Electrospinning-Data.org: A FAIR, Structured Knowledge Resource for Nanofiber Fabrication

This paper introduces Electrospinning-Data.org, a FAIR-aligned, structured knowledge resource that aggregates diverse and failure-inclusive electrospinning experimental data into a machine-readable format to overcome reporting inconsistencies and enable data-driven research, predictive modeling, and inverse design in nanofiber fabrication.

The Gist

Imagine you are trying to bake the perfect loaf of bread. You know that the taste and texture depend on a complex mix of ingredients (flour, water, yeast), your technique (kneading time, temperature), and your environment (humidity, altitude).

Now, imagine that every baker in the world who tries this recipe writes down their results in a different notebook. Some write in code, some just say "it was good," and most importantly, nobody writes down what happened when the bread burned or turned into a rock. They only share the winning recipes.

If you want to bake the perfect loaf, you have to read thousands of these messy, inconsistent notebooks, guess what the other bakers meant, and hope you don't accidentally burn your own bread. This is exactly the problem scientists face with electrospinning—a high-tech method used to make ultra-thin fibers for things like medical bandages, air filters, and energy batteries.

Here is how the paper explains the solution: Electrospinning-Data.org.

1. The Problem: A Library of Lost Recipes

Currently, electrospinning research is like a library where:

• The books are written in different languages: One scientist says "high voltage," another says "strong spark." It's hard to compare them.
• The "failures" are missing: If a scientist tries a setting and gets a tangled mess of fibers instead of a perfect sheet, they usually throw that data in the trash. They only publish the "successes."
• The result: Researchers waste years repeating the same mistakes because they can't see the "do not enter" signs that others have already discovered.

2. The Solution: A Universal, Smart Recipe Book

The authors built Electrospinning-Data.org, which acts like a universal translator and a strict librarian for fiber-making experiments.

Think of it as a giant, digital "Recipe Book" that forces everyone to write their recipes in the exact same format.

• Standardized Ingredients: Instead of saying "a bit of heat," the system forces you to enter "25°C."
• The "Burnt Bread" Section: This is the most important part. The system has a special section for failures. If your fibers broke or got stuck, you record it there. This tells future bakers, "Don't try this combination; it leads to a disaster."
• The Dictionary: They created a special dictionary (called Cogni-EMCV) so that when someone says "smooth fiber," the computer knows exactly what that looks like, rather than guessing based on a vague description.

3. How It Works: The "Quality Control" Line

You can't just dump any data into this book; it has a two-step security check to make sure the information is real and useful:

1. The Robot Check: A computer program instantly checks your entry. Did you type a negative number for temperature? Did you forget to say what material you used? If yes, the robot says, "Fix this!"
2. The Expert Check: If the robot is happy, a human scientist (an expert) reviews it. They make sure the experiment makes sense physically. Did the voltage actually work with that flow rate? If it looks suspicious, they ask for clarification.

Only data that passes both checks gets added to the public library.

4. Why This Matters: From Guessing to Knowing

Before this platform, finding the right settings for a new fiber was like playing a video game with a blindfold on, trying random buttons until you win.

With Electrospinning-Data.org, it's like having a cheat sheet.

• For a researcher: They can ask the computer, "Show me all the times people made PVA fibers (a common plastic) that were between 200 and 300 nanometers wide."
• The Result: The computer instantly gives them a list of 100+ experiments, including the ones that failed, so they know exactly which settings to try first.

The Big Picture

This isn't just a database; it's a community effort to stop reinventing the wheel. By organizing messy, scattered data into a clean, structured, and "failure-aware" system, the authors hope to speed up discoveries in medicine, energy, and technology.

Instead of spending years figuring out what doesn't work, scientists can finally focus on what does, using a shared map of the entire landscape of fiber-making. It turns a chaotic jungle of experiments into a well-organized garden where everyone can grow better results.

Technical Summary

1. Problem Statement

Electrospinning is a versatile nanofabrication technique used to produce continuous nanofibers for applications ranging from tissue engineering to energy storage. However, optimizing this process is hindered by a high-dimensional parameter space involving complex interactions between solution properties, processing parameters, and environmental conditions.

The field faces three critical data-centric challenges:

• Fragmentation and Inconsistency: Experimental data is scattered across thousands of publications with inconsistent reporting standards. A prior review found that only ~44% of papers contained sufficiently reported quantitative parameters for structured extraction.
• Publication Bias (Success Bias): Literature overwhelmingly reports successful outcomes, while negative results, instability regimes, and process failures are rarely documented. This lack of "failure data" introduces systematic bias, limiting the reliability of statistical analysis and machine learning models.
• Unstructured Reporting: Morphological descriptions often rely on subjective, narrative terminology rather than standardized, machine-readable formats, making cross-study comparison and automated data mining difficult.

Existing resources (e.g., FEAD, Cogni-e-SpinDB 1.0) are either static, limited to specific polymers, or lack continuous ingestion and failure-aware schemas.

2. Methodology

The authors developed Electrospinning-Data.org, a dynamic, FAIR-aligned (Findable, Accessible, Interoperable, Reusable) data infrastructure designed to aggregate, structure, and curate electrospinning experiments.

A. Data Representation and Schema

The platform utilizes a unified Process–Structure–Property framework:

• Relational Database: Built on MySQL, the schema links experimental inputs (materials, solution prep, process parameters, environment) to outcomes (fiber morphology, properties, instabilities).
• Hybrid Design: Uses a combination of relational tables for structured data and JSON fields for flexible equipment configurations and image metadata.
• Controlled Vocabulary (Cogni-EMCV): To standardize morphology, the authors introduced the Cognitive Electrospinning Morphology Controlled Vocabulary. This organizes morphology into seven orthogonal axes (Shape, Topography, Composition, Texture, Defects, Size, Size Variation), converting narrative descriptions into machine-readable, versioned codes.
• Failure-Aware Schema: Explicitly includes tables for "Process Instability" and "Defects," treating negative outcomes as first-class data points rather than omissions.

B. Data Ingestion and Governance

The platform operates on a two-stage moderation pipeline to ensure data quality:

1. Automated Validation (Cogni-EVVR): Submissions are checked against schema rules (mandatory fields) and physical constraint rules (e.g., voltage must be non-zero, humidity 0–100%).
2. Expert Review: Domain experts assess physical plausibility, parameter coherence, and documentation quality.

• Sources: Data is ingested via retrospective extraction from peer-reviewed literature and prospective direct submission by the research community via a web interface.

C. System Architecture

The platform is a modular, three-layer system:

• Data Storage: MySQL relational database with versioned releases.
• Backend: Spring Boot (Java) handling RESTful APIs, validation logic, and access control.
• Frontend: React.js interface for data exploration, filtering, and structured submission.
• Deployment: Containerized via Docker and served via Nginx over HTTPS.

3. Key Contributions

1. First Continuous, Failure-Aware Infrastructure: Unlike static datasets, Electrospinning-Data.org is a live platform that continuously ingests data and explicitly captures instability and failure regimes, addressing the "success bias" in literature.
2. Standardized Morphology Ontology: The introduction of Cogni-EMCV provides a versioned, controlled vocabulary for fiber morphology, enabling consistent semantic representation across studies.
3. Unified Process–Structure–Property Model: A machine-readable schema that links inputs (voltage, flow rate, humidity) directly to outputs (diameter, defects, mechanical properties), facilitating inverse design.
4. Governed Curation Workflow: A hybrid automated/expert validation system (Cogni-EVVR) that balances openness with rigorous data quality control.
5. Image Integration: Unlike previous datasets, this platform links experimental records to SEM and morphology images, providing visual provenance.

4. Results

• Initial Dataset Release: The platform launched with 809 experimental records derived from the Cogni-e-SpinDB 1.0 dataset, covering 12 polymer systems (e.g., PVDF, PVA, PAN) and 14 solvent systems from 57 publications.
• Data Coverage: Core process parameters (voltage, flow rate, distance) are present in 100% of records. Environmental parameters remain sparsely reported but are structurally supported.
• Demonstrated Utility: A use case involving Polyvinyl Alcohol (PVA) nanofibers demonstrated the platform's capability. By filtering for specific parameters (water solvent, flat collector, 180–380 nm diameter), the system instantly retrieved 108 records. This allowed researchers to identify median processing conditions (e.g., 20 kV, 0.30 mL/h) and practical bounds without manual literature extraction.
• Interoperability: The data is available in structured Excel formats and ZIP archives containing linked images, with full versioning and DOI-based provenance tracking.

5. Significance

Electrospinning-Data.org represents a paradigm shift from narrative-based research to data-driven materials science in the field of nanofiber fabrication.

• Enabling Machine Learning: By providing a structured, failure-inclusive corpus, the platform removes the "data wrangling" bottleneck, allowing researchers to train predictive models for fiber morphology and stability that were previously impossible due to data scarcity and bias.
• Accelerating Discovery: Researchers can perform systematic inverse design (finding process parameters for a target morphology) and map instability regimes, significantly reducing trial-and-error experimentation.
• Community Standardization: The platform establishes a governance model and vocabulary (Cogni-EMCV) that encourages the community to report negative results and standardize reporting, fostering reproducibility.
• Scalability: The architecture is designed to evolve, with future plans to integrate automated image analysis and expand the ontology, positioning the platform as a central, shared infrastructure for the global electrospinning community.

In summary, the paper presents a critical infrastructure solution that transforms fragmented, narrative electrospinning data into a structured, FAIR-compliant knowledge graph, directly addressing the reproducibility crisis and enabling the next generation of data-driven nanofabrication research.