摘要:

ANALYST, MAY 1992, VOL. 117 869 Rapid, High-purity Chemical Separation of Molybdenum From Iron Meteorites for Isotopic Analysis by Using Thermal Ionization Mass Spectrometry Qi-Lu and Akimasa Masuda Department of Chemistry, Faculty of Science, The University of Tokyo, Hongo 7-3- 7, Bunkyo-ku, Tokyo 773, Japan A chemical procedure has been developed, which combines both solvent extraction and anion exchange, so that microgram amounts of Mo can be cleanly, rapidly and efficiently separated from gram amounts of iron meteorites. Particular attention was directed to the complete separation of Mo from Zr and Ru. The isotopic abundance ratios of Mo can subsequently be determined with high accuracy by using thermal ionization mass spectrometry. The experiments indicate that the behaviour of Mo during solvent extraction and anion exchange is considerably different from that reported previously.In particular, it was found that there is a very narrow range of HCI concentrations within which it is possible to separate Mo from Fe by solvent extraction. The reproducibility and recovery of the method were examined by using inductively coupled plasma atomic emission spectrometry. Keywords: Molybdenum; solvent extraction; ion exchange; iron meteorite Molybdenum is a very interesting element as its seven stable isotopes reflect several effects related to nuclear physics. Investigations on the isotopic abundances of Mo in meteoritic samples can supply important information not only on the nucleosynthesis process of the early solar system but also on the decay of extinct radioisotopes of the elements adjacent in atomic number to Mo.1-5 However, for reasons discussed in the literature,5 few isotopic studies of Mo have been per- formed successfully in either terrestrial or meteoritic situa- tions.At present, information on the subject of Mo isotopes, particularly in meteorites, is limited. For these reasons, a fundamental study on the chemical and isotopic analysis of Mo in molybdenites has recently been carried out by using thermal ionization mass spectrometry (TIMS) .s In this connection, an efficient and clean chemical separation method to extract microgram amounts of Mo from gram amounts of iron meteorites was developed for the analysis of Mo isotopes with high accuracy. Basically, the choice of chemical separation procedure was made according to the concentrations of the primary chemical constituents of the iron meteorites6 and from previous work.5 It is known that the mean concentrations of Fe and Ni in different types of iron meteorite are about 91 and 8%, respectively, and that the concentration of Cu is about 0.1-0.01%. Kraus et al.7 have studied the anion-exchange behaviour of Mo"' with a strongly basic quaternary amine anion-exchange resin in HCI-HF mixtures, and Murthyl and Wetherills have applied this method to the separation of Mo from iron meteorites.Unfortunately, this method is unsatis- factory for the analysis of the isotopes of Mo in iron meteorites.9 Under the analytical conditions used, Zr and Ru could not be separated from Mo completely; in addition, the separation method is time consuming.Qureshi et al. 1 0 have carried out a systematic survey of the solvent extraction of many elements in HCI, HN03 and HCI04 at various concentrations by using the bis(2-ethyl- hexyl) hydrogen phosphate (HDEHP)syclohexane system, but few studies on the separation of Mo from a matrix such as Fe by use of this system have been reported. Therefore, in order to establish a suitable method for the isotopic study of Mo in cosmochemistry, the chemical separation of Fe, Zr, Mo and Ru by solvent extraction with the HDEHP-cyclohexane system and the anion exchange of Fe, Ni, Cu, Zr, Mo and Ru under different analytical conditions have been examined. In this work, particular attention was directed to the complete separation of Mo from Zr and Ru, as some isotopes of Zr and Ru produce isobaric interferences on the measurement of the isotope ratios of Mo.By using the analytical conditions described here, greatly improved separa- tion and recovery of microgram amounts of Mo from gram amounts of iron meteorites can be achieved for the isotopic analysis of Mo with high accuracy. Experimental Reagents The water, and the acids HF, HCI, HN03 and HCIO4 used in the analyses were purified by sub-boiling distillation. The following chemical reagents and materials were used as received: ammonia solution (containing less than 0.005 ppm of Mo), HDEHP and cyclohexane (high-performance liquid chromatography grade). Standard solutions for atomic absorption spectrometry (Aldrich, St. Louis, MO, USA), 1 mg ml-1 were used for Fe, Ni, Cu, Mo, Zr and Ru.The anion-exchange resin AG 1-X8,100-200 mesh, in the chloride form, was used. Apparatus An inductively coupled plasma atomic emission spectrometer (ICPS-5000, Shimadzu, Kyoto, Japan) was used as the detector for the solvent extraction and anion-exchange studies. An inductively coupled plasma mass spectrometer (PlasmaQuad, VG Elemental, Winsford, Cheshire, UK) was used to determine the concentrations of Mo in iron meteorites. Samples were dissolved by means of a microwave oven (MCD-81, CEM, Matthews, NC, USA). Isotopic abundances of Mo were measured with a thermal ionization mass spectrometer (VG Sector 54-30) with a single Faraday detector. Digestion of Iron Meteorites An accurately weighed amount (2-5 g) of the iron meteorite was placed in a 120 ml poly(tetrafluoroethy1ene) (PTFE) vessel and approximately 10 ml of 6 mol dm-3 HCI were added.The vessel was then sealed and heated in a microwave oven for about 20 min at a power of 2,4,5, 10 or 15%. After cooling, the vessel was opened and the clear solution was poured into a beaker. The residue was subjected to the above870 ANALYST, MAY 1992, VOL. 117 procedure several times until it was completely dissolved. The resulting solution was used for chemical separation. Solvent Extraction of Mo The experiments were carried out in an air-conditioned room at 21 k 1 "C. The HDEHP-cyclohexane system was selected for the separation of Mo from Fe, Ni, Cu and other elements in iron meteorites. Generally, after dissolution of the iron meteorites, Mo ions in HCI probably exist as various types of complexes of Mo"'.Molybdenum in the hexavalent state is usually stable; however, as it tends to hydrolyse and form an isopoly or heteropoly acid, its solvent extraction behaviour is complicated and a quantitative treatment of the extraction equilibria is not always possible. Considering that iron meteorites can be dissolved readily in 6 rnol dm-1 HCI, it was decided to investigate the effect of the HCI concentration on the extraction of the relevant elements. Normally, Zr ions are considered to be strongly partitioned into the organic phase as the Zr'" ion is a small and highly charged ion which forms a stable complex with HDEHP. Hence in subsequent work, the separation of Mo from Zr has to be performed.In order to achieve this, the back-extraction of Mo only was used. On the basis of the above studies, the procedure adopted for iron meteorites was as follows. (1) Following the dissolution of the iron meteorite, the solution was mixed with 25 ml of 0.75 rnol dm-3 HDEHP in cyclohexane in a 50 ml separating funnel. After shaking for 5 min, the aqueous phase was removed when separation of the immiscible phases was complete, and the organic phase was washed with three 25 ml portions of 5 rnol dm-3 HCIO4 and then with two 25 ml portions of 10 rnol dm-3 HN03. (2) For the back-extraction of Mo, 5 ml of 10 rnol dm-3 HN03 in 3% H202 were added to the organic phase and the mixture was shaken for about 5 min. The aqueous phase was transferred into a PTFE beaker and the back-extraction process was repeated.The aqueous phase was then evaporated to dryness at low temperature (about 70°C) and the residue was dissolved in 1 rnol dm-3 HF-O.01 rnol dm-3 HCl for further purification by means of anion exchange. Anion Exchange of Mo Further purification of Mo was undertaken by using an anion-exchange method after the solvent extraction. The following procedure was developed as a result of this study. The anion-exchange resin (AG 1 X-8, 100-200 mesh) was prepared by washing successively with concentrated HCl, water, concentrated ammonia solution, water, concentrated HN03 and finally 6 mol dm-3 HCl (the final 6 rnol dm-3 HCI stage was repeated). The resin thus prepared was placed in a 0.5 ml PTFE column for the meteoritic samples and in a 15 ml PTFE column for investigating the analytical conditions as a slurry in 1 rnol dm-3 HF-O.01 rnol dm-3 HCI.The end of the column was plugged with PTFE wool to prevent outflow of the resin. Before loading the Mo sample, extracted from the iron meteorite, onto a pre-treated 0.5 ml column, the anion- exchange resin was equilibrated with 0.01 rnol dm-3 HCI-1 .O rnol dm-3 HF by washing it with 3 ml of this solution. Then, 1 ml of 6 mol dm-3 HC1 was added to remove any Zr present. The Mo-containing fraction was subsequently collected by passing 2.0 ml of 7 rnol dm-3 HN03 through the column. For the 15 ml column, a synthetic solution was eluted by using different eluting solutions; the details are presented in the following section. Results and Discussion Solvent Extraction of Mo The solvent extraction behaviour of Mo, Fe, Zr and Ru at various concentrations of HCl with the HDEHP-cyclohexane r-- __ Zr 100 80 s .g 60 2 1 C + X w 40 20 0 4 8 12 Concentration of HCl/mol dm-3 Fig.1 Solvent extraction of Fe, Zr, Mo and Ru with the 0.75 rnol dm-3 HDEHP-cyclohexane system in different concentrations of HCI. The data are shown as percentages of each of these elements found in the aqueous phase compared with the total concentration of that element Table 1 Recovery of Mo and the effect of Fe by using solvent extraction with the HDEHP-cyclohexane system. Data were obtained by ICP-AES Mo found Fe found in a series of aqueous in back- phase washings (ppm) extract Sample (70 1 1 2 3 1. 1.oog 2. 1.01g * (iron meteorite) 88 42 1 7.3 - * - (Fe, 99.99%) 97 273 1.6 3.1.32g 4. NoFe 100 5. NoFe 99 6. NoFe 95 7. NoFe 99 * (Fe, 99.99%) 97 458 9.5 - * Below the detection limit ( 4 0 ppb). 7 9 11 13 15 Concentration of HN03/mol dm-3 Fig. 4 HCI; and B , 7 rnol dm-3 HN03 Comparison of the elution behaviour of Mo in A, 1 rnol dm-3 system is shown in Fig. 1. It can be seen that HCl has a marked effect on the extraction characteristics of Fe, only a small effect on Mo and Ru and no effect on Zr. Regardless of the concentration of HCI, the separation of Mo from the main matrix elements ( e . g . , Fe, Ni, Cu) of the iron meteorites is not possible. When the HCI concentration is about 6 rnol dm-3, Zr (100%) and Mo (>95%) will be extracted into the HDEHP-ANALYST, MAY 1992, VOL. 117 87 1 cyclohexane phase, whereas all of the Fe remains in the aqueous phase at this very narrow range of HCI concentration.The data thus obtained are not in agreement with those available in the literature,lO and these findings are of particular importance in achieving the chemical separation of 1 .o .g 0.8 L- B z 0.6 + 3 I C 0 2 0.4 .- c c c W f 0 0 0.2 0 80 160 240 Volume eluted/ml Effect of concentration of HN03 on the back-extraction of Mo Fig. 2 from the HDEHP-cyclohexane phase '0 0.4 .- e E 2 0.2 4- c e, 0 0 0.8 1.6 2.4 Volume el uted/ml Fig. 3 Study of the elution behaviour of a synthetic solution containing Fe, Ni, Cu, Mo, Z r and Ru by passing the following solutions through a 15 ml column: 1, 1 mol dm-3 HF4.01 mol dm-3 HCl; 2, 6 mol dm-3 HCl; 3, 1 mol dm-3 HCI; and 4, 14 mol dm-3 HNO3. 0.Fe; +. Ni; 0, Cu; A, Mo; X . Zr; and V Ru Table 2 Elemental abundances of Mo in iron meteorites. The results were obtained by ICP-MS Concentration of Sample Mo (PPm) Canyon Diablo 6.0 Hardesty 22.7 Odessa 15.6 not only Mo from Fe, but also Fe from the elements under consideration. Therefore, the solvent extraction of Mo from iron meteorites could be performed immediately after dissolu- tion of the iron meteorite in 6 rnol dm-3 HCI without the need for any other time-consuming chemical treatment. Inductively coupled plasma atomic emission spectrometry (ICP-AES) was used to measure the Fe present as a contaminant on the surface of the container; the Fe can be removed by washing the organic phase several times with HC104 or HN03, as shown in Table 1.Under the same conditions, Ru was partly and Zr com- pletely extracted into the organic phase. Ruthenium can be removed by washing the organic phase with HC104 o r HNO3. In the back-extraction of Mo by use of 10 rnol dm-3 HN03 in 3% H202, because Z r cannot be reduced to a low valence state, only Mo is back-extracted into the aqueous phase at this stage. The back-extraction of Mo in solutions of 3% H202 with 7, 10 and 14 rnol dm-3 HN03, respectively, is shown in Fig. 2. The results suggest that the optimum back-extraction of Mo is obtained in an aqueous phase of 10 rnol dm-3 HN03. The recovery of Mo separated by solvent extraction with the HDEHP-cyclohexane system is shown in Table 1. A 0.50 mg amount of Mo was taken as a tracer in all seven experiments, and there was no Fe in the last three samples.It can be seen that no significant matrix effect of Fe was found in the separation of Mo. It should be noted that for back-extraction using 5 ml of 10 mol dm-3 HNO3-3% H202, more than 75 and 22% of the Mo could be back-extracted in the first and second back-extractions, respectively. The whole separation process requires only about 1 h. Anion Exchange of Mo Although the solvent extraction method is capable of extract- ing Mo from iron meteorites, it is still necessary to remove isotopes of Z r (92Zr, 94Zr and 9"r) and Ru (96Ru, 98Ru and lOORu), which might affect the accuracy of the isotopic analysis of Mo (92M0, g4Mo, 96M0, 98M0 and I(H)Mo). An anion- exchange method was used for removing these isobaric interferences.It was found that earlier work138 on the measurement of Mo isotopes could be seriously affected by isotopes of Zr under the separation conditions used. In the present study, an improvement was made. In order to demonstrate the reliabil- ity of the results obtained in this work, a solution containing 1 mg each of Cu, Zr, Mo and Ru, 3.5 mg of Ni and 25 mg of Fe in 25 ml of 1.0 rnol dm-3 HF-O.01 rnol dm-3 HCI was passed through a 15 ml PTFE column, which had been' treated as described above for the 0.5 ml column. After removal of Fe, Ni and Cu with about 50 ml of 1.0 rnol dm-3 HF-O.01 rnol dm-3 HCI, Zr was removed by passing 20 ml of 6 rnol dm-3 HCI through the column. Then, about 40 ml of I rnol dm-3 HCI were added and thc eluate fraction was collected for the determination of Mo.Finally, the column was eluted with about 40 ml of 7 rnol dm-3 HN03 and this fraction was reserved for the determination of Ru. Aliquots of 4 ml were collected with an automatic fraction collector and the concentrations of these elements in the fractions were determined by ICP-AES. The experimental curves are presented in Fig. 3. Table 3 Mo isotope ratios in several iron meteorites. All isotope ratios are normalized to "MoPMo = 0.3802 by exponential law. Errors given are one standard error of the mean 95Mo/"8Mo yhMo/" Mo ')7Mo/yHMo lOOMo/WMo Sample 93 MoIYx M o Odessa 0.607961 -t 0.000054 0.656051 t 0.000050 0.688046 t 0.000070 0.394967 k 0.000039 0.400134 k 0.000054 Canyon Diablo 0.608082 k 0.000047 0.655930 +_ 0.000043 0.688156 ? 0.000044 0.394930 -t 0.000024 0.400135 f 0.000034 Hardesty 0.607900 f 0.000067 0.655954 f 0.oooO51 0.688156 k O.OOOOS6 0.394897 ? 0.000032 0.400333 t 0.000042 Molybdenite (Colarado, USA) 0.607924 k 0.000041 0.655994 -t 0.000016 0.688155 k 0.000040 0.394937 t 0.0()0022 0.400134 t 0.000016 Standard (Moo3.99.999%) 0.607926 -t 0.000013 0.655964 2 0.000013 0.688146 k 0.00001 1 0.394947 Ifr 0.000007 0.400129 ? 0.000012872 ANALYST, MAY 1992, VOL. 117 As Ru can only be eluted with concentrated HN03, no atomic interferences from Zr and Ru isotopes were found in the solution taken for the isotopic analysis of Mo, when monitored by inductively coupled plasma mass spectrometry This result shows that for a synthetic solution, the separa- tion of Zr, Mo and Ru can be achieved and the removal of Fe, Ni and Cu is also possible.However, for iron meteorites, it was found that the reproducibility and recovery of Zr, Mo and Ru were poor. This might be attributed to the effects of bulk matrix elements of the solution resulting from the use of gram amounts of the iron meteorite, which is necessary for extracting at least 20 pg of Mo for the accurate analysis of the seven isotopes of Mo by using a thermal ionization mass spectrometer with a single Faraday d e t e ~ t o r . ~ Further, accord- ing to the literature,',* and from our own experience, before loading the Mo sample onto a larger column the solution obtained from the dissolution of the iron meteorite must be evaporated to dryness at low temperature (about 70°C) and redissolved, which requires several days.Therefore, for additional purification of Mo extracted from iron meteorites after solvent extraction, a small PTFE column (0.5 ml) was used, as the purity and amount of the Mo sample obtained following solvent extraction was greatly improved. The elution curves of microgram amounts of Mo in 1 mol dm-3 HCI and 7 mol dm-3 HN03 are shown in Fig. 4. The data were obtained by ICP-AES by measuring the eluate dropwise (the 0.5 ml column was connected to the ICP-AES instrument). The use of 7 mol dm-3 HN03 rather than 1 mol dm-3 HCl is preferred because the chloride salt used for TIMS often generates more molecular and polyatomic inter- ferences than the nitrate salt. (ICP-MS). Application The abundances of Mo in a number of iron meteorites were determined by using ICP-MS.The isotopic abundances of Mo in several iron meteorites were determined by TIMS by use of the single collector method. Details of the TIMS technique have been described previously.5 The concentrations of Mo in three iron meteorites (Odessa, Canyon Diablo and Hardesty) and the isotopic abundances of Mo in the same three iron meteorites, a molybdenite sample and a standard sample (Moo3, 99.999%; Aldrich) are shown in Tables 2 and 3. The data obtained in this work for the elemental abundances of Mo in the Canyon Diablo sample show good agreement with the value reported by Murthyl (6.3 ppm). However, for the isotopic abundances of Mo in the Canyon Diablo sample, the data obtained in this work are different from those reported by Murthy. In the work of Murthy, the analytical error for the Mo isotope ratios was taken to be kO.6% of each ratio.In our previous work the standard Mo isotopic data from repeated analyses for seven runs varied by about k0.4 parts in 104. The small differences in the isotopic abundances of Mo among various samples are thought to be evidence of some nuclear effects of the early solar system; these will be discussed elsewhere. Conclusion The results obtained in this work indicate that the proposed chemical procedure could afford a rapid and highly efficient method not ony to extract microgram amounts of Mo from gram amounts of iron meteorites but also to separate Mo from Zr and Ru for the analysis of Mo isotopes with high accuracy. Moreover, by carrying out an additional back-extraction for Zr after back-extracting Mo, the former can be removed from the organic phase. Hence isotopic investigations of not only Mo but also Zr and Ru in iron meteorites can be carried out easily in comparison with the traditional ion-exchange method, as the isotopes of Zr and Ru have very similar nuclear backgrounds to Mo in terms of cosmochemistry. Further, the procedure would be applicable in nuclear chemistry and nuclear physics, because some isotopes of Zr, Mo and Ru are important products of the neutron-induced and spontaneous fission of U. 1 2 3 4 5 6 7 8 9 10 References Murthy, V. R., Geochim. Cosmochim. Acta, 1963, 27, 1171. Burbige, E. M., Burbige, G. R., Fowler, W. A., and Hoyle. F., Rev., Mod. Phys., 1957, 29, 547. Crouch, E. A., and Tuplin, T. A.. Nature (London), 1964,202, 1282. Howard, W. M., Meyer, B. S . , and Woosley, S. E., Astrophys. J . , 1991. 373, L5. Lu, Q., and Masuda, A., J . Am. Soc. Mass Spectrom., 1992.3, 10. Buchald, V. F., Handbook of Iron Meteorites, University of California Press, CA. 1975. Kraus. K. A., Nelson. F., and Moore, G., E., J. Am. Chem. Soc., 1955, 77, 3972. Wetherill, G. W., J. Geophys. Res., 1964, 69. 4403. Lu, Q., and Masuda, A.. Meteoritics, 1991. 26. 367. Qureshi, I. H., McClendon, L. T., and Lafleur. P. D., Radiochim. Acta, 1969, 12, 107. Paper 1 I05022 B Received October 1, 1991 Accepted November 14, 1991

ISSN:0003-2654    

DOI:10.1039/AN9921700869

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, MAY 1992. VOL. 117 873 Determination of Trace Amounts of Copper With Extraction-Photoacoustic Spectrometry Y. Deng and M. Ye The Centre of Analysis and Testing, Wuhan University, Wuhan, Hubei 430072, People’s Republic of China A sensitive method for the determination of copper in aqueous solution by pulsed laser-induced photoacoustic spectrometry after extraction is described. The copper is chelated with 1,5-diphenylcarbazide in chloroform and extracted into the chloroform phase. The optimum extraction conditions were found experimentally. By using the stepwise dilution method, the minimum absorbance that could be measured was 8.5 x 10-6 at 6.2 mJ per pulse; this corresponds t o 6.9 x 10-3 pg I-’ of copper in chloroform. The relative standard deviation for ten measurements of an absorbance of 7.8 x 10-3 at 0.7 mJ is 1.5%.Application of the method t o the determination of copper in biological samples is also described. The results obtained were compared with those given by atomic absorption spectrometry. Keywords: Extraction; photoacoustic spectrometry; copper determination; 1,5-diphenylcarbazide The ultra-high sensitivity of photoacoustic (PA) measure- ments for liquid samples has been demonstrated.’-3 In recent years, the PA determination of trace amounts of several solutes, such as cobalt, holmium, neodymium, benzene and cyclohexane, has been reported, and detection limits of 1 x 10-7-1 x 10-8 cm-1 were reached.- The determination of 1 x 10-4 pg 1 - 1 of cobalt by extraction-PA spectrometry has also been described.’ Copper is a common element that occurs widely in organisms as a necessary trace element possessing specific physiological functions.The determination of trace amounts of copper in biological material is receiving increasing attention in nutrition, and in medicinal and physiological studies. 1,5-Diphenylcarbazide (DPC) is a reagent mainly used for the selective colorimetric determination of chromium(v1). Recently, the use of this reagent for the colorimetric determi- nation of copper in basic medium by extracting the metal into chloroform as the Cu-DPC chelate was reported.8 However, no detailed studies of the extraction conditions were carried out. This paper describes the determination of sub-nanomolar amounts of copper in aqueous solution by PA measurement after extraction of the metal with DPC into chloroform.It is also demonstrated that the proposed method can be used for the determination of copper in biological samples such as grass carp and pig kidney. Experimental Apparatus A schematic diagram of the PA measurement system used is shown in Fig. 1. The excitation beam was the frequency- doubled (532 nm) output of an Nd:YAG laser (YG 581, Quantel) operating at 10 Hz with a duration of 8 ns. The power incident into the PA cell was monitored by a laser energy/ power meter (LPE-lA, Chinese Academy of Sciences). The beam radius, determined by an aperture, was 2 mm, unless stated otherwise. The PA cell for liquid PA measurements was laboratory-built and is shown schematically in Fig. 2. It was constructed by using a fused-quartz cuvette and two PZT-SH discs, which were fixed separately onto the outside walls of the cell.In order to enhance the PA signal response, the PZT discs were connected in series. In some experiments, a post-cell mirror was used for back-reflecting the pump light so that it passed through the sample again along the incident direction [Fig. 2(6)]. The cell with the post-cell mirror will be referred to as cell A, and the cell without the post-cell mirror as cell B. The PA signal was amplified ten-fold with a preamplifier (M115, PAR), from which the output was fed into a boxcar averager (M162/165, PAR) with a 20 ps aperture and 28% delay. A dual-pen x-y recorder was used to record the PA and power signals synchronously. A spectrophotometer (UV-240, Shimadzu) with a 1 cm cuvette was used for the absorbance measurements.Reagents A stock solution of copper(i1) (1.0 mg ml-1 in 1 mol dm-3 HCI) was prepared from high-purity copper wire. Working 532 nm - Iris I -I- I J Recorder Boxcar Pre-amplifier Fig. 1 Schematic diagram of the PA measurement set-up Aluminium foil Cell POdy Laser beam I c Laser beam874 ANALYST, MAY 1992, VOL. 117 solutions of copper at the ng 1-1 level were prepared daily by appropriate dilution of the stock solution with water o r the desired buffer solution. Analytical-reagent grade DPC was purified by three re- crystallizations from absolute ethanol, after which white crystals of DPC were obtained. A stock solution of DPC in chloroform (0.1Y0) was prepared by dissolving 100 mg of DPC in 3 ml of hot absolute ethanol and diluting to 100 ml with chloroform.The solution was stored in a brown flask. The DPC extractant used was obtained by further diluting the stock solution with chloroform. All other chemicals used were of analytical-reagent grade. Doubly distilled water was used throughout. Prior to solvent extraction, the buffer solution was extracted three times with chloroform containing DPC and then washed three times with chloroform to remove all the extractable impurities. Procedure Under optimum extraction conditions, an aliquot of an aqueous solution containing 10-80 ng of copper was trans- ferred into a 60 ml separating funnel and buffer solution of the desired pH was added so that the final volume of the aqueous phase was 20 ml. The solution was then shaken vigorously with 5 ml of chloroform containing 0.02% DPC for 3 min, after which the mixture was allowed to stand for 20 min so that the two phases could separate.After extraction, an aliquot of the extract (about 1.4 ml) was placed directly in the PA cell from the funnel for PA measurement. In preparing the absorption curve and for carrying out the experiments into the effect of pH, a solution containing microgram levels of copper was extracted and the absorbance measurements were performed with a spectrophotometer. Treatment of the Samples A known amount of the fresh sample (pig kidney, grass carp) was dried at 80 "C in an air oven for 8 h and then ground into a powder. A 1 g amount of the dried powder was digested with a mixture of nitric acid and hydrogen peroxide in a flask.After the organic matter had been destroyed, the residue was dissolved in 50 ml of 0.1 mol dm-3 HCI, transferred into a 100 ml calibrated flask and diluted to the mark with water. The sample solutions thus prepared were used for the determina- tion of copper. Results and Discussion Extraction of Copper(r1) With DPC The reaction of copper(i1) with DPC in basic aqueous solution forms a brown chelate which can be extracted into chloroform; in addition, copper(i1) in basic aqueous solution is also directly extractable as the Cu-DPC chelate by using chloroform containing DPC. The extraction was performed in various buffer systems, such as phosphate, tris(hydroxymethy1)- methylamine-hydrochloric acid and ammonium chloride- ammonia. It was found that the ammonium chloride buffer was superior to the other buffers examined for the extraction of copper with chloroform containing DPC.The dependence of the absorbance of the extract on pH is shown in Fig. 3; a pH of 8.6 was chosen. The effect of varying the DPC concentration over the range 0.002-0.1% on the degree of extraction, for 80 ng of copper in 20 ml of ammonium chloride buffer solution with 5 ml of DPC solution in chloroform, was examined. The degree of extrac- tion was found to be constant; hence, a DPC concentration of 0.02% was used as the optimum. Similar studies revealed that a shaking time of 1 min was sufficient to achieve quantitative extraction. Hence, for extracting less than 80 ng of copper in an aqueous phase volume of 20 ml, the optimum extraction 0.8 r 1 a, C 0.71 m 0.4 1 I I I I I I I 7.8 8.0 8.2 8.4 8.6 8.8 9.0 9.2 PH Fig.3 Effect of pH on the absorbance of the extract. The extraction was carried out in 10 ml of 0.1 mol dm-3 NH3-NH4Cl buffer containing 4 pg of Cu" with 10 ml of 0.1% DPC in chloroform 1.2 1 I g 1.0 C $ 0.8 $ 0.6 0 0.4 0.2 0 400 450 500 550 600 650 700 Wavelengthlnm Fig. 4 Absorption spectra of the copper chelate and the DPC reagent blank. A, Cu-DPC chelate against reagent blank; and B, DPC in chloroform against solvent conditions are as follows: sufficient 0.01 mol dm-3 ammonium chloride buffer of pH 8.6 to bring the final volume of the aqueous phase to 20 ml; 5 ml of 0.02% DPC in chloroform; shaking time, 3 min. Interference experiments were carried out under the conditions described above.Except for cobalt and nickel, none of the elements examined, i.e., Fe"', Cr"', Mn" 7 7 Zn" Cd", Hg", Mg" and Cr"', produced an extractable coloured chelate. Chelate of Copper With DPC The absorption spectra of the Cu-DPC chelate and of the reagent blank in chloroform are shown in Fig. 4. The wavelength of the absorption peak of the chelate at 545 nm closely matches that of the pump beam (532 nm) used here. The molar absorptivity of the Cu-DPC chelate in chloroform was found to be 7.8 x 1041 mol-1 cm-1 at 532 nm, whereas the absorbance of the DPC extractant at the same wavelength was negligible. Hence, this provides a low blank value in the absorption measurements, which is important for highly sensitive pulsed PA measurements. However, DPC is easily oxidized by air to produce a pink compound, which causes the blank to increase.Therefore, it was noted that careful purification of the extractant to obtain a sufficiently constant and low background is a prerequisite for highly sensitive analysis with the pulsed PA method. The colour system, after extraction, was found to be stable for at least 2 h. However, a gradual decrease in the PA signal was observed during irradiation with the pulsed laser. Curve A in Fig. 5 represents the situation in which the concentration of the chelate is slightly high; the PA signal decreases exponen-ANALYST, MAY 1992, VOL. 117 100 r I 875 I I I I I 1 0 4 8 12 16 20 4.25 1 1 Tim e/m i n 4.00 U s 3 3.75 3.50 3.25 0 2 4 6 8 1 0 Time/min Fig. 5 ( a ) Stability curve for the extracted Cu-DPC chelate during irradiation with a pulsed laser.Pulsed energy: 6.2 mJ. A. Absorbance = 4.4 x 10-2; and B , absorbance = 1.9 X 10-4. ( b ) Variation of In(&*) with irradiation time for curve A Time - Fig. 6 chelate. For details, see text PA signal generated from pulsed excitation of the Cu-DPC tially with irradiation time at 6.2 mJ per pulse. However, no decay of the PA signal is apparent at a low concentration of the chelate during the same irradiation time (curve B). In practice, an irradiation time of 1-2 min is sufficient for a single PA measurement. The deviation in the detection caused by the decay of the PA signal due to laser light irradiation is negligible, particularly for a weak absorbance or a low laser energy irradiation. PA Measurement The pulsed PA waveform of the extract is shown in Fig.6. The magnitude of the peak marked with an asterisk is directly proportional to the absorbance of the extract at a constant incident radiation power. It was observed that the waveforms were the same with both cell A and cell B. However, there is an almost 50% enhancement of the magnitude of the PA signal for cell A compared with cell B. It is known that the magnitude of the PA signal is linearly dependent on the pulsed energy. In order to increase the sensitivity for trace analysis, the pulsed laser energy applied 0 20 40 60 80 100 120 Po we r/m W Dependence of the magnitude of the PA signal on the incident Fig. 7 laser power Table 1 Application of the proposed method to the determination of copper in biological samples and comparison of thc results obtaincd with those given by AAS Cu foundlpg g- 1 Proposed Sample" method? AAS Pig kidney 25.0 26.5 Grass carp 6.93 7.50 * Dried powder.t Average of three determinations. should be as high as possible. However, a high pulsed energy will always induce many non-linear effects such as photodissociation and breakdown. The optimum pulsed energy value is that which not only maintains high sensitivity for PA detection, but also allows Beer's law to be obeyed over the concentration range to be examined. The graph of PA signal intensity versus incident laser power is shown in Fig. 7 . By using cell A, the optimum energy value found experimentally was about 6.5 mJ per pulse in a 3 mm beam radius for an absorbance of 0.04.The minimum absorbance that could be measured, which was obtained by using a series of solutions prepared by stepwise dilution of the Cu-DPC chelate with chloroform, was 8.5 x 10-6 at 6.2 mJ per pulse. This corresponds to 6.9 x 10-3 pg 1-1 of copper in chloroform solution. The relative standard deviation for ten replicate measurements of a chloroform solution of the chelate with an absorbance of 7.8 x 10-3 was 1.5% at 0.7 mJ per pulse. Because of the high content of copper in the biological samples examined here, a calibration graph was constructed in the range 0.04-0.4 pg 1-1 of copper in aqueous solution. The detection limit, defined as a signal-to-noise ratio of 3 at 1.9 mJ per pulse, was 0.022 pg 1- in aqueous solution. Applications of the Method The proposed method was applied to the determination of trace amounts of copper in pig kidney and fish flesh (grass carp).The results obtained with the proposed method were compared with those given by atomic absorption spectrometry (AAS) and are shown in Table 1. The relative standard deviation for six replicate extraction- PA measurements of 0.4 ml of a prepared solution was 7.8% for the grass carp sample. The recovery of copper from a sample (fish) was found to be 98.5% by extraction and PA measurement.876 ANALYST, MAY 1992, VOL. 117 Conclusion A sensitive method for the determination of copper, based on extraction and PA spectrometry, has been developed. The combination of PA measurement and separation by extraction considerably improves the selectivity of the PA measurement; further, the use of an organic solvent enhances the magnitude of the PA signal. The proposed method allows a Cu-DPC extract with an absorbance of 8.5 X 10-6 to be detected, and can also be applied to the determination of copper in biological samples containing little or no cobalt and nickel. This project was supported by the National Natural Science Foundation of China. References 1 Lahmann, W., Ludewig, H . , and Welling, H., Anal. Chem.. 1977. 49, 549. 2 Oda. S., Sawada, T., and Kamada, H., Anal. Chem., 1978.50, 865. 3 Kitamori, T.. Fujii, M., Sawada. T., and Gohshi, Y., J. Appl. Phys., 1985, 58, 268. 4 Yan, H.. Deng. Y . , and Zeng, Y., Chem. J. Chin. Univ., 1988, 4 , 25. 5 Yan, H.. Deng, Y., and Zeng, Y . , Kexue Tongbao (Engl. Trunsl.), 1989, 34. 790. 6 Zuo, B., Deng, Y . , and Zeng, Y., Chem. J. Chin. Univ., 1990, 11, 15. 7 Kitamori, T., Suzuki, K., Sawada, T., Gohshi. Y.. and Motojima, K., Anal. Chem., 1986.58, 2275. 8 Huang, X., Fenxi Huaxue, 1990, 18, 304. Paper I I0451 4H Received August 29, I991 Accepted December 4, I991

ISSN:0003-2654    

DOI:10.1039/AN9921700873

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, MAY 1992, VOL. 117 877 Simultaneous Determination of Acetylsalicylic Acid and Its Major Metabolites in Human Serum by Second-derivative Synchronous Fluorescence Spectrometry Dimitrios G. Konstantianos and Pinelopi C. loannou" Laboratory of Analytical Chemistry, Department of Chemistry, University of Athens, University Campus, Kouponia, Athens 75771, Greece A method is described for the simultaneous determination of acetylsalicylic, salicylic, gentisic and salicyluric acids (ASA, SA, GA and SU, respectively) in serum, based on their native fluorescence. The ASA-SA-GA-SU- containing serum samples are extracted with chloroform-1% acetic acid solution; ASA and SA are determined in the organic phase, and GA and SU in the aqueous phase, after removal of protein with trichloroacetic acid, at pH 5.0 and 11.6, respectively.The ASA-SA and GA-SU-SA mixtures are resolved using second-derivative fluorescence spectrometry and the appropriate empirical equations involving the effect of each acid on the signal of the other. Recoveries from sera spiked with ASA (1.0-10 vg ml-I), SA (25-50 1-19 ml-I), GA (0.05-0.2 pg ml-1) and SU (1.0-5.0 pg ml-1) ranged from 100 t o 104% (mean 101%), from 93 t o 99% (mean 97%), from 94 t o 104% (mean 99%) and from 94 t o 107% (mean 98%), respectively. Keywords: Second-derivative synchronous fluorescence spectrometry; acetylsalicylic acid; metabolites; human serum Monitoring of salicylates has been a field of active investiga- tion for more than 45 years.' This is due to the widespread and frequent use of salicylic acid (SA) derivatives, especially the acetylated form, aspirin (ASA), for a variety of medical conditions.Determination of salicylates in biological fluids is of interest both in emergency and routine testing24 and in pharmaco- kinetic investigations, as ASA and its major metabolites in the body, such as SA, gentisic acid (GA) and salicyluric acid (SU), have all been reported to show different pharmacological effects.5 Of the techniques used for determining salicylates, high-performance liquid chromatography (HPLC) is preferred for pharmacokinetic investigations because of the required level of specificity, sensitivity and simplicity. However, most of the HPLC methods proposed either fail to quantify ASA or fail to prevent its hydrolysis during sample preparation.6-8 Only a few methods provide a complete analysis for ASA and its major metabolites.9-11 Non-chromat- ographic methods for determining salicylates in biological fluids2.3 generally lack specificity, although this can be improved by using different techniques of sample preparation andor different spectral characteristics. Synchronous scanning derivative fluorescence spec- trometry, which combines high sensitivity with improved selectivity, compared with conventional fluorescence spec- trometry," is a very useful tool for resolving multicomponent mixtures without prior separation.Muiioz de la Peiia et a1.I3 determined SA and SU in urine by first-derivative fluores- cence spectrometry following extraction into diethyl ether and back-extraction of SA and SU into an appropriate buffer solution.Salinas et al. 14 improved the method to include GA; ASA cannot be determined by this method, and the ratios of metabolites examined do not cover their ratios in human urine found during pharmacokinetic studies by other methods.9.15~16 Recently, we reported a method for the simultaneous determi- nation of ASA and SA in serum and pharmaceuticals by second-derivative synchronous fluorescence spectrometry (SDSFS).l7 This study represents an improvement of the method to include SU and GA, hence allowing the simul- taneous determination of ASA and its major metabolites (SA, GA and SU) in a single serum sample. * To whom correspondence should be addressed. By using the method described here, ASA and SA were determined in the organic phase after simple extraction of the sample with chloroform-1% acetic acid; GA and SU were determined in the aqueous phase, after removal of proteins by treatment with trichloroacetic acid, so as to avoid interference from serum components and to liberate GA and SU, which are strongly bound to serum albumin.Mixtures of ASA-SA, GA-SU, SU-SA, GA-SA and GA-SU-SA were resolved by SDSFS. The proposed method was applied successfully to the determination of ASA, SA, GA and SU in serum samples. The analysis time for all the compounds was less than 15 min. Studies on samples containing various concomitantly adminis- tered drugs were performed in order to demonstrate the specificity of the proposed method. Experimental Apparatus A Model 512 fluorescence spectrometer (Perkin-Elmer, Norwalk, CT, USA), equipped with a 150 W xenon arc lamp and a magnetic stirrer under the cell holder, was used.All measurements took place in a standard 10 mm (pathlength) quartz cell, thermostatically controlled at 25.0 k 0.5 "C. Excitation and emission monochromators were locked together and scanned simultaneously with a constant differ- ence Ah = he, - Aex. The scan speed and response time of the spectrometer were set at 4 nm s-1 and 'Fast' mode, respec- tively. The digital read-out unit of the spectrometer was interfaced with an Amstrad CPC-6128 microcomputer (Brentwood, UK) for spectral acquisition, calculations of the spectrum deriva- tives and automatic evaluation and presentation of the analytical signals, with use of laboratory-made software.17 Smoothed and derivative spectra were defined by using the Savitzky-Golay method. 18 Reagents Spectroscopic-quality grade chloroform (Merck, Darmstadt, Germany) was used to make up 1% v/v acetic acid in chloroform. Henceforth, this mixture will be referred to as 'mixed solvent'. Stock solutions of ASA and SA (Fluka, Buchs, Switzerland) containing 4.00 and 2.00 mg ml-1, respectively, were prepared in the mixed solvent. The stock878 ANALYST, MAY 1992, VOL. 117 solution of ASA was prepared daily, whereas the stock solution of SA was stable for at least 1 year at room temperature. Aqueous stock solutions of ASA, SA, GA and SU (Sigma, St. Louis, MO, USA), containing 1.00, 0.50, 1.00 and 1.00 mg ml-1, respectively, were prepared in distilled, de- ionized water.The stock solution of ASA was prepared daily, whereas the stock solution of SA was stable for at least 1 month at room temperature. Stock solutions of GA and SU were stable for 1 week at 4 "C. A stock solution of bovine albumin (Sigma), containing 100 mg ml-1, was prepared in distilled, de-ionized water daily. Phosphate buffers (0.05 mol dm-3) of pH 7.2 and 11.6 were prepared; 10 mol dm-3 NaOH was used for pH adjust- ment. A 16% m/v solution of trichloroacetic acid in chloro- form was used for deproteinization. Procedures and Calculations Sample preparation Sample preparation and all measurement steps, are sum- marized in Scheme 1. Place 0.50 ml of serum containing 0.50-20.0 pg of ASA, 50-100 pg of SA, 0.025-0.4 pg of GA and 1.0-5.0 pg of SU into a test-tube.Add 2.00 ml of the mixed solvent, sonicate the mixture for 1 min and centrifuge for 3 min at 1500g. Use the organic phase for ASA and SA determination, and the upper aqueous phase for GA and SU determination. Transfer 0.40ml of the aqueous layer into another test- tube. Add 0.12 ml of trichloroacetic solution, sonicate for 2 min and centrifuge for 3 min at 1500g. Use the clear supernatant solution for GA and SU determination. Determination of ASA and SA Transfer 1.00 ml (0.50 ml for higher concentrations of SA) of the organic phase into the cell, add mixed solvent to a total volume of 2.00 ml and start the stirrer. Obtain the synchro- nous fluorescence spectra by scanning both monochromators simultaneously at a constant difference Ah = 60 nm (hex = 240-290 nm) and AL = 130 nm (Ibex = 290-320 nm) for ASA and SA, respectively. (Hereafter, all wavelengths referring to synchronous spectra are taken to be equal to those of the corresponding excitation wavelengths.) Evaluate the second- Sample preparation (test-tube) Measurement step (cell) Take 0.50 ml of serum, add mixed solvent (2.00 mi), Dilute (1 + 1 or 1 + 3) extract with mixed solvent ASA and SA Aqueous layer Add 10 mol dm-3 NaOH (0.03 ml) (final pH 11.6) Determination of SU Transfer 0.40 rnl, add trichloroacetic acid (0.12 mi), sonicate, centrifuge Take 0.30 ml, dilute (final pH 5.0) Determination of GA Scheme 1 Determination of ASA, SA, GA and SU in a single sample serum derivative signal of ASA, AZASA, within the spectral range 260-290 nm, and the signal of SA, AISA, within the spectral range 310-344 nm.Calculate the concentration of SA in the extract, cSA (pg ml-I), from the calibration graph for SA in the mixed solvent (cSA is required for the subsequent determination of ASA). Calculate the total concentration of SA in serum, c S ~ ( ~ ~ ~ ~ ~ ) , from a calibration graph obtained with control serum standards spiked with SA (concentration range 40-200 pg) and treated similarly. Calculate the total concentration of ASA in serum, c A S A ( ~ ~ ~ ~ ~ ) (pg ml-I), by using eqn. (1). This equation, and eqns. (2)-(6), were reported in previous work. 17 CASA(serurn) = where df is the actual dilution factor (8 or 16), and SAsA is the slope of the calibration graph for ASA in the mixed solvent. Determination of GA and SU For the determination of GA, transfer 0.30ml of the clear supernatant solution into the cell, add 1.70ml of buffer solution (pH 7.2) and start the stirrer. Obtain the synchronous fluorescence spectra by scanning both monochromators simul- taneously at a constant difference Ah = 120 nm (Aex = 250-390 nm).Evaluate the analytical signal of GA, AZGA, and that of the interferent, AIint,l, within the spectral ranges 320-360 and 264-304 nm, respectively. Correct AZGA with respect to the signal of the interferent AZint, 1, by using eqn. (2). Calculate the concentration of GA in serum from a calibration graph obtained with control serum standards spiked with GA (concentration range 0.1-2.4 pg ml-1). For the determination of SU, place 0.03 ml of 10 mol dm-3 NaOH in the same cell to bring the pH to 11.6.Obtain the synchronous fluorescence spectra by scanning both mono- chromators simultaneously at a constant difference of Ah = 70 nm (hex = 250-290 nm). Evaluate the analytical signal of SU, AZsu, and that of the interferent, AZint,2, within the spectral ranges 320-356 and 260-300 nm, respectively. Cor- rect AIsu with respect to the signal of the interferent, AZint,2, by using eqn. (2): A ~ G A or SU(corr) = d f AIGA or SU ( 0.348 log AIGA or su 10.760) (2) where df is the dilution factor (6.67). Calculate the concentra- tion of SU in serum from a calibration graph obtained with control serum standards spiked with SU (concentration range 0.4-8.0 vg ml-1). All instrumental parameters are summarized in Table 1. AIint 1 or2 Table 1 Instrumental parameters for the determination of ASA, SA, GA and SU Corn po und Parameter Slit-widthkx, em (nm) AUnm Savitzky-Golay filter size/points Synchronous spectrum scanning range/h,, (nm) AI evaluation range/h,, (nm) A1 evaluation range for interference/h,, (nm) ASA SA GA SU 20.20 20.20 10.20 10,20 60 130 120 70 9 9 11 11 240-320 290-370 250-390 250-390 260-290 310-344 320-360 320-356 (A~ASA) ( A ~ S A ) ( A ~ G A ) (Alsu) - - 264-304 272-312 (Afint, I ) (AJint.2)ANALYST, MAY 1992, VOL.117 879 Results and Discussion The simultaneous determination of ASA and SA by SDSFS has been reported previously.17 The measurement was per- formed in chloroform-1% acetic acid solution ('mixed sol- vent') in order to avoid hydrolysis of ASA, and to make use of the different fluorescence properties of ASA and SA in this solvent.160 1 .o 0.8 0.4 0.2 / I 240 280 320 360 L i / " r n 2 4 6 8 PH Fig. 1 Effect of pH (a) on the excitation spectrum (he,,, = 400 nm) of SU (pH: A , 5.0; B, 7.3; C, 8.2; D, 8.9; and E, blank) and ( b ) on the second-derivative signals of the excitation spectra: he, = 300 nm (1) and 332 nm (2) v) a 4- .- a 90 - C a v) ? 60 - 3 a - Y- .- 2 3 0 - - a U I (dl U U 5 -0.3 al v) The optimum wavelength differences, AA, for the synchro- nous scanning were found to be 60 nm for ASA and 130 nm for SA. The determination was performed by evaluating the second-derivative signals within the spectral ranges 260- 290 nm for ASA and 310-344 for SA (see Table 1). Empirical eqn. (1) for calculating the concentration of ASA at a given concentration of SA was proposed in order to overcome the influence of SA on the determination of ASA.For the determination of these compounds in serum a single extraction into mixed solvent was proposed. In this paper, the method was extended to the determina- tion of GA and SU in ASA-SA-GA-SU mixtures. As the extractibility of GA and SU into chloroform has been reported to be poor,'9 the possibility of separating an ASA-SA-GA- SU mixture into ASA-SA and GA-SU mixtures was studied. It was found that the use of mixed solvent instead of chloroform causes a slight increase in the extractibility of GA. However, the extraction of GA and SU from albumin solutions or serum into mixed solvent was negligible, owing to their strong binding with albumin. Hence, by a simple extraction with the mixed solvent, ASA and SA are trans- ferred into the organic phase, while GA and SU remain in the aqueous phase.It should be noted that whereas extraction of ASA from albumin solutions o r serum is almost quantitative, SA is extracted only as an unbound fraction16 (of about 30% of the total amount of SA in serum). The bound fraction of SA is liberated from serum albumin after deproteinization. There- fore, GA and SU have to be separated in the serum supernatant phase as GA-SU-SA mixtures. For separating binary or ternary mixtures of GA, SU and SA, the SDSFS technique was used in combination with the dependence of the fluorescence properties of the three compounds on the pH. As has been reported previously by Salinas el al. ,I4 at pH >11.0, the only fluorescent species are SU (Aex = 332 nm, A,, = 400 nm) and SA (hex = 300 nm, he, = 400 nm), whereas at pH (6.0 only GA (hex = 324 nm, A, = 444 nm) and SA (Aex = 300nm, he, = 400 nm) exhibit fluorescence.A pH of 6.0 and of 11.6 for the determination of GA and SU, respectively, was therefore proposed. However, a detailed study on the influence of SU on the analytical signal of GA showed significant interference from SU, especially at 260 320 380 260 320 380 hln m 260 320 380 Fig. 2 Synchronous tluorescence spcctra [(a)-(c)] of A, GA; B, SA; C , SU; D, GA-SA (a); GA-SU (b); and SU-SA (c) mixtures. and E, a blank; and their second-derivativc spectra [(d)-(f)]. ( a ) : cGA = 0.030 pg ml-I, CSA = 0.15 pg ml-I (Ah = 120 nm, sensitivity 10. pH = 5.0).( b ) cGA = 0.025 pg ml-1, csu = 1.00 pg ml-1 (Ah = 120 nm, sensitivity 10, pH = 5.0). (c): csu = 0.060 pg ml-1, cSA = 0.48 pg ml-1 (AI. = 70 nm, sensitivity 10. pH = 11.6)880 ANALYST, MAY 1992, VOL. 117 high SU : GA ratios. The influence is caused by the existence of a second excitation maximum of SU (hex = 300 nm) at 4 < pH < 8 [Fig. l(a)], which results in negative errors when GA is determined by the SDSFS technique at the optimum difference (Ah = 120 nm). The influence of pH on the second-derivative excitation signal of SU, AIsu(exc), at two maxima is shown in Fig. l(b). An optimum pH of 5.0 was selected for the determination of GA, where the AISU(ex,-) signal at hex = 300 nm is almost constant (pH = 4-6), and the AISU(exc) signal at hex = 332 nm is the lowest.The optimum pH of 5.0 for the determination of GA was obtained by addition of 0.05 mol dm-3 phosphate buffer of pH 7.2 to the acidified supernatant phase (after deproteiniza- tion with trichloroacetic acid). For the determination of SU, a pH of 11.6 was selected, where the fluorescence intensity of GA is negligible. The signals of SA and SU at pH 5.0 and 11.6 are referred to as interference signals, Alint,l and Alint,2, respectively. Comparison of Spectra The synchronous spectra and their corresponding second derivatives obtained for GA, SU, SA and GA-SA, GA-SU and SU-SA mixtures, at the optimum pH for the determina- tion of GA and SU, are shown in Fig. 2. As can be seen from Fig. 2, the combination of synchronous and derivative fluorescence techniques results in adequate resolution of the mixtures.However, the vicinity of large SA and/or SU bands to the band of GA, and the vicinity of the SA band to the band of SU, resulted in a decrease of AIGA [Fig. 2(d) and ( e ) , curve C] and AIsu signals [Fig. 2(R curve C], compared with the pure solutions (curves A). This decrease, which is purely a mathematical artifact, inevitably results in a loss of AZGA and AIsu signals in the presence of a large excess of SA and/or SU, which must be taken into account. The absence of any quenching effect has been confirmed by the fact that the spectral bands of the mixtures are equivalent to the sum of their individual bands. I c 345 E 330 1 2 31 5 300 1 I I I I I 0.15 I 1 0.12 ’ 0.09 ’ -x a 0.06 . 0.03 . 0 40 80 120 160 200 AUn m Fig.3 Effects of Ah on the second-derivative synchronous spectra of A, SA; B, GA; and C, SU. (a) Effect on the wavelength correspond- ing to the minimum of the second-derivative peaks; and (b) effect on the analytical signals Selection of AI for Synchronous Scanning In order to select the optimum wavelength difference between excitation and emission monochromators (Ah) for the deter- mination of GA and SU with SA by SDSFS, a wide range of Ah values (20-240nm) was examined. The position of the minimum of the second-derivative spectrum and the signal for GA and SU as functions of Ah are shown in Fig. 3(a) and (b), respectively. As optimum wavelength intervals for GA and SU, Ah values of 120 and of 70 nm, respectively, were selected, so as to minimize the spectral interference caused by each com- pound in the mixture and to minimize the loss of sensitivity.The same experiment for ASA and SA in the mixed solvent has been reported previously. 16 Other Instrumental Parameters For recording the synchronous spectra, a scan speed of 4 nm s-1 and a ‘Fast’ response time were selected. The sampling rate was defined to be 1 point every 4 nm. For the calculation of the second derivatives of the synchronous fluorescence spectra of GA and SU by the Savitzky-Golay method,18 filter sizes of 11 and 11 points, respectively, were selected. General Analytical Characteristics The linear concentration ranges were 0.003-6.50 and 0.006- 10.5 vg ml-1 for GA and SU, respectively. Pearson’s correla- tion coefficients20 ( r ) for the calibration graphs were 0.9998 and 0.9992 for GA and SU, respectively.The detection limits obtained by SDSFS, defined as three times the standard deviation of the lowest concentration, were 0.001 and 0.002 pg ml-1 for GA and SU, respectively. In order to test the precision of the method, three series of samples, covering the ranges of interest for GA (0.050, 0.50 and 5.00 pg ml-1) and for SU (0.090,0.80 and 7.00 pg ml-1) were analysed, and the corresponding relative standard deviations (RSDs) (n = 10) were found to be 3.2, 1.5 and 1.5% for GA, and 3.8,2.3 and 2.0% for SU. Determination of GA and SU in Binary or Ternary Mixtures with SA In order to apply the SDSFS technique to the simultaneous determination of GA and SU in binary or ternary mixtures with SA, a detailed study on the influence of the signal for each acid on the analytical signal of the other was performed.The effect of the analytical signal for SA and SU on the analytical signal for GA (curves A and C), and the effect of the 0.2 0.4 0.6 0.8 1.0 0 &t Fig. 4 the analytical signal [Al,,,,,,] of GA (A and B) and SU (C) Effect of the signals of SA and SU as interferences (Alint) onANALYST, MAY 1992, VOL. 117 88 1 Table 2 Determination of 0.050 pg ml-1 of GA in the presence of an excess of SA in aqueous solutions GA found*/pg ml- Recovery f SD (%) SA : GA mass ratio Uncorrected? Corrected$ Uncorrected Corrected 1 : l 0.040 0.050 80 k 14 100 f 2 2 : 1 0.035 0.050 70+ 11 1 0 0 f 3 4: 1 0.031 0.050 62 k 5 100 f 3 6 : 1 0.028 0.051 5 6 f 15 102 f 3 8 : 1 0.022 0.050 4 4 f 2 100 f 5 10: 1 0.016 0.048 32 k 8 96 k 1 12: 1 0.018 0.050 36 + 2 100 f 3 14: 1 0.013 0.045 26 f 6 90 f 5 16: 1 0.010 0.049 20 f 2 98 k 6 Mean 9 8 f 4 * Average of three measurements.t Without using correction equation. $ After using correction eqns. (3) and (6). Table 3 Determination of 0.050 pg ml-L of GA in the presence of exccss of SU in aqueous solutions GA found*/pg ml- Recovery f SD (%) SU : GA mass ratio Uncorrected? Corrected$ Uncorrected Corrected 10: 1 20: 1 30: 1 40 : 1 50: 1 60: 1 70: 1 80: 1 0.045 0.042 0.036 0.035 0.034 0.036 0.035 0.029 0.051 0.050 0.048 0.049 0.051 0.051 0.052 0.046 * Average of four determinations. t Without using correction equation. $ After using correction eqns. (4) and (6). 90 f 8 102 f 1 84 f 9 100+6 72 f 9 96 k 3 70 f 4 98 f 3 68 f 4 102 f 2 7 2 f 12 102+6 70 f 9 104 f 5 58 f 6 91 f 5 Mean 99+4 Table 4 Determination of 0.050 pg ml-1 of SU in the presence of excess of SA in aqueous solutions SU found*/pg ml- Recovery k SD (%) SA : SU mass ratio Uncorrectedt Corrected$ Uncorrected Corrected 2 : 1 4: 1 6 : 1 8 : 1 10: 1 12 : 1 16: 1 20: 1 0.039 0.033 0.028 0.024 0.020 0.019 0.010 0.012 0.050 0.050 0.051 0.050 0.051 0.049 0.049 0.051 78 f 5 66 k 6 56 f 6 48 f 4 40 k 1 38 f 4 20 f 2 2 4 f 8 1OOf3 100+4 102 f 4 100 k 2 102 f 1 98 k 1 98 f 5 102 -t 5 Mean 1OO-t2 * Average of four measurements.t Without using correction equation. $ After using correction eqns. ( 5 ) and (6). Table 5 Determination of GA and SU in synthetic GA-SU-SA mixtures Conccntratiodpg ml- Added Found* Recovery (%) massratio GA SU GA SU GA SU 7:5: 1 0.050 0.250 0.048 0.240 99.0 96.0 5:3: 1 0.080 0.240 0.087 0.230 108.8 95.8 3:5: 1 0.080 0.400 0.078 0.416 97.5 104.0 2: 10: 1 0.050 0.500 0.049 0.485 98.0 97.0 Mean 100.8 98.2 SA: SU: GA * Average of three measurements, using eqns.(3) and (6). analytical signal for SA on the analytical signal for SU (curve B), obtained by SDSFS at AL = 120nm and AIL = 70nm, respectively, is shown in Fig. 4. The analytical signals for GA and SU in the presence of the SA and/or SU (for GA) signal, obtained by SDSFS, decreased considerably with an increase in the SA or SU signal. It was found that this decrease, expressed as a correction factor for the analyte in the presence of the interferent, CFanal(int) (i.e., the ratio of the signal for GA or SU in pure solutions, AI,,,,, to its signal in the presence of the interferent, AIobs), is linearly related to the ratio of the observed signal for the analyte AIGA or AIsu to the signal for the interferent AIsA and/or AIsu, according to the equations: + 0.760( k0.023) AIGA CFc;A(SA) = 0.348( k0.005) log - AISA (Y = 0.999) + 0.905( k0.028) AIGA A k u CFGA(SU) = 0.253( k0.024) log - ( r = 0.993) AISU CFsu(sA) = 0.356( k0.012) log ~ + 0.776( k0.028) AISA (Y = 0.998) The corrected (true) signal in each instance is given by the equation: Each of the eqns.(3), (4) and (5) is the mean of three equations obtained at three different concentrations of GA (0.050, 0.40 and 3.00 pg ml-1) and at three different concen- trations of SU (0.080, 0.50 and 5.00 yg ml-I), at increasing signals for the interferents, SA and/or SU.Results for the determination of GA and SU in synthetic GA-SA, GA-SU and SU-SA mixtures, calculated from the calibration graphs with and without use of the appropriate correcting equations are summarized in Tables 2-4. As can be seen from these tables, by using the appropriate correcting equations, GA could be determined in the presence of up to a l6-fold excess of SA, or up to an 80-fold excess of SU, and SU could be determined in the presence of up to a 20-fold excess of SA. At higher excess of the interferent the signal for the analyte Alobs is completely hidden. As can be seen from eqns. (3) and ( 5 ) , the influence of the signal for SA on the signals for GA and SU is similar even though they were obtained at different AL (120 and 70 nm, respectively), and at different pH (5.0 and 11.6, respectively).Table 5 summarizes the results obtained for synthetic GA- SU-SA mixtures at ratios expected in serum samples.”1s The corrected signals for GA in ternary mixtures were calculated without using eqn. (4) because the influence of the signal for SU, with respect to the influence of the signal for SA, is negligible. The concentrations of GA and SU in SA-GA-SU mixtures can be calculated from eqn. (2), which is a combination of eqns. (3) and (6), and from the appropriate calibration graph. Serum Samples Serum samples or albumin solutions containing GA and SU gave signals smaller than those obtained with aqueous standard solutions, owing to some type of binding with precipitated proteins.As has been found, the maximum recovery of GA and SU was obtained when a chloroform solution of trichloroacetic acid was used to precipitate proteins, and the mixture was sonicated for 2 min. Recovery experiments on albumin solutions containing GA or SU at several albumin and component concentrations gave882 ANALYST, MAY 1992, VOL. 117 Table 6 Determination of quaternary mixtures of ASA, SA, GA and SU in synthetic serum mixtures Serum concentration/yg ml-1 Concentration found*/yg ml-1 Recovery k SD (%) ASA SA GA SU ASA SA GA SU ASA SA GA su 10.0 35.0 0.05 1 .o 10.4 34.0 0.049 0.94 1 0 4 f 2 97k 1 98+2 94+4 5.0 50.0 0.10 2.0 5.0 47.8 0.097 1.89 100+2 96+2 97+2 94+4 10.0 40.0 0.20 3.0 10.1 39.5 0.207 3.22 101 f 3 99+ 1 104f 1 107k2 1.0 30.0 0.20 3.0 1.0 28.0 0.202 2.89 100k 1 93+ 1 101+ 1 96+2 - 25.0 0.10 5.0 - 24.7 0.094 5.02 - 9 9 f 2 9 4 f 2 100+1 Mean 101.2 96.8 98.8 98.2 * Average of three measurements.values of 70.5 k 0.9% ( n = 5 ) and 73.4 k 3.1% (n = 6) for GA and SU, respectively. It was also found that the recoveries of GA and SU are not dependent on albumin concentrations in the range 2-5%. Recovery data for ASA-SA-GA-SU synthetic mixtures added to serum are shown in Table 6. The selected concentra- tions for quaternary mixtures are typical for ASA-SA-GA- SU levels in serum, during the first 6 h, from a typical subject, following an oral dose of 650 mg of aspirin.5.15 Interference Studies The determination of SA, ASA and its major metabolites GA and SU is not affected by endogenous substances usually found in the sera of healthy subjects.Interference from other drugs was studied by analysing synthetic mixtures of ASA, SA, GA and SU in serum where amounts of the drug under examination had been added. None of the drugs tested (antipyrine, ibuprofen, amilorid, imipramine, amitriptyline, indomethacin, amoxycillin, levodopa, caffeine, naproxen, carbamazepine , phenace tin, chlorpromazine , theophylline and dithranol) interferes with the determination of ASA, SA, GA and SU at concentrations higher than those achieved therapeutically. Conclusions The proposed method is the first non-chromatographic method for the simultaneous determination of ASA, SA, GA and SU in a single serum sample. The ASA-SA and GA-SU-SA mixtures in a wide range of ratios are resolved by using SDSFS and the appropriate equations, which involve the influence of each acid on the analytical signal of the other.These empirical equations are not instrument dependent and can be used without knowing the concentration of the analytes o r the interferents. The proposed scheme, which incorporates an extraction step for the separation and determination of ASA and SA, thus minimizing hydrolysis of ASA, and a single deproteiniza- tion step for the determination of GA and SU, might appear complex. However, the total analysis time (including separa- tion, deproteinization and measurement) does not exceed 15 min. The usual analysis time for all four components by HPLC methods varies between 25 and 35 min. The proposed method is fairly sensitive and specific for all four compounds and can be used for pharmacokinetic studies of aspirin in serum as an alternative to the HPLC technique.In addition, the method can be applied to the determination of SA, GA and SU in urine. The optimum conditions for this determination are under investigation. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 References Stewart, M. J., and Watson, I. D., Ann. Clin. Biochem.. 1987, 24, 1552. Trinder, P., Biochem. J . . 1954, 55, 301. Kang, E. S., Todd, T. A., Capaci, M. T., Schwenzer, K., and Jabbow, J. T., Clin. Chem. (Winston-Salem, N.C.), 1983, 29. 1012. Ram, N. G., and Mohebullah, Z., Clin. Chem. (Winston-Salem, N.C.), 1990, 36. 1690. Rumble, R. H., and Roberts, M. S.. J. Chromatogr., 1981,225, 252. Tenveij-Groen, T., Vahlkamp, T., and Krak, C., J. Chromat- ogr., 1978, 145, 115. Peng, J . W., Gadalla, A. F., Smith, W., Reng, A., and Chiou. W. L., J. Pharm. Sci., 1978,67, 710. Blair, D., Rumack, B. H., and Peterson, R. G., Clin. Chem. (Winston-Salem, N. C.), 1978. 24, 1543. Mays, D. C., Sharp, D. E., Beach, C. A., Ketshaw, R. A., and Bianchine, J. R., J . Chromatogr., 1984, 311, 301. Ogunbona, F. A., J. Chromatogr., 1986, 377, 471. O’Kruk, R. J., Adams, M. A.. and Philp, R. B., J. Chromat- ogr.. 1985, 310, 343. Rubio, S., Gomez-Hens, A., and Valcarcel, M., Talanta, 1986. 33, 633. MuAoz de la Peiia, A., Salinas, F., and Duran-Meras, I., Anal. Chem., 1988,60, 2493. Salinas, F., Muiioz de la Peiia. A.. Duran-Meras, I., and Soledad Duran, M., Analyst, 1990, 115, 1007. Reile, U.. J . Chromatogr.. 1983, 272, 325. Amick, E. N., and Mason, W. D., Anal. Lett., 1979, 12, 629. Konstantianos, D. G., Ioannou, P. C., and Efstathiou, C. E., Analyst. 1991, 116,373. Savitzky, A., and Golay, M. J. E.. Anal. Chem., 1964.36.1927. Bakar. S. K., and Niazi, S., J. Pharm. Sci., 1983, 72, 1020. Miller, J. C., and Miller, J. N.. in Statistics for Analytical Chemistry. eds.. Chalmers, R. A., and Masson, M., Ellis Horwood, Chichester. 2nd edn., 1988, pp. 85-90. Paper 11042 77G Received August 15, 1991 Accepted October 23, I991

ISSN:0003-2654    

DOI:10.1039/AN9921700877

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, MAY 1992, VOL. 117 883 Simultaneous Determination of Atmospheric Nitric Acid and Nitrous Acid by Reduction With Hydrazine and Ascorbic Acid With Chem i lu m i nescence Detect ion Yukio Kanda and Masafumi Taira NationaI Laboratory for High Energy Physics, Oho, Tsukuba, Ibaraki-ken 305, Japan A continuous-flow system for the simultaneous determination of atmospheric HN03 and HN02 has been developed that consists of two sets of dual-channel flow systems: one is for the measurement of total HN03-HN02 and the other is for the measurement of HN02. Nitric acid and HN02 are continuously stripped from the atmosphere into an NaOH solution by drawing the air sample and the solution through a glass coil. The nitrate is reduced with hydrazine sulfate in the presence of Cull t o nitrite, which is then reduced t o NO with ascorbic acid solution, and the NO is detected by a chemiluminescence NO, analyser.This measurement system can also be used t o determine HN02, thus giving the total concentration of HN03 and HN02. The measurement system without the hydrazine reduction procedure determines HN02 alone. The concentration of HN03 is determined by the difference between the two measurements. Of the common pollutants, NO2 and peroxyacetyl nitrate showed positive interferences. In order t o correct for these positive interferences, each of the measurement systems utilizes a dual flow system and a dual-channel NO, analyser. Keywords: Flow system; gaseous nitric acid and nitrous acid; h ydrazine and ascorbic acid reduction; chemiluminescence NO, analyser Atmospheric gaseous HN03 has received considerable atten- tion because of its role as a major sink for nitrogen oxides in photochemical air pollution* and its contribution to the acidity of precipitation .2-3 Most analytical methods for the determination of gaseous HN03 are based on collection of HN03, extraction of nitrate by a suitable solvent and subsequent analysis by colorimetry or ion chromatography.A variety of collection techniques have been developed including the use of a nylon an NaC1-impregnated filte1-6-7 and diffusion denuder tubes coated with sodium carbonate.8-1') A tungstic acid denuder technique was also developed in which the collected HN03 is thermally decomposed to NO and the NO is detected by a chemi- luminescence NO,. analyser.11 The methods combined with these techniques produce average or integrated HN03 con- centrations. A specific and direct method has been reported by Tuazon et al., I 2 which uses a 1 km folded path cell coupled to a Fourier transform spectrometer. Although this method may be the best approach because of its selectivity and capability of continuous monitoring, it requires complex instrumentation and its detection limit of about 5 ppb is often above ambient HN03 levels. A chemiluminescence method for continuous monitoring of HN03 has been developed by Joseph and Spicer,I3 using a modified commercial NO, analyser and a nylon filter. This method measures HN03 as the difference between total NO, including HN03 determined directly and the NO, determined after selective removal of HN03 on a nylon filter.The detection limit of the technique (0.8 ppb) is sufficient for studies in urban areas, but more sensitive techniques are required for studies in non-urban areas. We have recently developed a chemiluminescence method for the continuous monitoring of ambient gaseous HN02.14 The method is based on collection of HN02 by concurrently drawing the air and scrubbing solution through a glass coil, and the subsequent reduction of nitrite to NO by ascorbic acid followed by detection with a chemiluminescence NO, ana- lyser. The sensitivity of the method is a function of the ratio of sampling flow rate to carrier gas flow rate of sweeping the evolved NO into an NO, analyser. This permits readily the highly sensitive measurement of HNO?.The advantage of the technique led us to apply it to the continuous monitoring of gaseous HN03. Several chemiluminescence methods based on the direct reduction of nitrate to NO have been developed for the determination of trace amounts of nitrates in aqueous samples using reducing agents such as iron(ii)-molybdatels~l~ and vanadium(ii1). 17 These methods require highly acidic condi- tions andor operation at a relatively high temperature (80-100 "C). The method described here consists of two steps for the reduction of nitrate to NO. Nitrates are reduced to nitrites with hydrazine sulfate under mild alkaline conditions in the presence of CU" at a temperature in the range 30-50°C and the nitrites are then reduced to NO with ascorbic acid under moderately acidic conditions.The measurement systems with and without the hydrazine reduction procedure determine total HN03-HN02 and HN02, respectively, and HN03 is calculated from the difference. This paper describes the results of investigations on the reduction of nitrate by hydrazine and some preliminary results of ambient measure- ments. Experimental Choice of Reductant Many reducing agents have been used for the spectropho- tometric determination of nitrate based on the reduction of nitrate to nitrite followed by diazotization-coupling reactions. These include hydrazine sulfate, 18-*O titanium(ii1) chloride," amalgamated zinc,*' cadrnium,23.24 amalgamated cadmium,2s copper-coated cadmium ,24.*h-*X and copper-coated cadmium- silver.'"28 Of these reducing agents, copper-coated cadmium is the most widely used.It was therefore decided to investigate the application of the copper-coated cadmium reduction technique to a continuous monitoring system for gaseous HN03. A reduction coil was made from 1.2 mm i.d. poly( tetrafluoroethylene) (PTFE) tubing. A cadmium wire (diameter 1 mm, length 50 cm) was inserted into the tubing and was treated with dilute copper sulfate solution. The quantitative reduction of nitrate to nitrite was obtained with a stream of 5 x 10-3 mol dm-3 ethylenediaminetetraacetic acid (EDTA) in borate buffer of pH 8 at flow rates ranging from 0.15 to 0.30 ml min-1. However, the reducing power deteriorated with use because of contact with air bubbles accidentally introduced into the coil, and the reproducible884 ANALYST, MAY 1992, VOL.117 regeneration of the reductant was difficult. A homogeneous reduction procedure appears to be more suitable for a continuous-flow system. In this study, alkaline hydrazine solution was therefore used for the reduction of nitrate to nitrite. Reagents All chemicals used were of analytical-reagent grade from Wako Chemical Industries. Reagent solutions were prepared with high-purity water from a Millipore Milli-Q purification system. Stock standard solutions containing 100 mg 1-1 of nitrate-N and 100 mg 1-1 of nitrite-N were prepared in water from potassium nitrate and sodium nitrite dried at 11O"C, respectively. Working standard solutions were prepared in 0.05 rnol dm-3 NaOH by appropriate dilution of the respect- ive stock standard solutions.Stock solutions of 0.1 rnol dm-3 hydrazine and 0.01 mol dm-3 CU" were prepared by dissolving N2H4aH2S04 and CuS04.5H20 in water, respectively. A 0.2 rnol dm-3 ascorbic acid solution was prepared in 0.1 rnol dm-3 H2S04. Standard gas sources were the same as those used pre- viously.*4 They included a diffusion tube device for HN03, a continuous generation system for HN02 and a cylinder gas mixture for NO. The carrier gas was room air freed of NO, compounds by passage through a cobalt(iI1) oxide coated denuder. Apparatus A schematic diagram of the measurement system is shown in Fig. 1. The system consists of two sets of dual-channel flow systems. One is for total HN03-HN02 measurement and the other for HN02 measurement. The system for total HN03- HN02 is identical with that for HN02 except that it includes reaction coils for the reduction of nitrate to nitrite.In each system the second channel measures the interference effects from NO2 and peroxyacetyl nitrate (PAN). Gaseous HN03 and HN02 were stripped from the air sample into the solution by concurrently drawing the air and F G r Air sample scrubbing solution through a glass stripping coil. The reaction coil for the reduction of nitrate to nitrite by hydrazine was made from a 1 m length of 0.68 mm i.d. PTFE tubing and was immersed in a constant temperature water-bath. A gas-liquid separating coil, made from microporous PTFE tubing, was used for the reduction of nitrite to NO and the successive scrubbing of NO. Details of the stripping and gas-liquid separating coils have been described elsewhere.14 Air sample and carrier gas flow rates were controlled by calibrated mass flow controllers (SEC-410, Stec Corp.). Constant and con- tinuous supplies of reagent solutions were delivered by four- and two-channel peristaltic pumps (Gilson Minipuls-2, Gilson Medical Electronics). Measurements of HN03, HN02, and NO were made with Monitor Labs Model 8840 NO, analysers. The instrument is a dual-channel monitor with two reaction chambers and two photomultiplier tubes. In the measure- ments of ambient HN03 and HN02, the NO, flow line in the instrument was modified to lead directly to the reaction chamber by by-passing the converter to measure simul- taneously the NO from the dual flow system. Ambient Measurement The experimental conditions for measurements of ambient HN03 and HN02 were: sampling flow rate, 2.0 1 min-1; scrubbing solution, 0.10 ml min-1 of 0.05 rnol dm-3 NaOH; reduction of nitrate to nitrite, 0.10 ml min-1 of 2.0 X 10-3 rnol dm-3 hydrazine-1.6 x 10-5 rnol dm-3 Cu" at 30°C; reduction of nitrite to NO, 0.10 ml min-1 of 0.2 rnol dm-3 ascorbic acid in 0.1 rnol dm-3 H2S04; purging flow rate, 0.29 1 min-* in the total HN03-HN02 system and 0.35 1 min-* in the HN02 system.Results and Discussion Optimization of the Conditions for the Reduction With Hydrazine The effects of several parameters on the reduction of nitrate to nitrite by hydrazine were investigated. The parameters were: Fig. 1 Schematic diagram of the measurement systems for (a) HN07-HN02 and (h) HNO?. A. Stripping coil; B, gas-liquid separating coil; C, reaction coil; D, peristaltic pump; E, de-bubbler; F.mass flow'controller; G, air pump; H, clean air input; I , NO, analyser; and J. recorderANALYST, MAY 1992, VOL. 117 60 1 I 40 - 1 9 P 0 z - 20 . 0 1 .o 2.0 3.0 Hydrazine sulfate/l0-3 rnol dm-3 Fig. 2 Effect of hydrazine concentration on the reduction of nitrate (open symbols) and nitrite (closed symbols). Nitrate and nitrite concentration. 0.2 pg ml-I in 0.05 rnol dm-3 NaOH; Cu" in hydrazine solution, 1.6 x 10-5 mol dm-3; pumping rate, 0.10 ml min-1; ascorbic acid, 0.2 mol dm-3 in 0.1 mol dm-3 H2S01; and temperature, 30°C NaOH, hydrazine and Cull concentrations, reaction time and temperature. Previous work14 showed that nitrites can be quantitatively reduced to NO with acidic ascorbic acid. The effects were therefore evaluated by determining the NO evolved by the subsequent reduction by ascorbic acid of the nitrite produced.Measurements were performed by varying one parameter over a range of values while holding the others constant, using nitrate and nitrite standard solutions, and a 1 m length of reaction coil. Effect of NaOH concentration For assessment of the effect of the NaOH concentration on the reduction of nitrate to nitrite, each of a series of 0.2 yg ml-1 ni trate-N standard solutions prepared in different concentra- tions of NaOH and a 2.0 x 10-3 rnol dm-3 hydrazine solution containing 1.6 x 10-5 rnol dm-3 Cu" were separately pumped into the reaction coil at the same flow rate of 0.10 ml min-1. The mixed solution was then introduced into the gas-liquid separating coil, together with a 0.1 ml min-1 flow of a 0.2 rnol dm-3 ascorbic acid solution, and the NO evolved from the coil was measured.The maximum reduction efficiency was obtained when the NaOH concentration was between 0.02 and 0.16 rnol dm-3. A 0.05 rnol dm-3 NaOH solution was chosen as the scrubbing solution. Effect of hydrazine concentration As shown in Fig. 2, the yield of NO from nitrate steadily increased with hydrazine concentration and became equal to that from nitrite at 1.6 X 10-3 rnol dm-3. At higher concentrations, the yield of NO from both nitrate and nitrite showed a slight decrease, which apparently indicates that nitrite produced by reduction of nitrate is simultaneously being further reduced. The magnitude of the decrease in the NO yield caused by this over-reduction was about 20.7% at 3.0 x rnol dm-3 hydrazine. Effect of Cu" concentration Mullin and Riley18 have reported that reduction with hydraz- ine requires a small amount of Cu" to catalyse the reaction.In order to determine the effect of Cu" concentration on the reduction of nitrate to nitrite, a 0.2 yg ml-1 nitrate-N solution and each of a series of 2.0 x 10-3 rnol dm-3 hydrazine solutions containing different amounts of Cu" were separately pumped into the reaction coil at the same flow rate of 0.10 ml min- 1, and the NO evolved by the subsequent reduction by ascorbic acid was measured. The results showed that the system requires more than 1.4 x 10-5 mol dm-3 Cu" for the quantitative reduction of nitrate and that the reduction is not affected by Cull concentrations up to 3.0 x 10-5 rnol dm-3.60 - 40 s P 0 z Y 20 0 20 30 40 50 Tern peratu rePC 885 Fig. 3 Effect of temperature on the reduction of nitrate (open symbols) and nitrite (closed symbols). Nitrate and nitrite concentra- tion, 0.2 pg ml-1 in 0.05 rnol dm-3 NaOH; hydrazine, 2.0 x 10-3 mol dm-3 containing 1.6 X 10-3 mol dm-3 Cu"; ascorbic acid, 0.2 rnol dm-3 in 0.1 rnol dm-3 H2S04; and flow rate, 0.10 ml min-l 100 80 - 9 Q Q Z 1 0 6o 40 20 0 0.10 0.20 Pumping rate/ml min-1 Fig. 4 Effect of reaction time on the reduction of nitrate (open symbols) and nitrite (closed symbols). The reaction time is the residence time of the reaction mixture in the coil, which is inversely proportional to the pumping rate. In this system, a 0.10 ml min-1 pumping rate corresponds to a residence time of 2 min Effect of temperature Kamphake et a1.19 reported that the reduction of nitrate and nitrite with hydrazine was very sensitive to temperature; a quantitative reduction of nitrate to nitrite was obtained in a very narrow temperature range of 35-38 "C and increasing the temperature led to rapid reduction of nitrite.As shown in Fig. 3, however, the influence of temperature was not so significant in the present system. The yield of NO from nitrate became equal to that from nitrite at 30°C and remained almost constant up to 50°C. Effect of time Another important parameter is reduction time. Measure- ment of the effect of time was carried out by varying the residence time of the reaction mixture in the coil, in practice, by varying the pumping rates of the reagent solutions using a fixed length of coil.Each of a series of standard solutions of 0.2 pg ml-1 of nitrate-N and 0.2 yg ml-1 of nitrite-N, and a 2.0 x 10-3 rnol dm-3 hydrazine-1.6 X 10-5 rnol dm-3 Cu" solution were separately pumped into a 1 m length of coil at the same rates ranging from 0.05 to 0.20 ml min-1, corre- sponding to residence times of approximately 4-1 min. The results are shown in Fig. 4. If the reduction of nitrate to nitrite proceeds quantitatively in the residence times examined, a linear relationship should be obtained between the yield of NO and the pumping rate of886 1.0 ANALYST, MAY 1992, VOL. 117 ( b ) - the reagent solutions when the evolved NO is purged at a fixed rate.Fig. 4 indicates that at a pumping rate in the range 0.10-0.14 ml min-1 the reduction of nitrate to nitrite is quantitative; below 0.10 ml min-' over-reduction of nitrate occurs and above 0.14 ml min-1 the reduction of nitrate to nitrite is not quantitative because of the insufficient reaction time. Collection of HN03 and HN02 The collection efficiency of the stripping coil for gaseous HN03 and HN02 was determined by monitoring the re- spective standard gas flows upstream and downstream of the coil with an NO, analyser. Measurements were carried out by varying the sampling rate from 0.5 to 4.0 1 min-1, while pumping a 0.05 mol dm-3 NaOH solution into the coil at a fixed rate of 0.10 ml min-1. The collection of HN03 and HN02 was found to be quantitative at sampling rates up to 3.0 1 min-1 for both HN03 and HN02.Interferences Of the common air pollutants concomitant with HN02 and HN03, NO2 and PAN yield nitrate and/or nitrite ions in alkaline solution and, hence, would produce positive interfer- ences. 2N02 + 20H- -+ NO2- + NO3- + H 2 0 (1) CH3C(O)OON02 + 20H- + CH3COO- + N02- + 0 2 + H20 (2) A previous study14 on the interference of NO2 and PAN with the measurement of HN02, in which 5 x 10-3 mol dm-3 Na2C03 was used as the scrubbing solution, revealed that NO2 and PAN produce only a 0.7 and 1.9% positive interference, respectively. Peroxyacetyl nitrate decomposes in alkaline solutions to yield only nitrite ion, whereas NO2 yields equimolar amounts of nitrite and nitrate ions. Hence the effect of the interference from NO2 on the total HN03-HN02 measurement system is twice that on the HN02 measurement system; it is estimated to be 1.4%.This was confirmed by interference experiments using a scrubbing solution of 0.05 mol dm-3 NaOH. In order to correct for these interferences, both the total HN03-HN02 and HN02 measurement systems were made up of two equivalent flow systems with two stripping coils connected in series. The interfering effects were measured with the respective second flow system. Sensitivity and Calibration Graphs As demonstrated in the previous paper,l4 the analytical sensitivity of the technique, expressed as the slope of the straight line calibration graph, is proportional to the ratio of sampling flow rate to purging flow rate of the NO produced by chemical reduction and is given by rn = (qs/qphlr)2r)3 where qs is the sampling flow rate, qp is the purging flow rate, ql is the collection efficiency of gaseous HN03 or HN02, q 2 is the yield of NO2- from reduction with hydrazine (= 1.0 for the HN02 measurement system) and q3 is the yield of NO from reduction with ascorbic acid.In the present study, q l was greater than 0.99 for both HN03 and HNO2, q 2 was 0.86 and q3 was greater than 0.99. The equation obviously shows that high sensitivity can be obtained when the sampling flow rate is high and the purging flow rate is low; however, a high sampling flow rate results in a decrease in the collection efficiency and a low purging flow rate is limited by the sample flow rate demanded by the NO, monitor. In order to check the analytical performance and linearity of the systems, the calibration graphs were generated by using a sampling flow rate of 2.0 1 min-1 and a purging flow 0 1 ' 1 I I I 1.0 I I I I I I I I 1 12 18 24 6 12 Time of day/h Fig.5 Ambient HN03 and HN02 measured in Tsukuba: (a) June 17-18, 1991; curve 1. first channel - second channel; and curve 2, second channel; and (b) June 20-21. 1991. cL indicates the limit of detection rate of 0.29 1 min-1. Plotting the NO in the total HN03-HN02 measurement system against HN03 concentration in the range 6.5-26.5 ppb gave a linear calibration with a slope of 5.88 and a correlation coefficient of 0.998, which was the same, within statistical variation, as that obtained for HN02 in the range 2.1-20.5 ppb. A slope of 6.85 was obtained for the calibration graph for HN02 in the HN02 measurement system.These slopes were in good agreement with the values of 5.94 and 6.90 expected from the equation for the sensitivity when the sampling flow rate is 2.0 1 min-1 and the purging flow rate is 0.29 1 min-1. In practice, it is desirable that the calibration graphs for the total HN03-HN02 and HN02 measurement systems have the same slope in order to simplify the calculation of the HN03 concentration, which is determined by subtracting the value for HN02 from that for total HN03-HN02. The above equation for the sensitivity clearly indicates that a coincidence of the slopes of the calibration graphs for the two measure- ment systems can easily be obtained by setting the purging flow rate in the HN02 system at a rate of l/q2 (1.16 in this system) times that in the total HN03-HN02 system.The limit of detection was calculated to be approximately 0.13 ppb, using k = 3 in the equation cL = ksB/m,2g where s6 is the standard deviation of the blank and rn is the slope of the calibration graph. Ambient Measurement In order to assess the performance of the system, air was sampled at a rural site, Tsukuba, located approximately 60 km north-east of Tokyo. After removal of nitrate particles byANALYST, MAY 1992. VOL. 117 887 passage through a PTFE filter, the air stream was divided equally between the two measurement systems. Typical time profiles for gaseous HN03 and HNOZ obtained with a single continuous 24 h measurement are shown in Fig. 5. The concentrations of HN03 and HN02 showed very characteristic diurnal profiles; the HN03 concentration exhi- bited a maximum during the day and a much lower (sometimes below the detection limit) level at night.In contrast, the HNO2 level built up during the night and decayed rapidly after sunrise. During this study, the NO2 concentrations measured were in the range 4-23 ppb and the measurement of PAN was not carried out. In the measurement of total HN03-HNOZ on June 17-18, 1991, as shown in Fig. 5(a), the maximum contribution of NO2 to the NO evolved from the second channel was about 1.5 ppb, which corresponds to 0.13 ppb of HN03 and 0.13 ppb of HN02. In the measurement of HN02, no observable amount of NO was evolved from the flow system of the second channel. The proposed method has been demonstrated to be a sensitive technique for the continuous monitoring of HN03 and HNOZ.The accuracy is, however, affected by artefact formation of HN03 caused by the separation of gaseous and particulate nitrates. As the proposed method cannot discrimi- nate between gaseous and particulate nitrates, it is necessary to remove particulate nitrates from the air stream prior to the measurement of gaseous HN03. Although a PTFE filter was used in this study, the removal of particulate nitrate on a filter is known to result in the loss of particulate nitrate and generation of artifact HN03 by the reaction between H2S04 aerosols and deposited nitrate salts.7.”)J1 This problem appears to be solved by use of the diffusion denuder technique, which is capable of separating gases from particles without particle-particle interactions.A continuous sampling technique based on the principle of the diffusion denuder is currently under development in this laboratory. References Spicer, C. W.. Environ. Sci. Technol.. 1983. 17, 112. Galloway, J. N., and Likens, G. E., Atmos. Environ., 1981,15, 1081 * Durham, J . L., Overton, J. H., and Aneja. V. P.. Armos. Environ., 1981. 15, 1059. Appel, B. R.. Wall, S. M., Tokiwa, Y., and Haik, M., Atmos. Environ.. 1980, 14, 549. 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 30 31 Shaw, R. W., Jr., Steven, R. K., and Bowermaster, J., Atmos. Environ., 1982, 16, 845. Okita, T., Morimoto. S., and Izawa, M., Armos. Environ., 1976, 10, 1085. Forrest, J.. Tanner, R. L., Spandau, D., D’Ottario, T., and Newman, L., Armos.Environ., 1980, 14, 137. Forrest, J . , Spandau, D. J . , Tanner, R. L., and Newman. L., Atmos. Environ., 1982, 16, 1473. Levin. E. E., and Hansen, K. A.. Anal. Chem., 1984, 56,842. Ferm, M., Atmos. Environ., 1986, 20, 1193. Braman, R. S . , Shelly, J. J., and McClenney. W. A., Anal. Chem.. 1982,54, 358. Tuazon. E. C., Graham, R. A., Winer, A. M., Easton. R. R., Pitts. J. N., Jr.. and Hanst, P. L., Atmos. Environ., 1978, 12, 865. Joseph, D. W., and Spicer. C. W., Anal. Chem., 1978,50, 1400. Kanda, Y., and Taira, M., Anal. Chem., 1990, 62, 2084. Cox, R. D., Anal. Chem., 1980, 52, 332. Yoshizumi, K., Aoki, K., Matsuoka, T., and Asakura, S., Anal. Chem.. 1985, 57, 737. Braman, R. S., and Hendrix, S. A., Anal. Chem., 1989, 61, 2715. Mullin, J. B., and Riley, J. P., Anal. Chim. Acta, 1955,12.464. Kamphake, L. J . , Hannah, S. A., and Cohcn, J. M.. Water Res., 1967, 1, 205. Madsen, B. C., Anal. Chim. Acta, 1981, 124, 437. Al-Wehaid, A., and Townshend, A., Anal. Chim. Acta. 1986. 186, 289. Bajic, S. J.. and Jaselskis, B., Talanta. 1985, 32, 115. Margeson, J . H., Suggs, J . C., and Midgett. H. V., Anal. Chem., 1980, 52, 1955. Koupparis, M. A., Walczak, K. M., and Malmstadt, H. V., Anal. Chim. Acta. 1982. 142. 119. Morris, A. W.. and Riley, J. P., Anal. Chim. Acra. 1963, 29, 272. Anderson, L., Anal. Chim. A m , 1979, 110, 123. Gine, M. F., Reis, B. F., Zagatto, A. E. G., Krug, F. J . , and Jacintho, A. O., Anal. Chim. Acta. 1983. 155. 131. Willis, R. B.. Anal. Chem.. 1980, 52, 1376. Long. G . L., and Winefordner, J. D., Anal. Chem.. 1983, 55, 712A. Harker, A. B.. Richards. L. W., and Clark, W. E., Atmos. Environ.. 1977. 11, 87. Appel, B. R., andTokiwa. Y.. Armos. Environ.. 1981,15, 1087. Paper 1105189J Received October 14, 1991 Accepted December 2, I991

ISSN:0003-2654    

DOI:10.1039/AN9921700883

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, MAY 1992, VOL. 117 889 Application of Dowex-2 Loaded With Sulfonephthalein Dyes to the Preconcentration of Copper(ii) and Cadmium(ii) Ajai Kumar Singh and Surendra Kumar Dhingra Department of Chemistry, Indian Institute of Technology, New Delhi-I 10 016, India Dowex-2 loaded with sulfonephthalein dyes, viz., Pyrocatechol Violet (PV) and Xylenol Orange (XO), was prepared and investigated for the preconcentration of Cull and Cd" prior to their determination by flame atomic absorption spectrometry. At pH 7.0--8.0,1 g of the resin sorbs 23-26 mg of PV or XO. Copper(r1) (2 0.01 ppm) is sorbed quantitatively at pH 5.0-10.0 and 6.0-10.0 by the PV- and XO-loaded resin, respectively. The corresponding pH ranges for maximum sorption of Cd" (30.01 ppm) are 9.5-10.0 and 8.0-10.0. Copper and Cd were desorbed quantitatively with 1 mol dm-3 HCI and 1 mol dm-3 HN03, respectively.The metal ion sorbing capacity of the two resins was found to be <I00 pg of metal per 100 mg of resin. The sorption decreases when the concentration of NaCl exceeds 0.1 mol dm-3. The applicability of both resins to the preconcentration of Cu and Cd prior to their determination by flame atomic absorption spectrometry was demonstrated by analysing Yamuna river water samples for the two ions (relative standard deviation = 4.8-5.7%). Keywords: Sulfonephthalein dye; Dowex-2; preconcentration; copper(ii); cadmium(ii) There is continued interest among analytical chemists in the development of chelating polymeric resins,' which can pro- vide more flexible working conditions together with good stability and a high capacity for metal ions.This is because such resins can be used for preconcentration purposes in trace metal analysis, particularly for water systems. Several types of chelating reagent have been loaded on a polymeric matrix in order to prepare such resins,2-7 either by forming a covalent bond between the reagent and the matrix or by an ion- exchangeln-n interaction. The systems prepared by ion exchange involve an anionic group which does not participate in chelation, e . g . , a sulfonate group. The main advantage of an ion exchange based sorption system is the ease with which the active part can be varied, without recourse to the often difficult and time-consuming procedures needed for covalently linking the reagent to the skeleton of the resin.However, such systems do not always exhibit very good stability. If ion exchange were to be supported by n-n dispersion forces between the reagent and the polymer matrix, the stability would be improved. R = -CH,COOH xo H03S ' I 9- OH HO PV Therefore, the loading of sulfonephthalein dyes, by ion exchange, on a polymeric matrix having x-electron systems might result in a more stable chelating resin. In this work, Pyrocatechol Violet (PV) and Xylenol Orange (XO), which have a conjugated x-electron system and a sulfonate group, were immobilized on the anion-exchange polystyrene-based resin Dowex-2 and the resulting chelating polymers were investigated for the preconcentration of Cu" and Cd" prior to their determination by flame atomic absorption spectrometry (AAS).In order to demonstrate the utility of these resins, Cull and Cd" present in river water samples were preconcentrated and then determined by AAS. Experimental Apparatus and Reagents Atomic absorption spectrometric measurements were made with a Pye Unicam SP 191 atomic absorption spectrometer, using an air-acetylene flame (air and acetylene flow rates, 4 and 1 dm3 min-1, respectively). The pH measurements were made on an Elico Model LI 120 digital pH meter. Spectropho- tometric measurements were made on a Perkin-Elmer Lambda-3 ultraviolet/visible spectrophotometer. The stock solutions (0.01 rnol dm-3) of Cu" and Cd" were prepared by dissolving appropriate amounts of analytical- reagent grade copper(I1) sulfate and cadmium(i1) chloride, respectively, in doubly distilled water acidified with a small amount of the corresponding acid and standardized8 before use.The pH adjustments were made with 0.01-1.0 mol dm-3 HCl and NaOH. In some instances, hexamine buffer was also employed for adjusting the pH to about 6.0. The anion- exchange resin Dowex-2 (moderately basic, chloride form, 8% cross-linked) (200400 mesh) was obtained from Sigma. Pyrocatechol Violet [BDH (now Merck)] was recrystallized from a 10% v/v ethanol-water mixture before use. Xylenol Orange (BDH) was used as received. All other reagents used were of analytical-reagent grade. The glassware used was soaked in 5% HN03 for 1 week before use and cleaned with doubly distilled water. The water samples from the Yamuna river (near the Okhla industrial area, New Delhi, India) were collected in clean polyethylene bottles and concentrated HN03 (2 cm3 per dm3 of water) was added to each sample immediately after its collection.890 ANALYST, MAY 1992, VOL.117 Procedure for Regenerating Dowex-2 Dowex-2 (2 g) was equilibrated with 6 rnol dm-3 HCl(20 cm3) for 3 h and then washed successively with water (50 cm3), ammonia solution (40 cm3) and water until the filtrate was colourless towards phenolphthalein. The resin was finally washed with methanol, dried and stored in a desiccator in vacuo. Procedure for Loading PV on Dowex-2 Dowex-2 (100 mg) was mixed with 20 cm3 of a solution of PV (0.6 mmol dm-3). After adjusting the pH to 7.5-8.0, the mixture was shaken on a mechanical shaker for 8 h. The resin was then filtered, washed successively with water and methanol, dried and stored in vacuo.Procedure for Loading XO on Dowex-2 Dowex-2 (300 mg) was mixed with 75 cm3 of XO (0.14 mmol dm-3). After adjusting the pH to 7.0-7.5, the mixture was shaken for 3 h. The resin was subsequently filtered, washed successively with water and methanol, dried and stored in vacuo. Recommended Procedure for Preconcentration and Determination of Cd A mixture containing 100 mg of PV-loaded chelating resin and 25 cm3 of a solution containing not more than 2 pg cm-3 of Cd" was taken and its pH was adjusted to 9.5-10.0 with 0.1 rnol dm-3 NaOH. The mixture was equilibrated for 20 min on a mechanical shaker after which the resin was filtered and washed with 20 cm3 of water. Cadmium(xx) ions were desorbed by treating the resin with two 2.5 cm3 aliquots of 0.05 rnol dm-3 HN03.The acidic solution containing desorbed Cd" was aspirated into an air-acetylene flame for measurement of the absorption at 228.8 nm. The absorbance values were corrected for the blank and read from a previously constructed calibration graph. The same procedure was used for the XO-loaded Dowex-2, except that the working conditions were as follows: volume of Cd" solution, 40 cm3; pH, 8.0-8.7; shaking time, 10 min; and desorbing solution, 10 cm3 of 1 rnol dm-3 HN03. Recommended Procedure for Preconcentration and Determination of Cu The pH of a mixture containing 100 mg of PV-loaded chelating resin and 25 cm3 of a solution containing not more than 2 pg 6171-3 of Cu was adjusted to 6.0 with hexamine buffer.The mixture was equilibrated for 30 min on a mechanical shaker after which the resin was filtered and washed with 20 cm3 of water. Copper(I1) ions were desorbed by equilibrating the resin with two 2.0 cm3 aliquots of 1 rnol dm-3 HCl for 1 h. The acidic solution containing desorbed Cu" was aspirated into an air-acetylene flame for measurement of the absorption at 324.8 nm. The absorbance values were corrected for the blank and read from a previously constructed calibration graph. The same procedure was used for the XO-loaded resin, except that the pH was adjusted to 9.0 with 0.1 rnol dm-3 NaOH and the concentration of Cu" in the solution was not more than 4 pg cm-3. Results and Discussion Pyrocatechol Violet and XO are among the most versatile chelating agents capable of forming complexes with a variety of metal ions.Both can be sorbed on Dowex-2 apparently by exchange of the C1- ion of the resin matrix with the sulfonate group attached to the dyes and by means of n-rc dispersion forces arising from the aromatic nature of the resin and the reagent. The sorption of the two dyes on Dowex-2 was monitored spectrophotometrically at the maximum wavelength of absorption and was found to be pH dependent. At pH 7.0-8.0, maximum sorption is obtained if the resin is equilibrated for 2 and 7 h with XO and PV, respectively. Longer shaking times, however, do not adversely affect the sorption process. One gram of resin was found to sorb 25.7 mg of XO or 23 mg of PV. By using the batch method, both chelating resins were investigated €or the preconcentration of Cu" and Cd". Optimum Conditions for Sorption The sorption of Cull and Cd" on PV- and XO-loaded resins was studied at different pH values, keeping the other parameters constant. Optimum pH values for maximum and constant sorption of the two metal ions are given in Table 1; pH values above 10.0 were not investigated because the mixtures became turbid.Hexamine buffer was found to be suitable for adjusting the pH to about 6.0 for Cu". Sorption of both metal ions on PV- and XO-loaded Dowex-2 was studied by varying the equilibration time from 5 to 60 min. The optimum amounts of Cu" and Cd" sorbed on 100 mg of each of the two resins are also given in Table 1. Copper(i1) and Cd" (10 pg each) were first sorbed on 100 mg of PV- or XO-loaded Dowex-2 and then desorbed by using the recommended procedure.The average percentage recovery of three experi- ments was determined and the values are given in Table 1. Effect of NaCl on Sorption Sorption of both metal ions decreases slowly when the concentration of NaCl is increased beyond 0.1 rnol dm-3. In the presence of 0.2,0.4 and 0.6 rnol dm-3 NaCl, the sorption was found to be 86, 81 and 74%, respectively, for both ions. Limit of Preconcentration The two metal ions can be collected effectively (94-97%) on either of the two chelating resins from their solutions having a concentration of the order of 0.01 ppm. Sorption of the two ions was neither quantitative nor reproducible at lower concentrations. Desorption of Metal Ions From the Resins The desorption of Cd from both resins was instantaneous with 1 rnol dm-3 HN03. However, in order to desorb Cu quantitatively, equilibration of the resins with 1 rnol dm-3 HCI for 1 h on a mechanical shaker is required.Determination of Cu and Cd in River Water Samples One gram of PV- or XO-loaded chelating resin was shaken successively with five 100 cm3 aliquots of a filtered water sample as described in the recommended procedure. The resin was separated by filtration and treated successively with two aliquots of 25 cm3 of 1 rnol dm-3 HN03 and 1 rnol dm-3 HCI. Table 1 Optimum conditions for sorption of metal ions PV-loaded resin XO-loaded resin Parameter Cu11 Cd" CU" Cd" pH range 5.0-10.0 9.5-10.0 6.0-10.0 8.0-10.0 Equilibration time/min 20 10 30 10 Metal ion sorbed per 100 mg of resin/yg 50.0 62.5 100.0 80.0 Recovery (%) 97.4 98.3 96.7 98.2ANALYST, MAY 1992, VOL.117 89 1 Table 2 Determination of Cu and Cd (pg per 100 cm3) in Yamuna river water samples PV-loaded resin XO-loaded resin Sample Method CU RSD(%) Cd RSD(%) CU RSD(%) Cd RSD(%) 1 Direct 4.9 5.6 8.0 5.2 5.2 5.4 8.7 5.5 SA* 5.3 4.8 8.8 4.9 5.6 5 .O 8.2 5.3 2 Direct 5.2 5.6 8.1 4.8 4.8 5.7 8.8 5.1 SA" 5.8 5.0 8.7 4.8 5.3 5.6 8.4 5.2 * SA = Standard additions method. All the desorption solutions were mixed and the metal ions in the mixture determined by AAS as described above. The results are given in Table 2 and compared with the values obtained by the standard additions method. The relative standard deviation (RSD) for six determinations was calcu- lated.The results are presented in Table 2. The optimum pH range for the sorption of Cu" is wider than that for Cd". Dowex-2 loaded with PV offers a better pH range for the collection of Cu than either XO-loaded Dowex-2 or the other known preconcentration systems.' This is a significant advantage. Sorption of Cd on the resins prepared in this work is much faster than on other commonly used ion-exchange and chelating resins. 1 Dowex-2 loaded with XO has a greater sorbing capacity than the PV-modified system. Recoveries are comparable to those obtained' for ion exchangers, other chelating resins, polyurethane foam and polystyrene beads. The greater compatability between the size of the hydrated Cd" ion and the pores of the resins prepared in this work is probably responsible for the greater sorption of this ion compared with Cu".The instantaneous desorption of Cd" is another advantage of these modified resins. They are comparable to other functionalized resins1 in terms of the lower limit of preconcentration. For the analysis of river water, the performances of the direct and standard additions methods are similar, although the latter gives higher results and a lower RSD than the former. The discharge of industrial effluent into the river at a site very close to that at which the sampling was performed is probably responsible for the high Cu and Cd contents of the water samples analysed in this work. Financial support of this work from the Council of Scientific and Industrial Research (India) is gratefully acknowledged. References 1 Kantipuly, C., Katragadda, S., Chow, A . , and Gesser, H . D . , Talanta, 1990,37,491. 2 Nakayama, M., Chikurna, M., Tanaka, H . , and Tanaka, T., Talanta, 1982, 29, 503, and references cited therein. 3 Pesavento, M., Profumo, A . , and Biesuz, R . , Talanta, 1988,35, 431. 4 Chwastowska, J . , and Kosiarska, E., Talanta, 1988, 35, 439. 5 Brajter, K., Olbrych-Sleszynska, E., and Staskiewicz. M., Talanta, 1988, 35, 65. 6 Singh, A. K . , and Kurnar, T. G. S . , Microchem. J . , 1989, 40, 197. 7 Mendez, R., and Pillai, V. N. S., Analyst, 1990, 115, 213. 8 Vogel, A. I . , A Text-Book of Qualitative Inorganic Analysis, Longman, London, 3rd edn., 1973. Paper 1100702E Received February 14, 1991 Accepted November 6, 1991

ISSN:0003-2654    

DOI:10.1039/AN9921700889

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, MAY 1992, VOL. 117 893 Use of Hydrous Iron(iii) Oxide in a Concentration Step for the Determination of Trace Amounts of Organophosphorus Compounds in Aqueous Solutions Toshitaka Hori and Masahito Sugiyama" Department of Chemistry, College of liberal Arts and Sciences, Kyoto University, Kyoto 606, Japan Adsorption onto hydrous iron oxide (HIO) was compared as a function of pH for a variety of organophosphorus compounds (OPs), including phosphate esters of ethanolamine, hydroxyamino acids and sugars, phosphonates with methyl and aminoethyl substituents, and nucleotides. The percentage adsorption versus pH curves could be classified into four types according to an empirical rule, viz., that the adsorptivity of OPs depended primarily on the number of unsubstituted P-0 moieties in the tetrahedral structure around the P atom of the compound. The rule predicted that a large group of OPs containing more than three unsubstituted P-0 moieties should be collected quantitatively by HI0 from waters of pH 5.0-6.5.The OPs collected by adsorption onto HI0 did not show appreciable degradation during storage for at least 2 weeks. In addition, they could be released from the HI0 by using pentane-2,4-dione [acetylacetone (Hacac)] so that they entered a small-volume aqueous phase which was derived from the HI0 by the following reaction:Fe203-nH20- (OPs) + 6Hacac -+ 2Fe(a~ac)~ + (n + 3)H20 + (OPs). The whole procedure, involving adsorption of OPs onto HI0 from a 1 I water sample, separation of the HI0 from water by filtration and release of the OPs from the HI0 into a 2.5 ml aqueous phase, realized a 400-fold concentration with efficiencies ranging from 45% (for adenosine-5'-triphosphate) to 92% (for 2-aminoethylphosphonate).Keywords: Organophosphorus compounds; adsorption; hydrous iron oxide It is known that the use of 4-amino-3-hydroxynaphthalene-l- sulfonic acid as a reducing agent for the Heteropoly Blue method greatly improves the determination of orthophos- phate concentrations.' The so-called Fiske-Subbarow pro- cedure' not only provides reliable analytical results for the phosphate content of blood and urine samples but has also found wide application in the discovery of new types of organophosphorus compounds (OPs) ,273 many of which are nowadays known to act as cell membrane components, intermediates in energy metabolism and essential constituents of genes.Ishibashi and Tabushi4 proposed a method for the sensitive determination of orthophosphate concentrations by a series of procedures involving the adsorption of orthophosphate onto hydrous iron oxide (HIO), the addition of acidic molybdate solution to dissolve the H I 0 and obtain molybdophosphate, and the measurement of the absorbance of molybdophosphate extracted into butyl acetate. The method has since been used to determine orthophosphate concentrations down to 1 x 10-8 mol dm-3; orthophosphate is found in the hydrosphere as an essential nutrient for micro-organisms.5 Owing to the difficulty in designing an automated procedure, however, the method of Ishibashi and Tabushi4 has now been replaced by that of Murphy and Riley,6 which is based on the formation of heteropoly blue with the use of ascorbic acid and antimony1 tartrate.Nevertheless, the former method revealed that HI0 can adsorb some OPs in addition to orthophosphate,7 although systematic adsorption studies on a large group of OPs have not yet been performed. Hori et al.8 measured the percentage adsorption onto H I 0 of a selection of oxophosphorus compounds, such as hypo- phosphite (H,PO,), dimethylphosphate [(MeO)2P02], phos- phonate (HP03), monomethylphosphate (MeOP03), a- and P-glycerophosphates (a- and P-glyOP03, respectively), ortho- phosphate (PO,), pyrophosphate (P207) and triphosphate (P~O~O), each of which was dissolved in buffer solutions with various pH values. It was founds that the shape of the percentage adsorption versus pH curves (adsorption curves) that were obtained appeared to reflect.the structural charac- * To whom correspondence should be addressed. teristics of the phosphorus molecules tested. It was suggested that when the structural formula (and hence the number of P-0 moieties free of any substituents) of a given phosphorus compound was known, its adsorption curve could be predic- ted. In addition, it also seemed possible that measurement of the adsorption curve of an unknown phosphorus compound, even at a concentration as low as 1 x 10-8 mol dm-3, could allow the structural characteristics of the compound to be ascertained from the shape of the curve. In the present paper, the characteristics of the adsorption curves measured for 13 types of OPs are reported. As was anticipated, it is shown here that the adsorption of these OPs, all of which are biologically important and have a fairly sophisticated molecular structure, was primarily governed by the number of free P-0 moieties and was scarcely affected by the bulk of the substituent groups and the oxidation state of the P atoms.A procedure for releasing OPs adsorbed onto an HI0 matrix into a small-volume water phase is proposed, from which a concentration method for the analysis of a group of OPs dissolved in water at submicromolar concentrations was developed. Experimental Reagents A solution of 100 mg ml-1 of Fe in 5 mol dm-3 HCI was prepared by dissolving FeCl3.6H20 in a calculated amount of HCI and diluting with water. If the FeCI3.6H20 reagent contained more than 0.002 mol-% of PO4 as impurities, it was purified according to the procedure of Dodson et aZ.9 before use.Toluene, chloroform and pentane-2,4-dione [acetylacetone (Hacac)] were all commercially available products of guaran- teed-reagent grade, and were used after washing them with water. The standard materials used for preparing the OP stock solutions are listed in Table 1 together with the abbreviations adopted and the content of various impurities such as PO4, AMP and ADP. The level of PO, impurities, and the AMP impurities in ADP and also the ADP in ATP, were deter- mined chromatographically under the conditions given in894 ANALYST, MAY 1992, VOL. 117 Table 1 Standard materials used for preparing OP stock solutions OP standard material Abbreviation Monomethylphosphonic acid O-Phosphorylethanolamine 2-Aminoethylphosphonic acid (+)-l-Aminoethylphosphonic acid D-Fructose-l-phosphate, sodium salt ~-Fructose-6-phosphate, disodium salt a-D-Glucose-l-phosphate, disodium salt D-Fructose-l,6-diphosphate, tetrasodium salt O-Phospho-DL-threonine O-Phospho-L-serine Adenosine-5’-monophosphoric acid Adenosine-5’-diphosphate, Adenosine-5‘-triphosphate, disodium salt disodium salt MeP03 PEA 2-AEP 1-AEP F-1-P F-6-P G-1-P F-l,6-P P-Thr P-Ser AMP ADP ATP OP stock solution/mmol dm-3 10 10 10 10 10 10 10 5 10 10 10 5 3 Impurities found (mol-%) PO4, 0.14 PO4, 0.78 PO4, 0.18 PO4, 1.70 PO4, 1.18 PO4, 1.20 PO4, 0.67 PO4, 0.24 PO4, 0.031; ADP, 17 P04,2.28;AMP,7 PO4, 0.78; ADP, 2 P04,0.76 P04,0.47 Table 2 High-performance liquid chromatography conditions used for OP analysis OPs analysed* Separation column Eluent buffer and flow rate HP03, a-glyOPO3, p-glyOP03.MeOP03, MeP03, F-1-P, F-6-P, G-1-P P207, F-l,6-P AMP, ADP, ATP PEA, 1-AEP, 2-AEP, P-Ser, P-Thr TSKgel-IC Anion-SW, 5 cm x 4.6 mm i.d. Shim-pac IC-A2, lOcm x 4.6 mm i.d. Finepak SIL C18, 25 cm x 4.6 mm i.d. TSKgel IEX-215L1, 7.5 cm x 7.5 mm i.d. 1 mmol dm-3 tartaric acid solution of pH 3.35,1.5 ml min-1 0.75 mmol dm-3 potassium hydro- gen phthalate solution of pH 4.2, 1.5 ml min-1 30 mmol dm-3 KH2P04 containing 3 mmol dm-3 Bu4NBr and 8% v/v EtOH, 1 ml min-1 67 mmol dm-3 citric acid solution of pH 2.80 containing 0.2 mol dm-3 LiCl, 0.5 ml min-I Method of detectiont Ref. This COND work This COND work ABS (254 nm) 10 ABS (570 nm) 11 * Simultaneous separation of the groups of OPs was not necessarily successful, but the individual OP could always be separated from its t COND and ABS denote the measurements of conductivity and absorbance, respectively.possible moiety derived through hydrolytic decomposition and from PO4. Table 2. Stock solutions of the OPs were obtained by dissolving the corresponding standard materials in water or dilute NaOH solution and were standardized spectropho- tometrically against an orthophosphate standard solution, after being converted into orthophosphate by incineration with HC104. This was followed by correction for the PO4, AMP and ADP impurities. These solutions were freshly prepared every 2 weeks and stored in polyethylene bottles at 1 “C.The stock solutions of P207, HP03, MeOP03, and a- and p-glyOP03 and the buffer solutions of HOAc-NaOAc (pH 4.0-5.9), 2-(N-morpholino)ethanesulfonic acid-NaOH (pH 6.0-7.0), N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid-NaOH (pH 7.1-8.2) and NH3-NH4C1 (pH 8.3-10.2) were prepared as described previously.8 A 0.5 mol dm-3 acetate buffer solution (pH 6.0), 1 1 of which contained 20 g of NaOH and 29 ml of glacial acetic acid, was also prepared. Measurements of Adsorption Curves for Various OPs Unless stated otherwise, the procedures and conditions used were the same as those reported previously.8 Briefly, 30 pmol of each OP were equilibrated with a fixed amount of H I 0 ( 5 mg as Fe) in a 100 ml aliquot of 0.01 mol dm-3 buffer solution adjusted to the desired pH, and the percentage adsorption was calculated from the amount of OP found in the H I 0 and the amount used initially.The adsorption curves thus obtained are shown in Fig. 1. Apparatus A Shimadzu 6A high-performance liquid chromatograph was used for the analysis of OPs, and the conditions employed are summarized in Table 2. A Taiyo EP-1 mechanical shaker, a Kokusan H-103N centrifuge and a Shimadzu UV-200 spectrophotometer were also used. Results and Discussion Characterization and Classification of the Adsorption Curves by the Molecular Structure of OPs It was previously reported that the adsorption curves for the nine oxophosphorus compounds can be classified into four types as follows.8 (i) The first type of adsorption curve was observed with PO4 (curve 1 in Fig.1). The adsorption of PO4 was quantitative from pH 4.0 to 8.5 after which it started to decrease with a further increase of pH, so that no adsorption was observed above pH 11. From the adsorption curve, the pHs value ( i . e . , the pH at which half of the PO4 is adsorbed) could be determined as 9.5. This curve can be used as a convenient reference for the classification of the other eight OPs. (ii) Adsorption curves of the second type were observed with MeOP03, a- and p-glyOP03, and HP03 (curve 2 in Fig. 1). It can be seen from Fig..l that compounds with one less P-0 moiety than PO4, owing to substitution, commonly have an adsorption curve that recedes into the acidic region byANALYST. MAY 1992, VOL. 117 895 100 80 60 40 20 0 100 80 60 40 20 0 100 t 0 'G 80 2 a 2 60 Q) m 2 40 ? Q) a" 20 0 100 80 60 40 20 0 100 80 60 40 20 0 4 5 6 7 8 9 1 0 1 1 12 PH Fig.1 Classification of the adsorption curves of 13 types of OPs, with reference to the representative curves measured for PO4 (curve l ) , MeOP03 (curve 2), (MeO)2P02 or H2P02 (curve 3), and P207 or P3Ol0 (curve 4). (a) A, MeP03. (6) 0, PEA; 0. 2-AEP; and X, 1-AEP. ( c ) 0. F-1-P; V, F-6-P; V, G-1-P; and + F-1, 6-P. (d) +, P-Thr; and A, P-Ser. ( e ) 0, AMP; 0 . ADP; and M, ATP approximately 1.5 pH unit compared with curve 1. It is also noteworthy that the curve depended little on the nature of the substituent groups or the oxidation state of the P atoms in the compounds tested. (iii) The third type was observed with (Me0)2P02 and H2P02 (curve 3). These compounds have two less P-0 moieties than PO4 and show no appreciable tendency to be adsorbed. (iv) The fourth type of adsorption is illustrated by curve 4, and was found with P207 and P3OIO.The adsorption curve for both compounds extends towards the alkaline region by 0.5 pH unit compared with curve 1, indicating that they were adsorbed more strongly than PO4 (curve 1) against the increasing concentration of hydroxide ions which might compete for the adsorption sites on HIO. Such stronger adsorption was consistent with the larger numbers of P-0 moieties in P2O7 and P3OIO than in PO4. The adsorption curves measured for the 13 OPs are summarized in Fig. 1, and each is superimposed on the four types of typical adsorption curves (curves 1-4). Fig. l(a) shows that the adsorption curve for MeP03 accords well with curve 2.The adsorption curves for PEA, 1-AEP and 2-AEP, OPs which commonly have amine groups as the substituents, are compared in Fig. l(b). Among these OPs, the number of free P-0 moieties is the same and, as expected, the adsorption curves closely resemble each other. The fact that these three OPs have very similar adsorption behaviour indicates that neither the position of the amine residue (cf. 1-AEP and 2-AEP) nor the oxidation state of the P atoms (cf. 1-AEP and PEA) plays a significant part in the adsorption phenomenon. The slight deviations of these adsorption curves from curve 2 are attributable to the amine groups, i.e., below pH 5 the positive charge on the amine groups acts to suppress adsorp- tion onto the positively charged surface of HIO, whereas at pH 7.5-9.5 the positive charge on the amines actually facilitates adsorption onto the negatively charged HIO.On the basis of these findings, it is expected that a group of aminoalkyl phosphates and phosphonates could be collected by HI0 from solutions with pH values between 5.0 and 6.5. Sugar phosphates such as F-1-P, F-6-P, G-1-P and F-1,6-P exhibit adsorption curves similar to that shown in Fig. l(c). The curves for the first three of these OPs are in good accord with curve 2, suggesting again that the bulk and chemical nature of these sugar-substituents had little effect on the adsorption curve. In contrast, the curve for F-l,6-P appears in a more alkaline region (by 1.5 pH unit) than those for the other three compounds.This is due to a pair of PO3 groups, each of which is bonded to C1- and C6-carbons in the fructose skeleton. Although the detailed mechanisms of the adsorption of these compounds are not known, the adsorption curves of sugar monophosphates and diphosphates are thus different. Nevertheless, all these sugar phosphates can be collected on H I 0 in the pH range 4.0-6.8. By using P-Ser and P-Thr, the effect of the amino acid substituents of OPs was examined. As shown in Fig. l(d), the effects of the amino acid substituents were small and the adsorption curves for both P-Ser and P-Thr were identical with that of MeOP03, except for a slight deviation appearing above pH 7. As mentioned above [Fig. l(b)], this deviation is caused by the amine groups of the amino acid substituents.Hence, it appears that OPs possessing amino acid substituents could be collected by H I 0 from waters at pH 4.0-6.5. By using AMP, ADP and ATP as representative examples, adsorption curves for nucleotides were obtained and the results are shown in Fig. l(e). It can be seen that AMP, which is formally regarded as a monosubstituted compound of PO4, is adsorbed more favourably than MeOP03 (curve 2). Hence, the adenosine group appears to have a role in facilitating the adsorption of the AMP molecule. It can also be seen that ADP is adsorbed more favourably than AMP owing to the presence of a P206 group. Similarly, the effect of the P309 group is seen in the adsorption of ATP. Among the nucleotides studied here, ATP was adsorbed most favourably, but this was still slightly less than for P207 or P3OI0 (curve 4).The poly- phosphates and their derivatives can therefore be differen- tiated by their adsorption curves. It can be expected that various OPs belonging to this group could be collected simultaneously by HI0 at pH 4-7.896 I 1 ANALYST, MAY 1992, VOL. 117 t 1 I I 1 10-6 10-5 10-4 10-3 Amount of phosphorus/mol Fig. 2 Maximum amounts of PO4 collected by 100 (curve l), 200 (curve 2) and 300 mg of Fe in H I 0 (curve 3). For each experimental run, a series of 1 1 solutions of pH 6.0 impregnated with increasing amounts of PO4 were treated Amount of HI0 Needed for Collecting OPs From the results presented above, it can be seen that OPs with adsorption curves of type ( i ) , (ii) or (iv) could be collected from waters of pH 5.0-6.5, leaving behind OPs of type (iii).The amount of HI0 necessary to collect OPs from 1 1 of water was determined for various OP concentrations. The results are shown in Fig. 2, where the amounts of HI0 are indicated in terms of the mass of Fe. It can be seen from Fig. 2 that 100,200 and 300 mg of Fe can collect up to 50,100 and 150 pmol of PO4, respectively, at pH 6.0. However, it should be noted that after adsorbing the maximum amounts of PO4, HI0 becomes colloidal and cannot be filtered, leading to serious clogging of the membrane filters. In order to avoid this problem, it is recommended that, for instance, 100 mg (1.8 mmol) of Fe should be used to collect 30 pmol of PO4. Similar experiments were performed using HP03 and P207, and it was found that 100 mg of Fe can collect 30 pmol of HP03 and 15 pmol of P207 (Fig.3) without the formation of non-filtrable colloids. The Fe : P atomic ratio of 60, which is calculated as the ratio of the amount of Fe required to that of P in the OPs to be collected, was found to be a convenient measure of the HI0 necessary for collecting a certain amount of OP. Procedure for Concentrating OPs From Water Using HI0 (HI0 Procedure) Preparation of the H I 0 suspension A 1 ml aliquot of the 100 mg ml-1 Fe solution was taken in a 50 ml centrifuge tube and mixed with 20 ml of 1 mol dm-3 NH3 solution. The HI0 thus formed was washed twice with 20 ml of 0.5 mol dm-3 acetate buffer (pH 6.0), centrifuged for further washing and then dispersed into 20 ml of pure water by vigorous shaking.Adsorption of OPs With magnetic stirring, the HI0 suspension was added to a 1 1 water sample, the pH of which had been brought to approximately 6.0 by the addition of 20 ml of 0.5 mol dm-3 acetate buffer. After allowing to stand for 2 h, the H I 0 was collected by filtration and washed with pure water. It was then air-dried to produce flakes. The still wet HI0 flakes were transferred into a glass-stoppered 10 ml centrifuge tube using a spatula. Release of OPs from H I 0 To the 10 ml centrifuge tube containing the HI0 flakes were added 2 ml of Hacac. The mixture was shaken for 1 h at 20-30°C; the HI0 flakes turned to a reddish emulsion and finally to blood-red Fe(acac)3 which was extracted with 3 ml of toluene. After discarding the toluene extract, the remaining 10-6 10-5 10-4 10-3 Amount of phosphorus/mol Fig.3 of Fe. Other conditions as in Fig. 2 Collection of HP03 (curve 1) and P207 (curve 2) using 100 mg ADP 3 ADP 1 ' 0 10 20 0 10 20 Time/m in Fig. 4 Recovery of ADP with the H I 0 procedure. A 2.5 ml aliquot of a 0.479 mmol dm-3 ADP standard solution was diluted to 1 1 and returned to a 2.5 ml sample by the H I 0 procedure. (a) Before and ( b ) after the H I 0 procedure; 20 p1 aliquots of the ADP solution were chromatographed under the conditions given in Table 2 lower phase was mixed with a further 2 ml of Hacac and shaken for 12 h at 20-30 "C, followed by washing twice with 2 ml of toluene and once with 2 ml of chloroform. Washing with chloroform also facilitated the changeover of the aqueous phase from the bottom layer to the top layer, into which the OPs were released.The approximately 1.2 ml of aqueous phase thus obtained were transferred into a vial and made up to 2.5 ml by adding the washings of the centrifuge tube. When polyphosphates or their analogues were involved, the final aqueous phase was slightly coloured owing to the inclusion of trace amounts of Fe, otherwise it was colourless. Recovery of OPs With the H I 0 Procedure Phosphonates such as MeP03, 1-AEP and 2-AEP do not undergo hydrolytic decomposition, whereas the other OPs have a greater or lesser tendency to undergo degradation and liberate PO4. By comparing the contents of OPs before and after the H I 0 procedure, the percentage recovery of each OPANALYST, MAY 1992, VOL. 117 897 Table 3 Recovery of OPs and some inorganic oxophosphorus compounds with the H I 0 procedure Compound* Ck,-glyOPO3 (J-glyOPO, PEA HP03 P-Ser P-Thr 1-AEP 2-AEP MePo, MeOP03 Recovery (YO) (min-max)t 73.8-79.4 78.7-79.5 85.1-88.1 87.9-96.6 85.8-92.3 84.4-92 .O 87 4-90.4 80.3-92.2 84.3-92.8 64.3-77.9 Recovery after storage$ (YO) 74.6 85.2 88.5 78.2 88.3 - - - - 62.6 Compound* F-1-P F-6-P G-1-P G-6-P PO4 p207 AMP ADP ATP F-l,6-P Recovery (%) Recovery after (min-max) -t storage$ (%) 82.2-84.0 81.1 73 .O-77.8 - 75.7-82.8 - 87.0-90.8 90.0-96.8 92.5 60.0-61.9 - 83.7-90.7 - 67-77 - 92-96 - 45-60 - - * OPs (1-3 pmol) were treated according to the H I 0 procedure.t The minimum and maximum values of triplicate experiments are indicated. $ The H I 0 with adsorbed OPs was stored for 2 weeks at 1 "C and then treated with Hacac to release the OPs (for details see text).was evaluated. The OP concentrations were determined chromatographically under the conditions shown in Table 2. The typical chromatograms recorded for the ADP standard solution before (a) and after (6) the HI0 procedure are compared in Fig. 4. The peaks appearing at 8.4 and 12.0 min are due to AMP and ADP, respectively. It can be seen that ADP is recovered with an efficiency of 72%. The efficiencies thus evaluated for the individual OPs are summarized in Table 3. The OPs examined here were recovered with efficiencies ranging from 45% (for ATP) to 92% (for 2-AEP). Because of their resistance to hydrolytic degradation, relatively low efficiencies were observed for ATP and ADP.It can also be seen from Table 3 that the recoveries evaluated with and without storage of OPs on HI0 flakes under cold conditions (1 "C) showed good agreement within experimental error. The method therefore appears to offer a safe and convenient procedure for the compact storage of OPs adsorbed onto HI0 rather than as bulky solutions. At each of three different concentrations, viz., 1 x 10-7, 5 x 10-8 and 1 x 10-8 mol dm-3, five (1 I) standard solutions of 1-AEP, which was taken as a representative OP, were treated by the HI0 procedure and then analysed by high- performance liquid chromatography. The mean and standard deviation of the recoveries were found to be 89.3 k 2.8,91.5 k 6.4 and 86 k 18% at the respective concentrations. From these values the precision (relative standard deviation) of the proposed method was calculated to be 3.1 and 7.0% at concentrations of 1 x 10-7 and 5 x 10-8 mol dm-3, respectively, and the limit of detection (signal-to-noise ratio = 5 ) to be 1 x 10-8 mol dm-3. The authors thank Dr. F. Yamamoto of Kyoto University and Dr. T. Okada of Shizuoka University for their helpful suggestions in optimizing the conditions for high-performance liquid chomatography . This research was carried out under financial support (No. 01540474) from the Ministry of Education, Science and Culture, Japan. References 1 2 3 4 5 6 7 8 9 10 11 Fiske, C. H., and Subbarow, Y., J. Biol. Chem., 1925,66,375. Pontis, H. G., and Leloir, L. F., in Analytical Chemistry of Phosphorus Compounds, ed. Halmann, H., Wiley-Interscience, New York. 1971, pp. 617458. Fiske, C. H., and Subbarow, Y., J. Biol. Chem., 1929,81,629. Ishibashi, M., and Tabushi, M., Bunseki Kagaku, 1959,8,588. Fujinaga, T., and Hori, T., Environmental Chemistry on Lake Biwa, Japan Society for the Promotion of Sciences, Tokyo, 1982. Murphy, J., and Riley, J. P., Anal. Chim. Acta, 1962, 27, 31. Ishibashi, M., and Tabushi, M., Bunseki Kagaku, 1957, 6, 7. Hori, T., Moriguchi, M., Sasaki, M., Kitagawa, S., and Munakata, M., Anal. Chim. Acta, 1985, 173,299. Dodson, R. W., Forney, G. J., and Swift, E. H., J. Am. Chem. Soc., 1936,58, 2573. Pennings, E. J. M., and van Kemper, G. M. J., J. Chromatogr., 1979, 176,478. Hori, T., and Kihara, S., Fresenius' 2. Anal. Chem., 1988,330, 627. Paper I i04648I Received September 6, 1991 Accepted December 6, I991

ISSN:0003-2654    

DOI:10.1039/AN9921700893

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST. MAY 1992, VOL. 117 899 Denuder Tube Preconcentration and Detection of Gaseous Ammonia Using a Coated Quartz Piezoelectric Crystal Zulfiqur Ali, C. L. Paul Thomas and John F. Alder* Department of Instrumentation and Analytical Science, UMIST, P.O. Box 88, Manchester M60 IQD, UK Geoffrey 6. Marshallt National Power Technology and Environmental Centre, Kelvin Avenue, Leatherhead, Surrey KT22 7SE, UK The feasibility of using a cylindrical denuder tube for sampling gaseous ammonia, followed by detection with a piezoelectric quartz crystal, was investigated. Gaseous ammonia was sampled with a tungsten oxide-coated cylindrical denuder tube and then thermally desorbed onto a piezoelectric quartz crystal coated with pyridoxine hydrochloride-Antarox CO-880. A linear calibration graph of peak area response versus ammonia concentration sampled was obtained for ammonia concentrations between 3.1 and 8.2 pg 1-1.A concentration of 29 ng 1-1 of ammonia in air was detected with a signal-to-background ratio of 14: 1 by achieving an enrichment ratio of 900 with the tungsten oxide denuder tube. Keywords: Ammonia; piezoelectric crystal; denuder tube; preconcentration Ammonia is one of the most important trace gases in the atmosphere and is the only one that is basic. It i s water-soluble and can react with aerosols, thus influencing atmospheric acidity. Most ammonia emissions are released into the atmosphere by biological processes, primarily through the decomposition of organic matter. 1 The main industrial source is from fertilizer and ammonia production plants.’ In determining the precise role of ammonia in the atmo- sphere, it is important to distinguish free ammonia from ammonium particulates. Filtration techniques have been used for separation of the gaseous phase from particulates, although their use can cause errors by the introduction of artefacts. For instance, an over-estimation of the ammonia concentration can be obtained by release of ammonia from ammonium nitrate on the filter-paper.Equally, an under- estimation can be obtained by reaction of gaseous ammonia with acids deposited on the filter. Denuder tubes have been shown to be effective for the separation of gases from particles, and their theory and application for the determination of gaseous species have been reviewed.3-4 Air is drawn under conditions of laminar flow through a tube coated with a selective adsorbent. Gaseous species diffuse to the collection surfaces.Particulates, having much lower diffusion velocities, cannot migrate to the walls and hence pass through unabsorbed and do not contribute to the final measurement. Gormley and Kennedy5 have derived a solution describing diffusion from a stream flowing through a cylindrical tube: - = 0.819 exp (14.6272A) + 0.0976 exp (-82.22A) C( ) (1) where c is the mean concentration of gas leaving the tube, and co is the gas concentration entering the tube. + 0.01896 exp (-2 12A) nD1 4F A = - where D is the diffusion coefficient of the target gas in air, f is the length of coated tube, and F is the flow rate. For A 3 0.05 only the first term in eq.( I ) is significant. c = 0.819 co exp (-14.62726) ( 3 ) * To whom corrcspondcnce should be addressed. -t Present addrcss: Chcmistry Departmcnt, Birkbcck Collcgc, 20 Gordon Square. London WClH OPP, UK. The gas flow is, therefore, depleted by diffusion of the gaseous component to the walls, where it is adsorbed. The particulate phase, being heavier, passes through the tube. An oxalic acid-coated denuder tube has been used for the determination of free ammonia in ambient air.h.7 The method described requires washing the sorbed ammonia from the tube and determining it by potentiometry with an ammonia-sensi- tive probe. A tungsten oxide-coated denuder tube has also been used for sampling ammonia, which was then thermally desorbed and determined by means of a chemiluminescence nitrogen detector.8 Hlavay and Guilbault’ have used a pyridoxine hydrochloride-coated piezoelectric quartz crystal for the detection of ammonia.Pyridoxine hydrochloride reversibly adds ammonia to the phenolic 3-OH: NHj + OH /OH HCI- HCI -- A reported frequency change of 1190 Hz for 1 ppm of ammonia in air was obtained, while Moody et af. extended the useful lifetime of the detector by supporting the pyridoxine hydrochloride in a nonyl phenoxypolyethoxylate matrix of high relative molecular mass (Antarox CO-880). This was, however, at the expense of decreased sensitivity. Antarox CO-880 did not itself adsorb ammonia. The present paper describes the application of a tungsten oxide-coated denuder tube to the collection of gaseous ammonia, which is subsequently thermally desorbed and detected by a pyridoxine hydrochloride-Antarox CO-880 coated, 14.9 MHz piezoelectric quartz crystal.The frequency responses of the piezoelectric quartz crystal detector to various concentrations of ammonia were measured. Experimental Piezoelectric Quartz Crystal Oscillator The piezoelectric quartz crystals used were 14.9 MHz, fundamental mode, AT-cut with 3.8 mm diameter circular gold electrodes, on both sides of the 8 mm diameter quartz slab, supplied by Cathodeon Crystals, Cambridge, UK. The reference and detector crystals were made part of two pre-fabricated crystal oscillator circuits (Cathodeon Crystal)900 1 I. ANALYST, MAY 1992, VOL. 117 powered from a 5 V d.c. power supply. The oscillator outputs were mixed in the exclusive-OR gates of a 74LS86 operational amplifier (RS Components, Corby , Northamptonshire, UK), and a low-pass filter was used to select the difference frequency, which was passed to a 120 MHz range, 0.01 Hz resolution digital frequency counter (Philips PM6671; Philips, Cambridge, UK), equipped with a digital-to-analogue con- verter.The sensor crystal was housed in a double-impinger type cell with a swept volume of 0.7 ml, sample gas being injected perpendicular to each face of the coated crystal. Ammonia Determination A continuous-gas flow regime was used for the determination of ammonia. Cylinder air (dried over silica gel) was used as the carrier gas. Poly(tetrafluoroethy1ene) (PTFE) tubing (6.5 mm o.d., Omnifit, Cambridge, UK) was used for all supply and waste lines.Gas flows were controlled by needle valves and were measured by rotameters calibrated against a soap-film bubble meter. Low concentrations of ammonia in air were produced by two-stage dilution of a 100 ppm ammonia-in-air standard mixture (BOC Special Gases, London, UK). A four-way valve (Omnifit) was used to switch from the reference air flow to the sampling flow. A flow rate of 20 ml min-1 was maintained in the piezoelectric quartz crystal detector cell by means of a needle valve. Pyridoxine hydrochloride-Antarox CO-880 coated pie- zoelectric quartz crystals were prepared by brush-coating a 1 + 1 (vh) mixture of 0.02% m/v pyridoxine hydrochloride in 50% v/v aqueous ethanol and a 0.2% m/v solution of Antarox CO-880 in acetone. The coated quartz crystals were then kept in an oven at 80 "C for 2 h and allowed to cool in a desiccator.Preparation of Tungsten Oxide Denuder Tube Quartz tubes were coated with tungsten oxide by using a modification of the procedure described by Braman et al.8 The quartz tubing (3 mm i.d., 5 mm o.d., 35 cm length) was prepared for coating by first washing with benzene and then with 50% sodium hydroxide solution. The tubes were then treated with 40% m/v hydrofluoric acid before being finally rinsed with high-purity water and dried in an oven at 150°C for 2 h. Flow rates of less than 1 1 min-1 were used for much of the gas sampling work. A 10 cm subduction zone was made by pipetting a solution of trimethylsilane in dichloromethane to the 10 cm mark, draining and then washing with high-purity water and drying at 150°C for 2 h.The denuder tubes were coated by means of the apparatus shown in Fig. 1. The quartz tube to be coated was connected between two Swagelok T unions with 0.25-0.125 in reducing PTFE ferrules. Tungsten wire (0.5 mm 0.d.; Goodfellow Metals, Cambridge, UK) was spot-welded to stainless-steel rods (0.125 in 0.d.) at each end. Several welds were necessary to achieve a mechanically strong join between the wire and rod. The stainless-steel rods and tungsten wire were threaded through the denuder tube and the T union compression fittings. A gas-tight seal was formed by using the PTFE reducing ferrule (Phase Separations, Deeside Industrial Estate, Clwyd, UK). The denuder tube was evacuated to 267-400 Pa with a rotary vacuum pump.The tungsten wire was heated to approxi- mately 1000 "C by passing through it a current (a.c.) of 12 A. Heating the tungsten wire caused expansion, and the tension was maintained by carefully pulling the stainless-steel rods with insulated pliers. The current was controlled by a variable transformer to provide a slow coating rate. A light coating of the blue oxide was obtained after passage of 12 A for 30 min; a heavier coating was obtained after 2 h. Denuder tubes coated rapidly at higher currents resulted in mechanically less stable coatings. The blue tungsten(1v) oxide was obtained by heating the tube to 350 "C at a pressure of approximately 650 Pa. The denuder tube was heated with a Kanthal resistance wire heater (Scientific Wire Co., London, UK). The resistance wire was wrapped around the quartz tube and held in place with hose clips at both ends.The temperature of the denuder tube was monitored by placing in the tube a thermocouple connected to two strands of Kanthal wire, with output to a digital thermometer. Chemiluminescence Nitrogen Detector A chemiluminescence nitrogen detector (Monitor Labora- tories, Model 8440E) was used to characterize the tungsten oxide denuder tubes. Operation of the detector was based on the chemiluminescence reaction between nitrogen oxide and ozone. The nitrogen oxide and ozone reaction yielded stable nitrogen dioxide, primarily in the electronic ground state. A small fraction of the reaction yielded excited nitrogen dioxide, which emitted an infraredhisible continuum with a maximum intensity at approximately 1.1 pm.11 Determination of ammo- nia was achieved by chemically converting the sample into nitrogen oxide in a stainless-steel converter at 850 "C, prior to its entry into the reaction cell.The instrument was calibrated against standard nitrogen oxide mixtures (BOC Special Gases) and calibrated for ammonia by passing standard ammonia-air mixture into the converter. Characterization of the Tungsten Oxide Denuder Tube The collection efficiency of the tungsten oxide denuder tube was established by measuring the decrease in the steady-state signal, when the denuder tube was inserted between the standard ammonia sample stream and the chemiluminescence nitrogen detector. Desorption of the adsorbed ammonia was carried out by heating the denuder tube from ambient temperature to 350"C, with a desorption time of 1 min.The recovery efficiency was measured by dividing the area under I , Stainless-steel contact Tungsten wire Fig. 1 gauge: and 3, rotary vacuum pump Apparatus used for coating the denuder tubes with tungsten oxide. For details, see text. 1, A.c. voltage controller: 2, Pirani vacuumANALYST, MAY 1992, VOL. 117 90 1 V1 Vent Fig. 2 Apparatus for the preconcentration of ammonia on the denuder tube, and subsequent desorption onto the piezoelectric crystal 1, Digital thermometer; 2, a.c. voltage controller; 3, piezoelectric crystal detector cell; 4, oscillator; 5, frequency counter; recorder; 7, power supply; B, ammonia, 100 ppm in air; 9, air; and 10, drying tube 1 U the thermal desorption peak by the area under the collection profile.This ratio corresponds to the mass of ammonia desorbed from the denuder tube divided by the mass sorbed from the air stream. The linearity of response of the denuder system for ammonia detection was measured by introducing known mixtures of increasing ammonia concentration into the tungsten oxide denuder tube. Ammonia standards were prepared by dilution of standard 100 pprn ammonia-in-air mixtures. The adsorption capacity of the denuder was ascertained by subjecting it to an ammonia concentration of 60 pg m-3 in air at a flow rate of 100 ml min-1, noting the time taken for breakthrough to the NO, detector. Sampling and Detection of Gaseous Ammonia by Denuder Tubes and Piezoelectric Quartz Crystals A schematic diagram of the system used to sample and detect gaseous ammonia is shown in Fig.2. Low ammonia concentra- tions were generated from permeation tubes, prepared in the laboratory, having typical permeation rates of 180 ng min-1 at a temperature of 28°C. Standard ammonia gas streams were routed to the tungsten oxide denuder tube through a four-way valve V1. The collected ammonia was thermally desorbed into a dry air stream. Detection of ammonia was carried out by an AT-cut 14.9 MHz piezoelectric crystal coated with pyridoxine hydrochloride-Antarox CO-880. Results and Discussion The effect of temperature on the preparation of the denuder tube, and the desorption temperature, were examined. Tungsten oxide denuder tubes subjected to a desorption temperature of 500 "C showed degradation in response with successive sampling cycles, using chemiluminescence nitrogen detection.Desorption products included nitrogen oxide, nitrogen dioxide and ammonia. Tungsten oxide denuder tubes subjected to temperatures of less than 350 "C gave ammonia as the only desorption product, and 11 successive sampling and analysis cycles of a 3 ppm ammonia standard gas stream resulted in peak heights with a relative standard deviation of 3%. The collection and recovery efficiencies were found to be 97 and 96%, respectively. The linearity of quantitative adsorption of ammonia by the tungsten oxide denuder tube was demonstrated over ammonia mass loadings of 0.2-3.3 pg, with a correlation coefficient of 0.998. These results show that the tungsten oxide denuder tube is capable of reproducible and quantitative sampling of ammonia.Ammonia thermally desorbs at a temperature near to that where degradation of the denuder occurs. The reliability of the technique is, therefore, dependent on accurate temperature control. 100 5 80 5 60 40 20 N m (II 3 0 detector. 6, chart 0 14 28 42 56 Ammonia concentration in air/pg I-' Fig. 3 Calibration graph for the coated piezoelectric crystal detector over the range 0.743 yg I-' (1-90 ppm) of ammonia. The inset shows the calibration in the range 0-0.7 pg I-1 (0-1 ppm) of ammonia obtained at a different time Fig. 3 shows the response curve for the direct exposure of the piezoelectric crystal to increasing concentrations of ammonia in dry air. The sensitivity obtained is lower than that reported by Hlavay and Guilbault.9 A linear change in frequency with sorbed mass is expected for low mass loadings on the piezoelectric crystal,l2-15 hence the response curve will represent the ammonia sorption isotherm.The curve is similar to a Langmuir isotherm, and this is a reasonable model for a surface such as pyridoxine hydrochloride with clearly identifi- able active sites. Thermal desorption cycles of the tungsten oxide denuder tube, without added ammonia, resulted in frequency de- creases of the sensor piezoelectric quartz crystal. The fre- quency decreases were thought to be as a result of an elevation in the temperature of the piezoelectric quartz crystal and expansion of the desorption gas volume. The effects could be decreased by increasing the desorption gas flow rate.A plot of this blank peak area versus desorption gas flow rate (Fig. 4) showed an exponential decrease of the peak with increasing desorption volume flow. A compromise high-desorption flow will, however, result in increased dilution of the desorbed analyte. A desorption flow rate of 100 ml min-1, which minimized the peak, but maintained practical sensitivity, was used for all further studies. The linearity of response of the piezoelectric quartz crystal denuder system for ammonia sampling and detection was tested by exposing the tungsten oxide denuder to ammonia concentrations between 3.1 and 8.2 pg 1-1. A linear plot of902 46 42 N Eu g 34 : 3 38 2 Y m .- 30 E 26 e g 22 n 18 16 N 14 Eu $ 12 C 0 $ 10 2 2 8 z 6 m Y Q Y - 6 4 m 2 - - - - - - - - ANALYST, MAY 1992, VOL.117 0 100 200 300 Volume flow rate/ml min- Fig. 4 versus air flow rate through the denuder tube Plot of coated piezoelectric crystal blank response peak area 0 300 600 900 Enrichment ratio Fig. 6 Plots of peak height (0) and peak area (0) for the coated piezoelectric crystal response versus enrichment ratio calculated for the tungsten oxide-coated denuder tube I I I I I 2 4 6 8 1 0 Ammonia concentration in air/pg 1-1 Fig. 5 Plot of ammonia response peak area from the coated piezoelectric crystal versus concentration of ammonia sampled in the denuder tube. The correlation coefficient of the straight line is 0.97; the ordinate axis intercept is 7.1 peak area response versus concentration of ammonia is shown in Fig. 5. The linearity of response of the piezoelectric quartz crystal/denuder system was further tested by varying the exposure times of a 6.1 pg 1-1 ammonia gas stream to a tungsten oxide denuder tube.Exposure times of between 30 and 210 s resulted in 0.3-2.1 pg calculated ammonia mass loadings. A least-squares fit analysis of the peak area response versus calculated mass loading yielded a straight line between 0.6 and 2.1 pg, with a slope of 25 cm2 pg-1 and with a y-axis intercept of -2.0 cm2 pg-1. The correlation coefficient for the straight line was 0.99. Below 1 pg of ammonia, the line curved towards the origin. The adsorption capacity was calculated to be 2.9 pg of ammonia, which compares favourably with a value of 2.5 pg in earlier work16, when using the same method, and with 3 pg from breakthrough studies.Ammonia thermal-desorption profiles show the initial blank response followed by the ammonia peak. The tailing of the analyte peak increased with the mass of ammonia exposed. Such tailing could be due to a number of factors. The collection surface of the tungsten oxide denuder tube is not homogeneous, and the adsorption sites will have a range of binding energies, resulting in tailing of the desorption peak. Enrichment ratio Fig. 7 Plot of frequency response versus enrichment ratio 4000 0 300 600 900 Enrichment ratio Fig. 8 Plot of integrated peak area versus enrichment ratio Diffusional broadening and adsorptive retardation by the sampling lines and fittings could also lead to tailing. Preconcentration of the Analyte The denuder tube, in addition to separating gaseous ammonia from particulates, also serves to preconcentrate the analyte and hence increase the sensitivity of the system.The extent ofANALYST. MAY 1992. VOL. 117 903 preconcentration is determined by the enrichment ratio (ER), defined as the ratio of the concentration of ammonia in the desorbed gas (cD) to that in the ambient air [co; see eqn. (3)]. If the denuder tube preconcentrator is exposed until it is saturated, a significant part of the sample can be lost, but the ER is optimized. If the sampling time is restricted such that breakthrough does not occur, higher sampling efficiency, but lower enrichment ratios, are observed. A study was carried out to illustrate the improvement in sensitivity of the system with an increase in the enrichment ratio.An ammonia-in-air concentration of 29 ng I-' was passed over the tungsten oxide denuder tube for times sufficient to yield enrichment ratios of 300, 600 and 900. The data obtained are shown in Figs. 6-8 as plots of peak frequency response and integrated area response for the blanks and three enrichment ratios. It is clear that this procedure significantly enhances the over-all sensitivity of the determination. Conclusions The feasibility of using a tungsten oxide denuder tube preconcentrator with a pyridoxine hydrochloride-Antarox CO-880 coated piezoelectric quartz crystal for the sampling and detection of ammonia has been demonstrated. Analysis for low concentrations of ammonia can be carried out if high enrichment ratios are used. The enrichment ratios could be further increased by increasing the adsorption capacity of the denuder tubes by using an annular denuder, which would also permit much shorter sampling times.The use of denuder preconcentrators thus leads to the significant advantage of being able to desorb the sampled vapour in dry carrier gas, hence overcoming the effect of relative humidity on the piezoelectric crystal. Although the temperature pulse has an effect on the piezoelectric crystal, it can be minimized and useful sensitivity can be achieved. The data in Figs. 6-8 represent a signal-to-background ratio of about 14 : 1 in integrated area terms, with an enrichment ratio of 900 for an ammonia concentration of 29 ng 1-1, which represents a limit of detection of the order of 6.5 ng 1-1 of ammonia. The effect of interfering species on this technique has not been studied in this work.Previously published data indicate that alkylamines, nitrogen dioxide, peroxyacetyl nitrate and nitric acid adsorb into denuder tubes with tungsten(v1) oxide coatings.3 However, none of these species in the gas phase has been reported as adsorbing onto coatings of pyridoxine hydrochloride, and hence the second-stage selectivity of the piezoelectric crystal detector should minimize their action. It is also expected that aerosol and particulate interference will be minimal. The authors are grateful to SERC and the then CEGB for support of Z . A. under the CASE scheme. The contribution of G. B. M. is published by permission of National Power plc. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 References Bujisman, E., Moss. H . F . , and Asman. W. A. H. Afmos. Environ., 1987, 21, 1009. Dawson, G . A.. J. Geophys. Res., 1977, 82, 3125. Ali, Z., Thomas, C. L. P., and Alder, J. F., Analyst, 1989, 114, 759. Murphy, D. M., and Faheg, D. W., Anal. Cliern.. 1987, 59, 2753. Gormley, P. G., and Kennedy, M., Proc. R. tr. Acacf., Sect. A . , 1949,52A, 103. Ferm, M.. Amos. Environ., 1979, 13, 1385. Dimmock. N. A., and Marshall, G. B., Anal. Chirn. A d a , 1986, 185, 159. Braman, R. S . , Shelley, T. J., and McClenny. W. A., Anal. Chem.. 1982. 54. 365. Hlavay, J.. and Guilbault, G . G., Anal. Chern.. 1978.50, 1044. Moody, G. J., Thomas, J . D. R.. and Yarmo, M. A., Anal. Chirn. Acta, 1983, 155, 225. Clyne, M. A. A.. Thrush, B. A., and Wayne, R. P . , Discuss. Faraday Soc.. 1964, 60, 359. Sauerbrey, G. Z . , Z . Phys. Chern. (Leipzig), 1959, 155, 206. Miller, J . G., and Bolef, D. I . , J . Appl. Phys., 1968, 139. 5815. Lu, C. S . , J . Vuc. Sci. Technol., 3975, 12, 578. Cumpson, P. J., and Seah. M. P., Meas. Sci. Technol.. 1990, 1. 544. Ali, Z., PhD Thesis, UMIST, University of Manchester, 1989. Paper 0l05024E Received May 23, 1991 Accepted November 5, 1991

ISSN:0003-2654    

DOI:10.1039/AN9921700899

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, MAY 1992, VOL. 117 905 Automated Gravimetric Management of Solutions Part 1 High-performance Microcomputer-controlled Gravimetric Burette lldenise B. S. Cunha and Celio Pasquini* lnstituto de Quimica, Universidade Estadual de Campinas, Caixa Postal 6154, CEP 13081, Campinas, Sao Paulo, Brazil A versatile gravimetric burette and the necessary interface that allows it t o be controlled by an IBM-PC microcomputer are described. The burette employs an electronic balance that holds three 30 ml flasks. The flasks are used for delivering different titrants or standard solutions and are connected t o the sensor through the bottom of the balance. The addition of the solution is controlled by poly(tetrafluoroethy1ene) electromechanical valves housed inside the unit. The flasks can be refilled automatically from larger reservoirs placed outside the case.Solution level sensors are used t o realize automatic refill when necessary. The mass delivered from the flasks is read by the computer through an RS232C interface. The burette can, when driven by the appropriate software, perform potentiometric, biamperometric and spectrophotometric titrations, standard additions procedures and the preparation of standard solutions. Keywords : G ra virne tric burette; titra ti0 n; standard additions; a u torn a ted titra tion The gravimetric addition of a solution in titration procedures has been described in the literature. 1-5 The advantages of such an approach to titrant addition are always emphasized: for example, it is claimed that gravimetric additions do not require any calibration of the glassware, and are free from errors caused by solution viscosity and by contraction/dilata- tion of the volume caused by a change in the ambient temperature.Moreover, additional advantages are provided by the use of modern electronic balances: for example, rapid weighing, absence of moving parts that can be affected by mechanical stress, and the balance frequently has a standard RS232 interface providing prompt communication with micro- computers. In addition to these advantages the potential of the gravimetric approach has not yet been fully exploited. In fact, the gravimetric manipulation of solutions in the laboratory could lead to more precise and reliable results. On the other hand, there is a need for a versatile, cost effective, gravimetric apparatus to deal with the variety of analytical procedures that demand solution management.Some of the instruments described previously do not guard against air currents which can affect the mass measurement. Therefore, high sensitivity balances cannot be used.1.3.4 The instruments described also employ a large reservoir (and sometimes a control valve) placed on the balance plate. As the load on the balance is high, the usable sensitivity is only 0.01 g in those less expensive balances which can handle, for example, only 100 g with a sensitivity of 1 mg. Previously described instruments do not allow the simul- taneous use of more than one solution in the same gravimetric addition unit. The use of more than one solution with only one balance would improve the cost to benefit ratio of constructing the unit and at the same time would extend its ability to perform, for example, a back-titration, and simple or multiple standard additions procedures.This paper describes a versatile microcomputer-controlled gravimetric burette that extends the application of such units and allows for the use of more than one solution per balance which can operate with a higher sensitivity. * To whom correspondence should be addressed. Experimental Gravimetric Burette The gravimetric burette is depicted in Fig. 1. The unit was constructed using a metallic frame 50 cm high, 25 cm wide and 25 cm deep. Three sides of the metallic frame were closed with metallic sheets and the front was closed with an acrylic sheet that can be lifted if access to the flasks inside the unit is desired.The metallic frame has a platform on which an Acatec Model BCM 1003 electronic balance with capacitive sensor was placed. The balance can operate with a sensitivity of 1 mg (maximum load 100 g) or 10 mg (maximum load 1000 g). At the top of the case there is a small metallic door through which it is possible to reach the balance plate. A metallic rod was connected to the sensor through a hole in the base of the balance case. At the end of the rod, an acrylic plate sustains three 30 ml polyethylene flasks held on the plate by inner plastic screws. At the base of the flasks a hole of 0.1 mm diameter was drilled for admission of air during solution delivery. The outlet of the flasks contains a 2.0cm long, 1.6 mm o.d., 1.2 mm i.d., glass tube, to which is attached a 1.5 mm i.d.Tygon tube. The length of the Tygon tube is twice the distance between the glass tube and the three-way valves (NResearch Model 161T031) to which the tubes are connec- ted to the common port. The normally closed outlets of the valves are connected to poly(tetrafluoroethy1ene) (PTFE) tubes (20 cm X 0.9 mm i.d.). These tubes are taken out from the metallic case through small holes drilled in the front acrylic sheet and are used for solution delivery. The normally open outlets of the valves are connected to two-way valves (NResearch Model 161TOll) placed outside the case. These valves are connected to 11 solution reservoirs placed on the outside of the case. The manifold used to couple the solution reservoir to the flasks inside the burette case is shown in Fig.2. When any of the two-way valves is turned on the stock solution is delivered to the respective burette flask. When any of the three-way valves is turned on the solution present in the respective flask is delivered outside the case, through the PTFE tube. The mass, delivered during the period in which the valve is kept on, is monitored by the computer. Around the small glass tube fitted at the outlet of each inner flask there is an infrared opto-switch (PCST-2103). The switches are used to detect when a flask is empty, requiring immediate action by the computer to turn on the respective two-way valve to refill the flask. The refill operation is906 n ANALYST. MAY 1992, VOL.117 G Fig. 1 Proposed gravimetric burette. A, External solution reservoir flasks; B, electronic balance; C, two-way electromechanical valves; D, internal flasks; E, three-way electromechanical valves; F, solution outlet; and G, metallic case performed by switching the respective two-way valve on and by following the increase in the mass of the whole set of inner flasks. For the flasks used, a maximum mass increase of 20 g is allowed before the valve is switched off, stopping the operation. Therefore, no sensor for the ‘full’ status is required. The time spent on this operation can be reduced if a nitrogen pressure is applied to the external stock solution reservoir. Interface and Electronic Circuits The communication of the balance with the microcomputer (DICOM, IBM-PC XT compatible, 640 kbyte RAM, Win- chester of 20 Mbyte and Floppy of 360 kbyte, Display CGA-monochrome) is made at 9600 bits s-1 using the serial RS232 interface present in both the balance and computer. The balance sends a string containing the mass reading every 0.3 s.The control of the gravimetric unit is effected by using the asynchronous interface described previously .6 The interface communicates with the microcomputer through a user port based on the CI 8255.7 The card, containing the user port, is plugged into an extension slot inside the microcomputer. Two handshake signals (strobe and acknowledge), the ground line and the eight parallel data lines are exchanged between the computer user port and the interface. A circuit diagram of the interface employed for control of the burette and for data acquisition is shown in Fig.3. The CI 74LS373 (I) is used as an address decoder and up to eight electronic devices can be accessed.6 Two of the addresses (254 and 253 corresponding to bit 0 and 1) are used for a dynamic B t Fig. 2 Manifold to connect the external reservoir flasks to the internal burette flasks. A, External reservoir flask; B, balance support; C, burette tlask; D, opto-switch; E, Tygon tube; F, burette outlet; V1. two-way valve; and V2, three-way valve analogue-to-digital converter based on an 8 bit DA (ZN428) and an 8 bit AD (ZN448).8 The address line 251 (bit 2) is connected to another 74LS373 (111) used as an 8 bit latch. Six of the output lines are used to source the base current to 2222N transistors, used as switches for the electromechanical valves.Address line 243 (bit 3) is used to enable another CI 74LS373 (II), used to input the logic state of the three level sensors. A circuit diagram of the optical sensors is shown in Fig. 4. The average change in the voltage at the collector of the opto-transistor is 2.0 V (from 6.0 to 4.0 V). This change is caused mainly by the difference in the refractive index between the full and the empty glass tube. Therefore, approximately the same change is observed for coloured solutions such as 0.01 mol dm-3 KMn04 and 0.1 mol dm-3 K2Cr207. The reference voltage for the 741 comparators is set to the middle of the range, about 5.0 V. When the level of liquid in the flask is above the limit the TTL logic level at the comparator output is high.The existence of a low level signs the ‘flask empty’ condition. Analogue signals to be monitored are passed through an analogue input pre-conditioning stage. This stage (which is not shown in Fig. 3) is based on two OP07 operational amplifiers and is designed to provide a high versatility in terms of the dynamic range and polarity of the signal to be followed. The first OP07 can supply a positive or negative gain to the signal. Its output is sent to the next stage which will sum a positive or negative offset voltage9 to change a bipolar signal to a positive only signal which can be followed by the dynamic analogue-to- digital converter, shown in Fig. 3. Main Software to Control the Gravimetric Burette General use software to drive the user card and the interface, and for reading the dynamic analogue-to-digital converter, has been described previously.7.x In addition to this software, a number of QUICKBASIC 4.5 sub-programs were written specifically to control the gravimetric unit.These sub-pro- grams are shown in the Appendix. Five sub-programs are necessary to control the gravimetric burette. The SUB valveon(nv%) and SUB valveoff(nv%) are used to open or close. respectively, any of the six valves present in the gravimetric unit; nv% is the valve number (1-6). The sub-program SUB fillburette(n% ,ma) is used to fillANALYST, MAY 1992, VOL. 117 907 Fig. 3 Electronic circuit diagram of the interface employed for control of the burette and data acquisition 12v T 0 ’7 2 M 1 K 4= u &- 20K1 12 v Glass tube Electronic circuit diagram for solution level detection Fig.4 the burette flask n% (1-3). The flask n% will be empty to the optical sensor level and then refilled until a mass change equal to ma is observed. If the flask is already empty the refill operation is executed directly. The sub-program SUB addtime(n% ,t,madd) can be used to deliver the solution present in flask n% (1-3) for a fixed time interval t (in seconds). The sub-program checks if the flask is not empty and, in that situation, the SUB fillburette(n% ,ma) is automatically called. The mass actually added is returned in the variable madd. The SUB addmass(n%,m, madd) is used to deliver a mass approximately equal to m from the flask n% ; it also checks for the level of the solution. Again the added mass is returned in madd.The SUB readbalance(stb,flaginst- % ,mass) is used to read the balance through the RS232 serial interface. Mass is the variable containing the mass reading, stb is the value for the reading stability test. If two consecutive readings agree in between this value the sub-program isexited and the flaginstyo is set to 0. If more than ten readings are made the last value is returned in the ‘mass’ variable and the flaginst% is set to 1, indicating that the balance is unstable. The above set of sub-programs are sufficient for the implementation of most of the titration or standard additions procedures. Results and Discussion The gravimetric burette was evaluated for its accuracy in transferring mass from the burette flask to the reaction flask outside the case.The burette flasks were filled with water and the sensitivity of the balance was set to 1 mg. The mass was transferred to a previously weighed glass flask. The mass transferred was determined with an analytical balance with a sensitivity of 0.1 mg. The values of mass accessed by the microcomputer from the burette balance(mt) were compared with those obtained with the analytical balance(mr). Results for 120 comparisons (ten trials for each delivery time interval: 0.5, 2.0, 5.0 and 10 s, for each burette flask) showed that the differences (mt - mr) were never higher than +0.002 g or lower than -0.002g. The mt values are related to the mr values by the equation: mt = (-8.17 x 10-4 k0.34) + (0.9989 k 1.2 x 10-3)mr the correlation coefficient is 0.99998 and the error of the estimate is 9.6 X lO-4g.No systematic proportional or constant error between the two sets of measurements could be detected at the 95% confidence level.10 The flow rate of the gravimetric burette changes slightly as a function of the height of the liquid column inside the flasks. The change is not significant and does not alter the true measurement of the mass transferred. At the level of the bench the flow rate for a diluted aqueous solution is about 3.0 ml min-1, if the manifold described under Experimental is employed. If it is necessary, an increase in the flow rate can be achieved by placing the case higher than the level of the bench or by using larger bore tubing to assemble the manifold. The potentiometric titration curve for the titration of 5 ml of 0.01 mol dm-3 Fell solution with 0.001667 mol kg-1 K2Cr207 solution, monitored by using a Pt electrode and an Ag-AgCI reference electrode, is shown in Fig.5. The end-point was found using the second-derivative method. Note that the908 ANALYST, MAY 1992, VOL. 117 1063.92 ~ + + + + + + + 378.31 0 , 3.97 7.94 Masslg Fig. 5 Potentiometric titration curve obtained using the prototype gravimetric burette. Titrant, 1.667 X 10-3 mol kg-I K2Cr207; titrand. 0.01 mol dm-3 Fe” solution driven software is capable of slowing down the mass added near the end-point, permitting an increase in the precision of the determination. A sensitivity of 1 mg was employed in the burette balance in carrying out such titrations. The average of ten determinations and the standard deviation were (1.007 k 0.007) X 10-2 mol dm-3.The precision is good considering the small (5 ml) sample volume titrated. The gravimetric burette described here presents some advantages over previously described instruments. The unit is isolated from air currents, allowing the use of more sensitive balances. The same balance can be used with more than one flask, improving the cost to benefit ratio of constructing the unit. The total load of the accessories (support + empty flasks) is about 45 g. Therefore, it is possible to use high sensitivity balances in the construction of the burette. Refill of the burette is achievable and is performed automatically by the control program. The flasks do not need to be removed from the unit. Therefore, the over-all performance of the instrument is improved as factors affecting the weighing precision, such as a change in the condition of the tube connecting the flask to the valve, are not present.3 The tube is always ‘relaxed’ as is necessary for good precision in mass transference.3 The gravimetric unit can also be used to implement a totally gravimetric approach to a given analytical procedure.The mass of the sample (liquid or solid) can be found using the same balance of the unit, gaining access to the balance plate through the door at the top of the case. In so doing, it is important to keep the total load on the balance below the higher limit for the selected sensitivity. Although three flasks were used in the prototype described here, it is possible to add more flasks to the unit.By using a sensitivity of 0.01 g, the total mass of the support plus flasks can be as high as 1000 g. The atmosphere in the case can be kept inert (free of oxygen, for example) by maintaining a slight positive pressure of nitrogen. An inert atmosphere can be useful if one wishes to use the burette to perform standard additions in polarographic determinations. for example. If compared with the modern volumetric addition units commercially available, the gravimetric unit can offer the same performance at low cost per individual solution. No moving parts are present and less care need be expended in keeping the instrument in working order. One disadvantage of the gravimetric approach is, perhaps, that it does not enable the continuous addition of titrant to be performed when the reaction is sufficiently fast to permit the use of this type of procedure.However, a quasi-continuous procedure with rapid gravimetric additions followed by reading of the mass and potential without any stability test is being evaluated for use in routine determinations. Finally, more complex titration procedures requiring the addition of more than one reagent, back-titrations and multiple standard additions (as are necessary for the general- ized standard additions methods”) can be carried out by using the proposed gravimetric burette driven by the appropriate software. These applications of the gravimetric burette in routine laboratory tasks will be described in a subsequent paper. APPENDIX LISTING 1 Main QUICKBASIC 4.5 Sub-programs to Control the Gravimetric Burette ’ The following constant values associated with the ’ ’ should be declared in the main module of the program.’ The following sub-programs assume that the user card has ’ been initialized and that the asynchronous interface is in ’ step. For details of the interface and user card operation ’ see refs. 6-8. interface control and addresses of the devices CONST contr% = 1003, pa% = 1000, pb% = 1001, pc% = 1002 CONST ohstr% = 3, olstr% = 2, ihstr% = 5 , ilstr% = 4 CONST iack% = 32, oack% = 128, hill% = 255 CONST adc% = 254, dac% = 253, vadd% = 251, senadd% = 247 CONSTiof% = 1 ’ enableddisables 1/0, 1 = enabled, ’ 0 = disabled CONST tmax = 400 ’ time in seconds necessary to empty a ’ full burette flask COMMON SHARED dout% ,ma,stb ‘ initialization of the user card (see ref.7)ANALYST, MAY 1992, VOL. 117 OUT contr%,l93: OUT pa% ,hill%: OUT pb% ,ohstr% 909 ’ the following sub-programs are necessary to control the ’ gravimetric ‘ burette and to perform analytical procedures SUB fillburette(n% ,ma) I fill the burette n% (l<n%<4) empty% = 2 (n%-1) nr% = n% + 3 CALL readinter(247,di%) ’ checks for level sensor state IF (di% AND empty%) = 0 THEN GOSUB fill : EXIT SUB ’ flask is empty CALL valveon(n%) ’ flask is being emptied ’ waiting for an empty flask flask is not empty WHILE (di% AND empty%) > 0 CALL readinter( 247,di%) WEND CALL valveoff(n% ) GOSUB fill EXIT SUB fill: CALL readbalance(flaginst% ,mass,stb) mi = mass CALL valveon(nr%) WHILE ABS(mass-mi) < ma CALL readbalance(flaginst% ,mass,stb) WEND CALL valveoff(nr%) RETURN END SUB SUB addtime(n% ,t,madd) ’ delivers solution from burette n% ’ during the ’ time interval t (s).Returns madd, the true added mass told = t IF t > tmax THEN EXIT SUB CALL readbalance( flaginst% ,mass ,stb) mi = mass: maddi = 0 empty% = 2 (n%-1): ti=TIMER CALL valveon(n%) WHILE (TIMER - ti) < t CALL readinter(247 ,di% ) CALL valveoff(n%) CALL readbalance(flaginst% ,mass, stb) maddi = ABS(mass - mi) dt = TIMER - ti: t = t - dt CALL fillburette(n% ,ma%) CALL readbalance(flaginst% ,mass,stb) mi = mass ti = TIMER CALL valveon(n%) IF di% AND empty% = 0 THEN END IF WEND CALL valveoff(n%) CALL readbalance(flaginst% ,mass,stb) madd = maddi + ABS(mass - mi) t = told END SUB SUB addmass(n% ,m,madd) ’ adds solution from burette n% ’ until the mass m is ’ approximately added.Returns ’ madd, the true mass added910 ANALYST, MAY 1992, VOL. 117 mold = m IF m > ma% THEN EXIT SUB CALL readbalance (flaginst% ,mass,stb) mi = mass: empty% = 2 (n%-1): maddi = 0 CALL valveon(n% ) CALL readinter(247 ,di%) CALL readbalance(flaginst% ,mass,stb) IF di% AND empty% = 0 THEN CALL valveoff(n%) CALL readbalance(flaginst% ,mass,stb) CALL fillburette(n% ,ma%) CALL readbalance(flaginst% ,mass,stb) CALL valveon(n%) WHILE ABS(mass - mi) < m maddi = ABS(mass-mi): m = m - maddi mi = mass ENDIF WEND CALL valveoff(n% ) CALL readbalance(flaginstY0 ,mass,stb) madd = maddi + ABS(mass - mi) m = mold END SUB SUB readbalance(flaginst% ,mass,stb) ’ reads the balance ’ through RS232 in COM2 IF iof% = 0 THEN EXIT SUB OPEN “COM2: 9600,N,7,2,RS,DSO” FOR INPUT AS #1 n=O: mass1=0: mass:!= 2*stb DO UNTIL n > 10 OR ABS(mass2 - massl) =< stb GOSUB readmass mass1 = mass GOSUB readmass n=n+l mass2 = mass LOOP flaginst% = 0 IF n > 10 then flaginst% = 1 ELSE IF n =< 10 THEN dout% = dout% and 191 CALL outda(251 ,do%) CLOSE #1 EXIT SUB readmass: INPUT #1, mass$ mass = VAL(mass$) RETURN END SUB SUB valveon(nv% ) ’ IF nv% >6 OR nv% <1 THEN EXIT SUB dout% = dout% OR (2 (nv%-1)) turns the valve nv% on CALL outda(251 ,dout%) END SUB SUB valveoff(nv% ) ’ IF nv% > 6 OR nv% < 1 THEN EXIT SUB dout% = dout% AND (255-(2 (nv%-1))) CALL outda(251 ,dout%) turns the valve nv% off END SUB SUB outda(ad% ,bytetosend%) ’ used for output data to ’ the interfaceANALYST, MAY 1992, VOL.117 911 IF iof% = 0 THEN EXIT SUB OUT pb% ,ohstr% OUT pa%,ad% OUT pb% ,olstr% OUT pb?” ,ohstr% OUT pa% ,bytetosend% OUT pb% ,olstr% OUT pb% ,ohstr% WHILE (INP(pc%) AND oack% = 0) : WEND WHILE (INP(pc%) AND oack% = 0) : WEND OUT pa% ,hi% END SUB SUB readinter(ad% ,di%) ’ for input data from the ’ interface. This sub-program ’ can access the sensor state ’ and the ADC converted value IF iof% = 0 THEN EXIT SUB OUT pb% ,ohstr% OUT pa% ,ad% OUT pb% ,olstr% WHILE (INP(pc%) AND oack%) = 0: WEND OUT pb% ,ohstr% OUT pb% ,ihstr% OUT pb% ,ilstr% di% = INP(pa%) WHILE (INP(pc%) AND iack%) = 0: WEND END SUB References 1 Tamberg, T., Fresenius’ 2. Anal. Chem., 1978, 291, 124. 2 Luft, L., Talanta, 1980,27,221. 3 Kratochvil, B . , and Nolan, J . E.. Anal. Chem., 1984, 56, 586. 4 Chipperfield, J . R., and Webster, D. E., Anal. Chim. Ada, 1987, 197, 373. 5 Mak, W. C., andTse, R. S., J . Chem. Educ., 1991, 68, A95. 6 Souza, P. S., and Pasquini, C., Lab. Micro, 1990, 9, 77. 7 Malcolme-Lawes, D. J . , Lab. Micro., 1987, 6, 16. 8 Malcolme-Lawes, D. J . . Lab. Micro., 1987, 6. 122. 9 Horowitz, P., and Hill, W., The Arc of Electronics, Cambridge University Press, Cambridge, 8th edn., 1987. 10 Eckshlager, K., Errors, Measurements and Results in Chemical Analysis, Van Nostrand Reinhold, London, 1972, p. 144. 11 Saxberg, B. E. H . , and Kowalski, B. R . , Anal. Chem., 1979,51, 1031. Paper 1 I04873 B Received September 23, 1991 Accepted December 2, 1991

ISSN:0003-2654    

DOI:10.1039/AN9921700905

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, MAY 1992, VOL. 117 913 Colorimetric Determination of Iron(ii) and Iron(lli) in Glass* Elizabeth W. Baumann Westinghouse Savannah River Co., Savannah River Site, Aiken, SC 29808, USA A simple method t o determine relative amounts of Fell and Fell' was developed for the analysis of glass containing nuclear waste. The pulverized glass sample was dissolved in a sulfuric acid-hydrofluoric acid mixture with ammonium metavanadate added t o preserve the Felt content. The Vv-VIv couple, which is strongly acid-dependent, caused the Fell to be oxidized to Fell' in the highly acidic dissolution and then to be regenerated when the solution was adjusted t o pH 5 for formation of the magenta Fell-Ferrozine complex. Boric acid was added t o complex the fluoride and permit regeneration of Fell, otherwise Fell' would be stabilized as a fluoride complex. The Fell content was determined from the absorbance at 562 nm, and total Fe was determined from the absorbance of the same solution after ascorbic acid had been added to reduce Felllto Fell.The relative standard deviation was 5%, and the range was 3-97% for Fell in total Fe. The method was validated by analysing a glass frit spiked with ammonium iron(1r) sulfate and/or ammonium iron(ll1) sulfate solutions. The method is applicable to glasses (or minerals) that contain more than about 5% total Fe. Keywords: Glass analysis; iron(ii) determination; iron(iii) determination; iron in glass; Ferrozine At the Savannah River Site, a process has been developed for incorporating defence nuclear waste in a glass matrix for disposal.In this process, the redox state in the glass melter is an important process control parameter. If the mixture of the glass frit and waste in the glass melter is too reducing, nickel sulfide and other metals may precipitate within the melter.1 If the mixture is too oxidizing, the durability of the glass product is affected.' A recognized indicator of redox conditions in systems such as glasses and minerals is provided by the FeI1l-Fe1I couple, which is reversible and responsive to redox conditions. Hence the relative amounts of Fell and Fe"' will define the redox state. A simple and reliable procedure for this determination was needed for routine process control analysis of the radioactive glass product. The determination of both Fe" and Fe"' in a sample is challenging because Fell may be converted into Fell1 through inadvertent oxidation prior to the final analytical measure- ment.A non-destructive method such as Mossbauer spectro- scopy3 was attractive for this application, but the technique is not suited to routine analyses because of the long data collection time, sophistication, and limited sensitivity for low concentrations of either oxidation state. Many colorimetric methods have been described, in which differentiation between oxidation states is accomplished with chromogens specific for either the Fell or Fellr ion."-" For colorimetric methods the sample must be dissolved, which can provide an opportunity for significant oxidation of Fell to occur. Iron(i1) is readily oxidizable in the acidic fluoride medium common to most glass dissolutions.Techniques for rapid dissolution and/or inert environments have been recom- Chemical stabilization by vanadate has also been used.-' This paper describes the development and application of a simple and rapid method suitable for the remote routine analysis of glass that contains nuclear waste.'OThe method has been used successfully for comparison studies with Mossbauer determinations3 and in other studies," but has never been fully described in the literature. This paper describes the method and its validation. * This paper was prepared in connection with work done under Contract No. DE-AC09-89SR18035 with the US Department of Energy. Experimental Description of the Method Features The procedure combined several features to provide an integrated method for the reliable determination of Fell and Fell'. The glass sample, in a polystyrene vial, was dissolved in sulfuric-hydrofluoric acid as described by Jones et a1.S The mixture was not heated, except for the heat generated by the dissolution reaction.Because of the preservation of Fell by vanadate, discussed below, no inert gas blanket or protection from air was required. Iron(I1) was preserved by taking advantage of the acid dependence of the VV-Vv couple, as described by Wilson4 and illustrated by the reversible equation: acid dissolution pH 5 buffer V5+ + Fez+ -V4+ + Fe3+ In a strongly acidic solution such as the dissolution mixture, Vv is a strong oxidizing agent and quantitatively oxidizes Fe" to Fell', producing VIv.The oxidation potential is substantially reduced at pH 5 , and Fe" is regenerated from the V" preservative. Ferrozine (Hatch) was used as the chromogen because of its rapid colour formation, stability and selectivity.12 The com- plex has a broad peak with a maximum absorbance at 562 nm and a molar absorptivity of about 28000 1 mol-1 cm-1. Because this application required that only the relative amounts of Fell and Fe"' be determined, volumes were not closely controlled. Measurements of Fe'l and total Fe were made on the same solution before and after addition of solid ascorbic acid. The procedure can easily be made quantitative if desired with controlled volumes and use of standards. Procedure The stepwise analytical procedure, which assumes about 10% total Fe in the sample, is as follows. (1) Pulverize the glass to a free-flowing powder (mesh size was not determined) in a Brinkmann vibratory mill with an agate container and balls; avoid using Fe-bearing components.Keep the time for pulverization to a minimum to avoid possible oxidation of the Fell in the glass.914 ANALYST, MAY 1992. VOL. 117 (2) Weigh, to 0.1 mg, 10-30 mg of the pulverized sample into a polystyrene vial (typically 1 in diameter and 3 in high). (3) Add 100 yl of 0.85 mol dm-3 ammonium metavanadate solution (0.3 g of NH4V03 dissolved in 1 ml of concentrated H2S04 + 2 ml of H20). (4) Add 500 pl of concentrated (95-98%) sulfuric acid ( 5 ) After swirling to slurry the glass into the mixture, cautiously add, dropwise, 1000 pl of concentrated (48-51%) hydrofluoric acid (HF).The solution will heat up and effervesce as the glass dissolves. (6) When dissolution is complete, add 20 ml of saturated boric acid solution (about 6.5 g of H3B03 in 100 ml of H20). (7) Transfer an aliquot (typically 200 yl) containing 20-50 pg of total Fe into a second container. (8) Add 20 ml of buffered Ferrozine solution {prepared by diluting the following to 1000 ml: 50 ml of pH 5 buffer [270 g of sodium acetate trihydrate (CH3COONa.3H20) and 60 ml of glacial acetic acid (CH3COOH) diluted to lo00 ml] and 50 ml of 1 % Ferrozine [3-(2-pyridyl)-5,6-bis(4-phenylsulfonic acid)-1,2,4-triazine] (1 g of Ferrozine dissolved in 100 ml of (9) After 10 min, measure the absorbance of the solution in a 1 cm cuvette at a wavelength of 562 nm to determine the Fe" content.(10) Dissolve about 5 mg of solid ascorbic acid (C6Hg06) in the remaining solution and allow the colour to develop for about 2 mins. (11) Measure the absorbance at 562 nm to determine the total Fe content. Matched 0.01 mol dm-3 ammonium iron(ii) sulfate and ammonium iron(ii1) sulfate solutions were prepared for the validation studies. These solutions were made up in 0.1 mol dm-3 sulfuric acid to help retard the oxidation of Fe". (H2S04) - H20)). Comments The aliquots of the dissolved glass, which contain concen- trated acids, should be kept as small as is reasonable and must not exceed the capacity of the buffer added. A pH of 5-9 is essential for proper colour formation. Table 1 Composition of glass frit 16513 Component Content (% d m ) Si02 Na20 Li20 Zr02 B203 MgO 68 10 13 7 1 1 Also, in order to ensure quantitative preservation of the Fe", the vanadate should be approximately equivalent to the Fe" expected in the sample.A large excess of vanadate causes the generation of excess of Fe" and drifting readings, as discussed later. Because Ferrozine detects very small concentrations of Fe, trace amounts of Fe in the reagents or apparatus may cause high blanks, high results or irreproducible results. Vials and pipette tips should not be reused. Results and Discussion Validation of the Method Glass standards with known Fell and Fe"' contents cannot be reliably prepared. Therefore, validation of the method was carried out by analysing ammonium iron(ir) sulfate and ammonium iron(ii1) sulfate solutions spiked into an Fe-free glass frit.The composition of the frit, which was also used in other aspects of the glass programme," is given in Table 1. The Fell and/or Fe"' solutions and the spiked glass frit were used to study the recovery of Fe". The typical results, presented in Table 2, demonstrate the effectiveness of the addition of vanadate in preserving Fell. Further confirmation that the recovery of Fe" was improved by including vanadate in the procedure is provided by the results shown in Table 3. Experimental glass samples were analysed both with and without the addition of vanadate. All samples without vanadate showed a negative bias. The results presented in Table 4 show that the addition of vanadate preserved the Fell in the H2S04-HF solutions used for dissolution for at least 5 d.These samples were stored in closed vials in air. These observations agree with those of Wilson,4 who reported that it was possible to leave dissolved samples in air for up to 72 h to attain dissolution, without loss of Fe". Precision, Accuracy and Range The Fell : Fe"' ratios obtained using the proposed method were compared with those obtained with Mossbauer spectroscopy by Karraker3 and by Hunter et aZ.11 The ratios determined with the two methods agreed well. Precision and accuracy values are given in Table 5. The relative standard deviation was less than 5% for three or more determinations. The relative accuracy, from the ammonium iron(ii) sulfate and ammonium iron(ii1) sulfate spike studies, is also calculated to be 5%.The range of the method is judged to be from 3 to 97% Fe" in total Fe. One limitation lies in establishing the magnitude of the blank value. It is shown in Table 2 that the results for individual ammonium iron(i1) and ammonium iron(ii1) salts were within 3% of the values expected for the pure salts. Table 2 Validation with ammonium iron(ii) sulfate and ammonium iron(iii) sulfate solutions Fe" in total Fe (%) Spike Fri t Vanadate Expected Fell Fellr 1 : I t 1:l 1:2 1 : l 1:l 1 : 2 1 : l 1 : l 1 : 2 No No No No No No No No Yes Yes Yes Yes Yes Yes No No Yes No No Yes No No 100 0 50 50 33 50 50 33 50 50 33 Found 96.5 2.0 49.2 48.1 32.0 51.5 44.7 28.6 48.1 41.9 28.1 * In addition to buffered Ferrozine and ascorbic acid for colour formation. t Mixtures are Fe" : Fell'.ANALYST, MAY 1992.VOL. 117 915 Table 3 Analysis of glass samples with and without the addition of vanadate Glass No. 1 2 3 4 5 6 7 8 With vanadate 76.9 32.0 12.5 52.4 32.4 43.8 51.0 90.3 Without vanadate 63.4 28.1 6.1 42.2 28.1 39.4 46.5 79.5 Fell in total Fe (YO) Re I at ive bias (YO) - 17 - 12 -51 - 19 -13 - 10 -9 - 12 Table 4 Effect of time on glass dissolution Time since Sample Vanadate dissolution Fell : Fell' (1 : 1) with glass frit Yes 0 Yes 1 d in air Yes 5 d in air No 2 h in air Yes 2 d in air No 2 h under N2 GlassNo. 1 No 0 Glass No. 9 Yes 0 Glass No. 10 No 0 Fell in total Fe (Yo) 48.1 50.2 52.6 63.4 49.3 18.3 18.1 16.0 16.0 Table 5 Precision of measurements Fell in total Fe (%) Set 1 Standard Results Mean deviation 17.2 20.8 17.6 18.5* 18.5 k1.6 18.3 18.9 19.1 18.8 k0.4 17.6 18.8 18.3 18.2 k0.6 Average: 18.5 k0.9 * Analysed 2 d later. The effect of other redox-sensitive elements on the method was not systematically studied.Ferrozine itself has few interferences, but the effect of other oxidizable species on the method is open to question. The experimental glasses analysed here contained formate, uranium, manganese and nickel. Limitations of the Method The method has proved to be satisfactory for the analysis of glass samples that contain >5% Fe.3-11 However, limitations have been encountered as discussed below. Control of vanadate concentration Vanadate affects the kinetics of the formation of the Fell- Ferrozine complex. Although complete colour formation in the absence of vanadate was rapid, in the presence of vanadate about 10 min were needed for full colour development.Further, a significant excess of vanadate led to drifting readings and excessive colour formation. 0.36 0.35 I I Q) 0.34 0.33 n 0.31 0.3 I I I I 0 5 10 15 20 25 Time/min Fig. 1 Effect of vanadate on colour development. Vanadate to Fe ratio: A, 0; B , 0.5; C, 1.2; and D, 5. Ammonium iron(u) and ammonium iron(r1i) sulfate, 10 x 10-6 rnol dm-3 0.07 0.06 r 0.05 0.04 0.03 0.02 0.01 u C f Q t 0 5 10 15 20 25 Time/mi n Fig. 2 Effect of vanadate on colour development. Vanadate to Fe ratio: A, 0; B , 0.5; C, 1.2; and D, 5.0. Ammonium iron(ri1) sulfate, 20 x 10-6 rnol dm-3 25 d 0 20 7 X 7 15 t .- E g 10 . -? -3 5 0 1 I I I I I I I I 2 4 6 8 10 12 14 16 Vanadate c~ncentration/lO-~ mol dm-3 Fig.3 Effect of vanadate concentration on rate of colour develop- ment. Total dilution of matrix [(initial volume x final volume)/ aliquot]: A, 31; 0,62; 0,125; A, 125 and 62; +, 250; 0,500; and 0, 2000. Line shown without matrix The kinetic effect, and also the effect of excess of vanadate, on colour formation is shown in Fig. 1. The solution contained 1 x 10-5 rnol dm-3 Fell and 1 x 10-5 rnol dm-3 Fell'. The rate of colour development increased as the proportion of vana- date relative to Fe was increased beyond equivalence. The absorbance continued to increase beyond that corresponding to the Fell in the original solution. The V'" and the Ferrozine in solution apparently slowly drive the reduction of Fell1 to Fe", with formation of the Fell-Ferrozine complex and the abnormal absorbances ob- served. This action is illustrated in Fig.2 where the original916 ANALYST, MAY 1992, VOL. 117 solution contained Fell1 only. The rate of colour generation increased with the amount of vanadate in the system. Effect of glass dissolution mixture The glass dissolution mixture consists of H2SO4, HF, H3B03 and sometimes acid neutralizers such as NaOH or CH3COONa at significantly higher concentrations than the Fell sought. The presence of too much of this matrix affected colour development. It was found that the method worked satisfactorily for the application described here, which invol- ved a nominal total dilution {TD = [(initial volume) x (final volume)]/(aliquot volume)} of 1000-2000. For glasses with lower Fe concentrations, larger aliquots had to be used to attain the required sensitivity.Absorbance values continuously drifted upward at TDs below about 500. This behaviour is illustrated in Fig. 3, which compares the rate of colour formation in water and in dissolution mixtures at various TDs. The vanadate and Fe concentrations were approximately equal at 1 x 10-5 mol dm-3. The drifting phenomenon was attributed to an unidentified component from the dissolution matrix solution, which reduces Fell1. This drift at lower TDs limits the applicability of the method. The information in this paper was developed during the course of work done under Contract No. DE-AC09-89SR18035 with the US Department of Energy. 1 2 3 4 5 6 7 8 9 10 11 12 13 References The Treatment and Handling of Radioactive Wastes, eds. Blasewitz, A. G., Davis, J . M., and Smith, M. R., Battelle Press, Columbus, OH, 1983. Jantzen, C. M., and Plodinec, M. J., J. Non-Cryst. Solids, 1984. 67, 207. Karraker, D. G., Adv. Ceram. Mater., 1988. 3,337. Wilson, A. D., Analyst, 1960, 85, 823. Jones, D. R., IV, Jansheski, W. C., and Goldman, D. S., Anal. Chem., 1981,53,923. Begheijn, L. T.. Analyst, 1979, 104, 1055. Abe, S., Saito, T., and Suda, M., Anal. Chim. Acta, 1986,181, 203. Fritz, S. F., and Popp, R. K., Am. Miner., 1985, 70, 961. Close, W. P., and Tillman, J. F.. Glass Technol.. 1969, 10,134. Baumann, E. W., Coleman, C. J., Karraker, D. G., and Scott, W. E., 194th National Meeting of the American Chemical Society, New Orleans, LA, September 1987; Abstract ANYL- 175, American Chemical Society, Washington, DC, 1987; and US Department of Energy Report DP-MS-87-18, Aiken, SC, 1987. Hunter, R. T., Edge, M., Kalivretenos, A., Brewer, K. M., Brock, N. A., Hawkes. A. E., and Fanning, J. C., J. Am. Ceram. SOC., 1989, 72, 943. Stookey, L. L., Anal. Chem.. 1970,42, 779. Soper, P. D.. Walker, D. D., Plodinec, M. J., Roberts, G. J., and Lightner, L. F., Am. Ceram. SOC. Bull., 1983, 62, 1013. Paper 1 /053 79E Received October 22, 1991 Accepted December 18, 1991

ISSN:0003-2654    

DOI:10.1039/AN9921700913

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, MAY 1992. VOL. 117 917 Conductimetric and Spectrophotometric Determination of the Volatile Acidity of Wines by Flow Injection" Flavio Guimaraes Barros and Matthieu Tubinot lnstituto de Quimica, Universidade Estadual de Campinas, CP 6154, 13081 Campinas, Sao Paulo, Brazil Usual methods for the determination of the volatile acidity of wines are relatively slow, as about 40 min are necessary t o perform one analysis. In this work, a method was developed which provides results in a much shorter time. About 60 analyses can be performed in 1 h. The conductimetric analysis consists of the injection of the wine sample into a de-ionized water stream which then flows past a poly(tetrafluoroethy1ene) membrane separator. The acetic acid diffuses through the membrane into another water stream that passes through a conductivity cell.The spectrophotometric method is similar. The acetic acid diffuses into a stream of Bromocresol Purple solution, at pH 7.0, which passes through a flow cell in a spectrophotometer set at 540 nm. For comparison, analyses were also carried out by the method of Jaulmes. Keywords: Flow injection; volatile acidity; wine; conductimetry; spectrophotometry Fonzes-Diacon and Jaulmes' defined volatile acidity for wines as: 'The volatile acidity is the assembly of the fatty acids of the acetic series that are in the wine. The lactic and succinic acids and also carbonic acid and the free and combined sulfurous anhydride are excluded from the volatile acidity'. The most recent official definitions are usually very close to this concept.The known methods for the determination of volatile acidity are based on distillation processes. Among these methods the more usual ones are those of Duclaux,* Ferr6,3 Jaulmes4 and the Cash-Still method.5 All these involve distillation and are slow, in addition to being subject to a variety of errors. The first method is essentially an ordinary distillation whereas the other three are steam distillations. The method of Jaulmes, however, involves use of a column containing a helicoidal band of inox (stainless steel) screen to avoid distillation of non-volatile acids such as succinic and lactic acids. In all these methods the distilled acid is titrated with standard NaOH solution, after which it is necessary to carry out a titration with I2 to determine the concentration of SO2 that must be subtracted from the result obtained in the NaOH titration.Flow injection (FI),b which is essentially the introduction of a sample into a solution stream continuously passing through a detector, is a very important methodological innovation in analytical chemistry, which is customarily characterized by simple chemical processes, low-cost apparatus, easy manipu- lation and ability in yielding results that are usually of good quality. The use of an FI system, with a gas-diffusion membrane,' to separate volatile acids from wine, allows very good quality results in a rapid process. About 60 (spectropho- tometric) or 80 (conductimetric) samples can be analysed in 1 h. Similar systems have been used previously to determine the volatile acidity of vinegars8 and spoilt beer.9 Experimental Materials A standard solution of approximately 1% m/v of acetic acid (analytical-reagent grade) was prepared and titrated.Solu- tions of 0.02, 0.04, 0.06, 0.08 and 0.10% m/v in acetic acid were obtained from this solution by dilutions to 100.0 ml. Bromocresol Purple (BCP) solution (1 X 10-4 mol dm-3). Prepared by dissolving 0.27 g of BCP in 10 ml of ethanol, the * From the M.Sc. Thesis of F. G. B. f To whom correspondence should be addressed. volume being diluted to 500 ml with water; 50ml of this solution were diluted to 500 ml to obtain the working solution. The pH was adjusted to 7.0 by dropwise addition of dilute NaOH solution. In order to avoid absorption of C 0 2 from the air, the BCP solution was kept in a bottle protected by a tube containing CaCI2-NaOH-CaCI2 (respite).The solution was pumped to the FI system by another tube. Water used in the experiments was always boiled (de- gassed) and de-ionized. Apparatus Peristaltic pump. Ismatec mp13 GJ4. Conductimeter. Micronal Model B-331, connected to a Spectrophotometer. Single-beam Carl Zeiss Model PM2D Gas diffusion cell. This cell has been described previouslylo Conductimetric cell. This has been described previously. 1 0 Sampling valve. This has been described previously. 11 Volatile acidity is expressed, in this work, as mass (8) of chart recorder. set at 540 nm, connected to a chart recorder. and is similar to that of van der Linden.' acetic acid in 100 ml of solution (wine). ml min-' L'\ c, r M e t e r A1 A2 I I W Fig.1 Conductimetric FI manifold: T, ion-exchange resin column; P, peristaltic pump; S, sample inlet; V, sampling valve system; B, water-bath; M, diffusion cell; C, conductance flow cell; W, waste; and A, and A2. de-ionized water streams ml min-1 - Fig. 2 Spectrophotometric FI manifold: P, peristaltic pump; S, sampling inlet; V, sampling valve system; M, diffusion cell; E, spectrophotometer; W, waste; A l , de-ionized water stream; and I , BCP solution stream918 ANALYST, MAY 1992, VOL. 117 Methods Conductimetric A schematic flow diagram for the conductimetric method is shown in Fig. 1. The injected sample (S), de-gassed wine, is combined with the de-ionized water carrier stream (A1), pumped at a flow rate of 1.26 ml min-1, and is passed through the coil L2 kept in a temperature-controlled bath.In the diffusion cell (M), acetic acid and other volatile components of the wine diffuse into another de-ionized and de-gassed water stream, which passes through the conductimetric cell (C). In order to guarantee de-ionization of the water stream A2, an additional ion-exchange resin treatment is carried out in a column (T) introduced into the system. The temperature control of the streams Al and A2 is performed in the coils L1 and L2. The diffusion (M) and conductimetric cells (C) are also temperature controlled by the bath B, which is simply a circulating tap water-bath. The temperature (25 "C) was kept constant within k 0 . 5 "C or less during the entire day. Spectrophotometric The flow system used for the spectrophotometric analysis is shown in Fig.2. It is similar to that used for the conductimetric analysis. The ion-exchange column and water-bath are not necessary, and the detector is a spectrophotometer set at 540 nm. The stream A2 carries the BCP (1 x 10-4 mol dm-3) solution. Results and Discussion Flow rates of the donor and acceptor streams were adjusted at 1.26 ml min-1 as a function of the height of the signal, relative standard deviations and frequency of analysis. The same situation was found to be adequate for the spectrophotometric and conductimetric methods. The optimum concentration for BCP was found to be 1 X 10-4 mol dm-3, at 540 nm, with the pH of the acceptor stream adjusted to 7.0. Other acid-base indicators were tested, such as Bromothymol Blue in an aqueous solution adjusted to pH 8.0.However, the results obtained were invariably higher than expected. In both the spectrophotometric and conductimetric methods, many different sampling loops (the volume of the loop is the volume of sample injected by the valve), of various materials, volumes and internal diameters were tested, as retention of acetic acid on the walls of the loops was observed. This phenomenon was responsible for an increase in the peak height of the decreasing calibration graph.8 As a complete study of the materials, sizes and diameters of the loops would be an exhaustive task, an empirical selection was made. However, special attention had to be paid to the choice of the size and material of the sampling loops for initiating either of the two methods.In this work, it was empirically established that either a polyethylene loop (volume 120 p1; 0.8 mm i.d.) or a poly- (tetrafluoroethylene) (PTFE) (volume 62 p1; 0.9 mm i.d.) can be used in the conductimetric method without significant interference from retained acetic acid. For the spectropho- tometric method, a polyethylene loop (volume 240 PI; 1.6 mm i.d.) afforded the best results. In order to minimize the influence of the retention of the acetic acid in the loop, air was passed through it after each sampling. Carbon dioxide must be eliminated from samples before performing the analysis, as the gas diffuses through the PTFE micro-porous membrane, enhancing the resulting signal. This elimination is easily performed by subjecting the wine sample to low pressure, for about lOmin, with a water aspirator vacuum pump.As wines usually contain SO2 and sulfite it is necessary to oxidize these species to sulfate. Some drops (about 5) of H202, 0.5 volumes in about 10 ml of sample, are sufficient. The H202 must not be more concentrated than this, otherwise ethanol will be oxidized to acetic acid, with a consequent increase in the values obtained. A typical FI profile for the conductimetric method is shown in Fig. 3. The FI profile for the spectrophotometric method can be seen in Fig. 4. The calibration graphs that correspond to these profiles are shown in Fig. 5. Five wines were analysed by each method, four being the same wines in both instances. The measurements of the concentrations can be obtained graphically or by fitting first- and second-order equations for the spectrophotometric and conductimetric methods, respec- tively, by using an electronic calculator. The non-linearity of the calibration graphs in the conductimetric procedure can be Time - Fig.3 Calibration and sample runs for the determination of volatile acidity (conductimetric system). From left to right: triplicate signals for acetic acid standards (0.02, 0.04, 0.06, 0.08 and 0.10 g of acetic acid per 100 ml of solution); triplicate signals for wines; standards in the reverse order. Polyethylene sampling loop. 120 pl (0.8 mm I.d.) t I _ (D 0 v) .- 0.10 ,I0 min, 1 Time - Fig. 4 Calibration and sample runs for the determination of volatile acidity (spectrophotometric system). From left to right: triplicate signals for acetic acid standards (0.02, 0.04, 0.06, 0.08 and 0.10 g of acetic acid per 100 ml of solution); triplicate signals for wines; standards in the reverse order.Polyethylene loop (240 PI; 1.6 mm i.d.)ANALYST, MAY 1992, VOL. 117 919 Table 2 Comparison between values obtained by the FI conductimet- ric, FI spectrophotometric and Jaulmes methods, using the statistical Student’s t-test. Tabulated r value for the degree of freedom (Y) 4 is 2.776 (a = 0.05); Y = nl + n2 - 2 and nl = n2 = 3 in this instance Sample tl* f 2 t t3* t4§ t 5 l l t6ll 0.50 0.00 1.26 0.50 0.95 1.26 0.50 0.50 0.32 1.00 0.00 0.32 0.00 1.50 0.00 1.50 0.00 0.95 0.50 2.00 1.50 2.50 2.00 0.50 0.50 - 2.75 - 2.35 - - 0.50 1(R) 2(R) 3(R) 4(W) 5(W) 6(R) - - - - * ti ( i = 1-6) is the calculated Student’s t values: tl = Lol versus Lo2.t t2 = Lol versus Loj. * f3 = Lol versus Jaulmes. t4 = Lo2 versus LO^. 7 rs = Lo2 versus Jaulmes. 11 t6 = Lo3 versus Jaulmes. 0 0.02 0.04 0.06 0.08 0.10 0.12 Acetic acid/g per 100 ml of solution Fig. 5 Calibration graphs for the determination of volatile acidity. Details as in Figs. 3 and 4. h = Peak height in millimetres. A , Conductimetric and B. spectrophotometric system ~~ ~ Table 1 Volatile acidity of wines (g dl-1 acetic acid in wine) determined by the conductimetric and spectrophotometric FI systems and the method of Jaulmes4 Spectro- Conductimetric photometric LOl* LO?? o = f 0.002g 0.063 0.064 0.065 0.064 0.055 0.055 0.034 0.033 0.041 0.042 Sample (YO m/v) W ) l I 2( R) 3(R) 4( W)Il 5(W) 6(R) - - Lo3* 0 = k0.002 (Yo m/v) 0.063 0.066 0.052 0.038 0.035 - Jaulmes to (% m/v) 0.067 f 0.004 0.064 k 0.004 0.055 k 0.004 0.037 k 0.002 0.048 k 0.003 0.034 -t 0.002 * Lol = Loop 1, volume = 120 yl.i.d. = 0.8 mm, polyethylene. t Lo2 = Loop 2, volume = 62 yl, i.d. = 0.9 mm, PTFE. $ Lo3 = Loop 3, volume = 240 yl, i.d. = 1.6 mm, polyethylene. 5 o = Estimates of standard deviation. 1 (R) = Red wine. I( (W) = White wine. explained by the fact that acetic acid is a weak electrolyte and, as a consequence, conductivity is not linearly related to concentration. In the spectrophotometric method, deviations from linearity, as a consequence of the limitations of BFer’s law, were observed only above an acetic acid concentration of 0.12% m/v, which is outside the working range. In Table 1, the values for the volatile acidity obtained by the FI conductimetric and spectrophotometric methods, and also by the known method of Jaulmes4 are reported.The statistical t-test9 was used to compare the results obtained by the proposed methods and also by the method of Jaulmes4 (Table 2). Except for the t value (2.75, t3 in Table 2), obtained when comparing the FI conductimetric result for sample 5 with the Jaulmes result, which is close to the tabulated limit (2.776),12 no significant differences were observed between results at the 95% confidence level. Comparing operationally the spectrophotometric and con- ductimetric methods, the former appears to be simpler as it is not necessary to control the temperature of the detector or to use an ion-exchange resin column in the flow system. However, the conductimetric method allows a slightly more rapid sampling rate (about 80 h-1) than the spectrophoto- metric (about 60 h-1) method.The authors are grateful to FundaG3.o de Amparo a Pesquisa do Estado de S3o Paulo, FAPESP, and to Conselho Nacional de Desenvolvimento Cientifico et Technologico, CNPq, for financial support. The authors also thank Professor Roy E. Bruns for the English revision of the manuscript. 1 2 3 4 5 6 7 8 9 10 11 12 References Fonzes-Diacon, and Jaulmes, P., Bull. Pharm. Sud-Est, 1930, 34, 170. Duclaux, E., Ann. Chim. Phys., 1874, 2,289. FerrC, L., Ann. Falsif. Fraudes, 1930, 23, 323. Jaulmes, P., Bull. SOC. Chim. Fr., 1937,4, 157. Official Methods of Analysis of the Association of Official Analytical Chemists, ed., Horowitz, W., 11th edn., 1970. sect. 11036-11039, p. 187. RfiiiEka, J., and Hansen, E. H . , in Chemical Analysis, ed. Winefordner, J. D., Wiley, New York, 1988, vol. 62, and references cited therein. van der Linden, W. E., Anal. Chim. Acta, 1983, 151, 359. Tubino, M., and Barros, F. G.. J. Assoc. Off. Anal. Chem., 1991, 74, 346. Williams. J . G., Holmes, M., and Porter, D. G., J. Autom. Chem., 1982, 4. 176. Pasquini, C., and Faria, L. C., Anal. Chim. Acra, 1987,193,19. Tubino, M., and Barros. F. G., Quim. Nova, 1991, 14. 49. Eckschlager, K . , Errors, Measurements and Results in Chemical Analysis, ed., Chalmers, R. A., Van Nostrand Reinhold, New York, 1972, pp. 107-112. Paper 1 /03543 F Received July 12, I991 Accepted November 6, 1991

ISSN:0003-2654    

DOI:10.1039/AN9921700917

出版商:RSC    

年代:1992

数据来源: RSC