KIER cracks seawater electrolysis deposit problem with dual electrode system

Dual-electrode architecture enables repeated precipitate formation and removal, allowing complete elimination without external washing or other cleaning processes

National Research Council of Science & Technology

FacebookXLinkedInWeChatBlueskyMessageWhatsAppEmail

 

image: 

The research team’s dual-cathode seawater electrolysis system

view more 

Credit: KOREA INSTITUTE OF ENERGY RESEARCH

A research team led by Dr. Ji-Hyung Han from the Convergence Research Center of Sector Coupling & Integration at the Korea Institute of Energy Research (President Yi, Chang-Keun, hereinafter “KIER”) has developed a new seawater electrolysis system that overcomes the precipitate formation issue long blamed for performance degradation and process interruptions, while also presenting a new direction for further technology advancement.

Water electrolysis is a technology that produces hydrogen, an eco-friendly energy source, by splitting water. Recently, amid the global freshwater shortage, seawater electrolysis using seawater has been gaining attention as a promising alternative.

However, seawater electrolysis has often been considered inefficient because precipitates formed from magnesium and calcium ions in seawater accumulate on electrode surfaces, leading to performance degradation. It has also been pointed out that continuous hydrogen production is difficult because the deposited precipitates must be removed through acid washing or mechanical cleaning.

To address this issue, KIER researchers developed a new system architecture incorporating two electrodes for the first time in the world. While one electrode produces hydrogen and accumulates precipitates, the other, where precipitates have already built up, temporarily halts hydrogen production and dissolves the deposits using seawater that becomes naturally acidified during operation.

Once the precipitates are completely dissolved, the two electrodes switch roles, enabling hydrogen production and precipitate removal to proceed simultaneously. Through experiments, the researchers confirmed that by simply alternating the roles of the electrodes every 48 hours, precipitate formation and complete removal could be repeated continuously.

In conventional single-electrode seawater electrolysis systems, energy consumption increased by about 27% after 200 hours of operation due to precipitate buildup. By contrast, the system developed by the research team showed only a 1.8% increase in energy consumption even after more than 400 hours of long-term operation, delivering 15 times higher performance than the single-electrode system.

In addition, after 400 hours of operation, the hydrogen evolution catalyst content decreased by only 20% from its initial level, demonstrating superior stability compared with the single-electrode system, which showed a 53% reduction.

Dr. Ji-Hyung Han, the principal researcher of the study, said, “This study demonstrates that the precipitate issue, a major bottleneck in seawater electrolysis, can be controlled solely through system architecture design.” She added, “In particular, by being the first in the world to propose the concept of ‘self-cleaning,’ in which electrodes recover on their own using acidified seawater, this work presents a new direction for future seawater electrolysis technology development.”

Meanwhile, this research was carried out as a collaborative study with Professor Joohyun Lim’s team at Kangwon National University, with support from the Convergence Research Group Project of the National Research Council of Science & Technology (NST). The findings were published in the March issue of Chemical Engineering Journal (IF 13.2), a prestigious international journal in the fields of energy and chemical engineering.

---

Journal

Chemical Engineering Journal

DOI

10.1016/j.cej.2026.174360 

Article Title

Self-cleaning dual cathode for enhanced durability of bipolar membrane-based direct seawater electrolysis

 

Researchers show dinos hatched eggs less efficiently than modern birds

Research using dinosaur body model suggests that – unlike modern birds – bird-like dinosaurs may have used the sun’s warmth to help hatch eggs, shedding light on the evolution of avian-style incubation.

Frontiers

FacebookXLinkedInWeChatBlueskyMessageWhatsAppEmail

 

image: 

Lateral view of the clutch. The eggs were molded from casting resin. 

view more 

Credit: Chun-Yu Su.

What do we really know about how oviraptors – bird-like but flightless dinosaurs – hatched their eggs? Did they use environmental heat, like crocodiles, or body heat from an adult, like birds? In a new Frontiers in Ecology and Evolution study, researchers in Taiwan examined the brooding behavior and hatching patterns of oviraptors. They also modelled heat transfer simulations of oviraptor clutches and compared hatching efficiency to modern birds. To do so, they experimented with a life-sized oviraptor incubator and eggs.

“We show the difference in oviraptor hatching patterns was induced by the relative position of the incubating adult to the eggs,” said senior author Dr Tzu-Ruei Yang, an associate curator of vertebrate paleontology at Taiwan’s National Museum of Natural Science.

“Moreover, we obtained an estimate of the incubation efficiency of oviraptors, which is much lower than that of modern birds,” added first author Chun-Yu Su, who attended Washington High School in Taichung when the research was conducted.

Building a dinosaur

The reconstructed oviraptor Heyuannia huangi lived between 70 and 66 million years ago in what today is China. Estimated to be around 1.5 meters long and weighing around 20kg, it built semi-open nests made up of several rings of eggs.

The incubating oviraptor’s trunk was made from polystyrene foam and wood for the skeletal frame and cotton, bubble paper, and cloth for the soft tissue. Eggs were molded from casting resin. In the two clutches used in the experiments, eggs were arranged in double-rings based on real oviraptor clutches.

“Part of the difficulty lies in reconstructing oviraptor incubation realistically,” said Su. “For example, their eggs are unlike those of any living species, so we invented the resin eggs to approximate real oviraptor eggs as best as we could.”

When the team ran experiments to find out if clutch attendance of a brooding adult or different environmental circumstances may have impacted hatching patterns, they found that in colder temperatures, where a brooding adult attended the clutch, the eggs’ temperatures in the outer ring differed by up to 6°C, which could have resulted in asynchronous hatching, a pattern where eggs in the same nest hatch at different times. In warmer conditions, the difference in egg temperatures in the outer ring was just 0.6°C, suggesting that oviraptors living in warmer conditions may have exhibited a different pattern of asynchronous hatching because they could use the sun as an additional, powerful heat source.

“It’s unlikely that large dinosaurs sat atop their clutches. Supposedly they used the heat of the sun or soil to hatch their eggs, like turtles. Since oviraptor clutches are open to the air, heat from the sun likely mattered much more than heat from the soil,” Yang explained.

Better hatchers?

The team also investigated how oviraptor incubation efficiency compares to that of modern birds. Most birds use thermoregulatory contact incubation (TCI), where adults sit directly on the eggs to transfer heat. TCI requires three prerequisites – the adult bird must be in contact with every egg, be the main heat source, and maintain all eggs within a constrained temperature range – which oviraptors didn’t fulfil. For example, their egg arrangement prevented the adult from making full contact with all eggs in the clutch.

“Oviraptors may not have been able to conduct TCI as modern birds do,” said Su. Instead, these dinosaurs and the sun may have been co-incubators – a less efficient incubation behavior than that displayed by modern birds. Yet, the combination of adult incubation and an ambient heat source – perhaps a behavioral adaptation associated with the evolution from buried to semi-open nests – isn’t necessarily worse.

“Modern birds aren’t ‘better’ at hatching eggs. Instead, birds living today and oviraptors have a very different way of incubation or, more specifically, brooding,” Yang pointed out. “Nothing is better or worse. It just depends on the environment.”

The team pointed out that their findings are specific to the reconstructed nest and are limited by the fact that today’s climate does not resemble the Late Cretaceous climate, which may have impacted the results. Oviraptors also exhibited a longer incubation period than modern birds.

Yet, the study advances our understanding of oviraptor brooding strategies through innovative approaches. It represents an important bridge between physics-based simulations and paleontological interpretations, potentially enabling paleontologists to investigate topics for which approaches were limited until now.  

“It also truly is an encouragement for all students, especially in Taiwan,” concluded Yang. “There are no dinosaur fossils in Taiwan but that does not mean that we cannot do dinosaur studies.”

Lateral view of the clutch with the incubator on top

Photograph of the generalized clutch after Experiment III.

Dorsal view of the incubator.

Credit

Chun-Yu Su.

The arrangement of thermometers in the incubation experiments. Thermometers 1 (with thicker outlines) were used in Experiment II. Thermometers 2 (with lighter outlines) were the additional thermometers used in Experiment III. The schematic presents a lateral view of the clutch and the incubator.

Credit

Su et al.,2026.

---

Journal

Frontiers in Ecology and Evolution

DOI

10.3389/fevo.2026.1351288 

Method of Research

Experimental study

Subject of Research

Not applicable

Article Title

Heat transfer in a realistic clutch reveals a lower efficiency in incubation of oviraptorid dinosaurs than of modern birds

Article Publication Date

17-Mar-2026

 

What flocking birds can teach AI

Researchers draw from flocking bird patterns to help AI produce more reliable outputs

New York University

FacebookXLinkedInWeChatBlueskyMessageWhatsAppEmail

Among the primary concerns surrounding artificial intelligence is its tendency to yield erroneous information when summarizing long documents. These “hallucinations” are problematic not only because they convey falsehoods, but also because they reduce efficiency—sorting through content to search for mistakes of AI outputs is time-consuming.

To help address this challenge, a team of computer scientists has created an algorithmic framework that draws from a natural phenomenon—bird flocking—by mimicking how birds efficiently self-organize. The framework serves as a preprocessing step for large language models (LLMs), helping them produce more reliable summaries of large documents.

The work is reported in the journal Frontiers in Artificial Intelligence. 

The researchers created the bird-flocking algorithm by first unpacking how AI agents make mistakes. 

These systems are built on LLMs that are designed to autonomously research, write, and summarize. But while they may write well, they do not always produce accurate or faithful summaries. 

“One contributing factor is that when input text is excessively long, noisy, or repetitive, model performance degrades, causing AI agents and LLMs to lose track of key facts, dilute critical information among irrelevant content, or drift away from the source material entirely,” explains Anasse Bari, a computer science professor at NYU’s Courant Institute School of Mathematics, Computing, and Data Science and director of the Predictive Analytics and AI Research Lab, which conducted the work.

Drawing from the cause of this shortcoming, Bari and co-author Binxu Huang, an NYU computer science researcher, turned to an orderly and time-tested method of gathering disparate parts—bird flocking—and applied it as a preprocessing step to generative AI.

Their method considered each sentence in a long document—a scientific study or a legal analysis—as a virtual bird. In yielding a simplified outcome, it evaluated the document’s sentences based on their position, thematic centrality, and topical relevance, then grouped them into clusters that mirror how birds self-organize into flocks.

This grouping reduced each cluster to its most representative sentences, with the goal of minimizing redundancy and preserving key points. The resulting curated summary was then passed to an LLM as a structured, concise, and reduced input. 

“The intention was to ground AI models more closely to the source material while reducing repetition and noise before generating a final summary,” says Bari, who previously turned to natural phenomena in devising an algorithm to improve online searches.

Here is how it works in greater detail:

Phase 1: Score Every Sentence

• Each sentence is cleaned by keeping only nouns, verbs, and adjectives, while stripping out articles, prepositions, conjunctions, and punctuation. Among other natural language processing techniques, multi-word terms are also merged (“lung cancer” becomes “lung_cancer”) so single concepts stay intact. 

• Each sentence is then converted into a numerical vector by fusing lexical, semantic, and topical features. Sentences are scored on document-wide centrality, section-level importance, and alignment with the abstract, with a numerical boost for key sections like the Introduction, Results, and Conclusion.

Phase 2: Bird Flocking for Diversity

• Taking only the top-scored sentences risks repetition—and stymies flocking. For instance, in a cancer research paper, the five highest-ranked sentences might all discuss treatment outcomes. Instead, the framework treats each sentence as a bird positioned in an imaginary space according to its meaning. Much like real birds in nature, which self-organize into flocks by following three simple rules known as cohesion (stay close to nearby birds), alignment (move in the same direction as neighbors), and separation (avoid crowding), sentences with similar meanings naturally cluster together while maintaining distinct groupings. Leaders emerge within each cluster and followers attach to their nearest leader. 

• From each final flock of similar sentences, only the highest-scoring ones are selected, so the summary covers background, methods, results, and conclusions, rather than echoing one theme—thereby reflecting a document’s diversity of content without repeating it. The chosen sentences are reordered and fed to an AI agent powered by an LLM, which synthesizes them into a fluent summary grounded in the original source content.

• The researchers evaluated the algorithm on over 9,000 documents, examining whether this approach produced better outputs compared to an AI agent powered by an LLM alone. The framework, including its bird-flocking-inspired algorithm, combined with LLMs, helped generate summaries with greater factual accuracy than did LLMs producing content without the algorithm.

“The core idea of our work is that we developed an experimental framework that serves as a preprocessing step for large texts before it is fed to an AI agent or LLM and not as a competitor to LLMs or AI agents,” Bari says. “The framework identifies the most important sentences in a document and creates a more concise representation and summary of the original text, removing repetition and noise before it reaches the AI.” 

However, the authors acknowledge that their approach is not a panacea.

“The goal is to help the AI generate summaries that stay closer to the source material,” notes Bari. “While this approach has the potential to partially address the issue of hallucination, we do not want to claim we have solved it—we have not.”

# # #

Editor’s Note: In November 2025, NYU announced the establishment of the Courant Institute School of Mathematics, Computing, and Data Science. The newly established school recognizes the storied history of the Courant Institute of Mathematical Sciences—and its strengths in both applied and pure mathematics—while encompassing NYU’s Center for Data Science and linking the computer science departments at Courant and the Tandon School of Engineering.

 

---

Journal

Frontiers in Artificial Intelligence

DOI

10.3389/frai.2026.1703769 

Method of Research

Data/statistical analysis

Article Title

A Bird-Inspired Artificial Intelligence Framework for Advanced Large Text Summarization

Article Publication Date

17-Mar-2026