ISSN:0003-2654    

DOI:10.1039/AN98611FX029

出版商:RSC    

年代:1986

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98611BX031

出版商:RSC    

年代:1986

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98611BP033

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST AUGUST 1986 VOL. 11 I 865 Enzymatic Determination of Urea in Water and Serum by Optosensing Flow Injection Analysis Tracy D. Yerian Gary D. Christian* and Jaromir RdLiEkat Center for Process Analytical Chemistry Department of Chemistry BG- 70 University of Washington, Seattle WA 98795 USA Reflectance spectrophotometry was applied to the flow injection measurement of pH in order to perform an assay of urea in water and serum samples via the enzymatic (urease) degradation of urea to ammonia and hydrogen carbonate. Detection is based on commercial non-bleeding acid - base indicator papers situated in the flow stream of an integrated micro-conduit at the tip of a fibre-optic bundle. The sample and reagent are injected via a split-loop injection technique. The serum is analysed by a stopped-flow kinetic measurement to avoid errors in measurements due to the variability in the background sample pH.Keywords 1 Urea assay; optosensing; reflectance flow injection analysis; enzymatic degradation Two general types of optical fibre sensors have been devel-oped for the analysis of chemical species.’ Fibres can be used simply as light carriers with no transducers at the end. More recently sensors using the immobilisation of reagents and the integration of optical fibres have been designed. When a single fibre is used it is necessary to distinguish between the incident and detected radiation either temporally or by wavelength. The main advantage of using a single fibre in a “probe” sensor is that an extremely small probe size becomes possible and this has advantages in bioanalytical techniques.* Other potential advantages of using optical fibres for in vivo sensing include their mechanical flexibility the absence of risk of electric shocks inexpensive construction transmission of only low-energy radiation and relatively good biocompatibility .3 Single fibre systems are also easily applicable to fluorescence measurements as the wavelengths of the detected and incident radiation are different. A pH sensor based on the fluorescence of cellulose-immobilised fluoresceinamine has been reported4; non-fluorescent reagents that form fluores-cent complexes with metals have also been incorporated into fibre-optic sensors.536 In bifurcated optical sensors separate fibres transmit incident and detected radiation and only that part of the reagent phase that falls within both the cone of incidence and the cone of detected radiation is observed.Chemical sensors based on fibre optics may be classified as reversible in which the reagent phase is not consumed by its interaction with the analyte or non-reversible in which the reagent phase is consumed. Optical sensors that are based on the use of reagents immobilised on a solid support have been developed for a number of chemical species. When solid support matrices are employed in sensors it is difficult to measure the transmitted light. In these instances the intensity of the reflected light can be used as a measure of the change in colour of an immobilised reagent phase. The feasibility of using solid-phase immobilised dyes as reversible reagents has been demonstrated by Goldfinch and Lowe in their optoelec-tronic sensors for serum albumin urea penicillin and glu-cose.7,8 pH probes have been developed based on reflectance measurements from indicator dyes immobilised on micro-spheres or polymer supportsslJ; this same type of sensor design has also been adapted to O2 and glucose measure-ment.1 2 ~ 3 * To whom correspondence should be addressed. i Present address Chemistry Department A Building 209, Technical University of Denmark DK-2800 Lyngby Denmark. The fibre-optic sensor coupled to an immobilised chemical reagent has been studied by Kirkbright et al. in a flow system. 14 RfiiiZka and Hansen have incorporated reflectance spectrophotometry via optical fibres into a flow injection analysis (FIA) scheme “optosensing.”ls The solid support at the end of the fibre is situated in a flow path and this flow cell is incorporated into the integrated microconduits described by RfiiiCka in earlier work.16 In this study the integrated microconduit design is used with some improvements on the first microconduit design that incorporated the solid support interfaced to a fibre-optic bundle.The flow of solutions through the cell serves two purposes to transport the sample zone through the detector in the FIA mode and to renew the reagents which cannot be or intentionally are not immobi-lised within the optosensor. This illustrates a basic advantage of the integrated microconduit over a probe design irrevers-ible chemistry can be performed without resorting to “dispos-ables”; also rather than a continuous signal there is a base line avoiding any problems resulting from base-line drift and allowing for easy calibration; and sample pre-treatment can be integrated into the system.For most clinical measure-ments ex vivo techniques in which a single compact portable system can be used to monitor many patients are far more cost effective than in vivo sensors.17 An ex vivo analyser such as ours can also be connected to an invasive conduit serving as a single-patient dedicated instrument. From the viewpoint of process chemistries the incorpora-tion of the fibre allows sampling to be remote from the instrumentation without incorporating a long process time or dispersion of the signal.The solutions have been separated from the expensive instrumentation; a plastic fibre bundle and PVC microconduit are inexpensive and easily replaced. The solid support or “active surface,” implemented in our system is the ColorpHast (non-bleeding) indicator strips. These are cellulose fibres on which acid - base indicators have been covalently immobilised. They are Merck products; the different strips available collectively span the pH range 0.0-13.0 and each individual cellulose pad will change colour over a range of 2.0-3.5 pH units. A detection system developed around a pH change is very non-specific. Enzymes have been recognised in analytical chemistry as potential reagents for over 50 years because of their extreme specificity in a complex matrix.J8.l9 Conve-niently many enzymes also have a pH change associated with the reaction that they catalyse. Enzymatic reactions have been coupled to glass electrode detection systems over the last 20 years in many different ways.20-*3 However optical systems have some inherent advantages over electrodes the signal is not subject to electrical interference; the reagent phase doe 866 ANALYST AUGUST 1986 VOL. 111 not have to contact the fibre optic physically (so it is a simple matter to change the reagent phase); an optical sensor can offer cost advantage over electrodes especially if a single spectrophotometer is used with several sensors; the optical system can take advantage of multi-wavelength and temporal information; no “reference” is required; liquid junction potentials and streaming potentials are avoided; the response time is faster especially in low buffer capacity solutions; an optical system is less sensitive to temperature pressure and large background interferences such as a high sodium ion concentration; and sensors can be developed for which electrodes are not available.24 Urease is an extremely specific enzyme reacting to only one substrate urea (one source reacts to a slight extent with hydroxyurea) with the following stoicheiometry (at pH 7.4): urease NH2CONH2 + H20- 2NH3 + CO2 2NH3 + H20 - 2NH4+ + 20H-C02 + H20- HC03- + H+ Urea has many industrial uses but the bulk of urea analyses are performed on human body fluids as adjunct to the investigation and monitoring of renal and hepatic disease or in the crude determination of nitrogen balance in patients fed on elemental diets.Non-enzymatic determinations exist for urea; in fact a large number of clinical laboratories use direct, spectrophotometric techniques.25 However most spectro-photometric techniques suffer from interferences and the preferred most selective procedure involves the use of unpleasant volatile reagents high temperature and stabilisers. Enzymes provide an appealing alternative as the conditions are necessarily mild. The determination of urea in serum has been performed using flow injection analysis with a glass pH electrode.26 In this work urea was determined in water and serum using flow injection analysis and the “optosensing” detection system. Experimental Instrumentation All the experiments were carried out with ii Bifok-Tecator FIAstar 5020 flow injection analyser.Manifolds with inte-grated flow-through detectors and miniaturised injection valves were incorporated into the FIAstar system replacing the original injection valve. The manifold (described below) was made from a black PVC block; the light source was a relatively weak tungsten lamp and ambient light had to be eliminated. Optical communications were made with plastic optical fibres (Du Pont Crofon fibre 1610; Optronics Cam-bridge UK) consisting of 64 individual acrylic strands randomly bifurcated at the remote end. The fibre bundles were protected from ambient illumination by black PVC sleeving. The optical fibres were glued in place at all ends with opaque black epoxy adhesive F156 (Tra-Con Medford, MA USA) rather than a clear fibre-optic adhesive.The black adhesive eliminated the large blank signal due to the source illumination being piped directly to the detector via a clear epoxy - fibre channel. At the remote end the fibres were fixed perpendicular to the flow channel with the black epoxy hence eliminating any problems of alignment and the movement of individual fibres and also minimising any stray light contribu-tion. The fibre epoxy interface was sanded and polished to allow the fibre-optic bundle to serve as a cell window. Reflectance measurements were performed with the Bifok-Tecator FIAstar 5023 scanning spectrophotometer which was modified in the following manner the 40&700 nm blue filter was removed; the flow-through cuvette was replaced by a black plastic block that held the bifurcated ends of the optical fibre bundle one end in front of the internal light source the signal is weaker than the conventionally measured absor-bance the signal of the non-attenuated reference beam was balanced electronically with built-in potentiometers.The reflectance measurements (A,.) were registered auto-matically on the FIAstar 5023 spectrophotometer and digitally displayed and printed on the FIAstar 5020 spectropho-tometer. The results were concurrently fed to a recorder (Radiometer Servograph REC-61 furnished with an REA-112 high-sensitivity interface). Reagents All chemicals used were of analytical-reagent grade; all water was doubly de-ionised. The carrier solution for the buffer pH determinations was 5 x 10-4 M HC1.The carrier solution for urea determinations was a dilute buffer solution containing 1 X lop3 M Tris - HCl in 0.140 M NaCl adjusted to pH 7.0 for the preliminary aqueous urea analyses and pH 7.4 for the serum analyses. The carrier was prepared from a stock buffer solution by SO-fold dilution with 0.140 M NaCl. The stock buffer solution was made by mixing SO ml of 0.1 M Tris (12.114 g 1 - I ) with 4.0 ml of 1.0 M HCl and diluting to 100 ml with water. The urease (Sigma U-2125) contained 71000 p~ units g-l and was dissolved in the carrier solution at a concentration of 60 mg per 100 ml. Sigma unit definition is that one micromolar unit will liberate 1.0 pmol of ammonia from urea per minute at pH 7.0 at 25 “C under Sigma assay conditions.Volumes of 25 pl of this solution were injected per sample corresponding to 1.08 units per analysis. Bovine serum albumin (Sigma A 6003) solutions were prepared by dissolving albumin in carrier adding stock urea diluting to volume and adjusting the pH directly before use. The urea standards were prepared by dilution of a stock urea solution with the carrier. Stock urea solution (100 mM) was prepared by dissolving 0.600 g of urea (Sigma) in 100 ml of de-ionised water. Serum samples were prepared by diluting 1 + 20 with the carrier solution (100 p1 of serum and 2.00 ml of carrier solution) immediately before use. Microconduit The design of the microconduit is illustrated in Fig. 1. A schematic flow system is shown in Fig.2. The black PVC blocks measure 70 x 45 x 10 mm and channel patterns were I Fig. 1. Integrated microconduit for the measurement of pH. consisting of a split loop injection valve. a mixing coil and an optosensor ( L ) . S Sample solution; P and P, tubes leading t o Deristaltic Dumtx; W waste other end aligned with the detector; and a; the reflected . f i ANALYST AUGUST 1986 VOL. 111 I 867 I I I I I -D Fig. 2. Manifold for pH measurement consisting of two peristaltic pumps (P and P?) a timer (T) and the microconduit Shown in Fig. 1 (boxed area). S Sample; C carrier stream; M mixing coil (100 PI); D, source plus detector; and W waste. The flow cell communicates with the light source and detector of the spectrophotometer by means of optical fibres impressed or engraved on the underside of the block.The channels are closed by transparent plastic using a pressure-sensitive polymeric glue and this yields the semicircular channels typical of the integrated microconduits. The chan-nels are accessed from the top of the block by perpendicular holes into which the PVC tubing is glued. The flow cell compartment is 3.2 mm in diameter the same diameter as the fibre-optic bundle it faces. A hole is cut into the transparent plastic and tape 3 mm or more in diameter to accommodate the cellulose pad containing a covalently bound acid - base indicator. This pad is 3.2 mm in diameter cut from the Merck non-bleeding indicator strips (ColorpHast cata-logue number 9583). A diffuse reflector (off-white Formica) is glued behind the cellulose pad to seal the channel and reflect any unreacted source illumination.The circular pads are cut 3.2 mm in diameter and if unrestricted swell to a depth of 1.8 mm; in the manifolds the pads are allowed to swell to a depth of only 0.4 mm (or 0.2 mm see Flow Cell Optimisation). For simple pH determinations the injection valve is used in the conventional mode with a variable sample volume. For urea determinations using soluble urease a split loop injection technique is used.15 This technique allows the sample and reagent to be injected simultaneously using only one injection valve and one pump; both volumes can be varied by changing the length of the sample or the reagent loops. The effect is similar to a merging zones configuration but requires a much simpler system and is potentially more sensitive as the dispersion of the sample and reagent before merging is minimised.In the earlier optosensing paper urea was determined via its enzymatic degradation to ammonia. 15 The ammonia was detected by diffusion across a gas diffusion membrane to an acid - base indicator solution. The indicator was renewed after each measuring cycle. In the present design ammonia is detected directly in the reacting solution which is in contact with the cellulose-immobilised acid - base indicator. The system is therefore simplified and the response is no longer dependent on diffusion across a membrane. This should allow for a faster response as the immobilised indicator response is essentially instantaneous and also a greater sensitivity as 100% of the ammonia now has the potential to be detected.To allow for variability in the choice of dispersion values and mixing times the manifold in this system incorporated a length of knotted microline tubing as a mixing coil. For pH 0.4( d 0.20 0 1 1.25 8.98 .79 8.50 I I -Time Fig. 3. pH response of the optosensor with Merck indicator 9583 to standard buffer solutions (0.1 M) injected into a 5 x lop4 M HCI carrier stream. The decrease in reflected signal measured as increase in absorbance (A,) at 610 nm. Negative deflections observed at leading edge of buffer responses were eliminated on decreasing the dead volume of the flow cell and increasing the amount of light incident on the indicator pad measurements the mixing coil was represented by the shortest possible length of tubing (ca.2.5 cm) in order to minimise the dispersion of the sample. Depending on the buffering capacity of the sample to be analysed dispersion can be a critical parameter. For urea determinations with soluble urease the sample and reagent need time to mix adequately while minimising the dispersion of the product. A 100-pl (50 cm) length of tubing provides adequate mixing (peak distortions due to variations in refrative index were eliminated) and the tight knots in the microline tubing minimise dispersion.27 The flow-rate through the microconduit is 1.2 ml min-1. The sample is pumped at 3.0 ml min-1 and the reagent at 0.8 ml min-1 for 8 s before each injection. Sensor Characterisation Reflectance spectrum of Merck 9583 acid - base indicator To characterise the response of the indicator with respect to wavelength a scan of the “yellow” (acidic) form of the pad was performed from 400 to 700 nm and stored by the spectrophotometer.The “blue” (basic) form was then scanned and the yellow spectrum was subtracted from the blue by the spectrophotometer. The resulting spectrum is a map of the change in response from 400 to 700 nm exhibiting a maximum at 608 nm. The band pass of the spectrophotometer is 19 nm at 700 nm; in all subsequent analyses the colour change was monitored at 610 nm. Merck indicator 9583 response to ApH To determine the magnitude of the indicator response in the pH range 6.5-10.0 and the shape of the A vs.pH curve a series of buffers (100 vl) were injected in duplicate into a carrier of 5 x M HCI (Fig. 3). The buffers were prepare 868 ANALYST AUGUST 1986 VOL. 111 according to the recipes in Tables 10.37 (0.1 M boric acid -NaOH) 10.32 (0.1 M Tris - HC1) and 10.25 (0.1 M phosphate) in Perrin and Dempsey’s book.28 The response is linear with pH (as measured by a Beckman glass pH electrode) in the pH range 7.3-9.0 corresponding to an A of 0.088-0.400 (Fig. 4). Optimally the urea calibration graph will be designed to fall within this linear region. To perform these pH determinations the shortest possible distance between the point of injection and the flow cell was utilised (corresponding to approximately a 5-pl volume) to minimise the dispersion and mixing of the sample with the carrier.The change in response due to these factors is strongly dependent on the buffer capacity of the samples injected so any change due to mixing with the carrier stream must be minimised for sample pH determinations. Determination of pK of bound dye To determine the pK of the bound dye on the cellulose pad it is assumed that the change in the reflected signal is equivalent to an absorbance change. According to the equations derived by Bishop,*y the ratio of the undissociated and dissociated indicator can be determined and used to calculate the pK, of the indicator without knowing the concentration of the indicator. The transition range the transition point and the conditional formation constant for the indicator can then be calculated from the pK and [H+] if applicable.The ratio of the undissociated and dissociated forms of the indicator immobilised on the pad is determined from spectrophoto-metric data using the equation [In-l/[HInl = [Amix - AHInl/[AIn- - Am,,] * * . * (1) where AHIn is the absorbance of the undissociated indicator, AI is the absorbance of the dissociated form of the indicator and Amix is the absorbance of the undissociated and disso-ciated forms. Absorbance values were measured after pump-ing solutions of the appropriate pH through the flow cell until a steady-state signal was reached. The pH values of the solutions used were obtained potentiometrically (using a Radiometer Research pH meter); absorbance values from eight different intermediate pH buffers (pH 7.29-8.25) were obtained and the pK given as calculated from the following equation is an average of the eight values: The pK of the indicator immobilised on ColorpHast 9538 is determined to be 8.15 F 0.12.. . . . . . pK = pH - log[In-]/[HIn] * . (2) Lifetime Study In order to determine the stability of a single indicator pad, 1200 replicate injections were performed two sets of 300 on two consecutive days (Table 1). Buffers (50 pl) of pH 8.0,9.0 and 7.0 were injected through a simple manifold (no split loop) 99 times each. Standard deviations and relative standard 0.40 d 0.20 0 1 I 1 I I I 7.5 8.0 8.5 9.0 PH Fig. 4. Peak height vs. pH linear region (pH 7.29-8.98) deviations were calculated on each set of 99 injections. Before the data were collected each day 99 injections were per-formed to stabilise the instrument (about 40 min).Whereas the average signals showed some large changes (up to 20%) from one run to the next the relative standard deviations for the fourth set of injections were actually slightly better than the first set of injections. This indicates that the pads are stable for at the very least 1400 injections but the system should be calibrated before each set of measurements. Changes in flow-rate due to changes in pump tubing behaviour on ageing, changes in temperature slight changes in pad position and changes in source intensity are some of the possible contribu-tors to the changes in average signal intensity. Results and Discussion Dispersion and Retention Data The initial data on dispersion were collected by two different methods using the 0.4 mm flow cell with an indicator pad 2.9 mm in diameter.To study solution dispersion bromothymol blue (BTB) dye was injected into a borax carrier. ColorpHast indicator number 9590 a pad that remained yellow at the pH of the carrier (9.0) was placed in the flow cell. A 50-1.1 volume of BTB was first injected through the manifold containing a 5-1.11 mixing coil (the design used for buffer pH determina-tions) and then into the same manifold containing a 100-yl mixing coil (used for sample analysis) to determine the effect of the mixing coil on dispersion. The simple injection system was then changed to the split loop configuration with a volume of 50 1.11 in the first loop (the sample loop) and 25 p1 in the second loop (the reagent loop).BTB was injected through the 50-yl loop with the carrier in the 25-y1 loop and then through the 25-1-11 loop with the carrier in the 50-1.11 loop. The change in the dispersion value for a 5O-pl sample in the split loop injection system compared with the simpler injection system indicates an increase in the sample dispersion of 0.056 due to the split configuration and another 0.087 increase can be attributed to the mixing coil (Table 2). The reagent loop shows a higher dispersion (0.60 larger) due to the smaller Table 1. Lifetime data on a single indicator pad Relative Average Standard standard signal deviation deviation O/O 0.242 0.442 0.064 0.214 0.398 0.062 0.221 0.399 0.055 0.196 0.369 0.055 6.03 7.77 2.08 5.80 5.98 2.43 7.62 5.07 1.52 2.92 3.61 1.90 2.5 1.7 3.2 2.7 1.5 3.9 3.5 1.3 2.8 1.5 0.98 3.4 * ( l ) (2) (3) and (4) refer to the four sets of 99 consecutive injections of each pH buffer; sets (1) and (2) were collected during one day and sets (3) and (4) the following day.Table 2. Dispersion data for bromothymol blue Conditions 50 pl BTB 5-pl coil . . . . . . . . . . 50 pl BTB 100-pl coil . . . . . . . . 50 p1 BTB 100-pl coil split loop . . . . 25 p1 BTB 100-pl coil split loop . . . . 50 pl BTB 100-pl coil split loop 3.2-mm pad 25 pl BTB 100-pl coil split loop 3.2-mm pad 50 pl NH3 100-p1 coil split loop 3.2-mrn pad 25 p1 NH3 100-pI coil split loop 3.2-mm pad D . . . . . . 1.15 .. . . . . 1.25 . . . . . . 1.32 . . . . . . 2.11 . . . . . . 1.45 . . . . . . 2.89 . . . . . . 1.22 . . . . . . 2.3 ANALYST AUGUST 1986 VOL. 111 869 volume and the greater distance travelled. The indicator pad was then replaced by a pad 3.2 mm in diameter and the dispersion of the 50- and 25-p1 volumes in the split loop system increased by 0.098 and 0.37 respectively. Retention times, defined as the time elapsed between the sample injection and the peak maximum,30 for the 50- and 25-pl volume BTB are 7.6 and 9.0 s respectively. Similarly dispersion data were collected by the injection of ammonia into a dilute acid carrier ( 5 x 10-4 M HCI) using a Merck indicator pad number 9583, in order to compare the pad response data with BTB solution reflectance data (Fig.5). The retention time data for ammonia were the same within the error of day-to-day measurements. The BTB shows a higher dispersion some distortion and greater peak widths than ammonia probably owing to the larger BTB molecule within the pad taking longer to flush out than the smaller NH4+ ion. However BTB response has the advantage of being independent of the buffer capacity of the injected sample hence simplifying the interpretation of the results. Flow Cell Optimisation The flow cell first designed by RGiiCka and Hansenls involved placing the optical fibres against a window of the same plastic as used in construction of the fibres. This window was glued into the manifold above the flow channel and the indicator pad was placed into the channel with the reflector glued behind.To decrease the path length that the light had to travel thereby increasing the reflectance signal the window was removed for the present studies. The fibres were glued together (and to the PVC manifold) with the same black epoxy adhesive used at the bifurcated end. The fibres were brought to the base of the manifold rather than the top of the flow channel for two reasons the fibres are now accessible for polishing and a flow cell with a much smaller path length can be constructed. The indicator pad is placed directly against the fibre bundle (insert Fig. 1). Steady state T A A, - T Fig. 5. Peaks resulting from the injection of NH into a carrier stream of 5 X 10-4 M HCI through the split loop injection system; peak A is from a 5O-pl sample volume and peak B is from a 25 pI sample volume.Steady-state response to the NH3 (approximately 2 mM) is superimposed near the peaks; arrow indicates time of injection In the first construction of this flow cell design two layers of plastic and adhesive were built up around the pad with the adhesive and the diffuse reflector behind the indicator pad. This resulted in a flow cell approximately 0.4 mm in depth. In an improved design one layer of plastic and adhesive was removed decreasing the flow cell thickness to 0.2 mm. With this modification the average signals of the three buffers showed an increase of 28% (pH 8.0) 21% (pH 9.0) and 25% (pH 7.0) with correspondingly improved relative standard deviations 1.13% (pH 8.0) 1.08% (pH 9.0) and 2.04% (pH 7.0).The speed of response was also improved the sampling rate of the first flow cell was 126 samples h-1 (50 pl of pH 8.0 buffer) compared with 144 samples h-1 for the second flow cell under the same conditions. Not only is the dispersion reduced but the pad also has less room for movement in the flow cell contributing to the improved relative standard deviation for the buffer measurements. In principle reflectance by an opaque layer is independent of its thickness d provided that a minimum d value has been reached.31 This is one of the conditions under which the reflected signal can be equated with an absorbance signal i.e., over a relatively short range of concentrations the Beer -Lambert law will be obeyed. The two thicknesses of the opaque layer investigated in these experiments were 0.4 and 0.2 mm.To determine the efficiency of the reflecting system, four flow cell designs were compared (1) a cell 0.4 mm in depth with one indicator pad; (2) a cell 0.4 mm in depth with two indicator pads; (3) a cell 0.4 mm in depth with one indicator pad and one pad of white filter-paper; and (4) a cell 0.2 mm in depth with one indicator pad. To compare these systems three buffers of pH 7.29 8.25 and 9.25 were injected in triplicate. As was observed above, one pad in a 0.2-mm cell shows approximately a 20% increase in signal for all three buffers (Table 3). However one pad, with filter-paper behind it shows an even larger signal increase (28-44%) with the following two phenomena probably contributing to the increased reflectance signal for a given ApH; the indicator pad is forced closer to the end of the fibre-optic bundle resulting in increased collection efficiency, and the filter-paper is itself a good diffuse reflector increasing the amount of light that is available to interact with the bound dye molecules.The response to two pads (c) in Table 3 showed an increase in reflected signal of 90% for pH 7.29 66% for pH 8.25 and 28% for pH 9.25. While the signals are higher the slope of the calibration graph is actually decreased owing to the non-linear change in response. The reflectance signal as described by the Kubelka - Munk equation is more complex than an absorbance signal; a scattering coefficient is included, which is non-linearly affected by the changes in optical path length.31 When systems ( b ) and (c) were compared using the urea-soluble urease systems no increase in signal was observed for the NH3 produced in contrast to the measure-ment of buffers.It can be concluded that for urea determina-tions using soluble urease one pad is sufficient. For buffer determinations depending on the range of ApH of the samples there may be an advantage to using a second indicator pad. Both systems may show an improved signal using the larger flow cell and incorporating filter-paper as a diffuse reflector but the sampling rate will be decreased. Table 3. Effect of flow cell design on reflected signal 4 0.4 mm (d), PH one pad one pad two pads filter-paper ?.29 0.132 0.113 0.214 0.155 8.25 0.395 0.322 0.536 0.465 9.25 0.678 0.589 0.756 0.766 0.2 mm (a) 0.4 mm (b) 0.4 mm ( c ) one pad 870 Optimisation of carrier p H In order to determine the optimum carrier pH the following must be considered the enzyme activity the indicator region of colour change the sample pH and the sensor response to NH3.The indicator has been demonstrated to respond from pH 6.5 to 10.0. The carrier is 1 x 10-3 M Tris 140 mM NaCl; the pH was adjusted to 6.7,7.0,7.3,7.5 and 7.7 with 0.1 M HCI and 0.1 M NaOH. A 2 mM NH3 solution was prepared by dilution of concentrated ammonia solution with de-ionised water and 50 pl were injected into each carrier. The maximum response was observed at pH 6.7 with a 2% decrease at pH 7.0 a 10% decrease at pH 7.3 a 19% decrease at pH 7.5 and a 22% decrease at pH 7.7.NH3 equilibrium favours a larger signal with increasing pH but the higher the carrier pH the greater the background colour of the pad so the smaller the response to the pH change with the injection of Enzyme solutions were prepared from the different carriers by 1 + 9 dilution of a stock enzyme solution (60 mg of enzyme in 10 ml of de-ionised water). Calibration graphs of 0-1 mM urea were generated (Fig. 6) to establish the pH of maximum apparent enzyme activity. At pH 6.7 and 7.0 the enzyme behaviour is essentially the same. At pH 7.3 however the apparent enzyme response to urea is suppressed by an average of 11% (relative to the response at pH 6.7) which reflects the behaviour of the NH3 at this pH. At pH 7.5 and 7.7 however, enzyme behaviour is clearly inhibited (43% suppression of signal at pH 7.5 48% suppression at pH 7.7) compared with the 19 and 22% for suppression of the signal of injected NI€3 solutions.The indicator is non-linear below pH 7.3. Hence in order to maximise the enzyme response to urea and also stay near to the region of linear indicator response the carrier pH used in the aqueous urea experiments was 7.0. NH3. Optimisation of ionic strength In order to maximise the enzyme activity 1 x 10-3 M Tris carrier streams at pH 7.0 with 0.00 17.5 35 70 140,210 and 280 mM NaCl were compared €or urea analysis. Calibration graphs of &2 mM urea were established in each carrier stream (Fig. 7). An NaCl concentration of 140 mM the concentration of NaCl in serum resulted in the steepest line below 1 mM urea.Large changes in the background A result when the ionic strength of NaCl in the carrier is changed (Table 4). This is probably due to some extent to the change in refractive index of the solution in combination with an indicator salt error.29 In conclusion the NaCl concentration should be the same for the sample reagent and carrier. Na+ is reported to 0.5 Concentration of ureairnM 1 .o Fig. 6 . Response graphs for urea (50 PI) at 0.0-1.0 mM with urcase (25 pl 60 mg dl-‘) as a function of pH lines A B. C D and E represent pH 6.7,7.0,7.3,7.5 and 7.7. respectively for the carrier and sample; solutions were 140 mM in NaCl ANALYST AUGUST 1986 VOL. 111 have an inhibitory effect on the urease,-32 but 140 mM NaCl resulted in the fastest rate under the conditions of this experiment.Also to avoid a blank response to differences in the sample and carrier ionic strength the carrier stream and reagent stream should be adjusted to the same ionic strength as the samples being injected. Conditions for Urea Determination Using Soluble Urease Stopped flow The stopped flow approach was selected for enzymatic measurements because of the following advantages noise caused by any vibration of the cellulose pad by the flowing stream would be eliminated; the sample and reagent could be allowed a longer time to react without increasing the dispersion with a long length of mixing c o i P ; a kinetic measurement could be performed as the FIAstar has the option to measure only the change in absorbance generated during the stop time-this will eliminate the need to know the background pH of the system and so the problems due to the variable sample pH will be minimised.The duration of the stop time was somewhat variable (10-30 s) as the system was not thermostated. At the typical Seattle summertime temper-atures of 23-28 “C the stop time used was 20 s. The actual time that the “stop” was initiated had to be calibrated for a given run; it depended primarily on the age of the peristaltic pump tubing and varied from 6 to 8 s for a flow-rate of 1.2 ml min-1. By viewing the output of the recorder the position with respect to maximum NH4+ signal was easily determined. I€ the flow was stopped prior to the position of the signal maximum then when the flow resumed a sharp “spike” characterised the remainder of the signal.The relative magnitude of this spike provides information about how far from the signal maximum the stop position is; a peak that simply trails off when the flow is resumed indicates that the stop is past the largest signal. ~~ ~ ~ Table 4. Effect of ionic strength on background reflected signal [ NaCl]/mM A‘% 280 0.064 210 0.055 140 0.000 70 -0.016 35 -0.037 17.5 - 0.060 0 -0.164 0.3 qL 0.15 0 I 1 1 .o 2.0 Concentration of ureairnw Fig. 7. Response graphs for 0.0-2.0 mM urea with soluble urease (25 PI. 60 mg dl-1) at FH 7.0 as function of [NaCI] in carrier stream and sample solutions. [NaCI] for lines A. B C. D E. Fand G are 0 17,35. 70 140 210 and 280 mM.respectively. The line for 280 r n M is not shown; it shows the same response as 210 mM NaCI ANALYST AUGUST 1986 VOL. 111 87 1 The flow-rate was chosen simply as the fastest that the manifold could accommodate without leaking in order to maximise the turbulence of mixing the speed of product washout and the amount of product formed during the stop time rather than before stopping. At different positions along the sample - reagent plug, different amounts of NH4+ are produced hence unique calibration graphs are generated at each stop time (Fig. 8). For this work the region of the maximum response to low levels of urea was chosen; for other types of analyses for example the determination of urea in urine the leading edge of the plug, where the signal shows a linear response from 2 to 4 mM urea, may be more ideal.Enzyme concentration A series of calibration graphs were generated with 50 pl of &2 mM urea samples and 25 pl of 15,30,45,60 and 75 mg urease per 100 ml of 1 x 10-3 M Tris (pH 6.5) as reagents (Fig. 9). At concentrations of urease above 45 mg dl-1 there is no increase in the signal for 2 mM urea but the signal did increase for 0-1 mM urea. As this is the region of interest a concentration of 60 mg dl-1 was chosen for subsequent analyses. With such a small reagent volume this corresponds to only one unit of enzyme per sample. Sample size The choice of sample size is governed by the desire for a maximum signal for a given sample volume and the preferred use of the smallest volume of serum possible.As 25 pl of enzyme solution are used two volumes of sample were tried, 25 and 50 pl. Although twice the size of the smaller sample the larger volume sample increases the dilution of the enzyme, and the enzyme activity is non-linear with concentration. The actual gain in the signal is 20%. As 50 pl of injected sample correspond to only 2.5 p1 of serum 50 p1 of sample were used for this study. Entrapped (adsorbed) enzyme Once the cellulose pad with covalently bound indicator has been exposed to the enzyme in solution the pad will exhibit enzyme catalysis in the absence of any injected enzyme solution for some time the duration depending on how long the pad has been exposed to the enzyme [Fig. lO(a)]. The magnitude of the response will resemble the solution enzyme response and the pad enzyme yields a calibration graph of the same shape as soluble urease.Urea determinations using entrapped enzymes show reliability at least as good as when using soluble urease (Table 5 ) . To characterise this behaviour the responses of three pads were compared (a) a cellulose pad soaked overnight in enzyme solution; (b) a cellulose pad soaked in a vigorously stirred enzyme solution for 1 h; and (c) a cellulose pad placed in the microconduit flow cell with the enzyme solution flowing through the flow cell for 10 min. The pad soaked in the enzyme solution overnight showed negligible enzyme activity in the presence of urea. The pad stirred with the enzyme solution showed a small but stable response and the pad placed in the flow cell showed a signal roughly equivalent to the response obtained when enzyme (60 mg dl-1) is injected with urea [Fig.lO(b)]. These results indicate that the phenomenon is flow-related which supports an “entrapment” hypothesis. The response of the entrapped enzyme to changes in NaCl concentration and changes in pH is very similar to the behaviour of the soluble enzyme. The apparent activities of both the soluble enzyme and the entrapped enzyme diminish slightly if the original enzyme solution is filtered or centri-fuged verifying that the pad is not simply trapping “colloidal junk” from the enzyme solution. Also this entrapment phenomenon is not dependent on the enzyme source; urease obtained from Millipore (51 U mg-1) exhibits similar behav-iour to urease from Sigma.The use of the pad as an entrapment bed for the enzyme could be exploited. It would eliminate the need for the split loop injection system hence simplifying the injection system. It would allow faster sampling rates and smaller sample volumes to be used as no mixing with the reagent would be necessary and no dilution of the sample by the reagent or dispersion in the mixing coil would result. In this system two Table 5. Recovery of urea in aqueous solutions Urea Urea taken/mM f o u n d h 0.0925 0.088 0.375 0.350 0.0315 0.025 0.0925 0.085* 0.187 0.184* 0.375 0.387* * Obtained using adsorbed urea. Recovery, Yo 95 93 89 92 98 103 0.10 4 0.05 I 1 I 1 I 0 1 .o 2.0 3.0 4.0 Concentration of urea/mM Fig. 8.Response raphs for 0-4.0 mM urea with soluble urease (25 @,60 mg per 100 mlf pH 7.0 at different stop points (times) along the sample - reagent plug line A is obtained by stopping 5 s after injection of sample and reagent via the split loop injection system; B is 6 s and C is 7 s after injection 0.40 1 0 0.5 1 .o 1.5 2.1 Concentration of ureairnw Fig. 9. Response graphs for urea at 0 . s 2 . 0 mM as a function of soluble enzyme concentration lines A B C D and E represent 15, 30 45 60 and 75 mg dl-1 respectively. The pH of the carrier and sample is 7.0 and the NaCl concentration is 140 m 872 0.30 d 0.15 ANALYST AUGUST 1986 VOL. 111 -4 min --. B -0. k 0.05 I I I I 0 25 50 75 No. of injections ( b ) 0.15 I No. of injections Fig.10. (a) response of adsorbed urease to 2 mM urea at pH 7.0, atter flowing the enzyme through the flow cell for 5 min (A) and 15 min (B). Each point is the average of five injections. ( b ) Response of adsorbed urease to 2 mM urea after soaking the pad overnight in enzyme (A) stirring with enzyme for 1 h (B) and flowing enzyme through flow cell for 10 min (C) pads should double the amount of enzyme entrapped hence increasing the lifetime of the entrapped signal. Research in the field of solid-phase reagent chemistry has demonstrated that solid phases particularly cellulose have a stabilising effect on labile reagents. The storage stability of some biochemical systems has been prolonged for up to 2 years at room temperature.33 These are fibre-impregnated systems; the enzyme we are accumulating on the active surface placed in the flow cell is an equivalent system.It is probable that this system could be used as a system for “loading” enzymes on to the cellulose pads and that these pads could be used immediately or stored for later analysis. Each pad could be calibrated and then used to perform a limited number of assays; this would be convenient for a laboratory performing a relatively small number of enzymatic determinations on a routine basis. Loading could possibly be increased or the system stability improved by first activating the cellulose with glutaral-dehyde borrowing an established technology from membrane electrode te~hnology.3~ This will initiate cross-linking of the enzyme with the solid support.If the adsorbed enzyme is not to be exploited it is important to realise that it will still be present if the pad has been exposed to enzyme. No “blank” injections of sample without enzyme would be reliable. This enzyme catalysis has been observed at any feasible base-line pH (6.5-7.6) for the indicator. Aqueous urea determination Urea standards and samples were prepared by the dilution of a 100 mM urea stock solution with a carrier of 1 x 10-3 M Tris -HC1 (pH 7.0) and 140 mM in NaCl. These solutions (50 pl) were injected with the enzyme reagent (25 pl) at 60 mg enzyme per 100 ml of carrier (Fig. 11). The response to the standards was plotted vs. urea concentration and the concentrations of the “samples” were calculated from this calibration graph. Recovery of the expected sample concentrations ranged from 89 to 103%.This is fairly good as the graph is non-linear above 0.125 mM urea under these conditions and the recovery data were collected directly from all regions of the graph. O L -r Fig. 11. Response graphs for urea determination at pH 7.0. Replicate injections of 2.00 1.00 0.500 0.250 0.125 0.063 0.031 and 0.015 mM urea with soluble urease (60 mg dl-I). The samples following were injected in triplicate except the first sample which was injected five times to obtain a relative standard deviation of 1.56% 0.40 aJ c (D e 0.20 a v) 0 0.5 1 .o 1.5 2.0 Concentration of ureaimM Fig. 12. Response graphs for 0-2.0 mM urea with soluble urease (25 p1,60 mg dl-I) as a function of [HC03- 1; HC03- is at a concentration of A 0.00; B 1.25; C 2.50; and D 5.00 mM in carrier stream and sample.The NaCl concentration is 140 mM and the pH is 7.0 Serum Analysis Effect of hydrogen carbonate The hydrogen carbonate buffer system supplies approximately 95% of the buffering capacity of serum.35 A 20-fold dilution of serum would result in a sample hydrogen carbonate concen-tration of 1 mM. To determine the effect of hydrogen carbonate on the calibration graph Tris carrier streams 0.00, 1.25 2.50 and 5.00 mM in NaHC03 all 140 mM in NaCl (pH 6.5) were used for the analysis. Suppression of the signal at the low end of the calibration graph (normal serum) was slight for 1.25 and 2.50 mM hydrogen carbonate but all three graphs containing hydrogen carbonate established a lower signal for 2 mM urea essentially a new steady-state value for the enzyme (Fig.12). Without hydrogen carbonate in the standard solutions the signals due to serum urea would be lowered for high concentrations of urea. Hence calibration standards should be 1 mM in NaHC03 for serum analysis. Effect of albumin Human serum is approximately 7 g dl-1 in proteins. Although the contribution of albumin to the solution buffering capacity is sma11,36 the effect on the sample viscosity and refractiv ANALYST AUGUST 1986 VOL. 11 I 873 index may be significant and an indicator protein error is another possible source of deviation from the aqueous calibration graph.29 In order to determine the effect of this protein on the calibration graph for urea analysis urea standards of 0.0-2.0 mM were made up of 0.35 g dl-1 in bovine serum albumin 1.0 mM NaHC03 diluted to volume with Tris carrier the pH being adjusted to 7.4.These standards were compared with urea standards with no albumin 1.0 mM NaHC03 and diluted to volume with Tris carrier (pH 7.4). The two resulting calibration graphs differed significantly; the presence of albumin increased the signal resulting from the enzyme reaction. Both calibration graphs have the same zero intercept indicating that this is not a blank response to the albumin. The calibration graph is linear to 2.0 mM urea (correlation coefficient 0.9996); usable data can be obtained to 4 m M urea encompassing any probable serum value. Serum samples Serum samples and glycerol-based BUN standards (Monitrol I and 11) were obtained from the University Hospital Seattle, WA.Hospital urea measurements had been performed on an ASTRA system the previous day and the results were compared with FIA data (Table 6). Agreement is excellent Table 6. Serum values Astra FIA optosensing Urea/mM UreairnM Reported (diluted (diluted BUN sample) A sample) 81 25 41 36 23 5 43 10 13 51 50t 1.38 0.43 0.69 0.61 0.39 0.085 0.75 0.17 0.22 0.085 0.85 0.088 0.026 0.045 0.040 0.026 0.003* 0.048 0.008 0.015 0.003* 0.056 1.44 0.43 0.74 0.66 0.43 0.059 0.79 0.14 0.25 0.059 0.92 * Net signal less than three times the standard deviation of the ?- Monitrol standards. blank. 1 .o a a K I-v) 0 1 .o A, Fig.13. Comparison of results for serum urea FIA optosensing results vs. University Hospital ASTRA results (enzymatic determina-tion with conductimetric detection). The precision of both methods is 3 % when the albumin-containing calibration graph is used to calculate the urea concentrations of samples (Fig. 13). Our results tend to be a few per cent. high probably owing to the elevated pH of the serum samples resulting from bacterial degradation. This will be measurable a few hours after blood drawing; our undiluted serum sample pH values were as high as 8.0 at the time of measurement. The authors thank Tecator Inc. for the loan of the FIAstar system and Winnie Lee University Hospital for graciously providing serum samples. 1. 2. 3.4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. References Narayanaswamy R . Anal. Proc 1985 22 204. Coleman J. T. Eastham J. F. and Sepaniak M. J. Anal. Chem. 1984 56 2246. Smith A. M. Anal. Proc. 1985 22 212. Saari L. A and Seitz W. R. Anal. Chem. 1982 54 821. Saari L. A. and Seitz W. R. Anal. Chem. 1983 55 667. Saari L. A and Seitz W. R. Analyst 1984 109 655. Goldfinch M. L. and Lowe C. R. Anal. Riochem. 1980, 109 216. Goldfinch M. J . and Lowe C. R. Anal. Riochem. 1984,138, 430. Edmonds T. E . and Ross I. D. Anal. Proc. 1985 22 206. Peterson J . I . Goldstein S. R . and Fitzgerald R. V. Anal. Chem. 1980 52 864.Kirkbright G . F. Narayanaswamy R. and Welti N. A . , Analyst 1984 109 1025. Peterson J . I. Fitzgerald R. V. and Buckhold D. K. Anal. Chern. 1984 56 62. Schultz J. S . Mansouri S. and Goldstein I. J. Diabetes Care 1982 5 245. Kirkbright G . F. Narayanaswamy R. and Welti N . A., Analyst 1984 109 15. RGiiEka J . and Hansen E. H. Anal. Chim. Acta 1985 173, 3. RfiiiCka J. Anal. Chem. 1983 55 1040A. McKinley B. A . Houtchens B. A . and Janata J . Zon-Sel. Electrode Rev. 1984 6 173. Marconi W. Bartoli F. Gulinelli S. and Morisi F. Process Biochem. 1974 May 22. Fishman M. M. and Schiff H. F. Anal. Chem. 1972 44, 543R. Guilbault G. G . and Montalvo J. G . Anal. Lett. 1969 2, 283. Johansson G . and Ogren L. Anal. Chim. Acta 1976,84,23. Ripamonti M. Mosca A Rovida E . Luzzana M. Luzi L., Ceriotti F. Cottini F. and Rossi-Bernardi L. Clin. Chem., 1984 30 556. Joseph J. P. Mikrochim. Acta 1984 11 473. Seitz W. R. Anal. Chem. 1984 56 16A. Butler A . R. Trends Anal. Chem. 1982 1 120. Rfiiii-ka J. Hansen E. H. Ghose A. K. and Mottola H. A., Anal. Chem. 1979 51 199. Engelhard H . Neu U. D. Chromatographia 1982 15 403. Perrin D. D. and Dempsey B. “Buffers for pH and Metal Ion Control,” Chapman and Hall London 1974. Bishop E. “Indicators,” International Series of Monographs on Analytical Chemistry Volume 51 Pergamon Press Oxford, New York 1972. Rgiitka J . and Hansen E . “Flow Injection Analysis,” Chemistry and its Applications Volume 62 Wiley New York, 1981. Kubelka P. and Munk F. Z. Tech. Phys. 1931 12 593. Guilbault G . “Handbook of Enzymatic Methods of Analysis, Clinical and Biochemical Analysis,” Volume 4 Marcel Dek-ker New York and Basle 1976. Rocks B. and Sherwood. R. Talanta 1984,31 879. Tran-Minh C. and Broun G. Anal. Chem. 1975 47 1359. Szabo L. L. and Kaplan A. “Clinical Chemistry Interpreta-tions and Techniques.” Third Edition Lea and Febiger, Philadelphia 1983. Paper A6148 Received February 17th 1986 Accepted March 13th 198

ISSN:0003-2654    

DOI:10.1039/AN9861100865

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, AUGUST 1986, VOL. 111 875 Urea pH Electrodes: Characterisation and Optimisation for Plasma Measurements Pankaj Vadgama Department of Clinical Biochemistry and Metabolic Medicine, The Medical School, Newcastle upon Tyne NE2 4HH, UK A potentiometric urea enzyme electrode is reported where consumption of H+ by the urease-catalysed degradation of urea is followed at a glass pH electrode. The response of the electrode in both weak buffers and diluted plasma is described, and the special conditions needed for plasma assay are outlined. Response times were 1-3.5 min. The use of 0.002 M phosphate assay buffer gave linear calibration graphs up t o 5 x 10-4 M urea. The response to plasma diluted 1 + 49 or 1 + 999 initially proved unreliable, but corresponded to that in aqueous standards when allowance was made for background pH variation using an auxiliary pH electrode.Keywords: Enzyme electrode; urea determination; plasma assay; pH buffer capacity; pH electrode Enzyme electrodes based on pH sensors are, in principle, an attractive proposition for the analysis of biochemical species. 1 Most enzyme reactions involve the direct consumption or generation of protons, and the pH sensor remains one of the most sensitive and selective of the commercially available ion-selective electrodes. However, relatively few reports have been published on the practical application to biological fluids of enzyme electrodes based on pH measurement. Nilsson et aZ.2 have described the use of a penicillin pH electrode in fermentation broth, and we have reported on a urea pH electrode for plasma.3 The problems encountered with the use of such electrodes in biological fluids are primarily those relating to variable sample pH and buffer capacity.4 Here we report on the performance of a urea pM electrode employing urease to catalyse the hydrolysis reaction Urea + 2H20 + H+ += HC03- + 2NH4+ and the subsequent optimisation of this system for plasma urea measurements.Several groups have reported on urea pH electrodes,I,5-7 but the operational conditions used require modification before plasma assay can be contemplated. Experimental Apparatus and Reagents A Corning (4702200) Triple-Purpose pH electrode and a Beckman (541263) pH electrode were used against a calomel reference electrode (Type CRR; Russell pH, Auchtermuchty , Fife, UK).An electrically shielded junction box was used to enable alternate measurements on the two pH electrodes against a single calomel reference to be made. Electrode input was to a Radiometer PHM 72b Digital Acid - Base Analyser (Radiometer, Copenhagen, Denmark) with the signal output recorded on a Servoscribe recorder (Type 511.20; Smith Industries, London, UK). AnalaR-grade reagents (BDH Chemicals, Poole, Dorset, UK) were used to prepare buffer solutions, acrylamide gels and urea stock standard solutions (0.1 M). Jackbean urease (E.C. 3.5.1.5, type 111, Sigma, Poole, Dorset, UK) used for enzyme electrodes had an activity of 2.8 U mg-1. Quality control serum was obtained from Wellcome Diagnostics (Dartford, UK). Enzyme Electrode Assembly and Use Acrylamide gelling solution was prepared by dissolving N,N'-methylenebisacrylamide (1.15 g), acrylamide monomer (6.60 g), riboflavin (5.5 mg) and potassium persulphate (5.5 mg) in distilled water.8 The required amount of urease was dissolved in the gelling solution and the mixture purged with N2 until a soft translucent gel had formed. The enzyme gel was then applied to the tip of the Corning glass electrode over which a 60 pm thick nylon mesh had been mounted as a spacer.The gel was then covered with a Cuprophan dialysis mem- brane (Enka, Wuppertal, FRG) to furnish a thin, uniform coating over the pH sensor. Measurements were made in magnetically stirred 10-ml volumes of assay buffers at room temperature (21 5 2 "C). After a stable base-line potential had been obtained, urea standards were added by means of automatic pipettes and the steady-state readings recorded. Electrode responses were recorded as the total change in pH (ApH) registered by the enzyme electrode, rather than e.m.f.change. In some instances, the non-enzymic component of ApH due to changes in solution pH was subtracted from the response using the response of an auxiliary pH electrode dipped in the same solution. Plasma assays on samples stored in lithium heparin tubes were compared with measurements made by a standard spectrophotometric method based on diacetyl monoxime.9 Results Response times ( t 98%) showed considerable variation between electrodes [Fig. l(a)]; they were generally in the range 1-3.5 min, but increased significantly with enzyme loadings above 100 U ml-1 of urease.The magnitude of the response to urea, however, increased with enzyme loading only up to 40 U ml-1 of urease [Fig. 1 ( b ) ] , where it reached a plateau value which was independent of the amount of enzyme used. The buffer capacity of the assay solutions strongly affected the calibration graphs, which shifted to lower urea concentration ranges as the buffer strength was lowered (Fig. 2). The calibration graphs had linear portions with super- Nernstian slopes (90 mV decade-'), but over the concentra- tion range for urea likely to be found in the plasma assay solutions (10-4 to 10-3 M), responses were both small and non-linear. Curvature over the lower concentration range was reduced by resorting to a linear concentration range for urea (Fig.3). It was also possible to modify the shape of calibration graphs by altering the assay pH, or by including altern- ative buffers such as N-tris( hydroxyniethy1)methyl-2- aminoethanesulphonic acid (TES, pK, 7.5), or tris(hydroxy- methy1)methylamine (Tris) (pK,, 8.2) (Fig. 4). However, the key factor which enabled adequate linearity to be achieved was the use of a restricted concentration range for urea (0-4 X M).876 5 4 - (a) 0 I I I 1 I , 20 40 60 80 100 120 140 5 0.4 0.2 a 0.1 Om3/ 20 40 60 80 100 120 140 Urease concentration/U ml-1 Fig. 1. Effect of urease concentration in the enzyme layer on (a) the response time (t98%) and (b) the magnitude of response to 2 X M urea. Measurements were made in 0.001 M phosphate buffer (pH 6.86). A different enzyme gel was used for each point 2.0 I I -4 -3 -2 Log[urea/mol 1-11 Fig.2. Urea calibration graphs in phosphate buffer (pH 6.86) of varying concentrations: 1, 0,001; 2, 0.005; 3, 0.01; and 4, 0.025 M. Urease concentration. 84 U ml-1 1.5 I Q. 1.0 0.5 4 8 12 16 20 [Ureal/10-4 M Fig. 3. Urea calibration gra hs in phosphate buffer (pH 6.86) of varying concentrations: 1 , 0.011; 2, 0.002; 3, 0.004; 4, 0.006; and 5 , 0.01 M. Urease concentration, 69.5 U ml-1 The results for plasma assayed in weak buffer solutions did not match the expected values based on the spectrophoto- metric assay [Fig. 5(a)], and agreement remained poor when quality control serum was substituted for aqueous calibration standards [Fig. 5(b)]. Addition of plasma at 1 + 49 dilution to the weak assay buffers produced an upward shift in solution pH [Fig. 6(a)], and even at 1 + 999 dilution [Fig.6(b)], ANALYST, AUGUST 1986, VOL. 111 1.4 1.2 1 .o 5 Oe8 0.6 0.4 0.2 2 4 6 [Ureal/lO-4 M Fig. 4. Urea calibration graphs in (1) 0.002 M phos hate buffer (pH 6.4), (2) 0.001 M phosphate + 0.001 M Tris (pH 6.867 and (3) 0.001 M phosphate + 0.001 M TES (pH 6.86). Urease concentration, 69.5 U ml-1 1.5 ( a ) 1 0.2 0.4 0.6 I 0 Urea/rnrnol I-' Fig. 5. (a) Electrode response to (0) plasma at 1 + 99 dilution in 0.002 M phosphate buffer (pH 6.86) compared with (0) aqueous standards. (b) Electrode response to (0) plasma at 1 + 99 dilution in 0.001 M phosphate buffer (pH 6.86) compared with (0) Wellcontrol assayed sera. Urease concentration, 69.5 U ml- readings for plasma demonstrated a significant zero error.As indicated by the non-zero intercepts in Fig. 6, the average change in bulk solution pH due to plasma added to 0.002 M phosphate buffer (pH 6.86) was +0.04 pH at 1 + 49 dilution and +0.01 pH at 1 + 999 dilution. There was no advantage to the use of 1 + 999 dilution; the error arising from the pH change amounted to 31% of the response to 10 mM urea compared with 30% at 1 + 49 dilution, and the smaller responses led to significant imprecision in the measurements [Fig. 6(b)]. When measurements at an enzyme electrode were made against an auxiliary pH electrode to allow for pH changes in the bulk solution, both the scatter and the zero error were eliminated (Fig. 7). The agreement between responses in plasma and aqueous solutions indicates that at the dilution used (1 + 99), the bulk pH change due to plasma does not perturb the assay buffer capacity, and the intrinsic plasma buffers have no effect.ANALYST, AUGUST 1986, VOL.111 877 0.4 1 0.3 0.2 0.1 I 1 1 0.2 0.4 0.6 Urea/mmol I-’ Ip a (6) 0.08 L I I I Urea/pmol I-’ 4 8 12 16 Fig. 6. ( a ) Electrode response to plasma diluted 1 + 49 in 0.002 M phosphate buffer ( H 6.86); urease concentration, 69.5 U ml-I. (b) Conditions as in (UP using plasma diluted 1 + 999 r a 0.4 0.2 0.5 1 .o 1.5 2.0 [Ureal/lO-4 M Fig. 7. Electrode response to (@) 1 + 99 diluted plasma compared with (0) aqueous standards. pH’ is enzyme electrode response minus auxiliary electrode response; difference in electrode output in base-line buffer was 0.2 pH. 0.002 M phosphate buffer (pH 6.86) was used; urease concentration, 69.5 U ml-* Discussion Enzyme layer activity is a key determinant of substrate sensitivity in enzyme electrodes.10 The zero-order depen- dence in enzyme layer activity observed here [Fig. l(b)] and commonly seen with immobilised enzyme systems attests to the importance of substrate diffusion as the rate-limiting process for catalysis; a similar relationship between response and enzyme loading has been found for the urea NH4+ electrode.11 Higher enzyme loading would be expected to accelerate the transient electrode responsel2; the tendency for responses to be slower [Fig. l ( a ) ] may have been due to the additional diffusion barrier posed by the enzyme protein at higher loading. H+ diffusion through the enzyme layer is likely to have been particularly strongly retarded; electrode response times with NH4+ as the detected species have been much shorter (t 98%, 25-60 s) despite the use of thick enzyme layers.” Studies of H+ diffusion through a glycoprotein matrix indicate that proton diffusion may be more retarded than the diffusion of other monovalent cations through protein-containing layers.13 The enhancement in sensitivity at low buffer capacity has been demonstrated by previous workers.6.7 The sigmoid nature of the calibration graphs (Figs.3 and 4) may, in part, have been due to the approach towards optimum pH for urease in the enzyme layer with increasing rate, followed by reduced catalytic activity at higher pH values. Thus, responses commenced on the acid side (pH 6.86) of the pH activity profile for urease, optimum pH 7.3,14 with eventual approach to the critically low enzyme activity as the pH increased further at high urea concentration.Also important would have been a reduction in net proton comsumption at high pH; the urease reaction actually liberates protons at pH > 9 as a result of NH4+ and HC03- dissociation,l5 and before the dissocia- tion, the net consumption of protons diminishes towards pH 9. Buffer pK, also influences curvature; increased buffering at pH =z pK, would diminish the response. With the appropriate choice of buffers (Fig. 4) to counterbalance changes in pH, and therefore buffer capacity in the enzyme layer, it may be possible to extend the linearity beyond the range observed in this study.The key to reliable measurements on plasma (Fig. 7) was the use of high sample dilution. This ensured that (i) there was minimal perturbation of buffer pH, and therefore of the buffer capacity of the assay buffer, (ii) sample buffer capacity had a negligible effect on solutions and (iii) narrow, linear segments of calibration graphs could be utilised to cover the full clinical range of plasma urea concentration. The additional use of an uncoated pH electrode enabled the pH rise in the assay buffer induced by venous plasma samples to be corrected (pH reference range 7.38-7.4).16 The adapta- tion of a urea pH electrode to plasma measurement demon- strates how controlled assay conditions may permit the analysis of s mples of unknown pH and buffer capacity. The urea pH i ectrode can therefore take its place alongside urease-coated NH3,17 NH4+18 and C0219 sensors as an acceptable tool for urea measurement.It should be possible to extend such analysis to other substrates where enzymic degradation is associated with pH change. The author gratefully acknowledges support from the MRC. References 1. Nilsson, N., Akerlund, A., and Mosbach, K., Biochim. Biophys. Acra, 1973, 320,529. 2. Nilsson, N., Mosbach, K., Enfors, S.-O., and Molin, N., Biotechnol. Bioeng., 1978, 20, 527. 3. Vadgama, P. M., Covington, A. K. , and Alberti, K. G. M. M. , Anal. Chim. Acta, 1982, 136, 403. 4. Bucher, T., Hofner, H., and Rouayrenc, J. F., in Bergmeyer, H. U., Editor, “Methods of Enzymatic Analysis,” Volume 1, Verlag Chemie, Weinheim, 1974, p. 258. Kobos, R. K., in Freiser, H., Editor, “Ion-selective Electrodes in Analytical Chemistry,” Plenum Press, New York, 1980, p. 31. Alexander, P. W., and Joseph, J. P., Anal. Chim. Acta, 1981, 131, 103. Szuminsky, N., Chen, A. K., and Liu, C. C., Biotechnol. Bioeng., 1984, 26, 642. Guilbault, G. G., in Weetall, H. H., Editor, “Immobilized Enzymes, Antigens, Antibodies and Peptides,” Marcel Dek- ker, New York, 1975, p. 293. Marsh, W. H., Fingerhut, B., and Miller, H., Clin. Chem., 1975, 11, 624. Blaedel, W. I., Kissell, T. R., and Boguslaski, R. C., Anal. Chem., 1972,44,2030. Guilbault, G. G., and Montalvo, J. G., J. Am. Chem. SOC., 1970, 92, 2533. Carr, P. W., Anal. Chem., 1977, 49, 799. Vadgama, P., and Alberti, K. G. M. M., Experientia, 1983,39, 573. 5. 6. 7. 8. 9. 10. 11. 12. 13.878 ANALYST, AUGUST 1986, VOL. 111 14. Fishbein, W. N., J. Biol. Chem., 1969, 244, 1188. 15. Adams, R. E., and Carr, P. W., Anal. Chem., 1978, 50, 944. 16. Gambino, S. R . , Am. J . Clin. Pathol., 1959, 32, 294. 17. Mascini, M., and Guilbault, G. G . , Anal. Chem., 1977,49,795. 18. Guilbault, G . G., Nagy, G., and Kuan. S. S., Anal. Chim. Acta, 1973,67, 195. 19. Guilbault, G. G., and Shu, F. R.. Anal. Chem., 1972,44,2161. Paper A6137 Received February I Oth, I986 Accepted February 26th, I986

ISSN:0003-2654    

DOI:10.1039/AN9861100875

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, AUGUST 1986, VOL. 111 879 Application of the Reductive Flow Injection Amperometric Determination of Iodine at a Glassy Carbon Electrode to the lodimetric Determination of Hypochlorite and Hydrogen Peroxide Antoine Y. Chamsi and Arnold G. Fogg Chemistry Department, Loughborough University of Technology, Loughborough, Leicestershire LEI1 3TU, UK Dissolved molecular oxygen is insignificantly reduced at a glassy carbon electrode that is being held at -0.2 V vs. SCE for the determination of iodine. The tendency for molecular oxygen to oxidise iodide to iodine in acidic solutions, however, is a serious problem in the development of flow injection amperometric methods based on the use of iodine reactions in which iodine is monitored at -0.2 V. In the work described here, iodine (I2), has been determined at concentrations above 6 x 10-6 M in neutral, unsparged potassium iodide solution by injection through PTFE transmission tubing into a nitrogen-sparged 0.4% potassium iodide eluent in a flow injection system.By sparging the sample solution with nitrogen for 30 s before injection and shielding the PTFE transmission tubing with nitrogen, it was possible to determine iodine at concentrations above 1 X 10-7 M, with no significant loss of iodine being caused by the sparging. A nitrogen-sparged eluent, consisting of 0.04% potassium iodide in 0.3% acetic acid solution, remains free of iodine for several hours when stored under nitrogen. Hypochlorite was determined directly on-line in the range 0.1 x 10-4-2.5 x 10-4 M by injecting nitrogen-sparged sample solutions into this eluent.The reaction is fast and complete on-line. Hydrogen peroxide was determined on-line by injecting nitrogen-sparged sample solutions, 2.5 M in sulphuric acid, into a nitrogen-sparged 5% potassium iodide eluent. The reaction of hydrogen peroxide with iodide was incomplete on-line. Iodine formed in the molybdate-catalysed and the uncatalysed reactions of hydrogen peroxide with iodide off-line was determined by injection into a 1 M sulphuric acid eluent. Keywords : Flow injection analysis; amperometric detection; iodine determination; h ypochlorite determination; hydrogen peroxide determination Iodine titrations are useful in analytical chemistry because of the relatively low standard reduction potential of the 12 - 21- couple (0.54 V vs.NHE). This low reduction potential means that iodine reacts more selectively than many other oxidising agents and also that iodide is oxidised to iodine by a wide range of oxidising agents. The couple is reversible and very sharp end-points are possible in titrations with thiosulphate using visible means of detecting iodine, starch indicator or amperometric or biamperometric end-point detection. In iodimetric methods, the reduced forms of couples with lower reduction potentials are determined by monitoring the amount of iodine used to oxidise them, whereas oxidised forms of couples with higher standard reduction potentials are determined by the amount of iodine produced on reaction with iodide. The possibility of carrying out iodimetric reactions on-line using a flow injection technique with amperometric detection of iodine at a glassy carbon electrode has been investigated.Previous work in this laboratory has included on-line bromina- tion reactions and hypochlorite has been determined by injection into acidic bromide solutions.* Lee and Pollard3 have developed a spectrophotometric flow injection method of determining the iodine values of fatty acids. Miller et ~ 1 . ~ have described two flow injection methods using continuous and stopped-flow spectrophotometric detection of iodine; in their work, exclusion of air from the flow system was essential in order to avoid oxidation of iodide by oxygen. Experimental The flow injection analysis was carried out in a single-channel system similar to that described previously.5 The eluent flow was produced by means of a Metrohm pressure bottle (EA l l O l ) , which has a maximum nitrogen pressure limit of about 0.8 bar.The sample (a fixed volume of about 100 pl) was injected with a low-pressure Rheodyne valve (5020) connected between the pressure bottle (which contained the eluent) and a laboratory-built detector cell by means of 0.58-mm bore PTFE tubing. The detector cell only holds the glassy carbon electrode, the eluent being introduced to the stationary electrode in a wall-jet configuration. The detector cell was partly immersed in 0.1 M potassium chloride solution. The platinum counter and saturated calomel reference electrodes were placed in this electrolyte to complete the three-electrode system. A newly polished Metrohm glassy carbon disc electrode (EA286,3 mm diameter) was used.For the determination of iodine, the potential of the glassy carbon electrode was held at -0.2 V vs. SCE by means of a PAR 174A polarographic analyser (Princeton Applied Research). Values of currents were recorded on a Tarkan 600 y - t recorder. Eluents were sparged with nitrogen [which had been passed through a vanadium(I1) scrubber] in the pressure vessel before use; nitrogen pressure systems are ideal for maintaining air-free eluents in the reservoir. The values of background currents (the base lines from which current signals were measured) were not normally noted as they did not affect measurements at the levels of iodine monitored here. The background current at -0.2 V vs. SCE that was obtained for the nitrogen-sparged 0.4% potassium iodide solution was 10 nA.Other solutions were sparged where indicated in the text. For determinations of 12 at levels lower than 6 X 10-6 M, sample solutions were sparged with nitrogen for 30 s. At 1 2 levels lower than 1 x 10-6 M, nitrogen sparging of the sample solution was carried out for 30 s directly in the syringe by a method similar to that of Lloyd6; the PTFE transmission tubing was contained within PVC tubing of larger bore and nitrogen was passed through the space between the two tubes. In the applications described in this paper, it would not normally be necessary to work at such low concentrations of iodine.880 ANALYST, AUGUST 1986, VOL. 111 Eluents and Reagents Standard iodine (12) solution, 0.05 M.Dissolve 2g of potassium iodide in about 4 ml of water in a glass-stoppered 100-ml calibrated flask and add, through a small dry filter funnel, about 1.27 g of iodine (analytical-reagent grade). Insert the stopper and agitate the flask until all the iodine has dissolved. Allow the solution to reach room temperature and dilute to 100 ml with water. Store the solution in a cool dark place in a well stoppered glass bottle. Prepare more dilute standard solutions by diluting this solution with 0.4% potassium iodide solution. It is not necessary to de-oxygenate the water used to prepare this solution. The solution was standardised titrimetrically with standard sodium thiosulphate solution7 before use. Flow Injection Determination of Iodine A 0.4% nitrogen-sparged solution of potassium iodide was chosen as the eluent for this work.A 1-m length of 0.58-mm bore transmission tubing and a flow-rate of 6.0 ml min-1 were used. Currents obtained at various fixed potentials for the injection of nitrogen-sparged and unsparged solutions of the same composition as the eluent and for a 30 s-sparged 1.1 x 10-5 M solution of I2 in 0.4% potassium iodide solution are given in Table 1. The effect of not sparging the injected solution clearly becomes more significant at potentials more negative than -0.2 V. At -0.2 V, however, the increase in current when sparging of the injected solution is not carried out is less than 1% of the signal for the 1.1 x 10-5 M I2 solution. No loss of signal by contamination of the electrode was observed during over 100 injections of iodine at this level and the coefficient of variation was less than 1%.Calibration graphs were rectilinear over the range 1 X 10-7-1 x 10-3 M 12. The further precautions to exclude air described above were taken in determining 12 levels below 6 x 10-6 M. Currents at -0.2 V vs. SCE, corrected for the blank at the 4 X 10-7, 1 x 10-6 and 1 x 10-5 M I2 levels, were 24.8, 59.2 and 600 nA, respectively. Standard sodium hypochlorite solution, about 0.05 M. Dilute 20 ml of sodium hypochlorite solution (about 10% available chlorine) and 25 ml of 2 M potassium hydroxide solution to 500 ml with water. This solution was freshly prepared and standardised by reaction with iodide and titration of the liberated iodine with standard sodium thiosulphate solution as required.’ Standard hydrogen peroxide solution, about 0.07 M.Dilute 10 ml of 20 volume hydrogen peroxide solution to 250 ml with water. This solution was freshly prepared and standardised iodimetrically as required, allowing 15 min for reaction with iodide before titrating the iodine formed with standard sodium thiosulphate solution.7 Potassium iodide (0.04%) in 0.3 M acetic acid solution (eluent). Add 17.25 ml of glacial acetic acid to about 500 ml of water, add 0.4 g of potassium iodide dissolved in water and dilute to 1 1. This eluent was freshly prepared as required and was sparged with nitrogen for at least 15 min before use. On-line Determination of Hypochlorite Commercial hypochlorite products are usually alkaline, as hypochlorite is more stable under these conditions.In this study, hypochlorite was injected as solutions 0.1 M in potas- sium hydroxide; these solutions were sparged with nitrogen for 30 s before injection. Optimisation studies indicated that a transmission tube 1 m in length and with a 0.58 mm bore and a flow-rate of 7-9 ml min-1 should be used and the lower flow-rate of 7 ml min-1 was adopted. Currents associated with the electrochemical reduction at various measurement poten- Table 1. Effects of measurement potential on the values of the current signals obtained for the injection of 100 pl of nitrogen-sparged and unsparged blanks with the same composition as the eluent and of a 1.1 X 10-5 M solution of I2 in 0.4% potassium iodide solution (that had been sparged with nitrogen gas for 30 s) into a nitrogen-sparged 0.4% potassium iodide eluent Potential vs.SCE/V . . . . . . 0.1 0 -0.1 -0.2 -0.3 -0.4 -0.5 Signal for unsparged blank/nA . . 4 4 4 4 6 9 14 Signal for sparged blank/nA . . 6 6 8 10 18 58 182 Signal for sparged iodine solution/nA . , . . . . . . 292 456 578 664 680 680 685 Table 2. Effects of measurement potential on the values of current signals obtained for the injection of 100 p1 of 5.75 x lops M hypochlorite in nitrogen-sparged 0.1 M potassium hydroxide into sparged eluents of various acidities all containing 0.4% potassium iodide. The signals obtained with sparged blanks are also given: these values have been subtracted from the values given for injection of hypochlorite PotentialiV 0.2 Signal in 0.1 M potassium hydroxide*/pA .. . . . . <O. 00 1 Blank/pA . . . . . . . . <0.001 Signal in 0.1 M dipotassium hydrogen phosphate*/pA . . 0.016 Blank/pA . . . . . . . . 0.004 Signal in 0.1 M potassium dihydrogen phosphate*/pA . . 0.12 Blank/pA . . . . . . . . 0.024 Signal in 0.3 M acetic acid*/pA . . . . . . . . 0.94 Blank/pA . . . . . . . . 0.010 Signal in 0.1 M sulphuric acid*/pA . . . . . . . . 1.84 Blank/pA . . . . , . . . 0.014 * Containing 0.4% potassium iodide. ~~ 0.1 <0.001 <0.001 0.008 0.006 0.84 0.024 1.65 0.010 2.60 0.004 0 <0.001 <0.001 0.032 0.008 1.60 0.028 2.55 0.008 2.95 0.004 -0.1 0.002 <0.001 0.076 0.008 2.43 0.028 2.94 0.004 3.04 0.004 -0.2 0.006 <0.001 0.108 0.012 2.46 0.024 3.10 0.004 3.14 0.004 -0.3 0.010 0.002 0.148 0.016 2.74 0.016 3.10 0.006 3.20 0.005 -0.4 0.012 0.010 0.108 0.076 2.74 0.012 3.08 0.010 3.20 0.007 -0.5 0.036 0.020 0.056 0.140 2.88 0.008 3.05 0.018 3.10 0.012ANALYST, AUGUST 1986, VOL.111 881 tials of the iodine formed were obtained for the injection of 100 pl of 5.75 X 10-5 M hypochlorite in nitrogen-sparged 0.1 M potassium hydroxide solution into eluents of various acidities, all of which contained 0.4% potassium iodide and had been sparged with nitrogen for at least 15 min before use. These currents and those obtained for the injection of sparged blanks are given in Table 2. From the values of these currents, it is apparent that a reasonably high acidity is required for the full formation of iodine although the signals obtained with 0.3 M acetic acid were not significantly lower than those with 0.1 M sulphuric acid.Because the formation of iodine by the reaction of iodide with dissolved molecular oxygen is much slower in acetic acid solution than in sulphuric acid solution, the use of the 0.3 M acetic acid eluent containing potassium iodide was preferred. The effect of the potassium iodide concentration of the eluent was next studied. The current signals obtained at -0.2 V vs. SCE for identical injections of nitrogen-sparged hypochlorite into sparged 0.3% acetic acid eluent containing 0, 0.004, 0.04, 0.4 and 4% potassium iodide were 0.22, 4.16, 4.04,3.76 and 3.04 PA, respectively. The reduction in current at higher potassium iodide concentrations probably arises owing to the increased viscosity and density of the solutions. A potassium iodide concentration of 0.04% was adopted here to ensure an excess of iodide at the higher levels of hypochlorite determined.Calibration graphs were shown to be rectilinear Table 3. Values of current signals at -0.2 V vs. SCE obtained for the injection of 100 p1 of 6.4 X M hydrogen peroxide in various concentrations of nitrogen-sparged sulphuric acid into sparged 1 YO potassium iodide eluent. Blank signal for 5 M sulphuric acid = 0.17 PA; others lower but not recorded. Blanks not subtracted Sulphuric acid concentrationh . . 0.01 0.1 0.5 1 2.5 5 Signal/yA . . . . . . 0.76 1.14 2.04 3.68 8.64 13.9 Table 4. Values of current signals (FA) at -0.2 V vs. SCE obtained for the injection of 100 pl of 6.4 x M hydrogen peroxide in nitrogen-sparged 1 and 2.5 M sulphuric acid solutions into sparged eluents consisting of solutions of various concentrations of potassium iodide.Blank signal for 5% potassium iodide/2.5 M sulphuric acid eluent = 0.16 FA. Blanks not subtracted Current/pA Potassium iodide concentration/% m/V . . 0.002 0.01 0.1 1 5 Sulphuricacideluent (1 M) . . 0.16 0.19 0.34 3.68 9.6 Sulphuric acid eluent (2.5 M) 0.17 0.19 0.76 8.68 16 over the range 0.13 x 10-4-2.6 X 10-4 M and the coefficient of variation (five determinations) was less than 1% at the 0.65 x 10-4 M level. On-line Determination of Hydrogen Peroxide Optimisation studies indicated that a 0.58 mm bore trans- mission tube of 3 m length and a flow-rate of 1 ml min-1 were satisfactory and these values were therefore adopted. The formation of iodine was incomplete in this on-line determi- nation of hydrogen peroxide; however, increasingly greater amounts of iodine were formed when higher concentrations of sulphuric acid and iodide were used.This is illustrated in Tables 3 and 4, which show the currents obtained when 100-pl volumes of 6.4 x 10-4 M hydrogen peroxide as sparged solutions with different concentrations of sulphuric acid were injected into a 1% potassium iodide eluent which had been sparged for 15 min. The tables also show the currents obtained when hydrogen peroxide was injected as 1 and 2.5 M sulphuric acid solutions into sparged eluents consisting of various concentrations of potassium iodide. Calibration graphs were obtained for the injection of hydrogen peroxide in sparged 2.5 M sulphuric acid solution into a 5% potassium iodide eluent, which had been sparged for 15 min.These were found to be rectilinear over the range 0.06 x 10-3-6 X 10-3 M (Table 5) with coefficients of variation (five determinations) typically less than 1%. Flow Injection Determination of Hydrogen Peroxide with Off-line Formation of Iodine As the formation of iodine was determined to be only about 50% complete in the on-line determination of hydrogen peroxide, some laboratories may prefer to use flow injection amperometry only to replace the step involving titration of the iodine formed in an off-line reaction of iodide and hydrogen peroxide. This simple application of the flow injection amperometric determination of iodine is illustrated here. The titration methods given by Vogel7 were adapted for use with FIA.A transmission coil length of 1 m (0.58 mm bore) and a flow-rate of 6.0 ml min-1 were used. A volume of 2.5 ml of 20% potassium iodide solution and and 0-5 ml of sample solution (0.05 x 10-2-6 X 10-2 M in hydrogen peroxide). were added to 40 ml of 1 M sulphuric acid contained in a 50-ml calibrated flask. After diluting with 1 M sulphuric acid, the flask was firmly stoppered and the solution allowed to stand for 15 min in order to allow the iodine to form fully before injecting the solution into nitrogen-sparged 1 M sulphuric acid eluent. The same reaction was carried out adding 2 drops of 3% ammonium molybdate tetrahydrate to the reaction solution as a catalyst. In this latter instance, the iodine forms rapidly and the injection is made immediately on dilution.Table 5. Values of current signals at -0.2 V vs. SCE obtained for the on-line determination of hydrogen peroxide by injection in sparged 2.5 M sulphuric acid solution into sparged 5% potassium iodide eluent. Blank not subtracted Hydrogen peroxide concentration ( X 1 0 - 5 M ) . . . . . . . . o 6.2 18.6 43.4 62.2 434 622 Signal/pA . . , . . . . . 0.16 1.68 4.72 11.4 15.7 113 154 Table 6. Values of current signals at -0.2 V v s . SCE obtained for the off-line determination of hydrogen peroxide using the uncatalysed reaction and the reaction catalysed by ammonium molybdate Hydrogen peroxide concentration Signal for uncatalysed Signal for catalysed ( X 10-4 M) . . . . . . . . o 0.031 0.093 0.23 0.62 1.00 1.86 2.0 4.34 6.00 18.6 62.2 * * * * 4.40 * 9.20 * 28.8 92 300 reaction/pA .. . . . . . . 0.1 reaction/wA . . . . . . . . 0.08 0.19 0.46 1.10 300 * 8.70 * 19.6 28.7 88.0 296 * Not determined.ANALYST, AUGUST 1986, VOL. 111 This might be expected to be a more satisfactory method but, as the molybdate also catalyses the oxidation of iodide to iodine, it is necessary to de-oxygenate sample and reagent solutions when using this catalysed method.’ Currents obtained for the construction of a calibration graph are given in Table 6. These show good rectilinearity and coefficients of variation (five determinations) were typically less than 1%. Discussion Methods are given for the on-line iodimetric determination of hypochlorite and hydrogen peroxide using flow injection analysis, with amperometric detection by means of a glassy carbon electrode held at -0.2 V vs.SCE to monitor the iodine formed. These illustrate the possibilities of carrying out iodimetric determinations on-line in FIA systems. Compari- son of the signals with those obtained by the direct injection of standard iodine solutions under the same conditions indicate that complete reaction of hypochlorite with iodide occurs on-line. Hence maximum sensitivity and reliability is assured. Highly reproducible signals are also obtained for the on-line determination of hydrogen peroxide, but the reaction of hydrogen peroxide and iodide is incomplete. In using this latter method, it is important to ensure that other constituents of the sample are not catalysing the reaction. The possibility of determining by FIA the iodine formed by the complete reaction of iodide and hydrogen peroxide off-line has been demonstrated. A major disadvantage of this off-line approach is that the oxidation by dissolved molecular oxygen of iodide to iodine before injection is readily catalysed by both acidic solutions and light. The use of fully on-line reactions largely overcomes this difficulty and is also more convenient. The work described in this paper is being extended to the development of iodimetric methods in which reproducible amounts of iodine are produced on-line by the acidification of iodate - iodide solution and in which the amount of iodine required to oxidise the determinand is monitored. A.Y.C. thanks the Lebanese University for leave of absence and the Lebanese Government for financial support. References 1. 2. 3. 4. 5. 6. 7. Fogg, A. G., Ali, M. A . , and Abdalla, M. A., Analyst, 1983, 108, 840. Fogg, A. G., Chamsi, A. Y . , Barros, A. A . , and Cabral, J. O., Analyst, 1984, 109, 901. Lee, C. C., and Pollard, B. D., Anal. Chim. Acta, 1984, 158, 157. Miller, K. G., Pacey, G. E., and Gordon, G., Anal. Chem., 1985, 57, 734. Fogg, A. G., and Summan, A. M., Analyst, 1984, 109, 1029. Lloyd, J. B. F., J. Chromatogr., 1983, 256, 323. Vogel, A. I., “A Textbook of Quantitative Inorganic Analy- sis,” Fourth Edition, Longmans, London, 1978, p. 381. Paper A51370 Received October 16th, 1985 Accepted March 3rd, 1986

ISSN:0003-2654    

DOI:10.1039/AN9861100879

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, AUGUST 1986, VOL. 111 883 Simplified Technique for the Preparation of Glassy Carbon Electrodes Russell Moy* Union Carbide Corporation, Battery Products Division, 25225 Detroit Road, Westlake, OH 44 145, USA Glassy carbon electrodes, encapsulated i n a chemically inert Teflon FEP resin, have been prepared. The preparation of these electrodes is simple, fast and inexpensive. These electrodes are believed to be better suited than wax- or epoxy-impregnated electrodes when used in organic electrolytes or hostile environments. Electrodes prepared in this manner were found to be free from memory effects and leakage. Three identically prepared electrodes were found to behave similarly, facilitating the comparison of data obtained from different electrodes. These electrodes can be used in the evaluation of electrocatalysts, as sensors and in conventio na I vol tam metric analyses.Keywords: Glassy carbon electrodes; voltammetry Carbon electrodes are routinely employed in voltammetric and polarographic analyses as indicator electrodes. One advantage that carbon has over metals such as gold and platinum is its ability to resist surface oxide passivation.1 Metallic surface oxides can passivate the electrode surface and change its electrochemical properties. Different types of carbon have been used as indicator electrodes , with glassy carbon and pyrolytic graphite offering the most consistent behaviour, i.e., their electrochemical properties are not heavily dependent on their origin.2 The surface of the pyrolytic graphite was found to be susceptible to mechanical cleavage.Cleaved graphite can entrap materials, resulting in electrodes exhibiting memory effects. Glassy carbon has been used in the evaluation of electro- chemically active functional groups covalently bound to the electrode.3 Some covalently attached functional groups exhibit electrocatalytic activity,4 and the electrochemical evaluation of these functional groups is usually accomplished by standard voltammetric and polarographic techniques. The chemical modification of porous carbon substrates such as acetylene black and porous graphite can be accomplished, but the electrochemical evaluation of these materials is difficult. Voorhies and Davis5 and Elving and Smith6 reported that large background currents were obtained when these porous materials were used as working electrodes.Similar results were obtained in this laboratory when acetylene black and spectroscopic graphite electrodes were used in cyclic voltam- metry. These voltammograms are shown in Figs. 1 and 2. Wax impregnation is effective in reducing these currents.6 It should be noted that the wax may be soluble in, or reactive with, a number of non-aqueous electrolytes. Conventional glassy carbon electrodes available from PAR (Princeton Applied Research Corp. , Princeton, NJ, USA) or prepared by Panzer and Elving2 consist of amorphous carbon sealed in a glass tube with epoxy or melted polyethylene. These sealants may be attacked by aggressive or organic solvents. Other investigators have forced glassy carbon rods into tight-fitting Teflon sleeves'-9; however , electrolyte seep- age may occur when this procedure is employed. Electrolyte seepage may also occur when single-layer Teflon shrink tubing is used to seal the electrodes.Experimental The electrodes prepared for use in corrosive solvents consist of 3 mm diameter amorphous carbon rods (Grade V10, * Present address: Department of Chemical Engineering, University of Michigan, 2135 Dow Building, Ann Arbor, MI 48109, USA. Atomergic Chemetals, Plainview, NY , USA). The carbon is sealed to a two-layer Teflon shrink tube (Grade S-130, Norton/Chemplast, Wayne, NJ, USA). The outer layer is a heat-shrinkable Teflon TFE shell and the inner layer is a meltable Teflon FEP resin. On heating, the outer layer contracts and the inner layer forms a hermetic seal to the carbon. This procedure eliminates electrolyte seepage into the electrode.The carbon rods are cut into 1-cm lengths and dried for several hours at 150 "C. The carbon is inserted into 8-cm lengths of tubing. Electrical contact to the carbon is accom- plished with a nickel wire inserted between the carbon and tubing so that the end of the wire is at the midpoint of the carbon. In order to promote the adhesion of the FEP to the carbon, the electrodes are sealed under vacuum in a tube furnace. The carbon is polished with silicon carbide paper and diamond paste. Final polishing of the electrodes is carried out with a 0.25 pm diamond paste. The electrodes are then cleaned with methanol in an ultrasonic bath. Electrochemical pre-treatment of these electrodes was not used as this step has been reported to be deleterious to the reversibility of the hexacyanoferrate(I1) - hexacyano- ferrate(II1) couple on carbon electrodes.8,' Potassium hexa- cyanoferrate(I1) electrolytes were used for the evaluation of these electrodes. 3.0 2.0 1.0 0.0 -1.0 -2.0 PotentialiV vs.Ag - AgN03 Fig. 1. Voltammetry of acetylene black - Teflon electrode in a supporting electrolyte-. Supporting electrolyte: 0.1 M tetraethylammo- nium tetrafluoroborate in acetonitrile. Reference electrode: Ag - 0.1 M AgN03. Counter electrode: Pt gauze. Scan rate: 0.100 v sANALYST, AUGUST 1986, VOL. 111 884 t c 2 3 0 TI .- 5 0 Fig. 2. I I I , I I 2.0 1.0 0.0 -1.0 -2.0 -3.0 PotentiaVV vs. Ag - AgN03 Voltammetry of Teflon-sealed spectroscopic graphite in a supporting electrolyte.Supporting electrolyte: 0.1 M tetraethylammo- nium tetrafluoroborate in acetonitrile. Reference electrode: Ag - 0.1 M AgN03. Counter electrode: Pt gauze. Scan rate: 0.100 V s-l t c 2 3 V n .- 5 0 I I I 1 I I I I -0.4 -0.2 0 0.2 0.4 0.6 0.8 1.0 PotentiaW vs. SCE Fig. 3. Voltammetry of freshly prepared glassy carbon electrodes in a supporting electrolyte. Supporting electrolyte: 0.1 M H2S04 in water (deaerated). Reference electrode: SCE. Counter electrode: Pt gauze. Scan rate: 0.050 V s-1 I I I I I I I I -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 PotentialN vs. SCE Fig. 4. Voltammetry of glassy carbon electrodes after K4Fe(CN 6 soaking. Supporting electrolyte: 0.1 M H2S04 in water (deaerated]. Reference electrode: SCE. Counter electrode: Pt gauze.Scan rate: 0.050 V s-1 Results and Discussion Voltammograms for five randomly selected electrodes are shown in Fig. 3. The supporting electrolyte was a deaerated 0.1 M solution of sulphuric acid. Two of the electrodes, designated A and E , were found to exhibit significant background currents. These currents could possibly be attri- buted to incomplete polishing of the electrode surface. Additional polishing of these electrodes should reduce the background current. Electrodes B, C and D produce feature- less voltammograms with low background currents. For comparison, the voltammogram for a commercially available mV s-’ m1 /loo c t 4- 2 3 0 V TI .- z 0.0 0.2 0.4 0.6 0.8 1.0 PotentialiV vs. SCE Fig. 5. Anodic oxidation of Fe(CN),4- on electrode B. Electrolyte: 1.0 mM K,Fe(CN)6 - 0.1 M H2S04 - HzO.Reference electrode: SCE. Counter electrode: Pt gauze 0 0.1581 0.3162 0.4743 Fig. 6. Anodic peak current versus (scan rate)$ for the oxidation of Fe(CN) 4- on glassy carbon electrodes. H, Electrode B; A, electrode C; and b, electrode D (Scan rateN s-1)i glassy carbon electrode (PAR) is also shown in Fig. 3. The larger background current obtained with the PAR electrode is expected as its surface area is 5.5 times greater than that of those electrodes prepared in-house. The electrodes were stored overnight in a saturated solution of potassium hexa- cyanoferrate(I1) in order to test for memory effects. The electrodes were rinsed with distilled water and additional cyclic voltammograms were obtained (Fig. 4). These voltam- mograms do not differ significantly from those shown in Fig.3, indicating that the electrodes prepared in the above manner can be considered to be free from memory effects. Linear sweep and cyclic voltammograms were obtained for electrodes B, C and D at different scan rates in an aqueous solution consisting of 1 .O mM potassium hexacyanoferrate( 11) and a 0.1 M sulphuric acid supporting electrolyte. A typical voltammogram is shown in Fig. 5. A graph of the anodic peak height as a function of the square root of the scan rate is linear. An estimate of the theoretical slope and intercept of this line can be obtained from the equation ip = (2.69 x lo5) n:A D& v4 C,, . . . . (1) where n is the number of equivalents per mole of reactant, A cm2 is the electrode area, Do cm2 s-1 is the diffusion coefficient of the reacting ion, v V s-1 is the scan rate and Co mol cm-3 is the bulk solution concentration of the reactant.lo The diffusion coefficient of the hexacyanoferrate(I1) ion has been reported to be 3.9 x 10-6 cm* s-1.11 The theoretical slope of the peak current as a function of the square root of theANALYST, AUGUST 1986, VOL. 111 885 Table 1. Statistical analysis of voltammetric data Anodic peak height/A X 106 Scan rate/V s- l (Scan rate); Electrode B Electrode C Electrode D 0.010 0.100 2.76 2.72 2.99 0.020 0.1414 3.62 3.48 4.06 0.050 0.2236 5.20 4.80 5.55 0,100 0.3162 6.93 6.57 7.09 0.200 0.4472 9.25 8.66 9.13 Slope/A(Vs-’)-* x 105 . . . . 1.86 1.72 1.74 Intercept/A x lo7 . . . . . . 9.68 10.2 14.9 R . . . . . .. . . . . . 0.9997 0.9996 0.9973 Student t-ratio (intercept) . . 13.62 12.53 7.34 Student t-ratio (slope) . . . . -70.28 -68.74 -27.33 Expected slope: 3.76 x Expected intercept: 0 A A (V s- I)-$ t3.0.05 = 2.353 H,: experimental slope (or intercept) does not significantly differ from theoretical H I : experimental slope (or intercept) is significantly different than theoretical slope Test: reject H, if It1 > t3, o,Os (at the 90% level of significance) slope (or intercept) (or intercept) scan rate is therefore calculated to be 3.76 x 10-5 A(V s-+)-i and the theoretical intercept is zero. Student t-tests at the 90% level of significance showed that the experimental slope and intercept differed from the predicted values. The statistical calculations are shown in Table 1.This deviation from the predicted line suggests that there is some electrochemical irreversibility in the system. Wightman and co-workers12J3 have reported on the irreversi- bility of hexacyanoferrate(II1) reduction on carbon electrodes and similar results have been reported for hexacyano- ferrate(I1) oxidation.* Different polishing procedures may improve the reversibility of this couple on these electrodes.9 Nevertheless, Fig. 6 indicates that the peak heights for the different electrodes are almost identical. Hence, a comparison of data obtained from different electrodes will be meaningful. The preparation of these electrodes is simple, fast and inexpensive. They may be useful in situations requiring expendable electrodes. These electrodes can also be used for the evaluation of electrocatalysts and as indicator electrodes in conventional voltammetry. Electrolyte seepage into the elec- trode is eliminated when two-layer Teflon shrink tubing is used.Encapsulated electrodes should, however, be checked for memory and leakage effects in the manner described above. The author thanks Andrew Webber for helpful comments used in the preparation of this manuscript and Paul Dunn for his assistance in the preparation and evaluation of these electrodes. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. References Gunasingham, H., and Fleet, B . , Analyst, 1982, 107, 896. Panzer, R. E., and Elving, P. J., J. Electrochem. SOC., 1972, 119, 864. Rocklin, R. D., and Murray, R. W., J . Electroanal. Chem., 1979, 100,271. Yamana, M., Darby, R., and White, R. E., Electrochim. Acta, 1984, 29, 329. Voorhies, J . D., and Davis, S. M., Anal. Chem., 1960, 32, 1855. Elving, P. J., and Smith, D. L., Anal. Chem., 1960, 32, 1849. Rusling, J. F., Anal. Chem., 1983, 55, 1719. Engstrom, R. C., and Strasser, V. A., Anal. Chem., 1984,56, 136. Kamau, G. N., Willis, W. S., and Rusling, J. F., Anal. Chem., 1985, 57, 545. Bard, A. J., and Faulkner, L. R., “Electrochemical Methods,” Wiley, New York, 1980, p. 218. Eisenberg, M., Tobias, C. W., and Wilke, C. R., 1. Electro- chem. SOC., 1954, 101,306. Wightman, R. M., Kovach, P. M., Kuhr, W. G., andstrutts, K. J . , in Sarangapani, S . , Akridge, J. R., and Schumm, B . , Editors, “The Electrochemistry of Carbon,” Proceedings of the Electrochemical Society, 84-5 Electrochemical Society, Pen- nington, NJ, 1984, p. 510. Wightman, R. M., Deakin, M. R., Kovach, P. M., Kuhr, W. G., and Strutts, K. J., J . Electrochem. SOC., 1984,131,1578. Paper A51445 Received December loth, 1985 Accepted February 17th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9861100883

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, AUGUST 1986, VOL. 111 887 Use of a Multi-purpose Solid-state Ion-selective Electrode Body and an Agl-based Membrane Hydrophobised by PTFE for the Determination of I- and Hgz+ Josipa Komljenovic and Njegomir Radic* Faculty of Technology, University of Split, 58000 Split, Yugoslavia The design of a multi-purpose "all-solid-state" electrode body is described. The electrode body, which provides solid contact with the ion-selective membrane, can be used with either pressed pellet or metal disc membranes. The preparation and performance of a silver iodide-based pellet hydrophobised by PTFE are described. The pellet (Agl 25%, Ag2S 25% and PTFE 50% m/m) was mounted in the electrode body and examined as an ion-selective membrane for the determination of I- and Hg*+ ions. Keywords: Potentiometry; membrane electrodes; iodide; mercury(//) Commercial ion-selective electrodes (ISEs) with solid-state membranes are usually designed with a permanently fixed membrane.Except with the Rfiiieka-type Selectrode,' elec- troactive material cannot be exchanged or replaced. This paper describes the design of an inexpensive and easy to construct "all-solid-state" electrode body. This body can be used with either pressed pellet or metal disc membranes. As observed by Bagg and Rechnitz,* Ag2S - AgI electrodes show some roughness of and deposits on exposed surfaces after prolonged use in iodide solutions. It has also been shown3 that the potential of silver iodide-based silicone- rubber membrane electrodes is constant after soaking in 1 M KI, irrespective of variations in the activity of iodide ions in the sample solution.Iodide ISEs are most frequently employed for the determi- nation of Hg2+ on the basis of the formation of mercury(I1) - iodo complexes with the liberation of Ag+ ions.&' The relevant equilibrium constants indicate that Hg12 should be formed with the release of two silver ions per mercury ion, giving a response slope of 29.5 mV decade-'. However, it has been reported by various workers that the response slope can be 59.1 mV decade-1,8 or both non-linear and non- Nernstian .9 The preparation of Ag2S-based mercury(I1) ISEs has been reported in several papers.171+)-12 None of these electrodes gave a reproducible Nernstian response to Hg2f. In this paper, the successful development of an AgI-based membrane hydrophobised by PTFE is described.The membrane mounted in the electrode body showed very good mechanical properties and excellent response to the I- ion. The same electrode can be used as a sensor for the Hg2+ ion. Experimental Construction of the Electrode Body The two main parts of the electrode body (Fig. 1) were machined from PTFE. A stainless-steel disc and a coaxial cable provided electrical connection between the pellet and millivoltmeter. A ring of silicone-rubber was mounted between the pellet and the electrode body for two reasons: (i) silicone-rubber prevents leakage around the pellet and (ii) it improves contact between the pellet and steel disc. When the electrode was immersed vertically in solution, air bubbles were trapped on the active electrode's surface, and no stable potential values were recorded.This problem was solved by using a cell in which the electrode could be held in a sloping position. Preparation of the Membrane Silver salts were prepared by the addition of silver nitrate solution to sodium sulphide or potassium iodide solutions in a stoicheiometric ratio. The precipitates were washed with hot water and acetone, filtered and dried in air at about 373 K. Dry silver salts were mixed with PTFE powder. Approxi- mately 0.5 g of membrane material (AgI 25%, Ag2S 25% and PTFE 50% mlm) was pressed using a KBr die with a 13 mm diameter plunger. The pellets were pressed at 740 MPa for 1-2 h at room temperature. Before mounting in the electrode body, both surfaces of each pellet were polished with an Orion polishing strip (Orion catalogue No.948201). Reagents All chemicals were of analytical-reagent grade. Solutions were prepared with doubly distilled water. A stock solution of 0.1 M //-Coaxial cable Membrane * To whom correspondence should be addressed. 'Silicone rubber Fig. 1. Electrode body888 -240 ANALYST, AUGUST 1986, VOL. 111 - // Hg(I1) was prepared by using mercury(I1) nitrate and was adjusted to ca. 0.05 M in the corresponding acid. The total ionic strength was kept constant by the addition of potassium nitrate solution. Procedure All potentiometric measurements were made with an Orion 901 Microprocessor Ionalyzer using double-walled seals main- tained at 298 & 0.1 K. The reference electrode was an Orion 90-02 double-junction reference electrode (DJRE), with 10% potassium nitrate solution as the outer filler.The pH values were checked with an Orion 91-15 glass electrode and 801 pH - mV meter. During measurements the solution was stirred with a PTFE-coated magnetic bar. The stirring speed and electrode distance were both kept constant. The response measurements for I- and Hg2+ were made by serial dilution of the cell solution or by standard addition. The pH and ionic strength of the cell solution were kept constant during the measurements. Results and Discussion The solid-state ion-selective body provides solid contact between the metal conductor and the ion-selective membrane. This contact gives the solid state ion-selective body several advantages over ion-selective electrodes with liquid fillings.13 The use of the noble metals is not required if many solid contacts are present. A reversible transition from electronic metal conductivity to the ionic conductivity of the measured solution can be carried out in ISEs with membranes of the Ag2S type. These membranes have mixed ionic and electronic conductivity, and electron exchange was indicated as the potential-determining process at the solid - solid interface.14 When silver wire is embedded in an AgX or Ag2S pressed pellet, an “all-solid-state’’ selective electrode is formed, and its response is reversible to solution activities of Ag+ and X- or S2-. Similar behaviour is observed when there is direct contact between the pressed pellet and the inert electronic conductor or with the use of metals whose salt formation free energies are positive relative to silver.15 An electrode consist- ing of a surface coating of AgI on silver wire also responds reversibly to ionic activities, and its potential in solution with iodide is given at 298 K by E = -0.150-0.05910gaI- . . . . (1) The standard potentials of all-solid-state devices, with apparently the same sensitive materials, can differ signifi- cantly from electrodes of the second kind, because of the electronic conductor - salt interfacial potential. The electrode studied here is of the silver sulphide type. In an attempt to improve its mechanical properties and decrease its reactivity with iodide at higher concentrations, AgI - Ag2S powder was mixed with PTFE powder before pressing. PTFE as an organic phase does not serve as a reservoir of the electroactive material, but prevents penetration by the aqueous sample solution.At this stage of investigation, it has been found that PTFE (50% mlm) improves mechanical properties and decreases the formation of silver iodide complexes. The response of the electrode with a partially hydropho- bised pellet to I- is shown in Fig. 2, line B. The maximum response time of the potential cell with the electrode was 1 min, even at low concentrations of iodide ions. The electrode reached a steady-state potential (drift 0.2 mV min-1) within the time of the experiments. In all experiments, the potential drifts were parallel and relatively small. The experimental slope of 59 mV (PI)-’ obtained was in excellent agreement with theoretical values.The standard potential for the reaction AgI(s) + e Ag(s) + I- . . . . (2) + 400 > +200 E 0 - 200 -Log[ Hg2+] 7 6 5 4 3 2 1 I I I I I I I 0 . , 0 . ./ I I I t I 1 1 7 6 5 4 3 2 1 -Log [I -1 Fig. 2. Response of the electrode with a silver iodide-based membrane hydrophobised by PTFE to mercury(I1) ion and iodide ion in aqueous solution at constant ionic strength. A, Response to mercury(I1) ion; and B, response to iodide ion 1 2 3 4 Time/h Fig. 3. Dynamic response curve of the electrode with hydrophobised membrane in 1 M KI solution was found by extrapolating the appropriate potential line, -log aI-, to log ar- = 0. Activities were calculated by use of the Debye - Huckel equation. If the junction potential is ignored after a correction of 242 mV (the potential of the reference electrode at 298 K16), a value of EO = -41 mV is obtained, which is more positive than the value predicted in equation (1), -150 mV vs.NHE, and the published value17 when the AgI membrane was sealed with silver wire. The standard potentials are often more positive than theoretical values when silver or another metal is employed as the internal contact to the Ag2S pellet.18.19 After one week of use in iodide solutions with concentra- tions from 10-1 to 10-7 M , we observed no visual changes to the electrode surface. During this period the electrode showed Nernstian behaviour with a rapid response. The exposed surface of the AgI - Ag2S membrane pellet without PTFE was slightly tanned, even after only one experiment in the above concentration range.We suspect that at a high iodide concentration, AgI is leached out via the formation of soluble iodide complexes. PTFE presumably prevents penetration by the aqueous sample and hence decreases this process. When the membrane was soaked for 20 h in 1 M KI, a lower potential slope and narrower linear range were observed. However, after soaking, the electrode showed no loss in response to the activity of iodide in solution, as has been found in experiments with an iodide-based silicone-rubber membrane.3 At the iodide concentrations underANALYST, AUGUST 1986, VOL. 111 889 + 400 , +zoo E G 0 -200 15 11 7 3 Fig. 4. Relationship between the calculated activities of silver ion at the membrane surface and the experimental potential values of electrode in iodide solution (closed circles) and mercury(I1) solution (open circles) consideration, the electrode potential was unstable and a steady-state potential was reached only after about 1.5 h (Fig.3). The change in potential indicates some process at the surface of the electrode and, according to the direction of the potential change, this process can increase the activity of the silver ion at the phase boundary. After prolonged soaking in concentrated I- solutions, the electrode lost its Nernstian behaviour, but showed satisfactory potential response after the membrane was polished. Metal ions which form less soluble iodides than AgI or stable metal - iodide complexes will interfere. When the electrode is immersed in a solution containing Hg2+ ions the reaction PAg AgI(s)+ Hg2+ SHgI,(2-X)+ + x Agf .. (3) may be expected to occur. The formation constants of HgI(2-x)f species are so large that a vanishingly small Hg*+ ion concentration remains near the electrode surface. The HgIx formed diffuses into the solution with silver ions. The electrode phase which has started to dissolve responds to the Ag+ ions in solution near the electrode surface. Conse- quently, if the concentration of the mercury(I1) ion in solution and the solubility constant are sufficient for reaction (3) to proceed to the right, the activity of silver ion at the phase boundary is determined by the activity of mercury(I1) ions in solution and the potential of the electrode can be expressed in simplified form: 0.059 E = K + -log[Hg‘+] . . . . (4) X where K is a conditional standard electrode potential.In another experiment, the potential response of an electrode with a hydrophobised membrane to the Hgz+ ion was examined. The recorded potential values (Fig. 2, line A) indicate HgIx(2-x)+ (1 < x < 2) formation at the exposed surface of membrane. After experiments with Hg2+ the membrane responded perfectly, without polishing, to the I- ion in the experimental concentration range and the potential values have been acceptably recorded. However, in repeat experiments with mercury(I1) solution, the potential values increased to 15 mV, although the slope of the experimental curve [46 f 1 mV (pHg)-l], in addition to its linear range, were acceptably constant. The response time of the electrode in mercury(I1) solution was less than 1 min.In both the experiments, with iodide and mercury(I1) ions, the electrode potential is determined with an activity of Ag+ ion on the electrode surface. The experimental potential values are plotted against calculated -log aAg+ in Fig. 4. The calculation of pAg is based on the equation pAg=pK$#’-pl . . . . . ( 5 ) For the experimental concentration range of iodide (closed circles in Fig. 4), the iodide activities have been calculated from molar concentrations and activity coefficients at the experimental ionic strength. Activities of iodide in the mercury(I1) concentration range (open circles in Fig. 4) have been calculated by using equation (1). Instead of the standard potential (- 150 mV), the experimental standard potential of the electrochemical cell (-283 mV) was used in calculations.The calculated activities of silver ions in solution with the Hg(I1) ion are slightly shifted from the Nernstian slope and indicate higher values than expected. The standard potential of the electrochemical cell, which was determined in iodide solution, does not absolutely fit the electrode behaviour in mercury(I1) solution. These results indicate that there is a possibility that the mercury(I1) ion in reaction with the membrane forms a thin film at the surface of the membrane. However, this coating may be fairly complicated and unstable in solutions without the Hg2f ion. This suggests that AgI plays an important role in the response of the electrode potential to the Hg2+ ion. We also found in preliminary experiments that electrodes with membranes consisting of 33% AgI, 33% Ag and 34% mlm PTFE showed a Nernstian slope for monovalent elec- trodes in an Hg(I1) concentration range from 10-2 to near 10-5 M.The experimental slope was 58 mV (pHg)-1. In addition, in experiments with mercury, the slope of electrodes with AgI - Ag2S - PTFE membranes increases with decreasing amount of organic phase. When the PTFE content was reduced by up to 33%, the experimental slope was about 50 mV. The financial support for this work, including a graduate research assistantship from the Republican Council for Science (SIZ 11) of Croatia for J. K., is gratefully acknowledged. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. References RiiiCka, J., and Lamm, C. G., Anal. Chim. Acta, 1971, 53, 206. Bagg, J., and Rechnitz, G. A., Anal. Chem., 1974, 46, 943. Malissa, H., Grasserbauer, M., Pungor, E., Toth, K., Papay, M. K., and Pdos, L., Anal. Chim. Acta, 1975, 80, 223. Vesely, J., Weiss, D., and Stulik, K., “Analysis with Ion- selective Electrodes,” Ellis Horwood, Chichester, 1978, p. 193. Somer, G., Anal. Chem., 1981, 53, 4143. Aomi, T., Denki Kagaku, 1980, 48, 491. Arnold, M. A., and Meyerhoff, M. E., Anal. Chem., 1984,56, 20R. Orion Newsl., 1970, 9a, 10. Morf, W. E., Kahr, C., and Simon, W., Anal. Chem., 1974,46, 1538. Anfalt, T., and Jagner, D., Anal. Chim. Acta, 1971, 55, 477. van de Leest, R. E., Analyst, 1977, 102, 509. van der Linden, W. E., and Oostervink, R., Anal. Chim. Acta, 1979, 108, 169. Nikolski, B. P., and Materova, E. A., Ion-Select. Electrode Rev., 1985, 7, 3. Buck, R. P., Anal. Chem., 1974,46,28R. Buck, R. P., in Covington, A. K., Editor, “Ion-Selective Electrode Methodology,” Volume I, CRC Press, Boca Raton, FL, 1979, p. 203. Crombie, D. J., Moody, G. J., and Thomas, J. D. R., Anal. Chim. Acta, 1975, 80, 1. Papeschi, G., Bordi, S . , and Carla, M., J . Electrochem. SOC., 1978, 125, 1807. Vesely, J., Jensen, 0. J . , and Nicolaisen, B . , Anal. Chim. Acta, 1972, 62, 1. Koebel, M., Anal. Chem., 1974,46, 1559. Paper A616 Received January 6th, 1986 Accepted February I7th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9861100887

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST AUGUST 1986 VOL. 111 891 Mercury( II) and Silver( I) Ion-selective Electrodes Based Crown Ethers Ming-Tain Lai and Jeng-Shang Shih" Department of Chemistry National Taiwan Normal University Taipei Taiwan 1 17, on Dithia Republic of China Mercury (Hg2+) and silver (Ag+) ion-selective PVC membrane electrodes based on 1,4-dithia-I 2-crown-4 and 1,4-dithia-l5-crown-5 as neutral carriers were successfully developed. Both electrodes exhibited good linear responses of 30 and 40 mV decade-1 for Hg2+ and Ag+/ respectively within the concentration ranges 10-2-10-6 M Hg(N03)2 and 10-1-10-6 M AgN03. Some other crown ethers and cryptands were also investigated as neutral carriers for both ions. Both the mercury and silver electrodes exhibited comparatively good selectivities for mercury(l1) and silvertl) ions in comparison with alkali metal alkaline earth metal and some heavy metal ions.These crown ether ion-selective electrodes are suitable for use with aqueous solutions at pH 3 2. They were applied as sensors in titrations of Br- and CI- with Ag+ and of Hg2+ with 1- and Cr2072- and in the determination of the solubility products of AgCl in aqueous solutions. Keywords Dithia crown ethers; mercury(//) and silver(/) ion-selective electrodes Crown ethers have been demonstrated to be highly selective complexing agents for many metal and can be applied in their separation5-7 and determination.sl0 In general a crown ether can form a stable complex with a metal ion that fits well in the cavity of the crown ether. Taking advantage of their ion-discriminating ability crown ethers have been shown11-19 to be suitable neutral carriers for ion-selective electrodes (ISEs) especially for potassium as common crown ethers such as 18-crown-6 and 15-crown-5 can form stable complexes with alkali metal ions especially potassium ions.However crown ether-based ion-selective electrodes for transition metal ions such as Ag+ and Hg2+ have not been reported as they would suffer interference from alkali metal ions. On the other hand by using thia crown ethers in which the oxygen atoms are partly replaced with sulphur the interference from alkali metal ions can be expected to be significantly reduced as Hg2+ and Ag+ ions form stronger complexes with thia crown ethers than do alkali metal ions.20.21 There is no commercially available mercury( 11) ion-selective electrode.Even in the literature there is no report of a specific electrode for Hg2+ ions but a few other electrodes, such as the iodide electrode (AgI - Ag2S) which respond to mercury ions have been described.22-24 For Ag+ ions, homogeneous and heterogeneous silver sulphide solid elec-trodes are commercially available. The compacted crystal silver electrode gives a good performance but it is difficult to prepare; the crown ether silver electrode prepared in this work is more easily fabricated. Systematic studies performed with various crown ethers and cryptands in this study showed that PVC membranes containing 1,4-dithia-12-crown-4 and 1,4-dithia-15-crown-5 were suitable as neutral carriers for mercury(I1) and silver( I) ion-selective electrodes respec-tively.The electrochemical selectivities for various ions and the effects of the membrane matrix crown ether concentration, internal solution and pH for both Hg2+ and Ag+ electrodes were investigated. In addition both electrodes were used in the titration experiments for the determination of Hg2+ and Ag+ ions. Experimental Chemicals All chemicals were of analytical-reagent grade. 1,4-Dithia-15-crown-5 1,4-dithia-12-crown-4 and 1,7-* To whom correspondence should be addressed. dithia-12-crown-4 were synthesised by methods reported in the literature.25 1,4-Dithia-15-crown-5 was prepared by vacuum distillation (200 "C/3 mmHg) of a mixture of l,ll-dichloro-3,6,9-tri-oxundecane (25 g) ethane-1,2-diol (9.5 g) and NaOH (9 g) in ethanol and recrystallisation from a 1 + 1 benzene - hexane mixture.Yield 11%; m.p. 51-52 "C. NMR 6 = 2.63(t), 2.78(s) 3.55(s) and 3.69(t). GC - MS M+ = 252. 1,7-Dithia-12-crown-4 was obtained by vacuum distillation (200 "C/3 mmHg) of a mixture of bis(2-mercaptoethyl) ether (9.2 g) bis(2-chloroethyl) ether (9.5 g) and NaOH (6.0 g) in ethanol. Yield 19%; colourless liquid. NMR 6 = 2.77(t), 1,4-Dithia-12-crown-4 was prepared by vacuum distillation (260 "C/3 mmHg) of a mixture of 1,2-bis(2-chloroethoxy)-ethane (14 g) ethane-1,2-dithiol (7 g) and NaOH (6.0 g) in ethanol and purification by passage through a silica gel column with 3 + 1 benzene - hexane as the eluent. Yield 16%; colourless liquid. NMR 6 = 2.61(t) 2.91(s) 3.56(s) and Cryptand 222 and Cryptand 22 were obtained from E.3.47(t). GC - MS M+ = 208. 3.77(t). GC - MS:M+ = 208. Merck. Electrode Preparation A mixture of about 100 mg of PVC,26,27 30 mg of crown ether and 50 mg of dibutyl phthalate was dissolved in 2 ml of tetrahydrofuran (THF). In some instances especially for the mercury( 11) electrode sodium tetraphenylborate (STPB) was also added. The PVC - THF solution was poured into a glass dish of diameter 5 cm and the THF was evaporated at room temperature for about 24 h. A semi-transparent membrane about 0.03 mm thick was obtained. A good linear correlation between the thickness of the membrane and the amount of PVC was found. A piece about 12 mm in diameter was cut from the PVC membrane and attached to a polyethylene cap by wetting the membrane with the PVC - THF solution mentioned above.The diameter of the exposed membrane was about 7 mm. The polyethylene cap with the membrane was then incorporated into an Ag - AgCl wire electrode or Hg - Hg2C12 electrode. After filling with internal solution contain-ing 5 X 10-4 M Hg(N03)2 or AgN03 and 10-2 M HN03 the electrode was conditioned by soaking in 0.1 M Hg(N03)2 or AgN03 solution for 24 h. A salt bridge containing KN03 was prepared in each instance. The electrochemical system for this study was as follows: Ag - AgCl or Hg - Hg2C12 1 internal solution ( 5 X M AgN03 or Hg(N03)2 + 10-2 M HN03) 1 PVC membrane 892 ANALYST AUGUST 1986 VOL. 111 testing solution 1 salt bridge (1 M KN03) I satd. KCl 1 Hg2C12 -The e.m.f. measurements were made with a Basic Model 321 digital pH/mV meter.The response times of these crown ether electrodes were checked and were fairly short ( S l min). In this work the potential measurements were taken 3 min after the introduction of the test solutions which were stirred during the measurements. Hg Evaluation of Electrode Selectivity The selectivity coefficients (KPAOgfM or for various ions were evaluated graphically by the mixed-solution method.28 The electrical potentials of solutions of the Ag+ or Hg2+ ions alone (10-5 M) and of a mixed solution containing a fixed amount of Ag+ or Hg2+ ions (activity uHg or uAg = M) and a varying amount of the interfering ion M*+ (activity a“) were measured as El and E2 respectively and can be expressed as follows28: (I) KPA”,’M OE = { exp[ ( ~ 2 - E ~ ) F / R T ‘ I } ~ ~ - aAg .. and a$ = {exp[(E2 - EI)F/RT])aHg - aHg * (2) where z is the charge of the interfering ion. KP2iM can be evaluated as the slope of the graph of {exp[(E2 - EI)F/ R g } a A g - U A ~ against a#. miM can be obtained in a similar manner but with a# above being replaced by a%. Results and Discussion Various macrocyclic polyethers containing the crown ethers 1,4-dithia-12-crown-4 1,7-dithia-12-crown-4 1,4-dithia-15-crown-5 and monobenzo-15-crown-5 and cryptands such as cryptand 222 and 22 were tried as neutral carriers for the Ag+ electrode. The e.m.f. potential responses of PVC - Ag+ electrodes based on these macrocyclic polyethers are illus-trated in Fig. 1. It appears that the crown ethers give a better potential response than the cryptands which show almost no response to the presence of the Ag+ ion.It is well known that macrocyclic polyethers are very sensitive to the size of the metal ion. Among these macrocyclic polyethers cryptand 222 and 22 have the most suitable cavity 150 100 > E Li 50 size (0.28 nm)29 to fit the Ag+ ion which has a size of 0.26 nm.30 Even though both crown ethers and cryptands have very similar cavity sizes cryptands generally can form stronger complexes than crown ethers with the Ag+ ion.2 It is reasonable to predict that both the cryptands studied here will form very strong complexes2 with Ag+ ions which results in a difficulty in exchanging the Ag+ ions of the complexes in the PVC membrane with the Ag+ ions in the test solution.Hence the lack of response with both cryptands can be understood. In contrast as shown in Fig. 1 the electrodes based on crown ethers seem to exhibit a linear response to the activity of Ag+ ions. Among these Ag+ - crown ether electrodes that based on 1,4-dithia-15-crown-5 seems to show the best sensitivity with a slope of 40 k 2 mV pAg-1 and the widest linear range within the concentration range 10-1-10-5 M AgN03. Therefore in this work further studies on the Ag+ ion-selective electrode based on 1,4-dithia-15-crown-5 were carried out. Although the sensitivity with a slope of 40 mV pAg-1 for this electrode is good enough the response is non-Nernstian. The non-Nernstian slopes were observed even though we changed the composition of the membrane (e.g.by adding sodium tetraphenylborate which usually increases the sensi-tivity of ion-selective electrodes) changed the matrix of the membrane by using polystyrene and poly(viny1 acetate) instead of PVC and also changed the concentration of all the components in the membrane and the concentration of the internal solution in the electrode. In contrast the response of the Hg2+ electrode based on crown ethers is Nernstian (30 mV pHg-I) as shown in Fig. 2. In general the non-Nernstian slope was considered to be an “anion effect” or “anion response.”31 According to Boles and Buck’s theory,32 the response slope at 25 “C can be expressed by where UAgC+ and Ux- are the mobilities of AgC+ (the complex of Ag+ and crown ether C) and X- (Nos- in this instance) in the membrane.Only when uAgC+ >> UX- or UX-= 0 is the response slope Nernstian (60/n where n is the charge on Ag+). If the N03- ions are partially extracted as \ A IS O I 0 s w 0’ I I 1 I I 7 6 5 4 3 2 1 PAg + Fig. 1. Potential responses of Ag+ ion-selective electrodes based on various crown ethers and cryptands. (A) 1,4-Dithia-15-crown-5; (B) 1,4-dithia-12-crown-4; (C) monobenzo-15-crown-5; (D) 1,7-dithia-12-crown-4; (E) cryptand 22; and (F) cryptand 222. Internal solution 5 X lop4 M AgNQ + 5 X M KN03. Silver nitrate solution was used as the test solutio ANALYST AUGUST 1986 VOL. 111 I.,.,, 6 5 4 3 2 1 P M 2 + Fig. 2. Potential responses of Hg2+ ion-selective electrodes based on (A) 1,4-dithia-15-crown-5 (B) 1,7-dithia-12-crown-4 and (C) 1,4-dithia-12-crown-4 with the addition to the membranes of sodium tetraphenylborate (STPB) at an STPB to crown ether ratio of 0.5.Mercury(I1) nitrate solution was used as the test solution t Lu 6 5 4 3 2 1 pHg2+ Fig. 3. Effect of sodium tetraphenylborate (STPB) in membranes on potential rewonses of an Hg2+ ion-selective electrode based on i,4-dithia-12-'crown-4 with STPB to crown ether ratios of (A) 0 (B) 0.5 (C) 1.0 and (D) 2.0 AgC+ - NO3- ion pairs into the membrane this results in a reduction in the average mobility of the Ag+ ion or an increase in the average mobility of the NO3- anion. According to equation (3) this effect eventually leads to a reduction in the slope of the calibration graph Ux- # 0. If this is true UAFc+ = 5UNO3- can be obtained from the calculation according to equation (3) and the slope of 40 mV pAg-1.For the mercury(I1) ion-selective electrode as shown in Fig. 2 among the thia crown ethers examined 1,4-dithia-l2-~rown-4 seems to be the best neutral carrier exhibiting a Nernstian response with a slope of 30 k 1 mV pHg-1 and the widest h e a r range within the concentration range 10-6-10-3 M Hg(N03)2. However sodium tetraphenylborate (STPB) must be added to the membranes of all these Hg2+ electrodes, otherwise very poor electrical responses will be observed as shown in Fig. 3A. The addition of STPB results in a large improvement in the response slope. According to the litera-ture,31-33 the addition of STPB to a membrane can increase the electrical conductivity and the sensitivity of the electrode.t UI I I 1 1 1 I I 7 6 5 4 3 2 1 7 6 5 4 3 2 7 P&l+ Fig 4. Effects of (a) crown ether and (b) PVC contents in 5 cm diameter membranes of Ag+ ion-selective electrodes based on 1,4-dithia-15-crown-5. Amount of crown ether (A) 10; (B) 20; (C) 30; (D) 40; (E) 80; and (F) 100 mg t Lu 11° I I I I 1 1 I 1 7 6 5 4 3 2 1 PA^ + . -Fig. 5. Effect of membrane supports on otential responses of Ag+ ion-selective electrodes based on 1,4-ditka-15-crown-5. (A Poly-(vinyl chloride) (PVC); (B) poly(viny1 acetate) (PVA); (C] poly-styrene In addition according to Boles and Buck's theory [equation (3)] when the TPB- anion replaces NO3- in the membrane, because of its higher relative molecular mass its mobility is expected to be lower than that of NO3- and the reduction in the mobility of the anion ( Ux-) 1x1 the membrane can lead to an improvement of the slope.As illustrated in Fig. 3C and D with STPB to crown ether ratios of 1.0 and 2.0 the slopes are super-Nernstian (>30 mV pHg-1). At an STPB to crown ether ratio of 2.0 (Fig. 3D) the response slope reaches 60 mV pHg-1 which is a typical monovalent response and may be attributed to the formation of Hg(TPB)+ ions in the membrane. However Hg2+ elec-trode membranes with an STPB to crown ether ratio >0.5 exhibited poor reproducibility and salting out occurred after about 1 month whereas the membrane with an STPB to crown ether ratio of 0.5 showed good reproducibility and a Nernstian bivalent (Hg2+) response. In contrast the addition of STPB did not cause any improvement and in fact caused a deterioration of the response slopes of the silver( I) ion-selective electrodes based on thia crown ethers.This could be due to the formation of AgC+TBP- ion pairs in the membrane which leads to an increase in the mobility of anions or a decrease in the mobility of Ag+ ions and eventually a reduction in the response slope. Maximum electrical responses are found in all Hg2+ electrodes and some Ag+ electrodes based on thia crown ethers. According to Simon's theory,31 this maximum concen-tration (amax) at the maximum response is inversely propor-tional to the stability constant (KeJ of the complex of the metal ion (ML+) with the ligand (crown ether in this instance) and can be expressed as amax = constant x Kex-l/(z + 1).I 894 ANALYST AUGUST 1986 VOL. 111 t Lu I I 5 4 3 2 1 PH Fig. 6. Effect of pH of test solutions on potential responses of an Hg2+ ion-selective electrode based on 1,4-dithia-12-crown-4. (A) [Hg2+] = 10-3 M; and (B) [Hg2+] = M in solutions + Table 1. Selectivity coefficients ( K T t J for various ions with a silver(1) ion-selective electrode based on 1 .f-dithia- 15-crown-5 KPot Ag M Ion Li + . . . . <10-4 Na+ . . . 5 x 10-2 K+ . . . . 5 x 10-3 Ca2+ . . . . 4 0 - 4 sr2+ . . . . 4 0 - 4 Ba2+ . . . . 4 0 - 4 Mg*+ . . . . 2.8 x 10-2 Fe3f . . . . 0.87 Ion Pb2+ . . . . co2+ . . . . zn2+ . . . . Cd2+ . . . . cu2+ . . . . Ni2+ . . . . Hg2+ . . . . K g M < 10-4 < l o - 4 2.5 x 10-4 4 x 10-4 4 x 10-4 < 10-4 0.77 Table 2.Selectivity coefficients (KFgtM) for various ions with a mercury(I1) ion-selective electrode based on 1,4-dithia-l2-crown-4 Li + Na + K+ Rb+ cs+ Mg2+ Ca2+ Sr2+ Ba2+ Ion K g M . . . . 4 . 7 ~ 10-4 . . . . 7 . 3 ~ 10-4 . . . . 1.2 x 10-4 . . . . 1 . 2 ~ 10-3 . . . . <lo-5 . . . . <10-5 * . . . 7.2 x 10-4 . . . . 1.6 x 10-3 . . . . 6 . 0 ~ 10-5 Fe3 + cu2+ Cd2+ Zn2 + Pb2+ Ni2+ C02+ Ag+ Ion Kf$ M . . . . 3.5 x 10-2 . . . . 8.9 x 10-4 . . . . 4.4 x 10-4 . . . . 7.5 x 10-4 . . . . < 10-5 . . . . 2.5 x 10-4 . . . . 7.6 x 10-4 . . . . <10-5 7 6 5 4 3 2 1 PM Fig. 7. Selectivity measurements for various metal ions with an Ag+ ion-selective electrode based on 1,4-dithia-15-crown-5.[Ag+] = 10-5 M in solutions t UI t Lu ,x-x \ x/ v a/ 0 h C02+ Ag+ VZn l2+ YNi2, I I I 1 I 6 5 4 3 2 1 PM Fig. 8. Selectivity measurements for various metal ions with an Hg2 + ion-selective electrode based on 1,4-dithia-12-crown-4. [Hg2+] = 10-5 M in solutions other words among these thia crown ethers 1,4-dithia-12-crown-4 may form the weakest complex with Hg+ ions and gives the largest value of amax and the widest linear concentra-tion range. Unfortunately insufficient experimental data in the literature could be applied to confirm this suggestion. The effect of the composition of the PVC membranes containing PVC crown ethers and dibutyl phthalate as a plasticiser for both Ag+ and Hg2+ electrodes was also investigated.For example the effects of the contents of crown ether and PVC in a membrane with a diameter of 5 cm for the Ag+ electrode are shown in Fig. 4(a) and ( b ) respectively. As shown in Fig. 4(a) the electrode membranes containing 20-30 mg of crown ether exhibit better responses. In addition the pxA X 6 5 4 3 2 1 P M Fig. 9. Potential responses of an Hg'+ ion-selective electrode based on 1,4-dithia-12-crown-4 for the testing solutions containing (A) Hg(NOd2 (B) HgN03 and (C) HgCl, electrode with 30 mg of crown ether seems to have the lowest detection limit (10-6 M). Similar results were found for the Hg*+ electrode and therefore in subsequent work 30 mg of the crown ether was generally applied. In general the thickness of the membrane depends on the content of PVC in the membrane.A good correlation between the thickness and the PVC content of the membrane was found. As shown in Fig. 4(b) the sensitivity and the detection limit of the electrode seem to increase with a decrease in the PVC content. However if the membrane is too thin it is easily broken. Therefore in this study 100 mg of PVC were used to prepare the membranes for all membranes for both the Ag+ and Hg2+ electrodes. Polystyrene (PS) and poly(viny1 acetate) (PVA) were tried as alternatives to PVC but as shown in Fig. 5 PVC was much better. The effect of the pH of test solutions on these ion-selective electrodes was also studied. As illustrated in Fig. 6 for the Hg2+ electrode the potential responses of 10-3 and 10-4 M Hg2+ in solutions seem to indicate no significant changes at pH > 2 but at pH d 2 the potential responses are drastically reduced.Similar results were also found with the Ag+ selective electrode This could be due to protonation of the crown ether in the membrane at pH d 2 which results in a los ANALYST AUGUST 1986 VOL. 111 895 (a) I T 1 2 3 4 5 6 7 1 2 3 4 5 1 6 7 8 9 1 0 Titrant volumeiml Fig. 10. A plications of Ag+ and Hg2+ ion-selective electrodes based on ditiia crown ethers in titrations of (a) Br- (5 mmol) with Ag+ M); ( c ) Hg2+ (250 mrnol) with I- (0.1 M); and ( d ) Hg2+ (250 mmol) with Cr2072- (0.1 M) M); ( b ) Ag+ (5 mmol) with C1-of their complexing ability with the metal ions (Ag+ or Hg2+). The selectivities of both the Hg2+ and Ag+ ion-selective electrodes were investigated by making potential measure-ments in solutions containing a fixed amount of Hg2+ or Ag2+ ions (10-5 M) and varying amounts of interfering ions.The interferences of alkali metal alkaline earth metal Pb2+ and some common transition metal ions such as Fe3+ Co2+ Ni2+, Cu2+ Zn2+ and Cd2+ were studied. As shown in Figs. 7 and 8, except for Fe3+ these common transition metal ions (M”+) at [Mn+] < 10-3 M in 10-5 M Hg2+ or Ag+ solutions cause virtually no interferences with either the Hg2+ or Ag+ electrodes. The selectivity coefficients (WlM or for various ions (Mz+) were evaluated as the slopes of the graphs of {exp[(E2 - E1)FIRTl}aAg - aAg or {exp[E2 - E#’l RTJ}aHg - aHg against a# or a$ (for the Ag+ and Hg2+ electrodes respectively) as mentioned above and in previous papers.18Jg As shown in Tables 1 and 2 for both electrodes the selectivity coefficients for most metal ions are fairly small (10-2-10-5).In other words most common ions cause only small interferences with either Hg2+ or Ag+ electrodes. The Hg2+ electrode based on 1,4-dithia-l2-crown-4 was very sensitive to the anion as shown in Fig. 9; nearly no potential response to the Hg2+ ion was found for HgC12 (C), whereas a Nernstian response was observed for Hg(N03)~ (A). This could be attributed to the formation of C1- Hg2+ C1- or Hg(OH)+CI- ion pairs34 in HgCI2 solution (C). In addition the Hg2+ electrode also showed a response to the Hg+ ion not only to the Hg2+ ion. However as illustrated in Fig. 9B the electrical response of the Hg2+ electrode to Hg+ ions is non-Nernstian with a slope of 28 mV pHg+-l compared with 30 mV (pHg2+)-1 for the Hg2+ ion.The non-Nernstian (bivalent) response for the Hg+ ion can be understood as this ion is well known to exist as Hg22+ in aqueous solutions. In other words the Hg2+ ion-selective electrode shows a similar degree of response to both Hg2+ and Hg+ ions. Both the Ag+ and Hg2+ ion-selective electrodes were applied as sensors in titrations of Br- with Ag+ and of Ag+ with C1- [Fig. lO(a) and (b)] and of Hg2+ with I- and Cr2072-[Fig. 1O(c) and ( d ) ] . In addition the Ag+ electrode was also applied in the evaluation of the solubility products (K:p) of AgCl in aqueous solution [Fig. 10(b)] as 1.6 x 10-10 which is in good agreement with the literature value35 of 1.7 X lo-*().Because of the low solubilities of the thia crown ethers in water these crown ether - PVC membranes can be used repeatedly for at least 1 month. It can be concluded that both silver and mercury ion-selective electrodes based on dithia crown ethers exhibit good sensitivities detection limits (10-5 M for Ag+ and 10-6 M for Hg2+) reproducibilities and selectivities. The authors express their appreciation to the National Science Council of the Republic of China for financial support of this study. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. References Pedersen C.J. J . Am. Chem. SOC. 1967,89 7017. Christensen J. Eatough D. J . and Izatt R. M. Chem. Rev., 1974 74 351. Frensdorff H. K. J. Am. Chem. SOC. 1971,93 600. Kolthoff I. M. Anal. Chem. 1979 51 1R. Danesi P. R. J. Inorg. Nucl. Chem. 1975 37 1479. Blasius E. Jansen K. P. Adrian W. Klantke G . . Dorschneider R. Mauner G. P. Nguyen V. B. and Strockenen J. Fresenius 2. Anal. Chem. 1977 284 337. Jepson B. E. and Dewitt R. J. Inorg. Nucl. Chem. 1976,38, 1175. Pedersen C. J. and Frensdorff H. K. Angew. Chem. Int. Ed. Engl. 1972 11 16. Hogen-Esch T. E. Kopolow S . and Smid J . Macromol-ecules 1973 6 133. Kopolow K. Machacek Z . Takaki Z . and Smid J. J. Macromol. Sci. Chem. 1973,7 1015. Ryba O. and Petranek J. J . Electroanal. Chem. Interfacial Electrochem.1973,44 425. Petranek J. and Ryba O. Anal. Chim. Acta 1974 72 375. Ryba O. and Petranek J . J. Electroanal. Chem. 1976 67, 321. Tamura H . Kirnura K. and Shono T. J. Electroanal. Chem. 1980 115 115. Kirnura K. J. Electroanal. Chem. 1979 95 91. Shono T. Okahara M. Ikeda I . Kirnura K. and Tamura, H. J . Electroanal. Chem. 1982 132 99. Kamata S . Higo M. Kamibeppn T. and Tanaka I . Chem. Lett. 1982 287. Jeng J. and Shih J. S. Analyst 1984 109 641. Wang D . and Shih J . S . Analyst 1985 110 635. Vogtle F. and Weber E. Angew. Chem. 1974,89 126. Sedvic D. and Meider H. J. Inorg. Nucl. Chem. 1971 39, 1409. Orion Newsl. 1970 9a 10. “Analytical Method Guide,” Fifth Edition Orion Research, Cambridge MA 1972. Oehme F. and Dolzalova L. Fresenius Z . Anal. Chem., 1970 251 1. Bradshaw J. S . Hui J. Y. Haymore B. L. Christensen J. J., and Izatt R. M. J . Heterocycl. Chem. 1973 10 1. Shatkay A . Anal. Chem. 1967 39 1056. Moody G. J. Oke R. B. and Thomas J. D. R. Analyst, 1970 95 910. Srinivasan K. and Rechnitz G. A . Anal. Chem. 1969 41, 1203. Dietrich B. Lehn J. M. and Sauvage J. P. Tetrahedron Lett. 1969 2885. Cotton F. A . and Wilkinson F. R. “Advanced Inorganic Chemistry,” Fourth Edition Wiley New York 1980 p. 14. Simon W. Anal. Lett. 1974 7 9. Boles J. H. and Buck R. P. Anal. Chem. 1973 45 2057. Koryta J . Anal. Chim. Acta. 1977 61 329. Lamb J. D. J. Am. Chem. SOC. 1980 102 3039. Peters D. G . Hayes J. M. and Hieftje G . M . “Chemical Separations and Measurements,” Saunders Philadelphia, 1974 p. 205. Paper A51280 Received July 30th 198.5 Accepted January loth 198

ISSN:0003-2654    

DOI:10.1039/AN9861100891

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, AUGUST 1986, VOL. 111 897 Determination of Copper, Iron, Lead and Zinc in Complex Sulphide Materials by Flame Atomic Absorption Spectrometry James F. Alvin and Frances R. Gardiner CSIRO Division of Mineral Engineering, PO Box 3 12, Clayton, Victoria 3 168, Australia A procedure has been developed that uses atomic absorption spectrometry to analyse sulphide materials for Cu, Fe, Pb and Zn. The measurements are performed on solutions containing ammonium acetate, which is used to dissolve lead sulphate formed by oxidation during decomposition of the samples in the presence of nitric acid. To avoid calibration errors it is necessary to use standard Fe solutions that contain sulphate, prepared from sulphide samples of known Fe content. When more than 0.1% m/m of the metal is present in the sample the relative accuracy of the analysis is within +2%.Keywords: Atomic absorption spectrometry; air - acetylene flame; sulphide materials; ammonium acetate; Cu, Fe, Pb and Zn determination The most important economic sources of Cu, Pb and Zn occur in nature as sulphide deposits.' Iron is always present either as a separate sulphide (e.g., pyrite, FeS2) or in combination with a valuable mineral (e.g., chalcopyrite, CuFeS2). Several sulphide minerals containing these four metals may occur together in a single ore deposit. The quantitative determination of these elements is often required in connection with the normal separation and concentration of the individual minerals and also in research studies of the fundamentals of mineral separation and extractive metallurgy.Atomic absorption spectrometry potentially affords a rapid and convenient method of quantitative analysis for these four elements, and a study has been made of some factors influencing the determination of each individual element. Experimental Apparatus All results were obtained using a double-beam Varian- Techtron Series 875 atomic absorption spectrometer in conjunction with Varian-Techtron hollow-cathode lamps. The instrument was operated in the concentration mode and is capable of accepting up to five standards for calibration and curve correction. Dissolution of Sulphide Materials The most common method of dissolving ores and products containing high concentrations of sulphide minerals is to use an oxidising attack such as bromine - nitric acid or aqua regia.2 During such a decomposition sulphide is oxidised to sulphate and, if lead is present, lead sulphate is formed.Lead sulphate is almost insoluble in acids but it can be rendered soluble by EDTA3 or ammonium acetate. If the pH is adjusted to approximately 5 with ammonium acetate, lead sulphate is readily soluble and iron can be held in solution by complexing it with tartaric acid. Under these conditions analyses for Cu, Fe, Pb and Zn can be performed on a single solution. Inter-element Effects To investigate the possible inter-element effects, solutions of the individual elements were spiked with the four other major elements likely to be present in a complex sulphide ore. The concentrations of both analyte element and interferent were chosen at two levels.The first of these levels was equivalent to a high concentration of the element of interest (20 p.p.m.) spiked with 20 p.p.rn. of a second element. The second level represented a low level of concentration of the element of interest (2 p.p.m.) in a matrix of a typical concentrate of a second element ( i e . , several hundred p.p.m.). Copper was determined at a high level of 20 p.p.m. in solutions containing separately Fe, Pb, Zn and S at the 20 p.p.m. level. Using the conditions set out under Procedure, this would represent four samples containing 20% of Cu and 20% of one of the other four elements. Sulphuric acid diluted to the appropriate concentration was used for S spiking. No interference in excess of normal experimental error occurred and these results are typical of those obtained also with both Pb and Zn.However, Fe presented problems as the other elements interfered in the Fe determination. When a lean air - acetylene flame was used with the optical path length of 3 mm above the burner, spiking four separate 25 p.p.m. Fe solutions prepared from pure Fe powder with 20 p.p.m. of Cu, Pb, Zn and S solutions gave apparent concentrations of 27.0, 25.6, 26.2 and 34.3 p.p.m. of Fe, respectively. At an Fe concentration of 2.5 p.p.m., larger apparent concentrations were measured in the presence of each of the other elements. Altering the flame composition and/or the burner height resulted in larger divergences from the expected values. After substituting the dinitrogen oxide - acetylene flame, matrix effects became less troublesome. However, interferences were still at an unacceptable level, especially at low concentrations of Fe and high concentrations of the other elements.Roos and Price4 have also noted the critical setting of both burner height and flame composition when using an air - acetylene flame. Two methods were tried to overcome these interferences. When the burner height was lowered in small increments, a position could be found where an Fe standard prepared from pure Fe metal gave an absorbance reading identical with that of a sulphide sample of the same Fe concentration. A better alternative was to prepare calibration standards from sulphide samples whose Fe concentration had been determined by potassium dichromate titration. A range of standards covering 5-25 p.p.m.of Fe were prepared by either weighing two different masses of one or more samples, or choosing a number of samples with varying Fe concentrations. Although all four elements enhanced the Fe signal, experience showed that sulphate ion was of crucial importance and if it was present in the calibration standards, most interferences disappeared. If higher accuracy for iron is required a dinitrogen oxide - acetylene flame is recommended, still using sulphide samples for calibration standards.898 ANALYST, AUGUST 1986, VOL. 111 Table 1. Precision of AA measurements over a 10-min time span Cu - Fe Pb Zn x . . . . 9.04 11.69 19.44 11.28 N . . s . . . . 0.09 0.16 0.17 0.11 . . 10 10 10 10 Ammonium Acetate Solutions Ammonium acetate solutions are viscous relative to most dilute aqueous solutions and it would be expected that this could affect the transport efficiency from the sample to the nebuliser.When the calibration standards were prepared in a 2% V/V ammonium acetate matrix, typical results for un- known samples containing 20% ViV ammonium acetate (i.e., no dilution), 5% V/V ammonium acetate (1 + 3 dilution) and 0.5% VlV ammonium acetate (1 + 39 dilution) were -8, -2 and +2%, respectively, relative to the correct result. It is clear that, without dilution, both standards and unknowns must be matched in ammonium acetate concentration. However, it is doubtful if the errors at the 1 + 3 and 1 + 39 dilutions were significant as they just fell within the 95% confidence level of the accuracy of the measuring system.Precision The statistics mode of the spectrometer allows the recording of the mean (m, standard deviation (S) and the number of readings ( N ) for any given sample. Using this mode and taking readings at 1-min intervals, the spread of results that is likely over a 10-min time span was established (see Table 1). At the 95% confidence level an instrument relative error of 22% can be expected. At high concentration levels the elements listed in Table 1 can be titrated with higher precision. However, atomic absorption offers a great advantage in speed, and if this compromise with precision can be tolerated, much time and effort can be saved by using this technique. Testing The following samples were used to set up a programme to test the validity of the method.(a) Two international standard materials: CANMET Zn concentrate CZN-1 and CANMET Zn - Sn - Cu - Pb ore MPla. ( b ) Two research samples, one high-grade chalcopyrite sample and one high-grade pyrite sample. All group (b) samples were analysed by well known and accepted titration methods. (c) Three yearly composite samples from a Pb - Zn mine. The results given by the mine were taken as accepted values. (d) Synthetic samples, produced by separately weighing different masses from groups ( a ) , ( b ) and ( c ) . A range of samples were obtained to cover most of the likely combinations found in complex Cu, Fe, Pb and Zn sulphide materials. Table 2 compares the results found by atomic absorption measurements against the accepted values. Analytical Method Reagents Ammonium acetate solution.Mix in the following order equal volumes of distilled water, 53% mlV analytical-reagent grade ammonia solution and 100% mlV glacial acetic acid, producing 39% mlV ammonium acetate solution. Calibration Standards All calibration solutions are prepared in a base solution of 2% V/V ammonium acetate and 0.04% miV tartaric acid. Lead. Dissolve 500 mg of pure Pb foil in 15 ml of dilute nitric acid (1 + 3) and evaporate to dryness. Dissolve the residue in distilled water and transfer into a 1-1 calibrated flask; 1 1 of solution z500 p.p.m. of Pb. Take 2.5,5,10 and 25 ml, transfer into 250-ml calibrated flasks, add 5 ml of ammonium acetate solution, 10 ml of a 1% miVsolution of tartaric acid and dilute to 250 ml with distilled water.These solutions represent 5 , 10, 20, 35 and 50 p.p.m. of Pb, respectively. Zinc. Dissolve 250 mg of pure Zn metal in 15 ml of dilute nitric acid (1 + 3) and evaporate to dryness. Dissolve the residue in 50 ml of warm (ca. 50 "C) dilute hydrochloric acid (1 + 1) and transfer into a 1-1 calibrated flask; 1 1 of solution z250 p.p.m. of Zn. Take 1 , 4 , 8, 12 and 16 ml, transfer into 250-ml calibrated flasks, add 5 ml of ammonium acetate solution, 10 ml of a 1 o/o m/V tartaric acid solution and dilute to 250 ml with distilled water. These solutions represent 1, 4, 8, 12 and 16 p.p.m. of Zn, respectively. Copper. Dissolve 250 mg of copper foil in a similar manner to Zn, preparing standards of 1 , 5 , 10,15 and 20 p.p.m. of Cu. Iron. Determine, by potassium dichromate titration, the iron content of several unknown sulphide samples.Using these samples as calibration standards, process them in the same way as under Procedure. Take sample masses so that when diluted 1 + 9, they will represent nominally 5, 10, 15,20 and 25 p.p.m. of Fe. Procedure Weigh a nominal 250-mg sample into a 150-ml beaker and dissolve it in 15 ml of aqua regia or 15 ml of dilute nitric acid (1 + l), and a few drops of bromine, depending on which solvent is more efficient for the sulphide in question. Evaporate to dryness and dissolve the residue in 10 ml of hot (ca. 90 "C) dilute hydrochloric acid (1 + 2). Add 10 ml of 10% m/V tartaric acid solution followed by 50 ml of ammonium acetate solution. Stir to ensure that all Pb salts have dissolved and transfer into a 250-ml calibrated flask.Dilute to the mark with distilled water, mix and filter approximately 50 ml through a Whatman No. 42 paper, collecting the filtrate in a dry beaker. Dilute 1 + 9 or 1 + 19 and measure the concentration of each element by atomic absorption spec- trometry. If the sample is suspected to contain less than 1% of the element being determined, substitute a nominal 1000 mg sample mass, add 20 ml of ammonium acetate solution, 10 ml of 10% mlV tartaric acid solution, transfer into a 100-ml calibrated flask, dilute to the mark with distilled water and dilute 1 + 9. Atomic Absorption Measurement A lean air - acetylene flame with the image of the hollow- cathode lamp 3 mm above the centre of the burner is used for all readings. At least three standards are required for calibration with the highest standard reading between 0.68 and 0.72 absorbance unit.With Cu, Fe and Zn this is achieved by rotating the burner to reduce the number of absorbing atoms. Depending on the concentration being measured, two ranges of cali- bration standards may be required, using the three lowest standards for one range and the three highest standards for the other. Element concentrations are read using the concentra- tion mode of a Varian Model 875 atomic absorption spec- trometer, and the above range of standards ensures both high sensitivity and linear calibration points. Measurements of Cu,ANALYST, AUGUST 1986, VOL. 111 899 Table 2. Comparison between AA results and expected values No 1 2 3 4 5 6 7 8 c u , Yo Fe, Yo Pb, Yo Zn, o/o * Sample identification Expected CANMET CZN- 1 ., . , . . 0.144 CANMET MPla . . . . . . 1.44 Mine Pbconcentrate . . . . 0.903 Mine residue . . . . . . . . 0.013 Chalcopyrite . . . . . . . . 33.9 Pyrite . . . . . . . . . . <0.01 Research sample . . . . . . 20.4 Research sample . . . . . . I1 .0 1 + 3 . . . . . . . . . . 0.52 1 + 5 . . . . . . . . . . 16.9 2 + 5 . . . . . . . . . . 17.7 1 + 3 + 4 . . . . . . . . . . 0.35 3 + 4 + 5 . . . . . . . . . . 11.6 1 + 3 + 5 . . . . . . . . . . 11.7 Found 0.15 1.48 0.91 0.015 34.8 20.8 11.3 0.5 17.2 17.9 11.9 12.0 - 0.41 Expected 10.93 5.97 1.83 - 30.3 45.2 18.7 10.1 6.4 20.5 18.2 - - 14.3 Found 10.4 6.09 1.84 - 32.5 45.9 18.4 10.0 20.2 17.5 6.29 - - 14.0 * A dilution of 1 + 39 is necessary for zinc when its concentration exceeds 30% mlm.Expected 7.45 4.33 75.6 0.34 <0.01 <0.01 11.8 38.7 41.4 3,76 2.16 27.9 25.4 27.8 Found 7.73 4.56 0.35 77.2 - - 12.4 39.6 41.7 4.0 2.27 28.4 25.9 28.4 Expected 44.7 19.0 3.19 0.52 0.022 <0.01 - - 24.1 22.6 9.5 16.1 1.25 15.9 Found 45.5 19.4 3.3 0.54 - - - - 24.4 23.0 9.5 16.3 1.34 16.1 Fe, Pb and Zn are made at 324.8,248.3,283.3 and 213.9 nm, respectively. 1. Conclusions 2. Atomic absorption spectrometry offers a rapid method of A wide range of concentrations from 0.1% mlm to the maximum concentration occurring in pure sulphide minerals for the four elements can be measured. Measurements of lower values are possible, but because of the high viscosity of ammonium acetate, a separate calibration is necessary to counteract the reduction in nebulisation of the undiluted acetate solution. determining Cu, Fe, Pb and Zn in complex sulphide materials. 3. 4. References Read, H. H., “Rutley’s Elements of Mineralogy,” Twenty- sixth Edition, Thomas Murby, London, 1970, pp. 246,299 and 456. Dolezal, J . , Povondra, P., and Sulcek, Z., “Decomposition Techniques in Inorganic Analysis,” Iliffe, London, 1968, pp. 46-59. Fernandez Sanchez, M. L., Garcia Ortiz, C., Arribas Jimeno, S., and Sanz-Medel, A., A t . Spectrosc., 1984, 5, 198. Roos, J . T. H., and Price, W . J., Spectrochim. Acta, Part B, 1971, 26, 279. Paper A51454 Received December 18th, 1985 Accepted February 17th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9861100897

出版商:RSC    

年代:1986

数据来源: RSC