X-Ray Fluorescence in Biological Sciences. Группа авторов

the tea matrix, confirmed by studies of samples of biological origin [13, 89]. In addition, for P, S, and K, the effect of inhomogeneity in the sample distribution on the carrier was more significant when the suspension was used. It should also be noted that the determination of Cl and Br using the acid decomposition method is difficult due to the volatility of these elements, and therefore the use of suspension is recommended for them. The best measurement results for most elements are obtained by analyzing solutions after acid decomposition of tea leaves, whereby this sample preparation method is used to analyze samples of Krasnodar tea. To determine Cl and Br, samples were prepared as suspensions. The average value of the relative standard deviation of the analysis results from the certified values of the CRM for the method used in the verification was not more than 6% for Mg, P, S, Mn, Ni, Cu, Zn, Br, Rb; for K, Ca, Sr, Ba, and Pb – not more than 16%. The largest discrepancies were obtained by determining Ti and Fe. The total uncertainty of TXRF results, considering all stages of analysis (sample preparation and measurement), for all elements except for Ti was on average no more than 16% (n = 3). The uncertainty of applying a sample onto the carrier averaged 7% (n = 7).

Table 3.5 shows the ranges of element concentrations, average concentrations, and standard deviations from the average (n = 3) obtained from the analysis of 19 samples of tea leaves of Krasnodar tea.

      The Student's test was used to evaluate the two quantitative data sets obtained by TXRF and wavelength dispersive X‐ray fluorescence (WDXRF). Good results convergence was obtained for Cl, K, Ca, Mn, Fe, Cu, Zn, Rb, Sr, and Ba. However, the concentrations of P, S, Ni observed in the data had relatively low convergence (74% on average), which is explained by the possible influence of unconsidered factors, which demands further research. Advantages of TXRF over WDXRF include using fewer CRM to quantify data, reducing detection limits for most elements by about an order of magnitude, and not having matrix effects when emitters meet the thin layer criterion.

3.7 Interelement Effects and Procedures of their Accounting

      The present level of development of the theory of X‐ray fluorescence excitation allows researchers to accurately calculate fluorescence intensities for homogeneous samples. In this case software enables us to take into account different matrix effects: the effect of enhancement of element atoms, primary and fluorescent radiation scattered from the sample atoms, enhancement by sample Auger‐ and photoelectrons, cascade transitions, etc. [7790–92].

Table 3.5 Range, mean (C_mean ) and standard deviation (S) of elements' concentrations in tea leaves for the set of Krasnodar tea samples.

Element	Concentrations in tea leaves, mg/kg
Range	Mean and standard deviation
P	2497–5083	3629 ± 616
S	3057–4822	3830 ± 496
K	13 202–29 710	19 984 ± 4385
Ca	4173–6922	5371 ± 787
Mn	480–2007	1222 ± 451
Fe	96.1–327	177 ± 71
Ni	2.96–12.6	7.90 ± 2.45
Cu	10.2–33.8	19.8 ± 6.28
Zn^a	20–40.7	33 ± 7.0
Br^b	1.8–6.23	3.24 ± 1.24
Rb	15.2–166	67.1 ± 46.6
Sr	10.6–51.3	20.8 ± 9.58
Ba	8.89–63.2	40.5 ± 14.2
Pb	n/d^c–1.23	0.305 ± 0.316

      ^a ‐ In sample M6 Zn with concentration of 1223 mg/l is found, it was not considered at C _mean estimation.

      ^b ‐ Analysis of suspension

      ^c ‐ n/d – it is not determined.

In the 1970s the contribution the secondary electrons, scattered and fluorescent radiation, and the influence of the divergence of the primary beam to the intensities of X‐ray fluorescence were estimated by Irkutsk X‐ray physicists [77, 93]. The influence of inaccuracies in the fundamental parameters, inaccuracies in the description of the distribution of the energy of the primary radiation, etc. were evaluated [42, 75, 76, 94].

      We must list some Russian researchers who have contributed to the solution to these problems: N.F. Losev, G.V. Pavlinsky, V.P. Afonin, A.G. Revenko, Yu.I. Velichko, B.I. Kitov, V.Ya. Borkhodoev, A.L. Finkel'shtein, et al. At that time, many investigations were conducted by our foreign colleagues. First, one has to mention J. Sherman, H. Ebel, T. Shiraiwa, N. Fujino, L.S. Birks, M. Mantler, J. Criss, J. Gilfrich, B. Vrebos, K. Nielson, and others. These researches are being successfully continued by B. Kanngiesser, B. Beckhoff, W. Malzer, R. Sitko, and others.

      The possibility of the application of theoretical intensities was used in the Analytical Center at Institute of the Earth's Crust SB RAS (Irkutsk) to select specific CRMs suitable for calibration to convert measured intensities of analytical lines into concentrations of analyzed elements for different types of geological samples [3495–99].

The estimates of inter‐elemental effects on the intensity of analytical lines for tea, coffee, and some plants which were used in calibration and validation of XRF were presented by Revenko and Sharykina [15]. Table 3.6 show the estimates of the change of the relative specific intensities (I_rel) calculated by the authors of this chapter for the analytical lines of some elements. Variations of the intensity of coherently and non‐coherently scattered radiation of the Rh Kα‐line anode of the X‐ray tube are also given. The minimum and maximum I_rel values are set in bold. Calculations of the intensities were carried out using the program developed by Finkelstein and Afonin [100]. The program algorithm includes the contributions of secondary and tertiary excitation effects as well as the contribution of the radiation scattered by the sample. CRM of the gabbro SGD‐2 [101], diluted in the ratio of 19 : 1 water was taken to calculate the relative intensities of all analytical lines as the reference sample. The calculated intensities were normalized to the ratio of the concentrations of C_samp/C_ref, where, C_samp and C_ref are the concentrations of the element in the CRM or the sample plants and in a reference sample, respectively. For all analytical lines, I_rel for a reference sample is 1.000.

Table 3.6 The relative specific intensities