Ultrafast, high resolution, phase contrast imaging of impact response with synchrotron radiation

Understanding and predicting the dynamic response of materials at extreme conditions requires experimental investigations of their time, rate, and microstructure dependencies. Requirements for such experiments are the need for real-time, in situ, spatially resolved measurements which are challenging. For high strain-rate loading, traditional experiments have relied on optical techniques^1,2 (velocity and displacement interferometry) that monitor interface motion, or stress gauges;³ dynamic proton radiography shows high penetration power but is currently limited in spatial resolution.⁴ The development in synchrotron X-ray photon sources (high coherency and flux) and detection/measurement techniques (e.g., phase contrast imaging or PCI^5–8) offers unique opportunities for ultrafast, high-resolution measurements to examine dynamic materials response. Dynamic PCI measurements with synchrotron X-rays have been performed on the microsecond timescale;⁸ however, impact events typically occur on the sub-microsecond timescale requiring higher time resolution in the picosecond to nanosecond range. In this Letter, we report ultrafast (<100 ps), high resolution (∼3 μm), dynamic PCI measurements on representative materials/processes using a single synchrotron X-ray bunch during impact loading. Such measurements are expected to be valuable for revealing novel phenomena under high rate loading and studying the underlying mechanisms responsible for material failure,⁹ jet formation in metals,¹⁰ and hotspot formation in explosive¹¹ as well as more fundamental studies of dynamic material properties including phase transitions and equation-of-state.

Ultrafast PCI measurements were performed at the 32ID beamline¹² of the Advanced Photon Source (Argonne, IL). As shown in Fig. 1, the PCI measurement used a “white” beam which was transmitted in sequence through two shutters, a 2D slit, the sample, a scintillator (Lu₃Al₅O₁₂:Ce) with ∼55 ns decay time¹³ positioned 660 mm away from the sample, and then imaged onto the detector using standard optics. The optical components consisted of a 45° mirror to relay the optical emission from the scintillator into the PI-MAX ICCD camera (Princeton Instruments; ∼20 μm pixel size, and 1300×1340 pixels) and a 10× Mitutoyo infinity-corrected long working distance objective (Edmund NT46-144; 33.5 mm working distance and 3 μm focal depth). The field of view on the sample was approximately 1.6 mm×1.6 mm. Note that the sample-scintillator distance was not optimized due to the constraints of the current translation stages, and such constraints will be eliminated for better image contrast in future experiments. The storage ring was operated in the standard mode (24 bunches), with pulses of 80 ps (fwhm) duration spaced 153.3 ns apart. The beam intensity and spectra were adjusted by varying the undulator gap with typical operation in the 11-30 mm range. Most of our measurements used an undulator gap of 26 mm. For this gap, the majority of the intensity was located in the peak centered around 11.9 keV with a bandwidth of 0.6 keV FWHM, and the peak flux (per 1mm2 aperture) was about 1.7× 10¹⁴ photons/s/0.1%bw.

Materials were subjected to impact loading using a 12.6-mm bore light-gas gun capable of achieving velocities up to 1 km/s and designed specifically for use at a synchrotron source. A schematic of the experimental configuration is shown in Fig. 1. The gun system consisted of a gas breech, a launch tube (or barrel), and a target chamber all mounted on a mobile support structure (not shown) to allow for insertion and alignment within the X-ray beam. The X-ray beam entered through a side port, was transmitted through a sample, and exited through a second side port where the detector system was located. The side ports were sealed using Lexan windows (approximately 0.01 inch thick) to allow the X-rays to pass through while maintaining vacuum prior to the experiment. During the experiment, the projectile accelerated down the launch tube and impacted the target, generating a compressive wave in the sample. Projectile velocities were measured using standard photonic Doppler velocimetry (PDV).² 

Synchronization of the impact event, the incident X-ray beam, and the detector was achieved using piezoelectric impact pins (Dynasen, Inc.), two electromechanical shutters (slow and fast response time) placed between the X-ray source and the sample, and the gun control system (not shown). A 12 VDC signal obtained from the gun control system which corresponds to the projectile launch was used to trigger the slow shutter allowing the X-ray beam to pass through the sample. At impact, the electrical signal from piezoelectric impact pin was used to externally trigger a Stanford Research Systems (SRS) DG535 delay generator. Output signals from the delay generator (with appropriate delays) were used to trigger both the ICCD (using a gain of 200 and a gate width of 150 ns) to acquire the image and the fast shutter to interrupt the X-ray beam. The operation of these two shutters yielded a sample exposure time of about 60 ms, protecting both the sample and the scintillator/mirror components downstream from the sample while providing enough time to bracket the shock event. An example of the various timing signals generated during an experiment are shown in Fig. 1 (inset).

Three exploratory experiments were performed in this work to illustrate our ability to obtain spatially resolved images of dynamically compressed materials. Schematics of the experimental configurations are shown in Fig. 2. Experiments 1 and 2 used aluminum projectiles to launch 300-μm stainless steel cylinders into vitreous carbon (VC) and boron carbide (BC) target plates, respectively. The goal of these two experiments was to observe cylinder deformation during impact and the subsequent material response of the target plate (cracking, spall, etc.). Experiment 3 used an aluminum projectile to impact a micro-truss foam sample that was prepared using an optical waveguide method¹⁴ to observe real-time compression of an engineered material. Nominal sample dimensions for all experiments along the beam axis (z-axis) was 9 mm and the measured projectile velocities were 0.619 km/s, 0.657 km/s, and 0.35 km/s for experiments 1, 2, and 3, respectively.

The experimental results are shown in Figs. 3 and 4. Pre-shot images are shown in the upper left corner of each figure. The apparent layered structure visible in two of the pre-shot images (Fig. 3) is an artifact caused by sample tilt/rotation with respect to the beam direction. All images were background-corrected and scale bars were produced using a calibrated gold grid (63.5 μm) placed in the beam. The image for experiment 1 (Fig. 3 top) shows the cylinder penetrating the plate and resulting in visible plastic deformation of the cylinder along with evidence of spallation and ejecta in the vitreous carbon. In contrast, the image for experiment 2 (Fig. 3 bottom) shows significant plastic deformation of the cylinder with minimal penetration in the boron carbide plate. Additional material response is evidenced by the propagation of cracks through the carbide. In experiment 3, an aluminum projectile impacted the micro-lattice foam with approximate cell dimensions of 0.9-mm in width and 1.1-1.4 mm in length. A portion of the truss lattice is visible in the pre-shot image (Fig. 4; inset). The dynamic image (Fig. 4) shows the projectile compacting the foam resulting in the collapse of void 1 and partial collapse of voids 2 and 3.

The experimental results shown here clearly establish the dynamic phase contrast imaging technique using a single X-ray bunch (80-ps fwhm) to capture synchronized dynamic events with 3 μm spatial resolution on nanosecond timescales. This new capability is expected to reveal novel phenomena and to allow the examination of the rich underlying physics for materials subjected to high strain rates. Applications include studies of material strength¹⁵ and failure,⁹ compaction,^16,17 and hotspot formation in energetic materials.¹¹ Experiments are underway to use dynamic PCI to examine jet formation in metals, to perform high strain-rate Taylor cylinder impact, to observe compaction of idealized borosilicate spheres, and to develop a multi-frame detector capability. With such synchrotron-based platforms, dynamic shock experiments can fully exploit the unique advantages of synchrotron X-ray sources for ultrafast imaging, diffraction and spectroscopy, and for developing the necessary knowledge base for the best use of the next-generation photon sources such as X-ray free electron lasers.

This work was performed at Argonne National Laboratory. Jim Esparza, Chuck Owens, and Alex Deriy are gratefully acknowledged for their help with experimental setup and shot execution. T. Schaedler and W. Carter (HRL, Santa Barbara, CA) are both thanked for supplying the foam for one of the experiments. Use of the Advanced Photon Source, a user facility operated for the U.S. Department of Energy (DOE) Office of Science by Argonne National Laboratory, was supported by the U.S. DOE under Contract No. DE-AC02-06CH11357. We are grateful for the support from the Science Campaign and LDRD programs at LANL. Los Alamos National Laboratory is operated by Los Alamos National Security, LLC, for the National Nuclear Security Administration of the U.S. Department of Energy under contract DE-AC52-06NA25396.