Phase retrieval from noisy X-ray diffraction patterns of single molecules

単粒子分子のノイズの大きいX線回折データからの位相回復

河野 秀俊; 池田 思朗*; 徳久 淳師*

Kono, Hidetoshi; Ikeda, Shiro*; Tokuhisa, Atsushi*

A new phase retrieval method for noisy X-ray diffraction imaging is proposed. Conventional phase retrieval methods effectively solve the problem for high signal-to-noise ratio measurements, but would not be sufficient for single biomolecular imaging which is expected to be realized with femto-second X-ray free electron laser pulses which is soon available at SPring8. The new phase retrieval method is based on the Bayesian statistical approach. The features of the method are that (1) it does not need to set the object boundary constraint that is required by the commonly used hybrid input-output (HIO) method, instead a sparse prior is used for the object estimation, and we call it SPR (sparse phase retrieval) method and (2) it is about 10 times robust to noise compared with HIO method. The simulation results demonstrate that the SPR method can reconstruct the object efficiently under a realistic condition.