- Email: [email protected]
Determination of arsenic by inductively-coupled plasma atomic emission spectrometry enhanced by hydride generation from organized media
0039-9140/92 $5.00+ 0.00 Copyright0 1992PergamonRess Ltd
Tahto, Vol. 39,No. 11,pp. 1517-1523, 1992 Printedin GreatBritain.,411 rightsreserved
DETERMINATION OF ARSENIC BY INDUCTIVELY-COUPLED PLASMA ATOMIC EMISSION SPECTROMETRY ENHANCED BY HYDRIDE GENERATION FROM ORGANIZED MEDIA B. AIZPUN FERNANDEZ, C. VALDES-HEVIA Y TEMPRANO, M. R. FWNANDEZ DE LA CAMPA and A. SANZ-MEDEL*
Department of Physical and Analytical Chemistry, Faculty of Chemistry, University of Oviedo, C/Julian Claveria 8, Oviedo, Spain P. NEIL
Unicam Analytical Systems, Cambridge, England, U.K. (Received 15 March 1992. Accepted 31 March 1992) Summary-A method is described for the determination of arsenic, which combines a continuous flow hydride generation technique with an inductively coupled plasma atomic emission detection system. Some atomic absorption preliminary studies are described as well. Arsine is generated with NaBH, from a didodecyldimethylammonium bromide (DDBA) vesicular medium. The analytical performance of this vesicles-enhanced method is superior to the generation of the hydride from aqueous media: the detection limit (0.6 ppb) is improved by a factor of 2 and greater tolerance to interferences is observed for arsine generation from DDBA vesicles. Precision of As determinations is also improved. The proposed method has been validated for low As levels determinations in two Certified Reference Materials (CRM) sediments with satisfactory results. The potential of organized media to improve hydride generation is addressed.
The toxicity of arsenic to humans is widely recognised and consequently its determination in the environment and foodstuffs is of great importance. In sea and estuarine water, most arsenic is present as arsenate [As(V)], although low levels of arsenite [As(III)] and organoarsenicals (monomethylarsonic acid, dimethylarsinic acid, arsenobetaine, arseno-sugars, etc.) are also found.’ These species are also found in marine organisms and terrestrial plants. Although it is widely recognised that inductively coupled plasma atomic emission spectrometry offers adequate sensitivity for most environmental and clinical analyses, arsenic concentration levels in water samples are too low to be detected by ICP-AES.* The necessarily low efficiency of nebulization, to prevent quenching of the plasma, is a serious drawback in the determination of toxic metals by ICP-AES due to the high sample consumption and its wastage before the measuring process. The introduction of samples in the vapour phase to the plasma provides a solution to these problems. Sensitivity is improved due to the increase of the transport efficiency and inter*Author for correspondence.
ferences may be overcome by the vapour generation. Thompson et al. 3-5 first reported the combination of hydride generation techniques with ICP. Since then, a number of reports have appeared that discuss fundamental studies and applications”‘0 and theoretical aspects of hydride generation with ICP-AES. Particularly, the determination of arsenic by ICP based on arsine generation has been demonstrated to be subject to interferences, particularly from transition metal ions, which may affect the hydride An approach to imgeneration process. 4~‘o-12 prove the selectivity could be a previous liquid-liquid separation of arsenic from interfering elements,13 or a continuous on-line separation/hydride generation in the so-called “tandem on-line” technique used for As determination by ICP-AES and where the hydride is generated directly from the organic phase.i4 The introduction of organic solvents into the ICP, however, continues to present particular difficulties’s and therefore the use of micelles and other organised media, with a hybrid aqueous/organic character, could prove advantageous. In fact, we found that the presence of liquid solutions of “organized media”, e.g., 1517
B. AIZPUNFERNANDEZ et al.
1518
micelles and vesicles, may enhance the kinetics of the hydride generation.“j The singular characteristics of “organized media” such as solubilizing power, ability to build up a new “microenvironment”, organizing reactants at molecular level, capacity to change the kinetics of a process, could provide a powerful tool for enhancing the sensitivity of arsenic and other hydride forming elements. In addition, their availability to react selectively with opposite charged reagents could also be used to improve the selectivity. The purpose of the present work was to examine the ability of “organized media” to improve the analytical performance of hydride generation-ICP-AES technique by using arsenic as a model element. As a result of our studies, a sensitive and selective analytical method for the determination of arsenic by ICP-AES has been developed. This method is based on continuous flow arsine generation from vesicles of didodecyldimethylammonium bromide (DDBA). The validity of the method for real sample analysis has also been tested through the successful determination of low levels of arsenic in some environmental samples. EXPERIMENTAL
Apparatus ICP-AES measurements. A Unicam ICP, model PU7000 spectrometer was used for emission measurements. It has a 40.68 MHz freeradio-frequency generator, which running includes an internal voltage regulator. A fixed quartz torch is used. The spectrometer uses a coarse-ruled echelle grating, and cross dispersion with a quartz prism for order sorting, resulting in a two dimensional spectrum, at the focal plane. I7The nebuliser used consists of two platinum screens (dual platinum grid nebuliser): as the sample is nebulised from the first screen a primary aerosol is formed; the particle size is further reduced by impaction on a second screen. The nebuliser can act as an on-line gas-liquid separator when properly adjusted. As the liquid flows over the screens the hydride gases are entrained in an argon stream and are carried to the torch. The liquid is pumped through the drain to waste (Fig. 1). (This system is similar to that used by Watting and Collier’). Flame AAS measurements. A Perkin-Elmer Model 2280 atomic absorption spectropho-
Argon
Perstaltic pump Borohydride -,
To ICP
I I
T-piece Sample+ KI+HCI+BDDA
I Baffles
-brain Fig. 1.
tometer equipped with an electrodeless discharge lamp operated at 7 mA from an external power supply and a deuterium background corrector was used for AAS measurements. In this case, a laboratory-made hydride generator/gasliquid separator system previously describedI was used throughout for sample introduction of arsine (Fig. 2). A Perkin-Elmer MS-10 hydride generator was used for non-continuous (batch) measurements. An ultrasonic device Sonics & Materials, Model VC500, 250 Watts was used to prepare the vesicle solutions. Reagents A stock solution of arsenic (1000 p g/ml) was prepared by dissolving 1.320 g of As,O, in 25 ml of O.lM sodium hydroxide, hydrochloric acid was then added until pH 7.0 and this solution was made up to 1.0 litre with redistilled water. The working solutions were prepared fresh by diluting appropriate aliquots from the stock solution. Sodium tetrahydroborate (III) solution (1.5% w/v) was prepared by dissolving tetrahydroborate (III) powder (Carlo Erba) in demineralized Milli-Q water stabilized by 0.1% sodium hydroxide. Working solutions were prepared weekly and filtered before use. Potassium iodide stock solution (10% w/v) was prepared by dissolving 25 g of potassium iodide in 250 ml of demineralized Mini-Q water. Hexadecyl-trimethylammonium bromide (CTAB) solution (IO-‘M) was prepared by dissolving the surfactant in water by gentle warming. The other surfactants assessed were prepared in a similar way. Didodecyldimethylammonium bromide @DAB) solution (10e2M) was prepared by dissolving the surfactant powder in water and then sonieating for 12 min in an ultrasonic device, in order to obtain vesicles. All reagents used were of analytical-reagent grade and redistilled or Milli-Q water was used throughout.
iteration
Procedures Sample preparation. A 0.5-g weight of the sediment sample is weighed and placed into a nickel crucible, with 4 grams of potassium hydroxide. The crucible is then introduced into a furnace and heated at 500” for 30 min. The ashes are dissolved in 50 ml of 1M hydrochloric acid, and filtered before analysis. This solution is made up to 100 ml and analysed by ICP-AES. Non -con~jnuo~s (batch) me~urements. The sample is transferred into a lo-ml flask and made up to volume with demineralized Milli-Q water. An aliquot is placed into the reaction vessel and hydrochloric acid and potassium iodide are added. It is then mixed with a 8% NaBH, solution in the MS-lO/HG system. The generated arsine is swept into the heated quartz T-tube by a continuous stream of argon. Continuousmeasurements. An aliquot of 5 ml is placed into a 50-ml flask and 4.2 ml of hydr~hlo~c acid, 1 ml of 10% potassium iodide and 4.2 ml of 10W2MDDAB are added and this solution is made up to volume with demineralized Mill&Q water. This solution was continuously pumped through one of the channels of the peristaltic pump, while a 1.3% solution of NaBH,, was pumped through the second channel. Both flows were mixed at a T-piece where the hydride formation takes place. The gaseous and liquid mixture is passed through the grid nebuliser; the liquid is drained and the gaseous products are swept into the plasma, by the argon flow across the nebuliser. RESULTS AND DISCUSSION
AAS studies In our study of hydride generation methodology using organized media instead of aqueous media, the ability of different “organized media”16 to enhance the sensitivity of As determinations was investigated. Arsine generation was carried out in a Perkin-Elmer MS10 commercial hydride generator. Optimum instrumental conditions selected for such determination are summarized in Table 1.
of arsenic
1519
The following organised media were tested in concentrations below and above their critical micellar concentrations (cmc’s): -Cationic surfactant: CTAB (Hexadecyl-trimethyla~oni~ bromide). (cmc = 9.2 x 10-W) -Anionic surfactant: SDS (Sodium Dodecyl Sulphate). (cmc = 8.1 x 10b31M) -Non-ionic surfactant: TX-100 (TritonX-100). (cmc = 2 lo-4M) -Vesicles: DDAB (Didod~yldimethyl~monium Bromide). The net signals obtained generating the arsine from these four media, were compared with those obtained from aqueous media. Results showed that maximum AbsJAbsb ratios (where Abs, is the absorption due to the sample and Abs, the absorption due to the blank), were obtained with DDAB vesicles. To evaluate the relative merits of this methodology, we compared the sensitivity under optimized conditions of amine generation AAS from water, DDBA and TX-100 (the three media which provided higher signals). These experiments were performed in both, continuous and non-continuous hydride generation systems, and the results observed were as follows: Generationfrom a batch device (Perkin-Elmer MHS-IO). An aliquot of the sample solution is placed in the sample reservoir, along with hydrochloric acid and potassium iodide. An 8% borohydride solution and an argon stream were used to generate and transport the arsine to a quartz tube placed on the flame of the AAS spectrometer. Results obtained for the three media are given in Table 2. As shown in the Table, the precision was relatively poor when using a non-continuous generation system. Better results seemed to be obtained using vesicles of DDAB. To increase the precision of measurements a continuous generation system was investigated. Table 2. Comparison of analytical performance of amine generation-AAS in various media Precision (%)
L.D. Table 1. Inst~mentaI conditions measurements for As Lamp current Spectral band width Wavelength selected Air,/acetilene flow ratio
in AAS
7mA 0.7 mn 193.7 nm 4
Media Water BDDA TX-100
A
36 ng I7 ng 136 ng
A: Batch. lk Continuous.
3
4 ppb 1 ppb -
A
3
;: 39
6.7 6.9 -
1520
B. AIZPUNFEXNANDEZ et al,
Generation from a continuous hydride generation device. A schematic diagram of this hydride generator is shown in Fig. 2. The liquid streams of sample solution, or blank, in 0.2% potassium iodide solution [to ensure that all arsenic is in the As (III) form], 1M hydrochloric acid solution and the reducing reagent, 1% NaBH,, were pumped continuously and merged at a T-piece where the hydride formation takes place. The mixture of liquid and gas then arrives to an interface vessel (filled with 3-mm o.d. Pyrex beads) which provides a smooth separation of liquid and gas. The gaseous products, arsine and hydrogen, are swept through the vessel into the quartz tube of the AAS system by a continuous stream of argon. This argon stream was optimized and fixed at 0.045 l./min. The waste liquid is drained via a U-tube system, as can be seen in Fig. 2. Using this system, we compared the analytical characteristics of arsine generation, in terms of detection limit and precision, from the two better media, that is DDAB vesicles and water. As shown in Table 2, detection limits (3 Ob where bb is the standard deviation of blanks) found were: 1 ng/ml of As using BDDA and 4 ng/ml generating ASH, from aqueous media, the precision being similar in both cases (around f 7% for 12 ppb solutions of As).
Thus, organized media do enhance the analytical characteristics of hydride generationAAS determination of As. The use of micelles and other organized media are particularly useful in ICP-AES, because they provide a microenvironment similar to that of organic solvents for reactions.i6 However, they offer superior performance and are more compatible with plasma operation; in fact, extinguishing, de-stabilization and background levels of the ICP, are all minimized by using micellar solutions instead of organic solvents.i3*‘s Therefore the use of organized media was also tested for hydride generation-ICP-AES determination of As. HG-ICP
studies
The first step was to optimize the experimental generation of ASH, and its continuous introduction into the plasma using a Philips HG system (see Fig. 1). In a second step, the following organised media: Triton X-100, CTAB, and DDAB were assayed and compared with aqueous ASH, generation. Preliminary’ studies showed that CTAB produces a slight precipitate with potassium iodide, while TX-100 did not produce any improvement in the ICP-AES signal (as observed by AAS). Therefore, an
-Flame
Fig. 2.
Determination of arsenic
Table 3. Optimum
1521
operating conditions measurements
Plasm0 Coolant gas flow Pressure of Ar in nebuliser Forward power
in ICP-OES
11 l./min 35 psi 1.5 kW
Hydriak generation system 0.2
0.4
0.6
0.6
1
1.2
1.4
1.6
Fig. 3.
organized media of vesicles of DDAB finally selected for further experiments.
The criterion for optimization was taken to be the best signal-to-background ratio, SBR = (1, - &)/lb (where 1, is the intensity for the sample, and Ib for the blank). Selection of the analytical line. The two most sensitive analytical emission lines for arsenic were tested: 193.695 nm and 197.198 nm. The observed values of SBR were 7.2 for the former and 6.5 for the latter analytical line, using a 50-ppb standard. Optimization of chemical parameters for ASH, generation. Optimum hydrochloric acid and NaBH, concentrations were investigated using 50 ppb arsenic, a 0.5% potassium iodide solution as reducing agent and working at both emission lines sequentially. A univariant search was used in this instance to establish the optimum hydrochloric acid and NaBH, concentrations in the aqueous phase. The emission intensities obtained for each hydrochloric acid and NaBH, concentrations, are given in Fig. 3. It can therefore be seen why 1M hydrochloric acid and 1.3% NaBH, were selected for the subsequent work. It was observed that the ICP signal became higher with higher DDAB concentration. However, above 8.4 x 10W4M of the surfactant a precipitate formed in the presence of potassium iodide. Therefore, reductants other than potassium iodide were assayed in order to reduce As (V) to As (III). The behaviour of potassium 30
0’ 0
x
x-x
0.2
0.4
0.6 0.8 56 Reductor
Fig. 4.
concentration uptake rate acidity uptake rate cont.
1.3% m/v in 0.1% NaOH 2.1 ml/min 1M in HCl 4.2 ml/min 8.4 x lo-‘M
was
Optimization of the instrumental parameters
y 25 F
NaBH, NaBH, Sample Sample DDAB
iodide, titanium (III) chloride and sodium sulphite compared in vesicles of DDAB. Figure 4 summarizes the observed results which showed that the signal obtained when using TiCl, or Na,SO, was always lower than that obtained for potassium iodide. For this reason, a potassium iodide concentration of 0.5% was selected to ensure that all arsenic is in the As (III) form. The overall optimum conditions found are summarized in Table 3, and they happened to be the same in both media tested (water and DDAB). Analytical performance, characteristics of the DDAB-HG-ICP-AES determination of arsenic. Using the optimum conditions previously obtained (Table 3), sensitivity, selectivity, and precision of As determination by HG-ICP-AES using aqueous or DDAB media were evaluated. Calibration graphs obtained in both media are shown in Fig. 5 and show that the sensitivity (slope of calibration curve) is twice when using vesicles; the linear analytical range obtained extended from 6 ng/ml to, at least 100 pg/ml using DDAB and from 13 ng/ml to 100 pg/ml for water media, while the detection limts (3 bi,) were 0.6 ng/ml for BDDA and 1.3 ng/ml for aqueous media. The precision was evaluated by analysing ten replicates of a solution containing 50 ng/ml of arsenic with ASH, generation from aqueous and DDAB media. Relative standard deviation observed were f 3.2% using aqueous and f 2.3%
x
1
, 1.2 Cont. As (ppb)
Fig. 5.
B. AUPUN FERNNiDEz
1522
in DDAB media. Therefore, the precision of amine generation (see Table 2 for similar AAS measurements) seems to be clearly improved by adding vesicles to aqueous solutions. Interference
Table 4. Interferences of some elements in As determination by HG-ICPAES
Bi
Ca co Cr cu Fe
K Li Mn Na Ni Pb Sb Sn Te Ti Zn
Relationship 1:lOOO 1:500 1:300 1:lOO 1:50 1:lOOO 1:500 1:lOOO I:1000 1:500 1:lOOO 1:500 1:lOOO I:500 I:300 I:1000 1:500 1:300 1:lOOO 1:lOOO I:500 1:lOOO 1:lOOO 1:500 1:lOOO 1:500 1:lOOO I:300 1:200 1:lOO 1:500 1:300 1:lOOO 1:lOOO 1:sOO I:300 1:lOOO 1:500
Table 5. Results of the analysis of some sediments (reference lll&lidS)
Sediment CRM 211 CRM 320
studies
In order to investigate potential interferences from aqueous and DDAB media, solutions containing 25 ng/ml of arsenic (III) with various amounts of foreign ions were prepared. These solutions were analysed for As following the recommended procedure in the presence and absence of DDAB vesicles. All the results obtained are summarized in Table 4 in a comparative manner. As it can be seen in this table, some interferences are reduced (as in the case of Bi, for instance) or eliminated (in the case of Ti) when hydride generation is performed from DDAB vesicles. There is, however, an important interference in DDAB: when
Interferent
et a/.
Recovery (1) (Water) 50 52 54 55 85 110 101 101 108 93
Recovery (2) (DDBA medium) 15 98
98
91
:: 103 101 100 115 106 105 100 111 96 91 111 99 99 100 102 105 97 95 122 102 100 65 15 103 100
102 100
90 100 100 101
Observed value Wg) 41.6 f 1.0 14.1 f 3.3
Certified value 01B/B) 41.3 f 1.6 16.1 f 3.4
Sb is present in concentrations above 20 pg/ml, it produces precipitates (with DDBA) which seriously inhibit the introduction of the sample due to blocking of tubing. Results show that in general terms the method proposed using DDAB seems to be more selective than the method using aqueous solution and thus it can be favourable for the determination of As in real samples. Application to real samples
In order to test the analytical usefulness of the proposed method for the determination of low levels of As in environmental samples, we analysed two different certified sediments, namely, CRM 227 and CRM 320 (estuarine and river sediment respectively) from B.C.R. (E.C.). Sample dissolution and preparation was carried out as described elsewhere” with hydride production and determination using the developed methodology. The results obtained are given in Table 5 (each value given is the average of three determinations) and demonstrate that the DDABHG-ICP-AES procedure proposed here provides good accuracy and precision for this type of analytical problem. CONCLUSIONS
Long-chain surfactants have been proposed to improve the sensitivity of many analytical methods of atomic spectrometry. Sample transport in nebulizers and atomization efficiency enhancements in the flame had been invoked to rationalize the observed sensitivity improvements. Other workers, however, found very little or no sensitivity enhancement by surfactants, while recent research on the topic’s~‘9tends to indicate that transport efficiency does not increase on addition of most varied surfactants.” Micelles thus appear to have limited application for ICP-AES conventional nebulisation.” Conversely, micelles, and other organized media such as vesicles, could be most helpful in atomic methods based on hydride generation. As we have shown in a preliminary communication”j the hydride generation is a chemical
Determination
reaction which can be improved in the presence of such “organized” media, at least from two different points of view: (a) Micelles, vesicles, etc., are able to “organize” reagents at a molecular level, as has been repeatedly shownM for coloured and photoluminiscent reactions. This could affect sensitivity (effective concentration of reactants is higher in the microenvironment created by the organized medium than in water) and selectivity of hydride generation (i.e., by selecting oppositely or identically charged micelles for the analyte and interferences respectively). (b) Not only thermodynamic constants but kinetics of the reactions can also be altered by organized media. Our results for As have shown that, at least modestly, all the expected improvements (detection limits, selectivity and precision) have been realized using vesicles of DDAB in the continuous (and batch) generation of arsine. The “organized medium”-enhanced technique for hydride generation allows a fast and precise determination of As by ICP-AES and could be extended to other hydride forming elements. Moreover, analysis of environmental samples using the developed method gives results close to the certified values. Speciation of As is an important requirement for this application; this challenge could be approached using micellar chromatography” for separation of species and the proposed DDAB-arsine generation ICP-AES for detection. Research in this direction of As speciation is currently in progress in our laboratory.
of arsenic
1523 REFERENCES
1. A. G. Howard and M. H. Arbab-Zavar, Analyst, 1981, 106, 213. 2. R. J. Watting and A. R. Collier, ibid., 1988, 113, 345. 3. M. Thompson, B. Pahlavanpour, S. J. Walton and G. F. Kirkbright, ibid., 1978, 103, 568. 4. M. Thompson, B. Pahlavanpour, S. J. Walton and G. F. Kirkbright, ibid., 1978, 103, 705. 5. H. Thompson and B. Pahlavanpour, Anal. Chim. Acta, 1979, 109, 251. 6. E. De Oliveira, J. W. McLaren and S. S. Berman, Anal. Chem., 1983, 55, 2047. 7. R. R. Liversage, J. C. Van Loon and J. C. De Andrade, Anal. Chim. Acta, 1984, 161, 275. 8. G. S. Pyen and R. F. Browner, Appl. Spectrosc., 1988, 42, 508. 9. F. Nakata, H. Sunahara, H. Fujimoto, M. Yamamoto and T. Kumamaru, J. Anal. At. Spectrom., 1988,3,579. 10. T. Nakahara, Appl. Spectrosc., 1983, 37, 539. 11. J. Wilson Wershey and P. Keliher, Spectrochim. Acta., 1986, 41B, 713. 12. L. Halic and G. H. Rusell, Analyst, 1986, 111, 15. 13. A. Menendez Garcia, J. E. Sanchez Uria and A. SanzMedel, J. Anal. At. Spectrom., 1989, 4, 581. 14. J. Aznarez, F. Palacios, M. S. Ortega and J. C. Vidal, Analyst, 1984, 109, 713. 15. P. Changkan, Z. Guangzuan and R. F. Browner, J. Anal. At. Spectrom., 1990, 5, 537. 16. R. Femlndez de la Campa, C. Garcia Ortiz, M. Diaz Garcia and A. Sanz-Medel, Conference presentation, XII National Congress of Spectroscopy, Barcelona (Spain), 1990. 17. T. C. Dymott, International Laboratory, March 26, 1989. 18. J. Mora, A. Canals and V. Herandis, J. Anal. At. Spectrom., 1991, 6, 143. 19. Z. Yin Yan and Z. Wuming, ibid., 1989, 4, 797. 20. A. Sanz-Medel, P. L. Martinez and M. E. Diaz Garcia, Anal. Chem., 1987, 59, 774. 21. W. L. Hinze and D. W. Armstrong (ed.), Ordered Media in Chemical Separations, ACS Symposium Series 342, Washington 1987.
Tahto, Vol. 39,No. 11,pp. 1517-1523, 1992 Printedin GreatBritain.,411 rightsreserved
DETERMINATION OF ARSENIC BY INDUCTIVELY-COUPLED PLASMA ATOMIC EMISSION SPECTROMETRY ENHANCED BY HYDRIDE GENERATION FROM ORGANIZED MEDIA B. AIZPUN FERNANDEZ, C. VALDES-HEVIA Y TEMPRANO, M. R. FWNANDEZ DE LA CAMPA and A. SANZ-MEDEL*
Department of Physical and Analytical Chemistry, Faculty of Chemistry, University of Oviedo, C/Julian Claveria 8, Oviedo, Spain P. NEIL
Unicam Analytical Systems, Cambridge, England, U.K. (Received 15 March 1992. Accepted 31 March 1992) Summary-A method is described for the determination of arsenic, which combines a continuous flow hydride generation technique with an inductively coupled plasma atomic emission detection system. Some atomic absorption preliminary studies are described as well. Arsine is generated with NaBH, from a didodecyldimethylammonium bromide (DDBA) vesicular medium. The analytical performance of this vesicles-enhanced method is superior to the generation of the hydride from aqueous media: the detection limit (0.6 ppb) is improved by a factor of 2 and greater tolerance to interferences is observed for arsine generation from DDBA vesicles. Precision of As determinations is also improved. The proposed method has been validated for low As levels determinations in two Certified Reference Materials (CRM) sediments with satisfactory results. The potential of organized media to improve hydride generation is addressed.
The toxicity of arsenic to humans is widely recognised and consequently its determination in the environment and foodstuffs is of great importance. In sea and estuarine water, most arsenic is present as arsenate [As(V)], although low levels of arsenite [As(III)] and organoarsenicals (monomethylarsonic acid, dimethylarsinic acid, arsenobetaine, arseno-sugars, etc.) are also found.’ These species are also found in marine organisms and terrestrial plants. Although it is widely recognised that inductively coupled plasma atomic emission spectrometry offers adequate sensitivity for most environmental and clinical analyses, arsenic concentration levels in water samples are too low to be detected by ICP-AES.* The necessarily low efficiency of nebulization, to prevent quenching of the plasma, is a serious drawback in the determination of toxic metals by ICP-AES due to the high sample consumption and its wastage before the measuring process. The introduction of samples in the vapour phase to the plasma provides a solution to these problems. Sensitivity is improved due to the increase of the transport efficiency and inter*Author for correspondence.
ferences may be overcome by the vapour generation. Thompson et al. 3-5 first reported the combination of hydride generation techniques with ICP. Since then, a number of reports have appeared that discuss fundamental studies and applications”‘0 and theoretical aspects of hydride generation with ICP-AES. Particularly, the determination of arsenic by ICP based on arsine generation has been demonstrated to be subject to interferences, particularly from transition metal ions, which may affect the hydride An approach to imgeneration process. 4~‘o-12 prove the selectivity could be a previous liquid-liquid separation of arsenic from interfering elements,13 or a continuous on-line separation/hydride generation in the so-called “tandem on-line” technique used for As determination by ICP-AES and where the hydride is generated directly from the organic phase.i4 The introduction of organic solvents into the ICP, however, continues to present particular difficulties’s and therefore the use of micelles and other organised media, with a hybrid aqueous/organic character, could prove advantageous. In fact, we found that the presence of liquid solutions of “organized media”, e.g., 1517
B. AIZPUNFERNANDEZ et al.
1518
micelles and vesicles, may enhance the kinetics of the hydride generation.“j The singular characteristics of “organized media” such as solubilizing power, ability to build up a new “microenvironment”, organizing reactants at molecular level, capacity to change the kinetics of a process, could provide a powerful tool for enhancing the sensitivity of arsenic and other hydride forming elements. In addition, their availability to react selectively with opposite charged reagents could also be used to improve the selectivity. The purpose of the present work was to examine the ability of “organized media” to improve the analytical performance of hydride generation-ICP-AES technique by using arsenic as a model element. As a result of our studies, a sensitive and selective analytical method for the determination of arsenic by ICP-AES has been developed. This method is based on continuous flow arsine generation from vesicles of didodecyldimethylammonium bromide (DDBA). The validity of the method for real sample analysis has also been tested through the successful determination of low levels of arsenic in some environmental samples. EXPERIMENTAL
Apparatus ICP-AES measurements. A Unicam ICP, model PU7000 spectrometer was used for emission measurements. It has a 40.68 MHz freeradio-frequency generator, which running includes an internal voltage regulator. A fixed quartz torch is used. The spectrometer uses a coarse-ruled echelle grating, and cross dispersion with a quartz prism for order sorting, resulting in a two dimensional spectrum, at the focal plane. I7The nebuliser used consists of two platinum screens (dual platinum grid nebuliser): as the sample is nebulised from the first screen a primary aerosol is formed; the particle size is further reduced by impaction on a second screen. The nebuliser can act as an on-line gas-liquid separator when properly adjusted. As the liquid flows over the screens the hydride gases are entrained in an argon stream and are carried to the torch. The liquid is pumped through the drain to waste (Fig. 1). (This system is similar to that used by Watting and Collier’). Flame AAS measurements. A Perkin-Elmer Model 2280 atomic absorption spectropho-
Argon
Perstaltic pump Borohydride -,
To ICP
I I
T-piece Sample+ KI+HCI+BDDA
I Baffles
-brain Fig. 1.
tometer equipped with an electrodeless discharge lamp operated at 7 mA from an external power supply and a deuterium background corrector was used for AAS measurements. In this case, a laboratory-made hydride generator/gasliquid separator system previously describedI was used throughout for sample introduction of arsine (Fig. 2). A Perkin-Elmer MS-10 hydride generator was used for non-continuous (batch) measurements. An ultrasonic device Sonics & Materials, Model VC500, 250 Watts was used to prepare the vesicle solutions. Reagents A stock solution of arsenic (1000 p g/ml) was prepared by dissolving 1.320 g of As,O, in 25 ml of O.lM sodium hydroxide, hydrochloric acid was then added until pH 7.0 and this solution was made up to 1.0 litre with redistilled water. The working solutions were prepared fresh by diluting appropriate aliquots from the stock solution. Sodium tetrahydroborate (III) solution (1.5% w/v) was prepared by dissolving tetrahydroborate (III) powder (Carlo Erba) in demineralized Milli-Q water stabilized by 0.1% sodium hydroxide. Working solutions were prepared weekly and filtered before use. Potassium iodide stock solution (10% w/v) was prepared by dissolving 25 g of potassium iodide in 250 ml of demineralized Mini-Q water. Hexadecyl-trimethylammonium bromide (CTAB) solution (IO-‘M) was prepared by dissolving the surfactant in water by gentle warming. The other surfactants assessed were prepared in a similar way. Didodecyldimethylammonium bromide @DAB) solution (10e2M) was prepared by dissolving the surfactant powder in water and then sonieating for 12 min in an ultrasonic device, in order to obtain vesicles. All reagents used were of analytical-reagent grade and redistilled or Milli-Q water was used throughout.
iteration
Procedures Sample preparation. A 0.5-g weight of the sediment sample is weighed and placed into a nickel crucible, with 4 grams of potassium hydroxide. The crucible is then introduced into a furnace and heated at 500” for 30 min. The ashes are dissolved in 50 ml of 1M hydrochloric acid, and filtered before analysis. This solution is made up to 100 ml and analysed by ICP-AES. Non -con~jnuo~s (batch) me~urements. The sample is transferred into a lo-ml flask and made up to volume with demineralized Milli-Q water. An aliquot is placed into the reaction vessel and hydrochloric acid and potassium iodide are added. It is then mixed with a 8% NaBH, solution in the MS-lO/HG system. The generated arsine is swept into the heated quartz T-tube by a continuous stream of argon. Continuousmeasurements. An aliquot of 5 ml is placed into a 50-ml flask and 4.2 ml of hydr~hlo~c acid, 1 ml of 10% potassium iodide and 4.2 ml of 10W2MDDAB are added and this solution is made up to volume with demineralized Mill&Q water. This solution was continuously pumped through one of the channels of the peristaltic pump, while a 1.3% solution of NaBH,, was pumped through the second channel. Both flows were mixed at a T-piece where the hydride formation takes place. The gaseous and liquid mixture is passed through the grid nebuliser; the liquid is drained and the gaseous products are swept into the plasma, by the argon flow across the nebuliser. RESULTS AND DISCUSSION
AAS studies In our study of hydride generation methodology using organized media instead of aqueous media, the ability of different “organized media”16 to enhance the sensitivity of As determinations was investigated. Arsine generation was carried out in a Perkin-Elmer MS10 commercial hydride generator. Optimum instrumental conditions selected for such determination are summarized in Table 1.
of arsenic
1519
The following organised media were tested in concentrations below and above their critical micellar concentrations (cmc’s): -Cationic surfactant: CTAB (Hexadecyl-trimethyla~oni~ bromide). (cmc = 9.2 x 10-W) -Anionic surfactant: SDS (Sodium Dodecyl Sulphate). (cmc = 8.1 x 10b31M) -Non-ionic surfactant: TX-100 (TritonX-100). (cmc = 2 lo-4M) -Vesicles: DDAB (Didod~yldimethyl~monium Bromide). The net signals obtained generating the arsine from these four media, were compared with those obtained from aqueous media. Results showed that maximum AbsJAbsb ratios (where Abs, is the absorption due to the sample and Abs, the absorption due to the blank), were obtained with DDAB vesicles. To evaluate the relative merits of this methodology, we compared the sensitivity under optimized conditions of amine generation AAS from water, DDBA and TX-100 (the three media which provided higher signals). These experiments were performed in both, continuous and non-continuous hydride generation systems, and the results observed were as follows: Generationfrom a batch device (Perkin-Elmer MHS-IO). An aliquot of the sample solution is placed in the sample reservoir, along with hydrochloric acid and potassium iodide. An 8% borohydride solution and an argon stream were used to generate and transport the arsine to a quartz tube placed on the flame of the AAS spectrometer. Results obtained for the three media are given in Table 2. As shown in the Table, the precision was relatively poor when using a non-continuous generation system. Better results seemed to be obtained using vesicles of DDAB. To increase the precision of measurements a continuous generation system was investigated. Table 2. Comparison of analytical performance of amine generation-AAS in various media Precision (%)
L.D. Table 1. Inst~mentaI conditions measurements for As Lamp current Spectral band width Wavelength selected Air,/acetilene flow ratio
in AAS
7mA 0.7 mn 193.7 nm 4
Media Water BDDA TX-100
A
36 ng I7 ng 136 ng
A: Batch. lk Continuous.
3
4 ppb 1 ppb -
A
3
;: 39
6.7 6.9 -
1520
B. AIZPUNFEXNANDEZ et al,
Generation from a continuous hydride generation device. A schematic diagram of this hydride generator is shown in Fig. 2. The liquid streams of sample solution, or blank, in 0.2% potassium iodide solution [to ensure that all arsenic is in the As (III) form], 1M hydrochloric acid solution and the reducing reagent, 1% NaBH,, were pumped continuously and merged at a T-piece where the hydride formation takes place. The mixture of liquid and gas then arrives to an interface vessel (filled with 3-mm o.d. Pyrex beads) which provides a smooth separation of liquid and gas. The gaseous products, arsine and hydrogen, are swept through the vessel into the quartz tube of the AAS system by a continuous stream of argon. This argon stream was optimized and fixed at 0.045 l./min. The waste liquid is drained via a U-tube system, as can be seen in Fig. 2. Using this system, we compared the analytical characteristics of arsine generation, in terms of detection limit and precision, from the two better media, that is DDAB vesicles and water. As shown in Table 2, detection limits (3 Ob where bb is the standard deviation of blanks) found were: 1 ng/ml of As using BDDA and 4 ng/ml generating ASH, from aqueous media, the precision being similar in both cases (around f 7% for 12 ppb solutions of As).
Thus, organized media do enhance the analytical characteristics of hydride generationAAS determination of As. The use of micelles and other organized media are particularly useful in ICP-AES, because they provide a microenvironment similar to that of organic solvents for reactions.i6 However, they offer superior performance and are more compatible with plasma operation; in fact, extinguishing, de-stabilization and background levels of the ICP, are all minimized by using micellar solutions instead of organic solvents.i3*‘s Therefore the use of organized media was also tested for hydride generation-ICP-AES determination of As. HG-ICP
studies
The first step was to optimize the experimental generation of ASH, and its continuous introduction into the plasma using a Philips HG system (see Fig. 1). In a second step, the following organised media: Triton X-100, CTAB, and DDAB were assayed and compared with aqueous ASH, generation. Preliminary’ studies showed that CTAB produces a slight precipitate with potassium iodide, while TX-100 did not produce any improvement in the ICP-AES signal (as observed by AAS). Therefore, an
-Flame
Fig. 2.
Determination of arsenic
Table 3. Optimum
1521
operating conditions measurements
Plasm0 Coolant gas flow Pressure of Ar in nebuliser Forward power
in ICP-OES
11 l./min 35 psi 1.5 kW
Hydriak generation system 0.2
0.4
0.6
0.6
1
1.2
1.4
1.6
Fig. 3.
organized media of vesicles of DDAB finally selected for further experiments.
The criterion for optimization was taken to be the best signal-to-background ratio, SBR = (1, - &)/lb (where 1, is the intensity for the sample, and Ib for the blank). Selection of the analytical line. The two most sensitive analytical emission lines for arsenic were tested: 193.695 nm and 197.198 nm. The observed values of SBR were 7.2 for the former and 6.5 for the latter analytical line, using a 50-ppb standard. Optimization of chemical parameters for ASH, generation. Optimum hydrochloric acid and NaBH, concentrations were investigated using 50 ppb arsenic, a 0.5% potassium iodide solution as reducing agent and working at both emission lines sequentially. A univariant search was used in this instance to establish the optimum hydrochloric acid and NaBH, concentrations in the aqueous phase. The emission intensities obtained for each hydrochloric acid and NaBH, concentrations, are given in Fig. 3. It can therefore be seen why 1M hydrochloric acid and 1.3% NaBH, were selected for the subsequent work. It was observed that the ICP signal became higher with higher DDAB concentration. However, above 8.4 x 10W4M of the surfactant a precipitate formed in the presence of potassium iodide. Therefore, reductants other than potassium iodide were assayed in order to reduce As (V) to As (III). The behaviour of potassium 30
0’ 0
x
x-x
0.2
0.4
0.6 0.8 56 Reductor
Fig. 4.
concentration uptake rate acidity uptake rate cont.
1.3% m/v in 0.1% NaOH 2.1 ml/min 1M in HCl 4.2 ml/min 8.4 x lo-‘M
was
Optimization of the instrumental parameters
y 25 F
NaBH, NaBH, Sample Sample DDAB
iodide, titanium (III) chloride and sodium sulphite compared in vesicles of DDAB. Figure 4 summarizes the observed results which showed that the signal obtained when using TiCl, or Na,SO, was always lower than that obtained for potassium iodide. For this reason, a potassium iodide concentration of 0.5% was selected to ensure that all arsenic is in the As (III) form. The overall optimum conditions found are summarized in Table 3, and they happened to be the same in both media tested (water and DDAB). Analytical performance, characteristics of the DDAB-HG-ICP-AES determination of arsenic. Using the optimum conditions previously obtained (Table 3), sensitivity, selectivity, and precision of As determination by HG-ICP-AES using aqueous or DDAB media were evaluated. Calibration graphs obtained in both media are shown in Fig. 5 and show that the sensitivity (slope of calibration curve) is twice when using vesicles; the linear analytical range obtained extended from 6 ng/ml to, at least 100 pg/ml using DDAB and from 13 ng/ml to 100 pg/ml for water media, while the detection limts (3 bi,) were 0.6 ng/ml for BDDA and 1.3 ng/ml for aqueous media. The precision was evaluated by analysing ten replicates of a solution containing 50 ng/ml of arsenic with ASH, generation from aqueous and DDAB media. Relative standard deviation observed were f 3.2% using aqueous and f 2.3%
x
1
, 1.2 Cont. As (ppb)
Fig. 5.
B. AUPUN FERNNiDEz
1522
in DDAB media. Therefore, the precision of amine generation (see Table 2 for similar AAS measurements) seems to be clearly improved by adding vesicles to aqueous solutions. Interference
Table 4. Interferences of some elements in As determination by HG-ICPAES
Bi
Ca co Cr cu Fe
K Li Mn Na Ni Pb Sb Sn Te Ti Zn
Relationship 1:lOOO 1:500 1:300 1:lOO 1:50 1:lOOO 1:500 1:lOOO I:1000 1:500 1:lOOO 1:500 1:lOOO I:500 I:300 I:1000 1:500 1:300 1:lOOO 1:lOOO I:500 1:lOOO 1:lOOO 1:500 1:lOOO 1:500 1:lOOO I:300 1:200 1:lOO 1:500 1:300 1:lOOO 1:lOOO 1:sOO I:300 1:lOOO 1:500
Table 5. Results of the analysis of some sediments (reference lll&lidS)
Sediment CRM 211 CRM 320
studies
In order to investigate potential interferences from aqueous and DDAB media, solutions containing 25 ng/ml of arsenic (III) with various amounts of foreign ions were prepared. These solutions were analysed for As following the recommended procedure in the presence and absence of DDAB vesicles. All the results obtained are summarized in Table 4 in a comparative manner. As it can be seen in this table, some interferences are reduced (as in the case of Bi, for instance) or eliminated (in the case of Ti) when hydride generation is performed from DDAB vesicles. There is, however, an important interference in DDAB: when
Interferent
et a/.
Recovery (1) (Water) 50 52 54 55 85 110 101 101 108 93
Recovery (2) (DDBA medium) 15 98
98
91
:: 103 101 100 115 106 105 100 111 96 91 111 99 99 100 102 105 97 95 122 102 100 65 15 103 100
102 100
90 100 100 101
Observed value Wg) 41.6 f 1.0 14.1 f 3.3
Certified value 01B/B) 41.3 f 1.6 16.1 f 3.4
Sb is present in concentrations above 20 pg/ml, it produces precipitates (with DDBA) which seriously inhibit the introduction of the sample due to blocking of tubing. Results show that in general terms the method proposed using DDAB seems to be more selective than the method using aqueous solution and thus it can be favourable for the determination of As in real samples. Application to real samples
In order to test the analytical usefulness of the proposed method for the determination of low levels of As in environmental samples, we analysed two different certified sediments, namely, CRM 227 and CRM 320 (estuarine and river sediment respectively) from B.C.R. (E.C.). Sample dissolution and preparation was carried out as described elsewhere” with hydride production and determination using the developed methodology. The results obtained are given in Table 5 (each value given is the average of three determinations) and demonstrate that the DDABHG-ICP-AES procedure proposed here provides good accuracy and precision for this type of analytical problem. CONCLUSIONS
Long-chain surfactants have been proposed to improve the sensitivity of many analytical methods of atomic spectrometry. Sample transport in nebulizers and atomization efficiency enhancements in the flame had been invoked to rationalize the observed sensitivity improvements. Other workers, however, found very little or no sensitivity enhancement by surfactants, while recent research on the topic’s~‘9tends to indicate that transport efficiency does not increase on addition of most varied surfactants.” Micelles thus appear to have limited application for ICP-AES conventional nebulisation.” Conversely, micelles, and other organized media such as vesicles, could be most helpful in atomic methods based on hydride generation. As we have shown in a preliminary communication”j the hydride generation is a chemical
Determination
reaction which can be improved in the presence of such “organized” media, at least from two different points of view: (a) Micelles, vesicles, etc., are able to “organize” reagents at a molecular level, as has been repeatedly shownM for coloured and photoluminiscent reactions. This could affect sensitivity (effective concentration of reactants is higher in the microenvironment created by the organized medium than in water) and selectivity of hydride generation (i.e., by selecting oppositely or identically charged micelles for the analyte and interferences respectively). (b) Not only thermodynamic constants but kinetics of the reactions can also be altered by organized media. Our results for As have shown that, at least modestly, all the expected improvements (detection limits, selectivity and precision) have been realized using vesicles of DDAB in the continuous (and batch) generation of arsine. The “organized medium”-enhanced technique for hydride generation allows a fast and precise determination of As by ICP-AES and could be extended to other hydride forming elements. Moreover, analysis of environmental samples using the developed method gives results close to the certified values. Speciation of As is an important requirement for this application; this challenge could be approached using micellar chromatography” for separation of species and the proposed DDAB-arsine generation ICP-AES for detection. Research in this direction of As speciation is currently in progress in our laboratory.
of arsenic
1523 REFERENCES
1. A. G. Howard and M. H. Arbab-Zavar, Analyst, 1981, 106, 213. 2. R. J. Watting and A. R. Collier, ibid., 1988, 113, 345. 3. M. Thompson, B. Pahlavanpour, S. J. Walton and G. F. Kirkbright, ibid., 1978, 103, 568. 4. M. Thompson, B. Pahlavanpour, S. J. Walton and G. F. Kirkbright, ibid., 1978, 103, 705. 5. H. Thompson and B. Pahlavanpour, Anal. Chim. Acta, 1979, 109, 251. 6. E. De Oliveira, J. W. McLaren and S. S. Berman, Anal. Chem., 1983, 55, 2047. 7. R. R. Liversage, J. C. Van Loon and J. C. De Andrade, Anal. Chim. Acta, 1984, 161, 275. 8. G. S. Pyen and R. F. Browner, Appl. Spectrosc., 1988, 42, 508. 9. F. Nakata, H. Sunahara, H. Fujimoto, M. Yamamoto and T. Kumamaru, J. Anal. At. Spectrom., 1988,3,579. 10. T. Nakahara, Appl. Spectrosc., 1983, 37, 539. 11. J. Wilson Wershey and P. Keliher, Spectrochim. Acta., 1986, 41B, 713. 12. L. Halic and G. H. Rusell, Analyst, 1986, 111, 15. 13. A. Menendez Garcia, J. E. Sanchez Uria and A. SanzMedel, J. Anal. At. Spectrom., 1989, 4, 581. 14. J. Aznarez, F. Palacios, M. S. Ortega and J. C. Vidal, Analyst, 1984, 109, 713. 15. P. Changkan, Z. Guangzuan and R. F. Browner, J. Anal. At. Spectrom., 1990, 5, 537. 16. R. Femlndez de la Campa, C. Garcia Ortiz, M. Diaz Garcia and A. Sanz-Medel, Conference presentation, XII National Congress of Spectroscopy, Barcelona (Spain), 1990. 17. T. C. Dymott, International Laboratory, March 26, 1989. 18. J. Mora, A. Canals and V. Herandis, J. Anal. At. Spectrom., 1991, 6, 143. 19. Z. Yin Yan and Z. Wuming, ibid., 1989, 4, 797. 20. A. Sanz-Medel, P. L. Martinez and M. E. Diaz Garcia, Anal. Chem., 1987, 59, 774. 21. W. L. Hinze and D. W. Armstrong (ed.), Ordered Media in Chemical Separations, ACS Symposium Series 342, Washington 1987.