Comparison of rock density determination methods used in South African platinum mines for resource planning purposes

Abstract

Rock density is critical for determining the tonnage of an orebody and therefore impacts on the total resource of a deposit. Density is defined as the concentration of matter, and is expressed as mass per unit volume (g/cc; g/cm3 or t/m3). The density that is calculated will depend on the nature of the rock, and whether the volume calculated includes the open and/or closed pore volume of the rock. The pore volume will depend on the rock’s internal and external characteristics. This study looks at two methods commonly used to determine the rock density of samples taken from boreholes drilled for platinum mines on the North Eastern limb of the Bushveld Complex, South Africa. The first method is a gas pycnometer, which is almost exclusively used by laboratories. A Grabner Minidens air gas pycnometer was used. The second method is a hydrostatic immersion method, using water as the Archimedean fluid. An adapted Snowrex NH – 3 scale that can weigh a rock sample in air and in water was used. The first part of the study investigates the possible differences between conducting rock density measurements on finely milled core in the Grabner Minidens air gas pycnometer or on solid halved core samples using a hydrostatic immersion method, and the implications thereof. The second part of the study, not only investigates the differences between conducting density measurements on solid core samples or on milled core samples, but also looks at how the type of method used and how location affects the density measurement obtained. The location is important because changes in temperature and atmospheric pressure have been shown to produce small, but measurable changes in density. The density of pure water at 4 °C is approximately 1 g/cm3, increases or decreases in temperature will marginally decrease the density of water. The density of pure water at room temperature (21 °C) is 0.998 g/cm 3. Changes in atmospheric pressure have been shown to have a negligible effect on the density of most solids. The diamond drill core samples were taken from boreholes targeting the platinum group element (PGE) rich Merensky reef (MR) and Upper Group 2 (UG2) chromitite layer of the Upper Critical Zone. Samples were taken from the hangingwall (HW), reef and footwall (FW) of the MR and UG2. These rocks are made up of closely interlocking minerals, typical of cumulates. There are generally no visible pore spaces apart from highly fractured and altered samples. In part one of the study, 18,430 samples were used. The halved core samples were first measured using the hydrostatic immersion method at the exploration offices close to where the boreholes were drilled, referred to as the “Driekop” method. The samples were then sent to a laboratory in Johannesburg. Each sample was first milled to a fine powder (40 μm), and then a small portion of the milled sample (4 cm3) measured using the Grabner Minidens air gas pycnometer, referred to as the “Grabner Milled” method. For quality control, 811 of the remaining halved core samples were re-measured using the hydrostatic immersion method.The Grabner Milled results were found to be consistently higher than the Driekop results, with a mean average relative difference (AVRD) of approximately 5 % for all stratigraphic units. The difference observed can be accounted for, from the way in which the sample is prepared and the type of density that is measured. The Driekop method calculates the bulk density of the solid halved core sample, which includes all the open and closed pores of the rock. The Grabner Milled method calculates the true density of the finely milled sample, which through comminution, has excluded all open and closed pores that were in the rock. The quality control repeat measurements on the remaining halved core samples showed a good correlation with the original measurements, with a mean AVRD of only 0.33 %. In part two of the study, 82 randomly selected samples were used. The density of each solid sample was first determined using the hydrostatic immersion method. The same hydrostatic immersion method used in part one was applied at the same location; therefore it is also referred to as the “Driekop” method. The same hydrostatic method was then conducted on the samples at the laboratory in Johannesburg, referred to as “Lab water solid”. The gas pycnometer method was only conducted at the laboratory. The samples were first measured as a solid, referred to as “Grabner solid”. The samples were then milled to 40 μm and remeasured in the Grabner Minidens, referred to as “Grabner Milled”. The three solid methods results showed good correlation, with an average AVRD of only 0.01 % for the two hydrostatic immersion methods. On the other hand, there was a marked difference in results between the solid methods and the Grabner Milled method, the most significant difference being between the Grabner Milled and Grabner solid method (AVRD = 3.42 %). The resource model parameters for a project within the study area were used to illustrate the effect of density on resource planning. The average density used in the resource calculation will depend on what density method is used. The AVRD between the two methods for the mining cut density was approximately 5 %. The resource calculation showed that the difference in tonnage and 4E ounces between the two methods was also approximately 5 %. Changes in density result in equal changes in tonnage and metal content (4E ounces). Increases in dilution or overbreak from 10 to 30 cm above the optimal mining cut showed increases in tonnage and decreases in metal content. Due to similarities in rock composition between the HW, reef and FW of the MR, further dilution caused only a marginal decrease in density. The UG2 was found to be much more sensitive to dilution because of the distinct differences in rock composition between the reef, which is a chromitite layer and the HW and FW, which are both made up of plagioclase pyroxenite. Emphasis is commonly placed on the effect of dilution on grade; however this shows that the effect of density can be as important. The hydrostatic method of density determination is a very practical way of determining rock density at a remote exploration site. The whole sample can be measured and it is not restricted by the size or shape of the sample. Modern gas pycnometers have a higher degree of accuracy and precision, but need to be operated in a laboratory controlled environment, and are only capable of measuring a small amount of sample. With the correct application of quality control, both are suitable methods of density determination. The selection will depend on what type of density is required, the nature of the rock and whether the method must include or exclude pore spaces in the rock. Copyright