Since the discovery of the quantum nature of light and matter, researchers have devoted great effort to investigating complex dynamics phenomena arising from their interactions. Rational understanding of light-matter interactions has enabled the development of a wide range of technologies including LEDs and light-harvesting devices. It has also offered the opportunity to synthesise materials imbued with new properties, such as conducting electricity. Funded under the Marie Skłodowska-Curie Actions programme, the StrongLights project focused on furthering understanding of the effects of shooting a quantum material with ultrashort laser pulses at the near-infrared and ultraviolet. Electrons in quantum materials strongly interact with vibrations in the crystal lattice and cause photoinduced phase transitions. “Our main goal was to investigate the key microscopic factors that govern photoinduced phase transitions and controlling them by ultrastrong light pulses,” notes Angel Rubio, StrongLights coordinator.
Modelling the electronic structure of a photosensitive material
Researchers carried out first-principle calculations for periodic molecular systems. The analysis of electronic and structural changes in certain materials at the nanoscale almost reached the processing speeds of supercomputers. Special focus was placed on the (MeBr-DCNQI)2Cu molecular crystal that changes its phase when excited by light. This low-dimensional system exhibits a Peierls-type distortion at low temperatures, switching from an insulating phase to a metallic one. This observation served as a benchmark for comparing the crystal’s electronic band structure at high and low temperatures. The crystal also presented a unit cell three times larger than that at high temperatures.
Key factors that trigger photoinduced phase transitions
The team demonstrated that Coulomb repulsion forces play a significant role in properly characterising the crystal’s structural and magnetic properties in the low-temperature phase transition. “We have shown the need for a surprisingly high Hubbard (U) value for the conjugated molecular systems (or pi systems) to reproduce the experimentally observed charge separation phase at low temperatures,” notes Rubio. The value of the on-site Coulomb interaction U was determined from standard trial-and-error methods as well as from recently proposed first-principle calculations. Results also showed that a non-standard U value for the conjugated systems of the crystal’s organic molecules is necessary for localising lone pairs. “Although we did not calculate the specific electron-phonon couplings, we identified the geometrical changes involved in the crystal’s photoinduced phase transitions. This step is crucial for interpreting the vibrational modes that initiate phase transition when the system is heated or cooled,” adds Rubio.
Determining the electronic properties in excited states
Researchers extended their local density approximation methods to determine the crystal’s electronic excitation properties, such as the transition dipole moments and transition densities. The team proved their new scheme is crucial for determining exciton couplings, and therefore, excitation energy transfer dynamics that govern phase transitions from the insulator to the metal. The newly developed method has been implemented using the Octopus code. Researchers are working on machine learning techniques such as kernel ridge regression and deep neural networks to determine the molecular excitation spectra in an efficient and predictable manner. “Ultrastrong light pulses are powerful tools for modifying strongly correlated materials. We have taken a step forward by controlling not only the ground-state electronic properties but also the excited-state electronic dynamics of photosensitive crystals. By mixing different light pulses, we create new meta-stable states that demonstrate intriguing properties that have far-reaching implications for materials science,” concludes Rubio.
StrongLights, photoinduced phase transitions, metal, electronic properties, insulator, ultrastrong light pulses, excited state, machine learning