摘要:

ANALYST, JANUARY 1991, VOL. 116 53 Determination of Manganese(i1) by a Photoactivated, Catalytic Method Juergen Mattusch and Gerhard Werner Department of Chemistry, Karl-Marx-University Leipzig, Talstr. 35, Leipzig 7010, Germany Helmut Mueller Depa rtm en t o f Ch em istr y, Tech n ica I University "Carl Sc h o rlem me r, " L e u n a -Me rse b urg, Otto- Nusch ke-S tr., Merseburg 4200, Germany A method is described for the determination of Mn" based on the photoactivated oxidation of sulphite using the specific photosensitizer Rose Bengal. The reactivity of Mn" is approximately ten times higher for the photosensitized, acetylacetone-activated reaction than for the system that is only chemically activated. The determination of Mn" is not interfered with by a large excess of Col, Cull, Nill, Crlll, Cdll or Fell'.The detection limit of the proposed method is 0.3 ppb with a relative standard deviation of about 5.5%. A proposal for a possible reaction mechanism is made on the basis of spectrophotometric measurements. The catalytic method described was used for the determination of Mn" in drinking water. Keywords : Kinetic catalytic method; photosensitization; chemical activation; trace analysis; manganese (11) determination Kinetic or reaction-rate methods of chemical analysis utilize the dynamic properties of reaction systems. Catalytic determi- nations are the most widely used of the kinetic methods. In the last 10 years many such methods have been developed.'-5 Using a catalytic reaction one can determine extremely small concentrations of the catalyst through an increase in the reaction rate because a catalytic species may participate in a large number of cycles of the catalytic reaction. The best way to increase the rate of a catalysed reaction (and thereby the sensitivity while decreasing the limit of detection) is through the use of activators.j.5 Activation is understood from a catalytic point of view as the increase in the rate of a catalysed reaction resulting from the action of a chemical species (the activator) that takes part in a step for which the activation energy is lower than that involving the catalyst only (classical chemical activation). The effect of irradiation on catalysed reactions has been studied and the influence of light on their rate and selectivity (photoactivation) has been shown.6 Such photoactivation applied to the photosensitization of catalysed reactions can be used as an alternative to classical activation.Photoactivation can be subdivided into: direct photoactivation of a reactant; indirect photosensitization; and catalyst photoactivation. In this work, the analytical utility of the activation of the catalyst Mn" by photosensitization has been examined. Veprek-SiSka and co-workers778 have demon- strated that the auto-oxidation of sulphite is catalysed by metal ions. Both the thermal and the photochemical or photo- initiated oxidation of sulphite proceed via the same mechan- ism.7-10 The reactive intermediates of the thermal reaction are probably ternary complexes of the metal ion, e.g., Cu", with sulphite and oxygen.Also, in the photo-initiated auto-oxida- tion of sulphite the catalytic action of Fe"1 is connected with the existence of sulphitoiron(ii1) c0mplexes.6~11-13 In the present paper the photosensitized auto-oxidation of sulphite, catalysed by Mn" ions, was examined in detail. Experimental Materials All chemicals and reagents were of analytical-reagent grade. Acetylacetone (acac) (Merck) was distilled before use. The other chemical activators, i.e., 1,lO-phenanthroline (phen), ethylenediamine (en), nitrilotriacetic acid (nta), 2,2'-bipyr- idine (bpy), histidine (his) and oxalate (ox) were used as received. A 0.3 mol dm-3 stock solution of Na2S03 was prepared fresh daily by dissolving 3.78 g of Na2S03 in 100 ml of doubly distilled water. The 0.1 mol dm-3 buffer solution was obtained by dissolution of sodium tetraborate deca- hydrate (38.14 g 1-1).The Mn" stock solution was prepared by dissolution of Mn(N03)2. The exact amount of Mn" was determined by complexome tric ti tra tion. l4 Apparatus Continuous photolysis was performed with a 500 W super- pressure mercury lamp (HBO 500). Excitation was at 546 nm with a 10 nm bandpass. The reaction solution was contained within a 50 ml glass vessel and was stirred magnetically. The continuous oxygen consumption during the reaction time was measured with an oxygen-selective electrode of the Clark-type (Meinsberg, Germany). The reaction rate is expressed as oxygen consumption during a defined time (v = -A[O,]/At) in mg O2 1-1 s-1 in all instances. Detailed information about the absorbance of the photosensitizer and the reaction intermediates was obtained with a Specord spectra+ photometer (Jena, Germany).Procedure A 2 ml volume of 0.1 mol dm-3 buffer solution, 6 ml of a 0.01 rnol dm-3 solution of acac, 0.02 ml of 4 x 10-4 mol dm-3 Rose Bengal (RB) and appropriate amounts of Mn" were mixed with doubly distilled water to obtain a volume of 50 ml. The injection of 0.2 ml of sulphite stock solution was carried out at the beginning of the irradiation of the reaction solution in order to initiate the reaction. The oxygen consumption was measured for a fixed time interval. Results and Discussion Photosensitized Activation Preliminary experiments were performed to determine the general influence of the photosensitizer on the catalysed and uncatalysed auto-oxidation of sulphite.The use of Methylene Blue (MB) as a photosensitizer at a wavelength of 575 nm, which approximately agrees with the absorption band of the dye, i.e., 665 nm, has shown that non-optimum excitation conditions existed leading to a fast uncatalysed reaction and a54 ANALYST, JANUARY 1991, VOL. 116 1 0 1 2 3 4 5 6 7 8 [Sensitizer]/lO-8 rnol dm-3 Fig. 1 Dependence of the reaction rate on the concentration of the sensitizer with 1-3, the catalysed and 4-6, the uncatalysed reaction. 1 and 4, RB; 2 and 5, MB; and 3 and 6, Rh. 1.2 mmol dm-3 S032-; 15.4 wmol dm-3 Mn"; and HBO 500, h. = 546 nm r I ? 0.3 I - 6 0.2 1 +- E 5 0.1 u o .- 4- m 1 I 7 8 9 10 1 1 PH Fig. 2 pH-dependence of the reaction rate with different buffer systems: solid line, borate buffer; and broken line, non-complexing buffer.1 and 3, catalysed reaction; 2 and 4 uncatalysed reaction. Experimental conditions: 1.2 mmol dm-3 SO3*-; 4 X 10-8 rnol dm-3 RB; HBO 500, h. = 546 nm; and 15.4 pmol dm-3 Mn" low catalytic activity of Mn". Unfortunately, the catalysed reaction was not accelerated by increasing the concentration of MB. Therefore, it was necessary to find the optimum combination of photosensitizer, irradiation wavelength and catalyst. Hence, a number of relevant triplet sensitizers and irradiation with the intensive emission band of the lamp at 546 nm were tested. The reaction-rate dependence on dye concentrations for the catalysed and uncatalysed reactions is shown in Fig. 1. Of the dyes investigated only Rhodamine B (Rh), MB and RB sensitized the catalytic reaction.The most effective photosensitizer was RB, firstly, because its absorp- tion band corresponds to the irradiation wavelength at 546 nm, and secondly, because RB is highly effective for inter- system crossing (ISC) (quantum yield @ = 0.8 for ISC) and other irradiationless de-activation processes. 15 It can be concluded from Fig. 1 that the activating influence of the photosensitizer attains a maximum at a concentration of about 4 x 10-8 rnol dm-3. The calibration graphs over the range 0-1 x 10-8 rnol dm-3 Mn" in unbuffered solution for different RB concentrations were not suitable for analysis because of the non-linearity of the graphs. Another disadvantage was the fast uncatalysed reaction compared with the catalysed reaction.The pH dependence of the photosensitized auto-oxidation of sulphite catalysed by Mn" was determined in the pH range 7-10. The results for two buffer systems are shown in Fig. 2, which indicates that the catalytic activity of Mn" attains a maximum at pH 8.4 and drops sharply for pH values different from 8.4 using the borate buffer system. No catalysed dark reaction could be observed in the pH range investigated. Finally, with optimized reaction conditions for the Mn"- catalysed reaction [1.2 x lo-* rnol dm-3 sulphite, 4 X 10-8 rnol dm-3 RB, pH 8.4 (borate), T = 20 "C and HBO 500, h = 546 nm] a linear calibration graph was obtained from 8 to 50 ppb of Mn". The relative standard deviations (RSDs) for 8 and 42 ppb of Mn" are typically 10 and 6.4%, respectively. Table 1 Comparison of the action of chemical activators and photosensitizers Concentration/ Reaction rate*/ Sensi tizedactivator mol dm-3 .mg O2 1-1 s-1 Rose Bengal, hv 4 x 10-8 0.069 Rose Bengal, hv 4 x 10-7 0.240 1,lO-Phenanthroline 8 x 10-4 0.031 Histidine 1 x 10-3 0.018 Acet ylacetone 1.2 x 10-3 0.015 Nitrilotriacetic acid 8 x 10-7 0.108 * Reaction rate = V,,ta)ysed - v , , , , ~ ~ ~ ~ ~ ~ ~ .Reaction conditions: 1.2 mmol dm-3 SO3*-, pH = 8.4, 0.62 pmol dm-3 Mn". - I v) 0.06 1 0" 0) Q, 0.04 4- h C 0 Q, 0.02 .g a I * I I 1 I 1 4 8 12 16 20 24 [L1/10-4 rnol dm-3 Fig. 3 Comparison of the reaction rate on complex formation of the different forms of the MnII-acac complex: solid line, reaction rate; and broken line, concentration of the different complex forms. Experimental conditions: 0.62 pmol dm-3 Mn"; pH = 8.4; and 1.2 mmol dm-3 S032- Chemical Activation The detection limit of the catalytic method should be improved further by combining the photoactivated system with the chemical activation of Mn".At the same time the chemical activator should be capable of decreasing the rate of the uncatalysed photosensitized reaction. The influence of different ligands which are known to increase the catalytic activity of Mn" was compared with the action of light and RB. A comparison of the reaction rates presented in Table 1 indicates that only the activation by nta is comparable to the activation by photosensitization. The activation properties of the ligands can be expressed in order of decreasing efficiency as follows: nta > phen > his > acac >> ox==en=bpy.The dependence of the reaction rate on the concentration of nta shows that the most catalytically active form is the 1: 1 complex. This complex is able to coordinate oxygen or sulphite or both. Nitrilotriacetic acid also stabilizes the +3 state of the catalyst [log /3 (Mn"', nta) = 20.35],16 which probably initiates the radical chain reaction of sulphite. The use of nta as an activator for Mn" using the indicator reaction of Malachite Green and periodate has been described by Mottola and Heath." Similar results to those for nta have been obtained with acac, where again the 1 : 1 complex was the most effective form of the catalyst (Fig. 3). On the basis of the following equilibrium constants, log f3 (MnII, acac) = 4.24 and log p2 (Mn", acac) = 7.35,16 plots of simulated complex equilibria (COMICSlS) of the Mn"-acac complexes indicate that the 1 : 1 complex is the most effective form.The slight difference between the maxima of reaction rate and complex formation of the 1 : 1 complex can probably be attributed to a kinetic hindrance. The activation of the catalytic reaction by en, bpy or ox is negligible. With the activation of the auto-oxidation of sulphite by ligands such as phen, his and acac, the determination of Mn" is possible with a detection limit of greater than 2 ppb. The ligand nta is apparently the most effective activator for Mn" with a detection limit of 1.1ANALYST, JANUARY 1991, VOL. 116 55 c h 0.6 r Table 2 Influence of interfering ions on the photosensitized, acac- activated and thermal, acac-activated methods Ratio of to thermal Ratio of [Mn"] to [interfering ion] photosensitized Interfering ion Photosensitized" Thermall- method Co" 1 : 100 Cu" 1 : 500 Nil1 1 : 650 C P 1 : 585 Cd" 1 : 1000 Fe"' 1 : 2000 Sulphate Nitrate 1: 1.3 x 105 1 : 3 x 106 Phosphate - 1 : 0.08 1 : 2.5 1 : 23 1 : 20 1 : 1000 1:13 1 : 3 x 106 1 : 6500 1 : 1300 1250 200 28 29 1 160 1 20 - * 3.1 x 10-8 rnol dm-3 Mn".Conditions for the photosensitized, aeac-activated method: 1.2 mmol dm-3 SO3*-; 1.2 mmol dm-3 acac; 1.6 X lo-' rnol dm-3 RB; pH = 8.4; and HBO 500, h = 546 nm. t 6.2 x rnol dm-3 Mn". Conditions for the thermal, acac-activated method: 0.6 mmol dm-3 SO3;?-; 1.2 mmol dm-3 acac; and pH = 8.4. c 0.8 .- I - 0" 0.6 3 0.4 2 F 4- g 0.2 , ' I I I I I .- c.m 0 4 12 29 28 [RB]/10-8 mol dm-3 Fig. 4 Dependence of the reaction rate on the RB concentration in the presence (1 .and 3) and absence (2 and 4) of acac: solid line, catalysed reaction; and broken line, uncatalysed reaction. Experimental conditions: 1.2 mmol dm-3 SO3*-; 0.62 pmol dm-3 Mn"; and pH = 8.4 ppb, but the linear range of the calibration graph depends strongly on the Mn : nta ratio. It is also possible to determine the nta concentration by this method. The interference levels for several metal ions and anions are summarized in Table 2. Data in the second column of Table 2 indicate that CoII catalysed the indicator reaction more effectively than did Mn". The interference levels of Cull, Ni", CrIIl and F e I I I are fairly close to the Mn" concentration.A comparison of the method proposed above with the Mn"-catalysed reaction between iodide and periodatel9 shows only a small improve- ment in the selectivity of the proposed method in most instances. Combination of Photosensitized and Chemical Activation The first two parts of this paper have shown the improvement of the sensitivity of catalytic methods by chemical activation of MnI1 and also by photosensitization. Therefore, it was of interest to investigate the simultaneous influence of irradia- tion with light and chemical activation by the ligand. Prelimi- nary experiments for the characterization of the catalytically most active form of the catalyst in the photoactivated system provided the same results as for chemical activation alone. The combination of photoactivation and chemical activation led to a change in the efficiency of the ligands used as follows: acac > nta > his > phen >> bpy > o x z e n .The chemical activator acac has two significant attributes, shown in Fig. 4. Firstly, the formation of a complex between MnI1 and this ligand improves the catalytic action of the catalyst and secondly, the uncomplexed ligand decreases the t - n z , " - n T 0 1 2 3 4 [RB]/10-7 mol dm-3 Fig. 5 Dependence of the reaction rate on the concentration of the sensitizer for different chemical activators. Reaction rate = Vc,t,ly& - vUncatal sed (reaction rate in mg O2 1-1 s-1). 1,1.2 mmol dm-3 acac; 2,2 pmol d;n-3 nta; 3,O.S mmol dm-3 phen; 4,0.4 mmol dm-3 his; and 5 , en, ox, bpy. Ex erimental conditions: 1.2 mmol dm-3 SO3*-; 0.62 pmol dm-3 MnI{ and pH = 8.4 Table 3 Comparison of the analytical parameters of the thermal, photosensitized and combined method with acac as chemical activator and RB as photosensitizer Parameter Detection Method limit (ppb) RSD (%) Thermal 1.8 3.3 Photosensitized 8.0 9.5 Combined 0.3 5.5 rate of the uncatalysed reaction.Therefore, it is possible to increase the RB concentration in order to enhance the detection limit further (Fig. 5 ) . The influence of the light intensity on the acceleration of the catalysed reaction was investigated with a continuous argon laser. The excitation occurred at 514.5 nm. The experiments have shown that the reaction rate depends more strongly on the light intensity than on the concentration of the photosensitizer.Unfortunately, because the use of powerful light sources in analytical chemistry is limited, an increase in the photosensitizer concentration was necessary in order to improve the sensitiv- ity. With the optimized reaction conditions [1.2 mmol dm-3 S032-, pH 8.4 (4 mmol dm-3 borate), T = 20 "C, HBO 500, h = 546 nm] a linear correlation between the reaction rate and the concentration of Mn" exists from the detection limit up to approximately 9 ppb of Mn". The RSDs for 0 and 8.5 ppb of Mn" are 5.5 and 1.3%, respectively. The average and RSD values were obtained from five parallel measurements. With the RSD of the uncatalysed reaction the detection limit for MnII using the photosensitized, chemically activated reaction was calculated to be 0.32 ppb of MnlI.A comparison of the analytical parameters of the three proposed methods is presented in Table 3. It can be seen from the results that the combination of photoactivation with chemical activation yields the maximum sensitivity. The first column of Table 2 lists the ratio of MnII to interfering ion, when the interference began (average k 2 RSD). A comparison of the first and second columns shows the further improvement of selectivity in most instances by the choice of a selective excitation of the catalyst by photosensitization. The reason for the differences in selectivity appears to be due to the different photochemical reaction pathways used to oxidize the sulphite radicals. As described in the introduction the photo-reduction process, e.g., for the Fe"1 ion, occurs via complex formation of Fell with sulphite, absorption of light by this complex at approxi- mately 300-350 nm and electron transfer in the excited state.The irradiation of the reaction solution with light which cannot be absorbed by this complex leads to the same behaviour of the FeIII ions as in the non-photoactivated, acac-activated reaction. The enhanced reactivity of MnII in contrast to Fe"156 ANALYST, JANUARY 1991, VOL. 116 or CoII arises from the possibility of excitation of the MnIIL acac complex by photosensitization with RB and the accelera- tion of the MnlI1 reduction process by sulphite. The chain reaction of the sulphite radicals formed in the previous reaction increases the catalytic reactivity of Mn". It should be noted that the reaction mechanism was not investigated in detail.However, the formation of MnII during the catalytic reaction was detected by complex formation with pyrophos- phate and by optical absorption spectrometry of the MnIII- acac complex. Fig. 6 shows that the conversion of MnIII-acac into [Mn(H2P207)3]3- led to a decrease in the catalytic activity of Mn2+ at [P2074-] > 1 X 10-5 mol dm-3. Pyrophosphate is known to stabilize manganese in the +3 state.20 Further support for MnIII is provided by spectrophotometric observa- tion. The slow oxidation of Mn" by oxygen can be accelerated by using acac and a ligand.21 In aqueous solution (pH 8.4) the MnILacac complex shows an absorption band at 294 nm. In the presence of oxygen (0.2 mmol dm-3) and perchlorate (1.5 mmol dm-3) the absorption of the MnILacac complex was changed.The 294 nm band disappeared and was replaced by a band at 222 nm. This band originates from MnIII(aq) which is stable in the presence of perchlorate. This value shows good agreement with that previously measured by Wells and Davies22 and Bielski and Chan.23 Without oxygen no MnIII was observed. The photosensitized oxidation of MnII to MnIII was five times faster than the corresponding dark reaction. Other photochemical experiments suggested that the uncatalysed reaction occurs via singlet oxygen (lo2). The reaction can be inhibited by a powerful quencher of 1 0 2 such as his or acac. Another fact is that the triplet state of chlorophyll or benzophenone is quenched by MnII.24,25 The triplet energies (Et) of chlorophyll and RB of 13 200 and 13 800 cm-1, respectively, are similar, which means that a similar energy transfer from the photosensitizer to the catalyst should be possible.With the experimental data obtained the catalysed reaction mechanism of the photosensitized auto-oxidation of sulphite seems to be a combination of two amplification processes; firstly, the catalytic cycle of MnII-MnIII and secondly, the chain reaction of sulphite to sulphate initiated by MnIII.26 The uncatalysed reaction occurs via sulphite oxida- tion by singlet oxygen. The reaction mechanism is shown schematically below (L represents indicates the activated complex). Catalysed reaction scheme: hv RB - 3RB + MnIIL- 3RB. . .MnIIL- *MnIIL + O2 - MnIII(aq)+ S032- - SO3-- + O2 - SO5.- (chain reaction) - Mn"(aq) + L - 0 2 Uncatalysed reaction scheme: 3RB + 302 - 1 0 2 + so32- - S032- + 3 0 2 - SO5.- (chain reaction) - 0 2 the ligand and the asterisk 3RB 3RB.. .MnIIL RB + *MnIIL MnIII(aq) + 02- + L MnII(aq) + SO3.- SO5.- S042- Mn" L 1RB + ' 0 2 SO3.- + 02- soy - S042- The applicability and accuracy of the proposed photosensi- tized, acac-activated method for the determination of MnII were tested on simulated and natural drinking water. The simulated drinking water (SDW) was prepared by mixing the individual components. The composition of the water based on World Health Organization (WHO) standard values is summarized in Table 4. The average values of five parallel 5 4 3 -Log ([P20,4-I/mol dm-3) Fig. 6 Dependence of the reaction rate on the pyrophosphate concentration; reaction was recorded by: solid line, oxygen consump- tion and broken line, changes of the absorption of sulphite at 231 nm.Experimental conditions: 1.2 mmol dm-3 pH = 8.4; 4 x 10-8 mol dm-3 RB; 1.5 X 10-5 mol dm-3 Mn"; and HBO 500, k = 546 nm Table 4 Composition of the SDW and WHO standard values* Component Pb As Cr Cd Ba Fe Mn Zn c u Ag WHO1 mg 1-1 0.05 0.05 0.05 0.005 1.0 1 .o 0.5 0.05 0.1 15 * See reference 27. SDW/ mg 1-1 0.1 0.05 0.05 0.02 0.7 1.0 1.1 0.13 0.1 10 Component Ca Ni c o F- NO3- c1- SO& ~ 0 ~ 3 - Mg p205 WHO/ mg 1-1 220 125 - - 1.7 45 5 600 400 - SDWI mg 1-1 80 75 0.02 0.015 0.95 50 5 350 290 0.2 Table 5 Comparison of the photosensitized and thermal acac-activated methods Method Parameter Photosensitized Thermal Average concentratiodmg 1-1 1.145 0.850 RSD (%) 2.9 1.4 Deviation from the theoretical SDImg 1-1 0.033 0.010 Theoretical concentratiodmg 1-1 1.100 1.100 value (%) +4.1 -27.7 measurements are given in the first column of Table 5.For comparison the data of the acac-activated (thermal) reaction are given in the second column. Only the photosensitized, acac-activated method makes it possible to determine Mn" in this complex matrix with sufficient accuracy. A 0.1 ml volume of natural drinking water was directly introduced into the reaction solution. The concentration of Mn" was in the range 0.09-0.13 mg 1-1 depending on the source of the water and the day on which the sample was taken. Conclusion The proposed method for the determination of MnII is based on a combination of photoactivation with chemical activation of the catalyst by ligands.The reactivity of Mn" is approxi- mately ten times higher for the photosensitized, acac-acti- vated reaction than for the system that is only chemically activated. For monitoring the reaction rate, the oxygen consumption was measured with an oxygen-selective Clark electrode. The determination of Mn" is possible in the presence of a 100-1000-fold excess of Co", Cu", Ni", CrIII and Cd". Iron(1Ir) interferes at about a 2000-fold excess. The detection limit of the proposed method is 0.3 ppb with an RSDANALYST, JANUARY 1991, VOL. 116 57 of about 5.5%. The energy transfer from the triplet RB to the Mn"-acac complex represents the photoactivation. The oxi- dation of the excited Mn" complex by oxygen is therefore accelerated.The reaction of MnIII with sulphite initiates the chain reaction to form sulphate. The catalytic method described was used for the determination of Mn" in drinking water. 1 2 3 4 5 6 7 8 9 10 11 References Mueller, H., Otto, M., and Werner, G., Catalytic Methods in Trace Analysis, Modern Trace Analysis, Akademische Verlags- gesellschaft Geest & Portig, Leipzig, 1980, vol. 4. Kopanica, M., and Stara, V., Kinetic Methods in Chemical Analysis, Wilson and Wilson's Comprehensive Analytical Chem- istry, Elsevier, Amsterdam, 1983, vol. 18. Mottola, H. A., Kinetic Aspects of Analytical Chemistry, Chemical Analysis, Wiley, New York, 1988, vol. 96. Perez-Bendito, D., and Silva, M . , Kinetic Methods in Analytical Chemistry, Wiley, New York, 1988.Mueller, H., CRC Crit. Rev. Anal. Chem., 1982, 13, 313. Werner,,G., Quim. Anal. (Barcelona), 1983, 2, 68. Veprek-SiSka, J., LunBk, S., and El-Wakil, A., 2. Naturforsch, B , Anorg. Chem., Org. Chem., 1974, 29, 812. Lunak, S., El-Wakil, A., and Veprkk-SiSka, J., Collect. Czech. Chem. Commun., 1978, 43, 3306. Hayon, E., Treinin, A., and Wilfs, J., J. Am. Chem. SOC., 1972, 94, 47. Gmelins Hundbuch der Anorganischen Chemie, Sulfur (Part B ) , Verlag Chemie, Weinheim, 1963. Balzani, V., and Carassiti, V., Photochemistry of Coordination Compounds, Academic Press, New York, 1970. 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Lunhk, S., and VeprCk-SiSka, J., React. Kinet. Catal. Lett., 1978, 8, 483. VeprCk-Siska, J., and Lunak, S., J. Photochem., 1978,8, 391. Thiel, A., and Peter, O., 2. Anorg. Allg. Chem., 1928,173,174. Becker, G. O., et al., Einfiihrung in die Photochemie, Deutscher Verlag der Wissenschaften, Berlin, 1976. Inczedy, J. , Analytical Application of Complex Equilibria, AcadCmiai Kiado, Budapest, 1976. Mottola, H. A., and Heath, G. L., Anal. Chem., 1973,44,2322. Otto, M., personal communication. Tigiuyanu, Y. D., and Oprya, V. I., Zh. Anal. Khim. , 1973,28, 2206. Gmelins Handbuch der Anorganischen Chemie, Manganese C9, Springer-Verlag, Berlin, Heidelberg and New York, 1983. Yamasaki, K., and Sone, K., Nature (London), 1950,166,998. Wells, C. F., and Davies, G., J. Chem. SOC., A , 1967, 1858. Bielski, B. H. J., and Chan, P. C., J. Am. Chem. SOC., 1978, 100, 1920. Lindschuetz, H., and Sarkanen, K., J. Am. Chem. SOC., 1958, 80, 4826. Ledger, M. B., and Porter, G. , J. Chem. SOC., Faraday Trans 1, 1972, 68,, 539. Veprkk-SiSka, J., Lunhk, S., Lederer, P., and Mach, J., Oxid. Commun., 1984,6, 381. Carlson, S., Die Begrenzung von Wasserinhaltsstoffen in Empfehlungen, Standard und Normen der WHO und Ver- schiedener Lander. Schriftenreihe Ver. Wasser-BodenLuft- hygiene, Berlin-Dahlem, 1973, 40, 209. Paper Ol01526A Received April 4th, 1990 Accepted July 13th, 1990

ISSN:0003-2654    

DOI:10.1039/AN9911600053

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, JANUARY 1991, VOL. 116 59 Determination of Citric Acid Based on Inhibition of the Crystal Growth of Calcium Fluoride F. Grases, A. Costa-Bauza and J. G. March Department of Chemistry, University of Balearic Islands, 07071-Palma de Mallorca, Spain Inhibition of the growth of calcium fluoride crystals in the presence of citrate was followed using a kinetic-potentiometric technique and a calcium ion-selective electrode, and as a consequence, a method for the determination of citrate in the range 0.5-2.4 pg ml-1 has been developed. The method was successfully applied to the determination of citrate contained in pharmaceutical products and urine. Urine analysis requires prior separation of phosphate, sulphate and magnesium(l1). Elimination of these interferences was studied and accomplished using precipitation processes.Magnesium and phosphate were jointly eliminated in basic media by the addition of ammonium ions. Phosphate and sulphate were eliminated with barium(i1). Phosphate was also eliminated as a lithium salt. Keywords: Citric acid determination; calcium fluoride; crystal growth; urine analysis The determination of citric acid in various biological and non-biological samples has aroused increasing interest in recent years. A range of determinations in food'-11 and pharmaceutical preparations12.13 have been described. The dctermination of citrate in urine is also very important clinically because of its effective inhibitory action on calcium oxalate urolithiasis. 14-16 Common methods for the determination of citric acid can be classified into three main groups: gravimetric and volumetric methods;1"17-19 enzymic analyses;20-23 and chromatographic determinations, mainly high-performance liquid chromato- graphy (HPLC).1-11 Enzymic and HPLC procedures present notable advantages when the determination of citric acid is performed in a complex matrix. Owing to the specificity and selectivity of these techniques, sample preparation is minimal. A very simple spectrophotometric procedure has been pro- posed for the determination of citric acid in urine through the formation of a yellow complex with Fe"' after the prior separation of phosphate through its precipitation using magnesium and ammonium ions before the complexation .24 In the present work, the inhibition of the growth of calcium fluoride seed crystals by citrate was investigated.As a consequence, a kinetic method for the determination of citrate is proposed; this method was applied to several samples. The use of inhibition processes in crystal growth has also been demonstrated by determining phytic acid using seed crystals of calcium oxalate monohydrate25 and by determining phosphate ions using calcite (calcium carbonate) seed crystals.26 Experimental Reagents and Apparatus Sodium fluoride, calcium chloride, acetic acid and ammonia were purchased from Panreac (Barcelona, Spain), and sodium citrate (disodium salt) and calcium fluoride from Probus (Barcelona, Spain). All reagents used were of analytical- reagent grade. A suspension of calcium fluoride seed crystals was prepared by mixing 5 g of calcium fluoride and 25 g of water, then stirring magnetically for 48 h or more (aged seed).Enzymic citric acid analyses were performed using a Test- combination supplied by Boehringer Mannheim (Catalogue No. 139076). Potentiometric measurements were obtained using a Crison (Barcelona, Spain) 2002 micropotentiomcter equipped with a calcium ion-selective electrode [Ingold (Urdorf, Switzerland)] coupled with a silver-silver chloride reference electrode separated from solution with an intermediate junction con- taining potassium chloride. The seed crystals were charac- terized by using a Hitachi S-530 scanning electron microscope. Procedure for the Determination of Citric Acid To a 250 ml glass beaker, the following solutions were added: 8 ml of acetic acid-ammonium acetate buffer solution (total concentration 5 mol dm-3, pH 6.0); 2 ml of 0.10 mol dm-3 calcium chloride; and sufficient citrate solution to give a final concentration in the final volume (200 ml made up with water) of 0.5-2.4 pg ml-1.This solution was stirred magnetically at 500 rev min-1 with a synchronous motor stirrer (controlled agitation is always essential in order to keep the slurry in suspension), and 1 ml of aged calcium fluoride suspension was added. The electrodes were immersed in the resulting suspension and, when the potentiometer gave a constant reading (after a few seconds), 0.70 ml of a 0.80 mol dm-3 sodium fluoride solution was added, which initiated crystal growth, and consequently a decreasing electrical response which was measured continuously on a chart recorder. The calibration graph was obtained from the difference, in millivolts, between the suspension that contained citrate and the suspension that did not (blank), 5 min after the addition of fluoride.All experiments were carried out at room tempera- ture (25 "C). Procedure for Urine Analysis A 2 ml aliquot of urine and 1 ml of 5 mol dm-3 ammonia solution were placed in a test-tube, the resulting turbid solution was warmed at 90 "C for 30 min and after cooling, filtered through a disposable 0.22 pm filter membrane coupled to a syringe. The filtering system was washed out with 1 ml of water, then 1 ml of 0.35 mol dm-3 barium chloride was added and the warming and filtering procedure repeated. An aliquot of the resultant solution was used for the determination of citrate using the procedure described earlier.Results and Discussion Effects of Polycarboxylic Acids on the Crystal Growth of Calcium Fluoride The growth of calcium fluoride seed crystals was studied in the presence of several di- and tricarboxylic acids. As can be seen in Fig. 1, the growth of calcium fluoride crystals was notably delayed by the presence of some carboxylic acids, such that citric acid and oxalic acid delayed crystal formation for the longest and shortest periods, respectively. However, succinic, glycocholic and malonic acids had no significant effect under the experimental conditions used. These results demonstrate that the presence of two or more adjacent carboxylic acid groups permits very effective adsorption of the organic acids60 > E : -10 .- c) Q, a 0) 4- .- 4- - -30 Q, a ANALYST, JANUARY 1991, VOL.116 ’ . 5 - - 50 > 0 2 4 Timehin Fig. 1 Calcium fluoride crystallization in the presence of various carboxylic acids: A, citric acid; B , oxalic acid; C, D and E, succinic, glycocholic and malic acids; and F, crystallization in the absence of a foreign carboxylic acid. The initial concentrations of these carboxylic acids were, in all instances, 2 X 10-5 rnol dm-3. All other experimental conditions are given under Procedure for the Determi- nation of Citric Acid 0 I - Y U 8 7 6 > € 5 U 4 3 2 1 2 3 4 5 [F-1/10-3 rnol dm-3 Fig. 2 Effect of initial fluoride concentration on AmV. after 5 min, between the blank and a sample containing citrate at a concentration of 5 X 10-6 mol dm-3.All other experimental conditions are given under Procedure for the Determination of Citric Acid on the crystal surface. The adsorption of the polycarboxylic acids onto the active growth sites is likely to be responsible for the reduction in the crystal growth rates. In order to find the optimum conditions for the develop- ment of a kinetic procedure for the determination of citric acid, based on its inhibition of calcium fluoride crystal growth, the influence of supersaturation (initial concentrations of fluoride and calcium), pH and ionic strength (total concentra- tion of buffer solution) on the rate of reaction was studied. Fig. 2 shows the voltage difference in millivolts, after 5 min, between sample and blank corresponding to several fluoride concentrations.As can be seen for a total initial fluoride concentration of 2.8 x 10-3 rnol dm-3 the inhibitory effect of citrate was at a maximum. Consequently this concentration was chosen to establish the calibration graph. The results of the study of the influence of pH are summarized in Fig. 3. As is shown, the difference in electrical potential is only slightly affected for pH values between 4.5 and 7.0, thus indicating that some protonation of the citrate molecule did not affect its inhibitory capacity. When the influence of the ionic strength is considered it can be seen (Fig. 4) that again citrate inhibition is only slightly affected. The weak influence of the pH and ionic strength on the inhibitory effects of citrate can be explained by the observa- 8 1 7 1 4 5 6 7 PH Fig.3 Effect of pH on AmV, after 5 min, between the blank and a sample containing citrate at a concentration of 5 x 10-6 mol dm-3. The different pH values were obtained by adding distinct buffer solutions prepared by mixing suitable volumes of 5 rnol dm-3 acetic acid and 5 rnol dm-3 ammonia. The final pH of the suspension was checked with a pH-meter. No significant pH change was detected during crystal growth. All other experimental conditions are given under Procedure for the Determination of Citric Acid 8 7 > E U 6 5 0 0.1 0.2 [AcOH + AcO-l/mol dm-3 Fig. 4 Effect of total concentration of buffer solution (acetic acid + acetate) on AmV, after 5 min, between the blank and a sample containing citrate at a concentration of 5 x 10-6 mol dm-3.All other experimental conditions are given under Procedure for the Determi- nation of Citric Acid tion that the interactions between citrate oxygen atoms and calcium ions are not exclusively of an electrostatic nature. It is known that the morphology and surface characteristics of seed crystals can, in some instances, have considerable effects on the rate of crystal growth” and consequently can modify the influence of adsorbing inhibitor molecules in crystal growth reactions. Therefore, to evaluate the effect of size and morphology of the calcium fluoride seed on crystalli- zation runs, experiments were carried out using seeds with two different surface characteristics: with and without ageing treatment. As can be seen in Fig. 5 , noticeable variations were observed depending on the seed, i.e., recently prepared or aged seed, indicating that citrate was more effective in reducing the rate of growth of the most perfect crystals.This can be explained as a consequence of a relatively small number of active growth sites that can be blocked by the inhibitor while increasing the perfection of the crystals. With the aim of attaining maximum sensitivity values (lower detection limit), an aged seed was chosen for the preparation of the calibration graph. Characteristics of the Analytical Methods The fixed-time method ( 5 min) was applied to the potential (mV) versus time curves, recorded in the presence of different amounts of citrate under selected conditions in order to obtainANALYST, JANUARY 1991. VOL. 116 61 -10 - > E : m C P) .- U U - 50 I I I 0 2 4 6 Tim e/m i n Fig.5 Scanning electron micrographs (SEM) and crystallization graphs of calcium fluoride obtained from different aged seed crystals. Stirring time: A, 30 min; B , 24 h; C. 48 h; and D. 5 d the calibration graph, which was linear in the range of citratc concentrations between 0.5 and 2.4 pg m1-I. The relative standard deviation was 2.0% (n = 11, a = 0.05). The selectivity was tested by obtaining the rate curves in the presence of several species that in some instances accompany citrate in biological samples such as urine. The results are Table 1 Effect of foreign ions on the determination of 6 X mol dm-3 citrate Ion Na', Ball, LiI, C1-, urea Succinate Sulphate Phosphate Ethylenediaminetetraacetic acid Oxalate Tartrate Malate * The error was always positive.Concentration tested/mol dm-3 (1-2) x 10-3 1 x 10-3 1 x 10-4 1 x 10-3 1 x 10-4 i x 10-5 1 x 10-4 1 x 10-3 1 x 10-5 1 x 10-5 1 x 10-3 1 x 10-4 1 x 10-4 1 x 10-5 1 x 10-3 1 x 10-4 1 x 10-5 5 x 10-7 1 x 10-5 1 x 10-6 1 x 10-6 4 x 10-6 1 x 10-6 1 x 10-6 1 x 10-6 1 x 10-6 Error* (Yo ) 0.1-1.5 50 134 56 300 60 456 220 34 10 268 156 207 34 414 230 44 4 1.6 1.7 1.7 1.9 1.1 1.9 1.7 1.8 summarized in Table 1. Most of the interfering species caused a decrease in the difference of electrical potential. As can be seen, the main interferences were caused by phosphate, sulphate and magnesium ions. Elimination of these interfer- ents was extensively studied and accomplished through precipitation processes as described below. It is interesting to compare the selectivity of the proposed procedure with that described in previous papers25326 using seed crystals of a different nature.Thus it can be observed that, in general terms, the selectivity obtained when using calcium fluoride seed crystals is considerably inferior to that obtained when using calcium oxalate monohydrate or calcium carbonate (calcite) seed crystals. These observations can be explained by considering that owing to the simpler crystal structure of calcium fluoride seed crystals, the number of molecules or ions that can cause some blockage of the active growth sites is noticeably increased. When working with more complex crystalline structures the number of species that can adapt to the morphology of the active growth sites is restricted.Elimination of Interferences The elimination of the interference from phosphate, sulphate and magnesium(i1) was studied in order to establish the reliability of the determination of citrate in human urine. The separation processes were carried out on solutions prepared in the laboratory; the composition of these solutions was considered to be representative of human urine.28 Ordinarily a solution containing S042- (10.5 mmol dm-3), H2P04- (22 mmol dm-3), HP04'- (3.0 mmol dm-3), Cl- (240.2 mmol dm-3), C6HSO73- (4.0 mmol dm-3), Na+ (170.7 mmol dm-3), NH4+ (43.4 mmol dm-3), K+ (81.3 mmol dm-3) and Mg2+ (3.0 mmol dm-3) was used as synthetic urine. The elimination of Mg" to a concentration level that did not interfere in the determination of citrate in urine-was accom- plished by adding 1 ml of 5 mol dm-3 ammonia to 2 ml of synthetic urine.Under such conditions, ammonium magne- sium phosphate precipitated. After the precipitate had been digested at 90 "C over a period of 30 min, the solid phase was62 ANALYST, JANUARY 1991, VOL. 116 Table 2 Comparison of results obtained for the determination of citric acid in commercial formulations and urine. Results, the average of three separate determinations, are expressed as mg I-' with standard deviation (%) given in parentheses Spectro- Enzymic Sample method method method Benadryl (Parke-Davis)* 8800 (2.2) 8450 (1.9) 8700 (6.2) Benylin (Parke-Davis)? 4300 (3.2) 4300 (2.0) 4400 (6.5) Proposed photometric (reference) Pharmaceutical product- Urine- Synthetic$ Human 1 2 3 4 5 6 7 8 700 (3.7) 600 (4.0) 500 (4.2) 940 (4.0) 1110 (3.9) 550 (3.5) 730 (3.2) 640 (3.0) 700 (3.1) 725 (2.6) 720 (6.1) 819 (3.1) 561 (5.9) 679 (3.5) 474 (6.2) 1162 (3.2) 910 (6.3) 1148 (3.7) 1140 (7.0) 600 (2.9) 530 (5.7) 809 (3.1) 715 (5.4) 700 (2.8) 650 (6.7) 715 (2.5) 716 (5.0) * Value claimed by the manufacturer, 8350 mg I-'.t Value claimed by the manufacturer, 5183 mg I-'. $ Amount added, 750 mg 1 - l . S042- Precipitate Solution Ba3(PO4I2 A Ready for citrate BaSO, analvsis Fig. 6 normally present in human urine Schematic diagram for the elimination of interferent species separated from the solution by filtration through a disposable 0.22 pm filter. Phosphate and sulphate were then separated jointly from the solution, which did not at this stage contain Mg", by adding 1 ml of 0.35 mol dm-3 barium chloride.The new solid phase formed (barium sulphate and barium phos- phate) was removed from the solution in a manner similar to that used for the ammonium magnesium phosphate, i. e., digestion and filtration. After such treatment, the determina- tion of citrate by applying the proposed procedure was reliable with an error of 6.7%, as can be seen in Table 2, confirming that the elimination of interferents was satisfactory. Further experimental details are given under Procedure for Urine Analysis. Fig. 6 shows a schematic diagram of the separation processes. It is obvious that if the sample did not contain phosphate, the elimination of magnesium would not take place by the addition of ammonia only; under such circum- stances the addition of phosphate would be necessary. If phosphate was the only interferent present in the sample, its elimination could be carried out, either by adding barium(iI), as described, or by adding lithium chloride and sodium hydroxide (1 ml of 3 mol dm-3 lithium chloride and 2 ml of 2 mol dm-3 sodium hydroxide were added to 2 ml of synthetic urine without magnesium and sulphate).The remaining lithium(1) in solution did not affect the subsequent determina- tion of citrate. Finally, if sulphate was the only species to be eliminated, its quantitative precipitation as a barium salt from a synthetic urine which contained neither phosphate nor magnesium would take place without the addition of ammo- nia. However, when phosphate was the species being sep- arated, the satisfactory precipitation of barium phosphate took place only in basic media and the addition of ammonia, before the addition of barium(n), was necessary.Application The proposed method was applied to the determination of citrate in pharmaceutical products, synthetic urine and several samples of human urine. It should be noted that the determination of citrate in pharmaceutical products was carried out without any previous treatment of the sample (except appropriate dilution). Nevertheless, application to urine samples required previous separation of phosphate, sulphate and magnesium(II), as mentioned earlier. In order to confirm the reliability of the proposed procedure the results were compared with those obtained by spectrophotometry based on the formation of the yellow FeIII-citrate complex,24 and with those obtained with the enzymic procedure based on the transformation of citrate to oxaloacetate and acetate catalysed by the enzyme citrate lyase (CL): CL Citrate 4 oxaloacetate + acetate In the presence of the enzymes malate dehydrogenese (MDH) and lactate dehydrogenase (LDH), oxaloacetate and its decarboxylation product, pyruvate, are reduced to L-malate and L-lactate, respectively, by reduced nicotinamide adenine dinucleotide (NADH) : MDH Oxaloacetate + NADH + H+ + L-malate + NAD+ LDH Pyruvate + NADH + H+ - L-lactate + NAD+ Then, the amount of NADH oxidized is stoichiometric with the amount of citrate.The diminution in the concentration of NADH is evaluated spectrophotometrically at 340 nm.22 A kit containing all the reagents mentioned is commercially avail- able.(See under Reagents and Apparatus.) This enzymic procedure was considered as a reference method. The results are summarized in Table 2. As can be seen, the relative difference (as a percentage) between the proposed method and the enzymic method, ranged between 1.5 (sample 7) and 7.0 (sample 1). Nevertheless, when a spectrophotometric method was applied, samples 4-8 gave an acceptable relative difference (from 0.1 to 13.2), but, samples 1, 2 and 3 gave an unacceptable relative difference (46.0, 43.2 and 27.7, respec- tively), indicating that the spectrophotometric method could give erroneous results depending on the sample. The discrep- ancy can be explained by taking into consideration the fact that the formation of the yellow FeIII-citrate complex is strongly dependent on the pH of the medium, and that the amounts of substances with acid-base behaviour contained in real urine samples vary widely (depending on the sample).Thus, the addition of the same amount of hydrochloric acid to different samples (after the elimination of interferences) could give different final pH values, which affect the formation of the Fellkitrate complex and the absorbance value corre- sponding to the FeIlI and urine blanks. Financial support by the 'Direccion General de Investigacion Cientifica y Tecnica' through grant 86-0002, is gratefully acknowledged. One of the authors (A. C.-B.) thanks the 'Fundacio Joan Muntaner' for a fellowship. References 1 Bushway, R . J . . Bureau, J .L., and McGann, D. F., J . Food Sci., 1984, 49, 75.ANALYST, JANUARY 1991, VOL. 116 63 2 3 4 5 6 7 8 9 10 I 1 12 13 14 15 16 Steiner, W., Mueller, E., Froehlich, D., and Battaglia, R., Mitt. Geb. Lebensmittelunters. Hyg., 1984, 75, 37. Marsili, R. T., Ostapenko, H., Simmons, R. E., and Green, D. E., J. Food Sci.. 1981,46, 52. Ashoor, S . H.. and Knox, M. J., J . Chromatogr., 1984,299,288. Picha, D. H., J. Agric. Food Chem., 1985,33, 743. Schneyder, J.. and Flack, W., Mitt. Kloesterneuburg, 1981, 31, 57. Shimazu, Y., and Watanabe, M., Nippon Jozo Kyokai Zasshi, 1984, 75, 37. Andersson, R., and Hedlund, B . , 2. Lebensm. Unters. Forsch.. 1983, 176,440. Hamakawa, H., Shimazaki, K . , Sukegawa, K., and Kato, I . , Rakuno Kagaku Shokuhin no Kenkyu, 1983, 32, 139.Wilson, Ch. W., Shaw, P. E., and Campbell, C. W., J. Sci. Food Agric., 1982. 33, 777. Shaw, P. E., and Wilson, Ch. W., J . Sci. Food Agric., 1983,34, 109. de Souza. N. E., Godinho, 0. E. S., and Aleixo, L. M., Analyst, 1985, 110, 989. El-Tarras, M. F., Pungor. E., and Nagy, G., Anal. Chim. Acta, 1976, 82, 285. Ryall, R. L.. Harnett, R. M., and Marshall, V. R.. Clin. Chim. Acta, 1981, 112, 349. Grases, F., Genestar, C.. March. P., and Conte, A . , Br. J . Urol., 1988. 62, 515. Conte, A., Roca, P., Gianotti, M., and Grases, F., Znt. Urol. Nephrol.. 1989, 21, 369. 17 18 19 20 21 22 23 24 25 26 27 28 Encyclopaedia of Industrial Chemical Analysis, eds. Snell , F. D., and Ettre, L. S . , Interscience, New York, 1970, vol. 10, p. 92. Tan, H. S., and Szopa, M. E., Anal. Lett., 1983, 16, 573. Bhosale, S. N., and Khopkar, S. M., Talanta, 1985,32, 155. Moellering, H., and Gruber, W., Anal. Biochem.. 1966, 17, 369. Zender, R., de Torrente, C., and Schneyder, U., Clin. Chim. Acta, 1969, 24, 335. Welshman. S. G., and McCambridge, H., Clin. Chim. Acta, 1973,46, 243. Methods of Enzymatic Analysis, ed. Bergmeyer, H. V., Academic Press, London, 1974, vol. 3, pp. 1563 and 1568. Millan, A., Conte, A., Garcia-Raso, A., and Grases, F., Clin. Chem., 1987, 33, 1259. Grases, F., and March, P., Anal. Chim. Acta, 1989, 219, 89. Grases, F., and March, J. G., Anal, Chim. Acta, 1990,229,249. Singh, R. P., Gaur, S . S . , White, D. J., and Nancollas, G . H., J. Colloid Interface Sci., 1987, 118, 379. Robertson, W. G., and Scurr, D. S . , J . Urol. (Baltimore), 1986, 135, 1322. Paper Ol00599A Received February 9th, I990 Accepted August 7th, 1990

ISSN:0003-2654    

DOI:10.1039/AN9911600059

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, JANUARY 1991, VOL. 116 65 Determination of Trace Amounts of Fluorine, Boron and Chlorine From a Single Sodium Carbonate Fusion of Small Geological Sample Masses Alfons Hofstetter and Georg Troll lnstitut fur Mineralogie und Petrographie der Universitat Munchen, Theresienstr. 4 1, 8000 Miinchen 2, Germany Diet mar Matt h ies Lehrstuhl fur Forstliche Arbeitswissenscha f i und Verfahrenstechnik, Hohenlindenerstr. 5,8000 Miinchen 80, Germany Trace amounts of fluorine, chlorine and boron (all below 100 ppm) in a single sodium carbonate fusion of 150 mg of geological material were determined potentiometrically (F) and spectrophotometrically (CI and B). The accuracy and precision of this method were tested on six international reference samples. Keywords: Fluorine, chlorine and boron determination; geological material; potentiometry; spectro- photometry Despite the importance of fluorine and boron in geochemical processes (for example in lowering of the solidus temperature for crystal growth during magma fractionation or as complex- ing agents in ore deposit forming events), the database for these elements is still small compared with other major and trace elements.This is particularly true for extra-terrestrial material or related matter, such as tektites or impact rocks. There are several reasons for this: ( i ) concentrations of fluorine, boron and chlorine in tektites or impact rocks are usually a factor of 10-100 lower than in igneous, metamorphic or sedimentary rocks (common values for tektites are 1240 pprn of fluorine, 6-50 ppm of boron and 6-15 ppm of chlorine according to Moore et al.,* Matthies and Koberl2 and Mills3), (ii) very small amounts of these materials are available (usually substantially less than 1 g) and (iii) analytical techniques are generally complex and time consuming. Before a systematic investigation of the fluorine, boron and chlorine distribution in tektites and meteorites was carried out, the analytical method described in this paper was developed. Modern analytical techniques, such as prompt gamma neutron activation analysis (PGNAA), inductively coupled plasma atomic emission spectrometry (ICP-AES), X-ray fluorescence (XRF) , atomic absorption spectrometry (AAS) , ion chromatography, and others, were unable to determine all three elements at the level of sensitivity required for the current application.Fluorine and chlorine cannot be deter- mined by PGNAA, nor can boron be measured by this technique to the accuracy and precision required, as the sample mass of 1-2 g that is routinely required for analysis (Shaw et al.4) was not available in this application. X-ray methods such as electron microprobe and XRF do not meet the required detection limits (Schafer and Medunas); in addition, insufficient amounts of the rock or sample powder were available for routine XRF analysis (Hahn-Weinheimer et a1.6). Interferences from iron and silica, and the mass of sample required for analysis precluded the determination of boron by ICP-AES (Owens et al.7 and Dins). The AAS technique was considered unsuitable as it had a low sensitivity for boron and there is a strong tendency for boron to form carbides.Several papers have been published, dealing with trace analytical procedures for single-element determinations, e.g. , Langer and Baumanng for fluorine and Aruscavage and for chlorine. In most of these instances however, a reagent containing boron or fluorine is used for sample decomposition (H3B03 or HF, respectively), which invali- dates the use of these techniques for the proposed method. Therefore, it was necessary to develop and apply an alterna- tive analytical technique, in order to fulfil completely the requirements of the proposed procedure. The analytical procedure described here offers the possi- bility of high accuracy and precision for the determination of all three elements at low concentration levels (all below 50 ppm) and with a minimum of sample consumption (<200 mg).Furthermore, uncertainties due to possible inhomogeneity in the distribution of elements are avoided as all the determina- tions are made on a single sample split. Experimental Apparatus For the sodium carbonate fusion an electric muffle furnace, suitable for continuous operation at 950 "C, is required. Polyethylene beakers (100 ml) and polytetrafluoroethylene separating funnels from Nalgene are used for boron extrac- tion. Fluorine is measured potentiometrically with an ion- selective fluoride combination electrode (Orion 96-09-00) with a digital pWmV-meter (Orion, Model 701A). Boron and chlorine are determined spectrophotometrically using a double-beam spectrophotometer (Perkin-Elmer, Model 552).Reagents The reagents used for the determination of each element are listed below the appropriate heading: Fluorine Buffer solution. Tris(hydroxymethy1)aminomethane (242 g) (No. 93349, Fluka) and 230 g of sodium tartrate (No. 6663, Merck) are dissolved in 800 ml of H20 and adjusted to pH 5.25 by the addition of approximately 125 ml of concentrated HCl Standard sodium fluoride solution. Standard solutions of 1 X 1O-3,1 X lO-4,l x lO-5,5 X 10-5 and 1 X 10-6 mol dm-3 sodium fluoride are prepared by dilution of 0.1 mol dm-3 sodium fluoride solution (Orion 96-09-06). (3638%). Chlorine Ammonium iron(II1) sulphate solution. Ammonium iron(m) sulphate (12.055 g) [FeNH4(S04)2.12H20] (No. 3776, Merck) is dissolved in 100 ml of 9 mol dm-3 HN03.Mercury(n) thiocyanate solution. Mercury(I1) thiocyanate (0.35 g) [Hg(SCN),] (No. 4484, Merck) is dissolved in 100 ml of methanol (95%) for approximately 12 h.66 ANALYST, JANUARY 1991, VOL. 116 Standard sodium chloride solution. Sodium chloride (0.8242 g) (dry) is dissolved in 1 1 of doubly distilled water. A 10 ml aliquot is diluted to 1 1 (CI- concentration = 10 pg ml-1). Boron 2-Ethylhexane-I,3-diol (EHD) solution. 2-Ethylhexane-l , 3-diol (200 ml) (No. 820032, Merck) is mixed with 800 ml of chloroform. Carmine reagent. Carminic acid (100 mg) (No. 21 1, Merck) is dissolved in 200 ml of concentrated H2S04 (95-97%). Boron standard solution. Boric acid (10.5716 g) (Suprapur, No. 765, Merck) is dissolved in 1 1 of distilled H20. A 5 ml aliquot is diluted to 500 ml with distilled H20 (boron concentration = 1 pg ml-1).Procedures Sodium carbonate fusion Sodium carbonate (0.9 g) and zinc oxide (0.1 g) are added to 150 mg of the powdered sample and transferred into a 25 ml Pt crucible, which is covered with a lid. The crucible is placed in a muffle furnace at 600 "C and the temperature increased to 950 "C. After 30 min at 950 "C, the crucible is removed and cooled to room temperature. The fusion cake is then dissolved and washed out of the crucible with 28 ml of dilute nitric acid (1 + 30) into a 100 ml polyethylene beaker. Three drops of ethanol are added to reduce any Mn that may be present. The dissolution should last about 12 h at 40 "C after which the pH is adjusted to between 5.5 and 6.0 before the solution is filtered into a 250 ml polyethylene bottle; the solid residue is rinsed with doubly distilled water.The volume of the solution should not exceed 100 ml. Blank solutions are treated in the same manner. Plastic gloves should be worn during the entire procedure in order to minimize contamination (particularly for the determination of chlorine). Fluorine determination Fluorine is determined potentiometrically . A buffer solution (20 ml) is added to 20 ml of the sample solution and stirred continuously, while the electrochemical potential is measured with an ion-selective fluoride electrode. Using the standard additions technique the concentration of fluorine in the sample can be calculated using the Nernst equation. Chlorine determination Chlorine is measured spectrophotometrically using the method described by Huang and Johns." A 20 ml sample solution, 2 ml of ammonium iron(1n) sulphate solution and 2 ml of mercury(I1) thiocyanate solution are mixed and diluted to 25 ml with doubly distilled water in a flask.The time for complete development of the iron(m)-thiocyanate complex is 10 min. Blank and calibration solutions are treated similarly. The absorbance is measured against water at a wavelength of Table 1 Fluorine, chlorine and boron concentrations in six international reference samples and corresponding data from Govindaraju13 Fluorine- Chlorine- Reference sample DTS-1 BIR-1 B HVO- 1 AGV-1 MAG-1 NIM-L RUIV DTS-1 BIR-1 BHVO-1 AGV-1 MAG- 1 NIM-L RUIV Boron- DTS-1 BIR-1 BHVO-1 AGV-1 MAG-1 NIM-L RlIIV Data from this work* Mean 16,16,18,18,21 43,44,47,50,50 365,372,373,381, 400 431,440,443,459, 460 756,775 3898,3992,4057, 4356 289,295,307,312, 315,315,323,327 25,28,31,33,35 24,27,27,31,36 96,99,107,108, 107,110,115,117, Not determined 1157,1190,1211, 200,207,216,220, 114 117 1289 223,225,23 1,240 10,11,11,13,14 <Detection limit 2,3,5,5,6 8,10, 10,11,12 126,134 18,18,19,21 10,21,11,11,12, 12,13,15 t P P 4 18 47 378 447 766 4076 310 30 29 105 113 1212 220 12 4 10 130 19 12 Standard deviation (PP@ 2.0 3.3 13.3 12.3 - 198.0 13.0 4.0 4.6 7.3 4.5 56.1 12.7 1.6 1.6 1.5 1.4 1.6 - Relative standard deviation (Yo ) 11.5 7.0 3.5 2.8 - 4.9 4.2 13.1 16.0 6.9 4.0 4.6 5.8 14.0 39.1 14.5 7.4 13.0 - Compiled data from reference 13 131- 447 3851: 4251: 7701: 44003: 10.5f 26 t 119$ 12003: 0.50s 0.339 2.57 7.87 1361: - * = Data in ascending order 1' = Information values.3 = Proposed values. 3 = Recommended values.ANALYST, JANUARY 1991, VOL. 116 67 460 nm. The chlorine concentration is calculated from a calibration graph. Boron determination The remaining sample solution (between 50 and 60 ml) is used for the spectrophotometric determination of boron following the method of Troll and Sauerer.12 Five drops of concentrated HCI and 20 ml of EHD are added to the sample solution. After intensive shaking of the mixture for 45 min, the EHD solution containing boron is separated from the aqueous sample solution. The solvent extraction is repeated, this time using only 10 ml of EHD solution. The re-extraction is carried out with 10 ml of 0.5 mol dm-3 NaOH.Blank and calibration solutions are treated similarly. For the determination of boron, 1 ml of the sample (NaOH), blank and standard solutions , respectively, is transferred into quartz beakers and two drops of concentrated HC1 are added to each. While cooling in a water-bath, 5 ml of concentrated HzS04 and 5 ml of carmine reagent are added to the solutions (caution, violent reaction!). After 1 h the carminic complex is completely developed, and the absorbance is measured at a wavelength of 610 nm. The boron concentration in the sample is calculated from a calibration graph. Results and Discussion The accuracy and precision of the proposed method were tested by the analysis of six international reference samples of silicate rocks: USGS (United States Geological Survey) DTS-1 (Dunite), BIR-1 (Basalt), AGV-1 (Andesite), BHVO-1 (Basalt), MAG-1 (Marine Mud) and NIM-L (Lujavrite) (obtained from the National Institute of Metallurgy, South Africa) and an internal reference sample (RlIIV, Volcanic Glass, Whalers Bay, Deception Island, Antarctica).The results are given in Table 1 and are in good agreement with the recommended values. Obviously the precisions (expressed as the relative standard deviations) decrease as concentrations fall towards detection limit levels. However, at higher concentrations, the following limiting precisions were measured: fluorine, 2% at >800 ppm of fluorine; chlorine, 6% at >lo0 ppm of chlorine; and boron, 4% at >lo0 ppm of boron. Typical total blank values were measured as 2 ppm for fluorine, 27 pprn for chlorine and 3 ppm for boron.The detection limits, calculated as 30 confidence levels, following the International Union of Pure and Applied Chemistry (1UPAC)l4 recommendations, were < 1 ppm flu- orine, 4 ppm chlorine and 1 ppm boron. These data refer to an analysis of 150 mg of sample powder. In order to achieve reliable results at low concentration levels, several conditions must be fulfilled. For fluorine, the readings of the electrochemical potential must be left to attain constant values for a time which depends on the fluoride concentration as follows: 30 min <1 x 10-5 mol dm-3 NaF, 10 min between 1 x 10-5 and 1 X 10-4 mol dm-3 NaF and 5 min >1 x 10-4 mol dm-3 NaF. (For further information on electrode analysis refer to Nicholson .15) The determination of chlorine is the most problematic.To avoid contamination, the analytical equipment must not be touched without gloves. In addition, the time span between mixing of the solutions and spectrophotometric measurements should not exceed 30 min; after this time, the colour of the iron(I1r)-thiocyanate complex begins to fade. This fading is due to the oxidation of thiocyanate by iron(II1) in daylight (Huang and Johnsl'). The limiting factor in the determination of boron is the impurity of the sodium carbonate used for the fusion (Troll and Sauerer'z). It is, therefore, necessary to analyse a blank with each batch of samples. The authors thank Dr. A. Sauerer and T. Dorfner for their helpful discussions and technical advice. 2 3 4 5 6 7 8 9 10 11 12 13 14 15 References Moore, C. B., Canepa, J. A., and Lewis, C. F., J. Non-Cryst. Solids, 1984, 67, 345. Matthies, D., and Koberl, Ch., Meteoritics, in the press. Mills, A. A., in Origin and Distribution of the Elements, ed. Ahrens, H., Pergamon Press, Oxford, 1968, p. 521. Shaw, D. M., Truscott, M. G., Gray, E. A., and Middleton, T. A., Can. J. Earth Sci., 1988, 25, 1485. Schafer, H. P., and Meduna, J. U., Fresenius 2. Anal. Chem., 1987, 326, 558. Hahn-Weinheimer, P., Hirner, A., and Weber-Diefenbach, K., Grundlagen und Praktische Anwendung der Rontgen- Jluoreszenzanalyse (RFA), Viehweg Verlag, Germany, 1984. Owens, J. W., Gladney, E. S., and Knab, D., Anal. Chim. Acta, 1982, 135, 169. Din, V. K., Anal. Chim. Acta, 1984, 159, 387. Langer, K., and Baumann, P., Fresenius Z. Anal. Chem., 1975, 277, 359. Aruscavage, P. J., and Campbell, E. Y., Talanta, 1983,30,745. Huang, W. H., and Johns, W. D., Anal. Chim. Acta, 1967.37, 508. Troll, G., and Sauerer, A., Analyst, 1985, 110, 283. Govindaraju, K., Geostand. Newsl., 1989, 13, 1. IUPAC (International Union of Pure and Applied Chemistry), Pure Appl. Chem., 1976,48, 127. Nicholson, K.. Chem. Geol., 1983. 38, 1. Paper 9105397B Received December 19th, 1989 Accepted September I4th, 1990

ISSN:0003-2654    

DOI:10.1039/AN9911600065

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, JANUARY 1991, VOL. 116 69 Enhancement of Precision and Accuracy in Derivative Spectrophotometry of Highly Absorbing Samples Isabel Dol, Moises Knochen and Carmen Altesor Universidad de la Republica, Facultad de Quimica, Catedra de Analisis Instrumental, Av. Gral. Flores 2 124, Casilla 1157, I1800 Montevideo, Uruguay The extension of relative absorptiometric methods to derivative spectrophotometry is proposed as a way of improving precision and accuracy for extremely high absorbance values. The concept was developed by applying the transmittance-ratio method to second-derivative measurements of two model systems where the total absorbance is kept very high, leading to a significant enhancement of precision and accuracy. The variation of the signal to noise ratio and of the total error as a function of absolute and relative absorbance was studied.Some of the factors affecting the analytical behaviour of this method are discussed. Keywords: Derivative spectrophotometry; precision; noise; transmittance-ratio method; high absorbances Derivative spectrophotometry1-6 is often used as a means of overcoming various types of additive interference in absorp- tion spectrophotometry and other spectrometric techniques. The number of analytical applications in different areas has increased considerably in the last few years, and many reports have appeared in the literature reflecting the widespread interest .7-17 The basis for its application is simple. Let A(h) be the mathematical expression of the absorbance of a given analyte as a function of wavelength, and let B(h) be the corresponding expression for an overlapped spectrum.Assuming the addi- tive property of absorbances, the following expression for the total absorbance, AT, can be written: AT(h) = A(h) + B(h) As the derivative is an additive operator it follows that: dnAT dnA dnB - +- dhn dhn dhn Hence, if a convenient order, n , of derivation can be found so that the second term of the right-hand side of equation (1) becomes negligible, the contribution of the interference to the total analytical signal will be cancelled out. The effective cancellation of the interference depends on the relative amplitudes of the two terms in the right-hand side of equation Despite its inherent simplicity, this technique is not trouble-free, and its application is not as straightforward as it might seem.The influence of such factors as height- and width-ratios of interfering to analyte bands should be taken into account when predicting the systematic and random errors to be expected in a given determination. This point was studied by O’Haver and Green18 by means of a model based on Gaussian bands. One of the aspects to take into consideration is the influence of noise. The so-called ‘analytical noise,’ i. e., the random variations generated in the various steps of sample processing, concerns all types of analytical determinations and will not be considered here. Instead we will focus on instrumental noise, the type of random variability originating in the spectropho- tometer itself. As it depends on the operating conditions, it is important to have some knowledge of the factors that affect it.Instrumental noise originates mainly in the photoelectric detector and in the electronic amplification. 1 9 ~ 0 Detector noise is dominated by shot, or quantum, noise, while electronic noise has a high proportion of Johnson noise. Detector noise depends to a large extent on the amount of light striking the photodetector surface. This in turn depends on such factors as the intensity of the light emitted from the (1). lamp, slit-width and the total absorbance of the cell plus analyte plus solvent and matrix. In digital instruments, where an analogue to digital ( N D ) converter is employed, this converter is also a source of variability. Raw data [transmittance (T), absorbance ( A ) , etc.] must first be obtained before they can be processed, hence it is important to ensure the best possible signal to noise (S/N) ratio during data acquisition in order to generate usable derivatives.It is well known from the theory of classical spectrophotometry that there are optimum ranges of trans- mittance, and hence of absorbance, where measurements with the lowest possible spectrophotometric error can be made, leading to the lowest analytical error .21,22 In those situations where the absorbance is either very high or very low, differential, or relative, methods23-26 can be used to improve the S/N ratio and therefore to lower the analytical error. These methods were developed mostly during the 1950s as a means of improving the precision of the measurement under conditions of high spectrophotometric error, i.e. , extremely high or low absorbances. However, with the advent of highly sophisticated spectrophotometers, usable spectro- photometric ranges have widened, and absorbances in excess of 2 can now be measured with a reasonable error, while the precision of low-absorbance measurements has also been considerably improved. Hence, precision methods have fallen into general disuse, and a few words should be said about them at this point. Three precision methods are described in the literature: transmittance-ratio, trace-analysis and ultimate-precision methods.26 When the absorbance is too high, the transmittance-ratio (or high-absorbances) method is employed. In this method, a solution with an absorbance lower than that of the sample, but higher than that of the blank, is used instead of the blank for making the 100% T (i.e., zero absorbance) adjustment, thus obtaining a ‘relative absorbance’ which is lower than the true absorbance.Usually, a dilute solution of the analyte is used for this purpose. This provides an easy way of improving the S/N ratio in those situations where a dilution cannot be carried out, such as the measurement of the absorbance of solid samples. Similarly, the trace-analysis (or low-absorbances) method consists of using a solution with a transmittance different from zero for the ‘dark-current’ (0% 7‘) adjustment. The 100% T adjustment is made as usual. The relative absorbance thus measured is higher than the actual absorbance. As suggested by its name, this method should be employed when the absorbance is very low, a situation often found in trace analysis.Finally, the ultimate-precision method involves a combina-70 ANALYST, JANUARY 1991, VOL. 116 tion of both of the above-mentioned procedures, and is supposed to achieve the best possible precision. Using any of these methods, a convenient expansion of the scale and a gain in precision of the measurement are attained. Further details of these methods are given in the Appendix at the end of this paper. Unfortunately, in most digital spectrophotometers the 0% T adjustment is performed automatically and is not accessible to the user, hence it is not possible to carry out the low-absorbances and ultimate-precision methods with this type of instrument.In derivative spectrophotometry on the other hand, one is often faced with the situation of trying to determine a given analyte in the presence of a highly absorbing matrix or extreme turbidity. The over-all shape of the spectrum of the matrix or turbid solution may be such that its interference on the analyte spectrum could be overcome by a derivative of a certain order, but, as the total absorbance becomes higher, the noise increases, and may become intolerable, thus precluding the use of the derivative method, which otherwise would be useful. In addition, one cannot resort simply to diluting the sample as this would lead to a loss of analytical signal (derivative amplitude) , a situation not desirable when the amplitude obtained is already small. In this paper, the use of an adaptation of the transmittance- ratio method in derivative absorption spectrophotometry is explored.Only this method was investigated, because the other two precision methods cannot be carried out with the instrument used for the reasons mentioned above. Measurement of High Absorbances in Derivative Spectrophotometry Adaptation of the transmittance-ratio method to derivative spectrophotometry warrants some comment. If the detector response is assumed to be linear with light intensity, the general equation for the relative transmittance in differential methods26 can be written The subscripts R , x and ref refer to relative, sample and reference, respectively, whereas the subscript o refers to the object used for the 0% T adjustment. ‘Reference’ is the equivalent of ‘blank,’ i.e., the solution or object used for the 100% T adjustment.For the transmittance-ratio method, To = 0, hence equation (2) becomes and hence (3) where AR, A, and Aref are the corresponding absorbances. Differentiating n times in the wavelength domain (4) In the original version of this method, a less concentrated solution of the same analyte was used as the reference to obtain the desired scale expansion. This procedure cannot be followed when working with derivative spectrophotometry which, as opposed to a simple absorbance measurement, is performed over a wavelength range instead of at a single wavelength. In classical transmittance-ratio measurements, only the transmittance (or absorbance) value of the reference solution is important, whereas in derivative spectropho- tometry its spectral shapes become of major concern.It is clear from equation (4) that the use of a dilute solution of the same analyte would result in a loss of analytical sensitivity. Hence, the absorption spectrum of the object used for compensation purposes in the reference beam should be such that its derivative equals zero at the wavelengths to be used for measurement. If the absorption spectrum is flat (i.e. , Aref (A) = k ) , this condition is met for all orders of derivation. In fact, this approach applies also to conventional zero-order relative measurements, a fact not often mentioned in the literature. Experimental Instrumentation A Shimadzu UV-240 double-beam, microprocessor-con- trolled spectrophotometer with an OPI-2 spectral processing unit (Shimadzu) was used.This instrument is fitted with an R-928 photomultiplier, and is capable of performing digital derivation of spectra either in real time or from data previously stored in random-access memory (RAM). In this work the real-time mode was used in order to save time. Other instrumental conditions were as follows: mode, 721 (second derivative, 1 nm wavelength increment); scan speed, fast; slit, 1 nm; wavelength scale, 3; wavelength range, 510-530 nm (didymium filter) or 240-280 nm (homatropine methylbro- mide); and TIA range, -0.25 to +0.25 (didymium filter) or -0.5 to +0.5 (homatropine methylbromide). Second-deriva- tive negative-peak values were printed out by means of the ‘peak pick’ function. A Shimadzu UV-160 spectrophotometer was also used for some complementary measurements.Solutions were measured in 1 cm silica cells. In order to simulate the analyte, a didymium-glass filter (Shimadzu No. 202-30242) was used. A variable-beam attenuator (Beckman No. 104 186) was employed. Other similar laboratory-built attenuators were also used. The spectral flatness of the attenuators was checked. These attenuators were installed in the path of the reference beam of the spectrophotometer. In order to be able to achieve the desired levels of total absorbance in one of the experiments, a slide filter consisting of a variable number of layers of wire mesh (17 wires per centimetre; diameter, 0.4 mm) was constructed. The spectral behaviour of the mesh was checked and it was found to have a virtually flat spectrum within the wavelength range of interest, the absorbance of a single layer being about 0.4.Its second-derivative spectrum followed the zero line. Henceforth, this device will be referred to as the neutral filter. Reagents Homatropine methylbromide (HMB) of pharmaceutical grade, and analytical-reagent grade barium sulphate and glycerine were used. Methods Two different types of experiment were carried out. In the first, the spectrum of a didymium filter (Fig. 1) was scanned in the visible range (510-530 nm) and the second derivative was obtained under a number of different conditions of total and relative absorbance (Fig. 2). The variable neutral filter was used to simulate the ‘additive interference.’ When the didymium filter was placed in the sample-cell holder, the neutral filter was affixed next to it in the light path, and the number of layers of mesh was varied from 0 to 7 in order to obtain values of total absorbance in the range 0.4-3.7.The variable-beam attenuator was set in each instance to obtain a given relative absorbance. One hundred and seven different combinations of total and relative absorbance were investigated. The amplitude of the second-derivative spec- trum of the didymium filter was measured at the minimum at about 525 nm, corresponding to a shoulder in the zero-order spectrum.ANALYST, JANUARY 1991, VOL. 116 71 In order to simulate more closely an actual analytical application, in a different experiment, the second-derivative spectrum of a solution of HMB in water-glycerine (1 + 1) (625 mg 1-1) was scanned in the ultraviolet (UV) range (240-280 nm) in the presence of a turbidity obtained by suspending various amounts of barium sulphate (1.2-6.0 g 1-1).The glycerine was added to stabilize the suspension. Forty different combinations of total and relative absorbance were investigated, the total absorbance level being in the range Measurements were made at the minimum at about 258 nm, which corresponds to the main maximum of the HMB absorption band (Figs. 3 and 4). This model system can be considered to be representative of many real-world situations, because of the spectral region where the benzenoid absorption bands of HMB occur, and because of the presence of an additive interference which does not present a flat spectrum.1.8-3.7. 3.00 a, C ; 2.00 s s 1 .oo A B 4 51 0 520 530 hlnm Fig. 1 Absorption spectra of A, the didymium glass + neutral filter; B , didymium glass + neutral filter, with reference-beam attenuation; and C, didymium glass. The didymium glass was used as a simulated analyte, while the additive spectral interference was simulated by means of a neutral filter consisting of seven layers of wire mesh (see text) +0.25 -0.25 a ) In both experiments, the spectral baseline was memorized previously using air as the reference. Measurements were made with various degrees of optical compensation by placing the variable-beam attenuator in the reference-cell holder. Ten second-derivative scans were then made and the average, standard deviation, S/N ratio, relative bias and total error27 were calculated as follows: average standard deviation SIN ratio = Bias = average - true value bias Relative bias (%) = true value x 100 /bias/ + 2 (standard deviation) true value Total error (%) = x 100 where ‘true value’ refers to the corresponding measurement performed without interference.b) 2.50 2.00 8 1.50 m e s n Q 1.00 0.50 0 240 260 280 ?Jn m Fig. 3 Absorption spectra of A, 625 mg 1-I aqueous HMB solution + BaS04 suspension, uncompensated; B, HMB + BaS04 suspension, with optical compensation; and C, HMB without interference I 510 520 530 510 520 530 510 520 530 h/nm Fig. 2 Second-derivative spectra of (a) didymium glass + neutral filter; (b) didymium glass + neutral filter, with reference-beam attenuation; and (c) didymium glass. The didymium glass was used as a simulated analyte, while the additive spectral interference was simulated by means of a neutral filter consisting of seven layers of wire mesh (see text)72 ANALYST, JANUARY 1991, VOL.116 -0.50 240 260 280 240 260 280 hln m 240 260 280 Fig. 4 suspension, with optical compensation; and ( c ) HMB solution without interference Second-derivative spectra of ( a ) aqueous HMB solution (625 mg 1-1) + BaS04 suspension, uncompensated; ( 6 ) HMB + BaS04 Table 1 Variation of relative bias, S/N ratio and coefficient of variation (CV) with relative absorbance (AR) for the didymium-glass spectrum (amplitude of the second derivative at 525 nm), at a total absorbance level of 3.7 AR* Bias (YO) S/N ratio cv (Yo j 0.2 2.2 0.8 1.3 1.1 2.9 1.5 3.5 1.8 1.2 2.2 4.9 2.4 8.2 2.7 16.0 3.1 29.2 * Measured at 525 nm.16.9 18.3 23.0 41.4 30.1 18.3 12.0 7.1 6.6 5.91 5.47 4.35 2.42 3.32 5.45 8.35 14.10 15.23 Table 2 Variation of relative bias, S/N ratio and coefficient of variation (CV) with relative absorbance ( A R ) for the HMB spectrum (ampli- tude of the second derivative at 258 nm), at a total absorbance level of 3.7 AR* Bias (Yo) S/N ratio cv (Yo) 0.4 -1.39 30.5 3.3 0.6 2.59 12.6 7.9 1.1 3.02 9.9 10.1 1.6 -0.88 18.4 5.4 2.1 2.94 12.8 7.8 2.5 6.74 8.8 11.3 2.9 2.85 5.5 18.1 * Measured at 258 nm. Results and Discussion The evaluation of the experimental data demonstrates that a significant reduction in noise and analytical bias in the derivative can be attained by using differential techniques. Tables 1 and 2 exemplify the S/N ratio and bias values found for high total absorbances, at different levels of relative absorbance (measured at the same wavelengths as the second derivative), for the two experiments.As a demonstration of the gain in precision, during the measurements with the didymium glass, at a total absorbance level of 3.7 (at 525 nm), no usable data could be acquired from the conventional derivative method, as the noise in the second derivative was so high that the peak-finding algorithm of the instrument became inoperative. However, through the use of differential measurements, a value for the S/N ratio of 41, equivalent to an instrumental coefficient of variation of 2.4%, could be attained at a relative absorbance level of 1.5. 120 A L I I 0 1 .o 2.0 AR Fig.5 Variation of S/N ratio as a function of relative absorbance (AR) for the second derivative at 525 nm (didymium glass), at total absorbance levels of A, 1.6; B, 2.5; C, 2.8; D, 3.0; and E, 3.7 80 0 .- c : 60 0 0 - 40 m C u) v) .- .- 20 0 1 .o 2.0 3.0 AR Fig. 6 Variation of S/N ratio as a function of relative absorbance (AR) for the second derivative at 258 nm (HMB), at total absorbance levels of A, 1.8; B, 2.4; and C, 3.7 Signal to noise ratio enhancements ranging from 2.5 to more than 13-fold were observed in this work. The enhancement to be expected depends on such factors as the wavelength region, slit-width and the intensity of the light reaching the detector. In addition, it will also depend on the spectral shape. At high absorbance levels, the absorbance range spanned by a given absorption band may correspond to a variation of several decades in % T , and hence in light intensity, because of the logarithmic nature of absorbance.Therefore, the S/N ratio at the wavelength of the maximum may be considerably worseANALYST, JANUARY 1991, VOL. 116 73 60 1 s I L 0 40 - m U 20 0 Fig. 7 Variation of total error as a function of relative absorbance (AR) for the second derivative at 525 nm (didymium glass), at total absorbance levels of +, 2.5; X , 2.7; 0, 2.8; 0, 3.0; and 0, 3.7 1 .o 2.0 3.0 AR Fig. 8 Variation of total error as a function of relative absorbance (AR) for the second derivative at 258 nm (HMB), at total absorbance levels of 0, 1.8; X, 2.0; +, 2.4; and 0, 3.7 than at other wavelengths. The absorbance values reported here correspond to the wavelengths where the derivative measurements were carried out in each instance.Figs. 5 and 6 are plots of S/N ratio versus relative absorbance at various levels of total absorbance for the didymium glass and HMB, respectively, while Figs. 7 and 8 show similar plots for the total error as defined by McFarren et aL.27 From these plots, the existence of optimum ranges of relative absorbance can be observed. It could be stated tentatively that for the present instrument, the optimum S/N ratio is obtained for relative absorbances of about 1. This could be expected as the error pattern of the derivative follows that of the original data. The total error increases considerably at absorbances above 2, as a consequence of the increase in both noise and bias.The above results show that an increasing positive analytical bias appears as relative absorbances become higher, following the same over-all trend as the noise. At high values of total absorbance, the bias could be diminished by applying the transmittance-ratio method to lower the relative absorbance to values of about 1. This phenomenon, which attracted our attention, could not be explained on the basis of the considerations put forward by O’Haver and Green.18 As it was not clear whether this behaviour could be general or specific to the particular instrument being used, some complementary measurements were carried out using a different spectropho- tometer (Shimadzu UV-160), which is fitted with photodiodes instead of a photomultiplier.Despite the fact that the number of measurements made was not as high as with the UV-240 instrument, a similar trend was found for the total error. This bias could probably be assigned to the influence of stray light, although this requires further investigation. Digital smoothing is often used to improve the S/N ratio in spectrophotometric data, hence it is valid to compare both methods. Digital smoothing is usually performed on the data previously stored, and depending on the algorithm employed it may involve a compromise between precision and analytical sensitivity. If the noise level is very high, such as with the measurements carried out at absorbances of 3 and above, the degree of smoothing required would be so high that an appreciable loss of spectral resolution and analytical sensi- tivity would occur.Similar reasoning could be applied to other procedures used for lowering the noise level, such as the use of wider slits, as this would result in a loss of spectral detail and hence of derivative amplitude in most instances. On the other hand, optical compensation using the transmittance-ratio method does not cause any degradation, either in the signal amplitude or in the resolution attained. In fact, the data collected in this work show that under some circumstances it can make the difference between usable information and no information at all. In any event, spectrophotometric compen- sation and data smoothing can be considered as complemen- tary strategies aiming at the same objective. Baseline correction was made with air as the reference. As with most digital spectrophotometers, the instrument used in this work memorizes the spectral baseline in the working wavelength range.These data are stored in RAM and later subtracted automatically from the absorbance values measured in the sample position to obtain the true absorbance value, which is displayed subsequently. Therefore, it is not always necessary to use a cell with the solvent or blank in the path of the reference beam. Instead, one can place a cell with solvent (or blank) only in the path of the sample beam during the process of baseline memorizing, which is then made with air as the reference. This procedure wu3 followed in both experiments, because, as the variable-beam attenuator had to be installed in the reference-cell position during measure- ments, there was no possibility of simultaneously placing a reference cell when measuring the aqueous solutions.The didymium-glass filter is usually measured against air, hence this makes no difference. It should be pointed out that the procedure explained previously for baseline memorizing may not be adequate in certain instances, when the dynamic range of the instrument may be exceeded. The choice of the second derivative for this investigation was based on the spectrum of the neutral filter, which was flat, and of the barium sulphate suspension, which displayed a small slope within the wavelength range scanned. The corresponding second-derivative spectra followed the base- line. In addition, the second derivative is perhaps the most widely used derivative in analytical applications.Further, considerations of error propagation allow an extrapolation to derivatives of other orders to be made, because the noise reduction attained in the derivative originates in the gain in precision in the acquisition of the absorption spectra. Conclusions The transmittance-ratio method may be used for enhanced precision and accuracy in derivative spectrophotometry in situations of extremely high absorbances. The method is no more complex than the ordinary method, and it provides a significant reduction in noise and bias without loss of spectral resolution or analytical sensitivity. This technique can be used advantageously in the presence of a highly absorbing interference or intense light scattering, hence the field of application of derivative spectrophotometry is widened considerably.74 ANALYST, JANUARY 1991, VOL. 116 Although the proposed methodology was applied only to model systems, its utility in actual analytical situations is obvious, and some applications are currently being developed in our laboratory.The existence of optimum ranges of relative absorbance, which could be expected from theory, was also demonstrated. As to the limitations, the two most important limitations are detector sensitivity and stray light. When working with absorbances of 3, it should be remembered that this is equivalent to a transmittance of 0.1%. If an attenuator is placed in the reference beam, the amount of light reaching the detector through this path may be very small.For photomulti- pliers, this means that the maximum voltage available may be applied to the dynodes and still not be sufficient to reach the 100% T (zero absorbance) level. Stray light may cause a severe distortion of the spectra, thus rendering the measure- ments obtained under these conditions useless. The first of these problems was not found during this work, while stray light could be responsible for the bias effects noted at high absorbances; however, the quality and performance (state of the lamps, photodetector, etc.) of a given instrument will determine whether or not these problems will be of concern. In this work only the method for high absorbances was investigated. Extension of these concepts to the other two precision methods requires further investigation, and this is currently under study.Appendix For the benefit of those not familiar with relative absorptio- metric methods, a brief explanation of the properties of these methods, with the emphasis on the transmittance-ratio method, is presented below. Let us assume that the response, R , of a spectrophotometer is a linear function of the power, P, of the light falling on its detector. Then R = kP + k’ where k is the sensitivity of the instrument and k’ is an additive constant adjustable by means of the ‘dark current’ control. The zero adjustment is made by interposing an absorbing system (cuvette, shutter, etc.) such that the light power reaching the detector is Po, while the full-scale (usually 100% 7‘) adjustment is made with a cuvette containing blank or solvent, such that the light power falling on the detector is PB.If this adjustment is made with an absorbing system rather than the blank or solvent, the power reaching the detector will be Pref. When a given sample is measured, the light power reaching the detector will be P,. Obviously, Po < P, < Pref < PB. The respective instrumental responses, R , will then be R, = 0 = kPo + k’ * k’ = -kP, Rref = 1 = kPref + k’ = k(Pre, - Po) R, = kP, + k’ = k(Px - Po) The ‘relative transmittance’ can then be defined as If we divide all the powers by PB, the respective trans- mittances are substituted in place of the powers and hence T, = TRTref - TRT, + To (A2) In the ‘ordinary method’ the zero adjustment is made with a shutter interposed in the light path, and the full-scale response is adjusted with blank or solvent, hence To = 0 (0% T ) and Tre* = 1 (100% 7‘); therefore, equation ( A l ) simplifies to the usual form, T, = TR When using the high-absorbances method, To = 0 and Tref f l , hence Tx = T~Tref In the low-absorbances method, the zero adjustment is not made with a shutter, but rather with a system whose transmittance, To, is different from 0, and the full-scale adjustment is made with a blank or solvent T, = TR - TRT, + To In the ultimate-precision method, both the 0 and 100% T points are adjusted with solutions, and equation (A2) fully applies.On the other hand, according to the Bouguer-Beer law, --log T, = A , = abc where a is the absorptivity, b the length of the light path and c the concentration of the analyte, and, considering equation (A2) -log T, = -log(TRTref - TRT, + To) = A , = U ~ C (A3) For the estimation of the error, this equation will be differentiated, taking into account that To and Tref are cons tan t Dividing by equation (A3) and considering the approximation of increments to be sufficiently small, or which is the same.equation applies, substituting AAIA in place of AcIc. If the relative error in absorbances is sought, the same If the S/N ratio is preferred, we have the reciprocal Txlog Tx 1 S/N ratio = 0.4343(Tr,f - To) ATR From equations (A5) and (A6) the corresponding analytical errors can be calculated. These are the usual forms for the expression of the relative error of precision methods. In the older literature, the gain in precision attained is assigned to the scale expansion and hence to the lower reading error. Ingle28 has correctly pointed out that this is only part of the truth, as other sources of uncertainty are present.The enhancement of precision however will depend on which of the noise sources will be significant in a given instrument. If the development of theoretical equations is required, the noise characteristics of the detector should be known, something that is not always available. Classically, error equations in spectrophotometry have been calculated assuming a given behaviour for the depen- dence of the spectrophotometric error, AT, on T . This in turn depends mainly on the nature of the photodetector. For quantum detectors such as those used in UV-visible spectro- photometry, it has been postulated that AT can be considered to be proportional to T+ or some more complex form such as (F + T ) wANALYST, JANUARY 1991, VOL.116 75 l 2 tY 0 1 2 3 4 AR Fig. 9 Theoretical variation of S/N ratio as a function of relative absorbance (AR) assuming a dependence of the form K‘P between spectrophotometric noise and transmittance [equation (AlO)], for the transmittance-ratio method, at values of Aref of: A , 1.3; B, 1.0; C, 0.7; D, 0.4; E, 0.2; F, 0.1; and G, 0 As an example, the following equation can be derived from equation (A7) for the fl dependence: where K is a proportionality constant. In the high-absorbances method To = 0, hence equation (A8) becomes (A9) T log T, - (S/N ratio)’ - 0.4343K(S/N ratio) = Tref (TRY (S/N ratio)’ = (TR)~ log (TRTref) Fig. 9 shows the plot of S/N ratio as a function of AR for the transmittance-ratio method, for different values of Aref, assuming a simple dependence for the spectrophotometric error of the form Ti.It can be seen that the error can be diminished by application of the precision method, to an extent that depends on the values of Aref. However, these types of error equation, although useful, are not rigorous, as they depend on certain assumptions about the error that are not necessarily true, as shown by Youmans and Brown,22 who, for this reason, preferred to develop error equations based only on statistical processing of experimental results. Their considerations were further developed by Bense and D o P for application to the precision methods.The authors thank Tomas Bense, of Compafiiia Industrial de Tabacos Monte-Paz S.A., for permission to use the UV-160 spectrophotometer. 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 26 27 28 29 References Morrey, J. R., Anal. Chem., 1968, 40, 905. O’Haver, T. C., Anal. Chem., 1979,51, 91A. Talsky, G., Mayring, L., and Kreuzer, H., Angew. Chem. Int. Ed. Engl., 1978, 17, 785. Cahill. J . E., Am. Lab., 1979, 11, 79. Traveset, R . , Such, V., Gonzalo, R., and Gelpi, E., J. Pharm. Sci., 1980, 69, 629. Levillain, P., and Fompeydie, D., Analusis, 1986, 14, 1. Fell, A. F., Jarvie, D. R., and Stewart, M. J.. Clin. Chem., 1981, 27, 286. Davidson, A. G., and Hassan, S. M., J . Pharm. Sci., 1984, 73, 413. Fasanmade, A. A., and Fell, A. F., Analyst, 1985, 110, 1117. Lawrence, A. H., and Kovar, J., Analyst, 1985, 110, 827. Meal, L., Anal. Chem., 1986, 58, 834. Knochen, M., Bardanca, M., and Piaggio, P., Boll. Chim. Farm., 1987, 126,294. Shijo, Y., Nakaji, K., and Shimizu, T., Analyst, 1988, 113,519. Morelli, B., Analyst, 1988, 113, 1077. Dol, I . , and Altesor, C., SAFYBI (Buenos Aires), 1988, 28, 2660. Dol, I . , Knochen, M., and Altesor, C., Boll. Chim. Farm., 1989. 128, 18. Knochen, M., Altesor, C., and Dol, I., Analyst, 1989, 114, 1303. O’Haver, T. C., and Green, G. L., Anal. Chem., 1976,48,312. Ingle, J . D., and Crouch, S. R., Anal. Chem., 1972,44, 1375. Rothman, L. D., Crouch, S. R., and Ingle, J. D., Anal. Chem., 1975,47, 1226. Ewing, G. W., Instrumental Methods of Analysis, McGraw-Hill Kogakusha, Tokyo, 4th edn., 1975, p. 42. Youmans, H. L., and Brown, V. H . , Anal. Chem., 1976, 48, 1152. Skoog, D. A., and West, D. M., Analisis Instrumental, Interamericana, Mexico City, 1st edn., 1975, p. 92. Hiskey, C. F., Anal. Chem., 1949, 21, 1440. Hiskey, C. F., and Firestone, D., Anal. Chem., 1952, 24, 342. Reilley, C. N., and Crawford, C. M., Anal. Chem., 1955, 27, 716. McFarren, E. F., Lishka, R. J., and Parker, J. H., Anal. Chem., 1970,42, 358. Ingle, J. D., Anal. Chem., 1973, 45. 861. Bense, T . , and Dol, I., Estirnacion del Error Analitico en Espectrofotometria de Precision, XI11 Latin American Con- gress of Chemistry, Lima, Peru, 1978. Paper 0/00788I Received February 20th, 1990 Accepted July 30th, 1990

ISSN:0003-2654    

DOI:10.1039/AN9911600069

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, JANUARY 1991, VOL. 116 77 Sensitive Colorimetric Determination of Ammonium Ion in Water by Laser Photothermal Detection Eugen Strauss Department of Physics and Astronomy, University of Georgia, Athens, GA 30605, USA Jan-Paul Favier, Dane Bicanic” and Kees van Asselt Department of Agricultural Engineering and Physics, Agricultural University, Duivendaal I , 670 I AP Wag en ing en, Th e Netherlands Marcel Lubbers Department of Soil Science and Geology, Agricultural University, Duivendaal I , 670 I AB Wageningen, The Netherlands Photothermal beam deflection was combined with the established and proven indophenol colorimetric method for the measurement of trace amounts of ammonium ion in water. A 7 mW 632 nm He-Ne laser and a 4 mW 780 nm Al-Ga-As semiconductor laser provided a limit of detection below 0.25 mmol m-3.Keywords: Ammonium determination; trace detection; photothermal method; aluminium-gallium-arsenic diode laser; helium-neon laser Ammonia produced in large amounts from organic nitrogen compounds ( e . g . , proteins, amino acids and urea) by reductive chemical reactions and under anaerobic conditions by micro- biological processes plays an active part in the chemistry of soils. A considerable fraction is also released into the air; its characteristic odour does not go unnoticed in areas with a livestock industry under most meteorological conditions. Ammonia contributes significantly to various atmospheric processes and to water pollution.1 In particular, the ammo- nium ion dissolved in rain-water is the main base contributing to the neutralization of atmospheric acidity and is eventually returned to the ground with precipitation.Water bodies are also directly polluted by ammonia and nitrate because of the excessive use of nitrogen fertilizers in agriculture and from the discharge of sewage treatment plants. Ammonium is found at increasing concentrations in surface- and ground-water and in drinking water, which is a source of increasing public concern. As for other pollutants, sensitive, fast and low-cost analy- tical methods for the determination of both ammonia and ammonium ion are required for the control of pollution. The same is true for the investigation of ammonia deposition, propagation and decomposition in the environment and its consumption by plants and micro-organisms and by chemical processes.Various detection methods and techniques are presently known; of these, the classical colorimetric (both manual and automated) methods for the determination of ammonia and ammonium ion in sewage effluents, and in raw and potable waters have received much attention. After Nessler’s reaction,2 the methods based on the spectropho- tometric measurement of the absorbance of the coloured complexes formed (related to Indophenol Blue)3,4 are the most well established. In particular, the indophenol reaction is the standard method for the examination of water recommen- ded for example by the Environmental Protection Agency (EPA)5 and Deutsche Industrie Norm (DIN).6 The indophe- no1 method is recommended for use at concentration levels ranging from 2 to 70 mmol m-3, but the practical detection limit reported is close to 0.6 mmol m-3.4-6 Alternative methods include titration, ion-selective electrodes and spec- trofluorimetric detection.Under controlled conditions a gas-selective ammonium electrode may reach a limit of detection of < l O mmol m-3, if the non-linear part of the response curve is included.’ Spectrofluorimetric detection * To whom correspondence should be addressed. involves a reaction with o-phthalaldehydes and fluorescence at about 430 nm; it has been demonstrated to be particularly useful with ion chromatography9 and shown10 to be sensitive to <1 mmol m-3. Of the methods for the detection of ammonium, standard indophenol colorimetry appears to be advantageous for most of the routine applications mentioned above.It requires the least sophisticated instrumentation and is relatively cheap and simple in its chemistry. Furthermore, the possible interference of various salts and other substances frequently present in water has been studied and documented.4.5 Extending the range of reliable detection for indophenol colorimetry to below the 0.25 mmol m-3 level using a simple instrument suitable for field work would be useful in soil science studies and in agricultural work, hence the present investigation of the potential of laser-based photothermal detection in indophenol colorimetry. Photothermal laser spectroscopy is considered here because it is particularly suitable for the measurement of small absorptions, mainly because the signal is zero with no absorption and in addition weak turbidity and scattering do not interfere.Recently, the suitability of this technique for the detection of <0.1 mmol m-3 of orthophosphate in Molyb- denum Blue colorimetry has been shown.’’ Here, some details of photothermal detection are presented and a reliable detection method for ammonium in water at concentration levels below 0.25 mmol m-3 is described. Experimental Photothermal Absorption Measurement Laser-based photothermal detection techniques have been employed in recent years for a wide range of applications, in particular they have been shown to be capable of ultrasensitive trace detection. 12 Photothermal techniques are based on detection of the various effects of heat released in a sample as a result of the absorption of intensity-modulated radiation.Photoacoustics, a technical variant that measures, either by a gas microphone or a piezo-transducer, the acoustic wave generated by thermal expansion, is commonly known and various other photothermal effects can be employed for detection. The photothermal effect on the optical index of refraction has proved to be particularly useful.13 With laser sources it is experimentally advantageous to measure the gradient of the refractive index rather than the refractive index itself. A technical variant called ‘collinear’ photothermal78 ANALYST, JANUARY 1991, VOL. 116 beam deflection spectroscopy14 (BDS) is obviously particu- larly suitable for the detection of small absorptions in transparent liquid samples at a fixed wavelength because the entire path length in the sample contributes to the signal and residual small air bubbles do not interfere.In BDS two different laser sources are used. The wavelength of one laser (pump laser) must, as closely as possible, match the absorp- tion wavelength of the species to be detected. The energy absorbed from the periodically modulated pump laser beam heats, optically, the excited region of the sample and produces a time varying spatial profile of the refractive index that acts as a thin diverging 'thermal lens.' A second laser (probe laser), if possible emitting at a different wavelength that is negligibly absorbed by the sample, is used to detect the refractive index gradient. The thermal lens deflects the probe beam and the a.c.deflection is observed at the modulation frequency (with no absorption in the absence of the substance to be measured) using a synchronous lock-in detection technique. Apparatus With indophenol colorimetry, ammonium is determined as an emerald green complex which is formed by the reaction of ammonia at pH 12.8-13.0 with hypochloride and salicylate in the presence of sodium nitroprusside as a catalyst. The reagents were prepared following standard procedures:6 standard NH4+ solutions for calibration were prepared from (NH4)2S04 [Merck, pro analysi (p.a.)]; sodium phosphate buffer (Merck, p.a.) using a nitrogen-free NaOH solution (Baker) and potassium sodium tartrate solution (Merck, p.a.); the sodium salicylate-sodium nitroprusside (Merck, p.a.) solution was kept in the dark and the sodium hypo- chlorite (BDH) solution was prepared fresh daily.The indophenol complex has an absorption band between 600 and 700 nm, with a maximum at 650 nm. This band closely matches the red line of the He-Ne laser at 632.8 nm (Fig. 1). A readily available, moderate power (7 mW) laser suffices as the pump laser for the sensitivity required. A cheap, low power Al-Ga-As semiconductor laser (4 mW, 780 nm) equipped with a collimation lens was used as the probe beam source. At 780 nm the absorption of the complex and of the reagents is small and water absorption is negligible. Both beams have Gaussian profiles and are aligned to intersect in the middle of the sample cuvette at a small angle (about 3") in order to maximize their spatial overlap (Fig.2). Two suitable lenses 0.6 0, 6 0.4 e s 2 0.2 500 650 800 Fig. 1 A, The absorbance of the indophenol complex recorded with a 10 mmol m-3 ammonium solution in the 1 ern sample cell; and B, the residual absorption of the blank reagents solution (multiplied by a factor of ?to enable distinction from the x-axis). The arrow indicates the wavelength of the He-Ne pump laser Unm focus the beams into the cuvette; the probe beam to a tighter focus (l/e diameter 50 pm; where e is the base of the natural logarithm) than the pump beam (l/e diameter 150 pm). The probe beam is displaced vertically from the pump beam such that it intersects with the thermal lens in the region of the maximum gradient of the refractive index. This relative displacement is easily aligned by adjusting (with microposi- tioners) the lens which focuses the pump laser. The deflection of the probe beam by the thermal lens was measured by use of a quadrant position sensor made of silicon diodes (Centronix DQ-50).A 780 nm interference filter was used to block the light from the pump beam which was scattered by the sample and cuvette windows. The photothermal effect was measured in the frequency domain. The intensity of the pump laser was mechanically chopped at 14 Hz. The deflection signal was then modulated at that frequency and fed to a standard lock-in amplifier for de-modulation. Transmission at the He-Ne wavelength through the cuvette for the range of ammonium concentration of interest was 30.5, i . e . , the sample was 'optically thin.' The deflection signal is proportional to the absorption coefficient, i.e., the indophenol complex concentration , provided the d.c. temper- ature (only the modulating component of the temperature is being detected) power of both laser powers and the modula- tion frequency are constant. It was noted that the residual absorption of the probe beam in the sample did not distort the measured deflection signal when it was normalized to the power intensity reaching the position sensor. The reagents had a small residual absorption at the 632 nm pump wavelength (Fig. l), which contributed to the deflection measured. It is therefore necessary in practice to subtract the signal generated by a blank reference sample prepared with ammonium-free water from the same reagents as the sample.The relatively simple experimental set-up was assembled from standard mounts (provided with a magnetic base) and optical components available in a spectroscopy laboratory and arranged on a laboratory-built vibration-damped table. The deflection is typically of the order of 1 x 10-3" at the concentration level of 2-3 mmol m-3 and the displacement at the detector is about 1 pm, i.e. , the measurement is prone to noise from mechanical vibrations. However, no special vibrational insulation is necessary if a single sturdy bench carrying all the components is used. Standard solutions and samples could be pipetted directly into a 10 mm photometer cuvette without removing the cuvette from the set-up. Fluid turbulences and rising air bubbles disappeared almost immed- iately.A lock-in amplifier integration time constant of 10 s was sufficient. Several successive amplifier readings were averaged and the statistical deviations calculated. Results and Discussion Determination of Ammonium in Standard Solutions The sensitivity of photothermal indophenol colorimetry was tested with several dilution series prepared from the standard NH4+ solution (see above). A low concentration (below 20 PrB(780 nm) Ch sc IF Fig. 2 Experimental set-up for photothermal beam deflection detection of ammonium. C, Sample cuvette; Ch. mechanical chopper; PSD, position-sensitive detector; PrB, probe beam; PUB, pump beam; SC, sample cuvette; L1 and L2, lenses; M, plane mirror; and IF, interference filterANALYST. JANUARY 1991, VOL. 116 79 I I I I I 5 10 15 20 25 30 Ammonium co ncentrati o n/m mo I m - 3 Fig.3 concentration in photothermally detected indophenol colorimetry Variation of the signal as a function of the ammonium mmol m-3) in the range of interest was chosen for all further studies. At higher concentrations, a conventional spectro- photometer can be used. However, such samples are also photothermally accessible using dilution or a thinner cuvette to ensure that the sample is optically thin. The result of five replicate runs is depicted in Fig. 3. The reproducibility of the absolute values of deflection was good and limited only by the mechanical stability of the fairly simple experimental set-up. The net deflection signal, i.e., the difference between the signal measured for the sample and the blank reagents solution, is linear in the concentration range 0.25-25 mmol m-3.This detection limit is not restricted by the signal to noise ratio of the deflection signal but rather by the residual absorption near 650 nm observed in the blank solution (the linear absorption coefficient was found to be about 0.02 cm-I), see Fig. 1. A typical value for the absorption of the blank reported in the literature is 0.06 A for a cell 40 mm in length .4 The indophenol complex has a comparable linear absorption coefficient at a concentration of 0.25 mmol m-3. Hence, the residual absorption of the blank prevents the full utilization of the potential of photothermal detection in the determination of ammonium. The band shape and position of the residual absorption in the red region are about the same as those of the indophenol complex so that the use of a different pump laser wavelength would not alleviate the problem.This similarity may indicate the presence of ammonium impurities in the stock solutions. Conclusion The investigation shows that photothermal detection is a very suitable technique for the sensitive colorimetric determination of ammonium in water. Taking into account the interfering residual absorption, the data obtained imply an improvement in the practical detection limit (<0.25 mmol m-3) when compared to that attainable6 with standard spectrophotomet- ers. An important advantage of this technique is that it is not affected by the slight turbidity present in most ‘real’ samples. Furthermore, the set-up is basically very simple and construc- ted from readily available and inexpensive components.A compact instrument15 suitable for routine laboratory and field-work based on beam deflection detection is feasible. The design could include a 780 nm Al-Ga-As probe laser and a 670 nm semiconductor pump laser instead of an He-Ne laser. A red region semiconductor laser is also absorbed by the indophenol complex and can be expected to become a low-cost item in the near future. The instrument could also be used for phosphate colorimetry by simply interchanging the pump and probe beams of the two lasers.” Colorimetric determination of ammonium in water by means of photothermal detection has been shown to be a sensitive and reliable technique with a practical detection limit of C0.25 mmol m-3.Recent developments in the technology of inexpensive diode lasers16 indicate their potential for practical and analytical work, in particular for trace detection, in the near future. We thank staff members of the electronic and mechanical workshop, R. Baldew and K. Rijpma of the local university, and Dr. Nick van Hulst, Technical University of Twente, for their support. A grant from the FOM-STW foundation in Utrecht for E. S. is gratefully acknowledged. I 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 References Apostolache, S . , Rev. Chim. (Bucharest), 1962, 13, 615. Scheiner, D., Water Res., 1979, 10, 31. Standard Methods for the Examination of Water and Wastewater, American Public Health Association, Washington, DC, 1985. Ammonia in Waters, HM Stationery Office, Linneys of Mans- field, London, 1981. Deutsche Einheitsverfahren zur Wasser, Abwasser und Schlam- muntersuchung (procedure DIN 38 406-E5-1 and ISO/DIS 7150), VCH Verlagsgesellschaft, Weinheim, 1988. Degner, F., Honold, F . , and Putzien, J . , Vom Wasser, 1987,69, 33, Roth, M., Anal. Chem., 1971, 43, 880. Merz, W., and Oldeweme, J . , Vum Wasser, 1987, 69. 95. Zhang, G., and Dasgupta, P. K., Anal. Chem.. 1989, 61,408. Bicanic, D. D., Kunze, W., Sauren, H . , Jalink, H . , Lubbers, M.. and Strauss, E., Water, Air Soil Pollut., 1989, 45, 121. Bicanic, D. D., Favier, J. P., Strauss, E., Lubbers, M., and Fleuren, G., Int. J. Environ. Anal. Chem., 1990, 37, 623. Photothermal Investigations of Solids and Fluids, ed. Sell, J. A., Academic Press, New York, 1988. Boccara, A. C., Fournier, D., Jackson, W., and Amer, N. M., Opt. Lett., 1980, 5 , 377. Jackson, W., Amer, N. M., Boccara, A. C., and Fournier, D., Appl. Opt., 1981, 20, 1333. Charbonnier, F., and Fournier, D.. Rev. Sci. Instrum., 1986,7, 1126. Imasaka. T., and Ishibashji. N., Anal. Chem., 1990, 62, 363A. Paper 01’00091 D Received January 5th, 1990 Accepted September 18th, 1990

ISSN:0003-2654    

DOI:10.1039/AN9911600077

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, JANUARY 1991, VOL. 116 81 Spectrofluorimetric Flow-through Sensor for the Determination of Beryllium in Alloys M. de la Torre, F. Fernandez-Gamez, F. Lazaro, M. D. Luque de Castro" and M. Valcarcel Department of Analytical Chemistry, Faculty of Sciences, University of Cordoba, 14004 Cordoba, Spain A spectrofluorimetric sensor for the determination of beryllium based on the use of morin, immobilized in a resin exchanger located in a flow cell, is proposed. The determinative method developed with this sensor has a linear range between 1 and 40 ppb, with a relative standard deviation of 1.71 % and a sampling frequency of 30 h-1. Its selectivity allows the determination of beryllium in simulated alloy samples with excellent results. Keywords: Spectrofluorimetric sensor; beryllium determination; alloy samples; flow injection The presence of small concentrations of beryllium in manufac- tured materials (e.g., in aluminium-based alloys to retard the oxidation of the melt1 and in nickel alloys to produce high- strength alloys resistant to corrosion and wear2) calls for sensitive, accurate analytical methods for the appropriate control of these concentrations. A variety of methods have so far been proposed for this analyte.3 Because of the lack of specific methods for beryllium, its determination may require a prior separation step depending on the sample matrix.Conventional atomic spectrometric techniques, which are very selective, do not provide sufficiently low determination ranges.4 On the other hand, spectrofluorimetry is more sensitive and affords lower determination ranges.Morin (2',4',3,4,7-pentahydroxyflavone) is one of the spectroflu- orimetric reagents for berylliums-8 that has been studied most often. Several flow injection (FI) methods for this analyte have also been proposed, particularly spectrophotometric methods (determination of beryllium in copper-based alloys using Xylenol Orangeg) and atomic methods (liquid-liquid extrac- tion with acetylacetonate prior to introduction into an inductively coupled plasma atomic emission spectrometerlo). The principles behind the proposed sensor have only been used partly, although not successfully, by other workers. Saari and Seitz" developed a spectrofluorimetric probe for beryl- lium based on the immobilization of morin on cellulose.A thin layer of support plus reagent on cellophane was located at the end of an optical fibre, which was then introduced into the solution being measured. The sensor had a response time of 2 min and a restoration period of 3 min. No practical application has been reported. Recently, Capitan et a1.12 have reported an ion-exchange batch spectrofluorimetric method for the deter- mination of beryllium with morin. A series of tedious steps makes this method impractical for routine analysis. The proposed sensor uses morin, previously immobilized on an ion-exchange resin located at the spectrofluorimetric flow cell, which eliminates all manipulation as the sample is injected into a carrier-eluting stream (HN03) and, after measurement, the sensor is regenerated, and hence rendered ready for subsequent analyses, according to the following reactions: In the support In solution In the support HO doH + Be2+- OH- H o d o H \ HN03 7 / OH / 0- OH 0 o.-..-i)e2+ * To whom correspondence should be addressed.Experimental Reagents Beryllium (1 g dm-3) and morin (1 g dm-3 in a 40% solution of ethanol) were used. More dilute solutions of Be11 were prepared in 0.02 mol dm-3 NaOH. Dowex 1 X2-100, 1 X4-100, 1 X8-100, 1 X4-50, 1 X4-200, 1 X4-400 and 50 X4-100; Sephadex QAE 50-120 and 25-120 and diethyl- aminoethyl dextran (DEAE) 50-120 and 25-120 mesh resins, obtained from Sigma, were also used. Apparatus Fluorescence measurements were made on a Kontron SFM 25 spectrofluorimeter equipped with a Knauer x-t recorder, a Gilson Minipuls-3 peristaltic pump, a Rheodyne 5041 injec- tion valve of variable volume and a fluorescence flow cell with a path length of 1.5 mm packed with the appropriate resin.Procedure Solid samples (alloys) must be dissolved in nitric acid and diluted in order to bring the beryllium concentration within the linear range of the calibration graph (1-40 ppb). Ethylene- diaminetetraacetic acid (EDTA) was added up to a concentra- tion of 1 x 10-2 mol dm-3. A 1 ml volume of this solution was injected into a single-channel FI manifold through which an eluent stream (2 x 10-2 mol dm-3 HN03 + 2 x 10-2 mol dm-3 NaN03 + 20 ppb morin) carried the sample plug to the flow cell, where the resin support housed the immobilized ligand (morin). The transient signal (FI peak) thus obtained rapidly reached the baseline as a result of the dissociation of the complex and elution of the analyte.A blank was also used under the same working conditions. The difference between both signals was compared with the data from the calibration graph which had been constructed, under the same conditions, using standards in an acidic medium similar to that used for the samples. Results and Discussion Study of Variables This study was aimed at finding the optimum conditions (sensitivity, stability and regenerability) for the spectroflu- orimetric sensor. The variables influencing the system were divided into three groups, namely: those typical of the reaction-detection unit, chemical and FI. Variables of the reaction-detection unit The most suitable flow cell was a Hellma Model 176.052 QS with an inner volume of 25 p1 and a 1.5 mm path length.This path length was sufficient to ensure adequate sensitivity without the detector capacity being saturated by the fluor- escence of the resin plus morin it contained.82 ANALYST, JANUARY 1991, VOL. 116 The basic variables of the solid support were its chemical nature, mesh size and degree of cross-linking. Both cationic and anionic resins of different types (Sephadex and Dowex) were studied. The most suitable type of support was Dowex because of its high capacity to retain and elute the anionic complex. This is a result of the electrostatic charges of the complex and exchanger being involved in aromatic interaction with the support. Cationic Dowex supports were unsuitable as they neither immobilized morin nor allowed the formation of the Be-morin complex.The excellent immobilization of morin on the anionic Dowex exchanger was a result of the resonance forms of the reagent present in the basic medium used for immobilization. The time taken for equilibrium to be reached between the solid phase, on which the ligand is immobilized, and the beryllium solution, increases with decreasing mesh size, hence Dowex 1 X4-200 and 1 X4-400, with smaller mesh sizes, resulted in less effective Be-morin contact than Dowex Increasing degrees of cross-linking (between 2 and 8%) resulted in decreasing fluorescence signals. The most suitable support (Dowex 1 X4-100) was thus used in subsequent experiments. The level of the resin in the flow cell was a key variable because the maximum formation of the complex took place at the solidliquid interface.This phenomenon was most intense at the top of the packed support as the flow impinged from above; thus the top of the resin was kept as close as possible to the path length, all of which should be occupied by the resin. The optimum level was 8.0-8.5 mm. Other variables affecting the reaction-detection unit were those related to the reagent immobilization, i.e., morin concentration in the resin, ethanol content and time required for the immobilization. The baseline fluorescence signal increased with the concen- tration of morin retained in the resin up to a constant fluorescence intensity (If) of 68% (on a full scale of 100% If), which resulted in a narrow range for the occurrence of the analytical signal.On the other hand, high concentrations of morin resulted in uneven formation of the Be" complex possibly because of the difficulty in accessing active sites. This gave rise to an erratic elution equilibrium and, hence, to major irreproducibility of the measurements. A solution of morin at a concentration of 2.0 yg ml-1 provided adequate immobiliza- tion with good reproducibility and an acceptable range within which to perform the spectrofluorimetric measurements of the analyte. Increasing the ethanol content of the aqueous solution of morin (between 0 and 100%) resulted in no change in the amount of morin immobilized and produced similar baselines in all instances. Nevertheless, the analyte signal was most clearly defined when the immobilization was carried out with morin in a 40% solution of ethanol. The shape of the transient signal was significant when measurements of the slope of this signal were made.The influence of the time of contact between the resin and the morin solution was studied for periods between 5 and 420 min. When the period of contact was completed the resin was washed with distilled water and stored in 0.05 rnol dm-3 HN03 for 24 h. The baseline obtained was constant for times equal to or longer than 20 min; 45 min was chosen as the time producing the most reliable results. 1 X4-100. Chemical variables The chemical variables studied included sample pH, type and concentration of eluent, ionic strength and morin replenish- ment. The pH of the sample was changed with 1 x 10-3 and 1 x 10-1 rnol dm-3 NaOH for samples containing 10 ppb of Be".The maximum signal was obtained for an NaOH concentra- tion of 2 x 10-2 rnol dm-3, which resulted in a broadened linear analytical range. An eluent that could also be used as carrier was the best for this type of sensor as it simplified the FI manifold dramati- cally. An acid carrier-eluent allowed the Bell-morin complex to be decomposed very rapidly, the Be" being separated and the morin being kept in the resin, ready for the next determination. A 2 x 10-3 mol dm-3 solution of HN03 provided a lower baseline and higher analytical signal with effective elution. The ionic strength of the eluent was a key factor because of the degree of hydration, on which the compactness of the resin depends markedly. A decrease in the ionic strength yielded a higher fluorescence intensity; thus, NaN03 at a Concentration of 2 x 10-2 rnol dm-3 was used in the eluting carrier. A study of the stability of the immobilized morin was performed by repetitively injecting Be" under the optimum working conditions.There was a slight decrease in the height of the transient peaks with time, which was indicative of a small loss of the reagent caused by the eluting carrier. A morin concentration between 0 and 30 ppb in the carrier was investigated to compensate for this loss. The concentration of ligand in the carrier providing stable signals was 20 ppb. Flow injection variables The FI variables studied were the sample volume, which influenced the sensitivity of the method, and the flow-rate, which determines the rate at which the analyte can be eluted and hence the sampling frequency.Another typical FI variable such as the reactor length had no influence on the performance of this method as no chemical reaction was involved during the transport of the sample to the detection system. Its length must be as short as possible in order to minimize dispersion of the sample into the eluting carrier. The flow-rate was varied between 0.44 and 2.48 ml min-l. These changes indicated that the retention of the Be" (complex formation) was not instantaneous as the transient signal increased as the flow-rate decreased. Nevertheless, very low flow-rates were incompatible with short residence times and rapid baseline restoration; thus, a flow-rate of 1.2 ml min-1 was selected as a compromise.Sample volumes smaller than 100 yl yielded signals that were similar to that of the blank. The signal obtained with sample volumes larger than 100 yl increased linearly up to a sample volume of at least 2.5 ml, which leads to the conclusions that (i) the immobilized reagent was in large excess and (ii) the rate of formation of the Bell-morin complex was fairly low. As large sample volumes resulted in decreased sampling frequency, an injected volume of 1.0 ml was selected. Features of the Method The calibration graph was constructed under the optimum working conditions. The data obtained revealed a linear Table 1 Tolerated ratios of interferents in the determination of beryllium Tolerated ratio of Be" : foreign ion (m/m)* Ion added Ca" 1 : 100 - Cu" 1 : 100 - Co" 1 :40 (1 : 400) MglI 1:10 (1 : 300) * Ratios in parentheses correspond to the tolerance in the presence FelI1 1 :40 (1 : 400) Al"' 1 : l (1 : 100) of 1 x 10-2 rnol dm-3 EDTA.83 ANALYST.JANUARY 1991, VOL. 116 Table 2 Determination of beryllium in simulated typical alloys Composition (%) Alloy Be Co Ni Ag Copper-based 4.00 - - - 1.90 0.24 - - 3.40 0.24 - - 0.50 2.50 - - 0.40 1.60 - 0.95 0.40 - 1.50 - 2.10 0.50 - - 2.45 - 1.10 - Aluminium-based 5.00 - - - 0.25 - - - Nickel-based 1.80 - - - 2.70 - - - * k standard deviation ( n = 3). Uses Master alloy Electrical terminals, springs Welder bar Good electrical conductor Bearings Welding equipment Casting alloy Casting alloy Master alloy Aircraft alloy Springs Castings Be found (%)* 4.00 f.0.04 2.00 k 0.05 3.30 k 0.03 0.51 f. 0.04 0.43 k 0.07 0.40 k 0.02 2.30 k 0.03 2.44 k 0.06 4.97 k 0.04 0.25 f. 0.02 1.85 k 0.03 2.73 k 0.03 analyte concentration range between 1 and 40 ppb. Two equations were obtained according to whether the height of the transient signal or its rising slope was measured between 15 and 30 s. The equations and their regression coefficients are as follows: Amax: S1 = 10.48X + 6.767 (r2 = 0.9960) Tg (15-30 s): S2 = 0.50X + 0.440 (r2 = 0 9964) where S1, S2 and X are % fluorescence intensity, % fluorescence intensity s-1 and ppb of beryllium, respectively; Tg is the slope of the rising portion of the analytical signal, S, and is given by Tg = (S30 - S1s)/(t30 - tls) (the subscript numbers correspond to time, in seconds).The reproducibility of the method was studied on 11 samples containing 20 ppb of Be" injected in triplicate. The precision obtained, expressed as per cent. relative standard deviation (RSD), was 1.71. The sampling throughput was 30 samples h- 1 . A study of potential interferences in the determination of Bell using the proposed method was performed. The determi- nations were carried out at a Be" concentration of 20 ppb and the maximum Be11 : foreign ion ratio was 1 : 100. Table 1 lists the tolerated ratios of the species determined. A given species was considered not to interfere when the signal obtained lay in the interval S, k o (S, = fluorescence intensity of the analyte without interferent, o = RSD). Only one of the ions assayed (Al"') interfered at the same level as the analyte, while most of the others did not interfere in excesses of up to 100-fold over the analyte concentration. Nevertheless, the selectivity of the method was inadequate for application to the determination of beryllium in alloys.Thus, the masking of these species with EDTA at different concentrations was performed in order to increase their tolerated level.6 The tolerated levels achieved by masking with 1 X lo-' mol dm-3 EDTA are listed in parentheses in Table 1. Application of the Method to Simulated Alloys As the most common samples for the determination of beryllium are alloys, and taking into account the lack of standards of this type, simulated alloys,13 with similar compo- sitions to the most common alloys and with an acidic medium to assimilate the matrix of real samples, were prepared.The acid concentration in the samples was that required for the dissolution of the alloys, taking into account that the mass of the sample would be 2.5 mg. Table 2 lists the composition of these samples and the use of each, and also the percentage of Be" obtained for the different samples that were analysed. The procedure used is described under Experimental. These results show good agreement between the experimental and actual values for all the alloys studied. Conclusions The results obtained by applying the proposed method to synthetic samples similar to alloys lead to the conclusion that the proposed sensor is suitable for this type of analysis as the determination is fast, selective, inexpensive and simple. The proposed method has the following features: ( i ) it uses a very simple single-channel manifold with a conventional spectro- fluorimetric detector, (ii) requires no sample pre-treatment, (iii) has low reagent consumption, (iv) provides low determi- nation limits and a good linear determination range, ( v ) allows good sample throughput, and (vi) provides excellent selectiv- ity, which allows the determination of the analyte in real samples.The authors thank the Comision Interministerial de Ciencia y Tecnologia (CICyT) for financial support (Grant No. PA86- 0146). 1 2 3 4 5 6 7 8 9 10 11 12 13 References Bass, N. W., in The Metal Beryllium, eds. White, D. W., Jr., and Burke, J. E., The American Society for Metals, Cleveland, Burns, W. R., Chemical Spectroscopy, Wiley, New York, 1956, Snell, F. D., Photometric and Fluorimetric Methods of Analysis. Metals. Part I . Wiley, New York. 1978, pp. 681-690. Vanhoe, H., Vandecasteele, C., Desmet, B., and Dams, R., J. Anal. At. Spectrom., 1988, 3, 703. Will, F.. Anal. Chem., 1961, 33, 1360. May, R., and Grimaldi, F. S . , Anal. Chem.. 1961,33, 1251. Sill, C. W., Willis, C. P., andFlygare, J. K., Anal. Chem., 1961, 33, 1671. Fletcher, M. H., Anal. Chem.. 1965, 37, 551. Mochizuki. M., and Kuroda, R., Fresenius 2. Anal. Chem., 1981. 309, 363. Yamamoto. M., Obata, Y., Nitta, Y . , Nakata, F., and Kumamaru, T., J. Anal. A t . Spectrom.. 1988, 3, 441. Saari, L. A . , and Seitz. W. R.. Analyst, 1984, 109, 655. Capitan, F., Manzano, E., Navalon, A., Vilchez, J. L., and Capitan-Vallvey, L. F., Analyst, 1989, 114, 969. Kjellgren, B. R. F., Schwenzfeier, C. W., Jr., and Melick, E. S . , in Treatise on Analytical Chemistry, eds. Kolthoff, I. M., and Elving, P. J., Wiley, New York, 1st edn., 1964, vol. 6, part 2, pp. 12 and 13. OH, 1955, ch. II-C. p. 449. Paper 0102468F Received June 4th, 1990 Accepted August 29th, I990

ISSN:0003-2654    

DOI:10.1039/AN9911600081

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, JANUARY 1991, VOL. 116 85 Thiophosphoryl Compounds as Novel Inducing Agents in the Iodine-Azide Reaction Witold Ciesielski, Wtodzimierz Jedrzejewski and Zbigniew H. Kudzin Institute of Chemistry, University of Lodz, Narutowicza 68, 90- 136 Lodz, Poland Piotr Kietbasinski and Marian Mikotajczyk Centre of Molecular and Macromolecular Studies, Polish Academy of Sciences, Sienkiewicza I 12, 90-363 I: o dz, Po land The application of organophosphorus compounds as inducing agents in the iodineazide reaction was investigated. The induction activity was exhibited by thiophosphoryl compounds; their induction coefficients were dependent on the number and nature of sulphur atoms in the P(S), function. These relationships can be used for the group differentiation of organophosphorus compounds, for example phosphates, thiophos- phates and dithiophosphates. On the basis of their induction activity the microdetermination of thiophos- phoryl inductors on a nmol scale was elucidated.Keywords : Induced iodine-azide reaction; organophosphorus indue tors; microde term ination of thio- phosphoryl inductors The induced iodine-azide reaction, originally carried out by Raschig, 1 has been explored extensively for decades in analytical chemistry .2-10 This reaction, represented schemat- ically in equation ( l ) , occurs in the presence of inducing agents (inductors), usually divalent sulphur-containing compounds: inductor I , + ~ N , - - 21- The greatest induction was exhibited by dithiocarba- mates,8.11 thiolamino acids, e.g., cysteine4.9.12 or reduced gluthathione ,9,13 thioureasl4 and also compounds bearing thiopurine15 or thiopyrimidinel6 moieties.However, of the many analytical papers published on the iodine-azide reaction there is only one describing the induction activity of the P-S-containing compounds, and this is limited to thin-layer chromatographic detection of phosphorothiono ester-based pesticides. 17 Taking into account the industrial importance of organic thiophosphoryl compounds (e.g., plant protection agents, extraction and flotation agents, rubber industry) ,18-20 the large positive effect caused by thiophosphoryl inductors can constitute the basis for a novel procedure for their determination. In this paper the results of studies on the inductive effect of some thiophosphoryl compounds on the iodine-azide reaction are presented.The efficiency of thiophosphoryl compounds as inductors has been characterized and compared on the basis of their induction coefficients ( F J , defined by Fi = nI/ni where nI = millimoles of iodine consumed in the induced reaction and ni = millimoles of the inductor. Experimental The iodine determinations were performed using a VSU-2P spectrophotometer (Carl Zeiss, Jena) equipped with quartz cuvettes of path length (I) 1 or 5 cm. Reagents and Solutions All compounds 1-12 (Table 1) were synthesized according to the literature and were of the same purity as reported previously.21 They were also found to be homogeneous according to 31P NMR data. Organophosphorus compounds 1 and 2 were used as 0.001 mol dm-3 aqueous solutions and 3-12 as 0.001 mol dm-3 ethanolic solutions.All aqueous solutions were prepared using re-distilled (in glass apparatus) water. Iodine solution. A 0.01 mol dm-3 aqueous stock solution of iodine, containing 4 g 1-1 of potassium iodide, was prepared and a 0.0015 mol dm-3 solution was prepared by dilution of the stock solution. Potassium iodide solution. A 0.35 mol dm-3 aqueous solution was prepared. Reaction solution. The solution contained 20 g 1-1 of sodium azide and was buffered to pH 6 with hydrochloric acid. Determination of the Inductive Effect on the Iodine-Azide Reaction The consumption of iodine in the induced iodine-azide reaction was determined spectrophotometrically by measur- ing the absorbance at 350 nm, characteristic of the triiodide complex (I3-).The molar absorptivity of the triiodide was dependent on the concentration of the iodide anion owing to the reversibility of the formation reaction I2 + I- e 13- (3) with a plateau where the potassium iodide concentration is about 3 x 10-2 mol dm-3. At this point the molar absorptivity reaches a maximum value of 2.5 x 104 dm3 mol-1 cm-1. However, at the same time the induced reaction is adversely affected by the presence of the iodide anion and is highly suppressed when the iodide concentration reaches about 1 x 10-2 mol dm-3. To reconcile these two opposing effects, all experiments were carried out at the lowest possible iodide concentration (about 1.2 x 10-4 mol dm-3), resulting only from the applied solution of potassium triiodide. Sub- sequently, the concentration of potassium iodide was adjusted to 3 X 10-2 mol dm-3, i.e., the optimum for the subsequent spectrophotometric determination of iodine.Procedure for Determination of the Induction Coefficient A 50 pl volume of a solution of iodine was injected into the reaction cell containing 10 ml of a stirred reaction solution and a sample of the organophosphorus inductor (0.5-160 nmol) . The reaction mixture was allowed to stand for the time indicated in Table 1, 1 ml of potassium iodide solution was added and the mixture was measured spectrophotometrically at 350 nm, giving the total absorbance of unconsumed iodine (A). A similar measurement on a blank solution without organophosphorus inductor gave the total absorbance (Ao) .86 ANALYST, JANUARY 1991, VOL.116 The amount of iodine consumed in the induced iodine- azide reaction was calculated from the equation (4) where V is the volume of the reaction mixture (11.1 ml), I is the light path of the cuvettes (1 cm) and E ~ ~ - is the molecular ab- sorptivity of the triiodide anion (2.5 X lo4 dm3 mol-1 cm-1). The induction coefficients of the organophosphorus inductors (F'i) were calculated from equation (2). Procedure for Determination of Thiophosphoryl Inductor The determination of the thiophosphoryl inductor ( e . g . , disulphide 7, see Table 2) was performed under conditions identical with those described for the determination of the induction coefficients. For the determination of compound 7 at levels of 0.05-0.25 nmol a 0.0015 mol dm-3 iodine solution was used, followed by ultraviolet spectrophotometric measurements in cuvettes with a light path 1 = 5 cm.The relationship AA = A. - A as the function of the amount of inductor was used in the construction of the calibration graph (linear in the ranges 0.050-0.25 and 0.25- 2.00 nmol of compound 7) and was applied to the spectro- photometric determination of compound 7. Results and Discussion The results obtained for the induction activities of organo- phosphorus compounds 1-12 in the induced iodine-azide Table 1 Induction coefficients (F,) of thiophosphoryl compounds* No. Structure 1 (EtO),PS,K 2 (EtO),PSOK 3 (EtO),P=S 4 (EtO),P(O)SEt 5 (EtO),P(S)SEt 6 (EtS)3P=S s s II I1 7 (EtO),PSSP(OEt), s s II II 8 (EtO),POP(OEt), s o II II 9 (EtO),POP(OEt), 10 Bu~P=S 11 Ph3P=S 12 Ph,(EtO)P=S Induction Range of Reaction coefficient determination/ time/min (Fi) nmol 3 220 0.5-4 15 6 20-160 0 15 20 5-40 5 40 2.5-25 10 8 15-120 3 450 0.25-2.0 10 12 10-80 15 6 20-160 1 190 0.6-5 10 213 0.5-4 12 105 1.0-8 * Phosphoric acid, diethylphosphoric acid and triphenylphosphine oxide did not exhibit the induction.Table 2 Results for the determination of disulphide 7 (n = 6) Relative Cuvette standard Determination path length, Taken/ Found/ deviation No. llcm nmol nmol (%) 1 2 3 4 5 6 7 8 9 10 1 1 1 1 1 5 5 5 5 5 0.25 0.50 1.00 1.50 2.00 0.050 0.100 0.150 0.200 0.250 0.26 0.52 1.02 1.48 2.01 0.047 0.101 0.153 0.198 0.245 5.2 2.4 3.6 3.0 2.1 7.3 4.6 3.1 3.5 2.4 reaction are given in Table 1. They indicate that the induction activity of thiophosphoryl compounds is strongly dependent on the structure, especially on the nature of the P-S bonds.Thus, potassium diethyl phosphorodithioate (1) exhibits a high induction activity (Fi = 220), apparently due to the presence of the thiolate function in the PS2- anion. For disulphide 7, which may be formally considered to be a dimer of compound 1, the induction activity (Fi = 450) is approxi- mately double that of compound 1, probably owing to the facile cleavage of 7 under the reaction conditions. However, the induction activity of potassium diethyl phosphorothioate (2) (Fi = 40) is only about 20% of that of compound 1. Conversion of a free thiolate function in compound 1 into a stable thioester function causes a decrease in the induction activity of the resulting compounds.Thus, comparison of the induction activities of triethyl thiophosphates, compounds M, reveals a lack of activity of the phosphorothiolate 4 (Fi = 0), low activity of triethyl phosphorothionate (3) (Fi = 6) and triethyl phosphorotetrathiolate (6) (Fi = 8) and a slightly higher activity of triethyl phosphorodithioate ( 5 ) (F, = 20). In contrast, phosphine sulphides 10 and 11, containing the P-S bond, exhibit remarkably high induction effects, comparable to that of potassium diethyl phosphorodithioate ( l ) , with induction activity coefficients of 190 and 220, respectively. However, the replacement of the phenyl group in phosphine sulphide 11 by the ethoxy group leads to ethyl diphenyl- phosphinothionate (12) and to a substantial decrease in the induction activity of this compound (Fi = 105).Tetraethyl monothiopyrophosphate (9) and tetraethyl dithiopyro- phosphate (8) also exhibit low induction activities (F', = 6 and 12, respectively). The proposed method can also be used for the determina- tion of thiophosphoryl compounds by application of the calibration procedure described for C-S inductors.9 Thus, the determination range of disulphide (7) (Fi = 450) is between 0.25 and 2.0 nmol (or 0.05-0.25 nmol using a cuvette with a light path I = 5 cm) (Table 2). The determination ranges for other thiophosphoryl compounds tested are higher, corres- ponding to their lower induction activities expressed by their Fi values (Table 1). The analytical evaluations of the induction activities of thiophosphoryl compounds will be published elsewhere.The determination of the induction coefficients can also be used to distinguish between phosphates, phosphorothioates and phosphorodithioates and also phosphine sulphides from phosphine oxides. This project was supported financially by the Polish Academy of Sciences, grant CPBP 01.17 (analytical part) and Grant CPBP 01.13 (synthetic part). 1 2 3 4 5 6 7 8 9 10 11 12 13 14 References Raschig, F., Chem. Ztg.. 1908, 32, 1203. Feigl, F., and Chargraff, E., Fresenius 2. Anal. Chem., 1928, 74, 376. Kurzawa, Z., Chem. Anal. (Warsaw), 1960, 5 , 551. Strickland, R. D., Mack, P. A., and Childs, V. A., Anal. Chem., 1960,32,430. Weisz, H . , and Ludwig, H., Anal. Chim. Acta, 1972, 60, 385. Kiba, N., Suto, T., and Furusawa, M., Talanta, 1981, 28, 115. Kiba, N., and Furusawa, M., Tulanta, 1981, 28,601. Ciesielski, W., and Jqdrzejewski, W., Mikrochim. Acta, Part 111, 1984, 177. Ciesielski, W., Chem. Anal. (Warsaw), 1986, 31, 39 and 469. Wronski, M., and Goworek, W., Analyst, 1987, 112, 333. Kurzawa, Z., and Kubaszewski, E., Chem. Anal. (Warsaw), 1974, 19, 263. Ishii, K., J. Jpn. Biochem. Soc., 1952, 24, 118; Chern. Ahstr., 1953,47, 12125. Whitmann, D. W., and Whitney, R. M., Anal. Chem., 1953,25, 1523. Suzuki, S., Bunseki Kagaku, 1962, 65, 326 and 898.ANALYST, JANUARY 1991. VOL. 116 87 15 Kurzawa, Z., Matusiewicz, H., and Matusiewicz, K., Chem. Anal. (Warsaw), 1974, 19, 1174. 16 Kurzawa, J., Chem. Anal. (Warsaw), 1987, 32. 875. 17 Cserhati. T., and Orsi, F., Period. Politech., Chem. Eng., 1982, 26, 111. 18 Hassall, K. A., The Chemistry of Pesticides, Verlag Chemie, Weinheim, Deerfield Beach, Bade, 1982. p. 67. 19 Reid. E. E., Organic Chemistry of Bivalent Sulfur, Chemical Rubber Publishing, New York, 1962, vol. 1, p. 304. 20 21 Handley, T. H., Talanta, 1965, 12, 893. Methoden der Organischen Chemie (Houben- Weyl). 4. Auflage, Band XII, Georg Thieme, Stuttgart, 1963. Paper 0101 I71 A Received March 16th, 1990 Accepted August 3rd, I990

ISSN:0003-2654    

DOI:10.1039/AN9911600085

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, JANUARY 1991, VOL. 116 89 Spectrophotometric Determination of Trace Amounts of Copper([) and Reducing Agents With Neocuproine in the Presence of Copper(i1) Esma Tutem and Regat Apak* Department of Chemistry, Faculty of Engineering, University of Istanbul, Vezneciler 34459, Istanbul, Turkey Fikret Baykut Department of Chemical Engineering, Faculty of Engineering, University of Istanbul, Vezneciler 34459, Istanbul, Turkey The existing spectrophotometric method for the determination of total Cu with neocuproine (Nc) does not allow the differentiation of Cul and Cull. It is shown here that the use of a dilute (3.0 x 10-3 rnol dm-3) Nc solution in weakly acidic or neutral media makes the determination of Cul feasible at the 1 x 10-5 rnol dm-3 level in the presence of up to 0.1 rnol dm-3 Cu".For the determination of trace amounts of Cul in the presence of 1 rnol dm-3 Cull, the latter can be masked by NH3-NH4CI buffer a t pH 10, giving almost the same molar absorptivity for Cul, i.e., 7.5 x lo3 dm3 mol-1 cm-1. The ability to measure the absorption due to Cul in the presence of an excess of Cull was exploited as the basis of an indirect method for the determination of trace amounts of reductants. Superior to the existing individual methods of determination, which are usually specific for a given reducing agent, the proposed system consists of treating the reductant with the Cull-Nc reagent in ammonium acetate buffered media, followed by measurement of the absorbance of the Cul-Nc chelate at 450 nm. The quantification of a given reductant in the concentration range 1 x 10-6-1 x 10-4 rnol dm-3 usually takes 3 min with a mean relative standard deviation of 3%.The molar absorptivity of an n-electron reductant which has reacted stoichiometrically is approximately 7.5n x lo3 dm3 mol-1 cm-1, i.e., n times that of Cul-Nc. Hence, hydrogen peroxide, ascorbic acid, cysteine, hydroxylamine, hydrazine, thiosulphate, dithionite, mitoxantrone, glutathione, iron(i1) and thiourea were determined with theoretical molar absorptivities; carminic acid, sulphite, tin(ii) chloride, 2,4-dinitrophenylhydrazine, sodium tetrahydro- borate(iii) and 2,3-dimercaptopropan-1-ol were determined with the aid of empirical linear absorbance- concentration plots. The proposed spectrophotometric method is rapid and allows the quantification of biologically important reducing agents.Keywords: Copper(1) determination; reducing agent determination; copper(i1)-neocuproine reagent; neo- cuproine; spectrop hotometry The formation of the charge-transfer complex' between Cur and neocuproine (Nc) (2,9-dime thyl- 1,lO-phenanthroline) is the basis of the existing spectrophotometric method for the determination of total Cu.2-5 However, this method cannot differentiate between the mono- and divalent forms of Cu. Copper ion-selective electrodes are also useless for analysing mixtures of CuI and CuIr.6 The CuI-CuII redox couple is the focus of much attention in the studies of metalloenzyme modelling7 and Cu speciation in environmental waters.8 Therefore, the quantification of trace amounts of CuI in the presence of an excess of Cu" is important from an analytical standpoint.Considering the relative stabilities of the Nc and ethylene- diaminetetraacetic acid (EDTA) complexes of CuI and Cu" under specified experimental conditions, Ulibarri et aZ.9 concluded that the selective determination of CuI, stabilized by ally1 alcohol in the mixture, should be possible in the presence of CulI by masking the CuII with EDTA. However, the present study showed that such a high concentration of Nc (0.1% in ethanol), as used in the above work,9 would inevitably result in a positive interference from CuII. More- over, the unexpectedly low absorbances at 450 nm reported by the same workers9 of mixtures containing CuI, CuII, Nc and EDTA, which could have arisen from incomplete conversion of Cul into the CuI-Nc complex via mixed-ligand complex formation, suggest that there is a need to develop more satisfactory analytical procedures.Moffett et a1.10 were able to measure the absorbance due to Cu1-bathocuproine while preventing the reduction of 1 x 10-5 mol dm-3 Cu" by NH20H by masking the CUT' with * To whom correspondence should be addressed. ethylenediamine; however, concentrations of CuII were not tested. Hence a simple and reliable analytical method for Cu' has been developed in the present study which does not require the use of masking agents for CuII up to a concentration of 0.1 mol dm-3; most masking agents lead to negative errors in the quantification of Cur. The proposed method is also suitable for the determination of reducing agents with high sensitivity. Various spectrophotometric procedures exist for the deter- mination of biologically important reductants: hence, thiol groups can be determined with sulphanilamide-naphthylethy- lenediamine,ll N-ethylmaleimide,12 p-chloromercuriben- zoate,l3 N, N-dimethyl-l,4-phenylenediammonium chloride and chloramine-T,14 aminophenols and iron(111)15 or by using l-chloro-2,4-dinitrobenzene and the Ellman reagent.16.17 Glu- tathione can be determined with alloxan;l8 1,Zdithiols [e.g., 2,3-dimercaptopropan-l-o1 (BAL)] with benzidine;lg thiourea with iron(Ir1) thiocyanate" or tetraiodomercurate;2O ascorbic acid with phosphomolybdate,ll 2,6-dichlorophenolindophe- no111 or sulphanilamide-naphthylethylenediamine;11 and hydroxylamine by using the sulphanilic acid-1-naphthylamine azo coupling21 or aldoxime22 reaction.Reducing agents can be determined by reduction of a metal ion exhibiting multiple valencies, followed by treating the lower valency of the metal with a chromogenic reagent. Ascorbic acid was determined by oxidation with FeIII- 1,1O-phenanthroline23-25 or by oxidation with CuII and subse- quent colour development with Nc in a mono-26 or biphasic system.27 Schilt and Di Tusa28 used a chromogenic reagent for FelI such that the reductant to be determined reduced Fe"1 to a coloured product. However, these redox reactions required elevated temperatures (60-90 "C) and considerable periods of time (30-240 min) and were mostly incomplete,28 so that the90 ANALYST, JANUARY 1991, VOL. 116 calibration graphs did not correspond to the theoretical n-electron oxidations of the reductants being determined.In this work the ability to measure trace amounts of CuI in the presence of an excess of Cu" was exploited with the aim of indirectly quantifying reducing agents that could reduce the CuII-Nc reagent at a suitable pH to the coloured CuI-Nc chelate. Hence, an n-electron reductant that reacted stoi- chiometrically would exhibit an effective molar absorptivity of approximately 7.5n X 103 dm3 mol-1 cm-1. Experimental Reagents All chemicals were obtained from Merck and were of analytical-reagent grade with the exception of mitoxan- trone, which was synthesized according to the procedure of Murdock et al.29 All solutions and distilled water were de-aerated with oxygen-free nitrogen before use.Copper(r1) solutions of 0.1, 0.01 and 0.001 rnol dm-3 were prepared by appropriate dilution of a 1 mol dm-3 CuC12 stock solution. Copper(1) solution, 1 X 10-3 mol dm-3. Prepared from a newly opened pack of CuCl which had been kept in a de-aerated desiccator. A 0.0099 g amount of CuCl was dissolved in 25 ml of 7% NH4C1 solution and diluted to 100 ml with water, the operation being carried out under a flow of nitrogen. This solution can only be used for four measure- ments within a period of 15 min after which it must be discarded, otherwise probable oxidation and disproportiona- tion reactions of Cut would cause absorbance decreases. When Cut alone is to be used rather than a synthetic mixture of both Cut and Cu", a 1 x 10-3 rnol dm-3 CuI solution can be prepared from the corresponding amount of CuCI2 in 1 x 10-2 rnol dm-3 NH20H solution.Stable CuI solution in the form of CU(NC)~CI, 1 x 10-3 rnol dm-3. Copper(I1) chloride (0.05 mmol) was dissolved in 50 ml of 1 X lo-* rnol dm-3 NH20H.HC1 solution and 10 ml of ethanolic 0.015 rnol dm-3 Nc solution were rapidly added. The solution was shaken with consecutive 10 ml portions of CHCI3 until the CHCI3 phase was colourless. The combined organic phases were evaporated under a gentle stream of nitrogen, and 25 ml of ethanol were added before the CHC13 solution became dry. The addition of ethanol and the evaporation procedure were repeated twice under nitrogen, ensuring that the ethanolic solution did not evaporate to complete dryness. The ethanolic residue was taken up in 50 ml of 96% ethanol and stored in a refrigerator. The resulting solution is stable and has a CuI concentration of 1 X 10-3 rnol dm-3.Ammonia-ammonium chloride concentrated buffer of p H 10. Prepared from commercial (25% m/m) NH3 solution and solid NH4CI as 7.2 k 0.1 rnol dm-3 NH3 and 1.3 mol dm-3 NH4CI. The variation in NH3 concentration was caused by bubbling nitrogen through the buffer occasionally. Working solutions (5.0 x 10-4 mol dm-3) of the reducing agents (including NH20H.HC1) were prepared from the corresponding 1 x 10-2 mol dm-3 stock solutions. A 1 x 10-2 rnol dm-3 hydrogen peroxide solution was prepared from commercial hydrogen peroxide (35%) and standardized by titration with KMn04 solution. A 5.0 X 10-4 rnol drn-3 2,4-dinitrophenylhydrazine solution was prepared in ethanol.A 1 x 10-2 mol dm-3 NaBH4 solution was prepared in 1 X 10-2 rnol dm-3 NaOH immediately before use. The pH 7 buffer was 1 rnol dm-3 ammonium acetate. The pH 7.4 phosphate buffer was obtained by mixing equimolar (1 X 10-3 rnol dm-3) Na2HP04 and NaH2P04 solutions in appropriate proportions. A 0.3 rnol dm-3 solution of EDTA was prepared from Na2EDTA. A 3.0 X 10-3 mol dm-3 Nc solution was prepared in 96% ethanol. Apparatus The absorptions at 450 nm were recorded with a Beckman DB-GT ultraviolet-visible spectrophotometer in silica cuvettes. The pH measurements were made with a Metrohm E-512 pH meter. A Texas TI 58C programmable calculator was used for regression analysis of the absorbance-concentra- tion data. Spectra Copper(1) chloride, or CuI obtained by reduction of Cu" with NH20H in Cu" solutions up to 0.1 rnol dm-3, absorbs at 450 nm against water in the Nc procedure. As for the NH3 buffered 1 rnol dm-3 Cull solutions, both blank and Cur- containing samples have a relatively high absorption at 450 nm; however, the CuI sample exhibits maximum absorbance at 450 nm against the reagent blank.Procedure for Studying the CuJ-CuII System A 1 ml aliquot of a CuC12 solution of the desired concentration (1.0, 0.1, 0.01 or 0.001 rnol dm-3) was placed in a test-tube and a flow of nitrogen through a capillary tube was started. After de-aerating for 1 min, either x ml of 1.0 x 10-3 rnol dm-3 CuCl (0.1 d x d 1.0) or x ml of 5.0 x 10-4 rnol dm-3 NH20H-HCI (0.1 d x d 1.0) was added. In the latter instance, where CuI is produced in the mixture by the addition of NH20H, bubbling of nitrogen was continued for a further 1 min to allow the reduction to go to completion.Then, (4 - x ) ml of water was added for the unbuffered medium studies o r (4 - x - y ) ml of water for the buffered work, followed by y ml of concentrated NH3 buffer. Finally, 2.5 ml of 3.0 x 10-3 rnol dm-3 Nc in ethanol were added. Thirty seconds after the Nc had been added, the capillary tube was removed and the contents of the test-tube were trans- ferred into a silica cuvette. The absorbance was measured at 450 nm against water 2 min after the Nc had been added. A reagent blank (containing no CuCl or NH20H) was run for all the experiments. Air must be removed from all the solutions and from the distilled water by purging with nitrogen prior to the experiments and all additions of reagents must be carried out under a flow of nitrogen.The total aqueous and ethanolic volumes in the final solution are 5.0 and 2.5 ml, respectively. For studies with pure CuCl (without CuII) either (5 - x) or (5 - x - y ) ml of water was added depending on whether the unbuffered or buffered medium was being employed. Recommended Procedure for the Determination of Trace Amounts of CuI in the Presence of CuII Procedure A or B should be selected depending on the CuII concentration in the sample. If the total Cu concentration is not known before analysis, it should be measured by the conventional NH20H-Nc method.2-4 If the Cur1 concentration is SO.1 rnol dm-3,.procedure A should be used; if the CuII concentration is of the order of 1 rnol dm-3, procedure B should be used.Procedure A Place an aliquot of a weakly acidic or neutral sample solution (x ml) containing preferably 0.5 x 10-4-10 x mmol of CuI in a test-tube. Dilute the sample with 7% NH4CI solution before analysis if necessary to bring the Cu1 concentration within the desired range. Add ( 5 - x) ml of water and 2.5 ml of 3.0 x 10-3 rnol dm-3 Nc in ethanol. Measure the absorbance at 450 nm against water after 2 min and make the necessary correction for the reagent blank. (Carry out the passage of nitrogen throughout the procedure as described above.) The molar absorptivity for Cut, obtained by reduction of CuI1 with suitable amounts of NH20H, is 8.0 X 103 dm3 mol-I cm-I.For the determination of CuI in the range 1.33 X 10-5 Q cCui dANALYST, JANUARY 1991, VOL. 116 91 1.6 x 10-4 mol dm-3 in the final solution (11 different concentrations), Beer's law is obeyed, and the equation of the straight line is: A = 8.0 x lO3c + 0.044 (correlation coefficient = 0.9995), where A is the net absorbance and c the Cu1 concentration (mol dm-3) in the final solution. The relative standard deviation is 4.2%. The limit of detection and the limit of quantification for Cut in the final solution are 1.5 X 10-6 and 5.0 x 10-6 mol dm-3, respectively. However, the molar absorptivity for 'unprotected' CuI, i.e., Cul used in the form of CuCl rather than produced in situ by the addition of NH20H, might be lower (4.5 X 103-7.5 X 103 dm3 mol-1 cm-1) depending on the Cull concentration. Therefore, in such instances the Cut concentration can be found from a calibration graph obtained with variable amounts of standard CuCl in a standard Cull solution with a Cu2+ concentration identical with that of the solution to be analysed.Procedure B Place an aliquot of a weakly acidic or neutral sample solution ( x ml) containing preferably 0.5 X 10-4-5 X mmol of Cut in a test-tube. Dilute the sample with 7% NH4CI solution before analysis if necessary to bring the Cut concentration within the desired range. Add (4.35 - x ) ml of water, 0.65 ml of pH 10 buffer and 2.5 ml of 3.0 x 10-3 rnol dm-3 Nc in ethanol in this order. Measure the absorbance at 450 nm against water after 2 min and correct for the reagent blank. (Carry out the passage of nitrogen throughout the procedure and stop the flow of nitrogen 30 s after the addition of Nc.) When the amount of Cult is 1 mmol, the reagent blank has an absorbance of 0.34 against water.The effective molar absorptivity for Cul produced in situ by reduction with NH20H is 7.7 x 103 dm3 mol-1 cm-1. The equation of the straight line is: A = 7.7 x 10% + 0.037 (correlation coefficient = 0.998), where A is the net absorbance and c the CuI concentration (mol dm-3) in the final solution. However, the molar absorptivity for 'unprotected' Cu1 (used in the form of CuCI) is 4.4 x 103 dm3 mol-1 cm-1 under the described experimental conditions. Therefore, in such instances the Cur concentration can be found from a calibration graph construc- ted by preparing solutions containing various CuCl concentra- tions in a standard 1 rnol dm-3 CuCI2 solution.If the Cull concentration of the sample is between 0.1 and 1.0 mol dm-3, the amount of NH3 (in mol) should be 4.6 times that of Cut1 (15% in excess of the stoichiometric amount required for the tetraammine complex) and hence the volume of buffer should be adjusted accordingly. Recommended Procedure for the Determination of Reducing Agents A 1 ml aliquot of a 0.1 mol dm-3 Cu11 solution was placed in a test-tube and the flow of nitrogen through the capillary tube was started. Then, 2.5 ml of Nc, 1 ml of ammonium acetate, (3 - x ) ml of water and x ml of reductant (5.0 X mol dm-3) were added in this order. The absorbance at 450 nm was recorded against water 2 min after the addition of the reductant.The measured absorbances were corrected for the absorbance of the reagent blank containing no reducing agent. If the n-electron reductant did not yield a molar absorptivity of approximately 7.5n x lo3 dm3 mol-l cm-l ( i e . , n-times that of Cut-Nc) in the final (7.5 ml) solution, it was concluded that the redox reaction was not stoichiometric and the procedure was repeated by using an initial Cu" concentration of ,0.01 rnol dm-3 and/or a period of 20 min prior to measuring the absorbance. The absorbances for ten different concentrations of each reductant were measured within the concentration range in which Beer's law was obeyed. The linear equation for the quantification of each reductant was found by regression analysis of absorbance-concentration data.The mean molar absorptivities [E = (Anet/c)] and the standard deviations (s) were obtained. The value of E at the 95% confidence level was calculated from the equation: E = S k t s / G , where tO.975 = 2.26 and n = 10. Results and Discussion By using an excess of the Nc reagent in a weakly acidic or neutral solution, EDTA can prevent Cu" from forming the yellow CuII-Nc complex only for a limited period of time, whereas the CuI itself can be masked. Afterwards both CuI and Cult tend to exhibit similar absorbances. The effectiveness of EDTA in masking CU" selectively in preference to CuI is a function of pH, buffer concentration, EDTA concentration, Nc concentration and time. Ethylenediamine also masks Cull; however, selective decoloration of CuII-Nc from equimolar mixtures of CuI and Cult was not possible with ethylene- diamine at pH 7.2-7.5 (achieved with phosphate buffer) when an excess of Nc was present.At the working pH used by Ulibarri et al. ,9 i.e., pH 6-7, the value of log PI for Cu"-EDTA is actually comparable to the value of log 8 2 for CuII-Nc (log PI = 11.5 for Cu"-HY3-; log p2 = 11.7 for CulI-Nc).30 Hence, EDTA does not hinder the formation of CuIt-Nc at this pH, and this was confirmed experimentally in the present work. The suppression of the absorption due to CU" was not possible when a concentration of Nc as high as 0.1% in ethanol was maintained. Standard mixtures of CuI and CU" were prepared by adding a measured amount of NH20H (a two-electron reductant for CuII)21 to an excess of Cu".The molar absorptivity of Cul produced indirectly by this method is slightly higher than that obtained with pure CuCl (E = approximately 8.0 x 103 and 7.5 x 103 dm3 mol-1 cm-1, respectively). Similar observations have been reported in the literature, where NH20H gave rise to a molar absorptivity which was higher than expected for a two-electron reductant .28 The method developed in this work is based on the control of complex formation by the ligand by using 3.0 x 10-3 rnol dm-3 Nc in ethanol as the ligand. The values of log p2 for CuI-Nc and Cu11-N~ are 19.1 and 11.7, respectively.30 Hence CuII-Nc cannot be formed in a concentrated Cu't solution within 2 min with such a low ligand concentration, allowing the absorption of Cu1-Nc at 450 nm to be measured.If the CU" concentration is about 1 mol dm-3, masking with NH3 buffer at pH 10 is to be preferred. Masking of Cull with NH3 buffer is unnecessary for Cu" concentrations SO.1 rnol dm-3, and has Table 1 Variation of the molar absorptivity of CuI (as CuCl) with the concentration of CU" Cut] concentratiodmol dm-3 0 0.001 0.01 0.1 d103 dm3mol-1 cm-1 7.5 5.8 4.7 4.5 0.6 I I I 0.6 1.0 1.4 1.8 Buffer/mI Fig. 1 Variation of absorbance with volume of NH3 buffer. 1.0 x mmol of NH20H added to 1 mmol of Cu". Nc: 2.5 ml of a 3.0 X 10-3 mol dm-3 solution in ethanol, total volume = 7.5 ml. A , Sample; B, reagent blank; and C, net absorbance at 450 nm against waterTable 2 Determination of reductants with CuII-Nc Reductant Hydrogen peroxide$ Ascorbic acid$ CysteineO Hydroxylamine hydrochlorides Sodium dithionites Mitoxantrone dihydrochloride$ Glutathiones Hydrazine dihydrochloride$ Fell (in the form of Mohr's salt)$ Carminic acid$ Thiourea7 ThiosulphateQ SulphiteP Tin(n) chloride$ No.of electrons per reductant 2 2 1 2 2 2 1 4 1 4 4 1 2 2 Sodium tetrahydroborate(ni)§ 8(?) 2,3-Dimercaptopropan-l-ol$ 2 Absorbance- concentration line equation* A = 1.48 X 1 0 4 ~ - 0.010 Az1.56X 104~-0.005 A = 7.48 X 103~ - 0.006 A = 1.76 x 104c + 0.004 A = 1.54 X 10% + 0.042 A = 8.51 X 10% + 0.007 A = 7.49 x 10% + 0.007 A = 2.55 X 10% + 0.023 A = 3.19 x 104c + 0.041 A = 7.41 x 103c + 0.002 A = 1.36 X 104c + 0.029 A = 7.69 x 103~ + 0.045 A = 1.63 X 104~ - 0.021 A = 3.07 X 1 0 4 ~ - 0.003 A = 1.84 x 10% + 0.074 A = 4.31 x 104c + 0.035 A = 1.19 x 104c + 0.032 Correlation coefficient 0.9999 0.9999 0.9992 0.9990 0.9991 0.9999 0.9994 0.9992 0.9997 0.9996 0.9995 0.9999 0.9999 0.9998 0.9990 0.9995 0.9995 * A = net absorbance; c = concentration of reductant in the final solution (mol dm-3).t With respect to whether the reaction can be used for quantification of the reductant. ddm3 mol-1 cm-1 (1.56 -t 0.02) x 104 (7.45 f 0.23) x 103 (1.75 k 0.06) x 104 (1.59 f 0.12) x 104 (1.62 -t 0.06) x 104 (8.70 k 0.33) x 103 (3.05 f 0.08) x 104 (7.51 k 0.22) x 103 (2.66 f 0.09) x 104 (3.40 -t 0.12) x 104 (7.55 * 0.36) x 103 (1.43 f 0.08) x 104 (8.29 k 0.52) x 103 (1.44 f 0.04) x 104 (2.05 k 0.16) x 104 (4.54 f 0.27) x 104 (1.31 _t 0.07) x 104 $ Reaction' conditions for colour development: 0.01 moidm-3 Cull medium; absorbance measured within 2 min.$ Reaction conditions for colour development: 0.1 mol dm-3 Cull medium; absorbance measured within 2 min. 7 Reaction conditions for colour development: 0.01 mol dm-3 Curi medium; absorbance measured within 20 min. Relative standard deviation 3.9 1.2 2.9 2.8 7.3 4.0 4.5 2.1 3.1 3.9 3.7 4.5 4.5 5.0 (S/E) (Yo) 6.2 5.6 6.2 Linear range/ mol dm-3 0-1.4 x 10-4 0-1.3 x 10-4 0-2.7 x 10-4 6.67 x 10-6 G c 9 1.1 x 10-4 0-1.0 x 10-4 0-1.1 x 10-4 0-6.3 x 10-5 0-2.6 x 10-4 0-6.7 x 10-5 0-2.7 x 10-4 6.0 x < c d 2.3 x 10-4 04.7 x 10-5 6.6 x 10-6 d c d 1.3 x 10-4 6.6 x 10-6 d c d 2.0 x 10-4 e 9 . 3 x 10-5 0-4.0 x 10-5 6.6 x 10-6 d c d 1.0 x 10-4 Comments? Quantitative Quantitative Quantitative Quantitative Quantitative Quantitative Quantitative Quantitative Quantitative Incomplete, but suitable Quantitative Quantitative Quantitative Incomplete reaction; can be quantified by an empirical equation Incomplete reaction; can be quantified by an empirical equation Incomplete reaction; can be quantified by an empirical equation Incomplete reaction; can be quantified by an empirical equationANALYST, JANUARY 1991, VOL.116 93 an adverse effect on the molar absorptivity. The half-cell potential of the [CU(NH~)~]~+-[CU(NH~)~]+ couple becomes more negative as the [CU(NH~)~]+ or NH3 concentration increases, favouring the oxidation of Cul species. Hence, the decrease in the molar absorptivity accompanied by deviations from linearity of the absorbance versus concentration graphs observed with increasing CuI or NH3 concentration might be due to oxidation before chelation by Nc.With synthetic mixtures obtained by the addition of trace amounts of CuCl to an excess of CuCI2, the absorbance decreases and deviations from linearity become more marked as NH3 is added. Moreover, the molar absorptivity for CuI decreases with increasing Cull concentration in the absence of NH3 (Table 1). For a fixed amount of Cu1, i.e., 2.0 X 10-4 mmol, in excess of CU" (1 mmol) solution, the variation of the blank, sample and net absorbances with the volume of NH3 buffer added is illustrated in Fig. 1. The net absorbance decreases signifi- cantly when the NH3 concentration is greater than the stoichiometric concentration required for the formation of [ C U ( N H ~ ) ~ ] ~ + . A slight excess of NH3 buffer was found to be optimum for analysis in order to prevent precipitation.(The use of 0.65 ml of NH3 buffer allows a value for E of 7.7 X lo3 dm3 mol-1 cm-1 to be obtained for Cu'.) Copper(1) alone, as CuCl without Cull, in 0.65 ml of NH3 buffer gives an increase in absorbance with concentration that deviates slightly from linearity; however, when CuI (1.0 X 10-44.0 x 10-4 mmol) is either protected by a 10-fold excess of NH20H or used in the form of a CU(NC)~C~ stock solution, the maximum molar absorptivity of 7.7 x 103 dm3 mol-1 cm-1 can be attained even in the presence of the same amount of buffer (0.65 ml). This observation shows that Cul is not masked, but is probably oxidized in NH3 buffer when NHZOH is not used. Of the potential reducing agents, H202, ascorbic acid, cysteine, NH20H, N2H4, S2O42-, S2032-, mitoxantrone, glutathione, FeII and thiourea yield molar absorptivities corresponding to stoichiometric n-electron oxidations of these substances (Table 2).Carminic acid, sulphite, tin(I1) chloride, 2,4-dinitro- phenyl h y drazine , NaBH4 and 2,3-dimercaptopropan- 1-01 (BAL) are incompletely oxidized, but can be quantified by the aid of empirical equations in the linear absorbance-concentra- tion range. Citric acid, an interferent in the procedure of Lau et a1.,31 which hinders the Cu"-catalysed oxidation of ascorbic acid, does not pose a problem in the CuII-Nc system. Ascorbic acid was determined successfully in the presence of a 20-fold amount of citrate. Cysteine undergoes a stoichiometric one-electron oxidation in the CuII-Nc system, and the rapid oxidation of cysteine by the Cull-Nc reagent is not affected by the Cu-catalysed auto-oxidation of cysteine.32 The ease of oxidation of the potent anticancer drug, mitoxantrone, with the Cu11-N~ reagent might provide sup- port for its oxidative metabolism; the mechanism of its antitumour action is still not fully known.However, mitoxan- trone has recently been shown to be oxidized enzymically to the iminoquinone form via irreversible and reversible steps, the whole oxidation process involving two electrons.33 The slow reaction of CuII-Nc with thiourea might be due to the formation of a stable complex between this substance and the reduction product of the system, i.e., Cu', with log p3 = 13 and log p4 = 15.4.34 Nevertheless, the more stable Cul-Nc complex predominates given sufficient time (20 min), thus allowing the theoretical molar absorptivity typical of a four-electron reductant to be attained.Hexacyanoferrate(I1) and iodide could not be quantified in the CuII-Nc system because of their undesirable reactions with CU" and C U ~ , respectively. Glucose (and reducing sugars with the -(CHOH),-CHO group), oxalate, citrate, phosphate, arsenite and thiocyanate did not react, demonstrating that the Cull-Nc system is selective to a limited extent and might be of potential use in the quantification of other reductants in biological fluids with minimal interference from the major cons ti tuen ts . The potential utility of the proposed system has not been investigated fully as other reducing agents capable of being oxidized by Cu"-Nc might also be quantified.In principle, the Cu"-Nc system allows the spectropho- tometric determination of a reducing agent, Ared, provided that the redox reaction nCu2+ + 2nNc + Ared $ n[Cu(Nc)2]+ + A,, (1) is complete with the formation of an equivalent amount of [CU(NC)~]+ with respect to the n-electron reductant, Ared. Copper(I1) is a strong oxidizing agent only when its reduction product, CuI, is stabilized by a strong complex-forming ligand, e.g., Nc. The standard potential of the Cu2+-Cu+ couple (0.17 V) is shifted to more positive values by preferential complexation of Cu'. Even strong oxidizing agents such as H202, which are weak reductants, can be oxidized in such a system. For example, when 7.5 x 10-3 mmol of Nc and 5.0 X 10-5 mmol of reductant are used, the cell potential appears to be less favourable and the reduction is incomplete for H202 in the presence of 0.1 mmol of Cull.The same reasoning applies to the stronger reductant, NH20H, in the presence of 1.0 mmol of Cu". The oxidizing power of CuII in a solution containing Nc is dependent on the ease of formation of [CU(NC)~]+. A large excess of CuJI can exhibit an affinity for Nc, thereby preventing the preferential quantitative formation of [CU(NC)~]+. The stronger the reductant, the more quantitative will be the reduction of CuII with the subsequent formation of a stoichiometric amount of this complex. On the other hand, weak reductants should be determined either by masking the excess of Cu" so that it will not compete with CuI for complex formation or by using a more dilute solution of Cu".Most n-electron reductants give a molar absorptivity that is approximately n-times that of [Cu(Nc)z]+, demonstrating that the reduction is essentially complete within 2 min. Empirical methods of determination were established for those reducing agents whose oxidation was not complete within the pre- scribed period of time provided that their absorbance- concentration plots were linear. Usually there was a useful range of Cu" concentrations for reductants, possibily as a result of favourable oxidation potentials; hence the theoretical (expected for an n-electron reduction of Cu"-Nc) molar absorptivities could be achieved. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 References Mullikan, R. S., J.Am. Chem. SOC., 1952, 74, 811. Kurbatova, I., Zavod. Lab., 1966, 32, 1064. Smathers, J. B., Duffy, D., and Laksmanan, S., Nucl. Appl. Technol., 1969, 7, 84. Larsen, E. R., Anal. Chem., 1974, 46, 1131. Organic Reagents for Trace Analysis, eds. Fries, J., and Getrost, H., Merck, Darmstadt, 1977, p. 133. Blaedel, W. J . , and Dinwiddie, D. E., Anal. Chem., 1974, 46, 873. Kwik, W. L., and Ang, K. P., J. Chem. SOC., Dalton Trans., 1981, 452. Moffett, J. W.. and Zika, R. G., Mar. Chem., 1983, 13, 239. Ulibarri, G., Ogura, T., Tzeng, J . , Scott, N., andFernando, Q., Anal. Chem., 1982, 54, 2307. Moffett, J . W., Zika, R. G., and Petasne, R. G., Anal. Chim. Acta, 1985, 175, 171. Handbook of Analytical Chemistry, ed. Meites, L., McGraw- Hill, New York, 1st edn., 1982, pp.3-210, and 3-222. Broekhuysen, J . , Anal. Chim. Acta, 1958, 19, 542. Boyer, B. D., J . Am. Chem. SOC., 1954, 76, 4331. Reddy, B. S . . and Sastry, C. S. P., Microchem. J.. 1987,36,159.94 ANALYST, JANUARY 1991, VOL. 116 15 16 17 18 19 20 21 22 23 24 25 26 27 Sastry, C. S., Satyanarayana, P., and Tummuru, M. K., Analyst, 1985, 110, 189. Simplicio, P. D., Naldini, A., and Bianco, M. T., Bull. SOC. Ital. Biol. Sper., 1984, 60, 1161. Ellman, G. L., Arch. Biochem. Biophys., 1959, 82, 70. Patterson, J. W., Lazarov, A., Lemm, F. J., and Levey, S., J. Biol. Chem., 1949, 177, 197. Aldridge, W. N., Biochem. J., 1948, 42, 52. Hernandez, G., Znf. Microquim., 1956, 1, 26; Chem. Abstr., 1961, 55, 20759h. Encyclopedia of Industrial and Chemical Analysis, eds. Snell, F. D., and Ettre, L. S., Interscience, New York, 1971, vol. 14, p. 446. Ferriol, M., and Gazet, J., Anal. Chim. Acta, 1985, 174, 365. Besada, A., Talanta, 1987, 34, 731. Lau, 0. W., and Luk, S. F., J. Assoc. Off. Anal. Chem., 1987, 70, 518. Yamane, T., and Ogawa, T., Bunseki Kagaku, 1987, 36, 625. Baker, W. L., and Lowe, T., Analyst, 1985, 110, 1189. Contreras-Guzman, E. S., and Strong, F. C., Quim. Nova, 1984, 7,60. 28 Schilt, A. A., and Di Tusa, M. R., Talanta, 1982,29, 129. 29 Murdokk, K. C., Child, R. G., and Fabio, P. F., J. Med. Chem.. 1979, 22, 1024. 30 Critical Stability Constants, eds. Smith, R. M., and Martell. A. E., Plenum Press, New York, 1975, vol. 2. 31 Lau, 0.-W., Luk, S.-F., and Wong, K.-S., Analyst, 1986, 111, 665. 32 Hanaki, A., and Kamide, H., Bull. Chem. SOC. Jpn.. 1983,56, 2065. 33 Lown, J. W., Reszka, K., Kolodziejczyk, P.. and Wilson, W. D., in Molecular Mechanisms of Carcinogenic and Anti- tumor Activity, eds. Chagas, c., and Pullman, B., Pontificiae Academiae Scientiarum, Scripta Varia, October 21-25, 1986, University of Alberta, Edmonton, Canada, pp. 243-274. Lunge’s Handbook of Chemistry, ed. Dean, J. A., McGraw- Hill. New York, 12th edn., 1979, pp. 5-68. 34 Paper 0101 705A Received April 18th, I990 Accepted August 29th, 1990

ISSN:0003-2654    

DOI:10.1039/AN9911600089

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, JANUARY 1991, VOL. 116 95 Highly Sensitive Spectrophotometric Determination of Trace Amounts of Uranium(v1) With the Thiocyanate-Basic Triphenylmethane Dyes-Gum Arabic System Liu Zhongfan and Liu Shaopu Department of Chemistry, Southwest China Teachers University, Chongqing 630715, People's Republic of China Two highly sensitive spectrophotometric methods for the determination of trace amounts of uranium(v1) have been developed, based on its colour reactions with thiocyanate and triphenylmethane dyes (TPMD) in aqueous solution in the presence of gum arabic. Uranium(vi) reacts with thiocyanate and TPMD to form ion association complexes of composition (TPMD)2[U02(SCN)4]. The molar absorptivities are between 0.81 x lo5 and 5.71 x 105 dm3 mol-1 cm-1, the highest value being found with Crystal Violet.Suitable conditionsforthe reactions, the effects of foreign ions and a pre-concentration procedure for uranium(v1) were investigated. The methods can be applied to the spectrophotometric determination of trace amounts of uranium(vi) in water and some ores. Ke yw o rds : Ura n ium (vi) determination; spectrop h o tome try; thioc yana te-trip hen ylm et h an e d ye-g um ara bic system The spectrophotometric determination of molybdenum(v) , I zinc,*J cobalt(11),4 indium(Iir),s selenium(rv),6 iron(111)7 and vanadium(v)* as the ion association complexes of the metal ion-thiocyanate complex and some basic dyes in aqueous solution in the presence of a surfactant has been reported. However, the possibility of the spectrophotometric determi- nation of uranium(v1) using an ion association complex of this type has not been studied to date.Experimentally it was found that uranium(v1) forms an anionic [U02(SCN)4]2- complex in the presence of a large excess of thiocyanate. This anionic complex reacts with some basic triphenylmethane dyes (TPMD) such as Crystal Violet (CV), Malachite Green (MG), Brilliant Green (BG), Iodine Green (IG) and Ethyl Violet (EV) to form association complexes of the type (TPMD)2[U02(SCN)4]. These complexes can be kept in solution, avoiding precipitation in the presence of gum arabic. These colour reactions have very high sensitivity, their molar absorptivities are between 0.81 X 10s and 5.71 x l o 5 dm3 mol-1 cm-1 depending on the dyes used and the experimental conditions; the highest sensitivities were obtained with the CV and MG systems.Suitable conditions for the colour reactions, the effects of foreign ions and the pre-concentration procedure for uranium(v1) were studied and a highly sensitive spectrophotometric method for the determination of trace amounts of uranium(v1) in water and some ore samples is proposed. Experimental Reagents Sulphuric acid, 0.5 and 1.0 mol dm-3. Sodium thiocyanate solution, 15% mfv. Gum arabic solution, 1% mlv. Basic triphenylmethane dye solutions. Crystal Violet (analytical-reagent grade, Beijing Chemical Plant), 0.05% ; MG (chemically pure reagent, Beijing Chemical Plant), 0.05% ; BG (chemically pure reagent, Shanghai Third Chemical Reagent Plant), 0.05%; IG (BDH), 0.05%; and EV (Fluka), 0.025%.Potassium aluminium sulphate solution, 8.8% mlv. Dissolve 22 g of KAI(S04)2.12H20 in water and dilute to 250 ml with water. This solution contains 5 mg ml-1 of aluminium. Standard uranium( vz) solution. Dissolve 0.1782 g of uranyl acetate (guaranteed-reagent grade, BDH) in water. Transfer this solution into a 1000 ml calibrated flask, then dilute to the mark with water. Dilute further to 5 pg ml-1 as required. Apparatus A 721 spectrophotometer (Third Analytical Instrument Fac- tory, Shanghai), wavelength range 360-800 nm, was employed for absorbance measurements. Procedures Place 10 pg of uranium(v1) in a 25 ml calibrated flask. Add suitable amounts of sulphuric acid (1 mol dm-3) and sodium thiocyanate solution and set aside for 10 min. In the following order, add 1 ml of gum arabic solution and 4-7 ml of basic dye solution depending on the dye used.Dilute to the mark with water and mix. Measure the absorbance in a 1 cm cell at the maximum absorbance wavelength against the reagent blank solution. 1.0 A 450 500 550 600 650 700 ?Jn m Fig. 1 Absorption spectra of the ion association complexes. A, CV system; B, MG system; C, BG system; D, IG system; and E, EV system, [UVI] = 10 pg per 25 ml. A 1 cm cell was used and measured against the blank96 ANALYST, JANUARY 1991, VOL. 116 1 .o 0.8 8 .e 8 5 0.6 a 4 0.4 0.2 0 x c b -X \x I 0.1 0.2 Acidity of HsSOdmol dm-3 Fig. 2 MG system; C, BG system; D, IG system; and E, EV system Effect of solution acidity on absorbance. A, CV system; B, Results and Discussion Spectral Characteristics Fig.1 shows the absorption spectra of the ion association complexes of uranium(vI) with thiocyanate and five basic triphenylmethane dyes. The maximum absorption wavelengths are at 540 and 565 nm for the ion associate systems of CV and EV, respectively. However the MG, BG and IG systems also have two absorption peaks, the maximum absorption wavelengths are at 640 (MG system) and 600 nm (BG and IG systems). 1 .o 0.8 b, m 2 0.6 e $ 2 0.4 0.2 1 0 10 20 30 40 50 60 Ti me/m in Fig. 3 Effects of temperature on the absorbance of the CV system. A, 15; B, 18; C, 20; D, 25; and E, 30 "C Effect of Solution Acidity Fig. 2 shows the dependence of the absorbance of the ion association complexes on solution acidity. All the ion asso- ciation reactions are carried out in acidic media, but the acidities differ according to the system.The optimum acidity is between 0.10 and 0.16 mol dm-3 sulphuric acid for MG and IG systems, 0.03 and 0.08 mol dm-3 sulphuric acid for CV and BG systems and 0.06 and 0.08 mol dm-3 sulphuric acid for the EV system. Other acids such as hydrochloric, nitric and phosphoric acids were tested, but were found to be unsuitable. The use of nitric acid caused a rapid decrease of the absorbance because of the oxidation of the dyes by this acid. With hydrochloric and phosphoric acids, the chloride and phosphate ions formed complexes with uranium(v1). The highest and most stable absorbance was obtained with sulphuric acid; hence sulphuric acid was used in this work. The MG and IG systems have the widest and the EV system the narrowest acidity range for the systems mentioned above.Effects of Reagent Concentration A higher thiocyanate concentration is required for complete formation of the anionic [U02(SCN)4] complex. The optimum concentration of sodium thiocyanate is between 0.15 and 0.30 mol dm-3 for all systems with the exception of the MG system for which the optimum concentration is between 0.45 and 0.60 dm-3 mol. The optimum concentrations of the basic dyes are as follows: (2.0-2.5) x 10-4 mol dm-3 of CV; (0.84-1.7) X mol dm-3 of MG; (2.1-2.9) X 10-4mol dm-3 of BG; (1.8-2.4) x 10-4 mol dm-3 of IG; and (4.0-8.4) x 10-5 mol dm-3 of EV, separately. Effects of Surfactants and Stability of the Absorbance In the absence of a surfactant, the ion-association complexes will precipitate out of the aqueous solution due to their hydrophobicities. When a surfactant is added or a colloid 0 20 30 40 50 60 Time/min Fig.4 Effects of temperature on the absorbance of the MG system. A, 15; B, 20; C, 25; and D, 30 "C stabilizer such as gum arabic, gelatin, poly(viny1 alcohol), Triton X-100 and Tween 20, the solution remains clear and there is a marked colour change. The effects of surfactants on the absorption spectra characteristics are not marked, but their effects on the sensitivity and stability of the colour reaction are different. Gum arabic is the best in all systems, however, a gum arabic-Triton X-100 solution is used in the EV system. Figs. 3 and 4 show the effects of temperature on the speed and stability of colour development for the CV and MG systems in aqueous solution in the presence of gum arabic.With an increase in the temperature, colour development is faster and the stability of the absorbance is reduced. For example, for the colour to develop fully a period of 5 min is necessary; the absorbance is stable for about 40 min when the temperature is at 15 "C. However, colour development is instantaneous, but the absorbance is only stable for 5 min at 30 "C for the CV system. For the MG system, the effect of temperature was similar to that of the CV system (see Fig. 4). Therefore, the measurement of the absorbance should be carried out immediately at higher than ambient temperatures. Composition of the Ion Associates The molar ratio of uranium(v1) to SCN was 1 : 4 as established by the equilibrium shift method and uranium(v1) to CV or MG 1 : 2 as established by Job's method of continuous variation and the equilibrium shift method. Owing to the uranium(v1) ion, the main species in acidic solution is U022+, therefore the composition of the ion association complexes in the CV and MG systems may be inferred to be CV2[U02(SCN)4] and MG2[U02(SCN)4].ANALYST, JANUARY 1991, VOL.116 97 Table 1 Sensitivity of various methods for the spectrophotometric determination of uranium Reagent* Arsenazo( 111) Chlorophosphonazo(II1) PAR TAR Pyrogallol Red Thorin Dibromophenylfluorone- 5-Br-PAD AP-sal-CPB Benzoic acid-Rhodamine B CAS-CTM AC-PY Tween 20 Benzoic acid-MG SCN-CV-gum arabic SCN-MG-gum arabic pHlacidit y 5.0-7.5 m ~ l d m - ~ HN03 or HC1 0.5-4.0 mol dm-3 HCl 6.8-10.7 7.5-7.8 6.0-7.0 4.0-5.5 3.0-7.0 6.5-9.0 5 5 5 .9 4.5 6.0 0.03-0.08 mol dm-3 0.10-4.16 mol dm-3 H2S04 H2S04 ~ r n a h m 655 670 550 540-545 650 625 600 565 595 555 635 545 640 &/lo4 dm3 mol-l cm-l 6.0 7.36 4.25 (Extracted into CHCI3) 3.30 2.9 10.0 2.12 17 8.8 10.3 (Extracted into benzeneaieth yl ether-2-methylpentan-1-one 8.3 (Extracted into cyclohexane) 57 37.4 Reference 9 9 10 11 12 13 14 15 16 17 18 This work This work * PAR = 4-(2-pyridylazo)resorcinol; TAR = 4-(2'-thiazolylazo)resorcinol; CAS = Chrome Azurol S; CTMAC = cetyltrimethylammonium chloride; PY = pyridine; 5-Br-PADAP = 2-[2-(5-bromopyridyl)azo]-5-diethylaminophenol; sal = salicylic acid; and CPB = cetylpyridinium bromide. Table 2 Tolerated amounts of foreign ions in the determination of uranium(v1) with the CV system (Uvl: 5 pg; relative deviation t5%) Foreign Amount Foreign Amount ion tolerated/yg ion tolerated/yg 2000 1000 1000 200 200 200 100 100 100 100 50 50 50 50 30 30 20 20 20,100* 20, so* 10 10 10 10 10,100* 10, 40* 5,400" 5 - loo* - 20* - 1000* - 20* - 20* * 1 ml of 0.1 mol dm-3 EDTA solution and 1 ml of 5% thiourea solution added.Sensitivities and Selectivities of the Methods Different amounts of uranium(v1) were used for colour development under optimum conditions, the measurement of absorbance and the construction of calibration graphs. The molar absorptivities of the colour reactions were calculated and the following values found: E = 5.71 x 105 dm3 mol-1 cm-1 for the CV system; E = 3.74 x l o 5 dm3 mol-1 cm-1 for the MG system; E = 3.27 x lo5 dm3 mol-1 cm-1 for the BG system; E = 2.74 x 105 dm3 mol-1 cm-1 for the IG system; and E = 8.03 x 104 dm3 mol-1 cm-1 for the EV system.The con- centration ranges of uranium(v1) obeying Beer's law were 0-15 pg per 25 ml for CV, MG, BG and IG systems and 0-20 yg per 25 ml for the EV system. As can be seen from Table 1 the methods are highly sensitive for the spectrophotometric determination of ur- anium. The selectivity of the method using the CV system was investigated in the determination of 5 pg of uranium(v1) in the presence of a series of other ions. Table 2 shows the tolerance of foreign ions in the determination of uranium(v1). The main interfering ions are iron(m) , molybdenum(v1) , zinc, cobalt(Ir), indium(m), vanadium(v), mercury(n).The addition of 1 ml of 0.1 mol dm-3 EDTA solution and 1 ml of 5% thiourea solution can improve the selectivity of the method. If the absorbance is measured against a solution containing the same amount of aluminium, then, up to 20 mg of aluminium were found not to interfere. Hence, pre-concentration of uranium(v1) by the Al( OH), coprecipitation method greatly increases the selec- tivity of the method and it can be applied to the spectropho- tometric determination of trace amounts of uranium(v1) in the presence of many metal ions. The selectivity of the MG system was also investigated and found to be very similar to that of the CV system and the selectivity can also be improved by similar procedures. Spectrophotometric Determination of Trace Amounts of Uranium(v1) in Synthetic Water Samples Place 100 ml of the synthetic water sample in a 250 ml beaker, add 1 ml of nitric acid (1 + 1) and heat to boiling, then add 3 ml of 8.8% m/v potassium aluminium sulphate solution and 1 drop of Methyl Red solution.Add ammonia solution (1 + 1) dropwise until the solution changes colour from red to yellow and set aside €or 3-4 h. Filter the precipitate and wash it several times with hot water. Crystal Violet method Dissolve the precipitate in a 25 ml calibrated flask by adding 2.5 ml of 0.5 mol dm-3 sulphuric acid and heating, and wash a few times with up to 10 ml of hot water. In the following order add 1 ml of 0.1 mol EDTA solution, 1 ml of 5% thiourea solution, 3 ml of sodium thiocyanate solution, mix and then set aside for 10 min.Then add 1 ml of gum arabic solution, 4 ml of CV solution and dilute to the mark with water and mix. The absorbance of the solution was measured in a 1 cm cell at 545 nm, against the reagent blank. Malachite Green method The procedure is similar to that for the CV system, but differs in that the precipitate is dissolved in 3.0 ml of 1 rnol dm-3 sulphuric acid. Then 6 ml of sodium thiocyanate solution and 5 ml of MG solution are added. Finally, the absorbance of the solution is measured at 640 nm. The results obtained with both of these methods are summarized in Table 3.98 ANALYST, JANUARY 1991, VOL. 116 ~~ Table 3 Results for the determination of uranium(v1) in synthetic water samples Uranium(v1) content/mg 1-1 Sample Added 1 0.025 2 0.050 3 0.100 4 0.125 5 0.150 * Results of the CV method. 1 Results of the MG method.Found 0.023* 0.026t,O.O27t 0.0231 0.051*, 0.050* 0.049* 0.0521, 0.053 0.0491 0.102*. 0.101* 0.099* 0.0991,O.lOlt 0.1021 0.124*, 0.126* 0.123" 0.149*, 0.148* 0.147" 0.1477,O. 1461 0.151T Elements other than uranium(vI)/mg 1-1 PbII 0.6, Cd1I 0.6, Wvl 0.2, Ca" 50 MglI 9, Bill1 0.08, TilV0.6, Hg" 0.2 TI111 0.3, ThIV 0.4, ZrIV 0.5, Be 1, CrIII 5 , InlI1 0.2, TeIV 0.3, Ga1I1 0.4, CulI 0.1, Mn" 1, As111 0.15, Zn" 0.2, Fe"12, NiII0.5, MoVI0.1, Sb"I0.3, Bat110 FeIII 3, Cr"I4, SnIV 0.4, Vv0.6 Pb" 0.7, Cd" 0.5, Wvl 0.7, Ca" 60, Mg" 8, BPI 0.06. TiIV 0.1, HgI1 0.1, T11110.4, ThIV 0.5, ZrIV 0.6, Be"0.2, InIII0.3, Cu"0.2 Cut1 0.3, FelI1 5 , CrlI1 4, SnIV 0.5 Vv 0.1, Pb" 0.8, Cd" 0.2, WvI 0.3, Call Mg" 7, Bill1 0.1, TiIV0.2, HgII 0.3, TI111 0.5, ThIV 0.6, Z P 0.1 , Be11 0.2, In111 0.3, TeIV 0.5, Ga"' 0.2, MnII 2.5, As111 0.5, Zn" 0.05, Nil1 0.3, MoV1 0.03, SbIII0.2, Ba"5 Cu" 0.45, FeIII 4, Cr"' 2, SnIV 0.6 Vv 0.4, Pb" 0.4, Cd" 0.5, Wvl 0.6, Ca" 1, Mg" 0.4, Bill1 0.25, TiIV 0.5, Hg" 0.5, T P 0.2, ZrIV 0.4, ThIV 0.1, Be" 0.6, In111 0.1, TeIV 0.2, GalI1 0.5, Mn" 2, As111 0.1, Zn" 0.1, Nil1 0.1, Sb"' 0.5, BaII 1 CuII 0.4, Fell' 7, CrIII 4.5, SntV 0.2 VV 0.2, Pb" 0.3, Cd" 0.4, WvI 0.4, CaII7, MgII5, Bi1110.15,Ti1V0.3, HgII 0.4, T P , 0.08, ThIV 0.5, ZrIV 0.2, Be11 0.3, In111 0.4, TeIv 0.5, GaIII 0.2, Mn" 2.5, AslI1 0.5, ZnlI 0.15, Ni" 0.2, MoVI 0.15, Ba" 7 Table 4 Results for the determination of uranium in ores Uraniumcontent (YO) Uranium content (YO) Sample Certified Found* Sample Certified Found* 85-1-01 0.0005 0.00055 rock 72-12-01 0.0011 0.0010 85-1-02 0.0009 0.00092 rock 72-12-02 0.0058 0.0054 Granite Phosphonate Granite Phosphonate * Average of three determinations.Spectrophotometric Determination of Uranium in Ores With the Crystal Violet System Dissolve 0.1-1 g of sample ore (containing 1-10 pg of uranium) in a platinum crucible with 5-10 ml of nitric acid and 5 ml of hydrochloric acid and heat. Add 2-5 ml of hydrofluoric acid and heat gently to near dryness. Add 1-2 ml of sulphuric acid and heat until sulphuric acid fumes are no longer evolved. Repeat the treatment with hydrofluoric acid and sulphuric acid once or twice. To the residue add 1 ml of nitric acid and 5-10 ml of hot water and pass the solution through a filter into a 250 ml beaker, then wash the residue several times with hot water.Dilute to about 150 ml with water and add 2 ml of 8.8% m/v potassium aluminium sulphate solution and 1 drop of Methyl Red solution. Then heat the solution to boiling, add ammonia solution (1 + 1) dropwise until the solution changes colour from red to yellow and set aside for 3 4 h. Filter the precipitate and wash several times with hot water. The remainder of the procedure is the same as for the water samples. Results are presented in Table 4. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 References Liu, S.-p., Liu, Y., and Lihua, J.-y., Phys. Test. Chem. Anal., Part B, 1980, 16.24. Liu, S.-p., and Liu, Z.-f., Huaxue Shiji, 1981,3, 20. Liu, S.-p., and Liu, Z.-f., Huaxue Tongbao, 1982,4, 208. Zhan, C-q., and Liu, S-p., Huaxue Shiji, 1983,5, 117. Li, Z., Xu, Q.-h., and Yejin, F.-xi. Metufl. Anal., 1987, 7, 4. Li, X.-m., and Liu, Y., Fenxi Huaxue, 1988, 16, 25. Liu, Z.-f., Liu, S.-p., and Xie, C., J. Southwest Chin. Teachers Univ., 1989, 14, 105. Liu, S.-p., and Liu, Z.-f., Chem. J . Chin. Univ., 1989, 10, 800. Busev, A. I . . Tiptsova, V. G., and Ivanov, V. M., Analytical Chemistry of Rare Elements, MLR Publishers, Moscow, 1981, Shijo, Y., and Sakai, K . , Bull. Chem. Soc. Jpn., 1978.51,2574. Sommer, L., and Ivonov, V. M., Talanta, 1967, 14, 171. Sucmanova-Vondrova, M., Have], J., and Sommer, L., Collect. Czech. Chem. Commun., 1977,42, 1812. Shijo, Y . , and Takeuchi, T., Bunseki Kugaku, 1971,20, 297. Shibata, S . , and Matsumae, I., Bull. Chem. SOC. Jpn., 1959,32, 279. Wang, D.-j., Xiao, Z.-j., Wu, P., and Pan Q.-h., Chem. J. Chin. Univ., 1985, 6, 980. Sun, S.-s., and Shen, X.-t., Huaxue, 1984, 6, 48. MoeKen, H. H. ph., and Van Neste, W. A. H., Anal. Chim. Acta, 1967. 37, 480. Dubey, S . C., and Nadkarni, M. N., Talanta, 1977, 24,266. pp. 129-133. Paper 0101252A Received March 21st, 1990 Accepted July 27th, 1990 This project was supported by the National Natural Science Foundation of China.

ISSN:0003-2654    

DOI:10.1039/AN9911600095

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, JANUARY 1991, VOL. 116 99 Fluorescence Reaction of Sodium 7-Phenylazo-8-aminoquinoline-5- sulphonate With Gold and its Analytical Application Ruan Chuanmin and Xu Qiheng" Department of Chemistry, University of Yunnan, Kunming 650091, People's Republic of China Sodium 7-phenylazo-8-aminoquinoline-5-sulphonate has been synthesized and its identity confirmed by infrared spectrometry, thermogravimetry and elemental analysis. This compound reacts with gold(iii) in slightly acidic media forming a red complex which has an intense fluorescence at h,,/h,, = 380 nm/325 nm. The fluorescence intensity is proportional to the concentration of gold(ii1) in the range 0-320 ppb with a detection limit of 0.5 ppb. The method has been used for the determination of trace amounts of gold in mines.Keywords: Gold determination; sodium 7-phen ylazo-8-aminoquinoline-5-suIphonate; spectrofluorimetry Both 8-hydroxy- (8HQ) and 8-mercaptoquinoline (8MQ) are known reagents which are widely used as chelating agents in analytical chemistry. 1 However, the 8-amino derivative of quinoline (8AQ) has received little analytical attention, resulting in only a few published papers.2.3 The 8-amino derivative of quinoline with (N,N) as its chelating atom is more selective than 8HQ(N,O) and 8MQ(N,S). In recent years, a series of 8-aminoquinoline-5-azo derivatives have been synthesized and used in spectrofluorimetric and spectro- photometric analysis.4.s As part of an investigation of the 8AQ chelating system, this paper reports the synthesis of the fluorescent reagent, sodium 7-phenylazo-8-aminoquinoline-5-sulphonate (SPAAQ) , its fluorescence reaction with gold(II1) and its use for the spectrofluorimetric determination of trace amounts of gold(m).Experimental Apparatus The fluorescence spectra and intensities were measured with a Shimadzu (Japan) RF-540 spectrofluorimeter in 10 x 10 mm quartz cells. Reagents 8-Aminoquinoline-5-sulphonate ( A Q S ) . This was syn- thesized from 8-hydroxyquinoline .6 Sodium 7-phenylazo-8-aminoquinoline-5-sulphonate. This was synthesized by the following procedure. Aniline (0.5 g) was dissolved in 3 ml of ice-cold concentrated hydrochloric acid and 10 ml of ice-cold doubly distilled water and slowly diazotized with a solution of 0.4 g of sodium nitrite in 5 ml of water. The diazotized solution was then added dropwise with stirring to an ice-cold solution of AQS (1.1 g) in 40 mi of 2 mol dm-3 acetic acid; 2 mol dm-3 sodium hydroxide was added to keep the pH constant and the mixture left for 1 h, with stirring in the ice-bath, then neutralized with sodium hydroxide and filtered.The red precipitate was recrystallized several times from 95% ethanol to give a yield of 20% with a melting point of 239-241 "C. Elemental analysis was as follows, calculated: C, 48.91; H, 3.53; N, 15.21; and S, 8.69. Found: C, 48.34; H, 3.71; N, 14.81; and S, 8.72%. The data obtained from thermogravimetry (TG), infrared (IR) and nuclear magnetic resonance (NMR) spectra confirmed the structure of SPAAQ to be as shown in Fig. 1. SPAAQ solution, 1 x 10-4 rnol dm-3. This was prepared by dissolving 0.0368 g of the reagent in 1 1 of ethanol.The solution is stable for several months. * To whom correspondence should be addressed. S03Na N H2 Fig. 1 IR and NMR spectra Structure of SPAAQ confirmed by data obtained from TG, Gold(111) standard solution, 100 pg ml-1. This was prepared by dissolving gold in aqua regia. Working standards were prepared from this solution as required. Buffer solution (PH about 3.6). A mixture of 50 ml of 0.2 rnol dm-3 potassium hydrogen phthalate solution and 6 ml of 0.2 mol dm-3 hydrochloric acid. All other chemicals used were of analytical-reagent grade. Procedure for the Determination of Gold(m) To a sample solution in a 25 ml calibrated flask, add 0.4 ml of 1 x 10-4 rnol dm-3 SPAAQ and 3 ml of buffer solution (pH 3.6).Dilute to the mark with distilled water and mix well, heat for 20 min in boiling water and cool for 5 min, then measure the relative fluorescence intensity at he, = 380 nm and he, = 325 nm. Results and Discussion Fluorescence Spectra The excitation and emission spectra with maxima at 325 and 380 nm, respectively, of the reagent blank and the complex are shown in Fig. 2. Effect of pH The effect of the pH on the relative fluorescence intensity was studied; the optimum pH range was found to be between 2.5 and 4.0, hence the pH was fixed in the optimum interval with a buffer solution of potassium hydrogen phthalate-hydrochloric acid of pH 3.6. Effect of Amount of SPAAQ The maximum fluorescence intensity was obtained in the concentration range 1.2 x 10-7-2.0 x 10-7 mol dm-3 of SPAAQ solution. Higher reagent concentrations caused a decrease in the fluorescence intensity, when 0.4 ml of 1.0 x rnol dm-3 of reagent was used.The reason for this effect has not been studied.100 ANALYST, JANUARY 1991, VOL. 116 t 250 300 350 400 Wavelengthhm Fig. 2 Excitation and emission spectra of the gold complex (A and B) and the reagent blank (A' and B') Effect of Heating Time Without heating, the SPAAQ-gold(m) complex has a low fluorescence intensity. When heat is applied for 15-40 min the relative fluorescence intensity attains a constant maximum value. In this experiment a time of 20 min was selected. Composition of Complex The composition of the complex was determined by Job's method of continuous variation and by the molar ratio method.The molar ratio of gold to SPAAQ was found to be 1 : 1. The apparent stability constant of the complex was calculated from the results of the molar ratio and Job's method. An average value of log K = 8.1 k 0.1 was obtained at 20 "C. Effect of Foreign Ions The effect of foreign ions on the determination of 2 pg of gold(m) is summarized in Table 1. The limiting value of the concentration for each ion was taken as that value which caused an error of not more than 5% in the fluorescence intensity. The positive interference can be attributed to the fact that those elements also form complexes with SPAAQ in slightly acidic solutions. Platinum(II), Fe" and Fe"' cause a serious positive interference. Calibration Graphs If the recommended conditions are used, a linear relationship is found between the emitted fluorescence intensity and gold concentration in the range 0-320 ppb, with a detection limit of 0.5 ppb.Determination of Gold in Minerals A 2-5 g amount of sample was transferred into a 200 ml beaker and 30-50 ml of concentrated hydrochloric acid were added. The beaker was covered and heated gently to dissolve the sample. About 50 ml of distilled water were added and the mixture was filtered. The residue was dissolved in aqua regia and boiled twice almost to dryness with distilled water to Table 1 Interference of other ions in the determination of gold. [The concentration of gold(rrr) was 2 pg per 25 ml] Tolerance limit (M"+/Au) Ion 2 Pt2+, Fez+, Fe3+ 4 Ga3+ SO 100 Co2+, Pb2+ 500 lo00 Mn2+, Cu2+, Cr3+, Zn2+, Ag+ Hg2+, Ni2+ , Wv', MoV1, Al3+ Sn4+, CrVI, F-, Pod3-, C2042- Table 2 Determination of gold in minerals Sample 1 Sample 2 Gold content*/g per tonne 4.33 34.20 Gold found/g per tonne 4.37 34.40 * Values obtained by atomic absorption spectrometry.reduce the acidity; the solutions were then transferred into 50 ml calibrated flasks. To an aliquot of this solution was added 0.4 ml of 1 x 10-4 mol dm-3 SPAAQ and the pH was adjusted (2.5-4.0) with buffer solution. Iron was masked with 1 ml of 1% NaF. The results obtained are given in Table 2. Conclusion A number of spectrofluorimetric methods have been reported for the determination of trace amounts of gold, e.g., rhoda- mines,7,8 rhodanines,g kojic acid,lO bipyridylglyoxal diphenylhydrazone,ll and 2-phenylbenzo[8,9]quinolizino- [4,5,6,7-fed]phenanthridinylium perchlorate12 have been used.However, these methods are not always convenient. One of the major problems seems to be the use of organic solvents. The proposed method can be used to determine gold(II1) directly in the aqueous phase. 1 2 3 4 5 6 7 8 9 10 11 12 References Burger, K., Organic Reagents in Metal Analysis, Pergamon Press, Oxford, 1973. Blabco, M., and Maspoch, S . , Mikrochim. Acta, 1983,111, 11. Maspoch, S . . Bartroli, J., and Blanco. M., Mikrochim. Acra, 1983, 111, 95. Fang, R., and Xu, Q.-h., Chin. J. Appl. Chem., 1988, 5 , 11. Fang, R . , and Xu, Q.-h., Chem. Reagents, 1988, 10, 218. Vorozhtzov, N. N., and Kogan, I. M., Chem. Ber., 1932, 65B, 142; Chem. Abstr., 1932, 26, 2457. Blyum, I. A., Pavlova, N. N., and Kalaupina, F. P., Zh. Anal. Khim., 1977, 26, 55. Marinenko, J . , and May, I . , Anal. Chem., 1968, 40, 1137. Podberezskaya, N. K., Shilenko, E. A., and Sheherbov, D. P., Zavod. Lab., 1970, 36, 661. Mureta, A., and Ujihara, T., Bunseki Kagaku, 1961, 10, 497. Grases, F., Garcia-Sanchez, F., and Valcarcel, M., Anal. Lett., 1979, 12, 803. Tomas, P.-R., Concepcion, S.-P., Ortuno, J. A., and Molina- Buendo, P., Analyst, 1983, 108, 733. Paper 0101 954B Received May 2nd, 1990 Accepted July 25th, 1990

ISSN:0003-2654    

DOI:10.1039/AN9911600099

出版商:RSC    

年代:1991

数据来源: RSC