The ABINIT Users Mailing List ( CLOSED )

[Xavier Gonze <xavier.gonze@uclouvain.be>]

Dear Walter,

On 22 Jul 2008, at 21:18, Walter Lambrecht wrote:

> In Veithen et al. PRB 71, 125107 (2005) it is described how to  
> calculate the
> linear electro-optic
> tensor in the unclamped condition, i.e. including the piezoelectric
> contribution under zero external
> stress. From eq. A15, it appears that what is needed is the Raman  
> tensor, the
> force in response
> to strain internal strain tensor for the second term and the  
> derivatives of the
> high frequency dielectric  tensor  vs. strain components in the  
> first term.
> At present it appears like I would be able to extract all the  
> necessary
> pieces from the anaddb results for the second term but not the  
> first term. Are
> the
> third derivatives of the type: two vs. electric field and one vs.  
> strain being
> calculated by abinit
> and where can I find them in the DDB files?

Unfortunately, this third-order derivative of the energy has not been  
coded (yet ...).
Marek performed that computation by finite difference of dielectric  
tensor for different
primitive cells.

> or what variable needs to be set in
> the input
> to generate these? Or better, has this procedure been implemented  
> in anadbb?
> Alternatively is the anadbb capable of generating the elasto-optic  
> tensor
> directly?
> Sincerely,
> Walter Lambrecht

Best regards,
Xavier