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 10

How far ahead can we predict the weather?

What will the weather be like tomorrow, or next week or, indeed, in a month’s time? Today’s numerical weather predictions can easily answer the first two parts of this question, but they fall short beyond about 14 days. New work on predictability limits by scientists at the University of Miami and the NOAA Cooperative Institute for Marine and Atmospheric Studies (CIMAS) say that this maximum could be extended to around 129 days – but probably no further. Whether this limit can be reached in practice, however, will depend on future improvements in forecasting technologies, they say.

Numerical weather prediction was developed in the late 1940s. Indeed, it was one of the first major applications of electronic computers. As these have advanced, so have the predictions. Indeed, thanks to ever more sophisticated mathematical models capable of analysing increasing amounts of observational data, we can now predict, for example, when hurricanes will occur as far as eight days in advance – something that was deemed impossible even a few decades ago. But how far ahead can such predictions be made?

Previous attempts to determine this predictability limit have largely relied on analysing the behaviour of very small perturbations in the atmosphere. These tiny errors, however, are inaccessible either observationally or via numerical modelling, explains Zoltan Toth of the NOAA who led this new study together with his colleague Wei Zhang. “These past studies therefore had to make some assumptions about the behaviour of the small errors, and these assumptions are necessarily somewhat arbitrary.”

Analysing the energetic balance of the atmosphere

The method developed by Toth, Zhang and their colleagues makes no reference to this error behaviour and instead goes back to analysing the energetic balance of the atmosphere, which continually absorbs solar radiation on the molecular level. In their approach, the researchers began by considering how our planet’s atmosphere would behave as a closed system, the deterministic dynamics of which preserves information about its initial state, assumed to be perfectly known. In theory, such a hypothesis would allow forecasts of the atmosphere’s behaviour right out to infinity. This is not possible in reality, they explain, because the atmosphere is obviously not a closed system: it receives and emits radiation. In the real atmosphere, the phases of photons in sunlight are completely random and cannot be determined, injecting quantum-scale uncertainty into the atmosphere.

These unknown quantum characteristics act as noise and destroy any retained information, beginning first of all on the smallest scales, explain the researchers. As these scales become bigger, noise eventually affects all parts of the system. When the total energy in the atmosphere is entirely replaced, the ability to make any prediction is completely lost.

“Using the relatively well-measured quantities of total energy in the atmosphere and the incoming and outgoing solar radiation fluxes at its upper boundary, we estimate that the range of skilful forecasts could potentially be extended from 14 days as at present to 129 (±7) days at the most,” explains Toth.

Very different from the mainstream discourse

“This result is very different from the mainstream discourse about atmospheric predictability, which to this day is strongly influenced by early publications in the field,” adds Toth. “More recently even, the authors of a 2018 article in the Bulletin of the American Meteorological Society speculated on whether current forecast systems are reaching their limits for when it comes to predicting tropical cyclones. In 2020, we argued that this limit is at least decades, if not much farther away.”

The theoretical foundation for the new methodology to determine the time limit of predictability is rather simple, he says. “At its core is a conceptual realization, which came to us intuitively. This is that the energy turnover time (that is, the time it takes for all energy in the atmosphere to be replaced by incoming solar radiation) is equivalent to the upper limit of predictability.”

While there is no complex mathematics involved, Toth notes that he and his colleagues had to find a way to describe the technique in a readily understandable way – something that was not easy, he admits, given the complexity of the subject.

Looking ahead, the researchers, who detail their present work in Advances in Atmospheric Sciences, are now looking into several other independent ways to estimate the limit of predictability. “There is no doubt, this limit has a theoretical feel to it,” Toth tells Physics World. “Yet there are a number of very practical implications for both traditional equation- and AI-based modelling of the atmosphere following on from the theory of predictability that we are also exploring in a series of ongoing studies.”

Skyrmion topology makes long-distance optical communications more robust

Researchers at the University of the Witwatersrand (Wits) in Johannesburg, South Africa, have succeeded in harnessing a relatively unexplored property of light – its topology – to robustly transfer information through the atmosphere. The work could advance the development of more reliable long-distance optical communications, including space-based ones, and connect remote and underserved communities worldwide.

Their work is detailed in two separate studies, one based on classical and the other on quantum optics. These studies show that the topological information can remain remarkably robust – even when the states of light carrying this information are subject to the severe distortions caused by atmospheric turbulence. Instead of relying entirely on increasingly sophisticated hardware or computational algorithms to correct atmospheric distortions before or after light transmission, the new concept provides a fundamentally different way of thinking about robust optical communication in complex environments, such as air or water, they say.

Both studies are based on the same fundamental idea: creating an optical topological structure known as a skyrmion in a light field and testing whether it remains robust when it passes through a distorting environment, explains Andrew Forbes, who is head of the Structured Light Lab at the Wits School of Physics. Skyrmions are quasiparticles with a two-dimensional, knot-like structure and were first observed in magnetic materials. They have, however, recently been seen in electromagnetic fields and the electric field of light waves. Here, the swirling structures are created by twisting the polarization or spin vectors of light in space so that every polarization is found exactly once, twice, three times and so on – the count being the skyrmion number. Skyrmions are topologically stable, which means that external perturbations do not affect them in any way.

Real-world test

In the first study, Forbes and colleagues used laser light with a wavelength of 532 nm shaped into various vectorial beams using a spatial light modulator and a modified Mach–Zehnder interferometer to generate optical skyrmions with skyrmion numbers of 1, 2 and 3. These numbers, which are always whole integers, describe how twisted the vortex-like structure of the skyrmion is. They then sent the laser light through a 270-m-long real-world free-space optical link located on Wits’ Braamfontein campus in the centre of Johannesburg.

The researchers measured the topology of the skyrmions with a Stokes polarimetry setup and found that the skyrmion number remains robust in a wide variety of atmospheric conditions, ranging from “calm, cool morning air to the highly erratic and intense distortions at midday” when it is hotter. “This result holds true even when the vectorial polarization of the underlying laser beam has been highly scrambled by atmospheric turbulence effects – and was therefore unrecognizable,” says Forbes.

In-built correlations

In the second work, the team created the topology through the in-built correlations between two photons entangled in their optical angular momentum (OAM). The OAM quantum states are not themselves stable in distorting environments and the information they encode is easily lost. The researchers investigated the robustness of a skyrmion with a skyrmion number of 1 in a controlled laboratory environment by studying photons in media with varying levels of turbulence.

The result? Despite local distortions, the skyrmions maintained topological numbers close to 1. This, they say, suggests the topological number is fundamentally decoupled from modal crosstalk induced by the turbulence and the spreading of an individual photon’s OAM.

Despite the very different physical systems, both experiments reveal a similar result, says Forbes: the conventional properties of optical fields can be strongly degraded by distortions, while the topological information encoded in those states remains remarkably robust. Together, the two works demonstrate that topology can be harnessed in both classical and quantum optical systems as a robust carrier of information in very different communication and information-processing scenarios.

Robust protection

Topology is a concept that appears throughout physics, from fundamental particle physics to magnetism and condensed matter systems. In many of these systems, explains Forbes, topology comes with a physical protection mechanism – such as an energy barrier – that explains why a topological structure is resistant to external disturbances. In optics, however, the situation is rather different. Optical topologies can be created and manipulated with relative ease, but their robustness is not necessarily guaranteed by any particular underlying physical protection mechanism. “We therefore wanted to address a fundamental question: how robust is optical topology when light encounters a real-world, highly distorting environment?

“The atmosphere provides a particularly interesting testbed because it is highly relevant to optical sensing and communications and represents an extreme case where the distortions vary rapidly both in space and time,” he tells Physics World. “Demonstrating robustness in this setting was therefore both a fundamental test of the dynamics and resilience of optical topologies and an indication of its potential practical value.”

Hurdles and challenges

One of the major challenges was simply doing precision optical experiments outdoors, he adds. “We had to contend with the unpredictability of nature: changes in sunlight and temperature, wind, rain and even the thermal expansion of buildings, which can affect optical alignment, making it difficult to maintain stable measurements over long periods.”

Another hurdle to overcome was to find a way of measuring the optical topology itself. Existing methods to do this can be quite sensitive to experimental noise and imperfections, so the researchers say they had to carefully distinguish genuine changes in the optical channel from apparent changes introduced by the measurement process. Ultimately, this meant developing new ways to both obtain and process the experimental data so that they could reliably track the topological properties of the light despite the noise that was present.

Applications include free-space optical, satellite and space communications, and linking remote or underserved communities, where atmospheric turbulence and the lack of physical infrastructure can present significant challenges, says Forbes. “More broadly, topology could be useful beyond communication,” he adds. “Indeed, its underlying robustness could potentially be exploited for optical sensing, information processing and other applications where light has to propagate through complex or noisy environments.”

The quantum experiment adds another possibility, he notes. “We know that entanglement decays in real-world conditions that are not pristine.  Our work shows that despite this, the quantum information stored in the topology remains intact.  This may allow us to distribute information across a quantum network in a more reliable and robust manner.”

Looking ahead, the Wits researchers say they will now assess whether the robustness they have observed extends to other kinds of challenging environments, such as underwater and biological tissue, where scattering, birefringence and other phenomena can be much more prevalent than in the atmosphere.

“We also want to develop detectors and techniques for measuring optical topology,” says Forbes. “Improving how we characterize these structures will be important for moving from proof-of-principle experiments towards practical applications.

“Finally, we would like to actually exploit the robustness of the topology we have identified in high-speed communication experiments. This means implementing realistic communication protocols and investigating how much information can be reliably transmitted using a ‘topological alphabet’.”

Bilayer approach ramps performance of all-perovskite triple-junction solar cells

Researchers at the Helmholtz-Zentrum Berlin (HZB) in Germany have overcome a long-standing barrier in next-generation photovoltaics by developing a new type of all-perovskite triple-junction solar cell. Perovskites are a class of synthetic materials with a unique crystal structure that excel at absorbing light, while a triple-junction design stacks three distinct perovskite sub-cells together, each tuned to capture a different part of the solar spectrum to maximize energy harvesting. With this combination, the device achieved a power conversion efficiency of 27.3%.

Stacking three such individual sub-cells allows multi-junction devices to bypass the theoretical efficiency ceiling of traditional single-junction solar cells. However, their real-world implementation has been severely constrained by optical and electrical losses at the buried interfaces of their narrow-band-gap sub-cells. By introducing a novel graphene oxide and self-assembled monolayer (SAM) bilayer strategy, the HZB team has effectively suppressed these losses, unlocking both high efficiency and unprecedented operational stability.

The challenge of buried interfaces

The main bottleneck in multi-junction all-perovskite architectures centres on the narrow-band-gap tin-lead bottom sub-cell, which captures near-infrared light. For years, the field has relied on an organic polymer called PEDOT:PSS as the hole-transport layer to extract positive charges from this sub-cell. However, PEDOT:PSS suffers from a high degree of parasitic light absorption, robbing the underlying bottom cell of valuable incoming photons. Furthermore, its highly acidic and hygroscopic nature chemically degrades the sensitive perovskite material, creating a critical compromise between efficiency and device lifespan.

This results in a severe material trade-off, whereby a layer intended to extract electric charge ends up blocking the very light needed to generate that charge, all while slowly destroying the underlying solar cell structure.

To move away from PEDOT:PSS, researchers have frequently looked to SAMs, molecularly engineered alternatives that perform exceptionally well in pure lead-based perovskites. Yet, when deposited directly onto tin-lead perovskite layers, SAMs present an unexpected physical anomaly, triggering severe internal electrical field screening and inducing uneven grain growth at the buried interface. Using fast-hysteresis measurements and bias-assisted charge extraction measurements, the HZB team discovered that standard carbazole-based SAMs lead to massive ion accumulation, which severely hinders the cell’s ability to separate and extract photogenerated charge carriers.

Unlocking charge flow with a molecular bilayer

Led by Steve Albrecht, the HZB team designed a synergistic bilayer to circumvent this destructive ion accumulation and field-screening mechanism. They deposited an ultrathin, uniform layer of hydrophilic graphene oxide directly onto the transparent indium tin oxide conductive base, before introducing a specialized SAM molecule, MeO-2PACz. The phosphonic acid head groups of the SAM anchor strongly onto the oxygen-containing functional groups of the graphene oxide via hydrogen and covalent bonding.

This chemical arrangement prompts a favourable molecular reorientation that deepens the layer’s electronic work function, markedly enhancing its electrical conductivity. Additionally, the graphene oxide provides a smooth, hydrophilic foundation that accelerates crystal nucleation, generating a uniform, high-quality perovskite film completely free of the microscopic empty spaces (nanovoids) that plague SAM-only devices in this study.

When integrated into single-junction tin-lead solar cells, the new bilayer achieved a standalone power conversion efficiency of 22.1%, heavily outpacing the 12.0% efficiency of SAM-only equivalents. Using optoelectronic characterization, the team confirmed that this massive improvement was driven almost entirely by the mitigation of internal electronic and ionic extraction losses.

Record stability and the path to 30%

The researchers demonstrated the true potential of this technique when they integrated the bilayer into a monolithic triple-junction stack featuring sub-cells with band gaps of 2.00, 1.60 and 1.25 eV. By replacing the conventional gold and PEDOT:PSS interconnecting layers with an optimized indium tin oxide and graphene oxide/SAM bilayer configuration, they minimized parasitic light absorption in the near-infrared spectrum. This structural optimization raised the short-circuit current density of the narrow-band-gap bottom cell to 10.3 mA/cm2, driving the overall triple-junction efficiency to 27.3%.

Beyond the efficiency milestone, the chemically benign and robust nature of the graphene oxide/SAM interface yielded exceptional longevity. While traditional PEDOT:PSS-based cells degrade rapidly due to chemical interactions, the encapsulated bilayer devices successfully retained 90% of their initial performance after 770 h of continuous operational tracking under 1-sun illumination at room temperature. This sets a new stability record for triple-junction all-perovskite configurations.

The team notes that the current triple-junction performance is still bound by slight current mismatches and series-resistance transport losses across the interconnecting sub-cell junctions. By further optimizing the quality of the wider-band-gap perovskite layers and fine-tuning the band alignments of the intermediate contacts, the researchers project that the efficiency of this multi-junction solar architecture can comfortably clear the 30% barrier in the near future.

The solar cell design is detailed in Joule.

Word flower puzzle no. 8

How did you get on?

28 words Warming up nicely

38 words Getting hot, hot, hot

48 words Top dog!

Fancy some more? Check out our puzzles page.

Universities should judge quantum-computing investments by what students learn

The best moment in my quantum-computing class often comes when something fails. A student writes a short quantum circuit – a program that tells qubits what operations to perform – then runs it on an ideal simulator and gets the clean answer theory predicts. Yet when the same circuit runs on real quantum hardware, unexpected results appear.

What once looked certain now becomes messy. The student may think the machine has failed but this is, in fact, when the real lesson begins. A run on a quantum processor is a physical experiment that produces a computational result.

Quantum education is expanding rapidly. A recent analysis of course catalogues from 1456 US institutions identified more than 8000 courses the titles or descriptions of w hich mentioned “quantum”. That growth is welcome, but it also carries a risk. Many students on such courses first meet quantum computing as a perfect cartoon: ideal qubits, flawless gates and clean probabilities.

Such idealization can be useful with computer simulators – ordinary software that imitates a quantum computer without real qubits. Simulators make excellent first classroom tools because they let students grasp three basic quantum ideas: superposition, where several outcomes remain possible before measurement; interference, where those possibilities strengthen or cancel one another; and entanglement, where two qubits become linked so measuring one helps predict the other.

Students should feel the shock of an imperfect result

Simulators also let students debug code and see the mathematics clearly. That matters in my classroom, because many students arrive with more confidence in Python than in quantum mechanics.

Useful as those abstractions are, they are not enough: students should learn not only the algorithm but also the machine that runs it. Real hardware exposes the gap between an ideal circuit and a physical device. A qubit is not a tiny coin showing heads and tails at once. It is a fragile physical system that gives different answers with varying probabilities when measured. Qubits lose information, gates make small mistakes, measurements misread states and nearby qubits can disturb one another.

The educational goal is to connect a quantum circuit to a real device – one in the lab, or one reached through the cloud. Yet that same circuit can behave differently after a recalibration, or after the device has “drifted”.

Students should see this as soon as possible. Teaching noise early shows students that quantum computing is an experimental science. If universities buy or rent quantum computers merely so that students can press “run”, they miss the best part of the lesson.

Intel inside

At Florida International University, this view shapes how I teach quantum computing. My students use Python, Qiskit, Jupyter notebooks, simulations and cloud quantum systems. My students have cloud access to an IonQ quantum computer that uses trapped-ion qubits, charged atoms confined by electromagnetic fields and manipulated with lasers. For now, our teaching uses cloud access to a quantum processing unit (QPU), raising a useful question: what does that access teach well, and what does it hide?

What cloud access teaches best is comparison. Students can run the same circuit on a simulator and a real QPU, then ask why the histograms – bar charts of measured output strings – differ. On platforms that expose device details, students can compare connectivity maps and calibration data such as coherence times, readout error rates and gate error rates.

Cloud systems, however, can still hide too much. A polished interface can make a QPU look like a web service rather than a delicate machine. Students may not see the lasers that control ions or atoms, the cryostat that cools a superconducting chip, the microwave electronics that drive gates, or the human work behind calibration, where engineers tune the machine to keep it reliable. Cloud access teaches students to submit and analyse circuits. Access to hardware, control settings and error data helps them understand how the processor operates.

On-campus systems matter when students can inspect what sits behind the screen. In 2024 Rensselaer Polytechnic Institute in New York became the first university campus to host an IBM Quantum System One, initially powered by a 127-qubit Eagle processor. It is a striking example of a university putting a major machine within students’ reach.

Yet the educational value of any system depends less on qubit count than on what students can inspect, control and measure. A modest, accessible teaching platform – such as a photonic testbed that uses light to process information, a cryogenic set-up for testing devices at extremely low temperatures or a control-electronics lab that shows how signals drive qubits – can often teach more than a larger system that students cannot inspect.

Universities then should judge quantum-computing investments by learning outcomes, not headlines. Can students compare simulation and hardware? Can they test how a quantum computer changes over time by running the same circuit in the morning, afternoon and again the next day? Can they keep a notebook of circuits that worked in simulation but failed on hardware? Can they explain why one qubit pair works better than another? Can they connect algorithms to measurement, control, optics, microwaves, cryogenics and error reduction?

The experience has an emotional dimension too: a beautiful circuit on the whiteboard can become a difficult experiment on the chip. Students need to ask, “What algorithm did I write, and what physical experiment did the machine actually run?”

Students should feel the shock of an imperfect result, the frustration of a circuit that should have worked and the satisfaction of tracking down the physical cause. That experience turns quantum computing from a diagram into a craft.

A university quantum computer should be a teaching instrument that reveals the machine behind the calculation. Its purpose is to train people who can question claims, understand the hardware, and build the technology with sound judgement. The field needs graduates who understand the algorithm, the machine and the gap between them.

Using AI to plan for climate variability in renewable energy systems

Using solar and wind power for clean energy generation relies on the weather, which can be unpredictable. In addition to seasonal changes and daily variability, energy planners must also account for rare and extreme weather events. Traditional planning methods often fail to capture these events accurately, resulting in systems that are either overly expensive or insufficiently reliable.

In this work, the researchers aimed to develop a more effective way of planning energy systems that can cope with real climate variability. They built an artificial intelligence model trained on 30 years of wind and solar data from Pingtan in China. The model generated thousands of realistic weather scenarios, including both typical conditions and extreme events.

The researchers then used these scenarios to determine the most cost-effective and reliable combination of energy technologies. Their Integrated Energy System included solar panels, wind turbines, batteries, hydrogen production and storage, and a connection to the national electricity grid. The optimal system configuration was identified by testing its performance across all of the AI-generated scenarios.

The optimisation showed that the most effective renewable energy system combines several technologies working together. Wind and solar provide the main source of electricity, batteries manage short-term fluctuations in supply and demand, hydrogen storage provides backup during extended periods of low renewable generation, and the grid acts as an additional safety net.

The key finding is that batteries and hydrogen have complementary roles: batteries are best for balancing daily variations, while hydrogen is better suited to storing energy over weeks or months. Together, they help create a renewable energy system that is both reliable and cost-effective.

Read the full article

A climate-driven generative scenario framework for optimizing integrated energy systems with hybrid storage solutions

Jiawei Tan et al 2026 Prog. Energy 8 025008

Do you want to learn more about this topic?

Energy policy and public opinion: patterns, trends and future directions by Parrish Bergquist, David M Konisky and John Kotcher (2020)

Real-number quantum theory can be more wrong than you thought

Imaginary numbers first appeared in the 16th century as a mathematical invention, introduced to solve equations that real numbers could not. As the name implies, many treated them as kind of trick to get results, rather than an underlying truth about the nature of reality. Although highly controversial at the time, these numbers are now ubiquitous in many fields of physics and mathematics.

One such field is quantum mechanics. Researchers routinely use complex numbers, which include both imaginary and real parts. For many simple quantum systems, you could think this is a mathematical convenience rather than a physical necessity. Predictions made with complex numbers can often be reproduced using only real numbers, albeit in a larger mathematical space.

The key question is: are complex numbers merely just a convenience, a trick, or are they necessary to understand complicated quantum systems? And if they are necessary, to what extent?

A team of researchers from France, Poland and Spain have been working on this problem and they’ve now come up with an answer. Their work relies on the composition postulate – a standard starting point for formulations of quantum mechanics which describes how quantum systems are combined mathematically.

They set out by studying a star-shaped quantum network. In this network, a number of outer parties (N) each receive one part of a quantum system, while a central party (Eve) receives the other parts from independent sources.

Each outer party chooses between several simple two-outcome measurements, while the central party performs one measurement with many possible outcomes.

Their strategy was to build a specially designed, conditional Bell test. The most well-known Bell test compares classical physics with quantum mechanics, with experiments demonstrating the latter is a better description of reality.  However, Bell tests can compare the correlations predicted by many other different theories.

The two theories studied here were standard quantum theory using complex numbers and a different formulation of the theory only using real numbers. While it was already shown that complex numbers are necessary, just how much the predictions of complex and real quantum theories differ is still an open question.

They found that the ratio between the predictions of the two theories increased proportional to N-1. So as the network gets larger, real-number quantum theory becomes increasingly bad at reproducing the results of standard complex-number quantum theory.

The work therefore gives a clear answer to the original question. Namely, in sufficiently large quantum networks, complex numbers provide an advantage that can become arbitrarily large. If the composition postulate is respected, they are physically required.

Read the full article

Gap between quantum theory based on real and complex numbers is arbitrarily large – IOPscience

S. Sarkar et al 2026 Rep. Prog. Phys. 89 070503

Optical vortices provide a natural way to create topologies

Structured light is light with a designed spatial pattern. It can be shaped into standard topological structures, such as a twist in its phase or a changing polarisation across the beam. These patterns are useful because they can encode information, probe small distances and create topological field structures.

However, the issue is that is that these carefully engineered patterns often spread or degrade as the beam travels. This is a problem for real-world applications, such as precision metrology and super-resolution sensing.

In a new work, a team of researchers from Singapore and Spain went down a different route. They identified a spin structure which is both a topological texture and topological defect that is not laboriously engineered but appears automatically at the dark core of an optical vortex because Maxwell’s equations require it.

An optical vortex is a dark core in a light beam where the electric field vanishes. Because the field is zero there, the phase cannot be defined. The researchers found that the light around this core has a tiny, ordered spin pattern.

The called this a meronic spin defect. A meron is a texture that covers half of a sphere of possible spin directions, while a defect is a point where the direction becomes undefined. Here, both features occur together: the spin vanishes at the centre but forms a half-sphere-like pattern around it.

Crucially, these defects do not spread out during propagation. The beam might diffract as usual, but the normalised spin and polarisation patterns remain confined on subwavelength scales. They are also more stable against turbulence than standard engineered structures, because they arise intrinsically from Maxwell’s equations rather than from carefully balanced beam components.

Although still a fundamental result, the finding suggests that vortex cores could provide unusually stable and compact building blocks for future structured-light technologies.

Read the full article

Non-spreading meronic spin defects around optical vortices – IOPscience

N. Mata-Cervera et al 2026 Rep. Prog. Phys. 89 077901

Cling film helps stamp two-dimensional materials onto patterned surfaces

Since the discovery of graphene over two decades ago now, two-dimensional (2D) materials have revolutionized almost all areas of science and technology thanks to their unique properties compared with bulk materials. These include high electronic conductivity, high mechanical strength, transparency and flexibility, to name just a few. The problem since the beginning, however, has been to produce sheets of these materials large enough for real-world applications and, importantly, to be able to transfer them onto other substrates without cracking the sheets. A team of researchers in Amsterdam, the Netherlands, has now succeeded in doing just this, with the help of an unexpected material – kitchen cling film. Their work could help in the continuing development of next-generation electronic and optoelectronic devices.

“The method we developed allows us to, for the first time, pick up, transfer and place large (roughly 1-mm-sized) 2D layers on almost arbitrarily patterned surfaces,” explains Jorik van de Groep of the 2D Nanophotonics group at the UvA-Institute of Physics, who led this research study. “Being able to do this is crucial since it enables the integration of 2D materials with structures like larger electronic devices and photonic coatings. Most importantly, the transfer is no longer probabilistic (as were most other methods) but has a near-unity yield.”

In the early days of 2D materials, researchers obtained micron-sized mono- or multi-layers of these structures by mechanically shaving off, or exfoliating, flakes from the bulk material using sticky tape. This was first famously done for graphene, a sheet of carbon just one atom thick. More recently, a technique known as gold-assisted exfoliation has allowed them to produce high-quality centimetre-sized layers of technologically important materials like transition-metal dichalcogenides (TMDCs) and transfer these onto flat glass surfaces.

For the field of nanophotonics in particular, explains van de Groep, we need to be able to transfer large-area 2D materials onto both flat and patterned substrates – such as those with electrical contacts and optical coatings on them, for example – if we are to fabricate functional devices. The problem is that existing transfer techniques are limited in this respect.

Kitchen cling film to the rescue

The Amsterdam team – which also includes researchers from UvA’s Van ‘t Hoff Institute for Molecular Sciences, ARCNL and AMOLF – has now developed a simple method that allows them to reliably transfer both large-area monolayers of TMDCs and hexagonal boron nitride/monolayer heterostructures onto patterned or non-patterned substrates, ranging from flat surfaces to high-aspect ratio and low adhesion patterned interfaces. In their process, which is detailed in ACS Nano, they used low-density polyethylene (LDPE), or kitchen cling film, an inexpensive and widely available polymer that has a low melting temperature.

In their work, the researchers transferred a gold-assisted-exfoliated large-area monolayer of the TMDC tungsten disulphide (WS2) from a silica (SiO2) substrate to another target SiO2 substrate. They began by fabricating a stamp made of a heat-resistant half-sphere covered by LDPE cling film. They then mounted the stamp on an xyz-micron precision stage and moved it slowly towards the WS2 monolayer, which was heated to 70 °C, at a speed of 0.5 µm/s until it contacted with the LPDE.

“We used force sensors in the stage to measure the forces involved during the stamping procedure, both in the plane of the 2D material (Fx and Fy) and perpendicular to it (Fz),” explains van de Groep. “This not only offers better control and repeatability of the transfer, it also provides crucial information on the contact and friction dynamics throughout the process.”

During initial contact with the substrate, the normal force (Fz) increases to 120 mN, at which point the researchers stop moving the stamp. Once it has contacted the monolayer, they heat the system to 140 °C. This induces a phase transition at which the LDPE melts, so allowing it to strongly adhere to the monolayer. They then cool the ensemble back down to 70 °C to solidify the LDPE. Finally, they pull off the monolayer from the LPDE with the stage moving at 0.5 µm/s and clean it to remove any polymer residue.

Material retains its good photoluminescence properties

To assess how the transfer process affects the monolayer, the researchers imaged it before and after the procedure. Before transfer, cracks covered 8% of the material. This figure increased to 14% after, with the cracks mainly generated at the edge of the stamp contact area. “As such, the procedure only modestly increases the cracked area fraction and largely preserves the monolayer topography,” says van de Groep. “More importantly, the material retains its good photoluminescence properties.”

The new work could help 2D materials scientists working in a variety of different fields to build larger and more complex devices. “Already, we are using it to develop atomically thin optical elements, optical modulators, single-photon emitters and other optoelectronic devices,” van de Groep tells Physics World. “Also, and more fundamentally, the method we have developed could allow for the fabrication of complex heterostructures of interest for quantum materials.”

The Amsterdam team is now busy optimizing its method further by, for example, building a humidity-controlled environmental enclosure around the stamping setup. “2D materials and the (van-der-Waals) adhesion between them strongly depends on the surface chemistry and, as such, on the relative humidity during the stamping procedure,” explains van de Groep. “We’re also working on additional methods to characterize the thickness of the materials in situ, that is, during the stamping process.”

Missing decay at BESIII points to long-sought glueball

Physicists at China’s BESIII experiment in Beijing have found the strongest evidence yet for the existence of glueballs – subatomic particles made of gluons alone. The glueball-like particle is called X(2370) and it was first spotted at BESIII 15 years ago.

Gluons carry the strong nuclear force that binds quarks into protons and neutrons. Gluons also pull on each other as hard as they pull on quarks, which means they ought to be able to clump into particles called glueballs.

“So far, all ordinary matter that we know of is composed of quarks and leptons,” says Colin Morningstar of Carnegie Mellon University, whose 1990s supercomputer simulations helped predict what glueballs should look like. “A glueball is a totally new form of matter made up entirely of gluons – no quarks, no leptons,” explains US-based Morningstar, who is not a member of the BESIII team.

For four decades, detecting glueballs have been an important goal of BESIII and its predecessor – according to Shan Jin of China’s Nanjing University. BESIII is on the BEPCII accelerator, where electrons collide with positrons to create large numbers of J/ψ particles. A J/ψ can sometimes decay by shedding a flash of light and dumping the rest of its energy into gluons. The gluons could then bunch together to create a glueball.

Lightest glueball

In 2011, Yanping Huang was Jin’s PhD student and observed the production of the short-lived X(2370) particle at BESIII. “The most important next step was to determine its spin-parity quantum numbers, and it turned out to be one of the most difficult steps for us due to large background processes,” says Huang, now at China’s Institute of High Energy Physics in Beijing. Those numbers describe how a particle behaves when turned around and seen in a mirror – a kind of fingerprint. Only in 2024, with an exceptionally clean chain of decays and all 10 billion J/ψ particles, could the team read it, and it matched the prediction for the lightest glueball.

Now, Huang, Jin and the rest of the BESIII team have gone further and searched in vain for a specific X(2370) decay channel. They argue its absence suggests that X(2370) contains no quarks and the lightest glueball is its dominant constituent.

Quarks come in several types, or flavours, and a glueball contains none, so it cannot favour one flavour over another. That even-handedness forbids the X(2370) from breaking up into a particular pair of kaons, which are lightweight particles carrying a strange quark. BESIII looked for that break-up and found that the decay occurs at a rate that is at most 8% of the rate of a much more common X(2370) decay channel.

Huang explains that an ordinary quark-based particle should undergo the kaon-pair decay far more often. “The lightest glueball is essential for a natural and complete explanation of all these properties, while all other interpretations can hardly explain them simultaneously”.

Like a duck

Francesco Giacosa at Jan Kochanowski University in Poland is enthusiastic about the BESIII research, which he was not involved with. “We have an object, X(2370), that walks like a duck, quacks like a duck and, in addition, does not do something that the duck we are searching for should not do.”

Jin says that the BESIII result has important implications. “This is the direct proof of self-coupling among gluons predicted by QCD about 50 years ago,” where QCD is quantum chromodynamics, the theory of the strong force. It also illuminates the origin of mass, “since the gluon mass is zero and the glueball mass is totally from the strong interaction”. Glueballs, he says, are “a unique kind of matter made of pure force”.

Morningstar observes, “The results are the strongest evidence yet that particles dominated by a glueball component can exist in nature. This work is certainly an experimental triumph.”

Not everyone is ready to call the 50-year search over. “In my opinion, an independent experimental confirmation would be extremely important,” says Giacosa. A glueball can also blend with ordinary quark particles, he notes, so “even if X(2370) is predominantly a glueball, we do not know precisely how large its non-glueball component is”.

Upgrade needed

Gentler collisions elsewhere could shed further light on this, says Jin, “but it seems difficult for them to collect 10 billion J/ψ events to perform systematic studies as BESIII in a short time, unless we upgrade BEPCII itself again”.

That upgrade is on the collaboration’s wish list. Huang plans further measurements of the particle’s properties, which “will also certainly help us to identify more glueballs”. Reaching them may take ten times more data.

Morningstar also looks forward to further discoveries, “Our theoretical computations have suggested other glueballs that might exist, so with this successful study, I believe the search for other glueballs will intensify”.

Giacosa thinks the prize is worth it. “Glueballs are, in my view, among the most fascinating composite objects predicted within the Standard Model,” he says, meaning physicists’ theory of the fundamental particles and forces. “Establishing one experimentally would be much more than adding one more particle to the list – it would confirm one of the most remarkable predictions of the theory of the strong interaction.”

The results were presented in August at the International Conference on High Energy Physics and are described in a preprint on arXiv.

Copyright © 2026 by IOP Publishing Ltd and individual contributors