Extending the Methodology of X-ray Crystallography to Non-Crystalline Specimens

Since the emergence of powerful X-ray sources such as synchrotron radiation, the area of X-ray structure determination gradually evolved into two fields. The relatively small one is X-ray Microscopy which employs X-ray lenses such as zone plates, compound refractive lenses, curved mirrors and glass capillaries to focus X-rays to determine the structures of non-crystalline specimens. The resolution of X-ray Microscopy is limited by the resolution of the X-ray lenses and, for biological specimens, radiation damage. While the radiation damage problem can be alleviated somewhat by cooling the specimens down to liquid nitrogen temperature, the resolution of X-ray lenses is limited by fabrication difficulties. At present, the highest resolution achievable is around SO-SOnm'. The other far more successful field is Xray crystallography. By employing crystalline specimens, where constructive interference among the large number of identical unit cells generates strong Bragg peaks, X-ray crystallography can achieve atomic resolution without serious radiation damage to the specimens. The major difficulty of X-ray crystallography, however, is that the specimens should be crystalline, while most complex biological specimens, for example, can not be or are too big to be crystallized. The idea of extending X-ray crystallography to imaging non-crystalline specimens, using a combination of X-ray Microscopy and X-ray Crystallography, was first proposed by Sayre in 1980. However, during almost twenty years, the experimental progress of pursuing the idea was not successful due to two difficulties. First of all, when a specimen is non-crystalline, the diffraction pattern is weak and continuous, which makes the data acquisition challenging. The other is the well-known phase problem. That the oversampling technique could be used to retrieve the phase problem in X-ray diffraction studies of non-crystalline specimens was made by Sayre. By employing this phasing technique, we recently demonstrated for the first time that a soft X-ray diffraction pattern from a micron-size non-crystalline specimen can be recorded and inverted to form a high-resolution image'. With the experimental setup shown in Fig. 1, we recorded high quality diffraction patterns from non-crystalline specimens. By using a 10 |iim pinhole and also placing the specimen only about 30 Jim from the corner of the silicon nitride membrane, we obtained a clean diffraction pattern on the detector in three quadrants, free from background due to scattering from the pinhole. The detector was a back-thinned and liquid nitrogen cooled CCD with 512 x