Technical Report No. 3
Environment Australia, May 2002
ISBN 0 6425 4781 5
3. Methods used for heavy metals determination
Determination of metals content in airborne particulate matter involves three major steps. The first is the collection of suspended particulate matter onto appropriate filter papers, the next is the preparation of the filter sample for analysis and the last is the instrumental analysis of the filter contents to determine the chemical composition of the particulate matter. This section will review several methods/techniques reported in the literature on the collection and analyses of particulate matter for their metals composition.
Historically, measurement of particulate matter in air has concentrated on the determination of total suspended particle (TSP) concentration, with no requirements for size selection. However, since researchers started to focus on the health effects of small particles it became necessary to collect particles in different size ranges.
Samplers can be fitted with size-selective inlets (SSI), which exclude particles with aerodynamic sizes greater than 10 µm and retain the others (PM10). Other SSIs retain particles with aerodynamic diameters of 2.5 µm or less (PM2.5), or 1.0 µm or less (PM1). It is thus possible to collect airborne particle matter as TSP, PM10, PM2.5 or PM1 for chemical analysis.
A high volume sampler (hivol) is used for the gravimetric determination of TSP or PM10 in ambient air. The instrument comprises of a protective housing, a high-speed electric motor blower, a filter holder for holding a 203 x 254-mm filter and a flow-controller which controls the flow rate at 1.13-1.70 m³/min. The sampler is normally used to determine total aerosol loading in air, by drawing ambient air through the filter over a period of 24 hours. The filter from the high volume sampler is weighed before and after exposure. Prior to weighing, the filter is conditioned for several hours in a humidified chamber. The flow rate indicator, which is calibrated against a reference orifice meter, determines the total volume of air drawn through the filter-this volume is corrected to standard atmospheric conditions, and used with gravimetric data to determine the concentration of the particles in air (US EPA, 1999b).
Commercial hivols used for collecting particulate matter from ambient air are constructed to meet the minimum requirements of the US EPA reference methods for TSP or PM10 determination (US EPA, 1999b).
The Australian Standard methods AS3580.9.6 (1990) and AS2724.3 (1984) respectively set out procedures for determining PM10 and TSP using hivols. When these samplers are used for PM10 determination, they are fitted with inertia type impact size-selective inlets, which are operated by drawing air through impaction chambers to remove larger particles. The efficiency of removal depends critically on the flow rate.
The dichotomous sampler (dichot) is another particulate sampler that meets the sampling requirements of the US EPA standard reference method (US EPA 1999b). The instrument separates particulate matter into coarse (2.5-10 µm) and fine (<2.5 µm) fractions, and it is a low flow rate sampler (16.7 L/min). During sampling, the inlet removes particles larger than 10 µm by making them impact onto a plate, and inhalable particles are separated into fine and coarse fractions by using a virtual impactor technique. The filter paper used for sample collection can be analysed for the chemical composition of the collected suspended particulate matter.
The dichot operates at a lower flow rate, thus although it may collect a smaller fraction of particles for analyses, it enables the use of filter media such as Teflon, which would otherwise clog quickly at higher flow rates. Both hivols and dichots deposit the particulate matter uniformly across the surface of the filter paper.
Australian Standard Method AS3580.9.7 (1990) describes the gravimetric determination of suspended particulate matter using a PM10 dichotomous sampler.
The Partisol air sampler is a microprocessor-controlled low-volume sampler (16.7 L/min), which measures and logs flow rates and meteorological parameters such as ambient temperature and pressure. The instrument can be fitted with a cyclone type PM10 or PM2.5 SSI. Its microelectronics system provides the user with menu-driven programming, diagnostics and data storage capabilities. However, for gravimetric analysis, the filter has to be conditioned and pre-weighed as was done for filters used in high volume and other samplers. Post-collection chemical analysis of the filter contents can be conducted on several different analytical instruments. The Rupprecht and Patashnick Low-Volume Partisol Air Sampler meets the US EPA requirements for gravimetric determination of suspended particulate matter (US EPA, 1999b).
The MOUDI (Micro-Orifice Uniform Deposit Impactor) is a multi-stage (up to 12) cascade impactor with stages having 50% cut-off points ranging from 0.056 µm to 18 µm in aerodynamic diameter. It collects particles in discrete size ranges by passing the aerosols through a series of impaction stages. Higher flow velocities in each subsequent stage ensure the collection of particles smaller than those collected by the previous stage. The impaction, hence particle collection, depends upon the inertia of the particles. The flow rate, normally 30 L/min, is determined with high precision. Polycarbonate filters, 47mm in diameter with 0.4 µm pore size are used as collection substrates for all the stages except the last. The last stage is a high collection efficiency Teflon-backed filter (37mm) with a 1 µm pore size. (The pore size bears no relationship to the collection efficiency of the Teflon filters for atmospheric particles). The MOUDI enables the simultaneous collection of aerosol fractions over the whole size range that constitutes TSP, PM10, PM2.5 and PM1. Thus mass and chemical composition data can be obtained for all four aerosol-parameters from a single sample, and the sum of all the MOUDI stage collections is a true measure of TSP (Ayers et al., 1998).
Stacked filter samplers draw ambient air sequentially through a stack of filters with different pore sizes, to collect particles with different sizes. The porosity of the filter determines the particle size collected. The ANSTO-SFU (Stacked Filter Unit) was an example of a sequential filter-pack sampler that was used in some Australian studies. This unit operated at 16 L/min, and was fitted with a PM10 inlet to exclude particles larger than 10 µm. It had two filters to collect respectively the PM10-PM2.5 size fraction (on a 47 mm polycarbonate filter with 8 µm pore size), and the PM2.5 fraction (on a 47 mm polycarbonate filter with 0.4 mm pore size) (Ayers et al., 1998).
'Rain gauge type' passive fallout collectors can be used to collect atmospheric particles which are in excess of 10-20 µm in diameter (Noller et al., 1981). Settleable particulate matter is sampled in a high-density polyethylene (HDPE) funnel connected to a HDPE bottle. The containers are acid washed before deployment (Munksgaard and Parry, 1998). Sometimes deposited particulate matter has been collected on Petri dishes. The mass and the composition of the collected suspended particulate matter can be determined. This method is inexpensive and is especially suited for long term monitoring in remote areas. However, it provides information on atmospheric deposition rates of air pollutants, and not on ambient concentrations (Erisman et al., 1998). Deposited matter can be collected using procedures outlined in the Australian Standard Method AS2724.1 (1984). Gulson et al. (1995) have outlined the advantages and disadvantages of this type of monitoring.
The choice of filters for the samplers depends on the type of sampler, the analytical technique and the characteristics of the filter. Filters normally display different particle collection efficiencies, different mechanical and thermal stabilities, varying concentrations of blank impurities, different resistances to airflow and different loading capacities (US EPA, 1999b). All these characteristics will determine the costs and availability of the filter.
The characteristics of common filter types used for particle sampling and post collection chemical analyses has been summarised in Appendix D (adapted from Chow, 1995). The ideal filter type will have high particle-collection efficiency, low blank levels, low flow resistance, low hygroscopic tendencies and reasonable costs.
Filter holders are constructed from polycarbonate, polypropylene, aluminium, stainless steel, Teflon or perfluoroalkoxy Teflon (Chow, 1995). The filters are held in place by screens (support base) and/or silicon or viton O-rings. Large-sized filter papers (203 x 254 mm) are normally used in high volume samplers.
If sample preparation is required prior to chemical analysis, the filter is subjected to hot acid extraction and or microwave extraction to put the metals in solution. Some of the analytical techniques do not require sample extraction from the filters; the filters are cut to fit into the instrument and analysis is performed without the need for any of the destructive techniques described below.
The filter is cut into strips and extracted in a hydrochloric acid (8%)/nitric acid (3%) solution using a microwave oven. The method of extraction, described in the US EPA compendium of inorganic methods (US EPA, 1999b), requires that the microwave oven power should be calibrated so that an identical absolute delivered power can be used by other microwave ovens. Extraction is completed in a digestion vessel using a small volume of water.
Hot acid extraction is recommended when microwave technology is not available. The filter strips are placed in a beaker and extracted by refluxing on a hot plate, using 10 mL hydrochloric acid (8%)/nitric acid (3%) solution. The digestate is filtered before analysis (US EPA, 1999b).
The Australian Standard Method for the determination of lead in particulate matter (AS2800, 1985) recommends digestion with nitric acid (40%) for samples collected in urban areas, which have low levels of insoluble lead.
Airborne particulate matter retained on sampling filters, whether TSP, PM10, PM2.5, PM1 or dichotomous size fractions, can be examined by a variety of instrumental methods for their chemical composition. The common techniques used for metal analyses include:
- Flame Atomic Absorption Spectroscopy (FAAS)
- Graphite Furnace Atomic Absorption Spectroscopy (GFAAS)
- X-ray Fluorescence Spectroscopy (XRF)
- Particle Induced X-ray Emission Spectroscopy (PIXE)
- Inductively Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES)
- Inductively Coupled Plasma-Mass Spectrometry (ICP-MS)
- Neutron Activation Analysis (NAA)
- Cold Vapour Atomic Fluorescence Spectrometry (CVAFS) for Hg
- Ion chromatography for soluble metals
- Electron Microscopy for qualitative analysis
These methods generally provide data on the elemental composition of particulate matter. The analytical techniques differ in terms of sensitivity, requirements for sample preparation, sample throughput and costs of analyses. The US EPA compendium on inorganic analysis has standard procedures for some of these analytical techniques (US EPA, 1999b). The analytical techniques are outlined below, and a comprehensive list of the detection limits for each of the metal species, for the instrumental techniques listed above, can be found in Appendix E.
Atomic absorption spectroscopy (AAS) is a standard laboratory analytical technique in which trace metals in solution are detected in elemental form in a flame. There are two main AAS techniques, flame atomic absorption spectroscopy (FAAS) and graphite furnace atomic absorption spectroscopy (GFAAS). Both techniques are based on similar principles used for measuring metals in solution. However, they differ in the method used for sample introduction into the instrument. In FAAS, the sample is atomised with a nebuliser and introduced into a flame, normally an air/acetylene flame. A graphite furnace electrothermal atomiser is used in GFAAS.
Upon introduction of the metal solution into the instrument, the solution is vaporised by the flame or a furnace, and the trace metal to be detected is dissociated from its chemical bonds into its elemental form. A hollow cathode or electrodeless discharge lamp provides characteristic radiation energy for the metal. The wavelength of this emitted radiation must match the absorption wavelength of the metal to be determined. The amount of energy absorbed by the metal atoms is related to their concentration. Since each metal absorbs light at a characteristic wavelength, analysis for each metal requires a different light source, and only one element can be determined at a time.
Atomic absorption spectroscopy measurements are subject to interference from a number of confounding influences: background, spectral, ionisation, chemical and physical interferences have all been identified. Appropriate choice of filter media and matrix matching of the samples to standards tend to minimise interference (US EPA, 1999b). Overall, AAS has less interference than other techniques used for measuring metals in air.
High-volume samplers are normally used for sampling when FAAS or GFAAS analysis is planned. Both techniques are destructive and require that the sample be extracted or digested before introduction into system as a solution. The detection limit of the GFAAS is normally about two orders of magnitude better than the FAAS (Appendix E).
X-ray fluorescence spectroscopy (XRF) is a very powerful and comparatively inexpensive method for determining elements in airborne particulate matter collected onto filters. The sample on the filter is irradiated with a beam of X-rays. This primary radiation interacts with the elements in the sample to produce vacancies in the inner atomic shells, which then de-excite to produce characteristic secondary X-ray radiation. The wavelengths detected indicate which types of elements are present, and the quantity is determined from the intensity of the X-rays at each characteristic wavelength (US EPA, 1999b).
X-ray fluorescence spectroscopy can be used to determine all elements with atomic weights from 11 (sodium) to 92 (uranium). A typical commercial instrument uses up to seven fluorescers to determine up to 44 chemical elements, and it is normally calibrated with thin metal foils and salts (US EPA 1999b).
In a modification of the XRF technique, called wavelength dispersive analysis, X-ray are used to excite the samples and crystal spectrometers are used to disperse and analyse the characteristic secondary X-rays according to their wavelength. A semiconductor detector converts the energy of the incident secondary X-ray into a voltage pulse whose amplitude is proportional to that energy. The resolution of the semiconductor detector is adequate enough to separate X-ray lines from elements of adjacent atomic numbers. Thus the instrument is capable of performing simultaneous multiple element analysis for typical aerosol samples (Jaklevic, 1977).
The XRF technique is non-destructive and requires minimal sample preparation – the filter (sample) can be inserted directly into the instrument for analysis. Although it is a relatively inexpensive technology, the detection limits are normally higher than other analysis techniques. The detection limits depend upon the filter types used, and concentrations are corrected for filter blanks. A comprehensive list of detection limits for metals is in Appendix E. As a result of the higher detection limits of XRF, Partisol or dichotomous samplers, which sample onto Nylon or Teflon filters, are normally used for sampling when samples are to be analysed by XRF. High-volume samplers normally use quartz-filters and these have high background levels of several elements.
The Particle Induced X-ray Emission spectroscopy (PIXE) technique is used for elemental analysis and is similar to the XRF method. It is an ion beam analysis (IBA) method, which uses beams of energetic ions (eg, high-energy protons) produced by an accelerator, to create inner electron shell vacancies in the target atoms. These inner shell vacancies in the atoms are filled by outer shell electrons, resulting in the emission of characteristic X-rays, which are detected by wavelength detection.
A typical PIXE analysis system consists of an accelerator that produces a proton beam in the range of 2 to 5 MeV, and an optical system and a detector. Energy dispersive detectors are commonly used to determine the mass of elements from sodium to uranium simultaneously. The method detection limits for metals on a Teflon filter are shown in Appendix E. PIXE is relatively inexpensive, it is non-destructive and the filter can be preserved for additional chemical analysis.
There are several other IBA techniques that can be used to complement elemental analysis using PIXE. These techniques are based on the analysis of gamma radiation emitted when the accelerated ion beam interacts with the nucleus of the atom of the target material. Since these techniques target mainly the lighter elements, such as the hydrogen atom (H), they will not be covered here. These techniques have been reviewed in detail in the literature (Cahill, 1992; Cohen, 1992).
Inductively Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES) is used to simultaneously determine the concentration of several trace elements in an acid solution (US EPA, 1999b). The technique is based on the measurement of atomic emission by optical spectroscopy. The sample is introduced into the instrument in solution form, and the atoms in the sample solution are excited with an argon plasma 'torch'. When the excited atoms return to their normal sate, each element type emits a characteristic wavelength of light. The intensities of the wavelengths detected indicate the presence and amounts of specific elements. The plasma is produced by a radio frequency generator, which sends an oscillating current through a coil placed around a quartz tube. The oscillating current produces an oscillating magnetic field that interacts with ions formed in a flowing stream of argon gas in the quartz tube. This results in the formation of a plasma in the form of a toroid or doughnut. The emission spectrum from the plasma is resolved by dispersion with a grating spectrometer, and the relative intensities and concentrations of the elements present calculated.
Up to sixty-one elements can be analysed simultaneously by ICP-AES, at a rate of one sample per minute. The technique allows analysis over a large range of concentrations-up to 5 orders of magnitude (US EPA, 1999b). As with FAAS and GFAAS, the airborne particulate matter sample must be extracted and digested before introduction into the instrument as a solution. Typically, the ICP-AES detection limits for many metals are equal to or just better than those of FAAS, but GFAAS detection limits are better than ICP-AES for most metals (Appendix E). High-volume samplers are normally used for sampling when ICP-AES analysis is planned.
The Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) technique involves the use of an argon plasma torch to generate elemental ions for separation and identification by mass spectrometry (MS). More than 60 elements can be determined simultaneously, including their isotopes (US EPA, 1999b). The sample, in solution, is introduced by nebulisation into a radiofrequency plasma where energy transfer processes take place. The ions are extracted from the plasma through a differentially pumped interface and separated on the basis of their mass-to-charge ratio by a quadrupole mass spectrometer. These mass spectrometers normally have a minimum resolution capability of 1 atomic mass unit peak width at 5% peak height and can determine the isotopes of the elements.
As with FAAS, GFAAS and ICP-AES, the sample has to be extracted and digested before introduction into the instrument in solution form. An ICP-MS instrument has the lowest detection limit of all the instruments described (Appendix E).
Neutron Activation Analysis (NAA), as described in the US EPA compendium of inorganic methods (US EPA, 1999b), involves the bombardment of a sample and a standard with a high neutron thermal flux in a nuclear reactor or accelerator. The elements in the sample are transformed into radioactive isotopes that emit gamma and beta radiation. The gamma rays are more frequently monitored since they are discrete and characteristic of the emitting isotopes. The measured intensities of the gamma rays are proportional to the amounts of elements present. NAA is a simultaneous, multi-element method and does not require significant sample preparation such as extraction or digestion. It is highly sensitive with low detection limits for most metals (Appendix E), but does not quantify elements such as silicon, nickel, cobalt and lead.
Analysis by NAA is compatible with sampling by high-volume, dichotomous and Partisol samplers.
Cold Vapour Atomic Fluorescence Spectrometry (CVAFS) is used to analyse elemental mercury collected from ambient air. Mercury (Hg) exists in air in the vapour and particulate phase. Ambient air is drawn through gold-coated bead traps, at a low flow rate of 0.3 L/min, resulting in the collection of vapour phase Hg by amalgamation. However, as a result of its very low ambient concentration, particulate mercury is collected by trapping onto a glass-fibre filter, through which air is drawn at a higher flow rate of 30 L/min, over 12-24 hours (US EPA, 1999b).
Particulate phase mercury, collected on the glass-fibre filter, is extracted with nitric acid then all forms of mercury in the extract are oxidised to Hg2+ ions using BrCl. The Hg2+ ions are next reduced to volatile Hgo using SnCl2. The liberated mercury is collected on a gold-coated bead trap and determined by dual-amalgamation CVAFS. This involves the thermal desorption of the mercury from the gold-plated bead trap into the CVAFS detector cell where the mercury absorbs incident ultraviolet radiation and fluoresces. The fluorescence signal is detected by a photomultiplier tube that converts the signal to a voltage proportional to the amount of mercury present. Detection limits for particulate mercury is 30 pg/m3 for a 24-hour sample (US EPA, 1999b).
Ion Chromatography (IC) is used for determining cations and anions in solution. It is a chromatographic technique, which uses an ion-exchange column to suppress the detection of the ions of the solution used as the mobile phase. This process leaves only the species of interest to pass to the detector, which is normally a conductance cell. The metal ions that have been routinely determined by IC are K+, Na+, Mg2+ and Ca2+. Thus this method has not been used extensively for heavy metals analysis of suspended particulate matter. It is very suitable for determining polyatomic ions such as sulfate, nitrate, ammonium and phosphate (Sawicki et al., 1978).
Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) can be used to determine the presence of metals in particulate matter collected on filters. Very high magnification of the sample can be obtained when the sample is irradiated with an electron beam. As the electron beam strikes the sample, various signals (such as Auger electrons, X-rays and photons) are released. These signals can be collected to provide detailed information on the morphological features and chemical composition of the particles.
In SEM, high-resolution images of the morphology or topography of the surface of the specimen is produced. In TEM, the electron beam is transmitted through the sample. Thus SEM and TEM can be used to characterise particulate matter in terms of size, shape, distribution and elemental composition. Unfortunately the information obtained is qualitative, since only a limited number of particles are characterised. A sufficient number of particles must be analysed to relate the microscopic characteristics to the bulk properties of the sample in order to achieve a quantitative analysis (Casuccio et al., 1983).
The analytical techniques described above are not the only ones that have been used for determining metals in particulate matter. Electron Spectroscopy for Chemical Analysis (ESCA) also known as X-ray Photoelectron Spectroscopy (XPS); Anode Stripping Voltammetry (ASV)-an electrochemical technique; and Thermal Ionisation Mass Spectrometry (TIMS) can also be used for determining metals in environmental samples such as rocks and soil, and hence particulate matter, to high precision and at low concentrations (LBL, 1979). However, since samples for these techniques require tedious chemical processing before analyses, these techniques are not normally used for the routine analysis of particulate matter in ambient air.
There are no reports of continuous measurement methods that have been used routinely for real-time measurement of multiple metals in ambient air. However, there are continuous emissions measurement instruments used for measuring specific metallic pollutants emitted from industrial stacks at relatively high concentration levels. One such method, Laser-Induced Breakdown Spectroscopy (LIBS), has recently been tested in real-time quantitative analysis of magnesium-based particles in ambient air (Hahn and Lunden, 2000). LIBS involves the vaporisation of airborne particles and molecules in ambient air using a powerful laser beam. A plasma cloud formed is analysed temporally and spectrally with a spectrometer to determine the size of the particles and the concentration of their constituents. Magnesium particles as small as 0.2 µm with a detectable mass a low as 3x10-9 µg were measured in ambient air after a fireworks display (Hahn and Lunden, 2000).