– Losing is not an excuse for quitting, it simply meant I hadn't worked hard enough
As told by Eva Y. Andrei

– Math, algebra and riddles as soon as I could speak

I wasn't exactly planned. My parents felt they were too old and too poor to bring a second child into the uncertain world of communist Romania. That they decided to take a chance on me anyway gave me my first life lesson: to welcome the unplanned and befriend serendipity.

The living room of our government-assigned, third-floor Bucharest apartment—shared with a revolving string of other families—was a constant hive of activity. Between my family’s looms clattering as they knocked out shawls and carpets, and our neighbor hammering away at his furs, the air was thick with industry. We children built castles out of chairs amidst the clinking of dinner plates, contributing to a never-ending cacophony. To keep pace with the grown-ups, I spent hours practicing my own "craft," hammering nails into my small wooden chair until the surface was transformed into a shiny, perfect lattice of nailheads. It was, perhaps, my first fascination with the order of crystals.

Starting from scratch
In fifth grade, my family fled Romanian antisemitism to join my grandparents in Israel. While my parents struggled to find their footing, I stayed in my grandparents' tiny house in Ramat Hachayal. Enrolled in the local school without knowing the language, my classmates helped me piece together everyday Hebrew while I spent evenings in self-study out of arcane texts I found in my grandparents’ home. By the time we moved into a government-assigned, Pardes Rubin barrack in sixth grade, my Hebrew was fluent and school had become easy.

Competitive swimming taught me that losing is not an excuse for quitting; it simply meant I hadn't worked hard enough.

I spent my free time training and competing with the Hapoel Tel Aviv swim team. The pool became my second home, and my rumbunctious, hilarious, and fun-loving teammates became my second family. Competitive swimming taught me that losing is not an excuse for quitting; it simply meant I hadn't worked hard enough. I learned to get up and train until I hit the top spot. That mindset worked until 10th grade, when younger, faster girls began to outpace me. Realizing that raw physical effort had its limits, I decided it was time to change focus.

One of seven women in a class of a hundred
That summer I joined the Soreq Science Youth Camp, a program designed to train future scientists. We were a cohort of teenagers living on a rustic campus, spending our mornings in the lab and afternoons studying physics and mathematics under the supervision of volunteer scientists. It was my first time working on "real" physics—measuring the Mössbauer effect, studying wave interference, and experimenting with nuclear magnetic resonance. While these camps are often credited as the nurseries for Israel’s future as a Startup Nation, for me, it was where the idea of becoming a scientist first felt real.

I enrolled in Tel Aviv University’s Physics and Mathematics program, an intense three-year track where only a small fraction of my first-year peers reached graduation. Being one of seven women in a class of a hundred never really bothered me. It was there that I met Natan, the love of my life, and the most brilliant student in our cohort. By the time we graduated, we had married and begun our lives together as physicists: he as a theorist and I as an experimentalist.

Undergraduate paper that continues to be cited
For my undergraduate thesis at the Soreq Research Institute, under the guidance of Yehuda Zuss, I used electron irradiation to create nitrogen vacancies in hexagonal Boron Nitride (hBN). My first paper grew out of this work, describing the discovery of a radiation-induced 3-Boron paramagnetic center. At the time, hBN was mainly used industrially for its exceptional thermal stability and excellent lubrication; years later, it would serendipitously emerge as the "insulating cousin" of graphene and an indispensable building block of moiré materials. It amazes me that this early undergraduate paper continues to garner citations.

We moved to the United States for our graduate studies, settling into the neighboring academic hubs of Rutgers and Princeton. While Natan focused on high-energy theory at Princeton, I joined the experimental condensed matter group at Rutgers. My research, in the Glaberson group centered on superfluid helium—a frictionless quantum fluid that can flow through the tiniest pores without losing momentum. To probe the limits of this state, I built a rotating dewar to spin the fluid to high velocities and monitor the breakdown of quantum motion using “fourth sound” detection.

My research in the Glaberson group centered on superfluid helium—a frictionless quantum fluid that can flow through the tiniest pores without losing momentum.

Change of direction
In one of the weekly colloquiums at Rutgers I heard a talk by Mike Grimes from nearby Bell-Laboratories that changed my direction. He described "spraying" electrons onto the surface of superfluid helium to form a two-dimensional electron layer. By squeezing these electrons electrostatically, he could order them into a perfect lattice - a "Wigner crystal". I was transfixed; it was the microscopic version of the nailhead array I had hammered into my wooden chair so many years before.

However, Grimes’s electron crystal was classical, governed by the competition between Coulomb ordering and thermal fluctuations. It lacked the quantum fluctuations that Eugene Wigner had predicted in 1930, which should cause the crystal to melt as it is squeezed to higher densities. Hooked by this challenge, I turned down lucrative industry offers to pursue postdoctoral work on the quantum 2D electron crystal at Bell Labs. At the time, Bell Labs was a vibrant, innovative cradle of science; I shared a corridor with six future and current Nobel laureates including Philip Anderson and Steven Chu.

To reach the densities required for quantum melting, the standard helium setup was useless; the highly charged surface would simply collapse, letting the electrons sink to the bottom of the cell. I needed a different approach. I realized that thinning the helium down to a sub-micron, film would stiffen the surface enough to sustain the necessary charge. I built a bellows-based apparatus to gently reduce the helium depth from bulk to a thin film. The experiment did not work as expected. Instead of quantum melting, the electrons interacted with corrugations on the underlying substrate, dressing themselves in the surrounding helium to become heavy objects known as polarons. Nevertheless, this was the first time a polaronic transition had been observed, and the work was published in Physical Review Letters.

The Eureka moment
While reporting these results at a conference in Santa Fe, I met Francis "Tito" Williams, the inventor of an ingenious radio-frequency (rf) technique for probing 2D electron systems. He made me an offer I could not refuse: to join his Saclay group near Paris as a visiting scientist. Using Tito's rf technique we measured the shear modulus of the Wigner crystal, its specific heat and thermal conductivity. Our experiments established that the melting of the classical 2D electron crystal—a subject of long-standing debate—is a continuous transition mediated by the unbinding of dislocation pairs, precisely as predicted by the Kosterlitz-Thouless mechanism.

These were important results, but they did not get us closer to the elusive quantum transition predicted by Wigner. It was clear to me that electrons on helium could not be pushed into the quantum regime; we needed a new approach. The only other 2D electron system at the time, clean enough to consider, was trapped at the interface of GaAs/AlGaAs heterostructures grown by molecular beam epitaxy. The problem was that the electron density in this system was far too high, placing it deep in the quantum liquid regime, nowhere near the Wigner transition. Reducing the density wasn't an option, as the electrons would get stuck on interface defects long before they could form a crystal. It looked hopeless.

What if we applied a strong magnetic field to squeeze the electronic wavefunctions, reducing their overlap enough for the Coulomb energy to win out and drive crystallization?

Then came a eureka moment: what if we applied a strong magnetic field to squeeze the electronic wavefunctions, reducing their overlap enough for the Coulomb energy to win out and drive crystallization? To get there, we needed a massive magnetic field, a high-mobility GaAs/AlGaAs heterostructure, and a way to detect the transition. While our rf technique could be adapted to this experiment, heterostructures were outside our expertise.

Bernard Étienne from the nearby CNRS lab, provided us with a high mobility GaAs/AlGaAs sample, and we booked a coveted time slot at the Grenoble High Magnetic Field Laboratory, which had just reached a world record of 31.6 T. With the help of cryogenics wizard, Patrick Pari, we built a specialized, portable dilution refrigerator. It featured a plastic mixing chamber designed to survive a violent magnet quench and was compact enough to fit inside the trunk of Tito’s Alfa Romeo for the drive from Paris to Grenoble. We spent a grueling, exhilarating week working uninterrupted 24-hour days, setting up, battling technical setbacks, and finally taking data. By the week’s end we had observed the unmistakable signatures of the magnetically induced Wigner transition. It was finally time to catch up on sleep and celebrate. For this work our team won the prestigious Physics prize awarded by the French Commission of Atomic Energy (CEA).

– What we saw blew us away
I returned to the United States to join the faculty at Rutgers University, establishing my own laboratory dedicated to pushing the boundaries of low-dimensional electronic systems. Moving from a researcher to a professor meant building an infrastructure from scratch, but my time in the trenches—from designing bellows-based cells to portable fridges—had made me comfortable with setting up a new lab.

With the tenure clock ticking, however, I had to make a pragmatic choice. Fascinating as Wigner crystals were, the experiments were too hard and the publication pace too slow for attaining tenure. I decided to pivot to the much more active field of vortices in superconductors, where progress was faster and publications more abundant. Superconducting vortices are thin filaments of magnetic flux that penetrate a superconductor; much like electrons, they can organize into a vortex crystal when packed closely together.

I realized that to stand out, we needed to do something entirely different.

Because the field was crowded with dozens of researchers studying the collective motion of vortices, I realized that to stand out, we needed to do something entirely different. If we could take atomic-scale snapshots of individual vortices as they entered the material and crystallized, we could unlock physics others could only theorize about. To accomplish this, we needed a low-temperature scanning tunneling microscope (STM) capable of operating inside a high magnetic field. No commercial instrument manufacturer could provide one at the time, leaving us entirely on our own. We already possessed a low-temperature setup with a 15 T magnet; we just needed a microscope head to fit inside the magnet bore and the measurement electronics. Alongside my then-postdoc Guohong Li—who remains my close colleague and collaborator to this day—we built the entire instrument from scratch on a shoestring budget. To my knowledge, it was the first high-field, low-temperature STM of its kind in the world.

As any STM builder knows, the first order of business is calibrating the piezo motors to atomic precision. We chose a graphite sample—the gold standard for calibration due to its precise 0.142 nm carbon-carbon bond length. Once we achieved atomic resolution, we measuerd the differential conductance spectrum to test how the microscope performed under a magnetic field. What we saw blew us away. Instead of the familiar, evenly spaced Landau levels characteristic of a conventional, non-relativistic electron system, the spectrum revealed a series of prominent peaks at energies that presented a square-root dependence on both the level index and the magnetic field. It was entirely wrong for normal electrons.

Hindsight in a new light
I then recalled learning about just such a square-root dependence at the 2004 APS March Meeting in Anaheim, California. Having grown tired of vortices, I had wandered into various sessions outside my main focus to search for a new research direction. I happened to stumble into a small room where a dozen attendees were listening to Andre Geim—who would later win the Nobel Prize for this work— speaking about a new two-dimensional material he called "graphene," a single atomic layer of carbon exfoliated from graphite. At the time, I found it fascinating, but being so far removed from my area of expertise, I simply filed it away as a beautiful piece of physics.

The pieces suddenly clicked: we were looking at the Landau levels of ultra-relativistic electrons. But how was this possible?

Remembering Geim’s talk, the pieces suddenly clicked: we were looking at the Landau levels of ultra-relativistic electrons. But how was this possible? Our sample was bulk graphite, not graphene. Here again serendipity played a role: because we had purchased the least expensive grade of graphite available, the sample’s layers were imperfectly bonded; we were, in fact, imaging a naturally occurring, single layer of graphene that was floating, electronically decoupled, on top of the graphite surface. From that point on, we dropped all vortex work and focused our entire laboratory's efforts on studying graphene.

This was the first time that Landau levels had been observed with an STM, and because the result was so new, the road to publication was nightmarish. We encountered an exceptionally hostile reviewer who tried to derail us, forcing us to carry out an endless series of control experiments to prove our point. After nearly a year of fighting this battle, we withdrew the manuscript from Nature and published it in Physical Review Letters.

This was the first time that Landau levels had been observed with an STM, and because the result was so new, the road to publication was nightmarish.

– A delicious twist of irony
My theory colleagues at Rutgers, who at the time were focused on strongly correlated electron systems, showed little interest in graphene. The prevailing consensus was that interactions in graphene were too weak to display any interesting, correlated phases. Indeed, while the integer quantum Hall effect could be readily observed, its fractional counterpart—the hallmark of strong correlations—consistently eluded observation. We decided to take a closer look.

Based on our STM measurements, we knew that SiO2, the standard substrate of the time, was far too invasive, creating charge pools that masked graphene’s intrinsic electronic properties. We had to get rid of the substrate. By chemically etching a trench underneath the graphene flake, we isolated it from substrate interference. However, suspending the graphene restricted our sample to an unprecedentedly small scale, where standard four-lead Hall bar measurements failed because the Hall voltage was shorted out by hot spots near the contacts. To circumvent this, we replaced the traditional geometry with a two-lead configuration designed to avoid these hot spots.

This gamble paid off rewarding us with the discovery of the fractional quantum Hall effect (FQHE) in graphene. Not only was it present, demonstrating beyond a doubt the strongly correlated nature of these ultra-relativistic electrons, but it also turned out to be remarkably robust. Whereas observing the FQHE in conventional GaAs/AlGaAs heterostructures required massive magnetic fields exceeding 20 T and millikelvin dilution refrigerator temperatures, in graphene it already appeared at a comfortable temperature of 0.3 K in a modest field of 2 T readily accessible with standard laboratory equipment. We published the discovery in Nature. In a delicious twist of irony, Science magazine—whose editors had refused to even send our manuscript out for peer review—selected our paper as one of the Top 10 Scientific Breakthroughs of the Year for 2009.

Our almost perfect Landau level spectrum showed a mysterious gap at charge neutrality that kept me awake nights. This could have been a straightforward consequence of the underlying graphite substrate breaking graphene’s sublattice symmetry. Or, more tantalizingly, it could be evidence of "magnetic catalysis"—a phenomenon from quantum field theory where an external magnetic field triggers spontaneous symmetry breaking. This question was the spark that launched our line of work ultimately leading to the Kavli Prize.

Our almost perfect Landau level spectrum showed a mysterious gap at charge neutrality that kept me awake nights.

A tall order
To find the answer, we first had to get rid of the graphite substrate. However, suspending the sample directly from electrical leads—the technique we developed for the FQHE measurements—was out of the question; the resulting structure would be far too small and fragile for an STM experiment. Instead, we needed large-area graphene samples mounted on a transmission electron microscope (TEM) grid. The rigid mesh of the grid could provide the necessary structural support, while its microscopic holes would leave regions of graphene suspended cleanly in mid-air.

It was a tall order. Synthesizing large graphene sheets required chemical vapor deposition (CVD), and transferring those sheets onto a TEM grid required infrastructure that we didn't have. I contacted Jing Kong at MIT, one of the world’s leading experts in CVD growth, to see if her group could synthesize the large-area samples, and I emailed Andre Geim to ask for his advice on how to mount them. Geim suggested that his postdoc, Konstantin Novoselov, could help us with the TEM mounting. And so, this precious sample conceived in a chemical furnace at MIT, journeyed across the Atlantic to Manchester to be mounted on the grid, and finally back to the United States to be examined by our Rutgers STM.

But what we saw was a complete surprise—not even close to what we expected. The STM topography showed bold triangular lattices with periods orders of magnitude larger than that of atomic graphene. It turned out that, unbeknownst to us, Jing’s student had used a nickel substrate rather than the standard copper to grow the sample. As a result, in a stroke of pure serendipity, instead of monolayer graphene, the sample contained numerous twisted bilayers, each with a different twist angle.

"The magic angle"
What we were observing were moiré patterns. While these patterns caused by twisted layers had been known to exist in graphite since the early days of STM, in all those years no one had looked to see how they affected the electronic properties. We discovered that moiré patterns caused a dramatic reconstruction of the electronic band structure. Instead of the clear, V-shaped spectra we observed on monolayer graphene, the twisted bilayer spectra displayed two large peaks—van Hove singularities (VHS)—flanking the charge neutrality point. Because our sample contained many different twist angles, we were able to track the evolution of the VHS as a function of the angle, revealing that their energy separation steadily decreased as the moiré period grew with the shrinking twist angle.

We discovered that moiré patterns caused a dramatic reconstruction of the electronic band structure.

At a twist angle of 1.07°—coined the “magic angle” in a 2011 paper by my Kavli Prize co-recipient Allan MacDonald—the VHS merged into one giant peak, signaling what is now known as a “flat band” where the quenched electron kinetic energy provides fertile ground for the emergence of correlated states. Indeed, we observed a clear signature of such a state—a charge density wave—accompanied by a spectral gap when the Fermi level entered the VHS.

– I couldn’t help but suspect what it might be about
I relayed our results to Antonio Castro Neto, the only theorist at the time to have studied the band structure of twisted bilayer graphene. Excited by the data, Antonio and his colleagues calculated the twist angle dependence of the VHS, finding a perfect match to our experimental results. In the final paragraph of our 2009 Nature Physics paper, we stated: “This opens exciting opportunities for inducing and exploring correlated electronic phases in graphene.”

In the final paragraph of our 2009 Nature Physics paper, we stated: “This opens exciting opportunities for inducing and exploring correlated electronic phases in graphene.”

Unfortunately, the base temperature of our experiment—4.2 K—was a bit too high to observe the superconducting state later reported by the MIT group led by Pablo Jarillo-Herrero, my co-recipient of the Kavli Prize, whose 2018 discovery of superconductivity in magic-angle twisted bilayer graphene officially launched the twistronics revolution.

On March 23, 2026, when I received a Zoom invitation from Eirik Lislerud to “discuss a special symposium in Nanoscience as part of the Prize Week,” I couldn’t help but suspect what it might be about.

I owe a huge debt of gratitude to my amazingly brilliant, dedicated and creative students and postdocs who worked alongside me enduring the long nights, the instrument quenches, and the peer-review battles before moving on to establish their own distinguished careers. This prize belongs to all of us—it simply could not have happened without you!