You are on page 1of 11

Science of the Total Environment 408 (2010) 4999–5009

Contents lists available at ScienceDirect

Science of the Total Environment

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s c i t o t e n v

Impact of fugitive emissions in ambient PM levels and composition

A case study in Southeast Spain
M. Santacatalina a,⁎, C. Reche b, M.C. Minguillón b, A. Escrig c, V. Sanfelix c, A. Carratalá a, J.F. Nicolás d,
E. Yubero d, J. Crespo d, A. Alastuey b, E. Monfort c, J.V. Miró e, X. Querol b
Department of Chemical Engineering, University of Alicante, P. O. Box 99, 03080 Alicante, Spain
Institute of Environmental Assessment and Water Research, IDAEA, CSIC, Jordi Girona 18-26, 08034 Barcelona, Spain
Instituto de Tecnología Cerámica, Universitat Jaume I, Campus Universitario Riu Sec, 12006 Castellón, Spain
Laboratory of Atmospheric Pollution (LCA), Miguel Hernández University, Av. de la Universidad s/n, Edif. Alcudia, 03202 Elche, Spain
Conselleria de Medi Ambient, Aigua, Urbanisme i Habitatge, C/ Francisco Cubells, 7-46011 Valencia, Spain

a r t i c l e i n f o a b s t r a c t

Article history: The results of this study show the high impact that anthropogenic fugitive emissions of mineral dust have on
Received 13 May 2010 air quality (levels of PM10, PM2.5 and some metals) in a region in SE Spain named L'Alacantí. This could be
Received in revised form 12 July 2010 extensive to other areas of Europe with similar characteristics. Fugitive emissions, such as those arising from
Accepted 14 July 2010
large public construction works, cement and ceramic manufacturing, mining, heavy industries, handling and
Available online 11 August 2010
transport of powdered raw materials and road dust, are very often left out of emission monitoring and
inspections in Europe. The comparative study of daily PM10 series in the area shows how the increase of annual
Atmospheric pollution average PM10 concentrations over 40 μg/m3 is due to extreme episodes occurring in 2006 and 2007, at a
Aerosols regional scale, given the simultaneous recording of PM episodes at distant monitoring sites. The annual average
PM10 values of the PM10 concentrations were close to or slightly higher than 40 μg/m3 (limit value of Directive 2008/
Source apportionment 50/CE) during 2006–2007 (Alicante-University 39–41, Agost 40–42, Sant Vicent 42–46, Alicante-El Plà 40–
Cement 42 μg/m3). The main PM10 sources in the zone were identified with the assistance of the PMF receptor model.
Ceramic Six common factors were determined, mineral as a main source (37% at Agost and 32% at Sant Vicent), road
Trace elements
traffic, secondary sulfate, petroleum coke, sea spray and industry. Mineralogical studies, with XRD and SEM-
Fugitive emissions
EDX techniques, support the hypothesis that the highest PM episodes are associated to fugitive emissions of
mineral matter. Despite the fact that L'Alacantí region is a heavily industrialized area with two cement plants
and a significant number of ceramic manufacturing plants, the fugitive emissions may have accounted for the
exceedances of the PM limit values during these two years, part of them caused by the construction of a
highway. These results may contribute to the interpretation of prior studies on source apportionment carried
out in Southern Europe, with very high loads of anthropogenic dust in PM10 and PM2.5.
© 2010 Elsevier B.V. All rights reserved.

1. Introduction including traffic and industrial hotspots, as well as an obligation to

reduce future PM2.5 levels (a variable reduction percentage depending
Atmospheric particulate matter (PM) is a key parameter in air on the initial concentrations; in the worst case the target would be
quality monitoring, not only because of its potential health and 18 μg/m3). Concerning the chemical composition of PM10, Directives
environmental impacts (WHO, 2003; EPA, 2006; Bates et al., 2005; Reis 2004/107/CE and 2008/50/CE fixed annual target values for As (6 ng/
et al., 2002), but also for its important role in global climate and m3), Cd (5 ng/m3) and Ni (20 ng/m3), and an annual limit value for Pb
atmospheric chemistry (IPCC, 2007). PM standards have been (500 ng/m3).
established and modified in recent years in many regions of the Specific anthropogenic sources may make important contributions
world. Air Quality Directive 2008/50/CE established an annual limit to ambient PM levels, especially when urban agglomerations and
value of 40 μg/m3 for the annual arithmetic mean of PM10 concentra- several production plants concentrate in a region, as happens in this
tions and a daily limit value for PM10 of 50 μg/m3 not to be exceeded study. Furthermore, natural emissions may considerably influence PM
more than 35 times per year (90.4 percentile). Furthermore, an annual levels around large natural PM emission sources (mainly arid and
limit for PM2.5 (25 μg/m3) was set up for all types of environments, semi-arid regions), such as the Mediterranean Basin (Rodríguez et al.,
2001, Escudero et al., 2007).
In Spain, more than 30,000 tons of industrial primary PM10 were
⁎ Corresponding author. Tel.: + 34 965903400; fax: + 34 965903826. emitted into the atmosphere in 2007, of which 6% corresponds to
E-mail address: (M. Santacatalina). cement industries and 19% to ceramic industries (EPER, The European

0048-9697/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
5000 M. Santacatalina et al. / Science of the Total Environment 408 (2010) 4999–5009

Pollutant Emission Register, A number of handled for this project, constituting a potential source of fugitive dust
these industries are located in an area of around 625 square kilometers emissions.
in southeast Spain called L'Alacantí (Fig. 1). There are two cement The study area (25 × 20 km) is located in a semi-arid region
production plants located 2 km to the NW of the city of Alicante and (annual precipitation less than 300 mm), bordered by the Mediter-
their production is above 1.5 million tons per year. This, together with ranean Sea to the east (Fig. 1) and surrounded from north to south by
the production in the neighboring Valencian region, accounts for 10% mountains forming a basin. Mild climatology and long periods
of Spanish cement production, which is 16% of the total European without complete renovation of air masses, derived from the
production (EPER). The ceramic plants, producing mainly bricks and prevalence of anticyclonic meso-scale scenarios and the dominance
roof tiles, are located around the city of Agost, where production of sea-breeze circulations, favor the accumulation of local emissions
approaches 2 million tons per year. Table 1 shows a summary of the and the development of episodes of regional contamination (Millán et
main emissions arising from these industries, as well as the major al., 1997, 2000). Hence, an analysis of the sources contributing to PM
control techniques and their effectiveness (IPTS, 2006; IPTS, 2007; levels in the study zone is especially necessary given the complex
IPTS, 2010; Minguillón et al., 2009; Monfort et al., 2009; US, EPA, 1995). mixture of anthropogenic and potential natural emissions in the area.
As it can be seen in the table several types of industrial emissions were The objective of the present work is to interpret the variability of
identified. Hence, depending on the type of source, these emissions levels, composition and source contributions to ambient particulate
have been divided into channeled/ducted and fugitive/diffuse. The matter (PM10 and PM2.5) through 2005–2008 period, during which
former are released into the atmosphere at a single point through a PM levels varied widely and showed a significant reduction in 2008.
duct or chimney, and the major control techniques used to reduce the Identifying the causes of ambient PM concentration in an area with
particulate matter emissions are bag filters or electrostatic precipita- complex emission scenarios, including anthropogenic emissions from
tors, being important to consider the emission temperature to select industrial activities such as cement, ceramic and other related
the most appropriate cleaning system. The latter are released into the industries, large public construction works and urban areas, as well
atmosphere in a disperse way from a surface or volume. The control as natural emissions such as African outbreaks and soil erosion, is
techniques to avoid these fugitive emission are based on enclosure of considered an example of great interest for the scientific community
the dusty operations (screening, crushing, etc.), control of the bulk and air quality managers.
storage and handling areas (perimetral fences acting as a wind
barriers, irrigation of piles with chemical dust suppressor, pile 2. Methodology
formation with telescopic feeders, etc.), and minimization of the raw
materials transport emissions (road paving, irrigation of unpaved 2.1. Sampling sites
road, etc.).
Moreover, and related to the previous activities, reservoirs of clay, Four sampling sites were used in this study (Fig. 1 and Table 2),
gypsum, and calcareous rocks are present in the study area, and named Agost, Sant Vicent, Alicante-El Plà, and Alicante-University. The
activities such as quarrying, grinding and transporting these also first three sites belong to the air quality monitoring network of the
occur in this area. The number of these kinds of facilities in the area is Autonomous Government of Valencia (Generalitat Valenciana), while
about ten and data regarding their production and contribution to PM the last belongs to the University of Alicante. More details follow.
emissions is scarce. Agost: the site is located in the center of the town (4766
Moreover, a large transect of a highway was constructed in the study inhabitants) bearing the same name, surrounded by several local
area during 2006–2007, and many tons of soil and dusty materials were ceramic emission sources (1–4 km away). This can be considered an

Fig. 1. Map of the study area with indication of sampling sites and industrial activities location.
M. Santacatalina et al. / Science of the Total Environment 408 (2010) 4999–5009 5001

Table 1
Industrial emissions in the study area and major control techniques used.

Process Stage Type of emissiona Major control technologies and theoretical efficiency

Ceramic manufacturing Raw materials preparation Fugitive Partial road paving (90%)
Irrigationb (50–75%)
Milling Channeled cold Bag filters (99%)
Drying Channeled hot No significant measures
Firing (using petroleum coke, biomass and/or natural gas as fuel) Channeled hot No significant measures
Cement manufacturing Raw materials preparation Fugitive Perimetral fences/wind barriers (30%)
Partial road paving (90%)
Irrigationb (50–75%)
Enclosure of dusty operations (70–99%)
Pile forming with telescopic feeders (75%)
Water spraying with chemical dust suppressors (90%)
Firing Channeled hot Electrostatic precipitator (99%)
Colder Channeled hot Bag filter (99%)
Milling Channeled cold Cyclone + bag filter (99%)
Cold: emission at room temperature. Hot: emission above room temperature, usually from combustion processes.
The only abatement technology applied to large public works.

urban-industrial background site, influenced by road traffic emissions University (82 samples from July 2004–July 2005, and 38 samples from
from several roads and highways crossing the area (the one under January–December 2006). At Sant Vicent and Agost, one random filter
construction during 2006–2007 was at a distance of 5 km from this was selected every week, approximately, in order to obtain a rep-
station). resentative PM composition. In contrast, the samples from Alicante-
Sant Vicent: an urban-industrial background site located in a University were selected on the days of highest PM levels and different
school in the center of the city of Sant Vicent del Raspeig (49,341 meteorological conditions, in order to study the possible extreme
inhabitants), located less than two kilometers from the cement plants. concentrations of some trace elements. These different selection criteria
Alicante-El Plà: an urban background site located in the northern should be kept in mind when comparing period-averaged concentrations
coastal part of the city of Alicante (322,673 inhabitants). In addition to of the measured species.
the urban emissions, it can be influenced by emissions from the cement One-half of each filter was acid digested with HF:HClO4:HNO3 (Sant
plants transported by the prevailing wind directions, NW winds. Vicent and Agost samples) or HNO3:H2O2 (Alicante-University); major
Alicante-University: a suburban-industrial background site locat- and trace elements were determined in the resulting solution by
ed on the university campus between the residential areas of Sant Inductively Coupled Plasma Atomic Emission Spectroscopy (ICP-AES)
Vicent del Raspeig and Alicante. It is quite close to Sant Vicent but with and Inductively Coupled Plasma Mass Spectrometry (ICP-MS). Condi-
less urban character and a different orientation with respect to the tions used during the measurement for ICP-MS were 14 bar for cool gas
prevailing wind direction and cement plants' locations. flow, 0.95 bar for auxiliary gas flow and 0.90 bar for nebulizer gas flow;
low power was 1000 w. Conditions used during the measurement for
2.2. Sampling instrumentation, schedule and chemical analyses ICP-AES were 15 bar for plasma gas flow, 0.2 bar for auxiliary gas flow
and 0.85 bar for nebulizer gas flow; low power was 1300 w. CCT
Daily 24-h sampling was carried out from July 2004–December (Collision Cell Technology) technique avoids and control the interfer-
2008 at Alicante-El Plà and Alicante-University; from August 2005– ences in the ICP-MS. A one-quarter fraction of each filter was used for
December 2008 at Sant Vicent; and from September 2005–December water extractions, the resulting solution being used to analyze soluble
2008 at Agost by high-volume samplers (30 m3/h): DIGITEL DL77 ions (NO− 2−
3 , SO4 , Cl

and NH+ 4 ) by Ion Chromatography (IC) HPLC
(Alicante-University), and DIGITEL DH80 (remaining sites). At Agost, (High Pressure Liquid Chromatography) using a DIONEX DX-20 with
Sant Vicent and Alicante-El Plà, PM10 and PM2.5 inlets were switched AS9-HC 250x4 mm, with Na2CO3 9 mM as eluent, running at 1.4 mL/min
according to a sampling schedule: PM2.5 was measured during one (anions). Ammonia was performed also by IC using a CS12A 250 × 4 mm
week out of every four weeks, and PM10 was measured during the cationic column and 20 mM metanesulfonic eluent at a flow rate of
other three weeks. At Alicante-University, PM10 was measured during 0.8 mL/min (Sant Vicent and Agost) or colorimetrically indophenol
the whole sampling period. method (Grasshoff et al., 1999) (Alicante-University). The total carbon
Circular (15 cm) quartz microfiber filters were used. They were was determined by means of a LECO elemental analyzer. Fractions of
placed before and after sampling for 48 h at 50% relative humidity and blank filters were analyzed in the batches of their respective filter
21 °C, and subsequently weighed gravimetrically. samples and the corresponding blank concentrations were subtracted
Chemical analyses were done to a selection of the collected filters from each sample. A few milligrams of the reference material NIST
from September 2006–September 2008 at Sant Vicent (67 samples) and 1633b (Sant Vicent and Agost) or NIST 1648 (Alicante-University) were
Agost (67 samples), and from July 2004–December 2006 at Alicante- added to a fraction of a blank filter so that the resulting concentration is
in the range of the samples concentration to confirm the analysis
accuracy of the acidic digestions.
Table 2 The silica content was estimated according to the relation SiO2 =
Location and type of the sampling sites. 3 ⁎ Al2O3 (Dulac et al., 1992; Molinaroli et al., 1993; Minguillón et al.,
2007a). Carbonate concentration was calculated based on stoichiomet-
Station Longitude Latitude Altitude Type of site
ric relations (CO2−
3 = 1.5 ⁎ Ca+ 2.5 ⁎ Mg). Organic matter + elemental
Agost 0° 38′ 17″ West 38° 26′ 11″ North 312 m Urban-industrial
carbon (OM + EC) was calculated from the fraction of organic +
Sant Vicent 0° 31′ 40″ West 38° 23′ 51″ North 123 m Urban-industrial elemental carbon (OC + EC), determined by subtracting the mineral
background carbon (calculated CO2−3 concentrations) from the total carbon. Then to
Alicante-El Plà 0° 28′ 16″ West 38° 21′ 31″ North 45 m Urban background account for the mass of unmeasured hydrogen and oxygen in the
Alicante- 0° 30′ 57″ West 38° 23′ 14″ North 120 m Urban-industrial organic material, equation OM+ EC= 1.2 ⁎ (OC + EC) (Eatough et al.,
University background
1996; Putaud et al., 2000) was applied.
5002 M. Santacatalina et al. / Science of the Total Environment 408 (2010) 4999–5009

The mineralogy of the collected material was determined for some in the regression, owing to the relatively low reconstruction of the PM10
samples by X-ray diffraction (XRD). To decrease the interferences mass by the analysis of the measured species.
caused by quartz filter in the XRD prior to analysis, particulate matter PMF was applied separately to the Sant Vicent and Agost data sets.
was extracted from the filters with ethanol in an ultrasonic bath (3 s), Alicante-University was not selected since the days selected for
then deposited on a silver filter. chemical analyses did not represent the average conditions of the site,
The morphology of the particles was studied using a scanning but rather extreme values. The limited size of the input data matrices,
electronmicroscope (SEM) JEOL6400. Analyses were performed man- together with the profusion of missing values for some species
ually on gold coated samples using an energy dispersive X-ray (particularly for NO− − +
3 , Cl and NH4 ), constrains the PMF results to be
microanalysis system (EDX) with aspectrum acquisition time of 30 s merely exploratory, rather than of quantitative nature.
live time, all particles present in several randomly chosen areas of the
sample were analyzed in its centre. Microscope working distance was 3. Results and discussion
14 mm, accelerating voltage 15 kV, and beam current was approxi-
mately 1.00 μA. 3.1. PM levels

2.3. Additional tools The annual average values of the PM10 concentrations were close
to or slightly higher than 40 μg/m3 (limit value of Directive 2008/50/
For each day of the sampling period, 5-day back-trajectories of the air CE) during 2006–2007 (Alicante-University 39–41, Agost 40–42, Sant
masses arriving to the L'Alacantí area were calculated using the HYSPLIT Vicent 42–46, Alicante-El Plà 40–42 μg/m3), and quite lower in 2008,
4 model at three different altitudes (500, 1500 and 2500 m.a.s.l.) (Alicante-University 30, Agost 31, Sant Vicent 30, Alicante-El Plà
(Draxler and Rolph, 2003) in order to identify the days with African dust 35 μg/m3) (Table 3). Thus, the annual PM10 limit value is exceeded at
episodes. all sites in some years. African intrusions are relatively common in
L'Alacantí region, so calculations of their influence were carried out.
Even if the African contribution is not taken into account (Escudero et
2.4. Source apportionment al., 2007), the mean PM10 levels exceeded the limit several years at
Agost, Sant Vicent and Alicante-El Plà, with a maximum value of
The main PM10 sources in the zone were identified with the 42 μg/m3 at Sant Vicent. On average, 37% of the exceedances of the
assistance of the Positive Matrix Factorization (PMF) receptor model daily limit value are due to African dust outbreaks, which is within the
(Paatero and Tapper, 1994). This technique was selected because of its range of contribution of other regions in the SE Mediterranean (31–
superior properties with respect to traditional factor analysis methods, 39%, Minguillón et al., 2007a). The levels' high homogeneity across all
such as the minimum variance property of the estimated parameters or sites is remarkable for the years of maximum coverage (2006–2007),
the efficient handling of missing values. The PMF calculations were with a variation range for the annual average of 39–46 μg/m3
computed with ME-2 software (Paatero, 1999). The uncertainties (Table 3) and a large correlation between the daily levels of different
connected with each data value, used as weights for the PMF least stations (R2 = 0.6) (Fig. S1). This suggests that emission sources have
squares scheme, were estimated in a manner similar to the one a regional impact, since local influence would lead to larger
described by Escrig et al. (2009). The robust mode was used for the differences.
calculations (Paatero, 1997). The scale of the factors was established by Regarding the attainment of the daily limit value (Directive 2008/
regression of the factor scores against total mass. An intercept was used 50/CE), the 90.4 percentile exceeded this limit (50 μg/m3) at all sites,

Table 3
Mean levels of PM10 and PM2.5, exceedances of the daily limit (50 μg/m3), with and without North African advections (NAF), 90.4 percentile and number of samples for each site
during the study period (July 2005–July 2008). (*) Without data.

Year PM10 PM10 without NAF PM2.5 Exceedances Exceed without NAF P90.4 P90.4 without NAF n PM10 n PM2.5

Agost 2005 29 29 21 5 5 45 45 84 18
2006 42 38 27 66 37 69 57 259 54
2007 40 37 26 52 21 64 60 211 83
2008 31 28 16 17 0 49 43 250 83
2005–2008 36 33 23
Sant Vicent 2005 37 36 21 19 19 61 61 99 22
2006 46 42 26 79 51 80 74 255 54
2007 42 36 25 55 29 73 60 219 57
2008 30 28 17 21 5 51 43 351 87
2005–2008 39 37 22
Alicante-El Plà 2005 37 37 19 38 34 58 58 261 46
2006 42 40 22 48 41 67 59 251 48
2007 40 36 23 73 23 60 54 230 43
2008 35 33 19 31 11 55 52 131 30
2005–2008 39 37 21
Alicante-University 2005 41 39 * 83 72 64 64 316 *
2006 39 36 * 66 55 63 58 360 *
2007 40 37 * 92 68 63 59 363 *
2008 30 28 * 34 24 49 47 360 *
2005–2008 38 35 *
Suburban sites 25–30 12–17
Urban sites 30–40 20–25
Agost 6 7
Sant Vicent 9 6
Alicante-El Plà 9 2
Alicante-University 8 –
M. Santacatalina et al. / Science of the Total Environment 408 (2010) 4999–5009
Fig. 2. PM10 daily levels at a) Agost and Sant Vicent; and b) Alicante-El Plà and Alicante-University during the study period.

5004 M. Santacatalina et al. / Science of the Total Environment 408 (2010) 4999–5009

even if the days with African dust contributions are excluded from this
count. The exception of this is 2005 at Agost and 2008 at Agost and
Sant Vicent and Alicante-University are closest to the cement plants,
whereas Agost is closer to the ceramic industries (Fig. 1). However, the
time evolution of the PM10 levels recorded at these three sites is very
similar. On some days, the PM10 levels registered at Agost were much
higher than those at Sant Vicent and Alicante-University, though. The
PM10 levels during these days (mostly from April 2006–May 2007) were
the highest in the period, surpassing 100 μg/m3 (Fig. 2). Highway
construction in the Agost neighborhoods (approximately equidistant
5 km from Agost and San Vicent) took place from 10 July 2006–8 April
2007. The fact that the highest PM levels were recorded simultaneously
along with the highway construction works points to this source as one
of the main contributors to the PM10 daily limit exceedances.
Thus, the lower concentrations recorded during 2008 are interpreted
to be mainly a consequence of the end of the highway construction
project. However, late spring rains in 2008 could have disfavored
resuspension (unpublished data), and industrial production in general
(INE, 2008), including that of the cement and ceramic industries, was
also reduced this year, which could also have influenced this reduction.
The annual PM2.5 levels varied between 16 and 27 μg/m3 across sites Fig. 3. Chemical composition of PM10 (μg/m3) at a) Sant Vicent; b) Agost; and
and years, hence in some cases above the limit for PM2.5 (25 μg/m3, c) Alicante-University during the period of study.
Directive 2008/50/CE). PM2.5 annual levels varied between 16–21 μg/m3
for 2005 and 2008, whereas levels for 2006–2007 were higher (22 to
27 μg/m3). The time variation in these levels is in agreement with that of Finally, marine aerosol accounts for 4% of the PM10 mass, with
PM10 levels. levels of 1.6 μg/m3 at Agost and Sant Vicent, and 2.0 μg/m3 at Alicante-
The PM2.5/PM10 ratios were calculated from the mean concentra- University.
tions of PM2.5 and PM10. Since the sampling of these size fractions was The unaccounted mass fraction could come from aerosol bound
not conducted simultaneously, a little bias may arise. The PM2.5/PM10 water, minor components and measurement errors (Vecchi et al.,
ratios were between 0.5 and 0.6, which fall in the lower range of the 2008; Almeida et al., 2006).
usual ratios measured at Spanish sites (0.5 and 0.7) (Querol et al., The comparison of the registered concentrations with the usual
2008). This fact suggests the relevance of coarse PM contributions in levels in urban non-industrialized areas of Spain (Querol et al., 2008)
the study area. shows that the PM10 mineral matter levels in the area are between 9–
In general, PM10 and PM2.5 levels in the study area can be considered 13 μg/m3 higher (Fig. 4), whereas the remaining components are
high, especially in 2006 and 2007. Comparing the PM10 levels in the area within the usual ranges in non-industrialized urban areas of Spain. In
with usual ranges in urban non-industrialized areas of Spain (Querol et the case of OM + EC, they are even below this range. This excess is
al., 2008), there was an excess of 6–9 μg/m3, which can be attributed to comparable to the 6–9 μg/m3 found for bulk PM10, which suggests
the specific sources of the area: highway construction works, cement that the high PM10 levels are due to the high mineral contribution.
plants, ceramic plants, soil resuspension in the arid area. Similarly, PM2.5 This hypothesis is backed up by the high correlations between mineral
levels exceeded the usual urban range by 2–7 μg/m3. These excesses in concentrations and PM10 levels (R2 = 0.7–0.9), compared to the
the PM10 and PM2.5 fractions occurred mainly in 2006 and 2007. relatively low correlations of bulk PM10 with levels of sulfate or OM +
EC (R2 = 0.01–0.3) (Fig. 5).

3.2. PM10 chemical composition 3.3. Trace elements

Major and minor components analyzed in the selected filters were Trace elements do not greatly contribute to total mass; however,
grouped as follows: mineral (sum of SiO2, Al2O3, CO2− 3 , Ca, K, Mg, Fe, Ti they are very relevant to specific trace sources (Querol et al., 2007a).
and P), organic matter and elemental carbon (OM + EC), secondary
− +
inorganic compounds (SIC: SO2− 4 + NO3 + NH4 ), sea spray and trace
elements and the average is represented in Fig. 3.
The major component of PM10 in the study area is mineral matter,
with contributions between 44, 52 and 47% of the total PM10 mass (16,
20 and 26 μg/m3) for Sant Vicent, Agost and Alicante-University,
respectively (Fig. 3). This indicates the influence of PM emissions of
mineral composition: highway construction, cement and ceramics
industries, PM resuspension emitted by transporting and processing
raw materials or soil resuspension given the semi-arid nature of the
area. The high Ca and Al2O3 levels are remarkable.
Organic matter and elemental carbon account for 12–14% of the
PM10 mass at all sites. The levels are relatively low because these sites
are not largely influenced by traffic compared to the influence of other
emission sources.
Secondary inorganic compounds (SIC) account for 16–21% of the Fig. 4. Major component concentrations in the study area compared to usual ranges in
mass of PM10. The levels of sulfate (4.1–4.9 μg/m3) are higher than those other urban non-industrialized areas of Spain.
of nitrate (2.3–3.6 μg/m3). Querol et al., 2008.
M. Santacatalina et al. / Science of the Total Environment 408 (2010) 4999–5009 5005

sectors. Comparing the Sr levels with other ceramic areas in Spain shows
that the values are higher in the study area than in the ceramic tile
manufacturing area, but similar to the levels in another brick
manufacturing area (Querol et al., 2007a,b, Minguillón et al., 2007b).
Remarkable are the high Tl concentrations at the three sites, coinciding
with the high levels recorded at other ceramic production areas (Querol
et al., 2007a,b). However, Cs levels were found slightly lower than in
other ceramic production areas.

3.4. Source apportionment

Six common factors were determined by PMF at Agost and Sant

Vicent. The explained variation (Paatero, 2004), the percentages of
each species represented by each source, is depicted in Fig. 7. The
main characteristics of the identified factors are the following:

The mineral factor explains most of the variance of the levels of Al,
Ca, K, Mg, Mn, Ti, Rb and Sr. This factor is attributed to the
contribution of several sources: the highway construction, ceramic
and cement emissions (diffuse and channeled), soil resuspension,
dust emissions from the processing of mineral raw materials
(quarries, among others).
The road traffic factor is characterized by a high load of carbonaceous
particulate matter from vehicle exhaust emissions; Cu, Zn, Sn and Sb
are products of tire and brake abrasion (Querol et al., 2007a; Thorpe
and Harrison, 2008). NO3− is a result of NOx oxidation, also emitted
with exhaust gases.
The secondary sulfate factor is characterized by SO2− +
4 and NH4 . It is a
consequence of the oxidation of SO2 to H2SO4 and later neutraliza-
tion with NH+ 4 or NH3. The origin of the gaseous precursors can be
emissions of fossil fuels with high sulfur contents, such as petroleum
coke or coal, or emissions coming from sulfide oxidation contained in
the ceramic argillaceous material. It can also have a non-regional
origin; due to the stability of ammonium sulfate, it can be
transported long distances (Leck and Persson, 1996). However, the
large levels of sulfate compared with other Mediterranean stations
(Querol et al., 2008) suggest a special relevance from local sources.
However the large levels of sulfates in the higher range of other
Mediterranean stations, 3–5 μg/m3, (Querol et al., 2008) suggest a
special relevance of local sources.
The petroleum coke factor derives from the primary emissions from
the combustion of this fuel in the study area and it is characterized by
V and Ni. These two elements are enriched in the heaviest residual
Fig. 5. Correlation between PM10 levels with the levels of crustal material, sulfate and
fraction of the petrochemical industrial process. At Sant Vicent, this
OM + EC during the period of study.
factor is also characterized by Tl and may represent stack kiln
emissions from cement and ceramic factories using this heavy oil
Levels of Pb, As, Cd and Ni are below the normative values of
fraction as a fuel.
Directives 2008/50/CE and 2004/107/CE.
The sea spray factor is a consequence of the transport of marine
The levels of trace elements in the study area were compared to
those registered in non-industrialized urban areas of Spain (Querol et aerosol with the breeze. It is characterized by Na, Na+, Cl−, Mg and
al., 2008) (Fig. 6). Some of the elements were found in higher Mg+ (ionic forms indicate the concentration of the water-soluble
concentrations, especially V, Ni, Sr, Tl and Cs, and to a lower degree Rb fraction of the considered element).
and Li. V and Ni are correlated (R2 N 0.7 at all sites) and are associated The factor characterized by Pb, Tl, K and K+ is attributed to the
to fuel use, in particular petroleum coke or fuel oil in the study area industrial influence, and in particular, to kiln ducted emissions. The
industries (cement and ceramics) (Querol et al., 2007a). The rest of presence of soluble K+ in this factor can indicate industrial use of
the elements can be associated to the mineral matter; for example, Sr biomass as co-fuel. Cement industry uses approximately 20% of
can replace Ca in minerals like calcite and gypsum, and Rb, Li and Cs residues derived fuels (Conesa et al., 2008).
can be constituents of aluminum silicate minerals (Goldschmidt,
1954), probably clay minerals in this area.
V and Ni levels are higher than those recorded at other ceramic or At both sites, the major contributor to PM10 levels is the mineral
cement production areas (Querol et al., 2007a,b; Schumacher et al., factor, with an average contribution of 15 μg/m3 (37%) at Agost and
2002), indicating that this is a special characteristic of the local 12 μg/m3 (32%) at Sant Vicent. Road traffic accounted for 15% of PM10 at
industries, which probably use fuel oil or petroleum coke in different Agost (6 μg/m3) and 23% at Sant Vicent (9 μg/m3). The contributions of
5006 M. Santacatalina et al. / Science of the Total Environment 408 (2010) 4999–5009

Fig. 6. Levels of trace elements in the study area compared to other ceramic and cement industrial areas and to non-industrial areas of Spain. (1) Schumacher et al., 2002; (2) Querol
et al., 2007a; (3) Querol et al., 2007b; and (4) Querol et al., 2008.

the remaining factors are similar at both sites: 13% (5 μg/m3) for the indicates that the channeled emissions from cement and ceramic kilns
sulfate factor, 7–8% (3 μg/m3) for the fuel factor, 6% (2.5 μg/m3) for the have a significant influence in mineral matter levels (black dots in
sea spray factor and 6% (2.2–2.6 μg/m3) for the industrial factor. Fig. 8). Nevertheless, for some high concentrations of Al2O3 or Ca,
Research has been focused in the identification of the sources of relatively low levels of V were found (white dots in Fig. 8), indicating the
the atmospheric particles and the results may contribute to the influence of the fugitive or cold-channeled emissions mentioned earlier.
interpretation of prior studies on source apportionment carried out in Two different types of samples were selected for SEM-EDX
Southern Europe, with very high loads of anthropogenic dust in PM10 analyses: samples with high V levels (above 80 ng/m3) and high PM
and PM2.5, where mineral contribution is behind large number of levels (50–80 μg/m3), along with samples with low V (below 30 ng/
exceedances in specific periods. A revision of PM temporal series and m3) and high PM (50–80 μg/m3 at Sant Vicent or 100–150 μg/m3 at
diffusive sources of PM is suggested. Agost). Six random areas were selected per sample and all particles
shown in each area were analyzed. The EDS generated spectra
3.5. PM sources for extreme episodes provided valuable information on the elemental composition of the
particles analyzed and have also permit to calculate Si/Ca ratios. Some
Samples with high levels of PM from Sant Vicent were selected for samples from the first group (30/03/2007 at Agost and 6/11/2006 at
a mineralogical study in order to discriminate the influence of specific Sant Vicent) showed individual particles with Si/Ca ratios similar to
sources of atmospheric pollution (especially cement plants). In most cement composition and also significant C and gypsum levels, thus
of the samples, calcite was a predominant mineral. The identification being identified as cement particles and indicating the influence of
of gypsum (especially in the samples from 14/11/06 and 19/12/06) cement emissions in ambient PM on these days. Cement particles have
can have its origin in emissions from gypsum-consuming industries or been previously identified by SEM-EDX in other works (Khwaja et al.,
fugitive soil emissions (from the highway construction or natural 2009). Some samples from the second group (20/03/2007 at Sant
erosion), since the area's soil contains clays and gypsum in addition to Vicent and 23/01/2007 at Agost) showed individual particles with Si/
calcite and dolomite. Similarly, the presence of clay, illite and kaolinite Ca ratios similar to those of calcite and dolomite (instead of cement
indicates the influence of the emissions of raw materials used in the ratios) and significant levels of aluminum and potassium (compo-
ceramic industries and/or the soil. nents of clay), indicating that these episodes of high PM levels were
A main result is that the occurrence of clinker crystalline phases mainly associated to fugitive mineral emissions and not to warm-
was not detected by XRD, suggesting that the influence of kiln cement channeled emissions.
emissions in ambient PM was not very significant on the selected days
with high PM levels. 4. Conclusions
The sources of PM during episodes of high PM levels were also
investigated based on chemical results. Since mineral matter has been The annual mean PM10 levels in the study zone in 2006 and 2007
identified as the major component of PM, and because this can come exceeded the usual concentration range for urban areas of Spain by 6–
from different sources (cement plants, ceramic plants, highway 9 μg/m3 (2–7 μg/m3 in the case of PM2.5). The main component of PM10
construction and soil erosion), further investigation was required in was mineral matter (44–52% of PM10). The annual mean levels of V, Ni,
order to distinguish between these different sources. Sr, Rb, Li, Tl and Cs were higher than in other non-industrial areas of
Channeled emissions from both cement and ceramic industries are Spain. V and Ni are attributed to the use of petroleum coke or residual oil
mainly of mineral composition, but they also have high levels of V, by the area industries (cement and ceramic industries). Furthermore, V
related with the fuel used. On the other hand, emissions from open air and Ni concentrations variations depend significantly on PM10 levels,
storage of raw materials (or from cold-channeled) in cement and indicating that the hot-channeled emissions from cement and ceramic
ceramic plants and emissions from related industries do not have any V kilns have a significant influence on mineral matter levels. The
in their composition. Hence, the concentrations of V, Al2O3, Ca and PM10 remaining elements (Tl, Sr, Rb, Li and Cs) are attributed mainly to a
were evaluated and a clear relation was found, i.e., V concentrations mineral origin, and some of them are in concentrations similar to those
increase when Ca, Al2O3 and/or PM10 do so, as shown in Fig. 8. This found in other cement and ceramic production areas.
M. Santacatalina et al. / Science of the Total Environment 408 (2010) 4999–5009 5007

Fig. 7. Explained variation of each species by the different identified factors.

Source contribution analysis suggested that the main sources XRD and SEM-EDX results. It is relevant that its impact was recorded
influencing ambient PM10 in the area are mineral sources (attributable at a regional scale, given the simultaneous recording of PM episodes at
to the highway construction, ceramic emissions, cement emissions, distant monitoring sites.
soil erosion and resuspension, quarrying, transport of raw materials, This study shows the high impact that anthropogenic fugitive
etc.), road traffic, secondary sulfate, petroleum coke combustion, sea emissions of mineral dust have on air quality in some areas of Europe.
spray, and an industrial factor. More attention should be paid to controlling these fugitive emissions for
Studying the highest PM episodes suggests that these events were fulfilling air quality standards in the area (L'Alacantí region); the results
mainly associated to fugitive emissions of mineral matter. These of this study conducted to the development of a management plan.
episodes mainly took place in 2006–2007 when the highway These results may contribute to the interpretation of the chemical
construction occurred. Although (fugitive and channeled) industrial composition of particulate matter in Southern Europe, with very high
emissions have a clear incidence on the area's air quality, emissions loads of mineral matter in PM10 and PM2.5 and they show that, even in
from the public construction works probably caused the pollution heavily industrialized areas, fugitive emissions from different sources
episodes. This hypothesis was supported by the chemical analysis, may account for the exceedances of PM limit values.
5008 M. Santacatalina et al. / Science of the Total Environment 408 (2010) 4999–5009

Fig. 8. Concentrations of V versus concentrations of Al2O3, Ca and PM10 at a) Sant Vicent; b) Agost; and c) Alicante-University during the study period. White dots indicate the days
with predominant PM sources different from warm-channeled emissions from cement plants.

Acknowledgements Bates TS, Quinn PK, Coffman DJ, Jonson JE, Middlebrook AM. Dominante of organic
aerosols in the marine boundary layer over the Gula of Maine during NEAQS 2002
and their role in aerosol light scattering. J Geophys Res 2005;110:D18202.
This study was financed by the Conselleria de Medi Ambient, Aigua, Conesa JA, Gálvez A, Mateos F, Martín-Gullón I, Font R. Organic and inorganic pollutants
Urbanisme i Habitatge of the Generalitat Valenciana (CMAAUH) and by from cement kiln stack feeding alternative fuels. J Hazard Mat 2008;158:585–92.
Directive 2004/107/EC of the European Parliament and of the Council of 15 December
the Vicerrectorado de Infraestructuras, Espacios y Medio Ambiente of 2004 relating to arsenic, cadmium, mercury, nickel and polycyclic aromatic
the University of Alicante. PM data from Alicante-El Plà, Sant Vicent and hydrocarbons in ambient air.
Agost was provided by the CMAAUH. We also thank the GRACCIE- Directive 2008/50/EC of the European Parliament and of the Council of 21 May 2008 on
ambient air quality and cleaner air for Europe.
CSD2007-00067 project, supported by the Spanish Ministry of Science Draxler RR, Rolph GD. HYSPLIT (HYbrid Single-Particle Lagrangian Integrated
and Innovation for assisting in this work. Trajectory) Model access via NOAA ARL READY Website. http://www.arl.noaa.
gov/ready/hysplit4.html 2003 NOAA Air Resources Laboratory, Silver Spring, MD.
Dulac F, Tanré D, Bergametti G, Buat-Ménard P, Desbois M, Sutton D. Assessment of
Appendix A. Supplementary data African airborne dust mass over the Western Mediterranean sea using meteosat
data. J Geophys Res 1992;97:2489–506.
Eatough DJ, Eatough DA, Lewis L, Lewis EA. Fine particulate chemical composition and
Supplementary data associated with this article can be found, in light extinction at Canyonlands National Park. J Geophys Res 1996;101:19515–31.
the online version, at doi:10.1016/j.scitotenv.2010.07.040. EPA (Environmental Protection Agency. 40 CFR Parts 53 and 58, revisions to ambient air
monitoring regulations, final rule. Retrieved November 20, 2006 from EPA Web
site: 2006.
References EPER
Escrig A, Monfort E, Celades I, Querol X, Amato F, Minguillón MC, Hopke PK. Application
Almeida SM, Pio CA, Freitas MC, Reis MA, Trancoso MA. Approaching PM2.5 and PM2.5- of optimally scaled Target Factor Analysis for assessing source contribution of
10 source apportionment by mass balance analysis, principal component analysis ambient PM10. Journal of Air and Waste Management Association 2009;59(11):
and particle size distribution. Sci Total Environ 2006;368:663–74. 1296–307.
M. Santacatalina et al. / Science of the Total Environment 408 (2010) 4999–5009 5009

Escudero M, Querol X, Pey J, Alastuey A, Pérez N, Ferreira F, et al. A methodology for the Monfort E, Celades I, Sanfelix V, Gomar S, López JL, Calpe V. Estimación de emisiones
quantification of the net African dust load in air quality monitoring networks. difusas de PM10 y rendimiento de MTD's en el sector cerámico. Bol Soc Esp Ceram
Atmos Environ 2007;41:5516–24. Vidr 2009;48(1):15–24 (in Spanish).
Goldschmidt VM. Geochemistry. Edited by Alex Muir: Oxford Clarendon Press; 1954. Paatero P, Tapper U. Positive matrix factorization: a non-negative factor model with
Grasshoff K, Ehrhardt M, Kremling K. Methods of seawater analysis. 3nd edn. optimal utilization of error estimates of data values. Environmetrics 1994;5:
Weinheim: Verl. Chem; 1999. 597 pp. 111–26.
INE. Instituto Nacional de Estadística. 2008. Paatero P. Least squares formulation of robust non-negative factor analysis. Chemom
IPTS. European Commission. Reference document on best available techniques on Intell Lab Syst 1997;37:23–35.
emissions from storage. July 2006. Final Draft. 2006. Paatero P. The multilinear engine — a table-driven least squares program for solving
IPTS. European Commission. Reference document on best available techniques in the multilinear problems, including the n-way parallel factor analysis model. J Comput
ceramic manufacturing industry. 2007 August 2007. Graphic Stat 1999;8:854–88.
IPCC. Climate change 2007: the physical science basis. Contribution of Working Group I Paatero P. User's Guide for positive matrix factorization programs PMF2 and PMF3, part
to the Fourth Assessment Report of the IPCC; 2007 (ISBN 978 0521 88009-1 2; 2004.
Hardback; 978 0521 70596-7 Paperback). Putaud JP, Van Dingenen R, Mangoni M, Virkkula A, Raes F, Maring H, et al. Chemical
IPTS. European Commission. Reference document on best available techniques in the mass closure and assessment of the origin of the submicron aerosol in the marine
cement, lime and magnesium oxide manufacturing industries. boundary layer and the free troposphere at Tenerife during ACE-2. Tellus 2000;52B:
2010 May 2010. 141–68.
Khwaja HA, Parekh PP, Khan AR, Hershey DL, Naqvi RR, Malik A, et al. An in-depth Querol X, Viana M, Alastuey A, Amato F, Moreno T, Catillo S, et al. Source origin of trace
characterization of urban aerosols using electron microscopy and energy dispersive elements in PM from regional background, urban and industrial sites of Spain.
X-ray analysis. Clean-Soil Air Water 2009;37(7):544–54. Atmos Environ 2007a;41:7219–31.
Leck C, Persson C. Seasonal and short-term variability in dimethyl sulfide, sulfur dioxide Querol X, Minguillón MC, Alastuey A, Monfort E, Mantilla E, Sanz MJ, et al. Impact of the
and biogenic sulfur and sea salt aerosol particles in the arctic marine boundary implementation of PM abatement technology on the ambient air levels of metals in
layer during summer and autumn. Tellus 1996;48B:172–299. a highly industrialised area. Atmos Environ 2007b;41:1026–40.
Millán M, Salvador R, Mantilla E, Kallos G. Photooxidant dynamics in the Mediterranean Querol X, Alastuey A, Moreno T, Viana MM, Castillo S, Pey J, et al. Spatial and temporal
basin in summer: results from European research projects. J Geophys Res 1997;102 variations in airborne particulate matter (PM10 and PM2.5) across Spain 1999–
(D7):8811–23. 2005. Atmos Environ 2008;42:3964–79.
Millán M, Mantilla E, Salvador P, Carratalá A, Sanz MJ, Alonso L, et al. Ozone cycles in the Reis MA, Oliveira OR, Alves LC, Rita EMC, Rodríguez F, Fialho P, et al. Comparison of
Western Mediterranean Basin: interpretation of monitoring data in complex continental Portugal and Azores islands aerosol during a Sahara dust storm. Nucl
terrain. J Appl Meteorol 2000;Vol 39(N° 4):487–508. Instrum Methods Phys Res B 2002;189:272–8.
Minguillón MC, Querol X, Alastuey A, Monfort E, Mantilla E, Sanz MJ, et al. PM10 Rodríguez S, Querol X, Alastuey A, Kallos G, Kakaliagou O. Saharan dust contributions to
speciation and determination of air quality target levels. A case study in a highly PM10 and TSP levels in Southern and E. Spain. Atmos Environ 2001;35:2433–47.
industrialized area of Spain. Sci Total Environ 2007a;372:382–96. Schumacher M, Bocio A, Agramunt MC, Domingo JL, de Kok HAM. PCDD/F and metal
Minguillón MC, Querol X, Alastuey A, Monfort E, Mantilla E, Miró JV. PM source concentrations in soil and herbage samples collected in the vicinity of a cement
contributions in a highly industrialised area in the process of implementing PM plant. Chemosphere 2002;48:209–17.
abatement technology. Quantification and evolution. J Environ Monit 2007b;9: Thorpe A, Harrison RM. Sources and properties of non-exhaust particulate matter from
1071–81. road traffic: a review. Sci Total Environ 2008;400:270–82.
Minguillón MC, Monfort E, Querol X, Alastuey A, Celades I, Miró JV. Effect of ceramic US, EPA. AP 42, fifth edition; volume i section 13.2: introduction to diffuse dust sources.
industrial particulate emission control on key components of ambient PM10. 1995.
J Environ Manage 2009;90(8):2558–67. Vecchi R, Chiari M, D'Alessandro A, Fermo P, Lucarelli F, Mazzei F, et al. A mass closure
Molinaroli E, Gerzoni S, Giacarlo R. Contribution of Saharan dust to the Central and PMF source apportionment study on the sub-micron sized aerosol fraction at
Mediterranean Basin. In: Johnson NJ, Basu A, editors. Processes controlling the urban sites in Italy. Atmos Environ 2008;42:2240–53.
composition of the clastic sediments, vol. 284. Geological Society of America Special WHO. Health aspects of air pollution with particulate matter, ozone and nitrogen
Paper; 1993. p. 303–12. dioxide. World Health Organization; 2003.