You are on page 1of 10

ARTICLE IN PRESS

Atmospheric Environment 38 (2004) 1631–1640

Emission of polycyclic aromatic hydrocarbons from light-duty


diesel vehicles exhaust
Rui de Abrantesa,*, Jo*ao V. de Assun@*aob, Ce! lia R. Pesquerob
a
* Paulo, Brazil
Vehicular Emission Laboratory, Cetesb, Sao
b
* Paulo, Sao
School of Public Health, University of Sao * Paulo, Brazil
Received 24 July 2003; received in revised form 24 November 2003; accepted 24 November 2003

Abstract

Standardised tests were performed on four light-duty diesel vehicles running in a chassis dynamometer at a vehicular
emission laboratory, using the FTP-75 test cycle procedure. The aim was to characterise emissions of Polycyclic
Aromatic Hydrocarbons (PAHs), substances that create health hazards and are, as yet, unregulated. The pollutants
were analysed in both solid and gaseous phases using high-performance liquid chromatography. Total PAH values
ranged from 1.133 to 5.801 mg km1. Naphthalene, phenanthrene, fluoranthene, pyrene and chrysene were detected in
all tests. In addition, PAH emission was observed to be inversely related to emission of CO2.
r 2003 Elsevier Ltd. All rights reserved.

Keywords: PAHs; Aromatic pollutants; Vehicular emission; Air pollution; Toxic pollutants

1. Introduction emissions of hydrocarbons and 77% of particulate


matter emissions, corresponding to 70,600 and
All over the world, the numbers of motor vehicles 19,700 tons, respectively (CETESB, 2002).
have been progressively increasing, constituting a global Over the last 10 years, use of school vans, public
threat due to the thousands of tons of pollutants they transportation and small delivery services in the
spew into the atmosphere daily. In addition to worsen- Metropolitan Area of S*ao Paulo have increased. As a
ing global problems such as the greenhouse effect, these consequence, the fleet of light-duty commercial diesel
pollutants have created conditions for the occurrence of vehicles1 has shown an impressive growth rate of 7.39%
critical air pollution episodes. These episodes occur per year, considerably higher than the global average of
mainly in the larger urban centres, affecting the quality 2.6% over the same period (ANFAVEA, 2003).
of life and compromising public health, increasing This rapid growth rate, compared with recent
morbidity and mortality rates, mainly in children and epidemiological studies, creates new public health
seniors (Mage and Zali, 1992; Sher, 1998; Saldiva et al., concerns. There is currently no specific regulation of
2001; WHO, 1996a). the emission of Polycyclic Aromatic Hydrocarbons
In 2001, the Metropolitan Area of the city of S*ao (PAHs) from motor vehicles (WHO, 1998) despite the
Paulo, in the state of S*ao Paulo (Brazil), had a fleet of fact that emission of total hydrocarbons and particulate
seven million motor vehicles. Diesel vehicles represented matter are already regulated in Brazil and in most
6% of this total. However, in addition to contributing to industrialised countries (IBAMA, 1998).
elevated concentrations of other pollutants, emissions
from diesel vehicles constituted 18.4% of total vehicular
1
Light commercial vehicles in Brazil are defined as vehicles
with a maximum total mass of up to 3856 kg, designed to
*Corresponding author. Fax: +55-11-3030-6695. transport cargo, travel off-road or carry more than 12
E-mail address: ruia@cetesb.sp.gov.br (R. de Abrantes). passengers (Ibama, 1998).

1352-2310/$ - see front matter r 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.atmosenv.2003.11.012
ARTICLE IN PRESS
1632 R. de Abrantes et al. / Atmospheric Environment 38 (2004) 1631–1640

Consisting of hundreds of compounds that contain benzo[a]pyrene over a lifetime, 87 persons will die from
two or more aromatic rings (benzene rings); PAHs in the cancer related to that exposure.
air are more commonly in a solid, rather than a gaseous Considering the importance of these compounds in
phase. Some PAHs present carcinogenic potential, relation to human health, the aim of this work was to
capacity to accumulate in fatty tissue and in other characterise and quantify the more toxic PAHs in the
organic materials, such as particulate matter emitted by exhaust of some light-duty diesel vehicles currently in
vehicles. These attributes complicate their removal from use, which were not equipped with catalytic converters.
the human body (WHO, 1998).
Most of the various PAHs have been analysed by the
International Agency for Research on Cancer. Ben- 2. Materials and methods
zo[a]anthracene, benzo[a]pyrene and dibenzo[a,h]an-
thracene were classified as probably carcinogenic to 2.1. Sampling
humans. Naphthalene, benzo[b]fluoranthene and ben-
zo[k]fluoranthene were classified as possibly carcino- Between June of 2001 and January of 2002, four light-
genic to humans. Researchers were unable to classify duty commercial diesel vehicles, three vans and one
fluorene, phenanthrene, anthracene, fluoranthene, pyr- pick-up, were submitted to standardised tests using
ene, chrysene or benzo[ghi]perylene as to their carcino- a model ECE-50 light-duty hydraulic dynamometer
genicity to humans. Acenaphthene has not been (Clayton) at the Vehicular Emission Laboratory of
analysed (IARC, 2003). Cetesb, the Environmental Sanitation Agency of the
Measurements of PAHs in the atmosphere of the city state of S*ao Paulo. Considering the number of diesel
of S*ao Paulo taken by Martinis et al. (2002) showed vehicles in Metropolitan Area of the city of S*ao Paulo,
concentrations of 6400 pg m3 of benzo[a]anthracene the number of four vehicles represents an error of 40%
and 2700 pg m3 of benzo[a]pyrene, while Vasconcellos with a confidence interval of 64%, so these results must
(1996) found 1950 pg m3 of benzo[a]pyrene and be considered preliminary (Provost, 1984). The char-
1031 pg m3 of benzo[a]anthracene and Mendon@a acteristics of the vehicles tested are shown in Table 1.
(1991) found 3600 pg m3 of dibenzo[a,h]anthracene. In The vehicles were tested following the cold start
addition, in the atmosphere of the city of Rio de Janeiro United States Federal Test Procedure 75 (FTP-75), this
(Brazil), Azevedo et al. (1999) found 10670 pg m3 of standardised conduction cycle is divided in three stages
benzo[a]pyrene and 2110 pg m3 of benzo[a]anthracene. and simulates urban driving conditions.
These values are higher than those found in rural In accordance with CFR 40 recommendations
areas, where values typically ranged from 100 (United States, 1997), exhaust gases were collected after
to 1000 pg m3 (WHO, 1998). In addition, taking being mixed and homogenised with the ambient air. A
benzo[a]pyrene as an example, these values were higher Critical Flow Venturi-Constant Volume Sampler (Hor-
than those found in other urban atmospheres, such as iba CFV-CVS model 20A) system and a full-flow
those of Los Angeles, USA (590 pg m3), La Spezia, dilution tunnel were used. The dilution tunnel consisted
Italy (2200 pg m3), Julich, Germany (1100 pg m3) and of a stainless-steel duct, whose diameter promoted
Leeds, England (1040 pg m3) (Martinis et al., 2002). turbulent flow conditions in order to homogenise the
The WHO (1996b) has adopted the concept of the mixture of emission and ambient air.
Unit Risk Factor, which is used to calculate the risk of In addition, an ambient air-mixing device that filters
death from cancer after a lifetime of exposure to a the dilution air through three filters (two for particles
certain concentration level of a particular carcinogen in and one for hydrocarbon compounds) was used. In
the air. For instance, in a population of 1 million order to avoid water condensation in the sampling
inhabitants exposed to a cncentration of 1000 pg m3 of system, the device was placed as close as possible to the

Table 1
Vehicles characteristics

Vehicle Type Vehicles characteristics

Manufacturer/model Year Accumulated Cylinder Power


mileage (km) capacity (l) (kW)

A Pick-up Toyota/Hilux 4  4 CS 1995 31,182 2.8 64


Ba Van Asia/Topic 1995 2831 2.7 55
C Van Mercedes Benz/Sprinter F 310 D 1997 22,888 3.1 85
D Van Mercedes Benz/Sprinter 1998 43,697 3.1 85
a
After engine had been rectified.
ARTICLE IN PRESS
R. de Abrantes et al. / Atmospheric Environment 38 (2004) 1631–1640 1633

Fig. 1. Scheme of PAH sampling.

vehicle. The sampling point was positioned at the end of The greater part of the mixture, after being filtered,
the dilution tunnel, in the centre, facing upstream, using was stored in inert Tedlars plastic bags for subsequent
isocratic-sampling velocity. determination of the dilution ratio through quantifica-
The system collection scheme is shown in Fig. 1. For tion of CO2 emissions. Ambient air was also sampled
each stage of FTP-75, a glass fibre filter Gelman, and, in order to obtain an accurate assessment of the
porosity 1 mm and 47 mm diameter, was used for emission of the vehicles, the concentration of each
the collection of particulate matter. The maximum face compound was discounted. The ambient air sampling
velocity (20 cm s1) was adopted, in order to have point was placed just behind the ambient air filters in the
sufficient concentrations for each compound of interest, mixing device. Immediately after the end of each test,
according method J 1936 (SAE, 1998). Therefore, a the ambient air was sampled at a flow rate of 2 l min1
sampling flow rate of 10 l min1 was used, so as to for a length of time equal to or greater than that of the
achieve a collected sample volume of approximately full test (32 min).
108 l in stages one and three and 295 l in stage two. Two tests were performed on each vehicle, at intervals
The particulate matter mass in the filters was ranging from 12 to 36 h. Vehicle fuel consumption was
determined gravimetrically by the difference in mass also measured by the fuel-weighing method, in accor-
before and after each test. Before weighing, the filters dance with the Brazilian norm NBR 7024 (ABNT,
were pre-conditioned in a vacuum desiccator with silica 1990). In order to characterise PAHs emitted into the
gel for at least 16 h to eliminate humidity. The filters atmosphere in real situations, all tests were carried out
were weighed on an Ohaus analytical electronic balance, with commercial automotive diesel fuel, rather than with
model AP250D. After the second weighing, the filters diesel fuel designed for standardised laboratory tests.
were placed in Petri dishes, wrapped in aluminium foil to
avoid sample degradation from UV light and stored at 2.2. Analysis
17 C until time of extraction.
Gaseous-phase PAHs were collected with pre-washed Extraction and analysis were performed according to
XAD-2 Supelpak 20U adsorbent resins (100/50 mg- the recommendations of the method TO-13 (United
Supelco) and the sampling flow rate was 2 l min1, as States, 1988). Prior to extraction, samples were spiked
recommended by the manufacturer. Behind the filter with a known mass of surrogate2 standard Decafluor-
holders, where the sampling lines fork, two calibrated obiphenyl (Aldrich Chemical Co., purity X99.5%, CAS
dry-gas meters (DGMs) registered the volume sampled 434-90-2). Each sample, including filters and XAD-2
in each branch. After sample collection, the sorbent sorbent, was separately extracted by soxhlet over 8 h, at
tubes were sealed, wrapped in aluminium foil and stored
at 17 C until time of extraction. The samples were 2
Surrogate standard—a solution containing a compound
never stored for more than 10 days. that it is not expected to be found in the sample.
ARTICLE IN PRESS
1634 R. de Abrantes et al. / Atmospheric Environment 38 (2004) 1631–1640

10–15 min reflux time, with 120 ml methylene chloride concentrations of PAHs were not corrected for surro-
(J.T. Baker, HPLC grade, purity >99.8%, CAS 75-09-2). gate recoveries.
During extraction, the system was wrapped in alumi- The analytical detection limit for each PAH was
nium foil to protect the samples against UV light. Blank considered to be three times the noise level. The linearity
filters and sorbents were also extracted for analysis. interval of each PAH in the fluorescence detector was
All extracts were concentrated to 1 ml using a assessed through analysis of nine progressively more
Kuderna–Danish evaporator containing two boiling dilute standards, using three injections of each one. By
chips and 3-ball macro Snyder Column and then linear regression, the response of each PAH was
increased to a volume of 5 ml with acetonitrile (Carlo determined, using an r2 higher than or equal to 0.99 as
Erba Reagent, HPLC grade, purity >99.9%, CAS the criterion. Results are shown in Table 2. To check
75-05-8). instrument performance, a standard solution, with a
The analysis was performed by reversed phase high- concentration greater than 10 times the detection limit
performance liquid chromatography (HPLC). The for each PAH, was prepared daily. Analysis of blanks,
HPLC system was equipped with two model PU-980 including filters and XAD-2 tubes, showed no significant
Jasco pumps, a model LC-1250 GBC fluorescence contamination.
spectrophotometer and a Waters temperature controller
to maintain the column temperature at 25 C. Separa-
tions were achieved using a reversed-phase Supelcosil 3. Results and discussion
LC-PAH, type O.D.S. - C 18 (15 cm  4.6 ID  5 mm)
and a Supelcosil guard column (Supelco). A mobile- Table 3 summarises the results. The total of the 14
phase solvent gradient of acetonitrile and deionised PAHs analysed ranged from 1.133 to 5.801 mg km1,
water (50:50) was held isocratically for 5 min, changed with an average of 2.81571.846 mg km1. The emission
linearly to 100% acetonitrile over 40 min and held in of PAHs in relation to fuel consumption showed that the
100% acetonitrile for 5 min. In all analyses, the flow rate simple mathematical average ranged from 8.948 to
was 1 ml min1. Excitations and emission wavelengths 55.239 mg l1, with an average of 24.102717.260 mg of
that would improve sensitivity and selectivity detection PAHs per litre of diesel fuel.
for each PAH were selected. The analytical parameters Furthermore, with reference to the particulate matter,
utilised in HPLC analyses are shown in Table 2. values from 0.915 to 3.580 mg of PAHs adsorbed per
Fourteen PAHs were quantified, namely: naphtha- gram of particulate matter were found. The data shown
lene, acenaphthene, fluorene, phenanthrene, anthracene, in Table 3 reveal that concentrations of particulate
fluoranthene, pyrene, benzo[a]anthracene, chrysene, matter mass emitted during the second test of vehicle C
benzo[b]fluoranthene, benzo[k]fluoranthene, benzo[a] and during the first test of vehicle D were very low.
pyrene, dibenzo[a,h]anthracene and benzo[ghi]perylene These results should be considered with caution since
(Supelco standards, purity from 93.4% to 99.9%). they have an outlying behaviour, affecting the sum
Analysis of extracts was performed in duplicate and values of the PAHs in relation to particulate matter and

Table 2
Analytical parameters utilised in HPLC for the analyses of PAHs

Analytical parameters in HPLC

PAHs Excitation (nm) Emission (nm) Linearity interval (mg l1) Detection limit (mg l1)

Naphthalene 224 330 10.6–531.0 5.31


Acenaphthene 224 330 1.0–50.0 0.50
Fluorene 224 330 24.8–1242.0 12.42
Phenanthrene 252 370 3.0–302.0 1.51
Anthracene 252 402 1.6–82.0 0.82
Fluoranthene 252 402 34.5–345.0 6.90
Pyrene 238 398 15.2–152.0 3.04
Benzo[a]anthracene 252 402 55.0–550.0 11.00
Chrysene 252 402 5.3–265.0 2.65
Benzo[b]fluoranthene 268 398 23.4–233.5 4.67
Benzo[k]fluoranthene 268 398 5.5–277.0 2.77
Benzo[a]pyrene 268 398 5.8–288.5 2.89
Dibenzo[a,h]anthracene 268 398 6.4–319.0 3.19
Benzo[ghi]perylene 234 420 7.5–377.0 3.77
ARTICLE IN PRESS
R. de Abrantes et al. / Atmospheric Environment 38 (2004) 1631–1640 1635

Table 3
Fuel autonomy, CO2, particulate matter and PAHs emission

Emission rates

Vehicles Test Fuel CO2 emission Emission of Emission of Emission of Amount of


autonomy (g km1)a particulate PAHs PAHs relative PAHs
(km l1)a matter (mg km1)a to fuel adsorved in the
(g km1)a consumption particulate
(mg l1) matter
(mg g1)

Vehicle A 1st Test 9.618 309.9 1.352 5.801 55.239 2.085


2nd Test 9.011 306.5 1.551 5.345 45.888 2.374
Average 9.315 308.2 1.452 5.573 50.564 2.230

Vehicle B 1st Test 6.688 438.2 0.377 1.342 12.828 1.838


2nd Test 6.745 456.5 0.301 1.268 11.936 0.915
Average 6.717 447.4 0.339 1.305 12.382 1.376

Vehicle C 1st Test 7.732 420.4 0.141 1.133 8.948 3.361


2nd Test 7.795 422.5 0.056b 3.112 25.132 12.065b
Average 7.764 421.5 0.141 2.123 17.040 3.361

Vehicle D 1st Test 7.096 427.8 0.025b 2.728 19.366 19.248b


2nd Test 7.127 429.3 0.111 1.793 13.476 3.580
Average 7.112 428.6 0.111 2.261 16.421 3.580

All vehicles Average 7.727 401.4 0.639 2.815 24.102 2.359


Standard 1.071 58.6 0.640 1.846 17.260 0.993
deviation
a
These results are weighted average of three steps of FTP-75, according CFR 40 (United States, 1997).
b
Values considered outlyers.

thereby negating those values, the average was Analysis of variance supported the assumption that
2.35970.993 mg g1. PAH emission varied among the vehicles (ANOVA:
As can be seen in Table 3, vehicle A presented the F ¼ 11:3 > Fc ¼ 6:6). In addition, vehicles A and B
highest fuel autonomy: average of 9.315 km l1, a value displayed more repeatable behaviour, presenting, re-
considerably higher than the average of 6.717 km l1 spectively, 5.8% and 4.0% relative standard deviation in
obtained by vehicle B and the averages of 7.764 and PAH emissions, whereas vehicles C and D presented
7.112 km l1 obtained by vehicles C and D, respectively. greater variation: 66.0% and 29.3%, respectively.
Although vehicle A had presented such good fuel Despite these differences, the correlation coefficient
autonomy, contrary to what had been expected, it observed between emissions of PAHs and those of CO2
presented the highest particulate matter emission (aver- was 0.94, showing that there is an inverse correlation
age of 1.452 g km1). It also presented the highest between CO2 emissions and PAH emissions. Therefore,
emission of PAHs per kilometre and PAHs per litre of considering that high CO2 gas emission is directly
fuel consumed (average of 5.573 g km1 and 50.564 g l1, related to appropriate combustion, it can be assumed
respectively). According to Bosch (1996), diesel vehicles that efficient combustion proportionates low PAH
are normally adjusted to function with more air than is emissions (Sher, 1998).
necessary to create combustion. This situation is known
as a poor mixture because it is poor in fuel. However, as 3.1. Individual polycyclic aromatic hydrocarbons results
these tests were made with unmodified in-use vehicles,
we can suggest that vehicle A was working with a richer The results from the analysis of each PAH are shown
air/fuel ratio than the others. In other words, vehicle A in Table 4. All 14 PAHs studied were detected and
was receiving more fuel than necessary to create efficient quantified in at least three of the eight tests performed.
combustion, producing partially combusted fuel and Five PAHs (naphthalene, phenanthrene, fluoranthene,
thereby causing increased formation of particulate pyrene and chrysene) were present in all tests performed.
matter and increased PAH emission. The other PAHs exhibited irregular behaviour and, in
ARTICLE IN PRESS
1636 R. de Abrantes et al. / Atmospheric Environment 38 (2004) 1631–1640

Table 4
Emission rates of Individual PAHs, in mg km1

Emission rates of Individual PAHs (mg km1)

Compound Vehicle A Vehicle B Vehicle C Vehicle D

1st test 2nd test 1st test 2nd test 1st test 2nd test 1st test 2nd test

Naphthalene 2503.29 1450.47 738.35 963.52 403.64 2012.74 1702.78 1005.15


Acenaphthene 35.84 16.55 2.03 14.70 ND ND 1.52 ND
Fluorene 242.32 157.51 ND 15.75 ND ND ND ND
Phenanthrene 844.26 972.71 159.52 131.58 310.54 867.91 668.83 518.28
Anthracene 33.28 40.30 20.31 11.47 ND ND 0.28 0.82
Fluoranthene 1557.70 2073.66 88.77 29.08 137.20 100.92 138.18 138.99
Pyrene 108.75 51.63 144.44 67.14 277.48 126.01 177.04 121.65
Benzo[a]anthracene 239.68 326.66 35.55 ND ND ND 28.25 ND
Chrysene 35.11 74.71 43.51 7.72 2.19 2.61 4.37 4.24
Benzo[b]fluoranthene 100.63 103.86 17.88 ND 0.74 0.89 1.89 0.42
Benzo[k]fluoranthene 7.41 3.82 14.47 16.16 0.41 ND 1.52 ND
Benzo[a]pyrene 21.20 20.78 28.25 4.38 ND ND 1.22 ND
Dibenzo[a,h]anthracene 53.68 21.05 28.11 ND 0.38 ND 2.11 3.33
Benzo[ghi]perylene 17.43 31.03 20.92 6.34 ND 1.30 ND ND

ND—Not detected.

Naphthalene Table 5 shows some PAHs identified by other authors.


Phenanthrene
Fluoranthene As can be seen, of the 14 PAHs studied in this work,
Average of naphthalene plus fluoranthene
80 only a few of them were detected in previous studies,
70
68

whereas this study was able to detect them all, although


66

70
65

63

60

Rogge et al. (1993) and Westerholm et al. (1991)


57

60
evaluated just the PAHs adsorbed in the particulate
Percentage

50
40 matter. The six gasoline powered vehicles without
35

33

catalytic converters tested by Rogge et al. (1993) were


28

26

30
evaluated in same test cycle (FTP-75), anthracene,
16

20
11

10 fluoranthene, benzo[a]anthracene and dibenzo[a,h]


6
6
5

0 anthracene presented the same order of magnitude than


Vehicle A Vehicle B Vehicle C Vehicle D the lower values founded in this work, but phenanthrene
Fig. 2. Relative participation of most frequent PAHs in and pyrene presented lower values than this work, by
relation to the sum of PAHs studied. contrast benzo[k]fluoranthene, benzo[a]pyrene and ben-
zo[ghi]perylene showed higher values, probably due to
the fact that gasoline-powered vehicles operate on richer
some cases, were not detected. On average, the sum of mixtures than diesel vehicles, although diesel fuel has a
naphthalene, phenanthrene and fluoranthene emission higher PAHs content compared to gasoline. As ex-
rates amounted to 87% of total PAH emissions. pected, the seven gasoline powered vehicles with
Fig. 2 shows the relative participation of the most catalytic converters, evaluated in same conduction cycle,
frequent PAHs in relation to the sum of 14 PAHs presented lower PAHs values.
studied. Regularity in the percentile emission of the most Trucks tested by Rogge et al. (1993) and by
frequent PAHs in relation to the sum of PAHs is Westerholm et al. (1991) also presented lower values
evident. Naphthalene participation averaged 54% in and in some cases showed the same order of magnitude
relation to the sum of PAHs. Similarly, phenanthrene than the lower values founded in this work, these
and fluoranthene averaged 20% and12%, respectively. deviations can be attributable to differences among in
Again, only vehicle A presented behaviour which was motor power, piston displacement, accumulated mileage
quantitatively different than the other vehicles, emitting and especially to the conduction cycle.
a lower percentage of naphthalene and a higher The eight different diesel fuels tested in a bus and
percentage of fluoranthene. However, the sum of relative in a truck by Westerholm and Li (1994), showed wide
participation of naphthalene and fluoranthene in rela- variations in PAH emissions, indicating that fuel
tion to the sum of PAHs was comparable for all vehicles, chemical composition influences PAH emission. Max-
with an average sum of 66.6% and relative standard imum values of anthracene, fluoranthene, pyrene,
deviation of 4.7%. chrysene and benzo[b]fluoranthene also presented the
ARTICLE IN PRESS
R. de Abrantes et al. / Atmospheric Environment 38 (2004) 1631–1640 1637

Table 5
Comparison of PAH emission with results from others studies, in mg km1

Emission rate (mg km1)

Compound This worka Rogge et al., 1993b,d Westerholm et al. (1991)c,d Westerholm and Li, 1994e

b1 b2 b3

Phenanthrene 131.5–972.7 17.3 0.88 12.2 29 0.12–86


Anthracene 0.2–40.37 5.1 0.11 1.6 3.3 0.011–15
Fluoranthene 29.1–2073.7 48.3 2.0 13.0 17 8.8–45
Pyrene 51.6–277.5 31.0 2.5 22.6 11 10–54
Benzo[a]anthracene 28.2–326.7 73.8 1.9 3.6 0.47 0.49–4.1
Chrysene 2.2–74.7 2.8 0.1–3.2
Benzo[b]fluoranthene 0.4–103.9 0.021–1.1
Benzo[k]fluoranthene 0.4–16.2 40.7 2.0 2.7
Benzo[a]pyrene 1.2–28.2 43.5 1.9 1.3 o0.06
Dibenzo[a,h]anthracene 0.4–53.7 8.3 0.33 ND
Benzo[ghi]perylene 1.3–31.0 145.4 4.70 1.6 o0.13 0.012–0.34

ND—Not detected.
a
Three Vans and one pickup fuelled with commercial diesel, cycle FTP-75.
b
(b1)—six Non catalyst cars fuelled with gasoline, cycle FTP-75. (b2)—seven Catalyst car fuelled with gasoline, cycle FTP-75. (b3)—
two Trucks fuelled with diesel, cycle Idle 40 and 70 km h1.
c
One Truck fuelled with diesel, bus cycle.
d
Authors that Informed just PAHs present in solid phase.
e
One bus and one truck fuelled with 8 different diesel, bus cycle.

same order of magnitude than the lower values detected deviation: 39% and 44%, respectively. In addition, only
in vehicles tested in this study. However, the differences the ratio between benzo[ghi]perylene and benzo[a]pyrene
above cited, like motor characteristics and test condi- (1.13735%) was very close to that found by Martinis
tion, must be taken into account. et al. (2002) (1.11), whose sample site was in the city of
Emission profiles of PAHs can be affected by fuel S*ao Paulo, where vehicular emissions probably are the
chemical composition and test cycle (Westerholm and main source of PAHs. We suggest more studies to
Li, 1994; Mi et al., 2000). In addition, atmospheric PAH confirm if these ratios can be considered good ‘‘finger-
profiles can be affected by meteorological variables such prints’’ for motor vehicles.
as UV light and temperature, furthermore half-lives of
these compounds in the atmosphere vary (Martinis et al., 3.2. Distribution of PAHs between the solid and gaseous
2002; WHO, 1998). Despite these facts, some authors phase
consider the ratios between some of these compounds as
‘‘fingerprints’’ of emission sources (Khalili et al., 1995; Fig. 3, in log scale, shows the average emission
Li and Kamens, 1993). of PAHs by test found in solid and gaseous phases.
Table 6 shows the ratios found in this work and those On average, the PAH mass adsorbed in the particulate
calculated from other studies. The gasoline-fuelled matter was 21.5 mg, while 31.4 mg was present in
vehicles evaluated by Rogge et al. (1993) presented the gaseous phase corresponding to 40.6% and 59.4%,
generally lower ratios; only the ratios between benzo[- respectively. Lower molecular weight PAHs, such
ghi]perylene and benzo[a]pyrene (1.62 for vehicles with as naphthalene and acenaphthene, were in higher
catalytic converters and 3.84 for vehicles without) were concentrations in the gaseous phase (99.8% and
higher than the average ratio found in this study (1.13). 73.0%, respectively). Those of intermediate molecular
Diesel-powered vehicles presented different ratios too, weight were present in both phases; relative values in the
including those tested by Westerholm and Li (1994), solid phase were: fluorene: 56.7%; phenanthrene:
with wide variability in these ratios in relation to fuel 64.1%; anthracene: 83.7%; fluoranthene: 99.0%; and
composition, in short the PAH ratios as ‘‘fingerprints’’ pyrene: 78.7%. Higher molecular weight PAHs were
of motor vehicles must be taken with care. predominantly adsorbed in the soot particles, such
In spite of disparities among the vehicles in this benzo[a]anthracene: 100%; chrysene: 98.6%; benzo[b]
study, the ratio between benzo[a]anthracene and ace- fluoranthene: 97.5%; benzo[k]fluoranthene: 100%; ben-
naphthene (15.63) and the ratio between chrysene and zo[a]pyrene: 100%; dibenzo[a,h]anthracene: 96.5%;
benzo[a]pyrene (2.43) showed a small relative standard benzo[ghi]perylene: 100%. Following IARC (2003)
ARTICLE IN PRESS
1638 R. de Abrantes et al. / Atmospheric Environment 38 (2004) 1631–1640

Table 6
Comparison of PAH ratios with others studies

PAH ratios

This worka Rogge et al. (1993)b Westerholm Westerholm and Martinis


et al. (1991)c Li (1994)d et al. (2002)e
Average RSD (%) b1 b2 b3

Benzo½aanthracene 15.63 39
Acenaphthene
Fluoranthene: 5.06 37 0.65 1.05 3.61 36.17 10.98–36.54
Benzo½aanthracene
Pyrene: 34.71 179 0.82 0.86 7.79 4.70
Benzo½apyrene
Benzo½aanthracene: 12.87 71 1.95 0.66 1.24
Benzo½apyrene
Chrysene: 2.43 44
Benzo½apyrene
Dibenzo½a; hanthracene: 1.57 46 1.15 0.66 0.45
Benzo½apyrene
Benzo½ghiperylene: 1.13 35 3.84 1.62 0.55 1.1
Benzo½apyrene
Anthracene: 1.50 29 0.04 0.02 1.00 0.32–533.33
Benzo½ghiperylene
Chrysene: 1.95 23 0.38–91.67
Benzo½ghiperylene
a
3 Vans and 1 Pickup fuelled with commercial diesel, cycle FTP-75.
b
(b1)—calculated from Rogge et al. (1993) data—6 non catalyst cars fuelled with gasoline , cycle FTP-75.(b2)—calculated from
Rogge et al. (1993) data—7 catalyst car fuelled with gasoline, cycle FTP-75. (b3)—calculated from Rogge et al. (1993) data—2 trucks
fuelled with diesel, cycle Idle 40 and 70 km h1.
c
Calculated from Westerholm et al. (1991) data—1 truck (14.19 l) fuelled with diesel, bus cycle.
d
Calculated from Westerholm and Li (1994) data—1 bus (11 l) and 1 truck (9.6 l) fuelled with 8 different diesel fuels, bus cycle.
e
Calculated from Martinis et al. (2002) data—Atmospheric sampling urban site.

100000
Solid phase
10000 Gaseous phase
Average (ug/test)

1000

100

10

1
Benzo[a]anthracene

Benzo[b]fluoranthene

Benzo[k]fluoranthene

Benzo[a]pyrene

Dibenzo[a,h]anthracene
Acenaphthene

Phenanthrene

Anthracene

Fluoranthene

Pyrene
Naphthalene

Fluorene

Chrysene

Benzo[ghi]perilene

Fig. 3. PAHs average emission in mass by test, separated by gaseous and solid phase.
ARTICLE IN PRESS
R. de Abrantes et al. / Atmospheric Environment 38 (2004) 1631–1640 1639

classification, we were able to verify that the more highly measurement of fuel consumption. ABNT, Rio de Janeiro.
carcinogenic PAHs, with the exception of naphthalene, Available at www.abnt.org.br.
are predominantly adsorbed in the particulate matter. Anfavea: Associa@*ao Nacional dos Fabricantes de Ve!ıculos
Automotores, 2003. Anu!ario estat!ıstico da industria !
automobil!ıstica brasileira statistical annual of the Brazilian
automobile industry. Report on line. Available at
4. Conclusions URL:http://www.anfavea.com.br/Index.html [accessed in
2003 May 27].
This study has profiled the principal PAH emissions Azevedo, D.A., Moreira, L.S., Siqueira, D.S., 1999. Composi-
tion of extractable organic matter in aerosols from urban
of four light-duty commercial vehicles powered by
areas of Rio de Janeiro city, Brazil. Atmospheric environ-
commercial diesel fuel and without catalytic converters.
ment 33, 4987–5001.
The sum of fourteen PAHs presented an average Bosch, 1996. Automotive Handbook. 4th Edition. Robert
of 2.81571.846 mg km1, ranging from 1.133 to Bosch GmbH, Stuttgart.
5.801 mg km1. Moreover, PAH emissions ranged from Cetesb: Companhia de Tecnologia de Saneamento Ambiental,
0.915 to 3.580 mg per gram of particulate matter and !
2002. Relatorio de qualidade do ar no Estado de S*ao Paulo
from 8.948 to 55.239 mg/l of fuel consumed. 2001, report of air quality in the S*ao Paulo State 2001.
All 14 PAHs studied were detected and quantified in Cetesb, S*ao Paulo. Available at www.cetesb.sp.gov.br.
at least three of the eight tests performed. Naphthalene, IARC: International Agency for Research on Cancer, 2003.
phenanthrene, fluoranthene, pyrene and chrysene were Evaluations of carcinogenicity to humans, available at
detected in all tests. The sum of median emission rate of http://www.iarc.fr/ [accessed in 2003 March 03].
Ibama: Instituto Brasileiro de Meio Ambiente e dos Recursos
naphthalene, phenanthrene and fluoranthene amounted
Naturais Renov!aveis, 1998. Programa de controle da
to 87% of the PAHs detected. In addition, the sum of
polui@*ao do ar por ve!ıculos automotores—Proconve,
naphthalene and fluoranthene emissions proved to be control program for the air pollution by self-driven
generally constant for all vehicles, with an average of vehicles—Proconve. 2nd Edition, Ibama, Bras!ılia.
66.6% and relative standard deviation of 4.7%. Khalili, N.R., Scheff, P.A., Holsen, T.M., 1995. PAH source
In general, comparisons with studies by other authors fingerprints for coke ovens, diesel and gasoline engines,
(Rogge et al., 1993; Westerholm et al., 1991; Westerholm highway tunnels, and wood combustion emissions. Atmo-
and Li, 1994), showed that the vehicles evaluated in this spheric Environment 19, 533–542.
work presented higher PAH emission, except the gaso- Li, C.K., Kamens, R.M., 1993. The use of polycyclic aromatic
line-powered vehicles without catalytic converters that hydrocarbons as sources signatures in receptor modeling.
emitted practically the same order of magnitude (Rogge Atmospheric Environment 27A, 523–532.
Mage, D.T., Zali, O., 1992. Motor vehicle air pollution. World
et al., 1993). Furthermore, it was possible to verify that
Health Organisation, Geneva.
conduction cycle affects PAH emission and also the
Martinis, B.S., Okamoto, R., Kado, N.Y., Gundel, L.A.,
ratios between some PAHs. ‘‘Fingerprints’’ of PAHs Carvalho, L.R.F., 2002. Polycyclic aromatic hydrocarbons
emission from motors vehicles can be affected by many in a bioassay-fractionated extract of PM10 collected in S*ao
variables, and only could be considered after detailed Paulo, Brazil. Atmospheric Environment 36, 307–314.
evaluation. In this work, only the ratio between Mendon@a, M.C.G., 1991. Hidrocarbonetos polic!ıclicos arom-
benzo[ghi]perylene and benzo[a]pyrene was close to the a! ticos em aerossois
! atmosf!ericos: quantifica@*ao de baixas
ratio observed in the atmosphere of the city of S*ao Paulo concentra@oes* por cromatografia l!ıquida de alto desempen-
by Martinis et al. (2002) (1.1), indicating that further ho (Polycyclic aromatic hydrocarbons in atmospheric
studies are needed to evaluate if this ratio could be a aerosols: quantification of low concentrations for high
good ‘‘fingerprint’’ for diesel vehicles. performance liquid chromatograph). M.Sc Thesis, Chem-
istry Institute, University of S*ao Paulo.
On average, emission by test was 21.5 mg in the solid
Mi, H.H., Lee, W.J., Chen, C.B., 2000. Effect of fuel aromatic
phase and 31.4 mg in the gaseous phase, corresponding
content on PAH emission from a heavy-duty diesel engine.
to 40.6% and 59.4%, respectively. The PAHs with Chemosphere 41, 1783–1790.
higher carcinogenic potential are predominantly ad- Provost, L.P., 1984. Statistical methods in environmental
sorbed in the particulate matter. sampling. In: Schweitzer, G.E., Santolucito, J.A. (Eds.),
A negative correlation of 0.94 was observed between Environmental Sampling for Hazardous Wastes. 2nd
CO2 emission and PAH emission. Edition. ACS Symposium Series, vol. 267. American
chemical society, Las Vegas, pp. 79–96.
Rogge, W.F., Hildermann, L.M., Mazurek, M.A., Cass, G.R.,
1993. Sources of fine organic aerosol: noncatalyst and
References catalyst-equipped automobiles and heavy-duty diesel trucks.
Environmental Science and Technology 27, 636–651.
ABNT: Associa@*ao Brasileira de Normas T!ecnicas, 1990. NBR SAE: Society Automotive engineers, 1998. J 1936: Chemical
7024: Ve!ıculos rodovi!arios automotores leves: Medi@*ao methods for the measurement of non regulated diesel
do consumo de combust!ıvel light self-driven vehicles: emissions. In: Society Automotive Engineers. Surface
ARTICLE IN PRESS
1640 R. de Abrantes et al. / Atmospheric Environment 38 (2004) 1631–1640

Vehicles Emissions Standards Manual, Warrendale, e florestais (A study about the characterisation of poly-
pp. 278–287. cyclic aromatic hydrocarbons and its derived, and saturated
Saldiva, P.H., Concei@*ao, G.M., Miraglia, S.G., Kishi, H.S., aliphatic hydrocarbons in particulate matter originat-
Singer, J.M., 2001. Air pollution and child mortality: a time- ing from atmospheric urban sites, suburban and forest).
series study in Sao Paulo, Brazil. Environmental Health Ph.D. Thesis, Chemistry Institute, University of S*ao
Perspectives 109 (Suppl. 3), 347–350. Paulo.
Sher, F. (Ed.), 1998. Handbook of Air Pollution from Internal Westerholm, R., Li, H., 1994. A multivariate statistical analysis
Combustion Engines. Academic Press, London. of fuel-related polycyclic aromatic hydrocarbon emissions
United States. Code Federal Regulations 40, Part 86 sub part B, from heavy-duty diesel vehicles. Environmental Science and
1997. CFR 40. Emission regulations for 1977 and later Technology 28, 962–972.
model year new light-duty vehicles and new light-duty Westerholm, R.N., Alm!en, J., Li, H., 1991. Chemical and
trucks; test procedures, Available at http://www.access. biological characterisation of particulate, semivolatile and
gpo.gov/nara/cfr/waisidx 99/40cfr86 99.html [accessed in gas phase associated compounds in diluted heavy duty
2000 April 12]. diesel exhaust. Environmental Science and Technology 25,
United States. Environmental Protection Agency, 1988. 332–338.
Method TO-13: the determination of benzo[a]pyrene and WHO: World Health Organisation, 1996a. Diesel Fuel and
other polynuclear aromatic hydrocarbons in ambient air Exhaust Emissions. WHO, Stuttgart, Environmental health
using gas chromatographic and high performance liquid criteria, 171.
chromatographic analysis. In: Compendium of Methods for WHO: World Health Organisation, 1996b. Update and revision
the Determination of Toxic Organic Compounds in of WHO air quality guidelines for Europe, Report on
Ambient Air, 4th Edition. Research Triangle Park. line. Available at URL: http://www.who.nl/index1.htm
Vasconcellos, P.C., 1996. Um estudo sobre a caracteriza@*ao [2001 Out 17].
de hidrocarbonetos polic!ıclicos arom!aticos e seus derivados, WHO: World Health Organisation, 1998. Selected non-hetero-
e hidrocarbonetos alif!aticos saturados em material particu- cyclic polycyclic aromatic hydrocarbons. WHO, Stuttgart,
lado atmosf!erico proveniente de s!ıtios urbanos, suburbano (Environmental Health Criteria, 202).

You might also like