Wilson Cloud Chamber Muon Track Study

ISEF Category: Physics and Astronomy

Subcategory: Nuclear and Particle Physics  ·  Difficulty: Advanced  ·  Setup: Home Setup  ·  Time: Full Year

The Hook

A muon can pass through you right now and leave no trace. In a cloud chamber, though, that same particle can draw a tiny white line in vapor. You can turn that into a real research project with physics, coding, and altitude data. The hard part is not seeing tracks, it is proving what they mean.

What Is It?

A Wilson cloud chamber is a box where charged particles leave visible trails as they pass through supersaturated vapor. Think of it like fog that only appears where something disturbs the air. When a particle zips through, it helps tiny droplets form along its path, and you see a line.

Different particles make different track shapes. Muons often leave long, straight tracks because they are heavy and fast. Beta particles, which are electrons, usually scatter more and make thinner, messier paths. That gives you a chance to compare track topologies, which means the shapes and patterns of the trails.

The machine learning part adds another layer. You can train a CNN, a convolutional neural network, on simulated tracks from Geant4, then test whether it can classify your real chamber images. That lets you compare what your model learned from simulations with what happens in your setup.

Why This Is a Good Topic

This topic works well for a science fair because you can test a real physics idea, collect your own data, and add a strong analysis layer. You can change altitude, track type, image quality, or chamber conditions and measure how the results shift. It connects to cosmic rays, particle detection, and computer vision, all with a clear physical signal you can photograph. A student can learn experimental design, image analysis, and basic machine learning without needing a university detector.

Research Questions

• How does altitude affect the observed muon track rate in a Wilson cloud chamber?
• What is the effect of chamber lighting on the number of tracks a CNN can classify correctly?
• Does training a CNN on Geant4 simulations improve classification of real cloud chamber tracks?
• To what extent do beta tracks from a mentor-supervised thoriated welding rod differ from muon tracks in length and curvature?
• Which image preprocessing steps most improve track detection in cloud chamber photos?
• How does camera angle affect measured track topology features in a cloud chamber?

Basic Materials

• Fish tank or clear acrylic container
• Metal tray or cold plate for the chamber base
• Dry ice
• Isopropanol, high purity if possible
• Black felt or foam lining
• LED light source with steady output
• Dark background material
• Smartphone or digital camera
• Tripod or phone stand
• Ruler or calibration grid
• Mentor-supervised thoriated welding rod source
• Insulated gloves and tongs
• Safety goggles
• Timer
• Data notebook

Advanced Materials

• Geant4 simulation software
• Access to a desktop or laptop that can run CNN training
• Python development environment
• ImageJ for image inspection and measurement
• OpenCV for image preprocessing
• USB camera with manual exposure control
• Weather and altitude logging tools
• More stable cold plate or Peltier setup
• Shielded sample holder for source imaging under mentor supervision
• Statistical analysis software or Python libraries

Software & Tools

• Python: Runs image processing, feature extraction, and model training for your track classifier.
• Jupyter Notebook: Keeps your code, plots, and notes in one place while you test ideas.
• ImageJ: Lets you inspect chamber images, measure track length, and compare track contrast.
• Geant4: Simulates particle tracks you can use to train and validate a CNN.
• OpenCV: Helps you crop, denoise, threshold, and segment cloud chamber images.

Experiment Steps

1. Define the one claim you want to test first, such as altitude effects, track classification, or both together.
2. Choose the image features or track shapes your chamber can measure reliably, then match those features to your research question.
3. Plan a repeatable imaging setup so your camera, lighting, and background stay consistent across sessions.
4. Build a labeled dataset by separating clearly seen muon-like tracks, beta-like tracks, and ambiguous tracks.
5. Create a simulation-to-reality plan so your Geant4 data and chamber photos use comparable labels and image formats.
6. Decide on your statistics before you collect data, so you know how you will compare track rates across altitude and model outputs.

Common Pitfalls

• Using changing room light or flashlight angle, which makes the same track look brighter or dimmer across trials.
• Mixing up condensation streaks, dust, or frost lines with actual particle tracks.
• Collecting too few clear images, which leaves the CNN with weak training data and unstable results.
• Comparing altitude runs without keeping the chamber setup, camera settings, and cold-plate conditions consistent.
• Treating all long tracks as muons, which ignores background artifacts and weakens the physics claim.

What Makes This Competitive

A strong version of this project goes beyond making a chamber and counting streaks. You would need clean controls, a clear labeling method, and a thoughtful way to compare simulation with real images. A better project also checks whether your classifier works across different image conditions, not just on the data it saw first. If you can pair altitude measurements with error bars and a careful discussion of false positives, the work starts to look much stronger.

Project Variations

• Use a smartphone camera and compare how image resolution changes track classification accuracy.
• Focus only on altitude and test whether track rate changes between low, mid, and high elevation sites.
• Skip machine learning and build a rule-based track classifier from length, straightness, and brightness features.

Learn More

• NASA Cosmic Ray resources: Search NASA sites for educational pages on cosmic rays and how they are detected.
• CERN Geant4 documentation: Find the official Geant4 user guides and examples for particle simulation.
• PubMed: Search for review articles on radiation tracks, particle detectors, and cloud chamber imaging.
• MIT OpenCourseWare, Introduction to Nuclear and Particle Physics: Use course materials to review particle interactions and detector basics.
• NOAA altitude and weather data: Use NOAA resources to compare altitude, pressure, and atmospheric conditions across your test sites.