CHEMICAL COMPOSITION

The major, minor and trace elements of tourmalines from Paraiba and Mozambique were analyzed and sorted accordingly to the color of the individual growth zones. The most pronounced changes in color representative for the chemical evolution of the tourmalines are accompanied by a significant change of the minor element concentrations of Cu, Mn, Bi (See Fig. Par36a, 46a-b, 55a-b, 59a, 62a, 66a, 75a-b, 78a-b, 90a-b and 91c-d). This chemical transition can also be extracted from the Pb-Zn-Mg correlation (Fig. Par34, 59b, 62b, 66b-c, 76b and 91b). These trends can further be supported by the elements such as Be, and Fe and Ti (Fig. Par33b, 41a, 43a-b, 49b, 59c,e, 62b,f, 66b,d, 76a, 78f, 84, 85, 86 and 87). In addition to the major and minor elements, 1-5 µg/g concentrations were determined for Sr, Nb, U, Th, La, Ce, Pr, V and Ta. Concentrations between 0.01 and 1 µg/g were determined for Sn, Cr, Co, Rb, Zr, Mo, Ag, Cs, Ba, Nd, Sm, Eu, Gd, Tb, Er, Ho, Tm, Yb and Lu, whereas 0.01 represents the lowest limits of detection determined throughout this study (See table Par02-05 and corresponding figures on pages 25-40 and pages 48-53).

ACCURACY

Testing of the accuracy was carried out using EPMA concentration and LA-ICPMS concentration profiles across the Paraiba Tourmaline sample 2976 (compare Fig. Par28b with Fig. Par36a-b). The accuracy of the quantification approach using NIST 610 for calibration applied for LA-ICP-MS was carried out using different spatial resolution for the two independent, but complementary techniques (EPMA approx. 10 µm and LA-ICP-MS approx. 60 µm). Based on the difference in the spatial resolution, 65 data points (LA-ICP-MS) and 306 data points (EPMA) were acquired on the same traverse shown in Fig. Par28b-c and 36a. It can be seen that the concentration profiles (Na, Ca, Ti, Fe, Cu, Bi and Mn) are in very good agreement between the two independently calibrated techniques. For example, the profile of Mn starts for both methods at very similar concentrations (EPMA 12000 mg/kg, LA-ICP-MS 11500 mg/kg). For further comparison, the average concentrations of the most homogeneously distributed elements (n=306 and n=65, RSD < 5 % for both techniques) were calculated (Data from Tab. Par01 and Tab. Par03). Using this approach, Na concentration of 13500 + 570 mg/kg and Si concentration of 196800 + 2460 mg/kg (EPMA) were closely matched to Na 13800 + 740 mg/kg and Si 181100 + 8140 mg/kg. Even the copper concentration was determined to be approx within 8 % when comparing LA-ICP-MS and EPMA.

LIBS ANALYSIS

Laser induced breakdown spectroscopy (LIBS) was used for direct analysis of the tourmalines. Therefore, a 30 mJ frequency double 266 nm Nd:YAG (Continuum, USA) with a pulse width of 6 ns was coupled to an in-house built ICCD/Echelle spectrometer with a working range of 190-900 nm. The ICCD detector works in a wavelength range between 180-800 nm. The crater ablated on the sample and every element was acquired using 20 spectra (single pulse). Single element spectra were acquired for Cu (324.754 nm), Be (313.042 nm) and Mn (403.076 nm). Ablation crater were generated closely matched to the position of LA-ICP-MS analysis. The concentrations were correlated using the LA-ICP-MS data. Crater diameters using LIBS were selected in the same order of magnitude as applied for LA-ICP-MS to ensure similar destruction of the sample. The results on the LIBS-measurement in comparison to the LA-ICP-MS of same samples are shown in Fig. Par85 and 86.



ED-XRF ANALYSIS

ED-XRF analysis are carried with a Fischer Instrument (XAN-DP) using a 50kV acceleration voltage, a 1000micron Al-Filter, using different beam collimators (0.2mm-2mm, smaller collimators with longer measuring time) and a drift correction mode. The measurement was carried out for minimum of 50-100 seconds. A standard-less element calculation procedure provided by the manufacturer was used to obtain element concentrations of Al, Cu, Mn, Bi, Fe, Zn and Ga. For validation purpose, we measured 12 LA-ICP-MS standards with ED-XRF analyses. The sampling volume of the LA-ICP-MS and ED-XRF are different and the number of measured points by LA-ICP-MS had to be averaged (Tab. Par02-05), before they could be compared to ED-XRF analyses. A linear regression was made for each element measured by 2 different methods (LA-ICP-MS versus ED-XRF) on the basis of 12 standards. A consistent difference between ED-XRF and LA-ICP-MS was detected for the different elements of approx. 30%. This is not surprising because lighter elements (atomic weight below Al) are not measured with this relatively in-expensive ED-XRF instrument. A correction of the ED-XRF data was carried out based on the linear regression coefficient to lower values (See Fig. Par82d). We plotted the corrected ED-XRF results of 112 tourmalines of known origin and compared the results derived by the different methods (See Fig. Par82a-c, Par90a-b and Lit. Par01).