摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 735 Electrothermal Atomic Absorption Spectrometric Determination of Antimony in Metal Chloride Matrices Using Probe and Tube-wall Atomization Paavo Peramaki and Lauri H. J. Lajunen Department of Chemistry University of Oulu S F-90570 Oulu Finland Two atomization modes were used in the determination of antimony by electrothermal atomic absorption spectrometry namely tube-wall and probe atomization. The reciprocal sensitivity of the tube-wall atomization was approximately twice that of probe atomization when the peak height measurement was used. The interfering effects of different metal chlorides on the determination of antimony were investigated using both atomization modes. The results showed that probe atomization is useful when antimony is determined in samples containing alkali and alkaline earth metal chlorides (especially MgCI and CaCI,).Severe background over-compensation interferences were noticed with the deuterium background correction system used when antimony was determined in matrices containing iron and cobalt. To a lesser extent over-compensation interference was also noticed with nickel. The over-compensation interferences were avoided when 30 pg of magnesium as chemical modifier and a slit-width of 0.2 nm were used in the determinations. Keywords Atomic absorption spectrometry; probe atomization; antimony determination; chloride interference; background correction Non-isothermal conditions in the atomizer during the atomization step can give rise to interferences when for example antimony is determined in a complex sample matrix by electrothermal atomic absorption spectrometry (ETAAS).Atoms can re-form into molecules in the gas- phase during the atomization especially in samples con- taining large amounts of halides. In the normal tube-wall atomization the sample is vaporized at relatively low temperatures which often results in incomplete atomiza- tion. Interatomic interferences in the gas phase can be minimized by the use of the L'vov platform technique together with chemical modification for further delay of the atomization step.' In principle the use of a constant- temperature furnace would give the best conditions for isothermal at~mization.~*~ In order to provide sufficiently high atomization temper- atures and less gas-phase interferences in ETAAS a commercial probe atomization system has been introduced re~ently.~?~ In this system the graphite cuvette alone is first heated to the atomization temperature.When the cuvette temperature has stabilized the graphite probe (and an ashed sample on it) is automatically inserted into the constant-temperature environment for isothermal atomiza- tion. The first applications of the probe atomization technique in AAS included the dispensing of the sample on a tungsten filament or alternatively on a tungsten wire mounted on the autosampler arm.6*7 After the samples had been dried (and charred) the devices were introduced into the graphite tube which was pre-heated to the atomization temperature. Further a modifed graphite tube (T tube) was constructed and the tungsten wire was replaced with a graphite probe that is more resistant at higher temperatures.The system has been applied even to the determination of refractory element^.^^^ Generally the graphite probe is introduced into the graphite tube through a slot cut in the side of the tube but the application of end-entry probe insertion has also been studied.loJ1 In earlier probe measurements the probe was inserted manually into the graphite tube. However for routine work the automatic system is most ~onvenient.'~J~ Although the sample is introduced on the probe usually as a liquid droplet solid sampling is also po~sible.'~J~ The aim of this work was to compare the tube-wall and probe atomization techniques when antimony is deter- mined in different chloride matrices by ETAAS.Antimony is among the elements for which gas-phase interferences can be expected to occur.16 Although a comparison with platform atomization would have been more interesting platform tubes are unfortunately not available for the instrument used. Experimental Instrumentation The measuring system consisted of a Pye Unicam PU 9200 atomic absorption spectrometer equipped with a PU 9390 electrothermal atomizer and a PU 9380 furnace autosam- pler. A PU 9385 Autoprobe was mounted on the graphite furnace system. The details of the probe atomization system have been described el~ewhere.~.~ An antimony hollow cathode lamp (Cathodeon Cambridge UK) was used as a light source and background correction was made with a background corrector equipped with a deuterium hollow cathode lamp.The results were printed out on an Epson FX-800 printer. The measuring wavelength for antimony was 2 17.6 nm and the operating current of the hollow cathode lamp was 9.0 mA (the maximum current recommended by the manufacturer is 15 mA but a lower current was used to increase the lifetime of the lamp). Slit-widths of both 0.2 and 0.5 nm were employed in the measurements and a sample volume of 20 mm3 was dispensed into the graphite furnace by the furnace autosampler. In addition for some of the probe measurements 5 mm3 of chemical modifier solution were added into the graphite furnace. The temper- ature programmes for antimony are given in Table 1. All tube temperatures used in the experiments are set values on the instrument.Tube-wall and probe atomization were used. In tube-wall atomization uncoated graphite cuvettes made from normal graphite ('electrographite') were used as recommended by the manufact~rer.~~ In probe atomization electrographite probes and totally pyrolytic graphite probe cuvettes were used. The change in the atomization mode also requires a change in the graphite contacts holding the graphite cuvette. The special furnace alignment jigs were inserted in place of both types of graphite tubes to ensure the correct optical and graphite furnace alignment before the measure- ments.I8 Samples were digested using an MDS-8 1 D microwave oven (CEM Matthews NC USA) equipped with Teflon736 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 ~~ Table 1 Furnace programmes for the determination of antimony Atomization mode Phase Probe Dry Dry Ash§ Atomize5 Clean Cool Internal gas flow rate/ Temperature/"C* Ramp/"C s-' Hold/s cm3 min-'? Commandst 200 25 10 200 - 250 1 10 300 - 200 - 800 (1000) 50 5 (15) - 3 0 TC RD - 60 300 - 2700 (2550) 2700 - 3 300 TC 20 Tube wall Dry 105 10 Ash 1000 50 Atomize 2200 - 2800 - Clean 10 5 3 3 200 - 200 - 300 TC 0 TC RD *Temperature values programmed on the instrument.?Argon (99.99%) was used as an inert purge gas. STC = temperature control; RD = read. §Values in parentheses were used when reference materials were analysed and 30 pg of magnesium modifier were employed PFA (perfluoroalkoxy) digestion vessels of 120 cm3 capacity and a removable 12-position digestion turntable.The full microwave power of the instrument is 630 k 70 W. Scanning electron micrographs were obtained with a Jeol JSM-6400 scanning electron microscope (SEM). Procedures Analytical-reagent grade reagents and de-ionized distilled water were used throughout. All reagents were obtained from Merck (Darmstadt Germany). The test solutions containing antimony and the possible foreign ions were prepared in 50 cm3 calibrated flasks. Blank solutions were also prepared in order to detect any contamination by antimony of the reagents used. The acid matrix in the test solutions was 0.1 mol dm-3 nitric acid unless stated otherwise. Dilute solutions were measured on the same day as they were prepared. During the measurements at least every third site in the sample tray was occupied by an antimony standard solution in order to detect the drift of the graphite components and hence changes in analytical sensitivity. A geological reference material (0.2-0.25 g) was accu- rately weighed into a Teflon PFA digestion vessel.Diges- tion acids were added and a pressure-relief valve was placed on the top of the digestion vessel. The cap was placed on the vessel and first turned until it was finger-tight then further tightened using a capping station (CEM). Digestion vessels were placed on a turntable which was rotated at about 6 rev rnin-' to ensure uniform microwave heating of each sample. In the first method (HCl-HN03) 3 cm3 of concentrated hydrochloric acid and 1 cm3 of concentrated nitric acid were added to the digestion vessels and the following power settings were used for geological reference samples 20% for 2 min 50% for 10 min and 100% for 3 min In the other ; 0.20 8 0.05 400 1200 2000 2800 Ternperature/"C Fig.1 Ash-atomize plots for antimony A tube-wall atomization 0.5 ng of antimony; and B probe atomization 1 .O ng of antimony method (HF-HN03) 3 cm3 of hydrofluoric acid (40%) and 3 cm3 of concentrated nitric acid were employed and the following power settings were used 25% for 3 min 40% for 3 min 60% for 2 min 25% for 10 min and 75% for 10 min. When the digestion was completed the solutions were transferred into calibrated flasks (glass or plastic) and if necessary further diluted. Undissolved residue was allowed to settle before an aliquot was taken for measurement. Neither of the two methods described above was capable of dissolving the whole sample matrix.Total dissociation of geological material was achieved by lithium tetraborate fusion in a platinum crucible (20 min at 1000°C) and dissolution of the melt in dilute hydrochloric acid.19 Results and Discussion Ash-Atomize Plots and Sensitivity In principle the gas-phase temperature during the atomiza- tion step should be as high as possible to achieve complete atomization of the analyte. Generally higher atomization temperatures must be programmed on the instrument when probe atomization is used because the graphite probe is heated by convection and radiation. The use of the two-line absorption method for evaluating the gas-phase tempera- tures experienced by the analyte atoms suggested that only when the tube-wall temperature settings were higher than 2200°C were the gas-phase temperatures higher in probe atomization than in tube-wall atomization.20 The ash-atomize plots for the both atomization systems are shown in Fig.1. An atomization temperature of 2700 "C was selected for the probe work in order to minimize possible interference effects although improved integrated absorbance sensitivites were obtained for atomization temperatures below 2700 "C. Generally atomization temperatures higher than 2700 "C were not employed during the measurements because otherwise the thin walls of the probe cuvettes were gradually deformed (especially from the ends of the tube; the average lifetime of the probe cuvettes was found to be about 400 atomizations). When the graphite probe is not cooled by convection it remains hot for a long time after atomization. Therefore a cooling time of 60 s with the maximum gas flow was included in the furnace programme in addition to a 20 s cooling time programmed in the furnace software by the manufacturer (the probe stays inside the outer tube during the cooling period).The maximum ashing temperature for antimony with the probe system (800 "C without chemical modification) was found to be lower than that in tube-wall atomization. TheJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 737 reason for this is not known. It is possible that antimony molecules condense on the cooler tube parts during the ashing step and contribute to the signal during wall atomization. For probe atomization these condensed spe- cies would not contribute to the signal as recording of the signal commences after the probe cuvette has been pre- heated to the atomization temperature.The spreading of the sample droplet along the probe may be a problem if higher nitric acid concentrations are used owing to the lower surface tension of the sample. Therefore Carroll et decided to use a nitric acid concentration of 0.05 or 0.5% v/v and a sample volume of 10 mm3 in probe work. However in their system there was no ridge in the probe to prevent the spreading of the sample. The earlier results with the ridge probe suggest that the reproducibility should be better than 2% relative standard deviation (RSD) even up to a concentration of 20% v/v of nitric acid for a 10 mm3 ample.^ In this work it was found that both the probe surface structure (and hence its age) and the nitric acid concentra- tion of the sample had a considerable effect on the performance of the system.Occasionally an unwanted situation occurred as the sample droplet spread over the ridge of the probe out of the graphite tube which resulted in poor recovery and reproducibility. Studies by scanning electron microscopy showed that the surface of the graphite probe became gradually more porous during use owing to sublimation of smaller graphite particles (Fig. 2).22 An old probe with a roughened surface behaved differently to a new probe when a sample was injected on to it. The spreading of a sample droplet on a hot probe surface was not a problem when the sample was in a dilute nitric acid matrix (0.1 mol dm-3) and an old probe (over 200 atomizations) was used.The improved performance was seen in the complete recovery of antimony although no additional cooling time was included in the furnace Fig. 2 Surface of the graphite probe 250 times magnification. (a) A brand new probe; and (b) after 310 atomizations at 2700°C (samples in dilute nitric acid) Table 2 Recovery of 1 ng of antimony in the presence of 50 pg of iron(II1) chloride. A slit-width of 0.2 nm and an argon flow rate of 200 cm3 min-' during the cooling period were used. The signal for antimony alone at each cooling time is 100. Probe age 3-50 atomizations Relative absorbance* Cooling time/s Peak height Integrated absorbance 10 46 * 2.1 49 k 2.8 15 62 +_ 3.7 59 k 5.0 20 73 f 2.9 69 1- 3.7 30 93 f 3.6 73 +_ 6.7 45 96+ 1.9 74 1- 5.9 *Average of four measurements f SD.programme before the introduction of the next sample. A short 10 s cooling time was needed when a less aged probe (about 70 atomizations) was used. However when the concentration of nitric acid was increased the behaviour of the sample was more dependent on the temperature of the probe. Therefore cooling times of about 30 and 60 s (in 0.5 and 1.0 mol dm- nitric acid matrices respectively) were needed in order to obtain a complete recovery of antimony (about 100 atomizations were made with this probe). The effect of a short cooling time was also clearly demonstrated when measurements were made from matrices containing iron(II1) chloride (Table 2). The low recoveries obtained even with long cooling times especially with integrated absorbance measurements are due to over-compensation with the deuterium background correction system used.This is discussed later. All results described above were obtained using a sample volume of 20mm3. If a smaller sample volume is used fewer difficulties are likely to occur. The sensitivity of tube-wall atomization was roughly twice that of probe atomization. The characteristic mass (1% absorption peak height) for tube-wall atomization was 9.5 pg and for probe atomization 24 pg. Generally when determinations were made using dilute nitric acid solutions and the probe became more aged the sensitivity for antimony was slightly improved (with both peak height and integrated absorbance measurements).This was probably partly due to less spreading of the sample on a roughened probe surface. Because the probes are made from electro- graphite they are not as resistant as pyrolytic graphite coated graphite parts. Hence the work with corrosive matrices (e.g. solutions containing FeCl and HNOJ caused rapid wearing of the thin probe head which resulted in poorer analytical sensitivity. The reason for the poorer sensitivity obtained with the probe system is probably the loss of atoms by diffusion through the slot cut in the side of the probe ~ u v e t t e . ' ~ ~ ~ ~ 0.16 A I 3 0.14 8 0 5 0.12 ' % 2 0.10 ' 0 2 0.08 ol & 0.06 8 4- U 0.04 I I I I 0 400 800 1200 1600 2000 Tern peratu rePC Fig. 3 Ashing plot for 1 ng of antimony with tube-wall atomiza- tion.A Antimony alone; and B antimony together with I00 pg of CaC1,738 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 The higher atomization temperature used in probe atomiza- tion decreases the residence time which in part is a reason for a lower signal. Also the larger inside diameter of the probe cuvette results in poorer sensitivity with probe atomization (the i.d. is 6.5 mm for the probe cuvette and 5.0 mm for the normal cuvette). The poorer sensitivity obtained with the same type of probe system described here was also observed by Brown' in the determination of manganese and lead. The poorer sensitivity in probe atomization requires a larger sample volume. In our experience the maximum sample volume in probe mea- surements (including modifier solution) is 20-25 mm3; this requires very careful adjustment of the probe and autosam- pler tip positions.Interference Effects of Some Metal Chlorides During the studies the nitric acid concentration in the test solutions was kept at 0.1 mol dm-3 to ensure that the excess of the acid does not drive chlorides away during the ashing step. Also a short ashing time was used (hence it could be seen if any difference existed between the two atomization methods in the presence of a chloride matrix). Because there is always a risk of losing antimony as volatile chloride compounds during the ashing stage lower ashing tempera- tures were also employed. Fig. 3 shows an ashing plot obtained with an ordinary graphite tube in the presence of 100 ,ug of calcium chloride in a sample matrix.It can be seen that antimony is probably not lost as a volatile chioride during the ashing phase under these conditions but there is a depressive effect on the signal due to the presence of calcium chloride. The influences of the different metal chlorides on the signal for antimony are presented in Fig. 4. For compari- son the effects of some nitrate salts were also investigated. It can be seen that fewer interferences exist with probe atomization and signal quantification by integrated absor- bance. The chlorides of magnesium calcium and copper caused the strongest signal depression in tube-wall atomiza- tion. According to these findings the use of a probe atomization technique seems to be advantageous when antimony is determined in matrices containing alkali and alkaline earth metal chlorides.The studied matrix elements had a much greater influence on the peak shape in probe atomization than in tube-wall atomization (Fig. 5). Com- parison of the peak profiles indicates a slower atomization rate using the probe system. When for example a greater amount of a common chemical modifier palladium was used a very stable compound was formed resulting in a significant reduction in the peak height sensitivity (Table 3). It can be seen from Fig. 5 that the signal does not return to the baseline during the 3 s atomization sequence in the presence of manganese and palladium and hence lower peak areas are obtained. Additionally a double peak was occasionally observed with palladium nitrate and copper nitrate (and also with copper chloride).It is obvious that the probe atomization technique described here is only suitable for the determination of easily or medium volatile ele- ments because the graphite probe has a cooling effect when introduced into the hot atomizer tube and the gas-phase temperature is not as high as it should ideally be.20 Therefore the atomization rate for those elements which form thermally stable compounds is slow which results in poorer sensitivity in the measurements. It is also known Fig. 4 Influence of some metal chlorides (a) and (b) NaCI; (c) and (6) KCI; (e) and cf) MgC12; (g) and ( h ) CaC1,; (i) and 0) CuCI,; ( k ) and (/) ZnC1,; (m) and (n) MnCI,; and (0) and @) Mn(N03) on the determination of 1 ng of antimony using A tube-wall atomization ash 1000°C; €3 tube-wall atomization ash 500°C; and C probe atomization ash 800°C.A cool time of 30 s was used when the effect of Mn(NO,) was studied 80 80 60 60 - 120 120 60 60 120 1 8om; 1 40 I O O E i :"Fl 80 80 60 60 0 40 80 120 0 40 80 120 Metal chloride/pg 100 120Fl j---Tl 80 80 60 60 0 40 80 120 0 40 so 120 Manganese nitrate/pgJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY. AUGUST 1992 VOL. 7 739 I 0 . 4 0 0 h _ 0.200 I I \ I \ \ 0 0.160 \ a m 0.400 5 s ," 0.200 -0 c 0 a c Y .- g o 0 0.6 1.2 1.8 2.4 0.400 0.200 n-1 I \ Tirne/s Fig. 5 Absorption signals for 1 ng of antimony with tube-wall atomization (broken line) and probe atomization (solid line) for ( a ) antimony alone (integrated absorbances probe 0.080 wall 0.143); (6) in the presence of 5 pg of Cu (probe 0.072 wall 0.141); (c) in the presence of 2.5 p g of Mn (probe 0.069 wall 0.133); (6) in the presence of 1 pg of Pd (probe 0.064 wall 0.133); (e) in the presence of 5 pg of Pd (probe 0.052 wall O .l O l ) and (f) an enlarged signal from Fig. 5 ( d ) that carbide-forming elements are only slowly vaporized from an electrographite surface. During the studies it was observed that the well known over-correction interferences existed with the deuterium background correction system used when certain elements were present in the sample m a t r i ~ . ~ ~ ~ ~ ~ The most serious effects were found with iron and cobalt but the presence of nickel also caused an interference. The over-correction interference had a much smaller effect on the peak height than on the integrated absorbance measurement. The signals obtained i.e.background and specific absorbance are separated to some extent. A narrow slit-width might reduce spectral interferences owing to a less structured background falling within the spectral bandpass of the spectrometer. When a slit-width of 0.2 nm was used the spectral interferences were diminished and the effect of 100 pg of cobalt chloride on the determination of 1 ng of antimony was almost completely avoided with tube-wall atomization (in both peak height and integrated absorbance measurement modes). The presence of ascorbic acid might shift the atomic absorption pulses and reduce spectral interferences during the atomization step. 16726v27 The interference effects caused by chlorides can be reduced by using chemical modification Table 3 Effect of the presence of palladium on the determination of 1 ng of antimony. An ashing temperature of 800°C was used.The signal for antimony alone is 100 Relative absorbance* Pd/pg Peak height Integrated absorbance 0.5 75 k 0.7 101 f 1 . 1 1 .o 73 k 0.3 98k 1.8 2.0 72 +- 0.3 93-t- 1.4 5.0 62+ 1.3 82 k 3.6 7.5 43 + 2.3 59 f 4.3 *Average of at least three measurements f SD. and a long pyrolysis time at high temperature. Therefore chemical modification in the presence of ascorbic acid was tried for reducing the interferences in the determination of antimony by probe atomization. For example using copper and palladium (as nitrates) the ashing temperature could be raised in comparison with the use of a 0.1 mol dm-3 nitric acid matrix alone (Table 4).When ascorbic acid was added to the samples (75 pg in 10 mm3) and 5 mm3 of palladium- magnesium chemical modifier solution ( 1 pg of Pd and 2 ,ug of Mg) and a higher pyrolysis temperature (1 250 "C) were used the shapes of the atomic absorption pulses were more regular than those obtained with palladium alone (Fig. 5). The peak height and integrated absorbance sensitivities were also improved. However although a 0.2 nm slit-width was employed larger amounts of iron reduced the inte- grated absorbance sensitivities and satisfactory results for geological reference materials were obtained only when the peak height measurement mode was used. When ascorbic acid was present in samples a shiny surface similar to pyrolytic graphite was gradually formed on the graphite probe.This increased the lifetime of the probe considerably but its surface also became more slippery. Therefore the dispensing volume was decreased to 10 mm3 to prevent the spreading of the sample. If larger amounts of ascorbic acid (> 100 pg in 10 mm3) were used more carbonaceous residue rapidly built up on the probe. This could even hinder the movement of the probe through the slot. It was also observed that owing to the presence of ascorbic acid a deposit of palladium metal was gradually formed inside the autosampler dispensing tip. This problem was avoided when 1 mol dm-3 nitric acid was used as a flushing solution in the autosampler. Because satisfactory results for geological reference samples were obtained only when non-ideal conditions i e .peak height measurements were used for quantification palladium-magnesium chemical modification and ascorbic acid were omitted and the use of a magnesium modifier alone was studied. Sharp absorption pulses and improved740 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 Table 4 Stability of 1 ng of antimony in different sample matrices with probe atomization Relative absorbance* Ashing temperature/"C 800 1000 1200 1300 1400 1500 0.1 mol dm- HNO - Peak Integrated height absorbance 100 k 2.6 100k3.2 96 + 4.1 95 k 3.4 36+ 18 42 k 9.2 15k5.6 2 0 2 0 - - - - 800 1000 1200 1300 1400 1500 Peak Integrated height absorbance 76 + 2.6 96 k 4.2 7 4 2 12 90+ 10 23 + 7.3 22 k 3.2 48+ 12 5 8 2 15 - - - - 0.5 mol dm- HN03 Peak Integrated height absorbance 109f6.5 112k5.3 101 k9.0 10924.8 35 k 3.2 42 + 4.2 14k2.2 19k4.6 - - - - 0.5 p g Pd Peak Integrated height absorbance 95 k 2.0 102 k 3.5 97 k 5.3 87 k 5.6 94 f 4.3 95 f 3.0 86 k 5.1 66 k 7.4 - - - - Peak height 75 + 0.7 77 k 0.6 78k 1.0 78k0 73+ 1.6 62 f 2.4 ' Integrated absorbance 101 k 1.1 104k 1.4 102+ 1.8 102t- 1.4 98 +- 2.4 80 f 7.7 Peak 1 n tegrated height absorbance 24+ 1.2 36k 2.3 87+ 1.8 114k2.7 83 k 2.0 116k2.1 116kO 81 k 1.5 102k0.9 71 k3.3 - - *Average of at least three measurements 2 SD.t 5 mm3 of palladium modifier solution were added separately. An irregular peak was obtained at an ashing temperature 800"C.26 peak height and integrated absorbance sensitivities were obtained when a large amount of magnesium modifier was introduced on the probe (30 pg of magnesium as nitrate).While a longer pyrolysis time (1 5 s at 1000 "C) and a slightly lower atomization temperature (2550 "C) were employed the over-compensation interferences seemed to be resolved (Fig. 6). The detection limit for antimony in aqueous 120 z 5 n m 100 E -c 90 8 80 4 110 UJ r 3 W .- 70 - W - 4 I\ LL 60 120 Q) ( b ) 0 g 110 t B U 70 a 0 20 40 60 80 100 120 M eta I c h I o r i de/pg Fig. 6 Influence of some metal chlorides (A FeCl,; B CoCI,; and C NiCI2) on the determination of 1 ng of antimony showing (a) relative peak height absorbance and (6) relative integrated absor- bance. Slit-width is 0.2 nm 0.060 I I 0.040 3 0.020 Q 0 0.6 1.2 1.8 2.4 " Time/s Fig. 7 Signal trace for antimony recorded from material GXR-5 dissolved in HF and HNO (integrated absorbance 0.026 s) solution using the conditions described above correspond- ing to the mean of the blank plus three times its standard deviation was 0.05 ng (ten successive measurements).The usefulness of the probe system was tested by determining antimony in several US Geological Survey reference materials. The sample volume was 10 mm3 except for material GXR-5 (15 mm3) (Fig. 7). The results obtained with the integrated absorbance measurement mode and aqueous calibration are given in Table 5. All the recommended values except for GXR-3 were obtained by neutron activation analysis.28 Therefore these values prob- ably represent the total antimony contents in these ma- terials. Our results are lower than the recommended values and is believed to be due to incomplete dissociation of the sample matrix as noticed before.29 For example the amount of iron in GXR-2 is low (1.81% Fe203 and 0.76% FeO) which is well below the 1 pg of iron in the final sample dispensed into the graphite furnace. It is unlikely that this low value causes any spectral interferences.Otherwise GXR-3 contains a high level of iron (27.1% Fe20,) and a good recovery was obtained from it. When the samples of CXR- 1 were fused with lithium tetraborate it is possible that some antimony escaped owing to the high fusion temperature. Sample solutions of dissolved materials GXR- 1 and GXR-2 were also spiked with an antimony standard. These experiments showed complete recovery of antimony from the spiked samples.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL.7 74 I Table 5 Concentration of antimony measured in geological reference samples Antimony foundlpg g-I Sample GXR- 1 GXR-2 GXR-3 GXR-5 _____ _ _ _ _ _ _ ~ Recommended material HCI-HNO3 HF-HNO3 Fusion technique value? Jasperoid 68.6-t 1.8 85.8 k 15 96.8 +- 7.8 122k I8 (n=3)* (n=4) (n=4) Soil - 35.1 k 1.5 - 49k5 (n=4) Fe-Mn-W-rich - 36.4 k 1.7 - 38a I7 Hot Springs Deposit (n=4) Soil (n=4) - 1.1 kO.01 - 1.63 f 0.28 *n = Number of weighings. ?See ref. 28. Conclusions The interference effects of some chloride salts (especially magnesium and calcium chloride) can be considerably reduced by using probe atomization when compared with those in ordinary tube-wall atomization. However the completely interference-free determination of antimony is not achieved.The lower sensitivity of the probe system is a drawback especially when the surface tension of the sample is low and a decreased sample volume must be used. Also an impractically long cooling time must be employed to achieve good reproducibility in probe atomization. The surface of the probe cannot be coated with pyrolytic graphite because it would then be too slippery. Therefore there is a possibility that work with corrosive solutions will cause a rapid wearing of the thin probe head. The most serious drawbacks of probe atomization are the poor chemical and mechanical strength of the probe and the limited sample volume. When antimony is determined in real samples one should also take into account the well known over- correction effects that appear with the deuteripm back- ground correction system especially with samples contain- ing iron.The ideal situation that probe atomization should overcome gas-phase interferences without chemical modifi- cation is affected by these non-analyte absorptions when a deuterium background correction system is used. The background correction problems could be eliminated to a great extent by using Zeeman-effect background correc- tion. However if an ordinary deuterium background corrector is employed the spectral interferences can be reduced considerably by using a narrow slit-width (0.2 nm) and magnesium as chemical modifier. References L’vov B. V. Spectrochim. Acta Part B 1978 33 153. Hageman L. Mubarak A. and Woodriff R. Appl. Spectrosc. 1979 33 226. Frech W. and Jonsson S.Spectrochim. Acta Part B 1982 37 1021. Littlejohn D. Lab. Pract. 1987 36 12 1. Brown A. A. J. Anal. At. Spectrom. 1988 3 67. Garnys V. P. and Smythe L. E. Anal. Chem. 1979 51 62. Manning D. C. Slavin W. and Myers S. Anal. Chem. 1979 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 51 2375. Manning D. C. and Slavin W. Anal. Chim. Acta 1980 118 301. Slavin W. and Manning D. C. Spectrochim. Acta Part B 1982 37 955. Carroll J. Marshall J. Littlejohn D. and Ottaway J. M. Fresenius’ Z. Anal. Chem. 1985 322 145. Ottaway J. M. Carroll J. Cook S. Corr S. P. Littlejohn D. and Marshall J. Fresenius’ Z. Anal. Chem. 1986 323 742. Littlejohn D. Marshall J. Carroll J. Cormack W. and Ottaway J. M. Analyst 1983 108 893. Littlejohn D. Cook S. Durie D. and Ottaway J. M. Spectrochim. Acta Part B 1984 39 295. Chakrabarti C. L. Karwowska R. Hollebone B. R. and Johnson P. M. Spectrochim. Acta Part B 1987 42 1217. Chakrabarti C. L. Xiuren H. Shaole W. and Schroeder W. H. Spectrochim. Acta Part B 1987 42 1227. Niskavaara H. Virtasalo J. and Lajunen L. H. J. Spectro- chim. Acta Part B 1985 40 1219. Atomic Absorption Databook Philips Scientific Cambridge 1988. Brown A. A. Anal. Chim. Acta 1985 175 319. 1989 Annual Book of ASTM Standards American Society for Testing and Materials Philadelphia PA 1989. Corr S. P. and Littlejohn D. J. Anal. At. Spectrom. 1988 3 125. Carroll J. Miller-Ihli N. J. Harnly J. M. Littlejohn D. Ottaway J. M. and O’Haver T. C. Analyst 1985,110 1153. Ortner H. M. Schlemmer G. Welz B. and Wegscheider W. Spectrochim. Acta Part B 1985 40 959. Giiell 0. A. and Holcombe J. A Spectrochim. Acta Part B 1988 43 459. Falk H. and Schniirer C. Spectrochim. Acta Part B 1989,44 759. Martinsen I. Radziuk B. and Thomassen Y. J. Anal. At. Spectrom. 1988 3 1013. Voth-Beach L. M. and Shrader D. E. J. Anal. At. Spectrom. 1987 2 45. Gilchrist G. F. R. Chakrabarti C. L. Byrne J. P. and Lamoureux M. J. Anal. At. Spectrom. 1990 5 175. Gladney E. S. and Roelandts I. Geostand. Newsl. 1990 14 21. Peramaki P. and Lajunen L. H. J. Analyst 1988 113 1567. Paper 0/016808 Received April 17 1990 Acceated ADril2. 1992

ISSN:0267-9477    

DOI:10.1039/JA9920700735

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 743 Platform Wall and Probe Electrothermal Atomization for the Determination of Aluminium in Clinical Fluids Juan M. Marchante Gayon Juan Perez Parajon and Alfred0 Sanz-Medel* Department of Physical and Analytical Chemistry Faculty of Chemistry University of Oviedo Julian Claveria s/n 33006 Oviedo Spain Craig S. Fellows Unicam Analytical Systems Unicam Limited York Street Cambridge CB 1 ZPX UK A detailed examination of three electrothermal atomization techniques (wall platform and probe) for the determination of aluminium in clinical fluids was carried out. The optimum instrument conditions and atomization technique were compared for two commercial instruments. Platform atomization using integrated absorbance measurement was the method chosen for the determination of aluminium in tap water dialysis fluids and serum samples.The serum samples were diluted 1 + 1 with 0.2% v/v Triton X-1 00. No chemical modifier or sample pre- treatment was necessary. The concentration of aluminium was evaluated directly by aqueous standard calibration. The detection limits were similar for both instruments (about 1.5 pg dm-3). Probe atomization was convenient for reducing interference effects. Some practical drawbacks were observed for serum analysis. Keywords Aluminium determination; clinical fluids; atomic absorption; wall platform and probe atomization It is recognized that aluminium is implicated in the pathogenesis of some disorders observed in patients with chronic renal failure undertaking haemodialysis.These disorders include dialysis encephalopathy dialysis osteo- dystrophy parathyroid disfunction and microcytic The European Community has recommended4 minimizing patient exposure to aluminium by restricting the final dialysis solution concentration to 30 pg dm- and controlling aluminium concentrations in serum. The monitoring of dialysis fluids and serum samples for alu- minium is essential to prevent toxic effects in uremic patient^.^ Electrothermal atomization atomic absorption spectro- metry (ETAAS) and inductively coupled plasma atomic emission spectrometry have been used successfully in our laboratory for the last seven years to monitor and control aluminium concentration in serum. For the determination of aluminium in serum ETAAS is the preferred technique6 but there are several different sample pre-treatments each of which have their advocates.The common trend seems to be the use of simple dilutions without chemical modifica- t i ~ n . ~ - l ~ The water used for the dilutions should be previously checked for aluminium contamination. The use of a chemical modifier can prove inconvenient because of the risk of contamination. Systematic studies have indicated that with some atom- izers there are chemical interferences. 1 9 1 5 9 1 7 Metal chlor- ides are of particular importance in the determination of a l u m i n i ~ m . ~ ~ - l ~ The introduction of the L'vov platform and the concept of a stabilized temperature platform furnace have reduced these problems ~onsiderably.~.~ * 7 1 9 An im- provement to the L'vov platform was suggested by L'vov and Palievazo~Z1 and consisted of a probe which is intro- duced into a graphite furnace already set at the atomization temperature. This probe atomization has been investigated by several research g r ~ u p s ~ l - ~ ~ but its use in the analysis of clinical samples for aluminium has not yet been reported.In this work the use of wall platform and probe atomization in the determination of aluminium in water dialysis fluids and blood serum is investigated. Two commercial instruments were used and the comparative performance is reported. *To whom correspondence should be addressed. Experimental Instrumentation A Philips Analytical PU94OOX atomic absorption spectro- meter equipped with a PU938OX furnace a PU9385X furnace autoprobe and a PU939OX autosampler was used.A Philips aluminium hollow cathode lamp (HCL) was used at a current of 8 mA. Pyrolytic graphite coated graphite cuvettes were used with and without totally pyrolytic graphite L'vov platforms. Totally pyrolytic graphite cu- vettes with graphite probes were used for probe atomiza- tion. A Perkin-Elmer 3030 (PE-3030) atomic absorption spec- trometer equipped with an HGA-500 furnace and an AS-40 autosampler a Perkin-Elmer aluminium HCL at a current of 25 mA and pyrolytic graphite coated graphite tubes with totally pyrolytic graphite L'vov platforms were used. Instrumental conditions were optimized and are detailed in Table 1. Argon was used as the inert gas for all analyses. Reagents A stock solution of 1000 mg dm-3 was prepared by dissolving 1 .OOO g of aluminium foil (analytical-reagent grade Merck) in 20 ~ m - ~ of sulfuric acid (1 + 1) (Carlo Erba) and diluting to 1000 cm- with ultrapure water and 50 of concentrated nitric acid (Suprapur grade Merck).A 100 mg dm- standard solution in 5% v/v nitric acid was prepared from the stock solution. Solutions of potential interferents NaNO CaCO Fe(N03)z and KN03 were prepared by dissolution of solids or dilution of concentrated solutions as appropriate. Mg(N03)2 H3P04 HCl Zn(NO312 HZS04 Cu(NO3)29 Results and Discussion Analytical Conditions A comparison of the ashing-atomization curves and atomic signals for 50 pg dm- of aluminium in water dialysis fluid and serum using the PU94OOX and PE-3030 was made. Serum samples were diluted 1 + 1 with 0.2% v/v Triton X- 100. It has been shown6 that the platform gives a superior performance to wall atomization for the determination of744 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL.7 80 60 40 20 0 Table 1 Comparative instrumental settings for determinations of aluminium '+ - \+ - - - I I General conditions- PE-3030 Wavelengthhm 309.3 Slit/nm 0.7 Intensity (HCL)/mA 25 Integration time/s 4.0 Background correction On Signal processing Integrated absorbance Tube type Pyrolytic graphite coated graphite with L'vov platform PU94OOX 309.3 0.5 8 4.0 On Integrated absorbance PI Pyrolytic graphite coated graphite; W pyrolytic graphite coated graphite with L'vov platform; Pr totally pyrolytic graphite with graphite probes Graphite furnace programme- PE-3030 PU94OOX Temperature/ Internal gas Temperature/ Internal gas Step "C Ramp/s Hold/s flow/cm3 min-' Step "C Time/s Ramp/"C s-' flow/cm3 min-' 1 100 20 15 300 1 2 150 15 15 300 3 800 15 15 300 4 1200 15 15 300 2 5 2600 o* 4 10 6 2700 1 2 300 7 20 1 5 300 3 4 5 6 7 * Maximum power heating.t Commands TC = temperature control; and RD= read. 100 (PI) 200 (Pr) 150 (Pl) 100 (W) 15 5 300 200 (Pr) 800 15 50 300 1000 15 25 300 2600 4 ot 0 2700 2 2000 300 20 20 0 300 80 (W) 15 10 300 120 I Temperature/"C Fig. 1 Ashing curves for solutions of aluminium in water with platform atomization using A PE-3030; B PU94OOX; and C PU94OOX with magnesium nitrate as chemical modifier 120 P 0' I I I I I 900 1100 1300 1500 1700 TemperaturePC Fig. 2 Ashing curves for aluminium in dialysis fluids with platform atomization using A PE-3030; B PU94OOX; and C PU94OOX with magnesium nitrate as chemical modifier g s it > K Fig.120 900 1100 1300 1500 1700 Temperature/"C 3 Ashing curves for aluminium in human serum with platform atomization using A PE-3030; B PU94OOX; and C PU94OOX with magnesium nitrate as chemical modifier aluminium in biological samples. The results for aqueous solutions are shown in Fig. 1. The maximum signal obtained in every case was taken to be 100% recovery. The PE-3030 allowed a higher ashing temperature of about 1200 "C to be used but the addition of 4 pg of magnesium (as magnesium nitrate) into the furnace compensated for the rapid volatilization when using the PU94OOX. Both instruments behaved in a similar manner for dialysis fluids and serum samples (these samples contain 10-50 pg cm-3 of magnesium) with an optimum ashing temperature of about I200 "C (Figs.2 and 3). The addition of 4 pg of magnesium (as magnesium nitrate) to the furnace increased the maximum ashing temperature to about 1500 "C for all sample types tested (Figs. 1-3). Magnesium was found to increase the peak height absorbance while the integrated absorbance remained constant. The absorbance values were found to increase with increasing atomizationJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY 2ot 01 1 I I 1 I Tern pe rat u rePC 2100 2200 2300 2400 2500 2600 2700 Fig. 4 Atomization curves for solutions of aluminium in water A PE-3030; B PU94OOX (platform); C PU94OOX (wall); and D PU94OOX (probe) AUGUST 1992 VOL.7 745 > 5 60- 8 40 - 20 - U 900 1100 1300 1500 1700 Te m peratu rePC Fig. 5 Ashing curves for aluminium in water solution using PU94OOX A wall atomizatiion; and B probe atomization temperature and the lifetime of the cuvette was found to decrease. A compromise temperature of 2600 "C was chosen for sample analysis (Fig. 4). The ashing-atomization plots for wall and probe atomi- zation are shown in Figs. 5 and 6. Maximum ashing temperatures were 1000 "C for water and 1200 "C for dialysis fluids and serum samples. The effect on the absorbance signal for aluminium of potentially interfering elements at physiological levels,28 or slightly higher was investigated. No modifier and an ashing temperature of 1000 "C were used. These results are summarized in Table 2.Using integrated absorbance measurement only wall atomization in the presence of a high chloride concentration was effected. The presence of Ca2+ Mg2+ and Fe3+ was found to increase peak height 20 " 900 1100 1300 1500 1700 TemperaturePC Fig. 6 Ashing curves for aluminium in dialysis fluids using PU94OOX A wall atomization; and B probe atomization absorbance and C1- and Zn2+ to decrease peak height absorbance. No interference effect on the peak height or integrated absorbance measurement was found to occur if probe atomization was used. Study and the Performance of Probe Atomization The stages required for probe atomization are essentially the same as for wall and platform atomization. Slightly higher temperatures were required for each stage as the probe is only exposed to radiative heating rather than a mixture of radiative and conductive heating.Immediately before the atomization stage the graphite probe is with- drawn from the graphite cuvette. The cuvette is heated up to the atomization temperature and allowed to stabilize. The graphite probe is then inserted allowing the sample to atomize into a thermally stable environment. It was found that the analysis of serum samples left a small amount of carbonaceous residue that affected the positioning of the next sample. The use of Triton X-100 eliminated this problem for platform atomization as the carbonaceous residue formed a thin layer that did not affect sample deposition. Triton X-100 could not be used for probe atomization as it caused the sample to run off the graphite probe.The introduction of air during the ashing stage removed the carbonaceous residue but had the effect of seriously reducing the lifetime of the probe. The introduction of air during other s t a g e ~ ~ ~ ~ ~ ~ and the use of a desorption stage was investigated but was not found to prevent the reduction in lifetime. The detection limit sensitivity linearity and precision for aqueous solutions for the different types of atomization and instruments are shown in Table 3. Table 2 Influence of the individual elements in serum and dialysis fluids on the aluminium signal (peak height and integrated absorbance) Absorbance using PU94OOX (Oh) Absorbance using Maximum PE-3030 (platform) (To) Platform Wall Probe concentration/ Element mg dm-3 Peak height Integrated Peak height Integrated Peak height Integrated Peak height Integrated Na K Ca Mg P c1 S Zn c u Fe 0 400 200 80 200 10 000 24 4 2 2.4 None None + 60 + 60 None - 30 None - 10 None + 10 None None None None None None None None None None None None + 30 + 30 None - 10 None - 10 None + 10 None None None None None None None None None None None None + 30 + 30 None - 10 None - 10 None + 10 None None None None None - 10 None None None None None None None None None None None None None None None None None None None None None None None None746 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL.7 Table 3 Detection limit sensitivity linearity and precision for aqueous solutions Atomization Absorbance method measurement Detection Iimitlpg Sensitivityjpg PE-3030- Platform Integrated 16 Peak height 10 PU94OOX- Platform Integrated 15 Wall Integrated 16 Peak height 8 Probe Integrated 32 Peak height 22 Peak height 1 1 12 8 11 4 10 2 24 9 Within-run Linearityhg precision (Yo) 0.9 1 0.9 2 0.3 2 0.3 1 0.3 2 1 .o 3 1 .o 2 0.3 1 I Table 4 Results in ppb (meankstandard deviation) for the determination of aluminium using both instruments; n= 6 Aluminium concentration Sample Aqueous 6007 6008 6020 7018 601 5 7007 7016 701 9 245 248 260 270 Dialysis fluid Human serum PE-3030 66.1 f 1.9 9.9 f 6.4 33.5 & 1.3 45.2k 1.1 32.5 f 0.8 24.9 f 0.7 15.9 k 3.4 29.3 f 2.5 40.4 t 1.7 6 1.2 f 0.9 93.1 1-0.5 92.7 f 0.4 PU94OOX 68.3 f 4.7 12.9k 1.0 32.4t- 1.0 40.1 ? 0.9 36.9 f 0.5 28.9 f 0.6 38.3 f 1.0 60.8 k 2.5 83.3 f 1.8 84.1 k 2.8 23.6 f 0.4 19.4 f 0.5 TEQAS value 66.4 f 19.2 9.2 f 2.0 33.3 f 11.6 44.2 f 13.6 36.7f 11.6 24.3 f 7.9 18.4 f 6.6 25.9 f 10.0 38.6f 10.5 61.3f 11.3 88.6 k 15.7 89.6 k 13.2 Determination of Aluminium in Real Samples The results for real samples (Table 4) indicate a preference for the use of platform atomization unless high levels of interferents are present.The determination of aluminium in water and dialysis solutions was accomplished without dilution but serum samples were diluted 1 + 1 with 0.2% v/v Triton X- 100. The calibration was made using aqueous standards. A 10 mm3 aliquot of the sample solution was injected into the PE-3030 and 5 mm3 into the PU94OOX. No modifiers were used. Samples were obtained from the Trace Elements Quality Assessment Scheme (TEQAS) organized by The Robens Institute University of Surrey Guildford UK.Conclusions A comparison of different atomization techniques and two commercial instruments was made for the determination of aluminium in water dialysis fluids and serum samples. Platform atomization appears to be the most convenient method for both instruments. Atomization from the wall can present problems due to interferences (particularly if peak height measurement is used). Probe atomization exhibits poorer detection limits and is more laborious than other techniques. Integrated absorbance measurement is less affected by interferences. Chemical modifiers are not necessary for the determination of aluminium in the samples investigated. Probe atomization can reduce interference effects and increase the analytical linear range for aluminium but is not recommended for samples with a high biological organic content because of practical difficulties in eliminat- ing the carbonaceous residue. ’The results detailed in Table 4 show good agreement with the certified values for both instruments.Detection limits sensitivies and the within-run precision for aqueous solu- tions are also similar (Table 3). Thus both instruments are capable of giving good results for these types of samples. The PU94OOX dynamic range is shorter than that observed with the PE-3030. However a wider analytical range can be obtained with the former by using the curved portion of the calibration graph made possible by employing for analysis the segmented parabolic curve correction algorithm.(It should be noted however that the attainable accuracy can be reduced by using this curved portion.) Financial support for this research from the Fundacion para el Foment0 en Asturias de la Investigacion Cientifica Aplicada y la Tecnologia’ and ‘Comision Interministerial de 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 Ciencia y Tecnologia’ is gratefully acknowledged. References Aluminium in Food and the Environment eds. Massey R. and Taylor D. Royal Society of Chemistry Cambridge 1989 publication No. 73. Fell G. S. Nephrologie 1986 VI 26. Wills M. R. and Savory J. Lancet 1983 ii 29. Proposal for a Council Directive Relating to the Protection of Dialysis Patients by Minimising the Exposure to Aluminium Off J. Eur. Comm. No. C202 1983 and Amended Proposal No.C150 1985. Cannata J. B. Briggs J. D. Junor B. J. R. Fell G. S. and Beastall G. Lancet 1983 im 501. Sanz-Medel A Rodriguez Roza R. Gonzalez Alonso R. Nova1 Vallina A. and Cannata J. J. Anal. At. Spectrom. 1987 2 177. Slavin W. J. Anal. At. Spectrom. 1986 1 281. Gardiner P. E. and Stoeppler M. J. Anal. At. Spectrom. 1987 2 401. Rollin H. B. Theodorou P. and Kilroe-Smith T. A. Microchem. J. 1987 35 373. Lugowski S. Smith D. C. and Van Loon J. C. J. Biomed. Mat. Res. 1987 21 657. Fagioli F. Locatelli C. and Gilli P. Analyst 1987 112 1229. Gitelman H. J. and Aldeman F. R. Clin. Chem. (Winston- Salem N.C.) 1989 35 1517. Tatro M. E. Spectrosc. Znt. 1989 1(4) 22. Slavin W. and Manning D. C. Prog. Anal. At. Spectrosc. 1982 5 243. Manning D. C. and Slavin W. Anal.Chem. 1978,50 1235. Matousek J. P. Prog. Anal. At. Spectrosc. 198 1 4 247. Erspamer J. P. and Niemczyk T. M. Anal. Chem. 1982 54 538. Leung F. Y. and Henderson A. R. Clin. Chem. 1982 28 2 129.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 747 19 Frech W. Cerdegren A. Cederberg C. and Vessman J. Clin. Chem. 1982 28 2259. 20 L’vov B. V. and Palieva L. A. Zh. Anal. Khim. 1978 33 1572. 21 L’vov B. V. and Palieva L. A. Zh. Anal. Khim. 1980 35 1744. 22 Slavin W. and Manning D. C. Spectrochim. Acta Part B. 1982 27 955. 23 Ottaway J. M. Anal. Proc. 1984 21 55. 24 Chakrabarti C. L. Wu S. Karwowska R. Chang S. B. and Betels P. C. At. Spectrosc. 1984 5 69. 24 Marshall J. Baxter D. C. Carroll J. Cook S. Corr S. P. Giri S. K. Durie D. Littlejohn D. Ottaway J. M. Stephen S. C. and Wright S. Anal. Proc.,’ 1985 22 371. 26 Brown A. A J. Anal. At. Spectrom. 1988 3 67. 27 Corr S. P. and Littlejohn D. J. Anal. At. Spectrom. 1988 3 125. 28 Sanz-Medel A. Rodriguez Roza R. Gonzalez Alonso R. Nova] A. and Cannata J. ICP Znf Newsl. 1986 11 708. 29 Ekerlin R. H. Hoult D. W. and Carnrick G. R. At. Spectrosc. 1987 8 64. 30 Neve J. and Molle L. Acta Pharmacol. Toxicol. 1986 59 606. Paper 2/00521B Received January 30 1992 Accepted March 13 1992

ISSN:0267-9477    

DOI:10.1039/JA9920700743

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 749 Determination of Aluminium in Human Brain Tissue by Electrothermal Atomic Absorption Spectrometry Ning Xu Vahid Majidi,* William D. Ehmann and William R. Markesbery Departments of Chemistry Pathology Neurology and The Sanders Brown Center on Aging University of Kentucky Lexington KY 40506 USA Brain tissue is rich in phosphorus and alkali and alkaline earth metals. A synergistic interference has been observed with the determination of aluminium by electrothermal atomic absorption spectrometry when phosphorus is present in the sample along with alkali or alkaline earth metals. For the determination of aluminium in human brain samples potassium dichromate was found to be an effective chemical modifier. The concentration of aluminium in National Institute of Standards and Technology (NIST) Standard Reference Materials (SRMs) Citrus Leaves (SRM 1572) Oyster Tissue (SRM 1566a) and in National Institute for Environmental Studies (NIES) Certified Reference Material (CRM) No.6 Mussel were determined. Values of 76.6k4.2 86.820.5 and 55.5k10.1 pg g-' were obtained before the use of the modifier and 88.226.2 197.7k6.8 and 215.0-tl2.5 pg g-' after the use of potassium dichromate as chemical modifier for Citrus Leaves Oyster Tissue and Mussel respectively. The concentrations of aluminium determined with the use of potassium dichromate as chemical modifier are in good agreement with the certified values in the standard samples. The aluminium contents of several human brain samples were also determined.Values ranged from 6.2 to 9.8 pg g-' of aluminium (dry mass basis). Keywords Aluminium determination; human brain tissue; electrothermal atomic absorption spectrometry The concentration of aluminium in brain tissue and its role in the pathogenesis of human neurodegenerative diseases such as Alzheimer's disease and amyotrophic lateral sclero- sis have been under investigation for the past few Markesbery et a1.I reported a mean aluminium concentra- tion in normal adult brain tissue of 0.467 k 0.033 pg g-I on a wet mass basis (equivalent to 2.21 kO.16 pg g-I on a dry mass basis) by instrumental neutron activation analysis. Crapper et aL2 determined a mean aluminium content of 2.0-tO.1 pg g-' (dry mass basis) in normal human brain tissue using the procedure developed by Krishnan et aL3 for electrothermal atomic absorption spectrometry (ETAAS).McDermott et aL4 reported a mean value of 2.5 2 0.3 pg g-' of aluminium in normal brain tissue (dry mass basis) and suggested a threshold level of 10-20 pg g-l for human aluminium neurotoxicity. Since the indigenous concentration of aluminium in brain tissue is very low development of sensitive analytical methods is necessary. Electrothermal atomic absorption spectrometry (ETAAS) has become one of the most com- monly used techniques for the determination of trace element^.^ However when aluminium is determined in samples other than clean aqueous solutions chemical interferences can complicate the analysis. In some in- stances the detection of aluminium has been performed by ETAAS without the use of chemical modifiers in order to avoid the risk of sample c ~ n t a m i n a t i o n ~ ~ ~ whereas others confirmed that desirable analytical signals for real samples can only be obtained by using various chemical modi- f i e r ~ .~ - ' ~ Some of the most commonly used chemical modifiers for the determination of aluminium include magnesium nitrate,8 ammonium phosphate l 4 ammonium nitrate,I5 potassium dichromate16J7 and palladium chlo- ride.I8 Phosphorus has been shown to interfere seriously with the determination of aluminium by AASk9 Brain tissue is very rich in phosphorus and alkali and alkaline earth metals.20 In a typical fresh brain sample the concentration of phosphorus is 2.49 f 0.03 mg g-'.*' The concentrations of potassium and sodium are 0.03 and 0.02 mg g-l respec- *To whom correspondence should be addressed.tively.20 The purpose of this work is to investigate the possible effect of phosphorus on the determination of aluminium and to find an analytical method that can minimize this interference. Experimental Apparatus A Varian atomic absorption spectrometer (Model 475 Sunnyvale CA) equipped with a graphite furnace atomizer (Model GTA-95) was used for the determination of aluminium. The wavelength and bandpass of the mono- chromator were set at 309.3 and 0.5 nm respectively. Argon was used as an inert gas with no flow during the atomization cycle and the integrated absorbance mode was employed to integrate the entire absorption profile. A 20 mm3 sample was deposited on the wall of a pyrolytic graphite coated graphite tube by use of an autosampler.Reagents and Chemicals The standards were prepared daily by serial dilution of the National Institute of Standards and Technology (NIST) standard aluminium stock solution (NIST 2 127- 1 Gaith- ersburg MD USA) with distilled de-ionized water. Nitric acid was added to the sample to produce a 1% v/v HN03 solution. The HN03 was of Ultrex purity (J. T. Baker Phillipsburg NJ USA) and K2Cr207 KH2P04 (NH4)2HP04 and Mg(N03)2 were all of Instra-Analyzed purity (J. T. Baker). The palladium stock solution was a 10 mg cm-3 NIST 2128-3 standard. Procedures Contamination control Aluminium is a ubiquitous element therefore brain samples with low concentrations of aluminium are easily contaminated. In order to avoid this problem a strict operational procedure was followed similar to those sug- gested by other w o r k e r ~ .~ * ~ J ~ Distilled de-ionized water ( 16 klZ ern-' Milli-Q water purification system Millipore Milford MA USA) was used throughout these studies. All7 50 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 Table 1 by ETAAS Operating conditions for the determination of aluminium Step Temperature/"C Ramp time/s Hold time/s Drying 120 20 25 Ashing 1400 10 40 1 Atomizing 2600 1 2 Cleaning 2600 - containers were soaked in 5% HNO for at least 48 h before rinsing thoroughly with large amounts of distilled de- ionized water. All containers were tested for background aluminium values and showed no detectable levels. Sample digestion Fresh brain (from the motor cortex and superior parietal lobe of 54-68 year old subjects) were obtained from autopsy with quartz and titanium knives at the University of Kentucky Medical Center.Talc-free gloves and virgin polyethylene storage vials were used. The samples were kept at -40 "C before freeze drying. The samples were chopped with a quartz knife into small pieces during freeze drying. Approximately 20 mg of brain sample were placed in a Teflon vessel (microwave acid digestion bomb Parr Mol- ine IL USA) l .O cm3 HN03 was added into the vessel and the sealed vessel was heated in a microwave oven until the samples appeared to be completely digested. The samples were heated stepwise at 200 W for 4 min 350 W for 4 min 250 W for 8 min and lastly the solutions were heated an additional four times at 100 W for 10 min each.'After each heating step the samples were allowed to cool to room temperature prior to the next heating step. The cooled digestion solution was then diluted to 5 cm3 with and without the chemical modifier. Determination of aluminium Calibration graphs were established daily prior to sample analysis. Operating conditions for both the sample and the standard solutions were identical (Table 1). The aluminium signal was measured at the atomization stage by integrating the total absorption profile. The concentrations of alumi- nium in the test solutions were then calculated from the Cali brat ion graph. Results and Discussion Effect of Phosphorus Species on Ashing Temperatures Without a chemical modifier and in the ,absence of phosphorus a solution containing 30 pg dm-3 of alumi- nium realized an ashing temperature as high as 1400 "C without any sample loss.When KH2P04 at a concentration of 1000 mg dm-3 was added to the aluminium solution the maximum ashing temperature was reduced to 1000 "C. For this mixture ashing temperatures above 1000 "C led to a reduced analyte signal due to sample loss. While the addition of phosphorus reduced the maximum ashing temperature the addition of K2Cr20 (1000 mg dm-3) to the system gave a stable ashing temperature of up to 1400 "C without any reduction of the aluminium signal. Further- more K2Cr207 mixed with KH2P04 (1000 mg dm-3) can also support an ashing temperature as high as 1400 "C. This demonstrates that K2Cr20 can minimize the loss of aluminium signal caused by the presence of phosphate which is one of the most common constituents in the digestion solution of brain tissue.It is believed that this increase in ashing temperature is due to the interaction of K2Cr20 with phosphate species. An oxygen ashing technique was also tested for maximiz- ing ashing temperatures and removing chemical interfer- ences. This technique has been demonstrated to produce excellent results in the determination of metals in blood samples.22 The technique however did not enhance the aluminium signal and significantly reduced the lifetime of the furnace. Effect of Phosphate Species on Aluminium Signal In an investigation of the effect of phosphorus species on the determination of aluminium a substantial amount of molecular absorption was found at the high concentration level of phosphate and alkali and alkaline earth metals.This molecular absorption interferes with the absorption signal of aluminium and cannot be removed by deuterium background correction. This effect was not observed in ammonium phosphate solution implying that it is the presence of phosphorus in addition to the alkali or alkaline earth metal cations that produced this interference. This is in agreement with previous studies that have shown that KH2P04 leads to the formation of species that generate substantial molecular absorbance in electrothermal atomiz- e r ~ . * ~ This hypothesis was further tested by the addition of Na+ Mg2+ and Ca2+ to ammonium phosphate solutions and similar interference with the aluminium signal was observed Table 2 Concentration of aluminium in biological samples in pg g-I (dry mass) Concentration of aluminium Concentration of aluminium with KZCr20 Certified Sample without modifier modifier value Citrus Leaves Mussel Oyster Tissue Brain No.1 Brain No. 2 Brain No. 3 Multi-element (NIST SRM 1572) 76.6 f 4.2 55.5 f 10.1 86.8 f 0.5 (NIES No. 6)* (NIST SRM 1566a) 3.3 f 0.3 3.4 & 0.3 2.4 f 0.3 28.5 f 0.7 Mix A Solution (NIST SRM 3 17 1) 88.2 f 6.2 215.0f 12.5 197.7f6.8 6.2 & 0.5 7.7 a 0.9 9.8 f 0.5 31.0a 1.8 92.0 220 202.5 12.5 - 30.0 f 0.2 *NIES= National Institute for Environmental Studies (Japan Environment Agency Ibaraki Japan).JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 7 5 1 Table 3 Recovery of aluminium added to the SRMs with and without chemical modifier No chemical modifier Certified value/ A1 added/ Total Al/ Sample Pug g-' pg g-' Citrus leaves 92.0 100 (NIST SRM 1572) 200 250 Mussel 220 100 (NIES No.6) 200 250 Oyster Tissue 202.5 +- 12.5 100 (NIST SRM 1566a) 200 250 Adding K2Cr207 ( 1000 p g dm-3) as the chemical modifier to the above solution can alleviate the molecular absorp- tion. Thus K2Cr207 helps to minimize the interference caused by the comprehensive effect of phosphorus species and alkali or alkaline earth metals. It is important to emphasize that copious amounts of these cations in addition to various forms of phosphorus exist in brain tissue. An alternative approach might require the use of plat- form atomization in conjunction with Zeeman-effect back- ground correction.Determination of Aluminium in Biological Samples Owing to its effectiveness in eliminating the interference caused by phosphorus species K2Cr207 was employed as the chemical modifier for the determination of aluminium in biological samples. The concentrations of aluminium in several brain samples and NIST biological Standard Refer- ence Materials (SRMs) were determined with and without the K2Cr207 modifier and are listed in Table 2. Accurate results were obtained for all reference samples when K2Cr207 was utilized as the chemical modifier. It should be noted that the ashing temperature used for these studies (1 400 "C) leads to no loss of sample only when K2Cr20 is employed as the chemical modifier. Subsequently although molecular absorption is observed along with the atomic signal the sample with no chemical modifier will generate a lower absorption signal due to substantial loss of sample at this ashing temperature.In order to investigate the recovery of the aluminium signal standard solutions containing 10,200 and 250pg g-' of aluminium were added to the SRMs with and without chemical modification (Table 3). A quantitative recovery of 92-104% was obtained when K2Cr207 was used as the chemical modifier whereas poor recovery was achieved when no chemical modifier was used. Further investigation of the recovery of the aluminium signal from Oyster Tissue was carried out by intentionally adding potassium phos- phate to the system so that the concentration of phosphate was three orders of magnitude larger than that of alumi- nium.After the analysis with ETAAS using K2Cr2.01 as the chemical modifier the concentrations of aluminium recovered were within 5.8% of the certified value. These results confirmed that K2Cr207 is a convenient chemical modifier for the determination of aluminium in biological samples. Conclusion The interference of phosphorus species can alter the aluminium signal obtained by ETAAS. A K2Cr207 solution pg g-' 170.8 266.6 303.0 296.8 372.6 422.0 269.0 339.8 403.1 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 ' 17 18 19 20 21 '22 23 K2CrZ07 modifier Recovery (O/O) 79 87 84 77 76 80 67 69 80 Total Al/ pg g-I Recovery (O/O) 184.7 93 275.9 92 324.9 93 321.9 02 427.7 04 471.8 01 300.9 98 39 1.8 95 440.8 95 References can minimize this interference by providing an optimum ashing temperature at which the phosphorus species are eliminated.Both the SRMs and the brain tissues provided adequate aluminium signals and excellent recovery was obtained when K2Cr207 was used as the chemical modifier. Markesbery W. R. Ehmann W. D. Hossain T. 1. M. Alauddin M. Ann. Neurol. 1981 10 5 1 I. Crapper D. R. Krishnan S. S. and Quittkat S. Brain 1976 99 67. Krishnan S. S. Quittkat S. and Crapper D. R. Can. J. Spectrosc. 1976 21 25. McDermott J. R. Smith A. J. Iqbal K. and Wisniewski H. M. Neurology. 1979 29 809. Miller-Ihli N. J. Spectrochim. Acta Part B 1989 44 122 1. Carrondo M. J. T. Lester J. N. and Perry R. Anal. Chim. Acta 1979 111 291. Parkinson I. S. Ward M. K. and Kerr N. S. Clin. Chim. Acta 1982 125 125.Manning D. C. and Slavin W. Appl. Spectrosc. 1983 37 1. Pegon Y. Anal. Chim. Acta 1978 101 385. Manning D. C. Slavin W. and Carnrick G. R. Spectrochim. Acta Part B 1982 37 33 I. Leung F. Y. and Henderson A. R. Clin. Chem. 1982 28 2 139. Leung F. Y. and Henderson A. R. At. Spectrosc. 1983 4 1. Lewis S. A. O'Haver T. C. and Harnly J. M. Anal. Chem. 1985 57 2. Quinonero J. Mongay C. and de la Guardia M. Microchem. J. 1989 39 344. Wawschinek O. Petek J. L. Pogglitsch H. and Holzer H. Mikrochim. Acta 1982 I 335. Pinel R. Benabdallah M. Z. Astruc A. and Astruc M. Anal. Chim. Acta 1986 181 187. Shan X.-Q. Luan S. and Ni Z.-M. J. Anal. At. Spectrom. 1988 3 99. Liu P. and Keiichiro F. Anal. Chim. Acta 1985 171 279. Delves H. T. Buchak B. and Fellows C. S. Aluminium in Food and the Environment Royal Society of Chemistry Cambridge 1988 pp. 52. Markesbery W. R. Ehmann W. D. Alauddin M. and Hossain T. I. M. Neurobiol. Aging 1984 5 19. Mao T. X. Ehmann W. D. Markesbery W. R. Nucl. Instrum. Methods Phys. Rex Sect. B 1987 24125 1003. Eaton D. K. and Holcombe J. A. Anal. Chem. 1983,55,946. Majidi V. Ratliff J. and Owens M. Appl. Spectrosc. 1991 45 473. Paper 1/03 I35J Received June 24 1991 Acceoted Februarv 26 I992

ISSN:0267-9477    

DOI:10.1039/JA9920700749

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 Spectral Interferences on the Determination of Selenium Atomic Absorption Spectrometry* A. Javier Aller and Conception Garcia-Olalla Department of Biochemistry and Molecular Biology University of Leon E-24071 753 by Electrothermal Leon Spain The effects of chlorides and nitrates of magnesium calcium and aluminium on the atomization of selenium from both the tube wall and L'vov platform (microboat) are compared. Different concentrations of these salts were used under the same conditions in order to characterize the shapes of the absorption peaks charring profiles and spectral interferences of selenium. Ashing studies using chemical modification were then performed in order to overcome these chloride spectral interferences. Keywords Spectral interferences; magnesium calcium and aluminium chloride; aluminium nitrate; electro- thermal atomic absorption spectrometry Selenium is an important micro-element in humans and its determination at trace levels in a variety of matrices such as environmental samples and biological fluids is neces- sary.A procedure commonly used in such instances is electrothermal atomic absorption spectrometry (ETAAS). However the determination of selenium using graphite atomizers is subject to interferences which reduce the analyte signal. Non-spectral interferences in ETAAS have been widely They can be due to several different processes (i) gas-phase formation of analyte species which can diffuse from the optical path resulting in loss prior to dissociation into (ii) condensed-phase reaction to form an analyte compound which is volatilized and diffused out of the absorption tube prior to atomi~ation,~ (iii) co-volatiliza- tion or thermal expulsion of the analyte (in a solid liquid or vapour phase) together with rapidly expanding matrix gases or by a carrier (or occlusion) mechanism prior to (or after) reaching the atomization temperat~re~?~ and (iv) changes in the atomization path modifying analyte population~.~J~ However the formation of gaseous analyte monohalides has been reported as the most frequent chemical interfer- ence in ETAASlJJ l 7 l 2 and particularly alkali and alkaline earth chlorides are among the most common and trouble- some interferents in the determination of selenium by The strong depression in the selenium signal caused by some metal chlorides can be explained by a retention of the chlorine during the pyrolysis stage which is later evapo- rated during the atomization stage.I4 Consequently the concentration of chlorine during atomization is high and diatomic molecules of the SeCl type can be formed.This sort of interference is a result of vapour-phase interactions during which the gaseous selenium chlorides diffuse from the optical path prior to atomization. l5 Some explanations of these interferences have suggested that they depend more on the concentration of chloride than on that of the associated metal although influence from both has also been suggested.16 However these gas-phase interferences are only present at very high partial pressures of chlorine in the graphite tube.13J7 Thus a concentration of about 1% metal halide in the solution has been evaluated as an interferent in the atomization of selenium.16 In order to overcome this interference some workers have speculated that either the presence of matrix chlorides with bond dissociation energies higher than that of the analyte chlorides or the use of a long path (long residence ETAAS."J3 *Presented in part at the XXVII Colloquium Spectroscopicum Internationale (CSI) Pre-Symposium on Graphite Atomizer Tech- niques in Analytical Spectroscopy Lofthus Norway June 6 4 1 9 9 1.time) electrothermal ~ystem,~ or higher ashing tempera- t u r e ~ ~ ~ . ~ ~ could remove this type of interference. Other analytical problems related to the determination of selenium using ETAAS are spectral interferences.Some of these interferences have been studied especially by back- ground correction methods using continuum sources or the Zeeman effect. The structural background of line-rich electronic excitation spectra of molecules and the atomic lines of a matrix element are two types of spectral interferences encountered with continuum-source back- ground correction systems,20 but not usually for most Zeeman-corrected systems.21 However the magnetic field applied to the atomization cell produces a splitting of atom lines which might increase the probability of spectral interferences in d i r e ~ t * ~ ~ ~ or i n v e r ~ e ~ ~ - ~ ~ Zeeman AAS. The depressed selenium signal caused by some metal alkaline earth chloride^^^^^^ cannot be explained by the formation of the selenium chloride during atomization.In this paper the mechanism behind this interference is explained using a Smit h-Hieftj e background correct ion system. Experimental Apparatus The AAS measurements were made with a Thermo Jarrel Ash SH 11 spectrometer equipped with a controlled temperature furnace (CTF) Model 188 and a Smith-Hieftje background correction system. A selenium hollow cathode lamp (Visimax XI) was used for atomic absorption measure- ments. Time-resolved absorbance data were displayed with an Epson LX-800 printer. Standard pyrolytic graphite coated graphite tubes were used. For the platform atomization rectangular tubes and a pyrolytic graphite coated graphite microboat were used. The atomization cells were purged with argon (>99.9% by volume). Sample introduction wasperformedwith theFastac aerosol deposition system for the wall atomization and with a micro-syringe for the platform atomization. The furnace parameters listed in Table 1 were used duringmeasurements.Reagents and Standard Solutions Stock solutions of selenium (1000 mg dmW3) magnesium (3000 mg dme3) calcium (3000 mg dm-3) and aluminium (3000 mg dm-3) were prepared from high-purity (Suprapur Merck) salts (chloride or nitrate as necessary) dissolved in the appropriate acid and/or distilled de-ionized water. The acids hydrochloric acid 36% v/v and nitric acid 70% v/v were of Suprapur quality (Merck). Measurements The analyte-matrix element pairs were examined to reveal cases of over-compensation for the wall and platform7 54 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL.7 Table 1 Instrumental settings C l I I 1 I - Spectrometer- Background correction Beam mode Wavelengthlnm Slit-width/nm Lamp current/mA Integration time/s Graphite furnace- Dry Temperature/"C 150 Ramp time/s 2 Hold time/s 0 Purge position I Read - Sampling- Sampling mode Sample volume/nm3 Fastac parameters- Delay time/s Deposition temperature/"C Deposition time/s Smith-Hieftje Single beam 196.0 1 .o 5.8 5.0 Pyrolysis 1 Pyrolysis 2 Atomization Clean 400 500- 1900 2300 2500 0 0 5 0 2 2 0 3 - 20 20 0 - - Integration - Automatic 10 6.0 10.0 150 atomization modes. Thus 10 mm3 of solution (0.0 1 - 1 .O% m/v) of the matrix element were introduced into the atomization cell and the selenium was atomized using the furnace parameters given in Table 1.The parameters shown in Table 1 as pyrolysis 1 and pyrolysis 2 were used in all programmes assuming that pyrolysis 2 is varied from 500 to 1900°C. In order to avoid memory effects a new graphite tube was used for each matrix element studied. The appropriate acids were used as blanks. Results pre- sented in this paper are based on measurements of both peak height and integrated absorbance. Results and Discussion Atomization of Selenium in Different Matrices The influence of some metal chlorides on the signal profile of selenium was studied and the results from the wall atomization are shown in Figs. 1-3. From these figures two types of peaks can be observed positive peaks due to the absorption of selenium and negative peaks due probably to an error in over-compensation in the background correc- tion step as a result of the molecular absorption of some matrix species.A negative peak before the absorption signal of selenium is obtained when magnesium chloride is present together with selenium (Fig. 1). However if calcium rn 0.3 \ 0 1 I p 0.1 - b W 0 0.5 1 .o 1.5 2.0 2.5 Time/s Fig. 1 Absorbance versus time profiles for the wall atomization of A selenium alone (0.5 ng) and in the presence of B 0. I and C 1 .O% magnesium chloride; and D 0.01 E 0.1 and F 1.0% magnesium nitrate. The profile G shows the total absorbance (element plus background) of the signal B. Ashing temperature at 500°C. Note that for C no positive peak appeared or aluminium chloride is added with selenium the negative peaks appear at the same time and after the appearance time of selenium respectively (Figs.2 and 3). In order to characterize the effects of the respective matrix salts on the selenium signal and to propose a mechanism for the interference of these metal chlorides each of the matrix salt solutions was analysed as the nitrate. The results from wall atomization of the pitrates are also shown in Figs 1-3. No negative peaks were observed when equivalent concentrations of either magnesium or calcium nitrate were vaporized with selenium. Important differ- ences are noted for the magnesium and calcium nitrates compared with the chlorides but not for the aluminium salts which show the same negative peaks. The absorp- tion-time profiles presented in Figs.1-3 show important differences in the appearance temperature of the negative peaks (Table 2). Comparable results were obtained for all the metals studied using the platform atomization mode. Nonetheless a delay in the appearance temperature of the negative peaks was noted (Table 2). The reaction mechanisms of both the magnesium and calcium salts are very similar because the nitrates of both metals improve the atomization of selenium whilst their chlorides produce negative peaks. This could be owing to the fact that the nitrate anion modifies the graphite surface I '. F/ \. Fig. 2 Absorbance versus time profiles for the wall atomization of A selenium alone (0.5 ng) and in the presence of B 0.1% calcium chloride; and C 0.01 D 0.1 and E 1.0% calcium nitrate. The profile F shows the total absorbance (element plus background) of the signal B.Ashing temperature 500°C. Note that for B no positive peak appearedJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 755 0.3 UJ \ 2 0.2 e 2 0.1 P 0 0 0 !!? * UJ c. - Fi ' \ V'2.A I 0 0.5 1 .o 1.5 2.0 2.5 Time/s Fig. 3 Absorbance versus time profiles for the wall atomization of A selenium alone (0.5 ng) and in the presence of B 0.1% aluminium chloride; and C 0.01 D 0.1 and E 1 .O% aluminium nitrate. The profile F shows the total absorbance (element plus background) of the signal B. Ashing temperature 500 "C. Note that the positive peak obtained for C overlaps with those for D and E such that the atomization of selenium is improved. How- ever if the influence of the nitrate anion is assumed the signal for selenium would be the same for both metals. But this is not true because the appearance time of the selenium signal is shifted later for calcium nitrate than for magne- sium nitrate for equivalent concentrations.This suggests the active participation of the metal. The explanation of this large delay in the appearance time of the selenium signal in the presence of calcium nitrate compared with magnesium nitrate could be based on the occurrence of an occlusion process. Thus during the pyrolysis stage ther- mally stable CaO and MgO can be formed in which some of the selenium species might be in~orporated.~' However the differences between the effect of magnesium and calcium nitrate could be due to the differences in the MgO and CaO solids formed which would probably act as mini-plat- forms.32 The absorbance-time profile of selenium in the presence of the chloride and nitrate of aluminium is similar for both salts because the negative peaks appeared just after the selenium signal.However another effect complicates the atomization of selenium as aluminium chloride increases the signal intensity for selenium whilst aluminium nitrate decreases it slightly. These responses could be explained by the large reducing capacity of aluminium chloride solutions compared with those of aluminium nitrate. From alumi- nium chlorides the formation of A1203 with the subsequent participation of aluminium chloride species in the atomiza- tion mechanism of selenium is not favoured. This mecha- nism is absent in solutions of aluminium nitrate and a larger occlusion of some of the selenium species could in addition be responsible for the lower signal intensity of selenium. Spectral Interferences The negative peaks observed during the atomization of selenium in the presence of the different metal chlorides could possibly be explained by the formation in the gaseous phase of partly undissociated selenium chlorides.This could be especially true for calcium chloride for which the negative peak appeared at a temperature (Fig. 2 Table 2) similar to that of selenium alone.33 Thus as noted earlier,I4 the chlorine might be available from earlier volatilized material that later condensed on the cooler ends of the furnace tube and then re-volatilized during the atomization of selenium resulting in the formation of a vapour-phase selenium chloride.However selenium is probably not involved in the appearance of these negative peaks for the following reasons (i) with some matrices [MgCI2 AIC13 and AI(N03),] and at certain concentrations both a positive and a negative peak appear in the absorbance-time profile of selenium; (ii) the positive peak due to the atomic absorp- tion of selenium for both selenium alone and selenium in the presence of any metal chloride appears and when it does it is at the same peak time (i,e. time of maximum absorbance) and with a similar intensity; (iii) the intensity of these negative peaks grows with the concentration of metal chloride; and (iv) the appearance temperature of the negative peaks differs for each of the matrices studied corresponding with the boiling-point data and the dissocia- tion energy of the metal chlorides (Tables 2 and 3).Moreover negative peaks do not appear when selenium is atomized in the presence of either other metal chlorides (PdCI2 and HgCI#O or hydrochloric acid33 for the equiva- lent concentrations. Contrarily these metal chlorides en- hance the absorption signal of selenium. This enhancement of the selenium signal is probably a consequence of the lower dissociation energy of these metal-chloride bonds34 compared with the interfering metal chlorides. Thus if chloride can be eliminated during the pyrolysis stage selenium is probably atomized as selenide. In order to understand the participation or otherwise of selenium in the appearance of the negative peaks a solution containing chlorides and nitrates of magnesium calcium or aluminium without selenium were vaporized.Similar negative peaks to those reported in the presence of selenium were obtained from the chlorides of magnesium calcium and aluminium and from the nitrate of aluminium. This indicates that the negative peaks for magnesium and calcium chloride are a consequence of the influence of both the halide and the concomitant metal. The presence of these negative peaks suggests that the cause is a spectral interfer- ence where a strong absorption from matrix compounds exists. Thus although a pulsed source is used for the background correction and this is measured at the same Table 2 Appearance temperature (Tap,) and extinction temperature ( Tex)* of the negative peaks at different ashing temperature ranges for two atomization modes (wall and microboat); number of replicates > 3 Ashing Matrix? Atomization temperature (0.1 Yo) mode range/"C TapdOC Te,I"C MgCl2 Wall 500-700 700 900 Microboat 500-900 900 1300 CaCl Wall 500-900 1000 1300 Microboat 500- 1300 1300 1600 AlCI3 Wall 500- I 100 1100 1300 M i croboa t 500- 1 500 1400 1800 ANN 0 3 1 3 Wall 500- 1 100 I100 1300 Microboat 500- 1 500 1400 1800 *Extinction temperature is the temperature at which the peak returns to the baseline.?A 0.1% matrix was used in each instance.756 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 Table 3 Boiling-points and dissociation energy for the matrix metal compounds* M-CI bond dissociation Temperature energy?/ range of Metal salt Boiling-point/"C kcal mo1-I observation$/"C MgCh 1412 7 6 k 3 1 100- 1 900 CaCI 1600 9 5 k 3 1500-2300 A1203 AlC13 1789 118k3 - SeCI 2887 7 7 k 3 - - 2980 - *See ref. 34.t 1 kcal mol-I =4.184 kJ mo1-l $See ref. 15. §Sublimated. 7 Decomposed. wavelength as the analyte the emission line of the hollow cathode lamp is broadened during the high-current pulse. As the bandpass of the monochromator is larger than the width of the emission line some additional molecular absorption can occur during the high-current pulse at a wavelength very close to the selenium line resulting in an over-compensation error. Therefore it is reasonable to consider that this molecular interference can be avoided by increasing the ashing temperatures (Tables 2 4 and 5).The direct relationship between absorption and temperature suggests that the absorbing species were simple molecules resulting from the dissociation of matrix compounds. These absorption peaks were probably from the MgCl and CaCl molecules that under the experimental conditions used resulted from the vaporization and dissociation of unhydro- lysed metal chlorides during the atomization step. Suppor- tive evidence for MgCl CaCl as the strong absorbing species around the wavelength of selenium is shown e l ~ e w h e r e ~ J ~ ~ ~ ~ and in molecular spectra tables.36 This observation is additionally supported by the fact that the alkaline earth cations do not suffer any hydrolysis in aqueous solutions. Moreover MgC12 can remain on a graphite surface up to temperatures of about 590"C3' (Table 2) whilst CaCl also exhibits a high atomization temperature2 (about 1550 "C).This would also be explained by the high vapour pressures of these gaseous metal chlorides which are directly correlated with temperat~re~~ in the range of 590-1 500 "C. This is based on the fact that the equilibrium constant pressure basis Kp for the reaction MeCl,(g)=Me(g)+ 2Cl(g) is important at temperatures only above 1600°C. As a result these chlorides will probably stay in the gaseous phase until selenium is atomized. The large differences between magnesium and calcium chlorides cannot be due to their similar volatility but probably to different retention times on the graphite surface as is suggested by the experimental results shown in Table 2.It has been suggested that the type of magnesium salt is not ~ i g n i f i c a n t ~ ~ ~ ~ ~ with respect to the volatility of the analyte because all the magnesium salts (chloride and flitrate) are presumably reduced to Mg0.31 This suggestion is based on the fact that MgO is a common intermediate in the decomposition of the ~ h l o r i d e ~ ' ~ ~ at 600 "C However it appears that this reaction does not occur under the experimental conditions used here because the results for the atomization of selenium do not confirm this hypothesis. On the contrary they are in agreement with the results of O t t a ~ a y ~ ~ which showed that magnesium chloride interfered in the atomization of lead. However this interference was interpreted as a consequence of char losses rather than vapour-phase interactions.As a result for the magnesium and calcium salts the interferent metal is just as important as the chloride accompanying it. The negative peaks observed in the atomization of Table 4 Peak times* of the selenium (50 pg dm-3) signals for wall atomization and in the presence of various matrices with and without different chemical modifiers at various ashing temperatures number of replicates > 3 Peak timeh Matrix (0.1%) None MgC12 Mg(N03)2 CaCI Chemical modifier (0.1 O/o) None None HgC12-PdCl2 PdC12 HgClz CdC12-PdC12 CdC12 None None HgC12-PdC12 PdClz HgC12 CdCI,-PdC12 CdC12 None None HgC12 CdCI2 None Ashing temperature/"C 1100 s t s t 0.8 0.9 s t 0.8 s t 0.5 0.8 1.2 1.2 0.8 I .2 s t 1 .o s t 4 st,$ s t 4 st,$ 900 s t 0.7 1.25 1.3 0.8 1.1 0.55 0.9 st,$ 1.6$ 1 .O$ st,$ st$ s t 4 st,$ s t 4 s t 4 1.64 1.4 700 0.9 0.7$ 1.5 I .55 0.9 0.9$ 0.6$ 1.2 1.3$ st,$ st,$ st,$ 1.43 1 .O$ 1 .O$ 0.95$ 1 .O$ 1.3$ 1.6 500 1.I 0.9$ 1.4$ 1.4$ 0.9$ 1.6$ 1 .O$ 1.4 1.9 1.1$ 1.1$ 1.1$ 1.05$ *Peak time is the time of maximum absorbance. ?There were no positive peaks at these temperatures for these matrices. $A negative peak is produced at this temperature in either the presence or absence of a positive peak.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 757 Table 5 Peak heights and integrated absorbances for the wall atomization of selenium (50 pg dm-3) in the presence of various matrices with and without different chemical modifiers at various ashing temperatures; number of replicates > 3 Absorbance* Chemical modifier (0.1 %a) None Hg-Pd Pd Hg Cd-Pd Cd None Hg-Pd Pd Hg Cd-Pd Cd None None Hg-Pd Pd Hg Cd-Pd Cd None None Hg Cd None Ashing temperature/"C 1300 s f 0.0 10 0.005 0.004 0.004 s t 0.004 0.004 st st st st s t st sl- st s? st 0.0 17 0.017 0.02 1 0.026 s t st st s t s t A st,$ st,$ st,$ 1100 st 0.058 0.038 0.046 0.070 0.017 0.030 0.048 0.039 0.015 0.026 0.002 0.004 0.084 0.042 0.077 0.066 s? 0.080 0.072 s t 0.05 1 0.050 0.068 0.050 0.067 0.05 1 0.066 0.04 1 0.065 0.040 0.04 1 0.037 s t 0.048 0.046 st,$ st,$ st,$ st38 900 st 0.062 0.034 0.047 0.085 0.024 0.027 0.055 0.038 0.02 1 0.029 0.0 10 0.0 12 0.109 0.068 0.123 0.092 0.084 0.039 0.104 0.1 14 0.094 0.076 0.169 0.1 19 st,$ 0.0453 0.030 0.043$ 0.028 st,$ 0.040$ 0.05 1 st,$ 0.106 0.034 st,$ st,$ s7-4 st,$ 700 0.026 0.025 0.060 0.039 0.035 0.080 0.023 0.028 0.042 0.030 0.02 1 0.027 0.053$ 0.048 0.1 10 0.073 0.1 12 0.085 0.136 0.080 0.050 0.140$ 0.055 0.1 72 0.075 s t 4 0.038$ 0.020 st,$ s l - 4 0.047$ 0.024 0.0 13$ 0.007 0.074 0.023 0.1 14$ 0.160 0.055$ 0.0 18 0.083$ 0.080 0.0 13$ 0.013 0.101$ 500 0.03 1 0.03 I 0.056 0.04 1 0.03 I 0.073 0.028 0.033 0.036 0.045 0.027 0.034 0.026$ 0.0 I 5 0.095$ 0.053 0.I 003 0.069 0.120$ 0.035 0. 1 004 0.069 0.07 1 $ 0.030 0. I26 0.046 st,$ 0.026% 0.030 s t A st,$ st,$ st,$ 0.048 0.02 1 0.125$ 0.250 0.125$ 0.250 0.1 10$ 0.200 0.020$ 0.025 *The first value in each pair refers to peak height and the second value to integrated absorbance in seconds. ?There were no positive peaks at these temperatures for these matrices.$A negative peak is produced at this temperature in either the presence or absence of a positive peak. selenium in the presence of both aluminium chloride and aluminium nitrate originated in the presence of some volatile absorbing species. These could probably be oxides of aluminium (A10,). This is explained by the fact that both the chloride and nitrate of aluminium are transformed into the aluminium oxide. Thus A13+ in aqueous solution is generally present as a hydrate Al(OH)4- (independently of the anion present) and the following reactions can occur during the dessication step 2AI(OH)4- + 2H+~A1203 + 5H2O 2AlC13*3HZO*A1203 + 6HC1 The aluminum dihydroxide Al(OH),+(g) has also been observed44 at 1573 "C probably produced by the following reaction AlO+(adsorbed) + Hz0sA1(OH)2+ This hypothesis is also supported by both the very small vapour pressures of AlC13 even at high temperatures (> 1500 0C)38 and the very high boiling-pgint of Alz03. Ashing Studies It is known that differences in the volatilities of selenium and the matrix elements will reduce or eliminate interfer-758 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL.7 ences allowing analytical methods to be used for the determination of selenium in the presence of interfering element^.^^-^^ The isolation of the absorption signal of selenium from these interferences can be performed by a combined study of the ashing temperature and chemical modification. Owing to the presence of negative peaks the integrated absorbance measurements are not always usable because the absorbance is balanced between positive and negative peaks during the integration time.As a result peak height measurements could be more adequately used in all instances. In the presence of a negative peak the value of integrated absorbance for the positive peak can be obtained by carefully setting the integration time. The interference from MgC12 decreases at higher ashing temperatures and is completely eliminated at 900 "C (wall atomization) and 1000 "C (platform atomization) (Fig. 4 Table 4). These results confirm the hypothesis that the halides of magnesium (Mg-Cl) are dissociated at this temperature being responsible for interferences. Nonethe- less although at ashing temperatures above 900 "C interfer- ences from magnesium chloride do not exist the effective- ness of this salt for increasing the absorption signal of selenium is poorer than the magnesium nitrate (Table 5).The negative peak due to CaC12 obtained for selenium is eliminated at 1 100 "C (wall atomization) and 1400 "C (platform atomization) (Fig. 5 Table 4). Both the peak height and integrated absorbance of the signal for selenium in the presence of calcium chloride are greater than in the presence of calcium nitrate at this ashing temperature (Table 5). However the peak time and the appearance time of the selenium signal in the presence of calcium chloride is earlier than in the presence of calcium nitrate (Fig. 5 and 0 1 2 3 4 5 Time/s Fig. 4 Absorbance versus time profile for the wall atomization of selenium (0.5 ng) in the presence of 0.1% of A magnesium chloride (ashing temperature 900OC) and B and C magnesium nitrate (ashing temperatures 900 and 1 100°C respectively).Signals D and E are the corresponding background plus analyte signals for magnesium nitrate. Line F represents the temperature profile D y c I \ i t 0 1 2 3 4 5 Timels Fig. 5 Absorbance versus time profile for the wall atomization of selenium (0.5 ng) in the presence of A 0. I % of calcium chloride and B calcium nitrate. Signals C and D are the corresponding background plus analyte signals. Line E represents the temperature profile. Ashing temperature 1 I00 "C Table 4). This is probably due to either the formation of an easily atomizable calcium selenide in the presence of the calcium chloride or the occlusion of the selenium in the CaO lattice formed from calcium nitrate. These results suggest that the effect of calcium on the stabilization of the selenium signal is also affected by the counter ion.The effect of aluminium salts (chloride and nitrate) on the signal for selenium is fairly similar at increasing ashing temperatures. Thus a selenium peak occurs just before the negative peaks produced by the matrix (Fig. 3). However the signal for selenium obtained with aluminium chloride is larger than that produced in the presence of aluminium nitrate (Fig. 3 and Table 5) suggesting a direct influence of the aluminium chloride in the atomization path of sele- nium. The absorption signal for selenium diminished at higher ashing temperatures as a result of pyrolysis losses.The negative peaks appear sooner during atomization and their intensity decreases with ashing temperature disap- pearing at an ashing temperature of 1300 "C (wall atomiza- tion) and 1800 "C (platform atomization) (Tables 2 and 5). Chemical Modification Studies As the negative peaks show comparable behaviour for the two modes of atomization and they are extinguished earlier for wall atomization all the chemical modification studies were performed using this atomization mode. Selenium can be determined in the presence of magne- sium chloride at 500 and 700 "C using peak height measure- ments and any of the following chemical modifiers Hg Cd Pd Hg-Pd and Cd-Pd. However better results are ob- tained with Hg Pd and Cd-Pd at 500 "C or with Cd and Hg at 700°C.Any of these modifiers can also be used at 900-1 100°C to determine selenium in the presence of magnesium chloride using either peak height or integrated absorbance measurements (Table 5). However the best characteristic mass for an ashing temperature of 1100°C was obtained with the Cd-Pd modifier (Table 6) whilst the Hg-Pd modifier showed the best ratio of peak height to integrated absorbance. ~~ ~ Table 6 Characteristic mass (m,) for the wall atomization of selenium (50 pg dm-3) in the presence of various matrices with and without different chemical modifiers at two ashing temperatures; number of replicates >3 Chemical Matrix modifier (0.1%) (0.1%) mJpg temperature/"C Ashing MgWO,) None 44.00 I100 18.49 900 MgCl2 None 550 I100 183.33 900 PdCI2 33.33 1100 23.9 I 900 HgC12 56.40 900 HgC12-PdC12 52.38 1100 32.35 900 CdC12-PdC12 30.55 1100 19.30 900 CdC12 28.95 900 Ca(N03)2 None 33.33 1100 64.70 900 CaCl None 75.86 1100 PdC12 53.65 1100 78.57 900 HgC12 55.0 1100 HgC12-PdC12 43.14 1100 73.33 900 CdC12-PdC12 59.46 I100 43.13 900JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL.7 759 Although the negative peak from calcium chloride inter- ference is eliminated at 1 100 "C (Tables 4 and 5) selenium can be determined at lower ashing temperatures using a chemical modifier and peak height measurements (Table 5). At an ashing temperature of 700 "C a chemical modifier of 0.1 O/o CdC12-PdC12 chloride can be used but at 1 100 "C the best results were obtained by using a mixture of 0.1% HgClZ-PdCl2 as shown by the characteristic mass (Table 6).Selenium can be determined in the presence of aluminium chloride at ashing temperatures of 500 and 700°C using peak height measurements and without any chemical modifier (Table 5). The peak of the selenium signal in the presence of calcium using Pd as a modifier appeared 1.0 s into the atomization cycle after charring at 900 "C (Table 4) whilst no signal for selenium appeared using Hg as a modifier. However in the presence of calcium but using Pd-Hg as modifier the peak for the selenium signal appeared at 1.6 s suggesting that the influence of the chemical modifier upon selenium is not exclusively due to the presence of Pd since Hg also acts in the atomization mechanism of selenium. At an ashing temperature of 1100 "C it can be seen that HgC12 is as efficient as the PdC1,.However using a mixture of both modifiers is more effective than using each alone because the integrated absorbance obtained for the mixture is highef15 (Table 5). Similarly the mixing of CdCl and PdC12 is more efficient than each separately at 700 and 900 "C as has been However this mixed modifier shows less effectiveness than the PdC1 alone at 1 100 "C and in the presence of a calcium chloride matrix. At an ashing temperature of 700 "C the HgC1 modifier is as efficient as PdC1 for the atomization of selenium in the presence of magnesium chloride. At 900"C the most efficient modifier appears to be PdC12 although the Cd-Pd mixture also shows good results for the determination of selenium in the presence of the magnesium chloride.29 At 1 lOO"C the best results are obtained according to the following sequence Cd-Pd Pd Hg-Pd for measurements of integrated absorbance but the sequence giving the best results using peak height measurements is Hg-Pd Cd-Pd Pd (Table 5).As integrated absorbance and peak height can be correlated with the atomization and atomization rates of selenium respectively it can be deduced from Table 5 that both the atomization and atomization rates of selenium differ with modifier matrix composition and ashing tem- perature. As a result for the magnesium chloride matrix with and without any of the modifiers used the best atomization and atomization rate for selenium are usually achieved at 700-9OO0C whilst for the calcium chloride matrix they were best at higher temperatures (900- 1 100 "C).However at the highest possible ashing temperature ( 1 100 "C) the best atomization of selenium in the presence of magnesium chloride occurred in the presence of the CdC12-PdC12 mixture whilst the best atomization rate took place with the HgC12-PdC12 mixture. Nonetheless in the presence of a calcium chloride matrix both the atomization and atomization rate of selenium are better with the HgC12-PdC12 mixture. These results suggest that Hg exhibits a more efficient effect upon the atomiza- tion rate of selenium than Cd does. Comparing the peak height and integrated absorbance from either magnesium chloride or calcium chloride in the presence of any of the modifiers it can be seen that the magnesium chloride favours the atomization of selenium because both values of peak height and integrated absor- bance are larger than those obtained in the presence of calcium chloride.The same results are obtained for a matrix of aluminium chloride for corresponding situations. As a result an increased participation of matrix can be established for the sequence CaC1 <A1C1,<MgC12 in the atomization of selenium in the presence of any mbdifier. However this sequence is reversed in the absence of modifiers showing clearly that the atomization path of selenium is strongly dependent on the composition of the matrix. Conclusion The results presented show that similar spectral interfer- ences are present for both wall and platform atomization. However platform atomization does not show any advan- tage over wall atomization because higher ashing tempera- tures appear to be needed in the former to overcome these interferences.This is because with the temperature pro- gramme given in Table 1 the platform temperature lags behind that of the tube wall and a hold time for the pyrolysis step would be necessary so that the platform could reach the cited temperature. Vaporization of interferent metal chlorides appears to be the predominant mechanism for the interference of the magnesium and calcium chloride. The presence of these vapour-phase chlorides results in a strong molecular ab- sorption with the subsequent appearance of a strong negative peak in the absorption-time profile of selenium. However from the data in this paper the identity of the interferences from the aluminium salts can only be spec- ulation.This interference could be on the basis of the formation of some oxides of aluminium that exhibit a strong absorption at a wavelength very close to the selenium line. Negative peaks can be eliminated by a proper combination of ashing temperature and chemical modifier. As a result chemical modification with wall atomization can successfully eliminate some interferences in the deter- mination of selenium in the presence of chloride-containing matrices. The authors appreciate the financial support of the Conse- jeria de Cultura y Bienestar Social de la Junta de Castilla y Leon under award Ref. 1121/90 during the course of this research. 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 References Fuller C. W. Electrothermal Atomization for Atomic Absorp- tion Spectrometry The Chemical Society London 1977.L'vov B. V. Spectrochim. Acta Part B 1978 33 153. Hageman L. Mubarak A. and Woodriff R. Appl. Spectrosc. 1979 33 226. Erpsamer J. P. and Niemczyk T M. Anal. Chem. 1982,54 538. Welz B. Akman S. and Schlemmer G. J. Anal. At. Spectrom. 1987 2 793. Kantor T. Bezur L. Pungor E. and Winefordner J. D. Spectrochim. Acta Part B 1983 38 581. Churella D. J. and Copeland T. R. Anal. Chem. 1978 50 309. Krasowski J. A. and Copeland T. R. Anal. Chem. 1979 51 1843. Sturgeon R. E. Chakrabarti C. L. and Laqgford C. H. Anal. Chem. 1976,48 1792. Hydes D. J. Anal. Chem. 1980 52 959. Carnrick G. R. Manning D. C. and Slavin W. Analyst 1983 108 1297. Welz B. Schlemmer G. and Mudakavi J. R. Anal. Chem. 1988,60 2567.Frech W. Lundberg E. and Cedergren A. Prog. Anal. At. Spectrosc. 1985 8 332. Slavin W. Myers S. A. and Manning D. C. Anal. Chim. Acta 1980 117 267. Shekiro J. M. Skogerboe .R. K. and Taylor H. E. Anal. Chem. 1988,60 2578. Slavin W. Carnrick G. R. and Manning D. C. Anal. Chem. 1984 56 163. Sedykh E. M. and Belyaev Yu. I. Prog. Anal. At. Spectrosc. 1984 7 373. Yasuda S. and Kakiyama H. Anal. Chim. Acta 1976 84 291.760 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 Yasuda S. and Kakiyama H. Anal. Chim. Acta 1977 89 369. Fernandez F. J. and Giddings R. At. Spectrosc. 1982 3 61. Fernandez F. J. and Beaty M. M. Spectrochim. Acta Part B 1984 39 519. Massmann H. Talanta 1982 29 1051. Kurfurst U. Fresenius' Z. Anal.Chem. 1983 315 304. Wibetoe G. and Langmyhr F. J. Anal. Chim. Acta 1984 165 87. Wibetoe G. and Langmyhr F. J. Anal. Chim. Acta 1985 176 33. Wibetoe G. and Langmyhr F. J. Anal. Chim. Acta 1986 186 155. Carnrick G. R. Barnett W. and Slavin W. Spectrochim. Acta Part B 1986 41 991. Wibetoe G. and Langmyhr F. J. Anal. Chim. Acta 1987 198 81. Garcia-Olalla C. Robles L. C. Alemany M. and Aller A. J. Anal. Chim. Acta 1991 247 19. Garcia-Olalla C. Robles L. C. and Aller A. J. Anal. Sci. 1991 7 61 1. Slavin W. Carnrick G. R. and Manning D. C. Anal. Chem. 1982 54 621. Slovak Z. and Docekal B. Anal. Chim. Acta 1981 130 203. Garcia-Olalla C. and Aller A. J. Fresenius' J. Anal. Chem. 1992 342 70. 34 CRC Handbook of Chemistry and Physics ed. Weast R. C. Boca Raton FL 59th edn. 1978-1 979. 35 Brumbaugh W. G. and Koirtyohann S. R. Anal. Chem. 1988 60 d 05 1. 36 Pearse R. W. B. and Gaydon A. G. Identijcation of Molecular Spectra Whitefriars London 1963. 37 Kantor T. and Bezur L. J. Anal. At. Spectrom. 1986 1 9. 38 Schron W. Spectrochim. Acta Part B 1989 44 965. 39 Manning D. C. and Slavin W. Appl. Spectrosc. 1983 37 1. 40 Slavin W. Manning D. C. and Carnrick G. R. At. Spectrosc. 1981 2 137. 41 Fuller C. W. At. Absorpt. Newsl. 1975 14 73. 42 Smeyers-Verbeke J. Michotte Y. van den Winkel P. and Massart D. L. Anal. Chem. 1976 48 125. 43 Ottaway J. M. Proc. Anal. Div. Chem. Soc. 1976 13 185. 44 Styris D. L. and Redfield D. A. Anal. Chem. 1987 59 289 1. 45 Garcia-Olalla C. and Aller A. J. Anal. Chim. Acta 1992,259 295. Paper 1/03413H Received July 8 1991 Accepted April 13 1992

ISSN:0267-9477    

DOI:10.1039/JA9920700753

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 76 1 Determination of Gallium in Coal and Coal Fly Ash by Electrothermal Atomic Absorption Spectrometry Using Slurry Sampling and Nickel Chemical Modification Shan Xiao-quan,* Wang Wen and Wen Bei Research Centre for Eco-Environmental Sciences Academia Sinica P. 0. Box 2877 Beijing 100085 China A method was developed for the determination of gallium in coal and coal fly ash by electrothermal atomic absorption spectrometry using slurry sampling and nickel chemical modification. In the presence of nickel modifier the integrated absorbance for gallium was improved by a factor of 2.46 and matrix interferences with the absorption signals were eliminated and hence similar atomization profiles were obtained for gallium in slurries and in aqueous standard solutions.Therefore calibration using aqueous standards could be applied. The recommended method was applied to the determination of gallium in real samples. The data obtained by slurry sampling were in good agreement with the certified values or with those obtained after pressure attack of the samples. The relative standard deviations for ten replicate determinations were in the range 2.1-5.7% for gallium in coal and coal fly ash for gallium concentrations of 1.05-59.7O/0 pg g-l. Keywords Gallium determination; coal and coal fly ash; electrothermal atomic absorption spectrometry; slurry sampling; nickel chemical modifier Gallium is an important element used in such high- technology applications as computer chips and fibre-optic systems and its use is expected to increase significantly during the next decade.' Gallium is usually present in a variety of samples in trace concentrations. Therefore electrothermal atomic absorption spectrometry is fre- quently chosen for the determination of gallium owing to its high sensitivity the simplicity of the instrumentation and the relatively low cost.However the method is plagued by serious matrix interferences. In order to eliminate or at least to reduce the matrix interferences various techniques such as the standard additions m e t h ~ d ~ . ~ platform at~mization,~ zirconium- impregnated graphite tube^,^^^ chemical modification7-I1 and organic extraction and back-extraction in combination with the chemical modificationl2 have been reported.In a previous study both nickel and ammonium sulfate were used to modify the matrix and to remove the perchloric acid interferen~e.~ The mechanism of gallium loss during pre-atomization was ascribed to the formation of gaseous oxides of gallium. This was confirmed by mass spectrome- To our knowledge very few studies have been reported on the direct determination of gallium by electrothermal atomic absorption spectrometry using solid ~amp1ing.I~ Solid aluminium oxide was suspended in ethanol-water and the slurry was introduced on to a graphite wall or platform. However aqueous standard calibration was impractical for that particular determination. It is well recognized that slurry sampling for electrother- mal atomic absorption spectrometry offers the following advantages (i) the possibility of omitting the usually time- consuming digestion or decomposition step and hence increasing the speed of the whole analytical procedure; (ii) low contamination risk; (iii) fewer possibilities of analyte losses during the sample pre-treatment or retention by insoluble residues; (iv) no corrosive or hazardous chemicals are used; and ( v ) sample introduction with a micropipette and autosampler.The most critical factor in the slurry sampling technique is probably the requirement to main- tain a stable and homogeneous slurry during the time required for sample introduction. Recently a comprehen- sive review of solid sampling in electrothermal atomic try.13 *To whom correspondence should be addressed. absorption spectrometry was pub1i~hed.l~ The advantages and disadvantages associated with direct solid and slurry sampling were compared. The aim of this study was to develop a method for the determination of gallium in coal and coal fly ash by electrotherma! atomic absorption spectrometry using the slurry sampling technique.Various chemical modifiers were compared in terms of maximum pyrolysis temperature sensitivity enhancement effect and background absorption. The accuracy and precision of the method using aqueous standard calibration were evaluated by the determination of gallium in standard reference materials. Experimental Apparatus A Perkin-Elmer Model 3030 atomic absorption spectrometer equipped with a Model HGA-400 graphite furnace and a Model 56 chart recorder was employed for the measurements of gallium absorbances at the resonance line of 287.4 nm under the conditions of'gas stop' and 'maxi-mum power'.The spectral bandwidth was set at 0.7 nm. A gallium hollow cathode lamp was operated at 15 mA. Deut-erium arc background correction was used throughout. A 20 mm3 Eppendorf micropipette fitted with disposable poly(propy1- ene) tips was employed to introduce sample solutions into the pyrolytic graphite coated graphite tube atomizer. Reagents Gallium stock standard solution 1000 ,ug cme3 was prepared by dissolving 0.100 g of gallium (99.999%) (Shanghai Chemical) in 10 cm3 of 6 mol dm'3 nitric acid. The solution was boiled to expel nitrogen oxide and diluted to 100 cm3 with de-ionized water. Working standard solutions were prepared by appropriate dilution with 0.1 mol dm-3 nitric acid. Nickel and other chemical modifier solutions were prepared by dissolving suitable amounts of compounds (analytical-reagent grade) in 0.1 mol dm-3 nitric acid. All other chemicals used were of analytical-reagent grade.Prepbration of Slurry Amounts of 10-300 mg of coal or coal fly ash depending on the concentration of gallium in the samples were weighed762 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 into a 20 cm3 beaker 10 cm3 of ethanol-water (2 + 8) were added. The contents of each beaker were stirred with a magnetic stirrer during sample analysis. Decomposition of Sample Accurately weigh 0.100-0.300 g of the samples into a 30 cm3 Teflon container add 0.5 cm3 of 0.1 mol dm-3 nitric acid to moisten the sample thoroughly followed by 1.5 cm3 of concentrated nitric acid (67%).Allow the mixture to stand overnight. Add 1 cm3 of concentrated perchloric acid (72%) and 3 cm3 of hydrofluoric acid (40%) and cover the container with a Teflon cover. Place this container in a stainless-steel bomb and seal it with a screw closure to avoid gas leakage. Place the bomb in an oven increase the temperature to 170°C over a period of 0.5 h and maintain this temperature for about 6 h. Remove the bomb from the oven and cool it to room temperature before opening. Remove the Teflon container from the bomb and remove the Teflon cover carefully. Wash the interior surface of the cover with 0.5 cm3 of concentrated nitric acid into the container and heat the latter on a hot-plate at about 120°C until the appearance of perchloric acid fumes then increase the temperature to 160°C until the sample is nearly dry.Finally add a sufficient volume of 0.1 mol dm-3 nitric acid to dissolve the residue with gentle heating and transfer the soution into a 25 cm3 calibrated flask. Repeat this proce- dure several times. Dilute the solution to the mark with 0.1 mol dm-3 nitric acid. Determination of Gallium in the Slurry and Decomposed Solution A volume of 20 mm3 of slurry of decomposed solution obtained by the above pressure attack method was intro- duced into a pyrolytic graphite coated graphite tube atomizer together with the same volume of 1000 ,ug cm-3 nickel modifier solution. The solution was dried at 110°C for 30 s pyrolysed at 1100°C for 30 s and atomized at 2500°C for 5 s using the maximum power mode.The gallium absorbances were measured under the conditions of 'interrupted argon gas flow'. Finally the tube was cleaned at 2650°C for 4 s. Results and Discussion Importance of Nickel Modification In order to eliminate or reduce matrix interferences chemical modificaton in combination with wall or platform atomization has frequently been used in the determinations of gallium in a variety of samples by electrothermal atomic absorption spectromet~y.~-'~ However this technique was only used in work where the solid samples were digested o r decomposed prior to the final determinations. In order to evaluate the analytical merit of chemical modification for the determination of gallium in coal and coal fly ash by electrothermal atomic absorption spectrome- try using the 'slurry sampling technique various chemical modifiers were examined in terms of maximum pyrolysis temperature sensitivity enhancement effect and back- ground absorption for aqueous standards and slurries.The results are given in Table 1. It can be seen that the sensitivity for the determination of gallium in aqueous solution was slightly higher in the presence of cobalt and molybdenum than in the presence of other chemical modifiers. However more serious interferences were en- countered owing to the low pyrolysis temperature of 900 "C. The background absorptions of the chemical modifiers were of the same order except for a very low background absorption of palladium. In the presence of magnesium aluminium and nickel a maximum pyrolysis temperature of 1 100 "C was obtained and the integrated absorbances for gallium were also improved by a factor of more than two. The sensitivity enhancement effect of palladium was less significant.With a slurry of coal fly ash the maximum pyrolysis temperature for gallium was in the range 1000-1200°C in the absence or presence of modifiers. It was not surprising that the background absorptions were much higher owing to the complicated sample matrices. However the background absorption for slurries in the 0.20 I 0 800 1600 Tern peratu re/"C 2600 Fig. 1 Effect of pyrolysis temperature (atomization at 2500 "C) and atomization temperature (pyrolysis at 800 "C) on the absor- bance of gallium in the absence or presence of nickel A 0.8 ng of Ga; B 0.8 ng of Ga + 20 pg of Ni; C 0.8 ng of Ga in 82-20 1 Coal Fly Ash slurry; and D 0.8 ng of Ga in 82-201 Coal Fly Ash+20 ,ug of Ni Table 1 Comparison of various chemical modifiers for the determination of gallium in aqueous standards and slurries Aqueous standard (40 ng cm-3 of Ga) Slurry of coal fly ash (1.4 mg cm-3 slurry =40 ng cm-) of Ga) Chemical modifier None co2 + MoV1 Mg2+ Pd2 + Ni2+ A13+ Concentration of modifier as metal ion/mg cma3 1 .o 1 .O 1 .o L .O 0.1 1 .o - Maximum pyrolysis temperature/ "C 900 900 900 1100 1100 I100 1100 Sensitivity enhancement effect (-fold) 1 .oo 2.45 3.06 2.12 1.97 I .37 2.46 Background integrated absorbancek 0.005 0.045 0.003 0.02 1 0.037 0.005 0.06 1 Maximum pyrolysis temperature/"C 1000 1000 1000 1100 1100 1100 1200 Background integrated absorbance 0.290 0.185 0.200 0.263 0.288 0.242 0.260JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGI presence of cobalt and molybdenum were slightly lower than that in the presence of other modifiers.The reason for this is unclear because even a lower pyrolysis temperature of 1000°C was tolerated and less matrix was removed before atomization and hence higher background absorp- tion was expected. As the highest pyrolysis temperature for gallium was achieved using nickel the chemical modification effect of nickel was studied in greater detail. The effect of pyrolysis and atomization temperatures on gallium absorbances in the absence or presence of nickel is shown in Fig. 1. When no nickel modifier was used and the atomization temperature was set at 2500°C the gallium absorbance increased with increase in the pyrolysis temper- ature from 200 to 600"C then levelled off in the range 600-900°C and finally decreased with further increase in pyrolysis temperature.When a pyrolysis temperature of 800 "C was used the absorbances for gallium increased with atomization temperature from 1400 to 1800 "C and almost constant absorbances were obtained over the range 1800-2600 "C. In the presence of nickel a higher pyrolysis temperature of 1 100 "C was obtained and the sensitivity for the determination of gallium was improved by a factor of 2.46. The atomization temperature was also shifted to higher values. When a coal fly ash slurry was introduced into the graphite furnace there was a dip in both the pyrolysis and atomization curves if no nickel modification was used.In contrast these dips disappeared when nickel was used as a chemical modifier. The slurry matrix can strongly influence the shape of the 0.40 0.20 . ... . . .. 0.20 . . .* *.* ... 0 1.5 Time/s 3.0 Fig. 2 Comparison of absorption signals for gallium using nickel as a chemical modifier. In each instance the symbols apply to A decomposed solution; B slurry; and C aqueous standards and A' B' and C' apply to the corresponding background absorbance. (a) NIST SRM 1633a Trace Elements in Coal Fly Ash A 57.0; B 59.7; and C 60.0 ng ~ m - ~ of Ga. (b) Uncertified coal fly ash A 57.0; B 57.0; and C 60.0 ng cmV3 of Ga. (c) NIST SRM 1635 Trace Elements in Coal (Sub-bituminous) A 32.0; B 45.0; and C 40.0 ng cm-3 of Ga.(d) 82-201 Coal Fly Ash A 28.0; B 40.0; and C 40.0 ng ~ 3 1 1 ~ ~ of Ga ST 1992 VOL 7 763 gallium atomization signal. This effect makes it difficult to use aqueous standards to quantify the atomic absorption signal. As was pointed out above all chemical modifiers enhanced the gallium absorbance to various extents. Ad- ditionally larger or smaller amounts of these elements were present in coal and coal fly ash samples. Therefore calibration using aqueous standards for slurry sampling was obviously impractical and the use of a chemical modifier was necessary in order to match the absorption signals produced by the analyte when in an aqueous standard and when in thefslurry. If similar atomization profiles can be obtained then calibration using aqueous standards is reliable.This calibration method is simpler than the standard additions method or calibration using standard slurries. The atomization profiles produced by gallium in pure aqueous standards and slurries of coal and coal fly ash and the decomposed solutions are shown in Fig. 2(a)-(d). Curves A B and C represent the absorption signals of gallium in the decomposed solutions slurries and aqueous standards respectively. Fig. 2(a) and (6) indicate that similar atomization profiles were obtained for gallium in slurries of National Institute of Standards and Technology (NIST) Standard Reference Material (SRM) 1633a Trace Elements in Coal Fly Ash and uncertified coal fly ash and in aqueous standard solutions. Owing to the slightly lower gallium concentration in the decomposed solutions of these samples curlres A in Fig.2(a) and (6) were slightly lower but they were very similar to curves B and C. For 82-201 Coal Fly Ash (Research Centre for Eco-environmental Science China) the absorption signals of gallium in the slurry and in the decomposed solution were almost identi- cal. However the absorption signal for gallium in the decomposed solution was shifted to the right along the time axis. This phenomenon was also observed for NIST SRM 1635 Trace Elements in Coal (Sub-bituminous) where the absorption signal of gallium in the slurry appeared earlier than gallium in the decomposed solution. The reason for this shift was probably due to perchloric acid interferen~e.~.'~ It was not surprising that the background absorption decreased in the order sample slurry > decomposed sample solution > pure aqueous standard.For the sample slurry no chemical pre-treatment was applied and more matrices remained before atomization so the greatest background absorption was observed. Silicate and organic compounds should be expelled after pressure attack with a mixture of nitric acid perchloric acid and hydrofluoric acid and medium background absorption was obtained. The back- ground absorption for pure aqueous gallium standards was probably due to the nickel modifier and therefore the least background absorption was observed. It should be pointed out that all background absorptions were under the correc- tion capability of the deuterium arc background corrector used. In order to verify further the possibility of using aqueous standards for calibration a statistical comparison was made for two calibration graphs constructed with aqueous stan- dards and various amounts of slurry of 82-201 Coal Fly Ash.The results indicated that the linearity ranges inter- cepts slopes and correlation coefficients for the two calibration graphs were similar. This comparison demon- strated that no serious interferences from slurry matrices occurred and the application of aqueous standard calibra- tion was feasible for the determination of gallium in coal and coal fly ash if nickel chemical modification was used. The results of this study indicate that the use of chemical modification is extremely important. Only in the presence of nickel modifier was the sensitivity for the determination of gallium improved by a factor of 2.5 and were the matrix interferences eliminated and hence similar absorption7 64 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL.7 Table 2 Determination of gallium in real samples Gallium concentration/pg g-' This work Sample Slurry* NIST SRM 1633a Trace Elements in 59-t-2.1 Coal Fly Ash Coal Fly Ash (uncertified) 57.8 -t 1.29 82-201 Coal Fly Ash (China) 28.3 -t 0.6 NIST SRM 1623 Trace Elements in Coal (Sub-bituminous) 1.05 -t- 0.06 *Average values of ten replicate determinations ? 0. After decomposition 57.9 57.8 55.2 56.6 28.4 28.8 1.04 1.06 Information Reference value value Ref. 58 58.7 9 56.3 - 28.1 9 28.3 1.05 - - signals of gallium in slurries and aqueous standards were obtained. Therefore calibration for slurry sampling can be accomplished by using aqueous standards. Determination of Gallium in Real Samples As there were no matrix interferences affecting the shape of the atomization curve for gallium in slurries in the presence of nickel modifier the direct determination of gallium in coal and coal fly ash with slurry sampling was practical.The recommended method was applied to the determination of gallium in standard reference materials and the results are summarized in Table 2. Good agreement was achieved between the data obtained in this work and the certified values or reference values reported by other researchers or between the values obtained by slurry sampling and injection of decomposed solution. The relative standard deviations were in the range 2.1-5.7% for gallium in coal and coal fly ash at gallium concentrations of 1.05-59.7 ,ug g-? References 1 Davidson R. A. Harbuck D. D. and Hammargreu D. D. At. Spectrosc. 1990 11 7. 2 Barron D. C. and Haynes B. W. Analyst 1986 111 19. 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Nakamura K. Fujimori M. Tsuchiya H. and Orii H. Anal. Chim. Acta 1982 138 129. Hasegawa S. Kobayashi T. Hirose F. and Okochi H. Bunseki Kagaku 1987 36 37 1. Kugo K. Bunseki Kagaku 1981 30 529. Kugo K. Ooyu S. Kitazume E. and Tsujii K. Bunseki Kagaku 1984 33 E29. Botha P. V. and Fazakas J. Anal. Chim. Acta 1984 162 413. Mandjukov P. B. and Tsalev D. L. Microchem. J. 1990,42 339. Shan X.-q. Yuan Z.-n. and Ni Z.-m. Anal. Chem. 1985,57 857. Takekawa F. and Kuroda R. Talanta 1988 35 737. Hayashibe Y. Kurosaki M. Takekawa F. and Kuroda R. Mikrochim. Acta 1989 2 163. Clark J. R. J. Anal. At. Spectrom. 1986 1 301. McAllister T. J. Anal. At. Spectrom. 1990 5 171. Marecek J. and Synek V. J. Anal. At. Spectrom. 1990 5 385. Bendicho C. and de Loos-Vollebregt M. T. C. J. Anal. At. Spectrom. 1991 6 353. Koirtyohann S. R. Glass E. G. and Lichte F. E. Appl. Spectrosc. 1981 22 35. Paper I /06 1 92E Received December 10 1991 Accepted March 3 1 I992

ISSN:0267-9477    

DOI:10.1039/JA9920700761

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 765 Systematic Investigation of Aluminium Interferences on the Alkaline Earth Elements in Flame Atomic Absorption Spectrometry Part I. Behaviour of Beryllium Werner Luecke Institute of Petrography and Geochemistry University of Karlsruhe 75 Karlsruhe I Kaiserstrasse 12 Germany The flame atomic absorption spectrometric determination of Be using a dinitrogen oxide-acetylene flame in the presence of Al suffers interference from the latter at 10 mg dm-3. To remove this interference the absorbance of typical test solutions containing 0.5 1 .O 2.0 mg dm- of Be and increasing Al concentrations with a Be:Al mass ratio up to 1 :4000 was systematically examined together with specific reagents that form stable compounds selectively with A1 in the flame.Excess additions of KF LaCI ethylenediaminetetraacetic acid and quinolin-8-01 showed that quinolin-8-01 is the most effective agent to compensate for the Al interference thereby releasing Be for AAS. Keywords Flame atomic absorption spectrometry; beryllium determination; aluminium interference; dinitrogen oxide-acetylene flame; quinolin-8-01 suppressant When using an air-acetylene flame the number of metals detectable by atomic absorption spectrometry (AAS) is limited to about 30 because the flame conditions with temperatures of about 2200 "C are not sufficient to convert thermally stable compounds of any further elements into the ground state Mo required for AAS. To overcome this restriction in element determination Amos and Willis' introduced the hotter dinitrogen oxide-acetylene flame (about 29OO0C) so yielding a higher heat capacity and a strongly reducing atmosphere in the red feather part of this flame where the oxygen partial pressure from the decom- posing dinitrogen oxide is lower than that of the oxygen in the air-acetylene flame. The alternative gas mixture is able to atomize into Mo about a further 20 elements which form in the cooler air-acetylene flame together with e.g.oxygen aluminium and silicon refractory compounds. The possibil- ity of extending the determination range to about 65 metallic elements was therefore the basis for the general acceptance of the flame AAS method. Despite these favourable starting conditions studies of element determinations in different matrices soon revealed that in the flames applied in AAS chemical and physical interferences occur preventing the quantitative thermal dissociation of many elements.2 This usually leads to a decrease in element absorption compared with solutions of the same element concentration but without an interfering bulk as is commonly the case with simple aqueous calibration solutions.In analytical geochemistry these interference problems in AAS also arise because of the different bulk chemistry of the silicates and have not yet been completely examined quantitatively. They can be localized and explained only with the help of investigations on systematically mixed element combinations. Different matrix interferences re- sulting from the different compositions of rocks influence e.g.the concentration data measured for trace element^.^^^ Chemical and physical interferences on elements of the alkali metal group are attributed to the presence of other elements in alkali metal rich matrices of geological ma- t e r i a l ~ . ~ This study is concerned with chemical interferences on elements of the alkaline earth group; this part considers 'Be Part I1 will cover Mg and Ca6 and Part 111 Sr and Ba. As the investigations were related to geochemical aspects the influence of A1 which causes the most important chemical interferences in geological samples was therefore studied in detail. The well known interference of Si on the alkaline earths was not considered as in the generally applied hydrofluoric acid decomposition technique it is converted into silicon tetrafluoride and evaporates and hence is no longer present in dissolved silicate material. Experimental and Results Measurements of Be in the Dinitrogen Oxide-Acetylene Flame The strong binding of Be to oxygen and the high thermal stability of Be0 (melting-point ~ 2 5 3 0 "C) and possibly also of BeOH7 indicate that for AAS with an air-acetylene flame the degree of atomization of the Be is relatively insig- nificant 1000 mg dm-3 of Be yield an absorbance of only about 0.13 and the characteristic concentration is about 30 mg dm- of Be for 1% absorption. However with a dinitrogen oxide-acetylene flame the heat capacity causes a high rate of volatilization of the sub- pm particles generated by the aerosol so that under optimum reducing conditions in the red burning zone the Be atoms are easily liberated from the solid compound.In this flame Be shows excellent sensitivity with a character- istic concentration of 0.02 mg dm-3 for 1% absorption for the spectrometer used. Similarly to the behaviour of the other alkaline earth metals in the air-acetylene flame the determination of Be is negatively influenced by A1 even in the hotter dinitrogen oxide-acetylene flame.' At first in the presence of only small amounts of Al this interference on Be is small but with increasing A1 concentrations at a constant Be concen- tration the decrease in absorbance accelerates significantly. This is shown in Table 1 for test solutions containing Be and increasing A1 concentrations of up tQ 8000 mg dm-3. In the literature there are different opinions about the mecha- nism of this interference in the flame.*-lo Taking into account the well known chemical influence of A1 on Mg and Ca in the air-acetylene it must be presumed that it is the same for Be that thermally stable beryllium aluminates are produced in the dinitrogen oxide-acetylene flame being noticeable at A1 concentrations above about 10 mg dm-3.The corresponding refractory Be-Al-0 phases which ori- ginate from these aluminates may extend from Be0.A1203 (the spinel BeAl,04) to at least BeO.3Al2O3," with Be:Al mass ratios of 1:6-1:18 (see Table 1). The formation of766 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 Table 1 Decrease in Be absorbance in the dinitrogen oxide-acetylene flame in the presence of increasing A1 concentrations.The absorbances (in O/o) are related to the corresponding non-interfered absorption of Be without A1 (= 100%). The registered absorbance of Be is a function of the absolute Al content there is no relationship between the absorbance and the corresonding Be:Al mass ratio 0.5 mg dm-3 of Be 1.0 mg dm- of Be 2.0 mg dm- of Be Al/mg dm- Be:AI Absorbance (Yo) Be:Al Absorbance (Yo) Be:AI Absorbance (O/O) 0 l:o 1 oo* l:o 1 oot l:o 1 OO$ - - - 0.5 1:l 100.5 - 2.5 1:5 99.0 - - - 1:l 99.5 - - 1 2 5 1:lO 99.5 1:5 99.5 - - - 1 1 99.4 - - - - - - 10 20 25 50 100 200 250 500 1000 2000 4000 8000 - - 1:50 1:lOO - - 1:500 1:lOOO 1 :2000 1:4000 - - *Gives an absorbance of 0.206= 100°/o. ?Gives an absorbance of 0.402 = 1009'0. +Gives an absorbance of 0.780= 100%.- - 96.1 97.1 - - 94.2 90.8 74.8 61.2 - - 1:lO - - 1:50 1:lOO - - 1500 1:lOOO 1:2000 1:4000 - 98.3 - - 97.0 96.5 - - 91.5 76. I 61.9 57.2 - 1:5 1:lO - - 1:50 1:lOO - - 1:500 1:lOOO 1:2000 1:4000 97.4 95.6 - - 94.5 92.2 - - 75.8 61.9 57.2 50.4 these new compounds removes absorbing Be atoms from the flame and leads to a decrease in the Be absorbance. With further increase in A1 concentration (Al>>Be) the forma- tion of different A1-0 phases not containing any Be (A10 AlO A120 A1202 and A1203 11) may occur. This increase in refractory A1-0 compounds in the sample matrix causes an additional negative influence on the Be absorption due to physical interferences. The steep decrease in the Be absor- bance above about 500 mg dm- of A1 could be explained in this way (see Table 1).Contrary to the results of Amos and Willis' and Bokow- ski,8 Be can be detected free from interferences only in the presence of A1 at concentrations lower than about 10 mg drn-,. With respect to the present AAS determination range of 0.5-2.0 mg dm- of Be and in relation to Al-free reference solutions (asumed 100% absorbance see Table 1 ) the decrease in absorbance is not a consequence of the actual Be concentration or of the corresponding Be:Al mass ratios but is simply a function of the increasing A1 content in the flame. Different Be concentrations with a constant amount of A1 in solution yield the same degree of interference on Be (Table 1). Consequently Be can be determined accurately if there is a sufficiently high dilution of the interfering A1 concentration which may however bring the Be content of the sample below the AAS detection limit.Therefore valid methods to remove the interference of A1 on Be e.g. in geological or ceramic materials had to be tested in order to establish an efficient method to compensate for A1 contents over several orders of magni- tude. Discussion of the Methods Applied to Remove the Interference of A1 on Be To overcome the interference of A1 on the determination of Be many investigations by different workers have already been performed but with limited success. Four promising attempts were tested here and partly re-examined under modified conditions. It was the aim of this study to reach the suppression of A1 by excess additions of a specific reagent to the solutions the consequent reactions in the flame converting the interfering A1 into unreactive Al- rqagent compounds and so releasing Be as MO.Test series were established with additions of F- ions9 with the aim of masking A1 as a thermally stable fluoro complex; additions of La as LaC1 with respect to its well known releasing function removing chemical interferences on e.g. Mg or Ca6 in the cooler air-acetylene flame; additions of ethylenediaminetetraacetic acid (as Na,EDTA) with the aim of isolating A1 and Be as individual chelate complexes thereby protecting Be because of the separate formation of stable A1 chelate compounds in the flame; and with the same aim additions of quinolin-8-01.'~ The addition of F- ions (e.g. as KF) proposed by Ramakrishna et al.,9 was not successful in removing A1 interferences in alkaline earth metal solutions.The absor- bances of three series of Be concentrations (0.5 1 .O and 2.0 mg drn-,) with increasing A1 content representing Be:Al mass ratios of l:O 1:I l:lO 1:100 1:lOOOand 1:4000 were compared with the absorbances of the same solutions containing additionally 10 g dm- of KF. The test solutions with KF which were expected to be at least at the beginning free from A1 interferences (because of masking of A1 in the [A1F6I3- complex) yielded the same disadvanta- geous results as the solutions in Table 1 without KF. Excess additions of La ions cause a change in equilibrium in the original salt composition during the evaporation of the sample aerosol in the flame. As La shows similar chemical behaviour to the alkaline earth elements this reaction presumably leads to the production of stable LaAlO thereby releasing Be.The process can be described generally by the following equation Be0 + A1203 + 2LaC1 + 2H20=BeC1 + 2LaA10 + 4HC1 The addition of excess of LaC1 shifts the chemical equilibrium to the right and forces the production of thermally unstable BeCl (melting-point 405 "C). With respect to the LaC1 releasing solution the scheme for the non-equilibrium reaction that occurs is assumed to be flame oxygen xBe+yAl+zLaCl xBeC1 + yLaAlO,+(z- y)LaCI,JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 767 trations to a certain extent. The addition of e.g. 5% of La (Table 3) yields on the one hand the known disadvantage of a decrease in absorbance for Be because of the high La content (about 25% in the test series in Table 3) but on the other an extension of the A1 suppression up to 500 mg dmv3 of Al.Because of the excellent Be sensitivity a dilution to (0.5 mg dm-3 of Be is recommended thereby lowering high A1 concentrations possibly below 500 mg drn- together with the required amount of La releaser. Additions of EDTA to remove A1 interferences have been reported,2 but there have been few examinations of this method with regard to quantitative and element-specific aspects. According to Schwarzenbach and Flaschka,'* EDTA binds the di- and trivalent metals (including A1 and the alkaline earths) as chemically stable water-soluble and undissociated inner complexes (chelates) isolating the individual cations with an organic protecting cover.When EDTA is present in sample and calibration solutions in the evaporating sample spray the formation of refractory compounds between the alkaline earth metals and A1 in the flame is hindered. The chelate complex is thermally destroyed by the flame and releases Be as MO to be determined immediately. Possibly only in a higher part of the flame outside the HCL beam may the formation of Be-A1 oxides occur To 50 cm3 flasks containing 25 50 and 100 pg of Be (10.5 1.0 and 2.0 mg dm- of Be) and increasing amounts of Al 10 cm3 of a 0.5 mol dm-3 EDTA solution were added which corresponds to a maximum binding capacity of about 137 mg of Al neglecting the very small amounts of Be. Table 4 shows that this EDTA concentration is sufficient to remove the A1 interference up to at least 1000 mg dm-3 of A1 (= 50 mg of A1 in 50 cm3).For the 2.0 mg dm-3 Be series and at an A1 concentration of 2000 mg dme3 (= 100 pg of Be+100 mg of A1 in a 50 cm3 flask) the absorbance increased only by 86% with respect to a corresponding Be solution without Al; this shows that the EDTA concentra- tion applied is no longer adequate although still being present slightly in excess. Whereas the addition of 10 cm3 of 0.5 mol dm-3 EDTA to 50 cm3 flasks e.g. 25 pg of Be+250 mg of A1 (=0.5 mg dm-3 of Be+ 5000 mg dm-3 of Al) resulted in a relative Table 2 Decrease in Be absorbance (in Oh) in the dinitrogen oxide-acetylene flame in the presence of increasing A1 concentra- tions. An addition of 1% of La compensates for the negative A1 interference on Be up to an A1 concentration of about 200 mg dm-j (cf Table 1) Absorbance (Oh) 0.5 mg dm-3 1.0 mg dm-j 2.0 mg dm-3 Al/mg dm-j Be+ 1% La Be+ 1% La Be+ 1% La 0 0.5 1 2 2.5 5 10 20 25 50 100 200 250 500 1000 2000 4000 8000 1 oo* 99.5 1 oot 100.3 - 1 OO$ - - 99.9 _.- 100.0 100.0 100.3 100.0 - 99.9 100.0 100.5 100.5 - 100.5 100.8 - 100.0 99.7 - 95.5 84.9 73.9 65.3 - 87.1 74.4 65.9 60.7 - 75.0 66.8 62.0 56.7 *Gives an absorbance of 0.199= 100%. ?Gives an absorbance of 0.387= 100%. $Gives an absorbance of 0.744 = 100%. with z>>y>x. Table 2 shows the absorbances (in Yo) of the three Be series in Table 1 with increasing A1 content but now containing additionally 1% of La as dissolved LaCl,. A series of tests with Be and different concentrations of La revealed the following results.The addition of 1% of La generally induces a decrease in absorbance of 4% in the absence of A1 because of the increase in the total salt concentration in the sample (ie. bulk interferences). In contrast to the disadvantageous results in Table 1 the addition of 1% of La as releaser achieves complete suppression of A1 interferences up to about 200 mg dm-3 (Table 2). Higher A1 concentrations are not compensated for by the addition of 1% La. If the A1 content is known by preliminary determinations to exceed the 200 mg dm-3 limit larger amounts of La can still remove these A1 concen- ~ ~~ Table 4 Decrease in Be absorbance (in O/o) in the dinitrogen oxide-acetylene flame in the presence of increasing A1 concentra- tions. An addition of 0.1 mol dm-' EDTA compensates for the negative A1 interference on Be up to an A1 concentration of about 1000 mg dm-3 (cf Table 1) Table 3 Decrease in Be absorbance (in O/o) in the dinitrogen oxide-acetylene flame in the presence of increasing A1 concentra- tions.An addition of 5% of La compensates for the negative A1 interference on Be up to an A1 concentration of about 500 mg dm-3 (cf Table 1 ) Absorbance (Yo) 2.0 mg dm-3 Be+ EDTA 1 OO$ - - 100.0 - - 99.8 - - 100.3 - - 85.8 47.0 - 0.5 mg dm-3 1.0 mg dmA3 Al/mg dm-j Be+ EDTA Be+ EDTA 0 1 oo* 1 oot 0.5 99.4 - 100.0 - 1 Absorbance (O/O) 0.5 mg dmd3 1.0 mg dm-3 2.0 mg dm-j AVmg dm-3 Be+ 5% La Be+ 5% La Be+ 5% La - - 100.0 - - 99.7 - - - 100.0 - 100.3 99.4 - - 99.7 57. I - - 45.4 - - - - L 5 10 20 50 100 200 500 1000 2000 5000 6000 0 50 100 200 250 500 1000 2000 4000 1 oo* 1 oot 1 OO$ 100.6 - - - 100.3 - 100.2 100.6 - - 100.0 99.7 - - - 97.5 96.3 95.0 71.3 69.7 69.2 - 55.6 53.8 *Gives an absorbance of 0.1 57 = 100%.?Gives an absorbance of 0.300 = 100%. $Gives an absorbance of 0.5 78 = I 00%. *Gives an absorbance of 0.154= 100%. ?Gives an absorbance of 0.3 1 5 = 100%. $Gives an absorbance of 0.600= 100%.768 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 Table 5 Decrease in Be absorbance (in O/o) in the dinitrogen oxide-acetylene flame in the presence of increasing A1 concentra- tions. An addition of 2% of quinolin-8-01 compensates for the negative A1 interference on Be up to an A1 concentration of about 2000 mg dm-3 ( c j Table 1) Absorbance (%) 0.5 mg dm-3 1.0 mg dm-3 Be + 2% Be + 2% Al/mg dm-3 quinolin-8-01 quinolin-8-01 0 0.5 1 2 2.5 5 10 20 25 50 100 200 250 500 1000 2000 4000 8000 1 oo* 100.5 - - 100.0 100.5 - - 100.0 100.5 - - 100.5 100.0 100.0 100.5 - - 1 oot 99.7 - - - 100.0 99.7 - - 99.5 100.0 - - 99.7 100.0 99.7 91.4 - *Gives an absorbance of 0.190= 100%.?Gives an absorbance of 0.372= 100%. $Gives an absorbance of 0.705= 100%. 2.0 mg dm-3 Be + 2% quinolin-8-01 1003 99.7 - - - - 99.9 99.7 - - 100.0 100.1 - - 100.1 99.3 90.5 52.9 absorbance of only 57% the addition of 20 cm3 of 0.5 mol dm-3 EDTA (binding capacity about 270 mg of Al) yielded an enhanced absorbance of 80%. However investi- gations with even larger amounts of EDTA were not successful particularly as deposits of pyrolytic carbon in the burner slit strongly interfered.Therefore dilutions of Be concentrations down to (0.5 mg dm-3 and correspond- ing small EDTA additions to remove the A1 interference can be recommended. Finally the chelating agent quinolin-8-01 which was applied many years ago by Fleet et al.I0 to overcome A1 interference on Be was examined under the present test conditions. The element combinations in Table 1 were prepared together with 2% of quinolin-8-01 from a stock solution of 100 g of quinolin-8-01 in 250 cm3 of HCl ( 1 + 1) diluted to 1000 cm3 with water (= 10%). The results of Be determinations are given in Table 5. Comparing the data in Tables 2-5 it can be seen that quinolin-8-01 is the most effective in suppressing A1 interferences for up to 2 mg dm-3 of Be the presence of 2000 mg dm-3 of A1 can be compensated for quantitatively by quinolin-8-01 additioqs. Only in the presence of quinolin-8-01 in the Be solutions does a bulk-related interference with a low absorbance depression occur in the present instance of 8-9% when compared with the corresponding Be absorbance of the solutions without A1 in Table 1.Conclusion Lanthanum EDTA and quinolin-8-01 which are well known as releasing and protecting agents in AAS to remove interferences on alkaline earths such as Mg or Ca in an air- acetylene flame can also be successfully applied to compen- sate for remaining A1 interferences on Be in a dinitrogen oxide-acetylene flame. However only excess of these reagents in solutions causes bulk effects in the flame and forces the Be absorbance to decrease; 1% of La and 2% of quinolin-8-01 yielded decreases of about 4 and 8-9% respectively and 5% of La and 0.1 mol dm-3 EDTA resulted in a decrease in the Be absorbance of about 23-25% each in relation to Be solutions free from A1 and without additions of removal agents.The releaser addition of 1% of La and the protector addition of 2% of quinolin-8- 01 were found to be effective in suppressing the A1 interference on Be. Tests with higher La or quinolin-8-01 concentrations (> 5 and >2% respectively) to extend the A1 compensation range were unsuccessful because in addition to an increased negative bulk effect on the Be absorbance the added substances tended to form deposits in the burner slit preventing correct measurements. Based on the results presented here accurate AAS determination of Be in the presence of A1 can be achieved by applying the following steps.The Be and A1 contents of the sample solution are roughly determined by a routine method. If the Be concentration is >2 mg dmP3 the sample solution should be diluted to about 0.5 mg dm-3 thereby effecting a proportional decrease in the A1 concentration also. If now the approximate Be:Al mass ratio is > 1:4000 an accurate determination of Be with addition of 2% of quinolin-8-01 can be performed (see Table 5); if the Be:Al mass ratio is > 1:400 the addition of 1% of La is also recommended (see Table 2). If in the Be concentration range 0.5-2.0 mg dm-3 the Be:Al mass ratio is <1:4000 (ie. >2000 mg dm-3 of A1 in the sample solution) chemical separation methods must be applied e.g gradient elution chromatography where A1 can be fixed on the stationary phase resin and Be separately released by acid washing. The author thanks B. Welz for critical discussions of the manuscript. 1 2 3 4 5 6 7 8 9 10 1 1 12 References Amos M. D. and Willis J. B. Spectrochim. Acta 1966 22 1325. Welz B. Atomabsorptionsspektrometrie Verlag Chemie Weinheim Deerfield Beach and Basle 3rd edn. 1983 p. 527. Luecke W. Chem. Geol. 1977 20 265. Luecke W. Fresenius' Z. Anal. Chem. 1987 329 57 1. Luecke W. Chem. Erde 1979 38 1. Luecke W. Fresenius' 2. Anal. Chem. in the press. Alkemade C. Th. J. in Flame Emission andAtomic Absorption Spectrometry ed. Dean J. A. and Rains T. C. Marcel Dekker New York and London 1969 p. 10 1. Bokowski D. L. At. Abs. Newsl. 1967 6 97. Ramakrishna T. V. West P. W. and Robinson J.W. Anal. Chim. Acta 1967 39 8 1. Fleet B. Liberty K. V. and West T. S. Talanta 1970 17 203. Barin I. Thermochemical Data of Pure Substances Parts I and IZ VCH Weinheim 1989 p. 1739. Schwarzenbach G. and Flaschka H. Die kompleuometrische Titration Enke Stuttgart 1965 p. 339. Paper 2/0039 1 K Received January 24 1992 Accepted March 18 I992

ISSN:0267-9477    

DOI:10.1039/JA9920700765

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 769 Sample Introduction and Atomization of Volatile Alkyllead Compounds in Flame Atomic Absorption Spectrometry Gyula Bagdi Janos Lakatos and lstvan Lakatos" Research Laboratory for Mining Chemistry Hungarian Academy of Sciences P.O. Box 2 H-35 15 Miskolc- Egy e tem va ros Hung a ry Flame atomic absorption spectrometric (FAAS) analysis of organic liquids containing volatile organometallic compounds is discussed. lsobutyl acetate acetone and a light hydrocarbon mixture (gasoline) were used as solvents and tetramethyllead (TML) and tetraethyllead (TEL) served as model compounds. Detailed analysis of the sample introduction process revealed that the analyte reaches the flame mostly as a vapour. The amount and composition of the vapour change significantly with the uptake rate.At low uptake rates the sample completely evaporates whereas the composition of the vapour phase gradually approaches the value determined by the actual liquid-vapour equilibrium at higher uptake rates. The differences in efficiency of sample introduction and hence in the analytical signals of different lead compounds can be explained by their relative volatilities. The behaviour of volatile lead compounds in the nebulizer and the atomizer allowed a rapid and accurate FAAS method to be developed for the determination of total lead content and the concentrations of TML and TEL in light hydrocarbon mixtures. Keywords Flame atomic absorption spectrometry; sample introduction; solvent effect; alkyllead compound determination; hydrocarbon analysis The difference between the analytical responses in the determination of the lead content in gasoline containing different alkyllead compounds as antiknock agents e.g.tetramethyllead (TML) or tetraethyllead (TEL) is well Therefore conventional flame atomic absorption spectrometric (FAAS) techniques can be applied only when standards with a certified alkyllead composition are used for calibration. Therefore in the past various methods have been proposed in order to eliminate this problem e.g. destruction of the matrix and then dissolution of the ash or its extraction with an aqueous solvent direct introduction of samples after treatment with i ~ d i n e ~ emulsification of the samples treated with or without iodine6y7 and applica- tion of combined analytical systems such as gas chromato- graphy-AAS8 or high-performance liquid chromatogra- phy-AAS9 to separate and enrich the analyte from the bulk sample.Mauri and co-workersI0.l I developed a method based on the direct volatilization of samples. Despite the succesful determination of lead in gasoline they could not explain properly the analytical problem with volatile compounds. This phenomenon was explained theoretically by Bratzel and ChakrabartiI2 who taking the results of Kashiki et aL5 and Trent3 into account concluded that the analytical problem can be traced to the complete evaporation of volatile lead compounds from the sample in the spray chamber. Later similar conclusions were drawn by Shimomura et a1.I3 based on the determination of GeC1 in strongly acidic media and they also demonstrated the anomalies with the introduction processes for samples containing highly vola- tile analytes.They established that the main feature of sample introduction is the enhancement of the 'efficiency of nebulization' viz. the sample reaches the flame mainly as vapour and the atomizaton take place virtually without solid particulate formation in the flame. This coincides with another ob~ervation,'~ which revealed a significant discre- pancy in the atomization process when samples containing organic (alkyl) and inorganic [e.g. Pb(N03)2] lead com- pounds are analysed using identical aerosol flow rates and flame properties. The objectives of this work were to demonstrate the ~~ *To whom correspondence should be addressed.irregular behaviour of volatile lead compounds in the sample introduction process using different solvents and to determine the effects of vapour- and aerosol-type sample introduction on the analytical signals. In addition the experimental results offer a new possibility of developing a relatively simple analytical method. Experimental A Zeiss AAS- 1 atomic absorption spectrometer equipped with a concentric nebulizer was used. The nebulizer was operated with an air flow of 750 dm3 h-l. Acetone isobutyl acetate and a 1 + 1 mixture of regular and unleaded gasoline were used as solvents. When nebulizing lead-containing solvents the absolute amount of lead (Fpb cm3 min-') reaching the flame as TML vapour was determined indi- rectly from the waste (Fw cm3 min-I).The amount of sample introduced into the flame as an aerosol (Fa cm3 min-l) was obtained by separate experiments using a non-volatile organometallic compound (copper oleate) which was readily soluble in all the solvents and a single- stage impactor combined with a glass-fibre reinforced filter (Whatman GF/F r=0.5 pm).Is The waste (F,) was deter- mined by measuring its mass while F (total amount of sample reaching the flame as vapour in cm3 min-l) was calculated from the mass balance of the sample introduc- tion process F = Fa + F" + F where F is the rate of nebulization in cm3 min-l. As the atomization of Pb(N03)2 and alkyllead com- pounds had to be compared under identical conditions (flame stochiometry) a dual nebulizer system was devel- oped.16 The flow rates of air and acetylene were 430-500 and 100 dm3 h-' respectively.To enhance the resolution of the analytical signal a diaphragm with a narrow angular aperture was introduced into the optical path. The analyti- cal signal was measured at the Pb I 2 17.0 nm line between 0 and 20 mm above the burner and between + 5 and - 5 mm from the centre of the flame. A single-slot burner (50 mm long and 1 mm wide) was employed in all experiments. When the atomization of alkyllead compounds was studied one nebulizer introduced TML- or TEL-containing solutions with a nebulization rate770 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 (F,=0.27 cm3 mine') that ensured complete evaporation of both the analyte and the solvent in the mixing chamber at ambient temperature.Simultaneously distilled water was nebulized by the second nebulizer at Fl=2.0 cm3 min-l. A different arrangement was used to study the atomization characteristics of inorganic lead compounds. In this in- stance pure organic solvent (e.g. acetone) was introduced by the first nebulizer and an aqueous solution of 5Opg cme3 Pb(N03)2 by the second nebulizer using the same flow rates as in the previous tests. The absolute amount of lead introduced into the flame was obtained by the following products acetone solutions Flcalkyllead; and aqueous solu- tions FacrlMNO>) where c is the concentration in pg ~ m - ~ . The contribution of the second nebulizer to the total sample flow was obtained as mentioned earlier. Results and Discussion Nebulization and Atomization of Alkyllead Compounds The different liquid fluxes (Fw Fa and F,) are illustrated for all the solvents in Figs. 1-3 as a function ofthe nebulization rate.Here the absolute amount of analyte (f'pb TML) reaching the flame as vapour can also be compared directly with other flow fluxes. The starting point of all the curves (f'pb) corresponds to a specific condition viz. when waste appears in the sample introduction process. Consequently below this nebulization rate both the solvent and the 0 2 4 6 8 F,/cm3 min-' Fig. 1 Sample TML in isobutyl acetate. See text for definitions Dependence of flow rates on rate of nebulization F,. 0 2 4 6 8 FJcm3 min-' Fig. 2 Sample TML in acetone. See text for definitions Dependence of flow rates on rate of nebulization Fl.volatile alkyllead compound completely evaporate in the mixing chamber ( F p b = F l ; broken lines in Figs. 1-3). On the one hand it is characteristic that the total amount of sample reaching the flame is virtually independent of the nebuliza- t'ion rate. On the other hand it is surprising that F p b is always greater than Fa even though Fa changes with the solvent type. These results prove the fundamental role of vaporization in the sample introduction process as against the conventional aerosol formation and transport model. As the efficiency of the analyte transport with the alkyllead compounds is basically controlled by the distillation pro- cess in the nebulization chamber the primary aerosol distribution has a negligible effect on the analyte flux reaching the flame.If the 'distillation' mechanism is accepted as the primary factor then it is reasonable that the relative enhancement of different alkyllead compounds in the atomization source reflects their relative volatilities. In addition the nebulization efficiency (and the primary aerosol distribution) is only slightly influenced by the type of organic solvent on account of their similar surface tensions. As expected the amounts of vaporized analyte obtained with the different solvents which increase in the order isobutyl acetate<acetone<gasoline reflect the relative volatilities of the solvents. (It is probable that the high pentane content and the lower cooling effect of evaporation jointly lead to the highest F value with gasoline; the heats of evaporation are 521 and 368 J g-' for acetone and pentane respectively.) Iff' is compared with Fpb it can be seen that F p b is lower than F for gasoline and higher for the other two solvents.It is also characteristic particularly for gasoline where F p b decreases with increase in F that F p b is determined not only by the sum of Fa and the vaporized solvent. The phenomena is clearly demonstrated in Fig. 4 where the ratio of the total amount of analyte ( F p b ) to the sum of Fa and F is plotted against the rate of nebulization. As can be seen the analyte is distilled into the flame not at a constant rate but according to the relative volatility of the alkyllead compound with respect to the given solvent. That is why the ratio is greater than unity for the less volatile isobutyl acetate whereas the analyte concentration in the vapour phase is smaller for gasoline which is rich in low-boiling components (the boiling-points are pentane 36.2 "C ace- tone 56.5 "C isobutyl acetate 1 16.5 "C and TML 1 10°C).Using a low nebulization rate waste does not form viz. Fpb = F = Fa+ F In this instance Fpd(Fa+FV) is equal to unity (broken line in / / 1 I I I I J 2 4 6 8 F,/cm3 min-' Fig. 3 Dependence of flow rate on rate of nebulization Fl. Sample TML in gasoline. See text for definitionsJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY. AUGUST 1992 VOL. 7 77 1 2 2' urn ' u" n 0 2 4 6 f/cm3 min-' Fig. 4 Dependence of the Fpb:Fa+Fv ratio on rate of nebulization F using different solvents A isobutyl acetate; B acetone; and C a 1 1 mixture of regular and unleaded gasoline Fig.4). However the composition of the vapour leaving the mixing chamber gradually approaches the value determined by the liquid-vapour equilibrium as F increases. This trend can be explained by more effective liquid rinsing of the nebulizer accompanied by the fact that Fa is virtually independent of Fl. The phenomena observed on nebulization of both TML- and TEL-containing solutions are similar except that less TEL than TML reaches the flame. Generally the Fpb versus Fl relationship is identical with that for the TML-gasoline system even though TEL is less volatile than either of the solvents or of TML. Hence the sample introduction of TEL- containing solutions resembles the aerosol-type process. Accordingly the major portion of sample-containing volatile analytes is introduced into the flame as vapour completely mixed with the working gases.In contrast the sample appears in the atomization source through aerosol separation if the analyte is a non-volatile compound [e.g. Pb(NO,),]. The analytical consequence of the latter pheno- menon is the relative enrichment of atoms in the central zone of the flame which was confirmed theoretically by L'vov et all7 Consequently there will be a significant difference between the analytical signals of volatile (TML) and of non-volatile [Pb(NO,),] analyte-containing solutions (in other words vapour- or aerosol-type sample introduc- tion occurs) despite the fact that the same nebulizer and burner are used and the same amount of analyte reaches the flame.The two chemical forms of lead may serve as excellent models to demonstrate the effect of aerosol separation on the distribution of atoms in flames. Vapour- and aerosol- type sample introduction were studied with the dual nebulizer described previously. l6 Using this experimental arrangement the lateral distributions of analyte determined for Pb(N0,) and TML at identical Fa but at different observation heights are illustrated in Fig. 5. The character- 0.50 0 C 0.25 s 2 0 -5 0 5 Lateral d ist ri but ion/mm Fig. 5 Lateral distribution of lead absorption at different observa- tion heights A 3; B 5; C 10; D 15; and E 20 mm using ( a ) Pb(N03)2 and (b) TML as analyte istic parameters of the curves namely the integrated area below the curves (T in cm2) the half-width of the peak (a in cm) and their ratio (are,) and the relative absorbances measured in the central zone (A,,) are listed in Table 1 .In the evaluation of the experimental results however not only the effect of aerosol separation but also the difference in the atomization of Pb(N03)* and TML should be considered. It is clearly seen that the maximum absorbance is recorded for Pb(N03)2 not in the central zone (observa- tion height h = 3 mm) but at higher observation points. It can be concluded that the difference in the local atom concentration measured at low observation points ( h t 5 mm) is influenced jointly by aerosol separation the rate of aerosol vaporization and diffusion. As the relative absor- bance in the central zone is higher than unity the experimental results suggest a rapid transformation of aerosol and the effectiveness of aerosol separation at low observation heights.However it is curious that the inte- grated absorbance which reflects the total atomization in the flame is greater for Pb(N03)2 than for TML. This finding can probably be explained by the axial flow of aerosol in the central flame zone which offers a favourable condition for atomization whereas in vapour-type sample introduction the formation of free atoms is suppressed by the intense recombination process at the edge of the flame. As shown by the data in Table 1 only vapour exists above a certain flame height in both instances. Therefore the dynamic parameters of the flame have the same effect on the distribution of atoms independent of the chemical form of the analyte (ore changes slightly with h).The negligible decrease in Arel can probably be attributed to diffusion. All these phenomena are well demonstrated in Fig. 6 in which the relative lateral absorption profiles are shown at different observation heights. It is obvious that despite diffusion the aerosol separation is still considerable at h=20 mm. Table 1 Parameters of the absorption profiles at different observation heights hlmm TTMLlcm2 TPb(N0,)I/Cm2 ~ T M L I C ~ uPb(NO,)I/Cm crel* A r e i t 3 29.6 33.0 5.3 3.5 0.66 1.40 5 36.8 45.0 6.5 5.0 0.77 1.53 10 40.4 49.0 7.2 6.6 0.92 1.33 15 44.6 52.8 8.0 7.2 0.90 1.29 20 47.6 55.2 8.6 8.0 0.93 1.22772 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 2 { g 1 z P 0 C I -5 0 5 Lateral distribution/mm Fig.6 Lateral distribution of relative absorbance (APMNOl1]ATML) at different observation heights A 3; B 5; C 10; D 15; and E 20 mm. Volumetric flow rates air 930; and acetylene 100 dm3 h-' Similarly the absorption profiles were recorded for both TML and TEL. As expected there was no measurable difference between the curves thus demonstrating the similarity of the atomization. Determination of Total Lead Content and Alkyllead Composition of Gasoline The efficiency of sample introduction is determined by the volatility of the solvent (and ofthe solute) and the distillation process taking place in the mixing chamber. If the sample to be analysed contains a cation with different ligands (TML and TEL) the efficiency of sample introduction for indivi- dual analytes will also be different and the measured analytical signal will be an additive function of composition at a constant rate of nebulization.At an appropriately low nebulization rate however the efficiencies of sample introduction becomes identical because of the complete evaporation of both the solvent and the analyte. The plots in Fig. 7 demonstrate that the dependence of the analytical signals obtained for TML and TEL on the rate of nebulization can be utilized for the determination of the total lead content. Below a nebulization rate of 0.65 cm3 min-' the analytical graph is virtually independent of the chemical form of the analytes and is linear with the nebulization rate. At a certain point however the graphs deviate from each other.The graph obtained for TML remains linear because evaporation of TML from the solvent is complete. Most of the TEL however condenses into the waste resulting in a gradually decreasing analytical signal. Hence the total lead content can be easily deter- mined at low nebulization rates whereas measurements at high nebulization rates can be used for calculation of the alkyllead composition of the sample. In order to perform the analysis an analytical graph taken at F, with TML or TEL is required for the determination of total lead content and two calibrations at F132 with TML and TEL are needed to obtain the alkyl composition (Fig. 7). If the calibration graphs in Fig. 8 are used for analysis the corresponding alkyllead concentrations are calculated with the following equations CTML = A -ct ~ T E L ~ T M L - ~ T E L and CTEL = ct - CTML where cTML is the concentration of lead bound to TML (in 0.50 n qL 0.25 0 FJcm3 min-' Fig.7 Dependence of the analytical signal on rate of nebulization F using solutions of isobutyl acetate containing A TML and B TEL 0.3 0.2 0.1 0 5 10 15 c/pg cm-3 Fig. 8 Calibration graphs obtained for A TML and B TEL using isobutyl acetate as solvent ,ug ~ m - ~ ) c is the total lead concentration (in pg ~ m - ~ ) mTML is the slope of the calibration graph obtained at a nebulization rate Fl,2 for TML and m T E L is the slope of the calibraton graph obtained at a nebulization rate F,,2 for TEL. The analytical performance of the proposed method was evaluated for isobutyl acetate solutions with different alkyllead concentrations.The experimental conditions given in Table 2 were found to be optimum. The actual lead content of the standard samples was determined by a certified procedure (MSZ 1 1 735).18 The values are given in the top part of Table 3 and serve as a basis for comparison. According to the measured data it can be concluded that the method has excellent precision for the determination of total lead content. The RSD is however less favourable when the alkyllead concentration is determined by the proposed technique. This can probably be explained partly by the additivity of errors and partly by less efficient sample introduction of TEL at the given nebulization rate. Table 2 Analytical conditions for the determination of total lead content and alkyllead composition of gasoline Acetylene flow rate Air flow rate Observation height 15 mm Slit width 0.05 mm Rate of nebulization 3.0 cm3 min-' Fl I 0.5 cm3 min-l 1.6 cm3 min-l 80 dm3 h-I (0.1 MPa) 750 dm3 h-I (0.1 MPa) f i .2JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 773 Table 3 Determination of total lead content and alkyllead composition of isobutyl acetate Composition of standard solution- Certified concentration/pg cm-3 Sample ct 1 10.75 2 10.50 3 10.25 Anal-vtical results- CTML 2.50 5.00 7.50 Concentration obtained/pg cm-3 CTEL 8.25 5.50 2.75 Sample ct RSD* (O/O) cTML RSD* (%) c + E L ~ RSD* (O/O) 1 10.85 0.93 2.42 3.2 8.33 0.97 2 10.36 1.33 5.13 2.6 5.37 2.36 3 10.17 0.78 7.98 6.4 2.27 17.45 *n = 5. t&L = ct - cTM L. Table 4 Determination of total lead content and alkyllead composition of two commercial gasolines and a synthetic blend by the proposed FAAS technique; values are given f SD for n= 5 Lead contentlpg cmd3 Gasoline sample ct CTM L CTEL Composition- No.86 372 f 6 Unknown No. 92 385 f 7 Unknown Synthetic 600flO 200+3 40027 No. 86 375 + 7 - 375 & 12 No. 92 3 9 0 f 8 285f10 105&5 Measured composition- Synthetic 605f 12 229k 10 376+ 15 The analytical method was finally tested on different commercial gasolines bought in Hungary and a synthetic blend of unleaded gasoline and pure alkyllead compounds. The total lead contents of the commercial gasolines and the synthetic sample determined by the standard ASTM proce- dure,I9 are given in Table 4. As the samples had high lead contents they were diluted with isobutyl acetate prior to FAAS analysis.Hence significant matrix effects between the pure isobutyl acetate and isobutyl acetate-gasoline test series were not found and the results were comparable. As shown by the data in Table 4 the accuracy of the analysis is the same and the practical reliability of the proposed method is further demonstrated. In conclusion a rapid labour-effective and direct analyti- cal method has been developed utilizing the selective vaporization of different alkyllead compounds. The proce- dure is similar in principle to the method elaborated by Mauri et al.," but it is simpler and does not require the use of a heating unit as an evaporator. The phenomena discussed in this paper are relevant to similar analytical methods for the analysis of samples containing volatile organometallic compounds.Conclusions Volatile lead-containing analytes reach the flame mostly as vapour. The nebulization rate significantly influences the characteristics of sample introduction. At. low nebulization rates the volatile analytes completely evaporate in the mixing chamber and reach the flame exclusively as vapour; at high uptake rates the composition of the vapour phase gradually approaches the value determined by the liquid-vapour equilibrium. The differences in efficiency of sample introduction can be explained by the relative volatilities of TML and TEL. The difference in the analytical signals for the experimental conditions have been explained theoretically. The distribu- tions of atoms in the flame arising from volatile and non- volatile analytes are different.In the latter instance aerosol separation does not affect the lateral distribution of analytes in the atomization source. There is no measurable difference in the atomization of TML and TEL. The total analytical signal (integrated absorption profile) for inor- ganic lead compounds was found to be greater than for alkyllead compounds. Greater recombination at the edge of the flame is probably responsible for the phenomena when vapour-type sample introduction is achieved. The behaviour of volatile alkyllead compounds in the nebulizer and the atomizer made it possible to develop a rapid efficient and accurate method for the determination of total lead content and alkyllead composition in a light hydrocarbon mixture.References 1 Robinson J. W. Anal. Chim. Ada 1961 24 451. 2 Mostyn R. A.,andCunningham A. F.,J. Inst. Pet. 1967,53,10 1. 3 Trent D. J. At. Absorpt. Newsl. 1965 4 348. 4 Epstein M. S.,'At. Spectrosc. 1983 4 62. 5 Kashiki M. Yamazoe S. and Oshima S. Anal. Chim. Acta 197 I 53 95. 6 Berenguer V. Guinon J. L. and de la Guardia M. Fresenius' 2. Anal. Chem. 1979 294 416. 7 Diez L. P. Mendez J. H. and Penalva F. P. Analyst 1980 105 37. 8 Coker D. T. Anal. Chem. 1975 47 386. 9 Massman J. D. and Rains T. C. Anal. Chem. I98 I 53 1632. 10 de la Guardia M. Mauri A. R. and Mongay C. J. Anal. A t . Spectrom. 1988 3 1035. 1 1 Mauri A. R. de la Guardia M. and Mongay C. J. Anal. At. Spectrom. 1989 4 539. I2 Bratzel M. P. and Chakrabarti C. L. Anal. Chim. Acta 1972 61 25. 13 Shimomura S. Sakurai H. Morita H. and Mino Y. Anal. Chim. Acta 1978 96 69. 14 Dagnall R. and West T. S. Talanta 1964 11 1553. 15 Lakatos J. Lakatos I. and Bagdi Gy. in Proceedings ofXXXI Magyar Szinkepelemzo Vandorgyiiles Szolnok ed. Szilvassy Z . University Press Veszprem VeszpCm 1988 p. 21 I . 16 Bagdi Gy. Lakatos J. and Lakatos I. Acta Chim. Hung. 1991 128 319. 17 L'vov B. V. Katskov D. A. Kruglikova L. P. and Polzik L. K. Spectrochim. Acta Part B 1975 31 49. 18 Vajta L. and Kerenyi L. Koolajipari Vizsgalatok Muszaki Kiado Budapest 1978. I9 American Society for Testing and Materials (ASTM) Standard Test Methods for Lead in Gasoline by Atomic Absorption Spectrometry ASTM D-3237-79 p. 12 I . Paper 1/05062A Received October 4. I991 Accepfed March I O I992

ISSN:0267-9477    

DOI:10.1039/JA9920700769

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 775 Determination of Ultratrace Amounts of Gold in Geological Materials by Arc Emission Spectrometry Using Gas Chamber Profile Electrodes Xu Peiqing Central South Institute of Metallurgical Geology 1 7 Victory Road Yichang People's Republic of China A new method is described for the determination of ultratrace amounts of gold in geological samples. The method is based on arc emission spectrometry using gas chamber profile electrodes that have been developed in our laboratory. A liquid reagent (NHJ) was successfully used as a spectrochemical carrier. The technique involves separation of gold by sorption on a foamed plastic column. The detection limit of the method is 0.2 ng g-I and the limit of quantification is 0.5 ng g-l.The concentration of gold in the analysed samples ranged from 0.5 to 50 ng g-I. Studies on the analytical precision showed that the relative standard deviation of this method is 5% (3,-50 ng g-I) and 20% (0.5-3 ng g-'). Results obtained for gold in two Chinese geological standard samples were in good agreement with the certified values. The technique is characterized by its simplicity rapidity high sensitivity and improved analytical precision and accuracy. Keywords Ultratrace gold determination; arc emission spectrometry; gas chamber method; analytical precision and accuracy; sorption on foamed plastic column Gold in geological samples can be determined by a variety of techniques. Fire assay is employed to effect the separa- tion and preconcentration of gold prior to analysis by atomic absorption spectrometry (AAS) inductively coupled plasma or direct current plasma emission spectrometry (ICP-AES or DCP-ES) or ICP mass spectrometry (ICP-MS).Aqua regia (HC1-HNO3 3+ 1) or HBr-Br extraction followed by solvent extraction using isobutyl methyl ketone with a variety of finishes being another Although these methods have their merits considerable time is required for the extraction and chemical preparation steps. Various methods of spectrochemical analysis have been applied to the determination of ultratrace amounts of gold in geological When determining gold in exploration samples results are usually needed immediately in the field and a technique that offers high sensitivity and low cost is required. In this paper the characteristics of an arc emission spectrometric method using gas chamber profile electrodes are described.Gas chamber profile electrodes were developed some years ago in this laboratory9 and constitute the basis of the determination of ultratrace amounts of gold in geological samples. As gold is usually present at very low concentrations in a complex matrix a concentration and separation step is necessary prior to determination. Many methods have been described for the sorption of gold including ion exchange or coprecipitation with tellerium1° and anion exchange.l 1*12 In recent years there has been increased interest in the use Fig. 1 Apparatus used for the method of continuous sorption Table 1 Conditions of spectral analysis Spectrograph 2 m plane grating Slitwidth 25 nm Arc source Current 15 A Analytical line 267.59 nm D.c.with anode excitation Exposure time 3 s of sulfhydryl cotton fibre (foamed plastic) in China. The mechanism of sorption can be roughly divided into two types dynamic sorption and static sorption. In general 100% sorption cannot be achieved. In order to maximize recoveries a method of continuous sorption using a foamed plastic column is adopted employing the apparatus shown in Fig. 1. The continuous sorption onto a foamed plastic column is an innovation in this study. Experimental Reagents A series of standard solutions containing 5 2.5 1.0 0.5 0.05 and 0.01 pg 1-I of gold in 8% aqua regia was prepared. All acids were of analytical-reagent grade and were diluted with distilled water. A solution of 8% NHJ was used as the spectrochemical carrier.Apparatus A 2 m plane grating spectrograph (Beijing WSP-1) with a dispersion grating of 600 mm-1 was used. The conditions of spectral analysis are shown in Table 1. Sample Dissolution A 2 or 8 g portion of finely ground sample (400 mesh) was roasted in a porcelain dish (6.0 x 3.0 x 1.5 cm) for 1 h at 450 "C. The temperature was raised to 650 "C and maintained for a further 2 h. If the sample has a high organic and sulfide content it must be roasted for 3 h at 700 "C in a muffle furnace. The sample was cooled and transferred into a 250 ml glass beaker and 30 ml of aqua regia were added with mixing. The beaker was placed on a hot-plate for 30 min and the solution in the beaker was evaporated to about 10 ml.Distilled water was added and the solution was stirred then allowed to settle before sorption of the gold.776 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 Table 2 Detections limits obtained using various analytical techniques Analytical technique* Detection limithg g-I Nebulization? -1CP-MS 0.5 ETAAS? 0.4 ETV-ICP-MS$ 0.2 Proposed method 0.2 *See ref. 13. 92 g of sample in 10 ml of solution. $2 g of sample in 2 ml of solution. The foamed plastic was placed in a U-shaped glass tube and the sample solution poured into the sorption apparatus and caused to pass back and forth through the foamed plastic plugs to facilitate the continuous sorption of gold. The foamed plastic was then transferred into a 10 ml porcelain crucible which was heated to burn off the plastic and then the residue was ignited.Aqua regia (1 ml) was added and the mixture was heated on a water-bath. The solution was evaporated nearly to dryness. A 0.2 ml aliquot of 8% v/v aqua regia was added to the residue. A 0.05 ml amount of this sample solution was then pipetted into the electrode cavity in duplicate and dried prior to spectro- graphic analysis. Results and Discussion The determination of ultratrace amounts of gold in geologi- cal materials is one of most difficult analytical tasks. This work shows that two main factors influence sensitivity precision and accuracy. The first factor is the character of the gas chamber. As far as we know the sensitivity of the gas chamber method is the highest of any of the spectroche- mica1 methods described in literature being comparable with ICP-MS with electrothermal vaporization (ETV) (Table 2).Various shapes of electrode for arc emission spectrometry have been discussed in the literature. The shape adopted for the upper electrode is usually that of a pointed cone or a platform. The shape of the lower electrode varies according to the preference of individual workers. The selection of electrode shape is very important in the gas chamber method. The design of the electrode used in this work is shown in Fig. 2. This design has not previously been reported in the literature and offers the advantage that it can reduce the wandering of the arc around the surface of upper electrode. In this way the reproducibility and the intensity of spectral lines can be improved because the analytes volatilized from the lower electrode remain in the arc for a longer time increasing the chance of excitation. Usually solid powder is used as a spectrochemical carrier; in contrast NH41 solution is used in this work.It is dropped into the electrode cavity and then dried. After such pre- conditioning the sample solution is dropped into the cavity and again dried. When an arc is propagated between the electrodes the analyte rapidly volatilizes into the arc column. The total exposure time is only about 5 s (see Fig. 3) thus reducing spectral backgrounds and consequently enhancing sensitivity. The second factor that influences accuracy and precision is the mass of sample taken for analysis and the distribution of gold grains within the sample.In this respect the homogeneity of gold grains within a sample is not usually well known. This causes difficulties in analytical precision and accuracy. An initial study has shown that to reduce these problems the sample needs to be crushed to grain sizes of 0.04 mm. In order to achieve this a newly developed rod mill technique was used. One grain of gold i 5 t - I I 6 1 8 T Fig. 2 Design of the electrode used; dimensions in mm * o 1 2 3 4 5 Time/s Fig. 3 Graph of signal intensity versus time for Au (60 ng g-l) Fig. 4 One grain of gold (0.4 mm in diameter) divided into about 70 gold particles using the rod mill method (0.4 mm) added into gangue mineral could be divided into about 70 gold particles by the rod mill method (Fig. 4). Results for the determination of gold in three replicate samples show good agreement for general samples pro- cessed to a grain size of 0.03 mm and siliceous samples processed to 0.01 mm (Table 3).If the sample is homoge- neous satisfactory determination of gold can be obtained using a 2 g aliquot as shown in Table 4. It is apparent from these data that the rod mill method grinds and disperses discrete gold grains present thus overcoming effects of sample inhomogeneity. l 4 However very little work has been undertaken here on samples ground to a more normal mesh size of 200-250 to identify the mass of sample required for analysis; these points have not been addressed here. When analysing soils pre- treatment with HF of the 10 g sample mass was ineffective in increasing the amount of gold leached.However at the 1 g level complete recovery was accomplished with HF-aqua regia and indeed aqua regia alone was much moreJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 777 Table 3 Results for the determination of gold in general samples. A sample mass of 8 g was used n=3 Amount of gold determinedjng g-' Size of sample grain/mm Sample matrix 0.01 0.03 0.05 0.07 Dispersion halo 9.0 8.5 13.5 9.0 8.5 7.5 7.0 5.0 8.5 - Granite 7.5 7.5 6.5 6.5 5.0 5.5 9.0 4.0 6.0 - Carbonic slate 11.0 16.0 14.5 13.5 16.0 11.0 22.5 14.0 11.0 - Carbonati te - 48.0 50.0 - 25.0 30.0 32.0 ~ ~~~~~~~~~~~~~ ~ ~ Table 4 Determination of gold using different amounts of sample n=3 Measured value/ng g-' Amount of sample/g Size of 2 4 8 value f SD*/ 0.0 1 7.5 8.0 9.0 10.0 11.0 7.0 8.5 6.5 8.0 8.0 2 0.3 Certified sample grain/mm ng g-' *SD = standard deviation.~~ ~ ~ Table 5 Results for two certified reference materials Concentration of gold Concentration (certified value) found/ SD/ Sample t- SD/ng g-' ng g-' ng g-' RSD* (O/O) GBW07208 8.0 f 0.3 7.5 0.4 5.0 GBW07209 51 -+3 49.0 2.0 3.9 *RSD = relative standard deviation. effective (75-100%) at this reduced sample mass. It has been proposed that the major sources of error at the 10 g mass level or greater are caused by non-wetting of the samples inefficient mixing during digestion and the protec- tion of gold by unattacked gangue;I2 similar conclusions are drawn from this work. The proposed method was applied to two certified reference materials gold ores GBW07208 and GBW07209 (Geological Standard for Gold and Silver China) and the results show good agreement with the certified values (Table 5).Conclusions The proposed method for the separation of gold by sorption onto a foamed plastic column has been successfully em- ployed for the determination of gold in geological materials with high sensitivity good analytical precision and accu- racy. If the sample is homogeneous a 2 g aliquot gives satisfactory accuracy and precision. Detection limits are better than 0.5 ng g-l. The author thanks Wang Jiaen who carried out the sampling processing. References I Diamantatos A. Talanta 1987 34 736. 2 Hall G. E. M. and Bonham-Carter G. F. J. Geochem. Explor. 1988 30 255. 3 Meier A. L. J. Geochem. Explor. 1980 13 77. 4 Brooks R. R. and Naidu S. D. Anal. Chim. Acta 1985 170 325. 5 Xu P. Fenxi Huaxue 1982 10 61 1. 6 Vall G. A. Voddudnayn L. V. Yudeleviya I. G. and Zolotov Yu A. Z. Anal. Chern. 1979 34 885. 7 Opiev E. Ya Ognev V. R. and Karotada I. Ya Zavod. Lab. 1985 6 27. 8 Date A. R. and Gray A. L. Spectrochim. Acta Part B 1985 40 115. 9 Xu P. unpublished work. 10 Hall G. E. M. Valve J. E. Coope J. A. and Wetilan E. F. J. Geochem. Explor. 1989 34 157. 1 1 Sengupta J. G. Talanta 1989 36 651. 12 Ascensio Trancoso M. and Barros J. S. Analyst 1989 114 1053. 13 Hall G. E. M. Pelchat J. C. and Dunn C. E. J. Geochern. Explor. 1990 37 1. 14 Koratacva I. Yu Kusakina L. V. ZanHava V. Z. and Lontish S. V. Zavod. Lab. 1979 1 66. Paper I /03223B Received June 27 I991 Accepted January 7 I992

ISSN:0267-9477    

DOI:10.1039/JA9920700775

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 779 Abell Ian 53 Abollino Ornella 19 Adamik Miklbs 707 Ajayi Olubode O. 689 Akman Siileyman 187 Aller A. Javier 753 Al-Maawali Sabah 579 AL-Rashdan Amel 55 1 Anderson Mats 26 1 Anglov Thomas 329 Ansari Tariq M. 689 Badini Raul G. 481 Baeten Wilhelmina L. M. Bagdi Gyula 769 Bai Jian 35 425 Barnes Ramon M. 653 Barrett Jon F. R. 109 Baxter Douglas C. 141 405 Bekkers Mirjan H. J. 599 Berglund Michael 14 1 46 1 Bhattacharyya Shuvendu S. Biffi Claudio 409 Bolshov Mikhail A. 1 99 Bosch Mossi F. 47 Bosch Reig F. 47 Boutron Claude F. 99 Bozsai Gabor 505 Braverman Diane S. 43 Brumme Mathias 28 I Bulska Ewa 201 405 Byrne John P. 371 579 Cabon J. Y. 383 Campbell Michael J. 6 I7 Carlini Enzo 19 Carroll John 533 Caruso Joseph A.551 Chakrabarti Chuni L. 37 1 579 Chaudry Muhammad Mansha 29 701 Chekalin Nikolai V. 225 Chenery Simon 647 Chiba Koichi 115 Christensen Jytte Molin 329 Cobo Isabel Gutierrez 247 Coedo Aurora Gomez 247 Coles Barry J. 587 Cornett Claus 629 Cox Rosamund J. 635 Crain Jeffrey S. 605 Dams Richard 6 17 Das Arabinda K. 41 7 Dean John R. 229 Demeny Dezsd 545 707 Dedina Jiii 307 Docekal Bohumil 52 1 Dong Liping 293 439 Dorado Lopez M. Teresa 1 I Duan Yixiang 7 Duckworth Douglas C. 71 1 Duller Geoff A. T. 53 Ebdon Les 23 51 1 719 Ehmann William D. 749 Emteborg HBkan 405 Escudero Baquero Esther Falk Heinz 255 Fang Zhaolun 293 439 Farnsworth Paul B. 89 197 599 461 727 417 461 247 247 CUMULATIVE AUTHOR INDEX FEBRUARY-AUGUST 1992 Feldmann Ingo 121 Fellows Craig S. 315 743 Feng Jianxing 17 1 Fischer Werner 239 Fisher Andrew S.51 1 Ford Michael J. 719 Frech Wolfgang 141 405 Freedman Philip A. 57 1 Fuge Ronald 53 595 61 1 Galley P. J. 69 Gallimore David L. 605 Garcia-Olalla Conception Gilmutdinov Albert Kh. 675 Glick M. 69 Gluodenis Thomas J. Jr. Gomez Coedo Aurora 11 Gorlach Ufsula 99 Grazhulene Svetlana S. 105 Gregoire D. C. 371 579 Grosse-Wilde H. 343 Guntur Daru 239 Guo Tiezheng 667 Gutierrez Cobo Isabel 11 Giiqer Seref 179 Guell Oscar A. 135 Giinther Detlef 25 1 Hakala Erkki 191 Han Heng-bin 447 Han Myung S. 641 Hanselman D. S. 69 Hansen Steen Honore 629 Harnly James M. 533 Harrison W. W. 75 Hartley James 23 Haschke Michael 28 1 Haug Hermann O. 451 Heckel Joachim 28 1 Heitkemper Douglas T. 55 1 Hermann Gerd M. 457 Hernandez Cordoba Manuel Hieftje Gary M.69 335 Hill Steve J. 23 5 1 I 7 19 Hinds Michael W. 685 Hoffmann Erwin 727 Holcombe James A. 135 Huang Benli 287 Huie Carmen W. 353 Hunt Andrew 647 Hutton Robert C. 623 Hiitsch Bruno 1 Inamoto Isamu 115 Ivanov V. P. 675 Iwamoto Etsuro 42 1 fnce Hiirrem 187 Jacksier Tracey 653 Jahl Matthias J. 653 Jakubowski Norbert 12 1 Jansen Elisabeth B. M. 127 Jian An-bei 515 Jiang Gui-bin 447 Jimenez Seco Jose L. 1 1 Jin Qinhan 7 Katoh Takashi 539 Kato Takunori 15 Kantor Tibor 2 19 Khvostikov Vladimir A. 105 Kibble Helen A. B. 315 Kishimoto T. 343 753 30 1 247 247 529 Kitagawa Kuniyuki 539 Kleiner Joachim 433 Klockenkamper Reinhold Knipscheer Joop H. 127 Koloshnikov Vsevolod G. 99 Kong Xjangxing 7 Koklii Unel 187 Krishna Prabhu R. 565 Krivan Viliam 155 52 1 Kujirai Osamu 661 Kumamaru Takahiro 42 1 Kumpulainen Jorma 165 Kuss Heinz-Martin 25 1 Kuzua Mikio 493 Lakatos Istvan 769 Lakatos Janos 769 Lajunen Lauri H.J. 735 Lamoureux Marc 371 579 Larkins Peter L. 265 Larsen Erik H. 629 Lasnitschka George F. 457 Le Bihan A. 383 Lee Kwang W. 641 Lee Milton L. 197 Lei Zhu 425 Li Ang 447 Li Bingwei 425 Li Ke 141 Li Qinguan 13 1 Li Wenchong 131 Li Yongquan 425 Li Zhikun 425 Lin Fan 175 Lin Yuehe 287 Littlejohn David 29 533 689 695 701 727 Liu Jun 7 Liu Mingzhong 667 Lopez Garcia Ignacio 529 Lopez M. Teresa Dbrado Louie Honway 557 Liidke Christian 727 Luecke Werner 765 Lupke G. 343 Ly T. 371 Mahalingam T. R. 565 Majidi Vahid 749 Marchante Gayon Juan M. Marcus R. Kenneth 71 1 Markesbery William R. 749 Marowsky G. 343 Marshall John 229 Mathews C. K.565 Ma Yizai 35 425 McLeod Cameron W. 66 1 Mentasti Edoardo 19 Mikami Osamu 493 Miller-Ihli Nancy J. 533 Moder Ralph 457 Moen ke-Blankenburg Lieselotte 25 1 Mohammad Bashir 695 Morimoto Satoru 2 1 1 Morita Masatoshi 15 Mouillere Delphine 701 Mu Huiling 175 Mudakavi Jayateerth R. 499 Nagahiro Tohru 183 Naghmush Abdulmakid M. 273 247 743 323 Nagtegaal Mario 127 Nakahara Taketoshi 2 1 1 Naoumidis Aristidis 239 Nickel Hubertus 239 Ni Zhe-ming 447 5 15 Nichol Robbin 727 Oftley Stephen G. 3 15 O’Gram Samantha J. 229 OGregoire D. Conrad 579 O’Haver Thomas C. 533 Ohlsson K. E. Anders 357 Okochi Haruno 661 Olbrych-Sleszynska Ewa 323 Omenetto Nicolo 89 Ottaway Barbara J. 701 Park Chang J. 641 Park Sang R. 641 Pastor Garcia A. 47 Patterson Clair C.99 Paul Michael 251 Pearce Nicholas J. G. 53 595 61 1 Peramaki Paavo 735 Perez Parajon Juan 743 Peris Martinez V. 47 Perkins William T. 53 595 Petrucci Guiseppe A. 481 Pickford Christopher J. 635 Porta Valerio 19 Poulsen Otto Melchior 329 Pritzl Gunnar 629 Puri Be1 Krishan 183 Pyy Lauri 191 Qian Haowen I3 1 Quevauviller Philippe 6 17 Rademeyer C. J. 347 RadiC-PeriC Jelena 235 Radziuk Bernard 389 397 Raith Angelika 623 Ramsey Michael H. 587 Rankin Andrew H. 587 Rosen Arne 261 Rudnev Sergei N. I 99 Saarela Kjell-Erik 165 Sack Brigitte 121 Saeki Masao 115 Sanz-Medel Alfredo 743 Sarzanini Corrado 19 Satake Masatada 183 Schindler Rolf 28 1 Schlemmer Gerhard 499 Schrader Werner 66 7 Schumann Thomas 25 1 Seare Nichola J. 3 15 Shan Xiao-quan 394 447 Shi Huiming 175 Shimazu Hiromichi 42 1 Sieverdes F.343 Sinemus Hans-Werner 433 Skole Jochen 727 Slaveykova Vera I. 147 365 Smith Benjamin W. 89 Soo Susan Yoke-Peng 557 Sorokin Mikhail V. 105 Sperling Michael 505 Stabel Hans-Henning 433 Starn Timothy K. 335 Stuewer Dietmar 12 1 Sturgeon Ralph E. 339 Sun Di-jun 35 61 1 433 505 76 17 80 Sychra Vaclav 389 Szardening Thomas W. 457 Szucs Laszlo 707 Tan Jingyuan 131 Tao Keyi 171 Thomassen Yngvar 397 Thompson Michael 635 647 Tittarelli Paolo 409 Tomlinson William R. 229 Trojanowicz Marek 323 Tsalev Dimiter L. 147 365 Tyson Julian F. 301 315 Uehiro Takashi I 5 Ure Allan M. 695 van de Weijer Peter 599 Van Grieken RenC 81 Vela Nohora P. 551 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 Vermaak I. 347 Vermeir Gerda 617 Vijayalakshmi S.565 Vifias Pilar 529 Vlasov Igor I. 225 Voloshin A. V. 675 von Bohlen Alex 273 Vullings Peter J. M. G. 599 Kykhristenko Nina N. 105 Walder Andrew J. 571 Wang Jiazhen 425 Wang Wen 761 Wang Xiaoru 287 Wang Xinsheng 175 Wasa Tamotsu 21 1 Wei Jizhong 175 499 505 Welz Bernhard 307 389 Wen Bei 761 Wenzel W. 343 Whitley John E. 29 Whittaker Paul G. 109 Wilkinson Jamie J. 587 Williams John G. 109 Willie Scott N. 339 Winefordner James D. 89 Wu Mingin 197 Wu Nian 353 Xhoffer C. 81 Xiao Jian 131 Xu Fu-chun 515 Xu Ning 749 Xu Peiqing 775 48 1 Xu Shukun 293 Yamada Kei 661 Yaman Mehmet 179 Yang Pengyuan 287 5 15 Yang Seok R. 641 Yasuhara Akio 15 Yokota Kayoko 421 Yuan Dongxing 287 Zakharov Yu. A. 675 Zhang Baogui 171 Zhang Hanqi 7 Zhang Li 447 Zhang Zhanxia 13 1 Zheng Hui 425 Zhu Lei 425 Zhuang Zhi-xia 287 5 I 5

ISSN:0267-9477    

DOI:10.1039/JA9920700779

出版商:RSC    

年代:1992

数据来源: RSC