Juan A. V. A. Barros

The selection of an analytical method featuring multielemental capability, wide linear range, and ease of operation would be ideal for rare earth element (REE) determination in complex samples. This work targets the determination of Ce, Dy, Er, Eu, Gd, Ho, La, Lu, Nd, Pr, Sc, Sm, Tb, Th, Tm, Y, and Yb in geological samples, fertilizer, and agricultural gypsum samples using inductively coupled plasma–optical emission spectrometry (ICP-OES). Samples were prepared using microwave-assisted acid digestion and the accuracy was evaluated by addition and recovery experiments for spiked samples at three concentration levels. A careful choice of the plasma viewing position and appropriate dilution factors led to accurate and sensitive determinations of REEs. The proposed procedure using external calibration presented good accuracy and sensitivity.   

	Rare earth elements (REEs) comprise the lanthanides series, as well as Sc and Y. Nowadays these elements are essential for high-technology and military applications (1), and some ores can be classified as a rich source of REEs, such as Ce, Dy, Er, Eu, Gd, Ho, La, Lu, Nd, Pr, Sc, Sm, Tb, Th, Tm, Y, and Yb. The most abundant REEs in the Earth's crust are Ce, La, Nd, and Y (31–66 µg/g), while Tm and Lu (0.5–0.8 µg/g) are found in lower contents (1,2).

	The chemical analysis of REEs in geological materials is critical for better exploration of mineral reserves as well as for studies in geology and geochemistry because these elements present unique properties and can be used as geochemical tracers (3). Moreover, the predominant ionic form of some REEs is trivalent, which can replace Ca in phosphate minerals and may reach soil together with phosphorus fertilizers and agricultural gypsum (4,5).

	Among the spectrochemical methods available, instrumental neutron activation analysis (INAA) and inductively coupled plasma–mass spectrometry (ICP-MS) are commonly used for REE determination in mineral samples because of their multielement capability, high sensitivity, and low detection limits. However, these methods rely on expensive instrumentation and are prone to interferences caused by long irradiation times and spectral overlaps, respectively. In addition, X-ray fluorescence (XRF) spectroscopy may also be applied in the direct solid analysis of these samples; however, XRF lacks sensitivity and thus is not suitable for the determination of REEs at low concentrations (6). In this context, inductively coupled plasma–optical emission spectrometry (ICP-OES) has been reported in the literature as a feasible alternative for REE determination because of the multielement capacity, adequate sensitivity, wide linear dynamic range, and ease of operation (6–8).

	Considering these aspects, the goal of this study was to evaluate the performance of ICP-OES for the determination of REEs in geological samples, phosphorus fertilizers, and agricultural gypsum samples. The ICP-OES instrument was evaluated for measurements in both radial view and dual view (axial and radial) modes.

	An ICP-OES system with a dichroic spectral combiner for simultaneous collection of data in radial and axial viewing mode was used to perform all measurements (5100 SVDV, Agilent Technologies). The instrument arrangement involves a vertical argon plasma. The plasma operating conditions and parameters of the sample introduction system are presented in Table I. Samples were prepared using microwave-assisted acid digestion with an Ethos 1 oven (Milestone).

	All glassware was decontaminated by immersion in 10% v/v HNO

 for at least 24 h and rinsed with distilled–deionized water (resistivity ≥ 18.2 MΩ-cm) obtained from a Milli-Q Water system (Millipore). All solutions and analytical blanks were prepared with ultrapure water and nitric acid obtained using a sub-boiling distillation apparatus (Milestone). Concentrated nitric acid (65% m/m) and hydrochloric acid (37% m/m) were used to prepare the aqua regia solution used to digest samples. Analytical calibration solutions in the 0.05–5 mg/L range were prepared by appropriate dilution of a multielement stock solution (Agilent Technologies) of 10 mg/L Ce, Dy, Er, Eu, Gd, Ho, La, Lu, Nd, Pr, Sc, Sm, Tb, Th, Tm, Y, and Yb. Limits of detection (LOD) were obtained by considering background equivalent concentrations (BEC) and relative standard deviations (RSD) for 10 measurements of digested blank solutions.

Geological and Agricultural Samples Preparation 

	Approximately 100 mg of two geological samples (G1 and G2) were accurately weighed directly in the PTFE-PFA digestion vessels, followed by the addition of 9.0 mL of aqua regia. The loosely capped vessels were left overnight for a predigestion step in a laminar flow clean hood. The digestion was further carried out in a closed-vessel microwave cavity oven according to the following three-step heating program: 10-min ramp to 120 °C, 20-min ramp to 220 °C, and a 5-min plateau at 220 °C. After the digests cooled to room temperature, they were transferred to previously decontaminated 50-mL polypropylene flasks and the volume was made up with distilled–deionized water. Seven fertilizer samples (A–G) and two agricultural gypsum samples (Gy1 and Gy2) were provided by Embrapa Pecuária Sudeste. Samples A, B, and C were a mixture of mineral fertilizers with different NPK contents with B, Cu, Mn, and Zn in their compositions. Samples D, E, and F were a mixture containing B, Cu, Mn, and Zn in a 20-fold higher concentration than samples A, B, and C. Sample G was an organomineral fertilizer. The fertilizer samples were submitted to a 2-h predigestion step at room temperature, followed by microwave-assisted digestion. Microwave-assisted digestions were performed by adding 6 mL of HNO

(30% w/w) to approximately 200 mg of sample. The heating program for fertilizer samples was implemented in two steps: a 20-min ramp to 220 °C, followed by a 20-min plateau at 220 °C. After the digested samples cooled down, they were transferred to polypropylene flasks and volumes were made up to 50.0 mL with distilled–deionized water. The accuracy of the proposed procedures was checked by addition and recovery experiments at a single level at 2.5 mg/L for geological samples and two concentration levels (0.1 and 1 mg/L) for fertilizers. Spikes were added before digestion to check the sample preparation step. Both digestion and recovery procedures were performed in triplicate for each sample.

	The determination of REEs in geological samples, fertilizers, and agricultural gypsum was carried out using external calibration. Limits of detection were calculated considering BEC, signal-to-background ratio (SBR), and relative standard deviation (RSD) for 10 measurements of a blank solution, and the values are shown in Table II. Calibration curves with correlation coefficients better than 0.9990 were obtained for all analytes, showing the feasibility of the REE measurements by ICP-OES.

	Eid and colleagues (9) reported LODs ranging from 0.009 to 0.45 mg/mL for eight REEs determined by ICP-OES in phosphate samples. Lichte and colleagues (10) described a procedure for the determination of REEs in geological materials by ICP-MS. The LODs were in the 2–11 ng/g range with precision (RSD) of 2.5%. However, spectral interferences had to be overcome, leading to the formation of oxides and promoting drifts in sensitivity.

	Addition and recovery experiments were performed to check the accuracy of the proposed procedure. Samples of geological materials, fertilizers, and agricultural gypsum were spiked.

	Geological sample G2 was spiked at 2.5 mg/L and the results demonstrated that suitable accuracies were attained in both viewings. Recoveries for all elements, except Ce, ranged from 90.1% to 107% for both viewing modes. The only exception was for Ce (418.659 nm), where recoveries presented a positive error, 110% and 118%, for radial and dual view modes, respectively. Applying a 

-test, no statistical differences at a 95% confidence level were observed between the radial view and dual view modes. Lanthanum, Nd, and Pr were presented at high concentrations in both samples (G1 and G2), thus a fourfold additional dilution was necessary and reading in radial view mode was the best choice.

	The REE concentrations in fertilizers were lower than in the geological samples; therefore spikes were added at lower levels in the fertilizer samples. Addition and recovery experiments (Table III) were applied to three fertilizer samples (B, D, and G) with different compositions. Due to the low REE concentrations in the fertilizers, the dual viewing mode was used because of its higher sensitivity. The procedure presented suitable accuracy, except for Ce, La, Pr, Th, and Y at 0.1 mg/L in spiked sample B, because the concentrations of these REEs in sample B were higher than the spiked concentrations.

	For the agricultural gypsum sample, spikes at 0.1 and 1.0 mg/L were performed. The best results were reached after a 10-fold additional dilution and using radial view configuration because of the higher concentrations of REEs in this sample, except for Ce, La, and Nd-the added concentrations for these elements were higher than the concentrations naturally present in the samples. Using the dual view mode, good accuracy was attained for lower level spikes (0.1 mg/L) except for Ce, La, and Nd; however, for spikes at the higher level (1.0 mg/L), recoveries were not satisfactory, because the added concentrations were too high when compared to the real concentrations in the sample. Thus, the careful choice of the viewing position and additional dilution led to proper accuracies and sensitivities for REE determination in both viewing modes. For agricultural gypsum samples, radial viewing mode was better.

Determination of REEs in Geological Samples, Fertilizer, and Agricultural Gypsum Samples

	The developed procedure using dual view mode was applied to the determination of REEs in fertilizers and geological samples (Table IV). Geological sample 1 contained higher concentrations of all the analytes than sample 2, except for Sc. Rare earth element determination by ICP-OES was fast and suitable recoveries were obtained. The addition and recovery experiments showed that spectral interferences were not an issue. Three emission lines for each of the 17 elements were measured in less than 2 min in both viewing modes and required a sample volume of approximately 3 mL, demonstrating a high sample throughput for the procedure. Some concentrated samples were further diluted before the analysis, which was not a problem considering the wide linear range of the calibration curves. In the agricultural gypsum samples, analytes were determined using the radial viewing mode (Table V).

	Lutetium and Tm were not found in any fertilizer sample, and the REEs presented at higher concentration levels were the same as those previously cited by Tyler (2). Agricultural gypsum samples presented concentrations fivefold higher than those found in fertilizer samples.

	Spectral and matrix interferences were reported by Eid and colleagues (9) for REE determination, where suppression in La II 408.6-nm and Nd II 406.1-nm emission intensities were observed because of high Ca concentrations. Jaron and colleagues (11) reported spectral interferences caused by Fe and Ti overlaps and matrix interferences caused by the presence of Ca and Mg during the analysis of geological samples. However, those interferences were not found in the present work. The ICP-OES method was fast and suitable for the determination of rare earth elements in a wide concentration range in geological and agricultural samples.

	The ICP-OES method presented multielement capability and selectivity that was compatible with accurate REE determination in complex materials such as ores and agricultural samples. The high sample throughput and adequate accuracy and precision of the developed procedure indicates that ICP-OES is a suitable alternative for REE determination in geological and fertilizer samples. The feasibility of fast measurements of emission signals in radial and dual viewing modes was demonstrated here.

	The determination of REEs in geological and agricultural materials can provide valuable information about the geochemical formation, plant nutritional status, the need for supplementation, and possible contamination. Moreover, in some cases it is important to have a fast sample profile to check for possible interfering elements on other target analytes when applying different methods-for example, the magnitude of interferences caused by double-charged species of 

	The authors are grateful to grants 2013/26672-5 and 2014/18393-1 of São Paulo Research Foundation (FAPESP), for the scholarships provided to R.C.M. and A.V., and for research grant 2015/14488-0. The authors are also thankful to the Conselho Nacional de Desenvolvimento Científico e TecnolÓgico (CNPq, Grants 443771/2014-6 and 303107/2013-8), and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES – Grant 15/2104) for fellowships and financial support. We also would like to thank Agilent Technologies for their support.

	(3) K.H. Johannesson, K.J. Stetzenbachand, and V.F. Hodge, 

	(4) A. Bakou, C. Buser, G. Dandulakis, G. Brudvigand, and D.F. Ghanotakis, 

	(5) C.H.R. Saueia, F.M. Le Bourlegat, B.P. Mazzilli, and D.I.T. Fávaro, 

	(6) B. Zawisza, K. Pytlakowska, B. Feist, M. Polowniak, A. Kita, and R. Sitko, 

	(7) M.S. Navarro, H.H.G.J. Ulbrich, S. Andrade, and V.A. Janasi, 

	(8) A.K.G. Silva, J.C. de Lena, R.E.S. Froes, L.M. Costa, and C.C. Nascentes, 

	(9) M.A. Eid, J.A.C. Broekaert, and P. TschÖpel, 

	(10) F.E. Lichte, A.L. Meier, and J.G. Crock, 

	(11) I. Jaron, B. Kudowska, and E. Bulska, 

	(12) B.P. Jackson, A. Liba, and J. Nelson, 

Clarice D. B. Amaral and Raquel C.Machado

 are with the Group for Applied Instrumental Analysis in the Department of Chemistry at Federal University of São Carlos in São Carlos, Brazil and Embrapa Pecuária Sudeste in São Carlos, Brazil. 

Juan A. V. A. Barros, Alex Virgilio, and Joaquim A. Nóbrega

 are also with the Group for Applied Instrumental Analysis in the Department of Chemistry at Federal University of São Carlos. 

 is also with Embrapa Pecuária Sudeste. 

 is with Agilent Technologies in Barueri, Brazil. Direct correspondence to: 

Articles

This method demonstrates that ICP-OES is a suitable alternative to ICP-MS for the determination of rare earth elements in geological and agricultural samples

Juan A. V. A. Barros

Articles

Determination of Rare Earth Elements in Geological and Agricultural Samples by ICP-OES