METHOD FOR DETERMINING CRYSTAL STRUCTURE AND APPARATUS FOR CARRYING OUT THE SAME - Patent application
A computer-implemented method using dynamic diffraction theory and three-dimensional electron diffraction processing refines crystal structures, addressing inaccuracies in existing electron diffraction methods by enhancing precision and reducing radiation sensitivity, especially for organic materials.
Patent Information
- Application Number
- JP2024513360
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2021-09-01
- Filing Date
- 2022-09-01
- Publication Date
- 2024-11-06
- Estimated Expiration
- 2042-09-01
AI Technical Summary
Existing methods for determining crystal structures using electron diffraction struggle with dynamic diffraction effects, particularly for non-precessional electron diffraction data, leading to inaccurate structural models, especially for organic materials sensitive to electron radiation.
A computer-implemented method utilizing dynamic diffraction theory to refine crystal structures, incorporating three-dimensional electron diffraction data processing, virtual diffraction frame generation, and structural model optimization, which accounts for dynamic diffraction effects, enabling accurate determination of crystal structures, including those of non-centrosymmetric crystals.
The method achieves accurate crystal structure determination with improved precision, particularly for organic materials, offering faster data processing and reduced electron radiation sensitivity, comparable to precessional electron diffraction results but without the need for specialized equipment.
Smart Images

Figure 2024540800000001_ABST
Abstract
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application claims the benefit of priority to Czech Patent Application No. CZ2021-403, filed September 1, 2021, and U.S. Patent Application No. 17 / 565,624, filed December 30, 2021, both filed by Fyzikalni ustav AV CR, vvi, Czech Academy of Sciences, a public research institution located in Prague, Czech Republic. [Technical field]
[0002] The present invention relates to a computer-implemented invention. More particularly, the present invention relates to processing rotation electron diffraction data of crystalline materials. In one embodiment, the present invention relates to a method for refining a crystal structure from rotation electron diffraction using dynamic refinement of the data, which allows the atomic structure to be accurately determined. [Background technology]
[0003] Structural analysis is a method to obtain information about the spatial arrangement of atoms in a crystal structure. The analysis includes several steps, such as crystal preparation, data acquisition, finding an initial structural model, and optimizing the structural model.
[0004] The structural model is optimized by a mathematical procedure called the least-squares method. In this procedure, experimental diffraction data are measured on a crystal and compared with theoretical diffraction data calculated based on the actual structural model. By adjusting the structural model, the differences between the experimental and theoretical data are minimized. For this procedure to be successful, the calculation of the theoretical data from the structural model must be accurate.
[0005] If electrons are used in the data acquisition step, it is known as an electron diffraction experiment. In this case, the calculation of theoretical diffraction intensities is a complicated process, due to the existence of the so-called "dynamic diffraction effect".
[0006] Methods for calculating accurate theoretical diffraction data from electron diffraction, so-called refinements, have been designed for the special case of precession electron diffraction data; no comparable methods yet exist for general electron diffraction data. Summary of the Invention [Problem to be solved by the invention]
[0007] In view of the above mentioned drawbacks, it is an object of the present invention to provide a method for refining a crystal structure, the steps of which are performed by a computer, and thus the method is suitable for obtaining good structure optimizations using general electron diffraction data, and not only for the special case of precession electron diffraction.
[0008] It is a further object of the present invention to provide a computer-implemented method capable of using a wide range of crystallographic calculations to determine the electrostatic potential distribution and crystal structures of a wide range of chemical compounds and crystal types. [Means for solving the problem]
[0009] In the first aspect of the present invention, a method for determining crystal structure model is provided.The method preferably includes, but is not limited to, organic crystals.Crystal structure determination can be performed for inorganic crystals such as metal-containing compounds, for example hydrated cobalt and aluminum phosphate, or minerals such as quartz, natrolite, and organic crystals such as amino acids, carbohydrates, and their derivatives, and many drugs such as abiraterone acetate.However, it is particularly advantageous for determining protein structure, and determining the crystal structure of antibodies, genes, and drug delivery samples.
[0010] The method comprises: acquiring electron diffraction data from a crystal by three-dimensional electron diffraction, the data comprising information about crystal diffraction patterns, each diffraction pattern comprising information about the scattered electron intensity and crystal orientation in each direction relative to an incident electron beam during data acquisition, the crystal being rotated or tilted during recording of the diffraction data; processing the data to determine crystal lattice parameters, crystal orientations at the time of acquisition of the diffraction patterns, indices of each reflection, and scattered electron intensities of each reflection on each diffraction pattern, thereby generating processed data; A step of generating virtual diffraction frames by a computer program, each virtual diffraction frame including a list of integrated scattered electron intensities in each direction, summing the contributions from all experimental patterns forming a virtual frame, the virtual diffraction frames being formed by calculation from the provided three-dimensional electron diffraction data.
[0011] Providing an approximate crystal structure model; and Refining the approximate crystal structure model using dynamic diffraction theory wherein the refining step comprises: calculating a modeled intensity of the scattered electrons using dynamic diffraction theory based on an approximate crystal structure model; and minimizing the difference between the experimentally determined scattered electron intensities and the modeled intensities, the refined crystal structure model being the model with the least deviation. Includes.
[0012] The method according to the invention allows to determine the crystal structure, preferably the absolute structure of non-centrosymmetric crystals, which can be obtained by dynamic refinement. Compared to the kinematic approximation, the method according to the invention does not neglect dynamic diffraction effects in the diffraction data, and therefore makes it possible to obtain more accurate information about the crystal structure. The combination of steps according to the method of the invention is adapted to generate a so-called virtual diffraction frame, in which theoretical data corresponding to the experimental data in the virtual diffraction frame can be calculated more easily from the crystal structure model, compared to the prior art methods in which no virtual diffraction frame is generated. As a result, a structure model can be obtained with an accuracy similar to that of the model established by precessional electron diffraction, and with an accuracy much better than the procedures currently used for non-precessional electron diffraction data. The advantages of the method using non-precessional data compared to precessional data are easier access without the need for special equipment for the precessional method, faster data processing, and a lower overall illumination of the electron crystal during data acquisition, which is particularly advantageous for the analysis of organic materials that are sensitive to electron irradiation. Such optimized structural models provide better insight into the crystal structure and provide important information such as the absolute structure of crystals containing chiral molecules or the positions of light atoms that are difficult to detect but are often important for materials functions and applications.
[0013] In a preferred embodiment, a method for determining the absolute structure of a non-centrosymmetric crystal is provided, the method comprising the steps of: generating a model of an absolute structure that is inverse to the current approximate crystal structure model, thereby generating an inverted model; refining the approximate crystal structure model using dynamic diffraction theory on the inversion model; and comparing quality indicators of the inverted models, the correct structure being the one which results in the smallest deviation between the experimental data and the data calculated based on the refined structural model; Includes.
[0014] Another solution to the technical problem of obtaining a crystal structure is provided in a second aspect of the present invention. An alternative method for solving the same technical problem for determining the absolute structure of a non-centrosymmetric crystal capable of electron diffraction is now disclosed, which method comprises: acquiring electron diffraction data from a crystal by three-dimensional electron diffraction, the data comprising information about crystal diffraction patterns, each diffraction pattern comprising information about scattered electron intensity and crystal orientation for each direction relative to an incident electron beam during data acquisition, the crystal being rotated or tilted during data acquisition; and processing the data to determine crystal lattice parameters, crystal orientations at the time of acquisition of the diffraction patterns, indices of each reflection, and scattered electron intensities of each reflection on each diffraction pattern, thereby generating processed data; generating virtual diffraction frames by a computer program, each virtual diffraction frame comprising a list of integrated scattered electron intensities in each direction summing contributions from all experimental patterns forming one virtual frame, the virtual diffraction frames being computationally generated from the provided three-dimensional electron diffraction data; providing an approximate crystal structure model; calculating a quality index of the approximate crystal structure model using dynamic diffraction theory and calculating the model strength; generating a model of an absolute structure that is inverse to the current approximate crystal structure model, thereby generating an inverted model; Calculating a quality index of the inverted crystal structure model using dynamic diffraction theory and calculating the model strength; and comparing quality indicators of the inverted and non-inverted models, the correct structure being the one which results in the smallest deviation between the experimental data and the data calculated based on the refined structural model; Includes.
[0015] Stepwise or continuous rotation means that the sample rotates around the rotation axis within a certain preselected range, for example from -50° to 50° in discretely defined steps, for example 1°, or continuously. Diffraction patterns are exposed either in "step" mode after each rotation or in continuous mode during the crystal rotation, see for example Gemmi M, Mugnaioli E, Gorelik TE, Kolb U, Palatinus L, Boullay P, et al. 3D Electron Diffraction: The Nanocrystallography Revolution. ACS Cent Sci. 2019 Aug 28; 5 (8): 1315-29.
[0016] A diffraction pattern is a record of the direction and intensity of electrons scattered by a crystal and is usually acquired by an electronic device, i.e. an electronic detector, arranged such that the crystal is located between the electron source and the detector. The diffraction acquisition takes the form of an image, i.e. a two-dimensional recording, which contains information about the intensity of the reflections. A typical example of a diffraction acquisition is shown in Figure 1.
[0017] Those skilled in the art understand reflection as the point where a diffracted beam strikes a detector, see for example the text Vaclav Valvoda, Milena Polcarova, Pavel Lukac, Zaklady strukturni analyzy, Karolinum, Prague, 1992, ISBN 80-200-0280-4.
[0018] An approximate crystal structure model is understood to be a list of atoms with their approximate positions in the elementary cell of the crystal and the determination of the atom types. This model can be obtained from the literature if already published, or directly from diffraction data by established structure analysis methods, see for example the text of Giacovazzo, C. (editor), Fundamentals of Crystallography, Third Edition, Oxford University Press, 2011, ISBN 9780199573653. In another embodiment, the skilled person can provide a prior analysis of the data obtained from experimental diffraction patterns and provide an educated guess of the approximate crystal structure model.
[0019] The necessary information obtained from the diffraction data is the parameters of the unit cell of the crystal, also known as the crystal lattice parameters, and a list of the scattered electron intensities and standard deviations of all reflections measured during the experiment. The reflections are characterized by three (and in exceptional cases more than three) integer indices that determine their position in the inverse crystal lattice, see for example the text by Vaclav Valvoda, Milena Polcarova, Pavel Lukac, Zaklady strukturni analyzy, Karolinum, Prague, 1992, ISBN 80-200- 0280-4.
[0020] The refinement quality index is a number that estimates how good and reliable the refined structural model is. This quality index may be advantageously defined as the R-coefficients R1, R2 and wR2 as follows:
[0021]
number
[0022] Where: TIFF2024540800000003.tif6114
[0023] each TIFF2024540800000004.tif6114
[0024] are observed, which are the calculated intensities of a particular diffracted electron beam characterized by a diffraction vector g, where TIFF2024540800000005.tif6114
[0025] and TIFF2024540800000006.tif5114
[0026] The following values were determined: TIFF2024540800000007.tif6114
[0027] The particular choice of quality metric to be used is application dependent and can be made by the user of the invention based on their expert knowledge.
[0028] The present invention is particularly suitable for determining the absolute structure of crystals containing light atoms such as hydrogen, lithium, beryllium, or boron.
[0029] A further advantage of the present invention is that precise atomic positions within a structure can be achieved with an average error of less than 0.005 nm.
[0030] Three-dimensional electron diffraction is a method in which the electron beam and the crystal are rotated relative to each other to obtain data, particularly the intensity of the electrons scattered on the crystal under study. Rotation can be achieved by rotating the electron beam using the magnetic coils of a transmission electron microscope, or by rotating the crystal using a goniometer, or by a combination of both methods. In some embodiments, the rotation can be tilted.
[0031] In a preferred embodiment of the rotational diffraction data acquisition method, the electron scattering intensity is recorded in a series of successive images, and after acquisition of each diffraction pattern the crystal is rotated through an angle about the axis of the goniometer. The magnitude of the rotation is typically between 0.1° and 1°. The total tilt of the crystal can typically be in the range of ±60°, but in some embodiments can be greater.
[0032] In a preferred embodiment, the step of acquiring rotational electron diffraction data combines rotation of the electron beam and tilt of the crystal, and several diffraction patterns for different electron beam tilts are recorded for each crystal tilt. The preferred embodiment therefore has the advantage that the mutual orientation of the crystal and the beam is more accurate due to the fact that in some cases the electron beam tilt can be controlled with greater precision than the goniometer tilt.
[0033] In another preferred embodiment, continuous crystal tilt is combined with continuous diffraction data acquisition. The crystal goniometer rotates continuously around the tilt axis during electron diffraction imaging. Each rotational electron diffraction image covers an angular range determined by the goniometer rotation speed and exposure time. Unlike sequential rotations, continuous rotation records the integrated electron diffraction intensity. This preferred embodiment using continuous tilt of the crystal offers the advantage of a simpler experimental design, minimizing electron exposure to the crystal and acquiring integrated intensities.
[0034] All the above data acquisition methods by rotational electron diffraction are described in detail in M. Gemmi and A. Lanza: 3D electron diffraction techniques, Acta Crystallographica B, Vol. 75, pp. 495-504, year of publication 2019.
[0035] Data processing is - finding the maxima of all the diffraction patterns and determining their coordinates in reciprocal space; - finding the parameters of the crystal lattice and determining the orientation of the crystal in space, i.e. determining the so-called orientation matrix; determining the diffraction intensities of all potentially excited reciprocal lattice points on all diffraction patterns; Includes.
[0036] The so-called virtual diffraction frame is generated as follows: A suitable number of consecutive experimental diffraction patterns are selected and combined into one virtual frame, such that the total angular range covered by the virtual diffraction frame is preferably about 1°-3°.
[0037] ·It can be seen that all points of the reciprocal lattice lie within the angular range covered by the virtual diffraction frame.
[0038] For each point found, all the diffraction intensities belonging to that point on all experimental diffraction patterns are summed.
[0039] The list of acquired intensities belonging to the points of the reciprocal lattice forms a data set called a virtual diffraction frame.
[0040] The orientation of the virtual diffraction frame is determined as the average of the orientations of all the experimental diffraction patterns that compose it.
[0041] The angular difference between subsequent virtual diffraction frames is preferably selected such that the volumes of reciprocal space covered by adjacent virtual diffraction frames partially overlap.
[0042] In a preferred embodiment, the data processing also includes refinement of the diffraction geometry, i.e. the orientation of the crystal relative to the incident electron beam.
[0043] Refinement is a general term used to describe all the operations required to develop a test model into one that best corresponds to the observed data, in particular the intensity of electrons scattered by the crystal. Refinement can be viewed as a series of mathematical procedures carried out by a computer. The refinement of the crystal structure prediction of the present invention utilizes dynamical diffraction theory, which takes into account multiple scattering of electrons by atoms in the crystal.
[0044] A step of determining the quality of the refined crystal structure model is to determine the deviation between the observed quantities and the quantities simulated according to the crystal structure model, preferably using the parameter R. If the deviation is small enough, the crystal structure model is considered to be accurate.
[0045] In another embodiment of the present invention, an apparatus suitable for determining a crystal structure model according to the present invention is described.
[0046] The device comprises: an electron source capable of emitting an electron beam; a crystal holder adapted to perform a rotational or tiling movement with the crystal; A detector that detects electrons scattered from a crystal The detector comprises: Data Storage and Processing Unit where: The processing unit is adapted to determine the crystal lattice parameters, the crystal orientation at the time of acquisition of the diffraction patterns, the index of each reflection, and the scattered electron intensity of each reflection on each diffraction pattern, thereby generating processed data; and generating virtual diffraction frames by a computer program, each virtual diffraction frame including a list of integrated scattered electron intensities in each direction summing contributions from all experimental patterns forming a virtual frame, said virtual diffraction frames being computationally generated from said provided three-dimensional electron diffraction data; the data storage is configured to provide an approximate crystal structure model or the processing unit is adapted to perform an analysis of the processed data leading to the acquisition of the approximate structure model; The processing unit is adapted to refine the structural model.
[0047] In a preferred embodiment, the processing unit is adapted to refine the approximate crystal structure model using dynamic diffraction theory, wherein during refining, the processing unit comprises: calculating a modeled intensity of the scattered electrons using a dynamic diffraction theory based on an approximate crystal structure model; and Minimize the difference between the experimentally determined scattered electron intensity and the modeled intensity where the refined crystal structure model is the model with the smallest deviation.
[0048] In another preferred embodiment, the processing unit is configured for determining the absolute structure of a non-centrosymmetric crystal, wherein the processing unit comprises: Generate a model of the absolute structure that is opposite to the current approximate crystal structure model, thereby generating an inverted model; Refining the inverse crystal structure model using dynamic diffraction theory, and Compare quality metrics of inverted and non-inverted models where the correct structure is the one which results in the smallest deviation between the experimental data and the data calculated based on the refined structural model.
[0049] In another preferred embodiment, the processing unit is adapted to refine the approximate crystal structure model and the apparatus comprises: Calculate a quality index of the approximate crystal structure model; and Compare quality metrics where the correct crystal structure is taken as the modelled crystal structure which results in the smallest deviation between the experimental data and the data calculated on the basis of the refined structural model.
[0050] In another preferred embodiment, the processing unit is adapted to generate a virtual model of a crystal structure opposite to the structure of the current approximate crystal structure model, thereby generating an inverted model; the step of comparing the quality indicators is a step of comparing the quality indicators of the inverted model and the experimental data, and the correct structure is the structure having the smallest deviation between them.
[0051] Preferably, the detector is set up to collect data simultaneously with the rotation of the crystal.
[0052] In another preferred embodiment, the detector is set up to collect data stepwise during the rotation of the crystal. [Brief description of the drawings]
[0053] [Figure 1] An example of a diffraction pattern. This section shows a magnified image of one reflection with the specified refractive index, intensity, and standard deviation of the intensity. [Diagram 2] Schematic illustration of the generation of virtual diffraction frames from experimental diffraction patterns. In particular, Fig. 2 is a view along the rotation axis. The blue lines define the region of space covered by one experimental pattern (eo) exposed during the crystal rotation. The blue arcs with arrows indicate the angular range covered by the experimental patterns. The figure shows three such experimental patterns. Several experimental patterns (in Fig. 2) are always combined into one virtual frame (indicated in the figure by red arcs with arrows). [Diagram 3] Basic diagram of the device allowing the refinement of structures according to the invention. Figure 3 shows an electron source 1 emitting electrons 2, from which a condenser lens system 3 forms a focused electron beam 4 which is incident on a crystal sample 5, from which scattered electrons 6 propagate, which are incident on a detector 7 connected to a computing unit 8 and propagate further. [Figure 4] Flowchart of the method for refining and / or determining crystal structure according to the present invention. [Diagram 5]Diagram of the natrolite crystal used to determine the structure described in the Experimental Results section, with illuminated areas circled. The ring diameter is 750 nm. [Figure 6] Diagram of one of the abiraterone acetate crystals used to determine the structure described in the Experimental Results section, with illuminated regions circled. The ring diameter is 850 nm. [Figure 7] Structural formula of abiraterone acetate molecule DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0054] A corresponding embodiment of an apparatus for carrying out the method according to the invention will now be described with reference to FIG.
[0055] Fig. 3 shows a schematic representation of a part of a transmission electron microscope further connected to a data storage and processing unit 8 using the determination or refinement method according to the invention. Fig. 3 further shows a transmission electron microscope with an electron source 1, which can be for example an electron gun. The electron source 1 generates a primary electron beam 2. The electron beam 2 passes through a system of lenses 3 which focus the primary electron beam 2 into a focused electron beam 4, which is directed towards a crystal 5 capable of electron diffraction. The source 1 of the primary electrons 2, the lenses 3 and preferably also the tilt of the crystal 5 can be controlled by a control system. The scattered electrons 6 are detected by a detector 7 which is further connected to the data storage and processing unit 8. The data storage and processing unit 8 further comprises a database and a computer program for processing the data to determine and / or refine the crystal structure of the measurement crystal 5 by the method according to the invention.
[0056] In parallel with the measuring step or following the measuring step, the calculation unit 8 can generate an initial inaccurate model of the crystal structure by analyzing the data obtained. In another embodiment, the computer program and the database in the unit 8 can contain information on the structural model of the crystal 5 deduced from other sources such as X-ray diffraction. More specifically, the database can further contain information on the crystal structure of the crystal 5 determined by X-ray diffraction and / or theoretically predicted models. The structural model of the crystal structure 5 obtained by these methods can be incomplete and / or inaccurate, and therefore the technical challenge is to refine this initially inaccurate structural model of the crystal structure 5. The refined structural model of the crystal 5 can be used to understand the physical, chemical and pharmaceutical properties of materials and is therefore used whenever knowledge of the properties of materials is useful or necessary. The use of rotational electron diffraction offers advantages over precession electron diffraction, especially when the experimental setup is simple, the experimental equipment is inexpensive and the process of the experiment itself is simple. Thanks to the speed of data acquisition and the efficient use of the electron dose, rotational electron diffraction is also suitable for the study of materials that are very sensitive to electron irradiation. Dynamic diffraction theory is used to refine the crystal structure of crystal 5. In one embodiment, the Bloch wave method for calculating the diffraction intensities can be used to refine the structure estimation using dynamic diffraction theory. This method is described, for example, in the text JCH Spence and JM Zuo, Electron microdiffraction, Plenum press, 21992, p.35.
[0057] In this way, the intensity of the scattered electrons 6 is determined according to a given model of the crystal structure of the crystal 5. Alternatively, the diffraction intensity can be calculated using the dynamical diffraction theory using the multislice method, which gives results equivalent to the Bloch wave method. The multislice method is described in more detail in the text, for example, DB Williams and CB Carter, Transmission Electron Microscopy, 2nd Edition, Springer, 2009, p. 533.
[0058] The calculated intensity values are then compared to the measured data, in particular the scattered electron intensity 6.
[0059] Intensity calculated during rotational data acquisition in a particular embodiment TIFF2024540800000008.tif5114
[0060] can be obtained by integrating all calculated intensities at each possible orientation of the crystal 5 into one virtual diffraction frame. The integration can be performed numerically as the sum of the intensities calculated at a finite number of crystal orientations 5.
[0061] The refinement of the structure mainly consists of determining the structural parameters such as the positions of the atoms, the types of atoms and their scattering parameters. These parameters are modified in order to minimize the difference between the calculated and the experimental intensities. The calculation may also include parameters related to the crystal and its orientation, i.e. the thickness of the crystal 5 and the mutual orientation of the primary electron beam 4 and the crystal 5. The calculation is carried out based on the model intensities TIFF2024540800000009.tif6114
[0062] It is also affected by parameters that affect the calculation of
[0063] A least squares method can be used to refine the crystal structure model of the crystal 5. In one embodiment, a standard Gauss-Newton algorithm can be used. This approach is sufficient for small residual problems where the starting point is close to the solution, i.e. the model is nearly accurate.
[0064] Determining the quality of the refined model comprises verifying the correspondence between the refined model of the crystal structure and the observed quantities using refinement quality indicators. In one embodiment, the validation of the crystal structure model can be evaluated using one and / or the entire set of refinement quality indicators. These indicators are subject to the following formula:
[0065]
number
[0066] Where: TIFF2024540800000011.tif6114
[0067] each, TIFF2024540800000012.tif6114
[0068] are observed, which are respectively the calculated intensities of a particular diffracted electron beam 5 characterized by a diffraction vector g, TIFF2024540800000013.tif6114
[0069] and TIFF2024540800000014.tif6114
[0070] The following decision values: TIFF2024540800000015.tif5114
[0071] is the standard deviation of
[0072] If the coefficient wR2 reaches a minimum value, the crystal structure is considered refined. The R coefficient is also sensitive in determining the correct absolute crystal structure. By comparing the coefficient R of two variants of the absolute structure and selecting the variant with the smaller value of R, it is possible to unambiguously determine which of the two variants is correct. This makes it possible to determine the absolute structure of the crystal and therefore the absolute configuration of the molecules contained in the crystal.
[0073] Thus, the present embodiments provide a method for determining or refining the crystal structure of a crystal, typically with an accuracy better than 0.05 Å.
[0074] The method according to the present invention will now be described in detail in accordance with the best mode of operation of the invention which is contemplated, and those skilled in the art will further appreciate that the general approach shown in FIG.
[0075] Function of the invention Experimental Results The present invention has been tested with a number of materials, two of which are listed here: inorganic crystals of sodium silicate (the mineral natrolite) and organic crystals of abiraterone acetate.
[0076] Natrolite Natrolite is Na2(Al2Si3O 10 It is a mineral with the chemical composition ZnO(H2O). Its structure has the space group Fdd2, which means that it is reasonable to determine the absolute structure for this material.
[0077] A sample of natural natrolite from Mariánská Skála, Ústí nad Labem, Czech Republic, was ground to a fine powder in an agate mortar. The powder was applied to a copper grid coated with a carbon film. The grid was placed in a transmission electron microscope specimen holder and placed in an FEI Tecnai G2 20 microscope equipped with a LaB6 electron source and a SIS Veleta CCD detector at an accelerating voltage of 200 kV.
[0078] The electron diffraction data from the crystals was obtained by the three-dimensional electron diffraction method as follows: Several crystals were tested and suitable measurement candidates were selected by visually inspecting their diffraction patterns. The crystal goniometer was rotated to the -50° position. The crystal was rotated 0.6° at an angular velocity of 0.3° per second and its diffraction pattern was recorded on the detector throughout the rotation. The experimental diffraction pattern obtained was stored on the computer hard disk. This diffraction pattern procedure was repeated with an additional rotation of 0.6°. The total rotation of the crystal was 99.6°, thus obtaining a total of 166 experimental diffraction patterns.
[0079] The obtained data was analyzed using the computer program PETS2 (<http: / / pets.fzu.cz> ) The data processing included the following steps (for details see e.g. Palatinus, L., Brazda, P., Jelinek, M., Hrda, J., Steciuk, G. & Klementova, M. (2019) Specifics of the data processing of precession electron diffraction tomography data and their implementation in the PETS2.0 program, Acta Cryst.B75, 512-522): - finding the maxima of all the diffraction patterns and determining their coordinates in reciprocal space (see Figure 1); - determining the orientation matrix, i.e. finding the parameters of the crystal lattice and determining the orientation of the crystal in space. The determined lattice parameters were a=18.273 Å, b=18.646 Å, c=6.617 Å, α=90°, β=90°, γ=90°; - determining the diffraction intensities of all potentially excited reciprocal lattice points on all experimental diffraction patterns (see Figure 1); - generating a virtual diffraction frame, whereby: · One virtual frame covers an angular range of 1.2° and was therefore generated by combining two experimental diffraction patterns (see Fig. 2); ·All points of the reciprocal lattice that lie within the angular range covered by the virtual diffraction frame are found.
[0080] For each point of the reciprocal lattice found, all diffraction intensities belonging to this point were summed over all experimental diffraction patterns.
[0081] The list of intensities thus obtained belonging to the points of the reciprocal lattice forms a data set called a virtual diffraction frame.
[0082] The orientation of the virtual diffraction frame was determined as the average of the orientations of the two experimental diffraction patterns that compose it.
[0083] Each subsequent virtual diffraction frame was shifted by one experimental diffraction pattern compared to the previous one. Thus, the angular overlap of two successive virtual diffraction frames was 0.6°.
[0084] The result of the whole procedure was a list of 48847 reflections on 165 virtual diffraction frames.
[0085] The output from the PETS program in the form of a list of reflections with refractive index, intensity, and standard deviation was used to find and refine the crystal structure model using the Jana2006 program (<http: / / jana.fzu.cz> ) The method comprises the steps of: - Finding an approximate structural model that contains a list of atoms in the structure and their approximate positions - The dynamic diffraction theory was used to calculate the theoretical diffraction intensities required for refinement, while the step refinement, refining the structural model by the least squares method, resulted in residual coefficients R1=0.0947 and wR2=0.0968.
[0086] - determining the correct absolute structure, comprising the steps of: generating an inverted structural model, i.e. a model in which the coordinates of all atoms in the structure are replaced by coordinates having the same absolute value but opposite signs, A step of refining the structural model similar to the refinement of previous models.
[0087] The refinement of the inverted model resulted in residual coefficients R1 = 0.1559 and wR2 = 0.1663. These values are more than 0.01 greater than those of the non-inverted model, so this procedure clearly determines that the original non-inverted model corresponds to the correct absolute structure.
[0088] The resulting refined structural model was compared to the known reference structure of natrolite, determined by X-ray single crystal diffraction, with an average difference in interatomic distances of 0.0125 Å.
[0089] Abiraterone acetate Abiraterone acetate is an organic substance with medicinal properties. Its chemical formula is C 26 H 33 NO2. The molecule of abiraterone acetate is shown in Figure 7. The molecule is chiral and crystallizes in the P212121 space group, which means that it makes sense to determine the absolute structure of this material.
[0090] A white powder of abiraterone acetate was dissolved in distilled water. After 1 min, a drop of the solution was placed on a carbon-coated copper grid. The grid was placed in a cooled transmission electron microscope specimen holder and placed in an FEI Tecnai G2 20 microscope with a LaB6 electron source and a SIS Veleta CCD detector at an accelerating voltage of 200 kV. The sample was cooled to 100 K.
[0091] The data was acquired as follows: The crystals were examined and their diffraction patterns were visually inspected to select suitable crystals for measurement. A total of five crystals were selected for measurement and further processing. For the first crystal, the goniometer containing the crystal was rotated to the -28.3° position. The crystal was then rotated 0.4° at an angular velocity of 0.465° per second, and its diffraction pattern was recorded on the detector throughout the rotation. The experimental diffraction pattern obtained was stored on the computer's hard disk. This diffraction pattern procedure was repeated with an additional rotation of 0.4°. The total rotation of the crystal was 80°, thus obtaining a total of 200 experimental diffraction patterns.
[0092] The same procedure was repeated for the other four crystals. The rotation speed and rotation angle range in one experiment remained the same. The total range of rotation was 87.2°, 78.4°, 50.4°, and 54.0° for crystals 2, 3, 4, and 5, respectively.
[0093] Data obtained from each crystal were analyzed using the computer program PETS2 (<http: / / pets.fzu.cz> ) The data processing involved the following steps (for details see, e.g., Palatinus, L., Brazda, P., Jelinek, M., Hrda, J., Steciuk, G. & Klementova, M. (2019) Specifics of the data processing of precession electron diffraction tomography data and their implementation in the PETS2.0 program, Acta Cryst.B75, 512-522): - finding the maxima of all the diffraction patterns and determining their coordinates in reciprocal space (see Figure 1); - determining the orientation matrix, i.e. finding the parameters of the crystal lattice and determining the orientation of the crystal in space. The determined lattice parameters were a=7.470 Å, b=9.689 Å, c=30.20 Å, α=89.96°, β=89.98°, γ=89.99°; - determining the diffraction intensities of all potentially excited reciprocal lattice points on all experimental diffraction patterns (see Figure 1); - generating a virtual diffraction frame, whereby: · One virtual frame covers an angular range of 2° and was therefore generated by the combination of five experimental diffraction patterns (see Figure 2); ·All points of the reciprocal lattice that lie within the angular range covered by the virtual diffraction frame are found.
[0094] For each reciprocal lattice point found, all diffraction intensities belonging to this point were summed over all experimental diffraction patterns.
[0095] The list of intensities thus obtained belonging to the points of the reciprocal lattice forms a data set called a virtual diffraction frame.
[0096] The orientation of the virtual diffraction frame was determined as the average of all the orientations of the five experimental diffraction patterns that compose it.
[0097] Each subsequent virtual diffraction frame was shifted from the previous one by two experimental diffraction patterns for crystals 1, 2, and 3, and by three experimental diffraction patterns for crystals 4 and 5. Thus, the angular overlap of two successive diffraction patterns was 1.2° and 0.8°, respectively.
[0098] The result of the whole procedure was a list of 10957 reflections on 285 virtual diffraction frames.
[0099] The output from the PETS program in the form of a list of reflections including their refractive indices, intensities, and standard deviations was used in the Jana2006 program (<http: / / jana.fzu.cz> ) All data were read together and the structure model was refined against the data from all five crystals. The method involves the following steps: - Finding an approximate structural model that contains a list of atoms in the structure and their approximate positions - The dynamic diffraction theory was used to calculate the theoretical diffraction intensities required for refinement, while the step refinement, refining the structural model by the least squares method, resulted in residual coefficients R1=0.1242 and wR2=0.1354.
[0100] Determining the correct absolute structure involves the following steps: - generating an inverted structural model, i.e. a model in which the coordinates of all atoms in the structure are replaced by coordinates having the same absolute value but opposite signs; - A step of refining the structural model similar to the refinement of previous models.
[0101] - Refinement of the inverted model resulted in residual coefficients R1 = 0.1560 and wR2 = 0.1721, which are more than 0.01 greater than those of the non-inverted model, making this procedure a clear determination that the original non-inverted model corresponds to the correct absolute structure.
[0102] The resulting refined structural model was compared to the known reference structure of abiraterone acetate, determined by X-ray single crystal diffraction, with an average difference in interatomic distances of 0.0493 Å.
[0103] Industrial Applicability The invention finds application in the field of computational crystallography, more precisely, it can be applied to the determination of both the crystal structures of inorganic crystals and of the structures of organic crystals, knowledge of the crystal structures being usable in many fields such as the metallurgical or pharmaceutical industries. [Explanation of symbols]
[0104] 1 electron source 2. Electron Beam 3 Lenses 4. Focused electron beam 5. Measuring Crystal 6 Scattered electrons 7. Detector 8 Processing Unit
Claims
1. 1. A method for determining a crystal structure model, the crystal being capable of electron diffraction, acquiring electron diffraction data from the crystal by three-dimensional electron diffraction, the data comprising information about crystal diffraction patterns, each diffraction pattern comprising information about scattered electron intensity and crystal orientation for each direction relative to an incident electron beam during data acquisition, the crystal being rotated or tilted during recording of the diffraction data; and processing the data to determine crystal lattice parameters, crystal orientations at the time of acquisition of the diffraction patterns, indices of each reflection, and scattered electron intensities of each reflection on each diffraction pattern, thereby generating processed data; generating virtual diffraction frames by a computer program, each virtual diffraction frame comprising a list of integrated scattered electron intensities in each direction summing contributions from all experimental patterns forming a virtual frame, said virtual diffraction frames being computationally generated from the provided three-dimensional electron diffraction data; Providing an approximate crystal structure model; and refining said approximate crystal structure model using dynamic diffraction theory wherein the refining step comprises: calculating a modeled intensity of scattered electrons using dynamical diffraction theory based on said approximate crystal structure model; minimizing the difference between the experimentally determined scattered electron intensities and the modeled intensities, the refined crystal structure model being the model with the least deviation; A method comprising:
2. A method for determining the absolute structure of a non-centrosymmetric crystal, comprising the steps of claim 1, generating a model of an absolute structure that is inverse to the current approximate crystal structure model, thereby generating an inverted model; In the inversion model refining said approximate crystal structure model using dynamic diffraction theory; and comparing quality indicators of the inverted and non-inverted models, the correct structure being the one which results in the smallest deviation between the experimental data and the data calculated based on the refined structural model; A method comprising:
3. 3. The method of claim 1 or 2, wherein the step of obtaining electron diffraction data from the crystal by three-dimensional electron diffraction is a measurement of continuous crystal rotations.
4. 4. The method of claim 1, wherein the step of obtaining electron diffraction data from the crystal by three-dimensional electron diffraction comprises rotating the crystal.
5. 5. The method of claim 1, wherein the step of minimizing the difference between the experimentally determined scattered electron intensity and the modeled intensity comprises a least-squares calculation.
6. The method of claim 1 , wherein the step of providing an approximate crystal structure model is based on analysis of experimental data.
7. An apparatus for determining a crystal structure model, comprising: an electron source capable of emitting an electron beam; a crystal holder adapted to perform a rotational or tiling movement with the crystal; A detector for detecting electrons scattered by the crystal, Data Storage and Processing Unit Detector connected to Including, the processing unit is adapted to determine crystal lattice parameters, crystal orientations during acquisition of the diffraction patterns, indices of each reflection, and scattered electron intensities of each reflection on each diffraction pattern, thereby generating the processed data; and generating virtual diffraction frames by a computer program, each virtual diffraction frame comprising a list of integrated scattered electron intensities in each direction summing contributions from all experimental patterns forming a virtual frame, said virtual diffraction frames being computationally generated from the provided three-dimensional electron diffraction data; The data storage is configured to provide an approximate crystal structure model or the processing unit is adapted to perform an analysis of the processed data leading to the acquisition of an approximate structure model; and the processing unit is adapted to refine the structural model; Device.
8. The processing unit is adapted to refine the approximate crystal structure model using dynamic diffraction theory, and during the refining, the processing unit: calculating a modeled intensity of scattered electrons using dynamic diffraction theory based on said approximate crystal structure model; and Minimize the difference between the experimentally determined scattered electron intensity and the modeled intensity where: the refined crystal structure model is the model with the smallest deviation; 8. The apparatus of claim 7.
9. The processing unit is configured to determine an absolute structure of a non-centrosymmetric crystal, the processing unit comprising: A model of an absolute structure opposite to the structure of the current approximate crystal structure model is generated, thereby generating an inverted model; refining the approximate crystal structure model using the dynamic diffraction theory of the inversion model; and Compare quality indicators of the inversion model and experimental data In this case, The correct absolute structure is the one with the smallest deviation between them.
9. Apparatus according to claim 7 or 8.
10. The processing unit is adapted to refine the approximate crystal structure model, and the apparatus further comprises: Calculating a quality index of the approximate crystal structure model; and Compare the quality indicators where: the correct crystal structure is taken as the modeled crystal structure which has the smallest deviation between the experimental data and the data calculated based on the refined structural model; 10. Apparatus according to any one of claims 7 to 9.
11. 11. The apparatus of claim 7, wherein the processing unit is adapted to generate a model of a crystal structure that is inverse to a current approximate crystal structure model, thereby generating an inverted model, and the comparing step is a step of comparing quality indicators of the inverted model and the non-inverted model, wherein the correct structure is the structure that has the smallest deviation between the experimental data and data calculated based on the refined structural model.
12. 12. Apparatus according to any one of claims 7 to 11, wherein the detector is configured to collect data simultaneously with the rotation of the crystal.
13. 13. Apparatus according to any one of claims 7 to 12, wherein the detector is configured to collect data stepwise during the rotation of the crystal.
14. 1. A method for determining a crystal structure model, preferably the absolute structure of a non-centrosymmetric crystal, said crystal being capable of electron diffraction, comprising: acquiring electron diffraction data from said crystal by three-dimensional electron diffraction, said data comprising information about crystal diffraction patterns, each diffraction pattern comprising information about scattered electron intensity and crystal orientation for each direction relative to an incident electron beam during data acquisition, said crystal being rotated or tilted during data acquisition; and processing said data to determine crystal lattice parameters, crystal orientations at the time of acquisition of the diffraction patterns, indices of each reflection, and scattered electron intensities of each reflection on each diffraction pattern, thereby producing processed data; generating virtual diffraction frames by a computer program, each virtual diffraction frame comprising a list of integrated scattered electron intensities in each direction summing contributions from all experimental patterns forming a virtual frame, said virtual diffraction frames being computationally generated from the provided three-dimensional electron diffraction data; Providing an approximate crystal structure model; calculating a quality index of the approximate crystal structure model using dynamic diffraction theory and calculating model strength; generating a model of an absolute structure that is inverse to the current approximate crystal structure model, thereby generating an inverted model; calculating a quality index of the inverted crystal structure model using dynamic diffraction theory and calculating model strength; and comparing quality indicators of the inverted and non-inverted models, the correct structure being the one which results in the smallest deviation between the experimental data and the data calculated based on the refined structural model; A method comprising:
15. 15. The method of claim 14, wherein the step of obtaining electron diffraction data from the crystal by three-dimensional electron diffraction is a measurement of continuous crystal rotations.
16. 16. The method of claim 14, wherein the step of obtaining electron diffraction data from the crystal by three-dimensional electron diffraction comprises rotating the crystal.
17. 17. The method of any one of claims 14 to 16, wherein the step of providing an approximate crystal structure model is based on analysis of experimental data.
Citation Information
Patent Citations
Scanning transmission electron microscope and method for high throughput acquisition of electron scattering angle distribution images
EP3379557A1
Analysis method, analyzer, and program
JP2014119434A
Scanning Transmission Electron Microscope
US20180269031A1
A computer-implemented method and system for measurement of absolute structure of a crystal
WO2021093908A1
