 |
Center for Nonlinear Studies |
Discovery of generation of very-high-order harmonics from semiconductors with band gap of 2.5 eV and larger [1] followed by studies of bias-free ultrafast photocurrent generation [2] have stimulated intensive research in the field of semiconductor interactions with high-intensity infrared few-cycle laser pulses. Nonlinear absorption, associated electron excitations, and laser-driven ultrafast electron dynamics are the key processes of ultrashort high-intensity laser-solid interactions that serve as a basis for those applications. Simulation of those processes meets a few fundamental challenges. Advanced numerical methods, e. g., ab initio simulations or direct numerical solving of the Schrodinger equation either suffer from technical issues, e. g., limited domain of simulation time and total neglecting by electron-phonon collisions. Simplified semi-analytical approaches based on a cycle-averaged rate equation are among the most popular simulation approaches, but they also suffer from inadequate assumptions and internal contradictions. Many issues of the simplified models result from a priori assumptions about a dominating type of laser-driven electron dynamics [3]. In absolute majority of simulations, high-rate collision-driven dynamics is assumed to dominate over non-collision coherent electron dynamics, i. e., laser-driven oscillations. Under that assumption, the rate-equation models properly fit experimental data at values of electron-particle collision time about 1 fs. This fact is considered as a justification for the high-collision-rate approximation and use of corresponding models, e. g., the Drude model [4]. However, that collision time is significantly smaller than duration of a single optical cycle at infrared laser wavelengths and prevents use of the non-collision photoionization models, e. g., the Keldysh formula. Moreover, it is smaller than the fundamental lower limit of electron-particle collision time in a crystal [4].
We discuss a model of the ultrafast laser-driven inter-band and intra-band electron excitations [3,4] that is non-perturbative and is valid at high intensity; does not use a quasi-monochromatic approximations for laser radiation and can treat broad-band ultrashort laser pulses; does not a priori assume high electron-particle collision rate; and is simple enough to replace the usual rate-equation models. It is built by the assumption about domination of the laser-driven electron oscillations over the electron-particle collisions. The latter means the collision time must be appreciably larger than single laser cycle. This assumption is justified by numerous experimental data [4] to be overviewed in this talk. Correspondingly, the effects attributed to cycle-averaged energy of the oscillations (ponderomotive energy) are included into this approach by analytical relations, and time-dependent transient energy bands are introduced [3]. The Drude equation for the intra-band absorption rate is replaced by non-perturbative Vinogradov equation [4]. The physics of this new approach and novel predictions by this model, e. g., cycle-averaged bias-free photocurrent [3] and transient inversion of population of conduction band [4] are extensively discussed in this talk. Obtained results are compared with the regular rate-equation models [4] and publications of other researchers [2]. This material is based upon work supported by the Air Force Office of Scientific Research under award numbers FA9550-16-1-0069 and FA9550-15-1-0254.
REFERENCES:[1] S. Ghimire, A. D. DiChiara, et al., Nature Physics 7, 138 (2011).[2] T. Paasch-Colberg, A. Schiffrin, et al., Nature Photonics 8, 214-218 (2014).[3] V. Gruzdev and O. Sergaeva, Phys. Rev. B 98 (11) 115202 (2018).[4] O. Sergaeva, V. Gruzdev, et al, J. Opt. Soc. Am. B 35 (11), 2895-2905 (2018).
Host: Andrei Piryatinski