The Complex, Unique, and Powerful Imaging Instrument for Dynamics (CUPI²D) at the Spallation Neutron Source (invited)

Attenuation-based neutron radiography and computed tomography have contributed for decades to a broad range of scientific research applications at both continuous^1–8 and pulsed sources.^9–12 At research reactors and pulsed sources, research areas such as energy applications,^13–23 materials science,^24–29 engineering,^30–35 geosciences,^36–42 plant physiology,^43–52 archaeometry,^53–60 and medicine^61–64 have flourished over the past 20+ years. More recently, technological prowess in detector apparatus has demonstrated spatial resolution reaching a few μm.^65–67 Advanced techniques such as polarized neutron imaging^68–72 and neutron grating interferometry^73–80 have achieved new contrast mechanisms and detection sensitivity for features far below the pixel resolution by merging small angle scattering and neutron imaging. Pulsed sources provide a unique contrast mechanism based on the time-of-flight (TOF) information, i.e., the capability to determine the neutron’s wavelength based on the TOF of the neutron. Novel wavelength-dependent techniques have enabled measurements of microstructure,^{9,30,31,81–85} strain,^{30,35,83,86–95} and elemental and/or isotopic content in materials,^96–103 such as engineered components and geomaterials. Over the past ten years, the interest generated by these novel TOF capabilities has led to the nascence of a new class of wavelength-dependent neutron imaging facilities at worldwide pulsed sources such as IMAT¹² at ISIS (Rutherford Appleton Laboratory) in the United Kingdom and RADEN¹⁰⁴ (Japanese Proton Accelerator Research Complex) in Japan; the development of dedicated capabilities at other facilities such as Los Alamos Neutron Science Center (LANSCE);¹⁰ and the construction of VENUS (Versatile Neutron Imaging Instrument at SNS) at Oak Ridge National Laboratory in the United States,¹⁰⁵ ODIN at the European Spallation Neutron Source (ESS),^106,107 and ERNI at the Chinese Spallation Neutron Source (CSNS).¹⁰⁸ 

The SNS is an accelerator-based pulsed neutron source that is outfitted with 19 state-of-the-art neutron scattering instruments optimized to mostly use thermal neutrons at its First Target Station (FTS) that are geared to provide unique capabilities across a broad range of scientific disciplines. Plans are underway to construct a Second Target Station (STS),¹⁰⁹ providing a high intensity beam of cold neutrons that will support complementary instruments, including one instrument optimized for TOF imaging. This beamline, named the Complex, Unique, and Powerful Imaging Instrument for Dynamics (CUPI²D), will be equipped with the necessary optics and detectors to perform real time in situ Bragg edge imaging (BEI)⁸³ and neutron grating interferometry (nGI).^75,77,110 The combination of these capabilities and the high source intensity will allow for unprecedented fast measurements across a broad range of length and time scales currently unavailable at existing neutron imaging beamlines. CUPI²D will be designed to complement the existing and future neutron imaging instruments at the Oak Ridge National Laboratory (ORNL): the HFIR CG-1D Multimodal Advanced Radiography Station (MARS)^111,112 and the Versatile Neutron Imaging Instrument at SNS (VENUS)¹⁰⁵ presently under construction with an anticipated completion date of summer 2024. MARS provides a high flux of cold neutrons and is being upgraded to provide both high spatial resolution imaging capabilities and white beam (i.e., wavelength-independent) neutron grating interferometry. VENUS is designed to offer a broad range of neutron wavelengths over a large fully illuminated field-of-view of 20 × 20 cm², with neutron energies ranging from epithermal to thermal, hence permitting the measurement of resonances and Bragg edges, respectively. The CUPI²D imaging beamline will extend the range of materials and phenomena to be studied with the help of neutron imaging techniques. The cold neutron spectrum of the STS will extend the wavelength range beyond 10 Å needed for various studies, which require long data collection times at the existing flight paths due to the low flux at these wavelengths.

CUPI²D encompasses a diverse scientific portfolio and will play a significant role in a broad range of scientific topics, including materials science, energy (batteries, fuel cells, and nuclear), engineering, life sciences, biology, geosciences, medicine (cancer, biomaterials for scaffolds, and biofilm formation), dentistry (implants), and archaeometry.⁸ Since CUPI²D’s capabilities permit fundamental science studies as well as applied research, the expected user base for the instrument will include researchers from academia, national laboratories, and industry. This paper briefly highlights a few scientific applications that will benefit from the future capabilities at CUPI²D.

With a predicted cold neutron flux 20 times higher than that available at FTS VENUS¹⁰⁵ (see the discussion in the technical section), CUPI²D will be designed to perform real-time investigations of in situ dynamic processes in advanced materials such as energy storage devices^113–115 and superalloys such as those utilized in aerospace,^84–86 under extreme environments generated by furnaces, load frames, pressure and gas cells, etc. Both the materials’ phases and crystalline properties can be characterized using BEI, an atomic-level technique, as illustrated in Refs. 30, 83–85, 89, and 116. Briefly, the combination of BEI and nGI increases the probing length scale, from the atomic level to a few μm, with each technique providing enhanced contrast at a relevant scale. While BEI is sensitive to a material’s phases and crystalline properties, nGI is based on scattering effects similar to small angle neutron scattering (SANS).⁷⁷ Most importantly, the instrument will enable the simultaneous performance of transmission imaging/tomography, BEI, and nGI, allowing us to quantify material and system kinetics at multiple length scales and at a time resolution ranging from seconds to a few minutes. CUPI²D will be capable of associating several properties needed to be studied in situ and in real time, such as microporosity (nGI),^117–119 preferred microstructure orientation,²⁵ and strain (BEI)^{84,86,88,94,120} in materials such as additively manufactured (AM) metals and other superalloys. The beamline will provide experimental data that, in turn, inform predictive models capable of anticipating stress-induced porosity changes and fracture evolution. CUPI²D will contribute to the validation and scalability of AM components in failure-critical areas, such as aerospace parts and biomedical implants. The degradation and failure mechanisms in structures such as suspension bridges will be an opportune example of in situ experiments. Progress has been made on understanding the internal mechanics of multi-body wires in suspension bridges using neutron diffraction, but to date, it has not been possible to perform spatially resolved live and in situ imaging of degradation, overload, and plastic flow evolution of components (the latter of which can only be detected through imaging since plastic deformation mechanisms are independent of atomic lattice spacing resolved in diffraction).^121,122

CUPI²D will allow the imaging of full-size energy storage cells in operando over a full range of cycling conditions. The extension of conventional tomography to 4D (3D spatial and time) imaging capabilities will make available dynamic volumetric studies of the distribution and movement of light elements such as lithium and hydrogen in material components and in complete devices. While nGI focuses on mapping spatial non-uniformities in constituent materials and components, BEI measures the chemical phases in the functioning battery, as demonstrated in the literature.^113,114,123 Localized changes in material morphology, such as particle pulverization and changes in porous regions due to material expansion, will be observed with nGI. Localized state-of-charge, inhomogeneities, and other crystal structure changes that may arise during operation will be observed with BEI. Together, these capabilities will provide a multiscale understanding of the electrochemical processes in novel battery systems that will be transformational in developing new technologies in the field of energy storage.

Biological materials, such as plants, tissues, and microbes, are difficult to image using x-ray methods due to their low interaction with photons and potential for biological tissue damage. In addition, while neutrons have been successfully used to track water content and exchange between soils and roots,^36–52 it has been particularly difficult to image plant root–soil chemical interactions in situ due to their high water content. CUPI²D will provide powerful new capabilities in “indirect” high resolution imaging using nGI of light elements [e.g., carbon (C), oxygen (O), nitrogen (N), and minerals], the basic building blocks of biological materials. This capability allows the in situ dynamic visualization of many natural phenomena that, to date, have been studied using other indirect methods. The BEI mode and nGI capabilities will allow the dynamic assessment of specific compounds (e.g., root nutrient uptake and root C release) and their interactions with water and soil surfaces to provide a map of crystalline phases (minerals) and porosity quantities. This research will revolutionize our understanding of plant/soil/water/nutrient/C relations and soil C sequestration capacity, leading to improved climate models and crop/land management strategies.

CUPI²D will be a multimodal, multi-length scale instrument capable of characterizing both natural and engineered materials, from the Å to the cm length scale at acquisition times ranging from seconds to minutes. These capabilities are achieved by providing BEI and nGI transformative capabilities simultaneously, as needed. Table I displays the source parameters, capability, and equipment requirements based on the instrument’s anticipated scientific portfolio. The equipment requirements were collected from the scientific community during the competitive proposal process and are not necessarily explained throughout this section. Table I illustrates the types of capabilities required to successfully perform the proposed science, including detector technology and sample environment (SE). As such, this beamline’s preliminary design will evolve with the advances in the various technical fields of study that will benefit CUPI²D.

The desired performance capabilities will be achieved by selecting a moderator with the highest flux of cold neutrons, utilizing guides and advanced optical components while keeping the source-to-detector distance sufficient for a Bragg-edge resolution of δλ/λ ≈ 0.003.^{31,82,83,87,89,91,92,120,124}

BEI is an ideal technique for a pulsed neutron source^30,125 since it focuses on measuring the abrupt changes in neutron transmission due to the crystalline structure or phase of the sample as a function of neutron wavelength (with a wavelength resolution on the order of δλ/λ = 0.2%–0.5%). In fact, a sharp change in transmission occurs when the neutron wavelength, λ, is equal to twice the atomic spacing, d, for a specific <hkl> atomic plane. The height and position of the Bragg edge can provide phase and strain information, as illustrated in Refs. 27, 31, 81, 86, 89, 92, and 94. The combination of Bragg edge transmission with the small pixel size of an imaging detector is ideal for measuring phase/strain variations that would be averaged over a bulk measurement using, for example, diffraction.

The particle–wave duality of neutrons allows the measurement of a phase shift of a neutron beam due to the real part of an object’s refraction index. Measurements of the neutron phase shift can be performed using a grating-induced coherent (in time and space) source of neutrons.^73,74 Materials such as titanium (Ti)- and nickel (Ni)-based superalloys have similar linear attenuation coefficients, and thus, an attenuation-based measurement cannot separate the two materials. However, their phase shifts are negative and positive, respectively, allowing them to be separated in 2D and 3D phase measurements. This technique is called Differential Phase Imaging (DPI), and the contrast comes from the refraction at interfaces between materials. Moreover, the grating interferometry system provides a high angular resolution that can be exploited to detect ultra-small-angle-scattering (USANS) effects. With this technique—called Dark Field Imaging (DFI)—measurements of structures from nm to μm are possible, thus tremendously enhancing the imaging spatial information, which is not achievable when taking attenuation-based radiographs using conventional pinhole geometry systems. A parameter called the autocorrelation length, also known as the dark field length, is directly related to the microscopic sample structures that are being measured. Hence, by tuning the dark field length, ξ, one can probe through different length scales. ξ is given by ξ = λL_s/p₁, where λ is the neutron wavelength, L_s is the sample-to-detector distance, and p₁ is the G₁ phase grating period. Commonly, neutron grating interferometers, such as the Talbot–Lau nGI system, are comprised of three gratings: G₀, the source grating; G₁, the phase grating; and G₂, the analyzer grating.¹²⁹ G₀ constructs the source coherence, while G₁ measures the phase/amplitude shift. G₂ is the absorption grating that is stepped perpendicularly to the beam to measure the intensity oscillations that are smaller than a pixel on the detector. Since p₁ is fixed, for a fixed wavelength, L_s must be increased to cover a significant correlation length range, leading to a decrease in image resolution that can be compensated using advanced neutron optics. This is not necessary at CUPI²D since it will provide a broad range of neutron wavelengths simultaneously.

A conceptual design of a combined nGI/BEI setup is displayed in Fig. 1. The three gratings (G₀, G₁, and G₂) are placed close to the aperture such that the flux is maximized. The sample environment and micro-channel plate (MCP) Timepix (TPX) detector are moved upstream, accordingly. Since the source is pulsed and the MCP TPX is capable of timestamping neutrons as they arrive, BEI is effectively performed simultaneously by acquiring wavelength-resolved radiographs. All radiographs are acquired at the wavelength resolution required by BEI, which has the most stringent resolution requirement, and can be later binned into larger wavelength bins for nGI analysis.

CUPI²D’s prime directive is to provide the highest possible flux of wavelength-resolved cold neutrons—achievable only with STS’s uniquely intense cold neutron flux—to perform both BEI and nGI with a sufficient temporal resolution to capture dynamic processes. Hence, wavelengths shorter than ∼2 Å do not need to be propagated through the beamline path. The wavelength resolution is driven by BEI and should be on the order of δλ/λ ≈ 0.003, as demonstrated in previous work.^{31,82,83,87,89,91,92,120,124}

STS is designed to operate at 15 Hz, a lower repetition rate than FTS, which ensures broad wavelength bands. Two moderators are proposed for STS: (1) a cylinder moderator with a viewed area of 3 × 3 cm² that offers the best wavelength resolution (needed for BEI) and (2) a tube moderator with a 3 cm diameter viewing area and a comparatively higher flux than the cylinder moderator. While the tube moderator provides, on average, ∼50% more neutrons, it produces a broader pulse than the cylinder moderator and, thus, a compromise in the wavelength resolution is made unless the instrument is much longer than on a cylinder moderator. The instrument would have to be longer than 50 m, as shown in Fig. 2, which illustrates the different wavelength resolutions achieved at both STS moderators for different source-to-detector distances. The overall frame width is inversely proportional to the moderator-to-detector distance, reducing the beamline’s capability to measure fast kinetics over a broad range of wavelengths as the moderator-to-detector distance increases. Longer instruments also require longer guide systems, increasing the risk of unwanted inhomogeneities in the beam profile at the sample, resulting from misalignments and other imperfections (section gaps) in the guide system. The quality of the measured radiograph and potential neutron CT (nCT) reconstruction depends critically on the uniformity of the beam profile.

CUPI²D is designed to have its most downstream detector position at ≈ 33.5 m, which corresponds to an overall wavelength band of Δλ ≈ 7.89 Å and a wavelength resolution of δλ/λ ≈ 0.3% on a cylinder moderator, matching the wavelength resolution requirements listed in the instrument specifications (Table I). This large instantaneous wavelength band is essential to the CUPI²D scientific case, for example, to use Bragg edges to measure the chemical phases in batteries, which occur over several Å (from 3 Å to ≈ 12 Å), at cold wavelengths that are not available at VENUS due to the lack of flux beyond ≈ 5 Å. A narrow wavelength band, Δλ, would compromise the kinetic measurements of the same time event since not all appropriate wavelengths can be measured at the same time. Selecting a tube moderator jeopardizes the overall Δλ since δλ/λ of 0.3% can only be achieved if the CUPI²D imaging detector is placed at ≈ 54 m, thus corresponding to Δλ ≈ 4.89 Å. Moreover, because the CUPI²D detector can move as close as 21.5 m, the wavelength resolution on a tube moderator of δλ/λ ≈ 1% (see Fig. 2) would be insufficient for BEI for materials science applications that require strain mapping with an accuracy of ≈100 με.^116,120,130 In comparison, VENUS’s detector position is at 25 m, yielding a Δλ ≈ 2.64 Å and δλ/λ ≈ 0.2%.

The simulated averaged brightness at CUPI²D is expected to be at least 20 times higher than the VENUS’s simulated brightness for neutron wavelengths higher than 2.5 Å, as illustrated in Fig. 3. Hence, CUPI²D will be capable of achieving a much higher time resolution on the order of seconds to minutes, as compared to VENUS, which will require minutes to hours per measurement at similar cold neutron wavelengths.

Disk or bandwidth choppers are devices that prevent frame overlap between sequential pulses of neutrons. They act as mechanical bandpass filters and open to let only neutrons of the desired wavelength pass. Most FTS beamlines have three disk choppers to prevent frame overlap. Since the pulse frequency at STS is 15 Hz (4 times slower than FTS), fewer disk choppers are required. Since CUPI²D has a variable moderator-to-detector distance, the disk choppers must allow different neutron wavelength frames, which can be realized by installing a double disk chopper with disks that spin in opposite directions, each with adjustable phases. Most of the beamline optics and components that produce background are installed in the bunker, with the double disk chopper installed at 6 m, followed by the T₀ chopper at 7.5 m. The T₀ chopper decreases the prompt gamma pulse and fast neutron intensities by several orders of magnitude. One double disk chopper is sufficient to stop the frame overlap of neutrons with wavelengths shorter than ∼50 Å, as illustrated in Fig. 4.

Unlike FTS, the first 13.2 m (from the moderator) of an STS instrument is designed with common shielding called the bunker, which is shared by all beamlines located on the same side of the target building. This inevitably limits the floor space available to install imaging optics and introduces complexity in accessing the front-end area of the beamline in case of repair (several beamlines would need to be shut down to access CUPI²D front-end optics). Thus, it is prudent to limit the CUPI²D beamline components in the bunker to components that are absolutely necessary: the T₀ chopper, the frame-defining bandwidth choppers, and components that require heavy shielding, such as the beamline shutter, guides, beam collimators, and filters. These components are placed as far upstream as reasonable since they generate a significant background of neutrons and gammas that should be avoided close to the instrument cave. The STS moderator is bright but small (3 × 3 cm²), which is amenable to “transporting” the neutron source to a downstream location named the “virtual source.” For the abovementioned reasons, CUPI²D is designed to create a virtual source 18 m away from the moderator using an elliptical guide system.

CUPI²D can be divided into three main sections: (1) the bunker, shared by several instruments, starting at the source and ending at 13.2 m; (2) the instrument cave with a back wall at ∼40 m; and (3) the radiological material area, the control hutch, and the user laboratory located directly behind the instrument. Figure 5 is a three-dimensional rendering of the beamline showing the bunker and cave sections and some of their respective aforementioned components.

The instrument cave is designed to allow space for implementing various mission-specific sample environments at the beamline. Complementary modalities are essential for short-lived phenomena such as those in the geosciences and plant systems. In fact, complementary nCT and x-ray computed tomography (xCT) have successfully been employed in many scientific fields, such as soil, batteries,¹³¹ reservoir rocks,¹³² and archaeological samples.^132,133 Magnetic Resonance Imaging (MRI) is another emerging capability that has recently been utilized to visualize plant rhizosphere in 3D.¹³⁴ 

CUPI²D’s largest aperture matches the size of the virtual source and is located upstream of the instrument elevators. A variable aperture system can vary L/D ratios without moving the source-to-detector position (where L is the distance from the pinhole aperture, of diameter D, to the detector). CUPI²D is designed for three main imaging modes, as displayed in Fig. 6. The high-intensity mode corresponds to the most upstream detector position at ∼ 22 m (the original concept was 21.5 m) and, thus, corresponds to the lowest L/D of 100 and a maximum field of view (FOV) of ∼ 4 × 4 cm², with a wavelength band of Δλ ≈ 12.3 Å. The complementary mode is a high wavelength-resolution configuration with a detector positioned at ∼ 34 m with the largest L/D of 1000 and a FOV of 15 × 15 cm² and a wavelength band of Δλ ≈ 7.89 Å. Finally, the intermediate mode at 28 m, with a FOV of 9 × 9 cm² and an L/D of 300, allows for a balance between intensity, FOV, and wavelength resolution. Advanced optics, such as Wolter mirrors, gratings, and polarization, are installed on elevators similar to the current elevator system at the RADEN imaging beamline at J-PARC,¹⁰⁴ which is equipped with interchangeable optics, grating interferometers, polarization, and flight tubes. The installation of these optics on elevators (Fig. 5) allows for swift configuration changes between detector positions and imaging modalities (e.g., nGI may be used for selected measurements only, as it sacrifices flux).

To maximize the neutron flux, a straight, elliptical guide system is adopted to transport the source from the moderator and focuses on the virtual source aperture position at 18 m. The elliptical neutron transport system consists of two half-ellipses that are the left and right halves of two nearly identical ellipses: both are centered at around z = 9 m. Both ellipses have their first focal points slightly before the moderator and their second focal points slightly behind the aperture. Such a design provides the phase space (3 cm in height and width and an ∼0.7° divergence) required for the three L/D collimation settings illustrated in white, black, and blue in Fig. 6. To estimate the performance of the optical design, simulations were performed by combining MCViNE^135,136 and McStas^137–139 simulation scripts to estimate the performance of the optical design. As a representative example, the performance of the neutron transport for L/D = 100 is presented in Fig. 7. Nearly 90% of the source brilliance is transferred for neutrons with wavelengths longer than 2.5 Å. The maximum flux is about 4 × 10⁸ neutrons per second per cm² per Å at λ = 2.5 Å.

These initial simulations have yielded homogeneous intensity distributions within the desired field of view for all three modes. Figure 8 (top) shows the two-dimensional intensity patterns integrated over 1–9 Å at z = 22 m, z = 28 m, and z = 34 m detector positions for L/D = 100, 300, and 1000, respectively. In the case of L/D = 100, the FOV is limited by a region that comprises all pixels with an intensity that is > 90% of the maximum intensity (at the center). In the cases of L/D = 300 and 1000, the FOV is 9 × 9 and 15 × 15 cm², respectively. Further simulations and optimization are underway to account for the gaps introduced by the T₀ and double disk bandwidth choppers, to be reported elsewhere.

As illustrated in Fig. 8 (bottom), the modes corresponding to L/D = 300 and 1000 display less structure within the FOV than the simulation at L/D = 100 because they allow narrower neutron trajectories, i.e., those with a lower divergence through the aperture. As the detector moves closer to the variable aperture system, it can see neutrons with more reflections and a higher divergence.

The future CUPI²D imaging beamline will broaden the science portfolio of the neutron imaging program at Oak Ridge National Laboratory by enabling material characterization across broad length and time scales. CUPI²D will have a transformational impact on scientific areas such as energy storage and conversion (batteries, fuel cells to transform energy, and transportation fields), materials engineering (additive manufacturing and superalloys), nuclear materials (novel fuel cladding and moderators), cementitious materials, biology, and ecosystems (in situ soil–plant fluid/nutrient dynamics), and medical/dental applications (3D printed adaptive implants). The broad and diverse science portfolio is inherently concomitant with a wide-ranging sample environment.

The innovation of this instrument lies in the utilization of a high flux of cold neutrons for performing real time in situ neutron grating interferometry and Bragg edge imaging, with a wavelength resolution of δλ/λ ≈ 0.3%, simultaneously when required, across a broad range of length and time scales. The proposed timeline for the construction of the STS and CUPI²D is 10–15 years from today. There are three main modes at CUPI²D that correspond to three different L/Ds, FOVs, wavelength resolutions, and fluxes at each position. These are called high flux (L/D 100, closest), intermediate (L/D 300), and high wavelength resolution (L/D 1000, farthest). Advanced optics such as Wolter mirrors are being considered at CUPI²D. Further optimization of the beamline concept is underway and will be published in a future article.

A portion of this research used resources at the Spallation Neutron Source, a DOE Office of Science User Facility operated by the Oak Ridge National Laboratory. This research used resources of the Spallation Neutron Source Second Target Station Project at Oak Ridge National Laboratory (ORNL). ORNL is managed by UT-Battelle, LLC, for DOE’s Office of Science, the single largest supporter of basic research in the physical sciences in the United States. This material was based upon the work supported by the US Department of Energy, Office of Science, Office of Biological and Environmental Research. This work was also supported by the US Department of Energy (DOE), Office of Energy Efficiency and Renewable Energy, Bioenergy Technologies Office via the Systems Development and Integration program, and the US Department of Energy, Office of Energy Efficiency and Renewable Energy, Vehicle Technologies Office via the Advanced Combustion Engine Systems program. This research was supported by the DOE Office of Fossil Energy and Carbon Management. This work was also supported by Los Alamos National Laboratory LDRD Project No. 20200109DR.

The authors would like to thank Dr. Ken Andersen, Dr. Ken Herwig, Mrs. Cristina Boone, Mr. Scott Dixon, and Mr. William (Bill) Turner for their invaluable contributions to the concept and engineering design of the CUPI²D beamline. The team would also like to thank colleagues at ORNL for brainstorming discussions about CUPI²D: Dr. Yuxuan Zhang, Dr. Jean-Christophe Bilheux, and Mr. Erik Stringfellow. A special thank you to Dr. Franz Gallmeier for providing the source terms for the VENUS and CUPI²D beamlines. The authors would also like to thank Mr. Harley Skorpenske for sharing his expertise in many fruitful discussions on the implementation of in situ load frames and furnaces in neutron diffraction and their potential implementation on an imaging beamline.

This manuscript has been authored by UT-Battelle, LLC, under Contract No. DE-AC05-00OR22725 with the US Department of Energy (DOE). The US government retains and the publisher, by accepting the article for publication, acknowledges that the US government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for US government purposes. DOE will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan).

The authors have no conflicts to disclose.

Adrian Brügger: Conceptualization (supporting); Project administration (supporting); Writing – original draft (equal); Writing – review & editing (equal). Hassina Z. Bilheux: Conceptualization (lead); Project administration (lead); Writing – original draft (lead); Writing – review & editing (lead). Jiao Y. Y. Lin: Conceptualization (lead); Writing – original draft (supporting); Writing – review & editing (supporting). George J. Nelson: Conceptualization (supporting); Writing – original draft (supporting). Andrew M. Kiss: Writing – original draft (supporting). Jonathan Morris: Writing – original draft (supporting); Writing – review & editing (supporting). Matthew J. Connolly: Writing – original draft (supporting); Writing – review & editing (supporting). Alexander M. Long: Writing – original draft (supporting). Anton S. Tremsin: Conceptualization (supporting); Writing – original draft (supporting). Andrea Strzelec: Writing – original draft (supporting). Mark H. Anderson: Writing – original draft (supporting). Robert Agasie: Writing – original draft (supporting). Charles E. A. Finney: Writing – original draft (supporting). Martin L. Wissink: Writing – original draft (supporting). Mija H. Hubler: Writing – original draft (supporting). Roland J.-M. Pellenq: Writing – original draft (supporting). Claire E. White: Writing – original draft (supporting); Writing – review & editing (supporting). Brent J. Heuser: Writing – original draft (supporting). Aaron E. Craft: Writing – original draft (supporting). Jason M. Harp: Writing – original draft (supporting). Chuting Tan; Writing – original draft (supporting). Kathryn Morris: Writing – original draft (supporting); Writing – review & editing (supporting). Ann Junghans: Writing – original draft (supporting). Sanna Sevanto: Writing – original draft (supporting). Jeffrey M. Warren: Writing – original draft (supporting); Writing – review & editing (supporting). Fernando L. Esteban Florez: Writing – original draft (supporting); Writing – review & editing (supporting). Alexandru S. Biris: Writing – original draft (supporting). Maria Cekanova: Writing – original draft (supporting); Writing – review & editing (supporting). Nikolay Kardjilov: Writing – original draft (supporting); Writing – review & editing (supporting). Burkhard Schillinger; Writing – original draft (supporting); Writing – review & editing (supporting). Matthew J. Frost: Writing – original draft (supporting). Sven C. Voge: Writing – original draft (supporting); Writing – review & editing (supporting).

Data sharing is not applicable to this article as no new data were created or analyzed in this study.