Good day everyone,
I m confused with the reliability of my output file. For my system, I have used using 6-31G+(d) and 6-31G+ basis set and now which out put is more reliable and fits for DFT studies? Or 6-31G+(p) is more reliable? ( I also did run my calculation in 6-31G(d,p) but it is not supporting for further procedures)
Dear Rinaa
In principle, greater basis sets should yield better and therefore more reliable results as they become closer to the infinite basis set limit. However due to computational cost we are often restricted to the use of "small" basis sets.
In your case you have to consider the usefulness of each basis function and if it is worth including it for your systems:
-The "6-31G" is the standard, split-valence double-zeta basis set, the functions you will use to describe your core and valence orbitals.
-Functions in parentheses "(d)" or "(d,p)" are polarization functions, those used to properly describe chemical bonds (higher angular momentum than the valence orbitals). If there is no comma, such as "(d)" or it means that you are including polarization functions in all "heavy" atoms (all bar hydrogen); if there is a comma, like "(d,p)" it means that you are including the first functions in heavy atoms and the second ones in H. Polarization functions are practically always useful
-Functions identified with "+" are diffuse functions, useful to describe anions and/or long-range interactions (dispersion, hydrogen bonds...). "+" means diffuse functions for heavy atoms, "++" means diffuse functions for all atoms.
Now, if your system is not an anion or it does not contain critical long-range interactions, use the 6-31G(d) if you cannot afford the 6-31G(d,p). If you have any of those properties, include diffuse functions. However, there are some considerations:
1) It is somehow strange including diffuse functions without including polarization ones, like "6-31+G", because your basis set will overdescribe long-range interactions while forgoing short-range interactions (i.e. chemical bonds). At least use the 6-31+G(d)
2) There is no practical reason to include additional functions only on hydrogen atoms, like "6-31G(p)"
Hope you find it helpful
Hi Rinaa,
The choice of basis set depends on the system you ara working with (size, composition ...) and properties of interest. I suggest trying a bunch of these ones suggested, comparing it with experimental data (if available) and see which ones give you the best results.
For DFT or any ab initio studies, you should use a basis set as large as possible. since you have not specified your system, so it is difficult to say what basis set or functional is useful for you.
the notation + or ++ represent diffuse functions on heavy elements (i. e. other then hydrogen) and on hydrogen and heavy atoms. They are useful when you have a system with large electron density like halogen ions/electronegative elements or radicals. (d, p) represent polarization functions and they must be the part of all the basis set as they provides flexibility to otherwise regid basis set. I mean if you do not use polarization functions, size of your molecular orbital will remain fixed and this will not represent a real system.
My suggestion is to use at least 6-311++G(d,p) if your system has less then 100 atoms if your system consist of elements from period 1 to 3 of periodic table. for such system 6-31++G(d,p) will not produce good enough data to publish in a research paper (if you are using DFT with traditional functions like B3LYP, wB97XD etc).
If your computing anions, the use of diffuse functions (+) is mandatory. If not, the first extension of the basis set from 6-31G should be including polarization functions for all non-hydrogen atoms: 6-31G(d).
It is difficult to give some advice when you did not specify, what kind of results are you trying to get with the DFT. Generally, larger basis set of the series should be more reliable, which however doesn't have to mean that its results are closer to reality in each specific case.
As far as I know, in some cases - e.g. energy calculations, Dunning basis sets (e.g. cc-pVDZ, which is comparable to your basis sets) are performing better compared to Pople basis sets, even when having less functions.
Also, selection of functional and empirical dispersion correction probably have greater effect on your results than the small variations in basis sets you mention.
any one suggest I have 42 atoms but it takes more running gaussian software. Which basis set is good for optimizing structure I used 6-311+G(d,p).
Krishnamoorthy, that depends on the type of atoms in your structure, and which type of calculations you are running.
Krishnamoorthy gurushankar Sankar If you're really concern about fast & perfect approach, I would suggest you to consider B97-3C with def2-mSVP. But as far as I know it is available for only ORCA and Turbomole.
I need that software. Is it possible kindly give me link
Krishnamoorthy gurushankar Sankar
https://orcaforum.kofo.mpg.de/app.php/portal
You will need to register (for free) to download.
- Badges
- Science topic
- Similar topics
- Filing