MADMEC semifinals: Greenhouses, garbage, and garments

MADMEC is sponsored by Saint Gobain and the Dow Chemical Company.
Categories: Students

Energy-efficient climate control for greenhouses. A garbage bin that identifies and classifies waste. Clothes made from biodegradable nanofibers. Those were the projects presented by the three teams in the MADMEC semifinals earlier this month.

With three cash prizes up for grabs, “You’ve already won money,” said Mike Tarkanian, senior lecturer in DMSE and coordinator of MADMEC. “You have five weeks to sort out how much month you’re going to win.”

MADMEC is a competition hosted by DMSE in which students design materials solutions to sustainability problems. The first-place prize is $10,000. The teams will present the final results of their research at the finals, on Oct. 11.

Beginning last spring, teams were given guidance, access to equipment, and $1,000 funding to kick-start their projects. They gathered in the Chipman Room on Sept. 9 to share the progress they’ve made.

This year marks a return for the contest after the pandemic shutdowns of 2020. “We are hoping future contests can rebuild momentum and draw more teams to MADMEC, after a few years of covid postponements,” Tarkanian said.

Here are descriptions of the teams and their projects:


SmartClime is made up of Isabella Caruso, Chris Eschler, and Eric Lee. All three DMSE graduate students like gardening, so their project took on a particularly verdant hue, focusing on greenhouses—specifically, the covering that goes over the frame and plays a key role in heat retention. The traditional covering is glass; today the most common ones are polyethylene or polycarbonate plastics.

“Greenhouses are great because they enable a longer growing season; you can better control the climate inside,” said Caruso. “The problem’s that because of the greenhouse covering materials used, greenhouse growers have to artificially manipulate the temperature inside the greenhouse. They have to either heat it in the winter with furnaces in some cases and then cool it in summer.”

The team’s idea was to create a greenhouse covering that could control the climate inside and get better energy efficiency. Two options are being tested. One is thermochromic—so if the temperature rises, the material will change color, letting in fewer wavelengths of light that cause it to heat up inside. The other is electrochromic, which changes color in response to applied voltage of electricity. 

The team has designed small samples of both materials and seen some success. The next step is building larger samples that will fit on miniature model greenhouses to measure the temperature throughout the day.

“Eventually we would like to try to simulate what, say, a year would look like in one of these cells,” said Eschler, “and how it would change the amount of heating and cooling energy that you would need relative to existing materials.”


WasteAway is DMSE undergraduates Marilyn Meyers, Sam Song, Melissa Stok, and grad student Vineet Nair, from the Department of Mechanical Engineering. The team’s project addresses waste and recycling contamination in waste streams—so general waste items like leftover food going into a recycling bin, or a plastic container in a trash bin. 

Stok pointed to an Environmental Protection Agency statistic: 45% of materials in landfills are recyclable, and 39% is biomass, which could be composted or turned into useful resources. “It’s important to properly segregate our streams so that we can reprocess the maximum amount of material and come up with efficient solutions and better materials to reuse,” Stok said.

The WasteAway project relies on bin attachments with motion-activated cameras that take pictures of what gets thrown out. That information gets uploaded to the cloud, and machine learning is run on it to determine which items should be in a garbage or recycling bin and which shouldn’t.

But to identify what’s what, the team’s machine learning model needs a lot of good data. Most of the data used was from MIT’s Waste Watchers initiative—pictures of things like envelopes and plastic bags in trash and recycling bins—and open source data, which is separated into categories like glass and paper. “Machine learning is almost completely driven by what data you have available,” Song said.

Once the team gets better-quality data and refines its model, it plans to do a rollout of its bin attachments in Building E38, which houses the MIT Office of Sustainability.

But the goal for WasteAway is not building machines that sort trash and recycling. It’s about providing a tool that can be used to test behavior. 

“So does eyeing better signage by these waste bins cause a decrease in contamination? And then we can run that type of experiment,” Meyers said. “We’re hoping teaching people the right way to recycle is more cost-effective than getting the fancy waste bins that sort it for you.”


Yarnz stands for “Yarns are really nanofiberz,” and its members are DMSE grad students Natalie Mamrol and Veera Panova, and Nathan Ewell, a grad student in the Department of Chemical Engineering.

The driving force behind the team’s project is the global impact of the textile industry, which accounts for 5% to 10% of total greenhouse gas emissions—and a projected 49% by 2030. Plus, less than 15% are recycled or reused. 

The biggest source of pollution is polyester. “This material has little to no biodegradation,” said Mamrol. “You can think of all the plastic bottles; it’s made out of the same material. They end up in landfills.”

Enter biodegradable nanofibers—they have greater surface-to-volume ratio, so they biodegrade in less than a year, and they’re less energy-intensive to process than traditional fibers. And they’re strong—they won’t tear like paper napkins.

How to produce durable, biodegradable, environmentally friendly fibers? Electrospinning, which uses electric force to produce polymer fibers less than one micron in diameter, or one millionth of a meter. It starts with a precursor solution, or base compound, so there’s no need to melt plastics, as with traditional fibers. Advantages to electrospinning include scalability—it can be set up easily and cheaply. And it’s versatile: electrospun polymers can have targeted properties such as biodegradability and mechanical strength.

The team started electrospinning polycaprolactone, a biodegradable polyester, using a fiber extrusion needle into a bath of water. Then those fibers, forming a film in the water, can be twisted into a yarn.

“We started out hand winding the yarn, and now we’re working on developing spinnerets to automatically draw the yarn in a continuous fashion,” said Ewell.

The team is working on electrospinning a spool of yarn and then weaving a piece of fabric from it. Further, the process could be used to spin any kind of fiber, including biomass-based materials, to further reduce greenhouse emissions from the textile industry. 

MADMEC is sponsored by Saint Gobain and the Dow Chemical Company.