Page 4 – Physics World

Home » Page 4

Radiotherapy trials shape future treatments of brain metastases

At this week’s ASTRO 2026 meeting in Boston, MA, researchers shared the results of clinical trials addressing important and longstanding questions in cancer care. In a dedicated news briefing highlighting noteworthy research studies, two speakers presented new evidence that could help inform how we deliver care to patients with brain metastases.

Brain metastases – cancerous cells that spread to the brain from another site in the body – are typically removed surgically or treated with radiotherapy. For most types of cancer, the preferred approach is stereotactic radiosurgery (SRS), in which high-dose radiation is precisely directed to individual lesions while limiting dose to surrounding healthy brain tissue.

Randomized trials have shown that SRS results in superior cognitive preservation and quality-of-life compared with whole-brain radiotherapy (WBRT), where the entire brain is irradiated to ensure treatment of both visible tumours and microscopic disease. There’s an exception, however. Historically, for brain metastases arising from small cell lung cancer (SCLC), a rare, fast-growing cancer, WBRT remains the standard-of-care. Results of a new trial challenge this approach.

Chad Rusthoven from the University of Colorado Anschutz School of Medicine shared the findings of the first phase III randomized trial to compare SRS versus WBRT for brain metastases from SCLC. The trial found that after adjusting for patient and disease characteristics, SRS was associated with a 40% lower risk of death than WBRT.

Rusthoven noted that recent advances have improved outcomes for patients receiving WBRT. “This includes addition of memantine [a drug that helps protect cognitive function], as well as hippocampal avoidance techniques,” he explained. “As a result, HA-WBRT plus memantine is now the preferred whole-brain treatment strategy for cognitive preservation.”

As such, the NRG-CC009 trial – which included 151 patients in the US and Canada – randomly assigned participants to receive either focused SRS to visible metastases or HA-WBRT plus memantine. All subjects had a median of two SCLC metastases, although some had more than 20.

For the primary endpoint – the time before a patient experienced cognitive decline – the study found no significant difference between the two treatment groups. However, patients treated with SRS had a median overall survival of 17.4 months, compared with 8.6 months for those in the HA-WBRT group. Rates of serious treatment-related side effects or death from neurologic causes were similar for the two cohorts.

Although no differences in neurocognitive failure were observed, the superior overall survival with SRS addresses a fundamental historical objection to SRS alone for small cell lung cancer brain metastases. That is, the concern that omission of WBRT would lead to worse overall survival in this population. Overall these findings support SRS as a treatment option for patients with brain metastases from small cell lung cancer,” Rusthoven concluded.

The power of fractionation

Speaking at the same briefing, Ayal Aizer from Mass General Brigham/Dana-Farber Cancer Institute explained how fractionated radiosurgery can reduce the risk of cancer recurrence after surgery for large brain metastases, without significantly increasing side effects. Fractionation – in which the radiation dose is divided into several treatments – gives healthy tissue time to recover between doses while still effectively treating the surgical cavity.

Large metastases, which can cause pressure on the brain leading to symptoms such as seizures, headaches or difficulties with speech or movement, can be removed via surgery. This is usually followed by radiotherapy to destroy any microscopic cancer cells left behind in the surgical cavity. Aizer explained that current guidelines recommend treating the cavity with SRS, following two pivotal trials demonstrating favourable outcomes for SRS compared with postoperative WBRT and observation.

For larger surgical cavities, however, the challenge, is how to deliver enough dose to eliminate any remaining cancer cells without exposing too much surrounding healthy brain to radiation. “In both of these trials where single-fraction SRS was utilized, the one-year surgical bed recurrence rates were high: 40% in one trial, 28% in the other,” he explained. “Meanwhile, alternative approaches such as fractionated SRS to the surgical bed had been emerging.”

Aizer described Alliance A071801, the first randomized phase III trial to directly compare single-fraction versus fractionated SRS for resected brain metastases. The multi-centre trial enrolled 242 patients with one to four brain metastases, one of which was larger than 2 cm and surgically removed. The metastases arose from different primary tumour types, the most common being lung cancer.

The team randomly assigned patients to receive either a single session of SRS, with dose based on target size, or fractionated SRS: 27 Gy in three fractions for smaller targets or 30 Gy in five fractions (9%) for larger ones.

One year after treatment, 87% of patients who received fractionated SRS were free from recurrence at the surgical site, compared with 81% of those who received single-session radiosurgery. Side effects such as radiation necrosis and cerebral oedema were low in both treatment groups, with no significant differences between the two.

Siemens Healthineers unveils ultrafast Accela radiotherapy system

The overall survival was also improved after fractionated SRS, with patients surviving a median of 29 months, compared with 20 months for those treated in a single session. The researchers note that this survival difference was unexpected and it is not yet clear why this occurred.

“In conclusion, we can say that fractionated SRS improved surgical bed control compared with single-fraction SRS after resection of a brain metastasis,” said Aizer. “The improvement in surgical control was achieved without an apparent increase in toxicity. For patients undergoing resection of a larger brain metastasis, postoperative fractionated SRS should be considered the standard-of-care.”

Demon in the AI machine: bias, bad data and the burden of proof

Picture a room where every photon bouncing off a doorbell camera, every acoustic reflection captured by a passing delivery drone, and every drop of ambient thermal noise is fed directly into a Bayesian inference engine. This is the courtroom depicted in director Timur Bekmambetov’s 2026 film Mercy, which delves deep into the murky boundary between artificial intelligence (AI) and human behaviour.

Set in 2029, the film features Los Angeles police detective Chris Raven, who is also the defendant, played with an increasingly desperate, sweat-soaked intensity by Chris Pratt (Jurassic World, Guardians of the Galaxy). He is starting his trial at a staggering 97.5% probability of guilt.

The judge sitting across from him isn’t a human in robes, but an AI avatar named Judge Maddox, brought to life with icy, terrifying calm by Rebecca Ferguson (Silo, Dune). Raven has precisely 90 minutes to look into the municipal network, cross-reference any telecommunication data recorded by the system, and drive that probability score down below 92%, or a focused acoustic shockwave will end his life by rupturing his internal organs on the spot.

Mercy has not filled critics with delight (it got a score of just 24% on the Rotten Tomatoes “Tomatometer”) but audiences are finding much to enjoy in it (81% on the Rotten Tomatoes “Popcornmeter”). Beneath the reasonably compelling visuals and a home-movie-night-worthy whodunnit lies a fascinating (and possibly accidental) masterclass in the limits of computational physics.

The premise of the “Mercy Capital Court” rests entirely on a classic piece of scientific hubris: the concept of “Laplace’s Demon”. The concept dates back to 1814, when French mathematician Pierre-Simon Laplace postulated that if an intellect knew the exact position and momentum of every atom in the universe, it could calculate the past and future with absolute certainty. Judge Maddox is Hollywood’s latest incarnation of that demon, vacuuming up petabytes of data from the urban landscape under the assumption that total observation inevitably yields objective truth.

As the film unfolds within its high-stress 90-minute limit, reality asserts itself through the unavoidable laws of measurement and noise. The court treats its data streams as pure, classical bits; but the reality is that the physical world generating those feeds is still fundamentally chaotic. A camera feed misses a vital angle because of atmospheric diffraction; a digital packet drops over a high-latency mesh network; and a corrupted timestamp introduces entropy into the dataset.

Ferguson’s AI character acts less like an objective judge and more like the ultimate “black box” algorithm: impenetrable, unyielding and bound by the feedback loops of its own training data. This is not just cinematic exaggeration but mirrors the real-world crisis of predictive risk-assessment tools like COMPAS.

Used across American bail and sentencing hearings, COMPAS was famously exposed for generating disproportionately high risk scores for Black defendants, regardless of actual reoffending rates. It was a stark lesson in how automated systems codify historical bias under the guise of neutral maths. So when Pratt’s character desperately feeds fresh evidence into the system to shift his 97.5% guilt probability, he too is fighting a version of algorithmic bias. This is a feedback loop where an initial prediction, once generated by a closed-source formula, becomes nearly impossible for a human to contest.

The absurdity of the courtroom’s execution chair – where a single computational calculation yields an instantaneous, violent sentence – amplifies the real-world stakes of automated misidentification. In recent years, away from the cinema screen, law-enforcement reliance on automated tools, particularly facial-recognition software, has already led to disastrous false arrests.

With innocent citizens including Robert Williams and Christopher Gatlin being jailed for crimes they didn’t commit due to faulty algorithm matches, we have already seen what happens when police and courts outsource critical human judgement to automated systems with known failure rates. Mercy takes this terrifying precedent to its absolute extreme, demonstrating what happens when we strip away human discretion, constitutional due process, and the right to meaningful cross-examination in favour of algorithmic convenience.

Ultimately, the film succeeds as both a relentless, clock-ticking thriller and a thought-provoking slice of near future paranoia. Bekmambetov’s direction keeps the 2029 world feeling real enough to suspend disbelief, while Ferguson provides a wonderfully nuanced counterweight to Pratt’s frantic energy.

More than anything, the film serves as a stark reminder that no matter how much data we feed an AI, an algorithm cannot understand context, empathy or justice. As AI continues to creep into real-world courtrooms, the film’s message is clear: substituting automated probability for human reason doesn’t remove human bias, it merely automates it.

2026 Amazon MGM Studios 1 hr 41 min

Exploring efficiency and durability in fuel cell catalysts

An important low-carbon technology is the hydrogen fuel cell, which can be used in vehicles, backup power systems and other applications. Fuel cells convert hydrogen into electricity, producing only water as a by-product. A key process inside the cell is the oxygen reduction reaction, in which oxygen from the air is reduced at the cathode. This reaction is relatively slow and is one of the main factors limiting fuel-cell performance.

To speed up this step, platinum catalysts are used. While highly effective, platinum is expensive and gradually degrades during operation. Much research therefore focuses on reducing the amount of platinum required while improving performance and catalyst lifetime. One approach is to use platinum alloys, where a cheaper metal is mixed with platinum to improve efficiency. However, these alloying metals can gradually leach out, causing the catalyst to lose activity over time.

In this work, the researchers aimed to address the trade off between catalytic activity and long-term durability by investigating two alloy catalysts, platinum-copper (PtCu) and platinum-iron (PtFe). These were attached to a two-dimensional material, similar to graphene, called a MXene, which can improve performance while helping to hold the nanoparticles in place and enhance their stability. PtCu/MXene showed the highest catalytic activity, as copper modifies the platinum surface to improve oxygen reduction and accelerate the reaction. PtFe/MXene was less active but more durable, with strong Fe-O-MXene interactions helping to stabilise the catalyst during long-term operation.

By using in-situ X-ray spectroscopy, the researchers were able to observe the catalysts during operation and identify the structural and electronic changes responsible for their performance. This revealed how alloy composition and catalyst supports influence activity and durability, the study provides a roadmap for developing improved catalysts for future fuel cells.

Read the full article

Decoupling activity and durability in MXene-supported Pt–M (M = Cu, Fe) alloy catalysts for oxygen reduction via in-situ XAS and DFT

Sharon Benny Alex et al 2026 Prog. Energy 8 035003

Do you want to learn more about this topic?

Catalyst designing strategies for electrochemical CO2 reduction: a perspective by Shreya Sarkar and Sebastian C Peter (2022)

Drones, AI and the future of wildlife conservation

From tracking elusive animals to protecting some of the world’s most endangered species, the latest episode of Physics World Stories explores how cutting-edge robotics, drones and ecological science are coming together on the front lines of global wildlife conservation.

You’ll hear from Blair Costelloe, leader and principal investigator of the EU-funded WildDrone and WildBotics doctoral training networks, and director of research and conservation at Kenya’s Mara Elephant Project. Costelloe explains how drones, machine learning and other physics-based technologies are opening new possibilities for conservationists, from monitoring animals across large areas to collecting data that would be difficult to obtain on the ground.

Show host Andrew Glester also catches up with Camille Rondeau Saint-Jean, a PhD researcher at the University of Southern Denmark. She is developing and operating drones to track animal body data in real time, offering new insights into population dynamics and conservation. It’s the latest chapter in a career that has already taken her from musicology to drone piloting and conservation biology.

Finally, William Njoroge of Kenya’s Ol Pejeta Conservancy discusses how drones are being integrated with more traditional conservation methods, as well as the practical challenges involved, from technical constraints to regulatory requirements. Set against the backdrop of Mount Kenya, Ol Pejeta is an integrated conservation site and home to the world’s remaining two northern white rhinos.

The episode highlights the fascinating science that can emerge from collaborations across disciplines and cultures, united by a common goal: protecting wildlife.

Graviton modes survive on the lattice

In some materials, when many electrons interact very strongly, they behave collectively like a quantum fluid with unusual properties. One example is a Fractional Quantum Hall state, which occurs in a two-dimensional system under a strong magnetic field. In this state, the electrons form a topologically ordered quantum fluid that supports unusual collective excitations.

A Fractional Chern Insulator aims to recreate the same type of physics without requiring such a strong magnetic field. Instead, the crystal lattice and electronic band structure are engineered to mimic the effects of the magnetic field.

The electron fluid has an internal geometric structure that describes how electrons are correlated with one another. When this internal geometry oscillates collectively, it produces a special excitation known as a graviton mode.

In this work, the authors developed new mathematical tools and performed large-scale computer simulations to search for graviton modes in lattice-based systems. They then smoothly transformed a known Fractional Quantum Hall state into a Fractional Chern Insulator and showed that the graviton mode persists throughout the transition.

Graviton modes were already known to exist in Fractional Quantum Hall systems and have recently been observed experimentally. However, it was unclear whether they could survive in a Fractional Chern Insulator because the crystal lattice breaks some of the symmetries that are thought to protect these excitations.

Min Long from Meng’s research group at the University of Hong Kong and Zeno Bacciconi from Dalmonte’s group at the ICTP, Università di Bologna. Picture taken by Cheng Huang during Zeno’s visit in Hong Kong. (Courtesy: Dalmonte/Università di Bologna)

The authors found that graviton modes do survive on the lattice. They showed that the Fractional Chern Insulator graviton is not a completely new excitation, but is continuously connected to the graviton found in Fractional Quantum Hall systems. They also found that the graviton decays much more slowly than many researchers expected, meaning it remains a well-defined excitation. Finally, the work suggests that graviton modes could provide a valuable experimental signature for identifying exotic topological phases such as Fractional Chern Insulators.

“It is remarkable how geometric excitations can govern the behavior of inherently discrete lattices, demonstrating unexpected universal features of topological quantum matter.“ – Zi Yang Meng, The University of Hong Kong, and Marcello Dalmonte, Università di Bologna.

Read the full article

Chiral graviton modes in fermionic fractional Chern insulators

Min Long et al 2026 Rep. Prog. Phys. 89 078001

Do you want to learn more about this topic?

Fractional charge and fractional statistics in the quantum Hall effects by D E Feldman and Bertrand I Halperin (2020)

Most comprehensive study yet of the Higgs boson at CMS

The Higgs boson is associated with the Higgs field, which gives many fundamental particles their mass. Although it was predicted in the 1960s, it was not discovered until 2012. Since then, researchers have been studying whether the Higgs behaves exactly as predicted by the Standard Model of particle physics or whether it shows signs of physics beyond the Standard Model. This involves measuring how often Higgs bosons are produced, how they are produced, what they decay into, and how they interact with other particles.

In this work, the researchers combined measurements from proton-proton collisions recorded between 2016 and 2018 by the CMS experiment. These collisions can produce Higgs bosons, but such events are rare, so a huge dataset was needed. The study used 138 fb-1 of data, representing an enormous sample of proton-proton collisions on which to perform measurements of the Higgs. The researchers examined many Higgs decay channels, including decays into photons, bosons, and leptons. They also searched for invisible Higgs decays, in which the Higgs would produce undetectable particles such as dark matter candidates, and studied rare off-shell Higgs production, where the Higgs is produced as a virtual state with a mass different from its physical mass.

A display of a collision event from each of the Higgs boson decay channels included in this measurement. (Courtesy: CMS collaboration)

The researchers found that, with the current experimental precision, the Higgs boson behaves as predicted by the Standard Model. They measured an overall Higgs signal strength of 1.01 ± 0.05, where a value of 1 represents perfect agreement with the Standard Model. The result is therefore fully consistent with theoretical expectations. Many theories that extend the Standard Model predict that undiscovered particles or forces could subtly alter how the Higgs boson is produced or how it decays. Because the CMS measurements agree with Standard Model predictions, there is currently little evidence for such effects, significantly constraining many proposed theories of new physics.

This study provides the most comprehensive test of the Higgs boson to date using data from the CMS experiment. By narrowing the range of possible deviations from the Standard Model, it helps guide future searches for new physics at the Large Hadron Collider and other particle physics experiments.

“It is amazing to see how far we have come in just over a decade since discovery. We are now able to measure the Higgs boson production rate at the LHC with a 5% precision, confirming our theories regarding the origin of mass in the Universe.” – Dr Jonathon Langford from Imperial College London, a lead analyst for this paper.

Read the full article

Combined measurements and interpretations of Higgs boson production and decay in proton-proton collisions at √ s = 13 TeV

The CMS Collaboration 2026 Rep. Prog. Phys. 89 087801

Do you want to learn more about this topic?

Prospects for Higgs physics at energies up to 100 TeV by Julien Baglio, Abdelhak Djouadi and Jérémie Quevillon (2016)

Scientists explain how a 300 TeV photon could reach Earth

On 9 October 2022, a gamma-ray burst called GRB 221009A sent an enormous amount of high-energy radiation toward Earth. The Large High Altitude Air Shower Observatory (LHAASO) detected thousands of photons from the event, including some above 10 TeV. More than an hour later, the Carpet cosmic ray detector recorded something much harder to explain: a photon-like event with an estimated energy of 300 (+43/-38) TeV. At that energy, according to standard physics, the photon should have disappeared during its journey to Earth.

Giorgio Galanti of INAF and Marco Roncadelli of INFN have now examined what could allow such a photon to survive. Their explanation, reported in Physical Review Letters, involves two hypothetical effects: axion-like particles, or ALPs, and a possible violation of Lorentz invariance.

The space between galaxies is filled with background light, including the cosmic microwave background (CMB), the faint radiation left from the early universe. A 300 TeV gamma ray can collide with these low-energy photons and produce an electron and a positron. At this energy, the CMB is the dominant source of absorption.

“The CMB is extremely dense,” Galanti explains. A 300 TeV photon therefore has a mean free path far shorter than the distance from GRB 221009A, making the universe effectively opaque to it.

The numbers are difficult to put into perspective. Galanti and Roncadelli calculate that conventional propagation would result in only about 10^-96 expected photons in the Carpet energy range. The Carpet event itself is also unusual because the detector did not find the muon signal expected from a hadronic cosmic-ray shower. The collaboration estimated the probability of a hadron being misidentified as the event at about 3 × 10^-4.

Galanti says the researchers are taking the published Carpet result at face value while leaving open the question of whether it will ultimately prove to be a signal of new physics.

Two options to consider

One possible explanation involves ALPs, hypothetical particles that can mix with photons in magnetic fields. A photon can convert into an ALP, travel through a region where photons would be absorbed, and later convert back into a photon. This mechanism could help explain the LHAASO photons detected between roughly 10 and 20 TeV. But it runs into trouble at 300 TeV.

“ALPs alone fall short by roughly two orders of magnitude at 300 TeV,” Galanti says. They remain useful at lower energies, but the CMB absorption becomes too severe at the Carpet energy. The researchers found that within the range of ALP parameters under consideration, the expected number of Carpet photons stays below about 10^-4, lower than the 0.0513 required at the 95% confidence level.

That led the researchers to consider another effect: a possible violation of Lorentz invariance, one of the foundations of relativity.

In the scenario they studied, the Lorentz invariant violation (LIV)-induced modifications of photon propagation change the threshold for photon–photon absorption. A 300 TeV photon can then interact with higher-energy background photons, which are much less abundant than the CMB photons responsible for ordinary absorption. The universe then becomes more transparent to the photon.

The researchers calculate upper limits of 1.22 × 10²¹ GeV for the linear LIV scale and 2.03 × 10¹³ GeV for the quadratic case, at 95% confidence.

The interesting part is that they don’t use LIV as a replacement for ALPs. As Galanti explains, ALPs function at LHAASO energies but not at Carpet energies, while LIV operates at Carpet energies but not at LHAASO energies. Their model incorporates both effects into a single theoretical framework.

There is another clue. A separate analysis by Dmitry Ofengeim and Tsvi Piran found that the more-than-one-hour delay between the LHAASO and Carpet events can also be explained with quadratic LIV. Their estimated scale is consistent with the limit obtained by Galanti and Roncadelli.

But Galanti is careful about what the result means.

“A single candidate photon from a single cosmic event cannot be claimed as an undisputed discovery,” he says. Confirmation would require repeated observations from multiple distant sources showing the same energy-dependent signatures.

Future observations from facilities including ASTRI Mini-Array, CTAO, SWGO and the expanded LHAASO could test whether similar high-energy signatures appear in gamma-ray bursts and distant active galaxies.

For now, the 300 TeV event remains an unusual observation with a possible explanation that goes beyond standard photon propagation.

DNA computer settles in thermodynamic equilibrium

A DNA computer that calculates as it settles into thermal equilibrium has been unveiled in Ireland. Drawing together several earlier lines of research, Damien Woods and colleagues at Maynooth University created a DNA-based system that ran calculations far more quickly than they expected, and which could be easily implemented and reused with a relatively simple setup.

Much like a living organism, a computer runs far from thermal equilibrium. It expends energy to maintain encoded bits of information, and to switch between them quickly and reliably as it executes algorithms.

However, in the 1970s physicist Charles Bennett showed that any computation can, in principle, be modified to have reversible dynamics. Such a computer could operate close to thermal equilibrium, slashing its demand for energy.

Taking Bennett’s ideas a step further, others suggested that the output of an algorithm could be delivered in a system’s equilibrium state itself. Starting from a given set of inputs, the computer would then naturally drift towards the solution as it settles into its most stable state.

Multiple challenges

“The idea presents multiple challenges that include finding a physical implementation that is computationally expressive and programmable, has easily prepared initial states and has a controllable energy landscape for rapid navigation to target outputs with high probability,” Woods explains.

One promising platform is the DNA computer, which was first introduced in the 1990s. It replaces silicon chips and electrical currents with DNA strands and biochemical reactions.

Until now, much of the work on DNA computers has involved out-of-equilibrium systems, which are either driven by molecular fuel, or require carefully prepared initial states and tightly controlled experimental conditions. But because DNA strands bind to each other according to simple, predictable rules, researchers could engineer the energy landscape directly. Incorrectly bound strands would then detach and be replaced by better-fitting ones, until the system settles into its most stable state.

To build such a system, Woods’ team was inspired by “DNA origami”. In a 2006, US-based Paul Rothemund showed how a long, single-stranded DNA scaffold could bind to a chosen set of shorter strands. Depending on the strands chosen, the scaffold could fold into a wide array of shapes as it settled into thermal equilibrium.

Tile calculations

Applying these ideas to DNA computing, Woods’ team designed a DNA scaffold carrying an engineered sequence of binding domains, which they mixed into a solution of shorter DNA strands. Dubbed tiles by the researchers, these strands represent the possible values at each step of a calculation. They were also inspired by the notion of molecules contending to bind to the scaffold, a form of molecular competition first devised by US-based Erik Winfree in the 1990s.

After mixing, the tiles compete with each other to bind to the scaffold. At each binding domain, the “winning” tile is the one that binds most favourably to both the scaffold and its neighbouring tiles. In this way, the role of bits in a conventional computer is played by the specific strands that win each position.

“Eventually, the system settles down into its energetically preferred state which encodes the answer to the computation: a sequence of tiles each bound to the scaffold and to neighbouring tiles on its left and right,” Woods describes. “The competitive process of binding executes the computation.”

Altogether, the team demonstrated 10 different programs with their DNA computer, across more than 700 individual computations: including addition, multiplication, and division. Simple computations took as little as a minute to complete, though larger ones took considerably longer.

Just add water

“The same molecular computer can be reused to perform new calculations, running the same program up to 25 times on different inputs,” explains Maynooth’s Abeer Eshra. In one case, the team repeated an experiment after 15 months, simply adding water to a setup that had partially dried out, and still obtained correct answers.

“Our system uses just a handful of different kinds of molecules, never really following an organized process of steps, never making irreversible steps, and yet ending up with the right answer,” adds Maynooth’s Constantine Evans. “When thinking about computation at a molecular level, reliably making even those seemingly simple computations is very hard.”

The programs turned out to be surprisingly fast and easy to implement. To run a computation, the researchers simply mixed all the strands together, heated the solution, and let it cool as the system settled towards equilibrium, without any need for molecular fuel or specially prepared initial states. Remarkably, the team created what is arguably one of the most complex DNA computers built to date, paving the way for a whole host of new directions.

“Although we optimized some aspects of the design, there remain many ways it could be improved and generalized,” Woods says. “The work opens the door to new ways of thinking about computation in a wet environment, and about energy use in computation overall.”

The computer is described in Nature.

Siemens Healthineers unveils ultrafast Accela radiotherapy system

The annual meeting of the American Society for Radiation Oncology is the place to be to discover the latest radiotherapy technology developments and new product launches. And this year was no exception. The ASTRO 2026 meeting, held this week in Boston, MA, saw Siemens Healthineers unveil its newest radiotherapy system – Accela – which combines precision delivery with speeds fast enough to enable stereotactic treatments in a single breath-hold.

Accela achieves this by implementing a new delivery technique known as NeoArc, or dynamic modulated arc therapy (DMAT). DMAT is a type of arc therapy, in which the treatment beam rotates continuously around the patient during treatment, with the beam shape and intensity controlled to deliver the prescribed treatment plan.

But while conventional arc-based approaches such as volumetric-modulated arc therapy (VMAT) primarily optimize treatment plans for the dose distribution, DMAT evaluates dosimetric objectives and delivery efficiency simultaneously. This enables clinicians to tailor treatments for both speed and precision, to optimize the balance between them and, ultimately, to reduce the patient time on the treatment table.

Sasa Mutic, head of radiation oncology solutions at Siemens Healthineers, tells Physics World that the product launch at ASTRO was a great success. “The response to Accela from customers and the medical community has been overwhelming and overwhelmingly positive,” he says. “It was extremely gratifying to see that the ASTRO audience instantly grasped the significance of what we’ve accomplished and the impact it will have on clinicians and patients.”

Done in 60 seconds

Accela integrates high-speed gantry rotation of up to 2.5 rpm, dose rates of up to 40 Gy/min and a dual-layer multileaf collimator (MLC) with leaf speeds of up to 6.5 cm/s. This combination means that, as well as single breath-hold stereotactic treatments, Accela can deliver multi-site stereotactic body radiotherapy (SBRT) in as little as 60 s and go from imaging to beam-off in as low as 90 s – including on-couch imaging with a large HyperSight cone-beam CT panel, couch shifts and the treatment delivery.

An additional benefit of ultrafast radiotherapy is the potential to actually alter the body’s biological response.

“Because the treatment times are potentially as quick as a single breath hold, there is much less blood irradiated,” explains Dee Khuntia, chief medical officer at Varian, the Siemens Healthineers cancer-care business. “This allows for preservation of lymphocytes, which are necessary for helping the body heal, fight infection and fight cancer. It is also well documented that patients undergoing radiation or chemo-radiation who have lymphopenia [not enough lymphocytes in the blood] have lower survival rates.”

The first patient treatments with Accela are scheduled for December in Europe. By this time, the system will have the CE mark, with FDA clearance expected early next year. One of the first centres to use Accela will be Maastro in the Netherlands, which plans to initially treat small lung tumours, lung metastases, brain tumours and metastases in the vertebrae – sites that require a high radiation dose administered extremely precisely, while sparing healthy tissue and nearby vulnerable organs.

Updated on 6 October 2026 to amend the imaging to beam-off time and add a clarification from Siemens Healthineers: Accela and NeoArc are 510(k) pending. Not available for sale or quotation. No guarantee of commercialization or feature availability.

Polymer-tuned perovskite creates high-efficiency underwater solar cells

Terrestrial solar cells rely on capturing a broad spectrum of sunlight, particularly the low-energy, long-wavelength red and infrared regions between 800 and 1150 nm. Emulating this concept underwater fails because water acts as a fierce optical filter, strongly absorbing light at wavelengths of 630 nm or more. Furthermore, descending just a few metres deeper completely attenuates longer wavelengths. This forces the aquatic environment into a twilight zone, where virtually all red light has been absorbed, leaving only a narrow, faint blue-orange glow spanning 400 to 600 nm as the sole source of illumination.

Consequently, conventional silicon, cadmium telluride or standard 1.55 eV perovskite solar cells are poorly suited for underwater operations, as their optimal absorption regions match photons that never reach them. While these constraints have long hindered the development of underwater solar harvesting, stalling the deployment of autonomous marine infrastructure like deep-sea sensors and environmental monitors, a research team in China has now charted a clearer path forward.

In a leap toward self-sustaining underwater solar harvesting, a research team led by Wen-Hua Zhang at Yunnan University has developed a functionalized solar cell for the deep sea. As detailed in Joule, the team built a wide-bandgap perovskite solar cell that thrives in dim environments by capturing the specific blue-green light that penetrates deep water. This unique design allowed the device to generate electricity underwater with a record-breaking 34.71% power conversion efficiency at a simulated depth of 10 m.

Tuning the physics of perovskites

The researchers achieved this breakthrough by engineering a mixed-anion lead halide perovskite cell with an optimized, wide optical bandgap of approximately 1.96 eV, precisely tuned to match the narrow blue-green window of deep water. While wide-bandgap perovskites are theoretically ideal for this application, they suffer from severe material vulnerabilities on land. Exposure to intense light and heat induces volatile halide-ion migration, phase separation and high concentrations of structural defects, causing rapid degradation.

This is where the unique physics of the underwater environment and intelligent molecular engineering come together. The naturally cool temperatures (typically below 25°C) and reduced light intensities beneath the waves act as an inherent shield against thermal degradation. To address the remaining internal defects and ion migration, the researchers introduced a specialized crystallization additive, polyhexamethylene guanidine hydrochloride (PHMG), into the lead halide perovskite matrix.

The addition of PHMG completely rewrites the crystallization dynamics and surface electronic properties of the perovskite film. On a molecular level, the polymer combines a water-repelling backbone with highly stable, chemically active ions (guanidinium cations). These ions act as multisite anchors that lock into empty gaps within the crystal structure, stabilizing and strengthening the entire framework. At the same time, they act like a chemical glue, binding tightly to lead and halide ions through hydrogen bonding to lock the crystal together and prevent structural decay.

This dual-action mechanism raises the activation energy barrier for ion migration from 0.07 to 0.21 eV, effectively freezing phase segregation in its tracks. Advanced testing (using ultraviolet photoelectron spectroscopy) revealed that the PHMG shifts the film’s electronic behaviour from hole-carrying (p-type) to electron-carrying (n-type) conductivity. This shift creates a favourable energy slope (band bending) at the surface, significantly accelerating electron extraction while dramatically driving down energy losses from trapped charges (known as interfacial non-radiative recombination).

From lab simulation to the open sea

These physical adjustments translated to extraordinary performance gains. Under standard terrestrial sunlight, the modified cell achieved a certified power conversion efficiency of 16.79%. But beneath the surface, it truly shone. Tested using an underwater solar simulator containing custom-engineered multilayer interference optical filters, the small-area (0.0895 cm²) cell registered an impressive 34.71% efficiency under 10 m-deep light illumination.

Scaling up the technology, the team built large-area modules with an active area of nearly 29 cm². Shielded by a robust multilayer encapsulation of polyisobutylene, cover glass and protective epoxy resin, these modules withstood extreme hydrostatic pressures and salt-water corrosion.

Deployed on a miniature robot in the open waters of the South China Sea near Weizhou Island, the modules successfully harvested solar energy across varying depths. Even at a depth of 10 m, with severely attenuated illumination, the system maintained stable electricity production, successfully generating 324 mWh of energy to charge lithium-ion batteries underwater. These batteries were later used to illuminate a custom LED panel.

“In contrast to the lab-based static underwater solar simulator, waves, turbidity and suspended organic matter in real oceans can change the light intensity and spectral composition. This leads to substantial fluctuations in output power and reduces the power‑generation stability of the cells,” Zhang tells Physics World. “Therefore, in-field experiments of the large-area solar modules, as we carried out in the South China Sea, are indispensable to assess their real-world power generation.”

The team also performed high-temperature stress testing (accelerated thermal aging tests) and molecular decay modelling (Arrhenius degradation kinetics). These tests predicted a remarkable operational lifespan in which the device retains 80% of its initial efficiency (T₈₀ lifetime) for approximately 48,094 h, translating to 5.49 years of continuous underwater operation at a standard temperature of 25 °C. Furthermore, mass spectrometry confirmed negligible lead leakage into the surrounding water, verifying the environmental safety of the design.

By mastering the unique physics of underwater light transmission and overcoming the lattice instability of wide-bandgap perovskites, the researchers have advanced a frontier for photovoltaics. This development moves solar energy beyond its traditional terrestrial borders and lays down a path toward long-lasting, self-sustained energy networks beneath the ocean surface. Future marine biology sensors, deep-sea research cameras and submerged Internet-of-Things infrastructure could one day owe their autonomy to these high-performance underwater power grids.

When asked about how the device handles biofouling over extended periods, Zhang emphasizes the need for cross-disciplinary scaling: “The current research focuses on high-efficiency and stable solar cells, as well as the integrated application of photovoltaic energy storage systems. Indeed, biofouling is a critical challenge for the long-term operation of underwater photovoltaics. However, this issue involves multidisciplinary efforts and requires collaborative research from more scientists.”

Browse articles by content type