X-Ray Fluorescence in Biological Sciences. Группа авторов

0.2 ppm to 1.56%. The K concentration in grains is higher than in pulp (~1.5 and 1%, respectively) and significantly lower in cascarilla (0.07%). A typical Ca concentration is 0.15% in grains, 0.5% in pulp and 0.06% in cascarilla. Elements other than Zn and Fe show no significant changes in concentrations. The Zn content is generally higher in pulp (60–70 ppm) than in grains and cascarilla. Fe concentrations vary from 50 to 92 ppm (in the text for some reason up to 1300 ppm) in various grain samples, while Cu concentrations range are from 15 to 20 ppm. It is obtained that the As content is in the range of concentrations between 0.2 to 10 ppm and Pb is about 1 ppm. In some pulp samples, extremely high (~180 ppm) concentrations of Pb were found, but these coffee trees were treated with insecticides for experimental purposes. Concentrations of As and Pb, elements which are of interest were close to their minimum detection limits in some samples, and therefore the results obtained for these elements should be taken with caution. For these cases, the authors pre‐tested two methods of concentration: chemical decomposition followed by precipitation and low temperature ashing. The first method gives particularly good results for all non‐volatile elements such as Fe, Ni, Cu, Zn, Cd, Hg, and Pb. On the other hand, low temperature ashing can be effectively used for both high Z metals and volatile elements, among which As is the most important in coffee analysis.

Table 3.2 Concentrations of elements in some coffee products (ppm).

Element	Sample
Grain 1	Grain 2	Grain 3	Pulp	Cascarilla
K	15 645	15 304	14 700	10 370	725
Ca	1482	1372	1514	5390	621
Ti	—	—	1.2	11	—
Mn	38	36	22	30	11
Fe	54	54	92	45	70
Ni	0.3	0.5	—	0.3	0.6
Cu	18	17	18	27	22
Zn	5.8	5.9	10	57	6
As	—	0.2	1.2	—	0.6
Se	—	0.2	0.4	—	0.1
Br	1.1	1.0	1.2	1.4	1.9
Rb	39	40	38	—	3.3
Pb	—	0.9	0.3	17.2	—

      Haswell et al. [11] discusses the use of the EXTRA II TXRF spectrometer (Germany) for the classification of coffee and wine. In the first step of analysis preparation, the purified quartz reflectors were scanned to check for contamination. The quartz glass reflectors were processed with silicone solution (to prevent wetting on the disc and to aid droplet formation). For the analysis, samples of Cabernet Sauvignon wine from different countries and instant coffee of several brands were used. In preparing the specimens, a solution containing element V (internal standard) and deionized water was added to the wine and coffee samples. An aliquot of the sample with the added internal standard was then pipetted onto the prepared reflector. Measurements for each sample were performed for five hundred seconds, six times. Relative standard deviation for Ni turned out to be more (10 to 50%) than for other elements – K, Ca, Cr, Mn, Fe, Cu, Zn, Br, Rb, and Sr (up to 10%). Based on the results of the determination petal diagrams and dendrograms were built that showed a unique shape depending on chemical composition. The proposed technique allowed researchers to accurately determine the country of wine production and the brand of coffee.

      Ninomiya [83] used TXRF to test the quality of instant coffee. A drop of 1 μl of coffee sample was pipetted onto the Si substrate, then dried and measured. Excitation conditions: 40 kV and 30 mA, Mo Kα radiation of X‐ray tube anode, sample measurement time – 1000 seconds. Spectra obtained during measurement of two samples of instant coffee showed the presence in the coffee spectrum of a doubtful quality of emission line P and Cl, while in the spectrum of ordinary coffee line P is absent, and a weak peak Cl is visible. The Si peak due to the use of the Si substrate was noted. According to the gas chromatography and inductively coupled plasma mass spectrometry method, the difference between the spectra for the two coffee samples is due to the presence in the coffee sample of a doubtful quality of dichlofos, containing P and Cl.

Akamine et al. [84] carried out XRF analyses for 102 samples of fried and green coffee beans from Brazil, Colombia, Guatemala, Indonesia, Tanzania, and Vietnam. The Epsilon 5 polarized ED spectrometer was used. In combination with a 100 kV high voltage generator, this spectrometer allows the determination of high Z element contents with better sensitivity. To build a calibration graph, CRM plants were used: tomato leaves (SRM 1573a), spinach leaves (SRM 1570a), apple leaves (SRM 1515), wheat flour (SRM 1567a), white rice flour (CRM 7502a), and rice flour‐unpolished (CRM No.10‐a). Preparation of the samples for analysis consisted of lyophilization, grinding, and pressing of the material. Authors investigated the effect of tablet thickness (0.5–4.5 g) on the fluorescence intensity of elements with high Z. It was found that the radiation intensities of the Kα‐lines of Rb and Sr increase with increasing tablet weight and become constant at 2.5 g weight (tablet thickness – 6 mm). When the fluorescence intensities were normalized to the intensity of non‐coherently scattered primary radiation, it is obtained that the intensities of the Kα‐lines Rb and Sr do not change practically at the weight of the sample of 1.5 g (thickness – 4 mm). The effect of the secondary target material on fluorescence intensity has been investigated. It was obtained that the optimum secondary target for Mn, Fe, Ni – Ge (sixty minutes), for Rb and Sr – Mo (three hundred seconds), for Ba – Al₂O₃ (sixty