ISSN:0003-2654    

DOI:10.1039/AN98207FX041

出版商:RSC    

年代:1982

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98207BX043

出版商:RSC    

年代:1982

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98207FP113

出版商:RSC    

年代:1982

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98207BP117

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

November 1982 The Analyst Vol. 107 No. 1280 pH Cells for Over-all Temperature Compensation in the Measurement of the pH of Boiler Feedwater D. Midgley and K. Torrance Central Electricity Research Laboratories, Kelvin Avenue, Leatherhead, Surrey, KT22 7SE Strict control of the water chemistry in modern high-pressure boilers often requires that the measured pH of boiler feedwater is referred to a standard temperature of 25 "C. At present most of the pH meters installed in power stations cannot meet this requirement unless the sample temperature is controlled to 25 OC, because their temperature compensation circuits cannot correct for temperature-induced changes in the pH of alkaline feedwater. These changes can be considerable because the temperature coefficient of feedwater is of the order of -0.033 pH O C - 1 .Experimental glass electrodes have been developed which, when used in conjunction with a specified reference electrode, can provide over-all tempera- ture compensation in ammonia-dosed feedwater over the range 15-35 "C. The chemistry of these electrodes has been arranged such that the cell e.m.f. only responds to changes in pH brought about by changes in alkalinity. At constant ammonia concentrations equivalent to pH values in the range 9-9.3, the temperature-induced variations in the cell e.m.f. over the tempera- ture range 15-35 "C were equivalent to less than 0.05 pH for a pH electrode containing N-glycylglycine in its internal reference solution and using silver - silver chloride/O. 1 moll-' potassium chloride or calomel/3 mol 1-1 potassium chloride reference electrodes.Keywords : PH; temperature compensation ; glass electrode ; reference electrode ; boiler feedwater The CEGB specification for the measurement of the pH of ammonia-dosed feedwater in once-through boilers requires that the pH is referred to a standard temperature of 25 "C. This has been considered necessary because under conditions of varying sample tempera- tures the measurement of pH by instruments which display the correct pH at the temperature of the sample will record changes in pH that are due solely to the effects of temperature on the chemical equilibria in the sample. These changes can be considerable because the temperature coefficient in boiler feedwater is of the order of -0.033 pH "C-l. Without additional information, plant operators can be in doubt as to whether a difference in pH arises from a change in alkalinity or a change in sample temperature.Most pH meters installed in plant display the correct pH a t the temperature of the sample. What is required is a meter that will display the pH at the standard temperature of 25 "C and for our purposes such a pH measuring system would be described as having over-all temperature compensation. This paper describes experiments in which the temperature-dependent properties of glass pH electrodes and silver - silver chloride and calomel reference electrodes were investigated with the aim of producing a pH cell that was self-compensating for the measurement of pH to a reference temperature of 25 "C in ammonia-dosed feedwater over the range 15-35 "C.The e.m.f., and hence the indicated pH, of such a cell would be independent of temperature under conditions of constant ammonia dosage. As such it would have distinct advantages over existing methods of providing over-all temperature compensation that depend on a measurement of temperature and the application of a linear correction factor to a temperature- dependent e.m.f. obtained from the pH cell. 12971298 Over-all Temperature Compensation potentials of the glass (Eglass) and external reference (Eref) electrodes : MIDGLEY AND TORRANCE : pH CELLS FOR TEMPERATURE Analyst, VoZ. I07 The e.m.f. of a pH cell (Ecell) is conventionally written as the difference between the Ecell = Eglass - Eref .. .. .. - * (1) Equation (1) can be expanded to include terms for the e.m.f.of the internal reference electrode (Eint), which is inside the glass electrode, and the pH of the solution in which it is immersed (pHint), together with the pH of the external solution (PHext) and the potential of the liquid junction (E,) : Ecell = (Eint + kpHint)glass - KpHext - Eref + EJ - - * * (2) where k is the Nernst coefficient, equal to 2.303RTIF. When the temperature of the pH cell changes, each of the terms in equation (2) alters and the temperature dependence of the pH cell may be expressed by differentiating equation (2) with respect to temperature : dEce11 - (ddE;Ft + kdpHint dk kdpHex t dk dEref dEj dT pHl?xt'--- + dT (3) dT dT + p H i n t ' z ) glass -~ - d T d T - No attempt was made to compensate for the temperature dependence of the liquid junction as it is considered to be very sma1l.l On examination of equation ( 3 ) , it can be seen that, excluding the term dE,/dT, there are three pairs of terms (dE/dT, kdpH/dT and pH-dk/dT) each pair of which is identical except for sign and subscript.The most direct approach to self-compensation was first attempted by equating the terms in each of these pairs. For the measurements in feedwater, this requires that the pH of the filling solution inside the glass electrode is typical of that of feedwater and also has a similar temperature coefficient (dpH/dT). The value of the latter depends to some extent on the temperature dependence of the dissociation constant of the base, but in dilute solutions the variation of the dissociation constant of water with temperature dominates the relation- ship.Ammonia was an obvious choice because it is the base most commonly added to control the pH of feedwater. Two other bases, tris(hydroxymethy1)- aminomethane (Tris) and N-glycylglycine, were selected because calculations indicated that dpH/dT in their solutions is very close to that expected in feedwater (-0.033 pH "C-l). In addition, these bases are non-volatile and as such could have some advantage over ammonia, which might be gradually lost from the internal filling solution of the glass electrode. The two terms in equation (3) of the general form dE/dT involve the temperature depen- dence of the reference electrodes and as the silver - silver chloride electrode is used almost exclusively as the internal reference electrode in glass electrodes, it was the first choice as the external reference electrode.Experiments were carried out using silver - silver chloride electrodes in contact with 3 moll-' potassium chloride as internal and external reference electrodes in the same cell. Subsequently, attempts to equate the terms dE/dT were dis- continued and external electrodes differing in either concentration of electrolyte or type were used in an attempt to achieve the desired compensation, i.e., of dEcell/dT = 0. The most successful results were obtained with calomel (3 rnol 1-1 potassium chloride) and silver - silver chloride (0.1 mol 1-1 potassium chloride) electrodes. Three bases were used. Experimental Apparatus and Procedure A stream of de-ionised water was dosed with ammonia by diffusion of the gas through a length of silicone-rubber tubing from concentrated ammonia solution. Under conditions of constant temperature, flow-rate of de-ionised water and diffusion of ammonia, ammonia solutions of constant conductivities, comparable to those observed in feedwater, were produced.The temperature of these simulated feedwaters could be adjusted to known values by passing them through a temperature-controlled heat exchanger before measuring the changes in e.m.f. they produced in a number of pH electrodes contained in a flow cell. The apparatus was housed in an air-conditioned room whose temperature was controlled at 25 1 "C. Dilute ammonia solutions were prepared using the apparatus shown in Fig. 1.November, 1982 COMPENSATION IN MEASURING pH OF BOILER FEEDWATER 1299 In these investigations the pH values of the ammonia solutions were calculated from accurate measurements of their temperatures and conductivities.In order to facilitate simultaneous readings of temperature, conductance and the e.m.f. of the pH cells, signals from sensors for each of these parameters were fed to the central data acquisition and process- ing system (CDAPS) at CERL. These measurements were logged on a PCD-DCH70/72 cassette recorder and could be simultaneously displayed on a VDU. The output impedance of the conductance bridges and the impedance of the pH cells were too high to be fed direct to the measuring system and it was necessary to interpose an amplifier between these signal sources and CDAPS.Conductivity 3-way Electrodes cell valve (to amplifier) De-ionised water 200 mi min-' rubber 1 coil w To de-ionisation unit Fig. 1. General layout of apparatus. Measurement of Temperature The temperature of the stream of dilute ammonia solution was monitored at two points: directly after the ammonia diffusion cell and inside the conductivity cell that precedes the pH cell (see Fig. 1). Both of these measurements were used in the calculation of pH from the temperature - conductivity relationship in dilute ammonia solutions. In both of these positions, the temperature was sensed by a 100-kQ precision thermistor (Type YSI 44011, Swift-Sasco, Crawley). In the first position the thermistor was mounted in a small glass flow cell through which the total output of about 200 ml min-l from the diffusion cell flowed.The second thermistor was incorporated inside the conductivity cell by the manufacturer, where it measured the temperature of that portion of the total flow that was directed through the pH cell. The thermistors were powered from a 9.1-V d.c. power supply and the voltage across a 10042 resistor, placed in series with the thermistor, was measured by CDAPS. At 25 "C this was of the order of 9 mV. The thermistors were calibrated in flowing solution over the temperature range 15-35 "C against certificated mercury-in-glass thermometers. The measured voltages were fitted to an equation of the general form mV = aebT, where a and b are constants. The equations so obtained had correlation coefficients greater than 0.99 and temperatures could be measured with an accuracy of better than 0.1 "C.Measurement of Conductivity The conductivity of the solutions could be measured at three places (see Fig. 1) : directly before and after the ammonia diffusion cell and immediately before the pH cell. All of the conductivity cells were made from stainless steel and were of the flow-through type. Those on either side of the ammonia diffusion cell were Type EFD 001 (cell constant ca. 0.01 cm-l) and that in the water-bath Type ESC/OO5/2OOE/lOOK (cell constant ca. 0.05 cm-l), both types being manufactured by Electronic Instruments Limited (EIL). Accurate values of the cell constants of these cells were obtained by calibrating them against a standard conductivity cell in a flowing solution whose specific conductivity was comparable to that of feedwater.The conductances of the solutions were measured using Wayne Kerr Auto-1300 MIDGLEY AND TORRANCE : pH CELLS FOR TEMPERATURE AlzaZyst, VoZ. 107 balance Universal Bridges, Model B642 (Wilmot Breeden Electronics Limited, Bognor Regis) with an accuracy of the order of 0.1%. The conductances of the dilute ammonia solutions were monitored and changes in conduc- tivity were registered (by CDAPS) as changes in the potentiometric recorder output from the Wayne Kerr bridge. Each bridge was calibrated on the range applicable to the conductivity of feedwater (about 2-8 pS cm-l) using a standard resistance box; thus the voltage measured by the data logging system could be converted to a conductance by a linear equation of the form pS = a(mV) + c, where a and c are constants.Using this procedure, conductances could be measured with an error of less than 0.5%. Measurement of the E.m.f. of pH Cells It was decided to control the temperature of the pH cell at as near to 15, 20, 25, 30 and 35 "C as possible, as most of the constants used in calculating the pH from measurements of conductivity had been accurately determined at these temperatures. The simplest way to obtain temperature control to better than kO.1 "C was by the use of a high-quality pro- portional heater such as a Tecam Model TU-14 in an insulated water-bath. Some difficulty was experienced in measuring the e.m.f. due to the presence of earth loops, as the dilute ammonia stream was connected to the earth terminal of the mains electrical supply through the stainless-steel heat exchanger and metal conductivity cell immersed in the water-bath.These earth loops were eliminated by using a differential potentiometric measurement with the water-bath (or more correctly the water-bath potential) as the common point. A Keithley, Model 604, differential amplifier made the high-impedance e.m.fs from the pH cells acceptable to the CDAPS measuring system. The stability of this system was such that the standard deviation of the e.m.f. measured in buffer solutions over a period of 10 min was of the order of 0.03 mV. In simulated feedwater the precision was considerably reduced, but even then it was usually less than 0.5 mV when measured over a similar time period. During a logging sequence] the e.m.fs of all of the electrodes (glass and reference) were measured with respect to one reference electrode.Subsequently, the e.m.f. of any glass - reference electrode combination could be obtained by difference. Experimental Glass Electrodes Refillable glass electrodes were first tried but these proved unsuitable because the in- adequacies of the electrical screening gave high signal noise and poor stability. The alternative approach, which was completely satisfactory, was to obtain from EIL a supply of unfilled glass electrodes of the same type of glass as that used in their current range, and to fill these with the appropriate experimental internal reference solutions. These solutions contained 3 mol 1-1 potassium chloride saturated with silver chloride, together with a base present at a concentration in the range 10-3-10-2 moll-l.The pH of these solutions was adjusted] where necessary, to 9.0 & 0.1 by the addition of potassium hydroxide solution. External Reference Electrodes The basis of designing a self-compensating pH cell was that the pH and external reference electrodes would be subjected to the same temperature changes. This requires that both of these electrodes are immersed in the sample solution, unlike some industrial pH cells where a remote junction is favoured. A Corning Type 476029 silver - silver chloride reference electrode, which had a side-arm for connection to a remote reservoir, was emptied of its 4 mol 1-1 potassium chloride solution and refilled with 3 mol 1-1 potassium chloride solution saturated with silver chloride. This electrode was allowed to equilibrate with its new filling solution for about 10 d until its potential became constant.A second silver - silver chloride electrode of similar construction was used but in this instance the original electrolyte was replaced with 0.1 moll-1 potassium chloride solution. The electrode was again allowed to attain a steady potential before being used. An EIL RJ23 calomel electrode, also modified to take a remote reservoir, was the third reference electrode used. This was filled with 3 moll-1 potassium chloride solution. All of the electrodes were used in conjunction with remote reservoirs providing about 500 mm head of electrolyte solution and in every instance the liquid junctions were formed at ceramic frits.November, 1982 COMPENSATIOK IN MEASURING pH OF BOILER FEEDWATER 1301 pH Flow Cell The pH flow cell was made from Perspex and was designed for use with eight electrodes (see Fig.2). The cross-section of the cell was rectangular, 50 mm wide by 40mm high, and the overall length was 122 mm. The electrodes were mounted vertically in the cell and held in position in tapered holes by silicone-rubber bungs, cast to fit the taper of the holes. A support lid was held in position on top of the flow cell by eight 4BA studs (see Fig. 2). This rectangular piece of 10-mm Perspex sheet was necessary to prevent any movement of the stems of the electrode that could lead to ingress of water from the water-bath. The inlet and outlet of the cell were $-in Drallim bulkhead couplings (Drallim Tube Couplings Limited, Bexhill-on-Sea), which were connected by 50-mm lengths of $-in PTFE tubing to Chemcon three-way PTFE valves (Production Techniques Limited, Fleet, Hampshire).These valves were necessary to isolate that part of the circuit containing the dilute ammonia solution from the flow cell during standardisation with buffer solutions. Three reference Salt bridge electrodes tube to remote Fig. 2. Perspex flow cell. Operating Procedure The ammonia diffusion cell was designed to produce solutions of constant conductivity at an accurately known temperature of about 25 "C at a flow-rate of about 200 ml min-l. A portion of the total flow, usually 15-20 ml min-l, was passed through the heat exchange coil, conductivity cell and pH flow cell, all contained in the temperature-controlled water-bath.The apparatus was run overnight to reach constant conditions of conductivity and tempera- ture. During this period the temperature of the water bath was set at 15.0 "C, the lowest temperature of the range investigated. In the morning, readings were logged at l-min intervals for a period of 30 min. The temperature in the water-bath was then re-set to 20.0 "C and after approximately 90 min data were again logged for 30 min. This procedure was repeated at 5 "C intervals up to 35 "C. Only small differences (0.1-0.2 mV) were observed between the e.m.fs in flowing and static buffer solutions. Because of the operational simplicity of the latter, the pH electrodes1302 MIDGLEY AND TORRANCE: pH CELLS FOR TEMPERATURE ArcaZyst, VoZ.107 were standardised in the flow cell using static buffer solutions. During this procedure the flow cell was isolated from the circuit containing dilute ammonia solution by the use of the Chemcon three-way valves. The same temperature sequence described above was followed for the buffer solutions. At each temperature the responses of the electrodes were monitored and logged. Calculation of pH from the Conductivity of Dilute Ammonia Solutions insignificant errors arise from the use of limiting equivalent conductances and the assumption that the activity coefficients are unity. In the conditions described for the diffusion cell, only three ionic species are considered to be present and, therefore, the specific conductance, K (pS cm-l), can be written as At the ionic strengths considered here ( I < 5 x where the terms in square brackets are the concentrations (moll-l) and the A" terms are their corresponding limiting equivalent conductances (cm2 S equiv.-l).From consideration of the equations for electroneutrality and the dissociation constant of water, Kw, equation (4) can be rearranged and written as Equation (5) can be solved for [H+] (the smaller positive root). At the temperatures of the experiments, values of the limiting equivalent conductances were obtained directly or inter- polated from those given by Robinson and Stokes2 Values for K, at the same temperatures were calculated from the following polynomial : - log& = 4471*33 - - 6.0846 + 0.017054 T T where T is the absolute temperature.Results Initial Stability of Experimental Electrodes At intervals over a few weeks before starting the experiments in simulated feedwater, the e.m.fs of three experimental electrodes vcmm a Radiometer KlOO saturated calomel electrode were measured in standard phosphate and borax buffers. Between measurements, the electrodes were stored at 25 "C in a small volume of distilled water. The variations in e.m.f. in phosphate buffer for the N-glycylglycine-, ammonia- and Tris-filled electrodes are shown in Fig. 3. Changes in e.m.f. could be due to a number of factors: variations in the boundary potentials of the sensing glass, changes in the standard potential of the internal reference electrode or changes in the pH of the internal filling solution.None of these can be measured directly. The e.m.f. of an unused EIL E,7 glass electrode was monitored in a similar manner and, as this electrode was the same as the experimental electrodes except for the filling solution and had been assembled for at least 12 weeks before use, changes in its e.m.f. were taken as a measure of the changes which occur following hydration of the external surface of the glass. Initially its glass sensing surface was dry and on contact with aqueous solution small changes in the electrode potential were observed, much smaller than the changes observed with the experimental electrodes over the same period of time. Therefore, it was probable that the changes that occurred with the experimental electrodes were due to changes in the standard potentials of the internal reference electrodes or changes in the pHs of the internal filling solutions.However, after a period of about 3 weeks (see Fig. 3), the rate of change of potential of all of the experimental electrodes was only about -0.5 mV d-l. This was acceptable for the temperature compensation experiments as measurements at any one level of ammonia were completed within 24 h. Slope Factor of the Experimental Glass Electrodes During the sequence of tests with pure ammonia solutions the slope factors, k = 2.303RT/F, of the glass electrodes were determined about once per week at 15, 20, 25 and 35 "C byNovember, 1982 COMPENSATION IN MEASURING pH OF BOILER FEEDWATER 1303 measuring the e.m.fs in NBS standard borax and phosphate (1 + 1) buffer solutions.These measurements were made in static solutions in the flow cell and the average reading calculated from ten measurements a t 1-min intervals. The values of k obtained for all of the experi- mental electrodes were always greater than 99% of the theoretical values at 25, 30 and 35 "C. At the two lowest test temperatures (20 and 15 "C), values slightly less than 99% were found, particularly at 15 "C where a value as low as 98.4% was calculated for the Tris- filled electrode on one occasion. 7 14 21 Days after assembly Fig. 3. Variation of e.m.f. in phosphate buffer of experi- mental pH electrodes versus S.C.E. A, Tris-filled; B, N-glycylglycine-filled ; C, ammonia filled ; and D, EIL Ell7 Stability of Silver - Silver Chloride Electrodes in Alkaline Solutions The internal silver - silver chloride electrodes in conventional glass pH electrodes are in contact with neutral or acidic solutions containing chloride ions and under these conditions the internal reference electrode potential is stable.In the experimental glass electrodes the internal filling solutions were alkaline and although the stability was satisfactory for measure- ments over a 24-h period additional information was required on the variation of their e.m.f.s over a longer period. As it was not possible to carry out unambiguous tests on these internal reference electrodes because they were part of the glass electrodes, separate experi- ments were made using chloridised silver wire electrodes. Pairs of silver - silver chloride electrodes were tested for long-term stability in solutions of 3 mol 1-1 potassium chloride containing either ammonia or N-glycylglycine (ca.10-3- 10-2moll-1) adjusted to pH 9. The performance of these electrodes was compared with that of two similar electrodes in neutral 3 mol 1-1 potassium chloride solution. All of these solutions were saturated with silver chloride. Pairs of electrodes and their respective solutions were sealed in 50-ml Pyrex glass flasks and their e.m.f.s measured at intervals over a period of 3 weeks by introduction of a separate reference electrode. Their variations in e.m.f. over this period are shown in Fig. 4 and it can be seen that no significant differences were detected between the variations in e.m.f. in the neutral and basic potassium chloride solutions.Further, comparison of the variations in e.m.f. for the two alkaline solutions with those of the experimental glass electrodes in Fig. 3 suggests that the internal filling solutions and their associated silver - silver chloride electrodes were not the major cause of the initial irregular behaviour. Temperature Dependence of pH Cells in Flowing Solutions of Ammonia The variations of the e.m.f. of pH cells consisting of the experimental glass electrodes versus a stated reference electrode in flowing ammonia solution were measured over the temperature range 15-35 "C. During any one run the ammonia level was kept constant at a concentration1304 MIDGLEY AND TORRANCE: pH CELLS FOR TEMPERATURE AnaZyst, VoZ. 107 giving a pH in the range 8.6-9.3. Duplicate experiments were carried out at each ammonia level, although owing to the nature of the apparatus it was not always possible exactly to reproduce the ammonia levels.~ 7 14 21 Days after assembly Variation of e.m.f. of pairs (0, A) of silver - silver chloride electrodes in 3 mol 1-1 potassium chloride solution and bases: A, no base; B, ammonia (pH 9); and C, Tris (pH 9). :Results for each pair in solution were normalised with respect to the mean of the first readings of e.m.f. Fig. 4. Temperature Dependence of pH Cells with a Silver - Silver Chloride/3 moll-1 Potassium Chloride External Reference Electrode The temperature dependences of p H cells with ammonia-, Tris- and N-glycylglycine-filled experimental glass electrodes were measured in solutions containing from 0.09 mg 1-1 (pH 8.59) to 0.47 mg 1-1 (pH 9.17) of ammonia.This combination of glass and reference electrodes did not produce the temperature independence expected from the symmetry of the cell. The general trend in results for any one electrode combination was the same at all ammonia levels. Examples of the typical behaviour are given in Figs. 5 and 6. It was not possible to say whether the positive coefficient arose because the temperature coefficient of the glass electrode was too high or that of the external reference too low. In order to resolve this uncertainty, tests The pH cells had temperature coefficients of 0.5-0.6 mV "C-l. t + E ui Io 15 25 35 Temperatu re/"C Fig. 5. Temperature de- pendence of experimental pH electrodes versus silver - silver chloride/3 mol 1-1 potassium chloride external reference electrode in ammonia solution; pH = 8.59 a t 25 "C.A, Ammonia- filled; B, Tris-filled; and C, N-glycylglycine-filled pH electrode. I 15 25 35 Tern peratu re/"C Fig. 6. Temperature de- pendence of experimental pH electrodes versus silver - silver chloride/3 mol 1-1 potassium chloride external reference electrode in ammonia solution; pH = 9.17 a t 25 "C. A, Ammonia- filled; B, Tris-filled; and C, N-glycylglycine-filled pH electrode.November, 1982 COMPENSATION IN MEASURING pH OF BOILER FEEDWATER 1305 were made with a calomel/3 mol 1-1 potassium chloride reference electrode whose tempera- ture coefficient (ca. 0.5 mV "C-l) is higher than that of the silver - silver chloride electrode.1 Temperature Dependence of Cells with a Calomel/3 moll-1 Potassium Chloride External Reference Electrode In general, the cell temperature coefficients were much smaller than those obtained with the external silver - silver chloride/3 mol 1-1 potassium chloride reference electrode, con- firming that the coefficients of the glass electrodes were higher than those calculated from equation (3).The results shown in Fig. 7 indicate that over-all compensation was essentially achieved, with a variation in e.m.f. equivalent to only 0.05 pH over the span of 20 "C. Temperature Dependence of Cells with a Silver - Silver Chloride/O.l moll-1 Potassium Chloride External Reference Electrode Although a pH cell containing a calomel/3 mol 1-1 potassium chloride eFternal reference electrode gave essentially the desired over-all temperature compensation, a silver - silver chloride reference electrode is preferred in circumstances where changes in temperature may occur, as its response to, and recovery from, temperature excursions is much more rapid.3 In order to decrease the cell coefficient from the value found with the silver - silver chloride/ 3 mol 1-1 potassium chloride external reference electrode, it is necessary to reduce the concentration of the potassium chloride solution to about 0.1 mol 1-1 in the external reference e1ectrode.l t 1 x/-x-xA/ B C + E ui 15 25 35 Tern peratu rePC Fig.7. Temperature de- pendence of experimental pH electrodes veysus calomel/ 3 mol 1-1 potassium chloride external reference electrode in ammonia solution; pH = 9.17 at 25 "C. A, Ammonia- filled; B, Tris-filled; and C, N-glycylglycine-filled pH electrode.15 25 35 Tern peratu re/"C ~~ Fig. 8. Temperature de- pendence of experimental pH electrodes versus silver - silver chloride/O. 1 mol 1-' potassium chloride external reference electrode in ammonia solutions: A, pH 9.0; B, pH 9.16; and C, pH 9.26 at 25 "C. (a) N-Glycyl- glycine-filled pH electrode ; and (b) Tris-filled pH electrode. The temperature dependences of pH cells with an external reference electrode containing this electrolyte solution were measured at ammonia levels of 0.28 mg 1-1 (pH 9.0), 0.46 mg 1-1 (pH 9.16) and 0.63mgl-1 (pH 9.26). Results of e.m.f. versus temperature for pH cells containing the Tris- and N-glycylglycine-filled experimental electrodes are shown in Fig. 8 and the variations in pH with temperature are given in Tables 1-111.It can be seen that the Tris-filled electrode cell had slightly the higher temperature dependence but in the worst instance this was only equivalent to a variation of 0.08 pH over the range of 20 "C. The cell with the N-glycylglycine-filled electrode had a variation in pH usually less than &0.03.1306 MIDGLEY AND TORRANCE: pH CELLS FOR TEMPERATURE Analyst, VoZ. I07 TABLE I VARIATION IN INDICATED pH WITH TEMPERATURE IN CELLS HAVING AN EXTERNAL SILVER - SILVER CHLORIDE/O.~ moll-1 POTASSIUM CHLORIDE REFERENCE ELECTRODE (PH = 9.00 AT 25 "c) Tris-filled experimental electrode pH calculated h \ T/"C from pS cm-l E.m.f./mV Indicated pH* 15 9.35 -92.5 8.96 20 9.17 -93.8 8.98 25 9.00 - 94.8 9.00 30 8.84 - 96.0 9.02 35 8.70 -95.5 9.01 N-Gl ycylgl ycine-filled experimental electrode E.m.f./mV Indicated pH - 127.9 9.00 - 129.5 9.02 - 128.1 9.00 - 128.0 9.00 - 125.9 8.96 * Values of pH a t T "C are normalised with respect to the conductivity value at 25 "C.Discussion From the theoretical expression for the temperature dependence of the pH cell [equation (3)], the best over-all temperature compensation was expected from a combination of the experimental glass electrodes and a silver - silver chloride reference electrode each containing the same concentration of potassium chloride in the reference electrolyte. Under these conditions, according to equation (3) the terns associated with the inside of the glass electrode should exactly balance those of the sample and the external reference electrode, and thus the cell should have a temperature coefficient of zero.However, with electrodes that contained 3 mol 1-1 potassium chloride a coefficient of about 0.6 mV "C-l was observed. A closer examination of the individual terms of equation (3) was made to ascertain where this difference could have arisen. TABLE I1 VARIATION IN INDICATED pH WITH TEMPERATURE IN CELLS HAVING AN EXTERNAL SILVER - SILVER CHLORIDE/O.~ mol 1-1 POTASSIUM CHLORIDE REFERENCE ELECTRODE (pH = 9.16 AT 25 "C) Tris-filled experimental electrode pH calculated A \ T/"C from pS cm-1 E.m.f./mV Indicated pH* 15 9.51 - 103.4 9.11 20 9.32 - 105.2 9.14 25 9.16 - 106.3 9.16 30 8.99 - 107.1 9.17 35 8.84 - 107.7 9.18 N-Gl yc ylgl ycine-filled experimental electrode E.m.f./mV Indicated pH* I A > - 139.0 9.14 - 140.1 9.16 - 140.0 9.16 - 139.4 9.15 - 138.8 9.14 * Values of pH at T "C are normalised with respect to the conductivity value at 25 "C.One possible source was a difference in the temperature coefficients of the silver - silver chloride electrode in the alkaline solution inside the glass electrode and in the neutral solution of the external reference electrode. Experiments showed that the coefficients differ by only 0.1 mV "C-1 and therefore this difference was not solely responsible for the cell coefficient described above. Another possibility was a difference between the solution coefficient (dpH/dT) inside the glass electrode (where the electrolyte is mainly 3 mol 1-1 potassium chloride), and in the dilute ammonia solution.Measurements of pH in 3 mol 1-1 potassium chloride solution adjusted to pH 9 with ammonia were made with a standard pH cell cali- brated with buffer solutions at 25 and 35 "C. The value of -0.033 pH O C - l found for dpH/ d T corresponded to that found in dilute ammonia solution containing no added potassium chloride. Another possible explanation of the unexpected cell coefficient was that there was a reaction between the internal filling solution of the glass electrode and the glass, which produced a change in the pH of the filling solution. If the known values of the coefficientsNovember, 1982 COMPENSATION IN MEASURING pH OF BOILER FEEDWATER 1307 are substituted in equation (3) and the observed value of 0.6 mV "C-l used for dECell/dT, then the equation can be solved for pHlnt.From these calculations the internal filling solution would be required to change from pH 9 to 12. This is in contradication to the observed trend in change of potential of the experimental electrode (see Fig. 3) where two electrodes demonstrated a change equivalent to a slow decrease in internal pH while the other had an initial change equivalent to a small increase (+0.3 pH) but was followed by a decrease in pH. On the basis of these results it is concluded that equation (3), derived from a simple equation for the e.m.f. of a glass electrode pH cell, does not fully describe the temperature dependence of the cell. Although these cells had significant self-compensation characteristics (the coefficient being 0.01 pH "C-l compared with -0.033 pH "C-l for ammonia solutions) they were still too temperature dependent for satisfactory application in power st ation f eedwater. TABLE I11 VARIATION IN INDICATED pH WITH TEMPERATURE IN CELLS HAVING AN EXTERNAL SILVER - SILVER CHLORIDE/O.1 mol 1-1 POTASSIUM CHLORIDE REFERENCE ELECTRODE (pH = 9.26 AT 25 "C) Tris-filled experimental electrode pH calculated T/"C from p S cm-l E.m.f./mV Indicated pH 15 9.61 - 110.4 9.20 20 9.43 - 112.4 9.23 25 9.26 - 114.0 9.26 30 9.11 - 114.7 9.27 35 8.96 -115.1 9.28 N-Glyc ylgl ycine- filled experimental electrode r 1 E.m.f./mV Indicated pH* - 145.8 9.22 - 146.7 9.25 - 147.4 9.26 - 146.7 9.25 - 145.6 9.23 * Values of pH at T "C are normalised with respect to the conductivity value at 25 'C. Given the positive over-all temperature coefficient observed for cells expected to be self- compensating] an improvement in the over-all compensation was anticipated using the same experimental electrodes but with an external reference electrode whose coefficient was ca.0.5 mV "C-l greater than the 0.2 mV "C-l of the silver - silver chloride/3 mol 1-1 potassium chloride electrode at 25 "C. The calomel/3 mol 1-1 potassium chloride electrode has a coefficient of 0.5-0.6 mV "C-l at 25 "C and therefore the over-all coefficient of a cell incorporating such a calomel electrode was expected to be nearer zero. At an ammonia concentration equivalent to pH 9.17, the experimentally determined coefficients using the calomel external reference electrode were between 0.1 and 0.15 mV "C-l, which is equivalent to a pH change over the 20 "C range of G0.05.The main disadvantage of the calomel electrode is that it responds less rapidly and reversibly to changes in temperature than the silver - silver chloride ele~trode.~ For this reason, an alternative form of the silver - silver chloride electrode was investigated. In order to increase the temperature coefficient of a silver - silver chloride external reference electrode to give an over-all cell coefficient of zero with the same experimental glass electrodes, the concentration of potassium chloride in the filling solution of the external reference electrode must be reduced to 0.1 moll-l. With this solution in the external reference electrode, the over-all cell coefficient in ammonia solutions of pH 9.0-9.26 was (0.1 mV OC-l for the N-glycylglycine-filled pH electrode and -0.2 mV "C-l (0.003 pH "C-l) for the Tris-filled pH electrode (Fig.8). Even in the latter instance the compensation would be acceptable for pH measurements in boiler feedwater. Self-compensating cells incorporating commercially available glass electrodes are possible in principle] but as such electrodes usually have internal filling solutions of pH 7 an external reference electrode with a high temperature coefficient would be needed. With silver - silver chloride reference electrodes, the filling solution would need to be so dilute to achieve this high temperature coefficient that a stable liquid junction would be difficult to obtain. The calomel electrode would give higher temperature coefficients than the silver - silver chloride electrode at the same potassium chloride concentration and would enable reasonable potassium chloride concentrations to be used if the glass electrode had a filling solution of1308 MIDGLEY AND TORRANCE pH 7.For example, with EIL Type 1070 glass electrodes, which have an internal pH of 7.0 0.3, a calomel electrode containing potassium chloride at a concentration of 0.1- 0.5 mol 1-1 would be required. As noted previously, these reference electrodes would be less reliable in circumstances where the sample temperature was fluctuating. The type of pH meter required for a self-compensating electrode system is much simpler than those normally used to measure pH in conditions where the sample temperature is changing. The major simplification is the absence of any feedback circuit from a tempera- ture sensor in the pH cell since the temperature compensation is brought about chemical changes in the electrodes themselves.A meter suitable for use with these electrodes need only have the facility for manual alteration of the slope factor together with a control which offsets the e.m.f. to the value corresponding to the standardisation pH. A limitation of this measuring system might arise if conventional buffer solutions are used for standardisation because the electrode system treats all solutions as feedwater and so applies an incorrect compensation for dpH/dT. This effect can be avoided by buffering at the reference tempera- ture (25 “C). For the most accurate control of pH, the standardisation is best made by means of dilute ammonia solutions of known temperature and cond~ctivity,~ and if this technique is used for standardisation the limitation would disappear.In most instances, the pH meters currently installed in power stations are unsuitable for use with self-compensating electrodes without some alteration to their temperature compensa- tion circuits. In particular the isopotential control must be set at zero and the resistance thermometer removed and replaced by a fixed resistor whose value is equal to that of the thermometer at 25 “C. Conclusions Experimental glass pH electrodes were constructed with filling solutions formulated to compensate for the temperature-induced changes in both the e.m.f. of the external reference electrode and the pH of alkaline feedwater. When both the internal silver - silver chloride electrode of the glass electrode and the external silver - silver chloride reference electrode contained 3 mol 1-1 potassium chloride solution in addition to a weak base the cell coefficient in simulated feedwater was about 0.01 pH “C-l. This was not only greater than expected from theory but was not considered to be adequate for accurate measurements in feedwater. Over-all compensation suitable for an accuracy of A0.05 pH over a 20 “C range in tempera- ture was obtained by selecting an external reference electrode whose temperature coefficient exactly compensated for the combined effects of temperature on the experimental pH electrodes and feedwater. Suitable reference electrodes were shown to be a calomel/3 mol 1-1 potassium chloride electrode or a silver - silver chloride/O.l mol 1-1 potassium chloride electrode. For application in conditions where the temperature can vary rapidly, the latter reference electrode is preferred as its response is more rapid and reversible. This work was carried out at the Central Electricity Research Laboratories and is published The glass electrodes were kindly The authors thank Mr. by permission of the Central Electricity Generating Board. supplied by Mr. A. E. Bottom of EIL Analytical Instruments. N. A. Dimmock for his assistance with the practical work. References 1. 2. 3. 4. Mattock, G., “pH Measurement and Titration,” Heywood, London, 1961. Robinson, R. A., and Stokes, R. H., “Electrolyte Solutions,” Second Edition (Revised), Butter- Midgley, D., and Torrance, K., Analyst, 1976, 101, 833. Toms, D., Note SSD/SW/SO N92, CEGB South Western Region, Bedminster Down, 1980. worths, London, 1968, p. 465. Received May l l t h , 1982 Accepted June 23rd, 1982

ISSN:0003-2654    

DOI:10.1039/AN9820701297

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

Analyst, November, 1982, Vol. 107, pp. 1309-1315 1309 Kinetic Determination of Nitrite in Waters by Using a Stopped-flow Analyser M. A. Koupparis," K. M. Walczak and H. V. Malmstadtt School of Chemical Sciences, University of Illinois at Urbana-Champaign, Urbana, I L 61801, USA Automatic kinetic methods for the determination of nitrite in waters with a stopped-flow analyser are described. The methods are based on the diazoti- sation of sulphanilamide, the product being coupled with N- (1-naphthy1)- ethylenediamine dihydrochloride to form a highly coloured azo dye, which is measured at 540nm. A single-point kinetic procedure uses a delay time of 10 s and one measurement of 0.7 s. A multi-point reaction rate method uses a delay time of 0.8 s and a measurement time of 1.5 s.The methods are fast, sensitive, accurate and precise, and without serious interference. The sample throughput for routine analysis can be up to 360 samples per hour in the range 0.025-2.00 p.p.m. of nitrite-nitrogen. Keywords Nitrite determination ; water analysis ; kinetic method; stopped- flow analyser ; N- (1-naphthylfethylenediamine dihydrochloride Nitrite, an intermediate stage in the nitrogen cycle, occurs in water as a result of the bio- logical decomposition of proteinaceous materials. When correlated with the concentration of other nitrogen forms, trace amounts of nitrite can indicate organic pollution. The role of nitrite ion as an important precursor in the formation of nitrosamines, many of which have been shown to be potent carcinogens, has been studied.1*2 Therefore, sensitive and rapid methods for the routine determination of nitrite are desirable.Many spectrophotometric method^^-^ have been reported for the determination of small amounts of nitrite, most of them being based on the Griess reaction, i.e., the reaction of nitrite with a primary aromatic amine to form a diazonium salt, which is then coupled with another aromatic compound to form an azo dye whose absorbance is measured. Many of these methods have been adapted to automatic air-segmented continuous-flow systems,7-10 e.g., the Technicon AutoAnalyzer system, and flow injection ana1yser~ll-l~ for the determina- tion of nitrite and also nitrate after its reduction by various reductor columns. The stopped-flow technique for the rapid mixing of chemical reagents has had widespread use in basic studies on the rates of rapid chemical reactions and more recently it has been shown that a modified automated stopped-flow system is a very valuable tool for analytical purposes.14 High precisions are possible because of the very precise delivery of sample and reagent and rapid, quantitative mixing of the solutions.Reaction rate methods used with the stopped-flow system15 demonstrate the potential advantages over other automatic analysers. Recently, we have described the construction and analytical applications of a compact automated microprocessor-based stopped-flow analyser.16 Results are shown here using this system for an automated fast kinetic determination of nitrite in waters using the Shinnl7 reaction, i.e., the formation of a reddish purple azo dye produced by coupling diazotised sulphanilamide (SAA) with N-( 1 -naphthyl)ethylenediamine dihydrochloride (NEDD) .This reagent was chosen because of its high sensitivity and relative freedom from interferences. The determination of nitrite must be carried out promptly on fresh samples to prevent bacterial conversion of nitrite into nitrate or ammonia. For short-term preservations for 1-2 days, the sample should be frozen at -20 "C or 40 mg of mercury(I1) chloride per litre of sample should be added, with storage at 4 "C. Experimental Apparatus laboratory.16 144, Greece. The automated, microprocessor-based stopped-flow analyser (SFA) was developed in our The entire system is automated using a Rockwell AIM 65 microcomputer for * Present address : Laboratory' of Analytical Chemistry, University of Athens, 104 Solouos Str., Athens, t To whom correspondence should be addressed.1310 KOUPPARIS et al.: KINETIC DETERMINATION OF Analyst, Vol. 107 control of all operations, data acquisition and reduction, display and printout of results. The sampling - mixing module is used to sample 150-pl volumes of reagent and standards or samples on the turntable with a reproducibility better than O.lyo, to mix them efficiently and transfer them into the observation cell. The photometric system is an easily constructed module with a 1-cm flow cell, automatic shutter, light source, photomultiplier and interference filters. An interference filter with a 10-nm band pass at 540 nm was used in the photometer.An investigative program was used to evaluate the optimum parameters for the analytical procedure and a routine reaction rate program for the calibration graph and the analysis of the samples. All measurements were carried out in an air-conditioned laboratory maintained a t a nominal temperature of about 25 "C. Reagents All chemicals used were of analytical-reagent grade and de-ionised water was used in all experiments. Nitrite standard solution, lo00 p.p.m. of N02--N. A 4.926-g amount of sodium nitrite (Baker Analysed, 99.7y0), oven-dried at 100-105 "C for 2 h, was dissolved in de-aerated, de-ionised water and diluted to 1 1. A pellet of sodium hydroxide and 1 ml of spectroscopic- grade chloroform were added in order to prevent liberation of nitrous acid and bacterial growth.It was kept in a refrigerator and replaced every 2 weeks. Working standard solutions in the range 0.025-2 p.p.m. of NO,--N were prepared daily by appropriate dilution. The colour reagent was prepared by dissolving 10.0 g of SAA, 0.50 g of NEDD and 30 ml of 85% orthophosphoric acid in water and diluting to 1 1. This solution was stored in an amber-glass bottle in a refrigerator. Interfering ion solutions. These were prepared so as to contain 1000 p.p.m. of each ion to be tested. Reagent solution. Samples Waste-water samples from East-Central Illinois, USA, collected and preserved as recom- mended,3 were analysed. Before analysis, only samples containing suspended solids were filtered and highly coloured organic matter was removed by adding an aluminium suspension (2 ml per 100 ml), stirring and filtering.Procedure The turntable is loaded with the blank (de-ionised water), three to five standards in the range 0.025-2 p.p.m. of N02--N and the samples in 5-ml disposable polystyrene micro- beakers. One channel of the SFA is used to take aliquots of the reagent solution and the other the standards and the samples. The appropriate BASIC and machine language pro- grams are loaded from the cassette recorder into the computer's memory. Then, in execution of the program, the blank is injected into the flow cell, an integration time of 0.7 s is selected and the dark current and lo0 yo transmittance are measured automatically. The program then prompts the operator to select the time parameters. A delay time of 10s and a measurement time of 0.7 s (one integration) are used for the single-point kinetic procedure, and a delay time of 0.8 s and a measurement time of 1.5 s for the multi-point kinetic pro- cedure.The number of standards and samples to be measured, the number of measure- ments to be averaged and the number of flushes are then requested by the computer. The program then sequences through each standard, flushing the system between each standard (four flushes are needed) and prompting the operator for its concentration. After the standards have been measured, the microcomputer calculates the linear least-squares regression line and prints its slope, intercept, correlation coefficient and the standard error of the estimate. Samples are automatically measured, after which the concentration of the nitrite in the sample is calculated and printed.A dedicated program can also be used with all the information, time parameters and con- centration of standards contained in the software. Once the operator has input the number of samples, the analysis will be completed automatically.November, 1982 NITRITE IN WATERS USING A STOPPED-FLOW ANALYSER Results and Discussion Optimisation of the Procedure 131 1 The behaviour of the reaction using various concentrations of the colour reagent constitu- ents was investigated, using an investigative program that prints out 50 absorbance values during a pre-selected time period. The reaction curve was plotted on the paper printer and the reaction rate was calculated at various parts of the curve together with linearity factors.From these data the optimum reagent concentration, the delay time and measurement time were evaluated. The effect of the nitrosated species (SAA) concentration on the reaction is shown in Fig. 1. Both the reaction rate and the final absorbance values increase with increasing SAA con- centration. An SAA concentration of 1% was chosen as the optimum. 1.5 W 2 1.0 2 s 9 a 0.5 0 1.50% 1 .OO% 0.50% 0.10% I I 5 10 tis 15 Fig. 1. Effect of sulphanilamide concentration on the reaction. NO,--N, 1 p.p.m.; NEDD, 0.05%; and HsPO,, 10% V / V . ____ ~~ ~ 0.05% 0.10% 0.1 5% a 0.5 0.01% I I 0 5 10 r/s 5 Fig. 2. Effect of N-( 1-naphthy1)ethylenediamine dihydrochloride concentration on the reaction. NO,--N, 1 p.p.m. ; sulphanilamide, 1 "/o ; and H,PO,, 10% V / V .The effect of NEDD concentration was also investigated and the reaction curves for 15 s are shown in Fig. 2. As illustrated, the initial reaction rate increases with increase in NEDD concentration but the final absorbance value decreases. Because of the simultaneous addition of nitrosated species (SAA) and the coupling reagent (NEDD) there is a competition for nitrite ion. At low concentrations of NEDD the conversion of diazoniuni ion into the coloured pigment is slow and incomplete. Conversely, at high concentrations appreciable amounts of nitrite react with NEDD and pigment formation is reduced. The resultant effect is that there is a point of maximum pigment production with varying NEDD con- centration.18 Under our experimental conditions a 0.05% concentration was found to be the optimum.This allows a 30-fold excess of SAA. I 1 I J 0 5 10 15 ti s Fig. 3. Effect of phosphoric acid concentration on the reaction. NO,--N, 1 p.p.m. ; sulphanilamide, 1%; and NEDD, 0.05%.1312 KOUPPARIS et a,?. : KINETIC DETERMINATION OF Analyst, Vol. 107 The effect of orthophosphoric acid concentration was also investigated in the range 1-20%. As shown in Fig. 3, there is only a small dependence of the reaction rate OP the acid concentra- tion. Small increases in the reaction rate and the final absorbance value are initially observed as the acid concentration increases but at higher concentration they both decrease. The optimum concentration was 3%, which provides a pH of 1.43 in the mixed solution. This effect is a balance between too little nitrous acid (and then of the nitrosating species, N,OJ at higher pH values, and protonation of the NED (which causes the decrease in the reaction rate) and/or acid decomposition of the pigment at lower pH values.l8 2.5 2.0 al & 1.5 2 2 2 1 .o 0.5 0 5 tls 10 Fig.4. Reaction curves for various nitrite concentrations used for the calibra- tion graphs. Sulphanilamide, 1 :h ; NEDD, 0.05%; and H,PO,, 3% V / V . Using the optimum concentrations of reagents found, the reaction curves shown in Fig. 4 were obtained for various nitrite concentrations. The reaction is almost completed after 10 s so both multi-point reaction rate and single-point kinetic procedures can be used. For the multi-point reaction rate procedure a delay time of 0.8 s and a measurement time of 1.5 s gave the best results.The calibration graph obtained using this procedure was linear TABLE I NITRITE CALIBRATION GRAPHS Single-point kinetic method* A I \ NO,--” p.p.m. Absorbance RSD, 1% 0.025 0.035 4.0 0.050 0.071 3.2 0.100 0.138 0.5 0.500 0.708 0.2 1 .oo 1.413 0.2 1.50 2.112 0.1 Multi-point reaction rate method? f A \ NO,--N, p.p.m. Rate/mA s-l RSD, $% 0.100 26.8 3.4 0.500 184.6 1.0 1 .oo 376.6 0.2 1.50 563.9 0.2 2.00 747.8 0.1 Slope = 1.410 Intercept = 0.000 2 Correlation coefficient (r) = 0.999 995 Slope = 379.3 Intercept = -6.9 Correlation coefficient (r) = 0.99990 * Delay time = 10 s, measurement time = 0.7 s. t Delay time = 0.8 s, measurement time = 1.5 s. 3 Relative standard deviation (n = 5).November, 1982 NITRITE IN WATERS USING A STOPPED-FLOW ANALYSER 1313 from 0.100 to 2.00 p.p.m.of NO,--N. For the single-point procedure a delay time of 10 s and one measurement of 0.7 s gave a calibration graph that was linear from 0.025 to 1.50 p.p.m. of NO,--N. Using a delay time of 5 s a calibration graph linear up to 2.00 p.p.m. can be obtained. Table I shows typical results for calibration graphs obtained using both procedures. Accuracy and Precision of the Method The accuracy of the proposed method was examined by measuring the nitrite concentration of waste water samples before and after a standard addition. Recovery was calculated after addition of 100 pl of 10 p.p.m. NO,--N standard to 10 ml of each assayed sample. Because of the low concentrations of the samples the more sensitive single-point procedure was used.The results obtained (Table 11) show an average recovery of 98.6%. The precisions of both procedures are given in Table I. Similar results were obtained for the analysis of samples. TABLE I1 RECOVERY DATA FOR THE DETERMINATION OF NITRITE IN WASTE WATERS (SINGLE-POINT PROCEDURE) Nitrite content, p.p.b.* NO,--N Sample No. 1 2 3 4 5 6 7 8 9 10 Before standard addition 156 38 129 232 185 37 32 175 55 25 After standard addition c - A F , Recovery, Expected Determined % 253 252 99 137 138 101 227 225 99 329 325 99 282 274 97 136 133 98 131 128 98 272 272 100 153 150 98 124 120 97 Average: 98.6 * Parts per lo8. Interference Study The results of these experi- ments are shown in Table 111. Major interference is caused by reducing or oxidising ions.The low results caused by copper(I1) ion because of its catalysis of the decomposition of the diazonium salt in methods with long reaction times3 are avoided with the proposed method. The serious sulphite interference that appears during the analysis of nitrite in beet sugar factory juiceslS with the Shinn method is also almost eliminated in this method. The sulphide interference can be eliminated by adding excess of cadmium ions and filtering. Addition of 200 p.p.m. of Cd2+ to a waste water sample containing 1 p.p.m. of NO,--N and 50 p.p.m. of sulphide eliminated the error. The effect of various potential interferents was investigated. Sample Throughput Four flushes are used to change from one solution to another (flush volume 1 5 0 ~ 1 for sample and reagent and instrument cycle time about 1.5 s) and, assuming one measurement per sample, 1 s for the turntable position increment and 0.5 s for the computer calculation and printing time, the analysis rates for the aforementioned methods are 200 samples per hour for the single-point procedure (270 samples per hour if a 5-s delay time is used for the range 0.1-2.0 p.p.m.) and 360 samples per hour for the multi-point reaction-rate procedure.Conclusion The proposed kinetic determinations of nitrite are sensitive, accurate and precise. The The ions commonly useful analytical range can be extended by using a shorter delay time.1314 KOUPPARIS et al. : KINETIC DETERMINATION OF Analyst, “01. 107 Ion investigated Acetate . . .. Bromide . . .. Bromide .. .. Bromide . . .. Carbonate .. Chloride . . .. Dichromate . . Dichromate . . Dichromate . . Fluoride . . .. Iodate . . .. Iodide . . .. Iodide . . .. Iodide . . . I Nitrate . . .. Oxalate . . .. Perchlorate . . Phenol . . .. Sulphate . . .. Sulphide . . .. Sulphide . . .. Sulphide . . .. Sulphite . . .. Cyanide . . .. Aluminium .. Ammonium (NH,Cl) Cadmium .. .. Cobalt (CoC1,) . . Copper . . .. Iron(I1) . . .. Iron(II1) . . .. Iron(II1) . . .. Iron(II1) . . .. Lead . . .. Magnesium (MgC1,) Mercury(I1) . . Tin(I1) . . .. Tin(I1) . . .. Tin(I1) . . .. .. .. .. .. .. .. .. .. * . .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .... .. .. .. .. .. .. .. TABLE I11 EFFECT OF DIVERSE IONS Nitrite concentration : 1 p.p.m. Error,* % Ratio of ion to nitrite (m/m) 200 200 20 2 200 200 200 200 20 2 200 200 200 20 2 200 200 200 200 200 200 20 2 200 200 200 200 200 200 200 200 20 2 200 200 200 200 20 2 I Single-point procedure n -23.5 - 4.9 n n n n +32.3 +4.7 n n n -41.7 - 10.9 n n n n n n - 78.9 - 16.3 -2.2 n n n n n n -3.1 - 30.3 - 4.0 n n n $. 2.4 - 98.8 - 12.5 n 1 Multi-point procedure n -27.3 -4.7 n n n n +5.3 n n n n -46.7 -9.1 n n n n n n - 78.4 - 5.9 n + 3.4 n -4.2 n - 5.0 n n - 33.1 - 7.0 n n - 6.9 + 3.2 -99.7 - 13.4 n * n = negligible (less than 2%). present in waters do not interfere and the interferences from copper(I1) and sulphite ions are eliminated. Only a small amount of sample is needed, the sample throughput is high (up to 360 samples per hour) and the analyst effort is small.Compared with the standard Technicon AutoAnalyzer method it has the same sensitivity, a superior rate of analysis, better precision and accuracy and less interference from copper(I1) ions. This relatively simple automated stopped-flow analyser can be used for a wide variety of constituents in environmental samples. This project was partially supported by a grant from PHS-5RO1 GM 21984. References 1. 2. 3. American Public Health Association, American Water Works Association and Water Pollution Control Federation, “Standard Methods for the Examination of Water and Wastewater,” Fifteenth Edition, American Public Health Association, New York, 1980, p, 380. Kolthoff, I.M., and Elving, P. J., “Treatise on Analytical Chemistry,” Part 11, Volume 5, Inter- science, New York, 1961, p. 275. Lijinsky, W., and Epstein, S. S., Nature (London), 1970, 225, 21. Wolff, I. A., and Wasserman, A. E., Science, 1972, 177, 4043. 4.November, 1982 NITRITE IN WATERS USING A STOPPED-FLOW ANALYSER 1315 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. Streuli, C. A., and Averell, P. R., “The Analytical Chemistry of Nitrogen and its Compounds,” Sawicki, E., Stanley, T. W., Pfaff, J., and D’Amico, A., Talanta, 1963, 10, 641. Bernard, M., and Macchi, G., “Automation in Analytical Chemistry, Technicon Symposia, 1965,” Britt, R. D., Jr., Anal. Chem., 1962, 34, 1728. Armstrong, F. A. J., Stearns, C. R., and Strickland, J. D. H., Deep Sea Res., 1967, 14, 381. Henriksen, A., and Selmer-Olsen, A. R., Analyst, 1970, 95, 514. Anderson, L., Anal. Chim. Acta, 1979, 110, 123. Gin6, M. F., Bergamin Fo., H., Zagatto, E. A. G., and Reis, B. F., Anal. Chim. Ada, 1980, 114, 191. Zagatto, E. A. G., Jacintho, A. O., Mortatti, J., and Bergamin FO., H., Anal. Chim. Acta, 1980, 120, Krottinger, D. L., McCracken, M. S., and Malmstadt, H. V., Am. Lab., 1977, 9 (3), 51. McCracken, M. S., and Malmstadt, H. V., Talanta, 1979, 26, 467. Koupparis, M. A., Walczak, K. M., and Malmstadt, H. V., J . Autom. Chem., 1980, 2 (2), 66. Shinn, M. B., Ind. Eng. Chem., Anal. Ed., 1941, 13, 33. Fox, J. B., Jr., Anal. Chem., 1979, 51, 1493. Lew, R. B., Analyst, 1977, 102, 476. Part I, Wiley-Interscience, New York, 1970, p. 121. Mediad, New York, 1966, p. 255. 399. Received December lst, 1981 Accepted June llth, 1982

ISSN:0003-2654    

DOI:10.1039/AN9820701309

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

1316 Analyst, November, 1982, Vol. 107, PP. 1316-1319 Simplified Procedure for the Determination of Chemical Oxygen Demand Using Silver Nitrate to Suppress Chloride Interference A. Lloyd Southern Water Authority, East Sussex Water and Drainage Division, Construction House, Menzies Road, St. Leonards-on-Sea, East Sussex, TN38 9BD A sealed-flask procedure for the determination of the chemical oxygen demand of wastewaters is described. Samples are digested at 150 "C in a glass- stoppered flask using springs to retain the stopper. The use of sealed-flask conditions offers two advantages over reflux conditions : simplified apparatus and experimental procedure and improved suppression of chloride interference. The proposed procedure is similar to the standard procedure in accuracy and reproducibility of results.Keywords : Chemical oxygen demand ; sealed flask ; wastewaters The determination of chemical oxygen demand (COD) in wastewaters, using silver nitrate to suppress chloride interference, has been recently described.l The procedure employed the standard reflux conditions commonly used in Water Authority laboratories. The sealed- tube COD procedure2 offers several advantages over the reflux procedure, notably a saving in bench space and cost of equipment. In spite of these advantages and general compara- bility with reflux p r ~ c e d u r e s ~ ~ ~ in results, the sealed-tube procedure has not been adopted as a standard method. It is understood that the Department of the Environment's Standing Committee of Analysts decided not to endorse a sealed-tube COD procedure because of safety considerations.Thermal degradation of the tube caps and liners is also found to occur. This creates difficulties in unscrewing caps and occasionally causes contamination of samples, particularly if an oven is used to heat the tubes. The apparatus described below overcomes the problems cited above and offers several advantages over the silver nitrate reflux procedure. It is not necessary to use a weaker dichromate solution for COD levels of 200 mg 1-1 or less. The suppression of chloride inter- ference is improved and lower concentrations of silver nitrate are used. Experimental Apparatus Flasks Thick-walled Pyrex conical flasks, of 100-ml capacity, with ground-glass stoppers (24/20 socket) were used. Hooks on the stoppers and flask shoulders allow the stoppers to be retained by a pair of wire springs. Clean new apparatus before use by shaking the stoppered flasks containing a sulphuric acid - dichromate mixture and then heating in the oven at 150 "C for 2 h.Dry the flasks and stoppers before use. After heating, rinse the flasks and stoppers with low COD water. Oven Use a mercury-in-glass thermometer to check any variation in temperature within the working space. A fan convection oven, capable of heating the flasks at 150 & 2 "C, was used. Reagents de-ionised water with a suitably low COD for all reagents and blanks. to the mark in a 100-ml calibrated flask. batch within 2 weeks. Use analytical-reagent grade reagents, except where stated otherwise. Sulphuric acid.Silver nitrate solution, 25% m/V. Use distilled or AnalaR grade (d2,, = 1.84). Dissolve 25 & 0.5 g of silver nitrate in water and dilute Store this solution in amber glassware and use eachLLOYD 1317 Dissolve 6.129 g of potassium dichromate (previ- Renew BDH Chemicals Ltd. 1 ,lo-Phenanthroline - iron( 11) sulphate com- Dissolve 9.80 0.01 g of ammonium Carefully add 20 & 0.5 ml of sulphuric acid, Standardise this solution before starting Potassium dichromate solution, 0.125 N. ously dried at 140 "C for 1 h) in water and dilute to the mark in a 1-1 calibrated flask. this solution after 4 weeks. Ferroin indicator. plex solution, 0.025 M. Ammonium i r o n ( I 1 ) sulphate solution, 0.025 N. iron(I1) sulphate in about 100 ml of water. cool and dilute to the mark in a 1-1 calibrated flask.each batch of analyses, using the procedure described below. Procedure Caution-Oxides of sulphur and other toxic gases are emitted during addition of sulphuric acid. Add Examine flasks carefully before use and reject any showing cracks. sulphuric acid in a fume cupboard. Standardisation of 0.025 N ammonium iron (11) sulphate solution Dispense potassium dichromate (5 & 0.05 ml of 0.125 N) into a conical flask and dilute to approximately 50 ml with water. Carefully add sulphuric acid (15 & 0.25 ml), mix and cool to ambient temperature. Add Ferroin indicator (2 drops) and titrate with the ammonium iron(I1) sulphate solution. The end-point colour change is pale blue to red. Calculate the normality ( N ) of the ammonium iron(I1) sulphate solution from 0.625 V N = - where V = ammonium iron(I1) sulphate titre.Samples containing up to 4000 mg 1-1 of chloride, irrespective of COD level Do not allow sample or reagents to wet the neck of the flask during addition or mixing (NOTE 1). Pipette 10 0.1 ml of sample having a COD of less than 400 mg 1-1 (or an appropriate volume of stronger sample diluted to 10 ml) into a flask. Add silver nitrate (1.0 & 0.1 ml of 25% m/V solution) (NOTE 2). Add 5 & 0.05 ml of 0.125 N potassium dichromate solution, then mix and carefully add 15 & 0.25 ml of sulphuric acid. Mix carefully, stopper the flasks (immediately securing the stoppers with springs) and place in the oven which has previously been heated to 150 & 2 "C. Leave for 2 h (-&5 min).Remove the flasks from the oven and allow to cool to room tempera- ture (NOTE 3). Remove the stoppers, add 25 -& 1 ml of water and titrate the residual di- chromate as described above. Carry out duplicate blanks using 10 -& 0.1 ml of water in place of the sample (NOTE 4). Mix and allow to stand for between 5 and 15 min. Calculation 8 OOON s v COD (mgl-l) = - (VS - V S ) where VB = blank titre, Vs = sample titre and S , = sample volume. NOTES- 1. 2. Use silver nitrate (1.0 3. 4. Crystallisation of dissolved salts has been found to cause the stoppers to stick. 0.1 ml of 15% m/V solution) for samples having an expected chloride to Do not titrate the solution while still warm. In warm solution, iron(I1) is oxidised by oxides of Duplicate blanks should not differ by more than 0.3 ml and the average blank should not differ by COD ratio of 2 or less.nitrogen to iron(III), causing reversion of the end-point (see under Discussion). more than 1 ml from the volume of ammonium iron(I1) sulphate used in standardisation. Performance Characteristics The precision of the proposed procedure and bias with respect to the standard procedure4 were estimated by the method described ear1ier.l Within-batch relative standard deviation in sewage analysis ranged from 2.2% at levels of 60 mg 1-1 of COD to 0.8% at 380 mg 1-1 of COD (4 degrees of freedom). Total relative standard deviation in analysis of a 300mg1-11318 LLOYD: PROCEDURE FOR DETERMINING COD USING Analyst, VoZ. 107 COD potassium hydrogen phthalate solution was 1.0% (8 degrees of freedom).These results did not differ significantly (9 = 0.05) from data obtained using the standard pro- cedure or the silver nitrate reflux COD pr0cedure.l TABLE I ANALYSES OF WASTEWATERS BY STANDARD AND PROPOSED PROCEDURES (1 ml OF 25% m/V SILVER NITRATE SOLUTION) Sewage effluent: . . .. COD*/mg 1-1 r Sample Chloride/mg 1-l < 100 Settled sewage . . .. . . { Z88 2 900 < 100 2 800 Settled sewage: .. .. 4 000 Sewage effluent .. .. { <loo Standard proceduret 398 329 490 361 68 38 143 339 Proposed procedure 382 313 484 366 63 39 145 327 Mean bias of proposed procedure, yo -4.0 -4.9 - 1.2 + 1.4 - 7.4 +2.6 + 1.4 -3.5 4 000 91 92 +1.1 { 1000 74 78 4-5.4 * Results are the means of two determinations. Determinations were made using mercury(I1) sulphate (1 ml of 20% m/V solution in 10% V/V sulphuric acid) at chloride levels of 500 mg 1-' or less.At higher chloride levels a mass of mercury(I1) sulphate 40-times greater than the mass of chloride in the aliquot was used. : Samples spiked with chloride. Analyses of saline sewage and chloride-spiked sewage are presented in Table I. Waste- waters having a chloride to COD ratio of 2 or less were analysed by the proposed procedure but using silver nitrate (1 ml of 15% m/V solution). These results are given in Table 11. Standard solutions of potassium hydrogen phthalate, spiked with chloride, were also analysed. These results are compared with those given for the standard procedure in Table 111. TABLE I1 ANALYSES OF LOW CHLORIDE WASTEWATERS BY STANDARD AND PROPOSED PROCEDURES (1 mi OF 15% m/V SILVER NITRATE SOLUTION) COD*/nig 1-1 Sample Chloridelmg 1-l 70 * .{ 1;: Sewage e@uent . . .. .. Settled sewage . . .. . . .. 112 (200 mg 1-1 COD) . . . . .. 0 Potassium hydrogen phthalate standard Trade efluents- Preserves and pickles . . .. .. 71 Abattoir .. .. .. .. 94 Laundry .. .. .. .. 139 Laundrette . . .. .. .. 52 Frozen foods . . .. .. .. 48 Dairy .. .. .. .. .. 79 r Standard proceduret 75 68 54 300 202 6 790 305 59 1 363 768 287 1 Proposed procedure 78 63 56 307 199 6 460 302 587 374 735 283 Mean bias of proposed procedure, yo + 4.0 - 7.4 + 4.0 +2.3 - 1.5 -4.9 - 1.0 -0.7 f3.0 -4.3 - 1.4 * Results are the means of two determinations. t Determinations were made using mercury(I1) sulphate (1 ml of 20% m/V solution in 10% V / V sulphuric acid) a t chloride levels of 500 mg 1-1 or less.At higher chloride levels a mass of mercury(I1) sulphate 40-times greater than the mass of chloride in the aliquot was used.November, 1982 SILVER NITRATE TO SUPPRESS CHLORIDE INTERFERENCE 1319 TABLE I11 ANALYSES OF CHLORIDE-SPIKED POTASSIUM HYDROGEN PHTHALATE SOLUTIONS BY STANDARD AND PROPOSED PROCEDURES (1 d OF 25% m/v SILVER NITRATE SOLUTION) Expected A \ Observed COD a t given chloride level*t/mg 1-’ C1 COD/mgl-l 0 500 1000 1500 2 000 0 - 15 (10) 19 (23) 21 (41) 32 (59) 100 102 112 (103) 115 (105) 121 (109) 128 (121) 200 199 203 (206) 209 (214) 221 (215) 217 (228) 300 295 300 (308) 297 (308) 311 (311) 316 (318) 400 391 401 (404) 403 (404) 405 (412) 407 (419) * Results are the means of two determinations. t Values in parentheses are means of results obtained using the standard proced~re.~ Discussion As noted above, it is important to cool the diluted digest to room temperature before titration.This is attributed to oxidation of iron(I1) by nitrogen ox’ides, formed by a brown-ring type reaction of the nitrate ion. It is assumed that these oxides are expelled under reflux conditions. The reaction does not affect the end-point, as it does not occur in the presence of ~hromate.~ I t is possible, however, that the reversion might cause an inexperienced analyst, titrating a hot solution, to overshoot the end-point. Table I shows that the sealed-flask COD procedure extends the range of comparability with the standard procedure up to 4000 mg 1-1 of chloride.None of the observed differences are significant ( p = 0.05). Table I1 demonstrates the possibility of using a lower concentra- tion of silver nitrate solution when the chloride to COD ratio is 2 or less. The observed differ- ences are again not significant ($ = 0.05). Table I11 confirms the general comparability of the proposed and standard procedures over a wide range of chloride concentrations. The proposed procedure may be considered more economical and versatile than reflux procedures using mercury(I1) sulphate. Current costs for silver salts in the proposed procedure at 100 and 4000 mg 1-1 of chloride are 10 p and 17 p per test, respectively. Equivalent costs for silver and mercury salts in a reflux procedure are 12 p and 28 p, respectively. The versatility of the proposed procedure arises from the use of a single reagent to suppress chloride interferences. When using mercury(I1) sulphate it is necessary to add this salt in a 40-fold excess over the mass of chloride ion in the sample aliquot. Thus the mass of mercury(I1) sulphate required depends on the chloride concentration in the sample and the volume taken for analysis. Above 30 “C, the indicator reverts from red to blue within a few minutes. Acknowledgements are made to the Director of Operations, Southern Water Authority, and the Divisional Manager, East Sussex Water and Drainage Division, Southern Water Authority, for permission to publish this paper. The author also acknowledges technical assistance from D. Ballinger. References 1. Ballinger, D., Lloyd, A., and Morrish, A., Analyst, 1982, 107, 1047. 2. Jirka, A. M., and Carter, M. J., Anal. Chem., 1975, 47, 1397. 3. Ballinger, D., Jamison, A., Lloyd, A., Morrish, A., and Stone, D., Water Pollut. Control, 1982, 81, in 4. Standing Committee of Analysts, Editors, “Chemical Oxygen Demand (Dichromate Value) of 5. Vogel, A. I., “Macro and Semimicro Qualitative Inorganic Analysis,” Fourth Edition, Longmans, Received May 19th, 1982 Accepted June 28th. 1982 the press. Polluted and Waste Waters,” HM Stationery Office, London, 1977. London, 1973, p. 363.

ISSN:0003-2654    

DOI:10.1039/AN9820701316

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

1320 Analyst, November, 1982, Vol. 107, j5#. 1320-1329 Extraction of Iron(ll1) from Aqueous Solution with Mixtures of Aliquat 336 and Ferron in Chloroform S. Przeszlakowski and E. Habrat Department of Inorganic and Analytical Chemistry, Medical School, 20-08 1 Lublin, Poland Ferron is efficiently extracted by Aliquat 336 in chloroform and the ion pair (R,N+HL-) forms dark green complexes with iron(III), in the organic phase. On the basis of solvent extraction data for iron, as well as absorbance measurements, it is assumed that two iron complexes, one with ferron and one with the alkylammonium cation, can be present in the organic phase [FeL53-(R4N+)5 and FeL,-R4N+]. The latter complex can be formed at high iron concentration in aqueous solution, where part of the ferron is also stripped from the organic phase and forms coloured complexes with iron in the aqueous phase.The absorbance of solutions of Aliquat 336 and ferron in chloroform is very low, in the visible region. The absorbance of iron complexes with ferron and Aliquat 336 is higher than the absorbance of iron complexes with ferron in aqueous solutions and three absorbance maxima were found: 370nm ( E = 7.7 x 1031mol-1cm-1), 465nm (E = 6.86 x lo3 1 mol-l cm-l) and 610 nm (e = 5.95 x lo3 1 mol-l cm-l). The absorb- ances at 465 and 610 nm can be utilised in the extraction - spectrophotometric determination of iron. Beer's law is obeyed in the concentration range 0.1- 10 mg 1-1 of iron in aqueous solution. The method is selective for the deter- mination of iron and only copper interferes when present in amounts that exceed the iron concentration.The proposed method was used to determine iron in two samples of natural waters, and the results of 0.48 and 5.7 mg 1-1 are in good agreement with results obtained by atomic-absorption spectro- metry and a spectrophotometric method using thiocyanate. Keywords : Alkylammonium salt extractants ; Aliquat 336 extraction ; ferron extraction ; iron determination in water ; spectrophotometry The extraction of anionic complexes of metals with sulphonated chelating reagents by solutions of amines with high relative molecular mass or of quaternary alkylammonium salts has been utilised in the extraction - spectrophotometric determination of metals, because extraction often increases the sensitivity and selectivity of conventional methods used for the determination of metals with such reagents in aqueous solutions.lS2 Ziegler et aZ.394 improved the colorimetric method for determining iron(III), based on the extraction of the coloured complex of the metal with ferron (7-iodoquinolin-8-ol-5-sulphonic acid) by isoamyl alcohol, by adding tributylamine acetate to the aqueous solution.Although this method is more selective, in comparison with conventional spectrophotometric methods for the deter- mination of iron with ferron in aqueous solution, the relatively high extraction of ferron by higher alcohols5 and the solubility of tributylamine salts in water make it difficult to deter- mine the extraction mechanism of iron and to investigate the complex formation in the organic phase. Our earlier experiments on chromatographic methods (paper extraction chromatography and moist-paper techniques)6 as well as static extraction7y8 indicated that ferron and other sulphonated metal chelating reagents can be effectively extracted with solutions of ternary amines, their salts or quaternary alkylammonium salts in organic diluents and that Aliquat 336 (methyltricaprylylammonium chloride, where caprylyl is an alkyl C8-C,,) appeared to be the most effective extracting agent.Therefore, the extraction of ferron by Aliquat 336 and the extraction of iron(II1) by the ion pair produced by Aliquat 336 and ferron was investigated in order to utilise the extraction in the extraction - spectro- photometric determination of iron(II1). Experimental All experiments were performed at ambient temperature (23 & 2 "C).PRZESZLAKOWSKI AND HABRAT 1321 Apparatus A Zeiss VSU 2P (Jena, GDR) spectrophotometer with 1-cm silica cells was used for the absorbance measurements.The pH measurements were made with a Mera-Elwro N 517 (Wroclaw, Poland) direct reading pH meter with a glass - calomel electrode assembly. A Pye Unicam SP 192 single-beam atomic-absorption spectrometer was used for iron deter- mination in the aqueous phase after extraction. Reagents Aliquat 336 (General Mills Chemicals, Inc., Kankakee, IL, USA) containing 93.3% m/m quaternary alkylammonium chloride was purified, to ensure that no iron was present, by shaking a 0.01 M solution in chloroform with equal volumes of doubly distilled water five times and then filtering the organic phase through a cellulose filter.All other reagents were of analytical-reagent grade. Procedure Extraction of ferron was performed by shaking equal volumes (5 or 10 ml) of aqueous ferron solution with a chloroform solution of Aliquat 336 of the same concentration for 10min in a cylindrical separating funnel. The organic phase was then filtered through a cellulose filter to remove any remaining aqueous solution and the aqueous phase was centrifuged. Ferron was determined in the aqueous phase spectrophotometrically at 286 nm (or at 435 nm, for higher reagent concentrations), using appropriate calibration graphs. Extractions of iron(II1) by Aliquat 336 plus ferron solutions of equimolar concentrations in chloroform were performed in a similar manner.Aqueous iron(II1) solutions were pre- pared from Titrisol-grade standard iron(II1) chloride (E. Merck, Darmstadt, FRG) solution adjusted to an appropriate pH value by adding dilute hydrochloric acid, acetate buffer or ammonia solutions. Iron(II1) was determined by atomic-absorption spectrometry in the aqueous phase after extraction, and using calibration graphs. Spectrophotometric measure- ments were performed usually at 465 or 610 nm against a blank solution of Aliquat 336 plus ferron in chloroform. Results and Discussion Earlier preliminary experiments indicated that Aliquat 336 is a stronger extracting agent for ferron than trioctylamine or trioctylammonium chloride ; therefore, this quaternary alkylammonium salt was used as a liquid anion exchanger in all the extraction experiments I I 0 1 2 3 4 Initial ferron concentration in aqueous SolutioniM x I O - ~ Fig. 1.Concentration of ferron in the organic phase us. initial ferron concentration in aqueous solution. (A) Extractant M Aliquat 336 in chloroform. (B) Ferron extraction by Aliquat 336 after subtraction of the results of ferron extraction by chloroform.1322 PRZESZLAKOWSKI AND HABRAT: EXTRACTION OF FE(III) FROM Analyst, VoZ. 107 performed in this work. Chloroform was chosen as a diluent owing to the better phase separa- tion obtained than with benzene solutions of Aliquat 336 when the tendency to form an emulsion was observed. Pure chloroform partially extracts ferron, with extraction co- efficients of 0.1-0.12.8 The extraction results for ferron presented in Fig.1 suggest that only the sulphonic group of the reagent is bound in an ion pair with the alkylammonium cation (taking into account the partial extraction of an excess of ferron by chloroform) and thus ferron should maintain the complexing ability for metal ions. It was found that the extraction process is very rapid, a 2-min shaking time was sufficient to reach equilibrium, and the organic phase after extrac- tion was colourless, in contrast with the intense yellow colour of aqueous ferron solutions. Ferron is almost quantitatively extracted by Aliquat 336 and extraction coefficients of higher than 400 were obtained when a low excess (10%) of liquid anion exchanger, relative to ferron, was used for the extraction. As was found previously,* chloride ions pass almost quantitatively into the aqueous phase after extraction of ferron by Aliquat 336, and the extraction of the reagent can be described by an anion exchange reaction : OH OH It is thus presumed that the ion pair composed of alkylammonium cation and ferron should form complexes with iron(II1) in the organic phase.It was found that the iron extraction process with Aliquat 336 and ferron in chloroform is not as rapid (10 min shaking time was necessary to reach equilibrium) as the extraction of ferron by Aliquat 336 (Fig. 2). The organic phase after extraction of iron is dark green and the colour intensity does not change even after 7-d storage of the organic phase in a closed vessel. It is well known that ferron forms complexes with iron in aqueous solutions very rapidly and for the maximum stable intensity of colour, a large excess of reagent is req~ired.~ Owing to the strong extraction of ferron by Aliquat 336, the possibility of iron complexation in the aqueous phase after contact with the organic solution of the reagent seems to be of low probability.Thus, the low rate of iron extraction can be explained by complex formation at the interface (additionally, the stepwise complexation of iron cannot be excluded) and then the complex is transported into the organic phase; each of these processes can limit the rate of the extraction. However, this supposition needs additional investigations concerning the extraction kinetics. 0.8 0.6 $ 0.4 01 + a 0.2 0 5 10 15 Tim e/m i n Fig. 2. Effect of phase contact time on iron extraction and on absorbance of the organic phase. Concentration of iron(II1) in initial aqueous solution (pH 2.5), 5 mg 1-l.Extractant, 1 0 - 3 ~ Aliquat 336 + ferron in chloroform. Absorbance was measured a t (A) 375 nm, (B) 465 nm and (C) 610nm.November, 1982 SOLUTION WITH ALIQUAT 336 AND FERRON IN CHLOROFORM 1323 M Aliquat 336 and ferron solutions in chloro- form are presented in Fig. 3 in the form of an extraction isotherm. Although the graph becomes significantly less steep at iron concentrations in the organic phase equal to approxi- mately 3.3 x M, an excess of iron is extracted relative to the amount corresponding to the complex FeL33-(R4N+) at higher initial metal concentrations in the aqueous solution. The results for iron(II1) extraction by 0 2 4 6 8 10 Iron concentration in aqueous phase/M X Fig.3. Extraction isotherm for iron(II1). Extractant, 1 0 - 3 ~ Aliquat 336 + ferron in chloroform; pH of aqueous phase, 2.4-2.6. Because it is known that ferron in aqueous solutions with iron(II1) forms the complexes FeL, FeL, and FeLZ0 and that only anionic complexes can be extracted by liquid anion exchangers, the extraction of iron(II1) by an ion pair composed of ferron and alkylammonium cation can be described by the two following reactions: The complex anion FeL33- is predominant in the organic phase if an excess of extrac- tant is used for iron extraction: (2) Fe3+ + 3R4N+HL- + FeL33-(R,N+)3 -I- 3H+ . . - * (2) (ii) Besides the complex FeL,3-, the anionic complex FeL,- can be present in the organic phase if an excess of iron is present in the aqueous solution, which also contains chloride ions : Fe3+ + C1- + 2R4N+HL- + FeL2-R4N+ + R4N+C1- + 2H+ - * (3) It should be noted that after extraction the aqueous phase was bright green (the charac- teristic colour for iron complexes with ferron in aqueous solutions) when a large excess of iron relative to the extractant was present in the aqueous solution, which suggests the possibility of the following additional reactions : Fe3+ + C1- + R4N+HL- + R4N+C1- + FeL+ + H+ .. - ' (4) Fe3+ + 2Cl- + 2R4N+HL- $ 2R4N+C1- + FeL,- + 2H+ . . * * (5) Fe3+ + 3C1- + 3R4N+HL- + 3R4N+C1- + FeL33- + 3H+ - * (6) The possibility of such reactions seems to be indicated by the reaction of the ion pair composed of alkylammonium cation and strongly hydrophyllic tiron (1,2-dihydroxybenzene- 3,5-disulphonic acid disodium salt) with aqueous iron(II1) solution.In this instance the blue complex of iron with tiron was formed only in the aqueous phase.ll Iron(II1) is almost quantitatively extracted when a sufficient excess of extractant is used at pH values higher than 1.6 (see Fig. 4) and the slope of the straight line of the relation- ship between the logarithm of extraction coefficient and pH indicates that in this instance the complex Fe(R,N+L-), is predominant in the organic phase.1324 PRZESZLAKOWSKI AND HABRAT: EXTRACTION OF FE(III) FROM Analyst, VoZ. I07 100 8 80 75 tj 60 E 40 - 20 +- C 2 0 100 c C Q, 0 .- .- 5 lo s C 0 .- c z CI $ 1 1 2 3 0 1 2 PH Fig. 4. Percentage of iron extracted and extraction co- efficients for iron as a function of pH of the aqueous phase.Initial iron(II1) concentration in aqueous solution, 5 mg 1-1. Extractant, M Aliquat 336 + ferron in chloroform. Absorption Spectra Absorption spectra for aqueous ferron solutions as well as for organic solutions containing equimolar concentrations of ferron and Aliquat 336 in chloroform are shown in Fig. 5. The absorption maximum at 440 nm (characteristic of aqueous ferron solutions) disappears after extraction of the reagent by Aliquat 336 in chloroform and an absorbance of lower than 0.002 was found at wavelengths higher than 440 nm for freshly prepared 4 x 10-3 M Aliquat 336 plus ferron solutions in chloroform. It should be noted that the absorbance of organic ferron solutions in the wavelength range 400-600 nm was not changed after storage for 48 h.Wavelengthinm Fig. 5. Absorption spectra of aqueous ferron solutions (concentrations: C, 4 x M ; and D, 4 x 1 0 - 4 ~ ) and Aliquat 336 + ferron solutions in chloroform (ferron con- centrations: A, 4 x 10-5 M ; and B, 4 x 10-3 M). Blanks: water or Aliquat 336 in chloroform (4 x or 4 x M).November, 1982 SOLUTION WITH ALIQUAT 336 AND FERRON IN CHLOROFORM 1325 Absorption spectra for complexes of iron with ferron in aqueous solutions and for organic phase containing iron complexes with ferron and Aliquat 336 are shown in Fig. 6; the results were obtained at comparable experimental conditions (iron concentration 5 mg l-l, pH of initial aqueous iron solutions 2.6, ferron concentration in aqueous solution or in organic phase 4 x M and Aliquat concentration 4 x M).The lines for iron complexes with ferron in water (A) and in organic extracts (B and C) are parallel in the wavelength 1.5. e, 1.0 ? 2 a 0 C (0 a 0.5 - . A 4 . \ \ I I I I I 0 400 500 600 700 80( Wavelengthinm Fig. 6. Absorption spectra of iron (5mg1-l) complexes with ferron. A, Aqueous solution (ferron concentration 4 x 1 0 - 3 ~ ) measured against aqueous reagent blank : B, organic phase after extraction of iron complex with ferron with 4 x M Aliquat 336 in chloroform, measured against Aliquat 336 in chloroform; and C, organic phase after extraction of iron with 4 x M Aliquat 336 + ferron in chloroform measured against reagent blank. pH of aqueous iron solutions, 2.6. range 450-780 nm and show maxima a t 465 and 610 nm; however, the absorption coefficients for iron corresponding to these absorption maxima are higher for the iron complexes with ferron plus Aliquat 336.The strong absorbance at 465 nm, which is not suitable for iron determination using ferron in aqueous solutions, can be utilised for the extraction - spectro- photometric determination of iron, owing to the very weak absorbance of ferron in a chloro- form solution of Aliquat 336. The absorption coefficients for iron extracted by Aliquat 336 plus ferron in chloroform, corresponding to the three absorption maxima (370, 465 and 610 nm) decrease with the wavelength. However, at 370 nm a relatively strong absorbance for the extractant (Aliquat 336 plus ferron) was observed, see Fig.5. One experiment was also performed for the iron complex with ferron extracted from aqueous solution containing an excess of the ferron with 4 x M Aliquat 336 in chloroform; the strong absorbance between 360 and 370 nm (line B in Fig. 6) indicates that besides the complex anion FeL33- the free anion HL- is also extracted by the quaternary alkylammonium salt. Effect of pH on Absorbance The graphs of absorbance veysus pH shown in Fig. 7 indicate that the pH values for the development of maximum absorbance are between 1.6 and 2.9 and that the absorbance strongly decreases at pH values higher than 3, contrary to the absorbance versus pH graphs reported by Ziegler et a,?.* for iron complexes with ferron and tributylamine salt extracted with isoamyl alcohol (they found the constant, maximum absorbance to occur between pH 2.5 and 5 ) .1326 0.6 $ 0.4 c m -E 5: 4 0.2 a 0 PRZESZLAKOWSKI AND HABRAT: EXTRACTION OF FE(III) FROM Analyst, VoZ.107 A 1 2 3 4 5 PH Fig. 7. Influence of pH on absorbance of iron (5 mg 1-1) complex with Aliquat 336 + ferron measured at (A) 465nm and (B) 610nm. Extractant, 1 0 - S ~ Aliquat 336 + ferron in chloroform. 1.4 1.2 1 .o 6 0.8 + 3 0.6 a 0.4 0.2 a) 0 A 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 [Fe3+] [Fe3+ I+ [Aliquat 336 + ferron] Fig. 8. Stoicheiometry of iron complex with Aliquat 336 + ferron in the organic phase (continuous variation method). Measurements at (A) 465 nm and (B) 610 nm. Composition of Complexes in the Organic Phase The results obtained with the method of continuous variation (Fig.8) suggest the forma- tion of the complex FeL,&(R,N+), in the organic phase, in agreement with extraction data for iron (see Figs. 3 and 4) and with reaction 2. However, the shape of the graph for the extraction - spectrophotometric titration (Fig. 9) suggests the formation of two iron com- plexes with Aliquat 336 plus ferron, and to obtain a constant and maximum value of absorbance a large excess of extractant (about 15 times greater than the iron concentration in the initial aqueous solution) is required. It is therefore supposed that the extraction process can be described by reactions of stepwise formation of iron complexes with ferron and Aliquat 336 in the organic phase: Fe3+ + 2R4N+HL- + C1- + FeL,-R,N+ + R4N+C1- + 2H+ '. (7) FeL2-R,N+ + R,N+HL- + R,N+Cl- + FeL3&(R4N+), + H+ + C1- .. (8) The green colour of the aqueous phase [whose intensity increases with the iron concentra- tion in the initial aqueous solution (Fig. lo)], with an excess of iron relative to the extractant o.6/ 0.5 f--- 0.3 0.2 a / 0.1 G-' I 0 5 10 15 20 Aliquat 336 + ferron concentrationiM x Fig. 9. Absorbance of the organic phase measured a t 610nm us. Aliquat 336 + ferron concentration. Iron concentration in initial aqueous solution (pH 2.4), 10-4 M.November, 1982 SOLUTION WITH ALIQUAT 336 AND FERRON IN CHLOROFORM 1327 concentration, was also observed in these series of experiments, which confirms the possibility of the additional reactions (4), (5) and (6). Therefore, in further experiments the concentra- ted (4 x 10-3 M) solutions of Aliquat 336 plus ferron in chloroform were used for iron extrac- tion in order to ensure constant and maximum absorbance values.1.8 1 I I 1.6 1.4 Q) 1.2 .f! 1.0 s n 4 0.8 0.6 0.4 0.2 0 (D Iron concentrationipg ml-' Fig. 10. Absorbance of the organic phase measured at 610nm as a function of iron con- centration in aqueous solution (pH 2.4). Extractant, ~ O - , M Aliquat 336 + fenon in chloroform. Interferences The effect of some common cations and anions on the absorbance of the organic phase is given in Table I. The tolerance limits given in Table I could probably be increased, especially for metal ions forming complexes with ferron, if more concentrated solutions of Aliquat 336 plus ferron were used for iron extraction. Negative errors in the determination of iron occurred in the presence of amounts of magnesium, sulphate, phosphate and citrate ions larger than the limiting TABLE I EFFECT OF SOME CATIONS AND ANIONS ON THE DETERMINATION OF IRON Iron concentration, 5 ng 1-l; pH, 2.3-2.8; extractant, 4 x lo-, M Aliquat 336 plus ferron in chloroform.Foreign ion ~ 1 3 + . . .. Ca2+ . . .. Cd2+ . . .. Co2+ . . .. cu2+ . . .. K+ . . .. Mn2+ . . .. Na+ .. .. Nia+ . . .. Mg2+ . . . . Form added AWO,), COCl, cuso, MgC1, CaCl, CdC1, KNO, MnC1, NaCl NiCI, Tolerance limit* / mg 1-1 1000 2 000 50 2 000 5 2 000 200 1000 5 000 2 000 Foreign ion Zn2+ . . . . c1- . . . . F- . . . . HP0,Z- . . NO,- . . . . so,,- . . .. CH,COO- .. Citrate . . Tartrate . . Form added ZnC1, NaCl NH,F Na,HPO, NaNO, Na2S0, CH,COOH Trisodium citrate Disodium tartrate Tolerance limit*/ mg 1-l 50 5 000 190 10 5 000 80 5 000 40 1500 * Corresponding t o a 2% change in absorbance measured a t 610 nm.1328 PRZESZLAKOWSKI AND HABRAT : EXTRACTION OF FE(III) FROM Analyst, VoZ.I07 amounts indicated in Table I, whereas larger amounts of copper, cadmium and zinc enhanced absorbance. It is interesting that contrary to the method for the determination of iron with ferron in aqueous solution and contrary to the method recommended by Ziegler et aZ.,4 large amounts of cobalt, nickel and aluminium do not interfere in this determination of iron even without the addition of masking agents and only copper interferes if present in amounts minimally exceeding the iron concentration. Extraction - Spectrophotometric Determination of Iron in Natural Waters The proposed method was used for the determination of iron in two samples of natural waters and the results were compared with those obtained with conventional methods, i.e., atomic-absorption spectrometry and a thiocyanate spectrophotometric method (Table TABLE I1 DETERMINATION OF IRON IN WATER 11).The values are the means of five determinations on separate samples. Iron found, mg 1-1 A r \ Spectrophotometric methods , 1 Sample Proposed method Thiocyanate method AAS Drinking water (hardness at 20 “C) . . . . 0.484 f 0.007* 0.45 f 0.03 0.46 f 0.11 Mineral water, “Zuber”? .. .. . . 5.67 f 0.09 5.55 f 0.18 5.65 & 0.10 * Iron was extracted from a 50-ml sample into 10 ml of 4 X M Aliquat 336 plus ferron in chloro- t Declared composition: K+ 343, Na+ 7240, Li+ 12.8, Ca2+ 133, Mg2+ 307, Fe2+ 6.9, C1- 1180, Br- 2.42, form.S042- 77, HC0,- 19700 and H,SiO,- 37.1 mg 1-1. It should be noted that the precision of the proposed method for the determination of iron in drinking water (containing low amounts of iron) was markedly higher in comparison with the thiocyanate or the atomic-absorption spectrometric method owing to the distinctly higher values of absorbance obtained when extraction was performed at a phase volume ratio (Vaq: VOrg) equal to 5: 1. Recommended Procedure Prepare a 4 x 10-3 M solution of Aliquat 336 in chloroform and shake it five times with equal volumes of doubly distilled water in a separating funnel. Shake the purified organic solution with an equal volume of a 4 X M aqueous solution of ferron and after separating the phases filter the lower organic phase through a quantitative cellulose filter.The solution of extractant so prepared is stable for at least 1 week. Transfer 5-35 ml volumes of the stock solution into a series of 50-ml calibrated flasks and add 5 ml of acetate buffer solution (pH 2.2-2.8) to each. Make the solutions up to the mark with doubly distilled water. Treat the solution to be analysed with a few drops of 3% V/V hydrogen peroxide solution [to oxidise Fe(I1) to Fe(III)], warm to the boiling-point and cool to the ambient temperature. Adjust an appropriate volume to a pH value of between 1 and 4, transfer into a 50-ml cali- brated flask, add 5 ml of acetate buffer solution and make the solution up to the mark with doubly distilled water.Check the pH with a pH meter. Transfer equal volumes (5 or 10 ml) of aqueous iron(II1) solution and extractant solution into a small separating funnel, shake for 10 min and filter the lower organic phase through a quantitative cellulose filter. Measure the absorbance at 465 nm or at 610 nm (the latter is preferable if the analysed sample contains copper, zinc or cadmium) against the reagent blank. If the iron concentration in the analysed solution is lower than 1 mg 1-1, a somewhat modified procedure is recommended. A 45-ml aliquot of solution [iron(II) should be previously oxidised with hydrogen peroxide or ammonium persulphate] is transferred into Prepare a stock solution of iron(II1) chloride containing 10 mg of iron in 11. Check the pH with a pH meter.Determine the iron concentration with the aid of a calibration graph.November, I982 1329 a 100-ml separating funnel, 5 ml of acetate buffer solution and 5 or 10ml of extractant solution are added and the extraction is performed in the manner described above. Determine the iron concentration with the aid of calibration graphs obtained by the extraction of iron from aqueous working standard iron solutions (in the concentration range 0.1-1 mg 1-1 of iron) under analogous experimental conditions, i.e., at volume ratios of aqueous phase to organic phase equal to 5 : 1 or 10: 1. SOLUTION WITH ALIQUAT 336 AXD FERRON IN CHLOROFORM Conclusions The proposed extraction - spectrophotometric method seems to be useful for the deter- mination of iron owing to its simplicity, higher sensitivity (the molar absorption coefficients calculated from 30 independent measurements and corresponding to the three absorption maxima at 370, 465 and 610 nm have the following values: 7.7 x lo3, 6.86 x lo3 and 5.59 x 103 1 mol-1 cm-1) in comparison with the method using ferron in aqueous solutions.I t should be noted that the phases separate well after extraction and the absorbance of the free extractant (Aliquat 336 plus ferron in chloroform) is very low in the visible region. I t was found that Beer’s law is obeyed in the concentration range 0.1-10mg1-1 of iron and the precision of the method is sufficiently high. This work was supported by the Institute of Chemistry of the Maria Curie-Sklodowska University in Lublin (Grant No. MR.I.14.40.81). Thanks are due to Professor Dr. Edward Soczewihski for valuable discussions and for his interest in this work and to Mr. Jerzy Zukowski for partial assistance in experimental work. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. References Star$, J., KyrS, M., and Marhol, M., “SeparaCni Metody v Radiochemii,” Academia, Prague, 1975, Marczenko, Z., “Spektrofotometryczne Oznaczanie Pierwiastk6w,” PWN, Warsaw, 1979. Ziegler, M., Glemser, O., and Petri, N., Fresenius 2. Anal. Chem., 1956, 153, 415. Ziegler, M., Glemser, O., and Petri, N., Fresenius 2. Anal. Chem., 1957, 154, 170. Kurmaiah, N., Satyanarayana, D., and Pandu Ranga Rao, V., Anal. Chim. Acta, 1966, 35, 484. Przeszlakowski, S., and Wydra, H., Chromatographia, 1981, 14, 685. Przeszlakowski, S., Soczewinski, E., Kocjan, R., Gawecki, J., and Michno, Z., Polish Pat., P 219 538, Przeszlakowski, S., Kocjan, R., and Habrat, E., Chem. Anal. (Warsaw), in the press. Yoe, J. H., and Hall, R. T., J . Am. Chem. Soc., 1937, 58, 872. Lingaiah, R., Mohan Rao, J., and Seshaiah, U. V., Curr. Sci., 1967, 36, 197. Habrat, E., Diploma Dissertation, Medical School, Lublin, Poland, 1980. p. 138. 1979. Received December 7th, 198 1 Accepted March lSth, 1982

ISSN:0003-2654    

DOI:10.1039/AN9820701320

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

1330 Analyst, November, 1982, Vol. 107, pe. 1330-1334 Water Analysis by Inductively Coupled Plasma Atomic-emission Spectrometry after a Rapid Pre-concentration Michael Thompson, Michael H. Ramsey and Behrooz Pahlavanpour" Applied Geochemistry Research Group, Department of Geology, Imperial College of Science and Technology, University of London, London, SW7 2BP A method is described in which large batches of 10-ml water samples are pre-concentrated by evaporation and rapidly analysed for 16 elements at average river concentrations, by simultaneous inductively coupled plasma spectrometry. The effects of background interference, and its on-peak correction, on realistic detection limits of 30 elements were studied on solutions with high levels of calcium and magnesium and were found to place minor con- straints on the determination of some elements. Recoveries of 32 elements during pre-concentration were examined and 24 were found to be quantitative.The applicability of this method to the analysis of fresh water is considered in comparison with average river water concentrations and the EEC 1980 Council Directive. Keywords : Water analysis ; pre-concentration ; background interference ; inductively coupled plasma ; atomic-emission spectrometry We present a rapid method for the pre-concentration of fresh waters by evaporation, for subsequent simultaneous multi-element analysis by inductively coupled plasma atomic- emission spectrometry (ICP) . Ten-millilitre samples of water are evaporated in test-tubes in batches of up to 250 in an aluminium block bath held at 99 "C until about 1 ml remains.The remainder is analysed by ICP, variations in the final volume being compensated for by an internal standard of lanthanum added before the evaporation. The recovery of a wide range of elements at low levels is essentially quantitative. A batch of samples can be concentrated within a working day because of the small sample and the high working temperature. The method has the attractive feature that the sample has contact only with the one vessel from the time of collection to the final analysis. The ICP can provide a direct and very rapid analysis of surficial fresh waters for the following ten elements : sodium, potassium, magnesium, calcium, strontium, barium, iron, boron, sulphur and silicon. Other elements, including many toxic heavy metals that are subject to legislative control, have average abundances that are too close to their ICP detec- tion limits for satisfactory determination.Pre-concentration by a factor of ten or more brings a wider group of elements into a useful range, and methods which have been reported in connection with the ICP are solvent extraction,l ion e~change~-~ and coprecipitation.6~~ These methods, while possessing the virtue of selectivity against the major ionic constituents of waters and therefore minimising interference, are undoubtedly time consuming and exclude analyte species which are not in a labile ionic form [e.g., manganese(I1) and chromium(III)]. Pre-concentration by evaporation is generally regarded as an inferior method because of several factors, viz., possible loss of volatile species, possible contamination through the long exposure of the sample to the atmosphere in an open vessel, the lengthy procedure and the non-selectivity.However, we have shown that the scope of the ICP in the analysis of fresh waters can be considerably expanded by a ten-fold pre-concentration by evaporation. Experimental Spectrometer The ICP spectrometer used was an Applied Research Laboratories ARL 34000C with a 1-m vacuum spectrometer fitted with 36 lines, of wavelengths given in Table I. Data pro- cessing and instrument control were carried out by a dedicated PDP 11/04 computer. The ICP source was run a t a forward power of 1250 W with a viewing height above the load coil * Present address : Higher Institute of Technology, Brack, Libya.THOMPSON, RAMSEY AND PAHLAVANPOUR 1331 of 14mm.The nebuliser used was of the glass concentric type (Meinhard TR-30-3A) taking up 0.9 ml min-l of sample with a flow of humidified argon at 1.0 1 min-l. The liquid flow-rate was restricted by the use of a flexible polyethylene capillary uptake tube 350 mm long and of 0.5 mm i.d. No peristaltic pump was used. The tip of the nebuliser was washed between samples with 0.5 ml of de-ionised water containing 1% V/V Photo-flo to prevent a build-up of solids at the nebuliser tip and to reduce nebuliser noise. The spray chamber was a double-pass Scott type and the torch was a Fassel type with argon flow-rates of 12 1 min-l of coolant and 0.4 1 min-l of plasma gas. Heating Block The aluminium heating block, made by Scienco-Western Ltd.(Cambridge), was controlled thermostatically to h0.5 "C. For the fastest evaporation rate, the holes in the block should fit the sample tubes closely and expose less than 10 mm of the top of the tube to minimise reflux action. The block, with 250 holes, fitted into a typical fume chamber. To minimise corrosion of the block and sample contamination, the top of the block was covered with fresh aluminium kitchen foil for each batch of samples. Tubes used. followed by rinsing with pure water. Graduated 15-ml centrifuge tubes with conical bottoms, made of borosilicate glass, were They were pre-cleaned by prolonged refluxing of nitric acid in the heating block, Reagents water. chloric acid. acid. spectroscopy.Hydrochloric acid. Spectosol, 3676 m/m, diluted as appropriate with distilled, de-ionised Lanthanum chloride solution. Spectrosol, 100 g 1-1, diluted to 10 mg 1-1 with 1 M hydro- Calcium carbonate and magnesium carbonate. Specpure grade dissolved in hydrochloric Trace element solutions. Spectrosol, 1 g l-l, standard solutions for atomic-absorption Tip wash additive. Photo-Flo 600 (Kodak, Catalogue No. 3256773). Procedure Pipette 10.0 ml of each water sample or test solution into a centrifuge tube containing 1 .OO ml of the lanthanum solution. Place the tubes in the hot block pre-heated to 99 "C and remove them when the volume has decreased to between 0.5 and 1.0 ml (about 8 h). Make the volume up to 1.0 3 0.1 ml with pure water and mix. Analyse the resulting solution by ICP.Water samples should be filtered and acid-stabilised soon after collection. Calibration of the ICP The trace elements were calibrated in units of micrograms per litre using a 1 M hydro- chloric acid blank and a lo4 pg 1-1 multi-element standard. For reasons of chemical stability, elements in anionic form were segregated into a separate standard solution, as were the elements supplied as sulphates. Calcium and magnesium calibrations were prepared with six solutions each, from blank to 2500 mg 1-1 for calcium and to 300 mg 1-1 for magnesium, and fitted using the splined polynomial provided. The analytical procedure required a 20-s pre-flush followed by 3 x 5-s integration using a total volume of 0.6 ml of solution. Blank readings and sensitivities were checked after every ten solutions.Calibration and correction calculations were carried out via a specially written program. On-peak interference data for the effect of magnesium and calcium on trace elements were recorded at the same time as the solutions were run for major element calibrations. Only those values of apparent analyte concentration that were above the respective detection limit were used in the calculation of the interference coefficients. Linear or polynomial regression was used as appropriate to relate interference to major element concentration. Results were corrected for interference effects before ratioing to the internal standard.1332 THOMPSON et al. : WATER ANALYSIS BY Results and Discussion Analyst, Vol. 107 Detection Limits In order to investigate the performance and likely sources of error in the proposed method, synthetic trial solutions resembling fresh waters were concentrated and analysed by the procedure outlined above.Realistic estimates of the effect on detection limits of increasing background corrections were obtained by concentrating and analysing ten replicate samples of de-ionised water and ten of the synthetic fresh waters containing high levels of calcium (200 mg 1-I) and magnesium (30 mg 1-l) only. The resulting detection limits are given in columns 4 and 5 of Table I. The detection limits in column 3 are those estimated from ten contiguous sequential integrations of a single blank solution with no pre-concentration. Despite the more rigorous measure of detection limit for the pre-concentration method (Le., using replicate samples and not only replicate measurements), an approximately ten-fold improvement in detection limit was evident for most elements.TABLE I APPLICABILITY OF ICP TO WATER ANALYSIS Column 3 : detection limit (20) from 10 readings of blank solution by direct nebulisation. Column 4 : detection limit (20) from 10 replicate blank preparations by pre-concentration. Column 5: detection limit (20) from 10 replicate samples with 200 mg 1-l of calcium and 30 mg 1-1 All values of detection limits are approximate and can vary by 100% by random fluctuations. Column 6: the background interference from 200 mg 1-l of calcium and 30 mg 1-1 of magnesium Column 7: * elements giving low recoveries on spikes a t 50 and 500 pg 1-l; t elements giving low Column 8: average of median river concentrations from WedepohP and *Rose et d9 Question marks Columns 9 and 10: EEC guide levels (GL) and maximum admissible concentrations (MAC) of 1980.10 Column 11 : elements for which determination at average river levels is *suitable or tmarginal. Column 12: elements for which determination below EEC levels is *suitable or tmarginal.of magnesium prepared by pre-concentration. expressed in pg 1-1 of analyte; 0 signifies no measureable interference. recoveries on spikes at 50 pg 1-' only. signify uncertain or unknown values. Detection limits/yg 1-l Concentrations/vg 1-I Applicability L L r > / > - Wavelength/ By pre-concentration Average EEC Average nm --'-- Interference/ river (-'-, river Element (order) Direct Soft Hard Kg1-l Recovery water G 1, MAC water EEC 1 3 As Ba Be Bi Ca c o Cd Cr cu Fe HKg Li Mg Mn Mo Na Xi P Pb S Sb Se Sn Sr Te Ti V Zn Zr 2 328.1 x 2 308.2 x 2 193.8 x 2 455.4 x 1 313.0 x 2 223.1 x 2 317.9 x 2 228.6 x 3 226.5 x 3 267.7 x 2 324.8 x 2 259.5 x 2 194.2 x 1 766.4 x 1 670.8 x 1 279.0 x 2 257.6 x 2 281.6 x 2 559.0 x 1 231.6 x 2 178.3 x 2 220.3 x 2 180.7 x 3 206.8 x 2 196.1 x 2 190.0 x 2 407.8 x 1 214.3 x 2 337.3 x 2 311.1 x 2 202.5 x 3 319.6 x 2 3 2 50 30 4 0.1 30 60 5 2 3 2 40 4 100 1 100 10 5 50 8 20 30 70 80 80 7 2 30 60 2 7 3 4 0.3 15 2 0.4 0.02 2 5 0.6 0.2 0.2 0.2 8 0.6 9 0.1 10 1 0.6 30 0.8 3 4 8 5 8 0.6 0.2 2 5 0.2 0.8 0.2 5 0.5 6 1 0.3 0.03 4 0.5 0.3 0.8 0.3 5 1.5 9 0.1 2 4 20 0.9 2 4 57 6 12 0.4 0.1 5 7 0.1 1.4 0.9 - - 6 1.3 0 0 0 0.04 7.7 0.87 0.52 3.0 2.0 0 3.4 0 0 0 31 0 2.5 8.6 12.3 8.3 8.1 1.6 5.8 0 2.8 4.8 4.6 - - 1500 45 8 9 0.3 400 50 2 10 100 0.4? 0.005* 1.5 x 104 1 x 105 0.2 0.03* 1 7 100 loo* 50 0.07 2300 1 X lo1 4100 3 X lo4 3 7 20 1 6300 2 X lo4 1.5' 20 175 3.7 x 3 103 8.3 x 108 1 0.2 ? 3 0.9 ? ? 56 20 100 10 10 200 50 5 50 200 1 1.2 x 104 5 x 104 1.8 x 105 50 50 2 182 50 10 10 8.3 x 104 11 1 2 ' * * 1 t * * 1 : * * * * * * * * I * * * * * * * * * *November, 19S2 ICP - AES AFTER RAPID PRE-CONCENTRATION 1333 The effect of high levels of calcium and magnesium on the practical detection limit can be seen by a comparison of columns 4 and 5.The detection limits in column 5 are calculated in a way that includes both inaccuracy and imprecision in the background correction.To the normal two standard deviations of noise over ten samples has been added the absolute value of the correction bias. For example, sulphur has a detection limit of 70 pgl-1 by direct nebulisation, which is improved to 8 pg 1-1 by the pre-concentration method, in the absence of interfering elements. The interference, mainly from calcium, produced 1505 pg 1-I of apparent sulphur with a standard deviation of 20 pg 1-l. The total calculated interference correction is 1488 pg l-l, leaving an uncorrected residual of + 17 pg 1-1 of sulphur. The detection limit is therefore recorded as (2 x 20) + I +17 1 = 57 pg 1-l. A deterioration of the detection limit due to this cause is evident in a number of elements, notably sulphur and molybdenum. Although the uncertainties in the detection limits make rigorous interpre- tation difficult, there appears generally to be an increase in the detection limit equal to approximately 10% of the total background interference.The calcium and magnesium levels used in this study are approximately ten times higher than those in average river waters. The interference effects in river water analysis will therefore be proportionally reduced. Precision The precision measured using the ten replicate samples expressed as twice the coefficient of variation and averaged over 20 analytes was 8.0% at the 50 pg 1-1 level and 7.0% at the 500 pgl-1 level. The presence of 200 mg 1-1 of calcium and 30 mg 1-1 of magnesium increased these figures to 8.8 and 7.8y0, respectively, after interference correction. Loss of Elements During Sample Pre-concentration In order to study the recovery of 21 elements during sample pre-concentration, two levels of trace element spikes were added to ten-fold replicates of pure water.At the 500 pg 1-1 level only silver* (33% low) and antimony (17% low) showed recoveries that are signifi- cantly low at 95% confidence limits. At the 50 pg 1-1 level manganese (17% low), arsenic (16% low) and molybdenum (8% low) also showed values significantly below the spike added. Three other elements studied separately showed low results: titanium (up to 70% low), zirconium (up to 16% low) and beryllium up to (16% low). For these elements with high ionic potentials the low results probably represent loss of analyte by chemisorption or hydrolysis. For other elements low returns could be due to statistical inaccuracy and not losses (e.g., manganese).and was found to be quantitative in separate experiments with the pre-concentration of No significant loss of hydrochloric acid occurred during the evapora- Recovery of the principal ionic constituents (A13+, Ca2+, Fe3+, K+, Mg2+, Na+, natural water samples. t ion. Acidity Variation The maximum possible variation in the acid concentration in the final solution was -& loyo, and the effect of this both on the blank response and on sensitivity levels was examined for all of the analytes. By monitoring the effects in the range 0.5-1.5 M the error in the 0.9- 1.1 M region was established to be less than 1% relative error for all elements. Hencethe variation in acid concentration cannot affect detection limits or sensitivity.Applications of the Method Excluding elements that are partially lost during pre-concentration and those present a t too low a level, 16 elements can be determined by the proposed method in average river water, and are shown in column 11 of Table I. Of the other elements, molybdenum in soft water has a detection limit below the average river water concentration but interference (mainly from magnesium) and apparent small losses in preparation (8%) make the deter- mination impracticable by this method. Cliromium at average levels is similarly too close to the detection limits for reliable determination. higher than the level suggested by elementary solubility product calculations. * The solubility of silver in 1 M hydrochloric acid is greater than 1000 pg lkl, which is considerably1334 THOMPSON, RAMSEY AND PAHLAVANPOUR The pre-concentration method enables waters to be screened for 17 elements (column 12) for levels above the EEC 1980 Guide Levels (GL) and Maximum Admissible Concentrations (MAC).10 Arsenic has a detection limit well below the EEC MAC and correction for losses might be possible, but other methods, such as that of Thompson et aZ.,6 would be preferable for arsenic and also give values for antimony, bismuth, selenium and tellurium that also cannot be determined at the appropriate level by the proposed method.Conclusions The method described provides a rapid technique for measuring large numbers of small- bulk water samples for a wider range of major and trace elements than is possible by direct nebulisation into the ICP.Mercury (as HgC1,) and selenium (as H,SeO,) were recovered quantitatively from spikes during pre-concentration. In natural waters, however, they may be present in part as volatile methylated compounds and therefore more prone to loss unless additional pre-treatment of the sample is employed. This work was supported by a grant from the Natural Environment Research Council. 1. 2 . 3. 4. 5. 6. 7. 8. 9. 10. References McLeod, C. W., Otsuki, A., Okamoto, K., Haraguchi, H., and Fuwa, K., Analyst, 1981, 106, 419. Sturgeon, R. E., Berman, S. S., Desaulniers, J. A. H., Mykytiuk, A. P., McLaren, J. W., and Berman, S. S., McLaren, J. W., and Willie, S. N., Anal. Chem., 1980, 52, 488. Kerfoot, W. B., and Crawford, R. L., ICP Inf. Newsl., 1977, 2, 289. Smits, J., Nelissen, J., and van Grieken, R., Anal. Chim. Ada, 1979, 111, 215. Thompson, M., Pahlavanpour, B., and Thorne, L. T., Water Res., 1981, 15, 407. Hiraide, M., Ho, T., Baba, M., Kawaguchi, H., and Mizujp, A., Anal. Chem., 1980, 52, 804. Wedepohl, K. H., Editor, “Handbook of Geochemistry, Volume 1, Springer-Verlag, Berlin, 1969, Rose, A. W., Hawkes, H. E., and Webb, J. S., “Geochemistry in Mineral Exploration,” Academic Council Directive of 15 July 1980 relating to the quality of water intended for human consumption, Received April 5th, 1982 Accepted June 9th, 1982 Russell, D. S., Anal. Chem., 1980, 52, 1585. pp. 313-316. Press, London, 1979, pp. 549-581. 80/778/EEC, Off. J . Eur. Commun., No. L229/11-29.

ISSN:0003-2654    

DOI:10.1039/AN9820701330

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

Analyst, November, 1982, Vol. 107, pp. 1335-1342 1335 Rapid Flow Analysis with Inductively Coupled Plasma Atomic-emission Spectroscopy Using a Micro-injection Technique P. W. Alexander, R. J. Finlayson, L. E. Smythe and A. Thalib Department of Analytical Chemistry, University of New South Wales, P.O. Box 1, Kensington, 2033 N.S. W., Australia An introduction system for liquid micro-samples in inductively coupled plasma atomic-emission spectroscopy is described that allows the injection of 5-500-pl volumes into a rapidly flowing carrier reagent stream leading to the nebuliser. The effect on analyte signal was studied as a function of flow-rate, injection volume and sample concentration. It is shown that the carrier flow-rate determines the response time, sensitivity, precision and sample carry-over in the nebuliser.By the use of relatively rapid flow-rates of up to 7.5 ml min-l, fast injection of 10-pl samples is achieved a t an injection rate of 240 h-l with a relative standard deviation of 1.5% for a single-element analogue readout. Digital readout is used for multi-element determinations with similar or better precision. Detection limits of the order of 0.1 mg 1-1 are obtained for 10-pl injections, limited by the volume injected, with a proportionate decrease in detection limit for increasing volumes. Keywords ; R a p i d injection ; inductively coupled plasma ; seyum electrolyte Aqueous or organic liquid nebulisation techniques in common use for inductively coupled plasma atomic-emission spectroscopy (ICPAES) require several millilitres of solution and uptake rates are of the order of 1 ml min-l.When an adequate sample is available such usage rates present no special problems. However, in many environmental, clinical, forensic, toxicological, solid-state and other investigational areas, only micrograms to milligrams of the solid or microlitres of liquid are available. Techniques for the representative direct introduction of solid micro-samples into an ICP are being developed,l thereby avoiding more time-consuming sample processing. There are still several problems to be solved with the solid sample approach, including difficult representative sampling, matrix and volatilisation effects, pulse effects and plating out and memory effects in the transfer line to the plasma. However, solution processing that can be adapted to micro-samples has many advantages, including sub-sample homogeneity, possibilities of separation or concentration of the analyte, the use of internal standards or standard additions, the use of dilution techniques and various chemical modifications, including addition of matrix elements. Developments for the introduction of liquid micro-samples for ICPAES commencing in 1969 have recently been reviewed.2 Approaches include vaporisation of dried samples from graphite or metal filaments, introduction via micro-capillaries, cups or funnels, direct intro- duction into the plasma region after drying in a graphite cup and automatic flow methods.Automatic flow methods have been developed using flow injection analysis (FIA)3-6 and rapid flow analysis (RFA) .7 These approaches have the potential of very rapid multi-element analysis for microlitre samples.The FIA method described by Ito et aL3 used volumes of between 0.5 and 50pl and the determination of boron, copper and zinc in NBS orchard leaves was reported. Greenfield4 investigated the effects of controlled dispersion and delineated concentration - time profiles using FIA. Calcium was determined in Portland cement and the relative standard deviation (RSD) was 3%. Our preliminary report on the application of RFA5 has been extended and is now detailed. The novel aspect of our approach concerns the method of injecting samples into a peristaltic pump in order to damp out effects of pressure surges on the ICP pneumatic nebuliser system and simultaneous computer data processing.In previous electrochemical studies using RFA,7 electrode response times were shown to improve with increasing flow-rate. In this study with ICPAES, the transient signal pro- duced after each sample injection is also shown to be dependent on flow-rate, either by1336 ALEXANDER et al. : RAPID FLOW ANALYSIS WITH Analyst, VoZ. 107 monitoring with a recorder or by computer (multi-element mode). This approach of rapid flow into the nebuliser is shown to be potentially useful for fast multi-element analysis using ICPAES. Experimental Reagents and Standards Metal stock and standard solutions were prepared from Specpure metal samples (Matthey Garrett Pty. Ltd.) according to reference 8, yielding multi-element solutions in dilute (6% V / V ) aqua regia.All reagents were of analytical-reagent grade from BDH Chemicals Ltd. Instrumentation The ICPAES instrumentation and operating conditions are listed in Table I. TABLE I ICP INSTRUMENTATION AND OPERATING CONDITIONS Plasma power supply . . Xebuliser . . .. . . Labtest GMK Nebulisers (modified Babington type). Nominal sample flow-rate 0.2-2.0 ml min-1 Spectrometer . . .. . . Labtest V25 vacuum spectrometer, I-m Paschen - Runge, recipro- cal linear dispersion 0.46 nm mm-l, 21 channels Detector electronics . . Simultaneous A/D conversion of all 21 phototube signals and data processing performed in Labtest CRT 100 A multi-processor minicomputer system. Any one channel monitored via Labtest auto ranging amplifier to chart recorder (Houston Omniscribe, Type EB 5117-X5R) Sample argon, 0.9 1 min-l Forward power, 1400 W Reflected power, <5 W .. Labtest, Model 2000, 0.4-2 kW, 27.12 MHz, crystal controlled . . Plasma parameters . . . . Coolant argon, 12.0 1 min-l A conventional septum-injection valve (1 mm i.d.) from a normal liquid chromatograph was attached to the peristaltic pump, as shown in Figs. 1 and 2. The septum-injector was fitted to a Swagelock T-junction so that the sample could be injected into two separate reagents to give a single carrier stream, or the T-junction could be removed and the injector outlet attached directly to the pump to allow injection of the sample into a single carrier stream. SGE (Melbourne, Australia) micro-syringes (5-500 p1) were used throughout. The computer was programmed to print integrated emission counts for each element a t intervals of 3 s after sample injection for a total of ten discrete readings.A digital time response profile for each of the 21 elements could therefore be compiled. The computer was also programmed to calculate calibration plots, detection limits and precisi.on for each element after sampling a range of concentrations (1-100 mg 1-l) and a blank. Calibration could be performed using peak height or area, with and without internal standardisation. Sample injector 1 Carrier Reagent To :u m ,er Fig. 1. Injector for rapid flow analysis - ICPAES.NovcmbeY, 1982 ICP - AES USING A MICRO-IN JECTION TECHNIQUE 1337 Carrier reservoir I + i:?~i2 Septum injector I + Peristaltic pump i GMK nebuliser I + Plasma A "25 polychromator 750 monochromator (21 elements) Anv one A All 22 elements element to to CRT 1000A chart recorder computer for simultaneous data processing Fig.2. Schematic diagram of rapid-flow analysis with ICPAES showing the sequencing of sample and carrier stream p r i o r t o the peristaltic pump. The spectrometer was also fitted with an analogue emission signal readout, coupled to a chart recorder (Houston OmniScribe, Type EB 51 17-X5R) for single-element analogue monitoring. For this study, zirconium was selected as a test metal for monitoring the response of the nebuliser plasma system to change in aspiration flow-rate. A zirconium solution (5 mg 1-l) was injected into the nebuliser with the fixed polychromator wavelength set at 343.8 nm, and the zirconium emission intensity was monitored continuously on the chart recorder or the digital printer was used to print out the integrated emission signal at 3-s intervals.Flow Analysis Procedure The pump used was a Gilson peristaltic pump with ten rollers, three of which were in constant contact with the pump tubing at any given time to reduce pulsing. Liquid samples of volumes ranging from 10 to 500 pl were injected into the carrier stream, which was pumped continuously into the nebuliser. The pump tubing used in the peristaltic pump was of 1.02 mm i.d., but was replaced with tubing of 1.42 and 2.05 mm i.d. when higher flow-rates were required. The connecting tube to the pump was of 1.02 mm i.d. The flow-rate of the carrier stream was varied by use of the speed controller on the pump in conjunction with the above tubes.Flow-rates were varied by this method in the range 2.0-7.5 ml min-l. For a single metal determination in liquid samples, the samples were injected into the carrier stream and the resulting emission pulse was recorded on the chart recorder. The carrier stream used for all measurements was triply distilled water. Multi-element data, except for the recorder monitored element, was also printed out via the computer for the same sample slug injected into the nebuliser. For these studies, we investigated the effect of varying the injection volume on the emission intensity for a single metal sample, and then studied the effect of flow-rate on the emission response signal. At the maximum sampling rate, the precision, carryover, sensitivity and detection limits were determined for a single element injected in a small sample volume.The same parameters were then studied for multi-element analysis of the same injected sample. Results Effect of Sample Injection Volume continuously into the nebuliser at a fixed flow-rate of 2.0 ml min-l. The flow diagram is shown in Fig. 2. Fig. 3 shows the steady-state emission signal from aspiration of zirconium (5 mg 1-l) The effect of injecting1338 ALEXANDER et al. : RAPID FLOW ANALYSIS WITH Analyst, Vol. I07 I Time Fig. 3. Effect of sample injection volume on peak heights for 5 pg ml-1 of zirconium at a flow-rate of 2.0 ml min-'. small sample volumes into the carrier stream is also shown. The peak heights were found to approach the steady-state signal asymptotically as the injection volume was increased from 20 to 500 pl.Because of dilution and dispersion of the sample when injected into the stream, the peak height for small volumes (20 p1) was reduced by a factor of almost 10 in this system, giving reduced sensitivity. However, the use of micro-volumes is advantageous in many applications if the sensitivity and precision are sufficient, as the non-attainment of steady-state conditions means that washout is rapid, increasing the speed of analysis ; compare Fig. 4(a) with Fig. 4 ( b ) . ! C Ti I C e Fig. 4. Effect of flow-rate on peak-height precision at an injection rate of 240 h-l. (a) 10 p1 of 100 pg ml-l of zirconium injected at a samp- ling rate of 240 h-l. A, 2.0 ml min-l (RSD = 3.177;); B, 4.0 ml min-l (RSD = 2.0Sy0); and C, 7.5 ml min-l (RSD = 1.45%).(b) 300 p1 of 5 pg ml-1 of zirconium as in (a) A, RSD = 8.850,G; B, RSD = 2.43%; and C, RSD = 1.40%.November, 1982 ICP - AES USING A MICRO-IN JECTION TECHNIQUE 1339 Injection of a similar mass of analyte in 10- and 300-pl volumes indicates little change in precision (Fig. 4), except with a high injection volume and low flow-rate, where sample carryover becomes unacceptably high. Effect of Flow-rate on Response Time As in previous studies of electrode response times,' we found that the nebuliser response is considerably faster when the carrier flow-rate is increased. Fig. 5 shows emission response after injection of zirconium samples at flow-rates varying from 2.0 to 7.5 ml min-l. Con- siderable dispersion and dilution of the injection slug occurred a t 2.0 ml min-l, and approxi- mately 40 s were required for virtual (99%) washout of the sample from the nebuliser. The peak width was found to be inversely proportional to the flow-rate in the range studied, and also indicates improved sensitivity at the higher flow-rate due to less dispersion in the stream.> v) C a, C a, > m a, Lz c .- + .- .- c - Time Fig. 5. Effect of flow-rate on response time. 300 pl of 5 pg ml-l of zirconium injected at various flow-rates: A, 7.5; B, 4.0; and C, 2.0 ml min-l. These results show the advantage of using a high flow-rate with spectroscopic detectors, and indicate the possibility of faster sampling rates, because the rate of sampling is dependent only on the dispersion in the carrier stream and not on the detector characteristics.Precision and Carryover The precision of replicate peak-height measurements was also found to be dependent on the flow-rate of the carrier stream. Fig. 4(a) shows replicate peaks recorded for rapid injec- tions of zirconium (1.0 pg) at different flow-rates with 10-p1 injections made every 15 s. The slow flow-rate of 2.0 ml min-1 shows a poorer precision of 3.2% (RSD) compared with 1.5% at 7.5 ml min-l. The results in Fig. 4 also indicate the greater peak overlap occurring at the slow flow-rate. This effect caused carryover between consecutive high and low samples of approximately 5% at slow flow-rates, whereas Fig. 4(a) shows complete return to the base line for sample emission at a flow-rate of 7.5 ml min-l.These results indicate that an over-all sample injection rate of at least 240 samples h-l is permissible with rapid flow ICPAES analysis with sample injection volumes as low as 10 p1. Fig. 4(b) shows increasing carryover at higher injection volumes (300 pl). Multi-element Determination The ICPAES used in this study was programmed to give direct readout of 21 elements. The 13 elements listed in Table 11, contained in one calibrating solution, were used as model elements. The other elements on the polychromator (arsenic, potassium, phosphorus, sulphur, selenium, tungsten and mercury) in other solution groupings were not investigated in detail when it became apparent that the behaviour of all elements was similar and thus predict able. The spectrometer was programmed for rapid sample analysis to integrate the emission over ten successive 3-s periods, with a time delay after injection to synchronise a central 3-s period with the peak maximum (6-s delay for 2 ml min-l; no delay for 4 and 7.5 ml min-I).1340 ALEXANDER et al.: RAPID FLOW ANALYSIS WITH Analyst, Vol. I07 TABLE I1 MULTI-ELEMENT PRECISION (yo RSD) VeYSUS CARRIER FLOW-RATE Sample: 10 p1 of 100 mg 1-1 solution. RSD, % Element 2.0 ml min-l 4.0 ml min-l 7.5 ml min-l A1 . . .. . . 7.89 4.44 3.39 Ba . . .. . . 3.42 1.99 2.53 Ca . . .. . . 3.47 2.41 4.77 Cr . . .. . . 5.99 2.16 2.28 Fe . . .. . . 5.30 2.38 1.74 Mg . . .. . . 3.90 2.01 2.23 Mo . . .. . . 2.89 1.40 3.44 Na . . . . . . 6.01 6.32 1.53 Si . . .. . . 9.88 8.35 15.95 Sn . . .. . . 3.25 3.13 3.55 Ti .. .. . . 5.28 0.99 2.01 v . . . . . . 5.88 2.52 1.61 Zr . . .. . . 6.97 2.42 2.15 The net peak height was obtained by subtracting the first (base-line) integration from the maximum (peak maximum) integration. The peak height printed out by this method yielded precisions as low as 2-4% (RSD) (Table II), similar to the precision of the chart readout (Fig. 4). Hence, with 10-pl samples injected at the rate of 240 samples h-l, most of the 13 elements were determined with a precision of 1-3% (RSD). Improved precision could be obtained by measuring peak area (ca. 2% RSD) or by use of internal standardisation (see Table 111). The precision obtained (o.7-2y0 RSD) is signifi- cantly better than the 3-5y0 RSD reported for the micro-sampling technique of Knisely et aZ.10 for single-element determinations.TABLE I11 EFFECT OF QUANTITATION METHOD ON PRECISION Internal standard: 5 mg 1-I of barium. Flow-rate: 2.0 ml min-1. Quantitation method RSD,* yo Peak-height calculation . . . . .. . . . . 4.39 Peak-area calculation . . . . . . .. . . . . 1.93 Internal standard (Zr peak height/Ba peak height) . . 0.70 Internal standard (Zr peak area/Ba peak area) . . . . 0.61 * Six replicates of 300 p1 of 5 mg 1-1 zirconium solution. Detection Limits and Sensitivity The drawback of this injection technique is the loss of sensitivity as a result of dispersion in the carrier stream, as shown in Fig. 3. We determined the detection limit for a 300-,~1 injection of sample solution containing 13 elements as twice the standard deviation of the background emission counts.The results shown in Table IV range from about 0.01 to 0.2 mg 1-1, depending on the particular element, for concentration detection limits. Although the concentration detection limits are much poorer than those reported by Knisely et aZ.10 and Greenfield and Smithll for their micro-sampling technique, the absolute detection limits were better than 10 ng for many elements so that many of the samples suggested by Green- field4 and Greenfield and Smithll could be rapidly analysed by our rapid flow method. The linear working range is not degraded to any great extent owing to dispersion of the sample in the stream, and the useful linear range is approximately four orders of magnitude. Improved detection limits and a wider working range could be obtained in this system by further reducing dispersion and dilution of the stream.This would require tubing of small internal diameter throughout and a more rapid pump motor. Injection after the pumpATovember, 1982 ICP - AES USING A MICRO-IN JECTION TECHNIQUE TABLE IV 1341 DETECTION LIMIT FOR MULTI-ELEMENT DETERMINATIONS WITH MICROLITRE INJECTIONS 300-pl injection volumes with a carrier flow-rate of 4.0 ml min-l. Element Line/nm 396.1 A1 233.5 Ba 317.9 Ca 267.7 Cr 259.9 Fe 202.0 Mo 588.9 Na 288.1 Si 189.9 Sn 334.9 Ti 309.3 V 343.8 Zr 279.5 Mg Detection limit (2 x S.D.) - pg 1-1 ng 80 24.0 16 4.8 23 6.9 24 7.2 (120)* (36.0)* 3 0.9 25 7.5 200 60.0 700 210.0 36 10.8 13 3.9 36 10.8 17 5.1 Continuous nebulisation/ CLg 1-1 23 4 7 3 3 21 10 80 200 30 3 17 5 * High value due t o iron contamination from the syringe needle.(Fig. 6) just before the nebuliser does afford less dispersion and very rapid sample throughput (600 samples h-l), but at the cost of perturbing the plasma and even extinguishing it at higher injection volumes (100-300 pl). The use of a higher powered ICP as reported by Greenfield and Smithll should overcome the plasma instability problem. I-60 s-l B Time C Fig. 6. Injection after pump into a carrier stream of 7.5 ml min-1. 5 p1 of 100 pg ml-l of zirconium injected after pump. Injections per hour: A, 240 (RSD = 3.21%); B, 360 (KSD = 3.170,;); and C, 600 (RSD = 2.56%). Serum Electrolyte Analysis Application of the technique to serum electrolyte analysis using small sample volumes (10 p1). permitted a high throughput (240 samples h-l).Calibration against a 5% aqueous albumin matrix was necessary to overcome differences in nebuliser efficiency between the serum and the simple diluted acid standards.1342 ALEXANDER, FINLAYSON, SMYTHE AND THALIB TABLE V SERUM ELECTROLYTE ANALYSIS Injection . . .. . . 10 pl a t 240 injections per hour Standard . . .. . . 5% aqueous albumin matrix Carrier flow-rate . . . . 4.0 ml min-1 Concentration in control serum*/mg 1-1 Element RFA/ICP RSD,$ % Assigned value range* Na . . .. . . 3752 (4070)t 1.61 3 522-3 712 K . . .. . . 276 (156) 26.7 281-289 Ca . . .. . . 139 (120) 2.94 133-138 Mg . . .. . . 48.9 (44) 2.90 47-48 Fe . . .. . . 4.8 (-) 8.01 2.88 * Hyland Control Serum 11, Lot No. 0368N003AA. t Calibration against acid only (6% V/V aqua regia).$ From the mean of five determinations. Results on a control serum (Table V) indicate good agreement with established values even when using the least precise peak-height quantitation procedure. The required accuracy in this application did not require the use of the more precise peak area or internal standardisa- tion procedure. The poor precision obtained for potassium was caused by the low counts observed for the injected volume of 10 pl, for which the potassium concentration was close to the detection limit. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. References Cruz, R. B., Abercrombie, F. N., Murray, A. D., and Barringer, A. R., I C P Inf. Newsl., 1979, 4, 350. Smythe, L. E., Rev. Anal. Chem., 1982, in the press. Ito, H., Kawaguchi, H., and Mizuike, A., Bunseki Kagaku, 1980, 29, 332. Greenfield, S., Ind. Res. Dev., 1981, 23, 140. Alexander, P. W., Finlayson, R. J., Smythe, L. E., and Thalib, A., “6th Australian Symposium on Analytical Chemistry,” Australian National University, Canberra, August 23-28, 1981, Abstracts, Jacintho, A. O., Zagatto, E. A. G., Bergamin, H. F., Krug, F. J., Reiss, B. F., Bruns, R. E., and Alexander, P. W., and Seegopaul, P., Anal. Chem., 1980, 52, 2403. Purdue, L. J., I C P Inf. Newsl., 1980, 6, 49. Geiss, K. C., McKinnon, P. J., and Knight, T., I C P Inf. Newsl., 1981, 6, 481. Knisely, R. N., Fassel, V. A., and Butler, C. C., Clin. Chem., 1973, 19, 807. Greenfield, S., and Smith, P. B., Anal. Chim. Ada, 1972, 59, 341. p. 75. Kowalski, B. R., Anal. Chim. Ada, 1981, 130, 243. Received November 27th, 1981 Accepted June 4th, 1982

ISSN:0003-2654    

DOI:10.1039/AN9820701335

出版商:RSC    

年代:1982

数据来源: RSC