Coming dissertations at Uppsala university

The brilliant femtosecond pulses of an X-ray free electron laser could enable coherent diffractive imaging of single biomolecules, producing a weak and noisy diffraction pattern from a single perspective, before being disintegrated by the energetic X-rays. By taking advantage of the principle of "diffraction before destruction'' it is possible, with sufficient signal strength, to reconstruct the electron density of the particle using phase retrieval algorithms on the diffraction intensities. Combining diffraction patterns from serial measurements of identical samples taken with unknown particle orientations is called cryptotomography. It is possible to reconstruct a model of the three-dimensional diffraction intensities by using orientation recovery algorithms, like expand, maximize, and compress (EMC), enabling a three-dimensional phase retrieval. I tested the performance of the EMC algorithm using simulated diffraction patterns and real background measurements, combined in various proportions, to investigate the impact of background on orientation recovery, and I proposed a signal- and background-dependent limit on three-dimensional phase retrieval. I also simulated diffraction using a preferential alignment scheme based on dipole orientation in an electric field, which I took advantage of in EMC, enhancing the algorithm with biased orientations. I found that biased orientations are very advantageous, especially for diffraction patterns with added incoherent background noise. Finally, I discuss some of the conditions for successful imaging of single proteins, for future experiments.