Machine learning for ultrafast X-ray diffraction patterns on large-scale GPU clusters

Polychromatic synchrotron undulator X-ray sources are useful for ultrafast single-crystal diffraction under shock compression. Here, simulations of X-ray diffraction of shock-compressed single-crystal tantalum with realistic undulator sources are reported, based on large-scale molecular dynamics simulations. Purely elastic deformation, elastic–plastic two-wave structure, and severe plastic deformation under different impact velocities are explored, as well as an edge release case. Transmission-mode diffraction simulations consider crystallographic orientation, loading direction, incident beam direction, X-ray spectrum bandwidth and realistic detector size. Diffraction patterns and reciprocal space nodes are obtained from atomic configurations for different loading (elastic and plastic) and detection conditions, and interpretation of the diffraction patterns is discussed.

Download Full-text

THE STRUCTURE OF ACTIN-RICH FILAMENTS OF MUSCLES ACCORDING TO X-RAY DIFFRACTION

The Journal of Cell Biology ◽

10.1083/jcb.2.1.71 ◽

1956 ◽

Vol 2 (1) ◽

pp. 71-85 ◽

Cited By ~ 86

Author(s):

Cecily Cannan Selby ◽

Richard S. Bear

Keyword(s):

Large Scale ◽

Adductor Muscle ◽

X Ray Diffraction ◽

Cell Width ◽

X Ray ◽

Filament Axis ◽

Repeat Structure ◽

Diffraction Patterns ◽

The Individual ◽

Macromolecular Component

From analysis of moderate- to small-angle x-ray diffraction patterns, in the light of similar experience with paramyosin, has been derived the following description for the structure of actin-rich filaments in "tinted" portions of the adductor muscle of the clam, Venus mercenaria: 1. Some 11 diffraction maxima, widely streaked along layer lines and occurring at moderate diffraction angles (spacings 7 to 60 A) appear to be accounted for as (hk) reflections of a net whose cell elements are, for dry material: a ≑ 82 A, b = 406 A (filament axis identity period), and γ ≑ 82° (angle between a and b axes). These reflections follow a selection rule which indicates that the net cell is non-primitive and contains 15 equivalent locations (nodes) arranged as shown in Fig. 5. An alternative net has b' = 351 A and 13 nodes per cell. 2. Another interpretation rolls the net into a large-scale helix and places the 15 (or 13) nodes along 7 (or 6) turns of a helical locus projecting 406 (or 351) A along the filament axis. Whether considered to be built of planar-net or helix-net cells, the individual filament contains a single cell width transverse to its axis. Transverse filament dimensions are, therefore, in either case similar (50 to 100 A). 3. Consideration of existing electron-optical, physicochemical, and x-ray diffraction data regarding isolated actin suggests that the net cell is built of rods, each containing in cross-section from one to four actin molecules which run parallel to or twisted about rod axes that extend at 12° to the filament axis along the (21) diagonals of the cell. Depending on monomer shape, 2 to 15 monomers furnish length to reach across two cells, and the actin molecules are built into each rod in such a way as to repeat (or nearly repeat) structure 15 (or 13) times along the double cell length. Further details of intra-rod structure cannot be suggested because of lack of wide-angle diffraction information. 4. The actin system is sensitive to treatment of the muscle with ethanol. Concentrations of 5 per cent or greater abolish the net reflections. Other solvents—water, benzene, ether, pyridine, acetone—do not alter the pattern materially. 5. Two other reflections, occurring at the first and second layer lines of an axial periodicity of about 400 A, do not clearly belong to the actin-net system. They represent either a superstructure built upon the filaments by parts of the actin molecules themselves or by incorporated other molecular species, or they arise from an additional macromolecular component (possibly myosin, or its homologues or fractions) of similar axial periodicity.

Download Full-text

Symmetry prediction and knowledge discovery from X-ray diffraction patterns using an interpretable machine learning approach

Scientific Reports ◽

10.1038/s41598-020-77474-4 ◽

2020 ◽

Vol 10 (1) ◽

Author(s):

Yuta Suzuki ◽

Hideitsu Hino ◽

Takafumi Hawai ◽

Kotaro Saito ◽

Masato Kotsugi ◽

...

Keyword(s):

Machine Learning ◽

Crystal Structure Analysis ◽

Group Classification ◽

X Ray Diffraction ◽

Space Group ◽

X Ray ◽

Crystal System ◽

Machine Learning Approach ◽

Diffraction Patterns ◽

Xrd Patterns

AbstractDetermination of crystal system and space group in the initial stages of crystal structure analysis forms a bottleneck in material science workflow that often requires manual tuning. Herein we propose a machine-learning (ML)-based approach for crystal system and space group classification based on powder X-ray diffraction (XRD) patterns as a proof of concept using simulated patterns. Our tree-ensemble-based ML model works with nearly or over 90% accuracy for crystal system classification, except for triclinic cases, and with 88% accuracy for space group classification with five candidates. We also succeeded in quantifying empirical knowledge vaguely shared among experts, showing the possibility for data-driven discovery of unrecognised characteristics embedded in experimental data by using an interpretable ML approach.

Download Full-text

Cover art

University of Western Ontario Medical Journal ◽

10.5206/uwomj.v85i1.4191 ◽

2016 ◽

Vol 85 (1) ◽

Author(s):

Rakesh Gudimella

Keyword(s):

Large Scale ◽

Molecular Model ◽

Essential Medicines ◽

Chemical Formula ◽

X Ray Diffraction ◽

X Ray ◽

X Ray Crystallography ◽

Chemical Structures ◽

Diffraction Patterns ◽

Cover Art

Cover art by Rakesh Gudimella. In 1964, Dorothy Hodgkin won the Nobel Prize for the discovery of the structure of penicillin using the emerging technique of x-ray crystallography. The original x-ray diffraction patterns and the subsequent molecular model she created is shown in the foreground. Although the chemical formula of penicillin was known, its structure was not, making it difficult to produce on a large scale. Her discovery set us on the path to understanding antibiotic mechanisms and opened the door for the synthesis of cephalosporins and other important medications. The background shows the chemical structures of several lifesaving and influential drugs on the WHO List of Essential Medicines.

Download Full-text

Data-driven approach for synchrotron X-ray Laue microdiffraction scan analysis

Acta Crystallographica Section A Foundations and Advances ◽

10.1107/s2053273319012804 ◽

2019 ◽

Vol 75 (6) ◽

pp. 876-888 ◽

Cited By ~ 1

Author(s):

Yintao Song ◽

Nobumichi Tamura ◽

Chenbo Zhang ◽

Mostafa Karami ◽

Xian Chen

Keyword(s):

Machine Learning ◽

National Laboratory ◽

Machine Learning Algorithms ◽

Data Driven ◽

Computational Pipeline ◽

X Ray Diffraction ◽

X Ray ◽

Data Driven Approach ◽

Labeling Algorithms ◽

Diffraction Patterns

A novel data-driven approach is proposed for analyzing synchrotron Laue X-ray microdiffraction scans based on machine learning algorithms. The basic architecture and major components of the method are formulated mathematically. It is demonstrated through typical examples including polycrystalline BaTiO3, multiphase transforming alloys and finely twinned martensite. The computational pipeline is implemented for beamline 12.3.2 at the Advanced Light Source, Lawrence Berkeley National Laboratory. The conventional analytical pathway for X-ray diffraction scans is based on a slow pattern-by-pattern crystal indexing process. This work provides a new way for analyzing X-ray diffraction 2D patterns, independent of the indexing process, and motivates further studies of X-ray diffraction patterns from the machine learning perspective for the development of suitable feature extraction, clustering and labeling algorithms.

Download Full-text

Optimization of Material Composition of Li‐Intercalated Metal–Organic Framework Electrodes Using a Combination of Experiments and Machine Learning of X‐Ray Diffraction Patterns

Advanced Materials Technologies ◽

10.1002/admt.202000254 ◽

2020 ◽

pp. 2000254

Author(s):

Hirofumi Hazama ◽

Daisuke Murai ◽

Naoyuki Nagasako ◽

Masaki Hasegawa ◽

Nobuhiro Ogihara

Keyword(s):

Machine Learning ◽

Metal Organic Framework ◽

Material Composition ◽

X Ray Diffraction ◽

X Ray ◽

Metal Organic ◽

Diffraction Patterns ◽

Organic Framework

Download Full-text

Deducing density and strength of nanocrystalline Ta and diamond under extreme conditions from X-ray diffraction

Journal of Synchrotron Radiation ◽

10.1107/s1600577518017216 ◽

2019 ◽

Vol 26 (2) ◽

pp. 413-421 ◽

Cited By ~ 3

Author(s):

Y. Y. Zhang ◽

M. X. Tang ◽

Y. Cai ◽

J. C. E ◽

S. N. Luo

Keyword(s):

Shock Wave ◽

Molecular Dynamics ◽

Large Scale ◽

Density Measurement ◽

Dynamics Simulation ◽

Extreme Conditions ◽

X Ray Diffraction ◽

X Ray ◽

Diffraction Patterns ◽

Dynamics Simulations

In situ X-ray diffraction with advanced X-ray sources offers unique opportunities for investigating materials properties under extreme conditions such as shock-wave loading. Here, Singh's theory for deducing high-pressure density and strength from two-dimensional (2D) diffraction patterns is rigorously examined with large-scale molecular dynamics simulations of isothermal compression and shock-wave compression. Two representative solids are explored: nanocrystalline Ta and diamond. Analysis of simulated 2D X-ray diffraction patterns is compared against direct molecular dynamics simulation results. Singh's method is highly accurate for density measurement (within 1%) and reasonable for strength measurement (within 10%), and can be used for such measurements on nanocrystalline and polycrystalline solids under extreme conditions (e.g. in the megabar regime).

Download Full-text

Evaluation of Freezing Methods by Low Temperature X-Ray Diffraction

Proceedings, annual meeting, Electron Microscopy Society of America ◽

10.1017/s0424820100079176 ◽

1977 ◽

Vol 35 ◽

pp. 330-333

Author(s):

T. Gulik-Krzywicki ◽

M.J. Costello

Keyword(s):

Biological Materials ◽

Molecular Organization ◽

X Ray Diffraction ◽

Glass Capillary ◽

X Ray ◽

Rate Dependent ◽

Before And After ◽

Freeze Etching ◽

Diffraction Patterns ◽

Set Up

Freeze-etching electron microscopy is currently one of the best methods for studying molecular organization of biological materials. Its application, however, is still limited by our imprecise knowledge about the perturbations of the original organization which may occur during quenching and fracturing of the samples and during the replication of fractured surfaces. Although it is well known that the preservation of the molecular organization of biological materials is critically dependent on the rate of freezing of the samples, little information is presently available concerning the nature and the extent of freezing-rate dependent perturbations of the original organizations. In order to obtain this information, we have developed a method based on the comparison of x-ray diffraction patterns of samples before and after freezing, prior to fracturing and replication.Our experimental set-up is shown in Fig. 1. The sample to be quenched is placed on its holder which is then mounted on a small metal holder (O) fixed on a glass capillary (p), whose position is controlled by a micromanipulator.

Download Full-text

Examination of Rabbit Fc Crystals

Proceedings, annual meeting, Electron Microscopy Society of America ◽

10.1017/s0424820100100445 ◽

1968 ◽

Vol 26 ◽

pp. 140-141

Author(s):

J. P. Robinson ◽

P. G. Lenhert

Keyword(s):

Molecular Weight ◽

Flat Plate ◽

Phosphate Buffer ◽

Asymmetric Unit ◽

X Ray Diffraction ◽

X Ray ◽

Neutral Ph ◽

Small Crystals ◽

Carbon Coated ◽

Diffraction Patterns

Crystallographic studies of rabbit Fc using X-ray diffraction patterns were recently reported. The unit cell constants were reported to be a = 69. 2 A°, b = 73. 1 A°, c = 60. 6 A°, B = 104° 30', space group P21, monoclinic, volume of asymmetric unit V = 148, 000 A°3. The molecular weight of the fragment was determined to be 55, 000 ± 2000 which is in agreement with earlier determinations by other methods.Fc crystals were formed in water or dilute phosphate buffer at neutral pH. The resulting crystal was a flat plate as previously described. Preparations of small crystals were negatively stained by mixing the suspension with equal volumes of 2% silicotungstate at neutral pH. A drop of the mixture was placed on a carbon coated grid and allowed to stand for a few minutes. The excess liquid was removed and the grid was immediately put in the microscope.

Download Full-text