Materium Technologies, a startup company with deep NJIT roots, is bringing data science innovations into the slowly evolving field of solar energy panels.
Startups are always a gamble, but the Materium team has a good hand, with two pair of Highlanders — recent alumni Sheldon Fereira (M.S. ‘23) and Scott Daniel (M.S. ‘24), advised by Professor Nuggehalli Ravindra and Adjunct Instructor Michael Jaffe. Their collective scientific expertise spans the worlds of artificial intelligence, applied physics, biomedical engineering, and semiconductors.
Their idea evolved throughout 2023 and the company was incorporated in January 2024. Fereira had been studying applied physics and data science, while researching how the latter could help engineers choose optimal ingredients for new materials that have unique requirements. He then met Ravindra, known as “Dr. Ravi” to his students, and learned that solar panels lose efficiency over time.
Today’s best solar panels are 24% efficient at turning light into energy, but most products are only 21%, Ravindra said. Their efficiency erodes over years or decades, due to the effects of infrared light frequencies. (This also means that solar panels lose less power in colder environments, not where it’s hottest, which is the opposite of what a non-physicist might expect.)
Fereira explained that a precisely engineered coating based on his own data science research, whether it’s a film, paint or spray, could function like the sunblock you wear at the beach in letting through light and stopping radiation. Additional materials that are also derived from data science methods could help as a second phase, acting on whatever heat does get through and drawing it from the panels.
Meanwhile, Daniel had been looking at NJIT’s intellectual property portfolio and was advised that he could work with Fereira and Ravindra. Together, they all understood that there might be a viable business in designing new materials informed by machine learning methods. Daniel has a background in corporate and investment banking, already wanted to become an entrepreneur and had his own mentors in Ying Wu College of Computing professors David Li, Vincent Oria and Keith Williams.
While continuing their technical and market research, the trio attended National Science Foundation Innovation Corps training last January. Next, “We were selected by the U.S. Department of Energy to participate in the Cradle-to-Commerce program, which started up a couple of weeks ago, and will go through the summer,” Daniel said.
That’s the fun of being a startup,” Fereira added. “You can pivot quickly.”
“Basically, just like NJIT wants to take patents out of academia and put them out into the real world, the national laboratories want to take all of their research and development patents and IP that they produce, and get it out to the real world and solve real-world problems. So they are partnering with entrepreneurs to go in, sift through the IP, see what’s complementary to what you’re doing, and then license it from them and take that and produce it, commercialize it. We’ll be spending the summer picking the IP that we think is complementary. We may or may not license it, and then take it over the next 12 months into production.”
Materium’s hand might get even hotter soon. In May, Daniel became the first person at NJIT to graduate with the new M.S. in AI degree and said the company is close to receiving investments that will fund further research along with new employees to help develop prototype layers and coatings. The employees would likely also be Highlanders. In addition, Materium will collaborate with Martin Tuchman School of Management this fall in MGMT-680, Entrepreneurial Strategy, taught by Associate Professor Cesar Bandera.
The company is also considering the application of such films and coatings to computer chips. In that field, performance also degrades due to heat, which is why ordinary desktop computer CPUs have heat sinks and why data centers that run the Internet have massive air conditioning bills.
“Thermal management is one big critical problem the industry is facing, that they need in solar cells or semiconductors. They are struggling because either it’s very expensive, it’s invisible, or there is very limited research,” Fereira said. “So that’s where we started thinking, maybe we can use machine learning to make some processes easier, do validation, or do different types of material testing. And that’s how we can address this thermal management.”
“That’s the fun of being a startup,” Fereira added. “You can pivot quickly.”
This post was originally published on 3rd party site mentioned in the title of this site
