摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, OCTOBER 1987, VOL. 2 723 Mechanisms of lonisation and Atomisation of Barium in Graphite Furnace Atomic Absorption Spectrometry Etsuro Iwamoto, Shuichi Ohkubo, Manabu Yamamoto and Takahiro Kumamaru Department of Chemistry, Faculty of Science, Hiroshima University, Hiroshima 730, Japan The mechanisms of ionisation and atomisation of barium in graphite furnace atomic absorption spectrometry were studied, from the viewpoint of activation energies for ion and atom formation, and the effect of additions of alkali metals to the analyte solution on sensitivity of barium depending on pyrolytic graphite (PG) and non-pyrolytic graphite (NPG) tubes. It was found that the addition of caesium chloride to the barium solution suppresses the barium ionisation for the NPG tube but not for the PG tube as the caesium in the PG is dissipated before the ionisation of barium begins, while for the NPG tube it was present in the tube during the ionisation of barium.An activation energy (€,) of 770 kJ m0l-l for the ionisation of barium was obtained and it was attributed to an ionisation mechanism involving barium carbide, unlike theatomisation with an E, value of 540 kJ mol-1 which correlates with the breakage of the Ba-0 bond. The reason why the ionisation suppressant does not lead to a consequent increase in the atomic absorption signal was explained in terms of the differing mechanisms for the atomisation and ionisation of barium. Keywords: Barium; ionisation; atomisation; graphite furnace atomic absorption spectrometry; activation energy It is generally recognised that the effect of the analyte ionisatian on atomic absorption in the graphite furnace is small compared with that in a flame because of the relatively low temperature experienced by the atomic vapour, different time and temperature dependencies of the atom and ion populations and the high background concentration of free electrons generated in the furnace.Ottaway and Shawl studied atomic and ionic absorptions and emission signals for barium in the graphite furnace, and the effects on these signals of adding an excess of caesium to the analyte. They observed that the sensitivity of barium was 40-fold higher for atomic absorption than for ionic absorption. They concluded that the extent of barium ionisation was negligible. Sturgeon et a1.2 determined the concentration of free electrons in a graphite tube atomiser and concluded that the analyte ionisation in the graphite furnace should be negligible for elements having ionisation potentials greater than 4.6 eV, which is the work function for graphite.They reported a sensitivity difference of 10-fold between atomic and ionic absorptions.3 In these experiments mainly non-pyrolytic graphite (NPG) tubes and a heating cycle programme with a ramp time (350-650 "C s-1) at the atomisation step were used. Recently, however, it was found that pyrolytic graphite (PG) coated tubes enhanced the sensitivity for refractory elements such as Mo, Ti and V4-7 and a rapid heating programme with a ramp time of 0 s on atomisation (ca.1800 "C s-1, maximum power heating mode)8 is effective for the enhancement of sensitivity. Although it has been pointed out that the electron concentration in the graphite furnace plays an important role, the mechanism of ionisation is not always resolved ade- quately. A comparison of the NPG and PG tubes and the use of the maximum power heating mode would be expected to give some indication of the ionisation phenomena, as the quality of the graphite surface and matrices in analyte solutions greatly influences the sensitivity. In this study, activation energies for the formation of barium ions and atoms were obtained from the Arrhenius plots obtained with the PG and NPG tubes. The difference in mechanisms between the ionisation and atomisation of barium is discussed with respect to the effects of the addition of alkali metals to the analyte solution, ashing temperatures and purge-gas flows on the sensitivity of barium.Experimental Apparatus A Perkin-Elmer Model 5000 atomic absorption spectrometer equipped with a tungsten background corrector and a Model HGA-500 graphite furnace was used with a Model AS-40 autosampler. A Perkin-Elmer Data System 10 and a Wata- nabe Model WX4657 plotter were used to record the absorbance - signal profiles. Hamamatsu TV hollow-cathode lamps were used as the light source and argon was used as the purge gas. The analytical wavelengths and spectral band widths were 553.6 and 0.14 nm for barium atoms, 455.4 and 0.14 nm for barium ions and 852.1 and 0.4 nm for caesium atoms, respectively. The tube background and caesium emission signals were measured under the same gain.The caesium emission profile was obtained by subtraction of the background emission signal from the total emission signal at each time. A typical furnace programme is given in Table 1. Maximum-power heating was used for the atomisation step. The graphite furnaces used were of two types with different Lot Nos. (i) NPG tube (Perkin-Elmer, Part No. 070699: Lot Table 1. Instrumental parameters Ashing Parameter Drying Atomisation Conditioning . . . . . . . . . . Step No. 1 2 3* 4 5 Temperature/"C . . . . . . 140 Variable 700 2800 2700 Ramp time/s . . . . . . . . 40 30 1 0 1 Hold timeis . . . . . . . . 5 5 10 4 1 Internal gas flowiml min-1 . . 300 300 300 20 300 * This step was used only for examination of ashing temperature effects (Fig.2).724 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, OCTOBER 1987, VOL. 2 Nos. VI/168/510 and VIII/69/305); and (ii) PG tube (Perkin- Elmer, Part No. 091504: Lot Nos. VI/215/628 and 6-08050- 069). A Chino pyroscope Q and a two-wavelength emission intensity ratio method9 were used to calibrate the temperature settings. Reagents All solutions were prepared from analytical-reagent grade chemicals and de-ionised water, and stored in polyethylene bottles. A standard barium solution (1000 pg ml-1) was prepared by dissolving barium chloride in 1 mol 1-1 hydro- chloric acid. Results and Discussion Typical transient signals for the atomic and ionic absorption of barium are shown in Fig. 1 in which the axes are labelled in such a way that zero time has elapsed before the initiation of the atomisation stage of the heating cycle.Profiles with a sharp peak characteristic of the ionisation phenomenon were observed when amounts of more than 0.5 ng of barium for PG tubes and 2 ng for NPG tubes were atomised at temperatures above 2600 "C. The peak height due to the sharp peak gave a lower precision in some instances. Therefore, peak area (absorbance x seconds) was used throughout this work. The effect of ashing and atomisation temperatures on the atomic and ionic absorption signals were examined with both the PG and NPG tubes. The results are shown in Figs. 2 and 3. It should be noted that at higher ashing temperatures the relative ionic absorbance is higher than the relative atomic absorbance for both furnaces (Fig.2). The appearance temperatures (Tapp) for atomic and ionic absorption signals for PG tubes were 2100 and 2250 "C, respectively, in contrast to 1290 "C for the atomic signal and 1890 "C for the ionic signal reported by Sturgeon and Berman.3 The principal difference between the two studies was the amount of barium taken: in this study 1 ng of barium was taken for both atomic and ionic absorption measurements whereas in the study of Sturgeon and Berman3 100- and 600-ng amounts were used for atomic and ionic absorption, respectively. Although no signal could be ob- served with 1 ng of barium at atomisation temperatures below 1900 "C (Fig. 3), it was found that when 100 ng of barium were 0.8 a c $ 0.4 2 0 2 0 1 2 Time/s Fig. 1. Atomic and ionic absorption signals for barium.(a) PG tube (Lot No. VI/215/628), 1 ng of Ba; and ( b ) NPG tube (Lot No. VI/168/510), 4 ng of Ba. A, Ionic absorption; and B, atomic absorption. Ashing temperature, 700 "C I I 700 1300 1900 700 1300 1900 Ashing temperaturePC Fig. 2. Effects of ashing temperature on barium absorbances. (a) and (b) are the same as in Fig. 1. A, Ionic absorption; and B, atomic absorption. Atomisation temperature, 2800 "C taken, atomisation at 1500 "C gave an absorbance of 0.017 for atomic absorption, and when 300 ng were taken, atomisation at 1900 "C gave an absorbance of 0.036 for ionic absorption. Thus, the differences in the Tapp values mentioned above may be due to the differences in the amounts taken. The dependence of appearance temperatures on the amount taken was also observed for caesium atomisation, as seen later (Fig.6). Excessively increasing the amount of analyte present leads to a lowering of the appearance temperatures. Fig. 3 shows that the sensitivities are dependent on the quality of graphite (PG and NPG) with different Lot Nos. There are temperatures above which the sensitivity for ionic absorption is higher than that for atomic absorption, when the tubes with Lot Nos. VI/215/628 and VI/168/510 were used. The graphite tube which gives a higher sensitivity for atomic absorption produces a lower sensitivity for ionic absorption. Atomisation at 2800 "C was carried out to investigate the ionic absorption phenomenon throughout the work concerning the Arrhenius plots. Analyte ionisation is known to be a mass-action process in flame and graphite furnace atomic absorption and to be suppressed by the addition of easily ionised elements to the analyte solutions.l.3~10 Figs.4 and 5 show the effect of the addition of alkali metal salts on the atomic and ionic absorbances for NPG and PG tubes, respectively. The atomic absorbances for NPG tubes increase slightly when less than 200 ng of an alkali metal are present whereas those for PG tubes remained approximately constant (Fig. 4). It is interest- ing that, although the ion signals obtained with NPG tubes decrease markedly with 5-fold excess amounts of the alkali metals over barium, as expected, those obtained with PG tubes do not decrease even with 250-fold excess amounts (Fig. 5). The difference between NPG and PG tubes can be explained by the difference in the residence times of the alkali metal atoms during atomisation, as discussed below.The time characteristics of the caesium atomic absorption and emission, and barium ionic absorption signals are shown in Fig. 6. For PG tubes which give higher and narrower caesium atomic absorption and emission signals, compared with NPG tubes, the signals are weakened when the ionic absorption signal for barium begins to appear. The matrix salts deposited on the PG surface do not penetrate the interior and are easily dissipated compared with the NPG surface. 0.4 0.2 m 9 2 a 9 0.4 0.2 2200 2500 2800 0.4 0.2 0 2200 2500 2800 Atomisation temperature/"C Fig. 3. Effects of atomisation temperature on barium absorbances. (a) and (b) are the same as in Fig.1; ( c ) PG tube (Lot No. 6-08050-069), 1 ng of Ba; and (d) NPG tube (Lot No. VIII/691305), 4 ng of Ba. Ashing temperature 700 "C. A and B, are the same as in Fig. 2JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, OCTOBER 1987, VOL. 2 725 1 a, m C n Added metalhg Fig. 4. Effects of addition of alkali metals on barium atomic absorbances. (a) and ( b ) are the same as in Fig. 1. A, Na; B, K; C, Rb; and D, Cs 0.5 0 Added metal/ng Fig. 5. Effects of addition of alkali metals on barium ionic absorbances. (a) and ( b ) are the same as in Fig. 1. A, Na; B, K; C, Rb; and D. Cs I I 0 1 2 3 2.0 1.5 1 .o 0.5 0 0 1 2 3 Time/s Fig. 6. Atomic absorption (AA) and emission (AE) signals versus time characteristics for caesium and ionic absorption signals versus time characteristics for barium.(a) and ( b ) are the same as in Fig. 1. A, Cs AA, 2 ng; B, Cs AA, 1000 ng; C, Cs AE, 2 ng; D, Cs AE, lo00 ng; E, Cs AA, 2 ng; F, Cs AA, 500 ng; G, Cs AE, 2 ng; and H, Cs AE, 20 ng It was clearly shown by flame atomisation that the addition of an easily ionised .metal to the barium analyte inhibited ionisation and led to an increase in the atomic absorption signal which compensated for a decrease in the ionic absorp- tion signal.I0 However, such a marked effect has not been observed for the graphite furnace (Fig. 4). Although an addition of a 25-fold excess of caesium suppresses the ionic signal when using the NPG tube, the addition of alkali metals leads to an increase in sensitivity of only ca. 10% for barium atomic absorption.Ottaway and Shawl and Sturgeon and Berman3 also reported that for an atomisation temperature of 2300 "C there was a 40- or 10-fold difference, respectively, in sensitivity between barium atomic and ionic absorption, and suppression of the ionisation did not produce a consequent increase in the atomic signals; at an atomisation setting of 2700 "C the sensitivity for the ionic line was only 17% smaller than that for the atomic line but the addition of a 40-fold excess of caesium increased the atomic absorption signals by only 12% .3 This was attributed to the small degree of ionisation and the negligible ion population at the time of measurement of the atomic signals.1 An ionisation of over 90% was reported for the dinitrogen oxide - acetylene flarne,lO while only about a 10% ionisation has been reported for the graphite surface.3 In this work, a comparable sensitivity was obtained between atomic and ionic absorption and the residence times of atom and ion populations were comparable.It was pointed out that the important factors which govern sensitivity are the oscilla- tor strength fl1-13 and the lower level populations of the absorbants.12 The f-values for the atomic and ionic absorption of barium are 1.58 and 0.727, respectively.14 A higher sensitivity may be expected for atomic absorption. This is supported by the results in flame atomic absorption: in the dinitrogen oxide - acetylene flame, the total absorbance for the atomic and ionic absorbances increased when the ionic absorbance was decreased by the addition of the ionisation suppressant K.10 As the decrease in the ionic signals from the726 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, OCTOBER 1987, VOL.2 graphite furnace does not lead to a corresponding increase in the atomic signals, it is unlikely that the decrease is only due to the small degree of ionisation and the difference in residence time. Therefore, another explanation will be proposed from the results of the kinetic data below. Chungls proposed a method for the analysis of the atomisation mechanism with Arrhenius plots, taking the dissipation function into account, in which the rate constant of atom formation ( k ) at a given temperature was calculated by using the corresponding absorbance (A,) and the maximum absorbance (Amax): k = kdAt/(Am,, - A,) where kd is the dissipation constant. This equation gives Arrhenius plots with a better linear relationship compared with the equation offered by Sturgeon et al.16 The Chung method15 was applied to the analysis of the present atomisation and ionisation systems. When the absorption has a horn-like peak at the top of the signal, the absorbance at the beginning of the horn-like peak was taken as A,,,. Fig. 7 shows typical Arrhenius plots for the barium atomic and ionic absorption. Good linear relationships were obtained. Activation energies (E,) obtained for atom and ion formation are given in Tables 2 and 3, respectively, together with Tapp values. It can be seen that the activation energy does not depend on the ashing tempera- tures, the type of acid or the nature of the surface (PG and NPG with different Lot Nos.) of the tubes for either atomic or ionic absorption.Further, the E, value for ion formation obtained in the presence of 20 ng of caesium, as an ionisation suppressant, is nearly the same as that in the absence of caesium. Therefore, neither of the mechanisms of atomisation and ionisation is changed by the various parameters and conditions used. It is worth noting, however, that a marked difference in activation energies between atomic and ionic absorption was observed. Recently the atomisation mechanism for barium was investigated by Styrisl7 using graphite furnace atomic absorp- tion spectrometry and mass spectrometry simultaneously, and several species [Ba(OH)2, BaC1, BaO, BaC2, etc.] were found for vitreous carbon and pyrolytic graphite furnaces.Frech et al. 18 also reported the calculated distribution of barium species as a function of temperature and showed that gaseous barium oxide occurred above ca. 1430 "C. The average E, value obtained for atomic absorption was 540 t 20 kJ mol-1, ~- ~ Table 2. Activation energies (E,) of atom formation for barium* HCl Ashing temperature/'C T,,,I"C EJkJ mol-' 300 2030 596 700 2020 509 1100 2150 503 1600 2180 538 PG tubet- 2050 544; Mean: - 538 f 37 HN03 - T,,,/"C E,/kJ mol-- 1 21 10 595 2100 529 2050 528$ 2220 540 2130 561 - 551 t 28 T,,,,I"C E,,/kJ rnol-1 2060 595 2050 545 2050 547; 2150 49 1 2150 530 - 542 k 37 NPG tube§- 300 2050 509 1980 474 2050 503 700 2100 547 2040 53 1 2100 547 1100 2090 582 2000 533 2000 600 1600 2000 48 I 2040 547 2100 539 1920 5057 1950 5397 1980 53917 Mean: - 525 k 40 - 525 2 29 - 546 t- 35 * Concentrations of acids: HC1 and HN03, 0.05 rnol 1 - 1 ; H2S04, 0.025 rnol 1 - I .1' Lot No. VI/215/628. 3 Lot No. VI/168/510. 17 Lot No. VIII/69/305. $ Lot NO. 6-08050-069. Table 3. Activation energies (E,) of ion formation for barium* HCl HN03 Ashing temperature/"C TaPp/"C E,/kJ mol- Ta,,/"C EJkJ rnol-1 PG tubet- 300 2270 746 2220 74 1 700 2260 805 2260 819 1100 2330 807 2270 766 1600 2240 761 2240 788 2230 706$ 2230 738$ Mean: - 765 k 42 - 770 ?c 34 NPG tube§- 300 2220 782 2250 740 700 2220 755 2260 745 2150 8201 2320 8561 2280 846 1100 2250 762 1600 2180 730 2180 687 2240 8041 I Mean: - 776 k 33 - 775 * 73 * Concentrations of acids are as in Table 2.t, $, 9, 7 Lot Nos. as in Table 2. I/ With 20 ng of added caesiurn. H2S04 T,,,/"C 2190 2250 2260 2290 2220 21 10 2240 2230 2300 2180 EJkJ rnol- I 745 804 805 $ 814 743 782 k 35 710 760 8397 908 685 780 -t 92JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, OCTOBER 1987, VOL. 2 727 1.5 t \ I I I I I 3.4 3.6 3.8 4.0 4.2 4.4 ' O " , K - l T Fig. 7. Arrhenius plots for barium atomic and ionic absorption. A , Atomic, NPG tube; B, atomic, PG tube; C, ionic, NPG tube; and D, ionic, PG tube regardless of the nature of the tubes and ashing temperature. This value correlates well with the dissociation of Ba-0 bonds (546 kJ mol-1)14 and it is proposed that the rate determining step is the breakage of the Ba-0 bond in the gaseous oxide: An average E, value of 770 f 20 kJ mol-1 for ionic absorption was obtained.The ionisation potential of barium is 509 kJ mol-1 and there is no elementary process correspond- ing to the E, value obtained. Sturgeon and Berman3 discussed mechanisms of ionisation and suggested that chemi-ionisation processes involving the dissociation of the metal oxide are representative of the reactions in the graphite furnace: C,,) + MO,,) + M+(,) + e-,,) + co,,, a (3) Energies for barium in reactions (2) and (3) are 523 and 695 kJ mol-1, respectively. Although reaction (3) has an energy close to the value obtained, it seems that the participation of BaO,,, in the ionisation is excluded by the ionisation- suppressant effect mentioned above and the purge-gas effect stated later. It has been pointed out that barium can form a carbide in the graphite furnace.19 The BaCz which was found above a temperature 1480 "C, by mass spectrometry measurements,17 was postulated to be produced by the reaction of barium atoms adsorbed on to the graphite surface with gaseous carbon monoxide or carbon in the solid state.Therefore, the following reaction, corresponding to the decomposition reac- tion for BaC2 as a bulk entity, with an energy of 764 kJ mol-1, is proposed to be the rate-determining step: (764 kJ mol-1) BaC,,,, 'Ba+(g) + 2C,q + e-(g) * * (4) Therefore, the difference in E, between atomisation and ionisation indicates that the ionisation mechanism involves barium with an origin different to that for the atomisation mechanism. This theory is supported by the effect of the purge-gas flow in Fig.8, which shows that the atomic absorption signal decreases with increasing flow-rate, while the ionic absorption signal is less influenced, particularly for the NPG tube. The melting- and boiling-points of BaO are 1918 and 2000 OC,14 respectively, and the melting-point of BaC2 is 3000 OC.20 At each appearance temperature for atomic a , ' + :: n m 0.5 C m al .- c m a, IT - 0 k; 0 20 50 100 200 300 0.5 0 Flow-rateiml min-1 Fig. 8. absorbances. ( a ) and ( b ) , and A and B are the same as in Fig. 2 Effects of the argon gas flow-rate on barium atomic and ionic and ionic absorption, BaO is in the gaseous state and BaC2 is in the solid state. For atomic absorption, the gaseous species BaO,,) is more easily dissipated in the rate-determining step than is the solid species BaC2(,) for ionic absorption.Further, the difference in the effects of the ashing temperature between ionic and atomic absorption signals (Fig. 2) indicates that the ionising species is more stable than the atomising species. The different origins of barium for atomisation and ionisa- tion gives another explanation of the effect of addition of easily ionised alkali metals to the barium analyte on barium atomic and ionic absorption signals: in reaction (4) the addition of alkali metal suppresses the ionisation but not the carbide formation. It is unlikely that the carbide changes into barium oxide and then to barium atoms. The sharp horn-like signal may result from mechanisms different to those at the leading edges of the signals. Speculative explanations are: either that barium atoms formed in reaction (1) ionise at a higher temperature, leading to an abrupt increase in the ionic absorbance, and some of the barium ions formed in reaction (4) are reduced by the surrounding electrons, leading to an abrupt increase in the atomic absorbance, or that the atoms and ions condensed in the cooler parts of the furnace are re-vaporised at higher temperatures.Conclusion In view of the fact that the ionisation suppressants increase, to some extent, the atomic signals, it is likely that the gaseous barium atoms [after reaction (l)] are partly ionised. There- fore, it is indicated that the barium oxide formation contri- butes predominantly to the atomic absorption signal and partly to the ionic absorption signal, and the carbide formation exclusively to the ionic absorption signal.Once barium carbide has been formed, there is a resultant decrease in the atomic absorption signal. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. References Ottaway, J. M., and Shaw, F., Analyst, 1976, 101, 582. Sturgeon, R. E . , Berman, S. S., and Kashyap, S . , Anal. Chem., 1980,52, 1049. Sturgeon, R. E . , and Berman, S. S . , Anal. Chem., 1981, 53, 632. Thomson, K. C., Godden, G. R., and Thornerson, D. R., Anal. Chim. Acta, 1975, 74, 289. Manning, D. C., and Ediger, R. D., At. Absorpt. Newsl., 1975, 15,42. Sturgeon, R. E., and Chakrabarti, C. L., Anal. Chem., 1977, 49, 90. Slavin, W., Manning, D. C., and Carnrick, G. T., Anal. Chem., 1981, 53, 1504. Slavin,W., Carnrick, G. R., and Manning, D. C., Anal. Chem., 1982, 54, 621. Tsujino, R., Kishimoto, T., and Ikeda, M., Anal. Chim. Acta, 1976,84, 283. Manning, D. C., and Capacho-Delgado, L., Anal. Chim. Acta, 1966, 36, 312.728 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, OCTOBER 1987, VOL. 2 11. 12. 13. 14. 15. 16. Walsh, A., Spectrochim. Acta, 1955, 7 , 108. L’vov, B. V., “Atomic Absorption Spectrochemical Analysis,” Adam Hilger, London, 1970. Margoshes, M., Anal. Chem., 1967,39, 1093. Weast, R. C . , Editor, “Handbook of Chemistry and Physics,” . 64th Edition, CRC Press, Boca Raton, FL, USA, 1983. Chung, C. H., Anal. Chem., 1984, 56, 2714. Sturgeon, R. E., Chakrabarti, C.L., and Langford, C. H., Anal. Chem., 1976, 48, 1792. 17. Styris, D. L., Anal. Chem., 1984, 56, 1070. 18. Frech, W., Lundberg, E., and Cedergren, A., Prog. Anal. At. Spectrosc., 1985, 8, 257. 19. L’vov, B. V., Spectrochim. Acta, Part B, 1978, 33, 153. 20. Lagas, P., Anal. Chim. Acta, 1978, 98, 261. Paper J7l34 Received March 23rd, 1987 Accepted June 1 st, 1987

ISSN:0267-9477    

DOI:10.1039/JA9870200723

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, OCTOBER 1987, VOL. 2 729 SHORT PAPERS Determination of Barium, Calcium, Iron, Potassium, Magnesium and Sodium in High Purity Niobium Pentoxide by Flame Atomic Absorption Spectrometry Teresa Joanna Chruscinska Institute of Electronic Materiais Technology, Analytical Department, Konstruktorska 6, PL-02-673 Warszawa, Poland A trace analysis method for niobium pentoxide by flame atomic absorption spectrometry was developed. Determination of 10-400 pg g-1 of Ba, 1-80 pg g-1 of Ca, 4-400 pg g-1 of Fe, 0.5-40 1-19 g-1 of Mg (in a dinitrogen oxide -acetylene flame), 0.5-40 pg g-1 of K (in an air - hydrogen flame) and 0.3-40 pg g-1 of Na (in an air-acetylene flame) were carried out without matrix separation. The sample was decomposed with a H N 0 3 - HF mixture using a bomb technique.Hydrogen peroxide was added to transform niobium into the peroxide complex to prevent any hydrolysis subsequent to the binding of the fluoride ion by H3B03. Solutions containing 0.4% V/V HN03,4% V/V HF, 2% V/V H202, 2% m/V H3B03 and N b as 2% m/V N b205 were studied. The calibration graph method was used. Matrix interferences were overcome by preparing two separate sets of standards for the blank and the sample solutions (containing niobium). lonisation interferences were suppressed by 0.1% CsCI. Background correction was also found to be necessary. Keywords : Metal determination; flame atomic absorption spectrometry; niobium pentoxide The use of niobium pentoxide as a raw material for the electronic industry creates a requirement for the control of its purity.The analysis of this material has been the subject of numerous papers. Various techniques have been used, i. e . , radiometry,l-5 spectrography,6-8 polarography,"l2 extractive gravimetry,13 spectrophotometry,'k17 flame photometry17 and atomic absorption spectrometry,l8 to determine, among others, trace amounts of Ca,7 Fe,7,15,18 K17 and Na.2 Samples were decomposed by fusion with ammonium hydrogen sulphate, 19 ammonium sulphate,20 selenium dioxide21 and potassium pyrosulphate ,I7 or by dissolution with sulphuric acid5.17 or a hydrofluoric - nitric acid mixture.13J8722 Determi- nation of transition metals by flame (Fe) or electrothermal atomisation (Cu, Co, Ni, Mn, Cr) atomic absorption spec- trometry was preceded by ion-exchange separation of niobium to prevent the formation of niobium carbide in the graphite atomiser.18 The determination of trace amounts of common elements encounters difficulties due to contamination effects.Although in this instance direct instrumental methods would be prefer- able, these suffer from matrix interferences and sample inhomogeneity. As flame AAS is less prone to these interfer- ences laborious matrix separation is avoided and the system- atic errors due to the blanks are minimised, unless significant contamination ensues during the decomposition process. Decomposition should preferably be carried out in closed systems with a minimum surface area. The reagents should be very pure and the amounts used as small as possible. These requirements are met in the present method for the determi- nation of Ba, Ca, Fe, K, Mg and Na in high purity niobium pentoxide by using flame AAS without matrix separation.Experimental Instrumentation Studies were carried out using a double-beam Perkin-Elmer Model 430 atomic absorption spectrometer equipped with a background corrector (deuterium and halogen lamps). Per- kin-Elmer Intensitron hollow-cathode lamps were used as sources. The instrument has been modified to permit measurement of the height of observation. Reagents and Solutions Reagents were of electronic or spectral grade, except for CsCl, which was of analytical-reagent grade. De-ionised, doubly distilled water obtained from a quartz still was used through- out. Stock standard solutions of Ba, Ca, Fe, K, Mg and Na (1 mg ml-1 of each of the metals) were prepared as nitrates in 1% VlVHN03.Sample Preparation Treat the sample (1 g) in a Teflon pressure-decomposition vessel23 with 0.2 ml of 64% HN03 and 2 ml of 50% HF. Close the vessel, place it in a graphite heater for 2.5 h at 160 "C. Cool, introduce 1 ml of 30% H202 into the vessel and transfer the solution quantitatively into a 50-ml polyethylene calib- rated flask. Add 25 ml of 4% H3B03 and dilute to the mark. Results and Discussion Remarks on Sample Preparation Introduction of boric acid into the niobium fluoride solution in amounts sufficient to form fluoroboric acid causes hydrolysis of niobium salts, even though quantitative anion-exchange separation of niobium using hydrofluoric acid - boric acid media is possible.24 However, hydrolysis can be delayed for a period of several days when a peroxide - niobium complex has previously been formed (the yellow solution formed after the addition of hydrogen peroxide).Therefore, boric acid can be used as a masking agent in the spectrophotometric determina- tion of niobium using hydrogen peroxide in the presence of the fluoride ion? Increasing the boric acid concentration in order to inhibit the hydrolytic decomposition of fluoroboric acid26 leads to the hydrolysis of the niobium salt. The cationic constituents of the solutions (real samples or standards) contribute to an increase in the rate of hydrolysis; the standard solution without any metallic elements present is the most stable. Recovery The loss of elements during dissolution of the sample or due to the formation of insoluble fluorides and the occurrence of730 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, OCTOBER 1987, VOL.2 Table 1. Operation conditions of the atomic absorption spectrometer. Height of observation = h. A 10-cm long burner was used for the air - acetylene flame and 5 cm for other flames Sensitivity for 1% Wavelength/ Slit width/ Lamp current/ hi Oxidant, flow*/ Fuel, flow/ absorption/ Element nm nm mA mm 1 min-1 1 min-1 pgml-1 Ba . . . . . . . . 553.6 0.2 25 6 N20,4.3 C2H2,4.6 0.841 Ca . . . . . . . . 422.4 0.7 10 6 N20,4.0 CzH2,4.8 0.07t Fe . . . . . . . . 248.3 0.2 26 6 N20,4.0 C2H2,4.6 0.37 Mg . . . . . . . . 285.2 0.7 6 6 N20,4.6 C2H2,4.6 0.03 K . . . . . . . . 766.5 2 10 7 Air, 3.6 Hz, 12.2 0.04 Na . . .. . . . . 589.0 0.7 8 4 Air, 4.1 C2H2,2.3 0.01 4 Air, 5.4 CzH2,2.6 0.02.t * Plus an additional maximum flow of 5.3 1 min-1 (outside the flow meter; not measurable) for nebulisation of the solution. t In the presence of caesium. Table 2. Matrix and ionisation-suppression effects in the fluoroboric acid solution under the operating conditions given in Table 1. Expressed as percentage changes in signal height Interfering element Element Ba . . . . . . . . Ca . . . . . . . . Fe . . . . . . . . Mg . . . . . . . . K . . . . . . . . K . . . . . . . . Na . . . . . . . . * Air - hydrogen flame. t Air - acetylene flame. Concentration/ pg ml-1 4 0.8 4 0.4 0.4* 0.4t 0.4 Nb as Nb205 (2% m/V) Cs as CsCl (0.1% m/V) Without Cs With Cs -58 - 47 - 58 s 2 s 2 <2 <2 - 23 - 22 - 40 -13 -6 -7 Without Nb + 380 + 43 <2 <2 +2 +53 <5 With Nb + 130 + 13 <2 <2 +2 +116 <5 random errors, were studied to evaluate the recovery of the method.Known amounts of the elements were introduced into the vessels before the dissolution step or just before diluting the final fluoroboric acid solution, i.e., when the formation of fluorides became impossible. The solutions containing 4 pg of Ba, 2 pg each of K, Ca, Fe, 0.4 pg of Na and 0.2 pg of Mg per ml were compared by measuring their absorbance under the same conditions. The results of these measurements for three determinations agreed to within the error of measurement. Ionisation and Matrix Interferences The presence of elements that are easily ionised in flames can lead to erroneous results.In order to predict the possible interferences and avoid such errors, the action of 0.1% CsCl as a radiation buffer in fluoroboric acid solutions of the blank and sample was studied. Synthetic sample and blank solutions were prepared by a bomb technique in the same manner as the real sample solutions (two samples of each solution). A standard solution containing the elements to be determined was prepared and added to give final concentrations of: Ba and Fe, 4 pg ml-1; Ca, 0.8 pg ml-1; and Na, K and Mg, 0.4 yg ml-1 in solutions with and without caesium. The instrument was zeroed with the appropriate solution, i.e., one not containing the element under study. The flame operating conditions were varied to find an optimum corre- sponding to the maximum sensitivity (Table 1).The net signal responses of the test element, in the presence or absence of niobium and the radiation buffer, were measured to deter- mine the matrix and ionisation interferences (Table 2 ) . The results in Table 2 indicate that the ionisation effects cannot be excluded. The ionisation buffer need not be added for the determination of sodium in an air - acetylene flame, potassium in an air-hydrogen flame or magnesium and iron in a dinitrogen oxide - acetylene flame. It is required, however, in the determination of potassium in an air - acetylene flame and barium and calcium in a dinitrogen oxide - acetylene flame because of the significant ionisation that takes place. The niobium matrix depresses the optimum signal value in all instances except magnesium and iron in a dinitrogen oxide-acetylene flame: the signals of these elements in an air - acetylene flame are reduced by 83 and 1370, respectively. Depression of the sodium signal is small (i.e., is within the error limits). The matrix effects depend on the concentration of the niobium matrix; for example, the signal depression for potassium in an air - hydrogen flame decreases from 23 to 19% when the niobium and (equivalent) fluoroboric acid concen- trations are halved.Therefore, two separate sets of standards, for the blank and for the sample (containing niobium), must be prepared in order to overcome the matrix interference effects if the calibration graph method is to be used. Determination of Ba, Ca, Fe, K, Mg and Na All samples were prepared by using the technique described.Two sets of standard solutions were made, one of them containing amounts of niobium corresponding to 2% mlV of Nb205. All solutions included 0.4% V / V H N 0 3 , 4% V/V HF, 2%V/V H 2 0 2 and 2% m/V H3B03. The element concentra- tion ranges were 0-0.8 yg ml-1 for K, Mg and Na, 0-8 yg ml-1 for Fe and Ba, and 0-1.6 yg ml-1 for Ca. Calibration graphs were used to determine trace amounts of the elements in the blank and sample solutions of high-purity niobium pentoxide. The Fe, K, Mg and Na were determined without the ionisation buffer (K in an air - hydrogen flame) owing to the lack of CsCl of sufficient purity; Ba and Ca were determined in the presence of caesium by introducing 250 pl of a 2% CsCl solution to 5-ml portions of the sample, blank and standard solutions, respectively.The following measurement procedure was employed to avoid errors due to contamination of the standard materials.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, OCTOBER 1987. VOL. 2 73 1 Table 3. Determined metal concentrations in high purity Nb,O,. Results given as the average of seven determinations, with corre- sponding blank values given in parentheses Mean Standard Element vgg-‘ pg g-1 interval/pg g-1 concentration1 deviation1 95% Confidence Ba . . . . . . <10 Signal height at noise level Ca . . . . . . 5.7(2.0) 1.5 5.7 Ifr 1.4 Fe . . . . . . 177(<2) 3.8 177 k 3 K” , , . . . . 4.2(1.3) 0.25 4.2 k 0.2 Mg , . . . . . O.S(O.5) 0.04 0.5 k 0.04 Na . . . . . . 1.3 (5.1) 0.33 1.3 Ifr 0.3 * Air - hydrogen flame. The instrument was zeroed with the appropriate standard solution, i.e., one not containing the element under study, and calibration graphs for the blank and the sample were prepared. Blank and sample signals from real sample solutions were measured when the instrument was zeroed with water.In spite of insignificant background absorbance background correction was performed in all instances because positive errors due to the uncompensated background (€10 pg g-1 of Ba, <1 yg g-1 of Ca, 10 pg g-1 of Fe, 0.4 pg 8-1 of K, <2 pg g-1 of Mg and 2 pg g-1 of Na) were comparable to the contents of the elements to be determined in the sample. Concentra- tions of the elements being determined were read from the graphs and the concentration of the blank was subtracted from the concentration of the sample; results are given in Table 3.The calibration graphs were linear in the concentration ranges studied. Testing of the Method The results were verified by the method of standard additions in order to reveal the inter-element interferences from other, not investigated, elements. Background correction was used throughout. Results from the calibration graph and standard additions methods agreed for high-purity materials. However, for one of the samples studied the influence of the other elements was visible. The intense yellow solution of that sample suggested the presence of elements forming peroxide complexes; about 0.7% of Ti was determined by spark source mass spectrometry (SSMS) . Conclusion Binding the fluoride ion with boric acid allows the use of quartz glassware and glass AAS spray chambers if short contact times are maintained? However, fluoroboric acid was not found to provide an interference-free environment to such an extent as could be expected from studies of siliceous materials.26927 Ionisation of magnesium and iron atoms in a dinitrogen oxide - acetylene flame, sodium in an air - acetylene flame and potassium in an air - hydrogen flame does not occur, but ionisation effects may be significant in the determination of potassium in an air - acetylene flame or barium and calcium in a dinitrogen oxide - acetylene flame.These examples require the application of a radiation buffer. The niobium matrix depresses the signals from all elements except magnesium and iron. This reinforces the need to prepare two separate sets of standards for the analysis of the blank and sample (with niobium), in both the calibration graph method and the standard additions method.28 Using two separate sets of standards for the blank and the sample is not a usual practice in the calibration graph method, as opposed to the standard additions method, but it avoids laborious and troublesome matrix separation and prevents the increase of blanks, especially for common elements.The fact that no cations except caesium were introduced by the reagents contributes to the signal stability of the atomic absorption measurement and to a decrease in instrumental background readings.26 Although certified samples have not been analysed, the method was tested for systematic error from the sample pre-treatment and analytical procedure, and was found to be reliable for high-purity materials because of the recovery and agreement between results obtained from the calibration graph and standard additions methods.Recovery may be accepted as a measure of the reliability of the method on condition that the decomposition of the sample is complete and that the added analyte element occurs in the same form, i.e., behaves in an analytically identical manner with the element present in the sample. This seems to be the case here. If significant amounts of foreign ions are present in the solution the calibration graph method fails and application of the standard additions method is necessary. Background correction is required. 1. 2. 3. 4. 5 . 6. 7. 8. 9.10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. References Bilimovich, G. N., Alimarin, I. P . , and Tikhonova, T. V., Zh. Anal. Khim., 1971, 26, 122. Shemanenkova, G. I . , Frisov, V. I . , Shechelkova, V. P., and Shchulepnikov, M. N . , Zh. Anal. Khim., 1973, 28, 323. Andreev, A. V., and Golubchikov, V. V., Zh. Anal. Khim., 1975, 30, 2150. Ziegler, J., J. Vac. Sci. Technol., 1973, 10, 151. Abe, S . , and Takano, T., J . Radioanal. Chem., 1978,46,229. Skotnikov, S. A., and Lazareva, I. Yu., Zavod Lab., 1970,36, 420. Moroshkina, T. M., Kovrizhnykh, V. M., and Mel’nikov, Yu. A . , Zh. Anal. Khim., 1971, 26, 1352. Antonov, A. V., Mitropol’skaya, N. A., Odintsova, I. N., and Shtenke, A. A., Nauchn. Tr. Nauchno-fssled. Projekt. fnst. Redkomet.Prom-sti, 1977, 82, 105. Varavko, T. N., Vladimirskaya, I. N . , and Kaplan, B. Ya., Zavod Lab., 1976, 42, 659. Kurbatov, D. I . , and Il’kova, S. B., Zh. Anal. Khim., 1974,29, 1430. Dolmanova, I. F., Zolotova, G. A.. Voronina, R. D., and Peshkova, V. M., fnd. Lab., 1973, 39, 386. Il’kova, S. V., Zarimko, E . A . , and Rivkina, K. K., Zavod Lab., 1976, 42, 658. Goryanskaya, G. P., and Merisov, Yu. I., Zavod Lab., 1974, 40, 934. Elinson, S . Vi., Nevzorov, A. N., Belogortseva, M. V., Mirzoyan, N. A., and Mordvinova, S. N., Zh. Anal. Khim., 1974, 29, 1234. Malyutina, T. M., Orlova, V. A , , and Spivakov, B. Ya., Zh. Anal. Khim., 1974, 29, 790. Antonovich, V. P., Shelikhina, E. I., Serbinovich, V. V., and Chukhrii, Yu. P., Zh. Anal. Khim., 1983, 38, 1808. Gibalo, I. M.. “Analiticheskaya Khimia Elementov. Niobii i Tantal,” Nauka, Moscow, 1967. Pletneva, T. I., Dorfeev, V. S . , Chupakhin, M. S . , Vol’nya- gina, A . N., and Mikhailov, N. S . , Khim. Prom-st, Ser.: React. Osobo Chist. Veshchestva, 1979, 3, 26. Ken, M., Kiyonori, H., and Kikuo, T., Fresenius Z . Anal. Chem., 1982, 313, 562. Hashiba, M., Miura, E . , Nurishi, Y., and Hibino, T., Bunseki Kagaku, 1980, 29, 323. Mulder, B. J . , Anal. Chim. Acta, 1974, 72, 220. Vasnev, A . N., Kreingol’d, S . U., Zherebovich, A. S . , and Chupakhin, M. S . , Zavod Lab., 1976.42, 656. Kotz, L., Kaiser, G., Tschoepel, P., and Toelg, G., FreseniuJ Z . Anal. Chem., 1972, 260, 207. Adachi, T . , Bull. Chem. SOC. Jpn, 1982, 55, 1824. Danzaki, Y., and Takeyama, S . , Bunseki Kagaku, 1983,32, T 57. Bernas, B., Anal. Chem., 1968, 40, 1682. Price, W. J . , and Whiteside, P. J . , Analyst, 1977, 102, 664. Cammann, K., Fresenius 2. Anal. Chem., 1982, 312, 515. Paper J7f57 Received October 13th, 1986 Accepted May 12th, 1987

ISSN:0267-9477    

DOI:10.1039/JA9870200729

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, OCTOBER 1987, VOL. 2 733 Determination of Yttrium in Zirconia Matrices by Atomic Absorption Spectrometry Azzeddine Samdi and Jacques Piiris Laboratoire de Chimie Minerale Ill, U.A. 7 76, Universite Claude Bernard L yon I, 43 boulevard du 1 1 Novembre 19 18,69622 Villeurbanne Cedex, France Jean-Pierre Deloume and Gerard Duc Laboratoire de Chimie Analytique I!, Universite Claude Bernard Lyon I, 43 boulevard du I 7 Novembre 7978, 69622 Villeurbanne Cedex, France A rapid method is described for determining yttrium in solutions containing a large excess of zirconium using atomic absorption spectrometry. Samples are analysed in nitric acid solution, using a spectrochemical buffer which consists of potassium - EDTA and lanthanum to effect maximum absorption.A standard additions technique, in which yttrium solution is added t o the sample, is used t o avoid the preparation of matrix matched standards. The yttrium absorbance is measured at 410.2 nm using a dinitrogen oxide - acetylene flame. Synthetic solutions show that the procedure is sufficiently accurate to be used for the routine determination of yttrium in refractory oxides containing up to 3 mol-O/O of Y203 (5.7% mlm). Keywords: Atomic absorption spectrometry; yttrium determination; zirconium matrix; refractory materials; spectrochemical buffer Zirconia (Zr02) can exist as one of three phases depending on the temperature : monoclinic from ambient temperature to 1170°C, tetragonal up to 2370°C and cubic at higher temperatures up to melting point.1 In order to obtain desirable thermo-mechanical properties the tetragonal phase needs to be retained at the operating temperatures.For this purpose, metallic oxides are added to zirconia to form a partially stabilised zirconia (PSZ); when yttria (Y203) is added as a stabilising agent, the best results are obtained between 2.5 and 3 mol-% of Y2O3.2 We have developed a method for the preparation of PSZ, by the precipitation of mixed salts from aqueous solutions.3~4 As the solubilities of such salts are unknown, it has been necessary to determine yttrium in the precipitates in order to optimise the experimental conditions. In this paper a rapid method developed for the determination of yttrium in the range 0-3 mol-% of Y203 in such samples is described.Few papers have been published on the determination of yttrium by atomic absorption spectrometry (AAS). Wise and Solsky5 have described a method for the determination of yttrium in oxides where Y is present as the major constituent. They reported that a spectrochemical buffer containing potassium - EDTA was useful, but did not give information on either the enhancement or the sensitivity achieved. The main difficulty encountered by these workers was the dissolution of the oxide samples which was achieved by alkaline fusion or other such drastic treatment. We did not encounter this problem in this work as the salts we prepared were readily soluble in nitric acid. Initially we used the spectrochemical buffer as proposed by Wise and Solsky. It rapidly became apparent that zirconium did not enhance the yttrium absorp- tion signal as they had suggested, but depressed it.Therefore we systematically investigated the addition of releasing agents to the spectrochemical buffer to improve both the reliability and accuracy of the determination of yttrium by AAS. Experimental Apparatus A Perkin-Elmer 303 flame spectrometer was employed using a slightly reducing N 2 0 - C2H2 flame with a red cone height of 1.5 cm. The other instrumental parameters are shown in Table 1. Tabie 1. Instrumental parameters Wavelength . . . . . . 410.2 nm Lampcurrent . . . . 10mA Spectral band width . . 0.2 nm N2Oflow . . . . . . 6 1 min-1 C2H2flow . . . . . . 12lmin-1 Reagents All solutions were prepared with analytical-reagent grade chemicals (Prolabo) and distilled water unless indicated otherwise.Potassium - EDTA buffer solution. Following the procedure suggested by Wise and Solsky5 the buffer solution was prepared by adding 36.6 g of EDTA (free acid form), 300 ml of water and 10.7 ml of 45% KOH solution (sp. gr. 1.46) to a 500-ml calibrated flask. The solution was stirred, and am- monia solution was added slowly until all of the EDTA dissolved. The solution was then cooled to room temperature and diluted to volume (0.25 mol 1-1 of K + 0.25 mol 1-1 of EDTA). Yttrium stock solution. A 1500 mg 1-1 solution was prepared by dissolving 0.9524 g of Y2O3 (Riedel de Haen 99%) in 25 ml of hot concentrated HN03, which was then cooled and diluted to 500 ml.5 Zirconium stock solution. A 15000 mg 1-1 solution was prepared by dissolving 26.49 g of ZrOC12.8H20 (Riedel de Haen, analytical-reagent grade) in 25 ml of hot concentrated HN03, which was then cooled and diluted to 500 ml.Lanthanum stock solution. A 5000 mg 1-1 solution was prepared by dissolving 127.2 g of LaC13 in 1000 ml of acidified water (50 ml l-1 of concentrated HN03). Results and Discussion Optimisation of Sensitivity for Yttrium The influence of selected reagents on the yttrium atomic absorption signal was investigated for yttrium solutions having concentrations in the range 0-80 mg 1-1. The results are summarised in Fig. 1. The sensitivity of the dilute stock solution of yttrium is 1.12 x 10-3 1 mg-1 (line A). The presence of lanthanum ions (1600 mg I-l), commoniy used as a buffer, increases the sensitivity to 1.32 X 10-3 1 mg-l (18%JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, OCTOBER 1987.VOL. 2 734 0.2 a, C m + n" 0.1 a 0 20 40 60 80 Yttrium concentrationhng I-' Fig. 1. Influence of the composition of the yttrium solution on the AAS signal: A, nitric acid; B, nitric acid + La3+; C, nitric acid + HC104; D, nitric acid + potassium - EDTA buffer; and E, nitric acid + La3+ + HC104 + potassium-EDTA buffer enhancement, line B). He et a1.6 in a general study on the dinitrogen oxide - acetylene flame reported that perchloric acid increases the absorbance of yttrium. As Wise and Solsky also used perchloric acid to improve the solubility of the oxides, we added a few drops of perchloric acid to the dilute stock solutions, but this had very little effect on the signal (line C).However, the potassium - EDTA buffer (with a concentra- tion of 0.01 mol 1-I) resulted in a significant increase in sensitivity to 1.84 x 10-3 1 mg-1 (line D), i.e., 1.6 times greater than the nitric acid solution. Although this increase in sensitivity appears to be beneficial, it is still not sufficient to counterbalance the depressive effect of the zirconium (see below). It is suggested that this buffer causes yttrium to complex with EDTA, so reducing the concentration of hydroxo complexes in the solution, and thus the formation of oxide in the flame. At the same time, potassium acts as an ionisation suppressor, the .net effect being to increase the yttrium absorbance signal. As lanthanum has an enhancing effect on the atomisation of yttrium in nitric acid solution, it should result in the same behaviour when used jointly with the K-EDTA buffer.This hypothesis was tested by preparing solutions of yttrium containing a mixture of the EDTA buffer, lanthanum ions and a few drops of perchloric acid. The corresponding yttrium absorbance signal was then measured. The resulting sensitivity (line E) is much higher than in the previous experiments. Indeed the value of 2.5 X 10-3 1 mg-1 is higher than the product obtained by multiplying the aqueous yttrium sensitivity figure (1.2 x 10-3) by the product of the individual enhancement factors (1.2 x 10-3) x 1.18 x 1.6 = 2.1 x 10-3 1 mg-1. In this way we selected the composition of spectrochemical buffer suitable for optimum determination of yttrium by AAS. Solutions for analysis must be prepared as follows: the sample solution is introduced into a 100-ml calibrated flask, followed by rinsing water; then 3.2 ml of a 50000 mg 1-1 solution of La3+ and 4 ml of the K-EDTA buffer (0.25 M) are added followed by rinsing water.A known amount of yttrium is then added along with 6-8 drops of perchloric acid, methyl orange and sufficient concentrated ammonia solution to produce the methyl-orange end-point. The mixture is then diluted to the mark. The order of introduction of the buffer components and the rinses is important to avoid the formation of precipitates by contact between solutions of different pHs. Solutions prepared this way remain stable for up to one month. Influence of the Zirconium Matrix The influence of the zirconium has been studied in the presence of the spectrochemical buffer, by preparing four 0.2 a, K m e n" 0.1 a I D I I C B I I I I I I 0 600 800 1500 2000 Zirconium concentration!mg I - Fig.2. Yttrium concentration: A, 20; B, 40; C, 60; and D, 80mgl-1 Depressive effect of zirconium on yttrium AAS signal. Table 2. Influence of zirconium on the sensitivity of the yttrium AAS signal Zirconium concentration/ Sensitivity/ mgl-1 , . . . . . 0 600 800 1500 2000 Img-lofyttrium . . 2.8 2.2 2.1 1.8 1.5 series of yttrium solutions with the concentrations 20, 40, 60 and 80 mg 1-1. For each yttrium concentration, the zirconium concentration was varied from 0 to 2000 mg 1-1 (0, 600, 800, 1500 and 2000 mg 1-1). The results are given in Fig. 2. They indicate that at the highest concentrations, zirconium causes a near-linear reduction in the yttrium absorption signal.In the less concentrated solutions, the depressive effect deviates from linearity, especially between 0 and 800 mg 1-1. This result demonstrates that the preparation of matrix matched standards is not possible if the zirconium concentration is unknown. The relative difference in the zirconium concentra- tions of an unknown solution and the standards can be lowered with standard solutions containing 2000 mg 1-1 of zirconium, by a systematic addition of zirconium to the unknown to make it 2000mg1-1 more concentrated than it was. In this range of zirconium concentrations the sensitivity of the yttrium absorbance signal is the smallest (Table 2). Determination of Yttrium Based on these results, it is possible to devise a method for the determination of yttrium in zirconia matrices. As the composi- tion of the mixed precipitate is not known, it is not possible to prepare matrix matched standards to standardise the absor- bances.However, it is apparent from the results presented above that a standard additions technique can be used because the Beer - Lambert law is valid for a constant zirconium concentration. Standardised amounts of the yttrium stock solution were added to the sample solution to prepare series of solutions containing 10, 20, 30 and 40 mg 1-1 more yttrium than the original solution, providing the total concentration remains below 80 mg 1-1. If the results show that the 80 mg 1-1 limit has been exceeded, simple dilution of the mother liquor may be performed. The feasibility of this technique was first tested for yttrium alone. As the concentration of yttrium used was known, as it is prepared from Y203, we prepared a series of solutions with varying amounts of yttrium which were analysed by theJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, OCTOBER 1987, VOL. 2 735 Table 3.Determination of yttrium by AAS Yttrium concentration/mg 1-1 Theoretical 8.4 16.9 16.9 8.4 9.8 19.6 20.0 Found 8.8 16.2 16.4 8.4 9.7 19.1 19.4 Table 4. Determination of yttrium by AAS in the presence of zirconium Theoretical Theoretical Yttrium concentration concentration/ concentration, found/ mg 1-1 mgl- mol-% mgl-1 676 10.0 0.7.5 9.9 676 20.0 1.50 20.1 218 18.7 4.20 19.2 303 16.9 2.78 16.8 720 8.0 0.57 8.1 Zirconium yttrium y 2 0 3 concentration proposed method.The values obtained are given in Table 3. The results are in good agreement with the known concentra- tions. Finally we tested this technique on solutions containing an excess of zirconium. Test solutions were prepared using the method described above. The results are given in Table 4 and show that the method gives good agreement over a range of zirconium matrix concentrations. The sensitivity remains at an acceptable level of 1.8 x 10-3 1 mg-1 to 1.6 x 10-3 1 mg-1 according to the zirconium concentration (for the solutions without zirconium, the mean sensitivity was 2.1 X 10-3 1 mg-1). These solutions match the sample solutions prepared from the mixed precipitate (containing the same proportions of yttrium and zirconium and the same components) which is readily soluble in nitric acid.Results obtained in further experiments3 correlate well with the X-ray powder diffraction patterns of pyrolysed precipitates in the Y203 - Zr02 phase diagram. Conclusion This investigation indicates that yttrium can be determined in solutions containing a large excess of zirconium. The proce- dure requires neither a knowledge of the zirconium concentra- tion, nor any separation stages. Optimum analytical condi- tions occur when the yttrium concentration lies between 8 and 20 mg 1-1 and the zirconium concentration is below 800 mg 1-1. References 1. Subbaro, E. C., in Heuer, A. H . , and Hobbs, L. W., Editors, “Science and Technology of Zirconia,” Advances in Ceramics, Volume 3 , American Ceramic Society, Columbus, OH, 1981, Lange, F. F., J . Mater. Sci., 1982, 17, 22.5. Samdi, A . , Piiris, J., and Grollier Baron, T., unpublished results. Van de Graaf, M. A. C. G., and Burgraaf, A. J . , in Claussen, N., Ruhle, M., and Heuer, A. H., Editors, “Science and Technology of Zirconia 11,” Advances in Ceramics, Volume 12, American Ceramic Society, Columbus, OH, 1984, p. 744. Wise, W. M., and Solsky, S. D., Anal. Lett., 1976, 9, 1047. He, Z.-D., Zhang, L.-D., Yu, Y.-G., Qiu, S.-Y., Liu, H.-S., and Zhu, P.-S., Fenxi Huaxue, 1980, 8, 88. p. 1. 2. 3. 4. 5. 6. Paper J6195 Received October 7th, 1986 Accepted May 5th, 1987

ISSN:0267-9477    

DOI:10.1039/JA9870200733

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

736 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, OCTOBER 1987, VOL. 2 ERRATUM Atomic Absorption Spectrometric Determination of the Elemental Contamination of Plant Material Massimo Ottaviani and Paola Magnatti J. Anal. At. Spectrom., 1986, 1, 243 The name of the second author should be as given above (Paola Magnatti) and not as given previously (Paulo Magnatti).

ISSN:0267-9477    

DOI:10.1039/JA9870200736

出版商:RSC    

年代:1987

数据来源: RSC