- Email: [email protected]
Comparison of ultrasound-assisted extraction and microwave-assisted digestion for determination of magnesium, manganese and zinc in plant samples by flame atomic absorption spectrometry
Talanta 53 (2000) 433 – 441 www.elsevier.com/locate/talanta
Comparison of ultrasound-assisted extraction and microwave-assisted digestion for determination of magnesium, manganese and zinc in plant samples by flame atomic absorption spectrometry A.V. Filgueiras, J.L. Capelo, I. Lavilla, C. Bendicho * Departamento de Quı´mica Analı´tica y Alimentaria, Uni6ersidad de Vigo, Facultad de Ciencias (Quı´mica), As Lagoas-Marcosende s/n. 36200 Vigo, Spain Received 21 April 2000; received in revised form 6 July 2000; accepted 6 July 2000
Abstract In this paper, a sample preparation method based on acid extraction of magnesium, manganese and zinc from plant tissue by means of high intensity probe ultrasonication is described. Acid extracts obtained upon sonication were directly nebulised into an air–acetylene flame for fast metal determination by atomic absorption spectrometry. Parameters influencing extraction such as sonication time, ultrasound amplitude, sample mass, particle size, extractant composition and volume were fully optimised. Optimum conditions for metal extraction were as follows: a 3-min sonication time, a 30% ultrasonic amplitude, a 0.1-g sample mass, a particle size less than 50 mm, a 0.3% m/v HCl concentration in the extractant solution and a 5-ml extractant volume. Six plant samples used in the human diet were analysed, the concentration range of the three metals approximately being in the range of 1500 – 3000 mg g − 1 for Mg, 30–735 mg g − 1 for Mn and 20–45 mg g − 1 for Zn. Limits of detection corresponding to the ultrasound-assisted extraction method were 0.10, 1.26 and 0.65 mg g − 1 for Mg, Mn and Zn, respectively. Between-batch precision, expressed as R.S.D., was about 0.5, 1.5 and 1% for Mg, Mn and Zn, respectively. Analytical results for the three metals by ultrasound-assisted extraction and microwave-assisted digestion showed a good agreement, thus indicating the possibility of using mild conditions for sample preparation instead of intensive treatments inherent with the digestion method. The advantages and drawbacks of ultrasound-assisted extraction in respect to the microwave-assisted digestion are discussed. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Ultrasonic extraction; Plant analysis; Magnesium; Manganese; Zinc; Flame atomic absorption spectrometry
1. Introduction * Corresponding author. Tel.: +34-986-812281; fax: + 34986-812382. E-mail address: [email protected] (C. Bendicho).
Metal analysis of plants is an essential feature of environmental, biological and chemical research. Metals in plants display biological activity
0039-9140/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 0 0 ) 0 0 5 1 0 - 5
434
A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
as essential or toxic agents, hence being important to establish their normal concentration range and evaluate their role as part of the food chain. On the other hand, contamination processes of medicinal, aromatic and seasoning plants from heavy metals, although representing a small part of the diet, is part of their quality control [1]. Atomic absorption spectrometry is the likeliest technique used for metal determination in these kind of samples, combined with wet or dry ashing procedures for sample preparation [2]. An alternative to intensive sample preparations, which usually involve high temperatures and pressures, is acid extraction under mild conditions [3]. Solid– liquid metal extraction can be enhanced with the use of ultrasound irradiation [4,5]. Ultrasound energy causes its chemical effects through the phenomenon of cavitation, which consists of the production of microbubbles in a liquid when a large negative pressure is applied to it [6]. When a suspension of solid particles is irradiated by ultrasound, particle disruption can takes place, which in turn, causes an increase in the surface area available for reaction. When extraction applications of ultrasound are concerned, particle fragmentation can enhance the ability of the extractant to leach metals. Furthermore, leaching is facilitated by the generation of reactive species, e.g. radicals, as a result of the ultrasonic irradiation. Thus, when water is sonicated H and OH radicals are formed, which in turn result in the formation of oxygen gas and hydrogen peroxide [7]. Metals may be encapsulated within cell walls that must be broken down by the combined effects of dilute acid attack and ultrasonication in order to bring metals into the liquid extractant. Several workers have successfully used ultrasonic extraction of metals from biological samples, quantitative extraction being usually obtained with high intensity probe sonication. Thus, quantitative iron extraction from vegetal matter using an ethanolic solution containing hydrochloric acid has been reported [8]. Calcium, magnesium, potassium, phosphorus, sodium, iron, zinc, copper, and manganese have been extracted with 1 M hydrochloric acid in an ultrasonic bath [9]. Solid– liquid extraction of metals from powdered biolog-
ical samples have been reported with 5 ml of 1 M nitric acid in combination with an ultrasonic cleaner [10]. Cadmium, copper and lead have been extracted with 1.5 ml of 3% v/v nitric acid solution from several biological materials by means of an high intensity probe sonicator [11–13]. Ultrasound-assisted extraction of metals from plant tissue, although not yet sufficiently exploited, could be an attractive alternative to microwave-assisted digestion, thus avoiding concentrated acids, pressure reactions and time consumption involved, since apart from the time required for digestion, cooling of the reactors needs to be accomplished before opening [14]. In this work, parameters influencing ultrasound-assisted extraction such as ultrasound amplitude, sonication time, sample mass, particle size and extractant (volume and composition) are fully investigated. Metal determination in the extracts is carried out by flame atomic absorption spectrometry (FAAS), the results being compared with those obtained by microwave-assisted digestion.
2. Experimental
2.1. Apparatus 2.1.1. Atomic absorption spectrophotometer A Perkin–Elmer atomic absorption spectrophotometer model 2380 (Norwalk, CT, USA) equipped with 10-cm burner head was used for metal determination. A Cathodeon (Cambridge, UK) hollow cathode lamp of Ca and Mg was used as radiation source. The elements were measured under optimised operating conditions by FAAS with an air–acetylene flame. The instrumental parameters are shown in Table 1. 2.1.2. Agate ball mixer mill A MM 2000 Retsch mixer mill (Haan, Germany) was used for grinding the plant samples. Sieves made of nylon with mesh sizes of B25, B 50, B 100, B150 mm were used to study the influence of particle size on extraction.
A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
2.1.3. High intensity ultrasonic processor A 100 W, 20 kHz VC-100 Sonics and Materials (Danbury, CT, USA) high intensity probe sonicator equipped with a Ti – Al – V microtip (6 mm i.d.) was used for ultrasound-assisted extraction. The amplitude control of the ultrasonic processor allowed the ultrasonic vibrations at the probe microtip to be set at any desired level in the 10–100% range of the nominal power. An Alresa (Barcelona, Spain) centrifuge was used for rapid separation of the extracts. 2.1.4. Microwa6e digestion system A MDS-2000 Microwave Sample Preparation System (CEM, Matthews, NC, USA) was used for microwave-assisted digestion of plant samples. The digestions were carried out in CEM advanced composite vessels. 2.2. Reagents The concentrated acids, nitric acid, hydrochloric acid and hydrofluoric acid were of analytical reagent grade (Carlo Erba, Milano, Italy). Deionised distilled water was used throughout the work. All glassware and plasticware was washed with 5% m/v nitric acid and rinsed with deionised distilled water. Stock standard solutions of manganese and zinc containing 1000 mg l − 1 of each element were obtained by dissolving 0.25 g of the pure metal (Carlo Erba) in the minimum amount of (1+ 1) v/v nitric acid and (1+1) hydrochloric acid (Carlo Erba), respectively, and diluting to 250 ml. A stock standard solution of magnesium was obtained by dissolving magnesium chips (Aldrich, Milwaukee, WI, USA) in 1% v/v hydrochloric acid. Calibration standards of each element were obtained by appropriately diluting the stock solutions.
2.3. Samples The plant samples analysed were cinnamon, nutmeg, parsley, sweet paprika, hot paprika and black tea. All them were obtained from the market in the form of powder or leave. To reduce particle size they were ground in the agate ball mixer mill using a power between 20 and 80% and a grinding time in the range 2–5 min. The powdered samples were sieved so that the required particle size was obtained. Plant samples were dried for 24 h at 70°C as recommended elsewhere [15,16]. Once the samples were powdered and dried, they were kept in labelled capped glass flasks [17] inside a desiccator.
2.4. Ultrasound-assisted extraction procedure A portion (0.1 g) of sample was weighed into polyethylene centrifuge tubes (50 ml capacity) and 5 ml of 0.3% m/v hydrochloric acid were added. Then, the sample was sonicated for 3 min at 30% ultrasound amplitude. After sonication, the supernatant liquid was separated from the solid phase by centrifugation for 4 min at 4500 rpm. Determination of magnesium, manganese and zinc was carried out in the supernatant. Blanks were treated in the same way.
2.5. Microwa6e-assisted digestion procedure The microwave-assisted digestion was optimised elsewhere [14], and can be summarised as follows: about 0.1 g of sample were weighed into the Teflon vessel, and then, a mixture of 5 ml 69.5% m/m nitric acid and 0.5 ml 48% m/m hydrofluoric acid was added. The digestion vessel was closed and heated in the CEM microwave oven for a preselected program (i.e. two stages of
Table 1 Instrumental parameters for determination of Mg, Mn and Zn by flame atomic absorption spectrometry Element
Wavelength (nm)
Mg Mn Zn
285.2 279.5 213.9
Lamp current (mA) 4 15 15
435
Bandpass (nm)
Air/acetylene flow-rate (l min−1)
0.7 0.2 0.7
11/1 11/1 11/1
A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
436
Table 2 Operating conditions for extraction of Mg, Mn and Zn from plant samples Variable
Sonication amplitude (%) Sample amount (g) Particle size (mm) % m/v HCl (extractant) Extractant volume (ml)
Studied interval
10–60 0.05–0.2 25–150 0–5 3–9
Optimum extraction conditions 30 0.10 B50 0.3 5
1 min at 40 and 80 psi, and a final stage of 5 min at 120 psi). Once the digestion program was finished, the reactor was cooled in an ice bath before opening. The contents of each vessel were heated to dryness and dissolved with 1 ml of 37% m/m hydrochloric acid. The solution was quantitatively transferred into a 5-ml volumetric flask and made up to volume with deionised distilled water. The blanks were treated in the same way. Both procedures were validated against CRM GBW07605 (GSV-4) tea leaves (National Research Center for Certified Reference Materials, China), which was previously ground to pass through a nylon sieve (50 mm mesh size).
Variables influencing the extraction process were optimised within the intervals shown in Table 2. Each result was the average value of three determination performed in separate batches.
3.2. Influence of sonication time Temperature of the extraction medium increases with increasing sonication time. Usually, as the temperature increases up to 50°C, extraction efficiency is increased as a result of the larger number of cavitation nucleus formed [18,19]. As the temperature approaches the boiling point of the liquid, ineffective sonication occurs as a result of the decrease in surface tension and increase in vapour pressure within the microbubble, which in turn cause the damping of the shock wave [20]. The influence of sonication time on metal extraction is shown in Fig. 1(a). For the three metals, extraction efficiency increased with increasing sonication time from 1 to 2 min. Extraction efficiency was slightly worse (i.e. 95% for Mg) when the temperature of the medium exceeded 50–60°C, which occurred after a 3 min sonication time. A sonication time of 3 min was seen to be suitable for metal extraction.
2.6. Analytical determinations
3.3. Influence of sonication amplitude
Three sub-samples of each plant sample were used for analytical determinations with the digestion and extraction procedures. With each series of digestions or extractions a blank was measured. All measurements were run in triplicate for the sample solutions and standard solutions. Calculations of metal contents in samples are based on a calibration graph obtained from aqueous standards.
Intensity of ultrasound transmitted to the medium is directly related to the vibration amplitude of the probe. Usually, an increase in intensity will provide for an increase in the sonochemical effects. However, at high vibrational amplitude a great number of cavitation bubbles are generated in the solution, which may dampen the passage of sound energy through the liquid [6]. The influence of ultrasound amplitude in the range 10–60% is shown in Fig. 1(b). The efficiency of the process was maximum for an amplitude of 20%. For the three metals, extraction efficiency increased with increasing amplitude from 10 to 20% and remained constant for higher amplitude values. Results pointed out that the optimum extraction efficiency was attained in the range of 20–60% amplitude.
3. Results and discussion
3.1. Optimisation of the ultrasound-assisted extraction method Cinnamon was used for optimisation purposes.
A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
3.4. Sample amount Sample amount used largely depends on the procedure followed. In this work, the 0.05–0.20 g
437
mass interval was investigated. A sample mass of up to 0.2 g [7] has been used for extraction with ultrasonic probe equipment. A sample mass of up to 0.5 g has been reported in the work with an
Fig. 1. Optimisation of the main parameters influencing ultrasound-assisted extraction (, Mg; , Mn; , Zn). (a) Time (min); (b) Amplitude (%); (c) Sample amount (g); (d) Particle size (mm); (e) Hydrochloric acid concentration (%, m/v); (f) Nitric acid concentration (%, m/v).
438
A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
ultrasonic bath for extraction [21]. Solid–liquid extraction of Cd and Pb from biological materials has shown that a sample mass above 10–20 mg cause the extraction efficiency to decrease [11–13] when using typical autosampler cups for electrothermal – atomic absorption spectrometry. As can be seen in Fig. 1(c), a significant decrease in metal recovery from plant tissue is obtained when the sample mass is larger than 0.10–0.15 g.
3.5. Particle size Particle size was among the more critical parameters influencing ultrasound-assisted extraction. As was to be expected, reactions were enhanced on increasing the contact surface. However, from a certain particle size, a further decrease in particle size was not accompanied by an increase in sonochemical effects [6,18]. Only a few studies have included particle size among the experimental conditions to be optimised for ultrasound-assisted extraction of metals. Thus, some workers [21] employed a particle size less than 50 mm. Solid–liquid extraction of Pb from biological materials showed that a particle size larger than 150 mm caused a decrease in extraction efficiency [12]. The particle size attempted for metal extraction in this work with the use of ultrasonic probes ranged from less than 25 up to 150 mm. The results obtained in this study are shown in Fig. 1(d). As can be observed, extraction efficiency decreased when the particle size was larger than 50 mm for the three metals. When the particle size fraction of 50–100 mm was used for extraction a decrease in extraction efficiency of about 9, 11 and 20 and 16% was observed for Mg and Zn, respectively, in comparison with the 25 – 50-mm fraction. For further experiments, the fraction with particle size less than 50 mm was chosen, since the fraction with particle size less than 25 mm was more difficult to achieve by sieving.
3.6. Composition of acid extractant Acid concentration in the liquid extractant was seen to be the most critical parameter affecting
ultrasound-assisted extraction. Extraction efficiency was about 10% for Mn and Zn in the absence of acidity in the liquid extractant. Different acid mixtures such as 15% m/m nitric acid+ 1% m/m hydrochloric acid [22], 1 M hydrochloric acid+ 20% v/v ethanol [8], 1 M hydrochloric acid [21] and 1 M nitric acid [10] were reported to be used as extractants for solid–liquid extraction of metals. In this work, hydrochloric acid, nitric acid and a mixture of hydrochloric acid+ nitric acid were attempted for ultrasound-assisted extraction. The single acids were employed in the range 0–5% m/v. The acid mixtures used as extractants were 0.3% m/v hydrochloric acid+ 0–1% m/v nitric acid and 0.3% m/v nitric acid + 0–1% m/v hydrochloric acid. Blanks obtained from the different extractants were lower than the limit of detection (LOD) for the metals studied. Extraction results obtained with the use of a single acid (i.e. hydrochloric acid or nitric acid) are shown in Fig. 1(e) and (f). Extraction efficiency increases with increasing acid concentration, a steady extraction efficiency being reached for an acid concentration higher than 0.2% m/v. Likewise, the use of an acid mixture where at least one component is at a concentration higher than 0.2% can be used for extraction. The three metals studied display a similar extraction behaviour with all acids and their mixtures. A 0.3% m/v hydrochloric acid concentration was chosen for extraction since it has not oxidant properties, being more convenient than nitric acid for sample introduction by pneumatic nebulisation into the flame.
3.7. Extractant 6olume Variable volumes of extractant such as 50 ml [21] or 5 ml [10,21], have been employed for solid–liquid extraction of metals. In this work, the extractant volume was studied in the range of 3–9 ml. The influence of extractant volume on extraction efficiency was negligible for the three metals studied. An extractant volume of 5 ml was chose so that the required number of replicates could be performed without exhaustion of the sample solution.
A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
439
Table 3 Validation of the ultrasound-assisted extraction (UAE) and microwave-assisted digestion (MAD) methods against CRM GBW 07605 tea leaves Element
Certified value (X9 ts/ N)a
MAD value (X 9 ts/ N)b
UAE value (X9 ts/ N)b
Mga Mn Zn
0.1790.01 1240940 26.390.9
0.16 90.01 1256 9 17 26.8 90.6
0.15 9 0.02 1214 925 25.2 9 0.5
a b
Magnesium concentration was expressed as mg g−1. Average value 9 confidence interval (P= 0.05; N =5) (mg g−1).
3.8. Analytical results using microwa6e-assisted digestion and ultrasound-assisted extraction The detection and quantification limits were defined as 3s/m and 10s/m, respectively. Both values were calculated from the S.D. (s) of ten blank measurements and the slope of calibration graph (m). Limits of detection (LODs) for the ultrasound-assisted extraction method were 0.10, 1.26 and 0.65 mg g − 1 for Mg, Mn and Zn, respectively, being similar to those attained with microwave-assisted digestion when a 0.1-g sample mass was used for digestion. Limits of quantification (LOQs) were 0.4, 4.2 and 2.2 mg g − 1 for Mg, Mn and Zn, respectively. The equations for the linear range of the calibration graphs were as follows: Abs = 0.6768[Mg],
r= 0.9998
Abs =0.0794[Mn] +0.0002, Abs =0.2012[Zn] −0.0002,
r = 0.9999 r =0.9999
Where the concentration was expressed in each case as mg l − 1. Validation of the ultrasoundassisted extraction and microwave-assisted digestion is shown in Table 3. A good agreement between the found and certified metal contents can be observed for the three metals studied. Analytical results obtained by microwave-assisted digestion and ultrasound-assisted extraction corresponding to the plants analysed are shown in Table 4. Analytical results were expressed as average concentration 9 confidence interval (mg g − 1). Metal Recovery was calculated as follows:
R (%) = ([metal found with ultrasonic extraction] /[metal found with microwave digestion]) ×100 Average recoveries were 1029 7, 999 8 and 98911 for Mg, Mn and Zn, respectively, thus indicating that there was a good agreement between both sets of results. Between-batch precision, expressed as R.S.D. for N= 3 independent extractions, was as average, 0.5, 1.5 and 1%, for Mg, Mn and Zn, respectively, being similar to that of the microwave-assisted digestion method.
4. Conclusions The method described offers a rapid and efficient sample preparation for direct determination of Mg, Mn and Zn in plants by FAAS. The advantages of ultrasound-assisted extraction over microwave-assisted digestion are the following: (i) Ultrasound-assisted extraction is faster than microwave-assisted digestion. The time needed for one extraction was approximately of 7 min (i.e. 3 min sonication time and 4 min for separation of the extract). This time was much lower than that involved in the acid digestion procedure (i.e. 47 min); in this case, only 7 min were required for acid digestion, but it is also necessary to take into account the time required for the reactor to cool before opening (i.e. 20 min with an ice bath) and
b
a
23.490.1 19.6 9 1 31.49 1.3 29.2 9 1.3 33.99 0.7 40.19 1.4
Microwave digestiona X9 ts/ N
Zn
20.1 9 0.5 20.1 90.7 26.9 9 0.5 28.9 92.3 32.9 9 0.2 46.5 90.2
Ultrasonic extractiona X9 ts/ N 86 103 86 99 97 116 98 9 11
495 917 42.7 90.1 108 92 26.0 91.3 29.8 9 1.7 735 9 26
Recoveryb (%) Microwave digestiona X 9ts/ N
Mn
499 914 38.9 9 0.9 95.6 9 1 28.4 9 3 31.2 9 1.7 731 92
Ultrasonic extractiona X 9 ts/ N
101 91 88 109 105 99 999 8
1435947 22189 105 2799989 24449 52 29279117 24589111
Recoveryb (%) Microwave digestiona X 9 ts/ N
Mg
1439914 2472 9 17 2544 9 11 2653 940 2989 9 31 2501 9 72
Ultrasonic extractiona X 9ts/ N
100 111 91 109 102 102 102 9 7
Recoveryb (%)
Average value 9 confidence interval (P= 0.05; N= 3) (mg g−1). Recovery was expressed as the following ratio: ([metal content using ultrasound-assisted extraction]/[metal content using microwave-assisted digestion])×100.
Cinnamon Nutmeg Parsley Sweet paprika Hot paprika Black tea Average recovery (%)
Plant sample
Table 4 Analytical results for Mg, Mn and Zn as determined by ultrasound-assisted extraction and microwave-assisted digestion
440 A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
the time needed for heating to dryness so that the excess of acid can be eliminated (i.e. 20 min). (ii) The consumption of reagents is diminished. (iii) The procedure is safer than acid digestion as neither pressure nor high temperature are present during the extraction procedure. Furthermore, the use of corrosive concentrated acids is avoided. A proof sound chamber is recommended for intensive work with ultrasonic processors. (iv) The whole procedure is simpler since a lesser number of operations is involved that minimises contamination risks. Some drawbacks arise with ultrasound-assisted extraction, which are listed below: (i) The maximum sample amount that can be employed is lower than that used for acid digestion. (ii) The suitable particle size for ultrasound-assisted extraction is a critical parameter, a small size (e.g. 50 mm) being necessary in order to achieve quantitative extraction. (iii) Digestion procedures may be more robust than extraction procedures, since ageing of the ultrasonic probe surface can change the extraction efficiency.
Acknowledgements This work has been financially supported by the Ministerio Espan˜ol de Educacio´n y Cultura (Comisio´n Interministerial de Ciencia y Tecnologı´a) and the Vigo University (projects PB98-1081 and 64502.C801).
References [1] R. Chizzola, Ch. Franz, Angew. Bot. 70 (1996) 52. [2] M. Hoenig, Sci. Total. Environ. 176 (1995) 85.
.
441
[3] E. Wieteska, A. Zio´ek, A. Drzewinska, Anal. Chim. Acta 251 (1996) 251. [4] C. Bendicho, I. Lavilla, Ultrasound-assisted extraction, in: I.D. Wilson (Ed.), Encyclopedia of Separation Science, Academic Press, London, 2000. [5] C. Bendicho, I. Lavilla, Applications of ultrasound-assisted metal extractions, in: I.D. Wilson (Ed.), Encyclopedia of Separation Science, Academic Press, London, 2000. [6] T.J. Mason, Practical Sonochemistry: User’s Guide to Applications in Chemistry and Chemical Engineering, Ellis Horwood, Chichester, 1991, pp. 20 – 25. [7] S. Mamba, B. Kratochvil, Int. J. Environ. Anal. Chem. 60 (1995) 295. [8] F. La´zaro, M.D. Luque de Castro, M. Valca´rcel, Anal. Chim. Acta 242 (1991) 283. [9] I. Matejovic, A. Durackova, Commun. Soil Sci. Plant Anal. 25 (1994) 1277. [10] H. Minami, T. Honjyo, I. Atsuya, Spectrochim. Acta B 51 (1996) 211. [11] J.L. Capelo, I. Lavilla, C. Bendicho, J. Anal. At. Spectrom. 13 (1998) 1285. [12] L. Amoedo, J.L. Capelo, I. Lavilla, C. Bendicho, J. Anal. At. Spectrom. 14 (1999) 1221. [13] J.L. Capelo, A.V. Filgueiras, I. Lavilla, C. Bendicho, Talanta 50 (1999) 905. [14] I. Lavilla, A.V. Filgueiras, C. Bendicho, J. Agric. Food Chem. 47 (1999) 5072. [15] I.V. Kubrakova, T.F. Kudinova, E.B. Stavnivenko, N.M. Kuz’min, J. Anal. Chem. 52 (1997) 522. [16] J.E. Rechcigl, G.G. Payne, Commun. Soil Sci. Plant Anal. 21 (1990) 2209. [17] B. Markert, Sci. Total Environ. 176 (1995) 45. [18] J.P. Lorimer, T.J. Mason, Chem. Soc. Rev. 16 (1987) 239. [19] B. Agranat, M. Dudrovin, N. Javski, G. Eskin, Limpieza ultraso´nica. Fundamentos de la fı´sica y la te´cnica de los ultrasonidos, Mir, Moscow, 1987. [20] T.J. Mason, J.P. Lorimer, J.P. Moorhouse, Educ. Chem. 1 (1989) 13. [21] I. Lavilla, B. Pe´rez Cid, C. Bendicho, Int. J. Environ. Anal. Chem. 72 (1998) 47. [22] P. Riesz, T. Kondo, C.M. Krisha, Ultrasonics 28 (1990) 295.
Comparison of ultrasound-assisted extraction and microwave-assisted digestion for determination of magnesium, manganese and zinc in plant samples by flame atomic absorption spectrometry A.V. Filgueiras, J.L. Capelo, I. Lavilla, C. Bendicho * Departamento de Quı´mica Analı´tica y Alimentaria, Uni6ersidad de Vigo, Facultad de Ciencias (Quı´mica), As Lagoas-Marcosende s/n. 36200 Vigo, Spain Received 21 April 2000; received in revised form 6 July 2000; accepted 6 July 2000
Abstract In this paper, a sample preparation method based on acid extraction of magnesium, manganese and zinc from plant tissue by means of high intensity probe ultrasonication is described. Acid extracts obtained upon sonication were directly nebulised into an air–acetylene flame for fast metal determination by atomic absorption spectrometry. Parameters influencing extraction such as sonication time, ultrasound amplitude, sample mass, particle size, extractant composition and volume were fully optimised. Optimum conditions for metal extraction were as follows: a 3-min sonication time, a 30% ultrasonic amplitude, a 0.1-g sample mass, a particle size less than 50 mm, a 0.3% m/v HCl concentration in the extractant solution and a 5-ml extractant volume. Six plant samples used in the human diet were analysed, the concentration range of the three metals approximately being in the range of 1500 – 3000 mg g − 1 for Mg, 30–735 mg g − 1 for Mn and 20–45 mg g − 1 for Zn. Limits of detection corresponding to the ultrasound-assisted extraction method were 0.10, 1.26 and 0.65 mg g − 1 for Mg, Mn and Zn, respectively. Between-batch precision, expressed as R.S.D., was about 0.5, 1.5 and 1% for Mg, Mn and Zn, respectively. Analytical results for the three metals by ultrasound-assisted extraction and microwave-assisted digestion showed a good agreement, thus indicating the possibility of using mild conditions for sample preparation instead of intensive treatments inherent with the digestion method. The advantages and drawbacks of ultrasound-assisted extraction in respect to the microwave-assisted digestion are discussed. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Ultrasonic extraction; Plant analysis; Magnesium; Manganese; Zinc; Flame atomic absorption spectrometry
1. Introduction * Corresponding author. Tel.: +34-986-812281; fax: + 34986-812382. E-mail address: [email protected] (C. Bendicho).
Metal analysis of plants is an essential feature of environmental, biological and chemical research. Metals in plants display biological activity
0039-9140/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 0 0 ) 0 0 5 1 0 - 5
434
A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
as essential or toxic agents, hence being important to establish their normal concentration range and evaluate their role as part of the food chain. On the other hand, contamination processes of medicinal, aromatic and seasoning plants from heavy metals, although representing a small part of the diet, is part of their quality control [1]. Atomic absorption spectrometry is the likeliest technique used for metal determination in these kind of samples, combined with wet or dry ashing procedures for sample preparation [2]. An alternative to intensive sample preparations, which usually involve high temperatures and pressures, is acid extraction under mild conditions [3]. Solid– liquid metal extraction can be enhanced with the use of ultrasound irradiation [4,5]. Ultrasound energy causes its chemical effects through the phenomenon of cavitation, which consists of the production of microbubbles in a liquid when a large negative pressure is applied to it [6]. When a suspension of solid particles is irradiated by ultrasound, particle disruption can takes place, which in turn, causes an increase in the surface area available for reaction. When extraction applications of ultrasound are concerned, particle fragmentation can enhance the ability of the extractant to leach metals. Furthermore, leaching is facilitated by the generation of reactive species, e.g. radicals, as a result of the ultrasonic irradiation. Thus, when water is sonicated H and OH radicals are formed, which in turn result in the formation of oxygen gas and hydrogen peroxide [7]. Metals may be encapsulated within cell walls that must be broken down by the combined effects of dilute acid attack and ultrasonication in order to bring metals into the liquid extractant. Several workers have successfully used ultrasonic extraction of metals from biological samples, quantitative extraction being usually obtained with high intensity probe sonication. Thus, quantitative iron extraction from vegetal matter using an ethanolic solution containing hydrochloric acid has been reported [8]. Calcium, magnesium, potassium, phosphorus, sodium, iron, zinc, copper, and manganese have been extracted with 1 M hydrochloric acid in an ultrasonic bath [9]. Solid– liquid extraction of metals from powdered biolog-
ical samples have been reported with 5 ml of 1 M nitric acid in combination with an ultrasonic cleaner [10]. Cadmium, copper and lead have been extracted with 1.5 ml of 3% v/v nitric acid solution from several biological materials by means of an high intensity probe sonicator [11–13]. Ultrasound-assisted extraction of metals from plant tissue, although not yet sufficiently exploited, could be an attractive alternative to microwave-assisted digestion, thus avoiding concentrated acids, pressure reactions and time consumption involved, since apart from the time required for digestion, cooling of the reactors needs to be accomplished before opening [14]. In this work, parameters influencing ultrasound-assisted extraction such as ultrasound amplitude, sonication time, sample mass, particle size and extractant (volume and composition) are fully investigated. Metal determination in the extracts is carried out by flame atomic absorption spectrometry (FAAS), the results being compared with those obtained by microwave-assisted digestion.
2. Experimental
2.1. Apparatus 2.1.1. Atomic absorption spectrophotometer A Perkin–Elmer atomic absorption spectrophotometer model 2380 (Norwalk, CT, USA) equipped with 10-cm burner head was used for metal determination. A Cathodeon (Cambridge, UK) hollow cathode lamp of Ca and Mg was used as radiation source. The elements were measured under optimised operating conditions by FAAS with an air–acetylene flame. The instrumental parameters are shown in Table 1. 2.1.2. Agate ball mixer mill A MM 2000 Retsch mixer mill (Haan, Germany) was used for grinding the plant samples. Sieves made of nylon with mesh sizes of B25, B 50, B 100, B150 mm were used to study the influence of particle size on extraction.
A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
2.1.3. High intensity ultrasonic processor A 100 W, 20 kHz VC-100 Sonics and Materials (Danbury, CT, USA) high intensity probe sonicator equipped with a Ti – Al – V microtip (6 mm i.d.) was used for ultrasound-assisted extraction. The amplitude control of the ultrasonic processor allowed the ultrasonic vibrations at the probe microtip to be set at any desired level in the 10–100% range of the nominal power. An Alresa (Barcelona, Spain) centrifuge was used for rapid separation of the extracts. 2.1.4. Microwa6e digestion system A MDS-2000 Microwave Sample Preparation System (CEM, Matthews, NC, USA) was used for microwave-assisted digestion of plant samples. The digestions were carried out in CEM advanced composite vessels. 2.2. Reagents The concentrated acids, nitric acid, hydrochloric acid and hydrofluoric acid were of analytical reagent grade (Carlo Erba, Milano, Italy). Deionised distilled water was used throughout the work. All glassware and plasticware was washed with 5% m/v nitric acid and rinsed with deionised distilled water. Stock standard solutions of manganese and zinc containing 1000 mg l − 1 of each element were obtained by dissolving 0.25 g of the pure metal (Carlo Erba) in the minimum amount of (1+ 1) v/v nitric acid and (1+1) hydrochloric acid (Carlo Erba), respectively, and diluting to 250 ml. A stock standard solution of magnesium was obtained by dissolving magnesium chips (Aldrich, Milwaukee, WI, USA) in 1% v/v hydrochloric acid. Calibration standards of each element were obtained by appropriately diluting the stock solutions.
2.3. Samples The plant samples analysed were cinnamon, nutmeg, parsley, sweet paprika, hot paprika and black tea. All them were obtained from the market in the form of powder or leave. To reduce particle size they were ground in the agate ball mixer mill using a power between 20 and 80% and a grinding time in the range 2–5 min. The powdered samples were sieved so that the required particle size was obtained. Plant samples were dried for 24 h at 70°C as recommended elsewhere [15,16]. Once the samples were powdered and dried, they were kept in labelled capped glass flasks [17] inside a desiccator.
2.4. Ultrasound-assisted extraction procedure A portion (0.1 g) of sample was weighed into polyethylene centrifuge tubes (50 ml capacity) and 5 ml of 0.3% m/v hydrochloric acid were added. Then, the sample was sonicated for 3 min at 30% ultrasound amplitude. After sonication, the supernatant liquid was separated from the solid phase by centrifugation for 4 min at 4500 rpm. Determination of magnesium, manganese and zinc was carried out in the supernatant. Blanks were treated in the same way.
2.5. Microwa6e-assisted digestion procedure The microwave-assisted digestion was optimised elsewhere [14], and can be summarised as follows: about 0.1 g of sample were weighed into the Teflon vessel, and then, a mixture of 5 ml 69.5% m/m nitric acid and 0.5 ml 48% m/m hydrofluoric acid was added. The digestion vessel was closed and heated in the CEM microwave oven for a preselected program (i.e. two stages of
Table 1 Instrumental parameters for determination of Mg, Mn and Zn by flame atomic absorption spectrometry Element
Wavelength (nm)
Mg Mn Zn
285.2 279.5 213.9
Lamp current (mA) 4 15 15
435
Bandpass (nm)
Air/acetylene flow-rate (l min−1)
0.7 0.2 0.7
11/1 11/1 11/1
A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
436
Table 2 Operating conditions for extraction of Mg, Mn and Zn from plant samples Variable
Sonication amplitude (%) Sample amount (g) Particle size (mm) % m/v HCl (extractant) Extractant volume (ml)
Studied interval
10–60 0.05–0.2 25–150 0–5 3–9
Optimum extraction conditions 30 0.10 B50 0.3 5
1 min at 40 and 80 psi, and a final stage of 5 min at 120 psi). Once the digestion program was finished, the reactor was cooled in an ice bath before opening. The contents of each vessel were heated to dryness and dissolved with 1 ml of 37% m/m hydrochloric acid. The solution was quantitatively transferred into a 5-ml volumetric flask and made up to volume with deionised distilled water. The blanks were treated in the same way. Both procedures were validated against CRM GBW07605 (GSV-4) tea leaves (National Research Center for Certified Reference Materials, China), which was previously ground to pass through a nylon sieve (50 mm mesh size).
Variables influencing the extraction process were optimised within the intervals shown in Table 2. Each result was the average value of three determination performed in separate batches.
3.2. Influence of sonication time Temperature of the extraction medium increases with increasing sonication time. Usually, as the temperature increases up to 50°C, extraction efficiency is increased as a result of the larger number of cavitation nucleus formed [18,19]. As the temperature approaches the boiling point of the liquid, ineffective sonication occurs as a result of the decrease in surface tension and increase in vapour pressure within the microbubble, which in turn cause the damping of the shock wave [20]. The influence of sonication time on metal extraction is shown in Fig. 1(a). For the three metals, extraction efficiency increased with increasing sonication time from 1 to 2 min. Extraction efficiency was slightly worse (i.e. 95% for Mg) when the temperature of the medium exceeded 50–60°C, which occurred after a 3 min sonication time. A sonication time of 3 min was seen to be suitable for metal extraction.
2.6. Analytical determinations
3.3. Influence of sonication amplitude
Three sub-samples of each plant sample were used for analytical determinations with the digestion and extraction procedures. With each series of digestions or extractions a blank was measured. All measurements were run in triplicate for the sample solutions and standard solutions. Calculations of metal contents in samples are based on a calibration graph obtained from aqueous standards.
Intensity of ultrasound transmitted to the medium is directly related to the vibration amplitude of the probe. Usually, an increase in intensity will provide for an increase in the sonochemical effects. However, at high vibrational amplitude a great number of cavitation bubbles are generated in the solution, which may dampen the passage of sound energy through the liquid [6]. The influence of ultrasound amplitude in the range 10–60% is shown in Fig. 1(b). The efficiency of the process was maximum for an amplitude of 20%. For the three metals, extraction efficiency increased with increasing amplitude from 10 to 20% and remained constant for higher amplitude values. Results pointed out that the optimum extraction efficiency was attained in the range of 20–60% amplitude.
3. Results and discussion
3.1. Optimisation of the ultrasound-assisted extraction method Cinnamon was used for optimisation purposes.
A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
3.4. Sample amount Sample amount used largely depends on the procedure followed. In this work, the 0.05–0.20 g
437
mass interval was investigated. A sample mass of up to 0.2 g [7] has been used for extraction with ultrasonic probe equipment. A sample mass of up to 0.5 g has been reported in the work with an
Fig. 1. Optimisation of the main parameters influencing ultrasound-assisted extraction (, Mg; , Mn; , Zn). (a) Time (min); (b) Amplitude (%); (c) Sample amount (g); (d) Particle size (mm); (e) Hydrochloric acid concentration (%, m/v); (f) Nitric acid concentration (%, m/v).
438
A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
ultrasonic bath for extraction [21]. Solid–liquid extraction of Cd and Pb from biological materials has shown that a sample mass above 10–20 mg cause the extraction efficiency to decrease [11–13] when using typical autosampler cups for electrothermal – atomic absorption spectrometry. As can be seen in Fig. 1(c), a significant decrease in metal recovery from plant tissue is obtained when the sample mass is larger than 0.10–0.15 g.
3.5. Particle size Particle size was among the more critical parameters influencing ultrasound-assisted extraction. As was to be expected, reactions were enhanced on increasing the contact surface. However, from a certain particle size, a further decrease in particle size was not accompanied by an increase in sonochemical effects [6,18]. Only a few studies have included particle size among the experimental conditions to be optimised for ultrasound-assisted extraction of metals. Thus, some workers [21] employed a particle size less than 50 mm. Solid–liquid extraction of Pb from biological materials showed that a particle size larger than 150 mm caused a decrease in extraction efficiency [12]. The particle size attempted for metal extraction in this work with the use of ultrasonic probes ranged from less than 25 up to 150 mm. The results obtained in this study are shown in Fig. 1(d). As can be observed, extraction efficiency decreased when the particle size was larger than 50 mm for the three metals. When the particle size fraction of 50–100 mm was used for extraction a decrease in extraction efficiency of about 9, 11 and 20 and 16% was observed for Mg and Zn, respectively, in comparison with the 25 – 50-mm fraction. For further experiments, the fraction with particle size less than 50 mm was chosen, since the fraction with particle size less than 25 mm was more difficult to achieve by sieving.
3.6. Composition of acid extractant Acid concentration in the liquid extractant was seen to be the most critical parameter affecting
ultrasound-assisted extraction. Extraction efficiency was about 10% for Mn and Zn in the absence of acidity in the liquid extractant. Different acid mixtures such as 15% m/m nitric acid+ 1% m/m hydrochloric acid [22], 1 M hydrochloric acid+ 20% v/v ethanol [8], 1 M hydrochloric acid [21] and 1 M nitric acid [10] were reported to be used as extractants for solid–liquid extraction of metals. In this work, hydrochloric acid, nitric acid and a mixture of hydrochloric acid+ nitric acid were attempted for ultrasound-assisted extraction. The single acids were employed in the range 0–5% m/v. The acid mixtures used as extractants were 0.3% m/v hydrochloric acid+ 0–1% m/v nitric acid and 0.3% m/v nitric acid + 0–1% m/v hydrochloric acid. Blanks obtained from the different extractants were lower than the limit of detection (LOD) for the metals studied. Extraction results obtained with the use of a single acid (i.e. hydrochloric acid or nitric acid) are shown in Fig. 1(e) and (f). Extraction efficiency increases with increasing acid concentration, a steady extraction efficiency being reached for an acid concentration higher than 0.2% m/v. Likewise, the use of an acid mixture where at least one component is at a concentration higher than 0.2% can be used for extraction. The three metals studied display a similar extraction behaviour with all acids and their mixtures. A 0.3% m/v hydrochloric acid concentration was chosen for extraction since it has not oxidant properties, being more convenient than nitric acid for sample introduction by pneumatic nebulisation into the flame.
3.7. Extractant 6olume Variable volumes of extractant such as 50 ml [21] or 5 ml [10,21], have been employed for solid–liquid extraction of metals. In this work, the extractant volume was studied in the range of 3–9 ml. The influence of extractant volume on extraction efficiency was negligible for the three metals studied. An extractant volume of 5 ml was chose so that the required number of replicates could be performed without exhaustion of the sample solution.
A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
439
Table 3 Validation of the ultrasound-assisted extraction (UAE) and microwave-assisted digestion (MAD) methods against CRM GBW 07605 tea leaves Element
Certified value (X9 ts/ N)a
MAD value (X 9 ts/ N)b
UAE value (X9 ts/ N)b
Mga Mn Zn
0.1790.01 1240940 26.390.9
0.16 90.01 1256 9 17 26.8 90.6
0.15 9 0.02 1214 925 25.2 9 0.5
a b
Magnesium concentration was expressed as mg g−1. Average value 9 confidence interval (P= 0.05; N =5) (mg g−1).
3.8. Analytical results using microwa6e-assisted digestion and ultrasound-assisted extraction The detection and quantification limits were defined as 3s/m and 10s/m, respectively. Both values were calculated from the S.D. (s) of ten blank measurements and the slope of calibration graph (m). Limits of detection (LODs) for the ultrasound-assisted extraction method were 0.10, 1.26 and 0.65 mg g − 1 for Mg, Mn and Zn, respectively, being similar to those attained with microwave-assisted digestion when a 0.1-g sample mass was used for digestion. Limits of quantification (LOQs) were 0.4, 4.2 and 2.2 mg g − 1 for Mg, Mn and Zn, respectively. The equations for the linear range of the calibration graphs were as follows: Abs = 0.6768[Mg],
r= 0.9998
Abs =0.0794[Mn] +0.0002, Abs =0.2012[Zn] −0.0002,
r = 0.9999 r =0.9999
Where the concentration was expressed in each case as mg l − 1. Validation of the ultrasoundassisted extraction and microwave-assisted digestion is shown in Table 3. A good agreement between the found and certified metal contents can be observed for the three metals studied. Analytical results obtained by microwave-assisted digestion and ultrasound-assisted extraction corresponding to the plants analysed are shown in Table 4. Analytical results were expressed as average concentration 9 confidence interval (mg g − 1). Metal Recovery was calculated as follows:
R (%) = ([metal found with ultrasonic extraction] /[metal found with microwave digestion]) ×100 Average recoveries were 1029 7, 999 8 and 98911 for Mg, Mn and Zn, respectively, thus indicating that there was a good agreement between both sets of results. Between-batch precision, expressed as R.S.D. for N= 3 independent extractions, was as average, 0.5, 1.5 and 1%, for Mg, Mn and Zn, respectively, being similar to that of the microwave-assisted digestion method.
4. Conclusions The method described offers a rapid and efficient sample preparation for direct determination of Mg, Mn and Zn in plants by FAAS. The advantages of ultrasound-assisted extraction over microwave-assisted digestion are the following: (i) Ultrasound-assisted extraction is faster than microwave-assisted digestion. The time needed for one extraction was approximately of 7 min (i.e. 3 min sonication time and 4 min for separation of the extract). This time was much lower than that involved in the acid digestion procedure (i.e. 47 min); in this case, only 7 min were required for acid digestion, but it is also necessary to take into account the time required for the reactor to cool before opening (i.e. 20 min with an ice bath) and
b
a
23.490.1 19.6 9 1 31.49 1.3 29.2 9 1.3 33.99 0.7 40.19 1.4
Microwave digestiona X9 ts/ N
Zn
20.1 9 0.5 20.1 90.7 26.9 9 0.5 28.9 92.3 32.9 9 0.2 46.5 90.2
Ultrasonic extractiona X9 ts/ N 86 103 86 99 97 116 98 9 11
495 917 42.7 90.1 108 92 26.0 91.3 29.8 9 1.7 735 9 26
Recoveryb (%) Microwave digestiona X 9ts/ N
Mn
499 914 38.9 9 0.9 95.6 9 1 28.4 9 3 31.2 9 1.7 731 92
Ultrasonic extractiona X 9 ts/ N
101 91 88 109 105 99 999 8
1435947 22189 105 2799989 24449 52 29279117 24589111
Recoveryb (%) Microwave digestiona X 9 ts/ N
Mg
1439914 2472 9 17 2544 9 11 2653 940 2989 9 31 2501 9 72
Ultrasonic extractiona X 9ts/ N
100 111 91 109 102 102 102 9 7
Recoveryb (%)
Average value 9 confidence interval (P= 0.05; N= 3) (mg g−1). Recovery was expressed as the following ratio: ([metal content using ultrasound-assisted extraction]/[metal content using microwave-assisted digestion])×100.
Cinnamon Nutmeg Parsley Sweet paprika Hot paprika Black tea Average recovery (%)
Plant sample
Table 4 Analytical results for Mg, Mn and Zn as determined by ultrasound-assisted extraction and microwave-assisted digestion
440 A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
A.V. Filgueiras et al. / Talanta 53 (2000) 433–441
the time needed for heating to dryness so that the excess of acid can be eliminated (i.e. 20 min). (ii) The consumption of reagents is diminished. (iii) The procedure is safer than acid digestion as neither pressure nor high temperature are present during the extraction procedure. Furthermore, the use of corrosive concentrated acids is avoided. A proof sound chamber is recommended for intensive work with ultrasonic processors. (iv) The whole procedure is simpler since a lesser number of operations is involved that minimises contamination risks. Some drawbacks arise with ultrasound-assisted extraction, which are listed below: (i) The maximum sample amount that can be employed is lower than that used for acid digestion. (ii) The suitable particle size for ultrasound-assisted extraction is a critical parameter, a small size (e.g. 50 mm) being necessary in order to achieve quantitative extraction. (iii) Digestion procedures may be more robust than extraction procedures, since ageing of the ultrasonic probe surface can change the extraction efficiency.
Acknowledgements This work has been financially supported by the Ministerio Espan˜ol de Educacio´n y Cultura (Comisio´n Interministerial de Ciencia y Tecnologı´a) and the Vigo University (projects PB98-1081 and 64502.C801).
References [1] R. Chizzola, Ch. Franz, Angew. Bot. 70 (1996) 52. [2] M. Hoenig, Sci. Total. Environ. 176 (1995) 85.
.
441
[3] E. Wieteska, A. Zio´ek, A. Drzewinska, Anal. Chim. Acta 251 (1996) 251. [4] C. Bendicho, I. Lavilla, Ultrasound-assisted extraction, in: I.D. Wilson (Ed.), Encyclopedia of Separation Science, Academic Press, London, 2000. [5] C. Bendicho, I. Lavilla, Applications of ultrasound-assisted metal extractions, in: I.D. Wilson (Ed.), Encyclopedia of Separation Science, Academic Press, London, 2000. [6] T.J. Mason, Practical Sonochemistry: User’s Guide to Applications in Chemistry and Chemical Engineering, Ellis Horwood, Chichester, 1991, pp. 20 – 25. [7] S. Mamba, B. Kratochvil, Int. J. Environ. Anal. Chem. 60 (1995) 295. [8] F. La´zaro, M.D. Luque de Castro, M. Valca´rcel, Anal. Chim. Acta 242 (1991) 283. [9] I. Matejovic, A. Durackova, Commun. Soil Sci. Plant Anal. 25 (1994) 1277. [10] H. Minami, T. Honjyo, I. Atsuya, Spectrochim. Acta B 51 (1996) 211. [11] J.L. Capelo, I. Lavilla, C. Bendicho, J. Anal. At. Spectrom. 13 (1998) 1285. [12] L. Amoedo, J.L. Capelo, I. Lavilla, C. Bendicho, J. Anal. At. Spectrom. 14 (1999) 1221. [13] J.L. Capelo, A.V. Filgueiras, I. Lavilla, C. Bendicho, Talanta 50 (1999) 905. [14] I. Lavilla, A.V. Filgueiras, C. Bendicho, J. Agric. Food Chem. 47 (1999) 5072. [15] I.V. Kubrakova, T.F. Kudinova, E.B. Stavnivenko, N.M. Kuz’min, J. Anal. Chem. 52 (1997) 522. [16] J.E. Rechcigl, G.G. Payne, Commun. Soil Sci. Plant Anal. 21 (1990) 2209. [17] B. Markert, Sci. Total Environ. 176 (1995) 45. [18] J.P. Lorimer, T.J. Mason, Chem. Soc. Rev. 16 (1987) 239. [19] B. Agranat, M. Dudrovin, N. Javski, G. Eskin, Limpieza ultraso´nica. Fundamentos de la fı´sica y la te´cnica de los ultrasonidos, Mir, Moscow, 1987. [20] T.J. Mason, J.P. Lorimer, J.P. Moorhouse, Educ. Chem. 1 (1989) 13. [21] I. Lavilla, B. Pe´rez Cid, C. Bendicho, Int. J. Environ. Anal. Chem. 72 (1998) 47. [22] P. Riesz, T. Kondo, C.M. Krisha, Ultrasonics 28 (1990) 295.