Abstract
Lack of memory (locality in time) is a major limitation of almost all present time-dependent density-functional approximations. By using semiclassical dynamics to compute correlation effects in the time propagation of the density matrix, we incorporate memory, including initial-state dependence, as well as changing occupation numbers, and predict more observables in strong-field applications.
Original language | English (US) |
---|---|
Article number | 113002 |
Journal | Physical review letters |
Volume | 105 |
Issue number | 11 |
DOIs | |
State | Published - Sep 7 2010 |
Externally published | Yes |
All Science Journal Classification (ASJC) codes
- General Physics and Astronomy