The atoms of amorphous solids like glass have no ordered structure; they arrange themselves randomly, like scattered grains of sand on a beach. Normally, making materials amorphous — a process known as amorphization — requires considerable amounts of energy. The most common technique is the melt-quench process, which involves heating a material until it liquifies, then rapidly cooling it so the atoms don't have time to order themselves in a crystal lattice.
Now, researchers at the University of Pennsylvania School of Engineering and Applied Science (Penn Engineering), the Indian Institute of Science (IISc) and the Massachusetts Institute of Technology (MIT) have developed a new method for amorphizing at least one material — wires made of indium selenide, or In2Se3 — that requires as little as one billion times less power density, a result described in a new paper in Nature . This advancement could unlock wider applications for phase-change memory (PCM) — a promising memory technology that could transform data storage in devices from cell phones to computers.
In PCM, information is stored by switching the material between amorphous and crystalline states, functioning like an on/off switch. However, large-scale commercialization has been limited by the high power needed to create these transformations. "One of the reasons why phase-change memory devices haven't reached widespread use is due to the energy required," says Ritesh Agarwal , Srinivasa Ramanujan Distinguished Scholar and Professor in Materials Science and Engineering (MSE) at Penn Engineering and one of the paper's senior authors.
For more than a decade, Agarwal's group has studied alternatives to the melt-quench process, following their 2012 discovery that electrical pulses can amorphize alloys of germanium, antimony and tellurium without needing to melt the material.
Several years ago, as part of those efforts, one of the new paper's first authors, Gaurav Modi , then a doctoral student in MSE at Penn Engineering, began experimenting with indium selenide, a semiconductor with several unusual properties: it is ferroelectric, meaning it can spontaneously polarize, and piezoelectric, meaning that mechanical stress causes it to generate an electric charge and, conversely, that an electric charge deforms the material.
Modi discovered the new method essentially by accident. He was running a current through In2Se3 wires when they suddenly stopped conducting electricity. Upon closer examination, long stretches of the wires had amorphized. "This was extremely unusual," says Modi. "I actually thought I might have damaged the wires. Normally, you would need electrical pulses to induce any kind of amorphization, and here a continuous current had disrupted the crystalline structure, which shouldn't have happened."
Untangling that mystery took the better part of three years. Agarwal shipped samples of the wires to one of his former graduate students, Pavan Nukala , now an Assistant Professor at IISc and member of the school's Centre for Nano Science and Engineering (CeNSE) and one of the paper's other senior authors. "Over the past few years we have developed a suite of in situ microscopy tools here at IISc. It was time to put them to test — we had to look very, very carefully to understand this process," says Nukala. "We learned that multiple properties of In2Se3 — the 2D aspect, the ferroelectricity and the piezoelectricity — all come together to design this ultralow energy pathway for amorphization through shocks."
Ultimately, the researchers found that the process resembles both an avalanche and an earthquake. At first, tiny sections — measured in billionths of a meter — within the In2Se3 wires begin to amorphize as electric current deforms them. Due to the wires' piezoelectric properties and layered structure, the current nudges portions of these layers into unstable positions, like the subtle shifting of snow at the top of a mountain.
When a critical point is reached, this movement triggers a rapid spread of deformation throughout the wire. The distorted regions collide, producing a sound wave that moves through the material, similar to how seismic waves travel through the earth's crust during an earthquake.
This sound wave, technically known as an "acoustic jerk," drives additional deformation, linking numerous small amorphous areas into a single one measured in micrometers — thousands of times larger than the original areas — just like an avalanche gathering momentum down a mountainside. "It's just goosebump stuff to see all these phenomena interacting across different length scales at once," says Shubham Parate , an IISc doctoral student and co-first author of the paper.
The collaborative effort to understand the process has created fertile ground for future discoveries. "This opens up a new field on the structural transformations that can happen in a material when all these properties come together. The potential of these findings for designing low-power memory devices are tremendous," says Agarwal.
This study was conducted at the University of Pennsylvania School of Engineering and Applied Science, the Indian Institute of Science and the Massachusetts Institute of Technology and supported by the U.S. Office of Naval Research Multidisciplinary University Research Initiatives Program (N00014-17-1-2661), the U.S. National Science Foundation (NSF) Future of Semiconductors competition (#2328743), the U.S. Air Force Office of Scientific Research (FA9550-23-1-0189), the NSF Materials Research Science and Engineering Centers Division of Materials Research (MRSEC/DMR-2309043), and the Anusandhan National Research Foundation Science and Engineering Research Board (CRG/2022/003506) from the Government of India, as well as the facilities at CeNSE and the Advanced Facility for Microscopy and Microanalysis (AFMM), IISc, and the democratized system of usage.
Additional co-authors include Anudeep Tullibilli of IISc; Choah Kwon and Ju Li of MIT; and Andrew C. Meng, Utkarsh Khandelwal, James Horwath, Peter K. Davies and Eric A. Stach of Penn Engineering.
