The challenges of CO2 sensing

The health benefits of fresh air are irrefutable, and the lack of it affects our health and the ability to focus. Unfortunately, in this busy world, few of us are able to spend as much time outdoors as we might want to. On average, we spend up to 90 % of our lives indoors, and most of that time we reside in our homes. Even though we leave our houses regularly, a lot of the time spent outside the home is in another indoor location, such as an office, restaurant or shop. This clearly shows the importance of high quality indoor air and proper monitoring of CO₂ levels, especially in crowded venues.

Although high levels of CO₂ and bad air quality are not synonymous, raised concentrations of CO₂ can be a good indication that the room requires additional ventilation. In addition, an increase in CO₂ often comes hand-in-hand with a rise in volatile organic chemical (VOC) concentrations, as both are emitted by humans. It is commonly known that bad air – especially high VOCs – can be detrimental to health and increase the risk of transmission of airborne viruses, such as SARS-CoV-2. Additionally, a lack of fresh air also has a great impact on productivity and the ability to focus – a claim that is backed up by numerous studies.

Extremes are never good, and this is also true for ventilation. HVAC systems that are constantly working at maximum capacity will lead to high energy consumption and, consequently, exorbitant electricity bills, especially during very hot or cold periods. It is therefore not a surprise that demand-controlled ventilation is currently seen as the gold standard for HVAC systems, and the CO₂ concentration is often used as a control parameter, as it correlates closely with air quality. This application relies on sensors providing accurate information on CO₂ levels, activating the system when a specified limit has been reached. Although comfort norms vary around the world, there is a consensus that the CO₂ levels should always be kept below 1,000 ppm, and not exceed 1,500 ppm for long periods. A good compromise is to measure and adjust the CO₂ levels every 30 seconds, which keeps the air fresh and the energy bills low.

A common CO₂ sensor design consists of a light source and two detectors (Figure 1). As light passes through the measurement chamber, filled with ambient indoor air, it is absorbed by the molecules present. One detector has a filter with a window at around 4.3 µm – which corresponds to a peak in the CO₂ absorption spectra – meaning it only registers extinction of light due to the presence of CO₂ molecules. In contrast, the reference detector measures the unfiltered light intensity, making it possible to determine the CO₂ level by comparing the two measurements. The dual sensor design also helps to counter the drop in light intensity originating from light source degradation or small dust particles. To further enhance the robustness of the sensors, they should be fitted with a dust cover that stops particles from interfering with the detectors.

Although the dual channel approach is considered to be accurate, it alone cannot guarantee stable long-term measurements, as the baseline can start to drift over time due to aging of sensor components. This can be fixed through automatic baseline correction (ABC), which constantly tracks the sensors lowest reading, and corrects for any drift that is detected. This approach works well for buildings that are unoccupied for periods of time such as offices that are closed during the weekend. However, this drift is not as easy to identify and address in venues that are occupied 24/7, for examples hospital emergency rooms, logistic centres or factories. It is therefore crucial to use robust sensors that provide accurate long-term measurements without the need for constant calibration, allowing them to be used in all applications, regardless of occupancy patterns.

A room sensor needs to be able to accurately measure CO₂ levels under any conditions, which means it needs to have a good resistance to both gradual and acute changes in pressure, temperature and humidity. The pressure differences at different altitudes also need to be accounted for, as even an elevation of 400 m above sea level results in a 70 ppm offset in the measured CO₂ concentration. Considering that some regulatory bodies – for example several state governments in the USA – only allow a tolerance of ±75 ppm, this leaves almost no margin for error. Any high performance CO₂ sensor should therefore include absolute pressure compensation (Figure 2)

Extended testing should be performed to ensure that the sensor can operate under various conditions in order to guarantee long-term stability and function. Sensors should therefore be tested for an extended period – spanning a number of weeks – covering all possible weather conditions and focusing on those known to put a lot of stress on the device. For example, non-condensing wet heat performance can be tested at 95 % relative humidity and 35 °C to ensure the sensor exhibits corrosion resistance and can maintain its performance. On the other hand, dry heat measurements should be performed at higher temperatures – 60-70 °C – to confirm that no drift occurs due to the difference in expansion coefficients of the materials. As internal temperature gradients can also play a role in the overall device performance, the sensor elements must be built in a way that minimizes self-heating.