Comparison of Emissivity , Transmittance , and Reflectance Infrared Spectra of Polycyclic Aromatic Hydrocarbons with those of Atmospheric Particulates ( PM 1 )

Polycyclic Aromatic Hydrocarbons (PAHs) are a group of various complex organic compounds composed of carbon and hydrogen, and two or more condensed benzene rings. They are released into the atmosphere by the incomplete combustion or pyrolysis of organic matter. Some of the major sources of PAHs are burning of coal, wood, oil or gas, vehicle engines exhaust, and open burning. PAHs are of great concern to human health mainly because of their known carcinogenic and mutagenic properties. Consequently, it is very important to study atmospheric PAHs, especially those associated with ultrafine particles. This study aims to identify the spectral features of PAHs in samples of particulate matter < 1 μm (PM1) using infrared spectrometry. Emissivity and transmittance spectra of PAHs were obtained by infrared spectroscopy. PAHs standards spectra contributed to effectively identify PAHs in PM1 samples. Emissivity and transmittance spectra in the range of 680–900 cm exhibited the largest number of bands due to C–C out-of-plane angular deformations and C–H out-of-plane angular deformations. Bands of medium intensity in 2900–3050 cm region were also observed due to C–H stretching typical of aromatic compounds, although with lower intensity. This study compared the emissivity and transmittance spectra acquired using two different infrared spectrometers in order to identify PAHs in samples of atmospheric particulate matter and analyzed the capability and advantages of each of the infrared spectrometers. In addition, it was confirmed that the PAHs under study can be distinguished by their infrared spectral fingerprints.


INTRODUCTION
Particulate matter (PM) is considered as one of the major pollutants that affect the health of humans, especially for the fine and ultrafine fraction, which can adsorb greater concentrations of toxic compounds, e.g., polycyclic aromatic hydrocarbons-PAHs (Silva et al., 2009;Ribeiro et al., 2010;Kronbauer et al., 2013;Dias et al., 2014;Garcia et al., 2014;Pérez et al., 2014).Atmospheric aerosols contribute to adverse health and environmental effects, visibility degradation, and radiative forcing (Jacobson and Hansson, 2000).In order to understand and control the impacts of atmospheric aerosols, it is important to determine the physical and chemical properties of aerosols (Coury and Dillner, 2009).Various studies have pointed to vehicular emissions as the main source of fine and ultrafine particles in urban environments (Harrison et al., 1999;Shi and Harrison, 1999;Shi et al., 1999).PAHs are a group of several complex organic compounds consisting of carbon and hydrogen, and two or more condensed benzene rings and represent one of the most stable families of organic compounds known (Ravindra et al., 2008).PAHs may exist in large number of different structures, depending on the complexity and the number of isomers.PAHs represent a great public health concern (Panther et al., 1999) and they constitute a group of widespread pollutants of great environmental interest.PAHs and their metabolites are among the most toxic, carcinogenic, and mutagenic atmospheric contaminants known (Nagabalasubramanian and Periandy, 2010).
The key sources of atmospheric emission of PAHs are incomplete combustion or pyrolysis of organic matter (Li et al., 2005).Vehicular exhausts and combustion of coal, wood, or oil are the major contributors of PAHs in the atmosphere.In an urban environment, vehicular exhausts, especially those from diesel engines, are the main sources of PAHs (Tsapakins et al., 2002;Guo, 2003;Fang et al., 2004;Manoli et al., 2004;Ravindra et al., 2006;Ströher et al., 2007).PAHs released into the air may react with gases present in the atmosphere that might transform them into products that are probably more dangerous than they are in their native forms (i.e., nitrated PAHs).Moreover, organic compounds may enhance or inhibit the water-uptake characteristics of the fine particles (Blando et al., 2001).
In general, planar PAHs are more stable, less reactive, and biologically less toxic (Cerqueira et al., 2011;Silva et al., 2011;Cerqueira et al., 2012;Oliveira et al., 2012a, b;Quispe et al., 2012;Silva et al., 2012;Arenas-Lago et al., 2013;Oliveira et al., 2013;Ribeiro et al., 2013;Cutruneo et al., 2014;Saikia et al., 2014).Whereas, substitution of methyl or other groups into the ring system of PAH can result in molecular distortion from planarity that sometimes translates into more reactivity and carcinogenicity.The chemistry of PAHs allows for an extreme diversity in their molecular structure.Also, the arrangement of aromatic rings and the possibility of numerous substitutions and sidechains allow for numerous permutations in the structure of PAHs (Izawa et al., 2014).The effects of changes in the molecular structure of PAHs and their composition on solid-state reflectance spectra (ultraviolet, visible, and nearinfrared) were investigated using pure PAHs.Researchers also studied the fate of PAHs in the environment (transformation, degradation) and their effect on the health (toxicity) that depends of environmental factors (He et al., 2009).
Many studies have explored the analysis of PAHs in PM using gas chromatography associated with mass spectrometry (GC/MS) (Dallarosa et al., 2005a, b;Teixeira et al., 2012;Agudelo-Castañeda and Teixeira, 2014).Some of the disadvantages associated with GC/MS analysis are the requirement of a relatively large amount of aerosol mass, extraction, complex derivatization procedures for detecting compounds of varying polarity, and it is not possible to detect large molecules (Yu et al., 1998).Spectroscopic techniques has some advantages over chromatographic techniques in aerosol particle analysis.Advantages includes the ability to perform chemical fingerprinting; the ability to detect compounds in smaller concentrations without extraction or derivatization (nondestructive method) and the possibility of using the instrumentation in the sampling site.Thus the losses may be diminished of the samples during their transportation, cooling and storage; also, smaller sample sizes may be used with no requirement for sample preparation (Allen and Palen, 1989;Allen et al., 1994;Marshall et al., 1994;Coury and Dillner, 2008;Navarta et al., 2008).In addition, in spectroscopic techniques, several samples can be selected in order to perform GC/MS quantitative analysis at a later stage, thus reducing economic costs.
Moreover, Fourier transform infrared spectroscopy (FTIR) can be used to identify a compound or investigate the chemical composition of a sample using its vibrational spectrum, which is considered as one of its most characteristic physico-chemical properties, particularly in organic compounds.The vibrational modes depend on the type of the internal structure of the constituents, the size of the ionic radius, and the bonding forces and ionic impurities contained in the matter (Meneses, 2001).The total energy contained in a material is composed of the components associated with the electronic, vibrational, rotational, and translational energy.The infrared spectrum covers only the spectral region corresponding to vibrational and rotational modes giving intrinsic information on the chemical and structural composition of the material.Infrared spectroscopy has been widely used to determine the composition of submicron particles (Maria et al., 2003;Liu et al., 2009;Russell et al., 2011), although these works do not study PAHs.
In the present study, the Fourier transformed infrared emissivity and transmittance spectra of particulate matter of size < 1 µm (PM 1 ) and that of solid standards of PAHs were compared to study the efficiency of these techniques in the identification of PAHs in PM 1 samples collected on filters without prior treatment.This study gains significance since these PM 1 samples are hazardous than larger particles because they can enter the respiratory tract (Slezakova et al., 2007).Moreover, these particles possess larger specific area and hence can adsorb high concentrations of toxic, mutagenic, and/or carcinogenic compounds.

Sampling of Atmospheric Particulate Matter
PM 1 sampling was performed in an automatic sequential particle sampler model PM162M from Environment S.A. at a volumetric flow rate of 1.0 m 3 /h.PM 1 samples were collected using Zefluor™ membrane PTFE (polytetrafluorethylene) filters of 47 mm diameter, specifically designed for organic sampling (Peltonen and Kuljukka, 1995).The equipment for collection of particulate matter < 1 µm (PM 1 ) was installed at the following geographic coordinates (UTM) 29°49′06′′S/ 51°09′34′′W (Sapucaia do Sul) and 29°55′50′′S/51°10′56′′W (Canoas).Sapucaia do Sul and Canoas are located at south Brazil, in Rio Grande do Sul State.Canoas is under a strong vehicular influence, daily traffic congestions, Canoas air base and industries (oil refinery) upstream of the prevailing winds that have a medium influence in this sampling site.Sapucaia do Sul site has a greater vehicular influence vehicle: light and heavy fleet, traffic congestions and slow speeds.This site also has low industrial influence (oil refinery, steel mills that do not use coke, Canoas Air Force Base) upstream of prevailing winds.Sampling was carried out at a constant flow rate of 1 m 3 /h for 72 and 12 hours for the determination of emissivity (FTIR) and transmittance spectra.These sampling times were found to be appropriate for a good resolution of bands corresponding to organic species.Several studies show that PTFE filters are more adapted for FTIR tests and have lower absorption bands (overlapping peaks) in infrared analyses than nucleopore or quartz filters (Ghauch et al., 2006).

Instruments
Emissivity spectra were obtained using a Model 102F portable FTIR spectrometer by Design and Prototypes at Centro Estadual de Pesquisa em Sensoriamento Remoto e Meteorologia (CEPSRM) (State Research Center for Remote Sensing and Meteorology) of the Federal University of Rio Grande do Sul (UFRGS).The FTIR spectrometer consists of an optical/electronic module, Michelson interferometer, and infrared detectors.It is equipped with two detectors, an InSb (Indio-antimony) and a MCT (Mercury-cadmium-tellurium) detector covering 3330-2000 cm -1 and 2000-700 cm -1 spectral ranges, respectively, cooled by liquid nitrogen (Salisbury, 1998), and two blackbodies that can be set at different temperatures in the range of 5-60°C during the instrument calibration.The spectra were analyzed in the range of 1660-700 cm -1 , with a spectral resolution of 4 cm -1 and spectral accuracy of ± 1 cm -1 .The 1660-3330 cm -1 range was not analyzed due to low signal/noise ratio of the equipment in this range (Korb et al., 1996).Furthermore, within this range (1660-3330 cm -1 ) transmittance is low due to the presence of absorption bands of methane, CO 2 , and water vapor (Korb et al., 1996).The data were obtained from an average of 100 scans, i.e., 100 co-added interferograms.The measurements were performed at a distance of less than 50 cm to minimize atmospheric attenuation (Korb et al., 1996) using a foreoptic of diameter 2.54 cm to ensure that the field of view (FOV) was smaller than the sample with a diameter of 47 mm.In addition, all emissivity measurements were performed under conditions of clear sky without clouds with low to moderate relative humidity (< 60%).
Transmittance spectra were obtained by a BOMEM MBseries FTIR-Hartmann & Braun Michelson spectrometer equipped with DTGS detector at the Institute of Physics of Federal University of Rio Grande do Sul (UFRGS).Fifty scans of transmittance spectra with a resolution of 4 cm -1 were performed in the range of 400-4000 cm -1 to obtain an appropriate signal/noise.Transmittance spectra were collected from each sample of PM 1 and are presented in arbitrary units (a.u.), using a background spectrum of blank filter without sample.

Calibration and Operation of the FTIR Spectrometer
Emission and responsiveness of the instrument were calibrated with blackbodies under controlled temperatures (Salisbury, 1998).Calibration consisted of measuring the radiation of two blackbodies at two known temperatures and the background emission of the instrument was deducted.In addition, downwelling radiance calibration was also performed.Downwelling radiance of the hemisphere above the target is measured by a reflector located at the target, a gold plate with emissivity of ε = 0.040 in the spectral range.More details on the radiometric calibration of the instrument can be found in the studies by Hook and Kahle (1996) and Korb et al. (1996).
The measured radiance actually consists of the sum of various radiances entering the equipment, and for the conditions used in the present study the relation between the radiances can be written as (Salvaggio and Miller, 2001): where, L S (λ) is the total spectral radiance entering the sensor; B (λ,T S ) is the radiance of a blackbody at the sample temperature (T S ); (1 -ε s ) is the sample reflectance; and L DWR is the downwelling radiance.
Sample emissivity can be determined using Eq. ( 2), where the radiance measured must be divided by the Planck function of blackbody radiance at the sample temperature (T S ), which is unknown.
where, ε s (λ): emissivity of the sample surface as a function of the wavelength L s (λ): calibrated radiance of the sample L DWR (λ): calibrated radiance of the incident radiance B(λ,T s ): Planck function at the sample temperature Sample temperature was defined as described by Johnson et al. (1998) and Ribeiro da Luz and Crowley (2007).Initially, temperature measured by an infrared digital thermometer Minipa MT-350 with accuracy of ± 2°C was selected.Fig. 1 shows an example of emissivity spectra obtained as a function of the diverse temperatures chosen.This spectrometer calculates the spectra in µm, which was converted to frequency (cm -1 ) in order to compare the results with other studies.Water bands at 8.18, 8.24, and 12.58 µm show the residual characteristics of the downwelling radiance (Ribeiro da Luz, 2005), as observed in Fig. 1 indicated by arrows.If the temperature chosen is higher than the actual temperature of the sample (overestimated temperature), the residual characteristics of the downwelling radiance will point downwards.If the temperature chosen is lower than that of the sample, the residual characteristics will point upwards (underestimated temperature), similarly to the downwelling radiance (Horton et al., 1998).The correct emissivity spectrum was obtained through a trial-and-error process, i.e., in which the residual characteristics of the downwelling radiance were minimized.The spectrum obtained at 32°C did not show the residual characteristics of downwelling radiance indicating the appropriate temperature.The configuration geometry of the emissivity measurements was as follows: distance of < 50 cm between the sensor and the sample; sample size of 47 mm diameter; optical angle of 90°.

Preparation and Analysis of the PAHs Solid Standards
For transmittance spectra, the PAH solid standards of fluoranthene, pyrene, benzo [α]pyrene and benzo[α]anthracene with 99% purity were obtained from Sigma Aldrich and were prepared in solid KBr pellets to obtain transmittance spectra.For emissivity spectra, PAHs solid standards were used as received and placed in 50 mm diameter dishes.In both cases, PM 1 samples were analysed directly without any preparation, and the molecular vibrations of the two classes of spectra were identified using sources in the literature (Semmler et al., 1991;Carrasco Flores et al., 2005;Onchoke et al., 2006;Onchoke and Parks 2011).

Spectra of PAHs Solid Standards Emissivity
Figs. 2(a) and 2(b) show the emissivity spectrum of fluoranthene and pyrene, respectively.The emissivity of fluoranthene spectrum (Fig. 2(a)) was observed to have peaks of higher intensity than that of pyrene spectrum (Fig. 2(b)).Table 1 presents the representative peaks observed in the experimental emissivity spectra of fluoranthene and pyrene.The absorption bands corresponding to CO 2 and atmospheric water vapor in the range of 700 cm -1 and 1646 cm -1 were excluded from Table 1 (Jellison and Miller, 2004).The vibrations from C-C out-of-plane angular deformation and C-H out-of-plane (Table 1) were identified in the range of 700-850 cm -1 for both the standards.Previous studies (Semmler et al., 1991) have shown that at least one band is present in the range of 600-900 cm -1 in the spectra of PAHs.
Intense and moderate intensity bands were observed in the spectra of fluoranthene at 830 cm -1 and 734 cm -1 corresponding to C-C out-of-plane angular deformation and C-H out-of-plane angular deformation (Hudgins and Sandford, 1998b;Semmler et al., 1991).In the case of pyrene, peaks at 849 cm -1 and 734 cm -1 were assigned to C-C out-of-plane angular deformation and C-H out-ofplane angular deformation (Hudgins and Sandford, 1998a;Semmler et al., 1991).The spectral range of 1000-1300 cm -1 contains various peaks of low intensity due to C-H in-plane aromatic angular deformations (Table 1), except for 1100 cm -1 , where the peaks relating to C-C stretch and C-H in-plane angular deformation can be observed (Kubicki, 2001).In the range from 1350-1570 cm -1 , peaks of high intensity were observed for the fluoranthene spectrum, whereas they appear with low intensity ones in the pyrene spectrum.In the fluoranthene spectrum (Table 1), the peaks were mostly due to the presence of five rings of cyclopentadienyl (Hudgins and Sandford, 1998a).For pyrene (Table 1), the peaks were observed mainly in the range of 1313-1323 cm -1 , resulting from the C-H in-plane angular deformation.The peaks in the range 1358-1488 cm -1 appear due to C-C stretch and C-H in-plane angular deformation only in the fluoranthene spectra.

Spectra of Samples of Atmospheric Particulate Matter (PM 1 ) Emissivity
Figs. 4(a) and 4(b) shows the emissivity spectra of samples (PM 1 ) collected in Canoas and Sapucaia do Sul, respectively.In the emissivity spectra (Fig. 4), peaks can be observed in the range of 730-850 cm -1 resulting from C-C out-of-plane angular deformation and C-H out-ofplane angular deformation.Also, peaks in the range of 1350-1500 cm -1 resulting from C-C stretch vibrations and C-H in-plane angular deformation can be observed.The emissivity spectra of PM 1 samples (Fig. 4) displayed different peaks occurring in the 1620-1650 cm -1 spectral range.An array of compounds absorb radiation in this region, including -OH present in water, alcohols, and carboxylic acids, and the carbonyl stretch (C = O) such as amides that are more conjugated than aldehydes, ketones and acids (Reff et al., 2005).For example, the vibrational mode v 2 of liquid water occurs at 1640 cm -1 .Thus, these bands overlap and hinder the identification of organic compounds.Moreover, the emissivity spectra exhibit a peak at ~1100 cm -1 which corresponds to the carbon-fluorine bond (C-F); consequently, bands at this frequency may overlap and hence, cannot be identified unambiguously (Ghauch et al., 2006).With reference to the identification of PAHs, a peak may be observed at ≈830 cm -1 in both the spectra (Figs.4(a) and 4(b)), indicating a C-H out-of-plane angular deformation, possibly associated with an aromatic group (Wang et al., 2007).Other peaks can also be observed in the range of 600-900 cm -1 , a spectral region where vibrations of C-C and C-H out-of-plane angular deformation can occur, including that of the PAHs (Semmler et al., 1991).Various authors have shown that the spectral region in the range of 700-900 cm -1 is most characteristic and distinctive for the identification of PAHs molecules (Langhoff, 1996); and that in PAHs spectra, at least one band is found in 680-900 cm -1 spectral range (Semmler et al., 1991).Moreover, several others have defined 770-900 cm -1 spectral range as typical of PAHs, that corresponds to the C-H out-of-plane angular deformation (Szczepanski and Vala, 1993;Hudgins et al., 1994;Langhoff, 1996).Consequently, the bands observed in the emissivity spectra of PM 1 samples (Figs.4(a ≈767 cm -1 , ≈774 cm -1 , ≈815 cm -1 , ≈830 cm -1 , and ≈850 cm -1 are possibly related to molecular vibrations of PAHs.The band at ≈732 cm -1 may be due to four neighboring C-H units; the bands at ≈767 cm -1 /≈774 cm -1 due to three neighboring C-H units; and those at ≈815 cm -1 /≈830 cm -1 /≈850 cm -1 to two neighboring C-H units (Semmler et al., 1991).Bands were also observed in 1000-1500 cm -1 spectral range due to vibrations of C-C stretch and C-H in-plane angular deformation.PAHs spectra also presented at least one band in 3000-3100 cm -1 spectral range due to C-H aromatic stretch.

Transmittance
Figs. 5(a)-5(c) shows the transmittance spectra of PM 1 samples collected at Sapucaia do Sul on May 11, 2013 (Fig. 5(a)), at Canoas on May 27, 2013 (Fig. 5(b)) and May 4, 2013 (Fig. 5(c)).Canoas transmittance spectrum collected on May 27, 2013 (Fig. 5(b)) showed less strong peaks than the others did (Figs.5(a) and 5(c)).Probably, the particulate matter concentration in this sample and, consequently, of organic compounds and PAHs, was lower than that in the other samples.All the spectra (Figs.5(a)-5(c)) showed different peaks in 600-900 cm -1 spectral range, corresponding to the vibrations of the aromatic rings, which were also identified in the spectra of PAHs standards.Some peaks were also centered in 1000-1500 cm -1 spectral range, corresponding to the C = C of aromatics, in addition to the C-H out-of-plane angular deformation.Nevertheless, in the transmittance spectra, the peaks in 1250-1300 cm -1 range correspond to the carbon-fluorine (C-F) bond.In this region, a high intensity peak was observed due to the influence of the filter (PTFE), and hence, bands of the compounds at this frequency cannot be identified unambiguously.
As discussed in Section "Spectra of PAHs solid standards", the aromatic organic compounds have low intensity peaks in 3000-3100 cm -1 range resulting from the vibrations from C-H of aromatic groups.Consequently, the bands observed in the PM 1 samples showed low intensity in this spectral region, probably due to low organic compound concentrations (few nanograms per cubic meter), especially PAHs (Allen et al., 1994;Teixeira et al., 2013).For example, the bands in 3000-3100 cm -1 spectral range were observed only in Figs.5(a) and 5(c), but not in Fig. 5(b).
Fig. 5(a) shows the transmittance spectrum of the particulate matter in the sample collected in Sapucaia do Sul on May 11, 2013.The transmittance spectrum (Fig. 5(a)), in 600-900 cm -1 spectral range, showed a peak at 634 cm -1 that was identified in the spectrum of Benzo[a]pyrene (Table 1).Although, at this frequency (634 cm -1 ) molecular vibrations from sulfite ions (SO 3 2-) present in the particulate matter can be observed (Tsai and Kuo, 2006).Peaks were observed at 715, 783, and 856 cm -1 , corresponding to 5, 3, and 2 C-H neighboring units (connected) in the PAHs, respectively (Semmler et al., 1991).At 1000-1500 cm -1 spectral range, the transmittance spectrum presented peaks at 1049, 1157, 1255, and 1467 cm -1 .At 1049 cm -1 and 783 cm -1 frequencies, overlapping may occur due to SiO 4 -4 ion, which also generates frequencies in this region, typical of silicate ions (Tsai and Kuo, 2006).Some of these bands were identified in the fluoranthene spectrum and benzo[a]pyrene spectrum (Table 1), due to C-H out-of-plane angular deformations or C-C stretch.A peak was also identified at 3041 cm -1 due to aromatic C-H stretch, observed in the spectra of diverse PAHs, including pyrene.Peaks at 2920 cm -1 and 1722 cm -1 suggest the presence of aliphatic and carbonyl, respectively.688 cm -1 , 736 cm -1 , and 854 cm -1 due to C-H out-of-plane angular deformations of 5, 4, and 2 CH neighboring units, respectively.Nonetheless, at 688 cm -1 molecular vibrations arising from sulfite ions (SO 3 2- ) may occur (Tsai and Kuo, 2006).Bands were also identified at 1153, 1239, and 1338 cm -1 due to C-H in-plane angular deformations, and at 1454 and 1522 cm -1 due to C-C stretch and C-H in-plane angular deformations, respectively.As explained above, bands in the 3000-3100 cm -1 spectral range were not found.Vibrations from carbonyl and aliphatic groups were observed in bands 1731 and 2929 cm -1 , respectively.Fig. 5(c) shows the transmittance spectrum of the particulate matter in the sample collected in Canoas on May 4, 2013.The transmittance spectrum (Fig. 5(c)) shows peaks at 694, 752, and 819 cm -1 .In this case, the corresponding bands were observed in the benzo[a]pyrene, benzo[a]anthracene, fluoranthene, and pyrene standards (Table 1).Also, at 694 cm -1 , molecular vibrations from sulfite ions (SO 3 2-) may occur (Tsai and Kuo, 2006).The transmittance spectrum presented peaks at 1377 cm -1 and 1504 cm -1 (1000-1500 cm -1 spectral range).A peak in the 3000-3100 cm -1 spectral range, at 3088 cm -1 , was observed.This spectral region presents molecular vibrations arising from C-C in-plane angular deformations and C-H out-of-plane aromatic.

DISCUSSION
Firstly, it may be observed that the emissivity/transmittance spectra of the PAH solid standards (Figs. 2 and 3) showed clear and stronger bands when compared with those of PM 1 samples.This is probable because the intensity and width of the bands depend on composition and density (Kubicki, 2001).Broader bands in the PM 1 sample spectra (Figs. 4  and 5) can be attributed as a result of the interaction between PAH and the surface (Dabestani and Ivanov, 1999); and also to the fact that the particulate matter was collected in an heterogeneous media (e.g., adsorbed onto a filter).In addition, the forms and intensities depend on the relative masses and bonding forces (Hamilton, 2010).Secondly, when comparing emissivity and transmittance spectra obtained by different spectrometers, it might be observed that the peak intensities are different.The intensity of peaks in the transmittance spectra is higher than it is in emissivity spectra at lower frequencies, e.g., < 1000 cm -1 .Consequently, several clear peaks were observed in 600-900 cm -1 range, a spectral region with molecular vibrations typical of PAHs.Thirdly, emissivity spectra (both for PAHs standards and PM 1 samples) identified more peaks in the 1000-1500 cm -1 range than do transmittance spectra, which only showed few peaks.Finally, emissivity spectra showed more peaks in the 600-900 cm -1 in PM 1 samples (Fig. 4) than do transmittance spectra (Fig. 5).However, the transmittance spectra technique has the advantage of a larger spectral range (400-4000 cm -1 ), including 3000-3100 cm -1 region.This region contains the frequencies of the C-H stretch m(C-H), in which some frequencies that are essential to the identification of PAHs may occur.
Concisely, samples of atmospheric particulate matter can be characterized by diverse methods using infrared spectroscopy, either in laboratory or in situ.The FTIR spectrometer (used to obtain emissivity spectra) is advantageous since it can be carried to the field, thus reducing errors caused by transportation or cooling of the samples, whereas measurements of transmittance spectra using conventional FTIR spectrometer must be performed in laboratory, which increases the probability of errors.Nevertheless, transmittance technique has some advantages over emissivity techniques in that it has; (i) a larger spectral range; (ii) no interferences in the measurements taken in situ due to the presence of atmospheric gases.Thus, this study contributes to the knowledge on both the techniques with regards to the identification of PAHs in atmospheric particulate matter collected on filters.The choice of an appropriate technique depends on the goals of measurement.Each of the methods discussed in this study offers unique experimental resources and requires further studies that would analyze the merits and demerits on the quantification of compounds in the samples.

CONCLUSIONS
We conclude that the measurements of both emissivity and transmittance spectra using infrared spectroscopy are useful methods for the analysis of atmospheric particulate matter.Moreover, the spectra of PAHs standards contributed more effectively towards the identification of PAHs in the atmospheric particulate matter collected on filters.These measurements contributed heavily to the successful comparison of two methods in the identification of PAHs.This study confirms that PAHs may be differentiated by their infrared spectral fingerprints using the two methods described here albeit they are poorly studied.We expect that infrared spectroscopy becomes a technique not only to study functional groups, organic or inorganic composition of particle matter, but also PAHs.This study strives to contribute to the knowledge and understanding of these techniques for the analyses of PAHs, since these pollutants are of great health concern; especially, benzo[a]pyrene, classified as carcinogenic to humans according to the International Agency for Research on Cancer (IARC, 2010).Also, the IARC report, Air Pollution and Cancer (IARC, 2013) emphasizes the importance of studying these pollutants using different techniques such that they may be implemented anywhere in the world, including the undeveloped countries.

Fig. 1 .
Fig. 1.Emissivity spectra obtained for a sample temperature range of 30°C to 41°C.Water bands at 8.18, 8.24, and 12.58 µm show the residual characteristics of the downwelling radiance indicated by arrows.

Table 1 .
Designation of features identified in the experimental emissivity spectra of fluoranthene and pyrene in the 700-1660 cm -1 spectral range and in the transmittance spectra of Fluoranthene, Pyrene, Benzo[a]Pyrene and Benzo[a]Anthracene in the 400-4000 cm -1 spectral range.Unit: cm -1 .