Cloud Point Preconcentration and Flame Atomic Absorption Spectrometry Application to the Determination of Cadmium Journal of Analytical Atomic Spectrometry CARMELO GARCIA PINTO JOSE LUIS PEREZ PAVON AND BERNARDO MORENO CORDERO EMILIO ROMERO BEATO AND SOLEDAD GARCIA SANCHEZ Departamento de Quimica Analitica Nutrici6n y Bromatologia Facultad de Quimica Universidad de Salamanca 37008 Salamanca Spain Servicio General de Andisis Quimico Aplicado Universidad de Salamanca 3 7008 Salamanca Spain Cloud point methodology has been successfully used for the preconcentration of trace amounts of cadmium as a prior step to its determination by flame atomic absorption spectrometry. A procedure based on the formation of a complex with 142- pyridylazo)-2-naphthol (PAN) is used for the preconcentration of cadmium in a surfactant-rich phase of Triton X-114.The chemical variables affecting the preconcentration step and the viscosity of the solution affecting the detection process have been optimized. Under the optimum conditions a precision of 3.0% was achieved. The preconcentration of only 15 ml of sample with 0.05% Triton X-114 permits the detection of <0.4 ppb of cadmium with a concentration factor of 120. Keywords Cloud point preconcentration; flame atomic absorption spectrometry; cadmium; tap water; sea water The use of micellar solutions in different areas of analytical chemistry has attracted much attention in recent In particular its use in high-performance liquid chromatography solvent extraction gel filtration ultracentrifugation and capil- lary electrokinetic chromatography has opened up new pos- sibilities for the separation of metal species of biological and environmental interest.6 Aqueous solutions of almost all non-ionic surfactants become turbid when heated to a temperature known as the cloud point.Above this temperature the isotropic micellar solution separates into two transparent liquid phases a surfac- tant-rich phase of very small volume composed mostly of the surfactant plus a small amount of water and an aqueous phase in equilibrium with the former which contains a surfac- tant concentration close to its critical micellar concentration. The exact mechanism through which phase separation occurs remains to be fully This unique surfactant-phase solution-phase separation phenomenon permits the design of simple schemes for extrac- tion preconcentration and purification and has recently been reviewed by Hinze and Pramauro." Any hydrophobic species originally present in water is able to interact with and bind to micelles and become concentrated in a small volume of the surfactant-rich phase.The cloud point methodology has been used for the precon- centration of organic compounds of different types. Recently the compatibility of this method with HPLC both with optical detection (~ltraviolet'~-'~ and fluore~cence'~,~~) and electrochemical d e t e c t i ~ n ' ~ . ~ ~ . ' ~ has been demonstrated. The phase separation phenomenon has also been used for the extraction and preconcentration of metal cations after the formation of sparingly water-soluble complexes." Watanabe * To whom correspondence should be addressed.and Tanaka16 used this method for the first time to preconcen- trate zinc(Ir) using 1-(2-pyridylazo)-2-napththol (PAN) as the hydrophobic ligand and the surfactant PONPE 7.5 as the extracting system. It was later applied in the extraction of metallic chelates for the spectroph~tometric'~ and flow injec- tion" analysis of trace metals in a variety of different samples (tap water sea water soils et~.).''-~l In the present work we report on the results obtained in a study of the cloud point preconcentration of cadmium after the formation of a complex with PAN and later analysis by flame atomic absorption spectrometry using Triton X-114 as surfactant. The proposed method is applied to the determination of cadmium in tap and sea water samples.EXPERIMENTAL Apparatus A Hitachi Model 28000 atomic-absorption spectrometer equipped with Zeeman background correction and a cadmium hollow-cathode lamp as the radiation source was used. A Hitachi 180-0410 micro-sampling device was employed for the injection of 50 or 1OOpl of the sample. The instrumental parameters were adjusted according to the manufacturer's recommendations. A Kokusan H-103 N centrifuge was used to accelerate the phase separation process. Dynamic viscosity was measured with an Ostwald viscometer. Reagents The non-ionic surfactant Triton X-114 was obtained from Fluka and was used without further purification. The stock standard cadmium solution (1000 ppm) was prepared from pure cadmium nitrate (Merck).Working standard solutions were obtained by appropriate dilution of the stock standard solution. mol I-') of PAN (Merck Darmstadt Germany) in Triton X-114 were prepared from the commer- cially available product. Stock buffer solution 0.05 mol l-l was prepared by dissolving appropriate amounts of Na2B407 10H20 (Panreac Barcelona Spain) in water. All the other reagents were of analytical-reagent grade. All solutions were prepared in ultra-high-quality water obtained from an Elgastat UHQ water purification system. The materials and vessels used for trace analysis were kept in 10% nitric acid for at least 48 h and subsequently washed four times with ultra-high-quality water before use. Solutions (5.7 x Journal of AnaEytical Atomic Spectrometry January 1996 Vol. 11 (37-41) 37Procedures Cloud point determination The cloud point for Triton X-114 in the absence and in the presence of 0.05 mol 1-1 borax was determined by observing the appearance of the two phases upon heating different aqueous solutions of surfactant in a thermostated bath.Ratio of phases The volumes of the respective phases were measured in cali- brated tubes (0.4 cm i.d.) for the different amounts of surfactant under the same experimental conditions as those used for phase separation (heating at 40 "C and centrifugation at 3500 rpm). 34 32 - - 28- 18' 1 I 0 5 10 15 [Triton X-l14](%) Cloud point preconcentration Aliquots of 15.0 ml of the cold solution containing the analyte Triton X-114 and PAN buffered at a suitable pH were kept for 5 min in the thermostatted bath at 40 "C.Separation of the two phases was accomplished by centrifugation for 5 min at 3500 rpm. On cooling in an ice-bath the surfactant-rich phase became viscous. The aqueous phase could then be separated by inverting the tubes. Later a given volume of a solution of methanol containing 0.1 mol 1-1 of HNO was added to the surfactant-rich phase. The samples were introduced into the flame by conventional aspiration or by a device designed for microsamples that permits the introduction also by aspiration of a volume of 50 or loop1 of previously diluted surfactant- rich phase. Extraction of cadmium from spiked water samples Tap and sea water samples were filtered using a 0.45 pm pore size membrane filter to remove suspended particulate matter and were then stored at 4 "C in the dark.Aliquots of the water samples studied were 'cloud point preconcentrated' using 0.10% Triton X-114. After phase separation 1.0ml of a methanol solution containing 0.1 mol 1-1 of HNO was added to the surfactant-rich phase. The sample was introduced into the flame by conventional aspiration. RESULTS AND DISCUSSION Phase Diagram of Triton X-114 At a concentration above the critical micellar concentration Triton X-114 in aqueous solution displays a consolution curve above which the micellar solution separates into two isotropic phases. The temperature at which phase separation occurs depends on the concentration of surfactant and the presence and concentration of both organic and inorganic additives. Fig. 1 shows the phase diagram (temperature-concentration) for Triton X-114 in the absence (1) and in the presence (2) of 0.05 mol 1- borate buffer.Across the concentration range studied the presence of the salt decreases the cloud point temperature of Triton X-114. This decrease is more marked for surfactant concentrations below 1 .O% whereas at higher concentrations the difference between the cloud point temperatures is approximately 1 "C. However in both cases the phase diagram shows identical shape. The cloud point temperature remains almost constant (23-25 "C) within the 0.5-5.0% concentration range. Experiments carried out with PAN concentrations similar to those to be used in the preconcentration step did not lead to significant changes in the phase diagram of Triton X-114. Fig. 1 Phase diagram of the surfactant Triton X-114 in aqueous solution in the absence (l) and in the presence (2) of Na,B,O - 10H,O 0.05 mol 1- '.L Single isotropic phase region and 2L two isotropic phase regions Effect of pH The separation of metal ions by the cloud point method involves prior formation of a complex with sufficient hydro- phobicity to be extracted into the small volume of surfactant- rich phase (200 pl) thus obtaining the desired preconcen- tration. Extraction yield depends on the pH at which complex formation is carried out. Fig. 2 shows the effect of pH on extraction yield. It may be seen that for pH values above 8 yield is almost constant and close to 100%. Effect of the PAN Cadmium Molar Ratio The effect of the molar ratio between the complexing agent and the cation was studied for values ranging between 1 and 15 and a cadmium concentration of 3.0 x lop6 mol 1-l.The results obtained on preconcentra ting 15 ml of a solution containing the analyte with a Triton X-114 concentration of 0.25% then adding 1.0 ml of 0.1 mol 1-' HNO after phase separation show that at least a 5-fold excess of PAN over the cadmium concentration was required to obtain maximum and constant recovery. Preconcentration of 15 ml of the solution in the absence of cadmium did not give rise in any case to a cadmium signal. 2 4 6 8 10 12 PH Fig.2 Influence of pH on the extraction recovery (R%) of the Cd-PAN complex. Preconcentration step 3.0 x lop6 mol I-' Cd 3.4 x lo-' moll-' PAN and 0.25% Triton X-114. Other experimental conditions described in text 38 Journal of Analytical Atomic Spectrometry January 1996 Vol.11Effect of Viscosity on the Analytical Signal The small volume (80-400 pl) of surfactant-rich phase obtained after cloud point preconcentration contains a high concen- tration of Triton X-114 (-30%). The solution is therefore highly viscous. Accordingly after phase separation it is neces- sary to decrease the viscosity of the sample in order to facilitate introduction of the sample into the atomizer. Fig. 3(a) shows the variation in dynamic viscosity and in the analytical signal as a function of the Triton X-114 concen- tration initially placed in the solution. In this instance after phase separation 1.5 ml of 0.1 mol 1-l HNO were added to the surfactant-rich volume (between 200 and 300 pl depending on the concentration of surfactant) and the sample was introduced into the flame by conventional aspiration.An increase in surfactant concentration of between 0.10 and 0.50% elicits a strong increase in viscosity reflected in a 70% loss of the analytical signal This decrease can be attributed almost exclusively to the increase in viscosity since the dilution factor between both surfactant concentrations is negligible. However when dilution is carried out by adding 1.5ml of a solution of methanol containing 0.1 mol I-' HNO the 15 10 5 - . 8 0 - P - 1 I I I I I ( b ) l o t \ 1 4 2 - - I I I I I [Triton X-l14](%) 0.25 0.20 0.15 0.10 0.05 0.00 -c 0.20 0.15 0.10 0.05 0.00 Fig. 3 Variation in analytical signal of the Cd-PAN complex (1) and of the viscosity of the sample (2) as a function of the Triton X-114 concentration.Dilution of the surfactant-rich phase with water (a) or methanol (b) containing 0.1 mol 1-' HNO,. Preconcentration step 2.7 x lop7 mol 1-' Cd 2.6 x mol 1-' PAN and 0.10% Triton X-114. Other experimental conditions described in text 1 .oo 0.80 0.60 c 0.40 0.20 0.00 0 300 600 900 1200 1500 Added methanol/pl Fig. 4 Variation in normalized analytical signal (h) corresponding to the Cd-PAN complex as a function of the volume of methanol added. Theoretical plots as a function of the viscosity of the sample (1) and of the dilution factor (2) and experimental curve obtained (3). Preconcentration step 2.7 x lop7 mol 1-' Cd 2.6 x lop6 moll-' PAN and 0.10% Triton X-114. Other experimental conditions described in text increase in viscosity is much less pronounced and the decrease in the analytical signal is only 8% [Fig.3(b)]. Furthermore the effects of the organic solvent on the flame produce a signal enhancement factor that allows an additional increase in sensitivity. Table 1 shows the phase ratios and the signal enhancement factors obtained under different experimental conditions. It may be seen that the presence of methanol produces an increase of approximately 2 in the analytical signal in all cases. The influence of the combined effects of viscosity and dilution are shown in Fig. 4. The figure shows the normalized signals corresponding to the preconcentration of 15 ml of sample with a surfactant concentration of 0.10% as a function of the volume of methanol added to the surfactant-rich phase. Curves 1 and 2 represent the theoretical variations in the signal due to the decrease in viscosity and to the increase in the dilution factor respectively. The experimentally obtained curve (curve 3) depends on the combined effect of these two variables.As can be seen for added volumes of methanol of < 100 pl an important increase in the analytical signal occurs because the effect of decreased viscosity is very strong and clearly predominates over the dilution. However for higher added volumes the decrease in viscosity of the sample is lower and it is essentially the effect of dilution that predominates. Calibration Precision and Detection Limits Calibration curves were constructed by preconcentrating 15 ml of sample with Triton X-114 concentrations of 0.10 and 0.05%.Samples were introduced into the flame by conventional aspiration or using a device for microsamples. When conven- tional aspiration was used the surfactant-rich phase was Table 1 Ratio of phases (solution phase/surfactant-rich phase) and enhancement factor [Triton X-114) ("/.I Dilution of surfactant-rich phase 0.10 0.1 mol I-' HNO in water (1.0 ml) 0.10 0.1 mol 1-1 HNO in methanol (1.0 ml) 0.10 0.1 mol 1-' HN03 in methanol (200 pl) 0.05 0.1 mol 1-' HN03 in methanol (100 p1) 0.10 0.1 mol 1-' HNO in methanol (200 pl) 0.05 0.1 mol I-' HNO in methanol (100 pl) Sample introduction Aspiration Aspiration 50 pl 60 p1 100 pl 100 pl Ratio of phases 11 11 27 60 27 60 Enhancement factor * 10.8 22.7 50 123 48 120 * Relationship of absorbance peak height of preconcentrated samples ( 15 ml) to that obtained without preconcentration.Journal of Analytical Atomic Spectrometry January 1996 Vol. 1 1 39Table 2 Analytical characteristics of the method* Conditions Without preconcentration5 Preconcentration (0.10% Triton X-114)" Preconcentration (0.10% Triton X-114)" Without preconcentration§ Preconcentration (0.10% Triton X-114)** Preconcentration (0.05% Triton X-114)tt Without preconcentration5 Preconcentration (0.10% Triton X-114)** Preconcentration (0.050/ Triton X-114)?? Sample introduction Aspiration Aspiration Aspiration 50 p1 50 p1 50 pl 100 pl loo 1.11 loo p1 Range (PPb) 50- 1000 5.0-100 2.0- 100 100-1000 5.0-100 1 .O-50 50-1000 2.0-100 0.5-50 Slope (9.8k0.1) x lo-' (1.04f0.04)~ (2.09 0.04) x (3.03f0.06)~ lo-' (1.48 k0.06) x (3.1 kO.1) x (4.19k0.05) x lo-' (2.00-tO.04) x (5.06-0.09) x lop3 Intercept (5.0f0.2) x lop3 (5.5f0.2) x lop3 (1.2k0.4) x lop3 ( 2 f 2 ) x 10-3 (4+2) x 10-3 (4rt_4) x 10-3 (2f 1) x 10-3 (2f2) x 10-3 (2.4k0.6) x lop3 r2 0.998 0.998 0.999 0.999 0.997 0.997 0.999 0.998 0.999 s (%)t 2.7 (300) 2.8 (10) 3.1 (5.0) 3.0 (300) 2.9 (10) 3.4 (5.0) 2.8 (100) 3.5 (5.0) 3.2 (5.0) LODS ( PPb) 25 2.3 1.1 80 1.6 0.6 48 1 .o 0.4 * Samples 15 ml.Duplicate injection. t Values in parentheses are the cadmium concentrations for which s was obtained. LOD limit of detection (calculated as twice the noise). Standard solutions of cadmium in 0.1 mol 1-' HNO medium. Dilution of the surfactant-rich phase with 1.0 ml of a solution of 0.1 mol 1-' HN03 in methanol.7 Dilution of the surfactant-rich phase with 1.0 ml of 0.1 mol 1-' HNO,. ** Dilution of the surfactant-rich phase with 200 pl of 0.1 mol 1-' HNO in methanol. tt Dilution of the surfactant-rich phase with 100 pl of 0.1 mol 1-' HNO in methanol. diluted with 1.0 ml of a solution of methanol containing 0.1 mol 1-l HN03. When the device for microsamples was used the surfactant-rich phase was diluted with 100 or 200 pl of the 0.1 moll-' HN03 in methanol solution and 50 pl or 100 pl of the diluted solution were introduced into the flame. In all instances linear relationships were obtained between peak height in units of absorbance and the concentration of cad- mium. The least squares fitting parameters are shown in Table 2 for all conditions studied together with the relative standard deviation for 10 samples to which the complete procedure was applied and the calculated detection limits (twice the noise).The same table shows the calibrations obtained with standard cadmium solutions not subjected to the preconcentration step. Preconcentration of only 15 ml of sample with a Triton X-114 concentration of 0.05% permits a detection limit below 0.4 ppb with a 120 fold increase in the analytical signal. Interferences Two types of interferences affecting the preconcentration pro- cess can be distinguished in the proposed methodology; the cations reacting with PAN and species that form complexes with cadmium including anions and humic acids (potential interferents in water samples). Interference by the foregoing cations that form complexes with PAN can readily be avoided by increasing the concen- tration of PAN.However studies were conducted with the cations alu- minium calcium magnesium lead@) iron(m) nickel and zinc for interferent:cadmium ratios of 1 10 and 100 ([Cd]=18 ppb). The preconcentration step was performed with 0.25% Triton X-114 and a PAN concentration of 4.8 x lop5 mol 1-l. No interferences (error < 3%) were detected in the presence of these cations at the levels described. The results obtained in the study of the interferences pro- duced by the presence of anions and complexing agents for the interferent cadmium ratios previously mentioned shows that sulfate chloride cyanide and ammonia did not cause significant interferences (error < 3 Yo). However the presence of EDTA did lead to a negative interference up to 80% owing to the formation of a water-soluble cadmium complex that competes with the formation of the insoluble Cd-PAN com- plex thereby preventing the preconcentration of cadmium in the surfactant-rich phase.In order to study the interference produced by humic acid in the cloud point preconcentration of cadmium an optically standardized solution (- 10 ppm organic carbon content) was prepared following the procedure described by Johnson et ~ 1 . ~ Different dilutions of this solution spiked with cadmium (10 ppb) were used and the signal obtained after the preconcen- tration step was compared with that corresponding to a solution at the same concentration of cadmium in ultra- pure water. The recoveries obtained for the two different PAN concen- trations are shown in Table 3.For the lowest concentration of reagent the presence of humic acids leads to a decrease in the analytical signal of up to 42%. However when the PAN concentration is higher for dilutions of the humic acid solution between 1 1 and 1 3 no appreciable effect on the cadmium signal can be detected. A decrease of 15% in the signal occurs for a dilution of 1 2 and for an undiluted solution of humic acid the loss in signal is 25%. Although such high concen- trations of humic acids are not usually found in real water samples a sample pretreatment with peroxodisulfate similar to that described by R o y ~ e t ~ ~ was performed. The recoveries obtained in both cases for dilutions of 1 2 and 1 1 were 98 and 97% respectively (see Table 3).Table3 determination of cadmium Effect of humic acid solutions on the preconcentration and Recovery (YO) Dilution [PAN]=2.3 x mol 1-' [PAN]=4.8 x lop5 mol 1-' 1 100 100 1 50 94 99 1:20 98 1 10 85 101 1:5 74 104 1:3 69 102 1:2 60 (98) 85 (97) 1 l 58 (98) 75 (97) * Values in parentheses correspond to recoveries obtained after oxidation with peroxodisulfate of the organic material. 40 Journal of Analytical Atomic Spectrometry January 1996 Vol. 1 1Table 4 Slopes and intercept of the calibration and of the standard additions in tap and sea water Slope Intercept r2 Calibration (2.09k0.04) x (1.2f0.4) x 0.999 Tap water (2.08+0.03)x (4.0f0.6) x lop3 0.999 Sea water (2.05 k 0.04) x (5.7 f 0.7) x 0.998 Determination of Cadmium in Tap and Sea Water In order to test the reliability of the proposed methodology it was applied to samples of tap water (from the drinking water system of Salamanca Spain) and sea water (from the Cantabrico Sea Santander Spain).For this purpose 15 ml of each of the samples were precon- centrated with 0.10% Triton X-114 and a PAN concentration of 2.0 x The samples were spiked at concentrations ranging between 1.0 and 25 ppb obtaining a straight line calibration for the standard addition in the tap water the sea water and ultra- pure water matrices. Table 4 shows the parameters obtained under the different experimental conditions. The slope of the calibrations in ultra-pure water coincides with those of the straight lines obtained by standard addition for both samples; this shows that no matrix effect exists.The highest values of the intercept corresponding to the standard additions indicate that the samples contained cadmium. However it is not possible to quantify these levels properly since the signal obtained for the unspiked samples is almost the limit of detection under these conditions. These results show that the cloud point preconcentration method can be applied to the determination of cadmium in tap and sea water. mol I-' following the proposed method. CONCLUSION This paper shows that the cloud point method can be used to preconcentrate metal cations before their detection by flame atomic absorption spectrometry. The technique offers a simple alternative to other separation and/or preconcentration tech- niques with good yields as regards extraction and a good standard deviation.This work was supported by DGICYT (Project PB94-1393) and the Consejeria de Cultura y Turismo de la Junta de Castilla y Leon (Project SA68/93). REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Pelizzetti E. and Pramauro E. Anal. Chirn. Acta 1985 169 1. Hinze W. L. Sing H. Baba Y. and Harvey N. G. Trends Anal. Chem. 1984 3 193. Moreno Cordero B. PCrez Pavon J. L. and Hernandez Mendez J. Quim. Anal. 1989 8 231. McIntire G. L. Crit. Rev. Anal. Chem. 1990 4 257. Armstrong D. W. Sep. Purg Methods 1985 14 213. Hinze W. L. in Ordered Media in Chemical Separations ACS Symposium Series 342 Hinze W. L. and Armstrong D. W. eds. American Chemical Society Washington DC USA 1987. Degiorgio V. Piazza R. Corti M. and Minero C. J. Chem. Phys. 1985 82 1025. Blankschtein D. Thurston G. M. and Bebedek G. B. J. Chem. Phys. 1986 85 7268. Lindman B. and Wennerstrom H. J . Phys. Chem. 1991,95,6053. Hinze W. L. and Pramauro E. Crit. Rev. Anal. Chem. 1993 24 133. Saitoh T. and Hinze W. L. Anal. Chem. 1991 63 2520. Garcia Pinto C. Perez Pavon J. L. and Moreno Cordero B. Anal. Chem. 1992 64 2334. Moreno Cordero B. PCrez Pavon J. L. Garcia Pinto C. and Fernandez Laespada M. E. Talanta 1993 40 1703. Garcia Pinto C. PCrez Pavon J. L. and Moreno Cordero B. Anal. Chem. 1994 66 874. Garcia Pinto C. PCrez Pavon J. L. and Moreno Cordero B. Anal. Chem. in the press. Watanabe H. and Tanaka H. Talanta 1978 25 585. Kawamorita S. Watanabe H. and Haraguchi K. Anal. Sci. 1985 1 41. Fernandez Laespada M. E. Perez Pavon J. L. and Moreno Cordero B. Analyst 1993 118 209. Saitoh T. Kimura Y. Watanabe H. and Haraguchi K. Anal. Sci. 1989 5 577. Hoshino H. Saitoh T. Taketomi H. and Yotsuganagi T. Anal. Chim. Acta 1983 147 339. Buhai L. and Rigan M. Talanta 1990 37 885. Johnson W. E. Fendinger N. J. and Plimmer J. R. Anal. Chem. 1991,63 1510. Royset O. Anal. Chim. Acta 1986 185 75. Paper 5104083 C Received June 23,1995 Accepted September 24,1995 Journal of Analytical Atomic Spectrometry January 1996 Vol. 1 1 41