摘要:

ANALYST, FEBRUARY 1992, VOL. 117 157 Simultaneous Determination of Theophylline and Guaiphenesin by Third-derivative Ultraviolet Spectrophotometry and High-performance Liquid Chromatography Mohamed H. Abdel-Hay,* Mohie Sharaf El-Dint and Mustafa A. Abuirjeie Department of Medicinal Chemistry, Faculty of Pharmacy, Jordan University of Science and Technology, Irbid, Jordan Two methods are described for the simultaneous determination of theophylline and guaiphenesin in com bined pharmaceutical dosage forms. The first method depends on third-derivative ultraviolet spectro- photometry, with the zero crossing technique of measurement. Third-derivative amplitudes at 222 and 278 nm were selected for the assay of guaiphenesin and theophylline, respectively. The second method is based on high-performance liquid chromatography on a reversed-phase column using a mobile phase of 0.01 mol dm-3sodium dihydrogen phosphate-methanol-acetonitrile (8 + 2 + I ) (pH 5.5) with detection at 245 nm.Both methods showed good linearity, precision and reproducibility. The proposed methods were successfully applied to the determination of these drugs in laboratory-prepared mixtures and in capsules or elixir. Keywords: Theophylline and guaiphenesin determination; high-performance liquid chromatography; third-de riva tive ultra violet spec tro p h o tometr y; pharmaceutical form ula tions Theophylline is a xanthine derivative that relaxes smooth muscles, relieves bronchospasm and has a stimulant effect on respiration.' Guaiphenesin (glyceryl guaiacolate) is reported to reduce the viscosity of tenacious sputum and is used as an expectorant. 1 Theophylline and guaiphenesin in combination induce bronchodilation and assist the patient in coughing up viscid mucus,2 and have been used in the symptomatic treatment of bronchial asthma and other bronchospastic conditions.Theophylline and guaiphenesin together with their dosage forms, including combinations with other drugs, have been listed in various pharmacopoeias.3-5 The official compendia describe non-aqueous titrimetry, spectrophotometry and high-performance liquid chromatography (HPLC) for the determination of theophylline as the bulk drug and in dosage forms. Titrimetric, spectrophotometric and gas chromato- graphic (GC) procedures are described in various pharmaco- poeias for guaiphenesin in the bulk drug and in dosage forms.In combination with other drugs, guaiphenesin has been determined using second-derivative6 and differential7 spectro- photometry, colorimetry,s ~pectrofluorimetry~9.10 densi- tornetry,Il GC12 and HPLC. 13-16 Theophylline in combination with other drugs has been determined by conventional ultraviolet (UV) ,I7 differential's and derivative spectropho- tometry,1"2" colorimetry,21 differential thermal analysis,22 densitometry,23 thin-layer chromatography,24 GC25 and HPLC.2629 Some of the reported methods are not specific for the two drugs and some require extensive sample manipula- tion. To our knowledge, no methods have been described for both drugs in pharmaceutical dosage forms, except the assay method reported in the United States Pharmacopeia (USP),3 which involves the HPLC determination of theophylline and guaiphenesin in capsules.Hence it was considered desirable to develop a simpler and faster procedure that would serve as an alternative to the current official method.3 In this work two methods, based on selective derivative UV spectrophotometry and HPLC, are reported and the optimum experimental parameters for each method are described. * On leave from the Department of Pharmaceutical Analytical Chemistry. Faculty of Pharmacy, Alexandria University, Alexandria, On leave from the Department of Analytical Chemistry, Faculty Egypt. of Pharmacy, Mansoura University, Mansoura, Egypt. Experimental Materials Authentic samples of theophylline monohydrate, guaiphen- esin and phenacetin (employed as an internal standard) were kindly donated by Alexandria Co.for Pharmaceuticals and Chemical Industries (Alexandria, Egypt) and were used as received. Methanol and acetonitrile (Carlo Erba, Milan, Italy) were of HPLC grade; water was de-ionized and doubly distilled. All other chemicals were of analytical-reagent grade. Apparatus Spectrophotometric analysis was carried out on a Shimadzu UV-240 recording spectrophotomer in 1 cm matched quartz cells. The instrument parameters were spectral slit-width 2 nm, scan speed 10 nm s-l? recorder chart speed 10 nm cm-', wavelength range 190-320 nm and ordinate maximum and minimum settings k0.03. Third-derivative UV spectra were obtained with a Shimadzu attachment (optional programme/ interface, Model OPI-2) giving from first to fourth derivatives.Wavelength calibration was checked by using a holmium oxide filter, against air. The high-performance liquid chromatograph was composed of a Model 114 M single-piston pump (Beckman, Geneva, Switzerland), a Model 165 variable-wavelength UV detector (Beckman), an injector with a 20 pl loop (Beckman) and an SP 4270 integrator-plotter (Spectra-Physics, Basle, Switzerland). The detector wavelength was set at 245 nm at 0.1 a.u.f.s. Procedure for Derivative Spectrophotometry Calib ra ti0 n Standard solutions of theophylline and guaiphenesin were prepared in distilled water (10-30 pg ml-I). The third- derivative spectra were recorded over the wavelength range 190-320 nm, and appropriate third-derivative amplitudes were measured graphically (Table 1) and plotted against the corresponding concentration to obtain the calibration graph. Analysis of capsules and elixir An accurately weighed portion of the powder (mixed contents of 20 capsules) or an accurately measured volume of elixir, equivalent to about 150 mg of theophylline (90 rng of guaiphenesin), was transferred into a 100 ml calibrated flask158 ANALYST, FEBRUARY 1992.VOL. 117 and extracted (or diluted for elixir) by shaking for 10 min with 50 ml of distilled water. The resulting suspension was filtered, by washing through a filter-paper, into a 100 ml calibrated flask and then diluted to volume with distilled water (solution A). Then, 1.50 ml of solution A were pipetted into a 100 ml calibrated flask and the resulting solution was subjected directly to spectrophotometric analysis using distilled water as a reference.Procedure for HPLC Chromatographic conditions Routine analysis was carried out isocratically on a 5 pm reversed-phase Alltech-Macrosphere 300 C18 column (250 X 4.6 mm i.d.) using a mobile phase of 0.01 mol dm-3 sodium dihydrogen phosphate-methanol-acetonitrile (8 + 2 + l ) , adjusted to pH 5.5 with phosphoric acid, pumped at a flow rate of 1.2 ml min-1. Calibration Standard solutions of theophylline and guaiphenesin (2.5-15 pg ml-1) containing a fixed concentration (5 pg ml-1) of phenacetin (internal standard) were prepared in the mobile phase. Triplicate 20 pl injections were made for each solution and the peak area ratio of each drug to the internal standard was plotted against the corresponding concentration to obtain the calibration graph.2.0 Q, C t3 ; 1.0 $ 0 220 270 320 Wavelengthhm Fig. 1 theophylline (20 pg ml-1) in distilled water Zero-order spectra of A, guaiphenesin (20 pg m1-1) and B, + 0.03 m x TI 3 0 U -0.03 190 220 270 320 Wavelengthhm Fig. 2 Third-derivative spectra or A, guaiphenesin (20 pg ml-l) and B, theophylline (20 pg ml-l) in distilled water Analysis of capsules and elixir A 0.750 ml aliquot of solution A (prepared as above) was added to 1 .O ml of internal standard solution (500 pg ml-1) and the volume was adjusted to 100 ml with mobile phase. A 20 p1 volume of the final solution was injected into the chromato- graph . Results and Discussion Derivative Spectrophotometry The absorption (zero-order) UV spectra of theophylline and guaiphenesin in the 190-320 nm wavelength region are shown in Fig.1. Theophylline exhibits a peak and a shoulder at about 270 and 225 nm, respectively. Guaiphenesin, however, also absorbs over this wavelength region, with two peaks at about 270 and 220 nm. Because of the extensive overlap of the spectral bands of the two drugs, conventional UV spectropho- tometry cannot be used for their individual determination in a mixture. When third-derivative UV spectra are recorded, sharp bands of large amplitudes (Fig. 2 ) are produced, which may permit more selective identification and determination of the two drugs. As discussed elsewhere,3@-3* the choice of the optimum wavelength is based on the fact that the contribution of each component to the over-all derivative signal is zero at the wavelength at which the other component has the maximum absorption.Therefore, the third-derivative ampli- tudes at 278 nm (zero crossing of guaiphenesin) and at 222 nm (zero crossing of theophylline) were chosen for the simul- taneous determination of theophylline and guaiphenesin, respectively, in a binary mixture. Fig. 3 shows the third- derivative spectra of theophylline and guaiphenesin at several different concentrations; as can be seen, the position of the iso-differential point for each component is as stated above. Linear relationships between selected amplitudes from the third-derivative spectra and drug concentration were observed (Table 1). Least-squares regression analysis was carried out on the slope, the intercept and the correlation coefficient ( r ) .The +0.03 m x U U 3 0 -0.03 190 220 270 Wavelengthhm 320 Fig. 3 Third-derivative spectra of guaiphenesin (broken lines) and theophylline (solid lines) at several different concentrations in distilled water: guaiphenesin, 10, 15, 20, 25 and 30 pg ml-l; theophylline, 10, 15,20,25 and 30 pg ml-1ANALYST. FEBRUARY 1992, VOL. 117 159 Table 1 Analytical data for the calibration graphs (n = 5) for the determination of theophylline and guaiphenesin by third-derivative UV spectrophotometry and HPLC Regression equation Correlation Linearity range/ coefficient, Drug Method pg ml-l Slope Intercept r RSD(%)* Theophylline 3D2781 10-30 0.152 -0.011 0.9998 0.89 HPLC 2.5-15 0.066 -0.007 0.9998 1.13 Guaiphenesin 3D222t 10-30 0.148 -0.009 0.9996 1.11 HPLC 2.5-15 0.056 0.008 0.9993 1.53 * Relative standard deviation.t Third-derivative amplitude measured at 278 or 222 nm for theophylline and guaiphenesin, respectively. Table 2 Determination of theophylline and guaiphenesin in labora- tory-prepared mixtures by third-derivative UV spectrophotometry Table 3 Determination of theophylline and guaiphenesin in labora- tory-prepared mixtures by HPLC Theo- phylline to guaiph- Theoph ylline Guaiphenesin enesin ratio Taken/ Found/ Relative Taken/ Found/ Relative (dm) pg ml-1 pg ml-1 error (%) pgml-1 pgml-1 error (%) 1:2 10 9.86 -1.4 20 20.70 +3.50 2:3 20 20.20 +1.0 30 29.10 -3.00 3:4 15 14.54 -3.07 20 19.40 +3.00 1: 1 20 19.64 -1.80 20 20.00 0.00 5:4 25 25.39 +1.56 20 19.50 -2.50 3:2 30 30.33 +1.10 20 20.09 +0.45 A cc 0 4 8 Ti me/m i n Fig.4 HPLC trace of a 20 pl injection containing A, 10 pg ml-l of theophylline (3.37 min); B, 10 pg ml-1 of guaiphenesin (4.52 min); and C, 5 pg ml-1 of phenacetin (6.61 min) relative standard deviation calculated for the separate deter- mination of each drug was 0.89-1.11%, indicating good precision and reproducibility. In order to assess the validity of the proposed method for assaying each drug in the presence of the other, synthetic mixtures with different proportions of the two drugs were prepared and then assayed using the proposed derivative method. Satisfactory results were obtained for the recovery of both drugs (Table 2). Theo- phylline to guaiph- Theophylline Guaiphenesin enesin ratio Taken/ Found/ Relative Taken/ Found/ Relative (m/m) pg ml-1 pg ml-1 error (%) pgml-1 pg ml-1 error (%) 1 : 6 2.5 2.52 +0.80 15.0 14.91 -0.60 1:2 5.0 5.04 +0.80 10.0 9.83 -1.70 1:1.67 7.5 7.34 -2.13 12.5 12.78 +2.24 1.67:l 12.5 12.82 +2.56 7.5 7.39 -1.47 2:1 10.0 9.98 -0.20 5.0 5.07 +1.40 6 : l 15.0 15.21 +1.40 2.5 2.47 -1.20 Chromatography In order to effect the simultaneous elution of the two component peaks under isocratic conditions, the mobile phase composition was optimized.Phosphate buffer was chosen as the aqueous component. A satisfactory separation was obtained with a mobile phase consisting of the ternary mixture phosphate buffer (0.01 mol dm-3)-methanol-acetonitrile (8 + 2 + 1). The pH for the optimum resolution of the two drugs was 5.5. At lower pH values (3 or 4.5) a slight reduction in peak symmetry was observed, and the partial replacement of methanol with acetonitrile improved the resolution.Under the described chromatographic conditions, the analyte peaks were well defined, resolved and almost free from tailing. At a flow rate of 1.2 ml min-1, the retention times for theophylline, guaiphenesin and phenacetin (internal standard) were 3.37, 4.52 and 6.61 min, respectively (Fig. 4). Initial studies were performed while the effluent was monitored at 280 nm; the detector response (theophylline : guaiphenesin) was found to be in the ratio 10 : 1 for equal concentrations of the two drugs. Other wavelengths were therefore tried, and it was observed that on decreasing the wavelength below 280 nm the detector response for theophylline was decreased whereas that for guaiphenesin was increased.The optimum wavelength for detection was 245 nm, at which much better detector responses for both drugs were obtained (Fig. 4). The proposed method allows the determination of both drugs in capsules (labelled to contain 150 mg of theophylline and 90 mg of guaiphenesin per capsule) using the same dilution and the same injection volume and with reasonable responses for the two well resolved peaks. This is an advantage over the current USP3 procedure, which involves monitoring of the effluent at 280 nm, where the detector response of theophylline was found to be still much higher than that of guaiphenesin despite the different chromato- graphic conditions adopted in the USP procedure. For quantitative applications, linear calibration graphs were obtained with correlation coefficients better than 0.999 (Table 1).The good precision of the HPLC procedure was indicated by the relative standard deviation (1.13-1.53%). Results for the HPLC analysis of laboratory-prepared mixtures with different proportions of the two drugs are given in Table 3.160 ANALYST, FEBRUARY 1992, VOL. 117 Table 4 Results for the determination of theophylline and guaiphenesin in commercial formulations by third-derivative UV spectrophotometry and HPLC methods Theoph ylline" Guaiphenesin* Capsules t Elixirt Capsulest Elixir? Found RSD Found RSD Found RSD Found RSD Method (%) (%) (%) (%) (Yo) (Yo) (Yo) (Yo) Third-derivative UVspectrophotometry 100.02 0.95 100.34 1.13 100.15 1.54 100.97 1.03 HPLC 101.43 1.94 101.27 2.33 100.97 2.18 98.08 2.68 * Mean and relative standard deviation of five determinations given as a percentage of the claimed content.t Capsules were prepared in the laboratory to contain 150 mg of theophylline and 90 mg of guaiphenesin per capsule. Quibron elixir was claimed to contain 150 mg of theophylline and 90 mg of guaiphenesin per 15 ml. Analysis of Pharmaceutical Formulations The validity of the proposed methods for pharmaceutical preparations and the effect of possible interferences were studied by assaying Quibron elixir (labelled to contain 150 mg of theophylline and 90 mg of guaiphenesin per 15 mi) and laboratory-prepared capsules. The latter contained 150 mg of theophylline and 90 mg of guaiphenesin together with common additives and excipients, e.g., lactose, starch, talc and magnesium stearate. The results are given in Table 4. The results are accurate and precise, as indicated by the recovery (98.0&101.43%) and the relative standard deviation (0.95- 2.68% ) . Conclusions Derivative UV spectrophotometry and HPLC are suitable techniques for the reliable analysis of commercial formula- tions containing combinations of theophylline and guaiphen- esin. The most striking features of the derivative method are its simplicity, sensitivity and rapidity, which render it suitable for routine analysis in control laboratories. The HPLC method was shown to be a versatile reference method and may offer advantages over the derivative method for the selective determination of the two intact drugs in the presence of their degradation products or in a variety of matrices.The authors gratefully acknowledge the Deanship of Research at Jordan University of Science and Technology for financial support of this work through project No. 15/91. References 1 Reynolds, J. E. F., Martindale: the Extra Pharmacopoeia, Pharmaceutical Press, London, 29th edn., 1989, pp. 910 and 1532. 2 Gennaro, A. R., Remington's Pharmaceutical Sciences, Mack, Easton, PA, 17th edn., 1985, p. 874. 3 The United States Pharmacopeia, Twenty-First Revision, The National Formulary, X V I Edition, US Pharmacopeial Conven- tion, Rockville, MD, 1985, pp. 473 and 1042. 4 British Pharmacopoeia, HM Stationery Office, London, 1988, vol. 1, pp. 278 and 564. 5 European Pharmacopoeia, 11, Maisonneuve, France, 1971, p.382. 6 Yang, Q., Meng, Y., and Zhang, G., Yaowu Fenxi Zazhi, 1984, 4, 148; Anal. Abstr., 1985,47, 10E70. 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Tan, H. S. I., and Salvador, G. C.,Anal. Chim. Acta, 1985,176, 71. Rao, G. R., Avadhanulu, A. B., Giridhar, R., and Kokta, C. K., East. Pharm., 1988, 31, 141; Anal. Abstr., 1989, 51, 6E69. El-Yazbi, F. A., and Korany, M. A., Spectrosc. Lett., 1985,18, 543. Wahbi, A. A. M., and El-Omar, S. S., Alexandria J. Pharm. Sci., 1988, 2, 125. Tomankova, H., and Vasatova, M., Pharmazie, 1989,44, 197. Bambagiotti-Alberti, M., Pinzauti, S., and Vincieri, F. F., Pharm. Acta Helv., 1987, 62, 175. McSharroy, W. O., and Savage, I. V. E., J. Pharm. Sci., 1980, 69, 212. Muhammad, N., and Bodnar, J. A., J. Liq. Chromatogr., 1980, 3, 113. Carnevale, L., J. Pharm. Sci., 1983,72, 196. Heidemann, D. R., LC-GC, 1987, 5,422. Nowakowska, Z., Farm. Pol., 1987,43,14l;Anal. Abstr., 1988, 50, 5Ell. Saushkina, A. S., Vergeichik, E. N., Kompantseva, E. V., and Kilyakova, G. M., Farmatsiya, 1980,29,63; Anal. Abstr., 1981, 41, 3E16. Arnoudse, P. B., and Pardue, H. L., J. Autorn. Chem., 1986,8, 75. Hu, J., Wang, Y., and Kong, Q., Yaowu Fenxi Zazhi, 1988,8, 217; Anal. Abstr., 1989,51, 3D40. Aliev, A. M., and Guseinov, B. M., Farmatsiya, 1983, 32, 75; Anal. Abstr., 1984,46,4E15. Wesolowski, M., Int. J. Pharm., 1982, 11,35. Salama, 0. M., and Walash, M. I., Anal. Lett., 1989, 22, 827. Gaitonde, R. V., and Rivankor, U., Indian Drugs, 1987, 24, 486. Majlat, P., Pharmazie, 1984,39, 325. Juenge, E. C., Gurka, D. F., and Kreienbaum, M. A., J. Pharm. Sci., 1981, 70, 589. Chem, T.-M., and Chafetz, L., J. Pharm. Sci., 1981, 70,804. Roberts, S. E., and Delaney, M. F., J. Chrornatogr., 1982,242, 364. Low, G. K. C., Haddad, P. R., and Duffield, A. M., J. Chromatogr . , 1983, 261, 345. Garcia, S. F., Carnero, R. C., Marquez, G. T. C., Hernandez, L. M., and Heredia, B. A., Analyst, 1990, 115, 1121. Carnero, R. C., Heredia, B. A., and Garcia, S. F., J. Agric. Food Chem., 1990,38,178. Abdel-Hay, M. H., Elsayed, M. A., Barary, M. H., and Hassan, E. M., J. Pharm. Belg., 1990, 45,259. Paper 1 I02751 D Received June 10, 1991 Accepted August 26, 1991

ISSN:0003-2654    

DOI:10.1039/AN9921700157

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1992, VOL. 117 161 Determination of Tricarbonyl(2-methylcyclopentadienyl)manganese in Gasoline and Air by Gas Chromatography With Electron-capture Detection Virindar S. Gaind, Kusum Vohra and Fong Chai Occupational Health Laboratory, Ontario Ministry of Labour, 10 I Resources Road, Weston, Ontario, Canada M9P 3T7 Gas chromatography with electron-capture detection provides a highly sensitive technique for quantifying tricarbonyl(2-methylcyclopentadienyl)manganese (MMT). Airborne MMT can be collected by drawing a known volume of air through tubes containing XAD-2, and MMT concentrations as low as 0.001 mg m-3 can be monitored using a 10 I sample. No significant breakthrough was observed when 60 I of air were sampled at 1.0 I min-1 in the presence of a large excess of gasoline.The MMT in gasolines or other hydrocarbon fuels can be quantified by direct injection after dilution (1 + 99 or more) with hexane. Keywords: Tricarbony1(2-methylcyclopentadienyl)manganese determination; gas chromatography with electron-capture detection; gasoline; air samples Tricarbonyl(2-methylcyclopentadieny1)manganese (MMT), CH3C5H4Mn(C0)3 (Fig. 1), is an organometallic additive that improves the octane rating of gasoline. It is also used as a smoke abatement additive in fuels used for conventional reciprocating internal combustion engines and gas turbine engines, where 10-100 mg of MMT per litre of oil reduces smoke and particulate emissions by as much as 50-90% . l Tricarbonyl( 2-methylcyc1opentadienyl)manganese is highly toxic by all routes of exposure, i.e., inhalation, ingestion and skin absorption.Experimental data on animals have shown that exposure to MMT produces severe injury to the kidneys, liver, lungs and central nervous system.24 The American Conference of Governmental Industrial Hygienists (ACGIH) has adopted a Threshold Limit Value- Time Weighted Average (TLV-TWA) of 0.2 mg m-3 for MMT, determined as Mn.5 The TLV-TWA for other Mn compounds and Mn dust, however, is 5 mg m-3 as Mn. The considerable difference between the TLV-TWA of MMT and those of other Mn compounds suggested the need for a specific and sensitive analytical procedure capable of determining low concentrations of MMT in workplace atmospheres and in gasoline and other fuels. A number of analytical procedures have been reported for the determination of MMT, i.e., simple determination of total elemental Mn by atomic absorptions or gas chromatographic separation followed by flame-ionization detection .7 Gas chromatography (GC) coupled with an atmospheric pressure helium microwave plasma emission system has been shown to provide highly sensitive analytical capabilities8 as has gas chromatographic separation followed by atomic absorption from a d.c.argon plasma9 or atomic absorption in a slotted quartz tube analyser.10 Aue et aZ.*1 have described the application of atomic emission through a modified flame- photometric detection system in which the chemiluminescence from Mn is measured in order to quantify MMT. High- performance liquid chromatography coupled with a laser- excited atomic fluorescence spectrometric detector has also been described recently.12 0" Fig.1 Structure of MMT Although some of these procedures are specific and sensitive, their usage entails considerable modifications or interfacing of the available analytical systems, some of which are beyond the capabilities of many environmental analytical laboratories. This paper describes a simple, highly sensitive and specific gas chromatographic analytical procedure based on the use of electron-capture detection (ECD) for the quantification of MMT in gasoline and air. The unequivocal confirmation of the MMT peak was carried out by gas chromatography-mass spectrometry (GC-MS) in the electron impact (EI) or chemical ionization (CI) mode. Experimental Chemicals Tricarbonyl(2-methylcyclopentadienyl)manganese was ob- tained from Pfaltz & Bauer (Waterbury, CT, USA).Hexane and isooctane were of pesticide grade and the other chemicals were of analytical-reagent grade from Caledon Chemicals (Georgetown, Ontario, Canada). Apparatus The gas chromatograph used was a Hewlett-Packard (Avon- dale, PA, USA) Model 5840 instrument equipped with a 63Ni electron-capture detector. The mass spectrometer was a Hewlett-Packard Model 5985 instrument with an HP 7920 data system. The column in the GC-MS system was a fused silica capillary column (25 m x 0.32 mm i.d.) with a chemically bonded DB-5 methylphenylsilicone stationary phase, 1 pm thick, from J & W Scientific (Rancho Cordova, CA, USA). The air sampling pumps were portable Bendix (Ronceverte, WV, USA) Model 44 pumps, and the sorbent sampling tubes were from SKC (Eighty Four, PA, USA).The Tenax tubes had a front section with 50 mg of sorbent and a back-up section with 35 mg of sorbent. The XAD-2 sampling tubes had a front section containing 80 mg of sorbent and a back-up section with 40 mg of sorbent. The Test Atmosphere Generating System (TAGS) was from SRI (Menlo Park, CA, USA). Gas Chromatographic Conditions Calibration standards were prepared by diluting a known amount of MMT with isooctane and were prepared daily. The conditions used for the gas chromatographic separation of MMT were as follows: column, 10 ft x a in stainless steel162 ANALYST, FEBRUARY 1992, VOL. 117 0 z L r------ Timelm in Fig. 2 Gas chromatogram for ( u ) injection of 2 ng of MMT; and (b) gasoline containing 30 mg 1-l of MMT injected (2 pl) after a 1 + 249 dilution with hexane 100 80 f 60 2 40 20 A 0 Y 0) 0 4 8 12 16 20 MMT injectedlng Fig.3 Linearity of electron-capture detector response [peak area (%) (peak area for a 20 ng injection = 100%)] versus amount of MMT injected (2 p1 injection from a solution containing 0.1-10 pg m1-I of MMT). Correlation coefficient = 0.9986 with 10% FFAP on Chromosorb W, 80-100 mesh; oven temperature, 130 "C; injector temperature, 220 "C; and detector temperature, 240 "C. A typical chromatogram obtained by injecting 2 yl of MMT solution with a concentration of 1.0 yg ml-l(2 ng of MMT) is shown in Fig. 2(a). The linearity of the detector response to MMT was established by carrying out replicate injections (n = 6) of serially diluted standard solutions at various concentrations.The average area obtained for each concentration was plotted against the amount of MMT injected in order to establish the linearity of the electron-capture detector response in the range likely to be encountered in samples (Fig. 3). The lower detection level was calculated by serial dilution of standard solutions until the area of the peak obtained was three times the background noise level. Mass Spectra of MMT The EI mass spectrum of MMT [Fig. 4(a)] was obtained at 70 eV and scanned from 35 to 400 u. The CI mass spectrum was obtained at 2400 eV, scanning from 100 to 400 u and using 80 120 160 200 G t 40 .- 2 100 .- + - a3 80 a 60 40 20 b) 162 191 I I 0 I I 219 I 160 180 200 220 240 mlz Fig.4 MMT (a) EI mass spectrum of MMT; and (b) CI mass spectrum of methane as the reagent gas [Fig. 4(6)]. The oven temperature was held at 100 "C for 1 min and then raised to 150 "C at 10 "C min-1. Preliminary Evaluation of Sample Collection Media for Spiked MMT Five common air sampling media were evaluated for the collection of MMT, i. e., (1) glass or stainless-steel sampling bulbs; (2) sorbent tubes containing charcoal; (3) sorbent tubes containing Tenax; (4) sorbent tubes containing XAD-2; and (5) impingers containing isooctane. A known amount of MMT was spiked onto each of the sampling media. After equilibration for 10 min, an aliquot of air was withdrawn from the glass and stainless-steel bulbs with a glass syringe and 0.1 ml of the air sample was injected into the gas chromatograph in order to quantify MMT. The MMT in the sorbent tubes was extracted by placing the front and back-up sections in separate vials, adding 1.0 ml of hexane, sealing the vials with PTFE-lined caps and shaking for 30 min.The MMT spiked in the impingers containing isooctane was quantified by injecting 2 yl of the solution. Through a second set of spiked sorbent tubes and impin- gers, containing identical amounts of MMT, a measured volume of air was drawn for 50 min at 0.2 I min-1, using a sampling pump. The amount of MMT remaining in each sample was quantified in order to compare the effect of passage of air on the MMT collected. The percentage recovery in all of the above experiments was calculated by comparing the peak areas against freshly prepared standard solutions of MMT in hexane or isooctane. Stability of MMT Spiked on Sampling Tubes The stability of MMT spiked on XAD-2 sorbent tubes was evaluated by spiking replicate tubes with a known amount ofANALYST, FEBRUARY 1992, VOL.117 163 lMMT and storing them, wrapped in aluminium foil, at 4 "C in the dark. Duplicate tubes were analysed at various intervals in order to evaluate the stability of MMT adsorbed onto the tubes. Evaluation of Sample Collection Procedure for Airborne MMT The actual sample collection was evaluated by using a dynamically generated atmosphere of MMT in a TAGS having multiple sampling ports. The MMT (10% v/v) mixed with gasoline and maintained at room temperature in the sampling bulb of the TAGS was vaporized with a carrier stream of nitrogen at 0.1 1 min-1 and further diluted with air at 61 1 min-1 in a mixing chamber.The sampling ports were fitted with critical orifices, which allowed a precise sampling flow rate of 1.0 zk 0.05 I min-1. The XAD-2 and Tenax sampling tubes (which showed promising recovery of adsorbed MMT in the spiking experiments) were connected to the sampling ports. Each tube had a front section separated from a back-up section in order to assess whether there was any breakthrough during sample collection. Quantification of MMT in Gasoline Gasoline was diluted with hexane (1 + 99 or more) and 2.0 yl of the solution were injected into the gas chromatograph. The MMT furnished a distinct peak, usually without interferences. Results and Discussion Linearity, Precision and Minimum Detectable Amount of MMT Using ECD The electron-capture detector response was linear in the range tested, i.e., 0.2-20 ng of MMT. Fig.3 shows a plot of the average peak area versus the amount of MMT injected (2.0 pl injections of solutions containing 0.1-10 pg ml-1 of MMT). The correlation coefficient of the plot was 0.9986. The relative standard deviations for replicate injections of the above solutions ranged from 2.4 to 5.07%, showing the fairly good repeatability of the electron-capture detector response to MMT. The minimum detectable amount (MDA) of MMT using ECD was calculated to be 0.02 ng (2.0 yl injection of a 0.01 yg ml-1 MMT solution). This makes it possible to detect air- borne MMT at a concentration of 0.001 mg m-3 with a 10 1 air sample.The MDA for MMT when using ECD compares favourably with those reported for other analytical techniques, i.e., 12 ng of MMT by atomic emission from a d.c. plasma; 0.8 ng of MMT for atomic absorption in a slotted quartz tube atomizer; and 2 ng of MMT for atomic emission from the modified flame-photometric detector developed recently by Aue et al. 1 1 Stability of MMT Solutions and Spiked Sampling Tubes The standard solutions of MMT, at concentrations of 1.0-10.0 yg ml-1, were stable for a period of 4 d when kept in a refrigerator and protected from light by wrapping the contain- ers in aluminium foil. These solutions were, however, not very stable on exposure to strong daylight at room temperature and solid particles separated overnight, indicating decomposition of MMT.Ten XAD-2 sampling tubes, each spiked with 5.0 yg of MMT, showed no significant change in the amount of MMT present when kept in the dark at 4 "C for up to 7 d. This implies that when MMT samples collected on XAD-2 tubes are protected from strong light by wrapping them in aluminium foil and stored in a refrigerator, they can be kept for up to 7 d without any deterioration in the integrity of the samples. However, the calibration standards should be freshly pre- pared before analysing each batch of samples. Mass Spectra of MMT The mass spectrum of MMT in the EI mode [Fig. 4(a)] detected the molecular ion (mlz 218) with a relative abun- dance of 45%. The other expected fragments, mlz 190, 162 and 134, representing loss of three successive carbonyl (CO) fragments, were also present.The base peak had mlz 134, indicating loss of all three CO functions. The CI mass spectrum of MMT [Fig. 4(6)] showed M + 1 (mlz 219) as the only significant ion with very minor ions at mlz 191 and 162, representing loss of two CO units. Either of the two ionization modes can be used for quantifying MMT levels as low as 0.1 ng using the ion at mlz 218 or 219 with selected ion monitoring. Evaluation of Sample Collection Media for Spiked MMT The addition of MMT to a glass bulb produced erratic results; the MMT was not stable and showed a 90% loss over a period of 48 h. The stability of MMT in air did not improve when the samples were kept in a stainless-steel sampler, and attempts at direct sampling were abandoned. The results of recovery experiments on MMT spiked on various other sampling media are summarized in Table 1.The charcoal tubes gave erratic recoveries in the range 33-51%0. These were considered unacceptable. The recovery of MMT spiked on Tenax tubes was better, ranging from 60 to 85%, and showed little change after the passage of air. The XAD-2 tubes and the impingers containing isooctane, however, gave nearly quantitative recovery of the spiked MMT. The passage of air did not affect the recovery of MMT from Tenax, XAD-2 or the impingers containing isooctane. This suggests that tubes containing XAD-2, and impingers containing isooctane, are suitable for the collection of Table 1 Recovery of spiked MMT from various sampling media before and after passage of air MMT recovered (%) Sampling medium Isooctane Charcoal Tenax impinger XAD-2 MMT spikedlpg NA* A t NA* A t NA* AT NA* A t 10 50 51 85 81 100$ 103 100 100 1 .o 35 33 60 62 loo$ 101 100 93 0.1 35 35 80 78 1001 108 95 93 * NA: Not aerated.t A: After passage of 10 1 of air. + Solutions used for calibration.164 ANALYST, FEBRUARY 1992, VOL. 117 Determination of MMT in Gasoline The maximum permissible amount of MMT in unleaded gasolines is 18 mg 1-1 (as Mn), which corresponds to approximately 72 mg 1-1 as MMT. As MMT can be detected at very low concentrations through the use of ECD, the gasoline samples were diluted (1 + 99 to 1 + 499) with hexane in order to monitor the level of MMT in gasoline. The chromatogram shown in Fig. 2(b) represents a gasoline with an MMT level of 30 mg 1-1 and was obtained after a 1 + 249 dilution of the gasoline with hexane.The peaks due to most of the other constituents of the gasoline show little interference. However, in some gasolines containing excessively large concentrations of sulfur compounds, the peak due to MMT might be subject to interference from the sulfur-containing components if these are not completely separated. For such samples, it is necessary to quantify the amount of MMT by using a mass spectrometer in the CI mode in order to confirm the identity of the peak unequivocally. Table 2 Comparison of the amount of MMT collected on XAD-2 and Tenax tubes from a dynamically generated atmosphere MMT collected/p.g Sample No. Front 1 18.5 2 16.4 3 17.4 4 16.7 5 15.4 6 15.2 Mean: SD : RSD : Back 0.7 0.6 0.6 0.5 0.5 0.5 XAD-2 _ _ _ - Total 19.2 17.0 18.0 17.2 15.9 15.7 17.2 1.3 7.6% Tenax Front Back 2.4 1.1 1.6 0.8 2.5 1.3 1.7 0.7 2.7 1.4 3.1 1.4 Total 3.5 2.4 3.8 2.4 4.1 4.5 3.5 0.9 25.5% airborne MMT samples.However, as the sorbent sampling tubes are preferable for personal sampling, Tenax and XAD-2 tubes were subjected to further evaluation for sample collec- tion from dynamically generated atmospheres of MMT as both showed adequate recoveries of spiked MMT. Evaluation of Sample Collection on XAD-2 and Tenax From a Dynamically Generated MMT Atmosphere Table 2 shows the amounts of MMT found on XAD-2 and Tenax sampling tubes when a dynamically generated MMT atmosphere was used for multiple sample collection at a flow rate of 1.0 1 min-1 for a period of 60 min.The total amount of MMT collected on XAD-2 ranged from 15.7 to 19.2 pg, with an average value of 17.2 pg, a standard deviation (SD) of 1.3 pg and a relative standard deviation (RSD) of 7.6% , indicating fairly good precision. The amount of MMT detected in the back-up section of the XAD-2 sampling tubes was less than 4% of that detected in the front section in all instances, showing that there was no breakthrough even with an air sampling rate of 1 1 min-1 and in the presence of a large excess of gasoline vapour and an MMT loading of up to 17 pg. The Tenax tubes, however, showed poor efficiency for sampling airborne MMT as evidenced by the small amounts of MMT collected, the high RSD and the very high breakthrough to the back-up sections for all the samples (Table 2).This discrepancy between the fairly good adsorption efficiency of MMT spiked on Tenax tubes (Table 1) and the very poor collection efficiency of MMT from a dynamically generated atmosphere is probably due to poor adsorption and over- loading of Tenax with gasoline vapour. The air concentration of MMT in the dynamically generated atmosphere was about 0.3 mg m-3. The ACGIH adopted TLV-TWA concentration of MMT is 0.2 mg m-3 (expressed as Mn) and corresponds to about 0.7 mg m-3 of MMT. The results of sampling a dynamically generated MMT atmosphere show that XAD-2 tubes with two sections are the most suitable for the collection of airborne MMT. Conclusions Gas chromatography with ECD provides a simple and highly sensitive analytical procedure for quantifying MMT in air or gasoline.Airborne MMT can be collected by drawing air at 1.0 I min-1 through XAD-2 sampling tubes for 10-60 min. The MMT in gasoline can be quantified by appropriately diluting the gasoline with hexane or isooctane and injecting the solution directly into a gas chromatograph equipped with an electron-capture detector. An unequivocal confirmation of the presence of MMT can be carried out by GC-MS in either the EI or CI mode. 1 2 3 4 5 6 7 8 9 10 11 12 References Craig, P. J., Organometallic Compounds in the Environment: Principles and Reactions, Longman, London, 1986, ch. 10. McGinley, P. A., Morris, J. B., Clay, R. J., and Gianutsos, G., Toxicol. Lett., 1987, 36, 137. Browning, E., Toxicology of Metals, Butterworth, London, Hanzlik, R. P., Bhatia, P., Stitt, R., and Traiger, G. J., Drug Metab. Dispos., 1980, 8, 428. Threshold Limit Values and Biological Exposure Indices, for 2990-2991, American Conference of Governmental Industrial Hygienists, Cincinnati, OH, 1990, p. 26. Smith, G. W., and Oalmby, A. K., Anal. Chem., 1959,31,1798. DuPuis, M. D., and Hill, H. H., Anal. Chem., 1979, 51,292. Quimby, B. D., Uden, P. C., and Barnes, R. M., Anal. Chem., 1978, 50, 2112. Uden, P. C., Barnes, R. M., and DiSanzo, F. P., Anal. Chem., 1978, 50, 852. Coe, M., Cruz, R., and van Loon, J. C., Anal. Chim. Acta, 1980, 120, 171. Aue, W. A., Miller, B., and Sun, X.-Y., Anal. Chem., 1990,62, 2453. Walton, A. P., Wei, G.-T., Liang, Z., Michel, R. G., and Morris, J . B., Anal. Chem., 1991, 63, 232. 1966, pp. 185-196. Paper lf00869B Received February 22, 1991 Accepted October 16, 1991

ISSN:0003-2654    

DOI:10.1039/AN9921700161

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1992, VOL. 117 165 Fourier Transform Infrared Spectroscopic Studies on the Interaction Between Copper(il), Amino Acids and Marine Solid Particles Wang Xiulin Department of Chemistry, Fudan University, Shanghai 200433, People's Republic of China Fourier transform infrared (FTIR) spectroscopy was employed t o characterize the interaction between Cull, amino acids (AAs) and solid particles. The ion exchange between Cull and marine solid particles causes a stepwise change in v(0H) of surface H-bonding hydroxyl groups with an increase of the 'ion-exchange amount' of Cull at room temperature. The bands of the -COO- and -NH3+ vibrational modes of adsorbed AA and Cu"-AA surface complexes in a solid matrix are first isolated from the overlapping solid particle matrix by use of the spectral subtraction approach.With respect t o the corresponding free AA, v(NH3+) and 6(NH3+) of the adsorbed AA shift 10-39 and 13-49 cm-1, respectively, toward higher and lower frequencies, whereas the variation of v(C00-, as) and v(C00-, s) is less than 4 cm-1, nearly equal t o the resolution of the IR spectrometer. This indicates that the surface hydroxyl group associates with the amino group of the amino acid in a solid matrix rather than with the carboxyl group. Similarly, not only do thev(NH3+) and 6(NH3+) of the CuIl-AA surface complex in the solid matrix shift 16-27 and 34-42 cm-1, respectively, toward higher and lower frequencies, but also v(C00-, as) and v ( C 0 0 - , s) shift 15-32 and 10-25 cm-1, respectively, toward higher and lower frequencies, indicating that as a 'bridging' reagent the AA joins Cull and the surface hydroxyl group, respectively, through the amino group and the carboxyl group t o form a Model II ternary surface complex (TSC).The IR bands at 287-360 cm-1, due t o the Cu-0 stretching mode, were detected and thus further confirmed the formation of the TSC model between Cull, AAs and marine solid particles. Keywords: Fourier transform infrared spectroscopy; marine solid particle; copper(//); amino acid; adsorption Marine solids include suspended particles and surface sedi- ment, and consist mainly of oxides (such as Mn, Fe, Si and Al oxide), clays (such as montmorillonite, illite and kaolinite), and calcium carbonate.'-3 In general, bare metal ions and hydroxyl groups are two examples of the types of sites on the surface of solid particles.4 It has been observeds--7 that the isotherms of the ion exchange between trace metal ions and solid particles in sea-water can be classified as the 'plateau' type.Zhang and Liuh suggested that the plateau isotherms can be interpreted in terms of the interfacial stepwise ion- exchange theory they have proposed.6-7 However, so far the interfacial stepwise ion-exchange theory has not been demon- strated experimentally. Amino acids (AAs), which are found at a concentration of about 0.06 mg 1-1 in sea-water,x are one of the most important marine organic materials identified so far. It appears that AAs react with surface hydroxyl groups on oxides mainly through the amino group.9 However, infrared (IR) spectra have rarely been reported for AAs in a solid particle matrix, although 1R spectra of simple organic compounds, such as NH3, carboxylic acids and amines, in a solid matrix have been measured.lO-l4 Further, there are major limitations in the spectrometric examination of AAs in a solid matrix, including low concen- tration of AAs, and spectral interferences from the solid particle matrix.It is generally accepted that the formation of a ternary surface complex (TSC) causes some organic substances to promote the ion exchange between trace metal ions and solid particles when there is an interaction between trace metal ions, solid particles and organic rnatter.h.ls--'7 However, the IR spectra have rarely been reported for the same reasons as those given above for the amino acid-solid particle system.Leckie and co-workers16.1X proposed three models of TSC and Bowers and Huangly suggested that these models can be distinguished by comparing the graph of percentage adsorp- tion of metal ions versus pH with that of percentage adsorption of organic compound versus pH. However, the method of Bowers and Huang may be without value because neither of the pH curves was simple and both varied with the experimental conditions.20 In the present work, emphasis is placed on the IR spectroscopic evidence for the interfacial stepwise ion-exchange theory,6.7 and obtaining, by use of the spectral subtraction approach, the distinctive IR bands of AAs in a solid matrix for 'amino acid-solid particles' and 'amino acid-copper( ii)-solid particles' s ys tems .Experimental Reagents and Materials The CuC12.2H20 used was of analytical-reagent grade and all AAs used were of spectroscopic grade. The a-SiO2 was supplied by Beijing Chemical Reagent Co. (Beijing, China). Solid particle samples were prepared as described previously for goethite,s y-AIOOH6 and CaC03.3 The natural clays of montmorillonite and illite (Zhejiang Mineral Co., Hangzhou, China) were re-purified and transformed from calcium-clay to sodium-clay according to a previous method.6 All the solid particle samples were iddtified by using X-ray diffraction graphs. The specific surface area of these solid particles in aqueous solution, determined by the approach developed by Wang,2s are 742 m2 g-1 for y-AIOOH, 451 m2g-1 for goethite, 58 m2 g-1 for SO2, 39 m2 g-1 for illite, 59 m2 g-1 for montmorillonite and 296 m2 g-1 for CaC03.Pre-treatment Approximately 50 mg of the solid sample powder were accurately weighed and placed in an 80 ml centrifuge tube containing 50.0 ml of a solution of either Cu", or AA, or Cu"-amino acid of known concentration. The mixture was shaken for 3 h at 25.0 "C, and then the solid powder was separated from the supernatant by using an LXJ-64-01 centrifuge (Beijing Medicine Instrument Factory, Beijing, China). Measurements of pH in the supernatant were made with a Corning combined pH-reference electrode with the use of a Radiometer pHM84 meter (Copenhagen, Denmark). The concentration of CU" in the supernatant was determined by using a PE-3030 atomic absorption spectrometer (Perkin- Elmer, Norwalk, CT, USA), and of the AA by using a Hitachi 801 amino acid autoanalyser (Hitachi, Tokyo, Japan).In order to obtain a relatively reliable spectrum, revealing the actual surface species in aqueous solution, the solid powder166 ANALYST, FEBRUARY 1992, VOL. 117 was dried by evacuation at ambient temperature from atmospheric pressure to 8.8 x 104 Pa for 24 h. This procedure was carried out because some change in the IR spectra arose when the solid powder was pre-desiccated by evacuating to high vacuum, as occurred by heating in air at higher temperatures. 11~3922 Sample Preparation and Infrared Spectral Measurements The desiccated solid powder (about 20 mg) was mixed with dry potassium bromide (about 100 mg), ground in an agate mortar and subjected to a pressure of 8 x 106 Pa in an evacuated die, to produce a clear transparent disc with a diameter of 12 mm.Infrared spectra of the discs were obtained with 4 cm-1 resolution and a medium interscan correlation, using a Nicolet 10-DX Fourier transform (FT) IR spectrometer (Nicolet , Madison, WI, USA), equipped with a Model 1280 acquisition system. Fifty-four cumulative scans yielded an adequate signal-to-noise ratio in the spectra. Results and Discussion FTIR Spectra of the Copper(u)-Marine Solid Particle System Fig. 1 shows F-TIR spectra of illite with various ion-exchange amounts (IEAs) of Cull. Absorption, observed at 3441 cm-1 for all copper(i1)-illite samples (Table l), had the same band as the v(0H) of the surface H-bonding hydroxyl group (SHHG) on pure illite.The only band at 3441 cm-1 appears for an IEA of Cull of 0.52 pmol m-2. The bands at 3523,3574, 3316 and 3361 cm-* appear consecutively with a gradually increasing IEA of CulI from 18.1 to 158 pmol m-2, while the band at 3361 cm-1 disappears when the IEA of Cu" increases up to 202 pmol m-2. These bands may be assigned to v(0H) of SHHG as has been done in similar earlier analyses.12J3.24 It should be noted that the weaker band at 3316 cm-1 could be overlapped by the band at 3574 cm-1, which only appears as the IEA of Cu" is somewhat smaller (Fig. 1). These are the characteristic bands of v(0H) of surface hydroxyl groups on hydrous CuO (Fig. 2). The hydroxyl bands were chosen to examine the influence of CulI ion exchange upon hydroxyl 4000 3500 3000 25004000 3500 3000 25004000 3500 3000 2500 Wavenumberlcm - l Fig. 1 Stepwise change in v(0H) of SHHG with the increase of the IEA of Cu" at room temperature for: (a) Cu"-illite systems with an IEA of Cull of: A, 0 (pure illite); B, 0.52; C, 18.1; D, 20.1; E, 49.5; F, 158; and G, 202 pmol m-2; (b) Cu1I-rnontmorillonite systems with an IEA of Cull of: A, 0 (pure montmorillonite); B, 1.87; C, 10.8; D, 11.9; E, 13.1; F, 14.6; G, 18.8; and H, 22.9 pmol m-2; (c) Cu"-CaC03 systems with an IEA of CuI1 of: A, 0 (pure CaC03); B, 1.70; C, 10.6; and D, 26.1 pmol m-2 Table 1 Stepwise change in v(0H) of SHHG on clays and CaC03 due to the ion exchange between Cull, clays and CaC03 Mite IEA of Solid Cu"/ Y( OH)*/cm- 1 particle pH pmolm-2 6.38 0.52 3441m,br - - - 6.05 20.1 3441 m 3523m - 5.97 49.5 3441 m 3 5 2 3 ~ 3 5 7 4 ~ ~ - - 5.57 158 3441 m 3523m 3574m 3316m 3361m 8.11 202 3441 m 3523m 3574m 3316m - - - Montmorillonite 2.04 5.56 4.76 5.76 5.95 6.06 5.74 CaC03 6.15 5.82 5.65 1.87 10.8 11.9 13.1 14.6 18.8 22.9 10.6 26.1 1.70 3427 s,br 3427 m 3427 m 3427 m 3427 m 3427 m 3427 m 3427 m , br 3427 m 3427 s - - 3523m - 3523m - 3523m - 3523m 3574w 3523m 3574w 3523m 3574w 3359m - - 3359m 3574w 3316s - - * br = Broad, s = strong, m = medium, w = weak, and vw = very weak.ANALYST, FEBRUARY 1992, VOL.117 167 4000 3500 3000 2500 Wavenurnbertcm- Fig. 2 Characteristic bands at 3316 and 3574 em-* due to surface hydroxyl groups on hydrous CuO 4000 3500 3000 25004000 3500 3000 2500 Wavenurnbertcm-l Fig. 3 Stepwise change in v(0H) of SHHG with an increase of the IEA of Cu" at room temperature for: (a) Cu"-y-AlOOH systems with an IEA of Cu" of: A, 0 (purey-A100H); B, 1.32; C, 3.61; and D, 8.06 pmol m-2; (b) Cu"-goethite systems with an IEA of Cu" of: A, 0 (pure goethite); B, 2.37; C, 3.56; D, 4.01; and E, 8.38 pmol m-2 groups on illite, indicating that with an increase in the IEA of Cull, copper(i1)-illite ion exchange results in a stepwise change in the v(0H) of SHHG on illite, but no change in v(0H) of the free surface hydroxyl group (FSHG).Similarly, owing to the copper(I1)-montmorillonite ion exchange, a stepwise change in v(0H) of SHHG on montmorillonite was observed, but no change in v(0H) of FSHG [Fig. l(b)]. The band, observed at 3427 cm-1 for all copper(i1)-montmorillonite samples with various IEAs of Cu" (Table 1), had the same IR absorption as the v(0H) that was assigned to SHHG on pure montmorillo- nite.25 Only the band at 3427 cm-1 appears for an IEA of 1.87 pmol m-2 of Cu".Bands at 3523 and 3574 cm-1, due to v(0H) of SHHG , appear consecutively with gradually increasing IEAs of Cu", i . e . , the band at 3523 cm-1 appears when the IEA of CuI1 is 10.8,11.9 and 13.1 pmol m-2, while the bands at 3523 and 3574 cm-1 appear when the IEA of Cull increases to 14.6, 18.8 and 22.9 pmol m-2. It was shown from FTIR spectra25 that CaC03 had similar bands of surface hydroxyl groups to clays. Fig. l(c) reveals that CuI1-CaCO3 ion exchange results in a stepwise change in v(0H) of SHHG but no change in v(0H) of FSHG on CaC03. The description for the spectra [Fig.l(c)] is similar to that for the copper(i1)-illite and copper(i1)-montmorillonite systems, i.e., the bands at 3427, 3359, 3574 and 3316 cm-1 appear consecutively when the IEA of Cu" gradually increases from 10.6 to 26.1 pmol m-2 (Table 1). Fig. 3(a) shows FTIR spectra of y-A100H with different IEAs of Cu", indicating that Cu"-y-AlOOH ion exchange causes a stepwise change in v(0H) of SHHG rather than framework hydroxyl groups (FHG) in y-AlUOH, and that a Table 2 Stepwise change in v(0H) of SHHG on oxides due to the ion exchange between Cu" and oxides IEA of Cu"/ v(OH)*/crn - pH wmolm-2 4.95 3.61 3312s,br 3427w - 5.53 8.06 3312s,br 3427w 3512w 4.46 3.56 - 5.17 4.01 - 3427w 3512w 5.54 8.38 - 3427w 3512w Oxide y- AIOOH 5.73 1.32 3312s,br - - - - - Goethite 5.18 2.37 3 4 2 7 ~ - * br = Broad, s = strong, and w = weak.band at 3312 cm-1, due to v(0H) of SHHG on y-A100H,25 appears for all CuII-y-AIOOH samples with various IEAs of Cu" (Table 2). Only the band at 3312 cm-1 was detected for a smaller IEA of Cu" (1.32 pmol m-2); however, bands at 3427 and 3512 cm-1 appear consecutively when the IEA of Cu" gradually increases, i.e., the band at 3427 cm-1 appears when the IEA of Cull is 3.61 pmol m-2, while bands at 3427 and 3512 cm-1 occur when the IEA of Cu" increases up to 8.06 pmol m-2. Fig. 2(b) shows FTIR spectra of goethite with various IEAs of Cu", indicating that the bands at 3427 and 3523 cm-1, due to v(0H) of SHHG,12.23?24 appear consecu- tively when the IEA of Cu" gradually increases from 3.56 to 8.38 pmol m-2 (Table 2), and that the intense and broad band at 3133 cm-1, due to v(0H) of FHG in goethite,25 appears in the spectrum. However, as the IEA of Cu" was 2.37 pmol m-2 only an intense and broad band at 3133 cm-1 was detected, consistent with the finding that the band due to v(0H) of SHHG on goethite was overlapped by its intense and broad band at 3133 cm-1.25 Summarizing the spectra described above, it can be concluded that with an increase of the IEA of Cu", Cu" ion exchange causes a stepwise change in v(0H) of SHHG on solid particles, but no change in v(0H) of either FSHG on illite, montmorillonite and CaC03, or FHG in goethite and y-AIOOH, and that over the range of the IEA of Cu", two stepwise changes in v(0H) of SHHG were detected except that four were observed for the copper(ii)-illite system, presumably due to different IEAs of Curl.Further, illite, montmorillonite and CaC03 had essentially the same stepwise change in v(0H) of SHHG due to copper(i1)-solid particle ion exchange. The band at 3523 cm-1 for clays or 3359 cm-1 for CaC03 appears first, then the bands at both 3574 and 3316 cm-1 and finally the band at 3316 cm-1, and then disappears. Similarly, goethite and y-A100H had essentially the same stepwise change in v(0H) of SHHG due to the ion exchange between Cu" and oxides. The first stepwise change is indicated by the appearance of a band at 3427 cm-1, and then at 3512 cm-1. The second change for oxides is different from that for clays and CaC03, i.e., a band at 3512 cm-1 appears for oxides, whereas bands at 3574 and 3316 cm-1, due to the characteristic v(0H) of SHHG on CuO, appear for clays and CaC03.Consequently, these results provide the first IR evidence for the interfacial stepwise ion-exchange theory,6 and have elucidated the mechanism of the stepwise ion exchange between Cu" and marine solid particles more clearly. FTIR Spectra of Adsorbed Amino Acids in a Solid Matrix for Amino Acid-Marine Solid Particle Systems The FTIR spectra were measured for free glycine, alanine, histidine, glutamic acid and aspartic acid, and subsequently some band frequencies were assigned to the -NH3+ stretching and bending modes v(NH3+) and 6(NH3+), and the -COO- symmetric and asymmetric stretching modes v(C00-, S) and v(C00-, as) (Table 3) according to a similar study.26-29168 ANALYST, FEBRUARY 1992, VOL.117 It is difficult to recognize the IR bands of adsorbed AAs in a solid particle matrix from the IR spectra of amino acid-solid particle systems due to the low concentration of AAs and spectral interferences from the solid matrix. In order to obtain the distinctive AA bands, particularly v(NH3+), 6(NH3+), v ( C 0 0 - , as) and v ( C 0 0 - , s), the spectral subtraction approach was used to isolate the adsorbed AA bands. The spectrum of the amino acid-solid system was obtained first and then the solid particle background absorption was subtracted. The IR spectra of adsorbed AAs are given in Fig. 4(a) for the glycine-, glutamic acid- and alanine-illite systems, in Fig. 4(b) for the glycine-, histidine-, glutamic acid- and alanine-montmorillonite systems, and in Fig.4(c) for the amino acid-CaC03 and y-A100H systems and the Table 3 Infrared vibrational bands generated from -NH3+ and -COO- groups of free amino acids v(NH3+)/ v(C00-, as)/ 6(NH3+)/ v(C00-, s)/ Amino acid cm-1 cm-1 cm-1 cm-1 Glycine 3186 1592 1522 1413 Alanine 3086 1594 1506 1413 Histidine 3130 1578 1539 1409 Aspartic acid 3141 1617 1501 1420 Glutamic acid 3059 1616 1516 1422 glycine-Si02 system at different pH values. Note that 6(NH3+) and v ( C 0 0 - , s) of an amino acid in a CaC03 matrix are difficult to obtain owing to very strong IR interference from the CaC03 matrix near 1500 cm-1, and also v(NH3+) of AAs in the y-A100H matrix because of very strong interfer- ence from the y-A100H matrix in the range 3000-3500 cm-1. The FTIR spectra of adsorbed AAs (Fig.4) are different from those of the corresponding free AA; therefore, for amino acid-clay , amino acid-oxide and amino acid-CaC03 systems, v(NH3+), 6(NH3+), v ( C 0 0 - , s) andv(CO0-, as) of the adsorbed AAs were assigned as in a similar analysis26-29 and chosen to examine the interaction between AAs and solid particles, as the main IR absorption comes from amino and carboxyl groups. Their frequency shifts with respect to free AAs, Av(NH3+), A6(NH3+), Av(C00-, s) and Av(C00-, as), are listed in Table 4. For the amino acid-illite, amino acid-montmorillonite, amino acid-y- AlOOH and amino acid -CaC03 systems, Av(C00-, as) and Av(C00-, s) are less than 4 cm-1, nearly equal to the resolution of the IR spectrometer (4 cm-*), whereas the variations of v(NH3+) and 6(NH3+) are 10-39 and 13-49 cm-1, respectively, moderately more than the resolution.For the glycine-Si02 system, the variations of v(NH3+), 6(NH3+), v ( C 0 0 - , s) and v ( C 0 0 - , as) are similar to those of the above systems when the pH ( 4 . 5 9 ) is more than the pH,,, of SiO2, which can be defined as the pH value at zero net adsorption of H+ and OH- ions as these ions are presumably the potential-determining species, ( a ) al c ro t 5 4- c .- 2 I- I I I 3400 2800 2200 1600 3400 2800 2200 1600 3400 2800 2200 1600 Wavenumberkm Fig. 4 FTIR spectra of adsorbed amino acids in a solid particle matrix at room temperature for: (u) A , illite-histidine; B, illite-glycine; and C, illite-glutamic acid systems; ( b j A, montmorillonite-glycine; B. montmorillonite-histidine; C, montmorillonite-glutamic acid; and D, montmorillonite-alanine systems; (cj A, CaC03-glycine; B , y-A100H-glycine; and C, y-A100H-aspartic acid systems, and Si02-glycine system with D, pH 6.59 and E, pH 5.23 Table 4 Infrared bands of adsorbed amino acids in a solid particle matrix, and the frequency shifts (cm-1) with respect to the corresponding free amino acid Amount of amino acid Solid Amino adsorbed/ particle acid pmol m-2 PH v(NH3+) Av(NH3+) v(C00-, as) Av(C00-, as) 6(NH3+) A6(NH3+) Illite Glycine 1.79 6.26 3178 10 1596 4 1509 - 13 Histidine 0.82 7.64 3148 18 1582 4 1496 - 46 Glutamic acid 0.84 3.22 3098 39 1620 4 1498 - 18 Mont- Glycine 0.93 8.30 3180 12 1596 4 1497 - 24 morillonite Histidine 0.59 7.92 3160 30 1581 3 1494 - 45 Glutamic acid 0.52 3.35 3072 13 1620 4 1497 - 19 Alanine 0.58 6.53 3121 35 1596 2 1480 - 26 CaC03 Glycine 0.33 7.87 3148 16 1596 4 y-A100H Glycine 0.084 6.60 - - 1596 4 1487 - 35 Aspartic acid 0.049 2.98 - - 1619 2 1459 - 49 SiOz Glycine 0.99 6.59 3180 12 1596 4 1501 -21 GI ycine 1.08 5.23 3164 - 1596 4 1488 - 34 3266* 3332* *: v(NH2).ANALYST, FEBRUARY 1992, VOL.117 169 whereas Av(CO0-, s) and Av(C00-, as) are also less than 4 cm-1. However, three bands at 3164,3266 and 3332 cm-1, due to v(NH2), were detected when the pH ( ~ 5 . 2 3 ) is less than Compared with the resolution of the IR spectrometer the variations of the -COO- stretching modes are meaningless for interpreting the surface species of carboxyl groups in a solid matrix. In contrast, the change of the -NH3+ stretching and bending modes is sufficiently large to be detected, indicating its great value to elucidation of the interaction mechanism of the AA with the solid matrix.Therefore, it is reasonable to infer that the amino group of an amino acid in a solid matrix, rather than the carboxyl group, reacts with the solid particle. Sokoll and ~o-workers13~30 found that when the N-H bending mode shifts toward lower frequencies, by, e.g., 13-49 cm-1, as has been observed in this work, the amino acid presumably reacts with the surface hydroxyl groups on a solid particle through its amino group in the following manner: pH,,,. 1- I H 3400 2200 1200'400 250 Wavenumberlcm- Fig. 5 FTIR spectra of copper(1i)-amino acid complexes in a clay matrix at room temperature for the copper(ri)-histidine-illite system with an IEA of Cu" of: A, 0.20; and B, 5.56 pmol m-2; for the copper(ii)-histidine-montmorillonite system with an IEA of Cull of: C, 0; D.1.07; E, 2.40; F, 6.67; and G, 7.14 pmol m-2; and for the copper(ii)-glycine-montmorillonite system with H. an IEA of Cu" of 5.39 pmol m-2 If this is the mechanism, the N-H bond is reinforced and v(NH3+) should shift to higher frequencies. Actually, the increase in v(NH3+) by 10-39 cm-1 (Table 4) further supports the mechanism illustrated in eqn. (1). When pH > pH,,, the mechanism of the interaction between glycine and Si02 is the same as the other systems. However, when pH < pH,,, three characteristic (N-H) bands of the metal ion-amino acid complex*s8.31 were observed. It can be inferred that the amino group of an AA in the Si02 matrix coordinates with the bare metal ion on the surface of SO2: / R' (2) 4 9 MM--NH2-C-C0O- \ R!, Furthermore, for AA-Si02 systems (Table 4), the decrease in v(NH3+), owing to the formation of the surface complex at pH < pH,,, [eqn.(2)], is 13 cm-1, more than that of the surface species at pH > pH,,, [eqn. (l)]. This is consistent with the finding that the amount of surface hydroxyl groups on Si02 is the lowest in the marine solid particles (such as Al, Fe, Mn and Si oxides, clays and CaC03) and so its surface hydration level is assumed to be much weaker.*' 4000 3500 30004000 3500 30004000 3500 3000 2500 Wavenumberkm - 1 Fig. 6 Change in v(0H) of SHHG due to the interaction among copper(Ii), AAs and marine solid particles at room temperature for: ( a ) copper(1i)-illite-histidine with an IEA of Cull of: A, 15.0 pmol m-2; B, copper(i1)-illite-glycine with an IEA of Cu" of 16.3 p o l m-*; and C, 17.4 pmol m-*; (b) A, copper(i1)-montmorillonite- glycinc with an IEA of Cu" of 1.57 pmol m-*; and B, 6.72 pmol m-2; C, copper(i1)-montmorillonite-aspartic acid with an IEA of Cull of 1.28 ymol m-2; and D, 6.59 pmol m-*; (c) Cu1l-y-A1O0H-histidine with: A, an IEA of Cull of 4.09 ymol m-2; B, Cull-y-AIOOH-aspartic acid with an IEA of Cu" of 3.33 pmol m-2; and C, 6.79 pmol m-* FTIR Spectra of the Copper(r1)-Amino Acid Complex in a Solid Matrix for Amino Acid-Copper(r1)-Clay and Oxide Systems Similarly, owing to the low copper(r1)-amino acid complex concentration and the spectral interferences from the solid particle matrix, IR spectra of the copper(I1)-amino acid Table 5 Infrared bands of the copper(~~)-amino acid complex in a clay matrix for coppcr(i1)-amino acid-clay systems, and the frequency shifts (cm-1) with respect to free amino acid Illite-histidine Montmorillonite-histidine Montmorillonite-glycine PH IEA of CuIVprnol m-? Amino acid<oppcr( ii)/ mol mol-I V(NH3 +) Av(NH~+) v(CO0-,as) Av(C00-, as) W H 3 + ) Ab(NH3+) v(CO0-, s) Av(C00-, s) v(Cu-0, as) v(Cu-0, s) 4.16 0.20 3.58 5.56 2.31 0 4.91 1.07 7.04 2.40 9.00 6.67 3.57 7.14 3.06 5.39 1.20 3156 26 1599 21 1504 - 39 1391 -18 356 287 11.8 3187 19 1607 15 1483 -39 1388 -25 360 3 14 1 .oo 3148 18 1610 32 1501 - 38 1392 - 17 355 295 1.20 3146 16 1603 25 I504 - 35 1393 - 16 355 295 2.54 3156 27 1582 4 1501 -38 1409 0 - 2.54 3152 23 1606 28 1504 - 34 1395 - 14 356 287 12.6 3148 18 1610 32 1497 - 42 1399 - 10 356 287 0.06 3148 18 1610 29 1497 - 42 1399 - 10 356 287170 ANALYST, FEBRUARY 1992, VOL.117 Table 6 Stepwise change in v(0H) of surface hydroxyl groups due to the interaction among Cu", amino acid and marine solid particles Amino acid- Solid Amino IEA of CuV CU"/ Y( OH) */cm- 1 particle acid PH pmol m-2 mol mol-1 Mite Histidine 7.00 15.0 0.06 3441 w 3523 w - Glycine 6.29 16.3 0.06 3441 m 3523 m - Aspartic acid 6.63 17.4 0.06 3441 m 3523 m Montmorillonite Glycine 8.07 GI ycine 7.24 Aspartic acid 8.19 Aspartic acid 8.02 y-Al00H Aspartic acid 2.57 Histidine 4.25 Aspartic acid 3.66 * br = Broad, s = strong, m = medium, and w = weak.1.57 12.6 3427 s , br - - 6.72 0.06 3427 w 3523 w - 1.28 12.6 3427 s , br - - 6.59 0.06 3427 w 3523 w - 3.33 0.11 3312 m 3427 m - 4.d9 0.12 3312 w 3427 w - 6.79 0.06 3312 m 3427 m 3512 w surface complex in a solid matrix were obtained by use of the spectral subtraction approach. The resulting spectra are given in Fig. 5 for the amino acid-illite and montmorillonite- copper(i1) systems with different IEAs of Cull, in which bands of v(NH3+), 6(NH3+), v(C00-, as) and v(C00-, s) were assigned and chosen to characterize the interaction among AAs, CuI1 and marine solid particles. As shown in Table 5, v(NH3+) increases 16-27 cm-1 and 6(NH3+) decreases 34-42 cm-1, while v(C00-, as) increases 15-32 cm-1 and v(C00-, s) decreases 10-25 cm-1. However, v(C00-, as) and v(C00-, s) vary less than 4 cm-1 as the IEA of Cull approaches zero, similar to adsorbed amino acids.It can be inferred from Tables 4 and 5 that in the interaction among AAs, clay and Cull, the amino group of an AA in a solid matrix has the same surface species [eqn. (l)] as the adsorbed AA, which associate with the surface hydroxyl groups on clay, because the variation of v(NH3+) and 6(NH3+) of the copper(i1)-amino acid surface complex is sufficiently large to be detected with respect to the resolution of the IR spec- trometer. However, unlike the adsorbed AA, the carboxyl group coordinates with Cut' because v(C00-, as) increases 15-32 cm-1 and v(C00-, s) decreases 10-25 cm-1, which is greater than the resolution of the spectrometer. If the above mechanism is correct, the Cu-0 stretching mode should be detected over the range 250-400 cm-1 for the copper(n)- amino acid surface complex.3~31~32 Two new bands at 315 and 295 cm-1 (Fig.5) were recognized for the copper(1i)-histidine -illite system by comparison with free histidine; the former may be assigned to ~(CU-0, as) and the latter to ~(CU-0, s) by analogous analysis.3.31J2 Similarly, ~(CU-0, as) and ~(CU-0, s) are 356 and 287 cm-1, respectively, for the histidine- montmorillonite-copper(1i) system, and 360 and 314 cm-1, respectively, for the glycine-montmorillonite-copper(i1) system. These are all consistent with the values reported previously.8.11.12 The change in v(0H) of SHHG due to the interaction among amino acids, Cu" and solid particles is shown in Fig. 6. At an IEA of Cull of 15.0, 16.3 and 17.4 pmol m-2, two bands at 3441 cm-1, attributed to the v(0H) of SHHG on pure illite, and at 3523 cm-1, were observed for the copper(1i)-illite-histidine system.Similarly, at an IEA of CulI of 1.57 and 1.28 pmol m-2, only one band at 3427 cm-1, due to the v(0H) of SHHG on pure montmorillo- nite, was detected,25 while at an IEA of Cu" of 6.59 and 6.72 pmol m-2 two bands at 3427 and 3523 cm-1 were detected for the glycine-copper(I1)-montmorillonite and aspartic acid- copper(I1)-montmorillonite systems. When the IEA of Cu" was 3.33 and 4.09 pmol m-2 two bands at 3312 cm-1, attributed to the v(0H) of SHHG on pure y-AlOOH,*5 and 3427 cm-1 were observed, while when the IEA of Cull increased up to 6.79 pmol m-2 three bands at 3312,3427 and 3512 cm-1 were detected for the amino acid-copper(n)-y- AlOOH system.Consequently, a small change in v(0H) of SHHG was observed for copper(I1)-amino acid-illite, mont- morillonite and copper(n)-amino acid-y- AlOOH systems, analogous to copper(1i)-solid particle systems. The variations of v(NH3+), 6(NH3+), v(C00-, as) and v(C00-, s) and the change in v(0H) of SHHG have led to the conclusion that as a 'bridging' reagent the amino acid joins Cull and the surface hydroxyl group, respectively, through the carboxyl and amino groups: namely, the Model I1 of TSC proposed by Leckie and co-workers16,18 results from the interaction between AAs, Cu" and clays. Conclusions Fourier transform IR spectroscopic studies were carried out in order to characterize the interaction among AAs, Cull and marine solid particles.The variations of hydroxyl bands show that the ion exchange between Cu" and solid particles and the interaction between AAs, solid particles and Cu" causes a stepwise change in v(0H) of SHHG on solid particles with an increase of the IEA of Cu". However, no change in v(0H) of either FSHG on illite, montmorillonite and CaC03 or FHG in goethite and y-A100H was observed. This provided the first evidence for the interfacial stepwise ion-exchange theory. The spectral subtraction approach was used to isolate the main IR bands of adsorbed AAs and copper(1i)-amino acid complexes in a solid matrix from the spectra of amino acid-solid particle and amino acid-copper(Ii)-solid particle systems, respectively. For adsorbed AAs, v(NH3+) increases 10-39 cm-1 and 6(NH3+) decreases 13-49 cm-1 with respect to free AAs, moderately more than the resolution of the IR spectrometer, whereas variations of v(C00-, s) and v(C00-, as) are less than 4 cm-1, nearly equal to the resolution, indicating that surface hydroxyl groups associate with the amino group of AAs in a solid matrix rather than the carboxyl group.For the copper(i1)-amino acid surface com- plex, v(NH3+) increases 10-27 cm-1 and 6(NH3+) decreases 34-42 cm-1, analogous to the adsorbed AA, while v(C00-, as) increases 15-32 cm-1 and v ( C 0 0 - , s) decreases 10-25 cm-1, which are moderately more than the resolution of the IR spectrometer, and bands at 287 to 360 cm-1, due to ~(CU-0, s) and ~(CU-0, as), appear. This reveals that as a 'bridging' reagent AAs in a clay matrix join the Cu" and surface hydroxyl group, respectively, through its carboxyl and amino groups.The author thanks Professor Zhang Zhengbin (Department of Chemistry, Ocean University of Qingdao, China) for helpful suggestions and encouragement, and Associate Professor XiaANALYST, FEBRUARY 1992, VOL. 117 171 Zongfong (of the same address) for assistance with the spectroscopic work. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 References Huang, J . , and Zhang, S . , Haiyang Xeubao, 1983,5, 604. Liu, B., Haiyang Huzhao, 1981, 2, 20. Zhang, Z., Estuarine-Marine Chemistry of Huanghe Estuary, Springer-Verlag, Berlin, 1991, ch. 2. Daland, F., Buffle, J., and Haerdl, W., Environ. Sci. Technol., 1984, 18, 135. Wang, X., Zhang, Z., and Liu, L., Chin.J. Oceanol. Limnol., 1988, 6 , 258. Zhang, Z., and Liu, L., Theory of Interfacial Stepwise Ion1 Coordination Particle Exchange and Its Application, Ocean Press, Beijing, 1985, pt. I. Zhang, Z., and Liu, L., Haiyang Yu Huzhao, 1978, 9,51. Riley, J . P., and Skirrow, G., Chemical Oceanography, Academic Press, London, 1972, vol. 1, ch. 3. Zhang, Z., Wang, X., Liu, L., and Liu, X., Haiyang Yu Huzhao, 1989,20,34. Buckland, A. D., Rochester, C. H., and Topham, S. A., J. Chem. SOC., Faraday Trans. I , 1980, 76, 302. Ishikawa. T., Nitta, S., and Konda, S., J. Chem. SOC., Faraday Trans. , 1986.82. 2401. Lewis, D. G., and Farmer, V. C., Clay Miner., 1986,21, 93. Marx, U., Sokoll, R., and Hobert, H., J. Chem. SOC., Faraday Trans. I , 1986, 82,2505. Rochester, C. H., and Topham, S. A., J. Chem. SOC., Faraday Trans. I , 1979, 75, 1259. Bourg, A. C. M., and Schindler, P. W., Chimica, 1978,32,166. Davis, J . A., and Leckie, J. O., Environ. Sci. Technol., 1978, 12, 1309. Pleysier, J., and Cremers, A., J. Chem. SOC., Faraday Trans. I, 1975, 71,256. 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Benjamin, M. M., and Leckie, J . O., Environ. Sci. Technol., 1981,15, 1050. Bowers, A. R., and Huang, C. P., J. Colloid Interface Sci., 1985, 105, 197. Bowers, A. R., and Huang, C. P., J . Colloid Interface Sci., 1986,110, 575. Moreals, P. C., Broersm, C., and Badot, C., Clay Miner., 1979, 14, 307. Anderson, M. A., and Rubin, A. J.,Adsorption of Inorganicsat SolidlLiquid Interface, Ann Arbor Science Publishers, Ann Arbor, MI, 1981, pp. 183-217. Cambier, P., Clay Miner., 1966,21, 191. Inskeep, W. P., and Baham, J., Soil Sci. SOC. Am. J., 1983,47, 1109. Wang, X., Ph.D. Thesis, Ocean University of Qingdao, China, 1989. Bellamy, L. J., The Infrared Spectra of Complex Molecules, Chapman and Hall, London, 2nd edn., 1980, ch. 4. Heilinger, A. W., and Long, T. V., J. Am. Chem. Soc., 1970, 92, 6474. Larsson, L., Acta Chem. Scand., 1950, 4, 27. Parker, P. S., and Kirschenbaum, D. M., Spectrochim. Acta, 1960, 16,910. Sokoll, R., and Hobert, H., J. Chem. SOC., Faraday Trans. I , 1986,82, 1527. Misra, B. H., and Kripal, R. K., Indian J. Pure Appl. Phys., 1987, 22, 430. Condrate, R. A., and Nakamoto, K., J. Chem. Phys., 1964,20, 2590. Paper 1 I01 301 G Received March 18, 1991 Accepted September 17, 1991

ISSN:0003-2654    

DOI:10.1039/AN9921700165

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1992, VOL. 117 173 Rapid Indirect Method for Determining the Sodium Content of Table Olives Pedro Garcia, Manuel Brenes and Antonio Garrido" lnstituto de la Grasa y sus Derivados (CSIC), Apartado 1078, 41012 Seville, Spain Methods for the determination of sodium in olives were studied. Direct measurement by photometric and ion-selective electrode methods did not show systematic errors, with the latter being more precise and easier t o apply. The correlation ( p s 0.001) between NaCl concentration in the surrounding brine (calculated using the official Volhard procedure) and Na+ content in the olive flesh (determined using an ion-selective electrode, the most practical of the two methods tested) was of particular interest and permitted the determination of the latter from determinations of the former.This indirect measurement enables producers t o determine the sodium content in packed olives non-destructively using an established method. It also avoids the problems of obtaining representative samples and the tedium of sample preparation. Keywords: Sodium determination; ion-selective electrode; flame photometer; table olives The sodium content in foods is receiving increasing attention because a high sodium intake has been linked with hyperten- sion in certain sensitive individuals. 1.2 This trend will continue in the future as consumers are becoming more conscious of the close relationship between health and diet. There is concern in some food industries which use salt or fermentation in brine as traditional conservation procedures, where the sodium con- centrations in the final products are generally still too high.Green table olives, Sevillian or Spanish style, are one of these items and the equilibrium percentages of salt in the brines offered to consumers range from 4.0 to 6.0% m/v. In flesh, the corresponding proportion is about 1.2-2.0 g of Na+ per 100 g of flesh.3.J Other types of olive also contain high levels of salt, e.g., natural black olives in brine (about 8% m/v in brine) and natural olives in solid salt (about 1&14% mh), but their volume in the international market is low. In contrast, the brine of ripe olives (darkened by alkaline oxidation) contains only a low amount of salt (about 2% m/v), which means that the proportion in the flesh could reach 0.4--0.8Y0 m/v, a considerably lower level.5 Hence although all table olive products must be carefully monitored for sodium content, green table olives in brine are of most concern.The Spanish table olive industry is trying to adapt its production methods to this situation and to improve the image of the different packed commercial presentations (plain olives; pitted olives; olives stuffed with pimento, anchovies, almonds, etc.) by the introduction of new packaging technol- ogy, a stricter inspection of sodium chloride content and the development of new products with lower sodium contents. Hence simple, rapid, inexpensive and accurate analytical procedures are needed for routine measurements of salt concentrations during the successive phases of production, and control of the sodium level in the packaged olives.Two procedures, based on the flame photometric determi- nation of sodium and on the determination of sodium o r chloride using ion-selective electrodes, have previously been applied successfully to the evaluation of the salt content in table olive brines. Comparison with the Volhard method, the official method prescribed by the International Olive Oil Council, showed that neither of them had constant o r proportional systematic errors. Both had the same precision as the reference procedure.6 However, when the salt concentra- tion in green table olive brines was determined either by flame photometry o r with a sodium ion-selective electrode, a * To whom correspondence should be addressed positive displacement with respect to the Volhard value was observed, owing to an excess of Na+ ions with respect to CI-.This additional sodium comes from the preliminary treatment with sodium hydroxide solution7.8 to eliminate oleuropein, a polyphenol responsible for the bitter taste of olives, and which also has a certain bactericidal effect that sometimes interferes with the normal lactic fermentative process. Obviously, the deviations observed are caused by the determination of the different ions which form salt, i.e., Na+ with photometric and sodium ion-selective methods and CI- with the chloride ion-selective and Volhard procedures, which in these products are not present in the proportions corresponding to the formula NaCl. As the determination of salt in olive flesh (the part of the product actually eaten by the consumer) has received little attention until now, investigation of suitable methods would be of interest.Direct analysis by some of the procedures used for brines, including the recently introduced flame photo- metric and ion-selective electrode methods, is one option. However, another is the use of an indirect measurement that takes advantage of the possible correlation between the sodium content in the flesh and in the surrounding brine. This relationship is very useful as it would provide producers with a non-destructive method that avoids both the problems of obtaining representative samples and the tedium of sample preparation, and would be particularly applicable to packed olives. The aim of this work was to study and compare the behaviour of flame photometric and sodium ion-selective electrode methods for the determination of sodium in olive flesh.Emphasis is given to the investigation of the correlation between sodium content in the flesh of the final packed product, determined with a sodium ion-selective electrode (the most practical of the two tested procedures) and the salt concentration in the brine determined by the Volhard method, currently used routinely in the industry for fermenta- tion and packing control. Experimental Flame Photometry Apparatus A Meteor Model NAK-1 flame photometer with a scale reading from 0 to 200 mequiv dm-3 (PACISA, Madrid, Spain), an LIC Instruments Model 346 automatic diluter (PACISA), an electric heater plate and an electric oven were used.174 ANALYST, FEBRUARY 1992, VOL.117 Reagents A standard solution containing 100 mequiv dm-3 of sodium (Meteor, Cat. No. 91834; PACISA), a 25% m/v solution of magnesium nitrate [Mg(N03)2.6H20] in 96% v/v ethanol and 6 mol dm-3 hydrochloric acid were used. Determination Calibration was achieved by verifying the electrode slope with standard solutions. Concentrations in samples were calculated on the basis of a reading of 1000 ppm corresponding, according to the mass of sample and dilutions mentioned, to 31.45 mg of Na+ per gram of flesh. For correct working, the electrodes must be rinsed with the electrode rinse solution after each reading and recalibration should be carried out every 2 h. The Volhard method was applied according to Fernandez Diez et al. 10 Sample preparation This operation was similar to that described previously for the determination of iron in olives.9 A 100 g amount of size- calibrated pitted olives from homogeneous fruits commer- cially packed (250 g glass containers) was mixed with 100 ml of distilled water and homogenized with a blender.A 4 g amount of the mixture (equivalent to 2 g of fresh) was placed in a quartz capsule containing 0.2 ml of the magnesium nitrate solution. The capsule was heated to 350 "C on an electric heater plate. After 2 h at 350 "C, the capsule was placed in an electric oven and the temperature was first raised rapidly to 250 "C and then slowly to 550 "C, at which it was maintained for 8-10 h. The greyish white ash was dissolved in three portions of 2 ml of 6 mequiv dm-3 HCI, filtering each time through filter-paper of known ash content, using a suction bell.The solubilization of the ash was improved by gentle heating of the capsule after each addition. Finally, the three portions were combined and the volume was made up to 50 ml by addition of distilled, de-ionized water. Determination Samples (0.3 ml) were diluted automatically to 10 ml with distilled, de-ionized water and the solution was introduced into the flame photometer. The apparatus was calibrated to express Na+ concentra- tion in milligrams per gram of flesh by adjusting the 100 mequiv dm-3 solution reading to 57. If the sodium content in flesh was less than 10 mg g-1, the standard sodium solution used was 10 mequiv dm-3. The reading was also adjusted to 57 and the sodium concentration was deduced by dividing the scale readings by 10.Ion-selective Electrode Apparatus An Orion Model 501 specific ion meter with a sodium ion-selective electrode (ref. No. 97-1 1) and a double-junction reference electrode (ref. No. 90-02), a magnetic stirrer, a 0.001 g precision balance and a mixer were used. Reagents In order to prepare an ionic strength adjuster (ISA), ana- lytical-reagent grade NH&I (20 g) was dissolved in about 50 ml of distilled water in a 100 ml calibrated flask, 27 ml of concentrated ammonia solution were added and the mixture was diluted to volume with distilled water. Standard solutions containing 100,200,500 and lo00 ppm of NaCl and with 2 ml of ISA per 100 ml were prepared. The reference electrode filling solutions were Orion ref.No. 90-00-02 (inner chamber) and 0.1 mol dm-3 NH4Cl (outer chamber). Electrode storage solution was prepared by adding 2 ml of ISA per 100 ml of 5 mol dm-3 NaCl solution. Electrode rinse solution was obtained by diluting 20 ml of ISA to volume in a 1 dm3 calibrated flask with distilled water. Sample preparation The flesh and distilled water were mixed as described previously, then 5 g of the paste were mixed with 195 ml of distilled, de-ionized water and 4 ml of ISA. The mixture was homogenized for 5 min in a magnetic stirrer and assayed immediately. Design of Experiments The proportional systematic errors of both procedures when used for the determination of sodium in olive flesh were investigated using standard additions experiments. If the confidence interval of the adjusted straight line includes unity it demonstrates the lack of such errors.11 Precision comparisons were made by triplicate analyses of the sodium content in the flesh of green Spanish-style olives and ripe olives (by alkaline oxidation).Standard deviations were calculated from these data and the precisions of the two methods were compared by calculation of the corresponding experimental F values and their probabilities. 11 Correlation between the sodium contents in flesh and brine was achieved by determination of Na+ in olives using the ion-selective electrode method, and determining its concen- tration by the determination of chloride in brine according to the Volhard procedure. In all instances, samples were analysed after 15 d to allow equilibrium of Na+ between the flesh and the brine to be reached.Results and Discussion As mentioned previously, neither procedure showed system- atic constant or proportional errors when applied to table olive brines.8 In order to investigate the presence of systematic proportional errors if the two assay methods are used to determine sodium in olive flesh, increasing amounts of Na+ were added to the olive flesh (standard additions experi- ments). Table 1 shows the average values of the concentration of sodium ( x ) added. The regression and confidence limits for the slope (b) showed they included 1.00 in both instances, thus confirming the absence of systematic proportional errors in the determi- nation of sodium in olive flesh by either the flame photometric or ion-selective electrode method.These results allow the use of either method with this new matrix, and also with the fermentation brine, as has already been demonstrated.8 Table 1 Sodium content in olive flesh with different amounts of added Na+ ( x ) by flame photometric and ion-selective electrode methods Na+ found*/mg g-I Sodium added Flame 0 11.66 3.93 16.00 7.86 20.00 15.72 26.33 23.59 35.33 mg per 100 g of flesh photometry Regression parameters 0, = a + bx) Confidence limits (p d 0.05) * Average of three replicates. b = 0.9816 1.0276 < b < 0.9356 Ion-selective electrode 10.23 14.16 18.13 25.96 33.86 b = 0.9977 1.0022 < b < 0.9931ANALYST, FEBRUARY 1992, VOL. 117 175 Table 2 Sodium contents (mg of Naf per gram of flesh) in the flesh of green and ripe olives in brine determined by flame photometric and ion- selective electrode methods. Comparison of the respective precisions Green olives Ripe olives Flame Ion-selective Flame Ion-selective photometry electrode photometry electrode X(1) s x 10-2 x(1) s x 10-2 x(l) s x 10-2 x(1) s x 10-2 3.70 10.0 3.21 0 2.60 10.0 2.72 11.0 2.39 0 2.83 5.7 3.28 10.4 4.16 5.7 3.75 3.8 8.53 5.7 8.58 7.7 3.36 5.3 7.93 5.7 8.02 4.1 2.65 9.2 8.30 10.0 8.24 2.6 2.64 0.6 7.33 5.7 7.24 1.7 2.50 0 7.90 10.0 7.94 3.4 2.86 4.0 7.56 11.5 7.55 5.3 3.29 3.2 4.18 7.6 Estimated standard errors ~1 = 7.86 X 10-2 ~2 = 5.14 X 10-2 ~3 = 8.48 X 10-2 ~4 = 4.59 X 10-2 Variance comparison One-sided critical values for the F-test (1) Results are the average of three replicates.F* = ~ 1 ~ / ~ 2 ~ = 2.34 F(16,16; 0.05) = 2.33 F(16,16; 0.01) = 3.37 P" = s32Is42 = 3.49 F(12,12; 0.05) = 2.69 F(12,12;0.01) = 4.16 0 1 2 3 4 5 6 7 8 Salt content in packing brine (% m/v NaCI) Fig.1 Relationship between sodium content in the packing brines (% m/v NaCl) and in the flesh of the corresponding olives (mg of Na+ per gram of flesh), and the confidence interval of the regression line for one future determination. Green table olives stuffed with pimento. Regression line equation: y = -0.13 + 3.17~ Table 2 shows the percentages of sodium in the flesh of green Spanish-style olives and ripe olives and their standard deviations, using both methods. Comparison by F test of the combined variance for each method shows a tendency that was significant at the p d 0.05 level for the ion-selective electrode, indicating that the results were more precise than the flame photometric results. This fact, together with the simpler procedure, requiring considerably less manipulation and labour, makes the ion-selective electrode method the method of choice.However, flame photometry could also give a sufficient precision for most requirements. In order to investigate the correlation between the sodium content in brine and in the flesh of the packaged final product, the concentrations of sodium in each medium were deter- mined by the Volhard and ion-selective electrode procedures, respectively. Different types of table olives with a wide range of salt concentrations were examined, although results are given only for pimento-stuffed green olives in 0.3 kg jars, and ripe olives in 0.5 kg cans.Variance analysis of the corresponding regression demon- strated a significant correlation (p d 0.001) between both procedures. The adjusted straight lines (Y = 0.998 and 0.988, respectively) and their confidence limits for one future ; o v) 1 .o 2.0 3.0 4.0 Salt content in packing brine (% m/v NaCI) Fig. 2 Relationship between sodium content in the packing brines (% m/v NaCI) and in the flesh of the corresponding olives (mg of Na+ per gram of flesh), and the confidence interval of the regression line for one future determination. Ripe olives. Regression line equation: y = 2.85 + 2.14~ analysis are shown in Figs. 1 and 2. According to the statistical inferences of the confidence intervals, the sodium content in flesh could be determined with a precision of 1 mg of Na+ per gram of flesh for stuffed green olives and 0.5 mg of Na+ per gram of flesh for ripe olives.Hence, in spite of the sodium and chloride imbalance in most table olive products, the existence of a good correlation between sodium content in the flesh and chloride in the brine (determined by the Volhard method) permits this to be a useful routine method for sodium monitoring during the production and inspection of the final product for the consumer. Nonetheless, the laboratory must prepare or be provided with adequate correlation graphs or equations for the main variables involved, viz., combined acidity, propor- tion of juice in olives and fruit-to-brine ratio. This work was supported by the Spanish Government through the Comision Interministerial de Ciencia y Tecno- logia (CICYT) under project Ali-88-0115-C02-01.References 1 2 The Food Labelling Regulations 1984, SI 1984 No. 1305, HM Stationery Office, London, 1984. Prevention of Coronary Heart Disease, WHO Technical Report Series, No. 678, World Health Organization, Geneva, 1982.176 ANALYST, FEBRUARY 1992, VOL. 117 3 Castro Ramos, R., Nosti Vega, M., and Vazquez Ladron, R., Alimentaria, 1980, 115, 21. 4 Nosti Vega, M., Castro Ramos, R., and Vazquez Ladron, R., Grasas Aceites, 1982, 33, 5. 5 Nosti Vega, M., and Castro Ramos, R., Grasas Aceites, 1985, 36, 203. 6 Official, Standardized and Recommended Methods of Analysis, Society for Analytical Chemistry, London, 1973, pp. 350-351. 7 Garrido Fernandez, A., and Garcia Garcia, P., Grasas Aceites, 1985, 36, 134. 8 Garcia Garcia, P., Brenes Balbuena, M., and Garrido Fernan- dez, A., Grasas Aceites, in the press. 9 Garrido Fernandez, A., Albi Romero, M. A., and Fernandez Diez, M. J . , Grasas Aceites, 1973, 24, 287. 10 Fernandez Diez, M. J . , Castro Ramos, R., Garrido Fernandez, A., Gonzalez Cancho, F., Gonzalez Pellisso. F., Nosti Vega, M., Heredia Moreno, A., Minguez Mosquera, M. I . , Sanchez Roldan, F., and Castro Gomez-Millan, A . , Biotecnologia de la Aceituna de Mesa, CSIC, Instituto de la Grasa, Madrid, 1985, p. 414. Massart, D. L., Dijkstra, A., and Kaufman, L., Evaluation and Optimization of Laboratory Methods and Analytical Pro- cedures, Elsevier, Amsterdam, 1978, p. 39. 11 Paper 11024801 Received May 28, 1991 Accepted October 11, 1991

ISSN:0003-2654    

DOI:10.1039/AN9921700173

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1992, VOL. 117 177 Ion-selective Electrode With Fixed Quaternary Phosphonium Ion-sensing Species Marie-Josee Rocheleau" and William C. Purdyt Department of Chemistry, McGill University, Montreal, Quebec, Canada H3A 2K6 The response properties and selectivity of polymer membrane electrodes with fixed quaternary phosphonium ion-sensing groups for nitrate are described. A poly(trioctylvinylbenzylphosphonium nitrate) ( PTOVBPN03) membrane electrode demonstrated a selective response t o nitrate in the concentration range 5 x lo-5-0.1 mol dm-3, with a slope of -53.4 ? 0.5 mV decade-'. The selectivity of this electrode for nitrate ions can be favourably compared t o conventional nitrate ion-selective electrodes based on quaternary ammonium ion-exchange sites. Notably, a significant improvement of selectivity for nitrate by two orders of magnitude was obtained in the presence of perchlorate.The PTOVBPN03 membrane electrode was used for the determination of nitrate in samples of commercially available fertilizers. Keywords: Ion-selective electrode; ionic polymer membrane; fixed ion-sensing species; quaternary phosphonium groups; nitrate determination Nitrogen is an essential constituent of natural ecosystems. However, an undesirable level of nitrate in natural waters as a result of man's activities is a serious source of pollution. The extensive use of artificial fertilizers in agriculture has been implicated as a major cause of the increasing concentration of nitrate in natural waters and as an important factor in the growing problem of eutrophication of lakes.This high consumption of artificial fertilizers has stimulated the de- velopment of efficient, inexpensive analytical methods for the determination of nitrate. The use of an ion-selective electrode is the method of choice for measurements of nitrate. This is mainly attributable to the rapidity of the sample preparation and to the simplicity of the potentiometric measurements. Conventional nitrate ion- selective electrodes are based on quaternary ammonium salts entangled in a polymeric matrix.' In order to improve the lifetime and robustness of these polymeric ion-selective electrodes, membranes prepared by covalent attachment of the ion-sensing groups to a polymer matrix have been reported.2.3 These polymers with fixed ion-sensing groups are commonly referred to as ionic polymers.4 The covalent attachment of the ion-sensing species prevents the deteriora- tion of the membrane through the leaching of the ion-sensing groups.Furthermore, membranes with covalently bound ion-sensing species display interesting features such as an enhanced adherence to solid substrates.3 This feature is particularly attractive for the future integration of ionic polymer membranes with semiconductor devices. So far, problems of compatibility between polymeric ion-sensitive membranes and the electronic component have seriously limited the expansion and application of ion-sensitive field effect transistors.5 Unfortunately, membranes prepared from covalent attach- ment of quaternary ammonium groups to a polymeric matrix have demonstrated rather poor selectivity for nitrate ions.Several inorganic anions such as hydroxide and chloride seriously interfere with the measurement of nitrate.* This paper reports the application of quaternary phosphonium functionalized polymer membrane electrodes to the measure- ment of nitrate. The response properties and selectivity of poly( trioctylvinylbenzylphosphonium nitrate) ( PTOVBPN03), and poly( triphenylvinylbenzylphosphonium nitrate) (PTPVBPN03), were investigated. Experimental Preparation of Polymeric Membranes Polymeric membranes were prepared by direct functionaliza- tion of poly(vinylbenzy1 chloride) (PVBC) with quaternary phosphonium groups. Poly(vinylbenzy1 chloride) and tri- phenylphosphine were obtained from Aldrich (Milwaukee, WI, USA), while trioctylphosphine was obtained from Alfa (Ward Hill, MA, USA).The choice of phosphines was guided by the fact that phosphines with short hydrocarbon chains are pyrophoric. Triphenylphosphine and trioctylphosphine are both stable in air. In the first reaction scheme, 1.5 g of PVBC were dissolved in 30 ml of a 25% v/v solution of trioctylphosphine in chloroform. The solution was then heated to reflux, to about 70-80 "C. In the second reaction scheme, 1.5 g of PVBC were suspended in 30 ml of a 0.4 mol dm-3 solution of triphenyl- phosphine in methanol. This solution was also heated to reflux. In this instance, the end-point of the reaction was indicated by the dissolution of PVBC, as poly- (triphenylvinylbenzylphosphonium chloride) (PTPVBPCI) is soluble in methanol while PVBC is not.Phosphines are characterized by their high nucleophilic reactivity with alkyl halides to produce phosphonium salts. A reflux time of 5-6 h was usually required for completion of the reaction. A quantitative elemental analysis performed on the modified polymers indicated a yield of immobilization of the quaternary phosphonium groups of >8O% (mole : mole quaternary phosphonium : benzyl chloride units) in both instances. These analyses were performed by Guelph Chemical Laboratories (Guelph, Ontario, Canada). After quaternization of PVBC with the phosphine, the modified polymers were precipitated and filtered off. The products obtained were redissolved and purified by two successive precipitation steps. Functionalized PVBC membranes about 100 pm thick were cast on a carbon-support electrode.The construction of the carbon-support electrode has been described previously.h Membranes can be cast easily at room temperature from a solution of the polymer dissolved in a volatile organic solvent. Prior to coating the carbon-support electrode with the polymeric membrane, the counter ions of the bound qua- ternary phosphonium groups were exchanged for nitrate through a liquid-liquid extraction procedure. Chloroform was used to cast PTOVBPN03 membranes, while methanol was used to cast membranes of PTPVBPN03. * Present address: Department of Chemistry, University of -F To whom correspondence should be addressed. Alberta, Edmonton, Alberta. Canada T6G 2G2. Calibration The potentiometric measurements were made with a Fisher Accumet Model 805MP pH/ion meter (Fisher Scientific,178 ANALYST, FEBRUARY 1992, VOL.117 Montreal, Canada). A Servogor 120 recorder [BBC Goerz Metrawatt (Fisher Scientific)] was used to monitor potential drifts. All potential measurements were made with reference to a saturated calomel electrode (SCE). The temperature of the analyte solutions was maintained at 25 "C with a Heto Type 623 thermostatic bath (Heto Lab. Equipment, Birkerod, Denmark). The membrane electrodes were stored dry and were pre-conditioned in 0.1 rnol dm-3 potassium nitrate for 30 min prior to reuse. Results and Discussion The response of the PTOVBPN03 membrane electrode is illustrated in Fig 1. The linear response of the PTOVBPN03 membrane electrode extended from 5 x 10-5 to 0.1 rnol dm-3 NO3- with a slope of -53.4 k 0.5 mV decade-'.Note that each calibration point of Fig. 1 represents the average of three potential readings. The standard solutions were buffered with 0.1 rnol dm-3 phosphate at pH 7.0. Phosphate ions had no observable effect on the response of the PTOVBPN03 membrane electrode. The influence of pH on the potential of this electrode was also investigated. No variation of the electrode potential was measured in the pH range 4-10. Above pH 10, the presence of hydroxide ions interferes with nitrate measurements. This PTOVBPN03 membrane electrode demonstrated an enhanced sensitivity to nitrate ions compared with a mem- brane electrode prepared from poly(trihexylvinylbenzy1- ammonium nitrate), PTHVBAN03 (see Fig.1). Further- more, the detection limit of the PTOVBPN03 membrane electrode for nitrate is significantly improved. The construc- tion and the use of the PTHVBAN03 membrane electrode are both described in ref. 3. The PTOVBPN03 membrane electrode showed a fast response and rapid recovery; the response time is typically <30 s. This membrane electrode continued to function well for several months. Very little deterioration of the membrane response was observed after more than 10 months of use. Comparatively, a typical lifetime of 2-3 months has been reported for a coated-wire electrode based on Aliquat 336s.' The selectivity coefficients for the PTOVBPN03 and PTHVBAN03 membrane electrodes were measured by the fixed interference method; the concentration of interfering anions was fixed at 1 mmol dm-3, while the concentration of nitrate was varied from 1 x 10-5 to 0.1 mol dm-3.The selectivity coefficients for both electrodes are presented in 300 > 200 E . - m a CL .- w +I 100 't -5 -4 -3 -2 - 1 0 Log([NO3-1/mol dm-3) Fig. 1 Potentiometric response of the PTOVBPN03 and PTHVBAN03 membrane electrodes: A , PTOVBPN03 membrane (slope = -53.4 k 0.5 mV decade-', linear response range = 5 x 10-5-0.1 rnol dm-3) and B, PTHVBAN03 membrane; (slope = -43.2 k 0.6 mV decade-', linear response range = 1 x 10-3-0.1 rnol dm-3) Table 1. The PTOVBPN03 membrane electrode clearly exhibits an enhanced selectivity for nitrate ions compared with the PTHVBAN03 membrane in the presence of nitrite ions. The selectivity coefficients of both electrodes in the presence of other anions are comparable.The performance of this PTOVBPN03 membrane elec- trode and another nitrate ion-selective electrode with covalently bound sites reported by Ebdon et a1.2 were also compared. The latter was prepared from a poly(styrene-b- butadiene-b-styrene) matrix cross-linked with ally1 substituted quaternary ammonium salts (SBS-QAS membrane). This SBS-QAS membrane displayed near-Nernstian response, but poor selectivity for nitrate in the presence of common interfering ions. For example, chloride interferes seriously with the response of the SBS-QAS membrane electrode, kNO3-, cl- = 0.16, while the response of the PTOVBPN03 membrane is less seriously affected by the presence of chloride, kNO3-, cl- = 0.008. The response properties and the selectivity of the PTOVBPN03 membrane can also be favourably compared to conventional nitrate ion-selective electrodes based either on quaternary ammonium1 or phosphonium8.9 ion-sensing spe- cies.Perchlorate constitutes one of the most serious interfer- ents of commercially available nitrate ion-selective electrodes based on quaternary ammonium ion-sensing species, typically k ~ 0 3 - , ~ 1 0 ~ - = 1000 for a Corning No. 476134 ion-selective electrode (Corning, NY, USA).' A significant improvement of selectivity for nitrate in the presence of perchlorate was obtained with the PTOVBPN03 membrane electrode, kNO3-, clod- = 10. This represents an improvement by two orders of magnitude. While the behaviour of a poly(tri- hexylvinylbenzylammonium chloride) (PTHVBAC) mem- brane electrode can be related to the lyotropic interactions (Hofmeister series) of the analyte ion with the ion-sensing species and can be described by simple thermodynamics,3 it is not clearly understood why this PTOVBPN03 membrane electrode provides such a selective response for nitrate in the presence of perchlorate ions.Unfortunately, very little information is available concerning the selectivity for nitrate of liquid membrane electrodes using phosphonium ion-sens- ing species. Therefore, their selectivity for nitrate in the presence of interfering ions cannot be compared with the selectivity demonstrated by the PTOVBPN03 membrane electrode. Despite the affinity of quaternary phosphonium sites for nitrate ions, a PTPVBPN03 membrane electrode did not demonstrate any sensitivity to nitrate ions.Furthermore, this membrane electrode did not demonstrate any response to other anions such as thiocyanate and salicylate. Not only is selectivity involved in the membrane response mechanism but so is the mobility of the counter ion species within the membrane phase. In conventional polymeric ion-selective membranes, the strong association of an ion-exchange group and a counter ion species results in the formation of a neutral Table 1 Selectivity coefficients for nitrate ion-selective membrane electrodes with fixed ion-sensing sites k??' '31 Interfering anion Phosphate Acetate Sulfate Chloride Nitrite Bromide Iodide Perchlorate PTOVBPN03 o.00010 k 1 x 10-5 0.0081 k 1 x 10-4 0.00051 k 3 x 0.00320 k 1 x 10-5 0.052 f 0.001 0.252 k 0.003 3.17 t 0.07 10 k 1 PTHVBAN03 0.00051 k 2 x lW5 0.00079 f 1 x 0.0051 k 1 x 0.0089 f 1 x 0.251 k 0.001 0.316 k 0.001 0.79 If: 0.01 15.8 k 0.8ANALYST, FEBRUARY 1992, VOL.117 179 Table 2 Potentiometric determination of nitrate in commercial fertilizers N03-N (Yo) Sample Claimed Found* RaPidGro evergreen 2.9 2.7 t- 0.1 Jobe's sunsplash 0.4 0.44 * 0.02 * Average of three measurements k standard deviation. pair which is still mobile within the membrane phase because of the presence of a solvent/plasticizer.'O On the other hand, the more strongly an ion is preferred by the fixed ion-exchange sites in an ionic polymer membrane, the more poorly it moves within the membrane. There are, therefore, opposing effects between affinities and mobilities of ions in ionic polymer membranes, and the strong association of the sensing species with analyte ions constitutes the principal limitation of this membrane system.The lack of sensitivity demonstrated by the PTPVBPN03 membrane electrodes may also be related to the poor physical properties of the membrane. Membranes prepared from PTPVBPC were found to be porous, hard and brittle. Determination of Nitrate in Fertilizers The PTOVBPN03 membrane electrode was applied to the determination of nitrate-nitrogen in two samples of fertilizers. These fertilizers were obtained from local stores. The RaPid- Gro evergreen fertilizer (RAPIDGRO, Danville, NY , USA) is a granule concentrate, while the Jobe's sunsplash fertilizer (International Spike, Lexington, KT, USA) is a dilute aqueous preparation. Both contain all three of the main plant nutrients, i.e., nitrogen, phosphate and potassium.They also contain micronutrients such as boron, copper, iron, man- ganese and zinc. A 5 g sample of the RaPidGro evergreen fertilizer was finely ground and dried in an oven for 2 h at 100 "C. At this temperature, no degradation of the sample occurs. Three portions of about 0.1 g of the dry sample were then dissolved in 50 ml of 0.1 rnol dm-3 phosphate buffer, pH 7.0. Three aliquots of the Jobe's sunsplash fertilizer were used without dilution. An interference suppressor composed of 0.01 mol dm-3 aluminium sulfate, 0.01 rnol dm-3 silver sulfate, 0.02 rnol dm-3 boric acid and 0.01 rnol dm-3 sulfamic acid, and adjusted to pH 3.0 with 0.1 rnol dm-3 sulfuric acid was used.This solution effectively reduces the interfering effect of many common inorganic ions. Chloride ions are quantitatively precipitated as silver chloride. Trace amounts of bromide, cyanide, sulfide and phosphate are also removed by precipita- tion with Ag'. Further, aluminium ion strongly complexes anions of organic acids, and nitrite is quantitatively destroyed by reaction with sulfamic acid. Equal volumes (2 ml) of this interference suppressor and the sample solution were mixed. After precipitation of the interfering species, the solution was filtered and diluted to 50 ml with 0.1 rnol dm-3 phosphate buffer at pH 7.0. The use of this interference suppressor was found to be particularly useful for the analysis of the Jobe's sunsplash fertilizer.When the suppressor was not used, an unspecified component of the sample caused a positive interference with nitrate measurements. The nitrate-nitrogen content of three aliquots of each sample was quantified by use of the standard additions method, and the results are reported in Table 2. In both instances, these results are in good agreement with the amount of nitrate-nitrogen guaranteed by the manufacturer. The electrode potential obtained by repeated measurements on a 1 mmol dm-3 nitrate solution falls within 1 mV. This variability in the electrode response and sampling errors account for the standard deviation observed. Thus, the electrode described here proved to be a simple, reliable and sensitive means for the determination of nitrate. The authors are indebted to the Natural Sciences and Engineering Research Council of Canada for financial support of this work. 1 2 3 4 5 6 7 8 9 10 References Davies, J. E. W., Moody, G. J., andmomas, J. D. R., Analyst, 1972,97, 87. Ebdon, L., King, B. A., and Corfield, G. C., Anal. Proc., 1985, 22,354. Rocheleau, M. J., and Purdy, W. C., Electroanalysis, 1991, 3, 929. Ionic Polymers, ed. Holliday, L., Halsted Press-Wiley, New York, 1975, ch. 1. Janata, J., and Huber, R. J., in Zon-Selective Electrodes in Analytical Chemistry, ed. Freiser, H., Plenum Press, New York, 1980, ch. 3, vol. 2. Rocheleau, M. J., and Purdy, W. C., Talanta, 1990, 37, 307. James, H., Carmack, G., and Freiser, H., Anal. Chem., 1972, 44, 856. Skobets, E. M., Makovetskaya, L., and Makovetsii, Y., Zh. Anal. Khim., 1974,29,2354; Chem. Abstr., 1975,82,164394a. Hopirtean, E., Stefaniga, E., Liteanu, C., and Gusan, I., Rev. Chim., 1976, 27, 346; Chem. Abstr., 1976, 85, 1533622. Eisenman, G., in Zon-Selective Electrodes, ed. Durst, R. A., National Bureau of Standards, Special Publ. 314, US Govern- ment Printing Office, Washington, DC, 1969, ch. 1. Paper 1 I02031 E Received May 3, 1991 Accepted October 8, 1991

ISSN:0003-2654    

DOI:10.1039/AN9921700177

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1992, VOL. 117 181 Determination of Trace Amounts of Estriol and Estradiol by Adsorptive Cathodic Stripping Voltammetry Shengshui Hu, Qong He* and Zaofan Zhao Department of Chemistryf Wuhan University, Wuhan, China Estriol and estradiol are electroinactive in the potential range from -200 to -1000 mV versus a silver-silver chloride electrode at a mercury electrode. The conversion of these estrogens into electroactive nitro derivatives of estrogens, which are used for voltammetric determination, was studied. Such nitro derivatives give a well defined cathodic stripping wave at -600 mV in pH 10.5 borate buffer. Estriol and estradiol are determined in the ranges 1 x 10-9-1.5 x 10-6 and 5 x 10-9-2 x 10-6 rnol dm-3, respectively, by differential-pulse adsorptive stripping voltammetry at a hanging mercury drop electrode.Some steroids, such as estrone, interfere because the three estrogens have almost the same molecular structure and have similar nitro derivatives, but progesterone does not interfere and is reduced at significantly more negative potentials than the nitrated estrogens. It can be determined simultaneously with estriol or estradiol. A method was developed for the assay of estriol in pharmaceutical preparations. Keywords: Estriol; estradiol; adsorptive stripping voltammetry Estriol and estradiol are steroids of importance in biological processes and are substances of pharmaceutical importance. Many methods have been proposed for the determination of trace amounts of steroids, e.g., colorimetric,1-4 spectropho- tometric,s gas chromatographic610 and gas chromato- graphic-mass spectrometric" techniques.Electrochemical methods have previously been applied to the determination of steroids,12-16 but the detection limits are not sufficient for determinations at low concentration levels. Some electro- analytical procedures involve the formation of derivatives to obtain polarographically or voltammetrically usable peaks; for instance, Buecher and Franke17.18 converted 17-oxosteroids into the P-vinylhydrazone derivative by heating in 15% hydrochloric acid at 100 "C, whereas Starka and Brabencovalg used oxidation with periodic acid to determine acetaldehy- drogenic steroids in the presence of formaldehydrogenic steroids in urine. Estrogens such as estriol and estradiol cannot be reduced at a dropping mercury electrode because they do not contain the ketonic groups or a,P-unsaturated ketonic groups.However, some derivatives of estriol and estradiol enabled these compounds to be determined by polarography or stripping voltammetry. For example, Wolfe et aZ.20 reported that a water-soluble and reducible hydrazone is formed after estrone reacts with trimethylammonium acetohydrazide chloride. An immunoassay of estriol labelled with mercury( 11) acetate has been employed for the determination of estriol by electro- analysis.*' Wehmeyer et aZ.22 developed another immuno- assay system to determine estriol labelled with nitro groups by differential-pulse polarography . Recently, an experimental procedure based on a sensitive polarographic method has been employed for the determination of estrogens at the 4 x 10-8 rnol dm-3 level by linear-sweep polarography.23 These estrogens were made electroactive with sodium nitrite, and nitro derivatives of estrogen were obtained.24 These methods have been used to determine estriol in the urine of pregnant women.Adsorption is often considered a nuisance in an electro- analytical experiment, to be avoided, when possible, by changing the solvent, concentrations, etc. However, adsorp- tion of a species is sometimes a prerequisite for rapid electron transfer, and can be of major importance in many processes of practical interest ( e . g . , the oxidation of aliphatic hydrocar- bons or the reduction of proteins). Adsorptive stripping * Present address: Department of Chemistry, Qujing Normal School for Professional Training, Yunnan, China.voltammetry is a very sensitive electroanalytical method for the determination of some compounds that can be adsorbed at the electrode surface by adsorptive accumulation and then reduced. 6-Aminopenicillanic acid has been determined at the 1 x 10-9 rnol dm-3 level after accumulation for 4 min at a hanging mercury drop electrode.25 The adsorptive stripping voltammetric method was also used for measuring estriol and estradiol in this study. The experiments indicated that nitro derivatives of estriol or estradiol are adsorbed strongly on the mercury electrode and are reduced in the stripping step. By using this phenomenon, extremely sensitive and rapid adsorp- tive stripping procedures were achieved.Adsorptive deposi- tion periods of 15 min were employed for the determination of estriol and estradiol at the 8 x 10-10 and 2 x 10-9 rnol dm-3 levels, respectively. This method is also fairly simple because it was unnecessary to separate related compounds and the nitrated estrogen was determined directly in the reaction mixture. In this respect, the method differs from the determi- nation by adsorptive stripping voltammetry of other estrogens such as estrone.26 Nitrosation has also been employed for the determination of 1- and 2-naphthols27 and morphine.28 Experimental Adsorptive stripping voltammetry was carried out with a Model 174A polarographic analyser [Princeton Applied Research (PAR), Princeton, NJ, USA] with a Model 0089~-y recorder (PAR). The three-electrode system employed was a Model 303 static mercury drop electrode (PAR), a silver- silver chloride reference electrode and a platinum auxiliary electrode.A medium-sized drop of surface area 0.016 cm2 was used. pH measurements were made with a Model pHs-203 pH meter (Wuhan Electric and Technical Instrumental Factory, Wuhan, China). Estriol, estradiol and other biochemicals were obtained from Sigma (St. Louis, MO, USA). The estrogen was dissolved in 100 ml of absolute ethanol to give stock solutions of estriol or estradiol. All chemicals were of analytical-reagent grade or better. All solutions were prepared with doubly distilled water (from quartz). Borate buffer solutions (0.05 rnol dm-3) were prepared by dissolving sodium tetraborate in distilled water.A 2 rnol dm-3 solution of sodium nitrite was prepared by dissolving sodium nitrite in 1 x rnol dm-3 sulfuric acid. Pharmaceutical preparations of estriol were commercially available. The determination procedures were as follows: 0.2 ml of estriol (or estradiol) was placed in a 10 ml calibrated flask and182 ANALYST, FEBRUARY 1992, VOL. 117 2 ml of 2 mol dm-3 sodium nitrite solution were added. The mixture was heated at 100 "C on a boiling water-bath. After heating for 30 min, the solution was allowed to cool to room temperature, then 2 ml of 0.05 mol dm-3 borate buffer solution were added, the pH was adjusted to 10.5 by addition of hydrochloric acid or sodium hydroxide solution and the solution was diluted to 10.0 ml with distilled water. The resulting solution was transferred into an electrolytic cell, the stirrer was started and the solution was purged with nitrogen for 10 min.After forming a new hanging mercury drop, the accumulation potential (-0.10 V) was applied to the working electrode for a selected time while the solution was stirred. At the end of the accumulation period, the stirrer was stopped and 20 s were allowed for the solution to become quiescent. The differential-pulse cathodic stripping scan was then started, the peak height being measured at -0.60 V. Results and Discussion The differential-pulse cathodic stripping voltammograms of nitrated estriol and estradiol were investigated at various pH values. The pH-dependent profiles of both estrogens were similar. The maximum height was observed at pH 10.5, but little difference in height was observed at pH >10.8.The peak height obviously decreased between pH 8.5 and 10.0. A small double wave was observed when the pH of the solution was <6.0, and it was not suitable for determining very low concentrations. The use of other buffer solutions that cover this pH range was examined. The best results, with respect to peak enhancement and shape, were obtained using 0.01 mol dm-3 borate solution. The effect of the solution pH on the peak potential was also investigated. The results indicate that the peak potential is pH dependent and the peak shifts towards negative potentials with increase in pH. The accumulation potential applied to the electrode during the period of adsorption strongly affects the peak height obtained, as shown in Table 1.An accumulation potential between -0.05 and -0.1 V is optimum. At more negative potentials the peak height decreases rapidly as the reduction potential of nitrated estrogen is approached. However, at more positive potentials the peak height diminishes because a large amount of mercury(1) hydroxide was deposited on the electrode surface, and this deposited film may inhibit the adsorption of nitrated estrogens and the stripping response by blocking the electrode surface. The greatest peak height was obtained at -0.1 V. The peak height of nitro derivatives of estriol and estradiol was measured by adsorptive stripping voltammetry as a function of the estrogen concentration. The peak height increased with increasing concentration of estriol and estra- diol up to about 2 x 10-6 mol dm-3.At higher concentrations of both estrogens, curvature of the calibration graph occurred. Table 1 Effect of varying the accumulation potential on peak height of nitrated estrogens at pH 10.5 in 0.01 mol dm-3 borate buffer. Accumulation time, 30 s Peak height/nA PotentiaW +0.5 0 -0.05 -0.1 -0.2 -0.3 -0.4 -0.5 5 x lo-' rnol dm-3 5 x mol dm-3 estriol estradiol 290 270 300 290 320 300 320 300 280 270 240 230 150 130 100 90 The curvature presumably indicates that a limiting value of the amount of nitro derivative on the electrode surface has been achieved under the prescribed conditions. Further increases in concentration did not increase the amount of nitro derivatives at the electrode owing to surface saturation, hence the stripping peak height remained constant.For convenient measurement of concentrations ranging from 1.5 x 10-6 to 1 X 10-9 rnol dm-3 for estriol and from 2 x 10-6 to 5 x 10-9 mol dm-3 for estradiol, 0.5-10 min is usually sufficient. Fig. 1 shows the dependence of the adsorptive stripping peak height on the accumulation time for 1 x 10-7 mol dm-3 estriol and estradiol. These profiles represent the correspond- ing adsorptive isotherms (as the peak height depends on the amount adsorbed). With increasing accumulation time the adsorption of nitro derivatives on the mercury electrode is enhanced. This increased adsorption is non-linear as a function of accumulation time, as the electrode surface rapidly becomes saturated with nitrated estrogens. At adsorption times longer than 4 min, the peak height remains constant.For a 15 min accumulation, the detection limit is 8 X 10-10 and 2 X 10-9 rnol dm-3, respectively, based on a signal-to-noise ratio of 3. The sensitivity for the low concentration is improved by increasing the accumulation time, but the linear range is then diminished. Cyclic voltammograms obtained for a 5 X 10-7 mol dm-3 solution of estradiol at the hanging mercury drop electrode are shown in Fig. 2. Only a cathodic peak appears without the corresponding anodic peak. The measured parameter of these 0.3 0.2 f .> 0.1 1 I I I 0 200 400 600 Accumulation time/s Fig. 1 mol dm-3 estradiol (A) and estriol (B) Effect of accumulation time on the peak height for 1 X 0.4 0.3 5. k 0.2 0.1 -0.4 -0.6 -0.8 PotentialN Fig.2 Cyclic voltammograms of 5 x 10-7 mol dm-3 estradiol at pH 10.5 in 0.01 mol dm-3 borate buffer; scan rate = 50 mV s-l. A , First scan, accumulation time = 30 s; and B, second scan, without accumulationANALYST, FEBRUARY 1992, VOL. 117 183 Table 2 Results obtained for estriol in injection samples by the proposed method. Each determination was carried out in triplicate Estriol contendpg per ampoule Recovery experiments Sample Reference Proposed Estriol added/ Estriol foundl Recovery (% ) No. value method pg dm-3 pg dm-3 (mean k SD*) 1 500 465 15.0 14.0 93.3 k 6.5 2 500 443 18.0 15.9 88.3 k 9.6 3 500 448 18.0 16.1 89.4 k 4.7 * SD = standard deviation. 80 60 2 40 .> 20 0 50 100 150 200 Potential scan rate/mV s-1 Fig. 3 Effect of varying the otential scan rate on the peak height for 1 X mol dm-3 estriol (AT and estradiol (B); accumulation time = 4 0 s t ...I I I I -0.4 -0.8 -1.2 -1.6 PotentialN Fig. 4 Stripping voltammograms for A, 1 X 10-6 mol dm-3 estriol; and B, 5 X lo-’ mol dm-3 progesterone at pH 10.5 in 0.01 mol dm-3 borate buffer. Accumulation for 1 min at -0.1 V. The broken line represents the blank at 0.01 mol dm-3 borate buffer i-E curves is the ratio of the peak heights, ip,a/ip,c. Deviation of the ratio iP,& from unity is indicative of an irreversible electrode process. A large cathodic peak is observed after adsorption accumulation at -0.3 V for 30 s; the second scan without accumulation reveals only a small peak in the cathodic branch and the peak potential shifts towards the positive direction.This is a characteristic feature of adsorption of nitro derivatives. A similar cyclic voltammetric response was observed for estriol. The effect of varying the potential scan rate on the peak height for 1 x 10-8 mol dm-3 estriol and estradiol is shown in Fig. 3. The cathodic stripping peak height increases recti- linearly with scan rate, as expected for the reduction of an adsorbed species. These results are in agreement with electrochemical theory.29,30 Trace amounts of estrone can interfere if, under the conditions used, a nitrated estrone is formed with sodium nitrite which is adsorbed on the electrode and produces a reduction peak close to that of the nitro derivatives of estriol and estradiol. The nitrated estriol and estradiol have almost the same adsorptive stripping peak potential and cannot be t .- I I , - 0.4 -0.6 -0.8 -0.4 -0.6 -0.8 PotentialN Fig.5 Stripping voltammograms for (a) injection sample; and (b) injection sample plus 5 x 10-8 rnol dm-3 estriol. Accumulation time = 2 min. Other conditions as in Fig. 4 determined simultaneously without prior separation. No interference in the determination of 1 X 10-6 mol dm-3 estriol or estradiol was observed when the recommended procedures were applied after the addition of 1 x 10-7 mol dm-3 progesterone, because its reduction is at a more negative potential than the nitrated estrogens; the voltammograms are shown in Fig. 4. The presence of surfactants may interfere by competitive adsorption, which can diminish the surface area of the electrode available for adsorption of nitro derivatives.It has been shown that the peak height of the nitro derivative was decreased by adding a cationic surfactant (tetrabutylammo- nium bromide) or an anionic surfactant (sodium lauryl sulfate). However, no interference for 1 x 10-6 mol dm-3 estriol or estradiol was observed with natural organic surface- active materials, such as the non-ionic surfactant poly(viny1 alcohol) (PVA; 2 ppm). This phenomenon may be due to synergistic adsorption or a decrease in the adsorption of non-ionic PVA at this potential. Because of its high sensitivity, estriol was determined repeatedly in samples of pharmaceutical preparations by using the recommended method. A 0.1 ml estriol injection sample was placed in a 200 ml calibrated flask with 150 ml of ethanol and the flask was stoppered, shaken manually for 2 min and then the contents were diluted to 200.0 ml with distilled water. A 0.2 ml aliquot of the sample solution was transferred into a 10 ml calibrated flask and 2 ml of 2 mol dm-3 sodium nitrite solution were added.The mixture was heated on a boiling water-bath for 30 min and the solution allowed to cool to room temperature. Then, 2 ml of 0.05 rnol dm-3 borate buffer were added, the pH was adjusted to 10.5 and the solution was diluted to 10.0 ml with distilled water. Adsorptive cathodic stripping voltammetry was then applied as described above. The results for the determination of estriol in three injection samples are given in Table 2. The voltammograms from a typical analysis for estriol in an injection solution are shown in Fig.5 . This work was supported by the National Natural Science Foundation of China.184 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 References Forist, A. A., and Theal, S., J. Am. Pharm. Assoc. Sci. Ed., 1958,47, 520. Mader, W. J., and Buck, R. R., Anal. Chem., 1952, 24, 666. Clarke, I., Nature (London), 1955, 175, 123. Ansari, S., and Khan, R. A., J. Pharm. Pharmacol., 1960, 12, 122. Gross, J. M., Eisen, H., and Kedersha, R. G., Anal. Chem., 1952, 24, 1049. Domsky, I. I., and Perry, J. A., Recent Advances in Gas Chromatography, Marcel Dekker, New York, 1971. Sanghoi, A., and Wight, C., Clin. Chim. Acta, 1974, 56, 49. Gardiner, W. L., and Horning, E. C., Biochim. Biophys. Acta, 1966, 115, 524. Horning, M. G., Anal. Biochem., 1968,22, 284.Wotiz, H. H., Biochim. Biophys. Acta, 1963, 69,413. Stillwell, W. G., and Zlatkis, A., J. Steroid Biochem., 1972, 3, 699. Bond, A. M., Heritage, I. D., and Briggs, M. H., Anal. Chem., 1984, 56, 1222. Wang, J., Farias, A. M., and Mahmoud, S. J., Anal. Chim. Acta, 1985, 171, 195. Schaar, J. C., and Smith, D. E., Anal. Chem., 1982,54, 1589. Fogg, A. G., Fayad, N. M., and Burgess, C., Anal. Chim. Acta, 1979,110, 107. Hu, S.-S., Yan, Y.-Q., and Zhao, Z.-F., Anal. Chim. Acta, 1991, 248, 103. 17 18 19 20 21 22 23 24 25 26 27 28 29 30 ANALYST, FEBRUARY 1992, VOL. 117 Buecher, H., and Franke, R., Abh. Dtsch. Akad. Wiss. Berlin, Kl. Med. Wiss., 1965, 1, 93. Buecher, H., and Franke, R., Acta Biol. Med. Ger., 1965,14,1. Starka, L., and Brabencova, H., Clin. Chim. Acta, 1960,5,423. Wolfe, K . J., Hershberg, B. E., and Fieser, F. L., J. Biol. Chem., 1940, 136, 653. Heineman, W. R., Anderson, C. W., and Halsall, H. B., Science, 1979, 204, 865. Wehmeyer, K. R., Halsall, H. B., and Heineman, W. R., Clin. Chem. (Winston-Salem, N. C.), 1982, 28, 1968. Hu, S.-S., He, Q., and Zhao, Z.-F., Anal. Chim. Acta, in the press. Konyvey, I., and Olsson, A., Acta Chem. Scand., 1964,18,483. Hu, S.-S., and Zhao, Z.-F., Anal. Lett., 1991, 24, 827. Hu, S.-S., He. Q., and Zhao, Z.-F., Chem. J. Chin. Univ., in the press. Davidek, J., and Seifert, J., Sb. Vys. Sk. Chem.-Technol. Praze, E, 1971, 30, 7. Noninska, K. I., Dryanovska, L., and Iliev, L. S., Farmatsiya (Sofia), 1969, 19,24. Bard, A. G., and Faulkner, L. R., Electrochemical Methods, Wiley, New York, 1980. Laviron, E., J. Electroanal. Chem., 1974,52, 355. Paper 1102388H Received May 22, 199I Accepted September 16, 1991

ISSN:0003-2654    

DOI:10.1039/AN9921700181

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1992, VOL. 117 185 Differential-pulse Polarographic Microdetermination of Reactive Organohalides via In Situ Generation of S-Alkylisothiouronium Salts Wing Hong Chan* and Albert Wai Ming Lee* Department of Chemistry, Hong Kong Baptist College, 224 Waterloo Road, Ko wloon, Hong Kong Pei Xiang Cai De pa rtm en t o f Chemistry, Zh o ngs ha n U n ive rsit y, Gua ngzh ou, People's Republic o f Ch ina Reactive organohalides, after in situ derivatization to the corresponding S-alkylisothiouronium salts in the presence of an excess of thiourea, were determined in aqueous solution by differential-pulse polarography. A single differential-pulse polarographic peak was obtained at pH 12.6 for all the organohalides under investigation. The calibration graphs were rectilinear over the range from 1 x 10-6 to 1 x 10-4 mol dm-3 in the sample solution. Each of the S-alkylisothiouronium derivatives has a characteristic polarographic peak potential, thus allowing the qualitative and quantitative determination of different organohalides.For the microdetermination of reactive organohalides, the method was found to be precise and the detection limit was 1 pg of halide. Keywords: Reactive organohalides; differential-pulse polarography; organic analysis; in situ S-alk yliso th iou ron ium sa It genera ti0 n The halide group is one of the most fundamental organic functional groups. Many organohalides are important inter- mediates in organic reactions and are used extensively in organic synthesis. Other significant uses of this class of compound are as anaesthetics, refrigerants, and grain and fruit fumigants.1 Although the determination of organohal- ides has been the focus of many investigations, simple and sensitive methods are still in great demand.24 Many organic substances that are polarographically inactive are amenable to indirect determination by production of an electroactive derivative.5 However, few organic reactions give quantitative derivatization; this impedes the application of such a strategy to the polarographic determination of organic species. We previously developed a poly(viny1 chloride) membrane S-alkyl- isothiouronium-selective electrode for the determination of alkyl halides via the generation of the corresponding S-alkyl- isothiouronium salts.6 Smyth and Osteryoung7 reported that when subjected to differential-pulse polarographic studies, S-benzylisothiouronium produces a characteristic wave.As the direct polarographic determination of organohalides is sometimes not feasible , this paper describes the development of a sensitive indirect polarographic method for the determi- nation of organohalides via in situ generation of their polarographically active derivatives, viz . , Polarographically inactive Polarographically active Under the defined conditions, all reactive organohalides can be quantitatively converted into water-soluble and electro- active S-alkylisothiouronium salts, thus allowing their analy- tical determination. Experimental Apparatus Differential-pulse polarographic measurements were made by means of a Metrohm E-506 polarograph coupled with an E-505 polarographic stand.A three-electrode combination was used, consisting of a saturated calomel electrode (SCE) as the reference and platinum as the counter electrode. A Metrohm Model EA-87620 cell equipped with high-purity * Authors to whom correspondence should be addressed. nitrogen was used throughout. A pulse amplitude of 40 mV was used with a scan rate of 1.24 mV s-1, a force drop time of 2 s, and at a mercury head height of 60 cm. Reagents All chemicals were of analytical-reagent grade. S-Benzyliso- thiouronium chloride was prepared according to the pro- cedure described in the literature8 and was recrystallized once from 95% ethanol prior to use. Buffer solutions (pH 11-13), which contained sodium hydroxide, disodium hydrogen phosphate and potassium chloride, as defined by the National Institute of Standards and Technology (formerly the National Bureau of Standards) ,9 were all prepared using distilled water and were used as the background electrolyte.In Sifu Derivatization of Reactive Alkyl Halides For investigations with large amounts of sample, about 0.5 g of alkyl halides was accurately weighed in a 100 ml round- bottomed flask and 1.3 equiv of thiourea were added. The mixture was dissolved in 25 ml of 95% ethanol and the solution was refluxed for 2 h. After refluxing, the solvent was removed under reduced pressure. The residue was dissolved in distilled water, then made up to 100 ml in a calibrated flask. For the polarographic measurement, 1 ml of the solution was further diluted to the mark with distilled water in a 100 ml calibrated flask.benzyl chloride and thiourea were prepared separately by dissolving an appropriate amount of the pure substance in 95% ethanol in a calibrated flask. An appropriate amount of each standard solution, such that thiourea is 2-3-fold in excess of the halide, was transferred by pipette and mixed in a 10 ml round-bottomed flask. The mixture was then refluxed for 3 h to effect the formation of the S-alkylisothiouronium salt. The residue was redissolved in distilled water, transferred into a calibrated flask and diluted to the mark with distilled water. Further dilution may be required so that the concentration of the final solution is preferably in the range from 1 x 10-6 to 1 x 10-4 mol dm-3.For the microdetermination study, standard solutions of ' Determination of Organohalides by Differential-pulse Polarography A 20 ml volume of supporting electrolyte at pH 12.6 was placed in the cell of the polarograph and flushed with186 ANALYST, FEBRUARY 1992, VOL. 117 oxygen-free nitrogen for 5 min. Under a nitrogen atmosphere, a known amount of sample or standard solution was intro- duced. The mixture was flushed by bubbling nitrogen through the solution for a further 15 s. The cell was attached to the three-electrode assembly and a nitrogen flow maintained over the solution. A potential scan was then performed over the range from -0.40 to -0.80 V versus SCE at a rate of 1.24 mV s-1 in order to obtain a differential-pulse polarogram. Calibration Graphs for the Microdetermination of Benzyl Chloride A series of samples containing 0.001-0.01 mg of benzyl chloride in 1 ml of ethanol was mixed with 2 mg of thiourea.The mixture was refluxed for 3 h. After cooling, supporting electrolyte was introduced and the solution was quantitatively transferred into a 25 ml calibrated flask. A 20 ml aliquot of each of the solutions was then taken for polarographic measurement. Results and Discussion In Situ Derivatization Reaction The conversion of reactive organohalides into S-alkylisothiouronium derivatives by refluxing with thiourea has been shown to be quantitative by using the ion-selective electrode method.6 In order to determine the time required for the derivatization reaction, the reaction profiles for butyl bromide and benzyl chloride were re-established using the proposed method.By systematically increasing the derivatization time, the peak current of the resulting solution observed in the differential-pulse polarogram gradually increased until a maximum value was reached. The time required to achieve the maximum value of the peak current is likely to be the time needed for the quantitative derivatization 0.4 1 I 0.3 - f 2 ; 0.2 - 2 Y m a 0.1 - I of the organohalides. By using this assumption, the time required for the quantitative derivatization of benzyl chloride and butyl bromide was found to be very similar in both instances (i.e. , 2 h); this is in good agreement with results from previous work.6 In order to confirm that the derivatization reaction was quantitative , the differential-pulse polarographic calibration graphs of S-benzylisothiouronium solutions prepared both from the authentic salt and the in situ generated salt (from benzyl chloride) were constructed and compared.It was found that the two graphs were almost coincident (Fig. 1). Hence, the in situ generation of the S-benzylisothiouronium salt from benzyl chloride under the proposed conditions is quantitative. In order to ensure the completion of the derivatization reaction for small sample sizes, a longer reaction time and a greater molar equivalent of thiourea should be used. The derivatization conditions for different amounts of benzyl chloride (i. e., 0.1-100 mg) were established (Table 1). Quantitative derivatization could be achieved by refluxing benzyl chloride with an excess of thiourea in ethanol for 3 h.For even smaller amounts of sample (i.e. , <0.1 mg), a longer reaction time was required for complete derivatization. The effect of residual ethanol and thiourea, which act as the solvent and derivatization agent, respectively, on the determination of alkyl halides was also investigated. On the I I G 0 1 3 5 7 9 11 Concentration/lO-5 mol dm-3 Calibration graph obtained for S-benzylisothiouronium solu- Fig. 1 tion prepared from: A , the authentic salt; and B, in situ generation Table 1 In situ derivatization conditions for different amounts of benzyl chloride Mass of Molar equiva- benzyl lent of Amount of Con- chloride/mg thiourea ethanol/ml version (%) 100-500 1.3 15 loo* 10-100 3 10 loo* 1-10 5 5 loo* 0.1 2 mg 2 75 * >95t * Refluxing for 3 h; the extent of the derivatization was assessed by t Refluxing for 5 h; the extent of the derivatization was assessed by using a standard solution of S-benzylisothiouronium.using a standard solution of S-benzylisothiouronium. -0.4 -0.5 -0.6 -0.7 -0.8 E N versus SCE Fig. 2 Typical differential-pulse polarographic responses for the in situ determination of butyl bromide (a) at lower concentrations: A. 1.61 X B, 3.21 x 10-6; C, 4.80 x 10-6; D, 6.38 X 10-6; E, 7.94 x 10-6; F, 9.49 x 10-6; and G, 1.03 x 10-5 rnol dm-3. (6) At higher concentrations: A , 1.03 x 10-5; B, 1.33 X 10-5; C, 2.29 X D, 3.68 x 10-5; E, 5.59 x 10-5; F, 7.30 x 10-5; G, 8.86 x 10-5; and H, 9.82 x 10-5 mol dm-3ANALYST, FEBRUARY 1992, VOL. 117 187 Table 2 Calibration equations for different alkyl halides (concentration range from 1 the corresponding S-alkylisothiouronium salts Alkyl halide Equation* n Ally1 bromide i = 5.736 X 10% - 0.020 11 Benzyl chloride i = 2.987 X 103c - 0.003 12 Butyl bromide i = 3.108 X 10% - 0.011 12 Butyl iodide i = 3.539 x 10% - 0.008 12 Propyl bromide i = 3.393 X lO3c - 0.010 12 * i is expressed in pA.t r = Correlation coefficient. $ Concentration 1.0 x mol d r r 3 . x 10-6 to 1 x 10-4 rnol dm-3) and Ep values of EJV r t versus SCE 0.999 -0.582 0.997 -0.680 0.998 -0.642 0.999 -0.642 0.996 -0.620 -0.7 -0.6 w 0 a -0.5 v) 2 -0.4 9) < -0.3 -0.2 -0.1 0 11.0 11.5 12.0 12.5 13.0 PH Plot of E, versus pH for the differential-pulse polarographic Fig. 3 wave exhibited by: A, S-butylisothiouronium salt; and B, thiourea systematic addition of up to 1 ml of ethanol and up to 2 mg of thiourea to 25 ml of 1.00 x 10-5 rnol dm-3 S-benzylisothiouronium chloride solution, the peak shape and peak height of the differential-pulse polarographic wave of the solution remained unchanged, i.e., the presence of residual ethanol and thiourea did not affect the polarographic measurement in this determination. Characteristics of the Differential-pulse Polarographic Wave Butyl bromide was used as a representative example of an organohalide for detailed studies. Under the derivatization conditions, a single differential-pulse polarographic peak was obtained at 0.64 V versus SCE (for a concentration of 1 X 10-5 rnol dm-3 butyl bromide). The calibration graphs were rectilinear over the halide concentration range from 1 X 10-6 to 1 x 10-4 rnol dm-3 (in the sample solution) and the relative standard deviations were good (1.1% for three determinations). Typical differential-pulse polarograms used to obtain a calibration graph are shown in Fig.2. All the polarograms showed a narrow symmetrical peak which is suitable for quantification. The differential-pulse polaro- graphic waves of the solutions covering the lower concentra- tion range studied [Fig. 2(a)] exhibited some residual current in the pre- and post-wave regions owing to the reduction of thiourea and residual oxygen, respectively. However, this did not affect the accuracy of the determination. On the other hand, the peak current was found to be directly proportional to the square root of the height of the mercury reservoir.This indicated that the polarographic current was diffusion controlled. In addition, the choice of the electrolyte system was adequate for this study. The peak potential ( E p ) values of the differential-pulse polarographic wave shifted to more negative values on increasing the concentration of the alkyl halide. A plot of E, versus log [RBr] in the concentration range from 8 x 10-6 to 8 x 10-5 rnol dm-3 was linear with a slope of 29.2 mV. Owing to the simplicity of the sample treatment procedure, attention was directed initially to the determination of reactive organohalides. When several reactive organohalides were Table 3 Working calibration graph data for the microdetermination of benzyl chloride (microgram level) 10 pg level 1 pg level Mass of id Mass of id samplelpg nA 20.8 30.6 2.08 2.34 41.6 57.0 4.16 4.14 62.4 87.0 6.24 6.12 83.2 115.2 8.32 7.50 104.0 148.2 10.40 9.96 samplelpg nA r = 0.9993 r = 0.9971 subjected to the proposed polarographic method, good calibration graphs, covering two orders of magnitude (from 1 x 10-6 to 1 x 10-4 mol dm-3), were obtained in all instances.The calibration graphs are described by the equations given in Table 2. The effect of the substituent in the S-alkylisothiouronium salt on the E, is evident (last column of Table 2), thus allowing the qualitative and quantitative determination of different organohalides. Effect of pH Under alkaline conditions, the S-alkylisothiouronium solution exhibits a well-defined differential-pulse polarographic wave which is amenable to analytical investigation.However, at relatively high concentrations of benzylisothiouronium (i. e . , >1 x 10-4 rnol dm-3), a turbid solution is obtained in a strongly alkaline (pH >13) medium owing to the hydrolysis of alkylisothiouronium. In order to ascertain the working pH for the determination of alkyl halides, the effect of pH on the E, values of the differential-pulse polarographic waves exhibited by thiourea and the S-butylisothiouronium salt generated in situ from butyl bromide was studied. The results are shown in Fig. 3. The difference between the E, values of the two compounds over the pH range 12-13 is sufficiently large to minimize any possible interference caused by the residual thiourea from the derivatization reaction.In addition, the peak current (ip) of a standard solution of benzylisothiouro- nium salt is also constant over this pH range. Hence subsequent polarographic studies of organohalides were arbitrarily carried out at pH 12.6. Also, any interference effect caused by the presence of heavy metal ions will be completely eliminated by the formation of insoluble hydroxides under the alkaline conditions used. Working Calibration Graphs for the Microdetermination of Benzyl Chloride Under the described conditions, samples containing as little as 0.001 mg of benzyl chloride can be derivatized satisfactorily to the corresponding isothiouronium salt. The resulting reaction mixtures from different sample sizes were subjected to polarographic measurement after derivatization.Working calibration graphs with good linearity were obtained for halide sample sizes of the order of 1 and 10 pg (Table 3 ) . These can be188 ANALYST, FEBRUARY 1992, VOL. 117 Table 4 Microdetermination of benzyl chloride by in situ derivatiza- tion to the corresponding isothiouronium salt Experi- ment No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Mass of benzyl chloride used/mg 536.9 501.0 107.1 99.4 55.2 46.4 17.5 5.32 0.0194 0.0416 0.0624 0.0832 0.00194 0.00416 0.00832 Mass of benzyl chloride found*/mg 551.7 511.0 105.8 97.1 55.6 47.5 16.7 5.12 0.0212 0.0405 0.0620 0.0812 0.00212 0.00412 0.00775 Error 2.6 2.0 -1.3 -2.3 (Yo 1 0.7 2.4 -2.8 -3.8 -2.7 -0.6 -2.4 9.2 -1.0 -6.9 9.2 * Experiments 1-8 were carried out by using the standard additions method and experiments 9-15 by using working calibration graphs.Table 5 Reproducibility and precision of the proposed method as illustrated by the determination of benzyl chloride Mass of benzyl chloride/mg Experiment Recovery* No. Added Found (%) 1 2 3 4 5 6 7 8 9 10 0.1040 0.1040 0.1040 0.1040 0.1040 0.1040 0.1040 0.1040 0.1040 0.1040 0.1002 0.1082 0.1002 0.1041 0.1081 0.1085 0.1002 0.1039 0.1055 0.0984 * Relative standard deviation = 3.68% 96.2 104.0 96.3 100.1 103.9 104.3 96.3 99.0 101.4 94.6 used as the working calibration graphs for the subsequent microdetermination of benzyl chloride. Visibility of the Method for the Microdetermination of Reactive Alkyl Halides For the actual determination of halides, as illustrated by benzyl chloride, two different methods can be used depending on the size of the sample.For samples on the milligram scale, the standard additions method was found to be suitable (experiments 1-8 in Table 4). For samples on the sub- milligram scale, quantification of benzyl chloride can be achieved by using the working calibration graphs covering the appropriate concentration range (experiments 9-15 in Table 4). In order to demonstrate the reproducibility and precision of the proposed method, ten samples of benzyl chloride were subjected to voltammetric analysis. The average recovery was found to be 99.6% and the relative standard deviation was 3.68% (Table 5). Conclusion An indirect polarographic method for the determination of reactive organohalides has been developed. In the presence of thiourea, the quantitative in situ generation of the S-alkyliso- thiouronium salt from the corresponding electroinactive organohalide provides the basis for the viability of the method. The method can be carried out in aqueous solution and can be used to detect as little as 1 pg of halide. References Gessner, G. N., The Condensed Chemical Dictionary, Van Nostrand Reinhold, New York, 8th edn., 1971, p. 359. Olson, E. C., in Treatise on Analytical Chemistry, eds. Kolthoff, I. M., and Elving, P. J., Wiley, New York, 1971, vol. 14, pt. 11, Al-Abachi, M. Q., and Salih, E. S., Analyst, 1987, 112,485. Ware, M. L., Argentine, M. D., and Rice, G. W.,Anal. Chem., 1988, 60, 383. Smyth, W. F., Polarography of Molecules of Biological Signi- ficance, Academic Press, London, 1979, p. 28. Chan, W. H., Lee, A. W. M., and Cheung, Y. M., Analyst, 1991, 116,39. Smyth, M. R., and Osteryoung, J. G., Anal. Chem., 1977, 49, 2310. Fumiss, B. S., Hannaford, A. J., Smith, P. W. G., andTatchell, A. R., Textbook of Practical Organic Chemistry, Longman, London, 5th edn., 1989, p. 789. Dearr, J. A., Lange’s Handbook of Chemistry, McGraw-Hill, New York, 13th edn., 1987, pp. 5-101. p. 1. Paper 1 l03648C Received July 18, 1991 Accepted September 10, 1991

ISSN:0003-2654    

DOI:10.1039/AN9921700185

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1992, VOL. 117 189 Application of Ion-exchanger Phase Spectrofluorimetry to the Determination of Micro-amounts of Some Rare Earth Elements by Flow Analysis Kazuhisa Yoshimura and Shiro Matsuoka Chemistry Laboratory, College of General Education, K yush u Un iversit y, Ro pponmatsu, Ch uo- ku, Fukuoka 810, Japan Toyohisa Tabuchi and Hirohiko Waki Department of Chemistry, Faculty of Science, Kyushu University, Hakozaki, Higashi-ku, Fukuoka 812, Japan The fluorescence bands originating from d+f electron transitions, which can be used t o determine europium, terbium, dysprosium or samarium selectively, were enhanced only if these elements were sorbed in a weak-acid cation-exchange gel, i.e., CM-Sephadex. After a sample solution had been introduced into a fused-silica tube (1.5 mm i.d.) packed with 1 mg of CM-Sephadex, the fluorescence intensity increase, resulting from the rare earth elements sorbed in the ion-exchange gel, could be measured directly with good precision.For europium, the detection limit was 22 ng with an 8.3 cm3 sample solution. The sensitivity of the proposed method depended on the sample volume introduced. The cell could be used repeatedly after desorbing the target element with a solution of nitric acid. Keywords: /on-exchanger phase spectrofluorimetry; flow analysis; rare earth element determination; CM-Sephadex sorption An individual determination of rare earth elements is difficult because of the similarity of their chemical properties. Although the absorption bands or the fluorescence bands originating from f-f or d-f electron transitions are spec- trometrically specific for some rare earth elements, they are not very sensitive.One attempt to enhance the sensitivity was to measure directly light absorption by neodymium, which had been preconcentrated onto a cation exchanger.' On the other hand, it is known that the sensitivities are enhanced when some rare earth elements form complexes with ligands containing oxygen as the donor atom.2.3 Similar phenomena occur for europium, terbium, dysprosium and samarium, but only when they are sorbed in a weak-acid cation-exchange gel, i.e., CM-Sephadex. Ion-exchanger phase spectrofluorimetry, developed by Waki et af.4 and used for the determination of trace amounts of beryllium,4,5 aluminium,6 gallium7 and tungsten,8 is a very sensitive method consisting of the simultaneous ion-exchange concentration of a target element and direct fluorescence measurements of the ion-exchange gel.This method can be applied to the rare earth element-CM-Sephadex system. However, if the batch concentration method is used, this spectrofluorimetry involves a time-consuming procedure of separating the ion-exchange gel from the bulk solution and packing it into a cell. Moreover, not all of the target element sorbed in the ion-exchange gel can be used for measurement. A more rapid and sensitive way is to use a flow-through cell packed with a much smaller amount of the ion-exchange gel. In this work, this method of spectrofluorimetry was shown to be applicable to the determination of europium, terbium, dysprosium and samarium by flow analysis, using CM- Sephadex gel as the ion-exchange retention medium.Experimental Reagents All of the chemicals used were of analytical-reagent grade. De-ionized water filtered through a 0.45 vm Millipore filter was used for the dilution of samples and reagents. Each standard solution of europium, terbium, dysprosium and samarium was prepared by dissolving each of the respective chlorides in 0.1 mol dm-3 hydrochloric acid and then standardizing by titration with ethylenediaminetetraacetic acid (EDTA), with Xylenol Orange as the indicator. A carrier solution was prepared by diluting a mixture of 1.9 g of ammonium acetate, 1.4 cm3 of acetic acid and 0.11 g of calcium chloride with water to a total volume of 1 dm3. A desorbing agent solution was prepared by diluting 10 cm3 of concentrated nitric acid to 1 dm3 with water.A cross-linked dextran-type cation-exchange gel, CM- Sephadex C-25, was purchased from Pharmacia in the sodium form. Apparatus Fluorescence measurements were made using a Nippon Bunko Model FP-SOA, or a Shimadzu Model RF-5000 spectrofluorimeter. The carrier solution was pumped with a medium-pressure pump (GL Sciences, Model MPD-3MG, or Sanuki, Model DM2M-1024). The flow-through cell shown in Fig. 1 consisted of a fused-silica tube (i.d. 1.5 mm, 0.d. 4 mm) with a poly- In 1 Ion-exchanger . Poly(propy1ene) Ex filter / i out Fig. 1 Flow-through cell for - Detector ion-exchanger phase spectro- fluorimetry. (a) Flow-through cell: a fused silica tube (1.5 mm i.d.), packed with 1 mg of CM-Sephadex C-25; and (b) micro-cell holder (GL Sciences)190 ANALYST, FEBRUARY 1992, VOL. 117 D 8o 8 Fig.2 Schematic diagram of the flow analysis set-up. A, Pump; B and C, six-way rotary valves each with a PTFE tube loop; and D, spectrofluorimetric detector packed with ion-exchanger. Carrier solution, 0.05 mol dm-3 acetate (pH = 4.7,40 mg dm-3 calcium); flow rate, 1.5 cm3 min-1; ion-exchanger, CM-Sephadex C-25; flow- through cell, 1.5 mm i.d.; A,, = 395 nm, he, = 616 nm (propylene) filter tip at the bottom, packed with about 1 mg of the ion-exchange gel. The cell, held in an accessory designed for a 3 mm micro-cell (GL Sciences), was placed in such a way that the excitation beam could enter only the ion-exchange gel. A schematic diagram of the flow analysis set-up is shown in Fig. 2.A sample loop (8.3 cm3) was made by using a poly(tetrafluoroethy1ene) (PTFE) tube (1 mm i.d., 2 mm 0.d.). Each of the sample solutions and the desorbing agent solution were introduced into the flow system by means of a six-way rotary valve. The flow rate was maintained constant at 1.5 cm3 min-1. All of the tubing was made of PTFE. Measurement of Distribution Ratio After 200 cm3 of a sample solution containing 4 mg of europium had been equilibrated with 100 mg of CM-Sephadex C-25, the europium concentration of the supernatant solution was measured by the method described below. The distribu- tion ratio, D , of the component is defined by the equation: D = [(mol of the component sorbed)/(g of ion-exchange gel)]/[ (mol of the component in solution)/(cm3 of solution)] (1) Procedure for the Determination of Europium A sample solution containing 0.04-2 pg of europium was introduced into the carrier stream.The fluorescence emission intensity was measured continuously at 616 nm (20 nm slit-width) using an excitation wavelength of 395 nm (10 nm slit-width). The emission filter was a Shimadzu 0-56 high-path filter; the excitation filter, a Shimadzu B-390. The increase in fluorescence intensity from the background was measured on a chart recorder. After each measurement, the europium in the flow-through cell was desorbed by the introduction of about 3 cm3 of the desorbing agent solution into the carrier stream. Procedure for the Determination of Terbium Terbium (0.1-2 pg) was similarly determined.The fluor- escence intensity was measured at 544 nm, with an excitation wavelength of 351 nm. The emission filter was a Shimadzu Y-50 high-path filter; the excitation filter, a Shimadzu U-340. Procedure for the Determination of Dysprosium Dysprosium (0.2-2 pg) was also determined in a similar way. The fluorescence intensity was measured at 573 nm, with an excitation wavelength of 350 nm. The emission filter was a Shimadzu 0-56 high-path filter; the excitation filter, a Shimadzu U-340. Procedure for the Determination of Samarium The fluorescence intensity of samarium (0.4-2 pg) was measured at 596 nm, with an excitation wavelength of 401 nm. 2 60 .- v) a, t l .- a C a v) 40 3 G= a, > .- +I - a a 20 0 I 550 600 650 Wavelengthlnrn Fig. 3 Emission spectra at an excitation wavelength of 395 nm for acetate buffer solutions containing 1 X 10-3 mol dm-3 europium.A, 0; B, 0.1; C, 0.3; D, 1; and E, 4 rnol dm-3 acetate (pH 4.7) The emission filter was a Shimadzu 0-56 high-path filter; the excitation filter, a Shimadzu B-390. Results and Discussion Fluorescence Spectra of Europium in Ion-exchange Gel and in Solution The fluorescence spectra of europium in solutions containing different concentrations of acetate (pH 4.7) are shown in Fig. 3. In the visible region, two peaks were observed for each spectrum: the maximum intensities were at an excitation wavelength of 395 nm. The peak at 592 nm is assigned to the 5Do+7F1 transition, and the peak at 616 nm to 5Do-+7F2. With the deviation of europium from octahedral symmetry, the increase in intensity of the 5DO+7F2 transition is greater than that of the 5Do+7F1 transition.9 Europium, to which a large number of ligands containing oxygen donor atoms coordinate, yielded a high fluorescence intensity. By increasing the concentration of acetate, the fluorescence intensity at 616 nm is increased more than that at 592 nm, owing to the complexation of europium with carboxyl groups.Using 0.3 rnol dm-3 acetate (Fig. 3, curve C), the europium to carboxyl group ratio is estimated to be 1 : 2 or 1 : 3, by using reported stability constants of acetato com- plexes.10 The spectrum of europium sorbed in the weak-acid cation-exchange gel, CM-Sephadex, which has carboxyl groups as functional groups, is similar to that shown in Fig. 3 curve C, and therefore the degree of complexation in the cation-exchange gel corresponds to that in a 0.3 mol dm-3 acetate solution.Also, in the ion-exchange gel phase, exci- tation at 395 nm gave maximum intensities at two emission peaks. As shown in Fig. 4(a), the degree of complexation by the fixed functional groups did not change in the ion-exchange gel phase at low europium loadings and, therefore, the emission intensity was proportional to the amount of europium sorbed in the gel. However, the ratio of the fluorescence intensity at 616 nm to that at 592 nm decreased with an increase in europium loading [Fig. 4(6)]: the peak area ratio was 1.35 for curve A and 1.28 for curve D. This means that the degree of complexation is lower at high europium loadings.ANALYST, FEBRUARY 1992, VOL.117 191 80 > cn C 4- .- 2 60 .- a, C a, cn 2 2 40 w- a, .- c - a, LT 20 a1 E 150 100 50 D 550 600 650 550 600 650 Wavelengthhm Fig. 4 (a) Net emission spectra at low loadings and (b) emission spectra at high loadings, with an excitation wavelength of 395 nm. Solution, 0.05 rnol dm-3 acetate (pH 4.7, 100 cm3); ion-exchanger, CM-Sephadex C-25 (100 mg); micro-cell, 3 x 3 x 35 mm. (a) Europium concentration: A, 2; B, 4; C, 6; D, 8; and E, 10 mg dm-3. (b) Europium concentration: A, 10; B, 50; C, 100; and D, 200 mg dm-3 Table 1 Relationship between internal diameter of flow-through cell and relative fluorescence intensity Internal diametedmm 1 .o 1.5 2.0 Relative fluorescence intensity* 2.2 1 0.6 diameter cell. * Normalized value with respect to the intensity using a 1.5 mm Optimization of Measurements Geometry of the flow-through cell The sensitivity was compared using flow-through cells with different internal diameters (Table 1).Ion-exchange gel beads were packed in cells of equal height, and the linear velocity for each cell was kept constant. For a 2.5 mm diameter cell, the intensity was 40% lower than that for a 1.5 mm diameter cell. The intensity for a 1.0 mm diameter cell was about twice that for a 1.5 mm diameter cell, but the measurement time was too long: two samples per hour. The position of the flow-through cell in the cell compart- ment is critical, because, when the excitation beam hits the poly(propy1ene) filter, light scattering increases. As the width of the beam was about 6 mm, the ion-exchange gel column in the flow-through cell had to be about 5 mm in height for the excitation beam to strike the ion-exchange gel beads at the top and to cover the area packed as widely as possible.The change in the shape of the ion-exchange gel beads at the top gave rise to remarkable errors because the light scattering character- istics changed. It might be thought that the precision of ion-exchange phase spectrofluorimetry , which involves an unorthodox solid-phase optical medium, would be inferior to that of conventional solution spectrofluorimetry . However, the errors are not serious for such low sample concentrations, to which the conventional solution method cannot be applied directly. Selection of ion-exchanger Four different types of ion-exchanger were tested for polymer matrices, with different functional groups and degree of 0 30 60 Timehin Fig. 5 Fluorescence development profiles of europium obtained using the flow system with ion-exchanger phase spectrofluorimetry.Europium concentration: A, 0; B, 10; C, 20; D, 30; E, 40 and F, 50 pg dm-3. Sam le volume, 8.3 cm3; carrier solution, 40 mg dm-3 Ca in 0.05 rnol dm- s acetate (pH 4.7); flow rate, 1.5 cm3 min-l cross-linking. Cation-exchange resins of cross-linked poly- styrene, for example, Bio-Rad AG 5OW-X12 (100-200 mesh, hydrogen form), showed an intense background emission in the visible region because of fluorescent impurities present in the resin. A strong-acid cation-exchange gel of cross-linked dextran, SP-Sephadex C-25, did not increase the sensitivity.Although a weak-acid cation-exchange gel with a low degree of cross-linking, CM-Sephadex C-50, increased the sensitivity for 0.1-1 yg of europium, it was 20% lower than that for CM-Sephadex C-25. Therefore, CM-Sephadex C-25 seemed to be the best choice for this method. Effect of p H The relative fluorescence intensity varied at pH values lower than 4.5 owing to shrinkage of the ion-exchange gel and the decrease in the distribution ratio; the intensity was almost constant in the pH range 4.5-7. The pH of the solution was fixed at 4.7. After adjusting the pH of a sample solution, the determination of europium should be made within 1 h, otherwise some loss of europium from the solution, due to its adsorption on glassware vessels, will occur. Effect of acetate concentration In order to maintain a constant pH of the sample solution, an acetate buffer solution (0.05 mol dm-3) was used.With 0.2 mol dm-3 acetate, the fluorescence intensity was 25% lower than with 0.05 mol dm-3 acetate. This was because the distribution ratio was lowered owing to the formation of acetato complexes in the solution with high acetate con- centration. Calibration and Sensitivity Fig. 5 shows typical examples for fluorescence development of europium and continuous measurement of samples. The calibration graph obtained was linear but had a positive blank. Similar results were obtained for the samarium, terbium and dysprosium systems. Relative fluorescence intensity caused by the sorption of each rare earth element is summarized in Table 2; europium could be most sensitively determined.Enhancement of Ion-exchanger Phase Spectrofluorimetry by Using a Flow-through Cell Packed With Ion-exchange Gel For an ion-exchange gel layer prepared with m g of ion- exchange gel (previously equilibrated with V em3 of solution containing a sample component of concentration co mol dm-3), the fluorescence intensity, F , can be approxi- mately expressed as:192 ANALYST, FEBRUARY 1992, VOL. 117 Table 2 Sensitivity and detection limits Wavelengthhm Sample volume/ Detection limitt/ Element Excitation Emission cm3 RFI* pg dm-3 Sm 401 596 4.8 0.41 120 (n =5) Eu 395 616 8.3 55.3 2.7 (n = 5) Tb 35 1 544 8.3 13.9 5.9 (n = 4) Dy 350 573 8.3 6.10 9.6 (n = 6) * Relative fluorescence intensity caused by the sorption of each rare earth element (100 pg dm-3).t The concentration that produces a fluorescence intensity equal to twice the magnitude of the fluctuation in the background fluorescence intensity. if the concentration of the sample component is sufficiently low and D is large.3 The parameter I. is the excitation beam intensity; @, the quantum yield of fluorescence; a , the absorptivity; I , the light path; v , the specific volume in the equilibrated state (5 cm3 g-1 for CM-Sephadex C-25); and k , an instrumental proportionality constant. (The over-bar refers to the ion-exchange gel phase.) If V << mD as for the present systems [for the europium- CM-Sephadex system, D was 2.1 x 105 cm3 g-1 (pH 4.7,0.05 mol dm-3 acetate with 40 mg dm-3 calcium)], eqn. (2) can be simplified as follows: Eqn.(3) shows that there is a linear relationship between F and co, and that by increasing Vlmv (the solution to ion-exchange gel volume ratio) , the present method becomes much more sensitive. The sensitivity of the proposed method was compared with the solution method as follows. The intensity of the CM- Sephadex C-25 layer (0.1 g of the ion-exchange gel equili- brated with a 100 cm3 solution of 1 mg dm-3 europium) was compared with that of a 200 mg dm-3 europium solution (0.3 mol dm-3 acetate). The concentration of europium in the gel phase was the same as that in the solution. Under such conditions, ion-exchanger phase spectrofluorimetry by the batch method was 200 times more sensitive than the solution method. For the flow method, after packing 1 mg of the ion-exchange gel into the flow-through cell, and loading 10 cm3 of a 0.25 mg dm-3 europium solution, the intensity of the ion-exchange gel layer was compared with that of the solution described above: the flow method using a 10 cm3 water sample is much more sensitive than the solution method by a factor of 800.For the batch method, spectrofluorimetric measurements were made using a micro-cell 3 x 3 x 35 mm, packed with at least 30 mg (0.15 cm3) of the ion-exchange gel and the area where the light beam struck was about 1/4 of the total area of the ion-exchange gel packed in the micro-cell, and, therefore, 0.0375 cm3 of the ion-exchange gel beads in the cell was used for measurements. Therefore, the amount of europium irradiated in the two cells was 7.5 yg for the batch method and 2.5 pg for the flow method.Although the pathlength for the flow method is smaller than that for the batch method, higher sensitivity was obtained: the geometry for fluorescence measurement might be more favourable for the flow method than that for the batch method. Above all, with a flow-through cell packed with a small amount of ion-exchange gel, the feasibility of much higher volume ratios should greatly increase the sensitivity of ion-exchanger phase spectrofluorimetry . Detection Limit The detection limit, defined as the concentration producing a fluorescence intensity equal to twice the magnitude of the Table 3 Effect of foreign ions on the determination of europium. Sample, 4.8 cm3 (0.200 mg dm-3 Eu); carrier, 40 mg dm-3 Ca in 0.05 mol dm-3 acetate (pH 4.7) Concentration/ Europium found/ Foreign ion mg dm-3 mg dm-3 Ca 0 0.214 20 0.204 200 0.189 400 0.161 Fe 10 0.213 La 2.0 0.211 10 0.233 Sm 10 0.206 Tb 10 0.209 DY 2.0 0.188 10 0.178 Error +7.0 +2.0 -5.5 -19.5 +6.5 +5.5 +11.5 +3.0 +4.5 -6.0 -11.0 (Yo ) Table 4 Determination of samarium, europium and dysprosium in synthetic samples and a mineral Sample Concentration found/mg dm-3 Synthetic sampleslmg dm-3- Ca 40, Sm 2 Ca4O,FelO,La2,Sm2,Eu0.01, Sm2.14 Tb0.5, Dy2 Ca 40, Eu 0.05 Ca 40, Fe 0.5, La 0.05, Sm 0.05, Eu 0.05, Tb 0.05, Dy 0.05 Ca 40, Tb 0.05 Y50.6,LaOS,Cel.2,Pr0.2, Dy6.7k0.6(%m/m)(n=5) Sm 2.00 f 0.095 (n = 5) Eu 0.050 k 0.0038 (n = 5) Eu 0.0492 Tb 0.050 f 0.0012 (n = 5) Yttrium concentrate* (% m/m)- Nd0.8,Srn0.9,EuO.l,Gd2.9, Tb0.8,Dy6.8,Ho1.6,Er5.5, Tm 0.8, Yb 8.3, Lu 0.8 * Concentrated from xenotime (Shin-etsu Kagaku). Elemental analysis was carried out using X-ray fluorescence spectrometry.fluctuation in the background fluorescence intensity, is shown in Table 2. For the europium system, the detection limit was 22 ng, i.e., 2.7 pg dm-3, with an 8.3 cm3 sample solution. Effect of Sample Volume For the present system, D is sufficiently large to satisfy eqn. (3). This means that the sensitivity will be increased only by introducing a larger volume of sample solution. A variation in the sample volume from 2.1 to at least 16.6 cm3 resulted in a proportional increase in the fluorescence intensity. Much higher sensitivities can be achieved by employing larger volumes of sample solution. Effect of Foreign Ions With 8.3 cm3 sample solutions containing 50 pg dm-3 europium, the effects of the concomitant ions of calcium, iron, lanthanum, samarium, terbium and dysprosium were exam- ined.The presence of 40 mg dm-3 calcium or 2 mg dm-3 dysprosium gave remarkable negative errors, but other ions gave no interference at the 1 mg dm-3 level. The ion-exchange gel layer in the flow-through cell became contracted when the loaded ion was converted from a monovalent to a polyvalent cation. Therefore, the error induced by calcium might be due to shrinkage of the ion-exchange gel. In order to eliminate this effect, the counter ion of the ion-exchange gel was changed from ammonium to calcium by adding calcium chloride to a sample and the carrier solution. When 4.8 cm3 of a 0.20 mg dm-3 europium solution were introduced, in the presence of 40 mg dm-3 calcium, calcium no longer interfered up to 200 mg dm-3 (Table 3).The presence of 10 mg dm-3 dysprosium gave a negative error; the presence of 10 mg dm-3 lanthanum, a positive error.ANALYST, FEBRUARY 1992, VOL. 117 193 In addition to europium, samarium, terbium and dyspro- sium can be determined using the proposed method. Elements that interfered when present at a concentration of less than that of the target element were europium and terbium for the samarium system, samarium for the terbium system and terbium for the dysprosium system. Determination of Samarium, Europium, Terbium and Dysprosium in Synthetic Samples and a Mineral With 4.8 or 8.3 cm3 sample solutions, the precision was measured with a Shimadzu Model RF-5000 type spectro- fluorimeter.For five determinations, each relative standard deviation was within 10% (Table 4). The results in Table 4 also show that the method is applicable to the determination of dysprosium in an yttrium concentrate. The value obtained by the proposed method is in close agreement with that obtained by X-ray fluorescence spectrometry. Conclusion Although methods such as neutron-activation analysis and inductively coupled plasma optical emission spectrometry have been used for the determination of trace amounts of various rare earth elements, access to the necessary equip- ment is not always available. Without any fluorescent dyes, the proposed method affords an on-line, simple, rapid and fairly sensitive method to determine europium, terbium, samarium or dysprosium selectively using relatively inexpen- sive apparatus, because both concentration and spectro- fluorimetric measurements are carried out simultaneously. The fluorescence bands originating from the d-+f electron transition can be used to determine individual rare earth elements selectively and can be enhanced in a CM-Sephadex ion-exchange gel phase. The proposed method is applicable to the determination of trace amounts of a wide variety of target elements by employing the appropriate fluorescent dye. 1 2 3 4 5 6 7 8 9 10 References Yoshimura, K., and Taketatsu, T., Fresenius 2. Anal. Chem., 1987, 328, 553. Chrysochoos, J . , and Evers, A., Chem. Phys. Lett., 1973, 18, 115. Bunzli, J.-C. G., and Yersin, J.-R., Helv. Chim. Acta, 1982.65, 2498. Waki, H., Noda, S., and Yamashita, M., React. Polym., 1988, 7, 227. Capitan, F., Manzano, E.. Navalon, A., Vilchez, J. L., and Capitan-Vallvey, L. F., Analyst, 1989, 114, 969. Capitan, F., Manzano, E., Vilchez, J. L., and Capitan-Vallvey, L. F., Anal. Sci., 1989, 5 , 549. Capitan, F., Navalh, A., Vilchez, J. L., and Capitan-Vallvey, L. F., Talanta, 1990, 37, 193. Capitan, F., de Gracia, J. P., Navaldn, A., Capitan-Vallvey, L. F., and Vilchez, J. L., Analyst, 1990, 115, 849. Blasse, G., Bril, A., and Nieuwpoort, W. C., J . Phys. Chem. Solids, 1966, 27, 1587. Grenthe, I., Acta Chem. Scand., 1962, 16, 1695. Paper 01055681 Received December I I , I990 Accepted September 19, I991

ISSN:0003-2654    

DOI:10.1039/AN9921700189

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1992, VOL. 117 195 Spectrofluorimetric Determination of the Insecticide Azinphos-methyl in Cultivated Soils Following Generation of a Fluorophore by Hydrolysis F. Garcia Sanchez and A. Aguilar Gallardo Department of Analytical Chemistry, Faculty of Sciences, University of Malaga, 29071-Malaga, Spain A rapid and simple spectrofluorimetric method for determining residues of the pesticide and acaricide azinphos-methyl in soil is described. Soil is extracted with methanol. The method is based on the alkaline hydrolysis of the insecticide t o its main metabolite anthranilic acid. A detailed study of the parameters affecting the chemical hydrolysis is presented and the results are discussed. The calibration graphs were linear from 30 t o 1000 ng ml-1, with a limit of detection of 8 ng ml-1.The precision of the method is 2% at the 60 ng mi-1 level. Recoveries of spiked soil samples were between 105 and 11 1 %. Keywords: Azinphos-meth yl determination; spectrofluorimetry; soil analysis; insecticide; acaricide Analysis of pesticides in environmental matrices is of increas- ing interest because of their widespread use for both agricul- tural and industrial purposes. Although gas chromatography (GC) remains the major method of determination, high- performance liquid chromatography (HPLC) and discontinu- ous techniques for pesticides are often used. Methods for the determination of pesticide residues must offer high sensitivity. Molecular fluorescence is a detection technique that is highly regarded as an analytical tool because of its excellent sensitivity; however, its use has been limited because few pesticides fluoresce naturally and several prelimi- nary derivatizing steps may need to be carried out.Azinphos-methyl { S-(3,4-dihydro-4-oxobenzo[d][ 1,2,3]- triazin-3-ylmethyl) 0,O-dimethyl phosphorodithioate} is a non-systemic insecticide and acaricide that persists in the environment for long periods of time.' It is effective chiefly against biting and sucking insect pests and is used mainly on citrus, cotton, grapes, maize, some ornamental plants, top quality fruit and vegetables. Several workers have reported methods for the determination of azinphos-methyl using a variety of techniques such as titrimetry,* spectrophotometry,3 mass spectrometry4 and HPLC5-9 with ultraviolet (UV) detection or GC.10,Il The titrimetric 'method is based on hydrolysing azinphos-methyl in methanol with 3 rnol dm-3 NaOH in the presence of phenol followed by argentimetric titration of the 0,O-dimethyl hydrogen phosphorodithioate formed with dichlorofluorescein as indicator.The spectropho- tometric method involves the reaction of azinphos-methyl with 4-(4-nitrobenzyl)pyridine and oxalic acid in acetone medium to yield a solution with an absorption maximum at 560 nm. In a multi-residue method using reversed-phase HPLC with fluorimetric detection,l2 azinphos-methyl spiked in vegetables has been determined at 0.1-5 ppm with recoveries of between 93 and 115%. In the present paper the utility of spectrofluorimetry to determine the insecticide azinphos-methyl in soil samples by a simple method is described.The proposed method is based on the alkaline hydrolysis of the insecticide to give the fluoro- phore anthranilic acid, which is monitored at 394 nm with excitation at 314 nm. This provides a sensitive method (limit of detection, 8 ng ml-1) for the determination of azinphos- methyl that is applicable to residues in soils and gives recoveries ranging from 105 to 11 1%. Experimental Apparatus Emission measurements were carried out with a Perkin-Elmer LS-5 luminescence spectrometer (Perkin-Elmer, Beacons- field, Buckinghamshire, UK), equipped with a xenon dis- charge lamp (9.9 W) pulsed at the line frequency, F/3 Monk-Gillieson type monochromators, and 1 x 1 cm quartz cells. The spectrofluorimeter was operated in the computer- controlled mode via the RS232C serial interface by a Perkin-Elmer Model 3600 data station microcomputer. Instrumental control and data collection were achieved by using the commercially available Perkin-Elmer computerized luminescence software (PECLSII) .The system enables de- rivative spectra to be recorded. In order to ensure that all measurements could be compared and that repeatable measurements could be obtained, the LS-5 spectrofluorimeter was checked daily. A fluorescent sample of the polymer p-terphenyl(1 X mol dm-3) gave a relative fluorescence intensity (RFI) of 90% at A,, = 340 nm with A,, = 295 nm, with a slit-width of 2.5 nm and a sensitivity factor of 0.5973. For graphical recording, an Epson FX-85 printer-plotter (Seiko Epson, Suwa, Japan) was connected to the spectroflu- orimeter.All fluorescence spectra are uncorrected because no significant wavelength shifts were observed when the spectra were compared with corrected spectra. Ultraviolet absorption spectra were recorded with a Shimadzu UV-240 Graphicord recording spectrophotometer (Shimadzu, Kyoto, Japan). A rotary vacuum evaporator (W. Buchi Scientific Apparatus, Flawil, Switzerland) and an Ultrasons Selecta ultrasonic water-bath (Selecta, Barcelona, Spain) were used to homogenize soil samples. Reagents Stock solutions of azinphos-methyl [ >99% pure (Pestanal quality), Riedel-de-Haen , Seelze, Hannover, Germany] were prepared in ethanol at concentrations of 1.0 mg ml-1. Working solutions at concentrations of 100 pg ml-1 were prepared in ethanol.All solvents used were of analytical- reagent grade and were obtained from Merck (Rahway, NJ, USA). The NaOH was also of analytical-reagent grade (Merck). The water used was distilled and de-mineralized. Procedures Analytical procedure Transfer aliquots of a standard solution of azinphos-methyl (1 mg ml-1) in ethanol into 10 ml calibrated flasks in order to obtain a final concentration between 0.03 and 1 pg ml-1. Add 6 ml of ethanol and 2.5 ml of 0.2 rnol dm-3 NaOH (final concentration 5 x 10-2 rnol dm-3). Dilute to the mark with de-ionized water. Heat the resultant solutions for 10 min in a water-bath at 85 "C, cool for 3 min under running water and allow to stand for 7 min at room temperature. Measure the196 ANALYST, FEBRUARY 1992, VOL. 117 fluorescence intensity at 394 nm with excitation at 314 nm, against a solvent blank.The concentration of azinphos-methyl is determined from the conversion of RFI units by reference to the calibration graph. Extraction of soil samples Soil samples were obtained from three cultivated fields in Coin, Algarrobo and Campanillas (Malaga, Spain) by using a V-shaped shovel introduced into the ground to a depth of 20 cm. Twenty sub-samples were taken from different sites of the same field and combined to give a final mass of approximately 2 kg. The sample was spread in a dish, and large pieces and pebbles were removed. The sample was mixed thoroughly and divided into 300 g portions. Each soil sample was air-dried at room temperature and passed through a 2 mm sieve. Volumes of the stock solution of azinphos-methyl in ethanol were added to a 10 g portion of the soil sample in a beaker so that the percentage recovery obtained using the proposed method could be calculated.After thorough mixing, the sample was extracted with methanol in a proportion of 2 + 1 solvent to sample (v/m). In order to achieve a rapid homogenization of the sample with the extracting solvent, the beaker was placed in an ultrasonic bath for 1 min after which the sample was left to stand. The supernatant was filtered through a 30 ml Biichner funnel of medium porosity, and a vacuum was applied. The procedure was repeated three times. The contents of the filter flask were transferred quantitatively into a round-bottomed flask and taken to near dryness on a rotary evaporator at 45 "C.The residue was made up to a final volume of 10 ml with ethanol. This solution was used for the analytical determination. Results and Discussion The alkaline hydrolysis of azinphos-methyl leads to the formation of the metabolite anthranilic acid. The reaction is slow, and essentially depends on the pH of the medium and on the temperature. The fundamental chemical behaviour of the insecticide azinphos-methyl and its main metabolite anthranilic acid is shown in Scheme 1. This dynamic process is an irreversible hydrolytic process which leads to the generation of a fluoro- phore, and is a reaction that can be monitored by making fluorescence measurements. This behaviour facilitates the spectrofluorimetric determination of the non-fluorescent reagent.When the solution of azinphos-methyl is added to a strongly basic medium, the fluorescence emission spectrum observed shows a maximum at 394 nm when excited at 314 nm (Fig. 1). As the reaction is too slow at room temperature, even in basic medium, to be of use, the effect of temperature is significant in the optimization of the experimental procedure. Effect of the Experimental Variables The effect of the concentration of NaOH on the rate of hydrolysis was examined by monitoring the fluorescence emission of 3.15 x 10-6 rnol dm-3 solutions of azinphos- methyl in water-ethanol (40 + 60) at 394 and 314nm containing 0.01-0.10 mol dm-3 NaOH. The solutions were heated to 85 "C for 20 min and, after cooling, the fluorescence was measured. The results show that constant values of fluorescence intensity were obtained in the range 0.02-0.06 rnol dm-3 NaOH.Higher concentrations gave a diminution in fluor- escence readings, probably because of the decomposition of the hydrolysis products. An NaOH concentration of 0.05 rnol dm-3 was finally selected for the alkaline working solutions. Under these conditions the hydrolysate of azinphos- methyl is stable for at least 1 h. Temperature and heating time also have an important influence on the final conditions required for hydrolysis. The effect of these parameters was studied by measuring the fluorescence intensity for the total reaction at two different temperatures (65 and 85 "C) for 0.05 rnol dm-3 solutions of NaOH for heating periods of between 0 and 50 min. Fig. 2 shows that at lower temperatures longer heating times are 280 330 380 430 ?Jnm Fig.1 A, Excitation and B, emission spectra of anthranilic acid in water-ethanol(40 + 60). [Anthranilic acid] = 3.15 x rnol dm-3; LX = 314 nm, A,, = 394 nm s MeO,ll Me0 + ,P-s- CH20H 0 Azi n p h 0s- met h y I A H - s MeO, I I 1 COOH Anthranilic acid Scheme 1ANALYST, FEBRUARY 1992, VOL. 117 197 I 1 I I 0 10 20 30 40 50 Timehi n Fig. 2 Effect of temperature on hydrolysis at A, 85 "C and B, 65 "C. [Anthranilic acid] = 3.15 x 10-6 mol dm-'; [NaOH] = 0.05 mol dm-3; A,, = 314 nm, he, = 394 nm required in order to achieve constant readings in fluorescence intensity. From these experiments, it was deduced that a heating time of 10 min at 85 "C is sufficient to obtain optimum results. Analytical Parameters The calibration graphs were prepared using a set of standards individually hydrolysed and measured as described under Analytical procedure and plotting the RFI values against concentration in ng ml-1.Two concentration ranges of azinphos-methyl can be covered by linear calibration graphs, 100-1000 ng ml-1 and 20-100 ng ml-* , using direct fluorescence intensity measure- ments. Applying statistical treatment to the analytical data, the linear regression graphs obtained are as follows: IF = 0.0626 [azinphos-methyl] + 0.223 IF = 0.3770 [azinphos-methyl] + 1.390 Y = 0.997 Y = 0.997 where IF is the fluorescence intensity, Y the correlation coefficient and [azinphos-methyl] is in ng ml- 1. The sensitivity of the method is reported as the analytical sensitivity, sA = a/m, and has a value of 1.23 ng ml-1 where (J is the standard deviation of the analytical signal (n = 7) and rn is the slope of the calibration graph.13 The limit of detection, cL(K = 3) and the limit of quantification, cQ(K = 10) are reported as defined by IUPAC14 and have values of 8.13 and 27.11 ng ml-1, respectively; K is a numerical factor chosen in accordance with the confidence level desired.The limit of quantification, cQ, is employed to establish the lower limit of the linear dynamic range. The relative error of the method was 1.90% and a relative standard deviation (RSD) of 2% was obtained at the 60 ng ml-1 level ( n = 7). Analysis of Soil Samples Azinphos-methyl is used for the control of insect pests in fruits and vegetables at rates of 3 4 I ha-1; it persists in the soil for about 28 d.Soil samples from cultivated citrus fields in southern Spain (Coin, Algarrobo and Campanillas, Malaga) were used to demonstrate the applicability of the proposed spectrofluori- metric method. The first step in the determination of pesticide residues is usually the separation of the residues from the matrix material by solvent extraction. For efficiency, the solvent must remove the pesticide in a reproducible manner without removing large amounts of interfering compounds from the matrix. One of the most complicated procedures is the extraction from soil, because the extraction efficiency is affected by the type of soil, the properties of the extractant, method of extraction, etc. Ultrasonic techniques are generally preferred as the effect of water in the soil or type of organic matter present in the soil is avoided and the efficiency of the Table 1 Analysis of azinphos-methyl residues in ground soil; n = 3 Azinphos- Azinphos- methyl methyl added/ found/ Recovery recovery Mean Typeofsoil pgml-1 pgml-l (YO) (Yo ) Ground soil 0.9 0.984 109.3 0.965 0.970 107.7 107.2 108.1 k 1.1 Siliceous Clays 0.9 0.945 105.0 0.955 106.1 105.6 k 0.5 0.951 105.6 0.9 0.947 105.4 0.960 106.6 107.7 -t 3.0 1.000 111.1 extraction is increased. This is due to the breakdown of soil structure, allowing the extractant to work on a larger surface area.Therefore, the ultrasonic technique was selected for the extraction procedure. With respect to the solvent used to extract the pesticide from soils, a search of the literature showed that most workers have found that higher efficiency is achieved by using methanol as the solvent.15-'9 Interference problems are generally avoided with soil samples when the proposed extraction procedure is used, except for those interferents that are organic in nature and could fluoresce in an alkaline medium.The proposed method can be applied to the analysis of soil samples fortified with standard solutions of 0.9 pg ml-1 of azinphos-methyl, following the extraction procedure de- scribed above, The concentration of azinphos-methyl found in soil samples and the corresponding percentage recoveries obtained using the proposed method are given in Table 1 (n = 3). Blank signals corresponding to untreated soil samples were sub- tracted from the recovery data.The measurements obtained show that soil samples contain- ing azinphos-methyl at residue levels may be quantified by the proposed spectrofluorimetric method. No significant differ- ences in the percentage recovery were found for the three types of soils tested. The recoveries obtained are acceptable. No previous publications reporting recoveries of azinphos-methyl in soils have been found in the literature. Conclusion Determination of azinphos-methyl at residue levels in soil samples can be accomplished readily by using hydrolysis- induced fluorescence spectrometry. Despite the widespread use of chromatographic methods in the determination of pesticide residues, the development of alternative methods that are simpler and more rapid than those already in use would be useful to the analytical community.Recovery assays of azinphos-methyl in several soil samples show good results compared with those obtained by use of HPLC. References 1 Pesticide Manual, British Crop Protection Council, ed. Worth- ing, C. R., British Crop Protection Council, Croydon, 7th edn., 1983. 2 Muntaz, M., Nasir, N. E. R.. and Baig, M. M., Pak. J . Sci. Ind. Res., 1983, 26. 132. 3 Gunther. F. A., Iwata, Y., Papadopoulo, E., Berck, B . , and Smith, C. A., Bull. Environ. Contam. Toxicol., 1980, 24, 903. 4 Schulten, H. R., and Sun, S . , J. Environ. Anal. Chem., 1981, 10, 247. 5 Bushway, R. J . , J. Liq. Chromatogr., 1982, 5 , 49.198 ANALYST, FEBRUARY 1992, VOL. 117 6 Wilson, A. M., and Bushway, R. J., J. Chromatogr., 1981,214, 140. 7 Funch, F. H., Z. Lebensm. -Unters. Forsch., 1981, 173, 95. 8 Farran, A., and de Pablo, J., Znt. J. Environ. Anal. Chem., 1987,30, 59. 9 Marutoiu, C., Vlassa, M., Sarbu, C., and Nagy, S., HRC CC, J. High Resolut. Chromatogr. Chromatogr. Commun., 1987,10, 465. 10 Pressley, T. A., and Longbottom, J. E., Gov. Rep. Announce. Index (US), 1982, 82, 1544. 11 Allmaier, G., Goergl, A., Schmid, E. R., and Wagner, K., HRC CC, J. High Resolut. Chromatogr. Chromatogr. Com- mun., 1986, 9, 762. 12 Krause, R. T., and August, E. M., J. Assoc. Off. Anal. Chem., 1983, 66, 234. 13 Garcia Sanchez, F., and Cruces Bfanco, C., Anal. Chem., 1986, 58, 73. 14 Long, G. L., and Winefordner, .I. D., Anal. Chem., 1983, 55, 7 12A. 15 Klisenko, M. A., Scr. Fac. Sci. Nut. Univ. Purkynianae Brun, 1980, 1068. 16 McKone, C. E., J. Chromatogr., 1969, 44, 60. 17 Kahn, S. U., Greenhalgh, R. I., and Cochrane. W. P., Bull. Environ. Contam. Toxicol., 1975, 13, 602. 18 Cotterill, E. G., Pesric. Sci., 1980, 11, 23. 19 Peiia-Herasa, A., and Sanchez-Rasero, F., J. Chromatogr., 1986,358, 302. Paper 1 l02283K Received May 15, 1991 Accepted October 8, 1991

ISSN:0003-2654    

DOI:10.1039/AN9921700195

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1992, VOL. 117 199 Linear Titration Plot for the Determination of Boron in the Primary Coolant of a Pressurized Water Reactor Derek Midgley and Christopher Gatford National Power plc, Technology and Environmental Centre, Kelvin Avenue, Leatherhead, Surrey KT22 7SE, UK A linear titration plot method has been devised for the determination of boron as boric acid in partly neutralized solution, such as occurs in the primary coolant of pressurized water reactors. The total boron and the alkali in the sample are determined simultaneously. Although it is not essential to add mannitol in this method, it is more accurate when the solution is saturated with mannitol. Comparisons are made with other modes of titration: Gran plots, first and second differential potentiometric titrations and indicator titrations.None of these gives the total boron directly in partly neutralized solutions. Keywords: Potentiometric titration; Gran plot; linear titration plot; boric acid; pressurized water reactor Boric acid is added to the primary coolant of pressurized water reactors (PWRs) in nuclear power stations to provide fine control of the reactivity by means of the neutron-absorbing capability of the 1OB isotope, which constitutes 20% of natural boron. The boron concentration is reduced from 2000 mg 1-1 at the start of a fuel cycle to less than 100 mg 1-1 at the end. Daily analysis of the coolant is required. In addition to providing information on 1OB for neutron control, the boric acid concentration also determines the amount of lithium hydroxide to be added to maintain the pH at a level that minimizes corrosion and transport of corrosion products.Other solutions at PWR stations need analysing for boric acid, e . g . , in boric acid storage tanks, recovery liquors and auxiliary systems, but these analyses are less critical. Alkalimetric titration of boric acid is long-established, but in this instance the acid will always be partly neutralized to a small extent (=lYo) and some means of correcting for this is required. As the sample will be slightly radioactive, auto- mated procedures with a minimum of sample handling are desirable. Theory Boric acid is weakly dissociated (pK, = 9.23 at 25°C) and conventional titration is impractical, whether using a col- orimetric or potentiometric end-point.Adding a polyol that complexes with boric acid, as in eqn. (l), releases protons that can be titrated in the usual way; Belcher et al. 1 confirmed that mannitol, HOCH2(CHOH)4CH20H, was the most suitable reagent and the apparent pK of boric acid in its mixture of complexes is about 5.2 Stability constants in 0.1 moll-1 KCl medium at 25 "C have been determined as follows,3 where L represents mannitol: B(OH)4- + L = B(OH)2(H-2L)- : logpl = 4.000 B(OH)4- + 2L = B(H-2L)2- : 10gp2 = 4.888 The above constants are probably averages for a number of steric isomers. Linear titration plots475 can cope with acids as weak as boric acid, even without the addition of mannitol and can also be used to calculate the total and neutralized acidity.6 Applica- tion of the latter plot to data from the literature for boric acid produced results of reasonable but not great accuracy. A disadvantage of linear titration plots, apart from the original Gran plots,7 is the necessity of knowing the stability constants relevant to all the equilibria involved.The requirement for computing facilities has also been disadvantageous, but with the availability of personal computers and spreadsheets, this is no longer a problem. Linear titration plots have not been considered for complex systems such as borate-mannitol, but examination of the charge and mass balance equations shows that a simple equation is obtained if the mannitol concentration is kept constant. This can be achieved in practice simply by saturating the solution with mannitol. For a volume (Vo ml) of a solution of boric acid (total boron concentration CB moll-1) partly neutralized with CL, moll-1 of lithium hydroxide after the addition of V ml of sodium hydroxide titrant (rn moll-1) in the presence of mannitol: total boron, TB = CBVd(V0 + V) = [B(OH)3] + [B(OH),-] + [B(OH)2(H-2L)-l + [B(H-2L)2-1 + Pl[LI + P2ELI2) (1) + 2[B(H-2L)2-1 (2) (3) (4) = [B(OH),-]({H+}f/Ka + 1 total mannitol, TL = [L] + [B(OH)2(H-2L)-] total lithium, [Li+] = CLiVd(V0 + V) total titrant, "a+] = rnV/(Vo + V) charge, [H+] + [Li+] + "a+] = [OH-] + [B(OH)4-] +[B(OH)2(H-2L)-I + [B(H-2L)2-1 = [OH-] + TB - [B(OH)3] ( 5 ) where K, = {B(OH)4-} {H+}/[B(OH)3] is the dissociation constant and f is the univalent ion activity coefficient.If V, is the volume of titrant required to neutralize the sample, mV, = vO(cB - CLi) = m(vB - VLi), (6) where VL, and V, are the volumes of titrant equivalent to the lithium hydroxide and total boron, respectively, in the sample.Substituting in eqn. ( 5 ) , {H+}/f+ mVLi/(VO + V) + rnV/(Vo + V) = (1 - K,/{H+}f+ rnV~/[(Vo + v) X + (1 + PILL] + P2[L12)Ka/{H+Ifl-1}l On re-arrangement, (Vo + V)({H+} - K,/{H+})/rnf+ V = Ve - VB/[~ + (1 + PI[L] + P2[LI2)Ka/{H+}f] (7) The constants K,, Ka, P1 and P2 are known, as are the experimental quantities rn, Vo, V and {H+}, so the left-hand side of eqn. (7) can be plotted against 1/[1 + (1 + p1[L] + P2[L]2)Ka/{H+}f], if [L] is known, to give a straight line with a slope of -VB and an intercept on the ordinate of V,. Hence the original concentrations can be calculated: CB = rnVB/Vo and CLi = m(VB - Ve)/Vo200 ANALYST, FEBRUARY 1992, VOL.117 If the solution is saturated with mannitol, [L] is constant and the calculation is complicated only by the presence of activity coefficients. The titration can be carried out either at constant ionic strength, with a fixed value off, or f can be calculated from an equation such as the Davies equation,s with iterative corrections to the ionic strength, I . When I is equal to the left-hand side of eqn. ( 5 ) , in which "a+] is known and [H+] is initially approximated by {H+}; on the first cycle, [Li+] is neglected. After the first cycle, [Li+] is known to a good approximation and included in calculations of I until success- ive iterations agree. Experiment a1 Titrations were carried out with an Orion 960 titrator operating in Gran mode with approximately 10 mV spacing between points.Standard boron solutions were prepared from Aristar grade boric acid (BDH) and BDH ConvoL sodium hydroxide was used as the titrant. The glass electrode was calibrated each day with standard potassium hydrogen phtha- late and KH2P04-Na2HP04 buffers. Calculations were carried out using the Microsoft Excel 3.0 spreadsheet, but they are not dependent on any special feature of this product and almost any spreadsheet should suffice. Results Boric Acid Solutions Standard boric acid solutions equivalent to 0.46-9.25 ml of 0.1 moll-' NaOH were titrated in volumes of 20-25 ml. Results were calculated by the linear titration plot method on a spreadsheet and by the Gran plot software built into the Orion 960 titrator.The Gran weak acid plot, V{H+} versus V, was used for data before the equivalence point and the strong base plot, (Vo + V)/{H+} versus V, for data after the equivalence point. Segregation of the data was performed automatically by the Orion 960 titrator. The results are shown in Table 1 for boric acid solutions both with and without mannitol. The linear titration plot calculations were carried through with respect to the original volume of solution, i.e., without correction for volume changes produced by the Table 1 Results for unneutralized boric acid solutions by linear titration plot (LTP) and Gran procedures (25 ml sample, 0.1 mol 1-l titrant) LTP Gran error in VB (YO) Theo- retical Error in Pre-end- Post-end- Run VB/ml VB(%) VLi/ml point point Mean With mannitol- 1 9.251 2 9.251 3 4.626 4 4.626 5 2.313 6 2.313 7 0.463 8 0.463 -1.89 -0.74 -0.12 -1.27 - 1.56 -2.42 4.50 4.17 -0.028 -0.023 -0.005 -0.009 -0.004 -0.014 0.013 0.014 -0.97 0.06 0.12 -0.64 -0.55 -1.37 2.26 1.39 -0.30 0.63 0.57 -0.01 -0.08 -0.55 3.99 2.91 -0.64 0.35 0.35 -0.32 -0.32 -0.96 3.12 2.15 Mean -0.01 -0.007 0.04 0.89 0.47 Without mannitol- 9 9.251 0.12 10 9.251 0.40 12 4.626 1.00 13 2.313 0.87 14 2.313 0.76 15 0.463 4.17 16 0.463 7.02 Mean 1.56 11 4.626 -1.83 * ND = Not detected. 0.008 0.039 0.026 0.064 0.045 0.045 0.009 0.020 0.032 2.14 -11.11 ND* 1.72 2.91 1.74 9.61 7.88 2.13 - 1.46 -0.96 ND -2.54 -2.89 -3.75 -2.28 -0.12 -2.00 0.34 ND -0.41 0.01 -1.01 3.66 3.88 0.06 -6.04 dissolution of mannitol.The value of [L] was estimated, by interpolation of solubility data,9 to be 1.2 mol 1-1 on this basis, but this tended to produce small positive values of VLi even when only boric acid was present (an average of 0.03 2 0.01 ml for the runs in Table 1).The [L] was adjusted within each run until IVLiI <0.0005; the mean of these optimized concentrations was 1.17 +. 0.02, which was used to give the results in Table 1, where the mean VLi is -0.007 +. 0.015 ml. The plots are not shown here, but would be similar to those shown in Fig. 1 for partly neutralized solutions, except for a displacement on the ordinate. The mean error in VB was only 0.1% by the linear titration plot, although this was fortuitous to the extent that positive errors at the lowest concentration tested were balanced by negative errors at higher concentra- tions.The Gran weak acid plot gave still smaller errors, with a smaller spread and the Gran post-end-point plot relatively larger errors. Repeating the titrations without mannitol shows that the linear titration plot gave fairly good results, whereas those for the Gran weak acid plot were very scattered. The Gran post-end-point plot generally gave a slight underestimate of the titre. In one instance the software failed to find a segment sufficiently linear to qualify as a Gran plot, although the linear titration plot was satisfactory. Linear titration plots for these runs would be similar to those in Fig. 2, but displaced on the ordinate. In these calculations only one round of iteration was necessary. By working with a constant ionic background of potassium nitrate, iteration was avoided, but without an lo i 0 0.2 0.4 0.6 0.8 1.0 X Fig.1 Linear titration plots for partly neutralized boric acid solutions in the presence of mannitol: A, run 2 and B, run 5 from Table 1. Where Y = (Vo + V)({H+} - K,/{H+})/mf + V ; and X = 141 + (1 + PJLI + P2[LI2)Ka/{H+>fl 0 0.2 0.4 0.6 0.8 1.0 X Fig. 2 Linear titration plots for partly neutralized boric acid solutions without mannitol: A, run 9 and B, run 13 from Table 1. X and Y as defined in Fig. 1ANALYST, FEBRUARY 1992, VOL. 117 201 improvement in accuracy. The automatic titrator took read- ings rapidly, but no improvement in accuracy was obtained by slowly adding titrant and judging equilibrium from a display of the potential on a chart recorder. For titrations in which it is known that no neutralization of the boric acid has occurred, the Gran weak acid procedure would be preferred for its accuracy and simplicity, provided that mannitol is added.Note, however, that the post-end- point plot, in which base is in excess, appears to be better than the Gran weak acid plot itself. The linear titration plot allows the mannitol to be omitted, with little loss of accuracy, whereas the weak acid Gran procedure is rather unreliable in such circumstances. Partly Neutralized Boric Acid Solutions The above runs were re-calculated as if the titrations had started after a small volume, V,, of titrant had been added. Thus, V, is identical with VLi. The results are shown in Table 2. In these titrations, the pH data are the same as for the runs in Table 2 Results for partly neutralized boric acid solutions by linear titration plot procedure Theoretical Found Error in Error in Run VB/ml VLi/ml VB (Yo ) VL, (YO ) With rnannitol- 1A 2A 3A 4A 5A 6A 7A 8A Mean 9.251 9.251 4.626 4.626 2.313 2.313 0.463 0.463 Without mannitol- 9A 9.251 10A 9.251 11A 4.626 12A 4.626 13A 2.313 14A 2.313 15A 0.463 16A 0.463 Mean 0.655 0.655 0.555 0.555 0.555 0.504 0.302 0.302 -1.71 -0.52 0.21 - 1.02 - 1.22 -2.33 5.57 5.43 0.5.5 0.202 0.202 0.202 - 0.202 0.202 0.202 0.202 0.151 0.22 0.59 .1.35 1.31 1.57 1.42 6.76 7.02 2.19 -2.12 -0.94 1.41 0.15 0.46 -2.57 5.76 6.47 1.08 7.70 26.26 19.82 37.27 28.48 28.32 9.51 13.47 21.36 Table 1 and the volume data become Vo’ = Vo + V, and V’ = V - V,, where primed symbols refer to Table 2.Data corresponding to V < V, were omitted. For all but the lowest concentration (well below the normal operating range of the reactor) the accuracy when determining boron was almost the same as for unneutralized solutions. The lithium concentra- tion was also determined with reasonable accuracy. Typical plots are shown in Fig. 1: linearity over most of the plot is excellent, but as the abscissa tends to zero the ordinate increasingly deviates from the ideal value. These deviant points correspond to a pH >7 and the most likely cause is depression of the pH by the presence of carbon dioxide, whether arising from contamination of the titrant or absorp- tion during the titration. Points with an alkaline pH or abscissa ~ 0 . 0 1 should be excluded from the calculation and these are easily seen on the spreadsheet; no other points needed to be omitted from the data for the above results.Without mannitol, the boron concentration was still deter- mined fairly accurately, but the errors in the lithium concen- tration were considerable. Typical linear titration plots are shown in Fig. 2: the scatter is not very large, but greater than when mannitol is present. Results for Gran and conventional titrations are not shown, as even if perfectly executed they can yield only the boric acid concentration and not the total boron, i.e., they would have given an error in VB equal to VLi. In these examples the error would be 7% of the greatest boron concentration and 70% for the least. With these traditional titrations, a correction for lithium hydroxide must be made, either by first titrating with acid to pH 7 (before adding mannitol), so that VB is accurately determined in the subsequent titration with alkali.Alterna- tively, lithium can be determined by some other method (atomic spectrometry or with an ion-selective electrode) and a correction applied, but this is valid only in highly pure samples such as those used in this work, where it is safe to assume that there are no other sources of lithium or alkali in the boric acid solutions. Comparison of Conventional Titrations for Boric Acid Various instrumental methods of end-point detection are available in modern automatic titrators and these were tested for the titration of pure (unneutralized) boric acid solutions in the presence of mannitol.The results in Table 3 show that for the Orion 960 titrator the Gran mode was more accurate than the first and second derivative modes. For comparison, a further series of solutions (Table 4) was analysed with a Table 3 Results for different modes of end-point detection in titrations of boric acid in the presence of mannitol (Orion 960 titrator). All values for the error, within-batch (sW), between-batch (sb) and total relative standard deviations (st) are given in per cent; degrees of freedom are shown in parentheses Amount of B/mg (mg I - * ) Method First derivative Error Second derivative Error sw (2) sw (6) St (8) sw (6) st (8) sw (6) st (8) s b (2) Gran post-end-point (blank corrected) Error sb (2) Error sb (2) Gran weak acid, pre-end-point 10 (1997) 0.55 0.05 -0.60 0.39 NS* 0.50 0.10 0.10 NS 0.15 0.10 0.21 NS 0.29 5 (990.4) 0.92 0.05 1.98 0.08 0.44t 0.45 0.07 0.21 NS 0.29 -0.05 0.36 0 0.36 2.5 (500.8) 0.76 4.8 0.76 0.38 NS 0.52 0.50 0.16 0 0.16 0.38 0.30 0 0.30 l(200.4) 3.3 0.1 2.79 0.15 0.40t 0.40 0.50 0.45 0.50$ 0.70 1.25 0.65 NS 0.65 0.25 (49.6) 13.7 0.7 12.3 1.2 0 1.2 2.82 1 .o 2.4s 2.6 9.07 1.2 NS 1.4 * Not significant.t Significant at the P = 0.1% level. $ Significant at the P = 5% level. § Significant at the P = 1% level.202 ANALYST, FEBRUARY 1992, VOL. 117 Table 4 Boric acid titrations with the Metrohm 682 titroprocessor B takedmg 8 4 2 0.8 B takedmg 1-1 400 200 100 40 Mean error (YO) 0.35 0.58 1.6 2.6 sw* (%) 0.3 (4) 1.0 (4) 0.9 (10) 0.5 (4) * Within-batch relative standard deviation for a single result, with degrees of freedom in parentheses.Table 5 Titrations with visual end-point detection B takedmg 10 4 1 Error (YO) O* -0.25 1 .o B takedmg 1-I 500 200 50 s w t (%) 0.18 0.35 0.58 * Taken as standard. t Within-batch relative standard deviation (4 degrees of freedom). Metrohm 682 titroprocessor, which takes the end-point as the point of inflection of the titration curve and has a syringe burette instead of a calibrated pump, as with the Orion instrument. The precision and accuracy at any concentration were closely comparable for the two titrators. Finally, some titrations were carried out using visual end-point detection with phenolphthalein (Table 5), which produced results equal to those from the titrators. Tables 3-5 show the expected deterioration of accuracy and precision as the samples become more dilute.This is partly caused by loss of precision and accuracy in delivery of the titrant as the titre becomes smaller, but also by increasing errors in the approximations involved in the Gran functions4 or increasing discrepancy between the equivalence point and the point of inflection,lo however determined. In all of the techniques, the accuracy deteriorates more rapidly than the precision, because of these biases in the calculation of the equivalence point. Discussion The linear titration plot gives an accurate measure of the total boron concentration, even in partly neutralized solutions of boric acid, and also gives a good estimate of the amount of alkali added (identified with lithium in the present applica- tion).It is still advantageous to add mannitol, although not theoretically necessary. Without mannitol, the error in the estimate of added alkali becomes considerable. The calcula- tions are easily carried out on a spreadsheet and only one cycle of iteration is required to correct for variations in ionic strength. Alternatively, the ionic strength can be artificially maintained as constant, rendering iteration unnecessary. If the boric acid is known to be free of alkali, the Gran weak acid procedure is simpler and slightly more accurate than the linear titration plot, but the addition of mannitol is highly desirable. It is particularly convenient when the titrator is programmed to carry out the Gran plot, as in the Orion 960 titrator. Mannitol is essential for other modes of end-point indica- tion, such as first or second differential potentiometric titrations or the classical visual indicator technique, and these methods are then capable of giving excellent results in simple boric acid solutions. With these, as with both the Gran methods, a second determination is required to detect partial neutralization, necessitating additional sample handling. The linear titration plot thus provides a one-stop method for partly neutralized boric acid solutions and is particularly suitable when an automated method with minimal handling of samples is required. Submitted for publication by permission of National Power plc. 1 2 3 4 5 6 7 8 9 10 References Belcher, R., Tully, G. W., and Svehla, G., Anal. Chim. Acta, 1970, 50, 261. Dawber, J. G., and Matusin, D. H., J. Chern. Soc., Faraday Trans., 1982,78, 2521. Antikainen, P. J., and Pitkanen, I. P., Suom. Kernistil. B , 1968, 41, 65. Midgley, D., and McCallum, C., Talanta, 1974,21,723. Midgley, D., and McCallum, C., Talanta, 1976, 23, 320. Midgley, D., and McCallum, C., Fresenius 2. Anal. Chern., 1978, 290, 230. Gran, G., Analyst, 1952, 77,661. Davies, C. W., Ion Association, Butterworth, London, 1962. International Critical Tables of Numerical Data, Physics, Chem- istry and Technology, McGraw-Hill, New York, 1928, vol. IV. Meites, L., and Goldman, J. A., Anal. Chim. Acta, 1963, 29, 472. Paper 1/04397H Received August 22, 1991 Accepted October 10, 1991

ISSN:0003-2654    

DOI:10.1039/AN9921700199

出版商:RSC    

年代:1992

数据来源: RSC