IR spectr of X-D2O: discrepancy from band analysis of stretches

[jadefox337]

Hello,

I am getting IR spectrum from system dipole DFT (BLYP) for cellobiose
+ 1
D2O.  For D I used same basis sets as for H but with double the
mass.   I
also compute specific band stretches frequencies for D2O (from FFT of
the
bond length along the trajectory).  These should match quite nicely,
but
in fact there is quite a large discrepancy - about 100 cm-1.
Specifically:
(a)   DO-free (band analysis) ~ 2730 cm-1  versus DO-free (dipole
spectrum) ~ 2640 cm-1
(b)  DO - h-bonded (band analysis) ~2440 cm-1 versus DO-h-bonded
(dipole
spectrum) ~ 2340cm-1

Does anyone have any explanation ?  Could there be some error / bug in
dipole computation or in the way I used the basis set for D ?

Thanks much for your help,