“Only three or four in every hundred PhD students in the United Kingdom will land a permanent staff position at a university. Start planning your next step now!”
This eye-catching statistic – and the imperative that follows it – appeared front and centre on the registration page for a careers event organized by the Scottish Universities Physics Alliance (SUPA). Judging from the event itself, which took place yesterday in Edinburgh’s Dynamic Earth centre, it’s a message that finally seems to be getting through – at least among the organizers of careers events.
I’ve spoken at more than a dozen such events, flying the flag for careers in science communication and physics-based industries. At a few of them, I got the impression that I was there as a wild card among more “conventional” speakers (read: researchers in university or government labs). Other, more representative events were nevertheless advertised as promoting “alternative” careers, even though the statistics suggest that non-academic career paths for physicists are about as “alternative” as flannel-shirt-wearing Nirvana wannabes were in the 1990s.
At the SUPA event, though, I was pleased to see that all 10 of my fellow speakers had spent significant portions of their working lives outside universities. Mantas Butkus, for example, parlayed his PhD in quantum dot-based semiconductor lasers into a career at a laser manufacturer, Coherent. Similarly, Kirstin Hay’s astrophysics PhD focused on exoplanet transits, and she is now putting her data-science skills to use at Sainsbury’s Bank. Even Euan McBrearty, who is now a research fellow at the University of Glasgow, worked at a string of private companies (including QinetiQ, Helia Photonics and Clyde Space) before joining Glasgow’s Quantum Technology Hub for sensors and metrology.
The 11 of us spent the morning circulating between tables full of PhD students and answering questions about what we do. It’s a common format, and one that seems to work well; the students get to ask about things that interest them, and the speakers don’t have to address the whole group at once. Since there were more speakers than tables, I had the pleasure of accompanying another physicist, Ewan Hemingway, around the room and listening in on his answers.
As a PhD student at the University of Durham, Hemingway worked on computational fluid dynamics simulations, and his post-doc focused on modelling flow instabilities in particle physics. Nowadays, though, he’s an image analysis expert in the white-hot field of machine learning in medicine. The projects he works on typically involve training computers to do tasks such as segmentation (that is, delineating which parts of an image represent different organs or structures in the body) and registration (determining how those organs or structures change with time, over a series of images). The goal, he explained, is to develop products that help clinicians diagnose and treat disease.
The most common question the students asked Hemingway concerned how much experience he had in machine learning when he applied for his current job at Canon Medical Research Europe. His initial answer was “very little,” although he later acknowledged that he took an online course on artificial intelligence (AI) in his free time during his postdoc.
To me, this answer was both encouraging – it’s great to hear that employers are willing to look beyond tick-box lists of skills and hire physics PhDs based on talent, demonstrated interest and potential – and familiar. Certainly, the most valuable skill I gained during my PhD was a willingness to approach a difficult new topic and dive into it. And with luck, it’s also a skill that will help the 60 or so physics PhD students at yesterday’s event find their way into interesting careers – either in universities or outside them.
A new mechanoluminescent material can store information about mechanical events it experiences and then divulge its history under a laser “spotlight” as many as three days later. Demonstrated by Philippe Smet and colleagues at LumiLab, part of Ghent University in Belgium, the material derives its long memory from excited electrons that are trapped within defects in the material’s crystal lattice. The team’s findings could have several applications, including new ways to record damage to components or structures.
Mechanoluminescent (ML) materials emit light when they are subjected to mechanical stresses and deformations such as fracturing, bending, and compression. In some materials, ML effects can be induced by first irradiating the material with blue or ultraviolet light. While many of the electrons excited by this light will release their energy immediately, a fraction will become trapped in higher-energy states associated with defects in the material’s crystal lattice. Then, when the material experiences a mechanical stress, the trapped electrons return to their luminescent centres, emitting photons in the process.
This property could be useful for many applications, including pressure sensing and damage detection. However, because these ML materials emit light only when pressure is applied, materials used for such purposes must be monitored continually.
A range of trap depths
This shortcoming has hindered the development of ML materials, but Smet’s team got around it thanks to a ML crystal that contains defects with a wide range of trap depths. In this material, which has the chemical formula BaSi2O2N2:Eu2+, mechanical stresses push some of the excited electrons into deeper traps where they can be stored almost indefinitely – essentially “writing” information about the material’s history into its crystal structure. These deeply-trapped electrons can only escape and emit photons when they are subjected to infrared radiation, meaning information about past stresses can be “read out” optically.
To demonstrate the pressure-memory effect, Smet and colleagues dragged a rod across a sample of the material, then measured its luminescence after irradiating it with an infrared laser at different times. They found that the material’s memory of the location and intensity of the mechanical stress endured for more than three days. For the first time, this showed that mechanical actions can lead to a reshuffling of trap occupations in ML materials.
Smet and colleagues envisage that their memorable ML material could have applications in detecting damage to buildings, vehicles, and other structures for which continuous monitoring is impractical. Their findings could also inform studies of how charge carriers move into different types of defect within luminescent crystals, with further applications potentially including efficient LEDs and better glow-in-the-dark coatings. In future studies, the LumiLab team hope to improve the sensitivity of their material to optimize its memory storage and readout capabilities.
Proton therapy offers improved spatial dose delivery compared with photon- or electron-based radiotherapy. Meanwhile, recent studies suggest that FLASH radiotherapy – in which radiation is delivered at ultrahigh dose rates – can decrease normal tissue damage while maintaining anti-tumour activity. Combining the two could offer the intriguing possibility of enhancing proton therapy’s intrinsic spatial advantages with the unique temporal effects of FLASH.
With this goal, a team at the University of Pennsylvania has designed a novel treatment system that delivers FLASH proton therapy (FLASH-PRT) using double-scattered protons under CT guidance. In studies on mice, the researchers also report the first demonstration of FLASH-PRT-mediated normal tissue radioprotection (Int. J. Radiat. Oncol. Biol. Phys. 10.1016/j.ijrobp.2019.10.049).
Adapting the output
Using the ProteusPLUS accelerator employed for clinical treatments at the Roberts Proton Therapy Center, the team designed and constructed an apparatus that can deliver protons at either FLASH (60–100 Gy/s) or standard (0.5–1 Gy/s) dose rates.
“We worked extensively with the IBA engineers to achieve robust delivery for these extremely high dose rates,” explains co-senior author Keith Cengel. “This mostly involved tuning of the beamline and developing a control system to request/start/stop beam delivery.”
Cengel notes that the ability to adapt a commercial proton therapy system for FLASH depends on the type of accelerator, as well as beamline characteristics and nozzle type. He suggests that it may be easier to adapt for proton FLASH in a multi-room cyclotron-based facility.
The researchers used the single pencil beam delivered into the centre’s dedicated research room to create a double-scattering system with a uniform and expanded field size (thus removing the additional variable of spot-scanning speed). The apparatus controls dose rate using the proton beam current, with all other system elements identical, thereby ensuring that dose rate is the only difference between FLASH and standard conditions.
The team also developed dedicated dosimetry tools, an essential requirement for FLASH studies. “We built and tested a double-scattering apparatus with real-time verification of absolute dose rate and time–dose structure and ensured that this system was capable of accurately measuring and delivering the desired target dose within typical clinical tolerances,” says Cengel.
Depth–dose profiles of the both FLASH and standard proton beams demonstrated that the entrance region had relatively homogeneous dosimetry. The researchers also confirmed that dose rate in the entrance portion of the beam was linear with beam current, up to 300 nA. Measuring the 3D uniformity of the collimated beam using EBT3 film in solid water phantoms, for the 1 x 2 cm collimator used in animal studies, revealed that 100% of the target was covered by 90–95% of the desired dose.
In vivo advantages
Cengel and colleagues performed in vivo studies to compare the impact of dose rate on normal tissue damage and tumour growth. During irradiations, they used a NaI(TI) crystal detector to measure prompt gamma rays emitted at the beam line exit window and obtain the real-time delivered dose rate.
For the first study, they randomly assigned mice to receive 15 Gy whole-abdominal irradiation using either FLASH-PRT (78±9 Gy/s) or standard proton therapy (0.9±0.08 Gy/s). FLASH significantly reduced the loss of proliferating cells in intestinal crypts (required for self-renewal of the epithelium) compared with standard dose rate treatment, although both approaches decreased numbers compared with nonirradiated controls.
The researchers also evaluated long-term effects in mice exposed to 18 Gy focal intestinal irradiation. Eight weeks after treatment, they observed pronounced fibrosis on intestines irradiated using standard proton therapy. Such intestinal injury is a common side effect for patients with gastrointestinal tumours treated using focal radiotherapy, and can limit the deliverable dose. In contrast, FLASH-irradiated animals experienced significantly less fibrosis, with intestines qualitatively resembling nonirradiated tissues.
Finally, the team treated mice bearing pancreatic flank tumours with a dose of 12 or 18 Gy, using FLASH or standard proton therapy. The target encompassed the tumour and the upper intestine. Both approaches showed a similar dose-dependent inhibition of tumour growth, essentially translating to an increased therapeutic window for FLASH-PRT.
The team is now working to simultaneously improve the system, define the physical parameters of biological outcome (time–dose–fractionation, for example) and better understand the fundamental mechanisms by which FLASH retains tumour effect while sparing normal tissues.
As for when FLASH-PRT may possibly be ready to enter clinical trials, Cengel says that this will depend on the outcome of upcoming studies. “The time before human clinical trials is likely to be measured more in years than in months,” he tells Physics World. “But we are particularly excited about the possibility of starting clinical trials in veterinary oncology patients.”
Squeezed margins: A disassembled WiFi-connected Juicero Press reveals how much more complex this fresh-drinks maker was than it needed to be. (Courtesy: CC BY 2.0/Steve Jurvetson)
My friends and I were talking the other day about the craziest thing we’d seen over the last few years, discounting politics that is.
We began with Juicero – a $399 WiFi-connected juicer, billed by the company’s founder Doug Evans as “the first at-home cold-pressed juicing system”. Launched in 2016, the Juicero promised a glass of fresh juice every morning without you having to squeeze fruit with your bare hands. The firm, based in California, had received $120m in venture-capital funding from serious investors including Google’s parent firm Alphabet Inc. The Juicero was such a beautiful product that Apple, if they’d been in the juice market, would have been proud of it.
As a business model, a WiFi-connected juicer appealed to investors, offering health and lifestyle benefits for cash-rich, time-poor Silicon Valley hipsters, while guaranteeing a recurring revenue to the company through sales of its bags. All customers had to do was buy a machine, register the device, download an app to their phone, order a $5–7 “produce pack”, wait for it to arrive, scan a QR code and ensure their Internet-connected device had checked “the quality of the product”.
Then someone at Bloomberg squeezed a Juicero bag by hand and made a drink just as well as the fancy WiFi-connected device – all for a fraction of the price. Almost overnight the product was scorned as a symbol of Silicon Valley hubris and the answer to a question everyone realized that maybe, sort of, they hadn’t actually been asking. Undeterred, other firms tried to copy their device, with Juicero filing a complaint in federal court in April 2017 against a competing cold-press juicing device, the Froothie Juisir, for allegedly infringing its patent.
Sadly, the Juicero turned out to be one of the most amazing business flops of recent times
Sadly, the Juicero turned out to be one of the most amazing business flops of recent times. On 1 September 2017 the company announced it was suspending sales of the juicer and the packets. The business model (equipment + consumables + subscription) was not inherently crazy. After all, parts of this approach have been used to sell everything from coffee machines to razor blades. But applied to juice, which you can just as easily buy in the supermarket, it was way too complex and expensive.
Fruity flop
The key questions for any business are simple. What problem are you trying to solve for the customer? And is your solution the simplest and most cost-effective way of delivering that value proposition with the least effort for the customer? Subscription models can work extremely well of course – just look at the success of Netflix, Amazon Prime and Sky. But they have better thought-out value propositions and pricing than Juicero. Spending $5–7 on a single juice is silly when you could spend that amount of money on fruit and veg from your local store and make oodles of drinks using a plastic squeezer – or just buy it ready made.
The Juicero is part of the “delivery-on-demand” trend made possible thanks to our online connected world. Scan the website of any Silicon Valley venture capital firm and you’ll find all sorts of startups innovating and occasionally reinventing the stuff you used to take for granted. One I do like is Feather – a service that lets you “subscribe” to your furniture. On the face of it, subscribing to a settee or lamp sounds crazy. But it works as a business and is ideal for people renting houses and flats in cities. Customers get to pick the latest fashions and have their furniture delivered. Plus, you can change your furniture if you get bored and it gets taken away when you move – all for just $19 per month. Sounds a “no-brainer” doesn’t it?
Health check
While talking with my friends, Theranos also came up in conversation. It’s a private health-technology firm founded in 2003 by Elizabeth Holmes, a 19-year-old Californian who modelled herself on Steve Jobs (including wearing black polo necks). Her company, which claimed it had revolutionized blood tests using only a fingerprick of the stuff, raised more than $700m from venture capitalists and private backers. Investors and the media hyped Theranos as a breakthrough product in the US diagnostic-lab market, which is worth more than $70bn a year. At its peak, the firm was valued at $10bn.
The turning point came in October 2015, when John Carreyrou, an investigative reporter from the Wall Street Journal, questioned the validity of Theranos’ technology.
The turning point came in October 2015, when John Carreyrou, an investigative reporter from the Wall Street Journal, questioned the validity of Theranos’ technology, leading to a string of legal and commercial challenges. Then, on 14 March 2018 Theranos, Holmes and former company president Ramesh Balwani were charged with fraud by the US Securities and Exchange Commission (SEC). The SEC said Holmes had made many false claims including one in 2014 that her company had annual revenues of $100m, when the actual figure was barely $100,000.
Despite those problems, I believe that Theranos had a great business model. If the technology actually existed to deliver its claims, the firm would be on to a winner. Imagine being able to diagnose hundreds of health issues on the spot in your local supermarket or pharmacy with a simple prick-of-the-finger blood test. Mind you, the firm won’t be going out of the spotlight: apparently it’s going to be turned into a Hollywood movie called Bad Blood with Jennifer Lawrence in the starring role.
So Theranos didn’t work out. Fraud is a bad business model. But plenty of firms with mad business plans do succeed. And that’s the point: no matter how crazy a business plan sounds, if it works out commercially, it’s not crazy at all.
An illustration of the 2D ice from atomic force microscopy images. Courtesy: Y Jiang
It’s well known that water vapour in the air can transform directly into solid ice on cold days, forming a thin layer on surfaces such as windowpanes or windshields. This commonplace process is, however, little understood, and a team of researchers in China and the US has found a new way it can happen. Using a non-contact atomic force microscopy technique combined with theoretical calculations, the team studied how ice grows in two dimensions on a surface. Their findings could prove useful in designing materials from which ice can be removed more readily, and their technique might also be extended to other 2D materials – especially those with multiple metastable and intermediate edge structures.
Low-dimensional water exists when water is confined at the interface between two solid surfaces. It is ubiquitous in nature and plays a critical role in many fields, including materials science, chemistry, biology and atmospheric science. Knowing its structure is thus very important.
2D bilayer hexagonal ice
In their work, the researchers, led by Ying Jiang and Enge Wang of Peking University, in collaboration with Xiao Cheng Zeng of the University of Nebraska-Lincoln, imaged a 2.5 Å-thick layer of 2D bilayer hexagonal ice (which corresponds to two water layers) that they had grown on a hydrophobic Au(111) surface at a temperature of around 120 K. The non-contact atomic force microscopy (AFM) technique they employed is based on a “qPlus” sensor that uses a mechanical tip functionalized with a carbon monoxide molecule.
Thanks to the ultrahigh flexibility of the tip apex and the weak nature of high-order electrostatic forces between the tip and the ice, the tip interacts only weakly with the water sample. This limits the damage to fragile structures that form at the edge of the ice, and enables the researchers to obtain images of unstable intermediate edge structures in which the H and O atoms of the water are distinguishable. To extend the imaging time, they cooled the samples down to 5 K once ice growth had stopped, freezing them before imaging them with AFM. In this way, they were able to reconstruct ice growth processes with atomic precision.
Zigzags and armchairs
Hexagonal ice is the main type of naturally-occurring ice on Earth, and this arrangement of water molecules is the reason why all snowflakes have six-fold symmetry. The basal plane of this form of ice has a structure that is very similar to 2D ice, and it can terminate in two types of edges, known as “zigzag” and “armchair”. Zigzag edges usually predominate, while armchairs are normally absent because of the higher density of unsaturated hydrogen bonds, which increases the energy of these edges, thus reducing their stability.
In the current work, the researchers found that armchair edges can in fact become stable when ice grows in two dimensions. What is more, the growth of these edges follows a hitherto unseen reaction pathway in which five- and seven-membered rings of water molecules are present in an intermediate stage of ice growth. This result, which Zeng’s team confirmed in computer simulations, is quite different to the six-membered rings previously observed in zigzag growth patterns.
2D ice I
The result confirms the existence of the first genuine 2D ice, which the researchers have dubbed 2D ice I. It also opens the door to explore other 2D ice phases in nature, Jiang tells Physics World. This 2D bilayer hexagonal ice was first predicted to exist in 1997, but a direct image of its atomic structure, which has a fully saturated H-bonding structure, was lacking until now. The way it grows may completely change our conventional understanding of low-dimensional ices and may guide designers of anti-icing and superlubricative materials in the future, Jiang says.
Ice removal is critical for structures such as wind turbines, which cannot function when they are covered in ice. Understanding the interaction between this water ice and surface will help us develop new materials to make this ice removal easier, he adds. Jiang also suggests that the group’s technique could be extended to study the growth mechanisms of other 2D materials, opening up a new avenue for visualizing the structure and dynamics of low-dimensional matter.
The researchers, who report their work in Nature, are now busy studying how 2D ice evolves to 3D, which has a wider relevance to ice formation and growth in general. “We expect that the 2D growth mode can persist up to a certain thickness, after which the ice layer will undergo a structural transformation from a stacked flat bilayer conformation to an interconnected buckled bilayer one – similar to how graphite changes into diamond,” Jiang says.
Optics meets 2D materials: a simple probe light beam exiting an optical moiré lattice can be chopped up into diffuse streaks or confined into a single bright spot, depending on the twist angle in the moiré lattice. (Courtesy: Fangwei Ye)
When two periodic 2D patterns are stacked onto each other with a slight twist, a third larger pattern called a moiré lattice emerges. The study of moiré patterns in layered crystals has captivated scientists ever since it was demonstrated that two sheets of graphene can run the gamut of electronic properties – including superconductivity, magnetism and Mott insulation – simply by varying the twist angle between the sheets. Recently, the optics community has joined the bandwagon and now a team of researchers, led by Fangwei Ye from Shanghai Jiao Tong University, has replicated the moiré effect in 2D lattices made of light.
The researchers “stacked” two 2D optical lattices by interfering two light beams. Before interfering, one beam was intercepted with two prepatterned masks that were rotated with respect to one another. The masks altered the beam’s intensity and phase, imprinting the patterns onto the beam like a stencil. The pattern is then projected onto a strontium barium niobite crystal. This moiré pattern can be tuned by the relative intensity of the interfering light beams and rotation angle between the mask patterns.
“Our work was inspired by the ‘magic’ angle in bilayer graphene,” Ye says. “We wondered, how will light waves respond if we stack two optical lattices? We’re the first to study the physics behind photonic moiré lattices.”
Wave localization
The researchers investigated how a third light beam, the probe, evolved when it was passed through the moiré pattern. At sufficiently high intensity contrasts between the first two moiré beams, the probe maintained its initial shape with little angular spread, a process known as wave localization.
Light waves undergo diffraction and diffusion, resulting in their tendency to spread out like the rays from a flashlight. The ability to counter this spreading is important for preventing information loss in optical communications.
Moiré lattices offer a new way to localize light in 2D, Ye and his team have found. This mirrors why electrons superconduct or freeze in their tracks in twisted bilayer graphene: flat energy bands. Photons in flat energy bands are packed into a narrow spectrum of energies, which supports only the modes that resist diffraction.
Moiré studies to come
Given that moiré lattices in light and 2D crystals share similar flat-band phenomena, the researchers recognize that their optical system can be used as a proxy to study moiré physics in 2D crystals. Moiré photons approximately follow the Schrodinger Equation, the governing rule of electrons. Optical systems are also easier to work with compared to 2D crystals, which tend to restructure into more energetically favourable configurations.
“That’s exactly why optics guys [referring to his team] can ‘simulate’ some physics for the condensed matter guys in the optical lab,” Ye said.
This flat-band phenomenon is not infallible. For instance, light localization is lost at certain twist angles. To make their localization strategy more robust, Ye and his team aim to generate solitons – solitary waves that do not change shape as they move – in their moiré lattices. In fact, Ye suspects that soliton formation might be uniquely easier in moiré lattices.
“Since there’s almost no diffraction in a moiré lattice, we don’t need strong nonlinearity (such as high laser powers) to form solitons. That makes life easier.”
From the Brexit referendum to Trump’s election victory, it’s fair to say that traditional polling methods have been way off the mark in some recent cases. An emerging research field known as “sociophysics” could help to explain and predict why political outcomes sometimes seem to come out of nowhere. At the core of this research is the idea that the dynamics of opinions obey discoverable universal quantitative laws and can be modelled in the same way that scientists model the physical world.
For an introduction to sociophysics see the video above. For a more in depth analysis you can read ‘The physics of public opinion‘ by science writer Rachel Brazil, an article originally published in the January 2020 issue of Physics World.
An international research team has shown, for the first time, that music therapy applied to preterm infants can influence the structural maturation of their auditory and emotional brain areas. This finding raises the possibility of creating specialized brain-oriented care for improving preterm infants’ outcome (J. Neuroimage 10.1016/j.neuroimage.2019.116391).
Premature birth happens when a baby is born before 37 weeks gestational age, instead of the typical 40 weeks. It accounts for 11% of the world’s live births and it is one of the leading causes of neonatal mortality. Very preterm (VPT) birth, which refers to babies born before 32 weeks of gestation, can potentially affect the child’s development. In fact, up to 40% of VPT infants experience neurodevelopmental impairments in childhood and 25% of them exhibit behavioural impairments, such as inattention, anxiety and socio-emotional problems.
MRI studies on premature babies show that by term-equivalent age, these infants’ brains are structurally different from those of healthy full-term born babies. This evidence raises the question whether this could be due to the fact that these babies are exposed to different noxious events and deprived of meaningful sensory inputs whilst in the neonatal intensive care units (NICU) during a period of critical brain development. One possible solution is the use of music therapy to provide meaningful sensory stimulation, with the aim of improving early brain maturation.
Music therapy during NICU stay…
To understand whether music listening could have an impact on infants’ brain development, a team headed up at the Department of Woman, Child and Adolescent, University Hospitals of Geneva, performed a neuroimaging study on VPT babies who underwent music therapy during their NICU stay. The results revealed a maturation effect of the neuronal pathways responsible for auditory and emotional processing.
Senior author Petra Hüppi with her research team. (Courtesy: Joana Rita Alves Sa De Almeida, Child Development Laboratory)
The team recruited 30 VPT babies, of whom 15 received music intervention during their NICU stay and 15 received the standard-of-care. An additional 15 full-term babies were recruited as a control group. The researchers administered music therapy five times a week, with each session consisting of an eight minutes piece composed by Andreas Vollenweider specifically for this study.
Three songs were used, depending on the state of the child (waking up, falling asleep or being active), and consisted of sounds of harp, snake flute and bells. All babies underwent MRI scans at term-equivalent age (37–42 weeks gestational age), during which the team acquired both T2-weighted (T2w) anatomical images and diffusion tensor images (DTI) as part of the protocol.
…increases brain structural maturation
To determine whether the music therapy had an effect on the structural development of the infants’ brains, the researchers analysed and compared the acquired scans. They extracted the microstructural properties of white matter fibres from the DTI scans and analysed these for the three groups. Additionally, they extracted amygdala volumes from the T2w anatomical scans, a brain region known to play a major role in emotion processing and previously shown to be significantly smaller in size in babies born preterm compared with full-term infants.
The overall results showed that the VPT infants who underwent early music intervention had an increased maturation in brain regions responsible for acoustic and emotion processing (including the acoustic radiations, uncinate fasciculus and amygdala), in comparison with those receiving standard-of-care. This conclusion was supported through findings in both MRI modalities (T2w and DTI).
The study revealed that exposing VPT infants to music therapy during their NICU stay can have a positive impact on the maturation of brain regions known to be altered by prematurity and which hold key roles in emotional processing. As such, the question of using music as a way of improving standard-of-care arises. In fact, the authors of this study think that their findings support “the clinical use of such an intervention in mitigating later social-emotional difficulties associated with prematurity”.
Currently, the researchers are assessing the reproducibility of their study on a larger cohort. This involves an increased number of exposures to music therapy, an additional MRI acquisition before the start of the intervention, and a clinical follow-up to investigate the neurodevelopmental and cognitive outcomes of these children at long-term.
Surprising predictions: In the 2005 French referendum on establishing a European Constitution, the “no” result was a surprise to many, but not to physicist Serge Galam who had predicted it using sociophysics. (Courtesy: Olivier Hoslet/EPA/Shutterstock)
In August 2016 the French theoretical physicist Serge Galam published a paper explaining Donald Trump’s unexpected victory in that year’s US Republican primary election. His model also suggested that Trump could win the November presidential election – a view not then supported by analysts or polls (International Journal of Modern Physics B31 1742015). Galam, who is seen by some as the father of the emerging field of “sociophysics”, is convinced that the dynamics of opinions obey discoverable universal quantitative laws and can be modelled in the same way that scientists model the physical world.
Now based at the Paris Institute of Political Studies in France, Galam has spent most of his career in physics departments studying disordered systems. As a student in the 1970s, he began trying to apply statistical physics to social models, but initially this work was treated with suspicion by other physicists. “People were very hostile,” Galam recalls. “Until today, I don’t really understand why.” His faculty head even confiscated Galam’s first paper in the area to try and prevent him publishing it.
Galam’s track record in predicting political surprises goes back further than Trump. In 2005, foreshadowing Brexit, he correctly predicted the “No” result in that year’s French referendum on establishing a European Constitution (55% of voters rejected the treaty on a turnout of 69%). Galam warned before the vote that “even starting from a huge initial majority of people in favour of the European Union, an open and free debate would lead to the creation of a huge majority hostile to the European Union”. He says these same dynamics were at play with Brexit.
Physics models
Over the last 15 years many physicists have become interested in using physics models to analyse trends in public opinions. Indeed, there is now a growing sociophysics community, which is ultimately trying to explain and predict why political outcomes sometimes seem to come out of nowhere.
The growing sociophysics community is trying to explain and predict why political outcomes sometimes seem to come out of nowhere
One of those involved is Daniel Stein, a physicist from New York University in the US whose research focuses primarily on randomness and disorder in magnetic materials. He is working with social psychologist Mirta Galesic from the Santa Fe Institute in New Mexico to develop what Galesic describes as “cognitively enriched models from statistical physics”. As is the case for many sociophysics models, Stein and Galesic are using the mathematical modelling of ferromagnetism as the basis of their work – specifically, the Ising model.
The Ising model is typically depicted as a 2D lattice with – in the case of magnetic materials – a particle at each point on the grid. Each particle can be in one of just two states: spin “up” or spin “down”. The spins like to line up in parallel with their neighbours because doing so lowers the overall energy of the system. However, this action competes against thermal noise, which injects fluctuations that tend to destroy order. At low temperatures, long-range ordering is established, but above a critical temperature the system becomes disordered. Turns out this basic model works for many physical systems and has gone way beyond magnetism. “It is probably one of the most intensively used models in physics,” Stein points out. In sociophysics, for example, it is used as a simple model of opinion dynamics, where networks of people are interacting and influencing each other.
Another model that has been applied to sociophysics is percolation theory, which is used in disparate fields from physics to epidemiology. One of those to follow this route is Hernan Makse – a physicist from the City College of New York. He has a background in the oil industry where percolation theory is used to understand how water flows when it’s pumped into rocks to extract oil. “This is a typical percolation problem because you are solving a problem of connectivity – in this case it is the percolation of water through a porous media, which is the rock,” he says. In a similar vein, Makse now uses percolation models to describe the diffusion of information and its impact on people’s opinions.
Modelling opinion
So can these models really have any relevance to human behaviour and how we form our opinions? Stein admits that social systems are far more complex than individual atoms or spins, but there are commonalities. “It’s not just the properties of the agent themselves but the interactions between the agents and this is where the statistical mechanics comes in,” he explains. “That’s because the physical mechanics is all about the interactions of individual particles and spins and external fields.” Just as with atomic spins, the way in which human opinions change is linked to our surrounding networks and can be modelled as any collective phenomena with interacting “agents”, be they people or atoms.
Opinion-dynamics models look at interactions within networks and how these are likely to cause change – usually between two opposing opinions such as voting for a left or right wing political party. The models create successive rounds of “updates” with opinions shifting after each round. “There is of course a fine line between using a model as a framework to guide your exploration and interpreting this too literally,” says Galesic. “They shouldn’t be taken as a direct description of humans and human society – of course humans are not particles living in crystal lattices!” But the models do seem to produce patterns of opinion spread that resemble the real world. Galesic continues, “These very simple models have found a way to derive these patterns using only a few parameters.”
When Galesic and Stein use a form of the Ising model, opinions are modelled as up or down spins that can change in line with their coupling to surrounding spins – the network of people to which they are exposed. The likelihood of a spin change will also depend on intrinsic predispositions such as political leanings that affect the likelihood of changing beliefs. “Every particle can have a different internal magnetic field, which in our case is an analogy for all their [prior] beliefs and values,” says Galesic. She explains the psychology behind this: “If believing in anthropogenic climate change is an up or down spin, my internal magnetic field would be my political orientation – so if I am a Republican I would be less likely to accept such a scientific fact at all.”
To test their theoretical framework, Galesic and Stein used opinion data on politics and health collected from 80 students in an undergraduate dorm at the Massachusetts Institute of Technology (MIT) (Physica A519 275). One aspect they were looking to test was the mechanisms by which agents – the students – changed their opinions between successive surveys during 2008 and 2009. In sociophysics models there are different ways this can be modelled algorithmically, one of the most common being the “majority rule” where a person adopts the views of the majority within their local network. Alternatively, with the “expert rule”, people follow a particular individual whose opinion they value. Galesic and Stein showed that it’s not possible to use one rule in all circumstances and that how people behave is context-dependent. So in their study both rules were being used, depending on the kinds of questions being asked. “Things like playing sports once a week, eating salads versus unhealthy foods – we found that the majority rule worked best. So if your friends were going to exercise, you might too,” says Stein. But when asked about scientific questions, people are more likely to follow someone they consider an expert.
The types of networks modelled also make a difference. In physics it is common to use a Euclidean lattice (a 2D array of points), but that is not realistic for social networks. One model often used instead is the small-world network, which creates clusters and results in strangers being closely linked. In the MIT study, Galesic and Stein constructed real social networks from survey responses and by using Bluetooth and WiFi signals that indicated participants’ actual proximity. In fact, this Bluetooth network turned out to provide the most accurate model – even though it didn’t always match people’s own ideas of their networks.
Phase transitions and tipping points
The crucial aspect of the models borrowed from physics is their ability to identify phase transitions – when a system shifts from one phase of collective organization to another. A common example is how water changes state as a function of temperature and pressure. “These transitions are what we see in societies as well,” says Makse. However, the phase changes are not caused by temperature or pressure, but by changing distributions of opinions, which can be caused by various societal factors.
With colleagues from Brazil, Makse has looked for these kinds of phase transitions in real public opinion taken from over 10 years of data from the Pew Research Center in the US, on a range of political issues from immigration to abortion. The researchers were particularly interested in looking at how opinion becomes more extreme and if the numbers of people holding extreme opinions can cause “tipping points”, which – once reached – cause nonlinear changes and even greater take-up of extreme views across the whole population.
The conditions for a society to occupy moderate, incipient extremist, or extremist phases varies with the fraction of those who hold extreme views
Makse and his colleagues have created “social phase diagrams” that show the conditions for a society to occupy moderate, incipient extremist, or extremist phases (where a majority hold extremist views), and how this varies with the fraction of those who held extreme views. One example of such a transition was in opinions on how people rated the economic situation in various countries including France, Italy, Greece and Spain. Rather than overall views changing linearly as the proportion of those who considered the situation to be “very bad” increased, a point arose where the relationship between overall opinion and the proportion of those expressing extreme opinions was no longer linear. Instead, overall opinion on the economic situation became more negative than expected (figure 1). This occurred just after the 2009 European sovereign debt crisis (Sci. Rep. 5 10032).
1 A social tipping point
(Courtesy: Adapted from M Ramos, J Shao, S Reis et al. 2015 Sci. Rep.5 10032 )
Outcomes of 260 polls in 59 countries, where f is the fraction of people sharing an opinion (rating the economic situation in their country as either “bad” or “very bad”), and fe is the fraction sharing an extreme opinion (rating the economy only as “very bad”). The dotted line depicts the linear behaviour expected for a non-interactive group. The polls in Spain are highlighted with black squares to demonstrate time evolution. There is a clear tipping point where fe diverges from expectations in 2009.
A surprising finding was that to reach the tipping point, in some cases only 20% of a population need initially hold extreme views to sway others and eventually create an extreme majority. This seemed difficult to understand from both a physical and psychological perspective until the researchers factored in differences in people’s natures, similar to Galesic’s “internal field” concept. Specifically, they were able to model the onset of cascades of extreme views when they included “stubbornness” – clearly some people are less likely to change their minds than others, which becomes a crucial ingredient to the model. “We played with these thresholds to see how many stubborn people we needed in the model so that [people’s] ideas do not change,” says Makse, “All sorts of nice dynamics come up.”
Similarly, Galam’s model of opinion dynamics also incorporates stubbornness on each competing side, and is consequently able to predict the value of the corresponding tipping point with any initial proportion of opinions in a population. As few as 2% more stubborn agents on one side puts the tipping point at a very low value of around 17%, which leads to the unfortunate conclusion that to win a public debate, what matters is not convincing a majority of people from the start, but finding a way to increase the proportion of stubborn agents on your side (Physica A389 3619).
Majority rules yielding minority spreading
Galam has spent most of his career explaining the kind of unexpected results that seem to periodically arise in politics and public debate. He ascribes this to “minority opinion spreading”, where an initial vocal minority turns a majority towards its position. Rather than being a response to a specific political issue, Galam says this is a sociological phenomenon that can be modelled along a similar path used in statistical physics to reach global behaviour from local interactions.
Galam’s model uses a majority-rule-based approach, with repeated local small group discussions occurring within the network. Think of these as the real-life lunches, coffee breaks and dinner parties where people discuss the news. During these encounters, people change their views to follow the local majority, unless faced with an equal opinion split within the group. In the latter case, people are likely to support the position that offers the least change or is in tune with their unconscious prejudices (Eur. Phys. J. B25 403). As a result, over a number of updates and continuing debates, opinions shift to support the status quo or the opinion in tune with the activated prejudices.
An opinion that starts in the minority can quickly spread as long as it is above a base threshold. Galam suggests this threshold is just above 23% for groups of four people. In other words, although some groups of four start with no-one holding the minority view, over a series of updates a person will encounter split groups, causing a switch to the minority position. And after each update, the proportion of people holding the minority view will increase. Getting to a final point happens remarkably quickly. “In all cases the number of updates is less than 10,” Galam says. However, transposing that to real time would need an understanding of the speed of each update, which will increase, for example, as an election comes closer. One update could be between a few days and one to two weeks.
Strange surge: In the 2016 US presidential elections, evidence suggests that Donald Trump’s outrageous statements, though initially seen as repellent by most voters, managed to activate hidden or unconscious prejudices. (Courtesy: Suzi Altman/Zuma Wire/Shutterstock)
In the case of the 2016 US presidential elections, Galam says the prevailing factor was peoples’ “frozen prejudices”. He argues that Trump’s outrageous statements, though initially seen as repellent by most voters, managed to activate their hidden or unconscious prejudices. First, many Trump supporters shifted to Hillary Clinton, rejecting his statements with great outrage, leading to a decrease in support. But the initial outrage led to more public debates with an automatic increase in the number of local ties. At those points, “it’s like flipping a coin, but with a coin biased along the leading prejudice”, Galam says. Then many voters started to swing in favour of Trump.
However, the model does require a minimum support for Trump above the corresponding tipping point value. “To stay above the threshold, it is thus a combination of being able to modify the unconscious reality by activating old frozen prejudices, at the price of losing support, but still preserving a nucleus of people who are consciously OK with the prejudice,” explains Galam. Although the threshold value can be as low as 17%, the second requirement was not satisfied everywhere, which explains why Trump did not win in all US states.
Opinion polls, big data and statistical models
If Galesic has a criticism of past work in sociophysics, it is that many models were designed in isolation without reference to empirical measurements of real social groups. She carried out her own longitudinal survey of US public opinion to test her model, surveying 94 people in four waves in 2016. Participants were asked about their beliefs on guns, terrorism and vaccinations, and the survey followed how their views changed over the four waves. Galesic factored in information on peoples’ networks and their initial values to create a quantitative understanding of the likelihood that they would change their opinion. The resulting model could reproduce the real-world pattern of belief changes found in the survey.
For both the 2016 US and 2017 French presidential elections, Galesic asked interviewees not only who they would vote for, but also the voting intentions of other individuals in their social circle. Using this information calculated the likelihood that any individual would change their mind. Adding this information did seem to improve the accuracy of predictions (Nature Human Behaviour2 187) and Galesic also used this approach for the 2018 US mid-term elections, predicting a 9% Democratic lead over the Republicans. This turned out to be very close to the actual Democratic lead of 8.6% and a better prediction than just asking people about themselves (12%).
Physicists and computer scientists are also looking at social-media data, which is being used together with machine learning to make predictions about people’s opinions and voting intentions. This might in fact provide quicker and cheaper information than opinion polls. Prior to the 2016 US presidential elections, for example, Makse analysed over 170 million Tweets, using a combination of network analysis, natural language processing and machine learning (Scientific Reports8 8673). The results proved to be remarkably similar to the New York Times National Polling Average – an aggregate of hundreds of independent polls, but 10 days ahead – without the lag-time needed to complete a survey.
This approach does have its limitations says Makse. “You are just making statistical inferences, but you do not know why – it’s a black box, you won’t learn anything about society. If you want to actually change the situation, you need to do that with physics and understand the mechanism.” As he adds: “For a physicist it is more fun to find the fundamental laws behind those predictions.”
If you want to actually change the situation, you need to do that with physics and understand the mechanism
Galam agrees that statistical physics can provide unique insights, particularly when it comes to nonlinear effects and “phase changes”. “Big data will not show you a sudden change that’s rapid before it occurs because it is not there beforehand – you need a theory to do that.” Galam sees the unique power of sociophysics as its ability to predict the sort of nonlinear behaviour that opinion polling will never predict and that often triggers large political or social changes. “I think we will be able one day to predict on solid ground, co-operative phenomena like an election, or a social movement over an event timescale of a few weeks or a few months,” he says.
Opinion forecast: As sociophysics models improve, they should be able to reliably forecast the sort of nonlinear behaviour that opinion polling never predicts. (Courtesy: Shutterstock/Sandor Szmutko)
With Taksu Cheon, a physicist from Kochi University of Technology in Japan, Galam has succeeded in incorporating all his models into one unique update equation. “We now have all the parameters together so I have a kind of universal formula,” says Galam. “It gives me – for any proportions of inflexible individuals, contrarians, distribution of prejudices and for any size discussion groups – the relevant tipping point, which will eventually drive the final outcome, so this is a very strong step forward.” All possible opinion dynamics outcomes are now included within a phase diagram embedded in a parameter space of six dimensions (arXiv:1901.09622).
And while social circle perceptions have improved prediction, Galesic says they are a step away from creating a model that could predict how each individual surveyed is likely to behave. “That’s the ultimate frontier in social science, to try to predict human behaviour,” she suggests. But if this was ever possible, what does it tell us about our own free will? Perhaps our opinions are not formed as freely as we think. “We are taking in dynamics that we don’t see,” says Galam. “Only by identifying our hidden determinisms will tomorrow allow us to reach more free will, so this is why I think it’s very important to push in the sociophysics direction.”
As for Galesic, she thinks this kind of research can actually help us to have more freedom. “If these models show that our social contacts and our pre-existing moral values influence how likely it is that we will believe a potentially useful piece of information, then we should periodically re-examine who we socialize with and what we believe in.”
A laser-driven electron accelerator that is integrated on a chip has been created by Jelena Vučković and colleagues at Stanford University. The device was developed using inverse design algorithms and comprises a highly complex silicon waveguide that is driven by near-infrared pulses. Soon, the researchers hope that the technology can be used to accelerate electrons to energies of about 1 MeV.
Conventional accelerators use radio-frequency (RF) radiation to boost charge particles to relativistic speeds. While incredibly useful in both science and medicine, such accelerators are large and expensive – putting them out of reach of many universities, research institutes and hospitals.
Dielectric laser accelerators (DLAs) offer a promising way of creating smaller and cheaper electron accelerators. They work by firing pulses of visible or near-infrared light onto nanostructures such as silicon pillars or gratings. The nanostructures are in an evacuated channel through which a beam of low-energy electrons is sent to be accelerated by the light. By using radiation at much shorter wavelengths, DLAs can be around 10,000 times smaller than conventional RF accelerators.
Design challenges
However, DLAs also require bulky optical setups and this limits their scalability and robustness. Vučković’s team has shown that this problem can be overcome by using a photonic waveguide, which would enable DLAs to be integrated with compact photonic circuits. The challenge for the team was how to design a waveguide that is not degraded by scattering and reflection processes – something that was proving far too complex for a human to design.
The team addressed this issue by creating an “inverse design” algorithm, through which they could simply specify how much light energy they want to be delivered to electrons. From this input, the software calculates the precise silicon structure needed to deliver photon pulses to an incoming electron beam at just the right times and angles.
The bizarrely-shaped prototype waveguide suggested by the algorithm was manufactured on a silicon-on-insulator chip using electron-beam lithography. The accelerating nanostructure was also created using inverse design. When the device is operated, pulses of light interact with electron pulses, boosting the kinetic energy of the electrons by 1.21 keV over a distance of just 30 microns.
As a fully integrated photonic circuit, the researchers’ DLA is highly robust and scalable. Vučković and colleagues now hope that these characteristics will enable them to use around 1000 accelerator stages to reach energies of 1 MeV – around 94% the speed of light – by the end of 2020.
One use of the technology that the team is working on is to channel electrons from the accelerator through an evacuated catheter and into the body. This would allow radiation to be delivered directly to tumours while minimizing radiation damage to healthy parts of the body.
