Hydrogen is an abundant and clean energy source that can help decarbonise difficult-to-electrify sectors of our economies. However, its safe deployment relies on the availability of reliable and cost-effective hydrogen detection technologies. Existing hydrogen detectors are expensive, bulky, and power-hungry, limiting their broader use in emerging applications. In this talk, I will discuss the development of a hydrogen sensor based on printable organic semiconductors that can operate across wide temperature and humidity ranges and offer numerous advantages over commercial technologies. I will describe the fundamental sensing mechanism exploited and the prospects for further developments through new materials design and device engineering. Our organic sensors exhibit high responsivity, ultra-low power consumption, and record-short response and recovery times. Compared with commercial technologies, our printed hydrogen sensors demonstrate superior performance in diverse real-world sensing scenarios, making them strong candidates for health and safety applications in the emerging hydrogen industry.

Hole transport layers (HTLs) and electron transport layers (ETLs) are pivotal for emerging electronic and photonic devices. Wide bandgap, high charge mobility, and solution processability are essential properties to develop low-cost and commercialisable technologies. Metal-organic hybrids, such as organometallic compounds or coordination polymers (CPs), are promising due to the versatility of metal and organic ligand units, but they still largely lag behind in terms of development for device applications. Herein, we enhance charge-transport properties of hybrid semiconductors and consequently the performance of organic photovoltaics and photodetectors (OPVs and OPDs, respectively). For HTLs, we dope a p-type CP semiconductor, copper(I) thiocyanate (CuSCN), with copper(II) bromide (CuBr₂) to obtain synergistic effects of p-doping and defect healing, respectively contributed by Cu⁺ and Br^-. The improved hole-transport properties are demonstrated by the increased power conversion efficiencies (PCEs) of OPVs based on various active layers, achieving the highest PCE of 18.3% for CuSCN-based OPV devices (PM6:D18:L8-BO). Furthermore, our approach records a specific detectivity of 1.1 × 10¹² Jones (850 nm, -2 V) for single-component OPDs, i.e., using solely Y6 (non-fullerene acceptor) to convert photons to current.[1] For electron transport, ferrocenyl-bis-furyl-2-ketone (FcFk₂), an organometallic compound, is employed to form complexes with several ETLs, e.g., PFN-Br, PNDIT-F3N, and PDINN, for a variety of photoactive blends. The hybrid complexes exhibit faster electron extraction and reduce trap-assisted recombination of OPVs, yielding the champion PCE of 19.7% and 20.1% with anti-reflective coating. Upon 1-sun light illumination, FcFk₂-based OPV cells also retain 80% of the initial PCE up to 700 h under maximum power point tracking.[2] The universal device improvements effectively demonstrate hybrid HTLs and ETLs as a promising class of materials to enhance the performance and stability of a wide range of electronic and photonic devices.

Metal halide perovskites have emerged as a leading class of light-emitting semiconductors owing to their outstanding optoelectronic properties, high colour purity, and widely tunable bandgaps. In particular, near-infrared (NIR) perovskite light-emitting diodes (PeLEDs) are attracting increasing interest for applications including optical communication, machine vision, biomedical imaging, and night-vision technologies. While solution-processed NIR PeLEDs have now achieved external quantum efficiencies (EQEs) exceeding 30%, the development of thermally evaporated NIR perovskites has lagged significantly behind despite their inherent compatibility with mature OLED manufacturing infrastructure and large-area device fabrication. Progress in this area is further hindered by an incomplete understanding of phase stability, defect formation, and degradation processes in evaporated formamidinium (FA)-based perovskites, limiting both device efficiency and operational stability.

Here, we report a mixed-halide engineering strategy to control the structural and optoelectronic properties of thermally evaporated NIR perovskites. Such halide alloying stabilizes the photoactive α-phase, suppresses defect formation and improves charge-carrier transport compared to FAPbI₃ thin films. As a result, we demonstrate NIR perovskite LEDs (peak emission wavelength of 778 nm) with EQEs approaching 10%, and a low turn-on voltage of 2.4 V. Notably, the devices achieve a record peak radiance of 54 W sr^-1 m^-2, representing more than 18x improvement over the reported literature. To obtain deeper microscopic understanding of our films, we employ photoluminescence (PL) imaging and correlation cluster imaging (CLIM) to directly visualize static and dynamic defect populations, revealing spatial heterogeneity and carrier recombination pathways that have remained largely unexplored in vacuum-deposited perovskites. The optimized devices exhibit negligible electroluminescence peak shifts with increasing drive voltages and operational lifetimes (T₅₀) of several minutes were obtained. Furthermore, multimodal characterization of fresh and degraded device stacks using STEM-EDX, ToF-SIMS, and PL microscopy reveals the evolution of structure, morphology, composition, and defect landscape across the multilayer architecture, identifying interesting degradation pathways under continuous current-bias operation. Together, these findings establish evaporated mixed-halide perovskites as a promising platform for scalable and high-performance NIR light sources. 

In recent years, non-fullerene acceptor (NFA) materials have driven organic photovoltaic (OPV) efficiency from roughly 12% to certified single-junction values now exceeding 20%, with organic tandems surpassing 21%. However, these advances are most often achieved on small-area, spin-coated devices whose formulations are not readily transferable to scalable manufacturing, and which still face severe limitations in lifetime and operational stability under real-world conditions.

Closing this gap requires working on two fronts at once. Intrinsically, the organic semiconductors, formulations, and interlayers must be engineered for full compatibility with slot-die coating in roll-to-roll (R2R) production, without sacrificing the morphology behind high-efficiency blends. Extrinsically, module behaviour is equally critical: the physical and chemical compatibility of the active stack with adhesives, encapsulants, and barrier films, balanced against barrier performance, cost, and operating environment.

In this talk we present NFA-based modules and full-size panels manufactured by slot-die R2R coating, several of which are already installed and operating in outdoor environments. Drawing on this field experience, we discuss the practical requirements of moving from laboratory devices to manufacturable products, the recurring degradation modes observed under real operating conditions, and the factors that continue to slow the commercial uptake of these materials. We also consider application contexts where deployed OPV is beginning to show value, including surface temperature reduction.

Organic Photovoltaics (OPV) has recently reached Power Conversion Efficiency (PCE) above the 20% milestone for outdoor applications (1 Sun, AM1.5G), and at the same time, huge efforts have been made on pushing the performance for low light indoor applications, currently reaching above 30% PCE (for small cells at research scale) under such operating conditions. These devices have huge potential for powering up Internet-of-Things (IoT) devices down to 50 lux lighting conditions, and therefore, development of high-performance indoor OPV modules that can be manufactured and scaled at industrial compatible conditions is currently in focus. In this presentation, recent work on scalable OPV cells and modules will be presented, having a focus on Roll-to-Roll (R2R) techniques for development of scalable modules at ambient air conditions for outdoor [1], and in particular new indoor applications. Firstly, a summary of the processing techniques utilized in our work for scalable OPV device development is outlined and demonstrated, spanning from R2R-based vacuum sputtering to scalable slot-die coating, to cover the full range of functional layers embedded in these OPV device stacks.

Secondly, a particular focus will be paid to development of scalable OPV modules for indoor applications at low light conditions. As indoor light sources exhibit narrower emission spectra and lower light intensities, OPV donor and acceptor materials with wider HOMO-LUMO gap are utilized, and optimized towards improved visible light absorption, high shunt resistance and stable operation under indoor light conditions. In this work, we designed and developed OPV cells and modules using PTQ10 as the polymer donor and FCC-Cl as wide-bandgap non-fullerene acceptor, employed in various different device stacks, also using new PDIN-based molecules for the electron transport layer to reach high performance [2]. Optimized small-area cells achieved a power conversion efficiency (PCE) of above 26% under 1000 lux illumination, and high stability under continuous indoor illumination. Furthermore, the developed device architecture was successfully scaled up using sheet-to-sheet (S2S) slot-die coating at ambient air conditions, utilizing green solvents for the active layer processing. PTQ10:FCC-Cl OPV modules with an active area of 13.8 cm² fabricated under such conditions reached a high PCE of 20%, under the same indoor low light conditions. The results demonstrate the strong potential of this material system and processing approach for scalable, high-efficiency indoor OPV applications.

Optimizing organic photovoltaic (OPV) devices requires an accurate, multiscale understanding of exciton transport. However, modeling these processes from first principles introduces severe computational difficulties. In this presentation, I will outline our on-going efforts to overcome these difficulties by combining density functional theory, ab initio structure prediction, and machine learning models. 

We first discuss our efforts to model the surface of the coordination polymer copper thiocyanate (CuSCN), a wide-bandgap semiconductor often used as a hole-transporting layer in OPV devices. Finite slab models of CuSCN surfaces often suffer from artificial dipole moments, causing unphysical band energy drifts and bandgap closure (Figure). We introduce a novel passivation scheme utilizing "pseudohydrogen" atoms via the virtual crystal approximation (VCA) to counteract internal electric fields. While this VCA approach restores the bandgap, it induces an incorrect n-type behavior. By performing a comprehensive ab initio structure search, we demonstrate that surface reconstructions of CuSCN(001) offer a self-passivation mechanism that minimizes surface energy, reduces internal fields, and correctly reproduces experimental p-type semiconducting behavior. These results strongly support the use of reconstructed CuSCN surface slabs when modeling CuSCN-based OPV devices. 

We then discuss efforts to simulate exciton transport organic semiconductors, where calculating intermolecular couplings and other exciton transport parameters from first principles is excessively cost-prohibitive. We present a graph neural network (GNN) architecture that dramatically accelerates prediction times without sacrificing the accuracy needed for exciton transport simulations. Importantly, our GNN generalized across diverse molecular systems by utilizing atomic transition charges as an intermediate representation. This GNN successfully predicts exciton couplings for a broad class of fused-ring electron acceptors, enabling robust exciton diffusion simulations without repeated quantum chemical calculations [1], [2], [3]. 

Organic solar cells have recently broke the power conversion efficiencies limit of 20 %. This puts extra attention to increasing their stability as their last remaining weak point. Since organic solar cells consist of organic molecules, they are inherently susceptible to degradation via oxygen, light, heat, and humidity, which are all commonly found stresses in their working environment.

Different stabilizing additives, such as antioxidants, radical scavengers, hydroperoxide decomposers, UV absorbers and more, can be incorporated in active layers of organic solar cells to prevent or slow down their degradation. This is both inexpensive and easily upscalable intervention, and it also does not introduce further complexity into the solar cell device architecture.
Here we are presenting our recent results on non-fullerene based solar cells stability improvement using naturally occurring antioxidants that act as singlet oxygen quenchers and radical scavenging compounds in highly efficient organic solar cells. Using fluorescent probes, we also investigate and discuss the dominating degradation mechanisms in these systems.

Vertically stacked perovskite photodetectors enable accurate, filter-free full-color detection within a single pixel footprint, offering a promising route toward compact imaging systems.[1] However, the absence of scalable fabrication strategies for reliable vertical interconnect access (VIA) remains a critical bottleneck in the transition from proof-of-concept devices to integrated systems.

Here, we report the first back-end-of-line-compatible fabrication framework for the monolithic integration of vertically stacked perovskite photodetectors with an active readout backplane. By establishing perovskite-specific multilayer patterning and sidewall passivation protocols, a robust vertical interconnect strategy is realized, providing low-leakage electrical access from stacked perovskite devices to the underlying readout electrode.

Using this approach, we demonstrate the first 32 × 32 × 3 vertically stacked perovskite color image sensor integrated with a thin-film-transistor active-matrix backplane, showing improved light utilization with geometrical fill factors exceeding 70%. The sensor also exhibits high spectral selectivity and uniformity.

Beyond imaging, this work establishes a scalable integration strategy for perovskite optoelectronics, enabling co-localization of spectrally selective absorption, emission, and electronic functionality within a single pixel. This capability opens new opportunities for multispectral sensing, light-emitting devices, and in-pixel or neuromorphic computation, paving the way toward compact, high-efficiency, and multifunctional optoelectronic systems.

Metal halide perovskites have emerged as remarkable semiconductors for photodetectors, offering strong optical absorption, tunable bandgaps, and low-noise operation comparable to established silicon technologies.[1] Nevertheless, most reported thin-film devices remain at the level of stand-alone photodetectors, limiting their practical implementation in imaging systems. Integration onto CMOS readout platform is non-trivial, as the perovskite deposition process must be compatible with pre-fabricated CMOS surfaces[2], and photolithographic patterning of perovskites is still challenging due to their instability under conventional semiconductor fabrication conditions.[3]

Here, we present a scalable and reliable vacuum-based deposition protocol for the high performance perovskite photodiodes (PePDs), achieving a peak detectivity of 9.2 × 10¹² Jones at −0.5 V, among the top-performing devices reported to date. Using this strategy, we demonstrate solvent-free monolithic integration of PePDs onto a 180-nm CMOS readout integrated circuit. The resulting image sensors exhibit stable operation with low noise and uniform response across the pixel array, confirming the viability of the process for system-level implementation. In addition, we introduce a modified-pixel architecture that mitigates electrical crosstalk between neighboring pixels, thereby improving image sharpness. Our work highlights the importance of processing-controlled integration in translating perovskite semiconductors from discrete photodetectors to practical imaging systems.

The current success of organic semiconductor technology is mainly driven by the development of organic light-emitting diodes (OLED), which are now routinely employed in display technologies. In the last decade, however, organic photovoltaics (OPV), leveraging the impressive improvement in device efficiency and stability, have gradually moved from a lab curiosity to a niche market. Organic photodetector (OPD), a technology based on organic photodiodes and thus closely related to OPV, is a new exciting avenue.

OPDs capable of detecting near-infrared (NIR) and short-wave infrared (SWIR) light are emerging as promising candidates for low-cost, large-area optoelectronic applications. Achieving efficient NIR detection with solution-processed organic semiconductors remains challenging due to their inherently limited absorption bandwidths. However, the advent of nonfullerene acceptors (NFAs) and ultra-low bandgap polymers has enabled precise control over molecular energy levels and optical band gaps, paving the way for extended infrared responsivity > 1000 nm. Building on advances in organic photovoltaics, the design of optimised donor:acceptor blends and the mitigation of charge carrier recombination are now central strategies for achieving high-efficiency and durable NIR–SWIR OPDs. This talk outlines material and device design principles that drive effective light-to-current conversion in this spectral range, highlighting pathways for the next generation of infrared organic photodetectors.

Film inhomogeneities, such as thickness variations, are usually consider a caveat for reproducibility and upscaling of thin film-based technologies.  Controlled wedges, however, open up the possibility to evaluate how thickness affects a particular device in a continuous fashion, facilitating device optimization and accelerating material screening. Moreover, novel device concepts arise from the use of active layers exhibiting thickness gradients.

In this talk, I would first describe the fabrication of organic semiconductor films exhibiting controlled thickness gradients by evaporation [1] and by solution processing [2]. Then, I would show how this type of sample can be used for the high throughput optimization and materials screening for organic solar cells. The large body of data produced in this fashion can then serve as input to evaluate different theoretical frameworks for solar cells, such as drift-diffusion, kinetic Montecarlo, and machine learning models [2].

Besides device optimization, wedges open the opportunity to fabricate novel devices, such as position sensitive photodetectors [1] and miniature spectrometers based on microcavities [3]. As a final example, we will use optical microcavities with a thickness gradient in the core to tune the resonance and easily find the conditions to enhance the Raman signal of diluted molecules, thus opening the possibility to study solid state vibrations in isolated molecules. 

Nanoparticle exsolution - the growth of metallic nanoparticles directly from an oxide support in which host cations have been substituted to some degree by cations of the active component(s), has demonstrated great promise in the preparation of heterogeneous catalyst materials, producing fine, well-distributed nanoparticle catalysts that are anchored in their oxide host, endowing them with excellent stability against deactivation. In this work, we take a systematic approach to studying the exsolution behaviour of ruthenium, iron and their bimetallic alloy from defect fluorite-type yttrium zirconate, a host structure type that has not been extensively employed in exsolution studies, but which presents a particularly interesting alternative host structure for exsolution, as defect fluorites exhibit a high intrinsic concentration of oxygen vacancies, which are well established to play an important role in exsolution. We combine both ambient pressure and vacuum X-ray photoelectron spectroscopy, alongside electron microscopy, to probe how the electronic structure evolves as the reducing conditions are varied during exsolution, gaining valuable insight into the sequence of chemical state changes that take place in the initial stages of exsolution, and how the distribution of these states, and consequently, the extent of exsolution, depends on the conditions imposed during reduction.

Solution-processed molecular doping is a promising strategy to improve charge transport in organic semiconductors, but the relationship between dopant loading, thermal stability, and depth-dependent film morphology remains insufficiently understood. In this work, PM6 polymer thin films were doped with F4TCNQ and F6TCNNQ, and their electrical conductivity and vertical morphology were investigated as a function of dopant loading and thermal annealing.

Neutron reflectometry (NR) reveals that increasing dopant concentration leads to modest film densification and surface smoothing, while the fitted scattering length density remains close to that of pristine PM6 at dopant loadings of 1, 3, and 5%. These results indicate no clear evidence of macroscopic vertical stratification within the sensitivity of the measurements, suggesting that the doped films retain a largely stable depth profile after solution processing. Conductivity measurements after stepwise thermal annealing up to 150 °C were used to evaluate changes in doping efficiency and possible thermal dedoping. Complementary NR measurements on pristine PM6 show only minor changes in thickness and roughness after annealing to 150 °C, indicating that the PM6 matrix itself is vertically stable under these conditions. Therefore, annealing-induced changes in conductivity are more likely associated with local molecular rearrangement, dopant-polymer interactions, or changes in crystalline packing rather than large-scale vertical phase segregation.

Ongoing X-ray scattering and complementary thin-film characterization will further correlate conductivity evolution with molecular packing and nanoscale texture. Overall, this work provides a depth-resolved framework for evaluating dopant retention, morphology stability, and thermal robustness in solution-processed donor polymers for organic electronic applications.

Metal halide perovskite photovoltaics have reached remarkable power conversion efficiencies; however, their commercial deployment critically depends on replacing laboratory-scale fabrication with robust, scalable manufacturing processes. Among solution-processing techniques, slot-die coating and inkjet printing offer highly attractive pathways towards industrial production owing to its compatibility with continuous processing, low material waste, and large-area manufacturing. Nevertheless, successful implementation requires a detailed understanding of the interplay between ink formulation, coating dynamics, substrate engineering, drying kinetics, and interface formation.

This presentation will provide an overview of the printing and coating activities at the Helmholtz-Zentrum Berlin (HZB) carried out within the Helmholtz Technology Acceleration Platform Solar TAP, whose mission is to accelerate the transfer of emerging photovoltaic technologies from laboratory research to industrial implementation. The talk will illustrate how scalable process development is combined with close collaboration between academia, equipment manufacturers, material suppliers, and industrial end users to establish manufacturing-ready processes for printed photovoltaics. Recent achievements include scalable deposition of functional transport layers, optimization of perovskite absorber coating, process monitoring, and the translation of laboratory concepts to large-area devices and mini-modules [1,2,3]. Furthermore, examples of industrial training, technology transfer, and collaborative process development will demonstrate how shared infrastructure and application-driven research can significantly shorten the lab-to-fab cycle.

By highlighting both the scientific challenges of perovskite processing and the innovation ecosystem established through Solar TAP, this presentation will discuss how scalable coating technologies, standardized process development, and interdisciplinary collaborations are enabling the next generation of high-performance, manufacturable perovskite solar cells and modules.

Expeditious record-breaking power conversion efficiency of perovskite solar cells (PSCs) reaching 28% is owed to the scientific mainstream devoted to understanding its intricate scaffold of designing the layers architecture.¹The ease of processing and modifying the staked layers paves a pathway in realizing the targets of low-cost production, high efficiency, and better stability.  Most fabrication processes require inert environments, complex deposition methods of chemical layers, and high temperature.^2,3,4 Such conditions are majorly suitable for rigid substrates curbing the possibility to transfer the technology on a flexible base like polymeric substrates or thin metal foils. Flexible PSCs have its potential in the integration on curved surfaces and light weight enabling applications in wearable electronics, portable power sources, and building-integrated photovoltaics.⁵ In addition, replacement of hazardous and toxic solvents used in the fabrication processes is the need of the hour and address alternative solvents while reporting the device performance. This will positively impact in industrial production and commercialization advancements keeping in check the governmental regulations and legality of a country.^6,7 Augmenting the panoramic research of PSCs, this work emphasizes major 4 key areas in fabricating a flexible PSC. Starting from fabricating outside inert environment in a range of conditions, reducing high temperature processing requirement, scaling up from small area to larger modules, and using more sustainable and eco-friendlier benign solvents in the entirety of the fabrication, this work realized power conversion efficiencies of more than 21.36% on rigid substrates and 19.11 % on flexible PET substrates.  Moreover, scaling up with a blade coating deposition a mini module with PCE of 17.6% is achieved for the first time in a fully green fabrication process. A holistic pathway and minor changes essential in developing a fully green processed flexible perovskite solar cell are demonstrated in this work clearly marking the state-of-the-art in Industrial oriented PSCs research.

Metal-free halide perovskites are an inherently earth-abundant and non-toxic emerging class of materials, with broad application potential, from piezoelectric wearable devices to photodetectors. Much like their metal-halide perovskite counterparts, different properties can be accessed and tuned by introducing and mixing different ions in the ABX3 lattice. Yet, their potential is largely underexplored.

Mechanochemistry offers a powerful route into this parameter space: it is a processing strategy that is simple, fast, green, and easily upscalable. Surprisingly, beyond these practical advantages, it provides an additional dimension of control on the properties of metal-free perovskites.  Tuning the impact energy during processing directly governs visible light absorption in these traditionally wide bandgap materials, while the fundamental properties, such as the crystal structure, remain the same.

This establishes mechanochemical processing conditions as a new and powerful tuning strategy, opening new doors into a material class we are only just beginning to scratch the surface of.

Flexible perovskite solar cells (f-PSCs) have recently reached power conversion efficiency (PCE) above 25%. Although still lagging behind their rigid counterparts on glass, which in very short time have achieved 28% of certified efficiency,  the use of flexible substrates opens up to a wide range of applications, from sensors for the Internet of Things, to the retrofitting of existing buildings to improve their energy efficiency (building-applied PV), to space, thanks above all to the high power/to weight ratio generated which is the range of 29.4 W/g compared to 8.31 W/g for amorphous silicon and 0.254 W/g for ultra-thin CdS / CdTe.

In this presentation an overview of the fabrication of flexible perovskite solar cells and modules and of their use in two unconventional environments (indoor wearable and space) will be reported. In particular, for indoor applications I will report on our studies of combining perovskite solar cells and modules with energy storage systems (supercapacitors) introducing the concept of photocapacitor, a device where the energy photovoltaic (PV) generation and storage systems are combined in a single unit, offering an innovative approach to manage energy supply. For space applications, I will show the role of PSC in this emerging field of research, reporting our studies related to the resilience of flexible perovskite solar cells to neutrons and to other stressors such as AM0 illumination and vacuum.

Earth-abundant semiconductors are attracting increasing attention as candidates for sustainable and scalable photovoltaic technologies. Among them, zinc phosphide (Zn₃P₂) remains relatively underexplored despite its highly favorable optoelectronic properties, including a direct band gap of ~1.5 eV, a high absorption coefficient in the visible (>10⁴ cm⁻¹), carrier diffusion length in the range of 5−10 μm and intrinsic p-type.[1] In addition, its tolerance to stoichiometric variations enables a wide tunability of carrier concentration, making it a versatile platform for device design. However, the development of efficient Zn₃P₂-based solar cells has historically been hindered by limitations in both material quality and device architecture. For decades, device performance remained largely stagnant, with Schottky and heterojunction solar cells constrained by resistive losses, interface recombination, and incomplete carrier collection, even when improvements in crystallinity were achieved. [2]

Here, we present a device-focused approach to Zn₃P₂ photovoltaics, enabled by recent advances in thin-film growth at mild temperatures. Using molecular beam epitaxy (MBE) and selective area epitaxy (SAE), we achieve improved control over morphology, defect formation, and interfaces.[1,3] In particular, SAE enables nanoscale strain relaxation, leading to high-quality coalesced thin films with enhanced transport properties, including hole mobilities exceeding 500 cm²/V·s.[3,4] Building on this materials platform, we demonstrate the progression of Zn₃P₂-based photovoltaic devices from early heterojunction designs toward more optimized architectures.[4] Through combined electrical, optical, and nanoscale characterization, we identify the dominant loss mechanisms, including recombination at the front interface and limited carrier extraction. Leveraging these insights, we have recently achieved record device performance for Zn₃P₂-based solar cells, highlighting the impact of improved material quality coupled with refined device design.

Finally, we discuss ongoing strategies toward next-generation architectures based on selective contacts, aimed at enhancing carrier selectivity and suppressing interfacial recombination.[5] These approaches provide a clear pathway to further improve efficiency and establish Zn₃P₂ as a competitive, earth-abundant absorber for future photovoltaic technologies.