Anal. Chem. 2001, 73, 1515-1520
High-Temperature, Microwave-Assisted UV Digestion: A Promising Sample Preparation Technique for Trace Element Analysis Dieter Florian and Gu 1 nter Knapp*
Institute for Analytical Chemistry, Micro- and Radiochemistry, Graz University of Technology, Technikerstrasse 4, A-8010 Graz, Austria
A novel, microwave-assisted, high-temperature UV digestion procedure was developed for the accelerated decomposition of interfering dissolved organic carbon (DOC) prior to trace element analysis of liquid samples such as, industrial/municipal wastewater, groundwater, and surface water, body fluids, infusions, beverages, and sewage. The technique is based on a closed, pressurized, microwave digestion device. UV irradiation is generated by immersed electrodeless Cd discharge lamps (228 nm) operated by the microwave field in the oven cavity. To enhance excitation efficiency an antenna was fixed on top of the microwave lamp. The established immersion system enables maximum reaction temperatures up to 250-280 °C, resulting in a tremendous increase of mineralization efficiency. Compared to open UV digestion devices, decomposition time is reduced by a factor of 5 and the maximum initial concentration of DOC can be raised by at least a factor of 50. The system’s performance on a realtype sample was evaluated for the mineralization of skimmed milk (IRMM, CRM 151) and subsequent determination of trace elements using standard spectroscopic techniques. Recovery for Cd (109%), Cu (112%), Fe (99%), and Pb (96%) showed good agreement with the 95% confidence interval of the certified values. Trace analytical measurement techniques suffer from a significant decrease in detection power by matrix constituents. The potentially very low detection limits (LODs) can only be achieved when concentrations of dissolved inorganic and organic matrix compounds are kept to a minimum.1 Especially, the presence of dissolved organic carbon (DOC) may cause severe interferences for trace element determination by, for example, DPASV,2-4 ICPMS,1,5,6 and AAS.2,7,8 Thus, complete oxidation of organic matrix * Corresponding author: (fax) ++43/316/8738304; (e-mail) knapp@ analytchem.tu-graz.ac.at. (1) Krachler, M.; Radner, H.; Irgolic, K. Fresenius’ J. Anal. Chem. 1996, 355, 120. (2) Golimowski J.; Golimowska K. Anal. Chim. Acta 1996, 325, 111-133. (3) Kolb M.; Rach P.; Scha¨fer J.; Wild A. Fresenius’ J. Anal. Chem. 1992, 342, 341-349. (4) Wu ¨ rfels M.; Jackwerth E.; Sto¨ppler M. Anal. Chim. Acta 1989, 226 (1), 31-41. (5) Begerov J., Turfeld M., Duneman L. J. Anal. At. Spectrom. 1996, 11, 913916. (6) Begerov J.; Turfeld M.; Duneman L. J. Anal. At. Spectrom. 1997, 12, 10951098. 10.1021/ac001180y CCC: $20.00 Published on Web 03/02/2001
© 2001 American Chemical Society
constituents of biological samples is strongly recommended to achieve accurate and reproducible analysis results.1,2,4,9 Digestion of organic samples routinely is carried out in closed, pressurized devices, either microwave or conventionally heated, applying concentrated mineral acids. A major drawback of this digestion procedure is the introduction of relatively high background levels due to acid impurities. Even at the highest degree of purity commercially available and after additional subboiling distillation, the blank level of acids remains too high for certain ultratrace determinations.5,6 Furthermore, wet decomposition procedures lead to high acid concentrations in the resulting digestion solution, which are not tolerated by many analytical methods, especially by ICPMS, and thus have to be evaporated or diluted prior to analysis. The photochemical process of UV oxidation represents a promising alternative for the decomposition of DOC in liquid samples.2 Only minimum amounts of reagents are required compared to conventional acid digestion procedures. Hence, a potential source of contamination is eliminated. Furthermore, sample dilution is avoided after digestion, resulting in an improvement of the detection limit. Therefore, especially ultratrace determinations in the field of environmental analysis benefit from this contamination-free digestion technique.5,6,10 Typical UV digestion equipment, fundamentals of the degradation mechanism, and common fields of application were reviewed, comprehensively.2 So far, the majority of applications in trace element analysis were performed using so-called batch devices. These open systems work at atmospheric pressure and suffer from certain drawbacks such as limited reaction temperature (65-90 °C) to prevent loss of volatile analytes and solvent evaporation, limited initial concentration of organic compounds, extended degradation time up to several hours for a matrix of 100-300 mg L-1 C, and decreasing decomposition rates for samples of high turbidity (e.g., heavily polluted wastewater). In general, reaction temperature turned out to be the key parameter for UV mineralization. In fact, this phenomenon was observed for the decomposition of vanillic acid as a model substance for polyphenolic compounds in wastewater even within (7) Bauer K. H.; Gehron M.; Wiskamp V. Poster at Tagung der Fachgruppe Wasserchemie, Lindau, Germany, 1997. (8) Schwedt G.; Petri J. Labor Praxis 1992, (Dec), 12-23. (9) Mingorance, M. D.; Perez-Vazquez, M. L.; Lachica, M. J. Anal. At. Spectrom. 1993, 8, 853.
Analytical Chemistry, Vol. 73, No. 7, April 1, 2001 1515
a temperature range of 10-40 °C.11 Other authors reported considerable acceleration for the degradation of stable aromatic systems when temperature was increased from 65 to 90 °C.3 To contribute to the advancement in UV mineralization technology, now the main concern should be focused on the acceleration of the photodegradation process and its application to sample matrixes containing high DOC levels. Hence, the average decomposition temperature has to be increased considerably. This can be achieved easily by application of a closed, pressurized, digestion device. This way, reaction temperature is raised far above the solution’s boiling point without risking solvent evaporation or loss of volatile elements. By applying a microwave-assisted device, additional acceleration of the reaction rate by microwaves can be observed. At present, this effect is under investigation since it is assumed that microwave fields induce higher levels of chemical and physical activity.12 Further improvement of chemical reactivity was discussed recently for a combined microwave/UV process due to simultaneous irradiation by ultraviolet (causing excitation of valence electrons) and microwave (causing molecular rotation and polarization) energy.13,14 However, the effects observed during so-called microwave photochemistry are not investigated by now, since this is definitely a new field of chemistry. In this paper, a novel high-temperature microwave-assisted UV mineralization procedure, using a commercially available pressurized digestion device, is described.15This procedure combines the benefits of microwave digestion with those observed for UV mineralization. The system’s performance was evaluated for the mineralization of model substances and of skimmed milk (certified reference material CRM 151, IRMM) with subsequent determination of trace elements. EXPERIMENTAL SECTION Reagents. Nitric acid (69.5%), potassium hydrogen phthalate (C8H5KO4) pro analysis reagent grade, and hydrogen peroxide (30%). Suprapur reagent grade, were obtained from Merck (Darmstadt, Germany). Nitric acid was purified in-house by subboiling distillation using a quartz still. Standard solutions for Cd, Cu, Fe, Pb, and Sc were prepared from 1000 mg L-1 (2% HNO3) standard solutions (SPEX Plasma Standard, SPEX, Metuchen, NJ). For cleaning purposes, detergent Decon90 was obtained from Zinser Analytik GmbH (Frankfurt, Germany). The certified reference material CRM 151 (skimmed milk) was obtained from the European Commission Joint Research Center, Institute for Reference Materials and Measurements (IRMM) (B-2440 Geel, Belgium). Solutions were made up with distilled, deionized (18.2 MΩ cm-1) water (Barnstead Nanopure, Dubuque, IA). For the production of UV lamps, Hg, Cd, and Zn, pro analysis reagent grade, were obtained from Merck. Additional purification (10) Krachler, M.; Alimonti, A.; Petrucci, F. J. Anal. At. Spectrom. 1998, 13, 701-705. (11) Benitez F. J.; Heredia J. B.; Acero J. L. Toxicol. Environ. Chem. 1996, 56 (1-4), 199-210. (12) Fanslow G. E. Mater. Res. Soc. Symp. Proc. 1991, 189, 43-48 (Microwave Process. Mater. 2). (13) Chemat S.; Aouabed A.; Bartels P. V.; Esveld D. C.; Chemat F. J. Microwave Power Electromagnetic Energy 1999, 34, 55-59. (14) Cirkva V.; Hajek M. Microwave photochemistry in organic synthesis. In Proceeding of the Conference of Microwave and High Frequency Heating, Fermo, Italy, 1997; pp 153-154. (15) Zischka M.; Kettisch P.; Schalk A.; Knapp G. Fresenius’ J. Anal. Chem. 1998, 361, 90-95.
1516
Analytical Chemistry, Vol. 73, No. 7, April 1, 2001
was performed by distillation when transferred into the quartz bulb. Rare gases Ar, Kr, and Xe, 5.0 purity, were obtained from AGA Gas GmbH & Co KG (Hamburg, Germany). Instrumentation. A Multiwave pressurized microwave digestion device, equipped with 50-mL quartz vessels, were obtained from Anton Paar GmbH (Graz, Austria) and Perkin-Elmer (Norwalk, CT). A Cd low-pressure discharge microwave lamp (MWL) was developed in cooperation with Gesellschaft fu¨r lichttechnische Erzeugnisse mbH (G.L.E., Berlin, Germany). It was made from quartz glass,: type, Ilmasil; PN thickness, 1.0 mm; filler elements, 1 mg of Cd + 4 mbar Ar; bulb dimensions, 30 mm hight, 15 mm width, antenna length 35 mm. The emission domain in the UV region is located at 228 nm. The radiation intensity of the MWL during the digestion procedure depends on the microwave energy absorbed and varies from less than 1 W to more than 10 W. The actual microwave power input is regulated automatically by the Multiwave to maintain the preselected digestion temperature and is dependent on sample type and amount as well as on the number of digestion vessels employed. A Shimadzu TOC-5050 total organic carbon analyzer (Shimadzu GmbH Europe, Duisburg, Germany) was used for the determination of the remaining organic carbon content. ICP-AES (Optima 3000XL, Perkin-Elmer, Norwalk, CT) and GFAAS (AAnalyst 800, Perkin-Elmer, Norwalk, CT) were employed for element determination. Digestion Technique. A 7-mL aliquot of sample solution, 1 mL of H2O2, and 50 µL of HNO3 were mixed in the quartz vessel and digested by means of the Multiwave at 72 bar reaction pressure. The microwave energy is reduced automatically when the reaction pressure in one of the vessels exceeds the limit of 72 bar. The preadjusted parameters at the Multiwave were as follows: microwave power, 1000 W; digestion time, 30 min with cooling fan at level 1; cooling time, 15 min with cooling fan at level 3. Sample Preparation. UV mineralization of skimmed milk was performed for five samples and one blank simultaneously. An aliquot of 75 ( 2 mg of powder was transferred to each digestion vessel already containing an UV source. The powder was suspended in 7 mL of H2O (initial DOC concentration, ∼4000 mg L-1 C) and acidified with 50 µL of HNO3 + 50 µL of HCl. Immediately before sealing the vessel, 1 mL of H2O2 was added. Vessels were sealed and heated by following the established microwave program. Upon finishing the digestion step, each vessel was carefully opened in a fume hood at ambient temperature to release the reaction pressure. The resulting sample solution was transferred quantitatively into a 10-mL polyethylene (PE) tube. The certified reference material was not oven-dried prior to digestion to prevent loss of volatile compounds and possible contamination. The element concentrations measured were based on dry weight after correcting for moisture content determined from separate subsamples dried in an oven at 105 °C for 24 h. Cleaning. Each quartz vessel and its corresponding quartz lamp were treated like a unit to avoid any contamination of the UV source. Vessels and lamps were soaked in 5% Decon90 for 48 h followed by careful rinsing with distilled, deionized water. After this treatment, a pressure-controlled digestion was performed with 3 mL of HNO3 for a period of 20 min followed by careful rinsing with distilled, deionized water.
Analysis. Determination of the remaining DOC was carried out with a matrix solution containing 0.5 vol % HNO3. A linear calibration with up to three DOC standards was prepared. For calibration, C8H5KO4 was used as the standard compound. Calibration standards ranging from 0 to 200 mg L-1 were prepared daily by diluting a 1000 mg L-1 stock solution. Samples were diluted according to the calibrated range. Trace element determination was carried out with a matrix solution containing 0.5 vol % HNO3. For AAS measurements, the instrument manufacturer’s recommended temperature programs were applied; linear calibration with up to four element standards were prepared as follows: Cd, 0-5 µg L-1; Cu, 0-20 µg L-1; and Pb, 0-40 µg L-1. Fe was determined by ICP-OES. Linear calibration ranged from 0 to 1000 µg L-1, Sc was used as an internal standard (1000 µg L-1). Samples were diluted according to the calibrated range. All calibrations were verified by measuring control standards periodically. RESULTS AND DISCUSSION Description of the High-Temperature UV Digestion Device. In contrast to an open batch device equipped with one electrically driven UV source situated in the center surrounded by symmetrically distributed sample holders,2 the presented UV digestion device establishes a multiple immersion system using special, microwave-boosted, electrodeless discharge lamps. Single UV sources are directly immersed into the sample solution inside of the individual digestion vessels. Lamp operation is initiated and maintained by the oscillating microwave field distributed within the oven cavity. Immersed UV lamps simply behave like inert quartz bulbs, which continuously emitting UV irradiation. Consequently, this system does not require a separate power supply for the operation of the UV sources. Furthermore, pressurized vessels do no require a separate cooling device to prevent sample evaporation. Here, the microwave device automatically controls reaction temperature and pressure.15 In Figure 1 a schematic is given of the quartz pressure reaction vessel with the polyetheretherketone (PEEK) vessel jacket containing a MWL. The quartz vessel with the MWL and the appearance of the microwaveboosted plasma discharge is shown in Figure 2. The maximum number of samples treated simultaneously during one digestion run depends on the employed rotor system (vessel holding device). In this project, the number was limited to six vessels per run (Figure 3); however a new rotor generation is already available offering space for 12 vessels. Development of a Microwave-Boosted UV Source. The application of an electrodeless discharge lamp (EDL) located directly inside the closed digestion vessel is the only possibility to generate UV radiation inside a high-pressure, microwaveassisted digestion device. However, operation of a standard EDL is seriously impaired inside the applied kitchenlike microwave cavity. First, the observed electromagnetic field is of low homogeneity, since generated microwave energy is simply transferred by two waveguides into the oven cavity and is distributed by multiple reflections. Second, the amount of microwave-absorbing material (e.g., sample solution, number of digestion vessels, rotor system) and its orientation in the cavity additionally affect the electric field strength. Consequently, the electromagnetic field established inside such a cavity is by far less homogeneous and of lower electric field strength compared to standard microwave-
Figure 1. Schematic of a high-pressure digestion vessel with MWL: a, plug; b, safety disk; c, seal; d, screw cap; e, microwave radiation, hν1; f, PEEK vessel jacket; g, quartz pressure reaction vessel; h, electrodeless UV source; i, UV radiation, hν2; j, vessel base plate; k, air flow.
focusing antennas. Finally, the surrounding solution in the immersion system caused serious shielding for the microwave, thus impaired excitation efficiency of the EDL. To circumvent these problems, the development of an appropriate metal antenna was required to focus the microwave field into the EDL, thus to enhance the electric field constant in the lamp environment. A schematic of the established MWL is given in Figure 4. The antenna was made of a molybdenum foil positioned on top of the MWL, which was directly embedded into quartz to avoid contact between the metal and the sample solution. A W wire established the connection between the Mo foil and the discharge volume (quartz bulb). Considering the physical fundamentals of an alternating high-frequency radiation, maximum interaction between the microwave field and the antenna was assumed for an antenna length corresponding to 1/4 λMW. In this case (λMW ) 0.1224 m), the theoretical length was calculated to 30.6 mm and, thus, corresponded well with the obtained experimental result of 30 ( 3 mm for the optimum Mo foil length. Using this antenna construction, MWL operation and thus UV generation was maintained during the entire digestion run for a microwave power level of PMW g 100 W delivered by the magnetron. This means that the microwave energy distributed to the MWL is much lower. Analytical Chemistry, Vol. 73, No. 7, April 1, 2001
1517
Table 1. Selection of MWL Prototypes MWL type
system emission domain filler element fill gas % [DOC]remaining ( SD (n ) 3) bulb dimensions (height/width)
Hg
Cd
Zn
high-pressure metal vapor 254 nm 1.5 mg 4 mbar Ar 1.1 ( 0.9
low-pressure metal vapor 228 nm 0.5 mg 4 mbar Ar (a) 0.13 ( 0.04 (b) 0.03 ( 0.02 (a) 30/15 mm (b) 52/12 mm
low-pressure metal vapor 213 nm (weak) 0.5 mg 4 mbar Ar 5.4 ( 0.3 30/15 mm
30/15 mm
Kr
Xe
gas discharge
gas discharge
