426 Analyst, June, 1973, Vol. 98, pp. 426-431 Precise Coulometric Determination of Acids in Cells Without Liquid Junction Part IV.* The Assay of Primary Standard Sulphamic Acidt BY E. BISHOP AND M. RILEY: (Chemistry Department, University of Exeter, Stocker Road, Exeter, EX4 4QD) The precise (1 to 2 p.p.m.) location of the end-point in the pre-titration of the supporting electrolyte and in the titration of sulphamic acid has been examined, and d.c. differential electrolytic potentiometry gives excellent results. The methods and simple apparatus previously described in Parts I, I1 and I11 are applied to the assay of primary standard sulphamic acid previously collaboratively assayed by mass titrimetry, and give results with an accuracy and precision of 100 p.p.m., and a 95 per cent.confidence level of 0.02 per cent., but with a positive bias of 0.014 per cent., the reasons for which are canvassed. The method is of high merit; it is simple, fast and direct, and is capable of further refinement. EARLIER work has established that current sources of adequate stability and a timing device of adequate accuracy are available for high-precision amperostatic acidimetryl ; that suitable conditions can be chosen that permit the use of a level of deposition of more than 5000 C of bromide on a silver anode as the auxiliary reaction in the same compartment in which the cathodic determination of the acid is conducted2; and that the error arising from the solubility of silver bromide in the cell electrolyte and consequent deposition of silver on the working cathode can be evaluated with adequate accuracy and preci~ion.~ There remained for investi- gation first, a means of locating the end-point of the reaction with adequate accuracy and precision, and secondly, the testing of the whole method by the assay of an independently standardised primary standard acid. For the latter, sulphamic acid, which was collaboratively assayed by the Analytical Standards Sub-committee of the Analytical Methods Committee of the Society for Analytical Chemistry and recommended and accepted as an international primary was selected.For end-point location, d.c. differential electrolytic potentiometry has an adequate reserve of sensitivity.s EXPERIMENTAL The equipment used has been described previous1y.l The cell used in the final deter- minations is shown in Fig.1. The platinum-gauze cathode was the larger Model 72020, The silver anode comprised two electrodes, one each of sizes A and B, of total area 275 to 315 cm2. The twin antimony differential electrolytic potentiometric electrodes were mounted centrally so as to be out of the generating current field, and were held in a rubber bung, which was replaced with a plain bung when the indicating electrodes were removed from the cell. Pure oxygen, humidified by passage through a water bubbler, was passed over the surface of the solution, and served to exclude carbon dioxide and to oxygenate the solution for the benefit of the antimony electrodes.’ Excess of oxygen escaped through an exit tube that was provided with a spray trap. The platinum-gauze electrode was initially cleaned by immersion for 1 to 2 minutes in freshly prepared aqua regia, followed by very thorough washing with and storage in water.Subsequently, after “silver error” determinations, it was immersed in a 0.2 M solution of iron(I1) sulphate in 1.0 M sulphuric acid for at least 30 minutes before further use.3 Other electrodes were treated as described previously.1 The circuit, in which the cell is indicated in plan view, is shown in Fig. 2. The constant-current sources were run continuously, being switched to dummy loads when not in use. The selector switch on the P3 potentiometer was used to select the standard resistor that was required when power supplies were exchanged. Switching in power supplies to the cell automatically triggered the clock, and switching over to the dummy load stopped the clock.For particulars of Parts I, I1 and I11 of this series, see reference list, p. 431. t Presented at the Second SAC Conference, Nottingham, 1968. @ SAC and the authors. Present address: Electronic Instruments Limited, Hanworth Lane, Chertscy, Surrey.BISHOP AND RILEY 427 Twin antimony DEP electrodes Platinum -gauze cat hod e Fig. 1. Coulometric cell for acid determina- tion (DEP denotes differential electrolytic poten- tiometric) SULPHAMIC ACID- by the Analytical Standards Sub-committee. with reference to atomic-mass grade silver at 100.001 per cent. purity. dried under vacuum over fresh magnesium perchlorate for at least 24 hours before use. WEIGHING AND TRANSFER OF SAMPLES- Catch-weights of 4 to 5 g of dry sulphamic acid were weighed in small glass weighing bottles that had outside-fitting lids, with a hole drilled in each ground face so that pressure could be equalised by turning the lid to register the holes.Preliminary rough weighings to the nearest 1 mg were made on a CL3 balance, and the accurate weighings made on the Samples were taken from a 100-g batch of sulphamic acid provided in a sealed container The sample had been collaboratively assayed The samples were Fig. 2. Circuit for coulometric acidi- metry: c, plan view of coulometric cell; PSU (power supply unit), 2-A source AS 1411 ; OACS, operational amplifier constant- current source ; P, P3 potentiometer ; T, crys- tal clock; M, 39A pH meter; Vs, DEP voltage source; R,, 0.6-SZ standard resistor (two 1-SZ standards in parallel) ; Rz.104 standard resistor; R,, dummy load resistor, about 0.6 a; R,, 1042 dummy load resistor; and RB, DEP ballast resistor428 [Analyst, Vol. 98 special atomic-mass balance that had a standard deviation of 1.3 pg on 100 g and was provided with both 5 and 0-5-mg riders. Once the riders had been located in the correct notch, the balance was left for 30 minutes, then released and the swings were observed by telescope from a distance of 20 feet. The sequence of measurements were zero, weighing bottle PLUS sample, sensitivity, zero, empty weighing bottle, sensitivity, zero. Buoyancy corrections were made by using 8.0 g ~ m - ~ for the density of the weights and 2.126 g ~ m - ~ for the density of sulphamic acid.The sample was transferred into the cell, which had already been filled with pre-titrated electrolyte, with oxygen passing through it, via a small glass funnel inserted in the lid. Most of the crystals passed straight through into the cell. The lid of the weighing bottle was replaced and the bottle re-weighed. Before adding the sample to the cell, 20 ml of the neutral electrolyte were withdrawn into a hypodermic syringe, and this solution was used to wash the funnel thoroughly, which was then removed and the hole closed with a bung. END-POINT LOCATION- A conventional differential electrolytic potentiometric circuit8 was used as indicated in Fig. 2. The source voltage was 240 V and the ballast resistance was 960 MQ, giving a dif- ferentiating current density of about 4 pA cm-2.Differential potentials were measured on the 39A pH meter and recorded on a 10-mV recorder. The electrodes were inserted for pre-titration, then removed and replaced within 1 C of the end-point. Preliminary appraisal of the response was made by titrating 7 to 10mg of sulphamic acid at concentrations of 2.5 to 3-5 X M at currents of 10 mA in 0.03 M potassium bromide solution. Satisfactory differential peaks, 50 to 80 mV in amplitude, were obtained at differentiating current densities of 2.2 to 6.6 pA cm-2 and ballast loads of 3-6 x 1O1O to 1.1 x 1011 VQ, but with continuous generation the results showed a positive bias of 1 to 2 per cent. Bishop and Short6 determined similar amounts of perchloric acid by continuous generation at 2 to 10 mA but found a bias that was not greater than 0.2 per cent.The cause of the bias (electrolyte retention in the meshes of the electrode) was not immediately apparent, but it was concluded that incremental generation, with automatic time integration with the crystal clock, would be more apposite in high-precision work. With incremental generation, relatively long times were required for the differential potential to become stabilised near the end-point. Times of 4 to 5 minutes were usual in pre-titration (see below) and 6 to 10 minutes during the final end-point determination. The positive-going drifts were not caused by the electrodes, because glass indicator electrodes referenced to a saturated mercury( I) sulphate electrode showed similar drifts to higher pH values. The drift is due to slow diffusion of hydroxyl ions from the interstices of the gauze electrode into the bulk of the solution.Eckfeldt and Shafferg observed similar drifts in unbuffered solutions when using an electrode constructed in the form of a mesh of platinum strips, and found that the drift was eliminated when smooth platinum wire was used. Satisfactory end-points were achieved in sulphamic acid determinations by using incremental generation in the vicinity of the end-point, although the long equilibration times between increments made the procedure rather tedious. Differential potentials were con- sidered to be stable when a drift of less than 0.2 rnV min-l was recorded. A typical end-point determination is shown in Fig. 3 for incremental generation at 10mA. Exact end-points could be extracted from such curves geometrically,6 but could usually be estimated visually with sufficient accuracy; an error of 1 s in the end-point location corresponded to an over-all error of 2 p.p.m.in the determination. Curves for the pre-titration were identical, although peak heights were greater by about 10mV. Pre-titration of the cell electrolyte, before addition of the sample, was carried out at a current of 5 mA so as to correct for any residual carbon dioxide in the water or acidic impurities in the potassium bromide. This pre-titration required about 0.05 C, corresponding to a residual carbon dioxide concentration of 1.7 x M, assuming this to be the only impurity present, and is in good agreement with the earlier value6 of 1.6 x 10-6 M. PROCEDURE FOR SULPHAMIC ACID ASSAY- From the approximate mass of the sample, the theoretical amount of potassium bromide was calculated, and this amount plus an excess of 1.0 to 1.1 g was weighed.The power supply to the crystal clock was switched on at this stage: as the circuit is entirely of the low-power solid-state type, the clock reaches thermal equilibrium quickly and does not need BISHOP AND RILEY : PRECISE COULOMETRIC DETERMINATIONJune, 19731 OF ACIDS I N CELLS WITHOUT LIQUID JUNCTION. PART IV 429 0 0 0 0 0 0 410 430 450 Generation time a t 10 mA/s Fig. 3. Typical end-point location graph to be run continuously. The pre-conditioned cell was rinsed with carbon dioxide free water, the potassium bromide and magnetic follower were placed in it, the lid and electrode assembly fitted and the flow of oxygen was started.About 300ml of carbon dioxide free water were dispensed from the reservoir into the cell through a spare hole in the lid, which was then closed with a bung and the stirrer switched on. The operational amplifier constant- current source was set to give an output of 5 mA, and the potential drop across the standard resistor and the temperature of the latter were noted. Pre-titration was performed by generation in increments of 2 to 4 s, allowing the differential potential to become stabilised between increments. The values of the cumulative generation time and the differential poten- tial were noted after each increment, and generation was continued until the differential potential showed a marked fall.The quantity of electricity passed after the differential elec- trolytic potentiometric peak was calculated. The antimony indicator electrodes were removed and washed with and stored in water, the hole in the lid being then closed with a bung. The 2-A current source with output settings of 2.0 A and 39.9 V (the maximum setting) was monitored, and the sulphamic acid sample was transferred into the cell. The crystal clock was re-set to zero. The potential drop across the standard resistors and their tem- peratures were noted and generation at 2 A was started. The theoretical quantity of elec- tricity, QT, required for neutralisation was calculated, and generation at 2 A was continued to within 50 C of QT. The lid and walls of the cell, the electrode stems and the spray trap were carefully washed down, which was effected by fitting a 20-ml hypodermic syringe with a suitably bent needle, opening the spare hole in the lid, withdrawing 10 to 15ml of cell electrolyte into the syringe and using this solution for the washing down, further portions of electrolyte being withdrawn if necessary. Use of the bent needle enabled all of the washings to be carried out without removing the lid.Generation at 2 A was continued to within 5 C of QT, when the 2-A current source was disconnected, the generation time at 2 A was noted and the washing operations were repeated. Measurements of the current flowing were made at 2-minute intervals during electrolysis a t 2 A. The operational amplifier source, set to 10-mA output, was then monitored, the crystal clock re-set to zero and generation continued to within 1 C of QT.After a further washing operation, the antimony indicator electrodes were rinsed with carbon dioxide free water, inserted in the cell and the differential electrolytic potentiometric circuit activated. Genera- tion at 10 mA was continued in 10 to 20-s increments until the differential potential showed a significant increase, when a final washing down was performed. The final part of the generation was conducted in 4 to 5-s increments, noting the time and stabilised differential potentials after each increment, and was continued until four or five points after the end-point had been passed. The platinum-gauze electrode was then carefully removed from the cell, gently washed with water and drained several times, and immersed in water for 5 to 10 minutes.It was then mounted in a stripping cell (Fig. 1 in Part 1113) containing 0.1 M perchloric acid electrolyte, and the 39A pH meter was connected between the gauze electrode and the saturated mer- cury(1) sulphate electrode. The operational amplifier source was set to 5 mA, the crystal430 [finahst, vol. 98 clock re-set and the silver stripped from the gauze electrode as previously described,3 the time taken to reach a potential of +O-S V being measured by stopping the clock manually at this stage. CALCULATION OF RESULTS- The end-point was located graphically as in Fig. 3, for generation time at 10 mA, and the total number of coulombs used in the acid determination, QF, was calculated as follows- QF = QD + QE - Qs where QD is the number of coulombs passed at 2 A, calculated from the mean of the measured values of the current, QE is the number of coulombs passed at 10 mA in reaching the end-point PZZH any excess of generation at 5 mA in the pre-titration, and Qs is the number of coulombs required for the anodic stripping of silver.The result of the coulometric assay was therefore the ratio of QF to QT, expressed as a percentage. BISHOP AND RILEY : PRECISE COULOMETRIC DETERMINATION RESULTS AND DISCUSSION The results obtained on one batch of sulphamic acid, comprising eight assays of 4 to 5-g portions, are shown in Table I and are typical. The values of QT were calulated, to the nearest 0.01 C, by using the recommended value of the Faraday constantlo of 96 486.70 & 0-5 A s mol-1 and a relative molecular mass of 97.093 for sulphamic acid.Evidently, the method is capable of high precision: the 95 per cent. confidence limits are &Om02 per cent. The two results, 99.999 and 100.028, attenuate the precision rather severely, but cannot be excluded because they are, although only just, statistically significant. TABLE I COULOMETRIC ASSAY OF BATCH 3 SULPHAMIC ACID Sample masslg 4.850 03 4.496 97 4-702 20 6.138 46 4.325 46 4.320 23 4.683 28 4.895 63 Quantity of electricity h r \ Q D l C Q E l C QslC 4815.25 7.422 2.543 4468.38 4.645 2- 194 4672.73 3.131 2.246 5 104.42 5.675 2.637 4297.32 4.432 2.101 4289.51 6.3 16 2.173 4654.33 2.968 2.400 4864.78 3.863 2.506 Mean .. .. .. .. Relative standard deviation , .Assay, per cent. 100.008 99.999 100.01 6 100*021 100.028 100~009 100.0 18 100.022 100.0 15 0.009 per cent. It is clear that the results show a positive bias, being 0.014 per cent. higher than in the mass titrimetric assay, which in one respect is fortunate in that it shows that cancellation of errors had not occurred: it is probable that several factors contribute to the bias. A bias in the measurement of the 2 A current was adumbrated earlier,l and it is possible that the uncertainty of &SO p.p.m. in the values of the two 1-R standard resistors used for measure- ment of high currents could account for one third of the bias. The possibility that the measured silver errors, whose constancy reflects the fact that the total times occupied by the acid determinations were similar in all instances, were low because of some mechanical loss of silver from the gauze electrode cannot be entirely discounted, although such a loss might be expected to show larger variations in the value of Qs.It is not improbable that anodic reactions other than halide deposition could occur, which could be significant particularly towards the end of the determination when the pH is rising rapidly. These reactions might include the anodic directions of iC AgOH + e + Ag + OH- i a 0, + 4H,0+ + 4e + 6H,O 0, + 2H,O + 4e + 40H- iC i a i o i,June, 19731 OF ACIDS I N CELLS WITHOUT LIQUID JUNCTION. PART IV 431 and, although the solution is oxygenated it is not purged, so that some molecular hydrogen in solution may give a small anodic current due to its re-oxidation- i c 2H30+ + 2e + H, + 2H,O i* It is also possible that 1 or 2 p.p.m.of the anodic current produces bromine from the bromide, and the bromine would hydrolyse at pH above 5, thus producing hydrogen ions. The drifting differential potentials near the end-point made the finish of the determina- tions rather tedious, so that a complete determination required about 5 hours, although this is much quicker than the titrimetric assay, which takes 3 to 5 days. That the drifting is caused by diffusion of hydroxyl ions from the interstices of the platinum gauze is in agreement with previous finding^.^ It also explains the difference in equilibration times between pre- titration and final end-point location, because the concentration of residual hydroxyl ions on the electrode would be expected to be much larger near the end of the reaction.A coiled platinum rod or perforated heavy gauge sheet would be better than gauze, and could reduce the experimental time by 90 minutes. The measured values of the large current were found to change only very slowly over the 35 to 40 minutes of generation, and the maximum deviation was only 20 p.p.m. The cell resistance was less than 0.5 Q initially and rose to 2 Q at the end of the high-current generation period: the temperature of the cell electrolyte rose by 5 to 6 "C from the initial 20 "C. CONCLUSIONS Results obtained in the assay of primary standard sulphamic acid show that a precision and accuracy in the region of 100 p.p.m. can be achieved with the simple cell and equipment described. Refinement of the method, by using better current measuring equipment, smooth platinum instead of gauze and purging of the electrolyte with 1 + 6 oxygen - nitrogen or oxygen - argon mixture, could enhance the precision and reduce the time required. Operation in the differential mode, with two cells in series, for the intercomparison of chemical standards would certainly yield improved precision, and would give the method even more marked advantages over the conventional methods with multicompartment cells. One of us (M.R.) is deeply indebted 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. 6. 6. 7. 8. 9. 10. REFERENCES Bishop, E., and Riley, M., Analyst, 1973, 98, 305. 8 , Ibid., 1973, 98, 313. , Ibid., 1973, 98, 416. -- 3 -- Analytical Methods Committee, Ibid., 1967, 92, 587. International Union of Pure and Applied Chemistry, Analytical Division, Pure AppZ. Chem., 1969, Bishop, E., and Short, G. D., Analyst, 1964, 89, 687. Short, G. D., and Bishop, E., Analyt. Chem., 1965, 37, 962. Bishop, E., and Short, G. D., AnaEyst, 1962, 87, 467. Eclrfeldt, E. L., and Shaffer, E. W., Analyt. Chem., 1965, 37, 1634. Taylor, B. N., Parker, W. H., and Langenberg, D. N., Rev. Mod. Phys., 1969, 41, 376. NOTE-References 1, 2 and 3 are to Parts I, I1 and 111, respectively, of this series. 18, 443. Received December 18th. 1972 Accepted January 19tk, 1973