CCL:G: B3LYP and def2-TZVP

 Sent to CCL by: Uwe Huniar [huniar..cosmologic.de]
 Hi John,
partially correct. def2-TZVP is identical to def-TZVPP for some elements, but not for all. So generally it is not the same to use def-TZVPP or def2-TZVP.
The problem with the Turbomole (or Ahlrichs or Karlsruhe) basis sets is often, that people get confused with the names. What is def- or def2- standing for? Is TZVP the same as def-TZVP, and what about DZP, TZP, etc... I will try to answer some of the questions below. 
 Anyone who is not interested in those basis sets can stop reading here.
Alfredo: at the end of this post there is a link to the Turbomole basis set library download site. If you do not find def2-TZVP at PNL basis set library, you can download the basis set in different formats from the link given below. 
 ---
First, just forget about the old TZP, DZP, etc. basis sets (those which do not contain a V for valence in the name) which have been published before 1990, they are hardly used any more.
SVP stems from 1992, TZVP from 1994, QZVP from 2003 (references and details see below the link to the Turbomole basis set library website, there click on the big green [Readme] button).
To avoid confusion with the previously published basis sets with similar names, the Ahlrichs group called them def- which stands for 'default' about 15 years ago. So def-SV(P), def-TZVP, def-QZVP, etc. are the basis sets which have been used as default in Turbomole. Nowadays we recommend to use def2- instead of def- for reasons described below.
Perhaps those few lines from the Turbomole documentation can help to get an overview: 
 - def-SV(P) for routine SCF or DFT.
       Quality is about 6–31G*.
 - def-TZVP for accurate SCF or DFT.
       Quality is slightly better than 6–311G**.
 - def-TZVPP for MP2 or close to basis set limit SCF or DFT.
       Comparable to 6–311G(2df).
 - def-QZVP and def-QZVPP for highly correlated treatments;
        quadruple zeta + 3d2f1g or 4d2f1g (beyond Ne), 3p2d1f for H.
With the quadruple zeta basis sets from 2003 Weigend et al. have tested the accuracy of def-SVP and def-TZVP in comparison to def-QZVP. It turned out that for many elements the accuracy of TZVP and SVP basis sets have not been consistent, partly just additional polarization functions have been missing (this is what John was referring to).
In 2005 the paper with the new def2- basis sets have been published: "These basis sets are designed to give similar errors all accross the periodic table for a given basis set type." (F. Weigend, basis set download site).
For organic molecules and properties like energy or structures the difference is not really noticeable. For heavier elements or properties like NMR chemical shieldings def2-TZVP is highly recommended. 
 Here the promised link:
 http://bases.turbo-forum.com/TURBOMOLE_BASISSET_LIBRARY/tbl.html
 Supported formats: Turbomole, Gaussian, Dalton, Molpro
 Sorry for the lengthy post on CCL!
 Regards,
 Uwe
 Am 28.04.2010 21:57, schrieb John McKelvey jmmckel^^^gmail.com:
 
 Sent to CCL by: John McKelvey [jmmckel++gmail.com]
 Partial answer:  In the TURBOMOLE basis set directory it is pointed
 out that def2-TZVP basis is the same as TZVPP.  If that is not
 available in Gaussian it might be possible to get it from the PNL web
 site.
 John McKelvey
 On Wed, Apr 28, 2010 at 2:07 PM, Alfredo Tlahuice tlahuicef()
 yahoo.com<owner-chemistry:+:ccl.net>  wrote:
 
 Sent to CCL by: "Alfredo  Tlahuice" [tlahuicef===yahoo.com]
 Hi there
 I want to start a gaussian 03 calculation using the def2-TZVP basis.
 So my question is how can I define that basis set in G03?
 Thanks in advance.>
 
 
 --
 ------------------------------------------------------------------------------
 Dr. Uwe Huniar
 COSMOlogic GmbH & Co. KG