A calibration strategy for LA-ICP-MS analysis employing aqueous standards having modiÆed absorption coefÆcients F. Boue�-Bigne,a B. J. Masters,b J. S. Crightonc and B. L. Sharp*a aDepartment of Chemistry, Loughborough University, Loughborough, Leicestershire, UK LE11 3TU bLaporte Electronics, Amber Business Centre, Riddings, Alfreton, Derbyshire, UK DE55 4DA cBP Amoco Chemicals, Research and Engineering Centre, Chertsey Road Sunbury-on-Thames, Middlesex, UK TW16 7LL Received 7th July 1999, Accepted 15th September 1999 Laser ablation ICP-MS is a powerful technique for the direct elemental analysis of solids with spatial resolution down to a few microns.However, its range of application is limited by the lack of calibration standards. This paper describes a novel approach to calibration that employs aqueous standards whose absorption coefÆcients are modiÆed, by the addition of a chromophore, to produce the desired ablation yield. Chromophores for the important laser wavelengths at 193, 248 and 266 nm are given.The mechanism of ablation and parametric dependences for the modiÆed aqueous standards were investigated and it was concluded that ablation proceeds by a three step process leading ultimately to nebulisation of the bulk liquid. This is important as such a process should not involve fractionation between elements. Calibration curves produced using the aqueous standards were linear and reproducible, but internal standardisation is required to provide linked calibration for real samples.NIST standard Reference Material 613 Trace Elements in Glass and low density polyethylene were analysed and the results agreed with reference values or those obtained by other techniques. 1 Introduction Laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) allows the direct multi-element analysis of solid samples. Qualitative, semi-quantitative, quantitative analyses and isotopic ratio measurements can be performed on a localised spot or the bulk of a sample.The technique is of growing interest in many Æelds, such as the polymer and metallurgical sciences,1±3 the geosciences,4±6 environmental,7 biomedical8 and forensic9 sciences. Although LA-ICP-MS is very attractive, calibration remains a limiting factor when performing quantitative analyses on a wide variety of sample types. No universal method of calibration exists as yet and the ones that are currently used mostly attempt to match the standards to the sample matrix.Different approaches are used to obtain matrix-matched standards. CertiÆed reference materials are commercially available in different types of solid matrices (glass, ceramic, cement, metals, etc.). These reference materials are costly and the range of matrices available does not cover every type of sample (e.g., biological tissue, polymers). Some laboratories prepare their own matrix-matched standards,1,2,7 which are a mixture of an appropriate matrix/binder material and the analytes.Several problems are related to this approach. It is a time-consuming process (in some cases, the sample matrix may not be available, or is not in a pure state). Additionally, these home-made solid standards can have a very heterogeneous distribution of the elements through the matrix substrate. In addition to matrix-matching, internal standards are widely used in LA-ICP-MS for quantitative analyses. The internal standard compensates for the different ablation yields from the sample and the standard and enables the preparation of a calibration graph.It also compensates for the shot-toshot variation and therefore improves the precision of the measurement. Given that the use of an internal standard is required to obtain reliable results, the need for matrixmatching might be regarded as questionable. This paper describes the development of a new method of calibration for quantitative analysis using LA-ICP-MS.It involves the use of aqueous standard solutions whose absorption characteristics are modiÆed by the addition of a chromophore, in order to obtain the desired degree of coupling with the laser energy.10 Preliminary investigations have been carried out on the ablation mechanism of the standard solutions. The new method of calibration was used for the determination of analytes in National Institutes of Standard and Technology (NIST) (Gaithersburg, MD, USA) Standard Reference Material (SRM) 613 Trace Elements in Glass and in Low Density Polyethylene (LDPE). 2 Development and investigation of controlled absorption coefÆcient standard solutions 2.1 Experimental Reagents. The additives used for the modiÆcation of the standard solutions were poly(sodium 4-styrenesulfonate) (Aldrich, Gillingham, Dorset, UK), m-hydroxybenzoic acid (Aldrich), 2-thiobarbituric acid (GPR grade, BDH, Poole, Dorset, UK), 1,10-phenanthroline (ACROS, Fisher ScientiÆc, Loughborough, Leicestershire, UK), nitric acid (Aristar, BDH) and de-ionised water (18 MV).Instruments. Experiments related to the development of the standard solutions took place in two different laboratories (Loughborough University and BP Chemicals, Sunbury-on- Thames); therefore, two instrumental systems are described below. Loughborough University laboratory. An excimer laser, Model EX-742 (Lumonics Ltd., Rugby, UK), was used to ablate the samples. The laser was operated at the KrF transition: 248 nm.All control functions of the laser were performed from a hand-held keypad, which continuously displayed output pulse energy and other relevant operating J. Anal. At. Spectrom., 1999, 14, 1665±1672 1665 This Journal is # The Royal Society of Chemistry 1999parameters of the laser. The laser incorporated a `Stabilase' feature that automatically adjusted the voltage applied to the electrodes to maintain the required output energy.The rectangular shaped laser beam was focused down by a 1 m focal length plano-convex fused silica lens. The long focal length lens was chosen to provide a large depth of focus, which minimises chromatic aberrations, the laser beam having a 1 nm band pass. A mask was positioned after the focal point, in order not to be ablated, to select a section of the beam that provided a Øat energy distribution. A three-element objective was used to image the mask aperture onto the target.This lens combination was optimised to minimise aberrations and to provide high quality imaging at a 15 : 1 reduction. The lenses were coated to minimise reØection losses at 248 nm. Visual control of the ablation was possible with the aid of two systems. The Ærst one combined a retractable mirror, a tube correction lens and an eyepiece, to provide a microscopic view of the target while the laser was off; the second one allowed observation of the ablation of the standard solutions from below the transparent ablation cell, using a mirror positioned at 45� to reØect the image onto a CCD camera. A home-made ablation cell was constructed based upon the design of Arrowsmith and Hughes.11 The ICP-MS used during these studies was a prototype PQ1 instrument (VG Elemental, Winsford, Cheshire, UK).The plasma position, ion optics and gas Øows, were optimised daily to obtain a maximum and stable signal at m/z 115 for a 50 ng ml21 In standard solution. For this particular instrument, a typical count rate of 46106 counts per ppm was achieved.BP Chemicals laboratory. The laser system was originally a Nd:YAG laser (VG laserlab system), emitting at 1064 nm. The original Spectron SL402 (Spectron Laser System Ltd, Rugby, UK) laser was replaced with a Spectron SL282 system to make room for two doubling crystals (DCD and KP*P) in order to obtain emission in the visible (532 nm) and in the UV (266 nm) as well as in the IR. The ICP-MS instrument was a VG Elemental PQ IIz.Chromophore selection. Initial trials to produce a reliable standard solution, using various compounds, established appropriate criteria for the selection of the chromophore. Plasma condition when ablating standard solutions. One of the advantages that laser ablation has over solution nebulisation, as a means of sample introduction, is that l ablation creates a dry plasma. This reduces the amount of oxygen introduced into the plasma and should therefore minimise the occurrence of interfering oxide species.The ablation of solutions was investigated to determine if wet or dry plasma conditions were generated. A solution containing 9.2 g l21 of poly(sodium 4-styrenesulfonate) was ablated at 248 nm to estimate the amount of water being transported to the plasma. This relatively high concentration had previously been determined as being a level that would cause the solution to have an absorption coefÆcient equivalent to that of NIST SRM 613.In this case, the polymer was used without the addition of 2% nitric acid, which reduces its solubility to 4 g l21. Glass tubes of 15 cm length were packed with dried silica gel and sealed at each end with glass wool. The weight of the tubes was accurately measured and recorded. The aerosols generated by laser ablation of a modiÆed solution and conventional nebulisation were allowed to pass through the tubes. The tubes were then re-weighed and the amount of aerosol generated per minute, for each of the sample introduction methods, was recorded.Signal stability. In order to obtain a stable ablation yield, it was thought that either the selected chromophore would have to present a stable absorption at the lasing wavelength or that it would need to be UV stabilised by the use of quenchers. The stability of the chromophores under UV radiation was investigated. The absorption of a modiÆed solution was observed before and after exposure to a mercury pen lamp (l~254 nm). The stability of the ablation yield of the modiÆed standard solution was tested at 266 nm.Optimisation of the use of the modiÆed standard solutions. The inØuence of the laser energy on the ablation yield of standard solutions was studied for different chromophore concentrations. The solutions were ablated at 266 nm; the chromophore used was 2-thiobarbituric acid. Standard solutions were made up, containing 1 mg g21 of Co with 1, 1.8, 2.6 and 3 g l21 of chromophore in 2% nitric acid.Each solution was ablated for 60 s, at different laser energies. Investigation of the ablation mechanism of the standard solutions. To study the ablation mechanism of the modiÆed standards, solutions were prepared having different properties. The following experiments were carried out using the quadrupled frequency Nd:YAG laser (266 nm). The Ærst experiment involved two standard solutions where the additive, 2-thiobarbituric acid, was in solution with two different solvents: H2O and D2O.The two solutions were made up at 3 g l21, giving the same absorption at 266 nm. They were ablated for 1 min; the mass losses after ablation were compared. The mass losses due to the argon Øow above the solutions were monitored to allow the ablated mass to be calculated from the total (evaporationzablation) mass loss. A second experiment involved two different chromophores absorbing at 266 nm: 2-thiobarbituric acid (TBA) and 1,10- phenanthroline (PNT).The two solutions were made up in water to have the same absorption at 266 nm (i.e., 2.13 g l21 of TBA and 1.43 g l21 of PNT). As previously, the mass losses, after 45 s of ablation, were recorded and compared. For all these experiments, each mass loss measurement was repeated ten times in order to apply to them statistical tests to check the validity of the results. Observation and comment on the particular behaviour of mhydroxybenzoic acid as a chromophore in standard solution.It was observed that standard solution freshly made with mhydroxybenzoic acid would turn progressively yellow after a few days. Additionally, the slow kinetics of this reaction were found to be inØuenced by the presence of the analytes in solution, the effect occurring more rapidly with a greater analyte concentration. When ablated with the same energy, the yellow standard solution produced a higher ablation yield than a freshly made solution prepared with the same initial mhydroxybenzoic acid concentration.Determination of the changes that occurred in the modiÆed solution would be indicative of the chemical or physical parameters which are important in the ablation mechanism of the modiÆed standard solutions. The compounds present in the yellow solution were extracted from the acid medium and analysed by MS, 1H NMR and GC-MS. 2.2 Results and discussion Chromophore selection. It would seem logical that an ideal chromophore should be resistant to photodegradation, but as the results presented in the next section show, this is not the critical factor.More importantly the ideal chromophore should have the following characteristics. It should: absorb strongly at the lasing wavelength used for ablation; be soluble in 1±2% nitric acid; not precipitate when in contact with the analytes; and be non-toxic. A suitable chromophore combining these four characteristics was selected for each lasing wavelength.Our efforts were 1666 J. Anal. At. Spectrom., 1999, 14, 1665±1672focused on producing standard solutions for UV laser ablation as this results in a more efÆcient ablation of solid materials than IR ablation and is less prone to fractionation. The chromophores are listed below with their corresponding speciÆc absorptivity and maximum solubility in acid solution. For the 193 nm excimer laser two chromophores were found: (i) nitric acid, e193~3 l g21 cm21, used at 1±2% in solution; and (ii) poly(sodium 4-styrenesulfonate) (see Fig. 1a), e193 ~85 l g21 cm21, maximum solubility in acid solution~4 g l21. For the 248 nm excimer laser: m-hydroxybenzoic acid (see Fig. 1b), e248~30 l g21 cm21, maximum solubility in acid solution~3 g l21. For the 266 nm frequency quadrupled Nd:YAG laser: 2-thiobarbituric acid (see Fig. 1c), e266~80 l g21 cm21, maximum solubility in acid solution~4 g l21. Plasma condition when ablating standard solutions. The amount of aerosol generated by ablation, at 248 nm, of a standard solution containing 9.2 g l21 poly(sodium 4-styrenesulfonate) was 626 mg min21 against 23 550 mg min21 for conventional nebulisation.The plasma was much less loaded with water vapour when ablating the standard solutions. Importantly, it was found that the optimum operating conditions for the standards were similar to those for the solid samples suggesting that the plasma and interface conditions were broadly similar.If this were not the case, the internal standard would have to compensate for differential atomisation/ionisation effects in addition to those related to ablation yield and transport. Identifying such an internal standard for multi-element analysis would be difÆcult, if not impossible. Signal stability. The selected chromophores were found to be unstable under UV irradiation from the mercury pen lamp (257 nm). The absorbance of the standard solutions varied due to the formation of photo-products and mostly resulted in a decrease of the absorbance of the standard solution at the wavelength of interest.For example, Fig. 2 shows the decrease of absorption of a modiÆed solution (6 mg l21 of 2-thiobarbituric acid in 2% nitric acid) exposed to a mercury lamp. However, for poly(sodium 4-styrene sulfonate), a stable absorption was observed at 248 nm after irradiation with UV photons. The apparent stability at 248 and 266 nm was due to the fact that an absorption band with a peak at 224 nm broadened and Øattened on exposure to the UV radiation.Fig. 3 shows the signal stability of a solution containing 3 g l21 of TBA and Fe, Mo and W at 10 ppm in 2% nitric acid ablated at 266 nm. The signal is stable for much longer periods than this, provided that the laser focus is sustained. However, the period shown is relevant to typical acquisition times for calibration purposes. This shows that photo-stability is not critically important.There are two principal reasons for this; Ærst, most of the volume of aqueous standard strongly irradiated by the laser beam is efÆciently ablated and therefore removed, unless the energy threshold value has not been reached; second, the liquid state of the standard offers a renewable surface of fresh solution during the ablation event. Further to these points, making the reservoir volume (typically 1 ml) very much larger than the ablation volume ensures that a large excess of fresh chromophore is always present and also helps to reduce the effects of evaporation, which would otherwise cause a gradual increase in analyte concentration. Thus, the chromophore, used to modify the absorbance of the solution, does not have to be photo-stable to provide stable ablation yield over a long period of time.Given this observation, poly(sodium 4-styrenesulfonate) can also be expected to be an efÆcient chromophore at 193 nm where it has a high speciÆc absorptivity (85 l g21 cm21).Optimisation of the use of the modiÆed standard solutions. A simple aqueous standard solution containing nitric acid does not absorb the UV laser energy unless an ArF excimer laser that emits at 193 nm is used. Very little ablation occurs, if any, and the beam penetrates deeply into the solution and may ablate the containment vessel. The purpose of the chromophore is to improve coupling with the aqueous solution in such a way that the threshold ablation Øuence is reached within the surface layers of the liquid, thereby producing a Æne aerosol that is readily transportable to, and easily processed by, the plasma.The inØuence of chromophore concentration and laser energy on the ablation yield of the standard solutions is shown Fig. 4. The addition of chromophore to the standard solution reduces the threshold Øuence value. This reduction of threshold Øuence has also been observed in experiments involving the doping of solid materials12,13 in order to improve their ablation Fig. 1 Structures of (a) poly(sodium-4-styrenesulfonate), (b) m-hydroxybenzoic acid, and (c) 2-thiobarbituric acid. Fig. 2 Decrease of the absorption of a modiÆed solution (6 mg l21 of TBA in 2% nitric acid) exposed to a mercury lamp. Fig. 3 Stability of the ablation yield of a modiÆed standard solution (3 g l21 of TBA and Fe, Mo andWat 10 ppm) ablated at 266 nm, at the energy optimised for LDPE, and at 10 Hz. J. Anal.At. Spectrom., 1999, 14, 1665±1672 1667at selected wavelengths. In the case of the standard solutions, the addition of the chromophore offers the user the freedom to prepare standards that give a signal of the desired amplitude, when ablated with a Øuence pre-selected to be optimal for the ablation of the sample. The required concentration can be obtained directly from Fig. 4. For the analysis of a sample requiring a low laser Øuence for ablation, a relatively high chromophore concentration will be selected in order to decrease the threshold Øuence value and to obtain an efÆcient ablation (and hence an appropriate analyte count rate) of the standard solution at low energy.For a sample requiring a high laser Øuence to be ablated, a relatively low chromophore concentration will be selected. High laser Øuence was found to produce a violent explosive ablation event yielding large aerosol droplets that were visibly deposited onto the ablation chamber walls. An alternative strategy for determining chromophore concentration when analysing transparent materials is to measure the absorption coefÆcient of the target substrate and to match the solution absorption coefÆcient to this value. This was previously attempted for glass samples8 and surprisingly yielded similar analyte count rates, to within an order of magnitude, for the same analyte concentrations.This approach is of interest for comparing ablation yields, but is unlikely to be used for practical analyses.Investigation of the ablation mechanism of the standard solutions. The results of the H2O±D2O experiment are presented in Table 1. As the masses measured were of the order of 1 mg, a two-tailed t-test was applied to the results to check their validity. It was found that there was a signiÆcant difference between the two means at the 99.9% conÆdence level. The mass loss due to ablation from the aqueous solution was 2.6-fold greater than from the deuterated solution.The process by which the energy absorbed by the chromophore is coupled into the water matrix is complex and could proceed by a variety of routes. However, the direct conversion of several electronvolts of energy into translational motion of a single molecule is unlikely when that molecule is Ærmly embedded in the elastic water matrix. Whether the transfer route is via direct quenching, internal conversion or dissociation, the Ænal step in the energy cascade is most likely to occur through a vibrational process.The experiment with D2O seeks to verify this hypothesis. The physical properties of H2O and D2O are not very different. The boiling point of D2O is 101.4 �C compared with 100 �C for H2O and similarly, the enthalpy of vaporisation is 169.9 kJ mol21 for H2O and 174 kJ mol21 for D2O. These data derive from the fact that the O±H covalent bond is stronger in water than in D2O, which in turn leads to a weaker H bond in H2O compared with the equivalent bond in D2O.However, these properties are insufÆcient to account for the differences found in the experiments. Such differences could only be accounted for if the initial energy transfer from the chromophore to the H2O or D2O matrix is resonantly coupled through a principal or overtone vibrational transition of the solvent. H2O and D2O have signiÆcantly different absorption band patterns in the IR. An example of this highly speciÆc type of energy cascade has been described for the sensitizer [tetrakis(4-sulfonatophenyl) porphine] (TPPS) excited at 532 nm.14 In this case the triplet excited state of TTPS is quenched by O2, which in turn is excited to its singlet state.The measurement of the lifetime of the O2 singlet excited state showed the extent of its coupling with the solvent. This lifetime was found to be shorter when the solution was prepared in H2O than in D2O because H2O has an overtone which directly overlaps with an absorption band in O2.Attempts to maintain a degassed solution in our experimental arrangement proved unsuccessful. The results obtained for the experiment involving two chromophores, TBA and PNT, are shown in Table 2. As previously, a two-tailed t-test shows that there is a signiÆcant difference between these two means at the 99.9% conÆdence level. The solution containing PNT produced an explosive ablation compared to the one with TBA. Four times more solution was removed during the explosive ablation.Two reasons can be advanced to explain the difference found. TBA and PNT were, respectively, at 14.761023 mol l21 and 861023 mol l21 to give the same absorbance at 266 nm. Therefore the laser energy was absorbed by approximately twice the number of TBA molecules compared with PNT. This contributes to a more dispersed partitioning of the laser energy into the TBA solution. Additionally, as shown in Figs. 5a and 5b, TBA molecules present more sites of interaction with the solvent than PNT; therefore, the energy of its excited state is released in a more dispersed way to the solvent.This results in the creation of a Æne aerosol that is readily transportable to the plasma. PNT produces a more localised coupling to the solvent, resulting in the removal of large droplets that are not efÆciently transportable beyond the ablation cell. Contamination of the cell with large water droplets was readily observable under these conditions.The end result of coupling the laser energy to the liquid appears to be a rapid and local increase in temperature and pressure leading to nebulisation of the bulk solution. This is important as such a mechanism should be free of fractionation effects. These experiments underline the fact that the quality of aerosol produced (i.e., the transportable mass) is equally important to the initial ablation yield. Examination of the data in Fig. 6 indicates an approximately Fig. 4 InØuence on the ablation yield of the laser voltage for different concations of 2-thiobarbituric acid, at 266 nm. Table 1 Comparison of ablation yield of standard solutions with difference solvents Chromophore 2-Thiobarbituric acid 2-Thiobarbituric acid Solvent H2O D2O Ar Øow effect Total mass loss Ar Øow effect Total mass loss Average of mass loss/mg 1.5 2.8 1.4 1.9 RSD (%) n~10 12.8 9.2 5.6 9.3 Mass loss due to ablation/mg 1.3 0.5 1668 J. Anal. At.Spectrom., 1999, 14, 1665±1672exponential relationship between ablation yield and chromophore concentration and is suggestive of a process that might be modelled by simple application of Beer's law. This, however, is not the case. A simple form of Beer's law used for LA modelling13 is: For P0wPc: Pc~P0e{kn sLc Ö1Ü or Lc~ 1 kn s ln P0 Pc Ö2Ü where P0 is the incident power, Pc is the threshold power required for ablation to begin, kn s is the absorption coefÆcient of the sample, and Lc is then the depth of material ablated by the pulse.Clearly for constant Pc and Po, increasing kn s requires that Lc is reduced, that is, the volume of material ablated (Lcpr2) (where r is the beam radius) should be reduced as the absorption coefÆcient is raised. That this does not appear to apply here is indicative of other factors being signiÆcant, viz.: (i) this simple analysis assumes that power is absorbed more or less instantaneously before the ablation process begins, which may be true with ps lasers, but not with lasers such as those used here with 15 ns pulse widths; (ii) the power dissipated in the ablated volume, i.e., the power density, DPD~ÖP0{PcÜ=pr2Lc Ö3Ü does rise as the absorption coefÆcient is increased; (iii) there will be attenuation of the beam by coupling to the plasma formed above the liquid surface.The Ærst point to make is that, as shown in Fig. 6, Pc is reduced by increasing the absorption coefÆcient which will improve the yield.Conversely, for a transparent material, Pc increases asymptotically as kn s tends to zero15 so that no ablation occurs. Secondly, because the power dissipated per unit volume is increased, more energy becomes available to convert the sample to a transportable aerosol that will move away from the ablation surface and ultimately reach the plasma. The importance of local transport has recently been demonstrated by Frischknecht et al.16 who showed that exchanging Ar for He, a less dense gas, enhanced ablation yield, although this may also be partly due to modiÆcations of the ablation plasma.Given the relatively long laser pulse it can be anticipated that during the ablation process there will be a moving ablation boundary13 so that the simple absorption model will actually apply in a series of successive time increments with the target being gradually shielded from the beam as material is removed and an ablation plasma develops.Obviously a great deal more work would be required to provide a quantitative model of the ablation yield for such a complex process. Observation and comment on the particular behaviour of mhydroxybenzoic acid as a chromophore in standard solutions. Initial MS analysis of the aged yellow standard solution showed a peak at Mz46, which indicates the presence of a nitrated form of the initial chromophore. Comparative analyses were then performed by 1H NMR and GC-MS between m-hydroxybenzoic acid, a nitrated isomer of mhydroxybenzoic acid, 3-hydroxy-4-nitrobenzoic acid and the unknown yellow mixture.The samples were methylated using BF3±methanol to make the chromatography easier. Both types of analysis conÆrmed that the yellow solution was mainly a mixture of the initial chromophore m-hydroxybenzoic acid and 3-hydroxy-4-nitrobenzoic acid; various ketones and diesters were formed at very low concentrations. 3-Hydroxy-4-nitrobenzoic acid is much less soluble than m-hydroxybenzoic acid in 2% HNO3.The presence of the 3-hydroxy-4-nitrobenzoic acid in the solution may lead, like PNT previously, to a less dispersed release of energy from the excited chromophore to the solvent, producing more localised heating and a higher ablation yield. The nitration of m-hydroxybenzoic acid was not expected, but obviously occurs, and appears to be catalysed by the presence of metal ions. 3 Applications The new method of calibration was applied to the analysis of two samples presenting different matrices: the certiÆed Table 2 Comparison of ablation yield of standard solutions with different chromophores Chromophore 2-Thiobarbituric acid 1,10-Phenanthroline Solvent H2O H2O Ar Øow effect Total mass loss Ar Øow effect Total mass loss Average of mass loss/mg 0.97 1.14 1 1.64 RSD (%) n~10 9.32 4.8 0 9.56 Mass loss due to ablation/mg 0.17 0.64 Fig. 5 Interaction of (a) 2-thiobarbituric acid and (b) 1,10-phenanthroline with water.Fig. 6 Variation of the ablation yield with increasing chromophore concentration. Standard solutions (different chromophore concentrations and 5 ppm Co) ablated at the energy optimised for LDPE, and at 10 Hz. J. Anal. At. Spectrom., 1999, 14, 1665±1672 1669reference material NIST SRM 613, and low density polyethylene (LDPE).17 3.1 Experimental Instruments. The laser ablation analyses were carried out using the excimer laser operating at 248 nm. The analyses of the digested samples were performed with the VG PQ I prototype ICP-MS instrument and with a Varian 300 (Varian Ltd., Walton on Thames, UK) atomic absorption spectrophotometer with graphite furnace Model GTA96. Reagents.Reagents used for the LDPE sample digestion were a borate Øux (Johnson Matthey, Royston, Hertfordshire, UK; 120A: 80% lithium tetraborate, 20% lithium Øuoride), Aristar nitric acid (BDH), tartaric acid (Fisher ScientiÆc). Standard solutions were prepared from 1000 ppm stock reagents (Grade Specpure, Johnson Matthey, and Spectrosol, BDH).Dilution was made with 2% nitric acid solution and 18 MV de-ionised water. The additive used for the modiÆcation of the standard solution was m-hydroxybenzoic acid (Aldrich). The polymeric material was provided by the CSL Food Science Laboratory, Norwich, UK.16 Digestion procedure. As the polymeric material was not a certiÆed reference sample, it was digested to be analysed by ICP-MS and furnace AA. The results were compared to the ones found by LA-ICP-MS and from previous work.17 The digestion procedure used a lithium borate mixture to fuse inorganic samples that were not directly soluble in acidic solution.The sample (150 mg) was previously ashed in a platinum crucible with lid, above a gas burner Øame. The ashes were fused in the molten Øux at 950 �C. The fused solid, when cool, was dissolved in an acidic solution (4% nitric acid±0.5% tartaric acid) and made up to 50 ml to give an overall dilution of 333 : 1.Two replicates were made this way. Standard solutions at 30, 50, 100 and 200 ng g21 of the analytes were made up in the same lithium borate matrix for calibration of the ICP-MS instrument. 9Be, 89Y and 209Bi were used as internal standards. Preparation of trial calibration curves for LA-ICP-MS. Calibration curves were obtained by LA-ICP-MS for each element of interest in the LDPE sample. Their concentration range covered the analyte concentrations present in the sample.ModiÆed standard solutions were prepared with 0, 5, 10, 25, 50, 100 mg g21 of the following elements: Mg, Cr, Zn, Sb and Pb. All standards contained 50 mg g21 of Co, which was used as an internal standard. As the concentration range studied here is relatively high, the amount of chromophore added to the standard solutions had to be small enough in order to obtain analyte signals that would not saturate the detector. For this reason, a concentration of 0.15 g l21 of m-hydroxybenzoic acid was added to the solutions.Standard solutions used for quantitative analysis by LA-ICPMS. A blank solution and a standard solution (7.5 mg g21 of analytes), both containing 0.3 g l21 of m-hydroxybenzoic acid, were prepared. They were both ablated, as well as the LDPE sample and a LDPE blank, known to be free of the analytes of interest. This LDPE blank alwed correction for isobaric interferences due to carbon compounds. The instrument parameters for the calibration and the LDPE analyses by LA-ICP-MS are given in Table 3. 3.2 Results The aqueous calibration solutions were freshly made before use and were stable during a working day. When left for longer periods some precipitation occurred notably with elements such as Pb and Pd. Calibration curves by LA-ICP-MS. The analytes contained in the standard solutions are homogeneously distributed; therefore, unlike solid standards, the modiÆed standard solutions do not require ablation at several sites in order to give a representative response.Each standard solution was continually ablated for 5 min. The acquisition of the ablation signal was preceded and followed by 1 min of non-ablation. This non-ablation blank was subtracted from the ablation signal, before any calculations. The blank solution was treated as a point on the calibration curves. The 5 min ablation acquisition was divided into Æve contiguous portions of 1 min each. The analyte responses were normalised to the cobalt internal standard.The standard deviation was calculated for each point. A weighted regression line was calculated from the calibration data of each element. Errors in the slope and intercept of the regression lines are given in Table 4. All curves have a correlation coefÆcient greater than 0.993. Fig. 7 shows the calibration curve obtained for Mg, which gave the poorest correlation. LDPE analysis. A correction factor (CF) was calculated which allowed correction for differences in sensitivity betwen the standard and the sample. 59Co was the isotope used as internal standard. CF~âÖcpsCo sampleÜ|Öconc:Co stdÜä=âÖcpsCo stdÜ|Öconc:Co sampleÜä (4) Table 3 Laser and ICP-MS parameters Excimer laser parameters– Laser wavelength 248 nm Laser energy 60 mJ Frequency 10 Hz Crater diameter 200 mm Plasma conditions and data acquisition parameters– Rf power 1.25 kW ReØected power v0.5W Argon Øow rate: Coolant 12 l min21 Auxiliary 1 l min21 Carrier 1.1 l min21 Point per peak 1 Dwell time 10 ms Acquisition mode Time resolved analysis (peak jumping) Fig. 7 Calibration curve with the poorest correlation coefÆcient, obtained for 25Mg, from ablation of modiÆed standard solutions. The error bars indicate °3 standard deviations. 1670 J. Anal. At. Spectrom., 1999, 14, 1665±1672The analytes concentrations were then calculated by: conc:elt~âÖcpselt sampleÜ|Öconc:elt stdÜä=âCFxÖcpselt stdÜä Ö5Ü where elt~element, std~standard, CPS~counts per second. The analyte concentrations found by the three different analytical techniques are shown in Table 5a.Data found from previous work are shown in Table 5b. The results obtained by LA-ICP-MS are in good agreement with those found by digestion-ICP-MS, furnace AA and previous work.17 For the LA-ICP-MS analysis, n~6 represents the number of separate sites (200 mm diameter) ablated on the same sample to obtain a representative result for the bulk of the material. The high RSD for Sb indicates its heterogeneity in the LDPE matrix.These data illustrate the difÆculty of making polymer standards18 of known composition and even homogeneity. Previous to this study, a certiÆed NIST SRMglass sample was analysed using the modiÆed standard solutions10 by LAICP- MS. The additive used was poly(sodium 4-styrene sulfonate); the laser used was an excimer laser emitting at 248 nm. The results found for seven isotopes are presented in Table 6.These results demonstrate that modiÆed aqueous standard solutions are suitable for quantitative analysis of solid LDPE and glass material by LA-ICP-MS. 4 Conclusions The data presented have demonstrated that aqueous standards having a modiÆed absorption coefÆcient can be employed in the analysis of solid material by LA-ICP-MS. Samples with different types of matrix, such as glass and low density polyethylene, were successfully analysed. The mechanism of ablation appears to be different from that applying to the solid samples; nevertheless, the use of an internal standard enables practical analysis to be performed, as long as the analytes in the liquid standard and in the solid sample behave equally regarding the internal standard.One of the great advantages of the standard solutions is that simple variation of the additive concentration allows the standards to be ablated at an energy that is optimal for the sample. This characteristic makes their use applicable to a wide variety of sample types. Other practical beneÆts derived from the use of aqueous standard solutions are that they are easier to prepare than solid standards, their elemental composition is readily variable, the distribution of the elements within the standard is homogenous, they offer a renewable surface during the ablation event and they are inexpensive.A future publication will report the extension of this calibration technique to analysis at 266 nm and to other sample types, including stainless steel.References 1 P. J. Fordham, J. W. Gramshaw, L. Castle, H. M. Crews, D. Thompson, S. J. Parry and E. McCurdy, J. Anal. At. Spectrom., 1995, 10, 303. Table 4 Calibration curve statistics for various elements in modiÆed absorption coefÆcient aqueous standards Isotope Linear curve equation Corr. coeff. (R2) Error on slope Error on intercept 25Mg y~158xz50 0.9937 °12 °603 53Cr y~113x270 0.9988 °4 °184 64Zn y~87xz148 0.9993 °2 °110 121Sb y~112xz88 0.9991 °3 °152 208Pb y~383xz900 0.9987 °20 °941 Table 6 Results for analysis of NIST glass Isotope Calculated conc./mg g21 °s NIST certiÆed conc./mg g21 140Ce 36.3 2.6 39a 65Cu 36.5 1.4 37.7a 57Fe 44.7 6.3 51 208Pb Int.std. – 38.57 60Ni 35.2 3.1 38.8 88Sr 77.6 2.0 78.4 238U 35.8 0.4 37.38 aNot certiÆed, given for information only. Table 5a Results for analysis of LDPE DigestionzICP-MS/mg g21 RSD (%) DigestionzAA/mg g21 RSD (%) LA-ICP-MS/mg g21 n~6 RSD (%) 25Mg 57.6 5.6 52.3 6.8 55.5 7.2 53Cr 30.3 8.4 32.1 2.5 34.3 1.6 59Co 28.5 8.2 32.1 2.1 Internal std. – 64Zn 50.8 0.9 48.3 2.3 51.3 17.6 121Sb 62.3 4.8 64.8 1.9 62.6 61.2 208Pb 50.7 1.1 57.7 4.5 52.5 4.4 Table 5b Results for analysis of LDPE from another source17 Nominal concentrations/mg g21 MDa ICP-MS/mg g21 NAA/mg g21 Mg 51.6 61.6 51.7 Cr 51.9 40.8 26.4 Co 49.2 34.6 29.0 Zn 49.3 48.8 39.2 Sb 69.8 74.8 52.0 Pb 62.6 68.3 N/Ab aMD~ microwave digestion. bN/A~not available. J. Anal. At. Spectrom., 1999, 14, 1665±1672 16712 J. T. Westheide, J. S. Becker, R. Jager, H. J. Dietze and J. A. C. Broekaert, J. Anal. At. Spectrom., 1996, 11, 661. 3 A. Raith, R. C. Hutton, I. D. Abell and J. Crighton, J. Anal. At. Spectrom., 1995, 10, 591. 4 S. Chenery and J. M. Cook, J. Anal. At. Spectrom., 1993, 8, 299. 5 A. Moissette, T. J. Shepherd and S. R. Chenery, J. Anal. At. Spectrom., 1996, 11, 177. 6 A. Audetat, D. Gunther and C. A. Heinrich, Science, 1998, 279, 2091. 7 E. Hoffmann, C. Ludke, H. Scholze and H. Stephanowitz, Fresenius' J. Anal. Chem., 1994, 350, 253. 8 B. Masters, Ph.D. Thesis, Loughborough University, Loughborough, 1996. 9 J. Watling, European Winter conference on Plasma Spectrochemistry, 1999, Pau, France. 10 B. Masters and B. L. Sharp, Anal. Commun., 1997, 34, 237. 11 P. Arrowsmith and S. K. Hughes, Appl. Spectrosc., 1988, 42, 1231. 12 R. Srinivasan and B. Braren, Appl. Phys. A, 1988, 45, 289. 13 E. Sutcliffe and R. Srinivasan, J. Appl. Phys., 1986, 60, 3315. 14 M. A. J. Rodgers and P. T. Snowden, J. Am. Chem. Soc., 1982, 104, 5541. 15 T. J. Chuang, H. Hiraoka and A. Modl, Appl. Phys. A, 1988, 45, 277. 16 R. Frischknecht, T. Ulrich, C. A. Heinrich and D. Gunther, European Winter conference on Plasma Spectrochemistry, 1999, Pau, France. 17 P. J. Fordham, W. Gramshaw, L. Castle, H. M. Crews, D. Thompson, S. J. Parry and E. McCurdy, J. Anal. At. Spectrom., 1995, 10, 303. 18 A. M. Dobney, H. Klinkenberg, C. de Koster and A. Mank, European Winter conference on Plasma Spectrochemistry, 1999, Pau, France. Paper 9/905479k 1672 J. Anal. At. Spectrom., 1999, 14, 1665±1672