III IV 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

Nov 2024_EDFA_Digital

A RESOURCE FOR TECHNICAL INFORMATION AND INDUSTRY DEVELOPMENTS NOVEMBER 2024 | VOLUME 26 | ISSUE 4 ELECTRONIC DEVICE FAILURE ANALYSIS edfas.org FOUR-DIMENSIONAL STEM: PART III, PTYCHOGRAPHY SPECIMEN THINNING BY ARGON ION BEAM MILLING TEM SAMPLES WITH PFIB AND STEM EBIC NONDESTRUCTIVE 3D X-RAY MICROSCOPY SPEEDS THROUGHPUT 4 20 14 27

A RESOURCE FOR TECHNICAL INFORMATION AND INDUSTRY DEVELOPMENTS NOVEMBER 2024 | VOLUME 26 | ISSUE 4 ELECTRONIC DEVICE FAILURE ANALYSIS edfas.org FOUR-DIMENSIONAL STEM: PART III, PTYCHOGRAPHY SPECIMEN THINNING BY ARGON ION BEAM MILLING TEM SAMPLES WITH PFIB AND STEM EBIC NONDESTRUCTIVE 3D X-RAY MICROSCOPY SPEEDS THROUGHPUT 4 20 14 27

Decrease time to yield, manage manufacturing excursions and recover yield caused by systematic defects with Tessent yield learning software. By understanding how to manage and optimize the test and data collection environment, yield learning solutions can identify systematic defects that cause low yield, saving time. Tessent yield learning software leverages test and design data to uncover hidden yield limiters and help identify root cause of defect mechanisms. Scan diagnosis analysis maximizes failure analysis success rate. siemens.com/tessent-yield-learning Tessent Yield Learning from Siemens EDA Join us at booth #401

edfas.org 1 ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 4 ABOUT THE COVER See page 51 for a description of the contest images collage on the cover. A RESOURCE FOR TECHNICAL INFORMATION AND INDUSTRY DEVELOPMENTS NOVEMBER 2024 | VOLUME 26 | ISSUE 4 edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS Electronically Viable TEM Samples with PFIB and STEM EBIC William A. Hubbard Scanning TEM electron beam-induced current (STEM EBIC) imaging is a promising technique for providing high-resolution electronic and thermal contrast as a complement to TEM’s physical contrast. DEPARTMENTS Four-dimensional Scanning Transmission Electron Microscopy: Part III, Ptychography Aaron C. Johnston-Peck and Andrew A. Herzing The final part of this series covers ptychography, a form of computational imaging that recovers the phase information imparted to an electron beam as it interacts with a specimen. Author Guidelines Author guidelines and a sample article are available at edfas. org. Potential authors should consult the guidelines for useful information prior to manuscript preparation. 2 4 2 GUEST EDITORIAL Nicholas Antoniou 36 EDFAS AWARDS 38 ASM AWARDS 40 CALL FOR PAPERS 41 ISTFA 2024 EXHIBITORS LIST 42 ISTFA 2024 EXHIBITOR SHOWCASE 44 BOARD CANDIDATE PROFILES James Demarest 48 DIRECTORY OF FA PROVIDERS Rosalinda Ring 50 EDUCATION NEWS Navid Asadi 51 ABOUT THE COVER 52 PRODUCT NEWS Ted Kolasa 54 TRAINING CALENDAR Rosalinda Ring 56 ADVERTISERS INDEX Nondestructive 3D X-ray Microscopy Speeds Throughput in New Failure Analysis Workflows Cheryl Hartfield This article shows the application of 3D XRM to nondestructively detect non-optimized assembly processes that can influence local stresses and overall device reliability, making it useful for process development as well as FA. 14 For the digital edition, log in to edfas.org, click on the “News & Magazines” tab, and select “EDFA Magazine.” Celebrating 50 Years of ISTFA Nicholas Antoniou A look back at ISTFA over the years from the first gathering in 1975. 14 4 27 Precise Final Specimen Thinning by Concentrated Argon Ion Beam Milling of Plan View TEM Specimens C.S. Bonifacio, Y. Yu, M.L. Ray, M. Skowronski, and P.E. Fischione Xenon plasma focused ion beam specimen preparation is ideal for preparing plan view TEM specimens due to its large-volume-milling capabilities. 20 27 20

edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 4 2 There is so much to celebrate but first I’d like to thank those of you who started it all and set us on this long and fruitful path. ISTFA’s creation can be traced to a meeting held in 1975 in the basement of a TRW building. It lasted half a day, you had to bring your own lunch, and was attended mostly by IEEE Reliability Group folks. The reason they gave for having this meeting was captured in this statement: “FA is important to us; why is it hidden in dark rooms? Let’s talk about it.” The following year (1976) the meeting was already starting to look and feel like a conference. It lasted almost two days, was in a proper venue (Newport Beach Hotel), and had 20 formal papers and proceedings. They named this meeting ATFA (Advanced Techniques in FA). It was very much a local gathering, and the organizers came from TRW, Hughes Aircraft, Rockwell International, ITT Cannon, and the Aerospace Corp. The meeting lost money both years and IEEE lost interest. ATFA was created as a not-for-profit corporation by D. McCormac, Leon Hamiter, L. Kashar, Jim Richardson, and Bob Myers. This team changed the focus by bringing in exhibitors so attendees could try out equipment for FA, highlighting practical techniques and networking. In 1979, the name was changed to the International Symposium for Testing and Failure Analysis (ISTFA), participation increased, and it ran for three days with parallel sessions. Representatives from Japan, Europe, Africa, Taiwan, South Korea, India, Poland, Romania, and even Iran were involved. This was before the Berlin wall fell and before the Shah of Iran was overthrown. In 1986, ISTFA was transferred to ASM International where it has remained ever since. In 1998, largely due to the significant growth of FA interests and the microelectronics industry, ASM launched a new affiliate society, the Electronic Device Failure Analysis Society (EDFAS). ISTFA is the premiere event for EDFAS while still in the ASM family. ADVANCED FAILURE ANALYSIS TECHNIQUES FIRST ANNUAL SEMINAR An all-day seminar will be conducted on advanced techniques for failure analysis of state-of-the-art microelectronics, including semiconductors, thin and thick film hybrids, and joining and processing methods. This is the description of the first meeting held on January 18, 1975, of what was to become the International Symposium on Testing and Failure Analysis (ISTFA). A $6.00 charge for each attendee covered a steak lunch and refreshments during the breaks. NOVEMBER 2024 | VOLUME 26 | ISSUE 4 A RESOURCE FOR TECHNICAL INFORMATION AND INDUSTRY DEVELOPMENTS ELECTRONIC DEVICE FAILURE ANALYSIS GUEST EDITORIAL CELEBRATING 50 YEARS OF ISTFA Nicholas Antoniou nicholas.antoniou@kla.com edfas.org (continued on page 39) PURPOSE: To provide a technical condensation of information of interest to electronic device failure analysis technicians, engineers, and managers. Nicholas Antoniou Editor/KLA nicholas.antoniou@kla.com Mary Anne Fleming Director, Journals, Magazines & Digital Media Joanne Miller Senior Editor Victoria Burt Managing Editor Allison Freeman Production Supervisor ASSOCIATE EDITORS Navid Asadi University of Florida Guillaume Bascoul CNES France Felix Beaudoin GlobalFoundries Michael R. Bruce Consultant Jiann Min Chin Advanced Micro Devices Singapore Edward I. Cole, Jr. Sandia National Labs Michael DiBattista Varioscale Inc. Rosine Coq Germanicus Universitié de Caen Normandie Szu Huat Goh Qualcomm Ted Kolasa Northrop Grumman Space Systems Rosalinda M. Ring NenoVision Tom Schamp Materials Analytical Services LLC David Su Yi-Xiang Investment Co. Martin Versen University of Applied Sciences Rosenheim, Germany FOUNDING EDITORS Edward I. Cole, Jr. Sandia National Labs Lawrence C. Wagner LWSN Consulting Inc. GRAPHIC DESIGN Jan Nejedlik, jan@designbyj.com PRESS RELEASE SUBMISSIONS magazines@asminternational.org Electronic Device Failure Analysis™ (ISSN 1537-0755) is published quarterly by ASM International®, 9639 Kinsman Road, Materials Park, OH 44073; tel: 800.336.5152; website: edfas. org. Copyright © 2024 by ASM International. Receive Electronic Device Failure Analysis as part of your EDFAS membership. Non-member subscription rate is $175 U.S. per year. Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by ASM International for libraries and other users registered with the Copyright Clearance Center (CCC) Transactional Reporting Service, provided that the base fee of $19 per article is paid directly to CCC, 222 Rosewood Drive, Danvers, MA 01923, USA. Electronic Device Failure Analysis is indexed or abstracted by Compendex, EBSCO, Gale, and ProQuest. Antoniou

edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 4 4 EDFAAO (2024) 4:4-11 1537-0755/$19.00 ©ASM International® FOUR-DIMENSIONAL SCANNING TRANSMISSION ELECTRON MICROSCOPY: PART III, PTYCHOGRAPHY Aaron C. Johnston-Peck and Andrew A. Herzing National Institute of Standards and Technology, Gaithersburg, Maryland aaron.johnston-peck@nist.gov INTRODUCTION The final part of this series on four-dimensional scanning transmission electron microscopy (4D-STEM) covers the topic of ptychography. Ptychography is a form of computational imaging that recovers the phase information imparted to an electron beam as it interacts with a specimen, and which is subsequently lost during the detection process. STEM detectors are only sensitive to the amplitude of the electron exit wave. This amplitude information is encoded as intensity within a diffraction pattern. The phase information of the electron exit wave is lost in this process, the so-called “phase problem.” Ptychography can algorithmically recover this phase information to produce an image of the sample in the form of the transmission function. The computed transmission function represents both the modulus and the accumulated phase difference, relative to free space, of an electron wave transmitted through the sample. Its successful retrieval offers a route to several materials characterization methods since it contains information about the structure and properties of the specimen. In ptychography, the first step is to acquire electron scattering data from multiple points on the specimen with a known spacing and redundancy in the sampling. As in other 4D-STEM techniques, a 4D dataset is assembled by acquiring 2D diffraction patterns at each position of a 2D sampling grid. A key aspect of ptychography is that the spatial coverage of the 2D sampling grid is deliberately oversampled such that the illuminated area of the specimen at each position overlaps, as depicted in the schematic of Fig. 1, Panel I. This overlap in sampling eliminates ambiguities in the phase solution and aids convergence of the computation. It was Walter Hoppe who postulated this principal of using multiple diffraction patterns to eliminate ambiguities when solving the phase problem and who also coined the term ptychography,[1,2] although the current implementations of ptychography have evolved considerably from what Hoppe first described. More information on sampling requirements,[3,4] as well as additional information on the history, development, and fundamentals of ptychography can be found elsewhere.[5,6] Several algorithms have been used for ptychographic treatment of electron microscopy datasets.[7-13] This re- view limits the discussion to STEM data collected in the far-field where there are two primary experimental implementations defined by whether the electron probe is focused or defocused at the sample plane (Fig. 1, Panel II). There are considerations and benefits to each approach. For example, using a focused probe, as one would for conventional imaging, affords the possibility to acquire complementary datasets with high spatial resolution. This complementary signal could be a conventional imaging signal or a spectroscopic signal such as electron energy loss spectroscopy (EELS) or energy dispersive x-ray spectroscopy (EDX). By contrast, using a defocused probe illuminates a larger area of the sample, which can increase throughput and lower the applied dose rate to reduce radiation damage. A practical example of how ptychography functions is the commonly used extended ptychographical iterative engine (ePIE) algorithm. A schematic of the workflow is shown in Fig. 1, Panel III.[8] Essentially ePIE solves an optimization problem, where a model describing the interaction between the electron probe and the sample and the subsequent formation of the diffraction pattern is iteratively updated and compared to experimental data until a convergence criteria is satisfied. The exit wave, or the electron wave emanating from the sample, ψ(r, R), is modeled as the product of two complex functions describing the electron probe, P(r), and the object, O(r), as the probe is shifted relative to the sample by a distance, R. The propagation of the exit wave to the detector is then represented by a Fourier transform and accordingly the intensity at the detector is equal to I(u) = |F[O(r)P(r - R)]|2. Therefore, by identifying the correct phases for each diffraction pattern, each exit wave and the functions O(r) and P(r) can be determined. Guesses of the probe and object functions initiate the algorithm, from which an exit wave

edfas.org 5 ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 4 and diffraction pattern are then calculated. The modulus of the calculated diffraction pattern is then replaced with the square root of the experimentally measured data. An updated exit wave is back calculated and new values for O(r) and P(r) are determined using update functions. Each diffraction pattern is analyzed, and both the object and probe functions are updated at each step. A single cycle of the ePIE algorithm is complete once this process is conducted for all the diffraction patterns within the 4D dataset. This process is repeated until some convergence metric is satisfied. Upon convergence, the probe and object function are reconstructed. An example ptychographic result showing a reconstruction of the sample is shown in Fig. 2, Panel I, while examples of reconstructed probes can be viewed in references.[6,8] MEASURING THE PHASE SHIFT: PHASE CONTRAST IMAGING AND MORE The reconstructed transmission function quantifies the change in modulus and phase of the electron wave as it transmits through the sample. We will focus on the phase component of the reconstruction as it affords sev- eral benefits when compared to conventional STEM imaging modes. High angle annular dark field (HAADF) imaging is widely used because its incoherent character[14] makes image interpretation relatively straightforward and the signal intensity is sensitive to sample thickness and atomic number. However, HAADF-STEM lacks the necessary dynamic range to simultaneously image low and high atomic number elements, whereas ptychography can achieve this. For example, ptychography has been used to resolve both the anion and cation atomic columns in LaB6, [15] GaN,[16] and Li containing cathode materials.[17,18] This behavior is exemplified in Fig. 2, Panel I, where both the Ga and N atomic columns are observed in the phase image, while only the Ga atomic columns are observed in the HAADF image. Traditional phase contrast imaging modes (e.g., bright field) can also simultaneously resolve light and heavy atoms, but there can be ambiguities such as contrast reversals when interpreting the image, and Fig. 1 Panel I: A cartoon of a 2D array of probe positions where the area of illuminated specimen overlaps between each probe position. Panel II: A schematic of STEM with a thin sample using a defocused probe (a), a focused probe (b), on a thick sample (c), and the depiction of how a thick sample is modeled using the multislice algorithm. Panel III: A flowchart describing the iterative ePIE algorithm. Reproduced with permission from Maiden and Rodenburg.[8] (a) (b) (c)

edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 4 6 results should be compared to simulations to ensure accurate interpretation. By comparison, the phase signal resulting from ptychography tends to be linear and amenable to quantification. Furthermore, when comparing ptychography with conventional imaging techniques ptychography has been reported to be more dose effi- cient.[13,19-22] This dose efficiency facilitates the study of materials that are particularly sensitive to radiation damage, such as zeolites,[23] halide perovskites,[24] polymers,[25] and biological specimens.[26] The change in phase of the electron is sensitive to a host of phenomena beyond determining the atomic arrangement of a material. For example, charge redistribution due to changes in bonding, which might be a result of a defect, can be detected.[27] Magnetic potentials will also induce a phase shift and can be distinguished from the electrostatic potential of the sample.[28] Taking advantage of this, Chen et al. mapped the distribution of skyrmions in an FeGe sample.[29] Notably, the reported spatial resolution of the FeGe ptychography reconstruction was better than an induction map derived using center of mass (COM) data analysis (which was covered in Part II).[29] This ability of ptychography to surpass the resolution limits imposed by the optics of the microscope is a significant benefit and covered in the next section. SUPER RESOLUTION The spatial resolution of STEM is many times larger than the fundamental limit posed by the electron wavelength due to lens aberrations and diffraction from the probe forming aperture. Increasing the convergence semi-angle (α) mitigates diffraction effects[30] and increases the infor- mation limit.[31] However, α cannot be increased bound- lessly due to the geometric and chromatic aberrations of the microscope lens, which become resolution limiting as α is increased. As a result, the spatial resolution of microscopes without aberration correction is about 0.1 to 0.3 nm, depending on focal length of the objective lens and the electron wavelength set by the accelerating voltage.[30,32-34] Aberration corrected microscopes, which are equipped with advanced optical elements that reduce the lens aberrations, can achieve resolutions of 0.06 nm or better.[35] Aberration correctors are a hardware solution to improving spatial resolution, which comes at great cost and increased microscope complexity. Whereas ptychography offers a software-based approach, which has even been demonstrated to achieve “super resolution,” where the resolution of the reconstructed images surpasses the physical limit of the microscope. Super resolution is possible because the diffraction patterns in a ptychographic dataset are not corrupted by lens aberrations or other sources of incoherence and instability, which limit conventional image formation. The first demonstration of super resolution was by Nellist et al. where the reconstructed image possessed a spatial resolution of 0.136 nm using data collected on a microscope whose incoherent image resolution at best would be approximately 0.27 nm.[34] Due to the dependence of resolution on several factors, including wavelength, it can be difficult to make comparisons, and these quoted values start to lack context. Table 1 compares traditional imaging and ptychography through a set of performance metrics normalizing the reported resolution (d) as a function of wavelength (λ). In short, d/λ decreased with aberration correction and now even further with ptychography. Fig. 2 Panel I: A reconstructed modulus and phase image of GaN [210] compared to traditional ADF and annular bright field (ABF) images. The Ga and N atomic columns are clearly observed in the phase image. Reproduced under the terms of a Creative Commons CC BY license.[16] Panel II: Reconstruction of PrScO3 [001] (a) showing information transfer to 43.9 nm-1 in the FFT, (b) well beyond the information limit set by the microscope optics. The scale bar is equal to 0.2 nm. Reproduced with permission from Chen et al.[19] (a) (b)

edfas.org 7 ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 4 Since the report by Nellist et al., advances in detector technology, algorithms, and computing power have led to further gains in resolution. A reconstructed image of MoS2 reported spatial frequencies in the diffractogram of 25.6 nm-1 or approximately 0.039 nm. This value was a considerable improvement over the value of 10.2 nm-1 or approximately 0.098 nm observed in the diffractogram of a conventional annular dark field STEM image formed using the same imaging conditions.[21] This study also demonstrated the effect of the maximum scattering angle on the reconstruction by processing the same dataset whose maximum scattering angle was artificially truncated at different multiples of α. As the cutoff angle increased, the maximum frequency in the diffractogram increased and the atoms in the reconstructed image became sharper. This behavior mimics previously derived relationships for the resolution of ankylography and x-ray ptychography reconstructions; the resolution along the lateral (x and y) directions of the reconstruction is described by the relationship dx,y = λ / sin(θ max), where λ is the wavelength, and θmax is the maximum scattering angle in the diffraction pattern.[36] This equation suggests the spatial resolution of ptychography is limited by the maximum diffraction angle at which intensities can be measured with sufficient statistics. However, a study by Chen et al.[19] demonstrated that the motion of the atoms (due to thermal and zero-point effects) also constrains the spatial resolution. This phenomenon was identified by analyzing a reconstruction of PrScO3 (Fig. 2, Panel II) to quantify the different contributions that broadened the projected width of individual atomic columns. And while atomic motion poses a limit to the attainable resolution, sensitivity to the effects of atomic motion coupled with the spatial resolution of ptychography creates new measurement opportunities, such as extracting Debye-Waller factors of individual atomic columns. Such a level of sensitivity would be particularly useful around interfaces or defects, as existing techniques to measure Debye-Waller factors sample large regions thereby making it difficult to detect highly localized variations. OPTICAL SECTIONING AND THREEDIMENSIONAL STRUCTURE DETERMINATION Many electron ptychography studies to date involve nano or 2D materials, and this is not by chance but rather by design. The approximation introduced earlier to describe the exit wave as a product of the probe and the sample functions becomes untenable in thicker samples because it uses two dimensional functions to approximate three-dimensional objects. Conceptually, the breakdown Table 1 A compilation of references detailing resolution as a function of technique and accelerating voltage Description Accelerating voltage, kV Resolution, nm d/λ Reference High voltage TEM on an uncorrected instrument 1250 0.098 133.2 Ichinose, 1999[33] Intermediate voltage STEM on an uncorrected instrument 200 0.136 54.2 James, 1999[32] Intermediate voltage STEM with early generation geometric aberration correction 100 0.136 40.6 Delby, 2001[56] Early demonstration of electron ptychography 100 0.136 36.7 Nellist, 1995[34] Intermediate voltage STEM with geometric aberration correction 300 0.0405 20.6 Morishita, 2018[35] Low voltage STEM with geometric aberration correction 30 0.107 15.3 Sawada, 2015[57] Low voltage TEM with geometric and chromatic aberration correction 40 0.09 15.0 Linck, 2016[58] Ptychography, intermediate voltage, defocused probe, multislice algorithm 300 0.023 11.7 Chen, 2021[19] Ptychography, low voltage, focused probe, ePIE algorithm 80 0.039 9.3 Jiang, 2018[21]

edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 4 8 of this approximation can be visualized in Fig. 1, Panel IIc. In the ray diagram, the probe is focused at the entrance surface of the sample and, because the sample is sufficiently thick, the probe becomes increasingly divergent as it approaches the exit surface. As a result, a single probe function cannot accurately represent a range of diverse states. Relationships describing the limit of the multiplicative approximation have been proposed[9,37,38] and these models predict a value on the order of a few nanometers for commonly used experimental conditions. However, this simplistic picture does not account for dynamical scattering and electron-sample interactions, which modify the phase and modulus of the electron probe, meaning the multiplicative approximation can potentially fail in even thinner samples.[19] To address this problem, an extension to the single slice or 2D ptychography algorithms was introduced.[12,39] It is coined the “multislice” approach, deriving its name from the multislice image simulation technique, which treats the sample in a similar fashion[40] and models the specimen as a number of slices or layers, such that the multiplicative approximation remains valid for each individual slice. The algorithm then solves for a series of transmission functions rather than one. Accordingly, the exit wave of the first layer is calculated as the product of the probe and the transmission function, as we introduced for a 2D ptychography algorithm. It is then propagated to the next object slice and is used as the probe function for the next slice and so on until the final slice, as illustrated in Fig. 1, Panel IIc. Finally, the exit wave is propagated using a Fourier transform to calculate the intensity at the detector plane. The back calculation proceeds in a similar fashion. This approach was first demonstrated by examining two overlapping carbon nanotubes (CNTs).[41] A traditional STEM image would be a projection of the entire sample showing the two CNTs simultaneously; where the depth sampling of the ptychographic reconstruction was sufficiently fine that the upper and lower CNTs could be resolved separately and with sufficient spatial resolution to discern the structure of each (Fig. 3, Panel I). Importantly, this technique has been demonstrated to work on samples that are “thick” and would fail to reconstruct accurately in a single slice or 2D ptychography algorithm,[19] enabling ptychography to be used on a much broader set of samples. Thus far, the multislice approach has been used to characterize the depth dependent lattice variations due to strain and polarization around a dislocation core in SrTiO3 (Fig. 3, Panel II) [42] as well as the distribution of oxygen vacancies in a zeolite.[23] (a) (c) (b) Fig. 3 Panel I: Optical sectioning of two carbon nanotubes overlapping in projection. The phase images reconstructed at different depths can resolve the CNTs individually while reconstructing the lattice. Scale bars are equal to 10 nm. Reproduced under the terms of a Creative Commons Attribution 4.0 International License.[41] Panel II: Optical sectioning of a kinked edge dislocation in SrTiO3. The sum of all phase images (a), along with individual slices at 2.4nm, 6.4nm, and 12.0nm in depth (c). Schematics of the kink configuration (b). Scale bars are equal to 0.5 nm. Reproduced under the terms of a Creative Commons Attribution 4.0 International License.[42]

edfas.org 9 ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 4 Because STEM data are 2D projections of a 3D structure, it can be difficult to unambiguously detect or fully quantify features that vary through the thickness of the sample. Thus, 3D ptychographic reconstructions create many opportunities, such as improved accuracy in the dimensional metrology of rough interfaces or characterizing point defect complexes and extended defects, such as dislocations. These methods could also be used to isolate real sample features from artifacts induced during the sample preparation process. The true lattice strain of a thinned sample could be measured without complication from surface relaxations which inevitably occur to name but one example. The scope of applications realized will depend on the magnitude of the depth resolution. Compared to the lateral resolution, the depth resolution of multislice ptychographic reconstructions are at least an order of magnitude worse (e.g., depth resolution values of approximately 4 to 23 nm have been reported[19,23,41,42]) and are currently insufficient to resolve the 3D structure with atomic resolution. One potential pathway to improve the depth resolution of ptychographic reconstructions is to incorporate tomographic methods, where the 3D structure of a sample is reconstructed from a series of 2D images acquired at different orientations. Results from multislice ptychography advantageously showed that the phase change at different atomic positions was linear as a function of thickness[19] and therefore satisfy the projection requirement where the recorded dataset should exhibit a linear relationship to an integral of a physical property along the projection direction.[43] By comparison, even HAADF-STEM signals can exhibit non-linear behavior,[44] which suggests that ptychography will be a valuable addition to the 3D structure determination toolbox. Implementation could be done by serially recording and reconstructing ptychography data at multiple tilt angles and then combining with a tomography rou- tine.[45,46] This approach has been demonstrated to resolve individual atoms and point defects on datasets generated by numerical simulations[45] and was used to solve the structures of ZnTe filled CNTs[46] and a DNA origami scaffold.[47] This serial approach is not the only possible workflow and there are proposed algorithms that integrate both in a single reconstruction routine to generate the 3D reconstructed volume.[48] Furthermore, these examples were using 2D ptychographic reconstructions as input for the tomographic reconstruction. An alternative approach is to combine tilt-series acquisition with optical sectioning and multislice ptychography. Combining optical sectioning with a tilt-series has improved the resolution of the tomographic reconstruction for conventional STEM imaging.[49] Therefore, this approach when combined with ptychography could potentially lead to tomographic reconstructions with 3D atomic resolution. One such report on a Co3O4 nanoparticle has recently been made. [50] RECONSTRUCTION FIDELITY While the results discussed demonstrate the power of ptychography, some caution is warranted as there is no guarantee that a reconstructed image faithfully represents the structure or properties of the sample. This is exemplified in Fig. 4 where a 2D reconstruction algorithm is applied to an increasingly thick sample and fails because the algorithm did not appropriately model the physical parameters of the probe and sample. Other reports in the literature identify additional artifacts or limitations and attempt to address them.[6,19,51,52] In a systematic treatment, Cao et al. examined different experimental limitations that are present in the illumination, sampling, and detection processes and discussed algorithms to account for these imperfections.[53] Using these improved algorithms, a previously published experimental dataset,[54] whose reconstruction exhibited artifacts, was reprocessed and exhibited improved fidelity. Other approaches to facili- tate high fidelity reconstructions include statistical Fig. 4 Reconstruction of PrScO3 [001] at different depths comparing a multislice (top row) and 2D or single-slice algorithm (bottom row). The 2D reconstruction algorithm fails to faithfully reconstruct the structure at larger thicknesses. The scale bar is equal to 0.2 nm. Reproduced with permission from Chen et al.[19]

edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 4 10 methods to aid reconstruction parameter selection to ensure a specific algorithm is being used optimally.[55] SUMMARY This three-part series on 4D-STEM touched upon different imaging modalities, dimensional metrology, defect characterization, strain, electromagnetic field measurement and more. The intent was to convey the potential of 4D-STEM as an indispensable tool for materials characterization and highlight some applications that would benefit the electronic device characterization community. Keep in mind that 4D-STEM has not yet reached maturity, and the scope and accessibility of the techniques are still increasing. The authors expect adoption of 4D-STEM methods to grow and increasingly address scientific problems inaccessible to other techniques. REFERENCES 1. W. Hoppe: “Beugung Im Inhomogenen Primarstrahlwellen- feld. I. Prinzip Einer Phasenmessung von Elektronenbeungungsinterferenzen,” Acta Crystallographica Section A, 1969, 25(4), p. 495-501. 2. R. Hegerl and W. Hoppe: “Dynamische Theorie Der Kristallstrukturanalyse Durch Elektronenbeugung im Inhomogenen Primärstrahlwellenfeld,” Berichte der Bunsengesellschaft für physikalische Chemie, 1970, 74(11), p. 1148-1154. 3. T.B. Edo, et al.: “Sampling in X-Ray Ptychography,” Physical Review A, 2013, 87(5), p. 053850. 4. C. Gilgenbach, X. Chen, and J.M. LeBeau: “A Methodology for Robust Multislice Ptychography,” Microscopy and Microanalysis, 2024, 30(4), p. 703-711. 5. J.M. Rodenburg: “Ptychography and Related Diffractive Imaging Methods,” Advances in Imaging and Electron Physics, 2008, 150, p. 87-184. 6. J. Rodenburg and A. Maiden: “Ptychography,” Springer Handbook of Microscopy, 2019, p. 819-904. 7. J.M. Rodenburg and H.M.L. Faulkner: “A Phase Retrieval Algorithm for Shifting Illumination,” Applied Physics Letters, 2004, 85(20), p. 4795-4797. 8. A.M. Maiden and J.M. Rodenburg: “An Improved Ptychographical Phase Retrieval Algorithm for Diffractive Imaging,” Ultramicroscopy, 2009, 109(10), p. 1256-1262. 9. J.M. Rodenburg and R.H.T. Bates: “The Theory of Super-Resolution Electron Microscopy via Wigner-Distribution Deconvolution,” Philosophical Transactions of the Royal Society of London. Series A: Physical and Engineering Sciences, 1992, 339(1655), p. 521-553. 10. C.T. Putkunz, et al.: “Atom-Scale Ptychographic Electron Diffractive Imaging of Boron Nitride Cones,” Physical Review Letters, 2012, 108(7), p. 073901. 11. T.J. Pennycook, et al.: “Efficient Phase Contrast Imaging in STEM using a Pixelated Detector. Part 1: Experimental Demonstration at Atomic Resolution,” Ultramicroscopy, 2015, 151, p. 160-167. 12. A.M. Maiden, M.J. Humphry, and J.M. Rodenburg: “Ptychographic Transmission Microscopy in Three Dimensions using a Multi-Slice Approach,” J. Opt. Soc. Am. A, 2012, 29(8), p. 1606-1614. 13. Z. Chen, et al.: “Mixed-State Electron Ptychography Enables SubAngstrom Resolution Imaging with Picometer Precision at Low Dose,” Nature Communications, 2020, 11(1), p. 2994. 14. P.D. Nellist and S.J. Pennycook: “Incoherent Imaging using Dynamically Scattered Coherent Electrons,” Ultramicroscopy, 1999, 78(1), p. 111-124. 15. P. Wang, et al.: “Electron Ptychographic Diffractive Imaging of Boron Atoms in LaB6 Crystals,” Scientific Reports, 2017, 7(1), p. 2857. 16. H. Yang, et al.: “Electron Ptychographic Phase Imaging of Light Elements in Crystalline Materials using Wigner Distribution De- convolution,” Ultramicroscopy, 2017, 180, p. 173-179. 17. J.G. Lozano, et al.: “Low-Dose Aberration-Free Imaging of Li-Rich Cathode Materials at Various States of Charge using Electron Ptychography,” Nano Letters, 2018, 18(11), p. 6850-6855. 18. W. Song, et al.: “Direct Imaging of Oxygen Shifts Associated with the Oxygen Redox of Li-Rich Layered Oxides,” Joule, 2022, 6(5), p. 1049-1065. 19. Z. Chen, et al.: “Electron Ptychography Achieves Atomic-Resolution Limits Set by Lattice Vibrations,” Science, 2021, 372(6544), p. 826-831. 20. T. Seki, Y. Ikuhara, and N. Shibata: “Theoretical Framework of Statistical Noise in Scanning Transmission Electron Microscopy,” Ultramicroscopy, 2018, 193, p. 118-125. 21. Y. Jiang, et al.: “Electron Ptychography of 2D Materials to Deep SubÅngström Resolution,” Nature, 2018, 559(7714), p. 343-349. 22. T.J. Pennycook, et al.: “High Dose Efficiency Atomic Resolution Imaging via Electron Ptychography,” Ultramicroscopy, 2019, 196, p. 131-135. 23. H. Zhang, et al.: “Three-Dimensional Inhomogeneity of Zeolite Structure and Composition Revealed by Electron Ptychography,” Science, 2023, 380(6645), p. 633-638. 24. A. Scheid, et al.: “Electron Ptychographic Phase Imaging of BeamSensitive All-Inorganic Halide Perovskites using Four-Dimensional Scanning Transmission Electron Microscopy,” Microscopy and Microanalysis, 2023, 29(3), p. 869-878. 25. B. Hao, et al.: “Atomic-Scale Imaging of Polyvinyl Alcohol Crystallinity using Electron Ptychography,” Polymer, 2023, 284, p. 126305. 26. L. Zhou, et al., “Low-Dose Phase Retrieval of Biological Specimens using Cryo-Electron Ptychography,” Nature Communications, 2020, 11(1), p. 2773. 27. C. Hofer, et al.: “Detecting Charge Transfer at Defects in 2D Materials with Electron Ptychography,” doi.org/10.48550/arXiv.2301.04469. 28. J. Cui, et al.: “Antiferromagnetic Imaging Via Ptychographic Phase Retrieval,” Science Bulletin, 2023, doi.org/10.1016/j.scib.2023. 12.044. 29. Z. Chen, et al.: “Lorentz Electron Ptychography for Imaging Magnetic Textures Beyond the Diffraction Limit,” Nature Nanotechnology, 2022, 17(11), p. 1165-1170. 30. M. Weyland and D. A. Muller: “Tuning the Convergence Angle for Optimum STEM Performance,” FEI Nanosolutions, 2005, 1(24), p. 24-35. 31. Y. Peng, et al.: “Spatial Resolution and Information Transfer in Scanning Transmission Electron Microscopy,” Microscopy and Microanalysis, 2008, 14(1), p. 36-47. 32. E.M. James and N.D. Browning: “Practical Aspects of Atomic Resolution Imaging and Analysis in STEM,” Ultramicroscopy, 1999, 78(1), p. 125-139. 33. H. Ichinose, et al.: “Atomic Resolution HVEM and Environmental Noise,” Journal of Electron Microscopy, 1999, 48(6), p. 887-891. 34. P.D. Nellist, B.C. McCallum, and J.M. Rodenburg: “Resolution Beyond the ‘Information Limit’ in Transmission Electron Microscopy,” Nature, 1995, 374(6523), p. 630-632. 35. S. Morishita, et al.: “Attainment of 40.5 pm Spatial Resolution using 300 kV Scanning Transmission Electron Microscope Equipped with Fifth-Order Aberration Corrector,” Microscopy, 2017, 67(1), p. 46-50. 36. A. Suzuki, et al.: “High-Resolution Multislice X-Ray Ptychography of Extended Thick Objects,” Physical Review Letters, 2014, 112(5), p. 053903.

edfas.org 11 ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 4 37. P. Thibault, et al.: “High-Resolution Scanning X-Ray Diffraction Microscopy,” Science, 2008, 321(5887), p. 379-382. 38. E.H.R. Tsai, et al.: “X-Ray Ptychography with Extended Depth of Field,” Opt. Express, 2016, 24(25), p. 29089-29108. 39. T.M. Godden, et al.: “Ptychographic Microscope for ThreeDimensional Imaging,” Opt. Express, 2014, 22(10), p. 12513-12523. 40. E. Kirkland: “Computation in Electron Microscopy,” Acta Crystallographica Section A, 2016, 72(1), p. 1-27. 41. S. Gao, et al.: “Electron Ptychographic Microscopy for ThreeDimensional Imaging,” Nature Communications, 2017, 8(1), p. 163. 42. H. Sha, et al.: “Sub-Nanometer-Scale Mapping of Crystal Orientation and Depth-Dependent Structure of Dislocation Cores in SrTiO3,” Nature Communications, 2023, 14(1), p. 162. 43. P.A. Midgley and M. Weyland: “3D Electron Microscopy in the Physical Sciences: The Development of Z-Contrast and EFTEM Tomography,” Ultramicroscopy, 2003, 96(3), p. 413-431. 44. W. Van den Broek, et al.: “Correction of Non-Linear Thickness Effects in HAADF STEM Electron Tomography,” Ultramicroscopy, 2012, 116, p. 8-12. 45. D.J. Chang, et al.: “Ptychographic Atomic Electron Tomography: Towards Three-Dimensional Imaging of Individual Light Atoms in Materials,” Physical Review B, 2020, 102(17), p. 174101. 46. P.M. Pelz, et al.: “Solving Complex Nanostructures with Ptychographic Atomic Electron Tomography,” Nature Communications, 2023, 14(1), p. 7906. 47. Z. Ding, et al.: “Three-Dimensional Electron Ptychography of Organic– Inorganic Hybrid Nanostructures,” Nature Communications, 2022, 13(1), p. 4787. 48. J. Lee, et al.: “Multislice Electron Tomography using FourDimensional Scanning Transmission Electron Microscopy,” Physical Review Applied, 2023, 19(5), p. 054062. 49. R. Hovden, et al.: “Breaking the Crowther Limit: Combining Depth-Sectioning and Tilt Tomography for High-Resolution, WideField 3D Reconstructions,” Ultramicroscopy, 2014, 140, p. 26-31. 50. M.G.C. Andrey Romanov, et al.: “High-Resolution 3D Phase-Contrast Imaging Beyond the Depth of Field Limit via Ptychographic MultiSlice Electron Tomography,” arXiv, doi.org/10.48550/arXiv.2311. 02580. 51. C. Hofer and T.J. Pennycook: “Reliable Phase Quantification in Focused Probe Electron Ptychography of Thin Materials,” Ultramicroscopy, 2023, 254, p. 113829. 52. H. Yang, et al.: “Enhanced Phase Contrast Transfer using Ptychography Combined with a Pre-Specimen Phase Plate in a Scanning Transmission Electron Microscope,” Ultramicroscopy, 2016, 171, p. 117-125. 53. S. Cao, A.M. Maiden, and J.M. Rodenburg: “Image Feature Delocalization in Defocused Probe Electron Ptychography,” Ultramicroscopy, 2018, 187, p. 71-83. 54. M.J. Humphry, et al.: “Ptychographic Electron Microscopy using High-Angle Dark-Field Scattering for Sub-Nanometre Resolution Imaging,” Nature Communications, 2012, 3(1), p. 730. 55. M.C. Cao, et al.: “Automatic Parameter Selection for Electron Ptychography via Bayesian Optimization,” Scientific Reports, 2022, 12(1), p. 12284. 56. N. Dellby, et al.: “Progress in Aberration-Corrected Scanning Transmission Electron Microscopy,” Journal of Electron Microscopy, 2001, 50(3), p. 177-185. 57. H. Sawada, et al.: “Atomic-Resolution STEM Imaging of Graphene at Low Voltage of 30 kV with Resolution Enhancement by using Large Convergence Angle,” Physical Review Letters, 2015, 114(16), p. 166102. 58. M. Linck, et al.: “Chromatic Aberration Correction for Atomic Resolution TEM Imaging from 20 to 80 kV,” Physical Review Letters, 2016, 117(7), p. 076101. ABOUT THE AUTHORS Aaron C. Johnston-Peck is a materials research engineer at the National Institute of Standards and Technology where he characterizes materials using various electron microscopy techniques. Prior to NIST, he worked as a postdoctoral fellow at Brookhaven National Laboratory and received his Ph.D. in materials science and engineering from North Carolina State University. He has contributed to 50 peer reviewed publications. Andrew A. Herzing is a materials research engineer at the National Institute of Standards and Technology where he develops material characterization techniques with a particular interest in tomography and 4D-STEM. He received his Ph.D. in material science and engineering from Lehigh University in 2007 and is currently serving as the president-elect of the Microanalysis Society. Advertise in Electronic Device Failure Analysis magazine! For information about advertising in Electronic Device Failure Analysis: KJ Johanns, Business Development Manager 440.671.3851, kj.johanns@asminternational.org Current rate card may be viewed online at asminternational.org/advertise.

edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 4 12 Please see us at ISTFA Booth #605

edfas.org 13 ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 4

edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 4 14 NONDESTRUCTIVE 3D X-RAY MICROSCOPY SPEEDS THROUGHPUT IN NEW FAILURE ANALYSIS WORKFLOWS Cheryl Hartfield, FASM Carl Zeiss Microscopy LLC, Dublin, California cheryl.hartfield@zeiss.com EDFAAO (2024) 4:14-19 1537-0755/$19.00 ©ASM International® INTRODUCTION The market for high performance computing (HPC) has surged due to the increasing use of artificial intelligence (AI) and machine learning across a growing number of industries. This is fueling the adoption of heterogenous integration and “chiplet” computing architectures, driving a need for tightly packaged interconnects in 2D and 3D to increase feature density (Fig. 1). As the size and complexity of the die and packaging increase, so do the characterization and failure analysis (FA) challenges. It is becoming more challenging to identify defects and their locations quickly and nondestructively to effectively get to the root cause of the failure.[1,2] New FA workflows that leverage 3D x-ray microscopy (XRM) along with femtosecond lasers (fs-laser) and AI training models[3] can speed up nondestructive fault detection with repeatable recipes. This article shows the application of 3D XRM to nondestructively detect non-optimized assembly processes that can influence local stresses and overall device reliability, making it useful for process development as well as FA. The example discussed is thermocompression bonding (TCB) at 20 µm bump pitches. ZEISS Versa 3D x-ray microscopes have submicron resolution and are well-suited for imaging these structures.[4] When used along with AI training models, 3D XRM can achieve analysis of highly integrated packaging structures with reasonable throughput for process validation and error correction guidance. A second example shows the effectiveness of 3D XRM to guide and verify the precise and targeted sample preparation of a 3D package using a fs-laser integrated FIB-SEM, with the goal to enable functional testing and fault isolation with very little damage to the package and IC. In both use cases, combining 3D XRM and deep learning achieves a robust and repeatable nondestructive analysis method for faster insights on highly integrated packaging structures. 3D DATA FOR COMPREHENSIVE CHARACTERIZATION OF TCB PROCESSES As interconnect pitch transitions below 100 µm, thermocompression bonding (TCB) is required. TCB enables Fig. 1 Fan-out and 2.5D interconnects achieve high interconnect densities, while the recent commercial adoption of hybrid bonding has enabled further exponential increases in interconnect density per area.

edfas.org 15 ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 4 larger die as well as chip stacking, because localized reflow allows better control of chip gap height and tilt throughout the bonding process. High accuracy and repeatability in all dimensions are required to provide a reliable systemin-package.[5] To assess the qualitative and quantitative ability of 3D XRM to determine both the alignment accuracy of bump/ via stacks and the quality of TCB interconnects, a 15-die stack made by Interuniversity Microelectronics Centre (imec) was selected as the test vehicle. It consisted of a patterned test chip type-O (PTCO) die sequentially stacked vertically and bonded onto a patterned test chip type-P (PTCP) substrate chip with 20 µm solder bump pitch (Fig. 2). Each PTCO die contains a 5 μm high Cu bump 12.5 μm in diameter on the top side, connected to an 8.5 μm Cu/Sn bump 7.5 μm in diameter, connected by a 5 μm diameter Cu TSV. Alignment and bond validation involve analysis of 3828 peripheral bump/via stacks.[6] The highest-resolution scans performed in this study used voxels of 0.7 µm to scan narrow-pitch bumps located in each corner of the array. These are key for die-to-die alignment in the TCB tool and are identical in pattern, aside from any expected rotation due to their location. To speed up the workflow, an AI model was trained on a single corner of narrow-pitch bumps by doing a two-hour scan, followed by a 15-minute scan of another corner that applied a deep-learning algorithm. After all volumetric data was collected, the software rendered it into 3D virtual slices to achieve a large data set in a noticeably short amount of time while ensuring the analysis of the Cu pillar alignment and solder bump width and height would be statistically relevant. Fig. 2 TSV shift from die 1-15 can be noted by positional shift relative to the white vertical line in this colorized virtual XRM cross-section. The 0.7 μm-voxel-resolution 3D data volume from which the slice was extracted is shown in the colorized cross-sectional 3D rendering in the right-hand image. Fig. 3 Statistical data comparison for mean and standard deviation of TSV shift between all iterations of scan methodology. Bump diameter and height were also assessed (not shown).

edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 26 NO. 4 16 To validate the scans that applied a deep-learning algorithm, an additional 20.5-hour scan was run to serve as ground truth. It was also used to train an AI model that was applied on five- and two-hour scans. Representative images of a virtual slice revealing TSV misalignment and a 3D XRM image showing the density and volume of structures analyzed are shown in Fig. 2. Several representative interconnects were used as a benchmark to determine the threshold for analytics on bump width, height, and via alignment, and the algorithm was subsequently applied across the entire dataset from the FOV of the sample. A final comparison of the analytical data set was plotted across all scans to determine whether the deep learning-assisted model provided accurate information for via alignment and bump dimensions across all 15 layers in the stack. Figure 3 shows data for mean TSV misalignment (left) and standard deviation (right). The 15-minute scans were validated by their good agreement to 20-hour scans with a standard deviation of less than a half micron. This analysis of the entire chain from dies 1 to 15 showcases a novel 2D/3D x-ray microscopy alignment and inspection solution for TCB in a flip chip FOWLP with a workflow that allows for both quantitative evaluation of TSV and bump alignment in the x, y-plane as well as bump dimensions in the z-plane. By using 3D XRM along with AI models, it is possible to conduct analysis of highly integrated packaging structures with reasonable throughput for process development, validation, and error correction guidance. In addition, the scan times achieved by application of deep learning for reconstruction are practical for FA quality checks of failed products. Figure 4 shows a buried defect that was identified within the 3D XRM volume. Because the process in the use case was exclusively optimized for local interconnect alignment as the top die stack terminates at the final die level, global alignment was not a requirement. For future applications where the top die interconnect is required or global shift is of concern, an additional alignment step on a fixed reference point to the bottom substrate can be done quite easily to compensate. There is still opportunity to improve throughput by implementing emerging deep learning and data reconstruction solutions and validating scans on the larger FOV, which could potentially enable a full scan of complete arrays in 3D at throughput acceptable for an inline 3D x-ray platform. X-RAY-GUIDED LASER CUTS FOR PACKAGE-LEVEL FAULT ISOLATION In another application,[7] 3D XRM is used to guide and verify the precise and targeted sample preparation of a multi-die package using a fs-laser integrated FIB-SEM to enable functional testing and fault isolation while imparting minimal damage to the package and IC. The electrical connections going to the different components in 3D packages may have complex circuitry that could lead to challenging fault isolation routines to identify the failure sites. Isolating some of the components in these packages to determine failure sites with higher accuracy requires deactivating certain features or parts of the circuit. A baseband modem IC from the motherboard of a mobile phone, assembled as a 3D package consisting of one flip-chip die (baseband processor) connected to the substrate through solder bumps and another die (memory and/or analog) with wire bonds, illustrates this complexity, making it the ideal candidate for testing. With the right tools and workflows, it is possible to selectively break an interconnect or wire while maintaining sufficient integrity of the chip/package for functional testing. The targeted structures for physical alteration must be accessible either through the molding compound or other protective packaging materials. In the combined high-resolution nondestructive 3D XRM and LaserFIB workflow, a Fig. 4 A defect was detected in the 3D XRM scan. A 10 µm buried underfill void looks isolated in a virtual cross section slice (upper right) and is revealed to be more extensive by the plan-view slice (bottom left).

RkJQdWJsaXNoZXIy MTYyMzk3NQ==