Why we built AQUA-MIND

It started with problems we ran into during photocatalysis experiments.

WHAT IS AQUA-MIND

A photocatalysis lab notebook shared by researchers worldwide and AI

AQUA-MIND is a web platform where researchers around the world record and compare photocatalytic water purification experiments in the same way.

When conditions and measurements are collected in a common format, results from different labs can be compared side by side. This data becomes evidence for designing the next experiment and, once there is enough, for AI predictions.

AI joins from the start as a research partner: it finds data errors, explains results and suggests next experiments. Judgment and responsibility always stay with researchers.

HOW IT WORKS

How it works

  1. 01

    Run the same protocol

    Degrade methyl orange with ZnO following common protocol v1.0, together with three controls.

  2. 02

    Submit data with an AI quality check

    Upload the CSV template; it is checked for format, ranges and missing values, and AI points out easy-to-miss problems.

  3. 03

    Analyze with AI interpretation

    Code calculates concentration change, removal and the rate constant k; AI explains the result against the controls.

  4. 04

    Share and plan the next run

    Publish data with its source and license, and decide the next experiment with AI's suggestions as input.

Researchers experiment and decide. AI checks and analyzes.

  1. 01 · ResearcherRun the experimentExperiment and measure with the common protocol.
  2. 02 · ✦ AIData quality checkFinds unit errors, calibration range issues and outliers.
  3. 03 · ResearcherReview and shareFix what was flagged and publish the data.
  4. 04 · ✦ AIPatterns and interpretationExplains results and compares similar experiments.
  5. 05 · Researcher + ✦ AI suggestionDesign the next runAI suggests conditions; the researcher decides.
Researcher step✦ AI stepAs data from around the world grows, AI adds predictions and experiment suggestions.
WHO IT'S FOR

Who uses it, and how

Secondary-school research teams

Calculate and chart results and compare AI interpretations as evidence for inquiry reports.

Teachers and schools

Run lessons and joint projects that pool experiments from several classes or schools using one protocol.

University and institute labs

Share data in a standard format and check reproducibility against other labs.

Anyone curious about environmental science

Explore open experiment data and calculations to understand photocatalytic water purification.

Why ZnO and methyl orange?

ZnO (zinc oxide)
A widely studied photocatalyst that reacts under UV light and is easy to obtain, so school labs can work with it too.
Methyl orange
A colored dye whose concentration is easy to track by absorbance, so it is often used as a model pollutant in photocatalysis research.
Why start narrow
Fixing one catalyst and one pollutant makes results comparable across labs. We will expand to others as data grows.

For data you can trust

  • Every dataset is labeled measured, literature or virtual.
  • Values that were not measured are never replaced with 0; they stay empty with a reason.
  • Uploading alone never makes data 'verified'; format, source and scientific review are separate steps.
  • AI predictions stay off until the release criteria are met (100 measured experiments, 5 independent labs and more).

Coming next

  • Phase 2: accounts, server storage, discussion and review, live AI
  • Phase 3: joint experiment campaigns across countries, dataset DOIs
  • Phase 4: AI that predicts results with uncertainty, once enough measured data exists
Roadmap

Problems we want to solve

  • Conditions and results are hard to compare in a consistent format.
  • Without sources and methods, results are hard to reproduce and interpret.
  • Labs use different methods, so pooled results are hard to compare.
  • With little data or different conditions, AI predictions are hard to trust.

These are hypotheses to test through the service; we do not assume every researcher faces them.

Our story

[Written by the research team — interest in the local environment, the photocatalysis experiments actually performed, problems found along the way. Only what was really done and verified.]

Initial scope and exclusions

Included: ZnO × methyl orange, example data analysis, submission drafts, AI research partner demo, KO/EN

Excluded: guarantees about real rivers or seas, unvalidated AI predictions, recommendations for real-environment use

Frequently asked questions

Is AI really analyzing my data now?

In phase 1 the AI assistants are a demo: they insert code-calculated results into fixed sentences and carry an 'AI demo' label. Live AI comes after phase 2.

How is my data used?

The default terms are CC BY 4.0: anyone may use it with attribution, and you decide when it becomes public.

Can students take part?

Yes, under a teacher's or PI's supervision. Under-14s need guardian consent.

Can these results clean real rivers or seas?

Not directly. Catalyst recovery, by-products and effects on living things need separate safety validation.

Why no AI predictions yet?

With little data or different conditions, predictions are hard to trust. They will open once enough measured data from many labs exists.

Start analyzingJoin the research