1. Introduction
Air quality below set standard limits have been linked to a number of conditions including asthma, Chronic Obstructive Pulmonary Disease (COPD), pneumonia, and breast cancer [1-5]. Particulate matter (PM), a class of air pollutants defined as airborne particles, dust, droplets, biological materials and all other suspended solid and liquid substances have been shown over the years to be harmful to human health [6-11], causing specific conditions such as influenza virus infection, lung cancer, inflammation, DNA damage and cell death, and affect visibility [12-17]. PM is divided into three major classes based on their aerodynamic diameter. PM10, which refers to all particulate matter smaller than 10μm diameter, is called coarse particles. These particles are considered less dangerous because they can be sneezed or coughed out when inhaled by humans [18,19]. PM2.5 refers to all particulate matter with size smaller than 2.5μm diameter and is called fine particles. These are a lot more dangerous than coarse particles because they can penetrate deep into the human body including the lungs and even the blood when inhaled because of their small size. PM1 with sizes smaller 1μm is referred to as ultra-fine particles, which are also very dangerous to human health.
PM comprises mainly of sulphates and carbon (organic and inorganic) from natural and industrial processes including vehicular emissions, but studies have also indicated that other large components could be nitrates, inorganic elements, ions, radicals and biological materials like pollen. PM has been observed as products of nucleation and growth processes of aerosols [20,21]. Bacon et al., [22] investigated the vertical and horizontal distribution of lead polonium in sea water and PM at various distances from the sea floor. Thorpe and Harrison [23] reviewed the sources of PM from vehicular sources excluding vehicular exhausts. Such sources include brake materials and tyre wear. In Kai et al., [24], ground-based and satellite-based PM monitoring techniques were compared. A satellite PM monitoring system is however, very expensive to set up. Clark et al., [25] analysed PM data from four ground-based sites for organophosphate esters. Hao and Liu [26] found a relationship between the PM2.5 pollution in China and the Gross Domestic Product(GDP) per capita.
PM emission sources have also been classified as primary (direct emission) or secondary (such as gas-solid condensation) based on the generation process [20]. Ratios of ionic concentrations could also help to identify sources e.g. salts, and their precursor gases [27]. It has been established that air pollution is heterogeneous in space so localized monitoring techniques are generally unable to capture this information though some research work is ongoing to use satellite Aerosol Optical Depth (AOD) data [24,28-32]. Small sensors will be useful to unravel this spatial variation [26]. Personal exposure studies carried out using portable PM monitors such as those based on light scattering [33] and condensation particle counting techniques [34] are able to reveal both space and time variations, however they are still relatively expensive to purchase.
2. Past
Gravimetric methods, still very much in use especially in Air Quality Monitoring Centres represents the past of PM detection and monitoring in terms of miniaturized instrumentation development. This method is also still used for validation of measurements from other techniques as it is still the best mass concentration detection method given its high sensitivity to even the smallest particle mass. Different variations of this has been implemented mostly depending on the approach. Apart from direct mass measurements, electromagnetic wave absorption, extinction, scattering and resonance frequency shifts are other common methods. Examples of these include the Beta Attenuation monitor which measures how much Beta radiation is blocked off by the particles, the Dichotomous Aerosol Samplers which may have one or more impaction stages for particle segregation used for collecting particles on filter and other Total Suspended Particles collection techniques. The fundamental issue with these systems is their sheer size and cost, making them largely unsuitable for personal exposure monitoring, except for calibrating the smaller sensors. The implication of their huge cost is that the data they collect lack spatial resolution. In the United Kingdom, for example, there are only about 150 air quality monitoring stations, thus, roughly 150 of these equipment. The next discussion especially going forward is about whether a mass concentration accurately represents the ambient dangers posed by measured particles, especially when collected by filters for offline weighing. This is because smaller particles weigh less than bigger ones and thus their contribution may be underestimated. Whereas these particles are the most dangerous. A way to reduce this problem is to only collect particles of a size range together. However, this is only of little benefit as well since there is a large range of particle sizes in the PM10 range, for example. Thus, gravimetric detectors are useful if the measurement requirements do not include high spatial resolution and mobility.
3. Present
PM detection is mostly currently done with light scattering-based devices. Depending on the regime, Raleigh or Mie scattering is used for analysis. These devices are more portable, and less expensive than the traditional detectors. These devices may be pole-mounted, desktop or handheld. They usually need calibration with reference instruments, are susceptible to deviation and drift and rely on assumptions on particle type and refractive index. It is possible to segregate particles via an inlet control implemented before the particle laden airflow is allowed into the sensing chamber. Like the gravimetric detectors, this approach may be used in addition to current instruments. Another approach that is used in a currently commercially available miniaturised air quality monitoring system for indoor use is the prediction of other pollutants from one measured pollutant. In this case, the device detects the level of indoor CO2 and predicts particle levels based on this. It is not clear how accurate this device is in different environments. When collocated gravimetric equipment is used, filters are used to collect particles in the sensing chamber for verification. Filters are also used for size discrimination. Polytetrafluoroethylene (PTFE) filters, High efficiency particle arrestance (HEPA) filters, quartz filters, or tefloncoated glass filters [35,36] are some examples of well-known filters. Impactors [37], virtual impactors [38], and cyclones [39,40] are other size discrimination methods. Studies of constituents identification use scanning electron microscopy (SEM) usually in combination with energy dispersive x-ray spectroscopy (EDX) [41], and needing a coating for particles before analysis, thermal-optical analysis for both elemental and organic carbon constituents. Other techniques that have been used for constituents identification includes: ion chromatography [42,43], automated colorimetry, Proton Induced X-ray Emission (PIXE) and Proton Elastic Scattering Analysis (PESA) [44,45]. Light-based PM sensors are available commercially. It is becoming increasing research approach to incorporate these or other PM sensor as part of their instrumentation to add additional functionality during field campaigns and epidemiological studies. The peculiar branch of source apportionment has also been very active research wise. Several techniques broadly classified as dispersion or emission models have been developed including the HYbrid Single- Particle Lagrangian Integrated Trajectory model (HYSPLIT) [46], Chemical Mass Balance [47], Positive Matrix Factorization [48], Enrichment Factor Analysis [49] and Principal Component Analysis (PCA) [50]. Dispersion-based techniques such as the HYSPLIT model studies PM transport from the source with knowledge of the source and characteristics with spatial distance from it. In contrast, receptor based models try to locate sources based on information available from the ambient information, with or without knowledge of the sources using uncertainties or error minimization function.
Although, current research on PM detection focuses on the fine particles for obvious reasons, this is challenging to address with light scattering technique due to limitations related to refractive index. Further, the smaller the particle, the lesser light will be scattered, hence the smaller the signal measured. There is therefore a pressure on electronic readouts to overcome the current detection limits. Table 1 shows some of the most common PM sensors commercially available and their specifications.
4. Future
Miniaturised sensors for air quality monitoring is becoming a vibrant area of research driven especially by need for pervasiveness, affordability, and easy integration with Internet of Everything(IoE) devices for user health monitoring. In general, for PM2.5, most popular approaches have been through acoustics, electrical sensing, and microfluidics, even though the detection electronics still lags in miniaturisation. Inclusion of prediction methods such as using fuzzy, recurrent, and feed-forward neural networks [51,52], can prove to be very useful in addition to actual data obtained from sensors.
These PM sensors will be mostly based on the Impedance, and Acoustic techniques. That is, provided the readout electronics of the Acoustic methods become miniaturized. Solidly mounted resonators [53], Surface Acoustic Wave PM sensor [54], 3D printed Quartz crystal microbalance [55] are three sensors with miniaturised active material but mostly desktop readout. Impedance based techniques such as the capacitive detection [56] have shown possibility for miniaturization for readout electronics and are the most likely to achieve this.
5. Conclusion
This report has described air pollution, specifically particulate matter pollution and its effects on human health, current research in this area to address issues such as source apportionment, PM constituent identification, miniaturisation of PM sensors, current trends and outlook for the future. Data obtained from long term personal exposure monitoring will enhance the robustness of epidemiological studies. Further, electronic circuitry miniaturization should be a focus going forward. Finally, impedance-based methods show the most promise for readout miniaturization and can lead to a potentially non-intrusive miniaturized personal air quality monitor.
Competing Interests
The authors declare that they have no competing interests.