Novel Laser Sampling Technique for Inductively Coupled Plasma Atomic Emission Spectrometry KENNETH K. K. LAM AND W. T. CHAN* Department of Chemistry, T he University of Hong Kong, Pokfulam Road, Hong Kong A novel laser sampling technique, back-surface ablation, has through the substrate. As the sample at the sample–substrate interface is vaporized by the laser beam, enormous pressure is been developed for ICP-AES. Samples are coated onto a transparent substrate and a laser beam is irradiated onto the developed between the sample and the substrate at the laser spot.21 The sample at the laser spot is removed explosively by sample through the substrate instead of sampling directly from the sample surface.As part of the sample is vaporized by the a single laser pulse. A clean plug of sample is easily removed with moderate laser power density (#107 Wcm-2); therefore, laser beam, a high pressure develops at the sample–substrate interface. The sample at the laser spot is explosively and preferential vaporization is minimized.Furthermore, sampling efficiency is enhanced. The amount of material removed is up completely removed by the expanding vapour. Sampling efficiency is up to ten times higher than conventional ‘front- to 100 times larger than direct ablation of a sample (frontsurface ablation).21 The actual enhancement of laser sampling surface’ laser sampling. Also, preferential vaporization is minimized because of complete removal of the sample at the efficiency depends on the thickness of the sample film.A 10–20-fold enhancement was obtained in this work. Sensitivity laser spot. The risk of inaccurate chemical analysis associated with non-stoichiometric thermal vaporization in front surface enhancement (in terms of ICP emission intensity), however, is only about 3-fold, probably because of poor transport efficiency laser sampling is reduced. Two methods of calibration, viz., standard additions and calibration with standards in a of the laser-sampled material.The laser-sampled materials travel at high speed on taking off from the substrate, collide poly(vinyl alcohol ) matrix, were used for quantitative elemental analysis of household paints using front- and back- with the ablation chamber and break into small particles. Some of these particles are probably still too large for efficient surface ablation–ICP-AES. Internal standards were used to compensate for pulse-to-pulse laser energy fluctuation and transport to the ICP by the carrier gas.22 Household emulsion paints were analysed using both front- sample thickness variation across a sample.Ten elements with different thermal properties and at concentrations ranging and back-surface ablation to compare the sensitivity, precision, and laser-sampled material stoichiometry of these sampling from 10 to 1000 ppm were determined and the elemental concentrations were compared with those of microwave techniques.Quantitative elemental analysis using standard additions and calibration with standards in a poly(vinyl digestion/solution nebulization–ICP-AES. Back-surface ablation appears to be more accurate than conventional front- alcohol) (PVA) matrix is demonstrated. The results were compared with those of microwave acid digestion. surface ablation. Keywords: L aser sampling; back-surface laser ablation; inductively coupled plasma; atomic emission spectrometry EXPERIMENTAL Instrumentation Laser ablation is a versatile sampling technique for analytical A schematic diagram of the laser ablation set-up is shown excitation sources.It can be applied directly to a wide range in Fig. 1. The ICP spectrometer [Carl Zeiss (Jena, Germany) of sample types with little or no sample preparation.1–4 Sample Plasmaquant 110] uses a 1 m high-resolution echelle preparation, especially dissolution of solids, is time consuming and often a source of contamination and analyte loss.Laser sampling has been coupled with ICP-AES and ICP-MS as a tandem technique that allows independent optimization of the sampling and excitation processes.5–13 ICP is a mature high power density atomic source that can vaporize and atomize the laser-sampled materials efficiently. During laser sampling–ICP operation, the ICP parameters usually do not need much adjustment. However, laser–material interactions at the sample surface are complicated, the amount and stoichiometry of the laser-sampled materials vary with laser power density and pulse energy, gas atmosphere in the ablation chamber, as well as the properties of the sample.14–19 Preferential vaporization and laser-induced plasma shielding of the laser beam20 are especially troublesome for chemical analysis.Routine analyses using laser sampling have yet to be realized. A laser sampling technique without these problems is desirable. This paper describes a novel laser sampling technique, backsurface ablation, that makes use of the high pressure generated at the sample surface during laser sample vaporization to Fig. 1 Experimental set-upfor laser ablation–ICP-AES. During back- remove the sample at the laser spot. A film of a sample, a few surface ablation, the sample is placed at position 1 with the quartz tens of micrometres thick, is coated onto a transparent sub- substrate facing the laser beam. In conventional front-surface ablation, the sample is placed at position 2, facing the laser beam directly.strate. A laser beam is irradiated onto the sample surface Journal of Analytical Atomic Spectrometry, January 1997, Vol. 12 (7–12) 7spectrometer for wavelength selection. A fibre block with 132 optical fibres was positioned at the exit of the spectrometer; each fibre corresponds to a specific wavelength. Up to 12 emission lines can be monitored simultaneously. Optical fibres for the selected lines were connected to an array of 12 photomultiplier tubes (PMTs) for measurement.The operating parameters of the ICP-AES system and the selected spectral lines are given in Tables 1 and 2, respectively. Since the lasersampled materials take about 5 s to be completely swept into the ICP, ICP emission intensity was integrated for 6 s for each laser shot during quantitative analysis. A shorter integration time (0.03 s) was used for temporal study of the ICP emission. A KrF excimer laser [Lumonics 510 (Lumonics, Kanata, Fig. 2 Details of the glass laser ablation chamber. Canada) with stable optics, wavelength=248 nm, pulse duration= 12 ns FWHM] was used for laser ablation. Pulse ing is discussed below.) A sample-coated quartz plate was energy is approximately 60 mJ. A single plano-convex lens of attached to the front of the chamber with a plastic ring that focal length 200 mm at 248 nm was used to focus the laser fitted snugly to the laser ablation chamber so that the chamber beam onto the sample.The samples were always placed before was air-tight. The laser-sampled materials were carried to the the focal point to avoid laser breakdown of the atmosphere at ICP by an Ar stream via a Teflon tube (0.80 m×5 mm id). A the laser focus. Because of the stable cavity configuration, the three-way valve was placed between the chamber and the laser beam divergence is relatively large. A rectangular aperture torch. The bottom of the ICP torch was sealed from (30×10 mm) was placed between the excimer laser and the the atmosphere during sample changing.The Ar flow into the focusing lens to limit the laser beam size and select the central central tube of the ICP torch was resumed after the sample homogeneous portion of the laser beam. The laser energy after had been attached and the chamber was flushed with Ar the aperture is 20–30 mJ. Typical laser power density is in the for 1 min. range 107–108 Wcm-2, depending on the lens-to-sample dis- Conventional front-surface ablation using the same laser tance.Laser power density was determined from laser pulse and ICP operating parameters was performed for comparison. energy, pulse duration and measured spot size. Spot size was In this set-up, the chamber is rotated 180° so that the quartz also calculated from the lens-to-sample distance using geoplate at the end of the chamber now faces the laser beam and metric optics principles, which agreed reasonably well with the becomes the chamber window.A quartz plate coated with measured spot size. As all material at the laser spot is compaint sample was placed at the other end of the ablation pletely removed by a single laser shot during back-surface chamber, with the sample facing the laser. ablation, single laser pulse ablation was used. The layout of the laser ablation chamber for back-surface ablation is shown in Fig. 2. The glass ablation chamber is Preparation of Samples and Standards 30 mm in diameter and 30 mm long.One end of the chamber was sealed with a quartz plate. The chamber was mounted on Local latex emulsion paints (The China Paint MFG. Co., Hong Kong) of different colours were used. The water-based a xyz-translation stage to facilitate raster laser sampling of the sample and fine adjustment of the laser beam focus. Quartz paints can be coated on a substrate readily and reproducibly. Furthermore, aqueous standard solutions mix readily with the plates, 30 mm in diameter and 1 mm thick, were used as transparent substrates to hold the sample films.(Sample coat- samples for quantitative analysis using the standard additions method. The wet paint was diluted with an equal volume of distilled water to reduce its viscosity. A thin film of the diluted Table 1 ICP operating parameters paint was then coated onto a quartz plate by spreading a few ICP forward power 1.0 kW drops of the paint evenly on the quartz plate surface with a Observation height 10 mm above load coil glass rod.The paint was air-dried at room temperature. Coolant argon flow rate 12 l min-1 Different sample thicknesses were obtained by varying the Auxiliary argon flow rate 1.0 l min-1 number of drops of sample added to the substrate. A sample Carrier gas argon flow rate 1.0 l min-1 thickness of 35 mm was used during quantitative analysis. The Integration time 0.03–1.0 s thickness of the film, and thus the mass per unit area of the paint sample, varies slightly across the surface and from sample Table 2 ICP emission spectral lines used to sample.Variation in ICP emission intensity due to the variation in the amount of laser-ablated material was compen- Element Wavelength/nm sated using the internal standard method. For quantitative Al I 396.152 analysis using the standard additions method, titanium was Ba II 455.403 used as an internal standard because of its abundance in the Ca II 317.933 paint samples (about 5–10% TiO2 as pigment).Co II 228.616 Two calibration methods were used during quantitative Cr II 205.559 Cu I 324.754 elemental analysis: standard additions to the paint samples Fe II 238.207 and standards in a PVA matrix. In the standard additions Mg I 285.213 method, aqueous standard solutions in 1% nitric acid were Mn II 257.610 added directly to the wet paint samples and the resulting Ni I 341.477 mixture was coated onto the quartz plate for ablation. Pb II 220.351 Standards in a PVA matrix were prepared by dissolving 4 g Si I 212.412 Sn I 235.484 of PVA powder [Aldrich (Milwaukee, WI, USA), 88% hydro- Sr II 407.771 lysed, average molecular mass 80000–100 000] in 100 ml of Ti II 368.520 water.After dissolution, 4 g of titanium dioxide powder were V II 292.402 added and the mixture was stirred for 1 h with a magnetic Zn II 202.551 stirrer. Titanium dioxide absorbs strongly at the UV laser 8 Journal of Analytical Atomic Spectrometry, January 1997, Vol. 12wavelength. It is added to enhance laser energy absorption carrier gas flow rate.17 Since the operating parameters are the same except for the laser sampling method, ICP emission for and thus the ablation efficiency. In addition, PVA films become more brittle with titanium dioxide and the laser-sampled both back- and front-surface ablation has similar rise and fall times. materials from back-surface ablation can be fragmented into fine particles more efficiently.For quantitative analysis using The amount of laser-sampled material, however, is larger for back-surface ablation. With a single laser pulse of moderate standards in a PVA matrix, titanium was not used as an internal standard as its concentration varies in different paint power density (#107 W cm-2), the sample at the laser spot is completely removed using back-surface ablation. A small frac- samples and is not known. Ni (200 ppm) was used instead because it is absent from the paint samples and does not give tion of the sample is ablated using front-surface ablation at the same laser power density.The amount of material removed rise to spectral interference with other elements. A series of standards in a PVA matrix was prepared by adding aqueous by back-surface ablation can be 100 times larger than by frontsurface ablation.21 In this work, a 35 mm thick sample required standards (in 1% nitric acid) to the PVA solution. The standards were then coated onto quartz plates in a similar manner about ten laser pulses for complete removal of the sample at the laser spot using front-surface ablation (Fig. 4). Therefore, to the paint sample preparation. a 10-fold improvement in sampling efficiency is obtained using back-surface ablation at this sample thickness. Microwave Digestion It appears that enhancement of sampling efficiency using back-surface ablation can be further improved by increasing Quantitative elemental analysis of the paints using laser samthe sample thickness.However, there is an optimum sample pling–ICP-AES was compared with that of microwave digesthickness for maximum ICP emission intensity (Fig. 5). ICP tion and solution nebulization into the ICP. A modification of the method of Paudyn and Smith23 was used to digest the paints. Since the large amount of organic matter in the sample results in excessive pressure in the bomb and may cause damage to both the oven and the bomb, the sample was ashed before microwave digestion.A portion (3–4 g) of wet paint was weighed in an ashless filter-paper (Whatman No. 540) and ashed in a crucible with a Bunsen burner. About 1 g of ash remained after ashing. A portion (20–30 mg) of the ash was weighed and placed in a Parr Microwave Digestion Bomb [Parr Instrument (Moline, IL, USA) Model 4782], followed by 5 ml of concentrated HNO3 [analytical-reagent grade, Merck (Darmstadt, Germany)] and 2 ml of concentrated HF (analytical-reagent grade, Merck).The bomb was placed in the oven and irradiated for 2 min at about 600 W. After cooling Fig. 4 ICP emission intensity of Fe II 238.2 nm versus laser pulse for about 2 h, 20 ml of water were added to the vessel and the number using front-surface ablation. Laser power density, approxi- solution was transferred into a Teflon beaker and evaporated mately 5×107 W cm-2. to 1 ml on a hot-plate at 120°C. The resulting solution was diluted and filtered and made up to 100 ml with 1% nitric acid for spectrochemical analysis.The method of standard additions was used for quantitative elemental analysis of the digested samples. RESULTS AND DISCUSSION Sensitivity and Sample Thickness Typical ICP emission intensity for back- and front-surface laser ablation of a paint sample is shown in Fig. 3. The emission is transient for single pulse ablation. The peak shape and peak width are related to the volume of the ablation chamber, the length and diameter of the transfer tube, and the Fig. 5 ICP emission intensities of (a) Fe II 238.2 nm and (b) Mg I Fig. 3 Temporal ICP emission intensity of Fe II 238.2 nm for back- 285.2 nm versus paint sample thickness using back-surface ablation. Laser power density, approximately 5×107 Wcm-2. and front-surface ablation of a paint sample. Integration time, 0.3 s. Journal of Analytical Atomic Spectrometry, January 1997, Vol. 12 9Table 4 Oxide melting-points of Al, Ca, Cu, Fe, Mg, Mn and Sr25 emission intensity first increases and then reduces with sample thickness.Samples of all thicknesses used in Fig. 5 were Oxide melting-point/ completely removed by a single laser pulse during back-surface Element Oxide °C ablation. The initial increase in ICP intensity is probably Cu CuO 1326 related to the amount of sample ablated. Since back-surface Fe Fe2O3 1565 ablation removes all sample material at the laser spot, a thicker Mn Mn3O4 1564 sample means a larger sample mass and thus an increase in Al Al2O3 2072 ICP intensity.However, as the sample thickness increases, the Sr SrO 2430 Ca CaO 2614 efficiency of sample fragmentation into fine particles reduces. Mg MgO 2852 Larger amounts of paint fragments were found at the bottom of the laser sampling chamber as the sample thickness increased. Thin samples are probably more efficiently fragmented into fine particles than thick samples. Since large particles are not transported to the ICP efficiently,22 ICP strated.Stoichiometric sampling is possible only if critical or emission intensity decreases even when the mass ablated supercritical points are attained.26 However, if melting occurs increases with sample thickness. A sample thickness of 35 mm at the laser spot, the vapour pressure of low-melting oxides was used throughout this work. should be higher because of a larger difference between the melting-point and the molten oxide temperature.Differential vaporization occurs during front-surface ablation, leading to Preferential Vaporization enrichment of lower melting/boiling copper oxide (mp 1326 °C) With the optimum sample thickness of 35 mm, the ICP emission and depletion of high melting/boiling magnesium oxide (mp ratios of back-surface ablation to front-surface ablation range 2852°C) in the gas phase. The ICP intensity ratio therefore from 1 to 6 (Table 3), i.e., there is an enhancement of ICP reduces as the oxide melting-point increases (Fig. 6). In con- emission of up to 6-fold. The enhancement is smaller than that trast, preferential vaporization is minimal during back-surface of sampling efficiency (#10-fold, Fig. 4), probably because of laser ablation as all sample materials at the laser spot are the transport efficiency of the particles. However, the enhance- removed by a single laser shot. ment of ICP emission is also element-dependent, which may be related to preferential (thermal) vaporization of volatile elements during laser sampling.Preferential vaporization Precision during laser sampling leads to enrichment of volatile elements The typical RSD for single-pulse laser sampling reported in and depletion of refractory elements in the gas phase.14–19 the literature is 30–70%,1–4 which is mainly due to sample The extent of preferential vaporization during conventional heterogeneity and pulse-to-pulse variation in laser energy. front-surface laser sampling can be shown by the ICP-AES Laser power density is a major factor influencing the amount intensity ratios of front-surface ablation to back-surface of laser-sampled materials.27,28 In this work, the RSD for ablation for elements with a large difference in the melting- single-pulse front-surface ablation is 10–23%; it is reduced to point of their oxides (Fig. 6). Oxides are considered here 2–14% when Ti is used as an internal standard (Table 5). because the elements do not exist as metals in paint but mainly The RSD of the ICP-AES intensity for back-surface ablation as oxides.24 Oxides with melting-points25 ranging from 1326 is 9–16% (Table 5).The fluctuation represents the variation to 2852°C (Table 4) were studied. With such a wide range of in sample film thickness and laser pulse energy. Consistent melting-points, the effects of thermal vaporization of the molten and uniform thickness of the film from sample to sample is oxides at the laser spot during laser sampling can be demon- difficult to achieve using our sample-coating method.However, the variation in sample thickness can be compensated using Table 3 Enhancement of ICP emission intensity using back-surface an internal standard. The RSD of the ICP intensity is 2–10% ablation versus front-surface ablation with Ti as the internal standard (Table 5). Using an internal standard, the thickness of the film need not be strictly con- Intensity ratio, trolled and the preparation of sample films becomes simple Element back-surface ablation: front-surface ablation and straightforward. Cu I 1.0 Fe II 1.6 Si I 2.0 Al I 1.6 Sr II 2.8 Table 5 Precision (RSD %) of back- and front-surface ablation Ca II 5.7 Mg I 4.7 RSD (%) Back-surface Front-surface Back-surface Front-surface ablation ablation ablation ablation (without (without (with (with internal internal internal internal Element standard) standard) standard) standard) Ba 12 10 4 5 Co 13 14 3 3 Cr 11 11 4 5 Cu 13 11 5 3 Fe 11 14 2 3 Mg 9 23 8 5 Pb 15 16 9 10 Si 12 14 2 6 Sn 10 13 5 5 Sr 10 19 3 2 Fig. 6 ICP intensity ratios of front-surface to back-surface ablation Zn 16 21 10 12 versus oxide melting-points. 10 Journal of Analytical Atomic Spectrometry, January 1997, Vol. 12Fig. 7 Typical standard additions calibration graphs for (a) backsurface ablation and (b) front-surface ablation of a blue paint sample. Ti was used as an internal standard. Semi-quantitative Analysis of Paint Samples Using Standard Additions Method Fig. 8 Correlation of the elemental concentrations of Ba, Co, Cr, Cu, Paint samples were analysed using both back-surface and Fe, Mg, Mn, Pb, Sn, Sr, V and Zn in blue and red paints determined by the standard additions method using (a) back-surface ablation and front-surface laser sampling–ICP-AES. For comparison, the (b) front-surface ablation with those of the microwave digestion/ same samples were analysed using ICP-AES with microwave solution nebulization method.digestion/solution nebulization. Standard additions was used to compensate for matrix effects for all analyses. Ti was used as the internal standard to compensate for laser pulse energy dards are needed to ensure accurate analysis.29 However, fluctuation and sample thickness variation. Typical calibration matrix-matched solid standards are usually not available and graphs for laser sampling–ICP-AES are shown in Fig. 7. preparation of the standards can be tedious and time consum- Elemental concentrations of the paint samples determined ing.The standard additions method is applicable only for by back-surface and front-surface ablation–ICP-AES are com- certain samples and is also tedious. A laser sampling technique pared with those of microwave digestion/solution nebulization that does not require matrix-matched standards for calibration in Fig. 8. Good correlation between back-surface ablation and would, overall, simplify the analysis. microwave digestion at concentrations above 20 ppm for up Since sample matrix effects such as preferential vaporization to two orders of magnitude is observed [Fig. 8(a)]. The large are not significant in back-surface ablation, the use of non- deviation for elements of low concentration (<10 ppm) is matrix-matched standards for calibration seems feasible. probably due to the small amount of the elements present in Standards in a PVA matrix were prepared for the analysis of the sample.The mass of the paint material removed per laser paints using both front- and back-surface laser sampling. TiO2 pulse using back-surface ablation can be estimated from the was added to the PVA solution to enhance UV laser pulse laser spot size and the amount of sample added to the substrate. absorption by the PVA film and thus the laser sampling About 5 mg are removed by each laser pulse. If the elemental efficiency. Ni was used as an internal standard to compensate concentration is 10 ppm, the mass of metal sampled is #50 ng.for any film thickness variation. The sampling error for such a small amount of material may Calibration graphs using standards in a PVA matrix are be a major limiting factor for quantitative analysis of trace shown in Fig. 9. Again, the data of laser ablation using PVA elements. calibration were compared with those obtained by a microwave The difference in elemental concentration determined by digestion method (Fig. 10). There is a good correlation between front-surface ablation and microwave digestion is larger back-surface ablation and microwave digestion for major [Fig. 8(b)], probably due to preferential vaporization during elements [Fig. 10(a)]. On the other hand, the values obtained laser sampling. There is also a larger scattering of the data by front-surface ablation using PVA calibration and microwave points because of the smaller emission intensity and thus digestion do not agree so well [Fig. 10(b)]. Matrix-matched smaller S/N ratios for front-surface laser sampling. standards are needed for front-surface ablation calibration. Semi-quantitative Analysis of Paint Samples Using Standards in CONCLUSIONS PVA Matrix Back-surface ablation is potentially an effective laser sampling In laser sampling analysis, matrix effects strongly influence the amount of the laser-sampled material. Matrix-matched stan- method for coating materials, polymers and biological samples. Journal of Analytical Atomic Spectrometry, January 1997, Vol. 12 11The sensitivity and accuracy are improved compared with the conventional front-surface laser ablation method. Preferential vaporization is minimized as the sample at the laser spot is completely removed. Semi-quantitative analyses using the standard additions method and calibration using non-matrixmatched standards in PVA have been demonstrated. Elements of concentration from 1000 to 10 ppm were determined and the data correlate well with those of the microwave digestion/ solution nebulization method.Successful calibration using standards in a PVA matrix suggests that universal standards for a wide variety of sample matrices may be possible when back-surface ablation is used. Financial support from the Department of Chemistry and CRGC research grants of the University of Hong Kong is gratefully acknowledged. REFERENCES 1 Dittrich, K., and Wennrich, R., Prog.Anal. At. Spectrosc., 1984, 7, 139. 2 Darke, S. A., and Tyson, J. F., J. Anal. At. Spectrom., 1993, 8, 145. 3 Monenke-Blankenburg, L., Spectrochim. Acta Rev., 1993, 15, 1. 4 Russo, R. E., Appl. Spectrosc., 1995, 49, 14A. Fig. 9 Typical calibration graphs of standards in a PVA matrix (1% 5 Thompson, M., Goulter, J. E., and Sieper, F., Analyst, 1981, PVA+1% TiO2+standards) using (a) back-surface ablation and 106, 32. (b) front-surface ablation. Ni was used as an internal standard. 6 Carr, J. W., and Horlick, G., Spectrochim. Acta, Part B, 1982, 37, 1. 7 Ishizuka, T., and Uwamino, Y., Spectrochim. Acta, Part B, 1983, 38, 519. 8 Gray, A. L., Analyst, 1985, 110, 551. 9 Richner, P., Borer, M. W., Brushwyler, K. R., and Hieftje, G. M., Appl. Spectrosc., 1990, 44, 1290. 10 Pang, H. P., Wiederin, D. R., Houk, R. S., and Yeung, E. S., Anal. Chem., 1991, 63, 390. 11 Denoyer, E. R., Fredeen, K. J., and Hager, J. W., Anal. Chem., 1991, 63, 445A. 12 D’Silva, A. P., Baldwin, D.P., and Zamzow, D. S., Anal. Chem., 1994, 66, 1911. 13 Russo, R. E., Mao, X. L., Chan, W. T., Bryant, M. F., and Kinard, W. F., J. Anal. At. Spectrom., 1995, 10, 295. 14 Arrowsmith, P., Anal. Chem., 1987, 59, 1437. 15 Hager, J., Anal. Chem., 1989, 61, 1243. 16 Thompson, M., Chenery, S., and Brett, L., J. Anal. At. Spectrom., 1989, 4, 11. 17 Chan, W. T., and Russo, R. E., Spectrochim. Acta, Part B, 1991, 46, 1471. 18 Chan, W. T., and Russo, R. E., Appl. Spectrosc., 1992, 46, 1025. 19 Cromwell, E. F., and Arrowsmith, P., Anal. Chem., 1995, 67, 131. 20 Mao, X. L., Chan, W. T., Shannon, M. A., and Russo, R. E., J. Appl. Phys., 1993, 74, 4915. 21 Fletcher, T. R., J. Appl. Phys., 1993, 73, 5292. 22 Arrowsmith, P., and Hughes, S. K., Appl. Spectrosc., 1988, 42, 1231. 23 Paudyn, A. M., and Smith, R. G., Fresenius’ J. Anal. Chem., 1993, 345, 695. 24 Parsons, P., Surface Coatings, Chapman and Hall, London, 1993, vol. 1. 25 CRC Handbook of Chemistry and Physics, eds. Weast, C., Astle, M. J., and Beyer, W. H., CRC Press, Boca Raton, FL, 64th edn., 1984. 26 Ready, J. F., Effects of High-Power L aser Radiation, Academic Press, New York, 1971. 27 Imai, N., Anal. Chim. Acta, 1990, 235, 381. 28 McLeod, C. W., and Booth P. K., Mikrochim. Acta, 1990, 3, 283. 29 Darke, S. A., Long, S. E., Pickford, C. J., and Tyson, J. F., J. Anal. At. Spectrom., 1989, 4, 715. Paper 6/02822E Fig. 10 Correlation of the elemental concentrations of Ba, Co, Cr, Received April 23, 1996 Cu, Fe, Mg, Mn, Pb, Sn, Sr, V and Zn in blue and red paints Accepted August 27, 1996 determined by PVA calibration using (a) back-surface ablation and (b) front-surface ablation with those of the microwave digestion/ solution nebulization method. Standards in a PVA matrix were used for calibration. 12 Journal of Analytical Atomic Spectrometry, January 1997, Vol. 12