Thermal Stabilisation of Phosphorus During Electrothermal Atomic Absorption Spectrometry Using Sodium Fluoride as Chemical Modifier EDWIN HERNA� NDEZa, JOSE� ALVARADO*a , FREDDY ARENASb AND MARIANO VE� LEZc aUniversidad Simo�n Bolý�var, Departamento de Quý�mica, Apartado Postal 89.000, Caracas 1080-A, Venezuela bInstituto Universitario de T ecnologý�a ‘Dr. Federico Rivero Palacio’, Departamento deMateriales, Apartado Postal 40347, Caracas 1040-A, Venezuela cUniversidad Simo�n Bolý�var, Departamento deMateriales, Apartado 89.000, Caracas 1080-A, Venezuela The stabilisation mechanism of phosphorus during its was found to provide the maximum stabilisation eVect.8 However, a drawback to its use is the markedly low sensitivity determination by electrothermal atomic absorption spectrometry in the presence of sodium fluoride as a chemical of the measurements, approximately 38% less than that achieved using Pd as the modifier.modifier has been studied. The study includes atomisation from dual-cavity platforms and analysis by scanning electron The aim of this work was to gain information on the mechanism by which NaF stabilises phosphorus, and also on microscopy, energy dispersive X-ray spectrometry and X-ray photoelectron spectrometry of the material accumulated onto the cause of the reported low sensitivity.To this purpose, atomisation from dual-cavity platforms, scanning electron totally pyrolytic graphite platforms. The results obtained from this study point toward a stabilisation mechanism which is of a microscopy (SEM), energy dispersive X-ray spectrometry (EDS) and X-ray photoelectron spectrometry (XPS) were used physical nature, whereby the analyte is retained in the matrix of the modifier until temperatures around 1350 °C are reached.to study the morphology as well as the chemical composition of the material formed at diVerent stages of an ETAAS heating The eVect of the modifier on the sensitivity of measurements of diVerent phosphorus compounds was also studied.A possible programme for phosphorus in the presence of NaF. In the field of ETAAS, SEM has been mainly used to study the explanation of the comparatively low sensitivity of phosphorus measurements with this particular modifier, approximately morphology of the atomisers which in turn can be related to the behaviour of the analyte through the shape of the analytical 38% less as compared with palladium, is given.signal.9–11 Only a few authors have reported SEM analysis as Keywords: Electrothermal atomic absorption spectrometry; a tool for the elucidation of stabilisation mechanisms of phosphorus determination; sodium fluoride; modification; analytes by chemical modifiers.7,12–15 The lack of information stabilisation mechanism dealing with the use of XPS16–19 for stabilisation mechanism interpretation is also notable. However, these techniques are used in this work to put forward arguments supporting a Electrothermal atomic absorption spectrometry (ETAAS) is possible mechanism for stabilisation of phosphorus using NaF not the most sensitive technique for phosphorus determination, as a modifier in ETAAS. because of both the design of modern spectrometers and the nature of this particular element.In fact, since the resonance lines of phosphorus lie in the UV–vacuum (l<190 nm), where absorption from the light source by oxygen in the air occurs, EXPERIMENTAL it is necessary to resort to a non-resonant doublet at Instrumentation 213.5–213.6 nm, with the corresponding reduction in sensitivity, for the ETAAS determination of this element.1 A Perkin-Elmer (U� berlingen, Germany) Model 2100 atomic absorption spectrometer was used together with a HGA 700 Additionally, phosphorus losses in the form of molecular species have been shown to occur at low pretreatment tempera- graphite atomiser system and an AS-70 auto sampler.Measurements were made at the non-resonant doublet at tures.2 Even though a few authors have reported successful phosphorus determinations in the absence of chemical modifi- 213.5–213.6 nm with a Perkin-Elmer phosphorus hollow cathode lamp.Dual-cavity platforms made of pyrolytic graphite cation,3,4 the use of a modifier is practically mandatory for the ETAAS determination of this element. and commercial pyrolytic, as well as laboratory-made standard, graphite platforms, inserted into pyrolytically coated graphite Several studies have been performed to find the best modifier for phosphorus determination. Among the compounds tested, tubes were used.Eppendorf (Westbury, NY, USA) micropipettes with plastic disposable tips were used to manually inject palladium nitrate, alone or mixed with calcium nitrate5 and lanthanum nitrate,6 provides high sensitivity and allows for the phosphorus solutions onto the dual cavity platforms. A Philips Analytical (Eindhoven, The Netherlands) Model high-temperature pretreatment.Unfortunately these compounds are relatively expensive, and lanthanum has been XL 30 scanning electron microscope was used to study the morphology of the material accumulated onto pyrolytic graph- shown to considerably reduce the useful lifetime of the atomisers and platforms.7 ite platforms. In the dark region below the photomicrographs the following information is displayed from left to right: Recently, the use of fluoride compounds was proposed as an alternative for phosphorus chemical modification.8 It was acceleration voltage (kV), magnification, detector (SE), working distance (WD) and scale (mm).EDS was performed found that some fluoride compounds grant phosphorus a thermal stability comparable to that provided by Pd and La, with an ultra-thin windows (UTW) detector attached to the XL 30 instrument, allowing element detection analysis from as well as oVering the advantages of being less expensive, presenting no memory eVects and causing less deterioration of boron onwards.A Leybold Heraeus (Export, PA, USA) Model LH-11 X-ray the atomiser. Among the fluoride compounds studied, NaF Journal of Analytical Atomic Spectrometry, December 1997, Vol. 12 (1391–1396) 1391fluorescence analyser was used for identification of the com- obtained after heating, the platforms served as sample supports so that disturbance of the residue after its collection was pounds formed at some pretreatment temperatures.avoided. The platform, with the material deposited on it, was placed directly into the corresponding spectrometer analysing Reagents and Standards chamber and the analysis was carried out. No Au or Au–Pt sputtered coatings, usually applied to non-conductive materials Ammonium dihydrogen phosphate (Merck, Elmsford, NY, for SEM or EDS studies, were used in these experiments in USA, analytical-reagent grade), diammonium hydrogen phosorder to avoid any chemical disturbance of the material phate (Merck, analytical-reagent grade), sodium hydrogen deposited on the platform.phosphate, potassium dihydrogen phosphate (Mallinckrodt, St. Louis, MO, USA, analytical-reagent grade), ammonium hexafluorophosphate (Aldrich, Milwaukee, WI, USA, analyt- RESULTS AND DISCUSSION ical-reagent grade) and sodium hexafluorophosphate (Aldrich, analytical-reagent grade) were used to prepare phosphorus Experiments with Dual-cavity Platforms solutions.Sodium fluoride (Mallinckrodt, analytical-reagent The use of dual-cavity platforms, which allows separate injecgrade) was used as a chemical modifier. Distilled, de-ionised tions of the modifier and the analyte, is a suitable way to study water (18MV cm) was used throughout for sample preparation some processes that take place within the graphite furnace and dilution purposes. Nitric acid (J. T. Baker, Phillipsburg, during heating. If care is taken not to mix the solutions, NJ, USA, ULTRATREX II Ultrapure reagent) was used for information about either gas or condensed phase reactions can stabilising the phosphorus solutions.be obtained.7, 20–25 Fig. 1 shows the charring plots for 5 ml of a 100 mg ml-1 phosphorus solution atomised from dual-cavity platforms, using 5 ml of a 0.08% m/v NaF solution injected Procedure either in the same cavity as the or in a Preparation of standards diVerent cavity. This amount of modifier was found to be optimal for stabilisation of the amount of phosphorus con- Phosphorus solutions (1000 mg l-1) were prepared by direct tained in the test solutions used here.Greater amounts did weighing of the corresponding phosphorus compound. Nitric not favour either the stabilisation of the analyte or the sensi- acid was added to give a final HNO3 concentration of 0.2% tivity of the measurements. Smaller amounts were not suYcient v/v for stabilisation purposes. A 0.08% m/v NaF solution was for stabilising all the phosphorus present in the test solutions.prepared by direct weighing and dissolution. No acid was It is clearly seen that the maximum stabilising eVect is added to this solution. achieved when both the analyte and the modifier are placed in the same spot of the dual-cavity platform. From this, it is Experiments with dual-cavity platforms inferred that stabilisation is more eYcient in the condensed phase. However, it is worth noticing that even though a poorer Dual-cavity pyrolytic graphite platforms were used to detersensitivity is obtained when phosphorus and the modifier are mine if stabilisation occurs in the condensed or in the vapour atomised from diVerent cavities, stabilisation is still present phase.Charring plots for phosphorus and NaF placed either under these conditions. Bearing in mind that modifiers in P in the same or in a diVerent cavity were recorded. Five determinations by ETAAS are needed mainly to inhibit forma- microlitres of a 1000 mg ml-1 phosphorus solution and 5 ml of tion of volatile phosphorus species, which are lost even at a 0.08% m/v solution of NaF were atomised under the heating comparatively low ashing temperatures, it is possible that mass programme given in Table 1.transport processes, promoted by heating, allow these volatile species to reach the cavity where the modifier was placed. Study of the material accumulated onto graphite platforms Formation of stable phosphorus species will then occur in the bulk of the modifier.The possibility that these reactions occur Solutions of phosphorus and the modifier were injected onto with both reactants being in the gas phase is ruled out by the pyrolytic graphite platforms and heated up to one of the high boiling point of NaF (1695 °C)26 and by the fact that, following charring temperatures: 200, 400, 800, 1000 and according to Fig. 1, stabilisation occurs at much lower tempera- 1350 °C.These temperatures were chosen to be within the tures. The lower eYciency of stabilisation when both reagents range of charring temperatures where phosphorus analytical are placed in diVerent cavities could then be due to the fact signals exhibit maximum sensitivity, i.e. within the plateau range of the charring curve for this element. The procedure was repeated until suYcient material had been accumulated, #3 mg, at each temperature. Each platform was stored in a desiccator to minimise water adsorption.The analysis of the material was performed as soon as possible after accumulation. For the analysis by SEM, EDS and XPS of the residue Table 1 Heating programme for the determination of phosphorus by ETAAS Temperature/ Ramp/ Hold/ Gas flow/ Step °C s s ml min-1 Dry 90 5 10*/30† 300 Dry 120 10 10*/30† 300 Charring Variable 5 20 300 Cool-down 50 5 10 300 Fig. 1 Charring plots for phosphorus atomised from dual-cavity Atomisation 2650 0 5 0 Cool-down 20 1 8 300 platforms.Five microlitres of a 100 mg ml-1 NH4H2PO4 solution and 5 ml of a 0.08% m/v NaF solution. A, Analyte without modifier; B, analyte and modifier in diVerent cavities; and C, analyte and modifier * Holding time for dual-cavity platform experiments. † Holding time for conventional platform experiments. in same cavity. 1392 Journal of Analytical Atomic Spectrometry, December 1997, Vol. 12that only the portion of analyte that can, as volatile species, tubes made of the same material, any diVerence in the analytical signals should be related exclusively to the diVerent platform reach the modifier’s cavity is stabilised; the rest would be lost surfaces. during heating.As can be seen from Fig. 2, the use of NaF as a modifier does not eliminate the dependence of phosphorus sensitivity Dependence of the Phosphorus Absorption Signal on the Surface on the material of the atomisation surface. This is also true of the Atomiser when using Pd and La, as reported by Curtius et al.27 In fact, when phosphorus is atomised from a standard graphite plat- Another interesting feature in Fig. 1 is that, in the absence of form its sensitivity is slightly enhanced as compared to atomis- modifiers, the phosphorus analytical signal does not fall to ation from a pyrolytic surface. Moreover, although analyte zero even at charring temperatures as high as 1800 °C. losses start at the same charring temperature (ca. 1350 °C) However, in the presence of a modifier, the phosphorus signals regardless of whether a standard or a pyrolytic graphite completely disappear at charring temperatures equal to or platform is used, phosphorus signals can still be recorded at higher than 1700 °C. A similar behaviour was observed by charring temperatures as high as 1800 °C when the analyte is Curtius et al.,27 who attributed it to the formation of stable atomised from the standard platform. This behaviour of the phosphorus–carbon species, such as P2C6, or to an interstitial analytical signal, which contrasts with the one observed when compound.According to Oh and Rodriguez,28 phosphorus is atomising from a pyrolytic graphite platform in the presence capable of attacking the arm-chair sites of the graphite strucof a modifier, can be explained on the basis of the model ture to form stable compounds at temperatures higher than proposed by L’vov et al.,31 in which the atomisation of an 1050 °C.29 This might explain the stability of the analytical element from a standard graphite surface can be envisioned as signal of phosphorus at such high charring temperatures, when taking place from the bulk of the graphite. Thus, part of the no modifier is present, as shown in Fig. 1. In the presence of analyte, on injection, penetrates the pores of the surface and, a modifier, the possibility of having free phosphorus species to on heating, is converted into a phosphorus–carbon species,27,28 interact with graphite is greatly reduced in the range of which is retained more strongly and remains in the atomiser temperature in which the modifier is eVective.At higher at temperatures higher than when it is atomised from pyrolytic charring temperatures and/or at the beginning of the atomisgraphite surfaces in which penetration is considerably limited. ation stage, the modifier releases the analyte to an environment Formation of the stable phosphorus–carbon species after reac- in which its interaction with graphite is not favoured due to tion of the phosphate species and the carbon at the active sites the fact that gas expansion promotes losses of the analyte.of graphite, when atomisation occurs from a standard graphite Since no previous, i.e. at lower temperatures, interaction of surface, minimises formation of volatile molecular species phosphorus with graphite to form stable phosphorus–carbon which would be lost on heating.2 The increased sensitivity species occurred, due to the sequestering of the analyte by the observed when atomisation is done from standard graphite modifier, phosphorus is lost and therefore the P atomic absorpsurfaces can be attributed to this behaviour. tion signals do not exist at temperatures equal or higher than Retention of part of the analyte in the porous graphite 1700 °C, as shown in Fig. 1. surface is also indicated by the rising portions of the atomic Several workers have shown that the sensitivity of phosabsorption profiles of phosphorus shown in Fig. 3, which phorus measurements depends on the material of the atomisdenotes faster kinetics for the vaporisation of atoms from the ation surface, either in the presence5,30 or in the absence27 of pyrolytic graphite surface than from the standard graphite chemical modification. The best sensitivity is achieved when surface. Fig. 3 also shows that sweeping of phosphorus atoms the analyte is atomised from standard graphite surfaces (platfrom the observation volume of the atomiser is faster from the form or wall atomisation), because partial stabilisation of pyrolytic surface than from the standard one, in agreement phosphorus by the material of the surface is easier in these with the assumption that less interaction should be expected atomisers.27 In order to determine the eVect of the atomisation between the analyte and the pyrolytic surface. surface on the sensitivity of the phosphorus signal, when NaF is being used as a chemical modifier, ashing plots for phos- Study of the Material Accumulated Onto Graphite Platforms phorus, using phosphate solutions atomised from laboratorymade standard platforms and from commercially available Fig. 4 shows a micrograph of part of the residue left after pyrolytic graphite platforms, inserted into pyrolytically coated successive injections of solutions of NH4H2PO4 and NaF onto graphite tubes, were recorded.These plots are depicted in Fig. 2. Since the diVerent platforms were placed within graphite Fig. 2 Charring plots for the atomisation of a 100 mg ml-1 phos- Fig. 3 Peak profiles of the atomisation of 10 ml of a 100mg ml-1 phorus solution in the presence of 10 ml of a 0.08% m/v NaF solution in standard and pyrolytic graphite platforms. A, Atomisation from a phosphorus solution in the presence of 10 ml of a 0.08% m/v NaF solution. Atomisation from (a) pyrolytic graphite platform, and (b) stan- pyrolytic graphite platform; and B, atomisation from a standard graphite platform. dard graphite platform.Journal of Analytical Atomic Spectrometry, December 1997, Vol. 12 1393Table 2 Results of the EDS analysis of the particles formed at 200 °C (a) Marked with the ‘a’ arrow in Fig. 5 Element Mass (%) Atom (%) K ratio Z A F C K 0.29 0.48 0.0011 1.0715 0.2563 1.0005 O K 38.53 48.08 0.3948 1.0474 0.7021 1.0010 F K 11.26 11.84 0.0738 0.9817 0.4789 1.0013 Na K 33.18 28.81 0.3287 0.9718 0.7317 1.0007 P K 16.73 10.78 0.2016 0.9509 0.9102 1.0000 Total 100.00 100.00 (b) Marked with the ‘b’ arrow in Fig. 5 Element Mass (%) Atom (%) K ratio Z A F C K 0.24 0.41 0.0010 1.0942 0.2978 1.0004 O K 3.96 5.16 0.0417 1.0695 0.7666 1.0039 F K 40.72 44.73 0.4440 1.0025 0.8480 1.0025 Na K 53.85 48.85 0.4997 0.9922 0.8678 1.0000 P K 1.26 0.85 0.0136 0.9709 0.8678 1.0000 Total 100.00 100.00 Fig. 4 Micrograph of the residue left on a graphite platform after injection of NH4H2PO4 and NaF solutions and heating at 200 °C.maximum charring temperature for phosphorus using NaF as a platform that was later heated up to 200 °C. According to a modifier, the accumulated material adopts the shape of this figure, at this low pretreatment temperature the spherical particles dispersed through the platform (see Fig. 6). NH4H2PO4 melts (mp 190 °C)26 and penetrates into the bulk This phenomenon is probably caused by melting of the modifier of the modifier, which has a higher melting point (988 °C).26 at temperatures around 1000 °C during charring and the The rest of the NH4H2PO4, which does not find its way into sudden reduction of temperature from the end of the charring the bulk of the modifier, deposits above and around it. The step, #1350 °C, to the cool-down step, #50 °C, prior to ‘hole’ in the centre of the picture reveals the graphite substrate.atomisation (see Table 1). Melting of the modifier promotes a Fig. 5 shows isolated particles of NaF covered by molten more eYcient mixing, which in turn could lead to a more NH4H2PO4, as shown by EDS analysis of material deposited eYcient trapping, of that part of NH4H2PO4 which at the onto the same platform. The results of the EDS analysis of beginning of the heating programme penetrated its matrix. At these particles, Table 2 (a) and (b), indicate that at 200 °C and these temperatures, that part of the phosphorus which did not at lower temperatures there are no chemical reactions between penetrate the matrix of the modifier must have already been the analyte and the modifier, i.e., there is no indication of lost as some molecular, volatile species.2 formation of any stable P–F species in this range of temperature.Since part of the analyte is just deposited on and around EVect of the Presence of Fluoride as a Modifier on the the modifier and there is no evidence to show formation of a Sensitivity of DiVerent Inorganic Phosphorus Compounds P–F species which could stabilise phosphorus, it is reasonable to expect losses of this portion of the analyte in some molecular Ediger6 showed that in the absence of chemical modifiers the form as the heating programme proceeds.2 On the other hand, sensitivity of phosphorus determinations depended upon the that portion of the analyte which penetrates the bulk of the type of phosphorus compound introduced into the atomiser, modifier could remain trapped, and therefore stabilised, inside but in the presence of a 1% lanthanum solution the diVerences the modifier.This could explain stabilisation of the analyte in sensitivity were eliminated. In order to test the eVect of the and the lower sensitivity of phosphorus determinations with fluoride modifier on the sensitivity of measurements of diVerent NaF as a modifier compared to the sensitivity achieved with phosphorus compounds, solutions of diVerent inorganic phospalladium, as reported by Alvarado et al.8 These assumptions phorus salts were heated in the presence of NaF, and the will be further verified later in this paper.ashing plots were recorded (Fig. 7). At a temperature of 1350 °C, which is quite close to the The results show that the thermal stability of phosphorus Fig. 5 Micrograph showing molten NH4H2PO4; indicated by the ‘a’ Fig. 6 Micrograph showing spherical particles of material deposited on a graphite platform after heating at 1350 °C and sudden cooling arrow: covering the NaF modifier; indicated by the ‘b’ arrow: after heating at 200 °C.at 50 °C. 1394 Journal of Analytical Atomic Spectrometry, December 1997, Vol. 12Fig. 7 Charring plots for 10 ml of 100 mg ml-1 phosphorus solutions Fig. 8 Charring and atomisation curves for phosphorus: A and D, in introduced as diVerent phosphorus compounds, in the presence of NaF. the presence of NaF; B and C in the presence of NaOH. signals in the presence of NaF is practically independent of when using NaF, the presence of NaOH stabilises phosphorus.the type of phosphorus compound being heated. The phos- The maximum charring temperature with no loss of analyte is phates studied behave in a similar way. A definite diVerence is in the vicinity of 1100 °C. This is in agreement with the fact observed for the hexafluorophosphates in the temperature that Na2O sublimes at #1275 °C25 and that trapping is range between room temperature and #500 °C.Otherwise, the possible only when the modifier exists in the condensed phase. trend of the curves is similar to those of the phosphates. Interpretation of the results obtained after the NaOH tests However, the sensitivity of the phosphorus absorbance signals shows that trapping of phosphorus by Na2O formed after is drastically influenced by the type of phosphorus compound oxidation of NaF is not a possible mechanism, given that being heated. Phosphorus signals obtained after heating diVer- retention of phosphorus using NaF is eVective up to, at least, ent phosphates in the presence of NaF are up to four times 1350 °C, a temperature at which Na2O, if formed, should have more sensitive than for the PF6- salts. This behaviour can be completely sublimed.This surmise also finds support in that explained on the basis of the findings already discussed. Part the chemical analysis of the material accumulated onto graphite of the phosphate compound, on melting at temperatures platforms does not reveal the presence of Na2O, at any of the around 200 °C, penetrates the matrix of the modifier and, on diVerent temperatures tested.On the other hand, since NaF further heating, intimately mixes with the modifier, being has a reported boiling point of 1695 °C, trapping of phosphorus trapped in it. On the other hand, NaPF6 does not melt but in this compound at temperatures as high as 1350 °C is decomposes according to reaction (1), which starts at about possible.For this kind of stabilisation mechanism, the fact that 270 °C and is complete at around 650 °C: the release of the analyte occurs at temperatures lower than the boiling points of the trapping compounds, 1275 °C for NaPF6(s)�NaF(s)+PF5(g) (1) Na2O, 1695 °C for NaF and 3600 °C for MgO, could be Formation of the volatile PF5 molecule leads to losses of explained assuming that, at the charring temperatures at which phosphorus at relatively low temperatures, which is reflected losses of the analyte are first observed, the kinetic energy of in the relatively low sensitivity of the measurements for hexa- the trapped compound is such that it is able to free itself from fluorophosphate compounds shown in Fig. 7. trapping at that temperature even though the change in phase of the modifier has yet to occur. In the particular case of phosphorus, the analyte could be released as some molecular, Possibility of Formation of Na2O After Heating NaF rather than atomic, form at temperatures around 300 °C lower Stabilisation of an analyte by physical trapping in the structure than the boiling point of NaF.This assumption finds support of the modifier has been shown to operate for the use of in Fig. 9. The blank space observed between #1700 and Mg(NO3)2.32 This compound, on heating, transforms itself into MgOand the analyte is retained, trapped, in the bulk of the oxide. The possibility that a similar mechanism, i.e.conversion of NaF into Na2O and trapping of phosphorus in the bulk of the oxide, operates for stabilisation of phosphorus by NaF was considered and later discarded by the following tests. Aqueous phosphorus solutions, of the same phosphorus concentration as used in the previous tests, 100 mg ml-1, obtained by dissolution of NH4H2PO4, were atomised in the presence of NaOH solutions having the same concentration (0.08% m/v) as that of the NaF solutions previously used. Sodium hydroxide was chosen since this compound, at relatively low temperatures, transforms into Na2O.Formation of Na2O in the early stages of the heating will favour trapping of phosphorus in the structure of this molecule. Fig. 8 shows charring and atomisation curves obtained during this test. The corresponding curves using NaF as modifier are also shown for Fig. 9 Charring (A) and atomisation (B) curves for phosphorus in comparison purposes. According to Fig. 8, although the sensi- the presence of NaF as a modifier showing a blank space between #1700 °C and #2100 °C.tivity of the measurements is around 1.7 times lower than Journal of Analytical Atomic Spectrometry, December 1997, Vol. 12 13952100 °C in Fig. 9 has been found in every charring and REFERENCES atomisation graph for phosphorus plotted during this work. A 1 Manning, D. C., and Slavin, S., At. Absorpt. Newsl., 1969, 8, 132. similar behaviour was observed by Persson and Frech,2 in the 2 Persson, J.A., and Frech, W., Anal. Chim. Acta, 1980, 119, 75. temperature range of ca. 2000–2100 °C. Usually one expects 3 Prevo� t, A., and Gente-Jauniaux, M., At. Absorpt. Newsl., 1978, the atomisation curve to begin at temperatures slightly higher 17, 1. than the maximum charring temperature allowed without 4 Havezov, I., Russeva, E., and Jordanov, N., Fresenius’ Z. Anal. Chem., 1979, 296, 125. losing analyte. The atypical feature shown in Fig. 9 could be 5 Curtius, A.J., Schlemmer, G., and Welz, B., J. Anal. At. Spectrom., indicative of losses of analyte in this range of temperature, in 1987, 2, 115. the presence of this particular modifier. This blank space could 6 Ediger, R. D., At. Absorpt. Newsl., 1976, 15, 145. mean that at temperatures slightly higher than the maximum 7 Welz, B., Curtius, A. J., Schlemmer, G., Ortner, H. M., and charring temperature, phosphorus is released not as an atomic Birzer, W., Spectrochim. Acta, Part B, 1986, 41, 1175.species, as deduced from the lack of atomic absorption signals, 8 Alvarado, J., Cristiano, A. R., and Curtius, A. J., J. Anal. At. but most likely as a molecule. These molecules will dissociate Spectrom., 1995, 10, 483. 9 Ortner, H. M., Schlemmer, G., Welz, B., and Wegsheider, W., to produce the atomic species at temperatures close to 2100 °C, Spectrochim. Acta, Part B, 1985, 40, 959. at which atomic absorption is first observed, according to 10 Welz, B., Curtius, A., Schlemmer, G., and Ortner, H.M., Fig. 9. According to Persson and Frech,2 the main phosphorus Spectrochim. Acta, Part B, 1986, 41, 567. species existing in the temperature range #1200–2200 °C, with 11 Ortner, H. M., Birzer, W., Welz, B., Schlemmer, G., Curtius, A. J., the minimum value depending upon the type of phosphorus Wegsheider, W., and Sychra, V., Fresenius’ Z. Anal. Chem., 1986, compound originally heated, is the gaseous P2 molecular 323, 681. species.It is reasonable to expect, on the basis of the findings 12 Mu� ller-Vogt, G., and Wendl, W., Anal. Chem., 1981, 53, 651. 13 Bendicho, C., and de Loos-Vollebregt, M. T. C., Spectrochim. of these authors, that this could be the main species responsible Acta, Part B, 1990, 45, 679. for phosphorus losses during thermal pre-treatment. 14 Jianping, W., and Bo, D., Spectrochim. Acta, Part B, 1992, 47, 711. 15 Qiao, H., and Jackson, K. W., Spectrochim. Acta, Part B, 1992, 47, 1267.CONCLUSIONS 16 Shan, X. Q., and Wang, D. X., Anal. Chim. Acta, 1985, 173, 315. 17 Jianping, W., and Bo, D., Spectrochim. Acta, Part B, 1992, 47, 711. Stabilisation of phosphorus in the presence of NaF as a 18 Peng-yuan, Y., Zhe-ming, N., Zhi-xia, Z., Fu-chun, X., and modifier proceeds via physical trapping of the analyte in the An-bei, J., J. Anal. At. Spectrom., 1992, 7, 515. bulk of the modifier. Losses of phosphorus in some molecular 19 Xuan, W., Spectrochim. Acta, Part B, 1992, 47, 545.form, mainly as P2, are likely to be responsible for the lower 20 Welz, B., Akman, S., and Schlemmer, G., Analyst, 1985, 110, 459. sensitivity of the measurements as compared to that obtained 21 Dedina, J., Frech, W., Cedergren, A., Lindberg, I., and using Pd. The proposed mechanism for stabilisation of phos- Lundberg, E., J. Anal. At. Spectrom., 1987, 2, 435. phorus by NaF could also be applicable to other fluoride salts 22 Akman, S., and Doner, G., Spectrochim.Acta, Part B, 1994, 49, 665. 23 Akman, S., and Doner, G., Spectrochim. Acta, Part B, 1995, 50, 975. such as LiF (bp 1670 °C), KF (bp 1498 °C) and CsF (bp 24 Akman, S., and Doner, G., Spectrochim. Acta, Part B, 1996, 1253 °C), which have been previously shown to stabilise this 51, 1163. analyte.8 However, HF and NH4F, which do not leave solid 25 Doner, G., and Akman, S., Spectrochim. Acta, Part B, 1996, 51, 181. residues after heating and which also exert a stabilisation eVect 26 Handbook of Chemistry and Physics, ed. Weast, R. C., CRC Press, on phosphorus,8 must act in a diVerent way. Tests are being Boca Raton, FL, 53rd edn., 1972. carried out to check the possibility of stabilisation via chemical 27 Curtius, A. J., Schlemmer, G., and Welz, B., J. Anal. At. Spectrom., reaction of the analyte and these fluoride compounds. In spite 1986, 1, 421. of the lower sensitivity of phosphorus measurements, the use 28 Oh, S. G., and Rodriguez, N. M., J. Mater. Res., 1993, 8, 2879. 29 Holcombe, J. A., and Droessler, M. S., Fresenius’ Z Anal. Chem., of NaF as a modifier for phosphorus determination by ETAAS 1986, 323, 689. is an economical and convenient alternative which could be 30 Ediger, R. D., Knott, A. R., Peterson, G. E., and Beaty, R. D., At. useful when analysing samples with phosphorus concentration Absorpt. Newsl., 1978, 17, 28. levels which do not demand the highest sensitivity from this 31 L’vov, B. V., Bayunov, P. A., and Ryabchuck, G. N., Spectrochim. technique. Acta, Part B, 1981, 36, 397. 32 Slavin, W., Carnrick, G. R., and Manning,, Anal. Chem., 1982, 54, 621. The authors thank Dr. Bernhard Welz, from Perkin-Elmer Corporation, U� berlingen, Germany, for donating the dual- Paper 7/04689H cavity platforms. Thanks are also due to CONICIT for Grant Received July 3, 1997 MPS-RP-VII260076 and to the Decanato de Investigaciones at Universidad Simo�n Bolý�var. Accepted August 16, 1997 1396 Journal of Analytical Atomic Spectrometry, December 1997, V