Power when parked: EVs could help save money, reduce emissions by providing energy to homes

By relying on their vehicle's batteries for more than just transportation, EV drivers could save thousands on their energy bills and cut carbon emissions

University of Michigan

FacebookXLinkedInWeChatBlueskyMessageWhatsAppEmail

 

image: 

These maps show the median changes in life-cycle greenhouse gas emissions across the contiguous U.S. for electric vehicle charging scenarios explored by researchers from the University of Michigan and Ford Motor Company. In the maps shown here, the drops can be seen for (a) "smart charging" EVs when the power grid is cleanest, (b) incorporating vehicle-to-home, or V2H, charging that allows an EV's battery help power households and (c) using V2H in fully electrified homes (denoted by heat pump).

view more 

Credit: Jiahui Chen with data from J. Chen et al. Nature Energy. 2025 (DOI: 10.1038/s41560-025-01894-7). Made with Plotly.

Using electric vehicles batteries to power households could save their owners thousands of dollars in bills while cutting emissions from the power grid, according to new research from the University of Michigan and Ford Motor Company.

Published in the journal Nature Energy, the team investigated scenarios related to vehicle-to-home charging, or V2H. This emerging technology lets EV drivers tap into energy from their vehicles' batteries to help manage power to their homes. It's almost like using EVs that are parked in garages as generators, but instead of burning gasoline, they provide electricity from their batteries that have been charged by the grid.

"Putting vehicle batteries between the electricity grid and homes makes it possible for homes to buy electricity for all household uses when it is cheap and clean—for example, in the afternoon, when there is a lot of solar power—and to store it in the car’s battery for later use," said Parth Vaishnav, an assistant professor in the University of Michigan's School for Environment and Sustainability, or SEAS.

"If you're buying an EV because you want to cut greenhouse gas emissions—or if you're making an EV because you want to cut greenhouse gas emissions—this tells you that, in addition to reducing greenhouse gas emissions from transport, the EV could also help cut building sector greenhouse gas emissions." 

According to the study, supported by the Ford-University of Michigan Alliance Program, V2H could save EV owners 40 to 90% of their charging costs over the lifetime of the vehicle. That translates to between $2,400 to $5,600 in vehicle lifetime savings.

Furthermore, V2H could reduce lifecycle greenhouse gas emissions from a household’s electricity use by 70 to 250%, which would amount to cutting between 24 and 57 tons of lifetime carbon dioxide emissions. That'd be equivalent to driving a small gas-powered SUV for 80,000 to 190,000 miles, or 80 to 190 one-way flights between New York and Los Angeles. The reduction can surpass 100% when it more than makes up for emissions from the extra electricity needed to drive the car, Vaishnav said.

V2H in the U.S.A.

Vaishnav stressed that this idea is not new, but the discussion around V2H has been largely about its possibility and benefits in principle. What he and his colleagues have done is provided a more thorough and comprehensive outlook for the benefits of V2H in practice across the country.

The team evaluated the impact of V2H using a representative mid-sized SUV considering a variety of factors that vary by location. That included grid energy cost and emissions, housing stock and even the temperature outside, which affects energy efficiency. The team broke the contiguous U.S. into 432 regions defined by shared climates and grid conditions to map out the different impacts.

"We have a lot of geography-based insight," said Jiahui Chen, lead author of the study and a doctoral student in SEAS. For instance, not all regions saw the same benefit. 

But the research showed that V2H enabled greenhouse gas reductions that more than fully offset emissions from charging in regions that account for 60% of the U.S. population. In parts of Texas and California, the cost savings of V2H compared to conventional charging can be so great that it more than pays for the electricity needed for driving.

"When people think of EV charging, it's usually thought of as a burden, a cost that is added to your electric bill," Chen said. "But, with this kind of technology integration, we can make charging an asset." 

A work in progress

While the study's take-home message is that V2H has serious economic and environmental upside, the team also stressed that there are important caveats to consider. One way of looking at the study is that it provides decision-makers with an estimate of whether equipping homes for V2H is worth it, Vaishnav said.

“Another important factor is that the technology to control charging and maximize V2H isn't fully plug-and-play in the U.S. yet, but it is actively being demonstrated with local utilities in various U.S. markets” said Hyung Chul Kim, a research scientist at Ford and a coauthor of the new study.

"This capability is promising but still in its early stages. We're working with utilities to identify the best use cases for them, and we’re also determining ways to optimize overall battery lifetime."

Solutions have been developed and are being tested to deliver on that optimization with utilities and customers. 

"Ultimately, the goal is that drivers won’t have to change anything—they would park and plug in their EVs as normal, then technology running in the background automatically finds the best charging and discharging times,” Kim said.

While that infrastructure begins to scale, the team hopes its collaboration can also lead to a more immediate shift in the way people think about energy and their vehicles.

"We know that vehicles are parked the vast majority of the time and, so as this infrastructure develops, there's a great opportunity here," said Robb De Kleine, a life cycle research analyst with Ford and a coauthor of the new study. 

"As we try to decarbonize the grid, we need energy storage to be able to do that. A lot of the time, the first instinct is to build stationary storage. But EVs could serve as electricity storage devices," De Kleine said. "They just happen to have wheels on them."

The research team also included James Anderson, a technical leader of sustainability and environmental science at Ford, and Greg Keoleian, a professor with SEAS. The team also published a corresponding policy brief about the work.

---

Journal

Nature Energy

DOI

10.1038/s41560-025-01894-7 

Article Title

Vehicle-to-home charging can cut costs and greenhouse gas emissions across the USA

Article Publication Date

12-Dec-2025

 

Empress cicada wings help illuminate molecular structure

Coating nanostructures on cicada wings in silver can amplify molecule detection signals

American Institute of Physics

FacebookXLinkedInWeChatBlueskyMessageWhatsAppEmail

 

image: 

The nanostructures of an empress cicada wing before and after being coated with silver nanoparticles. 

view more 

Credit: Hong et al.

WASHINGTON, Dec. 16, 2025 — Zoom in far enough on an empress cicada wing, and a strange landscape materializes. At the nanoscale, densely packed spires rise from the surface, covering the wing in an endless grove of bowling pins. 

These spires, though, are more than just an eerie sight. The highly ordered, evenly spaced spikes can be modified to act as an optical metamaterial, using their tiny geometry to modify interactions between light waves and matter. 

In AIP Advances, by AIP Publishing, researchers from China Medical University and National Taiwan University showed that these natural nanostructures can be tuned to amplify signals in molecular detection techniques.

The scientists were interested in using the cicada wing structure to improve surface-enhanced Raman spectroscopy (SERS). Standard Raman spectroscopy illuminates a molecule with a laser beam and collects the unique spectrum of scattered light, which acts as a molecular fingerprint. However, the scattered light signal is often weak, prompting researchers to find ways to enhance it. 

SERS takes advantage of enhanced electric fields produced at gaps between metallic nanostructures to amplify Raman scattering signals. While conventional methods for fabricating these nanostructures can carefully calibrate feature size and geometry, they are costly and time-consuming. As such, the low-cost, uniform, and scalable Empress cicada wings appealed to the researchers as a ready-made nanostructure template.

“We wanted to demonstrate that by combining biology’s intrinsic nanoscale design with standard thin film techniques, it is possible to achieve SERS performance comparable to artificially fabricated structures, bridging the gap between nature-inspired design and practical sensing technology,” said author Chung-Hung Hong. 

The team coated cicada wing spires in silver nanoparticles using two methods: sputtering deposition and e-gun evaporation. Sputtering turned the nano-spires into cylindrical pillar-like structures. In contrast, e-gun coating made spires more conical. 

By analyzing the effect of different coating thicknesses on the performance of the SERS device, the team determined that the cylindrical nanostructures separated by five-nanometer gaps was most promising. 

The microscopic gap provided an optimal cranny for strong and consistent electromagnetic hotspots to form in, enhancing SERS performance by a factor of a million compared to non-coated cicada wings. 

Looking forward, the team hopes to extend this nature-inspired design to expand beyond the visible, infrared, and ultraviolet lights used in SERS to microwave and millimeter-wave resonator sensors that could detect biomolecules and environmental contaminants.

“In the future, this bio-templating approach could be extended to other natural micro–nanostructures, such as butterfly wings or plant leaves, and integrated with portable sensors for rapid detection of pathogens and pollutants,” said Hong. "We hope this research demonstrates how biological nanostructures can guide engineering design, opening a new path toward sustainable, low-cost, and highly sensitive sensing technologies.”

### 

The article "SERS Enhancement via Biotemplated Ag Nanostructures on Empress Cicada Wings: Effects of Sputtering and E-gun Deposition on Gap Geometry" is authored by Chung-Hung Hong, Cheng-Wei Kuo, and Hui-Hsin Hsiao. It will appear in AIP Advances on Dec. 16, 2025 (DOI: 10.1063/5.0291146). After that date, it can be accessed at https://doi.org/10.1063/5.0291146.

ABOUT THE JOURNAL 

AIP Advances is an open access journal publishing in all areas of physical sciences—applied, theoretical, and experimental. The inclusive scope of AIP Advances makes it an essential outlet for scientists across the physical sciences.  See https://pubs.aip.org/aip/adv. 

---

Journal

AIP Advances

DOI

10.1063/5.0291146 

Using sound waves to detect helium

An acoustic device created with traditional Japanese bamboo weaving lattice measures helium concentrations using frequency shifts

American Institute of Physics

FacebookXLinkedInWeChatBlueskyMessageWhatsAppEmail

 

image: 

Right: Helium detection device inspired by Kagome-biki. Left: The device's triangular structure helps determine the location of helium leaks in a 2D space.

view more 

Credit: Credit: Wang et al.

WASHINGTON, Dec. 16, 2025 — Helium leaks are hard to detect. Helium is odorless, colorless, tasteless, and does not react with other chemical substances.  Not only can we not see or smell it, but traditional gas sensors have trouble detecting the element because they rely on chemical reactions. Despite this, identifying a helium leak is still crucial, because excess helium can displace oxygen in a confined space, leaving less oxygen for people to breathe. 

In Applied Physics Letters, by AIP Publishing, researchers from Nanjing University developed a device that utilizes sound waves to detect helium.

The researchers built a device inspired by a traditional Japanese bamboo weaving technique called “Kagome-biki.” The resulting triangular Kagome structure consists of nine cylinders arranged in three sub-triangles that share their apexes. Microphones record the sound signal in the corner cylinders, and small tubes between each cylinder allow air to fill the device.

Speakers placed under the corner cylinders generate sound waves, which are localized at the structure’s corner cylinders. Sound waves are vibrations that carry energy through a medium, such as air or water. The shape of a sound wave determines its pitch, loudness, and speed — also called its frequency — amplitude, and sound velocity. In their helium-detection device, the researchers took advantage of how sound velocity changes in different media.

Sound waves travel faster in denser media — they are fastest in a solid, slower through air, and cannot transmit in a vacuum. All objects have a resonant frequency, which is the natural speed at which they vibrate, and adding energy to something at that resonant frequency drastically increases its amplitude.

When helium fills the device, the density of the gas in the device changes. The sound waves traveling through the device suddenly change speed and no longer vibrate the cylinders at their special resonant frequency. This causes a drastic change in the amplitude that the microphones record, and this shift in frequency tells the researchers the concentration of helium in the room.

“Because the relative sensitivity of our sensor remains constant and is not related to working conditions, such as temperature and humidity, the sensor can be applied at an extremely low temperature, which remains challenging for traditional gas sensors working with sensitive materials,” says author Li Fan.

The triangular device also allows the researchers to determine the location of helium leaks in a 2D space by measuring which corner experiences a frequency shift first.

The team hopes to expand the device to locate leakage points in 3D space and develop the system into a portable device.

###

The article, “A sensor for helium leakage detection and orientation based on a two-dimensional acoustic topological material,” is authored by Zhao-yi Wang, Zhan-tao Zhou, Li Fan, Xiao-dong Xu, Li-ping Cheng, and Shu-yi Zhang. It will appear in Applied Physics Letters on Dec. 16, 2025 (DOI: 10.1063/5.0288849). After that date, it can be accessed at https://doi.org/10.1063/5.0288849.

ABOUT THE JOURNAL 

Applied Physics Letters features rapid reports on significant discoveries in applied physics. The journal covers new experimental and theoretical research on applications of physics phenomena related to all branches of science, engineering, and modern technology. See https://pubs.aip.org/aip/apl.

---

Journal

Applied Physics Letters

DOI

10.1063/5.0288849 

Researchers discover bias in AI models that analyze pathology samples

The team created a new tool that reduces bias and improves cancer diagnosis across populations

Harvard Medical School

FacebookXLinkedInWeChatBlueskyMessageWhatsAppEmail

 

image: 

Pathology images of human tissue samples.

view more 

Credit: The Cancer Genome Atlas

At a glance:

• A new study reveals that pathology AI models for cancer diagnosis perform unequally across demographic groups.
• The researchers identified three explanations for the bias and developed a tool that reduced it.
• The findings highlight the need to systematically check for bias in pathology AI to ensure equitable care for patients.

Pathology has long been the cornerstone of cancer diagnosis and treatment. A pathologist carefully examines an ultrathin slice of human tissue under a microscope for clues that indicate the presence, type, and stage of cancer.

To a human expert, looking at a swirly pink tissue sample studded with purple cells is akin to grading an exam without a name on it — the slide reveals essential information about the disease without providing other details about the patient.

Yet the same isn’t necessarily true of pathology artificial intelligence models that have emerged in recent years. A new study led by a team at Harvard Medical School shows that these models can somehow infer demographic information from pathology slides, leading to bias in cancer diagnosis among different populations.

Analyzing several major pathology AI models designed to diagnose cancer, the researchers found unequal performance in detecting and differentiating cancers across populations based on patients’ self-reported gender, race, and age. They identified several possible explanations for this demographic bias.

The team then developed a framework called FAIR-Path that helped reduce bias in the models.

“Reading demographics from a pathology slide is thought of as a ‘mission impossible’ for a human pathologist, so the bias in pathology AI was a surprise to us,” said senior author Kun-Hsing Yu, associate professor of biomedical informatics in the Blavatnik Institute at HMS and HMS assistant professor of pathology at Brigham and Women’s Hospital.

Identifying and counteracting AI bias in medicine is critical because it can affect diagnostic accuracy, as well as patient outcomes, Yu said. FAIR-Path’s success indicates that researchers can improve the fairness of AI models for cancer pathology, and perhaps other AI models in medicine, with minimal effort.

The work, which was supported in part by federal funding, is described Dec. 16 in Cell Reports Medicine.

Testing for bias

Yu and his team investigated bias in four standard AI pathology models being developed for cancer evaluation. These deep-learning models were trained on sets of annotated pathology slides, from which they “learned” biological patterns that enable them to analyze new slides and offer diagnoses.

The researchers fed the AI models a large, multi-institutional repository of pathology slides spanning 20 cancer types.

They discovered that all four models had biased performances, providing less accurate diagnoses for patients in specific groups based on self-reported race, gender, and age. For example, the models struggled to differentiate lung cancer subtypes in African American and male patients, and breast cancer subtypes in younger patients. The models also had trouble detecting breast, renal, thyroid, and stomach cancer in certain demographic groups. These performance disparities occurred in around 29 percent of the diagnostic tasks that the models conducted.

This diagnostic inaccuracy, Yu said, happens because these models extract demographic information from the slides — and rely on demographic-specific patterns to make a diagnosis.

The results were unexpected “because we would expect pathology evaluation to be objective,” Yu added. “When evaluating images, we don’t necessarily need to know a patient’s demographics to make a diagnosis.”

The team wondered: Why didn’t pathology AI show the same objectivity?

Searching for explanations

The researchers landed on three explanations.

Because it is easier to get samples for patients in certain demographic groups, the AI models are trained on unequal sample sizes. As a result, the models have a harder time making an accurate diagnosis in samples that aren’t well-represented in the training set, such as those from minority groups based on race, age, or gender.

Yet “the problem turned out to be much deeper than that,” Yu said. The researchers noticed that sometimes the models performed worse in one demographic group, even when the sample sizes were comparable.

Additional analyses revealed that this may be because of differential disease incidence: Some cancers are more common in certain groups, so the models become better at making a diagnosis in those groups. As a result, the models may have difficulty diagnosing cancers in populations where they aren’t as common.

The AI models also pick up on subtle molecular differences in samples from different demographic groups. For example, the models may detect mutations in cancer driver genes and use them as a proxy for cancer type — and thus be less effective at making a diagnosis in populations in which these mutations are less common.

“We found that because AI is so powerful, it can differentiate many obscure biological signals that cannot be detected by standard human evaluation,” Yu said.

As a result, the models may learn signals that are more related to demographics than disease. That, in turn, could affect their diagnostic ability across groups.

Together, Yu said, these explanations suggest that bias in pathology AI stems not only from the variable quality of the training data but also from how researchers train the models.

Finding a fix

After assessing the scope and sources of the bias, Yu and his team wanted to fix it.

The researchers developed FAIR-Path, a simple framework based on an existing machine-learning concept called contrastive learning. Contrastive learning involves adding an element to AI training that teaches the model to emphasize the differences between essential categories — in this case, cancer types — and to downplay the differences between less crucial categories — here, demographic groups.

When the researchers applied the FAIR-Path framework to the models they’d tested, it reduced the diagnostic disparities by around 88 percent.

“We show that by making this small adjustment, the models can learn robust features that make them more generalizable and fairer across different populations,” Yu said.

The finding is encouraging, he added, because it suggests that bias can be reduced even without training the models on completely fair, representative data.

Next, Yu and his team are collaborating with institutions around the world to investigate the extent of bias in pathology AI in places with different demographics and clinical and pathology practices. They are also exploring ways to extend FAIR-Path to settings with limited sample sizes. Additionally, they would like to investigate how bias in AI contributes to demographic discrepancies in health care and patient outcomes.

Ultimately, Yu said, the goal is to create fair, unbiased pathology AI models that can improve cancer care by helping human pathologists quickly and accurately make a diagnosis.

“I think there’s hope that if we are more aware of and careful about how we design AI systems, we can build models that perform well in every population,” he said.

Authorship, funding, disclosures

Additional authors on the study include Shih-Yen Lin, Pei-Chen Tsai, Fang-Yi Su, Chun-Yen Chen, Fuchen Li, Junhan Zhao, Yuk Yeung Ho, Tsung-Lu Michael Lee, Elizabeth Healey, Po-Jen Lin, Ting-Wan Kao, Dmytro Vremenko, Thomas Roetzer-Pejrimovsky, Lynette Sholl, Deborah Dillon, Nancy U. Lin, David Meredith, Keith L. Ligon, Ying-Chun Lo, Nipon Chaisuriya, David J. Cook, Adelheid Woehrer, Jeffrey Meyerhardt, Shuji Ogino, MacLean P. Nasrallah, Jeffrey A. Golden, Sabina Signoretti, and Jung-Hsien Chiang.

Funding was provided by the National Institute of General Medical Sciences and the National Heart, Lung, and Blood Institute at the National Institutes of Health (grants R35GM142879, R01HL174679), the Department of Defense (Peer Reviewed Cancer Research Program Career Development Award HT9425-231-0523), the American Cancer Society (Research Scholar Grant RSG-24-1253761-01-ESED), a Google Research Scholar Award, a Harvard Medical School Dean’s Innovation Award, the National Science and Technology Council of Taiwan (grants NSTC 113-2917-I-006-009, 112-2634-F-006-003, 113-2321-B-006-023, 114-2917-I-006-016), and a doctoral student scholarship from the Xin Miao Education Foundation.

Ligon was a consultant of Travera, Bristol Myers Squibb, Servier, IntegraGen, L.E.K. Consulting, and Blaze Bioscience; received equity from Travera; and has research funding from Bristol Myers Squibb and Lilly. Vremenko is a cofounder and shareholder of Vectorly.

The authors prepared the initial manuscript and used ChatGPT to edit selected sections to improve readability. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.

---

Journal

Cell Reports Medicine

DOI

10.1016/j.xcrm.2025.102527 

Article Title

Contrastive Learning Enhances Fairness in Pathology Artificial Intelligence Systems

Article Publication Date

16-Dec-2025