A knowledge of the water content of petroleum products is important in refining, purchase and sale, and transfer of products, and is useful in predicting the quality and performance characteristics of the products.
There are numerous methods for the determination of water in petroleum products; many specific to a particular product line. Analytically the methods can be grouped as by distillation or by Karl Fischer titrations. Often the choice depends on the levels of water present in the sample. Generally, trace amounts of water must be determined by coulometric Karl Fischer methods. The Karl Fischer titration methods have been in use for over 60 years. They are extremely well documented methods with method interferences well characterized. Generally, these methods work well in "clean" matrices; but there are many potential chemical and physical problems with "real" and complex samples. ASTM Interlaboratory Cross Check Program conducted under the auspices of Coordinating Subcommittee D02.92 shows very poor reproducibility for lubricating oil and lube adpack crosschecks. Almost all participating laboratories in these programs show either high or low bias. In addition to the D02 sponsored Karl Fischer methods, there are several more methods under the jurisdiction of other ASTM committees. However, many of those are also applicable to the analysis of petroleum products. These methods consist of either manual, potentiometric, or coulometric titrations.
Interferences in Karl Fischer titrations could be of three types:
(1) Physical - Electrodes fouling; residual moisture in glassware and reagents; contamination of samples; separation of water layer in storage.
(2) Chemical - A number of substances and classes of compounds interfere by condensation and redox reactions. Many of these can be eliminated by adding specific reagents prior to titration. The interfering chemicals include: aldehydes, amines, ammonia, ketones, halogens, mercaptans, metallic O/OH compounds, oxidizers, sulfides.
(3) Other - The reagents may themselves react with certain species in the sample. Different brands of pyridine-free reagents have been found to give different results with different chemical samples. The discrepancy appears to be more pronounced in aromatic hydrocarbons than in nonaromatic hydrocarbons. Some of these interferences, in addition to be able to be suppressed by addition of certain reagents, can also be overcome by using vaporizer accessory which physically removes water from the heated matrix as water vapor and then titrates it coulometrically.
Tables below compare the main characteristics of alternative methods. This table includes some of the non-D02 committee test methods also, since they are relevant to the knowledge of this important technique.
LLP = Liquid Petroleum Products.
Interferences normally associated with KF titrations such as mercaptans, sulfides, ketones, aldehydes, etc.
_=Underlined methods of D2 origin or interest.