Artículo en PDF
How to cite
Complete issue
More information about this article
Journal's homepage in redalyc.org
Sistema de Información Científica
Red de Revistas Científicas de América Latina y el Caribe, España y Portugal
Rev. Int. Contam. Ambient. 13 (2), 81-85, 1997
COMPARATIVE STUDY OF ELEMENTAL CONTENTS
IN
ATMOSPHERIC AEROSOLS FROM THREE SITES
IN MEXICO CITY USING PME
Rosano PAREDES-GUTIÉRREZ, Alejandra LÓPEZ-SUÁREZ, Javier MIRANDA', Eduardo ANDRADE and J. Antonio
GONZÁLEZ
Instituto de Física, Universidad Nacional Autónoma de México, Ciudad Universitaria, Coyoacan 04510, D.F. México
(Recibido noviembre
1996,
aceptado diciembre
1997)
Keywords: atmosplieric aerosols, elemental contents, PIXE
ABSTRACT
A study of the elemental concentrations found in samples of atmospheric aerosols from three sites in the
Mexico City Metropolitan Area is presented. The samples were collected during the Winter and Spring
of 1994 using Stacking Filter Units (SFU), devised to separate particles with sizes between 2.5 pm and
15 pm (coarse fraction) and smaller than 2.5 pm (fine fraction). Elemental analyses were performed
using Particle induced X-ray Einission (PIXE), which allowed simultaneous detection of 14 elements
heavier than Al. By means of analysis of variance, comparisons among the elemental contents in the
three sites are giveii for both fractions, showing that, although variations are observed for some ele-
ments, in tlie fine fraction no defiiiite patteni is found witli respect to the site. On the other hand, in the
coarse fraction the soil-derived elemeiits Si, K, Ca, Ti, and Fe are the inost important contributors to
total mass.
RESUMEN
Se presenta un estudio de las coiiceiitraciones elementales encontradas en iiiuestras de aerosoles atmos-
féricos de tres sitios en la Zona Metropolitana de la Ciudad de México. Las muestras se colectaron
diirante e1 invierno y la primavera de 1994 utilizando Unidades de Filtros Apilados (SFU), disefiadas
para separar partículas con taiiiaíios entre 2.5 pin y 15 pm (~raccióii gruesa) y nienores que 2.5 pm
(fracción fina). Se efectuaron análisis elementales con la Eniisión de Rayos X Inducida por Partículas
Cargadas (PIXE), que pemiitió la detección siiiiultánea de 14 elementos iiiás pesados que Al. Por medio
de análisis de varianza se hacen las coniparacioiies entre los contenidos elementales de ambas fraccio-
iies para los tres sitios, iiiostraiido que, aunque se observan variaciones para algunos elementos, en la
fracción fina no hay un patrón definido en cuanto al sitio. Por otro lado, en la fracción gruesa se encuen-
tra que los elementos originados por el suelo Si, K, Ca, Ti y Fe son los contribuyentes más importantes
a la masa total.
INTRODUCTION
The atmosphenc pollution problein in tlie Mexico City Met-
ropolitan Area (MCMA) presents different aspects tliat need
to be tackled from equally distinct sides (Riveros 1994).
Among the major urban pollutants are atmospheric aerosols,
wliich are mostly responsible for the decrease in visibility,
altliough tliey also aEect tlie human respiratory tract, con-
tribute to acid rain, and cause damage to building (López-
Suárez 1996, Paredes-Gutitrrez 1996). Traditionally, Total
Suspended Particles (TSP) have been measured in many ur-
ban areas, altliougli the information from these studies is
strongly limited, as many of those particles can not enter the
human respiratory tract, and the scattenng of light is mostly
due to tlie fine fraction of tlie TSP. The Metropolitan Com-
mission for Prevention and Control of Pollution, through the
Automatic Network of Atmosplieric Monitonng (RAMA),
provides measurements of mass concentrations of particulate
'Author 10 whom al1 correspoiidence should be addressed
82
R. Paredes-Gutiemez el al.
matter with sizes below 10 pm (PM10). However, simple mass
concentration data does not give information about the pos-
sible toxicity or other effects of tliose particles, as no knowl-
edge of their composition exists. Therefore, furtlier studies
are required to find out what tlie aerosol composition is. In
this line, it is of interest to investigate the differences between
particles having different dimensions, because different sources
emit aerosols with various sizes. So, anthropogenic eleinents
typically have Mean Aerodynamic Diameters (MAD) below
2.5
pm,
while natural sources produce larger particles. There
is a need, thus, to determine whether any zones of the MCMA
are more aEected by sources of smaller particles, that is, an-
thropogenic, than others.
In order to determine elemental concentrations of a variety
of samples the use of Particle Induced X-ray Emission (PIXE)
is recoinmended (IAEA 1995a, Johansson and Campbell 1988)
as it presents advantages over other tecliniques tliat can be
used for aerosol analyses. It is a method based on charged
particle accelerators, tliat allows fast simultaneous analysis
of al1 elements with atomic number larger tlian 12, with a
good sensitivity and, in general, the sample is not destroyed.
PIXE does not offer information on chemical states or con-
tents of light elements, although it canbe complemented with
other techniques to achieve a more comprehensive view of
the aerosol cliaracteristics. Examples of them are ion chro-
matography (which gives information on compounds as ni-
trates and sulfates), proton elastic scattering analysis (for
measurement of hydrogen), and proton induced gainina-ray
einission (for detection of light elements, like carbon, nitro-
gen, oxygen or fluorine).
In the present work, a comparison of elemental contents in
samples of atmospheric aerosols collected during the first half
of 1994 from tliree sites in the MCMA is given. Elemental
concentrations were measured with PIXE, and examples of
further information that may be achieved after analyzing those
concentrations are shown. Results for both coarse and fine
aerosols are presented. Although several studies about elemen-
tal coinposition of atmosplieric aerosols in the MCMA have
already been publislied (Falcón 1988, Rosas 1995, Aldape
et
al.
1996, Miranda
et
al.
1996), they eitlier are more liinited
on tlie number of elements analyzed, on the nurnber of sites
wliere sampling occurred, or on the sampling periods.
TABLE
1.
SAMPLING SITES PARAMETERS
Characteristic
SW
W
NE
Latitude
19O18'N
19O24'N
19O33'N
bngitude
99OlO'W
99Ol6'W
99'2'W
Altitude
2300 mas1
2500 mas1
2300 mas1
Distance from
industry
2
km
5
km
2 km
Trafic density
Medium
bw
High
TY
pe
School
Residential
Residential
separates particles with MAD between 2.5 pm and 15 pm in
the coarse fraction, and particles with MAD below 2.5 pm in
the fine fraction. The coarse fraction is deposited onto 8 pm
pore size polycarbonate membrane filters (Nuclepore, Costar
Corp.), and the fine fraction is collected on 3 pm pore size
Teflon filters (Teflo, Gelman Sciences Inc.). Sampling was
done once a week, from 8:00 h to 14:OO h. simultaneously at
the three sites. The total mass of the deposited particles was
measured by pre- and post-weighing the filters with an Ohaus
200GD electrobalance (sensitivity 10 pg, resolution 10 pg).
Quality control procedures for sample preparation and lian-
dling are described elsewhere (Miranda
et
al.
1996). Air flow
in the SFU's was calibrated through an MKS 358C mass flow
meter (MKS Instruments, Andover, MA, USA).
PIXE analyses were carried out with the 5.5 MV Van de
Graaff accelerator at the Instituto de Física,
UNAM.
using a
2.2 MeV proton beam. The experimental setup is displayed
in figure 1. The resolution of tlie detection system is 180 eV
at 5.9 keV. Typical beam currents were 5 nA, and total inte-
grated charges were 1 pC. The detection system was cali-
brated by means of MicroMatter (Deer Harbor, Wasliington,
USA) thin film standards. A representative miniinum detec-
tion limits curve for the elemental contents deterniination in
the fine fraction is shown in figure 2, wliich presents data
only for elements with atomic number between 13 (Al) and
n
Van de Graaff
Accelerator
u
Proton beam
MATERIALS AND METHODS
Aerosols sarnples \vere collected ffom January 1994 to April
1994, in three different sites. Their characteristics are sliown
in
table 1.
The sites were chosen because tliey were expected
to be higlily polluted. Tlie Naucalpan site (W) is close to an
iinportant industrial zone; tlie Ecatepec site (NE) is close to a
major dust storm zone; wliile the Ciudad Universitaria site
(SW) should be strongly polluted due t0 particle transpOrt by
HG. l.
Experiniental setup for PIXE analyses of atmospheric aerosol saniples.
the wind. The sampling equipment was the Stacking Filter
The amplifier is an Ortec 672 and the multichannel analyzer is an
Unit (SFU) of the Davis design (Cahill et al. 1990), which
Ortec ACE add-on card attached to a personal computer
ELEMENTAL CONTENTS OF ATMOSPHERIC AEROSOLS FROM MEXICO CITY
83
1
10
15
20
25
30
35
z
IiIG
2. Sensitivity ciirve (pg of an element per m' of air) for a typical
PIXE analysis of an aerosol sample, using a 2.2 MeV proton bcaiii
and a 5 pC iiitegrated charge
35 (Br). Tliese eleiiients are analyzed witli tlieir K X-ray line.
Otlier liea~y
eleiiients, like Pb, are ineasured using tlieir L
lines, tlius corresponding to a different scnsitivity curve. How-
ever, tlie sensitivity for Pb is
-
45 pg/in3. Quality assurance
experiments for PIXE analyses were done previously with X-
ray Fluorescence (Miranda et al. 1994). X-ray spectra were
analyzed by ineans of tlie Quantitative X-ray Analysis Sys-
tein (QXAS) computer code (IAEA 1995b).
RESULTS AND DISCUSSION
About 1100 elemental concentrations were ineasured. Av-
erage eleinental contents in tlie fine fraction are displayed in
table
11,
wliile tliose for tlie coarse fraction are given in table
111.
Also, mass concentrations and the number of samples
are included for botli components.
Froiii tlie figures in table
11
it is possible to note th at co n -
centrations of the elements in tlie fine fraction are almost
constant in the three sites. In spite of sliowing different con-
centrations of some elements (S, Fe, Pb) within tlie uncer-
tainty. analysis of variance ata 95% confidente leve1 (Kreyszig
1985) of tlus sainple set does not sliow dflerences in the means
of tlie detected eleinents nor on tlie mass, thus supporting the
statement about the uniformity of aerosol composition, or otli-
envise pointing to a need for more samples. The results froin
tliis analysis of variance are given in table IV, botli for tlie
fine and the coarse fractions.
Regarding the coarse fraction, it is observed that the main
contribution comes from soil-derived elements (Si, K, Ca, Ti,
and Fe). Tliis Iias already been observed in previous size-
segregated studies in the MCMA (Miranda et
01.
1992, 1994,
1996). Again, analysis of variance shows no difference on
tlie eleiiiental concentrations, although it does indicate a lower
gravinietric inass concentration in tlie SW site (see table IV).
The explanation is probably the presence of organic com-
TABLE
11.
ELEMENTAL CONCENTRATIONS OF FlNE FRACTION
OF AEROSOL SAMPLES FROM THREE DIFFERENT
SlTES 1N 1994 (pg m-')
Element
SW
W
NE
Number of
samples
Mass
Al
Si
P
S
C I
K
Ca
Ti
v
Cr
Mn
Fe
Cu
Zn
Pb
O.12f 0.01
0.25f 0.05
0.10f 0.01
'ND
=
element was not detected
pounds, wliich cannot be measured witli PIXE. Also, S con-
tents in the coarse fraction is lower than in tlie fine one, as
anticipated from other studies. Rather surprising, though, is
the concentration of tliis element in the Westem site, wliich
was always below the sensitivity of the detection system. No
trace of Pb, typically anthropogenic, was measured in the
coarse fraction.
TABLE
111.
ELEMENTAL CONCENTRATIONS OF COARSE
FRACTION OF AEROSOL SAMPLES FROM THREE
DIFFERENT SlTES IN 1994
(pg
m")
Element
SW
W
NE
Number of
saniplcs
14
12
14
Mass
79f 10
l5Of 11
132f 17
Si
1
If 0 .8 0
17f 1.2
15f 1.1
S
1.lf 0.11
N R'
l.lf 0.10
C1
0.88I 0.06
0.089f 0.066
1.2f 0.08
K
0.91f 0.08
1.4f 0.14
1.4f 0.13
Ca
5.2f 0.29
6.5f 0.39
6.8f 0.37
Ti
O.3Of 0.03
0.3
1t 0.02
0.33f 0.02
V
0.042f 0.006
0.037f 0.006
0.033f 0.005
Cr
0.073f 0.01
0.068f 0.01
0.062+ 0.01
Mn
0.066f 0.01
0.0742 0.01
0.0755 0.01
Fe
3.lf 0.18
3.62 0.22
3.4f 0.19
Cu
ND"
0.010f 0.004
ND"
Zn
0.092f 0.02
0.12f 0.02
0.10f 0.02
'NR
=
element was not reliably measured
'ND
=
element was not detected
Remarkable differences are observed also between fine and
coarse components. First, Al, P, and Cu are more abundant in
the fine fraction, while C1, Cr, and Mn have Iiiglier concen-
trations in the coarse one. This is somewhat startling, as Cr
is normally among the anthropogenic elements, wliich are
most of the times present as fine aerosols. However, C1 and
Mn may be associated to soil derived elements. On the other
hand, concentrations of other elements, like V and Zn are
apparently uniform in both fractions. Again, higher values
R. Paredes-Gutiérrez
ef
al.
TABLE
1%'.
RATIO
va
O!
MEAN SQUARES FOR ANALYSIS OF
VARIANCE*
Element
Fine fraction
Coarse fraction
Mass
0.41
10.96
Si
O. 10
1.66
P
0.30
ND"
S
0.85
0.10
CI
0.15
2.81
K
0.26
2.89
Ca
1.48
1
.O7
Ti
0.14
O. 1 3
V
0.05
O. 1 9
Cr
0.54
0.60
Mn
1.58
0.32
Fe
2.75
0.32
Cu
1.45
0.86
Zn
0.90
0.72
Pb
1.14
ND
'The value
x
for which the F distribution function witli
(2,39)
degrees of
freedom has the value
0.95
is
x
=
3.24
(Kreyszig
1985).
If
va
Sx,
then the
mean concentrations are not different from each other.
"
ND
=
element not detected
were expected for these elements in tlie fine fraction.
Enrichment factors (EF) relative to Fe and to earth cmst
composition are also coinputed to estimate the possible origin
of the elements. Fe is chosen as reference element because it
is expected to have lower errors and uncertainties in the con-
centration measureinent, due to its higher X-ray energy (thus
reducing the uncertainty in the eficiency) and because there
are rather large amounts of tliis element in the samples. These
factors are obtained through the equation:
where CZ is tlie concentration of element Z, and tlie sub-
scripts S and EC refer to the sainple and to tlie average eartli
cmst composition, respectively. Figure 3 sliows tlie EF for
elements Si, K, Ca, and Ti. Values of elemental contents in
earth cmst were taken from the tables by Demayo (1984).
The EF for these particular elements is very close to one, as it
should be eqected from soil- derived elements. However, dif-
ferent behaviors are observed for K in the fine and coarse
fractions. EF for K is higher in the fine fraction, and this can
be explained by a contribution from smoke, as K is a typical
tracer of this aerosol source. Moreover, the high Zn contents
found in tlie coarse fraction are better understood wlien look-
ing at its EF, because their values for al1 three sites are around
25, wliile those for tlie fine fraction are between 250 and 300.
Based on the information given by the EF, i.e., that tlie
elements Si, K, Ca, Ti and Fe al1 have a common soil origin,
it is possible to compute a total soil contribution to the aero-
sol mass, following a tracer element method (Henry
et
al.
1984). Also, sulfate and non-soil K concentrations can be
-
----
-
OSW
NNE
-
;S
w
FIG. 3.
Enrichment factors for Si, K, Ca and Ti conipared to Fe, for (a)
coarse fraction and (b) fine fraction
evaluated. This is done by using the equations (Miranda
et
al.
1994):
Soil
=
2.20A1+ 2.49Si
+
1 .63Ca
+
1.97Ti
+
2.38Fe,
(2)
NSK
=
K
-
0.52Fe,
(3)
Sulfate
=
4.125S,
(4)
where Soil denotes tlie soil-derived aerosols concentration
(assuming that the elements are present as oxides and using
average earth cmst coinposition); NKS is tlie non-soil K con-
tribution (possibly from smoke); Sulfate is tlie concentration
of this compound, based on the assumption that S is present
as ammonium suifate, wliile Al, Si, S, K. Ca, Ti, and Fe
represent the measured concentration of the respective ele-
ment. Results of this calculation are presented in
table
V for
the fine fraction. Again, analysis ofvariance sliows no differ-
ence on the concentrations in the samples from the three sites.
Mass deficiency with respect to gravimetric mass after add-
ing al1 the contributions from PIXE analysis is explained by
other compounds not analyzed by this technique, like organic
matter and nitrates. The analysis of these species requires
complementary techniques, such as ion cliromatograpliy.
ELEMENTAL CONTENTS OF ATMOSPHERIC AEROSOLS FROM MEXICO CITY
TABLE V. SOIL, SULFATE AND NON-SOlL K CONCENTRATIONS IN
AEROSOL SAMPLES FROM THREE DIFFERENT SITES IN
1994
(
pg
m")
Element
SW
W
NE
Soil (fine)
13f 1
7f 0.5
8f 0.6
Soil
(coarse)
45f 1
63f3
56f 2
Sulfate (fine)
14f 1.2
llf 1.2
11f 1.4
NSK (fine)
10.27f 0.04
0.225 0.04
0.27f 0.04
CONCLUSIONS
This work represents the first attempt to establish an aerosol
monitoring network throughout the MCMA, with
multielemental and sensitive analytical capabilities. The
quantitative results regarding the concentrations did not al-
lowed to determine differences in the elements present in the
aerosols from the tliree sites, and only a gravimetric mass
difference was observed in the Southwestem site. There is
also a strong limitation in the number of samples considered
in this study. Other works are already in progress to support
or discard the uniformity hypothesis. Enrichment factors are
helpful in the identification of some elements as produced by
a specific source, as was the case of soil derived elements
;
moreover, the analytical methods are useful for the applica-
tion of simple tracer element methods for source apportion-
ment. Unfortunately, the lack of comparable information from
other urban areas (Miranda 1996) prevents the authors from
making any statement about the true leve1 of aerosol pollu-
tion in the MCMA.
ACKNOWLEDGEMENTS
Tlie autliors tliank the technical assistance of K. López,
J.C. Pineda, E. Pérez-Zavala, E. Santillana and M. Galindo.
They are also indebted with Dr. T.A. Cahill (UC-Davis) for
the loan of sampling equipment. This work was supported in
part by DGPA-UNAM under contract IN-100493.
REFERENCES
Aldape F., Flores M. J., García G. R. andNelson J.W. (1996).
PIXE analysis of atmospheric aerosols from a simulta-
neous three site sampling during the autumn of 1993 in
Mexico City. Nucl. Instr. and Meth.
B109/110,
502-505.
Cahill T. A,, Eldred R.A., Feeney P. J., Beveridge P. J. and
Wilkinson L.K. (1990). The stacking filter unit revis-
ited. Trans. AWMA 2 13-222.
Demayo A. (1984). Composition of average earth crust. In:
Handbook of Chemistty and Physics
(R.C. Weast, Ed.),
65th ed. CRC Press, Boca Raton, p. F-146.
Falcón Y.I., Ramírez C.R., Nuno L. and Sánchez M.L. (1989).
Mexico City aerosol study
:
PMlO sampling and analysis.
Proc. of the 82nd Ann. Meet. of the Air and Waste Man-
agement Assoc., Anaheim. 25-30 June 1989, pp. 89- 15 1.
AWMA, PA.
Henry R.C., Lewis C. W., Hopke P.K. and Williamson H. J.
(1984). A review of receptor model fundamentals. Atmos.
Environ.
18,
1507-1515.
IAEA (1995a). Final report of the advisory group meeting on
accelerator-based analytical techniques for characteriza-
tion and source identification of atmospheric aerosols. In-
ternational Atomic Energy Agenq, Vienna.
IAEA (1995b). Manual for QXAS. International Atomic En-
ergy Agency, Vienna.
Johansson S.A.E. and Campbell J.L. (1988).
PIXE: a novel
technique for elemental analysis.
Wiley, Chichester.
Kreyszig E. (1985).
Introductory niathematical statistics.
Wiley, New York, 505 p.
López-Suárez A. (1996). Analysis of atmospheric aerosols in
the southwest of Mexico City using nuclear analytical tech-
niques. Bachelor of Physics Thesis. Facultad de Ciencias,
UNAM, Mexico City. 92 p.
Miranda J., Morales J.R., Cahill T.A., Aldape F. and Flores
M. J. (1992). A study of elemental contents in atmospheric
aerosols in Mexico City. Atmósfera 5, 95-108.
Miranda J., Cahill T.A., Morales J.R., Aldape F., Flores
M. J. and Díaz R.V. (1994). Determination of elemen-
tal concentrations in atmospheric aerosols in Mexico
City using proton induced X-ray emission, proton elas-
tic scattering and laser absorption. Atmos. Environ.
28,
2299-2306.
Miranda J., Andrade E., López-Suárez A., Ledesma R., Cahill
T.A. and Wakabayashi P.A. (1996). A receptor model for
atmospheric aerosols from a southwestern site in Mexico
City. Atmos. Environ.
30,
3471-3479.
Miranda J. (1996). Aerosol studies in large urban areas: an
overview. Nucl Instr. and Meth.
R109/110,
450-456.
Paredes-Gutiérrez
R.
(1996). Comparative study of atmo-
spheric aerosols from three zones in the Mexico City Met-
ropolitan Area. Bachelor of Physics Thesis. Facultad de
Ciencias, UNAM, Mexico City. 63 p.
Riveros H.G. (1994). Contaminación ambiental. Ciencia y
Desarrollo
XH,
3 5-65.
Rosas I., Belmont R. and Jáuregui E. (1995). Seasonal varia-
tion of atmospheric lead levels in three sites in Mexico
City. Atmósfera
8,
157-168.
logo_pie_uaemex.mx