Researchers at the Hebrew University of Jerusalem have used intense millisecond flashes of light to control the crystal structure of bismuth oxide films while keeping the conducting glass beneath them below damaging temperatures. The resulting metastable semiconductor generated between ten and 50 times more photocurrent than its conventional stable phase, depending on how the film was prepared.
The process uses flash photonic heating, in which a thin coating absorbs powerful pulses of white light lasting from around a tenth of a millisecond to several milliseconds. Because the energy is delivered directly into the film over such a short period, the semiconductor can heat and cool much faster than thermal energy spreads into the underlying substrate.
Reported heating rates reached approximately one million to ten million degrees Celsius per second. In experiments using bismuth oxide deposited on fluorine-doped tin oxide conducting glass, the film could approach temperatures of around 2,000°C while the supporting substrate remained below approximately 100°C.
That thermal separation is the central result. Conventional furnace processing heats the coating and its substrate together, so transparent conducting glass reaches its practical temperature limit long before a film on top can experience the extreme thermal conditions required to form some crystal structures. Slow furnace cooling creates a second problem by giving atoms time to return towards their most stable arrangement.
The Hebrew University team used the short heating and cooling cycle to address both limitations. By delivering sufficient energy quickly and removing it before the substrate could follow, the researchers created conditions that temporarily favoured a different arrangement of the same material and then effectively froze that structure in place.
Bismuth oxide can exist in multiple polymorphs — materials with the same chemical composition but different atomic arrangements. Those structures can have markedly different electrical and optical properties even though no additional element has been added. Diamond and graphite provide the familiar bulk-material example; semiconductor polymorphs exploit the same general principle at much smaller scale.
The study focused on the stable alpha phase and the metastable beta phase of bismuth oxide. The researchers measured a bandgap of approximately 3.15eV for the alpha form and around 2.35eV for the beta phase. The narrower bandgap allows the beta material to absorb a broader part of the visible spectrum, contributing to its stronger photoresponse.
Dr Ronen Gottesman said: “The idea is to heat and cool the material so quickly that we can trap it in a crystal structure that would normally disappear. This allows us to access useful properties that conventional heating methods cannot easily preserve.”
The process resembles quenching in metallurgy, where rapid cooling prevents a high-temperature structure from fully relaxing into its equilibrium state. The difference is scale and selectivity: rather than heating a bulk component and then quenching it, the flash process concentrates the thermal event in a thin film while leaving the supporting material comparatively cool.
The researchers also found that total energy alone did not determine which phase formed. Pulses carrying the same overall energy could produce different structures depending on how rapidly that energy was delivered. Shorter, more intense pulses favoured the metastable beta phase, while longer pulses produced the stable alpha phase.
That makes processing time a critical control parameter alongside temperature and energy. Conventional thermal recipes are frequently described by peak temperature, dwell time, and cooling rate; flash processing compresses those variables into a regime where fractions of a millisecond materially change the structure that survives.
The beta phase produced between ten and 50 times more photocurrent than the alpha material under the reported conditions. The size of the improvement depended on film preparation and microstructure, but the researchers linked the stronger response to broader light absorption and more efficient movement of electrical charge through the material.
No change in chemical composition was required. The performance gain resulted from rearranging the same constituent atoms into a different crystal form, creating an attractive manufacturing proposition if the phase can be reproduced consistently over useful areas.
Transparent conducting glass is particularly relevant because it already forms part of photovoltaic, display, sensing, and photoelectrochemical systems. Its combination of optical transparency and electrical conductivity is useful precisely where light must pass through an electrode, but its thermal tolerance places limits on which materials can be deposited or transformed after the conducting layer is present.
Flash photonic heating offers a potential route around that restriction. A high-temperature film phase could, in principle, be formed late in the manufacturing sequence without subjecting every previously deposited layer to the same thermal cycle. That could expand the range of materials compatible with temperature-sensitive substrates and reduce the need to redesign the complete device stack around one furnace process.
The researchers also demonstrated reversible switching between the two bismuth oxide phases on the conducting glass. That is scientifically useful, but it should not yet be mistaken for a qualified industrial switching device. Repeated production would have to control film thickness, substrate condition, optical absorption, pulse uniformity, thermal history, phase stability, and device performance across substantially larger areas.
Scaling is likely to be the more difficult engineering problem. A laboratory film can be illuminated with a carefully controlled pulse over a small region. Industrial equipment would need to deliver sufficiently uniform optical energy across larger substrates or through a continuous process while monitoring enough variables to maintain the intended crystal phase from one product to the next.
Throughput will matter as much as peak performance. Millisecond processing is inherently fast, but production equipment still has to handle substrates, control pulse delivery, verify quality, and integrate with upstream and downstream steps without damaging adjacent materials.
The team is now examining whether the method can be extended to other materials and to plastic or flexible substrates with even lower temperature limits. Whether those systems respond in the same way will depend on their optical absorption, heat flow, transformation kinetics, and interaction with the surface beneath them.
The current work therefore establishes a processing principle rather than a universal manufacturing recipe. Its importance lies in showing that a semiconductor film can briefly reach temperatures that would normally destroy its substrate, transform into a more useful crystal phase, and cool again before the underlying glass has time to become hot enough to fail.
The reported 50-fold photocurrent improvement provides the eye-catching performance figure. The more consequential engineering result is that the improvement was achieved on transparent conducting glass without subjecting the whole device structure to a 2,000°C furnace cycle.
