Extent of inaccuracy in the hydrometer test

Abstract

Assessment of particle size distribution is a fundamental aspect of any soil characterization and is commonly used for soil classification. While it is relatively easy to classify the coarse fraction, classification of the soil fines remains problematic, particularly the fraction less than 2μm. Hydrometer analysis has been an international standard tool for fines assessment for many years but there may be serious flaws. A number of shortcomings facing the hydrometer have been alluded to by many researchers. Some of these have been addressed by laser scattering techniques. Others, including completeness of dispersion and the amount of clay carried down with coarser fractions, remain problematic.

Hydrometer analysis appears to underestimate clay content when compared to microscopic analysis and tends to overestimate clay fraction when compared to laser diffraction method. The analysis is based on Stoke’s law which estimates the rate at which particles settle in a suspension. The accuracy of hydrometer analysis depends critically on all of the clay present being dispersed into the suspension. Microscopic analysis, however, showed that, with the various types of dispersant specified in the hydrometer method, some samples experienced good dispersion while some were poorly dispersed.

Most of the commonly used methods of heave prediction in South Africa rely on an estimate of the clay fraction. Hydrometer analysis is not sufficiently reliable for a critical test such as heave prediction unless a means can be found of identifying which soils do not disperse effectively in the specified deflocculant.

Hydrometer and pipette analysis measures only the particle size distribution of the fine soil fraction, but some important aspects of soil behavior such as a change in volume depend on the clay mineral content. An assessment of clay mineral contents is therefore also required.

The necessity for a better method of estimating the clay fraction is illustrated by the many buildings and roads that have failed due to heave movements. A good example of this is a government subsidy housing site at Kimberley, where hydrometer analysis showed only 6% clay fraction. Van der Merwe’s method therefore predicted zero heave and houses were built on that assumption. Heave did in fact occur and one house became structurally unsound while still under construction, whilst many others became unsound shortly after construction.

This investigation assessed some aspects of the reliability of the hydrometer by separating the sand, silt and clay fraction of selected soil samples by sedimentation and the conducting tests on these fractions. Microscopic examination using a methylene blue staining technique was used to compare the composition of sediment layers with that expected from hydrometer theory.