In electronic technologies, key material properties change in response to stimuli like voltage or current. Scientists aim to understand these changes in terms of the material’s structure at the nanoscale (a few atoms) and microscale (the thickness of a piece of paper). Often neglected is the realm between, the mesoscale — spanning 10 billionths to 1 millionth of a meter.

Scientists at the U.S. Department of Energy’s (DOE) Argonne National Laboratory, in collaboration with Rice University and DOE’s Lawrence Berkeley National Laboratory, have made significant strides in understanding the mesoscale properties of a ferroelectric material under an electric field. This breakthrough holds potential for advances in computer memory, lasers for scientific instruments and sensors for ultraprecise measurements.

The ferroelectric material is an oxide containing a complex mixture of lead, magnesium, niobium and titanium. Scientists refer to this material as a relaxor ferroelectric. It is characterized by tiny pairs of positive and negative charges, or dipoles, that group into clusters called “polar nanodomains.” Under an electric field, these dipoles align in the same direction, causing the material to change shape, or strain. Similarly, applying a strain can alter the dipole direction, creating an electric field.

“If you analyze a material at the nanoscale, you only learn about the average atomic structure within an ultrasmall region,” said Yue Cao, an Argonne physicist. “But materials are not necessarily uniform and do not respond in the same way to an electric field in all parts. This is where the mesoscale can paint a more complete picture bridging the nano- to microscale.”

A fully functional device based on a relaxor ferroelectric was produced by professor Lane Martin’s group at Rice University to test the material under operating conditions. Its main component is a thin film (55 nanometers) of the relaxor ferroelectric sandwiched between nanoscale layers that serve as electrodes to apply a voltage and generate an electric field.

Using beamlines in sectors 26-ID and 33-ID of Argonne’s Advanced Photon Source (APS), Argonne team members mapped the mesoscale structures within the relaxor. Key to the success of this experiment was a specialized capability called coherent X-ray nanodiffraction, available through the Hard X-ray Nanoprobe (Beamline 26-ID) operated by the Center for Nanoscale Materials at Argonne and the APS. Both are DOE Office of Science user facilities.

The results showed that, under an electric field, the nanodomains self-assemble into mesoscale structures consisting of dipoles that align in a complex tile-like pattern (see image). The team identified the strain locations along the borders of this pattern and the regions responding more strongly to the electric field.

“These submicroscale structures represent a new form of nanodomain self-assembly not known previously,” noted John Mitchell, an Argonne Distinguished Fellow. “Amazingly, we could trace their origin all the way back down to underlying nanoscale atomic motions; it’s fantastic!”

“Our insights into the mesoscale structures provide a new approach to the design of smaller electromechanical devices that work in ways not thought possible,” Martin said.

“The brighter and more coherent X-ray beams now possible with the recent APS upgrade will allow us to continue to improve our device,” said Hao Zheng, the lead author of the research and a beamline scientist at the APS. “We can then assess whether the device has application for energy-efficient microelectronics, such as neuromorphic computing modeled on the human brain.” Low-power microelectronics are essential for addressing the ever-growing power demands from electronic devices around the world, including cell phones, desktop computers and supercomputers.

This research is reported in Science. In addition to Cao, Martin, Mitchell and Zheng, authors include Tao Zhou, Dina Sheyfer, Jieun Kim, Jiyeob Kim, Travis Frazer, Zhonghou Cai, Martin Holt and Zhan Zhang.

Funding for the research came from the DOE Office of Basic Energy Sciences and National Science Foundation.

“These findings add to growing evidence of increased cancer risk in post-Baby Boomer generations, expanding on previous findings of early-onset colorectal cancer and a few obesity-associated cancers to encompass a broader range of cancer types,” said Dr. Hyuna Sung, lead author of the study and a senior principal scientist of surveillance and health equity science at the American Cancer Society. “Birth cohorts, groups of people classified by their birth year, share unique social, economic, political, and climate environments, which affect their exposure to cancer risk factors during their crucial developmental years. Although we have identified cancer trends associated with birth years, we don’t yet have a clear explanation for why these rates are rising.”

In this analysis, researchers obtained incidence data from 23,654,000 patients diagnosed with 34 types of cancer and mortality data from 7,348,137 deaths for 25 types of cancer for individuals aged 25-84 years for the period Jan 1, 2000, to Dec 31, 2019, from the North American Association of Central Cancer Registries and the U.S. National Center for Health Statistics, respectively. To compare cancer rates across generations, they calculated birth cohort-specific incidence rate ratios and mortality rate ratios, adjusted for age effect and period effect, by birth years, separated by five-year intervals, from 1920 to 1990.

Researchers found that incidence rates increased with each successive birth cohort born since approximately 1920 for eight of 34 cancers. In particular, the incidence rate was approximately two-to-three times higher in the 1990 birth cohort than in the 1955 birth cohort for pancreatic, Kidney, and small intestinal cancers in both male and female individuals; and for liver cancer in female individuals. Additionally, incidence rates increased in younger cohorts, after a decline in older birth cohorts, for nine of the remaining cancers including breast cancer (estrogen-receptor positive only), uterine corpus cancer, colorectal cancer, non-cardia gastric cancer, gallbladder cancer, ovarian cancer, testicular cancer, anal cancer in male individuals, and Kaposi sarcoma in male individuals. Across cancer types, the incidence rate in the 1990 birth cohort ranged from 12% for ovarian cancer to 169% for uterine corpus cancer higher than the rate in the birth cohort with the lowest incidence rate. Notably, mortality rates increased in successively younger birth cohorts alongside incidence rates for liver cancer (female only), uterine corpus, gallbladder, testicular, and colorectal cancers.

“The increase in cancer rates among this younger group of people indicate generational shifts in cancer risk and often serve as an early indicator of future cancer burden in the country. Without effective population-level interventions, and as the elevated risk in younger generations is carried over as individuals age, an overall increase in cancer burden could occur in the future, halting or reversing decades of progress against the disease,” added Dr. Ahmedin Jemal, senior vice president, surveillance and health equity science at the American Cancer Society and senior author of the study. “The data highlights the critical need to identify and address underlying risk factors in Gen X and Millennial populations to inform prevention strategies.”

“The increasing cancer burden among younger generations underscores the importance of ensuring people of all ages have access to affordable, comprehensive health insurance, a key factor in cancer outcomes,” said Lisa Lacasse, president of the American Cancer Society Cancer Action Network (ACS CAN). “To that end, ACS CAN will continue our longstanding work to urge lawmakers to expand Medicaid in states that have yet to do so as well as continue to advocate for making permanent the enhanced Affordable Care Act tax subsidies that have opened the door to access to care for millions.”

The rapid advancement of AI technology, including applications like generative AI, has pushed the scalability of existing digital hardware (CPUs, GPUs, ASICs, etc.) to its limits. Consequently, there is active research into analog hardware specialized for AI computation. Analog hardware adjusts the resistance of semiconductors based on external voltage or current and utilizes a cross-point array structure with vertically crossed memory devices to process AI computation in parallel. Although it offers advantages over digital hardware for specific computational tasks and continuous data processing, meeting the diverse requirements for computational learning and inference remains challenging.

To address the limitations of analog hardware memory devices, the research team focused on Electrochemical Random Access Memory (ECRAM), which manage electrical conductivity through ion movement and concentration. Unlike traditional semiconductor memory, these devices feature a three-terminal structure with separate paths for reading and writing data, allowing for operation at relatively low power.

In their study, the team successfully fabricated ECRAM devices using three-terminal-based semiconductors in a 64×64 array. Experiments revealed that the hardware incorporating the team’s devices demonstrated excellent electrical and switching characteristics, along with high yield and uniformity. Additionally, the team applied the Tiki-Taka algorithm, a cutting-edge analog-based learning algorithm, to this high-yield hardware, successfully maximizing the accuracy of AI neural network training computations. Notably, the researchers demonstrated the impact of the “weight retention” property of hardware training on learning and confirmed that their technique does not overload artificial neural networks, highlighting the potential for commercializing the technology.

This research is significant because the largest array of ECRAM devices for storing and processing analog signals reported in the literature to date is 10×10. The researchers have now successfully implemented these devices on the largest scale, with varied characteristics for each device.

Professor Seyoung Kim of POSTECH remarked, “By realizing large-scale arrays based on novel memory device technologies and developing analog-specific AI algorithms, we have identified the potential for AI computational performance and energy efficiency that far surpass current digital methods.”

The research was conducted with support from the Ministry of Trade, Industry and Energy, the Public-Private Partnership for Semiconductor Talent Training Program supported by the Korea Planning & Evaluation Institute of Industrial Technology (KEIT) and Korea Semiconductor Industry Association, and EDA Tool of the IDEC.