(124a) Probing the Sensitivity of Point Charge Potentials in an Attempt to Improve Their Ability to Rank the Stability of Pharmaceutical Polymorphs

Temperature and entropic effects cannot be neglected when ranking the stability of organic polymorphs. Enormous effort has been put into structure prediction to seek out new crystalline phases for organic materials, with much of the stability ranking being an afterthought. Structure predictions have relied on lattice energy to determine the crystalline stability despite numerous experimental and computational studies that have shown that enantiotropic behavior is prevalent. In particular, one study has shown that 20% of a set of > 500 experimentally know polymorphs re-rank with temperature.

The bottleneck preventing free energy stability rankings in CSPs is largely due to the availability of accurate and efficient force fields. We have developed and tested molecular and lattice dynamic approaches to include temperature and entropic effects of enantiotropic polymorphs. For both approaches, we used point-charge potentials to compute the free energy of 12 systems. We found that the entropy differences using the point charge potential was the correct sign for 100% of the systems relative to experiment, but the enthalpic stability between polymorphs was only the correct sign for 58% of the systems. While more complex hamiltonians could be used to correct these issues the increase in computational expense becomes disfavorable, especially if added to a CSP.

We hypothesize that point charge potentials can be tuned with simple scaling factors improve the lattice energy, without negative implications on the entropic contributions. Initial results have shown that the entropic contributions to free energy stability are less sensitive to the potential used than the lattice energy. To better understand the tunability of point charge potentials we will look at how scaling the intermolecular interactions independently affect the enthalpy and entropy differences of the crystalline systems previously studied. The end goal would be to take the lattice energy rankings of a CSP performed with a highly level of electronic theory, tune the intermolecular interaction to minimize the error in the lattice energy stability, and then consider the effects of temperature and entropy. I propose to present the results tested to prove or disprove that hypothesis.