摘要:

ANALYST, MARCH 1986, VOL. 111 35 1 Examination of Metallochromic Indicators and Water-soluble Reagents for Metals by Planar Electrophoresis Marie M. Ferris and Michael A. Leonard Chemistry Department, The Queen's University of Belfast, Belfast BT9 5AG, UK Water-soluble azo, phthalein, sulphonphthalein and anthraquinone indicators and reagents were subjected to low-voltage electrophoresis on cellulose acetate sheets at pH 2-1 2. All dyes except low-sulphonated azo dyes yielded one or more sharp bands showing good mobility and resolution over some range of pH values. The performance exceeded that of other possible planar electrophoresis media. The addition of copper, lanthanum, cetyltrimethylammonium or lauryl sulphate ions caused a marked and often beneficial change in behaviou r.Keywords: Metallochromic indicators; planar electrophoresis; water-soluble dyes Electrophoresis is normally regarded as a technique of great value in biochemistry for the separation of protein and similar large molecule ions, but from many years of teaching planar electrophoresis and using it as a research tool, the senior author has been impressed by the excellence of the technique for the examination of comparatively small ions such as hydrophilic dyes. As an undergraduate project we examined the behaviour of a fairly comprehensive set of metallochromic indicators and water-soluble reagents at different pH values and in the presence of charged surfactants and metal ions. Cellulose acetate sheet was the principal medium used, but comparisons were made with paper, polyamide and various thin-layer media.We hope that we have discovered the best conditions under which to examine these various compounds. There is remarkably little information in the general literature on the planar electrophoresis of dyes, but valuable work that we have noted was that of Criddle et al.,lJ who separated food dyes at six different pH values on alumina, Kieselguhr, silica gel, cellulose acetate and paper. They commented in depth on the relationship between structure and mobility. Patuska and Trinks3 also examined synthetic food dyes and Sharma et a1.4 have looked at dyes in liquors and beverages as a means of detecting fakes. Experimental Apparatus Electrophoresis was carried out in Shandon equipment consisting of a Vokam power supply operated at 200 V and a simple tank suitable for sheet and plate material.The active electrophoretic distance was 6.5 cm. Contact was made between buffer solution and zone medium with Whatman No. 1 filter-paper rectangles. The zone medium principally used was 14 X 5.7 cm Cellogel sheets (Chemetron, Milan; supplied by J. W. Turner, Liverpool). This was impregnated with buffer solution using the recommended floating technique. Plastic sheet or glass plate thin-layer material was treated with buffer solution by dip or spray procedures, as appropriate. Sample solutions were applied using a Shandon single 5-mm slot applicator. Materials Bujjfer solutions Initial work on hydrophilic black inks showed that 0.05 M was the optimum buffer concentration. All major buffer com- ponents were accordingly of this value, pH adjustment being made with dilute sulphuric acid or sodium hydroxide solution as appropriate.The following buffer solutions were prepared: pH 2.0, sodium dihydrogen phosphate; pH 3.0, sodium hydrogen phthalate; pH 4.0, sodium oxalate; pH 5.0, sodium citrate; pH 6.0, maleic acid; pH 7.0, disodium hydrogen phosphate; pH 8.0, tris(hydroxymethy1)aminomethane; pH 9.0, monoethanolamine; pH 10.0, ethylenediamine; pH 11.0, piperidine; and pH 12.0, disodium hydrogen phosphate. Sample solutions These were prepared at a concentration of 1 mg cm-3 in water where possible. For dyes showing poor solubility, a few drops of 0.1 M ammonia solution were added, with subsequent reduction of the pH to 7. Most indicators and reagents were obtained from BDH Chemicals and used as received.Sodium alizarin-5-sulphonate was recrystallised Bayer material and sulphonated Alizarin Fluorine Blue (AFBS) was synthesised according to references 5 and 6. Procedure Five applications of different dye solutions were made across the cellulose acetate strip at about 30% of the strip length to give a long run towards the anode. The strip was placed in the tank and subjected to the electric field in the usual way as soon as possible. Results Ionophoretic mobilities of the indicators and reagents in the buffer solutions described above are given in Table 1. Experiments were repeated in 0.05 M maleic acid - maleate buffer (pH 6.0) with inclusion of 0.02 M copper sulphate, 0.02 M lanthanum nitrate, 0.02 M cetyltrimethylammonium bro- mide (CTMAB) or 0.02 M sodium lauryl sulphate (SLS).The results are given in Table 2. Discussion The cellulose acetate sheets mostly yielded sharp bands with little evidence of tailing or diffusion, especially below pH 10. They were far superior to paper, which gave diffuse bands, cellulose and silica gel, which showed tailing, and polyamide, which gave very low mobilities, presumably owing to excessive hydrogen bonding. Cellulose acetate was the most convenient to impregnate with buffer and gave very little sample diffusion at the origin. A disquieting feature was the way in which, for many compounds, the electrophoretic pattern changed with pH, e.g., Alizarin Fluorine Blue showed one line at all pH values except 8, where three were evident.Hence one cannot be sure whether a multiplicity of lines is due to an impure material, several ionised forms of one compound or some unsuspected complex formation. However, as with mostANALYST, MARCH 1986, VOL. 111 352 Table 1. Ionophoretic mobilities of the indicators and reagents in different buffer solutions* Ionophoretic mobility X 1O3/cm* min-I V-I No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Compound Eriochrome Blue Black R Arsenazo Hydroxynaphthol Blue Calmagite Fast Sulpon Black F Thoron Zincon Eriochrome Black T Solochrome Dark Blue BS Acid Alizarin Black SN HSN (Patton and Reeder’s) Arsenazo I11 (Calcon) Eriochrome Blue Black SE Eriochrome Blue Black B Xylenol Orange Methyl Thymol Blue Calcichrome Glycine Thymol Blue Catechol Violet Bromopyrogallol Red Alizarin Red S Alizarin Fluorine Blue- 5-sulphonate (AFBS) Alizarin-5-sulphonate Pyrogallol Red Alizarin Fluorine Blue PH 2 - 3.4 1.4D - - 3.0 - - - - - 0.1 2.7T 3.3 0.5 2.8T - 1.8T 2.2T 2.7 1.5T 2.I 0.4 0.8T 0.0 0.4 0.8T 1.9 0.4 0.2 1.5T 2.5 0.3T 1.5 0.8 - PH 3 - 5.9 3.0 - - 5.4 - - - - - 3.OT 3.5 0.9D 2.9T - 2.6TD 2.9T 3.5 1.8T 2.7 1.6 0.9 +4 others 1.4 0.4 0.6 3.2 1.5 1.0 1.5 PH4 - 6.1 1.5D - - 5.6 - - - - - 4.OT 4.7 0.2 1.8T - 2.9T 3.4T 4.1 2.5T 4.0 * 1.3 I .2 +4 others 1.4 0.2 0.2 1.5T 3.9 1.1 1.2 4.1T 1.3 PH 5 - 5.6 0.9T 1.6 2.2 - - 4.8 - - - - - 4.7 0.2 1.9T - 3.5T 4. I 2.3T 3.7 1.9 1.3 +6 others 0.9 0.4 0.5T 0.3 4.8 1.3 1.3 2.2T 1.8 PH 6 - 6.7 1.1T 2.0 3.3T - - 6.6 - - - - - 5.3 5.5 0.5 2.5T - 4.7D 0.6T 2.6T 4.9 4.6 1.7 +5 others 1.4 1.5 0.8T 1.1 5.8 1.5 1.8 2.3T 1.9 PH7 - 5.7 0.9 2.0 3.5T * - - 5.2 - - - - - 6.0 0.5D 2.8T - 4.8T 5.5 3.1T 5.4 5.4 1.8 +5 others 1.3 1.3TD 2.2 0.7T 1.6D 5.2 0.OT 2.0 2.1 2.9T 4.2T 1.4 PH8 -D 6.7 0.1T 1.6D 1.8 - - 5.8 -D - -D - - 6.3D 1.1 - 4.0 4.7 5.6 1.7 +4 others 1.6 1.2T 2.6 O.OT 2.2 7.5T 5.5 O.OT 3.7 2.8 3.7T 5.OT 6.4T 0.2 0.8 1.1T * PH 9 0.3T MO 5.8 2.2TD 2.7 MO 0.2T 0.6T 0.2 5.8 0.4 0.2 0.1T 0.2D - - 6.4T 6.8 3.1 5.3T - 4.3 5.4T 0.8T 3.5T 4.1 6.4 3.2 +7 others 2.8 1.2T 1.9T 2.8 5.2T 2.7 7.6T 6.1D 4.4 4.9 * * * 3.8D 1.1 pH 10 - 7.5 2.OD MO 0.5T 1.2T 0.3 5.8 O.OT 0.6 * - - MO 0.4T - 7.2T 7.8 0.7T 1.5T 4.6 7.5T MO 0.1 4.7 6.1T * 1.1T 4.2T 4.6 2.9T 6.1 3.7 +6 others 3.4 1.4T 3.2 5.4T * 2.3 5.3T 5.8 0.OT 5.0 3.4 1.3D pH11 pH12 0.2T 0.5TD MO MO 6.3 6.6 7.0 I .7 D 0.6D 2.6T 2.3T 3.5T 3.3T MOD 0.OT 0.2 1.4T - - 5.8 6.4 0.8D 2.2D 1.6D 0.1D 0.0 0.3 MOD 0.4T MO - - 0.0 0.2 7.OT 5.OT 7.4 5.7 - 4.7 3.8 5.OT 4.9T 6.OT 0.1D 0.1 * 4.4 4.2 5.OT 4.9T 1.8T 3.4T 3.7T 4.7T 3.9 4.9T 5.5 * 2.9T 3.OT 6.0 6.6 4.2 4.8 others 4.OT 4.4 6.5T 6.8 3.2 3.OT 5.2 3.9 +7 3.5TD 5.9TD 6.7TD 1.2D O.OD 2.4D 1.9D 6.OT 2.5D 6.1 5.0T 6.9 0.OT 0.OT 0.2T 3.4 2.4T 5.2 4.2 2.9T 4.9T 5.OD 6.5 6.7 1.4D 3.ODANALYST, MARCH 1986, VOL. 111 353 Table l-continued Ionophoretic mobility x lO3/cm2 min-1 V- No. Compound pH2 pH3 pH4 pH5 pH6 pH7 pH8 pH9 pH10 pH11 pH12 26 Thymolphthalein MO MO MO 1.3T MO 0.2 MO 0.9T MO 0.5T 1.4T complexone 1.7T 1.4T 1.2T 1.6 0.7T 0.5 3.8 4.3 0.5T 0.8T 1.7 2.3 2.0 1.7 2.5 0.8T 1.OT 1.4 2.2 2.6 4.3 * 2.7T * Key, -, zero mobility; T = trace; MO = main band has zero mobility; * = pH showing best resolution of bands; D = diffuse.Values in italics refer to principal band(s). A negative sign indicates migration towards the cathode. Table 2. Ionophoretic mobilities of the indicators and reagents in 0.05 M maleic acid - maleate buffer solution with different additives* Ionophoretic mobility (at pH 6) x lO3/cm2 min-1 V-1 Compound f No additive - 1 0.02 M cUso4 - 5.7D 6.5T 7.4T 0.8T 1.7 3.4T 5.5T 6.4T 7.2T MO 1 .OT MO 0.8T 5.2 - - - MO 0.6TD MO 1.8 4.3 4.6 5.0 0.3T 1.7T 2.7 -D 1.7D 2.4T 2.6T 3.2 0.6T 1.5T 2.1T 2.4T 3.4 2.5T 4.3D 0.7 1.0 0.3 1.7T 2.7T 2.7 3.2T 1.7TD 2.5D 3.5TD 0.02 M La(N03)3 0.02 M CTMAB 0.02 M SLS - 1.7T -1.9 3.1 0.0 +2.1 3.5TD -1.1 -3.2T 3.9T - 1.3T 0.7T -1.3 2.4 2.7T 3.2T 3.9T - 2 6.7 3 1.1T 2.0 3.3T 4 - 1.6T -1.8 -1.8 MO 0.9T MO 0.3TD 4.1 MO 3.2T 5 6 7 6.6 - -1.1 +0.4T -1.2D -2.4 -0.3T - 1.5T -2.6 -2.3 -2.4 -1.2T -1.5T -1.9 -1.2 -1.5 MO 1.9 8 9 10 - 0.6D 2.1T MO 1.2 11 12 5.3 5.5 MO 2.2T 1.7 13 0.5 2.5T 0.4T 1.2 3.2T 1.5 2.OT - -0.4D -1.1d - 1.6T -1.9 -1.7 -2.0t 14 15 - 4.7D 0.6T 1.1 16 0.6T 2.6T 4.9 0.9 1.6 -1.1 - 1.2T 17 18 4.6 2.OT 3.1D 0.8D 1.7D 2.7T -0.9 1.7 +5 others -1.1T -1.4T -1.7 0.2 1.7TD 1.8T -0.5 19 20 1.4 1.5 1.3 -1.2D 1.7T 3.7 3.8TD -1.1354 Table 2-continued ANALYST, MARCH 1986, VOL. 111 Ionophoretic mobility (at pH 6) x l03icrn2 min-1 V-I Compound? 21 22 23 24 25 26 * For key, see Table 1.t Compounds as in Table 1. No additive 0.8T 1.1 5.8 1.5 1.8 2.3T 1.9 MO 0.7T 2.5 0.02 M cUso4 0.0 0.2T 2.OD 2.2D 3.6T 1.9T 2.4T 3.I 1.3T 2.3 2.1T 2.9D 4.0D 1.4 1.7T 0.3D 0.02 M La(NO& 0.02 M CTMAB 0.02 M SLS 0.0 -1.2 0.5 2.5D MO 1.7T 3.3 1.4T 3.m 4.1 1.2 MO 0.9T 1.7 -1.1 -0.8 -0.9 -1.1 -0.7 +1.2T -0.4T -1.4 planar chromatographic techniques, careful comparison with pure standards should eliminate ambiguity. An increase in dye mobility with increasing pH as carboxy and phenolic groups ionised was clearly evident in many instances, such as with Xylenol Orange. For many compounds the mobility remained almost constant between pH 4 and 8 but then increased at higher pH to reflect the pKvalues of hydroxy groups. Although the mobilities were high at high pH, the sharpness of the bands often deteriorated, and pH 10 was selected as a good compromise. Dyes that had the highest electrophoretic mobilities at all pHs were Arsenazo , Thoron and Arsenazo 111.The behaviour of the azo dyes was interesting and followed a well defined pattern. Those with a ratio of number of SO3- plus As03H- groups to number of aromatic moieties of S1 showed zero or very low mobility at all but very high pH values, whereas those with a ratio >1 showed medium to high mobility at all pHs. Dyes of the former type tend to be poorly soluble at around neutral pH and are not readily amenable to this type of electrophoresis. In contrast, (su1phon)phthaleins and anthraquinone dyes with one or fewer sulphonate groups per aromatic moiety, although often rich in carboxy groups, showed good mobility at pH ca. 7. The obviously poor purity of many of these compounds (Glycine Thymol Blue is a bizarre example) illustrates the importance attached to purifying indicators and reagents prior to fundamental studies such as ionisation or stability constant determinations.Electrophoresis of the type described here would be an excellent means of following purification proce- dures. The presence of metal ions that are good complex formers with EDTA, e . g . , Cu2+ and La3+, drastically altered the electrophoretic patterns of the dyes, usually by reducing mobilities, e.g., AFBS, or increasing the number of observed bands, e.g. , Hydroxynaphthol Blue. Even some immobile dyes showed movement, e.g., HSN. Such marked changes could greatly assist the comparison of unknowns with stan- dards. However, a proper explanation of this behaviour shown by the metal complexes will require the production of pure dye materials.The presence of charged surface-active agents caused a marked change of pattern. The cationic agent cetyltrimethyl- ammonium bromide caused most bands to move towards the cathode and, notably, produced extensive and sharp band movement in azo dyes formerly immobile, often with resolu- tion of several bands, e.g. , Zincon. However, there seemed to be limited variation in band mobility. Owing to a shortage of cellulose acetate sheets, only a few experiments were carried out with the anionic agent sodium lauryl sulphate. Here, curiously, the mobility of normally highly mobile bands was greatly diminished but the bands remained very sharp and showed good resolution. We feel that high electroosmotic flow is probably the main reason for these phenomena. Selected dyes were run at pH 11.4 using an anionic buffer based on disodium hydrogen phosphate and a cationic buffer based on piperidine. The mobilities were greater, the bands were sharper and the resolution of the bands were superior in the anionic buffer. In conclusion, cellulose acetate sheets have been shown to be an excellent medium for the electrophoretic examination of phthalein, sulphonphthalein, anthraquinone and richly sul- phonated or arsonated azo water-soluble indicators and reagents. The addition of metal ions or charged surfactants frequently improves separations. References 1. 2. 3. 4. 5. 6. Criddle, W. J . , Moody, G. J., and Thomas, J. D. R., Nature (London), 1964,202, 1327. Criddle, W. J., Moody, G. J., and Thomas, J. D. R., J. Chromatogr., 1964, 16, 350. Patuska, G., and Trinks, H., Chem. Ztg., 1962, 86, 135. Sharma, V. K . , Sharma, I. C., and Tewari, S. N., Chromato- graphia, 1975, 9, 405. Leonard, M. A., and Murray, G. T., Analyst, 1974, 99, 645. Leonard, M. A., Analyst, 1975, 100, 275. Paper A51320 Received September 12th, 1985 Accepted September 24th, 1985

ISSN:0003-2654    

DOI:10.1039/AN9861100351

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, MARCH 1986, VOL. 111 35s Determination of the Silver Error in the Coulometric Titration of Acids in Various Media Adam Hulanicki, Stanis?aw Giqb and Wojciech Jedral Department of Chemistry, Warsaw University, Warsaw, Poland The dependence of the silver error in coulometric titrations of acids in a one-compartment cell on the solubility of silver halide has been established. The silver error is proportional to the total solubility of silver and inversely dependent on the current density. Because the main contribution to the total titration error is given by the silver error, it is necessary to choose conditions that produce the minimum value. The silver error is below 0.1% when the current density of the cathodic process is at least 10 mA cm-2. This relationship was tested in solutions of chloride, bromide and iodide in water, methanol, ethanol and acetone.For the aqueous solutions the bromide electrolyte is recommended and, for methanolic solutions, the best results are obtained with chloride ions. Keywords ; Silver error; coulometric titration; acid determination; non-aqueous solvents In coulometric titrations the anode and cathode compart- ments are commonly separated by a diaphragm to eliminate mutual interferences of electrolysis products formed at both electrodes. To ascertain that such interferences do not occur, special precautions must be taken; however, the decrease in diaphragm permeability increases the cell resistance and makes the generating current stabilisation more difficult. In addition, the increase in applied voltage in the generating circuit may disturb the functioning of the electrochemical indicating system.In addition, some electrolyte flow through the diaphragm occurs, which may influence the titration efficiency and complicate the determination. Szebelledy and Somogyi’ introduced the technique of one-compartment coulometric titrations for the determination of acids when the product formed at the anode is insoluble and does not interfere in the the cathodic processes. In this procedure a silver anode is placed in the halide-containing sample solution. A silver halide precipitate is formed at the anode, which should not, in principle, interfere in the determination of acids by the hydroxy ions formed at the cathode. Lingane and Small2 and Bishop and Riley3 stated that bromide is more advantageous than chloride, because of its lower solubility. The solubility of silver halide is not only governed by the solubility product, but also by the formation of soluble halide complexes, which may in turn be reduced at the cathode, decreasing the current efficiency of the genera- tion of hydroxy ions.It has also been observed4 that the addition of tetraphenylborate anions, which also form spar- ingly soluble silver salts, improves the results. Bishop and Riley3 calculated the silver error from the total charge necessary for the dissolution of silver deposited on the cathode during coulometric titration. For a current density of 30 mA cm-2 the error was 0.01% when 0.1 M sodium bromide solution was used as the electrolyte.The experimentally evaluated silver error was taken as a basis for the correction of the results.5 Because of the increasing importance of coulo- metric titrations in various solvents the effects of different parameters on the magnitude of the silver error have been studied. Experimental Reagents Doubly distilled water from a quartz still was used in all measurements. Solvents and salts of analytical-reagent grade * Presented at Euroanalysis V, Krakow, Poland, August 26-31, 1984. were used in all experiments. Perchloric acid solutions of approximately 0.01 M were prepared in water and methanol and standardised titrimetrically with sodium carbonate. Apparatus A Radelkis Model OH 404 coulometric analyser was used for all coulometric experiments. The generator electrodes were made of platinum foil with a surface area of 0.7, S or 20 cm2.The silver anode had a surface area of about 25 cm2. The indicator system consisted of a PHM 64 pH meter, G 202 B glass electrode and a K 401 calomel reference electrode, all from Radiometer. Titration end-points were evaluated using Gran’s method by plotting 10E’S vs. Q. Procedure The silver error in coulometric titrations of acids was determined in chloride, bromide and iodide solutions in water, methanol, ethanol and acetone (90%). The method followed was that of Bishop and Riley.3 The platinum electrode was thoroughly cleaned by anodic polarisation in 0.1 M perchloric acid and washing with nitric acid and water. This electrode was then placed in the coulometric vessel together with the silver anode.After addition of the sample of perchloric acid the electrolysis was performed with the given current density of the cathodic process. The electrolysis was stopped when a charge Q1, slightly less than that equivalent to the amount of added acid, passed through the solution. At this point the platinum electrode, with the small silver deposit, was removed immediately, washed with water and the amount of silver was anodically stripped in 0.1 M perchloric acid with a 1 mA current using an auxiliary platinum electrode. The potential was measured versus a saturated calomel electrode and from the recorded potential versus time graph the charge 122, necessary for the dissolution of silver, was determined. The silver error was calculated as the charge ratio Q2/(121 - Q2).Results and Discussion Dependence of the Silver Error on the Solubility of Silver Halide Satisfactory results for the determination of the acid can be obtained when the current efficiency of base ion formation and/or of direct reduction of hydrogen ion equals 100%. This is disturbed when the silver ions formed at the anode are not completely bound as insoluble precipitate at the electrode surface but enter into solution as free Ag+ ions, or as soluble complexed species that in turn may be reduced at the cathode.356 ANALYST, MARCH 1986, VOL. 111 Such an interfering process may be formulated as [Ag+, AgX, AgX2-, . . . , AgX,@-l)-] + e- e Ag + nX- (1) and its extent depends on the total concentration of dissolved silver species. The total silver solubility, S , is or after the introduction of respective stability constants S = Ks0([X-]-1 + p1 + Pz[X-] + .. . + pn[X-]n-l) = Ks0[X-]-1 (1 + pl[x-] + P2[X-]2 + . . . + PJX-1,) (3) The term in parentheses is equivalent to the side-reaction coefficient of silver ions under the given conditions. Thus, assuming that the actual concentration, [X-1, of the halide ions equals their total concentration, C,, the total solubility of silver is S = Kio/C, (4) where Kio is the conditional solubility product of the I 1.5 1 2.0 0.06 - 0.04 - 0 5 10 15 20 Solubility x I O V M Fig. 1. Dependence of the silver error in coulometric titration of acid in aqueous bromide solutions at a current density of the cathodic process of 25 mA cm-2. (a) Dependence on the bromide concentra- tion; and (b) dependence on the calculated molar solubility of silver bromide corresponding silver halide.From the values of the equilib- rium constants of the silver halides (Table 1) it follows that, except for very concentrated halide solutions, the most important term is the constant p2; however, where available, all constants were used in the calculations. The silver error was determined using a constant current density of the cathodic process under selected conditions of halide concentration and solvent used. The decrease in halide concefiirarion causes a significant decrease in the silver error [Figs. l(a) and 2(a)]. When the total solubility of silver was calculated according to equation (4) a linear dependence on the silvei error was found [Figs. l(b) and 2(b)].Some other experimentally determined values of the silver error at a current density of the cathodic process of 1 mA cm-2 and for 0.05 M solutions of halide ions are given in Table 2, and these are compared with the calculated values of conditional solubility products, pKio. Those values of the silver error are reproducible in most solutions. When acetone was used as a solvent a very large scatter of results was observed and a dependence on the electrolysis time was noted. This may be connected with the magnitude of the conditional solubility product of silver halides in acetone, and also with the kinetics of dissolution of the anodic deposit. When in the same solution subsequent determinations were performed, the silver error increased. The data in Table 2 are given for the first and fourth determinations.L al PCI 0 2 4 6 8 10 Solubility x IO'/M Fig. 2. Dependence of the silver error in coulometric titration of acid in methanolic chloride solutions at a current density of the cathodic process of 1 rnA cm-2. (a) Dependence on chloride concentration; and ( b ) dependence on calculated molar solubility of silver chloride Table 1. Equilibrium constants used in calculations of total solubility of silver halides Solvent Halide Water . . . . C1- Br- Methanol . . C1- Br- I- Acetone . . . . C1- Br- PKSO 9.75 12.3 13.0 15.2 18.2 16.4 18.7 Log P I Log P 2 3.4 5.3 4.2 7.1 7.9 10.9 14.8 16.7 19.7 Log P3 Log P4 Reference 5.48 6, 7 8.0 8.9 8 9 9 9 9 9357 ANALYST, MARCH 1986, VOL. 111 ~~ Table 2. Conditional solubility products for silver halides and the silver error in 0.05 M sodium halide solutions in various solvents.Cathodic current density, 1 mA cm-2 Conditional solubility product for 0.05 M X- Silver error, % Solvent c1- Br- I- c1- Br- I- Methanol . . . . . . 0.3 1.2 1-7.5 7.7 6.9 6.0 Ethanol . . . . . . 0.4 1 .o 1.7-10 Water . . . . . . 0.6 0.2 7.05 7.8 Acetone(90Yo) . . . . 2-22 3.7-10 2.3* 1.6* * Data for anhydrous acetone. Table 3. Comparison of the titration error and silver error in the coulometric titration of 0.009660 M perchloric acid in 0.05 M sodium chloride in methanol Current density/mA cm-2 0.1 1 .o 14.6 HC104 takedmmol . . . . . . 0.02415 0.0966 0.04830 Q(theoretical)/mC . . . . . . 2330 9320 4660 Q(experimental)/mC . . . . 2408 9361 4660 Relative standard deviation, YO (n = 5 ) 0.30 0.45 0.43 Titration error, YO .. . . . . +3.35 +0.43 <+0.01 Silver error, o/o . . . . . . . . +3.3 +0.3 <+0.01 . . . . . . . . . . Table 4. Comparison of the titration error and silver error in the coulometric titration of 0.01102 M perchloric acid in 0.05 M sodium bromide in water Current density/mA cm-2 0.5 1.0 10 . . . . . . 0.1102 HC104 taken/mmol 0.02755 0.0551 Q (theoretical)/mC . . . . . . 2658 5317 10633 Q(experimental)/mC . . . . 2668 5330 10638 Relative standard deviation, YO ( n = 5 ) . . . . . . . . . . 0.40 0.50 0.45 Titration error, YO . . . . . . +0.38 +0.25 +0.04 Silver error, Yo . . . . . . . . + 0.44 +0.21 +0.02 Also, irreproduci ble results were observed for iodide solutions in alcohols. Here, the first determination also exhibited the smallest error, which increased for the fourth in spite of relatively small values of conditional solubility products.The kinetics of the dissolution and the diffusion of ions may be responsible for those effects. The data in Table 2 indicate that the best results for acid determination, i.e., the smallest value of the silver error, are obtained in chloride solutions when methanol or ethanol is used as the solvent, whereas they confirm that in water the best electrolyte is bromide. Dependence of the Silver Error on the Current Density of the Cathodic Process The next parameter that influences the silver error for determination of acid is the current density at the working electrode. In these experiments the current density was changed from 0.05 to 30 mA cm-2 and the results were plotted logarithmically (Fig.3). For a single system (solvent - halide) a linear dependence was obtained with a slope close to -1. The relative positions of the lines for the systems studied are in accordance with the change of the solubility of the correspond- ing silver halides. For better results large current densities are advantageous, because the contribution of silver ions in the over-all cathodic process depends on the potential of the cathode. When the current density is small the electrode potential is more positive and corresponds to the reduction of silver ions. When the current is large in comparison with the diffusion current of silver ions at their given concentration, the \ 0.1 1 .o 10 1 O( Current density/mA cm-2 Fig.3. Dependence of the silver error on current density of the cathodic process for 0.05 M sodium halide in different su porting electrolytes. -, Methanol; - -, ethanol; and - - -, water. 8, C1; 0, Br; and A , I358 ANALYST, MARCH 1986, VOL. 111 cathode potential is more negative, because it is shifted to the potential of solvated protons or solvent reduction. When the current in the coulometric system is higher than the limiting reduction current of silver ions, the absolute amount of reduced silver is constant for a given period of time and is independent of the current density. However, the greater the current density the smaller is the relative contribution of silver reduction in the over-all cathodic process. Total Titration Error and Total Silver Error The total error of coulometric titration in a one-compartment cell depends on the silver error, which decreases the current efficiency of titration, on the apparatus error, which is dependent on the error of charge measurement, and on the error of end-point location during titration.For checking the magnitude of the titration error, known samples of perchloric acid were titrated and the total error was compared with the silver error in the same solutions. Such experiments were carried out under the optimum conditions, i.e., in methanolic 0 . 0 5 ~ sodium chloride solutions and in aqueous 0 . 0 5 ~ sodium bromide solutions (Tables 3 and 4). In both systems three different current densities were used. These results indicate that the main source of error in all experiments is the silver error and the difference between the total titration error and the silver error is less than the precision of determinations with the apparatus and experimental procedure used.Conclusions It is experimentally evaluated that there is an exact correlation between the solubility of silver halide in the electrolyte solution and the silver error of coulometric titration of acids with the silver anode in the one-compartment coulometric cell. The solubility is dependent on the concentration of the halide used as an electrolyte and on the solvent. The silver error is linearly dependent on the total solubility of silver and inversely dependent on the current density. To obtain silver errors of less than 0.1% the current density of the cathodic process should be at least 10 mA cm-2. Bromide is the best halide for aqueous solutions and chloride for methanolic solutions. Iodide electrolytes and acetone as a solvent are not recommended. By comparing the total titration error with the silver error it may be stated that the latter gives the main contribution to the total error of determination, indicating the importance of the proper choice of experimental conditions so as to minimise the silver error. This work was partially supported by the Committee on Analytical Chemistry of the Polish Academy of Sciences. 1. 2. 3. 4. 5 . 6. 7. 8. 9. References Szebelledy, L., and Somogyi, Z . , Fresenius 2. Anal. Chem., 1938, 112, 323. Lingane, J. J., and Small, L. A., Anal. Chem., 1949,21,1119. Bishop, E., and Riley, M., Analyst, 1973, 98, 313. Johansson, G., Talanta, 1964, 11, 789. Bishop, E., and Riley, M., Analyst, 1973, 98, 416 and 426. Davies, C. W., and Jones, A. L., Trans. Faraday SOC., 1955, 51, 812. Mironov, V. E., Radiokhimiya, 1962, 4, 707. Berne, E., and Leden, I., 2. Naturforsch., 1953, 84, 719. Luehrs, D. C . , Iwamoto, R. I., and Kleinberg, J., Znorg. Chem., 1966,5,201. Paper A51233 Received July lst, 1985 Accepted September 3rd) 1985

ISSN:0003-2654    

DOI:10.1039/AN9861100355

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, MARCH 1986, VOL. 111 359 Differential-pulse Stripping Voltammetry for the Determination of Soluble Iron in Simulated PWR Coolant K. Torrance and C. Gatford Central Electricity Generating Board, Central Electricity Research Laboratories, Kelvin Avenue, Leatherhead, Surrey KT22 7SE, UK A method has been developed for the determination of dissolved or ionic iron in pressurised water reactor (PWR) coolant using differential-pulse voltammetric stripping (DPVS) following adsorptive accumulation of an iron(l1l) - catechol complex at a hanging mercury drop electrode. The method does not discriminate between Fe(ll) and Fe(l1l) ions and the total ionic iron is determined in a Tris buffered solution at pH 8 & 0.1. Interference from the reaction of boric acid in the coolant with the catechol was overcome by limiting the total boron in the analytical solution to d 100 mg 1-1 and adding mannitol.No interference was observed from ions of common transition metals such as copper, nickel, cobalt and manganese. The effects of pH, catechol concentration, accumulation voltage and time were investigated in simulated coolant solution. The limit of detection was of the order of 0.1 pg 1-1 and the precision of a single determination based on a standard additions procedure was considered to be in the range 10-15%. When the results obtained by this technique were compared with those obtained by a spectrophotometric method, good agreement was obtained. The DPVS method for the determination of ionic iron in PWR should prove valuable in corrosion studies and it can complement alternative methods, such as electrothermal atomic absorption, which only gives the total iron in solution.Keywords : Iron determination; P WR coolan t analysis; adsorptive concentration; differentia I-pu Ise voltammetric stripping The boric acid solution used as the coolant in the primary circuit of a pressurised water reactor (PWR) contains trace amounts of metals derived from the materials of construction and periodic analysis of the coolant is required to assess the extent of these metals present with a view to adjusting, when necessary, the coolant chemistry. The most abundant metal has been shown to be iron whose concentration can vary widely according to the age and condition of the reactor. In general, the iron content has been reported as <50 yg 1-1 and at these concentrations the usual method of analysis is electrothermal atomic absorption spectrometry.This tech- nique determines the total rather than the soluble concentra- tion of any element unless a prior separation is used and its sensitivity is such that measurements of iron in the coolant are not very satisfactory below 1 pg 1-1. Differential-pulse voltammetric stripping (DPVS) following adsorptive accumu- lation at a hanging mercury drop electrode has been shown to be suitable for the determination of the low concentrations of soluble nickel and cobalt1 expected in PWR coolant and the same technique has now been extended to the determination of soluble iron. Catechol has been used as an adsorption enhancement reagent for the DPVS determination of uranium (VI) ,2,3 copper4 and iron5 in sea water.In the last application, samples were buffered to pH 6.9 in 10-3 mol 1-1 piperazine- N,N'-bis-2-ethanesulphonic acid (PIPES), and at this pH a reduction peak, whose current was proportional to the soluble iron concentration, was observed at -0.35 to -0.4 V. The sensitivity and apparent freedom from interference reported for this method of determination for soluble iron suggested that it could be applied to the analysis of PWR coolant. The difficulty associated with the use of catechol for the determination of iron in PWR coolant is its reaction with borate anions6 with which it forms complexes in a manner analogous to mannitol. An additional complication is the decrease in the boric acid content throughout the lifetime of a reactor fuel cycle from an initial value of >lo00 to a final value of <lo0 mg 1-1 of boron.This requires either a variation in analytical reagents corresponding to the decrease in boric acid or an excess sufficient to produce effectively constant analy- tical conditions throughout. Both these suggestions have their disadvantages and the procedure that was chosen was to dilute all the samples such that the final concentration of boric acid in the analytical solution was the equivalent of d 100 mg 1-1 of boron. Using this procedure there was no need to vary the volumes of reagents and, further, the effect of borate ion on catechol was immediately reduced. The resulting reduction in the iron concentration in the analytical solution was not expected to be a problem as the accumulation period can be increased to produce a measurable peak current.Experimental Reagents Catechol. Reagent-grade catechol was recrystallised twice from toluene. After drying at 40 "C it was stored in a blackened container. A 0.1 mol 1-1 ethanolic solution was stable for at least a week when also stored in a blackened container. Tris buffer solution. Analytical-reagent grade tris( hydroxy- methy1)methylamine was recrystallised twice from 3 + 1 ethanol - water. The original concentration of iron in the solid of 0.4-0.6 mg kg-1 was reduced to 0.03-0.05 mg kg-1 by this procedure. Additional recrystallisation did not improve the purity . A pH 7.9 (kO.1) buffer solution containing 0.5 mol 1-1 of Tris was prepared by the addition of isothermally distilled hydrochloric acid to the recrystallised base according to the proportions given by Perrin and Demp~ey.~ Attempts were made to reduce further the iron concentration of this buffer by adsorption on added manganese dioxide but the resulting solution always produced voltammograms with a broad distorted iron(II1) reduction peak.Boric acid. High-purity grade boric acid (Borax Consoli- dated Ltd.) was used to prepare simulated PWR coolant. No significant amounts of iron were detected in the boric acid at the concentrations (50 and 100 mg 1-1 of boron) used in the polarographic cell. Mannitof. Analytical-reagent grade mannitol was prepared as a 0.5 moll-1 solution and was purified by passing through a column of nuclear-grade cation-exchange resin in the360 ANALYST, MARCH 1986, VOL.111 H+ form. The iron content of the resulting solution was equivalent to 0.03 mg kg-1 in the solid mannitol. Standard iron(l1I) solution. A standard solution of Fe(II1) containing 1000 mg 1-1 was prepared from analytical-reagent grade iron(II1) ammonium sulphate. Working solutions con- taining 200-1000 pg 1-1 of Fe(II1) were prepared daily as required by sequential dilution. All standard solutions con- tained 10-2 mol 1-1 of sulphuric acid. De-ionised water. De-ionised water was prepared by circu- lating a bulk supply of water continuously through a mixed catiodanion-exchange column such that the conductivity was of the order of 0.06-0.07 pS cm-1 at 25 "C. Storage of reagents. Polythene containers were used exclu- sively.A cleaning procedure based on that described by Moody and Lindstroms was used in which the bottles were filled with hydrochloric acid (1 + l ) , let stand for a week, emptied, rinsed and refilled with nitric acid (1 + 1) and let stand for a further week. The bottles were subsequently rinsed and left filled with de-ionised water. Polarograph and voltammetric conditions An EG & G Princeton Applied Research Model 264 polarographic analyser in conjunction with a Model 303 mercury electrode assembly were used in these experiments. The voltammograms were recorded on a Model RE 0089 X - Y recorder (EG & G Instruments). The standard reference half-cell supplied with the apparatus was replaced by a borosilicate glass tube of the same dimensions but having a crack-junction1 rather than a ceramic-frit junction as in the original.Nitrogen (99.999%), which was first passed through a gas chromatographic oxygen trap (Phase Separations), was used to de-oxygenate the solutions in the polarographic cell. The sequence that was followed in most instances was a de-oxygenation period of 12 min, an accumulation period (usually 50 or 100 s) at -0.2 V, then a differential-pulse voltage scan at 5 mV s-l to -0.7 V. In all instances the drop size was set at medium (2.5 mg), the pulse height at 50 mV and the pulse interval at 0.5 s. Results Buffer Systems Voltammograms of the iron(II1) - catechol complex were initially investigated in a 0.1 mol 1-1 ammonium chloride solution whose pH was adjusted over the range 7-9.5 by the addition of isothermally distilled ammonia.Over this range a reduction peak was observed at -0.44 to -0.47 V whose current was proportional to the concentration of Fe( 111) and the accumulation period. These experiments indicated that the maximum sensitivity was obtained at about pH 8 and in the presence of boric acid there was a slight increase in peak current with time, which was possibly due to the borate - catechol reaction. In order to optimise the sensitivity, a buffer with a range encompassing pH 8 was required and two such systems were considered. A PIPES buffer at pH 8 was tried but it gave inferior voltammograms to those obtained with a Tris buffer at the same concentration (0.05 mol 1-1) and pH. The latter system was then investigated thoroughly. Effect of pH on the Peak Current and the Sensitivity in Tris Buffer The effect of pH on a solution containing 100 mg 1-1 of boron and 1 pg 1-1 of Fe(II1) was investigated in a 0.05 mol 1-1 Tris buffer whose pH was adjusted over the range 6.6-8.5 by the addition of small volumes of isothermally distilled hydro- chloric acid.Catechol was added to the extent of 80 pl of 0.1 mol 1-1 ethanolic solution per 10 ml of sample. Because a reduction in pH produced a shift (0.1 V per unit of pH) in the positive direction of the voltage of the peak, and thus Table 1. Effect of pH on peak sensitivity. Current range, 50 nA; accumulation time, 50 s; drop size, medium;pulse amplitude, 50 mV; voltage scan, -0.2 to -0.7 V; pulse interval, 0.5 s Sensitivity1 nA (pg 1- l ) - 1 PH 6.54 3.5 7.07 6.3 7.40 8.0 7.74 8.3 8.16 7.8 8.52 6.4 variations in the base line, it was considered that a measure- ment of sensitivity [i.e., nA (pg 1-1)-1] at each pH was more applicable than peak height.At each pH, a known addition of 1 pg 1-1 of Fe(II1) was made to the solution above and the change in peak height recorded as a sensitivity. The results in Table 1 show that the optimum sensitivity lies in the range pH 7.4-8, although this range could be extended by k0.5 unit with only about a 20% loss in sensitivity. The minimum volume of 0.5 mol 1-1 Tris solution (pH 7.9) required to take the pH of a solution containing 100 mg 1-1 of boron into the pH range for optimum sensitivity was shown to be 0.5-0.6 ml. In all of the experiments subsequently reported, 1 ml of 0.5 mol 1-1 Tris solution was added per 10 ml of analytical solution.Effect of Catechol Concentration on Peak Current The effect of varying the catechol concentration in the analytical solution from 10-5 to 10-3 moll-1 was investigated in a solution containing 100 mg 1-1 of boron, 2 pg 1-1 of Fe(III), 1.25 x 10-2 moll-1 of mannitol and Tris buffer at pH 7.8. The peak current gradually increased with addition of catechol up to a catechol concentration of about 2 x 10-4 mol 1-l and thereafter it was constant. The catechol concen- tration used in most of the subsequent experiments was 8 x 10-4 mol 1-1. Mannitol Addition and Peak Current Stability In the preliminary experiments in the ammonia - ammonium chloride buffer solution there was a slight increase in peak current with time, particularly at pH > 8.5, which was perhaps associated with the reaction of borate anion with catechol.Although this reaction was expected to be much reduced at the pH values in the Tris buffer (7.&8), where the proportion of free borate will be reduced, it could reduce the over-all accuracy. Mannitol forms stronger complexes with borate anion than does catecho16 and thus by adding an excess of it over the total boric acid present, the catechol - borate reaction will be reduced and the iron(II1) catechol reaction stabilised. Experiments were carried out on a solution containing 8 X 10-4 moll-1 of catechol, 100 mg 1-1 of boron and 2 pg 1-1 of Fe(II1) in the presence and absence of mannitol; the mannitol to boron molar ratio was approximately 1.5 : 1.Over a 40-min period there was no significant increase in peak height of the voltammograms measured in the solution containing mannitol whereas there was a 6% increase over the same period for the solution having no mannitol. For all of the experiments reported mannitol was added at 1.25 x 10-2 mol 1-1 in the analytical solution and no attempt was made to define the minimum mannitol concentration required. For determina- tions near the limit of detection (ca. 0.1 pg 1-1 of Fe) mannitol could be omitted or its concentration reduced as it introduces 0.1 pg 1-1 of iron in the analytical solution, but some loss of precision may occur. Effect of Accumulation Time and Potential The effect of accumulation time was investigated at two concentrations of iron(II1) in the analytical solution, 1.7 andANALYST, MARCH 1986, VOL.111 361 50 6o I (stirred) 4.7 pg 1-1 Fe (non-stirred) 1.7 pg I-' Fe (st i r red ) 10 - I I I I I 0 40 80 120 160 200 Accu mu lati on ti meis Fig. 1. Effect of accumulation time on peak current 4.7 pg 1-1 in the presence of 50 mg 1-1 of boron. The results are shown graphically in Fig. 1, where it can be seen that in stirred solution the onset of non-linearity occurs at ca. 120 s for the lower and at ca. 100 s for the higher concentration. For analytical purposes it is essential to operate at an accumulation time within the linear region when calibration is made by a standard additions procedure. One method of extending the linear region to solutions of higher concentrations is to use unstirred solution during the accumulation period and the effect of this on the solution containing 4.7 pg 1-1 was to extend the linearity to times greater than 200 s.Using accumulation potentials of 0, -0.1 and -0.2 V the peak current of a solution containing 1 pg 1-1 of Fe(II1) decreased by about 6% at the most negative potential. If copper and iron are to be determined simultaneously then it is necessary to use 0 V, but for iron alone a value of -0.2 V is more convenient. Effect of Iron(I1) A standard solution of iron(I1) (100 mg 1-1) was prepared by dissolving ammonium iron(I1) sulphate in de-ionised water that had been de-aerated with oxygen-free nitrogen. A second standard containing 1000 pg 1-1 of Fe(I1) was prepared from the first by dilution with oxygen-free water.A 10-pl portion of this second standard was added to 10 ml of a de-oxygenated solution containing Tris buffer and catechol. Voltammograms of this solution gave a peak at exactly the same voltage as that found for iron(II1). Multiple additions of iron(I1) gave the same peak current - concentration dependence as iron(II1) and it was concluded that at the voltage used for accumulation of the complex, -0.2 V, the iron(I1) ions diffusing to the electrode surface were oxidised to iron( 111). Consequently, under the prescribed conditions, the total ionic iron in solution is determined rather than the content of soluble iron(II1). Effect of Some Transition Metal Ions The effect of transition metal ions, which are known to be present in the primary circuit of a PWR, was investigated.Voltammograms of five portions of a solution containing 1.5 pg 1-1 of Fe and 50 mg 1-1 of boron were recorded and the mean peak current was compared with that obtained from five portions of the same solution containing in addition 10 pg 1-l each of Cr(VI), Mn(II), Ni(I1) and Co(I1). No significant difference in the peak currents was observed. Copper(I1) ions form a strong catechol complex that is adsorbed on the electrode and has a reduction peak at about -0.2 V in Tris buffer at pH 7.9. The separation between this peak and that due to iron is ca. 200 mV and no interference from peak overlap was observed (see Fig. 2). As the pH of the solution decreases, the iron peak moves closer to that of the copper and interference occurs. At pH 7 van den Berg and Huangs found it necessary to mask the copper with EDTA.A t .I -9 pg 1-1 Cu(1l) 0 -0.2 -0.4 -0.6 PotentialiV vs. 3 M Ag - AgCl Fig. 2. Voltammogram of Fe(II1) - catechol in the presence of Cu(I1) in a solution containing 50 mg 1-1 of boron. Accumulation time, 50 s; accumulation voltage, zero; pulse interval, 0.5 s; scan rate, 5 mV s-l; and pulse height, 50 mV 38 30 2 . 26t ,[ 14 / I I I I I 50 100 150 200 250 300 10 Accumulation timeis Fig. 3. Effect of copper(I1) on accumulation time/peak current of 2 pgl-IofFe(II1). A,2pg1k1Fe + 9 p g 1 - 1 C ~ ; B , 2 y g l - ~ F e + 4 p g l - ~ Cu; and C, 2 pg 1-I Fe less obvious form of interference can occur owing to competi- tion between the catechol complexes of copper and iron for adsorption sites on the mercury drop.This type of situation was observed with the nickel and cobalt complexes of dimethylglyoxime.1 One result of such site competition would be the onset of non-linearity in the peak current - accumu- lation period relationship at shorter accumulation times. No such effect was observed in a solution containing 2 pg 1-l of Fe(II1) in the presence of 9 pg 1-1 of Cu(I1) (Fig. 3), although for reasons that are not easy to explain there was a slight increase in sensitivity. This in itself presents no analytical problem as the method of calibration chosen was that of standard additions. Analytical Procedure, Calibration and Precision The analytical procedure, which was followed in the calib- ration and precision tests and also in the analysis of rig samples, was as follows.A known volume of sample, sufficient to give a final concentration of <lo0 mg 1-1 of boron when diluted to 10 ml, was added to the polarographic cell. De-ionised water was added to make the volume in the cell362 ANALYST, MARCH 1986, VOL. 111 t 0 .- 0.4 pg I-' Blank I I I I I I - 0.2 -0.4 -0.6 -0.8 PotentiallV vs. 3 M Ag - AgCl Fig. 4. Voltammograms of Fe(II1) - catechol complex from cali- bration in stirred solution containing 100 mg 1-l of boron. Conditions as in Fig. 2, except accumulation voltage, -0.2 V exactly 8.75 ml followed by 0.25 ml of 0.5 mol 1-1 mannitol solution. The polarographic cell was transferred to the electrode assembly and the contents of the cell were de- oxygenated for 8 min. At the end of this period 1 ml of 0.5 mol 1-1 Tris buffer and 80 pl of 0.1 mol 1-1 catechol solution were added and the solution was de-oxygenated for a further 4 min.The catechol was added at this stage because of its known instability in oxygenated solution. On completion of the de- oxygenation, the voltammetric sequence described previously was followed. A standard additions procedure was used for the determi- nation of iron in samples but this requires that there is a linear relationship between peak current, i,, and concentration at a particular accumulation time. The i, versus concentration relationship was investigated in synthetic coolant samples containing 100 mg 1-1 of boron with iron concentrations in the range 0.2-30 pg 1-l. Using a 50-s accumulation period at -0.2 V in stirred solution, voltammograms showed that there was a linear relationship between peak current and concentration up to about 10 pg 1-1.A regression analysis for the blank- corrected peak currents for 16 concentrations in the range 0.2-10 pg 1-1 gave the following equation: i, (nA) = 8.32~ + 1.01 with a regression coefficient r2 = 0.999, where c is the concentration in pg 1-1 in the analytical solution. Examples of these voltammograms are given in Fig. 4. The linearity of the calibration graph can be extended to higher concentrations by accumulating either for a shorter period or without stirring. A calibration was obtained at the same concentration of boron using an accumulation time of 75 s, without stirring, over the range 2-30 pg 1-1 of iron. A linear relationship was observed up to a concentration in the analytical solution of about 26 pg 1-1 and regression analysis of the data from 9 concentrations in the range 2-26 vg 1-l gave the following equation: i, (nA) = 4.88~ + 3.81 with a regression coefficient r2 = 0.997.The onset of non-linearity in both calibration graphs occurred at peak currents of about 80-90 nA. The linearity obtained in these calibration graphs indicated that a standard additions procedure, with a 50-s accumulation period in stirred solution, should be suitable for samples containing up to 6 vg 1-1 of iron. In unstirred solution, for an accumulation period of 75 s, standard additions calibration Table 2. Precision of determinations of peak current in the presence of 50 mg 1-1 of boron Concentration Accumulation Mean peak Standard of Fe/yg 1 - time/s currentt/nA deviation/nA Blank* 100 (stirred) 1.5 0.2 (13.3%) 1.7 100 (stirred) 18.1 1 .o ( 5.5%) 4.3 150 (stirred) 23.7 1.0 ( 4.2%) * No boric acid present.Accumulation voltage, -0.2 V; voltage t Mean of five determinations. scan, -0.2 to -0.7 V. Table 3. Soluble iron determined in samples from PWR rig Soluble Total Fe Sample Fe/yg I-' (on-1ine)lyg 1-1 A . . . . 7.3* 8.2 B . . . . 2.2t 2.6 * Analysed 90 min after collection. t Analysed 5 h after collection. can be at sample concentrations up to ca. 15 pg 1-1. These concentrations are based on samples containing 100 mg 1-1 of boron and a single standard addition, which will approxi- mately double the peak current. The standard deviations for a single measurement of peak current were obtained from voltammograms of a reagent blank solution (containing no boric acid) and solutions containing 1.7 and 4.3 pg 1-1 of Fe and 50 mg 1-1 of boron (see Table 2).The concentration of iron in the reagent blank solution was determined by standard additions to be 0.13 pg 1-1. Taking the statistical limit of detection at the 95% confidence level as 4.65 aB, where aB is the standard deviation of the blank, gives a limit of detection of 0.08 pg 1-1 (1.43 X 10-9 mol 1-I). Analysis of PWR Rig Samples The soluble iron content was determined in samples obtained, on two separate occasions, from a rig operating under PWR primary coolant conditions. The results of the voltammetric determinations were compared with those obtained from an on-line total iron monitor based on solubilisation with thioglycollic acid followed by a colorimetric determination using ferrozine.The samples were collected in polythene bottles to which had been added a small volume of hydro- chloric acid (ca. 1 ml of 5 moll-1 acid per 200 ml of sample) to stabilise the soluble iron. On each occasion, as the boron concentration was about 1100 mg 1-1, the sample was diluted 1 + 10 in the polarographic cell. The results are summarised in Table 3. Sample B was re-analysed 24 h after collection and the soluble iron content was 2.4 pg 1-1, indicating very little change on standing at ca. pH 2. A recovery test was carried out on this sample by adding to it sufficient Fe(II1) to increase the soluble iron by 2 pg 1-1; the resulting recovery was 104%. Discussion Purity of Reagents and Effect of Other Substances The iron concentrations in 10 ml of analytical solution resulting from the addition of 1 ml of 0.5 moll-1 Tris and 0.25 ml of 0.5 rnol 1-1 mannitol were 0.1-0.3 and S O .1 pg 1-1, respectively. Although some variations in the iron concen- tration of a reagent blank were observed from one batch of reagents to another, in no instance was there <0.1 pg 1-I. This relatively high concentration does influence the limit ofANALYST, MARCH 1986, VOL. 111 363 detection and should it be necessary to reduce the blank in the analytical solution the simplest alternative to improving the methods of purification would be to reduce the concentration of Tris and mannitol by a factor of two or three. The reduction in Tris will be accompanied by some small loss in sensitivity owing to the lowering of the pH of the analytical solution and the reduction in mannitol could lead to a loss in precision but there could still be an over-all gain in analytical range.The major interference with the complexing reaction of iron and catechol is from the reaction of borate; this was successfully removed by limiting the boric acid concentration in the analytical solution to ca. 10-2 mol 1-1, buffering the solution to pH 7.8 and adding a molar excess of mannitol. Under these conditions there was insufficient free borate ion to reduce the added catechol concentration of 8 x mol 1-1. There was no significant interference from the presence of other transition metal ions expected in PWR coolant.Iron(I1) ions were not expected to interfere on the basis of either their reduction potential or their complexation with catechol, but it was of interest to note that the bulk concentration of iron(I1) ions behaved polarographically in a manner identical with that for the same concentration of iron(II1) ions. This ability to detect total soluble iron is particularly important for PWR coolant sample where the main source of iron in solution is derived from magnetite in a hydrogenated boric acid medium such tnat Fe(I1) species are predominant. In any batch sampling procedure, the Fe(I1) will tend to be oxidised to Fe(II1) on exposure to the atmosphere, and therefore the capability of the technique reported here to determine both species as a single entity is a positive advantage.Calibration and Precision The sensitivity for iron in the sample, obtained for calibration using a simulated coolant containing 100 mg 1-1 of boron, was 7.3 nA (pg 1-1)-1 [equivalent to 4.1 nA per 10-8 mol 1-1 of Fe(III)] for an accumulation period of 50 s in stirred solution. Van den Berg and Huangs reported a sensitivity of 5.6 nA per 10-8 moll-’ in sea water for an accumulation period of 180 s. The poorer sensitivity was probably due to the high chloride concentration in the sample because when experiments were carried out in simulated sea water using Tris buffer the sensitivities we obtained were comparable to those reported.5 The relative standard deviations of the peak currents determined in simulated coolant of 5.5% at 1.7 pg 1-1 and 4.2% at 4.3 pg 1-1 were comparable to the range 5-7’/0 found for nickel and cobalt by DPVS.1 The standard deviation of a single determination of the iron concentration, based on the standard additions procedure, will obviously be greater than that obtained for a single determination of peak current as there are three current measurements (blank, sample and sample plus standard addition) to be considered.The three measurements of peak current that are used in a determination by the standard additions procedure are ib (peak current of the reagent blank solution), is (peak current of the sample solution) and is + a (the peak current of the sample plus a standard addition). Assuming that the peak currents are proportional to the concentration of the determi- nand in the analytical solution, then the concentration of the determinand in the latter, C,, can be expressed as or where Ca is the increment in concentration caused by the known standard addition.If bib, aiS and aiS + a are the standard deviation of ib, is and is + a, respectively, then an expression for the fractional standard deviation of the determinand can be derived from equation (1) by the method of propagation of errors9 as (6i: + 6ib2) (6is + 2 + 6s2) + * . (2) (is - ib)2 (is + a - i s ) 2 where 6cs is the standard deviation of a determination of the sample concentration. Under circumstances where ib is small relative to is, then (2r 2: Q + - (6is + a2 + 6is2) . . . . (3) i,2 (is + a - If the increment in concentration caused by the standard addition is made about equal to that of the sample and the assumption is made that the relative standard deviations of is and is + a are equal then equation (3) becomes (4) or From equation (5) the percentage relative standard devia- tion of a single determination by the standard additions procedure under these conditions is -- 100 6cs 6is -245- .. . . . . 0 L S I S On this basis the relative standard deviation of a single determination, from our precision results, would be of the order of 10-15%. The statistical limit of detection at the 95% confidence level, calculated from the standard deviation of the reagent blank solution, was 0.08 pg 1-1 and, as mentioned previously, any reduction in this would largely depend on lowering the iron concentration in the blank. However, these performance statistics are satisfactory for the levels of iron experienced in rig samples and anticipated in PWR coolant samples. The authors thank Dr. D. Midgley and Dr. C. M. G. Van den Berg for their helpful discussions on certain aspects of this work. This work was carried out at the Central Electricity Research Laboratories of the CEGB Technology Planning and Research Division and is published by permission of the Central Electricity Generating Board. 1. 2. 3. 4. 5. 6. 7. 8. 9. References Torrance, K., and Gatford, C., Talanta, 1985,32, 273. Lam, N. K., Kalvoda, R., and Kopanica, M., Anal. Chim. Acta, 1983, 154, 79. Van den Berg, C. M. G., and Huang, Z. Q., Anal. Chim. Acta, 1984, 164, 209. Van den Berg, C. M. G., Anal. Chim. Acta, 1984, 164, 195. Van den Berg, C. M. G., and Huang, Z . Q., J . Electroanal. Chem., 1984, 177, 269. Connor, J. M., and Bulgrin, V. C., J . Znorg. Nucl. Chem., 1967, 29, 1953. Perrin, D. D., and Dempsey, B., “Buffers for pH and Metal Ion Control,’’ Chapman and Hall, London, 1974. Moody, J. R., and Lindstrom, R. M., Anal. Chem., 1977,49, 2264. Parratt, L. G., “Probability and Experimental Errors in Science,” Wiley, London, 1961. Paper A51264 Received July 18th, 1985 Accepted October 9th, 1985

ISSN:0003-2654    

DOI:10.1039/AN9861100359

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, MARCH 1986, VOL. 111 365 Polarography-based Selective Titrations of Carboxylate and Phosphonate Ligands Used in Detergent Formulations Domenico Perosa, Maria Luisa Zanette, Franco Magno and Gin0 Bontempelli Department of Inorganic, Metallorganic and Analytical Chemistry, University of Padua, 35731 Padua, Italy Monoamperometric titration using dropping-mercury or platinum electrodes with periodic renewal of the diffusion layer of sequestrating agents used in detergent formulations has been performed with some metal ions. The choice of the titrant cations was made on the basis of their polarographic behaviour in comparison with that exhibited by the relevant complexes. Satisfactory results were obtained and are discussed. Keywords: Polarography; titrimetry; carboxylate complexes; phosphonate complexes; detergents In recent years, the need to reduce the content of poly- phosphates in detergents has increased owing to the eutrophi- cation phenomena caused by phosphorus present in waste waters.’ The consequent reduced use of polyphosphates implies their replacement in detergent formulations by other species that display similar properties.Carboxylate chelating agents such as ethylenediaminetetraacetic acid (EDTA) and nitrilotriacetic acid (NTA) are often employed for this purpose, as they are able to prevent trace amounts of heavy metal ions from catalysing the decomposition of perborates and, at the same time, maintain the cleaning activity (by complexing calcium and magnesium ions) of the reduced content of polyphosphates.More recently, the use of new chelating ligands containing phosphonate groups2J has been proposed, as they combine a high sequestrating power with good stability towards hydrolysis even at high temperatures and extremes of pH4 (important features for detergent components). Moreover, they exhibit very effective scale inhibition as the “threshold effect” operates for many differ- ent precipitating species (e.g., CaC03 and CaS04) with a performance much greater than that expected on the basis of mere sequestration. With the possible use of different complexing agents in detergent formulations, simple and rapid selective methods for their identification and determination are desirable. The analytical procedures reported so far are essentially based either on chromatographic procedures5.6 or on the formation of complexes with metal ions and subsequent detection by spectrophotometric or electroanalytical measurements.4>7-9 In this last instance, the potentiometric technique is generally employed, and polarography has surprisingly been neglected although it is characterised by selectivity, detection limits and reproducibility that are at least as good as those of poten- tiometry.In addition, a complete polarogram is also able to give much qualitative information about the simultaneous presence of different complexes whose reduction processes occur at different potentials depending on the stability constants. In view of these considerations, we undertook this investi- gation with the purpose of devising a convenient procedure for the titration of mixtures of sequestrating agents with potential use in detergent formulations; the detection of the end-point and the choice of the optimum experimental conditions are based on the use of polarography.The potential interferences from the presence of the different constituents of detergents (mainly tripolyphosphates) were also considered. Experimental Chemicals and Reagents All chemicals were of analytical-reagent grade, except the phosphonates, which were commercial products (Dequest 2010 and Dequest 2041; Monsanto). These last chemicals were repeatedly crystallised from aqueous acidic solution until a satisfactory purity was obtained as indicated by ion chromato- graphy.5 The solvent employed throughout was doubly distilled water and the different buffer media were prepared by adding hydrochloric acid to solutions containing a suitable weak base until the desired pH (in all instances near the value corre- sponding to half-neutralisation) was monitored by a glass electrode.When necessary, sodium perchlorate was added to these buffer solutions as a supporting electrolyte. A standard solution of copper ions, prepared by dissolving weighed amounts of pure metallic copper in nitric acid, was used (after suitable dilution) to standardise all the solutions of sequestrating agents. The solution of EDTA standardised in this way was then employed for the standardisation of Pb(NO& and FeC13 solutions. In all instances the end-points were detected monoamperometrically , employing either a dropping-mercury electrode (DME) or a platinum microelec- trode polarised at a suitable potential.Apparatus and Procedure Voltammetric and polarographic experiments were carried out in a three-electrode cell. The working electrode was either a platinum sphere, used with periodic renewal of the diffusion layer,lo or a DME with mechanical control of the drop time ( t = 3 s). The counter electrode was a mercury pool and the potential of the working electrode was probed by a Luggin capillary reference electrode compartment containing an aqueous SCE. The voltammetric unit employed was a three-electrode system assembled with MP-System 1000 equipment in con- junction with a function generator constructed in these laboratories. 11 The working potentials were monitored with a Keithley 168 Autoranging DMM digital voltmeter and the recording device was a Linseis LY-1800 X - Y recorder.The volumes of titrant, added from a 5-ml full-scale microburette, were measured to within 0.01 ml and the pH values were read on a Radiometer Model M84 pH meter. Unless stated otherwise, all the electroanalytical measure- ments were performed at room temperature and nitrogen, previously equilibrated to the correct vapour pressure, was used for the removal of dissolved oxygen. All titrations were performed in the concentration range 1 X 10-3 - 5 x 10-3 M and the relevant diffusion currents were suitably corrected for dilution. Results and Discussion Choice of Samples To test the potential of the proposed method, two different mixtures of sequestrating agents were analysed.366 ANALYST, MARCH 1986, VOL. 111 In the first .mixture (sample A), different proportions of EDTA, NTA and Dequest 2041 (ethylenediaminetetramethyl- enephosphonic acid; EDTMP) were present. This mixture was chosen on the basis of the following considerations: (i) EDTA is often used in detergent formulations to prevent the metal-catalysed decomposition of perborates and optical brightners; (ii) NTA is, at present, the most commonly used allowed replacement for polyphosphates; (iii) EDTMP is, according to Monsanto's suggestions,4 a very effective seques- trating species suitable for replacing polyphosphates in the future.0 0 - HO, II 11 ,OH Ho' II OH , CHzP \ HO/ pHPC\ HO\ PHzC/N(CH2)2N 0 0 EDTMP In the second type of mixture examined (sample B), diethylenetriaminepentaacetic acid (DTPA) and Dequest 2010 (1-hydroxyethylidene-1 , 1-diphosphonic acid; HEDP) were mixed in different proportions.DTPA was chosen because it gives complexes with metal ions characterised by the highest formation constants from aminocarboxylic acids12 and, conse- quently, its content in detergents may be kept low. Conver- sely, HEDP was chosen intentionally in view of the large number of complexes (mono- and polynuclear in nature) formed with several metal ions,13314 which are expected to make its determination by titration difficult. 0 II 0 II I 0-C-OH DTPA CH3 HEDP Choice of Titrant Cations The ligands employed can, in principle, be titrated by several cations. Consequently, a preliminary choice must be made by taking into account the following basic criteria: (i) the titrant cation should, with the chosen ligands, be able to give complexes that are much more stable than those formed with other potentially interfering species (e.g., tripolyphosphates) , in order to achieve a sharp end-point and a good selectivity; (ii) the cathodic reduction of the metal ion employed should occur at the lowest negative potentials possible so that a sufficiently large potential range is left available for the detection of the reduction displayed by the relevant complex, before the solvent discharge; in this way, the diffusion currents relative to these species can also be exploited for end-point detection, thus making the analysis of mixtures easier; (iii) a cation exhibiting reversible electrode behaviour is to be preferred as a high resolving power of the procedure can be achieved; (iv) cations that are soluble even at high pH should be employed to perform more selective titrations, thus taking advantage of the possibility of changing the pH over a wide range and of the masking effect of hydroxy ions.In order to satisfy more than one of these requirements simultaneously, Pb2+ , Cu2+ and Fe3+, exhibiting different features, were employed as titrants. 0 10 Q, 20 2- ? .- 30 40 0 10 Vlm I -2.0 -1.5 -1.0 -0.5 0 El V Fig. 1. (a Polarogra hic curves recorded during the titration of sample A dEDTAJ = k T A ] = [EDTMP] = 5 X M) in acetic acid - acetate buffer with Pb2+ solution. (1) Pb - EDTA reduction wave; (2) Pb - NTA diffusive reduction wave; (3) mixed wave from the Pb - NTA kinetic reduction and from free Pb2+ ions.( b ) Relevant monoamperometric titration plot. Currents were measured at -0.9 V vs. SCE 0 10 20 f -.. U . -2 30 40 0 10 Vlrn I -2.5 -2.0 -1.0 -0.5 EIV Fig. 2. (a) Polaro ra hic curves recorded during the titration of sample A ([EDTAf= fNT'A] = [EDTMP] = 5 x 10-3 M) in 6 x M NaOH solution with Pb2+ solution. (1) Pb - EDTA reduction wave; (2) HPb02 - reduction wave. (b) Relevant monoamperometric titration plot. Currents were measured at -0.85 V vs. SCE Titration of Sample A Inspection of the constants relative to complexes formed by the ligands employed with different metal ions,4,12,14 and their dependence on pH, suggested the use of Pb2+ as a titrant, also in view of its solubility as the hydrogen plumbite ion, HPbOZ-, in basic media.This cation should make possible, in principle, the titration of EDTA alone at pH values lower than 5, the determinationANALYST, MARCH 1986, VOL. 111 367 of the total EDTA and EDTMP content at pH values near 13 and the simultaneous titration of all the three sequestrating agents in an intermediate pH range (7-8). By carrying out titrations with lead ions on solutions of each individual sequestrating agent, and also on mixtures containing a known content of all the three ligands, it is possible to confirm these expectations only in acidic solution (pH 4.8, see Fig. 1) and in basic medium (pH 13, see Fig. 2). Conversely, the total ligand determination at intermediate pH values is precluded by the occurrence of a pre-wave (C.E. in naturel5) relative to the Pb - NTA complex that takes place at potentials very near to those appropriate for the reduction of free lead ions.This drawback can be overcome, however, by replacing Pb2+ with Cu2+ (pH 5.2; acetic acid - acetate buffer), the complex of which with NTA is reduced without exhibiting any kinetic complication. Therefore, the procedure we suggest for determining the individual ligands present in sample A requires the following three monoamperometric titrations: (i) the EDTA content is evaluated by titration with standard Pb2+ solution at pH 4.8 (0.1 M acetic acid - acetate ion buffer) employing a polarisation potential of -0.90 V vs. SCE for the DME; (ii) the total EDTA and EDTMP content is found at pH 13 (0.1 M NaOH) by titrating with standard Pb2+ solution and polarising the DME at a potential of -0.86 V vs.SCE; (iii) the total content of EDTA, EDTMP and NTA is determined by using Cu2+ as a 1 2 3 4 0 4 8 10 Vlm I Fig. 3. (a) Potentiometric and ( b ) monoamperometric titration plots for a solution of DTPA (2 x 10-3 M) in glycine buffer. Titrant: Fe3+ ions titrant at pH 5.2 (0.1 M acetic acid - acetate buffer) with the DME polarised at -0.06 V vs. SCE. In this way, very satisfactory results are obtained as shown in Table 1, where the reported values are the means of five replicate measure- ments. It is worth noting that copper ions cannot substitute lead ions totally in the procedure described as in basic solutions formation of the Cu - NTA complex also partially occurs, owing to the inadequate sequestrating power of ammonia, which must be used to keep Cu2+ ions in solution in alkaline media.Finally, it must be remarked that the proposed procedure is suitable for direct application to the alcohol-insoluble part16 of a real detergent. In fact, after dissolution of this part in alkaline solution and subsequent filtration to remove zeolites, the only species present in solution able, in principle, to interfere is tripolyphosphate. However, tabulated stability constants12 and literature reports8 indicate that this species does not cause overtitration under the experimental condi- tions adopted (titrant cations and pH values), and we verified this. Further, no problem arises from the very large excess of NTA with respect to EDTA found in real detergents as the two sequestrating agents present at lower concentrations (EDTA and EDTMP) are specifically titrated at suitable pH values.Titration of Sample B Also in this instance the choice of the titrant cation was made by first considering the formation constants of DTPA and HEDP complexes with different metal ions12J4 and their dependence on pH. Consequently, the use of the Fe3+ ion was studied in view of the high stability of the relevant complexes. Titrations of solutions containing DTPA alone in acidic medium (pH 2.8, glycine buffer) were satisfactory in that good monoamperometric or potentiometric end-point detection was easily achieved by employing a platinum indicator electrode (see Fig. 3). In contrast, titrations of HEDP gave incorrect results owing to the formation (at any acidic pH value) of different mono- and polynuclear complexes (par- tially insoluble), the relative concentrations of which appear to depend on the ligand content.Such behaviour is exhibited by HEDP with most metal ions13.14 and represents the main drawback to devising a suitable procedure for its determina- tion. This is the reason why it is recommended that the complexometric titration of this sequestrating agent is perfor- med with thorium ion17 (one of the few cations able to give only one complex with HEDP), even though this titrant cannot by employed in all analytical laboratories, its use being restricted by legal provisions. The use of Pb2f as a titrant at pH 6.7 (triethanolamine buffer) allows this problem to be solved by taking advantage of the very low solubility of the Pb2(HEDP) complex with respect to the other possible complexes, which leads to its preferential formation during the titration.On the other hand, this titrant when used at pH 6.7 is also suitable for the Table 1. Typical results obtained in the titration of sample A pH 4.8 Sequestrating Pb2+ Pb*+ EDTA . . . . . . . . 48.8 49.2 k 0.2 EDTMP - . . . . . . . . 120.0 NTA - . . . . . . . . 124.7 EDTA + EDTMP + NTA 293.5 - EDTA+EDTMP . . . . 168.8 - agent calculated/pmol found*/pmol pH 5.2 pH 13.0 cu2+ cu2+ calculated/pmol found*/pmol 24.5 - 22.6 - 25.5 - 72.6 72.6 f 0.1 47.1 - Pb2+ Pb2+ calculated/pmol found*/pmol 24.5 - 22.6 - 124.7 - 171.7 - 47.1 47.3 2 0.2 * Mean values of five replicate results.368 ANALYST, MARCH 1986, VOL. 111 ~~~ Table 2.Typical results obtained in the titration of sample B pH 5.1 pH 6.7 Sequestrating cu2+ cu2+ Pb2+ Pb2+ agent calculated/pmol found*/pmol calculated/pmol found*/pmol DTPA 32.0 31.9 & 0.1 128.0 . . . . . . . . . . . . - 54.1 - HEDP 135.5 DTPA + HEDP 167.5 - 182.1 181.7 ?I 0.2 . . . . . . . . . . . . - . . . . . . . . * Mean values of five replicate results. I -1.0 -0.5 E N Fig. 4. Polarographic curves recorded during the titration of sample B (DTPA = 5 x 1 0 - 3 ~ ; HEDP = 10-3 M) in triethanolamine buffer with Pb*+ solution. (1 Rising portion of the Pb - DTPA reduction wave; (2) Pb,(HEDP] reduction wave decreasing with time; (3) reduction wave for free Pb2+ ions quantitative determination of DTPA, so that the total DTPA and HEDP content can be easily evaluated under these experimental conditions, as illustrated in Fig.4. Before the equivalence point, this monoamperometric titration is not rapid owing to the slow precipitation of the phosphonate complex, which requires that the current measurements be made a few minutes after each titrant addition, thus allowing equilibrium conditions to be reached. It should be noted, however, that the more significant points of this titration are those beyond the equivalence point, which are not affected by this drawback. The separate contents of the two ligands might be determin- able by coupling the mentioned determination with a second monoamperometric titration carried out under the same experimental conditions, but at a working potential corre- sponding to the Pb - DTPA reduction (beyond -1.2 V; see Fig.4). However, the slow rate of formation of the Pb2(HEDP) precipitate again makes it necessary for a slow titration to be carried out (after the equivalence point in this instance). Therefore, it is preferable to determine the DTPA content by employing copper ions as the titrant, owing to the very large difference between the formation constants relative to the two sequestrating agents (KCuDTPA = 1022.5 and KCu2(HEDP) = 1012.5) and to the low stability of the phospho- nate complex in acidic solution. At pH 5 (0.1 M acetic acid - acetate buffer), the end-point for the titration of DTPA alone is in fact located correctly by polarising the working electrode corresponding to the nearly contemporaneous reduction of free copper ions and of the poorly formed phosphonate complex.Consequently, the content of the two sequestrating agents in sample B can be determined by coupling this copper-based determination of DTPA with the over-all titration of DTPA and HEDP performed with Pb2+ at pH 6.7 as mentioned above. Table 2 gives the results obtained with this procedure (means of five replicate measurements). Again, for the correct application of this procedure to real detergents, the presence of tripolyphosphate must be taken into account. In this instance, the selected pH value (6.7) does not allow their interference to be avoided. This problem can be eliminated, however, by resorting to the well established procedure that employes tin(I1) chloride to separate poly- phosphate ions by precipitation.16 On the other hand, Sn*+ ions do not give sufficiently stable complexes with DTPA and HEDP, so that they are unable to interfere in the subsequent titration of these sequestrating agents.Conclusion The results obtained in this paper clearly indicate that analytical procedures based on polarography are suitable for the qualitative and quantitative determination of sequestrat- ing agents employed in detergent formulations. Of course, better performances are expected from modern and more sophisticated electroanalytical techniques, particularly with regard to detection limits, and in order to verify this, and to evaluate the matrix effect, further work is in progress. Moreover, it must be remarked that two promising fields of application for this type of electroanalytical detection appear to be flow injection analysis and liquid chromatography. In particular, in the latter instance the information gained can be profitably exploited for enhancing the resolving power of ion chromatography by using voltammetric-type electrochemical detectors, thus maintaining the column separation capability with the selectivity appropriate for voltammetric analysis. Finally, it is worth comparing the non-equilibrium electro- analytical procedures proposed by us with the equilibrium method (potentiometry) widely used in this field.Potentiome- tric titrations are indeed easier to carry out even by untrained operators and, moreover, they require simpler apparatus. Notwithstanding this, non-equilibrium methods offer the following significant advantages: (i) they can also be used with non-reversible electrochemical processes; (ii) their response is faster because they do not require equilibrium conditions to be attained; (iii) the linear dependence of the current on the analyte concentration makes possible titrations at lower concentrations; (iv) the relatively more sophisticated apparatus used in monoamperometric titrations enables vol- tammetric tests to be carried out that provide qualitative information unattainable by potentiometric measurements.The supply of phosphonate ligands by Carini, Milan, is gratefully appreciated and financial aid from the Italian National Research Council (CNR) and the Ministry of Public Education is acknowledged.ANALYST, MARCH 1986, VOL. 111 369 1. 2. 3. 4.5. 6. 7. 8. 9. References Chiaudani, G., Gerletti, M., Marchetti, R., Provini, A . , and Vighi, M., “I1 Problema dell’Eutrofizzazione in Italia,” Quaderni dell’Instituto di Ricerca sulle Acque, No. 42, CNR, Rome, 1978. Kabachnik, M. J., Medved, I . Y . , Dyatlova, I. M., and Rudomino, M. V., Russ. Chem. Rev., 1974,43, 733. Maier, L., in “Proceedings of the 1st International Congress on Phosphorus Compounds, Rabat ,” Inst. Mond. Phosphates, Paris, 1977, p. 195. “Multifunctional Metal Ion Control Agents in Aqueous Solu- tions,” Technical Bulletin No. 53-39(E)ME-l, Monsanto, Brussels, 1983. “Determination of Sequestrating Agents,” Application Note No. 44, Dionex, Sunnyvale, CA, 1983. Rudling, L., Water Res., 1971, 5 , 831. King, T. M., and Mitchell, R. S., Special Report No. 7666, Monsanto, St. Louis, MO, 1970. Calapaj, R., Ciraolo, L., Corigliano, F., and Di Pasquale, S., Analyst, 1982, 107, 403, and references cited therein. “Trilon,” Technical Bulletin No. IT/P 2840, BASF, Ludwig- shafen/Rhein. 1983. 10. 11. 12. 13. 14. 15. 16. 17. Schiavon, G., Mazzocchin, G. A . , and Bombi, G. G., J. Electroanal. Chem., 1971, 29, 401. Magno, F., Bontempelli, G., Mazzocchin, G. A., and Patank, I., Chem. Instrum., 1975, 6, 239. Martell, A. E., and Smith, R. M., “Critical Stability Con- stants,” Volume 1, Plenum Press, London, 1974, p. 281. “Dequest 2010-Acide Phosphonique,” Technical Bulletin No. 53-34MEI(F), Monsanto, Brussels, 1980. Sillen, L. G., and Martell, A. E., “Stability Constants of Metal Ion Complexes,” Supplement 1, Chemical Society, London, 1971, p. 273. Heyrovsky, J., and Kuta, J., “Principles of Polarography,” Academic Press, New York, 1966, p. 367. Clinckemaille, G. G., Anal. Chim. Acta, 1968, 43, 520. King, T. M., and Maiec, E. J., Special Report No. 7158, Monsanto, St. Louis, MO, 1968. Paper A5186 Received March 7th, 1985 Accepted October 8th, 1985

ISSN:0003-2654    

DOI:10.1039/AN9861100365

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, MARCH 1986, VOL. 111 Sequential Micro-scale Determination of Sulphur in Organic Compounds Agostino Pietrogrande and Mirella Zancato Department of Pha rma ce u tical Sciences, University of Pad ua, 371 Chlorine (or Bromine) and 35131 Padua, Italy A comparison is made between the mercurimetric and the potentiometric determination of chlorides. As this comparison indicates the potentiometric titration with silver nitrate has higher accuracy and precision, this procedure has been combined with the bariometric method in order to achieve the sequential determination of halogens and sulphur in the same sample. The results obtained for several samples containing sulphur and chlorine (or bromine) confirm the effectiveness of the proposed method. Keywords: Sequential micro-scale determination; organo-chlorine; organo-bromine; organo-sulphur Simple and fast methods for the sequential micro-scale determination of halogens and sulphur in the same sample appear to be of practical utility as the presence of these elements in the same molecule is fairly frequently encountered in orgaqic compounds.These methods are expected to offer the advantages not only of a saving of time, but also of requiring the use of small amounts of sample. In previous studies on this subject, the combustion of the sample (either in a Schoniger flask,l-4 or in an oxygen stream5) was always followed by separate determinations of these elements. The bariometric titration is adopted for sulphur and halogens are determined as halides following different procedures, among which the mercurimetric method is the most widely used.In particular, a simple gravimetric determination was suggested by Hadzija and Kozarac.6 Some potentiometric titration procedures of considerable interest have also been proposed .7-10 This paper describes a convenient method for the sequential determination of halogens and sulphur, based on a combina- tion of potentiometric titration with silver nitrate for halogens and bariometric titration for sulphur, carried out successively directly in the combustion flask. The potentiometric titration of halogens is preferred to the mercurimetric method as higher accuracy and precision are obtained with the former approach. Experimental Equipment The combustion flask11 (300 ml) was modified in order to avoid the need for transfer of solutions and to make possible the easy titration of small volumes. The titrations were carried out with a Metrohm 636 Titroprocessor connected to a Model 635 automatic burette and a Metrohm EA 246 combined massive electrode. Procedure Determination of chlorine by the mercurimetric method The flask is charged with 5 ml of of 0.2 M sodium hydroxide solution and 0.5 ml of 3% hydrazinium sulphate solution, then 5-7 mg of sample are weighed in and the combustion is performed in the Schoniger flask in the usual way.After 30 min the flask is opened and 50 ml of propan-2-01 are added, carefully washing the stopper, the platinum basket and the inner side of the flask. After the addition of five drops of bromophenol blue (0.5% ethanolic solution), the solution is acidified with HC104 until a pale green colour persists, then 15 drops of diphenylcarbazone (0.5% ethanolic solution) are added and the titration of halides to a pale violet colour is performed with 0.005 N Hg(C10& with magnetic stirring. Determination of chlorine by the potentiometric method For the potentiometric titration after combustion, the stop- per, the basket and the inner side of the flask are washed with a mixture of 13 ml of propan-2-01 and 12 ml of distilled water.Three drops of methyl red (0.1% ethanolic solution), 0.1 M HN03 to a red colour, 3 ml of acetic acid (minimum 96%) and three drops of concentrated nitric acid are added. The electrode and the outlet tube of the burette are inserted to 1 cm above the bottom of the flask and magnetic stirring is started.The “control card operation” (programmed for the “dynamic titration kinetic T at value 3”) is inserted and automatic titration with 0.01 M silver nitrate solution is performed. Titration volumes and curves and potential break are printed out. The blank value must be subtracted from the titration volume. Sequential determination of chlorine (or bromine) and sulphur The absorption solution consists of 5 ml of distilled water and 1 ml of hydrogen peroxide (30%) and a sample containing not less than 0.7 mg of chlorine (or 1.5 mg of bromine) is weighed in. After the combustion, the stopper, the basket and the inner side of the flask are washed with 16 ml of propan-2-01, then two drops of Thorin (0.2 YO solution) and two drops of methylene blue (0.0125% solution) are added and the titration of sulphates is carried out to a pink colour with 0.01 M barium perchlorate with magnetic stirring.At the end of the titration, 5 ml of 0.2 M NaOH solution and 0.60 ml of hydrazinium sulphate (3% solution) are added. After waiting for 2 min, 5 ml of distilled water and three drops of methyl red (0.1% ethanolic solution) are added and the potentiometric titration of chlorides (or bromides) is performed by following the procedure described above. The blank value must be subtrac- ted from the titration volume. In the titration of bromides, correct values are obtained by subtracting the blank volumes from the second end-point. In this instance it is advisable to “activate” the electrode at the beginning of the analytical series by means of two preliminary titrations on solutions containing about 5 mg of potassium chloride or bromide.The salt must be dissolved in 30-35 ml of water - propan-2-01 (2 + l), to which all the mentioned reagents are added. Results and Discussion A preliminary comparison between the potentiometric and mercurimetric titration procedures was carried out by employ-372 ANALYST, MARCH 1986, VOL. 111 ing ten organic compounds characterised by different chlorine contents. The relevant results are reported in Table 1, which indicates a lower accuracy and a higher over-all relative standard deviation, i.e., a lower precision, for the mercuri- metric titration method (s = 0.15%) compared with that found for the potentiometric method (s = 0.08%).The results obtained for organic compounds containing both chlorine (or bromine) and sulphur are reported in Table 2. The solution employed for the absorption of the combustion products in the sequential determination of halogens and sulphur deserves some consideration. Some workers employ an alkaline solution of hydrogen peroxide,lJ>4 while the use of neutral H202 has been suggested in other reports.2.5J2 The aim of this reagent is to reduce free halogens and their related hypohalogenites (formed in the combustion and in the absorption process) to halide ions and, at the same time, to oxidise to sulphate the sulphur dioxide generated in the first of the mentioned processes. With reference to the standard potentials given in Table 3, it should be noted that an alkaline medium allows the oxidising power of hydrogen peroxide with respect to sulphur dioxide to be maintained, the potential of the H202 - H20 and Sod2- - S032- systems exhibiting the same dependence on pH. Similar arguments can also be employed for the reduction of hypohalogenites. Conversely, such an alkaline medium fav- ours the reduction of halogens to halides owing to the decrease in the potential of the O2 - H202 redox couple.Notwithstanding this, a quantitative reduction to halides is expected to occur also in the weakly acidic medium that originates during the combustion process, owing to the large reducing strength of hydrogen peroxide with respect to halogens and hypohalogenites. Consequently, the alkaline medium appears not to be essential. In spite of these considerations, and Burns and Maitin8 have recently suggested the replacement of hydrogen peroxide with hydrazine hydrate, which exhibits a stronger reducing power (see Table 3).The use of this reagent, however, makes the oxidation of sulphur dioxide difficult and consequently Mazzeo-Farina and Mazzeo3 suggested the elimination of the excess of hydrazine by an extended boiling procedure in the presence of hydrogen peroxide, which is slow and very tedious. In this investigation, we adopted the bariometric titration recommended by Fritz and Yamamura,13 which is reported to be strongly affected by the presence in solution of different ions, such as silver or alkali metal ions. Consequently, to allow the correct detection to the end-pont in the bariometric titration, the sulphate determinations were performed before the halide titration and an alkaline absorption solution was not employed.In conclusion, the recommended procedure involves the absorption of the combustion products (obtained in the Schoniger flask) by a neutral solution of hydrogen peroxide to which propan-2-01 is added at the end of the combustion process and the resulting solution is titrated directly for sulphates. At the end of this titration, hydrazinium sulphate is added to reduce possible residual traces of halogen or hypohalogenite and the solution is then titrated potentio- metrically with 0.01 M silver nitrate solution after acidification with nitric acid. Results for halogens and sulphur within +0.30% error are obtained, which confirm the effectiveness of the method.Of course, the measured volumes in the potentiometric titrations must be corrected for the chlorine contained in the filter-paper utilised in the combustion (ca. 35 pg). The amount of this correction (blank volume) can be easily determined by carrying out a combustion without the sample. No problems with this correction arise when the halogen to be determined is chlorine. Conversely, in the determination of bromine, the presence of chloride ions originating from the filter-paper implies that two distinct end-points are detected. In this instance the difference between these two end-points is greater than that expected on the basis of the amount of chloride ions determined in the independent filter-paper Table 1. Determination of chlorine Chlorine,% Mercurime try Potentiometry Compound Theoretical Found@) s,%* Found (2) s,%* Alprazolam, C7H13C1N4 .. . . . . . . . . . . 2-Amino-2’,5-dichlorobenzophenone, C13H,C12N0 . . Benzylisothiourea hydrochloride, C8HloN2S.HCl . . Chlordiazepoxide hydrochloride, C16H14C1N30. HCI . . p-Chlorobenzoic acid, C7HSC1O2 . . . . . . . . Diazepam, CI6Hl3ClN2O . . . . . . . . . . . . Furosemide, CI2HllC1N2OsS . . . . . . . . . . Temazepam, C16H13C1N202 . . . . . . . . . . l-Chloro-2,4-dinitrobenzene, C6H3C1N2O4 . . . . Levamisole, hydrochloride, C1 ]HI2N2S. HCl . . . . 11.49 26.67 17.49 21.11 22.65 17.50 12.46 10.73 14.74 11.80 11.70 26.46 17.60 20.98 22.64 17.57 12.50 10.99 15.00 12.00 0.14 0.19 0.14 0.19 0.17 0.19 0.20 0.05 0.15 0.04 11.41 26.56 17.64 21 .oo 22.68 17.44 12.36 10.79 14.71 11.85 0.07 0.08 0.21 0.15 0.02 0.03 0.02 0.07 0.03 0.10 *s = relative standard deviation (five replicate determinations).Over-all relative standard deviation for mercurimetry, 0.15% ; over-all relative standard deviation for potentiometry, 0.08%. Table 2. Sequential determination of chlorine (bromine) and sulphur Halogen,% Sulphur,% Compound Theoretical Benzylisothiourea hydrochloride, C8HIoN2S.HCI . . 17.49 Furosemide, C12HllC1N205S . . . . . . . . 10.74 Levamisole hydrochloride, CIIHI2N2S.HCl . . . . 14.74 Tetramisole hydrochloride, CllHI2N2.HC1 . . . . 14.74 Zambon 32/C 395, (halogen chlorine) . . . . . . 13.24 Bromocresol purple, Cl2Hl6Br2O5S . . . . . . 29.61 Bromothymol blue, C27H28Br20sS . . . . . . 25.60 Found (X) 17.65 10.76 14.84 14.80 13.35 29.51 25.50 s,% * (0.15) (0.09) (0.21) (0.13) (0.09) (0.19) (0.18) Theoretical Found (X) 15.82 15.78 9.68 9.61 13.33 13.25 13.33 13.18 11.97 11.82 5.94 5.89 5.13 5.07 s , O/O * (0.14) (0.14) (0.20) (0.11) (0.11) (0.10) (0.14) *s = relative standard deviation (five replicate determinations).Over-all relative standard deviation for halogens, 0.15% ; over-all relative standard deviation for sulphur, 0.13%.ANALYST, MARCH 1986, VOL. 111 373 Table 3. Standard potentials Redox couple H202- H20 . . . . 02-H202 _ . . . . . s o p - so 3 2- . . . . c12 - 2c1- . . . . . . c10- - c1- . . . . Br2-2Br- . . . . . . BrO- - Br- . . . , NHZNHZ-Nz . . . . Eo/V +1.77 +0.69 +0.15 +1.39 + 1 S O + 1.05 +1.19 -0.20 t Eq u iva lence point V - Fig. 1. Comparison between experimental (full line) and theoretical (broken line) generalised potentiometric curves relative to the sequential argentimetric titration of bromide and chloride ions correction and, at the same time, a negative error of the same order of magnitude is found for bromides if the titration volume relative to the first end-point is used.This result can be explained by considering that the sharp corner expected at the first equivalence point on the basis of theoretical considera- tions (see Fig. 1) is not found experimentally. In fact, it has a rounded form owing to the coprecipitation of AgBr and AgC1, which occurs near to this point,l4 the bromide concentration being suitably lowered. This roundness, occurring just where the maximum slope should be exhibited by the titration curve, therefore causes the first end-point to be detected with a negative error by the automatic device.However, as the second end-point always reflects the correct total content of Br- and C1-, the correct value for bromine can be easily found by subtracting “blank volume” from that corresponding to second end-point. In a previous paper11 dealing with the sequential poten- tiometric titration of bromides and chlorides in the same medium as in this work, we reported that the mentioned coprecipitation effect was not appreciable. In that instance, however, comparable concentrations of the two halides were employed, thus making negligible the relative error in the titration data caused by this phenomenon. Conclusions The potentiometric procedure allows the chloride determina- tion to be improved with respect to the mercurimetric methods.This result offers the possibility of combining such a potentiometric determination with the bariometric titration of sulphates for the effective sequential micro-scale determina- tion of organo-chlorine (or organo-bromine) and organo- sulphur with satisfactory accuracy and precision. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Boetius, M., Gutbier, G., and Reith, H., Mikrochim. Actu, 1958, 321. Giesselmann, G., and Hagedorn, I., Mikrochim. Actu, 1960, 390. Mazzeo-Farina, A., and Mazzeo, P., Microchem. J., 1978, 23, 137. White, D. C., Mikrochim. Actu, 1962, 807. Pella, E., Mikochim. Actu, 1961, 472. Hadzija, O., and Kozarac, Z . , Fresenius Z . Anal. Chem., 1975, 277, 191. Campiglio, A., and Traverso, G., Mikrochim. Actu, 1980, I, 485. Thorburn Burns, D., and Maitin, B. K., Analyst, 1983, 108, 452. Selig, W., Microchem. J . , 1976, 21, 291. Krijgsman, W., Griepink, B., Mansverld, J. F., and van Oort, W. J . , Mikrochim. Acta, 1970, 793. Pietrogrande, A., Zancato, M., and Bontempelli, G., Analyst, 1985, 110, 993. Dirscherl, A., and Erne, F., Mikrochim. Acta, 1963, 242. Fritz, J. S., andyamamura, S. S.,Anul. Chem., 1955,27,1461. Kolthoff, I. M., Sandell, E. B . , Meehan, E. J., and Brucken- stein, S., “Quantitative Chemical Analysis,” Macmillan, Lon- don, 1969. Paper A5i1.51 Received April 25th, 1985 Accepted September 2nd, 1985

ISSN:0003-2654    

DOI:10.1039/AN9861100371

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, MARCH 1986, VOL. 11 1 SHORT PAPERS 375 Determination of Trace Amounts of Copper, Lead, Thallium, Cadmium and Zinc in Pure Aluminium by Differential-pulse Anodic-stripping Voltammetry M. M. Palrecha, A. V. Kulkarni and R. G. Dhaneshwar Analytical Chemistry Division, Bhabha Atomic Research Centre, Modular Laboratories, Tromba y, Bombay 400 085, India A method for the determination of trace amounts of copper, lead, thallium, cadmium and zinc in pure aluminium metal by differential-pulse anodic-stripping voltammetry, employing a citrate buffer of pH 4.4 as the supporting electrolyte, is described. For the determination of copper, lead + thallium and cadmium, the pre-electrolysis was carried out at -0.80 V vs. an Ag - AgCl reference electrode. The overlapping of the lead and thallium anodic peaks was overcome by shifting the reduction potential of lead(l1) beyond -1.10 V by chelation with EDTA, while thallium(1) remained uncomplexed and was determined by carrying out deposition at -0.80 V.Copper and bismuth interference, if bismuth is present, can also be eliminated in the presence of EDTA, deposition of copper being carried out at -0.30 V. For the determination of zinc, the pre-electrolysis was carried out at -1 .I0 V. The effect of intermetallic compound formation between copper and zinc was investigated and the influence of probable impurities in pure aluminium was also studied. Keywords: Copper, lead, thallium, cadmium and zinc determination; aluminium analysis; differential-pulse anodic-stripping voltammetry There are diverse industrial and technological applications of pure aluminium metal, and the presence of trace impurities alters the the properties of pure aluminium considerably and adversely affects its utility A number of electrocLemica1 methods have been reported for the determination of various impurities in aluminium.Yoshimura1.2 analysed high-purity aluminium for trace amounts of copper, lead and zinc using square-wave polaro- graphy, however, the dissolution process is tedious and involved. An oscillopolarographic method for the determi- nation of copper, lead, cadmium and zinc was described by Beran et d . , 3 but their method does not take account of thallium if present. Neiman and Ponomarenko4 determined cadmium, indium, lead and copper in high-purity aluminium by inverse voltammetry with a graphite working electrode.Among different electroanalytical techniques, differential- pulse anodic-stripping voltammetry (DPASV) is the most sensitive and is eminently suitable for analyses at trace and ultra-trace levels. Although the use of a thin mercury film electrode (MFE) in conjunction with DPASV reduces the detection and determination levels of these metals to the p.p.b. and sub-p.p. b. range, it introduces interference effects owing to the formation of intermetallic compounds, particul- arly between copper and zinc as they are normally present in most matrices.5 This paper describes a method for the determination of trace amounts of copper, lead, thallium, cadmium and zinc in high-purity aluminium, taking into consideration two objec- tives: (i) the development of a procedure for decomposing a relatively large sample (about 2 g) that is simple and does not cause contamination or loss of the sample during its disso- lution and (ii) the choice of a suitable supporting electrolyte that prevents the hydrolysis of aluminium(II1) in solution and helps to separate interfering metal ions, namely lead(I1) and thallium( I).Experimental Apparatus A Model 174A Polarographic Analyser (Princeton Applied Research, USA) was used for the DPASV work. The electrode assembly consisted of a hanging mercury drop electrode (HMDE) (Metrohm, Switzerland), a platinum wire auxiliary electrode and an Ag - AgCl reference electrode. The cell assembly used was described earlier.6 All the experiments were carried out at 25 "C.Chemicals Aristar-grade hydrochloric acid (BDH Chemicals, UK) was used to dissolve aluminium metal. A standard aluminium solution was prepared by dissolving 99.9% pure aluminium metal (Johnson and Matthey, UK). Other chemicals were guaranteed reagents from E. Merck (FRG). Procedure About 2.0 g of high-purity aluminium filings were dissolved slowly in 50 ml of 6 M hydrochloric acid in a 200-ml Pyrex beaker. A few drops of 30% hydrogen peroxide and about 0.2 g of ammonium iron(II1) sulphate were added slowly to catalyse the dissolution. The mixture was initially boiled gently and then vigorously. After allowing the mixture to cool, a few drops of hydrogen peroxide were added and the mixture was boiled again. This procedure was repeated 2-3 times till the entire mass of aluminium metal had completely dissolved.The solution was then concentrated to a syrupy mass, with frequent stirring to avoid crust formation and bumping. After almost complete removal of hydrochloric acid, 50 ml of distilled water were added to the beaker and stirred to give a clear solution. The solution was then boiled for 5 min, cooled, transferred into a 100-ml calibrated flask and diluted to the mark with distilled water. The solution thus obtained was further purified by electro- lysing it for about 6 h at a mercury pool cathode at a constant potential of - 1.20 V vs. a saturated calomel electrode and a platinum wire auxiliary electrode. This solution served as the standard aluminium(II1) solution to which known amounts of metal solutions under investigation were added.Forensic samples of aluminium wire were dissolved in the same way. The pH of the electrolysis solution for DPASV work was maintained at 4.5 k 0.1. In order that aluminium (111) was not hydrolysed during the pH adjustment, the following pro-376 ANALYST, MARCH 1986, VOL. 111 cedure was adopted: a 2-ml aliquot of the standard aluminium solution was taken in a beaker and the pH was adjusted to 2 by adding dropwise 0.01 M sodium hydroxide solution. Then 2.5 ml of 1 M sodium citrate solution were added, the pH was adjusted to 4.5 _+ 0.1 and the solution diluted to 10 ml. Stripping Procedure After transferring the solution into the electrolysis cell and deaerating it with pure nitrogen (IOLAR grade, Indian Oxygen Ltd. , India), pre-electrolysis was carried out for 1 min at -0.80 V vs.Ag - AgCl while the solution was being stirred. After the solution had been allowed to rest for 30 s, an anodic scan at 5 mV s-l was applied, keeping the modulation amplitude at 50 mV. The anodic peaks for cadmium, lead + thallium and copper were obtained at -0.54, -0.34 and -0.08 V, respectively. A zinc anodic peak resulted at -0.88 V when the pre-electrolysis potential was kept at - 1.10 V (Fig. 1B). A corresponding reagent blank stripping current was also recorded (Fig. 1A) by applying a deposition potential of -0.80 V for cadmium, lead + thallium and copper and -1.10 V for zinc determination. These elements were determined by the method of standard additions to pure aluminium solutions (Fig. 1C). Results and Discussion As lead(I1) and thallium(1) ions have the same cathodic and anodic peak potentials in most non-complexing and weakly complexing media, they mutually interfere in each other’s determination. Dhaneshwar and Zarapkar6 used EDTA, which preferentially complexes lead(I1) and not thallium(1) , c u .c,d J v 10.5 PA (for Cd, Cu, Pb) >TI added to (B), EDTA added Pb I / - (for Cu, Pb, Cd, TI) , I I I -0.9 -0.7 -0.5 -0.3 -0.1 +0.1 E N VS. Ag - AgCl Fig. 1. Differential-pulse anodic-stripping voltammetric polaro- grams obtained with HMDE (pH 4.4). Eel = -1.1 V for Zn and -0.8 V for Cd, Cu, Pb and T1; electrolysis time = 1 min; rest period = 30 s: voltage scan = 5 mV s-1; pulse height = 50 mV. (A) Blank, 2.5 ml of sodium citrate solution (1 M) + HC1; (B) sample; and (C) standard additions: Cu, 0.574 pg; Pb, 0.5 pg; Cd, 0.045 pg; TI, 0.5 pg; and Zn, 1.0 pg in 10 ml of sample solution.Polarogram for TI recorded in EDTA to determine these metals by ASV in water and silicon. Bonelli et aZ.7 also complexed cadmium(I1) and lead(I1) with EDTA to eliminate their interference in the analysis of natural waters for thallium(1). After recording anodic-stripping peaks for copper, lead + thallium, cadmium and zinc in the citrate buffer (pH 4.4), about 0.5 g of the disodium salt of EDTA was added to the electrolysis cell solution and stirred to give a clear solution. Most of the EDTA was consumed by aluminium(II1) to form an Al(II1) - EDTA soluble complex. As the metal ions under investigation were present in trace amounts, the excess of EDTA was sufficient to complex them.Owing to strong complexation, the reduction potentials of lead(I1) and cad- mium(I1) shifted to potentials more negative than -1.20 V, while that of thallium(1) remained unchanged as it is not complexed. Hence the pre-electrolysis at -0.80 V and subsequent anodic stripping yielded an anodic peak that corresponded to thallium(1) alone. Synthetic samples were prepared by adding a number of metal ions to standard aluminium solution. Table 1 lists results for the determination of copper, lead, thallium, cadmium and zinc in synthetic samples to which, in addition to the metal ions under investigation, nickel (0.2), Cr (O.l), iron (1.0) arsenic (0.02), tin (0.1) and antimony (0.02 pg ml-1 in the electrolysis solution) were added.None of these elements was found to interfere. However, bismuth(III), if present in a 1 : 1 ratio or more relative to copper, interferes with the copper determi- nation, as the difference in their reduction potentials is only ca. 0.1 V. Normally, in high-purity aluminium, bismuth is present in extremely small amounts compared with copper; however, if as much bismuth as copper is present, the addition of EDTA again is helpful in overcoming the interference of bismuth on copper. In the presence of EDTA, copper and bismuth yield separate anodic peaks at -0.07 and +0.03 V, respectively, but the resolution of the peaks is not sufficient to make their determination quantitative. The cathodic peaks of copper(I1) and bismuth(II1) in the citrate buffer (pH 4.4) containing EDTA occur at -0.20 and -0.55 V, respectively.The pre-electrolysis potential, if kept at -0.30 V, will deposit copper alone, giving rise to a subsequent copper dissolution peak without any interference from bismuth.8 There are several reports dealing with the well known copper - zinc intermetallic formation and means of over- coming this interference effect in ASV applications.5>9JO Although the copper - zinc intermetallic formation is predomi- nantly noted when using a mercury film electrode (MFE), Copeland et aZ.5 observed it with even an HMDE, as they observed a slight diminution in zinc stripping current. This diminution was not sufficient to confirm the occurrence of copper - zinc intermetallic formation. During the analysis of Table 1. Recoveries of Cu, Pb, Ti, Cd and Zn added to synthetic aluminium solution.A1 concentration, 4 mg ml-1 Concentration Concentration Sample taken/ found/ Deviation, 1 Zn 0.1 0.1 - No. Metal pgml-1 pg ml-1 Yo Cd 0.045 0.04 11 Pb 0.05 0.06 20 TI 0.051 0.05 2 c u 0.029 0.029 - 2 Zn Cd Pb c u 3 Zn Cd Pb TI c u 0.2 0.09 0.102 0.057 0.4 0.18 0.20 0.204 0.115 0.23 0.091 0.102 0.059 0.41 0.18 0.19 0.21 0.108 15 1.1 3.5 2.5 5 5 6 - -ANALYST, MARCH 1986, VOL. 111 377 fly ash samples for copper, lead, cadmium and zinc, it was found that the zinc stripping current was affected by a change in the matrix of the sample rather than by intermetallic formation at an HMDE.11 The presence of Al(II1) was found to diminish the zinc peak up to an Al(II1) concentration about 70 times the Zn(I1) concentration in the solution; a further increase in Al(II1) concentration did not affect the zinc current.11 In this work, there was no change in the zinc stripping current although copper stripping currents were found to be 25-30% higher when deposition was carried out at -1.10 V (where both zinc and copper were deposited) than when deposited at -0.80 V. This could be attributed to the higher overpotential deposition in relation to copper in the former than in the latter experiment. By adopting a procedure of depositing cadmium, lead, thallium and copper at -0.80 V and zinc at -1.10 V, reliable results for these elements can be obtained. This also elimi- nates the possible interference due to irreversible oxidation of cobalt and nickel in the copper determination.This method was used satisfactorily to analyse various high-purity aluminium metal samples. A typical analysis of aluminium wire for source identification in forensic science is copper 21.0, lead 9, zinc 12 and thallium and cadmium less than 1 p,p.m. The copper concentration obtained by neutron- activation analysis was found to be 19.5 p.p.m. Conclusion Copper, lead, thallium, cadmium and zinc in pure aluminium were determined by differential-pulse anodic-stripping vol- tammetry in citrate buffer of pH 4.4. The dissolution of high-purity aluminium was catalysed by the addition of ammonium iron(II1) sulphate, which otherwise was tedious and elaborate. The elements other than zinc were deposited at -0.80 V and zinc was deposited at -1.10 V. The mutual interference of lead and thallium was overcome by carrying out pre-electrolysis at -0.80 V in the presence and absence of EDTA and for copper and bismuth the pre-electrolysis potential was required to be shifted to -0.30 V so that copper alone was deposited.The authors thank Dr. M. Sankar Das, Head, Analytical Chemistry Division, BARC, for his keen interest and encouragement and Dr. N. Chattopadhyay of CFSL (NAA Unit, Calcutta), BARC, for the neutron-activation analysis determination of copper. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. References Yoshimura, W., Bunseki Kagaku, 1972, 21, 4; Chem. Abstr., 1972, 77, 69753. Yoshimura, W., Bunseki Kagaku, 1981,30,347; Chem. Abstr., 1981, 95, 5415811. Beran, P., Dolezal, J., and Mrazek, D., J . Electroanal. Chem., 1963, 6, 381. Neiman, E. Ya., and Ponornarenko, G. B., Zh. Anal. Khim., 1973, 28, 1485; Chem. Abstr., 1974, 80,222988. Copeland, T. R., Osteryoung, R. A., and Skogerboe, R. K., Anal. Chem., 1974,46, 2093, and references cited therein. Dhaneshwar, R. G., and Zarapkar, L. R., Analyst, 1980, 105, 386. Bonelli, J. E., Taylor, H. E., and Skogerboe, R. K., Anal. Chim. Acta, 1980, 118, 240. Dhaneshwar, R. G., Kulkarni, A. V., and Zarapkar, L. R., to be published. Neiman, E. Ya., Petrova, L. G., Ignatov, V. I . , and Dolgo- polova, G. M., Anal. Chim. Acta, 1980, 113,277. Lazar, B., Nishri, A., and Ben-Yaakov, S., J . Electroanal. Chem., 1981, 125,295. Dhaneshwar, R. G., Kulkarni, A. V., and Zarapkar, L. R., paper presented at the Convention of Chemists, Jadhavpur University, October 1984. Paper A5157 Received February 13th, 1985 Accepted October 17th, I985

ISSN:0003-2654    

DOI:10.1039/AN9861100375

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, MARCH 1986, VOL. 111 379 Determination of Phenolphthalein in Pharmaceutical Preparations Using N-Bromosuccinimide Laila El Sayed, Loris 1. Bebawy and Mohmad M. Amer Analytical Chemistry Department, Faculty of Pharmacy, Cairo University, Kaser El Aini, Cairo, Egypt Phenolphthalein was determined by direct titration in 0.04 M sodium hydroxide solution with N-bromosuccinimide. Micro-elemental analysis and infrared spectroscopy of the reaction product indicated that it was tetrabromophenolphthalein. The method was compared with the US National Formulary method. Amounts of 0.318-1 1.13 mg of phenolphthalein were determined by this method, the recovery being 100.13 k 0.72%. The compound was determined in several pharmaceutical formulations. Other laxatives found in combination with it, such as aloin and podophyllin, do not interfere.Keywords: Phenolphthalein determination; titrimetry; bromination; laxative determination; N - bromosuccinimide Phenolphthalein is used as a laxative (cathartic) in many pharmaceutical formulations and has been determined gravi- metrically,l i~dimetrically,~J polarographically4~5 and spectrophotometrically.6~7 The method proposed in this paper is carried out by direct titration of phenolphthalein in 0.04 M sodium hydroxide solution using N-bromosuccinimide. The method was used for the determination of phenolphthalein in pharmaceutical preparations. Experimental Materials Authentic samples of phenolphthalein were obtained from BDH Chemicals and its purity was determined by the BP method.7 N-Bromosuccinimide was obtained from E.Merck and its purity was determined iodimetrically.8 Pharmaceutical preparations were obtained from the com- panies indicated in Table 2. Procedures Prepare a 0.001 M solution of phenolphthalein by dissolving 31.8 mg in 100 ml of 0.04 M sodium hydroxide solution. Prepare a 0.004 M solution of N-bromosuccinimide in distilled water and standardise it iodimetrically . For the determination of phenolphthalein in tablets, grind 20 tablets, mix well, weigh an amount equivalent to one tablet and extract it with 0.04 M sodium hydroxide solution in a 100-ml calibrated flask, using three 15-ml volumes of extractant, then dilute to volume with distilled water. For pills containing phenolphthalein, remove the outer coating, then proceed as described above. For emulsions and chocolate laxatives extracted according to the method of Allen and Johnson9 dissolve the residue obtained from each extraction in 100 ml of 0.04 M sodium hydroxide solution.Determine phenolphthalein in authentic samples and phar- maceutical preparations by taking a volume equivalent to 0.318-11.13 mg (1-3.5 pmol) of phenolphthalein and titrate with 0.004 M N-bromosuccinimide solution. The end-point is shown by a change in colour from pink to bluish green. Each 1 ml of 0.004 M N-bromosuccinimide solution is equivalent to 0.318 mg of phenolphthalein. A microburette may be needed for the titration. Results Tables 1 and 2 show the recoveries of authentic phenolphthalein’ samples and pharmaceutical preparations. The advantages of the method are that it is rapid, accurate and needs only simple equipment.It is also applicable to pharmaceutical preparations and other cathartics found, such as aloin and podophyllin, do not interfere. Discussion N-Bromosuccinimide has been widely used as a reagent for the quantitative bromination of phenols ,8 and the mechanism of bromination has been found to be electrophillic substitution reactions that take place in all the available ortho and para positions.8 Barakat et al. 10 described procedures for the micro-determination of salicylic acid , phenol, vanillin and thymol by direct titration with N-bromosuccinimide. However, nothing has been cited in the literature on its use for the determination of phenolphthalein, which possesses two phenolic moieties. In phenolphthalein there are four ortho positions available for bromination.It was found that phenolphthalein reacts with N-bromosuc- cinimide in a molar ratio of 1 : 4, four bromine atoms being consumed during the titration, as demonstrated by micro- Table 1. Determination of phenolphthalein by the National Formulary Method and the N-brornosuccinimide method National formulary method3 Taken/ Found/ recovery, 10 101 S O 15 99.30 25 99.10 50 100.90 100 101.00 150 98.90 mg mg YO - - - - - - - - - - Mean . . . . . . 100.12 Standard deviation . . 1.14 Fiducial limit . . . . k1.19 N-Bromosuccinimide method Taken/ Found/ Recovery, 0.320 0.317 99.06 0.640 0.639 99.81 0.960 0.953 99.27 1.280 1.274 99.53 1.600 1.580 98.75 3.200 3.180 99.38 4.800 4.840 100.83 6.400 6.410 100.16 8.000 8.140 101.75 9.600 9.700 101.04 11.200 11.400 101.79 100.13 1.07 k0.72 mg mg YO380 ANALYST, MARCH 1986, VOL.111 ~~ ~~ ~~ ~~ ~~~ ~ Table 2. Determination of phenolphthalein in some commercial preparations using the direct titrimetric method with N-bromosuccinimide Sample analysis Control experiment Amount per tablet/mg Preparation Normalin tablet* . . . . . . Labelled Found 50 48.07 48.59 48.27 Boldolaxinet . . . . 75 Alphalaxine tablet? . . . . . . 35 Brooklax . . . . . . 130 Paragar emulsiont . . . * Misr Company. t Kahira Company. 78.00 79.50 79.50 37.22 37.83 38.50 129.58 129.32 129.79 20 per mg 194.73 per 15 ml 197.66 197.86 Standard Found/ Recovery, addedlmg mg % 50 49.37 98.75 50.16 100.30 49.84 99.69 30 29.78 99.20 30.30 101.00 29.60 98.60 50 50.16 100.30 50.78 101 S O 50.35 100.70 30 30.30 101.02 30.40 101.30 30.56 101.80 Mean ( P = 0.5): 99.59 * 0.98 Mean ( P = 0.05): 101.10 5 0.61 50 48.59 97.18 50.16 100.32 48.98 97.96 30 30.30 101 .oo 29.78 99.27 30.10 100.33 50 50.68 101.36 50.55 101.10 50.16 100.32 30 29.78 99.27 30.30 101 .oo 30.39 100.30 Mean ( P = 0.05): 99.34 k 1.58 Mean ( P = 0.05): 100.56 * 0.71 32 31.70 99.06 32.32 101.00 32.32 101.00 Mean ( P = 0.05): 100.35 L 1.17 3800 3000 2000 1500 1000 500 Waven u m bericm - 1 Fig.1. minated product Infrared spectra of (a) phenolphthalein and (6) the bro- elemental analysis of the product of the reaction. The infrared spectrum of the product (Fig. 1) was characterised by a bromine peak at 740 cm-1. The reaction was carried out in Table 3. Statistical comparison of the results obtained by the proposed direct titrimetric method with that of the National Formulary3 Method of assay National method Formulary Proposed method Mean recovery, % No.of determinations . . . . 6 11 Variance . . . . . . . . . . 1.37 1.14 t . , . . . . . . . . . . - 0.1786 (2.131)* F . . . . . . . . . . . . - 1.202 (3.6)* 0.05. (P=O.O5) . . . . . . . . 100.03+1.23 100.13k0.72 * Figures in parentheses indicate theoretical values oft and Fat P = sodium hydroxide medium and the solubility of phenolph- thalein increased with increase in alkalinity. The optimum concentration of the sodium hydroxide solution used was 0.04 M, where no fading of colour was observed during the titration. At lower sodium hydroxide concentrations the solubility of the compound was poor.From a statistical comparison of the mean results obtained using the proposed method and the US Natiopal Formulary method,3 with a 95% confidence limit for the difference, it is clear that it lies within the limits that are in agreement with the test of significanceANALYST, MARCH 1986, VOL. 111 381 (Table 3). Therefore, one can conclude with 95% confidence that there is no significant difference between the accuracy of the two methods. Other laxatives found in combination with phenolphthalein in pharmaceutical preparations, such as aloin and podophyl- lin, do not interfere as being anthraquinone glycosides. Emulsions and coloured preparations gave good results (Table 2) using the extraction method of Allen and J o h n ~ o n . ~ Trials to detect the end-point by potentiometry using two platinum electrodes were carried out, but no sharp inflection in the titration curve was observed. 1. 2. 3. References Bickford, C. F., and Schoetzow, R. E., J . Am. Pharm. ASSOC., 1936,25, 1128. Warren, A. T., Logun, J. E., and Thatcher, R., J . Am. Pharm. Assoc., 1950, 39, 10. “The National Formulary,” Fourteenth Edition, American Pharmaceutical Association, 1975, p. 563. 4. 5 . 6. 7. 8. 9. 10. Kolthoff, I. M., and Lechmicke, D. I . , J. Am. Chem. SOC., 1948, 70, 1879. Suzuki, M . , J. Electrochem. SOC. Jpn., 1954, 22, 220; Chem. Abstr., 1954, 48, 134736. Gustafson, J. H., and Benet, L. Z., J . Pharm. Pharmacol., 1974, 26, 937. “British Pharmacopoeia 1980,” HM Stationery Office, Lon- don, 1980, p. 342. Mathur, N. K., and Narang, C. K., “The Determination of Organic Compounds with N-Bromosuccinimide and Allied Reagents,” Academic Press, London, 1975, pp. 54-59. Allen, J., and Johnson, C. A.,J. Pharm. Pharmacol., 1962,14, 73T. Barakat, M. Z., Fayzalla, A. S . , and El-Aassar, S . T., Analyst, 1972, 97, 470. Paper A5f88 Received March 11 th, 1985 Accepted September 16th, 1985

ISSN:0003-2654    

DOI:10.1039/AN9861100379

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, MARCH 1986, VOL. 111 BOOK REVIEWS 383 Modern Methods for the Determination of Non-Metals in Non-Ferrous Metals C. Engelmann, G. Kraft, J. Pauwels and C. Vandecasteele. Pp. xiv + 410. Walter de Gruyter. 1985. Price DM190; $76. ISBN 3 11 010342 7; 0 89925 010 6 During the past 10 years, a collaborative exercise sponsored by the Commission of the European Communities has been conducted with the co-operation of a few highly qualified University and industrial laboratories throughout Western Europe. The scope of this exercise has been directed towards the determination of non-metallic impurities in non-ferrous metals with the specific objective of improving the accuracy and precision of analysis. This volume summarises the collective efforts of the contributing laboratories in applying chemical and physico-chemical methods of analysis to the determination of carbon, sulphur, phosphorus, boron, oxygen and nitrogen in a wide range of non-ferrous alloys based on aluminium, copper, lead, nickel, molybdenum, niobium, tantalum, titanium, tungsten and zirconium.Chapter I outlines the influence of non-metals on various physical properties of non-ferrous alloys including conduc- tivity, hardness, tensile strength, ductility and embrittlement and considers the manner in which these factors can affect production/fabrication. This information will appeal to the practising analyst seeking technical data beyond hidher fortk. Chapter I1 considers nuclear methods of analysis with specific reference to neutron activation, charged particle activation and photon activation analysis.The theoretical bases for these techniques are discussed in depth and factors that affect accuracy, precision and sensitivity are evaluated. Chapter I11 deals extensively with sample preparation-a vital precursor to any method of analysis for trace element determination in alloys based on metals with a high affinity for oxygen and nitrogen-detailing optimum conditions to mini- mise contamination. Considerable attention is paid to machine tool parameters, compositions and coolants with specific recommendations for the preparation of millings/ chippings or solid pieces. Chapters IV and V consider methods used for the determi- nation of boron and carbon, respectively. Six spectropho- tometric methods in addition to flame emission and plasma excitation are described for boron in various matrix applica- tions.Combustion in aidoxygen provides the most satisfactory means of releasing carbon from non-ferrous matrices, with detector systems utilising coulometry, conductivity, man- ometry and titrimetry. Choice of flux material is discussed extensively together with the potential contribution to “blank” values derived from crucibles, boats, tubes, com- bustion gas, etc. Chapters VI (nitrogen) and VII (oxygen) constitute the bulk of the text. Methods for the determination of both these elements are based predominantly on inert gas or vacuum fusion principles. For nitrogen, chemical methods incorpor- ating Kjeldahl distillation followed by titrimetric or spectro- photometric determination are also included.Chapters VIII and IX discuss phosphorus and sulphur, respectively, in a brief but succinct manner. In general, nuclear techniques feature extensively through- out the entire text and tables of collaborative results abound. All individual methods of analysis described contain meticu- lous working detaiIs. The book is produced in an attractive cover and includes a detailed list of contents supported by an adequate subject index; 661 references are cited in a text that is both easy to read and absorb and is refreshingly free from typographical and publication errors. M. S. Taylor Mass Spectrometry in Environmental Sciences Edited by F. W. Karasek, 0. Hutzinger and S. Safe. Pp. xx + 578. Plenum Press. 1985. Price $75. ISBN 0 306 41552 6. This book consists of a series of reviews by “experts in the subject matter” covering the field of environmental applica- tions of mass spectrometry in some depth.It is laid out starting with a series of seven chapters on the general principles of mass spectrometric techniques such as gas chromatography - mass spectrometry (GC - MS), soft ionisation methods (positive and negative ion chemical ionisation, field desorp- tion), and data retrieval and interpretation. The next two chapters cover the general application of mainly GC - MS in water and air pollution studies followed by the bulk of the book, 16 chapters in 360 pages, containing detailed reviews on the analysis of particular classes of environmental pollutants. Unlike too many books on mass spectrometric applications, the early chapters do not set out to give a detailed description of a mass spectrometer and how it works, but are rather a series of illustrations of the state of the art around 1980 or so and its use in environmental analysis.This does not detract from the book, as exhaustive monographs of the theory of mass spectrometry are available elsewhere. The unusual chapter on data retrieval is almost wholly concerned with the computer interpretation of spectra using library matching and data enhancement. The techniques described are based on the HP5992 GC - MS - computer system but do provide a description of the various library searching methods used by most manufacturers. The chapter on water pollution studies by Pellizzari and Bursey is largely devoted to the development of the EPA’s general method for the detection and determination of organics in water.The next chapter, on air pollution, gives a much fuller coverage of the methods used to date. In neither chapter does there appear to be a reference to the EPA’s priority pollutants guidelines, which is a puzzling omission in a work such as this. The various in-depth reviews covering particular classes of compounds in general appear to give a comprehensive picture up to 1980. A few later references are given in some chapters but most fall into the period 1975-80. The individual reviews are variable in content and coverage. Thus, Sweetman and384 ANALYST, MARCH 1986, VOL. 111 Karasek’s chapter on polycyclic aromatic hydrocarbons deals almost exclusively with the mass spectral properties whilst the chapters on polychlorinated biphenyls and dibenzodioxins and dibenzofurans cover not only the mass spectral properties but also the metabolism and breakdown products plus the problems of analysis of interfering compounds in GC - MS.This is illustrated in the chapter on dioxins, which gives a table listing nine possible interfering compounds whose spectra contain ions with the same nominal mass as the molecular ions for TCDD. The chapters on volatile halocarbons, DDT, organophos- phorus compounds , pyrethroids and insect pheromones are also more extensive in their coverage than just a description of mass spectral properties, detailing in addition such aspects as quantitative methods, derivatisation prior to GC - MS and (pheromones) chemical manipulation to determine the posi- tions of any unsaturated bonds. An Appendix lists a number of chemicals with an indication of the ionisation method used and the page on which it appears. It is always difficult in a review such as this to publish quickly, and to some extent the work described is dated. Recent introductions such as fast atom bombardment ionisa- tion, target compound analyses using on-line computers and metastable peak monitoring are not discussed. Nevertheless, this book gives a useful and thorough coverage of the field up to 1980 and is sure to find a place on the bookshelves of laboratories carrying out environmental analyses. N . J . Haskins

ISSN:0003-2654    

DOI:10.1039/AN9861100383

出版商:RSC    

年代:1986

数据来源: RSC