Determination of Lead and Cadmium in Environmental Samples Optimized by Simplex Optimized Atomic Absorption Methods I Journal of 1 Analytical Atomic Spectrometry CONSTANTINE D. STALIKAS AND GEORGE A. PILIDIS European Environmental Research Institute Dodonis 42 45221 Ioannina Greece MILTIADES I. KARAYANNIS Laboratory of Analytical Chemistry University of loannina 451 10 loannina Greece Simplex optimized methods for the determination of cadmium and lead in environmental samples by use of electrothermal atomic absorption spectroscopy (ETAAS) are presented. The optimum experimental conditions were attained by applying the composite modified simplex optimization method after considering the following variables ramp to ashing temperature ashing temperature atomization ramping time atomization temperature and modifier concentration.The way of introducing the matrix modifier in the graphite furnace is one of the most important among the numerous factors that must be considered for the development of an ETAAS method. The matrix modifier was injected into a graphite tube provided with a L'vov platform prior to the injection of the sample leaving the modifier wet. The analytical investigations showed that the optimized systems are ideally suited to the analysis of cadmium and lead in samples of environmental origin offering good accuracy and precision. Keywords Lead; cadmium; environmental samples; simplex optimization; atomic absorption spectrometry Cadmium is a toxic element present at low concentrations in nature.' Cadmium occurrence stems from anthropogenic sources such as mining operations waste incineration and combustion of coal and oil while it occurs naturally in the environment as a result of volcanic emissions.2 The contami- nation from Cd has increased rapidly in recent years; it is commonly found in aquatic and terrestrial environments and is characterized by a long environmental per~istence.~?~ The monitoring of Cd in marine species (seafood) can serve as an indicator of variations in marine pollution.' Lead in the atmosphere comes mainly from the combustion of gasoline that contains tetraethyllead as an anti-knock agent.The accurate determination of lead in foodstuffs is important since the prolonged intake of even low concentrations of lead can cause serious toxic effect^.^.^ The low concentration levels of cadmium and lead during their assay in most of the samples unavoidably require a preconcentration step or a preliminary separation of cadmium in the bulk matrix.Most analytical procedures require a solution of the analyte and in principle solid samples must be dissolved prior to quantification. A number of methods of varying complexity have been used to determine trace amounts of Cd and Pb spectrophotometry atomic absorption spectrometry (AAS) inductively coupled plasma atomic emission spectrometry or mass spectrometry and electroanalytical methods. Among these the fastest and most sensitive that have conquered the field of analytical chemistry are electrothermal atomic absorption spectrometry (ETAAS) and inductively coupled plasma atomic emission spectrometry or mass spectrometry.The two techniques differ from each other in relative simplicity cost of analysis and interferences en~ountered.~.~ ETAAS combines high sensitivity simplicity low detection limits and a low sample volume requirement. All of these features have been responsible for its widespread use for a diverse range of samples. Since the performance of ETAAS is strongly affected by several instrumental parameters it is advisable to carry out optimization prior to the adoption of any procedure. The most convenient method of sample introduction into the graphite furnace is the injection of liquid samples. Thus solid samples require mineralization and destruction of the organic matrix before the implementation of any analytical technique.Volatilization losses can occur during wet oxidation if the temperature is allowed to exceed 250°C.10 The open- beaker wet ashing technique appears to be gaining popularity since it does not require complex and expensive instrumen- tation. Microwave digestion has received considerable atten- tion as an alternative to conventional wet Bomb digestion has been found to be preferable for the multi-element analysis of samples because problems of losses of volatile analytes have been observed using both microwave or hot plate dige~tion.'~ The direct introduction of solid samples or slurries is increasing in popularity because it combines the simultaneous destruction of the sample matrix and the atomiz- ation or excitation of the analyte but suffers from poor precision although the accuracy is good e n ~ u g h .' ~ - ' ~ Chemical modifiers of different types have been used but none of them seems to be suitable for general application. The combination of chemical modifiers and the L'vov platform in the interior of graphite tubes has largely eliminated inter- ferences on Cd and Pb signal in various matrices showing improved precision and ac~uracy.'~ In this paper the composite modified simplex (CMS) method is employed for the determination of Cd and Pb in seafood. ,Unlike the classical optimization procedures the CMS takes into account possible interactions among the parameters involved while minimizing the number of the required experi- ments accessing the optimum parameter value^.'^.'^ To the best of our knowledge no attempt has been made to date to optimize by use of the CMS method the analysis of these two important naturally occurring metals by ETAAS.However the method has been applied for the optimization of several analytical systems.20-2z EXPERIMENTAL Apparatus An Atomic Absorption Spectrometer Varian (Mulgrave Victoria Australia) SpectrAA-300 was used for this work linked with the Varian GTA-96 graphite furnace atomizer and Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 (595-599) 595the automatic sample dispenser. The results were recorded with an Epson (Nagano Japan) LX-400 printer. The hollow cathode lamps (SpectrAA) in conjunction with the deuterium background corrector were used for the lead and cadmium determinations. A lamp current of 4 mA at 228.8 nm was employed for the analysis of cadmium while the 283.3 nm lead resonance line was selected and used with a lamp current of 5 mA for the analysis of lead.The slit width in both cases was 0.5 nm. The atomization was performed with a pyrolytic graphite tube with a L'vov platform. The peak height was used for quantitative analysis. Argon was used as the inert gas for purging the graphite tubes. In obtaining the data a 15 pl volume of sample solution plus a 5 pl volume of modifier at appropriate concentration were used throughout. Reagents Used and Sample Treatment All solutions were prepared with particle-free deionized doubly distilled water. Merck standard solutions for AA were used to prepare the working solutions in 5% HNO,. Ammonium dihydrogenphosphate was selected as the appro- priate matrix modifier based on previous investigation^.^^ Analytical reagent grade crystals of the modifier were dissolved in water to give several solutions which were subsequently used as the chemical modifier.A mussel tissue RM 278 with certified Cd and Pb contents of 0.34k0.02 pg g-' and 1.91 f 0.04 pg g- ' respectively was used as the reference mate- rial (RM) for the validation of the methods. Marine Sediment (NRC-PACS l) Estuarine Sediment (CEC-CRM 277) and Tea (NRC-CRM CS5-05) were selected for the implementation of the developed methods. The certified reference materials were treated as follows OSOg dry mass was slowly digested by heating in a Teflon digestion vessel with 6ml of concentrated HNO,. The sedi- ments were then further subjected to HF-HC104 attack until the entire sample had been dissolved.The digestates were quantitatively transferred and diluted to a final volume of 20ml with 10% HNO,. Appropriate dilutions with the blank (10% HNO,) were made by the programmable sample dis- penser during the analysis. Optimization Procedure The simplex process was carried out using software which can run on any IBM compatible computer. The optimization was carried out on a sample of mussel tissue in order to establish the 'real' experimental conditions and to eliminate problems arising during the analytical procedure. This reference material has a matrix similar to most of the edible environmental samples and is thus suitable for the validation of the method. The effect of five experimental parameters on the analyses of the metals was investigated ramp to ashing temperature ashing tempera- ture atomization ramping time atomization temperature and modifier concentration were optimized by the CMS optimiz- ation method.Some preliminary univariate experiments were carried out prior to the CMS optimization in order to establish the boundaries of the values of each parameter. The lower and the higher values as well as the precision required for each variable were set accordingly as is shown in Table 1. The initial set of conditions was given and the program generated the rest of the experiments. For the five variable system six sets of experiments constitute the initial simplex. After the performance of each experiment the actual values of the variables and the peak height of the signal were entered.The first six experiments represent the first cycle of the simplex. The software continued the process of accessing the optimum by performing reflections (R) expansions (E) con- tractions (C) and lagrange interpolative fits (L). Each set of experiments was performed in three replicates. Between each experiment a blank corrective experiment was run to ensure stable and repeatable results. RESULTS AND DISCUSSION The optimum values of the parameters were deemed to have been reached after almost 30 experiments for Cd and Pb. Tables 2 and 3 summarize the optimization experiments carried out and the corresponding responses. Because of the volatility of many Cd and Pb compounds lower ashing temperatures are needed in comparison with those applied for the determination of most of the other metals. Higher temperatures result in losses of Cd and Pb during the charring step while lower temperatures retain matrix interferences Under the optimum ashing temperatures the ashing ramping times when restricted to low values do not appear to affect significantly the response on the analysis of Cd and Pb.High ashing ramping times in connection with high ashing tempera- tures decrease the absorbances due to the considerable loss of the analytes. The optimum atomization temperatures allow complete atomization of the analytes in the RM. By stopping the gas flow at this point the sensitivity increases while the lifetime of the graphite tube is not shortened. The atomization ramping time is a very critical parameter affecting the sensitivity of the system.It is advisable to heat the furnace rapidly to a preselected optimum temperature in order to achieve a sufficient temperature difference between tube and platform. Longer atomization ramping times during the atomization step led to a decrease in the peak height absorbance due to losses of the volatile analytes and to vapour phase interferences (physical interferences) when analysing samples with complex matrices. In the case of Pb the shorter ramping time was used whereas up to 2 s this variable does not seem to affect the analysis of Cd strongly. Optimum atomization temperatures have been established to ensure complete atomization and prolonged lifetime of the graphite furnace. The modifier concentration is critical when determining Cd and Pb in environmental samples although in these experi- ments it is not very pronounced.In both cases the concen- tration of matrix modifier was kept at 3.5 mg ml-I since the ammonium dihydrogenphosphate is a very cheap reagent thus keeping the cost of the analyses low. Moreover the analytes Table 1 of Pb and Cd Range and precision of experimental variables for the composite modified simplex (CMS) optimization of ETAAS for the determination Reverse boundary - - Variable Pb Cd Ashing ramping time/s 5 5 Ashing temperature/oC 300 300 Atomization ramping time/s 0.5 0.5 Atomization temperaturePC 1900 1600 Modifier concentration/mg ml- ' 0.5 0.5 Forward boundary Precision Pb Cd Pb or Cd 1000 800 50 2600 2400 100 50 50 5 8 8 0.5 7 7 0.5 596 Journal of Analytical Atomic Spectrometry August 1996 Vol.11Table 2 Simplex optimization of variables affecting the ETAAS response on the determination of Cd 0.16. Experiment no.* 1 2 3 4 5 6 7R 8E 9R 1 OE 11R 12R 13R 14C 15R 16R 17R 18C 19L 20R 21c 22R 23R 24C 25R 26C 27L 28R 29C 30R (a) Ashing ramping time/s 25 50 40 40 40 40 25 20 20 15 15 15 30 20 5 30 40 15 20 15 20 10 10 20 30 15 15 25 15 15 Ashing temperature/ "C 600 600 800 700 700 700 600 550 5 50 500 3 50 400 750 450 400 600 650 450 500 400 550 400 350 550 600 450 450 550 450 400 Atomization ramping time/s 5.0 5.0 5.0 8.0 6.0 6.0 2.5 0.91 1.5 0.61 1 .o 0.6$ 3.5 1.5 0.61 3.5 4.0 1.5 2.0 0.62 3.5 1.5 1 .o 2.0 2.0 1.5 2.0 2.5 1.5 2.5 Atomization temperature/ "C 2000 2000 2000 2000 2400 2100 2100 2200 1700 1600 1700 1600 1900 1800 1600 2000 2100 1700 1800 1600 1900 1600 1600 1900 2200 1700 1700 1900 1700 1800 Modifier concentration/ mg ml-' 1.5 1.5 1.5 1.5 1.5 4.5 2.0 2.5 2.5 3.5 3.5 0.5 0.5 2.5 2.5 4.0 2.5 2.5 2.5 3.5 2.0 3.5 3.5 2.5 1.5 3.0 3.0 2.5 3.0 3.5 Peak height? (arbitrary units) 0.160 0.143 0.122 0.11 1 0.1 14 0.137 0.175 0.169 0.180 0.172 0.141 0.154 0.127 0.184 0.157 0.180 0.138 0.185 0.171 0.160 0.165 0.177 0.171 0.173 0.1 72 0.190 0.193 0.170 0.193 0.195 * R Reflection; E Expansion; C Contraction.7 Average of three runs. $ The actual variable values used. The lowest possible value of ramping time under the instrumental specifications and the defined atomization temperature. are stabilized and the potential interfering constituents of the samples are presumably converted into more volatile com- pounds which are removed during the ashing process (see steps 3-5 in Table 4).The optimum furnace operating conditions for the analysis of both metals are given in Table4. The cooling down step was included in the furnace heating programme to ensure that the graphite had reached an adequately low temperature before the next injection. The recorded responses during the progress of the simplex as a function of each set of experiments are presented in Fig. 1. The method of introduction of the matrix modifier into the graphite furnace has proved to also affect the efficiency of the analysis of Cd and Pb in the environmental samples.24 There are two ways of introducing the matrix modifier into the graphite furnace when using a programmable sample dispenser in conjunction with ETAAS firstly simultaneous injection of the modifier and the sample (normal injection); and secondly injection of the modifier and optional drying prior to the injection of the sample (pre-injection). Pre-injection of the modifier can be performed by either proper drying of the modifier (ashing temperature) prior to the injection of the sample or wet deposition of the modifier at low temperature followed by injection of the sample. After the construction of the calibration curves by applying the specified parameters the accuracy the precision and the characteristic mass (defined as the mass of the analyte which yields a signal equal to 0.0044 absorbance) corresponding to the sensitivity of each injection method were evaluated and are shown in Table 5.All injection methods exhibited more or less similar and fairly good analyt- ical characteristics. More specifically the results indicate that pre-injection of the modifier by the wet deposition method proved to be slightly superior to the other two methods in AA 0*14t A A A ; I 8 0.12 e A Pb 0 0 5 10 15 20 25 30 35 I v a g 0.20 $ 0.18 0*16 t A A A A A A A O*12t A A A Cd 0.10; 5 10 15 20 25 30 Experiment no. Fig. 1 the determination of Pb and Cd Progress of the simplex during the optimization procedure for terms of the criteria which have been set for the evaluation. This could be attributed to the better mixing between sample and modifier and therefore to the higher efficiency of the reaction for stabilizing the analyte.Journal of Analytical Atomic Spectrometry August 1996 Vol. 1 1 597Table 3 Simplex optimization of variables affecting the ETAAS responst on the determination of Pb Experiment no.* 1 2 3 4 5 6 7R 8E 9R 1 OE 11R 12R 13R 14C 15R 16R 17C 18L 19R 20c 21L 22R 23R 24R 25C 26R 27C 28R 29R 30C 31R Ashing ramping time/s 10 35 25 25 25 25 25 25 25 25 40 30 20 30 10 20 25 25 40 20 20 20 15 15 20 15 20 10 5 20 10 Ashing temperature/ "C 600 600 950 700 700 700 400 300 3 50 300 3 50 300 300 500 450 5 50 3 50 400 400 450 450 300 300 400 500 500 450 450 550 400 450 Atomizatiomn ramping time/s 3.0 3.0 3.0 6.5 4.0 4.0 5.0 6.0 2.0 0.91 3.0 1.5 2.5 3.0 2.5 4.0 2.0 2.5 3.0 2.5 2.5 1.5 2.0 0.91 0.91 1 .o 1 .o 0.91 0.9$ 1.5 1 .O$ Atomization temperature/ "C 2000 2000 2000 2000 2300 2100 2200 2300 2300 2500 2500 2200 2600 2100 2000 2200 2200 2200 2500 2100 2200 2300 2300 2300 2300 2300 2300 2400 2400 2300 2500 Modifier concentration/ mg ml-' 1 .o 1 .o 1 .o 1 .o 1 .o 3.5 3.5 5.0 4.5 6.0 6.0 7.0 7.0 3.0 2.5 0.5 5.5 5.0 6.0 3.5 3.5 4.0 4.5 4.5 4.5 5.0 4.5 4.0 3.5 4.5 5.0 Peak height? (arbitrary units) 0.130 0.141 0.102 0.103 0.130 0.153 0.156 0.152 0.164 0.164 0.149 0.151 0.145 0.155 0.152 0.136 0.156 0.162 0.148 0.159 0.165 0.160 0.157 0.171 0.168 0.161 0.168 0.172 0.163 0.167 0.171 * R Reflection; E Expansion; C Contraction.? Average of three runs. $ The actual variable values used. The lowest possible value of rampirig time under the instrumental specifications and the defined atomization temperature. Table 4 Optimized instrumental parameters for the determination of Pb and Cd* Step 1 Drying 2 Drying 3 Ashing 4 Ashing 5 Ashing 6 Atomization 7 Atomization 8 Clean-out 9 Cooling down TemperaturerC Pb Cd - 150 200 500 500 500 2300 2300 2600 45 95 120 400 400 400 1800 1800 2400 45 'Time/s Pb Cd 25 35 10 10 10 15 1 .o 1 .o 2.0 2.0 0.9 2.0 2.0 2.0 2.0 2.0 12.8 11.8 Gas flow rate/] min-' Pb or Cd 3 3 3 3 0 0 0 3 3 Read command Pb or Cd No No No No No Yes Yes No No ~~ ~ ~~ * Injected volume 20 p1 (sample +matrix modifier).Table 5 The statistical results after the implementation of several injection techniques for the determination of Pb and Cd in the reference material Determined value/pg g- Accuracy 11%) Precision (RSD YO) Characteristic mass/pg* Injection method Pb Cd Pb Cd Pb Cd Pb Cd Normal injection 1.84 0.35 96 103 3.9 3.4 4.9 0.3 Pre-injection wet deposition 1.88 0.34 98 100 3.2 2.4 4.3 0.3 Pre-injection dry deposition 1.84 0.32 96 94 4.9 3.8 4.9 0.3 * Characteristic mass the mass of the analyte which yields a signal equal to 0.0044 absorbance.The conventional analytical procedures using working matrix effect was also investigated by performing recovery experiments on the RM. The results showed a satisfactory recovery for both metals ranging from 95 to 101%. The proposed techniques were also used to determine Cd and Pb content in certain certified environmental reference materials. The results are presented in Table 6 and indicate that the curves with standard solutions prepared in 5% HN03 were applied t o determine the concentrations of the analytes in the mussel tissue RM.Under the optimum experimental con- ditions these concentrations were found to be 0.34 pg g-' and 1.88 pg g-' respectively. The accuracy of dependence on the 598 Journal of Analytical Atomic Spectrometry August 199.6 Vol. 11Table 6 Results for the recovery of Pb and Cd from several environmental Certified Reference Materials applying the optimized techniques Type Certified value Found & s* Certified value Found & s* Marine Sediment (NRC-PACS 1) 404 & 20 392+ 17 2.38 k0.20 2.30 & 0.10 Estuarine Sediment (CEC-CRM 277) 146+ 3 152+5 11.9k0.4 1 1.4 & 0.5 Tea (NRC-CRM CS5-05) 1.06 _+ 0.10 1.01 f0.09 0.032 0.005 0.03 3 k 0.004 * Average of three runs &standard deviation. proposed optimized techniques are successfully applicable to a variety of environmental samples.CONCLUSION Taking into account the great importance and interest of Cd and Pb in nature as environmental pollutants we have pre- sented in this paper an optimized ETAA system for the analysis of these metals in environmental samples. The CMS optimiz- ation procedure has successfully been applied to improve the determination of Cd and Pb in samples with complex matrices. Even if the settings of the optimum experimental conditions cannot be exactly employed in other types of AAS the results are very promising regarding sensitivity accuracy and pre- cision. Since most of the ETAA systems can be automated the proposed techniques offer increased flexibility in developing procedures for the assay of various metals in real samples even if their matrix is unknown.REFERENCES 1 Aylett B. J. ‘Cadmium’ in Comprehensive Inorganic Chemistry ed. Trotman-Dickenson A. F. Pergamon Press Oxford 1973 p 254. 2 Fassett D. W. in Metals in the Environment ed. Waldrom H. A. Academic Press New York 1980 pp 61-1 10. 3 Chaney L. R. Beyer W. N. Gifford C. H. and Sileo L. Trace Subst. Environ. Health 1989 22 263. 4 Robards K. and Worsfold P. Analyst 1991 116 549. 5 Alikhan M. A. Bagatto G. and Zia S. Water Res. 1990,24,1069. 6 Nriagu J. O. Environ. Pollut. 1988 50 139. 7 tobinski R. and Adams F. C. ‘Lead and Organolead Compounds’ in Analysis of Contaminants in Edible Aquatic 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Resources ed. Kiceniuk J. W. and Ray S. VCH New York Fernandez F. J. and Giddings R. At. Spectrosc. 1982 3 61. Olesik J. W. Anal. Chem. 1991 63 1. Salisbury C. D. and Chan W. J. Assoc. Of. Anal. Chem. 1985 68 218. McCarthy H. T. and Ellis P. C. J. Assoc. Off. Anal. Chem. 1991 74 566. Cabrera C. Lorenzo M. L. Gallego C. Lopez M. C. and Lillo E. J . Agric. Food Chem. 1992 40 1631. Okamoto K. and Fuwa K. Anal. Chem. 1984,56 1758. Brown A. A. Lee M. Kullemer G. and Rosopulo A. Fresenius’ 2. Anal. Chem. 1987 328 354. Stephen S. C. Ottaway J. M. and Littlejohn D. Fresenius’ 2. Anal. Chem. 1987 328 346. L‘vov B. V. Specrochim. Acta Part B 1978 33B 153. May T. W. and Brumbaugh W. G. Anal. Chem. 1982,54 1032. Betteridge D. Wade A. P. and Howard A. G. Talanta 1985 32 709. Betteridge D. Wade A. P. and Howard A. G. Talantu 1985 32 723. Betteridge D. Taylor A. F. and Wade A. P. Anal. Proc. 1984 21 373. Stalikas C. D. Karayannis M. I. and Tzouwara- Karayanni S. M. Talanta 1994 41 1561. Pergantis S. A. Cullen W. R. and Wade A. P. Talanta 1994 41 205. Analytical Methods for Graphite Tube Atomizers ed. Rothery E. Varian Techtron Publication 1988. Johnson D. Vurian Analytical Instruments at Work AA-101 November 1990. 1994 pp. 115-157. Paper 6/01 4798 Received March 1 1996 Accepted May 8 1996 Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 599