MIT engineers 3D print the electromagnets at the heart of many electronics

MIT engineers 3D print the electromagnets at the heart of many electronics

Imagine being able to build an entire dialysis machine using nothing more than a 3D printer.

This could not only reduce costs and eliminate manufacturing waste, but since this machine could be produced outside a factory, people with limited resources or those who live in remote areas may be able to access this medical device more easily.

While multiple hurdles must be overcome to develop electronic devices that are entirely 3D printed, a team at MIT has taken an important step in this direction by demonstrating fully 3D-printed, three-dimensional solenoids.

Solenoids, electromagnets formed by a coil of wire wrapped around a magnetic core, are a fundamental building block of many electronics, from dialysis machines and respirators to washing machines and dishwashers.

The researchers modified a multimaterial 3D printer so it could print compact, magnetic-cored solenoids in one step. This eliminates defects that might be introduced during post-assembly processes.

This customized printer, which could utilize higher-performing materials than typical commercial printers, enabled the researchers to produce solenoids that could withstand twice as much electric current and generate a magnetic field that was three times larger than other 3D-printed devices.

In addition to making electronics cheaper on Earth, this printing hardware could be particularly useful in space exploration. For example, instead of shipping replacement electronic parts to a base on Mars, which could take years and cost millions of dollars, one could send a signal containing files for the 3D printer, says Luis Fernando Velásquez-García, a principal research scientist in MIT’s Microsystems Technology Laboratories (MTL).

“There is no reason to make capable hardware in only a few centers of manufacturing when the need is global. Instead of trying to ship hardware all over the world, can we empower people in distant places to make it themselves? Additive manufacturing can play a tremendous role in terms of democratizing these technologies,” adds Velásquez-García, the senior author of a new paper on the 3D printed solenoids that appears in the journal Virtual and Physical Prototyping.

He is joined on the paper by lead author Jorge Cañada, an electrical engineering and computer science graduate student; and Hyeonseok Kim, a mechanical engineering graduate student.

Additive advantages

A solenoid generates a magnetic field when an electrical current is passed through it. When someone rings a doorbell, for instance, electric current flows through a solenoid, which generates a magnetic field that moves an iron rod so it strikes a chime.

Integrating solenoids onto electrical circuits manufactured in a clean room poses significant challenges, as they have very different form factors and are made using incompatible processes that require post assembly. Consequently, researchers have investigated making solenoids utilizing many of the same processes that make semiconductor chips. But these techniques limit the size and shape of solenoids, which hampers performance.

With additive manufacturing, one can produce devices that are practically any size and shape. However, this presents its own challenges, since making a solenoid involves coiling thin layers made from multiple materials that may not all be compatible with one machine.

To overcome these challenges, the researchers needed to modify a commercial extrusion 3D printer.

Extrusion printing fabricates objects one layer at a time by squirting material through a nozzle. Typically, a printer uses one type of material feedstock, often spools of filament.

“Some people in the field look down on them because they are simple and don’t have a lot of bells and whistles, but extrusion is one of very few methods that allows you to do multimaterial, monolithic printing,” says Velásquez-García.

This is key, since the solenoids are produced by precisely layering three different materials — a dielectric material that serves as an insulator, a conductive material that forms the electric coil, and a soft magnetic material that makes up the core.

The team selected a printer with four nozzles — one dedicated to each material to prevent cross-contamination. They needed four extruders because they tried two soft magnetic materials, one based on a biodegradable thermoplastic and the other based on nylon.

Printing with pellets

They retrofitted the printer so one nozzle could extrude pellets, rather than filament. The soft magnetic nylon, which is made from a pliable polymer studded with metallic microparticles, is virtually impossible to produce as a filament. Yet this nylon material offers far better performance than filament-based alternatives.

Using the conductive material also posed challenges, since it would start melting and jam the nozzle. The researchers found that adding ventilation to cool the material prevented this. They also built a new spool holder for the conductive filament that was closer to the nozzle, reducing friction that could damage the thin strands.

Even with the team’s modifications, the customized hardware cost about $4,000, so this technique could be employed by others at a lower cost than other approaches, adds Velásquez-García.

The modified hardware prints a U.S. quarter-sized solenoid as a spiral by layering material around the soft magnetic core, with thicker conductive layers separated by thin insulating layers.

Precisely controlling the process is of paramount importance because each material prints at a different temperature. Depositing one on top of another at the wrong time might cause the materials to smear.

Because their machine could print with a more effective soft magnetic material, the solenoids achieved higher performance than other 3D-printed devices.

The printing method enabled them to build a three-dimensional device comprising eight layers, with coils of conductive and insulating material stacked around the core like a spiral staircase. Multiple layers increase the number of coils in the solenoid, which improves the amplification of the magnetic field.

Due to the added precision of the modified printer, they could make solenoids that were about 33 percent smaller than other 3D-printed versions. More coils in a smaller area also boosts amplification.

In the end, their solenoids could produce a magnetic field that was about three times larger than what other 3D-printed devices can achieve.

“We were not the first people to be able to make inductors that are 3D-printed, but we were the first ones to make them three-dimensional, and that greatly amplifies the kinds of values you can generate. And that translates into being able to satisfy a wider range of applications,” he says.

For instance, while these solenoids can’t generate as much magnetic field as those made with traditional fabrication techniques, they could be used as power convertors in small sensors or actuators in soft robots.

Moving forward, the researchers are looking to continue enhancing their performance.

For one, they could try using alternate materials that might have better properties. They are also exploring additional modifications that could more precisely control the temperature at which each material is deposited, reducing defects.

This work is funded by Empiriko Corporation.

Putting AI into the hands of people with problems to solve

Putting AI into the hands of people with problems to solve

As Media Lab students in 2010, Karthik Dinakar SM ’12, PhD ’17 and Birago Jones SM ’12 teamed up for a class project to build a tool that would help content moderation teams at companies like Twitter (now X) and YouTube. The project generated a huge amount of excitement, and the researchers were invited to give a demonstration at a cyberbullying summit at the White House — they just had to get the thing working.

The day before the White House event, Dinakar spent hours trying to put together a working demo that could identify concerning posts on Twitter. Around 11 p.m., he called Jones to say he was giving up.

Then Jones decided to look at the data. It turned out Dinakar’s model was flagging the right types of posts, but the posters were using teenage slang terms and other indirect language that Dinakar didn’t pick up on. The problem wasn’t the model; it was the disconnect between Dinakar and the teens he was trying to help.

“We realized then, right before we got to the White House, that the people building these models should not be folks who are just machine-learning engineers,” Dinakar says. “They should be people who best understand their data.”

The insight led the researchers to develop point-and-click tools that allow nonexperts to build machine-learning models. Those tools became the basis for Pienso, which today is helping people build large language models for detecting misinformation, human trafficking, weapons sales, and more, without writing any code.

“These kinds of applications are important to us because our roots are in cyberbullying and understanding how to use AI for things that really help humanity,” says Jones.

As for the early version of the system shown at the White House, the founders ended up collaborating with students at nearby schools in Cambridge, Massachusetts, to let them train the models.

“The models those kids trained were so much better and nuanced than anything I could’ve ever come up with,” Dinakar says. “Birago and I had this big ‘Aha!’ moment where we realized empowering domain experts — which is different from democratizing AI — was the best path forward.”

A project with purpose

Jones and Dinakar met as graduate students in the Software Agents research group of the MIT Media Lab. Their work on what became Pienso started in Course 6.864 (Natural Language Processing) and continued until they earned their master’s degrees in 2012.

It turned out 2010 wasn’t the last time the founders were invited to the White House to demo their project. The work generated a lot of enthusiasm, but the founders worked on Pienso part time until 2016, when Dinakar finished his PhD at MIT and deep learning began to explode in popularity.

“We’re still connected to many people around campus,” Dinakar says. “The exposure we had at MIT, the melding of human and computer interfaces, widened our understanding. Our philosophy at Pienso couldn’t be possible without the vibrancy of MIT’s campus.”

The founders also credit MIT’s Industrial Liaison Program (ILP) and Startup Accelerator (STEX) for connecting them to early partners.

One early partner was SkyUK. The company’s customer success team used Pienso to build models to understand their customer’s most common problems. Today those models are helping to process half a million customer calls a day, and the founders say they have saved the company over £7 million pounds to date by shortening the length of calls into the company’s call center.

The difference between democratizing AI and empowering people with AI comes down to who understands the data best — you or a doctor or a journalist or someone who works with customers every day?” Jones says. “Those are the people who should be creating the models. That’s how you get insights out of your data.”

In 2020, just as Covid-19 outbreaks began in the U.S., government officials contacted the founders to use their tool to better understand the emerging disease. Pienso helped experts in virology and infectious disease set up machine-learning models to mine thousands of research articles about coronaviruses. Dinakar says they later learned the work helped the government identify and strengthen critical supply chains for drugs, including the popular antiviral remdesivir.

“Those compounds were surfaced by a team that did not know deep learning but was able to use our platform,” Dinakar says.

Building a better AI future

Because Pienso can run on internal servers and cloud infrastructure, the founders say it offers an alternative for businesses being forced to donate their data by using services offered by other AI companies.

“The Pienso interface is a series of web apps stitched together,” Dinakar explains. “You can think of it like an Adobe Photoshop for large language models, but in the web. You can point and import data without writing a line of code. You can refine the data, prepare it for deep learning, analyze it, give it structure if it’s not labeled or annotated, and you can walk away with fine-tuned, large language model in a matter of 25 minutes.”

Earlier this year, Pienso announced a partnership with GraphCore, which provides a faster, more efficient computing platform for machine learning. The founders say the partnership will further lower barriers to leveraging AI by dramatically reducing latency.

“If you’re building an interactive AI platform, users aren’t going to have a cup of coffee every time they click a button,” Dinakar says. “It needs to be fast and responsive.”

The founders believe their solution is enabling a future where more effective AI models are developed for specific use cases by the people who are most familiar with the problems they are trying to solve.

“No one model can do everything,” Dinakar says. “Everyone’s application is different, their needs are different, their data is different. It’s highly unlikely that one model will do everything for you. It’s about bringing a garden of models together and allowing them to collaborate with each other and orchestrating them in a way that makes sense — and the people doing that orchestration should be the people who understand the data best.”

Generative AI for smart grid modeling

Generative AI for smart grid modeling

MIT’s Laboratory for Information and Decision Systems (LIDS) has been awarded $1,365,000 in funding from the Appalachian Regional Commission (ARC) to support its involvement with an innovative project, “Forming the Smart Grid Deployment Consortium (SGDC) and Expanding the HILLTOP+ Platform.”

The grant was made available through ARC’s Appalachian Regional Initiative for Stronger Economies, which fosters regional economic transformation through multi-state collaboration.

Led by Kalyan Veeramachaneni, principal research scientist and principal investigator at LIDS’ Data to AI Group, the project will focus on creating AI-driven generative models for customer load data. Veeramachaneni and colleagues will work alongside a team of universities and organizations led by Tennessee Tech University, including collaborators across Ohio, Pennsylvania, West Virginia, and Tennessee, to develop and deploy smart grid modeling services through the SGDC project.

These generative models have far-reaching applications, including grid modeling and training algorithms for energy tech startups. When the models are trained on existing data, they create additional, realistic data that can augment limited datasets or stand in for sensitive ones. Stakeholders can then use these models to understand and plan for specific what-if scenarios far beyond what could be achieved with existing data alone. For example, generated data can predict the potential load on the grid if an additional 1,000 households were to adopt solar technologies, how that load might change throughout the day, and similar contingencies vital to future planning.

The generative AI models developed by Veeramachaneni and his team will provide inputs to modeling services based on the HILLTOP+ microgrid simulation platform, originally prototyped by MIT Lincoln Laboratory. HILLTOP+ will be used to model and test new smart grid technologies in a virtual “safe space,” providing rural electric utilities with increased confidence in deploying smart grid technologies, including utility-scale battery storage. Energy tech startups will also benefit from HILLTOP+ grid modeling services, enabling them to develop and virtually test their smart grid hardware and software products for scalability and interoperability.

The project aims to assist rural electric utilities and energy tech startups in mitigating the risks associated with deploying these new technologies. “This project is a powerful example of how generative AI can transform a sector — in this case, the energy sector,” says Veeramachaneni. “In order to be useful, generative AI technologies and their development have to be closely integrated with domain expertise. I am thrilled to be collaborating with experts in grid modeling, and working alongside them to integrate the latest and greatest from my research group and push the boundaries of these technologies.”

“This project is testament to the power of collaboration and innovation, and we look forward to working with our collaborators to drive positive change in the energy sector,” says Satish Mahajan, principal investigator for the project at Tennessee Tech and a professor of electrical and computer engineering. Tennessee Tech’s Center for Rural Innovation director, Michael Aikens, adds, “Together, we are taking significant steps towards a more sustainable and resilient future for the Appalachian region.”

Cybersecurity software wins a 2024 Federal Laboratory Consortium Excellence in Technology Transfer Award

Cybersecurity software wins a 2024 Federal Laboratory Consortium Excellence in Technology Transfer Award

The Federal Laboratory Consortium (FLC) has selected MIT Lincoln Laboratory’s Timely Address Space Randomization (TASR) as one of the recipients of their 2024 Excellence in Technology Transfer Award. This cybersecurity technology was transferred in 2019 and 2021 to two companies that develop cloud-based services.

TASR has the potential to help harden many cloud-based servers and user applications against rampant information-leakage attacks. These attacks have been involved in several recent high-profile breaches in which cyber criminals used sensitive information to commit fraud or identity theft, steal financial assets, or gain unauthorized access to other restricted or mission-critical systems. TASR is the first technology that mitigates the impact of such attacks regardless of the attack mechanism or underlying system vulnerability.

A nationwide network of more than 300 government laboratories, agencies, and research centers, FLC helps facilitate the transfer of technologies out of research labs and into the marketplace to benefit the U.S. economy, society, and national security. On an annual basis, FLC confers awards to commend outstanding technology transfer achievements of employees of FLC member labs and their partners from industry, academia, nonprofits, and state and local governments. The Excellence in Technology Transfer Award recognizes exemplary transfer of federally developed technology.

“We are honored to receive this FLC award recognizing our excellence in such technology transfer — in this case, of a cutting-edge cybersecurity technology for protecting everyday users of cloud infrastructure,” says Lincoln Laboratory Chief Technology Ventures Officer Asha Rajagopal.

The Lincoln Laboratory team behind TASR initially developed the technology under sponsorship by the National Security Agency (NSA), following a survey of existing cyber defenses and their vulnerabilities. The three-year development of TASR led to a research prototype in 2015 and a U.S. patent in 2019. In 2020, the U.S. Department of Homeland Security (DHS) selected TASR for its Commercialization Accelerator Program, through which the team matured the technology and connected with commercial companies. Given the growing need for hardening cloud-based services, TASR offers an attractive solution, as it protects Linux-based applications and servers from cyberattacks. Originally developed for personal computers based on Intel’s x86 architecture, the Linux operating system now runs more than 80 percent of all internet servers, 90 percent of public cloud workloads, all 500 of the world’s fastest supercomputers, and the majority of smartphones using Android.

TASR works by automatically and transparently shuffling (rerandomizing) the location of code in memory every time an application processes an input-and-output pair. Information may leak to an attacker whenever the application sends an output, such as a file write or data packet transmitted over a network. But with TASR, the information that may be leaked during system output will have changed at the next point the attacker is able to act on such information (i.e., at system input). Through this moving-target approach, TASR addresses a significant problem contributing to information-leakage attacks: target homogeneity. Once attackers devise an attack against an application, they can easily compromise millions of computers at once because all installations of that application look alike internally. By continuously rerandomizing memory throughout the application’s execution, TASR prevents such action.

“From the first day we started working on TASR, our focus was on making the technology as practical as possible to facilitate its transition to real users. We are honored to be recognized by the FLC for the decade-long journey leading to the transfer of TASR,” says principal investigator Hamed Okhravi, senior staff in the laboratory’s Secure Resilient Systems and Technology Group. Okhravi led the nearly decade-long process of conception, NSA and DHS sponsorship, development, maturation, and transfer phases for TASR, with support from the laboratory’s Technology Ventures Office and MIT’s Technology Licensing Office. The other team members are David Bigelow, Jason Martin, and William Streilein, and former staff members Thomas Hobson and Robert Rudd. TASR was previously recognized with a 2022 R&D 100 Award, acknowledged as one of the year’s 100 most innovative technologies available for sale or license.

The TASR team and awardees in the other categories will be honored at an award ceremony on April 10 during the 2024 FLC National Meeting in Dallas, Texas.

New AI model could streamline operations in a robotic warehouse

New AI model could streamline operations in a robotic warehouse

Hundreds of robots zip back and forth across the floor of a colossal robotic warehouse, grabbing items and delivering them to human workers for packing and shipping. Such warehouses are increasingly becoming part of the supply chain in many industries, from e-commerce to automotive production.

However, getting 800 robots to and from their destinations efficiently while keeping them from crashing into each other is no easy task. It is such a complex problem that even the best path-finding algorithms struggle to keep up with the breakneck pace of e-commerce or manufacturing. 

In a sense, these robots are like cars trying to navigate a crowded city center. So, a group of MIT researchers who use AI to mitigate traffic congestion applied ideas from that domain to tackle this problem.

They built a deep-learning model that encodes important information about the warehouse, including the robots, planned paths, tasks, and obstacles, and uses it to predict the best areas of the warehouse to decongest to improve overall efficiency.

Their technique divides the warehouse robots into groups, so these smaller groups of robots can be decongested faster with traditional algorithms used to coordinate robots. In the end, their method decongests the robots nearly four times faster than a strong random search method.

In addition to streamlining warehouse operations, this deep learning approach could be used in other complex planning tasks, like computer chip design or pipe routing in large buildings.

“We devised a new neural network architecture that is actually suitable for real-time operations at the scale and complexity of these warehouses. It can encode hundreds of robots in terms of their trajectories, origins, destinations, and relationships with other robots, and it can do this in an efficient manner that reuses computation across groups of robots,” says Cathy Wu, the Gilbert W. Winslow Career Development Assistant Professor in Civil and Environmental Engineering (CEE), and a member of a member of the Laboratory for Information and Decision Systems (LIDS) and the Institute for Data, Systems, and Society (IDSS).

Wu, senior author of a paper on this technique, is joined by lead author Zhongxia Yan, a graduate student in electrical engineering and computer science. The work will be presented at the International Conference on Learning Representations.

Robotic Tetris

From a bird’s eye view, the floor of a robotic e-commerce warehouse looks a bit like a fast-paced game of “Tetris.”

When a customer order comes in, a robot travels to an area of the warehouse, grabs the shelf that holds the requested item, and delivers it to a human operator who picks and packs the item. Hundreds of robots do this simultaneously, and if two robots’ paths conflict as they cross the massive warehouse, they might crash.

Traditional search-based algorithms avoid potential crashes by keeping one robot on its course and replanning a trajectory for the other. But with so many robots and potential collisions, the problem quickly grows exponentially.

“Because the warehouse is operating online, the robots are replanned about every 100 milliseconds. That means that every second, a robot is replanned 10 times. So, these operations need to be very fast,” Wu says.

Because time is so critical during replanning, the MIT researchers use machine learning to focus the replanning on the most actionable areas of congestion — where there exists the most potential to reduce the total travel time of robots.

Wu and Yan built a neural network architecture that considers smaller groups of robots at the same time. For instance, in a warehouse with 800 robots, the network might cut the warehouse floor into smaller groups that contain 40 robots each.

Then, it predicts which group has the most potential to improve the overall solution if a search-based solver were used to coordinate trajectories of robots in that group.

An iterative process, the overall algorithm picks the most promising robot group with the neural network, decongests the group with the search-based solver, then picks the next most promising group with the neural network, and so on.

Considering relationships

The neural network can reason about groups of robots efficiently because it captures complicated relationships that exist between individual robots. For example, even though one robot may be far away from another initially, their paths could still cross during their trips.

The technique also streamlines computation by encoding constraints only once, rather than repeating the process for each subproblem. For instance, in a warehouse with 800 robots, decongesting a group of 40 robots requires holding the other 760 robots as constraints. Other approaches require reasoning about all 800 robots once per group in each iteration.

Instead, the researchers’ approach only requires reasoning about the 800 robots once across all groups in each iteration.

“The warehouse is one big setting, so a lot of these robot groups will have some shared aspects of the larger problem. We designed our architecture to make use of this common information,” she adds.

They tested their technique in several simulated environments, including some set up like warehouses, some with random obstacles, and even maze-like settings that emulate building interiors.

By identifying more effective groups to decongest, their learning-based approach decongests the warehouse up to four times faster than strong, non-learning-based approaches. Even when they factored in the additional computational overhead of running the neural network, their approach still solved the problem 3.5 times faster.

In the future, the researchers want to derive simple, rule-based insights from their neural model, since the decisions of the neural network can be opaque and difficult to interpret. Simpler, rule-based methods could also be easier to implement and maintain in actual robotic warehouse settings.

“This approach is based on a novel architecture where convolution and attention mechanisms interact effectively and efficiently. Impressively, this leads to being able to take into account the spatiotemporal component of the constructed paths without the need of problem-specific feature engineering. The results are outstanding: Not only is it possible to improve on state-of-the-art large neighborhood search methods in terms of quality of the solution and speed, but the model generalizes to unseen cases wonderfully,” says Andrea Lodi, the Andrew H. and Ann R. Tisch Professor at Cornell Tech, and who was not involved with this research.

This work was supported by Amazon and the MIT Amazon Science Hub.

Sadhana Lolla named 2024 Gates Cambridge Scholar

Sadhana Lolla named 2024 Gates Cambridge Scholar

MIT senior Sadhana Lolla has won the prestigious Gates Cambridge Scholarship, which offers students an opportunity to pursue graduate study in the field of their choice at Cambridge University in the U.K.

Established in 2000, the Gates Cambridge Scholarship offers full-cost post-graduate scholarships to outstanding applicants from countries outside of the U.K. The mission of the scholarship is to build a global network of future leaders committed to improving the lives of others.

Lolla, a senior from Clarksburg, Maryland, is majoring in computer science and minoring in mathematics and literature. At Cambridge, she will pursue an MPhil in technology policy.

In the future, Lolla aims to lead conversations on deploying and developing technology for marginalized communities, such as the rural Indian village that her family calls home, while also conducting research in embodied intelligence.

At MIT, Lolla conducts research on safe and trustworthy robotics and deep learning at the Distributed Robotics Laboratory with Professor Daniela Rus. Her research has spanned debiasing strategies for autonomous vehicles and accelerating robotic design processes. At Microsoft Research and Themis AI, she works on creating uncertainty-aware frameworks for deep learning, which has impacts across computational biology, language modeling, and robotics. She has presented her work at the Neural Information Processing Systems (NeurIPS) conference and the International Conference on Machine Learning (ICML). 

Outside of research, Lolla leads initiatives to make computer science education more accessible globally. She is an instructor for class 6.s191 (MIT Introduction to Deep Learning), one of the largest AI courses in the world, which reaches millions of students annually. She serves as the curriculum lead for Momentum AI, the only U.S. program that teaches AI to underserved students for free, and she has taught hundreds of students in Northern Scotland as part of the MIT Global Teaching Labs program.

Lolla was also the director for xFair, MIT’s largest student-run career fair, and is an executive board member for Next Sing, where she works to make a cappella more accessible for students across musical backgrounds. In her free time, she enjoys singing, solving crossword puzzles, and baking. 

“Between Sadhana’s impressive research in the Distributed Robotics Group, her volunteer teaching with Momentum AI, and her internship and extracurricular experiences, she has developed the skills to be a leader,” says Kim Benard, associate dean of distinguished fellowships in Career Advising and Professional Development. “Her work at Cambridge will allow her the time to think about reducing bias in systems and the ethical implications of her work. I am proud that she will be representing MIT in the Gates Cambridge community.”

Moving past the Iron Age

Moving past the Iron Age

MIT graduate student Sydney Rose Johnson has never seen the steel mills in central India. She’s never toured the American Midwest’s hulking steel plants or the mini mills dotting the Mississippi River. But in the past year, she’s become more familiar with steel production than she ever imagined.

A fourth-year dual degree MBA and PhD candidate in chemical engineering and a graduate research assistant with the MIT Energy Initiative (MITEI) as well as a 2022-23 Shell Energy Fellow, Johnson looks at ways to reduce carbon dioxide (CO2) emissions generated by industrial processes in hard-to-abate industries. Those include steel.

Almost every aspect of infrastructure and transportation — buildings, bridges, cars, trains, mass transit — contains steel. The manufacture of steel hasn’t changed much since the Iron Age, with some steel plants in the United States and India operating almost continually for more than a century, their massive blast furnaces re-lined periodically with carbon and graphite to keep them going.

According to the World Economic Forum, steel demand is projected to increase 30 percent by 2050, spurred in part by population growth and economic development in China, India, Africa, and Southeast Asia.

The steel industry is among the three biggest producers of CO2 worldwide. Every ton of steel produced in 2020 emitted, on average, 1.89 tons of CO2 into the atmosphere — around 8 percent of global CO2 emissions, according to the World Steel Association.

A combination of technical strategies and financial investments, Johnson notes, will be needed to wrestle that 8 percent figure down to something more planet-friendly.

Johnson’s thesis focuses on modeling and analyzing ways to decarbonize steel. Using data mined from academic and industry sources, she builds models to calculate emissions, costs, and energy consumption for plant-level production.

“I optimize steel production pathways using emission goals, industry commitments, and cost,” she says. Based on the projected growth of India’s steel industry, she applies this approach to case studies that predict outcomes for some of the country’s thousand-plus factories, which together have a production capacity of 154 million metric tons of steel. For the United States, she looks at the effect of Inflation Reduction Act (IRA) credits. The 2022 IRA provides incentives that could accelerate the steel industry’s efforts to minimize its carbon emissions.

Johnson compares emissions and costs across different production pathways, asking questions such as: “If we start today, what would a cost-optimal production scenario look like years from now? How would it change if we added in credits? What would have to happen to cut 2005 levels of emissions in half by 2030?”

“My goal is to gain an understanding of how current and emerging decarbonization strategies will be integrated into the industry,” Johnson says.

Grappling with industrial problems

Growing up in Marietta, Georgia, outside Atlanta, the closest she ever came to a plant of any kind was through her father, a chemical engineer working in logistics and procuring steel for an aerospace company, and during high school, when she spent a semester working alongside chemical engineers tweaking the pH of an anti-foaming agent.

At Kennesaw Mountain High School, a STEM magnet program in Cobb County, students devote an entire semester of their senior year to an internship and research project.

Johnson chose to work at Kemira Chemicals, which develops chemical solutions for water-intensive industries with a focus on pulp and paper, water treatment, and energy systems.

“My goal was to understand why a polymer product was falling out of suspension — essentially, why it was less stable,” she recalls. She learned how to formulate a lab-scale version of the product and conduct tests to measure its viscosity and acidity. Comparing the lab-scale and regular product results revealed that acidity was an important factor. “Through conversations with my mentor, I learned this was connected with the holding conditions, which led to the product being oxidized,” she says. With the anti-foaming agent’s problem identified, steps could be taken to fix it.

“I learned how to apply problem-solving. I got to learn more about working in an industrial environment by connecting with the team in quality control as well as with R&D and chemical engineers at the plant site,” Johnson says. “This experience confirmed I wanted to pursue engineering in college.”

As an undergraduate at Stanford University, she learned about the different fields — biotechnology, environmental science, electrochemistry, and energy, among others — open to chemical engineers. “It seemed like a very diverse field and application range,” she says. “I was just so intrigued by the different things I saw people doing and all these different sets of issues.”

Turning up the heat

At MIT, she turned her attention to how certain industries can offset their detrimental effects on climate.

“I’m interested in the impact of technology on global communities, the environment, and policy. Energy applications affect every field. My goal as a chemical engineer is to have a broad perspective on problem-solving and to find solutions that benefit as many people, especially those under-resourced, as possible,” says Johnson, who has served on the MIT Chemical Engineering Graduate Student Advisory Board, the MIT Energy and Climate Club, and is involved with diversity and inclusion initiatives.

The steel industry, Johnson acknowledges, is not what she first imagined when she saw herself working toward mitigating climate change.

“But now, understanding the role the material has in infrastructure development, combined with its heavy use of coal, has illuminated how the sector, along with other hard-to-abate industries, is important in the climate change conversation,” Johnson says.

Despite the advanced age of many steel mills, some are quite energy-efficient, she notes. Yet these operations, which produce heat upwards of 3,000 degrees Fahrenheit, are still emission-intensive.

Steel is made from iron ore, a mixture of iron, oxygen, and other minerals found on virtually every continent, with Brazil and Australia alone exporting millions of metric tons per year. Commonly based on a process dating back to the 19th century, iron is extracted from the ore through smelting — heating the ore with blast furnaces until the metal becomes spongy and its chemical components begin to break down.

A reducing agent is needed to release the oxygen trapped in the ore, transforming it from its raw form to pure iron. That’s where most emissions come from, Johnson notes.

“We want to reduce emissions, and we want to make a cleaner and safer environment for everyone,” she says. “It’s not just the CO2 emissions. It’s also sometimes NOx and SOx [nitrogen oxides and sulfur oxides] and air pollution particulate matter at some of these production facilities that can affect people as well.”

In 2020, the International Energy Agency released a roadmap exploring potential technologies and strategies that would make the iron and steel sector more compatible with the agency’s vision of increased sustainability. Emission reductions can be accomplished with more modern technology, the agency suggests, or by substituting the fuels producing the immense heat needed to process ore. Traditionally, the fuels used for iron reduction have been coal and natural gas. Alternative fuels include clean hydrogen, electricity, and biomass.

Using the MITEI Sustainable Energy System Analysis Modeling Environment (SESAME), Johnson analyzes various decarbonization strategies. She considers options such as switching fuel for furnaces to hydrogen with a little bit of natural gas or adding carbon-capture devices. The models demonstrate how effective these tactics are likely to be. The answers aren’t always encouraging.

“Upstream emissions can determine how effective the strategies are,” Johnson says. Charcoal derived from forestry biomass seemed to be a promising alternative fuel, but her models showed that processing the charcoal for use in the blast furnace limited its effectiveness in negating emissions.

Despite the challenges, “there are definitely ways of moving forward,” Johnson says. “It’s been an intriguing journey in terms of understanding where the industry is at. There’s still a long way to go, but it’s doable.”

Johnson is heartened by the steel industry’s efforts to recycle scrap into new steel products and incorporate more emission-friendly technologies and practices, some of which result in significantly lower CO2 emissions than conventional production.

A major issue is that low-carbon steel can be more than 50 percent more costly than conventionally produced steel. “There are costs associated with making the transition, but in the context of the environmental implications, I think it’s well worth it to adopt these technologies,” she says.

After graduation, Johnson plans to continue to work in the energy field. “I definitely want to use a combination of engineering knowledge and business knowledge to work toward mitigating climate change, potentially in the startup space with clean technology or even in a policy context,” she says. “I’m interested in connecting the private and public sectors to implement measures for improving our environment and benefiting as many people as possible.”

Explained: Carbon credits

Explained: Carbon credits

One of the most contentious issues faced at the 28th Conference of Parties (COP28) on climate change last December was a proposal for a U.N.-sanctioned market for trading carbon credits. Such a mechanism would allow nations and industries making slow progress in reducing their own carbon emissions to pay others to take emissions-reducing measures, such as improving energy efficiency or protecting forests.

Such trading systems have already grown to a multibillion-dollar market despite a lack of clear international regulations to define and monitor the claimed emissions reductions. During weeks of feverish negotiations, some nations, including the U.S., advocated for a somewhat looser approach to regulations in the interests of getting a system in place quickly. Others, including the European Union, advocated much tighter regulation, in light of a history of questionable or even counterproductive projects of this kind in the past. In the end, no agreement was reached on the subject, which will be revisited at a later meeting.

The concept seems simple enough: Offset emissions in one place by preventing or capturing an equal amount of emissions elsewhere. But implementing that idea has turned out to be far more complex and fraught with problems than many expected.

For example, projects that aim to preserve a section of forest — which can remove carbon dioxide from the air and sequester it in the soil — face numerous issues. Will the preservation of one parcel just lead to the clearcutting of an adjacent parcel? Would the preserved land have been left uncut anyway? And what if it ends up being destroyed by wildfire, drought, or insect infestation — all of which are expected to become more likely with climate change?

Similarly, projects that aim to capture carbon dioxide emissions and inject them into the ground are sometimes used to justify increasing the production of petroleum or natural gas, negating the intended climate mitigation of the process.

Several experts at MIT now say that the system could be effective, at least in certain circumstances, but it must be thoroughly evaluated and regulated.

Carbon removal, natural or mechanical

Sergey Paltsev, deputy director of MIT’s Joint Program on the Science and Policy of Global Change, co-led a study and workshop last year that included policymakers, industry representatives, and researchers. They focused on one kind of carbon offsets, those based on natural climate solutions — restoration or preservation of natural systems that not only sequester carbon but also provide other benefits, such as greater biodiversity. “We find a lot of confusion and misperceptions and misinformation, even about how you define the term carbon credit or offset,” he says.

He points out that there has been a lot of criticism of the whole idea of carbon offsets, “and that criticism is well-placed. I think that’s a very healthy conversation, to clarify what makes sense and what doesn’t make sense. What are the real actions versus what is greenwashing?”

He says that government-mandated and managed carbon trading programs in some places, including British Columbia and parts of Europe, have been somewhat effective because they have clear standards in place, whereas unregulated carbon credit systems have often been abused.

Charles Harvey, an MIT professor of civil and environmental engineering, should know, having been actively involved in both sides of the issue over the last two decades. He co-founded a company in 2008 that was the first private U.S. company to attempt to remove carbon dioxide from emissions on a commercial scale, a process called carbon capture and sequestration, or CCS. Such projects have been a major recipient of federal subsidies aimed at combatting climate change, but Harvey now says these are largely a waste of money and in most cases do not achieve their stated objective.

In fact, he says that according to industry sources, as of 2021 more than 90 percent of CCS projects in the U.S. have been used for the production of more fossil fuels — oil and natural gas. Here’s how it works: Natural gas wells often produce methane mixed with carbon dioxide, which must be removed to produce a marketable natural gas. This carbon dioxide is then injected into oil wells to stimulate more production. So, the net effect is the creation of more total greenhouse gas emissions rather than less, explains Harvey, who recently received a grant from the Rockefeller Foundation to explore CCS projects and whether they can be made to contribute to true emissions reductions.

What went wrong with the ambitious startup CCS company Harvey co-founded? “What happened is that the prices of renewables and energy storage are now incredibly cheap,” he says. “It makes no sense to do this, ever, on power plants because honestly, fossil fuel power plants don’t even really make economic sense anymore.”

Where does Harvey see potential for carbon credits to work? One possibility is the preservation or restoration of tropical peatlands, which he has received another grant to study. These are vast areas of permanently waterlogged land in which dead plant matter —and the carbon it contains — remains in place because the water prevents the normal decomposition processes that would otherwise release the stored carbon back into the air.

While it is virtually impossible to quantify the amount of carbon stored in the soil of forest or farmland, in peatlands that’s easy to do because essentially all of the submerged material is carbon-based. Simply measuring changes in the elevation of such land, which can be done remotely by plane or satellite, gives a precise measure of how much carbon has been stored or released. When a patch of peat forest that has been clear-cut to build plantations or roads is reforested, the amount of carbon emissions that were prevented can be measured accurately.

Because of that potential for accurate documentation, protecting or restoring peat bogs can also be a good way to achieve meaningful offsets for carbon emissions elsewhere, Harvey says. Rewetting a previously drained peat forest can immediately counteract the release of its stored carbon and can keep it there as long as it is not drained again — something that can be verified using satellite data.

Paltsev adds that while such nature-based systems for countering carbon emissions can be a key component of addressing climate change, especially in very difficult-to-decarbonize industries such as aviation, carbon credits for such programs “shouldn’t be a replacement for our efforts at emissions reduction. It should be in addition.”

Criteria for meaningful offsets

John Sterman, the Jay W. Forrester Professor of Management at the MIT Sloan School of Management, has published a set of criteria for evaluating proposed carbon offset plans to make sure they would provide the benefits they claim. At present, “there’s no regulation, there’s no oversight” for carbon offsets, he says. “There have been many scandals over this.”

For example, one company was providing what it claimed was certification for carbon offset projects but was found to have such lax standards that the claimed offsets were often not real. For example, there were multiple claims to protect the same piece of forest and claims to protect land that was already legally protected.

Sterman’s proposed set of criteria goes by the acronym AVID+. “It stands for four principles that you have to meet in order for your offset to be legitimate: It has to be additional, verifiable, immediate, and durable,” he says. “And then I call it AVID+,” he adds, the “plus” being for plans that have additional benefits as well, such as improving health, creating jobs, or helping historically disadvantaged communities.

Offsets can be useful, he says, for addressing especially hard-to-abate industries such as steel or cement manufacturing, or aviation. But it is essential to meet all four of the criteria, or else real emissions are not really being offset. For example, planting trees today, while often a good thing to do, would take decades to offset emissions going into the atmosphere now, where they may persist for centuries — so that fails to meet the “immediate” requirement.

And protecting existing forests, while also desirable, is very hard to prove as being additional, because “that requires a counterfactual that you can never observe,” he says. “That’s where a lot of squirrely accounting and a lot of fraud comes in, because how do you know that the forest would have been cut down but for the offset?” In one well-documented case, he points out, a company tried to sell carbon offsets for a section of forest that was already an established nature preserve.

Are there offsets that can meet all the criteria and provide real benefits in helping to address climate change? Yes, Sterman and Harvey say, but they need to be evaluated carefully.

“My favorite example,” Sterman says, “is doing deep energy retrofits and putting solar panels on low-income housing.” These measures can help address the so-called landlord-tenant problem: If tenants typically pay the utility bills, landlords have little incentive to pay for efficiency improvements, and the tenants don’t have the capital to make such improvements on their own. “Policies that would make this possible are pretty good candidates for legitimate offsets, because they are additional — low-income households can’t afford to do it without assistance, so it’s not going to happen without a program. It’s verifiable, because you’ve got the utility bills pre and post.” They are also quite immediate, typically taking only a year or so to implement, and “they’re pretty durable,” he says.

Another example is a recent plan in Alaska that allows cruise ships to offset the emissions caused by their trips by paying into a fund that provides subsidies for Alaskan citizens to install heat pumps in their homes, thus preventing emissions from wood or fossil fuel heating systems. “I think this is a pretty good candidate to meet the criteria, certainly a lot better than much of what’s being done today,” Sterman says.

But eventually, what is really needed, the researchers agree, are real, enforceable standards. After COP28, carbon offsets are still allowed, Sterman says, “but there is still no widely accepted mandatory regulation. We’re still in the wild West.”

Paltsev nevertheless sees reasons for optimism about nature-based carbon offset systems. For example, he says the aviation industry has recently agreed to implement a set of standards for offsetting their emissions, known as CORSIA, for carbon offsetting and reduction scheme for international aviation. “It’s a point for optimism,” he says, “because they issued very tough guidelines as to what projects are eligible and what projects are not.”

He adds, “There is a solution if you want to find a good solution. It is doable, when there is a will and there is the need.”

Study unlocks nanoscale secrets for designing next-generation solar cells

Study unlocks nanoscale secrets for designing next-generation solar cells

Perovskites, a broad class of compounds with a particular kind of crystal structure, have long been seen as a promising alternative or supplement to today’s silicon or cadmium telluride solar panels. They could be far more lightweight and inexpensive, and could be coated onto virtually any substrate, including paper or flexible plastic that could be rolled up for easy transport.

In their efficiency at converting sunlight to electricity, perovskites are becoming comparable to silicon, whose manufacture still requires long, complex, and energy-intensive processes. One big remaining drawback is longevity: They tend to break down in a matter of months to years, while silicon solar panels can last more than two decades. And their efficiency over large module areas still lags behind silicon. Now, a team of researchers at MIT and several other institutions has revealed ways to optimize efficiency and better control degradation, by engineering the nanoscale structure of perovskite devices.

The study reveals new insights on how to make high-efficiency perovskite solar cells, and also provides new directions for engineers working to bring these solar cells to the commercial marketplace. The work is described today in the journal Nature Energy, in a paper by Dane deQuilettes, a recent MIT postdoc who is now co-founder and chief science officer of the MIT spinout Optigon, along with MIT professors Vladimir Bulovic and Moungi Bawendi, and 10 others at MIT and in Washington state, the U.K., and Korea.

“Ten years ago, if you had asked us what would be the ultimate solution to the rapid development of solar technologies, the answer would have been something that works as well as silicon but whose manufacturing is much simpler,” Bulovic says. “And before we knew it, the field of perovskite photovoltaics appeared. They were as efficient as silicon, and they were as easy to paint on as it is to paint on a piece of paper. The result was tremendous excitement in the field.”

Nonetheless, “there are some significant technical challenges of handling and managing this material in ways we’ve never done before,” he says. But the promise is so great that many hundreds of researchers around the world have been working on this technology. The new study looks at a very small but key detail: how to “passivate” the material’s surface, changing its properties in such a way that the perovskite no longer degrades so rapidly or loses efficiency.

“The key is identifying the chemistry of the interfaces, the place where the perovskite meets other materials,” Bulovic says, referring to the places where different materials are stacked next to perovskite in order to facilitate the flow of current through the device.

Engineers have developed methods for passivation, for example by using a solution that creates a thin passivating coating. But they’ve lacked a detailed understanding of how this process works — which is essential to make further progress in finding better coatings. The new study “addressed the ability to passivate those interfaces and elucidate the physics and science behind why this passivation works as well as it does,” Bulovic says.

The team used some of the most powerful instruments available at laboratories around the world to observe the interfaces between the perovskite layer and other materials, and how they develop, in unprecedented detail. This close examination of the passivation coating process and its effects resulted in “the clearest roadmap as of yet of what we can do to fine-tune the energy alignment at the interfaces of perovskites and neighboring materials,” and thus improve their overall performance, Bulovic says.

While the bulk of a perovskite material is in the form of a perfectly ordered crystalline lattice of atoms, this order breaks down at the surface. There may be extra atoms sticking out or vacancies where atoms are missing, and these defects cause losses in the material’s efficiency. That’s where the need for passivation comes in.

“This paper is essentially revealing a guidebook for how to tune surfaces, where a lot of these defects are, to make sure that energy is not lost at surfaces,” deQuilettes says. “It’s a really big discovery for the field,” he says. “This is the first paper that demonstrates how to systematically control and engineer surface fields in perovskites.”

The common passivation method is to bathe the surface in a solution of a salt called hexylammonium bromide, a technique developed at MIT several years ago by Jason Jungwan Yoo PhD ’20, who is a co-author of this paper, that led to multiple new world-record efficiencies. By doing that “you form a very thin layer on top of your defective surface, and that thin layer actually passivates a lot of the defects really well,” deQuilettes says. “And then the bromine, which is part of the salt, actually penetrates into the three-dimensional layer in a controllable way.” That penetration helps to prevent electrons from losing energy to defects at the surface.

These two effects, produced by a single processing step, produces the two beneficial changes simultaneously. “It’s really beautiful because usually you need to do that in two steps,” deQuilettes says.

The passivation reduces the energy loss of electrons at the surface after they have been knocked loose by sunlight. These losses reduce the overall efficiency of the conversion of sunlight to electricity, so reducing the losses boosts the net efficiency of the cells.

That could rapidly lead to improvements in the materials’ efficiency in converting sunlight to electricity, he says. The recent efficiency records for a single perovskite layer, several of them set at MIT, have ranged from about 24 to 26 percent, while the maximum theoretical efficiency that could be reached is about 30 percent, according to deQuilettes.

An increase of a few percent may not sound like much, but in the solar photovoltaic industry such improvements are highly sought after. “In the silicon photovoltaic industry, if you’re gaining half of a percent in efficiency, that’s worth hundreds of millions of dollars on the global market,” he says. A recent shift in silicon cell design, essentially adding a thin passivating layer and changing the doping profile, provides an efficiency gain of about half of a percent. As a result, “the whole industry is shifting and rapidly trying to push to get there.” The overall efficiency of silicon solar cells has only seen very small incremental improvements for the last 30 years, he says.

The record efficiencies for perovskites have mostly been set in controlled laboratory settings with small postage-stamp-size samples of the material. “Translating a record efficiency to commercial scale takes a long time,” deQuilettes says. “Another big hope is that with this understanding, people will be able to better engineer large areas to have these passivating effects.”

There are hundreds of different kinds of passivating salts and many different kinds of perovskites, so the basic understanding of the passivation process provided by this new work could help guide researchers to find even better combinations of materials, the researchers suggest. “There are so many different ways you could engineer the materials,” he says.

“I think we are on the doorstep of the first practical demonstrations of perovskites in the commercial applications,” Bulovic says. “And those first applications will be a far cry from what we’ll be able to do a few years from now.” He adds that perovskites “should not be seen as a displacement of silicon photovoltaics. It should be seen as an augmentation — yet another way to bring about more rapid deployment of solar electricity.”

“A lot of progress has been made in the last two years on finding surface treatments that improve perovskite solar cells,” says Michael McGehee, a professor of chemical engineering at the University of Colorado who was not associated with this research. “A lot of the research has been empirical with the mechanisms behind the improvements not being fully understood. This detailed study shows that treatments can not only passivate defects, but can also create a surface field that repels carriers that should be collected at the other side of the device. This understanding might help further improve the interfaces.”

The team included researchers at the Korea Research Institute of Chemical Technology, Cambridge University, the University of Washington in Seattle, and Sungkyunkwan University in Korea. The work was supported by the Tata Trust, the MIT Institute for Soldier Nanotechnologies, the U.S. Department of Energy, and the U.S. National Science Foundation.