Anal. Chem. 2004, 76, 3525-3529
Microwave-Assisted Sample Combustion: A Technique for Sample Preparation in Trace Element Determination E Ä rico Marlon de Moraes Flores,*,† Juliano Smanioto Barin,† Jose´ Neri Gottfried Paniz,† Joa˜o Alfredo Medeiros,‡ and Gu 1 nter Knapp§
Departamento de Quı´mica, Universidade Federal de Santa Maria, 97105-900 Santa Maria, RS, Brazil, Instituto de Quı´mica, Universidade Federal do Rio de Janeiro 21949-900 Rio de Janeiro, RJ, Brazil, and Institute for Analytical Chemistry, Micro- and Radiochemistry, Graz University of Technology, Technikerstrasse 4, A-8010 Graz, Austria
A novel digestion procedure based on sample combustion ignited by microwave radiation is proposed for organic samples. Certified samples of bovine liver, pig kidney, and skim milk were used as examples to demonstrate the performance of the proposed procedure. Cadmium and copper were determined in these samples by electrothermal atomic absorption spectrometry. Samples (between 50 and 250 mg) were wrapped with paper and placed on a homemade quartz holder that was positioned inside to quartz vessels used in a commercial microwave oven. Ammonium nitrate solution was added to the paper, and vessels were pressurized with oxygen to 15 bar. The rotor containing four vessels was placed inside the oven, and microwave radiation was applied for 20 s at 1400 W. Combustion was complete in few seconds, and an additional reflux step, which was optional, was applied. The agreement to the certified values was between 96 and 105% for both analytes. Only with the combustion step, the residual carbon (RC) was below 1.3%. The RC decreased to less than 0.4% when an additional reflux step with concentrated nitric acid was applied. In past years, sample digestion has been considered as an Achilles' heel in the analytical sequence. It is attributed to some drawbacks that could occur in this step and could cause errors, mainly systematic, up to 100% in the final result for trace element determinations.1 Despite the recent and accelerated development of instrumentation for analytical techniques, the development of commercial sample digestion systems has been relatively slow. Consequently, sample digestion is often regarded as a weak link in sample analysis that provides much scope for improvement.2,3 The growing number of papers dealing with that topic shows the increasing interest in analytical chemistry.4-6 Depending on * Corresponding author. Fax: +55 55 220-8054. E-mail:
[email protected]. † Universidade Federal de Santa Maria. ‡ Universidade Federal do Rio de Janeiro. § Graz University of Technology. (1) Wagner, G. Sci. Total Environ. 1995, 176, 63-71. (2) Lamble, K. J.; Hill, S. J. Analyst 1998, 123, 103R-133R. (3) Arruda, M. A. Z.; Gallego, M.; Valca´rcel, M. J. Anal. At. Spectrom 1996, 11, 169-173. (4) Hoenig, M. Talanta 2001, 54, 1021-1038. (5) Richter, R. C.; Link, D.; Kingston, H. M. Anal. Chem. 2001, 73, 30A-37A. 10.1021/ac0497712 CCC: $27.50 Published on Web 05/11/2004
© 2004 American Chemical Society
sample matrix or kind of association of the analyte to matrix, digestion may not be complete or may be a very time-consuming step.7 In addition, the presence of residual carbon (RC) at high levels may strongly interfere in analysis performed by analytical techniques such as inductively coupled plasma mass spectrometry,8 inductively coupled plasma optical emission spectrometry,9,10 voltametry,11 and atomic absorption spectrometry.12 Decomposition of solid samples containing an organic matrix that could interfere with the analytical determination of an element can be achieved in many different ways, e.g., open-vessel hot plate digestion or a closed-vessel technique using a variety of commercially available digestion systems.13-15 Microwave digestion using closed systems has gained widespread acceptance as an effective method for sample preparation, reducing the digestion times and reagent amounts, avoiding contamination and loss of volatile species, and improving safety.16 In these systems, concentrated mineral acids are used to decompose the sample, which may increase the blank value caused by elements present as contaminants. In addition, the presence of high acid concentrations may not be supported by some analytical techniques and a subsequent step to remove (or dilute) the acid excess may be necessary. (6) No´brega, J. A.; Trevizan, L. C.; Araujo, G. C. L.; Nogueira, A. R. A. Spectrochim. Acta 2002, 57 B, 1855-1876. (7) Matusiewicz, H.; Sturgeon, R. E. Fresenius J. Anal. Chem. 1994, 66, 428433. (8) Krachler, M.; Radner, H.; Irgolic, K. Fresenius J. Anal. Chem. 1996, 355, 120-128. (9) Wasilewska, M.; Goessler, W.; Zischka, M.; Maichin B.; Knapp, G. J. Anal. At. Spectrom. 2002, 17, 1121-1125. (10) Knapp, G.; Maichin, B.; Baumgartner, U. At. Spectrosc. 1998, 19, 220222. (11) Wu ¨ rfels, M.; Jackwerth, E.; Stoeppler, M. Fresenius J. Anal. Chem. 1987, 329, 459-461. (12) Golimowski, J.; Golimowska, K. Anal. Chim. Acta 1996, 325, 111-133. (13) Kingston, H. M., Jassie, L. B., Eds. Introduction to Microwave Sample Preparation. Theory and Practice; American Chemical Society: Washington, DC, 1988. (14) Bock, R. A Handbook of Decomposition Methods in Analytical Chemistry; Wiley: New York, 1979. (15) Sulcek, Z.; Povondra, P. Methods of Decomposition in Inorganic Analysis; CRC Press: Boca Raton, FL, 1989. (16) Kingston, H. M., Haswell, S. H., Eds. Microwave-enhanced chemistry. Fundamentals, sample preparation, and applications; American Chemical Society: Washington, DC, 1997.
Analytical Chemistry, Vol. 76, No. 13, July 1, 2004 3525
In the past, sample combustion in open vessels has been extensively used to decompose samples containing high levels of carbon. However, some drawbacks, namely, losses of analyte due to volatilization, impaired its use for trace analysis. An alternative, to partially solve that problem, was the use of closed devices for sample combustion. The oxygen flask combustion (first introduced by Hempel and further improved by Scho¨niger),17,18 and combustion in calorimeter bombs (oxygen bombs) are old methods but are still in use.19 In these systems, samples are burnt in the presence of excess of oxygen and the gaseous products are absorbed in a solution into the same vessel used for combustion. The advantage of these procedures is the use of oxygen, which is pure in relation to its trace element content compared to the acids used in wet ashing. Moreover, the decomposition of organic matter is effective and may be performed in few minutes.20 The conventional oxygen flask combustion system is a very simple technique, using a flask with a ground-glass stopper in which a platinum wire is sealed. The wire can be connected to a platinum basket in which the sample is placed wrapped with a piece of paper. The paper is ignited manually and placed into the flask as quickly as possible, either ignited electrically or by means of focused infrared radiation. Normally, samples of 10-50 mg are burnt in a closed oxygen-filled flask with a volume of 250-500 mL.14 The major drawbacks of this technique are the limited sample mass and unfavorable ratio of vessel surface to sample. The loss of some elements due to alloy formation with platinum has also been reported.21 However, that method has been especially used for subsequent determination of halogens, sulfur, phosphorus, and some metals. Combustion of higher masses of sample, opposite of those supported by the oxygen flask, can be obtained in calorimeter bombs. The sample is placed in a crucible, and oxygen is allowed in up to a pressure of ∼25 atm in a stainless steel vessel with a screw or snap cap. A glowing platinum or Ni-Cr wire in contact with the sample starts its combustion. The main advantage of the method over the oxygen flask is the increasing amount of samples that can be combusted, while the ratio of vessel surface to sample mass is smaller. A disadvantage is the susceptibility to contamination due to the metallic surfaces of the bomb. However, despite being able to prepare just one sample each time, that method may be considered the state of the art for sample preparation of organic materials when coupled to ion chromatography for the determination of halogens and some other nonmetals. In this paper, a procedure involving the combustion of organic samples, in closed vessels, ignited by microwave radiation is demonstrated. Quartz vessels are pressurized with oxygen, and combustion is started by microwave radiation after the vessels are hermetically closed. This new concept for sample digestion is described for the first time and allows the combination of the advantages of classical combustion techniques with those from conventional closed systems heated by microwave radiation. In this procedure, a fast and complete digestion is performed in a (17) Hempel, W. Z. Ang. Chem. 1892, 13, 393-394. (18) Scho ¨niger, W. Mikrochim. Acta 1955, 123-129. (19) Souza, G. B.; Carrilho, E. N. V. M.; Oliveira, C. V.; Nogueira, A. R. A.; No´brega, J. A. Spectrochim. Acta, B 2002, 57, 2195-2201. (20) Iyengar, G. V.; Subramanian, K. S.; Woittiez, J. R. W. Element Analysis of Biological Samples-Principles and Pratice; CRC Press: Boca Raton, FL, 1997. (21) Hassan, H. N. A.; Hassouna, M. E. M.; Gawargious, Y. A. Talanta 1988, 35, 311-313.
3526 Analytical Chemistry, Vol. 76, No. 13, July 1, 2004
Table 1. General Technical Data and Operating Conditions for the Multiwave 3000 System Using High-Pressure Quartz Vessels parameter vessel volume (mL) maximum power (W) maximum operation pressure (bar) maximum operation temperature (°C) minimum filling solution volume (mL) maximum filling solution volume (mL)
80 1400 80 280 6 40
few minutes in a pressurized oxygen atmosphere with minimum acid consumption. The combustion procedure was developed using a commercial microwave system with quartz vessels without modifications in either the microwave oven or the device for vessel capping. To demonstrate the potential for sample preparation, the proposed procedure was applied to digestion of bovine liver, pig kidney, and skim milk samples. Cadmium and copper were chosen as an application example, and determinations were made by electrothermal atomic absorption spectrometry. The new digestion procedure was evaluated with certified reference materials, showing the high efficiency of the proposed combustion procedure. EXPERIMENTAL SECTION Reagents. Analytical-grade reagents (Merck, Darmstadt, Germany) were used unless otherwise indicated. Concentrated nitric acid (65%) was doubly distilled in a sub-boiling system (Milestone, model DuoPur). Milli-Q water (18.2 MΩ cm) was used to prepare all solutions. Working analytical solutions for Cd and Cu were prepared immediately before use by serial dilution of stock reference solutions containing 1000 mg L-1 (Merck, Titrisol, Darmstadt, Germany). Ammonium nitrate was dissolved in water, and that solution was used as igniter for the combustion procedure. A piece of paper (Colomy cigarette paper, density of 1.75 mg cm-2, Souza Cruz Co.) with low ash content was used to wrap the sample. The paper was treated with a 5% (m/v) HNO3 solution for 1 h and dried in an oven for 2 h at 60 °C before use. All glass apparatus was soaked in 10% (m/v) HNO3 for 48 h and thoroughly washed with water before using. The following reference samples were used in this work: bovine liver (NIST SRM 1577 and 1577b), lyophilized pig kidney (IRMM CRM 186), and spiked skim milk powder (IRMM CRM 151), with certified Cd concentrations of 0.27 ( 0.04, 0.50 ( 0.03, 2.71 ( 0.15, and 0.101 ( 0.008 mg g-1, respectively. The corresponding Cu concentrations in the reference samples were 193 ( 10, 160 ( 8, 31.9 ( 0.4, and 5.23 ( 0.08 mg g-1. Instrumentation. A Multiwave 3000 microwave sample preparation system (Anton Paar, Graz, Austria) equipped with up to eight high-pressure quartz vessels was used in this study. Technical details and recommended experimental conditions of this equipment are summarized in Table 1. Homemade holders were used to put the sample inside the quartz vessel. The holders were constructed in quartz, and their dimensions as well as some details of the combustion system are shown in Figure 1. The microwave energy program used for the combustion procedure was as follows: (1) 1400 W for 20 s, (2) 0 W for 2 min, (3) 1400 W for 8 min (optional step), and (4) 0 W for 20 min for cooling if step 3 was applied. During these steps, the cooling fan was
Figure 1. High-pressure quartz vessel with sample quartz holder (a) and details of holder (b) for the combustion microwave-assisted sample digestion procedure (dimensions in mm).
selected at level 2 (level 3 was applied only for cooling). The software version was v1.27-SYNT, and the microwave system was previously modified to run with a security pressure variation up to 3 bar s-1. In this work, each run was always performed with four vessels. Cadmium and copper determinations were carried out with a model AAS EA5 atomic absorption spectrometer (Analytik, Jena, Germany) equipped with a continuum source background correction system, MPE 5 furnace autosampler, and electrothermal atomizer with a transversely heated graphite tube. Argon with a purity of 99.996% (White Martins, Sa˜o Paulo, Brazil) was used as the purge and protective gas for the graphite atomizer. Hollow cathode lamps for Cd and Cu (NARVA, G.L.E., Berlin, Germany) as well as the temperature program were used under the conditions recommended by the manufacturer. All measurements were made in the integrated absorbance mode. The RC was determined by inductively coupled plasma optical emission spectrometry (ICP-OES, Optima 2000, Perkin-Elmer, Norwalk, CT), and measurements were made according to the conditions described in ref 22. For the determination of acid consumption for the proposed procedure, a digital potentiometer (Hanna, model pH211, Sarmeola di Rubano, Italy) equipped with a combined Ag/AgCl, a saturated KCl reference electrode, and a glass indicator electrode was used to perform the acid-base titration. Sample Preparation. For the proposed combustion procedure, test samples between 50 and 250 mg were weighed directly on the paper (3.5 cm × 2.3 cm, 14 mg). After weighing, samples were wrapped with paper and placed in the quartz holder. The quartz vessels were previously charged with 6 mL of concentrated nitric acid as absorbing solution. The holder containing the sample was positioned into the quartz vessel, and 50 µL of ammonium nitrate solution was immediately added to the paper. After the (22) Gouveia, S. T.; Silva, F. V.; Costa, L. M.; Nogueira, A. R. A.; No´brega, J. A. Anal. Chim. Acta 2001, 445, 269-275.
closure of the vessels and capping of the rotor, they were pressurized with oxygen between 5 and 15 bar for 2 min. Vessels were pressurized using the device originally designed for pressure release when using it for conventional acid sample digestion. After, the rotor with the vessels was placed inside the microwave cavity and the selected program for microwave radiation was started. The microwave oven was operated at 1400 W. After finishing the digestion, each vessel was carefully opened to release the pressure. The resultant solution was diluted with water and transferred into a 25-mL polypropylene vessel. After each run, holders were soaked in concentrated HNO3 for 10 min followed by rinsing with water. Residual Carbon. The residual carbon content of the digested samples of milk powder was determined by measuring the carbon emission by ICP-OES at 193.091 nm. The RC is expressed as a percentage of the original carbon content of the solid samples. To avoid interference of dissolved volatile carbon, the digests were degassed before the carbon measurement. Digests were degassed by passing argon through the digest solutions for 20 min according to ref 23. RESULTS AND DISCUSSION Optimization of Operating Conditions for the Microwave Combustion Procedure. Initial experiments were performed to start the combustion in a closed vessel, avoiding the use of focused light or electrical current as currently used in classical oxygen flask technique. The search for a convenient way to ignite the sample was made having in mind some characteristics such as the following: simplicity, minimum reagent consumption and contamination, reducing time for digestion, ease of coupling to a commercial microwave oven used, and, mainly, the possibility of ignition based only in the microwave radiation without the necessity of opening the vessel before the digestion was complete. Several attempts were made using graphite, paper, platinum, and other metals as igniters with quartz vessels pressurized initially with oxygen at 5 bar. However, several problems (sparks, vessel damages) occurred, and the use of these materials was discontinued.24 Based on the use of nitrates as aids for dry ashing procedures, ammonium nitrate was tried as igniter. Although magnesium nitrate has been currently used in these procedures,14 this reagent was changed to NH4NO3 in order to avoid contamination from an excess of magnesium in the digests. Fortunately, in the first attempt using this reagent as an aqueous solution, a complete combustion of the paper was achieved (oxygen pressure of 5 bar). In that experiment, no sample was placed in the quartz holder. After this, a systematic study was performed to optimize the conditions for the proposed microwave combustion procedure and its application to samples with an organic composition. The use of an NH4NO3 aqueous solution added to the paper was convenient in view of availiblity, easy purification, low price, and minimal residue introduction to the final digests that could cause interference during the determination step. The NH4NO3 solutions were prepared daily, and no other additives were tested in this work as igniters. Samples of bovine liver were weighed (23) Gra¨ber, C.; Berndt, H. J. Anal. At. Spectrom. 1999, 14, 683-691. (24) Barin, J. S. Determination of metals and nonmetals in pharmaceutical products after decomposition by microwave-assisted combustion in closed vessels. Dissertation, Universidade Federal de Santa Maria, Santa Maria, Brazil, 2003.
Analytical Chemistry, Vol. 76, No. 13, July 1, 2004
3527
Table 2. Mean and Standard Deviation for the Determination of Cd and Cu (µg g-1, n ) 4) in Certified Reference Samples by the Proposed Procedure sample bovine liver (SRM 1577) bovine liver (SRM 1577b) pig kidney (IRMM CRM 186) skim milk (IRMM CRM 151)
Cd Cu Cd Cu Cd Cu Cd Cu
reference value
without reflux
with HNO3 reflux for 8 min
0.27 ( 0.04 193 ( 10 0.50 ( 0.03 160 ( 8 2.71 ( 0.15 31.9 ( 0.4 0.101 ( 0.008 5.23 ( 0.08
0.28 ( 0.05 194 ( 6 0.49 ( 0.04 158 ( 6 2.85 ( 0.12 32.5 ( 0.7 0.099 ( 0.006 5.35 ( 0.10
0.27 ( 0.03 190 ( 4 0.50 ( 0.02 162 ( 4 2.82 ( 0.10 31.6 ( 0.3 0.098 ( 0.003 5.30 ( 0.12
(50 ( 3 mg) and wrapped with paper that was positioned on the quartz holder. A solution of NH4NO3 was immediately added to the paper, and vessels were capped and pressurized with oxygen. The microwave power was started, and after the end of the microwave program, vessels were depressurized and opened. In these initial experiments, 6 mL of concentrated HNO3 was used as absorbing solution and the oxygen pressure was maintained at 5 bar. Based on the previous experiments, the concentration of the NH4NO3 solution was tested with 1, 3, 6, 9, and 12 mol L-1. A volume of 50 µL was used for these tests. For all tests (n ) 3 for each concentration tested), combustion occurred. However, for the most concentrated solutions (6, 9, and 12 mol L-1), the beginning of combustion was very similar and more reproducible (variation between 2 and 3 s) than other tested solutions where observed ignition times were found to vary from 3 to 6 s. An NH4NO3 concentration of 6 mol L-1 was chosen for subsequent experiments. The beginning of combustion was detected by the high increase of internal pressure of quartz vessels recorded by the pressure sensor of the equipment. For other concentrations, combustion started later and a higher variation of beginning time was observed. That parameter was important to ensure the ignition of the paper and sample with a reproducible behavior. The influence of the volume of NH4NO3 solution was investigated. Volumes varying from 20 to 100 µL of the igniter were tested. For all tests, combustion occurred but a more reproducible behavior concerning the beginning of ignition was observed with 50 µL of 6 mol L-1 NH4NO3 (n ) 4). Under these conditions, the pressure typically increased from 5 to 12 bar and the beginning of combustion occurred between 3 and 5 s after microwave irradiation started. For further studies with 10 and 15 bar of oxygen, combustion was started at 3 s of microwave irradiation and maximum pressure observed was 17 and 22 bar, respectively. As the maximum pressure achieved during the combustion in the previous tests was 22 bar, the initial oxygen pressure of 15 bar was maintained for the subsequent studies. To verify the possible sample amount to be burnt in the proposed system, bovine liver masses of 100, 150, 200, and 250 mg were weighed and vessels were pressurized at 15 bar. Small black residues from incomplete sample combustion were observed in some tests for sample masses of 200 and 250 mg if no reflux step has been applied. For other sample masses, no residues were observed. During the combustion of 250 mg of sample, the maximum pressure achieved was 29 bar. That pressure is ∼36% of the maximum pressure supported by the used quartz vessels. Thus, even with 250 mg of sample, the procedure is far from the 3528 Analytical Chemistry, Vol. 76, No. 13, July 1, 2004
limit conditions for the equipment used so the combustion procedure was safe in the tested conditions used. The necessity of a reflux step, after combustion, was investigated for bovine liver, pig kidney, and skim milk samples. The elements Cd and Cu were determined in digests with and without reflux with concentrated HNO3. Table 2 shows the results obtained from 250 mg of reference samples. Blank were always low, and for both analytes, the agreement to the correspondent certified reference values were between 96 and 105%. Concerning comparison among results for combustion or combustion followed by a reflux step (8 min), it can be observed that for Cd and Cu the combustion alone is sufficient to ensure good agreement to reference values. Residual Carbon. Bovine liver was chosen to demonstrate the extent of the microwave combustion procedure due to the relatively high carbon content of that sample. After combustion of 250 mg of sample, the RC was ∼1.3 ( 0.4% of the original carbon content (for oxygen pressure of 15 bar). It is important because this low RC value was obtained only with combustion without any additional reflux with the absorbing solution. If concentrated HNO3 is used as absorbing (and as refluxing) solution and reflux is applied for 8 min, the RC quickly decreases up to less than 0.4% of the corresponding original carbon content. Probably, if more time for reflux is applied, the RC will be decreased but no additional tests were performed on this way. However, the obtained RC can be considered good when compared with other conventional procedures by wet digestion. These results show the high efficiency that can be achieved using the combustion procedure that takes ∼2 min if no reflux is applied. Comparing the results of the proposed procedure (without the reflux step) with those using conventional PTFE closed vessel, microwave-assisted digestion, it can be observed that the new procedure has a good advantage in relation to the time and efficiency of digestion.9,22 The RC was also evaluated for the digestion of skim milk samples. For this sample, the RC was below 1.1% for the combustion step alone and less than 0.3% if reflux was applied for 8 min with concentrated HNO3. Acid Consumption. The amount of residual acid in each solution after decomposition by microwave combustion followed by a reflux step with concentrated nitric acid was determined by acid-base tritration.25 According to these data (n ) 3), residual acid in the digests varied from 98 to 99%, indicating that concentrated nitric acid was not the main decomposition agent. (25) Arau´jo, G. C. L.; Gonza´lez, M. H.; Ferreira, A. G.; Nogueira, A. R. A.; No´brega, J. A. Spectrochim. Acta 2002, 57 B, 2121-2132.
Based on the results, it can be assumed that digestion was performed almost exclusively by the previous combustion step. This fact is important because it indicates the possibility of applying the proposed procedure using less acid as absorbing solution despite this test not being performed in this work. CONCLUSION The proposed microwave-assisted combustion procedure allows the digestion of bovine liver and skim milk samples in a faster way than conventional wet digestion using microwave for heating. Results for Cd and Cu were satisfactory, taking into account that a reflux step may not be necessary and the digestion can be performed in only a few minutes. The minimum reagent consumption and reduced time for digestion make this procedure promising for digestion of organic samples. The advantages offered by the proposed procedure address the possibility of ignition based only in the microwave radiation without opening the vessel before the completeness of digestion and, additionally, the facility to be (26) Flores, E. M. M.; Barin, J. S.; Paniz, J. N. G.; Medeiros, J. A.; Knapp, G. Verfahren und Vorrichtung zum Verbrennen von Proben in einer sauerstoffhaltigen Atmospha¨re. Patent Reg. 8618755-M/44010, Munich, 2003.
performed in commercial equipment with minimum changes.26 To the knowledge of the authors, no other digestion procedure presents the same performance in the same time and facilities. In view of its simplicity and efficiency, the procedure seems to be very attractive for routine analysis. The proposed microwave combustion technique is still in its infancy. The exact mechanism of ignition has not been identified and the effects of some parameters (e.g., holder design or kind of absorbing solution) were not properly investigated. Obviously, the procedure can be used for other applications in sample preparation. ACKNOWLEDGMENT We thank to Anton Paar GmbH (Graz, Austria) for providing the Multiwave 3000 microwave sample preparation system and Dr. M. Korn by special comments concerning the initial development of the procedure.
Received for review February 10, 2004. Accepted March 30, 2004. AC0497712
Analytical Chemistry, Vol. 76, No. 13, July 1, 2004
3529
