Sensor-based sorting has had a wide range of industrial use in automating and speeding up the process which requires substances or objects to be segregated from each other. The high demand for goods including raw materials, food, minerals, and waste recycling has increased the pressure for high-speed sorting. The first part of this paper presents a comprehensive theoretical and practical survey and comparison of current sorting methods relying on the use of the electromagnetic spectrum. Sorting methods are classified among other things, which portion of the EM spectrum is it, background noise rejection capacity, sample size limitations, sample chemistry limitations, sample surface cleanliness, spatial resolution capacity, spectral resolution, and feed rate limitations. The analysis focuses on color or visible light sorting, gamma-ray sorting, infra-red sorting, x-ray transmission-based sorting and x-ray fluorescence sorting, coupling the findings to the classification criteria outlined. We see a need to define a universal sorting scheme that will in general be applied to most sorting tasks. To do this, the final part of this paper re-looks at the x-ray transmission and x-ray fluorescence sorting scheme in line with the established limitations and proposes a dual x-ray transmission and fluorescence method to mitigate the challenge affecting the different schemes.

Modern-day diamond sorting is achieved through the application of X-ray luminescence (XRL) and X-ray transmission (XRT) techniques. Sorting with XRL is limited to the class range of 1.25 mm to 32 mm because of self-absorption associated with larger diamonds, greater than 32mm. The effect of self-absorption is also a high-energy phenomenon in XRL. XRT is limited to sorting large size diamonds as the technique suffers poor contrast for diamonds smaller than 10mm. XRT measurements are immune to self-absorption for all sample sizes, while XRL measurements have good contrast for particles smaller than 32mm. The applications of these techniques have hitherto been used independently of each other and have subsequently progressed mutually exclusively. Here we analytically show a new paradox of a dual-modality X-ray diamond sorting combining XRL and XRT techniques' strengths. Key features of our new paradoxical model performance are contrast mitigation for small particles and self-absorption rejection for a large particle at high energy as well as improved particle detectability and classification.

No abstract