788 Arcalyst, December, 1968, Vol. 93, $q5. 788-791 Determination of Chloride in Chloridecontaining Materials with a Chloride Membrane Electrode BY J. C. VAN LOON (Department of Geology, University of Toronto, Toronto 5, Ontario) The determination of chloride by using the membrane-type chloride electrode is described. Procedures are given for the analysis of chloride materials, including silver halides, both in the presence and absence of bromide and iodide. RECENTLY a solid-state silver chloride type membrane electrode has been developed com- mercially for the measurement of chloride activities. This electrode, together with an expanded-scale pH meter, can be used for the rapid determination of chloride concentration in solutions of constant ionic strength. Only the presence of OH-, Br-, I-, S2-, CN-, NH, and S,0,2- in the solutions to be measured interfere directly with the results.1 The avail- ability of this system for rapid studies in the field and laboratory when large numbers of analyses are required makes it potentially useful.Before the commercial development of the silver chloride membrane electrode the use of silver - silver chloride electrodes for the determination of chloride has been recorded for boiler waters,2 chloride in ground water,5 free chloride in the presence of many cations,4 also sweat, urine and miscellaneous solutions.6 This type of electrode developed redox potential errors in the presence of strong oxidising substances, a limitation not shared by the new silver chloride membrane electrode. In this paper procedures are recorded for the rapid determination of chloride in soluble substances and insoluble silver halide materials, both when bromide and iodide are present and in their absence.EXPERIMENTAL APPARATUS- An Orion Specific Ion Chloride Electrode Model 94-17 was used, in conjunction with a Beckman Expandomatic pH meter. A scale expansion of 200-mV full scale was sufficient for these measurements. REAGENTS- All the water used was distilled and checked for the absence of chloride in amounts that would cause interference. Chromic acid sohtiolz-Dissolve 40 g of analytical-reagent grade chromic oxide in 400 ml of sulphuric acid (1 + 2). Standard sodium chloride solution, 0.1 M. Dilute working standards for analysis of sample solutions were prepared by combining all the ingredients used to prepare the sample but substituting an aliquot of standard solution for the chloride sample in the graduated flask.All reagents were tested for the presence of chloride contamination and found to contain insignificant amounts. Solutions were stored in polythene bottles. PROCEDURES FOR SOLUBLE MATERIALS- (a) No bromide or iodide present-Weigh an appropriate amount of sample (final chloride concentration 10-2 to 1 0 - 4 ~ ) into a 250-ml beaker and dissolve it in water. Rinse it com- pletely into a 1000-ml graduated flask and dilute to volume with water. Millivolt readings for these slowly stirred solutions were checked against the appropriate standards (meter was set to read +lorn0 mV for M chloride standard solutions), with the nearest standards being read before and after each sample.0 SAC and the author.VAN LOON 789 (b) Bromide and iodide @resent-Weigh an appropriate amount of sample (final chloride concentration 10-2 to l o - 4 ~ ) into a 250-ml Erlenmeyer flask. Wash the sample into the bottom of the flask with 5 ml of water and tilt it to allow the sample to run into one edge. Add 20 ml of the chromic acid solution and bubble nitrogen into the mixture with a fritted bubbler for 15 minutes. Rinse the bubbler inside and out and remove it from the flask. Wash the solution in the flask carefully into a 500-ml graduated flask and dilute to volume with water. Millivolt readings for these slowly stirred samples were checked against the appropriate standards (meter was set to read +lO.O mV for M chloride solutions), with the nearest standards being read before and after each sample.PROCEDURES FOR SILVER HALIDE- to 10-4 M) gently in 0.7 g of potassium carbonate and a small amount of spectrographic carbon in a platinum crucible until all of the silver, as the metal, has formed a coherent coating on the bottom of the crucible. Place the crucible upright in a 250-ml beaker and cover with 200 ml of water. Place a small stirring rod in the crucible and stir gently to complete dissolu- tion of the melt. Remove and rinse the crucible and stirring rod, then wash the solution from the beaker into a 500-ml graduated flask and dilute to volume with water. With standards prepared as indicated above and the meter set to read +lO.O mV for the M chloride standard solution, the mV readings for the samples were obtained on slowly stirred solutions.The nearest standards are read before and after each sample. (d) Bromide and iodide present-Fuse the sample gently as above, then cool the crucible and add 5 ml of water to the upright crucible on a magnetic stirrer. Stir with a small stirring rod until the dissolution is complete, then rinse the contents of the crucible with a minimum of water (4 to 5 ml) into a 250-ml Erlenmeyer flask. Add 30 ml of the chromic acid solution and support the flask in a tilted position so that the contents run to one side of the flask. Place a bubbler in the solution and bubble nitrogen through the solution for 15 minutes. Rinse the bubbler inside and out and wash the solution from the Erlenmeyer flask into a 500-ml graduated flask. Dilute to volume and measure the millivolt readings of the slowly stirred samples against appropriate standards with the meter set to read +lO.O mV for the M chloride standards.It is important not to allow any metallic silver to contact the concentrated chromic acid solution as silver is easily oxidised to Ag+ in this medium, with the resultant consumption of chloride. The most common causes of inconsistent values when the above methods are used can be summarised as follows: failure to allow sufficient time for the electrode reaction to reach a steady state; the presence of small particles of silver metal in the fusion, which sub- sequently are oxidised by the chromic acid (can be observed because of a resulting cloudy final solution); and failure to check higher and lower standards before and after each sample.RESULTS AND DISCUSSION METER NOISE AND DRIFT- Meter readings were found to fluctuate by * 5 mV, because of the movement of the hands around the chloride electrode. This was corrected by wrapping the electrode and two thirds of its body in three layers of aluminium foil and grounding the foil, somewhere on the lower body, to an outlet. The electrode response to changes in chIoride concentration was not instantaneous and often drifted towards the correct value over an interval of 60 seconds. Response times became even longer after long exposure of the electrode to many samples containing chromic acid solution. This problem was easily rectified by a light buffing of the membrane with fine emery paper.Slow response in the concentration range below 1 0 - 3 ~ could be, to some degree, overcome by stirring slowly with a thin stirring rod. REFERENCE ELECTRODE- It was, of course, impossible to use the normal standard calomel electrode directly in contact with chloride sample solutions, hence a small diameter calomel electrode was inserted in a larger diameter glass casing from a broken reference electrode. The casing with its fibre removed was plugged with a small amount of agar-agar - 4 ~ ammonium nitrate and then filled with 4 M ammonium nitrate. This electrode was clamped to the side of the chloride electrode with the Orion Specific Ion Electrode Holder. (c) No bromide or iodide +resent-Fuse a sample (final chloride concentration The nearest standards are read before and after each sample.790 VAN LOON: DETERMINATION OF CHLORIDE I N CHLORIDE- [Arta&St, VOl. 93 INTERFERENCES- Serious interferences with the use of this electrode would be caused by OH-, S2-, CN-, NH,, S,OS2-, Br- and I-.Of these, Br- and I- are the only ions often encountered in solutions resulting from chloride minerals. In addition, complexing cations, which remove free chloride from solution, could cause serious errors if not kept at an insignificantly low concentration by the use of dilute solutions. An agreement, within the level required by the accuracy of the method, of ionic strengths between samples and standards must be maintained in order to determine chloride concentration directly. The bromide and iodide interference can be eliminated by anion exchange,6 but this method was found to be too slow to be practical in this instance.Evans’ recommended the use of chromic acid in 8 to 9 N sulphuric acid for the separation of bromide from chloride in wide ranges of bromide-to-chloride ratios. It was necessary to prove that the chromic acid system worked for the expected removal of iodide and that it did not cause serious problems with the operation of the chloride electrode. Aliquots containing chloride (so that the final chloride content will be ~ O - * M ) and (or) bromide and (or) iodide, in 1 + 1 + 1 ratios, were treated according to the proposed pro- cedure, to determine firstly if chloride was stable during the oxidation, secondly whether bromide and iodide were eliminated quantitatively by the proposed procedure, and thirdly if chromic acid solution interfered with the operation of the chloride electrode.Tests were also carried out to see if the resultant C9+, produced during the oxidation, interfered in these solutions. Results for these experiments are listed in Table I, and from these it is obvious that the proposed procedure can be used without encountering an error of greater than 10 per cent. Although CI3+ was shown to interfere, its molar concentration is usually less than chloride and at this level does not cause serious error. Test TABLE I TESTS WITH CHROMIC ACID OXIDATION Maximum deviation Number of times from standard, performed mV Stability of chloride in presence of chromic acid mixture (no Br- or I- added) .. .... .. 4 Removal of I- by oxidation . . 3 oxidation . . .. .. . . 5 Removal of Br- and I- together by Possible interference by CP+ : MC1- - MCra+ (1 + 3) . . * . .. 1 MCl- - MCra+ (1 + 10) .. . . 1 MC1- - MCrS+ (1 + 29) .. .. 1 - 0.7 f0-8 & 1.0 + 0-7 +3-1 + 5.0 Maximum error (approximate), per cent. 3 6 10 3 15 23 TEST OF PROPOSED PROCEDURE- Many weighed samples of potassium chloride and sodium chloride, analytical-reagent grade salts, were subjected to the procedures (a) and (b), with aliquots of bromide and iodide solution being added to the sample to test the procedure (b). These results are given in Table 11. Several naturally occurring chloride minerals were analysed by using procedures (a) TABLE I1 ANALYSIS OF “KNOWN” SALTS Weight of samples Procedure taken, Sample used mg KCl .... a 100 NaCl .. .. a 100 b 50.0 b 50.0 d 11 to 22 AgCl .. .. c 18 to 21 Theoretical chloride, per cent. 47.7 47.7 60.7 60.7 24-7 24-7 Mean of chloride found, per cent. 46.5 46.6 60-6 60.1 24.0 23.8 Number of separate determinations 6 4 6 3 4 5 Standard deviation 1.18 1.71 0.56 1.27 0.22 ‘0.66December, 19681 CONTAINING MATERIALS WITH A CHLORIDE MEMBRANE ELECTRODE TABLE I11 ANALYSIS OF CHLORIDE MINERALS 791 Sample Halite (NaC1) . . .. Sylvite (KC1) . . .. Procedure used a b b a b b b U U Carnallite (KMgC1,.6H20) U b b Weight of sample, mg 49-0 47-0 49.0 48.5 50.4 51.8 49.1 48.7 51.3 50.7 50.6 49.0 Theoretical amount of chloride, m€! 29.7 28.5 29.7 29.4 24.0 24.6 23.4 23.2 24.5 19.4 19.2 18.8 Amount of chloride found, Deviation, mg mg 30-5 + 0-8 29-8 + 1.3 28.7 - 1.0 28.4 - 1.0 23-0 - 1.0 23.2 - 1.4 24.1 + 0.7 21.6 - 1.6 25-7 + 1.2 18.4 - 1.0 18.8 - 0.4 18.6 - 0.2 and (b) and the results are listed in Table 111.Several analyses with precipitated silver chloride and silver chloride jh?zls equal amounts of precipitated silver bromide or silver iodide, or both, were subjected to procedures (c) and (d) and the results are listed in Table 11. From these results it is obvious that the proposed procedures can be used for the rapid deter- mination of chloride in the presence of bromide and iodide when the greatest accuracy is not required. The time required for a single analysis is about 20 to 30 minutes with pro- cedures (b) and (a), but this can be lowered to about 8 minutes for each when many samples are analysed by integrating the steps. Procedures (a) and (c) required 2 to 5 minutes for completion. The author thanks the Department of University Affairs of the Province of Ontario for financial assistance for this work, and Dr. J. Mandarin0 of the Royal Ontario Museum for the samples. REFERENCES 1. 2. 3. 4. 6. 6. 7. Instruction Manual, Halide Electrodes, Orion Research Corp., Cambridge, Mass. Seidel, R. L.. U.S. Atomic Energy Corp. KAPL-M-3MS-93, 1958, p. 35. Black, W., J . Amer. Wal. Wks Ass., 1960, 52, 923. Onashi, H., and Morozumi, I., Hokkaido Doigaku Kogat ubu Kenkyu Hokoku, 1967,42, 131. Stem, M., Shwachman, H., Licht, T. S., and deBethune, A. J., Analyt. Chem., 1958, 30, 1506. De Geiso, R. C., Rieman. W., and Lindenbaum, S., Ibid., 1954, 26, 1840. Evans, B. S., Analyst, 1931, 56, 590. For mode 94-17. Received May 7th. 1968