Skip to main content

Syukuro Manabe, Klaus Hasselmann and Giorgio Parisi win the 2021 Nobel Prize for Physics

Syukuro Manabe, Klaus Hasselmann and Giorgio Parisi have won the 2021 Nobel Prize for Physics. The prize is awarded “for groundbreaking contributions to our understanding of complex physical systems”.

The prize is worth 10 million Swedish krona ($1.1 million). Manabe and Hasselmann will share one half of the prize and Parisi takes the other half. Because of the COVID-19 pandemic, there will not be an award ceremony in Stockholm in 2021 and laureates will be presented with their Nobel medals in their home countries.

Parisi bagged his half of the prize “for the discovery of the interplay of disorder and fluctuations in physical systems from atomic to planetary scales”.

Manabe and Hasselmann share their half of the prize equally “for the physical modelling of Earth’s climate, quantifying variability and reliably predicting global warming”.

Common thread

While Parisi’s research is very different from that of Manabe and Hasselmann, the common thread in this year’s award is the study of disorder and fluctuations in complex systems. When announcing the award, the secretary general of the Royal Swedish Academy of Sciences, Göran Hansson, explained: “disorder, fluctuations, and how disorder and fluctuations together, if you understand it properly, can give rise to something you can predict”.

During the early 1970s Parisi began to work on the theory of phase transitions within solids. He was particularly attracted by the application of field theory techniques  from high-energy physics to condensed-matter physics. Indeed, it is in this area that Parisi says he made his most important scientific contribution – helping form the physical theory behind spin glasses. These are magnets with “frustrated” interactions caused by disorder, which give them a rich set of behaviours.

Steven Thomson, who studies spin glasses at Dahlem Center for Complex Quantum Systems at the Free University of Berlin says, “Parisi has had a hand in inventing some of the most powerful and important theoretical machinery we have for understanding the behaviour of spin glasses and other complex systems. His work has had far-reaching impact on a huge variety of other fields, including particle physics, quantum field theory, neural network theory and other mathematical methods. The depth and breadth of these achievements is remarkable, and it’s fantastic to see all of this work recognized by the Nobel Committee.”

COVID-19 and conference registrations

Speaking to journalists minutes after the prize was announced, Parisi said that he was currently studying the dynamics of how COVID-19 spreads. Indeed, he has looked at a wide range of complex systems, and in 2007 he co-authored a letter in Nature Physics that identified a universal behaviour in the dynamics of how people register for conferences and showed how the final number of attendees could be predicted from early registrations.

When asked how he felt about winning, Parisi said, “I was very happy. I was not expecting the phone call but I knew there was non-negligible possibility [of winning].” Commenting on the work of Hasselmann and Manabe, he said, “It is urgent that we take strong decision on the climate. We are in a situation where we have positive feedback and accelerating increase of temperature. We have to act in a fast way and without delay.”

Pioneering climate research

Manabe is a pioneer in the development of  physical models of the Earth’s climate. In the 1960s he was the first to study how the balance of energy absorbed and emitted by the Earth interacts with the vertical transport of air masses. This research continues to play a crucial role in the development of climate models. In a ground-breaking paper published in 1967, Manabe and colleague Richard Wetherald showed that a doubling of the carbon dioxide content of the atmosphere would result in a 2 °C increase in the temperature of the lower atmosphere.

According to physics Nobel prize committee member John Wettlaufer, Manabe was “gobsmacked” when he heard he had won. “I’m just a climate scientist!” protested Manabe.

About 10 years later, Hasselmann created a model that connects climate and weather, showing that climate can be reliably modelled despite the fact that weather is changeable and chaotic. He showed, for example, that variation in weather on a timescale of days can influence the ocean on a timescale of years.

Hasselmann also developed techniques for identifying signals, or fingerprints, that natural phenomena and human activities create in climate data. These techniques have been used to prove that the increasing temperature of the atmosphere is the result of the human emission of carbon dioxide.

Essential understanding

James Hansen, director of climate science, awareness and solutions at Columbia University’s Earth Institute, says that the Nobel Foundation has made a good choice in awarding the prize to Manabe and Hasselman. “Their award is very well earned – their climate models provide the essential basis for our understanding of global warming,” he says.

Hansen, who pioneered work in the 1980s on climate sensitivity and feedback and who is a  former director of NASA’s Goddard Institute for Space Studies, adds that Wallace Broecker, who died in 2019, would have also been a possible recipient given his foundational work in the paleoclimate and oceanographic studies. “It’s too bad that he did not live to be part of the award,” adds Hansen.

Tim Palmer, Royal Society Research Professor of Climate Physics at the University of Oxford, says that this year’s prize is “well deserved”. “[Manabe] carried out pioneering studies of climate change using 3D climate models,” says Palmer. “They predicted the cooling of the stratosphere, the Arctic hotspot as well as global warming.”

Palmer adds that there are also strong links between the work of Hasselman and Parisi. “How do you represent small-scale events such as clouds and turbulent eddies in a climate model that cannot resolve these processes?” asks Palmer. “Parisi carried out fundamental work in this area and Hasselmann pioneered the use of stochastic models for studying the climate.”

Nobel careers

Born in Rome in 1948, Giorgio Parisi graduated in 1970 with a PhD in high-energy physics from La Sapienza University, during which he studied the Higgs mechanism under Nicola Cabibbo. Parisi then worked at the Laboratori Nazionali di Frascati, on the outskirts of Rome, on the theory of positron and electron collisions, which were being performed at the National Research Council’s Adone accelerator, also in Frascati. From 1981 Parisi worked at the University of Rome Tor Vergata, and in 1992 returned to La Sapienza.

Parisi was awarded the Dirac Medal in 1999 from the Abdus Salam International Centre for Theoretical Physics (ICTP) and the Max Planck medal in 2012 from the German Physical Society. In 2021 he was also awarded the Wolf Prize for his “ground-breaking discoveries in disordered systems, particle physics and statistical physics”.

Syukuro Manabe was born in 1931 in Shingu, Japan. He received his PhD at the University of Tokyo in 1958 before heading to the US to work at the US Weather Bureau until 1997. He then moved back to Japan to work at the Frontier Research System for Global Change, serving as director of the Global Warming Research Division. In 2002 he returned to the US to Princeton University, where he is currently a senior meteorologist.

In 1992 Manabe was the first recipient of the Blue Planet Prize of the Asahi Glass Foundation and in 2015 he bagged the Benjamin Franklin Medal of Franklin Institute. In 2018 he received the Crafoord Prize in Geosciences jointly with Susan Solomon “for fundamental contributions to understanding the role of atmospheric trace gases in Earth’s climate system”.

Klaus Hasselmann was born in Hamburg, Germany, in 1931. He was awarded his PhD in physics from the University of Göttingen in 1957 before moving to the Institute of Naval Architecture at the University of Hamburg, where he remained until 1961. He then moved to the US to work at the Scripps Institution of Oceanography before moving back in 1964 to the University of Hamburg. In 1975 he became a director of the Max-Planck-Institute of Meteorology in Hamburg before retiring in 1999.

Oxford Instruments logo

Physics World‘s Nobel prize coverage is supported by Oxford Instruments Nanoscience, a leading supplier of research tools for the development of quantum technologies, advanced materials and nanoscale devices. Visit nanoscience.oxinst.com to find out more.

Introducing RadCalc 7.2 and its EPID Module

Want to learn more on this subject?

In this webinar we will be covering the new features released in the anticipated Version 7.2 of RadCalc.

These include new customizable features for intelligent automation with RadCalcAIR that allows you to customize which DICOM tags to trigger actions, new layout customizations for a cleaner workflow on the modules that you work with the most, and we will be covering the two configurations that our new EPID module is designed for: In-Air deliveries and In-Vivo deliveries.

For true 3D IMRT and VMAT verification and reconstruction with absolute dose using existing EPID on all conventional linacs.

RadCalc’s EPID module uses the acquired measurements from static and dynamic delivers and works backwards to compute the dose to the patient CT dataset. Providing a very accurate evaluation of the intended versus delivered dose.

Want to learn more on this subject?

Carlos Bohorquez, MS, DABR is the product manager for RadCalc at LifeLine Software, Inc., a part of the LAP Group. An experienced board-certified clinical physicist with a proven history of working in the clinic and medical device industry, Carlos’ passion for clinical-quality assurance is demonstrated in the research and development of RadCalc into the future.

A review of AAPM task groups 218 and 219

Want to learn more on this subject?

Pending CAMPEP CE credit.

The American Association of Physicists in Medicine (AAPM) has recently released new recommendations that dive deep into patient-specific quality assurance (QA), with task group (TG) 219 published in 2021 covering independent calculation-based dose/monitor-unit verification for intensity modulated radiation therapy (IMRT), which stems from TG 218 published in 2018 on tolerance limits and methodologies for IMRT measurement-based verification QA.

Benefits of attending:

  • Understand the limitations for each of the point, planar and volumetric secondary check processes and algorithms (TG 219).
  • Understand the recommendations from TG 218 with regard to IMRT QA and how pass/fail rates correspond to the delivery of the plans.

Want to learn more on this subject?

Vimal Desai, PhD, clinical instructor, Thomas Jefferson University Hospitals. Vimal’s introduction to the field of medical physics was through patient-specific QA projects utilizing log file analysis. This lead to his expanded interest in radiation dosimetry, quality assurance, and radiation measurements/simulations. His doctoral thesis was focused on linear accelerators and investigated calibrating radiation detectors closer to clinical delivery conditions. This has led to various collaborators investigating the efficacy of plan complexity metrics and maintains interests and ongoing projects in various topics related to patient-specific QA.

Carlos Bohorquez, MS, DABR, is the product manager for RadCalc at LifeLine Software Inc, part of the LAP Group. An experienced board-certified clinical physicist with a proven history of working in the clinic and medical device industry, Carlos’s passion for clinical quality assurance is demonstrated in the research and development of RadCalc into the future.

Personalized brain stimulation could treat untreatable depression

Depression is a common disorder, affecting an estimated 5% of adults worldwide, and a leading cause of disability. Although therapy and medications are effective in most patients, there’s a substantial minority who remain resistant to all available treatments.

“It was these patients who really drove us to do this research,” explains Katherine Scangos from the University of California, San Francisco (UCSF). Scangos and colleagues are developing a personalized treatment for severe depression based on deep-brain stimulation (DBS), in which implanted electrodes deliver electrical impulses to targeted structures in the brain.

The researchers have now demonstrated the feasibility of their new approach in a 36-year-old woman with severe, treatment-resistant depression, reporting their findings in Nature Medicine. The patient, Sarah, had previously tried all possible treatment options, including multiple antidepressant combinations and electroconvulsive therapy, with no success at lifting her depression, which restricted her daily life and led to unremitting suicidal impulses.

“When I first received the stimulation, I felt the most intensely joyous sensation and my depression was a distant nightmare for a moment,” said Sarah, speaking at a press briefing. “Months later, when the researchers implanted the chronic device and turned it on for the first time, my life took an immediate upward turn. Now a year into therapy, the device has kept my depression at bay, allowing me to return to my best self and rebuild a life worth living.”

Therapy on demand

The idea of personalized DBS arose some seven years ago in the laboratory of UCSF neurosurgeon Edward Chang, when he observed that stimulating brain areas in epilepsy patients also alleviated emotional symptoms such as anxiety and depression.

While DBS has shown some promise for treating depression, previous clinical trials have revealed highly variable responses between subjects. These inconsistent findings may be due to the use of open-loop DBS, which delivers constant electrical stimulation to a single brain structure. Recent work, however, showed that the effects of DBS are dependent on the emotional state of the patient. In addition, different neural circuits underlie different subsets of depression symptoms and may vary between different people.

“So we set out to see whether we could develop a personalized DBS strategy,” Scangos explains.

Sarah’s treatment was a two-stage process. To identify her unique depression circuits, the researchers first placed 10 temporary electrodes into various regions of her brain. For 10 days, they continuously recorded neural activity while Sarah rated her symptom severity. This brain mapping approach enabled the team to identify a personalized brain activity biomarker: the finding that high brain activity in the amygdala was correlated with most severe depressive symptoms.

The researchers also used the temporary electrodes to deliver small stimulation pulses to each brain region. They identified an area of the brain, the ventral capsule/ventral striatum (VC/VS), in which electrical stimulation consistently eliminated feelings of depression.

Armed with this information, the researchers then implanted a commercial DBS device (the NeuroPace RNS System) into Sarah’s brain. One of the device’s electrode leads was placed in the amygdala and the other in the VC/VS. They found that a 6 s stimulation at 1 mA was clinically effective and that Sarah could not feel the electrical stimulation at this level.

In addition to identifying a personalized neural biomarker, the other key breakthrough in this work was the use of the biomarker to perform closed-loop therapy, in which stimulation is only delivered when needed. To achieve this, the team programmed the implanted device to continually monitor Sarah’s amygdala for abnormal activity. When this activity was detected (representing a state of severe depression), it automatically triggered a 6 s stimulation pulse to the VC/VS.

The proof-of-concept study proved a notable success. “When we turned this treatment on, our patient’s depression symptoms dissolved and in a remarkably short time, she went into remission,” said Scangos.

Over two months, the device delivered an average of 468 stimulation pulses throughout the day, with few stimulations at night. The team capped the number of stimulations at 300 per day to minimize sleep disturbance from evening therapy.

Ongoing research

Looking at the longer-term implications of this new treatment, Scangos notes that it is too early to tell how long the device will need to remain in a patient, or whether it’s possible that it may somehow help the brain rewire its circuity. However, one advantage of the closed-loop approach is that it does not require continuous stimulation, providing a battery life of over 10 years and enabling the implanted device to provide long-term stimulation if needed.

The researchers also point out that treatments such as cognitive behavioural therapy are almost impossible when a patient is suffering severe depression. If the implanted device can treat the most extreme symptoms, then patients may be more able to employ such therapies. Sarah notes that once the DBS treatment had begun: “I was finally able to use the therapy skills I’d learned and never been able to apply.”

The researchers emphasize that this work is at a very early stage. They cannot determine whether the particular neural biomarker and depression circuit identified in this single-participant study would be present in all individuals. They have now enrolled two other patients in the trial and hope to add nine more.

“These results provide hope that much needed personalized biomarker-based treatment for psychiatric disorders is possible,” says Scangos.

Carbon fibres have directional electrical properties

The electrical properties of a carbon fibre are very different when measured across its width or along its length, according to a new study by Satoshi Matsuo and Nancy Sottos at the University of Illinois at Urbana-Champaign in the US. Using a technique designed to probe the electrical resistivity of 2D materials, the duo has shown for the first time that fibres are significantly less conductive in the transverse direction.

When carbon fibres are woven into interlocking sheets, the composite materials they produce can display a unique variety of electrical properties, with applications including electromagnetic shielding; sensing for structural damage in buildings; and protection against lightning strikes. To tailor these composites for specific uses, it is important to have accurate models of their electrical behaviours. However, the complex structure of these materials makes this extremely challenging.

Carbon fibres measure just a few microns in diameter and comprise bundles of smaller carbon filaments, which are themselves composed of crumpled sheets of carbon atoms. Within these sheets, strong covalent bonds between the atoms are aligned parallel to the axis of the fibre. On longer length scales, filaments are bonded together by far weaker van der Waal’s forces.

Difficult to characterize

Such hierarchical structures are very difficult to characterize reliably so Matsuo and Sottos turned to the “van der Pauw” method. This technique is commonly used to measure the resistivities of 2D materials and involves placing two separate pairs of electrodes around the perimeter of a sample. In their experiment, the duo connected the electrodes to 2D slices of carbon fibre, which they cut using a focused ion beam.

Across the diameter of the fibre, the duo measured an electrical resistivity roughly six times greater than that along its length. This means that the material is a poorer electrical conductor in the transverse direction than it is in the longitudinal direction – something that can have important implications for how electrical currents move through a carbon fibre component.

Their results are a step forward in efforts to better understand the electrical properties of carbon fibres – but researchers still have much to learn about the characteristics of far more complex composite materials.

Matsuo and Sottos are now making further progress towards this goal, by measuring the electrical contact resistance between two separate carbon fibres. This value is directly connected with the transverse resistivities of the fibres; as well as the area of contact between them, and the angle at which they cross each other. In the future, the researchers also hope to assess how electrical properties vary under different environmental conditions, such as temperature. In addition, they hope to carry out similar experiments on fibres made from other conductive materials, such as polymers or metals.

The research is described in Journal of Applied Physics.

High-resolution independent VMAT/IMRT patient QA: Clinical implementation and results

Want to learn more on this subject?

This presentation has been accredited 1 MPCEC hour by CAMPEP and submitted for approval by EBAMP as a CPD event for Medical Physicists at EQF Level 7.

Learn about the clinical implementation and routine use of the MatriXX Resolution detector array. Discover the latest features in the myQA software that allow for accurate measurement of volumetric modulated plans, as well as static IMRT, or any dose delivery method used by your clinic. Integration of the wireless Gantry Sensor+ with the device and angular correction will also be discussed.

The guest speaker Dr Raj Mitra from Ochsner Health Systems, Louisiana, USA, will walk participants through the initial setup and calibration of the device, testing and validation and finally, present clinical test results for various IMRT/VMAT cases. The presentation will conclude with expert answers to your live questions.

Benefits of attending your webinar include:

  • Learn about product improvements with the new MatriXX device
  • Discover new features of myQA software
  • Get introduced to the new wireless Gantry Angle Sensor +
  • Gain knowledge about the Look Up Table (LUT) integration for different beam energies (standards and FFF beams)
  • Learn about LUT and Angular correction validation
  • Understand calibration for different energy modalities
  • Discover clinical use and benefits of MatriXX Resolution

Want to learn more on this subject?

Dr Raj Mitra is the lead medical physicist for Ochsner Health System, New Orleans, Louisiana, USA. He has more than 25 years of experience in clinical radiation oncology physics and is board certified by the American Board of Radiology and American Board of Medical Physics. In addition to his work at Ochsner Health System, Raj Mitra has also had numerous research articles published.

Distillation method strengthens quantum entanglement in a single pair of photons

Quantum entanglement is a valuable resource, enabling spy-proof communications and allowing quantum algorithms to be faster than classical ones. But like other quantum phenomena, entanglement is also extremely delicate and sensitive to environmental noise. Because many quantum communication protocols require high levels of entanglement to operate properly, preserving that entanglement is crucial.

There is a solution, but it comes at a hefty price. By sacrificing some poorly entangled quantum objects, physicists can create a better-entangled pair out of the objects that remain – a little like reducing a weak broth into a hearty soup by boiling off the excess water. This method of increasing entanglement in quantum objects is known as entanglement distillation and was first described theoretically in the late 1990s. Since then, it has been demonstrated in all kinds of quantum systems, from superconducting circuits to photons. Now, however, researchers at the Institute of Quantum Optics and Quantum Information (IQOQI) at Vienna have demonstrated entanglement distillation using only a single pair of photons. By using the various quantum properties embedded in this photon pair, these researchers can generate and distribute entanglement more quickly, more easily and with greater protection than ever before.

Boil down your qubits

Entanglement allows pairs of quantum objects to communicate with each other, regardless of how separated they are in space. This property makes entangled pairs of photons extremely useful, as it can allow two parties to whisper secrets to one another, knowing that no one else can eavesdrop without disturbing their delicate quantum system. However, entanglement can be degraded by the environment over time, making it harder and harder for the entangled photons to “hear” each other clearly.

Cartoon showing a distillation flask partly full of weakly entangled particles with a "droplet" of a highly-entangled particle being distilled out

Entanglement distillation reverses this noise, reviving the entanglement and giving the pair of photons a new life via a protocol involving quantum logic operations known as controlled-not (CNOT) gates. Traditionally, this protocol is pretty wasteful: each distillation step sacrifices a good pair of photons, and even worse, there is no fail-proof way to guarantee the operation will succeed. The IQOQI researchers, however, found a better way. “You can perform these controlled-not gates not just between two photons, but between two properties of the same photons,” explains Sebastian Ecker, a PhD student at the IQOQI and first author of a report published in Physical Review Letters.

Leveraging degrees of freedom

Photons have many uniquely quantum properties, such as their polarization state, energy level and spatial mode. Collectively, physicists refer to these properties as “degrees of freedom” and all of them have been used independently to demonstrate entanglement. However, the IQOQI study is the first experiment to demonstrate entanglement distillation with different degrees of freedom.

In their experiment, the researchers generate entangled photons pairs using a nonlinear crystal, then send each photon in the pair to a different optical table. Each table holds a labyrinth of optical devices that perform the entanglement distillation step and interpret the results. At the core of this labyrinth is an unassuming optical device called a polarizing beamsplitter. This small glass cube transforms the state of the photons, changing the quantum state of the photon’s polarization only if certain conditions are met with the photon’s energy-time domain. That action exactly describes a CNOT logic gate, one of the basic logical building blocks of quantum computing. After this distillation process is complete, the researchers measure the properties of the photon pair and determine how much entanglement was recovered.

The researchers also verified that their distillation process is robust by intentionally inserting noise into the environment. Because that noise is very carefully controlled, they can quantify how well their new procedure works in noisy environments, showing that their method for entanglement distillation is faster than traditional two-pair methods by a factor of 100 million. “In our experience, these degrees of freedom are robust enough to revive entanglement after passing through long optical fibres or free-space links,” Ecker says.

A perfectly entangled world

Because polarization and energy-time are both frequently used in other aspects of quantum communication, the researchers are confident that their scheme will soon find many other applications. After considering how this method might improve on previously extracted quantum keys, their sights are now set even higher. “Wouldn’t it be nice if you could use the high dimensional entanglement to make your qubit entanglement noiseless? This would be really cool,” Ecker says.

Life beyond the Nobel: Andrea Ghez eyes up new research directions

Physicists around the world are gearing up for tomorrow’s big reveal of who has won the 2021 Nobel Prize for Physics. Part of the prize’s appeal is that no-one – apart from the members of the Nobel Committee for Physics – currently knows who this year’s winners will be. Even the Royal Swedish Academy of Sciences only grants final approval on the very morning the prize is announced, which sounds seat-of-the-pants, but that’s the way it is.

Once the winners are declared, however, their lives will change forever. To find out what impact the prize can have, I caught up with US astrophysicist Andrea Ghez, who shared one half of last year’s award with Reinhard Genzel for their work discovering a huge black hole lurking in the middle of the Milky Way (Roger Penrose bagged the other half for his theoretical studies of black holes and general relativity).

I have to work harder to maintain a balance between being a more public figure and carrying out research.

Angela Ghez

Ghez, 56, is still as active as ever, but admits her working life has certainly changed over the last year. “I’m receiving a lot more requests from all over the place,” she says. “So I have to work harder to maintain a balance between being a more public figure and carrying out research, which continues to be my first priority.”

She was, for example, co-author of a paper earlier this year describing plans for an infrared spectrograph to be used on the upcoming Thirty Meter Telescope. Ghez also recently gave a keynote address to students graduating from the International Centre for Theoretical Physics in Trieste, Italy. She even revealed she still gives introductory lectures to undergraduate students to “shape the next generation’s ideas about who can be a scientist”, knowing how vital it is for Nobel laureates to act as role models.

However, Ghez has had plenty of opportunity to do things she wouldn’t have had the chance to tackle without a Nobel under her belt. “The most rewarding experience that I would not otherwise have had was speaking to the Hawaii county council at a meeting when they gave me a lovely signed certificate acknowledging my Nobel-prize work that was carried out in Hawaii,” she reveals.

“I was deeply honoured and thrilled to express my gratitude for having had the opportunity to work in Hawaii at Keck Observatory, the largest telescope in the world.”

But are there are any new research directions Ghez wants to go into that would have not been possible before the prize?

“Yes!” she insists. “I’m really excited to take on a more ambitious and riskier research agenda that would not have been possible otherwise to explore how gravity works near supermassive black holes and how these exotic, but poorly understood, objects regulate the formation and evolution of galaxies.”

Oxford Instruments logo

Physics World‘s Nobel prize coverage is supported by Oxford Instruments Nanoscience, a leading supplier of research tools for the development of quantum technologies, advanced materials and nanoscale devices. Visit nanoscience.oxinst.com to find out more.

Green jobs for physics graduates: decarbonizing energy sources

Ann Davies, chief operations officer, Lightsource BP

Taking advantage of new technologies will be essential if the world is to achieve its net-zero goals. Renewable energy sources are perhaps the most obvious of these, and the transition to them is already under way.

“I don’t think there’s ever been a more exciting time to join the energy sector,” says Ann Davies, who is chief operations officer at Lightsource BP, a firm that has been developing solar projects since 2010, and now employs nearly 600 people. She took up her current job after studying physics at the University of Oxford, UK, and gaining experience in several engineering roles after graduating.

Ann Davies

Among the main reasons to choose a career in renewables, Davies cites the growth that the sector is experiencing due to the increased focus on climate change and demand for clean energy. “Along with the cost of solar and wind dropping significantly, that makes renewables a really sound economic proposition, which means that investment is at a record high,” she explains. “From a graduate perspective, that means there will be more and more opportunities as your career develops.”

Davies’ work involves leading teams of scientists and engineers working on the planning, implementation and operation of solar projects around the world. She notes that there are lots of problems for scientists to solve, from loading up the grid with renewables to managing intermittency issues.

When recruiting new scientists and engineers, Davies emphasizes the importance of technical grounding. “Physics teaches you how to break down complex problems into simple parts, and we need those skillsets,” she says. In the hiring process, she also values interpersonal skills. “It’s not one person who is going to solve this – it takes a team.”

Davies advises graduates who want to join these efforts to read widely and to connect with people in the industry. Since there are so many different areas of sustainability that physicists can contribute to, she believes it is important to find out what makes you tick as an individual, and how you want to apply your skills. For herself, she finds working in the energy sector rewarding, because it is so universal. “Energy touches everyone,” she says, “so being part of providing it in a responsible way really gets me going when I get up in the morning.”

Hari Chohan, nuclear radiation analyst, UK Atomic Energy Authority

While solar and wind energy might be the first clean-energy technologies that spring to mind, they are not the only low-carbon options. Another relevant area in which many physicists work is nuclear energy, both fission and fusion.

Hari Chohan

“I consider all nuclear energy to be green energy,” says Hari Chohan, who is a nuclear fusion radiation analyst at the UK Atomic Energy Authority (UKAEA). “My granddad was an engineer and worked on nuclear projects,” he adds, “so I’ve always been pro-nuclear energy.” Chohan developed a stronger interest in nuclear fusion while writing an article about it for one module of his physics degree at Imperial College London.

As a result of these two influences, Chohan decided to do a Master’s degree in physics and nuclear technology at the University of Birmingham, UK. He then did a nine-month internship at Fusion for Energy in Barcelona, a body that co-ordinates the EU’s contribution to ITER, which is the largest experimental fusion reactor under construction in the world.

During this internship, Chohan’s main area of work was neutronics, also known as neutron transport, which is the study of the motion of neutrons and how they interact with materials. This is not only important for developing appropriate shielding but also estimating the lifetimes of components. “Neutrons are produced in both fission and fusion,” he explains, “but they have a lot more energy in the case of fusion, so we need to ensure the materials and components we design can withstand them.”

Chohan continues to work on fusion neutronics in his current role at UKAEA, which involves programming, running simulations, analysing results and writing them up in reports. “As well as working on major international projects, we have a couple of fusion test machines at UKAEA, the Mega Amp Spherical Tokamak Upgrade (MAST-U) and EUROfusion’s Joint European Torus, and we’re designing a prototype fusion energy plant, the Spherical Tokamak for Energy Production, at the moment, so there’s a lot of active research going on,” he says.

Chohan believes accurate public communication about nuclear energy is essential. “In fusion, unlike fission, there will be no high-level nuclear waste generated by the reaction itself, but there will still be a lower level of nuclear waste generated by the interaction of the neutrons with the reactor components,” he says. “All technologies have advantages and disadvantages. To move away from fossil fuels, we will need a mixture of different energy sources, but we can’t do it without nuclear. And the prospect of nuclear fusion is truly exciting.”

Rhiann Canavan, scientific project manager, Crossfield Fusion

While fusion is a long-term goal with huge clean-energy potential once it’s achieved, we don’t have to wait until then to get something positive from the research going into it. There are many by-products along the way that can be useful more immediately, as Rhiann Canavan, who works at UK-based start-up Crossfield Fusion, has discovered.

Rhiann Canavan

While studying at the University of Birmingham, UK, Canavan was inspired by a nuclear-physics professor to go into nuclear power. “I knew I wanted to go into a job where I was delivering something useful,” she says, “and nuclear power seemed like it had the potential to change the world.”

After graduating with an MSci in physics, Canavan studied for a PhD in experimental nuclear physics with the University of Surrey, UK, and the National Physical Laboratory. Her project focused on understanding fast neutron-induced fission reactions, which can be done to make nuclear waste decay faster.

After finishing her PhD, Canavan did a summer placement with Crossfield Fusion, which she found through the South East Physics network (SEPnet) – an association of nine university physics departments that supports students in south-east England. “When I read the mission of the company, I really wanted to get involved,” she says. “The end goal is fusion, but there are also short- and mid-term goals, such as using the technology to produce radioisotopes for medical scans.”

After completing her internship, Canavan joined the company in a permanent role as scientific project manager. Crossfield Fusion is a start-up with just five employees, so her tasks vary widely. She began by helping to build the research reactor, and she now plans and carries out experiments with it. “Some days are lab days when everything has to be spot-cleaned because we’re installing components,” she says. “Other days I’m analysing data, computer programming or group brainstorming what to try next.”

Canavan says she feels a lot of ownership of the work, having seen the progress from the very early days. “Your heart is really in it and you want it to succeed,” she says. Since deuterium – a key ingredient in fusion reactions – is highly abundant and can be extracted from any type of water, Canavan also points out how much more environmentally friendly it would be to fuel a fusion reactor than to burn fossil fuels. “Instead of digging up coal,” she says, “we could just use sea water.”

‘CatGym’ algorithm predicts better catalysts

Designing efficient new catalysts is no easy task. In catalysts that contain more than one element, for example, researchers not only need to take into account all the possible elemental combinations, they must also add a number of other variables, such as particle size, shape and surface structure, as well as the degree of alloying or phase segregation. This ultimately leads to an overwhelmingly large number of potential candidates.

To address this challenge, scientists employ computational design techniques that focus on screening material components and alloy composition to optimize a catalyst’s activity for a given reaction and so reduce the overall number of prospective structures that would need to be tested and then developed. Such techniques require combinatorial approaches coupled with theory calculations, which can both be time-consuming and complex.

The best surface atom configurations

A team led by Zachary Ulissi of Carnegie Mellon University has now taken a different approach by developing a deep reinforcement learning (DRL) programme, dubbed CatGym, that iteratively changes the positions of atoms on the surface of a catalyst to find the best configurations from a given starting configuration.

The researchers showcased their technique by predicting the surface reconstruction pathways of a ternary Ni3Pd3Au2(111) alloy catalyst. Their results show that the DRL programme can not only be used to explore more diverse surface compositions than conventional methods, but that it can also generate new pathways based on how energetically favourable they are.

The team also demonstrated that the kinetic pathways that lead to a stable surface composition (with a low minimum energy surface composition) and the associated transition state predicted by the DRL programme agree well with the minimum energy path predicted by traditional “nudged elastic band” calculations done “by hand”.

A lot of human input

There has been much excitement in recent years for when it comes to using machine learning methods to accelerate catalysis simulations, says Ulissi. Such an approach reduces the computational cost of each step in the simulation, but the downside is that it requires a lot of human input to run the calculation. This is because scientists need to define what structure is used from the outset, what mechanisms should be investigated and if there is a better path to take to go from reaction A to reaction B. All these questions require a trained expert many days or weeks to answer.

“The new work is very exciting for us because it proposes using DRL methods to tackle these strategic questions,” Ulissi tells Physics World. “With our system, we can let the computer autonomously explore a number of different possible pathways.”

Representation and action space

DRL requires three things, he explains. “The first is a representation – that is, how do we show an atomic structure of a catalyst to the computer in a way that it understands? In our system we use a common representation from the literature. The second is an “action space”: what are we going to let the computer do? In our approach, it can move an atom, find an energy minimum, find a transition state or run a short dynamic simulation. Finally, how do we decide what action to take next? In our case, we tried many DRL schemes to answer this question.”

“One aspect that made this project really interesting was that the final goal was not clear,” explains Ulissi. “In a video game, for example, it is obvious what you want you want the DRL to do – maximize the final score. We thus spent a lot of time defining and identifying the goal the DRL would work well with.”

Double-checking results

Ulissi says he previously studied catalyst surface reconstruction mechanisms by hand, which can be very tedious. “A tool to automate and accelerate this process not only allows us to ask much more interesting questions, it can also be used to double-check the results obtained by human experts.”

The researchers, who report their work in Machine Learning: Science and Technology, are now using the method they have developed to predict how stable hypothetical catalyst surfaces are. “We also hope to apply our approach to better understand the mechanisms at play on these surfaces,” adds Ulissi. “Doing this will help us think creatively about what might happen to a catalyst during real-world reactions.”

It will not all be plain sailing, however, he admits. One of the major limitations of the current technique is that, like most DRL applications, it requires a lot of data input and training episodes. “Accurate simulations are extremely demanding computationally,” he explains, “and the simulations we performed in our work are fast but rather coarse approximations.” The researchers are trying to solve this problem by also using machine learning models to make them not only faster but also more accurate.

Copyright © 2026 by IOP Publishing Ltd and individual contributors