JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 81 Microanalysis of Individual Environmental Particles* Plenary Lecture Ren6 Van Grieken and Chris Xhoffer Department of Chemistry University of Antwerp (U.I.A.) 8-261 0 Antwerp- Wilrijk Belgium Applications of instrumental microanalysis techniques for the characterization of individual particles in environmental samples are reviewed. The principles of electron microprobe analysis and related techniques the micro-version of proton-induced X-ray emission laser microprobe mass spectrometry secondary ion mass spectrometry and Raman microprobe analysis are briefly reviewed and their published applications to aerosols and to aqueous suspensions are described. Keywords Microanalysis; environment; aerosols; electron microprobe; particle analysis Microanalysis and surface analysis techniques are nowa- days very popular in materials sciences corrosion studies microelectronics and other advanced fields of research.To study individual environmental particles however such analytical techniques are very seldom invoked. This is rather surprising in view of the importance of particulate matter in the environment particles carry most material through estuaries to the ocean bottom and through the atmosphere particles influence the global climate and the visibility and many constituents of environmental particles can have negative health effects. Suitable microanalysis techniques can reveal whether a specific element or com- pound is uniformly distributed over all particles of a population or whether it is a component of only a specific group.Sometimes even the element distributions within a particle and the surface enrichments can be inferred. This can facilitate the assignment of particles to specific sources and provide insights into source mechanisms and hetero- geneous surface reactions. However microanalysis techniques tend to be expensive and often it is very difficult to obtain truly quantitative results while analysis of a statistically relevant number of individual particles might be fairly time consuming. A major problem with most microanalysis techniques is that they operate in vacuum so that some loss or transformation of volatile or unstable compounds could occur between the sample introduction and analysis step; it is therefore expected that in the future non-invasive techniques such as Fourier-transform infrared or Raman microspectrometry will play a larger role.Of course each technique has its own specific constraints related to the principle on which it is based (such as sample-beam interaction and experimental set-up). As a consequence these techniques can comple- ment each other with respect to lateral resolution detection limits detectable elements etc. The present review article is based on a computer search of the literature carried out early in 1991 on individual environmental particle analysis by means of instrumental techniques. A brief overview of the most relevant tech- niques and their applications in environmental studies is given. Some aspects of micro- and surface analytical techniques for environmental studies have previously been reviewed by Grasserbauer.*Presented at the XXVIl Colloquium Spectroscopicum Interna- tionale (CSI) Bergen Norway June 9- 14 199 l . Electron Microprobe Analysis and Related Techniques Principles In both electron probe X-ray microanalysis (EPXMA) and scanning electron microscopy (SEM) an electron beam is focused to a nanometre-sized probe and used to excite various signals which can rapidly provide information about composition and surface topography in small areas of the specimen. Secondary electrons are mostly used for imaging and electron micrographs. Backscattered electrons give rise to two types of images a topographical image which shows the roughness of the sample and a composi- tional image which is a visualization of the variation in atomic number with location in the sample.Both the backscattered and secondary electron signals can thus be used for morphology studies. The X-ray photons emitted as a result of the interaction of the electron beam with the specimen atoms can be detected by wavelength- or energy- dispersive spectrometers (WDX and EDX respectively). The signals detected are transformed into electronic pulses and after amplification stored in a multichannel device according to the corresponding wavelength or energy. Characteristic X-rays are superimposed on a rather intense Brems-strahlung continuum background which is the result of non-characteristic emissions from incident electrons interacting with the electrostatic field of the atomic nuclei and inner electron shells. In Table 1 some of the major characteristics of EPXMA are summarized together with those of the other micro-analytical techniques which will be discussed in some detail below.It should be emphasized however that some of the values in this table can vary with the sample type the elements present the instrumental set- up and the goal of the analysis; they should only be regarded as an approximate indication. The theory for EPXMA and SEM-EDX analysis has been described in detail in several text-books. There is no longer a sharp distinction between EPXMA and SEM-EDX SEM was originally used for high-resolution imaging rather than for chemical analysis; EPXMA was primarily developed for achieving quantitative elemental information rather than for imaging purposes.This difference is more a matter of instrumental set-up and practical arrangement of the detectors. Both techinques are now converging to some extent for the purpose of chemical and morphological studies. The study of particulate samples by individual particle analysis requires measurements on a large population set in order to obtain statistically meaningful data. At the Univer-82 JOURNAL OF ANAL.YTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 Table 1 Comparison of the characteristics and performances of some microprobe techniques Parameter EPXMA Micro-PIXE* LAMMSt Excitation by- Detection via- In-depth resolu tion/pm Lateral resolution/pm Elemental coverage (2) Detection limit (ppm) Quantification Molecular information Element mapping Destructive Elect cons Photons 0.5-5 0.1-5 EDX 11-92 lo00 Yes No (EDX) Yes No *Micro-PIXE micro-proton-induced X-ray emission.tLAMMS laser microprobe mass spectrometry. SSIMS secondary ion mass spectrometry. Protons Photons 100 5- 10 EDX 11-92 10 Yes No (EDX) Yes No Photons Ions > 1 1 1-92 10 Difficult Sometimes No Yes SIMSS Ions Ions 0.00 1 1 1-92 1 Difficult Sometimes Yes Yes sity of Antwerp a JEOL Superprobe JXA-733 EPXMA unit is automated with a Tracor Northern TN-2000 system and controlled by an LSI 11/23 minicomputer. The following methodology is generally used for automatic particle recog- nition and characterization (PRC). An electron beam is raster-scanned over a pre-set sample area by means of a digital beam control. A particle is detected when the electron backscattered signal of the closed particle contour points exceeds a pre-set threshold value.The area perimeter and average diameter are calculated. An X-ray spectrum can be accumulated at the centre of the particle or while performing scans across the particle. Thus the PRC program is set up in three sequential steps localizing sizing and chemical characterization after which the beam scans for the next particle. Digital X-ray mapping of one or more elements is also possible by accumulating X-ray signals at a pre-set number of beam spots across the sample area. All the data can be stored for off-line crossing. Automated EPXMA is a very efficient method for analysing many individual particles within a short time. For example 500 particles can be analysed for about 12 elements in less than 2.5 h under optimized working conditions.A relative precision of about 5% can be obtained while the detection limit of EPXMA using EDX analysis is about 0.1 %. Individual particle analysis com- bined with multivariate techniques and/or cluster analysis constitutes a powerful method for discriminating different particle types. So far only a limited number of papers have been published on the use of automated EPXMA in the field of environmental research. Application to Aerosols Electron probe X-ray microanalysis and SEM have recently been applied to study the composition of aerosols with origins ranging from extremely remote to workplace environments. The Antarctic continent is probably the most remote location on earth and the most convenient place to study the composition of background aerosols.Initial analyses of single particles of Antarctic aerosol samples from different locations revealed the following particle types S-rich particles (which may be formed by gas-to-particle conver- sion) sea-salt particles (formed by the bursting of gas bubbles that arise through wave action) aluminosilicates (earth crustal dust or particles originating from local sources such as volcanoes geysers or other surface/ocean- floor disruptions) and particles whose X-ray spectra contain mostly Fe peaks (long-range transported anthropogenic or maybe meteoric dust particle^).^-^ Naturally their relative concentrations vary with sampling site season and meteo- rological conditions. Especially in summer sulfate particles tend to dominate the Antarctic aerosol by number and also by ItoS and Biggg detected H2S04 and (NH4)2S04 particles.Hierarchical and non-hierarchical cluster analyses were performed after automated EPXMA on numerous individual coastal Antarctic aerosols.I0 The results show a domination of marine components in both the fine and the coarse mode fractions. Only a minor crustal component was found. Bigs9 analysed individual particles from Cape Grim (Tasmania) Mauna Loa Observatory (Hawaii) and Point Barrow (Alaska) using SEM and performed chemical tests on them. This approach turned out to be very effective. The great majority of those particles were found to be composed of sulfuric acid or its reaction products with ammonia. The Barrow and Mauna Loa particles were predominantly sulfuric acid while the South Pole and Cape Grim particles were predominantly ammonium sulfate.Aerosols collected from an aircraft in remote continental and marine regions at altitudes ranging from the boundary sea-air interface to the troposphere were analysed with SEM-EDX by Patterson et al." The continental aerosol population consisted of crustal particles with ~ 0 . 5 pm and sulfate aerosols with r t 0 . 5 pm. No significant qualitative differences were noted as a function of altitude. Contrarily Pacific marine measurements showed large variations be- tween the boundary layer and the troposphere. A decrease in the crustal component was observed from the North towards the South. Electron probe X-ray microanalysis combined with an automated image analysis system has been used for the characterization of individual North Sea aerosols.About 2500 particles sampled from a research vessel were sized chemically analysed and classified.'* Sea-salt constituted the most abundant particle type when the collected air masses originated from over the Atlantic Ocean and travelled towards the continent. In contrast in air masses that spent longer residence times over the continent high concentrations of aluminosilicate particles (mostly spheri- cal fly ash particles) carbonaceous particles CaSO and spherical iron oxides were observed. Later analogous EI'XMA characterizations were performed on more than 25 000 individual aerosols collected over the North Sea and the English Channel from a research ve~se1.I~ Differences between samples were studied on the basis of abundance variations using principal component analysis.Nine differ- ent particle types were classified and they were all source- apportioned unambiguously. The release of sea-salt into the atmosphere is dominated by the process of breaking waves and this is more effective as the relative wind speed increases. Transformed sea-salt particles rich in C1 and S are formed by the conversion reaction of NaCl into Na,S04 implying the release of HCl into the atmosphere. Sulfur-rich particles of various composition namely H2S04 (NH4)2S04 (NH4)HS04 and (NH4)3H(S04)2 were assignedJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 83 to anthropogenic sources and they are probably formed by gas-to-particle conversion. The CaSO particles above the North Sea are emitted by anthropogenic sources such as combustion processes they can result from aeolian tran- sport or they can originate from the marine environment.Indeed various dissolved salts begin to crystallize sequenti- ally when sea-water evaporates. Calcite (CaCO,) and dolomite CaMg(C03)2 precipitate first followed by CaSO and the Mg salts. It is possible that CaMg(CO,) undergoes further reaction with gaseous S-rich components. Measure- ments by EPXMA support the existence of such trans- formed particle species. Particles of CaSO may be enriched in S and can therefore be partially identified as CaSO and (NH4)2S0 results from coagulation of CaSO with sub- micrometre sulfate aerosols. Aluminosilicate particles can- not be distinguished from fly ash particles on the basis of their chemical composition only morphology can some- times make the differentiation.Important differences of these typical nearly perfectly spherical fly ash particles were observed in samples taken over the North Sea as air- mass backtrajectories originating from above Eastern Eu- rope. A minor fraction of the quartz can be emitted during combustion processes of coal in power plants.' The titanium particles above the North Sea most probably find their origin on the continent and possible sources are paint spray soil dispersion asphalt production and power plants. In manual EPXMA of Central Pacific aerosols a large concentration of aluminosilicates present as an internal mixture with sea-salt aerosols was observed; these complex particles are more likely to be the result of coagulation of sea-salt and silicate particles within clouds including droplet coalescence1s rather than resuspension of silicate particles from the sea surface as a result of bubble bursting processes.Automated EPXMA of more .than 5000 indivi- dual particles from the Eastern Pacific14 showed that the most abundant particle type was rich in S (45% of all particles) and this in the absence of other detectable elements. Morphological inspection of these particles made it possible to differentiate between two groups namely one group of S-rich particles in the sub-micrometre range that are unstable under the electron beam and are most probably (NH4)2S04 and one S-rich group in the micrometre range (mean particle diameter of 2 pm) showing more spherical contours.The latter group is much less affected by electron irradiation. An important fraction of particles only yielded characteristic Ca and P X-rays; their mean diameter varied between 0.4 and 0.8 pm. Their abundance shows a slight tendency to increase as the sample location approaches the continent but this is insufficient to be able to predict terrestrial sources as being responsible. It is known that the Pacific Ocean is slightly supersaturated in hydroxyapatite [Ca5(P04),0H],16 but this is not sufficient reason to suggest that the ocean is responsible for the production of these aerosols rich in Ca and P. As part of the Global Tropospheric Experiment (GTE) of the US National Aeronauts and Space Administration (NASA) individual aerosol particles sampled over the Amazon Basin were analysed by automated EPXMA in order to study the processes of aerosol and gas emissions by the forest and to assess the chemical mechanisms occurring in the Amazon Basin atmo~phere.~OJ~ About 27% of all particles showed no detectable elements with 2> 10. Most of the particle types could be related to two prevalent local sources soil dust and biologically derived material.The former type is typically composed of Al Si and Ti. The latter type is identified by the high Bremsstrahlung back- ground (low-2-elements) and the presence of elements such as Na S K C1 P Ca and Zn or a combination of them. Particle types containing mainly S K and P can be related to aerosol emissions by vegetation.The composition of urban aerosols is of course highly variable and depends on the geographical location the activities performed locally and the industries surrounding the sampling site. A study involving automated EPXMA on 15 000 aerosol particles from Antwerp Belgium showed soil dust to be the most abundant particle type. Other particle types often found are sulfates (CaSO fine and coarse S-rich particles) automobile exhausts (lead halides and sulfate derivatives) and different anthropogenic par- ticles derived from various sources such as oil burning processes (S V and Ni) abrasion processes (iron and chromium oxides) and emissions from incinerators (Zn- Pb- Cu- Zn- and Sb-rich particle^).'^*^^ The principal source of particulate Pb in the urban atmosphere is the combustion of leaded petrol.A comprehensive study of particulate material in the 0.1-30 pm size range in the urban aerosol of Phoenix AZ USA was conducted by Post and Buseck.20 More than 8000 individual particles were analysed by analytical SEM. The coarse particle fraction (> 1 pm) was mainly crustal material eg. clays quartz feldspar and calcite. A minority of biological material and S compounds and Pb salts' from automobiles were also observed. The sub-micrometre aerosol fraction consists of S-containing particles (60-80%) presumably present as (NH,)2S04. Some of these particles contain various amounts of elements such as Zn Pb Cu Ca Na and K. Analyses of volcanic ash particles by SEM combined with XPS have been reported by several ~ o r k e r s .~ ~ - ~ ~ Major elements detected by SEM-EDX were Al Ca K and Si; and the minor elements were Fe Mg and Ti. Volcanic eruptions have been proved to be responsible for a fraction of the terrestrial particles released directly into the ~tratosphere.~ Particles present in the stratosphere can also be derived from sulfuric acid aerosols sapphires and meteorites.2S They contain A1 metal particles and A1203 spheres. The sub- micrometre regime is dominated by sulfate aerosols of terrestrial originz6 Relatively high concentrations of S are emitted during volcanic eruptions and the presence of thin sulfate gels on the surface of ash particles is probably the result of processes within stratospheric clouds.27 Most often however EPXMA and SEM have been applied to industrial and workplace aerosols and mostly on coal asbestos and fly ash.The name fly ash covers a variety of particles emitted by combustion processes mostly for the generation of electric power. Knowledge of the bulk composition of fly ash is often insufficient because of the important internal composition heterogeneity within a particle population The characteristics of fly ash particles depend on the mineral matter used the thermal behaviour of the coal in the furnace melting and decomposition temperatures of the mineral matter ,and possible chemical reactions and heterogeneous assemblages of the different emission products during their cooling in the atmosphere. Several workers have used SEM and/or EPXMA to deter- mine the morphological and chemical characteristics of fly ash particle^.^^*^^-^^ The main elements present in both the micrometre and sub-micrometre particles are Si Al K Fe Ti Mg and S while Ca P Na C1 and Ni are minor c~nstituents.~~ The major part of fly ash particles have a characteristic spherical geometry although irregularly shaped particles are also o b ~ e r v e d .~ ~ * ~ ~ One should differen- tiate between two fly ash types according to the material (oil or coal) used for the operation of a power plant. Differences between oil fly ash and coal fly ash have been reported by several w o r k e r ~ . l ~ - ~ ~ - ~ ~ Oil fly ash particles vary in morpho- logy from nearly spherical to lacy or spongy lumps which indicate a long exposure history to heat and These spongy structures easily break down to smaller aggregates.Over 90% of the mass fraction occurs in the fine fraction. Oil combustion particles contain considerably more S and substantial amounts of V and Ni. Coal fly ash predominantly consists of smooth mineral spheres and84 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 contains less cenospheres. Almost 90% of the mass fraction occurs in the coarse fraction.44 Application to Aqueous Sediment and Suspension Particles Suspended particulate matter from estuarine and marine environments is being investigated extensively in order to assess sedimentation processes the interactions between sediments and the water column and the physico-chemical reactions that particles undergo. J e d ~ a b ' ~ Skei and Mel~on,~ and Sundby et aL50 have successfully applied SEM-EDX.The common feature of these studies was that the particles were searched and analysed manually for their chemical and morphologi- cal characteristics. The first results of automated EPXMA of marine suspended particulate matter were reported by Bishop and Bi~caye.~' They applied this technique to individual particles from the nepheloid layer in the Atlantic Ocean and classified the analysed particles on the basis of their %:A1 ratio. Part of the data was also described by Lambert et aLS2 Subsequently various aquatic environments have been studied all by using automated EPXMA. For all estuarine systems EPXMA made it possible to evaluate the effect of mixing material from different origins and to separate the mixing process from other processes such as deposition and remobilization.Results of analyses on the Ems estuary (in Germany and The Netherlands) elucidated that the mixing with marine material occurs in the freshwater tidal area and that the suspended matter of marine origin is transported upstream of the tidal zone across the salt wedge.53 The same approach proved to be equally successful when applied to the sediment fraction of the Elbe estuary (Germany) and yielded comparable results.54 For the Garonne and Rhdne rivers (France) no evidence was found for a net flux of marine suspended particulate matter into the estuaries; this is a consequence of the different nature of these e s t u a r i e ~ . ~ ~ - ~ ~ For the Magela Creek river system (Australia) Hart et analysed by automated EPXMA both the suspension ( > 1 pm) and the colloidal (1-0.1 pm) particles the importance of which is increasingly being recognized whereas most previous work had concentrated on the larger sized suspended particulate matter (generally >0.45 pm).The inorganic mineral composition proved also to serve as an equally good tracer for the origin of suspended particu- late matter of non-fluvial estuarine environments. Automated EPXMA of 15 000 particles collected at different locations and depths in the Baltic Sea showed that the abundance variations of the particle types correlated with hydrographical/hydrochemical and bulk data and provided information about geochemical and physical processes that influence the levels and distribution patterns of certain particle types throughout the Baltic and the transient area to the North Sea.s8 A better insight into the sources and lateral/depth dispersal of suspended matter in Makasar Strait and Flores Sea and around the Sumbawa Island (Indonesia) was satisfactorily obtained by EPXMA. It was possible to differentiate between particles of terrestrial volcanic and biogenic rigi in.^^,^^ Manual examinations by EPXMA revealed particle asso- ciations e.g.BaSO formation in recently dead siliceous plankton6* and distinguished different structures e.g. for Mn,62 and different species e,g. for pyrite.63 Dehairs et Particle-induced X-ray Emission Principles Ion beams collimated to a micrometre size can become a very useful tool for individual particle analysis. In scanning proton microprobe (SPM) analysis a proton beam with an energy of 1-3 MeV is finely focused to a diameter of 0.5- 10 pm by means of magnetic quadrupoles and/or electrostatic lenses.A scanning system controls the beam positioning and a computer system collects data from several detectors and beam information. The high-energy proton beam is obtained with cyclotrons or with nuclear electrostatic accelerators such as Van de Graaffs. The recent advent of small commercially available accelerators has greatly expanded the number of experimental facilities capable of performing SPM analysis. In the micro-version of proton- or particle-induced X-ray emission (PIXE) analysis a Si(Li) detector collects X-rays generated for each beam position while the beam scans over the sample and an on-line sorting process makes it possible to obtain real time X-ray iptensity imaging.These elemen- tal maps of the sample being analysed are constructed by the computer on a graphic terminal; they are similar to those obtained with EPXMA instruments. In several SPM set-ups it is possible to observe simultaneously in real time 8-20 elemental maps. Point analysis is also possible with the X-ray spectra stored for off-line quantitative analysis. As the proton beam generally goes through the sample and is collected in a Faraday cup quantification is very easy and matrix effects are few. An accuracy of 10-20% is obtained for absolute analysis at detection limits down to 10 ppm i.c 2 or 3 orders of magnitude below those of EPXMA. Johansson and Campbell6 have reviewed PIXE in depth.Some of the characteristics of micro-PIXE are outlined in Table 1. There are several other processes occurring during the interaction of the proton beam with the sample. Backscat- tered particles provide information on light elements such as C N and 0 through 'Rutherford back scattering' (RBS) analysis,6S while the gamma rays generated from nuclear reactions make it possible to measure F Na and other elements by particle-induced gamma emission (PIGE).66 Frequently PIXE and RBS analysis are performed simulta- neously allowing the determination of C N and 0 together with 10-15 trace elements that are heavier than Na. Application to Aerosols The SPM is a recently developed technique and most of the applications are in biology archaeology geology and material sciences.Artaxo et al.67 showed the feasibility of a combined approach of SPM (using the facility at the University of Oxford) automated EPXMA and laser microprobe mass spectrometry (LAMMS) characterization of individual aerosol particles from the Amazon region. Using RBS it was also possible to make elemental maps of C N and 0 in real time simultaneously with elemental maps for the trace elements. This has been shown to be very useful in measuring the stoichiometry of compounds in atmospheric aerosol particles.67 Vis et a1.68 have analysed fly ash particles with the SPM system of the Free University of Amsterdam. A beam size of 7 x 10 pm was used with a current of 20-40 PA. It was possible to measure trace elements such as Se V Cr Ti and Cu and to obtain concentration profiles for large particles.The analysis was complemented by tube-excited X-ray fluorescence for bulk trace element measurements. Application to Aqueous Sediment and Suspension Particles The Hamburg SPM group has measured trace elements in particles from river sediments.69 Using a 2 MeV proton beam of 2.3 x 3.0 pm and beam currents of 0.3-3 nA the detection limits were about 10 ppm. It was possible to detect Si S K Ca Ti V Cr Mn Fe Ni Cu Zn As Br,JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 85 Rb Sr Zr and Pb. Using different absorbers the SPM analysis could be optimized for a certain range of elements further increasing the sensitivity for heavier elements. Laser Microprobe Mass Spectrometry Principles Laser microprobe mass spectrometry is based on the mass spectrometric analysis of ions formed by the interaction of the sample with a high power density pulsed laser beam.Verbueken et a/.70 have published an overview of LAMMS techniques. Several instruments have been commercialized; e.g. the LIMA-2A of Cambridge Mass Spectrometry Cambridge UK and the LAMMA- 1000 and LAMMA-500 of Leybold-Heraus Cologne Germany. In the last instrument a Nd:YAG laser generates very short and intense light pulses for vaporization and ionization of a microvolume of the sample. The power density is 1 x 1 07-l x 10’ W cm-2 for a 1 pm laser focus and it can be reduced to 2% of its initial value by a 25-step attenuating filter system. This is especially interesting for particle surface analysis. Depending on the spectrum polarity chosen positive or negative ions are accelerated by a potential of 3000 V into a field-free drift region of the time- of-flight mass spectrometer. The time-of-flight needed by an ion to traverse this region is related to its mass-to-charge ratio.The signal is then fed into a 32 kbyte memory transient recorder and digitized. Spectra are stored in a personal computer for off-line data handling. Software packages are available for data processing and include a baseline correction algorithm a peak integration routine and spectrum averaging facilities. The commercially avail- able instruments differ in the geometry for the collection of ions from the specimen. Generally the LAMMS technique has various interesting features it can detect all elements and compared with other microchemical techniques detection limits are fairly good. It can give indications concerning stoichiometry and information about several organic compounds of environ- mental importance.Disadvantages are the facts that the technique is destructive and rather irreproducible and that the theoretical aspects of ion formation and behaviour in the system are not yet eludicated. Some characteristics of LAMMS have been summarized in Table 1. Applications of LAMMS in medicine biology and environmental research have been reviewed by Verbueken et aL7’ Application to Aerosols The LAMMS technique has been applied to a number of representative particles from different environments. Much attention has focused on marine aerosols. The most typical marine aerosols are sea-salt particles formed by the bubble bursting mechanism.The LAMMS spectra of ‘pure’ sea-salt are dominated by Na K and typical Na-K-CI cluster ions. However in the North Sea environment sea-salt particles are often transformed to some extent nitrate and sulfate coatings are readily dete~table.~~ Otten et a/.73 found the relative abundances of ammonium-rich particles in the North Sea aerosol to increase dramatically under the influence of polluted air masses. Bruynseels et a/.74 also found the amount of nitrate coated sea-salt particles to increase significantly from a beach site in Brazil towards an industrialized area 30 km downwind from the ocean. The detection of methane sulfonate a biogenic airborne organic compound above the Sargasso Sea and the Bahamas area75 and the coast of constitutes an excellent illustration of the occasional ‘organic successes’ of the LAMMS technique.The marine aerosol of Cape Grim (Tasmania) was also analysed with LAMMS by Surkyn et al.,77 who found sea-salt derived and exceptionally crust derived particles. Sheridan and Mu~selman~~ performed LAMMS and EPXMA on particles sampled during flights over the Alaskan Arctic. Virtually all sub-micrometre particles yielded spectra that highly resembled those of an ammon- ium sulfate standard. Because the likelihood of finding appreciable amounts of ammonium vapour in the winter Arctic atmosphere is small they concluded that those particles were collected as sulfuric acid and gradually transformed in the laboratory. Potassium-rich particles in this aerosol were tentatively attributed to wood combustion.Surkyn et performed LAMMS on aerosol particles from the uplands of central Bolivia. The sampling station was located at 5230 m above sea level. The immediate surroundings are totally uninhabited and the ground over a wide area is stony partly snow covered year-round and without vegetation. Particles from this site were soil derived aluminosilicates and/or Ca-rich particles and in the smal- lest size fraction ammonium sulfates. Occasionally K- and C-rich particles were detected; most probably they resulted from forest burning. The LAMMS spectra of Amazon Basin aerosols are very complex due to the presence of different organic com- pounds fragmented to various extents sulfate salts of amines methane sulfonate and fragmentation patterns of hydrocarbons terpenes and phospholipids have so far been Part of this organic material was found to be associated with inorganic salt mixtures consisting of plant nutrients which points to a plant-transpiration origin.Another interesting result is the association of some trace elements (e.g. Pb and Zn) with the organics. Laser microprobe mass spectrometry was performed on individual particles containing Pb sampled near the city of Antwerp Belgium. l 9 The results indicated that partial conversion of lead halide containing particles into lead sulfates often occurs by the reaction with ammonium sulfate present in the urban atmosphere. The ammonium sulfate can also be present as a coating on the particles containing Pb.Barths et investigated airborne dust from 1 1 Euro- pean coal mines. Correlating LAMMS results with toxicity data they could confirm the role of quartz as a specific toxic agent for German but not for French coal samples. Cluster analysis of the element distribution patterns revealed factors which clearly modulate the quartz-related toxicity and also factors with their own toxic potency. These factors seem to be mine dependent or at least area dependent. The cytotoxicity of different silica dusts was found to be primarily determined by the incidence of Si-dominated particles. The latter turned out to be a better cytotoxic parameter than the quartz content as determined by bulk analysis. Their results support the idea that some fraction of the quartz is toxicologically ineffective.On the basis of LAMMS spectra of individual particles different asbestos types which are known to promote fibrosis and/or cancer can be distinguished.81 The tech- nique has also been applied to the analysis of organic impurities at the surface of asbestos fibres.82 The absorption behaviour of different asbestos varieties for various organ- ics was also s t ~ d i e d . ~ ~ . ~ ~ The LAMMS revealed preferential leaching of elements in for example biological liquids. These results are of importance in the sense that next to fibre geometry chemical properties and reactivities also determine the carcinogenic effect. In an attempt to identify Ni compounds emanating from pollution sources Musselman et succeeded in distingu- ishing different (standard) Ni species on the basis of their LAMMS spectra.Gondouin and Miillera6 and Poitevin et used86 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 LAMMS to infer the oxidation state of Cr in dust particles formed during stainless-steel machining and soldering operations. Their stoichiometric information which is important from a toxicological point of view was based on relative cluster ion intensities in the LAMMS spectra. Michaud,88 who analysed particles containing Cr from pigmentation soldering and plating industries showed that these ratios are extremely dependent on instrumental fluctuations so standards should be analysed on a regular basis. The anthropogenic Cr particles appeared most of the time in the hexavalent (k the most harmful) oxidation state.It should however be stressed that in general obtaining stoichiometric results with LAMMS is by no means straightforward. Polycyclic aromatic hydrocarbons were also detected using LAMMS on soot particles from an experimental oil- shale retort.89 Application to Aqueous Suspension Particles Laser microprobe mass spectrometry has been used to study the trace element composition and surface characteristics of suspended matter particles from the Atlantic Ocean and the Scheldt river estuary (Belgium and The nether land^).^^ The Fe-rich phase appeared to contain significant amounts of trace elements such as Ba Cr and Pb and of phosphate. Aluminosilicate particles from the Atlantic Ocean do not contain detectable amounts of Pb but those from the Scheldt do while Ba was found to be associated with the aluminosilicates from both samples.Inferring the compo- sition of surface layers using LAMMS in the 'desorption mode' appeared difficult; the spectra supported but could not prove the existence of for example surface coatings of CaCO on Si-rich particles. The use of model systems appeared to be necessary to identify surface layers unambi- guously. Secondary Ion Mass Spectrometry Principles Secondary ion mass spectrometry (SIMS) is based on the bombardment of a sample surface by primary ions (Ar+ F- 0- etc.) generated in a duoplasmatron (energies in the kiloelectron volt range). A small fraction of the sputtered atoms are charged. These secondary ions are attracted to a mass spectrometer where they are separated according to their mass to charge ratio.The mass spectrometer is based on electric/magnetic deflection fields or on the quadru- polehime-of-flight principle; the latter being cheaper but less satisfactory with respect to mass resolution. Lodding91 distinguished three classes of SIMS instrumentation (a) non-imaging probes (static SIMS) used for depth profiling on laterally homogeneous specimens or for surface analysis; (6) imaging ion microprobes (dynamic SIMS) which use a narrow (< 10 pm) beam of primary ions at energies of 5-20 keV and allow imaging and microscopy by rastering the beam over the sample surface; and (c) direct imaging microscope microanalysers which use wide (5-300 pm) primary beams. The absolute detection limit for SIMS analysis is about 1 x lO-I5 g for most elements and chemical compounds and for anions down to 1 x g.92 When molecular information of particles is needed both LAMMS and SIMS can be used as they exhibit qualitatively the same positive ions.Special capabilities are offered by SIMS for particle analysis. Ion specific images of elemental or molecular constituents can be obtained in the ion microscope or ion micr~probe.~~ The limiting lateral resolution is 0.5- I pm for the ion microscope and about 1 pm for the scanning ion microprobe. The determination of constituents with depth in a particle with a resolution in the nanometre range is another microstructural feature of interest. The capabilities of SIMS for the detection of all elements fingerprinting of compounds isotope ratio measurements depth profiling and ion imaging of specific constituents are described by N e ~ b u r y ~ ~ with special reference to particle studies.Some characteristics of SIMS are included in Table 1. Application to Aerosols Depth profile studies using SIMS of small coal fly ash particles by von Rosenstiel et ~ 1 . ~ ~ showed a significant surface enrichment of Pb. Similar studies by Linton and co- w o r k e r ~ ~ ~ ~ ~ ~ indicated strong surface enrichments of Pb and T1; this implies that coal fly ash may have a more deleterious environmental impact than is apparent solely on the basis of conventional bulk analysis. Cox et investigated particles from a coal-fired power plant with a digital imaging system interfaced to an ion microscope. The set-up they used permitted the simultaneous acquisition of spatially resolved mass spectral data for a number of single particles.These workers found substantial differences in the relative concentrations and/or depth profiles of Ba Pb Si 'Th T1 and U from particle to particle. Lead T1 and U were generally concentrated on the particle surfaces. The SIMS analysis of single oil-soot particles showed that they are characterized by high levels of 0 V C Na Ca and K.99 have reported elemental distributions as a function of the sputter depth by SIMS from large (> 10pm) particles of automobile exhausts. Enrichments of Pb and Br ;and less obvious of S at the particle surfaces were found. 'The SIMS technique has also been applied for the analysis of organic impurities at the surface of asbestos fibres.lol Keyser ef Micro-Raman Spectrometry Principles Raman spectrometry and microspectrometry are based on 1he Raman effect.When photons of frequency vo hit molecules most of them are scattered elastically (Rayleigh scattering). The Raman scattering is caused by the inelastic collision of photons and molecules resulting in photons with a series of frequencies related to the original fre- quency vo by the expression v o + f . The Raman frequency shift f is independent of the incoming radiation and corresponds to certain rotational vibrational and electronic levels of the molecules under investigation. Depending on the symmetry of these molecules vibrations are infrared active Raman active neither or both so sometimes infrared and Raman are complementary. Moreover water i s a suitable solvent for Raman spectroscopy but not for infrared spectroscopy.The advent of lasers as sources for the excitation of Raman spectra and the developments in instrumentation optics now allow analysis of discrete microsamples pro- vided one can meet some specific technical requirements as have been described in the literature. The scope and hmitations of single particle analysis by Raman microprobe spectrometry have been demonstrated by several work- ers.102-106 At the National Institute of Standards and Technology a Raman microprobe has been developedIo7 specifically for the analysis of micrometre sized particles. At the same time a commercial Raman microprobe/micro- scope MOLElo8 became available. Knoll et al.lo9 and E;ieferI1O reviewed micro-Raman spectrometry of particles.Ir is possible to obtain Raman spectra of particles whose sizes are of the order of or larger than the wavelength of the exciting light. When these particles have a well defined geometry the spectra can be seriously distorted by peaksJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 87 which arise due to magnetic vibrations of the particle morphology-dependent structural resonance modes occur. These additional peaks allow the user to determine the size of the particle accurately but on the other hand they complicate the assignment of peaks to molecular vibra- tional modes. Thurn and Kieferlll showed that this effect only occurs on particles with a well defined geometry and not on microcrystals. Applications to Aerosols As mentioned previously EPXMA has revealed that the major fine particle aerosol type in Antarctica especially in summer is rich in S.Raman microprobe measurements have shown that this component exists predominantly in the form of H2S04 (NH4)HS04 (NH4)2S04 or a mixture possibly including more complex species. Io3 Lang et af.Il2 studied individual dust particles (ranging from 10 to 50 pm) from an office-laboratory environment using infrared and Raman spectrometry. Raman spectra were obtained on dust specimens if the infrared spectra did not provide sufficient information to permit suitable characterization of the sample. Many of the dust particles identified could be linked to a paper product as the source. Etz and c ~ - w o r k e r s ~ ~ ~ J ~ ~ have made extensive measure- ments on particulate material from oil- and coal-fired power plants using a Raman microprobe.They found the exis- tence of V2OS as a principal component in the oil ash particles but not in the coal derived particles. In fact it is highly surprising that although Raman microprobe analysis has been available for more than a decade and offers a fantastic potential so few publications on its environmental application are available. Other Micro-analysis Techniques Although numerous other analysis techniques exist which provide microscopic analysis capability and which are used extensively in various research fields their applications to environmental samples are still extremely scarce or even non-existant. This includes Auger electron spectrometry X- ray photo-electron spectrometry and Fourier transform infrared spectrometry in the micro-version.One technique which has only recently been commer- cialized and is still at the experimental stage for environ- mental studies is electron energy loss spectrometry (EELS). Hitherto few applications of EELS analysis to individual particles have been r e p ~ r t e d ~ l ~ J l ~ but include the adsorp- tion behaviour of different varieties of asbestos for various organic substances. * This work was partially supported by the Belgian Ministry of Science Policy (under contract Eurotrac EU 7/08). References Grasserbauer M. Mikrochim. Acta 1983 111 4 15. Cadle R. Fisher W. Frank E. and Lodge J. J. Atmos. Sci. 1967 25 100. Parungo F. Bodhaine B. and Bortnak J. J. Aerosol Sci.1981 12 367. Shaw G. E. Atmos. Environ. 1983 17 329. Ito T. J. Atmos. Chem. 1985 3 69. Bodhaine B. A. and Murphy M. E'. J. AerosolSci. 1980,11 305. Bodhaine B. A. Deluisi J. J. and Hams J. M. Tellus Ser. B 1986 38 223. Cunningham W. and Zoller W. J. Aerosol Sci. 1981 12 367. B i g E. J. Appl. Meteorol. 1980 19 521. 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 Storms H. Ph.D. Thesis University of Antwerp 1988. Patterson E. M. Kiang C. S. Delany A. C. Artburg A. F. Leslie A. C. D. and Huebert B. J. J. Geophys. Res. 1980 8 736 1. Bruynseels F. Storms H. Van Grieken R. and Van Der Auwera L. Atmos. Environ. 1988 22 2593. Xhoffer C. Bernard P. Van Grieken R. and Van der Auwera L. Environ. Sci.Technol. 1991 25 1470. Xhoffer C. M.Sc. Thesis University of Antwerp 1987. Andreae M. O. Charlson R. J. Bruynseels F. Storms H. Van Grieken R. and Maenhaut W. Science 1986 232 1620. Riley J. P. and Chester R. Introduction to Marine Chemis- try Academic Press London and New York 197 1. Artaxo P. Storms H. Bruynseels F. Van Grieken R. and Maenhaut W. J. Geophys. Res. 1988 93 1605. Van Borm W. Adams F. and Maenhaut W. Atmos. Environ. 1989 23 1139. Van Borm W. Wouters L. Van Grieken R. and Adams F. Sci. Total. Environ. 1990 90 5 5 . Post J. E. and Buseck P. R. Environ. Sci. Technol. 1985,18 35. Fruchter J. Science 1980 209 1 116. Wightman J. P. Colloids Surf. 1982 4 401. Gooding J. L. Clanton U. S. Gabel E. M. and Warren J. L. Geophys. Res. Left. 1983 10 1033.Rampino M. R. and Self S. Sci. Am. 1984 250 48. Brownlee D. E. Microbeam Anal. 1980 199. Cadle R. D. and Grams G. W. Rev. Geophys. Space Phys. 1975 13 475. Mackinnon 1. D. R. Gooding J. L. McKay D. S. and Clanton U. S. J. Volcanol. Geotherm. Rex 1984 23 125. Bonafede; G. and Kiss L. T. Am. SOC. Mech. Eng. 73- Baker J. E. Evans C. A. Loh A. and Natusch D. F. S. Proc. Annu. C0n.f-Microbeam Anal. SOC. 1975 10 33A. Jan de Zeeuw H. and Abresch R. V. Proceedings of the Fourth International Ash Utilization Symposium ERDA Morganstown WV 1976 p. 386. Chiu A. S. Thesis Department of Chemical Engineering New Hampshire University Durham 1978. Gibbon D. L. Scanning Electron Microsc.. 1979 501. Carpenter R. L. Clark R. D. and Su Y.-F. J. Air Poflut. Control Ass. 1980 30 679.Lichtman D. and Mroczkowski S. Environ. Sci. Technol. 1985 19 274. Middleman L. M. and Geller J. D. Scanning Electron Microsc. 1976 17 1. Small J. A. and Zoller W. H. NBS Monogr. (US.) 1977 464,651. Fisher G. L. Prentice B. A. Silberman D. and Ondov J. M. Environ. Sci. Technol. 1978 12 447. Parungo F. Ackerman E. Proulx H. and Pueschel R. Atmos. Environ. 1978 12 929. Hayes T. L. Pawley J. B. and ,Fisher G. L. Scanning Electron Microsc. 1978 239. Capron R. Haymann P. and Pellerin F. C. R. Hebd. Seances. Acad. Sci. 1979 C289 313. Kaufherr N. and Lichtrran D. Environ. Sci. Technol. 1984 18 544. Wagman J. Colloid and Interface Science Volume II Aerosols Emulsions and Surfactants Academic Press New York and London 1976. Denoyer E. Mauney T. Natusch D. F. S. and Adams F.MAS 1982 191. Mamane Y. Miller J. L. and Dzubay T. G. Atmos. Environ. 1986 20 2125. Raeymaekers B. Ph.D. Thesis University of Antwerp 1987. McCrone W. C. and Delly J. G. The Particle Atlas Ann Arbor Science Publishers Ann Arbor MI 2nd edn. 1973 VOl. 2 p. 543. Dehairs F. Chesselet R. and Jedwab J. Earth Planet. Sci. Lett. 1980 49 528. Jedwab J. Earth Planet. Sci. Lett. 1980 49 55 1. Skei J. M. and Melson S. Estuarine Coast. ShelfSci. 1982 14 61. Sundby B. N. Silverberg N. and Chesselet R. Geochirn. Cosmochim. Acta 1984 45 293. WNCD-7 1973.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 88 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 Bishop J. K. B. and Biskaye P. E. Earth Planet. Sci.Lett. 1982 58 265. Lambert C. E. Bishop. J. K. B. Biscaye P. E. andchesselet R. Earth Planet. Sci. Lett. 1984 70 237. Bernard P. Van Grieken R. Eisma D. and Hudec B. Environ. Sci. Technol. 1986 20 467. Van Put A. Van Grieken R. and Wilken R. D. Environ. Sci. Technol. submitted for publication. Eisma D. Bernard P. Boon J. Van Grieken R. Kalf J. and Mook W. Mitt. Geol. Palaeontol. Inst. Univ. Hamburg 1985 58 397. Eisma D. Bernard P. Cadee G. Ittekkot V. Kalf J. Laane R. Martin J. Mook. W. Van Put A. and Schuma- cher T. Neth. J. Sea Res. 1991 28 193. Hart B. Douglas G. Beckett R. Van Put A. and Van G rie ken R. Hydrological Processes submitted for pu blica- tion. Bernard P. Van Grieken R. and Briigmann L. Mar. Chem. 1989 26 155. Eisma D. Kalf. J. Karmini M. Mook W.Van Put A. and Van Grieken R. Neth. J. Sea Res. 1989 24 383. Eisma D. Van Put A. and Van Grieken R. Neth. J. Sea Res. submitted for publication. Bishop J. Nature (London) 1988 332 341. Middelburg J. De Lange G. Van der Sloot H. Van Emburg P. and Sophiah S. Mar. Chem. 1989 23 353. Luther G. Meyerson A. Krajewski J. and Heres R. J. Sediment. Petrol. 1980 50 1 1 1 7. Johansson S. A. E. and Campbell J. L. PIXE-A Novel Technique for Elemental Analysis Wiley New York 1988. Finstad T. G. and Chu W. K. in Ion Beam Techniques Analytical Techniques for Thin Films eds. Tu K. N. and Rosemberg R. Academic Press New York and London 1988. Ion Beams for Material Analysis eds. Bird J. R. and Williams J. S. Academic Press New York and London 1989. Artaxo P. Van Grieken R. Watt F.and Jaksic M. Proceedings of the Second World Congress on Particle Techno- logy Kyoto Japan 1990 421. Vis R. D. Bos A. J. J. Valkovic V. and Verheul H. IEEE Trans. Nucl. Sci. Vol. NS-30 1983 2 1236. Grossmann D. Kersten M. Niecke M. and Puskeppel A. SCOPEKJNEP Sonderband He# 1985,58,6 1 9. Verbueken A. H. Bruynseels F. J. Van Grieken R. and Adams F. Inorganic Mass Spectrometry eds. Adams F. Gijbels R. and Van Grieken R. Wiley New York 1988 173. Verbueken A. H. Bruynseels F. J. and Van Grieken R. Biomed. Mass. Spectrom. 1985 12 438. Bruynseels F. and Van Grieken R. Atmos. Environ. 1985 19 1969. Otten P. Bruynseels F. and Van Grieken R. Anal. Chim. Acta 1987 195 1 17. Bruynseels F. Storms H. Tavares T. and Van Grieken R. Int. J. Environ. Anal. Chem. 1985 23 1.Kolaitis L. Bruynseels F. Van Grieken R. and Andreae M. Environ. Sci. Technol. 1989 23 236. Wouters L. Artaxo P. and Van Grieken R. Int. J. Environ. Anal. Chem. 1990 38 427. Surkyn P. De Waele J. and Adams F. Int. J. Environ. Anal. Chem. 1983 13,257. Sheridan P. J. and Musselman I. H. Atmos. Environ. 1985 19 2 159. Bruynseels F. Artaxo P. Storms H. and Van Grieken R. Microbeam Anal. 1987 356. Barths G. Schmidtz G. Kaufmann R. Bruch J. and Tourmann J. Inst. Natl. Sante. Rech. Med. [Colloq.] 1987. De Waele J. Van Espen P. Vansant E. and Adams F. Microbeam Anal. 1982. 37 1. De Waele J. Vansant E. Van Espen P. and Adams F. Anal. Chem. 1983 55 67 1. De Waele J. Vansant E. and Adams F. Mikrochim. Acta 1983 367. De Waele J. Verhaert I. Vansant E. and Adams F. SIA 85 Musselman I.Linton R. and Simons D. MAS 1985 337. 86 Gondouin S. and Muller J. Proceedings of the Third International Laser Microprobe Mass Spectrometry Workshop University of Antwerp Antwerp 1986. 87 Poitevin E. Muller J. Klein F. and Dechelette O. Analusis 1989 17 47. 88 Michaud D. personal communication I99 1. 89 Mauney T. and Adams F. Sci. Total Environ. 1984 36 215. 90 Wouters L. Bernard P. and Van Grieken R. Int. J. Environ. Anal. Chem. 1988 34 17. 91 Lodding A. Inorganic Mass Spectrometry eds. Adams F. Gijbels R. and Van Grieken R. Wiley New York 1988 p. 125. 92 Benninghoven A. Appl. Phys. 1982 1 3. 93 Morrison G. and Slodzian G. Anal. Chem. l975,47,932A. 94 Newbury D. E. Characterization of Particles NBS Spec. Publ. (U.S.) 533 Proceedings of the Particle Analysis Session of the 13th Annual Conference of the Microbeam Analysis Society Ann Arbor MI June 22 1978 ed.Heinrich K. F. J. 1980 p. 139. 95 von Rosenstiel A. P. Gay A. J. and Van Duin P. J. Direktabb. Ober-. 1981 14 153. 96 Linton R. W. Loh A. Natusch D. and Williams P. Science 1975 191 853 97 Linton R. W. Williams P. Evans C. A. and Natusch D. F. S. Anal. Chem. 1977 49 1514. 98 Cox X. B. 111 Bryan S. R. and Linton R. W. Anal. Chem. 1987,59,20 18. 99 McHugh J. and Stevens F. Anal. Chem. 1972 44 2187. 100 Keyser T. R. Natusch D. F. S. Evans C. A. Jr. and Linton R. W. Environ. Sci. Technol. 1978 12 768. 101 Van Espen P. De Waele J. Vansant E. and Adams F. Int. J. Mass Spectrom. Ion Phys. 1983 46 515. 102 Etz E. and Rosasco G. Proceedings of the 5th International Conference on Raman Spectroscopy Hans Ferdinand Schulz Verlag Freiburg 1976 pp.776-777. 103 Etz E. and Rosasco G. NBS Spec. Publ. (US.) 1977 464 343. 104 Rosasco G. J. Advances in Infrared and Raman Spectroscopy eds. Clark R. J. H. and Hester R. E. Heyden London 1980 vol. 7 223. 105 Etz E. and Blaha J. Characterization of Particles Proceed- ings of the Particle Analysis Session of the 13th International Conference of the Microbeam Analysis Society Ann Arbor MI June 22 ed. Heinrich K. 1980 p. 153. 106 Purcell F. and Etz E. Microbeam Anal. 1982 17 301. 107 Rosasco G. J. Etz E. S. and Cassatt W. A. Appl. Spectrosc. 108 Dhamelincourt P. Wallart F. Leclercq M. N'Guyen A. 109 Knoll P. March] M. and Kiefer W. Indian J. Pure Appl. I10 Kiefer W. Croat. Chem. Acta CCACAA 1988 61 473. I 11 Thurn R. and Kiefer W. Appl. Spectrosc. 1984 38 78. I12 Lang P. Katon J. and Bonanno A. Appl. Spectrosc. 1988 42 313. I13 Etz E. Rosasco G. and Blaha J. Environmental Pollutants Detection and Measurements eds. Toribara T. Coleman J. Dahneke B. and Feldman I. Plenum New York 1978 p. 413. 114 Etz E. Rosasco G. and Heinrich K. EPA Report EPA- 115 Wolf B. 2. Naturjosch. C Biosci. 1988 43 155. 1 16 Xhoffer C. Lathen C. Van Borm W. Broekaert J. A. C. Jacob W. and Van Grieken R. Spectrochim. Acta Part B 1992 47 155. 1 17 Xhoffer C. Berghmans P. Muir I. Jacob W. Van Grieken R. and Adams F. J. Microsc. 1991 162 179. 1975 29 396. and Landon D. Anal. Chem. 1979 51,414A. Phys. 1988 26 268. 600/2-78-193 August 1978. Paper I /0545 7K Received October 28 1991 Surf Interface Anal. 1983 5 186. Accepted December 30 1991