The Manufacturing Puzzle

Organic solar cells promise a revolution in renewable energy. Lightweight, flexible, and tunable in color, they could transform windows into power generators or integrate into clothing. Over the past few years, their efficiency has soared past 20%, rivaling established technologies.

But there's a problem that has plagued the field since its inception. Change the solvent used to process the active layer—the heart of the solar cell—and efficiency plummets. A cell that achieves 17% efficiency when made with chloroform might drop to 12% when made with chlorobenzene. Switch to toluene or xylene, and performance tanks further.

This isn't just inconvenient. It's potentially catastrophic for commercialization. Industrial production demands environmentally friendly solvents. Chloroform is toxic. Chlorobenzene is hazardous. Regulators increasingly restrict halogenated solvents for large-scale manufacturing. Yet these are precisely the solvents that produce the best solar cells.

The dilemma runs deeper than environmental compliance. Even among acceptable solvents, performance varies wildly. Manufacturers cannot simply switch between processing liquids based on cost or availability without risking catastrophic efficiency loss. This unpredictability strangles scalability.

Researchers have now solved this problem. By redesigning the molecular architecture of the light-absorbing materials, they created organic solar cells that maintain 19% efficiency regardless of which solvent processes them. The breakthrough reveals fundamental principles governing morphology control—how molecules arrange themselves during film formation—and provides a blueprint for materials design toward practical applications.

The Morphology Mystery

Organic solar cells consist of two types of semiconductor molecules mixed together in a bulk heterojunction. Electron donors absorb light and generate excitons—bound electron-hole pairs. Electron acceptors receive electrons from those excitons, separating the charges. Both components must arrange themselves into favorable nano-scale structures for the cell to work efficiently.

This morphology—the three-dimensional organization of donor and acceptor molecules in the solid film—determines whether excitons reach interfaces before recombining, whether separated charges escape their mutual attraction, and whether carriers reach electrodes before getting trapped. Get the morphology wrong and efficiency collapses.

The processing solvent profoundly affects morphology. As the solvent evaporates during film formation, molecules aggregate, crystallize, and phase-separate. Different solvents have different boiling points, different interactions with the semiconductors, and different evaporation kinetics. These variations alter which molecules precipitate first, how quickly domains form, and what final structures emerge.

For years, researchers optimized morphology empirically. Try different solvents. Add processing additives. Adjust spin-coating speed. Perform thermal annealing. Each material system required its own recipe. Even minor modifications to molecular structure demanded starting optimization from scratch.

This empirical approach provided no general principles. Why does PM6:Y6 (a popular donor-acceptor combination) achieve 17% efficiency in chloroform but only 12% in chlorobenzene? The answer involved aggregation in solution, precipitation kinetics, crystallization pathways—but no unified framework connected these phenomena.

Understanding remained microscopic. Researchers could see large aggregates in solution or measure domain sizes in films. But the relevant processes occur at molecular and nanometer scales where excitons diffuse, charges separate, and transport happens. Bridging this gap required new approaches.

The Side Chain Strategy

The breakthrough came from rethinking what controls molecular behavior in solution. Previous work focused on the semiconducting backbone—the conjugated core that absorbs light and transports charge. Researchers tuned energy levels, broadened absorption spectra, and improved transport properties.

But what about the side chains? These aliphatic appendages hang off the conjugated backbone, serving primarily to improve solubility. Standard acceptor molecules carry branched alkyl chains—flexible hydrocarbon tentacles that interact weakly with solvents through dispersion forces.

The research team asked: what if we replaced these passive side chains with active ones that interact strongly with solvents?

They designed BTP-TO2, a non-fullerene acceptor molecule based on the high-performing BTP core architecture. The key modification: attaching an oligo(ethylene glycol) (OEG) side chain to the central nitrogen atom of the benzotriazole unit. This OEG chain contains oxygen atoms that make it hydrophilic, polar, and flexible.

These properties proved transformative. OEG chains interact strongly with organic solvents through hydrogen bonding and dipole-dipole interactions. Nuclear magnetic resonance spectroscopy revealed that BTP-TO2 maintains essentially identical molecular conformations across different solvents—chloroform, chlorobenzene, toluene, and p-xylene. The strong solvent-OEG interactions stabilize the molecular geometry.

In contrast, the reference molecule BTP-TC8—identical except for standard branched alkyl side chains—showed significant conformational changes between different solvents. Weak alkyl-solvent interactions allowed the molecule to adopt different shapes depending on its environment.

This conformational stability matters because it propagates through the entire processing sequence. A molecule that maintains the same shape in different solutions will behave more predictably during film formation.

The Donor-Acceptor Dance

But stable acceptor conformation alone doesn't guarantee solvent-insensitive morphology. The donor polymer PM6 must also behave consistently. Here, a second design principle emerges: weakening donor-acceptor interactions in solution.

Two-dimensional nuclear magnetic resonance spectroscopy revealed a striking difference. In PM6:BTP-TO2 solutions, correlation signals arose primarily from intra-molecular interactions within BTP-TO2 molecules. Minimal signals indicated PM6-BTP-TO2 intermolecular interactions. The bulky OEG side chains acted as steric barriers, preventing donor and acceptor molecules from associating strongly in solution.

Computational simulations confirmed this picture. Cohesive energy density calculations showed that OEG-solvent interactions consistently exceeded solvent-solvent interactions across all solvents tested. This thermodynamic favorability drove strong solvation of the OEG chains.

In PM6:BTP-TC8 solutions, the situation differed dramatically. Correlation signals showed substantial PM6-BTP-TC8 intermolecular interactions. The extent varied between solvents. In some solutions, donor and acceptor associated extensively. In others, they remained more independent.

These solution-state differences propagate into film formation. Small-angle neutron scattering measurements quantified polymer conformation by fitting flexible cylinder models. PM6:BTP-TO2 solutions yielded large Kuhn lengths and large radii of gyration across all solvents—signatures of rigid, extended polymer conformations. PM6:BTP-TC8 solutions showed more variability.

The mechanism becomes clear. Strong OEG-solvent interactions and weak PM6-BTP-TO2 interactions maintain PM6 in a rod-like conformation regardless of solvent. The polymer doesn't twist or entangle differently in different solvents. It maintains a consistent architecture.

Preferential Precipitation

Solution behavior sets initial conditions. But film formation dynamics determine final morphology. As solvent evaporates, molecules must transition from solution to solid state. Which component precipitates first matters enormously.

In situ grazing incidence wide-angle X-ray scattering tracked this process in real time during spin coating. The technique captures X-ray diffraction patterns as the film forms, revealing crystallization dynamics at sub-second resolution.

For PM6:BTP-TO2 blends, a consistent pattern emerged across all solvents. In the out-of-plane direction, the π-π stacking diffraction peak of PM6 (at 1.75 Å⁻¹) consistently exceeded that of BTP-TO2 (at 1.70 Å⁻¹) during the initial seconds of solvent evaporation. This indicated PM6 deposited at higher rates than BTP-TO2 regardless of which solvent evaporated.

This preferential precipitation proved crucial. PM6 formed a matrix first, with its rigid rod-like conformation determining the overall structural template. BTP-TO2 filled remaining spaces without disrupting PM6's architecture. The result: consistent morphology independent of solvent.

PM6:BTP-TC8 blends showed different behavior. In chlorobenzene, BTP-TC8 tended to deposit before PM6. In toluene, PM6 deposited first. The non-uniform molecular packing and intermolecular interactions observed in solution carried through to determine which component precipitated when. Different solvents produced different deposition sequences and hence different morphologies.

Molecular dynamics simulations rationalized these observations by calculating cohesive energy densities at increasing concentrations. Strong BTP-TO2-solvent interactions persisted even as concentration increased to simulate film formation. Weak BTP-TC8-solvent interactions made component deposition sensitive to evaporation kinetics.

The preferential precipitation of PM6 in PM6:BTP-TO2 systems minimized dependence on evaporation dynamics. The solvent could evaporate slowly or quickly, but PM6 still formed the structural framework first.

Morphological Consistency

The combination of stable acceptor conformation, rigid donor conformation, weak donor-acceptor interactions in solution, and preferential donor precipitation produced remarkably consistent final morphologies.

Static grazing incidence wide-angle X-ray scattering quantified crystalline structures in finished films. For PM6:BTP-TO2 blends processed from different solvents, the maximum fluctuation in lamellar and π-π stacking peak positions was only 0.02 Å⁻¹. Coherence crystalline lengths varied by only 0.6 nm.

PM6:BTP-TC8 blends showed much larger variations: 0.04 Å⁻¹ for peak positions and 5.6 nm for crystalline lengths. These numbers quantify how much molecular packing changes between solvents.

For semicrystalline organic semiconductors, slight distortions in crystalline structure don't significantly affect electrical properties—a fortunate circumstance that provides robustness. But large distortions alter charge transport, recombination, and collection.

Resonant soft X-ray scattering characterized average phase separation and domain sizes by probing contrast between donor-rich and acceptor-rich regions. PM6:BTP-TO2 blends showed center-to-center domain spacings ranging from 47.6 nm to 49.8 nm across different solvents. Average domain sizes ranged from 29.4 nm to 29.7 nm.

PM6:BTP-TC8 blends exhibited much broader ranges: 53.3 nm to 76.6 nm for domain spacing and 22.9 nm to 52.2 nm for domain size. These large variations indicate dramatically different morphologies depending on processing solvent.

The molecular-scale control established in solution and during film formation manifested in nano-scale morphological consistency crucial for device performance.

Uniform Excellence

The proof came from device fabrication. Organic solar cells with conventional architecture ITO/2PACZ/active layer/C₆₀/BCP/Ag were made using different solvents to process the PM6:BTP-TO2 active layer.

Chloroform-processed devices achieved 19.07% power conversion efficiency—a champion value that was certified independently. Chlorobenzene: 18.83%. Toluene: 19.13%. P-xylene: 18.96%. The fluctuation ratio across 20 cells for each condition remained under 2%.

This uniformity is extraordinary. For PM6:BTP-TC8, the best efficiency (17.61%) came from chloroform processing. Switch to chlorobenzene and efficiency dropped to 14.65%. Toluene: 15.29%. P-xylene: 14.39%. The same pattern seen throughout the field: strong solvent dependence.

External quantum efficiency spectra confirmed that integrated photocurrents matched short-circuit currents within 3%, validating the measurements. Voltage, current, and fill factor all remained stable across solvents for PM6:BTP-TO2 devices while varying substantially for PM6:BTP-TC8.

The consistency extended to operational stability. Devices held at maximum power point under continuous 100 mW cm⁻² illumination in nitrogen atmosphere maintained 80% of initial efficiency for over 1,200 hours regardless of processing solvent. PM6:BTP-TC8 devices showed varying stability depending on solvent, with an additional 10% decay compared to PM6:BTP-TO2 even under the best conditions.

Scaling Up

Laboratory achievements on small-area devices (0.06 cm²) don't always translate to large-area modules. The research team blade-coated 5 × 5 cm² active layers in ambient atmosphere and assembled modules comprising eight series-connected cells with total active area of 15.64 cm².

PM6:BTP-TO2 modules processed from chloroform, chlorobenzene, toluene, and p-xylene all delivered efficiencies around 16.1%. The toluene-processed module achieved 16.26%—among the highest reported for non-halogenated solvent-processed modules exceeding 10 cm² active area.

PM6:BTP-TC8 modules showed the familiar pattern: 14.31% for chloroform, dropping to around 10% for other solvents. Light-beam-induced current mapping revealed that PM6:BTP-TO2 modules displayed stronger and more uniform photocurrent distribution than PM6:BTP-TC8 modules.

The consistency across lab-scale and module-scale devices, across halogenated and non-halogenated solvents, demonstrated that the morphological design principles enable practical manufacturing flexibility.

Generalizing the Principles

To test whether the design strategy extended beyond BTP-TO2, the researchers synthesized BTP-TO3—increasing OEG chain length from two to three oxygen atoms. PM6:BTP-TO3 devices showed the same solvent tolerance, maintaining high efficiency across different processing solvents.

They also tested ID-OEG-2F, a small-molecule acceptor based on the indacenodithiophene core with one OEG side chain attached. When blended with PM6, this acceptor also produced similar device parameters across different solvents in conventional device structures.

These results validate the generality of the morphological control strategy: enhance side chain-solvent interactions to stabilize acceptor conformation, weaken donor-acceptor interactions to maintain rigid donor conformation, ensure preferential donor precipitation.

The specific molecular architecture—whether benzotriazole-based or indacenodithiophene-based—matters less than adherence to these design principles.

What Makes This Different

Previous attempts to understand morphology focused on solution aggregation at microscale: measuring how many molecules cluster together in solution. But this microscale aggregation doesn't directly determine the nanoscale morphology relevant to device physics.

Two materials might both dissolve without large aggregates yet produce different morphologies because their molecular conformations differ. Or they might show similar conformations but different precipitation kinetics. The gap between microscale observation and nanoscale outcome remained unbridged.

The current work connects molecular interactions to macroscopic outcomes through a comprehensive multilevel picture:

Molecular level: Side chain chemistry determines solvent interactions and conformational stability.

Solution level: Molecular conformation determines polymer shape and donor-acceptor association.

Film formation level: Polymer conformation and component interactions determine precipitation sequence.

Nano-scale morphology level: Precipitation sequence determines domain organization and crystalline packing.

Device level: Nano-scale morphology determines charge generation, recombination, transport, and collection.

Understanding propagates causally through these levels. Change side chain chemistry and you can predict consequences all the way to device efficiency.

The Green Solvent Future

Industrial organic solar cell production cannot rely on chlorinated solvents. Environmental regulations increasingly restrict their use. Worker safety concerns limit exposure. Waste disposal costs money. Moving to green solvents isn't optional—it's mandatory for commercialization.

But green solvents like toluene, xylene, or even more environmentally benign alternatives produce terrible morphologies with most current materials. Efficiency drops. Reproducibility suffers. Manufacturing becomes unpredictable.

This work demonstrates that the solvent sensitivity results from materials design rather than fundamental limitations. Thoughtful molecular engineering can create materials that perform equivalently across a wide solvent range.

The certified 19.06% efficiency of toluene-processed small-area cells and 16.26% efficiency of toluene-processed large-area modules prove that non-halogenated processing need not sacrifice performance. The door opens to manufacturing with solvents acceptable for industrial scale-up.

Remaining Challenges

While the demonstrated solvent tolerance represents major progress, practical commercialization requires addressing additional factors.

First, the study focused on one donor polymer (PM6) paired with engineered acceptors. Other high-performance donor materials might require different optimization strategies. The design principles should apply broadly, but validating them across many material combinations demands extensive work.

Second, the processing conditions—spin coating at specific speeds, thermal annealing temperatures, layer thicknesses—were optimized for each solvent. Industrial processing uses blade coating, slot-die coating, or roll-to-roll printing with different fluid dynamics and drying kinetics. Translating lab recipes to industrial processes remains challenging.

Third, long-term stability under realistic operating conditions—elevated temperatures, humidity exposure, UV illumination—requires further testing. The 1,200-hour stability under nitrogen with LED illumination provides encouraging initial data but doesn't fully represent outdoor deployment.

Fourth, module fabrication in ambient atmosphere achieved good efficiencies, but scaling to even larger areas introduces additional challenges in film uniformity, defect density, and contact resistance.

Fifth, the economic analysis comparing manufacturing costs with halogenated versus non-halogenated solvents must account for regulatory compliance, waste handling, and safety infrastructure—factors that may favor green solvents even if raw solvent costs are higher.

Lessons for Materials Design

The research provides concrete guidance for designing next-generation organic photovoltaic materials:

Choose side chains that interact strongly with solvents. Oligo(ethylene glycol) chains work well, but other polar, flexible side chains might also succeed. The key is creating favorable thermodynamic interactions that stabilize molecular conformation across different solvent environments.

Minimize donor-acceptor intermolecular interactions in solution. Bulky side chains providing steric hindrance help. Alternatively, modifying the conjugated backbone to reduce π-π stacking or dipole-dipole attraction between donor and acceptor could achieve similar results.

Design for preferential donor precipitation. This likely requires maintaining polymer rigidity in solution while ensuring acceptors remain well-solvated during early-stage film formation. The balance depends on polymer-solvent and acceptor-solvent interaction strengths.

Validate solution behavior experimentally. Nuclear magnetic resonance spectroscopy, neutron scattering, and absorption spectroscopy provide complementary information about molecular conformation, aggregation, and interaction states in solution. Computational simulations can predict interaction energies and conformational preferences.

Track film formation dynamics. In situ X-ray scattering during spin coating or blade coating reveals which component precipitates first and how crystallization proceeds. These real-time observations connect solution behavior to final morphology.

Characterize finished films comprehensively. Static X-ray scattering, resonant soft X-ray scattering, and grazing incidence small-angle scattering quantify crystalline structures, domain sizes, and phase separation at different length scales relevant to device physics.

The combination of experimental characterization and computational modeling at multiple length scales provides the understanding needed to design materials rationally rather than stumbling onto good performance through trial and error.

Beyond Organic Photovoltaics

The insights developed here extend beyond solar cells. Any organic electronic device relying on solution processing confronts similar morphology control challenges: organic light-emitting diodes, organic transistors, photodetectors, thermoelectrics.

For organic LEDs, emission efficiency depends on exciton formation and radiative decay rates determined by molecular packing and orientation. Different solvents produce different morphologies affecting light output and color purity.

For organic transistors, charge carrier mobility depends sensitively on crystalline domain connectivity and π-π stacking quality. Solvent choice dramatically impacts transport properties.

For photodetectors, dark current and spectral response depend on domain purity and interfacial structure. Morphological variations alter detector performance.

The general principle applies broadly: enhance side chain-solvent interactions to stabilize molecular conformation, control component interactions to manage co-assembly, understand film formation kinetics to predict final structures. These strategies should transfer across organic electronic platforms.

Even beyond electronics, any application involving solution-processed functional materials confronts morphology control: sensors, catalysts, separation membranes, drug delivery vehicles. The fundamental lesson—that thoughtful molecular design can decouple morphology from processing conditions—has wide relevance.

What Comes Next

The research team's success with BTP-TO2 and related acceptors opens multiple research directions.

First, exploring different polar side chain chemistries. Oligo(ethylene glycol) works well, but so might other groups: sulfones, amides, phosphates. Each offers different interaction profiles with various solvents. Systematic investigation could expand the solvent compatibility range further.

Second, applying the principles to donor polymer design. This work focused on acceptor modification while using an established donor. Designing donors with enhanced solvent-interaction side chains might provide additional morphological control.

Third, developing ternary blends incorporating the engineered acceptors. Adding a third component can fine-tune absorption spectra, energy levels, or morphological features. Ensuring ternary systems maintain solvent tolerance requires careful component selection.

Fourth, investigating tandem architectures stacking multiple subcells. Tandem devices can exceed single-junction efficiency limits but require compatible processing of each layer. Solvent-tolerant materials simplify tandem fabrication.

Fifth, integrating the materials into flexible substrates. Organic photovoltaics' main advantage over silicon lies in mechanical flexibility and lightweight properties. Demonstrating that solvent-tolerant materials maintain performance on plastic substrates under bending stress would advance practical applications.

Sixth, extending operational stability testing. While 1,200 hours under continuous illumination provides initial validation, commercial deployment requires 20+ year lifetimes. Identifying and mitigating degradation pathways specific to OEG-containing materials remains important.

The Path Forward

Organic solar cells have achieved remarkable efficiency gains over the past decade. But efficiency alone doesn't guarantee commercial success. Manufacturing scalability, cost, stability, and environmental compatibility all factor into economic viability.

The solvent problem represented a significant barrier. Materials optimized for laboratory-scale fabrication with toxic solvents couldn't translate to industrial production with acceptable solvents. Each new material required starting the solvent optimization from scratch.

By establishing design principles for solvent-insensitive morphology, this work removes a critical roadblock. Materials can be engineered for robust processing across different solvents. Manufacturers gain flexibility in solvent choice based on cost, safety, environmental regulations, and processing requirements rather than being locked into whatever solvent happened to work for a particular material.

This doesn't mean organic photovoltaics will immediately replace silicon. But it means one fewer obstacle stands in the way. Combined with continuing efficiency improvements, stability advances, and manufacturing scale-up, the path toward commercial organic solar cells becomes clearer.

The broader lesson transcends any specific technology. Understanding materials at multiple length scales—from molecular interactions through nanoscale assembly to macroscopic properties—enables rational design. Empirical optimization gives way to predictive engineering. Performance stops being an accident of processing and becomes a controlled outcome of deliberate molecular architecture.

That represents progress.

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/10.1038/s41560-024-01678-5