Measurement of elemental sulfur in soil and sediments - Field sampling, sample storage, pretreatment, extraction and analysis by high performance liquid chromatography

Abstract

A rapid method for measuring elemental sulfur (S0) in soil and sediments is described that is accurate and precise down to a detection limit of 0.1 mg kg-l. Elemental sulfur is extracted into chloroform and measured by high-performance liquid chromatography. For moist samples, water is added to exceed their liquid limit. The method has been used to follow oxidation of both natural and fertilizer So in soil in the laboratory and field. Sediments analysed included some from lakes, and settling ponds containing coal mine wastes. Extraction within the plastic range was incomplete and recoveries as low as 5%-65% were found for five soils selected for a wide range of plastic limit. The plastic soil aggregates were not dispersed in the water-immiscible chloroform, preventing the dissolution of occluded S0; but at moisture levels above and below the plastic range, the soil was dispersed and extraction of S0 was quantitative. Extraction into a water-miscible solvent, chloroform-methanol (50 : 50), dispersed the soil at all moistures by dissolving the water thus preventing formation of plastic soil. This was less convenient because either volume correction for moisture was needed, or excess water added later to separate the chloroform. Measurement of total sulfur (S) in chloroform by inductively coupled plasma gave comparable results where samples contained S0 fertilizer, but at the lower natural levels they were much less precise and sometimes higher because of the additional S compounds extracted. Storage of moist soil below -10°C or air-dry soil at 20°C, gave no loss of S0 after one month. Moist soil at 20°C gave 40% loss after one week, but at 4°C the loss was only 5% after four weeks. Forced air-drying at 30°C, freeze-drying and microwave drying (under a narrow range of conditions) gave losses of about 5% for fertilizer S0 (<0.15 mm fraction), and of about 20% for smaller S0 crystals. In field trials with S0, sample variability was associated mostly with the difficulty of applying the S0 uniformly and with the small number of particles applied per unit area. For particles <0.25 mm, variability from mixing of cores and sub-sampling was not so great.