Page 28 – Physics World

Home » Page 28

Measuring the invisible: how internal dosimetry shapes radiation protection

<strong>Safety first:</strong> Internal Dosimetry is the practice of estimating intakes and risk from occupational exposures to radioactive materials (Credit: iStock/magicmine) — **Safety first:** Internal Dosimetry is the practice of estimating intakes and risk from occupational exposures to radioactive materials (Credit: iStock/magicmine)

Radioactive materials have become a vital tool in our modern society, enabling the delivery of powerful medical treatments, the detection of toxic substances, and the generation of energy on a massive scale. But they also pose significant risks to human health, which means that organizations handling radioactive isotopes must have robust procedures in place to manage and monitor the dose received by their workforce and the wider public.

Dosimetry badges provide a visible and straightforward way to measure external exposure to radiation, but they are unable to determine the amount of radioactive particulates that have entered the body. Quantifying these intakes requires internal dosimetry, which typically uses measurements of urine or faecal samples – or less commonly in vivo radiological counts – to estimate the intake of specific radioisotopes. While the internal dose cannot be measured directly, monitoring data from these bioassays can be combined with biokinetic models to assess the likely impact of an individual inhaling, ingesting or being injected with a radioactive substance.

But internal dosimetry is not an exact science, and organizations with limited experience of radiation safety can struggle to translate regulatory requirements and the associated guidance into monitoring programmes that provide appropriate protection for their employees. Luckily, help is at hand. Dade Moeller & Associates, a wholly owned subsidiary of NV5, is a consultancy company specializing in radiation safety and dosimetry monitoring. Dade Moeller’s team of senior and certified health physicists can help organizations to balance worker protection with regulatory compliance.

Fit for purpose

“We encourage people to think about what they are monitoring for,” says Dr Brett Rosenberg, a certified health physicist at Dade Moeller with expertise in internal dosimetry. “The dosimetry programme must be tailored to the activity and half-life of the materials being handled, how those particular radionuclides interact with the body, and the maximum dose that an individual is likely to receive.” All of these considerations are important for prescribing a monitoring programme that does not become too onerous in costs and implementation while still providing an accurate assessment of occupational exposures.

In low-dose environments, for example, Rosenberg would advise organizations to run a confirmatory programme, in which samples from a subset of the workforce are used to check for any unexpected intake that would warrant further investigation. In contrast, routine internal dosimetry becomes a necessary requirement when there is a greater risk of workers receiving doses that exceed monitoring criteria – or even limits – set by the regulations.

Such situations are likely to arise, for example, in facilities that develop and manufacture medical radioisotopes, such as highly volatile iodine-131 for treating thyroid conditions. Radionuclides with the potential to generate higher doses are also produced throughout the fuel cycle for nuclear power generation, from extracting and refining the uranium – which creates radioactive dust that is easily inhaled – through to the management of nuclear waste.

In the US, such commercial operations are overseen by the Nuclear Regulatory Commission (NRC), which sets maximum limits both for the total radiation dose and the intake that a worker can receive. Companies must demonstrate compliance with these limits, but the NRC does not dictate the methodologies that are used for internal monitoring, or for estimating the intake from lab-based measurements. “Programme development and design is not based on regulation, it’s based on guidance,” says Rosenberg.

Flexibility leads to variability

That leaves organizations with significant flexibility in how they design their monitoring strategies, which allows programmes to be tailored to operational realities rather than forced into a rigid template. However, it also leads to variability in monitoring approaches and in the interpretation of results. Annual audits required by the NRC offer an opportunity for experts like Rosenberg not just to verify compliance, but also to identify opportunities for improvement. “I look for technical shortfalls in internal dosimetry, and check that the measures in place are effective and appropriate,” he says.

Many companies value the input they receive through these regular evaluations. Dan Sowers, for example, is the Global Radiation Protection Manager for Westinghouse, a nuclear services company that, among other things, fabricates uranium fuel and maintains the inner workings of nuclear reactors when they are not being operated. “My priority is the health of our workers,” he says. “Ultimately, we want our internal dosimetry programme to demonstrate that workplace controls are effective.”

Regular input from external advisors helps Sowers to ensure that robust procedures are in place for contamination control, and for assigning a dose to an employee when there is an intake of radioactivity. “The experts at Dade Moeller have enough real-world experience to know what’s right, which can go beyond the guidance we get from the regulators,” says Sowers. “They bring a fresh set of eyes to our operation and share best practice from other sites.”

The team at Dade Moeller also provides expert input to the US Department of Energy (DOE), which runs the national labs as well as facilities for nuclear security. Unlike the NRC, the DOE prescribes structured programmes for internal dosimetry that are tailored to the radioactive materials being handled at each site. It mandates lower dose thresholds than specified by the NRC for workers to undergo routine internal dosimetry checks, and uses a newer set of recommendations for assessing intake and estimating the dose received by different parts of the body. (However, notes Rosenberg, neither the NRC nor DOE use the latest set of recommendations, which are being adopted internationally.)

Creating consistency

Many of those recommendations come from the International Commission on Radiological Protection (ICRP), an independent body that provides guidance on dose limits and monitoring methodologies. The ICRP also develops and publishes the biokinetic models that are used by dosimetrists to translate the radiation counts from bioassays into estimates of intake and effective dose. Specific models have been developed to assess the impact of inhaling or ingesting different radioactive species, but the results are open to interpretation and depend on the version of the model being used.

To avoid these ambiguities, national regulators in Europe have combined forces to create a unified framework for internal dosimetry. Based on the outcomes from a co-ordinated research project in the early 2000s, this framework defines a standard set of methods and assumptions to ensure that different organizations calculate similar dose estimates from the same monitoring data. However, no such consistency exists in the US, where the NRC does not specify the details of the monitoring programme or the modelling methods a company should use, or between regulators in other parts of the world.

That lack of consistency can create difficulties for international companies like Westinghouse, which operates across 21 countries in different geographic regions. “Our internal dosimetry programmes are slightly different in each location,” says Sowers. “Those differences don’t affect the fundamentals of how we protect the health and safety of the worker, but the fine print of the regulations often requires us to do things slightly differently to make sure we are compliant.”

Dealing with uncertainty

To improve consistency, Rosenberg is involved in both national and international standards initiatives that aim to define best practice. However, he says, even the most carefully designed monitoring programme cannot eliminate the inherent uncertainty in estimating the effective dose from lab-based measurements of excreted samples.

“Even the most up-to-date models may not be very accurate,” says Rosenberg. “While the bioassay results are taken from a real individual, the model assigns the dose to a reference person who weighs 70 kg, excretes 135 g of faeces per day, and has both ovaries and a prostate.” As a result, the effective dose calculations based on excreted samples can carry errors of 300% when considering the anatomy and physiology of an “average worker”, and the practical limitations around counting statistics and sampling. “For practitioners, the challenge is not to remove uncertainty, but to understand and manage it,” says Rosenberg.

Given these intrinsic uncertainties, Rosenberg’s mission is to help regulators and commercial organizations to deploy the most effective methodologies for improving the standard of care. As part of these efforts, he has developed a week-long training course that aims to demystify the process of internal dosimetry for managers and regulators who need to understand the guidance from regulatory and international bodies like the ICRP.

“We do see uncertainty associated with inappropriate dosimetry practices, and to advance the field both the regulators and the practitioners need to stay current with the latest recommendations and models,” he says. “It’s easy to follow NRC regulations, but putting structured monitoring programmes in place and using the most appropriate models will help companies to produce more accurate evaluations and improve the safety of their workers.”

Physics and the green economy: a slide puzzle

To play, click or drag one or more tiles to move them into the empty space. Repeat until you have reconstructed the image.

Image courtesy: iStock/Galeanu Mihai

Fancy some more? Check out our puzzles page.

Changes in structure and ionic resistance of lithium-ion battery graphite electrodes

Want to learn more on this subject?

Watch now

The fast-charging performance of lithium-ion batteries is strongly influenced by the structure of porous electrodes and its evolution during battery operation. In such electrodes, ionic transport is governed by the electrolyte conductivity as well as by the porosity, tortuosity and thickness of the electrode. The latter parameters undergo both reversible and irreversible changes, induced by the lithiation and delithiation of graphite particles and by the formation of the solid–electrolyte interphase (SEI).

We will discuss how SEI formation alters electrode thickness, porosity and tortuosity. Particular emphasis will be placed on how these structural changes depend on the initial electrode porosity and how SEI formation can be quantitatively linked to the evolution of transport properties.

Second, we will address state-of-charge–dependent effects. Graphite particle expansion during lithiation changes the pore structure, influencing electrode porosity, thickness and pore volume. It will be shown how these changes affect ionic resistance and electrolyte transport, and how they connect to macroscopic effects such as the recently described electrolyte-motion–induced salt inhomogeneity (EMSI).

Together, these insights provide a coherent picture of how SEI formation and state-of-charge-dependent expansion alter ionic transport in porous graphite anodes, offering guidance for electrode design and improved methodologies for realistic parameter determination used in battery modelling and fast-charging optimisation.

Want to learn more on this subject?

Watch now

Lennart Reuter received his PhD from the Technical University of Munich (TUM) in June 2025 under the supervision of Hubert Gasteiger and is now a postdoctoral researcher at Uppsala University. His research centres on identifying and quantifying material and electrolyte decomposition processes in lithium-ion batteries. He implemented OEMS for gas analysis in the presence of volatile commercial electrolytes and developed a novel three-electrode cell for the operando detection of parasitic side reactions. In addition, he investigated structural and transport changes in porous electrodes during operation, including the characterisation of porous transport properties by electrochemical impedance spectroscopy (EIS).

Jonas Dickmanns is a PhD candidate at the Chair of Technical Electrochemistry, supervised by Hubert Gasteiger at the Technical University of Munich (TUM), focusing on advanced lithium-ion battery cell chemistry and anode materials. He received his MSc in chemistry from TUM, where he worked on the electrochemical characterisation of graphite. He later expanded his research to other anode materials, including microscale silicon and silicon/carbon composites. His work investigates key electrode processes such as solid–electrolyte interphase (SEI) formation, structural evolution of anodes, and electrolyte interactions. The goal of his research is to enhance the performance, stability and lifetime of next-generation lithium-ion batteries.

Machine vision begins to work like the human eye

Artificial vision is employed in applications ranging from self-driving cars to advanced robotics. However, most “artificial eyes” do not work well in complex real-world conditions, such as sudden strong glare, dimness or uneven light levels. Such devices also rely on complex and rigid electronic circuitry to adjust their sensitivity to light and on post-algorithmic processing for accurate image recognition. A research team headed up at the University of Electronic Science and Technology of China and Penn State in the US has now developed a new adaptive photomemristor structure that does not suffer from any of these problems, allowing for machine vision that truly adapts to natural and complex light fluctuations, just like human eyes.

Sight is one of our most important senses: we acquire over 80% of the information from our environment through our eyes. Indeed, the human eye is so sensitive that it can process visual information under varied lighting levels with a photodetection dynamic range of more than 160 dB. This allows us to watch a film in a dark room or drive in the glare of oncoming headlights. Until now, however, artificial eyes have lacked this so-called dynamic self-adaptation.

Memristors for improving light adaptation

In recent years, researchers have been looking into neuromorphic devices that mimic the way biological structures work. These devices can be engineered to behave very much like neurons in the human brain, which learn by reconfiguring the strengths of the connections (synapses) between neurons. Memristors (or memory resistors) are excellent in this respect because they can bring this learning functionality to the connections in electronic circuits.

Rather than being fixed, the resistance of memristors changes depending on the current or voltage previously applied to them. This means that specific resistances can be programmed into the devices and subsequently stored. Importantly, the “remembered” value of the resistive state persists even when the power is switched off, making it a non-volatile form of electronic memory.

Photomemristors are a type of memristor that can change their resistance in response to both voltage/current and light and could therefore be used in advanced optical systems.

Reversible water absorption and desorption

In the new work, reported in Nature Communications, a team led by materials scientist Jia Zhu made a highly adaptive optoelectronic synapse from the composite material TiO₂/PEDOT:PSS. The TiO₂ absorbs light, converting it into a photocurrent that then passes through the conductive surface of the PEDOT:PSS polymer. This intrinsic photothermal effect, explains Zhu, triggers reversible water absorption and desorption on the polymer surface, which in turn impacts the device’s conductivity.

“This process allows the device to automatically suppress the photoresponse under intense bright light and amplify its sensitivity under dim weak light,” he says. “In extreme strong-light conditions, its photocurrent can even drop below the dark current level, allowing for an ultralow-sensitivity state similar to human eye pupil constriction.”

A high image recognition accuracy

Zhu tells Physics World that when integrated with artificial neural networks, the new adaptive vision system achieves a high image recognition accuracy of 91.3% in dynamically changing mixed-light scenarios, including glare, dim shadow and alternating brightness. It also eliminates the need for bulky auxiliary circuits and complex real-time computing algorithms, reducing power consumption and hardware volume.

The new work is a first step towards making next-generation neuromorphic vision systems that could be used in practical applications. “Our adaptive photomemristor vision system can effectively resist strong glare from oncoming vehicle headlights, accurately identify pedestrians, lane lines and traffic signals in complex day–night alternating light environments, so greatly improving the safety of autonomous driving,” Zhu explains.

Nanofluidic memristors compute in brain-inspired logic circuits

The technology could also be widely applied to humanoid robots, outdoor intelligent monitoring equipment, aerial reconnaissance and portable intelligent vision terminals, he adds.

Looking ahead, the researchers will now focus on optimizing the material formula of the compound studied in this work and improve its fabrication. They will also be working on complete system-level integration, including array device packaging, signal readout modules and adaptive control units, with the aim of constructing a fully functional bionic artificial eye prototype. “We will then verify whether this biomimetic dynamic regulation strategy can be extended to other oxide/polymer composite memristor systems to develop a universal design principle for high-performance light-adaptive neuromorphic devices,” says Zhu.

Advanced simulations of exotic quantum matter

When electrons are placed in a strong magnetic field and cooled to very low temperatures, they stop behaving independently and instead act as a collective fluid. This is known as a fractional quantum Hall (FQH) system. In this regime, the electrons lose their individual behaviour and act as one correlated system. The strong interactions between them produce quasiparticles called anyons. These are not real particles but effective ones that emerge from the collective behaviour, and they have unusual properties that do not occur in ordinary systems.

Anyons exhibit braiding behaviour: when they are moved around each other, the system remembers the path taken. In Abelian braiding, this only adds a simple phase, so the quantum state is effectively unchanged. In contrast, non-Abelian braiding transforms the quantum state into a different one, which can be measured and used to encode information.

A Moore–Read state of 20 electrons on a sphere, with two quasiholes pinned at fixed positions — A Moore–Read state of 20 electrons (solid dots) on a sphere, with two quasiholes pinned at fixed positions (red open circles). The lower panel presents a double stereographic projection, showing the electrons in the northern and southern hemispheres, respectively. (Courtesy: Ting-Tung Wang/University of Hong Kong)

To study these systems, physicists must compute measurable properties from complicated wave functions, which is computationally challenging, especially for large systems. The traditional method, Metropolis Monte Carlo, is slow and struggles with large numbers of electrons, limiting system size, accuracy, and the ability to test advanced theories.

In this work, the researchers developed a new simulation method, Hybrid Monte Carlo (HMC), for FQH wave functions. It includes global updates (updating many electrons at once) and double stereographic projection to more accurately sample particle positions, making it much faster and more efficient than the Metropolis method.

Using this improved method, they accurately computed edge properties that reveal the system’s topology, as well as high-quality braiding matrices of non-Abelian quasiholes, which are crucial because such braiding underpins topological quantum computing. Overall, this work enables fast, large-scale simulations, allowing more accurate study of exotic quantum states and their anyon behaviour, advancing both fundamental physics and topological quantum computing.

Read the full article

Hybrid Monte Carlo for fractional quantum Hall states

Ting-Tung Wang et al 2026 Rep. Prog. Phys. 89 068001

Do you want to learn more about this topic?

A review of the quantum Hall effects in MgZnO/ZnO heterostructures by Joseph Falson and Masashi Kawasaki (2018)

Controlling glycine polymorphs through nanoconfinement

Molecules such as glycine, the simplest amino acid, can crystallise into different polymorphs with the same chemical composition but different structures. The two phases of glycine are α-glycine, which forms easily and is stable in bulk conditions, and β-glycine, which is difficult to produce and unstable in the bulk. However, β-glycine is piezoelectric; it generates electricity when compressed or bent and is therefore technologically useful.

In this work, the researchers developed a new method to produce glycine crystals. Using an electric field, they sprayed a glycine solution to create nanoscale droplets in which crystals form. By controlling the spraying conditions, they confined crystal growth at the nanoscale, thereby controlling how the crystals form. Crystals below 120 nm in size were pure β-glycine, above 130 nm were mostly α-glycine, and sizes between 120-130 nm produced a mixture of both polymorphs. This shows that crystal size (nanoconfinement) is the key parameter controlling which polymorph forms, and β-glycine is stable between 5 and 120 nm.

The crystals form in two steps. Firstly, nanoclusters form, then these clusters rearrange into crystals. β-glycine forms first because it has a lower interfacial energy, making it easier to create surfaces, and a lower nucleation barrier, so it forms faster. Overall, β-glycine is kinetically favourable. However, α-glycine has a lower bulk free energy and better packing. As crystals grow larger, they can rearrange into α-glycine, which is thermodynamically more stable. Under nanoconfinement, molecular rearrangement is restricted, preventing the formation of the more ordered α-phase and stabilising β-glycine.
This research is important because it shows how to precisely control which crystal form a material takes using nanoscale confinement, rather than relying on trial and error methods. It also provides a general strategy for stabilising metastable but functional materials, with potential applications in sensors, energy harvesting, and advanced functional materials.

Read the full article

Electric-field-driven nanoconfinement and the 5–120 nm stability regime of piezoelectric β-glycine

Kexin Zhang and Zhengbao Yang 2026 Rep. Prog. Phys. 89 058001

Do you want to learn more about this topic?

Physical properties and atomic arrangements in crystals by W A Wooster (1953)

Antineutrino detectors could spot signs that a fusion reactor is producing weapons-grade plutonium

Telltale patterns of antineutrino emissions could reveal whether fusion reactors have been reconfigured to produce material for nuclear weapons, say physicists in the US. Although commercial fusion power plants are still some years away, members of the team at Virginia Tech and Princeton University argue that it is nevertheless useful to develop robust ways of monitoring their output now, to ensure the technology is not misused.

“Neutrinos cannot be shielded, their signatures cannot be spoofed and they can be detected from a distance, either onsite or offsite, allowing for nonintrusive monitoring of reactor operation,” says team member Patrick Huber, a physicist at Virginia Tech’s Center for Neutrino Physics. Better still, the team’s calculations show that existing antineutrino detectors such as Virginia Tech’s mobile MiniCHANDLER are sensitive enough to spot these signatures.

The DT process and nuclear proliferation

Most fusion technologies rely on the deuterium-tritium (DT) fusion process, which occurs when a deuterium nucleus (a proton bound to a neutron) and a tritium nucleus (one proton and two neutrons) fuse to produce helium-4, a free neutron and large amounts of energy. Because fissile materials are not directly involved in this reaction, the nuclear proliferation risk of fusion reactors is widely regarded as being much lower than that of fission power plants.

That said, if neutrons released in the DT process were combined with small amounts of material such as uranium-238 or thorium-232 in the “blanket” that surrounds the DT fuel in the fusion reactor, it would, in principle, be possible to use the reactor to produce plutonium-239 or uranium-233. Both materials are used in nuclear weapons, and a gigawatt-scale fusion reactor operating in this mode could produce tens of kilograms per week – enough for several first-generation bombs.

Checking for nuclear fingerprints

Crucially, though, this clandestine production would not necessarily go unnoticed. Whenever radioisotopes in the blanket absorb neutrons during the fusion process, they give off a distinctive pattern of antineutrinos. And according to the Princeton-Virginia Tech team’s simulations, the current generation of antineutrino detectors is already capable of distinguishing this pattern from the one produced during normal reactor operations.

In the current study, which appears in Phys. Rev. Applied, the researchers focused on the plutonium antineutrino signal. Using a Monte Carlo particle transport code called MCNP6, they modelled the output of a simplified toroidal fusion reactor with an outer radius of 6.2 m and an inner radius of 2.0 m. This reference reactor would operate at a typical power density, and Huber explains that it would produce a total of 1500 MW of power through plasma fusion via the DT reaction. The team also analysed two designs for the “blanket”: one featuring a molten salt (LiF-BeF2 or FLiBe) and one that relied on dual-coolant lead-lithium (DCLL).

Next, the two Princeton members of the team, Alexander Glaser and Robert James Goldston, defined and evaluated a reference fissile material production scenario. As part of this, they estimated both the plutonium production rate and the energy release that occurs as some of the covertly introduced nuclear material undergoes fission. Huber explains that this evaluation provided the first part of the antineutrino background signal.

Glaser and Goldston then analysed the neutron activation and antineutrino signatures driven by neutrons emanating from the fusion plasma. These provided the second part of the background. Finally, the team used these signatures to assess the detectability of covert fissile material production against the combined background from neutron activation and cosmogenic antineutrinos.

Based on these simulations, the team concluded that a detector like MiniCHANDLER could pick up on the production of a few kg of plutonium over 30 days – a result that Huber says should be applicable to any fusion power system that relies on DT fusion. Importantly, he points out that unlike some alternatives, antineutrino-based monitoring does not require inspectors to have access to reactor buildings and could allow plants to be monitored around the clock.

Practical advantages

The idea of using antineutrinos to detect illicit nuclear material production is not new. It dates back to the 1970s, and the two Princeton team members have previously worked on nuclear monitoring and verification for a variety of nuclear reactors and nuclear fuel-cycle facilities (including some in Iran). Given the rapid development of fusion energy systems, Hubner tells Physics World it was “a natural next step” to consider non-intrusive, antineutrino-based monitoring options for them, too. He adds that the group has been studying possible safeguards for fusion reactors since the early 2010s and was part of an IAEA consultancy on the topic in 2013.

As well as optimizing the design of detector systems, Huber says he and his colleagues are also interested in studying thorium as an alternative “fertile” material. “Even though the fission signature would be weaker compared with the uranium reference case analysed in this study, the results we have reported on already suggest that such an effort is well justified,” he says.

Evidence in action: how science helps us make better decisions

How do we make good, well-informed decisions? This is the central question of author Helen Pearson’s new book Beyond Belief: How Evidence Shows What Really Works. It examines evidence-based decision-making and compares it with “conventional wisdom and questionable opinion”. Pearson is interested in how we can use scientific reasoning to empower decision-making using empirical evidence, rather than choices based on opinion or public sentiment.

Pearson uses human stories and historical narratives to outline “the evidence revolution”, in which academics, policy-makers and civilians have tried to embed evidence into decision-making. She tackles disciplines where evidence is seldom used, such as people management, or where its absence is surprising, like environmental conservation. Throughout Beyond Belief, Pearson asks us to understand what makes for good evidence, and how evidence can be better used in daily life.

An award-winning science journalist and editor at the journal Nature, Pearson is particularly good at establishing clear narratives based on a host of different scientific studies. She begins with the field of medicine, focusing on Iain Chalmers, one of the founders of the Cochrane Collaboration. Formed in 1993, it is considered to be a trailblazer in evaluating evidence-based healthcare. Indeed, the institution has influenced important changes to treatment, including the use of corticosteroids to save the lives of premature babies.

It seems obvious, Pearson suggests, that doctors would choose treatments based on rigorous evidence. Instead, she shows us that evidence-based medicine is a young field that only emerged in the 1970s. One particularly personal example she gives is the standard practice of shaving knee cartilage in meniscal tears – a procedure I nearly underwent in 2015 – which was shown to be no better than exercise therapy after a landmark trial in 2016. Beyond Belief offers a glimpse into how recent, and as-yet fragile, our reliance on evidence-based treatment is; as well as how much opposition it has faced from institutional inertia and scepticism.

The book is not just about medicine. It includes examples ranging from the roll-out of the pioneering Mexican social welfare scheme, PROGRESA, which helped to decrease poverty and improve health and education in the country; to the ineffective use of bat bridges over the UK’s A11 dual-carriageway. She also discusses using evidence to make management decisions, as opposed to relying on corporate wisdom or management consultancy. The examples are often surprising, underscoring Pearson’s central argument that evidence allows for better decisions, but getting people to use good evidence consistently is challenging.

Throughout the book, Pearson is engaging and often witty, while still discussing the complexities of implementing evidence-based best practice. Beyond Belief is highly readable too – even when navigating difficult concepts in meta-analyses. Pearson outlines the difficulties of teaching people how to use evidence, how to design studies well, and how studies can easily be misleading. She presents the enormous challenge of encouraging institutions to use evidence correctly, while remaining hopeful about our direction of travel.

There are places where tensions arise, one being in the chapter on policing. Pearson describes how evidence can be used in effective policing, for example, to support increasing patrols or stop-and-search in crime hot-spots. This reveals an interesting question about what the role of successful policing is – crime reduction, retributive justice or rehabilitation – and whose definition of justice that evidence is being used to support.

In these cases, the author’s arguments would have been strengthened if she had engaged more with the central premise of the book: that evidence always makes for better decisions. There are nuances that arise between evidence-based policy and harm reduction, and broader questions of trust that cause people to rebuke evidence entirely. A more serious conversation about these reservations might have deepened the argument in favour of empirical evidence, particularly in questions of social policy.

This discussion is not entirely absent. One particular strength of Beyond Belief is Pearson’s examination of COVID, drawn from her experience as a Nature editor during the pandemic. In this chapter, close to the book’s end, Pearson presents the difficulties of using evidence-based trials in acutely developing crises, in which scientists and public-health bodies have to adapt quickly to minimize harm. At a time when discussions about public health responses to COVID are so polarized, Beyond Belief provides a nuanced analysis of how we can arrive at a scientific conclusion under immense pressure.

At the heart of this book is Pearson’s passion and advocacy for a scientific approach to decision-making. She includes a call to action, empowering readers to be champions of evidence in their lives. Although only present in the short chapters towards the book’s end, her anecdotes about childcare, artificial intelligence and using your vote to support evidence through democracy, are useful reference tools. She also succeeds in her goal of starting a conversation about how we can best use evidence.

Ultimately, Beyond Belief sets out to demonstrate the ways in which evidence enters our lives and improves them. Pearson asks us to advocate for science in a world of increasing misinformation, and to challenge conventional wisdom while equipping us with the tools to do so. For that alone, her book is well worth reading.

2026 Princeton University Press 368pp £25 hb

AI-enhanced rare-event sampling helps predict extreme weather

The frequency of extreme weather events can be predicted more accurately than presently possible by combining artificial intelligence (AI) with physical climate modelling, using a protocol called rare-event sampling. That’s the conclusion of a study from researchers in the US and France. The researchers used the approach to model extreme heat events such as the one currently roasting Europe, but they believe it could also be applicable to many other extreme events in climate science.

As the global climate warms, extreme weather is becomingly increasingly deadly and difficult to predict, and quantifying the risks of such events is important for climate change adaptation and mitigation. “Very often the rarest events carry the largest impacts,” says climate physicist Amaury Lancelin of France’s Laboratory of Dynamic Meteorology.

Full global climate model simulations will need to run for an infeasibly long time to produce sufficient data for an estimate of the precise frequencies of rare events. Alexander Wikner of the University of Chicago in Illinois gives the example of an event that, on average, occurs once every century: “Each year you’re going to flip a coin that has a one-in-a-hundred chance of coming up heads,” he says. “On average, once every hundred years, you’ll get one head, but you could very easily get no heads or two or three heads.” A meaningful estimation of the frequency therefore would traditionally require around a thousand years’ worth of simulation – which is not computationally feasible in a physical model.

Deep-learning algorithms, which ignore the underlying climate physics and simply train themselves using pattern recognition, require up to 10,000 times less computing power. However, the reliability of these in modelling extremely rare events is questionable. For example, there is evidence that an AI model will not predict more extreme cyclones than it has seen in its training data, “which is a quite unnerving idea if we imagine that tropical cyclones will be getting more extreme in the future and we want to use an AI model to forecast them,” Wikner explains. They also produce no insight into the physics of the events.

Rare-event sampling (RES) refers to techniques that preferentially allocate computing resources to specific rare events of interest, thereby avoiding modelling long periods of time in which none of those events occur. The problem, of course, is that one has to have some idea what kinds of conditions increase the probability of a rare event occurring in order to know which periods to model in more detail.

In the new research, scheduled for publication in Physical Review Letters, Lancelin, Wikner and colleagues developed the AI+RES framework. An AI algorithm runs climatic simulations repeatedly and selects those that it predicts are most likely to lead to the rare event. A full climate model then simulates only these. The researchers used this protocol to compare the frequency of heatwaves at mid-latitudes, using the predictions of a relatively coarse-grained direct numerical simulation called PlaSim as the ground truth.

They found that their technique produced similar results to PlaSim with up to 1000 times lower computational resources. The researchers now hope to apply the technique to predict other types of extreme events and apply it to more sophisticated models.

“The reason we used PlaSim is that it is computationally quite cheap compared to state-of-the-art climate models so that we can actually verify that our methodology works,” says Lancelin. “Otherwise there is no way to know whether we have made a garbage prediction or not without the ground truth.” Knowing that the AI+RES model reproduces the PlaSim results correctly, however, the researchers now hope to apply computationally expensive models in situations where it was previously impossible, with confidence that the results are likely correct.

Climate scientist Robin Noyelle of ETH Zurich believes the work marks a significant contribution to ideas that were already “floating out there”. He says that, although RES has been in use for nearly ten years, selecting the best candidates for full modelling has always proved challenging: “People were using things that we thought were reasonable – I did that during my PhD – so if I want to look at hot summers, then I select on temperature.”

This approach fails for short-term extreme events, however, because of the dynamic nature of the atmosphere. The AI model itself also failed to produce good accuracy when emulating the events, but it was “basically free” in terms of the computing power required, and provided a good starting point for the physical models. “This coupling between the two ideas is really new,” Noyelle says.

National Science Foundation’s X-Labs initiative draws fire as ‘terrible idea’

The National Science Foundation (NSF) has announced a new initiative to convert new research into products and industries. The $1.5bn programme, called X-Labs, however, has been criticized by some in the US scientific community given that financing the plan could come via cutting the budgets of current NSF directorates by as much as 30%.

The NSF says that the first round of X-Labs funding will focus on two areas – one on quantum sensing and AI-driven computational imaging while the second area will be in the development of quantum information and computing.

The projects that are chosen will receive $1.5m over a nine-month period to carry out further work. The NSF will then select projects to move onto the next phase in which they will receive a further $10–50m per year for 2–3 years.

According to an NSF spokesperson, X-Labs are designed to “spur progress on platform technologies critical to US competitiveness”, supporting high-risk, high-impact research that requires “substantial resources beyond traditional mechanisms”. A key goal of the handful of labs, the spokesperson adds, is to “reduce the time it takes to go from insight to impact through this new investment framework”.

While critics don’t have an issue with the concept, there are concerns that it is coming at the expense of basic research. Former presidential science adviser and former NSF director Neal Lane told Physics World that X-Labs is a “terrible idea”.

“There’s been a push to get more practical and get NSF involved in ensuring that science it funds actually makes it to the marketplace,” adds Lane. “But suddenly it’s been taken to extremes by the Trump administration. They’re trying to cut as much basic research as possible and put it all into an experimental programme.”

‘Without precedent’

The X-Labs programme comes as the NSF – one of the main science funding agencies in the US – still lacks a director. The previous head, Sethuraman Panchanathan, resigned in April 2025 after receiving orders from the White house to cut the agency’s budget by more than 50% (US Congress later restored much of the funding).

The Trump administration has announced Jim O’Neill, a Silicon Valley investor, as the next director, but he has yet to go through the necessary Senate confirmation.

There is also no further movement on the new personnel that will make up the National Science Board, the body that acts at the NSF in a way similar to a company’s board. In April the US administration announced that it was terminating the positions of all 22-member NSB “effective immediately” without disclosing the reasons for the move.

“[That removal of the NSB] is quite without precedent,” adds Lane. “It seems to be an effort to simply eliminate the NSF.”

Browse articles by content type