Anal. Chem. 1999, 71, 1728-1732
Source of the Difference between the Measurement of Water in Hydrocarbons As Determined by the Volumetric and Coulometric Karl Fischer Methods Sam A. Margolis
Analytical Chemistry Division, National Institute of Standards and Technology, Gaithersburg, Maryland 20899
The disparity between the volumetric and coulometric Karl Fischer methods for the measurement of water in hydrocarbons and hydrocarbon mixtures (oils) has been resolved. The amount of water detected by the volumetric method when the titration vessel solution is homogeneous is higher than that measured coulometrically or volumetrically when the titration vessel solution is heterogeneous. Water in oil that is not measured by the coulometric method when the oil is incompletely dissolved has been shown to reside in the oil phase of the heterogeneous (multiphasic) coulometric vessel suspension. Water that is not measured volumetrically under relatively low chloroform concentrations after the point of transition to a heterogeneous titration vessel solution is reached also appears to reside in the nonaqueous phase of the titration vessel solution. It appears that this water is retained in the hydrocarbon phase in a manner such that it is completely inaccessible to the volumetric or coulometric Karl Fischer reagents. Two basic techniques, volumetric and coulometric, are most widely used for the measurement of water by the Karl Fischer method. Both techniques measure the end point amperometrically by detecting the amount of iodine present. The volumetric method approaches this iodine level by adding reagent containing iodine to the reaction vessel. The coulometric method regenerates the iodine electrochemically from iodide that is formed by the Karl Fischer reaction with water in the anodal compartment of the twocompartment reaction vessel or in the single-compartment vessel. Considerable disagreement exists within and between the coulometric and volumetric Karl Fischer methods for the measurement of water in hydrocarbons and hydrocarbon mixtures.1-5 Studies conducted for ASTM Committee D27 on the measurement of water in transformer oils show large variations in the results (1575%) between laboratories using different coulometric instru* Corresponding author: (tel) (301) 975-3137; (fax) (301) 977-0685; (e-mail)
[email protected]. (1) Margolis, S. Anal. Chem. 1995, 67, 4239-4246. (2) Margolis, S. Anal. Chem. 1997, 69, 4864-4871. (3) Margolis, S. Third International Symposium on Humidity and Moisture; National Physics Laboratory: Teddington, England, 1998; Vol. 2, pp 133(4) 140. Margolis, S. Anal. Chem. 1998, 70, 4264-4270. (5) Cedergren, A.; Lundstrom, M. Anal. Chem. 1997, 69, 4050-4055.
1728 Analytical Chemistry, Vol. 71, No. 9, May 1, 1999
ments.6 When oil is added to the coulometric vessel anode solution, a suspension is usually formed.3,4 We have documented that the observed water titer in oil samples decreases at the point that the oil becomes insoluble in the solvent in the titration vessel of the volumetric Karl Fischer instrument.1,3,4 This decrease in solubility is a function of the amount of Karl Fisher titrant and the amount of cosolvent, such as chloroform, that is added to the titration vessel solution. It is characterized by a transient reduction of the observed water titer to zero or close to zero at the point of transition from a homogeneous solution to a heterogeneous suspension.1,4 The measured water of the subsequent samples is almost always less than but never greater than the amount of water measured before the transition to a heterogeneous solution. The magnitude of this decrease appears to be a function of the water content of the oil, the composition of the oil, and the composition of the titrating reagent.3,4 In comparative studies, the amount of water measured by the volumetric Karl Fischer method after the solution has become heterogeneous is similar to the amount of water measured by the coulometric method.3,4 The difference in the amount of water measured volumetrically under homogeneous conditions and that measured coulometrically under heterogeneous conditions will be referred to as “missing water”. However, the disposition of the missing water is not clear. It could be one of two possibilities as a result of either (a) the retention of the water in the oil layer such that it cannot react with the Karl Fischer reagent or (b) the alteration of the mole ratio of the water to (I2) from 1:1 for the standard Karl Fischer reaction (eq 1) in protic solutions such as
H2O + I2 + [RNH]SO3CH3 + 2RN f [RNH]SO4CH3 + 2[RNH]I (1) R ) CH3(CH2)n methanol toward that of 2:1 for the Bunsen reaction (eq 2) in aprotic solvents such as dimethyl formamide7
2H2O + SO2 + I2 f H2SO4 + 2HI
(2)
This study demonstrates that the missing water is retained in the upper oil layer of the coulometric vessel solution after the two phases were permitted to separate. The water in both phases 10.1021/ac981137z Not subject to U.S. Copyright. Publ. 1999 Am. Chem. Soc.
Published on Web 03/19/1999
after they separated was measured under homogeneous conditions by the volumetric Karl Fischer method. EXPERIMENTAL SECTION The Hydranal reagents and the octanol (ACS grade, 99+% purity, mass fraction of water
