Some Difficulties and Status in the Application of X-Ray Spectrometry in Geological Analysis: A Review

X-ray fluorescence spectrometry (XRF) analysis of geological samples is often performed using solids or powder, which can avoid the time-consuming complex pretreatment process of wet chemical technology and the use of a large number of toxic and harmful chemical reagents^[2-5]. Nowadays, XRF has been widely used in the fields of petrology, geochemistry, chronology, mineral resources and environmental geoscience^[9-11]. However, there are often some difficulties in the practical analysis of geological samples using XRF due to the complexity of geological samples and the characteristics of the elements themselves. Some difficulties and the corresponding possible solutions in geological analysis with the method of XRF are reviewed here, including small size samples or precious samples analysis, the application of scattering effect, the analysis of volatile elements, variable valence elements and rare metals.

　　1. Analysis of small size samples and precious samples

　　Preparation of small beads or pellets which are easy to preserve and repeated measurement is the key for XRF quantitative analysis of geological samples on a small scale. Sometimes, it is necessary to analyze small size samples or precious samples with XRF, such as meteorites or extraterrestrial samples (for example, lunar samples) as well as banded iron formation^[12], loess strata^[13] and periodically banded rocks within zebra rock^[14] which can be drilled from banded positions using micro drilling. Such samples can only be qualitative or semi-quantitative by scanning electron microscopy, micro X-ray fluorescence spectrometry or laser ablation-inductively coupled plasma-mass spectrometry as in previous studies. Some researchers have found that preparation of small beads or pellets which are easy to preserve and repeated measurements yield more accurate quantification with XRF.

　　2. Application of the scattering effects

　　Scattering effects are used to obtain more information about the chemical composition in samples with unknown elemental composition. The calibration curve of Compton-to-Rayleigh intensity ratio versus average atomic number can provide important information when samples with unknown compositions are studied under the same geometrical conditions and at the same energy. First, construct a calibration curve of Compton-to Rayleigh intensity ratio with respect to the average atomic number using reference materials with well-known chemical composition. Then, the concentration of the elements might be indirectly obtained from scattering X-ray peaks of the samples^[19]. This method of the evaluation of an unknown specimen is particularly sensitive for a light matrix.

　　Scattering effects are beneficial for matrix correction of heterogeneous sample in situ analysis. Researchers have studied the element distribution characteristics of bryophyte-soil-rock interface samples^[20]. It was found that the correction method with Ray*Ray/Com is suitable for Ca_Kα, Mn_Kα, Fe_Kα, Cu_Kα, Zn_Kα and Pb_Lα whereas the correction with Ray/Com is good for K_Kα. When the elements’ intensity was corrected by the respective suitable methods, the spectra of K, Ca, Pb, Zn and Cu reached the maximum peaks at the soil layer in interface samples. This exploration is useful for the study of biogeochemical interface processes of the interest elements.

　　3. Analysis of the volatile elements

　　Volatile elements such as C, S and halogen are difficult to analyze accurately with XRF in both the fusion method and pressed powder pellet in general conditions. The fusion method often leads to the measured value drop. Whereas the measurement accuracy is not ideal either using the pressed powder pellet because of the influence of mineral effect and particle effect. Researchers found that the mineral effect and particle effects can be minimized by ultra-fine grind, and that the uncertainty of the analysis results reduced in response^[21]. Besides, The addition of stabilizer can inhibit the volatilization of volatile elements such as S at high-temperature in the melting process, so the analytical accuracy can also be improved^[26].

　　4. Analysis of the variable valence elements

　　The characteristic X-ray spectral peak position, and the shape and relative intensity of the spectral line can be affected by the atomic valence states and coordination states^[32-33]. Some researchers used the relative intensity of Fe Kβ₅/Kβ_1,3 as an analytical parameter to analyze the content of divalent iron (FeO) in igneous rocks. Chubarov, et al.^[33] analyzed the valence state and form of manganese in various kinds of manganese ore by S8 Tiger-type wavelength dispersive X-ray fluorescence spectrometer, so that the quality of manganese ore can be accessed quickly without complex processes.

　　5. Analysis of the rare metals

　　Constructing and optimizing the work curve with artificial standard samples or secondary standard samples are effective ways to analyze the rare metals with XRF^[36]. Zhou et al.^[37] expanded the word curve linear range of La, Ce and Y by adding high purity rare earth oxides La₂O₃, CeO₂ and Y₂O₃ in the analysis of rare earth minerals and mineralized samples. Silva et al.^[38] added the secondary standard to establish the work curve and used the empirical coefficient method to optimize the calibration curve for analysis of La, Ce, Nd, Sm and Gd in Amazon cassiterite tailings. The results of precision and accuracy tests using the CRM (Diorite Gneiss-CCRMP) was satisfactory. Coefficients of variation of five analyzed elements except Gd were less than 5%. The analytical recovery of five elements were between 103% and 116%.

　　Preparing low-dilution (sample to flux ratios) glass beads facilitates the determination of rare metals using XRF. Nakayama et al.^[40] determined 42 components in felsic rocks using XRF. Low-dilution glass beads with 1∶1 of sample-to-flux ratios were prepared to measure Sc, Sn, Cs, Hf, Ta and rare earth elements. Calibration curves of the components showed good linearity (r=0.991−1.000). Ichikawa et al.^[41] developed a low-dilution glass-bead method to determine 34 components including rare metals in basaltic and granitic rocks using XRF. The calibration curves present good linearity (r>0.990).

　　High-energy excitation may help the determination of rare metals in geological samples. Generally, the L line of rare elements is selected as the analytical line because the K line of rare elements cannot be excited by conventional X-ray tubes. However, high-energy polarization XRF can effectively stimulate the K series of rare metals. The overlap interference of K line of rare metals is less and the excitation factor of the K line is larger than the L line, so the excitation efficiency is greatly improved. Researchers developed a method for multi-element analysis of rare earth elements in soil, rock and deposits using high-energy polarized energy dispersive X-ray fluorescence spectrometer. The excitation voltage of the rare earth elements is up to 100kV. The calibration lines show great linearity (the correlation coefficients r>0.99 for La, Ce, Pr, Nd and Y, the rest of rare earth elements r>0.969)^[44].

　　Overlap interference correction of the spectral lines sometimes plays a key role in rare metals analysis ^[42,45]. For example, Maritz et al.^[45] analyze the trace elements in soil. There were significant differences between the measured values and recommended values of Hf and Ta elements because of the ignoring of overlap interference between Cu Kα₁ lines and Hf Lα₁ and Ta Lα₁ lines. Some other researchers fully considered the overlap interference of Ta Lα₁ (0.1522nm) and Cu Kα₁ (0.1541nm) in the determination of Ta, Cs and other rare metals in rocks^[42]. The measurement bias was not significant according to the criterion of negligible error and student’s distributions.

　　6. Challenges and perspectives

　　The limitations and challenges of the application of XRF technology in geological analysis are still prominent: (1) The accuracy of analysis results depends on standard material, especially for geological samples with complex composition, which requires samples with similar physical and chemical properties to participate in the calibration curve; (2) It is still difficult to quantitatively determine the light elements whose atomic number Z<8. In the future, with advances in particle physics, optical systems and detection devices, it is believed that these limitations will be overcome. In addition, with the integrated development of earth sciences in recent years, XRF technology has also developed toward a broader and more dimension-focused direction. New technologies related to XRF, such as micro XRF, X-ray absorption spectrum and portable XRF, are playing an increasingly important role in geoscience field.