Moore’s Law, which observes that computer power doubles every two years, is under threat. As scientists continue to improve the semiconductors at the heart of our computers, maintaining the same exponential progress is increasingly difficult. The challenges are intricate, posed by the esoteric laws of nature itself. Moving forward means facing these challenges.
Mardochee Reveil is among those who delve into this arcane frontier. As a PhD student and member of Paulette Clancy’s research group, Chemical and Biomolecular Engineering, Reveil works at the intersection of materials engineering and computational chemistry. He investigates new materials and processes that will enable faster, smaller, and more efficient electronic devices.
“I was impressed by how much work goes into our having computers and smartphones that we can use on a daily basis,” says Reveil. Having spent two summers working for Intel, he grew to appreciate the complexity behind semiconductors, and the rough road ahead for this vital industry. “I became interested in contributing to the field,” he says.
For the past 50 years, most progress in electronics involves reducing the size of individual transistors (electronic switches that grant computers their computing ability) so that more can be packed into a given surface, enabling electronic devices such as smartphones to be as small and powerful as they are today. These transistors are made of silicon. For 50 years silicon served the industry well.
But that relentless progress has now forced the industry to reckon with silicon’s inevitable scaling limits, laid down by the fundamental laws of nature. For instance, with greater transistor density, the heat dissipation becomes too high, increasing cooling requirements and hindering the use of the transistors.
Looking for New Materials and Techniques
“The industry is reaching a point where we need to consider not only how to improve computing power, but also how to do it in a more energy-efficient way,” says Reveil.
New materials called III-V—alloys, containing elements from groups III and V in the periodic table—offer a potential alternative to silicon. Their properties are promising. “Their charge carriers can travel much faster than they can in silicon. Imagine a device 10 times faster and still dissipating the same or less heat and still of the same size,” Reveil says.
The problem is that, unlike in the case of silicon, not much is known about the exact behavior of these materials. This knowledge is necessary if manufacturing were to shift towards using these materials. Take the instance of annealing. Annealing is an important manufacturing step whereby the semiconductor material is heated to change its electric properties. Predicting the temperature change can be very difficult but necessary.
One promising annealing method magnifies the problem: laser-spike annealing utilizes bursts of laser radiation that merely last for sub-milliseconds. “This is a very, very fast process by which the sample is heated at a rate of a million kelvin. You can imagine that it can be very challenging to measure the temperature while it is changing so rapidly,” says Reveil.
Progress with CLASP
Cornell developed software to meet this need—CLASP, Cornell Laser Annealing Simulation Package—which is employed by companies, such as Novellus. As with most software in the semiconductor industry, however, CLASP is tailored for silicon.
In his first six months at Cornell, Reveil successfully expanded CLASP to include an array of new materials, such as Indium Gallium Arsenide (InGaAs) and other III-V materials. His success enables the predictability of laser-spike annealing with non-traditional materials, which is important for new materials’ integration into the silicon-laden industry. His modifications have been incorporated in CLASP’s iteration released on March 2015.
Solutions through Simulations
Reveil helped expand the Van der Pauw technique used to measure the resistivity of thin film sheet semiconductors, which are widely employed in today’s devices. The technique requires measured samples to meet a number of limiting conditions. However, actual samples often do not meet the necessary conditions.
Reveil provided a solution through his computational work. He conducted simulations to calculate the error made if the ideal conditions were not satisfied. He also evaluated a new way of making the same type of measurements. This modification to the technique allowed measurements to be made at the micro level, something that is extremely challenging with the original method.
His contributions won him the Best Poster award at the 2015 annual meeting of Cornell’s Nanofabrication Facility. Asked about the response to his work, Reveil paused. He rejected ‘amazed’ and ‘impressed’ before answering, “They were happy with it,” perhaps a testament to the utility of his computational investigations.
“This is why simulation is interesting,” says Reveil. “Instead of going to the lab and fabricating the materials every time, which is costly and time-consuming, you can use simulations to reproduce the exact conditions.”
“Imagine a device 10 times faster and still dissipating the same or less heat and still of the same size,” Reveil says.
The collaborative relationship between the simulation-oriented Clancy group and the experiment-oriented Thompson group led by Michael O. Thompson, Materials Science and Engineering, is built on this advantage. “The Thompson group will validate our approach, and in turn, we will provide insights that are almost impossible to get from experiments alone,” says Reveil. This collaboration was instrumental in Reveil’s breakthroughs, he says.
The extensive collaboration cutting across different fields and research groups, celebrated among Cornell researchers, is clearly pronounced in Reveil’s work. “My research is highly interdisciplinary,” he says. “I think this is the most interesting aspect of doing research here. In a lot of other places it can be difficult to move away from what your department is doing. And for some reason, it’s very easy here.” Adding, “My adviser is very open to ideas that involve collaboration.”
Exploring the uncharted is not an easy task. In this regard, Reveil is grateful for the resources provided by the campus. “We have facilities like the Cornell Nanoscale Science and Technology Facility (CNF), but also I find that a lot of people and offices will support you."
Always Looking for Solutions
As in research, Reveils looks for solutions in other aspects of his life, for example the challenges associated with being a minority student, and he says, “It’s great to have resources available to help.” As vice president of the Black Graduate and Professional Student Association, he provides community service to his fellow researchers. He also serves on the Diversity and Inclusion Committee of the Graduate and Professional Student Assembly, addressing issues faced by the highly diverse graduate and professional community.
Regardless of the challenges, Reveil persists in his work. “I have always been curious,” he says. He calls Haiti home, but in exploring the depths of knowledge, he finds himself at Cornell and at the forefront of research. If in his work he can appreciate the complexity behind progress, it is because he has personified that journey into the unknown.