Direct determination of rare earth impurities in lanthanum oxide by Øuorination assisted electrothermal vaporization inductively coupled plasma atomic emission spectrometry with slurry sampling Chen Shizhong, Peng Tianyou,* Jiang Zucheng, Liao Zhenhuan and Hu Bin Department of Chemistry, Wuhan University, Wuhan, 430072, China Received 3rd June 1999, Accepted 25th August 1999 Slurry sample introduction with electrothermal vaporization (ETV) has been applied to inductively coupled plasma atomic emission spectrometry (ICP-AES) for the direct determination of rare earth impurities in lanthanum oxide.A polytetraØuoroethylene (PTFE) emulsion was used as Øuorinating reagent to form volatile Øuorides rather than refractory carbides of rare earth elements (REEs). The Øow path of carrier gas between the graphite furnace device and the ICP torch was improved, and the main factors affecting the analytical signals, such as the Øow rate of carrier gas and auxiliary carrier gas, matrix concentration, exposure time, vaporization temperature and vaporization time, were studied systematically. Under the optimum operating conditions, the detection limits (DL) for 14 REEs were obtained in the range of 2 ng ml21 (Yb) to 130 ng ml21 (Ce), and the relative standard deviation (RSD) is less than 5%.The recommended approach has been applied to directly analyse lanthanum oxide without any chemical pretreatment. Introduction The applications of high-purity rare earth oxides in hightechnology Æelds depend not only on the characteristics of the basic substances, but also on the purity of the compounds.1 Therefore, the development of a rapid, sensitive and reliable method for the determination of trace rare earth elements (REEs) is very essential.Of the many analysis techniques for rare earth oxides, it is now accepted that inductively coupled plasma atomic emission spectrometry (ICP-AES) offers numerous advantages as a technique for the quantitative analysis of REEs.2±6 However, the conventional ICP-AES method suffers from the following problems: (1) limited sensitivity; (2) low nebulization efÆciency; (3) spectral interference and matrix effects.Electrothermal vaporization (ETV) as sample introduction device combined with ICP-AES provides very attractive features, including high sampling efÆciency, requirement of a small sample, low absolute detection limit and direct analysis of solid samples.Unfortunately, the REEs are difÆcult to vaporize from the graphite furnace at high temperature because of the formation of refractory carbides. In previous work with ETV-ICP-AES, halogenating reagents, such as CHF3,7 CCl4 or NH4Cl,8 and Freon-12,9 were utilized to promote the vaporization and transportation of the refractory elements, and improved the detection limits of refractory elements. Fluorination assisted ETV-ICP-AES using a polytetraØuoroethylene (PTFE) emulsion as a Øuorinating reagent has been successfully applied to directly analysing solid biological, environmental and high-purity materials.10±13 In the present study, successful efforts have been made in the analysis of highpurity lanthanum oxide.The conventional Øow path of the carrier gas between ETV and ICP was replaced by a new joint design, and the main factors affecting the analytical signals, such as the Øow rate of carrier gas and auxiliary carrier gas, matrix concentration, exposure time, vaporization temperature and vaporization time, were investigated in detail.The results showed that the proposed method has some advantages in sensitivity, rapidity, reliability, no requirement for chemical pretreatment and matrix matching. Experimental Apparatus and operating conditions A 2 kW, 27°3 MHz ICP generator (Beijing Second Broadcast Equipment Factory, Beijing, China) and an ICP torch (Chang Sha Quartz and Glass Factory, Changsha, China) according to Fassel were used.A modiÆed graphite furnace vaporizer was used as the vaporization device.10±13 The evolved compounds were swept by carrier gas into the plasma excitation source through a concentric glass tube with three openings (Fig. 1) connected with plastic tube (4 mm id). The analytical signals were recorded using a WPG-100 plane grating spectrograph with 1200 grooves mm21 blazed for 300 nm (dispersion 0.8 nm mm21 in Ærst-order spectrum, Beijing Second Broadcast Equipment Factory, Beijing, China).The spectrum was photographically recorded on a photographic plate (ultraviolet I, Tianjing, China) and the blackening values of the analysis lines were measured using a microphotometer (Zeis II model, Jena, Germany). Then the blackening values were converted to signal intensity values by the characteristic curve of emulsion. The operating conditions for ETV-ICP-AES are listed in Table 1. Standard solutions and reagents The stock standard solutions of REEs with a concentration of 1 mgml21 were prepared by dissolving their specpure oxides in Fig. 1 Block diagram of the joint of the gas Øow path.J. Anal. At. Spectrom., 1999, 14, 1723±1726 1723 This Journal is # The Royal Society of Chemistry 1999dilute HCl, followed by dilution to a certain volume with water. A 60% (m/v) PTFE emulsion (dv1 mm) was provided by the Institute of Shanghai Organic Chemistry (Shanghai, China). All other chemicals used in this work were of specpure grade.Twice-distilled water was used throughout. Slurry sample preparation The La2O3 powder (50 mg) was accurately weighed into a 5.0 ml test tube; 0.75 ml of 60% (m/v) PTFE emulsion and 3.0 ml of 0.1% (m/v) agar solution were added, and then adjusted to 5.0 ml with water. The resulting mixtures were dispersed with an ultrasonic vibrator for 20 min, and the test tube was shaken prior to any sampling. Recommended procedure After the ICP had stabilized, a 20 ml sample was pipetted into the graphite furnace.After being dried and ashed, the analyte was vaporized and carried into the plasma by the argon gas, and the emission signals of the analytes were recorded. Results and discussion The interface between ETV and ICP Aschematic diagram of the modiÆed Øow path of the carrier gas is shown in Fig. 1.13 A concentric glass tube with three openings (5 cm long, 1 mm internal tube id and 1 mm interval between internal and outer tube) was connected between the ETV device and the ICP torch.In this system, the conventional Øow path of the carrier gas was divided into two parts. One was a sample carrier gas from the ETV device to transport the vaporized sample into the ICP, and the other was an auxiliary carrier gas from the outer tube of the concentric glass tube to maintain the ICP channel and to obtain the optimum analytical signal. Optimization of the gas Øow rate The effect of the Øow rate of the carrier gas and the auxiliary carrier gas on the analytical signal of Yb is shown in Fig. 2. As can be seen from Fig. 2, a lower Øow rate of carrier gas and a higher Øow rate of auxiliary carrier gas gave the optimum signal-to-background (S/B) ratio because of the effective ICP channel and high transportation efÆciency. Similar effects were observed with other elements, in the absence and presence of PTFE. Therefore, a carrier gas Øow rate of 0.3 l min21 and an auxiliary carrier gas Øow rate of 0.7 l min21 were chosen in this work.Effect of matrix concentration Fig. 3 shows the effects of matrix (La) concentration on the determination of REE impurities. It can be seen that, when the concentration of matrix (La) varied from 0 to 10 mg ml21, no obvious emission signal intensity changes of the analytes were observed. However, once the matrix (La) concentration surpasses the tolerance, the analytical signal intensities begin to increase or decrease. The decrease in the signal intensity could be attributed to two competitive reactions.One took place between the REEs determined and the Øuorinating reagent PTFE, and the other between the matrix (La) and the Øuorinating reagent PTFE, leading to incomplete vaporization of the REE impurities. The increase in the signal intensity was mainly due to spectral interference from the emission lines of La in the vicinity of the REE analytical lines, causing an increase in the signal intensity of the analytes. In real sample analysis, a sample concentration of 10 mg ml21 was used.Choice of vaporization time and exposure time At the vaporization temperature of 2400 �C, the effects of the vaporization time on the analytical signals are shown in Fig. 4. The vaporization time has an obvious inØuence on the analytical signal intensities from 0.5 to 3 s with PTFE, and then the emission intensity reaches a plateau. This indicates that the REEs could be completely vaporized within a short time.In this study, 4 s was chosen as the vaporization time. The relationship between the signal intensity and the exposure time in the presence of PTFE is given in Fig. 5. In this technique, an integrated signal was detected, and the optimum exposure time is the time necessary for complete transportation of the vaporized analytes; therefore, an exposure time of 10 s should be chosen. Table 1 The operating conditions for ETV-ICP-AES Incident power/kW 1.2 Carrier gas/l min21 0.3 Auxiliary carrier gas/l min21 0.7 Coolant gas/l min21 16 Entrance slit width/mm 20 Drying temperature/�C 100; ramp 10 s, hold 20 s Ashing temperature/�C 1000; ramp 10 s, hold 20 s Vaporization temperature/�C 2400, 4 s Exposure time/s 10 Sample volume/ml 20 Fig. 2 Effect of the Øow rate of the carrier gas and the auxiliary carrier gas on the signal-to-background (S/B) ratio. Yb, 1.0 mg ml21. Flow rate of auxiliary carrier gas/l min21: curve 1, 0.3; curve 2, 0.5; curve 3, 0.7; curve 4, 0.8; curve 5, 0.9.Fig. 3 Effect of matrix concentration on the analytical signals with PTFE: Yb, 1.5 mg ml21; Eu, Gd and Ho, 3.0 mg ml21; Pr and Sm, 5.0 mg ml21. 1724 J. Anal. At. Spectrom., 1999, 14, 1723±1726Vaporization temperature The vaporization temperature has a critical effect on the analytical signal. From Fig. 6, it can be seen that, in the absence of PTFE, poor analytical signals are obtained and do not reach a maximum until 2800 �C. However, in the presence of PTFE, stronger analytical signals are observed, and reach a plateau above a temperature of 2400 �C due to the formation of more volatile Øuorides with similar vaporization characteristics.In this study, a vaporization temperature of 2400 �C was used for simultaneous multielement determination. Detection limits and precision The detection limit (DL) is deÆned as the analyte concentration yielding a signal equal to three times the standard deviation of the background noise.The detection limits and relative standard deviations (RSD) for the proposed method (C~2.0 mg ml21, n~9) are summarized in Table 2. Table 2 Fig. 4 The analytical signal as a function of the vaporization time with PTFE: Yb, 1.5 mg ml21; Eu, Gd and Ho, 3.0 mg ml21; Pr and Sm, 5.0 mg ml21. Fig. 5 Dependence of the analytical signal on the exposure time with PTFE: Yb, 1.5 mg ml21; Eu, Gd and Ho, 3.0 mg ml21; Pr and Sm, 5.0 mg ml21. Table 2 The detection limits (DL) and relative standard deviations (RSD) for this method DL This method/ PN14/ TC-ETV15/ Element Wavelength/nm ng ml21 ng ng pg RSD (%) Ce 413.380 130 2.4 110 – 4.3 Pr 390.844 30 0.6 98 – 3.4 Nd 406.190 90 1.8 98 – 3.8 Sm 359.260 40 0.8 60 – 3.1 Eu 397.199 6 0.1 4 12 2.8 Gd 342.447 14 0.3 66 – 2.5 Tb 350.917 20 0.4 16 – 2.7 Dy 353.170 8 0.2 7 – 2.3 Ho 345.600 5 0.1 5 – 1.8 Er 337.271 7 0.2 1 34 1.8 Tm 313.126 25 0.5 0.3 – 3.0 Yb 328.937 2 0.04 1 18 1.4 Lu 307.765 30 0.6 22 54 2.7 Y 321.669 6 0.1 4 25 2.0 Fig. 6 The analytical signal versus the vaporization temperature. (A) With PTFE: Yb, 1.5 mg ml21; Eu, Gd and Ho, 3.0 mg ml21; Pr and Sm, 5.0 mg ml21. (B) Without PTFE. Table 3 The analytical results of REEs in La2O3 powder (n~5) Element Calibration curve methoda/ mg g21 PN-ICP-AESa/ mg g21 Standard addition methodb/ mg g21 Calibration curve methodb/ mg g21 Ce 112.3°6.0 115.0°7.0 116.9°8.0 110.5°8.0 Pr 73.6°3.0 75.8°2.0 69.8°5.0 72.4°4.0 Nd – – – – Sm 55.4°3.0 56.7°5.0 52.8°3.0 53.6°4.0 Eu – – – – Gd 15.5°2.0 16.7°2.0 13.4°2.0 14.2°4.0 Tb – – – – Dy 4.85°0.6 5.11°0.6 4.58°0.5 4.73°0.3 Ho 2.54°0.3 2.92°0.4 2.48°0.4 2.77°0.2 Er 5.92°0.4 6.21°0.4 5.75°0.6 5.60°0.6 Tm 12.1°2.0 14.5°2.0 12.4°2.0 13.7°2.0 Yb 1.87°0.2 2.05°0.2 1.91°0.2 1.85°0.1 Lu – – – – Y 4.21°0.4 4.53°0.3 3.95°0.5 4.03°0.4 aAnalysis after dissolving sample with HCl.bDirect analysis with slurry sampling. J. Anal. At. Spectrom., 1999, 14, 1723±1726 1725also shows the comparison of the DL of the REEs obtained by this method, conventional pneumatic nebulization (PN)-ICPAES and tungsten coil (TC)-ETV-ICP-AES.As can be seen from Table 2, the DL of the REEs obtained are comparable or better than those for the TC-ETV-ICP-AES and PN-ICP-AES methods. Sample analysis The contents of the trace REE impurities in high-purity La2O3 were directly determined according to the described method. The sample was also analysed by PN-ICP-AES, and the analytical results are listed in Table 3.The recoveries of some REEs are summarized in Table 4, and were obtained by the calibration curve method with slurry sampling. Conclusion In brief, the application of slurry sampling ETV-ICP-AES for the determination of trace REE impurities in high-purity lanthanum oxide has numerous advantages, such as: (1) elimination of sample pretreatment; (2) small sample requirement; (3) simple and rapid operation; (4) reduction of sample contamination; (5) calibration with standard solutions without matrix matching.Therefore, it is likely to become an effective method for the direct determination of trace impurities in powder samples. Acknowledgements This work was supported by the National Science Foundation and the Education Ministry Foundation of China. References 1 W. Fey and K. H. Lieser, Fresenius' J. Anal. Chem., 1993, 346, 896. 2 Z. C. Jiang and B. Hu, Fenxi Kexue Xuebao, 1995, 11(2), 62. 3 J. Y. Li, J. Yang and J. R. Dong, Guangpuxue Yu Guangpu Fenxi, 1995, 15(4), 71. 4 Z. C. Jiang, H. Chen and S. X. Zhang, Gaodeng Xuexiao Huaxue Xuebao, 1990, 11(11), 1283. 5 L. Darbha and S. Gangadharan, Fresenius' J. Anal. Chem., 1994, 348(4), 284. 6 W. S. Li, C. L. Peng, P. Yuan, W. D. Qi, Z. X. Kuang and C. H. Xu, Fenxi Ceshi Xuebao, 1998, 17(1), 18. 7 G. F. Kirkbright and R. D. Snook, Anal. Chem., 1979, 51, 1938. 8 K. C. Ng and J. A. Caruso, Analyst, 1983, 108, 476. 9 J. M. Ren and E. D. Salin, Spectrochim. Acta, Part B, 1994, 49, 567. 10 T. Peng and Z. Jiang, J. Anal. At. Spectrom., 1998, 13, 75. 11 Z. Jiang, B. Hu, Y. Qin and Y. Zeng, Microchem. J., 1996, 53, 326. 12 T. Peng and Z. Jiang, Fresenius' J. Anal. Chem., 1998, 360, 43. 13 S. Cheng, F. Li, Z. Liao, T. Peng and Z. Jiang, Fresenius J. Anal. Chem., 1999, 364, 556. 14 X. Pu, A. Pei and B. Huang, Chin. J. Anal. Chem., 1989, 17, 61. 15 K. Dittrich, H. Berndt, J. A. C. Broekaert, G. Schaldach and G. Tolg, J. Anal. At. Spectrom., 1988, 3, 1105. Paper 9/04436A Table 4 Recoveries of some rare earth elements Element Added/mg ml21 Found/mg ml21 Recovery (%) Nd 0.0 0.0 – 3.0 3.24 108 5.0 5.37 107 Eu 0.0 0.0 – 0.1 0.096 96 0.2 0.207 104 Tb 0.0 0.0 – 0.5 0.47 94 1.0 0.96 96 Lu 0.0 0.0 – 0.5 0.46 93 1.0 0.95 94 1726 J. Anal. At. Spectrom., 1999, 14, 17