JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 227 Determination of Aluminium Barium Magnesium and Manganese in Tea Leaf by Slurry Nebulization Inductively Coupled Plasma Atomic Emission Spectrometry* Colin K. Manickum and Alistair A. Verbeekt Department of Chemistry University of Natal P. 0. Box 375 Pietermaritzburg 3200 South Africa Samples of commercially available tea were ground in an agate pestle and mortar. Particles having diameter greater than about 80 pm were removed from this initial grind material by transport of the finer material in a flow of carbon dioxide and trapping the flow in a tube cooled with liquid nitrogen. Analyses of both the initial grind and the finer material by slurry nebulization using aqueous solutions for calibration purposes and by conventional solution nebulization after dry ashing and acid dissolution were carried out.While the initial grind material gave results for Al Ba Mg and Mn which were up to 4.3% lower than those obtained when using conventional solution analysis slurry nebulization of the finer material gave results having relative standard deviations between 0.3 and 1.8% and not significantly different (95% confidence level) from those of conventional analysis. Keywords inductively coupled plasma atomic emission spectrometry; slurry nebulization; tea leaf Dissolution of solid samples for pneumatic nebulization of the solution into an inductively coupled plasma requires the use of reagents and procedures which are often hazardous and usually lengthy. During these procedures the undesirable risks of analyte contamination and loss of more volatile analyte material are always present.In addition incomplete dissolution of some sample constituents can lead to low results if further sample treatment is not carried out on residues left by the initial attack. The nebulization of finely divided sample mate- rial suspended in a liquid by using a high-solids pneumatic nebulizer provides a convenient method of sample introduc- tion into the plasma and usually requires no alteration to the existing spectrometer except the exchange of the nebulizer itself. This procedure of slurry nebulization is not without problems which can be serious enough to invalidate the use of the technique in some cases.1y2 The sample types that have been investigated using slurry nebulization cover a wide spectrum from plant material and coal to refractory samples such as firebrick and recently an abstract has appeared reporting the successful determination of some elements in plant leaves including tea leaf mussels and soil and rock samples3 using aqueous calibration stan- dards. There have been a number of other reports of successful determinations using slurry nebulization and such calibration standards the work of Ebdon and Goodall' and of McCurdy and Fry4 being particularly noteworthy.Much attention has been paid in these investigations to the importance of ensuring a small particle size of the solid material in the slurry and there is considerable evidence that particle sizes of less than about 6 pm are necessary if inefficiencies in sample transport and atomization are to be a ~ o i d e d .~ - ~ Relatively less attention has however been paid to the use of slurry nebulization into an ICP for plant material analysis as is clear from recent in this journal. This must partially be owing to the relative ease of achieving complete dissolution of such samples for analysis compared with for example geological samples. Nevertheless the problem of contamination and loss of material during the dissolution process remains and it might still be worthwhile to approach the analysis using a slurry rather than a solution nebulization procedure. The work described in this paper concerns the determination of Al Mg Mn and Ba in tea leaf using slurry nebulization ~~ ~ * Presented at the XXVIII Colloquium Spectroscopicurn Internationale (CSI) York UK June 29-July 4 1993.t To whom correspondence should be addressed. into an argon ICP with calibration by means of aqueous standards. The use of aqueous solutions of the elements for calibration purposes is the method of choice since it avoids the pitfalls that can be associated with other methods of calibration' and which lead to analytical inaccuracies. In addition the method of standard additions is not particularly suitable since the addition cannot necessarily be made in the same solid matrix as the sample which would ensure that it had the same transport and atomization properties as the sample slurry. Even if standard reference materials were to be used a very large range would need to be available to deal with all of the types of samples likely to be encountered and in many cases the elements of interest are not certified so that large uncertainties could exist in the mean values used. Experimental Instrumentation A sequential ICP emission spectrometer (Instrumentation Laboratory Plasma-100) in standard form except for the removal of the in-line sample filter was used for the calibration and analysis of samples that had been dissolved.For the slurry samples and their associated aqueous calibration standards the standard nebulizer in its end-cap to the mixing chamber was replaced by a van den Plas V-groove nebulizer in an end- cap. The spoiler cone in the mixing chamber was not removed. Particle size distributions were measured using a laser scattering particle sizer (Malvern Instruments Mastersizer) while sample grinding was performed with an agate electrically driven mortar and pestle (Glen Creston).Sample Treatment Separation of residual coarse-grained (> 100 pm) sample was achieved by transport of the material using carbon dioxide through a baffled glass tube at a flow rate of 1-2mlminV1 with subsequent collection by freezing out in a tube cooled in liquid nitrogen. The finer sample material was recovered following removal of the carbon dioxide by sublimation while the collecting tube was warmed to room temperature. A sample (approximately 0.5 g) for conventional analysis was ashed by slow heating in a furnace from room temperature to 480-490 "C. The residue was treated with concentrated (16 mol 1-') HNO (0.6ml) and evaporated to dryness before being heated at 490 "C for 15 min.The cooled residue was moistened with H 2 0 and dissolved by warming with concen-228 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 trated (10 mol 1 - I ) HC1 (0.25 ml) and H20 (10 ml). Triton- XlOO (25 pl) was added and the solution was diluted to 50 ml. For slurry nebulization ICP-AES sample (approximately 0.1 g) was mixed in an ultrasonic bath for 5 min with 10 ml of a solution containing 10 mol 1-1 HCl(1 ml) and Triton X-100 (0.1 ml) in H 2 0 (200ml). The slurry was then stirred using a magnetic stirrer whilst being nebulized. The wavelengths of the lines used for analysis are presented in Table 1. Each line was corrected for background; the lines for A1 and Mg were corrected on both sides and those for Mn and Ba on the high wavelength side only.Five 1 s counts were made at each wavelength. The optimum observation height for each line was checked for both aqueous solutions and slurries and each line was observed at this height. No difference in observation height was found except in the case of Ba which was measured 2 mm higher in the plasma when introduced as a slurry although this height increase may not be significant. It is only possible to vary the observation height of the Plasma-100 in 2 mm increments so a compromise value of 1 mm higher could not be chosen and consequently during the calibration procedure for the slurry measurements the aqueous standards were observed at the height found to be optimum for slurries.Conditions for the different nebulizer flow rates and the sample pumping rates were optimized by the fixed size simplex method and are noted in Table 1. Calibration of the spec- trometer was by means of the same set of three aqueous mixed element standards containing HC1 and Triton-X regardless of the nebulizer being used. Results and Discussion Any attempt to separate finer from coarser material in a sample introduces the possibility of discrimination between particles of a non-homogeneous mixture. This discrimination becomes less important as the homogeneity of the sample material increases. While such problems are known to be of great importance in for example geological samples where the different minerals present have differing degrees of hardness and resistance to grinding they are not necessarily a source of error in the analysis of relatively soft samples such as biological materials.Nevertheless it is known that leaves and bark or stalks of plants contain differing amounts of some elements as has been confirmed by analysis carried out in these labora- tories on tree samples and vegetable samples such as cabbage. Should a grinding process therefore introduce a particle size discrimination based on the different degrees of hardness of leaves and stalks of tea any subsequent separation of particle sizes would introduce a discrimination factor into the results. In order to confirm that the separation of a relatively small number of larger particles (but a larger proportion of the total volume and thus mass of the sample) would not introduce such errors into the analysis samples of the initial grind material and the finer separated material were analysed using a method based on one which had been shown to be suitable for the analysis of tree leaves and other botanical and animal Table 1 Instrumental parameters Element Wavelength/nm A1 396.15 Ba 455.40 293.65 257.61 Mn Mg Standard IL crossflow nebulizer Pressure 24 lb in-2 Sample pump rate 1.5 ml min-' Pressure 32 Ib in-' Sample pump rate 1.5 ml min-' van Den Plas V-groove nebulizer ,samples." The pertinent properties of the particle size distri- butions and the results obtained for the analysis of these samples are shown in Table 2.The results show that there has been no elemental discrimination in the separation process and the average value for all determinations for each element was used as a basis for comparison of the results obtained by slurry analysis.Table 3 shows the results obtained from the analysis of slurries prepared from both the initial grind and the separated finer material. While the determinations using initial grind material for Al Ba and Mg were significantly lower (95% confidence) than the control values for these elements the results from the separated material were not significantly different at the 95% confidence level as determined by use of the t-test. The results obtained for Mn from the initial grind material however were not significantly different from those of the controls; yet those obtained from the finer material were significantly different (calculated t = 2.89; 6 degrees of freedom) at the 95% but not at the 98% level of confidence. The correspondence between the values can be considered good.The fact that some of the slurry particles were consider- ably larger than the size of the cut-off diameter for transport through some sample introduction system^^,^-^ and yet did not lead to seriously low recoveries indicates that it may not be essential to obtain a very small particle size for slurry nebulization. The maximum particle size in the slurries of Liu and Li3 was determined by the 300 mesh sieve used and was shown to be 44 pm by particle size analysis. These workers also obtained good correspondence between their slurry analy- sis and the values for the certified reference materials they used. Whilst very few minerals have a density which is less than Table 2 of tea leaves (all results as pg g-' of dry leaf) Particle size data and analytical results for control analyses Sample A1 Ba Mg Mn D(v 0.9) = 51.3 pm* 826 39.2 2250 641 D(4,3) = 20.6 pmt 804 39.4 2244 639 D(u 0.5) = 8.1 pm$ Separated finer material D(u 0.9) = 36.1 pm 824 39.6 2269 645 D(4,3) = 13.4 pm 809 39.4 2255 649 D(u 0.5 j = 7.1 pm Mean of all results 816 39.4 2255 644 Initial grind * D(q0.9) is the diameter exceeded by 10% of the volume t D(4,3) is the mean diameter over the volume distribution. This is 4 D(u 0.5) is the median diameter of the distribution.distribution. also known as the Herdan or the Brouckere diameter. Table 3 Results from slurry analysis of tea leaves (all results in pg g-' of dry sample) Sample Initial grind material Separated finer material Mean of separated material results Mean of all conventional analysis results A1 776 & 12* 782 & 28 803 f 13 796 & 16 782 f 8 815 f 14 799 & 141- 816 f llt Ba 37.9 & 0.1 38.2 f 0.1 39.2 0.1 38.3 +_ 0.1 39.4 f 0.1 39.2 k 0.1 39.0 f 0.5 39.4 f 0.2 ~~~ Mg 2249 f 28 2249 & 65 2276 f 54 2276 +_ 40 2260 & 59 2267 & 41 2270 f 8 2255 f 11 Mn 650 k 8 642 k 9 651 _+ 6 648 f 8 656 & 6 651 f 8 652 k 3 644$.4 * Standard deviation within the set of intensities measured for each t Standard deviation of the set of four mean values.sample (n = 5).JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 229 2gcm-3 and the lowest value listed in a table of common mineral densitiesll is 1.6 g cmP3 the density of dried biological material can be much less.Thus transport of fine material of this type by a gas flow might be expected to be much more efficient than transport of denser materials. It would seem more reasonable to suggest that density differences also play an important role in the efficiency of sample transport than to suggest that differences in nebulizer and spray chamber design could account for the apparently much more efficient sample transport of larger particles despite the improvements noted4 in the efficiency of transport when tortuous paths through the system are eliminated. It is also known that the leaching of analytes from the sample occurs during the dispersion pro- cedure. Between 50 and 60% of the total A1 and Ba and between 70 and 77% of the total Mg and Mn were found to be extracted during this 5 min dispersion period.Thus any influence of particle size on the recovery of analyte from the slurry might be diminished considerably by this fact. However in order to achieve good correspondence between conventional and slurry analysis results it is still necessary that slurry particles even if partially leached of analyte must be efficiently delivered to and atomized in the plasma. It is interesting to note that similar amounts of analyte elements are leached in the usual 5 min brewing of tea made with boiling water. The authors are grateful to Dr. A. Pitchford of Hulett Aluminium Pietermaritzburg for provision of the particle size analyses. 1 2 3 4 5 6 7 8 9 10 11 References Ebdon L. and Goodall P. J . Anal. At. Spectrom. 1992 7 111 1. Verbeek A. A. and Brenner I. B. J . Anal. At. Spectrom. 1989 4 23. Liu W.-y. and Li J.-x. ICP Inf. Newsl. 1992 18 237. McCurdy D. L. and Fry R. C. Anal. Chem. 1986 58 3126. Ebdon L. and Collier A. R. J. Anal. At. Spectrom. 1988 3 557. Ebdon L. Foulkes M. E. and Hill S. J . Anal. At. Spectrom. 1990 5 67. Saba C. S. Rhine W. E. and Eisentraut K. J. Anal. Chem. 1981 53 1099. Cresser M. S. Armstrong J. Cook J. Dean J. R. Watkins P. and Cave M. J. Anal. At. Spectrom. 1993 8 1R. Cresser M. S. Armstrong J. Dean J. Watkins P. and Cave M. J. Anal. At. Spectrom. 1992 7 1R. Verbeek A. A. Spectrochim. Acta Part B 1984 39 599. Handbook of Chemistry and Physics ed. Weast R. C. The Chemical Rubber Company Cleveland OH 50th edn. 1970. Paper 3 fO3931 E Received July 7 1993 Accepted October 21 1993