On-line solid phase extraction system using PTFE packed column for the flame atomic absorption spectrometric determination of copper in water samples

Transcription

1 Talanta 54 (2001) On-line solid phase extraction system using PTFE packed column for the flame atomic absorption spectrometric determination of copper in water samples Aristidis N. Anthemidis, George A. Zachariadis, John A. Stratis * Department of Chemistry, Laboratory of Analytical Chemistry, Aristotle Uni ersity, Thessaloniki 54006, Greece Received 14 November 2000; received in revised form 30 January 2001; accepted 16 February 2001 Abstract A new flow injection on-line adsorption preconcentration system adapted to flame atomic absorption spectrometry (FAAS) for copper determination at the g l 1 level was developed. Polytetrafluoroethylene (PTFE) turnings packed in a mini-column were used as sorbent material. The copper ammonium pyrrolidine dithiocarbamate (APDC) complex was sorbed on the PTFE turnings, from which it could be eluted on-line instantly by isobutyl methyl ketone (IBMK) into the flame at a flow rate of 2.3 ml min 1. The system was optimized and offered good performance characteristics with practically unlimited life time, greater flow rates and improved flexibility, as compared with other sorbent materials and the knotted reactor preconcentration systems. With 1 min preconcentration time, and a sample frequency of 40 h 1, the enhancement factor was 340, which could be further improved by increasing the preconcentration time. The detection limit was c L =0.05 g l 1, and the precision was 1.5%, at the 2.0 g l 1 Cu level. The method has been applied successfully to the analysis of potable, river and seawater, and its accuracy was tested by the analysis of certified reference materials and by recovery measurements on spiked samples. No significant interferences exist from other substances usually occurring in natural water Elsevier Science B.V. All rights reserved. Keywords: Copper; On-line preconcentration; Atomic absorption spectrometry; PTFE 1. Introduction Flame atomic absorption spectrometry (FAAS) is one of the most extensively used techniques for various elements determination with significant * Corresponding author. Tel.: ; fax: address: jstratis@chem.auth.gr (J.A. Stratis). precision and accuracy. However, there are many difficulties to determine traces of heavy metals in environmental samples due to insufficient sensitivity or matrix interferences. In manual batch methods, analyte preconcentration and matrix separation is time and reagent consuming with a risk of losses or contamination [1]. A powerful technique to effectively enhance atomic absorption spectrometry s sensitivity and /01/$ - see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S (01)

2 936 A.N. Anthemidis et al. / Talanta 54 (2001) selectivity is flow injection on-line analyte preconcentration and matrix separation [2,3]. Ruzicka and Arndal [4] have shown that sorbent extraction may be used as an alternative to solvent extraction and also, they highlighted the versatility of FI systems coupled with sorbent packed columns. Sorbent extraction is achieved with a hydrophobic sorbent solid phase packed in a column, which collects on its surface the analyte complexes formed in the aqueous phase. The preconcentrated complex can be eluted later (by changing the polarity of the carrier) into a small volume of a water-immiscible organic solvent. Various attempts have been made to improve performance and efficiency of the on-line preconcentration systems for FAAS since their introduction by Ruzicka and Arndal [4]. During the last 10 years, a number of methods have been reported to adapt FI on-line preconcentration with FAAS, for the determination of copper and other heavy metals in various samples. These methods usually make use of a packed column (PC) filled with a suitable stationary phase like C 18 [4 6], various resins like Amberlite XAD-2 [7], Chelex [8,9] etc. Alternatively, a knotted reactor (KR), which consists from a long PTFE tube properly knotted, can be applied as a sorbent surface [10 12]. In both cases, the metallic cation forms a complex with a chelating agent which is quantitatively sorbed by the column material or on the surface of PTFE tube during a fixed sample loading period. Then, the retained complex is eluted by using MeOH, EtOH [5,6], CH 3 COONH 4, [8,9] in case of packed column reactor, or by using isobutyl methyl ketone (IBMK) [11] in case of knotted reactor. Finally, the eluent is injected in the flame or the graphite tube of an atomic absorption spectrometer. In case of copper determination by on-line preconcentration FI-AAS, the most efficient chelating agents are dithiocarbamates like diethyldithiocarbamate (DDC) [5,11,13,14], diethylammonium-n,n -diethyldithiocarbamate (DDTC or DDDC) [6,15], ammonium pyrrolidine dithiocarbamate (APDC) [12,17] and dithiophosphates like ammonium diethyldithiophosphate (DDPA or DDTP) [15,16]. In addition, the standard method for determination of trace metals in aqueous samples [18] recommends the use of APDC for complex formation, extraction of the metal complex with IBMK, and subsequent determination by FAAS. Thus, in this work we combined the advantages of both approaches using for the first time a packed column filled with PTFE turnings, and the system APDC IBMK to the on-line preconcentration of copper, and subsequent FAAS determination. The proposed packed column produces significantly lower back-pressure than other sorbent materials, so the sample loading flow rates that can be applied are much higher. Also, the problem of weak retention of sorbed complexes in knotted reactors is eliminated. In addition, due to the very large effective surface of PTFE turnings, practically there is no limitation for the sample loading volume. The developed manifold is simple, low cost, and very suitable for routine operation. Moreover, the proposed method is applicable to various water samples, with very good analytical figures of merit, comparable to those obtained even by using electrothermal atomic absorption spectrometry (ETAAS). 2. Experimental 2.1. Instrumentation A Perkin-Elmer model 5100 PC flame atomic absorption spectrometer equipped with a deuterium arc background corrector was used as a detector. A copper hollow cathode lamp, was used as light source operated at 15 ma. The wavelength was set at nm resonance line and the slit at 0.7 nm. A time-constant of 0.2 s was used for peak height evaluation. The flame conditions were slightly leaner than those recommended by the manufacture, to compensate for the effect of IBMK, which serves as additional fuel. The air flow rate was set at 8.6 l min 1 and acetylene flow rate at 0.8 l min 1, while the resulting free uptake rate was 6.4 ml min 1. The flow injection manifold, which is shown schematically in Fig. 1, was adapted to the nebulizer system of the above spectrometer. It was consisted from two peristaltic pumps (Watson

3 A.N. Anthemidis et al. / Talanta 54 (2001) Marlow model 205U/BA), a two-position sixport injection valve (Labpro, Reodyne, USA) and a PTFE packed column for the on-line preconcentration of the analyte. Tygon pumping tubes were used for delivery of all aqueous streams. A displacement bottle (Tecator, Hoganas, Sweden) was used for delivery of the IBMK stream, because this organic solvent is not compatible with Tygon tubes. All other tubes were made of 0.5 mm i.d. PTFE. The length of PTFE tube between the confluence point of sample with APDC stream and the injection valve was as short as possible (ca. 1 cm), because the PTFE material of the tube adsorbs the complex [14] thus affecting the precision of the determination. Moreover, to minimize the dispersion, the connection tubes between column edges and the injection valve were also short and the PTFE capillary, which connected the above valve with the nebulizer, was 12 cm (0.35 mm i.d.). The column (Jour Research) was made from polyetheretherketone (PEEK), 30/4.6 mm i.d. The solid sorbent material was 600 mg of PTFE turnings, with a mean width of 0.1 mm, which were mechanically produced in our laboratory, by using a lathe. The turnings were washed thoroughly by ethanol followed by 1 M HNO 3 and finally with deionized water. In order to avoid any instability while reversing the direction of aqueous and organic streams the material was firmly packed and polypropylene O-rings were placed at both sides to keep the packing material in place. The column was initially flushed with deionized water and subsequently with IBMK. The performance of the column was stable during all experiments with practically unlimited lifetime Reagents Fig. 1. FI manifold for on line sorption preconcentration FAAS system. APDC, 0.05% APDC solution; FAAS, flame atomic absorption spectrometer; P1, P2, peristaltic pumps; DB, displacement bottle for IBMK; PC, mini-column packed with PTFE turnings; W, waste. All chemicals were of analytical reagent grade and were provided by Merck (Darmstadt, Germany) and doubly deionized water was used throughout for dilutions. Working standard solutions of copper were prepared immediately before use, by stepwise dilution of a 1000 mg l 1 stock standard solution (Merck) to the required g l 1 levels. The chelating reagent, 0.05% mv 1 APDC was prepared fresh daily by dissolving the appropriate amount of ammonium pyrrolidine dithiocarbamate (Merck) in deionized water. Isobutyl methyl ketone was used, after saturation with deionized water, without any other purification. The following standard reference materials were used to verify the accuracy of the proposed method: NIST (National Institute of Standard

4 938 A.N. Anthemidis et al. / Talanta 54 (2001) Table 1 FI operation sequence for on-line adsorption preconcentration and elution Flow rate (ml min 1 ) Step Valve position Time (s) Medium delivered Pump 1 Pump 2 Loading Fill 60 Sample APDC Elution Injection 30 IBMK and Technology, USA) CRM 1643b Trace elements in water, and NRCC (National Research Council Canada, Canada) CASS 3 Coastal seawater Procedure The method consists of two main steps, which are shown in Fig. 1. Details for the operation sequence of the preconcentration system are given in Table 1. In the loading step, the injection valve was in the fill position, while sample or standard solution and APDC streams were propelled by pump 1 and merged together, before passing through the column. The Cu PDC complex formed was collected on the surface of the PTFE turnings in the column. During the elution step, the valve was in the injection position, IBMK is propelled to the column through activation of pump 1, and the eluted Cu PDC complex is delivered to the FAAS nebulizer. In order to minimize dispersion, the column was arranged in the manifold so that the IBMK eluent flowed through it in reverse direction than that of the sample and reagent. The peak height of the transient signal was proportional to copper concentration in the sample, and was used for all measurements. The recorded peaks were sharp and the baseline was stable. Five replicate measurements per sample were made in all instances. Natural water samples were filtered before analysis through 0.45 m filters and acidified to 0.01 mol l 1 HNO 3 (ca. ph=2). Standard solutions of copper were also made up to ph=2 with 0.01 mol l 1 HNO 3, while reference materials were already acidified close to this ph. 3. Results and discussion 3.1. Optimization of chemical and FI ariables It is well known that the metallic complexes of dithiocarbamates are retained efficiently on the walls of the PTFE knotted reactors. The mechanism of the retention of the Cu PDC complex on the surface of PTFE turnings should probably be molecular sorption and not mechanical entrapment, as was further investigated by Fang et al. for cadmium DDC complex by using Scanning Electron Microscopy [10]. Preliminary experiments showed that lengths of the column between 3 and 5 cm had no effect to the signal obtained for a 20 g l 1 of Cu, thus a 3 cm column was used throughout. Also no detectable amounts of copper could be retained in the packed column in absence of the chelating reagent. Variation of APDC concentration in the range of % m V 1 had no effect on the preconcentration efficiency. Therefore, a 0.05% m V 1 solution of APDC in water was selected, and the flow rate of the APDC stream was fixed at 0.5 ml min 1 for further study. In order to ensure complete elution of the copper complex, the elution time was fixed at 30 s. The chemical and FI variables of the manifold were optimized using a standard solution 20 g l 1 of Cu(II). The sample s acidity is one of the critical variables for the formation of Cu PDC complex, and its adsorption on the sorbent surface. The effect of the ph was studied over the range from 0.7 to 6.7 by adjusting it in the copper solution with dilute nitric acid. The ph of the waste solution

5 A.N. Anthemidis et al. / Talanta 54 (2001) was similar to that of the sample, indicating that the introduction of the APDC solution did not affect significantly the acidity. The absorption signal is maximum and practically stable in a wide ph window varied between 2 and 5. This fact enables the use of the method directly in many aqueous samples after common acid preservation, without any laborious precise ph adjustment. The optimum ph range found to be inside to the range of the standard batch method [18]. The decrease of the efficiency at higher and lower ph values is probably due to non-complete complex formation, and sorption, and not to the function of the column PTFE material, which is very resistant to any acidic or alkaline solution. The effect of the sample flow rate during the loading step was studied in the range of ml min 1. The absorption was linear up to the whole range of sample flow rate, implying that the complexation is complete and the contact time is sufficient. This is a significant advantage of the proposed column preconcentration system over knotted reactor ones, which enables more sample to be loaded during a given loading time resulting in higher concentration efficiency. A flow rate of 12.0 ml min 1 was selected as a compromise between high sensitivity and low sample consumption. The influence of sample loading time was studied in the range from 10 to 240 s. The absorption was increased linearly up to 180 s and practically up to 240 s. This fact proved that even at long loading times and prolonged sample flow, no negative effect is produced from partial leaching of the complex [10]. Finally, a 60 s loading time was chosen as a compromise between medium sample consumption, high sensitivity and sufficient sampling frequency. Although ethanol or methanol could be used as good eluents, IBMK is considered to be a better eluent for FAAS, producing higher and sharpest signals, because of more intensive enhancement of the flame atomization. On the other hand, IBMK is almost completely immiscible with water and less polar solvent, producing lower dispersion and better elution characteristics. The effect of elution flow rate was studied in the range of ml min 1. Peak height goes through a maximum, at IBMK flow rate of 2.3 ml min 1, and above this value slightly decreases, due mainly to the higher dispersion at these elution flow rates. Thus, this flow rate was chosen as optimal Analytical performance of the on-line sorbent extraction FAAS system Under the optimum conditions described above, the analytical performance of the method is described in Table 2. Peak height absorption signals were used for the calculation of the enhancement factor, and for the calculation of the calibration curve and the precision of the method. The detection limit was calculated by the 3s criterion and was c L =0.05 g l 1. The determination limit (10s criterion) was c Q =0.15 g l 1 and the linear range of the method was varied between 0.15 and 15.0 g l 1. The features of the proposed method, showed that it is a convenient, low cost and flexible one, and has also some good advantages over other on-line preconcentration FI-FAAS methods reported in the literature for copper determination. In Table 3, the figures of merit of this an other selected methods, which use FAAS or ETAAS as detectors, are summarized for comparative purpose. With FAAS as detector, the detection limit Table 2 Analytical performance of the FI on line preconcentration FAAS method for copper determination Sample volume Loading time Sampling frequency Enhancement factor Detection limit (3s) Determination limit (10s) Precision, relative standard deviation (n=20) Linear range Regression equation (14 standards, n=5) Correlation coefficient 12.0 ml 60 s 40 h c L =0.05 g l 1 c Q =0.15 g l 1 s r =1.5% (2.0 g l 1 ) g l 1 A= [Cu] r=0.9996

6 940 Table 3 Comparative data from recent references on copper determination by FI-FAAS and FI-ETAAS methods after on-line preconcentration on packed column or knotted reactors (KR) Reference Analyte Chelating reagent Eluent Sorbent material Detector EF b LT c (s) SF d h 1 c L e gl 1 s r f % a Cu APDC IBMK PTFE FAAS [11] Cu Na DDC IBMK KR PTFE FAAS [5] Cu, Cd, Pb DDDC MeOH or C 18 FAAS EtOH [6] Cu, Cd DDTC MeOH C 18 FAAS [7] Cu HCl Amberlite XAD-2 FAAS [9] Cu, Pb, Cd, Zn CH 3 COONH 4 Chelex 100 or 122 resin FAAS [9] Cu, Pb, Cd, Zn 8 HQ CH 3 COONH 4 Controlled pore glass FAAS [15] Cu, Pb, Cd DDPA MeOH C 18 FAAS 35 20? [16] Cu, Cd, Pb DDTP MeOH C 18 FAAS 86? 29 7? [19] Cu, Pb, Cd Dithizone MeOH C 18 FAAS [20] Cu, Co, Hg, Rh, N N DDC HNO 3, HCl Controlled pore glass FAAS 18? [21] Cu, Cd, Co 1, 10 Phenathroline EtOH C 18 FAAS [22] Cu APDC IBMK Zeolite (Na LTA) FAAS ? [22] Cu APDC IBMK Zeolite (Na Fau) FAAS ? [12] Cu, Mn, Ni APDC MeOH KR PTFE ETAAS [17] Cu, Pb APDC MeOH C 18 ETAAS? 120? 0.5? [13] Cu, Cd Na-DDC MeOH C 18 ETAAS ? 0.024? [14] Cu, Cd, Pb, Ni Na-DDC EtOH C 18 ETAAS ? [23] Cu, Cd, Pb HNO 3 Superlig ETAAS A.N. Anthemidis et al. / Talanta 54 (2001) a This work. b Enhancement factor. c Loading time. d Sampling frequency. e Detection limit. f Precision, relative standard deviation.

7 A.N. Anthemidis et al. / Talanta 54 (2001) and the precision of the proposed method are lower than those obtained using a knotted reactor, or a packed column with other sorbent materials and are comparable to some other methods using ETAAS. When ETAAS is used as detector and PTFE KR for on line sorption, the detection limit can be lowered [12]. However, it should be taken in account that the proposed FI-FAAS method consists of simpler apparatus with easier manipulation of the system, much lower equipment and running costs Interference studies APDC does not form complexes with alkali and alkaline earth metals, which are usually found in high concentrations in natural waters. However, in order to identify their potential interference on the on-line sorption of the Cu PDC complex, high concentrations of them were tested. Na(I) and K(I) up to 1000 mg l 1 and Ca(II), Mg(II), Ba(II) up to 500 mg l 1, did not caused any significant interference. On the other hand it is well known that transition metals form strong complexes with APDC, so the recovery of 5 g l 1 copper solution was tested in presence of these metals. The results are summarized in Table 4. Only Ni(II), Hg(II) and Cd(II) were found to interfere significantly, at concentrations higher than 0.5 mg l 1, but such levels are not likely to be found in natural waters. However, in case of wastewaters or heavily polluted waters, the interference effects could be minimized by using longer (5 cm) packed column without significant changes in the performance characteristics. In addition a number of common anions like Cl,Br,I,SCN,SO 4 2, HCO 3,NO 3 were tested and found that they did not interfere at concentrations up to 1000 mg l Application to aqueous samples The proposed method was applied to the analysis of tap water, river water and coastal seawater samples collected from rivers and gulfs of Northern Greece. The accuracy was tested by determining the copper content of standard reference Table 4 Effect of interferents on the determination of 5 g l 1 Cu(II) by the proposed FI-FAAS on-line preconcentration method Interferent Concentration (mg l 1 ) Recovery (%) Al(III) Cd(II) Co(II) Cr(III) Cr(VI) Fe(III) Hg(II) Mn(II) Mo(VI) Ni(II) Pb(II) Zn(II) seawater samples and also by calculating the recovery of copper from spiked tap, river and seawater samples. All the determinations were performed by using aqueous standard solutions for calibration, and without further matrix specific optimization. The certified values and the found results are presented in Table 5. It can be seen that the recovery obtained to natural water samples (94 102%) is comparable to that for the reference materials (97 98%), and the performance of the method is very good in all types of natural waters. Finally the relative standard deviation for copper determination in the examined real samples ranged between 3 and 5%.

8 942 A.N. Anthemidis et al. / Talanta 54 (2001) Table 5 Analytical results of copper determination in various reference materials and natural water samples Certified value ( g l 1 ) Added ( g l 1 ) Found* ( g l 1 ) Recovery (%) Reference material CASS CRM 1643b Water samples Tap Tap River River Coastal seawater Coastal seawater * Mean value standard deviation based on five replicate determinations. 4. Conclusions The introduction of PTFE as a column packing material in the on-line adsorption preconcentration system was found to improve significantly the flow injection FAAS determination of copper. The developed manifold permitted high sample loading rates in order to achieve higher preconcentration. The obtained detection limit was thus comparable to those obtained by ETAAS. The proposed method proved to be rapid, reliable and flexible for copper determination in a variety of natural water samples with limited interferences. Additional work is in progress on testing the PTFE as packing material and its performance with other chelating agents for on-line determination of other metals in water samples and in more. complex matrices such as biological liquids. References [1] Y.A. Zolotov, N.M. Kuzmin, Preconcentration of Trace Elements, Elsevier, Amsterdam, [2] Z. Fang, Flow Injection Atomic Absorption Spectrometry, John Wiley & Sons Ltd., West Sussex, England, [3] Z. Fang, Spectrochim. Acta Part B 53 (1998) [4] J. Ruzicka, A. Arndal, Anal. Chim. Acta 216 (1989) 243. [5] Z. Fang, T. Guo, B. Welz, Talanta 38 (1991) 613. [6] S. Xu, M. Sperling, B. Welz, Fresenius J. Anal. Chem. 344 (1992) 535. [7] S. Ferreira, V.A. Lemos, B.C. Moreira, A.C. Spinola Costa, R.E. Santelli, Anal. Chim. Acta 403 (2000) 259. [8] Z. Fang, S. Xu, S. Zhang, Anal. Chim. Acta 164 (1984) 41. [9] Z. Fang, J. Ruzicka, E.H. Hansen, Anal. Chim. Acta 164 (1984) 23. [10] Z. Fang, S. Xu, L. Dong, W. Li, Talanta 41 (1994) [11] H. Chen, S. Xu, Z. Fang, Anal. Chim. Acta 298 (1994) 167. [12] E. Ivanova, K. Benkhedda, F. Adams, J. Anal. At. Spectrom. 13 (1998) 527. [13] M. Wang, A.I. Yuzefovsky, R.G. Michel, Microchem. J. 48 (1993) 326. [14] M. Sperling, X. Yin, B. Welz, J. Anal. At. Spectrom. 6 (1991) 295. [15] Ma.W. van Mol, F. Adams, Anal. Chim. Acta 285 (1994) 33. [16] S.M. Sella, A.K. Avila, R.C. Campos, Anal. Lett. 32 (1999) [17] R.H. Atallah, G.D. Christian, A.E. Nevissi, Anal. Lett. 24 (1991) [18] Standard Methods for the Examination of Water and wastewater, APHA, AWWA, WEF, Washington, 19th Edition, [19] S. Kartikeyan, B. Vijayalekshmy, S. Chandramouleeswaran, T. Prasada Rao, C.S.P. Iyer, Anal. Lett. 30 (1997) [20] A. Townshend, K.A.J. Habib, Microchem. J. 45 (1992) 2. [21] A. Ali, X. Yin, H. Shen, Y. Ye, X. Gu, Anal. Chim. Acta 392 (1999) 283. [22] Y.P. de Pena, W. Lopez, J.L. Burguera, M. Burguera, M. Gallignani, R. Brunetto, P. Carrero, C. Rondon, F. Imbert, Anal. Chim. Acta 403 (2000) 249. [23] E. Hosten, B. Welz, Anal. Chim. Acta 392 (1999) 55.