Advances in Biomolecular Imaging with X-ray Free-Electron Lasers

Abstract: Utilizing X-rays to solve molecular structures has proven to be an immensely powerful and im- portant scientific technique. The invention of X-ray crystallography has allowed for countless breakthroughs in chemistry, biology and material science and remains the number one method used for structural determination today. Of particular interest is the structures of biomolecules, such as proteins, due to their medical relevance. Unfortunately, the need for crystals of sufficient size constitutes the biggest drawback to this approach. This is troubling since many of the im- portant biomolecules, in particular membrane proteins, have proven to be difficult or sometimes even impossible to crystalize. When limited to a small nanocrystal or even a single particle, con- ventional crystallography is no longer adequate to probe the structure at high enough resolution. Recent developments, most notably the introduction of X-ray free-electron lasers (XFELs), have opened up new possibilities for circumventing these limitations. The high intensities and ultra- short pulse lengths provided by XFELs allows for diffractive imaging of smaller crystals through Serial Femtosecond Crystallography (SFX) and can even be extended to single molecules, Single Particle Imaging (SPI). These methods are still in their infancies, and much research and refine- ment is needed before they can be properly established.The current work covers fundamental studies of X-ray interaction with biomatter carried out to aid and improve upon SFX and SPI. Three papers based on computer simulation studies are presented, related to mainly two central aspects faced when imaging molecules with XFELs. Pa- per I explores a novel approach using explosion mapping to facilitate spatial orientation of single particles, which is necessary to reconstruct the three dimensional structure from two dimensional diffraction patterns. Paper II concerns radiation damage of the sample in SFX experiments using a plasma model and studies the impact of different pulse profiles on these processes. Lastly, pa- per III outlines the details of an online database available to researchers worldwide that contains simulated data on damage development in samples exposed to an XFEL pulse.In the first study, molecular dynamics was adopted to map the XFEL-induced Coulomb explo- sions in SPI for biomolecules. Four proteins were investigated, each with three different levels of hydration, and it was found that explosion patterns for both carbon and sulfur ions are re- producible for all twelve systems. However, water bound to the protein surface seems to have a shielding effect on carbons, causing their trajectories to be favored toward the exposed parts of the sample. This complicates the adoption for orientation determination as the water content would have to be known. Sulfurs, on the other hand, showed no signs of water dependence and consistently produced similar explosion patterns regardless of hydration level. We speculate that this is because of their higher mass and ionization cross section and conclude that mapping of heavier ions could provide valuable information for spatial orientation.In the second study, radiation damage in terms of ionization and atomic displacement within a nanometer-sized crystal illuminated by an XFEL pulse was explored with a non-local thermody- namic equilibrium plasma code. Different temporal distributions of the same number of photons was employed to assess its impact of damage dynamics. The results show that the pulse profile is substantially important. A front-loaded pulse is more beneficial for imaging purposes since the bulk of the photons encounters an undamaged sample. If the majority of photons instead arrive late, early photons will already have initiated the crystal decay causing further contribution to the diffraction pattern to be degraded.In the third study, the free-electron laser damage simulation database (FreeDam) was estab- lished. It presents simulated time-resolved data for average ionization, ion and electron temper- atures and atomic displacement for various materials and XFEL parameters. Simulations were carried out using the same code as in paper II, and the data is freely available online.This thesis is aimed to provide one of the stepping stones toward atomic resolution imaging of nanocrystals and single particles with free-electron lasers. If realized, these techniques could well turn out to be one of the greatest scientific achievements of the 21th century.