# [molpro-user] normal mode

Eryin Feng fengbf at mail.ahnu.edu.cn
Tue Jul 26 09:34:55 BST 2011

```Dear users,

In my work, I try to get the normal mode coordinates of a molecule by molpro, but when i compare the result get from molpro with from  Gaussian98 code, I find the different results !

For example:  O--C--S  (linear molecule),  V3 antisymmetry normal mode (Q3) :

in Gaussian:  Q3:  -0.50 Oz + 0.86 Cz - 0.07 Sz      ( molecule located at Z axis)

in molpro:    Q3:  -0.14098 Oz + 0.2364 Cz -0.01822 Sz

To switch from Molpro  normal coordinate coefficient to gaussian one,  two molpro-users told  me two different methods as:

(1)   calculate the reduced mass (mured) corresponding to the normal coordinate of interest.
mured= 1/sqrt( sum over all cartesian displacements of the square of the molpro coefficient divided by the corresponding atomic mass )
Here mured =1/sqrt [ (0.14098)^2/15.99 + (0.2364)^2/12 + (0.01822)^2/31.97]= 13.006
multiplying the molpro coeff. by 1 / sqrt(mured) gives the corresponding gaussian normal coordinate coefficient

This method gives:
reduced mass: 13.008
Q3:  -5.084378723760641E-001 Oz  +    8.526378218576923E-001   Cz  -    6.570958769342700E-002   Sz

(2)  The reduced mass can be calculated from 1/(sum-of-squares of molpro  normal coordinate coefficient ):
mured= 1/( 0.14098^2 + 0.2364^2 + 0.01822^2 )= 13.14030
multiplying the molpro coeff. by 1 / sqrt(mured) gives the corresponding gaussian normal coordinate coefficient

This method gives:
reduced mass:  13.1403
Q3:  -5.110462461425969E-001 Oz  +      8.570120005877189E-001   Cz  -     6.604668918421140E-002    Sz

The difference between these two  method  seems to origin  from the different cognizance that the use or not of mass-weighted coordinates for Molpro normal coordinate coefficient . What is the true case in Molpro? (Which method is right ?)   Looking forward to your suggestions!