Indoor PM2.5 Infiltration Across Building Types

ISEF Category: Earth and Environmental Sciences

Subcategory: Atmospheric Science  ·  Difficulty: Advanced  ·  Setup: School Lab  ·  Time: Full Year

The Hook

Tiny particles do not stay outside just because you close the door. They slip through cracks, vents, and open windows, then build up inside like invisible smoke. If you can measure that leak, you can compare homes, classrooms, and apartments in a way that actually matters for health.

What Is It?

PM2.5 means tiny airborne particles that are 2.5 micrometers wide or smaller. That is about thirty times thinner than a human hair. These particles come from traffic, wildfire smoke, cooking, candles, and dust. Your project asks a simple question with a tricky answer, how much outdoor PM2.5 gets inside, and how much of the indoor reading comes from indoor sources?

Think of a building like a bucket with holes. Outdoor air pours in through the holes, but people also keep adding water from inside by cooking, burning candles, or running a vacuum. The infiltration coefficient tells you how much outside air makes it indoors. A state-space Kalman filter helps you separate the two streams, the outside contribution and the indoor-source contribution, so your estimate is not fooled by dinner or a birthday candle.

Why This Is a Good Topic

This makes a strong science fair topic because you can collect real data from real buildings and test a clear variable, building type. You can compare apartments, detached homes, and classrooms, then ask whether tighter buildings or different ventilation setups change infiltration. The project connects to air quality, asthma, wildfire smoke, and exposure science. You can learn sensor calibration, signal processing, and data modeling without needing a full university lab.

Research Questions

• How does building type affect the indoor-to-outdoor PM2.5 infiltration coefficient?
• What is the effect of window opening status on the estimated infiltration coefficient?
• Does HVAC operation change the gap between indoor and outdoor PM2.5 during the same outdoor event?
• To what extent do indoor sources bias infiltration estimates when you do not apply a state-space model?
• Which building features, such as floor level, visible drafts, or ventilation type, best predict higher infiltration?
• How does sensor pairing between PurpleAir and PMS5003 change the stability of infiltration estimates?

Basic Materials

• PurpleAir sensor or comparable outdoor PM sensor with cloud access.
• PMS5003 sensor or similar indoor PM sensor.
• Two data logging devices, or one device with time-synced logging for both sensors.
• Laptop with spreadsheet software.
• Measuring tape.
• Notebook or digital log for indoor events like cooking, candles, or vacuuming.
• Outdoor weather data from NOAA or a nearby station.
• Simple floor plan sketches for each building.
• Access to at least two building types for repeated sampling.

Advanced Materials

• PurpleAir sensor with documented calibration history.
• PMS5003 sensor with microcontroller-based logger.
• Indoor reference monitor, if available through a school or partner lab.
• HVAC runtime data or smart thermostat logs.
• Portable CO2 sensor for ventilation context.
• Particle calibration setup using a controlled aerosol source in a supervised lab.
• Reference gravimetric PM sampler, if available through a partner lab.
• Location-specific weather and smoke transport data from NOAA, EPA, or NASA satellite products.
• Building envelope notes, such as window type, mechanical ventilation, and occupancy schedule.

Software & Tools

• Excel or Google Sheets: Organizes paired sensor time series and helps you inspect gaps, spikes, and event timing.
• Python: Cleans data, aligns timestamps, and runs the Kalman filter model.
• Jupyter Notebook: Keeps code, notes, and plots in one place for analysis.
• ImageJ: Measures any simple visual checks you make on sensor housing or dust buildup, if needed.
• NOAA Air Quality or weather data pages: Provide outdoor context for wind, humidity, and regional particle events.

Experiment Steps

1. Define the building types and rooms you will compare, then decide how you will keep the sampling conditions similar across sites.
2. Choose one outdoor PM2.5 reference and one indoor PM2.5 sensor strategy, then plan how you will synchronize timestamps.
3. Map the indoor activities that create false spikes, then decide how you will label or flag those events in your dataset.
4. Build a calibration plan that converts raw sensor signals into comparable PM2.5 estimates before you model infiltration.
5. Design the state-space model so it separates outdoor-driven changes from indoor-source changes, then test it on a small pilot dataset.
6. Plan the comparison metrics you will report, such as infiltration coefficient, model error, and sensitivity to indoor events.

Common Pitfalls

• Treating indoor spikes from cooking as outdoor infiltration, which inflates the coefficient.
• Comparing sensors that were not time-synced, which shifts peaks and breaks the model.
• Using one building during one weather pattern, which makes the results look stronger than they are.
• Ignoring humidity effects on low-cost PM sensors, which can make particle counts look higher than they really are.
• Skipping a calibration step between PurpleAir and PMS5003, which leaves the two data streams on different scales.

What Makes This Competitive

A competitive version of this project does more than compare averages. You would build a clean calibration pipeline, separate indoor sources from outdoor infiltration, and test whether the pattern holds across multiple building types and weather conditions. Strong entries also report uncertainty, not just a single coefficient. If you can show when the model fails, and why, your project gets much stronger.

Project Variations

• Compare apartments, detached houses, and school classrooms to see whether building form changes PM2.5 infiltration.
• Replace the Kalman filter with a simpler regression model, then test how much error indoor sources add.
• Add wildfire smoke days versus non-smoke days to see whether infiltration behaves differently under heavy outdoor pollution.

Learn More

• EPA Air Sensor Guidebook: Background on low-cost particle sensors and where to find it on the EPA website.
• NOAA Air Resources Laboratory: Outdoor air and weather data useful for linking PM2.5 changes to local conditions.
• NASA Earthdata: Satellite and atmosphere data for smoke transport and aerosol context, searchable through NASA Earthdata.
• PubMed: Search for review articles on indoor PM2.5 infiltration, low-cost sensor calibration, and exposure science.
• USGS Science Database: Find environmental background data and methods articles that may help with outdoor particle context.
• MIT OpenCourseWare Environmental Engineering courses: Free lectures on air pollution and environmental measurement methods.