|
11. |
Microanalysis of individual environmental particles. Plenary lecture |
|
Journal of Analytical Atomic Spectrometry,
Volume 7,
Issue 2,
1992,
Page 81-88
René Van Grieken,
Preview
|
PDF (1372KB)
|
|
摘要:
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
ISSN:0267-9477
DOI:10.1039/JA9920700081
出版商:RSC
年代:1992
数据来源: RSC
|
12. |
Photon detection based on pulsed, laser-enhanced ionization and photoionization of magnesium vapour: quantum efficiencyversusion yield. Invited lecture |
|
Journal of Analytical Atomic Spectrometry,
Volume 7,
Issue 2,
1992,
Page 89-98
Nicolò Omenetto,
Preview
|
PDF (2179KB)
|
|
摘要:
JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 89 Photon Detection Based on Pulsed Laser-enhanced Ionization and Photoionization of Magnesium Vapour Quantum Efficiency Versus Ion Yield* Invited Lecture Nicolb Omenetto Benjamin W. Smith,? Paul B. Farnswortht and James D. Winefordnert Commission of the European Communities Joint Research Centre Environment Institute 2 1020 lspra (Varese) Italy The concepts of quantum efficiency and ion yield pertinent to a photon detector consisting of an atmospheric pressure oxygen-argon-acetylene (or hydrogen) flame containing magnesium atomic vapour and also some published experimental results obtained with this detector are critically discussed. In these experiments collision-induced ionization and direct photoionization of magnesium atoms are performed.In the former instance one dye laser excites the atoms from the ground state into the 3s3p-lPl0 state and another laser in temporal and spatial coincidence with the first brings the excited atoms into the 3s5d-'D2 state from which collisional ionization occurs. In the last instance direct photoionization from the lPlo level is accomplished either with the excimer laser radiation at 308 nm or by tuning the second laser into an autoionizing resonance at 300.9 nm. It is shown that the total loss rate from the lPlo level can be calculated by time resolving the resonance fluorescence waveform with and without the presence of the second laser and that by integrating this signal over the duration of the laser pulse the fluorescence dip obtained can be related to the ion yield of the excitation-ionization scheme.It is also stressed that because of the presence of a metastable level which can act as a trap for the excited atoms owing to quenching collisions in the flame the quantum efficiency of this detector will never exceed the 'ion branching ratio' i.e. the ratio between the ionization rate and the total loss rate from the excited level. By evaluating several experimental results obtained with different laser systems it is shown that in the collision-induced ionization mode both fluorescence and ionization measurements are necessary to derive the quantum efficiency whereas in the direct photoionization mode fluorescence data suffice. Keywords Photon detector; quantum efficiency; ion yield; laser-enhanced ionization and photoionization A conventional air-acetylene flame supported on a slot or a capillary burner at atmospheric pressure has been the most widely used atom reservoir in diagnostic and analytical applications of the technique of laser-enhanced ionization spectroscopy.'12 This is understandable because of the simplicity of operation of the flame its adequate versatility as it is capable of atomizing many different elements and its analytical sensitivity which may reach the upper range of pg ml-1 of concentration in aqueous solutions. An attractive use of the flame or of another atomizer in combination with the ionization technique was proposed and evaluated several years ago3v4 and also more The basic idea stems from the fact that when atoms are excited to a certain bound state by absorption of resonance photons they can be subsequently ionized either by pumping them into another higher energy level from which efficient collisional ionization occurs or by direct photoion- ization proceeding from the excited state into the ionization continuum.The number of charges created is proportional to the number of photons absorbed in the primary excita- tion step as a consequence the atomic system and the ionization technique combined together form what can be classified as a photon detector based on ionization. The practical implementation of such a detector can be as follows a given metal atom M emits resonance photons of ~ ~~ * Presented at the XXVrColloquiurn Spectroscopicum Interna- t Permanent address Department of Chemistry University of 3 Permanent address Department of Chemistry Brigham Young tionale (CSI) Bergen Norway June 9- 14 199 l .Florida Gainesville FL 3261 1 USA. University Provo UT 84602 USA. energy hvM as a consequence of some excitation process (which can be induced by thermal electrical or radiative means). These photons are collected and transferred to the flame containing pure M vapour mainly in its ground state. The absorbed photons will then create M* excited atoms. If the flame is simultaneously illuminated with one or two lasers which ionize M* the flame will act as an ionization detector for the primary photons emitted at hv,. As mentioned before these photons can be the result of thermal excitation of the metal atom M (e.g. another flame a d.c.arc or a plasma) electrical excitation (glow discharge) or radiative excitation (fluorescence or Raman photons) . In our previous theoretical evaluation of this ~ o n c e p t ~ ~ ~ several 'ionization detectors' were considered together with two possible ionization schemes involving the ele- ments lithium and lead assumed to be present as atomic vapours in an atmospheric pressure flame. No experimen- tal data were reported. Recently,* we have succeeded in measuring a partial Raman spectrum of carbon tetra- chloride by scanning the ultraviolet (UV) output of a dye laser focused in a cuvette filled with pure carbon tetra- chloride and transferring the scattered photons from the cell into an air-acetylene flame containing magnesium vapour which was simultaneously illuminated with a second dye laser. In the experiment the efficiency of the detection process (about 50%) was evaluated by ratioing the number of electrons created by the number of Raman photons entering the flame.Both quantities were measured experiment ally. The primary figure of merit of a photon detector must be its quantum eflciency QE ie. the ratio between its90 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 output event and its photon input. Here this quantity is identified as the ratio between the number of charges created in the detector and the number of incident photons. Whether the photons entering the detector are all in the probed ionization volume or are all absorbed by the ground-state atoms present in the interaction region de- pends on how well the experiment is devised.The same holds for the efficiency of collection of charges created. Such experimental constraints should therefore not be considered in the definition of the quantum efficiency. On the other hand as the detection is based on the generation of charges another pertinent parameter is the ‘ion yield’ i.e. the ratio between the number of charges created during the interaction and the number of atoms present in the probe volume. The two concepts are different as clearly pointed out by Havrilla et ~ 1 . ~ in an ionization study with continuous-wave lasers. As stated in their work the near- unity ionization yield attained for sodium should not be confused with the quantum efficiency or ion yield per photon absorbed as under their conditions a given atom could undergo about 1 x lo4 excitation and quenching collisions before a collisional ionization occurred.This explains the different basic requirements that the ionization method has to fulfil if it is used as an analytical method the ion yield should be unity irrespective of the number of absorption transitions (i.e. primary photons) necessary to form an ion pair; and on the other hand if it is used as a photon detector its quantum eficiency should be unity so that one ion pair is created for every primary photon absorbed. Both the ion yield and the quantum efficiency of the ionization process have been addressed experimentally and results have been reported for lithiurnlo and for lithium and sodium atoms in an air-acetylene flame.” In this paper a complete characterization of a photon detec- tor based on the ionization of magnesium atoms in an oxygen-argon-acetylene flame is described.The theory pertinent to both photoionization and collisional ioniza- tion schemes is given. From a consideration of the fluorescence and ionization data several relationships between the various parameters are derived and discussed. On the basis of this theory several experimental results obtained with different laser systems and published re- cently are then discussed. Theory General Considerations The atomic level diagram pertinent to the ionization detector used is shown in Fig. 1. In Fig. l(a) the magnesium atomic system is considered as a five-level scheme a ground state (level I) a metastable triplet of states (level m) a single excited state (level 2) a highly excited manifold of states of which one particular level (level 3) is reached by the laser and an ionization continuum (level i) or an autoionizing state (level ai) above the first ionization potential of the atom.As indicated ground-state atoms are excited into the lPlo state by absorption of photons at A12=285.213 nm. In the absence of the second laser the excited atoms may be lost by collisional ionization into the continuum (indicated by the rate coefficient k2i) by photoionization caused by absorption of a second photon of energy hvI2 and by downward collisions into the metastable triplet (indicated by the rate coefficient k2,,,). On the other hand when the second laser is also present the atoms in the lPlo level can be further excited into the ID2 state by absorption of photons at &=470.299 nm.From the ’D2 state collisional ionization (indicated by the rate of coefficient k3J or direct photoionization with another photon of energy hv23 occurs. ai b) -3-- * 3s 3P 3P 3s2 3s 3P 3P 3s2 Fig. 1 Simplified energy level scheme for the magnesium atom showing the two-step radiative excitation scheme followed by collisional ionization (a) without photoionization process; and (b) with schematized photoionization processes. Both radiative rates and upwards and downwards collisional rates are indicated by the appropriate coefficients (see text for discussion). The dashed arrows at 285.2 13 and 470.299 nm indicate that photoionization is neglected. Energy levels are numbered according to the increase in energy m and i represent the metastable level and the ionization continuum respectively and ai represents autoionizing level As repeatedly pointed out in the literature,11-14 it is difficult to define a state-specific collisional rate coefficient as in the flame atoms will be transferred to nearby levels that also efficiently ionize by collisions. Therefore kj1 has to be regarded as an ‘effective’ rate coefficient involving aJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL.7 91 manifold of levels (e.g. the np and nf levels) close to the 5d-ID2 level directly pumped by the laser. Downward collisional and radiative losses to level m are considered as indicated by the rate coefficients k3 and A3,. A second laser can of course directly photoionize the atoms from the *PI0 state if the photon energy provided is greater than the difference between the ionization contin- uum i and the excited state 2 or if a suitable autoionizing level [ai in Fig. l(b)] is available and can be reached by the laser.In this event a four-level scheme is considered. When the laser is tuned to a transition between bound states (i.e. 1-2 or 2+3) the transitions can be optically saturated so that the stimulated absorption and emission rates dominate all other rates between these levels. Reso- nance fluorescence (indicated by the coefficient of sponta- neous emission A2J can be monitored both in the absence and presence of the second excitation-ionization step. The temporal shape of the laser pulses is considered to be rectangular and spatial homogeneity of the beam in the interaction volume is also assumed.The duration of the laser pulses is of the order of 10-1 5 ns. As a consequence during this time scale neither recycling between the atoms trapped in the metastable manifold m back to ground level 1 is considered nor is recombination of the ions formed during the interaction assumed to play a role in the kinetic modelling of the temporal evolution of the system. The laser spectral bandwidth was different for the two excitation steps. In each instance because of the fast collisional relaxation occurring in the atmospheric pressure flame both radiative interactions address the entire atomic population irrespective of their velocity distribution with regard to the laser beams.Finally coherence effects which would require a density matrix rather than a rate equation approach,I5 are neglected in view of the relatively modest laser powers involved and of the high collisional rates characteristic of our flame system. Temporal Evolution of the Populations for the Case of Collisional Ionization from a Highly Excited State [Fig. l(a)] When the collisional ionization scheme indicated in Fig. l(a) is adopted for photon detection the 3p+d transition needs to be optically saturated. This is equivalent to saying that during the entire duration of the laser pulse the population densities ( ~ m - ~ ) of levels 2 and 3 are locked together and undergo the same temporal changes with proper account of the statistical weights of the levels.The following relationships then hold n2(t)+n3(l)=n23(t) (la) The temporal evolution of the ion density and of the metastable density are dn,ldt = Rln23 (2) dn,ldt= R,n23 (3) where R and R (s-I) are the total loss rate coefficients into the ionization continuum and the metastable state respec- tively (see below). The total atomic population nT is then given by nT= n + n2 + n + ni (4) It is easy to reduce the number of differential equations needed by realizing that n,lni = R,/Ri ( 5 3 ) and redefining a new variable n as follows n,+ni=ni 1+- =n ( 3 By substituted in eqn. (4) we have nT = n + n23 + n,' ( 6 ) The two differential equations needed are therefore [see eqns. (1) and (5b)J dn231dt=R,,n,-n,3(R~ +R,+R,) (7) (8) The various rate coefficients appearing in the above equations can be written explicitly as follows (9a) R k2 + k3 + A (9b) R12 = ~ 1 2 ~ 1 2 ( ~ V e l T ~ ~ ~ I 2 ) (9c) dn,'ldt = n23 (R + R,) R = R21 + RJi = k2 + o2,II2+ k31 + 63,123 (9d) g1 R i 1 3 ( g2 - + g3) R 2 +A2 I + k2 1 R21-421 +k2l Note that the collisional ionization from the ground level and also collisional population of the excited levels are neglected.In eqns (9a-e) the various values (cm2) represent the peak cross-section of the processes evaluated at the central frequency of the transition and and I 2 3 indicate the laser photon irradiances (s- I cm-2) integrated over the emission spectral bandwidth. The ratio in eqn. (9c) takes into account the spectral efficiency of the primary absorption process as 6ve is the effective absorption width of the magnesium atoms in the flame and 6vI2 is the laser bandwidth.The temporal behaviour of the ion density and the excited atom density can finally be obtained by solving the above system with standard mathematical procedures.16J7 The result is n T nl(t)= - ni(t)= - Ri Ri Ri+Rm R + R 1 (1 1 ) a2a3 n23( t) = nT- x -(e-*,'- e-a2t) i ~ 2 - ( ~ 3 R,+R where a2 and a3 are given by V It is now possible to express our definition of the quantum efficiency of the ionization detector in terms of the above rate coefficients. In the limiting case in which all incident photons are absorbed (see also ref. 1 1 )92 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 Note that in eqn. (1 3) the coefficient Rzl (rather than Ril) has been used because the photon detector must obviously be linear. As a result the photoionization term o ~ ~ I in eqn.(9a) has also been omitted. The quantity Ri ti* I- Ri + R can be defined as the 'ion branching ratio' i.e. the relative proportion of atoms which are lost by ionization with respect to the total number of atoms lost in the two traps (ionization + metastable). It is instructive to combine eqns. (1 3) and (1 4) to show that Clearly qi* cannot exceed ti*. In the absence of a metastable trap ti* is unity (Rm=O) and qi* correctly represents the fractional probability of ionization in complete analogy with the quantum efficiency of resonance fluorescence. It is also worth noting that this definition of qi* given by eqn. (1 3) as the ratio between the number of charges produced [ =(nlR12)] follows directly from eqn.(7) in the steady-state limit i.e. when dn2,1dt=0. However strictly eqn. (1 3) does not conform to our earlier general definition of quantum efficiency as the ratio between the charges produced and the photons incident on the detector. The two definitions are related by the following expression QE=tli* [1 -ex~(-a,2nlladI ( 1 5a) Here the quantity within brackets represents the fraction of photons absorbed by n ground-state atom density distri- buted over the absorption length labs. Clearly this fraction should be unity. Finally when the second laser (tuned at transition 2-3) is absent one simply needs to neglect level 3 (and the associated coefficients) in the above-derived equations so that Ri will be given only by k2 + o2,II2 R by k2m and so on.Ion yield The instantaneous yield of charges can easily be obtained from eqn (10). The coefficients a2 and a3 [eqn. (12)] can be simplified in the two cases of linear interaction between the atoms and the photons hvI2 and of optical saturation. In the former instance R is much smaller than the other coefficients whereas the opposite is true in saturation. For the linear case and using the definitions of eqns. (1 3) and (1 4) the instantaneous ion yield can be derived and is given by The integrated ion yield which is obtained by integrating over the probing time Tp [eqn. (1 l)] is given by the same expression [eqn. ( 16b)] after replacement of the variable t with Tp ie. In both equations the subscript 'lin' indicates linear interactions.Eqn. ( 16) emphasizes the difference between the parameters quantum efficiency and ion yield; indeed for ti* = 1 the ion yield can still approach unity even if the quantum efficiency qi* is much less than unity because the product R12qi*Tp can still be greater than unity. Of course the ion yield with a metastable trap present will never exceed the ion branching ratio. In the absence of the second laser the integrated ion yield is given by ,where the superscript 'off' indicates the absence of the second excitation step. In eqn. (17) ti and qi now without the asterisk as superscript are still given by eqns. (1 3) and 1:14) in which all the coefficients pertinent to level 3 have been set equal to zero. In a similar way one can derive the ion yield for the case where the 1-2 transition is optically saturated.With and without the second laser present one obtains and Again the subscript 'sat' represents optical saturation of the 11-2 transition and the superscript 'off indicates the absence of the second laser. The measured ionization signal is directly proportional to the ion yield through a calibration constant which takes into account the collection efficiency the gain of the amplifier and the shape of the current pulse. The ratio between eqns. (1 8a) and ( I 8b) can be defined as the enhancement E in the ionization signals measured with and without the second laser. Explicitly this ratio is given by The derivation of the linear counterpart of eqn. (19) is straight forward. Resonance fluorescence and fluorescence dip The spectrally integrated resonance fluorescence signal measured in a conventional way at 90" with respect to the excitation beam with a photomultiplier is always directly proportional to the number density of excited atoms in level 2.The fluorescence power (J s-l) IF(f) or energy (J) i.e. integrated over the probing time QF can easily be derived from eqns. ( I I ) and (lc). By proceeding as in the previous section and using C (J s- cm3) as a proportional- ity constant we can write the following eight equations:JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY. MARCH 1992 VOL. 7 93 [IF(f)]k7= Cn,( ") exp [ - (") (R2i + kzm) I ] g,+g2 g1 +g2 (23b) Eqns. (2 1 a) and (2 1 b) describe the time-resolved fluores- cence waveform in conditions of optical saturation of the 1-2 transition.It is worth noting that for t=O IF is not zero but rather reaches its maximum value. The outcome is the result of simplifying eqn. (1 1) because of the assump- tion of saturation. As eqn. (1 1) correctly shows at t=O n23 (and therefore n2) is zero. However early in the interaction the first exponential term in parentheses is negligible in comparison with the second so that the losses in the fluorescence photons occur only after a very short 'pumping time' which is given approximately by the reciprocal of RI2.l8 Therefore by time resolving the saturated fluores- cence waveform within the duration of the laser pulse the total loss rate from level 2 can be directly observed. This outcome is significantly different from that obtained by monitoring the time-resolved fluorescence waveform ex- cited by a very short laser excitation pulse after the pulse has ceased.In this instance in fact irrespective of whether the transition is saturated or not one obtains the effective lifetime of level 2 which includes the radiative and quenching de-excitations towards the ground state in addition to the losses to the ionization continuum and to the metastable level. When the laser photon irradiance is such that the transition is just saturated but the photoioni- zation rate ( 0 ~ ~ 1 ~ ) is still negligible then Rzi in eqn. (2 1 b) is given only by kZi. As the 3p level reached with this transition lies 3.3 eV below the ionization potential kZi is not expected to be high" and is certainly much less than k2 in the acetylene flame.The only term governing the loss rate from level 2 would then be k2". The usefulness of eqns. (20)-(23) can be appreciated when one monitors the decrease in the resonance fluores- cence signal caused by the depletion of n2 when the second laser is present i.e. the 'fluorescence dip'. As shown by Omenetto et af.,19 fluorescence dip spectroscopy can be an important tool in flame and plasma diagnostics and several papers have appeared confirming the above predic- From the above equations one can derive the time- resolved and the time-integrated fluorescence dips for both linear and saturated conditions from the following defini- tions tions. 13.2021 A = - QF (25a) A' ZE A/[ &Ioff (25b) The absolute fluorescence dips are defined by eqns. (24a) and (25a) whereas relative dips are defined by eqns.(24b) and (25b). The analytical expressions for the various fluorescence dips can be derived by the appropriate substitution of eqns. (21)-(23) into eqns. (24) and (25). When the time-resolved fluorescence waveforms are available the time-dependent absolute dip [eqn. (24a)] at t=O is given by Cn (26) 828 [A(O)Isat = (gl +gZ)kl +gZ+g3) The dip is then seen to increase with time and to reach a maximum at a time t,, given by 1 Eqns. (21a) and (21b) and eqn. (24a) are plotted in Fig. 2 where it is assumed that the sum R,+R exceeds R2,+kz by one order of magnitude. It is clear that the maximum in the temporal behaviour of A(t) occurs earlier for increasing values of R,+R,. On the other hand the relative time- resolved dip not shown increases monotonically towards unity.The attraction of monitoring the fluorescence dip is that it can be directly related to the ion yield (see under Ion yield). Within the limits of validity of our kinetic model it is clear that the rate of increase of the instantaneous ion yield is identical with the loss of the fluorescence signal [compare eqns. (16a) and (20a) respectively]. It is then interesting to relate the integrated fluorescence dip with the integrated ion yield with the purpose of infemng the last parameter from a single independent measurement of the former. This can be carried out by using eqns. (17) (18) (22) (23) and (25). Under saturated conditions we obtain the following relationship 0.80 r I I -. - \ 1 1 0 5 10 15 Timehs Fig. 2 Calculated time behaviour of the fluorescence dip at 285.213 nm occumng when the second laser beam depletes the population of the 3p level A [I&)]# (R,i+R2,=6x lo7 s-l); B [Id?)]% (Ri+R,=6x lo8 s-l); and C A(?).See text for the definitions of the symbols and for discussion. Ordinate values are normalized to nT94 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. (28) As eqn. (28) indicates however tempting and intuitively feasible one cannot directly evaluate the ion yield from an integrated fluorescence measurement. It is possible how- ever to simplify eqn. (28) as follows. For probing times Tp of the order of several nanoseconds it is likely that in the oxygen-argon-acetylene flame and for the 3p level of magnesium the following inequality holds In the absence of quenching collisions into the metastable level eqn.(29) is certainly valid. By using eqn. (29) eqn. (28) becomes or after using eqn. (1 8a) and some simple algebra From an inspection of eqn. (30b) several considerations (0 The highest value of Yi cannot exceed ti* i.e. ( Yi/Ci*)< 1. For Yi = ti* A'= 1. (i0 For Y,-0 A'=[g3/(gI+g2+g3)]. This outcome had to be expected because even in the absence of ionization the fluorescence signal decreased because of the redis- tribution of the population in the three levels. (iii) In the absence of a metastable level or in an atomic reservoir characterized by negligible quenching (z. e. negligible k2m) ti* = 1. An experimental measurement of A' would then give Yi directly. Eqn. (30b) is plotted in Fig.3(a) for several values of ti*. It is important to realize that owing to the poor sensitivity (slope) of the functional dependence of A versus Yi a small uncertainty in the measurement of A will result in a large uncertainty in the corresponding evaluation of Yi. are worth stressing Temporal Evolution of the Populations for the Case of Direct Photoionization from the 3p Level [Fig. l(6)) Here direct photoionization from the 3p level is consi- dered. This can be accomplished with the excimer laser (308 nm) or by tuning the frequency doubled laser to 300.9 nm which corresponds to the peak of a broad autoionizing transition of the magnesium atom.22 This excitation-ioni- zation scheme is obviously simpler than that treated in the previous section. Moreover it is the only scheme possible in collision-free atom reservoirs e.g.atomic beams in vacuum. In an atmospheric pressure flame and in an argon plasma direct photoionization of magnesium from the 3p level was successfully accomplished with the excimer laser at 308 nm.I3 The analytical expressions describing the ion yield and the fluorescence signals for the excitation-ionization scheme of Fig. l(b) can be derived in a straightforward manner from the expressions already given in the previous section. Indeed the modifications required are trivial because level 3 is now absent ( i e . all terms having 3 as 1 .oo (a) A B 0.40 1 7 0 0.20 0.40 0.60 0.80 1 .oo ( y)o:t 0.80 0.60 . .- .9 0.40 0.20 0 0.20 0.40 0.60 0.80 1 .oo ( YJ0:' Fig. 3 Theoretical dependence of the time integrated fluorescence dip obtained under saturation and the integrated ion yield for A (*=0.5 and B ?= 1 .(a) Collision-induced ionization case and (b) photoionization case. Adapted from ref. 23 subscript are zero) and eqns. (9a) and (9b) are modified as fdl0WS I n eqn. (31a) the last term on the right-hand side represents the photoionization rate from level 2 due to the excimer laser and the corresponding term in eqn. (3 1 b) indicates the photoionization rate through the autoionizing level. The terms IZi and 12(ai) are the photon irradiances (s-I cm-2) of the excimer and the dye laser respectively. With excimer photoionization the quantum efficiency [see eqn. (1 3)] now becomes where the superscript 'pi' indicates photoionization. The difference between this photoionization scheme and that treated in the previous section is evident if one considers the expressions for the absolute and relative fluorescence dips.By making the appropriate substitutions in eqns. (20) (21) and (27a) we obtain (-) 1 [A(f)]gk= R"(qi)Piexp [ -R12( 2) Pi I] 'Cn R2iJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 95 Rl’ in the above equations is given by eqn. (3 1 a) or (3 1 b) whichever applies. Unlike the case described by eqn. (26) here the dip at time zero is indeed zero as the atomic population is shared between levels 1 and 2 and no losses to the metastable or to the ionization continuum have yet occurred. In a similar way eqns. (22) (23) (28) and (30) are modified to obtain the relationship between the integrated fluorescence dip and the integrated ion yield as follows or after simplification [see eqn.(29)] The same considerations given in the discussion of eqn. (30b) hold here with the difference that for Yi=O A = O and therefore the relative integrated dip is more sensitive to variations of Y than in the corresponding collisional ionization case. This can be seen in Fig. 3(b) where eqn. (37) is plotted for several values of cl*. Again if k2 is negligible a single measurement of A’ gives Y,. It is worth stressing again that the photoionization scheme performed in a collision-free environment is indeed simpler to treat than the collisional scheme in our flame and that the experimental data obtained are more amenable to a direct interpretation. Evaluation of qi* From the definition of qi* the most direct approach to its experimental evaluation would obviously consist in mea- suring the number of photons absorbed and the total number of charges created in the (common) atomizer volume.This entails a simultaneous absorption and ioniza- tion measurement which can easily be arranged experimen- tally. If one performs absorption fluorescence and ioniza- tion measurements simultaneously then several combina- tions of experiments exist as illustrated below. Evaluation of ft,* for collisional ionization from a highly excited level The evaluation of quantum efficiency in this instance can be derived from the theoretical treatment given under Temporal Evolution of the Populations for the Case of Collisional Ionization from a Highly Excited State. For the sake of illustration and simplicity the necessary measure- ments are reported below in an itemized sequence (all measurements refer to the flame) A.B. C. D. E. Saturated time-resolved fluorescence waveform in ab- sence of the second laser i.e. eqn. (21b). Integrated ion yield in the linear interaction case i.e. eqn. (16b). Integrated ion yield in saturation conditions i.e. eqn. ( I 8a). Relative integrated fluorescence dip i.e. eqn. (30b). Saturated time-resolved fluorescence waveform with both laser beams i.e. eqn. (21a). The interplay and combination of the above measure- ments in order to derive qi* is now easy to find. For example from E one can obtain the ratio (Ri/ti*); subse- quently from C one derives ti* and finally from B qi* can be calculated if the laser parameters (i.e R,2 and At,) are known.Another similar sequence would be to start with C and D to obtain ti* which again inserted in B allows evaluation of qi*. As one can see in this instance the quantum efficiency of the ionization detector cannot be evaluated by using only fluorescence measurements because there appears to be no easy direct way of addressing experimentally k3i in the flame without resorting to ionization measurements. Evaluation of )I,* for direct photoionization from the 3p level By proceeding as in the previous section and refemng to the theory outlined under Temporal Evolution of the Populations for the Case of Direct Photoionization from the 3p Level we can define the following sequence of measurements A. B. C. D. E. Saturated time-resolved fluorescence waveform in the flame in the absence of the second laser i.e.eqn. (2 1 b) (modified as described under Temporal Evolution of the Populations for the Case of Direct Photoionization from the 3p Level). Integrated ion yield in the flame under linear condi- tions i.e. eqn. ( I 6b) (modified as above). Integrated fluorescence dip in an argon atmosphere (this could be for example an inductively coupled argon plasma). In this instance quenching from the 3p level into the metastable state should be negligible and therefore it is possible to derive the photoionization cross-section of the 3p level,13 i.e. aZi. Lifetime of the 3p level in the flame i.e. a direct measurement of the fluorescence decay after the laser excitation pulse. Measurement of the fluorescence quantum efficiency in the flame by a combination of absorption and fluores- cence measurements obtained under strictly linear interaction conditions or from the ‘saturation para- meter’ pertinent to the 285.21 3 nm transition which is obtainable from a saturation curve.12J6 By referring to the above sequence one would then obtain from A the collisional losses into the metastable (i.e.k2,) from C the photoionization cross-section and there- fore from both A and C <,*. The quantum efficiency qi* can then be derived from B D or E. It is worth pointing out that contrary to the collisional ionization case the ionization quantum efficiency can be obtained by resorting to fluorescence measurements alone. Discussion In the recent l i t e ~ a t u r e ~ ~ ~ ~ * ~ ~ the ionization of magnesium atoms in oxygen-acetylene and oxygen-hydrogen flames diluted with nitrogen and grgon has been reported.Collision-induced ionization,23 non-resonant photoioniza- tionI3 and photoionization via an autoionizing statez4 have been discussed although only in ref. 23 was the aim as in this work explicitly that of characterizing the system as a photon detector. These experiments can now be discussed on the basis of the theoretical considerations given in the previous sections. Collision-induced Ionization In this instance oxygen-argon-acetylene and oxygen-ni- trogen-acetylene flames were used and more experimental results in addition to those already published,z3 are96 JOURNAL OF ANA.LYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL.7 reported here. A fluorescence experiment was also per- formed in an inductively coupled argon plasma ix. in an atom reservoir characterized by a completely different quenching behaviour. The data were obtained with two tunable dye lasers (Model FL3002E Lambda Physik Gottingen Germany and a special eight-cuvette system Jobin-Yvon Longjumeau France) pumped by an excimer laser (Model LPX-200 Lambda Physik) operated with XeC1. The laser beams could be directed by folding prisms and mirrors into the flame supported on a laboratory made circular burner head. Beam expanders were also used whenever necessary to improve the uniformity of irradiation in the ionization volume. The fluorescence wavelength of the experiment (i.e. 285.213 and 470.299 nm) were monitored with a fast microchannel plate photomultiplier (Model R 1294 Hamamatsu TV Japan) characterized by a rise time of 283 ps displayed with a fast (500 MHz) analogue oscilloscope (Model 7904 Tektronix Beaverton OR USA) and then photographed.The fluores- cence was isolated by a small monochromator (Model H- 10 Jobin-Yvon). For the ionization measurements a water- cooled electrode negatively biased at 1500 V was im- mersed in the flame and the burner grounded through a 1 kZZ resistor. The capacitively coupled signal was fed into a fast transimpedance amplifier and the resulting voltage pulse was displayed on a digital storage oscilloscope (Model 2430 Tektronix). The absorption measurements were performed by sending the UV laser into a photodiode after attenuation and averaging the signal with and without magnesium atoms in the flame.As shown in Fig. 4 full absorption was obtained. Three typical waveforms representing the laser shape the resonance fluorescence in an argon ‘plasma and the resonance fluorescence in the acetylene flame diluted with argon respectively are shown in Fig. 5(a) (b) and (c). From an inspection of these shapes several considerations can now be made. (i) The laser pulse as scattered from a piece of diffuser (PTFE) positioned at the flame approximates fairly well a square pulse [Fig. 5(a)] as assumed in the theoretical section. However this assumption may be very misleading as it was found that the temporal profile of the laser pulse showed considerable instabil- ity even in a short time period ( ~ 3 0 min) with a _c wavefm s; -aE&*kG-&ie ~ ~ ~ ~ corresponding variation of its half-width and rise time.For a meaningful interpretation of the data it would then be necessary to monitor the laser pulse shape immediately before and after the experiment. In addition even for the pulse shown in Fig. 5(a) the rise time is too slow (see below). (ii) The shape of the fluorescence waveform shown in Fig. 5(b) clearly illustrates the optical saturation behaviour of the interaction. Indeed the temporal width of the fluorescence pulse is larger than that of the excitation pulse and there appears to be no decay of the signal during the laser pulse. The saturation spectral irradi- ance of the magnesium in the argon plasma at the erg s-l cm-2 cm-1 ( 1 erg= 1 x lo-’ joules) which is approximately ten times lower than that in the acetylene flame (see below).Although this outcome can only be considered on a qualitative basis it gives -28 5.-21 .$s. bc 2.7 {!-jl8 0 0 C 50 2 100 &scan - Fig. 4 Experimental scan of the resonance absorption signal at 285.2 13 nm showing that total absorption in the flame is achieved. The dye laser output was greatly attenuated with neutral density filters. The magnesium concentration aspirated into the flame was 20 p g ml- I . Here ‘absorbance’ represents ‘absorption’ E’ig. 5 Typical waveforms obtained with a fast microchannel plate photomultiplier (a) laser scatter; (b) resonance fluorescence of magnesium in an argon plasma; and (c) resonance fluorescence of magnesium in an oxygen-argon-acetylene flame.The time scale corresponds to that indicated on the lower right of each figure. Waveforms (a) and (b) recorded at 5 ns per division and (c) at 2 ns per divisionJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 97 convincing evidence that in an argon atmosphere quenching into the metastable state k2 can be safely neglected leading to a value for the ion branching ratio rl* close to unity. (iii) The shape of the fluorescence waveform shown in Fig. 5(c) was obtained in the argon-diluted flame with a laser pulse characterized by a rise time of 1.63 ns a width of 9.2 ns and an energy per pulse of 290 nJ. The laser had a cross-section of 7.1 x cm-2 in the flame and therefore its photon irradiance was 6 . 4 ~ lozo s-l cmW2. With a spectral bandwidth of 11 GHz (3 x 10-Io cm) similar to that of the absorption profile of magnesium atoms in the flame its spectral irradiance was 1.5 x 1019 erg s-l cm-2 cm-I corre- sponding to a radiative pumping rate R12 of 1 GHz.The corresponding pumping time," { [(gl +g2!lg2] R12/-l is therefore 0.75 ns indicating that the rise time of the laser pulse was too long for time resolving the onset of the saturation behaviour of the transition. In addition the saturation spectral irradiance for mag- nesium atoms in the oxygen-nitrogen-acetylene flame characterized by a quantum efficiency for the fluorescenceZS of 0.1 is calculated to be 1.98 x lOI9 erg s-' cm-2 cm-I which is higher than the laser spectral irradiance. For the acetylene flame diluted with argon the fluorescence quantum efficiency in- creases to 0.23 which is calculated by ratioing the time-integrated fluorescence signals obtained by re- placing nitrogen with argon and assuming equal atomization efficiencies for the two flame mixtures.In this instance the saturation spectral irradiance de- creases to 8 . 6 ~ 10'8 erg s-I cm-2 cm-I which is however only a factor of 1.7 lower than the laser spectral irradiance. In conclusion the ideal conditions assumed in the theory i.e. zero rise time and complete saturation were not met experimentally. Nevertheless the different behaviour in the flame and argon plasma can clearly be seen in the waveforms of Figs. 5(6) and (c) and an attempt to calculate the population losses of level 2 during the laser pulse can therefore be made. From the decay of the fluorescence during a time interval of about 6 ns a rate of 5.6 x lo7 s-l results.The rate [see eqn. (2lb)l contains the collisional ionization and photoionization rates out of level 2 in addition to k2,. The collisional ionization rate coefficient kzl is expected to be less than 1 x lo6 s-! for an energy defect (i.e. an energy difference between the 3p level and the ionization continuum) of 3.3 eV.lo?I1 The photoioniza- tion rate can be calculated from the value of the ionization cross-section at 285 nm which must be similar to that reportedI3 at 308 nm (2.03 x lo-" cm2) and the measured laser irradiance giving a value of 1.3 x 1 O4 s-]. It can therefore be safely concluded that the decay observed in Fig. 5(c) is entirely due to kzm. This rate coefficient together with the tabulated value of 5.3 x lo8 S-I for the radiative transition probability A2 gives the quenching coefficient to the ground state k 2 ] which is equal to 5 .8 6 ~ 108 S-I Again the conclusion is that the quenching of the fluores- cence in the oxygen-argon-acetylene flame is too severe and that a much higher laser power is needed. On the other hand for the evaluation of the quantum efficiency of our detector it is the value of kjl that matters. As stated before it would be difficult to measure directly this coefficient in our two-step excitation-ionization scheme and in fact no such measurements were reported.23 The time-integrated resonance fluorescence dip [see eqn. (30b)l was found to be 0.84. However the value of c,* is not known and therefore the ion yield cannot be derived directly from Fig.3(a). Of course the theoretical specula- tions given in ref. 23 showing that this ionization scheme would have a quantum efficiency of at least 0.73 still hold. The only experimental way of measuring ql* seems therefore to be that indicated under Evaluation of ql* for collisional ionization from a highly excited level. Apart from the inadequacy of the laser characteristics another experimental difficulty which can be easily anticipated is that it would be impossible to measure the ionization signal resulting from the two-step excitation scheme at the atom density used for the time-resolved fluorescence measure- ments. In addition the evaluation of the efficiency of detection of the charges created in the flame during the laser pulse is by no means a trivial matter.' Such measurements need still to be performed.Direct Photoionization from the 3p Level This scheme which is much simpler to treat theoretically (see under Temporal Evolution of the Populations for the Case of Direct Photoionization from the 3p Level) was used in recent experiments with pulsed laser excitation. In one experiment,13 direct photoionization into the continuum was achieved with the excimer laser at 308 nm after saturation of the 285.213 nm transition with a frequency- doubled dye laser. The aim of this experiment was to measure the non-resonant photoionization cross-section of the 3p level. With an excimer photon irradiance of 1.17 x lozs s-I cm-2 and a cross-section of 2.03 x 10-I7 cm2 R; [see eqn.(31a)l is calculated to be 2 . 4 ~ lo8 s-I neglecting again collisional ionization (k2J and photoioni- zation due to the dye laser ( 0 ~ ~ 1 ~ ~ ) . As quenching into the metastable and into the ground state is negligible [see also Fig. 5(6) and the related discussion] cl* is equal to unity. From eqn. (32) (ql*)pL=O.3 1. The same experiment would have resulted in <,* and (ql*)pi equal to 0.81 and 0.17 respectively if carried out in the oxygen-argon-acetylene flame. In these experiments it is clear that by making the excimer photon irradiance ten times greater than the over- all de-excitation rate of the 3p level the ionization quantum efficiency will increase to 0.9. This would be easy to accomplish experimentally. The same reasoning holds for the photoionization experi- ment described in ref.24. Here a frequency-doubled Nd:YAG laser (532 nm) was used to pump a dual dye laser system (Models YG58 1-30 and TDLSO Quantel Interna- tional Santa Clara CA USA). Both dye laser outputs were frequency doubled to reach the 285.213 and 300.9 nm transitions. This last transition reaches an autoionizing level above the ionization continuum. In this instance the flame was an oxygen-argon-hydrogen mixture supported on a conventional nebulizer-burner assembly. The first resonance transition was optically saturated and the auto- ionizing transition was investigated by spectrally scanning the second dye laser output while monitoring simultane- ously the resonance fluorescence and ionization signals; in this way the time-integrated fluorescence dip could be measured.The photon irradiance of the ionizing laser was 4.62 x lo2' s-l cm-2 and the autoionizing cross-section was 3 x 10-'6 cm-2 giving a rate of 1 . 4 ~ lo8 s-l. As no time- resolved data were reported one can only speculate that if the same quenching coefficients as reported for the acety- lene flame are used eqns. (31b) and (32) would give (ql*)pl=O.ll and <,*=0.71. The energy per pulse of the photoionizing laser was 396 pJ in a 10 ns pulse. An increase in the energy by a factor of 10 would result in a quantum yield of 0.54. As about 4 mJ per pulse is not commonly achievable with frequency-doubled dye lasers this scheme would be unsuitable for the use of a quenching flame as a practical ionization detector.98 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL.7 Conclusions An attempt has been made to characterize theoretically and discuss experimentally a system consisting of a flame mixture containing magnesium vapour as a photon detec- tor whose principle of operation is based on collision- induced ionization and direct photoionization. The main conclusions which can be derived from this work are the following (i) For the collision ionization case there is no experi- mental evidence yet that this detector has reached unity quantum efficiency of ionization. In fact ioniza- tion data obtained under strictly linear interaction at the first resonance transition and with a calibrated detection system are still lacking. Theoretical specula- tions indicate that such a system should have a quantum efficiency of no less than about 0.7.(ii) With photoionization the quantum efficiency of the detector can be experimentally obtained by perform- ing only fluorescence measurements and can easily be made close to unity with a sufficiently high ionization rate. (iii) A conventional flame system operated at atmospheric pressure is not an ideal detector because of its significant quenching behaviour; a thermal argon reservoir for the magnesium atoms would be the preferred choice. It is finally worth stressing that this paper has addressed only one parameter (although the most significant one) of this detector i.e. the quantum efficiency of the ionization process. Other important parameters such as its time response and linearity of operation have not been discussed and are currently the object of further investigation.References 1 Travis J. C. Turk G. C. de Voe J. R. Schenck P. K. and van Dijk C. A. Prog. Anal. At. Spectrosc. 1984 7 199. 2 Axner O. and Rubinsztein-Dunlop H. Spectrochim. Acta Part B 1989 44 835. 3 Matveev 0. I. J. Appl. Spectrosc. USSR 1983 38 561. 4 Matveev 0. I. Zorov N. B. and Kuzyakov Y. Y. J. Anal. Chem. USSR 1979 34 654. 5 Omenetto N. Smith B. W. and Winefordner J. D. Spectro- chim. Acta ‘Future Trends in Spectroscopy’ Special Suppl. 1989 91. 6 Smith B. W. Omenetto N. and Winefordner J. D. Spectro- chim. Acta ‘Future Trends in Spectroscopy’ Special Suppl. 1989 101. 7 Okada T. Andou H. Moriyama Y. and Maeda M. Opt. Lett. 1989 14 987. 8 Smith B. W. Farnsworth P. B. Winefordner J.D. and Omenetto N. Opt. Lett. 1990 15 823. 9 Havrilla G. J. Weeks S. J. and Travis J. C. Anal. Chem. 1982,54 2566. 10 Smith B. W. Hart L. P. and Omenetto N. Anal. Chem. 1986,58 2 147. 1 1 Axner O. and Berglind T. Appl. Spectrosc. 1989 43 940. 12 Alkemade C. Th. J. Hollander Tj. Snelleman W. and Zeegers P. Th. J. Metal Vapours in Flames Pergamon Press Oxford 1982. 3 Omenetto N. in Proceedings of RIS-88 Institute of Physics Conference Series No. 94 eds. Lucatorto T. B. and Parks J. E. IOP Publishing Bristol 1989 Section 3 p. 141. 4 Marunkov A. G. and Chekalin N. V. Opt. Spectrosk. 1986 61 461. 5 Pantell R. H. and Putoff H. E. Fundamentals of Quantum Electronics Wiley New York 1969. 16 Alkemade C. Th. J. Spectrochim. Acta Part B 1985,40 1331. I7 Omenetto N. Smith B. W. and Hart L. P. Fresenius’ Z. Anal. Chem. 1986 324 683. 18 Omenetto N. Spectrochim. Acta Part B 1982 37 1009. I9 Omenetto N. Turk G. C. Rutledge M. and Winefordner J. D. Spectrochim. Acta Part B 1987 42 807. 20 Axner O. Norberg M. and Rubinsztein-Dunlop H. Spectro- chim. Acta Part B 1989 44 693. ;!I Bonin K. D. Gatzke M. Collins C. L. and Kader-Kallen M. A. Phys. Rev. A 1989 39 5624. 2’2 Bradley D. J. Dugan C. H. Ewart P. and Purdie A. F. Phys. Rev. A 1976 13 1416. ;!3 Omenetto N. Smith B. W. and Farnsworth P. B. in Proceedings of RIS- 90 Institute of Physics Conferences Series No. 114 eds. Parks J. E. and Omenetto N. IOP Publishing Bristol 199 1 Section 9 p. 369. 24 Petrucci G. A. Stevenson C. L. Smith B. W. Winefordner J. D. and Omenetto N. Spectrochim. Acta Part B 199!,46 975. ;!5 Zeegers P. J. Th. and Winefordner J. D. Spectrochim. Acta Part B 1971 26 161. Paper I /04018I Received August I I991 Accepted August 26 1991
ISSN:0267-9477
DOI:10.1039/JA9920700089
出版商:RSC
年代:1992
数据来源: RSC
|
13. |
Detection of trace amounts of toxic metals in environmental samples by laser-excited atomic fluorescence spectrometry. Invited lecture |
|
Journal of Analytical Atomic Spectrometry,
Volume 7,
Issue 2,
1992,
Page 99-104
Mikhail A. Bolshov,
Preview
|
PDF (727KB)
|
|
摘要:
JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 99 Detection of Trace Amounts of Toxic Metals in Environmental Samples by Laser-excited Atomic Fluorescence Spectrometry* Invited Lecture Mikhail A. Bolshov,t Vsevolod G. Koloshnikov and Sergei N. Rudnev Institute of Spectroscopy USSR Academy of Sciences 142092 Troitzk Moscow Region Russia Claude F. Boutron and Ursula Gorlach Laboratoire de Glaciologie et Geophysique de I'Environndment du CNRS Domaine Universitaire 2 rue Moliere B.P. 96 38402 St. Martin d'Heres Cedex France Clair C. Patterson Division of Geological and Planetary Sciences 1 70-25 California Institute of Technology Pasadena CA 91125 USA Results for the direct determination of trace amounts of Pb and Cd in Antarctic and Greenland ancient ice and recent snow by laser-excited atomic fluorescence spectrometry (LEAFS) are presented. The whole procedure starting from field sampling mechanical decontamination of the samples in an ultra-clean laboratory and final analysis of the decontaminated samples is described.The measured concentrations varied in the ranges 0.1 -3 pg ml-l for Cd and 0.3-30 pg ml-l for Pb. The results for direct analysis by LEAFS agree favourably with those obtained by isotope dilution mass spectrometry and electrothermal atomic absorption spectrometry which require time-consuming pre-treatment and pre-concentration stages. Keywords Trace amounts of toxic metals; laser-excited atomic fluorescence spectrometry; Antarctic and Greenland ice and snow; direct analysis Over the last two decades there has been a continuing increasing interest in the investigation of the concentrations of toxic heavy metals (such as Pb Cd Hg and Bi) in the environment.It is well known that human industrial activities have strongly modified the natural occurrence of these metals. Correct estimation of the difference between the natural (pre-industrial) and modem heavy metal con- tent of the environment is of great scientific interest. The best approach to the problem is the investigation of toxic metal occurrences in the well preserved dated snow and ice layers deposited in the remote Antarctic and Greenland ice This is indeed a unique way to reconstruct the past natural tropospheric cycles of these metals and to assess their recent alteration by man in both hemisphere^.^^^ Despite the growing interest there are at present few reliable data for the Pb Cd and Zn contents in Antarctic and Greenland ice and snows because the minimum concentrations of these metals to be measured are ex- tremely low (down to 0.3 pg ml-' of Pb and less than 0.1 pg ml-I of Cd).Such low concentrations manifest a serious challenge for a complete analytical project starting from field sampling then transportation sample pre-treatment and the final analysis. The great difficulties of maintaining the highest levels of clean and uncontaminated conditions at every stage of the project were the reason for a large number of erroneous results being published at the very beginning of the exercise. Even with the cleanest field sampling procedures most samples especially deep ice cores are contaminated on the outside.Reliable and accurate results could only be ob- tained after developing efficient ultra-clean procedures to decontaminate the sample^.^*^*^ The decontaminated * Presented at the XXVII Colloquium Spectroscopicum Interna- 7 To whom correspondence should be addressed. tionale (CSI) Bergen Norway June 9- 14 199 l . samples must then be analysed using ultra-clean and ultra- sensitive techniques. Up until the last 2-3 years the analytical techniques which had been used for the final analysis were either isotope dilution mass spectrometry (IDMS)1-2*6 or electro- thermal atomic absorption spectrometry (ETAAS).3v7v8 Un- fortunately both techniques are not sufficiently sensitive for direct measurements of the decontaminated samples.For IDMS a complicated chemical pre-treatment and large sample volumes (30-500 ml) are required to extract a sufficient amount of analyte for the final analysis. For ETAAS a pre-concentration step and large sample volumes (30-100 ml) are also necessary. The ultra-sensitive technique of laser-excited atomic fluorescence spectrometry (LEAFS) with electrothermal atomization offered a very promising alternative for the direct determination of toxic metals at and below the pg ml" level. The first measurements of Pb9 and Cdlo in Antarctic and Greenland ice and snows demonstrated the great potential of LEAFS as the analytical technique eventually used in the programme for monitoring toxic metals in ancient ice and snows. Results of an international collaboration between scien- tists from France the USA and USSR in the direct determination of Pb and Cd in Antarctic and Greenland ancient ice and recent snows are presented.Experimental Samples and Reagents The samples analysed for the Pb and Cd contents were 14 sections of the 905 m Dome C deep Antarctic ice core 6 sections of the 2083 m Vostok deep ice core and 18 samples of fresh snow collected on a precipitation event basis at Dye 3 station South Greenland. The Dome C core was collected by thermal drilling at the French station Dome C in a dry100 JOURNAL OF ANAL.YTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 hole without a wall retaining liquid. It covers the past 40 000 years BP. The core from the Soviet Antarctic station at Vostok was thermally drilled in a hole filled with a wall retaining liquid (kerosene).It covers the past 155000 years BP. All these core sections were transported frozen in double sealed polyethylene bags from the Antarctic to Patterson's clean laboratory at the California Institute of Technology (Caltech Division of Geological and Planetaw Sciences Pasadena CA USA). The sections were about 10 cm in diameter and 15-30 cm in length. They were decontami- nated inside the Caltech clean laboratory by mechanically chiselling successive veneers of ice (each 6- 10 mm thick) in progression from the outside to the interior of the core using ultra-clean stainless-steel chisels. The whole proce- dure was performed on a cooled double-walled nitrogen- flushed ultra-clean polyethylene tray. The procedure has been described in detail elsewhere.6 The remaining inner cores were 2-4 cm in diameter.The chiselled ice from a particular veneer (or inner core) was allowed to melt overnight at room temperature in an ultra-clean conventional polyethylene beaker. The melted samples were then acidified to be 0.1 % in HN03 using ultra- pure HN03 prepared at the National Institute of Standards and Technology (NIST) Gaithersburg MD USA. The acidified solutions were allowed to stand for 2 h then 50- 100 ml aliquots were transferred into 250 ml precondi- tioned polyethylene bottles packed inside acid cleaned polyethylene bags and immediately frozen. Specific sub- samples of these prepared samples were then analysed at Caltech by IDMS. The residual samples were then transpor- ted frozen to the Laboratoire de Glaciologie et Geophysi- que de 1'Environnement (LGGE) France.Inside the LGGE clean laboratory the residual samples were allowed to melt at room temperature and 5-10 ml sub- aliquots were transferred into 30 ml preconditioned poly- ethylene bottles and packed inside acid cleaned polyethyl- ene bags. These sub-aliquots were then transported frozen to the Institute of Spectroscopy (ISAN) USSR. All of the polyethylene beakers and bottles used for pre- treatment and storage of the ice samples and for the synthetic standards had been previously cleaned both at Caltech and at the LGGE clean laboratories. The cleaning procedures were similar to those reported in refs. 2 9 and 11 and described in detail elsewhere.12 Aqueous standard solutions used in the experiments were synthetic acidified (0.1 Oh NIST HN03) multi-elemental standards.These standards contained 1 9 elements accord- ing to typical elemental concentration ratios expected in the Antarctic ice. They were prepared in the LGGE clean laboratory from Baker or Fisher certified atomic absorption standards (1 mg ml-' solutions). The procedure for the preparation of the standards has been described elsewhere9 and differed in some details from the typical procedure of successive dilutions. The concentrations of the standards used in the LEAFS measurements varied in the 1-250 pg ml-I range for Pb and 0.1-25 pg ml-I range for Cd. Apparatus Ancient Antarctic ice and recent snows were analysed using the laser atomic fluorescence analytical spectrometer (LA- FAS-1) instrument designed and constructed at ISAN.A detailed description of the LAFAS-1 has been given elsewhere,13 hence only a brief description of the main units of the spectrometer will be given here. Radiation source Tunable dye lasers (DL) and the associated optical har- monics have been the most widely used sources for LEAFS so far. In the LAFAS-1 instrument the DL was pumped by Table 1 Typical parameters of the DL and SHG pulses (for a pulse energy of the pumping XeCl laser of 10 mJ) Parameter Pb Cd Dye laser Rhodamine N Coumarine 47 EnergyImJ 0.7 0.8 Durationhs 10 10 Power/kW 70 80 Second harmonic KDP KB5 Energyljd 25 2 Durationhs 8 8 PowerIkW 3 0.25 a XeCl excimer laser with a moderate pulse energy of 10 rnJ with a pulse duration of 10 ns and repetition frequency of up to 25 Hz.The oscillator-amplifier scheme of the transverse- pumped DL was used. The wavelength of the DL was tuned and stabilized by the computer-controlled original grating assembly and control unit. The time required for setting a wavelength was 40 ms. The range of angle rotation of the grating controlled by the electrodynamic drive was 2". This range is sufficient to cover the entire tuning range of a particular dye. Tunable ultraviolet radiation below 300 nm was obtained by second harmonic generation (SHG) of the DL radiation in non-linear crystals. Two types of crystals were used KDP (40 mm in length 15x30 mm2 cross-section) for the determination of Pb and KB5 ( 1 7 mm in length 5 x 8 mm2 cross-section) for the determination of Cd. The parameters of the DL and SHG radiation for the two analytes are listed in Table 1.Lead atoms were excited at the 283.3 nm line and direct- line fluorescence was recorded at 405.8 nm. As the structure of the electronic levels of the Cd atoms does not allow fluorescence to be recorded at any shifted line the reso- nance scheme of Cd excitation and recording at the strongest line at 228.8 nm was used. Atomizer In the LAFAS- 1 instrument electrothermal atomization of the sample was used. The samples could be atomized both in a controlled inert atmosphere or under vacuum. Up to 50 pl of the sample can be introduced into a graphite cup pressed between the graphite electrodes. In the experiments with Pb laboratory-made and pyrolytic graphite coated graphite cups were used whereas in the experiments with Cd Ringsdorff pyrolytic graphite coated graphite cups were used.The latter proved to be cleaner and produced more stable signals with lifetimes approximately twice as long as the former. The temporal profile of the analytical signals was more reproducible with the Ringsdorff cups. The maximum temperature of the cup was 2700 "C and the temperature was controlled and stabilized by monitoring the radiation of the cup. The optimum heating programmes for Pb and Cd are listed in Table 2. The laser beam profile over the cup was shaped by a system of diaphragms and lenses in front of the input window of the chamber. The scattering of the laser radiation was a serious problem for the resonance fluores- cence of Cd. To minimize this scattering the beam formation system was improved in comparison with that described in ref.13. The major changes were made in order to decrease the extent of the optical surfaces prior to the chamber and reduce the scattering of the laser beam at large angles. The improved optical scheme is shown in Fig. 1 and described in detail in ref. 14.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 101 Table 2 Atomization programmes for Pb and Cd Temperature/ "C Step Pb Cd Dry 95 90 Ash 600 180 Atomize 1600 1000 Clean 2000 1200 Cool 20 20 Ramp rate/ "C s-' Hold time/ S Argon flow rate/ 1 min-' Pb Cd 40 30 80 80 700 500 700 500 - - Pb Cd 30 210 10 10 4 4 5 5 40 50 Pb Cd 3 6 3 3 3 3 0 0 0 0 Atomizer cell Fig. 1 Modified optical scheme of the LAFAS-1 spectrometer Ll-L3 lenses; PI and P2 prisms; D1 and D2 diaphragms; W 1 input window; and W2 output window Recording system Fluorescence photons were collected at an angle of 90" on the entrance slit of the monochromator (relative aperture 1:3.7 reciprocal linear dispersion of 6.3 nm rnm-l) by a telescopic system.Fluorescence was recorded by a photo- multiplier tube FEU- 100 and a charge sensitive analogue- to-digital converter (ADC). The ADC output numerical code was fed to the computer for further processing. Neutral light filters were used to attenuate the intensity of the fluorescence for high concentrations of analyte. Analytical methodology In order to reduce contamination of the standards and samples the LAFAS-1 instrument was located inside a specially designed room supplied with filtered air. In addition the atomizer recording system associated elec- tronics and a special table covered with polyethylene film for the standards and samples were placed inside a clean chamber.The ventilation system of the chamber provided high efficiency filtering of the air and a laminar vertical flow of clean air inside the chamber. Both the samples and standards were introduced into the cup manually with a 20 pl Eppendorf micropipette. Each new polypropylene tip of the micropipette was cleaned before the first use by dipping for 2-3 min into concen- trated laboratory-reagent grade HN03 followed by several consecutive rinsings with 1% HN03 (NIST doubly dis- tilled diluted in LGGE ultra-pure water). In the course of the experiments the tip was cleaned in three consecutive bottles of 1% HN03 before every sample insertion.Typi- cally one tip was used for one working day for all the standards and samples provided there was no accidental contamination. Results Calibration of the LAFAS-1 Instrument Detailed calibration of the LAFAS-1 instrument was per- formed using the synthetic acidified standards described above. The following sets of standards were used 1 2.5 5 10 25 50 100 and 250 pg ml-I of Pb and 0.1 0.25 0.5 1.0 2.5 5.0 10.0 and 25.0 pg ml-I of Cd. Each standard was measured 2-3 times. Full sets of the analytical signals measured (numerical values) have been published else- where .9* lo The regression lines were constructed for both analytes by the least-squares method in the form Log A,=log a+b log Ci where Ai is the analytical signal (ADC code arbitrary units) Ci is the analyte concentration in the ith standard and b is the slope of the regression line.The limit of detection (LOD) was obtained for both analytes by extrapolation of the regression line to three standard deviations of the background. The parameters of the regression lines and LODs for Pb and Cd are listed in Table 3. The concentra- tion of Pb in LGGE ultra-pure water was found from the regression line to be 0.28 pg ml-I which is in excellent agreement with the value of 0.27 pg ml-l previously determined by IDMS. The concentrations of Pb in the Caltech and Cd in the LGGE and Caltech ultra-pure waters were found to be below the LODs for the LAFAS-1 instrument. Outer-Inner Profiles of the Deep Ice Cores In spite of all the precautions taken during field sampling the outermost veneers of all the core sections were contami- nated to some extent during drilling and packaging.To obtain the true values of the toxic metals content of a particular ice section all veneers of the section should be analysed. If an 'outer-inner' profile exhibits a plateau for the analyte concentration from the second or third layer to the inner core it could be stated that the surface contamina- tion did not penetrate to the central part of the section. The plateau concentration then represents the metal concentra- tion in that section of the deep ice core. However if the profile exhibits a continuous decrease in the concentration from the outside to the central part the inner core concentration must be regarded as the upper limit. Profiles for the Cd content were measured using the LAFAS-1 instrument for 10 of the 14 sections of Dome C core and for all 6 sections of the Vostok core. Two types of Cd profiles Table 3 Parameters for the regression lines and LODs for Pb and Cd (log a values for C in pg ml-I) Analyte Loga 6 LOD/pgml-I Pb 2.36 0.92 0.18 Cd 3.71 1.03 0.07102 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL.7 1 10’ r - E cr $ 1 x 10‘ 0 .- c E c 2 10 8 2 2 ’ C 0 s (a) u 0 1 2 3 4 1 x lo-‘ i L 0 1 2 3 4 Radius of corelcm Fig. 2 External-internal profiles of Cd concentrations for the two deep ice core sections (a) Dome C 172 m (3846 years BP); and (b) Vostok 499 m (26200 years BP) with and without plateaus are shown in Fig.2. The same profiles have been measured for Pb previously by IDMS6 and some of the results were checked with the LAFAS- 1. Ancient Antarctic Ice Samples Initially contamination caused by the chiselling procedure and from the ultra-pure HN03 and the water the walls of the bottles and beakers the air inside the laboratory etc. were carefully determined in numerous blank measure- ments. To do this an artificial ice core made by freezing Caltech ultra-pure water was chiselled and sub-divided using the same procedure as described above. Thus pre- pared blank samples were analysed for the Pb content by IDMS at Caltech and for the Cd content using LAFAS-1 at ISAN. The blank values were found to be about 0.11 pg ml-I for Pb and about 0.06 pg m1-I for Cd.All the determined concentrations of the analytes in ice samples were corrected for these blank values. All of the measurements made using the LAFAS-1 instrument were made with sample volumes of 20 pl for Pb and 50 pl for Cd. The full sets of the experimental data for both analytes have been published else~here.~+’~ The confidence intervals (CIS) for the measured concentrations were calculated using eqn. 8 from ref. 15 or eqn. (3) from ref. 9. For most samples the CI values (at the 95% confidence level) ranged from about 30 up to 5Ooh of the measured concentration. Variations in the Pb and Cd concentrations with depth in the Dome C and Vostok ice cores are shown in Figs. 3 and 4 respectively. It should be noted that there is very good agreement between the IDMS and LEAFS data for Pb.For most sections the results coincided to within the limits of experimental error which strongly supports the accuracy obtainable with both analytical techniques. The LAFAS-1 data for Cd are the first accurately measured values with estimated CIS for all sections of the Dome C and the Vostok ice cores. Previously published16 Cd concentrations for the Dome C provided only upper limits for many sections because even after preconcentra- tion the Cd contents of these samples were below the LOD for ETAAS. The lowest concentration of Cd (0.10 pg ml-l) was obtained for the 500.5 m Dome C section (14 000 years BP). For a 50 pl sample volume this corresponds to only 5 fg of Cd in the sample. In Fig. 3 several data points (open circles) represent the Cd values that should be carefully checked at some future date.For these sections the ‘outer-inner’ profiles are not Age of sample/l O3 years BP 0 10 20 30 40 30 - E 0 20 0 .- c I 8 8 + n P 10 0 4 3 I - E 0 P 0 \ .- 2 E El c 0) C 0 0 1 0 Sample depth/m Fig. 3 Variations of the Pb and Cd concentrations in the Dome C deep ice core as a function of the age of the ice A IDMS Pb values; B LAFAS-1 Pb values; C LAFAS-1 Cd values (sections with a concentration plateau); and D LAFAS-1 Cd values (sections without a concentration plateau) monotonic as the Cd concentrations in inner cores are higher than in some of the outer layers. This could be due to possible contamination during sample pre-treatment and/ or sub-dividing of the samples. The result for the section at 602 m is particularly dubious as the Cd concentration in the inner core is 3.1 pg ml-1 but it is 0.6 pg ml-1 in the previous (sixth) layer.It should be pointed that the Pb concentration in this same 602 m section is fairly ‘normal’. Without this dubious point for 602 m the Dome C profile for Cd qualitatively matches the Pb profile. Matching of the profiles for the two analytes is even better for the Vostok ice core. Greenland Fresh Snow A comparative determination of Cd in Greenland fresh snow was carried out using the LEAFS and ETAAS techniques. Because of significant anthropogenic input the Cd concentrations are relatively high in the range 1-10 pg m1-I in recent Greenland snow. These Cd contents can be measured by ETAAS after preconcentrati~n.~~~~~ The :samples of fresh snow were collected on a precipitation event basis from January to August 1989 at Dye 3 South Greenland as part of the ‘Dye 3 Gas and Aerosol Sampling Program’ of the US National Science Foundation.Eighteen samples of snow were analysed by LEAFS and ETAAS. The result of direct analyses using the LAFAS-1 and the analysis of preconcentrated samples using a Perkin-Elmer 2380 :spectrometer (HGA-500 graphite furnace 50 pl for 3 x 50 pl injections) are listed in Table 4. It can be seen that LAFAS-1 and ETAAS data are again in excellent agree- ment,which strongly supports the accuracy obtainable with both techniques.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 103 Age of sarnp~e/l o3 years BP 0 40 80 120 160 1 1 1 B 40 7 30 - E 0 0. 0 \ .- c h 8 20 00 c c n P 10 0 500 lo00 1500 2000 Sample depth/m 4 3 - E n 0 -r 0 .C c 4- !! 00 2 C 0 0 1 0 Fig. 4 Variations of the Pb and Cd concentrations in the Vostock deep ice core as a function of the age of the ice A IDMS Pb values; B LAFAS-1 Pb values; and C LAFAS-1 Cd values Discussion The sensitivity of the LEAFS technique is excellent as it is several orders of magnitude better than IDMS and ETAAS for the determination of Pb and Cd.For all samples measured the results for both Pb and Cd using the LAFAS-1 instrument and by IDMS or ETAAS are in excellent agreement. It must however be emphasized again that the LAFAS-1 measurements required only 20 pl of sample for the Pb and 50 pl of sample for the Cd Table 4 Comparative determination of Cd in Greenland fresh snows by LEAFSo and ETAAS6 (with preconcentration) Cd concentration ( ? standard deviation n= 3)/pg ml-I date ( 1 989) LAFAS- 1 ETAAS Sampling January 7 February 23 April 10 April 15 April 19 April 21 April 25 April 29 June 5 June 10 June 15 June 25 June 26 July 2 July 16 July 18 August 5 August 6 1.6 f 0.9 2.1 2 1.2 0.88 2 0.50 3.02 1.7 1.1 k0.6 1.72 1.0 0.4 1 k 0.24 1.3 f 0.7 2.5f 1.4 7.3 f 4.2 2.3 k 1.3 8.0 f 5.1 1.520.9 0.94 k 0.53 3.5 f 2.0 0.31 20.18 0.47 k 0.27 2.92 1.6 1.4 f 0.4 4.0 f 0.6 1 .O f 0.3 4.2 A 0.6 1.5 k0.3 2.0 ? 0.4 0.2 5 0.2 1.7f0.4 2.9 f 0.5 8.95 1.0 2.4 k 0.4 14.5k 1.4 2.4 & 0.4 0.9 k 0.4 4.2 5 0.6 0.2 5 0.2 0.6 f 0.4 3.1 k 0.5 determinations.On the other hand for the IDMS measure- ments 50-200 ml of sample and a difficult and time- consuming chemical pre-treatment stage were required for analyses at the pg ml-l level.In addition one sample analysis can take at least half a day. For the ETAAS measurements 50 ml of sample and time-consuming preconcentration were necessary. In these first applications of the LEAFS technique to the direct determination of trace amounts of Pb and Cd in ancient ice and recent snow the main goal was to compare the results with those obtained by IDMS or ETAAS for as large a concentration range as possible and for samples taken from different sections' of deep ice cores and different geographic locations. There could have been a systematic difference as the speciation of the analytes may not have been the same during the different geologic epochs Holo- cene Last Glacial Maximum (LGM) Wisconcin and Last Interglacial. During this first stage of the project it was decided to analyse the maximum amounts of samples from different sections but most samples were measured only once.Only very unexpected and dubious results for a number of samples (such as the Cd content in the 602 m Dome C section) were repeated two or three times. The small number of measurements caused the relatively wide CIS obtained (30-50%). Obviously much narrower CIS could be obtained in the future with more extensive sets of measurements. The different sources of Pb in Antarctic ancient ice were carefully determined and discussed in previous ~ o r k . ~ J ~ It was shown that the main sources of atmospheric Pb are soil dust and volcano plumes. The contribution from the oceans is insignificant throughout the period studied.The time variations of the Pb concentration faithfully tracks varia- tions in the soil dust Pb concentration calculated from aluminium concentrations in ice thus proving that most of the natural Pb in the global atmosphere originated from soil dust during both the late Wisconsin (LGM) and the next to last ice age. The contribution to the Pb content from volcanoes calculated from sulphate concentrations in the ice was negligible during the late Wisconcin and next to last ice age but accounts for about half of the measured total Pb during the Holocene age. A detailed interpretation of changes in Cd concentration in the Antarctic is given ref. 20. These data for the Pb and Cd contents in deep ice cores confirm that prior to human industrial activities there were no excesses of the two toxic metals above that contributed by soil dust and volcanoes.Now more than 99% of Pb in the troposphere of the Northern Hemisphere originates from human activities. Anthropogenic Cd emissions to the atmosphere (about 7600 tonnes per year in 1983*'~~~) are now surpassing natural Cd emissions from rocks and soils sea-salt spray volcanoes and biogenic sources (about 1300 tonnes per year).22 References 1 Murozumi M. Chow T. J. and Patterson C. C. Geochim. Cosmochim. Acta 1969 33 1247. 2 Ng A. and Patterson C. C. Geochim. Cosmochim. Acta 1981,45 2109. 3 Wolf E. W. and Peel D. A. Nature (London) 1985,313,535. 4 Peel D. A. in The Environmental Record in Glaciers and Ice Sheets ed.Oeschger H. and Langway C. C. Jr. Dahlem Konferenzen Wiley New York 1989 p. 207. 5 Boutron C. F. and Gorlach U. in Metal Speciation in [he Environment ed. Broekaert J. A. C. Gucer S. and Adams F. Springer Verlag Berlin 1990 p. 137. 6 Boutron C. F. and Patterson C. C. Nafure(London) 1986 323 222. 7 Wolf E. W. and Peel D. A. Ann. Glaciol. 1988 10 193.104 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 8 Wolf E. W. and Peel D. A. Ann. Glaciol. 1985 7 61. 9 Bolshov M. A. Boutron C. F. and Zybin A. V. Anal. Chern. 1989,61 1758. 10 Bolshov M. A. Boutron C. F. Ducroz F. M. Gorlach U. Kompanetz 0. N. and Rudnev S. N. Anal. Chim. Acta 199 1 251 169. I 1 Boutron C. F. and Patterson C. C. Geochim. Cosmochirn. Acta 1983 47 1355. 12 Boutron C. F. Fresenius J. Anal. Chem. 1990 337 482. 13 Apatin V. M. Arkhangelskii B. V. Bolshov M. A. Ermolov V. V. Koloshnikov V. G. Kompanetz 0. N. Kuznrtsov N. I. Mikhailov E. L. Shishkovskii V. S. and Boutron C. F. Spectrochim. Ada Part B 1989 44 253. 14 Bolshov M. A. Hutsch B. and Rudnev S. N. J. Anal. At. Spectrom. 1992 7 1. 15 Bolshov M. A. Dashin S. A. Zybin A. V. Koloshnikov V. G. Mayorov I. A. and Smirenkina I. I. Zh. Anal. Khim. 1986 41 1862. 16 Batifol F. Boutron C. F. and de’Agelis M. Nature (London) 1989,337 544. 17 Gorlach U. and Boutron C. F. in Heavy Metals in the Environment ed. Vemet J. P. CEP Consultants Edinburgh 1989 p. 24. 18 Ducroz F. Gorlach U. Jaffrezo J. L. and Boutron C. F. unpublished data. 19 Boutron C. F. Patterson C. C. Petrov V. N. and Barkov N. I. Atmos. Environ. 1987 21 1197. :20 Nriagu J. O. and Pacyna J. M. Nature (London) 1988,333 134. :2 1 Nriagu J. O. Nature (London) 1989 338 47. Paper 1 /03030B Received June 17 I991 Accepted July 9 1991
ISSN:0267-9477
DOI:10.1039/JA9920700099
出版商:RSC
年代:1992
数据来源: RSC
|
14. |
Determination of lead in natural and waste waters using a non-dispersive atomic fluorescence spectrometer with a tungsten spiral atomizer |
|
Journal of Analytical Atomic Spectrometry,
Volume 7,
Issue 2,
1992,
Page 105-108
Svetlana S. Grazhulene,
Preview
|
PDF (438KB)
|
|
摘要:
JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 105 Determination of Lead in Natural and Waste Waters Using a Non-dispersive Atomic Fluorescence Spectrometer With a Tungsten Spiral Atomizer* Svetlana S. Grazhulene Vladimir A. Khvostikov Nina N. Vykhristenko and Mikhail V. Sorokin Institute of Microelectronics Technology and High Purity Materials USSR Academy of Sciences 742432 Chernogolovka Moscow District Russia A laboratory-made atomic fluorescence spectrometer with a capacitively heated tungsten spiral atomizer was applied to the determination of lead in waters. The optical system for fluorescence collection and the electronics for registration have been improved which made it possible to suppress the emission of the hot spiral and to measure the fluorescence directly from the inner volume of the spiral.The detection limit of lead (3.5 x lo-' g drn-3) is reduced by a factor of 30 compared with measuring the fluorescence above the spiral. The relative standard deviation is 0.05O/0 (n=lO). An acidity of <1% HN03 a magnesium content of (0.01 g dm-3 and a calcium content of (0.02 g dm-3 do not affect the lead fluorescence signal. Keywords Non-dispersive atomic fluorescence spectrorne ter; tungsten spiral atomizer; lead determination; waters Lead is a major toxicant and its ultimate concentration must be no more than 3 x g dm-3 in drinking water in Russia. The content of lead in USSR rivers may vary from 2 x lo-' to 4.5 x g dm-3 and highly sensitive methods are required for its determination. Moreover rapid tech- niques using simple instrumentation are required for such environmental analyses.There are numerous publications on the determination of lead in waters,'-' and methods involving spectrophoto- metry or stripping voltammetryd have most commonly been employed. However even with the use of time-consuming preconcentration proceduress the spectrophotometric detection limits (DL) are only 1 x 10-5-2 x g dm'3. Stripping voltammetry6 permits the DL of lead to be decreased by an order of magnitude. However this method is also time consuming. The lowest theoretical DL ( 1 x lo-* g dm-3) has been obtained using laser methods such as laser atomic ioniza- tion spectrometry' and laser flame atomic fluorescence spectrometry (AFS).' The real DL is about 1 x lo-' g dm-3.4 The same DL has also been achieved in the non- dispersive determination of lead by AFS using a hydride generation te~hnique.~ The detection limits of commer- cially available instruments for the determination of trace amounts of lead in waters are 9 x g dm-3 [flame atomic absorption spectrometry (AAS)) 7 x g dm-3 (electro- thermal AAS) and 3 x dm-3 [inductively coupled plasma atomic emission spectrometry (ICP-AES)]. Atomic fluorescence spectrometry is often superior to other atomic spectrometric methods as far as linear dynamic ranges and DLs are concerned.The use of a nondispersive optical registration system (ND-AFS) permits the design to be simplified and the cost of the device to be reduced. The potential of AFS using a spiral tungsten atomizer has been reported This paper presents a new model of the ND-AFS based on a tungsten spiral atomizer.The potential of the spectrometer was studied on aqueous solutions of different elements attention being focused on the determination of lead in waters. Experimental Apparatus A diagram of the non-dispersive atomic fluorescence * Presented at the XXVII Colloquium Spectroscopicurn Interna- tionale (CSI) Bergen Norway June 9-14 1991. spectrometer with a tungsten spiral atomizer is shown in Fig. 1. An electrodeless discharge lamp (A) excited by an r.f. generator (B) (frequency of excitation 100 MHz) serves as a fluorescence excitation source. Its radiation is collected by a lens (C) and directed to the tungsten spiral atomizer (D). The atomizer is a tungsten spiral 1.5 mm in diameter and 2 mm long; the wire diameter is 0.1 mm and the number of wire loops is ten.The resonance fluorescence signal is observed at 90" to the incident angle. Fig. 2 shows the design of the analytical unit which includes the tungsten spiral atomizer with a gas flow- forming system a fluorescence excitation source a photo- multiplier and optics for excitation irradiation and fluor- esence collection. If necessary an interference filter with a spectral bandpass corresponding to the wavelength of the element to be determined can be placed in front of the photomultiplier to suppress the emission of light from the hot spiral. The design differs from the previous one,1° in which the fluorescence signal was measured above the spiral by measuring the fluorescence directly from the inner volume of the spiral. In this instance the concentration and I I L To computer Fig.1 Schematic diagram of the ND-AFS instrument with a tungsten spiral atomizer A EDL; B excitation r.f. generator; C lens; D tungsten spiral atomizer; E objective; F diaphragm; G photomultiplier; H optical filter; I narrow-band lock-in amplifier; J pulse generator K peak detector L gated integrator; M indicators; N laboratory-made power source; and 0 control unit106 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 Fig. 2 Design of the analytical unit of the spectrometer A electrodeless discharge lamp; B excitation r.f. generator; C lens; D tungsten spiral atomizer E objective; F diaphragm; G photomul- tiplier; H optical filter; I gas flow-forming system; J light traps; and K sample injector temperature of the atomic vapour are higher.However the emision of light from the hot spiral increases by 2-3 orders of magnitude compared with measuring the fluorescence above the spiral. In order to suppress this effect and also excitation radiation scattered on the spiral loops a special optical system for fluorescence collection with an objective (E) and diaphragm (F) in its focal plane is used. The additional suppression of the emission from the spiral is made at the expense of modulation of the exciting radiation of the electrodeless discharge lamp (modulation frequency 10 kHz depth of modulation 80%) and is followed by the separation of a valid signal from the total signal at the output of the photomultiplier using a narrow-band lock-in amplifier (I) at the modulation frequency.Further the fluorescence signal the duration of which is 20-50 ms is separated by means of the high frequency filter from the scattered signal which is also modulated but with a constant amplitude. The resulting fluorescence signal is measured by the gated peak detector (K) and an integrator (L). The information on amplitude and integrated absor- bance is output to indicators (M). Output to a computer or other external device is also provided. The tungsten spiral atomizer is heated by means of a laboratory-made power source (N) which includes the electron current stabilizer voltage stabilizer capacitor commutator and discharge key. All of this is controlled by the control unit (0) according to the heating programme.The programme includes four cycles annealing drying ashing and atomization. The first three cycles involve heating the spiral in a direct current (at temperatures of up to 1900 K and time up to 100 s) and for the fourth cycle impulsive discharge of the condenser batteries through the spiral is applied (the maximum temperature is 3300 K with ;atomization time up to 100 ms and heating rate up to 1 x lo6 K s-l). The control unit involves the on-line memory which allows eight different heating programmes to be recorded. The possibility of working with an outside memory device is provided. Moreover the control unit synchronizes all the other units of the spectrometer. Reagents and Procedures !Standard and test solutions were prepared from high-purity metal nitrates and de-mineralized water.The stock solu- tions had a metal content of 1 g dm-3 and were diluted as required. Samples with concentrations less than 1 x lo+ g dm-3 were prepared directly before measurement to prevent precipitation on the walls. The solution may be applied to the spiral either by using a micropipette or by dipping the spiral in a glass tube containing the solution to be analysed in which the liquid is held in the vessel by capillary forces. The latter improves the precision from 10% for the former to 2%. A sample volume of 2 p1 was used in both instances. Coincidentally with the fluorescence measurement the emission from the spiral was measured by a photodiode placed near the spiral. The photodiode was calibrated against temperature which allowed the dependence of the lemperature of the spiral on time during heating to be obtained and comparison of the fluorescence signal with atomization temperature conditions.Parameters of the femperature conditions during heating of the spiral (ashing drying time and temperature heating rate and maximum atomization temperature) were chosen for each element to give the best signal-to-noise ( S / N ) ratio. Argon was used as a shielding gas. Results and Discussion The typical fluorescence signal shape for lead under optimum conditions of heating the spiral (the lead content in the solution is 1 x g dm-3) is shown in Fig. 3. The atomization temperature curve under the same conditions is also given. It can be seen that the majority of the lead atoms are atomized at 1200-1 300 "C.It is likely that the further decrease in fluorescence is associated with removal of the main part of the analyte from the spiral. The spiral heating rate is initially 1 x lo5 K s-' at the atomization temperature. As can be seen the fluorescence peak widens at lower spiral heating rates and its amplitude decreases whereas increasing the heating rate above the optimum 1300 c E !i 800 E z 30C I 1 0 0.050 0.100 Tim& Fig. 3 Shape of temperature curve (A) and fluorescence signal (B) of lead in aqueous solution. Conditions of determination lead concentration 1 x g dm-'; drying temperature 360 K; drying time 60 s; temperature of atomization 1200 K; and spiral heating rite 1 x lo4 K s-IJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL.7 107 10 .- a .- c 1 v) 9) c a3 C 3 $ 1x10-' E 3 = .- + - a3 K 1x10-2 I 1 1 ixio-' DL 1 x 1 0 ~ 1x10" iX10-' Lead concentration/g I-' Fig. 4 Calibration graph and detection limit for lead in aqueous solution 0.4 0.3 n v) 0.2 a 0.1 0 lX10-' l X l O d 1 ~ 1 0 - ~ iX10-' Lead concentration/g I-' Fig. 5 Concentration dependence of the RSD of lead in aqueous solution does not lead to an increase in fluorescence intensity. The atomization temperature is reached in about 10 ms under optimum conditions. In this event the fluorescence peak is sufficiently narrow (20 ms) and more intense. The calibration graph for the atomic fluorescence deter- mination of lead in aqueous solutions (the amplitude of fluorescence is measured) is shown in Fig. 4. About 10-20 measurements were camed out for each sample and the mean value of the fluorescence was determined.The calibration graph is linear over a concentration range of 2.5 orders of magnitude. The horizontal line in Fig. 4 corre- sponds to a three-fold background fluctuation. The DL obtained is 3 . 5 ~ lo-' g dm-3 (the absolute DL is 0 . 7 ~ g in a probe volume of 2 ~ 1 ) . Corresponding 1 1 1 I I I . 0 1 2 3 4 5 6 Concentration of HNO (% v/v) Fig. 6 Influence of the solution acidity on fluorescence of lead values of the relative and absolute DLs of lead determined previously with fluorescence measurements above the spiral are 1 x g dm-3 and 2 x g respectively.1° Hence the improvement in DL is 30-fold. Moreover additional measurements showed that the fluorescence signal is inde- pendent of the shielding gas consumption whereas an appreciable consumption of inert gas (7 dm3 min-I) to form the atomic vapour is required with fluorescence measure- ments above the spiral.The dependence of the relative standard deviation (RSD) of lead determination on concentration is shown in Fig. 5. The RSD is 0.3 at concentration levels near to the DL and 0.04 for the concentrations > IODL. The analysis of real water samples requires the investiga- tion of the influence of acidity and of other elements particularly calcium and magnesium. These elements are always present in natural and waste waters in amounts that usually exceed that of lead by some orders of magnitude and may affect significantly the analytical results. The investigations were performed on aqueous lead nitrite solutions containing 1 x g dm-3 of lead in the presence of magnesium nitrate calcium nitrate and nitric acid.It was found that magnesium and calcium at concen- trations less than 0.2 g dm-3 and acidities less than 1% do not affect the fluorescence signal. Higher acid concentra- tions give rise to depression of the fluorescence signal (Fig. 6). The real samples of waste water analysed for lead contain many different elements which were determined by ICP-AES (Table 1). The accuracy of the determination of lead in a waste water sample using the standard additions method is shown in Table 2. Table 2 Accuracy of the determination of lead in a waste water sample Lead added/ Lead determined/ g dm-3 g dm-3 RSD n 1 .o 0.82 f 0.07 0.10 7 2.0 1.8 f 0.06 0.03 5 5.0 4.4 f 0.6 0.1 I 5 8.0 7.1 f 0.4 0.02 4 Table 1 Total composition of the waste water sample analysed Concentration/ Concentration/ Concentration/ Concen trationl Element 1 0-6 g dm-3 Element g dm-3 Element g dm-3 Element g dm-3 A1 22 B 112 Ba 23 Ca 32000 c o 3.5 c u 8 .7 Mn 10 SC 1 1 Fe 33 Na 960 Si 3800 K 840 Pb 36 Sr 180 Li 12 Pt 62 Ta 21 Mg 1400 S 4900 Zn 2.6108 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 Table 3 Detection limits of various elements in aqueous solution Parameter Element Ag Bi Cd c u In Mn Te Zn DU 1 0-6 g dm-3 1.5 2.5 35 250 50 100 1000 0.5 Absolute DU10-'4 g 3.0 5.0 70 500 100 200 2000 1 .o Analogous measurements on aqueous solutions of other elements were also performed. In order to determine other elements one must replace the source electrodeless dis- charge lamp and change the temperature programme of heating the spiral. The DLs obtained are given in Table 3.It should be noted that the DL is limited mainly by the emission from the spiral. As the resonance lines of many elements lie in the ultraviolet region and the maximum emission of the spiral is in the infrared and visible regions the insertion of interference filters (filtering is 1 nm wide) makes it possible to suppress additionally this emission effect and may decrease the DL of the elements. Conclusion The non-dispersive atomic fluorescence spectrometer with a tungsten spiral atomizer has been used successfully in the determination of toxic elements especially lead in waters. The DLs of the elements investigated are at the same level as with electrothermal atomizers.The DL of lead is 3.5 x lo-' g dm-3. The RSD does not exceed 0.04 for concentrations 3 1 ODL. The content of calcium and magnesium present in waste water does not significantly decrease the fluorescence signal. The low cost small size and good analytical character- istics of the atomizer make it suitable for use in environ- mental analysis under stationary and field conditions. References 1 Holliday M. C. Houghton C. and Ottaway J. M. Anal. Chim. Acta 1980 119 67. 2 Sthapit P. R. Ottaway J. M. and Fell G. S. Analyst 1984 109 1061. 3 D'Ulivo A. and Papoff P. Talanta 1985 32 383. 4 Marunkov A. G. Reutova I. B. and Chekalin N. V. Zh. Anal. Khim. 1986 41 68 1. 5 Petrova T. V. Dgerajan T. G. and Sawin S. B. Zh. Anal. Khim. 1990 45 579. 6 Fedorina L. I. Risev A. P. and Solomonov V. A. Zh. Anal. Khim. 1989,44 2088. 7 Omenetto N. Human H. G. C. Cavalli P. and Rossi G. Analyst 1984 109 1067. 8 Arkhangelskii B. V. Gonchakov A. S. and Grazhulene S. S. J. Anal. At. Spectrom. 1987 2 829. 9 Gonchakov A. S. Arkhangelskii B. V. and Grazhulene S. S. Vysokochist. Veschestva (Ultrapure Substances) 1988 No. 6 153. 10 Grazhulene S. S. Khvostikov V. A. Sorokin M. V. Vykhris- tenko N. M. Korovyatnikov G. F. and Gonchakov A. S. in Proceedings of the X i Conference on Analytical Atomic Spec- troscopy. Nauka Moscow 1990 p. 37. Paper 1 /034 I5 D Received July 8 1991 Accepted October 21 1991
ISSN:0267-9477
DOI:10.1039/JA9920700105
出版商:RSC
年代:1992
数据来源: RSC
|
15. |
Precise determination of iron isotope ratios in whole blood using inductively coupled plasma mass spectrometry. Invited lecture |
|
Journal of Analytical Atomic Spectrometry,
Volume 7,
Issue 2,
1992,
Page 109-113
Paul G. Whittaker,
Preview
|
PDF (607KB)
|
|
摘要:
JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 109 Precise Determination of Iron Isotope Ratios in Whole Blood Using Inductively Coupled Plasma Mass Spectrometry* Invited Lecture Paul G. Whittaker and Jon F. R. Barrett University Department of Obstetrics Princess Mary Maternity Hospital Ne wcastle upon Tyne NE2 380 UK John G. Williamst NERC ICP-MS Facility Department of Geology Royal Holloway and Bedford New College Egham Surrey TW200EX UK The feasibility of measuring Fe absorption by incorporation of stable Fe isotopes into red blood cells has been assessed. The clinical protocol necessitated giving 10 mg of 57Fe orally and 0.5 mg of 58Fe intravenously with only two blood samples needed one basal and one after 14 d when most of the tracer is in the circulating red blood cells.Iron absorption was determined by comparison of 57Fe:56Fe and 58Fe:56Fe enrichments which required great confidence in the reliability of 56Fe measurements. Estimation of required precision by theoretical calculations suggested that a relative standard deviation of (0.9% would be required to detect the enrichment in isotope ratios which necessitated finding the optimum Fe signal in relation to the interfering polyatomic species present in both aqueous standards and whole blood samples diluted in buffer. It was found that Fe solutions of 10 ppm or greater routinely gave a precision of (0.4% for 57Fe:56Fe and <O.6% for 58Fe:56Fe approaching counting statistics. Assessment of dead time correction ensured that the concentration of Fe had a negligible effect on the isotope ratios and bias correction by running standards ensured comparability within and across assay occasions correcting for minor variations in blank subtraction for the less abundant isotopes.The use of a range of enriched aqueous and spiked whole blood samples showed that measured and calculated abundances correlated with a slope of unity. Blood samples from two subjects showed that after incorporation of enriched isotopes isotope ratios of 57Fe:56Fe and s8Fe:56Fe were clearly distinguishable (standard deviation >9) from the baseline. Inductively coupled plasma mass spectrometry with conventional aqueous sample introduction can be optimized to give precise measurement of all Fe isotope ratios in whole blood permitting clinical studies of Fe absorption. Keywords lnductively coupled plasma mass spectrometry; iron; blood; isotope ratios; absorption Stable isotopes are increasingly employed as tracers for studies of mineral metabolism in man.Methods have been developed for investigating elements such as Zn and Cu,' Mg2 and Ca.3 They have been applied to studies of Fe absorption in men,' n~n-pregnant'.~ and pregnant6 women and pre-term infant^.^ In recent years isolation of the factors that influence Fe absorption by individuals believed to be at the greatest risk of Fe deficiency (children and women of child-bearing age) have been hindered by ethical considerations that prohibit the use of radioisotopes. However the availability of stable isotope tracer methods should greatly facilitate research with these vulnerable groups.In a study of absorption of Fe during human pregnancy,8 inductively coupled plasma mass spectrometry (ICP-MS) was used with sample introduction by electrothermal vaporization (ETV) to determine Fe isotope ratios in blood serum without prior sample preparation. The ratios s4Fe:56Fe and 57Fe:56Fe were determined in serum taken from non-pregnant women following oral and intravenous administration of enriched 54FeS04 and 57FeS04 respec- tively. Sample introduction by ETV significantly reduces the levels of certain polyatomic ions in particular those associated with the interference of 54Fe ('OAr14N) and 56Fe (40Ar160). In addition ETV is an ideal method of sample introduction requiring minimal sample volume only 5 pl of sample are required for each analysis.A logical progression from the work of Whittaker et a1.8 is * Presented in part at the XXVII Colloquium Spectroscopicum Internationale (CSI) Bergen Norway June 9- 14 199 1 and the 4th Surrey Conference on Plasma Source Mass Spectrometry Guild- ford UK July 15-18 1991. t Invited Lecturer. to assess whether Fe absorption can also be measured by incorporation of stable Fe isotopes into red blood cells (RBC). A few recent studies have used ingestion of a single stable isotope such as 58Fe779 or 54Fe10 to determine Fe availability. However although the patient protocols were simple requiring only two samples of erythrocytes (one basal and one 10-14 d after ingestion when incorporation of a tracer Fe into RBC is complete) single isotope erythrocyte analysis requires an assumption to be made about the level of oral Fe incorporation into erythrocytes this level in fact alters from between 86'O and 93%'' in non- pregnant individuals to 65% in subjects during late preg- nancy.6 The double isotope method'* overcomes the need for these assumptions.It is based on the simultaneous adminis- tration of two isotopes one by the intravenous route and the other by the oral route. This method provides greater accuracy since the comparison of oral with intravenously administered (absorbed 100%) tracer allows compensation to be made for redistribution of the tracer in the body. The RBC method used extensively with two radioactive tracers has been well quantified against whole blood counting. I 3 The double isotope approach is used in the present study to assess Fe absorption.This procedure uses the two least abundant stable Fe isotopes and has potential for safe serial studies both during pregnancy and in the newborn requiring only two blood samples. Iron absorption can be determined by comparison of 57Fe:56Fe and 58Fe:56Fe isotope ratio enrichments. Therefore great confidence in the reliability of 56Fe measurements is required. By using ETV-ICP-MS it was possible to determine Fe isotope ratios in the small volumes of serum available ( I ml or less) containing approximately 1 pg ml-l of Fe by reducing to background levels the polyatomic species that occur at 54 56 and 57 mlz. In the present study Fe concentration and110 JOURNAL OF ANAL,XTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL.7 sample volume were not limitations. The feasibility of using a conventional nebulizer/spray chamber for liquid sample introduction into an ICP-MS instrument for the determina- tion of 57Fe:56Fe and 58Fe:56Fe ratios in the presence of interfering polyatomic species in both simple aqueous standards and whole blood samples is described. Theory Estimation of Required Precision for Dual Isotope Studies of Fe Absorption A normal adult non-pregnant female has a total circulating Fe mass of about 1.56 g; 92 mg as 54Fe 35 mg as 57Fe and 5 mg as 58Fe. Enhancement of these isotopes is limited by the cost of the enriched stable isotopes (54Fe $5 per mg 57Fe f 15 per mg and 58Fe f 100 per mg approximately) and also by the desire to give physiological doses of Fe comparable to dietary intake (about 12 mg per day).The natural abundance of 54Fe is 2.6 times that of 57Fe which means that 2.6 times as much 54Fe would be required to yield the same enrichment as using 57Fe. While the cost benefits are marginal the use of 54Fe would require too much extra Fe in the daily intake which cannot itself be reduced to less than about 5 mg per day from other dietary sources. Conversely the use of 58Fe for oral administration would be very expensive at least in the adult. The theoretical calculations are based on giving the patient an oral dose of 10 mg of 57Fe (average non-pregnant rate of absorption is and an intravenous dose of 500 pg of 58Fe (absorption 100%) and assuming that 93% of both tracers are in the RBC after 2 weeks.'' With this addition the basal 57Fe:56Fe ratio of 0.02389 should change to 0.02454 an enrichment of 2.7% and the basal 58Fe:56Fe ratio of 0.003606 should change to 0.003931 an enrich- ment of 9%.The required relative standard deviation (RSD) for detection [>3 standard deviations (SD) from basal] of these enrichments is <0.9% for 57Fe:56Fe and (3% for 58Fe:56Fe. If an RSD of 0.5% is obtained then the minimum detectable enrichment is 1.5% i.e. 0.02425 for 57Fe:56Fe and 0.003660 for 58Fe:56Fe. This is equivalent to 0.53 mg of 57Fe absorbed and incorporated or a minimum detectable absorption of 5%. Estimation of Fe Signal Required to Obtain the Necessary Precision In order that isotope ratios of Fe can be determined with optimum precision in the presence of polyatomic ions the signal from an isotope suffering from interference has to be significantly greater than the signal of the underlying polyatomic species. In order to obtain such a differential in a conventional ICP-MS system the polyatomic signal would have to be reduced with respect to a given isotope signal or the isotope signal would have to be large with respect to a given polyatomic signal or a combination of both.Fig. 1 shows the relationship between the integrated absorbance of 56Fe and theoretical counting errors of 57Fe:56Fe and 58Fe:56Fe ratios for natural isotopic abun- dance of Fe assuming no polyatomic interferences on the Fe isotopes. In order that counting errors of better than 0.5% could be obtained for both isotope ratios an integrated absorbance of over 1 x lo7 s-l above any background (polyatomic) peak would have to be obtained. The ion count rate can be increased to the required level simply by introducing Fe solutions of a sufficiently high concentra- tion.However if the background signal were large it would be impossible to count an isotope signal above this level as the ion detection system would become saturated and incapable of counting the total ion flux. Optimizing the 0 10 20 Integrated absorbance of 56Fe/106 s Fig. 1 Relationship between the integrated absorbance of 56Fe and the theoretical isotope ratio counting error; A 58Fe:56Fe; and B s7Fe 56Fe ICP-MS system according to the method of Gray and Mlilliams14 ensures that the polyatomic ion at 56 m/z has an equivalent concentration of t 2 0 ng ml-l i.e. the signal from the polyatomic ion is at a low level such that large 56Fe isotope signals can be counted without causing the detec- tion system to become saturated.To summarize in order to obtain Fe isotope ratios with a precision of <0.5% the ICP-MS instrument has to be optimized such that the 40Ar160 signal is reduced to a minimurn. In addition the concentration of Fe in the saLmples has to be sufficiently high to produce the signal necessary to obtain the required precision without saturat- ing the detection system. Experimental Inistrumentation A PlasmaQuad PQ2 -b (VG Elemental Winsford Cheshire UK) inductively coupled plasma mass spectrometer was used for these studies. Details of the instrumentation are given in Table 1. The analysis time for five blanks and ten sample replicates was 40 min.Preparation of Blood Samples and Standards Aqueous solutions of whole blood (1 +24) (which were colllected in a heparinized tube) were prepared according to the following method (modified from the method of Delves and Campbell15). To a 50 ml calibrated flask were added 10 ml of doubly distilled de-ionized water 2 ml of a chemical modifier (0.14 mol dm-3 ammonia solution 0.003 rnd dm-3 disodium dihydrate ethylenediaminetetracetate and 0.029 mol dm-3 ammonium dihydrogen phosphate in water) 2 ml of whole blood 10 ml of Triton X- 100 (5% v/v in water) solution and doubly distilled de-ionized water to vdume. Iron concentrations were generally between 10 and 201 pg ml-l. Blanks and Fe standards (Sigma Poole Dorset UIK) were also prepared using this method.Enriched 57Fe and 58Fe standards were obtained from the UIK Atomic Energy Authority Harwell and Techsnabex- port London UK and made up as iron(@ sulfate by the pharmacy at Northwick Park Hospital (Harrow UK). A miISS analysis is shown in Table 2. Known amounts of these enriched standards were added to aqueous solutions of normal whole blood to give blood solutions of known enrichment.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 111 0.025 0) LL 3% LL u) 0.024 Table 1 ICP-MS instrument details and operating conditions - - Nebulizer Spray chamber Sampling cone Skimmer cone Operating conditions- Forward r.f. power/W Reflected power/W Coolant gas flow/l min-I Auxiliary gas flow11 min-' Nebulizer gas flow Sample solution pump ratelm1 min-l Mass scan range No.of sweeps No. of channels Dwell time/,us Total scan time/s (mass flow controlled)/l min-' 0.0036 0.0035 0) LL 0.0034 e 0.0033 0.0032 De Galan V-groove Single pass in-house design Nickel 1 mm orifice Nickel 0.7 mm orifice - - - - - - 1300 (5 14 0.5 0.75 0.8 50.94-64.92 1500 1024 80 123 Table 2 Mass analysis of enriched isotope preparations At.-% Enriched isotope 54 56 57 58 57FeS04 0 3.00 95.10 1.90 57FeS04 0 0.57 95.93 3.50 58FeS04 0 0.21 6.56 93.20 Clinical Protocol Two normal women attended the research unit following an overnight fast and 10 ml of blood were taken into a lithium heparin tube for determination of basal isotope ratios. Then 250 pg of 58FeS04 were given by intravenous injection followed by 5 mg of 57FeS04 administered orally with 50 ml of fresh orange juice.No food tea or coffee was allowed for 2 h. The following morning the protocol was repeated. After 14 d a 10 ml sample of blood was taken from which the enriched isotope ratios were measured. Results Counting error is a major contributor to precision and it varies inversely with the square root of the number of ions collected. Theoretical calculations of counting errors can be made for both 57Fe:56Fe and 5sFe:56Fe ratios and are shown in Fig. 1. An integrated absorbance of 1 x 1 O7 s-l is required to achieve a counting error of 0.5% for 58Fe:56Fe. In practice this was achieved with about 10 pg ml-l of the Fe standard solution and the calibration line for Fe was found to be linear between 1 and 20 pg ml-l.Eight repeated runs of groups of five replicates of 10 pg ml-l Fe standard samples showed that precision for 57Fe:56Fe varied between 0.13 and 0.29% with an average of 0.21% close to the theoretical counting error of 0.19%. Precision for 58Fe:56Fe varied between 0.30 and 0.90% with an average of 0.64% close to the theoretical counting error of 0.51%. Figs. 2 and 3 show that appropriate dead time corrections of 20-25 ns could be made so that both isotope ratios appeared independent of concentration. 0.026 * 0.023 1 1 I I I 5 10 15 20 25 Dead time/ns Fig. 2 Effect of concentration and dead time on the 57Fe:56Fe ratio A 5; B 10; and C 15 ,ug ml-* of Fe. Error is 3SD 5 10 15 20 ' 25 Dead time/ns Fig. 3 Effect of concentration and dead time on the 58Fe:56Fe ratio A 5; B 10; and C 15 ,ug ml-l of Fe.Error is 3SD The effect of increasing Fe concentration on the precision of both isotope ratios (Table 3) showed that concentrations of greater than 5 pg ml-l of Fe were optimum and equally precise and accurate isotope ratios could be obtained for diluted blood and aqueous standards.112 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 Table 3 Effect of concentration on precision of isotope ratios n= 15 Fe concentration/ Standard- pg ml-I 57Fe:56Fe RSD (O/O) 58Fe:56Fe RSD (%) 1 0.023 1 1 0.79 0.003 2 2 1.90 5 0.02303 0.34 0.00323 0.72 10 0.02322 0.35 0.0032 1 0.34 15 0.0233 1 0.2 1 0.00325 0.35 Blood- 9 0.02320 0.32 0.00322 0.60 18 0.02320 0.20 0.00322 0.38 Measurement of solutions of known isotope ratios showed that results were consistently accurate for all isotope ratios over the whole range of enrichment (Table 4) with correlation of measured and expected abundances giving a slope of unity.In practice minor adjustments in bias were made for unknowns by comparing standards with accepted natural abundances.16 Thus the average bias for 56Fe was 0.19'0 for 57Fe 4.l0h and for 58Fe -2.9% all significant and reflecting minor differences in the accuracy of blank subtraction for the less abundant isotopes. For the isotope ratios this translated into an average of 3.8% for 57Fe:56Fe and - 3.1 Oh for 5*Fe:56Fe. Table 5 gives results for sample 1 (measured 15 times) which was an aliquot of a blood sample taken before the test began (Le. basal). Sample 2 was an aliquot of a blood saimple from the same patient two weeks after ingestion of 10 mg of 57Fe and injection of 500 pg of 58Fe.An Fe standard was then measured before sample 3 (basal) and saimple 4 (enriched) from a second patient were measured. Einrichment of both isotope ratios were clearly significant and therefore estimates of Fe absorption could be derived; about 20% of the given oral dose in these patients. The method of calculating the final absorption figure in a larger group of subjects will be the topic of a further paper. Table 4 Measurement of solutions of known abundance 57Fe abundance s8Fe abundance Sample Enriched standard 1 Enriched standard 2 Enriched standard 3 Enriched blood 1 Enriched blood 2 Basal standard Regression equation* *x=expected y= measured ratio. _ _ _ ~ ~ Expected Measured 95.1 95.13 95.1 95 3 6 95.1 94 78 95.93 95..69 95.93 95.68 6.56 6.60 6.56 6,34 7.01 6,82 2.82 2.84 2.19 2.23 ' Expected Measured 1.9 1.87 1.9 1.82 1.9 1.84 3.5 3.62 3.5 3.65 93.23 93.22 93.23 93.20 0.40 0.40 0.34 0.34 0.33 0.32 y=0.0102+ 1.002x Table 5 Iron isotope ratios in blood from two test patients n = 15 Detection Detection Sample type 57Fe:56Fe RSD (%) limit* 58Fe:56Fe RSD (Oh) limit* 1 Basal 0.02379 0.24 0.023915 0.00353 0.35 0.00 3 5 7 2 Enriched 0.02496 0.13 - 0.00398 0.40 standardt 0.02390 0.23 - 0.00360 0.78 3 Basal 0.0238 1 0.16 0.0239 I 0.00359 0.48 0.00364 4 Enriched 0.02529 0.15 - 0.00409 0.34 * Basal mean+ 3SD.t Standard had 5 replicates.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 113 Discussion The sample preparation method was simple requiring only dilution of whole blood in a buffer.A previous study of Fe in blood by ICP-MS7 used wet ashing with HNOJ before dilution to a similar concentration as that used in the present method. Because it is hoped that a clinical method applicable to human pregnancy can be developed assumptions of isotope incorporation into RBC could not be relied upon and therefore a protocol that compensated for tracer redistribu- tion was required i.e. the use of two isotopes. Janghorbani et aL7 used only one isotope 58Fe by the oral route and monitored 58Fe:57Fe to show that a precision below 1% could be achieved. It has been shown here that a precision of less than 0.6% for both 57Fe:56Fe and 58Fe:56Fe can be attained when Fe concentration and ion intensity are optimized.The high concentration of Fe in whole blood means that only 1 ml of whole blood is required (diluted 1 +24) for 10 replicate analyses at 2 min each. The long measurement times can be accommodated by use of an automatic sampler. Information about the accuracy of the method is also important. It has been shown here that the independence of the isotope ratios from Fe concentration is determined by the use of appropriate dead time correction which compen- sates for counting losses of 56Fe at the upper end of the concentration range. In addition there is close agreement between measured and calculated abundances over the entire range available. Minor correction of bias when measuring clinical samples is achieved by running Fe standards every third sample and using accepted natural abundances since a reference material with certified iso- tope ratios or abundances does not exist for Fe in blood. Thus one can overcome day to day variation in measured natural isotope ratios and thereby use the basal isotope ratio of each subject rather than an average of all subjects as used by Janghorbani et aL7 The difference in Fe isotope ratios before and after incorporation of isotopes into RBC are small but signifi- cant.The enriched samples have an SD39 from the basal ratios. It is therefore clear that the use of 57Fe administered orally gives sufficient RBC enrichment to permit quantita- tive studies of Fe absorption in adults. The reduced blood volume and body Fe stores in children and infants would mean that the protocol could be applied to these groups with resulting greater enrichment. Alternatively the use of 57Fe and 58Fe could be applied to the simultaneous within- subject comparison of food and aqueous Fe absorption.In conclusion ICP-MS with conventional aqueous sam- ple introduction can be optimized to give precise measure- ments of all Fe isotope ratios in whole blood permitting clinical studies of Fe absorption. The authors are grateful to the Royal Society for financial support of the project. The ICP-MS facility is supported by the Natural Environment Research Council. References I King J. C. Raynolds W. L. and Margen S. Am. J. Clin. Nutr. 1978 31 1198. 2 Schuette S. Vereault D. Ting B. T. G. and Janghorbani M. Analyst 1988 113 1837. 3 Fairweather-Tait S. J. Johnson A. Eagles J. Ganatra S. Kennedy H. and Gurr M. I. Br. J. Nutr. 1989 62 379. 4 Janghorbani M. Ting B. T. G. and Young V. R. J. Nutr. 1980 110 2190. 5 Cantone M. C. Molho N. Pirola L. Gambarini G. Hansen C. Roth P. and Werner E. Med. Phys. 1988 15 862. 6 Dyer N. C. and Brill A. B. in Nuclear Activation Techniques in the Life Sciences I.A.E.E. Vienna 1972 pp. 469-477. 7 Janghorbani M. Ting B. T. G. and Fomon S. J. Am. J. Hematol. 1986 21 277. 8 Whittaker P. G. Lind T. Williams J. G. and Gray A. L. Analyst 1989 114 675. 9 Fairweather-Tait S. J. and Minski M. J. Br. J. Nutr. 1986 55 279. 10 Lehmann W. D. Fischer R. and Heinrich H. C. Anal. Biochem. 1988 172 151. 1 1 Larsen L. and Milman N. Acta Med. Scand. 1975,198,271. 12 Saylor L. and Finch C. A. Am. J. Physiol. 1953 172 372. 13 Werner E. Roth P. Hansen C. and Kaltwasser J. P. in Structure and Function of Iron Storage and Transport Proteins ed. Urshizaki I. Elsevier Amsterdam 1983 pp. 403-408. 14 Gray A. L. and Williams J. G. J. Anal. At. Spectrorn. 1987 2 599. 15 Delves H. T. and Campbell M. J. J. Anal. At. Specctrorn. 1988 3 343. 16 Emsley J. The Elements Oxford University Press Oxford 1989. Paper 1/05006K Received September 30 I991 Accepted November 13 1991
ISSN:0267-9477
DOI:10.1039/JA9920700109
出版商:RSC
年代:1992
数据来源: RSC
|
16. |
Application of isotope dilution analysis—inductively coupled plasma mass spectrometry to the precise determination of silver and antimony in pure copper |
|
Journal of Analytical Atomic Spectrometry,
Volume 7,
Issue 2,
1992,
Page 115-119
Koichi Chiba,
Preview
|
PDF (550KB)
|
|
摘要:
JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 115 Application of Isotope Dilution Analysis-Inductively Coupled Plasma Mass Spectrometry to the Precise Determination of Silver and Antimony in Pure Copper* Koichi Chiba lsamu lnamoto and Masao Saeki Materials Characterization Laboratory Nippon Steel Corporation I6 18 Ida Nakahara-ku Kawasaki 2 1 I Japan isotope dilution analysis combined with inductively coupled plasma mass spectrometry was applied to the determination of ultra-trace levels of silver and antimony in pure copper. The precision and sensitivity for the analysis of pure metallic materials was investigated. This technique gives analytical values very close to those of the conventional standard additions method but has higher precision. By using a preconcentration technique it is possible to determine 20 ng g-l of silver and 5 ng g-l of antimony in pure copper while keeping the analytical errors to less than 10%.Keywords inductively coupled plasma mass spectrometry; isotope dilution analysis; pure copper analysis; silver determination; antimony determination Recently there has been a concerted effort to analyse high- purity metallic materials with high sensitivity and high precision. Copper is one of the most important components in electronic devices and the precise determination of ultra- trace levels of impurities especially silver and antimony is required in order to improve the quality of the devices. Inductively coupled plasma mass spectrometry (ICP-MS) makes it possible to determine sub-ng ml-l levels of elements in the aqueous phase.The remarkably high sensi- tivity of the technique has attracted the attention of many researchers since it was introduced by Houk et al. in 1980.' Isotope dilution analysis is noted for its high precision. It does not require calibration graphs for elemental analysis and is not affected by losses and errors during sample treatment procedures. The method has been used primarily with thermal ionization mass spectrometry (TIMS)2 and spark-source mass spectrometry (SSMS).3 Thermal ioniza- tion requires skilful instrument operation and has the limitation that some elements are undetectable. Spark- source MS needs complicated sample preparation for precise measurements. Isotope dilution analysis has found application primarily in the geological field.The combination of ICP-MS and isotope dilution analy- sis is expected to make possible the determination of ultra- trace levels of elements with high precision. Isotope dilution analysis-ICP-MS has begun to be applied to the analysis of environmental ~a'mples,~~~ biological sample^,^*^ food samples,* geological ~amples,~JO water and acids' l - I 2 and reference material ~amp1es.l~ However there have been few applications of the technique to the highly precise analysis of industrial m a t e r i a l ~ . ~ ~ J ~ The primary objective of this work was to determine ultra- trace levels of silver and antimony in pure copper using isotope dilution analysis-ICP-MS. Coprecipitation tech- niques were also applied to isotope dilution analysis in order to increase the sensitivity of this technique.This is valid because isotope dilution analysis is not affected by either recoveries or losses during the coprecipitation procedures. Experimental Instrumentation The ICP-MS instrument used was a PlasmaQuad (VG Elemental Winsford Cheshire UK). The operating condi- * Presented at the XXVII Colloquium Spectroscopicum Interna- tionale (CSI) Bergen Norway June 9- 14 199 l . Table 1 Operating conditions ICP- Power 1 . 3 kW Coolant gas flow rate Auxiliary gas flow rate Carrier gas flow rate 12.4 dm3 min-I 0.48 dm3 min-' 0.82 dm3 min-' Ag Sb Mass spectrometer- Mass range 106.0- 1 10.0 1 19.90- I3 1.56 Channel No. 512 512 Scanning No. 1000 I500 Exposure time per channel 80 ps 80 ps tions are given in Table 1. The total isotope ratio measure- ment time is 40 and 60 s for silver and antimony respectively.The measurement was repeated ten times for each sample and the analytical errors in the isotope ratio measurements were defined as a la deviation. Reagents All acids were of ultra-pure grade from Tama Chemical (Japan) and were used as received. The water used for all solutions was prepared by double distillation of deionized water. Enriched nuclides of silver and antimony were purchased from Oak Ridge National Laboratory (Oak Ridge TN USA). The enriched nuclide of silver consisted of 99.1 1% of Io7Ag and 0.89% of Io9Ag and that of antimony consisted of 97.48% of 123Sb and 2.52% of l2ISb. The spike solutions were prepared by dissolving the enriched nuclides in 1 mol dme3 nitric acid. In the preparation of a stock solution of silver about 0.1 g of silver was dissolved in 20 ml of nitric acid ( I + 1).The test solutions for the investigation of the analytical preci- sion of the isotope ratio measurement were prepared by diluting the silver stock solution. Samples Standard reference materials (SRMs) of copper from the National Institute of Standards and Technology (NIST) (Gaithersburg MD USA) were used as test samples. The SRMs 393 395 and 396 were analysed for silver and SRMs 393 395 and 398 for antimony. Two other copper samples were used to examine the validity of the method 99.99%116 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 Copper sample (1 g) e- HNO (1+1) (10 ml) Spike addition +Te4'; 50 mg addition I Decomposition Evaporation to dryness I - HCI (20 ml) Dissolution -a- SnCI2-2H,O-HCI I Reduction to precipitation Filtration I (Filtrate) (Precipitate) Rejection -1 - Water - HN03(1+1) (10 ml) Dissolution Test solution (50 ml) ICP-MS Fig.1 Preconcentration procedure for isotope dilution analysis of silver in copper and 99.9999% copper samples purchased from Koch-Light (Colnbrook Buckinghamshire UK). Sample Preparation Procedures for Copper Samples When the concentrations of silver or antimony were expected to be more than 1 pg g-l about 0.1 g of copper sample was dissolved in 10 ml of nitric acid (1 + 1) and diluted to 100 ml. The sample solutions were measured directly by ICP-MS. A coprecipitation technique was applied for preconcen- tration of silver and antimony before analysis when the concentrations were expected to be less than 1 pg g-l.The coprecipitation procedures for silver and antimony are summarized in Figs. 1 and 2 respectively. For the determination of silver samples (about 1 g) were decomposed in 10 ml of nitric acid (1 + 1) and diluted to 150 ml. A suitable amount of spike solution was added to the sample solutions followed by 10 ml of 1% tellurium(1v) chloride solution which contained about 50 mg of Te4+ as a coprecipitating reagent. This solution was evaporated to dryness in order to remove the nitric acid because it interferes with precipitation in the next step. The residue was dissolved again in hydrochloric acid and diluted to 100 ml. An aliquot (20 ml) of 20% tin(@ chloride in 6 mol dm-3 hydrochloric acid was added to the sample solution in order to reduce Te3+ for the formation of a metallic tellurium precipitate.The precipitate was filtered and then dissolved again in 10 ml of nitric acid (1 + 1) and diluted to 50 ml. This was the final test solution for ICP-MS measurements. The recovery of the over-all procedure was more than 70%. The concentration of silver was about ten times higher than that of the test solution without the preconcentration treatment. For the determination of antimony samples (about 1 g) were decomposed in 10 ml of nitric acid (1 + 1) and diluted to 50 ml. If some residue remained it was filtered and dissolved in a sulfuric acid-nitric acid mixture. The two solutions were combined to form the sample solution. A Copper sample (1 g) I - HNO (1+1) (10 ml) Decomposition Filtration - H,SO + HNO - Spike addition - La3+; 10 mg addition Heating pH adjustment to precipitation Filtration 1 I c N H l a q ) (1+1) (Filtrate) I (Precipitate) I +-NO ( i + i ) (10 mi) Dissolution c- Water I 1 Rejection Test solut,ion (50 ml) ICPiMS Fig.2 Preconcentration procedure for isotope dilution analysis of a.ntimony in copper suitable amount of spike solution was added followed by I0 ml of 0.2% lanthanum(II1) chloride solution which contained 10 mg of La3+ as a coprecipitating agent. The pH of the solution was adjusted to 9 in order to form a lanthanum(II1) hydroxide precipitate. The precipitate was dissolved in 10 ml of nitric acid (1 + 1) and diluted to 50 ml. This is the final test solution for ICP-MS measurements. The recovery of the over-all procedure was more than 80%.The concentration of antimony was at least ten times higher than that of the test solution without the preconcentration treatment. Isotope Dilution Analysis In isotope dilution analysis the amount of an element is calculated from the isotope ratio using the following equation (1) where x is the molar concentration of the target element in the sample P is its molar concentration in a spike solution R is the measured isotope ratio in the mixture of the sample and the spike a and b are the natural isotopic abundances and A and B are the isotopic abundances in the spike. The analytical error in isotope dilution analysis [the error magnification factor F(Oh)] is defined as follows (2) where R is the measured isotope in the sample solution r is the deviation when the isotope ratio is observed as R and X(R-tr) and X(R) are the molar concentrations of the element determined from the isotope ratio R k r and R according to eqn.( l ) respectively. From eqn. (2) it can be seen that the error magnification factor i.e. the precision in the isotope dilution analysis is affected both by the values of the isotope ratio and by the x= P(A - BR) /(bR - a) F= 1 OO[X(R k r) - X(R)] /X(R)JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 10 E n a 1.0 v) 117 - - W -8 - i 0 10 20 30 40 50 60 Isotope ratio ( R ) Fig. 3 Relationship between error magnification and isotope ratio (lo7Ag:'OgAg) in the determination of silver. RSD in measuring isotope ratio 1 5; 2 I; 3 0.5; and 4 0.1% 10 - ' I .- 0 10 20 30 Isotope ratio ( R ) Fig.4 Relationship between error magnification and isotope ratio ( 123Sb:121Sb) in the determination of antimony. RSD in measuring isotope ratio I 5 ; 2 1; 3 0.5; and 4 0.1% deviations in the measurement of the isotope ratio. The dependences of the error magnification factors on R and r for the determination of silver and antimony are shown in Figs. 3 and 4 respectively. There is an isotope ratio which makes the error magnification factor a minimum when the deviation of the isotope ratio measurements is assumed to be constant because F is a non-linear function of both R and r. The results suggest that the isotope ratio in a mixture of a sample and spike solution should be adjusted by the addition of the correct amount of spike solution in order to increase the precision of the method.Results and Discussion Mass Discrimination Mass discrimination which is the discrepancy between the natural isotope ratio and measured isotope ratio was also investigated. There may be two main causes of mass discrimination effects matrix effects and the bias of the detection system itself. Matrix effects of copper on the isotope ratio measure- ments of silver and antimony are summarized in Table 2. When the amount of copper matrix increases from 0 to 1000 pg ml-l the measured isotope ratios of both silver and antimony remain constant. A copper matrix of 1000 pg ml-1 was thought to be the upper limit of matrix concentration because shot noise was observed. This noise is thought to be due to the neutral particles of copper and/or copper oxide which may be condensed at the end of the plasma.The stability of ICP-MS measurements also de- creased when a solution with a higher dissolved solids 0.1 1 .o 10 100 Ag concentrationtng ml-' Fig. 5 Dependence of precision of isotope ratio measurement on the concentration of silver content was introduced into the plasma. It is concluded that there is no mass discrimination effect due to the copper matrix in measurements of silver and antimony isotope ratios,' because the mass of copper is much less than that of silver or antimony.16 The averages of the measured isotope ratios of silver and antimony were 1.044 and 0.7783 respectively their natural abundances being 1.076 and 0.7463 respectively. It was found that there were small but constant differences between the natural and the measured isotope ratios.These discrepancies are due to the efficiency of the mass detection system for each element. The ratio of the natural and the measured isotope ratios was defined as the mass discrimina- tion factor. Fortunately these types of mass discrimina- tions are easily corrected for by multiplication by the mass discrimination factors. Precision of Isotope Ratio Measurement The analytical precision of the isotope ratio measurement using ICP-MS was investigated. The dependence of the relative standard deviations of isotope ratio measurements on the concentration of silver is shown in Fig. 5 . In common with other analytical methods the precision is in inverse proportion to the concentration.For the determina- tion of silver the relative standard deviation (RSD) of measurement at the 10 ng ml-l level is about 1% and that at the 1 ng ml-I level is about 10%. These deviations together with the measured isotope ratio define the precision of isotope dilution analysis. 0.1 ' I I 0.1 1 .o 10 Ag concentration lpg g-' Fig. 6 Relationship between analytical precision and concentra- tion of silver in the analysis of copper. A Isotope dilution analysis-ICP-MS; and B standard additions method118 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 Table 2 Mass discrimination effect Cu matrix/pg ml-1 Concen tration/ Element ng ml-' 2 5 10 20 Average Natural abundance Io7Ag:'O9Ag Sb 10 20 Average Natural abundance 123Sb:121Sb 0 100 500 1000 1.064 f 0.053 1.057 f 0.046 1.023 f 0.046 1.033 -t 0.070 1.046 f 0.03 1 1.042 f 0.034 1.037 f 0.02 1 1.02 1 +- 0.033 1.048 f 0.0 1 7 1.05 1 2 0.02 1 1.046 f 0.007 1.039 f 0.009 1.048 f 0.01 1 1.049 2 0.0 1 3 1.044f0.011 1.076 0.778 1 f 0.0098 0.7780 f 0.0060 0.7795 ? 0.0069 0.7788 -+ 0.0096 0.7787 f 0.0039 0.776 1 f 0.0059 0.7798 f 0.0075 0.7773 k 0.0059 0.7783f0.0012 0.7467 1.053 f 0.021 1.047 f 0.02 1 Table 3 Determination of silver in copper samples.Mean results in pg g-l in= 10) Without preconcentration- Method* 99.99% SRM 395 SRM 396 copper sample I D-ICP-MS 12.19 f 0.20 3.15 f 0.06 10.36k0.08 ICP-MS (standard additions) lo7Ag 11.9f0.4 3.0 f 0.2 9.8 f 0.4 Certified value 12.2 3.30 - ImAg 12.1 f0.4 3.1 f 0.2 10.1 f0.5 With preconcen t ra t ion- Method* 99.9999% SRM 393 copper sample I D-ICP-MS 0.106 f 0.004 0.2 14 f 0.006 ICP-MS (standard additions) Io7Ag <O.i! 0.26 f 0.07 lwAg <O.i! 0.4 1 f 0.08 ETAAS 0.20 Certified value 0.10 f 0.02 - ID= isotope dilution analysis; ETAAS= electrothermal atomic abscwption spectrometry (graphite furnace). ~~ Table 4 Determination of antimony in copper samples.Mean results :in pg g-I (n= 10) Without preconcentration- Method* ID-ICP-MS Certified value With preconcentration- Method* ID-ICP-MS Certified value ID=isotope dilution analysis. SRM 395 SRM 398 8.0 7.5 7.82 1 f 0.07 1 7.429 f 0.072 SRM 393 39.99% copper sample 99.9999% copper sample 0.128 f 0.04 0.25 f 0.05 - - 0.256 k 0.0 1 1 0.0056 f 0.0006 Analytical Precision of Isotope Dilution Analysis The analytical precision of isotope dilution analysis as a function of silver concentration was investigated and the results are shown in Fig.6. In these experiments the copper samples were analysed after dissolution in order to investi- gate the precision of the over-all analytical procedure. The isotope ratio of the test samples was adjusted to be about 4 by the addition of a suitable amount of spike solution because this isotope ratio gives the best precision according to Fig. 3. The precision is compared with that of the standard additions method in Fig. 6. As can be seen the isotope dilution method has a much higher precision. When a 1 pg g-' concentration of silver in a copper sample is measured by ICP-MS the isotope dilution method gives an RSD of about 3% whereas using the standard additions method an RSD of about 15% is obtained.Isotope dilution analysis has another major advantage when it is necessary to concentrate a target element before determination isotope dilution analysis is free from prob- lems due to recoveries and losses in preconcentration processes. This is because the isotope ratio in a test solution never changes once it has reached a state of equilibrium. Even when the concentration of silver is much lower than 0.1 pg g-l in a copper sample it can be determined with high precision if a suitable preconcentration technique is applied. Determination of Silver and Antimony in Copper Samples Silver and antimony in some copper samples were deter- mined by isotope dilution analysis-ICP-MS. Samples con- taining small amounts of silver or antimony were analysed after applying the preconcentration techniques shown inJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY.MARCH 1992 VOL. 7 119 Figs. I and 2 and the others were analysed without preconcentration. The results for silver and antimony are summarized in Tables 3 and 4 respectively. In the determination of silver the NIST SRMs 395 and 396 and the 99.99% copper sample were analysed without preconcentration. The results obtained by isotope dilution analysis and the standard additions method are in close agreement with each other and both results also agree well with the certified values. The SRM 393 and the 99.9999% copper sample were analysed after preconcentration. The result for SRM 393 obtained by isotope dilution analysis agrees well with the certified value.The results for the 99.9999% copper obtained by isotope dilution analysis the standard additions method and electrothermal atomic absorption spectrometry are in very close agreement with each other. Table 3 shows that the analytical precision of isotope dilution-ICP-MS is much higher than that of'the other methods. With isotope dilution-ICP-MS the analyti- cal error is within 5% even for the determination of 0.1 pg g-l of silver in copper. In the determination of antimony SRMs 395 and 398 were analysed without preconcentration. The SRM 393 and the 99.99% and 99.9999% copper samples were analysed after preconcentration. The results for the SRMs with and without preconcentration agreed well with the certified values. The other results show that isotope dilution-ICP- MS can be used to determine 0.005 pg g-* of antimony within a 10% RSD even after preconcentration.Conclusions Analysis by isotope dilution-ICP-MS was applied to the determination of ultra-trace levels of silver and antimony in high-purity copper samples and was shown to give the same accuracy and much better precision than the conventional standard additions method. It is very easy to combine isotope dilution with preconcentration methods because it is not affected by recoveries and losses in the preconcentra- tion procedures. While keeping analytical errors to less than lo% it is possible to determine 20 ng g-' of silver and 5 ng g-l of antimony in high-purity copper. Hence the combination of ICP-MS with isotope dilution analysis makes very precise measurements of low concentrations of impurities in high-purity materials possible. This method is capable of producing excellent results in a very short time. References 1 Houk R.S. Fassel V. A. Flesch G. D. Svec H. J. Gray A. L. and Taylor C. E. Anal. Chem. 1980 52 2283. 2 Saito T. Shimizu H. and Masuda A. Geochem. J. 1987,21 237. 3 Beary E. S. Paulsen P. J. and Lambert G. M. Anal. Chem. 1988,60 733. 4 Beauchemin D. McLaren J. W. Mykytiuk A. P. and Berman S. S. Anal. Chem. 1987 59 778. 5 Usztity A. Viczihn M. Wang X. and Barnes R. M. J. Anal. At. Spectrom. 1989 4 76 1. 6 Ting B. T. G. and Janghorbani M. Anal. Chem. 1986 58 1334. 7 Ting B. T. G. Mooers C. S. and Janghorbani M. Analyst 1989 114 667. 8 Dean J. R. Ebdon L. and Massey R. J. Anal. At. Spectrom. 1987 2 369. 9 Hall G. E. M. Park C. J. and Pelchat J. C. J. Anal. At. Spectrom. 1987 2 189. 10 Ward D. B. and Bell M. Anal. Chim. Acta 1990 229 157. 11 Paulsen P. J. Beary E. S. Bushee D. S. and Moody J. R. Anal. Chem. 1988 60 971. 12 Haraldsson C. Westerlund S. and Oehman P. Anal. Chim. Acta 1989 221 77. 13 Beary E. S. Brletic K. A. Paulsen P. J. and Moody J. R. Analyst 1987 112 44 1. 14 Makishima A. Inamoto I. and Chiba K. Appl. Spectrosc. 1990 38 697. 15 Umeda H. Inamoto I. and Chiba K. Bunseki Kagaku 199 1 40 109. 16 Tan S. H. and Horlick G. J. Anal. At. Spectrom. 1987 2 745. Paper 1/03950D Received July 30 1991 Accepted November 5 1991
ISSN:0267-9477
DOI:10.1039/JA9920700115
出版商:RSC
年代:1992
数据来源: RSC
|
17. |
Analysis of conducting solids by inductively coupled plasma mass spectrometry with spark ablation |
|
Journal of Analytical Atomic Spectrometry,
Volume 7,
Issue 2,
1992,
Page 121-125
Norbert Jakubowski,
Preview
|
PDF (694KB)
|
|
摘要:
JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 121 Analysis of Conducting Solids by Inductively Coupled Plasma Mass Spectrometry With Spark Ablation* Norbert Jakubowski lngo Feldmann Brigitte Sack and Dietmar Stuewert lnstitut fur Spektrochemie und angewandfe Spektroskopie Postfach 10 13 52 W-4600 Dortmund 1 Germany Spark ablation was applied as an inexpensive technique for direct sample introduction in elemental analysis by inductively coupled plasma mass spectrometry (ICP-MS) with restriction to conducting samples. For these investigations a commercial spark ablation system developed for application in ICP atomic emission spectrometry was coupled with a laboratory-built ICP-MS system. The spectra show only small contributions from disturbing polyatomic species; in particular argides can be neglected.In the analysis of steel detection limits below 100 ng g-I were determined; the precision was about 3%. A calibration procedure was applied to a set of steel standard reference materials. The relative sensitivity factors correspond to values obtained by glow discharge MS with identical MS equipment. Keywords Spark ablation; inductively coupled plasma mass spectrometry; steel analysis From its origin and nature inductively coupled plasma mass spectrometry (ICP-MS) is primarily a solution analy- sis technique. Analysis of solid samples requires the preparation of a solution from the analyte sample which is always a problematic procedure as several sources of systematic errors are additionally included. This is particu- larly true for samples that require severe attack by digestion techniques.Therefore there is increasing interest in direct introduction techniques for solid samples in ICP-MS. For the direct elemental analysis of conducting and non- conducting solids laser ablation has proved to be a powerful and versatile technique of sample introduction. It permits not only bulk analysis but also local- and micro- distribution analysis. However laser ablation equipment is expensive which is an obstacle to its widespread accep- tance. Therefore an inexpensive alternative with similar analytical performance would be of interest for direct analysis by ICP-MS avoiding the problems involved with dissolution of the samples. It was the aim of this work to investigate the analytical capabilities of a spark ablation system for sample introduction in ICP-MS.By its nature this system is restricted to the analysis of conducting in particular metallic samples. Therefore this paper deals with metal analysis only but the investigations will be extended later to the analysis of non-conducting powders compacted with a conducting matrix. Arcs and sparks have a long history in atomic emission spectrometry (AES). Application of the direct introduction of solids to the ICP has been restricted. Human et a/.' were the first to apply spark ablation for the analysis of metallic samples by ICP-AES. Arc ablation was applied by Ohls and Sommer2 for the direct analysis of non-conducting powders by ICP-AES. Aziz et aL3 used a medium-voltage spark source for various types of aluminium samples and for aluminium oxide powders.A more detailed summary of the state of the art in this field has recently been given by Watters et aL4 For application with ICP-MS Jiang and Houk investi- gated arc nebulization for the analysis of metallic samplesS and compacted samples of non-conducting powders.6 Re- cently a spark ablation merging system has also been applied.' These approaches have been made with specially constructed laboratory equipment which precludes their general application. This is the main reason why after some preliminary experiments with an available laboratory spark * Presented at the XXVII ColloquiumSpectroscopicum Interna- t To whom correspondence should be addressed. tionale (CSI) Bergen Norway June 9- 14 199 l . generation system spark ablation with a commercial spark generation system attached to a custom-made ICP-MS system was investigated.The aim of this paper is to report the work carried out so far on the application of this system to the analysis of steel standard reference materials. A comprehensive evaluation of the technique must be post- poned until it has also been applied to the analysis of non- conducting materials. The results are sometimes discussed with reference to the application of glow discharge mass spectrometry (GDMS) because this is a very versatile and efficient technique for the direct elemental analysis of solids studied extensively in previous work.*v9 It is particu- larly useful as a reference here as a GDMS system with identical MS equipment is used in our 1aboratory.IO It should be mentioned that the combination of GDMS and ICP-MS with identical MS equipment is very attractive.This is demonstrated not only by the commercial introduc- tion of a combined instrument (Turner Scientific Wamng- ton UK) but also by the fact that two research groups have already extended their ICP-MS equipment with additional GDMS ion sources.LLJ2 Our approach of using identical MS equipment has the advantage of course that no delays due to adaption work arise when changing over from one technique to the other. Experimental As the ICP-MS system a laboratory-built system that has been described in detail previously was used.1° The system was designed to provide high flexibility for basic investiga- tions. It is a 40 MHz system including x- y- z-positioning of the torch relative to the'MS system and also the bias potential technique which permits simple optimization of the transmission with respect to differences in the ion kinetic energy distributions.10 As the spark ablation system the LISA (Spectro Analyti- cal Instruments Kleve Germany) was chosen which was developed especially for application in ICP-AES.The sample serves as the sealing part of a spark chamber switched as the cathode. The anode is formed by a tungsten tip with an inter-electrode gap of 3.2 mm. The spark frequency is 100 Hz. A ceramic ring with an inner diameter of 7.5 mm is used to restrict the sparking area on the sample surface. The design of the spark source renders the use of a special cyclone for removal of larger particles unnecessary unless the aerosol gas flow rate is adequately chosen.This is the reason why a gas flow rate maintained strictly at 0.6 dm3 min-I is recommended for operation of the spark chamber.122 JOURNAL OF ANAL,YTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 Table 1 Operational parameters for the different stages of the analytical programme Gas flow rate/ Sparking dm3 mine' Stage Voltage/V CurrentlA Source Bypass Time/s 5 Flush Pre-spark 500 2.7 0.6 0.5 10 Stabilize 490 1 .o 0.6 0.5 10 Measure 490 1 .o 0.6 0.5 30 5 Flush - 4 - - - 4 - - Therefore the gas flow is split into a source gas flow and a bypass gas flow which are combined behind the spark chamber to a total gas flow of 1.1 dm3 min-l for optimum ICP operation For cleaning the spark chamber is flushed with an argon flow of 4 dm3 min-I. A fiearly straight poly(tetrafluoroethy1ene) (PTFE) tube (2 m x 4 mm id.) serves as a connection to the torch.No substantial material deposition has been observed to occur inside the tube so far. Changing over from the usual pneumatic nebulization system to the spark ablation system requires less than 2 min. The operational parameters for the different stages of the procedure are given in Table 1. After introduction of the sample a first flushing period serves to clean the sparking chamber. During this period the gas flow is not fed to the torch but to a special 'waste' exit. After a pre-sparking period with higher energy per spark the system is operated under the analytical conditions but a stabilization phase precedes the data acquisition.At the end of the sparking phase the whole system is flushed again with the higher argon flow for cleaning as before which now includes not only the source but also the tubing and the torch. Cooling of the sample did not take place. A problem arises from the material deposition at the interface and also at the skimmer. This makes cleaning of the interface necessary after 10- 1 5 analyses. The material deposition may be reduced by changing the analytical programme so that flushing is not directed through the torch but to the waste exit as is also carried out in the first flushing cycle. Optimization of the parameters for ICP operation results in a sampling distance of 10 mm for the preselected total gas flow rate of 1.1 dm3 min-'.The optimum bias potentiallo was determined to be 0 V. For maximum ion intensity the ICP was operated with a forward power of 1.5 kW which is greater than comparable values for wet aerosols. For these investigations the National Institute of Stand- ards and Technology (NIST) Standard Reference Materials (SRMs) 1 162 1 164 and 126 1 - 1265 Steels and the Bunde- sanstalt fur Material priifung (BAM) certified reference material (CRM) 098-1 were used. As in GDMS analysis the disc-shaped samples were prepared by wet grinding fol- lowed by rinsing with de-ionized water and acetone and finally by drying with compressed air. Before clamping a new sample to the spark chamber the anode was always cleaned with a wire brush. In general registration was performed by a secondary electron multiplier (SEM) in the analogue mode.For direct comparison with GDMS registration by the Faraday cup method was applied to obtain absolute current values. In analytical measurements multiple ion detection was ap- plied with an integration time of 63 ms per data point. For evaluation of the measurements internal standardization was performed using 57Fe as reference. Results and Discussion For a coarse size characterization of the aerosol particles produced by spark ablation a single-jet cascade impactor (Battelle DCI-6) was used. The particle fraction impinging on the last stage and the fraction passing to the exit were collected by membrane filters with a pore diameter of 0.1 pm. The filter load was dissolved in dilute hydrochloric acid and the amount of Fe was determined by ICP-MS.According to the results of these measurements more than 95% of the total analyte mass carried in the aerosol has a diameter below 0.5 pm demonstrating that the spark ablation system produces an extremely fine aerosol which is nearly free of 'large' particles. The total mass flow carried in the aerosol is 2.1 pg s - I while the result of sample weighing indicates an ablation rate of 7 ,ug S - I . Hence the total transport efficiency is relatively high (35%). Fig. 1 particle (size > 10 pm) Scanning electron microscope photograph of a large aerosol 1200 I L 600 5 400 (F -. 2 = 200 120 150 180 0 30 60 90 Time/s Fig. 2 Single ion monitoring profiles of A Co and B V illustrating the measurement procedureJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL.7 123 lxloK A u) C c .- $ .- S 1x10' f! .- 5 = 1x10' - VI Q c 200 - - - I 0 10 20 30 40 mlz 4 - a ,o 3 - 2 0 I \ CI 2 - C 0 - 1 - Fig. 3 Blank spectrum with gas flow on but spark off For investigation of the nature of 'large' particles a conductive adhesive foil (W. Plannet Marburg Germany) was used to collect particles impinging on the last stage of the impactor. Such particles could therefore be submitted to inspection by scanning electron microscopy. A photograph of such a particle is shown in Fig. 1. The structure obviously suggests that it must be produced by gas-phase reactions and not by the ablation process itself. This is in agreement with observations of other w ~ r k e r s . ~ * I ~ J ~ The measurement procedure is illustrated in Fig. 2 showing single ion monitoring profiles for two analyte elements.The analytical signals show a certain increase during the measuring cycle but obviously the ratio is constant suggesting application of internal standardiza- tion. The sharp decay of the signals at the end of a sparking cycle demonstrates that the ablated material is carried out rapidly from the sparking chamber. As the first of a series of spectra Fig. 3 shows the blank spectrum with a logarithmic intensity scale measured with the SEM. The spectrum was obtained under the usual analytical conditions but without sparking. It does not include the main isotope of Ar so that 36Ar with an abundance of 0.3% should be taken as reference. The signals of C N and 0 show lower intensities than the reference peak and in particular disturbances from mole- cular species eg.H20 are of low significance when compared with the arc nebulization work of Jiang and H ~ u k . ~ Fig. 4 shows the spectrum of NIST SRM 1264 Steel measured with analogue registration by a Faraday cup. i - - ~ - ~ r- ~~ I I 0 10 20 30 40 50 60 70 mlz Fig. 4 Spectrum of NIST SRM 1264 Steel with registration by Faraday cup Apart from the isotopes of the matrix only the main isotope of Ar was found. No peaks of ArH or H20 can be seen as the aerosol does not carry a water load. The matrix ion current is about 0.5 nA from which the useful ion yield can be derived as 1 x A. The intensity of the main isotope of Fe is more than three times higher than the intensity of the main isotope of Ar.In GDMS the reverse situation exists* while the total ion currents of both techniques are nearly equal. Another peculiarity of spark ablation ICP-MS becomes obvious in Fig. 5 showing the low-mass region of the spectrum with analogue registration by a multiplier. Here not only the peaks of 1 10 pg g-' of B but also several peaks for doubly charged species in the region around m/z 28 were found so that the determination of Si may become troublesome. The appearance of doubly charged species arising from the main constituents of the analyte corre- sponds to a concentration equivalent of about 0.1 % and is more than one order of magnitude higher than in GDMS. However the often problematic elements P and S may be determined reliably from this spectrum. As an example of a higher mass region Fig.6 shows the spectrum of NIST SRM 1263 Steel from m/z 125 to 145 including the elements Te La and Ce with contents in the low pg g-l region. There is no problem with the evaluation and it can be estimated that in this measurement the detection limit will be below 1 pg g-l. lxloK c L I 0 10 20 30 40 mlz Fig. 5 Spectrum of NIST SRM 1264 Steel containing 1 10 pg g-' of B and 250 pg g-I of S in the low-mass region 2ooo 1 mlz Fig. 6 Part of the spectrum of N E T SRM 1263 Steel containing 9 p g g-I of Te 6 pg g-I of La and 14 pg g-I of Ce124 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 Table 2 Results of a calibration procedure performed with NIST SRMs 1 162 I 164 and I26 1 - 1265 Steels x=element; rn = mass of the isotope evaluated; n = number of standards in which element x is determined; a= slope of calibration graph (relative sensitivity); b=ordinate intercept of calibration line (pg g-'); sr.,,,=mean repeatabilrty i.e. average of relative standard deviations derived from nine scans for each standard (%); srGC=relative standard deviation of measured values from the calibration graph (Oh); and 6cc,,= range of certified contents (ug g-') X mi n a b 4. m Sr c 6CW" B 1 1 7 0.13 0.4.5 3.3 5.6 1.3-1 10 Al 27 6 0.4 1 54.8 2.1 3.2 7-2 400 Ti 48 7 1.17 92.3 2.6 1.7 6-2 400 V 51 7 1.06 17.1 2.6 0.4 6-3 100 Cr 52 6 1.02 97.9 1.8 2.3 72-7 400 c o 59 7 0.79 -2. I 2.2 0.4 70-3 000 Ni 60 7 0.66 99.9 2.0 0.5 410-20000 c u 63 7 0.58 3.6 2.0 0.7 58-5 100 As 75 7 0.23 7.12 3.5 3.1 2-950 Zr 91 6 1.25 42.8 4.7 1.6 90-2 000 Nb 93 5 1.47 - 10.6 4.0 1.0 220-3000 Mo 95 7 1.64 -46.1 2.7 1.3 50-4 900 Sn 120 6 0.89 21.2 3.7 1.2 80-1 040 Sb 121 4 0.46 2.93 3.4 0.7 20-340 W 184 5 0.62 -21.9 3.5 1.7 170-2000 A calibration procedure as used in previous work9 was applied to NIST SRMs 1 162 1164 and 1261-1265.The measuring conditions were chosen so as to realize calibra- tion over the region of contents covered by the certificate values keeping the limit of detection in the low pg g-' region. A sequence of 20 isotopes was chosen for each data scan in the multiple ion detection mode; nine scans were performed with an integration time of 63 ms for each data point. The results are given in Table 2. In particular elements in the mass region of Fe show convincing results and even the determination of As is not problematic as there are no interferences from ArCl in this direct sample introduction technique. Further it is remarkable that a reliable determination of W is possible although this element serves as the anode for sparking.The mean repeatability (cJ Table 2) is about 3.0% as an average for the elements under investigation which can be considered satisfactory with regard to the current state of this work. The same holds true for the scatter above the calibration graph. This is considerably lower than in previous GDMS work which must be attributed primarily to the fact that GDMS has a much lower material consumption of about 1 pg s-l. As the main result of the calibration procedure Table 3 contains the relative sensitivity factors (RSF) correspond- ing to the slope of the calibration graphs (a in Table 2).The sensitivity for B is unusually low but for this element considerable losses in the tubings have been observed previously with other techniques also e.g. owing to de~olvation.~~ Overall the RSF values are within one order of magnitude with an average of 0.83 indicating a good possibility of a semiquantitative procedure with limited accuracy such as may be of interest for analyses of components in the nanogram per gram region where usually no suitable standards are available. Higher positive values of the intercept on the ordinate may be interpreted as possible contributions from interferences as they may arise particularly from doubly charged species. On the other hand negative values may indicate non-linearity of the calibration.In general higher absolute values or the ordinate intercept b may also be a peculiarity of the calibration data and the regression applied. In order to obtain an estimate of the lowest realizable detection limit a reference material (CRM 098-1) was analysed under the most sensitive measuring conditions. This high-purity Fe standard is extremely useful for studying influences of interferences. As part of the resulting spectrum Fig. 7 shows results in the mass region of Mo. In this region there is usually at least in GDMS a problem with the strong inteference from the argide of the main isotope of Fe appearing at m/z 96 which otherwise can only be avoided by use of neon instead of argon as the working gas.16 However an evaluation of this measurement shows that the isotopic pattern is fully reproduced and no interference is affecting the determination of the abundance of 96Mo.Hence spark ablation has the significant advantage that interferences from argides can be excluded. It should be noted that from this measurement the detection limit can be calculated to be about 100 ng g-*. For evaluation of this result one should take into account additionally that our laboratory system was designed with respect to greater flexibility in favour of higher sensitivity. It is known from comparative experiments that with commercial instruments the sensitivity may be about one order of magnitude higher. Therefore we feel justified in expecting that with commer- caal ICP-MS instruments detection limits of about 10 ng g-l can be realized for many elements.Table 3 shows a comparison of the RSF values that have been obtained for some elements by analysis with spark ablation ICP-MS and with low-resolution GDMS including additional values obtained for solution analysis by ICP-MS with pneumatic nebulization. The measurements with GDMS were performed previously with our laboratory- built system which is based on identical MS eq~ipment.~ The results in Table 3 make a correlation between the techniques obvious. The GDMS results were obtained in a region of higher pressure for the argon working gas. The Table 3 Comparison of RSF values with 57Fe as internal standard fix low-resolution GDMS (argon pressure 420 Pa power 5.2 W) for ICP-MS with spark ablation and for ICP-MS with pneumatic nebulization of a solution ICP-MS Element B A1 Cr co Ni c u Pb GDMS 0.55 0.7 1.1 0.83 0.89 0.6 0.3 Spark 0.13 0.4 1 1.02 0.79 0.66 0.58 0.62 Solution 0.36 0.66 1.14 0.97 0.83 0.79 0.25JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL.7 125 I40 90 91 92 93 94 95 96 97 98 99 100 101 mJz Fig. 7 Part of the spectrum of the high-purity Fe standard CRM 098-1 containing 8 pg g-' of Mo strong correlation suggests that under these working condi- tions the behaviour of the sensitivities is mainly deter- mined by the transmission of the MS system and is not influenced by differences caused by ionization or atomiza- tion processes in the source region. The fact that the sensitivities with spark ablation are in general lower than those with the other techniques suggests that certain losses may appear here as have already been observed for a comparable experimental set-up with longer tubing.The agreement with the values obtained for spark ablation is remarkable and indicates that a semiquantitative evalua- tion with a calibration by solution analysis may also be considered if restricted accuracy would be acceptable. Glow discharge mass spectrometry likewise offers good prospects for a semiquantitative analysis and is often applied for this purpose but requires a longer total analysis time owing to the longer stabilization phase necessary to reach sputter equilibrium. Conclusion In comparison with ICP-AES ICP-MS offers the general advantage of a real multi-element technique with signifi- cantly lower detection limits for many elements.For direct introduction of conducting solids in ICP-MS analysis spark ablation has been shown here to be a promising alternative to laser ablation. The main advantage is the use of much less expensive equipment. A high sample throughput is guaranteed by a total analysis time of less than 2 min and no digestion or dissolution procedure is necessary. A satisfac- tory precision of about 3% has been obtained for steel analysis. Detection limits below 100 ng g-' can be realized. The technique hardly suffers from interferences; in particu- lar no interferences from argides contribute above the detection limit. In comparison with GDMS ICP-MS with spark ablation has the advantage of a much shorter analysis time. A disadvantage in comparison with laser ablation is that only conducting solids can be analysed.Further investigations on compacted samples prepared from non- conducting powders with a conducting matrix may demon- strate whether this will be a real restriction. This work was supported financially by the Bundesminister- ium fur Forschung und Technologie and by the Minister- ium fur Wissenschaft und Forschung des Landes Nord- rhein-Westfalen. 1 2 3 4 5 6 7 8 9 10 I 1 12 13 14 15 16 References Human H. G. C. Scott R. H. Oakes A. R. and West C. D. Analyst 1976 101 265. Ohls K. and Sommer D. Spectrochim. Acta Part B 1984,39 1091. Aziz A. Broekaert J. A. C. Laqua K. and Leis F. Spectrochim. Acta Part B 1984 39 1091. Watters R. L. DeVoe J. R. Shen F. H. Small J. A. and Marinenko R. B. Anal. Chem. 1989 61 1826. Jiang S.-J. and Houk R. S. Anal. Chem. 1986 58 1739. Jiang S.-J. and Houk R. S. Spectrochim. Acta Part B 1987 42 93. Hirata T. Akagi T. and Masuda A. Analysl 1990 115 1329. Jakubowski N. Stuewer D. and Toelg G. Int. J. Mass Spectrom. Ion Processes 1986 71 183. Jakubowski N. Stuewer D. and Vieth W. Anal Chem. 1987,59 1825. Jakubowski N. Raeymaekers B. J. Broekaert J. A. C. and Stuewer D. Spectrochim. Acta. Part B 1989 44 219. Kim H. J. Piepmeier E. H. Beck G. L. Brumbaugh G. G. and Farmer 0. T. Anal. Chem. 1990,62 639. Shao Y. and Horlick G. Spectrochim. Acta Part B 1991,46 165. Ono A. Saeki M. and Chiba K. Appl. Speclrosc. 1987 41 970. Raeymaekers B. van Espen P. Adams F. Broekaert J. A. C. Appl. Spectrosc. 1988 42 142. Jakubowski N. Feldmann I. and Stuewer D. Spectrochim. Acta Part B 1992 47 107. Jakubowski N. and Stuewer D. Fresenius' Z. Anal. Chem. 1989,335 680. Paper 1 /03 9828 Received July 31 1991 Accepted September I I I991
ISSN:0267-9477
DOI:10.1039/JA9920700121
出版商:RSC
年代:1992
数据来源: RSC
|
18. |
Rapid and accurate element determination in lubricating oils using inductively coupled plasma optical emission spectrometry |
|
Journal of Analytical Atomic Spectrometry,
Volume 7,
Issue 2,
1992,
Page 127-130
Elisabeth B. M. Jansen,
Preview
|
PDF (424KB)
|
|
摘要:
JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 127 Rapid and Accurate Element Determination in Lubricating Oils Using Inductively Coupled Plasma Optical Emission Spectrometry* Elisabeth B. M. Jansen Joop H. Knipscheer and Mario Nagtegaal Kuwait Petroleum Research and Technology B.V. P.O. Box 545 3190 AL Hoogvliet Rt The Netherlands The internal standard method has been applied to element determination in lubricating oils using inductively coupled plasma (ICP) optical emission spectrometry in order to achieve high accuracy. On-line dilution allows rapid determination without sample pre-treatment. Various lubricating oils can be handled with a single procedure because the effects of lubricant additive components are eliminated. A good analytical performance (relative standard deviation =5%) is achieved under routine operating conditions. Wear particle sizes up to 10 pm are tolerated by the ICP.Keywords Inductively coupled plasma optical emission spectrometry; element determination; lubricating oil; on-line dilution; infernal standard The spectrometric analysis of lubricating oils includes the determination of wear metals and contaminant and addi- tive elements. Engine condition monitoring through used oil analysis is common practice for systems where large amounts of lubricating oils are involved or where engine wear is critical. For this purpose samples are submitted for analysis at regular time intervals. Engine wear patterns are monitored by the wear metal analysis results. Used oil contamination originating from dirt and leaks can be traced by contaminant element determination.Additive packages for lubricating oils consist of among others antiwear agents antioxidants dispersants detergents and viscosity index improvers. Some of these additives consist of organometallic components. The additive element concen- tration which might be as much as several Oh m/m is determined for product characterization and for quality control. X-ray fluorescence (XRF) spectrometry is commonly used for the determination of additive elements in fresh oil. As inductively coupled plasma optical emission spectrome- try (ICP-OES) offers high sensitivity and multi-element analysis capability this technique is applied to wear contaminant and additive element determination in used oils as in the Kuwait Routine Analysis System.The viscosity of lubricating oils is too high for nebuliza- tion with the commercially available nebulizers. Algeo et al.' approached the problem by modifying the nebu- lizer with a sample heater. A more common solution is to dilute the sample with a low viscosity solvent prior to nebulizati~n.~-~ Comprehensive studies on the use of various organic solvents in the ICP have been published by Maessen et and by Nygaard et aL9 Botto7 examined several solvents for applications in the petroleum industry. For lubricating oil application kerosene proved to be the most practical because of its good solvent properties low volatility and toxicity. Large amounts of samples have to be analysed within short time periods therefore accurate results are required with a minimum of sample handling.Granchi et aLS have described a reduction in sample handling by the application of robotics. In our approach a more cost effective method to reduce sample pre-treatment was selected by using an on- line dilution system.6 A 5-fold dilution was applied to maintain sufficient detection power for routine trace ele- ment analysis. The internal standard method has to be applied to Presented at the XXVII Colloquium Spectroscopicum Interna- tionale (CSI) Bergen Norway June 9- 14 199 l . compensate for lubricating oil matrix effects. With this method calculations are performed by using the analyte to Co intensity ratio. The internal standard element Co which is not present in lubricating oils is added to the diluent to obtain analysis without sample pre-treatment.The effect of the viscosity index improver additive was also investi- gated. Experimental Instrumentation The ICP-OES measurements were obtained using a Jobin- Yvon (Longjumeau Cedex France) JY-70-plus system. The spectrometer consists of a monochromator and a dual polychromator. The latter was used for lubricating oil analyses to allow the rapid determination of 2 1 elements in the wavelength range of 179-800 nm. The dispersion is 0.5 nm mm-l in the first order. The spectrometer is equipped with a demountable torch which has an aluminium injector. The V-groove nebulizer was used for sample introduction and the generator operated at 40 MHz. Typical operating conditions for analysis of oil samples are presented in Table 1.The on-line dilution system includes two Gilson peristal- tic pumps and a double-needle system. The latter mounted on the arm of the computer driven sample changer consists of two concentric stainless-steel needles. The dilution solvent is pumped with the first peristaltic pump into the space between the two needles. The second peristaltic pump draws the sample and diluent through the inner tube to the nebulizer. For organic solutions Gilson isoversinic pump tubing was used. Day-to-day variations in dilution rate are compensated for by using standardization samples with each sample batch. The experiments on wear particle sizes were carried out with a Tribometrics (Berkeley CA USA) Model 56 wear particle analyser which collects wear particles as the samples are drawn through a magnetic filter.The increase Table 1 Instrument operating conditions Para meter Power Plasma argon flow rate Auxiliary argon flow rate Sheathing argon flow rate Camer argon flow rate Sample uptake rate Setting 1.2 kW 4 1 min-l 1.2 I min-' 0.3 1 min-' 1 1 min-I 4 ml min-I128 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. ? in magnetic flux is con-verted into micrograms of metallic Fe. Filter sizes vary from 1 to 10 pm. Reagents Multi-element organometallic standard solutions were used for calibration and standardization. For wear metal and contaminant element determination covering the concen- tration range between 1 and 200 mg kg-l Conostan S-21 (Conoco Ponca City OK USA) standards were used. The upper limit for Fe and Cu extended to 1000 mg kg-'.Additive elements were calibrated in the 0.01-2% m/m range using multi-element standards prepared from metal organic concentrates (Angstrom Belleville MI USA). Standards covering the appropriate concentration range were prepared from stock solutions by dilution with base oil (SAE 30 grade). Prior to analysis the standard and sample solutions were diluted with kerosene using the on-line dilutor. In the experiments where the internal standard method was applied calibration graphs were plotted for all elements by using the analyte to Co intensity ratio. Cobalt was added to the diluent at a concentration of 10 mg 1-I. Results and Discussion Viscosity Index Improver Incomplete element recovery was found in the direct analysis of multi-grade engine oils.Severe loss of accuracy occurred with the 5-fold dilution which was the dilution rate used when applying the on-line dilution system as shown in Table 2 for a 15W-40 engine oil. Accuracy improved when the dilution rate was increased however not sufficiently and at the sacrifice of a loss in detection power. Lubricating oils contain chemical additives which give different characteristics to the oils. A typical additive for multi-grade engine oils is the viscosity index (VI) improver. Viscosity index improvers are high relative molecular mass polymers that are added to the lubricating oils to influence the change in viscosity with temperature. Experiments showed that the VI improver had an effect on the nebulizer efficiency and therefore on the element recovery.This effect appeared to be independent of the nebulizer used. During engine operation shear forces will cause the break down of the polymer chains. In used oils therefore the effect on the nebulizer efficiency will decrease with operating time. However even small amounts of the VI improver diminish the accuracy of the analysis results (Fig. .l). This effect can be overcome by applying the internal standard method. The degradation of nebulizer efficiency caused by the VI improver will have a similar impact on the analyte and Co added to the sample by the diluent. The net effect on the analyte to Co intensity ratio is negligible (see Fig. 1). Table 2 Element recovery in the direct analysis of a multi-grade engine oil at various dilution rates Dilution factor Method Recovery (O/O) 100 Manual 92 10 Manual 87 5 On-fine 81 5* On-line 77 * Cross-flow nebulizer.105 A a 85 - 0 3 6 9 12 15 Viscosity index improver (%) Fig. 1 Internal standard (A) versus direct (B) on-line dilution method for multi-grade engine oils 77 I I L 1 L I & I A 0 5 10 15 20 25 30 35 40 -k 0.375 I (b E I 2s X 2s 0 5 0.275' 1 I I 0 0 5 10 15 20 25 30 35 40 Analysis number Fig. 2 Long term stability for wear metal and additive element determination (a) control chart for Fe; and (b) control chart for Ca Analytical Parameters Important analytical parameters for element determination in lubricating oil include accuracy precision stability viscosity and matrix effects. The proposed method showed precision data based on ten repetitive measurements of better than 1 O/o relative standard deviation (RSD).Repeata- bility data published in standard test methodslo-I1 show a 5% RSD for additive element determination in fresh oil and a 10% RSD for wear metals in used oil. Long term stability determined by consecutive measure- ments on a used oil sample over a period of 40 d showed an IUD of 5% both on wear metals and additive elements (Fig. 2). Viscosity effects were virtually absent when the on-line internal standard method was applied. The variation in analysis results between multi-element standard solutions made up in base oil with different viscosities was within the over-all standard deviation (SD) (Fig. 3). The viscosity for xegular used oil samples varied from 5 to 15 mm2 s-l at 100 "C. By contamination with foreign materials or as a result of oxidation processes by the oil itself used oils can have higher viscosities however even in higher viscosity ranges 110 significant effect on the results could be observed.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY. MARCH 1992 VOL.? 129 rn 40 i 30 C + .- 20 c E +I 5 10 s o V Viscosity at 100 "C/mmZ s" Fig. 3 Effect of viscosity on the recovery of various wear metals Most lubricating oils consist of a mineral base oil originating from crude oils. As well as mineral oils synthetic lube oils made by the chemical combination of low relative molecular mass components are also used. The main types of synthetic base fluid are hydrocarbons such as poly(a-olefins) or esters. By using the on-line internal standard method no significant effect could be observed by either type of synthetic fluid compared with the mineral oil (Table 3).This implies that various lubricating oils can be analysed using calibration graphs based on mineral oils. For the on-line direct method a slight effect was observed; however the deviation was still within reasonable accuracy limits. X-ray fluorescence spectrometry is the recognized technique for the determination of an additive element in fresh lubricating oils. As shown in Table 3 the direct XRF method shows large deviations especially with ester-type synthetic base fluids. Other possible interferences occurring in lubricating oil element determination are background shift and spectral overlap. Background shifts which will affect the low determination limits are dealt with by dynamic back- ground correction.The analytical element lines are chosen such that spectral interferences are minimal. Most samples contain Ca in very high concentrations up to 3.5% m/m. Therefore some minor interference by Ca on the analytes cannot be avoided. Calcium spectral interference can be expected on the Si 251.6 nm and on Fe 238.2 nm lines. Accurate interference correction is relatively simple to obtain as interfering elements are included in the analytical programme for lubricating oil element determination. Spectral interference correction both linear and polyno- mial and also dynamic background correction are provided in standard instrument software. Table 3 Lubricating oil matrix effects Recovery (O/O) Internal standard Direct Direct Matrix method (ICP) method (ICP) method (XRF) All mineral 100 100 100 Pol y( a-ole fin) 101 105 109 Ester type 103 102 83 Table 4 Fe particles >2 pm versus total Fe concentration Fe concentration/mg kg-' Total >2 pm 240 22 150 8 100 10 50 3 250 200 v) 0 - .- 5 150 Q .c ; 0 100 f z 50 5 10 15 20 25 0 Particle size/ pm Fig.4 Wear particle size distribution . 100 r " 0 1 2 10 Filter size/pm Fig. 5 ICP-OES measurements on Fe wear metals before and after filtration with the wear particle analyser unshaded area Fe on filter after filtration using the wear particle analyser; and shaded area Fe in sample after filtration Wear Particle Size Engine wear occurs to some extent during normal system operation. For many day-to-day samples the amount and size of Fe wear particles were recorded; some typical measurements are presented in Table 4.In general it can be stated that the average wear particle size is less than 2 pm. This is supported by measurements using an image analysis system (Vidas Kontron) (Fig. 4). It is however important to know the behaviour of the ICP with particles larger than 2 pm which can occur with serious engine failure. The ICP-OES measurements on samples before and after filtration using the wear particle analyser showed that particles up to 10 pm can be tolerated (Fig. 5). Conclusion When applying the internal standard method with on-line dilution element determination in various lubricating oils can be carried out in a single procedure. The effects of lubricant components on the nebulizer efficiency are elimi- nated giving high accuracy and precision data.Interferences are virtually absent or can easily be corrected with standard instrument software. The method was developed for used oil analysis. Additive element determination by XRF is currently the preferred method for fresh oil product characterization because of the high accuracy and the possibilities for direct analysis without sample handling. However these qualities of directI30 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. ? XRF analysis are not valid in certain instances when synthetic base oils are used. The ICP-OES analysis which is the only sensitive and rapid method for wear metal and contaminant element determination in used oil can also be applied to additive element determination. When used with the on-line internal standard method ICP-OES might fulfil the requirements for fresh oil additive element determina- tion. References 1 Algeo J. D. Heine D. R. Phillips H. A. Hoek F. B. G. Schneider M. R. Freelin J. M. and Denton M. B. Spectro- chim. Acta Part B 1985 40 1447. 2 Brown R. J. Spectrochim. Acta Part B 1983 38 283. 3 Mason P. R. Anal. Proc. 1983 20 471. 4 King A. D. Hilligoss D. R. and Wallace G. F. At. Spectrosc. 1984 5 5 . 5 Granchi M. P. Biggerstaff J. A. Hilliard L. J. and Grey P. Spectrochim. Acta Part B 1987 42 169. 6 Evans S. J. and Klueppel R. J. Spectrochim. Acra Part B 1985,40,49. 7 Botto R. I. Spectrochim. Acta Part B 1987 42 181. 8 Maessen F. J. M. J. Seeverens P. J. H. and Kreuning G. Spectrochim. Acta Part B 1984 39 9. 9 Nygaard D. D. Schleicher R. G. and Sotera J. J. Appl. Spectrosc. 1986 40 1074. :lo Deutsche Industrie Norm DIN 51391/3 1991 and DIN 51396/1 1990. 'I 1 American Society for Testing and Materials 1989 ASTM D 495 1-89 Philadelphia PA. Paper 1 /034 14F Received July 8 1991 Accepted September 18 1991
ISSN:0267-9477
DOI:10.1039/JA9920700127
出版商:RSC
年代:1992
数据来源: RSC
|
19. |
Matrix effects of easily ionized elements on the spatial distribution of electron number densities in an inductively coupled plasma using an optical fibre probe and a photodiode array spectrometer |
|
Journal of Analytical Atomic Spectrometry,
Volume 7,
Issue 2,
1992,
Page 131-134
Xiao Jian,
Preview
|
PDF (456KB)
|
|
摘要:
JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 Spectro- PDA detector - meter - (f=l m) ICP source - 131 Peltier cooling system Matrix Effects of Easily Ionized Elements on the Spatial Distribution of Electron Number Densities in an Inductively Coupled Plasma Using an Optical Fibre Probe and a Photodiode Array Spectrometer* Xiao Jian Li Qingyuan Li Wenchong Qian Haowen Tan Jingyuan and Zhang Zhanxia Department of Chemistry Zhongshan University Guangzhou People 's Republic of China The effects of alkali and alkaline earth metals on the spatial distribution of electron number densities (n,) in an inductively coupled plasma have been studied. An optical fibre probe was used to sample the spatial position of the plasma and the photodiode array spectrometer to detect the HP line.The resultant lateral profiles were subjected to an Abel inversion. An on-line intelligent background correction method was used to correct the background under the HP line whose full width at half maximum was then calculated automatically. Graphs of three-dimensional spatial distributions of n in the presence of K Na and Ca are compared with that of water. At an observation height of 7.2 mm above the load coil the n increased with decreasing ionization potentials of the elements. At an observation height of 9.6 mm no measurable increase of n is observed except for K. Keywords Easily ionized element; inductively coupled plasma; electron number density; photodiode array spectrometer; optical fibre probe In an inductively coupled plasma (ICP) the chemical effect of easily ionized elements (EIEs) on the analyte emission intensity has been discussed in the literat~re,'-~ but owing to different operating conditions and lack of explanation at the fundamental level its mechanism has not been eluci- dated convincingly.In previous ~ o r k ~ ~ some enhancement effects of EIEs on the line intensities of some analytes at observation heights (OBHs) lower than 10 mm above the load coil have been observed. For this phenomenon it has been suggested qualitatively that electron enhanced colli- sional excitation of the analyte predominates since elec- trons are responsible for or are involved in collisional excitation de-excitation ionization and three-body recom- bination. Evidently to study the mechanism in more detail it is also necessary to evaluate the variations of electron number densities (n,) at low OBHs. Furuta et aL5 and Blades and Caughlin6q7 have given precise maps of n in the ICP.Blades and Caughlin measured the spatial distribution of n at different OBHs above the load coil using a photodiode array (PDA) spectrometer to detect the HB line and by translating the ICP torch enclosure to 150 separate lateral spatial positions. The resultant lateral profiles were subjected to an Abel inversion and isopleths of n were plotted as a function of both radial and vertical positions. However these isopleths were obtained only under certain operating conditions and no report on the effects of EIEs on the spatial distributions of n was given. Huang and Hieftje8 have reported the simultaneous measurement on spatially resolved electron temperature (T,) n and gas temperature by laser light scattering from the ICP and in a later paper,9 they reported that the addition of Na or K has no measurable influence on either T or n in the plasma zones monitored i.e.at OBHs of 7 and 10 mm. Boumans and De BoerIO have reported that the introduction of a solution with 10 mg ml-I of K in the ICP would lead to an increase in the value of n of 4 x 1014 ~ m - ~ assuming complete ionization of K.I0 The aim of this fundamental study is to identify that the proposed electron enhanced collision process3 is respon- sible for the enhancement effects of EIEs on the analytes. The effect of alkali and alkaline earth metals on the spatial distribution of n in an ICP is evaluated.An optical fibre probe was used to sample the spatial position of the plasma * Presented at the XXVII Colloquium Spectroscopicum Interna- tionale (CSI) Bergen Norway June 9- 14 199 l . and the PDA spectrometer to detect the HP line. An on-line intelligent background correction method" was used to correct the background under the HP line whose full width at half maximum (FWHM) was calculated automatically. The effect of EIEs on the spatial distribution of n specifically within the axial channel of the plasma is discussed. Experimental Instrumentation and Operating Parameters A schematic diagram of the set-up for an optical fibre probe sampling system is shown in Fig. 1. The quartz optical fibre probe is of cylindrical form with a diameter of 0.6 mm and a length of 2 m.The output terminal of the probe is fixed on a three-dimensional adjustable mount and is pressed close to the entrance slit of the spectrometer. The input terminal of the probe is fixed on a two-dimensional adjustable Screen Optical fibre probe $ T P PDA spectrometer I 'I I ICP source xy mount xyz mount Fig. 1 Layout of the optical fibre probe sampling system -- 1 1 U U U Fig. 2 meter Block diagram of a laboratory-constructed PDA spectro-132 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 Table 1 Instrumentation and operating conditions Spectrometer Grating Reciprocal linear dispersion Entrance slit 25 pm WPG- 1 (The Beijhg Second Optical Factory China) Czerny-Turner 1200 grooves mm-' blazed at 300 and 560 nm 0.8 nm mm-I PCD S2304- 1024Q (Hamamatsu Japan) with C2325 drivedamplifier with 1 m focal length Photodiode array system circuit board Sensitive length 25.6 mm Element Wavelength coverage 200-800 nm Spectral window 20 nm Data processor C2890 (Hamamatsu) 1024 and 25 pm element spacing Peltier-based cooling sub-system -10 to -15 "C Computer Sigma compatable IBM PC CPU 8088 Dynamic RAM 512 K Math coprocessor 8087 R.f.generator Argon flow rates Outer gas flow Intermediate gas flow Injector gas flow Sample uptake rate Nebulizer Model EH 2.5-27-111-SDY-2 (Shi Jiazhuang Electronic Processing Institute China) Power set at 1.21 kW (reflected power 130 W) 10 1 min-I 0.3 1 min-I 0.5 1 min-' Meinhard concenlxic glass type (Institute of Non-ferrous Metals China) 2.2 ml min-I Observation height Set as required mount.The image of the plasma is focused on a screen by a quartz lens having a diameter of 36 mm and a focal length of 75 mm. The screen is movable and is placed in the optical axis only as a guide for sampling the required monitored lateral and vertical positions. Thus the orienta- tion of the required positions is straightforward. The block diagram of the laboratory-constructed PDA spectrometer is shown in Fig. 2. The specifications of the spectrometer and the computer and the parameters used in this study are listed in Table 1. The PDA spectrometer provides a convenient method for fast concurrent acquisition of the data required for on-line continuum correction and the Hp line profile. The lateral position was measured at an interval of 0.5 mm and 15-20 points of data were collected.The vertical position was measured at an interval of 2-20 mm. Results and Discussion Measurement of the FWHM of Hp The n has been determined by several w o r k e r ~ ~ v ~ J ~ - ~ ~ on the basis of Stark broadening of the Hp 486.1 nm line and applying the formula of Grieg et al.15 as follows ne= [Co+ C,(lnAA)+ Cz(lnAA)z+ C,(~XIAA)~]AA~/* 1 013 (1) where the coefficient Co=36.84 C1= - 1.430 Cz= -0.133 C3=0.0089. The AA is the FWHM of the Hp line in 8 (1 x lo-* cm) obtained only after Abel inversion. Evidently the accurate measurement of the FWHM of the Hp line is of vital importance. In general because of the Lorentzian shape the true value of the background is difficult to estimate thus leading to an incorrect estimation of the FWHM.In this work the on-line artificial intelligent background correction program' has been used to correct the background under the Hp line. With this program the background could be corrected automatically and the FWHM of the Hj? line could be read instantly from the screen as shown in Fig. 3. Although the Stark broadening of the Hj? line is relatively large compared with the contributions from other line- I 395 I I 3 c 370 .- C = 345 ? E .= 320 -e >. 295 5 270 245 220 - c .- v) c C - 484.52 486.1 o 487.55 Wavelengt hln m Fig. 3 Profiles of HP line obtained from a laboratory-constructed I'DA spectrometer. Peak parameters peak wavelength 486.10 nm; peak width 0.33 nm and 17.00 channel; FWHM 0.24 nm and 11 2.13 channel; peak width at baseline 1.09 nm and 56.00 channel; and peak height 150.69 broadening effects instrumental and Doppler broadening were corrected in this work.By assuming the excitation temperature of the plasma to be 5500 K the Doppler width was calculated to be 0.258 A. Since both entrance and exit slits were 25 pm the instrumental broadening was calcu- lated to be 0.283 A. The net FWHM of the Hj? line was AA=(AA~x,-0.066-0.080)1'2 (2) In order to assess the precision of the measurement of the FWHM of the Hp line three lateral positions namely at 0 (the axial channel) 2.6 and 4.1 mm were sampled five times each at an OBH of 9.6 mm. It was found that the relative standard deviation was 2.9% for the axial channel 2.8% at 2-6 mm and approximately zero at 4.1 mm. The error in the single determination of the FWHM was 0.1 A.By means of the error propagation theory the effect of the error in measuring the FWHM on the calculation of n wasJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 I33 1.34 1.48 1.18 1.3 7 n -z 0.86 -2 0.94 6 1.02 1.12 II) u-0 2 0.7 2 0.76 0.54 0.58 0.38 0.4 1 2 3 4 5 6 Radial position/mm Radial position/mm 1.45 1.27 el 'E 1.09 P 0.91 0 7 0.73 0.55 0.37 7.2 1.39 1.22 1.05 zv 0.98 0 7 0.71 0 1 2 3 4 5 61JTL Radial position/mm Radial position/mm Fig. 4 Variations of three-dimensional distributions of n in the ICP in the absence and presence of EIEs (a) water; (b) 4000 mg I-' of K (c) 4000 mg I-' of Na; and (dj 4000 mg I-* of Ca Table 2 Effects of elements with different IPS on the radial distributions of n at OBHs of A 7.2 and B 9.6 mm above the load coil.Concentrations of K Na Li Ca and Mg=4000 mg I-' Variations of n,(oh) Radial position 0 0.5 1 .o 1.5 2.0 2.5 3.0 3.5 K Na Li Ca A B A B A B A B A 37.6 27.9 20.0 8.4 12.3 4.1 8.5 -5.8 0 42.8 35.2 23.5 11.2 16.4 8.1 11.3 -2.8 5.3 41.8 36.5 23.2 12.7 17.0 10.1 11.6 -1.7 5.2 37.0 33.2 21.0 13.0 16.1 9.5 10.7 -2.0 5.5 30.1 27.3 17.6 12.6 13.8 10.8 9.1 -3.2 6.1 22.8 19.9 14.2 11.4 11.4 10.4 7.7 -4.8 1.1 15.7 11.8 11.0 9.7 9.2 9.1 6.2 -6.8 -4.0 9.0 3.8 8.6 7.4 7.6 7.9 5.3 -8.5 -10.0 B -2.9 6.6 11.2 9.9 9.7 3.3 -2.8 - 9.2 estimated to be in the range 2.2-10% for a FWHM in the range 1.5- 1 .O A. A discussion of the uncertainties caused by Abel inversion can be found in ref. 16. Spatial Distribution of n Potassium Na Li Ca and Mg whose ionization potentials (IPS) are 4.34 5.14 5.39 6.1 1 and 7.65 V respectively were selected to study their effects on the spatial distri- bution of n,.Solutions of these elements (4000 mg 1-l) were run under the operating conditions listed in Table 1. Typical graphs of the three-dimensional spatial distri- bution of n were obtained in the absence and presence of K Na and Ca respectively as shown in Fig. 4(a)-(d). It is notable that in the presence of these elements the over-all spatial distribution pattern of n changed to some extent in comparison with that of water Fig. 4(a) especially at lower OBHs and for elements with lower IP values such as K. In order to study further the variations of n within the axial channel plots of n versus radial positions at OBHs of 7.2 and 9.6 mm are shown separately in Fig.5(a) and (b) and Fig. 6(a) and (b) for K Na Li Ca and Mg. Apparently the n; around the axial channel at an OBH of 7.2 mm above the load coil increased to different extents according to the magnitude of the IPS of the elements. The strength of the enhancement effect follows the order K>Na>Li>Ca>Mg. However for an OBH of 9.6 mm only K exerted a measurable enhancement effect on n and Ca showed some negative effect The increased percentages of n at different radial positions for OBHs of 7.2 and 9.6 OBH are presented in Table 2. In view of the above phenomena it can be concluded that in the presence of EIEs the n at lower OBHs have increased to some extent. The greatest increase is 40% for K. Although this is not significant enough to support our previous s~ggestion,~ it at least confirms that the increase of n which might lead to electron enhanced collision excita-134 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL.7 60- 40 - 20 - D A E I I I 1 1 0 1 2 3 4 5 6 Radial position/mm Fig. 5 Radial distributions of n in the ICP discharge at an OBH of 7.2 mm above the load coil. (a) A water; B K; C Na; and D Ca. (6) A water; D Li; and E Mg 120 (a) 0 1 2 3 4 5 6 Radial position/mm Fig. 6 Radial distributions of n in the ICP discharge at an OBH of 9.6 mm above the load coil. (a) A water; B K; C Na; and D Ca. (b) A water; D Li; and E Mg tion is partly responsible for the enhancement effect of EIEs on the intensities of certain analytical lines. The variation of n in the central channel of the plasma as a function of OBH for K Na and Ca is shown in Fig.7. It 120 r 1 2o t 0 ' 1 I I Observation heighUmm Fig. 7 Variations of n in the central channel of the plasma as a function of OBH for A water; B K; C Na; and D Ca 5 10 15 can be seen that the effects of the elements on the enhancement of n decreased with the increase of OBH. A crossover point (i.e. an OBH where enhancement turns :into depression) could be found for K Na and Ca. It is surprising to find also that the crossover points are related -to the IPS of the elements. This phenomenon may support .the previous suggestion3 that different mechanisms pre- dominate at different OBH values. 'This work was supported by the National Natural Science Foundation (China). References 1 Kalnicky D.J. Fassel V. A. and Kniseley R. N. Appl. Spectrosc. 1977 31 137. 2 Prell L. J. Monnig C. Hams R. E. and Koirtyohann S. R. Spectrochim. Acta Part B 1985 40 1401. 3 Sun D. Zhang Z. Qian H. and Cai M. Spectrochim. Acta Part B 1988 43 391. 4 Sun D. Mo G. and Zhang Z. in Proceedings of the Fourth International Beijing Conference and Exhibition on Instrumen- tal Analysis C. Spectrometry eds. He S. and Tong A. Science Press Beijing 1991 C21. 5 Furuta N. Nojiri Y. and Fuwa F. Spectrochim. Acta Part B 1985 40 423. 6 Blades M. W. and Caughlin B. L. Spectrochim. Acta Part B 1985,40 579. 7 Caughlin B. L. and Blades M. W. Spectrochim. Acta Part B 1985,40 987. 8 Huang M. and Hieftje G. M. Spectrochim. Acta Part B 1989,44 291. 9 Huang M. Yang P. Y. Hanselman D. S. Monnig C. A. and Hieftje G. M. Spectrochim. Acta Part B 1990 45 51 1. 10 Boumans P. W. J. M. and De Boer F. J. Spectrochim. Acta Part B 1977 32 365. 1 1 Huang Y. Zhang Z. Qian H. and Li W. International Congress on Analytical Sciences '91 (Japan) Abstract 4BP3 p. 390. 12 Jarosz J. Mermet J. M. and Robin J. P. Spectrochim. Acta Part B 1978 33 365. 13 Montaser A. Fassel V. A. and Larson G. Appl. Spectrosc. 198 1 35 385. 14 Uchida H. Tanabe K. Nojiri Y. Haraguchi H. and Fuwa K. Spectrochim. Acta Part B 1981 36 71 1. 15 Grieg J. R. Lim C. P. Moo-Yourig G. A. Palumpo G. and Griem H. R. Phys. Rev. 1968 172 148. 16 Cremers C. J. and Birkebak R. C. Appl. Opt. 1966,5 1057. Paper 1 /04458C Received August 20 I991 Accepted January 14 I992
ISSN:0267-9477
DOI:10.1039/JA9920700131
出版商:RSC
年代:1992
数据来源: RSC
|
20. |
Monte Carlo study of analyte desorption, adsorption and spatial distribution in electrothermal atomizers. Invited lecture |
|
Journal of Analytical Atomic Spectrometry,
Volume 7,
Issue 2,
1992,
Page 135-140
Oscar A. Güell,
Preview
|
PDF (686KB)
|
|
摘要:
JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 135 Monte Carlo Study of Analyte Desorption Adsorption and Spatial Distribution in Electrothermal Atomizers* Invited Lecture Oscar A. Guell and James A. Holcombet Department of Chemistry and Biochemistry University of Texas at Austin Austin TX 78712 USA Monte Carlo simulations of electrothermal atomization are used to study the effects of surface re-adsorption and non-uniform spatial distribution as well as the interpretation of the activation energy of atomization and peak shapes. The work is based on an atomization model for Cu. The kinetic parameters are altered to follow their effect upon the measured activation energy of atomization and the spatial distribution inside the atomizer. For elements with similar desorption kinetics re-adsorption shifts the peak to higher temperatures (later times).The width of the peak increases with an accompanying decrease in peak height. Similar activation energies for atomization can be obtained from the linear region at the beginning ( i a early in time) of Smets and Arrhenius plots. These plots are methods that deal with the initial rates and neither linearity nor accuracy should necessarily be expected after the initial portion of the graphs. The activation energy for atomization gives an estimation of the energy barrier for desorption but the adsorption barrier is not available from these plots. Spatially resolved atomization profiles show defined trends during the rising portion as well as at the peak. Stronger interactions with the graphite surface produce a steeper gradient in the gas phase at the beginning of the atomization.The accuracy of the estimation of this energy barrier can be significantly affected by the viewing position of the spectrometer system. Keywords Monte Carlo simulation; atomization mechanism; electrothermal atomization; atomic absorption spectrometry The difficulties encountered in past efforts in developing an understanding of the kinetics and mechanisms for elec- trothermal atomization atomic absorption spectrometry (ETAAS) have slowed the development of models. They have been explained by the complexity of the geometry of the system and the chemical and physical processes occur- ring during atomization. An improved insight is provided by the use of Monte Carlo techniques to study the chemical and physical processes in ETAAS.lv2 Classical mechanistic studies generally start from the experimental determination of an activation energy of atomization (a,,,) for the analyte employed.However most theoretical approaches have neglected the presence of re-adsorption or the finite residence time required for diffusional loss in the absence of very fast removal (e.g. vacuum conditions) as well as the dependence of the absorbance signal on the radial position of observation and the existence of a non-isothermal environment inside the furnace (ie. temperature variations along the length of the atomizer are ignored). The role of re-adsorption has been proposed to be significant for analytes with fairly strong interactions with graphite.Several studies have been made suggesting that the broadening of the atomization peak3 as well as the existence of steeper gradients in the gas p h a ~ e ~ - ~ are a result of these interactions. Holcombe and c o - ~ o r k e r s ~ ~ ~ have shown that the distribution of the analyte in the gas phase is seldom uniform especially early in time. As discussed previously,1 the complex chemistry and geometry associated with the graphite furnace have made the theoretical approach to the application of conventional * Presented at the XXVII Colloquium Spectroscopicurn Interna- t To whom correspondence should be addressed. tionale (CSI) Bergen Norway June 9- 14 199 l . analytical solutions of equations nearly impossible for this system. Most approaches have included simplistic assump- tions and in some cases have ignored the chemical-physical basis for the system.This work presents a further extension to the ETAAS Monte Carlo algorithm and attempts to provide a better understanding of the effects of surface re-adsorption and non-uniform spatial distribution as well as the interpreta- tion of activation energies of atomization and peak shapes in ETAAS. Theory Determination of Activation Energies of Atomization The simplest atomization mechanism in ETAAS can be described by M(surface) = M(g in)-M(g out) (1) where M represents the analyte on the surface or in the gas phase inside or outside the furnace. The standard kinetic expression obtained for this case is where N represents the number of analyte atoms in the gas phase (proportional to the partial pressure of the analyte pM) t is time vd is a pre-exponential factor for desorption from the surface Ordm is the fractional surface coverage raised to an order of release a is the activation energy for desorption R is the gas constant T is the absolute temperature s* is the sticking coefficient for re-adsorption136 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL.7 upon collisions with the surface (O<s*< l) nads is an order of adsorption (unity in most cases) S,(@ represents the fraction of empty sites on the surface (close to unity at all times because of the small sample size used) a d s is an activation energy for re-adsorption and fdlr is a diffusional loss function dependent on the diffusion coefficient Do T and the gas phase concentration gradient Vws.This gradient is determined in part by the geometry of the furnace (e.g. tube length diameter and dosing hole size). The Ey,, values are obtained from absorbance and temperature data usually by Arrhenius-type plots. The basic assumption involved is that the absorbance at any given time is proportional to the rate of atomization:’ (3) at Then neglecting the re-adsorption and diffusion terms in eqn. (2) (i.e. early in time) (4) where r/ includes several constants. In keeping with a standard assumption made for vacuum thermal desorption studies (e.g. ref. 8) Smets9 suggested that for ETAAS it also holds that ( 5 ) and a plot of ___ ~ ~~ Table 1 Monte Carlo simulation conditions unless otherwise specified in the text Description of furnace- Furnace lengthkm Furnace diametedcm Dosing hole diametedcm Ramp rate/K s-’ Starting temperature/K Final temperature/K Sheath gas (argon)- Atomization step Heating program me (experim en tab- Description of analyte sample- Number of particles Distribution Spot diameterkm Offset from centrekm halyte kinetics and digusion- Desorption activation energy1k.I mol-I Pre-exponential factor Order of release Adsorption activation energy/U mol-’ Sticking coefficient Diffusion coefficient at 273 Wcm2 s-I Thermal order of diffusion .Description of simulation parameters- Simulation time incrementh Total time simulatedh 0.90 0.30 0.10 960 723 2373 stop flow 50000 Circular spot 0.10 0 126.0 1 .o 0 1 .o 0.070 1 S O 1 .3 ~ 105 LOX 1 0 - 4 3.0 (hereinafter referred to as a ‘Smelts plot’) should be a straight line with a slope of qlom (i.e.a,) for a first-order desorption. It is often assumed that 8 is constant early in time or that nd,=O and a plot of 1 Ln A versus - RT (hereinafter referred to as an ‘Arrhenius plot’) should yield a linear region with a slope of a,,,. Thus the value of a, gives an estimation of the desorption bamer qes. In reality the detected signal corresponds to the number of gas phase atoms AKNga (6) Hence the Arrhenius plot is a method of investigating initial rates and linearity should not necessarily be expected after the initial portion of the graph. However the cause of non-linearity on Arrhenius plots is still not very clear and has been attributed to multiple-generation functions,’*1° wall diffusi~n,~*l~ changes in particle size,12*13 changes in surface coverage“ and interferences with the supply func- tion by the removal function.1° Bass and H~lcornbe~~ studied the effects of diffusion errors in the assumed order of release and the presence of small systematic errors on Arrhenius plots.They found that these factors produced curvature in the plots similar to those commonly seen in the literature. However in all cases they showed a portion of the graph that was linear and produced the correct activation energy. Thus they did not invalidate the use of these plots. Nevertheless Rayson and Holcombe6 obtained absorbance data from isolated spatial regions within the furnace and found that the apparent a,,,,,. values increased in going from the bottom to the top of the furnace.Further considerations are discussed in this paper. Calculations ‘The program MCGFAA version 1.03.2 was used in the simulations presented here. Details of the algorithm em- ployed are given elsewhere.2 Table 1 describes the simula- tion parameters as well as other additional conditions employed in this study. The Varian CRA-90 furnace size was used for this study. It has a small 9 mm long tube that is heated at the sides with two graphite electrodes. The small volume makes it mostly isothermal [i.e. uniform qt)] and it has been chosen for several mechanistic The absorbance signal is assumed to be proportional to the number of particles in the gas phase throughout. All calculations were performed using a Cray X-MP/24 super- computer located at the University of Texas Center for High Performance Computing.,4tomization Parameters ‘This work is based on the atomization model for Cu (Table l ) which has been previously shown to give approximate experimental atomization profiles of this element.s-16 The kinetic parameters are altered in order to follow their effect upon the measured activation energy of atomization and the spatial distribution inside the atomizer. The value of a,=126 kJ mol-I for Cu is consistent with the work of Black et aL5 and Wang et u1.l’ for atmospheric pressure (ETAAS) and vacuum thermal desorption mass spectro- metry conditions. The pre-exponential factor n d a was determined with the new algorithm and the experimental appearance temperature of Cu. The Do value for Cu in Ar was calculated as described by Skelland’* with data from Hultgren et aI.l9 This value was reduced by 16% in order to obtain a closer fit to experimental profiles from integrated con tact cuvet t e furnaces.2oJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 137 12 000 cn ‘3 .- 10000 5 Q m r 8000 6000 m P c 2 4000 n z 0 5 2000 2 400 2 000 1600 2 .L.’ 1200 $ i? 800 400 n 0 0.5 1.0 1.5 2.0 2.5 3.01 Time/s Fig. 1 Atomization profiles for three similar elements that only differ in their activation barriers for adsorption (as) A 0; B 25.0; and C 126 kJ mol-l. The experimental temperature profile D is also shown. The relative areas under A-C are 265 1,3279 and 4023 respectively Results and Discussion Role of Re-adsorption Fig. 1 shows the atomization profiles from three similar elements that have the same desorption (i.e.a, vd, &.,) and diffusional (i.e. Do nd,f) properties but different activation barriers for adsorption. A low value of a d s is characteristic of a strong metal-graphite interaction. The intermediate value (2OOh of a,,) is typical of physisorption processes whereas a large ads indicates a large activation barrier that produces elastic collisions even at moderate temperatures. The presence of re-adsorption shifts the peak to higher temperatures (later times). Also the width of the peak increases with an accompanying decrease in peak height. These effects imply an increase in the average residence time of the analyte inside the furnace due to an increased time spent adsorbed on the surface. Fig.2 shows the distribution of the analyte particles during the atomization step. The initial sample is located on the furnace wall and after =0,7 s a significant transfer of material into the gas phase begins. The particles in the gas phase collide with the wall and re-adsorb or leave through the furnace openings (ie. the dosing hole or furnace ends). The relative rates of desorption collision and adsorption are also shown. The extreme cases of complete re-adsorp- tion upon collision (i.e. no adsorption barrier) and almost elastic collisions are compared. In the case of efficient re- adsorption ,[i.e. Fig. 2(a)] a gas phase concentration gradient along the tube length is noticed because of the earlier loss of particles through the dosing hole rather than through the furnace ends.The population on the surface remains constant for longer times with significant re- adsorption. It is also important to note the existence of a pseudo-equilibrium between the gas and the surface species because of the similarities in the desorption and re- adsorption rates i.e. E and G curves in Fig. 2(a). In contrast with no re-adsorption [Fig. 2(b)] the character- istics of the analyte are more similar to a diffusion- controlled process. The apparent atomization efficiencies (z.e. the maximum percentage of gas particles at the peak curve B) are 8 and 2 1% with the presence and absence of re- adsorption respectively . Understanding Peak Shape As discussed previously,*l the profile shape is determined by contributions from two Drocesses that split the peak into 50 000 40 000 30 000 20 000 fg 10000 5 - 0 .- a 0 50000 & n E c 40000 30 000 20 000 10 000 - 40000 ( a ) - 30000 - 20000 - 10000 100000 ‘i \ c a 80 000 60 000 40 000 20 000 0 0 0.5 1.0 1.5 2.0 2.5 3.0 Time/s Fig.2 Distribution of analyte particles during the atomization step for (a) total re-adsorption and (b) no re-adsorption. Particles A on the furnace wall B in the gas phase C total leaving the dosing hole and D total leaving the furnace ends. The relative rate of generation (E) wall collision (F) and re-adsorption (G) are also shown with the early profile shape being determined by the kinetics of desorption from the graphite; and a ‘loss’ region at the falling edge with the profile shape at later times governed by diffusional loss. These facts justify the use of a method of initial rates to determine at,,, but might disable further straightening of the Arrhenius-type plots because of the T3jZ dependence of the diffusional functionhif (Do T V,,).Arrhenius and Smets Plots and Eft,,, Values Fig. 3 compares Arrhenius and Smets plots for the situation with no re-adsorption barrier. The final temperature is 10 0 / Y -12 I 1 1 -4 1 ‘ 1 -14 0.050 0.070 0.090 0.110 0.130 (RT)-’/mol kJ-’ Fig. 3 Comparison between A Smets and B Arrhenius plots for the situation with no re-adsorption barrier. The data correspond to curve A in Fig. 1 two overlapping regions a ‘supply’ region at the rising edge I138 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 3 8.0 - Q) 0 - .- 6.0 - + L 2 4.0 - 5 - 5 2.0 - 0 - 0.040 0.070 0.100 0.130 0.160 (RT)-'/mol kJ-' Fig.4 Arrhenius plots obtained from the atomization profiles depicted in Fig. 1. A-C are the same as in Fig. 1. A straight line with a slope of 126 kJ mol-l is shown for reference obtained around 0.053 mol kJ-I producing an upward curvature of the Smets plot. Similar slopes (ato,,,= 126 kJ mol-I) are obtained from the linear region around 0.1 1 mol k J - I ( i e . at the beginning of the atomization). Both plots give a good estimation of the actual desorption energy a,. While the Smets plot is effective in extending the linear region to the left of the curve the extended linear region has a slightly smaller slope (1 14 kJ mol-I) because of the convolution of the loss function with the supply function near the atomization peak.Thus this part of the graph should not be used to obtain an a,,, value. A noticeably higher noise level is present at the useful portion to the right of the curves (around 0.12 mol H-l) resulting in uncertain- ties of 10-20 kJ mol-' in the &,, value. Arrhenius plots are used to obtain a value for a,,, in the remainder of this work but most of the conclusions apply to both methods. Fig. 4 depicts the Arrhenius plots obtained from the three profiles in Fig. 1. The value for a,,, is close to 126 kJ mol-1 at the beginning of all the curves. However by using most of the apparently linear portion values of 1 14 1 1 1 and 126 kJ mol-l are obtained for curves A B and C respectively. This shows that the Arrhenius method is valid only early in time [i.e.high values of (RT)-l] even though the later portion of the plot to the left might look straighter and less noisy. The surface interaction for the analyte in curve B is of intermediate strength and the analyte behaves as a weak interactor at low temperatures (k. as curve C) and like a strong interactor at higher temperatures (ie. as curve A). (a) Zone 10 Zone 7 Zone 4 Zone 1 Zone 10 Zone 1 Fig. 5 Spatially resolved observation zones in the CRA-90 furnace. (a) Magnified side view (0.090 mm2 cross-section each zone); and (b) lateral view the zone windows extend along the furnace length 6000 2400 /---I 0 0.5 1.0 1.5 2.0 2 5 3 Time/s 2000 1600 1200 800 400 * 0 s co 2400 & $ 2000 I- 1600 1200 800 400 0 .o Fig. 6 Spatially resolved atomization profiles using the observa- tion zones depicted in Fig.5; ( a ) and (b) correspond to the A and C analytes introduced in Fig. 1. The temperature profiles are also shown 10 "." 0.060 0.069 0.077 0.086 0.094 0.103 0.111 0.120 (RT) '/mol kJ ' Fig. 7 Spatially resolved Arrhenius plots from the atomization profiles shown in Fig. 6 (a) and (b) correspond to the A and C analytes presented in Fig. 1JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 139 Role of Spatial Distribution The furnace volume was sliced to obtain the ten 0.090 mm2 cross-sectional observation zones shown in Fig. 5. Zones 2-9 were selected for the study since zones 1 and 10 are too close to the furnace wall and would be difficult to observe experimentally. The spatially resolved atomization profiles obtained from analytes with different re-adsorption characteristics are illustrated in Fig.6. Defined trends during the rising portion of the curves as well as at the peak are observed. Similar tendencies have been observed in spatially resolved experimental measurements of several element^,^ although the effect of a single chemical interaction parameter of the analyte can only be studied through Monte Carlo tech- niques. Stronger interactions between the analyte and the graphite surface seem to increase the retention of the particles near the bottom of the furnace and to produce a steeper concentration gradient in the gas phase at the beginning. However the gradient at the signal peak is more prominent when the analyte interactions with the wall are weak. Subsequent to analyte diffusion from the bottom the strong interactors will fill up the volume of the furnace more uniformly since they have a longer residence time as a result of spending more time on the graphite surface.The presence of the dosing hole is partially responsible for the faster diffusional loss at the top of the furnace for the weak interactors. Actually if the dosing hole is removed as for B the gradient is only slightly affected on the rising edge but is 5 180 E c 140 .- Y Zones Fig. 8 Activation energies of atomization from spatially resolved Arrhenius plots corresponding to the three analytes presented in Fig. I . A B and C correspond to those in Fig. 1 respectively. The expected value of 126 kJ mol-l is depicted as a reference by the broken line (0% relative error is also shown as a broken line) .. . . * . . . . . . I ' . . . .. . . . . . * .- ,. . . . . . . .... . ' . .. ... . I .. . . . . . . . . .. -.. ... .;.. . . .. . . .' . . it;{. **.;:. . . . . . . - . noticeably reduced near the peak. Additional effects caused by the presence of the dosing hole have been discussed previously. I Fig. 7 shows the Arrhenius plots from the spatially resolved profiles in Fig. 6. The early gas-phase gradient clearly affects the accurate determination of the a,,, value (i.e. the estimation of aes) because of the increasing slope of the curves in going from the bottom to the top of the furnace. Thus the accuracy of the estimation of a can be drastically affected by the viewing position of the optical system. This is shown in detail in Fig.8. The analyte that is a weak interactor presents a more uniform distribution of E&,, values while the strong interactor presents a larger dependence on spatial viewing. The intermediate interactor behaves closer to the strong at the bottom and closer to the weak at the top of the furnace because of the temperature shift between the curves (cJ Fig. 6). Finally Fig. 9 shows the actual spatial distribution of the particles in the gas phase near the rising portion of the peak for the strong interactor.22 Most of the analyte is concen- trated at the centre of the furnace at this time. The gas concentration gradient from the bottom to the top is evident in this figure. Conclusions For elements with similar desorption kinetics the presence of re-adsorption shifts the peak to higher temperatures (later times) but does not alter the appearance temperature.The width of the peak increases with an accompanying decrease in peak height and apparent atomization effici- ency. These effects imply an increase in the average residence time of the analyte inside the furnace but during the increased time in the furnace the analyte is adsorbed onto the furnace wall. A strong graphite-analyte interaction might yield a pseudo-equilibrium between the species in the gas phase and those adsorbed on the graphite. Since the profile shape seems to be determined by contributions from two processes that split the peak into two overlapping regions*' (a 'supply' and a 'loss' region) the further straightening of the Arrhenius-type plots might be difficult. Similarly the q, value can be relatively ac- curately obtained from the linear region at the beginning (i.e.early in time) of the Smets and Arrhenius plots. The plots are methods of establishing initial rates and linearity should not necessarily be expected after the initial portion of the graphs. The Smets plot is effective in extending the linear region of the curve. However the extended linear region has a slightly smaller slope because of the convolu- tion of the loss function with the supply function near the atomization peak. The value of Es,, gives an estimation of the desorption barrier aes but a value for ads is not available from these plots. The spatially resolved atomization profiles show defined trends during the rising portion as well as at the peak.Stronger interactions with the graphite surface produce a steeper gradient in the gas phase at the beginning of the atomization. The gradient at the peak is more prominent for the weak interactors and is magnified by the presence of the dosing hole. Experimental values of & increase on going from the bottom to the top of the furnace. The accuracy of the estimation of aeS can be drastically affected by the viewing position of the spectrometer system. This work was partially supported by grants from the National Science Foundation (CHE 870424) and Cray Research. Fig. 9 Beginning of the atomization process for a strong interac- tor (a) absorbance-time profile; and (b) side view and; (c) lateral view of the furnace** References 1 Guell 0.A. and Holcombe J . A. Spectrochim. Ada Part B 1988 43 459.I40 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 2 Giiell 0. A. and Holcombe J. A. Anal. Chem. 1990 62 529A. 3 McNally J. and Holcombe J. A. Anal. Chem. 1987 59 1105. 4 Holcombe J. A. Rayson G. D. and Akerlind N. Jr. Spectrochim. Acta Part B 1982 37 3 19. 5 Black S. S. Riddle M. R. and Holcombe J. A. Appl. Spectrosc. 1986 40 925. 6 Rayson G. D. and Holcombe J. A. Spectrochim. Acta Part B 1983 38 987. 7 Sturgeon R. E. Chakrabarti C. L. and Langford C. H. Anal. Chem. 1976,48 1792. 8 Redhead P. A. Vacuum 1962 12 203. 9 Smets B. Spectrochim. Acta Part B 1980 35 33. 10 van den Broek W. M. G. T. and de Galan L. Anal. Chem. 1977,49 2 176. 1 1 L'vov B. V. Bayunov P. A. and Ryabchuk G. N. Spectro- chim. Acfa Part B 1981 36 397. 12 Sturgeon R. E. and Arlow J. S. J. Anal. At. Spectrom. 1986 1 359. 13 L'vov B. V. and Bayunov P. A. Zh. Anal. Khim. 1985,4O,6 14. 14 Guerreri A. Lampugnani L. and Tessari G. Spectrochim. Acta Part B 1984 39 193. 15 Bass D. A. and Holcombe J. A. Spectrochim. Acta Part B 1988,43 1473. 16 Guell 0. A. Ph.D. Thesis University of Texas at Austin 1990. I7 Wang P. Majidi V. and Holcombe J. A. Anal. Chem. 1989 61 2652. 18 Skelland A. H. P. DifSusional Mass Transfer Wiley New York 1974 ch. 3. :I9 Hultgren R. Desay P. Hawkins D. Gleiser M. Kelley K. K. and Wagman D. D. Selected Values of the Thermo- dynamic Properties of the E/ements American Society for Metals Metals Park OH 1973. ;!O Giiell 0. A. and Holcombe J. A. paper presented at the XV Meeting of the Federation of Analytical Chemistry and Spectroscopy Societies (FACSS) Chicago USA October 1-6 1989 paper No. 34. 21 Guell 0. A. and Holcombe J. A. Spectrochim. Acfa Part B 1989,44 185. 22 Giiell 0. A. and Holcombe J. A. Monte Carlo Simulation of A tom iza t ion Processes in Elect rot hermal A tom izers U n i ve r si t y of Texas Austin videotape ed. 1989. Paper I /03000K Received June 19 1991 Accepted July 18 1991
ISSN:0267-9477
DOI:10.1039/JA9920700135
出版商:RSC
年代:1992
数据来源: RSC
|
|