Dear Dalton team,
I'm currently doing calculations of TPCD rotatory strengths using Dalton 2020. I did my calculations using the following keywords:
**INTEGRAL
.DIPVEL
.ROTSTR
.ANGMOM
**RESPONSE
*QUADRATIC
.SINGLE RESIDUE
.TPCD
.ROOTS
6
and tried to compute the rotatory strength according to the formulas given e.g. in Friese, Hättig & Rizzo, PCCP 2016, 18, 13683, eq. (2)-(5).
Specifically, I took the relevant components of the P, M, and T tensors out of the output and used the formula given as eq. (19) in the paper, assuming that the quantities given as (DIPVEL ANGMOM) etc. are exactly the numbers provided in the output. For the calculation of the elements of T, I used eq. (8) of the paper, interpreting the whole expression "1/(2\omega^2) \sum_P \sum_n (T^+*{ac}^{0n} \mu_d^{p,0f})/(\omega-\omega*{0n}) " as the quantity provided in the output as, e.g., <<XZROTSTR; YDIPVEL

Is this correct? I'm unsure because test calculations on several molecules (H2O2, Tyrosine, Tryptophan) provided TPCD spectra that looked qualitatively comparable to those found in the literature, but when looking at specific numbers for the B1,B2,B3 coefficients of eq. (2), or directly at the tensor elements for P, M, and T, there tend to be large discrepancies (different signs, deviations up to a factor of 10).

I'll be very thankful to a clarification of how the TPCD rotatory strengths are to be computed from the output and a hint on the question if the results can be very sensitive on, e.g., small changes of the molecular geometry.

Thank you very much in advance,

yours,

Jens Petersen