摘要：

X-ray diffractive imaging is at the very heart of materials science and has been utilized fordecades to solve unknown molecular structures. Nowadays, it serves as the key method ofstructural biology to solve molecular structures of large biological molecules comprisingseveral thousand or even millions of atoms. However, x-ray diffraction from isolatedmolecules is very weak. Therefore, the regular and periodic arrangement of a huge numberof identical copies of a certain molecule of interest within a crystal lattice has been anecessary condition in order to exploit Bragg diffraction of x-rays. This results in a hugeincrease in scattered signal and a strongly improved signal-to-noise ratio compared todiffraction from non-crystalline samples. The major bottleneck of structural biology isthat many of biologically interesting molecules refuse to form crystals of sufficient sizeto be used at synchrotron x-ray lightsources. However, novel x-ray free-electron lasers(XFELs), which became operational very recently, promise to address this issue.X-ray pulses provided by XFELs are many orders of magnitude more intense than x-raypulses from a synchrotron source and at the same time as short as only several tens offemtoseconds. Combined with wavelengths in the nm–pm range, XFELs are well-suitedto study ultrafast atomic and molecular dynamics. Additionally, the ultrashort pulses canbe utilized to circumvent the damage threshold which set a limit to the incident intensityin x-ray diffraction experiments before. At XFELs, though eventually destroying the in-vestigated sample, no significant sample deterioration happens on the ultrashort timescaleof the XFEL pulse and the measured diffraction pattern is due to an (almost) unharmedsample.In the framework of this thesis, the approach of utilizing the highly intense XFEL pulsesfor x-ray diffraction of weakly-scattering non-crystalline samples was taken to the limitof small isolated molecules. X-ray diffraction was performed on a gas-phase ensembleof the prototypical molecule 2,5-diiodobenzonitrile (C7 H3 I2 N, DIBN) at the x-ray free-electron laser LCLS. The target molecules were laser-aligned along a common axis in thelaboratory frame by a Nd:YAG laser. Reaching a strong degree of molecular alignment,was an important step in this experiment. Therefore, a significant part of the work wasdedicated to gaining control of the molecular degrees of freedom. In order to reach ahigh degree of alignment, the target molecules were prepared in low rotational quantumstates by means of efficient cooling in a supersonic expansion from a pulsed valve followedby spatial quantum-state selection in an electrostatic deflector. Utilization of the deflec-tor significantly improved alignment of the DIBN molecules. Further applications of thedeflection technique such as, e. g., the spatial separation of several species of molecularcomplexes/clusters are presented in this thesis as well.The quantum-state selected and strongly laser-aligned samples were probed by the x-raypulses of LCLS and the obtained diffraction patterns show a significant difference whencomparing diffraction from aligned and isotropically-distributed DIBN which agrees wellwith theory. The results represent an important step in the effort of pushing diffractiveimaging of non-crystalline samples at XFELs towards the single-molecule limit. Conceptsand experimental requirements for future experiments of this kind are discussed, involving,e. g., the step towards imaging of laser-aligned large (bio)macromolecules or imaging ofultrafast fragmentation dynamics in femtosecond pump-probe experiments at XFELs.Stern, Stephan

展开