Notes about Pople basis sets
============================

There are a few types - 3-21G, 4-31G, 5-21G, 5-31G, 6-21G, 6-31G. Of
these, only 3-21G and the 6-31G subfamily are regularly used.

5-21G is only included in the original BSE due to Gaussian using it for
Li and B for the 4-31G basis set. I believe they are rarely used.

For hydrogen and helium, there is no split core/valence. Therefore,
4-31G, 5-31G, and 6-31G are all equivalent for H and He. These basis
sets are stored in the 31G-* files.

For beryllium, the 5-21G/6-31G basis sets given in dill1975a were
superseded by ones given in binkley1977a. The original data from dill1975a
is not included in the new BSE.

For 6-31G, the basis sets for potassium and calcium were originally
given in rassolov1998a, along with d-type polarization functions with an
exponent of 0.2000.  Rassolov2001a updates these basis sets, adding extra
d functions to the valence. In doing so, the d-type polarization functions
were also updated, to be 1/4 the smallest valence d function.

As far, diffuse functions for K,Ca have not been published. Gaussian
uses the diffuse functions from blaudeau1997a. But those functions
are for a different kind of basis set. This is also
true for Sc-Zn. Gaussian is also missing the updated data from Rassolov2001a for
K,Ca. In light of this inconsistency, I have removed K-Zn from v1.

There does not appear to be diffuse functions available for helium. The data
here is taken from Gaussian.

For Na-Ar (Z=11-18), 6-31G uses the same core as 6-21G (gordon1982a)
with a different valence (francl1982a).

For Li-Ne (Z=3-10), 3-21G uses the same valence as 6-21G (binkley1980a).

6-31G for helium/neon do not appear to be published. For v1, I have taken the data
from Gaussian 09.

In Gaussian, the added polarization function for He is the same as for H.
Neon uses the same as for C-F. The polarization functions for Li-B are of unknown
origin, but appear to be evenly-spaced.


Data sources
------------

v1 of the data comes from Gaussian09 and GAMESS. Both generally match the old BSE,
however it looks like GAUSSIAN as re-normalized contraction coefficients, so they
have more decimal places.

I have chosen GAMESS for the data for Ga-Kr, due to Gaussian using a different
basis set for Ga-Kr. See below for the reason.


Differences with Gaussian
-------------------------

Gaussian does not included the added d shell for K,Ca found in rassolov2001. I have added them
directly from the paper. The polarization functions for K,Ca that are in gaussian are of unknown
origin.

Gaussian choses a different basis set for Ga-Kr. They have chosen the basis sets
found in blaudeau1997a. This basis is not consistent with the typical terminology,
and therefore moved to 6-31G(C), as recommended in rassolov2001a.

The original BSE had the data for Ga-Kr, but only for the 6-31G*. The basis sets there
are consistent with my interpretation, and therefore differed from Gaussian.


3-21G basis
-----------

3-21G was developed later, and uses pieces of the Huzinaga MINI basis set for the
core and some valence.


Differences/Issues with the original BSE
----------------------------------------

The original BSE sometimes truncates to a fixed number of decimal places, removing
significant figures.

In the original BSE, the H basis for 6-31G and 4-31G differ slightly,
possibly due to transcription error. The 6-31G basis set for H in
GAMESS and Gaussion both agree with the basis set given in 4-31G in the
original BSE.

The original BSE was missing the modified potassium and calcium given in rassolov2001a.


-------------------------------------------------
 REFERENCES MENTIONED ABOVE
 (not necessarily references for the basis sets)
-------------------------------------------------
binkley1977a
    Binkley, J. Stephen, Pople, John A.
    Self-consistent molecular orbital methods. XIX. Split-valence
            Gaussian-type basis sets for beryllium
    J. Chem. Phys. 66, 879-880 (1977)
    10.1063/1.433929

binkley1980a
    Binkley, J. Stephen, Pople, John A., Hehre, Warren J.
    Self-consistent molecular orbital methods. 21. Small split-valence
            basis sets for first-row elements
    J. Am. Chem. Soc. 102, 939-947 (1980)
    10.1021/ja00523a008

blaudeau1997a
    Blaudeau, Jean-Philippe, McGrath, Mark P., Curtiss, Larry A., Radom,
            Leo
    Extension of Gaussian-2 (G2) theory to molecules containing third-row
            atoms K and Ca
    J. Chem. Phys. 107, 5016-5021 (1997)
    10.1063/1.474865

dill1975a
    Dill, James D., Pople, John A.
    Self-consistent molecular orbital methods. XV. Extended Gaussian-type
            basis sets for lithium, beryllium, and boron
    J. Chem. Phys. 62, 2921-2923 (1975)
    10.1063/1.430801

francl1982a
    Francl, Michelle M., Pietro, William J., Hehre, Warren J., Binkley, J.
            Stephen, Gordon, Mark S., DeFrees, Douglas J., Pople, John A.
    Self-consistent molecular orbital methods. XXIII. A polarization-type
            basis set for second-row elements
    J. Chem. Phys. 77, 3654-3665 (1982)
    10.1063/1.444267

gordon1982a
    Gordon, Mark S., Binkley, J. Stephen, Pople, John A., Pietro, William
            J., Hehre, Warren J.
    Self-consistent molecular-orbital methods. 22. Small split-valence
            basis sets for second-row elements
    J. Am. Chem. Soc. 104, 2797-2803 (1982)
    10.1021/ja00374a017

rassolov1998a
    Rassolov, Vitaly A., Pople, John A., Ratner, Mark A., Windus, Theresa
            L.
    6-31G* basis set for atoms K through Zn
    J. Chem. Phys. 109, 1223-1229 (1998)
    10.1063/1.476673

rassolov2001a
    Rassolov, Vitaly A., Ratner, Mark A., Pople, John A., Redfern, Paul
            C., Curtiss, Larry A.
    6-31G* basis set for third-row atoms
    J. Comput. Chem. 22, 976-984 (2001)
    10.1002/jcc.1058.abs