Skip to main content

Reset your password

Please enter the e-mail address you used to register to reset your password

Note: The verification e-mail to change your password should arrive immediately. However, in some cases it takes longer. Don't forget to check your spam folder.

If you haven't received the e-mail in 24 hours, please contact customerservices@ioppublishing.org

Registration complete

Thank you for registering with Physics World
If you'd like to change your details at any time, please visit My account

Close
Home » Page 369

W boson mass measurement surprises physicists

The most precise measurement to date of the mass of the W boson has yielded a result seven standard deviations away from that predicted by the Standard Model of particle physics. The stunning result was obtained by a painstaking analysis of data taken at the Fermilab Tevatron collider in the US before it closed in 2011. The particle physics community must now study the results carefully to work out whether it is an incredible statistical fluke, an unknown experimental error, a flaw in the Standard Model or a genuine indication of physics beyond the Standard Model.

The W boson is one of the most intriguing particles described by the Standard Model. Together with the neutral Z boson, the charged W boson mediates the weak interaction, which causes beta decay and several other important processes in particle physics. The weak interaction has long intrigued scientists searching for physics beyond the Standard Model, partly because it is the only force known to violate charge-parity symmetry. If particles in a process are exchanged for their antiparticles and the spatial co-ordinates are inverted, the weak interaction in this mirror image process are not always identical. This puzzle is not explained in the Standard Model.

The weak interaction is also unique in that, whereas the photons that mediate electromagnetism and the gluons that mediate the strong interaction are massless, the W and Z bosons have mass. The Standard Model does not predict the absolute mass of the W boson directly, but it does predict the ratio between the masses of the W and Z bosons. To make absolute predictions, however, physicists needed to know the mass of the Higgs boson, which was not known when Tevatron switched off in 2011. The very next year, however, CERN’s new Large Hadron Collider famously discovered the missing boson at a mass of 125.35 GeV.

Encouraging measurements

“The Z boson mass was measured already much more precisely than the W,” explains Ashutosh Kotwal of Duke University in the US, who is part of the Collider Detector at Fermilab (CDF) collaboration, which was one of two experiments at the Tevatron. “That’s one of the things that’s been driving us: if you know the Z boson mass precisely, that allows you to calculate the W boson mass precisely, because it’s the connection that the theory tells you.” The Higgs boson mass allowed physicists to infer that the mass of the W boson would be 80,357±6 MeV. This encouraged physicists to work to either confirm or refute this prediction.

Measurements of the W boson’s mass date back to 1983, and many of the most recent have been based on data from CDF and the Tevatron’s other detector D0. Until now, the most precise value – from the ATLAS collaboration on the LHC – stood at 80,370±19 MeV – just within agreement with the theoretical prediction.

According to Kotwal, the Tevatron offers two advantages over the LHC for measuring the W boson mass. First, its lower beam luminosity means that events of interest are less likely to be obscured by other events at the detector. More fundamentally, whereas the LHC collides protons with other protons, Tevatron collided protons with antiprotons: this makes quark-antiquark annihilation a simpler process with fewer errors to consider.

Four million collisions

The CDF collaboration’s new analysis is based on four million collisions in which a W boson was created by quark-antiquark annihilation, before decaying to produce a lepton or muon and a neutrino. The researchers had previously published results based on a subsets of the data and these were statistically consistent with the predicted mass value. For the new work, they incorporated a dataset four times larger, as well as improved calibration of many experimental parameters. Their result is a stunning 80,433±9MeV, seven standard deviations away from the theoretical prediction.

“At this point, there really isn’t much we can think to improve further without input from the outside,” says Kotwal, “We really need to share what we have done with the community, and it’s their feedback that will help us think more.”

If the result does hold up to scrutiny and if similar results appear in experiments at other detectors, Kotwal says that “from the theory perspective it is way out of our field of expertise”. However, he notes that theorists are forced to make approximations and assumptions to make calculations tractable. “It has happened historically every once in a while that something that was left out of calculations because, from experience, people assumed it would be a small effect turned out when actually calculated not to be such a small effect after all.”

“It’s both significant and surprising, and people will scratch their chins for some years to come, I suspect,” says Paul Grannis of Stony Brook University, who is part of the D0 and ATLAS Collaborations. “I know the researchers; I know how careful they are; I know how talented they are, so I am inclined to believe that they have been as thorough as they possibly can be. However, as their result is so far away from the Standard Model prediction, everybody is going to be looking at their result as hard as they possibly can.” He says that, compared to the theoretical uncertainties within the Standard Model, the uncertainties measured are “in the direction you would get from most models of beyond the Standard Model physics”. “However, most of those departures from the Standard Model have been exhaustively searched for unsuccessfully at the LHC,” he says. “So I suspect there is no cheap way to find an explanation in terms of beyond the Standard Model physics. I haven’t heard any exhaustive study of that yet and people will surely do that, but it’s going to be be tough.”

The research is described in Science.

Kirigami-inspired electrocardiogram sensors can be worn all day

The Japanese art form of kirigami has inspired a new heart sensor that can monitor electrical signals from the surface of the skin while being worn for long periods of time. Borrowing from the principles of kirigami, Kuniharu Takei and colleagues at Osaka Prefecture University cut a pattern of holes in an electrode-coated PET film, which allowed the material to conform to the skin’s surface more easily.

The latest advances in wearable technology include devices that can monitor aspects of a wearer’s health in real time. In the future, wearable technologies could identify diseases in their early stages or even predict the likelihood of future medical conditions.

Electrocardiography (ECG) is an important technique for monitoring the health of the heart. The heartbeat causes subtle electrical changes to the skin’s surface, and these changes are captured by placing pairs of electrodes on specific parts of the body. Information provided by an EKG can be used to diagnose heart conditions such as arrhythmias.

Comfort and stability

Most ECG monitoring is done in a matter of minutes, with most of the time devoted to placing and removing the electrodes. However, being able to monitor the heart continuously could benefit some people. An important challenge of doing continuous monitoring is creating electrodes that can remain on the skin for long periods of time without causing discomfort to the wearer. Such devices must be able to conform to the intricate and changing contours of the body, while remaining comfortable and stable, even during exercise.

Previous attempts at developing electrodes have focussed on making electrodes as thin and flexible as possible. So far, however, there has been less focus on making sensors smaller and more breathable. Reducing sensor size is a particular challenge because smaller devices tend to have higher levels of noise.

To address size and breathability, Takei’s team turned to kirigami, an art form whereby paper is cut before being folded to create 3D objects. To create their sensor, they first printed sheets of silver electrode on both sides of a plastic (PET) film. A second film of PET is then deposited on one of the silver sides. The remaining silver side of the device contacts the skin, while the PET protects the device.

Millimetre-sized holes

A laser was then used to cut a kirigami pattern of millimetre-sized holes and slits into this multi-layered structure. This allowed the relatively stiff PET material to stretch and bend far more easily, while also allowing the wearer’s sweat to readily pass through. To perform ECG’s, the electrodes were connected to a battery-powered sensor, which uses Bluetooth to transmit detected signals to an app on the wearer’s smartphone.

Takei and colleagues found that an optimal balance between comfort and noise could be reached through an electrode size of roughly 200 mm2 with the sensors placed 1.5 cm apart. This resulted in accurate, reliable heart data when the device was tested on people carrying out a range of everyday activities, including sitting and walking.

Through further improvements, the researchers now hope to integrate more sensors into their system. This could allow it collect multiple types of data from the skin surface and approve its ability to help doctors make non-invasive diagnoses.

The research is described in Applied Physics Reviews.

Laser plasma accelerators unlock potential for radiobiology studies with protons

Laser plasma accelerators (LPAs) propel particles to high energies over short distances using intense, ultrashort pulses of laser light. These accelerators can supply high-quality particle beams for radiobiological studies that will help scientists better understand how radiation leads to DNA damage, and ultimately, help optimize particle-based cancer treatments.

Proton beams produced by laser plasma acceleration have a broad energy distribution, an exponentially decaying energy spectrum, and are bunched into particle pulses containing up to one trillion protons. This unique combination of beam properties makes proton LPAs suitable for ultrahigh-dose-rate radiobiology studies. Before researchers can conduct these studies, though, they must demonstrate that they can control LPA beam parameters.

To that end, Florian Kroll, a physicist at the Institute of Radiation Physics of Helmholtz-Zentrum Dresden-Rossendorf, his colleagues, and experts from OncoRay Dresden spent the last few years testing a sophisticated proton LPA platform. The platform has reliably displayed accelerator stability for two years, delivers model-compliant dose distributions, and in a pilot study in mice, demonstrated precise dose delivery and dosimetry.

A new laser plasma accelerator platform

Though strides in laser plasma acceleration have been made over the years, one of its primary challenges has persisted: laser plasma accelerators, and the particles that they accelerate, are difficult to stabilize and control.

The LPA method that Kroll and his colleagues implemented to improve this reliability relies on a high-power laser, tightly focused onto a thin (220 nm) plastic target.

“The intensity of the laser is so enormous that, when it impinges on the target, it immediately ionizes the target material, turning it into a plasma,” explains Kroll, first author on the resulting Nature Physics study.

Stripped from their atomic cores, the target’s electrons are pushed through the plasma by the laser. Some of the electrons, unable to escape the now positively charged target, form a “sheath” on the back side of the target and create a quasi-static electric field that “pulls” on the target ions. This field accelerates the ions into the mega-electronvolt (MeV) range.

Dose delivery
Dose delivery: The Draco PW laser impinges on the target and drives a broadband LPA proton source. The protons are transported through an energy selection aperture (ESA), and traverse beam monitors and dosimeters installed downstream, before delivering dose to the planning target volume (PTV) in the mouse. (Courtesy: CC BY 4.0/Nat. Phys. 10.1038/s41567-022-01520-3)

The researchers optimized the shape of the resulting proton beam with several techniques, including a pulsed two-solenoid beamline that contains apertures and scatterers. To fine tune proton acceleration, they tailored the laser pulse shape temporally (the findings of a previous study demonstrated that the laser pulse’s “temporal shape” influences proton acceleration performance).

The LPA platform generates protons with a cut-off energy of up to 70 MeV and demonstrates stable accelerator performance, as well as daily delivery of prescribed volumetric dose distributions with percent-level margins.

In their accompanying radiobiological pilot experiments, the researchers irradiated human tumours on mouse ears using multi-shot LPA proton irradiation, clinical proton source irradiation, or standard 200 kV X-ray irradiation, comparing dosimetry and tumour control measures (the study design also included control and sham groups).

Although radiation-induced effects on tumour growth were observed, the focus of the pilot study was to demonstrate the feasibility of animal studies and to test the limits of dose delivery.

“We don’t want to speculate about the clinical applicability of laser-driven proton beams,” says Kroll. “In the early days of laser acceleration, many claims with respect to revolutionary, compact and cheap laser-driven therapy machines were made. In the end, everything turned out to be more complex than expected. Nevertheless, LPA machines have always been and will always be an interesting complementary accelerator technique to cyclotrons, synchrotrons and more.”

The future of LPA: FLASH and more

Some researchers note that the Nature Physics study is a breakthrough for the proton LPA community and that laser plasma accelerators are now ready for translational research. The study sets new standards for proton LPA sources, these researchers say, and paves the way for subsequent studies, such as those on the FLASH effect.

The FLASH effect occurs when a therapeutic radiation dose is delivered in a fraction of a second at an ultrahigh dose rate. In clinical settings, FLASH radiotherapy may reduce radiation damage in healthy tissues while remaining effective on tumours. Kroll notes that while the researchers were not trying to induce the FLASH effect in this study, they did find that single-shot sample irradiation “at full power” resulted in parameters that are expected to trigger the FLASH effect. As a result, future applications of their proton LPA platform may include verifying the FLASH effect with protons and investigating radiochemistry in the context of FLASH.

“[This study] was an amazing display of teamwork between technicians, engineers, physicists and biologists, all pulling in the same direction,” says Kroll. “We will continue to try answering radiobiological questions and look closer into the FLASH effect and its mechanisms. In parallel, we constantly strive for higher proton energies, particle numbers and source stability.”

Quantum computing meets machine learning, how motorsport could save the planet

This episode of the Physics World Weekly podcast features an interview with the physicist Maria Schuld, who is a senior researcher and software developer at Xanadu – a Toronto-based quantum technology company. She talks about the challenges and rewards of implementing machine-learning systems on quantum computers.

Also on hand is the author Kit Chapman, who chats about his latest book Racing Green: How Motorsport Science Can Change the World. He explains how the myriad technologies developed to make racing cars faster and safer have already benefitted society – and how they could help us combat climate change.

Optical chipmaker focuses on high-performance computing

How did Lightelligence get going?

Spencer Powers: I was an MBA student at MIT in 2015 when I met Yichen, who was doing a PhD with Marin Soljačić – who is also a co-founder of Lightelligence. We met in a class where half the students were doing an MBA and the other half were doing PhDs.

Everything started in 2017 when Yichen and Marin published a paper in Nature Photonics called “Deep learning with coherent nanophotonic circuits”. This describes the basis of our technology, which was Yichen’s PhD project.  

In 2018 we assembled a team of 15 employees and the company got its first round of funding. In 2019 Lightelligence produced its first optical artificial intelligence (AI) accelerator demo called COMET, which was able to accurately recognize handwritten digits.

By 2020 we had raised more than $100 million and today we have about 180 employees who have already created several chip designs for integrated optical computing.

Why is there a need for integrated optical computing systems? 

Yichen Shen: Starting around 2012, AI moved into its “modern era” in terms of performance and processing requirements. The processing power needed to run the most advanced AI algorithms is doubling every 3–4 months. So in the past decade, requirements have increased by a factor of about 150 000 – and conventional computers cannot keep up with this. People started to use graphical processing units (GPUs) – but now even GPUs can’t keep up. 

In about 2015, Moore’s law started to break down – transistors are still getting smaller, but the performance gains from the ongoing miniaturization are not as good as they used to be. What is more, the time it takes to develop the next generation of smaller devices is getting longer. 

The AI industry has responded by moving to increasingly specialized technologies – from central processing unit (CPU) to GPU to field-programmable gate array (FPGA) to application-specific integrated circuit (ASIC) – but it is running out of fuel. So what we really need is a new technology to further expand the computing power needed for AI.

And that’s where optical processors come in?

YS: Yes, at Lightelligence we want to power the next generation of AI technology with integrated photonic ASICs. Light provides an ideal solution to AI computing for three key reasons. First, you can get a very high data throughput using photonics. An optical fibre can carry more than one terabit per second, which is about 100 times greater than a copper wire. Second, photonics is much more energy efficient than electronics because there is no ohmic heating as light travels through a circuit. Finally, light has lower latency because it travels much faster through circuits than electrons. 

How does your technology solve the speed problem?

YS: We have invented an optical multiplication and accumulation (MAC) unit that uses nanoscale photonic devices to do matrix multiplications. We also have a separate network on chip (NOC) to carry and broadcast data to the electronic chips in the system. 

We use a combination of optical and electrical devices because researchers have not yet solved the “optical memory problem”, so we must use electronic memory storage. Our solution is a hybrid and we believe that this type of architecture will remain for at least the next few decades. 

Our optical circuits do not contain optical transistors. Instead, our chips carry out analogue optical computing. Light travels through the device via waveguides and the signals interfere with each other to perform the required computation. It’s a bit like light going through a lens, which does a Fourier transform on the light. 

While this is a passive effect, we can actively tune how this interference occurs. We have created a programmable chip where we control what kind of matrix multiplication the light is doing. This is done by applying voltages to different parts of the chip.

What do you plan to do with this technology?

YS: Using this technology, we can build an AI processor that is much higher speed and lower power than traditional digital electronics. Our potential market would be cloud AI, finance and smart retail. Digital assistants such as Siri and Alexa could benefit from better natural language processing – recognizing longer sentences and behaving more like real human beings. Other applications include computer vision for self-driving cars, optical character recognition, drug discovery and robotics. 

We can go to mass production volumes when needed – we are “lab-to-fab”. Our products do not use esoteric technologies and can be made using standard chip fabrication processes

In December 2021 you released a new product called PACE; what does it do?

YS: PACE, the photonic arithmetic computing engine has been developed over the past two years. We believe that PACE is the first optical system that has been shown to be superior to a digital system. PACE is also our first system that is designed for applications beyond AI and deep learning. It searches for solutions to some of the hardest computational problems including the Ising model, the graph max-cut problem and the graph min-cut problem. The Ising model is an NP (nondeterministic polynomial) complete problem that is very difficult to solve – all quantum computing companies use this as a benchmark.

PACE solves these problems using an optimization algorithm. Real world applications include bio-informatics, scheduling, circuit design and materials discovery. 

The system combines a photonic integrated circuit (PIC) with an electronic integrated circuit (EIC). The EIC contains the memory for PACE, and it also does some non-linear operations using conventional logic gates. The EIC is interfaced with the PIC, which does the matrix multiplication. PACE runs a recurrent heuristic algorithm, which converges on the solution after thousands or tens of thousands of iterations. 

How fast is PACE?

YS: The computing speed is dictated by how fast we can do the matrix multiplication. Our chip can finish a 64 × 64 matrix within 4 ns. In contrast, the best GPUs will take hundreds of nanoseconds. This is how PACE can achieve orders of magnitude advantage over digital electronics. 

PACE runs hundreds of times faster than high-end CPUs and GPUs and 25 times faster than electronic systems that are especially designed to solve the Ising model. 

What are the current technological challenges that you face in developing products like PACE?

YS: Today, the biggest limiting factor for our technology is the electronics. We can’t live without electronic memory – we need to use it. This means we must convert photonic signals back to electronics and vice versa, and that limits our performance. Another challenge is device size. Photonic devices are still much larger than electronic devices. But we are still better than pure electronics. 

What is next for Lightelligence?

SP: We will continue to grow the company and we have several projects running internally now. Later this year we will be unveiling an AI accelerator pilot project that will involve pilot customers. It is not a production-scale product, but it is a commercial-grade AI accelerator that gets its advantage from photonics. This will be an important milestone for us. 

Following that, we will release a production-level AI accelerator that will have high throughput and low latency compared to other AI accelerators. 

Can your chips be mass produced?

SP: Yes, we can go to mass production volumes when needed – we are “lab-to-fab”. Our products do not use esoteric technologies and can be made using standard chip fabrication processes. With PACE, mass volume production is not needed. We are producing in the order of 10 units for demonstration purposes. 

New algorithm is forever simulating bubbles

“Foamy flows” of bubble-filled materials occur in a host of industrial processes, including food and cosmetics production as well as drug development and delivery. They are very difficult to model mathematically, however, since the bubbles in these foams can be as tiny as microfluidic crystals or as large as ocean waves. Computer models also need to take into account interactions between the bubbles, which are separated by stable, thin films of liquid.

A team of researchers at Harvard University in the US and ETH Zurich in Switzerland has now overcome these difficulties with a new multilayer volume-of-fluid (multi-VOF) method that handles multiscale, non-coalescing bubbles with ease. The new technique can simulate thousands of bubbles and makes it possible to model the behaviour of foamy flows from the micro- to the macro scales.

Current methods of simulating foamy flows rely on colour-coding individual bubbles in a foam and tracking each of them. These methods are computationally costly, which limits the simulations to just a few tens of bubbles. Real foamy flows, meanwhile, may contain anywhere from thousands to millions of bubbles.

Breaking a foam into a grid

Instead of tracking individual bubbles, the new multi-VOF technique developed by Petros Koumoutsakos and colleagues breaks the foam down into a grid in which each cell of the grid contains parts of a maximum of four bubbles. Each bubble is assigned a unique colour and the colours are used to connect the parts from neighbouring cells. For example, if one part of a bubble is in a particular cell, team member Petr Karnakov explains that the remaining pieces of the bubble must be in neighbouring cells.

The researchers developed an algorithm that finds these remaining pieces by matching the corresponding colours. This approach does away with the need for tracking individual bubbles, thereby allowing the researchers to create predictive simulations across multiple length scales.

The Harvard team backed up its simulations with experiments and complemented them with an open-source software package known as Aphros. After reporting their work in Science Advances, the researchers now plan to apply the computational tools they developed to a variety of science and engineering problems. “These include predicting foam dynamics in food processing, controlling bubbles in microfluidic devices and designing membrane-less electrochemical reactors for hydrogen production,” Koumoutsakos tells Physics World.

Arab Physical Society hosts inaugural event to celebrate founding

The Arab Physical Society (APS) is hosting its first ever event today, 7 April 2022, just 10 months after the organization was founded. The one-day online conference features a welcome address by Prince El Hassan bin Talal of Jordan as well as talks by leading physicists, including three Nobel laureates.

The APS is a non-governmental, non-profit organization that seeks to promote physics in countries that have Arabic as an official language. Since its formation on 7 June 2021, the APS has been building its organizational structure, setting up a member-registration process as well as electing a seven-member advisory committee.

The founding president of APS is Shaaban Khalil, a high-energy physicist who is currently director of the Fundamental Physics Center at the Zewail City of Science and Technology in Egypt. The advisory committee consists of physicists at universities and institutes in Egypt, Jordan, Libya, Morocco, Palestine, Qatar and Syria.

At today’s event Takaaki Kajita, who shared the 2015 Nobel Prize for Physics with Arthur McDonald, will discuss his work on neutrino oscillations. Fellow Nobel laureate Gerard t’Hooft, who shared the 1999 prize with the late Martinus Veltman, will examine the foundations of quantum physics and the Standard Model of particle physics.

“I am excited about the launch of the Arab Physical Society,” says Kajita. “I hope that the Arab Physical Society will foster research and education in physics in the Arab world, which is very important.”

The Arab Physical Society and this inaugural event will help physics in the region gain international visibility and consolidate links between physicists.

John Ellis, King's College London

The third Nobel laureate to speak is Roger Penrose, who shared the 2020 Nobel prize with Reinhard Genzel and Andrea Ghez for their work on black holes. Other presenters include Claudia Felser from the Max Planck Institute for Chemical Physics of Solids in Germany and theorist John Ellis from King’s College London, UK.

Ellis told Physics World that physics in the Middle East has been advancing in recent years. The progress has been aided by the construction of the SESAME synchrotron near Amman, Jordan, as well as the founding of the King Abdullah University of Science and Technology in Saudi Arabia. Other initiatives include regional participation in experiments at the CERN particle-physics lab near Geneva, spearheaded by Morocco and Egypt.

“The establishment of the Arab Physical Society and this inaugural event will help physics in the region gain international visibility and consolidate links between physicists across the region,” says Ellis. The APS is, according to its website, also keen to foster diversity “to ensure that everyone has the same opportunities, regardless of gender, ethnicity, religion and culture”.

Lia Merminga becomes first woman to head Fermi National Accelerator Laboratory

The particle physicist Lia Merminga has been appointed the first female director of the Fermi National Accelerator Laboratory based in Batavia, Illinois. She will become the seventh director of the US particle-physics lab and succeeds Nigel Lockyer, who has headed Fermilab since 2013. Merminga will begin her term on 18 April.

Born in Greece, Merminga earned a bachelor’s degree in physics from the University of Athens in 1983 and then began a PhD in physics at the University of Michigan, Ann Arbor. As part of her PhD, Merminga working on the Tevatron at Fermilab as part of the lab’s new graduate progamme in accelerator physics.

After completing her PhD in 1987, Merminga went on to manage projects at the Thomas Jefferson National Accelerator Facility in Virginia, Canada’s TRIUMF accelerator, as well as the SLAC National Accelerator Laboratory in California. She also chaired the Fermilab Accelerator Advisory Committee for the influential Particle Physics Project Prioritization Panel “P5” panel report, which was released in 2014 and set the future pathway for US particle physics.

Merminga returned to Fermilab in 2018 as project director of the lab’s PIP-II superconducting accelerator – a major upgrade to the lab’s accelerator complex. When complete in 2028, it will produce the world’s most intense beam of high-energy neutrinos for Fermilab’s Long-Baseline Neutrino Facility, which involves sending the particles some 1300 km to a detector at the Sanford Underground Research Facility in Lead, South Dakota.

‘Visionary leader’

The search committee that unanimously recommended Merminga’s candidacy emphasized her commitment to international collaboration. “Lia has been instrumental in developing strong scientific collaborations with research institutions across America and throughout the world in her past positions,” says Eric Barron of the Universities Research Association, which manages Fermilab for the Department of Energy’s Office of Science. “We look forward to a continued partnership as she assumes the role of laboratory director.”

Lia is an accomplished accelerator physicist, an amazing project director and a visionary leader

Jill Hruby

That view is backed by Jill Hruby, director of the Department of Energy’s National Nuclear Security Administration and a former director of Sandia National Laboratories. “Lia is an accomplished accelerator physicist, an amazing project director and a visionary leader,” Hruby told Physics World. “She will bring new insights and perspective to the DOE laboratory complex. I am excited about having her as one of our laboratory directors and look forward to working with her.”

Merminga herself is excited about the future at Fermilab, which over a decade ago switched its focus towards neutrino physics following the closure of the Tevatron. “Fermilab has given me a very rewarding career,” she says. “My goal as Fermilab director is to successfully complete the profound and compelling vision [of PIP-II and related projects] while continuing to deliver groundbreaking science and technology innovation and realize the lab’s full potential in workforce development and diversity, lab operations, and in regional, national and international partnerships.”

Ultrastrong magnetic fields could prevent neutron stars from forming black holes

A massive and exotic type of neutron star could be formed by the merger of two neutron stars and avoid becoming a black hole – at least temporarily. That is the conclusion of Arthur Suvorov at Manly Astrophysics in Australia and Kostas Glampedakis at Germany’s University of Tübingen who have calculated that magnetically supramassive neutron stars could stave off gravitational collapse, despite lying above the theoretical mass limit for black hole formation.

In 2017 the LIGO–Virgo collaboration detected the first gravitational waves emanating from two neutron stars as they spiralled into each other, and eventually merged. This event provided important opportunities for astronomers to study the aftermath of the merger using a range of different telescopes, but key questions remain about the object that was created.

Neutron stars are expected to have masses that are at least 1.1 times the mass of the Sun, so when two neutron stars merge, they can create an object with a mass that is greater than the Tolman–Oppenheimer–Volkoff limit. Above this limit, which is believed to be 2–3 solar masses, a neutron star created in a merger is expected to collapse immediately to a black hole.

Magnetic effect

The Tolman–Oppenheimer–Volkoff limit applies to non-rotating neutron stars, and astrophysicists believe that more massive neutron stars can avoid collapse temporarily if they are spinning. In this latest theoretical analysis, Suvorov and Glampedakis have shown that a neutron star above the Tolman–Oppenheimer–Volkoff limit could also avoid collapse if it was created with a magnetic field some 1017 times stronger than that of the Sun.

The duo reckon that these “magnetically supramassive” neutron stars could then stave off collapse for roughly 1–10 years as the strength of the magnetic field wanes. The exact timescale depends on the star’s core temperature, internal field strength, birth mass, and the relation between its mass and density.

Suvorov and Glampedakis believe that the formation of a magnetically supramassive neutron star would be accompanied by short gamma-ray bursts, followed by an intense but short-lived X-ray afterglow. On its collapse to a black hole, the resulting magnetic shocks would accelerate surrounding electrons to relativistic speeds, producing a flash of fast radio bursts.

If their calculations are correct, the duo believes that these characteristic signals could be observed by existing telescopes. If so, they could be picked up in tandem with future observations of gravitational waves emanating from merging neutron stars. They say that such multimessenger observations would provide a “smoking gun” signature, offering concrete evidence that this exotic variety of neutron star can exist.

The research is described in Physical Review D.

Astronomers see star enter a ‘Maunder Minimum’ for the first time

For the first time, astronomers have observed a star that has entered a state of low, or flat, activity – analogous to the famous Maunder Minimum that gripped the Sun during the latter half of the seventeenth century.

Anna Baum, of Penn State University and Lehigh University, led a survey of 59 approximately Sun-like stars, combining over 50 years’ worth of observations drawn from the likes of the Mount Wilson Observatory HK Project that ran from 1966 until 2003 (though Baum’s team did not have access to data post 1995), and the California Planet Search at Keck Observatory that’s been running since 1996.

Baum’s team was searching for signs of magnetic activity on the stars – the same activity that produces sunspots, flares and coronal mass ejections on our Sun. They did so by observing absorption lines of ionized calcium in each star’s spectra. In particular, the calcium H and K spectral lines are sensitive to the strength of the magnetic field. This is quantified using the so-called S-value, which is a measure of the strength seen in the H and K lines. A higher S-value indicates a more magnetically active star.

Comparing these stars to the Sun enables astronomers to better determine how typical – or not – the Sun is as a star. The Sun’s magnetic activity is defined by its 11-year sunspot cycle. Of the 59 stars that Baum’s team surveyed, 29 appeared to also have starspot cycles, and the period of those cycles could be measured for 14 of them.

“Of these 14 stars, the average length of their cycle is just under 10 years, which is similar to the Sun’s 11-year cycle,” Baum tells Physics World. However, not all the stars adhered to this time frame. One star that was surveyed has a cycle lasting less than four years, while another star, HD 166620, had a cycle 17 years long.

Note the past tense. Sometime between 1995 and 2004, HD 166620’s starspot cycle simply stopped.

The star that switched off

The uncertainty in the timing is because it occurred during an instrument change – the transition from the Mount Wilson survey to the California Planet Search, which uses the HIRES spectrometer, designed not only to search for the radial velocity signals of exoplanets, but also to take more complex spectra of stars. The HIRES instrument received an upgrade in 2004, after which the flat activity of HD 166620 became apparent.

Despite this, “HD 166620 has been shown to have flat activity for over ten years,” says Baum. And although the beginning of this flat activity was missed, “It’s exciting to see exactly when it might return to its cycle.”

This return to activity may be some time coming. The Sun’s Maunder Minimum lasted from 1645 and 1715, during which its solar activity almost entirely fell off a cliff. Between 1672 and 1699, fewer than 50 sunspots were observed. Contrast that with the Sun’s normal magnetic activity: even at the minimum in the Sun’s 11-year cycle there are ordinarily a dozen or so sunspots per year, and at maximum, over 100.

The Maunder Minimum “was as though the Sun’s magnetic cycle mostly turned off for a period of about 70 years, then turned on again,” says Ricky Egeland of NASA’s Johnson Space Center, who was not involved in the Baum study.

Rotational reasons

Exactly what causes Maunder Minimum-type events remains unclear, though Egeland’s previous work found some clues. In a study with Travis Metcalfe of the Space Science Institute in Arizona, and Jennifer van Saders of the University of Hawaii, he showed that the rotation of a star is key to understanding Maunder Minimum events.

Stars begin life spinning fast, and gradually slow down over their lifetime, which in some way affects their dynamo that generates their magnetic field. As they age, Maunder Minimum events may become more regular, until “at some point in a star’s rotational evolution, they enter a grand minimum and never return [to an active cycle],” explains Egeland. Hence studying stars such as HD 166620, which are near this transition point, will help astronomers to better say when the Sun’s magnetic dynamo may shut down permanently. “It might take another billion years or so,” Egeland notes.

Of the other stars in the survey, some displayed chaotic activity, while others had none at all. One star, HD 101501, completely shut down its magnetic activity between 1980 and 1990 before reactivating again. And other surveys have found similarly lethargic stars. These include the star HD 140538, where a transition from flat activity to cycling was observed, and HD 4916, where the magnetic cycle became progressively weaker, but didn’t shut off. However, “HD 166620 is, to my knowledge, the first that we have observed to have clearly just entered its minimum,” says Baum.

Further monitoring to try to catch HD 166620 returning to activity will now take place. The flat activity “might be visible in the overall brightness of the star, which is something we’re looking into,” says Jason Wright, one of Baum’s co-authors from Penn State University.

The team publish the results in The Astronomical Journal.

Copyright © 2026 by IOP Publishing Ltd and individual contributors