摘要:

INTER-LABORATORY NOTE Determination of total tin in environmental and geological samples by electrothermal atomic absorption spectrometry using a tungsten furnace after pretreatment by solvent extraction and cobalt(III) oxide collection Tomohiro Narukawa Department of Chemistry, Chiba Institute of Technology, 2-1-1 Shibazono, Narashino, Chiba 275-0023, Japan Received 19th March 1999, Accepted 12th May 1999 Solvent extraction and cobalt(III) oxide collection has been studied for the determination of total tin in environmental and geological samples by electrothermal atomic absorption spectrometry using a tungsten furnace. Tin iodide was extracted into benzene under acidic conditions using sulfuric acid, and cobalt(III ) oxide powder was added to the benzene to collect the extracted tin.The cobalt(III ) oxide powder was separated from benzene by vacuum filtration and then suspended in 5 ml water. Part of the slurry thus obtained was introduced into the tungsten furnace and the determination of tin performed. A relative standard deviation of 1.8% (n=6) was obtained for the determination by the proposed method of 0.20 mg l-1 tin in the slurry.The calibration curve was linear up to 0.40 mg l-1 (4.0 ng per 10 ml ) tin, and the detection limit (3 s) was 6.00 mg l-1 (0.06 ng per 10 ml ) tin. The proposed method was applied to marine sediment, fish tissue and rock as examples of environmental and geological samples. The determination of total tin obtained was in good agreement with the certified value or reference value of the environmental and geological samples.Organotin compounds are widely used to inhibit marine tellurium23 and arsenic24 by ETAAS using a tungsten furnace and reported the eVects of using cobalt(III ) oxide powder as a organisms from adhering to the bottom of ships or fishing nets. These organotin compounds are eluted by seawater and chemical modifier and collector. In this study, cobalt(III) oxide powder is used to collect tin, and pretreatment by solvent concentrate in the marine sediment or in the bodies of marine organisms having a harmful influence upon the ecosystem.extraction is performed to determine traces of tin in the environment. Recently, some organotin compounds have been specified as being endocrine disrupters. Hence, the development of an easy, precise method for the determination of tin is desired. Experimental Organotin compounds are easily decomposed into inorganic Reagents tin in the presence of light, etc.; however, in the bodies of organisms, organotin compounds remain unchanged.Thus, Tin(II) standard solution. This was prepared by diluting the chemical nature of tin in the environment is complex. AAS-grade tin solution (1000 mg l-1, SnCl2 in 6.0M HCl, Therefore, it is important to determine the amount of total Wako, Osaka, Japan) with 3 M HCl. tin in a sample, in addition to the amount of each specific chemical species of tin.Tin(IV) standard solution. This was prepared as follows: The sensitivity of ETAAS and ICP-AES for the determi- Precisely 0.301 g of tin(IV) chloride pentahydrate nation of tin is relatively high, and that of ICP-MS is very (98.0%<SnCl4 5H2O; Wako) was dissolved in 6 M HCl to high.1 However, the determination of trace amounts of tin in make a 100 ml solution, and the resulting solution was estabcomplex matrices is diYcult due to large interference from lished as the 1000 mg l-1 standard solution of tin(IV); this matrix components.Therefore, pretreatment procedures such standard solution was diluted with 3 M HCl and used in the as organic solvent extraction2–7 and other methods8–10 are experiments. necessary to separate tin from its matrix components or to Organotin compound was prepared by diluting tri-n-butyltin determine tin in each chemical species present in the matrix. chloride, triphenyltin chloride and tri-n-pentyltin chloride stan- Ni et al.11 and Ferri et al.12 have studied organic solvent dard solution (1000 mg l-1, Kanto, Tokyo, Japan) with extraction methods for tin(IV) and organotin compounds.toluene. Since tin(II) easily oxidizes to tin(IV), most previous studies Cobalt(III ) oxide powder. 30 mg of cobalt(III ) oxide powder on inorganic tins have focused on pretreatment methods and (purity 99.5%;Wako) were passed through a 200 mesh (74 mm) methods for the determination of inorganic tin(IV); there are sieve, weighed and used for the experiments.few reports on tin(II). In the case of tin determination by The acids used were of poisonous metal analysis (PMA)- ETAAS using a graphite furnace, detailed reports on chemical grade (Wako), and benzene and potassium iodide used were modifiers13–15, and their eVect and behavior during the charof gravimetric analysis grade (Kanto). Caution: benzene is ring and atomization stages,16–21 have been studied. However, both toxic and carcinogenic and care should be exercised in in the case of tin determination by ETAAS using a tungsten its use.furnace only a limited number of reports are available on the Ultrapure grade water purified with a Milli Q-Labo filter eVects of chemical modifiers. We have studied the determination of lead,22 bismuth,22 (Nippon Millipore, Tokyo, Japan) was used throughout. J. Anal. At. Spectrom., 1999, 14, 1081–1085 1081Apparatus samples by acid, the residue was dissolved by adding 5 ml of 3 M HCl, and the resulting solution was poured directly into The analysis was conducted using an SAS 7500A atomic a separating funnel. absorption spectrometer connected to a PS200A electrothermal In addition, a blank test of the acids used to dissolve the atomizer (Seiko Instruments, Chiba, Japan).25 An L-233 tin solid samples, as well as measurement of water content in the hollow cathode lamp (Hamamatsu Photonics, Shizuoka, solid samples (85 °C, 4 h), was performed to compensate for Japan) was used.A deuterium lamp was used for the back- the determined values. ground correction, and the furnace was a high capacity (50 ml ) U-type metal boat made of tungsten. Results and discussion Procedure EVect of charring temperature Five ml of sample solution (total Sn!2.0 mg), which was Both organotin and inorganic tin are present in the environacidified with 3.0 M HCl, were introduced into a separating mental samples studied.In this study, organotin was comfunnel, to which 5 ml of 5 M KI were added. To this solution pletely decomposed when the solid sample was dissolved in 10 ml of 7.0M H2SO4 were added to give 20 ml of aqueous acid; therefore, to evaluate various experimental conditions, solution (concentration of H2SO4: 3.5 M). An equal volume inorganic tin was used. of benzene was added to this aqueous solution, and the funnel According to the procedure described above, a slurry was was placed on a shaker for 5 min to extract tin.After shaking, prepared using aqueous solutions which contained 1.0 mg of the benzene was separated from the aqueous solution into a either tin(II ) or tin(IV). The eVect of charring temperature on 100 ml PTFE beaker, to which 3 ml of 4-methyl-2-pentanone the slurry (1.0 mg per 5 ml=0.2 mg l-1) during tin atomization (MIBK) and 30 mg of cobalt(III) oxide powder were added. was studied when the atomization temperature was set at The beaker was placed in an ultrasonic bath, and the organic 2500 °C and the charring temperature was varied from 400 to phase was stirred for 10 min to collect tin that was adsorbed 1600 °C.At the same time, the following solutions were used onto the cobalt(III) oxide powder. Vacuum filtration was then for comparison. (i) 0.20 mg l-1 tin(II ) or tin(IV) standard performed using a PTFE membrane filter (diameter, 25 mm; solution (3 M HCl ); (ii) 5 ml of 0.20 mg l-1 tin(II) or tin(IV) pore size, 3.0 mm) to separate cobalt(III) oxide powder from standard solution into which 30 mg of cobalt(III) oxide powder the organic phase.The entire membrane filter including the was dispersed. The results obtained are shown in Fig. 1, and cobalt(III) oxide powder was inserted into a plugged test tube, the absorption signals of tin obtained using these solutions to which 5 ml of water were added. The test tube was shaken are shown in Fig. 2. by hand to obtain homogeneously dispersed slurry.Part of The absorbance was almost constant regardless of charring the slurry was introduced into the furnace, and the absorbance temperature, when tin(II) or tin(IV) standard solution was of tin (atomized under the conditions shown in Table 1) was used. In addition, the absorbance of tin(II) is similar to that measured. Manual pipetting was employed for injecting the of tin(IV). Sensitivity was also similar regardless of valence. slurry. In cases of tin standard solutions into which cobalt(III) oxide powder was dispersed, the absorbance obtained during Acid decomposition of solid samples atomization of tin(II) corresponds to that of tin(IV), which was also similar to that of the standard solutions.No influence To dissolve the solid sample, a mixed solution of HNO3, of cobalt(III) oxide powder on background absorption was HClO4 and HF was used. A precisely weighed solid sample observed. (from 0.5 to 3.0 g) was transferred into a PTFE beaker to On the other hand, in slurries prepared using the proposed which the mixed solution consisting of HNO3 (30 ml ), HClO4 method, the absorbance was lower than that of standard (10 ml ) and HF (3 ml ) was added.After the beaker was left solutions treated with cobalt(III ) oxide powder, due to the to stand for 15 min, it was placed on a hot plate (230 °C) to influence of iodide which was adsorbed onto the cobalt(III) dissolve the solid sample; the sample was then solidified oxide powder at charring temperatures of 1000 °C or less.through evaporation. If a sooty residue was obtained or the However, the absorbance almost reached that of the standard decomposition of the organic substance was incomplete, solutions with cobalt(III ) oxide powder at charring tempera- additional nitric acid was added for complete decomposition. The residue was then dissolved in 10 ml HCl (1+10) solution, and the resulting solution was poured into a measuring flask, to which 12.5 ml of concentrated HCl and a suYcient amount of water were added to obtain 50 ml of the solution.The molarity of HCl in the obtained solution was set at 3 M for convenience. If the amount of tin contained in the solid samples was extremely small after decomposition of the solid Table 1 Instrumental operating parameters for tin Parameter Ramp time/s Hold time/s Temperature/°C Dry 10 20 130 Char 10 15 1400 Atomize 0 2 2500 Clean 0 1 2600 Wavelength 224.6 nm Spectral bandwidth 0.5 nm Fig. 1 EVect of charring temperature. +,6: Sn standard solution Lamp current 20 mA (0.20 mg l-1); &,%: 5 ml of Sn standard solution (0.20 mg l-1) Gas flow rate Ar: 5.0 l min-1 containing dispersed cobalt(III ) oxide powder; $,#: slurry (Sn: 1.0 mg H2: 1.0 l min-1 per 5 ml ) of the proposed method; +,&,$: SnII; 6,%,#: SnIV; Injection volume 10 ml injection volume: 10 ml. 1082 J. Anal. At. Spectrom., 1999, 14, 1081–1085Fig. 2 Atomic absorption profiles of tin on atomization.(A): SnII Fig. 3 Recovery of SnII and SnIV from extraction process as a function standard solution; (B): SnIV standard solution; (C): SnII standard of H2SO4 concentration. Organic solvent: benzene, #,%: without KI; solution containing dispersed cobalt(III ) oxide powder; (D): SnIV $,&,6: with KI;$,#: SnII;&,%: SnIV;6: SnII in 3 MHCl aqueous standard solution containing dispersed cobalt(III ) oxide powder; (E): solution; cobalt(III) oxide powder: 30 mg; concentration of KI: 1.25 M.slurry of SnII for the proposed method; (F): slurry of SnIV for the proposed method; (G): blank of the proposed method; charring temperature: 1400 °C; concentration of Sn: 0.20 mg l-1. recovery of tin(II) increased to 100% (6). For HCl concentrations of 2.5 M or less and HClO4 concentrations of 2.0–4.0 M, the recovery was 70–98%. Tin is extracted by benzene in tures of 1300–1500 °C. The measurement of the slurry H2SO4 solution in the form of SnI4, upon the addition of absorbance was repeated (n=6) at charring temperatures of potassium iodide.Therefore, when tin(II) is present in an 1300–1500 °C. The results indicate that the highest reproducaqueous solution, recovery is considered to increase through ibility was observed at a charring temperature of 1400 °C, oxidation of tin(II) to tin(IV) by the addition of an oxidizing where the relative standard deviation (RSD) was 1.8%. The agent. However, in this study, recovery was increased upon optimal charring time was determined to be 15 s.the addition of potassium iodide, a reducing agent, to the HCl solution. EVect of atomization temperature When arsenic iodide was extracted by an organic solvent, the following phenomena were confirmed: the production of The eVect of atomization temperature on the absorbance of slurries containing tin was examined when the charring tem- free iodine was promoted with increasing HCl concentration, arsenic(III) was oxidized to arsenic(V) due to the I2 produced, perature was set at 1400 °C and the atomization temperature was varied from 2000 to 2700 °C.When the atomization and arsenic was extracted by the solvent in the form of arsenic(V).26 Similar to the phenomena observed in arsenic, temperature was 2300 °C or less, wide peak signals were obtained. As the atomization temperature increased, the width free iodine may promote the oxidation of tin(II ) to tin(IV) through the addition of potassium iodide to the HCl solution.of the peak signal decreased and the absorbance increased sharply; the highest absorbance was obtained at an atomization With respect to tin(IV), the recovery was constant regardless of when potassium iodide was added, however, the extraction temperature of 2700 °C (RSD=4.3%, n=6). However, RSD for a repeated 6 measurements was minimal (1.8%) at an process could not be clarified. Based on the above results, the concentration of HCl in the atomization temperature of 2500 °C.The optimal atomization time was determined to be 2 s. aqueous solution was set at 3 M, and the concentration of H2SO4 maintained at 3.5 M after potassium iodide was added. With respect to other measurement conditions, detailed analysis was performed and the optimal conditions shown in Table 1 were determined. Selection of organic solvent The following three organic solvents used for tin extraction, Extraction conditions of tin the specific gravities of which are lower than that of water, were studied: (i) diethyl ether, (ii) benzene and (iii) MIBK.Aqueous solutions containing 1.0 mg of tin(II) or tin(IV) and 1.25 Mpotassium iodide were acidified with 2.0–4.0 Msulfuric Fig. 4 shows the recovery of tin(II ) and tin(IV) at various concentrations of H2SO4 with a constant concentration of acid. The solutions were diluted to give a final volume of 20 ml. Tin was then extracted by benzene (20 ml ) and the 1.25 M of potassium iodide.In the experiment, the ratio of the aqueous to the organic phase was 151. extract treated with cobalt(III) oxide powder to recover the tin. Fig. 3 shows the results. The recovery of tin(II ), obtained with H2SO4 concentrations of 2.5–4.0M was 100%, when benzene was used. Recoveries When potassium iodide was added to the aqueous solution, the recovery of tin(II) for H2SO4 concentrations of 2.5–4.0 M of 100% and 62%, respectively, were obtained when diethyl ether was used with 2.5M H2SO4 and MIBK with 2.0M was approximately 80% ($).However, tin(IV) was extracted by benzene in the form of SnI4, and the recovery for H2SO4 H2SO4. Similarly, the recovery of tin(IV), obtained with H2SO4 concentrations of 2.0–4.0M was 100% (&). As explained above, the recovery of tin(II) diVered from concentrations of 2.0–4.0 M was 100%, when benzene was used. Recoveries of 69% and 38%, respectively, were obtained that of tin(IV) when potassium iodide was added to H2SO4 solutions containing tin.Based on the above results, we first when diethyl ether and MIBK were used with 2.0M H2SO4. Based on these results, it was found that a recovery of 100% acidified the aqueous solution with 1.0–6.0 M HCl or 2.0–4.0 M HClO4, and then controlled H2SO4 concentration could be obtained with benzene in the presence of H2SO4 at a wide concentration range (2.5–4.0 M) for both tin(II ) and after potassium iodide was added.As a result, by adding potassium iodide to 2.0–6.0M HCl and then controlling tin(IV). However, MIBK was superior to benzene in terms of operation since MIBK has higher viscosity and loss of H2SO4 concentrations of the solution at 2.5–4.0 M, the J. Anal. At. Spectrom., 1999, 14, 1081–1085 1083added to 50–200 ml of benzene containing 1.0 mg of both tin(II) and tin(IV), the recovery was 100%, indicating that 30 mg of cobalt(III) oxide powder can also be eVective when the volume of benzene is large.EVect of coexisting ions The eVect of coexisting ions on the recovery was studied in order to apply the proposed method to environmental samples. Various kinds of ions were added to the aqueous solution containing 1.0 mg of tin(II) or tin(IV). According to a previously determined procedure, a slurry was prepared by solvent extraction and collection of tin by cobalt(III) oxide powder. Ten ml of the slurry was introduced into the furnace and absorbance during atomization was measured.The obtained results were compared with absorbance obtained without coexisting ions, and recovery was determined. Fig. 4 Recovery of SnII and SnIV from extraction process using various To determine the amount of coexisting ions added, those organic solvents in the case of H2SO4 system. $,#: benzene; &,%: contained in 2 g of rock or sediment samples were used. diethyl ether; +,6: MIBK; $,&,+: SnII in 3 M HCl aqueous Chlorides of each ion were used as cations to be added.Table 2 solution; #,%,6: SnIV in 3M HCl aqueous solution; cobalt(III ) oxide summarizes the results. Because of pretreatment which powder: 30 mg; concentration of KI: 1.25 M. included solvent extraction and collection using cobalt(III) oxide powder, the recovery was 100±2%, regardless of the cobalt(III) oxide powder was suppressed during the recovery quantity of coexisting ions. operation such as a vacuum filtration. For these reasons, we employed the following procedure in this study: tin was first Calibration curve extracted by benzene to which 3 ml of MIBK was added; then, Calibration curves were obtained using slurries prepared tin was collected using cobalt(III) oxide powder.according to the determination procedure using an aqueous solution containing tin(II) or tin(IV). The results indicate that EVect of concentration of potassium iodide the calibration curve of tin(II) corresponds to that of tin(IV).Potassium iodide (0.1–2.0 M) was added to the aqueous The calibration curve was linear up to 0.40 mg l-1 (4.0 ng per solution containing 1.0 mg of tin(II) or tin(IV) to give 20 ml of 10 ml ), and the detection limit (3 s) was 6.00 mg l-1 (0.06 ng 3.5 M H2SO4 solution. per 10 ml ). A recovery of 100% was obtained in the presence of potass- When the calibration curves are compared with the standard ium iodide concentrations of 0.1–2.0 M. At this concentration solution, the absorbance of the slurry at tin concentrations of range, potassium iodide concentration did not aVect both the 0.25 mg l-1 (2.5 ng per 10 ml ) or higher is slightly lower than absorption signal and the background absorption, obtained that of the standard solution; the slopes of the two calibration by tin atomization during ETAAS using a tungsten furnace.Therefore, we set the potassium iodide concentration at 1.25 Table 2 EVect of coexisting ions on the determination of tin(II) M.and tin(IV) EVect of extraction time Found/mg l-1 bc Recovery (%) Approximately 100% of tin iodide can be extracted by benzene Ion Added/mga SnII SnIV SnII SnIV in approximately 1 min, when the ratio of the aqueous to the organic phase is 151. In this study, in order to completely SnII/SnIV — 200.0 200.0 — — extract tin in a single extraction operation, the extraction time NaI 150 197.4 198.9 99 99 KI 150 199.8 200.7 100 100 was set at 5 min. MgII 150 200.1 200.0 100 100 CaII 150 197.6 198.3 99 99 Collecting condition on cobalt(III ) oxide powder SrII 1 199.6 198.6 100 99 BaII 5 197.4 199.5 99 100 Tin(II) and tin(IV) were extracted into benzene; then 30 mg of MnII 3 199.9 197.4 100 99 cobalt(III) oxide powder were added.The thus-prepared sample CoII 0.1 200.0 200.1 100 100 was stirred for various time periods and recovery carried out. NiII 0.1 201.3 200.8 101 100 A recovery of approximately 70% was obtained for a stirring CuII 0.5 199.8 200.6 100 100 time of 1 min.Recovery was increased with increased stirring ZnII 0.2 201.1 199.7 101 100 CdII 0.01 199.9 200.2 100 100 time; a recovery of 100% was obtained with a stirring time of PbII 0.1 200.1 200.0 100 100 5 to 30 min. Considering the requirements for rapid operation, AlIII 200 199.4 202.6 100 101 the stirring time was determined to be 10 min. CrIII 0.2 202.8 201.4 101 101 Five to 50 mg of cobalt(III) oxide powder was added to FeIII 200 197.3 196.9 99 98 20 ml of benzene containing both tin(II) and tin(IV), and the AsIII 0.1 200.0 199.9 100 100 resulting organic phase was stirred for 10 min to obtain the SbIII 0.1 200.1 199.9 100 100 BiIII 0.1 200.1 200.0 100 100 recovery.The recovery was 100% when 10–50 mg of cobalt(III) TiIV 20 196.9 198.2 98 99 oxide powder was added; it was approximately 80% when SeIV 0.1 199.9 199.8 100 100 5 mg of cobalt(III) oxide powder was added. In this study, VV 1 200.3 201.8 100 101 since the volume of the slurry was set at 5 ml, an insuYcient MoVI 0.01 201.3 202.2 101 101 amount of cobalt(III) oxide powder was present in the slurry SiO32- 500 196.7 197.6 98 99 due to the decrease in the solid: water ratio; the organic phase PO43- 3 198.6 199.7 99 100 contained no tin. The amount of cobalt(III) oxide powder was aSample volume, 20 ml; bslurry, 5 ml; cinjection volume, 10 ml.set at 30 mg. When 30 mg of cobalt(III) oxide powder was 1084 J. Anal. At. Spectrom., 1999, 14, 1081–1085Table 3 Recovery of organotin compounds Organotin compound Taken/mga Found/mg l-1 bc Recovery (%) Tri-n-butyltin 1.0 200.0 100 0.5 100.5 101 Triphenyltin 1.0 201.6 102 0.5 99.7 100 Tri-n-pentyltin 1.0 199.8 100 0.5 99.4 99 Taken/mga Tri-n-butyltin Triphenyltin Tri-n-pentyltin Found/mg l-1 bc Recovery (%) 0.2 0.3 0.5 101.5 102 0.3 0.5 0.2 98.9 99 0.5 0.2 0.3 100.2 100 aAmounts as tin; bslurry, 5 ml; cinjection volume, 10 ml.Table 4 Analytical results for total tin in environmental and are used as pretreatment.The method can also determine the geological samples amount of inorganic tin with a diVerent valence number using the same determination procedure. The current pretreatment Mean/mg g-1 Reference value or method is proven to be eVective in the determination of tin, Sample ± (n=6) certified value/mg g-1 including organotin, in environmental samples. JG-1a 4.43±0.06 4.47 (Granodiorite) References JB-3 0.94±0.01 0.94 (Basalt) 1 A. B.Pandyn and J. C. Van Loon, Fresenius’ Z. Anal. Chem., JF-1 0.25±0.02 0.3 1988, 331, 707. (Feldspar) 2 H. N. Elsheimer, Anal. Sci., 1993, 9, 681. NIES No. 11 2.35±0.01 2.4±0.1 3 H. N. Elsheimer and T. L. Fries, Anal. Chim. Acta, 1990, 239, 145. (Fish tissue) 4 R. Pinel, M. Z. Benabdallah, A. Astruc and M. Astruc, J. Anal. NIES No. 12 9.38±0.04 10.7±1.4a At. Spectrom., 1988, 3, 475. (Marine sediment) 5 M. Chamsaz and J. D. Winefordner, J. Anal. At. Spectrom., 1988, 3, 119. aRef.No.: 27. 6 K. Tanaka, Bunseki Kagaku, 1962, 11, 332. 7 S. Terashima, Bull. Geol. Surv. Jpn, 1985, 36, 375. 8 E. Lundberg and B. Bergmark, Anal. Chim. Acta, 1986, 188, 111. curves do not agree. For this reason, in this study, we used 9 K. Ide, S. Hashimoto and H. Okochi, Bunseki Kagaku, 1995, the calibration curve prepared using the slurry. 44, 617. 10 J. Kuballa, R. D. Wilken, E. Jantzen, K. K. Kwan and Y. K. Acid decomposition and recovery of organotin Chau, Analyst, 1995, 120, 667. 11 Z.-M. Ni, H.-B. Hang, A. Li, B. He and F.-Z. Xu, J. Anal. At. Since organotins are contained in the environmental samples Spectrom., 1991, 6, 385. used in this study, the current pretreatment method was 12 T. Ferri, E. Cardarelli and B. M. Petronio, Talanta, 1989, 36, 513. applied to some organotin compounds. According to pro- 13 R. Pinel, M. Z. Benabdallah, A. Astruc and M. Astruc, Anal. cedure described above, recovery of tin was obtained through Chim. Acta, 1986, 181, 187. 14 M. Taga, H. Yoshida and O. Sakurada, Bunseki Kagaku, 1987, acid decomposition, solvent extraction and collection of tin 36, 597. using cobalt(III ) oxide powder. Recovery was almost 100% for 15 M. Tominaga and Y. Umezaki, Anal. Chim. Acta, 1979, 110, 55. every organic compound used, which means that the current 16 K. Yasuda, Y. Hirano, T. Kamino and K. Hirokawa, Anal. Sci., method can also be applied to environmental samples 1994, 10, 623. (Table 3). 17 K. Yasuda, Y. Hirano, K. Hirokawa, T. Kamino and T. Yaguchi, Anal. Sci., 1993, 9, 529. 18 Y. Terui, K. Yasuda and K. Hirokawa, Anal. Sci., 1991, 7, 599. Application to environmental and geological samples 19 K. Oishi, K. Yasuda and K. Hirokawa, Anal. Sci., 1991, 7, 883. The proposed method was applied to environmental samples 20 M. Taga, O. Sakurada and H. Takahashi, Bunseki Kagaku, 1988, 37, 164. from the National Institute for Environmental Studies (NIES) 21 W.-M. Yang and Z.-M. Ni, Spectrochim. Acta, Part B, 1997, 52, and geological standard rock samples from the Geological 241. Survey of Japan (GSJ). 22 T. Narukawa, W. Yoshimura and A. Uzawa, Bunseki Kagaku, First, these samples were precisely weighed, and one of the 1998, 47, 707. following procedures was carried out: (i) dissolution in 5 ml 23 T. Narukawa, J. Anal. At. Spectrom., 1999, 14, 75. of 3 M HCl solution, or (ii) dissolution in a suYcient amount 24 T. Narukawa, W. Yoshimura and A. Uzawa, Bull. Chem. Soc. Jpn., 1999, 72, 701. of acid to make 50 ml (3 M HCl solution) and 5 ml of the 25 T. Narukawa, A. Uzawa, W. Yoshimura and T. Okutani, J. Anal. resulting solution was used for a later experiment. At. Spectrom., 1997, 12, 781. The results revealed that values obtained in this study agree 26 S. Tagawa, Bunseki Kagaku, 1980, 29, 563. with the certified or reference value reported for each standard 27 J. Yoshinaga, H. Kon, T. Horiguchi, M. Morita and K. Okamoto, sample (Table 4). Anal. Sci., 1998, 14, 1121. Use of the proposed method enables the determination of trace amounts of tin in complex matrices when solvent Paper 9/02207D extraction and collection of tin using cobalt(III ) oxide powder J. Anal. At. Spectrom., 1999, 14, 1081–1085 1085

ISSN:0267-9477    

DOI:10.1039/a902207d

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

INTER-LABORATORY NOTE Determination of rare earth elements and yttrium in rocks by inductively coupled plasma atomic emission spectrometry after solvent extraction with a mixture of 2-ethylhexyl dihydrogenphosphate and bis(2-ethylhexyl ) hydrogenphosphate P. K. Srivastava* and A. Premadas Chemical Laboratory, Central Region, Atomic Minerals Division, Civil Lines, Nagpur-440001, India Received 22nd February 1999, Accepted 14th May 1999 The importance of rare earth elements in petrogenetic studies is well known.The inductively coupled plasma atomic emission spectrometry (ICP-AES) technique is mostly used for the determination of rare earth elements after its separation by the ion exchange method. An alternative, rapid, solvent extraction method involving a mixture of (2-ethylhexyl ) dihydrogenphosphate and bis(2-ethylhexyl ) hydrogenphosphate in kerosene suitable for the separation and preconcentration of REE in silicate rock is described. Major elements, such as Mg, Ca, Mn, Na and K, are not extracted appreciably from dilute hydrochloric acid solution.The extraction of the major fraction (>95%) of iron is prevented by its reduction with ascorbic acid. Titanium, Th, U and Zr are also extracted quantitatively and Al is extracted partially (~6%). The REE are selectively stripped from the organic phase after the addition of tri-n-butyl phosphate with 4.5 mol dm-3 hydrochloric acid. Major fractions of co-extracted interfering elements (residual Fe, Th, U, Zr and Ti) remain in the organic phase.The aluminium is also stripped along with the REE; however, aluminium does not cause any interference in the determination of REE by ICPAES. The method is applied for the preconcentration of REE in silicate rocks and its validity is checked by analyzing some Standard Reference Materials. The results are in good agreement with the recommended value. The method is simple, rapid, and suitable for the determination of REE in a silicate rock sample.exchange separation using a cation exchanger. The ion Introduction exchange procedure using hydrochloric acid elution tends to The rare earth elements (REE) form a coherent group of eluate La and Ce with substantial tailing. Also, in the presence elements of geochemical significance and their distribution of large amounts of Al and Fe, the use of 130 ml of 2 mol dm-3 in rock samples is important in petrogenetic studies.1,2 nitric acid as eluting agent often gives low recovery of the Various methods, such as isotopic dilution mass spectrometry middle REE(Sm, Eu and Gd).13 As a modification, a combi- (ID-MS),3,4 instrumental neutron activation analysis nation of oxalate precipitation and ion exchange separation is (INAA),4–6 inductively coupled plasma mass spectrometry suggested for the preconcentration of REE in geological (ICP-MS)7–11 and inductively coupled plasma atomic emission materials.22 Although cation exchange is reliable and accurate, spectrometry (ICP-AES)12–21 are available for the determi- an alternative, economical, simple and expeditious method is nation of REE.These methods can analyse REE at the low needed for the separation and preconcentration of REE. levels that are normally present in silicate rock. The first three Solvent extraction with tri-n-octyl phosphine oxide has been methods require costly equipment or a reactor facility (for also used for the group separation of REE because of its INAA) and most of the laboratories in developing countries rapidity.27 However, this method needs careful control of are not equipped with these instruments.Therefore, ICP-AES pH and treatment of solution after the extraction. is the widely used method for the determination of REE due Bis(2-ethylhexyl ) hydrogenphosphate (HDEHP, a diester) and to its simplicity and low cost of the equipment. However, a 2-ethylhexyl dihydrogenphosphate (H2MEHP, a monoester) separation and preconcentration of REE and Y are needed to are strong complexing agents and quantitatively extract the eliminate spectral and background continuum interferences REE from a low acid solution.28–30 A mixture of HDEHP and due to major matrix elements and to bring their concentration H2MEHP was used for the extraction and preconcentration above the determination limits of ICP-AES.of REE in sea-water and the REE were stripped from the Various group separation methods, such as precipitation as organic phase diluted with octyl alcohol using hydrochloric oxalate22 or fluoride23 using Ca as carrier, ion exchange using acid and determined by ICP-MS.31 In spite of better extraction cation12–17 and anion24 exchanger, solvent extraction using eYciency of H2MEHP and HDEHP, not much work is TTA25 and organophosphorus compounds,26 have been used.reported for the separation of REE from a silicate rock. This The separation of REE by precipitation methods yields an may be probably due to lack of selectivity and diYculty in unacceptably high salt load, whereas the REEs fraction complete stripping of Tm, Yb and Lu32 from the organic obtained after solvent extraction or ion exchange procedure is phase.A partial or complete extraction of FeIII, Ti, U, Th, Zr, suitable for ICP-AES analysis. The most frequently used method for the separation and preconcentration is ion etc., is reported.26 However, the reagents can be successfully J.Anal. At. Spectrom., 1999, 14, 1087–1091 1087used if the concentration of the interfering elements, stated Matthey, Royston, Hertfordshire, UK) by dissolution in hydrochloric acid. A combined reference solution was made above, can be reduced considerably, either at the extraction or at the stripping stage. Accordingly, the solvent extraction up containing: 15 mg ml-1 each of La, Pr and Sm; 25 mg ml-1 each of Ce and Nd; 2 mg ml-1 each of Eu, Ho, Tm, Yb, Lu studies involving a mixture of H2MEHP and HDEHP for the group separation of REE forming major, minor and trace and Y; and 5 mg ml-1 each of Gd, Dy, Tb, Er.It was prepared from the standard solution by dilution, maintaining a interfering elements present in the silicate rock samples were taken up for detailed investigation. 1 mol dm-3 hydrochloric acid concentration. For the calibration this reference solution was further diluted to 5-, 10-, Optimization of the extraction and stripping parameters of REE and Y in silicate rock are presented.A mixture of 25- and 50-fold with 0.5 mol dm-3 HCl. H2MEHP and HDEHP in kerosene is suitable for the quanti- Reagents tative extraction of REE and Y from an acidic solution (~0.12 mol dm-3 HCl) of silicate rock samples. Ascorbic acid Bis(2-ethylhexyl) hydrogenphosphate (Fluka, AG, Buchs, prevents the extraction of FeIII by its reduction to the FeII. Switzerland). The mixture contains 50% each H2MEHP and The REE are selectively stripped from the organic phase HDEHP (referred to as reagent A).containing interfering elements using 4.5 mol dm-3 hydrochloric acid in the presence of tri-n-butyl phosphate (TBP). 2-Ethylhexyl hydrogenphosphate (E. Merck, Darmstadt, Uranium, Ti, Th, Zr, and the residual extracted iron remain Germany). The mixture contains ~45% H2MEHP and ~55% in the organic phase. Aluminium is also partially extracted HDEHP (referred to as reagent B). A 10% solution of reagent and stripped along with REE. However, the quantity of it A or a 12.5% solution of reagent B was prepared in kerosene, present in the rare earths fraction does not interfere in the and it was equilibrated with 6 mol dm-3 hydrochloric acid determination by ICP-AES. The method is simple, rapid, and washed with 0.12 mol dm-3 HCl before use. The solution precise, accurate and suitable for the separation and preconcen- of the former corresponds to ~0.24 mol dm-3 H2MEHP and tration of the rare earth elements in silicate rock (for the ICP- ~0.16 mol dm-3 HDEHP (the sum of the molarity of the two AES determination).It can be applied to larger number of esters is equal to 0.4 mol dm-3 ), while the latter corresponds samples simultaneously, as compared with the ion exchange to ~0.26 mol dm-3 H2MEHP and ~0.21 mol dm-3 HDEHP procedure. (the sum of the molarity of two esters is equal to ~0.47 mol dm-3). The solutions of reagent A and reagent B were used for the Experimental initial optimisation studies for the extraction behaviour of Instrumentation REE and Y in a synthetic mixture and in rock samples, respectively. An Integra Model XM ICP-AES sequential spectrometer from GBC Australia (Dandenong, Victoria, Australia) was used.Bis(2-ethylhexyl) hydrogen phosphate (E. Merck). The instrumental parameters used are given in Table 1. Two Tri-n-butyl phosphate (Merck Ltd., Poole, Dorset, UK). replicate measurements were made for each element.For all All other reagents used were of either general reagent or lines the built-in AUTO background correction software was AnalaR grade. used. Sample decomposition Preparation of standards To a 1–1.5 g sample in a platinum dish 5 ml nitric acid, 10 ml The REE stock standard solution of 1000 mg ml-1 was prehydrofluoric acid and 2 ml of perchloric acid are added and pared from high purity oxides (99.99 or 99.999%, Johnson mixed with a polytetrafluoroethylene (PTFE) rod.The dish is placed in a water-bath and the contents are evaporated to incipient dryness. The above process is repeated twice. Then, Table 1 Operating parameters for ICP-AES system the dish is transferred to a sand-bath and the contents are evaporated to white fumes of perchloric acid; fuming is RF generator 40.68 MHz (crystal controlled) Forward power 1200 W continued until the complete termination of the evolution of Reflected Power <20 W white fumes. The dry residue is dissolved by warming in 25 ml of 4 mol dm-3 HCl.The solution is boiled and cooled. It is Gas flow 10 l min -1 (coolant) 0.6 l min -1 (sample) filtered through a Whatman No. 540 filter-paper and washed 0.3 l min -1 (auxiliary) with 1 mol dm-3 HCl. The residue, if any, is placed in a platinum crucible (30 ml capacity) and the filter paper is Monochromator Modified Czerny–Turner Focal length 750 mm ignited completely on a burner. The residue is fused with 1 g Light path medium Vacuum of sodium carbonate and a pinch of boric acid.The fused DiVraction grating 1800 grooves mm-1 mass is next dissolved in 25 ml of 4 mol dm-3 HCl having Ruled area 52 mm×52 mm been placed separately in another beaker. Then, 5 mg of FeIII Wavelength range 160–820 nm are added as the chloride and the REE are precipitated as hydroxides using dilute (151) ammonia solution. The precipi- Wavelength resettability ±0.002 nm tate is filtered through 540 filter paper, re-dissolved in 25 ml Dispersion 0.74 nm mm -1 (first order) of 4 mol dm-3 HCl and mixed with the original solution.Nebuliser Concentric Extraction Solution uptake 1.7 ml min -1 Slits 20 mm entrance The sample solution in the beaker is evaporated to dryness on 40 mm exit a water-bath and the residue is dissolved in 2 ml of 6 mol dm-3 Detectors Dual photomultipliers HCl and 20 ml of water. The solution is boiled briefly to PMT voltage 675 V obtain a clear solution and then cooled.FeIII is reduced to FeII Flush time 10 s (indicated by disappearance of the yellow colour) by the Integration time 3 s addition of 10 ml of 10% ascorbic acid solution and transferred Observation height 12 mm above load coil into a separating funnel previously marked at ~100 ml. The 1088 J. Anal. At. Spectrom., 1999, 14, 1087–1091volume of the aqueous phase is diluted to ~100 ml with water. Results and discussion Then, 10 ml of extracting solution mixture (0.47 mol dm-3 of Selection of proper combination of H2MEHP and HDEHP for reagent B) are added and the REE and Y are extracted by the extraction shaking for 5 min and allowing the layers to separate.The aqueous layer is transferred into another separating funnel Bis(2-ethylhexyl ) hydrogenphosphate (HDEHP) and containing 10 ml of extracting solution. Any aqueous droplets 2-ethylhexyl dihydrogenphosphate (H2MEHP) are used extenleft in the first separating funnel are also transferred with sively for the separation of actinides and lanthanides.26 Various 5–10 ml of 0.12 mol dm-3 HCl.The extraction process is factors, such as the concentration of the reagent, the acidity repeated to extract residual REE and Y and the aqueous layer of the aqueous phase, the oxidation state and the ionic radii is discarded. Both the organic layers are mixed and washed of metal ions, aVect the extraction of metal ions. A dilute briefly (30 s) once with 25 ml of 0.12 mol dm-3 HCl.solution (0.1 mol dm-3) of diester can extract heavier rare earth elements (HREE) quantitatively from an aqueous solu- Stripping of REE and Y tion of pH ~2. However, the extraction of lighter rare earth elements (LREE) is poor32 and the stripping of Yb and Lu is A 5 ml volume of TBP solution is added to the organic phase also diYcult. The major problem in silicate rock is to keep all and the REE are back-extracted with 10 ml of 4.5 mol dm-3 of its components in solution at pH ~2.Therefore, we need HCl (2 min shaking, twice). The aqueous phases are collected an extracting solution that can extract the REE from an acidic in a beaker and dried on a hotplate. Traces of organic matter solution having a pH around one. Accordingly, we took up are removed by HNO3 (10 ml ) and HClO4 (1 ml ) treatment. the solvent extraction studies involving a higher concentration The perchloric acid is completely fumed oV and the residue is of diester and a mixture containing mono- and diester in dissolved in 25 ml of 0.5 mol dm-3 HCl.varying combinations. The sum of the molarity of the two esters was maintained at 0.4 mol dm-3 in the extracting solu- ICP-AES determination tion. The results of the recoveries for LREE are shown in The wavelengths used and detection limits found under the Fig. 1. It was observed that the extraction of LREE was not experimental conditions are given in Table 2. Correction for quantitative, even with higher concentration of diester alone spectral interference of Ce and Nd on Pr was applied.All the (0.4 mol dm-3). As the concentration of monoester was REE and Y are analysed in one programme sequentially. increased the recovery of LREE was also increased, and was almost quantitative when the molar concentration of the Solvent extraction studies with synthetic mixture of REE and Y monoester in the mixture was equal to or greater than the diester. For the extracting solution containing a monoester A combined working solution containing 600 mg of La, 1300 mg concentration less than the diester, the recovery of LREE of Ce, 130 mg of Pr, 800 mg of Nd, 150 mg of Sm, 5 mg each of increased with the increase in atomic number. The extraction Eu, Tm and Lu, 70 mg of Gd, 15 mg each of Tb and Ho, 55 mg of HREE was quantitative in all the cases studied.of Dy, 25 mg each of Er and Yb and 200 mg of Y in dilute This study indicates that the monoester is a more powerful hydrochloric acid is used for the solvent extraction studies.extracting agent for LREE than the diester. Probably, diester The parameters studied are as follows: selection of a suitable acts as a polar diluent for monoester by decreasing its degree combination of H2MEHP and HDEHP; concentration of of polymerization.26 Therefore, any commercially available extracting reagent; aqueous phase acidity; and back extracmixture having a monoester concentration more than 45% can tion.The solution is diluted to ~100 ml, maintaining be used for the extraction of lanthanides and Y. Thus, reagent ~0.12 mol dm-3 HCl. The REE are extracted once with the A or reagent B can be selected for the solvent extraction 10 ml of extracting solution and stripped twice with of REE. 4.5 mol dm-3 hydrochloric acid after the addition of TBP (0–1.07 mol dm-3). The organic matter is destroyed as EVect of extracting reagent concentration described above and the REE are determined after suitable dilution by ICP-AES.The extracting reagent concentration is another important factor for the complete extraction of REE. Solvent extraction studies with various concentrations of extracting reagent A Table 2 Spectral lines used for emission measurement and the detec- between 0.1 and 0.4 mol dm-3 (the sum of the molarity of the tion limits two esters is given for the sake of simplicity) were carried out. The results show that all the rare earth elements and yttrium Crustal are quantitatively extracted with a reagent concentration equal Detection limit/ abundance/ Element Wavelength/nm mg l -1 mg g -1a or greater than 0.2 mol dm-3.Considering the presence of La 333.749 6 19 Ce 418.660 22 38 Pr 422.293b 50 4.3 Nd 430.358 45 16 Sm 442.434 23 3.7 Eu 381.967 16 1.1 Gd 342.247 7 3.6 Tb 350.917 15 0.64 Dy 353.170 6 3.7 Ho 345.600 10 0.82 Er 349.910 10 2.3 Tm 346.220 10 0.32 Yb 328.937 2 2.3 Lu 261.541 1 0.3 Y 371.170 1.5 20 aS.R.Taylor, and S.M.McLennan, Phil.Trans. R. Soc., London. A, 1981, 301, 381. 120 100 80 60 40 20 0 0.15 0.25 0.20 0.20 0.25 0.15 0.30 0.10 0.35 0.05 0.40 0.00 La Concentration of HDEHP (top) and H2MEHP (bottom)/mol dm–3 Ce Pr Nd Sm Recovery (%) Eu bCorrection for spectral interference due to Ce and Nd is applied. Fig. 1 EVect of H2MEHP on the extraction of LREE. J. Anal. At. Spectrom., 1999, 14, 1087–1091 1089other extractable elements (Fe, Al, Ti, Zr, etc.) in a silicate rock sample, the concentration of the extracting reagent is selected as 0.40 mol dm-3 for reagent A or 0.47 mol dm-3 for reagent B.However, a minimum of two extractions are needed for the complete recovery of REE in a silicate rock. At a higher concentration of the extracting reagent (>0.5 mol dm-3 ), the complete back-extraction of Tm, Yb, and Lu is diYcult. EVect of acid concentration The concentration of the acid used for solution preparation plays an important role in the extraction of REE.A better Fig. 2 EVect of TBP on the back extraction of Er, Tm, Yb and Lu separation factor is reported in a chloride medium in comparison with nitrate and sulfate media.32 Therefore, dilute hydrochloric acid is selected as a medium due to its better ability to dissolve most of the major cations in the silicate rock. We have carried out the solvent extraction studies by varying the hydrochloric acid concentration (0.12, 0.36 and 0.60 mol dm-3). The REE mixture was extracted with a solvent containing 0.4 mol dm-3 of reagent A.The results show that the REE and Y can be extracted quantitatively in single extraction at lower acidity (~0.12 mol dm-3 HCl). At higher acidity (>0.36 mol dm-3) the extraction of LREE was incomplete. At an acidity <0.08 mol dm-3 a tendency to 2.5 2.0 1.5 1.0 0.5 0 0 0.2 0.4 0.6 0.8 1.0 1.2 Concentration of TBP/mol dm–3 Mass of non-REE/mg Fe Ti U emulsion formation was observed. So, the acidity of Fig. 3 EVect of TBP on the back extraction of Fe, Ti and U.~0.12 mol dm-3 HCl is fixed for the extraction of REE from a rock sample solution. the organic phase. It was also noticed that thorium, zirconium and scandium were retained in the organic phase. Stripping of rare earth elements Because of the better extraction eYciency of the monoester, EVect of matrix element on the extraction of REE many interfering ions are also co-extracted. So, separating the lanthanides and Y from these elements at the stripping To understand the cumulative eVect of matrix elements (silicate rock) on the REE extraction, studies were carried out using stage is necessary to reduce the spectral interferences in ICPAES determination.From the preliminary studies it was optimum parameters for REE extraction described in the Experimental section (Extraction). The quantity taken for observed that with higher concentration of hydrochloric acid (4.5 mol dm-3) the quantitative stripping of Tm, Yb and Lu each major and REE element was as follows (amount given in parentheses): Al (50–100 mg); Fe (10–50 mg); Mn is diYcult, even after the dilution of the organic phase with equal volume of kerosene.However, recovery of LREE was (0.2–4 mg); Ti (0.2–10 mg); Ca(5–50 mg); Mg (5–75 mg); Na and K (25 mg each); U, Th and Zr (0.5 mg each); La, Ce, Nd quantitative. It is reported that the antisynergistic interaction of TBP with H2MEHP reduces the distribution coeYcient of and Y (10–200 mg); Pr, Sm and Gd (6–15 mg each); Tb, Dy, Er and Tm (3–10 mg each); and Eu, Yb and Lu (1–5 mg each).NpIV considerably.26 Advantage of the antisynergistic eVect is taken for the quantitative back extraction of HREE. The eVect The results show that the recovery of REE and Y was quantitative with RSD of ~3%. of TBP on the back extraction of Er, Tm, Yb and Lu is shown in Fig. 2. It shows that from 0.18 mol dm-3 TBP onwards, The percentage of matrix elements (shown in parentheses) found in the aqueous portion was as follows: Al (~6); Mn the stripping of REE is quantitative.A 0.6 mol dm-3 TBP is selected because at this concentration major portions of and Ti (~3 each); Fe and Mg (~0.2); Ca (~0.5); Na and K (<0.1); U (~4). co-extracted iron, titanium and uranium can be retained in Table 3 Comparative concentration of the REE and Y in rock SRMs in mg g-1 Canadian SY-2 Canadian SY-3 USGS AGV-1 USGS GSP-1 Syenite Syenite Andesite Granodiorite Element Found RVa Found RVa Found RVa Found RVa La 74 75 1310 1340 35 38 179 184 Ce 172 175 2200 2230 65 67 405 399 Pr 18 18.8 220 223 7 7.6 49 52 Nd 75 73 700 670 31 33 191 196 Sm 16 16.1 115 109 5.8 5.9 25 26.3 Eu 2.3 2.42 17.4 17 1.58 1.64 2.4 2.33 Gd 16 17 110 105 5.4 5 11.8 12.1 Tb 2.7 2.5 19 18 0.6 0.7 1.3 1.34 Dy 18 18 120 118 3.5 3.6 5.3 5.5 Ho 3.9 3.8 28 29.5 0.6 0.67 0.9 1.01 Er 13 12.4 70 68 2 1.7 2.5 2.7 Tm 2.1 2.1 11 11.6 0.4 0.34 0.4 0.38 Yb 17 17 63 62 1.64 1.72 1.66 1.7 Lu 2.7 2.7 7.7 7.9 0.23 0.27 0.2 0.21 Y 129 128 729 718 18 20 25 26 aRV=recommended value (Govindraju, Geostand.Newsl., 1994, special issue, vol. XVIII) 1090 J. Anal. At. Spectrom., 1999, 14, 1087–1091The FeIII ion is extracted quantitatively and saturates the analytical results. The advantages of solvent extraction separation are the increases in the analysis eYciency in terms of organic phase, and thus adversely aVects the extraction of LREE. However, the extraction of major portions (>95%) of sample throughput and simplicity of the method.iron can be prevented by reducing FeIII to FeII with ascorbic acid. The residual iron remains in the organic phase at the Acknowledgements time of stripping (Fig. 3). Although, titanium is quantitatively Authors are thankful to Dr. R. K. Malhotra, Associate extracted, the major part of it also remains in the organic Director, Chemistry Group, Shri S. C. Verma, Regional phase at the time of stripping and is separated from the REE.Director and Shri K. K. Dwivedi, Director, AMD, Hyderabad, The quantity of Ti found was ~3%. If the TiO2 content is for their constant encouragement in this work. Services more than 1% in silicate rocks, it creates problems in the rendered by Dr. Vijay Kumar are gratefully acknowledged. dissolution of the sample in 0.12 mol dm-3 HCl medium. These can be overcome, to a certain extent, by the addition of 0.5 g of oxalic acid.Addition of oxalic acid also takes care of References traces of Nb and Ta and does not aVect the extraction of REE. 1 L. A. Haskin, in Rare Earth Element Geochemistry, ed. P. The partially extracted aluminium was stripped along with Henderson, Elsevier, Amsterdam, The Netherlands, 1984, the REE and it does not aVect the analytical results. It was pp. 115–148. further confirmed experimentally that up to 500 mg ml-1 of 2 H. Elderfield and M. J. Greaves, Nature (London), 1982, 296, 214.aluminium does not aVect the detection limit of individual 3 M. F. Thirlwall, Chem. Geol., 1982, 35, 155. rare earth elements and yttrium. The background continuum 4 L. A. Haskin, T. R. Wildeman and M. A. Haskin, J. Radiat. due to Al is mainly in the wavelength region between 190 and Chem., 1968, 1, 337. 5 J. W. Jacobs, R. L. Koritev, D. D. Balchard and L. A. Haskin, 220 nm,33 whereas the wavelength used ranges from 261 to J. Radioanal. Chem., 1977, 40, 93. 445 nm in this study.It was also observed that as the quantity 6 J. G. Crock, F. E. Lichte and T. R. Wildeman, Chem. Geol., 1984, of Al increases (>75 mg), the dispersion of the organic phase 45, 149. in the aqueous phase increases and the time required for the 7 A. R. Date and A. L. Gary, Spectrochim. Acta, Part B, 1985, separation of the layers also increases. 40, 115. Beryllium was also co-extracted and stripped along with the 8 W. Doherty and A. Vander Voet, Can. J. Spectrosc., 1985, 30, 135. 9 T. Hirato, H. Shimizu, T. Akagi, H. Sawetari and A. Masuda, REE and it interferes in the determination of Tm at 313.1 nm. Anal. Sci., 1988, 4, 637. So, another equally sensitive line, 346.22 nm, is selected for its 10 W. Doherty, Spectrochim. Acta, Part B, 1989, 44B, 263. determination. 11 A. Stroh, F. Bea and P. Montero, Atom. Spectrosc., 1995, 16, 7. 12 J. N. Walsh, F. Buckley and J. Barker, Chem. Geol., 1981, 33, 141. 13 J. G. Crock, and F. E. Lichte, Anal. Chem., 1982, 54, 1329.Accuracy and precision 14 J. G. Crock, F. E. Lichte, G. O. Riddle and C. L. Beach, Talanta, 1986, 33, 601. The accuracy of the proposed method was checked by analys- 15 D. W. Zachmann, Anal. Chem., 1988, 60, 420. ing four international standard reference materials (SRMs), 16 P. J. Watkins and J. Nolan, Chem. Geol., 1992, 95, 131. and the results are shown in Table 3. The results obtained 17 J. C. Farinas, H. P. Cabrera and M. T. Larrea, J. Anal. At. after the proposed solvent extraction group separation of REE Spectrom., 1995, 10, 511.and Y are in good agreement with the recommended values. 18 P. Roychowdhury, N. K. Roy, D. K. Das and A. K. Das, Talanta, 1989, 36, 1183. The RSD obtained for four determinations is ~1–3%. It 19 M. A. Eid, J. A. C. Broekaert and P. Tschepel, Fresenius J. Anal. shows that the proposed method allows precise and accurate Chem., 1992, 342, 107. analysis of geological samples for their REE and Y contents. 20 J. C. Farinas, H. P. Cabrera and M. T. Larrea, J. Anal. At. The method is simple, expeditious and enables larger samples Spectrom., 1995, 10, 511. to be taken up for the group separation of REE and Y 21 K. Satyanarayana, At. Spectrosc., 1996, 17, 69. simultaneously as compared with the ion exchange method. 22 K. Iwasaki and H. Haraguchi, Anal. Chim. Acta, 1988, 208, 163. 23 J. DiYeld and G. R. Gilmore, J. Radioanal. Chem., 1979, 48, 135. The REE in complex matrices (zircon, ilmenite, rutile, 24 I.Roelandts, Anal. Chem., 1981, 53, 676. columbite–tantalite, uranium rich rock samples and ultra-basic 25 Y. Marcus and A. S. Kertes, Ion Exchange and Solvent Extraction rock) cannot be determined by this method because the major of Metal Complexes, Wiley Interscience, New York, USA, 1969. elements saturate the organic phase. Therefore, the REE and 26 A. K. De, S. M. Khopkar and R. A. Chalmers, Solvent Extraction Y cannot be completely extracted even with two extractions. of Metals, Van Nostrand, New York, USA, 1970. For such types of samples this method can be applied after 27 D.Weiss, T. Paukert and I. Rubeska, J. Anal. At. Spectrom., 1990, 5, 371. the oxalate precipitation of REE using calcium as carrier. 28 D. F. Peppard, G. W. Mason, J. L. Marier and W. J. Driscoll, J. Inorg. Nucl. Chem., 1957, 4, 334. 29 D. F. Peppard, S. W. Moline and G. W. Mason, J. Inorg. Nucl. Conclusion Chem., 1957, 4, 344. Solvent extraction using a mixture of H2MEHP and HDEHP 30 D. F. Peppard, G. W. Mason and S. W. Moline, J. Inorg. Nucl. in kerosene can be applied for the separation, suitable for the Chem., 1957, 5, 141. 31 A. B. Shabani, T. Akagi, H. Shimizu and A. Masuda, Anal. Chem., ICP-AES determination, of REE and Y in a silicate rock. The 1990, 62, 2709. REE at a trace level suitable for petrogenetic study can be 32 J. Rydberg, C. Musikas and G. R. Choppin, Principles and determined precisely and accurately. Back extraction of inter- Practices of Solvent Extraction, Marcel Dekker, Inc., New York, fering elements (Fe, Ti, Zr, U, etc.) can be prevented using USA, 1992, p. 398. TBP and the quantity of iron and titanium in the aqueous 33 M. Thompson and J. N. Walsh, A Handbook of Inductively strip can be controlled by selecting suitable concentration of Coupled Plasma Spectrometry, Blackie, London, 1983, p. 13. TBP. The matrix interferences accompanying separation of REE in a silicate rock were too small to cause any change in Paper 9/01437C J. Anal. At. Spectrom., 1999, 14, 1087–1091 1091

ISSN:0267-9477    

DOI:10.1039/a901437c

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

INTER-LABORATORY NOTE Analysis of ZrO2 powders by microwave assisted digestion at high pressure and ICP atomic spectrometry Dirk Merten,a Jose A. C. Broekaert,*b Rolf Brandtc and Norbert Jakubowskic aUniversita�t Dortmund, Fachbereich Chemie, D-44221 Dortmund, Germany bUniversia�t Leipzig, Institut fu�r Analytische Chemie, Linne�-Straße 3, D-04103 Leipzig, Germany cInstitut fu�r Spektrochemie und Angewandte Spektroskopie (ISAS), D-44013 Dortmund, Germany Received 25th January 1999, Accepted 6th May 1999 Microwave assisted digestion at high pressure was investigated for the dissolution of diVerent ZrO2-based ceramic powders and their subsequent analysis performed by inductively coupled plasma optical emission spectrometry (ICP-OES) and mass spectrometry (ICP-MS).For a fine grain size ZrO2 powder (median particle size <1.9 mm) the results were found to well agree with those obtained in the case of conventional digestion at high pressure, decomposition by fusion with NH4HSO4 and slurry nebulization ICP-OES.In the case of microwave assisted digestion at high pressure, up to 600 mg of ZrO2 could be dissolved within only 60 min, whereas by conventional digestion at high pressure up to 1000 mg ZrO2 powder could be dissolved; however, this required a time of 10 h. By fusion with NH4HSO4 it was not possible to dissolve all of the ceramic powders investigated completely. For all investigated elements excepted for B and Si, recoveries of 100% were obtained within the level of experimental error 3–13%.Detection limits, in the case of ZrO2 powders with high concentrations of Hf, Na and Y, were found to range from 0.03 mg g-1 for Mg, when applying conventional digestion at high pressure, over 0.4 mg g-1 for Fe, in the case of microwave assisted digestion at high pressure, to 92 mg g-1 for Na in the case of slurry nebulization ICPOES and 114 mg g-1 in the case of Y subsequent to decomposition by fusion. The results of analysis subsequent to the diVerent dissolution methods and those obtained with slurry nebulization ICP-OES agreed well for the elements Cr, Fe, Hf, Mg, Na, Ti and Y.With quadrupole based inductively coupled plasma mass spectrometry Na could be determined at the 600 mg g-1 level and the results agreed well with those obtained by ICP-OES, whereas for Al, Cr, Cu, Fe, Mg, Mn and Ni spectral interferences were found to hamper analyses. For Li, as well as for Ce, La, Pr and Th, it could be shown that the impurity levels in the samples analyzed were below 2 and 1 mg g-1, respectively.ZrO2-based ceramics are of great interest due to their high seems to be very successful.18 Also, mixtures of (NH4)2SO4 with H2SO4 have been used.19 The major advantage of alkali resistance to pressure variations and high thermal resistance. They are used in engines and also in electronics and electro- fusions lies in the short digestion time, which also applies in the case of ceramic powders.A disadvantage might lie in the technics1 and as sensors for diVerent applications.2,3 Furthermore, ZrO2 powders are used as stationary phases for use of open systems and of large amounts of reagents. To minimize contamination conventional digestion at high liquid chromatography, e.g., in the separation of proteins.4 The properties of advanced ceramics may be influenced pressure in closed vessels was applied. In the case of ZrO2, mainly the use of H2SO420 and HF,21 as well as of strongly by the addition of minor but also of trace elements.5 Therefore, the chemical composition of the powders used for HF–HCl–H2SO4,22 have been described.However, conventional digestion at high pressure is time-consuming. Recently, the production of advanced ceramics must be determined with high precision down to trace concentrations. Since many microwave assisted digestion at high pressure in a closed system also has been used for the dissolution of ZrO2-based elements are important for the properties of ZrO2 based ceramics only methods with high multi-element capacity should powders.23 Because of the very eYcient energy transfer to the sample short decomposition times could be realized.In this be applied, such as optical emission spectrometry or mass spectrometry with inductively coupled plasma (ICP-OES and work we investigated the use of microwave assisted digestion at high pressure for the dissolution of a ceramic powder on ICP-MS). ICP-OES may suVer from spectral interferences, since Zr has a line-rich atomic emission spectrum.Therefore, the base of ZrO2 which contains large concentrations of Y and Hf (>10% Y and >1% Hf w/w) and subsequently matrix removal procedures are of prime interest,6–8 as they also provide for an enrichment of the analytes. As the dissolu- determined the elements Al, B, Ca, Cr, Cu, Fe, Hf, Mg, Mn, Na, Ti, Si, V, Y and Zn by ICP-OES. The possibilities and tion of ZrO2 may be diYcult or it may suVer from analyte losses or contamination, slurry nebulization can be applied.9–11 limitations of the procedure were compared with those of conventional digestion at high pressure and with those of However, the application of slurry nebulization for ICP-OES is limited to powders with fine grain size.12,13 Recently, solu- fusion with NH4HSO4.As the dissolution of ceramic powders does not only depend on the chemical composition of the tions have been proposed involving electrothermal evaporation of slurries.14,15 sample to be dissolved, but also on the particle size distribution, 24 the latter was measured with the aid of laser light Generally, however, there is a need for the elaboration of digestion procedures enabling analyses of coarse powders.For scattering. Despite spectral interferences quadrupole-based ICP-MS ZrO2 fusions with Na2B4O7, Na2CO3, KHF2, NH4HSO4 and (NH4)2SO4,16 and LiBO2–H3BO317 can be applied.In particu- will be shown to be of use especially for attesting the degree of purity for a number of high mass elements. lar, decomposition by fusion with NH4HSO4 or (NH4)2SO4 J. Anal. At. Spectrom., 1999, 14, 1093–1098 1093Table 1 Instrumental parameters of the JY 24 ICP-OES and the VG Dissolution procedures and slurry nebulization PQ2 ICP-MS For the diVerent digestion methods the maximum sample mass ICP-OES— that can be dissolved, the time required for the digestion, the Generator 800 W, 40.68 MHz temperature and the amount of reagents used were optimized.Monochromator 0.64-m Czerny–Turner mounting grating: Further, blank digestions were carried out and the recoveries 2400 grooves mm-1 were determined. For both microwave assisted digestion at Peristaltic pump Perimax 12 (Spetec, Munich, Germany) high pressure and conventional digestion at high pressure four Nebulizer G.M.K. (Labtest, Ratingen, Germany) blank digestions were carried out, while for the fusion tech- Analysis of solutions: Analysis of slurries: Nebulizer pressure 2.5 bar 2.5 bar nique five blank digestions were performed.In the analysis of Viewing height 7 mm 10 mm the blank solutions by ICP-OES aqueous multi-element stan- Coater gas flow 0.3 l min-1 0.3 l min-1 dard solutions of Al, B, Ca, Cr, Cu, Fe, Hf, Mg, Mn, Na, Ni, Outer gas flow 16 l min-1 16 l min-1 Si, Ti, V, Y and Zn, with concentrations of 0, 12.5, 25, 50 and Sample uptake 2.5 ml min-1 2.8 ml min-1 100 mg l-1, respectively, were used for calibration.Recoveries ICP-MS— were determined for five standard solutions of 10 ml, each Power 1400 W containing 40 mg of Al, B, Ca, Cr, Cu, Fe, Mg, Mn, Na, Ni, Outer gas flow 14 l min-1 Nebulizer gas flow 0.9 l min-1 Si, Ti, V and Zn, 1 mg of Hf and 5 mg of Y. These solutions Intermediate gas flow 1.1 l min-1 were evaporated to dryness at 100 °C and the reagents for the Sample uptake Meinhard-nebulizer with peristaltic pump digestion were added.The digestion procedure was carried out (Gilson, USA) with four of these samples, whereas the fifth sample was used Sample uptake rate 0.8 ml min-1 for comparison. All sample solutions obtained were diluted to Interface Ni sampler (1.0 mm), Ni skimmer (0.7 mm) a volume of 100 ml. Vacuum 2.3×10-9 bar The microwave assisted pressurized digestion system ‘Pressurized Microwavn’ (PMD) (Paar, Graz, Austria) consisted of a microwave oven with a maximum Experimental power of 750 W.In the system two high-pressure vessels made of perfluoroalkoxyethylene (PFA) with a volume of 50 ml can Instrumental set-up be used simultaneously at up to a maximum pressure of 35 For the digestion of the ceramic powders we investigated the bar. The increase in the temperature of 1 kg of H2O as a use of a PMD (pressurized microwave digestion) system (Paar, function of time for diVerent power settings was determined Graz, Austria) and a DAB III (Berghof, Eningen, Germany) so as to evaluate the power uptake, as has been described by pressurized digestion system.In Table 1 the instrumental par- Kingston et al.26 After cooling, the sample was transferred to ameters and the plasma operating conditions of the sequential 100 ml flasks and the PFA vessels were cleaned for the next JY 24 ICP-OES (ISA Jobin-Yvon, Longjumeau, France) and digestion by applying blank digestions. the VG PQ2 Turbo Plus (Fisons, Winsford, Cheshire, UK) In the conventional pressurized digestion system ‘DAB III’ ICP-MS used are listed.(Berghof, Eningen, Germany) polytetrafluoroethylene (PTFE) vessels with a volume of 250 ml were used. The sample was Reagents placed in the PTFE vessels, which were subsequently sealed and transferred to a steel autoclave after adding the required Analytical grade chemicals were used. Solutions of Al, B, Ca, acids. After decomposition and cooling the sample solutions Cr, Cu, Fe, Hf, Mg, Mn, Na, Ni, Si, V, Y, Zn and Zr were were transferred to 100 ml flasks and the PTFE vessels were prepared from Titrisol solutions (1 g l-1) (Merck, Darmstadt, cleaned by applying blank digestions and subsequent rinsing Germany).Solutions of Ti were prepared from a Ti standard with 2% HNO3. solution (Riedel de Hae�n, Seelze, Germany). NH4HSO4 was The fusion with NH4HSO4 was carried out in quartz purchased from Fluka (Deisenhofen, Germany). The acids crucibles placed in an aluminium block (height, 70 mm; and HF (37%) (Baker, Griesheim, Germany), HCl (37%) (Riedel id of bores, 30 mm).After loading, the block was heated up de Hae�n), HNO3 (65%) (Merck) and H2SO4 (95–97%) (Riedel to 400–450 °C by means of a muZe oven, which was coated de Hae�n) were used. Nitric acid and sulfuric acid were purified with aluminium foil so as to avoid contamination. The temby sub-boiling distillation. Water used for the preparation of perature was controlled with a Ni/Cr–Ni thermoelement.After all solutions was obtained from a Milli-Q system (Millipore, fusion, the temperature was increased to about 500 °C so as Eschborn, Germany). The ZrO2 powder samples analyzed to remove excess NH4HSO4 by sublimation, and after cooling were supplied by CeramTec (Plochingen, Germany), CWH the sample was dissolved in 13.6% HNO3. (Marl, Germany), Friatec (Mannheim, Germany) and Tosoh For slurry nebulization ICP-OES the ceramic powder was (Tokyo, Japan).One was prepared at the Laboratory for transferred to 100 ml flasks and water was added while homo- Materials Science at the University of Dortmund (Germany).25 genizing for 15 min in an ultrasonic bath (Sonorex Super RK 510H, Bandelin, Berlin, Germany). Since a highly stable suspen- Characterization of the ceramic powders sion is recommended in order to obtain a representative aerosol for introduction into the ICP, the pH of the suspensions was The specific surfaces of the ceramic powders were determined varied.This is known as the electrostatic stabilizing mechan- by BET measurements (Gemini 2360, Micromeritics, Mo� nism. 27 Further, during the sample aspiration the slurries were chengladbach,Germany) using pressure changemonitoring after continuously stirred with a magnetic stirrer (Janke und Kunkel, admitting N2 in closed vessels. For the determination of the Staufen, Germany) while placed in a thermostated bath. particle size distribution laser light diVraction measurements Slurries (1% m/v Zr) of CH 411/1, Powder 2 and TZ-8Y were made with a Particle and Droplet Analyzer 2600 were prepared at pHs varying between 0 and 14 and the (Malvern, Herrsching, Germany).Therefore, slurries were stability of the slurries was evaluated by measuring the intensity prepared from the ceramic powders in an ultrasonic bath and of the Zr II 339.198 nm line as a function of time over 30 min. pumped through the measurement cuvette.With this technique the particle size can be determined from the diVraction angle, Line selection for analyses of ZrO2 by ICP-OES whereas the number of particles of a specific size between 0.5 and 188 mm can be calculated from the amplitudes at a For the selection of the atomic emission lines to be used for the analyses of ZrO2 by ICP-OES, both single element solu- given angle. 1094 J. Anal. At. Spectrom., 1999, 14, 1093–1098Table 3 Optimized conditions for the dissolution of ZrO2 powders by Table 2 Atomic emission lines and wavelengths for background correction for the analysis of solutions containing large amounts of Zr by means of microwave assisted digestion at high pressure (power uptake: 625±4 W) ICP-OES Wavelengths for background Ceramic powder Sample weight/mg Acids Element Wavelength/nm correction/nm Dynazirkon F 600 Q TZ-3Y 600 N B I 208.959 208.929 208.974 Cr II 205.552 205.537 205.583 TZ-8Y 600 N 1 ml HNO3 P9a 600 R + Cu I 223.008 222.978 223.038 Fe II 238.204 238.175 238.218 M 824 300 N 2 ml H2SO4 CH 282/1 300 N Fe II 239.562 239.529 239.595 Hf II 227.717 227.688 227.761 Powder 2 300 S Hf II 251.688 251.646 251.730 CH 411/1 200 Q Mg II 279.553 279.523 279.583 Dynazirkon MS 200 N 1 ml HF+ Mg II 280.270 280.226 280.300 Powder 3 300 R 2 ml H2SO4 b Na I 588.995 588.939 589.051 TZ-O 400 S Na I 589.592 589.539 589.662 aSeparate sample treatment necessary: 15 min with 3 ml HNO3.Si I 250.690 250.674 250.722 bAddition of 20 ml H3BO3 (sat.) for complexation of HF.Ti II 334.941 334.872 334.983 Y II 371.030 370.989 371.071 Y II 377.433 377.393 377.473 Zn II 206.196 206.181 206.211 for the conventional digestion at high pressure (Table 4). The concentrations added for the calibration by standard addition were 0, 12.5, 25, 50 and 200 mg l-1 for Cr, Cu, Fe, Mg, Ti and tions and multi-element solutions containing Al, B, Ca, Cr, Zn, and 0, 75, 150, 300 and 1200 mg l-1 for Na.For Hf, 0, 1, Cu, Fe, Mg, Mn, Na, Ni, Si, Ti, V, Zn (10 mg l-1 each), Hf 2 and 5 mg l-1 were added. Finally, the concentrations added (500 mg l-1), Y (1000 mg l-1) and Zr (10 g l-1) were used for Y were 0, 10, 25 and 50 mg l-1. and measurements were made in the wavelength range 165– 750 nm with the aid of the IMAGE system (ISA Jobin Fusion with NH4HSO4. Five digestions of 500 mg TZ-8Y Yvon).28 The wavelengths of lines and the related intensities were carried out by fusion with NH4HSO4.The concentrations (for 10 mg l-1) were taken from the data base in the IMAGE added for calibration by standard additions were 0, 12.5, 25, library and the lines were investigated with respect to their 50 and 100 mg l-1 for Cr, Cu, Fe, Mg, Ti and Zn, as well as freedom from interferences for the case of solutions of Zr. In 0, 75, 150, 300 and 600 mg l-1 for Na. For Hf 0, 1, 2 and Table 2 the atomic emission lines selected for the analysis of 5 mg l-1 were added.Finally, the concentrations added for Y solutions with high concentrations of Zr by ICP-OES and the were 0, 10, 20 and 40 mg l-1. wavelengths for background correction are listed. For Al, Ca, Mn, Ni and V no interference-free sensitive atomic emission Slurry nebulization ICP-OES. For the analyses by slurry lines were found. For Fe, Hf, Mg, Na and Y two and for B, nebulization ICP-OES, the optimized parameters for the analy- Cr, Cu, Si, Ti and Zn one atomic emission line, respectively, sis of slurries, listed in Table 1, were used.Again, the calibration could be found. was performed by standard additions. For the determination of B, Cr, Cu, Fe, Mg, Na, Si, Ti and Zn 1.35 g of TZ-8Y were ICP atomic spectrometric analyses suspended in 100 ml of water and the pH was adjusted to 2.3. The concentrations of the elements added were 0, 25, 50, 100 Quantitative determinate carried out on the ZrO2 and 500 mg l-1 for B, Cr, Cu, Fe, Mg, Si, Ti and Zn.For Na powder TZ-8Y, as this powder can be dissolved by all of the the concentrations added were 0, 150, 300, 600 and 3000 mg l-1. digestion methods investigated and as determinations by slurry For the determination of Hf and Y 100 mg of TZ-8Y were nebulization ICP-OES are possible. For all ICP-OES determisuspended in a volume of 100 ml. For this last case the added nations calibration was performed by standard additions. concentrations were 0, 2.5, 5, 10, 20 and 30 mg l-1 for Hf and To investigate the reproducibility of the digestion methods 0, 12.5, 25, 50, 100 and 200 mg l-1 for Y.several digestions were carried out and the concentrations of Semi-quantitative determinations by ICP-MS were carried Cr, Cu, Fe, Hf, Mg, Na, Ti, Y and Zn in the solutions obtained out with the solutions obtained by microwave assisted digestion were determined by ICP-OES. The solutions obtained with the at high pressure, conventional digestion at high pressure and diVerent digestion methods investigated were all diluted to a decomposition by fusion with NH4HSO4 for both blank volume of 100 ml with purified water. For the determinations of Hf and Y in TZ-8Y the sample solutions were diluted by a Table 4 Optimized conditions for the dissolution of ZrO2 powders factor of 50, and for the determination of the elements Cr, Cu, using conventional digestion at high pressure (T, 220 °C) Fe, Mg, Na, Ti and Zn the samples were diluted from 15 to 100 ml.For each atomic emission line 10 replicate measurements Sample were made. The operating conditions for ICP-OES optimized Ceramic powder weight/mg Acids Time/h for the analysis of solutions were used (Table 1). Dynazirkon F 750 Q 6 ml HNO3 TZ-O 750 N + Microwave assisted digestion at high pressure. Six digestions N TZ-3Y 1000 R 6 ml H2SO4 10 of 600 mg TZ-8Y were carried out using the optimized con- TZ-8Y 1000 N + ditions for the microwave assisted digestion at high pressure P9 1000 4 ml H2O (Table 3).The concentrations added for the calibration by Powder 2 500 S standard additions were 0, 12.5, 25, 50 and 100 mg l-1 for Cr, M 824 200 Q 8 ml HNO3 Cu, Fe, Mg, Ti and Zn, and 0, 75, 150, 300 and 600 mg l-1 CH 282/1 200 R + 20 for Na. For Hf 0, 1, 2 and 5 mg l-1 were added. Finally, the CH 411/1 200 S 8 ml H2SO4 concentrations added for Y were 0, 10, 20 and 40 mg l-1. Powder 3 500 L 3 ml HF+4 ml HNO3 Dynazirkon MS 300 K +4 ml HCl* 20 Conventional digestion at high pressure.Seven digestions of aAddition of 20 ml H3BO3 (sat.) for complexation of HF. 1000 mg TZ-8Y were carried out at the optimized conditions J. Anal. At. Spectrom., 1999, 14, 1093–1098 1095samples and for solutions of TZ-8Y. As an internal standard by pyrolysis of Zr propylate, a decomposition of the organic matrix with 3 ml HNO3 for 15 min was found to be necessary. 20 mg l-1 of Rh were added to all solutions. For the case of the TZ-8Y samples the matrix concentration was adjusted to Blank levels for Al, B, Hf, Si and Y were found, which might have stemmed from the vessels, whereas the blank values 250 mg l-1.Six digestions of 600 mg TZ-8Y and four blank digestions were carried out with microwave assisted digestion for Ca, Cr, Fe and Mg might have stemmed from the acids used. The addition of 20 ml of a solution saturated with at high pressure. Seven digestions of 1000 mg TZ-8Y and four blank digestions were carried out with the conventional diges- H3BO3 for the complexation of free HF was found to lead to an increase in contamination.tion at high pressure. Finally, five digestions of 500 mg of TZ-8Y and five blank digestions were carried out by fusion For Al, Ca, Cr, Cu, Fe, Hf, Mg, Mn, Na, Ni, Si, Ti, V, Y and Zn, recoveries of 100% were obtained within the exper- with NH4HSO4. imental error level of 3–8%. When applying microwave assisted digestion at high pressure with 1 ml of HNO3 and 2 ml of Results and discussion H2SO4, positive errors of 200 and 300% occurred for B and Si, whereas in the case of HF losses of about 40% were Characterization of the ceramic powders obtained for Si.The specific surfaces of the ZrO2 powders investigated are listed in Table 5 and found to range from 3 to 176 m2 g-1. The pH was found to significantly influence the particle size Conventional digestion at high pressure. All digestions were distribution. Whereas, at a pH of 5.5, particles of up to 188 mm carried out at a temperature of 220 °C.With conventional were found in the case of slurries of Powder 2, the maximum digestion at high pressure all ZrO2 powders investigated could particle size for a pH of 2.3 is only 35 mm (Fig. 1). This be completely dissolved within 10–20 h (Table 4). The dissolu- therefore demands a careful optimization of the pH prior to tion behaviour of the diVerent powders was found to agree slurry analysis. In Table 5 the median particle sizes for the well with the size of the specific surfaces (compare with the investigated ceramic powders and the pH-values at which they data in Table 5).For ZrO2 powders with a large specific were determined are listed. For Dynazirkon MS, M 824, P9 surface (such as Dynazirkon F, TZ-O, TZ-3Y, TZ-8Y, P9 and and Powder 3, stable slurries could not be obtained, whereas Powder 2) between 500 and 1000 mg can be dissolved within for TZ-3Y, TZ-8Y and Dynazirkon F all particles were found 10 h, whereas for the coarser powders (M824, CH411/1 and to be of a size below 1.9 mm.CH 282/1) only 200 mg can be dissolved within 20 h. Dynazirkon MS and Powder 3 could only be completely Dissolution procedures dissolved when using HF. For the complexation of the free HF 20 ml of a solution saturated with H3BO3 were added. Microwave assisted digestion at high pressure. All ceramic powders investigated could be dissolved within 60 min. The For Al, Ca, Mg, Mn, Ni, Zn, and especially for B, Cr, Fe and Hf, blank digestions showed that contamination takes maximum amounts which could be dissolved ranged from 200 to 600 mg (Table 3). For all digestions the power uptake was place.For Al, Cr, Fe, Ni and Zn the blank values obtained probably stem from a corrosion of the steel autoclaves. Ca, found to be 625±4W. For P9, being a powder manufactured Table 5 Specific surface of the investigated powders and median particle sizes at optimized pH obtained by diVraction of laser light Powder Distributor Specific surface/m2 g-1 Median particle size/mm pH Powder 2 Ceram Tec 6.2 4.0 2.3 Powder 3 Ceram Tec 2.8 P9 Univ.of Dortmund 176 Dynazirkon F CWH 71 <1.9 5.5 Dynazirkon MS CWH 2.9 CH 282/1 Friatec 4.3 4.6 5.5 CH 411/1 Friatec 3.1 5.6 5.5 M 824 Friatec 3.4 TZ-0 Tosoh 14 4.5 2.3 TZ-3Y Tosoh 16 <1.9 2.3 TZ-8Y Tosoh 14 <1.9 2.3 Fig. 1 Particle size distributions obtained for slurries of Powder 2 by diVraction of laser light at a pH of 5.5 (a) and 2.3 (b) (abundancy and cumulative abundancy are given). 1096 J. Anal. At. Spectrom., 1999, 14, 1093–1098Mg and Mn might be introduced by the acids and B, Hf and lowest median particle size. However, the recovery for Powder 2 is relatively low as compared with that for the other powders. Si may be released from the PTFE vessels used. For all investigated elements except for B and Si, recoveries For the powders Dynazirkon MS, M 824, Powder 3 and P9 no stable slurries could be obtained, whereas for other powders of 100% were obtained within the error level of 3–8%.For B the recovery was about 80% and for Si it was between 30 (see Table 5) the slurries were found to be stable. and 50%. Analyses by ICP-OES Fusion. Not all ZrO2 powders could be completely Detection limits. The detection limits (3s criterion) obtained dissolved by fusion with NH4HSO4. Powder M 824, which is for determinations in the ZrO2 powder TZ-8Y by ICP-OES stabilized by the addition of MgO, could not be dissolved at subsequent to microwave assisted digestion at high pressure, all and for the powders Dynazirkon F, Dynazirkon MS, conventional digestion at high pressure, fusion with NH4HSO4 Powder 2 and Powder 3 a residue below 1% of the original and by slurry nebulization ICP-OES (Fig. 3) vary consider- mass remained after digestion. It could be shown by X-ray ably. The lowest detection limit was obtained for Mg in the fluorescence (XRF) that this residue consists of SiO2, as has case of conventional digestion at high pressure (0.03 mg g-1) already been reported in the literature.16 Five hundred and the highest one for Y in the case of fusion (114 mg g-1).milligrams of the powders CH 282/1, CH 411/1, P9, TZ-0, For most elements detection limits of 1 mg g-1 or better were TZ-3Y and TZ-8Y can be completely dissolved when using a obtained when using sample dissolution. They were lowest in 14-fold excess of NH4HSO4.By increasing the temperature the case of conventional digestion at high pressure and worst to 500 °C, the excess of NH4HSO4 can be removed by at in the case of fusion, which shows that the power of detection least 99%. The heating of the sample, the removal of the increases with the amount of sample dissolved. However, excess of NH4HSO4 and the subsequent cooling of the sample despite the large amount of sample used in slurry nebulization, together take about 5 hours.the detection limits were found to be higher than those for Contamination from the crucibles was found to occur only sample dissolution. This eVect is probably due to a decrease for Si. However, for each of the elements investigated, blank in precision encountered in the analysis of slurries as a result contributions were found to be introduced by the NH4HSO4 of the evaporation of solid particles.29 used. Again, for all elements except for B and Si, recoveries of 100% were obtained within the error level of 5–13%.Results of quantitative determinations. The analysis results Positive errors of up to 50% were found to occur for Si as a of ICP-OES for Cr, Fe, Hf, Mg, Na, Si, Ti, Y and Zn in the result of blank contributions and losses by up to 80% were ceramic powder TZ-8Y obtained with the diVerent decompo- found to occur for B as a result of the high temperatures to sition methods and by slurry nebulization agree well within be applied.the error levels (Fig. 4). The errors of analysis, as calculated by a propagation of error, range from 5–8% for the elements Slurry nebulization ICP-OES present at concentrations well above the detection limits. After an optimization of the viewing height especially (see Table 1 for comparison with the parameters used for the Semi-quantitative analyses by ICP-MS analysis of solutions) the recoveries for Zr at diVerent pH As a result of spectral interferences by species stemming from values were determined.the decomposition reagents, Ca, Ti and V could not be For Powder 2, stable slurries could be obtained at pH values determined by quadrupole-based ICP-MS. In Table 6 the of 1.5–2.5 and of 11. For TZ-8Y, stable slurries were obtained detection limits obtained for Al, Ce, Cr, Cu, Fe, La, Li, Mg, at pH values of 1.5–2.5 and 9. For both powders no significant Mn, Na, Ni, Pr, Th, U and Zn are listed. For Al, Cr, Cu, Fe, diVerences in the recovery at these pH ranges were found. For Mg, Mn, Ni and Zn the concentrations determined were below powder CH 411/1, stable slurries were obtained at pH values the detection limits.For Na the results of 640±40 mg g-1 of 1.5–2.5 and 5–6. Here, the recovery for Zr is significantly (conventional digestion at high pressure), 580±40 mg g-1 higher at a pH of 5–6. The highest recoveries were obtained (microwave assisted digestion at high pressure) and for slurries of Powder 2 and TZ-8Y at a pH of 2.3 and for 650±60 mg g-1 (fusion) obtained by ICP-MS well agree with slurries of CH 411/1 at pH 5.5.The recovery obtained for those obtained by ICP-OES (see Fig. 4). For elements having TZ-8Y was higher than the one for slurries of Powder 2 and their masses in ranges free from spectral interferences it was CH 411/1 (Fig. 2). This agrees in principle with both the possible to obtain detection limits significantly below 1 mg g-1 specific surfaces of the powders investigated and with the with ICP-MS, and here semi-quantitative analyses without median particle sizes found for the slurries at the optimized pH.Powder TZ-8Y has the largest specific surface and the Fig. 2 Recoveries for Zr obtained at optimized conditions for the Fig. 3 Limits of detection (LOD) for the analysis of ZrO2 powder TZ-8Y by ICP-OES. analysis of slurries of ZrO2 by ICP-OES. J. Anal. At. Spectrom., 1999, 14, 1093–1098 1097the largest amount of sample can be used.With microwave assisted digestion at high pressure the digestion time and also the blank levels are lower than those obtained with all other digestion methods investigated. For conventional digestion at high pressure the dissolution behaviour of the diVerent powders could be shown to be related to the specific surfaces of the ceramic powders. It could also be shown that decomposition by fusion with NH4HSO4 was not suitable for ceramic powders containing significant concentrations of SiO2 and MgO.Acknowledgements The authors thank Anton Paar GmbH (Graz, Austria) for supplying the system for microwave assisted digestion at high pressure and CeramTec (Plochingen, Germany), CWH (Marl, Germany), Friatec (Mannheim, Germany) and Tosoh (Tokyo, Japan) for the donation of the ceramic powders analyzed. Financial support from the Ministerium fu� r Wissenschaft und Forschung des Landes Nordrhein-Westfalen and by the Bundesministerium fu� r Bildung, Wissenschaft, Forschung und Technologie is acknowledged by R.B.and N.J. Fig. 4 Analysis of the ceramic powder TZ-8Y by ICP-OES subsequent to microwave assisted digestion at high pressure (n=6), conventional References digestion at high pressure (n=7), decomposition by fusion (n=5) 1 J. A. C. Broekaert, T. Graule, H. Jenett, G. To� lg and P. Tscho� pel, and slurry nebulization ICP-OES. Fresenius’ Z. Anal. Chem., 1989, 332, 825. 2 Y. T. Yan, N. Miura and N. Yamazoe, Sens. Actuators, 1995, Table 6 Concentrations and detection limits (in brackets) for selected B24, 287.elements obtained by semi-quanitative analyses of the ZrO2 powder 3 H. Kurosawa, Y. Yan, N. Miura and N. Yamazoe, Chem. Lett., TZ-8Y by ICP-MS subsequent to dissolution 1994, 1733. 4 Y. Hu and P. Carr, Anal. Chem., 1998, 70, 1934. Conventional Microwave assisted 5 J. A. C. Broekaert and G. To� lg, Mikrochim. Acta, 1990, II, 173. digestion at digestion at high Dissolution by 6 R. Lobinski, J.A. C. Broekaert, P. Tscho� pel and G. To� lg, high pressure pressure fusion Fresenius’ J. Anal. Chem., 1992, 342, 569. (n=7) (n=6) (n=5) 7 K. Kimura, Bull. Chem. Soc. Jpn., 1961, 34, 63. /mg g-1 /mg g-1 /mg g-1 8 C. F. Coleman, C. A. Blake and K. B. Brown, Talanta, 1962, 9, 297. Detection limits— 9 J. A. C. Broekaert, C. Lathen, R. Brandt, C. Pilger, D. Pollmann, 27Al 5 19 50 P. Tscho� pel and G. To� lg, Fresenius’ J. Anal. Chem., 1994, 349, 20. 52Cr 7 5 3 10 H. Min and S.Xi-en, Spectrochim. Acta, Part B, 1989, 44, 957. 63Cu 1.3 1.0 7 11 L. Ebdon, M. Foulkes and K. Sutton, J. Anal. At. Spectrom., 56Fe 80 110 60 1997, 12, 213. 24Mg 8 2 5 12 D. Merten, P. Heitland and J. A. C. Broekaert, Spectrochim. Acta, 55Mn 1.3 0.1 0.3 Part B, 1997, 52, 1905. 60Ni 4 2 5 13 B. Raeymaekers, T. Graule, J. A. C. Broekaert, F. Adams and 68Zn 5 5 3 P. Tscho� pel, Spectrochim. Acta, Part B, 1988, 43, 923. 14 S. Hauptkorn, G. Schneider and V. Krivan, J. Anal. At.Spectrom., Semi-quantitative analyses— 1994, 9, 463. 140Ce <0.3 (0.3) 0.29±0.03 (0.01) 0.26±0.03 (0.19) 15 G. Schneider and V. Krivan, Spectrochim. Acta, Part B, 1995, 139La 0.35±0.03 (0.1) 0.40±0.04 (0.04) 0.37±0.07 (0.06) 50, 1557. 7Li 1.7±0.2 (0.3) 1.56±0.08 (0.3) 1.6±0.3 (0.9) 16 R. Bock, Aufschlußmethoden der anorganischen und organischen 23Na 640±40 (64) 580±40 (30) 650±60 (100) Chemie, VCH, Weinheim, Germany, 1972. 141Pr 0.10±0.01 (0.03) 0.06±0.03 (0.02) 0.07±0.02 (0.05) 17 W. M. Wise and S. D. Solsky, Anal. Lett., 1976, 9, 1047. 232Th 0.9±0.1 (0.2) 0.90±0.09 (0.15) 1.0±0.1 (0.1) 18 F. Kohl, N. Jakubowski, R. Brandt, C. Pilger and J. A. C. 238U 0.30&plusplusmn;0.09 (0.1) 0.3±0.1 (0.1) Broekaert, Fresenius’ J. Anal. Chem., 1997, 359, 317. 19 J. Ito, Bull. Chem. Soc. Jpn., 1962, 35, 225. 20 J. Dolezal, L. Lenz and S. Sulcek, Anal. Chim. Acta, 1969, 47, 517. calibration solutions could be performed on the solutions 21 K. Stulik, P. Beran and J. Dolezal, Talanta, 1978, 25, 363. obtained with conventional digestion at high pressure, micro- 22 N. Z. Baluch, K. Anwar, S. M. Ifzal and D. Mohammad, J. Radioanal. Nucl. Chem., 1990, 141, 417. wave assisted digestion at high pressure and fusion with 23 M. T. Larrea, I. Go�mez-Pinilla and J. C. Farin� as, J. Anal. At. NH4HSO4 (Table 6). Owing to the lack of comparative data, Spectrom., 1997, 12, 1323. however, no quality assurance of the analyses could be 24 G. Zaray, G. Konya, J. A. C. Broekaert and F. Leis, Chem. Anal. included. (Warsaw), 1990, 35, 311. 25 M. Willert-Porada, personal communication. 26 H. M. Kingston and L. B. Jassie, Introduction to Microwave Conclusions Sample Preparation. Theory and Practice, American Chemical Society, Washington, DC, USA, 1988. It has been shown that microwave assisted digestion at high 27 J. C. Farin� as, R. Moreno and J.-M. Mermet, J. Anal. At. pressure is very suitable for the digestion of diVerent ceramic Spectrom., 1994, 9, 841. ZrO2 powders prior to analysis with ICP atomic spectrometry. 28 D. Merten, J. A. C. Broekaert and A. Le Marchand, J. Anal. At. Both with microwave assisted digestion at high pressure and Spectrom., 1997, 12, 1387. conventional digestion at high pressure a complete dissolution 29 W. A. Van Borm and J. A. C. Broekaert, Anal. Chem., 1990, 62, 2527. was obtained for all the ceramic powders investigated. The detection limits obtained by ICP-OES were shown to be superior Paper 9/00676A in the case of conventional digestion at high pressure, as here 1098 J. Anal. At. Spectrom., 1999, 14, 1093&ndas

ISSN:0267-9477    

DOI:10.1039/a900676a

出版商:RSC    

年代:1999

数据来源: RSC