Non-homogenous matter under ionizing radiations, a challenge for fisrt principles modelling
Paris, Ecole Normale Supérieure or Université Paris Sorbone
The workshop will be centered on questions concerning the consequence of interaction of matter with ionizing radiations, on the mechanisms of energy deposition, relaxation and dissipation into the environment and on the subsequent ultrafast chemical physics taking place on the subnanosecond time scales. The workshop will provide a forum for exchanging ideas between experimental and theoretical/computational approaches. We will especially focus in identifying the current methodological locks toward the realization of realistic simulations of nonhomogeneous matter under ionizing radiations.
Radiation chemistry is defined as "the part of chemistry which deals with the chemical effects of ionizing radiation (IoR), as distinguished from photochemistry associated with visible and ultraviolet electromagnetic radiation." (cf. IUPAC). IoR may either be high energy photons (X-rays, y-rays) or high kinetic energy massive particles (protons, electron, positrons, alpha particles, muons or heavy atom nuclei...). Understanding the consequences of submission of matter to IoR is of paramount importance in several areas. In medicine for example IoR exhibit a dual-face. On one side they are responsible for cellular accelerated ageing and are responsible for the development of cancers because they cause deleterious chemical damages (lesions) to DNA and proteins. On the other hand they offer very promising avenues for XXIst century radiotherapies that will permit exquisitely localized energy delivering to tumorous cells.
In the field of astrochemistry, the IoR that compose the cosmic ray spectra (mainly 1 MeV protons, but bare ions are also present) are considered to shape the chemistry taking place in the interstellar medium. Upon interaction with isolated (gas phase) molecules IoR open cationic reactive channels that must be incorporated into chemical models. Recently experimental evidences have shown low energy electrons emitted upon ionization of icy grains found near star-forming regions of the Universe, may drive the chemistry taking on these grains, eventually leading to organic, prebiotic molecules. As the bombardment of the early Earth by comets that have captured ices grains is thought to have provided the molecular precursors at the origin of life on Earth, investigating radiolysis processes occurring on these grains appear as of tremendous interest.
Nuclear industry is also much concerned by the effects of ionizing radiations; not only in case of nuclear disasters as the one that took place in Fukushima , but more commonly to understand the effect of radiolysis of effluents and for dealing with radioactive wastes.
For years the early stages of matter submitted to such extreme irradiations have escaped a fine understanding. The situation is now changing. The availability of picosecond pulse-radiolysis systems in Europe, Japan, or North-America are bringing unanticipated insights on the chemical reactivity of unstable species like the radical cation of water (H2O°+) or of presolvated electrons with DNA bases. Clever experimental set-ups have been devised to investigate the reactivity of low energy electrons on DNA or models of icy grains. The discovery of High Harmonic Generation (HHG) or XFEL (X-Free Electron Lasers) sources that have opened the possibility to probe the atto- and femto-seconds time-domains with photons. Spectacular attosecond pump–probe experiments performed in small molecules have for instance allowed tracking charge dynamics in the natural time scale of electron motion.
In these research fields, theoretical and numerical approaches have indubitably a central role to play either to establish theories or to simulate the processes of interest, providing mechanistic insights that are not always accessible by experimental means. The simulations of ultrafast time scales pose serious difficulties though. Difficulties arise from the diversity of particles involved (electrons, nuclei, photons, other types of particles…) and from complexity of the physical systems of interest which often display non-periodic or non-homogenous topologies. In most cases system receives a perturbation and one is interested in the relaxations processes. The latter involves continuous flows of energy and information among the electronic and nuclear degrees of freedom that must be simulated realistically. Simulating this out-of-equilibrium dynamics represents a particularly difficults but stimulating challenge.
Brute force simulations of the electronic and nuclear degrees of freedom at the quantum level are still out of reach except for rather small systems. One usually adopt mixed quantum-classical schemes where a quantum treatment is retained only for the electrons, the time evolution of which are simulated by propagating the time-dependent Schrodinger equation or, in the framework of Density Functional Theory, by the time-dependent Kohn-Sham equations (TD-DFT). The nuclear motion are simulated by Newton's law. A difficulty arises to couple both types of motions and to define classical forces acting on the nuclei. Various alternatives like Ehrenfest, surface hopping, or the most recent exact factorization are possible mixed quantum-classical algorithms having their respective pros and cons.
TD-DFT has attracted a lot of interest for its potential ability to simulate large molecular system, at the expense on needed approximation in the time-dependent exchange-correlation potential. Many groups reported implementation of so-called Real-Time-TDDFT approach[12,13], eventually coupled to Ehrenfest MD. Some implementations are based on space grids (REF, Toulouse, e.g. Octopus 10.1002/pssb.200642067) while others rely on basis sets to expand the Kohn Sham orbitals[15,16]. These different methodologies have their own advantages regarding the computational efficiency of the algorithms (e.g., efficient parallelization; Fast Fourier transforms or density fitting/Resolution of the Identity to compute Coulomb interaction…).
The simulation of the interaction between the electron cloud and swift ions with kinetic energies in the keV-MeV range have been reported for small gas phase molecules by various research groups[17,18]. A recent publication reporting the irradiation of a DNA decamer represents a real tour de force. Such atomistic simulations give access the stopping power projectile from first principle. They further give access to valuable mechanistic details regarding the ionization process per se, the localization of secondary electrons, the mechanisms of energy relaxation and dissipation into the nuclear modes on the atto- to femtosecond time scales. RT-TDDFT based methodologies have been used to simulate Coulomb explosion of solvated molecules of biological interest such as uracil or desoxyribose .
As an alternative to RT-TDDFT simulations like the non-adiabatic Real-Time Nonequilibrium Green’s Functions propagations have been proposed and tested to simulate ultrafast charge migration of ionized biomolecules. Methodologies pertaining to wavefunctions methods like TD-Configuration interactions have been proposed.
Despite important progresses in simulation algorithms of ultrafast dynamics some pending challenges remain. They are expected to be discussed during the meeting.
- Adiabatic approximation and time-dependent current DFT
- Self-Interaction-Error (SIE) and various fixes proposed so far (average self-interaction-correction, inclusion of exact exchange, Hubbard corrections)
- Asymptotic behavior of exchange contributions.
- Alternative to DFT methodologies (wave function methods, or green function methods)
- Tools to extract physical and chemical insights from ultrafast electron-nuclear dynamics (populations analyses, time-dependent topological analyses, reconstruction of electronic spectra…)
- Environment effects in case of highly inhomogeneous systems (hybrid QM/MM or QM/QM')
- Coupling between electron and nuclear dynamics (over-coherences, demixing within mean-field schemes,).
- Efficiency of implementations for HP architectures (scalability, limitations in terms of system sizes…)
Mai Dinh (LPT, Toulouse) - Organiser
Marie-Anne Hervé du Penhoat (IMPMC, Université Pierre et Marie Curie-Paris6) - Organiser
Rodolphe Vuilleumier (Sorbonne Université - ENS-PSL) - Organiser
Aurélien de la Lande (CNRS-University of Paris-Sud, Orsay ) - Organiser