When you swipe your finger across a smartphone touchscreen, the transparent panel beneath it is coated with a conductive film—typically indium tin oxide (ITO) or fluorine-doped tin oxide (FTO). Solar cells rely on the same kind of glass. To get the best performance out of a semiconductor thin film deposited on top for converting light into electricity, high-temperature processing to improve crystal quality is unavoidable.
This has long created a contradiction. Fully developing the crystallinity of a semiconductor thin film typically requires temperatures above 800°C, but FTO glass and its conductive layer start to degrade above 500–600°C. Put the whole stack in a furnace, and the substrate fails before the film ever reaches its target temperature. Stop at a temperature the substrate can survive, and the film is left riddled with crystal defects, unable to efficiently convert light into electricity.
This constraint becomes even more serious when it intersects with a textbook concept in materials science: polymorphism. The same chemical composition can have wildly different properties depending on how its atoms are arranged—diamond and graphite, both pure carbon, are the classic example. In many semiconductor materials, the phase with the most desirable properties is only stable at high temperature; cool it down slowly, and it reverts to the ordinary, less useful phase. Bismuth oxide () is one such case. Multiple studies since 2010 have shown that while the monoclinic alpha phase (bandgap of about 2.8 eV) is stable at room temperature, the tetragonal beta phase (bandgap of about 2.5 eV) absorbs a broader range of visible light and shows higher photocatalytic activity. But the beta phase is metastable, and until now there was no way to create and preserve it on conductive glass.
It's Not the Total Energy, but the Delivery Speed, That Determines the Phase
A team led by Shahar Artzi and Dr. Ronen Gottesman at the Hebrew University's Institute of Chemistry broke through this double bind using millisecond-scale light pulses. In a paper published in Small Structures on July 28, 2026, the team used a technique they call flash photonic heating (FPH), firing intense pulses of white light—each lasting roughly 0.1 to a few milliseconds—at a thin film on FTO glass. The film absorbs the light directly and heats up almost instantaneously, and cooling finishes before the glass substrate beneath it has any time to catch up thermally.
The results were striking. The film reached roughly 2,000°C, while the FTO glass underneath stayed below 100°C—a temperature gap that's impossible to achieve with a conventional furnace, even in principle. Heating rates reached up to 10 million °C per second. For comparison, a typical furnace heats at around 10°C per minute—orders of magnitude slower.
An even more notable finding: even when the total energy delivered by a pulse was held constant, simply changing the delivery speed (the combination of pulse width and intensity) switched which phase formed. Longer pulses produced the stable alpha phase, while short, intense pulses locked in the metastable beta phase at room temperature. This isn't thermodynamic equilibrium at work—it's kinetic control. Cooling completes before the atoms have a chance to "want to go back," trapping a structure that would normally vanish.
As Dr. Gottesman put it in a press release: "The idea is to heat and cool the material so quickly that we can trap it in a crystal structure that would normally disappear... it gives us access to useful properties that are hard to preserve with conventional heating methods."
| Item | Conventional (furnace heating) | This study (FPH) |
|---|---|---|
| Film temperature reached | Up to 500–600°C (limited by FTO substrate) | ~2,000°C |
| Substrate temperature | Nearly identical to film | Below 100°C |
| Heating rate | ~10°C/min | Up to 10⁷ °C/sec |
| Cooling rate | Slow (atoms revert to equilibrium phase) | Extremely fast (locks in metastable phase) |
| Phase selectivity | Only thermodynamically stable phase | Alpha or beta phase selectable via pulse conditions |
| Reversible switching | Not possible (substrate can't survive it) | Demonstrated repeatedly on FTO glass |
A Color Change and a 50-Fold Jump in Current
The phase transition was visible to the naked eye. The alpha-phase film appears pale gray, but upon transitioning to the beta phase, it turns a vivid yellow. With a bandgap of about 2.5 eV, versus about 2.8 eV for the alpha phase, the beta phase absorbs a wider range of visible wavelengths. It captures light that the alpha phase would simply let pass through, boosting photocurrent (the current generated by light) by 10 to 50 times. The absolute photocurrent values used for this comparison haven't been disclosed in the materials published so far. The magnitude of the improvement varied depending on how the film was prepared, but in every case the beta phase substantially outperformed the alpha phase.
"We haven't changed the chemical composition at all," Dr. Gottesman explained. "We're just controlling how the atoms are arranged, and that dramatically changes how the material responds to light." The closest everyday analogy is a blacksmith quenching steel: cool it fast enough, and the atoms don't have time to settle back into their preferred arrangement. The team is doing essentially the same thing with light, on a millisecond timescale, on the exact same kind of transparent conductive glass found behind a touchscreen.
It's worth noting this isn't the team's first foray into this approach. While at the Helmholtz-Zentrum Berlin (HZB), Dr. Gottesman worked in Professor Roel van de Krol's group applying rapid thermal processing (RTP) to oxide thin-film photoelectrodes. In a 2022 paper published in ACS Energy Letters, the group flash-heated , , and thin films on FTO substrates to as high as 850°C, improving crystallinity without damaging the substrate. They also achieved a then-record photocurrent density of 1 mA/cm² in - using RTP. The current study builds directly on that work, extending RTP's scope from "improving crystallinity" to "reversibly controlling crystal polymorphism."
Reversible Switching: Uncharted Territory
What's particularly new about this result is the repeated demonstration of reversible switching between the alpha and beta phases on FTO glass. According to the research team, no one had previously shown this kind of reversible phase transition on the conductive glass actually used in real devices. By simply changing the pulse conditions, the same film can be flipped back and forth between gray and yellow, over and over. This opens the door to rewriting a material's optical properties in place, without swapping out the material itself.
Flash lamp annealing (FLA) itself is not new—it's been used in semiconductor manufacturing since the 1970s for dopant activation after ion implantation, and more recently has found applications in perovskite solar cells and flexible electronics. However, the team argues that most existing applications aim at crystallization or sintering, and that this study is the first to demonstrate selective, reversible control over crystal polymorphism on a conductive substrate.
What Comes Next
The team itself identifies extending this approach beyond to other material systems as a key next step. They're currently exploring its applicability to photoelectrode materials for solar water splitting, photocatalysts, and high-entropy oxides. At a 2026 conference presentation, Dr. Gottesman also mentioned applying FPH to high-entropy rare-earth oxides (HEREO), suggesting the research's reach extends well beyond bismuth oxide.
Several questions remain unresolved, however. First is the long-term stability of the beta phase. Being metastable, it thermodynamically "wants" to revert to the alpha phase, and the paper doesn't fully address how well it would hold up under real-device operating conditions—temperature fluctuations, humidity, prolonged light exposure. Second is applicability to plastic or flexible substrates, which have even lower heat tolerance than FTO glass; the team says this is currently under investigation. Third, how much of the reported 10-to-50-fold gain in photocurrent translates into actual efficiency improvements in a working solar cell device remains to be verified at the device level.
It's not the number on a furnace's thermostat, but the speed at which light arrives, that now determines a material's fate. That simple fact has opened up a new degree of freedom in how semiconductor thin films are made.