Computational discovery of bifunctional organic semiconductors for energy and biosensing

This study presents a high-throughput computational framework that successfully identifies seven bifunctional organic semiconductors, including an optimal candidate with 36.1% predicted power conversion efficiency and strong protein-binding affinity, by balancing photovoltaic performance against synthetic accessibility.

The Gist

Imagine you are a chef trying to invent a new, super-delicious, and healthy dish. You have a massive library of 17,458 potential ingredient combinations (molecules). Your goal is to find the perfect recipe that does two things at once:

1. Generates energy (like a solar panel).
2. Detects diseases (like a medical sensor).

The problem? Most recipes in the library are either impossible to cook (too complex to make in a lab) or they taste terrible (don't work well). This paper is the story of how the authors used a "super-smart kitchen assistant" (a computer program) to find the few perfect recipes that are both delicious and easy to cook.

Here is the breakdown of their discovery, explained simply:

1. The "Impossible" Search

Usually, finding a new material for solar panels or medical sensors is like looking for a needle in a haystack. Scientists often find a molecule that works amazingly well in a computer simulation, but when they try to build it in a real lab, they realize it's too expensive or too difficult to manufacture. It's like finding a recipe that requires a rare, extinct fruit and a 50-step process involving a rocket ship.

2. The "Smart Kitchen Assistant" (The Computer Model)

The authors built a computational framework (a digital filter) to sift through 17,458 molecules from a public database. They didn't just look for the "best" performance; they looked for the best balance.

They created a special scoring system called PCESAScore. Think of it like a "Value for Money" rating:

• PCE (Performance): How well does it generate electricity? (The "Taste").
• SAScore (Synthetic Accessibility): How easy is it to make? (The "Cost/Effort").

The Formula: Score = Performance - Difficulty
If a molecule is amazing but impossible to make, the score is low. If it's easy to make but useless, the score is low. They wanted the "Goldilocks" molecules: good enough to work, but simple enough to build.

3. The "Double-Action" Discovery

Most materials are specialists. Some are great at catching sunlight (solar panels), while others are great at sticking to viruses (biosensors). The authors were looking for bifunctional molecules—materials that are "Swiss Army Knives."

They found 7 molecules that could do both jobs.

• The Solar Star: One molecule (ID 17851) was predicted to be incredibly efficient at turning sunlight into electricity (up to 36% efficiency, which is huge!).
• The Medical Detective: Another molecule (ID 1712) was found to be excellent at "hugging" specific proteins, like the HIV virus or cancer-related proteins. This means it could be used to build sensors that detect diseases early.

4. The "Magic Ingredient" (Nitrogen)

The authors noticed a pattern in these 7 winners. They were all rich in Nitrogen.

• Analogy: Think of Nitrogen atoms as "universal connectors." In the world of solar panels, they help electrons move smoothly (like a well-paved highway). In the world of medicine, they act like "Velcro hooks" that can grab onto specific biological targets (like viruses).
• By tweaking the Nitrogen content, the molecules could be tuned to be both efficient energy harvesters and sensitive medical detectors.

5. Why This Matters (The "So What?")

This research changes the game in two ways:

• It saves time and money: Instead of chemists spending years building and testing thousands of molecules that fail, they can use this computer method to pick the top 7 winners immediately. It's like using a metal detector instead of digging the whole beach with a spoon.
• It enables "Smart" Devices: Imagine a future where a single piece of plastic film on a window does two things: it powers your house with solar energy and simultaneously monitors the air for viruses or pollutants. These 7 molecules are the first step toward making that "magic film" a reality.

The Bottom Line

The authors didn't just find a few good molecules; they built a new map for discovering materials. They proved that you don't have to choose between "high performance" and "easy to make." By using a smart computer filter, they found the sweet spot where science meets practicality, opening the door to cheaper, smarter solar panels and life-saving medical sensors.

Technical Summary

Here is a detailed technical summary of the paper "Computational discovery of bifunctional organic semiconductors for energy and biosensing."

1. Problem Statement

The discovery of organic semiconductors (OSCs) faces a critical bottleneck: identifying molecules that simultaneously exhibit exceptional photovoltaic performance (high Power Conversion Efficiency, PCE) and practical synthetic accessibility.

• The Trade-off: High-performance OSCs often require complex, multi-step syntheses that are cost-prohibitive or difficult to scale. Conversely, easily synthesized molecules often lack the electronic properties required for efficient energy conversion.
• The Gap: Existing high-throughput screening methods often prioritize theoretical performance (PCE) while treating synthetic feasibility as a post-hoc constraint, leading to candidates that are theoretically promising but experimentally unrealizable. Furthermore, few studies explore the dual functionality of OSCs for both energy harvesting (photovoltaics) and biosensing (protein recognition).

2. Methodology

The authors developed a comprehensive, high-throughput computational framework integrating Machine Learning (ML), Density Functional Theory (DFT), and Molecular Docking to screen 17,458 organic molecules from the PubChemQC dataset (specifically the CHNOPSFClNaKMgCa500 subset).

A. The Screening Workflow

The process involved four sequential filtering stages:

1. Frontier Orbital Alignment: Molecules were filtered based on HOMO/LUMO energy levels to ensure compatibility with benchmark acceptors (PCBM) and donors (PCDTBT).
   • Donors: HOMO < -5.02 eV, LUMO > -4.36 eV.
   • Acceptors: HOMO > -5.50 eV, LUMO < -3.45 eV.
2. Performance Prediction: Power Conversion Efficiency (PCE) was estimated using the Scharber model, enhanced with dynamic fill factor corrections.
3. Synthetic Accessibility Filtering: The SAScore (Synthetic Accessibility Score) was calculated. A composite metric, PCESAScore, was introduced:
   $$\text{PCESAScore} = \text{PCE} - \text{SAScore}$$
   Candidates were retained only if $\text{PCESAScore} > 0$, ensuring a balance between high efficiency and low synthetic complexity.
4. Multifunctional Assessment: Top candidates underwent:
   • Excited-state analysis: Using TDDFT (CAM-B3LYP/6-31+G*) and xTB/sTDA for charge transfer characteristics.
   • Protein Binding: Molecular docking (AutoDock Vina) against therapeutically relevant targets (HIV-1 Protease, Proteasome, COVID-19 related proteins).
   • Charge Transport: Calculation of reorganization energies ($\lambda$) for hole and electron transport.

B. Computational Details

• Geometry Optimization: PM6 semi-empirical geometries from PubChemQC were used as starting points, followed by single-point energy calculations at the B3LYP/6-31G* level.
• Validation: The workflow was validated against 25 literature benchmarks, achieving a correlation ($R^2$) of 0.626 for PCE predictions. Statistical robustness was confirmed via bootstrap resampling (1,000 iterations).

3. Key Contributions

• Novel Metric (PCESAScore): The first study to quantitatively combine PCE predictions with SAScore into a single selection metric, systematically prioritizing molecules that are both efficient and synthetically accessible.
• Dual-Function Discovery: Successfully identified OSCs capable of serving as both high-efficiency photovoltaic materials and high-affinity biosensors, bridging the gap between energy and healthcare applications.
• Hierarchical Screening Efficiency: Demonstrated a workflow that reduced the chemical space from ~17,500 molecules to 7 top candidates (a >99.9% reduction) while retaining the most promising multifunctional materials.
• Open Science: All data, scripts, and workflows are publicly available via GitHub and Zenodo, adhering to FAIR principles.

4. Key Results

The screening identified seven exceptional candidates that balance high PCE with low SAScore (< 4.0).

A. Top Candidates

• Molecule 17851 (Donor):
  • PCE: Predicted 36.11% (with PCDTBT).
  • Synthesis: SAScore = 7.62 (Moderate complexity).
  • Structure: Features a fused pyrrole-furan core with nitrogen-rich substituents.
  • Properties: High ligand efficiency and strong binding to multiple protein targets.
• Molecule 4550 (Donor):
  • PCE: 16.89% (with PCDTBT).
  • Ligand Efficiency: 0.340 kcal mol⁻¹/heavy atom (exceeding pharmaceutical standards).
  • Application: Identified as optimal for charge transport and versatile optoelectronic applications.
• Molecule 1712 (Acceptor):
  • PCE: 8.50% (with PCBM).
  • Biosensing: Exhibited the strongest binding affinities, specifically to the Proteasome (-9.7 kcal mol⁻¹) and HIV-1 Protease (-7.6 kcal mol⁻¹).
  • Application: Promising candidate for fluorescent biosensors.

B. Structure-Property Relationships

• Heteroatom Distribution: Acceptor-type molecules showed significantly higher oxygen content, while donor-type molecules favored nitrogen incorporation and structural flexibility (rotatable bonds).
• Flexibility vs. Rigidity: The study found that strategic conformational flexibility (often associated with rotatable bonds) enhances molecular recognition for biosensing without critically compromising charge transport, challenging the traditional view that OSCs must be strictly rigid.
• Nitrogen's Dual Role: Nitrogen-rich frameworks were found to simultaneously elevate HOMO levels (beneficial for PCE) and provide hydrogen-bonding sites for protein binding.

C. Statistical Validation

• Uncertainty Quantification: Bootstrap analysis confirmed the stability of the rankings, with standard errors of ±0.12 eV for HOMO-LUMO gaps and ±2.3% for PCE.
• Reorganization Energy: While some candidates showed high hole reorganization energies (suggesting potential mobility limits), others (like 17851) showed favorable characteristics for charge transport.

5. Significance and Future Outlook

• Paradigm Shift: This work shifts the materials discovery bottleneck from "computation" to "laboratory synthesis" by embedding synthetic feasibility as a hard constraint from the outset.
• Integrated Devices: The identification of bifunctional molecules opens the door for self-powered biosensors—devices that can harvest energy from sunlight while simultaneously detecting biological signals (e.g., viral proteins or enzymes).
• Experimental Roadmap: The authors propose a clear validation path for the top candidates (specifically Molecule 17851), including scale-up synthesis, thin-film photophysical characterization, and pilot device integration.
• Generalizability: The PCESAScore framework and the hierarchical screening workflow are applicable to other classes of multifunctional materials beyond organic photovoltaics.

In conclusion, this paper provides a robust computational pipeline that successfully navigates the trade-off between performance and synthesizability, delivering a shortlist of experimentally viable organic semiconductors ready for the next stage of materials development.