Arduino Colorimetric Reagent Dispenser Science Project

ISEF Category: Chemistry

Subcategory: Other  ·  Difficulty: Advanced  ·  Setup: School Lab  ·  Time: 1 to 2 Months

The Hook

A tiny change in how you mix reagents can flip a weak color into a strong one. That means the machine, not just the chemistry, can decide how good your result looks. If you can automate the mixing, you can also test which recipe works best. That turns a kitchen-scale gadget into a real research tool.

What Is It?

This project is about building a small automated system that mixes liquids for you and then helps choose the best recipe for a color test. Think of it like a smart blender for chemistry. Instead of guessing the right combination by hand, you let the machine try different reagent settings and measure which one gives the strongest or cleanest color signal.

The color signal matters because many chemistry tests turn a substance into a visible dye or pigment. Prussian blue, for example, is a deep blue material that can show up when certain chemicals react. You can measure that blue with a phone camera or a color sensor, then compare how different mixing plans change the result. Bayesian optimization means the software uses earlier results to suggest the next test, so you do not waste as many trials.

Why This Is a Good Topic

This is a strong science fair topic because you can test one clear thing, how mixing choices affect color output, and you can measure that with data. It connects chemistry, automation, and data analysis, which gives you room to show real problem-solving. You can also build it with accessible parts and still make it look original by designing your own control logic, calibration method, and comparison rules.

Research Questions

• How does reagent ratio affect the intensity of the Prussian blue color signal?
• What is the effect of pump dispensing order on final color uniformity?
• Does automated dispensing produce more repeatable color results than hand mixing?
• To what extent does Bayesian optimization reduce the number of trials needed to find the strongest color signal?
• Which measurement method, phone camera RGB values or a color sensor, gives the most stable readout?
• How does tubing length variation affect dosing accuracy in a peristaltic pump system?

Basic Materials

• Arduino board or compatible microcontroller board.
• Peristaltic pumps, one for each liquid line.
• Silicone tubing sized for the pump heads.
• Small reagent reservoirs or bottles with caps.
• Breadboard, jumper wires, and a basic wiring kit.
• MOSFETs or pump driver modules rated for the pumps.
• Stable power supply matched to the pumps.
• Phone camera or USB color sensor.
• White background or light box for consistent imaging.
• Digital kitchen scale with 0.1 g accuracy.
• Standard lab beakers, cups, or clear test vials.
• Distilled water.
• Safety glasses and disposable gloves.

Advanced Materials

• Arduino or other microcontroller with data logging support.
• Peristaltic pumps with known flow curves.
• Multi-channel pump manifold or custom tubing rack.
• Colorimeter or spectrophotometer for absorbance measurements.
• DSLR camera or fixed smartphone mount for image capture.
• Reference color standards or calibration targets.
• Magnetic stir plate and stir bars.
• Analytical balance.
• pH meter.
• Temperature probe.
• Lab-grade glassware.
• Software for Bayesian optimization and statistical analysis.

Software & Tools

• Arduino IDE: Programs the microcontroller that controls the pumps and logs basic output.
• Python: Runs the optimization loop, processes images, and analyzes trial data.
• ImageJ: Measures color intensity from images of each reaction sample.
• Jupyter Notebook: Keeps your code, plots, and notes in one place while you test ideas.
• Excel: Organizes trial data and helps you make quick charts and summaries.

Experiment Steps

1. Define the color reaction you will optimize and the one measurement that counts as success.
2. Map the hardware pieces you need, then decide how each pump will connect to the control board and power source.
3. Build a calibration plan so you can link pump output and color signal to real numbers.
4. Choose your variables, your controls, and the trial range you will let the optimizer explore.
5. Design the data pipeline that moves each test result into a score the software can use for the next recipe suggestion.
6. Plan a comparison that shows whether automation or Bayesian search improves repeatability, signal strength, or both.

Common Pitfalls

• Using pumps that do not dispense at a steady rate, which makes every recipe inconsistent.
• Measuring color with changing room lighting, which makes the same sample look different across trials.
• Skipping pump calibration, which hides dosing errors behind fake chemistry trends.
• Changing too many variables at once, which makes it impossible to tell what caused the color change.
• Letting the optimizer chase one noisy result, which can make the software learn bad recipes.

What Makes This Competitive

A strong version of this project does more than make a gadget work. You should show that your system produces better data than manual mixing, or that the optimizer finds good recipes faster than a simple grid search. You can raise the level by reporting uncertainty, repeatability, and failure cases, not just the best-looking sample. A clear comparison between different sensing methods or different optimization strategies can make the project stand out.

Project Variations

• Test the same automation system with food dyes instead of Prussian blue to compare how viscosity and color depth affect measurement.
• Swap the phone camera for a low-cost color sensor and compare which method gives tighter repeatability.
• Use the same setup to optimize a different reaction, such as a pH indicator system, and compare how easy each chemistry is to control.

Learn More

• MIT OpenCourseWare: Search for general chemistry, laboratory methods, and control systems lectures that explain measurement, calibration, and experimental design.
• NIH PubChem: Search compound pages and reaction-linked entries to check chemical properties and safety data.
• NIST Chemistry WebBook: Look up physical data, spectra, and reference values for common compounds.
• ImageJ: Search the official documentation for tutorials on measuring color intensity from images.
• Python documentation: Read the NumPy, SciPy, and scikit-learn docs for data handling, curve fitting, and simple Bayesian optimization workflows.