New lithium salts on the computer: fiction or fact?

The need for improved lithium salts for 'traditional' neat salt/polymer pol ymer electrolyte systems is still urgent. However, the development of new s alts is time-consuming and costly and we raise the question if it is possib le to find new lithium salts in an intelligent way and if their properties can be predicted a priori to some accuracy? We here present the use of ab i nitio quantum mechanical calculations as a primary selection and search too l in the quest for new lithium salts. To evaluate the validity of this appr oach we present results obtained for 'traditional' polymer electrolyte salt s like LiBF4, LiClO4, LiTFSI, etc. These results are then used as a referen ce system and compared to experimental studies with a focus on vibrational spectroscopy to elucidate the anion dependence of the local coordination st rength. Along the same lines the successes and failures of some more recent ly employed lithium salts are investigated and explained with our method. F inally, we also challenge the polymer electrolyte community to actually syn thesize and lest some new computer constructed salts that, judging from the current approach and results, seem to be worthwhile alternatives for neat salt/polymer systems. The a priori calculated effects on different substitu tions possible for a given basic system are exemplified. (C) 2001 Elsevier Science Ltd. All rights reserved.