The interaction of ionizing radiation with molecules often triggers internal electronic rearrangements governed by correlated processes such as shake-up, Auger decay, and Interatomic Collision Decay (ICD). Attosecond charge migration, the rapid movement of electron density along a molecular backbone, has been predicted as a consequence of these many-body interactions, occurring while the nuclei remain essentially stationary. Advances in ultrafast XUV and NIR sources have enabled the study of electronic migration, with evidence of few-femtosecond dynamics in aromatic amino acids. However, direct observation of the correlation-driven charge migration process, where the energy flows from a single excited electron to the coupled electrons in the system, remained elusive. This process is expected to exist only until nuclear motion begins, typically within 10 fs of ionization. Therefore, capturing this ultrafast energy redistribution offers a unique opportunity to control molecular dynamics. This study aims to provide experimental evidence of correlation-driven charge migration in adenine following ionization with an attosecond XUV pulse, demonstrating the potential for ultrafast control of molecular processes before non-adiabatic effects become significant.

Previous studies have investigated ultrafast charge migration in molecules, particularly aromatic amino acids, using attosecond spectroscopy. These studies have hinted at few-femtosecond charge dynamics, but a clear demonstration of the correlation-driven process, where the energy flow from a single excited electron to the coupled electrons is directly observed, has been lacking. Theoretical predictions have suggested the existence of such a process, but experimental verification was needed. The current study builds upon these previous works by employing advanced experimental techniques and theoretical simulations to directly probe the correlation-driven charge migration in a biologically relevant molecule, adenine.

The experiment involved ionizing adenine using an isolated 300-nm XUV pulse (photon energies from 15 to 35 eV) generated via high-harmonic generation. A waveform-controlled 64-fs NIR probing pulse was combined with the XUV pump pulse using an interferometric approach. Adenine was sublimated into a helium buffer gas and the produced ions were collected using a time-of-flight spectrometer as a function of the XUV-pump NIR-probe delay. The ion mass spectrum showed a dominance of ionic fragments, indicating a high probability of ionization in the energy range. The addition of the NIR pulse led to increased fragmentation. The key observation was the appearance of a doubly charged adenine ion at small positive delays, indicating a sub-3 fs delay. This observation was verified by comparing with similar measurements on krypton. First-principles time-dependent simulations using time-dependent density functional theory (TDDFT) and Ehrenfest dynamics were conducted to support the experimental findings. The simulations showed that the nuclei can be considered essentially frozen on the sub-3 fs timescale. A rate-equation approach was used to estimate shake-up transition times, providing a physical explanation of the experimental delay. More sophisticated ab initio calculations employing the non-equilibrium Green's function method were performed to account for electronic dynamics triggered by the XUV photoionization and subsequent NIR pulse absorption. These simulations provided a detailed view of orbital-resolved occupations, time-dependent electron density, and NIR-induced depletion of key states.

The experiment revealed a sub-3 fs delay in the formation of a doubly charged adenine ion after ionization with an attosecond XUV pulse, followed by a delayed NIR pulse. This delay is attributed to a correlation-driven charge migration process. Time-dependent density functional theory (TDDFT) simulations confirmed that nuclear motion is negligible within this timeframe. A rate equation approach estimated a shake-up transition time of 2.5 fs, consistent with the experimental delay, for a specific excited orbital (LUMO+6). More detailed ab initio calculations using the non-equilibrium Green's function method showed that this LUMO+6 state is populated through a shake-up process and that its subsequent depletion by the NIR pulse closely reproduces the experimental results for the doubly charged adenine ion yield. These calculations further revealed an out-of-plane charge migration associated with the LUMO+6 state, highlighting the role of electronic correlations in the delayed ionization. The simulations also showed that the LUMO+6 state is only accessed efficiently when the XUV pulse is polarized perpendicularly to the molecular plane.

The observed sub-3 fs delay in the formation of the doubly charged adenine ion directly supports the hypothesis of correlation-driven charge migration. The close agreement between experimental observations and the results from both simplified rate equations and sophisticated ab initio calculations strongly suggests that the phenomenon is due to a many-body effect and not an artifact of the experimental setup. The identification of a specific excited state (LUMO+6) and its associated out-of-plane charge migration mechanism provides valuable insights into the dynamics of this process. The ability to control the ionization process by manipulating the polarization of the attosecond pulse is a significant finding, opening the possibility for controlling molecular reactivity at the electronic timescale.

This study presents the first experimental observation of correlation-driven charge migration in adenine, occurring within a sub-3 fs timeframe. The findings, supported by theoretical simulations, reveal a mechanism involving shake-up excitation to a specific state (LUMO+6) and subsequent out-of-plane charge migration. Precise timing of a NIR control pulse can exploit this correlation-driven charge redistribution to prevent dissociative relaxation. This work demonstrates the potential for controlling molecular photo-reactivity at the electronic time scale, opening new avenues for research in attochemistry and molecular dynamics.

The theoretical simulations do not explicitly account for nuclear motion and non-adiabatic effects, which might become significant at longer timescales. While the simulations provide strong support for the proposed mechanism, the absence of nuclear dynamics in the calculations limits the ability to fully describe the complete molecular dynamics. Future studies could incorporate nuclear dynamics to refine the model and provide a more complete picture of the process.