Behaviour of Various Arsenic Species in Electrothermal Atomic Absorption Spectrometry I Journal of I Analytical Atomic Spectrometry I I VERA I. SLAVEYKOVA," FARAMARZ RASTEGAR AND MAURICE J . F. LEROY Ecole Europeenne de Chimie Polymeres et Materiaux de Strasbotirg Laboratoire de Chimie Analytique et Minerale U R A 405 CNRS 1 rue Blaise Pascal 67 008 Strasbourg Cedex France The behaviour of arsenite (As"') arsenate (As') monomethylarsonate (MMA) dimethylarsinate (DMA) arsenobetaine ( AsB) and arsenocholine ( AsC) in pyrolytic graphite coated graphite (pyrocoated) tubes was investigated. The influence of a tungsten carbide coating on the thermal pre- treatment losses of the analytes and on the analytical signals in aqueous and methanolic solutions was studied. Inorganic species MMA and DMA are less volatile in pyrocoated tubes; a tungsten carbide coating produces a good thermal stabilization but a marked 'dip' in the pyrolysis curves is observed in aqueous solutions.No pronounced stabilizing effect for the highly volatile AsB and AsC was observed in tungsten- treated tubes or in the presence of palladium chloride. The determination of these species requires the addition of palladium nitrate in both pyrocoated and tungsten-treated tubes. A comparison of the stabilizing action of palladium as its chloride and nitrate was made. Palladium nitrate exhibits efficient stabilizing action for each of the species studied whereas palladium chloride is efficient only for inorganic 4s species in pyrocoated tubes. The tungsten treatment of the tube and addition of palladium nitrate leads to a further increase in the pyrolysis temperature and better sensitivity for the As species.Tungsten treatment plays an important role in improving the performance of palladium chloride particularly in the determination of organically bound species. The effective stabilization and relative 'levelling-off' of the signal for each of the As species (except for AsC) in methanolic solution was observed in the presence of a palladium modifier in tungsten- treated tubes. The in situ separation and determination of As"' and AsB in tungsten-treated tubes was attempted. Howevtx because of the presence of significant amounts of AsB under the conditions used for As"' determination complete separation of these species was not possible.Keywords Arsenic species; electrothermal atomic absorption spectrometry; tungsten carbide coating; palladium modijie 1. Arsenic exists in environmental and biological samples in different oxidation and binding states. A knowledge of the speciation and transformation of As is important because each of the species possesses unique physical and chemical proper- ties which determine their bioavailability and physiological/ toxicological effects in the environment.''2 ETAAS provides limited potential for speciation. Usually the advantages of ETAAS are used for total As determination3 or off-line detection of As species after separation by HPI 1C4-7 and various extraction as well as the trapping of hydrides on the treated surface of a graphite This is reflected in the appearance of only a few publications dealing with the behaviour of As species in ETAAS. Krivan and Arpadjan13 have investigated the behaviour of As"' and AsV in a graphite furnace by means of an 76As * On leave from Faculty of Chemistry University of Sofia 1 lames Bourcier Sofia 1126 Bulgaria.radiotracer. The influence of different matrices such as HCl NaC1 HNO and urine and of various chemical modifiers including W and Pd was studied. A mixed W+Pd+citric acid modifier has been applied to determine As"' and AsV by ETAAS after extraction ~hromatography.'~ Tsalev et al.I5 showed that Ce" thermally stabilizes arsenite arsenate monomethylarsonate (MMA) and dimethylarsinate (DMA) up to 1100-1300°C and improves their response in ETAAS. LarsenI6 evaluated the analytical sensitivity for six As species using conventional and fast furnace programmes as well as a Pd + Mg modifier.Nevertheless systematic investigations of the behaviour of the individual As forms during the thermal pre-treatment and atomization stage are lacking. The aim of this work was to investigate the behaviour of arsenite (As"') arsenate (As") MMA DMA arsenobetaine (AsB) and arsenocholine (AsC) by ETAAS. Aqueous and methanolic solutions of these species were studied because they are the main components of the various extraction systems and HPLC effluents. The efforts were aimed at using the advantages provided by the W Pd and W+Pd modifiers" and by treatment of the graphite tubes with W and to compare their influence on the atomization signal and thermal pre-treatment losses of the As species.The W and W + Pd modifiers are very effective thermal stabilizers for inorganic As.'* Palladium alone and in combi- nation with other substances is one of the most popular and efficient modifiers used in ETAAS.I7 Pre-treatment of the graphite tube with W and other carbide-forming refractory elements has been widely employed in ETAAS offering certain advantages as reviewed elsewhere." Attention was directed to a detailed investigation of the pyrolysis stage which is the most critical step for the analysis of highly volatile species. The investigation of the behaviour of the various As species by ETAAS is of interest in both fundamental and practical aspects. In analytical practice it is important to apply a correct calibration procedure and to obtain reliable results for the total As in a sample of environmental or biological origin as well as with a view to developing methods for in situ speciation for some of the species.This preliminary study could be a basis for the development of ETAAS methods free of systematic errors for As determination in various undigested biological samples when As is present in a different form than in the reference solutions. EXPERIMENTAL Apparatus Measurements were made with a Varian ( Palo Alto CA USA) SpectrAA 400 Zeeman atomic absorption spectrometer equipped with a GTA 96 graphite atomizer and a programm- able sample dispenser. The operating parameters were set as recommended by the manufacturer except that a bandpass of Journal of Anizlytical Atomic Spectrometry October 1996 Vol.11 (997-1 002) 997Table 1 Temperature programme for the GTA 96 graphite atomizer loo-. 4 3 80 0 W 9) 6o s 40 a 20 Step No. 1 2 3 4 5 6 7 8 9 0 11 - - - - T/"C 90 120 120 var" var" 120 120 120 2300 2300 2600 t'ls 10 10 20 10 20 10 1 1 1 3 2 Ar flow rate/ 1 min-' 3.0 3.0 3.0 3.0 3.0 3.0 3.0 0 0 0 3.0 Read No No No No No No No No Yes Yes No *See figures. 0.5 nm was used. Pyrolytic graphite coated graphite (pyro- coated) partition tubes and integrated absorbance measure- ments were used throughout. The optimized heating programme for the GTA 96 graphite atomizer is shown in Table 1. A 'cool down' step was incorpor- ated in the heating programme in order to normalize atomiz- ation conditions for experiments at different pre-treatment temperatures.Reagents Stock standard solutions each containing 1000 mg 1-' of AS"' As" MMA DMA AsB and AsC in doubly distilled water were prepared using the following reagents sodium arsenite NaAsO,. (Rectapur; Prolabo Paris France); sodium arsenate Na2HAs0,.7H,0. (Rectapur; Prolabo); sodium monomethyl- arsonate CH3As0,Na2.6H,0. (Carlo Erba; Milan Italy); sodium dimethylarsinate (CH,),As02Na.3H,0. (Rectapur; Prolabo); arsenobetaine C,H1,As02*5H20. (Service Central de Microanalyse CNRS; Solaize France); and arsenocholine C,H,,AsOBr(CNRS). Working solutions containing 50 or 100 pg I-' of As were prepared daily from each of the stock standard solutions by dilution in doubly distilled water. The W chemical modifier (0.1% m/v) was prepared by dissolution of ammonium paratungstate (NH,),,H,( W207)6 (Fluka Buchs Switzerland) in doubly distilled water with gentle heating.Solutions (0.2 and 0.1% m/v) of Pd modifier as Pd(N03) in 15% HNO (Merck Darmstadt Germany) and PdCl in 10% HC1 (Riedel-de-Haen Hannover Germany) were used in the thermal pre-treatment study. Procedures Treatment of the graphite tube with the W modifier was realized by injection of 50 pl of a 0.1% m/v solution of the W modifier into the tube and thermal treatment by using the following temperature programme 90 "C for 20 s 120 "C (15+20s) 300°C (10+20s) 1200°C (lO+lOs) 2300°C (lo+ 2 s) 2500 "C for 2 s. The procedure was repeated three times. Such thermal treatment is believed to provide a smooth carbide coating with a minimized risk of high temperature volatilization of W03 and WC.,' The following procedure for in situ removal of As contami- nation from the W modifier was developed and applied in some thermal pre-treatment studies.A 20p1 aliquot of the modifier solution was injected and As was removed by pre- heating to 1700 "C for 10 s. After cooling to 40 "C the sample aliquot was placed in the graphite tube and the temperature programme in Table 1 was applied. A similar approach to in situ modifier purification has been used in Cd determination in serum in the presence of a Mg+ Pd modifier.21 Thermal Pre-treatment Study The maximum permissible pyrolysis temperatures were deter- mined by systematic variation of the pyrolysis temperature at a fixed atomization temperature. The effect of any drift in sensitivity was avoided by randomizing measurements and using the mean of two or three values at each temperature.Normalized absorbance measurements were used assigning a value of 100% to the plateau absorbance for the species. When a pronounced plateau in the pyrolysis curves was absent the maximum integrated absorbance signal was used to nor- malize the signals. The influence of the pyrolysis time on analyte losses and absorbance signals has been studied in detail p r e v i ~ u s l y . ~ ~ . ~ ~ On the basis of that study a pyrolysis time of 20 s was chosen as being the optimum and experiments were carried out at fixed pre-treatment times. RESULTS AND DISCUSSION Study of the Behaviour of As Species in Aqueous Solutions in the Absence of Modifier The behaviour of As in pyrocoated tubes depends on its oxidation and binding states.Typical examples of thermal pre- treatment curves for MMA and AsB are presented in Figs. 1 and 2 respectively. The inorganic and methylated species exhibit a maximum in the pyrolysis curves at temperatures of about 700-800 "C. At lower pyrolysis temperatures losses in the integrated absorbance signal were observed. The dimin- ution of the signal at low pre-treatment temperatures could be explained by redistribution of the sample in the tube as a result As is less efficiently atomized in the cooler region towards the ends of the tube. The temperature gradient in the '20 T fi 0 I 0 200 400 600 800 1000 1200 1400 1600 1800 Temperature PC Fig. 1 Pyrolysis curves for MMA in pyrocoated tubes (A) W-treated tubes (B) and in the presence of in situ purified W (20 pg) (C).T 2300 "C; aqueous solutions I2O T 04 I 0 200 400 600 800 1000 i200 1400 1600 1800 Temperature /OC Fig. 2 Pyrolysis curves for AsB in pyrocoated tubes (A) W-treated tubes (B) and in the presence of in situ purified W (20 pg) (C) and 6 pg Pd in W-treated tubes (D). T 2300 "C; aqueous solutions 998 Journal of Analytical Atomic Spectrometry October 1996 voz. 11Table 2 Characteristic masses m in picograms for different As species in aqueous solutions W-treated tubes Pd as its nitrate of chloride and W-treated tubes plus Pd modifier as its nitrate or chloride respectively ~ ~ As species Tube/modifier As"' AsV MMA DMA AsB AsC Pyrocoated 39 35 36 36 56 72 W 38 35 36 32 36 66 PdCl 40 43 51 45 48 65 WNO,) 25 26 26 25 26 40 W-PdCI 25 25 25 28 26 37 W-Pd(NO,) 20 19 20 20 20 28 central part of the furnace at low temperatures should favour condensation of both As vapours and compounds which vaporize at low temperature without decomposition.The vapo- rization of volatile As compounds at low temperature and their transformation into a less volatile form during thermal treatment could also be taken into account to explain the observed effect. By using MS at atmospheric pressure Styris et aL2' have shown that in the absence of a modifier As"' in HNO medium begins to lose As as AsO(g) at 490K; the losses are significant at 770K when As,(g) appears in the gas phase. AsB and AsC are more volatile and it is impossible to analyse these species without losses at ashing temperatures higher than 200 "C.Variation of the atomization temperature shows that the maximum integrated absorbances are obtained between 2000 and 2300 "C; however a temperature of 2300 "C was chosen because it gave the best ratio of peak height peak area and good response for all the species studied. The values of the characteristic masses m which correspond to the mass of analyte producing an integrated absorbance signal equal to 0.0044 s at pyrolysis temperatures correspond- ing to the maximum absorbance signal in the pyrolysis curves were calculated and are summarized in Table 2. The values shown are for aqueous solutions of the sodium salts of the As species without addition of acids. Chemical Modification With Tungsten Palladium and Tungsten Plus Palladium Efficient thermal stabilization is one of the prime requirements of a modifier.The utilization of higher pyrolysis temperatures has beneficial effects as has been reviewed in ref. 17. The potential of the applied modifiers was investigated with respect to their stabilizing power and improvement in the sensitivity. The stabilizing action of two forms of Pd viz. the nitrate and chloride in pyrocoated and W-treated tubes was compared. On the basis of a previous study,26 6 pg of Pd modifier was chosen as being the optimum. The effects of W tube treatment and in some experiments an in situ pre-purified W modifier were also studied. Aqueous Solutions The maximum pyrolysis temperatures for the studied systems in aqueous solutions are compiled in Table 3. Further details and specific features of the pyrolysis curves are presented below.In contrast to pyrocoated tubes signal losses for inorganic and methylated As species were not observed at low temperatures in W-treated tubes; however a minimum in the pyrolysis curves between 800 and 1000°C was clearly estab- lished. A similar effect has been observed for Se species uiz. Se" SeIV and SeV1 in the presence of 20 pg of Ni and 2.5 pg of Cu as m0difie1-s.~~ The observed losses were avoided when a sufficiently high thermal pre-treatment temperature was applied and the original absorbance signal was restored. Various factors could be responsible for the observed 'dip' in the pyrolysis curves. Insufficiently high atomization tempera- tures differences in the heating rates and different starting pyrolysis temperatures are unlikely to influence the signal because the 'dip' appears when higher atomization tempera- tures and a 'cool down' step temperature programme are used.The transformation of As species into volatile forms which would be expected to evaporate at temperatures higher than 500 "C probably accounts for the losses observed. The earlier appearance of losses of the As species AS'" AsV MMA and DMA in W-treated tubes in comparison with pyrocoated tubes could be explained by hindrance of the formation of intercal- ation compounds between As species and graphite in W-treated tubes. The original signal is restored at 1000-1200°C; this could be related to the specific properties of tungsten carbides. It is known that tungsten carbides are oxidized at temperatures higher than 1000°C; moreover they can play the role of oxygen carrier.Ig It is possible that the resulting tungsten oxide prevents further reduction of oxygen-containing As species by binding them and delaying their vaporization.The more volatile AsB and AsC are not thermally stable in the W-treated tube; however the sensitivity is still about 1.5 times better (see Table2) than that for the pyrocoated tube alone. It should be noted that W-treated tubes show a marked drift in sensitivity after about 60 firings. The original sensitivity is restored when the procedure for W tube treatment described under Experimental is re-applied. In order to provide additional insight into the reasons for the 'dips' in the pyrolysis curves in W-treated tubes experi- ments were carried out with an in situ purified W modifier.In the presence of 20 pg of W the pronounced 'dip' in the pyrolysis curves for inorganic As species MMA and DMA is not observed and the maximum loss-free pre-treatment tem- peratures are of the same order as in the W-treated tubes. Even the very volatile AsB and AsC are stable up to 1000 and 900°C respectively. The stabilizing effect could be due to the embedding of As species (by means of isomorphous substi- tution or simple embedding) in the modifier oxides or oxocar- bides which would lead to a decrease in the As partial pressure and reduce analyte losses during pyrolysis. Table 3 Maximum pyrolysis temperatures for As species in the presence of various modifiers Species P yrocoated * W - treated 20 Pg w PdCl Pd(NO3 )2 W-PdCl W-Pd (NO,) As"' 800 1600t 1600 1300 1300 1500 1500 AsV 800 1500t 1500 1300 1300 15002 1500 MMA 800 13001- 1300 400§ 1300 14002 1400 DMA 800 1200t 1250 4 w 1200 14002 1400 AsB 200 200 1000 400s 1200 1400 1400 AsC 120 200 900 40@ 1200 1400 1400 *Corresponding to the maximum integrated absorbance signal in the pyrolysis curves.?A 'dip' in the pyrolysis curves was observed. $A signal increase at low temperature was observed. $The signal diminished gradually with temperature. Journal of Analytical Atomic Spectrometry October 1996 Vol. 1 1 999The observed differences in the behaviour of inorganic As species MMA and DMA in pyrocoated and W-treated tubes compared with that of AsB and AsC could become a major source of systematic errors when undigested natural samples are analysed directly in a graphite furnace.As"' and AsV are usually used to prepare the standard solutions whereas AsB is the predominant species in some biological materials par- ticularly marine tissues. Differences in the behaviour of As species were observed in pyrocoated tubes in the presence of PdC1 and Pd(N03) as modifiers. Inorganic species are stabilized efficiently up to 1300"C regardless of the form in which the Pd modifier is introduced into the tube. For organically bound species a pronounced plateau up to 1200-1300 "C was found in the presence of Pd(N03) whereas a gradual decrease in the integrated absorbance signal with increasing pyrolysis temperature was found for MMA DMA AsB and AsC in the presence of PdC1 as is illustrated in Fig.3 for MMA. The low efficiency of the stabilizing action of Pd as its chloride could be explained by the formation of volatile chloride-containing analyte species which are not stabilized by the modifier.28 The experimental results of Kamiya et using thermogravimetry differential thermal analysis and X-ray diffraction indicate a transformation interval of 675-755°C for the decomposition of PdC1 to Pd. Thus residual chloride would be expected to have an effect on analyte stabilization and atomization. Depending on the form of the Pd modifier differences in the peak profiles were found for the various As species. In the presence of PdCl a small additional peak was observed which was negligible for the inorganic species and more pronounced for the organic species (see Fig.4). An increase in the pyrolysis temperature led to a reduction of this additional peak. Its appearance could be related to the action of residual chloride in the system; moreover the magnitude of this small additional peak correlates with the volatility of the chloride- 120 T t""l 8 "t $ 40 d 20 0 \ h,- A 0 - l I 0 200 400 600 800 1000 1200 1400 1600 1800 Temperature /OC Fig.3 Comparison of the stabilizing effect of Pd as its chloride or nitrate for MMA in pyrocoated tubes (A) and (B) and also W-treated tubes (C) and (D) respectively d) c e 8 2 -0 Pd as a chloride .05; 3.0 Time/s v . e 2 e F c a Fig. 4 Overlay of the absorbance profiles for 80 pg 1-' MMA in the presence of 6 pg Pd as its nitrate or chloride respectively; 20 pl injections methanol solutions containing species of the As forms and increases in the order As"' AsV < MMA < DMA < AsB < AsC.In the presence of Pd as its nitrate no double peaks were observed. In addition the absorbance signal was delayed in comparison with that in the presence of chloride. The effects discussed above are illustrated for MMA in Fig. 4. The sensitivity for the analytes in pure aqueous solutions using PdC1 was lower by approximately a factor of 1.5 as compared with that obtained in the presence of Pd(N03),. Similarly differences in the performance of Pd prepared from its chloride or nitrate have been reported for other element^.^^'^^'^^ The combination of W tube treatment and addition of Pd(N03) leads to a further increase in the pyrolysis tempera- ture for all the species studied and to an improvement in the sensitivity by a factor of 1.5-2.A considerable improvement in the performance of the PdCl modifier was obtained with W-treated tubes (Fig. 3). A pronounced enhancement of the maximum pyrolysis tempera- ture and decrease in the influence of residual chloride were found. It is difficult to explain the increase in the absorbance signal observed at low temperatures for AsV MMA and DMA in the presence of PdCl in a W-treated tube. Redistribution of the analyte during pre-treatment the availability of chloride in the system and the conversion of PdCl into an active metal form could be some of the reasons. However a complete explanation is not yet possible. The W treatment also results in an earlier appearance of absorbance profiles for As species in the presence of the Pd modifiers in comparison with pyrocoated tubes.This could be due to the formation of smaller Pd droplets during pyrolysis whereas in pyrocoated tubes more pronounced agglomeration of the particles on the graphite would be expected as has been shown in ref. 32. It should be noted that the maximum loss-free pyrolysis temperatures for As in W-treated tubes in the presence of Pd are of the same order as those for As"' in the presence of a mixed W+Pd modifier." This finding is not unexpected because Pd is thermally stabilized in W-treated platforms33 in a manner similar to that observed for the corresponding mixed W + Pd modifier.34 The analytical performance of a Pd modifier in the presence of chloride is of practical importance since many real matrices/ digests may contain an excess of chloride. Experimental results show that treatment of the tubes with W results in an improvement in their action as a thermal stabilizer and makes them less prone to chloride interferences particularly in the determination of AsB and AsC.The influence of the modifier on the characteristic masses of the As species studied is summarized in Table 2. For AS'" AsV MMA and DMA the sensitivity is similar for both pyrocoated and W-coated tubes. Palladium as its nitrate improves the sensitivity whereas Pd as its chloride produces a similar or poorer response for the As species in comparison with pyro- coated tubes. Tungsten treatment of the tube and addition of Pd further improves the characteristic masses and Pd(NO,) in combination with W provides the best sensitivity.The behaviour of the As species is very similar in the presence of the Pd(NO,) modifier in both pyrocoated and Table4 Characteristic masses m in picograms for different As species in methanolic solutions W-treated tubes (W) and W-treated tubes plus Pd modifier ( W-Pd) As species Modifier As"' AsV MMA DMA AsB AsC W 31 53 53 55 69 93 W-Pd 25 38 32 36 37 56 1000 Journal of Analytical Atomic Spectrometry October 1996 Vol. 11W-treated tubes; therefore Pd( NO3) is more suitable than the other modifiers studied for stabilizing the As species particularly in undigested biological samples. The role of Pd(NO,) in a graphite furnace as a thermal stabilizer also approaches the requirement of an 'analyte species isoformer'. =.'00 A Ej - a 60 8 3 * 20 a 40 D Methanolic Solutions The problems accompanying the analysis of organic scllutions in ETAAS are summarized in ref. 35. Taking into account the thermal pre-treatment resiilts for aqueous solutions only the role of W treatment and Pd(N03) modifier was studied. Tungsten treatment produces a s tabiliz- ing effect for inorganic and methylated As species. Analytical pre-treatment losses ('dips') were not observed and the pyrolysis curves showed a pronounced plateau (see Fig. 5). AsB and AsC in methanol were not stabilized in W-treated tubes and the maximum thermal pre-treatment temperature decreased in the order As"' As" > MMA > DMA >> As 13 AsC. The significant differences in the stabilizing action of Pd(NO,) in aqueous and methanolic solutions of the As species were not ascertained completely but the sensitic ity was impaired (see Table 4).The decrease in the sensitivity could be explained by spreading of injected sample further along the surface of the tube and as a result the As is less efficiently volatilized and atomized from the cooler region towards the ends of the tube. A decrease in the diffusion path of the species before leaving the tube36 could also be a factor. The addition of a Pd modifier in W-treated tubes not only assists the efficient thermal stabilization for all six As specie but also leads to a delay in the absorbance signal as illustrated in Fig. 6 for AsB. The peaks appear in a relatively short time and are narrower in W-treated tubes for each of the As species studied.The addition of Pd produces a shift in the peak maximum signal broadening and an increase in the integrated absorbance. - - - - - 120 T A'\ B 0 4 ..I---I 0 200 400 600 800 1000 1200 1400 1500 1800 Temperature /OC Fig.5 Pyrolysis curves for DMA in W-treated tubes (A) anct in the presence of 6 pg Pd in W-treated tubes (B). T 2300°C; methanolic solutions 0.50 3000 9 -. 2 4- 2 4 2 9 $ e F -0.05 * 0 0 30 Time/s Fig.6 Effect of W treatment and Pd modifier on the absorbance profiles for 100 pg 1-' AsB; 20 pl injections methanolic solutions Table 5 Determination of As"' and AsB in model solutions Concentration/pg I-' Total As As"' AsB Expected 50 57 107 Determined 61 50 111 Expected 10 99 109 Determined 29 82 111 Expected 90 11 101 Determined 92 16 108 Attempt at in situ Speciation The observed differences in the behaviour of AsB and AsC and of the inorganic and methylated As species suggested that in situ speciation of As in a W-treated graphite furnace might be possible.An attempt was made to separate and determine AsB and As"' in solutions containing these species in ratios of 1 1 1 9 and 9 1. AsB is the major As species in marine samples. As"' is the most toxic of the As species. These two species were chosen because of their similar sensitivity and a pronounced difference in their volatility during the pyrolysis stage in W-treated tubes. The procedure involves total As determi- nation in W-treated tubes with pyrolysis at 200°C and measurement of the As"' concentration after pyrolysis at 1400°C.The calibration is performed with AsB and As"' standards respectively. The results obtained are presented in Table 5. As can be seen the values obtained for total As are in good agreement with the expected values. However the results for As"' are systematically higher by about 20% for an As"' AsB ratio of 1 1 by about 50% for an As"' AsB ratio of 1 9 and in good agreement for an As"' AsB ratio of 9 1. The presence of significant amounts of AsB under the conditions used for As"' determination makes the complete separation and correct determination of As"' impossible. Unfortunately the tested procedure does not allow the in situ speciation of As"' and AsB in W-treated tubes. CONCLUSIONS The behaviour of several As species in ETAAS was investigated.It was found that their properties depend on the oxidation and binding states of As. On the basis of analyte behaviour in pyrocoated and W-treated tubes the As species can be classified conditionally into two groups. The first group includes inor- ganic and methylated As species. They exhibit a maximum in their pyrolysis curves at 600-700°C in pyrocoated tubes. A pronounced 'dip' in their pyrolysis curves between 800 and 1000°C in W-treated tubes and subsequent restoration of the original absorbance signal in aqueous solution is observed. The second group includes AsB and AsC which are more volatile in pyrocoated tubes and pre-treatment losses are possible even at drying temperatures as low as 120-200°C. As a result of the treatment of the tube with W the sensitivity can be improved but no significant stabilization is observed.In the presence of Pd as its nitrate or chloride the behaviour of the As species depends on the form of the modifier. Palladium as its nitrate efficiently stabilizes each of the species up to 1200-1300 "C whereas Pd as its chloride exhibits effective stabilizing action only for the inorganic species. The absorbance signal diminishes gradually at temperatures higher than 700°C for organic As species. The W treatment of the graphite tubes in combination with the Pd( N03)2 modifier produces excellent stabilizing action for all the species studied regardless of their oxidation and binding states and also improves and 'levels-off' the sensitivity (except for AsC). Journal of Analytical Atomic Spectrometry October 1996 VoE.11 1001Thus Pd(N03)2 in both pyrocoated and W-treated tubes is more suitable than the other modifiers for stabilizing As species in various undigested biological and environmental samples and is closer to the requirements of an 'analyte species isoformer'. A procedure for the in situ separation and determination of As"' and AsB in a model solution in W-treated tubes by ETAAS was tested; however because of the presence of signifi- cant amounts of AsB under the conditions used for As"' determination complete separation and determination of these species was not possible. REFERENCES 1 I 7 3 4 5 6 7 8 9 10 11 12 Seiler A. and Sigel H. Handbook on Metals in Clinical and Analytical Chemistry Marcel Dekker New York 1989.Florence T. M. in Trace Element Speciation Analytical Methods and Problems ed. Batley D. E. CRC Press Boca Raton FL Tsalev D. L. J. Anal. At. Spectrom. 1994 9 405. Brinkman F. E. Blair W. Jevett K. and Iverson W. J. Chromatogr. Sci. 1977 15 493. Chau Y. and Wong P. Fresenius' J. Anal. Chem. 1991 339 640. Environmental Analysis Using Chromatography Interfaced with Atomic Spectroscopy eds. Harrison R. M. and Rapsomanikis S. Ellis Horwood Chichester 1989. Ebdon L. Hill S. Walton A. P. and Ward R. W. Analyst 1988 113 1159. Chung C. H. Iwamoto E. Yamamoto M. and Yamamoto Y. Spectrochim. Acta Part B 1984 39 459. George G. M. and Frahm L. J. J. Assoc. Of. Anal. Chem. 1986 69 838. Sturgeon R. Willie S. Sproule G. Robinson S. and Berman S. Spectrochim. Acta Part B 1989 44 667.Zhang L. Ni Z.-m. and Shan X.-q. Spectrochim. Acta Part B 1989 44 339. Shuttler I. L. Feuerestein M. and Schlemmer G. J. Anal. At. Spectrom. 1992 7 1299. 1989 pp. 319-341. 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Krivan V. and Arpadjan S. Fresenius' Z . Anal. Chem. 1989 335 743. Russeva E. Havezov I. and Detcheva A. Fresenius' J. Anal. Chem. 1993 347 320. Tsalev D. L. Mandjukov P. B. and Stratis J. A. J. Anal. At. Spectrom. 1987 2 135. Larsen E. H. J. Anal. At. Spectrom. 1991 6 375. Tsalev D. L. Slaveykova V. I. and Mandjukov P. B. Spectrochim. Acta Rev. 1990 13 225. Slaveykova V. I. and Tsalev D. L. Anal. Lett. 1992 23 1921. Volynsky A. Zh. Anal. Khim. 1987 42 1541. Byme J. P. Hughes D. M. Chakrabarti C. L. and Gregoire D. C. J. Anal. At. Spectrom. 1994 9 913. Bulska E. Grobenski Z. and Schlemmer G. J. Anal. At. Spectrom. 1990 5 203. Slaveykova V. I. and Tsalev D. L. Spectrosc. Lett. 1991,24 139. Slaveykova V. I. and Tsalev D. L. J. Anal. At. Spectrom. 1992 7 365. L'vov B. V. Polzik L. and Yatsenko L. F. Talanta 1987,34 141. Styris D. Prell L. and Redfield D. Anal. Chem. 1991 63 503. Slaveykova V. I. and Tsalev D. L. Spectrosc. Lett. 1992,25,221. Welz B. Schlemmer G. and Voellkopf U. Spectrochim. Acta Part B 1984 39 501. Tesfalidet S. and Irgum K. Anal. Chem. 1988 60 2031. Kamiya N. Noshino K. and Ota K. Nippon Kagaku Kaishi 1988 12 1944. Voth-Beach L. M. Spectroscopy 1987 2 21. Voth-Beach L. M. and Shrader D. E. J . Anal. At. Spectrom. 1987 2 45. Qiao H. and Jackson K. Spectrochim. Acta Part B 1991 46 1841. Tsalev D. L. D'Ulivo A. Lampugnani L. Di Marco M. and Zamboni R. J . Anal. At. Spectrom. 1995 10 1003. Tsalev D. L. Slaveykova V. I. and Mandjukov P. B. in 5 Colloquium Atomspektrometrische Spurenanalytik ed. Welz B. Bodenseewerk Perkin-Elmer Uberlingen 1989 pp. 178-205. Komarek J. and Sommer L. Chem. Listy 1988 82 1151. Gilutdinov A. Kh. and Fishman I. S. Spectrochim. Acta Part B 1984 39 171. Paper 6/01 305 H Received February 23 1996 Accepted June 13 1996 1002 Journal of Analytical Atomic Spectrometry October 1996 Vol. 11