Recrystallization of shock-melted Si observed in real-time
Washington State University researchers have directly observed shock-induced melting and recrystallization of silicon on nanosecond timescales. As they report in a recent issue of Physical Review Letters, the researchers observed the melting through in situ, time-resolved x-ray diffraction (XRD) measurements at pressures above 30 GPa. This work adds new constraints to the high-temperature, high-pressure phase diagram of silicon and suggests that the technique could similarly reveal structural changes in other materials under shock-wave compression.
“Melting and freezing are two of the most ubiquitous materials phenomena,” says Yogendra Gupta, senior researcher on the project. “Typically, these phenomena are viewed as slow—because our familiarity with these phenomena is mostly based on observations at long time scales. Hence, some questions that have been around for a long time are: How fast can these phenomena occur? And what is the nature of the material state?” These are fundamental questions, Gupta says, but the answers also have applications in ballistics, debris-spacecraft collisions, and planetary formation, among other areas.
Shock experiments—in which a material is rapidly compressed by a shock wave resulting from high-velocity impact or energy deposition from a high intensity laser beam—enable researchers to study materials under extreme pressures and temperatures. However, generating the intense conditions is only part of the experimental challenge. Researchers must also be able to obtain measurements in these conditions on the appropriate time scales.
Traditionally, researchers have identified shock-induced phase transitions by monitoring the volume change in a sample. More recently, solid–solid phase transitions have been identified via in situ XRD measurements that reveal changes in crystalline structure. However, melting is usually accompanied by a small change in volume, according to the researchers, and the XRD patterns of a liquid can be difficult to differentiate from that of an amorphous solid. For these reasons, the direct and unambiguous detection of a shock-induced solid–liquid phase transition has remained elusive despite indirect evidence that it occurs.
Recent advances in XRD capabilities have enabled in situ, time-resolved, and real-time measurements on nanosecond timescales. Inspired by this, WSU’s Gupta, Stefan Turneaure, and Surinder Sharma studied the shock-melting and recrystallization of silicon at the Dynamic Compression Sector, a user facility for synchrotron-based compression studies that opened in 2016 at Argonne National Laboratory’s Advanced Photon Source. They chose silicon because of its cubic, monoatomic crystalline structure and its prevalence in modern technology, as well as the expectation that it melts on the nanosecond timescale at high pressures.
Using XRD, the researchers characterized silicon’s response to shock compression every 153.4 ns. The samples ranged in thickness from 0.5 mm to 1.6 mm and shock waves were created by impacting the sample with a lithium fluoride (LiF) flyer plate to pressures of 26–54 GPa, the range within which melting was predicted to occur. The samples were attached to a stiff backing of LiF so that the shock waves reflected back through the material.
At pressures under about 31 GPa, the diffraction line profile of a silicon sample included several sharp peaks, reflective of its simple hexagonal crystalline structure at these pressures. However, at 31–35 GPa, the shape of the profile morphed into that of a single, broad diffraction ring—the pattern characteristic of a liquid or an amorphous solid. Furthermore, when the sample was reshocked by the reflected wave, the silicon resolidified, this time with a hexagonal close-packed crystalline structure.
Although the transition of sharp peaks into a diffuse diffraction ring could be evidence of a liquid or amorphous solid phase, the calculated temperature profile during the transition supports a liquid phase. In addition, the silicon recrystallization was not accompanied by the large temperature increase normally required to achieve the transition from an amorphous solid to a ground state structure. Therefore, say the researchers, the results are unambiguous and this is the first known observation of the recrystallization of a shock-melted solid in real time.
The researchers are now focused on extending these measurements to faster timescales and other classes of materials. They want to collaborate with theorists to create a more complete picture of materials behavior under extreme conditions. “Silicon is a model system and can serve as a benchmark for testing our theoretical understanding of the time-dependence and the material state associated with melting and refreezing,” Gupta says.
“This work represents a major advance in understanding the high pressure phase diagram and kinetic response of silicon,” says Richard Kraus, an expert in shock physics at Lawrence Livermore National Laboratory who was not associated with this research. “Their utilization of the plate impact facility at the DCS is a novel method for obtaining shock-molten and re-solidified states of matter. I look forward to this capability extending to metals such as Fe, Ta, and Mo, where there is still significant debate as to the high pressure melt curve,” Kraus says.
Read the abstract in Physical Review Letters.