DSpace at VNU: Indoor-outdoor behavior and sources of size-resolved airborne particles in French classrooms

Building and Environment 81 (2014) 183e191

Contents lists available at ScienceDirect

Building and Environment
journal homepage: www.elsevier.com/locate/buildenv

Indooreoutdoor behavior and sources of size-resolved airborne
particles in French classrooms
Dinh Trinh Tran a, *, Laurent Y. Alleman b, c, Patrice Coddeville b, c, Jean-Claude Galloo b, c
a

Viet Nam National University, Ha Noi, Faculty of Chemistry, 19 Le Thanh Tong Street, Ha Noi, Viet Nam
Univ. Lille Nord de France, F-59000 Lille, France
c
Mines Douai, CE, F-59508 Douai, France
b

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 31 March 2014
Received in revised form
11 June 2014
Accepted 27 June 2014
Available online 5 July 2014

Indoor and outdoor airborne particles were monitored with a 5-s time resolution in three elementary
schools presenting different site typologies (rural, urban, and industrial) in the North of France. We

studied the influence of the children's activities, outdoor sources, temperature and relative humidity on
particle mass concentrations and particle massesize distribution, and estimated cancer risk regarding
particle composition.
The indoor weekly mean PM10 mass concentrations during teaching hours varied from 70 to
99 mg mÀ3, exceeding the French daily recommended value of 50 mg mÀ3, implying a potential impact on
the respiratory system. However, fine particles (<2 mm) were always below French daily recommended
value of 25 mg mÀ3 applied to PM2.5.
The results showed that children's activities impacted the suspended coarse fraction (2e10 mm) more
strongly than the fine one (<2 mm). The mass distribution of indoor PM10 was extremely variable in
association with occupant's activities in classrooms whereas the outdoor one seemed to be only lightly
variable. During lessons, average concentrations of indoor PM1, PM1e2, PM2e5, and PM5e10 increased
respectively by factors of 2.9, 3.1, 8.7 and 33.8 compared to unoccupied periods.
Indoor sources from continuous emission and occupant's activities may lead to lower density of indoor
PM10 compared to outdoor ones.
The estimation of some potential carcinogen elements such as As, Cd, Cr, and Ni in indoor PM2 showed
low concentrations in the range of 0.11e1.71 ng mÀ3. Consequently, the cancer risk of these elements was
estimated to be not significant for long-term exposure to both children and teachers.
© 2014 Elsevier Ltd. All rights reserved.

Keywords:
Indoor/outdoor
Particles
Real-time variation
Particle massesize distribution
Cancer risk assessment

1. Introduction
There has been an increasing interest in researches on indoor air
quality of school environment as children spend considerable time
of their school day indoor where, in certain cases, the air is more

polluted than outdoor [1e3]. Moreover children are more sensitive
to atmospheric pollutants than adults, due to a non-fully developed
respiratory system and high rates of acute respiratory infections [4].
Many epidemiological surveys have demonstrated that exposure to particulate matter (PM) causes adverse respiratory and
cardiovascular health threats [5e8]. Regarding the effect of PM to
children's health, studies showed an association between air
pollution and infant mortality (primarily due to respiratory deaths

* Corresponding author.
E-mail addresses: , (D.T. Tran).
/>0360-1323/© 2014 Elsevier Ltd. All rights reserved.

in the post-neonatal period), lung function development, and upper and lower respiratory symptoms [4].
Moreover, the health effects of particles depend strongly on
their size, number, morphology, specific surface area, and chemical
composition, e.g. their heavy metal contents [9,10].
In spite of the fact that researches on indoor air pollution have
lagged behind outdoor air, ANSES in France has suggested guideline
values of daily PM2.5 and PM10 concentrations for indoor air, similar
to the WHO guideline for outdoor ones, i.e. 25 mg mÀ3 and
50 mg mÀ3 respectively [11]. Moreover, it is generally accepted that
prolonged exposure to ultrafine particles can cause adverse health
effects [12].
A number of studies on school Indoor Air Quality (IAQ) have
focused on respirable suspended particles and their influencing
factors (using average mass concentration over a period of time)
[13,14] but only a few have worked both on particulate (size) distribution and real-time variation [15e19].


184

D.T. Tran et al. / Building and Environment 81 (2014) 183e191

Observations and conclusions drawn on indoor particle concentrations vary between studies due to the variability of conditions governing the indoor air pollutants such as number of
occupants and/or types of specific activities within the classroom.
In addition, sources from educational supplies, school furniture's,
painting and building materials, air renewal rate and outdoor
pollutant sources may also affect the indoor PM content and its
spatial and temporal evolution [3].
The characterization of indoor and/or outdoor sources of particles collected inside buildings has been already considered in the
literature. Most of these works used indoor/outdoor (I/O) ratios and
some employed correlations as basic factors to point out indoor and
outdoor pollution origins [19e22].
Although these approaches reveal details about indoor and
outdoor source relationships, they present some limitations as I/O
ratios or correlations give only general information about the global
pollutant origins, omitting to discriminate each specific source
[1,23].
Recently, some works aimed at identifying the particle origins
using factor analysis [1,3,24]. This approach brings richer and more
detailed information about indoor and outdoor pollutant sources.
The main purposes of the present paper are to (1) determine the
influence of children's activities on mass concentration of different
particle size fractions, (2) identify the origin of particles collected in
classrooms, and (3) estimate the concentration of some potential
carcinogenic elements and their related cancer risk in a school
environment.
2. Materials and methods
During the campaigns performed in three schools, the indoor
and outdoor mass concentrations were monitored in parallel according to their particle sizes (PM1, PM1e2 PM2e5 and PM5e10) along

with meteorological conditions. The information about the site
typology and the children's activities in each school during the
whole campaigns has been also recorded by questionnaires.
2.1. Sampling sites and school characteristics
The three elementary schools studied are located in the NordPas-de-Calais region (NPdC, France) near the North Sea coast and
the Belgian border.
Sampling sites have been selected to represent different environments: industrial and coastal (School 1 e S1), urban/traffic
(School 2 e S2) and rural (School 3 e S3). Schools have been also
chosen upon edifice characteristics such as building age, ventilation
system, internal covering including flooring, wall and ceiling, and
windows structure, as summarized in Tran et al. [3].
The volume of the selected classroom in schools 1, 2, and 3 are
respectively 201 m3, 234 m3 and 181 m3. The three schools are
naturally ventilated (by window opening and small cracks as most
schools in France). The number of occupants in each classroom was
rather comparable (from 24 to 27).
2.2. Sampling protocol and statistical analysis
The sampling campaigns lasting continuously for 2 successive
weeks for each school were divided into 2 periods: in presence and
in absence of pupils in the classrooms. The occupied period consisted of teaching hours during school days including Monday,
Tuesday, Thursday and Friday. Detailed sampling protocol was reported in Tran et al. [3]. Quickly, the sampling during teaching
hours lasted about 8-h per day (8h45e16h45) while 12-h per day
(from 19 h to 7 h of the next day) were attributed to unoccupied
period including the week-end when the sampling time was

accounted for 24-h per day. The sampling campaigns were conducted successively in each school from May 13th 2009 to July 1st
2009 under relatively cool and stable meteorological conditions
during each campaign (Table 1). Outdoor average temperatures
were 15 ± 2  C, 16 ± 3  C and 21 ± 3  C in S1, S2 and S3 respectively
during occupied period and 14 ± 3 c, 13 ± 3 c and 17 ± 4 c during

unoccupied period.
Two sets of identical equipments have been deployed indoors
and outdoors to measure particulate concentrations, CO, CO2 levels,
and comfort parameters (temperature, relative humidity). Realtime measurements of indoor and outdoor number concentrations of different particulate sizes (15 nominal size bins capturing
particles from 0.3 mm to 20 mm) have been monitored using two
pre-calibrated optical particle counter GRIMMS, model 1.108 at 5-s
intervals. These spectrometers worked by constantly drawing the
air sample via a volume controlled pump (1.2 l minÀ1) through a flat
laser light beam. The particle mass concentration was then obtained by conversion from the number concentration measurements hypothesizing that the particles are spherical with a density
equal to 1 g cmÀ3. Each PM counter was equipped with a 47 mm
teflon filter (Mitex, Millipore) to collect the particles with an optical
diameter smaller than 20 mm we corrected the reading mass concentrations obtained by GRIMMs (theoretical values) by weighing
these back-up filters after each school campaign. The measurement
of particles mass collected on the filters was conducted in a
conditioned room according to EN 12341 [3]. The reading/weighted
value ratio was defined as the correction factor (C-factor). This
factor was then used to calculate the adjusted PM10 concentrations.
it must be noted that the absolute concentrations of PM10 presented in this work are all corrected values.
Indoor and outdoor comfort parameters as well as CO, CO2
concentrations were continuously measured using two identical
TSI Q-Trak (model 7565 TSI Inc) at 5-s intervals to match the particle mass concentration database. Before each campaign, the TSI QTrak were calibrated with certified calibration gas (for CO and CO2).
MS Excel and the SPSS 16 statistical package were used for numerical evaluation. Indoor and outdoor concentrations were
compared using the ManneWhitney U test. Spearman correlations
were conducted to evaluate the relationship between indoor and
outdoor mass concentrations. A value of p < 0.05 was considered
significant in all the statistical calculations.
3. Results and discussion
3.1. Indoor and outdoor particle mass concentrations, CO, CO2 levels
and comfort parameters
Descriptive statistics of different particulate fraction concentrations (mg mÀ3) as well as CO, CO2 levels, and various comfort

parameters during occupied and unoccupied periods are summarized in Table 1.
The indoor and outdoor CO concentrations were very low (close
to the detection limit) except for a few spike identified as cigarette
smoking after the class in yards and corridors. This gas was thus
excluded from the data base. The mean indoor CO2 concentrations
were remarkably higher during teaching hours than in absence of
children, and higher at S1 (712e2900; mean: 1475 ppm) and S2
(406e3488; mean: 1484 ppm) than at S3 (372e2780; mean:
690 ppm). The lowest concentration of CO2 during lessons at S3 is
clearly associated with the frequent openings of windows and
doors in this school (confirmed by the questionnaire), resulting in a
higher air exchange with outdoor air. The mean values for unoccupied periods were respectively 504, 573, and 404 ppm (Table 1).
These CO2 levels observed during teaching hours at S1 and S2
regularly exceeded the recommended values [25] and could affect


D.T. Tran et al. / Building and Environment 81 (2014) 183e191

185

Table 1
Descriptive statistics of size distributed PM concentrations indoor and outdoor, and comfort parameters of the three schools during occupied (8-h of sampling) and unoccupied
periods (12-h, values in parentheses) over 2-week campaigns.
Particulate fractions (mm)

Indoor
S1
Mean
Std
Min

Max
S2
Mean
Std
Min
Max
S3
Mean
Std
Min
Max
Outdoor
S1
Mean
Std
Min
Max
S2
Mean
Std
Min
Max
S3
Mean
Std
Min
Max

Comfort parameters

0.3e1 mg mÀ3

1e2 mg mÀ3

2e5 mg mÀ3

5e10 mg mÀ3

0.3e10 mg mÀ3

CO2 (ppm)

T ( C)

RH (%)

11.2
3.1
6.7
16.9
11.4
3.4
8.0
16.6
14.3
10.0
7.4
30.8

(2.2)

(0.6)
(1.4)
(3.6)
(4.8)
(1.2)
(3.4)
(6.3)
(3.4)
(1.0)
(3.7)
(4.8)

7.1
1.6
4.6
9.2
6.7
2.2
4.8
10.1
4.8
0.8
3.7
5.7

(1.8)
(0.5)
(1.1)
(2.6)
(2.7)

(0.5)
(2.2)
(3.2)
(1.7)
(0.4)
(1.1)
(2.5)

59.9
18.2
35.9
94.2
48.4
9.6
37.5
62.3
29.0
6.2
20.8
35.9

(4.3)
(1.0)
(2.5)
(5.5)
(9.8)
(2.9)
(7.2)
(14.4)
(4.0)

(1.1)
(2.5)
(6.6)

54.4
18.1
31.1
90.4
58.0
9.2
45.2
67.4
41.1
10.6
25.7
49.1

(1.0)
(0.7)
(0.3)
(2.8)
(4.6)
(0.9)
(3.4)
(5.8)
(1.2)
(1.4)
(0.3)
(2.9)

132
40.7
78.3
210
124
23.6
95.5
156
93.3
24.5
57.6
113

(9.3)
(2.7)
(5.8)
(12.3)
(21.9)
(4.6)
(16.2)
(28.8)
(11.3)
(2.0)
(9.8)
(13.6)

1475
324
712
2900

1484
659
406
3488
690
375
372
2780

(504)
(130)
(377)
(1054)
(573)
(231)
(394)
(1534)
(404)
(51)
(357)
(867)

22.6
1.5
19.3
25.6
18.7
1.4
16.7
22.3

25.3
2.2
21.5
31.3

(20.9)
(1.4)
(18.3)
(24.3)
(17.9)
(0.9)
(16.4)
(20.1)
(24.3)
(3.0)
(18.3)
(34.7)

50.0
6.8
32.1
64.3
69.6
7.4
56.6
83.5
46.8
7.9
35.8
77.3

(49.8)
(5.2)
(37.8)
(66.4)
(64.0)
(3.3)
(58.5)
(75.5)
(48.1)
(6.2)
(25.3)
(69.5)

4.3
1.1
2.9
5.9
4.1
1.4
2.7
5.6
3.4
0.9
2.6
4.2

(4.7)
(1.4)
(3.8)

(7.7)
(4.0)
(1.4)
(2.4)
(6.2)
(6.4)
(2.6)
(2.7)
(9.1)

3.1
1.0
1.9
4.3
3.6
1.4
2.5
6.0
3.1
0.8
2.4
4.1

(2.4)
(1.3)
(1.5)
(5.1)
(2.6)
(1.1)
(1.4)

(4.2)
(4.2)
(1.9)
(1.6)
(6.5)

14.4
4.4
7.9
20.4
13.0
6.1
8.6
22.5
11.4
2.5
9.1
13.7

(9.0)
(3.2)
(6.7)
(16.0)
(7.1)
(2.8)
(4.0)
(10.8)
(14.3)
(7.2)
(5.4)

(23.3)

15.9
5.3
7.0
24.4
10.4
4.3
7.8
18.0
13.3
2.9
11.0
17.2

(5.0)
(1.0)
(3.6)
(6.3)
(3.0)
(0.9)
(1.8)
(4.9)
(5.7)
(2.3)
(2.9)
(8.7)

37.8
11.4

28.1
54.7
31.1
11.5
22.4
44.4
32.5
7.4
26.0
39.9

(21.1)
(6.5)
(15.9)
(35.2)
(16.6)
(6.0)
(10.1)
(25.8)
(30.6)
(13.9)
(12.5)
(47.1)

357
30
286
593
304
15

253
404
331
14.0
294
391

(336)
(17)
(278)
(498)
(304)
(27)
(260)
(1076)
(357)
(65.0)
(288)
(624)

15.9
2.7
10.8
24.0
21.6
4.1
13.3
30.2
16.5
3.3

11.1
25.5

(13.7)
(2.3)
(8.2)
(21.3)
(16)
(4.3)
(8.1)
(30)
(12.0)
(3.1)
(3.6)
(25.6)

47.0
20.1
16.8
83.3
64.4
15.9
18.4
83.4
40.3
18.7
10.9
84.0

(60.2)

(17.6)
(18.8)
(93.5)
(75.8)
(9.0)
(57.3)
(97.4)
(60.8)
(22.2)
(16.2)
(92.4)

Std: standard deviation.

the student performance during lessons. Outdoor CO2 levels were
comparable between the three schools at around 350 ppm.
As explained in the paragraph 2.2, we used correction factors to
estimate the PM10 mass concentrations in the 3 schools. The Cfactors for indoor air at S1, S2, and S3 were respectively 0.67, 0.72,
and 0.75. Note that according to the manufacturer GRIMM, for
outdoor PM, under normal conditions, the C-factor is usually in the
range of 0.8e1.2 and in our case, C-factor outdoor values were
indeed close to unity for the three schools. This implies that indoor
PM in this study present a density and/or a refractive index lower
than the PM model (density of approximately 1 g cmÀ3, and
refractive index of about 1.45) used for the GRIMM calibration, and
lower than outdoor ones. This could be due to a higher content of
organic compounds than outdoor [26]. These compounds could
reflect indoor VOC sources as well, such as paints, construction
materials, wood preservatives, solvents and cleaning products used
each day after school [26,27]. The same tendency of higher fraction

of organic compounds indoor than outdoor, (but with variable ratios) was observed for PM2.5 [22], and for PM10 [28]. Inversely, a
larger ion contents (ammonium, sulfates, nitrates …) with a higher
density than organic compound were observed in outdoor PM10
than indoor ones [28]. These chemical characteristics could induce
a difference in density that could results in lower C-factor for indoor PM.
In this study, the PM10 fraction comprise particles between
0.3 mm and 10 mm, a rather good estimate as particles below 0.3 mm
represents only a very small portion of total PM10 mass concentrations [29]. Globally, the corrected PM10 concentrations obtained
by the optical counters GRIMM in this paper were comparable
(within 5% in presence of occupants) with those obtained by
gravimetric methods [3]. The 8-h average PM10 concentrations
during occupancy were 88.4, 89.3 and 70.0 mg mÀ3 respectively at
S1, S2, and S3, while in absence of children, the 12-h average PM10
concentrations were respectively 9.3, 21.9, and 11.3 mg mÀ3. These
lower concentrations during unoccupied period suggested that

continuous indoor emission sources of PM such as paint, walls or
furniture were probably not significant. The questionnaire showed
that there was no specific source of particle emission as combustion
(cooking, household heating, smoking …) during the sampling
campaigns at these schools.
Outdoor, the PM10 concentrations varied from 31.1 mg mÀ3 (S2)
to 37.8 mg mÀ3 (S1) during teaching hours and from 16.6 mg mÀ3
(S2) to 30.6 mg mÀ3 (S3) when children were out of school.
Consequently, the presence/absence PM10 ratios varied from 3.4 to
5.5 indoors and from 1.3 to 1.8 outdoors, confirming the expected
larger impact of human activities in a closed environment. These
results are comparable with previous studies in classrooms
[16,23,30,31]. As indicated in Tran et al. [3], we didn't find any
significant influence due to the specific typology of the schools

selected (rural, urban and industrial) on indoor and outdoor PM10
mass concentrations. The higher outdoor PM10 concentrations
collected at S1 during the occupied period compared to S2 and S3
(Table 1) were attributed to more frequent dust-off of clouts near
the outdoor samplers in the courtyard of S1.
The number of pupils in the classrooms were comparable
(24e27 children), and the volume of the classroom at S3 (181 m3)
was slightly smaller than at S1 and S2 (201 m3 and 234 m3,
respectively). In absence of large continuous sources of PM beside
school activities during teaching hours, this last parameter should
favor higher indoor PM10 concentrations at S3 compared to the
other schools. Yet, indoor PM10 concentrations during teaching
hours at S3 were relatively lower than the others. This was linked
with the fact that the teachers opened frequently the windows and
doors at S3 (confirmed by the questionnaire), resulting in a dilution
of indoor air by a much lower content of PM10 from fresh ambient
air.
Based on the feedback form regarding daily activities on site, the
usual activities in classrooms were clearly the main sources of indoor PM10 as also noted in previous works performed in classrooms
[1,16,23,28,30].


186

D.T. Tran et al. / Building and Environment 81 (2014) 183e191

The PM10 values obtained during teaching hours largely exceeded the daily limit values (50 mg mÀ3) for indoor air (French recommendations), and were much higher than in absence of children
(Table 1). Outdoor, PM10 concentrations always respected these
daily limit values. On the other hand, the smaller PM2 fraction,
often considered more harmful, presented lower concentrations

than the PM2.5 daily limit recommendation from France
(25 mg mÀ3), both indoor and outdoor.
3.2. Indoor and outdoor particle massesize distribution
The relative percentages of different size fractions of indoor and
outdoor PM10 at S1, S2, and S3 are shown in Figs. 1 and 2.
Outdoors, the percentages of PM1, PM1e2, PM2e5, and PM5e10
during teaching hours were comparable and varied only slightly
from school to school. These small variations might be associated
with emission processes and/or the proximity of sources which
could modify the particle size distribution [32]. In addition, the
children's activities in the school's yards during breaks could also
influence the mass distribution of outdoor PM10 by adding up to the
coarser fractions. However, this phenomenon was partially masked
by the dilution of ambient air over weekly period probably
explaining why we did not observe much variation in outdoor PM10
concentrations. During unoccupied period, the tendency was
almost alike but with a decrease of the PM5e10 fraction and a
symmetrical increase of the finer fractions (Fig. 1). Indeed, the PM1,
PM1e2, PM2e5, and PM5e10 accounted respectively for 10e13%,
8e11%, 35e41%, and 33.4e42% of the total PM10 mass concentrations during teaching hours. When schools were vacant, these
values were respectively 21e24%, 11e15%, 42e46%, and 18e23%.
Indoors, the mass distributions variability of PM1, PM1e2, PM2e5,
and PM5e10 at S1, S2, and S3 were more remarkable. In presence of
pupils in the classrooms, the coarse fraction PM2e10 made up most
part of the total mass of PM10. For instance, this fraction accounted
for 86% (S1 and S2), and 75% (S3) of PM10 (Fig. 2). On the contrary, in
absence of children, there was a sharp drop of the coarse fractions,
in particular the PM5e10 fraction. Indeed, this last fraction represented 41.2e46.8% of PM10 during the teaching hours, but only
about 10% (S1 and S3), and 21% (S2) in absence of pupils. Inversely,
an increase of the fine fractions was observed from the occupied to

the unoccupied period, PM1 accounting for about 9% (S1 and S2)
and 19% (S3) during teaching hours and rising up to about 23% (S1
and S2) and 39% (S3) during unoccupied period (Table 1). These
results imply that the presence of occupants and their activities in

Fig. 1. Outdoor particle size-distribution during occupied and unoccupied periods.
PM10 (mg mÀ3) ¼ 37.8; 31.1; 32.5 for S1; S2; S3 during occupied period and ¼ 21.1; 16.6;
30.6 during unoccupied period, respectively.

Fig. 2. Indoor particle size-distribution during occupied and unoccupied periods. PM10
in mg mÀ3 ¼ 132; 124; 93.3 for S1; S2; S3 during occupied period and ¼ 9.3; 21.9; 11.3
during unoccupied period, respectively.

classrooms are at the origin of the coarse particles increase. In
absence of occupants and other significant emission sources of PM
in the classrooms, the relative fine fractions drastically increase in
the indoor environment. Similar observations on the impact of
occupants' activities indoors on coarse fractions were also reported
in previous works [16,18,23,30].
Interestingly, the relative percentage of the PM1 was higher at
S3 than at S1 and S2, both during teaching hours or vacant periods.
These observations are again linked to the large openings of doors
and windows during occupied period at S3 while doors and windows were systematically closed at S1 and S2. Moreover, the airtightness at S3 was lower than at S1 and S2 due to cracks on the
walls and around the windows and doors, resulting in a much
higher air exchange rate (estimated values by the CO2 decay
method are about 5 times higher at S3 than at S1 and S2 during
unoccupied period). Consequently, a higher air exchange rate could
promote a quicker dilution of the indoor coarse fraction by outdoor
air without influencing significantly the fine particles presenting
similar indoor and outdoor concentrations.

3.3. Real-time variation of indoor and outdoor particles
The real-time variation of indoor and outdoor PM concentrations combined with questionnaires information about teaching
activities allowed us to interpret their relative influence with time.
As a case of study, we only present here the indoor and outdoor fine
and coarse particles variations during 2 consecutive weeks at S1
(Figs. 3 and 4). Indeed, similar tendencies were evidenced at S2 and
S3; and while S1 presented a complete dataset, the others had
some missing data due to power outage at S2 and to instrumental
breakdown at S3.
In this paragraph, we grouped the fractions PM2e5 and PM5e10
into the so called coarse fraction PM2e10 as they showed similar
behaviors. For the same reasons, the PM1 and PM1e2 were grouped
to represent the fine fraction PM2.
As shown in Fig. 3, during the class, the mass concentration of
indoor PM2 was slightly higher than that measured outdoor
whereas in absence of occupants in the classrooms, it was the
opposite. The evolution of indoor fine particles did not seem to be
correlated to outdoor ones. It was also observed some relatively
small spikes of this size fraction. We associated them with occupant's activities in classrooms (walking, sitting, and playing), given
that during school days, the doors and windows were sometime
kept open which should have lead to similar concentrations indoor


D.T. Tran et al. / Building and Environment 81 (2014) 183e191

187

Fig. 3. Indoor and outdoor real-time variation of fine particle PM2 at S1 over a 2-week period.

Fig. 4. Indoor and outdoor real-time variation of coarse particles at S1 over a 2-week period.

and outdoor. The larger spikes were linked to sweeping activities
on the floor at the end of the school day.
The variation of indoor concentrations of the coarse fraction
(PM2e10) related to occupant's activities in classrooms was more
noteworthy (Fig. 4). Indeed, this size fraction increased sharply as
soon as the arrival of occupants (at about 8h40 AM) and phased out
quickly during both the breaks (at about 10h30 and 15h00) and the
lunch time (11h45e13h45) in weekdays (Fig. 5). The most intense
spikes of PM2e10 concentrations were observed during the cleaning
routine after the class. This type of activity could result in an increase of 50e70 times the indoor PM2e10 concentrations compared
to unoccupied period.

Outdoor, the variation of this fraction was not as significant and
sometimes was anti-correlated with indoor ones, particularly
during the breaks and lunch time when the pupils played in the
courtyards and close to the monitoring instruments.
3.4. Indoor/outdoor mass concentration ratios (I/O)
This paragraph evidences the integrative impacts of different
phenomenon such as the children's activities in the classroom, the
outdoor PM concentrations or the building air permeability on I/O
ratios for different particulate size fractions. The results are reported in Figs. 6 and 7. Note that the time-resolution data was used


188

D.T. Tran et al. / Building and Environment 81 (2014) 183e191

Fig. 5. Zoom on real-time variation of indoor coarse particles during teaching hours (June 8th 2009) at S1.

to calculate the average particle concentrations for both occupied
and vacant periods.
The Mann and Whitney U test show significant I/O ratio differences between presence and absence of children at p < 0.001 for
PM2 and PM2e10 (Figs. 6 and 7) whilst this ratio is comparable for
submicron particles PM1.
In presence of occupants, I/O ratios of the coarse fraction PM2e10
were comprised between 7.0 and 9.6 at S1, S2, and S3. These values
for the fine fraction PM2 were much lower and closer to 1, implying
the presence of indoor sources regarding the coarse fraction (Fig. 6).
These indoor sources were undoubtedly associated to children's
activities in classrooms which could emit PM directly with their
clothes, shoes, hair, and indirectly by the resuspension of previously deposited PM [33e36] The utilization of chalks during
teaching hours could also introduce an important source of coarse
PM [3,37].
In absence of occupants (evenings, nights, and week-ends), the
I/O ratios of all the particulate fractions were below 1, linked
probably to (1) the absence or negligible continuous indoor sources, (2) the particle filtration efficiency of the building envelope

[24,35,36,38e40], and (3) the higher indoor deposition factors than
outdoor due to the larger indoor surface/volume ratio [41].
Note that I/O ratios of the different size fractions at S3 were
always lower than at S1 and S2 (Fig. 7). This may seem contradictory with a higher air exchange rate at S3. In other words, S3 should
present a higher penetration of outdoor PM into the classrooms
resulting in higher I/O ratios at S3 (closer to one) during unoccupied period. That was probably linked to an artifact associated to
our sampling strategy. Actually, the measurements performed
during unoccupied period started about 2 h after the departure of
the children when the PM concentrations in indoor air could still be
relatively high, particularly the days when the cleaner swept the
floors, increasing de facto the I/O ratios. At S3 (rural school), the
doors were opened directly to the courtyard after class, allowing a

rapid equilibrium between indoor and outdoor PM which limited
the above-mentioned bias. By contrast, at S1 and S2, the doors were
only opened to the corridors, resulting in a lower air exchange with
outdoor and then in a slower decrease of the coarse fractions.
This bias was eliminated when integrating the sampling time
from midnight to the next day up to the arrival of children in the

Fig. 6. I/O ratios of different size fraction during occupied period (n ¼ 8, 6 and 4 at S1,
S2, and S3 respectively).

Fig. 7. I/O ratios of different size fraction during unoccupied periods (n ¼ 8, 6 and 4 at
S1, S2, and S3 respectively).


D.T. Tran et al. / Building and Environment 81 (2014) 183e191

classrooms (to make sure that most suspended particles were
deposited). By doing so, we found that the I/O ratios of different size
fractions were slightly higher at S3 compared to the other, but still
lower than 1. Indeed, the average I/O ratios for PM2 at S1 and S2
were about 0.30 against ~0.4 at S3. The corresponding values for
PM2e10 at S1 and S2 were 0.18 against 0.2 at S3. We demonstrate
here that when dealing with I/O ratios, one should carefully
consider the time resolution and sampling periods to avoid such
artifacts.
3.5. Contribution of PM emitted during occupied period to the
indoor PM sampled during unoccupied period
The class ended generally at 16h30 every school's day followed, 15e30 min after the departure of the children, by dusting
(twice a week), and sweeping/wet cleaning with detergents
(twice a week) the classrooms. The dusting/cleaning lasted

generally about 30 min, the cleaner leaving the classroom at
about 17h00e17h15. The movement of the occupants at the end
of the class as well as the vigorous sweeping activities resulted in
the suspension of a large amount of PM previously deposited on
the indoor surfaces (particularly on the floor) since the last
cleaning. In addition, the use of detergents for cleaning could
result in formation of new particles such as secondary organic
aerosols [42]. The high PM concentrations due to these activities
could modify the indoor particles behavior during unoccupied
period (Table 2). To examine this phenomenon, we calculated the
percentage of PM sampled from 19h00 (about 2 h after the departure of the cleaner) to 23h00 against the total PM sampled
during the whole night from 19h00e7h00 the next day (Table 2).
Based on the PM concentration decrease after the departure of
occupants, the 23h00 cap was set to make sure that most of the
fine and coarse PM previously suspended were quantitatively
deposited on the floor.
As shown in Table 2, the average mass of PM2 sampled during
the first third of the sampling period (19h00e23h00) at S1 and S2
accounted for 62% and 55% of the total PM2 collected during the
whole night (from 19h00e7h00, the next day). These values for the
PM2e10 fraction were significantly higher with 84% and 80% for S1
and S2, respectively. We can assume that the indoor PM concentrations present a rather constant decrease during the whole unoccupied period with a relative steady penetration of outdoor PM
and a similar deposition velocity (classrooms closed and no significant indoor emission sources). Although this hypothesis may be
questionable in terms of coagulation phenomenon and deposition
rate, this rough assumption may help us to estimate the impact of
the resuspension fraction. Theoretically, the PM2e10 sampled from
19h00e23h00 should represent about 33% of the total PM2e10
sampled during the whole unoccupied period. The same percentage should be found for the PM2 fraction.
By subtracting the theoretical concentration (33% for the three
schools) from the measured concentrations, we found that 51% (for

S1) and 47% (for S2) of the PM2e10 resuspended during occupied
period were accounted for unoccupied period. These values for the
PM2 were 29% and 22%, respectively for S1 and S2, confirming again
the larger influence of children's activities to the suspension of the
coarse fractions.

CDIðmg=kg=dayÞ ¼

189

Table 2
The percentage of PM sampled from 19h00e23h00 compared to the PM sampled
during the whole unoccupied period (sampling from 19h00e07h00 the next day).
Percentage of PM2
and PM2e10 (%)
PM2
PM2e10

Schools (n: Number of nights)
S1 (n ¼ 8)

S2 (n ¼ 6)

S3 (n ¼ 5)

62.3 ± 3.5
83.8 ± 2.8

54.5 ± 4.4
80.4 ± 2.6

30.1 ± 4.0
36.3 ± 4.5

It is interesting to note that at S3, the coarse and fine PM
sampled from 19h00e23h00 accounted for only one third of their
total mass sampled during the whole unoccupied period (Table 2),
matching the theoretical values and suggesting negligible impact of
the sweeping activities in this school. This is in agreement with the
strong air exchange rate during and after the teaching hours
resulting in a fast wash out of indoor PM by a cleaner outdoor air
and in a quick ex-filtration of indoor PM.
3.6. Heavy metal concentrations in classrooms and their risk
assessment
In this section, we reviewed the concentration of four elements
(As, Cd, Cr, and Ni) defined as carcinogens (Group 1) by the International Agency for Research on Cancer (IARC) in the PM2 as they
can reach the pulmonary alveoli and even penetrate into the
bloodstream [43]. Then, we estimated their potential risk during
teaching hours at S1 situated in an industrial zone. To do so, we
considered that the relative concentrations of these elements in
indoor PM do not change significantly during occupied period or
from day to day.
The concentration of As, Cd, Cr, and Ni were calculated using
mass concentration of PM2 obtained by GRIMM measurements
during teaching hours multiplied by the percentage of these elements in the PM2.5 reported in a previous work [3]. Consequently,
the percentage of As, Cd, Cr, and Ni in the PM2.5 may slightly differ
from PM2 discussed in this paper, somehow compensated by the
missing fraction from the GRIMM, unable to measure particles
below 0.3 mm.
As shown in Table 3, concentrations for As, Cd, Cr, and Ni were

calculated respectively at 0.46 ± 0.11 (ng.mÀ3), 0.11 ± 0.03 (ng.mÀ3),
1.27 ± 0.30 (ng.mÀ3), and 1.71 ± 0.40 (ng.mÀ3). These values are
relatively higher than that of S2 and S3 (results not shown here).
They are 3 times (for As) to more than 10 times (for Ni) higher
compared to the concentrations of these elements measured in 39
schools in Barcelona [44], although Cd concentrations are comparable. However, the concentrations of Ni and Cr were comparable
r et al. [45] conducted during five schools in
with the work of Molna
Stockholm, Sweden.
Regarding the risk assessment, we used the incremental cancer
risk factor as described in Feng et al. [46] considering a period of
five years (number of years that children study in an elementary
school). It can be calculated by multiplying the cancer potency
factor of a given carcinogen with the chronic daily intake (CDI). The
cancer potency factors of As, Cd, Cr, and Ni are 15; 6.3; 41; and 0.84
(mg/kg/day), respectively. The CDI is calculated using the following
equation.

À
 Á
À 
Á
C mg m3 $Intake rate m3 day $Exposureðdays=studied periodÞ
Body weightðkgÞ$5$ðy=studied periodÞ$365ðdays=yÞ


190

D.T. Tran et al. / Building and Environment 81 (2014) 183e191

Table 3
Estimated concentration of As, Cd, Cr, and Ni for indoor PM2 at S1 (n ¼ 8) and their incremental cancer risk for a period of five years.
Elements

Mean ± Std
(ng/m3)

As
Cd
Cr
Ni

0.46
0.11
1.27
1.71

±
±
±
±

0.11
0.03
0.30
0.41

Median (ng/m3)

Min (ng/m3)

Max (ng/m3)

Cancer potency
factor (mg/kg/day)

CDI for 5 years
(mg/kg/day)

Incremental cancer risk
for 5 years

0.47
0.11
1.28
1.71

0.29
0.07
0.79
1.05

0.66
0.16
1.81
2.42

15
6.3
41

0.84

4.92$10À8
1.17$10À8
1.35$10À7
1.81$10À7

0.74$10À6
0.07$10À6
5.55$10À6
0.15$10À6

where, C is the concentration of the studied element in the indoor
air, and intake rate is the amount of air that a child inhaled each day
for the studied period of five years.
According to the EPA [47], children in elementary schools aged
from 6 to 11 years have an average body weight of 31.8 kg. For an
average intensity of activities during teaching hours including
seating, playing, running, their inhalation rate is estimated at
1.32 m3 hÀ1 [47]. In this work, the teaching hours start at 08h45 and
end at 16h45 with a break of 1h30 for lunch. The exposed time of
children to indoor PM is therefore estimated at 6.5 (h). The average
volume of air inhaled by a pupil during occupied period (one school
day) will therefore be 8.58 m3. In addition, in France, the children
work normally 9 months/year and 4 days/week.
The estimated incremental cancer risk factor for As, Cd, Cr, and
Ni were respectively 0.74$10À6; 0.07$10À6; 5.55$10À6, and
0.15$10À6. These values, except for Cr, are lower than the limit that
is usually set at 1 Â 10À6 corresponding to a life time exposure to
unpolluted ambient environment [46]. This implies low cancer risk

regarding these elements in classrooms.
However, the cancer risks calculated for the children in the
above-equation in this section only takes into account 6.5 h/day,
4 days/week, and 9 months/year of time exposure to air containing
these trace elements representing only a small fraction of their real
total exposure (~11%). If we consider that the concentrations of As,
Cd, Cr, and Ni in the classrooms are similar to other indoor environments where children are present (home, transport facilities …)
and they spent 80% of their time indoor, the cancer risks regarding
these elements will be respectively 1.50$10À6; 0.15$10À6; 11.3$10À6;
0.31$10À6. These values are above (for As and Cr), or lower than the
limit 1 Â 10À6 (for Cd and Ni). However, when considering the
bioavailable form that permits to more accurately assess the environmental and health risks, the cancer risk factor could decrease
significantly (about 10 times for Cr) according to Feng et al. [46].
This suggests little cancer risk for the 4 studied elements when
using their bioavailable form in the PM2 in school environment.
Teachers spent more time in the classrooms during their active
life than children leading to longer exposure to indoor PM containing heavy metals As, Cd, Cr, and Ni. Nevertheless, calculated
values of cancer risk regarding As, Cd, Cr, and Ni in the indoor PM2 for
teachers were comparable with that of children, assuming parameters for a teacher as follow: 65 kg of body weight, 20 m3 of inhaled
air/day for 38 years of carrier and a life expectancy of 75 years.
Finally, in this study we focused on the estimation of cancer risk
regarding the fine fraction PM2 while occupants inhale also the
coarse PM2e10 presenting higher concentrations in the classrooms.
Indeed, this coarse fraction may ultimately be removed from the
pulmonary system by the mucociliary clearance, then swallowed
and digested, freeing pollutants to the systemic circulatory system.
This may represent a supplementary effect to cancer risk assessment that should be accounted for in future works.
4. Conclusions and perspectives
This paper demonstrated the influence of children's activities in
the classrooms on the indoor coarse and fine particles behaviors.

During teaching hours, when there were intense activities, the
PM10 concentrations in the classrooms of the 3 schools were
significantly higher than the daily recommendation values according to French standards that could cause health problems on
the fragile children respiratory systems. The occupant's activities
impacted essentially the coarse fractions whereas in absence of
activities, the fine fraction became more important in the air.
Regarding occupied period, indoor PM fate and sources were
primarily associated to occupant's activities whereas during unoccupied period, these behaviors and sources were still strongly
influenced by the PM resuspended during teaching hours, many
hours after the class. More generally, the influence of the occupant's activities may overcome various continuous indoor sources
as well as the impact of the building permeability on the indoor
PM concentrations. The PM emitted by pupil's during occupied
periods as well as cleaning activities after the class could account
for up to 50% for the coarse PM and 30% for the fine ones sampled
during unoccupied period. Consequently, attention to this phenomenon should be paid when one want to study the dynamic
of particles or more generally the air quality of an indoor
environment.
The occupant's activities in the classroom might also result in
the higher organic fraction in PM10 than outdoor ones which favor
in return the low density of indoor PM10.
Regarding some potential carcinogen elements such as As, Cd,
Cr, and Ni in the indoor PM2, the estimated results showed that the
cancer risk associated with these elements was not significant for
children except for Cr. Moreover, taking into account a time exposure in other indoor environment and considering the bioavailable
fraction of these metals, both children and teachers spending a lot
of time in classroom during their career present very low risk of
cancer due to PM2 inhalation.
It would be pertinent to quantify both the global contribution of
all emission sources as well as each independent source to the

coarse and fine PM concentrations during the teaching hours by
looking at a large variety of organic and inorganic compounds. In
addition, this information along with parameters of particle dynamic in indoor air such as penetration capacity, deposition velocity and air exchange rate may allow one to estimate the real-time
exposure of children to indoor PM and therefore to study the influence of specific indoor sources on children's health.
Acknowledgment
This research project was part of the “Institut de Recherche en
Environnement Industriel” (IRENI) and was financially supported
by the Nord-Pas-de-Calais Regional Council (Program IRENI, Action
gionale e CPER DT
1, Axe 1 & 2 e Convention ANR 2006 e Action re
cision n 06 e CPER n 093-01 e Convention FEDER,
n 15006 e De
Obj 1 e 2006/3 e 4.1 n 79/8559 (2006e2008)), by the French
Ministry of Higher Education and Research, by the European
Regional Development Funds (through the Regional Delegation for
Research and Technology) and Mines Douai-Armines Research
Centre. We thank Bruno Malet for its valuable and continuous
technical support.


D.T. Tran et al. / Building and Environment 81 (2014) 183e191

References
 a
nek J, Hytychov
[1] Branis M, Safr
a A. Indoor and outdoor sources of sizeresolved mass concentration of particulate matter in a school gym - implications for exposure of exercising children. Environ Sci Pollut Res 2011;18:
598e609.
[2] Franklin PJ. Indoor air quality and respiratory health of children. Paediatr
Respirat Rev 2007;8:281e6.

[3] Tran DT, Alleman LY, Coddeville P, Galloo JC. Elemental characterization and
source identification of size resolved atmospheric particles in French classrooms. Atmospher Environ 2012;54:250e9.
[4] WHO-Europe. Health aspects of air pollution results from the WHO project:
systematic review of health aspects of air pollution in Europe; 2004. p. 24.
[5] Englert N. Fine particles and human health - a review of epidemiological
studies. Toxicol Lett 2004;149:235e42.
[6] Lewtas J. Air pollution combustion emissions: characterization of causative
agents and mechanisms associated with cancer, reproductive, and cardiovascular effects. Mutat Res 2007;636:95e133.
[7] Pope CA, Burnett RT, Thun MJ, Calle EE, Krewski D, Ito K, et al. Lung cancer,
cardiopulmonary mortality, and long-term exposure to fine particulate air
pollution. J Am Med Assoc 2002;287(9):1132e41.
[8] Salonen RO, Pennanen AS, Vahteristo M, Korkeila P, Alm S, Randell JT. Health
risk assessment of indoor air pollution in Finnish ice arenas. Environ Int
2008;34:51e7.
€ rres RA, et al. Toxicity
[9] Oeder S, Dietrich S, Weichenmeier I, Schober W, Pusch G, Jo
and elemental composition of particulate matter from outdoor and indoor air of
elementary schools in Munich, Germany. Indoor Air 2011;22(2):148e58.
[10] Vousta D, Samara C. Labile and bioaccessible fractions of heavy metals in the
airborne particulate matter from urban and industrial areas. Atmospher Environ 2002;36(22):3583e90.
[11] WHO-Europe. Air quality guidelines global update 2005: particulate matter,
ozone, nitrogen dioxide and sulfur dioxide; 2006. p. 484.
[12] Donaldson K, Tran CL, Macnee W. Deposition and effects of fine and ultrafine
particles in the respiratory tract. Eur Respir Monogr 2002;21:77e92.
[13] Guo H, Morawska L, He C, Gilbert D. Impact of ventilation scenario on air
exchange rates and on indoor particle number concentrations in an airconditioned classroom. Atmospher Environ 2008;42:757e68.
[14] Janssen NAH, Hoek G, Harssema H, Brunekreef B. Childhood exposure to
PM10: relation between personal, classroom, and outdoor concentrations.
Occup Environ Med 1997;54:888e94.
[15] Alshitawi MS, Awbi HB. Measurement and prediction of the effect of students’

activities on airborne particulate concentration in a classroom. VAC&R Res
2011;17(4):446e64.
[16] Chithra VS, Shiva Nagendra SM. Indoor air quality investigations in a naturally
ventilated school building located close to an urban roadway in Chennai,
India. Build Environ 2012;54:159e67.
[17] Diapouli E, Eleftheriadis K, Karanasiou AA, Vratolis S, Hermansen O, Colbeck,
et al. Indoor and outdoor particle number and mass concentrations in Athens.
Sources, sinks and variability of aerosol parameters. Aerosol Air Qual Res
2011;11:632e42.
[18] Kingham S, Durand M, Harrison J, Cavanagh J, Epton M. Temporal variations in
particulate exposure to wood smoke in a residential school environment.
Atmospher Environ 2008;42:4619e31.
[19] Tippayawong N, Khuntong P, Nitatwichit C, Khunatorn Y, Tantakitti C. Indoor/
outdoor relationships of size-resolved particle concentrations in naturally
ventilated school environments. Build Environ 2009;44:188e97.
[20] Ho KF, Cao JJ, Harrison RM, Lee SC, Bau KK. Indoor/outdoor relationships of
organic carbon (OC) and elemental carbon (EC) in PM2.5 in roadside environment of Hong Kong. Atmospher Environ 2004;38:6327e35.
[21] Koponen IK, Asmi A, Keronen P, Puhto K, Kulmala M. Indoor air measurement
campaign in Helsinki, Finland 1999: the effect of outdoor air pollution on
indoor air. Atmospher Environ 2001;35:1465e77.
[22] Na K, Cocker III DR. Organic and elemental carbon concentrations in fine
particulate matter in residences, schoolrooms, and outdoor air in Mira Loma,
California. Atmospher Environ 2005;39:3325e33.
 wka I, Krupin
 ska B, Zwo
[23] Zwo
zdziak A, So
zdziak J, Nych A. Infiltration or indoor
sources as determinants of the elemental composition of particulate matter
inside a school in Wroclaw, Poland? Build Environ 2013;66:173e80.

191

[24] Poupard O, Blondeau P, Iordache V, Allard F. Statistical analysis of parameters
influencing the relationship between outdoor and indoor air quality in
schools. Atmospher Environ 2005;39:2071e80.
[25] Handbook: fundamentals. American Society of Heating, Refrigerating and AirConditioning Engineers. (ASHRAE) Inc; 1993, ISBN 0910110972.
lisation du comportement dyna[26] Tran DT. Identification des sources et mode
rieur des e
coles [PhD thesis]. University of
mique des particules dans l'air inte
Science and Technology of Lille; 2011.
[27] Alves CA, Urban RC, Pegas PN, Nunes T. Indoor/Outdoor relationships between
PM10 and associated organic compounds in a primary school. Aerosol Air Qual
Res 2014;14:86e98.
[28] Fromme H, Diemer J, Dietrich S, Cyrys J, Heinrich J, Lang W, et al. Chemical and
morphological properties of particulate matter (PM10, PM2.5) in school classrooms and outdoor air. Atmospher Environ 2008;42:597e605.
[29] Ntziachristos L, Ning Z, Geller MD, Sheesley RJ, Shauer JJ, Sioutas C. Fine, ultrafine and nanoparticle trace element compositions near a major freeway
with a high heavy-duty diesel fraction. Atmospher Environ 2007;41(27):
5684e96.
 
[30] Branis M, Safr
anek J, Hytychov
a A. Exposure of children to airborne particulate
matter of different size fractions during indoor physical education at school.
Build Environ 2009;44:1246e52.
[31] Fromme H, Twardella D, Dietrich S, Heitmann D, Schierl R, Liebl B, et al.
Particulate matter in the indoor air of classrooms-exploratory results from
Munich and surrounding area. Atmospher Environ 2007;41:854e66.
[32] Ehrlich C, Noll G, Kalkoff WD, Baumbach G, Dreiseidler A. PM10, PM2.5 and

PM1.0 emissions from industrial plants-Results from measurement programmes in Germany. Atmospher Environ 2007;41:6236e54.
[33] Goyal RL, Khare M. Indoor-outdoor concentrations of RSPM in classroom of
naturally ventilated school building near urban traffic roadway. Atmospher
Environ 2009;43:6026e38.
[34] Luoma M, Batterman SA. Characterization of particulate emissions from
occupant activities in offices. Indoor Air 2001;11(1):35e48.
[35] Thatcher TL, Lunden MM, Revzan KL, Sextro RG, Brown NJ. A concentration
rebound method for measuring particle penetration and deposition in the
indoor environment. Aerosol Sci Technol 2003;37:847e64.
[36] Thatcher TL, Layton DW. Deposition, resuspension and penetration of particles within a residence. Atmospher Environ 1995;29(13):1487e97.
[37] Majumdar D, Prince William SPM. Chalk dust fall during classroom teaching:
particle size distribution and morphological characteristics. Environ Monit
Assess 2009;148:343e51.
[38] Chao CYH, Wan MP, Cheng ECK. Penetration coefficient and deposition rate as
a function of particle size in non-smoking naturally ventilated residences.
Atmospher Environ 2003;37:4233e41.
[39] Lazaridis M, Smolík J, Kulmala M. Particle size characterization and emission
rates during indoor activities in a house. Atmospher Environ 2006;40:
4285e307.
[40] Nazaroff WW. Indoor particle dynamics. Indoor Air 2004;14(7):175e83.
[41] Howard-Reed C, Wallace LA, Emmerich SJ. Effect of ventilation systems and
air filters on decay rates of particles produced by indoor sources in an occupied townhouse. Atmospher Environ 2003;37:5295e306.
[42] Morawska L, He C, Johnson G, Guo H, Uhde E, Ayoko G. Ultrafine particles in
indoor air of a school: possible role of secondary. Environ Sci Technol
2009;43:9103e9.
[43] Schwarze PE, Ovrevik J, Lag M, Refsnes M, Nafstad P, Hetland RB, et al. Particulate matter properties and health effects: consistency of epidemiological
and toxicological studies. Hum Exp Toxicol 2006;25(10):559e79.

[44] Viana M, Rivas I, Querol X, Alastuey A, Sunyer J, Alvarez-Pedrerol
M, et al.

Indoor/outdoor relationships of quasi-ultrafine, accumulation and coarse
mode particles in school environments in Barcelona: chemical composition
and Sources. Atmospher Chem Phys 2013;13:32849e83.
€llsten G, Boman J. Indoor and outdoor concentrations
[45] Moln
ar P, Bellander T, Sa
of PM2.5 trace elements at homes, reschools and schools in Stockholm, Sweden. J Environ Monit 2007;9:348e57.
[46] Feng XD, Dang Z, Huang WL, Yang C. Chemical speciation of fine particle
bound trace metals. Int J Environ Sci Technol 2009;6:337e46.
[47] EPA/600/R-10/030. Available at: />