Talanta 59 (2003) 201 Á/205 www.elsevier.

com/locate/talanta

Solid-phase extraction of bismuth, lead and nickel from seawater using silica gel modified with 3aminopropyltriethoxysilane filled in a syringe prior to their determination by graphite furnace atomic absorption spectrometry
Nilgun Tokman, Suleyman Akman 1, Mustafa Ozcan
Faculty of Science and Letters, Department of Chemistry, Istanbul Technical University, Maslak, Istanbul 34469, Turkey Received 7 May 2002; received in revised form 3 September 2002; accepted 4 September 2002

Abstract In this study, the use of syringe filled with sorbent for the separation and enrichment of bismuth, lead and nickel prior to their analysis by graphite furnace atomic absorption spectrometry was described to substitute for batch and column techniques. The method proposed in this paper was compared with column technique with respect to easiness, fastness, simplicity, recovery and risk of contamination. The syringe was filled with 0.5 g of sorbent and in order to retain the analyte elements, 5 ml of sample solution (pH ]/5) was drawn into the syringe to 15 s and discharged again in 15 s. Then, 2.0 M of HCl, as the eluent, was drawn into the syringe and ejected back to desorb the analyte elements. At optimum conditions, the recoveries of Bi, Pb and Ni were 95 Á/99% with relative standard deviations (RSDs) of around 9/2%. Detection limit (d ) was 0.5 mg l( 1 for Bi, Pb and Ni, respectively. The elements could be concentrated by drawing and discharging several portions of sample successively but eluting only one time. Bi, Pb and Ni added to a seawater sample were quantitatively recovered ( /95%) with low RSD values of around 9/2 Á/3%. The risk of contamination is less than that with the column technique. In addition, it is much faster, simpler, easier, more practical and handy compared with column technique. # 2002 Elsevier Science B.V. All rights reserved.
Keywords: Sorbent in syringe; Solid-phase extraction; Graphite furnace atomic absorption spectrometry; Bismuth; Lead; Nickel

1. Introduction Even with the high sensitivity of graphite furnace atomic absorption spectrometry (GFAAS), the concentration of elements in many matrices, e.g. seawater, may be too low to be determined directly and consequently, a concen-

1 Corresponding author. Tel.: '/90-212-285-3160; fax: '/90212-285-6386 E-mail address: akmans@itu.edu.tr (S. Akman).

0039-9140/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 0 2 ) 0 0 4 7 9 - 4

10]. and dried at 150 8C for 1 day. not practical and not handy. pipettes. 3-aminopropyltriethoxysilane (Aldrich) was used as supplied [12]. Therefore. Bismuth. Instrumental parameters were adjusted according to the recommendations of manufacturer’s recommendations.1 M NaOH and controlled using WTW pH340-A/SET2 pH meter. As a result. and methanol was boiled off under (2) Two off-line methods have been used for the separation and enrichment of analyte elements prior to their analysis by AAS. and the concentration factor is limited. because of the low acidity of the silanol groups as well as the less pronounced donor properties of the surface oxygen atoms. / Talanta 59 (2003) 201 Á/205 tration procedure is often necessary. 35 Á/70-mesh) was used as the supporting material. Apparatus and chemicals A Perkin-Elmer Zeeman 3030 atomic absorption spectrometry equipped with an HGA 600 Graphite Furnace was used for the analysis of bismuth. All chemicals were of analytical reagent grade (Merck). sorption of trace elements on an ion-exchanger or chelating collector has been of greatest interest and various collectors have been proposed for this purpose. Among the preconcentration/separation techniques. Batch method is quite slow. lead and nickel prior to their analysis by GFAAS. but it is still an opened . Á/SH. Á/ CH has been a subject of interest [9.1 M HCl or 0. It does not swell and has good mechanical strength and thermal stability. In this study. Preparation of silica gel modified with 3aminopropyltriethoxysilane The silica gel modified with 3-aminopropyltriethoxysilane was prepared as described in a previous paper [13]. Much higher concentration factors can be reached.2. lead and nickel hollow cathode lamps were used as spectral radiation sources.7] and activated carbon [8] with or without chelating groups have been the most widely used collectors for this purpose. the collector. Stock solutions (1000 mg l (1) of Bi.1. In addition. interferences due to complex matrices of many samples cannot always be controlled even with stabilized temperature platform furnace (STPF) concept [1] together with a suitable modifier and separation procedure may be required. Silica gel can be directly used as an adsorbent. Besides. etc. Silica gel (Merck. Tokman et al. silica gel modified with 3-aminopropyltriethoxysilane. 2. the interactions with the silica surface are rather weak. 2. for most metal ions. Polymeric resins [1 Á/5]. vessels. was filled in a syringe for the separation and enrichment of bismuth. there is a risk of contamination because it is an opened system that uses glasses. Experimental 2. washed with distilled water.202 N. At first silica gel was prepurified with hydrochloric acid (1/1). was added to a silica gel/methanol suspension (1 mmol silane per g of silica gel). Pb and Ni were made from titrisol (Merck) and further diluted with distilled deionized water daily to prepare reference solutions. as an alternative method to the column and batch techniques. The silica gel modified with 3-aminopropyltriethoxysilane. silica gel modified with organic functional groups such as Á/NH2. flame and GFAAS combined with preconcentration/separation techniques have been widely used for the determination of trace elements. silica [6. (C2H5O)3Si(CH2)3NH2. lead and nickel in this work. There are numerous papers in the literature for the use of different collectors in the both techniques. ICP or other techniques. The pH adjustment was made by adding 0. However. Sorption mechanism of the modified silica gel for the metals was described by Koklu [11] as below: [M2' ] ' [NH2 ] (on surface) U[M2' NH2 ] (on surface) (1) system and uses glasses and pipettes. Column method is easier than batch method.

2. The seawater. 1). 3.0 M HCl in one time.0 M of HCl as the eluent was drawn into the syringe and ejected back to desorb the analyte elements. The residual silica was dried until the washing contained no dissolved materials. The pH of the sample was adjusted to the desired value with 0. Pb and Ni (analyte concentration: 0. 1.2. all the solution should be drawn above the top of sorbent. Picture of the syringe. The analyte elements in the discharge were determined by GFAAS and recoveries were calculated. The effect of pH of sample on the sorption of Bi. The syringe was filled with 0. 5 ml of different sample portions were passed through the sorbent successively but eluted with only 5 ml of 2. This shows that standards prepared in distilled water and most of the slightly acidic real samples (pH ]/5) could be passed through the sorbent without adjusting the pH. The elimination of pH adjustment step is advantageous because addition of reagents for pH adjustment and its control by soaking a pH electrode or a pH strip is not only a time-consuming procedure but may sometimes cause high blank values. i. If the sample was prepared in 0. For a better contact between sample solution (or eluent) and sorbent. flow rates of sample and eluent for quantitative sorption and desorption. The three Fig.5 g of (2 cm of height) silica gel modified with 3-aminopropyltriethoxysilane (Fig. Otherwise the analysis was not interferencefree and recoveries were low. 2. elements were quantitatively sorbed at pH ]/5. Bi. / Talanta 59 (2003) 201 Á/205 203 vacuum.05 mg l ( 1). In order to concentrate the analyte elements. We called this pull and push process as one cycle.e. Procedure for sorbent in syringe Cleaned porous disk filter was placed on the bottom of a 5 ml of syringe. . Pb and Ni is shown in Fig. 3.1.1 M HCl. 2. During this procedure. The washing step before elution process is especially important to remove the seawater matrix. and even irreproducible contaminations and consequently low precision. the analyte elements were sorbed (retained) by the sorbent. In order to achieve quantitative retention Fig. are very important parameters. Fivemilliliter of sample solution (for quantitative recovery pH ]/5) was drawn into the syringe in about 15 s and discharged again in 15 s. Tokman et al.1 mol l (1 of HCl or NaOH solutions. To find the retention yields.3. 2. Results and discussion 3. Effect of pH on sorption The sample was drawn into the syringe passing through the sorbent and discharged passing through the sorbent again one more time.N. Flow rates for sorption The adsorption and desorption kinetics. Then. standards and blanks were treated in the same manner. one or two times of distilled water was drawn and ejected to wash the collector. This is quite a low pH limit value for a quantitative sorption of cations for a sorbent. Before elution. Pb and Ni in the discharge were analyzed by GFAAS. The effect of the pH of sample on the sorption Bi. the retention for Bi was still 100% whereas it is reduced to 40 and 30% for Pb and Ni. respectively.

quantitative retention was achieved. 5 ml of different portions of matrix-free solutions were . lead and nickel are shown in Table 1.0100 0.05 mg l ( 1. sorption and desorption should be completed as fast as possible. i.01029/0. some small random changes of flow rates (i.0002 0. In addition. Small deviations from this rate did not influence the retention because for quantitative recovery the flow rate could be increased up to around 15 ml min (1 (10 s for drawing and 10 s for ejecting back for 5 ml) without problem. which is quite far from the applied flow rate of 10 ml min (1. analyte concentration: 0.01 mg l ( 1. / Talanta 59 (2003) 201 Á/205 Table 1 The effect of different eluents on elution of Bi. totally 30 s.e. since the flow rates in syringe method might not be precisely controlled.204 N.0100 0. Elution The eluent was then drawn into the syringe and ejected back to desorb the analyte elements. This means that 5 ml of solution contacted with the collector 15 s two times.0002 102 94 96 Five times of 5 ml 0. Elution of the three elements was successfully completed at a flow rate of 10 ml min (1. one pull and push procedure was enough to achieve quantitative desorption of analytes. If quantitative retention had not been obtained.4. several cycles of sampling but one cycle of elution enriched the analytes depending on the cycle ratios. pH 5.00969/0. The other acids could not provide a quantitative elution on which we did not make any overestimation. sample drawing and ejecting rates) should not influence the recovery.00949/0. In order to concentrate the analyte elements five times. The more concentrated acids did not change the elution yield or elution kinetics.e.e. flow rate) between sample (or eluent) and sorbent could be adjusted by pulling or pushing the syringe piston at different rates.0004 0. and elution of analyte elements even at fast flow (or drawing and ejection) rates.01 mg l ( 1 sample.e. The effects of different acids at different concentrations on desorption of bismuth. 3. N 0/3) Element Prepared (mg l( 1) Foundb (mg l ( 1) Recovery (%) Bi Pb Ni a b Ni 779/2 959/2 699/2 819/3 609/2 1 2 1 2 2 M M M M M HCl HCl HNO3 HNO3 CH3COOH 0.0100 0. the sample solution would have been passed through the sorbent more slowly or better successively a few times until the sorption was quantitative but in our system this was absolutely not needed.3. Enrichment and analysis in seawater If the concentrations of analytes are too low to be detected.e.0 M of HCl. i. which corresponds to a flow rate of 10 ml min (1. The contact time (i. we preferred the acidity of the eluent to be as low as possible not to cause any possible high blank values and interferences due to HCl. i. This is obtained only if the flow rates of sample and eluent are much slower than maximum limiting values for a quantitative sorption and desorption. N 0/3) Eluent Recovery (%) Bi Pb 649/2 1009/3 699/2 979/3 739/2 939/2 989/3 819/2 829/2 789/2 Table 2 Recoveries of the elements studied (sample volume: 25 mla. Although the higher concentrations of HCl achieved the same elution efficiency. Thus it is very important to know the fastest permissible flow rate limit for quantitative recovery. If 5 ml of sample solution (pH ]/5) was drawn into the syringe in 15 s and discharged again without waiting at the same rate. Tokman et al. Pb and Ni (analyte concentration: 0. 3. eluent: 5 ml of 2 M HCl. several portions of sample were drawn and discharged successively but the eluent was drawn and ejected only one time. but again there was no problem with a flow rate of 15 ml min(1. As can be seen from the results that the analyte elements retained by the collector sorbent could be quantitatively desorbed with at least 2. Only one cycle of elution. Average of three determinations with 95% confidence levels values after five times of concentration.

[11] U. H. Koklu. S. Atom. Fresenius’ J.0008 0. Soc. Tanada. Chem. Kiba. Soc. Anal. U. The method was applied to find the analyte elements in bosphorus seawater samples after five times of concentration. Bull. Anal. J. [13] H. 5 (1977) 1060 Á/1065. Anal. The Chemistry of Silica(Chapter 6). Seren. 25 ml of sample was passed through the sorbent and eluted with 5 ml of 2.7 96. It seems that the capacity of the modified silica. Jpn. It can be concluded that the syringe method is superior to the column technique in many respects in the separation and enrichment of bismuth. handy. Dog ˘ an. Ince. a Mean9/95% confidence level. 342 (1992) 560 Á/562. [9] I.032 mmol Cu g(1 modified silica [13]. five cycles for sorption and one cycle for elution). Soylak. [5] Y. Lett. New York. References [1] B. Akman. [3] M. Lett. Acta B 55 (2000) 1129 Á/1133. concentration factor 5 )/. which further shows that the matrix was successfully separated with this technique. lead and nickel. Ozcan. As shown from the results. 355 (1996) 665 Á/666. S. Chem. Inamura. Chem. It is possible to make the sorption process in the source of sample. drawn and ejected but eluted with 5 ml of 2. Koklu. Since five times of concentration was enough to detect the analyte elements in seawater. Chim. 622 Á/642. lead and nickel in seawater prior to their analysis by GFAAS. 26 (2000) 1169 Á/1174.0008 0. Sarac. Sci. Ekinci. Bull. S. Erbil. Basu. [4] S. M. The risk of contamination is quite low. G. more practical.0006 0. [12] C. 29 (1996) 635 Á/643. S. A. Inoue. [2] M.01549/0. Tokalıoglu. Anal.0 M of HCl only one time (i. Ozeroglu. Elci. 1979. eluent: 5 ml of 2 M HCl.K. Takai. S. the elements added to the seawater sample were quantitatively ( /95%) recovered which proves that the heavy matrix of seawater did not affect the quantitative separation of analyte elements. Akman. 348 (1994) 281 Á/283. The recoveries are given in Table 2. As can be seen from the results in Table 3.0006 Addeda (mg l ( 1) 0. Bakırcıoglu.010 0. Shinoda. 35 (2002) 1075 Á/1083. Akman. J. 4. T. L. Spectrochim. Fresenius’ J. Akman.00489/0.01609/0. Wiley Á/ Interscience. Anal. S. [7] S. Acta Turcica 12 (1984) 265 Á/272. Tokman et al. Conclusion Syringe filled with sorbent was perfectly applied for the separation and enrichment of trace bismuth. U. The background values of AA signals were in blank level. Pb and Ni in spiked seawater sample (sample volume: 25 ml (five times of 5 ml). Ozcan. Chem. S. which was found as 0. Jpn. Acta B 55 (2000) 1491 Á/ 1495.010 Founda (mg l ( 1) 0. Kartal. The syringe method is more .e. Totally.3 Concentration of spike in seawater.0 M HCl. pH 5.010 0. Koklu. Anal. 53 (1980) 1242 Á/1246.0010 Recovery (%) 102. Terada. Chem.00669/0. / Talanta 59 (2003) 201 Á/205 205 Table 3 Determination of Bi. N. U. Spectrochim. Akman. [10] K. 7 (1992) 187 Á/189.01529/0. [6] K. Ralph. C. [8] M. C. Fresenius’ J.00609/0.4 96. Konar. Koklu. Anal. was enough to collect the analyte elements from the seawater quantitatively without problem. pp. we did not attempt to reach higher concentration factors.0007 0.N. Ince. N 0/3) Element Bi Pb Ni Seawater (mg l ( 1) 0. Spectrom. quantitative recoveries were achieved with reasonable (9/2 Á/3%) RSD values. S. faster and easier than the column technique.