zero contraction coefficients in some basis sets?!

[mradon]

Dear All:

I see that in some basis sets there are single-primitive contractions with zero contraction coefficient!! For example, in my calculation using Fe def2-TZVP or Fe def2-TZVPP I see in the basis file:

| (...)
|    1  p
|  0.83380000000E-01   1.0000000000
|    1  d
|  0.91000000000E-01   0.0000000000 <------ NOTE !
|    1  f
| (...)

It seems that this zero coefficient comes from the "d.diff" function in the library. So, this is the relevant fragment of basen/fe file:

| *
| fe P(d.diff)
| # extrapolated from TZV FW03/05
| *
|    1  d
|    0.091
| *

It lacks any contraction coefficient (on the left of "0.091" exponent).

Am I right that this is the bug in the basis set library? So, should I correct my basis file manually (replace "0.00" with "1.00") before starting calculations?

best regards,
Mariusz Radon

[uwe]

Dear mradon,

this might look like a bug, but it is not. If the shell is not contracted, i.e. if you have just one exponent (1 s, 1 p, 1 d, etc.), the contraction coefficient is meaningless. Turbomole automatically normalizes the contraction coefficients to 1 - just the way it generally normalizes the contraction coefficients of the input basis sets.

In case of an uncontracted shell, the coefficient is not needed since it is one per definition. Hence:

  1  d
     0.091
  1  d
     0.091  0.00000
  1  d
     0.091  3.4994

is all the same.