H2N Me Me Me NH2 Me N N OH N O OH O OH O OH O ( a) ( b) Sorption–Catalytic Determination of Manganese Directly on a Paper-based Chelating Sorbent M. K. Beklemishev, T. A. Stoyan and I. F. Dolmanova* Analytical Chemistry Division, Department of Chemistry, Moscow State University, Moscow 119899 GSP, Russia. E-mail: mkb@analyt.chem.msu.su A hybrid technique based on a catalytic reaction carried out on the surface of a paper-based sorbent is proposed. It is shown that MnII exhibits its catalytic action in the oxidation of 3,3A,5,5A-tetramethylbenzidine with periodate in aqueous solution as well as on filter-paper with or without attached diethylenetriaminetetraacetate (DETATA) groups.Optimum conditions differ for the reaction in solution and on filter-paper. For equal catalyst concentrations, higher initial reaction rates are attainable on the filter-papers. Preconcentration of manganese on the DETATA sorbent combined with the subsequent catalytic reaction improves the selectivity, reduces the limit of determination down to 5 3 1026 mg l21 (as compared with 6 3 1025 mg l21 in solution) and expands the linear range by an order of magnitude (to 5 3 1026–2.5 3 1023 mg l21).The precision of manganese determination on the sorbent is high (the RSDs are @5% for !6 3 1024 mg Mn). Samples of tap and river water were analyzed by use of the proposed sorption–catalytic technique. A rapid procedure for the determination of manganese with visual detection was also developed. Keywords: Catalytic kinetic method of analysis; chelating sorbent; diethylenetriaminetetraacetate; preconcentration of manganese Catalytic procedures are appreciated by analytical chemists for their sensitivity and simplicity in realization.Catalytic indicator reactions can be applied to the determination of a large number of compounds including the catalyst, inhibitors, activators, and compounds which convert the catalyst into an active state (oxidizers, ligands1), and catalytically inactive metal ions through the use of the competitive complexation principle.Catalytic methods become especially powerful when combined with analyte separation/preconcentration, which allows the selectivity to be increased and the detection limits to be lowered. One of the approaches involves adsorption of metal ions on chelating sorbents with subsequent desorption and catalytic determination.2 However, this type of procedure, while being convenient in flow systems,2 is tedious in other cases owing to the need for desorption of the analyte.Here, we propose to eliminate the desorption stage by conducting the catalytic indicator reaction directly on the sorbent used for preconcentration purposes. Little information has been published to date on the foregoing principle.3 This novel area requires studies of the peculiarities of catalytic indicator reactions as they occur on the surfaces of sorbents, including kinetics, concentration effects, and the metrological characteristics of the procedures.The main purpose of this study was the practical realization of the aforementioned sorption–catalytic principle. For the investigation, the preconcentration of manganese on diethylenetriaminetetraacetate (DETATA) paper-based chelating sorbent was chosen as it combines effectiveness of metal sorption and simplicity of operation.2,4 As an indicator reaction, the oxidation of 3,3A,5,5A-tetramethylbenzidine (TMB) by potassium periodate (KIO4) catalyzed by MnII was selected.Various amines other than TMB were previously studied in this reaction;5–7 however, TMB is more stable in air, less carcinogenic and its oxidation provides higher absorbing products. In the present work, the catalytic effect of MnII on the reaction between TMB and KIO4 in solution was studied. The aim was to employ this system to ascertain the analytical possibilities of the sorption–catalytic approach for the determination of manganese with both instrumental and visual detection.A potential benefit of the sorption–catalytic approach may be realized in ‘rapid-test’ techniques, such as spot-tests, field tests and similar procedures,8 in which a visible signal appears on the surface of paper strips or other supports. The development of rapid tests for manganese may be a subject for further work originating from this study. Experimental Reagents, Solutions and Apparatus TMB ‘for analysis’ was obtained from Riedel-de H�aen, Hannover, Germany; boric, hydrochloric and sulfuric acids (‘special purity’) from Reakhim (Moscow, Russia) were used.All other reagents were of analytical-reagent grade. Humic and fulvic acids were obtained from Moscow river water by adsorption on poly(styrene–divinylbenzene)sorbent (XAD) with subsequent desalination. All aqueous solutions were prepared using distilled water. Ethanolic solutions of TMB (0.025 mol l21) were prepared every 10 d; less concentrated solutions were obtained by dilution with ethanol when necessary.A stock solution of periodate in water (1 g l21 KIO4) was diluted to an appropriate concentration (in most cases 0.1 g l21) every 4–5 d. A stock solution of anhydrous MnSO4 (1 g l21 Mn) was standardized by titration.9 Solutions with lower contents of the metal were prepared every 3 d (10 and 1 mg l21) or daily (0.01 mg l21) by dilution of the stock solution. All the manganese solutions were acidified with H2SO4 to pH 1.8–1.9; it was found that such acidification was necessary in order to obtain reproducible results.Hydrochloric acid was used to provide pH values ranging from 2 to 4; it was purified by isothermal distillation of the concentrated acid (reagent grade). Borate buffer of pH 6.8 was prepared by dropwise addition of Scheme 1 (a) TMB; (b) DETATA groups attached to filter paper. Analyst, October 1997, Vol. 122 (1161–1165) 11610.2 mol l21 KOH to a solution of 3.3 g of boric acid in 1000 ml of water until the desired pH value was reached.Buffers for the pH range 4–6 (0.1 mol l21 in sodium) were prepared from acetic acid and sodium acetate, and buffers for pH values of 7–11 (0.1 mol l21 in borate) were prepared from boric acid, Na2- B4O7·10H2O and KOH, by using standard procedures.10 The other buffer solutions with pH > 4 were purified by pumping through two consecutive DETATA filters placed in a special polyethylene holder (‘Biospectr’, St.Petersburg, Russia) at a rate of 8–10 ml min21, which removed transition metal impurities. The chelating sorbent was synthesized as described previously by chemical attachment of aminocarboxylic (diethylenetriaminetetraacetate, DETATA) groups to cellulose filterpaper. 11 The filter-paper was 2.5 cm in diameter. Water was distilled in a commercial apparatus (D3-4-2M, Ioshkar-Ola, Russia). A KFK-3 spectrophotometer (ZOMZ, Zagorsk, Russia) was used for measurements of absorbance. To spray the filter-papers with periodate solution, a hand-operated sprinkler for thin-layer chromatograms was used.Reaction in Solution Preliminary studies of various mixing orders of the components of the indicator reaction showed that maximum absorbance was achieved when MnII and borate buffer (pH 6.8) were mixed first, followed by addition of TMB and finally of KIO4. Based on this, the following procedure was used.In a 15 ml glass test-tube were placed the reagents in the following sequence (the amounts are given for the final procedure for manganese determination): 3.9 ml of the buffer (pH 6.8), 0.5 ml of the solution, containing MnII at a pH of 1.8–1.9, 0.1 ml of a 2.5 3 1024 mol l21 ethanolic solution of TMB and 0.5 ml of 4.3 3 1022 mol l21 periodate. The addition of periodate was taken as the time of the start of the reaction. After agitation, the reaction mixture was transferred into a 0.5 cm cell and the absorbance of the oxidation product was measured at 650 nm against water.The absorbance value at 3 min (A3) was taken as the analytical signal. The experiments were performed at ambient temperature, which was kept at 24 ± 1 °C. Reaction on Filter-paper In preliminary experiments the optimum mixing order of the reactants was sought witthe use of visual detection. Coloration of the filter-paper was most intense when TMB and the buffered manganese solution were mixed first and periodate was added at the last step.The role of the humidity of the filter-papers was also studied (the filter-papers were stored for up to 3 d in a desiccator with controlled 75–95% humidity); no effect of humidity on the reaction rate was observed. The final procedure for the reaction on filter-paper (with or without attached DETATA groups) is given below. MnII was first applied onto a filter-paper either (a) in the form of a 0.02 ml aliquot of a solution (0.4 mg l21 Mn) added to the centre of the filter-paper by an Eppendorf pipette, or (b) by pumping a buffered manganese solution of pH 6.8 (5 3 1025–0.05 mg l21 MnII, 0.08 mol l21 in borate) through the filter-paper by use of a peristaltic pump at a rate of 8 ml min21.The filter-paper was then dried with a gentle flow of pressurized air until no traces of moisture could be detected, after which TMB was pipetted onto the filter-paper (0.02 ml of a 4 3 1024 mol l21 solution) followed by an identical drying procedure.The oxidant was added by sprinkling the filter-paper with a 4.3 3 1024 mol l21 solution of KIO4, which was taken to be the beginning of the reaction. The coloured reaction zone was about 20 mm in diameter. Sprinkling rather than pipetting of periodate was necessary in order to obtain uniform coloration of the filterpapers; the sprinkling technique was applicable owing to a weak dependence of the analytical signal on the amount of KIO4 on the filter-paper, as shown below.The absorbance of a filter-paper was measured in its wet state against the same filter-paper without added periodate (i.e., not coloured with the reaction products) using the following procedure. On addition of TMB solution, but before drying, the filter-paper was placed between two glass plates fixed together with a clip and then placed in the cell compartment of the spectrophotometer perpendicular to the light beam. The absorbance of this filter-paper was taken as zero. The filter-paper was then dried, sprayed with KIO4 as described above, and the wet specimen was again placed between the glass plates for the measurements. The absorbance at 2 min (A2) was taken as the analytical signal.Calculation of Metrological Characteristics The limit of determination was defined as the concentration of manganese for which the RSD did not exceed 33%; this concentration was taken as the lower limit of the linear range.The limit of detection (cmin) was calculated as 3a/b; for the logarithmic plot y = a + bx, where y is absorbance and x = log(cMn), there are some complications. The following procedure was used: (1) shift of the graph y = a + bx to the origin (x = y = 0), i.e., transformation of the graph to the form Y = A + bX where A = 0 and b is the previous value. This required a recalculation: X = x + (a/b) [for example, X = log(cMn) + 4.82 for the reaction on DETATA filters]; (2) calculation of the regression parameters for the new graph Y = A + bX, including sA; (3) calculation of log(cmin) as 3sA/b.Results and Discussion Reaction Products At least two different coloured products are observed in the course of the studied reaction catalyzed by MnII in solution. A bluish green product (absorption maxima 370 and 650 nm) is formed under conditions where there is a deficiency of the oxidant (less than 1 3 1024 mol l21 KIO4 for a TMB concentration of 2.5 3 1024 mol l21, Fig. 1). According to the literature,12 this species may be a dimeric dication (lmax = 380 Fig. 1 Absorbance of the indicator reaction products as a function of KIO4 concentration. Curves 1 and 2, reaction in solution; curves 3 and 4, reaction on filter-paper with attached DETATA groups (KIO4 applied by sprinkling the solution onto the filter-paper). For the reaction in solution (curve 1): 2.5 3 1024 mol l21 TMB, 1 3 1023 mg l21 MnII; the measurements in solution (A3) were made at 650 nm for cKIO4 < 5 3 1025 mol l21 (bluish green product) or at 460 nm for cKIO4 > 5 3 1025 mol l21 (orange product). For the reaction on DETATA filter-paper (curve 3): 8 3 1028 mol TMB, 8 3 1023 mg l21 MnII; all the measurements of the filter-paper absorbances (A2) were made at 650 nm (bluish green product).For curves 2 and 4, the conditions are the same as for curves 1 and 3, respectively, but with no MnII. 1162 Analyst, October 1997, Vol. 122and 650 nm), probably in a mixture with a meriquinone (lmax = 655 nm).An increase in periodate concentration over 1 3 1023 mol l21 results in an orange product (lmax = 465 nm), which is probably the result of a more extensive oxidation and may be ascribed a quinonediimine structure12 (lmax = 465 nm). Intermediate concentrations of the oxidant provide a brown mixture with absorbance maxima at 380, 650 and 465 nm. As can be observed from Fig. 1, the maximum difference in the rates of MnII-catalyzed and non-catalytic reactions in solution corresponds to the formation of the bluish green product, whereas for higher periodate concentrations (and, consequently, for orange product formation) the difference decreases.For the determination of manganese, the bluish green product (370, 650 nm) was used. Kinetic Curves Kinetic curves for the TMB–MnII–periodate reaction are depicted in Fig. 2. In order to compare the data for the homogeneous and heterogeneous variants of the reaction, the catalyst concentrations should be presented in units which are common to both the solution and the filter-paper.For instance, the mass of manganese per square of the cross-section of the spectrophotometer light beam (mg cm22) would be a common value for both the sorbent and a cell with the solution. If the kinetic curves for the same concentration of manganese (in mg cm22, Fig. 2) are compared, it can be seen that the slope of the ascending portion of the curve is markedly higher for the reaction on filter-paper (with or without DETATA groups) than for that in solution.The optimum conditions for the reaction in solution and on the filter-papers differ, so it can only be noted that the highest attainable initial rate on the filter-papers is higher than that in solution. The absorbance of the filter-paper specimen (which corresponds to the formation of the bluish green product) increases rapidly for the first 0.5 min (counting from the start of the reaction), after which a slow decrease in absorption follows.In solution this decrease leads to the formation of a scarcelycoloured final product which requires a few hours. On the dry filter-paper the bluish green product is more stable (the coloration is not diminished during 24 h). The slow increase in the absorption of filter-paper specimens at 6–10 min is likely to be caused by processes in the cellulose filter-paper itself, e.g., swelling (wet filters without the reagents also exhibit an Fig. 2 Kinetic curves for the KIO4–MnII–TMB reaction at pH 6.8 in solution (1, 4), on filter-paper with attached DETATA groups (2, 5) and on filter-paper (3, 6). Curve 1, 2.5 3 1024 mol l21 TMB, 2.6 3 1025 mol l21 KIO4, 2.6 ng cm22 MnII; 2, 8 3 1028 mol TMB, 4.3 3 1024 mol l21 KIO4, 2.6 ng cm22 MnII. For curve 3, Mn solution (20 ml) was applied by pipetting. For curve 2, Mn solution (20 ml) was pumped through the DETATA filterpaper. The solution of KIO4 was applied onto the filter-paper (curves 2, 3, 5, 6) by sprinkling.For curves 4, 5 and 6, the conditions are the same as for curves 1, 2 and 3, respectively, but without MnII. The measurements were made at 650 nm. Fig. 3 Absorbance of the reaction products as a function of pH in solution (3, 4) (A3) and on filter-paper with attached DETATA groups (1, 2) (A2). Curve 1, 2.5 3 1024 mol l21 TMB, 2.6 3 1025 mol l21 KIO4, 1 ng ml21 MnII; 3, 8 3 1028 mol TMB, 4.3 3 1024 mol l21 KIO4, 8 ng ml21 MnII.For curves 2 and 4, the conditions are the same as for curves 1 and 3, respectively, but without MnII. The measurements were made at 650 nm. To study the effect of pH, manganese solution (0.04 mg l21) of the appropriate pH value was pumped through the DETATA filter-paper and the reaction was carried out as described under Reaction on Filter-paper. Fig. 4 Absorbance of the reaction products as a function of amount of TMB in solution (2, 4) (A3) and on filter-paper with attached DETATA groups (1, 3) (A2).Curve 1, 4.3 3 1024 mol l21 KIO4, 8 ng ml21 MnII; 2, 2.6 31025 mol l21 KIO4, 1 ng ml21 MnII. For curves 3 and 4, the conditions are the same as for curves 1 and 2, respectively, but without MnII. The measurements were made at 650 nm and at pH 6.8. Fig. 5 Relative standard deviations of absorbance of the reaction products as a function of manganese concentration for the reaction carried out in solution (1) and on filter-paper with attached DETATA groups (2).Curve 1, MnII amount denotes cMn/mg cm22 in solution (reaction conditions: 2.5 3 1024 mol l21 TMB, 2.6 3 1025 mol l21 KIO4); curve 2, MnII amounts were calculated as cMn30.02, where cMn is the MnII concentration in the solution that was pumped through the DETATA filter-paper and 0.02 l is the solution volume (conditions for the reaction on the filter-paper: 8 3 1028 mol TMB, 4.3 3 1024 mol l21 KIO4). The measurements were made at 650 nm and at pH 6.5.Analyst, October 1997, Vol. 122 1163increase in absorption). If the latter is not taken into account, the shape of the kinetic curves (Fig. 1) may be tentatively explained13 by reversible sequential–parallel reactions such as TMB"Product I (370, 650 nm)"Product II (colourless) or a set of two reversible reactions TMB"Product I (370, 650 nm) TMB"Product II (colourless) If one of these schemes is true, the steady-state portion of the absorbance–time plot will correspond to an equilibrium of the bluish green product with a colourless product.Effect of pH and Reagent Concentrations It was interesting to study whether the influence of reagent concentrations (TMB, KIO4) and pH on the signal differs for the reaction on filter-paper as opposed to in solution. As shown in Figs. 3–5, considerable differences exist. As can be seen from the pH curve (Fig. 3), the maximum amount of the reaction products is formed at pH 3.1 and 6.8 both in solution and on the DETATA filter-papers, but only on the sorbent is there a catalytic effect of manganese at pH 3.1.The signal on the DETATA filter-papers is higher at pH 3.1 than at pH 6.8 but the precision is poorer, viz., RSDs of 8 and 2%, respectively, are obtained (for 8 31028 mol TMB, 4.3 31024 mol l21 KIO4 and 8 ng ml21 MnII pumped through the DETATA filter-papers). One reason for the low reproducibility at pH 3.1 may be incomplete sorption of MnII at this pH value.Subsequent reactions, both in solution and on the filter-papers, were carried out at pH 6.8. A study of the effect of the periodate concentration on the reaction on the filter-papers showed that only the bluish green product is formed even at high concentrations of the oxidant. In solution (Fig. 1), the orange product was found at KIO4 :TMB ratios !1 : 1, whereas on the filter-papers it was never obtained. This implies some sort of stabilization by the filter-paper of the bluish green oxidation product; a possible explanation is the reducing properties of the filter-paper. Another property of the reaction on the filter-papers is a virtual absence of the effect of KIO4 concentration on the difference in absorbances for catalytic and non-catalytic reactions (Fig. 1). The signal is only slightly affected by periodate concentration in the range 1024–1023 mol l21. This permits the oxidant to be applied onto the filter-papers without strict control of the amount; sprinkling of the filter-papers with KIO4 was used.Sensitivity and Precision of Manganese Determination The absorbance of the TMB–KIO4 reaction products was found to be proportional to the logarithm of the manganese concentration for the reaction both in solution and on DETATA filterpapers. The metrological characteristics of the determination procedure are given in Table 1. In solution, the limit of determination (6 31025 mg l21) is close to that reported for the most sensitive reactions for manganese: viz., oxidation with periodate of N,N-diethylaniline [1 3 1025 (ref. 14) or 1 3 1024 mg l21 (ref. 15)], p-phenetidine [1 3 1024 mg l21 (ref. 16)] and o-dianisidine [2 3 1024 mg l21 (ref. 15)]. Preconcentration of MnII on DETATA filter-papers with the determination directly on the filter-paper makes it feasible not only to decrease the detection limit but also to expand the linear range for MnII from 1.5 orders (in solution) to over 2.5 orders of magnitude (on DETATA filter-papers) (Table 1).As regards the precision of the determination, it was thought that it would be fairly low on the filter-papers because of both the additional preconcentration operation and irregularities in the paper structure (hence, irregular colouring of the filter-papers). However, the RSD values for the reaction on the filter-papers are close to those in solution (Fig. 5), i.e., the precision of the determination remains fairly high. Interferences The criterion for interference was taken as a change of ±5% in the absorbance for 0.001 (reaction in solution) or 5 3 1025 (reaction on DETATA filter-papers) mg l21 of manganese in the analyzed aqueous solution.No interference results from the presence of a 1000-fold molar ratio of various ions (Table 2) or of 1 mg l21 humic and fulvic acids from river water. The selectivity for manganese in the reaction on DETATA filterpapers is higher than that in solution, and both procedures are no Table 1 Equations for the calibration graphs and linear ranges for the determination of MnII by oxidation of TMB with KIO4 in solution and on DETATA filter-papers (preconcentration of MnII from 20 ml of solution for DETATA filter-papers) Linear range/ Reaction a* sa b sb r cmin/mg l21 3 mg l21 In solution 0.455 0.037 0.104 0.024 0.987 1.531025 631025–231023 On DETATA filter-paper 0.385 0.015 0.077 0.006 0.996 2.531025 531026–2.531023 * For the equation y = a + bx, where x = log (cMn) and y = A3 (absorbance measured 3 min after the start of the reaction; blank absorbance was not subtracted).Table 2 Tolerance limits for foreign ions (cion : cMnII) in the determination of MnII by use of the catalytic reaction of TMB with KIO4 in solution and on filter-paper with chelating DETATA groups DETATA filter-paper (1 ng Mn; Solution (0.001 preconcentration Foreign ion mg l21 Mn) from 20 ml) FeII 5 50 ZnII 50 150 Cl2 500 500 FeII* 500 !1000 K, Na, Ca, Mg, Al, 700 700 FeIII, CuII, Br2, SO4 22, acetate !1000 !1000 * In the presence of 0.2 mg l21 KF.Table 3 Effect of FeII and fluoride on the absorbance of the products of the TMB–MnII–KIO4 reaction in solution 3 min after the start of the reaction (A3). [TMB] = 2.5 31024 mol l21; [KIO4] = 2.6 31025 mol l21; l = 650 nm; l = 0.5 cm Concentration of MnII/mg l21 KF/mg l21 FeII/mg l21 0.0002 0.002 0.02 0 0 0.056 0.204 0.223 0 0.28 0.053 0.086 0.097 0.2 0.28 0.058 0.206 0.220 1164 Analyst, October 1997, Vol. 122less (sometimes more) selective than those using the reactions of periodate oxidation of other amines.14–16 The only exception for the TMB–KIO4 reaction is FeII, which significantly interferes by decreasing the reaction rate (tolerance limit is 5 : 1 FeII :MnII for the reaction in solution): in p-phenetidine oxidation,16 a 100-fold amount of FeII was tolerated. At the same time, FeIII does not interfere in large amounts. The effect of FeII can be removed by adding 0.2 mg l21 potassium fluoride (Table 3).The mechanism of fluoride action is not clear, neither is the mechanism of FeII interference itself. Fluoride is not able to complex strongly with FeII ions; however, FeIII may be formed in situ, while fluoride can change the redox potentials of the pairs FeIII–FeII and MnIII–MnII simultaneously in such a manner that iron may no longer participate in the reaction. Rapid Determination of Manganese With Visual Detection One of the potential advantages of sorption–catalytic techniques is the feasibility of rapid determinations of analytes directly on the sorbents with no use of instrumental detection.The TMB– MnII–KIO4 reaction was conducted on DETATA filter-papers after preconcentration of manganese from 20 ml of aqueous solution, using the same procedure as for quantitative measurements (see under Reaction on Filter-papers). Instead of measuring the absorbance of the filter-paper after sprinkling it with periodate, it was dried with a stream of air (which required about 3 min) and the colour was observed visually.The colour remains stable in air for not less than 6 h. Various concentrations in the range from 1 3 1022 to 100 ng of manganese in 20 ml of solution were studied. It was found that confident discrimination of the colour intensities can be made for manganese concentrations which differ by not less than half an order of magnitude (i.e., 3 times).The determination is reliable for 0.1–10 ng of manganese (5 31026–5 31024 mg l21 for a pumped volume of 20 ml), which allows a colour scale to be constructed for the semiquantitative determination of manganese in this range. The whole procedure requires 6–7 min, starting with the pumping of the manganese solution through the DETATA filter-paper. Analysis of Tap and River Water For analysis, an aliquot of the sample (1.0 ml of tap water or 0.10 ml of river water preserved by adding sulfuric acid to pH 1.85 immediately after sampling) with 0.2 ml of KF (20 mg l21) added was diluted to 20 ml with borate buffer (pH 6.8).The analyses were performed as described under Reaction in solution and Reaction on Filter-paper. The results agreed with those obtained by spectrophotometry17 and/or flame atomic absorption spectrometry (Table 4). The high values obtained with the catalytic method in solution and by spectrophotometry might be due to the lower selectivity of these techniques.When manganese is preconcentrated on the DETATA sorbent, it is separated from interfering species and the results obtained agree with those obtained by another selective technique (atomic absorption). The authors thank Dr. G.I. Tsysin for providing the DETATA filter-papers and for fruitful discussions, Dr. N.M. Sorokina for AAS measurements, Dr. T.V. Polenova for the humic acid preparation, and the Russian Foundation for Basic Research for financial support (grant No. 96-03-08854). References 1 Dolmanova, I. F., and Peshkova, V. M., Vestn. Mosk. Gosud. Univ., Ser. 2: Khim., 1977, 18, 599. 2 Kolotyrkina, I. Ya., Shpigun, L. K., Zolotov, Yu. A., and Tsysin, G. I., Analyst, 1991, 116, 707. 3 Tikhonova, L. P., Bakay, E. A., Prokhorenko, E. P., Tarkovskaya, I. A., and Svarkovskaya, I. P., presented at the 5th International Symposium on Kinetics in Analytical Chemistry, September 25–28, 1995, Moscow, Russia; Abstracts of Papers, Nauka, Moscow, 1995, L24. 4 Varshal, G. M., Velyukhanova, T. K., Pavlutskaya, V. I., Starshinova, N. P., Formanovsky, A. A., Seregina, I. F., Shilnikov, A. M., Tsysin, G. I., and Zolotov, Yu. A., Int. J. Environ. Anal. Chem., 1994, 57, 107. 5 Naylor, F. J., and Saunders, B. C., J. Chem. Soc., 1950, 3519. 6 Hester, R. E., and Williams, K. P. J., J. Chem. Soc., Faraday Trans. II, 1981, 77, 541. 7 Makemoto, K., and Maysunaka, M., Bull. Chem. Soc. Jpn., 1968, 41, 764. 8 Zolotov, Yu.A., Zh. Anal. Khim., 1994, 49, 149. 9 Pöribil, R., Analytical Application of EDTA and Related Compounds, Mir, Moscow, 1975, p. 200 (in Russian). 10 Lurye, Yu., Handbook in Analytical Chemistry, Khimiya, Moscow, 1989 (in Russian). 11 Tsysin, G. I., Mikhura, I. V., Formanovsky, A. A., and Zolotov, Yu. A., Mikrochim. Acta, 1991, III, 53. 12 Saunders, B. C., and Watson, G. M. R., Biochem. J., 1950, 46, 629. 13 Denisov, E. T., Kinetics of Homogeneous Chemical Reactions, Vysshaya Shkola, Moscow, 1988, p. 57. 14 Nikolesis, D. P., Anal. Chem., 1978, 50, 205. 15 Dolmanova, I. F., and Yatsimirskaya, N. T., Zh. Anal. Khim., 1971, 26, 1540. 16 Gragorovich, F. G., Fresenius’ Z. Anal. Chem., 1974, 271, 5, 354. 17 Alimarin, I. P., Practical Recommendations on Physico-Chemical Methods for Analysis, Moskovskii Gosudarstvennyi Universitet, Moscow, 1987, p. 58 (in Russian). Paper 7/02595E Received April 16, 1997 Accepted June 30, 1997 Table 4 Concentrations of manganese in water (mg l21) found by using the TMB–KIO4 reaction and reference techniques.The RSDs were obtained from five parallel runs Catalytic method Reaction on Atomic DETATA absorption Sample Reaction in solution* filter-paper† spectrometry Spectrophotometry‡ Tap water (1.1 ± 0.3)31023 (0.7 ± 0.1)31023 (0.6 ± 0.1)31023 — River water (1.2 ± 0.1)31021 (0.9 ± 0.2)31021 (0.76 ± 0.03)31021 (1.4 ± 0.1)31021 * [TMB] = 2.5 3 1024 mol21; [KIO4] = 2.6 3 1025 mol l21; sample volume = 0.1–1 ml; [KF] = 0.2 mg l21; pH, 6.8; l = 650 nm, l = 0.5 cm.† The analyzed solution with added buffer (pH 6.8) and 0.2 mg l21 KF was pumped through the DETATA filter-paper and the reaction was carried out as described under Reaction on Filter-paper. ‡ Determined with formaldoxime.17 Analyst, October 1997, Vol. 122 1165 H2N Me Me Me NH2 Me N N OH N O OH O OH O OH O ( a) ( b) Sorption–Catalytic Determination of Manganese Directly on a Paper-based Chelating Sorbent M. K. Beklemishev, T.A. Stoyan and I. F. Dolmanova* Analytical Chemistry Division, Department of Chemistry, Moscow State University, Moscow 119899 GSP, Russia. E-mail: mkb@analyt.chem.msu.su A hybrid technique based on a catalytic reaction carried out on the surface of a paper-based sorbent is proposed. It is shown that MnII exhibits its catalytic action in the oxidation of 3,3A,5,5A-tetramethylbenzidine with periodate in aqueous solution as well as on filter-paper with or without attached diethylenetriaminetetraacetate (DETATA) groups.Optimum conditions differ for the reaction in solution and on filter-paper. For equal catalyst concentrations, higher initial reaction rates are attainable on the filter-papers. Preconcentration of manganese on the DETATA sorbent combined with the subsequent catalytic reaction improves the selectivity, reduces the limit of determination down to 5 3 1026 mg l21 (as compared with 6 3 1025 mg l21 in solution) and expands the linear range by an order of magnitude (to 5 3 1026–2.5 3 1023 mg l21).The precision of manganese determination on the sorbent is high (the RSDs are @5% for !6 3 1024 mg Mn). Samples of tap and river water were analyzed by use of the proposed sorption–catalytic technique. A rapid procedure for the determination of manganese with visual detection was also developed. Keywords: Catalytic kinetic method of analysis; chelating sorbent; diethylenetriaminetetraacetate; preconcentration of manganese Catalytic procedures are appreciated by analytical chemists for their sensitivity and simplicity in realization.Catalytic indicator reactions can be applied to the determination of a large number of compounds including the catalyst, inhibitors, activators, and compounds which convert the catalyst into an active state (oxidizers, ligands1), and catalytically inactive metal ions through the use of the competitive complexation principle. Catalytic methods become especially powerful when combined with analyte separation/preconcentration, which allows the selectivity to be increased and the detection limits to be lowered.One of the approaches involves adsorption of metal ions on chelating sorbents with subsequent desorption and catalytic determination.2 However, this type of procedure, while being convenient in flow systems,2 is tedious in other cases owing to the need for desorption of the analyte. Here, we propose to eliminate the desorption stage by conducting the catalytic indicator reaction directly on the sorbent used for preconcentration purposes. Little information has been published to date on the foregoing principle.3 This novel area requires studies of the peculiarities of catalytic indicator reactions as they occur on the surfaces of sorbents, including kinetics, concentration effects, and the metrological characteristics of the procedures. The main purpose of this study was the practical realization of the aforementioned sorption–catalytic principle.For the investigation, the preconcentration of manganese on diethylenetriaminetetraacetate (DETATA) paper-based chelating sorbent was chosen as it combines effectiveness of metal sorption and simplicity of operation.2,4 As an indicator reaction, the oxidation of 3,3A,5,5A-tetramethylbenzidine (TMB) by potassium periodate (KIO4) catalyzed by MnII was selected. Various amines other than TMB were previously studied in this reaction;5–7 however, TMB is more stable in air, less carcinogenic and its oxidation provides higher absorbing products.In the present work, the catalytic effect of MnII on the reaction between TMB and KIO4 in solution was studied. The aim was to employ this system to ascertain the analytical possibilities of the sorption–catalytic approach for the determination of manganese with both instrumental and visual detection. A potential benefit of the sorption–catalytic approach may be realized in ‘rapid-test’ techniques, such as spot-tests, field tests and similar procedures,8 in which a visible signal appears on the surface of paper strips or other supports.The development of rapid tests for manganese may be a subject for further work originating from this study. Experimental Reagents, Solutions and Apparatus TMB ‘for analysis’ was obtained from Riedel-de H�aen, Hannover, Germany; boric, hydrochloric and sulfuric acids (‘special purity’) from Reakhim (Moscow, Russia) were used.All other reagents were of analytical-reagent grade. Humic and fulvic acids were obtained from Moscow river water by adsorption on poly(styrene–divinylbenzene)sorbent (XAD) with subsequent desalination. All aqueous solutions were prepared using distilled water. Ethanolic solutions of TMB (0.025 mol l21) were prepared every 10 d; less concentrated solutions were obtained by dilution with ethanol when necessary. A stock solution of periodate in water (1 g l21 KIO4) was diluted to an appropriate concentration (in most cases 0.1 g l21) every 4–5 d.A stock solution of anhydrous MnSO4 (1 g l21 Mn) was standardized by titration.9 Solutions with lower contents of the metal were prepared every 3 d (10 and 1 mg l21) or daily (0.01 mg l21) by dilution of the stock solution. All the manganese solutions were acidified with H2SO4 to pH 1.8–1.9; it was found that such acidification was necessary in order to obtain reproducible results.Hydrochloric acid was used to provide pH values ranging from 2 to 4; it was purified by isothermal distillation of the concentrated acid (reagent grade). Borate buffer of pH 6.8 was prepared by dropwise addition of Scheme 1 (a) TMB; (b) DETATA groups attached to filter paper. Analyst, October 1997, Vol. 122 (1161–1165) 11610.2 mol l21 KOH to a solution of 3.3 g of boric acid in 1000 ml of water until the desired pH value was reached. Buffers for the pH range 4–6 (0.1 mol l21 in sodium) were prepared from acetic acid and sodium acetate, and buffers for pH values of 7–11 (0.1 mol l21 in borate) were prepared from boric acid, Na2- B4O7·10H2O and KOH, by using standard procedures.10 The other buffer solutions with pH > 4 were purified by pumping through two consecutive DETATA filters placed in a special polyethylene holder (‘Biospectr’, St.Petersburg, Russia) at a rate of 8–10 ml min21, which removed transition metal impurities. The chelating sorbent was synthesized as described previously by chemical attachment of aminocarboxylic (diethylenetriaminetetraacetate, DETATA) groups to cellulose filterpaper. 11 The filter-paper was 2.5 cm in diameter.Water was distilled in a commercial apparatus (D3-4-2M, Ioshkar-Ola, Russia). A KFK-3 spectrophotometer (ZOMZ, Zagorsk, Russia) was used for measurements of absorbance. To spray the filter-papers with periodate solution, a hand-operated sprinkler for thin-layer chromatograms was used.Reaction in Solution Preliminary studies of various mixing orders of the components of the indicator reaction showed that maximum absorbance was achieved when MnII and borate buffer (pH 6.8) were mixed first, followed by addition of TMB and finally of KIO4. Based on this, the following procedure was used. In a 15 ml glass test-tube were placed the reagents in the following sequence (the amounts are given for the final procedure for manganese determination): 3.9 ml of the buffer (pH 6.8), 0.5 ml of the solution, containing MnII at a pH of 1.8–1.9, 0.1 ml of a 2.5 3 1024 mol l21 ethanolic solution of TMB and 0.5 ml of 4.3 3 1022 mol l21 periodate.The addition of periodate was taken as the time of the start of the reaction. After agitation, the reaction mixture was transferred into a 0.5 cm cell and the absorbance of the oxidation product was measured at 650 nm against water. The absorbance value at 3 min (A3) was taken as the analytical signal.The experiments were performed at ambient temperature, which was kept at 24 ± 1 °C. Reaction on Filter-paper In preliminary experiments the optimum mixing order of the reactants was sought with the use of visual detection. Coloration of the filter-paper was most intense when TMB and the buffered manganese solution were mixed first and periodate was added at the last step. The role of the humidity of the filter-papers was also studied (the filter-papers were stored for up to 3 d in a desiccator with controlled 75–95% humidity); no effect of humidity on the reaction rate was observed.The final procedure for the reaction on filter-paper (with or without attached DETATA groups) is given below. MnII was first applied onto a filter-paper either (a) in the form of a 0.02 ml aliquot of a solution (0.4 mg l21 Mn) added to the centre of the filter-paper by an Eppendorf pipette, or (b) by pumping a buffered manganese solution of pH 6.8 (5 3 1025–0.05 mg l21 MnII, 0.08 mol l21 in borate) through the filter-paper by use of a peristaltic pump at a rate of 8 ml min21.The filter-paper was then dried with a gentle flow of pressurized air until no traces of moisture could be detected, after which TMB was pipetted onto the filter-paper (0.02 ml of a 4 3 1024 mol l21 solution) followed by an identical drying procedure. The oxidant was added by sprinkling the filter-paper with a 4.3 3 1024 mol l21 solution of KIO4, which was taken to be the beginning of the reaction.The coloured reaction zone was about 20 mm in diameter. Sprinkling rather than pipetting of periodate was necessary in order to obtain uniform coloration of the filterpapers; the sprinkling technique was applicable owing to a weak dependence of the analytical signal on the amount of KIO4 on the filter-paper, as shown below. The absorbance of a filter-paper was measured in its wet state against the same filter-paper without added periodate (i.e., not coloured with the reaction products) using the following procedure.On addition of TMB solution, but before drying, the filter-paper was placed between two glass plates fixed together with a clip and then placed in the cell compartment of the spectrophotometer perpendicular to the light beam. The absorbance of this filter-paper was taken as zero. The filter-paper was then dried, sprayed with KIO4 as described above, and the wet specimen was again placed between the glass plates for the measurements.The absorbance at 2 min (A2) was taken as the analytical signal. Calculation of Metrological Characteristics The limit of determination was defined as the concentration of manganese for which the RSD did not exceed 33%; this concentration was taken as the lower limit of the linear range. The limit of detection (cmin) was calculated as 3a/b; for the logarithmic plot y = a + bx, where y is absorbance and x = log(cMn), there are some complications.The following procedure was used: (1) shift of the graph y = a + bx to the origin (x = y = 0), i.e., transformation of the graph to the form Y = A + bX where A = 0 and b is the previous value. This required a recalculation: X = x + (a/b) [for example, X = log(cMn) + 4.82 for the reaction on DETATA filters]; (2) calculation of the regression parameters for the new graph Y = A + bX, including sA; (3) calculation of log(cmin) as 3sA/b. Results and Discussion Reaction Products At least two differe coloured products are observed in the course of the studied reaction catalyzed by MnII in solution.A bluish green product (absorption maxima 370 and 650 nm) is formed under conditions where there is a deficiency of the oxidant (less than 1 3 1024 mol l21 KIO4 for a TMB concentration of 2.5 3 1024 mol l21, Fig. 1). According to the literature,12 this species may be a dimeric dication (lmax = 380 Fig. 1 Absorbance of the indicator reaction products as a function of KIO4 concentration.Curves 1 and 2, reaction in solution; curves 3 and 4, reaction on filter-paper with attached DETATA groups (KIO4 applied by sprinkling the solution onto the filter-paper). For the reaction in solution (curve 1): 2.5 3 1024 mol l21 TMB, 1 3 1023 mg l21 MnII; the measurements in solution (A3) were made at 650 nm for cKIO4 < 5 3 1025 mol l21 (bluish green product) or at 460 nm for cKIO4 > 5 3 1025 mol l21 (orange product). For the reaction on DETATA filter-paper (curve 3): 8 3 1028 mol TMB, 8 3 1023 mg l21 MnII; all the measurements of the filter-paper absorbances (A2) were made at 650 nm (bluish green product).For curves 2 and 4, the conditions are the same as for curves 1 and 3, respectively, but with no MnII. 1162 Analyst, October 1997, Vol. 122and 650 nm), probably in a mixture with a meriquinone (lmax = 655 nm). An increase in periodate concentration over 1 3 1023 mol l21 results in an orange product (lmax = 465 nm), which is probably the result of a more extensive oxidation and may be ascribed a quinonediimine structure12 (lmax = 465 nm).Intermediate concentrations of the oxidant provide a brown mixture with absorbance maxima at 380, 650 and 465 nm. As can be observed from Fig. 1, the maximum difference in the rates of MnII-catalyzed and non-catalytic reactions in solution corresponds to the formation of the bluish green product, whereas for higher periodate concentrations (and, consequently, for orange product formation) the difference decreases.For the determination of manganese, the bluish green product (370, 650 nm) was used. Kinetic Curves Kinetic curves for the TMB–MnII–periodate reaction are depicted in Fig. 2. In order to compare the data for the homogeneous and heterogeneous variants of the reaction, the catalyst concentrations should be presented in units which are common to both the solution and the filter-paper. For instance, the mass of manganese per square of the cross-section of the spectrophotometer light beam (mg cm22) would be a common value for both the sorbent and a cell with the solution.If the kinetic curves for the same concentration of manganese (in mg cm22, Fig. 2) are compared, it can be seen that the slope of the ascending portion of the curve is markedly higher for the reaction on filter-paper (with or without DETATA groups) than for that in solution. The optimum conditions for the reaction in solution and on the filter-papers differ, so it can only be noted that the highest attainable initial rate on the filter-papers is higher than that in solution.The absorbance of the filter-paper specimen (which corresponds to the formation of the bluish green product) increases rapidly for the first 0.5 min (counting from the start of the reaction), after which a slow decrease in absorption follows. In solution this decrease leads to the formation of a scarcelycoloured final product which requires a few hours.On the dry filter-paper the bluish green product is more stable (the coloration is not diminished during 24 h). The slow increase in the absorption of filter-paper specimens at 6–10 min is likely to be caused by processes in the cellulose filter-paper itself, e.g., swelling (wet filters without the reagents also exhibit an Fig. 2 Kinetic curves for the KIO4–MnII–TMB reaction at pH 6.8 in solution (1, 4), on filter-paper with attached DETATA groups (2, 5) and on filter-paper (3, 6).Curve 1, 2.5 3 1024 mol l21 TMB, 2.6 3 1025 mol l21 KIO4, 2.6 ng cm22 MnII; 2, 8 3 1028 mol TMB, 4.3 3 1024 mol l21 KIO4, 2.6 ng cm22 MnII. For curve 3, Mn solution (20 ml) was applied by pipetting. For curve 2, Mn solution (20 ml) was pumped through the DETATA filterpaper. The solution of KIO4 was applied onto the filter-paper (curves 2, 3, 5, 6) by sprinkling. For curves 4, 5 and 6, the conditions are the same as for curves 1, 2 and 3, respectively, but without MnII.The measurements were made at 650 nm. Fig. 3 Absorbance of the reaction products as a function of pH in solution (3, 4) (A3) and on filter-paper with attached DETATA groups (1, 2) (A2). Curve 1, 2.5 3 1024 mol l21 TMB, 2.6 3 1025 mol l21 KIO4, 1 ng ml21 MnII; 3, 8 3 1028 mol TMB, 4.3 3 1024 mol l21 KIO4, 8 ng ml21 MnII. For curves 2 and 4, the conditions are the same as for curves 1 and 3, respectively, but without MnII.The measurements were made at 650 nm. To study the effect of pH, manganese solution (0.04 mg l21) of the appropriate pH value was pumped through the DETATA filter-paper and the reaction was carried out as described under Reaction on Filter-paper. Fig. 4 Absorbance of the reaction products as a function of amount of TMB in solution (2, 4) (A3) and on filter-paper with attached DETATA groups (1, 3) (A2). Curve 1, 4.3 3 1024 mol l21 KIO4, 8 ng ml21 MnII; 2, 2.6 31025 mol l21 KIO4, 1 ng ml21 MnII.For curves 3 and 4, the conditions are the same as for curves 1 and 2, respectively, but without MnII. The measurements were made at 650 nm and at pH 6.8. Fig. 5 Relative standard deviations of absorbance of the reaction products as a function of manganese concentration for the reaction carried out in solution (1) and on filter-paper with attached DETATA groups (2). Curve 1, MnII amount denotes cMn/mg cm22 in solution (reaction conditions: 2.5 3 1024 mol l21 TMB, 2.6 3 1025 mol l21 KIO4); curve 2, MnII amounts were calculated as cMn30.02, where cMn is the MnII concentration in the solution that was pumped through the DETATA filter-paper and 0.02 l is the solution volume (conditions for the reaction on the filter-paper: 8 3 1028 mol TMB, 4.3 3 1024 mol l21 KIO4).The measurements were made at 650 nm and at pH 6.5. Analyst, October 1997, Vol. 122 1163increase in absorption). If the latter is not taken into account, the shape of the kinetic curves (Fig. 1) may be tentatively explained13 by reversible sequential–parallel reactions such as TMB"Product I (370, 650 nm)"Product II (colourless) or a set of two reversible reactions TMB"Product I (370, 650 nm) TMB"Product II (colourless) If one of these schemes is true, the steady-state portion of the absorbance–time plot will correspond to an equilibrium of the bluish green product with a colourless product. Effect of pH and Reagent Concentrations It was interesting to study whether the influence of reagent concentrations (TMB, KIO4) and pH on the signal differs for the reaction on filter-paper as opposed to in solution.As shown in Figs. 3–5, considerable differences exist. As can be seen from the pH curve (Fig. 3), the maximum amount of the reaction products is formed at pH 3.1 and 6.8 both in solution and on the DETATA filter-papers, but only on the sorbent is there a catalytic effect of manganese at pH 3.1. The signal on the DETATA filter-papers is higher at pH 3.1 than at pH 6.8 but the precision is poorer, viz., RSDs of 8 and 2%, respectively, are obtained (for 8 31028 mol TMB, 4.3 31024 mol l21 KIO4 and 8 ng ml21 MnII pumped through the DETATA filter-papers). One reason for the low reproducibility at pH 3.1 may be incomplete sorption of MnII at this pH value.Subsequent reactions, both in solution and on the filter-papers, were carried out at pH 6.8. A study of the effect of the periodate concentration on the reaction on the filter-papers showed that only the bluish green product is formed even at high concentrations of the oxidant.In solution (Fig. 1), the orange product was found at KIO4 :TMB ratios !1 : 1, whereas on the filter-papers it was never obtained. This implies some sort of stabilization by the filter-paper of the bluish green oxidation product; a possible explanation is the reducing properties of the filter-paper. Another property of the reaction on the filter-papers is a virtual absence of the effect of KIO4 concentration on the difference in absorbances for catalytic and non-catalytic reactions (Fig. 1). The signal is only slightly affected by periodate concentration in the range 1024–1023 mol l21. This permits the oxidant to be applied onto the filter-papers without strict control of the amount; sprinkling of the filter-papers with KIO4 was used. Sensitivity and Precision of Manganese Determination The absorbance of the TMB–KIO4 reaction products was found to be proportional to the logarithm of the manganese concentration for the reaction both in solution and on DETATA filterpapers.The metrological characteristics of the determination procedure are given in Table 1. In solution, the limit of determination (6 31025 mg l21) is close to that reported for the most sensitive reactions for manganese: viz., oxidation with periodate of N,N-diethylaniline [1 3 1025 (ref. 14) or 1 3 1024 mg l21 (ref. 15)], p-phenetidine [1 3 1024 mg l21 (ref. 16)] and o-dianisidine [2 3 1024 mg l21 (ref. 15)]. Preconcentration of MnII on DETATA filter-papers with the determination directly on the filter-paper makes it feasible not only to decrease the detection limit but also to expand the linear range for MnII from 1.5 orders (in solution) to over 2.5 orders of magnitude (on DETATA filter-papers) (Table 1). As regards the precision of the determination, it was thought that it would be fairly low on the filter-papers because of both the additional preconcentration operation and irregularities in the paper structure (hence, irregular colouring of the filter-papers). However, the RSD values for the reaction on the filter-papers are close to those in solution (Fig. 5), i.e., the precision of the determination remains fairly high. Interferences The criterion for interference was taken as a change of ±5% in the absorbance for 0.001 (reaction in solution) or 5 3 1025 (reaction on DETATA filter-papers) mg l21 of manganese in the analyzed aqueous solution.No interference results from the presence of a 1000-fold molar ratio of various ions (Table 2) or of 1 mg l21 humic and fulvic acids from river water. The selectivity for manganese in the reaction on DETATA filterpapers is higher than that in solution, and both procedures are no Table 1 Equations for the calibration graphs and linear ranges for the determination of MnII by oxidation of TMB with KIO4 in solution and on DETATA filter-papers (preconcentration of MnII from 20 ml of solution for DETATA filter-papers) Linear range/ Reaction a* sa b sb r cmin/mg l21 3 mg l21 In solution 0.455 0.037 0.104 0.024 0.987 1.531025 631025–231023 On DETATA filter-paper 0.385 0.015 0.077 0.006 0.996 2.531025 531026–2.531023 * For the equation y = a + bx, where x = log (cMn) and y = A3 (absorbance measured 3 min after the start of the reaction; blank absorbance was not subtracted).Table 2 Tolerance limits for foreign ions (cion : cMnII) in the determination of MnII by use of the catalytic reaction of TMB with KIO4 in solution and on filter-paper with chelating DETATA groups DETATA filter-paper (1 ng Mn; Solution (0.001 preconcentration Foreign ion mg l21 Mn) from 20 ml) FeII 5 50 ZnII 50 150 Cl2 500 500 FeII* 500 !1000 K, Na, Ca, Mg, Al, 700 700 FeIII, CuII, Br2, SO4 22, acetate !1000 !1000 * In the presence of 0.2 mg l21 KF.Table 3 Effect of FeII and fluoride on the absorbance of the products of the TMB–MnII–KIO4 reaction in solution 3 min after the start of the reaction (A3).[TMB] = 2.5 31024 mol l21; [KIO4] = 2.6 31025 mol l21; l = 650 nm; l = 0.5 cm Concentration of MnII/mg l21 KF/mg l21 FeII/mg l21 0.0002 0.002 0.02 0 0 0.056 0.204 0.223 0 0.28 0.053 0.086 0.097 0.2 0.28 0.058 0.206 0.220 1164 Analyst, October 1997, Vol. 122less (sometimes more) selective than those using the reactions of periodate oxidation of other amines.14–16 The only exception for the TMB–KIO4 reaction is FeII, which significantly interferes by decreasing the reaction rate (tolerance limit is 5 : 1 FeII :MnII for the reaction in solution): in p-phenetidine oxidation,16 a 100-fold amount of FeII was tolerated.At the same time, FeIII does not interfere in large amounts. The effect of FeII can be removed by adding 0.2 mg l21 potassium fluoride (Table 3). The mechanism of fluoride action is not clear, neither is the mechanism of FeII interference itself.Fluoride is not able to complex strongly with FeII ions; however, FeIII may be formed in situ, while fluoride can change the redox potentials of the pairs FeIII–FeII and MnIII–MnII simultaneously in such a manner that iron may no longer participate in the reaction. Rapid Determination of Manganese With Visual Detection One of the potential advantages of sorption–catalytic techniques is the feasibility of rapid determinations of analytes directly on the sorbents with no use of instrumental detection.The TMB– MnII–KIO4 reaction was conducted on DETATA filter-papers after preconcentration of manganese from 20 ml of aqueous solution, using the same procedure as for quantitative measurements (see under Reaction on Filter-papers). Instead of measuring the absorbance of the filter-paper after sprinkling it with periodate, it was dried with a stream of air (which required about 3 min) and the colour was observed visually.The colour remains stable in air for not less than 6 h. Various concentrations in the range from 1 3 1022 to 100 ng of manganese in 20 ml of solution were studied. It was found that confident discrimination of the colour intensities can be made for manganese concentrations which differ by not less than half an order of magnitude (i.e., 3 times). The determination is reliable for 0.1–10 ng of manganese (5 31026–5 31024 mg l21 for a pumped volume of 20 ml), which allows a colour scale to be constructed for the semiquantitative determination of manganese in this range.The whole procedure requires 6–7 min, starting with the pumping of the manganese solution through the DETATA filter-paper. Analysis of Tap and River Water For analysis, an aliquot of the sample (1.0 ml of tap water or 0.10 ml of river water preserved by adding sulfuric acid to pH 1.85 immediately after sampling) with 0.2 ml of KF (20 mg l21) added was diluted to 20 ml with borate buffer (pH 6.8).The analyses were performed as described under Reaction in solution and Reaction on Filter-paper. The results agreed with those obtained by spectrophotometry17 and/or flame atomic absorption spectrometry (Table 4). The high values obtained with the catalytic method in solution and by spectrophotometry might be due to the lower selectivity of these techniques. When manganese is preconcentrated on the DETATA sorbent, it is separated from interfering species and the results obtained agree with those obtained by another selective technique (atomic absorption). The authors thank Dr.G.I. Tsysin for providing the DETATA filter-papers and for fruitful discussions, Dr. N.M. Sorokina for AAS measurements, Dr. T.V. Polenova for the humic acid preparation, and the Russian Foundation for Basic Research for financial support (grant No. 96-03-08854). References 1 Dolmanova, I. F., and Peshkova, V. M., Vestn. Mosk. Gosud. Univ., Ser. 2: Khim., 1977, 18, 599. 2 Kolotyrkina, I. Ya., Shpigun, L. K., Zolotov, Yu. A., and Tsysin, G. I., Analyst, 1991, 116, 707. 3 Tikhonova, L. P., Bakay, E. A., Prokhorenko, E. P., Tarkovskaya, I. A., and Svarkovskaya, I. P., presented at the 5th International Symposium on Kinetics in Analytical Chemistry, September 25–28, 1995, Moscow, Russia; Abstracts of Papers, Nauka, Moscow, 1995, L24. 4 Varshal, G. M., Velyukhanova, T. K., Pavlutskaya, V. I., Starshinova, N. P., Formanovsky, A. A., Seregina, I. F., Shilnikov, A. M., Tsysin, G. I., and Zolotov, Yu. A., Int. J. Environ. Anal. Chem., 1994, 57, 107. 5 Naylor, F. J., and Saunders, B. C., J. Chem. Soc., 1950, 3519. 6 Hester, R. E., and Williams, K. P. J., J. Chem. Soc., Faraday Trans. II, 1981, 77, 541. 7 Makemoto, K., and Maysunaka, M., Bull. Chem. Soc. Jpn., 1968, 41, 764. 8 Zolotov, Yu. A., Zh. Anal. Khim., 1994, 49, 149. 9 Pöribil, R., Analytical Application of EDTA and Related Compounds, Mir, Moscow, 1975, p. 200 (in Russian). 10 Lurye, Yu., Handbook in Analytical Chemistry, Khimiya, Moscow, 1989 (in Russian). 11 Tsysin, G. I., Mikhura, I. V., Formanovsky, A. A., and Zolotov, Yu. A., Mikrochim. Acta, 1991, III, 53. 12 Saunders, B. C., and Watson, G. M. R., Biochem. J., 1950, 46, 629. 13 Denisov, E. T., Kinetics of Homogeneous Chemical Reactions, Vysshaya Shkola, Moscow, 1988, p. 57. 14 Nikolesis, D. P., Anal. Chem., 1978, 50, 205. 15 Dolmanova, I. F., and Yatsimirskaya, N. T., Zh. Anal. Khim., 1971, 26, 1540. 16 Gragorovich, F. G., Fresenius’ Z. Anal. Chem., 1974, 271, 5, 354. 17 Alimarin, I. P., Practical Recommendations on Physico-Chemical Methods for Analysis, Moskovskii Gosudarstvennyi Universitet, Moscow, 1987, p. 58 (in Russian). Paper 7/02595E Received April 16, 1997 Accepted June 30, 1997 Table 4 Concentrations of manganese in water (mg l21) found by using the TMB–KIO4 reaction and reference techniques. The RSDs were obtained from five parallel runs Catalytic method Reaction on Atomic DETATA absorption Sample Reaction in solution* filter-paper† spectrometry Spectrophotometry‡ Tap water (1.1 ± 0.3)31023 (0.7 ± 0.1)31023 (0.6 ± 0.1)31023 — River water (1.2 ± 0.1)31021 (0.9 ± 0.2)31021 (0.76 ± 0.03)31021 (1.4 ± 0.1)31021 * [TMB] = 2.5 3 1024 mol21; [KIO4] = 2.6 3 1025 mol l21; sample volume = 0.1–1 ml; [KF] = 0.2 mg l21; pH, 6.8; l = 650 nm, l = 0.5 cm. † The analyzed solution with added buffer (pH 6.8) and 0.2 mg l21 KF was pumped through the DETATA filter-paper and the reaction was carried out as described under Reaction on Filter-paper. ‡ Determined with formaldoxime.17 Analyst, October 1997, Vol. 122 1165