ISSN:0003-2654    

DOI:10.1039/AN97398FX017

出版商:RSC    

年代:1973

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97398BX019

出版商:RSC    

年代:1973

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97398FP049

出版商:RSC    

年代:1973

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97398BP053

出版商:RSC    

年代:1973

数据来源: RSC

摘要:

MAY, 1973 Vol. 98, No. 1166 THE ANALYST Precise Coulometric Determination of Acids in Cells Without Liquid Junction Part I.* Introduction and Instrumentationt BY E. BISHOP AND M. RILEY: (Chemistry Dejbartnzent, University of Exeter, Stocker Road, Exeter, E X 4 4QD) The requirements of high-precision amperostatic coulometry in the absolute mode are summarised, and the apparatus and instruments used described. Three constant-current sources are appraised : a 10-mA operational amplifier circuit, a commercial coulornetric titrimeter and a commercial analogue computer high-current source. The first and last have shown noise and drift levels that are acceptable in high-precision work. COULOMETRY, particularly amperostatic coulometry, is capable of producing analytical results of very high precision and accuracy, provided that the following three conditions are fulfilled: firstly, that the desired electrode reaction should proceed at a known high current efficiency; secondly,.that no analyte species should escape from, and no electroactive species should enter, the working compartment of the cell; and thirdly, that a means of locating the point of completion of the reaction, of commensurate precision and accuracy, should be available. The only electrode processes theoretically capable of attaining exactly 100 per cent. current efficiency are those involving electrolysis of the medium (solvent or molten salt) and its ions, and then only if the products of reaction are totally prevented from reaching the other electrode. For such reactions in the amperostatic mode the solvent molecules act as the intermediate, and when the limiting current for, for example, the reduction of hydrogen ion in aqueous solution, is exceeded by the total current, as must happen as the reactionapproachescom- pletion, the potential of the working electrode will change by about 800 mV as reduction of water becomes the dominant process.This change must be taken into account when assessing the influence of potentially electroactive impurities in the solvent or supporting electrolyte. It must further be noted that un-ionised weak acids are directly reducible a t a working cathode and un-ionised weak bases are directly oxidisable at a working anode. Acid - base reactions in a pure supporting electrolyte dissolved in a pure solvent are therefore attractive, and have accounted for more than half of the high-precision coulometric determinations so far made.Most of these have been conducted in multi-compartment cells,l-5 but some simpli- fication has been reported.6 The complexity of the manipulation, the difficulties of main- taining an effective seal between compartments and the limitations imposed by the low currents used make these techniques unattractive for routine work on standards. The alternative to separating the anode and cathode compartments is to change the nature of the auxiliary electrode process to one that proceeds a t very high efficiency at a potential less negative or positive, as required, than the complementary reaction to the main reaction. Both electrodes can then be accommodated in the same vessel, liquid junctions and salt bridges are eliminated, and the cell resistance is greatly reduced.This alternative, together with the use of high generating currents, which reduce background errors, has been explored for high-precision cathodic acidimetry.' The main cathodic reaction in an acidic solution free from reducible impurities at a platinum electrode is8 i c i, .. .. (1) 2H,0+ + 2e + H, + 2H,O .. * For Part I1 of this series, see p. 313. t Presented at the Second SAC Conference, 1968, Nottingham. $ Present address : Electronic Instruments Limited, Hanworth Lane, Chertsey, Surrey. @ SAC and the authors. 305306 BISHOP AND RILEY : PRECISE COULOMETRIC DETERMINATION [Analyst, Vol. 98 until the cathodic current exceeds the diffusion limited value for the hydrogen ion, when reaction (1) is joined by- i C 2H,O + 2e + H, + 20H- .... (2) %a so that the solvent acts as the intermediate in reaching the equivalence point. At a platinum anode in the same solution, reactions (1) and (2) proceed in the reverse direction, and, in addition, are joined by the reverse directions of i C i a 0, + 2H,O + 4e + 40H- .. .. * (3) .. .. (4) 0, + 4H,0+ + 4e + 6H,O in alkaline media and i C .. it3 in neutral or acidic media or when the anodic current exceeds the diffusion limited value for hydroxyl ion. Reactions (3) and (4) will proceed in the forward direction at a clean platinum cathode. The presence of oxygen in the catholyte does not lead to a loss of current efficiency in the determination of acids when the platinum cathode is clean, but when the cathode becomes dirty it no longer catalyses the immediate decomposition of hydrogen peroxide, an intermediate in reactions (3) and (4), and there is then the risk of some hydrogen peroxide diffusing away from the electrode surface and upsetting the acid - base balance of the bulk of the solution.Hydrogen can be removed by purging with nitrogen. The auxiliary reaction must proceed at a working potential much less positive than is required for reactions (3) and (4), while the supporting electrolyte required for it must not affect the hydrogen-ion concentration of the solution and must be suitable for reactions (1) and (2). The electrode reaction must neither introduce nor remove any entity that can affect the acid - base balance of the solution and must not introduce any entity that would be active at the cathode, thereby reducing the current efficiency of reactions (1) and (2).The best reaction is one that gives a solid product that adheres completely to the auxiliary electrode, and an obvious choice is the deposition of halide on a silver anode. This reaction was used in the pioneering work of Szebell6dy and SomogyiQJO and has since been repeatedly used by other workers. It has also been used as the auxiliary system in multi-compartment cells.1 Apart from this last application, methods of determination that involve the use of the principle have not set out to attain particularly high accuracy, 0.3 to 3 per cent. being common. Location of the end-point in high-precision coulometry has mainly been effected by means of zero-current potentiometry.The sensitivity of this process is limited by the effective Q of the reaction under the end-point conditions,l1Pl2 which is rarely high enough for the parts per million precision level. The most sensitive and precise method of end-point location is d.c. differential electrolytic potentiometry, or, even better, time-biassed periodic differential electrolytic potentiometry. The major attenuation of the precision of the results usually depends on the measurement of the generating current. Absolute coulometry, in which the Faraday constant is used as the ultimate standard, is dependent on the following factors: an adequately exact determination of the Faraday constant (the current value13 of 96 486.70 & 0.5 A s mol-1 may be conservatively rated at about 5 p.p.m.) ; the rationalisation of electrode processes, the determination of electrode kinetic parameters and the evaluation of current efficiencies11,12,14~1~ ; and experimental verification by high-precision determina- tions, of which there is a considerable bodyl2J6 to which the present study contributes.The Bishop Report16 proposing the adoption of the Faraday constant covers all these points and has been accepted by I.U.P.A.C. An extensive study has been made17 of cell and electrode design, of constant-current instrumentation and measurement, of auxiliary reactions and the errors introduced thereby, and of the precise location of end-points in high-precision assays of reference standard sulphamic acid of grade C.1839 In this paper the experimental methods and instrumentation are described and the constant-current sources and timing device are evaluated.In further papers, the auxiliary electrode processes, the working electrode processes, including the “silver error,” and the high-precision assays will be discussed. The choice of auxiliary reaction is exacting.May, 19731 OF ACIDS I N CELLS WITHOUT LIQUID JUNCTION. PART I 307 EXPERIMENTAL Volumetric operations were performed with Grade A calibrated glassware ; less critical weighings were carried out on a single-pan, five-place, constant-load balance (Stanton CL3), while critical weighings were carried out on a specially constructed free-swinging balance that had a standard deviation of 1.6 pg for loads up to 100 g.AnalaR reagents were used except for the specially purified sulphamic acid. WATER- throughout the work unless otherwise specified. CARBON DIOXIDE FREE WATER^^--- Carbon dioxide free water was prepared in 2-litre batches by rapidly boiling water in a Pyrex glass vessel. The lower part of the vessel was then cooled, without stirring, so that the top 30 to 50-mm layer remained close to the boiling-point. The contents were then rapidly transferred to a conditioned polythene bottle under a stream of white-spot nitrogen. The bottle was closed with a soda-lime guard-tube and was furnished with a polythene outlet tap. Such water was stored for a maximum of 12 hours, any that remained unused then being discarded. This water, blanketed by nitrogen or oxygen, was used for the preparation of electrolyte solutions and for washing electrodes and cells.ELECTRODES- Platin~m-Platinum-wire electrodes were made by welding 35 mm lengths of 22 s.w.g. grade C platinum to tinned copper connecting leads, sealing them into sheaths of soda-glass tubing and trimming them to a length of 25 mm. Gauze electrodes were of the cylindrical Fischer type (Johnson Matthey). The larger electrode (72020) weighed 31 g, the cylinder having a diameter of about 45 mm and an apparent surface area of about 125 cm2; the smaller (72050) weighed 19 g, was 32 mm in diameter and of 70 cm2 surface area. When necessary, the electrodes were cleaned by immersion for 2 to 3 minutes in freshly prepared aqua regia, washed in several batches of boiling water, and stored in water.Pre-treatment will be dealt with in Part 111. Silver-Silver-wire electrodes were made from 22 s.w.g. mint silver. A thread of blue glass was wound round the heated metal to form a sealed bead, which was then sealed into soda-glass tubing and annealed. The exposed wire was trimmed to a length of 25 mm. Large silver electrodes were made by manually winding 10 s.w.g. mint-silver rod on a mandrel mounted in a lathe so as to give cylindrical, helical coils with a stem for electrical connection. The coils were wound as tightly as possible so that, on removal from the lathe, their inside diameters were only slightly larger than the outside diameter of the mandrels, while adjacent turns were slightly separated, leaving space for liquid to circulate between them.Two sizes, designated A and B, were made, having the characteristics shown in Table I. Water was distilled in an automatic still with quartz condensing surfaces2* and was used TABLE I SILVER COULOMETRIC ANODES Size A Size B Over-all length/mm . . .. .. .. 145 135 Cylinder length/mm . . .. .. .. 48to53 30 to 35 Cylinder internal diameterlmm . . .. 25 to28 50 to 55 Mass/g .. .. .. .. .. 115to 120 155 to 165 Apparent surface area/cm* . . .. .. 125to135 170 to 180 Number of turns in cylinder . . .. 13to 14 9 to 10 Apparent surface areas (neglecting the roughness factor) were calculated from the known length of wire originally used, corrected for the length .of stem not immersed in the solution, the density of silver and the mass at the time, the mass giving the amount of silver consumed during use.Electrodes of area greater than 300 cm2 were made by mounting a size A electrode inside and concentrically with a size B electrode and making a common electrical connection. The large platinum-gauze cathode was mounted concentrically with and between the two silver helices. Silver electrodes were cleaned by immersion for 30 s in 1 + 1 nitric acid solution, followed by thorough washing with water. Silver bromide coatings were removed308 BISHOP AND RILEY : PRECISE COULOMETRIC DETERMINATION [Analyst, Vol. 98 by immersion in 1.0 M potassium cyanide solution, and the electrodes then thoroughly washed with water. An amount of silver black, representing the excess of silver present in the bromide film, remained after the cyanide treatment; most of this deposit was removed during washing, and the residue dissolved in the nitric acid.After a final washing, the electrodes were stored in water. Antimony-Antimony-rod electrodes with glass sheaths were prepared by a modification of the casting technique.lg Suitable lengths of 4-mm outside diameter Pyrex tubing of 0.75-mm wall thickness were cleaned with chromic acid, thoroughly washed and pre-heated to 110 to 120 "C in an air-oven. Specpure antimony was heated in a crucible to just above the melting- point, a tube removed from the oven, clamped vertically with one end dipping in the molten antimony, and a 40 to 60-mm column of antimony drawn into the tube by suction from a pipette filler, all operations being carried out as quickly as possible.The tube was removed from the still molten antimony, and electrical connection made by dropping several short lengths of resin-cored solder into the tube, melting them by gently heating with a flame and inserting a length of tinned copper wire. Finally, the bottom 5 to 10 mm of the tubing was cut off square with a diamond wheel so as to expose a cross-section of antimony rod of area about 6 mm2. Temperatures are critical and the method as described gave leak-free electrodes in which the antimony rod was tight-fitting and bonded to an antimony mirror on the inner wall of the tubing. Higher or lower temperatures gave either a rod capable of sliding in the tube, or an antimony mirror together with a sliding core; leakage in such instances was obvious.The electrodes were stored in water and were never allowed to dry out. Prior to use, the antimony surface was wiped with moist paper tissue and rinsed with water. When necessary, a fresh antimony surface was exposed by cutting off another small slice of the electrode. Reference electrodes-Saturated calomel (S.C.E.) or saturated mercury(1) sulphate (S.M.S.E.) electrodes of high capacity, the surface area of the mercury being 30 cm2, were used (the S.C.E. had a potential of +245 mV and the S.M.S.E. +640 mV versus a standard hydrogen electrode at 20 "C). In each instance a double remote junction was used, and the salt bridge was terminated by a low-leakage ceramic plug that dipped into the electrolyte solution. COULOMETRIC CELLS- The vessels most often used were amber-glass reagent bottles of about 400-ml capacity that had their necks sawn off and the rims ground flat; these bottles were subjected to a prolonged leaching and conditioning process before being brought into use.They were then cleaned with chromic acid, very thoroughly washed with hot water and re-conditioned by storing them filled with water for 1 week, the water being changed twice a day. Subsequently, until further cleaning was required, cells were washed with hot tap water, well rinsed with water and stored filled with water when not in use. Close-fitting lids were machined from 11-mm thick Perspex sheet and drilled with suitable holes to accommodate rubber bungs carrying the various electrodes and other equipment. In some experiments, cells made from sawn-off 400-ml beakers were used.Magnetic stirring was employed, the follower being coated with PTFE. CONSTANT-CURRENT SOURCES- Separate sources were used for high currents up to 2 A and for low currents up to 10 mA. When in use, the sources were run continuously, dummy load resistors being switched into the circuits in place of the coulometric cell in the intervals between increments or deter- minations. The 2-A soztrce-The 2-A source was a mains-operated solid-state power unit (Solartron AS 141 1) that incorporated a silicon controlled rectifier (SCR or thyristor) circuit designed to operate either as a constant-voltage supply in the range 0 to 40 V, or as a constant-current supply in the range 0 to 2.2 A. In the latter mode, the output voltage is directly proportional to the external load resistance, up to the value pre-set on the voltage limit control.If the load resistance rises to too high a value, the unit switches over to the constant-voltage mode, supplying a current inversely proportional to the resistance. The maximum cell load for 2 A at the maximum voltage setting of 40 is therefore 20 0. A small coil of resistance wire, having a resistance of 0.6 Q, was used as the dummy load.May, 19731 OF ACIDS IN CELLS WITHOUT LIQUID JUNCTION. PART I 309 The 10-mA sowce-The 10-mA source was constructed from chopper-stabilised analogue computer valve amplifiers (Solartron AA 1023) of d.c. gain lo6 (120 dB) and maximum outputs of 100 V and 12 mA. The circuit is shown in Fig. 1. Amplifier A,, in the constant-voltage configuration, gives a reference output voltage without imposing a significant current drain on the standard cell VR (Mallory cell, Type 303114,l-35 V).The output voltage ( E ) is given by and can readily be set by means of the 15-turn precision Helipot, R,. This produces a constant current in Rs in the input ts amplifier A, in the constant-current configuration, which maintains a constant current in its feedback loop through the coulometric cell. The current, I , is obtained from and with the component values used (R, = Rs = lo3 Q, 2 W, 1 per cent. high stability; R, = lo4 a, 15-turn; R, = 8.2 x lo3 Q, 2 W, 1 per cent. high stability) gives a range of currents from 1.5 to 15 mA, the voltage being limited by R, to within the saturation voltage of the amplifiers.The power supply to the amplifiers was a Solartron AS 853.3 unit. .. .. - * (5) E = VR (1 + R,IR,) .. .. - * (6) I = E/Rs = (V/Rs) (1 + R,/R,) .. Fig. 1. Operational amplifier 10-mA constant-current source (component values are given in the text) MAINS POWER SUPPLY- The sources already described, the AS 853.3 power unit, the pH meters, the crystal clock, the recorders and all other mains-driven instruments were supplied from a very low distortion, saturable reactor, constant-voltage transformer (Volstat CVN 500A, Advance Electronics), which was loaded to within 1 per cent. of its full output by adding resistance wire elements to the load. This technique ensures that the transformer is operating under optimum conditions. TIME MEASUREMENT- Electrolysis times were measured on a solid-state quartz crystal controlled clock (Venner TSA 3314) that had an operating frequency of lo4 Hz and made use of frequency divider circuitry.The over-all measurement range was 0.1 ms to 99 990 s in six ranges with a carry pulse at overflow, which was used to operate a solid-state driven electromechanical counter (Venner TS 8). The clock was calibrated, via a Marconi CR 100 communications receiver, against the standard frequency broadcast of station WWVH operated by the US. National Bureau of Standards in Hawaii. Errors of h 0 . l s in lo4 s were largely due to human reaction time in triggering, but over a period of 10 days the error was less than 1 s, corresponding to an accuracy of 1 p.p.m. In use, the clock was triggered by the action of switching sources from the dummy load to the electrolysis cell, and vice versa.310 BISHOP AND RILEY : PRECISE COULOMETRIC DETERMINATION [Analyst, vol. 98 ELECTRICAL MEASURING INSTRUMENTS- For fiotentials-All cell potentials and working or indicator electrode or differential electrolytic potentiometric potentials were measured on a vibrating capacitor input electro- meter (E.I.L.39A pH meter) that had a measured input impedance greater than 1014 $2. Com- prehensive backing-off circuitry and a recorder output to which back-off and band-spread could be applied permitted the expansion of 10 mV to 28 cm of recorder chart width. Voltage outputs from the sources were monitored on Sangamo-Weston S82 multi-range 1000 V a-1 voltmeters. For currents-Currents were monitored on multi-range Sangamo-Weston S82 meters.Accurate measurements were made by means of the voltage drop across standard, calibrated four-terminal resistors (Croydon Precision RS. 1) immersed in transformer oil. Mercury-in- glass thermometers immersed in the oil were used to determine the running temperature for reference to the resistance - temperature calibration. The voltage drop across the standard resistors was measured with a Croydon Precision P3 potentiometer, on which potentials could be measured to &5pV after calibration of the galvanometer, which showed a maximum sensitivity of 0.3 mm pV-l. Recorders-Honeywell-Brown 153 X 17, 10 mV, 2 s, strip-chart recorders with a range of gear boxes, supplemented by a 0-25-s high-speed version when required, were used.The X - Y recorder was a Houston EHR 921, with a 280 x 215 mm platen. Currents were con- verted into potential inputs by the use of calibrated resistors. Potentiostat-A solid-state potentiostat (Southern Analytical Wadsworth A.1654) was occasionally used. Coulometric titrator-A commercial coulometric titrator (Thorn Electrical TE 110) was evaluated together with the constant-current sources. This titrator had a maximum output of 200 mA at 14 V, derived from a classical series-regulated circuit, and incorporated a machine integrator adjustable to give a read-out in micrograms of determinand. APPRAISAL OF CURRENT SOURCES- The current was allowed to flow through a chain of standard resistors, with or without the inclusion of the coulometric cell, and the voltage drop across a part of the chain was measured by using the 39A meter as a transducer.The major portion of the voltage drop was backed off, and the remaining small portion recorded on the 10-mV recorder. The calibrated backing-off supply of the 39A meter is electronic and is less stable than the mercury battery driven buffer controls. External Mallory RM 42 R batteries were used to supply most of the backing-off potential and the meter's buffer controls were used for fine control. The recorder was set to a centre zero and 4-45 mV of the total 1.35 V was recorded. Noise is defined as excursions from the mean trace of duration less than 1 minute, and drift is defined as a long-term departure from the zero setting. The inherent instability of the current measuring and recording system produced a noise of about -J=0.05mV over 24 hours.The supply noise and drift detectable above the background noise therefore corre- sponded to *5 and &lo p.p.m., respectively. The time constant of the measuring system was about 1 s, so that fast transients were not recorded. These, and the noise, were examined on a calibrated measuring X - Y oscilloscope (IIewlett-Packard 130 C), which, although designed for a maximum of 500 kHz, maintained a reasonable response to beyond 60 MHz. RESULTS AND DISCUSSION THORN TE 110 COULOMETRIC TITRATOR- Tests were made at three current levels, 10pA, 10mA and 200mA, over periods of 2 hours, during which the laboratory temperature was constant to within 0.5 "C. Noise at all levels was usually within h 0 .l per cent. of the current, and the drift was unidirectional upwards at the two higher currents and downwards at 10 PA, and fluctuated between 0.05 and 0.35 per cent. overall. The performance of the complete unit was also assessed by carrying out replicate determinations of ceriurn(1V) with electrogenerated iron( 11) and of arsenic(II1) with electrogenerated bromine ; the current efficiencies of the generation reactions were known to be essentially 100 per cent. For amounts of sample of the order of lO-*mol the relative standard deviation was not better than -+0*3 per cent. It was concluded that this instrument would be suitable only for uncritical routine use when a reproducibility of 0.5 to 1 per cent. would suffice.May, 19731 OF ACIDS IN CELLS WITHOUT LIQUID JUNCTION.PART I 311 THE OPERATIONAL AMPLIFIER SOURCE- A single-amplifier unit was first constructed, in which A, was replaced by either a 2-V accumulator or the Solartron AS 1411 power supply in the constant-voltage mode. The performance was not satisfactory, and the instability observed is attributed almost entirely to the instability of the reference voltage source under the current drain of 10 mA. The operational amplifier reference voltage source A, was constructed and showed an excellent performance at a current output of 10 mA. Over periods of 2 to 5 hours the noise was less than 5 p.p.m. and the drift less than 10 p.p.m. The two circuits, A, and A,, were then combined as in Fig, 1, and the over-all performance was found to be scarcely inferior to that of the reference voltage source, A,.Tests were carried out at currents of 5 and 10mA under ordinary open laboratory conditions without taking any precautions about temperature or draughts. Noise and drift levels were about &5 and &lo p.p.m., respectively, that is, hardly detectable above the background. This source is therefore entirely adequate for high-precision coulometry. THE SOLARTRON AS 1411 SOLID-STATE 2 A SUPPLY- Preliminary tests at a current of 1 to 2 A showed noise levels of less than &20 p.p.m. and drifts of less than &lo0 p.p.m. over periods of 1 hour. These tests showed that the instrument was highly sensitive to temperature changes (50 p.p.m. “C-l) and particularly sensitive to airflow caused by draughts, which altered the temperature of the heat sinks of the power transistors. An overnight “constant draught” test showed a typical bow-shaped curve, indicative of the slow drift with temperature changes.At a current of 2 A, noise levels were -+lo p.p.m.; drifts over 2 to 5 hours were &40 p.p.m., and over 1 to 2 hours were &25 p.p,m. The above test suggested that good performance could be expected if the temperature and airflow in the vicinity of the instrument could be maintained constant. However, when run in a large fan-circulation oven at ambient temperature, or in a large cupboard, the heat generated was sufficient to raise the air temperature to 40 “C, when, although the specification named an upper ambient limit of 50 “C, malfunctioning was apparent. A refrigerated constant-temperature room was not available at the time.Attempts were made to compensate for noise and drift by sensing the change in current level and by using an operational amplifier source to feed sufficient additional current into the circuit to give a constant summation current. Insoluble problems of interaction between sources were encountered, and it was concluded that the successful application of the additive technique required the interposition of some non-electrical current sensing system between them. Optical methods that make use of a mirror galvanometer and two photocells might be suitable, but the oscillation period of a galvanometer suspension is too long to sense noise. An attempt to improve the regulation of the 2-A source was made by using operational amplifiers to increase the gain of the automatic current control circuit, but the results were not encouraging.The best performance was obtained with the source in a corner of a seldom used laboratory, unenclosed but partially screened by other equipment. Drifts within &20 p.p.m. over periods of 90 to 150 minutes were attained. Thyristor circuits have inherently high noise levels, generated by the large switching pulses, and so the higher frequency noise was examined on a measuring oscilloscope at a sensitivity of 20 pV mm-l, the d.c. being backed off. With a current of 2 A flowing through a 1 Q standard resistor as the input, a 50-Hz ripple of rather distorted sine wave form and amplitude of 200 pV was observed. In each complete cycle at 50 Hz, two large positive- going and two large negative-going spikes of amplitude 50 mV were observed.Closer examination revealed that the spikes comprised bursts of oscillation at 1 MHz, the amplitude of which died to zero in 10 ps. The true noise was therefore greater than that indicated by measurements with sensors of long time constant. This, however, is not random “white” noise, but tends to average out close to zero over periods of 1 s or more; nevertheless, it is possible that potentiometric measurement of current may be subject to some bias. Overall, the unit performed satisfactorily over the range 20mA to 2 A, and it was decided to use it in acid assays, the benefit of the large current being that electrolysis times for 0-1-mol samples are reduced to about 40 minutes.The operational amplifier source, covering the current range up to 12 mA, would be used for the final 0-1 per cent. of the reaction, and differential electrolytic potentiometry for location of the end-point.312 BISHOP AND RILEY One of us (M.R.) is greatly indebted to the Charitable and Educational Trust of the Worshipful Company of Instrument Makers for financial support in the form of a Research S tuden tship. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. REFERENCES Taylor, J. K., and Smith, S. W., J. Res. Natn. Bur. Stand., 1959, 63A, 153. Marinenko, G., and Taylor, J. K., AnaZyt. Chem., 1968, 40, 1645. Marinenko, G., and Champion, C. E., Ibid., 1969, 41, 1208. Quayle, J. C., and Cooper, F. A., AnaZyst, 1966, 91, 355. Cooper, F. A., and Quayle, J. C., Ibid., 1966, 91, 363. Eckfeldt, E. L., and Shaffer, E. W., Analyt. Chem., 1965, 37, 1534 and 1581. Riley, M., and Bishop, E., Proc. Soc. Analyt. Chem., 1966, 3, 143. Bishop, E., Chemia Analit., 1972, 17, 511. Szebellkdy, L., and Somogyi, Z., 2. analyt. Chem., 1938, 112, 323. , Ibid., 1938, 112, 332. Bishop, E., in Shallis, P. W., Editor, ‘‘Proceedings of the SAC Conference, 1965, Nottingham,” W. Heffer & Sons Ltd., Cambridge, 1965, p. 291. -, “Coulometric Analysis,” Volume IID of Wilson, C. L., and Wilson, D. W., Editors, “Compre- hensive Analytical Chemistry, ” Elsevier Publishing Company, Amsterdam, 1973. Taylor, B. N., Parker, W. H., and Langenberg, D. N., Rev. Mod. Phys., 1969, 41, 375. Bishop, E., Analyst, 1972, 97, 761. -, “Report on the Status of the Faraday Constant as an Analytical Standard,” International Union of Pure and Applied Chemistry, Commission,,V-5, Pure A$@. Chem., in the press. Riley, M., “Some Studies in High Precision Coulometry, Ph.D. Thesis, University of Exeter, 1969. Analytical Methods Committee, Analyst, 1965, 90, 251. Bishop, E., and Sutton, J. R. B., Analytica Chim. Acta, 1960, 22, 590. Short, G. D., Ph.D. Thesis, University of Exeter, 1962. , -- -, Ibid., 1972, 97, 772. -, Ibid., 1967, 92, 587. Received December lath, 1972 Accepted January lst, 1973

ISSN:0003-2654    

DOI:10.1039/AN9739800305

出版商:RSC    

年代:1973

数据来源: RSC

摘要:

Analyst, May, 1973, Vol. 98,pp. 313-324 313 Precise Coulometric Determination of Acids in Cells Without Liquid Junction Part II.* The Silver - Silver Bromide Auxiliary Anodic Reaction? BY E. BISHOP AND M. RILEY: (Chemistry Department, University of Exeter, Stocker Road, Exeter, EX4 4QD) The solubilities of silver halides in solutions of their respective halides have been examined. The anodic deposition of silver bromide on silver electrodes has been studied, and mass-transfer limited current densities determined. The resistance of silver bromide films has been investigated, and the critical thickness at which the resistance suddenly rises was found to be inversely proportional to current density. The thickness of films in terms of coulombs per square centimetre and of micrometres was determined, and the specific resistance of various films measured. The magnitude of the “silver error” arising from solubility of silver bromide in bromide solutions and anodic stripping of silver ions to form a precipitate in the bulk of the solution has been established and found to be unacceptably high for precise work, although it can be reduced to less than 0-01 per cent.for less critical work. The behaviour of anodically deposited silver bromide films is explained on the basis of a porous deposit that becomes non-porous at the critical thickness. Conditions are chosen for determinations of 0.05 mol of a mono- basic acid a t high total currents that are well within the capacity of the silver anode. THE complexity of coulometric cells for high-precision work, and the degree of simplification achieved if the auxiliary electrode can be immersed directly in the test solution, have been mentioned in Part 1.l The requirements for this simplification were examined and found to be very restrictive; however, in one important determination, the cathodic assay of acids, it is possible to use a non-protolytic anodic auxiliary reaction of low potential.This reaction, which was first used by Szebellkdy and S ~ m o g y i , ~ ~ ~ involves the deposition of halide on a silver anode- i c is (equations are numbered in sequence from Part 11) where X- is chloride or bromide; iodide is not generally suitable as it tends to produce hard, non-porous films.* The limitation of this device is that the silver halide has a small but significant solubility in the supporting electrolyte, and its reduction to silver at the cathode competes with, and reduces the current efficiency of, the desired reduction of hydrogen ions and water molecules.Moreover, if the anodic current density exceeds the mass-transport limited current density of the halide ion, silver ion is anodically stripped and forms a precipitate with the halide in the bulk of the solution; the precipitate adheres to the cathode where it is reduced, thus further increasing the loss of current efficiency. The combined effect is hereafter called “the silver error.” Thus when the silver - silver halide auxiliary anode has previously been used in high-precision coulometric acidimetry, it has been kept isolated in a separate compartment .5 ~ 6 Szebellddy and Somogyi used chloride deposition in the determination of hydrochloric2 and sulphuric3 acids, but Lingane and Small’ preferred the use of the less soluble bromide, as did Carson and Ko8 for the determination of nitric acid. Lingane further used bromide in automatic determinations of hydrochloric acid,g as did Bishop and Shortlo in determining perchloric acid and in the micro-determination of perchloric and acetic acids. Results for .. - . (7) AgX + e + Ag + X- . . .. * For Part I of this series, see p. 305. t Presented a t the Second SAC Conference, 1968, Nottingham. @ SAC and the authors. Present address : Electronic Instruments Limited, Hanworth Lane, Chertsey, Surrey.314 BISHOP AND RILEY : PRECISE COULOMETRIC DETERMINATION [Analyst, Vol.98 the solubility of silver halides in solutions of their respective halide+ are presented in Fig. 1, which suggests that little advantage exists for bromide over chloride. The minimum solubility of silver chloride in chloride media occurs at chloride concentrations of between and 10-3 M, and of silver bromide in bromide media at bromide concentrations of between 10-3 and M. The solubility of silver bromide exceeds that of silver chloride when the respective halide concentrations become greater than 0.1 M, while silver iodide is more soluble in iodide media than the other halides in their respective media at all halide concentrations. 0 -2 -4 -6 Log,, IKXI Fig. 1. Solubilities of silver halides in solutions of the corresponding potassium halides. A ( x - - - x ) , X = Cl-, 25 "C; B (0-O), X = Br-,25"C; andC(0-.-0), X = I-, 20 "C.[KX] is measured in moll-1 Increase in solubility at high halide concentrations arises from complex formation. Assuming that the latter does not proceed beyond AgX,- formation, the reactions can be represented, together with their formation constants, as follows- KO .. .. * * (8) -k x- Agxsolid .. If So is the concentration of undissociated silver halide in the solution, and Ks is the solubility product of the silver halide, then .. .. .. .. . . (11) 1 KO = - KS and .. .. .. .. . . (12) SO KS K1 =- If no complexes higher than AgX2- are formed, then the solubility of the silver halide is12 S = - Ks + So + KIK,Ks[X-] . . .. . . (13) [X-I When the halide concentration [X-] is high (K, being small) the solubility of the silver halide increases linearly with [X-1.Forbes and Cole13 derived a similar equation for silver chlorideMay, 19731 OF ACIDS IN CELLS WITHOUT LIQUID JUNCTION. PART I1 315 in chloride media, and obtained values for So, representing the minimum, “intrinsic solu- bility” of silver chloride, of 6.1 x 1 0 - 7 ~ in hydrochloric acid and 6.3 x 10--7~ in sodium chloride solutions. The rate of dissolution of the silver halide may also be important in controlling the amount of dissolved silver in the cell electrolyte. This rate may be expected to depend on the surface area exposed to attack, that is, the particle size or film porosity, or both, of the halide, but, apart from work on photographic emulsions showing that the dissolution rate depended on both particle size and halide concentration,l4 no quantitative information is available.For the present purpose, the halide concentration must be kept low so as to avoid formation of soluble silver species, and yet sufficiently large to maintain the mass-transfer limited current above the working current so as to prevent formation of precipitate particles in the bulk of the solution: a large electrode area, with correspondingly reduced current density, is an obvious aid. For a variety of reasons, including reduced light sensitivity, the silver bromide system was chosen, and the purpose of this work was to establish conditions that involve the passage of 5000 C at a generating current of 2 A, which give the minimum “silver error,” and to examine possible means of further reducing the error.EXPERIMENTAL The equipment used has been previously described.l The coulometric cell and circuit are shown in Fig. 2. The salt bridge leads to an S.M.S.E., the platinum-gauze cathode was the large 72020 and the cell electrolyte consisted of a potassium bromide solution containing the aDDroDriate amount of suhhuric acid to simulate the acid. A kllelectrode potentials ’ire given with reference Clock r-3 Si I ,ver I anode to conditions for a determination o’f the standard hydrogen electrode -Platinum cat1 ,Stirrer bar lode Fig. 2. Coulometric cell and circuit for the study of the silver - silver bromide auxiliary anode system. V = Sangamo Weston S82 multirange voltmeter; and PSU = Solartron AS 1411 thyristor power supply unit, 2 A, 40 V RESULTS ANODIC OXIDATION OF SILVER IN BROMIDE~MEDIA- Tests were made of the electrolysis of potassium bromide solutions 0.08 M in sulphuric acid at 2 A by using silver anodes of areas 20 to 50 cm2. At bromide concentrations large316 BISHOP AND RILEY : PRECISE COULOMETRIC DETERMINATION [A%&?ySt, VOl.98 enough to support the anodic mass transport, the cell resistance rose rapidly until the output voltage limit of the power supply was exceeded and constant-current operation was no longer possible. At lower bromide concentrations the cell resistance remained relatively low, but the mass transport of bromide ions could not sustain the current and the potential of the silver anode rose rapidly to over + 1-8 V ; evolution of oxygen was observed and silver bromide was precipitated in the bulk of the solution. A potential - time curve plotted on the fast recorder is shown in Fig.3. At a current density of 70 mA cm-2 in 0.08 M bromide solution, after the initial almost instantaneous rise the potential rose slowly for about 10s as some silver bromide was deposited on the anode and some was precipitated in the solution. The potential of the anode then rose rapidly to a steady value governed by the evolution of oxygen [equation (4), Part I]. Repetition with a large silver anode of area 220 cm2 gave an initial rise of 30mV and then the potential of the silver anode rose steadily at about 1 mV s-1 for 500 s. > 1.5 > w. 5 0 p. in 0 ’ I +-- Chart travel Fig. 3. Anodic oxidation of silver a t 70 mA cm-2 0-05 M potassium bromide - 0.08 M sulphuric acid 0.2 0.4 0 6 Potential versus S.H.E./V Fig.4. Current - potential graphs for a silver anode: A, in 0.08 M sul- phuric acid; B, in 0.0038 M potassium bromide - 0.08 M sulphuric acid; and C , in 0.0067 M potassium bromide - 0.08 M sulphuric acid The limiting current for the discharge of bromide ion at the anode is15 nFDB,Ar [Br-IB I&. = -- .. .. .. . . (14) (l - tBr)aX E + + m &Is -+o where DBr 1 cm-1 s-1 is the thermal diffusion coefficient of bromide ion and 6% cm is the diffusion layer thickness. Subscripts S and B refer to concentrations at the plane of closest approach to the electrode surface, and in the bulk of the solution, respectively. The apparent electrode area is A cm2, the roughness factor is r and the transport number of bromide ion is tBr.A value of DBr = 2.06 x 1 cm-1 s-l was calculated from the Nernst expression16 . . (15) .. .. . o RT Z F 2 D = 10-3 - & . . where z is the charge number of the ion and A, IR-1 cm2 equiv-1 is the limiting conductance at infinite dilution. If a value of 6x = 3 x 10-3 cm in a well stirred solution17 is taken, then from equa$ion (14) the limiting current density for bromide ions at a bulk concentration of 0.01 M is 6.6 mA cm-2. Badoz-Lamblingls reports a value of 0.75 mA cm-2 in 0.001 M bromide solution, which is in fair agreement. Typical current - potential scans for a silver anode of A = 250 cm2 are shown in Fig. 4. Limiting current densities were proportionalMay, 19731 OF ACIDS IN CELLS WITHOUT LIQUID JUNCTION.PART I1 317 to bromide concentration: a mean value of 6.3 mA cm-2 in 0.01 M bromide solution is in good agreement with the calculated value. The calculated values for Fig. 4 are 2.5 and 4.4 mA cm-2 as against the measured values of 2.4 and 4.2 mA cm-2. The concept of 8% in well stirred solutions is a fiction, and the mass-transfer rate constant15 calculated from equation (15) can be used with some confidence in the selection of conditions; THE RESISTANCE OF ANODICALLY FORMED SILVER BROMIDE FILMS- A low film resistance is necessary to keep the over-all cell resistance within the limits of the power supply, and, in the interests of good current regulation, rapid changes in cell resistance must be avoided. At known constant currents, the cell resistance was monitored by means of the voltmeter, V, in Fig.2. Readings were taken at intervals of 20 to 200s and corrected for known resistance in the circuit. Electrolyses were performed at various current densities in solutions 0.3 M in bromide and containing 0.025 mol of sulphuric acid. Relatively small anodes of apparent area, A, 15 to 30 cm2 were used, so that large changes in total bromide concentration were avoided: the current densities of 4 , 7 , 10,20 or 40 mA cmW2 were calculated on the basis of the initial electrode area. The graphs of cell resistance versus film thickness in coulombs per square centimetre (means of several replicate runs in each instance) are shown in Fig. 5, and display a characteristic pattern. There is first a slow steady rise to approximately five times the initial resistance, followed by an accelerating increase to more than fifty times the initial resistance, and this rapid increase occurred at a critical film thickness that was inversely proportional to the current density.Similar curves were obtained at current densities of 10 to 40 mA cm-2 in 0.15 M bromide solution. The sharp rise coincided with the appearance of yellow patches or stripes on the previously dull green silver bromide film. The yellow areas, less than 1 per cent. of the total area, formed at places where the current density would be expected to be least, that is, at places most remote, or screened, from the cathode. The yellow deposit was softer than the green deposit, and dissolved very rapidly in the 1.0 M cyanide cleaning solution.Further, the cell resistance tended to show large irregular fluctuations during the rapid rise, suggesting that occasional cracking o€ the film occurred. On dissolution of the film in cyanide, a black, loosely adherent deposit remained on the electrode. The deposit dissolved rapidly in 1 + 1 nitric acid solution and was chemically identified as silver. Values for the critical film thick- ness were estimated by extrapolation of the two straight portions of the curves shown in Fig. 5 to a point of intersection. These values are plotted against the reciprocal of the current density in Fig. 6 and show a good fit. Estimates were made of the dimensional thickness of the films by direct measurement ILim. = - Kmass Br nFA [Br-IB. and by calculation. Direct measurements were made by micrometer gauge of the diameter Film thickness/[: cm-* Fig.5. Cell resistance v e w m film thickness graphs for the anodic formation of silver bromide at current den- sities of 40 (A), 20 (B), 10 (C), 7 (D) and 4 (E) mA cm-2318 BISHOP AND RILEY : PRECISE COULOMETRIC DETERMINATION [APUdySt, VOl. 98 I I I I 0.10 0.20 (Current density)-' /mA'-' cm2 Fig. 6. Variation of critical film thickness with current density of film formation of the coated rod and of the diameter after removal of the film; portions of the film removed from the anode were also measured; and such films after drying out were sandwiched between two thin glass plates, and the thickness of the glass plates was measured with and without the portion of film between them. Calculations were made on the assumption that no signifi- cant dissolution of the film occurred during its formation and that the density of silver bromide was 6.47 g cm-3.Results for films formed at a current density of 7 mA cm--2 are shown in Table I. The measured values were always 20 to 30 per cent. greater than the TABLE I THICKNESS OF SILVER BROMIDE FILMS Thickness/ p m 1 15.6 47 61 16.4 49 64 ThicknesslC cm-2 Calculated Measured calculated values, so the density of the deposit is less than the bulk density, thus supporting the contention that the deposited films are porous. B b 1 I 1 I I 40 80 120 F i I m t h ic kness/p m Fig. 7. Specific resistance of anodically formed silver bromide films at different current densities of formation. A, 20mAcm-2; and B, 4mA cm-2May, 19731 O F ACIDS I N CELLS WITHOUT LIQUID JUNCTION.PART I1 319 The calculated value for film thickness was used in conjunction with the appropriate value for cell resistance, corrected by subtraction of the cell resistance at zero time before any deposit had formed, in order to obtain estimates of the specific resistance of the films. The values were found to be in the region of lo4 to lo5 SZ cm and to vary as the film thickness increased. The variation for two films formed at different current densities is shown in Fig. 7. Attempts were made to measure the specific resistance of pre-formed films in potas- sium sulphate solutions of various concentrations. The output voltage of the source was measured with an anode current density of 5 mA cm-2, passed for the shortest possible time, before and after deposition of the film.Results are shown in Table I1 for the specific resistance of films pre-formed in 0.3 M bromide - 0.08 M sulphuric acid solution, measured in 0.2 and 0.02 M potassium sulphate solutions. The electrodes were immersed in turn in the two solutions, starting with the more dilute solution, and the process was repeated with fresh potassium sulphate solutions. The figures for the 30-pm film show that the specific resistance depends on the electrolyte concentration, again indicating that the films are porous. The 47-pm film was above the critical film thickness and caused the voltage limit of the power supply to be exceeded, so that it switched over to the constant-voltage mode; the current was rather unsteady, so that exact values could not be obtained.TABLE I1 SPECIFIC RESISTANCE OF PRE-FORMED SILVER BROMIDE FILMS, MEASURED I N POTASSIUM SULPHATE SOLUTIONS Specific resistancela cm Film thicknesslpm in 0:2 M K,SO, in 0.02 M SO, 30* (q 2.60 x 104 3-41 x 104 (2%) 2-73 x 104 3-28 x 104 47 > 4.5 x 106 > 4.5 x 106 * Estimated uncertainty of results = k0.13 x lo4 M cm. To check performance during a full acidimetric assay, the cell resistance was measured during the passage of 6000 C at 2 A by using a silver anode of area 255 cm2 in a 0.25 M potas- sium bromide solution containing 0.025 mol of sulphuric acid. A typical graph is shown in Fig. 8. The resistance increases steadily until a rapid rise occurs near the end-point, when the cathodic reaction switches from reduction of hydrogen ions to reduction of water molecules [equations (1) and ( 2 ) , Part 11] with consequent increase in the back e.m.f.of the cell. There- after, the cell resistance increases steadily at an enhanced rate, but is always well within the voltage output of the supply unit. In a similar full-scale experiment with a silver anode 10-3 OK Fig. 8. Cell resistance during the anodic formation of silver bromide a t 7-9 mA cm-a in a solution initially 0-25 M in potassium bromide and 0.08 M in sulphuric acid320 BISHOP AND RILEY PRECISE COULOMETRIC DETERMINATION [AIzdySf, VOl. 98 of area 263 cm2 in 0.1 M potassium bromide solution, the cell potential and the potential of the silver anode were measured. As shown in Fig. 9, both relationships are similar: the initial steady rise is followed by a more rapid rise as the limiting current density for the residual potassium bromide concentration is exceeded. The subsequent irregular behaviour indicates fracture of the anodic film. The magnitude and behaviour of the anode potential prior to breakdown of the film indicate an increasing overpotential due to film resistance.The measured anode potential was independent of the distance between the anode and the termination of the salt-bridge connection to the reference electrode, thus confirming that the IR drop between the working electrodes is caused by the resistance of the silver bromide film. Fig. 9. Silver anode poten- tial uersus S.H.E. (A) and power supply unit output voltage (B) during electrolysis a t an anodic current density of 7.6 mA cm-8 of a solution initially 0.1 M in potassium bromide and 0.08 M in sulphuric acid THE MAGNITUDE OF THE “SILVER ERROR”- The silver error was measured by chronopotentiometric anodic stripping, which will be described in Part 111, in order to ascertain its magnitude and the dependence on bromide concentration.Large anodes, of area 260 to 290 cm2, were used, and the acid determinand was about 0.025 mol of sulphuric acid. Most electrolyses were continued to the end-point, but some were of shorter duration in an attempt to discover whether the initial rate of dissolution of the halide film was greater than the average rate on account of the initially high but continuously decreasing bromide concentration. After the electrolysis, a visible film of silver black was present on the platinum cathode.The cathode was removed, carefully drained, washed by immersion in several successive portions of water so as to prevent loss of silver, and then transferred to a fresh cell for the determination of the silver error. This determination was carried out in an electrolyte consisting of 0.1 M potassium nitrate - 0.001 M nitric acid solution, at a current of 3 to 5 mA to a potential of +O.S V. The results in Table 111 show that the silver error is significant, being as high as 0-08 per cent., and is strongly dependent on the initial bromide concentration. The minimal error was attained with a low initial bromide concentration with subsequent addition of increments of potassium bromide at regular intervals so as to maintain the concentration at a level necessary to support the mass transport by bromide ions at the current of 2 A.The results also confirm that the rate of dissolution of the silver bromide film is higher than the average rate during the early stages of the electrolysis.May, 19731 OF ACIDS I N CELLS WITHOUT LIQUID JUNCTION. PART I1 TABLE I11 THE MEASURED “SILVER ERROR” 321 Quantity of electricity Initial [KBr]/mol 1-1 passed/C 0.33 4840 4750 0.25 4500 4500 1500 1010 Silver error/ C 3.63 3.28 1-48 1.48 0.57 0.47 Silver error, p.p.m. 750 690 330 330 380 470 0.22 4500 1.00 220 0-10* 4500 0.43 95 * KBr (10 mmol) added to cell electrolyte after each 500 s (1000 C ) of electrolysis. The results indicated that unacceptably large errors would arise in the assay of acid, even if the minimum initial bromide concentrations were used with periodic replenishment (0.0095 per cent.).Attention was turned to the cathodic reaction, and means of maintaining the cathode at a more positive potential during the determination, thus decreasing the silver deposition current. Saturation of the solution with oxygen so that reaction (4) (Part 11) would dominate without loss of cathodic current efficiency and the use of an “emerging” cathode were examined, The latter device was used by Hersch,lS who found that the electro- dissolution current of oxygen at silver cathodes was enhanced as much as twenty-fold by partially withdrawing an initially submerged electrode from the solutioq so as to expose a narrow band to the air. Chart travel - Fig.10. The potential of the platinum cathode in the vicinity of the end-point in the determination of 0.025 mol of sulphuric acid in 0-25 M potassium bromide solution. A, immersed electrode, solution exposed to air ; B, immersed electrode, solution saturated with oxygen; and C, “emerging” cathode, solution saturated with oxygen A series of three comparative experiments was performed in which the potential of the platinum-gauze cathode was recorded throughout an electrolysis that involved the passage of about 5500 C through a solution that was initially 0.25 M in potassium bromide and con- tained 0.025 mol of sulphuric acid. The area of the silver anode was 250 cm2, and the ceramic plug termination of the salt-bridge connection to the reference cell was replaced with a Luggin capillary placed very close to the cathode.The first experiment was performed with the solution exposed to the atmosphere (A in Fig. lo), the second with oxygen bubbling rapidly through a fine porosity sintered-glass disperser (B in Fig. 10) and the third with the top 3 to 5 mm of the cathode above the surface of the solution through which oxygen was322 BISHOP AND RILEY : PRECISE COULOMETRIC DETERMINATION [Analyst, VOl. 98 bubbled as in the second experiment (C in Fig. 10). The curve obtained with the “emerging” cathode was not significantly different from that obtained when purging with oxygen. In the first aerated solution, the cathode potential, initially about -0.4 V, decreased steadily to -0.7 V, then rapidly (as reduction of water molecules replaced reduction of hydrogen ions) to -1-5 V in the vicinity of the end-point of the neutralisation reaction. With the oxygen purge the curve was very similar, but the potential was about 0.1 V more positive throughout and the sharp drop occurred slightly later in the electrolysis.Purging with oxygen therefore had the desired effect qualitatively, but was disappointing quantitatively. Current - potential curves were prepared in oxygen-saturated and oxygen-free 0-05 M per- chloric acid - 0.1 M potassium bromide solutions. A Luggin capillary and a silver anode of area 240 cm2 were used, and the curves obtained (shown in Fig. 11) indicate identity at currents above 0.8 A, while at lower currents the presence of oxygen gives potentials that are rather more positive, but the difference is not more than 70 mV even at a current of 50 mA.DISCUSSION The distinct green colour of the anodically formed silver bromide films, together with the loose residue of silver black after treatment with cyanide, attest the migration of bromide into the silver metal, and that of silver atoms through the film to the liquid interface, a common occurrence. Indira and Doss20 later found silver to chlorine ratios to be as high as 1-1: 1 in chloride films formed at 10 to 20 mA cm-2, but thought that the excess of silver might be present in the form of interstitial silver ions, which involves the presence of an equal number of trapped electrons (presumably solvated) , and consequently that absorption in the visible region produced a colour different from that of the normal halide.However, such a colour would be brown, not green, and such a large accumulation of separated charge in a conducting and porous film is very unlikely. I - 0 5 -0.3 -0.1 Potential versus S.H.E./V Fig. 11. Current - potential curves for a platinum cathode in 0.05 M perchloric acid - 0.1 M potassium bromide, de-oxygenated (A) and saturated with oxygen (B) Porosity readily explains the characteristic shape of the cell resistance versus film thickness curves. Growth of the film reduces the effective surface area of exposed silver, and the more restricted mobility of electrolyte solution in the pores will lead to exhaustion of bromide ions and stripping of silver ions, which migrate to the film- liquid interface, depositing normal silver bromide in the pores.Hence the yellow colour appears in parts of the film that retain high porosity. Final blockage of the pores leads to an essentially non-porous film and a rapid rise in resistance. The values found for the specific resistance in potassium sulphate solutions further confirm that silver bromide films are porous until the critical thickness is exceeded. Although this finding conflicts with the results obtained by Jaenicke, Tischer and Gerischer,21 who found the specific resistance of silver chloride films to be inde- pendent of electrolyte concentration for thicknesses even as small as 3 pm, Briggs andMay, 19731 OF ACIDS I N CELLS WITHOUT LIQUID JUNCTION. PART I1 323 Thirsk22 considered such films to be porous, and calculated that if the total capillary area were only 0.1 per cent.of the electrode area, then 99 per cent. of the current would be carried by the liquid in the capillary pores in a 0.1 M potassium chloride solution. K ~ r t z ~ ~ first observed that anodically formed silver halide layers had specific resistances ten to one hundred times less than that of the solid crystallised from the melt; the value for solid silver bromide is about lo7 Q C I T L . ~ ~ The values found in the present work, 5 x lo4 to lo5 Q cm, agree with those found by Lal, Thirsk and Wynne- Jones,24 and Jaenicke et aZ.21 obtained curves similar to those in Fig. 7 for silver chloride films. In all three instances, the calculated values for specific resistance increased with decreasing current density of film formation.The higher conducti- vity of anodic silver halide films has been explained by Kurtz23 in terms of porosity, while Briggs and Thirsk22 thought that the silver chloride film was a crystalline material of specific resistance equal to that of the solidified melt, in parallel with a capillary network of pores. Jaenicke et aL21 concluded that silver chloride films were non-porous, while Indira and Doss20 ascribed the higher conductivity to a high concentration of interstitial ions. The inverse proportionality between critical film thickness and current density of formation (Fig. 6) parallels the work of La1 et u Z . , ~ ~ who observed a change of slope of about two-fold in the overpotential versus time for anodic formation of silver chloride. The calcu- lated film thickness at the point of change was approximately inversely proportional to current density of formation and was independent of chloride concentration in the electrolyte, as is demonstrated here for bromide films.It is clear that the structure of anodically formed silver halide films is strongly dependent on the current density of formation, and so also, therefore, is the coulombic capacity of the reaction. The small effect of oxygen on the cathodic reaction indicates that the charge-transfer overpotential for the reduction of oxygen on platinum, either clean or covered with a film of silver, is very large under the conditions used. CONCLUSIONS The results presented provide a basis for the selection of suitable values for the current density and bromide concentration for the auxiliary electrode reaction at given values of generating current and total quantity of electricity required.An anode area of 300 cm2 is chosen; at a current of 2 A the current density will be about 7 mA cm-2, the capacity of the anode for silver bromide formation will be about 9000 C and, even after ten determinations of 0.05 mol of a monobasic acid, the capacity should still be in excess of 6000 C. A residual bromide concentration at the end of a determination has been chosen to be 0.03 to 0.05 M, for which the limiting current density is 20 to 33 mA cm-2; when added to the amount of bromide consumed in the reaction, the initial bromide concentration should be 0.20 to 0.22 M. The “silver errors” are significant and attempts to minimise them were only partially effective.A means of correction for the silver error is necessary for work of the highest accuracy, and will form the subject of Part 111, but it is possible to make the error tolerably small for less critical work. One of us (M.R.) expresses deep gratitude to the Charitable and Educational Trust of the Worshipful Company of Scientific Instrument Makers for financial support in the form of a Research Studentship. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. REFERENCES Bishop, E., and Riley, M., Analyst, 1973, 98, 306. Szebelledy, L., and Somogyi, Z., Z. analyt. Chein., 1938, 112, 323. Bishop, E., and Dhaneshwar, R. G., Analyt. Chem., 1964, 36, 726. Taylor, J. K., and Smith, S. W., J . Res. Naha. Bur. Stand., 1959, 63A, 153. Eckfeldt, E. L., and Shaffer, E. W., Analyt. Chem., 1965, 37, 1534. Lingane, J . J., and Small, L. A., Ibid., 1949, 21, 1119. Carson, W. N., and KO, R., Ibid., 1951, 23, 1019. Lingane, J. J., Analytica Chim. Acta, 1954, 11, 283. Bishop, E., and Short, G. D., Analyst, 1964, 89, 587. Linke, W. F., “Solubilities of Inorganic and Metal Organic Compounds,” Fourth Edition, Volume 1, Laitinen, H. A., “Chemical Analysis,” McGraw-Hill, New York, 1960, pp. 114-116. Forbes, G. S., and Cole, J. I., J . Amer. Chem. SOC., 1921, 43, 2492. James, T. H., and Vanselow, W., J . Phys. Chem., 1958, 62, 1189. I , Ibid., 1938, 112, 332. -__ Van Nostrand, Princeton, 1958.324 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. BISHOP AND RILEY Bishop, E., Chemia Analit., 1972, 17, 511. Kolthoff, I. M., and Lingane, J. J., “Polarography.” Volume I, Second Edition, Interscience Publishers Inc., New York, 1952, p. 52. Bishop, E., Dhaneshwar, R. G., and Short, G. D., in West, P. W.. Macdonald, A. M. G., and West, T. S.. Editors, “Analytical Chemistry 1962,” Elsevier, Amsterdam, 1963, p. 241. Badoz-Lambling, J., Bull. Soc. Chim. Fr., 1959, 792. Hersch, P., in Reilley, C. N., Editor, “Advances in Analytical Chemistry and Instrumentation,” Volume 111, Interscience Publishers Inc., New York, 1964, p. 198. Indira, K. S., and Doss, K. S. G., J . Electroanalyt. Chem., 1968, 17, 145. Jaenicke, W., Tischer, R. P., and Gerischer, H., Z . Elektrochem., 1955, 59, 448. Briggs, G. W. D., and Thirsk, H. R., Trans. Faraday SOL, 1952, 48, 1171. Kurtz, L. J., Dokl. Akad. Nauk SSSR For. Lang. Edn, 1935, 2, 305. Lal, H., Thirsk, H. R., and Wynne-Jones, W. F. K.. Trans. Faraday SOC., 1951, 47, 70. NOTE-Reference 1 is to Part I of this series. Received December 121h, 1972 Accepted January lst, 1973

ISSN:0003-2654    

DOI:10.1039/AN9739800313

出版商:RSC    

年代:1973

数据来源: RSC

摘要:

Analyst, May, 1973, Vol. 98, $9. 325-328 325 Titration of Sulphate in Mineral Waters and Sea Water by Using the Solid-state Lead Electrode BY MARC0 MASCINI (Istituto di Chimica Analitica, Universith di IZoma, Rome 00185, Italy) A procedure is described for the determination of sulphate in the range 20 to 3000 p.p.m. in mineral and sea waters by using a lead-selective electrode. Chloride and hydrogen carbonate are separated from the sample by passing it firstly through a cation-exchange resin in the silver form, and secondly through a cation-exchange resin in the acid form. The solution recovered is titrated with standard lead nitrate solution. THE recent availability of a lead-selective electrode has facilitated the potentiometric deter- mination of sulphate by direct titration with a standard lead solution.This method was described by Ross and Frantl and the titration was adapted by Selig to the determination of sulphur in organic compounds after oxygen-flask combustion.2 Several anions, at a high concentration, are reported to interfere in this determination; among these the anions C1- and HCO, are most commonly present at a high level in the majority of mineral waters and sea water and limit the usefulness of the method in routine applications. Sulphate was separated from other materials, both organic and inorganic, by a cation- exchange resin supported on aluminium oxide.3 Recently, details of the separation of nitrate from chloride and hydrogen carbonate by using two cation-exchange resins, with which the sample was mixed consecutively, have been published.* The first resin, in the silver form, removes the chlorides and the second, in the acid form, removes the silver ions left in the sample by the first.At the exit from the second column the sample is acidic and hydrogen carbonate is eliminated. In the work described in this paper, the feasibility of separating sulphate from chloride and hydrogen carbonate by means of ion-exchange resins was established, making possible the use of the lead-selective electrode for the titration of sulphate in most mineral and saline waters. APPARATUS AND REAGENTS- was used. was used. 1 M sodium nitrate solution. pH meter-An instrument capable of giving a scale expansion of 1 mV per division Lead ion selective electrode-A Sens Ion, Model 201 Pb, made by AMEL, Milan, Italy, Reference electrode-This had a double junction, the outer chamber being filled with Titration cell-A polyethylene beaker was used for this purpose. Magnetic stirrer.Ion-exchange resin-Merck ion exchanger I (a strongly acidic cation exchanger) was used, Analytical-reagent grade reagents were used throughout. PROCEDURES PREPARATION OF THE CATION-EXCHANGE RESIN IN THE SILVER FORM- Wash about 50 g of the analytical-grade resin with about 200 ml of 2 M nitric acid and then remove the acid by filtration on a sintered-glass funnel, washing the solid with distilled water until neutral. Next, transfer the resin to a 500-ml flask, add 200 ml of 0.25 M silver nitrate solution, stirring magnetically for about 30 minutes, and then filter the resin from the liquid on a sintered-glass funnel and wash it several times with distilled water to displace the excess of Ag+ ions.Store the resin in a dark-coloured vessel away from direct light. A 50-g portion of resin is sufficient for at least fifty analyses. @ SAC and the author.326 [Autalyst, Vol. 98 PREPARATION OF THE CATION-EXCHANGE RESIN IN THE ACID FORM- Wash 50 g of the analytical-reagent grade resin with 200 ml of 2 M nitric acid and then remove the acid by filtration on a sintered-glass funnel, washing the solid with distilled water until neutral. Use the resins in bead or column form; in the latter case prepare a column about 15 cm in height and 0.5 cm in diameter. ANALYTICAL PROCEDURE- Pass 50 to 100 ml of the sample, containing 1 to 100 mg of chloride, at 1 to 3 ml min-1 through two columns in series, the first containing the resin in the silver form, and the second in the acid form.Discard the first 10 ml. The solution is now free from chloride, hydrogen carbonate and silver ions and is slightly acidic, and should be adjusted to a pH of 5 to 6 by adding a few drops of 1 0 - 2 ~ sodium hydroxide solution and internal indicator. The pH should now be at the optimum for the lead-selective electrode.6 Alternatively, place 50 to 100 ml of the sample in a beaker with 0.5 to 1.0 g of resin in the silver form and stir magnetically for about 20 to 30 minutes. Filter the mixture through a sintered-glass funnel and, to the clear liquid, add 0.5 to 1-Og of resin in the acid form, stirring magnetically for about 20 to 30 minutes.Then, filter the solution from the resin and adjust the pH to 5 to 6. Dilute an aliquot (e.g., 10 ml, measured with a pipette) with an equal volume of 1,4-dioxan. Insert the lead-selective and reference electrodes into the magnetically stirred solution and titrate it with lead nitrate solution ; the concentration of the lead nitrate solution should be about ten times that of the sulphate. Record and plot the potential values on a graph versus the volume of titrant added to reach the end-point. The procedure does not require that the sample should be completely recovered, but that the volume taken for the final titration should be accurately known. With an unknown sample a rough titration could be carried out first on a 10-ml aliquot in order to determine the concentration of titrant required and the approximate end-point.A second 10-ml aliquot would then be used for a more accurate titration in which the titrant was added in smaller increments in the vicinity of the end-point in order to increase the precision. RESULTS AND DISCUSSION The determination of sulphate in mineral and waste waters is frequently necessary. Fig. 1 shows a typical graph of the titration of a sample containing chloride and sulphate, MASCINI: TITRATION OF SULPHATE IN MINERAL WATERS AND > E --. W Pb (NO,),/ml Fig. 1. Titration of 50 p.p.m. of sulphate ion in the presence of 100 p.p.m. of chloride ion in 1,4- dioxan (50 per cent.) with 5 x lo-3rd lead nitrate solution. Sample volume 10 ml: A, without treat- ment ; and B, with the ion-exchange treatment described in the textMay, 19731 SEA WATER BY USING THE SOLID-STATE LEAD ELECTRODE 327 at levels normally found in mineral waters, with lead nitrate solution before and after passing the sample solution through the ion-exchange columns.The advantage of using the ion- exchange process is evident. TABLE I TITRATION OF STANDARD SOLUTIONS OF SULPHATE IN THE PRESENCE OF CHLORIDE IONS Sulphate taken, p.p.m. Chloride taken, p.p.m. Sulphate found, p.p.m. 100 35 100 100 98 98 100 1000 350 980 970 1000 1020 1000 100 3500 98 100 98 98 98 1040 1040 1000 1020 1020 1000 3500 Error, p.p.m. Nil Nil 2 2 Nil 20 30 Nil 20 Nil 2 Nil 2 2 2 40 40 Nil 20 20 Phosphates are occasionally present in mineral waters and interfere in the determination of sulphate.By passing the sample through the resin in the silver form the amount of phosphate is reduced, but not eliminated; in this instance, more complicated procedures for separation must be considered.6 Tables I and I1 show the results of several titrations carried out on standard solutions of sulphate and on sea water and mineral waters. The extent of the error indicates that the TABLE I1 SULPHATE TITRATIONS IN DIFFERENT WATER SAMPLES Sulphate added Size of (as sample/ Na,SO,), Water sample. ml p.p.m. Laboratory water 10 Nil Nil Nil Nil 10 10 20 20 30 50 100 10 Nil Nil Nil Nil 10 10 20 20 30 60 100 Mineral water sample 1 Sulphate recovered (as Na,SO,), p .p . m. 35 35 40 40 45 45 55 55 60 85 140 75 70 75 77 85 85 95 90 105 125 170 Water sample Mineral water sample 2 Sea water sample 1 Sea water sample 2 Sulphate added Size of (as sample/ Na,SO,), ml p.p.m.10 Nil Nil Nil Nil 10 10 20 20 30 50 100 1 Nil Nil Nil Nil 1 Nil Nil Nil Nil Sulphate recovered (as Na,SO,), p. p. In. 50 55 50 50 60 60 75 70 85 100 150 2350 2300 2300 2300 2500 2400 2500 2450328 MASCINI procedure is acceptable for routine analysis. The limit of sensitivity is about 10 p.p.m., and the amount of chloride can be up to ten times the amount of sulphate. The presence of nitrate in the sample does not affect the results when using this procedure, at least up to the level normally found in the mineral waters, and lead nitrate solution is recommended for the titration, as it is more readily available than lead(I1) perchlorate, recommended by Ross and Frant.l The time taken to carry out the titration is about 10 minutes.In the analysis of samples of low concentration (20 to 50 p.p.m.) the electrode response is sluggish near the equivalence point, as in other potentiometric titrations. In this instance it is advisable to add the titrant in small increments and to record the value after a delay (2 minutes). In this way the titration will be more reproducible. Another valuable technique is the automatic recording of the titration volume. The resins can be regenerated several times; the first, in the silver form, can be treated with 3 M ammonia solution and washed with distilled water and then with nitric acid, as described under Preparation of the cation-exchange resin in the silver form. If the resin turns brown in colour, the ion-exchange activity becomes very low and it is necessary to change the resin. This effect is probably caused by oxidation of the resin by Ag+ ions. The resin in the acid form can be regenerated by treating it with 1 M nitric acid. REFERENCES 1. 2. 3. 4. 5. 6. Ross, J. W., jun., and Frant, M. S., Analyt. Chem., 1969, 41, 967. Selig, W., Mikrochim. Acta, 1970, 168. Kirsten, W. J., Hansson, K., and Nilsson, S. K., Analytica Chim. A d a , 1963, 28, 101. Paul, J. L., and Carlson, R. M., J . Agric. Fd Chem., 1968, 16, 766. Mascini, M., Analytica Chim. Acta, in the press. Colson, A. F., Analyst, 1963, 88, 26. Received January 27th, 1972 Amended June 30th, 1972 Accepted January 8th, 1973

ISSN:0003-2654    

DOI:10.1039/AN9739800325

出版商:RSC    

年代:1973

数据来源: RSC

摘要:

Analyst, May, 1973, Vol. 98, pp. 329-334 329 Analysis of High-purity Water by Flameless Atomic-absorption Spectroscopy Part II.* Signal Integration with a Non-resonance Line Correction System for Spurious Absorption Phenomena BY C. J. PICKFORD? AND G. ROSSIS (Chemistry Division, Euratom-CCR. 21020-Ispra ( Varese), Italy) A polychromator has been used in conjunction with a multi-channel integration system and an automatic sample injection unit in graphite-tube flameless atomic-absorption spectroscopy. The precision of the system has been evaluated a t high and low absorbances, and its ability to compensate for spurious absorption and variable volatility effects examined. IN Part I,l we described an automatic sampling unit used in conjunction with flameless atomic-absorption spectroscopy as part of a project to provide simultaneous determinations of a number of elements present at parts per lo9 (p.p.b.) levels in water.This paper describes the spectrophotometer and the detection system used in conjunction with this sampling unit for the above purpose. The system has been tested by using one main channel and one reference channel sequentially ; the system used for multi-element determinations will contain up to five channels and one or more reference channels. SPECTROPHOTOMETER SYSTEM- In the past, several systems have been described for multi-element determinations by atomic-emission,2 fluorescence3 or absorption4 spectroscopy with the use of polychromators or interference filter systems to separate the light beam into its individual components.For atomic-absorption spectroscopy, band-pass requirements usually preclude the use of interference filters, so that a polychromator is the obvious choice. SPURIOUS ABSORPTION PHENOMENA- Interferences in flameless atomic-absorption spectroscopy from spurious absorption effects, whether caused by condensed vapours of inorganic salts or by carbon dust, have often been For this reason, in a fully automatic system with low absorbance signals from samples that might be expected to exhibit such phenomena, some way of correct- ing for this effect is obviously desirable. There are several methods that can possibly be used for this purpose. A nearby non-resonance line of the same or another element can be used to measure the absorption due to non-specific effects alone, and this value can then be subtracted from the signal obtained by measurement at the element resonance line.The correction can be effected simultaneously, by using a dual-channel system,* as is achieved with a number of modern atomic-absorption spectrophotometers, or can be applied sequen- ti all^.^ The latter method is, of course, much easier to apply instrumentally, and can be carried out with even the simplest atomic-absorption instrument but, clearly, for the purpose of a study such as that being considered, the choice of a simultaneous correction system is imperative. Perhaps the most elegant compensation system is the deuterium-arc corrector,lO which involves the non-specific absorption of light from a deuterium arc at the same wavelength as the resonance line in order to compensate for background absorption.A similar system has been describedll in which a hydrogen-arc lamp, and polarisers to separate the two light beams, are used. In choosing the method of correction that is most suitable for multi-element determinations, the advantages and disadvantages of the two-line and the deuterium-arc methods should be considered. The principal disadvantages of the two-line method are that Gebouw voor Analytische Scheikunde, Technische Hogeschool Delft, Jaff alaan 9, * For details of Part I of this series, see reference list, p. 334. t Present address : $ To whom requests for reprints should be addressed. @ SAC and the authors. Delft, The Netherlands.330 PICKFORD AND ROSSI: ANALYSIS OF HIGH-PURITY WATER BY [Analyst, Vol.98 selection of the non-resonance line (which must be very close to the resonance line) may be difficult, except when a source rich in intense emission lines is used, and that isolation of the two lines with exit slits may be rather tedious unless lines from overlapping orders, as with grating instruments, can be selected. The principal disadvantages of the deuterium-arc system are that the continuum emitted may have a low intensity at certain wavelengths, and is essentially zero above 360 nm. Also, it is necessary, for instrumental reasons, for intensities of the two signal sources to be exactly balanced before measurements can be made. When multi-element determinations are to be carried out, this latter requirement indicates the use of the two-line method, because other- wise the intensity of each resonance line must be attenuated to that of the deuterium arc at that wavelength, which attenuation would be difficult to achieve and wasteful of energy (thus giving a greatly increased noise level).TREATMENT OF THE ABSORPTION SIGNAL- Although the amplitude of an absorption signal is usually used for the purpose of deriving calibration results, the integral of the signal is more often recorded in multi-channel instru- ments, mainly because of the greater ease of signal handling and subsequent display. It has also been ~ l a i m e d ~ ~ ~ ~ ~ ~ 3 that integration may, in addition, give greater precision when used in conjunction with flame or flameless atomic-absorption spectroscopy. For the above reasons, signal integration was preferred to amplitude measurement involving peak-sensing circuitry with memory capabilities. EXPERIMENTAL POLYCHROMATOR AND SLIT ASSEMBLY- A modified Jarrell-Ash, Model 75-000, fl6.3, plane grating 0-75-m spectrograph was used, with the photographic plate assembly removed and a light-tight box containing a series of exit slits and associated photomultipliers firmly attached.Commercially available aluminised quartz plates with 100-pm slits (Jarrell-Ash Co.) were carefully positioned at the focal plane of the spectrograph, a refractory quartz plate being used in front of each slit for the correct alignment of the slits with the spectral lines. The slit widths used corresponded with the opening of the entrance slit of the spectrophotometer.The minimum degree of separation of two consecutive slits that could be achieved was about 5 mm, which corresponded to a wavelength difference of 5 nm in the first order. When a smaller degree of wavelength separation was required two plane mirrors were used in order to bend the light path of the reference line sufficiently to bring it into line with the photo- multiplier. With this procedure, lines as close as 0.6 nm could be used. Sufficient space was available to enable up to ten lateral window photomultipliers to be installed. E.M.I. 9783 B in addition to RCA 1P 28 photomultipliers fed by a Keithley 245 high-voltage supply were used in this study. The above system (with individual slits) will later be replaced by a single aluminised quartz plate marked with various slits at the selected positions that correspond to the spectral lines to be used.Preliminary experiments with a reference beam system were performed by using a quartz plate positioned immediately after the graphite oven and tilted in such a way as partly to reflect the light from the multi-element hollow-cathode lamp, The reflected light was filtered with an interference filter so that a narrow wavelength band in the proximity of the resonance line to be studied could be passed and monitored by a photomultiplier. However, the dif- ference in band pass between the interference filter and the spectrophotometer led to results that were not very satisfactory. INTEGRATION ASSEMBLY- This assembly is shown in a simplified form in Fig. 1 for one channel (A) with a reference channel (B).This scheme is also applicable to other channels with which the same or different reference channels are used. The sequence of operations is as follows: the incoming photomultiplier signals are first converted into logarithmic form and then subtracted. The auto-balance circuit furnishes a d.c. level to the amplifier, which performs the subtraction step, so that during intervalsMay, 19731 FLAMELESS ATOMIC-ABSORPTION SPECTROSCOPY. PART I1 331 Fig.. 1. Elock diagram of dual-beam integrator circuit. k l energised; Re-set: k l and k3 energised; and Integrate: k l and k2 energised Balance : no relays energised ; Hold : between analyses (when there is no absorption) the subtractor output will be adjusted auto- matically to zero. This procedure is essentially equivalent to adjusting the intensities of the two light beams so that they become equal, but is achieved automatically.The output of the subtractor (now in absorbance units) is then integrated. The Balance, Hold, Re-set and Integrate cycles are controlled by three magnetic reed relays. The sequence can be achieved automatically, by being operated from the previously described central programming unit, or manually. The circuit can also be operated without the integration circuit, and also with direct amplification instead of log conversion. The log conversion, subtractor and integrator circuits are conventional, and have been described elsewhere.14 The auto-balance circuit is based on an auto-balance circuit described for use with gas-chromatographic integrators.15 The operational amplifiers used were the Burr- Brown 1556115 amplifiers.The output from the integrator is connected across a ten-turn helical potentiometer (Beckmann) so that for the preparation of linear calibration graphs, the integrator output can be scaled down directly to give a value in concentration units, which is displayed either by the digital voltmeter or the recorder. In normal operation the relays were controlled so that the integrator was in the Balance position continuously except when on Hold, Re-set or Integrate. Re-set was operated for 5 s before the atomisation cycle, Integrate was operated during the atomisation cycle and Hold was used as an intermediate position whenever a change was made, and for 30 s after Integrate so as to allow sufficient time for the integral to be recorded.A direct-current system was used in order to reduce the cost and complexity of the system in the event that the maximum number of channels is required. The principal disadvantage of this system in atomic-absorption spectroscopy, i.e., susceptibility to source emission and noise, is not a severe problem when the graphite tube is used, and in fact, when the light beam was correctly centred and focused, there was almost no interference from light emission originating within the crucible. The signal output was displayed either on a recorder (as previously described) or with a digital voltmeter (Keithley, Model 160). In fully automatic operation, a printer will be connected across the digital voltmeter for output display.OPERATING CONDITIONS- The operating conditions for the hollow-cathode lamps and the methods used for the preparation of solutions have been described previous1y.l Calibration graphs were obtained by using manual sample injection with a 100-pl Eppendorf pipette and manual operation of the integrator unit. Repetitive sampling was carried out by using the automatic sampling unit, with the integrator output displayed on a recorder. For measurements made for comparison purposes with a single-channel system, the Beckmann 1301/DBG spectrophoto- meter described previously was used.332 PICKFORD AND ROSSI : ANALYSIS OF HIGH-PURITY WATER BY [Analyst, Vol. 98 Selection of reference lines was made easier by initially preparing a photographic plate (Kodak Spectrum Analysis No.1, developed under standard conditions) of the hollow-cathode lamp concerned, by using the normal plate attachment for the spectrograph. After comparison of the plate with reference tables, suitable reference lines were selected according to the criteria given by billing^.^ RE s ULTS CALIBRATION GRAPHS- Calibration graphs were prepared for the elements cobalt, chromium, copper, iron, manganese, nickel and lead by using the reference line that had been established previously for each element. The chosen line pairs are listed in Table I. As an example, calibration graphs for lead are shown in Fig. 2; the intensity axis is expressed in arbitrary units, depending on whether the graph for 0 to 20 or 0 to 100 p.p.b. is considered. As can be seen, the region of linearity (ie., for which the integrator output can be calibrated directly in concentration units) extends up to about 60 p.p.b.TABLE I LINE PAIRS AND REPRODUCIBILITY RESULTS Element line/nm CO 240.7 Cr 357.9 CU 324.7 Fe 248.3 Mn 279-5 Ni 232.0 Pb 283.3 Reference line/nm CO 241.4 Ne 352.0 CU 323.1 CU 249.2 CU 282.4 Ni 231.4 Pb 280.1 Relative standard deviation with dual-channel integration system & Per cent. p.p.b. 0.5 20 1.5 2 0.9 50 1.8 5 1.0 20 2.3 2 0.8 10 1.5 2 1-1 10 1.6 2 1.3 50 2.2 5 1.8 50 2.3 6 Relative standard deviation with single-channel a.c. system +--7 Per cent. p.p.b. 0.6 50 1.0 100 1.2 50 1.1 30 1.7 10 2.4 50 Other reported values5 & Per cent. p.p.b. 4.9 2 4.7 2 4.7 10 3.5 2 2.6 20 3.9 5 EFFECTIVENESS OF THE COMPENSATION SYSTEM- This was tested by observing the effect produced on the zero base-line when a substance with a non-specific absorption at the wavelength used was introduced into the light path. With a metal-wire mesh screen, the compensation was complete (up to 96 per cent.absorption) when reference lines within about 15 nm of the main resonance line were used. However, when concentrated solutions of sodium chloride were vaporised in the crucible, thus producing a white “fog” in the light path, it was necessary for the reference line to be closer to the main line, usually within 4 nm, in order to give complete compensation. This requirement is to be expected and results from the small particle size and the greater light-scattering effect of the sodium chloride; it does not constitute a serious limitation as it does not affect the choice of a suitable reference line, when a multi-element hollow-cathode lamp or two independent lamps can be used.The effect of added sodium chloride on the absorption signal due to the elements con- cerned (at a concentration of 1Opg1-l) was examined both for the dual-beam integration system and for the normal single-beam a.c. instrument. Fig. 3 shows the results obtained for lead. As can be seen, a reduction in the integrated intensity occurs initially, possibly because of a chemical interference effect, and then the signal is virtually constant up to an 80 000-fold excess (corresponding to 95 per cent. absorption in channel A) ; beyond thisMay, 19731 FLAMELESS ATOMIC-ABSORPTIOX SPECTROSCOPY.PART I1 333 limit, the signal increases rapidly. Without compensation, the effect on the absorption signa is very much greater. Similar behaviour has been observed for the other elements considered in this study. COMPENSATION FOR VOLATILITY CHANGES- Because of changes in the voltage applied to the graphite tube, or to changes in the wall thickness, the absorption signal obtained for a standard solution may not always be constant on a day-to-day basis. In addition, various matrices may cause a depression in the amplitude of the absorption signal, compared with purely aqueous solutions. The rate of volatilisation of 10 pg 1-1 of lead was therefore varied artificially (by changing the atomisation voltage) in order to determine whether this problem could be solved more readily by signal integration than by amplitude measurement.Fig. 4 shows that, up to a certain limiting voltage value, there is no significant variation of the integrated signal, although above this value there is a rapid decrease. At lower concentration levels, this decrease at higher voltages was not observed. The variation of the signal amplitude with a normal single-beam system can be seen to be critically dependent on the response time of the amplifier and recorder. 20 40 60 80 100 Concentration, p.p.b. Fig. 2. Calibration mraphs for lead by using dual-channel Tntegration system; 283.3 and 280.1-nm lines. A, 0 to 20 p.p.b. of lead; and B, 0 to 100 p.p.b. of lend 1.0 w .- 8 3 0 m 8 L u 2 u .- L- F .- + 0.5 +? L 0 QJ 8 m - 9 0, Q Ll 0 200 400 600 800 1000 Concentration of added NaCI, p.p.m.Fig. 3. Effect of added sodium chloride on the absorption signal of 10 p.p.b. of lead: 1000 p.p.m. of NaCl= 100 000-fold excess. A, signal obtained with single-beam ax. instrument; and B, signal obtained with dual-beam inte- gration system REPRODUCIBILITY- The reproducibility was tested, as previously described, for the same elements as before, and in addition for lead. For each element two concentration levels were examined and the standard deviation of the integrator signal was evaluated from repetitive evaporations (at least sixteen for each concentration) from samples injected automatically. In Table I, the line pairs used are given and the precision achieved is compared with that of results obtained with a single-channel ax.system and with other values reported in the literature. The improvement in precision compared with that of previous results is slight, but the precision is maintained down to a lower level. At the lower concentrations chosen, the precision is affected adversely by the integrator base-line errors, caused by both drift and electrical pick-up of the relay switching signals of the oven unit and the central programmer, although these effects had been minimised by careful shielding of the input leads and by connecting capacitors in parallel with the switch contacts. CONCLUSIONS The use of a non-resonance line correction system and integration of the absorption signal improves the precision of graphite-tube flameless atomic-absorption spectroscopy,334 PICKFORD AND ROSS1 particularly at low concentration levels.Some variations in signal intensity caused by changes in the volatility of the sample can be eliminated, and the compensation for spurious absorption phenomena is satisfactory up to 95 per cent. absorption, provided that the reference line is close to the resonance line of the main element. 4 5 6 7 8 9 10 Atomisation voltage Fig. 4. Effect of variation of atomisation voltag; on the absorption signal of 10 p.p.b. of lead. .4, signal obtained with dual-channel inte- gration system; B, signal obtained with single-beam a.c. instrument with 0.5 s recorder response time; and C, signal obtained as for R but with 1 . 5 s recorder response time With the automatic sampler previously described, the above system does represent a first approach to a fully automatic scheme for multi-element determinations by flameless atomic-absorption spectroscopy, The results achieved indicate the feasibility and the applicability of such a system, particularly when a limited number of elements have to be determined on a routine basis or when the limited size and the nature of the sample (toxicity and activity) could make cumbersome or impossible the determination of more elements.A greater degree of flexibility could be expected by the use of a more appropriate spectro- photometer and of a more sophisticated exit slit assembly. MTe are grateful to Mr. M. Mol for constructing the electronic system, and to Euratom for the award of a postdoctoral fellowship to one of us (C.J.P.). 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. REFERENCES Pickford, C. J., and Rossi, G., Analyst, 1972, 97, 647. Haagen-Smit, J. W., and Ramirez-Munoz, J., Analytica Chiwt. Acta, 1966, 36, 469. Mitchell, D. G., and Johansson, A., Spectyochinz. Acta, 1970, 25B, 175. Mavrodineanu, R., and Hughes, R. C., Appl. Opt., 1968, 7, 1281. Fernandez, F. J., and Manning, D. C., Atom. Absorption Newsl., 1971, 10, 65. Alder, J. F., and West, T. S., Analytica Chim. A d a , 1970, 51, 365. Amos, M. I)., Bennett, P. A., Brodie, K. G., Lung, P. W. Y., and MatovBek, J. P., Analyt. Chern., Massmann, H., “Mdthodes Physiques d’Analyse,” Volume 4, G.A.M.S., Paris, 1968, p. 193. Billings, G. K., Atom. Absorption Newsl., 1965, 4, 357. Kahn, H. L., lbid., 1968, 40, 7. Woodriff, R., Culver, B. R., and Olson, K. W., A p p l . Spectrosc., 1970, 24, 530. L’vov, B. V., “Atomic Absorption Analysis,” Hilger and Watts, London, 1970. Nishita, H., Farmer, R., and Peterson, S., Analytica Chim. Acta, 1972, 58, 1. “Applications Manual for Computing Amplifiers,” Philbrick Research Tnc., Nimrod Press, Dedhani, Riggs, W. A., Analyt. Chevn., 1971, 43, 976. NOTE-Reference 1 is to Part I of this series. 1971, 43, 211. Mass., 1966. Received November 30th, 1972 Accepted January 5th, 1973

ISSN:0003-2654    

DOI:10.1039/AN9739800329

出版商:RSC    

年代:1973

数据来源: RSC

摘要:

Analyst, May, 1973, Vol. 98, f@. 335-342 335 A Method for the Determination of Silver in Ores and Mineral Products by Atomic-absorption Spectroscopy* BY GEORGE WALTON (Department of Physical Science, Western New Mexico University, Silver City, New iVexico 88061, U..S.il.) An atomic-absorption spectroscopic method for the determination of silver in some ores and mineral products is reported. The effects of sorption of silver on the container walls, cleanliness of equipment, stability of standards and sample solutions, interference from silicate ions and background absorp- tion have been studied. The sample is treated with mixed acids (including hydrofluoric) and evaporated to dryness; the acid extract of the residue is then made strongly ammoniacal and the solution nebulised. Recoveries are, on average, about 10 per cent.higher than those obtained by using fire assay. Silver values obtained by the proposed procedure are almost quantitative for sulphide ores, but with other ores they are about 3 per cent. low, unless the final residue is re-examined. THE determination of silver by atoniic-absorption spectroscopy in water,l solution^,^^^ soil and rocks,*+ alloys,6 sulphide minerals,' mine and mill products, bullions and other materials has been reported. The studies reported have either been of low precision or have related to narrow and specific problems. The present study involves the use of atomic-absorption spectroscopy for the determination of silver in ores and mineral products, particularly in siliceous or sulphide ores in the range from 0.0 to 25 troy oz per net ton, and is based on the use of pure silver solutions as standards.For many years, the accepted and official method for the determination of silver in ores has been fire assay. The rapidity and simplicity usually associated with atomic-absorption measurements has aroused considerable interest in applying this technique to the determina- tion of silver. While its application to simple solutions is possible, substantial interferences have been noted in the determination of silver in ores. In order to obtain results that are comparable with those obtained by use of the fire-assay method, the method of simulation or standard addition had been used; in some instances, the standards used for atomic absorp- tion have been made from ore samples, the silver content of which was based on fire-assay values.The fire-assay technique has also been applied to the determination of gold by atomic-absorption ~pectroscopy.~ Results obtained by the atomic-absorption method are compared with fire-assay results obtained on the same samples. UNIFORMITY OF SAMPLE- A discussion of this subject can be found in the paper by Thompson, Nakagawa and Van Sickle.lo Because the distribution of silver (or any other element) in natural ores and mine products is rarely completely homogeneous, one of the advantages of the fire-assay method is the large sample size taken for analysis (from 7.29 to 29-17 g, or even larger), which tends to reduce the effect of sample inhomogeneity. For atomic-absorption spectro- scopy, the sample size should be chosen so as to give a read-out between 20 and 80 per cent.transmittance. With rich samples, this principle might involve taking such a small sample that inhomogeneity could become important, in which event the use of larger amounts of sample and final solution volumes would overcome the problem. The effect of sample inhomogeneity, if it occurs, would appear in the variation of results for a given sample. Currently, the fire-assay method requires that the sample pulp pass an 80-mesh Tyler screen, and we have based our method on a similar requirement. Extra- fine grinding as a means of providing a more homogeneous sample is not desirable, and may even cause plating-out of precious metals on the grinding surfaces. * Presented at a meeting of the New Mexico Academy of Science, October 13th and 14th, 1972, at Portales, New Mexico, U.S.A.@ SAC and the author.336 WALTON: A METHOD FOR THE DETERMINATION OF SILVER IN ORES AND [Analyst, Vol. 98 SOURCE OF ORE SAMPLES AND FIRE-ASSAY RESULTS- Ore samples used in this study (and two samples of blister copper) were obtained from mining operations in Texas, New Mexico and Arizona. Both oxide and sulphide ores are represented. The fire-assay values given in this paper are referee values; they therefore represent the best value taken from at least six determinations carried out by three different labora- tories and by experienced personnel. All samples were obtained as pulp, ground to pass a Tyler No. 80 sieve (U.S. Standard Sieve Series, W. S. Tyler Co., 0-0070-inch opening), and were examined without drying.DISSOLUTION OF SAMPLE- Several of the sample dissolution procedures reported in the literature were tried, but all yielded extremely low results for silver. We then examined the effect of introducing a fusion procedure before the determination, as described by Huff man, Mensik and Riley.ll Obviously, this procedure resulted in complete dissolution of the ores, but consistently low results were still obtained, which suggested that an interference effect was present. The procedure finally adopted for the dissolution of the samples involved the use of a mixture of acids; however, a re-dissolution procedure, included in order to prevent interference by silicate ions, was found to be essential before using the sample for atomic-absorption spectroscopy. We have found that 100 per cent.recovery of the silver present in the ores could not be obtained by a single dissolution step. As discussed later, a re-examination will facilitate the recovery of an additional 1.2 to 5.7 per cent. of silver and is essential for the highest accuracy to be obtained. PREVENTION OF INTERFERENCE BY SILICATE IONS- Rubeska, Sulcek and Moldan' also reported that sulphuric acid depressed the silver signal at 328.1 nm, which was confirmed by Belcher, Dagnall and West.12 Our work confirmed these observations, but in addition we noted a synergistic effect by silicate ions in the presence of strong acids (Table I); silicate ions alone reduced the silver signal to about 93 per cent., but when 10 per cent.V/V of concentrated nitric acid was present the signal was reduced to 17.3 per cent., with 10 per cent. V/V of concentrated sulphuric acid to 46.4 per cent. and with 10 per cent. V/V of concentrated hydrochloric acid to 72.4 per cent. The results in Table I were obtained for solutions containing 2.50 p.p.m. of silver (as the nitrate). Silicate ions were incorporated, where indicated, by adding 3.0 per cent. V/V of sodium silicate solution (40 to 42 .Be). The presence of sodium ions was shown to have no effect on the signal. In the absence of silicate ions, the strong acids exert a much smaller depressant effect on the silver signal, sulphuric acid at the same concentration as above reducing the signal to about 82 per cent. of its original value. Our method overcomes this interference effect by introducing a high concentration of ammonia in the final solutions.The results in Table I1 demonstrate the improved recovery of silver from solutions containing silicate ions plus 5 ml of concentrated hydrochloric acid and 30 ml of concentrated ammonia solution per 100 ml of solution. TABLE I EFFECT OF SILICATE IONS AND STRONG ACIDS ON THE ATOMIC-ABSORPTION Interference by silicate ions was noted by West, West and Ramakrishna.1 SIGNAL AT 328.1 nm FOR 2-50 p.p.m. OF SILVER Transmittance, per cent. Control (no silicate or acid) . . .. .. . . . . 63.5 10 per cent. V / V of Concentrated nitric acid . . . . . . 64.0 10 per cent. V / V of concentrated hydrochloric acid . . . . 63.4 10 per cent. V / V of concentrated sulphuric acid .. . . 69.0 Silicate ions added* . . . . .. . . . . . . 65-5 Silicate* + 10 per cent. V / V of concentrated nitric acid . . 92.4 Silicate* + 10 per cent. V/V of concentrated hydrochloric acid 72.0 Silicate* -b 10 per ccnt. V / V of concentrated sulphuric acid . . 81.0 * 3 per cent. V / V of sodium silicate solution (40 to 42 Absorbance 0.1973 0.1939 0.1979 0.161 1 0.1838 0.0342 0.1428 0.09 16 .Be). Per cent. of control 100.0 98.3 100.3 81.7 93.1 17.3 72.4 46.4May, 19731 MINERAL PRODUCTS BY ATOMIC-ABSORPTION SPECTROSCOPY TABLE I1 RECOVERY OF SILVER BY ATOMIC-ABSORPTION SPECTROSCOPY* FROM SOLUTIONS CONTAINING 3.50 p.p.m. OF SILVER IONS AND 3.0 PER CENT. V/V OF SODIUM SILICATE SOLUTION1 Run No. Absorbance spectroscopy, p.p.m. Ag+ by atomic-absorption Control (no silicate) 0-2660 3-50 1 0.2668 3.51 2 0.2690 3.54 3 0.2683 3.53 4 0.2660 3.50 5 0.2661 3.50 6 0-2645 3-48 7 0.2658 3.50 8 0.2652 3.49 9 0.2653 3-49 10 0.2659 3.50 Average .. .. . . 3.504 Standard deviation, per cent. 0.51 Range . . . . , . . . 0.060 * Silver signal a t 328.1 nm. 7 Including 5 ml of concentrated hydrochloric acid and 30 ml of concentrated ammonia solution per 100 ml of solution. EXPERIMENTAL REAGENTS AND REAGENT BLANKS- 337 Ammonia solution, concentrated-Mallinckrodt analytical-reagent grade, 58 per cent. m/m. Hydrochloric acid, concentrated-Du Pont reagent grade, 37.5 per cent. m/m. Nitric acid, concentrated-Du Pont reagent grade, 70.5 per cent. m/m. Perchloric acid, concentrated-Mallinckrodt analytical-reagent grade, 70 per cent. m/m. HydroJEuoric acid, concentrated-Mallinckrodt analytical-reagent grade, 48 per cent.m/m. Silver-Wire, 99.95 per cent. pure, D. F. Goldsmith Chemical and Metal Corp., Evanston, Sodium silicate solution40 to 42 "B6, City Chemical Corp., New York, U.S.A. The suggested dissolution procedures were carried out without any sample material but sufficient of the reagents were used for a 10-g sample. No detectable atomic-absorption signal was obtained at 328.1 nm, giving a zero reagent blank and establishing that the reagents were sufficiently pure for the intended purpose. IKSTRUMENTAL- The atomic-absorption spectrophotometer used was a Techtron, Model AA-100, instru- ment with a 4-inch solid stainless-steel burner, Type AB41, having a slit width of 0.020 inch and flame path length of 10.0 cm, and in which air at 15 p.s.i.g.and acetylene at 9 p.s.i.g. pressure were used. The acetylene was metered to the burner at a flow-rate of 5 on the panel-mounted flow meter. The burner height was set at 5 on the height adjustment. The lamp used was made by Atomic Spectral Lamps, Melbourne, Australia, and was operated at 4 mA with a coarse gain setting of 4. The 328.1-nm line of silver was used for all measure- ments. The air supply was cleaned and regulated with a Hymatic filter-regulator (Varian Techtron Corp., California, U.S.A.). All of the glassware used was Pyrex. CLEANLINESS OF GLASSWARE- In order to prevent the possibility of silver sorbed on the container walls being carried over into subsequent analyses, Nakagawa and Lakin4 advise a final rinse with dilute cyanide solution.The final sample solutions in our method contain 30 ml of concentrated ammonia solution per 100 ml, and silver is apparently not sorbed from such solutions on to the container walls to a significant extent. We found that one rinse with detergent solution, followed by three rinses with tap water and three with distilled water, is sufficient to remove all silver ions. Glassware cleaned in this manner and refilled with a solution containing 5 per cent. of con- centrated hydrochloric acid and 30 per cent. of concentrated ammonia solution repeatedly gave no detectable silver signal. Ill., U.S.A.338 WALTON: A METHOD FOR THE DETERMINATION OF SILVER IN ORES AND [AnaZyst, Vol. 98 PREPARATION, USE AND STABILITY OF STANDARDS- Standards with silver values similar to those of the samples were run on the atomic- absorption equipment immediately before and after the samples themselves.The standards were made by dissolving silver wire of known high purity in the minimum amount of nitric acid to make 1 litre of concentrated stock solution containing 0.9986 g 1-1 of silver. From this solution, a dilute stock solution was prepared, which contained 0.09986 g 1-1 of silver (0.09986 mg ml-l). The working standards were prepared from this dilute silver stock solution. Except when in actual use, all standards and stock silver solutions were kept in darkness. To prepare the working standards, a measured amount of the dilute silver stock solution was placed in a 100-ml calibrated flask together with 5 ml of concentrated hydrochloric acid; about 50 ml of water were added, then 30 ml of concentrated ammonia solution, and the solution was cooled to room temperature.Finally, the volume was made up to the mark with water and the solution was shaken to ensure homogeneity. Standards covering the range 0.0 to 10.0 p.p.m. of silver were prepared in this manner. A series of standards prepared by including the addition of acid and evaporation, as for actual ore samples, gave identical results. As found by Lockyer and Hames3 and others, the graphs of absorbance ueysus concen- tration also gave straight-line calibration graphs. However, sample results were calculated from the atomic-absorption spectroscopic results obtained by running suitable standards simultaneously with the samples, thus negating any changes in the absorbance to concen- tration ratio due to burner adjustments, etc.The standards prepared for this study containing 5 per cent. of concentrated hydrochloric acid and 30 per cent. of concentrated ammonia solution showed no measurable instability over a 7-month period. Fresh standards prepared at intervals up to 210 days later gave identical atomic-absorption spectroscopic results within the limit of the reading error involved ( 5 0 . 3 per cent. transmittance). In contrast, standards prepared for two solvent-extraction methods (dithizone in ethyl propionate, and triisooctyl phosphorothioate in benzene) gave substantially lower atomic-absorption signals within 4 hours. The excellent stability of the working standards prepared from pure silver is one of the main advantages of our method.PROCEDURE- For each 2 g of ore used, add, in a 260 or 400-ml beaker, 5 ml of concentrated hydrofluoric acid, 10 ml of concentrated hydrochloric acid, 10 ml of concentrated nitric acid and 5 ml of concentrated perchloric acid (see preca~tionsl~). Evaporate the mixture to dryness on a hot-plate, covering the samples with a watch-glass for the first 30 minutes. Add 5ml of concentrated hydrochloric acid plus 30 ml of water (a portion of this water can be used to rinse any residue that remains on the watch-glass into the beaker). Warm the mixture for 5 minutes on a hot-plate, then decant off the liquid into a 100-ml calibrated flask. Rinse the beaker and the residue with 5ml of water, allow the solid to settle and decant off the liquid into the flask.Rinse the beaker and the residue with 30 ml of concentrated ammonia solution, allow the solid to settle and decant off the liquid into the flask. Rinse the beaker and residue with one or two more 5-ml portions of water, allow the solid to settle and decant off the liquid into the flask. Mix the contents of the flask, cool to room temperature, make the liquid in the flask up to the 100-ml mark with water, stopper the flask and shake it. Allow any residue to settle. Adjust the instrument settings to read 100 per cent. transmittance on a solution con- taining 5 ml of concentrated hydrochloric acid and 30 ml of concentrated ammonia solution per 100 ml. Standardise the instrument at 328-1 nm by using standards of known silver content in the appropriate range of the sample values and containing 5 ml of concentrated hydrochloric acid and 30 ml of concentrated ammonia solution per 100 ml of final solution.Aspirate the clear supernatant sample solution directly into the flame of the atomic- absorption instrument, with a wavelength setting of 328.1 nm. (If desired, a small portion of the sample solution can be filtered and the filtrate aspirated into the flame.) The time required for this method is about 3 hours, most of which is devoted to evaporat- ing the solutions to dryness.May, 19731 NINERAL PRODUCTS BY ATOMIC-ABSORPTION SPECTROSCOPY 339 LOSS OF SILVER BY SORPTION- The possible loss of silver by sorption on the walls of the containing vessels has been considered by West, West and Ramakrishnal and by Rubeska, Sulcek and Moldan,’ who recommend the use of mercury(I1) ion in order to form complexes, thus inhibiting the sorption of silver.This method would have the greatest application when extremely small silver values are to be considered, as for example, in the determination of silver in natural waters. To several of the samples we added mercury(I1) ions in the recommended amount (30 mg) before the dissolution procedure. No difference was found in the silver values obtained with and without mercury(I1) ions, and we conclude that the sorption of silver from the sample solutions is negligible. The excellent long-term stability of the standards also supports this view. DETERMINATION OF SILVER BY ATOMIC-ABSORPTION SPECTROSCOPY- The atomic-absorption signal of many metals can be enhanced considerably by the use of organic solvent^.^^^^^-^^ The use of these solvents may be desirable or even necessary in certain instances (very low metal values, for example), but it is not essential to our method.On the contrary, the use of solvents as a concentration device or to enhance the signal intro- duces additional complications that are best avoided in a precision method. In our method, we always obtained straight-line relationships between the absorbance and the silver concentration in the range 0 to 10 p.p.m. of silver at 328.1 nm. In the region of 50 per cent. transmittance, our estimated sensitivity of 50.3 scale division corresponds to about &O-10 troy oz per net ton for a 2-g ore sample when using a final dilution to 100 ml.This sensitivity corresponds to a reading error of about h1.5 per cent. in the determination of silver, for an average sample. As the precision of the fire-assay method is usually con- sidered to be about 2.5 per cent. or more, our method has a comparable precision, even without the use of solvent-concentration or signal-enhancement techniques. The high concentrations of ammonia in our final sample solutions are preferable to high concentrations of acid in terms of equipment corrosion problems. In order to show that the atomic-absorption signal obtained a t 328.1 nm was derived only from silver ions, the final solutions from several samples (made 8 N with respect to nitric acid) were extracted with a 30 per cent. solution of triisooctyl phosphorothioate in benzene.According to Nakagawa and Lakir~,~ based on work by Handley and Dean,lG this extractant is not specific for silver, but is highly selective [only mercury(II), tantalum, vanadium and silver can be extracted under the conditions used]. After four successive extractioi~;, the sample solutions (aqueous phase) gave readings of 100 per cent. transmittance at 328-1 nm, showing that the element that generated the signal obtained at this wavelength was com- pletely extracted by the triisooctyl phosphorothioate in benzene. Significantly high con- centrations (1000-fold molar concentrations based on the silver originally present) of mer- cury(II), tantalum and vanadium ions added to the original sample solutions failed to generate a visible signal at this wavelength.We therefore conclude that the signal originally obtained from the sample solutions at this wavelength was due only to the presence of silver ions. In order to prove that further additions of silver to our sample solutions would give absorbance read-outs that are linearly proportional to the concentrations of silver added, we added known Concentrations of standard silver stock solutions to aliquots of the final sample solutions, then re-measured the absorbances of the solutions (the “method of standard addition”). The results obtained show that the “recovery” of the added silver by the atomic- absorption procedure in our sample solutions was linear, and very close to that expected by comparison with standards based on the pure metal.In twelve such experiments, a total of 39-179 p.p.m. of silver was added and a total of 39.060 p.p.m. of silver was “recovered” by the atomic-absorption procedure. The apparent net loss of 0.3 per cent. is within the error of measurement. We therefore conclude that any other substances present in the final solutions exerted neither an enhancing nor a depressing effect upon the atomic-absorption signal for silver under the conditions used. The fact that we were able to extract the silver from the sample solutions so as to give a reading of 100 per cent. transmittance, and the fact that we were able to obtain a linear recovery of the silver added to the sample solutions, is strong evidence that background absorption is not a significant factor in our method.340 WALTON: A METHOD FOR THE DETERMINATION OF SILVER IN ORES AND [Analyst, Vol.98 RE-EXAMINATION PROCEDURES- One of the major difficulties encountered in this study was that of obtaining complete dissolution of the silver present in the ores. In view of this problem, eight of the oxide ores used in this study were submitted to a re-examination procedure. In every instance, an additional amount of silver, ranging from 2.3 to 5.7 per cent. of the amount found in the first determination, was recovered (see Table 111). TABLE 111 COMPARISON OF RECOVERIES OF SILVER FROM ORES BY ATOMIC-ABSORPTION SPECTROSCOPY FROM FIRST DETERMINATION, RE-EXAMINATION AND SECOND RE-EXAMINATION PROCEDURES Ore Oxide ores- No. 1 No. 6 No. 9 No. 11 No. 13 No. 17 No. 18 No. 20 No. 2 No.12 No. 15 No. 16 Sulphide ores- Average silver content (troy oz per net ton) in 7 first first determination re-examination 12.90 16-05 8-12 2.19 4-13 7.35 4.39 3.65 2.87 2-56 5-22 9.9 1 Sulphide concentrate- No. 19 1-33 0.42 0.74 0.28 0.05 0.10 0.23 0.25 0.15 0-07 0.03 0.09 0.17 0.02 Per cent. of silver recovered on first re-examination 3.3 4.6 3.4 2.3 2.4 3.1 5-7 4.1 2.4 1.2 1.7 1-7 1.5 Average silver content (troy oz per net ton) in second re-examination 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 A second re-examination procedure performed on the same samples did not yield any further signals for silver. It therefore appears that for the highest accuracy to be obtained, it is essential to add the results from one re-examination procedure to those from the original determination, but that additional re-examinations are not necessary.Four sulphide ores gave additional silver values of 1-2 to 1.7 per cent. of the amount found in the first determination on re-examination (Table 111). In general, it is probable that this type of ore, which contains silver in association with galena, is more soluble than the siliceous oxide ores. Mill products, such as blister copper, which are completely dissolved by the method given, would of course require no re-examination. RESULTS We examined twenty samples, consisting of oxide and sulphide ores, sulphide concen- trate and blister copper, by the proposed method, with six to eight atomic-absorption spectro- scopic determinations being made on each sample. The results are presented in Table IV, which also gives the average and range of values and the percentage standard deviation.The results show an average silver content obtained by atomic-absorption spectroscopy that is 10.9 per cent. higher than that obtained by the fire-assay method. The average relative standard deviation is 2.2 per cent. DISCUSSION A substantial amount of recent information has shown that the fire-assay method for determining precious metals is subject to serious negative errors due to loss of the metals at various stages of the determination. For ruthenium, losses of up to 30 per cent. have been found,17 for rhodium up to 12 per cent.,ls iridium up to 63 per cent.19 and platinum as high as 89 per cent. (for neutral slags).20& . y , 19731 MINERAL PRODUCTS BY ATOMIC-ABSORPTION SPECTROSCOPY TABLE IV RESULTS FOR THE DETERMINATION OF SILVER IN ORES AND OTHER MINERAL PRODUCTS BY FIRE-ASSAY AND ATOMIC-ABSORPTION METHODS 341 Sample No.Type 1 Oxide ore 2 Sulphide ore 3 Oxide ore 4 Sulphide 5 Oxide ore 6 Oxide ore 7 Blister copper 8 Oxide ore 9 Oxide ore 10 Oxide ore 11 Oxideore 12 Sulphide ore 13 Oxide ore 14 Blister copper 15 Sulphide ore 16 Sulphide ore 17 Oxide ore 18 Oxide ore 19 Sulphide 20 Oxideore concentrate concentrate Average Best values for silver (troy oz per Standard Range (troy oz per net ton) deviation of atomic- value for silver net ton) by atomic- of atomic- absorption Difference, Percentage No. of by fire absorption absorption results B - A difference, deter- assay spectro- results, (troy oz per (troy oz B - A x 100 8 12.02 13.32 1.1 0-45 1-30 + 10.8 6 2.80 2-94 3.1 0.3 1 0.14 + 5.0 minations ( A ) scopy (R) per cent.net ton) per net ton) 7 8 6-73 7.14 2.1 0.70 0.41 + 6.1 6 8 8 8 7 8 6 6 6 6 6 7 7 6 6 0-64 1.55 16.20 3-2 1 14-6 8-00 8-85 2-10 2-40 3-87 2-04 5.02 9.25 7-32 4.05 0.80 1.64 16.78 3-56 19.8 8.40 9.79 2.24 2.5D 4.23 2-38 5.30 10.08 7-58 4-64 2.8 2.7 1.7 4.1 4.8 4-6 0.9 2.1 0.5 3-7 1.1 1.5 1.7 0.1 1.6 0.09 0.19 1.37 0.46 3.03 1-49 0.55 0-17 0.1 1 0-54 0.15 0.41 0.75 0.22 0.33 0-16 0.09 0-58 0.35 5.2 0.40 0.94 0.14 0.19 0.36 0.34 0.28 0-83 0.26 0.59 + 25.0 + 5.8 + 3.6 + 10.9 + 35.6 -/- 5.0 -+ 10.6 + 6.7 + 8.0 + 9.3 + 16.7 + 5.6 + 9.0 + 3.6 + 14.6 6 1.16 1.35 2.7 0.12 0.19 + 16.4 6 3.48 3.80 1-4 0-24 0-32 + 9.2 Average . . 2.2 Average .. +lorn9 A recent radiochemical study of the fire-assay method for determining silver was reported by Faye and InmanP2l who showed that the major loss of this metal occurs in the cupellation step, approximately 2.5 per cent. being lost to the cupel even at 890 to 900 "C. They noted that at 1000 "C this loss is doubled, and also mentioned that "under certain conditions often used in practice the total error due to losses may be as high as 5 to 10 per cent." They also mentioned that the loss of silver to the cupel is about 25 per cent. higher for bone-ash than for magnesia cupels, and indicated that buttons of less than 25 g will result in incomplete collection of the silver. In another radiochemical study, by Nakamura and Fukami,22 losses of silver (again, primarily to the cupel) from 5 to 30 per cent.were reported. In the above studies, loss of silver by volatilisation has been for the most part neglected, as it is small compared with other sources of error. However, even in 1911, F ~ l t o n ~ ~ reported losses of silver by volatilisation of about 5 to 6 per cent. from 750 to 1000 "C, and losses of up to 29 per cent. by volatilisation of silver in the presence of gold (silver to gold ratio of 2: 1). Fulton also quoted extensive work by Eager and Welch, Godshall, Kaufman and Hillebrand, and Allen on the loss of silver under various conditions of cupellation. Although these two processes appear to cause the greatest losses of silver during the fire assay, there are others. Losses of silver during crucible fusion and scorification have also been noted.The above studies support the view that fire-assay methods, despite their extensive usage and historical background, are obviously subject to losses of metal that can substantially affect their accuracy. Except for the operations of weighing the pulp and the final bead, which operations could be subject to either positive or negative errors, all other major errors inherent in the fire-assay method are negative. In view of this and of the foregoing studies on the various possible sources of error, it is hardly surprising that our atomic-absorption spectroscopic342 WALTON results yield silver values that are consistently higher than those obtained by the fire-assay method. Rawling, Greaves and Amos,24 working with lead concentrates, reported results obtained by an atomic-absorption method for the determination of silver that were, on average, about 0.5 troy oz per net ton higher than those obtained by the fire-assay method in the range 16 to 18 troy oz per net ton. CONCLUSIONS The proposed atomic-absorption method gives an average silver value in the ores and mineral products studied that is 10.9 per cent.higher than that given by fire assay. We have attempted to show that the signal obtained from the samples is due to silver alone, and that the presence of silver in the sample solutions gives a linear signal response for absorbance versus concentration. Under these circumstances, it can be concluded that there is actually a higher silver content in these samples, and that the bias of negative errors inherent in the fire-assay method is responsible for the comparatively low results obtained by that method.Much of the economics of mining is dependent upon the ultimate accuracy of determina- tions of precious metal values in the ore, and we suggest that more accurate results for such metal values are obtained by the proposed atomic-absorption method. Also, procedures for the recovery of metals are aimed primarily at the recovery of metal values as predicted from assays. If there is actually a higher silver content than is indicated by the fire-assay method, this fact must be recognised before serious attempts can be made to win the higher metal values from the ore by metallurgical procedures. We gratefully acknowledge assistance and special supplies from Jesse F. Bingaman and David M.Dennis of Western New Mexico University, Silver City, New Mexico; Harold E. Richard of Hawley & Hawley (Tuscon, Arizona) ; Dickenson Laboratories (El Paso, Texas) ; Walter Cox and the Hurley Laboratories of the Kennecott Copper Corp. (Hurley, New Mexico) ; W. R. Hein and W. C. Lashley of the Silver City Testing Laboratories; and J. Wilson and Ken Wallin of the Gila Analytic and Research Laboratory (Silver City). This research was financed in part by a grant from the Research Committee of Western New Mexico University. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. REFERENCES West, F. K., West, P. W., and Ramakrishna, T. V., Environ. Sci. Technol., 1967, 1, 717. Greaves, M. C., Nature, Lond., 1963, 199, 552. Lockyer, I<., and Hamcs, G. E., Analyst, 1059, 84, 385. Nakagawa, H. M., and Lakin, H. W., Prof. Pap. U.S. Geol. Surv., No. 525-C, 1965, C172. Huffman, C., jun., Mensik, J . D., and Rader, L. F., Ibid., No. 550-B, 1966, €3189. Wilson, L., Analylica Chim. Acta, 1964, 30, 377. Rubeska, I., Sulcelr, Z., and Moldan, B., Ibid., 1967, 37, 27. Hickey, L. G., Ibid., 1968, 41, 546. Olsen, A. M., Atom. Absovption Nezusl., 1965, 4, 278. Thompson, C. E., Nakagawa, H. M., and Van Sickle, G. H., Prof. Pap. U.S. Geol. Surv., No. 600-J3 Huffman, C., jun., Mensik, J . D., and Riley, L. B., Circ. U.S. Geol. Surv., No. 544, 1967, 1. Belcher, R., Dagnall, R. El., and West, T. S., Talanta, 1964, 11, 1257. Muse, L. A., J . Chem. Educ., 1972, 49, A463. Allen, J. E., Spectrochim. Acta, 1961, 17, 467. Lockyer, R., Scott, J. E., and Slade, S., Nature, Lond., 1961, 189, 830. Handley, T. H., and Dean, J . A., Analyt. Chem., 1960, 32, 1878. Thiers, R., Graydon, W., and Beamish, F. E., Ibid., 1948, 20, 831. Allen, W. F., and Beamish, F. E., Ibid., 1950, 22, 431. Barefoot, R. R., and Beamish, F. E., Ibid., 1952, 24, 840. Hoffman, I., and Beamish, F. E., Ibid., 1956, 28, 1188. Faye, G. H., and Inman, W. R., Ibid., 1959, 31, 1072. Nakamura, Y., and Fukami, K., Japan Analyst, 1957, 6, 687. Fulton, C. H., “A Manual of Fire Assaying,” 2nd Edition, McGraw-Hill, New York, 1911. Rawling, B. S., Greaves, M. C., and Amos, M. D., Nature, Lond., 1960, 188, 137. 1968, B130. Received November 14th, 1972 Accepted January M A , 1973

ISSN:0003-2654    

DOI:10.1039/AN9739800335

出版商:RSC    

年代:1973

数据来源: RSC

摘要:

Analyst, May, 1973, Vol. 98, $9. 343-346 343 A Simple Time-delay Accessory for Use with Micro-scale Sampling Atomic-absorption Techniques BY H. T. DELVES AND R. B. REESON (The Hospital for Sack Children and Institute of Child Health, Great Orwaond Street, London, W.C.l) A time-delay accessory for use with micro-scale sampling atomic- absorption spectrophotometry is described that facilitates the interpretation of chart recordings as a first step towards the automatic calculation of results. THE current concern with environmental pollution has resulted in large numbers of people being screened for excessive exposure to lead by means of blood-lead analyses. Micro-scale sampling and flameless atomic-absorption method~l-~ are becoming in- creasingly important for this application because of their low sample volume requirements.One such technique, the Delves cup method,3 has been used successfully at the Hospital for Sick Children for the past 3 years. This and other similar methods produce two or more signals from the combustion and atomisation of each sample, of which only one is the lead atomic-absorption signal. No difficulty arises in discriminating the lead atomic-absorption from the non-specific signals that precede it as they are adequately resolved in time. How- ever, the interpretation of chart recordings with multiple signals from each sample can be tiring and time consuming when the analysis rate approaches 200 tests per day. It was therefore decided to automate the interpretation of the chart recordings and the calculation of the results that are based not only on peak-absorption values but also on integrated absorption signals.The first stage in this project, which is reported here, was the development of a time-delay circuit to eliminate the recording of the non-specific absorption signals that are produced during the combustion of the sample. This elimination is essential for the successful integration of the lead signal and although it can be achieved by using background- correction accessories with a continuum source, these are expensive and require the use of double-beam instruments or single-beam instruments designed for that purpose. The device described here is simple, inexpensive and can be used with single- or double-beam instruments. This delay circuit will eliminate the recording of all the non-specific absorption signals from the combustion products of the sample that precede the lead signal.It will not eliminate any non-specific absorption signals that occur simultaneously with the lead signal. The delay circuit, which is shown schematically in Fig. 1, consists of two coupled inte- grated-circuit monostable multivibrators, which switch the mains supply to the recorder drive motor or to the recorder directly, and subsequently switch the input signal to the recorder. Initiation of the timing sequence is accomplished by closure of a microswitch that is actuated by the sample cup carrier. The total time of running of the recorder is fixed at 15 s, whereas the delay is continuously variable between 0-4 and 3-6 s; during this delay time a pre-set reference voltage is fed to the recorder to give a deflection equivalent to 100 per cent.transmission. A by-pass switch enables the chart recorder to be run continuously. RESULTS A number of blood-lead determinations were carried out by the method described previously .3 Fig. 2 shows that increasing the delay time before recording the absorption signals has no effect upon the peak lead atomic-absorption signal but decreases the non-specific signal to zero at delay times of greater than 2.4 s. The optimum delay time was 2.6 s, and this delay was used to establish the results shown in Fig. 3. These results show clearly that the chart recordings are simplified and are much easier to interpret. CONCLUSION The time-delay accessory described here facilitates the interpretation of chart recordings produced by micro-scale sampling atomic-absorption techniques.This development should @ SAC and the authors.344 DELVES AND REESON: A SIMPLE TIME-DELAY ACCESSORY FOR USE [Analyst, VOl. 98 TMay, 19731 WITH MICRO-SCALE SAMPLING ATOMIC-ABSORPTION TECHNIQUES 80 c; 5 70- u Q) a L < 60- C 2 g , 50- 5 m C - 2 40- 3 c(J w m 30- .- + ff $4 g 20- 10- 40 - 30 20 10 - - - 1.0 2-0 3.0 5 0 t ' " " ' " ' - Delay time/s Fig. 2. The effect of delay in the start of recording on absorption signals for blood-lead analysis. A, lead atomic-absorption signal ; and B, non-specific absorp- tion signal from combustion products of blood sample. Each point is a single determination on 10 1-11 of blood containing 140 pg of lead per 100 ml I A L 6 Sample I B !3 345 Sample II A Pb / \ L B Fig 3.Absorption signals from the micro-sampling atomic-absorption spectrophotometric determination of lead in blood. A, no time delay used; B, delay time of 2.6 s before recording signals; sample I, 140 pg of lead per 100 ml; and sample 11, 20 pg of lead per 100 ml346 DELVES AND REESON prove useful to workers carrying out large numbers of blood-lead analyses, and in particular, to those who use single-beam instruments that cannot be equipped with background-correction accessories with a continuum source. We thank Professor Barbara E. Clayton and Dr. G. Pampiglione for their interest in this work. One of us (H.T.D.) gratefully acknowledges financial support from the Medical Research Council. REFERENCES 1. 2. 3. Massman, H., Spectrochim. Acta, 1968, 23B, 215. West, T. S., and Williams, X. I<., Analytica China A d a , 1969, 45, 27. Delves, H. T., Analyst, 1970, 95, 431. Received October 16th, 1972 Accepted January 12th, 1973 Appendix LIST OF COMPONENTS Variable resistors- = 5 kS2, wire-wound = 1 kS1, pre-set VRl VR, Capacitors- Cl C, c37 c4 = 500 pF, 16 V working, electrolytic = 0-1 pF, 30 V working, ceraniic = 0.01 pF, 30 V working, ceramic Semicondwtors- IC,, IC, = Texas instruments, SN74121N (pin connections indicated) Tr,, Tr, = Texas instruments, 2N3904 D,, D2, D,, D,, D, = Texas instruments, 1N4001 Re jays- KLA = Radiospares, type 11A RLB = Radiospares, type “LP-Relay-12 V” Switches s,, s, = Radiospares, type “MT-SW-DPCO” s3 = Radiospares, type “Micro I1/T” Sundry componeflts- Nl = Radiospares, type “M-NEON-250 V-AMBER”

ISSN:0003-2654    

DOI:10.1039/AN9739800343

出版商:RSC    

年代:1973

数据来源: RSC