High-throughput three-dimensional X-ray imaging system using a transmission X-ray source

Abstract

The X-ray imaging system includes at least one detector and an X-ray source including an X-ray transmission vacuum window. The X-ray source is configured to produce divergent X-rays, which exit the vacuum window and propagate along an extended X-ray propagation axis, through a region of interest of an object, to at least one detector. The divergent X-rays have a propagation path within an angular divergence angle of more than 1 degree about the X-ray propagation axis. The system further includes at least one first motion stage configured to rotate the object about a rotation axis. The system further includes at least one second motion stage configured to move the X-ray source and at least one detector relative to the object to switch between a laminography configuration and a tomography configuration.

Description

Technical Field

[0001] Claim of Priority The present application is a continuation-in-part of U.S. Patent Application No. 18 / 311,558 filed on May 3, 2023, the entire content of which is incorporated herein by reference, and is also a continuation of U.S. Patent Application No. 17 / 540,608 filed on December 2, 2021, the entire content of which is incorporated herein by reference, and claims the benefit of priority from U.S. Provisional Application No. 63 / 122,354 filed on December 7, 2020 and U.S. Provisional Application No. 63 / 274,367 filed on November 1, 2021, each of which is incorporated herein by reference in its entirety. The present application also claims the benefit of priority from U.S. Provisional Application No. 63 / 517,448 filed on August 3, 2023, the entire content of which is incorporated herein by reference.

[0002] Background field The present application generally relates to tomographic and laminographic X-ray imaging systems. Background Art

[0003] Description of related technologies Three-dimensional (3D) X-ray imaging technology is useful for imaging the internal structure of objects. Generally, a tomographic dataset consisting of X-ray transmission images collected over a large angular range (e.g., about 180 degrees, about 360 degrees) is then reconstructed to obtain a 3D image. A large angular range is used to avoid (e.g., minimize) 3D image artifacts. A three-dimensional X-ray imaging system comprises an X-ray source configured to irradiate an object for imaging, a position-sensitive X-ray detector configured to record transmitted X-ray images, and an electromechanical system for manipulating the object relative to the X-ray source and the position-sensitive X-ray detector.

[0004] The X-ray flux incident on an object's region of interest is inversely proportional to the square of the distance of the region of interest from the X-ray source, and this distance can be called the Focus-Object Distance (FOD). To achieve high throughput in three-dimensional X-ray imaging, the FOD is selected to be small (for example, by placing the region of interest as close to the X-ray source as possible). For example, considering the small voxel volume used to achieve spatial resolution, placing the region of interest close to the X-ray source can be used in high-spatial-resolution three-dimensional X-ray imaging using laboratory microfocus X-ray sources. Furthermore, in the case of three-dimensional X-ray imaging of a small region of interest within a larger object (for example, a small region of interest within a laterally extended planar object, such as interconnects in a semiconductor integrated circuit (IC) package and fine structural details in a large fiber-reinforced composite panel, but not limited to these), the minimum FOD is limited by the dimensions of the object, and the object should be rotated 180 degrees, which substantially limits the achievable throughput.

[0005] However, prior art micro X-ray computed tomography (μXCT) and micro X-ray computed laminography (μXCL) systems have many limitations. For example, Figures 1A and 1B schematically show conventional tomography and laminography configurations, respectively, for laterally extended objects (e.g., printed circuit boards, wafers) where the region of interest (ROI) is at or near the center of the object. The X-ray source emits an X-ray beam (indicated by horizontal dashed lines), and the thickness of the object perpendicular to the plane of the paper in Figures 1A and 1B can be equal to or less than the dimensions of the object along the X-ray beam (e.g., so that there is sufficient space between the X-ray source and the X-ray detector to rotate the object around the axis of rotation). As seen in Figure 1A, in the case of tomography, the axis of rotation is substantially parallel to the surface normal of the object and substantially perpendicular to the X-ray beam. As seen in Figure 1B, in laminography, the axis of rotation is substantially parallel to the surface normal of the object and tilted by an angle β from a direction substantially perpendicular to the X-ray beam. Figures 1A and 1B show that such conventional tomography and laminography configurations are not suitable for imaging defects in a plane parallel to the surface of laterally extended objects (e.g., semiconductor IC packages) because the transmitted X-ray spectrum in the tomography / laminography dataset changes with the angle of the X-ray beam axis relative to the object. As a result, reconstructed images (e.g., computed tomography or CT images) are affected by radiation tolerance and photon-deficient artifacts, leading to a dependency of the fidelity of reconstructed features on their orientation relative to the object's rotation axis. The resolution and image quality of reconstructed images in a plane parallel to the surface of laterally extended objects are typically worse than those in the direction along the surface normal. In addition, neither technique is optimized to achieve a small FOD due to physical interference between the object and the X-ray source.The improvement in image resolution by reducing the X-ray source size with high flux, which may be desired in many applications, is severely limited in such prior art XCT and XCL systems and methods, and higher depth resolution along the surface normal of such extended objects and / or planar objects is particularly difficult to achieve. [Overview of the initiative] [Means for solving the problem]

[0006] overview In a particular implementation, a three-dimensional X-ray imaging system is configured to generate a transmission image of a region of interest within an object. The system comprises at least one position-sensitive X-ray detector having at least one active element. The system further comprises an X-ray source comprising an X-ray-transmitting vacuum window having an outer surface. The X-ray source is configured to produce divergent X-rays, at least some of which exit the vacuum window and propagate along an X-ray propagation axis extending from the X-ray source, through the region of interest of the object, to the at least one active element of the at least one position-sensitive X-ray detector. The divergent X-rays have a propagation path within an angular divergence angle of more than 1 degree with respect to the X-ray propagation axis. The X-ray propagation axis is at a first angle with respect to the outer surface of the vacuum window, and the first angle is in the range of 3 to 45 degrees. The system further comprises at least one sample motion stage configured to rotate the object about a rotation axis, wherein the rotation axis is configured to have a second angle with respect to the X-ray propagation axis, and the second angle is in the range of 45 to 90 degrees. The system further comprises a sample mount on the at least one sample motion stage. The sample mount is configured to hold the object and comprises a first portion located in at least some of the propagation paths of the divergent X-rays that propagate through the object to the at least one position-sensitive X-ray detector. The first portion has an X-ray transmittance of more than 30% for X-rays having an energy greater than 50% of the maximum X-ray energy of the X-ray spectrum of the divergent X-rays.

[0007] In a particular implementation, the three-dimensional X-ray imaging system comprises at least one position-sensitive X-ray detector. The system further comprises an X-ray source comprising an X-ray-transmitting vacuum window having an outer surface. The X-ray source is configured to produce divergent X-rays, at least some of which exit the vacuum window and propagate along an X-ray propagation axis extending from the X-ray source. The divergent X-rays propagate through a region of interest of an object to the at least one position-sensitive X-ray detector and have an angular divergence angle of more than 1 degree about the X-ray propagation axis. The X-ray propagation axis is at a first angle with respect to the outer surface of the vacuum window, and the first angle is in the range of 3 to 45 degrees. The system further comprises at least one sample motion stage configured to rotate the object about a rotation axis and to adjust the rotation axis to have a second angle with respect to the X-ray propagation axis, wherein the at least one sample motion stage has an unsystematic angular runout of less than 5 microradians.

[0008] In a particular implementation, the X-ray imaging system is configured to generate an X-ray image of a region of interest within an object. The system comprises at least one X-ray detector and an X-ray source having a transmissive vacuum window with an outer surface. The X-ray source is configured to produce divergent X-rays, at least some of which exit the vacuum window and propagate along an X-ray propagation axis extending from the X-ray source, through the region of interest of the object, to the at least one X-ray detector. The divergent X-rays received by the at least one X-ray detector have a propagation path within an angular divergence angle of more than 1 degree around the X-ray propagation axis. The system further comprises at least one first motion stage configured to move the object relative to the X-ray source and / or rotate the object about a rotation axis. The system further comprises at least one second motion stage configured to move the X-ray source and the at least one X-ray detector relative to the object to switch between a laminography configuration in which the X-ray propagation axis has a first non-zero angle with respect to the rotation axis and a tomography configuration in which the X-ray propagation axis has a second non-zero angle with respect to the rotation axis, wherein the second angle is different from the first angle.

[0009] In a particular implementation, the X-ray imaging system is configured to generate an X-ray image of a region of interest within an object. The system comprises at least one X-ray detector and an X-ray source configured to produce divergent X-rays. At least some of the divergent X-rays propagate along an X-ray propagation axis extending from the X-ray source, through a three-dimensional field of view (3D FOV) within the object, to the at least one X-ray detector. The divergent X-rays have a propagation path within an angular divergence angle of more than 1 degree with respect to the X-ray propagation axis. The system further comprises an optical microscope having a focus aligned with the X-ray source and configured to overlap with at least a portion of the 3D FOV.

[0010] In a particular implementation, the X-ray imaging system is configured to generate a three-dimensional X-ray image of a region of interest within an object. The system comprises at least one X-ray detector having at least one active element, and an X-ray source having a transmissive vacuum window with an outer surface. The X-ray source is configured to produce divergent X-rays, at least some of which exit the vacuum window and propagate along an X-ray propagation axis extending from the X-ray source, through the region of interest in the object, to at least one active element of the at least one X-ray detector. The divergent X-rays have a propagation path within an angular divergence angle of more than 1 degree with respect to the X-ray propagation axis. The system further comprises at least one sample motion stage configured to rotate the object about a rotation axis, the rotation axis being configured to have a non-zero angle with respect to the X-ray propagation axis. The above-mentioned X-ray source and the above-mentioned at least one X-ray detector are configured to remain stationary while the above-mentioned at least one sample stage rotates the object during data acquisition for the above-mentioned three-dimensional X-ray image. [Brief explanation of the drawing]

[0011] [Figure 1] Figures 1A and 1B schematically show conventional tomographic and laminographic configurations, respectively, of a laterally extended object where the region of interest (ROI) is at or near the center of the object. [Figure 2] Figure 2 schematically shows an exemplary X-ray 3D imaging system that conforms to a specific implementation described herein. [Figure 3] Figure 3A schematically shows a cross-sectional view of an X-ray spot generated by an electron beam impacting at least one X-ray target, according to a specific implementation described herein. Figures 3B and 3C schematically show top views of two exemplary configurations of the electron beam and at least one X-ray target in Figure 3A, according to a specific implementation described herein. [Figure 4] Figure 4 schematically shows another exemplary system comprising at least one aperture, relating to a specific implementation described herein. [Figure 5A] Figure 5A schematically shows an exemplary system relating to a specific implementation described herein, in which the sample stage and sample mount are configured to hold and rotate an object. [Figure 5B] Figure 5B schematically shows an exemplary system relating to a specific implementation described herein, in which the sample stage and sample mount are configured to hold and rotate an object. [Figure 6A] Figure 6A schematically shows an example of a system relating to a specific implementation described herein, in which at least one position-sensitive X-ray detector comprises a first position-sensitive X-ray detector and a second position-sensitive X-ray detector. [Figure 6B] Figure 6B schematically shows an example of a system relating to a specific implementation described herein, in which at least one position-sensitive X-ray detector comprises a first position-sensitive X-ray detector and a second position-sensitive X-ray detector. [Figure 7] Figure 7 schematically shows an exemplary system configured to perform multi-contrast X-ray imaging using Talbot interferometry or Talbot-Rau interferometry, relating to a specific implementation described herein. [Figure 8] Figure 8 schematically shows an exemplary system adapted for generating a translated laminography image, relating to a specific implementation described herein. [Figure 9] Figure 9 schematically shows an exemplary three-dimensional X-ray imaging system configured to generate a three-dimensional X-ray image of a region of interest within an object, according to a specific implementation described herein. [Figure 10A] Figure 10A schematically shows another exemplary X-ray imaging system in an exemplary laminography configuration relating to a specific implementation described herein. [Figure 10B] Figure 10B schematically shows another exemplary X-ray imaging system in an exemplary tomography configuration relating to a specific implementation described herein. [Figure 11] Figure 11 schematically shows an exemplary vacuum window for a transmission X-ray source relating to a specific implementation described herein. [Figure 12] FIG. 12A schematically illustrates a portion of an example system according to certain implementations described herein. FIG. 12B schematically illustrates an example X-ray source having a cone angle according to certain implementations described herein. [Figure 13] FIG. 13A schematically illustrates a side view of an example system including at least one example X-ray detector including a first X-ray detector and a second X-ray detector according to certain implementations described herein. FIG. 13B schematically illustrates the first and second X-ray detectors of FIG. 13A viewed along a direction substantially parallel to the X-ray propagation axis. [Figure 14] FIG. 14 schematically illustrates an example three-dimensional X-ray laminography system including an optical microscope according to certain implementations described herein. MODE FOR CARRYING OUT THE INVENTION

[0012] Detailed Description In certain implementations, a three-dimensional X-ray imaging system enables imaging of ROIs with very small FODs within large objects to reduce laminography dataset acquisition time (e.g., to improve imaging data acquisition speed). In certain implementations, the system also provides improved image quality (e.g., fidelity) and image resolution in a plane parallel to the laterally extended surface of the object, which can be important for many applications such as measurement, inspection, failure analysis, and process development of semiconductor IC packages (e.g., solder bumps and Cu interconnects). The X-ray source may be configured to achieve a small FOD and to improve spatial resolution. In certain implementations, the electron beam focus is configured to be "compressed" into one dimension using measurement geometry, thereby enabling the use of asymmetric, larger, and therefore higher-power focus that is symmetric or nearly symmetric after projection, which would be desirable for having isotropic spatial resolution. In certain implementations, the X-ray source is configured to include an X-ray generating material having a thickness configured to achieve high spatial resolution in a direction substantially perpendicular to the laterally extended surface of the object being analyzed. In addition, certain implementations include additional components and / or methods for realizing several modes of imaging contrast, including Talbot interferometry for obtaining absorption, phase, and dark-field (scattering) contrast, dark-field (scattering) contrast only, and enhanced absorption contrast.

[0013] FIG. 2 schematically illustrates an exemplary X-ray 3D imaging system 5 compatible with a specific implementation described in the present specification. The system 5 comprises an X-ray source 20 and at least one position-sensitive X-ray detector 50 comprising at least one active element 52. The X-ray source 20 comprises an X-ray transmissive vacuum window 29 having an outer surface 27, and the X-ray source 20 is configured to generate divergent X-rays. At least some of the divergent X-rays 60 exit from the vacuum window 29, propagate along the X-ray propagation axis 10 extending from the X-ray source 20, through the region of interest 31 of the object 30, to at least one active element 52 of the at least one position-sensitive X-ray detector 50. The divergent X-rays 60 incident on the at least one active element 52 have propagation paths within an angular divergence 12 of more than 1 degree centered on the X-ray propagation axis 10. The X-ray propagation axis 10 is at a first angle 11 relative to the outer surface 27 of the vacuum window 29, and the first angle 11 is within a range of 45 degrees or less (for example, 30 degrees or less, a range of 3 degrees to 45 degrees, between 5 degrees and 30 degrees, less than 3 degrees).

[0014] In certain implementations, System 5 further comprises at least one sample motion stage 80 (e.g., electrically powered and computer-controlled, comprising an electromechanical system). The at least one sample motion stage 80 is configured to rotate an object 30 about a rotation axis 19. The sample motion stage 80 is configured such that the rotation axis 19 has a second angle 16 with respect to the X-ray propagation axis 10, and the second angle 16 is in the range of 45 degrees or more (e.g., from 45 to 90 degrees). In certain other implementations, the second angle 16 is less than 45 degrees. In certain implementations, the second angle 16 may be in the range of 45 degrees or more, and the rotation axis 19 may be at a third angle with respect to the surface normal of the outer surface 27 of the vacuum window 29, and the third angle is in the range of less than 45 degrees (e.g., less than 30 degrees). In certain implementations, System 5 comprises a mechanism configured to change the third angle. For example, the mechanism may include at least one tilting stage (e.g., a goniometer, an electromechanical motion driver, a rotary motor, a stepping motor, a motor with an encoder, a linear motion driver with a worm drive) configured to tilt the X-ray source 20 with respect to the rotation axis 19 and / or the rotation axis 19 with respect to the X-ray source 20.

[0015] In a particular implementation, System 5 further comprises a sample mount 85 configured to hold an object 30 on at least one sample motion stage 80. The sample mount 85 comprises a first portion 86 that lies in at least some propagation paths of divergent X-rays 60 propagating through the object 30 to at least one position-sensitive X-ray detector 50. The first portion 86 has an X-ray transmittance of more than 30% (e.g., more than 50%) for X-rays having energies greater than 50% of the maximum X-ray energy in the X-ray spectrum of the divergent X-rays 60.

[0016] The exemplary system 5 in Figure 2 is configured to image a region of interest (ROI) 31 within a large or laterally extended object 30 (e.g., positioned substantially parallel to the outer surface 27 of the vacuum window 29) to minimize the foreseeable object disk (FOD) between the ROI 31 and the X-ray source 20. For example, to image an ROI 31 within a large three-dimensional object 30, the object 30 may be positioned close to the X-ray source 20 (e.g., the distance between the outer surface 27 of the vacuum window 29 and the surface 32 of the object 30 is less than 70 millimeters) and centered on the rotation axis 19 of at least one sample motion stage 80. A three-dimensional (3D) image dataset of the ROI 31 can be collected by recording a series of X-ray transmission images of divergent X-rays 60 transmitted through the ROI 31 of the object 30 to at least one active element 52 of at least one position-sensitive X-ray detector 50 while rotating the object 30 around the rotation axis 19 over a range of angles (e.g., between 180 and 360 degrees). To obtain a 3D image of ROI31, a computer laminography dataset can be reconstructed using known laminography reconstruction methods. In another example, to image ROI31 within a planar object 30 (e.g., solder bumps and / or interconnects in a semiconductor IC package), the object 30 can be positioned such that the surface 32 of the object 30 closest to ROI31 faces the vacuum window 29. For large / planar objects 30, the ROI31 can be positioned closer to the X-ray source 20 by using a small angle 14 between the outer surface 27 of the vacuum window 29 and the surface 32 of the object 30 to improve imaging throughput by increasing the X-ray flux to ROI31. In another example, the surface 32 of the object 30 faces away from the outer surface 27 of the vacuum window 29 (e.g., to reduce or minimize the dose of radiation to ROI31).

[0017] x-ray source As schematically shown in Figure 2, in a particular implementation, the X-ray source 20 comprises a vacuum chamber 21 including a vacuum region 23 and an electron beam source 22 within the vacuum region 23. The X-ray source 20 further comprises an electron optics system 25 (e.g., electrodes) configured to direct at least some electrons from the electron beam source 22 into an electron beam 24 focused onto at least one X-ray target 28. For example, the electron beam source 22 and the electron optics system 25 are configured to generate a focused electron beam 24 and irradiate at least one X-ray target 28 with the focused electron beam 24, with a selectable maximum focused electron energy in the at least one X-ray target 28 in the range of 10 kVp to 250 kVp. In a particular implementation, the vacuum chamber 21 comprises a vacuum-sealed tube housing the electron beam source 22, the electron optics system 25, and at least one X-ray target 28. In contrast to an open-tube X-ray source, the X-ray source 20 in a particular implementation is not actively evacuated.

[0018] At least one X-ray target 28 is located within a vacuum region 23 and is configured to generate divergent X-rays 60 in response to irradiation by a focused electron beam 24. The at least one X-ray target 28 comprises at least one X-ray generating material selected for its X-ray spectral generation characteristics (e.g., characteristic X-ray energy) and / or other properties (e.g., atomic number Z, electron density) that affect the X-ray generation capability of at least one X-ray generating material. The at least one X-ray generating material may have a sufficiently high thermal conductivity to dissipate the heat generated by irradiation by a high-power electron beam 24. Examples of X-ray generating materials include, but are not limited to, Cr, Fe, Co, Ni, Cu, W, Rh, Mo, Au, Pt, Ag, SrB6, LaB6, and CeB6. As schematically shown by the inset in Figure 2, at least one X-ray target 28 can be attached to (e.g., integrated with, a component of, or in contact with) a vacuum window 29, the vacuum window 29 separating a vacuum region 23 from a non-vacuum region outside the X-ray source 20. The thickness of at least one X-ray generating material along a direction substantially perpendicular to the outer surface 27 of the vacuum window 29 can be in the range of 0.1 to 15 microns (e.g., 0.1 to 10 microns), and the thickness of the vacuum window 29 in a direction substantially perpendicular to the outer surface 27 of the vacuum window 29 can be in the range of 0.05 to 3 millimeters. As described herein, the thickness of at least one X-ray generating material may be configured to be optimized for high spatial resolution (e.g., by minimizing electron beam scattering within the material) and / or high system throughput (e.g., by maximizing electron energy deposition within the at least one X-ray generating material). For example, the thickness of at least one X-ray generating material may be less than twice the image resolution along the rotation axis 19. In a particular implementation, at least one X-ray generating material has a plurality of regions that can be irradiated by the electron beam 24 (for example, by shifting the electron beam focus), each region having a corresponding thickness along a direction substantially perpendicular to the outer surface 27 of the vacuum window 29.The electron beam 24 can be directed by the electron optics system 25 to illuminate a selected region having a corresponding thickness that provides a chosen trade-off between throughput and resolution.

[0019] In a particular implementation, the vacuum window 29 is essentially composed of atomic elements with atomic number (Z) less than 14 and is substantially transparent to higher-energy X-rays produced by at least one X-ray generating material. For example, the vacuum window 29 may have sufficiently high thermal conductivity to provide a heat conduit to prevent thermal damage (e.g., melting) to at least one X-ray generating material (e.g., one or more materials selected from the group consisting of beryllium, diamond, boron carbide, silicon carbide, aluminum, and beryllium oxide (BeO)). The vacuum window 29 can further provide sufficiently conductive paths for dissipating charge from at least one X-ray generating material and / or the vacuum window 29. In a particular implementation, the vacuum window 29 is configured to have X-ray transmittance such that more than 50% of X-rays produced by at least one X-ray source 28 having an energy greater than half of the selected maximum focused electron energy pass through the vacuum window 29. In a particular implementation, divergent X-rays 60 emitted from the outer surface 27 of the vacuum window 29 are not blocked by the X-ray source 20 along the X-ray propagation axis 10. The X-ray source 20 emits X-rays in a solid angle of 4π, but Figure 2 schematically shows only the divergent X-rays 60 (e.g., X-rays contributing to image formation) that propagate along the X-ray propagation axis 10 within the angular divergence angle 12 to at least one active element 52 of at least one position-sensitive X-ray detector 50, and other divergent X-rays generated by at least one X-ray target 28 and emitted in other directions from the outer surface 27 of the vacuum window 29 are not shown in Figure 2.

[0020] In certain implementations, System 5 further includes a thermal cooling mechanism configured to reduce heating of the object 30 due to heat generated by the X-ray source 20. For example, the thermal cooling mechanism may include an infrared (IR) reflective material (e.g., a thin IR-reflective and highly X-ray-transmissive film or layer, one example being aluminized Mylar) between the vacuum window 29 and the object 30. The IR-reflective material is configured to reflect heat generated by the X-ray source 20 (e.g., resulting from the conversion of electron beam power into heat at at least one X-ray target 28) away from the object 30 and the sample mount 85 (e.g., to move the thermal energy away from the object 30 and the sample mount 85 in order to reduce or minimize heat transport to the object 30 and / or the sample mount 85). In this way, the IR-reflective material can protect the object 30 and the sample mount 85 from temperature changes in the sample, which could otherwise cause thermal expansion that negatively affects the accuracy of ROI selection and / or the fidelity of 3D volume reconstruction. The IR reflective material is thin enough (e.g., less than 1500 microns thick, less than 100 microns thick) so as not to substantially hinder the positioning of the object 30 near the electron beam focus of the X-ray source 20.

[0021] In specific implementations, the X-ray source 20 comprises a grounded anode transmission X-ray source (e.g., one with an electrically grounded vacuum housing), examples of which include, but are not limited to, the DAGE BrightHawk® X-ray source available from Nordson Corporation in Westlake, Ohio; the L10711-03 microfocus X-ray source available from Hamamatsu Photonics Co., Ltd. in Hamamatsu, Japan; the Excillum Nanotube N1 and N2 X-ray sources available from Excillum Corporation in Kista, Sweden; and X-ray sources available from X-ray Worx GmbH in Galbsen, Germany, and COMET Technologies in San Jose, California. These exemplary X-ray sources 20 have an X-ray source point integrated with a vacuum window 29 and may be configured to position the object 30 close to the X-ray source 20 in order to reduce (e.g., minimize) the FOD and thereby increase (e.g., maximize) the X-ray flux in the object 30. Due to electrical and mechanical constraints, the X-ray source 20 may have a large flat surface, which is the end of the vacuum vessel and coplanar with the vacuum window 29, but cannot be reduced in size without impairing the quality of the electron beam focus. In existing X-ray tomography and laminography imaging techniques utilizing such an X-ray source 20, this large flat surface has limited the ability to place a large and / or planar object 30 close to the X-ray source 20 and to rotate the object 30 (e.g., up to 180 degrees) around a rotation axis 19 substantially perpendicular to the large flat surface. The specific implementations described herein advantageously circumvent this major drawback of existing X-ray imaging systems.

[0022] In a particular implementation, the X-ray source 20 is configured to have a small X-ray spot size (e.g., having dimensions of less than 7 microns in at least one lateral direction substantially parallel to the outer surface 27 of the vacuum window 29) while generating sufficient X-ray flux to facilitate a sufficiently short image acquisition time. Generally, the X-ray spot size is approximately equal to the convolution of the focused electron beam spot size (e.g., radius) in at least one X-ray target 28 and the size (e.g., radius) of the X-ray generation volume inside at least one X-ray target 28 due to electron scattering inside at least one X-ray target 28. Thus, a larger focused electron beam spot size can facilitate higher electron beam power, along with the accompanying higher X-ray flux and shorter image acquisition time, at the expense of lower spatial resolution, while a smaller focused electron beam spot size can facilitate higher spatial resolution, at the expense of lower X-ray flux and longer image acquisition time. In addition, since a large portion of the incident power from the focused electron beam 24 (e.g., about 99%) is converted into heat in at least one X-ray target 28, it may be desirable to limit the incident electron beam power, which typically decreases linearly with the X-ray spot size.

[0023] In a particular implementation, a higher spatial resolution of the X-ray transmission image in a plane including the plane normal of the outer surface 27 of the vacuum window 29 and the axis of rotation 19 is provided by reducing the thickness t of at least one X-ray generating material of at least one X-ray target 28, which reduces the effective X-ray source size s. For example, the thickness t of at least one X-ray generating material along a direction substantially perpendicular to the outer surface 27 of the vacuum window 29 can be in the range of 0.1 to 15 microns. Figure 3A schematically shows a cross-sectional view of an X-ray spot produced by an electron beam 24 impacting at least one X-ray target 28 according to a particular implementation described herein. Figures 3B and 3C schematically show top views of two exemplary configurations of the electron beam 24 and at least one X-ray target 28 in Figure 3A according to a particular implementation described herein. As schematically shown in Figure 3A, the X-ray spot size viewed along the X-ray propagation axis 10 can be smaller than the width W (e.g., diameter) of the electron beam 24 on the X-ray target 28. For an X-ray generating material having a thickness t, the effective width s of the full-width at half maximum (FWHM) of the X-ray spot as viewed along the X-ray propagation axis 10 may be less than approximately t / 2 (Figure 3A is not a constant proportional scale). In a particular implementation, at least one X-ray generating material comprises a thin high-Z material layer (e.g., with a thickness t in the range of 0.1 to 3 microns) on or inside a low-Z material substrate (e.g., vacuum window 29) to achieve a small X-ray spot size (e.g., less than 5 microns) in the cross-sectional plane of Figure 3A along the X-ray propagation axis 10 at an angle of less than 30 degrees with respect to the outer surface 27 of the vacuum window 29. The effective FWHM width s of the X-ray spot size in the cross-sectional plane is given by s = {(t / 2) 2 +[W·sin(θ)] 2} 0.5It can be approximately equal to, where t is the thickness of the high-Z material and W is the FWHM width of the electron beam size in the cross-sectional plane. For example, when t = 1 micron, θ = 10 degrees, and W = 1 micron, the effective FWHM width s of the X-ray spot size in the cross-sectional plane is 0.53 microns, which is smaller than W. Therefore, the effective X-ray source size in the cross-sectional plane along the X-ray propagation axis 10 can be compressed compared to the width of the electron beam 24 in the cross-sectional plane.

[0024] If the effective X-ray spot size in the cross-sectional plane is substantially smaller than the electron beam width W, certain realizations can further achieve a small X-ray spot size in a direction substantially perpendicular to the cross-sectional plane in Figure 3A. For example, as schematically shown by Figure 3B, the electron beam 24 can be focused so as to be compressed in a direction substantially perpendicular to the cross-sectional plane. A focused electron beam 24 having an elongated (e.g., rectangular) shape or footprint in the X-ray target 28 may have a long side dimension (e.g., FHWM width W) in the cross-sectional plane of Figure 3A and a short side dimension (e.g., FWHM width w) in a direction substantially perpendicular to the cross-sectional plane of Figure 3A (e.g., in a plane including the X-ray propagation axis 10 and the surface normal of the outer surface 27 of the vacuum window 29), where the short side dimension is smaller than the long side dimension.

[0025] As another example, as schematically shown by Figure 3C, the width d of the high-Z material of at least one X-ray generating material in a direction substantially perpendicular to the cross-sectional plane of Figure 3A may be smaller than the FWHM width w of the focused electron beam 24 in a direction substantially perpendicular to the cross-sectional plane of Figure 3A. Since at least one X-ray generating material generates X-rays more efficiently than the vacuum window 29 (for example, the X-ray generation efficiency is approximately proportional to the average atomic number of the material), the portion of the electron beam 24 that does not collide with at least one X-ray generating material does not generate X-rays efficiently and does not contribute substantially to the X-ray spot size, thus limiting the X-ray spot size to the width d in a direction substantially perpendicular to the cross-sectional plane of Figure 3A. In a particular implementation, the width d of at least one X-ray generating material is in the range of 0.1 microns to 5 microns.

[0026] When viewed along the X-ray propagation axis 10 (for example, from the direction of at least one position-sensitive X-ray detector 50), and considering the extraction angle between the outer surface 27 and the X-ray propagation axis 10, the X-ray spots in Figures 3B and 3C may appear square or circular in at least one position-sensitive X-ray detector 50. For example, at an extraction angle of 10 degrees, an electron beam 24 with a rectangular focus having a 5:1 aspect ratio may appear substantially symmetrical. Because the electron beam 24 is elongated and impacts a wider area than if the electron beam 24 were narrowly focused in both directions, higher electron power can be used to reduce image acquisition time by increasing the X-ray flux (e.g., achieving higher X-ray brightness) while maintaining spatial resolution. For example, an electron beam 24 with a 5:1 aspect ratio can provide up to a 5-fold increase in apparent power density. Further reducing the extraction angle in conjunction with a focused electron beam with a high aspect ratio can facilitate further improvements in apparent power density. A specific implementation provides a spatial resolution in the cross-sectional plane, as shown in Figure 3A, which is higher than the spatial resolution in the direction perpendicular to the cross-sectional plane. In many applications, high depth resolution (e.g., in a direction substantially perpendicular to the surface 32 of a laterally extended object 30) is more important than lateral resolution (e.g., imaging of delamination of solder bumps parallel to the surface of a semiconductor package or stress-induced cracks in solder bumps).

[0027] aperture Figure 4 schematically shows another exemplary system 5 comprising at least one aperture 70, relating to a particular implementation described herein. In the particular implementation, the at least one aperture 70 comprises at least one orifice 72 (e.g., a slit, having a width of less than 100 microns) in at least one solid material (e.g., a plate), the at least one orifice 72 is positioned on the X-ray propagation axis 10 and between the vacuum window 29 and the object 30 (e.g., downstream of the vacuum window 29 and upstream of the object 30). At least one aperture 70 is configured not to attenuate divergent X-rays 60 (e.g., X-rays used for imaging) propagating along the X-ray propagation axis 10 within the angular divergence angle 12 before reaching the ROI 31 and / or at least one position-sensitive X-ray detector 50, but to attenuate at least some X-rays and / or scattered X-rays (e.g., X-rays not used for imaging) emitted in other directions from the vacuum window 29 before reaching the ROI 31 and / or at least one position-sensitive X-ray detector 50. In a particular implementation, at least one solid material of the at least one aperture 70 has a sufficiently high Z (e.g., W, Au), a sufficiently high electron density, and a substantially small thickness, so that the at least one aperture 70 is configured to attenuate X-rays without substantially limiting the FOD. In a particular implementation, at least one aperture 70 is configured to attenuate the amount of irrelevant X-ray flux outside the angular divergence angle 12 that is emitted by the X-ray source 20 and reaches the object 30. By attenuating this irrelevant X-ray flux, at least one aperture 70 can reduce the contribution of the irrelevant X-ray flux to the image as a harmful background, and / or reduce the amount of harmful X-rays irradiated to the object 30 by the irrelevant X-ray flux that does not contribute to imaging. The edge of at least one orifice 72 of at least one aperture 70 can define the angular divergence angle 12 by allowing only X-rays within the at least one orifice 72 to propagate further along the X-ray propagation axis 10.In a specific implementation, the distance from one aperture 70 to the vacuum window 29 is configured to achieve a small FOD between the X-ray source point and the ROI 31 (for example, a distance in the range of 0.3 to 5 millimeters).

[0028] Sample motion stage and sample mount In a particular implementation, at least one sample motion stage 80 is configured to move an object 30 relative to an X-ray source 20 and / or at least one position-sensitive X-ray detector 50. In a particular implementation, at least one sample motion stage 80 comprises at least one linear motion substage 82 configured to controllably adjust the position of the object 30 (e.g., along substantially vertical x, y, and z directions) and at least one rotational motion substage 84 configured to controllably adjust the orientation of the object 30 (e.g., to rotate the object 30 about a rotation axis 19). For example, at least one linear motion substage 82 may include one, two, or three electromechanical linear motion drivers (e.g., linear motors, stepping motors, motors with encoders, piezoelectric motors, screw-type rotary motors) configured to move the object 30 so that the ROI 31 is at a selected position along the X-ray propagation axis 10, and at least one rotational motion substage 84 may include at least one electromechanical motion driver (e.g., rotary motors, stepping motors, motors with encoders, linear motion drivers with worm drives) configured to rotate the object 30 and the ROI 31 around a rotation axis 19.

[0029] In a particular implementation, as schematically shown in Figure 2, a laterally extended object 30 may be mounted on at least one sample motion stage 80 such that the surface 32 of the object 30 parallel to the long side dimension of the object 30 is inclined at an angle β (e.g., in the range of 1 to 30 degrees) with respect to the X-ray propagation axis 10. In a particular implementation, at least one rotational motion substage 84 is further configured to controllably adjust the angle β, and at least one linear motion substage 82 is further configured to linearly translate the object 30 and / or at least one rotational motion substage 84 so that the ROI 31 is positioned on the rotation axis 19 and the X-ray propagation axis 10. In a particular implementation, at least one sample motion stage 80 is configured to controllably adjust the distance between the rotation axis 19 and the X-ray spot (e.g., electron beam focus).

[0030] In a particular implementation, the sample mount 85 is configured to hold the object 30 while it is being irradiated by the X-rays 60 so that the X-rays 60 pass through the ROI 31 to at least one position-sensitive X-ray detector 50. The sample mount 85 is configured so that the X-rays 60 interact minimally with the solid components of at least one sample motion stage 80 (e.g., at least one linear motion substage 82 and at least one rotational motion substage 84) (e.g., to be minimally scattered and / or absorbed). The sample mount 85 is configured to reduce (e.g., minimize) the portion of divergent X-rays 60 that interacts with the solid parts of at least one sample motion stage 80 as the object 30 rotates around the rotation axis 19.

[0031] In a particular implementation, at least one sample motion stage 80 and sample mount 85 are configured to hold and rotate the object 30 without compromising throughput. For example, since the throughput of microfocus transmission X-ray computed laminography is inversely proportional to the square of the focal-object distance (FOD), high throughput can be achieved by using a small FOD (e.g., bringing the object 30 as close as possible to the X-ray source 20).

[0032] Figures 5A and 5B schematically show an exemplary system 5 relating to a specific implementation described herein, in which at least one sample motion stage 80 and a sample mount 85 are configured to hold and rotate an object 30. In the exemplary system 5 of Figure 5A, the object 30 is located between the X-ray source 20 and at least one sample motion stage 80. At least one linear motion substage 82 and at least one rotational motion substage 84 are located on the opposite side of the object 30 from the at least one X-ray source 20, so that at least one sample motion stage 80 does not obstruct the X-rays 60 from reaching the object 60. However, as shown in Figure 5A, after propagating through the ROI 31, the X-rays 60 propagate through at least a portion of the sample mount 85 before reaching at least one position-sensitive X-ray detector 50. In the exemplary system 5 of Figure 5B, the X-ray source 20 and at least one sample motion stage 80 are on the same side of the object 30, and the X-rays 60 propagate through at least a portion of the sample mount 85 before reaching the object 30 (for example, the sample mount 85 can limit the closest approach of the X-ray source 20 to the object 30). Figure 5A shows an exemplary implementation in which the X-ray source 20 is above the object 30 and at least one sample motion stage 80, but in other implementations, the X-ray source 20, the object 30, and at least one sample motion stage 80 may have any orientation while maintaining the same relative positioning with respect to each other. Figure 5B shows an exemplary implementation in which the X-ray source 20 is below the object 30, but in other implementations, the X-ray source 20, the object 30, and at least one sample motion stage 80 may have any orientation while maintaining the same relative positioning with respect to each other.

[0033] In a particular implementation, the sample mount 85 is configured to offset the object 30 from at least one sample stage 80 such that at least one sample stage 80 (e.g., at least one linear motion substage 82 and at least one rotational motion substage 84) does not fall within the imaging field of view of at least one position-sensitive X-ray detector 50 (e.g., to reduce, avoid, or minimize X-ray scattering and / or absorption that would adversely affect the fidelity of image reconstruction). For example, the sample mount 85 may offset the object 30 from at least one sample stage 80 by a distance greater than 50 millimeters (e.g., greater than 100 millimeters, in the range of 100 millimeters to 500 millimeters, or in the range of 100 millimeters to 200 millimeters). A sample mount 85 in a particular implementation comprises a first portion 86 configured to be struck by at least a portion of the X-rays 60 as the object 30 and the sample mount 85 rotate around the rotation axis 19, and a second portion 87 mechanically coupling the first portion 86 to at least one motion substage (e.g., at least one linear motion substage 82 and / or at least one rotational motion substage 84). The first portion 86 and the second portion 87 are configured to offset the ROI 31 of the object 30 from at least one motion substage so that the divergent X-rays 60 do not strike at least one motion substage or the second portion 87 (e.g., as the object 30 and the sample mount 85 rotate around the rotation axis 19).

[0034] In a particular implementation, the first part 86 is essentially composed of low-Z elements (e.g., atomic elements with an atomic number less than 14) and / or thin material (e.g., less than 10 millimeters thick along the rotation axis 19). In a particular implementation, the first part 86 has an X-ray transmittance of more than 50% for X-rays having an energy greater than 50% of the maximum X-ray energy in the X-ray spectrum of the X-rays 60 (e.g., X-rays emitted by the X-ray source 20). A particular such implementation simultaneously provides a sufficiently high throughput with a sufficiently small amount of radiation damage to the object 30 (e.g., since the X-rays do not need to penetrate the absorbing material of at least one sample motion stage 80, the X-ray flux irradiating the object 30 can be kept low enough to avoid radiation damage while providing a sufficiently large amount of detected X-rays for high-throughput imaging). For example, the first part 86 may comprise a carbon fiber plate or a quartz plate (e.g., having a projected thickness of 2 millimeters or less along the X-ray propagation axis 10). The second part 87 is mechanically coupled to the first part 86 and to at least one linear motion substage 82 and / or at least one rotational motion substage 84. In certain realizations, the second part 87 is composed of essentially the same low-Z elements and / or thin material as the first part 86 (e.g., a low-Z rod or hollow tube such as a carbon fiber tube or quartz tube with a projected thickness of 2 mm or less along the X-ray propagation axis 10), but in certain other realizations, the second part 87 comprises any solid material (e.g., regardless of X-ray absorption and / or scattering of the solid material). In certain realizations, the second part 87 is part of at least one linear motion substage 82 and / or at least one rotational motion substage 84.

[0035] For example, as schematically shown in Figure 5A, the first portion 86 and the second portion 87 may extend along the axis of rotation 19 and be configured so that the X-rays 60 do not collide with at least one linear motion substage 82 and / or at least one rotational motion substage 84 for all laminography angles and all rotation angles of the object. For at least one sample motion stage 80 having a maximum dimension (e.g., radius) R, extending from the axis of rotation 19 and substantially perpendicular to the axis of rotation 19, the first and second portions 86, 87 may be configured to hold the object 30 at a distance z from at least one linear motion substage 82 and / or at least one rotational motion substage 84 along the axis of rotation 19 so that the envelope of the X-rays 60 does not collide with the portion of the at least one sample motion stage 80 (e.g., a portion containing at least one element with atomic number Z greater than 14, at least one linear motion substage 82, at least one rotational motion substage 84) that would scatter and / or absorb the X-rays 60.

[0036] As another example, as schematically shown in Figure 5B, the first part 86 is configured to hold (e.g., clamp) the sides and / or edges of the object 30, and the second part 87 mechanically couples the first part 86 to at least one sample motion stage 80 (not shown in Figure 5B). The at least one sample motion stage 80 in Figure 5B further comprises a clear aperture region 88 (e.g., a region at least partially defined by the second part 87) through which the X-ray source 20 is configured to pass at least partially. A particular such implementation allows the X-ray source 20 to be positioned arbitrarily close to the surface of the object 30, thereby providing high throughput while firmly mounting the object 30. An exemplary second part 87 compatible with the particular implementation described herein is a large-diameter, ultra-high-precision pneumatically bearing rotary stage available from PI (Physik Instrumente) in Auburn, Massachusetts.

[0037] The fidelity of image reconstruction depends on the precise rotation of the object 30 during measurement, and uncontrolled motion of the object 30 can result in a deviation of the actual recorded projection data from what the laminography reconstruction algorithm expects. As a result, these uncontrolled motions can cause blurring in the back-projected data, which reduces the resolution and contrast in the reconstructed volume. In a particular implementation, at least one sample motion stage 80 reduces (e.g., avoids, minimizes) deviations from pure rotation that cause translation and / or changes in orientation of the object 30 within the imaging field of view. At least one sample motion stage 80 may have sufficiently low unsystematic angular runout (e.g., uncontrolled angular motion of the axis of rotation 19 as a function of rotation about the axis of rotation 19), sufficiently low radial runout (e.g., uncontrolled translation of the axis of rotation 19 as a function of rotation about the axis of rotation 19, resulting in lateral movement of the object 30 substantially perpendicular to the axis of rotation 19), and / or sufficiently low axial runout (e.g., uncontrolled axial movement of the object 30 substantially parallel to the axis of rotation 19), such that the uncontrolled motion of the object 30 is less than one-fifth of the system resolution (e.g., uncontrolled motion of less than 0.1 microns for a system resolution of 0.5 microns). For example, an object 30 positioned at a distance L above at least one rotational motion substage 84 of at least one sample motion stage 80, and an unsystematic angular runout ω (e.g., less than 100 nanoradians, less than 200 nanoradians, less than 1 microradian, less than 5 microradians) for an image resolution (e.g., detector resolution divided by image magnification) δ may result in radial runout R = ωL < δ / 5 (e.g., less than δ / 3, less than δ / 2, less than 1 micron, less than 0.5 microns, less than 200 nanometers, less than 100 nanometers) and / or axial runout A < δ / 5 (e.g., less than δ / 3, less than δ / 2, less than 1 micron, less than 0.5 microns, less than 200 nanometers, less than 100 nanometers).In a particular implementation, at least one sample motion stage 80 has non-systematic angular runout (e.g., error) of less than 5 microradians (e.g., less than 1 microradian), radial runout repeatability better than 1000 nanometers, and axial runout repeatability better than 1000 nanometers. In a particular implementation, system 5 further comprises a measurement system configured to measure the angular runout of at least one sample motion stage 80 with accuracy better than less than 5 microradians (e.g., less than 1 microradian), measure the radial runout of at least one sample motion stage 80 with accuracy better than 1000 nanometers, and / or measure the axial runout of at least one sample motion stage 80 with accuracy better than 1000 nanometers.

[0038] In a particular implementation, at least one rotational motion substage 84 of at least one sample motion stage 80 may comprise an air-bearing rotational stage (e.g., the A-62X or A-688 rotational stage available from Physik Instrumente (PI) in Auburn, Massachusetts, or the ABRX00, ABRX150, or ABRX250 rotational stages available from Aerotech, Inc. in Pittsburgh, Pennsylvania) having a deflection angle of less than 5 microradians (e.g., less than 1 microradian, less than 200 nanoradians) and radial and axial runout of less than 100 nanometers. The position of the object 30 over an angular range of rotation around the rotation axis 19 (e.g., 360 degrees) may be more accurate than half the system's resolution.

[0039] Position-sensitive X-ray detector In a particular implementation, at least one position-sensitive X-ray detector 50 is configured to record an image of the X-rays 60 received after passing through the ROI 31. Examples of at least one position-sensitive X-ray detector 50 include flat panel detectors (FPDs) including photon counting detectors (e.g., including silicon, CdTe, and / or CdZnTe, configured to directly convert X-rays into electrons with or without energy discrimination; Eiger ASICs and Pilatus ASICs available from Dectris in Baden-Detville, Switzerland), scintillator materials (e.g., CdWO4, CsI, Gd2O2S, LSO, GAGG, and / or LYSO; Shad-o-Box HS detectors available from Teledyne Dalsa in Waterloo, Canada; 2315N detectors available from Varex Imaging in Salt Lake City, Utah; Athena and Onyx detectors available from Nordson Corporation in Westlake, Ohio; and 1412HR detectors available from Spectrum Logic Corporation in Boulder, Colorado). Examples include, but are not limited to, a detector, optical fiber plate and CMOS or CCD detector, a scintillator material (e.g., CdWO4, CsI, Gd2O2S, LSO, GAGG, and / or LYSO) configured to magnify an image onto the CMOS or CCD detector, and an objective lens. In a particular implementation, at least one position-sensitive X-ray detector 50 comprises a plurality of active elements 52 (e.g., pixels) having a lateral dimension (e.g., dimension along the surface of the detector 50) of less than 70 microns (e.g., less than 50 microns).

[0040] In a particular implementation, at least one position-sensitive X-ray detector 50 is configured to receive and image X-rays 60 transmitted through an object 30 containing an ROI 31, the X-rays 60 having a predetermined range of energy (e.g., X-ray spectrum) that promotes (e.g., optimizes) sufficient image contrast to distinguish features of an object within the ROI 31 and / or reduces the imaging acquisition time. For example, the predetermined X-ray spectrum can be generated by selecting the focusing electron energy and / or at least one X-ray generating material of at least one X-ray target 28 such that the generated X-rays 60 of the predetermined X-ray spectrum have a sufficiently large X-ray flux to promote image contrast and / or imaging acquisition time. In the case of X-ray imaging using absorption contrast, the predetermined X-ray spectrum may include energies in which the object 30 has an X-ray transmittance in the range of 5% to 85% (e.g., in the range of 8% to 30%). This range of X-ray transmittance can offer a favorable trade-off between image contrast (which prefers lower energy X-rays) and the transmittance of object 30 (which prefers higher energy X-rays).

[0041] In another example, at least one position-sensitive X-ray detector 50 may be configured to have at least one energy threshold for detecting X-rays (for example, at least one position-sensitive X-ray detector 50 may be configured to reject and / or suppress the detection of X-rays having energy below a first energy threshold and / or energy above a second energy threshold). For example, at least one energy threshold may comprise a threshold cutoff X-ray energy, and at least one position-sensitive X-ray detector 50 is configured to image only X-rays having energy below this threshold cutoff X-ray energy. The threshold cutoff X-ray energy in a particular implementation corresponds to X-rays for which the object 30 has an X-ray transmittance of less than 85% (e.g., less than 50%). For example, at least one position-sensitive X-ray detector 50 may comprise a photon counting detector configured to select at least one threshold cutoff X-ray energy (e.g., to controllably adjust the threshold cutoff X-ray energy). Photon counting detectors may further be configured to collect energy-dependent X-ray transmission images (for example, using multiple operator-selectable energy windows to reduce noise, image artifacts, and / or provide material distinction). In another example, at least one position-sensitive X-ray detector 50 may comprise a combination of a scintillation screen and material configured to image only X-rays below a threshold cutoff X-ray energy.

[0042] Figures 6A and 6B schematically illustrate two examples of a system 5 relating to a specific implementation described herein, wherein at least one position-sensitive X-ray detector 50 comprises a first position-sensitive X-ray detector 54 and a second position-sensitive X-ray detector 56. The second position-sensitive X-ray detector 56 may be configured to provide gain sensitivity to a different portion of the X-ray spectrum than that obtained from the first position-sensitive X-ray detector 54, to increase the overall throughput, and / or to provide measurements at a different spatial resolution than that of the first position-sensitive X-ray detector 54. For example, the first and second position-sensitive X-ray detectors 54, 56 may have different scintillation materials and / or different scintillator thicknesses.

[0043] For example, as schematically shown in Figure 6A, the first detector 54 may be configured to absorb and detect a first spectral portion of the X-rays 60 that have passed through the ROI 31, and the second detector 56 may be configured to absorb and detect a second spectral portion of the X-rays 60 that have passed through the ROI 31 but have not been absorbed by the first detector 54. The first detector 54 may be configured (e.g., optimized) for high resolution and capable of transmitting at least a portion of the X-ray flux received from the ROI 31, and the second detector 56 may be positioned behind the first detector 54 (see, for example, Figure 6A) and configured to detect at least a portion of the X-ray flux that has passed through the first detector 56. As another example, as schematically shown in Figure 6B, a first spectral portion of the X-rays 60 propagating along the X-ray propagation axis 10 may be absorbed by the scintillator screen 55 of the first detector 54, and a second spectral portion of the X-rays 60 propagating along the X-ray propagation axis 10 may pass through the scintillator screen 55 and collide with the second detector 56. The X-rays absorbed by the scintillator screen 55 may generate scintillation photons (e.g., visible light photons), which are reflected by the mirror 57 and imaged onto the position-sensitive photon detector 59 by the objective lens 58. The material and thickness of the mirror 57 may be selected to have high transmittance of X-rays 60 passing through the object 30.

[0044] Figures 6A and 6B schematically show a configuration in which two detectors 54, 56 are positioned to detect X-rays 60 propagating along the same direction from one another. However, in certain other realizations, the two detectors 54, 56 are positioned to collect X-rays propagating from at least one X-ray source 20 along different directions from one another (e.g., different directions at angles less than 45 degrees with respect to the outer surface 27 of the vacuum window 29). In certain realizations, the two detectors 54, 56 are used simultaneously from one another, while in certain other realizations, the two detectors 54, 56 are used separately from one another (e.g., sequentially). In certain realizations, the two detectors 54, 56 are configured to have different pixel resolutions in the object 30 (e.g., approximately equal to the pixel size of the detector divided by its geometric image magnification). Combining the outputs of the two detectors 54, 56 in certain realizations can improve the overall detection efficiency of the system 5, provide access to a wider range of object spatial frequencies, and thus improve throughput and reconstruction quality. In a particular implementation, the second detector 56 is configured (e.g., optimized) to be sensitive to different parts of the X-ray spectrum of the X-rays 60, which may enable beam hardening correction and / or material identification.

[0045] Figure 7 schematically shows an exemplary system 5 configured to perform multi-contrast X-ray imaging using Talbot interferometry or Talbot-Lau interferometry, relating to a specific implementation described herein. System 5 may be configured to provide high resolution and sensitivity and unique imaging capabilities (e.g., absorption, phase, and dark-field image contrast in 2D and 3D, dark-field, and enhanced absorption contrast imaging) for a wide range of applications. In a specific implementation, System 5 is configured to collect a 3D imaging dataset using Talbot-Lau interferometry, dark-field contrast, and enhanced absorption contrast, and to reconstruct the dataset to obtain a 3D image of an object 30 and / or an ROI 31 within the object 30. In a specific implementation, as schematically shown in Figure 7, System 5 further comprises at least one sample motion stage 80 and a sample mount 85 (see, for example, Figure 5A) configured to reduce (e.g., minimize) the interaction between a portion of the divergent X-rays 60 propagating along the X-ray propagation axis 10 and the solid portion of at least one sample motion stage 80.

[0046] As schematically shown in Figure 7, System 5 comprises a first grating G1 (e.g., a phase grating) configured to generate a Talbot self-image interference pattern at a first position along the X-ray propagation axis 10, and a second grating G2 (e.g., an analyzer grating) positioned at a second position along the X-ray propagation axis 10. The first and second gratings G1 and G2 are configured to adapt to Talbot interferometry so that the Talbot pattern is indirectly imaged by at least one position-sensitive X-ray detector 50. In certain other realizations, the second grating G2 may be omitted, and at least one position-sensitive X-ray detector 50 may have an active element 52 with a spatial resolution (e.g., size in the transverse direction substantially perpendicular to the X-ray propagation axis 10) small enough to adapt to Talbot interferometry so that the Talbot pattern is directly imaged by at least one position-sensitive X-ray detector 50. In certain implementations, System 5 further comprises a source grating and is configured to perform Talbot-Rau interferometry. In certain other implementations, the X-ray source 20 comprises a plurality of X-ray targets 28 in a regular array of one-dimensional or two-dimensional configurations, and the geometric parameters of the plurality of X-ray targets 28, a first grating G1, and a second grating G2 are configured to satisfy the requirements of a Talbot-Rau interferometer. Various configurations of the X-ray source 20 and the first and second gratings G1, G2 are disclosed in U.S. Patents 9,719,947 and 10,349,908, each of which is incorporated herein by reference in its entirety.

[0047] In a particular implementation, the first grating G1 comprises an absorbing grating. For example, one or both of the first grating G1 and the second grating G2 may comprise an array of patterned one-dimensional or two-dimensional substantially X-ray absorbing (e.g., greater than 50%) structures having a width in the range of 0.5 to 20 microns and being spaced apart from each other by substantially non-absorbing (e.g., less than 50% absorptivity) gaps having a width in the range of 0.5 to 20 microns. In addition to using techniques such as the phase step method for tricontrast imaging (e.g., absorption, phase, and scattering), the system 5 of a particular implementation may be configured to obtain dark-field (e.g., scattering) contrast imaging only by configuring the pitch, distance from the X-ray source 20, and alignment of the first and second gratings such that X-rays transmitted through the aperture of the first (e.g., upstream) grating G1 are incident on the absorbing portion of the second (e.g., downstream) grating G2. For example, the first grating G1 and the second grating G2 can be positioned along the X-ray propagation axis 10 between at least one X-ray source 20 and at least one position-sensitive X-ray detector 50 (for example, the first grating G1 is closer to at least one X-ray source 20 than to at least one position-sensitive X-ray detector 50), so that substantially non-absorbent structures of the second grating G2 are aligned with substantially absorbing structures of the first grating G1 (for example, in their shadows), and substantially absorbing structures of the second grating G2 are aligned with substantially non-absorbent structures of the first grating G1. In such a configuration, it is expected that in the absence of object 30, no X-rays will penetrate both the first grating G1 and the second grating G2; however, in the presence of scattering features of object 30, at least some of the scattered X-rays will penetrate the second grating G2, leading to imaging of the features of object 30 that caused the scattered X-rays. In a particular implementation, an enhanced absorption contrast image can be obtained by shifting the relative alignment of the first and second gratings by half a pitch from the configuration used in dark-field imaging, such that X-rays scattered by object 30 are reduced by the absorbing structure of the second grating G2.In a particular implementation, instead of having a second grid G2, a first set of active elements 52 (e.g., pixels) of at least one position-sensitive X-ray detector 50 is aligned with substantially absorptive structures of the first grid G1 (e.g., in their shadows), and a second set of active elements 52 of at least one position-sensitive X-ray detector 50 is aligned with substantially non-absorptive structures of the first grid G1. In such a configuration, when an object 30 is positioned along the X-ray propagation axis 10, the X-ray counts recorded by the second set of active elements 52 may be used to generate an absorption contrast image, and the X-ray counts recorded by the first set of active elements 52 may be used to generate a scatter / darkfield and / or refractive image.

[0048] In a specific implementation where a semiconductor IC package is to be imaged in 3D, System 5 is configured to obtain an X-ray transmission image using an X-ray propagation axis 10 at a small angle (e.g., within a range of less than 45 degrees) with respect to the surface normal of the semiconductor chip surface 32. For example, rotational laminography over a large angular range (e.g., from 180 to 630 degrees) or limited-angle translation laminography over a finite angular range (e.g., ±30 degrees) may be performed. The 3D image can be combined with a laminographic 3D image obtained using an X-ray propagation axis 10 at a large angle (e.g., greater than 60 degrees) with respect to the surface normal of the semiconductor chip surface 32 to generate a 3D image. Figure 8 schematically shows an exemplary System 5 adapted for generating a translation laminographic image according to a specific implementation described herein. System 5 comprises an additional X-ray detector 90 and at least one detector stage 92 configured to translate the detector 90 and the object 30 relative to the X-ray source 20. In a particular implementation, at least one detector stage 92 comprises at least one sample motion stage 80. For example, the detector 90 and the object 30 can be moved proportionally in the same direction (e.g., along a line from the X-ray source 20 through the ROI 31 to the center of the detector 90). In a particular implementation, the system 5 in Figure 8 is configured to achieve higher spatial resolution and / or better image clarity for features (e.g., the sidewalls of a copper interconnect) that extend substantially parallel to the plane normal of the semiconductor chip.

[0049] In a particular implementation, system 5 further comprises at least one motion mechanism configured to change the geometric scaling of an image of a region of interest of an object, generated by at least one position-sensitive detector. For example, at least one motion mechanism may include at least one first motion stage configured to move the X-ray source 20 relative to the object 30 (e.g., a linear motion stage, an electromechanical linear motion driver, a linear motor, a stepping motor, a motor with an encoder, a piezoelectric motor, a screw-type rotary motor), at least one second motion stage configured to move at least one position-sensitive detector 50 relative to the object 30 (e.g., a linear motion stage, an electromechanical linear motion driver, a linear motor, a stepping motor, a motor with an encoder, a piezoelectric motor, a screw-type rotary motor), and / or at least one third motion stage configured to move the object 30 relative to the X-ray source 20 and / or at least one position-sensitive detector 50 (e.g., a linear motion stage, an electromechanical linear motion driver, a linear motor, a stepping motor, a motor with an encoder, a piezoelectric motor, a screw-type rotary motor, at least one linear motion substage 82).

[0050] Figure 9 schematically shows an exemplary three-dimensional X-ray imaging system 5 configured to generate a three-dimensional X-ray image of a region of interest 31 within an object 30, according to a specific implementation described herein. System 5 comprises an X-ray source 20 and at least one X-ray detector 50 having at least one active element 52. The X-ray source 20 comprises a transmissive vacuum window 29 having an outer surface 27. The X-ray source 20 is configured to produce divergent X-rays 60, at least some of which exit the vacuum window 29 and propagate along an X-ray propagation axis 10 extending from the X-ray source 20, through the region of interest 31 of the object 30, to at least one active element 52 of the at least one X-ray detector 50. The divergent X-rays 60 have a propagation path within an angular divergence angle of more than 1 degree around the X-ray propagation axis 10. System 5 further comprises at least one sample motion stage 80 configured to rotate the object 30 about a rotation axis 19. At least one sample motion stage 80 is configured such that its rotation axis 19 has a non-zero angle with respect to the X-ray propagation axis 10. The X-ray source 20 and at least one X-ray detector 50 are configured to remain stationary while at least one sample stage 80 is rotating the object 30 during data acquisition for a three-dimensional X-ray image.

[0051] In a particular implementation, the X-ray imaging system 5 is configured to acquire (e.g., generate) data including a laminography dataset from multiple angular projections of a region of interest 31 within an object 30. In a particular implementation, the system 5 is configured to acquire the laminography dataset while the object 30 is rotating around a rotation axis 19 (e.g., a rotation axis 19 substantially perpendicular to the ground). The rotation axis 19 can be at a non-zero angle β with respect to the X-ray propagation axis 10 (e.g., defined as the distance from the center point of X-ray emission by the X-ray source 20 to the center point of at least one X-ray detector 50). The angle β can be in the range of 20 degrees or less, or greater than 45 degrees (e.g., between 45 and 85 degrees, or between 60 and 90 degrees).

[0052] In a particular implementation, the distance between the outer surface 27 of the vacuum window 29 and the surface 32 (e.g., the top surface) of the object 30, measured in a direction substantially parallel to the axis of rotation 19, is within the range of less than 7 millimeters (e.g., between 0.5 millimeters and 7 millimeters, between 1 millimeter and 5 millimeters, between 1 millimeter and 3 millimeters, or less than 1 millimeter). In a particular implementation, the system 5 further includes at least one sensor (e.g., a laser curtain, not shown) configured to detect a potential collision between the X-ray source 20 and the object 30 (e.g., to prevent the object 30 from potentially colliding with the X-ray source 20).

[0053] Figures 10A and 10B schematically illustrate another exemplary X-ray imaging system 5 in exemplary laminography and exemplary tomography configurations, respectively, relating to a particular implementation described herein. In a particular implementation, the X-ray imaging system 5 is configured to produce an X-ray image of a region of interest 31 within an object 30 and comprises an X-ray source 20 and at least one X-ray detector 50 (e.g., at least one position-sensitive X-ray detector). The X-ray source 20 comprises a transmissive vacuum window 29 having an outer surface 27 and is configured to produce divergent X-rays 60. At least some of the divergent X-rays 60 exit the vacuum window 29 and propagate along an X-ray propagation axis 10 extending from the X-ray source 20, through the region of interest 31 of the object 30, to the at least one X-ray detector 50. The divergent X-rays 60 received by the at least one X-ray detector 50 have a propagation path within an angular divergence angle of more than 1 degree around the X-ray propagation axis 10. System 5 further comprises at least one first motion stage 80 (e.g., at least one linear motion stage 82, at least one rotational motion stage 84) configured to move the object 30 relative to the X-ray source 20 and / or rotate the object 30 about the rotation axis 19.

[0054] In a particular implementation, System 5 further comprises one or more mechanisms configured to switch System 5 between at least two configurations (e.g., a laminography configuration and a tomography configuration). For example, as schematically shown by Figures 10A and 10B, System 5 may comprise at least one second motion stage 100 configured to move the X-ray source 20 and at least one X-ray detector 50 relative to an object 30 in order to switch between a laminography configuration (see, e.g., Figure 10A) in which the X-ray propagation axis 10 has a first non-zero angle β1 with respect to the rotation axis 19 and a tomography configuration (see, e.g., Figure 10B) in which the X-ray propagation axis 10 has a second non-zero angle β2 with respect to the rotation axis 19, where the second angle β2 is different from the first angle β1. The non-zero difference angle Δ=|β2-β1|, which is equal to the absolute value of subtracting the first angle β1 from the second angle β2, can be at least 10 degrees (e.g., at least 30 degrees, at least 45 degrees, at least 90 degrees). In a particular implementation, the first angle β1 is in the range of 20 degrees or less, or in the range of 45 degrees to less than 90 degrees (e.g., in the range of 45 degrees to 85 degrees, or in the range of 60 degrees to less than 90 degrees), and the second angle β2 is substantially equal to 90 degrees (e.g., in the range of 87 degrees to 95 degrees).

[0055] In a particular implementation (see, for example, Figures 10A to 10B), the motion of at least one second motion stage 100 for switching from a laminography configuration to a tomography configuration, or vice versa, may include rotation (e.g., swirling) of the X-ray propagation axis 10 (e.g., extending from the X-ray source 20 to at least one X-ray detector 50) around a point along the rotation axis 19 (e.g., a point within the region of interest 31). In a particular implementation, this motion can be performed without changing the distance between the vacuum window 29 of the X-ray source 20 and at least one X-ray detector 50 (e.g., without changing the distance between the vacuum window 29 and the region of interest 31, and without changing the distance between at least one X-ray detector 50 and the region of interest 31). In certain other implementations, this motion may be achieved by varying the distance between the vacuum window 29 of the X-ray source 20 and at least one X-ray detector 50 (e.g., by varying the distance between the vacuum window 29 and the region of interest 31, and / or by varying the distance between at least one X-ray detector 50 and the region of interest 31). For example (see, for example, Figures 10A to 10B), in a laminography configuration, the distance between the outer surface 27 of the vacuum window 29 and the object (e.g., the upper outer surface of the object 30) may be in the range of less than 5 millimeters, and in a tomography configuration, the distance between the outer surface 27 of the vacuum window 29 and the object 30 (e.g., the side or edge of the object 30) may be in the range of more than 5 millimeters.

[0056] In a particular implementation, at least one second motion stage 100 comprises at least one goniometer, and both the X-ray source 20 and at least one X-ray detector 50 are mechanically coupled to at least one goniometer (for example, both are mechanically coupled to a single goniometer). In a particular other implementation, at least one second motion stage 100 comprises at least one source motion stage (for example, at least one linear motion stage and / or at least one rotational motion stage) configured to move the X-ray source 20 relative to the object 30, and at least one detector motion stage (for example, at least one linear motion stage and / or at least one rotational motion stage) configured to move at least one X-ray detector 50 relative to the object 30. For example, at least one source motion stage can adjust both the orientation and distance of the X-ray source 20 relative to the region of interest 31 of the object 30, and at least one detector motion stage can adjust both the orientation and distance of at least one X-ray detector 50 relative to the region of interest 31 of the object 30, while the orientation and / or distance between the X-ray source 20 and at least one X-ray detector 50 remains unchanged.

[0057] In certain implementations, at least one second motion stage 100 is configured to keep the X-ray source 20 and at least one X-ray detector 50 stationary while at least one first stage 80 rotates the object 30 during data acquisition for a three-dimensional X-ray image (e.g., in a laminography configuration and / or tomography configuration). In certain implementations where the cone angle of the X-rays emitted from the vacuum window 29 is sufficiently large (e.g., greater than 130 degrees), at least one second motion stage 100 is configured not to rotate the X-ray source 20 while rotating at least one X-ray detector 50 (e.g., only 90 degrees) around the region of interest 31, so that the X-ray propagation axis 10 defined by the divergent X-rays 60 received by at least one X-ray detector 50 rotates around the region of interest 31. In a particular implementation, the X-ray source 20 is configured to adjust the tilt angle of the electron beam inside the X-ray source 20 (for example, to be substantially parallel to the X-ray propagation axis 10) so that the X-ray source 20 rotates only within a smaller range, such as between 30 and 45 degrees.

[0058] In a particular implementation, as schematically shown in Figures 10A and 10B, the vacuum window 29 is located at the part of the X-ray source 20 that tapers at the end closest to the object 30 (e.g., a tapered projection or "snout"), which allows switching (e.g., converting) between a laminography configuration and a tomography configuration without completely rotating the X-ray source 20 by 90 degrees, while still allowing the object 30 to be brought as close as possible to the vacuum window 29 in either a laminography configuration or a tomography configuration.

[0059] Figure 11 schematically shows an exemplary vacuum window 29 of a transmission X-ray source 20 relating to a specific implementation described herein. The vacuum window 29 of Figure 11 comprises a substrate 110 and at least one X-ray target 28. The substrate 110 substantially comprises low atomic number elements with atomic numbers less than 14 (e.g., less than 10, less than 5), examples of which include, but are not limited to, carbon (e.g., diamond) and boron. The at least one X-ray target 28 comprises at least one X-ray generating material consisting of at least one atomic element with an atomic number of 13 or greater (e.g., 26 or greater), examples of which include, but are not limited to, tungsten. Other examples of the vacuum window 29 and the at least one X-ray target 28 are also described herein (see, for example, Figures 3A–3C and the corresponding text).

[0060] In certain realizations, the rotation axis 19 of at least one sample motion stage 80 is at a large angle (e.g., 60 degrees or more) with respect to the electron propagation axis 120, which can result in a propagation path length (PL) of the electron beam 24 from the electron beam source 22 through at least one X-ray target 28. For example, the electron beam 24 may have a propagation PL substantially greater than the thickness d of the film, as shown in Figure 11. The PL can be expressed as follows: PL = d / sin(90 degrees - A), where d is the thickness of at least one X-ray target 28 and A is the angle between the rotation axis 19 and the electron propagation axis 120. For example, if angle A is 80 degrees, the PL is 1 / sin(10) = 5.76 times the thickness d. Due to the increase in PL, certain realizations may have at least one X-ray target 28 with a reduced thickness d compared to most previously disclosed transmission X-ray sources.

[0061] Figure 12A schematically shows a portion of exemplary System 5 relating to a specific implementation described herein. The X-ray source 20 in Figure 12A comprises an X-ray target 28 comprising an X-ray generating material having a thickness d equal to 5 microns and comprising a tungsten film on a diamond substrate 110, wherein the substrate 110 is substantially transparent to the generated X-rays having energies between 25 keV and 30 keV (see, for example, Figure 11). When the inclination angle A between the rotation axis 19 and the electron propagation axis 120 is 78.5 degrees (for example, corresponding to the X-ray propagation axis 10 being at an angle of 11.5 degrees with respect to the upper outer surface of the object 30), the 5 micron thickness becomes effectively a PL of 25 microns. The X-ray target 28 becomes self-attenuating, and the transmittance drops to a range of 20-35%, compared to 70-82% for a 5-micron film.

[0062] In situations where lower X-ray energies should be used, a transmission X-ray source may comprise an X-ray generating material having a thickness of 4 microns or less (e.g., 1 micron or less, 1 to 2 microns, or 2 to 4 microns).

[0063] In a particular implementation, the X-ray source 20 has a small X-ray spot with a full width at half maximum (FWHM) width of 1 micron or less in at least one dimension (e.g., in the range of 1 to 2 microns, and in the range of 2 to 3 microns) (see, for example, Figures 3A–3C and the corresponding text). In a particular implementation, the X-ray spot is rectangular or elliptical (e.g., one long side dimension and one short side dimension) such that a laminography dataset viewing the X-ray source 20 at a certain extraction angle uses substantially equivalent X-ray source sizes in both the X and Y directions.

[0064] In a particular implementation, as schematically shown in Figure 12B, the X-ray source 20 emits X-rays at a cone angle C within the range of 130 degrees or greater (e.g., within the range of 150 degrees or greater, within the range of 175 degrees or greater). In a particular implementation, the X-ray propagation axis 10 is within the range of 10 to 30 degrees from the outer surface of the vacuum window 29 of the X-ray source 20, and it may be possible to allow X-rays to pass through a region of interest 31 of an object 30 located near the X-ray source 20 (e.g., within a distance of 10 millimeters or less (e.g., within the range of 1 millimeter to 10 millimeters, within the range of 1 millimeter to 5 millimeters, within the range of less than 1 millimeter)). In a particular implementation, the cone angle C of the X-rays passing through the region of interest 31 of the object 30 is within the range of 30 degrees or less (e.g., within the range of less than 10 degrees, within the range of 10 to 20 degrees, within the range of 10 to 30 degrees).

[0065] Figure 13A schematically shows a side view of an exemplary system 5 comprising at least one exemplary X-ray detector 50, comprising a first X-ray detector 54 and a second X-ray detector 56, according to a particular implementation described herein. Figure 13B schematically shows the first and second X-ray detectors 54, 56 of Figure 13A as viewed along a direction substantially parallel to the X-ray propagation axis 10. In a particular implementation, the first X-ray detector 54 comprises a large field-of-view (LFOV) X-ray detector (e.g., a CMOS sensor) configured to provide a fast readout rate (e.g., greater than 10 fps, greater than 15 fps, greater than 20 fps, greater than 30 fps). For example, the first X-ray detector 54 may have a number of pixels in the range of more than 4 megapixels (e.g., more than 6 megapixels, more than 10 megapixels, more than 20 megapixels), and / or the first X-ray detector 54 may have a pixel size of more than 40 microns (e.g., more than 50 microns, more than 70 microns), so that the quotient of the number of pixels multiplied by the pixel size is 150 millimeters or more. The first X-ray detector 54 may comprise a plurality of first active elements 154 (e.g., pixels), each having a first width w1 (e.g., pixel size) in at least one direction substantially perpendicular to the X-ray propagation axis 10. In a particular implementation, the first width w1 is in the range of 20 to 75 microns (e.g., 40 microns, 50 microns), such as a flat panel detector with an numerical aperture of 1. In certain such implementations, the effective pixel size of system 5 may be within a range of less than 20 microns (for example, between 5 and 10 microns, between 10 and 15 microns, or between 15 and 20 microns).

[0066] In a particular implementation, as schematically shown in Figures 13A and 13B, the second X-ray detector 56 comprises a scintillator coupled to a sensor (e.g., a CCD or CMOS sensor) having a plurality of second active elements 156 (e.g., pixels). Each pixel has a second effective width w2 (e.g., pixel size) in at least one direction substantially perpendicular to the X-ray propagation axis 10. In a particular implementation, the effective second width w2 is less than 20 microns (for example, the first X-ray detector 54 comprises a CMOS sensor, a visible light objective lens 58 or an optical fiber, and a scintillator screen 55 such as CsI with a thickness of 100 microns or more). The second width w2 can be derived from the sensor pixel size divided (or multiplied) by any magnification (or reduction) effect of the visible light optical system, such as a reflective or refractive objective lens or an optical fiber. The second width w2 (e.g., 20 microns or less, 16 microns or less, 10 microns or less) may differ from the first width w1 (e.g., the second X-ray detector 56 may have a higher resolution than the first X-ray detector 54). The second X-ray detector 56 may be positioned alongside (e.g., side by side) the first X-ray detector 54, and the system 5 may further include at least one detector motion stage (e.g., at least one linear motion stage and / or at least one rotational motion stage) configured to move the first and / or second X-ray detectors 54,56 such that one of the selected first and second X-ray detectors 54,56 is struck by divergent X-rays 60. Certain such implementations may allow the user to select which X-ray detectors and which resolutions should be used based on the desired application.

[0067] Figure 14 schematically shows an exemplary three-dimensional X-ray laminography system 5 comprising an optical microscope 200 (e.g., equipped with an optical camera) according to a specific implementation described herein. System 5 comprises at least one X-ray detector 50 and an X-ray source 20 configured to produce divergent X-rays 60. At least some of the divergent X-rays 60 propagate along an X-ray propagation axis 10 extending from the X-ray source 20, through a three-dimensional field of view (3D FOV) 210 in an object 30, to at least one X-ray detector 50, and the divergent X-rays 60 have a propagation path within an angular divergence angle of more than 1 degree around the X-ray propagation axis 10. The optical microscope 200 is aligned with the X-ray source 20 and has a focus configured to overlap with at least a portion of the 3D FOV 210 (e.g., a portion of the 3D volume in the object 30 from which the divergent X-rays 60 from the X-ray source 20 are collected by at least one X-ray detector 50). System 5 may further include at least one sample motion stage configured to rotate an object about a rotation axis 19, wherein the rotation axis 19 is configured to have an angle with respect to the X-ray propagation axis 10 in the range of 60 to 85 degrees.

[0068] The optical microscope 200 may be configured to be used to align a selected portion of the region of interest 31 along both the X-ray propagation axis 10 and the rotation axis 19 (for example, within a 3D FOV 210, at the center 212 of the 3D FOV 210). In certain implementations, the optical microscope 200 is configured to facilitate alignment because, in X-ray computed laminography (also known as oblique computed tomography), the beam of diverging X-rays 60 is not perpendicular to the rotation axis 19 but intersects the rotation axis 19 at a certain angle, making it difficult to use X-ray projection information to align a particular internal layer of a flat object 30 to the center 212 of the 3D FOV 210. The optical microscope 200 may be positioned and aligned with respect to the X-ray source 20 to allow selection in the depth direction of a particular volume of interest within an object 30 having a finite thickness.

[0069] In a particular implementation, the optical microscope 200 is configured (e.g., adjusted and calibrated) so that the center 212 of the 3D FOV 210 is positioned at the intersection of the rotation axis 19 and the X-ray propagation axis 10. The optical microscope 200 may be pre-aligned and positioned so that the focal point of the optical microscope 200 coincides with the center 212 of the 3D FOV 210. In a particular implementation, the optical microscope 200 is configured to detect the surface morphology of an object 30 (e.g., whether the upper outer surface 32 of the object 30 is warped or curved).

[0070] In certain implementations, system 5 comprises at least one optical microscope stage (e.g., at least one linear motion substage and / or at least one rotational motion substage) configured to adjust the position and / or orientation of the optical microscope 200 relative to the X-ray source 20. For example, the optical microscope 200 may be moved to a position corresponding to (e.g., aligned with) the rotation axis 19. In certain other implementations, the optical microscope 200 may be fixed and centered along a predetermined rotation axis 19 of system 5 (e.g., having an optical axis collinear with it).

[0071] In a particular implementation, the distance between the optical microscope 200 and the object 30 is adjusted until the upper outer surface 32 of the object 30 is in focus with the optical microscope 200. The object 30 can then be positioned horizontally (or the X-ray source 20 and at least one X-ray detector 50 can be translated relative to the object 30) so that a predetermined region of interest 31 is located at the center 212 of the 3D FOV 210. If the depth of the region of interest 31 is known from the upper outer surface 32 of the object 30, the region of interest 31 can be easily aligned to the center 212 of the 3D FOV 210 by moving the object 30 along the rotation axis 19 by a known amount.

[0072] In a specific implementation, object 30 is a semiconductor wafer (e.g., a 300 mm wafer) having at least one region of a non-planar surface profile (e.g., wafer warp). In conventional laminography acquisition, the distance between the X-ray source 20 and the region of interest (ROI) 31 on object 30 is minimized. However, wafer warp can make it difficult to always maintain the minimum distance between the X-ray source 20 and the wafer (e.g., object 30) because the wafer may "collide" with the X-ray source 20. In a specific implementation, information from a wafer profilometer (e.g., contact or non-contact) is used as input to the computer of the laminography system, along with information about the 3D appearance of the X-ray source 20, to ensure that the distance between the ROI 31 and the X-ray source spot is reduced (e.g., minimized) without causing the wafer to collide with the X-ray source 20. In a specific implementation, no independent profilometer is used, and information from an optical microscope 200 is used to provide information about the surface profile to ensure the minimized source-sample distance. Alternatively, if object 30 comprises a sample other than a wafer (e.g., a PCB), a 3D model of object 30 (e.g., electronic CAD) may be used to reduce (e.g., minimize) the source-object distance while avoiding collisions between the X-ray source 20 and object 30.

[0073] Commonly used terminology is used to describe the systems and methods of specific implementations for ease of understanding, but these terms are used herein so as to have their broadest and most reasonable interpretation. Various aspects of this disclosure are described with respect to exemplary examples and implementations, but the examples and implementations disclosed should not be construed as limiting. Conditional language such as “can,” “could,” “might,” or “may” is generally intended to convey that a particular implementation includes a particular feature, element, and / or step, but other implementations do not, unless otherwise specifically noted or understood in a different manner within the context in which they are used. Thus, such conditional language is generally not intended to implicitly suggest that a feature, element, and / or step is absolutely necessary for one or more implementations. In particular, the terms “comprises” and “comprising” should be interpreted in a non-exclusive manner as referring to elements, components, or steps, indicating that the referenced elements, components, or steps may exist, be used, or be combined with other elements, components, or steps not explicitly referenced.

[0074] Conjunctive language, such as "at least one of X, Y, and Z," should be understood in the context of its general use to convey that an item, term, etc., is one of X, Y, or Z, unless otherwise specified. Therefore, such conjunctive language is not typically intended to implicitly suggest that a particular realization requires the presence of at least one X, at least one Y, and at least one Z.

[0075] The terms "approximately," "about," "generally," and "substantially," when used herein, represent values, quantities, or characteristics close to the mentioned value, quantity, or characteristic, while still performing the desired function or achieving the desired result. For example, the terms "approximately," "about," "generally," and "substantially" may refer to quantities within ±10%, ±5%, ±2%, ±1%, or ±0.1% of the mentioned quantity. As another example, the terms "approximately parallel" and "substantially parallel" refer to values, quantities, or characteristics that deviate from strictly parallel by ±10 degrees, ±5 degrees, ±2 degrees, ±1 degree, or ±0.1 degrees, while the terms "approximately perpendicular" and "substantially perpendicular" refer to values, quantities, or characteristics that deviate from strictly perpendicular by ±10 degrees, ±5 degrees, ±2 degrees, ±1 degree, or ±0.1 degrees. Furthermore, the scope disclosed herein includes any and all overlaps, sub-scopes, and combinations thereof. The words “up to,” “at least,” “greater than,” “less than,” and “between” include the number being described. The meanings of “a, an” and “said,” as used herein, include multiple references unless the context clearly indicates a different aspect. In this specification, structures and / or methods are discussed in terms of elements labeled by order adjectives (e.g., first, second, etc.), but order adjectives are used merely as labels to distinguish one element from another, and not to indicate the order of these elements or the order of their use.

[0076] Various configurations have been described above. It should be understood that the implementations disclosed herein are not mutually exclusive and may be combined with each other in various configurations. While the present invention is described with reference to these specific configurations, such descriptions are intended to be illustrative and not limiting. Those skilled in the art will be able to conceive of various modifications and applications without departing from the true spirit and scope of the invention. Thus, for example, in any method or process disclosed herein, the acts or actions constituting the method / process may be performed in any preferred sequence and are not necessarily limited to any particular disclosed sequence. The various implementations and features or elements from the examples described above may be combined with each other to create alternative configurations that are compatible with the implementations disclosed herein. Various aspects and advantages of the implementations are described as appropriate. It should be understood that not all such aspects or advantages can necessarily be achieved according to any particular implementation. Thus, for example, it should be recognized that various implementations may be performed in a manner that achieves or optimizes one advantage or group of advantages taught herein, without necessarily achieving other aspects or advantages that may be taught or suggested herein.

Claims

1. An X-ray imaging system configured to generate an X-ray image of a region of interest within an object, At least one X-ray detector, The system comprises an X-ray source having a transmissive vacuum window with an outer surface, wherein the X-ray source is configured to produce divergent X-rays, at least some of which exit the vacuum window and propagate along an X-ray propagation axis extending from the X-ray source, through the region of interest of the object, to at least one X-ray detector, and the divergent X-rays received by the at least one X-ray detector have a propagation path within an angular divergence angle of more than 1 degree with respect to the X-ray propagation axis, and the system further comprises A first motion stage configured to move the object relative to the X-ray source and / or rotate the object about a rotation axis, A system comprising at least one second motion stage configured to move the X-ray source and the at least one X-ray detector relative to the object to switch between a laminography configuration in which the X-ray propagation axis has a first non-zero angle with respect to the rotation axis and a tomography configuration in which the X-ray propagation axis has a second non-zero angle with respect to the rotation axis, wherein the second angle is different from the first angle.

2. The system according to claim 1, wherein the at least one second motion stage comprises at least one source motion stage configured to move the X-ray source relative to the object, and at least one detector motion stage configured to move the at least one X-ray detector relative to the object.

3. The system according to claim 1, wherein the first angle is in the range of 45 degrees to less than 90 degrees.

4. The system according to claim 1, wherein the second angle is substantially equal to 90 degrees.

5. The system according to claim 1, wherein the second angle and the first angle differ by at least 10 degrees.

6. The system according to claim 5, wherein the second angle and the first angle differ by at least 30 degrees.

7. The system according to claim 1, wherein the at least one second motion stage is configured to keep the X-ray source and the at least one X-ray detector stationary while the at least one first stage rotates the object during data acquisition for a three-dimensional X-ray image.

8. The system according to claim 1, wherein the at least one second motion stage comprises at least one goniometer.

9. The system according to claim 1, wherein in the laminography configuration, the distance between the outer surface of the vacuum window and the object is within a range of less than 5 millimeters.

10. The system according to claim 1, further comprising at least one sensor configured to detect a potential collision between the X-ray source and the object.

11. The system according to claim 1, wherein the X-ray source is a transmission-type X-ray source that produces a cone angle of at least 130 degrees.

12. The system according to claim 1, wherein the X-ray source comprises at least one layer on a diamond substrate, the at least one layer substantially comprises at least one atomic element having an atomic number of 13 or higher, and the at least one layer has a thickness in the range of 2 microns to 4 microns.

13. The system according to claim 12, wherein the vacuum window comprises the diamond substrate.

14. The system according to claim 1, wherein the at least one X-ray detector comprises a plurality of pixels having a pixel size in the range of 10 microns to 30 microns.

15. An X-ray imaging system configured to generate an X-ray image of a region of interest within an object, At least one X-ray detector, The system comprises an X-ray source configured to produce divergent X-rays, wherein at least some of the divergent X-rays propagate along an X-ray propagation axis extending from the X-ray source, through a three-dimensional field of view (3D FOV) within the object, to at least one X-ray detector, and the divergent X-rays have a propagation path within an angular divergence angle of more than 1 degree centered on the X-ray propagation axis, and the system further comprises: A system comprising an optical microscope having a focus that is aligned with the X-ray source and configured to overlap with at least a portion of the 3D FOV.

16. The system according to claim 15, further comprising at least one sample motion stage configured to rotate the object about a rotation axis, wherein the rotation axis is configured to have an angle between 60 and 85 degrees with respect to the X-ray propagation axis.

17. The system according to claim 16, wherein the optical microscope is fixed so as to have an optical axis collinear with the rotation axis.

18. The system according to claim 16, wherein the center of the 3D FOV is positioned at the intersection of the rotation axis and the X-ray propagation axis.

19. The system according to claim 15, further comprising at least one optical microscope stage configured to adjust the position and / or orientation of the optical microscope relative to the X-ray source.

20. An X-ray imaging system configured to generate a three-dimensional X-ray image of a region of interest within an object, An X-ray detector comprising at least one active element, The system comprises an X-ray source having a transmissive vacuum window with an outer surface, the X-ray source being configured to produce divergent X-rays, at least some of which exit the vacuum window and propagate along an X-ray propagation axis extending from the X-ray source, through the region of interest of the object, to at least one active element of at least one X-ray detector, the divergent X-rays having a propagation path within an angular divergence angle of more than 1 degree with respect to the X-ray propagation axis, and the system further comprises: A system comprising at least one sample motion stage configured to rotate the object about a rotation axis, wherein the rotation axis is configured to have a non-zero angle with respect to the X-ray propagation axis, the X-ray source and the at least one X-ray detector are configured to remain stationary while the at least one sample stage rotates the object during data acquisition for the three-dimensional X-ray image.

21. The system according to claim 20, wherein the non-zero angle is in the range of 60 to 85 degrees.

22. The system according to claim 20, wherein the non-zero angle is 20 degrees or less.

Images