Quantification of surface sensor representivity of primary crushed ore for bulk ore sorting

There are two fundamental methods for metalliferous sensing of primary crushed ore available for deployment at operations. These methods can be classed as surface measurement techniques (including XRF, LIBS, NIR, SWIR) or penetrative/semi-penetrative measurement techniques (including PGNAA, PFTNA, GAA, MR). While the individual sensing technologies exploit different physical phenomena to measure metal content, penetrative sensor measurements of grade are often regarded to be superior to surface measurement techniques due to the much larger sampling volume penetrative techniques provide. Surface sensing techniques often utilise simpler, established technologies that have considerable benefits in the speed of analysis, capital costs, ease of deployment, social license to operate and operational safety. This paper investigates whether surface sensing techniques achieve adequate sampling representativity of primary crushed rock mass to provide economic benefit over penetrative sensing technologies for bulk ore sorting applications. The paper details the methodology used to assess whether surface XRF measurements change throughout various stages of comminution, and increased surface exposure, from primary crushed ore (top size of -175mm) to finely crushed product for assay (top size -2mm). The methodology was applied to 2 ore mineralisation styles, across variable metals and grades, to identify any associated opportunities and limitations to surface sensing techniques. The results assist in developing a method to determine if surface sensing and measuring via XRF is an adequate representation of the rock mass for bulk ore sorting applications and identify any associated limitations in particle size distribution, mineralisation or natural deportment strength to guide suitable application of surface sensing techniques.