Anal. Chem. 2008, 80, 9369–9374
Microwave-Induced Combustion Coupled to Flame Furnace Atomic Absorption Spectrometry for Determination of Cadmium and Lead in Botanical Samples ´ rico M. M. Flores* Juliano S. Barin, Fabiane R. Bartz, Valderi L. Dressler, Jose ´ N. G. Paniz, and E Departamento de Quı´mica, Universidade Federal de Santa Maria, 97105-900, Santa Maria, RS, Brazil A procedure based on microwave-induced combustion coupled to flame furnace (FF) atomic absorption spectrometry (FF-AAS) was used for analysis of solid samples. Botanical samples were prepared as pellets and introduced into a quartz holder device. This device was fitted to a glass chamber that was used for the combustion step. The complete device was coupled to the flame furnace by using poly(tetrafluoroethylene) and quartz tubes. The glass chamber was placed inside a microwave oven in a position previously set to receive the higher power of microwave radiation. Ignition was performed by microwave radiation using a small piece of paper wetted with NH4NO3 solution. An oxygen flow was used to assist the sample combustion and also to transport the combustion products up to the heated FF positioned above an air/ acetylene burner. Flame furnace temperature, oxygen flow rate, flame stoichiometry, and sample mass range were evaluated. Cadmium and lead were determined in botanical samples as examples to demonstrate the potential of the proposed procedure for trace analysis. Sample masses up to 60 mg could be used, allowing a limit of detection as low as 0.003 and 0.24 µg g-1 for Cd and Pb, respectively. Integrated absorbance was used with an integration time of 30 s. Background signals were always low, and relative standard deviation (n ) 5) was below 9% for Cd and 11% for Pb. The throughput was 20 determinations/h, including the weighing step. Accuracy was between 94 and 105%, and calibration was performed using standard solutions. The combustion device could be easily adapted to conventional atomic absorption spectrometers. Flame atomic absorption spectrometry (FAAS) is a widespread analytical technique that has been successfully applied to the determination of many elements in several kinds of samples. The robustness and relatively low cost made FAAS very attractive, and a great number of laboratories have currently used this technique for routine analysis. However, FAAS presents some limitations especially regarding the sensitivity and limit of detection (LOD) for some elements that make this technique unsuitable for trace analysis in many matrixes. One of main drawbacks is that samples * To whom correspondence should be addressed. Fax: + 55 55 3220 9445. E-mail:
[email protected]. 10.1021/ac8015714 CCC: $40.75 2008 American Chemical Society Published on Web 11/01/2008
should be introduced, preferentially as liquids by nebulization. In this process, the larger drops are drained off while the smaller ones are carried out up to the flame. Thus, the efficiency of conventional nebulization systems is poor and contributes to the low sensitivity observed because, in general, less than 10% of the solution reaches the atomizer. In addition, the analyte dilution with the flame gases also contributes for the relatively high LOD for FAAS.1-3 Recently, a new method using a metallic tube heated by air/ acetylene flame was proposed to improve the sensitivity of FAAS.4-6 In this system, called thermospray flame furnace atomic absorption spectrometry (TS-FF-AAS), the liquid sample was transported through a flame-heated ceramic capillary producing a thermospray that was immediately introduced into a metallic tube (FF) positioned above an air/acetylene burner. The introduction of the whole sample and the increase of analyte density in the optical path resulted in a significant improvement of LOD for many elements.5 This procedure has been successfully applied for many samples and analytes.7-11 However, despite the better LODs, the application for solid samples requires a previous offline digestion or at least a dissolution step. Thus, some disadvantages related to the sample preparation step remain as analyte dilution, risks of losses/contamination, and low throughput. On the other hand, if solid samples could be directly analyzed or the analyte could be totally introduced into the FF, these disadvantages would be minimized and an increase in the sensitivity and also better LODs would be expected. Although flames are not particularly suitable for solid sample introduction, some attempts have been carried out to overcome (1) Matusiewicz, H. Spectrochim. Acta, Part B 1997, 52, 1711–1736. (2) Welz, B. Spectrochim. Acta, Part B 1999, 54, 2081–2094. (3) Welz, B.; Sperling, M. Atomic Absorption Spectrometry, 3rd ed.; Wiley-VHC: Weinheim, 1999. (4) Wu, P.; Liu, R.; Berndt, H.; Lv, Y.; Hou, X. J. Anal. At. Spectrom. 2008, 23, 37–42. (5) Ga´spa´r, A.; Berndt, H. Spectrochim. Acta, Part B 2000, 55, 587–597. (6) Ga´spa´r, A.; Berndt, H. Anal. Chem. 2000, 72, 240–246. (7) Pereira, M. G.; Pereira-Filho, E. R.; Berndt, H.; Arruda, M. A. Z. Spectrochim. Acta, Part B 2004, 59, 515–521. (8) Davies, J.; Berndt, H. Anal. Chim. Acta 2003, 479, 215–223. (9) Nascentes, C. C.; Kamogawa, M. Y.; Fernandes, K. G.; Arruda, M. A. Z.; Nogueira, A. R. A.; No´brega, J. A. Spectrochim. Acta, Part B 2005, 60, 749–753. (10) Nascentes, C. C.; Arruda, M. A. Z.; Nogueira, A. R. A.; No´brega, J. A. Talanta 2004, 64, 912–917. (11) Aleixo, P. C.; Santos Ju´nior, D.; Tomazelli, A. C.; Rufini, I. A.; Berndt, H.; Krug, F. J. Anal. Chim. Acta 2004, 512, 329–337.
Analytical Chemistry, Vol. 80, No. 23, December 1, 2008
9369
this limitation. Some authors tried to analyze solid samples by introduction as slurries using the conventional pneumatic nebulization.12-14 However, in some cases, problems such as nebulizer blockage were observed depending on the particle size.14 A previous grinding step can be used to overcome these problems,15 but other drawbacks related to contamination during the milling step could be observed if very low particle size is required.16 Special nebulizers could be used to minimize nebulizer clogging (e.g., Babington’s type nebulizers), but the calibration must be currently carried out using a standard addition procedure.17,18 Another way of sample introduction for FAAS is the use of a sampling cup directly heated by the flame. This approach was successfully used by Delves for blood analysis using a small nickel cup and a metallic tube heated by the flame.19 Another tentative procedure was performed by introducing small pieces of filter paper containing the solid sample into a graphite tube, which was heated with an air/acetylene flame. This procedure was proposed for lead determination in sediment samples. However, the lifetimes of graphite tubes were low and they could be used for less than 2 h of heating.20,21 More recently, a procedure for the direct sample introduction using flames as atomizers was proposed for the determination of several elements in biological and environmental samples.22-25 Using this procedure, powdered solid samples were transported as dry aerosols to a heated T-quartz cell and then dispersed into a conventional air/acetylene flame. Among the advantages of this procedure are the low LODs, relatively high throughput, and ease of coupling the introduction device to conventional flame atomic absorption spectrometers. However, the calibration step was not possible by using aqueous solutions, and it must be carried out using solid certified reference materials (CRMs). Another possibility for solid sample analysis using flame atomizers is to separate the analyte vaporization process from the atomization step. In this case, combustion or volatilization of solid samples has been used for further analyte introduction into the atomizer.26,27 Using this approach, is possible to vaporize relatively high sample masses, minimizing the risk of excessive background signals. Moreover, if the vaporized analyte is introduced into a heated tube instead of directly into the flame, better LODs could (12) Alves, F. L.; Cadore, S.; Jardim, W. F.; Arruda, M. A. Z. J. Braz. Chem. Soc. 2001, 12, 799–803. (13) Alves, F. L.; Smichowski, P.; Farias, S.; Marrero, J.; Arruda, M. A. Z. J. Braz. Chem. Soc. 2000, 11, 365–370. (14) Fuller, C. W. Analyst 1976, 101, 961–965. (15) Ferreira, H. S.; Santos, W. N. L.; Fiuza, R. P.; No´brega, J. A.; Ferreira, S. L. C. Microchem. J. 2007, 87, 128–131. (16) May, T. W.; Kaiser, M. L. J. AOAC Int. 1984, 67, 589–593. (17) Fry, R. C.; Denton, M. B. Anal. Chem. 1977, 49, 1413–1417. (18) Mohamed, N.; Fry, R. C. Anal. Chem. 1981, 53, 450–455. (19) Delves, H. T. Analyst 1970, 95, 431–438. (20) Alvarado, J.; Jaffe´, R. J. Anal. At. Spectrom. 1998, 13, 1297–1300. (21) Alvarado, J.; Jaffe´, R. J. Anal. At. Spectrom. 1998, 13, 37–40. (22) Costa, A. B.; Mattos, J. C. P.; Mu ¨ ller, E. I.; Paniz, J. N. G.; Dressler, V. L.; Flores, E. M. M. Spectrochim. Acta, Part B 2005, 60, 583–588. (23) Flores, E. M. M.; Costa, A. B.; Barin, J. S.; Dressler, V. L.; Paniz, J. N. G.; Martins, A. F. Spectrochim. Acta, Part B 2001, 56, 1875–1882. (24) Flores, E. M. M.; Paniz, J. N. G.; Martins, A. F.; Dressler, V. L.; Mu ¨ ller, E. I.; Costa, A. B. Spectrochim. Acta, Part B 2002, 57, 2187–2193. (25) Flores, E. M. M.; Paniz, J. N. G.; Saidelles, A. P. F.; Barin, J. S.; Dressler, V. L.; Mu ¨ ller, E. I.; Costa, A. B. J. Braz. Chem. Soc. 2004, 15, 199–204. (26) Berndt, H. Spectrochim. Acta, Part B 1984, 39, 1121–1128. (27) Campos, R. C.; Curtius, A. J.; Berndt, H. J. Braz. Chem. Soc. 1990, 1, 66– 71.
9370
Analytical Chemistry, Vol. 80, No. 23, December 1, 2008
be achieved.28 However, up to now only sample masses lower than 10 mg have been used and better LODs for solid samples could be expected if higher masses could be analyzed. In principle, the previous destruction of matrix, the vaporization of the analyte, and total signal integration should allow the calibration without solid CRMs. However, for flame atomizers, this kind of procedure was described only using a previous electrothermal vaporization step.29,30 In recent years, a system was proposed for the combustion of solid samples in oxygen-pressurized vessels using microwave radiation for ignition.31 This technique, called microwave-induced combustion (MIC), combines the advantages of classical combustion techniques with those from conventional closed systems heated by microwave radiation, allowing a complete digestion in few minutes with minimum acid consumption.32 In this system, samples are pressed as pellets and positioned on a small piece of low-ash filter paper that is previously placed on the quartz holder. About 50 µL of 6 mol L -1 ammonium nitrate solution is added as igniter, and the system is closed and pressurized with oxygen (15-25 atm). The ignition occurs within 3-10 s, and the time necessary for absorption of analytes may be only 5-10 min. The procedure was successfully applied for several kinds of samples for both metal and nonmetal determination.31,33-36 A special feature of the MIC technique is that the ignition can be carried out using microwave radiation, avoiding additional devices such as metallic electrodes, which can interact with sample and analyte.37 In this work, an online coupling of MIC to a flame furnace for analysis of solid samples by FAAS is proposed. Sample masses between 2.5 and 60 mg were burnt using oxygen flow to assist the combustion process and also to carry the products of combustion/vaporization step to the analysis by flame furnace atomic absorption spectrometry. To demonstrate the performance of the proposed microwave-induced combustion flame furnace atomic absorption spectrometry (MIC-FF-AAS) system for trace analysis, cadmium and lead were determined in botanical samples. Certified reference materials were used to check the accuracy. Operational conditions were evaluated and the results were compared to those using other devices proposed in the literature to analyze solid samples by FAAS. (28) Campos, R. C.; Curtius, A. J.; Berndt, H. J. Anal. At. Spectrom. 1990, 5, 669–673. (29) Lee, Y. I.; Kim, J. K.; Kim, K. H.; Yoo, Y. J.; Back, G. H.; Lee, S. C. Microchem. J. 1998, 60, 231–241. (30) Riter, K. L.; Matveev, O. I.; Smith, B. W.; Winefordner, J. D. Anal. Chim. Acta 1996, 333, 187–192. (31) Flores, E. M. M.; Barin, J. S.; Paniz, J. N. G.; Medeiros, J. A.; Knapp, G. Anal. Chem. 2004, 76, 3525–3529. (32) Flores, E. M. M.; Barin, J. S.; Mesko, M. F.; Knapp, G. Spectrochim. Acta, Part B 2007, 62, 1051–1064. (33) Mesko, M. F.; Moraes, D. P.; Barin, J. S.; Dressler, V. L.; Knapp, G.; Flores, E. M. M. Microchem. J. 2006, 82, 183–188. (34) Flores, E. M. M.; Mesko, M. F.; Moraes, D. P.; Pereira, J. S. F.; Mello, P. A.; Barin, J. S.; Knapp, G. Anal. Chem. 2008, 80, 1865–1870. (35) Barin, J. S.; Flores, E. M. M.; Knapp, G. Trends in sample preparation using combustion techniques. In Trends in Sample Preparation; Arruda, M. A. Z., Ed.; Nova Science Publishers: Hauppauge, 2006; Chapter 3, pp 73-114. (36) Moraes, D. P.; Mesko, M. F.; Mello, P. A.; Paniz, J. N. G.; Dressler, V. L.; Knapp, G.; Flores, E. M. M. Spectrochim. Acta, Part B 2007, 62, 1065– 1071. (37) Hassan, H. N. A.; Hassouna, M. E. M.; Gawargious, Y. A. Talanta 1988, 35, 311–313.
EXPERIMENTAL SECTION Reagents. Milli-Q water (18.2 MΩ cm) and analytical-grade reagents (Merck, Darmstadt, Germany) were used. Working standard solutions for Cd and Pb determination were prepared before use by serial dilutions of each respective stock reference solution of 1000 mg L-1 (Spex Certi Prep, Metuchen, NJ). Ammonium nitrate solution (6 mol L-1) was used as an ignition aid. A small piece of filter paper (1.5 × 1.5 cm, 25 mg) with low ash content (Black Ribbon Ashless, Schleicher & Schuell GmbH, Dassel, Germany) was used to wrap the sample and also as an aid for the combustion process. The filter paper was previously treated with 10% nitric acid in an ultrasound bath for 20 min and further dried in an oven by 2 h at 60 °C before use. Glass or quartz materials were soaked in 10% (m/v) HNO3 for 12 h and thoroughly washed with water before use. Samples were dried, ground, and pressed as pellets. The following CRMs were used in this work: aquatic plant (IRMM BCR 60), olive leaves (IRMM BCR 62), oriental tabaco leaves (ICHTJ-cta-OTL-1), and pine needles (SRM NIST 1575). For the optimization of the proposed system, a teatype sample (Ilex paraguariensis) obtained from a local market was employed. This sample is a native plant in South America, called erva-mate, which leaves are commonly used in hot beverages. Instrumentation. For the combustion step, a microwave oven (Panasonic, model NN-S52 B) with 900 W of nominal power was used. For Cd and Pb determination, a flame atomic absorption spectrometer (Varian, model SpectrAA 600), equipped with a deuterium lamp background correction system, was used. Hollow cathode lamps were operated at 4 and 10 mA for Cd and Pb, respectively. Wavelengths were set at 228.8 (Cd) and 283.3 nm (Pb), and a 0.5-nm spectral resolution was used for both elements. All measurements were carried out using an integrated absorbance mode with an integration time of 30 s. A conventional oven (Nova E´tica, model 400/2ND) was used for sample drying. An ultramicrobalance (Sartorius, model M2P) with a resolution of 1 µg and weighing range up to 2 g was used for weighing the samples. A hydraulic press, set at 8 tons (Specac, model hydraulic press, 15 ton), was used to prepare the pellets of samples. Infrared thermometer (Ircon, model Ultimax 20 infrared thermometer) was used for temperature measurements of the metallic tubes. For comparison of results, a microwave oven (Milestone, model ETHOS 1) equipped with up to 10 poly(tetrafluoroethylene) (PTFE)-perfluoroalkoxy vessels was used to sample decomposition. The determination of Cd and Pb after sample digestion was carried out by inductively coupled plasma mass spectrometry (ICPMS) and by graphite furnace atomic absorption spectrometry (GF AAS). The inductively coupled plasma mass spectrometer (PerkinElmer, model Sciex-Elan DRC II) was equipped with a concentric nebulizer (Meinhard Associates, Golden, CO), a cyclonic spray chamber (Glass Expansion, Inc., West Melbourne, Australia), and a quartz torch with a quartz injector tube (2-mm i.d.). Instrumental performance optimization, including ion lens voltage and torch alignment, was carried out according to the instructions of the manufacturer using conventional nebulization.38 Other operational conditions were as follows: rf power of 1400 W, plasma gas flow rate of 15 L min-1, auxiliary gas flow rate of (38) PerkinElmer. Software Guide, ICP Mass Spectrometry, ELAN Version 3.0, 2003.
Figure 1. Device used for the proposed MIC-FF-AAS procedure. Combustion chamber and expansion chamber have internal volumes of 40 and 20 mL, respectively.
1.2 L min-1, and nebulizer gas flow rate of 1.15 L min-1. The graphite furnace atomic absorption spectrometer (Analytik Jena, model ZEEnit 60) was equipped with a transversely heated graphite tube atomizer and a Zeeman-effect background correction system. In this case, the magnetic field was set at 0.8 T for both elements. Proposed Procedure. For the proposed procedure by MICFF-AAS, the microwave oven was modified to allow an oxygen inlet to the combustion chamber and transport of the combustion products up to the FF. The FF (10-cm length, 11-mm o.d., and 9-mm i.d.) was produced from a metallic bar (Camacam, Inconel 600, 72% Ni, 15% Cr, 8% Fe, 0.5% Cu, and 0.3% Tl). The distance between the burner and the FF was set at 4 mm. The combustion chamber was made with borosilicate glass and the sample holder with quartz (2.5-cm length). This chamber was connected by a PTFE glove (3-mm i.d., 2-cm length) to a second glass chamber (expansion chamber; details are given in Figure 1). The expansion chamber was necessary to minimize eventual excessive pressure during the beginning of combustion. The transport of combustion products from the expansion chamber up to the FF was performed using a PTFE tube (10-cm length and 3-mm i.d.) connected to a quartz tube (30-cm length and 3-mm i.d.). The end of this quartz tube was fitted to the FF. For the proposed procedure, pellets of samples were weighed directly on the filter paper. Sample masses from 2 to 80 mg were evaluated. After weighing, samples were wrapped and placed on the sample holder. Ammonium nitrate (5-50 µL of 6 mol L-1 solution) was added to the paper, and the quartz holder was placed into the glass chamber. The combustion device was introduced into the microwave oven, and oxygen was passed through the glass chamber. Further, irradiation with microwaves was applied until the start of sample ignition (∼8 s). After ignition, the microwave radiation was stopped and the atomic signals were recorded. Integrated absorbance was used throughout for the proposed MIC-FF-AAS procedure. For the calibration step, Cd and Pb standard solutions were added to the filter paper and the same procedure was used for combustion and transport to the FF. Sample Digestion. For the comparison of results, samples were digested using closed vessels with microwave heating for Analytical Chemistry, Vol. 80, No. 23, December 1, 2008
9371
Figure 2. Analytical signal profiles for (a) Cd (5 ng) and (b) Pb (256 ng) obtained for solid samples with the proposed procedure. Flame stoichiometry: air and acetylene, 14 and 2 L min-1 for Cd and 12 and 2 L min-1 for Pb. Oxygen flow rate: 1 L min-1. Flame furnace without holes and with 12 holes for Cd and Pb, respectively.
the subsequent determination of Cd and Pb by ICPMS and also by GF AAS. About 300 mg of sample was digested using 7 mL of concentrated nitric acid, 0.5 mL of hydrofluoric acid, and 1 mL of 30% hydrogen peroxide. The microwave heating program was set at 1000 W during 20 min, with 30 min for cooling (based on the recommendation of the manufacturer).39 Digests were diluted with water to 25 mL in polypropylene vessels. RESULTS AND DISCUSSION Different approaches were used to the combustion of samples during the development of the proposed procedure. In these initial studies, holders were constructed in glass, and contrarily for small sample masses, with the combustion of masses higher than 15 mg a slight melting of the holder was observed. Using quartz, no problems were detected even using higher sample masses and this material was used for further tests. Initially, the combustion was carried out using a spherical glass chamber. However, overpressure was observed when sample masses higher than 20 mg were combusted. Then, an additional chamber (expansion chamber) was fitted in the end of the original combustion chamber (Figure 1) to prevent excessive pressure during the combustion. The concentration of NH4NO3 solution used as igniter was chosen based on a previous work.31 Taking into account the different characteristic of the proposed system related to the original work, different volumes of 6 mol L-1 NH4NO3 solution were evaluated (between 5 and 50 µL). With volumes lower than 10 µL and higher than 30 µL, the combustion was not reproducible. With 20 µL of NH4NO3 solution, the ignition was more reproducible and this volume was selected for the subsequent experiments. Flame Furnace. The performance of the FF depends on the temperature and chemical environment inside the metallic tube. A simple way to improve these conditions is the use of holes in the bottom side of the FF that allows the entrance of small amount of hot flame gases.6,10 This approach allows the achievement of higher temperatures that improves the performance for many analytes. Thus, to evaluate the effect of additional holes, tubes without holes and tubes with 6 and 12 holes of 2-mm diameter were used. Another tube with a longitudinal slit was also tested only for Pb determination. The improvement in the characteristic mass (m0) with additional holes at the bottom of the FF was observed only for Pb, while for Cd, better results were obtained using tubes without holes. Probably, this improvement for Pb (m0 ) 15.9 ± 4.3 ng for tube without holes and m0 ) 4.2 ± 0.3 ng for tube with 12 holes) occurred due to the higher temperature in the tube with holes. (39) Milestone . Application notes for digestion, ETHOS 1, 2006.
9372
Analytical Chemistry, Vol. 80, No. 23, December 1, 2008
Table 1. Results for Cd and Pb Determination Using the Proposed MIC-FF-AAS Procedure (Results in µg g-1, n ) 5) analyte
CRM
certified value
MIC-FF-AAS
Cd
IRMM BCR 60 IRMM BCR 62 ICHTJ-cta-OTL-1
2.20 ± 0.10 0.10 ± 0.02 1.12 ± 0.12
2.13 ± 0.11 0.105 ± 0.009 1.05 ± 0.08
Pb
IRMM BCR 60 IRMM BCR 62 ICHTJ-cta-OTL-1 SRM NIST 1575
63.8 ± 3.2 25.0 ± 1.5 4.91 ± 0.79 10.8 ± 0.5
66.1 ± 4.1 24.2 ± 0.7 5.18 ± 0.42 10.2 ± 0.6
On the other hand, for highly volatile elements (as for Cd), the decrease in the atom density of the atoms in the optical path due to the flame gases reduces the sensitivity28 (m0 ) 82.4 ± 5.3 pg for a tube with 12 holes and 59.8 ± 3.5 pg for a tube without holes). Thus, FF without holes was chosen for Cd determination. The results observed for Pb using a FF with a longitudinal slit were similar to those using a tube with 12 holes. However, the tube lifetime was reduced from more than 500 h to only 50 h. Therefore, FF with 12 holes was used for Pb determination. Using the previous selected conditions for both analytes, the sensitivity was not significantly changed if standard solutions or solid samples were analyzed. This similarity is important to allow the calibration avoiding the use of solid CRMs. Probably, the previous combustion/vaporization step minimizes the differences between solid samples and standard solution. During this process, the solid matrix could be converted to more simple substances, mainly CO2 and H2O, which have low influence in the atomization environment. This could also be the reason for the low background signals observed for both elements, as shown in Figure 2. Therefore, the analysis of solid samples with air/acetylene heated FF and using standard solutions for calibration was feasible. Oxygen Flow Rate. The optimization of oxygen flow rate is an important parameter to ensure a suitable combustion process and also the transport of combustion products up to the FF. For oxygen flow rates lower than 1 L min-1, the ignition was not reproducible. The same situation was observed for flow rates higher than 2 L min-1. For this reason, the oxygen flow rate was evaluated from 1 to 2 L min-1. For both elements, the optimum oxygen flow rate was 1 L min-1. When O2 flow rate was increased to 1.5 or 2 L min-1, the characteristic masses were about 24 and 50% higher for Cd and Pb, respectively. This result was expected due to the decrease of atom density in the optical path caused by higher O2 flow rates. Moreover, the relative standard deviation of measurements was always lower than 11% (n ) 5) using an oxygen flow rate of 1 L min-1.
Figure 3. Influence of sample mass for (a) Cd and (b) Pb determination by MIC-FF-AAS (n ) 5). Oxygen flow rate, 1 L min-1; tube for Cd and Pb, without holes and with 12 holes, respectively. Air and acetylene flow rates for Cd and Pb, 14 and 2 L min-1 and 12 and 2 L min-1, respectively. The lines represent the mean and the standard deviation (sd) for the mass range without statistical differences. Table 2. Characteristic Mass, Limit of Detection, and Maximum Sample Mass Observed for Different Techniques Cd technique conventional F AAS27 FAAS28 TS-FF-AAS5 sol-GF AAS SS-F AAS27 SS-F AAS28 MIC-FF-AAS a
m0 (ng) 2.5 0.0019 0.001 0.2 0.006 0.05
Pb LOD (µg g-1)
sample mass (mg)
0.6a 6.8b 0.06b 0.003c 0.03 0.003
10 2 60
m0 (ng) 50 0.13 0.02 4.4 0.35 4.1
LOD (µg g-1) 20a 44b 0.48b 0.02c 1.62 0.24
sample mass (mg)
10 2 60
-1 b
µg L . Calculated value considering the introduction of 200 µL of sample, 250 mg of sample digested and diluted to 25 mL. c 250 mg of digested sample and diluted to 25 mL; SS-F AAS, solid sampling flame atomic absorption spectrometry.
Flame Stoichiometry. Air and acetylene flow rates of flame affect the temperature of the FF and the atomization environment inside the tube with additional holes.6,27,28 In this work, different air/acetylene flow rates were evaluated: 12/2.4, 12/2, 14/2, 16/ 2, and 18/2 L min-1, respectively. As expected for Cd, the flame stoichiometry did not have influence on the sensitivity, because the flame does not reach the atomization environment inside the tube without holes. However, lower standard deviations were obtained using air and acetylene flow rates of 14 and 2 L min-1, respectively. For Pb, changes in the flame stoichiometry caused a more intense effect. This behavior was expected, because a FF with holes was used for this element. Then, a small part of flame that could enters into the FF and probably contributes to the increase of the temperature atomization and consequently to the improvement of sensitivity. For Pb, the characteristic mass was 25% lower with 12 L min-1 air and 2 L min-1 acetylene. These air and acetylene conditions, 14 and 2 L min-1 and 12 and 2 L min-1, were used for the subsequent measurements of Cd and Pb, respectively. The temperature distribution along the wall for both FFs investigated was evaluated for the selected air and acetylene flow rates. The temperature for both FFs presented a similar profile. The hottest part for both tubes was the region around 1.5 cm from the center of FF (1100 and 1050 °C for FF with 12 holes and without holes, respectively). For other parts, the temperature was lower: 1040 and 970 °C in the middle part and 1020 and 950 °C in the ends, for FFs with 12 holes and without holes, respectively. The lower temperature in the middle part is due to the constant oxygen introduction that causes the temperature decrease. The extremity of FF is not well heated by the flame and this could explain the lower temperature for both FFs. Then, optimization of flame stoichiometry and FF type could be
important for other elements, especially for less volatile analytes, that could be determined by MIC-FF-AAS. Sample Mass Range. The influence of sample mass was evaluated by plotting the analyte concentration as a function of the sample mass.40 In this experiment, sets of measurements for Cd and Pb were divided in mass intervals of 10 mg (total of 55 measurements for Cd). Their respective standard deviations were compared using analysis of variance (ANOVA) and considering a confidence interval of 95% (Tukey-Kramer multiple comparisons test). The results are shown in Figure 3. For Pb, the optimum mass range was between 10 and 60 mg (total of 50 measurements). In this case, minimum sample mass of 10 mg must be used due to high dispersion of the obtained results. This variation could be related to the lack of homogeneity of analyte with sample masses lower than 10 mg. For Cd, sample masses up to 60 mg could be also burnt with success. Using sample masses higher than 60 mg, a black residue was observed in the glass chamber indicating incomplete sample combustion and a washing step must be applied after each run. With 60 mg of sample mass, up to 10 combustion rounds could be made without signal decrease. Therefore, the maximum sample mass was 60 mg for both analytes. Using 60 mg, the characteristic masses obtained were 0.055 and 4.1 ng and the LOQ were 0.009 and 0.72 µg g-1 for Cd and Pb, respectively. Determination of Cd and Pb in CRMs. The results obtained using MIC-FF-AAS procedure for Cd and Pb determination in different CRMs are described in Table 1. For both analytes, the agreement with the certified values was between 94 and 105%. For different tea samples, the Cd and Pb concentrations obtained (40) Rodrigues, L. F.; Mattos, J. C. P.; Dressler, V. L.; Pozebon, D.; Flores, E. M. M. Spectrochim. Acta, Part B 2007, 62, 933–938.
Analytical Chemistry, Vol. 80, No. 23, December 1, 2008
9373
using MIC-FF-AAS and GF AAS or ICPMS after digestion showed an agreement between 95 and 106%. Comparison of MIC-FF-AAS Proposed Procedure with Other Procedures. The characteristic mass, limit of detection, and maximum sample mass obtained by the proposed procedure were compared with other techniques as shown in Table 2. For Cd and Pb using the proposed MIC-FF-AAS procedure, the characteristic mass in relation to the conventional FAAS technique with liquid sample introduction was increased by a factor of 50 and 12, respectively. Moreover, the characteristic mass obtained by MIC-FF-AAS was higher when compared with other systems that perform the analysis of solid samples by FAAS using three infrared lamps and a T-tube.28 This result can be explained due to the relatively high dead volume of the combustion chamber of the proposed system (60 mL) that increases the dispersion of the combustion products during the transport up to the FF and also to the relatively low temperature of FF (∼1100 °C). On the other hand, for the MIC-FF-AAS technique, LOD was improved when compared with the other works.5,27 For Cd, the LOD was improved by a factor higher than 2500 times. For Pb, a factor of 183 times was achieved in relation to the LOD of conventional FAAS28 with introduction of liquid solution. When compared with conventional graphite furnace atomic absorption spectrometry using solutions (sol-GF AAS), the LOD for Cd was similar to this technique in comparison with MIC-FF-AAS and ∼10 times higher for Pb.
9374
Analytical Chemistry, Vol. 80, No. 23, December 1, 2008
CONCLUSION The proposed procedure using MIC-FF-AAS is simple and allowed the determination of Cd and Pb in solid samples. Relatively low LOD was obtained in view of the relatively high sample masses used. The calibration with standard solutions can be considered an important advantage when compared with other methods by solid sampling using FAAS.5,27,28 Moreover, the combustion device can be easily fitted to conventional atomic absorption spectrometers. Low background signals and good accuracy were obtained for both analytes. The procedure was exactly the same for different samples, and up to 20 determinations can be performed in 1 h (excluding the grinding and pressing steps). Finally, this procedure allows the reduction or elimination of hazardous substances and their consequent residues. ACKNOWLEDGMENT The authors thank to CAPES, CNPq, and Farmacope´ia Brasileira/ANVISA. The authors also thank Dr. Marco Aurélio Zezzi Arruda and Dr. Dário Santos Jr. for the important comments during the development of this work.
Received for review July 25, 2008. Accepted October 1, 2008. AC8015714
