This paper presents the assessment of three spectroscopic techniques, i.e., Raman spectroscopy (RS), ultraviolet fluorescence (UVF) spectroscopy, and energy-dispersive X-ray diffraction (ED-XRD), for measuring calcite (CaCO3) concentration in copper ore. The results for the three methods are compared in terms of their correlation coefficient R2 and prediction error (root-mean-square error of prediction, RMSEP). These approaches substantially shorten the measurement time with respect to current laboratory procedures from 300 min down to 1,15, and 9 min for RS, ED-XRD, and UVF, respectively, while preserving a good accuracy level (RMSEP <; 0.577). Prediction models were calibrated and validated for the three techniques, using the correlation coefficients Rc2 and Rv2, the error of prediction (RMSEP), and the error of calibration (root-mean-square error of calibration, RMSEC). Rv2 values of 0.90,0.88, and 0.94 were obtained for the RS, UVF, and ED-XRD models, respectively. The calibration and validation were performed for calcite concentrations ranging from 2.9% to 10.7%, and although the ED-XRD model performed slightly better than its RS counterpart, the time taken by the latter is significantly shorter than that by the other two, representing a good candidate for online estimation of calcite concentration.
Assessment of Three Spectroscopic Techniques for Rapid Estimation of Calcite in Copper Ore
2010
Abstract
This paper presents the assessment of three spectroscopic techniques, i.e., Raman spectroscopy (RS), ultraviolet fluorescence (UVF) spectroscopy, and energy-dispersive X-ray diffraction (ED-XRD), for measuring calcite (CaCO3) concentration in copper ore. The results for the three methods are compared in terms of their correlation coefficient R2 and prediction error (root-mean-square error of prediction, RMSEP). These approaches substantially shorten the measurement time with respect to current laboratory procedures from 300 min down to 1,15, and 9 min for RS, ED-XRD, and UVF, respectively, while preserving a good accuracy level (RMSEP <; 0.577). Prediction models were calibrated and validated for the three techniques, using the correlation coefficients Rc2 and Rv2, the error of prediction (RMSEP), and the error of calibration (root-mean-square error of calibration, RMSEC). Rv2 values of 0.90,0.88, and 0.94 were obtained for the RS, UVF, and ED-XRD models, respectively. The calibration and validation were performed for calcite concentrations ranging from 2.9% to 10.7%, and although the ED-XRD model performed slightly better than its RS counterpart, the time taken by the latter is significantly shorter than that by the other two, representing a good candidate for online estimation of calcite concentration.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
