摘要:

ANALYST, AUGUST 1991, VOL. 116 81 1 lndomethacin Ion-selective Electrode Based on a Bis(triphenylphosphoranylidene)ammonium-lndomethacin Complex Ralph Aubeck, Christoph Brauchle and Norbert Hampp* Institute of Physical Chemistry, Ludwig-Maximilians-Universitat, Sophienstrasse I I , 0-8000 Miinchen 2, Germany The properties of an ion-selective electrode for indomethacin based on the bis(tripheny1- phosphorany1idene)ammonium-indomethacin ion-pair complex in a poly(viny1 chloride) membrane are described. The detection limit found for indomethacin at pH 7.0 was 2 x 10-5 mol I-' (6.4 pg mi-1) and the linear range was determined to be between 5 x 10-5 and 310-3 mol 1-1. The selectivity coefficients observed for organic anions of different lipophilicity were 10-1.' for naproxenate, 10-0.75 for salicylate and <lo-4 for tartrate.Cationic organic and inorganic ions showed negligible interference with selectivity coefficients of <lo-4. A strong electrode response was observed for the inorganic anions lo4-, C104- and SCN- with selectivity coefficients of up t o lO+O-9. Very low detection limits, e.g., 6 x 10-7 mol 1-1 (60 ng ml-1) for C104- were found. The selectivity for inorganic anions was close to the Hofmeister series. Keywords: lndomethacin determination; ion-selective electrode; perchlorate/periodate determination The use of ion-pair based ion-selective electrodes (1SEs)l has several advantages in applications such as the determination of ionic drugs in vitro2 and in vivo3 or the recently reported monitoring of organic ions generated in situ .4 Ion-selective electrodes constitute a simple, rapid and inexpensive means of measuring the activity of an ionic analyte over a wide concentration range of, typically, 3-6 orders of magnitude without extensive sample preparation. Ion-selective electrode measurements provide a means of determining the ion-activity of a compound directly even in coloured or turbid solutions.Lipophilic quaternary amines, e.g., methyltrioctylammonium salts such as Aliquat 336s ,5 sodium tetraphenylborate6.7 or quaternary phosphonium compounds, e. g . , tetraphenyl- phosphonium chloride,s and bulky dyes such as Crystal Violet9 are commonly used as ion-pairing agents. This paper describes the use of an ion-pair complex formed from the bis(triphenylphosphorany1idene)ammonium (PNP) cation and the indomethacin (INDO) anion (Fig.1) in a drug electrode for the determination of the INDO activity at physiological pH values. Indomethacin is a non-steroidal anti-inflammatory analgesic which acts as a therapeutic agent for arthritis.10 In chemistry, PNP is used for the isolation and precipitation of labile organometallic anions. 11312 The high lipophilicity of PNP allows its use as a counter ion in liquid membrane electrodes. Experimental Reagents High relative molecular mass poly(viny1 chloride) (PVC), ortho-nitrophenyl octyl ether (o-NPOE), tetrahydrofuran (THF) and the sodium phosphate buffer salts were purchased from Fluka. Bis(triphenylphosphorany1idene)ammonium chloride (Cl-PNP, Aldrich) and INDO [ 1-(p-chlorobenzoy1)- 5-methoxy-2-methyl-3-indolylacetic acid; Sigma] were used as reagents in order to obtain the electroactive component of the electrode, the INDG-PNP complex.The following chemicals were used for the determination of the selectivity coefficients of the electrodes: (+)-sodium naproxenate (Sigma), sodium salicylate (Fluka) , potassium sodium tartrate (Fluka), glycine (Sigma), NaIO4, NaC104, NaN03, NaCl, NaSCN, KCl, MgC12 (all from Fluka), atropine sulphate (Roth), papaverine hydrochloride (Roth) , berberine hydrochloride (Roth) and * To whom correspondence should be addressed. INDO PNP Fig. 1 Structures of INDO and PNP glucosamine hydrochloride (Fluka). They were used as sodium phosphate buffered solutions (0.1 moll-*, pH 7.0) at concentrations of between 1 x 10-8 and 1 x 10-2 mol I-*.Reagents of the highest available purity and doubly distilled water were used for the preparation of the analytes. Preparation of the Electroactive Compound A 10 mmol 1-1 solution of the sodium salt of INDO was obtained by dissolving the free acid in sodium hydroxide solution. An equimolar solution of C1-PNP was added and the resulting, slightly yellow, precipitate was collected by centrifu- gation. The ion-pair complex obtained was washed thoroughly with distilled water and dried over P205 in a desiccator at room temperature for 18 h. Its composition was determined by elemental analysis (C, H, N). A 1 : 1 molar ratio of PNP to INDO was found with deviations of <0.3%. The melting- point of the INDO-PNP complex was determined to be 160-162 "C. Preparation of the Electrodes and Measurement Procedure A mixture containing 4.6% of the electroactive compound (INDO-PNP or Cl-PNP), 67.1% of plasticizer (o-NPOE) and 28.3% of PVC dissolved in THF was poured onto glass plates.After removal of the solvent, membranes of 100-150 pm in thickness remained. Pieces with a diameter of 9 mm were punched out of these membrane sheets and glued onto the front end of a PVC electrode body containing an inner Ag-AgC1 junction. A 3 moll-1 NaCl solution was used as the inner electrolyte. Potentiometric measurements were made at room temperature with a digital microprocessor pWion-meter (pMX 2000, WTW Weilheim). The electrode potentials were812 ANALYST, AUGUST 1991, VOL. 116 Table 1 Response characteristics of the electrodes containing the ion-pair complexes INDO-PNP and CI-PNP.Data were taken from measurements in 0.1 mol 1-1 sodium phosphate buffer at pH 7.0 Detection limit Electrode Ion INDO-PNP INDO Naproxenate Salicylate 104- c104- NO3- c1- SlopeImV decade-' 88 k 1.5 96 k 1.0 56 -L 1.5 57 k 0.5 56 k 1.0 =m* =59* moll-' 2 x 10-5 2 x 10-4 2 x 10-5 2 x 10-4 1 x 10-3 1 x 10-6 6 x lo-' yg ml-1 6.4 36.6 2.5 0.2 0.06 12.4 36 x 103 Linear range/molI-' 5 x 10-5 to 2-10-3 3 x 10-5 to 310-2 010-3)t (>10-*)t 6 x to 2 x 10-6 to 210-2 1 x 10-6 to 310-2 CI-PNP INDO 82 k 1.0 4 x 10-5 12.8 8 x 10-5 to 210-3 104- 54 k 1.5 8 x 1.6 2 x 10-5 to 310-2 c104- 57 f 1.0 9 x 10-6 0.9 3 x 10-5 to 210-2 NO3- -56* 2 x 10-3 124 (> 10-2)t c1- -56* 3 x 10-3 108 x 103 (>10-2)t Naproxenate 82 & 1.5 4 x 10-4 73.2 8 x 10-4 to 2-10-2 Salicylate 60 k 1.0 1 x 10-4 12.5 6 x to 2-10-2 * Extrapolated values owing to the low detection limit.t Extrapolated values. Table 2 Logarithmic selectivity coefficiencts, kE7' for the electrodes containing the ion-pair complexes INDO-PNP and Cl-PNP. Values were measured in 0.1 moll-' sodium phosphate buffer (pH 7.0) and calculated at the e.m.f. readings of 60 mV for each compound. Data <-4 were obtained with interferents which gave no response at concentrations of at least 1 mmol I-' Log kfq' Log kf7' Interfering Interfering ion INDO-PNP CI-PNP ion INDO-PNP Cl-PNP INDO 0 O* 104- +0.70 +0.50 Naproxenate - 1.10 - 1.05 C104- +0.90 + O M Salicylate -0.75 -0.75 N03- -2.00 - 1.95 Tartrate <-4 <-4 c1- -2.25 -3.05 G1 ycine <-4 <-4 SCN- 0 +0.05 K+, Mg2+, atropine, papaverine, berberine, glucosamine: log k c ' <-4 * INDO was taken as the primary ion for the Cl-PNP electrode for comparison with the INDO-PNP electrode.measured against a single junction Ag-AgC1 reference elec- trode (K 801, Radiometer) and recorded together with the temperature on a two-channel strip-chart recorder. All measurements were carried out in a flow cell (200-300 pl volume) under continuous-flow conditions. The flow rate of the analyte solutions was kept at 1 ml min-1 by means of a peristaltic pump (Microperpex, Pharmacia-LKB). A pro- grammable sample changer (Model 222, Gilson) was used to ensure reproducible sample handling. Most of the measure- ments were made at a neutral pH of 7.0, which is near to physiological conditions.In alkaline media, decomposition of INDO is observed.13 The potentiometric selectivity coeffi- cients, kfPf, were determined by using the separate solution method. 14 Results and Discussion Response Characteristics of the Electrodes Electrodes with membranes containing INDO-PNP or C1- PNP were characterized together in order to permit a direct comparison. The values obtained for the detection limit, linear range and slope of the electrodes are summarized in Table 1. The response time, i.e., the time taken to reach 90% of the final reading, for the different analytes varied from 5 to 10 s for high analyte concentrations and from 1 to 2 min for low analyte concentrations, which is the typical range for macro- scopic ISEs.7 The electrode slope for most of the analytes investigated was nearly Nernstian.For INDO and naprox- enate, super-Nernstian responses with slopes of about 90 mV decade-' were observed. This behaviour was observed for both membrane types, i.e., INDO-PNP and CI-PNP, both in Tris-HC1 [Tris = tris(hydroxymethyl)aminomethane] and in sodium phosphate buffer. Measurements in more alkaline media of pH 9 and 11 showed an almost Nernstian slope of 58 mV decade-'. This observation can be explained by reference to the pKa values of INDO (pK, = 4 3 , naproxen (pKa = 4.2) and salicylic acid (pK, = 3.0)13 at pH 7.0, when only partial dissociation of the carbonic acid groups of INDO and naproxen occurs, and by the formation of dimers, which is typical of carbonic acids, e.g., benzoic acid.15 The high sensitivity of the INDO-PNP electrode for C104- with a detection limit of 6 x 10-7 mol 1-1 and its wide linear range of between 1 x 10-6 and 310-2 mol 1-1 C104- is interesting as it also allows the use of this electrode for the determination of C104-.Previously reported C104- ISEs5716 showed lower sensitivities of the order of 10-5 mol 1-1. Selectivity of the INDO-PNP Membranes Several organic and inorganic compounds with different charges, lipophilicity and relative molecular mass were examined (Table 2). The relative order of the kfp', values measured for the INDO-PNP and CI-PNP electrodes was identical. However, the dynamic range of the INDO-PNP based electrode was extended by up to one order of magnitude in the direction of low concentrations. For lipophilic cationic interferents, e.g., berberine, no electrode responAe was observed, but inorganic anions such as 104- and C104- interfered strongly.The same behaviour was recently re- ported17 for an NO3- ISE with a nitrate-PNP ion pair dissolved in nitrobenzene. Logarithmic kf'yt values of 3.18 and 3.20 were reported for 104- and C104-, respectively.17 The relative order of the selectivity coefficients found for the NO3- electrode and the INDO-PNP and CI-PNP electrodes was identical, i.e., 104- = C104- > SCN- 3 INDO >ANALYST, AUGUST 1991, VOL. 116 813 5 min -3 1 1 I Time- Fig. 2 Stability of the INDO-PNP electrode response to alternating INDO and C104- activities measured in 0.1 moll-' sodium phosphate buffer at pH 7.0 under continuous operation in a flow cell at room temperature.The values of the logarithm of the INDO and C104- concentrations in the test solutions are given in the plot ([INDO] = 1 x 10-5-1 x 10-3 moll-1; [C104-] = 1 x 10-6-1 x 10-2 moll-1). Solid curve, 0 h; and broken curve, 48 h salicylate > NO3- > C1-, in agreement with the Hofmeister lipophilic series. 1 Significant differences were observed for Cl- and C104-. The sensitivities towards Cl- and C104- were approximately five times higher and three times lower, respectively, for the CI-PNP electrode compared with the INDO-PNP electrode. The results obtained using 0.1 moll-' Tris-HC1 buffer at pH 7.0 instead of sodium phosphate buffer were virtually identical, except for 1 0 4 - which oxidized the buffer. Reproducibility The reproducibility of the electrode signals was determined from continuous measurements at room temperature over periods of 12-48 h.Fig. 2 shows the response of the INDO-PNP electrode to INDO and C104- at pH 7.0. For the sample-to-sample reproducibility of the e.m.f. values an average deviation of k O . 1 mV was found. After a measuring period of 48 h, only a small decrease (3%) of the initial slope was observed. The measurement of increasing and decreasing concentration steps showed slightly different e.m.f. values in the increasing and decreasing part of the profile (see Fig. 2) for low concentrations, e.g., for 1 x 10-5 mol 1-1 C104-. This might be due to the reversible partial exchange of the INDO-PNP complex for C104-PNP at high C104- concentra- tions. This leads to a slightly extended linear range for low concentrations of C104-.This effect was found to be reproducible over the whole test period. Owing to the overproportional response of the INDO-PNP electrode to C104-, the possibility of an exchange of the INDO-PNP ion pair for the C104-PNP ion pair must be considered. During the 48 h test period even a slow exchange of this type should lead to the formation of a C104-PNP electrode. This would cause a change in the response curve. From Fig. 2 it can be concluded that INDO-PNP is the more stable complex and that any exchange with C104-PNP that occurs during the exposure to high C104- concentrations is only short-term. The equilibration times for decreasing concentration steps are of the order of less than 30 s, which is fairly fast for ion-pair based electrodes (see Fig.2). This is due to the use of a flow cell in which the analyte flows tangentially to the membrane surface. This accelerates any potential changes as no concen- tration gradients can build up at the membrane surface; such gradients would hinder the diffusion of ions into and out of the membrane. 18 Although a white precipitate can be obtained by mixing phosphate buffer with CI-PNP solution, the use of phosphate buffer is possible over long periods of time without any observable decrease in the electrode slopes for the analytes tested. Comparison With Other Analytical Methods The detection limit found for INDO with the INDO-PNP electrode was 6.4 yg ml-1. This is comparable to the detection limits of spectrophotometric methods which range from 1 to 20 pg ml-1; however, owing to the broad linear range (from 5 x 10-5 to 210-3 moll-1) of the electrodes described here, they might be more flexible in applications where a series of samples with widely varying INDO concentrations have to be analysed, e.g., in pharmaceutical manufacturing, or in appli- cations where continuous monitoring is required.Other methods reported20721 for the determination of INDO are more sensitive, but dilution and other time-consuming steps for sample preparation are required. Conventional high- performance liquid chromatographic methods20 show detec- tion limits of 0.1 pg ml-1. By using high-performance liquid chromatography with spectrofluorimetric detection,21 0.01 yg ml-1 of INDO can be detected in human serum.Owing to the detection limit of the INDO-PNP electrode for C104- of 6 x 10-7 mol 1-1 (60 ng ml-l), this electrode might also be used for the determination of C104- in analytes where high concentrations of INDO can be excluded. The INDO-PNP electrode is about one order of magnitude more sensitive to C104- than previously reported C104- ISEs; e.g., Hassan and Elsaiedl6 described a C104- ISE based on a nitron-perchlorate ion pair in nitrobenzene with a detection limit of 9 x 10-6 moll-' C104-. Conclusions The proposed INDO-selective electrode based on an INDO- PNP ion-pair complex as the electroactive compound might be a useful analytical tool for the in vitro determination of INDO in the range from 5 x 10-5 to 210-3 moll-1 and therefore an alternative to spectrophotometric methods.The interference from salicylate, which might be a problem in medical applications, can be compensated by a combination of the INDO-PNP electrode with a salicylate ISE.22 The INDO- PNP based ISE might also be useful for the determination of C104- or 104- because of its detection limit of 60 ng ml-1, which is approximately three times lower than that of the CI-PNP electrode. Interference from lipophilic organic cations is negligible. The high sensitivity to 1 0 4 - should make it possible to detect the kinetics of Malaprade reactions for the cleavage of glycol-type compounds in a more sensitive manner than previously reported.23 References Morf, W. E., The Principles of Ion-Selective Electrodes and of Membrane Transport, Elsevier, Zurich, 1981.Ma, T. S., and Hassan, S. S. M., Organic Analysis Using Ion-Selective Electrodes, Academic Press, London, 1982. Jaramillo, A., Lopez, S., Justice, J., Salamone, J., and Neill, D., Anal. Chim. Acta, 1983, 146, 149. Chan, W. H., Lee, A. W. M., and Chan, L. K., Analysf, 1990, 115, 201. Cammann, K . , Working with Ion-selective Electrodes, Springer. Berlin, 1979. Aubeck, R., Hampp, N., and Brauchle, C., Ber. Bunsenges. Phys. Chem., 1988,92, 1423. Aubeck, R., Brauchle, C., and Hampp, N., Anal. Chirn. Acta, 1990, 238,405. Amoroso, P., Campanella, L., Angelis, D. G., Ferri, T., and Morabito, R., J. Membr. Sci., 1983, 16,259. Ishibashi, N., and Kina, K., Bull. Chem. SOC. Jpn., 1973, 46, 2454.814 ANALYST, AUGUST 1991, VOL. 116 19 Sastry, C. S. P., and Rac, A. R. M., Analusis, 1987, 15, 569. 20 Brown, Y. L., Kandretas, R. J., Douglas, J. B., and Gal, P., J. Chromatogr., 1988,459,275. 21 Mawatari, K. I., Iinuma, F., and Watanabe, M., J. Chromatogr. Biomed. Appl., 1989, 03,389. 22 Mitsana-Papazoglou, A., Diamandis, E. P., and Hadjiioannou, T. P., Anal. Chim. Acta, 1984, 159, 393. 23 Efstathiou, L. E., and Hadjiioannou, T. P., Anal. Chim. Acta, 1977,89, 55. 10 11 12 13 14 15 16 17 18 Ullmann 's Encyclopedia of Industrial Chemistry, ed. Gerhartz, W., VCH, Weinheim, 1985, vol. A2, p. 269. Ruff, I., and Schlienz, W., Inorg. Synthesis, 1974, 15, 84. Oltmanns, P., and Rehder, D., J. Organomet. Chem., 1988, 345, 87. Merck Index, Merck, Rahway, NJ, 11th edn., 1989. Pungor, E., Toth, K., and Hrabeczy-Pall, A., Pure Appl. Chem., 1979,51, 1913. Nagaoka, S., Terao, T., Imashiro, F., Saika, A., Hirota, N., and Hayashi, S., Chem. Phys. Lett., 1981,80,580. Hassan, S. S. M., and Elsaied, M. M., Talanta, 1986,33, 679. Werner, G., Kolowos, I., and Senkyr, J., Talanta, 1989,36,966. Cammann, K., and Rechnitz, G. A., Anal. Chem., 1976, 48, 856. Paper 1 I00226 K Received January I6th, I991 Accepted March 25th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600811

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, AUGUST 1991, VOL. 116 815 Column Preconcentration of Trace Metals From Sea-water With Macroporous Resins Impregnated With Lipophilic Tetraaza Macrocycles Stephane Blain, Pierre Appriou and Henri Handel lnstitut d'Etudes Marines, Universite de Bretagne Occidentale, Faculte des Sciences et Techniques, 6 Avenue le Gorgeu, 29287 Brest Cedex, France Macroporous resins impregnated with lipophilic tetraaza macrocyclic derivatives were used for the uptake and enrichment of trace metals from sea-water prior to their determination by electrothermal atomic absorption spectrometry. The effects of the size of the macrocycle cavity, the support used for impregnation, the pH of the extraction and the parameters of the back-extraction were studied in a column process. The preconcentration of trace metals (Cd, Cu, Mn, Ni, Pb and Zn) from de-ionized water and sea-water was investigated.Satisfactory results were obtained for Cd, Mn, Pb and Zn with an average recovery of 98 ? 8%. Absolute blanks are in the range 1-10 ng and the detection limit varies from 0.7 ng for Cd to 15 ng for Pb. The accuracy of the method was demonstrated by replicate analyses of the National Research Council of Canada marine reference material NASS-2 (open ocean sea-water). The precision is better than 15% (lo). The selectivity of the resin for Cd, Cu, Mn, Ni, Pb and Zn over Nal, Call and Mg" was studied. Keywords: Tetraaza macrocycle; trace metal determination; sea-water; column preconcentration; electrothermal atomic absorption spectrometry The determination of trace metals in a complex medium such as sea-water generally requires a preconcentration step.One of the most widely used methods is based on extraction chromatography with metal chelating resins, the coated ligands being mostly iminodiacetate1.2 or 8-hydroxyquinoline (or its derivatives) impregnated on macroporous polymer~33~ or silica.5 Closed systems and reduced handling minimize the risk of contamination and produce a lower analytical blank than liquid-liquid extraction .6 Moreover, on-line preconcen- tration is possible, which, when coupled with an instrumental technique such as electrothermal atomic absorption spec- trometry (ETAAS) ,778 inductively coupled plasma atomic emission spectrometry9 or inductively coupled plasma mass spectrometry ,lo improves the sensitivity significantly.The determination of trace metals on board ship provides the opportunity to modify sampling strategies when the metal is a tracer (Mn in a study of a hydrothermal systemll) and also to identify and solve contamination problems. Of the various instrumental techniques, flame atomic absorption spec- trometry (FAAS) meets the difficult conditions that exist on board a ship." However, the measurement of the concentra- tion of trace metals requires a matrix that is as free as possible from alkali and alkaline earth elements in order to prevent interferences. Different solutions have been used depending on the chelating resin being employed. With the iminodiacet- ate ion exchanger Chelex, the selective elution of Ca and Mg is required; however, this step is thought to remove a certain amount of Mn.12 With more selective resins such as 8-hydroxyquinoline chemically bonded to silica, direct elution with a dilute acid is possible; however, high levels of interfering ions require the use of the standard additions method for the determination of Mn and Pb.5 More selective ligands should be able to reduce these problems significantly.The high selectivity of tetraaza macrocyclic compounds for transition metals over alkali and alkaline earth metal ions makes them potentially useful materials for providing a matrix as free as possible from interfering ions. However, such an application of an aza macrocycle requires a modification of the ligand in order to prevent the ligand itself or the metal complexes formed being soluble in water. The first polymers with tetraaza macrocycles were synthesized by reacting the macrocycle with poly- (chloromethyl)vinylbenzene ,13,14 leading to a cross-linked resin with swelling properties and complexation rates that were unacceptable for utilization ot a column for analytical purposes.15 Recently, the development of a new reaction for the monoalkylation of tetraaza macrocycles enabled a large number of substituted tetraaza macrocycles to be synthesized.In recent worki16 poly(vinylbenzyl)-l,4,8,11-tetraazacyclo- tetradecane was synthesized. Investigations of the properties of the resin showed that it was suitable for the preconcentra- tion of trace metals from sea-water with quantitative recovery of Mn (Mn is one of the most weakly complexed metals).The major problem tor utilizarion of this resin in a column is the back-extraction step which requires an unacceptably long period of time. Diffusion processes are usually the limiting factor in ion exchange, but rate control by complexation or decomplexation reactions can occur with some ligands. Moreover, improvement of the kinetic properties of the resin requires the investigation of two variables, viz., the size of the macrocycle and the nature of the support. Lipophilic derivatives, the synthesis of which can easily be achieved by using the monoalkylation technique described above, represent another very interesting way to obtain chelating resins by impregnating these derivatives on various organic sorbents. Owing to the highly cross-linked nature of the organic polymer matrices, high exchange rates and improvement of the complexation kinetics could be expected. In the present work, the macroporous resins XAD-4 and XAD-7 were used as adsorbents for various lipophilic tetraaza macrocycles.Experimental Apparatus A GBC SB 900 atomic absorption spectrometer was used for all metal measurements by FAAS. For measurements by ETAAS, a Varian AA 475 instrument equipped with a CRA 90 graphite furnace atomizer was used. Samples were injected manually with a 5 or 10 pl Eppendorf micropipette. Storage bottles, labware and tubing for the manifold and column were cleaned by soaking in a solution of hydrochloric acid overnight and then rinsing with high-purity water. Reagents All aqueous solutions were prepared with de-ionized water obtained from a Milli-ROMilli-Q apparatus (Millipore).816 ANALYST, AUGUST 1991, VOL.116 W I I \ S, R E lzl (3) (4) Fig. 1 Structures of the tetraaza macrocycles studied. (1) 1,4,7,10- tetraazacyclododecane (cyclen); (2) 1,4,8,11-tetraazacyclotetrade- cane (cyclam); (3) 1,4,8,12-tetraazacyclopentadecane; and (4) 1,5,9,13-tetraazacyclohexadecane The metal standard solutions and spikes were prepared from commercial standard solutions for AAS (lo00 mg 1-1) (Aldrich) (the nitrate salts are used for Cu", Cd", Mn", Ni" and Pb", and the chloride salt for Zn'I). Nitric and hydro- chloric acids, and ammonia solution were ultrapure (Merck). The macroporous resins XAD-4 and XAD-7 were ground in a ceramic mortar and sieved to obtain the 300400 pm fraction.The resins were then washed overnight with a solution of 4 mol dm-3 hydrochloric acid-methanol (1 + l ) , rinsed with water and dried under a laminar flow hood. Sea-water The coastal sea-water was sampled in hard polyethylene bottles and stored in similar bottles after filtration through a 0.45 pm Millipore filter using an all-polypropylene apparatus. The samples were then acidified to pH <2 with concentrated hydrochloric acid. The sea-water reference material NASS-2 (open ocean sea-water) was obtained from the Chemistry Division of the National Research Council of Canada. Synthesis of the Tetraaza Macrocyclic Derivatives (1)-(4) Compounds (l), (2) and (3) (see Fig. 1) are commercial products and were purchased from Parish, Aldrich and Strem Chemicals, respectively.Compound (4) (Fig. 1) was syn- thesized as described by Richman and Atkins.17 The mono N-alkylation of the different tetraaza macrocycles with dodecyl bromide was carried out by the method developed by Blain et al. 16 Preparation of the Impregnated Resin A 500 mg amount of resin was placed in a round-bottomed flask for XAD-7, or in a 1 1 beaker for XAD-4, and 10 ml of acetone containing the lipophilic tetraaza macrocycle (50 mg) were added. For XAD-7, the contents of the flask were heated under reflux for 15 min and the solvent was subsequently removed by evaporation in vacuu. For XAD-4, 800 ml of water were added dropwise to the organic solution under gentle agitation with a magnetic stirrer. The resins were filtered off, rinsed several times with water and stored in a 2 rnol dm-3 hydrochloric acid solution.Fig. 2 Schematic diagram of the manifold used for preconcentration of trace metals: S, sample; A, acid solution ( 2 rnol dm-3 HNO,); W, waste; V, three-way valve; P1 and P2, eight-channel peristaltic pumps; R, reagents (see text); and E, acidic effluent Before packing the resin into the column, it was washed successively with water, 2 mol dm-3 ammonia solution, 1 mol dm-3 KCN solution (see under Safety Considerations), several times with water, then with 4 mol dm-3 hydrochloric acid and finally with water. Preparation of the Column and Manifold The impregnated resin (400 mg) suspended in hydrochloric acid solution (4 rnol dm-3) was pumped into a poly- tetrafluoroethylene (PTFE) column (Ih in i.d.) fitted with glass wool.Two Gilson Minipuls eight-channel peristaltic pumps (Fig. 2) were used: the first pump was equipped with Tygon tubing (0.09 in i d . ) for aspiration of reagents and samples through the column, while the second was equipped with poly(viny1 chloride) (PVC) tubing (0.09 in i.d.) for elution of the acid solution. Flow lines were made of PTFE tubing (0.25 or 0.125 in i.d.). Samples and reagents were stored in PTFE tubes or bottles. The acidic effluent was collected in a PTFE tube. Safety Considerations In order to prevent accidents that can occur when using KCN solution for the decomplexation of the Cu", the following precautions were taken: (1) all the procedures using KCN solution were carried out under an aspirating hood; (2) the resin was systematically rinsed with water and 2 mol dm-3 ammonia solution before the treatment with the KCN solution; (3) water was systematically pumped through the column after the treatment with KCN; and (4) the KCN filtrate or column effluent was immediately destroyed by dilution with a basic solution (pH 8-9) and oxidation by H202.Capacity Measurements The capacity of the impregnated resin for Cu" was determined at pH 5.5. The impregnated resin (200 mg) was equilibrated by shaking it for 6 h with 40 ml of a buffered metal solution containing 0.25 g 1-1 of Cu". The resin was then filtered and rinsed with water. The filtrate and rinsing solution were collected and adjusted to 100 ml in a calibrated flask. The concentration of Cu" was determined, after dilution, by FAAS.Effect of pH on Batch Extraction The following buffered solutions were prepared: acetic acid-sodium acetate (pH 3-5); disodium hydrogen phos- phate-sodium dihydrogen phosphate (pH 5-7); and sodium borate-sodium hydroxide (pH 7-9). A 40 ml volume of aANALYST, AUGUST 1991, VOL. 116 817 buffered solution containing 1 mg 1-1 of a metal ion was shaken with an excess of impregnated resin for 6 h. The remainder of the procedure was as described under Capacity Measurements. Breakthrough Capacity Breakthrough studies were performed with Cult at pH 5.0 or with Mn" at pH 8.0. Solution concentrations were 1 mg 1-1. The metal ion solution was passed through the column at a specified flow rate of 1 ml min-1 with the peristaltic pump as described above. Fractions (5 ml) were collected in 10 ml tubes with an automatic collector (2112 Redirac, LSK) and then analysed by FAAS.Column Procedure Before using the column, the resin was rinsed with 20 ml of 1 mol dm-3 KCN solution (1 ml min-I) (see under Safety Considerations), 30 ml of water (1 ml min-I), 10 ml of 2 mol dm-3 HCI (0.1 ml min-1) and water (1 ml min-1). In the procedure, the flow rate during the aspiration of the reagents and samples through the column was fixed at 1 ml min-1. Firstly, 5 ml of water containing 0.2 ml of ammonia solution ( 5 % ) were passed through the column, which was sub- sequently rinsed with 5 ml of water. The sample, the pH of which was adjusted to 8.0 with a small amount of ammonia solution, was then passed through the column. The volume of ammonia solution to be added to the sample was determined on another aliquot of the sample in order to prevent contamination by the pH probe.Afterwards, the column was rinsed with 5 ml of water. The three-way valve was then turned to the elution position and 5 ml of a 2 mol dm-3 HN03 solution were passed through the column at a flow rate 0.1 ml min-1 and collected in a PTFE tube. The resin was washed with 5 ml of water and was then ready for a new preconcentra- tion step. Recoveries of Spikes From De-ionized Water and Sea-water The de-ionized water into which metal ions were spiked in order to obtain a concentration of 1 pg 1-1, was passed through the column, after adjustment of the pH with ammonia solution using the procedure described under Column Procedure. The volumes pumped were variable and were determined in order to obtain a significant signal during analysis by ETAAS.For each determination a blank was obtained by processing the same volume of de-ionized water without any spike. The recoveries of spikes from sea-water were determined with coastal sea-water. The sea-water was spiked with a metal ion in order to obtain a concentration two or three times the supposed initial concentration based on a preliminary deter- mination. The procedures for complexation-extraction and analysis of the blank were the same as described above. The concentrations in the extracts were determined either by direct calibration with standard solutions in 2 mol dm-3 nitric acid or by the standard additions method. Both methods show good agreement and either can be used to calculate the efficiency of the recovery of the spike.Results and Discussion Resin Properties Impregnation and batch capacity of the resin Two impregnation supports were tested: XAD-4 and XAD-7. Muller et al. 18 have previously described a technique for the impregnation of a tetraaza macrocycle derivative on XAD-7 and have studied the distribution of the ligand between the aqueous phase and the macroporous resin. The distribution is very favourable with respect to the solid phase, the (ligand),,lid to (ligand)so~ut~on ratio being 104. In this work, the amount of tetraaza macrocycle adsorbed on polymers (XAD-4 or XAD-7) during the impregnation step represented 92% of the total amount introduced. No significant bleeding of the ligand from the column was observed and the resin capacity was not affected after several months of continuous use.The results discussed below were obtained with compound (3), which was the ligand chosen for all subsequent studies (see under Influence of the Size of the Macrocycle Cavity). The resin capacity, determined by the batch method, is 340 ymol g-1 for Cull (pH 5.5) and 38 pmol g-1 for Mn" (pH 8.0). Muller et a1.18 have demonstrated that the thermodynamic equilibrium of the complexation of various metal ions (e.g., Zn", Co" and Ni") was not completely achieved even after several days. Hence the different behaviour of the two metals (CU" and Mn") can be explained by a kinetic effect. The equilibrium is rapidly attained with Cu; however, with Mn, total complexation is not achieved even after 6 h.p H dependence Fig. 3 shows the pH dependence of the recovery of different metals. Tetraaza macrocycles are tetrabases and protonation of the nitrogen atoms and complexation of metal ions are competing reactions. The pK, values of the different macro- cycles studied here have been reported in the literature,I9 but data for complexation constants with transition metals are incomplete and, in some instances, e.g., Mn, not available. This makes complete interpretation very difficult. However, the curves shown in Fig. 3 permit the experimental determina- tion of the optimum pH for the total complexation of Cd, Mn, Pb and Zn. The optimum pH is close to 8. Breakthrough curves In the utilization of a column the breakthrough capacity is an important parameter.It depends on a large number of factors and in particular on the ion concentration. In order to obtain the breakthrough curves, an ion concentration significantly higher than the concentrations encountered in practice during the analysis of sea-water was selected. The studies were carried out at pH 5.5 for Cu and at pH 8.0 for Mn, with a column loaded with 400 mg of chelating resin. The break- through curves are shown in Fig. 4. The ratio of the capacity of 60 l I l - I I 4- 7 - 100 8 - a < 60 P 3 c.' 20 I I I I 1 3 5 7 1 3 5 7 PH Fig. 3 pH dependence of the uptake of metals by XAD-4 impreg- nated with a tetraaza macrocycle in a batch system. (a) Cu; (b) Mn; ( c ) Cd; (d) Zn; ( e ) Pb; and (f) Ca818 .- Y C 0.8 0.6 0, c 8 ANALYST, AUGUST 1991, VOL.116 f: - 'd I t r-- 0 4 1 5 10 " 30 Metal ion/ymol Fig. 4 Breakthrough curves. Metal ion concentration, 1 mg 1-1; flow rate, 1 ml min-l. A, Mn containing 0.5 rnol dm-3 Na, 0.01 rnol dm-3 Ca and 0.05 rnol dm-3 Mg; B, Mn; and C, Cu. Relative concentration is the fraction of metal detected from the total amount loaded Table 1 Effect of the size of the cavity and of the support used for impregnation Recovery (%) c u Mn Macrocycle XAD-4 XAD-7 XAD-4 XAD-7 52 55 38 34 50 56 48 100 96 80 - 41 77 - (1) (2) (3) (4) - - 0.8 c 0 .- 4- 0.6 a, C 0 p 0.4 .- Y m a, - a 0.2 0 2 4 6 Volume of effluent/ml Fig. 5 Elution curves for Mn" eluted from the column using 0.2 mol dm-3 HCI. Flow rates: A, 1; and B, 0.1 ml min-1; and C, back-flushing with a flow rate of 0.1 ml min-l.Relative concentration is the fraction of metal detected from the total amount loaded the column for Cu to that for Mn is of the same order of magnitude under both batch and flow conditions. The introduction of Ca", Mg" and Nal at respective concentrations of 0.01, 0.05 and 0.5 rnol dm-3 into the metal ion solution, in order to simulate a sea-water matrix, decreases the break- through capacity of the column for Mn from 3 to 1 pmol. This capacity is fairly small but, owing to the extremely low levels of trace metals in sea-water, particularly in the open ocean, this should not be a problem. Column Preconcentration Procedure Copper and manganese were chosen for investigations into the effect of the different parameters that control the extraction and back-extraction steps.These two metals represent the extreme behaviour of transition metals towards tetraaza 0.2 111 b \ 0 5 10 Volume of effluent/ml Fig. 6 Elution curves of various trace elements eluted from the column using back-flushing with 2 rnol dm-3 HN03 (flow rate, 0.1 ml min-1). Relative concentration is the fraction of metal detected from the total amount loaded macrocycles. Copper is the most strongly and rapidly com- plexed metal,20 whereas Mn is probably one of the most weakly and slowly complexed metals. Influence of the size of the macrocycle cavity Only a few papers have described kinetic studies of the acidic decomplexation of aza macrocycles21 and there is no known relationship between the size of the macrocycle and the rate of complexation or decomplexation. In the present work four tetraaza macrocycles (Fig.1) were investigated using the experimental conditions described under Column Procedure. The results are shown in Table 1. For Cu", the decomplex- ation rate constants are reported only for reaction in an alkaline medium.22 However, they showed that the decom- plexation of the complex of Cut' with the macrocycle (4) is l o 4 times faster than for the other macrocycles studied [com- pounds (1) and (2)]. The experimental results for the recovery of Cu" could also probably be explained by the significant differences between the decomplexation rates of the different macro-cycles, but in all instances the recovery is not quantita- tive even with the largest macrocycle (4). For Mn", only the macrocycle (3) leads to a quantitative recovery.Hence this ligand was used for all subsequent studies. Influence of the support used for impregnation Parrish23 has reported that the resin with the highest water regain, viz., XAD-7, shows the highest rate of uptake of metal ions from an aqueous solution. More recently, Isshiki et aZ.3 have reported that the distribution between the resin and the aqueous phase is also an important factor in the extracting efficiency of the resin. In order to determine the effect of the sorbent, XAD-4 and XAD-7 were studied. No significant differences between these supports were observed. However, the XAD-4 resin was chosen because XAD-7 is a polyacrylate resin containing oxygen atoms. These atoms are hard bases according to Pearson's classification24 and prefer to bind hard acids such as alkali or alkaline earth elements.The selectivity might therefore be less satisfactory with XAD-7. Back-extraction step The elution curves for Mn, under different conditions, are given in Fig. 5. Quantitative recovery in 5 ml of an acid solution (0.2 mol dm-3 HC1) was obtained by back-flushing the column at a nominal flow rate of 0.1 ml min-1. Under these conditions only 30% of the sequestered Cu" is eluted from the column. Different acid solutions of various concen-ANALYST, AUGUST 1991, VOL. 116 819 Table 2 Recovery of trace metals from de-ionized water and sea-water Recovery (YO)* Table 5 Concentration of major elements in the acidic eluate Concentration of metaYmg 1-1 Metal De-ionized water Sea-water Cd 99 f 5 94 * 5 c u 61 f 4 NDt Mn 91 f 14 104f 12 Ni 55 f 12 NDt Pb 91 k 12 93 k 14 Zn 110 f 6 102 f 10 * Mean and precision expressed as 95% tolerance limit (five t Not determined.determinations). Table 3 Absolute blank and detection limit Limit of detection*/ Metal Absolute blankhg ng Cd 0.9 f 0 . l t 0.75 Mn 5 f 0.5 3.7 Pb 7 r t l 7.5 Zn 8 + 2 15 * For 400 ml of sample. t Standard deviation of ten determinations. Table 4 Analysis of NASS-2 Concentratiodpg 1-l Metal This work Certified value* Cd 0.028 f 0.003t 0.029 f 0.004 c u 0.102 k 0.023t 0.109 k 0.01 1 Mn 0.027 k 0.0061- 0.022 f 0.007 Pb 0.035 f 0.007$ 0.039 f 0.006 Zn 0.183 k 0.010$ 0.178 f 0.025 * Precision expressed as 95% tolerance limit. t Mean and precision expressed as 95% tolerance limit (four determinations).$ Mean and recision and tolerance as 95% tolerance limit (three determinations!. trations were also tested in order to improve the recovery of Cu. With the highest concentration (2 mol dm-3) compatible with analysis by ETAAS, the recovery was less than 70% when using a hydrochloric acid solution and less than 60% when using a nitric acid solution. In order to prevent any possible interference during analysis by AAS, 2 mol dm-3 nitric acid was chosen for elution. Typical elution curves are shown in Fig. 6. As all of the sequestered Cu" is not eluted under these conditions, it is necessary to rinse the column periodically (i.e., every 20 samples) with 5 ml of 1 mol dm-3 KCN solution and water to regenerate the resin completely. Recovery of spikes from de-ionized water and sea-water The efficiency of the recovery of spikes from de-ionized water and sea-water is shown in Table 2. Satisfactory results were obtained for Cd, Mn, Pb and Zn with an average recovery of 98 k 8% in de-ionized water and 98 k 5% in sea-water.The recovery was not quantitative for either Cu" or Ni" as discussed previously. In the absence of comprehensive theoretical data, it is not clear why such behaviour is observed but it is thought that, because the kinetic decomplexation is very slow, the metals are trapped in the resin during the swelling that occurs in an acidic medium. Blank and limit of detection The major problem encountered during the minimization of the analytical blank was the contamination from the tubing used for the peristaltic pumps.Basic and acidic solutions rapidly degrade such tubing, particularly Tygon, leading to a marked increase in the absolute blank. This problem was partly solved by the aspiration of the sample and reagent through the column. Hence, the solutions were only in contact with PTFE before passing through the column. Only acidic Resin Na Ca Mg Chelex* 320 350 63 Chelext 150 1800 1700 ChelexS 87 480 480 ChelexQ 195 300 330 8-Hydroxyquinoline on Si0,fl 17 13 65 Tetraaza macrocycle on XAD-411 45 3 9 * Reference 1; volume of sea-water, 100 ml; volume of acid, 7 ml. 7 Reference 25; volume of sea-water. 4 1; volume of acid, 30 ml. S Reference 26; volume of sea-water, 200 ml; volume of acid, 10 ml. § Reference 26; volume of sea-water, 1 1; volume of acid, 10 ml. Reference 27; volume of sea-water, 900 ml; volume of acid, 10 ml.This work; volume of sea-water, 400 ml; volume of acid, 5 ml. solutions were pumped through tubing which was not made from PTFE. Tubing made from PVC gives satisfactory results, as shown in Table 3. The absolute blanks are generally higher than those reported for 8-hydroxyquinoline on silica,6 but the limit of detection (Table 3), defined as three times the standard deviation of the blank, permitted the determination of Cd, Mn, Pb and Zn in open ocean waters. Accuracy and precision The accuracy of the proposed method was evaluated by the determination of Cd, Cu, Mn, Pb and Zn in the reference sea-water NASS-2. Aliquots of NASS-2 were preconcentrated 80-fold. The results, presented in Table 4, are statistically indistinguishable from the certificate values for Cd, Mn, Pb and Zn.The precision of the determination, expressed as the relative standard deviation, is better than 15%. The determi- nation of Cu in NASS-2 confirmed the low recovery of this metal; however, a satisfactory determination of the concentra- tion was obtained by using the recovery of 61% previously determined for de-ionized water. Selectivity of the resin One of the aims of this work was to improve the selectivity of the resins used for the preconcentration of trace metals from sea-water. Hence the concentrations of the major cations present in sea-water (Na', Call and Mg") were determined in the acidic eluate. The results are summarized in Table 5. The improvement in the selectivity is evident for Ca" and Mg"; however, the concentration of Na' is higher than expected. The reason for this is not known, but it is assumed that the XAD-4 is probably responsible.Mackey28 has reported that similar styrene-divinylbenzene copolymers, used as stationary phases in gas chromatography, have been found to contain small numbers of polar impurity sites and that carboxyl groups are probably responsible for the weak cation-exchange capacity of XAD-4. Experimentally, this assumption appears to be confirmed by the fact that when sea-water is passed through the impregnated column at an acidic pH (no cation is complexed by the macrocycle), the amount of Nal sequestered by the resin is similar to the value obtained using the normal procedure. Conclusion This work has demonstrated that 1,4,8,12-tetraazacyclopen- tadecane (3) is the most suitable tetraaza macrocyclic ligand for obtaining a resin with satisfactory properties for analytical applications.Good results were obtained for the determina- tion of Mn", Cd", Zn" and Pb" in sea-water. The selectivity of the tetraaza macrocycles towards transition metals gives a matrix with very low levels of alkali and alkaline earth elements.820 In order to improve the recoveries of Cu" and Ni.", new supports with a greater hydrophilic character are currently being developed. 1 2 3 4 5 6 7 8 9 10 11 12 13 References Kingston, H. M., Barnes, I. L., Brady, T. D., Rains, T. C., and Champ, M. A., Anal. Chem., 1978,50, 2064. Paulson, A. J., Anal. Chem., 1986,58, 183. Isshiki, K., Tsuji, F., Kuwamoto, T., and Nakayama, E., Anal.Chem., 1987,59,2491. Willie, S . N., Sturgeon, R. E., and Berman, S. S., Anal. Chim. Acta, 1983, 149, 59. Nakashima, S., Sturgeon, R. E., Willie, S. N., and Berman, S. S., Fresenius 2. Anal. Chem., 1988, 330, 592. Sturgeon, R. E., and Berman, S. S., Ann. Chim., 1988, 78, 1. Marshall, M. A., and Mottola, H. A., Anal. Chem., 1985,57, 729. Hirata, S., Honda, K., and Kumamaru, T., Anal. Chim. Acta, 1989, 221, 65. Kumamaru, T., Matsuo, H., Okamoto, Y., and Ikeda, M., Anal. Chim. Acta, 1986, 181,271. Beauchemin, D., and Berman, S. S., Anal. Chem., 1989, 61, 1857. Klinkhammer, G., Rona, P., Greaves, M., and Elderfield, H., Nature (London), 1985,314,727. Strachan, D. M., Tymochowicz, S., Schubert, P., and Kingston, H. M., Anal. Chim. Acta, 1989, 220, 243. Louvet, V., Handel, H., Appriou, P., and Guglielmetti, R., Eur. Polym. J . , 1987, 23, 585. 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 ANALYST, AUGUST 1991, VOL. 116 Szczepaniak, W., and Kuczynski, K., React. Polym., 1985, 3, 101. Percelay, L., Appriou, P., Handel, H., and Guglielmetti, R., Anal. Chim. Acta, 1988,209,249. Blain, S . , Appriou, P., Chaumeil, H., and Handel, H., Anal. Chim. Acta, 1990, 232,331. Richman, J. E., and Atkins, T. J., J. Am. Chem. Soc., 1974,96, 2268. Muller, F., Handel, H., and Guglielmetti, R., Helv. Chim. Acta, 1983, 66, 1525. Izah, R. M., Bradshaw, J. S., Nielsen, S. A., Lamb, J. D., and Christensen, J. J., Chem. Rev., 1985, 85, 271. Leugger, A. P., Herti, L., and Kaden, T. A., Helv. Chim. Acta, 1978,61,2296. Chen, L.-H., and Chung, C.-S., Znorg. Chem., 1988,27, 1880. Tabushi, I., and Fujiyoshi, M., Tetrahedron Lett., 1978, 25, 2157. Parrish, J. R., Anal. Chem., 1977,49, 1189. Pearson, R., J. Chem. Educ., 1968,45, 581. Bruland, K. W., Franks, R. P., Knauer, G. A., and Martin, J. H., Anal. Chim. Acta, 1979, 105,233. Pai, S., Whung, P., and Lai, R., Anal. Chim. Acta, 1988, 211, 257. Sturgeon, R. E., Berman, S. S., Willie, S. N., andDesaulniers, J. A. H., Anal. Chem., 1981, 53, 2337. Mackey, D. J., Mar. Chem., 1982, 11, 169. Paper Ol05239F Received November 21st, 1990 Accepted March 15th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600815

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, AUGUST 1991, VOL. 116 821 Determination of Trace Amounts of Tellurium by the Sub-superequivalence Method of Isotope Dilution Analysis Hiroe Yoshioka, Yoshinori Miyaki and Kunihiko Hasegawa Radiochemistry Research Laboratory, Faculty of Science, Shizuoka University, 836 Oh ya, Shizuoka 422, Japan Trace amounts of tellurium in synthetic samples were determined by the redox sub-superequivalence method of isotope dilution analysis (SSE-IDA) and the results were compared with those of sub- stoichiometric isotope dilution analysis (Subst-IDA) in order to demonstrate the usefulness of SSE-IDA. 125mTelV was separated from an HCI solution in which 125Sb and 125mTetv were in radioactive equilibrium and 125mTe was used as a tracer. The reactant (Tetv) was oxidized sub-stoichiometrically with a solution of K2Cr207 in 0.1 rnol dm-3 HCI, followed by the extraction of TeIv with 20% tributyl phosphatekerosene solution (after changing the concentration of HCI to 5 rnol dm-3) in order to separate TetV and Te"'. Activities of TeV1 in each aqueous phase were measured and the amount of tellurium was determined according to the graphical method of SSE or Subst-IDA. Under the conditions where Subst-IDA gives large errors, accurate determinations were possible using SSE-IDA with errors of less than 5-6 per cent.The limit of the determination was about 2 pg mi-'. Keywords: Redox sub-superequivalence method of isotope dilution analysis; tellurium- 125m; tellurium(lv) determination; tellurium(lv) and tellurium(v~) separation; extraction with tributyl phosphate-kerosene Sub-stoichiometryl-3 has been used widely in analytical chemistry. Compared with other methods such as gravimetric and volumetric analysis, it is not subject to interference from other concomitant substances.Since sub-stoichiometry was introduced in isotope dilution analysis, it has made much progress. In this laboratory, sub-stoichiometric isotope dilu- tion analysis (Subst-IDA) has been applied to the determina- tion of Sb in metals such as Zn, As, Sn and in a commercial 124Sb sample.4 One problem with the method is that the following strict conditions must be applied: the reagent must react completely with the analyte and be separated quantita- tively. If these experimental conditions are not observed, large errors occur. Reagents possessing such a property are rather few in number.Klas et a1.5 proposed a method termed the 'sub-super- equivalence method of isotope dilution analysis (SSE-IDA)', which does not require such strict conditions but which gives an accurate determination. In addition, it has the advantage of sub-stoichiometry mentioned above. This method was applied to the determination of Sbn1,6 T11,7 and deoxyribonucleic acid8 and the advantages were verified by experiment. In the work described in this paper, TeIV is determined by this method, as tellurium is widely used in the semiconductor industry and is becoming an environmental pollutant. Experimental Sample and Reagents 125mTe'" tracer solution. Prepared by separating 1*5mTerv from an HCI solution of 125Sb and 125mTe in radioactive equilibrium using an ion-exchange column.A 1.0 ml (about 0.3 g) portion of anion-exchange resin (CI- form) in concentrated HCI was packed in a column (0.6 cm i.d.) and then washed with concentrated HCl solution. A 0.1 ml aliquot of a solution containing 125Sb and 125mTe in 9 rnol dm-3 HCl and bromine water, 2-3% m/v (19 + l ) , was added at the top of the resin. Next, 3 ml of a solution of 9 rnol dm-3 HCI and bromine water (19 + 1) was added as eluent in order to elute 125mTev1, followed by 1 rnol dm-3 HCI solution in order to elute 125mTerv. The resulting contamination from 125Sb was 0.5%. Tel" and Te"' carrier solutions. Prepared by dissolving an accurately weighed amount of potassium tellurite (K2Te03) and potassium tellurate trihydrate (K2Te04.3H20) in 6 rnol dm-3 HCl and diluting to the desired concentration.Synthetic sample solution. A solution of TetV in 2 mol dm-3 HCI containing 125mTerv as a tracer was used. Tevl (with 125mTev') solution. Prepared by the oxidation of the synthetic sample solution with an excess of K2Cr207 solution and used for extraction, etc. Oxidizing reagent. Prepared by dissolving an accurately weighed amount of K2Cr207 (99.98%) and KBr03 in distilled, de-ionized water to the required concentration. Separating reagent. A 20% solution of tributyl phos- phate(TBP) in kerosene was used for the separation of TeIV from Te"I.9 All the solutions were prepared using chemicals of ana- lytical-reagent grade. Activity Measurement Activities of 125mTe (27.4 keV, K X-ray and 35 keV, y rays) were measured using an NaI(T1) well-type scintillation counter .Procedure Two series of solutions were prepared. First series Aliquots (0.1 ml) of the synthetic sample solution labelled with a radioisotope [TeIV (125mTelv) = 44.1 pg ml-1 in 0.2 rnol dm-3 HCI] were placed in six polypropylene micro- test-tubes equipped with stoppers. Then, 0.1 ml aliquots of various Te'" carrier solutions (0, 44.1, 88.2, 132.3, 176.3 and 219.7 pg ml-1 in 0.2 rnol dm-3 HCI) were added. Second series Pairs of 0.2, 0.3 and 0.4 ml aliquots of the synthetic sample solution used in the first series were placed in the test-tubes. It is not essential to repeat the analysis for the second series, but doing so will give a more reliable average value. No carrier solution was added. All the aliquots in both series were adjusted to the same volume and acid concentration by the addition of dilute HCl and then homogenized by shaking using a micro-test-tube shaker.The drops adhering to the stoppers and to the walls of the tube were collected by centrifugation, because they cause large errors in a small-scale experiment.822 ANALYST, AUGUST 1991, VOL. 116 Oxidation and separation The same operations using micro-tubes were carried out. A 0.1 ml aliquot of 3.44 X 10-4 mol dm-3 (0.34 pequiv ml-1) K2Cr207 solution was added to each sample aliquot. The total volume was adjusted to 0.9 ml, the HC1 concentration being 0.1 mol dm-3 at this stage of the oxidation. After it had been left to stand for 60 min, the acid concentration of the solution was altered to 1 rnol dm-3 by the addition of 0.1 ml of 10 mol dm-3 HCI, ready for TBP extraction.Then, 1.0 ml of a 20% TBP-kerosene solution was added to each micro-test- tube and the unreacted TeIV was extracted from the mixed solution of TeIV and TeV1. The radioactivity of TeV1, oxidized sub-stoichiometrically in each aqueous phase (0.8 ml), was then measured with an NaI(TI) well-type scintillation counter. Results and Discussion Effect of HCI Concentration on the Extraction Ratio of TeW and TeM The extraction of TeIV and TeV1 (mg ml-1 concentrations) from HCl solution with 20% TBP-kerosene has been reported previously.9 In the present work, the effect of the concentra- tion of HCI on the extraction ratio of TeIV and TeV1 (pg ml-1 concentrations) was examined. The results are shown in Fig.1. Tellurium(1V) was extracted satisfactorily (95%) using 5 mol dm-3 HCl but TeV1 was not extracted in the measured acid concentration range (0.1-5.0 mol dm-3 HCl). The same operation was carried out using trace amounts of 125mTerv and 125mTev1 and the extraction behaviour was almost the same as when micro-amounts were used. Fig. 1 shows that Tew can be separated quantitatively from TeIV at an acid concentration >5 mol dm-3. Effect of HCI Concentration on the Oxidation Ratio Two oxidants, K2Cr207 and KBr03, which have redox potentials far higher than that of TeIV, were compared. They are generally known to be strong oxidants and are chemically stable. (An oxidation ratio of 100% means, hereafter, that the oxidant is completely consumed.) Under the conditions shown in Fig.2, TeIV was only slightly oxidized by KBr03 over a wide concentration range of HCl. However, using K2Cr207 the oxidation curves had a maximum value at [HCl] = 0.1 rnol dm-3. Therefore, K2Cr207 was used as a suitable oxidant for TeIV and the oxidation was carried out in 0.1 mol dm-3 HCl . Effect of Tew Concentration at the Stage of Oxidation on the Oxidation Ratio The equivalence ratio of TeIV: K2Cr207 was kept constant at 1 : 0.5. The effect of the TeIV concentration at the oxidation 100 I I J 2o t / B A h h h U U u 0 1 2 3 4 5 Concentration of HCl/mol dm-3 Fig. 1 Effect of [HCl],, on the extraction ratio of Telv and TeV1. A, Vol,, and volorg, volumes of the aqueous and the organic phase used in the extraction stage, are both 0.6 ml.Extraction time, 30 s ex = 8.5 pg ml-I; and B, [TeV'],, = 86.4 pg ml-1. [HCl],, and ,,, concentration of HCl and TeIV used in the extraction stage. stage, [Telv],x, on the oxidation ratio was examined over a wide concentration range (1 X 10-3-1 X 102 pg ml-1) and the result is shown in Fig. 3. The oxidation ratio is very dependent on [TelvIox. For the quantitative oxidation of TeIV, more than 100 pg ml-1 of TeIV were necessary. When 10 and 5 pg ml-* of TeIV were used, the ratios were 60 and 30%, respectively. Below 0.1 pg ml-1, the oxidation reaction hardly proceeded. Effect of the Amount of Te'" on the Oxidation Ratio The effect of the amounts of TetV used on the oxidation ratio was examined in detail and the results were compared with the theoretical value (Fig.4). The difference between the experimental and the theoretical values is largest near the equivalence point, and the experimental value approaches the theoretical with increasing amounts of TeIV. The determina- tion of TeiV by SSE- and Subst-IDA was carried out in this slowly increasing region, where the oxidation reaction did not proceed quantitatively. Effect of Oxidation Time on the Oxidation Ratio The relationship between the oxidation ratio and oxidation time was examined under the conditions shown in Fig. 5, where a value of 100% on the ordinate means that K2Cr207 was completely consumed. When the concentration of TeIV was reduced, the ratio and rate of oxidation were also reduced. The time taken to attain equilibrium at low concentrations of TeIV was too long, so an oxidation time of 60 min was adopted for comparison with Subst-IDA. 20 D A A 0 1 2 3 Concentration of HCVmol dm-3 Fig.2 Effect of [HCl],, on the oxidation ratio using K2Cr207 and KBr03. A, [Tevl],, = 117.5 pg ml-1, [K2Cr207] = 1.8 pequiv ml-1, vol,, = 0.6 ml; B, [Telv],, = 10.9 pg ml-1, [K2Cr207] = 0.16 pequiv ml-1, VO~,, = 0.6 ml; C, [Telv],, = 8.7 pg ml-1, [K2Cr207] = 0.06 pe uiv ml-1, volpx = 1.3 ml; and D, [TeiV],, = 8.7 pg ml-1, [KBrO37 = 0.06' pequiv ml-l, vol,, = 1.3 ml. Vol,, and vole,, the volumes of aqueous phase used in the oxidation and extraction stages are both 0.8 ml. Oxidation time, 60 min; extraction time, 30 s; and [HCl],, = 5.0 mol dm-3 100 80 A &? c 60 0 x 40 0 Y .- I-' 20 0 10-3 10-2 10-1 100 101 102 103 [Telvl,,/pg ml- Fig.3 Effect of [Telv],, on the oxidation ~atio.[Te~~],, : [K2CrzO7] = 1:O.S (pequiv:pequiv) [HCl],, = 0.1 rnol dm-3, volox = 0.8 ml, oxidation time, 60 s. [HCI],, = 5 mol dm-3, vole, = 0.8 ml, volOrg = 0.9 ml, extraction time, 30 s.ANALYST, AUGUST 1991, VOL. 116 823 1 0 5 10 15 20 TeVVpg Fig. 4 Effect of the amount of TeIv on TeV' oxidized: 1, theoretical curve; 2, experimental curve. K2Cr207 = 0.08 pequiv (rTeIV = 5.08 pg); [K2Cr207] = 0.06 pequiv ml-1 (=TeIV = 4.06 pg ml-l); [HCl],, = 0.1 rnol dm-3, vol,, = 1.25 ml, oxidation time, 60 min; [HCl],, = 5 rnol dm-3, vole, = 1.1 ml, volOrg = 1.1 ml; and extraction time, 30 s 100 n V A 80 s c 60 x 6 40 - .- 4- 20 0 20 40 60 80 100 120 Time/min Fig. 5 Effect of the oxidation time on the oxidation ratio. A, [TelvIox = 117.5 pg ml-I; B, Telv ox = 17.3 pg ml-I: and C, Telv ,, = 2.0 pg ml-I.[K2Cr207] : [Telv/ox = 1 : 1 (pequiv: pequiv), [HCl],, = 0.1 rnol dm-3, VO~,, = 0.6 ml, HCI],, = 5 mol dm-3, vole, = 0.7 ml, volOrg = 0.7 ml, and extraction time, 30 s Optimum Amount of Reagent for the Determination by For determination by SSE-IDA, an optimum amount of the complex-forming reagent was calculated theoretically by Klas,lo but the validity of the theory was not demonstrated experimentally. Therefore, the effect of the amount ( R pequiv) of K2Cr207 on the unknown amount of sample (x pequiv) was examined. The result is shown in Fig. 6. These values were determined using eqn. (6) in reference 6, viz., x = j y / ( k - 1). The quantity x is obtained graphically by plotting ukx/ax+iy against iy and finding the abcissa value (iy) corre- sponding to ukx/ux + iy = k for any value of k ; iy is the amount of the carrier (TerV) added to each aliquot of sample in the first series (see Table 1).The experimental graphs (B, C, D, E) approach the theoretical line (A) for a decreasing R : x ratio i.e., [K2Cr2O7Iox : [Te'"],,, which agrees with Klas's theory. According to this theory, the smaller the R : x ratio, the closer the experimental graph comes to the theoretical line, resulting in a small error for the analysis. In practice, however, too little reagent caused a large error as shown in curve B, because of the large counting error due to the small activity. When the activity is labelled sufficiently, a satisfactory result can be obtained. In this experiment, the optimum amount of the reagent gave an R : x ratio of 0.5 : 1.SSE-IDA Comparison of SSE-IDA With Subst-IDA Synthetic samples were analysed by SSE-IDA under con- ditions where Subst-IDA does not give satisfactory results. 0 2~ 4~ 6~ 8 y Effect of amount of K2Cr207 on the value determined i v ( = 1.4 x ipg, i = 0-8) Fig. 6 K2Cr207: K2Cr207 Curve" TeIV takerdpequiv Found/pg Error (%) B 0.2 0.012 4.2 6.3 C 1.0 0.062 4.1 3.8 D 1.5 0.093 4.2 13.9 E 25.0 1.56 - - * A, theoretical curve. TelV taken = 3.95 pg, [Telv],, = 8.78 pg ml-l, [HCI],, = 0.1 rnol dm-3, VO~,, = 0.45 ml, oxidation time, 60 min, [HCI],, = 5 rnol dm-3, vole, = 0.7 ml, volOrg = 0.7 ml, and extraction time, 30 s Table 1 Data from sample No. 4 in Table 2. The results using Subst-IDA were calculated using eqn.(7), k = 1, in reference 6. For SSE-IDA, eqn. (6) in reference 6 was used, i = 0,1,2, . . . i and k = 1, 2, 3, 4. Conditions as for Fig. 7 [TelVlox /pg ml-l 4.4 9.8 14.6 19.6 24.4 29.3 Activity of separated TeV Amount of TelV/pg counts min-1 1st series 2nd series x + iy kx 4.4 - 8.8 8.8 13.2 13.2 17.6 17.6 22.0 - 26.4 - 1st series 2nd series ax + iy a X Y 200 - 146 293 115 352 102 400 - 92 88 - * Background, 30 k 0.27 counts min-1; and counting error is <1%. 0 1 Y 2Y 3Y 4Y 5 Y i y (= 4.4 x i pg, i = &6) Fig. 7 Determination of TeIV by SSE-IDA and Subst-IDA. TetV taken = 4.4 pg, K2Cr207 = 0.034 pequiv, YCl],, = 0.1 rnol dm-3, vol,, = 0.9 ml, oxidation time, 60 min, [HCI ex = 5 rnol dm-3, vole, = 0.1 ml, volorP = 1.0 ml, and extraction time, 30 s824 ANALYST, AUGUST 1991, VOL.116 Table 2 Comparison of the values found by the two methods Subst-IDA SSE-IDA Amount of k = l k = 2 k = 3 k = 4 TeIV Sample [TelV],x/ [KzCrz07]: present/ Found/ Error Found/ Error Found/ Error Found/ Error No. pgml-1 [TeIV] Pg (%) Pg (Yo) Pg (Yo) Pg (Yo) 1 100.8 0.9 70.5 >88 (>25) 70.1 (-0.6) - - - - 2 47.2 1 .o 52.9 >80 (>50) 49.5 (-6.4) - - - - 3 6.3 0.5 4.4 >13 (>200) 4.2 (-4.5) - - - - 4 4.9 0.5 4.4 >12 (>170) 4.3 (-2.3) 4.2 (-4.5) 4.6 (2.7) 5 2.2 0.5 1.1 >10 (>800) 1.4 (27.2) 1.7 (54.5) 1.9 (72.7) Table 1 shows the results obtained by the procedure men- tioned above. The symbols are the same as those used in reference 6; each aliquot of sample in the second series contains k times the amount of sample and radioisotope used in the first series.First series solutions were used for the determination by Subst-IDA, and for SSE-IDA, solutions from both the first and second series were used. The radioactivity of TeIV in the first series progressively decreased because these solutions were diluted with a carrier solution. On the other hand, the aliquots in the second series were not diluted with a carrier, so the radioactivities could be the same. However, it is obvious that the values increased with increasing values of k because the amount of the product, TeIV, depended on the amount of the reactant, TetV. The graphs in Fig. 7 were drawn using the data from Table 1. The results of the determination using these methods are shown in Table 2.As expected, a large error was introduced into the results using Subst-IDA, whereas SSE-IDA gave an accurate determination within the expected error limit (except for sample 5). Accordingly, these results confirmed that accurate determination was possible using SSE-IDA without observing the strict conditions regarding completion of the reaction and the separation that is required for Subst-IDA. Conclusion In order to fulfil the strict conditions required for Subst-IDA, more than 99% of the reagent must be consumed, i.e., the formation constant of the product must be sufficiently large. The SSE-IDA method did not require such strict conditions, and the determination of TeIV by SSE-IDA with an error of 5-6 per cent was possible. The range of determination using SSE-IDA was wider than that of Subst-IDA.Determination in the low concentration range is generally difficult, because the reaction rate becomes slower and more time is needed to reach equilibrium. However, determination using SSE-IDA was possible under non-equilibrium conditions. Adsorption of the ions on the walls of the container becomes serious in the low concentration range, but it does not affect the results provided that the amounts adsorbed are the same in all the aliquots. The optimum amount of the reagent required for the determination using SSE-IDA agreed with Klas's theory. The SSE method may be applicable not only to isotope dilution analysis and radiometric analysis" but also to analysis using radioactivity including neutron-activation analysis and iso- tope-exchange analysis. 1 2 3 4 5 6 7 8 9 10 11 References Suzuki, N., in Proceedings of the 2nd Conference on Radioiso- topes (Japan), Japan Atomic Industrial Forum, INC, Tokyo, 1958, p. 151. Zimakov, I. E., and Rozhavskii, G. S., Tr. Kom. Anal. Khim. Akad. Nauk SSSR, 1958,9,231. RGiiCka, J., and Stary, J., Talanta, 1961, 8, 228. Kambara, T., Suzuki, J., Yoshioka, H., and Watanabe, Y., J. Radioanal. Chem., 1980, 60, 121. Klas, J., Tolgyessy, J., and Klehr, E. H., Radiochem. Radio- anal. Lett., 1974, 18, 83. Yoshioka, H., and Kambara, T., Talanta, 1984,31,509. Yoshioka, H., and Hasegawa, K., Analyst, 1987, 112, 855. Yoshioka, H., Hisada, T., Yoshinaga, K., Hasegawa, K., and Ikeda, N., Radioisotopes, 1987, 36, 389. Inarida, M., J. Chem. SOC. Jpn. Pure Chem. Sect., 1968,79,696. Klas, J., Radiochem. Radioanal. Lett., 1978, 33, 31, Yoshioka, H., Hasegawa, K., and Kambara, T., J. Radioanal. Nucl. Chem.. 1987, 117,47. Paper 0104183A Received September 13th, 1990 Accepted March 22nd, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600821

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, AUGUST 1991, VOL. 116 825 Separation and Determination of Trace Amounts of Zinc and Cadmium by On-line Enrichment in Flow Injection Flame Atomic Absorption Spectrometry Rajesh Purohit and Surekha Devi* Department of Chemistry, Faculty of Science, M.S. University of Baroda, Baroda 390 002, India Synthesis and characterization of a series of tercopolymeric chelating resins involving 8-hydroxyquinoline and resorcinol-hydroquinone are discussed. The resins were used in the determination of zinc and cadmium after preconcentration, at batch level, and by flow injection (FI) with micro-columns of the resins. Flow rate, pH, equilibration time, microcolumn dimensions, and the type and concentration of eluents were optimized. It was observed that pH 3 for zinc and pH 3-6 for cadmium are desirable for effective preconcentration.The rate of metal exchange and the activation energy for the exchange were determined at 30, 35, 40 and 45°C. A column length of 2 cm and a flow rate of 2 ml min-1 was found to be optimum for the preconcentration of zinc and cadmium using the FI system. The lower detection limit for the metal ions was found to be 1 ng ml-1. Column chromatographic separations of Zn + Cd; Cu + Cd; Cu + Zn; Cd + Co + Zn; and Cu + Co + Cd were carried out using an eluent switching technique. Keywords: Zinc; cadmium; preconcentration; separation; flow injection The removal of trace metal ions from water by ion-exchange has held the interest of research workers for over a century. However, the batch and column methods used for this purpose have limitations such as the fact that they have a slower rate of removal of ions, hence they are tedious processes, and have lower enrichment factors and require large sample volumes.The advent of on-line ion-exchange enrichment of metal ions using flow injection (FI) has drawn much attentionl-7 because of its numerous advantages over conventional methods. The materials generally used for enrichment are activated alumina ,6 8-hydroxyquinoline immobilized on glass or silica4 and chelex 100.1 In the late 1950s chelating resins containing 8-hydroxyquinoline and resorcinol were used for the precon- centration and separation of metal ions. These resins"10 were reported to have low metal exchange capacities, slow exchange rates and poor chemical stabilities.Several workers attempted to overcome these drawbacks by controlling the water content and curing conditions."-14 Recently, 8-hydroxyquinoline was reported to have been grafted onto organic polymeric supports and successfully used for the preconcentration of Call, Cd", Cull, Mg" and Zn11 ions.15 In the present paper an attempt has been made to improve the chemical stability and exchange kinetics of resins contain- ing 8-hydroxyquinolin by incorporating hydroquinone with the 8-hydroxyquinoline in the polymeric resin. The effects of cross-linking agents on the physico-chemical properties of the resins have also been studied. The resins have been used in the determination of zinc and cadmium with batch and continuous flow processes after preconcentration. Experimental Analytical-reagent grade chemicals and high-purity de-ion- ized water were used to prepare metal ion solutions.Working solutions of zinc and cadmium were prepared from standard stock solutions whenever required. The buffer solutions were prepared from 0.2 mol dm-3 acetic acid and sodium acetate. Synthesis of Resins A series of chelating resins was synthesized from 8-hydroxy- quinoline, resorcinol-hydroquinone and formaldehyde-2- furaldehyde-benzaldehyde as cross-linking agents, then re- fluxed in dimethylformamide (DMF). The time required for gel formation vaned from 0.5 to 40 h depending upon the cross-linking agent used for the synthesis. All resins were ground after curing, and sieved into 60-80 mesh size particles, and Soxhlet extracted using DMF and methanol.The resins were then washed thoroughly with water and converted into the acidic (H+) form by treatment with 0.5 mol dm-3 HCI. Excess of acid was removed completely by washing with de-ionized water prior to further work. The measurements of moisture content, true density, sodium exchange, capacity, t4 for sodium exchange (time required for 50% exchange of metal ion), spectral analysis, elemental analysis and thermal analysis were carried out for characterization of the resins. The relevant results are given in Table 1. Table 1 Physico-chemical properties of the resins Resin Properties Moisture contentlg g-1 True density/g ml-1 Void volume/ml g-1 Sodium exchange capacity/mmole g-1 ti* for sodium exchange/min ti for zinc exchange/min t3 for cadmium exchange/min * ti = 50% exchange.8HOQFR 0.03 1.40 0.78 4.60 7 20 20 8HOQFHQ 0.08 1.50 0.66 5.00 4 9 14 8HOQFuR 0.64 1.25 0.53 2.80 10 35 20 8HOQFuHQ 1.98 1.16 0.49 3.30 12 16 10 8HOQBR 1.51 1.18 0.63 2.60 24 16 30826 ANALYST, AUGUST 1991, VOL. 116 Table 2 Column characteristics Column Interstitial Column voIume/ml volume/ml C&I*- SHOQFR 17.5 9.2 SHOQFHQ 17.5 9.8 8HOQFuR 9.0 4.8 8HOQFuHQ 10.5 5.2 SHOQBR 9.3 4.6 SHOQFR 17.5 9.2 SHOQFHQ 17.5 9.8 SHOQFuR 9.0 4.8 8HOQFuHQ 10.5 5.2 SHOQBR 9.3 4.6 * 1 mg ml-1 taken. -t BTC, breakthrough capacity; TC, total capacity. &II*- Breakthrough capacity/ mequiv. g-1 0.052 0.046 0.043 0.043 0.049 0.094 0.094 0.083 0.076 0.006 Total capacity/ mequiv. g-1 0.21 0.25 0.20 0.32 0.24 0.45 0.52 0.40 0.49 0.41 Degree of utilization (BTC/TC)t/ mequiv .g-1 0.25 0.18 0.22 G. 13 0.20 0.21 0.18 0.21 0.16 0.02 Preconcentration of Metal Ions The resins were used for the batch and column chromatographic studies of zinc and cadmium ion-exchange. Optimization of pH for preconcentration at batch level was carried out by following a literature procedure16 and using 0.2 mol dm-3 acetate buffer of pH 2-7. Zinc and cadmium were determined complexometrically from the supernatant solution and also by eluting the chelated metal from the resin. Rate constants and activation energies for the formation of zinc and cadmium chelates were calculated by measurement of the kinetics of the ion-exchange phenomena at 30,35,40 and 45°C. The mode of diffusion of metal ions through the solution towards the resin was determined by the interruption testl6.17 and confirmed by the method reported by Native et al.18 During the kinetics study using the interruption test, resin particles were interrupted by removing them from the solution for a brief period (10 min) and then re-immersing them.A graph of percentage exchange versus time was plotted and the nature of diffusion of the ions was determined as discussed by Helfferich. 16 Various eluents such as acidic or alkaline reagents, elec- trolytes and complexing agents of different strengths were tested for quantitative elution of zinc and cadmium with batch processes. Chromatographic Separations Chromatographic columns of 9-17 cm in length and 0.7 cm i d . were prepared from the resins 8HOQFR (8-hydroxyquino- line-formaldehyde-resorcinol), 8HOQFHQ (8-hydroxy- quinoline-formaldeh yde-h ydroquinone) , 8HOQFuR (8- hydroxyquinoline-2-furaldehyde-resorcinol), 8HOQFuHQ (8-h ydroxy quinoline-2-furaldeh yde-h ydroquinone) , and 8HOQB R (8-h ydrox yquinoline-benzaldeh y de-resorcinol) .The break-through and column capacities19 and void volume fractions16 were determined according to literature methods. The results are presented in Table 2. Binary and ternary mixtures of Zn + Cd; Cu + Cd; Cu + Zn; Zn + Cd + Co, and Cu + Cd + Co (400 ppm) in 1 + 1 or 1 + 1 + 1 proportions were passed through the columns at the relevant pH at a flow rate of 1 ml min-1. The column was then washed with de-ionized water. Chelated zinc, cadmium, copper and cobalt were eluted with appropriate eluents. Flow Injection for the Enrichment of Metal Ions An FI manifold for the preconcentration of Zn*I and Cd" at nanogram levels was constructed as shown in Fig.1, using 2 x 0.2 cm microcolumns, a Gilson Minipuls peristaltic pump, a Rheodyne RH 5020 injection valve and 0.5 mm i.d. poly- Table 3 Operational conditions for AAS Parameter Monochromator slit widthhm Lamp current/mA Air pressure/lb in-2 Acetylene pressure/lb in-2 Chart speedmm min-1 Recorder range (arbitrary units) Sensitivit y/mV Length of columnkm Column i.d./cm Flow rate/ml min-1 Wavelength for detectiodnm Cadmium 0.5 1.5 40 10 5 04.7 0.5 2 0.2 1.5 228.8 Zinc 1.0 2.5 32 5 5 0-1 .o 0.5 2 0.2 2.0 213.9 Metal solution n De-ionized water 2 Fig. 1 Flow injection manifold for the on-line preconcentration of ZnlI and CdlI using microcolumns containing chelating resins.1, Three-way key; 2, peristaltic pump; 3, injection valve; 4, column containing chelating resins; and 5, flame AA detector tetrafluoroethylene tubing. The detector was a Varian AA 775 flame atomic absorption spectrometer with recorder. All operational conditions are given in Table 3. The absorbance- time response was recorded on an x-t chart recorder. Zinc and cadmium were detected at 213.9 and 228.8 nm, respectively. Preconcentration conditions such as variation in pH, column length, flow rate and concentration were optimized for the determination of zinc and cadmium using the FI system. The chelated metals from the columns were eluted by injecting various reagents of different strengths. The eluted metal was analysed by flame atomic absorption spectrometry (AAS).The column was washed by injecting 50 pl of HCl twice and passing buffer solution through the system. A three-way key was used to control the flow of the standard solution and buffer through the system as necessary. Results and Discussion The physico-chemical properties of the resins under study are given in Table 1. The resins are insoluble in most organic solvents, acids and alkaline solutions of higher concentration and are stable up to 300°C. The infrared (IR) spectra of the resins showed a broad band for polymeric 0-H stretching atANALYST, AUGUST 1991, VOL. 116 827 3400-3200 cm-1, a sharp but weak band of aromatic tertiary C-N vibration at 1370 cm-1, methylene group bending vibration at 1460 cm-I, aromatic ring stretching at 500 cm-l and a weak band at 1600 cm-1 indicating the presence of C=N stretching.These resins did not show instability in 2-4 mol dm-3 acids as did those reported pre~iously.13~14 The sodium exchange capacities and rate of ion-exchange are found to be lower and comparatively slower for the resins prepared with the cross-linking agents 2-furaldehyde and benzaldehyde. The pH studies showed that the metal exchange capacity was at a maximum at pH 6 for zinc and at pH 3-6 for cadmium. The results obtained are presented in Fig. 2. A study of the nature of the diffusion phenomenon of metal ions using the interruption te~t16.l~ indicated that the exchange of metal ions is governed by particle diffusion. With the interruption test, the interruption allows time for the concen- tration gradients in the beads to level out.Therefore, in particle diffusion control .I6317 the rate increases immediately 1.6 1.2 0.8 c I a 0.4 - E E . 40 30 20 10 3 c 0, P) c .- Y O 20 ' 10 ' 0 - 1 4 8 1 4 8 10 12 PH Fig. 2 Effect of pH: (a) batch process for ZnII; (b) FI for Zn"; (c) batch process for Cdll; and (d) FI for Cd". Resin: A, SHOQFR; B, SHOQFHQ; C, 8HOQFuHQ; D, 8HOQFuR; and E, 8HOQBR after interruption. In film diffusion, no concentration grad- ients exist in the beads and the rate depends only on the concentration difference across the film, hence interruption does not affect the rate. In the present study the exchange of metal ions was observed to be a particle diffusion phenom- enon. The t4 values were calculated from the kinetics study and are given in Table 1.Thermodynamics and Rate of Exchange The reaction rates for the uptake of zinc and cadmium by the resins at four different temperatures were calculated and are given in Table 2. The rate constant k (for the initial exchange reaction) is calculated using the equation for the first-order reaction dc kt -= kcand-log(a-fl=- dt 2.303 where a is the initial concentration of metal ion a n d f i s the fraction of ions exchanged on the resin. The plots of -log(a - f) versus t are not straight lines passing through the origin. Hence, a mirror method20 was used where the rate constant was calculated from the initial rate of the reaction by using a mirror which, when oriented normal to the curve, shows a reflection which is a smooth continuation of the curve.A tangent is then drawn perpendicular to the normal, and k was calculated from the slope of the line. The activation energy required for complex formation was determined using the Arrhenius equation. k = Ae-AEIRT 1 T From the slope of the plot of log k versus -, the activation energy AE was calculated and is given in Table 4. The acid dissociation constants (kl and k2) for zinc-8-hydroxyquinoline and cadmium-8-hydroxyquinoline complexes in solution are reported to be 8.5 and 7.4, and 7.8 and 6.2, respectively.15 The higher values for zinc indicate the higher stability of the complex. Thus, the activation energy for the zinc-8-hydroxy- quinoline complex should be higher than that of the cadmium- 8-hydroxyquinoline complex. This is in agreement with our observations (Table 4).As the stability and activation energy of the zinc-8-hydroxyquinoline complex are higher, the ti values for zinc exchange should be higher than those for cadmium, which is found to be true in most instances. Study of Elution Systems Elution of chelated zinc and cadmium was carried out using 0.1-3 mol dm-3 HCI, H2S04, HN03 and CH3COOH, NaCl, sodium citrate, sodium tartrate, KSCN, thiourea and 5-50% m/m HC104. Cadmium was eluted quantitatively with 0.2 mol dm-3 HCI, 0.1 mol dm-3 H2S04 and 0.5 ml dm-3 Table 4 Rate of reaction and activation energy k Metal ion Resin CdlI SHOQFR SHOQFHQ 8HOQFuR 8HOQFuHQ SHOQBR SHOQFR SHOQFHQ 8HOQFuR 8HOQFuHQ SHOQBR 30°C 0.008 0.006 0.005 0.005 0.004 0.008 0.005 0.004 0.003 0.004 35°C 0.011 0.007 0.007 0.006 0.006 0.012 0.007 0.006 0.004 0.004 40°C 0.075 0.008 0.007 0.008 0.007 0.018 0.006 0.008 0.006 0.005 45°C 0.017 0.010 0.008 0.010 0.008 0.019 0.007 0.010 0.006 0.005 Activation energy1 J mol-I 96.2 29.3 37.7 100.4 37.7 142.3 46.0 37.7 405.9 58.6828 40 20 8 - - m c E - 0 C 0 .- 4- - w 40 20 0 -0.1 rnol dm-3 CH3 COOH-/----2 rnol dm-3 HCI - 0.1 rnol dmP3 CH3 COOHi I--0.5 rnol dmP3 CH3 COOH- 20 40 60 80 0.5 rnol dmP3 CH3 COOHi +2 rnol dmP3 HCI 4 20 40 60 80 40 20 ANALYST, AUGUST 1991, VOL.116 -0.1 rnol dm-3 H C l l 1- 2 rnol dm-3 HCI- 0 20 40 60 80 100 -0.1 rnol dmP3 CH3 COOH-2 rnol dm-3 HCI - 20 40 60 80 100 Effluent/ml Fig. 3 Chromato raphic separations: (a) A, Cdrl; and B, Zn", at H 4; (b) A, CdII; and B, CuII, at pH 4; (c) A, Zn"; and B, Cu", at pH 4; (df A, CdII; B, Cu"; and C, Zn", at pH 4; and (ey A, CdII; B, Cu"; and C, Co", at pH 4, all using 8HOQFHQ resin.The eluents used are as given on the figures Table 5 Efficiency of recovery from binary mixtures Amount of Zn2+/mg Amount of Cd2+/mg Resin Taken Recovered 8HOQFR 4 4.08 8HOQFHQ 4 3.99 8HOQFuR 4 4.10 8HOQFuHQ 4 3.90 8HOQBR 4 3.99 Amount of Cu2+/mg 8HOQFR 10 10.07 8HOQFHQ 10 10.12 8HOQFuR 10 10.16 8HOQFuHQ 10 9.77 8HOQBR 10 9.91 Amount of Cu*+/mg 8HOQFR 8HOQFHQ 8HOQFuR 8HOQFuHQ 8HOQBR 10 10 10 10 10 10.42 10.22 10.24 9.96 10.17 Recovery (% ) 102.0 99.7 102.5 97.5 99.7 100.07 101.15 101.60 97.65 99.07 104.2 102.2 102.4 99.6 101.7 Taken Recovered 4.0 4.0 4.18 4.34 4.46 Amount of Cd2+/mg 10 10.15 10 10.30 10 10.16 10 10.79 10 10.34 Amount of Zn*+/mg 10 10 10 10 10 10.77 10.18 10.38 10.27 10.18 Recovery (%) 100.0 100.0 104.5 108.5 111.5 101.5 103.0 101.6 107.9 103.4 107.7 101.8 103.8 102.7 101.8ANALYST, AUGUST 1991.VOL. 116 CH3COOH, and zinc with 3 rnol dm-3 HCI, 3 rnol dm-3 H2S04 and 0.5 rnol dm-3 CH3COOH from almost all resins. This evidence supports the higher stability of the zinc-8- hydroxyquinoline complex over the cadmium-8-hydroxy- quinoline complex. Separations of binary mixtures of Zn + Cd, Cu + Cd; Cu + Zn and ternary mixtures of Co + Zn + Cd,Co + Cu + Zn;andCo + Cu + Cdin 1 + 1 and 1 + 1 + 1 proportions were carried out using 400 ppm metal solutions. The column characteristics are given in Table 2. The solutions were passed through the columns at a flow rate of 1 ml min-1 at the appropriate pH as shown in Fig.2. After washing the column with water, the metal ions were eluted using the specific eluents given in Fig. 3. The efficiency of recovery is given in Table 5. No cross contamination was observed in binary mixtures. Separation of ternary mixtures of Cu + Zn + Cd, Co + Cu + Cd and Co + Cu + Zn showed cross contamination to some extent (results not shown). A typical separation required about 2 h, this time was considerably reduced when continuous FT was used. Flow Injection System By using the FI manifold shown in Fig. 1, without the microcolumn, a calibration graph for zinc and cadmium was constructed by injecting a series of standard metal ion solutions into the carrier streams and plotting the values obtained. These graphs were used to determine the per- centage recovery of metal after preconcentration.Zinc and cadmium were determined at nanogram levels using the FI manifold (Fig. 1). Zinc and cadmium were preconcentrated on the chelating resin microcolumns by passing known volumes of the metal solutions (15 ml of 1 X 10-7 rnol dm-3) through them. The column effluent did not show the presence of metal ions when continuously analysed by AAS, indicating quantitative chelation of zinc and cad- mium on the microcolumn. The column was washed with buffer and the chelated metal was eluted by injecting 50 p1 of 1 rnol dm-3 HN03. The dispersion due to the flow system was calculated by the nebulization method using 50 pg ml-1 of metal solution and injecting 50 pl of 50 pg ml-1 metal solution, through the FI manifold, with and without the introduction of a microcolumn.The introduction of a microcolumn into the manifold did not affect the peak height. The FI peak height in comparison with the steady-state response gave a dispersion coefficient of 2.1 for the manifold. The sharp peak obtained on elution of zinc and cadmium from the microcolumn indicates faster kinetics. Continuous elution of zinc and cadmium using 1 rnol dm-3 HN03 solution with the FI manifold gives a narrow peak of comparable height indicating that the analyte concentration lies in a relatively narrow eluent range. The dispersion in the system is critically dependent on the volume of sample injected and was observed to be higher for smaller volumes. When the results obtained by conven- tional flame AAS were compared with those obtained from the FI system (Fig.l ) , it was observed that the former gives a lower detection limit at the pg ml-1 level whereas the latter gives a lower detection limit at the ng ml-1 level. The recovery after preconcentration was found to be approximately 95%. The pH optimization for the preconcentration of zinc and cadmium was carried out by passing metal solutions in the pH range 1-11.5 through the columns and eluting the chelated metal ions with 50 pl of 1 rnol dm-3 nitric acid solution. When the results obtained using the FI system (Fig. 2) were compared with those from the batch method, no significant variation in the optimum pH for cadmium was observed. However, the pH for zinc shifted to the acidic range (6-3). The column length was optimized for each resin for both Zn and Cd ions using 2 , 3 , 5 and 6 cm long and 2 mm i.d.columns. A 2 cm long column was found to be sufficient for effective preconcentration. Column lengths greater than 6 cm increased the dispersion coefficient by 12%. Hence 2 cm columns were used throughout the study. 829 Zinc and cadmium were eluted using 0.01-1 mol dm-3 acids. The volume of eluent was vaned from 20 to 150 pl and it was found that 20 p1 of 0.05 rnol dm-3 nitric acid quantitatively eluted the metal ions from the resins. However, in order to ensure complete elution, 50 pl of 1 rnol dm-3 nitric acid was 80 E E 2 60 0, 0, s .- Y 40 n 0, 20 I 0 40 80 120 Ilb’ 110 Concentration of metahg ml-’ Fig. 4 Calibration ( a ) and ( b ) for Znl* after preconcentration using the FI manifold. Numbers on the peaks are ng ml-1 of metal.Resin: ., SHOQFuHQ; 0, SHOQFHQ; A, SHOQFuR; x , SHOQBR; and a, SHOQFR 8o Pa) 0 20 40 60 80 100 50 Concentration of metalhg ml-’ Fig. 5 Calibration ( a ) and ( b ) for Cd” after preconcentration, using the FI manifold. Numbers on the peaks are ng ml-1 of metal. Resin: ., SHOQFuHQ; 0. SHOQFHQ; A, SHOQFuR; x , SHOQBR; and a, SHOQFR830 ANALYST, AUGUST 1991, VOL. 116 used as the eluent. The flow rate of the solutions through the FI manifold was varied from 0.5 to 5 ml min-1. By using the continuous scanning process it was found that even at a flow rate of 3 ml min-1 the extraction of metal ions by the resins was quantitative. However, very high flow rates showed a decreased efficiency for the preconcentration process. Conclusion The calibration graphs for zinc and cadmium were obtained by passing 5 ml of metal solution through the FI system, eluting the metal with 50 p1 of 1 mol dm-3 HN03 and plotting the results obtained.All readings were obtained in triplicate. (The results for 8HOQFR are given in Figs. 4 and 5.) The lowest detection limit (three times the standard deviation of the blank signal for the carrier system) was found to be 1 ng ml-1 (when 15 ml of solution was used for preconcentration). The correlation coefficient and average relative standard deviation (RSD) values are given in Table 6. The resins can be used for the continuous preconcentration and determination of metals for more than three months without loss of activity. A typical analysis of the zinc-cadmium mixture (ng ml-1 detection limit) after preconcentration using the FI system requires 5-7 min.Table 6 Analytical data from the calibration graph Zinc Cadmium Correlation Average Correlation Average Resin coefficient RSD* (%) coefficient RSD" (%) SHOQFR 0.9957 1.6 0.9996 1.7 SHOQFHQ 0.9989 1.5 0.9972 1.5 8HOQFuR 0.9993 1.9 0.9995 1.8 8HOQFuHQ 0.9851 1.5 0.9926 1.6 SHOQBR 0.9976 1.7 0.9994 1.7 * Average RSD for n = 6 and five different concentrations. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 References Olsen, S., Pessenda, L. C. R., RfiiiEka, J., and Hansen, E. H., Analyst, 1983, 108, 905. Fang, Z., RfiiiEka, J., and Hansen, E. H., Anal. Chim. Acta, 1984, 164, 23. Hirata, S., Umezaki, Y., and Ikeda, M., Anal. Chem., 1986,68, 2602. Hartenstein, S. D., R6iiZka, J., and Christian, G. D., Anal. Chem., 1985,57, 21. Fang, Z . , Xu, S . , and Zang, S . , Anal. Chim. Acta, 1987,200,35. Zhang, Y., Riby, P., Cox, A. G., McLeod, C. W., Date, A. R., and Cheung, Y. Y., Analyst, 1988, 113, 125. Devi, S., Khalil, A. J. H., and Townshend, A., Quim. Anal., 1989,8, 159. Von Lillin, H., Angew. Chem., 1954.66, 649. Parrish, J. R., Chem. Znd. (London), 1955, 386. Parrish, J. R., Chem. Znd. (London), 1956, 137. Pennington, L. D., and Williams, M. B., Znd. Eng. Chem., 1956, 51, 759. Parrish, J. R., and Stevenson, R., Anal. Chim. Acta, 1979,70, 189. Vernon, F., and Nyo, K. M., Anal. Chim. Acta, 1977,93,203. Parrish, J. R., Anal. Chem., 1982, 54, 1890. Abollino, I., Mentasti, E., Porta, V., and Sarzanini, C., Anal. Chem., 1990,62, 21. Helfferich, F., Ion-Exchange, McGraw-Hill, New York, 1962, pp. 256. Kressman, Y. R. E., and Kitchener, J. A., Discuss. Faraday Soc., 1949, 1, 90. Native, M., Goldstein, S., and Schmuckler, G., Anal. Chem., 1975, 37, 1951. Inczedy, L., Analytical Applications of Ion Exchangers, Per- gamon Press, London, 1966. Frost, A. A., and Pearson, R. G., Kinetics and Mechanism, Wiley, New York and London, 1961, p. 46. Paper 01046570 Received October 1 Oth, 1990 Accepted March 12th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600825

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, AUGUST 1991, VOL. 116 831 Flow Injection Sample-to-standard Additions Method Using Atomic Absorption Spectrometry Applicable to Slurries lgnacio Lopez Garcia, Francisca Ortiz Sobejano and Manuel Hernandez Cordoba* Department of Analytical Chemistry, Faculty of Chemistry, University of Murcia, 30071 Murcia, Spain An extension of the flow injection (FI) sample-to-standard additions method using atomic absorption spectrometry for detection is discussed. Hypotheses are given for slurries injected into a simple FI manifold, with a standard solution of the analyte used as the carrier. An equation is derived allowing the calculation of the concentration of the sample from two signals, one obtained from the injection of the slurried sample and the other from a slurry prepared from a standard of known concentration.The reliability of the approach is confirmed by determining the concentrations of copper, manganese, chromium and zinc in slurries prepared from commercial iron oxide pigments. The results show good precision (relative standard deviation of less than 3.8%) and accuracy (relative errors in the 0.6-5.8% range). Keywords: Flow injection; sample-to-standard additions; slurry; atomic absorption spectrometry; iron oxide pigments As has been recently shown,1.2 the coupling of flame atomic absorption spectrometry (FAAS) with flow injection (FI) techniques provides recognized advantages for a number of analytical problems. With respect to the essential point of calibration, several procedures for the implementation of the standard additions method of calibration have been discussed.Thus, an FI analogue of the standard additions method with atomic absorption spectrometric detection has been devel- oped by Tyson and co-workers.3-7 In this approach, the sample was continuously pumped into the nebulizer, and a steady-state absorbance signal was thus obtained. Subse- quently, plugs of different standards were injected into the carrier, leading to transient absorbances, which were positive or negative depending on the relative concentrations of the analyte in the carrier stream and in the injected standards. The applicability of this procedure was later confirmed by Frenzels for spectrophotometric and potentiometric determinations. This FI approach was also studied in depth by Israel and Barnes,g who used a mathematical treatment in order to develop a conventional standard additions method using inductively coupled plasma atomic emission spectrometry for detection.Recently, Israel and Barnes10 demonstrated the analytical applicability of the sample-to-standard additions method, proving their hypotheses by means of two conventional FI methods with spectrophotometric detection. The equation proposed by Israel and Barnes permits a simple calculation of the concentration of the sample using only two injections, one from the sample and the other from a blank solution. The aim of this paper is to demonstrate the applicability of the sample-to-standard additions method when using slurries in FAAS. The equation proposed below, which is in essence an extension of that discussed by Israel and Barnes, is experimentally tested using aqueous slurries prepared from commercial iron oxide pigments.The determination of chromium, manganese, copper and zinc, using FI to introduce the slurries into an atomic absorption spectrometer, with calibration by means of matched standards, has recently been reported. 1 1 This approach is a useful alternative calibration method and can be extended to other slurried samples. Experimental A Pye Unicam SP1900 atomic absorption spectrometer was used together with a 7040A Hewlett-Packard recorder. The * To whom correspondence should be addressed. measurements were made at 324.8,279.5,213.9 and 357.9 nm for copper, manganese, zinc and chromium, respectively, using conventional hollow cathode lamps.The spectrometer was operated in accordance with the standard conditions recommended by the manufacturer for maximum sensitivity with an air-acetylene flame. The FI manifold used has been described previously. 12 A three-way connector was included in the manifold in order to provide an inlet for air. This T-piece allows air-compensation for the difference between the nebulizer uptake rate and the pumping flow rate, and acts as a pre-nebulizer. Analytical-reagent grade reagents and doubly distilled water were used throughout. Acidic stock solutions of copper, manganese, chromium and zinc (lo00 mg 1-1) were prepared as recommended by the Perkin-Elmer user manual.13 Sodium hexametaphosphate (HMP) was obtained from Fluka and used without further purification.Iron oxide samples were analysed beforehand by treating 1 g of sample with a mixture of hot hydrochloric and nitric acids (3 + 1 v/v), until total dissolution was achieved. The copper, manganese, chromium and zinc contents were then obtained by FAAS using the standard additions procedure. The slurries were prepared as described previously" by suspending the samples in water containing 1% HMP. The sample loops (135 PI) were filled while the suspensions were continuously stirred. The plugs were then injected into the FI manifold. A 1% HMP solution containing a fixed concentra- tion of the analyte being studied was used as the carrier and pumped at 2 ml min-1. Theory For clarity and comparison purposes, the terminology pro- posed by Israel and Barnes10 is used.Consider a standard solution of analyte continuously flowing to the detector, providing a steady-state response signal Ps. Eqn. (1) relates the steady-state response with the background signal Ib, the concentration of the standard solution 0, and K , the proportionality constant of the detector. P, = KCO, + Zb (1) When a discrete volume of a sample solution with a concentration of 0, is injected into the manifold, the detector response at peak height ZpXs is given by: where D is the dispersion at peak height. pxs = KCos - (K/D)CO, + (K/D)COx + l b (2)832 ANALYST, AUGUST 1991, VOL. 116 As Israel and Barnes demonstrated in detail,lO an equation for the determination of the unknown concentration Cox can be obtained: where Zp is the difference between Zpxs and lo,, i.e., the peak height obtained from the injection of the unknown sample, taking the steady-state signal of the carrier solution as the baseline of reference.Similarly, Zm is the difference between the constant signal from the carrier solution and the signal at the peak height when a blank solution is injected. In order to calculate the unknown concentration of analyte Cox, two measurements are required, ZP from the sample injection and Zm from the blank solution injection. This simple approach was outlined and experimentally tested by Israel and Barnes10 using molecular absorption spectrophotometry as the detection technique. Excellent agreement has been found in this laboratory between the values predicted by eqn. (3) and the experimental results when FAAS is used for detection.As stated above, eqn. (3) was obtained by assuming that only solutions are involved. A more general equation can be obtained, using atomic absorption for detection, that is applicable to heterogeneous systems such as slurries. Consider that while a standard solution of the analyte is continuously pumped to the atomic absorption spectrometer, a slurry prepared from a sample containing the analyte is injected into the manifold. Eqn. (2) can be rewritten as ZPXS = KCo, - (K1/D)COs + (K,/D)Co, + Zb (4) where Cox is the unknown concentration of analyte in the slurry, K1 is the proportionality constant of the detector response for the standard analyte solution in the presence of the solid matrix and K2 is the proportionality constant for the analyte supplied as a slurry.Subtracting eqn. (1) from eqn. (4), the peak height ZP with respect to the steady-state signal of the carrier solution is given by: ZP = (K2/D)Cox - (KI/D)CO, ( 5 ) If a slurry with an analyte concentration of Corn is injected, eqn. (4) is again applied, eqn. (1) is subtracted and the following expression for Zm is obtained: I" = (K2/D)Corn - (KJD)Co, (6) Combining eqns. (5) and (6) and rearranging: (I" - Zp)/Zm = (Corn - Cox)/Com - P C o s ) (7) where p is the K1/K2 ratio. The unknown concentration Cox of the analyte in the slurry is now obtained easily: Cox = pCO, (Zm - ZP)/Z" + ComZP/Z" (8) It is interesting to note that eqn. (8), derived for a heterogeneous system, is really an extension of eqn. (3). When the conditions K1 = K2 (p = 1) and Corn = 0 (i.e., the slurry used as reference is without analyte) are met, the equation for homogeneous systems given by Israel and Barnes is obtained.As occurs for homogeneous systems, working within the linear range of the detector, a plot of Cox versus ZP at a constant 0, must be a straight line, the slope being given by (Corn - PcO,)/Irn From the intersection of this line with the abscissa axis a value for COX of PCo, is obtained. So that eqn. (8) can be applied, two different injections are needed, one of the unknown slurry and the other using a slurry prepared from a sample having a known analyte content. Furthermore, the p value for each analyte must have been previously calculated in a given solid matrix. It is clear that P, being an empirical coefficient, depends on the experimental conditions and must be recalculated when the conditions are changed.Results and Discussion The small particle size of commercial iron oxide pigments allows the preparation of slurries which can be supplied to the flame of an atomic absorption spectrometer using a simple FI manifold.ll Slurries from these pigments were used in order to verify the above hypotheses. To prove eqn. (8), the value of the p coefficient must be calculated. There are several approaches that can be followed in order to obtain this value, and one based on the separate estimation of K1 and K2 was used. While a carrier stream containing 1% HMP solution and no analyte was being continuously pumped through the system, a slurry prepared from a previously analysed iron oxide was injected into the manifold five times.The mean signal thus obtained gave an estimate of the response of the detector to the analyte being supplied as a slurry (K2). Next, a known amount of aqueous standard analyte was added to the slurry and five injections were again performed. From the mean signal obtained, the signal used for the estimation of K2 was subtracted and this allowed the estimation of the proportionality constant of the detector for the standard analyte solution in the presence of the solid matrix (Kl). This procedure for the estimation of p was performed for copper, zinc, manganese and chromium Slurry (%) 0 0.1 2.4 Q, (D - 2 1 . 9 - 1.4 - I D I I 0.9 0 0.5 1 .o 1.5 Slurry (%) Fig. 1 Variation of p coefficient with the slurry percentage A 4 min - Time --c Fig.2 Flow injection sample-to-standard additions method for the determination of manganese in iron oxide slurries. A sample loop of 135 pl was used. A, Water; B, a 1% slurry prepared from a sample containing 108 pgg-1 of manganese; C, 138; D, 155; E, 199; F, 205; G, 214; H, 230; and I, 289 pg g-1. The carrier was a 1% HMP solution containing 2 pg ml-l of manganeseANALYST, AUGUST 1991, VOL. 116 833 using slurries containing different percentages of iron oxide, and the results are summarized in Fig. 1. As can be seen from the graph, there was a significant difference between the behaviour of copper and that of zinc, chromium and man- ganese. The p value for copper was nearly unity over the whole range studied proving, in accordance with previously reported data,ll that the detector response is the same irrespective of the analyte being supplied to the flame as an aqueous standard or as a slurry.On the other hand, p values for manganese, chromium and zinc, all of which have lesser atomization efficiencies than copper, increased with the slurry percentage and were near unity only for diluted slurries, suggesting a matrix effect. This means that eqn. (3) can only be applied for solutions containing a low percentage of slurry, which is too restrictive from a practical point of view. To prove eqn. (8) a number of measurements were made using 0.2 and 1% slurries. Fig. 2 shows the results obtained for 2 0 v) C w .- ; -2 9 2 -4 4 4 .- I a -6 I I I 1 100 200 300 400 500 Content of analyte in iron oxidedpg g-’ Fig.3 &content of analyte relationships. A , Steady-state signal due to 2 pg ml-1 of manganese or chromium; B, plot of the results shown in Fig. 2; and C, plot of the results obtained from the measurements of chromium. The carrier was a 1% HMP solution containing 2 pg ml-1 of manganese or chromium Table 1 Experimental verification of eqn. (8) Data obtained by plotting Cox versus I P , value 2 SD* Metal Slope? Intercept B fk! rfr SD Cu 0.047 f 0.002 204 f 3.7 1.02 rfr 0.019 0.99 rfr 0.015 Cr 0.036 t 0.004 439 t 5.5 2.19 -C- 0.026 2.20 rfr 0.020 Mn 0.037 2 0.002 268 rfr 1.7 1.34 f 0.017 1.35 k 0.012 Zn 0.048 k 0.001 225 k 1.5 1.50 -C- 0.018 1.50 rfr 0.015 * SD = Standard deviation. t Slopes are given in absorbance units per 100 pg g-1 of analyte content.Data for Zn were obtained using a 0.2% slurry and 0.3 pg ml-1 of analyte in the carrier. values obtained by using a separate estimation of K I and K2. $ Experimental manganese when 1% slurries, prepared from previously analysed samples, were injected into a carrier solution containing 1% HMP and analyte at a concentration of 2 pg ml-1. For comparison, the results for two injections of water are also shown. In accordance with the above hypoth- esis, when some slurries having a higher concentration than the aqueous standard solution used as the carrier were injected, negative peaks were obtained as expected, as in this instance K1 > K2. As predicted by eqn. (8), a plot of Cox versus IP must be a straight line which intersects the abscissa axis to give a Cox value of PCO,.These points are demonstrated in Fig. 3 for two sets of experiments using 1% slurries for the analytes manganese and chromium. Data on the x-axis are the analyte concentrations in the solid matrix because, as all the slurries were prepared using the same percentage and injected under identical experimental conditions, the data are directly related to COX. The perpendicular line from the point of intersection of the manganese data (line B) with the steady-state signal (line A) crosses the x-axis at 269 pg g-1, which agrees, taking into account the 0, value (2 pg ml-I), with the p value of 1.35 calculated for this analyte when using 1% slurries. In a similar manner, the line from the chromium data (line C) intersects the x-axis at 439 pg g-1, in agreement with the p coefficient calculated previously for this analyte.Close agreement was also found between the calculated p values for zinc and copper and those obtained from graphs similar to Fig. 3. The results obtained for 8 values, slopes of the graphs and intercepts for the four analytes studied are summarized in Table 1. In order to test the reliability of the whole calibration approach, eight samples of commercial iron oxide pigments were analysed for the four analytes studied, using the FI sample-to-standard additions method. Slurries of 1% were used for copper, manganese and chromium, and 0.2% slurries for zinc, and eqn. (8) was applied. As shown in Table 2, the results obtained using this procedure agree with those obtained using conventional dissolution in acids followed by AAS measurements as indicated under Experimental.The results were tested using a Student’s t-test and no statistically significant difference between the two procedures was found. Financial support from the Spanish Direccion General de Investigacion Cientifica y TCcnica (DGICYT) (Project 87- 0053) is gratefully acknowledged. References Flow Injection Atomic Spectroscopy, ed. Burguera, J. L., Dekker, New York, 1989. Tyson, J . F., Anal. Chim. Acta, 1990, 234, 3. Tyson, J. F., Anal. Proc., 1981, 18, 542. Tyson, J. F., and Idris, A . B . , Analyst, 1981, 106, 1125. Tyson, J . F., Appleton, J . M. H., and Idris, A . B . , Analyst, 1983, 108. 153. Table 2 Comparison of the results for eight iron oxide samples analysed by the acid-dissolution and slurry procedures Analyte concentratiodpg g-1* Copper Chromium Manganese Zinc Sample A-t BS A t BS A t BS A t Bf 112k4 113+3 2 4 0 f 6 2 3 7 f 5 135+3 132+3 8 0 2 2 84+1 26555 269-C-4 6 1 2 3 63+1 10554 10724 1 6 9 f 5 167-C-4 14325 1 4 2 t 2 2 6 7 f 8 263 2 5 156+5 160-C-3 176+5 17323 1 2 6 f 4 128k3 201 + 7 20625 2 1 5 f 6 219rfr5 138+4 13522 28928 285 +4 21428 217-C-5 155+6 1 5 7 f 4 19924 198+3 205 5 4 206+2 1 3 8 f 5 131 + 5 108-C-4 1 1 2 f 2 23026 237+3 367 f 6 295 f 9 882 f 12 6402 10 539 f 10 221 2 11 476 f 10 44O-C-9 362 + 3 293 2 3 881 -t 6 644-C-6 535 1- 5 216 k 7 480 f 7 439 k 6 * Mean value -C- standard deviation of three determinations. t Acid-dissolution procedure. S Slurry procedure.834 6 Tyson, J. F., Appleton, J. M. H., and Idris, A . B., Anal. Chim. Acta, 1983, 145, 159. 7 Tyson, J. F., and Idris, A. B., Analyst, 1984, 109,23. 8 Frenzel, W. , Fresenius Z . Anal. Chem., 1988, 329, 668. 9 Israel, Y., and Barnes, R. M., Anal. Chem., 1984, 56, 1192. 10 Israel, Y., and Barnes, R. M., Analyst, 1989, 114, 843. 11 Lopez Garcia, I . , Ortiz Sobejano, F., and Hernandez Cordoba, M., Analyst, 1991, 116, 517. ANALYST, AUGUST 1991, VOL. 116 12 Lopez Garcia, I . , Hernandez Cordoba, M., and Sanchez- Pedreiio, C., Analyst, 1987, 112, 271. 13 Analytical Methods for Atomic Absorption Spectrometry, Per- kin-Elmer, Norwalk, Connecticut, USA, 1982. Paper 0105821 A Received December 28th, 1990 Accepted March 26th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600831

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, AUGUST 1991, VOL. 116 835 Flow Analysis Method for the Determination of Silicic Acid in Highly Purified Water by Gel-phase Absorptiometry With Molybdate and Malachite Green Kazuhisa Yoshimura Chemistry Laboratory, College of General Education, Kyushu University, Ropponmatsu, Chuo-ku, Fukuoka 810, Japan Ushio Hase Resources and Environment Protection Research Laboratories, NEC Corporation, Miyazaki, Miyamae-ku, Kawasaki 216, Japan A flow analysis method based on the on-line concentration and direct absorptiometric measurement of the ion-associate species formed between molybdosilicate and Malachite Green on Sephadex LH-20 gel, is described for the determination of silicic acid in highly purified water. The increase in the attenuance of the coloured species, which was concentrated on-line onto the gel packed in a flow-through cell, could be measured continuously with a spectrophotometer at 627 nm.The coloured species on the gel was easily desorbed with a mixture of sulphuric acid and acetone, and the gel-packed flow-through cell could be used repeatedly. The concentrations of silicic acid in 4-5 samples of highly purified water were determined in less than 1 h. The detection limit was 0.1 yg dm-3 of Si using water samples with a volume of about 5 cm3. The total amount of silicate-silicon in highly purified water was determined after treating the samples with ammonia sol ut ion. Keywords: Solid-phase absorptiometry; flow analysis; silicic acid determination; Molybdenum Yellow; Malachite Green The demand for highly purified water is increasing steadily in such fields as the electronic industry for the production of very large-scale integrated circuits.Silicic acid is one of the most difficult components to remove and usually the first that begins to leak from an exhausted mixed-bed ion-exchange column during the high-purity water producing process. Although many sensitive analytical methods have been developed to measure increasingly lower concentrations of silicic acid ,14 almost all of these methods require expensive equipment andor some time-consuming preconcentration steps. A very sensitive method, viz., gel-phase absorp- tiometry, based on the direct measurement of the light absorption by Molybdenum Blue species concentrated on Sephadex G-25 gel, has been reported, whereby silicic acid can be determined at yg dm-3 levels in a conveniently short time using 100 cm3 water samples.5 The implementation of on-line detection with solid-phase retention, recently applied to the flow analysis of copper,6J bismuth,8 chromium,g molybdenum,lo iron11 and phos- phate ,12713 might afford a more convenient flow analysis method for the determination of silicic acid in highly purified water. Although complete desorption is one of the pre- requisites for repeated measurements, when a Molybdenum Blue-Sephadex gel systems was first used in flow analysis, the coloured species could not be completely desorbed from the Sephadex gel particles in flow-through cell.Subsequently the ion-associate species formed from molybdosilicate and Mala- chite Green and used for silicic acid determination in aqueous ~olutions,4~14 and the species similarly produced from molyb- dophosphate and Malachite Green,12 were found to be strongly absorbed on Sephadex LH gel, a hydroxypropyl derivative of Sephadex dextran gel, and to be easily and completely desorbed by using a mixture of sulphuric acid and acetone.The aim of this work was to show that, by using an ion-associate species (the final product from the reaction between molybdate and Malachite Green), gel-phase absorp- tiometry with Sephadex LH gel could be applied to the determination of silicic acid in highly purified water using a flow analysis method. Experimental Chemicals All chemicals used were of analytical-reagent grade. Water, purified with a Milli-Q SP reagent water system (Millipore), was assumed to be silicic-acid free and was used for the dilution of samples and reagents. Poly(tetrafluoroethy1ene) (PTFE), polypropylene or polyethylene containers, pre- treated with a hydrofluoric acid-sulphuric acid solution, were used throughout. Standard silicic acid solution, 1000 mg dm-3 of Si.Supplied by Wako Pure Chemicals. A working solution was prepared by diluting the standard solution with water. Ammonium molybdate solution (for procedure l ) , 5% (mlv). Prepared by dissolving 12.5 g of (NH4)6M07024-4H20 (special analytical-reagent grade, Wako Pure Chemicals) in 250 cm3 of water. Ammonium molybdate-sulphuric acid solution (for pro- cedure 2). Prepared by dissolving 5 g of (NH4)6M07024.4H20 in 250 cm3 of water containing 3.5 cm3 of concentrated sulphuric acid.A fresh solution was prepared every 3 d. Malachite Green solution, 0.01%. Prepared by dissolving 0.025 g of Malachite Green oxalate (Kishida) in 250 cm3 of water containing 42 cm3 of concentrated sulphuric acid. These last three solutions were filtered through 0.22 ym membrane filter-papers (Millipore). Carrier solution for Malachite Green, 3 mol dm-3 sulphuric acid solution. Prepared by diluting 41.7 cm3 of concentrated sulphuric acid with water, to a total volume of 250 cm3. Other carrier solutions. Prepared by diluting 2.8 cm3 of concentrated sulphuric acid to 1 dm3 with water (procedure 1, carrier solution for sample), or by diluting 3.5 cm3 of concentrated sulphuric acid to 250 cm3 with water (procedure 2, carrier solution for molybdate).Desorbing agent solution. Prepared by diluting 30 cm3 of concentrated sulphuric acid to 500 cm3 with water and mixing the solution with 500 cm3 of acetone. Dissolved gases were expelled from the solution using an ultrasonic bath. Sephadex LH-20 gel (Pharmacia).836 Apparatus Light measurements were made with a Nippon Bunko UVIDEC-320 double-beam spectrophotometer. A per- forated metal plate of attenuance" 2 was placed in the reference beam to balance the light intensities. The flow-through cell was supplied by Nippon Sekiei Glass; it had black sides, a path length of 10 mm and a diameter of 1.5 mm. In order to fill only a 3-5 mm length of the light-path portion of the cell with the gel, the cell was blocked with a small polypropylene filter disc .6,8-10,12 An internally mirrored tube (40 x 12 mm i.d.) was placed between the cell holder and the light-detector window to recover partly the light scattered from the cell.This reduced the background attenuance by about 1. ANALYST, AUGUST 1991, VOL. 116 Procedure for Sample Preparation for the Determination of the Total Amount of Silicic Acid in Highly Purified Water A 50 cm3 volume of sample solution in a 100 cm3 PTFE vessel was placed in a polypropylene container together with 200 cm3 of electronic industry grade ammonia solution (Kanto Chemi- cals) in a 200 cm3 polyethylene vessel, and allowed to stand for 1 d. Then, the ammonia solution was displaced by 200 cm3 of a 9 mol dm-3 solution of sulphuric acid and the container was allowed to stand for 1 d at 60 "C to eliminate the ammonia from the sample solution.The amount of water removed by evaporation during this procedure was estimated by measure- ments of the mass of the vessel containing the sample solution. Procedure for the Determination of Silicic Acid in Water by Flow Analysis After Molybdosilicate Formation as in the Batch Method (Procedure 1) To a water sample containing 10-250 ng of Si in a 50 cm3 polypropylene calibrated flask, 0.85 cm3 of a 3 rnol dm-3 sulphuric acid solution and 2.5 cm3 of a molybdate solution were added, and the mixture was diluted to 50 cm3 with water. The solution was allowed to stand for 10 min at room temperature. Then, 3.7 cm3 of this solution were introduced into the stream by means of a six-way rotary valve, 3 min after the introduction of Malachite Green (3.0 cm3) into the stream by means of another six-way rotary valve.The carrier solution for the sample was pumped at a flow rate of 1.25 cm3 min-1 and that for the Malachite Green solution at a flow rate of 0.25 cm3 min-1 with respective Sanuki DM2M-1024 pumps. The attenuance was monitored continuously at the maximum absorption wavelength of the coloured species (627 nm) and the attenuance was recorded on a strip-chart recorder set at 0.5 or 1.0 attenuance full scale. The increase in attenuance from the background (AA) was measured and the silicic acid concentration was determined by means of a calibration graph. Then, the desorbing agent solution (2 cm3) was introduced into the stream by means of another six-way rotary valve. Before the introduction of the next sample, the tube for the sample was washed with about 1 cm3 of 2 mol dm-3 sodium hydroxide solution.All tubing was made of PTFE (1 mm i.d.). A schematic diagram of the flow analysis system is shown in Fig. l(a). Procedure for the Determination of Silicic Acid in Water Using the Flow Method (On-line Colour Development, Procedure 2) First, 3.7 cm3 of a Malachite Green solution were introduced into the stream by means of a six-way rotary valve. The carrier * Attenuance: in solid-phase absorptiometry, the effect of light scattering on measurements is considerable; hence the term 'attenu- ance' is preferable to 'absorbance', although attenuance has essen- tially the same meaning as absorbance. H B Fig. 1 Schematic aiagram of (a) the double-line flow analysis system and (b) the manifold for FI: A, pump; B, six-way rotary valve for sam le introduction [with (a) a 3.7 cm and (b) a 7.9 cm3 PTFE tube loop? C, six-way rotary valve for molybdate solution introduction (with a 2.0 cm3 PTFE tube loop); D, six-way rotary valve for Malachite Green introduction [with (a) a 3.0 cm3 and (b) a 3.7 cm3 PTFE tube loop]; E, reaction tube (2 m of PTFE tubing of 1 mm i.d.); F, mixing tube (0.5 m of PTFE tubing of 1 mm id.); G, six-way valve for desorbing agent introduction; and H, gel-packed detector solution was pumped at a flow rate of 0.25 cm3 rnin-' with one side of a Sanuki DM2M-1024 double-plunger type pump.After 2 min, a sample solution (7.9 cm3) containing 2-50 ng of Si and a molybdate solution (2.5 cm3) were introduced at the same time into the stream by means of respective six-way rotary valves.The carrier solution for the sample was pumped at a flow rate of 1 cm3 min-1 with a Gasukuro Model MPD-3MG medium-pressure pump, and that for the molybdate solution at a flow rate of 0.25 cm3 min-1 with the other side of the Sanuki double-plunger type pump. The attenuance was monitored continuously and recorded as in procedure 1. Before the introduction of the next sample solution, the tube was washed with about 1 cm3 of 2 mol dm-3 sodium hydroxide solution by means of the six-way rotary valve used for sample introduction. Then, the desorbing agent solution was introduced into the stream by means of another six-way rotary valve. A schematic diagram of the flow analysis system is shown in Fig.l(6). Results and Discussion Absorption Spectra of the Ion Associate Species of Molybdenum Y ellow-Malachi te Green By analogy with the molybdophosphate system,12 the mono- valent cationic species of Malachite Green (MG+), showing maximum absorption at 618 nm, forms an ion associate with molybdosilicate, (MG+),H4-nSiMo12040 (n = 1 4 ) . The ion associate employed for the spectrophotometric determination of silicic acid [after the coloured species had been stabilized by adding poly(viny1 alcohol)14 or collected by flotation15 or filtration41 adsorbs well on Sephadex LH-20 gel with maxi- mum absorption at 627 nm (Fig. 2). Therefore, the molybdo- silicate-Malachite Green system was employed for solid- phase absorptiometry.ANALYST, AUGUST 1991, VOL.116 837 4.0 2.0 : 400 600 800 Wavelengthhm Fig. 2 Absorption spectra of the molybdosilicate-Malachite Green ion associate in the gel phase. Sample: volume, 100 cm3 (A and C, blanks; B and D. 0.5 pg of Si): 5% ammonium molybdate (1 mol dm-3 H2S04), 5 cm3; 0.01 rnol dm-3 Malachite Green (2 rnol dm-3 H2S04), 1 cm3; and 9 rnol dm-3 H2S04, A and B, 6 cm3; C and D, 2 cm3. Sephadex LH-20, 0.2 g; and cell length, 5 mm Contamination of Silicic Acid From Water, Reagents and Containers It is difficult to obtain water that is completely free from silicic acid. Water purified by sub-boiling distillation using a PTFE vessel5 or water purified with a Milli-Q system, in which the activated carbon, ion-exchange resins and membrane filter are renewed, can be regarded as being completely free from silicic acid.Contamination from the reagents used was negligible and hence further purification of the reagents was not necessary. Poly(tetrafluoroethy1ene) , polypropylene or poly- ethylene containers, pre-treated with a hydrofluoric acid- sulphuric acid solution, were used throughout. Concentration of Molybdate and Sulphuric Acid Silicic acid reacts with molybdate to form the (3-isomer of the yellow complex under acidic conditions: the final concentra- tions of molybdate and acid suggested by Iwasaki16 are 0.01-0.02 and 0.05 rnol dm-3, respectively. As shown in Fig. 3, because the absorbance of the reagent blank due to the Malachite Green is high in the Iow acid concentration range, the acid concentration for ion-associate formation is critical, and was therefore examined using the batch method.After allowing 15 min for the formation of the Molybdenum Yellow in a 100 cm3 sample solution under the conditions recommen- ded in the literature,16 sulphuric acid and Malachite Green solutions and then 0.2 g of Sephadex LH-20 were added to the solution. The mixture was stirred for 25 min. The attenuance at the maximum absorption of the species (627 nm), and that in a range where only the gel beads attenuate light (750 nm), were measured for the gel layer in a 5 mm ce11.12,17 The effect of sulphuric acid concentration is shown in Fig. 3. Although the net absorbance of silicic acid decreased as the sulphuric acid concentration increased, a concentration of 0.5 mol dm-3 sulphuric acid was finally adopted as a compromise between the decreased net absorbance of the sample species and a lower background attenuance.Application of Gel-phase Absorptiometry to Flow Analysis (Procedure 1) Colour development profile The colour development profile of a sample component sorbed in a gel phase can be be predicted according to the plate O U - u 0.2 0.3 0.4 0.5 H2S04 concentrationlmol dm-3 Fig. 3 Effect of acid concentration on colour development (batch method). Sample: volume, 100 cm3 (A, blank; and B, 0.4 pg of Si); 5% ammonium molybdate (1 rnol dm-3 &SO4), 5 cm3; and 0.01 mol dm-3 Malachite Green (2 rnol dm-3 H2S03), 1 cm3. Sephadex LH-20, 0.2 g; and cell length, 5 mm 1.0 t + c1 0.5 0 5000 10000 15000 n Fig. 4 Retention of components with different distribution ratios on a gel column in the flow-through cell.Distribution ratio: A, 10; B, 100; C, 500; D, 1000; and E, 10000 (for a definition of Q see text) theory.18.19 Here, rn, i, and V are the mass of gel in the flow-through cell, the interstitial solution volume of the gel, and the volume of the mobile phase of each plate, respec- tively. The distribution ratio D of the sample component in question is defined as the amount of sample component in 1 g of gel divided by the amount in 1 cm3 of solution after equilibration has been attained. For example, if nV cm3 of solution is introduced into a flow-through cell (a gel column consisting of 11 plates numbered 0-10) the fraction (Q) of the total amount of the sample species remaining in the column can be estimated as: where the sample volume loaded is 10 000 V cm3 and k 2 p .If n is smaller than 10000, then k = 1 instead of n - 9999. Because m/i for a column in which spherical beads are packed is approximately 2, plots of Q versus n can be constructed for given values of D (Fig. 4).? In systems with high distribution ratios, all the sample component introduced is retained in the cell; in systems with low distribution ratios, a steady state is reached after only a small amount of sample component has been introduced into the cell. In a previous experiment,12 because the value of D for the reagent blank was somewhat t Provided that the height equivalent of the theoretical plate is 0.05 cm and the inner diameter of the cell is 0.15 cm, Vcorresponds to about 0.3 mm3.838 ANALYST, AUGUST 1991, VOL.116 smaller than that for the Malachite Green-molybdophosphate ion-associate species, the blank species was eluted from the cell in a fairly short time; the Malachite Green associated with molybdophosphate was still retained after at least 20 cm3 of the carrier solution had passed through the cell. Therefore, in order to decrease the background attenuance and to lower the detection limit, the difference (AA) between the attenuance before loading a mixture of the Malachite Green and the sample solution and that after desorption of the reagent blank from the cell was used for measurements. The attenuance difference was measured in a similar way to that used for the phosphate system; however, when measurements are made after the complete desorption of the reagent blank from the cell, this procedure is somewhat time-consuming, and the ion-associate species is not as stable as that of the phosphate system.Because the value of D for Malachite Green is 40 cm3 g-1, the sorption of Malachite Green on the gel should reach a steady state shortly after its introduction into the flow system. The value for the Malachite Green associated with molybdosilicate could not be determined, but was expected to be much greater than that for the free Malachite Green. Therefore, as shown in Fig. 5, the time required for measurements would be much shorter, and the precision of this method much improved if the Malachite Green solution was introduced before the sample and the value of AA was obtained from the chart recorder by calculating the difference before and after loading the sample in the cell.Waiting period for the formation of the Molybdenum Yellow solution prior to the introduction of the sample into the flow system A waiting period of 2-20 min after mixing the sample and colouring agent solutions affected the results only very slightly. After 20 min, the absorbance for 2 pg dm-3 of Si decreased gradually; at the same time, the absorbance of the reagent blank increased. The waiting period was fixed at 10 min. Concentration of Malachite Green The dependence of the absorbance on the Malachite Green 'concentration is shown in Fig. 6 . A value of 0.01% was selected as the optimum. Malachite Green in neutral and 10.1 A 1 II 0 20 40 Ti me/mi n Fig. 5 Colour development profiles for the Malachite Green reagent blank and for the Malachite Green-rnolybdosilicate ion associate: 1 , introduction of Malachite Green; and 2, introduction of molybdo- silicate (2 pg dm-3 of Si) alkaline solution is easily adsorbed onto the walls of PTFE tubes and various types of containers.The reagent was dissolved in a 3 mol dm-3 solution of sulphuric acid so that mixing of the Molybdenum Yellow and Malachite Green solutions led to a 0.5 mol dm-3 sulphuric acid concentration in the flow system. Carrier solution The concentrations of sulphuric acid in the two carrier solutions were the same as those for the sample solution (0.05 mol dm-3) and the Malachite Green solution (3 rnol dm-3). Desorbing agent Attempts were made to apply the Molybdenum Blue- Sephadex gel system (previously used in gel-phase colori- metry, batch methods) to the present method; however, the coloured species could not be completely desorbed from the Sephadex G-25 gel.The ion-associate species on Sephadex LH-20 could be quantitatively desorbed with a 50% acetone solution containing sulphuric acid at a concentration of about 0.5 mol dm-3. 0 0.005 0.010 0.01 5 Malachite Green oxalate concentration (%) Fig. 6 Effect of Malachite Green concentration (procedure 1): Sample: volume, 3.7 cm3 (A, blank; and B, 7.4 ng of Si); carrier solution, 0.05 mol dm-3 H2S04; and flow rate, 1.25 cm3 min-l. Malachite Green solution: volume, 3.0 cm3; carrier solution, 3 mol dm-3 H2S04; and flow rate, 0.25 cm3 min-1. Gel, Sephadex LH-20; flow-through cell, 1.5 mm diameter; and wavelength, 627 nm k .1 A 0 30 I D E l 60 Ti mehi n / 90 120 Fig. 7 Colour development profiles of silicic acid using the flow system (procedure 1). Sample: volume, 3.7 cm3, A, blank; B, 3.7; C, 7.4; D, 11.1; and E, 14.8 ng of Si; F, water for thermal power generation 40 cm3 sample in a 50 cm3 calibrated flask); G, desorbing agent I 0.5 rnol dm-3 H2S04, 50% v/v acetone); carrier solution, 0.05 rnol dm-3 H2S04; and flow rate, 1.25 cm3 min-1. Malachite Green solution: volume, 3.0 cm3; carrier solution, 3 rnol dm-3 H2S04; and flow rate, 0.25 cm3 min-l. Gel, Sephadex LH-20; flow-through cell, 1.5 mm diameter; and wavelength, 627 nmANALYST, AUGUST 1991, VOL. 116 839 Memory effect Highly concentrated molybdosilicate in a sample solution was partly adsorbed onto the walls of the PTFE sample loop tube, leading to a higher analytical value for the next measurement. This problem could be overcome by washing the PTFE tube with sodium hydroxide solution prior to the introduction of the next sample solution into the sample loop tube.Calibration As shown in Fig. 7, the increase in the attenuance (AA), obtained from the chart recorder by calculating the difference between the attenuance before and after the loading of the sample, was used for calibration. Precision and detection limit The relative standard deviation was about 3.5% (four determinations) at the 2.3 pg dm-3 level of Si for water circulating in a thermal power generation plant (Fig. 7). The detection limit, defined as the concentration producing an absorbance equal to twice the standard deviation of the background attenuance, was 0.1 pg dm-3 of Si (four deter- minations). Effect of concomitant ions The effect of foreign ions is shown in Table 1.Germanate severely interfered when the molar ratio of this ion to silicic acid was 0.1 : 1 or greater. However, the amounts of interfer- ing ions normally present in highly purified water are generally tolerable. Recovery tests were performed by adding known amounts of silicic acid (Table 2); the recovery of silicic acid was almost quantitative. The proposed method can be applied directly to the analysis of highly purified water. Effect of sample volume on the sensitivity The sensitivity of the proposed method increases linearly with the volume of the sample solution introduced into the flow system.The use of loops of different length gives a wide concentration range for calibration. Application to Flow Injection (Procedure 2) The results obtained by following procedure 1 were employed as the conditions for flow injection (FI). After introducing the Malachite Green solution, the highly purified water sample and the molybdate solution were introduced into the stream at the same time. A reaction tube length of at least 2 m (PTFE tube, 1 mm i.d.) was required. Although the reaction did not proceed sufficiently for complete coloration, and therefore the sensitivity was lower than that for procedure 1, silicic acid could be determined at pg dm-3 levels or lower by FI using the gel-packed detector described here (Fig. 8). A comparison of the sensitivities of various analytical methods for the determi- nation of silicic acid is shown in Table 3.Recently, the Committee for Japan Industrial Standard has proposed a testing method for the determination of silicic acid in the range 1-20 p.8 dm-3 of Si in highly purified water: the method involves extraction of Molybdenum Blue from 200 cm3 sample solutions into butan-1-01 (25 cm3) , followed by absorptio- 0 30 60 90 Ti me/mi n Fig. 8 Flow injection profiles obtained with the flow system (procedure 2). Sample: volume, 7.9 cm3; A, blank: B, 4.0; C, 7.9; D, 11.7; E, 15.6; F, 19.5; and G, 23.4 ng of Si: H, desorbing agent (0.5 mol dm-3 H2S04, 50% v/v acetone); carrier solution, H,O; and flow rate, 1.00 cm3 min-1. Molybdate solution: volume, 2 cm3; carrier solution, 0.25 mol dm-3 H2SO4; and flow rate, 0.25 cm3 min-1. Malachite Green solution: volume, 0.25 cm3; carrier solution, 3.0 mol dm-3 H2S04; and flow rate, 0.25 cm3 min-1.Gel, Sephadex LH-20; flow-through cell, 1.5 mm diameter; and wavelength, 627 nm Table 1 Effect of foreign ions on the determination of silicic acid. Procedure 1; sample volume, 4.8 cm3 (4.0 pg dm-3 of Si) Table 3 Comparison of sensitivity Molar ratio (foreign Si found/ Foreign ion ion : Si) pg dm-3 Error( % ) P" 0.1 1 AsV 1 10 GelV 0.1 1 Fellr 10 CU" 10 4.0 5.4 4.1 4.3 5.1 9.3 3.9 4.0 0 +35 +3 +8 +28 + 132 -3 0 Table 2 Recovery tests (procedure 1) Sample A *- Si addedtlng 0 25 50 75 Si found/ng 56 84 109 131 Si addedtlng 0 20 40 60 Si foundlng 46 66 84 107 Sample BS- * Sample A: water circulating in a thermal power generation plant t Amount of Si in 50 cm3 calibrated flask.Volume introduced into S Sample B: doubly distilled water treated with ammonia solution (20 cm3). the system: A, 3.7; and B, 4.8 cm3. (40 cm3). Amount of sample Method taken/cm3 Absorptiometryt Molybdenum Yellow - Molybdenum Blue - Molybdenum Blue$ 200 Molybdenum Blue§ 100 Malachite Green7 Gel-phase absorptiometry Procedure 1 3.7 Procedure 2 7.9 Ion chromatography 0.1 FAAS11 - ETAAS** 0.01 ICP- AES1-t - Sensi- tivity*/ Detection pg dm.-3 limitlpg of Si dm-3 of Si Reference 3.5 0.2 5 1 .o 0.1 Thiswork 1 .o 0.2 Thiswork - 22 21 - 100 22 - 2 22 - 5 3 * Si concentration giving a final absorbance of 0.1. t Cell length: 1 cm. 5 Extracted into 25 cm3 of butan-1-01; cell length: 2 cm. 5 Cell length: 0.5 cm.7 Proposed method. (1 FAAS = Flame atomic absorption spectrometry. ** ETAAS = Electrothermal atomic absorption spectrometry. tt ICP-AES = Inductively coupled plasma atomic emission spec- trometry.840 ANALYST, AUGUST 1991, VOL. 116 Table 4 Analysis of highly purified waters SUpg dm-3 Purification process Soluble Total* Milli Q system 0.3 0.9 Milli Q SP system < O . l t 0.7 Milli Q system followed by sub-boiling distillation Ion exchange followed by double distillation using using a fused-silica still 0.2 1.2 a fused-silica still 0.2 1.1 * Treated with ammonia solution. t Less than the detection limit. metric measurements using a 2 cm ce11.20 The proposed method is much more sensitive and requires smaller amounts of sample solutions. Determination of Silicic Acid in Highly Purified Water The results of the analysis of highly purified water produced by four different purification processes are shown in Table 4.Miwa et aZ.3 have reported that the contents of soluble and total silicate-silicon in water purified with the Milli-Q system could be determined by an absorptiometric heteropoly blue method and inductively coupled plasma atomic emission spectrometry after a 50-fold preconcentration by freezing, and that water purified with the Milli-Q system followed by sub-boiling distillation using a fused-silica still was still contaminated with about 1 pg dm-3 of insoluble silicate- silicon. The data obtained with the proposed method compare very favourably with those reported by Miwa et aZ.3 The total amounts were determined after decomposing the insoluble silicate-silicon with ammonia solution.The electronic industry grade ammonia solution and the water samples were placed separately in a closed vessel and the ammonia vapour was allowed to diffuse into the water samples. Contamination of the silicate from the decomposing agent could therefore be suppressed.23 The ammonia could be eliminated from the sample solutions if sulphuric acid and the sample solutions were placed separately in a closed vessel for 1 d at 60 "C: the ammonia concentration in the sample solutions could be lowered from 3.7 to 0.001 mol dm-3. The applicability of the proposed method to the analysis of water samples treated according to the procedure described above is shown in Table 4. Although repeated measurements were not carried out, the errors are expected to be within k O . 1 pg dm-3 of Si.With the flow method of gel-phase absorptiometry, on-line concentration of the molybdosilicate-Malachite Green ion- associate species and on-line absorptiometric determination are carried out simultaneously. As a result, silicic acid can be determined at pg dm-3 or lower levels in 4-5 water samples in less than 1 h. The authors thank Dr. K. Matsumi, Dr. T. Yuasa and Dr. T. Okuda for helpful discussions and encouragement throughout this work. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 References Mykytiuk, A., Russell, D. S., and Boyko, V., Anal. Chem., 1976,48, 1462. Subramanian, K. S., and Chakrabarti, C. L., Prog. Anal. At. Spectrosc., 1979, 2, 287, Miwa, T., Noguchi, Y., and Mizuike, A., Anal. Chim. Acta, 1988, 204,339. Motomizu, S., Oshima, M., and Ikegami, T., Analyst, 1989, 114, 1679. Yoshimura, K., Motomura, M., Tarutani, T., and Shimono, T., Anal. Chem., 1984, 56,2342. Yoshimura, K., Anal. Chem., 1987, 59, 2922. Lazaro, F., Luque de Castro, M. D., and Valchrcel, M., Anal. Chim. Acta, 1988, 214, 217. Yoshimura, K., Bunseki Kagaku, 1987,36,656. Yoshimura, K., Analyst, 1988, 113, 471. Yoshimura, K., Matsuoka, S., and Waki, H., Anal. Chim. Acta, 1989,225, 313. Lazaro, F., Luque de Castro, M. D., and Valchrcel, M., Anal. Chim. Acta, 1989,219, 231. Yoshimura, K., Nawata, S., and Kura, G., Analyst, 1990, 115, 843. Lacy, N., Christian, G. D., and RfitiEka, J.,Anal. Chem., 1990, 62, 1482. Motomizu, S., Oshima, M., and Ojirna, Y., Anal. Sci., 1989,5, 85. Motomizu, S., Wakimoto, T., and Tdei, K., Analyst, 1983,108, 361. Iwasaki, I., Bunseki Kagaku, 1960, 9, 184. Yoshimura, K., and Waki, H., Talanta, 1985, 32, 345 (and references cited therein). Rieman, W., and Walton, H. F., Ion Exchange in Analytical Chemistry, Pergamon Press, Oxford, 1970. Hatano, H., Kagaku no Ryoiki, Zokan, 1969, 88, 2 . JIS KO555 (1990) , Testing Method for Determination of Silica in Highly Purified Water, Committee for Japan Industrial Stan- dard, Tokyo, 1990. Okada, T., and Kuwamoto, T., Anal. Chem., 1985,57,258. Suzuki, M., Atomic Absorption Spectrophotometry, Kyoritsu, Tokyo, 1984. Mizuike, A., Enrichment Techniques for Inorganic Trace Analysis, Springer-Verlag, Berlin, 1983. Paper 0/05567K Received December 11 th, 1990 Accepted March 25th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600835

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, AUGUST 1991, VOL. 116 841 Differential Conductimetry in Flow Injection. Determination of Ammonia in Kjeldahl Digests Jarbas Jose Rodrigues Rohwedder and Celio Pasquini" lnstituto de Quimica, Universidade Estadual de Campinas, CEP 13081, C. P. 6154, Campinas, Sa'o Paulo, Brazil A differential conductivity meter has been constructed for use in flow injection (FI) systems. The instrument employs a Wien bridge to generate a sinusoidal alternating electrical potential (2 V, 1.4 kHz) applied to a measurement bridge containing two twin conductance flow cells. The difference in the conductance between the cells is monitored. The instrument can be used to follow the small changes in conductance that occur in samples in which a high background ionic concentration is found. This facility was used for the determination of ammonia in Kjeldahl digests. The de-ionized water acceptor stream, previously used in the FI system, was replaced by a 5 x 10-4 mol dm-3 HCI solution and the decrease in the conductance caused by the reaction of hydrogen ions with ammonia was monitored.Linear calibration graphs were obtained in the range 0.5-25 mmol dm-3 of ammonium ion and samples can be processed at a rate of 90 h-1. The relative standard deviation for the peak heights of the FI signals is 1 % or lower. The differential system was briefly studied for the determination of carbon dioxide by using NaOH solution as the acceptor stream. Keywords: Differential conductivity meter; ammonia determination; flow injection; nitrogen in Kjeldahl digests; carbon dioxide determination Conductimetry has recently been used together with flow injection (FI) systems for the determination of compounds that can be converted into a molecular form and diffuse through a polytetrafluoroethylene (PTFE) membrane.1-4 The diffused species is received in a de-ionized water stream where it ionizes to a certain extent causing an increase in conduc- tance.The method is very suitable for the determination of ammonia and carbon dioxide, but it was desirable that some of the shortcomings observed should be eliminated. The calibration graphs for the determination of ammonia and carbon dioxide are not linear because these substances are weak electrolytes in aqueous solution. An exponentially decreasing curve is obtained for the change in the conductance caused by the increase in the concentration of the analyte in the sample. 1 An additional number of standard solutions should then be used in the calibration procedure if a wide range of concentrations are expected in the samples. This problem could be overcome if the change in the conductance was caused by a reaction which could quantitatively convert the diffused species into a product causing a change in the total ionic mobility of the solution.Further, the advantage of low-cost instrumentation? necessary to conductimetric me- thodology,1-4 is partly negated by the requirement for the use of good quality de-ionized water as the acceptor stream. It is, therefore, desirable to work with acceptor streams that could contain a high concentration of ions in order to apply selective and quantitative reactions to the diffused species or simply to allow for the use of water of lower quality.Commercial conductivity meters have a restricted use because they are usually designed to work with only one flow cell. The use of solutions with high ionic concentrations requires the use of a high range conductance scale in the instrument. Small changes cannot be observed owing to the high signal attenuation imposed. Further, small changes in the conductance cannot be amplified as these will take place over a high background signal. In order to find a solution to these problems, a differential conductivity meter was constructed. The instrument was used to amplify the difference in the conductance between two twin flow cells in an FI manifold which was employed for the determination of ammonia in Kjeldahl digests.This paper * To whom correspondence should be addressed. describes the instrument the construction of the conductance flow cells and the methodology for the determination of ammonia. Experimental Differential Conductivity Meter Fig. 1 depicts an electronic diagram of the differential conductivity meter. The circuit was supplied by a standard d.c., +12 V, 250 mA power supply and can be divided into three main parts. The first part is the source for the a x . sinusoidal signal applied to the conductance measurement bridge. This unit is based on a Wien bridge.5 The values selected for the capacitors and resistors allow the bridge to oscillate at a frequency of 1.4 kHz. The signal amplitude can be adjusted by the variable resistor P1 and a value of 2 V (peak-to-peak) was chosen for the experiments described here.The frequency of the sinusoidal signal can be easily modified by changing the passive elements of the Wien bridge. The operational 741 amplifier (OP,) employed can couple with frequencies in the range 50 Hz-50 kHz for a 2 V peak-to-peak signal amplitude. The second operational 741 amplifier (0P2) is simply a buffer stage to supply a current to the bridge. The maximum output current from this unit is about 20 mA as specified for the 741 amplifier. The second part is the measurement bridge associated with an instrumentation amplifier used to acquire the differential signal between the two bridge arms. One of the arms, which contains the cell C1, is the reference. The solution flowing through this cell must have a constant ionic composition. The other arm is the indicator arm and it contains the cell C2 used to sense any change in the ionic composition of the fluid during the determination procedure. The difference in the electrical voltage between the arms is monitored by the instrumentation amplifiers constructed using 3140 field-effect transistor (FET) operationals (OP3, OP4 and OPS).The voltage difference gain can be adjusted with the potentiometer P4. This stage also contains a circuit for the adjustment of the bridge sensitivity and for finding the balance point. The circuit is based on another 741 amplifier (OPb), this time used as a comparator,6 and on a simple rectifier circuit.Firstly, the user must adjust the maximum sensitivity by switching the circuit to point A. A comparison between the voltage at the reference arm of the842 ANALYST, AUGUST 1991, VOL. 116 1 (c) t I 10 kQ I kQ 34 kQ 470 kQ 180 kQ Fig. 1 Circuit diagram of the differential conductivity meter. (a) Wien bridge; (b) measurement bridge with instrumentation amplifier and circuit for sensitivity and balance adjustment; and ( c ) precision rectifier and output buffer. For details see text Fig. 2 Conductimetric flow cell: A, electrical contact wire; B, screw for coupling contact wires; C, O-ring; D, screws for coupling the FI manifold tubing; E , acrylic block; F, stainless-steel disc electrode; G, rubber spacer; H, O-ring for preventing contact of the electrodes with the thermal equilibration bath; and I , mounting guide pin bridge with half of the amplitude of the sinusoidal signal applied to the bridge can be accomplished.The potentiometer P2 is adjusted until the light emitting diode (LED) is on, if it was previously off, or, otherwise, until the LED is off. The balance point can then be found by switching to point B, comparing the voltage in the reference arm with that in the indicator arm. Potentiometer P3 is now utilized; the same LED and its state (ordoff) are used to indicate the equilibrium condition. The capacitor values used for this stage are suitable for the working frequency of 1.4 kHz. Finally, the third part consists of a circuit based on another 741 amplifier (OP7), capable of rectifying the a.c. signal obtained from the previous stage, avoiding the loss of small signals caused by the diode voltage drop.6 The rectified signal is then sent to a final buffer amplifier (OP8) and from there to an external signal monitor such as a chart recorder.Flow Cells Fig. 2 shows how the twin flow cells were constructed. The stainless-steel electrodes and the contact wires are housed inside two acrylic blocks, avoiding contact with the water of the isolated water-bath used in the FI manifold. Essentially, the cell constant is determined by a spacer (G) placed between the two metallic discs. The spacer used here was a 3cm diameter styrene-butadiene rubber disc, 0.5 mm thick, with a central hole 20mm long and 2mm wide. It is very easy to reproduce the spacer dimensions. Therefore, it is also easy to construct cells with matching cell constants. Other Apparatus and FI Manifold The FI manifold used for the determination of nitrogen in Kjeldahl digests is illustrated in Fig.3. It was constructed using polyethylene tubing of 0.8mm i.d. An Ismatec MP13 GJ4 peristaltic pump and Tygon pumping tubes were used. A 5 I isolated water-bath was used to allow a long period of operation, independent of the external temperature.' The same diffusion unit previously described, using the same commercial PTFE tape, was also used1 for the determination of ammonia. The FI manifold employed for carbon dioxide was the same as that for ammonia, except that it had an enlarged diffusion cell similar to that described previously.3 The conductimetric FI signals were registered on an ECB potentiometric reccrder.Reagents, Samples and Standard Solutions All reagents and standard solutions were prepared using freshly de-ionized water. Ammonium standard solutions were prepared daily by suitable dilution of a 0.05 mol dm-3 ammonium sulphate stock standard solution. Ammonium standard solutions used for the analysis of Kjeldahl digests were prepared so that they had the same sulphuric acid composition as the samples subjected to the digestion pro- cedure. Therefore, these solutions were 1.44 mol dm-3 in sulphuric acid. Carbonate standard solutions were prepared daily by suitable dilution with de-ionized water from a 0.1 mol dm-3 sodium carbonate stock standard solution. Samples of vegetable tissues were provided by the Instituto Agronomico de Campinas and were subjected to the following digestion procedure. To 200mg of the dry material in a digestion flask were added 8 ml of concentrated sulphuric acid and 2 g of a mixture prepared to contain 100 g of potassiumANALYST, AUGUST 1991, VOL.116 A R R 843 I 2.4 1.3 1.3 ml min-1 Conductivity meter - sulphate, 10 g of copper sulphate and 1 g of selenium. The flask was placed in a digestion block for 4 h. After digestion, the flask was cooled to ambient temperature and the volume made up to 100 ml. Results and Discussion The differential flow conductivity meter was first evaluated in relation to the equality of the reference and indicator arms. By following the signals from the two arms, with the help of an oscilloscope, it was observed that both the signals were in the same phase.This shows that there is a very good similarity between the capacitance component of the two cells and that of the two bridge arms. Small phase displacement was observed only when relatively very concentrated ionic solu- tions were pumped through the cells. At the selected HCI or NaOH solution concentration used, the signals were always in the same phase. Therefore, no capacitance adjustment was required, although the instrument can, if necessary, be used to apply a small variable capacitance in parallel with the indicator cell to correct for a small phase shift. The difference in the resistance of the two cells, when a 5.0 X 10-4 mol dm-3 HC1 solution is pumped through, is 0.8%, the values for the resistance being 394 and 397 Q.The performance of the circuit used to find the maximum sensitivity and the balance point of the measurement bridge was also evaluated. It was observed that the circuit could be used to find the maximum sensitivity condition accurately and the balance point could be reached with a maximum difference of 3.0 mV in relation to the true balance point as shown by the oscilloscope. The drift and noise of the baseline when a 5.0 X 10-4 mol dm-3 HCl solution was pumped through the system at a rate of 2.4 ml min-1 with the cells immersed in the isolated water-bath of the manifold shown in Fig. 3 were estimated to be 0.3 mV h-* and 0.01 mV, respectively. As the bridge employed in the measurement circuit has a restricted linear range of response in relation to the change in the resistance of the indicator flow cell, it was necessary to find the range of the output voltage given by the circuit which was directly proportional to the cell resistance. The minimum gain was utilized in the conductivity meter.The flow cells were replaced by potentiometers and the resistance was adjusted to the same value in order to resemble the resistance of the cells when filled with de-ionized water or 5 X 10-4 rnol dm-3 HCI. The potentiometers were then set to various values simulating the changes observed in real measurements (increasing the resistance when HCI was used as the acceptor stream or decreasing the resistance when de-ionized water was employed instead). The results show that, under the above conditions, the instrument output voltage is linearly related to the change in the resistance in the range 0-80 mV.This change in the output voltage corresponds to about a 40% increase or decrease of the cell resistance in relation to the value observed > 400 E 3 Q +- 300 L +- a E, 200 2 100 +- .- > .- +- 0 0 0 5 10 15 20 25 NH4+ concentrationhmol dm-3 I 1 > E 5 300 3 Q 0 4- & E 200 > .- 4- 2 100 D C 0 - 0 2 4 6 8 1 0 C032- concentration/mmol dm-3 100 > E 3 75 3 Q 0 a, L w 50 5 4- .- > .- 4- 25 C 0 Fig. 4 Calibration graphs for the determination of (a) ammonium and (b) carbonate using several acceptor streams in the differential system. A, De-ionized water; B, distilled water; and C, 5 x 10-4 mol dm-3 HCI solution for ammonium and 1 x 10-3 rnol dm-3 NaOH solution for carbonate when the bridge is initially balanced. Beyond this limit, the output voltage shows a positive or negative deviation from linearity with a further decrease or increase of the resistance, respectively. Fig.4 presents calibration graphs obtained for ammonium and carbonate standard solutions employing various acceptor streams in the FI manifold shown in Fig. 3. The reagent stream for carbonate was a 0.5 rnol dm-3 sulphuric acid solution. It can be seen that the use of a strong electrolyte (HCl or NaOH) which can promote a quantitative reaction with the diffused species can effectively linearize the calibration graph. The sensitivity of the measurement is, however, decreased as the change in relative conductance taking place in the acceptor stream is lower for a stream that contains an initially high ionic concentration than for that containing de-ionized or distilled water.Nevertheless, the conductance measurements, in the concentration range of interest for the Kjeldahl digests, can be obtained even for a minimum gain of the conductivity meter with very good reproducibility. The sensitivity for the determination of carbonate is lower than that for ammonia. In order to increase the analytical signal for carbonate, a large diffusion cell was employed. The lower sensitivity observed could be attributed to a smaller decrease in the conductance caused by the decrease of the OH- concentration in the acceptor solution, or to a decrease in the permeation of carbon dioxide through the membrane, or to the low rate of hydration of carbon dioxide.However, further studies are necessary in order to establish the relative contribution of each factor. The advantages of linearization are mainly related to the small number of standard solutions required in the calibration procedure and to the data treatment which can be carried out very easily. The linear range of the calibration graphs can be increased to higher concentrations by increasing the concen-844 ANALYST, AUGUST 1991, VOL. 116 0 0.5 1.0 1.5 2.0 H2S04 concentration/mol dm-3 Fig. 5 Effect of the sulphuric acid concentration on the peak height of the conductimetric signal of ammonium standard solutions. Ammonium concentration: A, 2.0; and B, 5.0 mmol dm-3 tration of HCI or NaOH in the acceptor stream. The upper limit of the linear range is determined by the acceptor concentration which, for ammonium in the FI manifold used here, was estimated to be approximately two hundredths of the highest ammonium concentration.The most suitable acceptor stream for the determination of low ammonia and carbonate concentrations is de-ionized water as it can provide very good sensitivity and because the calibration graphs tend to exhibit linear ranges when the concentration of the diffused species is reduced.3 The dif- ferential conductivity meter allows, however, the use of distilled water as the acceptor fluid with a decrease in the sensitivity that is still acceptable. The previous requirements for good quality de-ionized water and for an in-line polishing column can then be relaxed. The procedure for the determination of ammonia in Kjeldahl digests' was re-evaluated for differential conduc- tance measurements.The conductance change in this instance is due to the decrease in the total ionic mobility of the acceptor stream caused by the net reaction: H+(aq) + N H 3 W + NH4+(aq) The effect of various substances present in Kjeldahl digests was investigated. Copper sulphate (0.8-5 g 1-*), selenium (0.14.4 g 1-1) and potassium sulphate (10-40 g 1-1) had no significant effect on the peak height obtained for 2 and 5 mmol dm-3 ammonium standard solutions containing 1.4 mol dm-3 sulphuric acid. The results of changing the sulphuric acid concentration are shown in Fig. 5. This is the most important parameter that has to be considered in order to ensure good accuracy for the determination.This is because the analytical signal is partly affected by the acid concentration and because the acid is lost in different amounts using the various digestion procedures. The titration of eight solutions obtained from the standard digestion procedure described under Experimental showed that the acid concentration can range from 1.40 to 1.42 mol dm-3. An investigation into the effect of the acid content at a concentration ten times lower than in the original sample showed that its effect is negligible in the range 0.1- 0.25 mol dm-3. An alternative procedure could then be used to overcome the effect of changing the acid concentration of the sample in those situations where the final acid concentra- tion is not known or cannot be maintained in a narrow range.In this procedure, the ammonium standard solutions should be prepared to be 0.14 mol dm-3 in sulphuric acid and the samples should be diluted ten times before being injected into the FI manifold. Fig. 6 shows a typical calibration run for the determination of ammonium in Kjeldahl digests. The precision of the signals and the stability of the baseline can be evaluated. The samples can be processed at a rate of 90 h-1 with negligible carryover. 8 iii ii I 20 min Time - Fig. 6 Typical run for the analysis of Kjeldahl digests. From left to right are shown the signals for five ammonium standard solutions, the signals for ten samples and the signals for the same ammonium standards in reverse order. The numbers over the peaks are the ammonium concentrations in the standard solutions in mmol dm-3.All measurements were made in triplicate Table 1 Comparative results for the determination of nitrogen in Kjeldahl digests of plant tissue samples Nitrogen/mmol kg-1 Sample Interlaboratory* Proposed method? 1 314 f 22 311 f 19 2 1769 k 69 1720 & 69 3 1094 ? 55 1080 * 21 4 1092 f 46 1047 _+ 21 5 2379 f 99 2398 f 72 6 1224 f 41 1206 k 19 * Best mean results of 72 interlaboratory determinations together ? Results are expressed together with the calculated standard with the standard deviation.' deviation of five measurements. For details see text. The results for the determination of the nitrogen content in six samples of vegetable tissues are shown in Table 1. The nitrogen content in the digested samples was determined five times by employing the proposed conductimetric method using five calibration graphs obtained on different days over a period of 1 week.For comparison purposes the same number of significant figures found in an interlaboratory report7 were used when expressing the results of the conductimetric method. The correlation between the results obtained with the conductimetric method and the average of 72 interlaboratory values reported for the samples7 can be expressed by: c, = (-26.6 k 0.9) + (1.006 2 0.018) ci, where c, is the nitrogen concentration in mmol kg-1 found by the proposed conducti- metric method and ci that for the interlaboratory results. The correlation coefficient is 0.9993 and the error of estimate is 28.3 mmol kg-1. The mean results for the conductimetric method show a slight negative systematic error which can be attributed to the digestion procedure.In fact, the results of standard additions, carried out to increase the ammonium concentration of the sample solutions in increments of 1 .O mmol dm-3, showed a mean recovery of 101 .O% . This demonstrates that there is no matrix effect on the conducti- metric method. Conclusion The differential conductivity meter described in this work can improve the determination of molecular species such asANALYST, AUGUST 1991, VOL. 116 845 ammonia and carbon dioxide, after their diffusion through a membrane. The main advantage is the linearization of the calibration graph. Selective reactions are therefore possible in both sides of the FI manifold, which might lead to the development of new and more selective Flkonductimetric methodologies. Further, the differential conductivity meter could be used to extend the conductimetric determinations directly to samples with a high background ionic concentration. In order to perform this type of determination the signal-to-noise ratio should be improved in order to measure the very small conductance changes that will be observed. The authors are grateful to 0. C. Bataglia for providing the plant tissue samples. References 1 Pasquini, C., and de Faria, L. C., Anal. Chim. Acta, 1987,193, 19. 2 Jardim, W. F., Pasquini, C., Guimariies, J. R., and de Faria, L. C., Water Res., 1990, 24, 351. 3 de Faria, L. C., and Pasquini, C., Anal. Chim. Acta, 1991,245, 183. 4 de Faria, L. C., Pasquini, C., and Neto, G. O., Analyst, 1991, 116, 357. 5 Horowitz, P., and Hill, W., The Art of Electronics, Cambridge University Press, Cambridge, 8th edn., 1987. 6 Malcolme-Lawes, D. J., Microcomputers and Laboratory Instrumentation, Plenum Press, New York, 2nd edn., 1988. 7 International Plant-Analytical Exchange, Department of Soil Science and Plant Nutrition, Wageningen Agricultural Univer- sity, The Netherlands, Bimonthly Report 90.3 (May-June), 1990, p. 13. Paper ll00658D Received February 12th, 1991 Accepted April 17th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600841

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, AUGUST 1991, VOL. 116 847 Determination of the Platinum, Rhenium and Chlorine Contents of Alumina-based Catalysts by X-ray Fluorescence Spectrometry Rao. V. C. Peddy, G. Kalpana and Valsamma J. Koshy Research Centre, Indian Petrochemicals Corporation Limited, Baroda-397 346, India N. V. R. Apparao, M. C. Jain and R. V. Patel Research Centre, Indian Oil Corporation, Faridabad, India An X-ray fluorescence (XRF) method is described for the determination of the platinum, rhenium and chlorine contents of y-alumina supported catalysts. Calibration graphs were linear in the range 0.02-0.80, 0.20-1 .I0 and 0.1-1.0% m/m for platinum, rhenium and chlorine, respectively. In order to determine the accuracy of the proposed method, the results obtained were compared with those given by ultraviolet-visible (UVNIS) spectrophotometry, inductively coupled plasma atomic emission spectrometry (ICP-AES), electrothermal atomic absorption spectrometry (ETAAS) and energy dispersive X-ray analysis (ED-SEM).Statistical analysis showed no significant errors at the 95% confidence interval for the chlorine content obtained by XRF, UVNlS spectrophotometry and ED-SEM. The determination of platinum in monometallic catalysts was carried out by XRF and the results obtained were compared with those given by ICP-AES, ETAAS and UVNlS spectrophotometry. Similarly, for bimetallic catalysts, the results obtained using XRF and UVNlS spectrophotometry for the determination of platinum and rhenium were compared. Statistical evaluation showed that there was no significant bias between the analytical methods used.Keywords: X-ray fluorescence spectrometry; platinum; rhenium; chlorine; catalyst Platinum supported catalysts have the ability to rearrange and transform the molecular structure of hydrocarbons; hence they have become very important in the petrochemical industry. The determination of platinum in catalysts requires a higher accuracy and precision than the normal analysis of materials owing to economic considerations. In the produc- tion of reforming catalysts on a commercial scale, small variations in the platinum assay will lead to heavy financial losses either by the producer or by the customer. Because of its importance, there are many different methods and tech- niques described in the literature for the determination of platinum.Most of these methods involve instrumental tech- niques: e.g. , wavelength dispersive X-ray fluorescence spec- trometry,1-4 X-ray Diffraction (XRD) ,5 atomic absorption spectrometry (AAS) ,637 emission ~pectrography~s.9 neutron- activation analysis10 and spectrophotometry .11-14 However, little work has been carried out on the determination of the elements present in bimetallic reforming catalysts. For work on catalyst development, a fast and reliable method of analysis is required, particularly when there is a need to analyse various batches of catalysts. X-ray fluorescence (XRF) provides an analytical facility that meets most of the require- ments for the determination of the elements present over a wide range of concentration levels in alumina-based cata- lysts.15 The reliability of XRF spectrometric results depends on the extent of correlation between the measured XRF intensities of the sample and calibration standards. The main source of error is likely to be matrix differences between the samples and standards. The aim of this work was to develop an XRF method for determining the platinum, rhenium and chlorine contents of y-alumina supported catalysts. As there were no certified reference materials available for these elements, synthetic standards were prepared in this laboratory and the results obtained with the XRF method were compared with those given by other complementary techniques such as ultraviolet- visible (UVNIS) spectrophotometry , AAS, inductively coupled plasma atomic emission spectrometry (ICP-AES) and energy dispersive X-ray analysis (ED-SEM) .Experiment a1 Regents and Chemicals All reagents used were of analytical-reagent grade. Doubly distilled water was used throughout for dilution purposes. The y-alumina spheroids used for preparing the catalyst standards were 99.99% pure (Condea, USA). Stock solution of platinum, lo00 pg ml-1. Prepared by dissolving 0.5000 g of pure platinum wire in 20 ml of aqua regia [HC1-HNO3 ( 3 + l)]. Oxides of nitrogen were removed by three successive evaporations with concentrated HCl. The residue was then dissolved in 10ml of HCI, the solution transferred into a 500ml calibrated flask and diluted to volume with water. Stock solution of rhenium, lo00 pg ml-1. Prepared by weighing 1.3007g of Re207 (Alfa Products) in a beaker, adding 25 ml of concentrated HCI to dissolve the contents and diluting to 11 in a calibrated flask.Stock solution of chloride, 10mgml-1. Prepared by dis- solving 1.65 g of dried sodium chloride (Merck) in water and making up to 100 ml in a calibrated flask. Preparation of Standards y-Alumina was used as the support for preparing the catalyst standards. The support was dried at 200 "C for 4 h to remove moisture and volatile matter. Platinum standards were pre- pared to achieve a final concentration of platinum in the range 0.02-0.80% m/m on a 20 g batch of y-alumina when impreg- nated using the pore volume technique. 16 Rhenium standards were prepared in a similar fashion from the rhenium stock solution to achieve a final concentration of rhenium in the range 0.20-1.10% d m .For bimetallic standards, platinum and rhenium solutions were mixed in various proportions from the stock solutions in the working range 1 : 1 and 1 : 2 (by mass) of platinum and rhenium, respectively. They were made up to slightly more than the estimated pore volume of alumina and these solutions were impregnated on 20g batches of y-alu-848 ANALYST, AUGUST 1991, VOL. 116 mina. Each batch was then dried at 200 "C for 4 h. The XRD patterns of these samples were checked for uniform distribu- tion of the active elements and were found to match well with the diffractograms of commercial catalyst samples. For the preparation of chlorine standards , y-alumina was finely powdered in an agate mortar with a pestle, sieved (200 mesh) and dried at 200 "C for 4 h in an air oven. The dried samples (5.0 g) were placed in separate 100 ml round-bottomed flasks. To these, the appropriate volume of the sodium chloride stock solution was added to attain a final concentration of chlorine in the range 0.1-1 .O% d m . The total volume of the solution was adjusted to 15 ml using doubly distilled water.The contents of the flasks were maintained at 40 "C for 4 h with constant stirring and then dried on a Biichi Rotavapor. The final drying stage was carried out in an air oven at 100 "C for 4 h . These standard samples were stored in clean, air-tight bottles, kept in a desiccator. Instrumentation and Measurement Procedure All the measurements were made using a Philips PW 1410 XRF spectrometer and associated equipment.For determina- tions using UVNIS spectrophotometry , absorbance values were measured on a Varian Superscan-3 spectrophotometer with matched 1 cm silica cells. The AAS measurements were carried out using a Varian Techtron 1200 atomic absorption spectrometer with a Varian CRA-63 graphite furnace acces- sory. A Kevex Model 7000-77 energy dispersive X-ray analyser attached to a scanning electron microscope (Jeol JSM 35C) was used for ED-SEM measurements. A Jarrell-Ash Atom Comp 1100-Mark I11 emission spectrometer was used for ICP-AES measurements. All synthetic standards and catalyst samples were finely powdered and sieved through 200 mesh and dried at 200 "C for 2 h in an air oven. Finely ground samples (2.5 g; 200 mesh) were placed in individual spex cells (1.5 x 2.5cm i.d.) and covered with Mylar film.The samples were loaded into the X-ray spectrometer and were compacted in situ for analysis. A portion (2.5 g) of each of these powdered samples was placed in individual spex cells and covered with Mylar film. These cells were loaded into the X-ray spectrometer for analysis. The experimental conditions are summarized in Table 1. X-ray fluorescence intensity data were obtained for all the standards of platinum, rhenium and chlorine. Six replicate measure- ments were made on each sample. The mean measured XRF intensities for each set were subjected to linear regression against the concentration of the corresponding analyte. Catalyst solutions were prepared and the tin(I1) chloride method was applied12 for the determination of platinum by UVNIS spectrophotometry .The reference solutions were treated in the same way as the samples. The absorbance of the orange-yellow platinum(1v)-tin(1r) chloride complex was measured at 403 nm. Similarly, thiourea and tin(r1) chloride were used for the determination of rhenium17 in the platinum- rhenium bimetallic catalysts. The chloride content of the laboratory prepared standards was determined by a spectro- photometric method18 using iron(iI1)-mercury thiocyanate, and these values were used for further calculations. Platinum was separated as the platinum-dithizone complex by extrac- tion into isobutyl methyl ketone and determined by electro- thermal atomic absorption spectrometry (ETAAS). l9 The chlorine contents of the catalyst samples were determined by ED-SEM.20 The chlorine Ka X-ray energy falls between 2.58 and 2.68 keV (window).The determination of platinum in the monometallic catalysts was carried out using ICP-AES at 265.95 nm in the sequential mode. Results and Discussion Evaluation of Laboratory Prepared Standards Acquisition of the XRF data for the laboratory prepared standards was carried out under the experimental conditions listed in Table 1. The LiF analysing crystal was found to separate the platinum and rhenium peaks clearly. Although the normal method of preparing the powdered sample is first to grind it and then to form a pellet at high pressure, it was found in this work that when the finely ground samples were subjected to analysis, reproducible results were obtained [relative standard deviation (RSD) S 1 .O%].Hence the powdered samples were used as such for analysis. The bulk of the scattered continuum emanated from the Mylar window fitted to the base of the sample cup. The spectral background, which was found to be constant for the samples, was subtracted from the peak counts for each measurement. The counts corresponding to the alumina blank were negligible. New calibration graphs were constructed each time by plotting signal intensity against analyte concentration. The time required for the calibration procedure is about 40min. The calibration graphs for platinum, rhenium and chlorine were linear (Fig. 1) for the concentration ranges covered by the standards: the regression parameters are listed in Table 2. The bimetallic catalyst reference material prepared by the procedure described above was analysed by spectro- photometry.17 This reference material was also analysed by XRF using monometallic standards, and it was found that the values for platinum obtained with XRF and those given by UVNIS spectrophotometry were comparable. For rhenium, there was a slight deviation which might be due to inter- element interferences caused by platinum. The values for rhenium obtained by XRF were higher than those afforded by spectrophotometry. In order to account for this deviation, a correction factor ( K f ) was employed based on the equation: where XI and X2 are the rhenium concentrations obtained by XRF and the standard spectrophotometric method, respec- tively, and X3 is the concentration of platinum obtained by XRF.This factor was employed in the determination of rhenium in bimetallic catalyst samples. Table 1 Instrumental parameters of the XRF spectrometer Cr tube 20ldegrees Counting tirne/s Element kV mA Crystal Collimator Line Detector PK* BGS PK* BGS Pt 40 30 LiF (200) Coarse Pt La scs 38.15 -1.15 40 20 Re 40 30 LiF (200) Coarse Re Lor scs 41.74 +1.26 40 20 c1 30 20 PETS Fine CI Kor FCll 66.39 1-1.61 40 20 * PK: Peak. 7 BG: Background. $ SC: Scintillation counter. 5 PET: Pentaerythritol. 7 FC: Gas flow proportional counter.ANALYST, AUGUST 1991, VOL. 116 849 10.1 m z X 4.6 I 0.80 5 25 1 5 0.02 Pt (%) Pt/pg m1-I Wpg ml-1 I 0.76 0.06 0.45 1.05 Re (%) 2 25 0.1 0.9 Re/pg ml-' CI (%) a C 0.24 -(g) r = 0.9989 5.3 0.9 e 2 2 0.03 0.5 5.0 0.1 CVpg rnl CI (%) Fig.1 Pt (ETAAS); (d) Re (XRF); (e) Re (UVNIS); (f) C1 (XRF); ( g ) C1 (UVNIS); and (h) C1 (ED-SEM) Calibration graphs for Pt, Re and C1. (a) Pt (XRF); (b) Pt (UVNIS) (Pt : Re molar ratio, A, 1 :O; B, 1 : 1; and C, 1 :2); (c) Table 2 Regression parameters for calibration graphs Concentration range of No. of Method Element r standards* points XRF UV/VIS ETAAS UVNIS UVNIS XRF UVNIS XRF UVNIS ED-SEM Pt Pt Pt Ptt PtS Re Re c1 C1 Cl 0.9961 0.9999 0.9963 0.9994 0.9971 0.9953 0.9999 0.9945 0.9989 0.9935 0.02-0.8 5-25 1-5 5-25 5-25 0.20-1.1 2-25 0.10-1 .o 0.50-5.0 0.10-1 .o 6 5 5 5 5 6 6 4 6 4 * Values in % m/m for XRF and ED-SEM and in pg ml-* for the t For Pt : Re (1 : 1) bimetallic catalyst standards. S For Pt : Re (1 : 2) bimetallic catalyst standards.other methods. Table 3 Chlorine contents of commercial catalysts determined by using three spectrometric methods Chlorine (Yo m/m) Sample description* XRFt UV/VIS ED-SEM P- 1 P-2 P-3 P-4 P-5 P-6 PR- 1 PR-2 PR-3 0.83 0.14 0.12 0.60 0.12 0.14 0.76 0.72 0.67 0.77 0.11 0.10 0.65 0.11 0.13 0.74 0.78 0.66 0.78 0.13 0.09 0.60 0.10 0.15 0.77 0.71 0.69 * P-1-P-6 are Pt/y-A1203 monometallic catalysts. PR-1-PR-3 are Pt-Re/y-A1203 bimetallic catalysts. n = 6; RSD <1.1%0. Analysis of Catalyst Samples The acquisition of data for the catalyst samples was carried out by following the same procedure as that used for the standards. A comparison of the results obtained for the chlorine contents of commercial catalysts by XRF, UVNIS spectrophotometry and ED-SEM is shown in Table 3.The results obtained by XRF, ED-SEM and UV/VIS spectro- photometry were subjected to an F-test, and at the 95% confidence level the calculated F value did not exceed the tabulated value, indicating the absence of any significant error. A regression analysis21 was performed for the chlorine contents determined by XRF and ED-SEM, and also {or those obtained by XRF and UVNIS spectrophotometry. The values of the correlation coefficients (r) and slopes are given in Table 4. These values were found to be close to unity, which implies that the determination of chlorine in alumina-based catalysts can be performed equally well by any of the three methods. Commercial monometallic catalysts were analysed for their platinum content by using the calibration graph for the synthetic standards.The results obtained by XRF, ICP-AES, ETAAS and UVNIS spectrophotometry are summarized in Table 5. Application of the F-test at the 95% confidence level showed no significant bias. The parameters of a regression analysis of the data obtained by XRF with respect to those obtained with the other three techniques are summarized in Table 4. The value of r in all three comparisons was greater than 0.97, demonstrating that the methods are viable. For bimetallic catalysts, the platinum and rhenium contents were determined using the bimetallic reference material; the results are summarized in Table 6. The correlation coefficients for the platinum (r > 0.97) and rhenium (r > 0.99) data obtained with XRF and UVNIS spectrophotometry (Table 4) indicate that the methods are not significantly different.850 ANALYST, AUGUST 1991, VOL.116 Table 4 Correlation parameters for the different methods of catalyst analysis Slope 1 XRF versus UVNIS Pt* 0.9492 2 XRF versus UVIVIS Re 0.9970 3 XRF versus ICP-AES Ptt 0.9893 4 XRF versus UVIVIS Ptt 0.9756 5 XRF versus ETAAS Ptt 0.9899 6 XRF versus UVIVIS C1 0.9932 7 XRF versus ED-SEM C1 0.9969 Sample No. Method Element * Bimetallic catalyst. t Monometallic catalyst. Intercept 0.0130 0.0012 0.0050 0.0109 0.0049 0.0028 0.0014 r 0.9452 0.9960 0.9894 0.9720 0.9898 0.9924 0.9968 Table 5 Platinum contents of Pt/y-A1203 monometallic catalysts determined by using four different methods Platinum (YO m/m) Sample No. XRF* UVNIS ICP-AES ETAAS P- 1 0.61 0.59 0.60 0.60 P-2 0.36 0.36 0.34 0.37 P-3 0.34 0.32 0.36 0.34 P-4 0.60 0.59 0.60 0.60 P-5 0.33 0.39 0.37 0.38 P-6 0.42 0.40 0.40 0.41 * n = 6; RSD So.7%.Table 6 Platinum and rhenium contents of Pt-Re/y-Al*03 bimetallic catalysts determined by using different methods Platinum (Yo mlm) Rhenium (YO m/m) Sample No. XRF* UVIVIS XRF* UVNIS PR- 1 0.30 0.30 0.24 PR-2 0.30 0.30 0.72 PR-3 0.31 0.30 0.63 PR-4 0.34 0.31 0.26 PR-5 0.26 0.25 0.26 PR-6 0.25 0.21 0.36 PR-7 0.25 0.24 0.59 PR-8 0.27 0.29 0.26 * n = 6; RSD (Pt) S1.0%; RSD (Re) So.7%. 0.28 0.69 0.62 0.25 0.25 0.37 0.59 0.27 Conclusions The results obtained here show that the proposed XRF method is comparable to the conventional methods for determining the platinum, rhenium and chlorine contents of mono- and bimetallic catalysts using synthetic standards.The time required for the XRF analysis of various catalyst batches is very short (6 min for each sample after the calibration has been performed) compared with the other methods in which the solution preparation step is more time consuming. The standards, once prepared, can be used several times in the XRF analysis; this is an additional advantage of the method over UVNIS spectrophotometry, ETAAS and ICP-AES. For the determination of rhenium, the composition of the refer- ence material should be very close to that of the batch samples. The proposed method is non-destructive, sufficiently fast, sample preparation is relatively simple and the sensitivity is adequate for the determination of platinum, rhenium and chlorine in alumina-based catalysts; the method could be used as a quality control method in a production unit. The authors are grateful to Dr.I . S. Bhardwaj for permission to publish this work. They also thank R. H. Patel, N. R. Shah and M. Sunder for their technical assistance. References 1 2 3 Coombes, R. J., and Chow, A., Anal. Chim. Acta, 1977, 91, 273. Lincoln, A. J., and Davis, E. N., Anal. Chem., 1959,31, 1317. Kallmann, S., Talanta, 1976,23, 579. 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Beamish, F. E., Lewis, C. L., and Van Loon, J. C., Talanta, 1969, 16, 1. Van Norstrand, R. A., Lincoln, A. J., and Carnevole, A., Anal. Chem., 1964,36, 819. Potter, N. M., Anal. Chem., 1976, 48, 531. Rubeska, I., and Stupar, J., At. Absorpt. Newsl., 1966, 5, 69. Talalaen, B. M., Zh. Anal. Khim., 1964. 19, 1163. Cooley, E. F., Curry, K. J., and Carlson, R. R., Appl. Spectrosc., 1976,30, 52. Venovking, A. V., Gilbert, E. N., and Mckhailev, V. A., J. Radioanal. Chem., 1977, 36, 359. Sandell, E. B., Colorimetric Determination of Traces of Metals, Interscience, New York, 3rd edn., 1959. Ayres, G. H., and Meyer, A. S., Jr., Anal. Chem., 1951, 23, 299. Conrad, A. J., and Evans, J. K., Anal. Chem., 1960,32, 47. Okubo, T., and Kojima, M., Bunseki Kagaku, 1966, 15, 845. Labrecque, J. J., X-Ray Spectrom., 1980, 9, 28. Castro, A. A., Scelza, 0. A., Bencenuto, E. R., Baronetti, G. T., De Miguel, S. R., and Parere, J. M., Preparation of Catalysts IZZ, Elsevier, Amsterdam, 1983. Olivera, G., Garcia, S., and Bezombe, A., Rev. Fac. Ing. Quim., 1987,47, 35. Koshy, V. J., and Garg, V. N., Talanta, 1987,34,905. Kalpana, G., Koshy, V. J., and Garg, V. N., Indian J. Chem., submitted for publication. Koshy, V. J., Kalpana, G., Rao, K. V., and Garg, V. N., Talanta, in the press. Ripley, B. D., and Thompson, M., Analyst, 1987, 112, 377. Paper 1 I001 66 C Received January 14th, 1991 Accepted April 2nd, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600847

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, AUGUST 1991, VOL. 116 851 Determination of the Insecticide Promecarb by Fluorogenic Labelling With Dansyl Chloride F. Garcia Sanchez and C. Cruces Blanco Department of Analytical Chemistry, Faculty of Sciences, The University, 29071 Malaga, Spain The use of spectrofluorimetry to determine the fluorescent derivative of the insecticide promecarb, following hydrolysis to the corresponding phenol in basic media and subsequent coupling with the labelling agent dansyl chloride, is described and discussed. A study of media of different basicity and of different temperatures for both reactions gave optimum conditions of 20 min for the hydrolysis reaction and 10 min for the labelling reaction at 55 "C in 0.05 mol dm-3 sodium hydrogen carbonate solution with a reagent to insecticide ratio of 12: 1.The effect of the solvent on the formation of the dansyl derivative and on the extraction process was studied using nine and seven solvents, respectively. The use of a mixture of acetone and water (50 + 50, v/v) and an extraction into cyclohexane gave the best results. The minimum detectable concentration of promecarb in the experimental assays was 100 ng ml-1. The error and relative standard deviation at a concentration level of 0.6 pg ml-1 were 9.7 and 10.9%, respectively. Air samples containing promecarb at different concentration levels were analysed. Keywords: Promecarb insecticide determination; air samples; dansyl chloride labelling; spectrofluorimetry The N-methylcarbamate insecticide promecarb (3-isopropyl- 5-methylphenyl methylcarbamate) acts against many pests and is effective both as a contact and a stomach poison.The risks associated with exposure to this type of insecticide are caused by the accumulation of endogenous acetylcholine in human beings, due to the inhibition of cholinesterase. For these reasons, and taking into account the paucity of analytical methods described in the literature for this insecti- cide,' simple and rapid assays are required. Carbamate pesticides have frequently been analysed by thin-layer2 or gas chromatography.3 These procedures, however, are not entirely satisfactory in terms of the sensitivity normally required for residue analysis. Because promecarb does not show native fluorescence, the technique of spectrofluorimetry has not been previously applied to its determination.This difficulty has been over- come in the present work by introducing a highly fluorescent moiety into the pesticide molecule. Such an approach, called fluorogenic labelling, has found much application in the field of amino acid and peptide chemistry4 and also in pesticide residue analysis.5-9 The reagent most often used in such labelling techniques is dansyl chloride [5-(dimethylamino)naphthalene-l-sulphonyl chloride], which reacts with primary and secondary amino groups and phenolic hydroxy groups to form highly fluores- cent derivatives. The extensive application of dansyl chloride is due not only to its ability to react with compounds having an amine or phenol moiety but also to its ability to react with both the amine and phenol hydrolysis products of carbamates, resulting in two derivatives suitable for the quantification of carbamate residues in the low nanogram range.This proce- dure has permitted the determination of numerous carbamate insecticides, 1@12 organophosphate insecticides,*3,14 herbi- cides7.15 and fungicides.16 Similar procedures have been used for determining amino acids,17?18 food additives such as monosodium glutamatel9 or alkaloids in complex phar- maceutical dosage forms.20.21 In the present work, the use of dansyl chloride as a fluorogenic labelling reagent for the determination of the carbamate insecticide promecarb is described. The reaction, based on the formation of a fluorescent derivative of the phenolic hydrolysis product of the carbamate, has been studied in detail and can be carried out in less than 40 min.Experimental Apparatus All spectrofluorimetric measurements were performed on a Perkin-Elmer Model MPF-43A fluorescence spectrophoto- meter, equipped with an Osram XBO 150 W xenon lamp, excitation and emission grating monochromators, 1 X 1 cm quartz cells, a Hamamatsu R-777 photomultiplier and a Perkin-Elmer 023 recorder. Instrument sensitivity was adjus- ted daily, using a Rhodamine B bar as a reference standard. A water-bath circulator (Frigiterm S-382) was used for tempera- ture control. An electric shaker (Selecta) was used for extraction procedures. Reagents A stock solution (1 mg ml-1) of promecarb (>99% pure, Pestanal quality, Riedel-de Haen, Hannover, Germany) was prepared in analytical-reagent grade acetone.A working standard solution of 60 pg ml-1 was prepared from this stock solution by dilution with acetone. The dansyl chloride used (Sigma, St. Louis, MO, USA) was prepared as solutions (1 mg ml-1 and 470 pg ml-1) in ACS spectrophotometric grade acetone (Gold Label; Aldrich, Milwaukee, WI, USA). Sodium hydroxide, sodium carbonate and sodium hydrogen carbonate were of analytical-reagent grade (Merck, Darm- stadt, Germany); solutions of these salts were prepared in de-ionized water. All of the solvents used were of analytical- reagent grade. All pesticides tested were of 99% purity or better and were used without further purification. Reaction Procedure Different volumes (30,21,15,9 and 3 pl) of the stock solution of promecarb in acetone (1 mg ml-1) and 50, 40, 30, 20 and 10 pl of the working standard solution of the same reagent in acetone (60 pg ml-1) were placed in 15 ml test-tubes.A 0.5 ml volume of 0.1 mol dm-3 sodium hydrogen carbonate solution was added and the tubes were loosely stoppered and heated in a water-bath at 55 "C for 20 min. The tubes were then cooled to room temperature and 0.5 ml of a solution of dansyl chloride in acetone (470 pg ml-1) was added. The tubes were heated in the water-bath for a further 10 min at 55 "C and then852 ANALYST, AUGUST 1991, VOL. 116 allowed to cool to room temperature. A 3 ml volume of cyclohexane was added and the tubes were shaken for 1 min. Immediately after sample preparation, when both layers were clearly separated, the fluorescence intensity of the organic layer was measured at an excitation wavelength of 340 nm and an emission wavelength of 485 nm against a reagent blank.Standard Procedure for Air Samples Air samples were collected in a workroom environment (dry temperature, 29 "C; humidity temperature, 16 "C; relative humidity, 52%) of 40.5 m2 (4.5 x 9.0 m) with 37 mm three-body filter cassettes equipped with mixed cellulose ester membrane filters of 0.8 pm pore size [Mine Safety Applicans Co. (MSA), Pittsburgh, PA, USA]. Battery-powered per- sonal sampling pumps (Model Fixt-Flo, MSA) with the capacity to operate at between 1 and 2 1 min-1 were calibrated at flow rates of 1.00 1 min-1 over a period of 24 h. Each filter sample was treated with 25 ml of acetone. After the mixture had been shaken for 30 min at 25 "C and 160 rev min-1, the acetone extract was transferred into a round-bottomed glass funnel and evaporated to incipient dryness in a rotary evaporator at 45 "C.The dry residue was then dissolved in acetone and taken to a final volume of 10 ml with the same solvent. An aliquot of this solution (100 p1) was then treated as described under Reaction Procedure. Results and Discussion Unless indicated otherwise, the conditions employed in the optimization experiments were the same as those employed in the Experimental section. From earlier work in the field of carbamate pesticide analysis,7~"~14~*2~23 it is known that the labelling reaction with dansyl chloride involves two distinct steps: ( a ) hydrolysis of the carbamate; and (b) coupling of the hydrolysis product with dansyl chloride.The coupling reaction proceeds faster than either the hydrolysis of the carbamate or the hydrolysis of the reagent to the corresponding sulphonic acid. As the rate of hydrolysis of dansyl chloride is constant at a constant pH or temperature, only the rate of hydrolysis of the carbamate will govern the yield of the dansyl derivative. A reaction scheme is shown in Fig. 1 for the formation of the derivative of the N-methyl- carbarnate, promecarb. Several reaction procedures were attempted for the hydroly- sis of the insecticide to the corresponding phenol. Different concentrations (0.02, 0.04, 0.06, 0.08 and 0.10 mol dm-3) of sodium hydroxide, sodium hydrogen carbonate and sodium carbonate were tried using a heating time of 30 min (hydrolysis reaction) and 20 min (dansylation reaction) at 45 "C.The fluorescence intensity of the organic layer (benzene) was compared with that of a blank signal. The results (Fig. 2) indicated that, by using sodium hydrogen carbonate solution for the hydrolysis process, stable analytical measurements and a good signal-to-noise ratio were obtained with a constant pH of 9 in the reaction mixture (1 pg ml-1 of promecarb). The dansylation reaction is usually performed in a mixture of water and acetone at pH 9-11.15 This method proved to be favourable with respect to the competitive kinetic rates of the labelling reaction and the hydrolysis of dansyl chloride .24,25 The effect of the solvent on the dansylation reaction was investigated. For these experiments, 15 pl of the stock solution of promecarb in acetone (1 mg ml-1) and 0.5 ml of 0.1 mol dm-3 sodium hydrogen carbonate solution were placed in 15 ml test-tubes and the solution was heated in a water-bath at 45 "C for 30 min.Then, 0.5 ml of each pure solvent and 50 pl of dansyl chloride in acetone (1 mg ml-1) were added and the mixture was heated for a further 20 min at 45 "C. The fluorescence of the organic (cyclohexane) layer was measured and the results obtained are summarized in Table 1. Promecarb Hydrolysed promecarb A OH so2 I 0 I Dansyl chloride S03H Sulphonic acid B Fig. 1 Over-all reaction scheme for A, the hydrolysis; and B, the dansylation of promecarb It can be seen that solvents that are immiscible with water (such as chloroform and isobutyl methyl ketone) give no fluorescence, indicating that dansylation has not taken place.Solvents with high relative permittivities and hydrogen bonding capacity (such as water, ethanol and methanol) give high fluorescence both in the analyte and reagent blank, because of the absence of hydrolysis of the excess of dansyl chloride to the corresponding sulphonic acid. The fluor- escence emission occurs also at longer wavelengths in highly polar solvents as reported previously.26 As the extent of the solute-solvent interaction increases, the emission is shifted to shorter wavelengths. 1,4-Dioxane, despite having a low relative permittivity, is miscible with water and gives, together with acetonitrile and acetone, a good signal-to-noise ratio. It is deduced from this study that a change in the polarity of the reaction medium modifies, to a great extent, the characteristics of the labelling reaction.In order to obtain a good signal-to-noise ratio, the excess of dansyl chloride has to be completely converted into the corresponding sulphonic acid, which is more polar and remains in the aqueous phase instead of passing into the organic layer. After acetone had been chosen as the solvent for the reaction procedure, the ratio of water (buffer) to acetone in the reaction mixture was studied at values of between 20 and 90%. The ratio was found not to be critical, as reported previously.15.27 It was decided to use a mixture of 50% acetone (dansyl chloride solution) and 50% water (buffer solution) for further work. The reaction eliminates HCl but the pH is kept constant by the presence of 0.05 mol dm-3 sodium hydrogen carbonate as buffer.Owing to the partial hydrolysis of the dansyl chloride reagent by the buffer or its consumption as a result of theANALYST, AUGUST 1991, VOL. 116 853 60 a, v) 2 Table 1 Effect of the solvent on the dansylation reaction - Solvent 1,4-Dioxane Chloroform Isobutyl methyl ketone Ethanol Acetone Methanol Acetonitrile Dimethylformamide Water Relative permittivity at 25 "C 2.2 4.8 13.1 20.5 20.7 32.7 35.5 36.7 78.5 h,,/nm 338 345 345 345 340 345 343 338 350 h,,/nm 480 485 490 495 485 500 492 488 495 Ah*/nm 142 140 145 150 145 155 149 150 140 RFI? Anal yte 83 5 1 140 100 164 71 10 25 Blank 4 2 1 128 4 164 4 9 3 A RFI f 79 3 0 12 96 0 67 1 22 * Stokes shift = he, - hex. ? RFI = Relative fluorescence intensity.$ ARFI = RFI analyte - RFIblank- B 0 0.02 0.04 0.06 0.08 0.1 0.12 [NaOH]/mol dm-3 I 1 I 1 I I I 0 0.02 0.04 0.06 0.08 0.1 0.12 [Na2C031/mol dm-3 (c) I 30 A I T , - B 0 0.02 0.04 0.06 0.08 0.1 0.12 [NaHC031/mol dm-3 Fig. 2 Influence of (a) NaOH, (b) Na2C03 and (c) NaHC03 concentrations on the hydrolysis reaction and fluorescence intensity of A, promecarb; and B, a reagent blank formation of side-products, the amount of dansyl chloride added to the reaction mixture could influence the results. Fig. 3 shows the effect of varying the dansyl chloride concentration on the yield of the dansyl derivative for a final promecarb concentration in the reaction medium of 5 pg ml-1. A gradual increase in fluorescence intensity on increasing the dansyl chloride to analyte concentration ratio up to 10 : 1 is observed.A minimum of an 8-fold concentration excess of dansyl chloride over the carbamate is required. Use of a larger excess of dansyl chloride does not affect the results. The rate of hydrolysis of the carbamate increased at higher temperatures as did the rate of the reaction of the reagent with 2 6 10 14 18 Dansyl chloride: analyte Fig. 3 Effect of the amount of dansyl chloride on the formation of the phenol derivative of 5 pg ml-1 of promecarb (shaded area shows the working zone) the hydrolysis product. Fig. 4(a) shows the rates of hydrolysis at different temperatures. These were determined by forming the dansyl derivative of the liberated phenol in the hydrolysis mixture of the carbamate. The amount of dansyl derivative formed is directly proportional to the extent of carbamate hydrolysis.Although the rate of hydrolysis is increased on increasing the temperature from 25 to 55 "C, prolonged heating at 65 "C leads to decomposition of the products. Also, the fluorescence signal was found to be fairly stable for heating times of between 5 and 20 min at 55 "C, but for heating times longer than 20 rnin the signal started to increase. A heating time of 20 min was considered sufficient for complete hydrolysis of the carbarnate, thereby avoiding problems of decomposition. In order to unify the temperatures for the hydrolysis and dansylation reactions, a study of the effect of temperature on the latter was carried out. Fig. 4(b) indicates that at 45 "C, 35 rnin are required for complete stabilization of the fluorescence measured, whereas only 10 rnin are necessary at 55 "C.As a result, the working conditions chosen for the over-all reaction were 20 rnin for the hydrolysis reaction and 10 rnin for the dansylation reaction at 55 "C. These reaction times are considerably shorter than those normally employed by other workers for the determination of similar compounds with dansyl chloride (between 45 and 90 rnin); hence the optimization of the experimental variables has permitted a rapid method for the determination of promecarb (30 min), which is suitable for routine analysis. Fluorescence Phenomena The spectra of the dansylated promecarb hydrolysis products, extracted into cyclohexane, are shown in Fig. 5. The excitation and emission maxima were found to be 340 and 485 nm, respectively. In order to check that the excess of dansyl854 ANALYST, AUGUST 1991, VOL. 116 chloride is hydrolysed and that the hydrolysis products remain in the aqueous phase, a blank test was performed without promecarb, and the organic phase was analysed by measuring the fluorescence intensity in the same wavelength range.As can be seen in Fig. 5, no significant signal was found. The formation of the intensely fluorescent hydrolysis product (the sulphonic acid) makes it necessary to isolate this compound from the derivative of the hydrolysed promecarb. This isolation is usually carried out by solvent extraction into non-polar solvents such as benzene or hexane. The organic phase is then used for fluorescence measurements.Maximum emission of dansyl derivatives occurs between 450 and 580 nm, depending on the type of solvent used. The effect of the solvent on the extraction process was studied using seven solvents. All the solvents were of low polarity and had low relative permittivities and also low solubility in water. Table 2 indicates that an increase in the relative permittivity causes a bathochromic shift of the emission maxima (485 nm, cyclohexane; 504 nm, dichloro- 120 1 00 80 60 > C a .z 40 .G 20 $ 0 4- 0, C a 4 110 .- P 100 - r 4- - 90 80 70 60 50 - _ 0 5 10 15 20 25 30 35 tlmin Fig. 4 Influence of time and temperature on (a) the hydrolysis and (b) the dansylation of 5 pg ml-1 of promecarb with a 12-fold excess of dansyl chloride. A, 55; B, 45; C, 65; and D, 25 "C methane), and an increase in the blank signal.Solvents such as benzene and cyclohexane gave the best signal-to-noise ratio. Cyclohexane was chosen for further work as it is 30 times less toxic than benzene. The reaction proved to be reproducible from one day to the next although fresh standards were always carried through the reaction procedure. The dansylated derivatives were very stable, showing the same fluorescence intensity after storage in a refrigerator for 21 d. This is particularly useful for residue analysis as samples cannot always be extracted and analysed on the same day. Calibration Graph and Repeatability Experiments The relative fluorescence intensity of a 3 ml volume of the organic phase, obtained as described under Reaction Pro- cedure, was found to be a linear function of the promecarb concentration over the range 0.2-10.0 pg ml-1.The precision is good, as shown by the regression coefficient of 0.9980. The equation for this graph is: relative fluorescence intensity = 0.06 [promecarb] -2.70 for the range 0.2-10.0 pg ml-1. The minimum detectable concentration or detection limit, cL, and the lower limit of the dynamic range, cQ, defined by IUPAC28 as CL = 3sB/rn and CQ = 10s~/rn [where SB is the relative standard deviation (RSD) of the blank signal and rn A B 290 330 370 410 450 490 530 570 610 Wnm Fig. 5 Fluorescence spectra of the dansylated phenol moiety of promecarb (solid line) and a reagent blank (broken line) under the same conditions. Promecarb concentration, 5 pg ml- l; acetone-water (50 + 50, vlv); and a 12-fold excess of dansyl chloride.A, Excitation; and B, emission Table 2 Effect of the solvent on the extraction process Dipole Relative moment at permittivity Solubility Solvent 25 "CPD at 25 "C in water* Benzene 0 2.3 0.18 Hexane 0.1 1.9 0.001 Toluene 0.3 2.4 0.051 Dichloromethane 1.1 8.9 1.60 Chloroform 1.1 4.8 0.815 Ethyl acetate 1.9 6.0 8.70 C yclohexane 0 2.0 0.01 *Solubility in water expressed in % d m . 7 TLV = Threshold limit values. $ RFI = Relative fluorescence intensity. 8 ARFI = RFIanalyte - RFIblank. TLV-t lpg ml-1 10 300 100 200 500 25 400 hexlhem lnm 3451490 3381485 3381490 3451490 3451504 345/500 3301490 RFIS Analyte 96 116 74 91 90 87 77 Blank 10 10 5 18 32 22 102 ARFIP 86 106 69 73 58 65 -25ANALYST, AUGUST 1991, VOL. 116 855 Table 3 Characteristics of the proposed spectrofluorimetric method for the determination of promecarb Linear Mean dynamic concentration SA/ CLl rangel found/ Error? RSDS SS* yg ml-l ygml-l pgml-I ygml-I (Yo) (Yo) 4.0 0.1 O .l § O.l$-lO.O 0.6 9.7 10.9 * Expressed in units of relative fluorescence intensity. ? Relative error = 100ts/Xn+. $ RSD = Relative standard deviation. § Values calculated from the expressions: XB + 3sB/m and xB + 10sB/m, respectively. the slope of the calibration graph] and the sensitivity, sA,8 of the determination, together with other analytical characteris- tics, are summarized in Table 3. Repeatability assays were carried out by performing the hydrolysis reaction, and the subsequent dansylation, simul- taneously on seven separate carbamate samples of two different concentrations (0.6 and 5 pg ml-1).The RSD (1OOs/X) and the relative error (100stlVnX) are a measure of the precision and accuracy of the analytical determination, respectively, where s is the standard deviation (%‘X(x - X)2/n - 1) for n = 7, X the mean value of seven samples containing 0.6 or 5 pg ml-1 of carbamate and t the Student’s t value for a 95% confidence limit. An RSD of 5.4% and an error of 4.8% were obtained at the higher concentra- tion level of 5 yg ml-1. At lower concentrations (0.6 yg mi-*), both the RSD (10.9%0) and the error (9.7%) increased. The inherent error in the dansylation step was determined by analysing a set of seven replicate samples containing 5 pg ml- 1 of promecarb and prepared from the same hydrolysed solution.In order to evaluate the inherent error in the hydrolysis reaction, seven aliquots of a 5 pg ml-1 sample from individual hydrolysed solutions were analysed. The first step evaluates the intra-assay precision, whereas the second gives the inter-assay precision of the proposed method. The intra-assay RSD and error were 6.8 and 6.l%, respectively, whereas values of 5.4 and 4.8%, respectively, were obtained for the inter-assay precision. This indicates that the higher percentage of error is caused by the dansylation reaction, as in other types of labelling reactions. The between-day precision was determined by analysing samples containing 5 pg ml-1 of labelled promecarb (hydrolysis products) over a period of seven consecutive days. The RSD and error obtained were 5.8 and 5.2%, respectively. Specificity of Promecarb Determination Owing to the wide applicability of the dansylation procedure for pesticide analysis,10-16 it was considered of particular interest to determine the selectivity of the proposed method for the determination of promecarb in the presence of other pesticides that are susceptible to derivatization with this labelling reagent.The specificity of the proposed spectrofluorimetric method was determined with different concentrations of several pesticides having a chemical structure that made them susceptible to derivatization with dansyl chloride, with or without a previous hydrolysis reaction. The compounds used were the insecticides carbaryl, chlortoluron, fenitrothion, methyl parathion, dimethoate and l-naphthol, the herbicide propham and the plant growth regulator 1-naphthaleneacetic acid.Various volumes of stock solutions of the different potential interferents in acetone (1 mg ml-1) were added to 15 ml test-tubes together with 15 yl of the stock solution of promecarb (1 mg ml-1) in order to obtain different interferent to analyte ratios in the final solution. The solutions were then treated as described under Reaction Procedure. Table 4 Interference study. Analysis of synthetic mixtures Interferent * Chlortoluron (25) Fenitrothion ( 5 ) Propham (50) Methyl parathion (25) l-Naphthaleneacetic acid, NAA Dimethoate (50) l-Naphthol(50) Chlortoluron (25) + 1-naphthol Propham (50) + NAA (25) Dimethoate (50) + chlortoluron Propham (25) + dimethoate (25) Propham (50) + fenitrothion ( 5 ) Dimethoate (25) + fenitrothion Chlortoluron (25) + l-naphthol Carbaryl(2.5) (50) (25) (25) + chlortoluron (25) ( 5 ) + NAA(5) (25)+ NAA(25) Concentration of promecarb foundlyg ml- 5.4 5.3 4.3 4.2 5.6 5.3 4.6 5.2 4.7 5.5 4.7 5.6 5.5 4.6 5.5 Recovery k SD? (%) 101 t 5 107 k 4 86 f 3 83 k 3 114 k 8 107 k 5 92 k 2 104 k 5 95 t 5 110 k 7 95 * 4 112 k 4 109 * 5 92 f 2 110 5 * Promecarb concentration added: 5 pg ml-I.Values in parentheses ‘r Mean value of three determinations. are the concentrations of the interferents in pg ml-I. The criterion for evaluation of interference was a deviation of the fluorescence intensity of ;F k 3s,, where X is the mean concentration value found after a repeatability assay of samples (n = 7) containing 5 yg ml-1 of promecarb and s, is the standard deviation of these measurements.This criterion permits a confidence limit for the concentration measure- ments of between 4.2 and 5.8 yg ml-1, which is equivalent to a confidence of 83.7-116.3% recovery. Results are shown in Table 4. Low recoveries are obtained if the ratio of the labelling reagent to the synthetic mixture to be analysed is below 12. The values shown in Table 4 indicate that the concentration ratio for propham, 1-naphthaleneacetic acid, dimethoate and 1-naphthol is 10 : 1, for chlortoluron the ratio is 5 : 1 and for fenitrothion it is 1 : 1. Carbaryl and methyl parathion, both of which formed derivatives with dansyl chloride, gave a positive interference; hence these compounds must first be separated by chromatographic techniques.The maximum ratio tested for these pesticides was a 10-fold m/m ratio of interferent to promecarb. Analysis of Air Samples Because of their benefits, pesticides have become increasingly important in agricultural production; however, their prolifera- tion in the environment is causing concern for human health and because of their effect on other components such as soil, sediment, water, air, animals and vegetation. In an attempt to measure the amount of environmental contamination related directly to the application of promecarb856 ANALYST, AUGUST 1991, VOL. 116 Table 5 Analysis of environmental air samples Promecarb content k SDt/ Sample No. Volume*/l mg m-3 1 21.3 6 + 1 2 18.9 4.5 + 0.3 3 21.3 6.2 -t- 0.8 4 20.2 6.4 ? 0.9 5 22.5 8 - c l 6 19.5 5.2 + 0.4 * Volume of air sampled.t Mean value of three determinations. to different crops (potato, citrus fruits and fruit trees) by dusting, the analysis of air samples by the proposed spectro- fluorimetric method was carried out. Six air samples were collected from different areas of a workroom environment permeated with promecarb at similar flow rates. After a sampling time of 15 min, total air volumes of between 18 and 23 1 were obtained. The results, expressed in milligrams per cubic metre of promecarb in the environ- ment, and given as the mean values of three replicates of each environmental sample, are presented in Table 5. This investigation was supported financially by The Direccion General de Investigacion Cientifica y Tecnica (Project PB86- 0247).References Gunew, D. S., in Analytical Methods for Pesticides and Plant Growth Regulators, eds. Zweig, G., and Sherma, J., Academic Press, New York, 1980, vol. XI, p. 141. Mendoza, C. E., and Shields, J. B., J. Chrornatogr., 1970, 50, 92. Katz, S. E., and Strusz, R. E., J. Agric. Food Chem., 1969,17, 1409. Pataki, G., Techniques of Thin-layer Chromatography in Amino Acid and Peptide Chemistry, Ann Arbor Science Publishers, Ann Arbor, MI, 1968. Fink, D. W., TrAC, Trends Anal. Chem. (Pers. E d . ) , 1982, 1, 254. 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Seitz, W. R., CRC Crit. Rev. Anal. Chem., 1980, 8, 367. Lawrence, J. F., and Laver, G. W., J. Assoc. Off. Anal. Chem., 1974,597, 1022. Garcia Sanchez, F., and Cruces Blanco, C., Anal. Chem., 1986, 58, 73. Garcia Sanchez, F., Cruces Blanco, C., Hernandez Lopez, M., Marquez Gomez, J. C., and Carnero, C., Anal. Chim. Acta, 1988,205, 149. Lawrence, J. F., and Frei, R. W., Anal. Chem., 1972,44,2046. Frei, R. W., and Lawrence, J. F., J. Chromatogr., 1972,67,87. Lawrence, J. F., and Leduc, R., J. Chromatogr., 1978,152,507. Lawrence, J . F., Renault, C., andFrei, R. W., J. Chromatogr., 1976,121, 343. Frei, R. W., Lawrence, J. F., and LeGay, D. S., Analyst, 1973, 98, 9. Frei-Hausler, M., Frei, R. W., and Hutzinger, O., J. Chromat- ogr., 1973, 79, 209. Traore, S., and Aaron, J. J., Talanta, 1981, 28,765. Bayer, E., Grom, E., Kaltenegger, B., and Uhmann, R., Anal. Chem., 1976,48, 1106. Greco, B., J. Chromatogr., 1983, 255, 67. Rhys Williams, A. T., and Winfield, S. A., Analyst, 1982, 107, 1092. Nachtmann, F., Spitzy, H., and Frei, R. W., Anal. Chim. Acta, 1975, 76, 57. Frei, R. W., Santi, W., and Thomas, M., J. Chromatogr., 1976, 16, 365. Frei, R. W., and Lawrence, J. F., J. Chromatogr., 1971,61,174. Lawrence, J. F., and Frei, R. W., J. Chromatogr., 1972,66,93. Seiler, N., and Niechmann, M., Fresenius 2. Anal. Chem., 1966, 220, 109. Seiler, N., J. Chromatogr., 1971, 63, 97. Chen, R. F., Arch. Biochem. Biophys., 1967, 120,609. Seiler, N., and Wienchmann, M., Progress in Thin-layer Chromatography and Related Methods, eds. Niedenvieser, A., and Pataki, G., Ann Arbor Science Publishers, Ann Arbor, MI, 1970, vol. I, p. 133. IUPAC, Guidelines for Data Acquisition and Data Quality Evaluation in Environmental Chemistry, Anal. Chem., 1980, 52, 2242. Sommer, L., Langova, M., and Kuban, V., Scr. Fac. Sci. Nat. Ujep Brun. Chem., 1978, 8, 13. Paper 8104298E Received October 28th, 1988 Accepted March 7th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600851

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, AUGUST 1991, VOL. 116 857 Determination of Epinephrine, Norepinephrine, Dopamine and L-Dopa in Pharmaceuticals by a Photokinetic Method Carmen Martinez-Lozano, Tomas Perez-Ruiz, Virginia Tomas and Otilia Val Department of Analytical Chemistry, University of Murcia, Murcia, Spain A study of the photochemical reaction of the Rose Bengal (RB)-ethylenediaminetetraacetic acid system in the presence of epinephrine, norepinephrine, dopamine and L-dopa is presented. The rate of photoreduction of RB is dramatically retarded by small amounts of these catecholamines, which have an inhibitory effect on the excited state of RB, which is the activator of the process. Optimum conditions for the determination of catecholamines in the range of concentration between 5 x 10-6 and 1 x 10-4 rnol dm-3 are described.The proposed method has been applied with excellent results to the determination of catecholamines in p h a r maceu t ica Is. Keywords : Ca tech olamine determination; photo kinetic method Catecholamines are compounds with amines attached to a benzene ring bearing two hydroxy groups. The most import- ant endogenously produced compounds of this group are epinephrine (adrenaline), norepinephrine (noradrenaline) and dopamine. The main sites of production of the catechol- amines are the brain, chromaffin cells of the adrenal medulla and the sympathetic neurons. Epinephrine is quantitatively the most important substance produced by the adrenal medulla, whereas norepinephrine is the major substance liberated by the postganglionic sympathetic nerves.Dopam- ine and norepinephrine have, in general, a marked influence on the vascular system; and epinephrine, which is considered to be the true adrenal medullary hormone, influences meta- bolic processes, especially carbohydrate metabolism. Phar- maceutical preparations containing catecholamines as pure substances and in dosage forms have been available for many years for the treatment of certain neural disorders; for example, L-dopa is extensively used for Parkinson’s disease. The determination of catecholamines in biological speci- mens normally requires the use of trace analysis techniques, mainly chromatography with fluorimetric or electrochemical detection. 1 In contrast, catecholamines that form the active constituents in pharmaceutical preparations are present there- in in relatively large amounts, and increasing efforts are being directed towards the development of simple and reliable analytical methods.Recently, spectrophotometric, fluori- metric, titrimetric and kinetic methods2-9 have been widely used for the determination of catecholamines in pharmaceut- icals. The purpose of the present investigation was to develop a simple and sensitive assay for four catecholamines (epineph- rine, norepinephrine, dopamine and L-dopa) using a photo- kinetic method and to apply the procedure to various dosage forms. The method is based on the strong inhibitory effect of these catecholamines on the photochemical reaction between Rose Bengal (RB) and ethylenediaminetetraacetic acid (EDTA). Experimental Apparatus The light source was a halogen lamp (Sylvania, 24 V, 150 W).The illumination device is very simple and has previously been described.1°311 The photoreduction of RB was monitored with a photometric titration unit (EEL-Unigalvo 200). A Pye Unicam SP8’200 spectrophotometer was used for recording the spectra. Reagents Analytical-reagent grade chemicals and doubly distilled water were used throughout. An aqueous 0.001 rnol dm-3 solution of RB (3,4,5,6- tetrachloro-2’ ,4’ ,5’ ,7’-tetraiodofluorescein, CI 45440) was prepared by dissolving 1.02 g of the product (K and K) in doubly distilled water and diluting to 1 1 in a calibrated flask. Standard catecholamine (0.001 rnol dm-3 epinephrine, norepinephrine, dopamine and L-dopa) solu- tions were prepared from the Sigma product and stored frozen in a dark bottle before use.Working standards were prepared daily from this solution by diluting with doubly distilled water. General Procedure To a 15 ml calibrated flask, add 2 ml of 2 rnol dm-3 acetate buffer pH 5.5,s ml of 0.3 rnol dm-3 EDTA, 2 ml of 1.0 X 10-4 mol dm-3 RB and an appropriate volume of catecholamine solution (standard or sample) to give a final catecholamine concentration between 1 x 10-4 and 1 x 10-5 rnol dm-3. Dilute to the mark with water and transfer into the reaction cell, maintained at 25 k 0.5”C. Remove oxygen from the solution by bubbling pure (99.99%) nitrogen through it for 15 min. Switch on the halogen lamp and measure with a photometric titration unit the time required for the absor- bance to be reduced to 10% of its initial value.Construct a calibration graph by plotting catecholamine concentration versus txlto, where t, is the time required for the photoreduc- tion of the sample, and to the time required for the photoreduction of a catecholamine-free sample. The illumina- tion intensity chosen should give a value of about 100 s for to. Results and Discussion When a solution containing RB and EDTA is illuminated at suitable pH and in the absence of oxygen, photoreduction of the dye occurs and the pink colour disappears hv RBoxidized + EDTA -+ RBreduced + EDTA oxidation products The reaction proceeds at an adequate rate only if the light is sufficiently intense. If air is passed through the colourless solution, the dye is oxidized rapidly, returning to its original pink colour.Fig. 1 shows the absorption spectra of the dye before reduction and then after the oxidation with oxygen or hydrogen peroxide of the leuco RB formed by photoreduc- tion. The three spectra coincide, showing that the RB does not undergo irreversible breakdown during the photochemical reaction. The stoichiometry was determined by adding excess of EDTA, at various pH values, photolysing until the dye was completely decolorized, and titrating the surplus EDTA with standard ZnII solution. The molar ratio found was 1: 1 RB : EDTA.858 ANALYST, AUGUST 1991, VOL. 116 0.50 a C (0 2 2 0.25 2 0 300 400 500 600 700 Unm Fig. 1 Absorption spectra for 1.3 x mot dm-3 RB 1.3 x 10-2 rnol d ~ n - ~ EDTA-acetate buffer pH 5.5. 1, Before the photochemical process (solid line); 2 (line) and 3 (dotted line), after the photochem- ical process and re-oxidation with oxygen and hydrogen peroxide, respectively Effect of Experimental Variables The rate of photoreduction of RB by EDTA is pH dependent.Curve 1 of Fig. 2 shows the result obtained by plotting the rate of photoreduction (expressed in relation to its maximum value) as a function of pH. The initial rate and integration techniques were used for the determination of the order of reaction with respect to RB. The results of both methods show a first-order dependence on the concentration of RB. The over-all reaction order was also determined by the integration method. The results show an over-all reaction order of 2, hence the reaction is also first-order with respect to EDTA.Variations in temperature between 20 and 40°C have very little influence on the rate of the photochemical process. Reaction Mechanism The photochemical reaction between RB and EDTA in the absence of oxygen takes place through a series of collisions between EDTA molecules and the dye molecules in an excited state. The strong influence of pH on the reaction rate (Fig. 2, curve 1) can be explained by assuming that the only reacting species, as occurs with other aminopolycarboxylic acids and dyes,10,12-14 are the ionic forms of EDTA having at least one non-protoned nitrogen atom, and RB in its excited triplet state. The photoreactive state of the dye is probably the triplet state10.12J5 as EDTA does not quench the dye fluorescence although it is oxidized by the activated dye.Further, small amounts of substances such as p-phenylenediamine or iodide do not quench the dye fluorescence either, but both sub- stances dramatically retard the photoreduction of RB by EDTA. Obviously this is due to an energy transfer process between the dye in the triplet state, which has a relatively long life, and the inhibitor. Study of the Photochemical Reaction in the Presence of Catecholamines It was found that the photochemical reduction of RB is strongly inhibited by small amounts of epinephrine, nor- epinephrine, dopamine or L-dopa. The presence of these catecholamines does not quench the dye fluorescence but 2 3 3 5 7 9 11 PH Fig. 2 Rate of photoreduction of RB by EDTA (relative to the rate at pH = 9.5, assigned a value of 100) as a function of pH.1, Without catecholamine; 2, [epinephrine] = 3.3 x lo-’ mol dm-3; and 3, do dopa] = 3.3 x rnol dm-3 I 20 I 1 YA I - - 15 - - - c I v) * c - - I I D I 0.1 0.2 (EDTA]/mol dm-3 0.3 Fig. 3 Influence of EDTA concentration on the rate of the hotochemical rocess. Conditions: [RBI = 1.3 X 10-5 rnol dm-3; rcatecholaminef= 3.3 X 10-5 mol dm-3; acetate buffer pH 5.5. A, Epinephrine; B, norepinephrine; C, dopamine; and D, L-dopa dramatically retards the photochemical reaction. This can also be attributed to an energy transfer from the triplet (T,) state of RB to the singlet (So) state of the catecholamines. In accordance with these findings, it can be assumed that the inhibition is the only effect of the catecholamines on the photochemical process of the RB-EDTA system.There are a number of processes such as intersystem crossing, energy transfer and chemical reaction, whereby absorbed radiant energy can be dissipated by a molecule in the excited state in its return to the initial ground state. Each process has an associated rate constant. When the steady-state hypothesis is applied to the kinetic schemes with and without inhibitor, an expression similar to the Stern-Volmer relation- ship 10716 is obtained: a0 k, [inhibitor] -=1+ a k d + k , [EDTA] (1) where and @ are the quantum yields in the absence and presence of inhibitor, respectively, k, is the rate constant of the energy transfer (quenching) of the triplet state of the RB to the inhibitor, kd is the over-all rate constant of the intramolecular deactivation of the triplet state of RB (phos- phorescence and intersystem crossing) and k , is the rate of the photochemical reaction.For measurements made with a fixed light intensity then @0 V g k , [inhibitor] (2) - -_ - - -1+ @ V , k d + k,[EDTA] v, and vo being the reaction rates with and without inhibitor, respectively.ANALYST, AUGUST 1991, VOL. 116 859 Table 1 Photokinetic determination of catecholamines Inhibition constant/l 04 Concentration Regression No. of Regression Catecholamine dm' mol- * range/mol dm-3 line values coefficient Epinephrine 2.65 6.0 x 10-6-7.0 x lop5 y = 0.98 + 2.65 X 104~ 10 0.9994 Norepinephrine 2.28 1.0 x 10-5-7.0 x 10-5 y = 0.99 + 2.28 X 10% 9 0.9978 Dopamine 2.92 5.0 x 10-6-7.0 x 10-5 y = 0.97 + 2.92 X 104~ 10 0.9985 L-Dopa 1.64 1.0 x io- 5-1.0 x 10-1 y = 1.01 + 1.64 x iwc 10 0.9996 Table 2 Determination of catecholamines in pharmaceutical prepara- tions Amount?/mg Found by Nominal proposed Sample" Source Catecholamine value method Epistaxol Medical Adrenaline 0.5 0.50 f 0.02 Adrenalina Llorente Adrenaline 1.0 1.02 f 0.01 Oculos epilo Frumtost- Zyma Adrenaline 10.0 9.9 f 0 .2 Adrenor Llorente Noradrenaline 20.0 19.6 f 0.1 Dopamina Fides Dopamine 200.0 200.2 & 0.1 * Composition of samples. Epistaxol: adrenaline hydrochloride, 0.5 mg; vitamin P (rutin), 0.2 mg; naphazoline hydrochloride, 0.5 mg; antipyrine, 30 mg; and water 1 g. Adrenalina: adrenaline hydro- chloride, 1 mg; and water, 1 g. Oculos epilo: adrenaline hydrogen tartrate, 10 mg; pilocarpine hydrochloride, 40 mg; and water 1 g. Adrenor: noradrenaline hydrogen tartrate, 20 mg; and water 1 g.Dopamina: dopamine, 200 mg; sodium hydrogen sulphite, 50 mg; and water, 10 g. t Average of four determinations k standard deviation. By operating with a large excess of EDTA so that its concentration can be deemed constant, then (3) vo - = 1 + k [inhibitor] vx This equation allows the determination of epinephrine, norepinephrine, dopamine and L-dopa by measuring the rate of the photochemical process in the presence and absence of the inhibitor. The reaction conditions need to be optimized. Curves 2 and 3 of Fig. 2 show the variation of the rate of the photochemical process in the presence of epinephrine and L-dopa, respect- ively, at several values of pH (similar results are obtained with norepinephrine and dopamine).A pH 5.5 (acetate buffer) was selected as a compromise between a high degree of inhibition and a practical value for the rate of the photochemical reaction. The rate of the photochemical reaction increases with increasing concentration of EDTA, but there is no further increase for concentrations of EDTA >0.15 rnol dm-3 (see Fig. 3). The rate of photoreduction of RB by EDTA in the presence of the catecholamines is only slightly affected by temperature. A temperature of 25 k 0.5"C was chosen as suitable. To summarize, the best experimental conditions for the determination of epinephrine, norepinephrine, dopamine and L-dopa are an EDTA concentration of 0.16 mol dm-3, an RB concentration of 1.3 x 10-5 mol dm-3 and a pH of 5.5 at 25 k 0.5 "C. Under these optimized conditions the inhibition constants [eqn.(3)] for the four catecholamines are given in Table 1. The values of the over-all inhibition constant show that a sensitive determination of these compounds can be carried out by measuring the rate of photoreduction of RB by EDTA. The fixed concentration method was chosen on the criteria of simplicity, sensitivity and linear range. The time required to reduce the concentration to 10% of its initial value was measured. The concentration ranges obtained for the determination of catecholamines are given in Table 1. For higher levels the photolysis times are too long and it is advisable to dilute the sample. A study of the precision was performed by carrying out ten independent measurements on solutions of various concentrations of EDTA, RB and acetate buffer.The relative standard deviation was 1.9% for epinephrine, 2.8% for norepinephrine, 2.1% for dopamine and 1.8% for r,-dopa at concentrations of 1.5 x 10-5 mol dm-3. Analysis of Pharmaceutical Dosage Forms The method was applied to the determination of epinephrine, norepinephrine, dopamine and L-dopa in pharmaceuticals. Commercially available formulations were analysed and the results obtained are summarized in Table 2. As can be seen, for all the formulations examined, the assay results were in good agreement with the declared content. The accuracy of the method was verified by carrying out recovery studies. When synthetic preparations, with composi- tions identical with those of the commercial formulations, spiked with known amounts of the drug were analysed, quantitative recoveries (99.6-100.2%) were obtained for each of the catecholamines. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 References Pesce, A.J., and Kaplan, L. A., in Methods in Clinical Chemistry, ed. Bircher, S., The C. V. Mosby Company, St. Louis, MO, 1987, pp. 944-963. Sankar, D. G., Sastry, C. S., Reddy, M. N., and Prasad, S. N., Indian J. Pharm. Sci., 1987, 49,69. Fujita, Y . , Mori, I.. Fujita. K., and Takana, T., Chem. Phurm. Bull. , 1985, 38, 5385. Salem, F. B.. Talanta, 1987,34, 810. Canson, R. C., and Brown, M. J., Ann. Clin. Biochem., 1982, 19, 396. Mohamed, W. I., and Salem, F. B., Anal. Lett., 1984, 17, 191. Amin, D., Analyst, 1986, 111, 255. Pelizzeti, E., Mentasti, E., Pramauro, E., and Giraudi, G., Anal. Chim. Acta, 1976, 85, 161. Rodrigucz-Dapazo, M. J., Silva. M.. and Perez-Bendito, D., Microchem. J . , 1989, 39, 235. Pkrez-Ruiz, T., Martinez-Lozano, C., and Ochotorena, J . , Talanta, 1982, 29, 479. Perez-Ruiz, T., Martinez-Lozano, C., Tomas, V., and Yagiie, E., Mikrochim. Acta, Part I I , 1989, 129. Oster, G., and Wotherspoon, N., J. Am. Chem. Soc., 1957,79, 4836. Joussot-Dubien, J., and Faure, J., Bull. SOC. Chim. Fr., 1960, 3434. Sierra, F., Sanchez-Pedrefio, C., Perez-Ruiz, T., Martinez- Lozano, C., and Hernandez, M., Anal. Chim. Acta, 1975, 78. 227. Koizumi, M., Obata, H., and Hayashi, S., Bull. Chem. SOC. Jpn., 1964,37, 108. Well, C. H. J . , Introduction to Molecular Photochemistry. Chapman and Hall, London, 1972, p. 84. Paper 0105335J Received November 27th, 1990 Accepted March 26th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600857

出版商:RSC    

年代:1991

数据来源: RSC