Karl-Fischer Titrations Overview
C5 H5 N· 12 + C5 H5 N·SO2 + C5 H5 N + H2 O —>
2C5 H5 N· HI + C5 H5 N· SO3
There is a secondary reaction with the methanol (solvent):
C5 H5 N·SO3 + CH3 OH—>C5 H5 NH· SO4 CH3
It is noted that although this reaction is specific and quantitative, some difficulty may be caused by any admixed impurities which can react with the solvent to form water, interfere by redox reaction, or which bind iodine or react chemically with the other constituents. Usually these side reactions are much slower than the reaction with water, and although they might be discerned, such reactions cannot be eliminated. Altogether, the Karl Fischer titration technique is successful in most problems and its use has shown a steady growth over the past 40 years.
Various improvements of the original technique are also of long standing. Karl Fischer himself suggested that an electrometric endpoint detection could be used instead of the visual observation of the disappearance of the dark color of iodine. Almy and Griffin(3) used a potentiometric endpoint successfully. However, an even sharper endpoint detection was introduced by Wernimont and Hopkinson( 4 ) who employed the “dead-stop” method of polarized sensing electrodes.( 5 ) The various commercial instruments are all based on this method, by now considered classic.
However, in 1959, Meyer and Boyd demonstrated the possibility of coulometrically generating iodine in Karl Fischer titrations.( 6 ) This was a novel concept with great potential advantages. If the reagent mix contains iodide together with pyridine, sulfur dioxide and a solvent, the necessary iodine to complete the Karl Fischer reaction can be generated by passing electric current through the titration chamber. When the geometry of the cell and other conditions are properly chosen, 100% current efficiency can be maintained. When this is the case, according to the basic laws of coulmetry, 96,500 coulombs = 1 chemical equivalent (Faraday number). A Karl Fischer coulometer is then an absolute instrument, and the analysis requires no calibration or standardizaton. The coulometer can be governed by the classic dead-stop electrode. As the latter governs the reaction back to the original set-point, sample after sample can be titrated in the same solution. This system then can be essentially all-electronic, obviating the need for handling liquids, burettes, etc.; this is of particular advantage in this case because the Karl Fischer reagent is noxious and somewhat toxic. Additionally, an electronic system lends itself to automation. These advantages were recognized and a number of coulometric Karl Fischer titrators were employed in research.( 7 ) The first demonstration of such a system in routine analysis was by Shaw and Goode.( 8 ) Encouraged by this background information the first commercial instrument, the AQUATEST,( 9 ) was produced by Photovolt. It had a typical accuracy of 1% and sensitivity of about 20 micrograms of water. What made it particularly attractive as an industrial quality control instrument was that it read directly in micrograms of water and that the only control was a “Start” switch to initiate the titration. One vexing problem was keeping the reagent free from deterioration by ambient humidity; thus this model was designed with a completely closed titration vessel. However, whenever positive or negative pressure developed, the liquid was pumped from the cathode to the anode, or vice versa. It was then found that the vessel solution, which lacks iodine, is not nearly as hygroscopic as the usual Karl Fischer reagent. Consequently, a new version, the AQUATEST II, was built with a pressure-equalizing vent. The cathode chamber and configuration was also improved and the instrument exhibited superior performance.( 10 ) The AQUATEST II enjoys widespread use and a number of papers have been published on its performance in the field.( 11 ) It is also accepted in an official test method.( 12 ).
The AQUATEST IV was a newer version of these instruments. It is the result of a further study( 13 ) in which the limiting factors were scrutinized. Changes include an asymmetric sensing electrode, a cover fastened by a clamp which reduces still further the seepage of ambient humidity, and a stepping motor that provides more reliable performance. Together, these changes tend to reduce electrical noise.
The AQUATEST IV incorporated a microprocessor which offers the advantages pointed out above. The most important is, perhaps, that it serves to distinguish between the titration of water in the sample and water which is slowly generated by parasitic reactions, ambient humidity which seeps in, or any other effects which are oxidizing (or reducing) in character. All these effects, conveniently called “drift”, are determined subsequent to the titration proper and the result of the titration is automatically corrected. The microprocessor also performs the arithmetic operations. There are no controls other than the keyboard, which incorporates all the functions.
The latest version of the AQUATEST is the AQ1010.
The operation is based on the principles of all AQUATEST instruments. When the instrument is started and a sample is added to the vessel solution, a voltage arises across the polarized sensing electrode which indicates a “wet” condition. This triggers the coulometer and a constant current flow from the anode through the frit, which separates the vessel solution (anolyte) from the generator solution (catholyte), to the cathode. In consequence, iodine is developed at the anode by oxidation of iodide. The iodine completes the Karl Fischer reagent and is mixed by stirring throughout the vessel. When all the water has reacted, the voltage at the sensing electrode drops. This signals the coulometer to stop, and the electrical charge integrated during the titration process is stored in memory. During the subsequent time period of approximately one-minute, data is taken to establish current requirement, both positive and negative, to maintain the solution at equivalence. These data are then used to correct the initial titration value to reflect the net value due only to the water content of the sample without the background “drift”.
1. Mitchell, J. Jr.; and Smith, D.M., Aquametry John Wiley & Sons, New York, Second Edition 1977
2. Fischer, K., Angew, Chem. 48, 394 (1935)
3. Almy, E.G., Griffin, W.E., and Wilcox, C.S., Ind. Eng. Chem., Anal. Ed. 12, 392 (1940)
4. Wernimont, G., and Hopkinson F.J., ibid. 15, 272 (1943)
5. Foulk, C.W., and Bawden, A.T., J.Am. Chem. Soc. 48, 2045 (1926)
6. Meyer, A.S. Jr., and Boyd, C.M., Anal. Chem. 31, 215 (1959)
7. Swensen, R.F., and Keyworth, D.A., ibid 35, 863 (1963) also Pribyl, M., and Slovak, Z., Mikrochim. Acta, 6, 109 (1964), Lindbeck, M.R., and Freund, H., Anal. Chem. 37, 1647 (1965), Rechnitz, G.A., and Srinivasan, K.Z. Anal. Chem. 210, 9 (1965), Bizot, J., Bull. Soc. Chim. (1967), Vol 1, 151, Karlsson, R., and Karrman, K.J., Talanta, 18 459 (1971)
8. Shaw, R.G., and Goode, J.V., private communication
9. Seltzer, D.M., and Levy, G.B., Am. Lab. 1, (9), 30 (1969)
10. Levy, G.B., and Seltzer, D.M., Recent Analytical Developments in the Petroleum Industry, Hodges, D.R., Ed. Chapter XV, Applied Science Publishers, Essex England (1974)
11. Johnson, C.R., Power, March 1971 also Gedener, T.J., and Frey, R. Amer.Lab 7 (3) 47, (10) 43 (1975), 8 (7) 41 (1976), Draper, R., Bakers Dig., June 1977, p. 26
12. ANSI/ASTM D1533-77 13. Lindblom, K.A., et al. Paper #260 Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, February 1978.
13. Lindblom, K.A., et al. Paper #260 Pittsburgh Conference on Analytical Che