At a time when many Haverford students were looking forward to a relaxing week of naps and Netflix binges, Liana Alves and Aaron Schankler, both members of the Class of 2018, had a very different vision in mind for their fall breaks: conducting laboratory research.

Both chemistry majors, they landed positions at national and government-sponsored labs where they had the opportunity to work with some of the most sophisticated scientific tools available—Alves, for example, was in close proximity to a “super-cool $1,000,000,000 robot” known as HERMAN—and collaborate with some of the best minds in the field.

Though their paths ran parallel in some ways and both were sponsored by the Koshland Integrated Natural Sciences Center (KINSC), Alves and Schankler’s scientific concerns could not have been more different. While Alves flew out to California to synthesize organo-halide perovskites, a material “used for its applications to solar cells, LEDs, and other semiconductor electronics,” in Lawrence Berkeley National Lab’s famous Molecular Foundry, Schankler and his mentor, Associate Professor of Chemistry Joshua Schrier, received funding from the Defense Advanced Research Projects Agency, or DARPA, to attend a “working meeting” (a sort of collaborative research effort) in Washington, D.C., to discuss areas for possible improvement in protein-folding models with “other researchers using similar data-driven approaches” to achieve this same end goal.

When asked why she chose to spend her vacation week hard at work rather than relaxing at home, Alves quips, “Robots, why else?” As a chemist who has come to terms with the limits of human speed and efficiency in hands-on experiments, her enthusiasm for robotic technology is understandable. “High-throughput robots [are] liquid-dispensing robots that are super-fast and optimize and accelerate chemical procedures, as compared to doing a bench-top reaction that might take one hour for one reaction,” she says. “[With the robots,] I can do 96 reactions in one 96 well plate. And do multiple plates a day.”

How do these robots factor into Alves’s experimental end goal? Since perovskites themselves are synthesized from individual crystals, Alves explains, she—with the aid of her mentor at the Molecular Foundry, Emory Chan—is attempting to define the optimal conditions for crystal growth. “For example, [we need to consider factors like] concentration, temperature, inorganic to organic ratio, and...[the] type [and corresponding dimensionality] of… the perovskites.” It’s no easy task, and Alves likens the successful growth of single crystals to black magic.

Even frequent failures, though, have their value, as the large data sets generated from the quantified conditions of any given attempt at crystal growth are incorporated into machine-learning algorithms to produce a series of input and output values that indicate an ideal “formula” of environmental factors conducive to perfect single crystal growth.

Alves’s research may sound inaccessible, but it has many applications to everyday life, as the high crystallinity associated with perfect single crystals is “important for electronic devices.” Phones, computers, and tablets all owe some degree of technical form to the single crystals created by researchers like Alves.

As for Schankler, his research works to identify individual factors in proteins “that are not fully captured by current models.” As Schankler is, by his own admission, primarily interested in mathematical approaches to chemistry, the methodology of this research is right up his alley because it utilizes “a combination of machine learning and quantum chemistry” to analyze large data sets of synthetic peptides that had been tested experimentally for stability. Schankler and his mentor are currently in the process of creating a method “to use machine learning methods to identify differences between stable and unstable structures.”

Schankler’s work also has implications for the quality of human life. “There have been a lot of really cool advances [in this field] lately,” he says, citing as an example “de novo protein design, where scientists are designing artificial proteins that have a specific shape or a specific property. One lab at the University of Washington has designed proteins that bind to certain pathogens, and they’ve shown that if you infect mice with influenza and squirt these proteins up their noses they don’t get sick.” This discovery has a broad range of applications: for example, these small, lab-designed proteins could replace the natural antibodies—large proteins that require biological synthesis, a costly and time-consuming process—currently used as the biological basis for some classes of treatments. “That’s the eventual goal,” says Schankler, “to have more tailored protein-based therapeutics.”

Alves’s trip to the Berkeley Lab was actually a return trip. She first conducted research there last summer with Associate Professors of Chemistry Alexander Norquist and Joshua Schrier, who had suggested she join them in their work with perovskites after she returned from her semester abroad. And now, as a two-time research veteran, she has been offered an “affiliate user” position by the lab, which means she has the opportunity to continue her research at the Molecular Foundry after graduation.

Neither Alves nor Schankler, though, are looking quite so far ahead at the moment: both are immersed in the process of beginning their senior theses. Schankler’s will focus on a comparison of two models of protein folding, while Alves’s is concerned with the applications of high-throughput robots. Both Fords plan to incorporate the data collected over the course of their fall research into their theses.