The use of DFT calculation for salt-cocrystal distinguishing

M. Hušák1, E. Skořepová2, J. Rohlíček1

1Institute of Chemical Technology Prague, Technická 5, 166 28 Prague 6

2Institute of Physics CAS, Cukrovarnická 10/112,162 00 Prague 6

husakm@vscht.cz

To identify the hydrogen position is essential for distinguishing between salt and cocrystal. The difference between salt and cocrystal in pharmaceutical industry can have essential influence on legal and registration process. Unfortunately it is often hard to identify the precise hydrogen position when we have low-quality single crystal data or the structure was solved from powder. The other methods suitable for cocrystal/salt distinguishing as ΔpKa rule or ss-NMR are not always 100% reliable.

We had tested on 5 structures a method for hydrogen position verification based on DFT-d calculation. The idea is to place hydrogen atom in wrong position (close to acid oxygen for salt, close to acceptor oxygen for cocrystal) and to see whatever the result of geometry-optimization will shift the hydrogen somewhere. For the DFT-d calculation we had used following settings: CASTEP software, PBE functional with Tkatchenko-Scheffler dispersion correction, 520 eV energy cuf-off. The other setting had corresponded to the "Fine" CASTEP pre-set profile. The cell parameters were fixed during the geometrical optimization. The result are summarised in Tab. 1.

Table 1. The Oacid-H-O distances in all structures. Underlined are intentionally wrong artificial values.

 

Structure name

Single crystal data

DFT start-up data

DFT final results

 

Oacid-H

H-O

Oacid-H

H-O

Oacid-H

H-O

Agomelatine citric acid cocrystal

0.889

1.572

1.487

0.950

1.202

1.220

Agomelatine maleic acid cocrystal

0.871

1.652

1.502

0.950

1.103

1.358

Agomelatine bromide salt

1.635

0.841

0.950

1.497

1.375

1.087

Agomelatine phosphate LT salt

1.613

0.838

0.950

1.437

1.372

1.098

Agomelatine phosphate HT cocrystal

XRPD solution

XRPD solution

1.433

0.845

1.118

1.317

 

The result from Tab. 1. confirms the correct cocrystal/salt assignment for agomelatine maleic acid cocrystal, agomelatine bromide salt and agomelatine phosphate low-temperature salt. The high-temperature agomelatine phosphate should be to a big surprise a cocrystal. The single crystal data for agomelatine citric acid phase should be re-evaluated, because the results indicate a phase half/way between salt and cocrystal. Generally the whole method should be tested on more structures because it can probably detect interesting issues with data interpretation.