Vapor Generation Atomic Absorption Spectrometric Determination of

Apr 29, 2009 - ... in Environmental Samples with in-Line Anion Exchange Separation. Citing Articles; Related Content. Citation data is made available ...
0 downloads 9 Views 276KB Size
Anal. Chem. 2009, 81, 4249–4255

Vapor Generation Atomic Absorption Spectrometric Determination of Cadmium in Environmental Samples with In-Line Anion Exchange Separation Samuel B. Adeloju* and Yanlin Zhang NanoScience and Sensor Technology Research Group, School of Applied Sciences and Engineering, Monash University, Gippsland Campus, Churchill, Victoria 3842, Australia A novel in-line anion exchange separation method is described for the removal of Mn2+, Fe3+, Ni2+, Co2+, Cu2+, Zn2+, and Pb2+ from Cd2+ and for its subsequent determination by vapor generation atomic absorption spectrometry. High Cd2+ retention efficiency and maximum exclusion of other metal ions were achieved by using Cl-, Br-, or I- loaded strong base anion exchange resin and fiber in chloride medium. SO42- and NO3affected the retention of Cd2+ on Cl- or Br- loaded exchangers but were beneficial for elution from Iloaded exchangers. Fast and efficient elution from the exchangers was achieved by using ethylenediamine solution. The removal of Zn2+ was unnecessary as it prevents a decline in the sensitivity of cadmium response when ethylenediamine and high acid concentration were used. Under optimum conditions, the achievable detection limits (3σ) with sample loadings of 0.48 and 9.6 mL were 40 ng L-1 and 3 ng L-1, respectively. Interference of residual organic matrix, in sample digests, with the separation and retention of cadmium were eliminated by use of microwave assisted digestion. The method was successfully applied to the determination of Cd in river sediment, fish liver, and water samples. Cadmium is of a major environmental significance because of its high toxicity and its determination has attracted considerable interest.1 Atomic absorption spectrometry (AAS), especially by flame AAS (FAAS) and graphite furnace AAS (GFAAS), is a commonly used method for cadmium determination because of its simple sample treatment requirement and low interference from co-existing metal ions.2,3 However, FAAS is often not sensitive enough for the determination of cadmium in various environmental samples, and GFAAS is slow and requires special care to ensure * To whom correspondence should be addressed. E-mail: sam.adeloju@ sci.monash.edu.au. Phone: +61 3 9902 6450. Fax: +61 3 9902 6738. (1) Matusiewicz, H.; Kopras, M.; Sturgeon, R. E. Analyst 1997, 122, 331–336. (2) Lemos, V. A.; Santelli, R. E.; de Carvalho, M. S.; Ferreira, S. L. C. Spectrochim. Acta B 2000, 55, 1497–1502. (3) Vin ˜as, P.; Pardo-Martı´nez, M.; Herna´ndez-Co´rdoba, M. Anal. Chim. Acta 2000, 412, 121–130. 10.1021/ac802618u CCC: $40.75  2009 American Chemical Society Published on Web 04/29/2009

reliable determination.2-4 On the other hand, vapor generation AAS (VGAAS) is a fast and sensitive method for Cd determination.5-25 However, many base metal ions, such as Fe3+, Ni2+, Cu2+, Zn2+, and Pb2+, have been found to interfere with the vapor generation of Cd.8,13,14,17,20,21 Unfortunately, no effective method has been developed for the elimination of these interferences or their complete removal. As these interfering ions can be easily reduced by borohydride and their chemical properties are similar to that of Cd2+, their discrimination and separation, especially of Zn2+, are difficult. Duan and co-workers23 reported the use of Cyanex 923 as a selective chelating agent for the separation of interfering ions from Cd2+. They adequately separated Fe3+, (4) Bianchin, L.; Nadvorny, D.; Furtado da Silva, A.; Vale, M. G. R.; Messias da, Silva, M.; dos Santos, W. N. L.; Ferreira, S. L. C.; Welz, B.; Heitmann, U. Microchem. J. 2006, 82, 174–182. (5) Lima, E´. C.; Barbosa, F., Jr.; Krug, F. J. Anal. Chim. Acta 2000, 409, 267– 274. (6) Cacho, J.; Beltra´n, I.; Nerin, C. J. Anal. At. Spectrom. 1989, 4, 661–663. (7) Valde´s-Heviay Temprano, M. C.; Ferna´ndez de la Campa, M. R.; Sanz-Medel, A. J. Anal. At. Spectrom. 1993, 8, 847–852. (8) Ebdon, L.; Goodall, P.; Hill, S. J.; Stockwell, P. B.; Thompson, K. C. J. Anal. At. Spectrom. 1993, 8, 723–729. (9) Guo, X.; Guo, X. Anal. Chim. Acta 1995, 310, 377–385. (10) Guo, X.; Guo, X. J. Anal. At. Spectrom. 1995, 10, 987–991. (11) Sanz-Medel, A.; Valde´s-Hevia y Temprano, M. C.; Bordel Garcı´a, N.; Ferna´ndez de la Campa, M. R. Anal. Chem. 1995, 67, 2216–2223. (12) Ferna´ndez de la Campa, M. R.; Segovia Garcı´a, E.; Valde´s-Heviay Temprano, M. C.; Aizpu´n Ferna´ndez, B.; Marchante Gayo´n, J. M.; Sanz-Medel, A. Spectrochim. Acta B 1995, 50, 377–391. (13) Infante, H. G.; Ferna´ndez Sanchez, M. L.; Sanz-Medel, A. J. Anal. At. Spectrom. 1996, 11, 571–575. (14) Bermejo-Barrera, P.; Moreda-Pin ˜eiro, J.; Moreda-Pin ˜eiro, A.; BermejoBarrera, A. J. Anal. At. Spectrom. 1996, 11, 1081–1086. (15) Hwang, T.; Jiang, H. J. Anal. At. Spectrom. 1997, 12, 579–584. (16) Garrido, M. L.; Muno ˜z-Olivas, R.; Camara, C. J. Anal. At. Spectrom. 1998, 13, 295–300. (17) Goenaga Infante, H.; Ferna´ndez Sanchez, M. L.; Sanz-Medel, A. J. Anal. At. Spectrom. 1998, 13, 899–903. (18) Garrido, M. L.; Muno ˜z-Olivas, R.; Camara, C. J. Anal. At. Spectrom. 1998, 13, 1145–1149. (19) Vargas-Razo, C.; Tyson, J. F. Fresenius J. Anal. Chem. 2000, 366, 182– 190. (20) Feng, Y.-L.; Sturgeon, R. E.; Lam, J. W. Anal. Chem. 2003, 75, 635–640. (21) Lampugnani, L.; Salvetti, C.; Tsalev, D. L. Talanta 2003, 61, 683–698. (22) Sun, H.; Suo, R. Anal. Chim. Acta 2004, 509, 71–76. (23) Ritschdorff, E. T.; Fitzgerald, N.; Mclaughlin, R. L. J.; Brindle, I. D. Spectrochim. Acta B 2005, 60, 139–143. (24) Duan, T.; Song, X.; Jin, D.; Li, H.; Xu, J.; Chen, H. Talanta 2005, 67, 968– 974. (25) Manzoori, J. L.; Abdolmohammad-Zadeh, H.; Amjadi, M. Talanta 2007, 71, 582–587.

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

4249

Cu2+, and Pb2+ but were unable to separate Zn2+. Ion exchange, particularly strong base anion exchange resin, has been used for selective retention of Cd2+ from I- medium, followed by elution with HNO3 or NaNO3.26-32 However, the elution rate was too slow for integration with VGAAS. Despite several reported separation of Cd2+ from other metal ions,33-36 no successful application of such separation to vapor generation of Cd has been achieved. Apart from metal ions, the presence of organic matter has also been found to interfere with metal analysis by some methods. For example, it has been reported that lipid matrix interfered with electrochemical analysis of Cd.37 Also, organic substances in sediment sample digests have been found to interfere with the adsorption of metal ions on ion exchange resins,38,39 resulting in a decreased resin efficiency. It is therefore important that any separation method developed for Cd determination must have the means of reducing or preventing such organic interferences. In this paper, we report on a novel in-line method for the separation of Cd2+ from Mn2+, Fe3+, Co2+, Ni2+, Cu2+, Zn2+, and Pb2+ based on the use of halide loaded strong base anion exchangers and its subsequent determination by VGAAS. The influence of halide (Cl-, Br-, and I-) loading on the efficiency of the resin in removing interferences from Cd2+ determination by VGAAS was carefully investigated and optimized. The possible removal of Zn2+ from the column by washing the exchanger with dilute HCl prior to the elution of Cd2+ was also investigated. A comparison was also made between the use of fiber and the use of resin based anion exchangers. Furthermore, the effectiveness of conventional and microwave assisted acid digestion methods in reducing the effect of residual organic matrix on the separation and retention of Cd2+ is compared. In addition, the application of the method to the determination of Cd in sediment, fish, and water samples is investigated. EXPERIMENTAL SECTION Reagents and Instruments. All chemicals are reagent grade unless otherwise stated. AG 1-×8 strong base anion exchange resin of 400 mesh (Bio-Rad Laboratories, Richmond, CA, U.S.A.), Amberlite IRA 900 strong base anion exchange resin of 20 to 50 mesh (Sigma-Aldrich Chemie GmbH, Buchs, France), Dowex 1-×8 strong base anion exchange resin of 35 to 50 mesh (BDH Chemicals Ltd., Poole, England), and ZB-2 strong base anion (26) (27) (28) (29) (30) (31) (32) (33) (34) (35) (36) (37) (38) (39)

Matusiewicz, H.; Krawczyk, M. Microchem. J. 2006, 83, 17–23. Kallmann, S.; Oberthin, H.; Lin, R. Anal. Chem. 1958, 30, 1846–1848. Marcus, Y.; Eliezer, I. J. Inorg. Nucl. Chem. 1963, 25, 867–874. Korkisch, J.; Feik, F. Anal. Chim. Acta 1965, 32, 110–116. Korkisch, J.; Sorio, A. Anal. Chim. Acta 1975, 76, 393–399. Strelow, F. W. E. Anal. Chim. Acta 1978, 100, 577–588. Hayashibe, Y.; Sayama, Y.; Oguma, K. Fresenius J. Anal. Chem. 1996, 355, 144–149. Li, Q.; Ouyang, R.; Liu, G. Talanta 2004, 64, 906–911. Saeed, M. M.; Ahmed, M. Anal. Chim. Acta 2004, 525, 289–297. Itabashi, H.; Mesuda, Y. J. Flow Injection Anal. 2003, 20, 193–196. Peterson, D. P.; Huff, E. A.; Bhattacharyya, M. H. Anal. Biochem. 1991, 192, 434–440. Meucci, V.; Laschi, S.; Minunni, M.; Pretti, C.; Intorre, L.; Soldani, G.; Mascini, M. Talanta 2009, 77, 1143–1148. Ndung’u, K.; Franks, R. P.; Bruland, K. W.; Flegal, A. R. Anal. Chim. Acta 2003, 481, 127–138. Gue´guen, C.; Belin, C.; Thomas, B. A.; Monna, F.; Favarger, P.-Y.; Dominik, J. Anal. Chim. Acta 1999, 386, 155–159.

4250

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

exchange fiber with a linear density of 1.7-2.3 dtex, and exchange capacity g3.0 mmol g-1 (Guilin Zhenghan Science and Technology Developing Co. Ltd., China) were used for the investigation. Standard solutions (1000 mg L-1) of Cd, Mn, Fe, Co, Ni, Cu, Zn, and Pb were purchased from BDH Chemicals (Australia) Pty. Ltd., (Kilsyth, Victoria, Australia) and adequate dilution was made as required. Milli-Q water was used throughout the study. Br- or I- loaded strong base anion exchange column (Brcolumn or I-column) was prepared by placing 200 mg of exchange resin or fiber into a plastic cartridge of 9 mm ID, resulting in a micro column of about 4 mm in length. Both ends of the column were blocked with an inert macro-porous polymer disk. A 0.3 M solution of KBr or KI was percolated through the column for 30 s with a peristaltic pump at a flow rate of 9.6 mL min-1 (see below), and the excess KBr or KI was washed off by percolating water through the column for 30 s. Cl- column was not prepared prior to use because all the anion exchangers were purchased in Cl- form and the sample solution was in a chloride medium. Once adsorbed on the resin or fiber, Br- and I- can not be displaced by Cl- owing to their stronger retention. NaBH4 solution was prepared by dissolving 1 g of thiourea (TU) in water, followed by addition of 0.2 mL of 1000 mg L-1 Co2+ solution and neutralization with 2 mL of 10% NaOH solution. Then 3.5 g of solid NaBH4 was added into the solution and diluted to 100 mL with Milli-Q water. This solution was prepared daily prior to use. The river sediment reference material AGAL-10, collected from the Hawkesbery River in New South Wales, Australia was provided by Australian Government Analytical Laboratories. The fish liver reference material DOLT-4 was purchased from National Research Council, Canada. A lake water sample was collected from Lake Hyland, Churchill, Victoria, Australia. These materials and samples were employed for method validation. A FD3 freeze drier (Bettervac Pty Ltd. T/A, Victoria, Australia) was employed for the moisture removal from fish samples. A Mars 5 microwave digester (CEM Corporation, Matthews, U.S.A.) was used for sample digestion and comparison with the conventional digestion method. A Perkin-Elmer 3030 atomic absorption spectrometer (PerkinElmer Pty Ltd., Rowville, Victoria, Australia) was used with hollow cathode lamps as the irradiation sources for the measurement of Cd and other metals. The operational parameters were set as given in the instruction manual. For VGAAS detection of Cd, instrument was operated in background correction mode with a deuterium lamp as the continuum irradiation source. An absorption cell was assembled in our laboratory as previously reported.40 Detection of Cd was carried out at room temperature at a wavelength of 228.8 nm. The transient signals were recorded with a Perkin-Elmer 56 recorder. To establish the optimum separation and elution conditions, a multicommutation system with FAAS detection, as shown in Figure 1 as system “a”, was assembled in our laboratory. The system composed of a model Miniplus 3 peristaltic pump (Gilson, Villiers Le Bel, France) and four pinch valves made in our laboratory which were operated at 24 V DC. Four 7951 multifunc(40) Zhang, Y.; Adeloju, S. Talanta 2008, 74, 951–957.

Figure 1. Schematic diagrams of multicommutation systems. (a) For initial separation study and (b) for in-line separation/VGAAS determination. V1 to V5, pinch valves; P, peristaltic pump; MC, microcolumn; FAAS, flame atomic absorption spectrometer; S, sample; E, eluent; R, regeneration agent; a, b, tubular reactors; J1, J2, joining points; FM, flow meter; Ar, argon gas; GLS, gas-liquid separator; AC, absorption cell. Table 1. Operational Program for the Valves in the Multicommutation Systems step

valve 1

valve 2

valve 3

valve 4

valve 5

sampling washing eluting washing regenerating cleaning

on off off off off off

off on off on off on

off off on off off off

off off off off on off

off off on off off off

tional digital LCD programmable time delay relays (Trumeter Company Inc., Deerfield Beach, FL, U.S.A.) were used to run the timing program. The first time delay relay (TDR) was used to control the first valve and initialize the second TDR which in turn controls the second valve and third TDR. The third TDR controls the third valve, while the fourth TDR controls the fourth valve. Tygon pump tubing of 1.6 mm ID (Elkay Products, Inc. Worcester, MA, U.S.A.) was used to deliver the solutions. Flexible pump tubing and PTFE tubing of 1 mm in diameter (Cole-Parmer Instrument Company, Verson Hills, IL, U.S.A.) were used for the connection and liquid delivery. The pump was operated at 48 rpm to achieve a flow rate of 9.6 mL min-1. The four valves were activated in sequence by the TDRs, as outlined in Table 1 (valve 1 to valve 4). The in-line separation-vapor generation for AAS determination of Cd was carried out with a different multicommutation system (shown in Figure 1 as system “b”). The system was assembled in our laboratory by using the same components as in system “a” plus an extra pinch valve (valve 5) for intermittent input of HCl and a U tube as the gas-liquid separator (GLS). The valves in

this system had the same timing program as in system “a”, and valve 5 had the same sequence of operation as for valve 3 (Table 1). A separate pumping channel was used in the system to deliver HCl to acidify the eluate from the column, and a third channel was employed for the delivery of NaBH4 solution and its carrier stream. Tygon pump tubing of 1.6 mm, 0.42 mm, and 0.8 mm in diameter (Elkay Products, Inc.) were used for the sample, HCl, and NaBH4 channels, respectively. The pump was also operated at 48 rpm, achieving flow rates of 9.6, 2.5, and 4.8 mL min-1 for the sample, HCl and NaBH4 channels, respectively. PTFE tubing of 1 mm ID was used for connection. PTFE tubing of 10 cm in length and 1.6 mm in diameter was employed as the tubular reactors a and b. A flexible plastic tube of 40 cm in length and 10 mm in diameter was used for the gas delivery from the GLS to the AC. Sample Pretreatment. Sediment samples were dried at 70 °C in an oven, homogenized, and passed through 63 µm Nylon sieve. Fish samples were homogenized and freeze-dried at -40 °C for 48 h. Water sample was filtered through a 0.45 µm cellulose filter, acidified to 0.2 M HCl, and radiated for 30 min with an Analamp Model 80-1025-01 low pressure mercury lamp (BHK Inc., Claremont, CA, U.S.A.) in a 50 mL beaker. Two digestion methods were investigated as follows: (a) Conventional acid digestion. A 0.2 g of sediment sample was digested with 5 mL aqua regia in a 50 mL Pyrex beaker covered with a watch glass by heating on a hot plate to about 1 mL. Then, after cooling, 5 mL of 10 M HCl was added and again heated to about 2 mL. After allowing the beaker to cool, another 3 mL of 10 M HCl was added and heated until about 2 mL of solution was left. Finally, after cooling the beaker, 3 mL of 10 M HCl was added and heating continued for 5 min. At the end of this period, the digest was allowed to cool and then filtered through a filter paper, and the filtrate was collected into a 50 mL volumetric flask and made up to volume with Milli-Q water. An aliquot of the solution was transferred into a 10 mL volumetric flask, and 1 mL of 2 M HCl was added before diluting to mark with Milli-Q water for Cd determination. It is important to ensure that the final Cd concentration is less than 1 µg L-1. Otherwise further dilution may be required. Also for the fish sample, 0.2 g was weighed into a 50 mL Pyrex beaker, followed by the addition of 5 mL of HNO3 (16 M). After standing for 4 h, the sample was heated on a hot plate to about 2 mL. Then, after cooling, 5 mL of 10 M HCl was added and again heated to about 2 mL. After allowing the beaker to cool, another 3 mL of 10 M HCl was added and heated until about 2 mL of solution was left. Finally, after cooling the beaker, 3 mL of 10 M HCl was added and heating continued for 5 min. At the end of this period, the digest was allowed to cool and then filtered through a filter paper, and the filtrate was collected into a 50 mL volumetric flask and made up to volume with Milli-Q water. A 2 mL aliquot of the solution was transferred into a 10 mL volumetric flask, and 1 mL of 10 M HCl was added before diluting to mark with Milli-Q water for Cd determination. (b) Microwave assisted digestion. A 0.2 g of sediment or fish sample was weighed and digested using 10 mL of 16 M HNO3 for 20 min by ramping the oven temperature from 25 to 160 °C in the first 10 min, and then held at 170 °C for final 10 min. After cooling, the digest was transferred to a 50 mL beaker, Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

4251

and 5 mL of 10 M HCl was added before heating to about 2 mL. Then, after cooling, another 5 mL of 10 M HCl was added and again heated to about 2 mL. After cooling the beaker, a final 5 mL of 10 M HCl was added and heated for 5 min. After this period, the digest was allow to cool, the sediment digest was filtered through a filter paper, and the filtrate was collected into a 50 mL volumetric flask. Then 5 mL of 2 M HCl was added and made up to volume with Milli-Q water. It was not necessary to filter the fish sample digest as it contained little or no residue, but 5 mL of 2 M HCl was added and then diluted to 50 mL with Milli-Q water. The solution was used for Cd determination. It is important to ensure that the final Cd concentration is less than 1 µg L-1. Otherwise further dilution may be required. Operational Procedures of the Multicommutation Systems. A six-step operational program of the multicommutation system “a” was employed, as outlined in Table 1, for investigation of the separation performance of the method. Initially, each valve was normally closed. The first step was started by opening valve 1 to aspirate the sample solution (in 0.2 M HCl) through the micro column where Cd2+ ions were retained while the non-retainable metal ions passed through the column. Valve 1 was then closed and valve 2 was opened to flush out the remaining water-soluble matrix in the channel with 0.1 M HCl. Valve 2 was then closed and valve 3 was opened to aspirate 0.2 M ethylenediamine (En) to elute the retained Cd2+ ions from the column and introduce into the flame atomizer. Then, valve 3 was closed and valve 2 was opened again to flush out residual eluent in the channel with 0.1 M HCl. To remove the residual Cd2+ ions in the column and eliminate the memory effect, 0.4 M of En solution was aspirated through valve 4 in the fifth step followed by aspirating the 0.1 M HCl through valve 2 to clean the channel and the column for the next sample injection. The time intervals of the third to the sixth steps were fixed at 5, 5, 10, and 5 s, respectively, while the sample loading time in step 1 was regulated depending on Cd2+ concentration. In general, longer sampling time was used for lower Cd2+ concentration to improve sensitivity. The wash time in the second step was altered depending on concentrations of interferants. Prolonged washing time was required when high interferant concentrations were present to ensure complete removal. The minimum time used for step 2 in this study was 10s. A similar six-step operational program was employed for system “b”. In this case, while valve 1, valve 2, or valve 4 was opened during the first, second, fourth, fifth, or sixth step, valve 3 and valve 5 were opened to water, whereas in the third step, valve 3 was opened to the eluent and NaBH4 solution while valve 5 was opened to HCl (1.5 M). The eluate from the column was acidified by merging with the HCl input through valve 5 at J1 and then merged with the NaBH4 solution at J2 to allow formation of gaseous species of Cd in the tubular reactor a. The gaseous species were carried by argon gas at a flow at 570 mL min-1 through b to the gas liquid phase separator (GLS) where the gaseous phase was separated from the liquid phase and swept into the absorption cell to generate an absorption signal. By using valves 3 and 5 for the intermittent input, reagent consumption was minimized. 4252

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

RESULTS AND DISCUSSION Retention of Cd on Halide Loaded Resins and Fiber. The desired outcome for optimization of the separation of Cd2+ is to achieve maximum retention of Cd2+, while minimizing the retention of other metal ions. It was found that the presence of Br- or I- in the sample solution was beneficial for the retention of Cd2+, but resulted in lower elution rate. Similarly, when Br- or I- was loaded onto the resin or fiber, but excluded from solution, improved Cd2+ retention was obtained and a high elution rate was also achieved. The presence of >0.02 M SO42- (added as H2SO4 or Na2SO4) had significant impact on the retention of Cd2+ on Cl- or Brloaded exchangers. This is due to the displacement of Cl- or Br- by SO42- from the exchangers. However, the retention of Cd2+ on I- loaded exchangers was not affected by SO42-, but instead enhanced its elution. The FAAS signal showed that the elution of Cd2+ from I- loaded resin had a higher peak profile and less tailing in the presence of SO42- than in its absence. This may be due to partial displacement of I- on the resin by SO42-, resulting in the formation of the complex [CdICl3]2which is easier to elute than [CdI2Cl2]2- formed on the SO42free exchanger. A similar effect was observed in the presence of NO3-. For these reasons, Cl- or Br- loaded exchangers can only be adequately used for the retention of Cd2+ when low concentrations of SO42- and NO3- are present. Although Cd2+ concentrations below the mg L-1 level can be completely retained by any of the exchangers, it was found that the retention rate decreased in the following order: Iexchanger > Br- exchanger > Cl- exchanger. Obviously, the Cl- loaded exchanger was least effective for the retention of Cd2+. Two noted significant benefits of loading I- onto the anion exchangers include improved tolerance to other anions and improved elution performance in the presence of SO42-. Mn2+, Fe3+, Co2+, and Ni2+ were not retained by any of the exchangers and, hence, were completely separated from Cd2+. On the other hand, Cu2+ and Pb2+ were well retained by Iloaded resins and fiber, but not retained by Cl- and Br- loaded exchangers. The retention of Cu2+ on I- loaded resins and fiber was due to its reduction to Cu+ and subsequent complexation/ precipitation by I-.41 Likewise, the retention of Pb2+ is caused by its complexation with I-.41 Consequently, when both metals are present, it is preferable that SO42- and NO3- concentrations are low. As a low SO42- concentration will eliminate its benefit on Cd sensitivity when I- loaded resin is used, it is more desirable to use Cl- or Br- loaded exchangers. Alternatively, SO42- can be removed by precipitation with Ba2+, while NO3can be removed by heating in presence of HCl. However, considering the ubiquitous nature of Cu2+ in sediment samples and the lack of a simple method for its removal, the use of Clloaded resin was chosen for the separation of Cd2+ from Cu2+ and Pb2+ ions. Also, it was found that all the anion exchangers partially retained Zn2+, and the retention efficiency decreased in the following order: Cl- exchanger > Br- exchanger > I- ex(41) Dean, J. A. Analytical Chemistry Handbook; McGraw-Hill: New York, 1995, 2.10–12.11.

Figure 2. Influence of ethylenediamine (En) concentration on elution efficiency of AG 1-×8 anion exchange resin.

Figure 3. Effect of HCl concentration used for eluate acidification and Zn2+ concentration in eluent on Cd response.

changer. Evidently, the I- loaded resins were least effective in retaining Zn2+. Elution of Cd2+. It was found that Cd2+ retained on the halide loaded resins or fiber can be eluted with water, and the elution rate decreased in the following order: Cl- resin > Br- resin . I- resin. Elution from fiber was generally more difficult than from the resins. The use of reagents that are capable of forming positively charged complexes with Cd2+ was also found to be beneficial in accelerating its elution. A typical example of such reagent is ammonia. However, precipitation of Cd(OH)2 occurs in alkaline medium, reducing the recovery and resulting in memory effect. In contrast, the use of ethylenediamine enabled better elution, and much less precipitate was formed. Figure 2 shows that 0.1 M En can completely elute Cd2+ from Cl- or Iloaded AG 1-×8 resins and was therefore used as the eluent in other investigations. Taking into account the different elution rates from the various exchangers, 0.2 M En was used in the VGAAS method. Comparison of Elution Efficiency of Anion Exchangers. The retention and elution performances of a number of strong base anion exchangers were compared. It was found that the rate of elution of Cd2+ from anion exchange fiber was often slower than from anion exchange resins, and this tended to result in a more significant memory effect with the fiber. The slower elution from fiber may be caused by the difference in stoichiometry. For example, because of its flexibility, Cd2+ can be coordinated by three or four immobilized X atoms on the fiber, while the rigid surface of the resin can only offer two X atoms for coordination. Therefore, desorption from fiber is more difficult. A notable observation was that the effect of SO42- varies from one type of exchanger to another. Macro-porous resin (Amberlite IRA 900) and fiber were more susceptible to the effect of SO42- than meso-porous resins (Dowex 1-×8 and AG 1-×8). This may be because immobilized halide atoms on the former are kinetically easier to displace by SO42- than on the latter. The results indicate that Dowex 1-×8 (35-50 mesh) is less prone to the effect of SO42- than AG 1-×8 (400 mesh). This could also be attributed to a kinetic difference. For example, for the same pore size, the rate of ion exchange is slower with larger particle size resin than with smaller size resin.42 Therefore, displacement of X- by SO42- is slower on Dowex 1-×8 resin than on AG 1-×8 resin. Another notable difference between I- loaded Dowex 1-×8 and I- loaded AG 1-×8 resins is that the former can tolerate 2 M SO42-, whereas the latter can only tolerate 0.5 M SO42-. Furthermore, once SO42- has partially displaced the loaded I- on Dowex 1-×8 resin, the

performance of the resin became stable and the elution rate of Cd2+ remains unchanged irrespective of the presence of SO42-. In comparison, the presence of SO42- was always necessary to achieve improved elution rate when AG 1-×8 resin was used, and the constant elution rate can only be obtained when the SO42- concentration is between 0.2 and 0.3 M. Higher SO42concentration reduced signal intensity, possibly because of increased displacement of I-. However, AG 1-×8 resin had much better elution performance than all other exchangers owing to its kinetic advantage, and the sensitivity obtained with this resin for Cd by VGAAS was 150% higher than with other exchangers. It is important to note that the inclusion of a wash-out step prior to the elution of Cd2+was intended to ensure complete removal of residual metallic interferants. The use of 0.1 M HCl enabled complete removal of residual metal ions without any loss of adsorbed Cd2+. Attempts to lower HCl concentrations resulted in loss of adsorbed Cd2+. Complete removal of Zn2+ with 0.1 M HCl was more difficult but was achieved by extending the wash time. Also the inclusion of the regeneration step in the operation of the separation-vapor generation procedures eliminated the memory effect, and all the anion exchangers tested in this study gave satisfactory results for VGAAS determination of Cd, except for the higher sensitivity obtained using AG 1-×8 resin. Effect of Acid, En, and Zn2+ Concentrations. The resulting eluate from the use of En solution to elute cadmium from the columns was alkaline and, therefore, required acidification prior to merging the NaBH4 solution for vapor generation. For this reason, HCl was introduced at J1 in system “b” (Figure 1). Acidification of the alkaline eluate by 1.5 M HCl gave the best sensitivity, as shown in Figure 3. Signal depression by Zn2+ in VGAAS determination of Cd has been reported in a number of studies.7,15,25 However, it was found in this study that changes in HCl concentration used for acidification had different influence on the observed effects of Zn2+. Figure 3 shows that the optimum HCl concentration required for acidification of the eluate in the absence of Zn2+ increased from 1.5 to 1.7 M when 2 mg L-1 Zn2+ was present. More significantly, the presence of Zn2+ suppressed the cadmium response when 1.5 M HCl was used but enhanced the response when 1.7 M HCl was used. This observation highlights the need to consider the influence of Zn2+ in combination with HCl concentration. Figure 4 shows that, when 1.7 M of HCl was used for acidification, the sensitivity of Cd response increased with the addition of Zn2+ even when absent in the En solution. Also, the sensitivity increased with increasing Zn2+ concentration in the En solution, reaching an optimum sensitivity which remained constant when g2 mg L-1 Zn2+ was

(42) Rawat, J. P.; Chitra, Bull. Chem. Soc. Jpn. 1988, 61, 2268–2270.

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

4253

Figure 4. Effect of Zn2+ concentrations in sample and En solutions on Cd response.b: Zn2+ in sample solution was 0, O: Zn2+ in sample solution was 1 mgL-1.

present in En solution. Therefore, the addition of 2 mg L-1 of Zn2+ into the En solution acidified with 1.7 M HCl is beneficial for achieving optimum and constant sensitivity. Under such conditions, the tolerance for Zn2+ in the sample solution increased from 1 mg L-1 to 100 mg L-1. The above observations provide two distinct choices for Cd determination by the proposed VGAAS method: (a) at low Zn2+ concentrations (