Selenium powered the first solar cell ever built, back in 1883. Then it vanished from the spotlight for more than a century. Now, as researchers hunt for materials that can push solar efficiency higher and power everything from indoor devices to tandem solar panels, selenium is staging a quiet comeback. But despite rapid progress, selenium solar cells still deliver only about a third of their theoretical maximum efficiency. The main culprit? A voltage that falls far short of what the material's wide band gap should allow.
A new study combining advanced computer modeling with experimental analysis of state of the art selenium films has uncovered something unexpected. The defects that typically sabotage solar cell performance—vacancies, impurities, and displaced atoms—turn out to be remarkably benign in trigonal selenium. The real problem lies elsewhere: in the grain boundaries, interfaces, and disordered regions that riddle the material as it's currently grown.
This discovery shifts the focus. Instead of fighting an unavoidable defect problem, researchers can now target the extended structural flaws that are, in principle, easier to eliminate. The findings open a clearer path toward unlocking selenium's potential as a defect tolerant absorber for next generation photovoltaics.
Why Selenium Matters Again
Selenium's unique properties make it attractive for specific solar applications. Its band gap—the energy threshold electrons must cross to generate current—sits between 1.8 and 2.0 electron volts. That's wide enough to capture high energy photons efficiently, making it ideal for indoor photovoltaics, where artificial light dominates, and for tandem solar cells, where selenium could sit on top of a silicon layer to harvest different parts of the solar spectrum.
The material is also simple to work with. It melts at just 220 degrees Celsius, can be deposited from solution or vapor, and remains stable in air. Unlike many emerging photovoltaic materials, selenium is a single element system, which should simplify both manufacturing and recycling. Recent advances have pushed indoor selenium cell efficiencies to 18 percent, surpassing amorphous silicon and lead free perovskites. Outdoor solar cells have reached 8.1 percent efficiency, up from barely 1 percent for most of the past century.
Still, these numbers fall well short of the theoretical limits. Under ideal conditions, a selenium solar cell could hit 22.6 percent efficiency outdoors and around 50 percent under indoor lighting. The gap between experiment and theory is widest in the open circuit voltage, the maximum voltage a cell can deliver. Record devices reach just under 1 volt, about 0.56 volts below the radiative limit set by the band gap. Understanding where that voltage disappears has been a central puzzle.
Defects That Don't Defect
In most semiconductors, point defects—missing atoms, extra atoms, or foreign atoms sitting in the wrong place—act as traps that pull electrons and holes together before they can contribute to electrical current. This non radiative recombination saps voltage and efficiency. The new study used hybrid density functional theory, a high accuracy quantum mechanical method, to calculate how different defects behave in trigonal selenium's unusual helical chain structure.
The results were striking. Selenium interstitials, extra atoms inserted into the lattice, adopt a split interstitial geometry. An additional selenium atom slots into a helical chain, displacing neighboring atoms slightly but preserving the local bonding environment. The chains simply twist and buckle to accommodate the intruder. Bond lengths compress by only a few hundredths of an angstrom, and bond angles deviate by less than five degrees. This minimal disruption keeps the formation energy low, at 0.82 electron volts, and leaves the defect electrically neutral across most of the band gap. Neutral defects don't trap charge carriers, so they don't drive recombination.
Selenium vacancies behave differently. They can adopt multiple charge states and introduce deep electronic levels in the band gap, hallmarks of a recombination center. The neutral vacancy splits into a bipolaron, with two localized hole states on the dangling chain ends. Charged vacancies form bridging bonds between chains or split a single hole across two terminating atoms. The study found a metastable neutral vacancy structure, where the dangling chain ends reconnect across the void, lying only 27 millielectron volts higher in energy than the ground state, with a barrier of just 65 millielectron volts separating them.
Despite these deep levels, vacancies don't appear to limit device performance. Their formation energy is higher than that of interstitials, so their equilibrium concentration is lower. More importantly, detailed calculations of charge carrier capture rates revealed that while vacancies trap holes quickly, they capture electrons extremely slowly. Efficient recombination requires fast capture of both electrons and holes in sequence. Without that, isolated vacancies can't establish the recombination cycles that drain voltage. Even assuming a high non equilibrium vacancy concentration of 10^18 per cubic centimeter, carrier lifetimes would still exceed 100 microseconds, far longer than needed for efficient charge collection.
Impurities That Compensate Themselves
Real solar cells always contain unintended impurities. The researchers used time of flight secondary ion mass spectrometry on state of the art selenium films to identify what's actually present. Fluorine and chlorine showed up throughout the bulk material, originating from the tellurium substrates used to seed selenium's trigonal phase. Oxygen and tellurium appeared only at surfaces and interfaces. Bromine and iodine were hard to distinguish from selenium and tellurium isotopes but likely also present.
Computational screening covered hydrogen, halogens (fluorine, chlorine, bromine, iodine), pnictogens (nitrogen, phosphorus, arsenic, antimony), and chalcogens (oxygen, sulfur, tellurium). Chalcogen impurities substitute easily into the selenium chains, forming low energy neutral defects that don't contribute to doping or recombination. Tellurium substitution, for instance, costs only 67 millielectron volts. This explains why tellurium accumulates readily at selenium film interfaces but doesn't diffuse into the bulk—the tight in chain bonding keeps substitutional impurities from migrating, while negatively charged halogen interstitials, which behave more like free ions between chains, can move through the film.
Heterovalent impurities—elements with one more or one fewer valence electron than selenium—show amphoteric behavior. They stabilize in both +1 and −1 charge states, with the transition level near mid gap. This means they compensate their own doping. A halogen substitution that might donate holes in the +1 state will also trap holes in the −1 state, canceling any net change in carrier concentration. Fluorine interstitials came closest to contributing p type doping, with a transition level 0.42 electron volts above the valence band, but even this yielded predicted hole concentrations of only 10^12 per cubic centimeter.
Pnictogen and hydrogen defects behaved similarly, forming low energy amphoteric centers that self compensate. Nitrogen was an outlier, remaining high in energy and showing no stable charged states. Across the board, no extrinsic impurity examined could explain the hole densities of 10^15 to 10^16 per cubic centimeter reported in capacitance voltage measurements of polycrystalline selenium films.
The Real Culprits: Grain Boundaries and Disorder
If neither intrinsic defects nor common impurities cause the observed doping or recombination losses, what does? The study points to extended defects, grain boundaries, and remnant amorphous regions. Selenium is a soft, highly deformable solid. Its bulk modulus—a measure of resistance to compression—sits at just 15.0 gigapascals, among the lowest for any semiconductor. Several competing crystal structures and amorphous phases lie within thermal reach of the stable trigonal form. During typical film growth, amorphous selenium is deposited and then annealed just below the melting point to crystallize the trigonal phase. This process easily traps non equilibrium defect populations and incomplete crystalline regions.
Historical reports describe polycrystalline selenium as consisting of well conducting crystals embedded in poorly conducting amorphous layers. More recent work using critical melt annealing has shown that improved crystallinity correlates with better device performance and lower carrier concentrations, consistent with the idea that extended disorder, not point defects, drives both doping and recombination.
All charged intrinsic defects in selenium involve forming bridging bonds between chains. Such bonds are expected to concentrate at grain boundaries and other extended defects, making these regions electrically active. Recent deep level transient spectroscopy measurements found that defect signals scale linearly with pulse width, a signature of extended rather than point defects. Large differences between acceptor densities measured by capacitance voltage and drive level capacitance profiling—which probe bulk and interfacial defects differently—further implicate interfaces as a major source of the apparent doping.
The experimental device studied here showed a moderate Urbach energy of 44 millielectron volts, indicative of band edge disorder. The absorption onset deviates from the simple exponential tail expected for ideal crystals, suggesting structural variations. The combination of low elastic constants, high deformation potentials, and weak interchain bonding means thermal fluctuations can cause significant local variations in the band gap. Strain near interfaces could create spatially varying electronic potentials that trap carriers or accelerate recombination.
What This Means for Better Solar Cells
The findings reframe the challenge. Selenium is not inherently limited by point defect chemistry. Its helical chain structure accommodates interstitials and substitutions with minimal electronic penalty. Even vacancies, despite introducing deep levels, don't recombine charge carriers efficiently enough to explain voltage losses. Instead, the path to higher efficiency runs through better crystal growth and interface engineering.
Several strategies emerge. Minimizing extended defects and amorphous inclusions requires refining the annealing process. The recently reported critical melt annealing approach, which achieved 7.2 percent efficiency, supports this direction. Direct epitaxial growth of crystalline trigonal selenium on suitable substrates, using techniques like molecular beam epitaxy, could bypass the amorphous to crystalline transition and reduce disorder from the start. Long, slow anneals might allow selenium chains to reorganize more completely, though the low melting temperature and high vapor pressure constrain the thermal budget.
Reducing film thickness to 200 to 400 nanometers could limit the impact of extended defects while retaining sufficient absorption, especially if films are oriented along the chain direction to maximize optical absorption and in chain carrier mobility. Controlled texturing or antireflective coatings can compensate for the reduced thickness. The latest 8.1 percent record efficiency used substrate heating to enhance crystallographic orientation along the preferred axis, exactly the kind of directional control that theory predicts should help.
Halogen doping, counterintuitively, may also help. Although halogens don't contribute net doping due to self compensation, early experiments found that adding bromine or iodine to polycrystalline selenium improved conductivity without changing carrier concentration. The improvement came instead from reduced grain boundary resistance, possibly through passivation of dangling bonds or charged defects at these interfaces. Deliberate, controlled halogen incorporation during growth might extend this effect.
The study also highlights the value of strain engineering. Selenium's band gap shifts by 0.25 electron volts over a volume change of just 5 percent. Local strain near interfaces or in textured films could be tuned to optimize band alignment or reduce trapping, though careful control would be essential given the material's sensitivity.
A Path Forward
Selenium's resurgence as a photovoltaic material rests on advantages that go beyond performance alone. Its elemental simplicity, low temperature processing, stability, and recycling potential make it attractive for sustainable energy applications. The high vapor pressure allows closed space evaporation to recover selenium from spent devices with minimal energy input, a demonstrated path for indoor photovoltaics.
The new defect study clarifies what has held selenium back and what can be done about it. Point defects are not the enemy. Extended structural disorder is. Unlike intrinsic point defects, which form unavoidably due to entropy, extended defects depend on growth conditions and can, in principle, be engineered away. This shifts the development focus from fundamental materials limits to process optimization, a more tractable problem.
Achieving open circuit voltages closer to the radiative limit will require understanding and controlling the specific types of extended defects present in real films. High resolution transmission electron microscopy combined with nanoscale electronic probes could map where disorder concentrates and how it correlates with recombination. Deep level transient spectroscopy can fingerprint defect states and track how they respond to different annealing protocols or dopants. Multi modal microscopy techniques that correlate local structure with local electronic properties will be especially valuable.
Selenium won't replace silicon in rooftop solar panels. But for tandem cells, where a wide gap top cell pairs with a narrow gap bottom cell to capture more of the solar spectrum, and for indoor photovoltaics powering sensors and electronics under artificial light, selenium offers a compelling combination of performance, simplicity, and sustainability. With efficiencies already at 18 percent indoors and rising outdoors, and with point defect tolerance now confirmed, the main obstacles are tractable. The first solar cell was made from selenium. The next generation of high efficiency, low cost photovoltaics might bring it back.
Credit & Disclaimer: This article is a popular science summary written to make peer-reviewed research accessible to a broad audience. All scientific facts, findings, and conclusions presented here are drawn directly and accurately from the original research paper. Readers are strongly encouraged to consult the full research article for complete data, methodologies, and scientific detail. The article can be accessed through https://doi.org/10.1039/d4ee04647a






