Classical wet ashing versus microwave-assisted attacks for the determination of chromium in plants A. Sahuquillo, R. Rubio* and G. Rauret Departament de Química Analítica, Universitat de Barcelona, Avd. Diagonal 647, E-08028, Barcelona, Spain Received 5th November 1998, Accepted 5th November 1998 Microwave-assisted pre-treatments for the determination of Cr in plants are compared with classical wet-ashing procedures using open attacks in sand-baths. Four certified plant reference materials were analysed: pine needles (NIST SRM 1575), rye grass (BCR CRM 281), beech leaves (BCR CRM 100) and an aquatic plant (Trapa natans) (BCR CRM 596).The use of acidic procedures with HClO4 or H2SO4 yielded different Cr results for these materials when classical wet-ashing procedures were used, as the use of HClO4 caused losses of volatile chromium compounds. The shorter time of analysis required (60 min) in open-focused microwave-assisted attack allows the use of HClO4 for obtaining results very close to the certified values for CRM plant materials.This type of microwave digestion also led to good reproducibility values with relative standard deviations between 5 and 10%. Introduction Chromium is determined routinely to monitor pollution levels in both environmental and biological matrices. For the latter, Cr is usually determined in plants since its concentration level provides information about plant uptake from polluted soils and entry into the trophic chain.The chromium content in these matrices is generally lower than 5 mg kg21, which requires the use of sensitive analytical techniques.1 The technique most widely used for chromium determination is atomic absorption spectroscopy with electrothermal atomisation (ETAAS).2,3 Nevertheless, it is widely accepted that the measurement of chromium in complex matrices by AAS involves serious difficulties.2,4 Other powerful spectroscopic techniques, such as inductively coupled plasma mass spectrometry (ICP-MS), are subject to spectral and non-spectral interferences5,6 and consequently this technique is unsuitable for the certification of chromium in reference materials.4,7 Recent developments in ETAAS include different sample introduction procedures such as the preparation of slurries.However, this system presents homogeneity problems for sample intake at the milligram level, especially for field samples, and therefore sample pre-treatment is mandatory to obtain good reproducibility.8,9 With regard to the attack of sample materials for chromium determination, the use of classical open attacks can presumably lead to losses of volatile chromyl chloride or organic chromium compounds, due to the reduction of perchlorate anion to chloride and due to the oxidation of organic substances.10,11 Sample digestion is usually the most time-consuming step of the analysis.With the aim of shortening the time and the total volume of reagents, and to avoid contamination in the digestion process, the use of microwave sample pre-treatment is increasing, as shown by the large number of reviews dealing with the principles.6,10,12–14 The main microwave-assisted digestions described for chromium determination in plants consist of microwave PTFE closed-vessel attacks using different acidic mixtures for short periods of time.15–17 In this work, different digestion procedures for the determination of Cr in plants were studied with the aim of establishing a suitable method that ensures no losses of volatile chromium compounds, has the ability to use as little reagent as possible and is rapid.For this purpose, classical open wet ashing procedures with sand-baths were compared with open focused microwave-assisted pre-treatments optimised for the determination of chromium in plants. In both cases, HClO4 and H2SO4 were tested for sample mineralisation. The instrumental conditions for final determination by ETAAS with Zeeman effect background correction (ZETAAS) were optimised for the different extracts obtained after sample pre-treatment.Experimental Apparatus A Pselecta Recisplac sand-bath (Afora, Barcelona, Spain) was used for sample pre-treatment when heating by conduction. A Microdigest A301 open-focused microwave digestor (Prolabo, Paris, France) was used. The magnetron worked at a maximum power of 200 W (100%) and it could be regulated from 10 to 100% in steps of 5%.A Perkin-Elmer (Norwalk, CT, USA) Model 4100ZL atomic absorption spectrometer with longitudinal Zeeman effect background correction and transversal heating, equipped with an automated autosampler able to perform standard additions, was used for Cr determination. Pyrolytic graphite-coated tubes with a pre-built L’vov platform were used. Reagents All solutions were prepared using doubly de-ionized water (Culligan Ultrapure GS, 18.3 M½ cm21) in a class 100 work bench with vertical air flow, in accordance with the USA Federal Standard 209b/d norm.18 The class of the laboratory was checked annually.19 All the concentrated acids used for the attacks were of Suprapur quality (Merck, Darmstadt, Germany).Chromium(vi) standard solutions were prepared from certified National Institute of Standards and Technology (NIST) potassium dichromate of 99.984 ± 0.010% purity. The working calibrant solutions were prepared in 0.3 mol l21 HNO3. Analyst, 1999, 124, 1–4 1Otherwise, Cr(iii) commercially available calibrant solution was used and checked against K2Cr2O7 standard solutions.As chemical modifiers, 0.02 mol l21 Mg(NO3)2 and 0.01 mol l21 Pd of Suprapur quality were used (Merck). Samples Four certified plant reference materials were used: pine needles (NIST SRM 1575), rye grass [Bureau Community of Reference (BCR) CRM 281], beech leaves (BCR CRM 100) and an aquatic plant (Trapa natans) (BCR CRM 596). The certified chromium contents in these materials were 2.6 ± 0.2, 2.14 ± 0.12, 8.0 ± 0.6 and 36.3 ± 1.7 mg kg21, respectively, on dry mass.Further information on these materials is available in the certification reports.4,7,20 Sample preparation HClO4 digestions by conduction. A 5 ml volume of 14 mol l21 HNO3 was added to 0.5 g of sample in a PTFE beaker. The beakers were then placed in a sand-bath and heated to between 120 and 150 °C. Subsequent aliquots of HF–HClO4 (2 + 1) were added until the remaining residue did not have a siliceous aspect.A total volume of 33 ml of the mixture was added in four steps. A total volume of 6 ml of 8.8 mol l21 H2O2 was added to the sample in three steps for complete mineralisation. Finally, 2 ml of HClO4–H2O2 (1 + 2) were added and the sample was evaporated to dryness. The residue was dissolved in dilute HNO3 and diluted to 50 ml with doubly deionized water. The whole sample pre-treatment process lasted between 40 and 56 h.H2SO4 digestions by conduction. The above procedure was also used but HClO4 was replaced with H2SO4. At the end of the digestion, a residue of insoluble sulfates was obtained. The suspension was filtered through an ashless Whatman (Maidstone, Kent, UK) No. 42 filter-paper. The residue was washed carefully with doubly de-ionized water and the filtrate was diluted to 50 ml. When using H2SO4, the total time of attack increased to 90 h. Microwave-assisted digestion procedures.Table 1 gives the optimised programmes for the microwave digester used. As for the open attacks in sand-bath, the final solution was filtered through an ashless Whatman No. 42 filter-paper to ensure the separation of any fine particles, and the filtrate was diluted to 50 ml with doubly de-ionized water. Measuring conditions ZETAAS determination was performed at 357.9 nm with a slitwidth of 0.2 nm. Peak areas were considered for signal treatment. The standard additions method was used for calibration. A 5 ml volume of sample, diluted to a final volume of 20 ml with blank and Cr(vi) standard solution, was injected.The working linear concentration range was from 2 to 30 mg l21. Table 2 gives the optimised programme of temperatures and times, which was established for each matrix studied. All measurements were performed in duplicate. Results and discussion Establishment of graphite furnace heating programmes The acidic solutions obtained from each plant material were used for the optimisation of the final ZETAAS instrumental conditions. Each programme of temperature and times was established from the construction of pyrolysis and atomisation curves.The pyrolysis curve was obtained at a fixed drying and atomisation temperature by increasing the temperature of pyrolysis from 900 to 1800 °C in steps of 100 °C. For the atomisation curve, the drying and pyrolysis temperatures were fixed and the atomisation temperature was increased from 1900 to 2600 °C in steps of 100 °C.At each temperature, 20 ml of plant extracts and 20 mg l21 Cr(vi) calibrant solution were injected in duplicate. Fig. 1 shows the curves obtained for each of the materials studied. The criterion for choosing the working temperatures was a maximum peak area with a good peak shape, that is, non-tailing and reproducible peaks. Table 2 gives the final optimised programmes obtained from the pyrolysis and atomisation curves.No significant differences were observed between the final temperatures obtained for standard solutions in 0.3 mol l21 HNO3 and those established for plant extracts. Table 1 Microwave digestion procedures for plants Programme A (HClO4) Programme B (H2SO4) 5 ml HNO3; 30 W (10 min); 40 W (5 min) 5 ml HNO3; 30 W (10 min); 40 W (5 min) 5 ml HF–HClO4 (2 + 1); 30 W (10 min); 50 W (5 min); 60 W (15 min)a 5 ml HF–H2SO4 (2 + 1); 30 W (10 min); 50 W (5 min); 80 W (15 min)a 10 ml HNO3 (1 + 10); 20 W (5 min) 5 ml HNO3 (1 + 10); 20 W (5 min); 40 W (5 min)a 2 ml H2O2; 20 W (5 min); 40 W (10 min)a 10 ml HNO3 (1 + 10); 20 W (5 min) Total time: 50 min Total time: 75 min a Evaporation to dryness.Table 2 Graphite furnace programme for plant materials Step T/°C tramp/s thold/s Drying 1 110 10 20 Drying 2 130 15 30 Ashing 1400 1300a 10 20 Atomisationb 2200 0 4 Cleaning 2400 1 2 a Ashing temperature for beech leaves. b Stopped flow during atomisation. Fig. 1 Pyrolysis and atomisation curves for plant materials and Cr(vi) calibrant solution. 2 Analyst, 1999, 124, 1–4The ashing temperature was set to 1300 °C for beech leaves since slightly higher peak area values were obtained. The extracts resulting from HClO4 and H2SO4 digestions yielded the same pyrolysis and atomisation conditions. Atomisation profiles and analytical performance Two calibration methods were used: linear calibration graph and standard additions. Chromium(vi) and (iii) solutions were tested as calibrants.The results obtained using both types of calibrant solution and calibration method agreed closely, as also observed in previous studies with different matrices.3 Since slightly better reproducibility was obtained with the standard additions method, the final Cr determination was carried out with this method and using Cr(vi) as calibrant solution. Atomisation profiles and background correction signals were studied by adding some of the more commonly used chemical modifiers mentioned in the literature for chromium, such as Mg(NO3)2 21,22 and Pd.22 The characteristic masses obtained for rye grass with and without addition of Mg(NO3)2 were 7.10 ± 0.05 and 7.53 ± 0.18 pg Cr, respectively, with similar atomisation peak profiles.On the other hand, the addition of 0.01 mol l21 Pd increased the background signal significantly without improving either the atomic signal or the peak profile. Therefore, no chemical modifier was used for Cr quantification in these materials.Classical digestion procedures Two series of open attacks in a sand-bath examined, one using concentrated HClO4 as oxidant acid and the other using H2SO4 and maintaining the same acid volume-to-sample mass ratio. Addition of HF was necessary to remove the siliceous structural component of the plants. The initial volumes tested were 5 ml of each of the acids and the final volumes were established according to the aspect of the residue obtained.Complete mineralisation was obtained only after the addition of H2O2. Five replicates for each plant material were digested in different working sessions. Some differences were observed between extracts resulting from HClO4 and H2SO4 attacks. Whereas when HClO4 was used clear, colourless solutions were obtained, the use of H2SO4 led to a black, insoluble residue which required a filtration step and the final solutions were slightly coloured. Owing to the difficulty in removing H2SO4, the final time of attack was increased to between 30 and 50 h.Table 3 gives the results obtained by applying the two digestion procedures. The mean value and the standard deviation for five determinations are compared with the certified values. For Cr determination in pine needles, both attacks yielded concentration values that agreed closely with the certified value. However, a negative bias was observed for the other three materials when HClO4 was used.The losses of analyte detected were about 35%, with reproducibility ranging between 11 and 19% (RSD), showing different behaviour depending on the type of matrix. When samples were digested with conduction heating only H2SO4 gave good results. The main disadvantage of this acidic digestion was that a different procedure would have to be followed if other heavy metals such as Pb are to be determined. Microwave-assisted attacks With the aim of shortening the time of analysis and studying the behaviour of chromium when using other heating systems such as microwaves, two different programmes were optimised for sample pre-treatment by using HClO4 or H2SO4, both combined with HF.The same acidic mixtures as used for sand-bath procedures were tested in different steps. Before the addition of a new acidic mixture the aspect of the remaining residue was monitored until a crystalline residue was obtained. The microwave power was increased in subsequent heating processes. Microwave powers up to 80 W were only necessary in drying steps with H2SO4 owing to its high boiling-point.As can be seen in Table 1, 50 min were required when using the HClO4 and 75 min when using the H2SO4 procedure. In the same way as for the open attacks, the extracts were filtered and diluted to 50 ml with doubly de-ionised water. Although the microwave digester allowed the possibility of evaporating the sample to dryness, several drops of condensed HF remained on the walls of the PTFE tubes.Therefore, 10 ml of saturated H3BO3 were added to the digestion vessels in order to eliminate any remaining HF. The addition of H3BO3 did not have any effect on the heating temperature programmes in the graphite furnace. Table 4 gives the results obtained for pine needles, rye grass and beech leaves. The results obtained using the HClO4 and H2SO4 procedures were consistent with the certified values (recoveries ranged from 92 to 108%) for all matrices except rye grass when HClO4 was used (86% recovery). Since the digestion time when using microwaves was reduced by a factor of about 60, no significant losses of chromium occurred.The RSD obtained for HClO4 were lower than 5% for triplicate analyses and slightly higher values were obtained for H2SO4 attack (4–13%). Conclusions The use of HClO4 during the sample pre-treatment step in classical open attacks with conduction heating (sand-bath) led to losses of volatile chromium compounds for some of the plant Table 3 Comparison of open attacks with HClO4 and H2SO4 (conduction) (mg kg21 Cr, mean ± s, n = 5) Sample Certified value RSD% HClO4 attack RSD% H2SO4 attack RSD% Pine needles 2.6 ± 0.2 7.7 2.15 ± 0.26 12 2.45 ± 0.14 5.7 Rye grass 2.14 ± 0.12 5.6 1.48 ± 0.28 19 2.33 ± 0.31 13 Beech leaves 8.0 ± 0.6 7.5 5.12 ± 0.80 16 7.27 ± 0.44 6.1 Aquatic plant 36.3 ± 1.7 4.7 23.08 ± 2.63 11 32.01 ± 4.66 14 Table 4 Comparison of open attacks with HClO4 and H2SO4 (microwaves) (mg kg21 Cr, mean ± s, n = 3) Sample Certified value RSD% HClO4 attack RSD% H2SO4 attack RSD% Pine needles 2.6 ± 0.2 7.7 2.40 ± 0.07 2.9 2.45 ± 0.14 5.7 Rye grass 2.14 ± 0.12 5.6 1.84 ± 0.05 2.7 2.33 ± 0.31 13 Beech leaves 8.0 ± 0.6 7.5 8.12 ± 0.80 9.8 8.07 ± 0.80 9.9 Analyst, 1999, 124, 1–4 3materials studied. From this it can be concluded that losses of chromium are matrix dependent and therefore validation of the method is required for samples of different origins. When classical open attack is the only possibility, the use of H2SO4 is mandatory in spite of the longer time of analysis required.The open-focused microwave heating system is suitable for Cr determination in plants, even when HClO4 is used in the acidic mixtures. Moreover, the time of analysis and the volume of reagents can be reduced by factors of 60 and 6, respectively. Hence it can be considered a cleaner method than the open attack method.The proposed method provides good reproducibility. References 1 J. W. Moore, in Inorganic Contaminants of Surface Water. Research and Monitoring Priorities, ed. R. S. De Santo, Springer, New York, 1991, ch. 9, pp. 82–97. 2 R. Rubio, A. Sahuquillo, G. Rauret and Ph. Quevauviller, Int. J. Environ. Anal. Chem., 1992, 47, 99. 3 A. Sahuquillo, J. F. López-Sánchez, R. Rubio and G. Rauret, Mikrochim. Acta, 1995, 119, 251. 4 Ph. Quevauviller, J. G. Van Raaphorst and H. Muntau, Trends Anal.Chem., 1996, 15, 259. 5 S. Wu, X. Feng and A. Wittmeier, J. Anal. At. Spectrom., 1997, 12, 797. 6 F. E. Smith, Talanta, 1996, 43, 1207. 7 J. G. Van Raaphorst, Ph. Quevauviller and H. Muntau., The Certification of the Mass Fraction of Chromium in Rye Grass (CRM 281), Beech Leaves (CRM 100), Aquatic Plant (Trapa Natans) (CRM 596), Fly Ash (CRM 038) and Sewage Sludge (CRM 597), EUR 16840 EN Report, European Commission BCR Information. Reference Materials, Brussels, 1996. 8 R. E. Sturgeon, Spectrochim. Acta, Part B, 1997, 52, 1451. 9 A. Carlosena, M. Gallego and M. Valcárcel, J. Anal. At. Spectrom., 1997, 12, 479. 10 I. Kubrakova, Spectrochim. Acta, Part B, 1997, 52, 1469. 11 M. J. T. Carrondo, R. Perry and J. N. Lester, Anal. Chim. Acta, 1979, 106, 309. 12 M. Burguera and J. L. Burguera, Quim. Anal., 1996, 15, 112. 13 R. Chakraborty, A. K. Das, M. L. Cervera and M. de la Guardia, Fresenius’ J. Anal. Chem., 1996, 355, 99. 14 A. Sinquin, T. Görner and E. Dellacherie, Analusis, 1993, 21, 1. 15 D. H. Sun, J. K. Waters and T. P. Mawhinney, J. AOAC Int., 1997, 80, 647. 16 I. V. Kubrakova, T. F. Kudinova, E. B. Stavnivenko and N. M. Kuz’min, J. Anal.Chem., 1997, 52, 522. 17 R. Chakraborty, A. K. Das, M. L. Cervera and M. de la Guardia, J. Anal. At. Spectrom., 1995, 10, 353. 18 J. R. Moody, Anal. Chem., 1982, 54, 1358A. 19 Report of Verification and Control of Clean Laboratories, Sociedad de Validación de Sistemas (SVS), Sant Cugat, Barcelona, 1998. 20 Certificate of Analysis of SRM 1575 (Pine Needles), National Institute of Standards and Technology, Gaithersburg, MD, 1993. 21 M. Hoenig and A. M. de Kersabiec, L’Atomisation � Electrothermique en Spectrométrie d’Absorption Atomique, Masson, Paris, 1989, pp. 137–165. 22 A. Sahuquillo, R. Rubio and G. Rauret, in Quality Assurance for Environmental Analysis, ed. Ph. Quevauviller, E. A. Maier and B. Griepink, Elsevier, Amsterdam, 1995, pp. 39–62. Paper 8/08659A 4 Analyst, 1999, 124, 1&ndash