*3.2. Exposure to Submicron Particles*

In Table 3 and Figure 1, the submicron particle concentrations, in terms of particle number and lung-deposited surface area, to which the children attending the three different schools (S1, S2, and S3) were exposed to in the different microenvironments (sleeping, indoor day, outdoor day, school, transport, cooking & eating) are shown. In the box plots of Figure 1, exposure data not statistically different amongst the six different microenvironments for each group of children separately (S1, S2, S3) and amongst the three groups of children for each microenvironment separately are also indicated (*p* > 0.01) as resulting from the statistical analysis explained in Section 2.4 (Kruskal–Wallis test). Due to the huge amount of data available for each microenvironment, most of the exposure received in the six microenvironments by the same group of children as well as those received in the same microenvironment by the three groups of children resulted in statistically different results.

**Figure 1.** Statistics of (**a**) particle number and (**b**) lung-deposited surface area concentrations experienced by three groups of children (attending school S1, S2, and S3) in each microenvironment. Data was not statistically different within each group of children (S1, S2, S3) and amongst the same microenvironments of different groups (\*,+) are also indicated (*p* > 0.01).

The children's exposure to submicron particles in the "school" microenvironment presents a significant deviation amongst the three schools. Indeed, children attending school S1, S2 and S3 were exposed to median PN and LDSA concentrations of 1.57 <sup>×</sup> 104 part. cm−3/66 <sup>μ</sup>m2·cm−3, 2.13 <sup>×</sup> 10<sup>4</sup> part. cm−3/89 <sup>μ</sup>m2·cm−3, and 3.39 <sup>×</sup> 103 part. cm <sup>3</sup>/14 <sup>μ</sup>m2·cm 3, respectively. In particular, the concentration levels in the school S3 were much lower than S1 and S2 ones. This is due to the different outdoor concentrations, indeed, if no indoor submicron particle sources are in operation in the schools (as mentioned in the methodology section), the indoor

concentrations are just affected by the outdoor-to-indoor penetration factors [65,90,91]. Thus, the low concentrations measured in school S3 are just related to the low outdoor concentrations typical of the rural site under investigation and discussed in the methodological section (Table 1). Indeed, the median particle number and lung-deposited surface area concentrations in the "outdoor day" microenvironment were equal to 1.91 <sup>×</sup> <sup>10</sup><sup>4</sup> part. cm−3/<sup>79</sup> <sup>μ</sup>m2·cm<sup>−</sup>3, 2.58 <sup>×</sup> <sup>10</sup><sup>4</sup> part. cm−3/<sup>106</sup> <sup>μ</sup>m2·cm<sup>−</sup>3, and 4.22 <sup>×</sup> 10<sup>3</sup> part. cm−3/17 <sup>μ</sup>m2·cm−3, for children attending school S1, S2 and S3, respectively. The resulting "school"/"outdoor day" concentration ratios (considering the median concentrations) were equal to 0.80–0.83 and 0.82–0.86 in terms of PN and LDSA concentrations, respectively, then consistent with the typical penetration factors reported in the scientific literature for naturally ventilated schools [65,90,91]. The location of the children's schools and homes is then the most influencing parameters in their exposure to submicron particles in "outdoor day" and "school" microenvironments, in fact the highest correlations between average outdoor NO2 concentrations measured at the FSPs (Table 1) and PN concentrations measured during the experimental campaigns were determined for these two microenvironments (linear regressions with r2 >0.99). The correlation between outdoor and indoor concentrations gets weaker when it comes to non-school environments, indeed, here the possible presence of indoor sources (cooking, incense, candles, heating systems) can lead to high indoor concentrations. In this context, as expected, the most critical microenvironment is "cooking & eating" which presents median values of PN and LDSA concentrations of 4.20 <sup>×</sup> <sup>10</sup><sup>4</sup> part. cm−3/<sup>112</sup> <sup>μ</sup>m2·cm<sup>−</sup>3, 5.11 <sup>×</sup> 104 part. cm−3/136 <sup>μ</sup>m2·cm<sup>−</sup>3, and 4.91 <sup>×</sup> 104 part. cm−3/130 <sup>μ</sup>m2·cm<sup>−</sup>3, for children attending school S1, S2 and S3, respectively. The correlation with the average outdoor NO2 concentrations measured by the FSPs barely doesn't exist, indeed the concentrations are much larger than the outdoor ones, and also children attending school S3 are exposed to very high submicron concentrations in "cooking & eating" microenvironment and roughly comparable to the S1 and S2 ones despite the much lower outdoor concentrations.

In regard to the other indoor environments labelled as "indoor day" microenvironment, the children's exposure resulted in lower statistical rates than the "cooking & eating" ones for all the three children groups. Nonetheless, the exposure in the "indoor day" microenvironment, when compared to the "outdoor day" one, varied amongst the different children groups. Indeed, the exposure in the "indoor day" microenvironment resulted statistically similar results, slightly lower, and much larger than the "outdoor day" environment for S1 (1.85 <sup>×</sup> 10<sup>4</sup> part. cm−3/84 <sup>μ</sup>m2·cm−3), S2 (1.79 <sup>×</sup> <sup>10</sup><sup>4</sup> part. cm−3/<sup>81</sup> <sup>μ</sup>m2·cm<sup>−</sup>3), and S3 group of children (1.32 <sup>×</sup> 104 part. cm−3/<sup>60</sup> <sup>μ</sup>m2·cm<sup>−</sup>3), respectively. The huge "indoor day"-"outdoor day" difference in the exposure detected for S3 group of children is related to the very low outdoor concentration level; thus, even a minor indoor source can easily increase the indoor concentration to values higher than the outdoor ones. Regarding the exposure in the "sleeping" microenvironment, the concentrations resulted in 0.5–0.6-fold of the "indoor day" microenvironment for all the three groups of children. Finally, during the "transport" microenvironment, higher concentrations were measured for children attending school S1 (2.38 <sup>×</sup> 104 part. cm−3/106 <sup>μ</sup>m2·cm<sup>−</sup>3) and S2 (1.93 <sup>×</sup> 10<sup>4</sup> part. cm−3/86 <sup>μ</sup>m2·cm<sup>−</sup>3), which are close to trafficked roads. On the contrary, children attending school S3 were exposed to quite low concentrations (6.58 <sup>×</sup> <sup>10</sup><sup>3</sup> part. cm−3/<sup>29</sup> <sup>μ</sup>m2·cm<sup>−</sup>3), likely due to the location of the schools (rural area).

In summary, the daily exposure of the children is not only affected by the location of schools and homes, i.e., the proximity to outdoor sources, but also by the presence of indoor sources (mainly cooking); therefore, using outdoor concentration values as proxies of the daily exposure of the children could lead to serious under- or overestimation of the exposure. This is clearly highlighted by the daily median exposure data reported in Table 3; the concentrations, in terms of PN and LDSA, were equal to 1.44 <sup>×</sup> 10<sup>4</sup> part. cm−3/71 <sup>μ</sup>m2·cm<sup>−</sup>3, 1.55 <sup>×</sup> 104 part. cm−3/77 <sup>μ</sup>m2·cm<sup>−</sup>3, and 0.62 <sup>×</sup> 10<sup>4</sup> part. cm−3/34 <sup>μ</sup>m2·cm−3, for children attending school S1, S2 and S3, respectively. Indeed, such values were 0.75-, 0.60-, and 1.48-fold the outdoor PN concentration values and 0.90-, 0.73-, and 2.00-fold the outdoor LDSA concentration values for S1, S2 and S3 children groups, respectively.
