Understanding the Performance of Low-Cost PM2.5 Sensors for Cooking Exposure

By Malcolm Bricknell (Loughborough University) and Matthew Leach (Gamos Ltd.).

In a newly published report, Professor Prashant Kumar and a team from the University of Surrey’s Global Centre for Clean Air Research (GCARE) have assessed the comparative accuracy of new low-cost sensors for measuring particulate matter (PM) concentrations in kitchens.

Reference-grade PM equipment is expensive to deploy and often impractical for everyday use. Low-cost sensors present a more affordable alternative, enabling broader measurement of indoor air quality across a larger sample of households and over longer periods. In addition, these sensors provide near-real-time data, making it possible to directly observe the impact of cooking activities on indoor air quality. In contrast, filter-based reference-grade equipment measures the total mass of different PM size fractions over a set period (e.g., a day). During this time, household activities can vary, making actual exposure to cooking related pollution less certain. The features of low-cost sensors thus open up new possibilities, particularly for longitudinal health studies, by offering detailed exposure profiles during activities like cooking.

The study was conducted in the state-of-the-art ‘ENVILUTION^TM’ chamber at GCARE under controlled conditions for temperature, humidity, and PM levels, designed to replicate typical real-world environments across a series of predefined scenarios. The aim was to facilitate a controlled comparison between the research-grade reference monitor and each low-cost sensor, evaluating their accuracy and responsiveness under various conditions.

The team tested six low-cost sensors, including two based on optical sensing techniques and four utilising laser scattering sensors. Each scenario was configured with specific ranges for temperature, relative humidity (RH), and PM_2.5 concentrations, with sensors running for extended periods to capture a substantial dataset in each condition.

All sensors performed well at RH levels below 70%, with optimal performance in the moderate RH range (40-60%). Temperature had minimal impact, as shown by consistent performance across different temperature conditions. However, for all sensor types, high RH levels above 70% were found to reduce measurement accuracy due to hygroscopic growth, which affects particle size and scattering properties. This phenomenon is well documented, and correction methodologies exist to adjust PM concentration measurements, though these adjustments introduce some uncertainties.

As some sensors are more sensitive to environmental factors than others, selecting the appropriate sensor for specific conditions – while also considering cost and data handling methods (manual or web-based) – is crucial for obtaining accurate and meaningful air quality data. While the data quality from low-cost devices will always be lower than that of high-end equipment, the study suggests that with strict quality control and assurance, these sensors can still produce high-resolution spatial and temporal data of significant quality.

Machine learning methods are increasingly popular as a part of the quality control for aligning absolute particle count values from sensors with reference standards. Unlike linear correlations, these methods more effectively capture the temporal variations in sensor data. When larger datasets are available, employing such approaches can be highly beneficial.

Although the study was lab-based and explored a range of real-world conditions, field testing in actual homes is necessary to account for variability in conditions over time, variability in the types of cooking emissions, and PM from other indoor and external sources. Further testing in real cooking contexts will be important to understand the opportunities and limits to these approaches. However, the results of this study demonstrate that the new generation of low-cost sensors can be  viable tools for accurately measuring PM concentration changes in projects focused on transitioning to cleaner cooking technologies.

…………………………….

Featured image: The ‘ENVILUTION^TM’ chamber at the University of Surrey’s Global Centre for Clean Air Research (GCARE) (photo credit: GCARE, University of Surrey).