The information era, which lasted roughly 60 years, gave the world the internet, smart phones, and super-fast computers. This has been made feasible by about doubling the number of transistors that can be packed onto a computer chip every two years, resulting in billions of atomic-scale transistors that currently fit on a fingernail-sized device. Individual atoms may be observed and counted within such “atomic scale” lengths.
The physical limit
With this doubling reaching a physical limit, the United States Department of Energy’s (DOE) Princeton Plasma Physics Laboratory (PPPL) has joined industry efforts to prolong the process and discover new techniques to make ever-more competent, efficient, and cost-effective chips. In the first PPPL research under a Cooperative Research and Development Agreement (CRADA) with Lam Research Corp., a global producer of chip-making equipment, laboratory scientists properly anticipated a fundamental stage in atomic-scale chip creation using modeling.
“This would be one small piece of the puzzle,” said David Graves, associate laboratory director for low-temperature plasma surface interactions, professor in the Princeton Department of Chemical and Biological Engineering, and co-author of a paper describing the findings published in the Journal of Vacuum Science & Technology B. Modeling insights, he added, “may lead to all kinds of positive things, and that’s why our endeavor at the Lab has some potential.”
While the shrinking can’t last much longer, “it hasn’t fully stopped,” he says. “To far, industry has been successful in developing creative new processes mostly via empirical approaches, but a better basic knowledge will accelerate this process. Fundamental research demand time and skill that the industry does not always have “He said. “This provides labs with a significant incentive to take up the assignment.”
The PPPL researchers modeled “atomic layer etching” (ALE), a fundamental manufacturing process that attempts to remove single atomic layers from a surface at a time. This method may be used to etch intricate three-dimensional patterns onto a coating on a silicon wafer with crucial dimensions hundreds of times thinner than a human hair.
The fundamental accord
“As a first step, the simulations largely agreed with the results and might lead to a better understanding of the application of ALE for atomic-scale etching,” said Joseph Vella, a postdoctoral scholar at PPPL and main author of the journal publication. “It all begins with increasing our basic knowledge of atomic layer etching,” he said, adding that improved understanding would allow PPPL to explore topics like the level of surface damage and the degree of roughness created during ALE.
On an atomic scale, the model mimicked the sequential employment of chlorine gas and argon plasma ions to manage the silicon etch process. Plasma, also known as ionized plasma, is a gas composed of free electrons, positively charged ions, and neutral molecules. In contrast to the ultra-hot plasma utilized in fusion research, the plasma used in semiconductor device production is around room temperature.
“A surprising empirical discovery from Lam Research was that the ALE method became more efficient when the ion energies were quite a bit greater than the ones we began with,” Graves said. “So our next step in the simulations will be to see if we can understand what happens when the ion energy is significantly greater and why it’s so excellent.”
In the future, “the semiconductor industry as a whole is envisioning a significant increase in the materials and kinds of devices to be employed, and this expansion will also need atomic scale accuracy,” he added. “The United States’ mission is to lead the world in utilizing research to solve critical industrial challenges,” he said, “and our work is part of that.”
The DOE Office of Science provided some funding for this work. David Humbird of DWH Consulting in Centennial, Colorado, was one of the coauthors.