) Liptak and Shields, J. Am. Chem. Soc. 2001, 123, 7314-7319 ( https://doi.org/10.1021/ja010534f ).  In this paper, the free energy of solvation, DeltaG_s, is calculated using an implementation of the polarizable conductor model (CPCM) -- see the third page (p. 7316) of this paper.  Near the top left-hand side of that page, it says, "The CPCM calculations were performed as SPCs (single-point calculations) using the 6-31G(d) and 6-31+G(d) basis sets on the HF/6-31G(d) and HF/6-31+G(d) geometries for each of the six systems."  Based on that description, the HF geometry optimization was performed first and the CPCM single-point calculation was performed second.  So, if my reasoning from example (1) above is correct (i.e., interpretation (a) above), I'd expect the "double slash" notation in the paper to be HF//CPCM.  But, in fact, it's not.  In the Results section near the bottom of the left-hand side of the third page (p. 7316), the authors state, "CPCM/6-31G(d)//HF/6-31G(d), CPCM/6-31+G(d)//HF/6-31G(d), and CPCM/6-31+G(d)//HF/6-31+G(d) are denoted S1, S2, and S3 [in Table 1]."  This, I think, implies that interpretation (b) above is correct.  The strange thing, though, is that at the bottom of Table 1 on the fourth page [p. 7317], the definitions of S1, S2, and S3 are different from those stated in the text, by an "extra" "HF" to the left of the "//"; there, they state, "S1: CPCM/HF/6-31G(d)//HF/6-31G(d). S2: CPCM/HF/6-31+G(d)//HF/6-31G(d). S3: CPCM/HF/6-31+G(d)//HF/6-31+G(d)."

It's entirely possible, then, that there is no set convention for "double slash notation" in quantum chemistry (or that I'm misinterpreting these papers).  Could anyone shed some light on this, since I'm not a quantum chemist?

Thanks so much for any insight you can provide,
Andrew

Andrew DeYoung, PhD
Carnegie Mellon University