Nephelometer Sensitivities for the Determination of PM2.5 Mass Concentration in Ambient and Indoor Air

Received: April 20, 2020
Revised: October 2, 2020
Accepted: November 3, 2020

 Copyright The Author(s). This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are cited.

Cite this article:

Karali, D., Loupa, G., Rapsomanikis, S. (2021). Nephelometer Sensitivities for the Determination of PM_2.5 Mass Concentration in Ambient and Indoor Air. Aerosol Air Qual. Res. 21, 200159. https://doi.org/10.4209/aaqr.2020.04.0159

HIGHLIGHTS

• On line PM_2.5 mass concentration monitoring with nephelometry.
• Light scattering coefficient dependence on relative humidity.
• Indoor aerosol mass concentrations and densities.

Keywords: PM2.5 gravimetric mass concentrations, Aerosol light scattering coefficient, Nephelometric measurements, Indoor aerosol density

1 INTRODUCTION

Indoor and outdoor aerosol present a significant risk factor for the human health, as it has been asserted in numerous studies (Cohen et al., 2017; Nagel et al., 2018; Vardoulakis et al., 2019). The main effort nowadays is to reduce the atmospheric concentrations of the PM_2.5 (particulate matter with aerodynamic diameter < 2.5 µm), because it is deleterious to health and is associated with respiratory and cardiopulmonary diseases, as well as lung cancer (Kim et al., 2015; Feng et al., 2016; Melki et al., 2017). Such an effort is mirrored, for example, in the European project “Urban PM_2.5 Atlas - Air Quality in European cities”, which provides information on the levels of air pollution (specifically PM_2.5) in 150 European cities (Thunis et al., 2017). The determination of the atmospheric concentrations of PM_2.5 is regulated by national and international legislation and their monitoring is an imperative necessity.

Monitoring the atmospheric PM_2.5 can be an expensive, time consuming and/or tedious task. Gravimetric methods are the standard reference methods (EU, 1999; EN, 2005). However, proxy methods that need calibration have been used by scientific groups worldwide, because of the ease of use, low purchase price and more importantly their continuous operation (recording time step of minutes) (Holstius et al., 2014; Budde et al., 2018). Such instrumentation may be based on the principle of β radiation attenuation by the collected aerosol, change in the frequency of oscillation in tapered oscillating microbalance, or the intensity of the light that is scattered by the aerosol flowing through an optical cell (Huang, 2007; Allen et al., 2011; Kulkarni et al., 2011; Kim, 2015). Comparison of these techniques and instruments, against each other and against international and European standard norms, is crucial to facilitate the PM_2.5 monitoring (Waggoner and Weiss, 1980; Sioutas et al., 2000; Liu et al., 2002; Allen et al., 2003; Chow et al., 2006; Grimm and Eatough, 2009; Dinoi et al., 2017). In the present study, relationships between the gravimetrically measured PM_2.5 mass concentration by an impactor and the PM_2.5 light scattering coefficient as recorded by a nephelometer, were derived for three different environments: two outdoor locations (an urban and a suburban) and one indoors. The results will provide a simple and cost-effective method to estimate the PM_2.5 mass concentrations on line (with a time step of minutes), albeit significant only for the above named, geographically defined locations. The effect of the relative humidity on the nephelometer readings was also evaluated (Liu et al., 2002; Chow et al., 2006) and it was the trigger to introduce an almost constant RH stream into the nephelometer. Furthermore, in the indoor environment, the PM_2.5 mass concentrations (measured gravimetrically and estimated by the nephelometer data) were compared with the readings of another light scattering instrument (an aerosol light scattering spectrometer). This spectrometer provided also the PM_2.5 volume concentrations that permitted the calculation of the PM_2.5 density. The instrumental set up follows the philosophy of minimum cost and intervention.

2 METHODS

The relationship between the mass concentration of PM_2.5 and the aerosol light scattering coefficient of PM_2.5 (B_sp) were examined during three monitoring campaigns in the city of Xanthi, Greece. The first period of measurements was conducted in the city centre of Xanthi, Greece (September 2013 and November 2014; referred to as “urban”). A Radiance Research M903 nephelometer (hereafter referred to as RRN-903; Radiance Research, Seattle, WA, USA), without drying the aerosol laden stream, was operated in parallel with a high volume PM_2.5 sampler (TE-6001-2.5-I, Tisch Environmental Inc., Village of Cleves, OH, USA). The PM_2.5 mass concentrations and the B_sp were monitored at the adult breathing height (∼1.7 m), adjacent to the pavement of a busy road. Temperature and relative humidity (RH) were also recorded (Vaisala HUMICAP®, Helsinki, Finland). During this campaign the effect of RH on the nephelometer readings was established. During June and July 2014 the nephelometer recorded data every 5 min in the centre of the city (the high volume sampler was not used during this period). From these data, eleven days had RH ≤ 40% and were used in the ANOVA presented in the results, for the case of the urban site.

During 2019, the sampling station was installed in the University campus, in the suburbs of Xanthi. Initially, the sampling was conducted outdoors (referred to as “suburban”) and finally inside a laboratory (referred to as “indoors”). During, these two campaigns the nephelometer operated with a Nafion™ dryer in place (see Fig. S1, in the supplementary material). The whole length of ½” I.D. tubing together with the length of the Nafion™ dryer was 2.5 meters, as determined by the algorithm of von der Weiden et al. (2009). The truncation error for PM_2.5 is small, below 1.06 and for our data treatment was ignored (Anderson and Ogren, 1998; Müller et al., 2009).

Additionally, particle number, volume and mass concentrations were monitored inside the laboratory with a PALAS Promo 2000 light scattering spectrometer (which can classify particles in 40 size bins between 200–2500 nm, PALAS GmbH, Greschbachstraße 3b, 76229 Karlsruhe, Germany). Aerosol mass concentrations provided by the PALAS Promo 2000 are calculated by the software of the instrument with a default PM_2.5 density of 1 g cm^–3. The 24 h average mass concentrations, acquired gravimetrically, along with the PM_2.5 readings by the PALAS Promo, provided the opportunity to calculate the indoor PM_2.5 densities. Also, the correlation of the readings of the two collocated light scattering instruments was examined.

Details for the monitoring methods can be found in the supplementary material.

3 RESULTS AND DISCUSSION

3.1 The B_sp – RH Relationship

The relationship of the B_sp and the atmospheric water content, as expressed with the RH, was examined for the experimental data acquired during the first campaign, i.e., during 2013 and 2014. In Fig.1 the measured B_sp values are plotted against RH. Two models were applied: a simple linear model and an exponential model, presented in Table 1, along with the estimated fitting parameters of each model. The best fit model in our data was the linear model, probably due to the low RH levels (most of the time below 65% during this first campaign) and due to the small number of samples. The B_sp dependence on the RH was not unexpected and it has been extensively reported in the literature with several proposed algorithms to account for this dependence (Vincent, 2007; Zieger et al., 2013; Tryner et al., 2019). However, these relationships cannot be universal solutions, because atmospheric aerosol composition (and water affinity) varies with source, sampling point, season, day and minute. 

Fig. 1. The B_sp values dependency on RH (urban site, ambient RH). 

During the other two campaigns in our study, the aerosol laden air stream was dried by a Nafion™ drier in the inlet of the nephelometer to a RH less than 40% for a flow of 2 L min^–1. The flow rate and the size of the tubing (diameter and length) were calculated using the algorithms and the graphical user interface given by von der Weiden et al. (2009).

3.2 The B_sp Dependence on the Location of the Monitoring Station

The dependence of the B_sp values on the location of the nephelometer was evaluated through an Analysis of Variance (one-way ANOVA), which tested for significant differences between the means of the B_sp values for one categorical predictor (urban, suburban and indoors). For this analysis, the data from the urban environment (uncontrolled RH) that corresponded to days with a 24-h average RH ≤ 40% were used. The results of the ANOVA are statistically significant and corroborate that the nephelometer readings depended on the location of the monitoring station, (F = 6.38488 and p < 0.005194).

3.3 The Relationship between the B_sp and the Mass Concentrations of PM_2.5

In Fig. 2 the average B_sp values were plotted against the respective average, gravimetric mass concentrations of PM_2.5 for each sampling location, i.e., urban, suburban and indoors. The coefficient of determination (R²) for each linear model fitting was above 0.70, pointing out to a satisfactory relationship between the two variables. The slope of the regression line varied between 1.6 m² g^–1 to 4.5 m² g^–1, values that are similar to those reported in other studies (Liu et al., 2002; Chow et al., 2006; Hand and Malm, 2007). 

Fig. 2. The relationship B_sp-PM_2.5 (gravimetrically measured).

Also, the PM_2.5 mass concentrations acquired with the impactors can be used to convert nephelometer readings, in each case of the present study, into mass concentration in order to observe the PM_2.5 variation in a time step of few minutes. These equations are:

PM_2.5 (µg m^–3) =0.55 × B_sp (Mm^–1) + 1.13 (urban, ambient RH)                                                             (1)

PM_2.5 (µg m^–3) =0.42 × B_sp (Mm^–1) + 2.47 (suburban, RH ≤ 40)                                                             (2)

The respective equation for the indoor laboratory environment is presented below, in Fig. 3. 

Fig. 3. The indoor relationship B_sp with PM_2.5 mass concentrations, as measured gravimetrically and as calculated with the PALAS Promo, corrected for particle density.

3.3 Indoor PM_2.5 Densities

In the laboratory, the two light scattering instruments were run simultaneously with the impactor. Note that the laboratory was void of people during the experiment, except the instrument operator, hence there were not PM emission events, which could disturb the readings of the nephelometer (Liu et al., 2002).

This experiment provided the opportunity to calculate PM_2.5 densities. Assuming that the average PM_2.5 mass concentration obtained indoors, gravimetrically, is the same concentration that the PALAS Promo detects, then the measured mass of PM_2.5 divided by the total volume V of PM_2.5 particles recorded by the PALAS Promo, will result in indoor aerosol densities, the last column in Table 2.

The calculated PM_2.5 densities were used to correct the PALAS Promo PM_2.5 mass concentrations (calculated by the instrument with the default density of 1 g cm^–3). The PM_2.5 mass concentrations (second column in the Table 2) were corrected by multiplying each value with the respective density in the last column of the Table 2. As can be seen from Table 2, the calculated indoor particle densities differed (mean = 0.95 (SD = 0.17) g cm^–3) but not significantly from the arbitrary selected density of 1 g cm^–3, due to absence of indoor emissions and a relatively unpolluted outdoor air in this case. The results indicate that arbitrarily selected densities of indoor aerosol for the calculation of their indoor concentrations may be erroneous, especially if an indoor environment has enhanced human activity, i.e., variable indoor sources. 

The corrected PM_2.5 mass concentrations of the PALAS Promo were better related with the gravimetrically derived values. The Relative Standard Error was 3.15% for the uncorrected and 2.46% for the corrected values. In Fig. 3, the corrected PM_2.5 mass concentrations from the PALAS Promo and the PM_2.5 mass concentrations from the gravimetric method are plotted against the respective the B_sp values to highlight the good correlation of the three methods.

4 CONCLUSIONS

The mass concentration of PM_2.5 in a small time step is needed to trace aerosol variations. In the present study the PM_2.5 mass concentrations were determined gravimetrically and related to the aerosol light scattering coefficient obtained by a RRN-903, during three campaigns at three different environments, i.e., an urban, a suburban and indoors. The data from the first campaign, in the city center, confirmed, as expected, that the aerosol light scattering coefficient was sensitive to the ambient RH. Two models were tested in order to express this relationship and a simple linear model has been proved to provide the best fit in this case. Thus, in the following two campaigns (in the suburbs of the city and in a laboratory), a drying device before the aerosol entrance to the nephelometer was used. For these campaigns, relationships between B_sp and the PM_2.5 mass concentrations (gravimetric) were also derived. The presented algorithms can be applied in order to convert light scattering data to bulk PM_2.5 mass concentration, using the nephelometer RRN-903 as a PM_2.5 mass detector. However, as confirmed also by an Analysis of Variance, the B_sp is sensitive to the location of the monitoring station and there is a need for in situ calibration in each case. Furthermore, the densities of indoor particles with the same experimental set up with the cooperation of an aerosol spectrometer (PALAS Promo), which readings were well related with the nephelometer RRN-903, indicate their variability even in a supposedly controlled environment.

ACKNOWLEDGEMENTS

The present work was funded by Democritus University of Thrace (Greece) funds.