TEST RESULTS OF LOW COST SENSORS FOR PARTICULATE MATTER AND GASES FOR THE DETERMINATION OF OUT DOOR AIR QUALITY

Summary:
In air quality there is a trend to use low cost sensors instead of expensive measurement instruments. Uncertainty of the sensors can be reduced and determined with different quality assurance measures, by comparing the results of the sensors to reference instruments. Tests of low cost sensors for PM as well as gaseous compounds were performed in laboratory as well as outdoor in the ambient air. Results will be presented as well as the limitation of the use of sensors will be shown.
Introduction:
In the field of air pollution control, sensors have been increasingly used for emission measurements for a long time, in addition to the usual emission measuring devices. One should think of the millions of lambda sensors installed in vehicles with regulated catalysts. Sensors are also used for other applications. For studies in the ambient air, for a long time sensors were not sensitive enough to detect the usually very low concentrations of airborne contaminants in the outside air. In recent years, however, there has been a huge development in various sensors, which also makes them suitable for use in outdoor air.
Interesting are the low-cost sensors for different reasons: First, they are very inexpensive to purchase. The price is between 0.1 and a few percent of the cost of a commercially available professional measurement instrument, e.g an aerosol spectrometer for measuring PM. In addition, the sensors are small, light, require little power and can often be powered by batteries or accumulators. Due to the low price, larger networks can be built by installing sensors at many points in a city or area. This gives a very good overview of the spatial and temporal distribution of air pollution in an investigation area. The other advantages mentioned, such as the low weight and the battery operation, make it interesting for mobile measurements as performed by the authors. As mobile measuring platforms bicycles, tethered balloons, possibly drones and public transport such as buses, trams and rack railways are planned. Currently in many cities, sensor networks spring up (e.g. http://luftdaten.info), some of them with several hundred sensors to determine the air quality, mainly the PM10 and PM2.5 components. The network operators generally do not put much effort into quality assurance, either because of a lack of awareness of the need for quality assurance or due to the lack of the required sensor matching technology. Another reason is the notion that when using as many sensors as possible, the individual sensors do not necessarily deliver "accurate" results, but the uncertainties rebalance.
Methodology and Results:
The authors have acquired different low-cost sensors from different manufacturers for both particulate matter (PM10, PM2.5,
PM1, particle count) and different gases (NO, NO2, O3 and CO). Comparative measurements for PM with profession aerosol
spectrometer were performed both, in the laboratory as well as in the ambient air. On the one hand, the determination of the
measurement uncertainty of the sensors are subject. On the other hand, a concept was developed and comparison
measurements were done, where low-cost sensors of the sensor networks were compared with professional aerosol
spectrometers to reduce the uncertainty of the sensor results and increase the data quality. First simple correction approaches
have shown that this can improve the deviation of the measurement results of the sensors with each other and in comparison
to the aerosol spectrometers. An important influencing factor on the quality of the results of the particle sensors is the
humidity. This influence must be considered in the evaluation.
All sensors need a calibration with a reliable PM standard. After calibration, sensor no.1 delivered satisfying results for small particles in the range of 0.3 to 2μm (PM2.5 signal). For coarser particle in the range of 2 to 10 μm (PM10 signal) this sensor did not deliver satisfying results. Sensor no.2 did not deliver satisfying results, neither for PM2.5 nor for PM10 signal. Sensor no. 3 delivered satisfying results for all PM channels, PM10, PM2.5 and PM1. Even more this sensor delivers reliable results for 16 channels between 0.38 and 17 μm for the particle number concentration.

For gaseous compounds results of calibrations with calibration gases as well as comparative measurements between different electrochemical sensors and reference instruments (chemiluminescence, NDIR, UV absorption) will be presented.
Conclusions:
Low cost sensors can be used after the application of quality assurance measures prior to the measurements instead of expensive measurement instruments. But not all sensors deliver the same quality of data. Some are even inappropriate for usage in the ambient air.