Trace Metals in PM10 and PM2.5 Samples Collected in a HighlyIndustrialized Chemical/Petrochemical Area and Its UrbanizedSurroundings

Silvia dos Anjos Paulino • Rafael Lopes Oliveira • Josiane Loyola •

Alan Silva Minho • Graciela Arbilla • Simone Lorena Quiterio •

Viviane Escaleira

Received: 24 August 2013 / Accepted: 29 January 2014

� Springer Science+Business Media New York 2014

Abstract The aim of this study was to determine the

potential impact of a highly industrialized area on its

urbanized surroundings. The area studied is home to a

refinery, a thermoelectric plant and several petrochemical

facilities industries. The concentrations of twelve elements

were determined in PM10 and PM2.5 samples collected

along a busy highway and near the petrochemical complex.

Significantly higher concentrations of Ca, Mg, Mn, Fe, Cu

and Al were observed in the petrochemical zone, but

principal component analysis revealed similar patterns for

both the highway site and a site approximately 1.5 km from

the petrochemical complex, suggesting that the main pol-

lution source in the area is vehicular flux. Higher concen-

trations in the industrial area may be attributed to intense

diesel-powered truck and bus traffic movement, mainly due

to the transport of supplies, fuel and gas. The observed

concentrations of the elements Cr, Co, Ni, Cd and Pb were

always lower than the detection limits of the technique

used.

Keywords Fine particulate matter � Trace metals �Vehicular emissions � Petrochemical complex

Petrochemical industries are considered to be an important

emission source of organic and inorganic pollutants (Nadal

et al. 2004, 2007). Consequently, populations living in the

vicinity of these industrial facilities can be subject to an

increased risk of cancer and other adverse health effects

(Lin et al. 2001). One of the largest petrochemical com-

plexes in Brazil is located in Rio de Janeiro. A large

petroleum refinery, a thermal power station and a number

of important chemical and petrochemical facilities are all

located in the zone. The region is subjected to high con-

centrations of particulate matter, volatile organic com-

pounds and ozone, and has adverse topographic and

meteorological conditions, which impede pollutant dis-

persion. Public concern about possible adverse health

effects for the population living near this industrial com-

plex has increased in the last decade. Accordingly, we

initiated a wide survey focused on determining the current

levels of various inorganic and organic pollutants in the

area and characterizing the main emission sources. The

results concerning the studied metal pollutants are pre-

sented below.

Materials and Methods

Samples were collected in the petrochemical complex of

Duque de Caxias, an urban-industrial area in the northern

region of the Rio de Janeiro Metropolitan Area (RJMA),

Brazil, which is strongly affected by vehicular and indus-

trial emissions. The petrochemical complex consists of a

refinery, a thermoelectric plant and several petrochemical

industries that manufacture and sell plastics, rubber, resins,

S. dos Anjos Paulino � R. L. Oliveira � J. Loyola �A. S. Minho � G. Arbilla (&)

Instituto de Quımica, Centro de Tecnologia, Universidade

Federal do Rio de Janeiro, Predio A, Sala 408, Cidade

Universitaria, Rio de Janeiro, RJ 21949-900, Brazil

e-mail: gracielaiq@gmail.com

S. L. Quiterio

Instituto Federal de Educacao, Ciencia e Tecnologia do Rio de

Janeiro, Campus Rio de Janeiro, Rua Senador Furtado, 121,

Maracana, Rio de Janeiro, RJ 20270-020, Brazil

V. Escaleira

Centro Nacional da Pesquisa do Solo, EMBRAPA, Rua Jardim

Botanico 1024, Rio de Janeiro, RJ 22460-000, Brazil

123

Bull Environ Contam Toxicol

DOI 10.1007/s00128-014-1219-4

solvents, fluids, oils, and, in particular, chemicals and

intermediates.

Samples were collected at three monitoring stations:

Federal Road Police (FRP), Cora Coralina State School

(CC) and Adelina de Castro State School (AC). The FRP

Station (latitude 22�400 and longitude 43�170) is located by

the Washington Luiz Highway and is approximately 5 km

away from the petrochemical complex. The Washington

Luiz Highway is part of the BR-040 road, which links the

cities of Rio de Janeiro and Petropolis. The vehicular flux

from this highway contributes roughly 3 % of the partic-

ulate matter and volatile organic compounds emitted from

mobile sources in the RJMA. The CC (latitude 22�420 and

longitude 43�180) is located in a neighborhood approxi-

mately 4 km from the petrochemical complex and 2.5 km

from the Washington Luiz Highway. The Adelina de

Castro State School Station (latitude 22�420 and longitude

43�160) is approximately 1.5 km from the petrochemical

complex. The movement of diesel-powered trucks and

buses, and light vehicles is very intense near this moni-

toring station. A map with the location of the petrochem-

ical complex and the sampling sites is shown in Fig. 1.

Data collected from the Aeronautical Meteorology

Services Network website (REDEMET 2010) during the

sampling period showed a predominance of weak winds in

the southeast and northwest directions (Fig. 1). A plot of

the wind frequencies according to direction showed that the

AC and CC stations were heavily affected by industrial

emissions, especially in the early hours of the day.

PM10 and PM2.5 were collected using high volume sam-

plers (AGV PM2.5, Energetica, Rio de Janeiro, RJ, Brazil)

and quartz fiber filters, with an area of 20 9 25 cm and a

thickness of 0.5 mm (Millipore Corporation, Billerica, MA,

USA). The flow rates were 1.1–1.7 and 1.05–1.21 m3 min-1

for the PM10 and PM2.5 samplers, respectively. Samplings

were performed at a height of 4.5 m over a period of 24 h, as

recommended by the (EPA 1999a).

The sampling periods and number of samples were

reported in Table 1. The samples were collected on

weekdays. Because only one PM10 and two PM2.5 samplers

were available, simultaneous sampling in all three loca-

tions was not possible.

Levels of PM10 and PM2.5 were determined by gra-

vimetry using an electronic microbalance with a sensitivity

Fig. 1 Location of the petrochemical complex and the sampling sites, and a wind rose plot, showing a predominance of stronger wind movement

in a northwesterly direction

Bull Environ Contam Toxicol

123

of 1 lg. For the analysis of trace metals, the same proce-

dure as that used in our previous studies was followed

(Toledo et al. 2008; Loyola et al. 2009). Briefly, after

extracting the metals, nitric and hydrochloric acid were

added, and the samples were analyzed using ICP-OES

(inductively coupled plasma-optical emission spectros-

copy) following Method IO-3.4 (EPA 1999b). All samples

and SRM were determined in triplicate, and a difference of

[1 % was considered acceptable. Detection and quantifi-

cation limits were computed as three and ten times,

respectively, the standard deviation of the distribution of

outputs for ten repeated measurements of the standard,

which contained no metals. These limits were calculated as

30 ng m-3 for Ca, 20 ng m-3 for Mg and Ni, 35 ng m-3

for Al, 3 ng m-3 for Cu, 40 ng m-3 for Fe and Pb,

1 ng m-3 for Mn, Zn and V, 7 ng m-3 for Cr and Co,

2 ng m-3 for Cd and 9 ng m-3 for Mo.

The accuracy of the method was evaluated using a

standard reference material (SRM, 2783 Air Particulate on

Filter Media, NIST, Gaithersburg, MD, USA). Three

samples of the reference material were determined in

triplicate, and the results were compared with the con-

centration reported in the certificate of analysis. The dif-

ference was [8 %.

Filter and reagent blanks were processed following the

same treatment. The metal content of the blanks for Cr was

[8 % of the samples’ average content. For the other

metals, the metal content was [5 % of the samples’

average content.

The experimental data were analyzed using STATIS-

TICA 7.0 (StatSoft Inc., Tulsa, OK, USA). The data were

analyzed using descriptive statistics before extended anal-

yses. After removing outliers, a multivariate statistical

analysis was performed that included cluster analysis (CA)

based on Euclidian distances and Ward’s method and

principal component analysis (PCA). The calculations (CA

and PC) were achieved using the individual experimental

data. Also Mann–Whitney U test and Kruskal–Wallis test,

for two and multiple independent samples, respectively,

were used to compare the results. These methods were

selected because the results for each sampling site are

independent but they are not normally distributed.

Results and Discussion

The mean values for PM10 and PM2.5 at the FRP station

were 42.2 and 21.2 lg m-3, respectively. At the AC sta-

tion, the mean values for PM10 and PM2.5 were 80.7 and

36.0 lg m-3, respectively. The mean value for PM2.5 at the

CC station was 45.8 lg m-3. Higher levels at the AC and

CC sites were expected, as many streets are unpaved, and

activities such as backyard burning and improper handling

of construction waste are frequently observed in those

areas. Values for the AC station are high compared to PM10

Brazilian air quality standards (50 lg m-3 annual mean

and 150 lg m-3 daily value). These values are also high

compared to those obtained by Godoy et al. (2009) for ten

different sites in the RJMA, which were not directly

located near heavy traffic avenues or the petrochemical

complex. It should be noted that quartz-fiber filters have a

large specific surface on which adsorption of gases can

occur leading to positive biased results. Also filter integrity

may be compromised by handling, which causes pieces of

quartz filters to be lost causing a negative bias. The extent

Table 1 Statistical summary of metal concentrations determined in

PM10 and PM2.5

Concentration (ng m-3)

Ca Mg Mn Fe Zn Cu Al

FRP PM10 (n = 34) 11/2008–03/2009

Meana 241 80 9 394 78 19 148

SD 132 45 5 237 134 9 98

Min 45 18 1 52 2 10 20

Max 534 276 22 968 631 56 452

AC PM10 (n = 34) 05/2009–11/2009

Meana 503 149 16 775 135 34 424

SD 310 61 8 438 108 21 252

Min 90 59 5 146 23 8 49

Max 1,253 296 36 1,983 488 94 1,004

FRP PM2.5 (n = 34) 11/2008–03/2009

Meanb 37 \DL 3 101 38 13 \DL

SD 24 \DL 1 90 63 6 \DL

Min 4 \DL 1 8 \DL 4 \DL

Max 95 \DL 6 324 313 29 \DL

AC PM2.5 (n = 34) 05/2009–11/2009

Meanb 112 28 5 125 75 24 66

SD 85 16 3 121 75 16 51

Min 36 13 1 8 9 7 19

Max 394 78 16 610 371 78 204

CC PM2.5 (n = 42) 05/2009–11/2009

Meanb 229 93 10 458 43 52 230

SD 145 38 3 173 26 19 92

Min 26 6 3 72 8 14 23

Max 585 192 18 1,065 112 97 434

Samples were collected at FRP, AC and CC, as detailed in the text.

Sampling period, number of samples, standard deviations (SDs),

minimum (Min) and maximum (Max) values are also shown.

n number of samples, DL Detection Limita For metals in PM10 samples (FRP and AC) all p values for Mann–

Whitney U test were lower than 0.00005b For metals in PM2.5 samples (FRP, AC and CC) all p values for

Kruskal–Wallis test were lower than 0.007. For metals in PM2.5

samples (AC and CC) p values for Mann–Whitney U test were lower

than 0.00001 for Ca, Mg, Mn, Fe, Cu, and Al and equal to 0.05 for Zn

Bull Environ Contam Toxicol

123

of the bias is also dependent on season and location and it

is difficult to estimate (EPA 2001).

Twelve metals were analyzed in the samples: Ca, Mg,

Mn, Fe, Zn, Cu, Co, Ni, Al, Cd, Cr and Pb. The mean

concentrations, standard deviations, and minimum and

maximum values are shown in Table 1.

The concentrations of the elements Cr, Co, Ni, Cd and

Pb were always lower than their detection limits. In PM10

samples, all elements showed significantly higher concen-

trations at the AC station. As expected, Ca, Fe and Al were

the most abundant metals in PM10. These elements are

major components of crustal materials and soil and are

predominantly a result of the resuspension of dust. In PM2.5

samples collected at the FRP station, the Mg and Al con-

centrations were consistently below their detection limits.

Fe and Ca were the most abundant elements. Ca, Mg, Mn,

Fe, Cu, and Al showed lower concentrations in FRP site.

Metal levels may be compared with the results obtained

at the entrance of the Andre Reboucas Tunnel, in the

southern part of the city (Loyola et al. 2012). In the PM10

samples, the Ca and Mg values were at least five times

higher at the entrance of the tunnel; the Mn, Fe and Al

levels were similar, and the Cu and Zn concentrations were

lower. In the PM2.5 samples collected at the FRP site, all

metals were present in lower concentrations except for Zn.

In contrast, the values at the CC station were higher than

those obtained at the entrance of the tunnel, primarily for

Ca, Mg, Zn and Cu. The higher concentrations of Zn may

be attributed to the higher contribution of diesel vehicles in

the area of the petrochemical complex (Loyola et al. 2009).

No differences in particulate matter levels and metal

concentrations during the sampling period were observed.

Mean temperatures were fairly constant with values in the

range 22 (winter) to 26�C (summer). Differences between

the wet (summer) and dry (winter) period were also not

observed but these may be due to the limited number of

samples.

Enrichment factors (EF) relative to the Earth’s crust,

using Mg and Ca as a reference, and the relative concen-

trations of each metal in both the Earth’s crust and the

samples were calculated (Webelements 2013; Fang et al.

2004; Kwangsam and Cocker 2009). As shown in Table 2,

for PM10, the average metal contents were 24.9 %, 40.7 %

and 15.3 % for Ca, Fe and Al, respectively, at the FRP

station. For the AC station, the average metal contents were

24.7 %, 38.0 % and 20.8 % for Ca, Fe and Al, respec-

tively. The presence of these metals is mainly due to nat-

ural sources and non-tailpipe emissions from vehicles. The

fact that they are present in higher ratios in the samples

than in typical crustal material may indicate that vehicles

also contribute to the emissions of these elements. Zn and

Cu are present in all of the samples in much higher ratios

than those found in crustal materials, indicating that these

elements, which are anthropogenic in origin, are enriched

in the soil (Figueira et al. 2002). The EF for Zn and Cu

were higher than ten in all samples, indicating the existence

of an anthropogenic input source. Fe and Cu are present in

high concentrations in car brake linings and car brake dust,

and Zn and Ca are usually associated with tire wear

(Thorpe and Harrison 2008).

For PM2.5, Ca, Mn, Fe, Zn and Cu are present in higher

ratios than in crustal material. Because the presence of

these particles is mainly due to combustion processes, these

metals may have a non-soil origin, and their primary

contributors may be vehicular traffic and the nearby

refinery, and petrochemical and industrial facilities. The Zn

and Cu ratios are higher near the highway (FRP station),

suggesting a significant contribution from vehicles (Godoy

et al. 2009; Santos et al. 2011).

FACTOR2 = 1.6393E-16-2.7859E-10*x

-2 -1 0 1 2 3 4 5

FACTOR1

-2

-1

0

1

2

3

4

5

FA

CT

OR

2

FRP PM 10 FRP PM 2.5 AC PM2.5 AC PM 10

Fig. 2 Score plot of the first two PCs of metal concentrations in

samples collected in the FRP and AC sampling sites. See text for

details

Table 2 Typical relative concentrations of each metal in the earth’s

crust and in the samples: FRP PM10, FRP PM2.5, AC PM10, AC PM2.5

and CC PM2.5

Metals In earth’s

crust

FRP

PM10

AC

PM10

FRP

PM2.5

AC

PM2.5

CC

PM2.5

Ca 11.8 24.9 24.7 19.2 25.7 20.5

Mg 7.6 8.3 7.3 \0.01 6.5 8.4

Mn 0.2 0.9 0.8 1.4 1.1 0.9

Fe 13.8 40.7 38.0 52.6 28.8 41.1

Zn \0.01 8.0 6.6 19.9 17.3 3.9

Cu \0.01 1.9 1.7 6.8 5.4 4.6

Al 19.9 15.3 20.8 \0.01 15.3 20.6

The values are listed as % of metal

Bull Environ Contam Toxicol

123

Principal component analysis was applied to the matrix

of the samples from the three sites. The first two PCs

encompassed 82.9 % of the variance in the original data

set. Factor 1 is mainly characterized by Ca, Mg, Mn, Fe

and Al (loading factors higher than 0.8), while factor 2

included Cu and factor 3 included Zn. When only data

from the FRP and AC stations are considered, the same PC

values are obtained, and the first two account for 86.6 % of

the variance. The first factor thus explains 72.6 % of the

total variance, and the second factor explains 14.0 % of the

remaining variance. A score plot obtained from the first

two PCs is shown in Fig. 2. High factor loadings for Cu, as

obtained for the PM10 samples, have been used to char-

acterize vehicle traffic (Castanho and Artaxo 2001).

When only PM2.5 values are considered, the first three

factors account for 86.1 % of the variance. Factor 1

included Fe and Al, factor 2 included Zn, and factor 3

included Cu. A plot of factor 3 versus factor 1 is shown in

Fig. 3. The results suggest that the main contributions at

the CC site and the other two sites are different, revealing

the important contribution of Cu at the CC site. A similar

result is obtained when factor 2 is plotted versus factor 1.

The data for FRP, which is mainly impacted by vehic-

ular emissions, and AC, which is mainly impacted by

activities in the industrial area, do not differ significantly,

suggesting that the main pollution source in these areas is

vehicular flux. Higher concentrations in the AC area may

be attributed to the intense diesel-powered truck and bus

traffic movement in the vicinity of the industrial area due to

the transport of supplies, fuel and gas. The CC area appears

to be affected by other metal inputs. The area is home to

several non-regulated activities such as waste burning,

battery reforming and metal recovery.

Atmospheric trajectory models show that changes in

wind directions occur due to the area’s geographical

location on a coastal plain at the base of a mountain range.

Frequently, a sea breeze predominates from the south-

southeast direction during the day and polluted air emitted

by the vehicular and industrial facilities is carried into the

continent. At night, air flows from the continent to the

ocean (INEA 2009). The transport of air from the coast

area, where the petrochemical complex is located, into the

continent may also cause the mixing of air masses,

resulting in similar distributions between all the sites.

Trace metals, in PM10 and PM2.5, collected in a location

characterized by high vehicular flux and significant

industrial and petrochemical activities, were determined.

Despite the high magnitude of industrial activities in the

area studied, the current results suggest that the refinery

and industries are not relevant metal pollution sources. The

presence of metals in the particulate matter appears to be

largely caused by highway transport emissions and trans-

port sources within the industrial park rather than industrial

sources.

The present results suggest that for further monitoring,

the use of a more sensitive analytical technique (like ICP-

MS) should be employed to determine the levels of toxic

elements, including As, Cd, Hg and Pb. The use of such an

analytical technique will help ensure that industrial activ-

ities in the area do not represent a health risk to the

population.

Acknowledgments The study was funded in part by FAPERJ,

CNPq and CAPES. We would like to thank CENPES and ASSE-

CAMPE for providing the facilities to perform sampling at the

monitoring stations.

References

Castanho ADA, Artaxo P (2001) Wintertime and summertime Sao

Paulo aerosol source apportionment study. Atmos Environ

35:4889–4902

EPA (1999a) Method IO-2.1. Compendium of methods for the

determination of inorganic compounds in ambient air. EPA/625/

R-96/010a. Environmental Protection Agency, Cincinnati, OH

45268

EPA (1999b) Method IO-3.4. Determination of metals in ambient

particulate matter using inductively coupled plasma (ICP)

spectroscopy. EPA/625/R-96/010a. Environmental Protection

Agency, Cincinnati, OH 45268

EPA (2001) Air quality criteria for particulate matter. EPA 600/P-99/

002aB. Environmental Protection Agency, Cincinnati, OH 45268

Fang GC, Wu YS, Huang SH, Rau JY (2004) Dry deposition

(downward, upward) concentration study of particulates and

heavy metals during daytime, nighttime period at the traffic

sampling site of Sha-Lu. Taiwan. Chemosphere 56:509–518

FACTOR3 = -8.0665E-17+6.5632E-10*x

-2 -1 0 1 2 3 4 5

FACTOR1

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

FA

CT

OR

3

CC PM2.5 FRP PM 2.5 AC PM 2.5

Fig. 3 Score plot of factor 3 versus factor 1 for the PCA analysis of

metal concentrations in the PM2.5 samples collected at the three

sampling sites. See text for details

Bull Environ Contam Toxicol

123

Figueira R, Sergio C, Souza AJ (2002) Distribution of trace metals in

moss biomonitors and assessment of contamination sources in

Portugal. Environ Pollut 118:153–163

Godoy MLDP, Godoy JM, Roldao LA, Soluri DS, Donagemma RA

(2009) Coarse and fine aerosol source apportionment in Rio de

Janeiro, Brazil. Atmos Environ 43:2366–2374

INEA (2009) State Environmental Institute reports. http://www.inea.

rj.gov.br/downloads/relatorios/qualidade_ar_2009.pdf. Accessed

1 Dec 2012

Kwangsam N, Cocker DR III (2009) Characterization and source

identification of trace elements in PM2.5 from Mira Loma. S.C.

Atmos Res 93:793–800

Lin MC, Yu HS, Tsai SS, Cheng BH, Hsu TY, Wu TN, Yang CY

(2001) Adverse pregnancy outcome in a petrochemical polluted

area in Taiwan. J Toxicol Environ Health 63:565–574

Loyola J, Arbilla G, Quiterio SL, Escaleira V, Bellido AV (2009)

Concentration of airbone trace metals in a bus station with a high

heavy-duty diesel fraction. J Braz Chem Soc 20:1343–1350

Loyola J, Arbilla G, Quiterio SL, Escaleira V, Minho AS (2012)

Trace metals in the urban aerosols of Rio de Janeiro city. J Braz

Chem Soc 23:628–638

Nadal M, Schuhmacher M, Domingo JL (2004) Metal pollution of

soils and vegetation in an area with petrochemical industry. Sci

Total Environ 321:59–69

Nadal M, Schuhmacher M, Domingo JJ (2007) Levels of metals,

PCBs, PCNs and PAHs in soils of a highly industrialized

chemical/petrochemical area: temporal trend. Chemosphere

66:267–276

REDEMET (2010) http://www.redemet.aer.mil.br/. Accessed 2010

Santos DSS, Korn MGA, Guida MAB, Santos GL, Lemos VA,

Texeira LSG (2011) Determination of copper, iron, lead and zinc

in gasoline by sequential multi-element flame atomic absorption

spectrometry after solid phase extraction. J Braz Chem Soc

22:552–557

Thorpe A, Harrison RM (2008) Sources and properties of non-exhaust

particulate matter from road traffic: a review. Sci Total Environ

400:270–282

Toledo VE, Almeida PB Jr, Quiterio SL, Arbilla G, Moreira A,

Escaleira V, Moreira JC (2008) Evaluation of levels, sources and

distribution of toxic elements in PM10 in a suburban industrial

region, Rio de Janeiro, Brazil. Environ Monit Assess 139:49–59

Webelements (2013) http://www.webelements.com/geology.html.

Accessed 1 June 2013

Bull Environ Contam Toxicol

123