Source: Adapted from LBL (1979) and Chow (1995). " XRF = X-ray fluorescence analysis. 6 PIXE = particle-induced X-ray emission. ' Col = colorimetry.

'' AA = atomic absorption spectrometry. '' NA = neutron activation analysis. ' ASV = anodic stripping voltammetry.

MS = mass spectrometry. h ES = emission spectrometry.

' ESCA = electron spectroscopy for chemical analysis; also known as XPS = X-ray photoelectron spectroscopy. ' ICP = inductively coupled plasma spectroscopy.

Source: Adapted from LBL (1979) and Chow (1995). " XRF = X-ray fluorescence analysis. 6 PIXE = particle-induced X-ray emission. ' Col = colorimetry.

'' AA = atomic absorption spectrometry. '' NA = neutron activation analysis. ' ASV = anodic stripping voltammetry.

MS = mass spectrometry. h ES = emission spectrometry.

' ESCA = electron spectroscopy for chemical analysis; also known as XPS = X-ray photoelectron spectroscopy. ' ICP = inductively coupled plasma spectroscopy.

be quantified using conventional absorption spectroscopy, have been used rather extensively in the past. An example is the measurement of CI in aerosols from remote regions by the mercury thiocyanate method (Huebert and Lazrus, 1980). In this technique, chloride ions react with Hg(SCN)2 in a dioxane-ethanol solution to form HgCl2, HgCl^", and SCN". Upon addition of Fe3+ in a nitric acid solution, an orange solution due to FeSCN2"1" results, whose absorbance can be measured at its 460-nm peak (Iwaski et al., 1956).

While colorimetric methods have the advantages of being relatively inexpensive, simple to carry out, and applicable to a large number of elements, they are increasingly being replaced by other physical techniques. The major reason for this is that, as discussed earlier in this chapter, wet chemical methods are more likely to suffer from unrecognized interferences, particularly in complex environmental samples. However, when the aerosol composition is sufficiently well known that one can be confident of the absence of interfering species, colorimetric methods are useful.

X-ray fluorescence (XRF). The sample is irradiated with monochromatic X-rays that eject electrons from the inner shells of the elements. When an electron from an outer shell of the ion drops into the vacancy, it emits characteristic X-rays whose wavelength is used to identify the element and whose intensity is related to the amount present. XRF is used primarily for elements heavier than magnesium because of the weak fluorescence of lighter elements and absorption of the X-rays within the particles. The combination of transmission or scanning electron microscopy (TEM/SEM) with X-ray fluorescence, also known as energy-dispersive spectrometry (EDS), was discussed in Section B.2b.

Particle-induced X-ray emission (PIXE). Elements heavier than sodium can be analyzed using PIXE. In this method, the sample is bombarded with a beam of particles, usually protons, that excites the elements in the sample in a manner similar to that for XRF, causing them to emit X-rays at wavelengths characteristic of the elements (Johansson et al., 1975). Closely related methods of analysis use other ions such as a particles to bombard the sample and induce the X-ray emission. As in the case of XRF, the lighter elements (hydrogen through fluorine) cannot be easily measured with this technique; however, backscattering of a particles used to bombard the sample can be measured, and the energy lost in the nuclear recoil can be used to identify the scattering element for these lighter species. These ion-excited X-ray analytical techniques (IXA) are reviewed by Cahill (1980, 1981a, 1981b) and by Traxel and Watjen (1986). An example of the applica tion to PMI() and PM25 in Brisbane, Australia, is discussed by Chan et al. (1997).

Atomic absorption spectrometry (AA). This is a standard laboratory analytical tool for metals. The metal is extracted into a solution and then vaporized in a flame. A light beam with a wavelength absorbed by the metal of interest passes through the vaporized sample; for example, to measure zinc, a zinc resonance lamp can be used so that the emission and absorbing wavelengths are perfectly matched. The absorption of the light by the sample is measured and Beer's law is applied to quantify the amount present.

Emission spectrometry (ES). Emission spectrometry is based on the excitation of an element to an upper electronically excited state, from which it returns to the ground state by the emission of radiation. As discussed in Chapter 3, the wavelength emitted is characteristic of the emitted species, and, under the approximate conditions, the emission intensity is proportional to its concentration. Means of excitation include arcs and sparks, plasma jets (see ICP), and lasers.

Inductively coupled plasma spectroscopy (ICP). ICP has become a well-established analytical technique for a variety of trace metals. The sample is introduced into a plasma formed by a rf discharge in a gas such as argon. Ions and electrons generated in the plasma are induced to travel in annular paths by interaction with a fluctuating magnetic field generated by the rf induction coil. Elements in the plasma are excited and emit at their characteristic wavelengths. The particular elements can thus be identified from the emission wavelengths and the amounts of each from the emission intensity.

ICP can also be coupled with mass spectrometry (ICP-MS) for very high sensitivity and is finding increasing use for elemental analysis (e.g., Skoog et al., 1998).

Neutron activation (NA). The sample is bombarded with neutrons and the radioactivity induced in the sample is then measured. Both /3 and y radiation can be monitored, but y radiation is more frequently used because of the discrete wavelengths associated with emission that can be used to identify the emitter.

Anodic stripping voltammetry (ASV). This is an electrochemical technique in which the element to be analyzed is first deposited on an electrode and then redissolved, that is, "stripped," from the electrode to form a more concentrated solution. For example, a drop of mercury hanging from a platinum electrode in a solution containing the species to be measured has been used as the deposition electrode. A potential slightly more negative than the half-wave potential for the ion of interest is applied to deposit the element on the electrode. After deposition of the metal for a given time period, stirring of the solution is stopped and the voltage decreased at a constant rate toward the anodic potential while the anodic current is measured. The peak anodic current, corrected for the residual current, is proportional to the elemental concentration under controlled conditions, for example, fixed deposition time.

Mass spectrometry (MS). Mass spectrometry is a common method for detecting and measuring organics (vide infra), but it has also been used for certain inorganic elements and ions as well. For example, Schuetzle et al. (1973) volatilized ambient particulate matter into the source region of a high-resolution mass spectrometer by heating the sample continuously from 20 to 400°C. The elements sulfur, cadmium, and iodine were identified and measured using their masses, ion intensities, and vaporization temperatures, in combination with tCP, MS provides a powerful analytical technique for trace metals.

Secondary ion mass spectrometry (SIMS) and secondary neutral mass spectrometry (SNMS) have also been used for the surface analysis of atmospheric particles. In the SIMS approach, the sample is collected and bombarded with high-energy (keV) atoms or molecules, typically Ar+, causing ejection of material from the surface of the particles into the gas phase. The emitted species include positive and negative ions that are then measured by mass spectrometry. In the SNMS method, the sample is located behind an orifice that contains an rf plasma, for example in argon. The sample holder is held at negative potential, which extracts Ar+ from the plasma and accelerates them toward the sample where they eject surface materials. Neutral species ejected from the surface become partially ionized as they travel back through the plasma, and these are then detected by mass spectrometry (SNMS). The depth analyzed is typically a few monolayers. The application of these techniques to atmospheric particles is described by Klaus (1986) and, as discussed in more detail below, in a series of papers by Goschnick and co-workers (Goschnick et al., 1994a,b; Bentz et al., 1995a, 1995b; Faude and Goschnick, 1997).

Mass spectrometry has also been shown to be a promising method for differentiating the oxidation states of some metals. This is important because the oxidation state in some cases determines the toxicity of the element. For example, Cr(Vl) is a carcinogen whereas Cr(III) is not. Laser ionization mass spectrometry studies of the oxides of chromium and arsenic suggest that some cluster ions are characteristic of the oxidation state. For example, Neubauer et al. (1995) have shown using single-particle laser ionization mass spectrometry (vide infra) that the ratios of ions such as Cr,0,7/Cr20(7, HCr2()7/Cr2Of7, and Cr2057/Cr20,7

can be used to determine the relative amounts of Cr(III) and Cr(VI). However, the relative signal intensities also depended on a number of other parameters such as particle size, water content, laser irradiance, and counterions. As a result, at present this approach is applicable only to well-controlled situations such as process analysis rather than ambient air.

Similarly, Allen et al. (1996) have used laser ionization mass spectrometry to differentiate the oxidation states of arsenic, in this case, a bulk sample was collected on a sampling stage and inserted into the instrument (rather than using single particles). As203, i.e., As(ltl), was shown to give a characteristic As3Oj" ion whereas As205, i.e., As(V), gave an As3Of7 ion. This approach has also been shown to be promising for some organics. For example, it has been used to screen for the presence of nitro-PAHs (see Chapter 10) in diesel exhaust particles (Bezabeh et al., 1997). Quantification was not possible due to such factors as matrix effects.

While these techniques are promising for ambient air analysis, this is clearly going to be complex due to the many different species present in air and the many parameters that affect the ionization process.

Electron spectroscopy for chemical analysis (ESCA / X-ray photoelectron spectroscopy (XPS). The sample is irradiated with X-rays of a fixed frequency, causing ejection of electrons whose kinetic energy is measured. Conservation of energy dictates that the kinetic energy of the electron plus its binding energy must equal the energy of the exciting photon; since the latter is known and the kinetic energy of the electron is measured, the binding energy can be calculated. Since the binding energies are characteristic of each element, this can be used for elemental analysis. In addition, the binding energies of inner-shell electrons are influenced to some extent by the bonding electrons that determine the oxidation state of the element. For example, ESCA was used by Novakov and co-workers (1972) to elucidate the forms of nitrogen and sulfur in atmospheric particulate matter. Similarly, Faude and Goschnick (1997) used XPS to identify a variety of components of aerosol particles in the upper Rhine Valley in Germany, including sulfate and chlorine. Nitrogen in the form of ammonium and organo-nitrogen compounds (but, interestingly, not nitrate) was observed and carbon in organic or elemental as well as in oxidized forms attached to oxygen was noted. Other atmospheric applications are discussed in the review of Cox and Linton (1986).

Intercomparison studies. A number of intercompari-son studies have been carried out to determine the accuracy and precision of measurements of various elements found in particles. For example, Nejedly et al.

(1998) compared the analysis of ambient air particles using ion chromatography, PIXE, and X-ray fluorescence. Two samplers operated side-by-side with PIXE analysis of the filters to assess precision were generally in excellent agreement, to within ~ 10% for a series of elements including Na, Mg, Ti, Cu, Al, Si, Mn, Fe, Ca, Zn, and S as well as for mass of PM25 and the light absorption coefficient.

In another portion of the study, filters were analyzed both by X-ray fluorescence and by PIXE for samples collected in a remote area and in an urban area. The normalized percentage difference compared to the mean of the two was < 10% for the elements S, K, Ca, Mn, Fe, and Zn. However, at small concentrations of sulfur, the X-ray fluorescence data were about 20-30% higher than the PIXE analyses for unknown reasons. The agreement for Si was poorer, perhaps due to the greater absorption of X-rays by Si and/or matrix effects.

In the third portion of the study, the results using five different sampler and analytical method combinations were compared. When obvious outliers were excluded from the data, the normalized percentage differences compared to the mean value for sulfur varied from —21 to +23%. Pairwise comparisons for other elements showed similar variability. The agreement overall for X-ray fluorescence compared to PIXE was good, although there was scatter in the individual measurements, perhaps due to differences in sampling (Nejedly et al., 1998).

Similarly, the concentrations of 17 elements in particles sampled using a variety of methods at Mace Head, Ireland, were compared (Francois et al., 1995). Sampling was carried out using a Hi-Vol sampler, a stacked filter unit, Nucleopore filters, and cascade impactors. Analytical techniques included AA, ICP, NA, PIXE, and X-ray fluorescence. Concentrations obtained using the Hi-Vol sampler were higher, which was attributed to differences in collection efficiencies, particularly of larger particles. Ratios of the concentrations of elements determined using cascade impactors compared to stacked filter units ranged from 0.48 + 0.12 for Na to 1.31 ± 0.19 for Ti, but for most elements were in good agreement. A comparison of two different cascade impactors with PIXE analysis gave relatively good agreement for the total elemental concentrations, although the size distributions in the smaller size range differed for S and Pb. The latter may reflect differences in cutoff diameters for the states in the two impactors and/or bounce and particle reentrainment problems (Francois et al., 1995).

In summary, there is relatively good agreement overall between different methods of elemental analysis for atmospheric particles, with many of the observed dis crepancies due to differences in sampling rather than analysis.

b. Inorganic Ions

Inorganic ions such as NH4, S04~, and NO^ are major components of ambient particulate matter and a wide variety of methods have been used to measure their concentrations. A few of the methods most commonly used are summarized in Table If. 12 and discussed briefly in the following sections.

Colorimetry. A variety of colorimetric techniques have been used to measure ions such as NH|, S04 , and NO^ in ambient particles. For example, nitrate can be measured by reduction to nitrite using hydrazine in the presence of a copper catalyst, followed by its conversion to a colored azo dye, which can be measured by its absorbance at 524 nm (Mullin and Riley, 1955). Sulfate has been determined using an exchange reaction between sulfate and a barium-nitrosulfo-nazo(III) chelate in aqueous acetonitrile; the chelate has an absorbance peak at 642 nm and hence the decrease in this peak can be followed as a measure of the amount of sulfate present that has exchanged with the chelate (Hoffer et al., f 979). Similarly, NH J can be measured by the indophenol blue method (Weather-burn, 1967).

Ion chromatography (IC). Ion chromatography has become one of the most widely used methods for the determination of ion concentrations in ambient particles. As the name implies, ions are separated using ion exchange chromatography and are detected usually using electrical conductivity. For example, sulfate and nitrate can be separated on a column containing a strong basic resin using a carbonate solution as the eluant. To overcome the high conductivity of the elu-ant, which would mask the signal due to the sulfate and

TABLE 11.12 Some Common Methods of Measuring the Major Inorganic Ions in Atmospheric Particles


Analytical methods


Col," ESCA,'' IC,' SIE/' IR'





Source: Adapted from LBL (1979). " Col = colorimetry.

h ESCA = electron spectroscopy for chemical analysis. ' IC = ion chromatography. '' SIE = selective ion electrodes. ' IR = infrared spectroscopy.

' Chemical conversion followed by detection of the product of the NO," reaction.

Source: Adapted from LBL (1979). " Col = colorimetry.

h ESCA = electron spectroscopy for chemical analysis. ' IC = ion chromatography. '' SIE = selective ion electrodes. ' IR = infrared spectroscopy.

' Chemical conversion followed by detection of the product of the NO," reaction.

nitrate, the solutions then pass into a suppression column that contains a strong acid resin; this converts the carbonate into C02 + H20, which has a low conductivity, and the sulfate and nitrate into their acids, which have high conductivities and hence can be easily detected against the suppressed eluant background (Mulik et al., 1976). This eluant suppression was the key to the development of IC to measure sulfate and nitrate. Since this first application of IC in ambient aerosols, a variety of anions and cations in ambient aerosols have been separated and measured using this technique. An example of its application to the measurement of nitrate, sulfate, chloride, and ammonium in PMH) in Taiwan is discussed by Tsai and Perng (1998).

Selective ion electrodes (SIE). Selective ion electrodes are essentially variants of the well-known pH meter. They are membrane indicator types of electrodes in which a potential is developed across a membrane in the presence of the ion; the size of the potential is related to the concentration and hence can be used to quantitatively detect and measure the species. However, instead of a glass membrane, as in the pH meter, the membranes consist of organics that are immersible in water. For example, anion-sensitive electrodes use a solution of an anion exchange resin in an organic solvent; the liquid can be held in the form of a gel, for example, in polyvinyl chloride. The ion reacts with the organic membrane, setting up an equilibrium between the free ion in solution and the ion bound to the membrane, generating a potential difference, which is measured.

Membrane electrodes used to measure species such as NHj that are in equilibrium with the gaseous form (i.e., NH3) in solution are known as gas-sensing electrodes. In this case, the solution to be analyzed is separated from the analyzing solution by a gas-permeable membrane. The gas in the solution to be analyzed diffuses through the membrane and changes the pH of the internal solution, which is monitored using a standard glass electrode.

Infrared and Raman spectroscopy. Stephens and Price (1970, 1972) used infrared spectroscopy to examine both ambient and laboratory-generated aerosols. They identified sulfate, nitrate, and ammonium ion absorption bands in ambient particles as well as bands indicating the presence of organics in diesel exhaust (C-H) and oxidized organics in irradiated hydrocarbon-NOx mixtures. Since then, many studies using IR have been carried out and a variety of species identified, including CO2 , PO41-, and Si044~. See Chapter 9.C.2 and Figs. 9.49, 9.50, and 9.51 for some typical FTIR spectra of atmospheric particles.

A variety of infrared approaches have been used, including transmission IR, photoacoustic IR, diffuse reflectance IR, and attentuated total reflectance. The principles behind these methods and their application to atmospheric aerosols have been reviewed by Allen and Palen (1989).

Raman spectroscopy (reviewed by Schrader, 1986) has also been applied to single particles in laboratory systems. For example, Fung and Tang (1991, 1992a, 1992b) and Fung et al. (1994) have applied resonance Raman spectroscopy to particles containing nitrate and sulfate, both very common constituents of atmospheric particles. The detection limits for nitrate and sulfate in aqueous droplets of 15-/j,m diameter were reported to be about 0.0025 M (Fung et al., 1994), suggesting that this method might prove applicable to ambient particles as well.

A variant of Raman spectroscopy that has been used to probe interfaces in large aqueous particles (e.g., of the order of several hundred microns) in laboratory studies is nonlinear morphology-dependent stimulated Raman scattering (e.g., Zhang and Aker, 1993; Aker and Zhang, 1994). In this method, light generated inside the particle in effect undergoes internal reflections at the interface; when the wavelength of the light is an integral factor times the circumference of the particle, it gets "trapped," in effect increasing the optical path length and hence the net absorption by species dissolved in the particle. As with resonance Raman, this technique has not yet been applied to particles in ambient air.

Mass spectrometry. Laser microprobe mass spectrometry (LMMS) has also been applied to atmospheric particles to measure primarily inorganic elements and ions (e.g., see Bruynseels et al., 1985; Kaufmann, f 986; Wieser and Wurster, 1986; Dierck et al., 1992; and Hara et al., 1996). The particles are collected using techniques such as impactors described earlier and subsequently analyzed. A laser pulse, e.g., at 266 nm, is used to volatilize a selected particle or a group of particles and the gaseous fragments produced are analyzed by mass spectrometry. Both positive and negative ions can be analyzed. The mass spectra of ambient particles can be quite complex and include many fragments as well as clusters. It is because of the extensive fragmentation that specific organic compounds cannot be identifed, although clusters of carbon atoms from soot, for example, can be seen. As discussed shortly, this method has been applied more recently to single particles suspended in air, and typical positive and negative ion spectra are shown in Section B.4a.

c. Total Carbon: Organic versus Graphitic (Elemental)

The separate determination of organic and elemental carbon in atmospheric particles has been addressed in a number of ways by many workers over a period of years; despite this, there is still no accepted accurate and reliable standard method of sampling and analysis for these important aerosol species.

Four major methods have been used to separate the organic elemental carbon: thermal methods, digestion, extraction, and optical techniques. These are discussed in detail in the volume on particulate carbon edited by Wolff and Klimisch (1982) and in the article by Cadle et al. (1983).

In the thermal methods, the sample is heated to increasingly higher temperatures, with most steps being carried out in the presence of 02. The basis of this method is that volatile organics will vaporize first and then other organic compounds will be oxidized. Only at the highest temperatures will graphitic carbon oxidize. The carbon thus ejected into the vapor phase at various temperatures is detected in the form of C02 or, alternatively, after catalytic reduction, as CH4.

For example, in one thermal method, shown in Fig. 11.67, the sample is oxidized and volatilized with an 02-He mixture at 350°C; the volatilized carbon is oxidized to C02 in an Mn02 bed and reduced to CH4 so it can be measured using the sensitive technique of flame ionization detection (FID) (e.g., Huntzicker et al., 1982; Japar et al., 1984; Huffman, 1996). The purge gas is then replaced by pure He and the temperature is raised to 600°C; in this step the remaining organic carbon is volatilized, oxidized to C02 by the Mn02 catalyst, and reduced to CH4 for measurement. Finally, elemental carbon is determined by heating in an 02-He mixture from 400 to 600°C. In this particular apparatus, a light pipe, He-Ne laser, and photocell are used to monitor the reflectance of the filter as an indication of the changes in graphitic carbon on the filter (vide infra).

FIGURE 11.67 Schematic diagram of one type of thermal analyzer for organic and graphite carbon (adapted from Huntzicker et al, 1982).

Although relatively fast and simple, such thermal methods can suffer from the possibility of carbonization of organics during heating in an inert atmosphere; thus elemental carbon can be formed from organic carbon during the analysis, leading to significant errors. Corrections for this can be applied by following the sample reflectance during the heating (e.g., Huffman, 1996). Fung (1990) suggests that this error can be minimized by using another approach in which the sample is oxidized by MnOz during rapid heating to a maximum of 525°C, during which organic carbon is oxidized but elemental carbon is not. Heating to 850°C then leads to oxidation of elemental carbon by Mn02.

A second approach to analyzing organic and elemental carbon has been to digest the sample in a strongly oxidizing solution (e.g., nitric acid) to remove the organics. The remaining carbon on the filter is then measured using standard methods with the assumption that only graphitic carbon remains on the filter after digestion. Organic carbon is then the difference between the total carbon on the filter before and after digestion, respectively. However, it has been shown that during digestion, some elemental carbon is removed, in addition to organic carbon (Cadle et al., 1983). Thus digestion has no clear advantages over thermal methods.

Extraction of the organics from filters using various solvents has also been used. Total carbon analysis of portions of the filter after extraction gives graphitic carbon directly, and organic carbon is obtained by the difference between this and total carbon before extraction (e.g., Japar et al., f984). As with thermal and digestion techniques, there are problems in establishing that the organics and elemental carbon are clearly and accurately separated.

There are a variety of optical methods used to measure graphitic carbon alone, the most widely used being visible light absorption or reflectance techniques. Visible light absorption is the basis of what is known as the integrating plate method (IPM) (Lin et al., 1973), shown schematically in Fig. 11.68. Particles are collected on a Nucleopore filter and inserted between the light source and the detector; the light transmitted through the filter is compared to that transmitted through a clean filter, that is, one not containing particles. Opal glass is placed between the filter and the detector to transmit an isotropic light flux from scattered and transmitted light through the filter. Scattering of light by the particles, which would interfere with an absorption measurement, is minimized by using a filter with a refractive index approximately equal to that of the particles.

As might be expected for a measurement based on simple light absorption in a complex sample, there are

Filter s Particles,

Opal ^glass

Filter s Particles,

Was this article helpful?

0 0

Post a comment