*2.1. Study Area and Monitoring Site*

Children considered in the experimental campaign attended three naturally ventilated schools located in three different cities in Italy (Salerno, South of Italy, Roma, Central Italy, and Parma, North of Italy); the locations of the three schools within the urban contexts are completely different. In particular, the school in Salerno (S1) is placed in a suburban area as it is 1.6 km outside of the city centre, but quite close to a highway. The school in Rome (S2) is located in the urban area, and, in particular, in the proximity of highly trafficked roads, whereas, the school located in Parma (S3) is in the rural area, about 5 km from the city centre, and quite far from trafficked roads. The experimental campaigns in the three schools were performed from November 2018 to May 2019 for about two months in each school as summarized Table 1.

**Table 1.** School sites, sampling periods and summary of the meteo-climatic conditions (temperature, T, relative humidity, RH) and air quality parameters (NO2, PM10 and PM2.5) measured by the closest fixed sampling stations of the Italian environmental protection agency. The data related to every single period of the campaign are expressed as daily average values and their ranges (min–max).


In order to better describe the three sampling sites in terms of outdoor air quality, in Table 1, the distance of the closest fixed sampling point (FPS) installed by the Italian environmental protection agency to the schools are reported, as well as its definition in terms of type of station (background, industrial, or traffic) and type of site (urban, suburban rural sites). The closest FPSs to school S1 (100 m), S2 (1 km), and S3 (12 km), are defined as suburban/background station, urban/traffic station, and rural/background station, respectively. The parameters measured by the three FSPs during the three different sampling periods (November–December 2018, February–March 2019, and April–May 2019 for school S1, S2 and S3, respectively) clearly highlight the different outdoor air quality of the locations investigated: indeed, the highest NO2, PM10 and PM2.5 values were measured by the FSP close to the school in Rome (S2) (average values of 42, 29, and 16 <sup>μ</sup>g·m−3, respectively) whereas the lowest, as expected, were measured by the FPS close to the school in Parma (S3) (average values of 12, 12, and 7.5 <sup>μ</sup>g·m−<sup>3</sup> for NO2, PM10 and PM2.5, respectively). The authors, once again, point out that

the concentration of the different PM fractions cannot be considered as a good proxy for ultrafine or submicron particles. Indeed, the latter, along with NO2, are good markers of the tailpipe emissions of the vehicular traffic, whereas PM10 is only partially due to tailpipe emissions of vehicles (a significant fraction is due to the traffic-induced particle resuspension) and it is a good marker, amongst others, of biomass combustion for residential heating [75]. Therefore, an overall correlation between outdoor concentrations of PM10 and submicron particles can be found, but, in some conditions (e.g., co-presence of other sources) these two metrics could be poorly correlated. Actually, since the FSPs close to S1 and S2 are strongly affected by traffic sources, a good correlation between PM10 and submicron particles is somehow expected; this is partially confirmed by the fact that NO2 and PM fractions data shown in Table 1 present very good correlations (linear regressions with r2 equal to 0.95 and 0.99 for PM10 and PM2.5, respectively). Finally, regarding the meteo-climatic parameters, temperature and relative humidity values were found to be roughly similar in the three sites during the three measurement periods. This is a not trivial aspect—indeed, generally, the time of the year (e.g., season) can affect the children's exposure and doses both in terms of time-activity patterns and ventilation of the microenvironments since warmer conditions would have increased the time spent outdoor and the manual ventilation in indoor environments (e.g., schools and homes). Thus, the similar outdoor meteo-climatic conditions had a relatively negligible effect on the time of the year on the results.

### *2.2. Study Design*

To evaluate the surface area dose received by children attending the three schools considered in the present study, particle number (PN) and lung-deposited surface area (LDSA) concentrations and average particle sizes (Dp) were measured by means of a personal monitor, which is a handheld diffusion charger particle counter (NanoTracer, Philips). The children were equipped with the mobile monitor fixed to a belt at the hip for 48 h.

During the campaign, 20 children for each school (60 children in total) were monitored. In particular, children aged 6–10 years were monitored (both males and females). Measurements were performed only on school days; weekends were not considered in the study. The authors monitored such high number of children in each school in order to obtain sufficient data that could be representative of the exposure level in each microenvironment where they live/reside. Indeed, the exposure of the children in each microenvironment and during each activity was affected by several parameters, such as the outdoor concentration levels, the volume of the indoor environments, and the presence and the strength of indoor sources (e.g., cooking, smokers, incense, candles etc.). As an example, the children's exposure when they stay in the kitchen during parents' cooking activities is strongly affected by the kitchen volume and the different types of foods and stoves ([76–80]), thus having performed different measurements (on different children) allowed the authors averaging amongst all these influencing parameters. Similarly, the exposure during transport can vary significantly as a function of the transportation modes (i.e., car, walking, bus, etc.; [43,81]), thus, once again, multiple measurements allowed to take into account for all these conditions.

In order to estimate the dose, the children, with the support of their parents, were asked to fill in an activity diary to take note about the place, time, and activity performed. A pre-compiled form of the activity diary was prepared by the authors and given to the children along with the portable instrument; the form was prepared considering 15-min time slots (e.g., 00:00–00:15, 00:15–00:30, etc.) in order to make it easy to fill in the forms with the required information. The diary was then used during the data post-processing in order to evaluate the time spent in each activity (i.e., the time-activity pattern) and to determine the exposure during each activity and in each microenvironment. The daily dose of the children under investigation in terms of particle surface area in the tracheobronchial and alveolar regions of the lungs (δ), was calculated as sum of the dose received during the activities performed in the *j* microenvironments:

$$\delta = \sum\_{j=1}^{n} \left\{ IR\_{\text{activity},j} \cdot \text{LDSA}\_{j} \cdot T\_{j} \right\} \left( \text{mm}^{2} \right) \tag{1}$$

where IRactivity (m3·h<sup>−</sup>1) is the inhalation rate of the child, LDSA is the Lung-Deposited Surface Area concentration (μm2·cm−3), and *Tj* (h) is time spent in each microenvironment. The IRactivity is a function of the age and activity performed by the children; in particular, we have considered the IR data for 6–10-year-old children summarized in Buonanno et al., 2012 [63]. In Equation (1) the term "microenvironment" is used for the sake of simplicity: the activities performed by the children, obtained based on the time-activity patterns, were grouped in six main microenvironments, summarized in Table 2. Particular attention should be paid to the "Cooking & Eating" microenvironment; indeed, children do not perform cooking activities per se, thus, the exposure related to this microenvironment is due to cooking activities performed by the parents. To compare the received dose of the children in different microenvironments, the dose-intensity ratio (i\_δ, mm2·min<sup>−</sup>1), i.e., the ratio between the daily dose fraction and the daily time fraction characteristics of each microenvironment, was also evaluated [53].


**Table 2.** Classification of the activities performed by the citizens in seven main microenvironments.

#### *2.3. Instrumentation and Its Quality Assurance*

As mentioned above particle number (PN) and lung-deposited surface area (LDSA) concentrations and average particle sizes (Dp) were measured by means of a hand-held diffusion charger particle counter (NanoTracer Philips). It measures the particle number concentration and the average particle size in the range 10–300 nm, with a sampling time of 10 seconds. The operating principle of this instrument is based on the diffusion charging technique. In particular, the sampled aerosol is charged in a standard positive unipolar diffusion charger imparting an average known charge on the particles that is approximately proportional to the particle diameter of the aerosol. The number of charges, and thus the number of particles, is then detected by an electrometer [82–84]. Since over 99% of total particle number concentrations in urban environments are due to particles below 300 nm in diameter [85,86], the instrument was considered adequate for the experimental campaign. Actually, the lung-deposited surface area (LDSA) concentration cannot be considered, strictly speaking, a direct measurement, since it is provided by the instrument on the basis of built-in semi-empiric relationships allowing calculating the particle surface area deposited in the alveolar and tracheobronchial through the PN concentration and average particle size (Dp) measured data as described in details in Marra, et al. [87] and Fierz, Houle, Steigmeier and Burtscher [82]. Then, the LDSA concentration was evaluated as sum of the alveolar- and tracheobronchial-deposited contributions. Nonetheless, in order to take into account for calibrated PN concentrations and Dp values, we have used the semi-empiric relationships to calculate the LDSA concentrations on the basis of the calibrated values. In particular, the calibration of the device was performed before and after each experimental campaign. To this end, both a Condensation Particle Counter (CPC 3775, TSI Inc., Shoreview, MN, USA) and a Scanning Mobility Particle Sizer (SMPS 3936, TSI Inc.) were used to compare the devices in terms of number concentration and particle size, respectively. The SMPS consisted of an Electrostatic Classifier (EC 3080, TSI Inc.), a Differential Mobility Analyzer (DMA 3081, TSI Inc.), and a CPC 3775. The SMPS 3936 was used, with an aerosol/sheath flow ratio of 0.3/3.0 L·min<sup>−</sup>1, thus measuring particle number distributions in the range 14–700 nm. The calibration was carried out at the European Accredited Laboratory of Industrial Measurements (LaMI) of the University of Cassino and Southern Lazio (Italy) in a 150 m3-room, with a conventional mechanical ventilation system guaranteeing constant thermo-hygrometric conditions (20 ± 2 ◦C and 50 ± 5% RH). Comparisons were performed for two different aerosols: aged indoor aerosol and freshly emitted aerosol produced by incense burning. Tests were conducted for 2 h performing simultaneous measurements with the Nanotracer, the CPC 3775, and the SMPS 3936. CPC and SMPS sampling times were set at 1 s and 135 s, respectively. SMPS measurements were corrected for multiple charge and diffusion losses. The correction factors obtained by averaging the results of the two aerosols investigated before and after each experimental campaign were applied as correction factors for each campaign. The differences in correction factors measured before and after the campaigns were found lower than 10%.
