Search

This paper proposes combining Multiple Linear Regression (MLR) and Digital Image-Based Flame Emission Spectrometry (DIB-FES) to overcome the inherent spectral interference in Flame Emission Spectrometry (FES) when sodium and calcium are present in the same sample. Metals in the air–butane flame emit radiation which is revealed in the digital images captured by a webcam (detector), on the basis of the Red-Green-Blue (RGB) color system. MLR calibration models for sodium and calcium were both developed and validated. In order to illustrate the feasibility of this proposal, the validated models were applied for the determination of sodium and calcium in powder milk. The results were compared with those obtained using Flame Atomic Absorption Spectrometry (FAAS) as the reference method. No statistically significant difference was observed between the results for both metals by using the paired t-test at a 95% confidence level. The accuracy of the models was also attested to by their recovery factors which fell between 97 and 103% for both metals. The results confirm the models' discriminating power in isolating each analyte's information from a single digital image, even with spectral interference.

In this study a flow-batch analyzer (FBA) was utilized and its feasibility was demonstrated for the determination of Fe(III) in edible and lubricating oils without external pretreatment. The FBA method uses the reaction of this ion with thiocyanate in organic medium. The formed red complex was monitored by employing a light emitting diode (LED, λmax = 520 nm) and a phototransistor linked to the mixing chamber by optical fibers. All standard solutions were prepared in-line and all analytical processes were carried out by simply changing the operational parameters in FBA control software. Comparing with the reference method, no statistically significant differences were observed when applying the paired t-test at a 95% confidence level. The relative standard deviation, analytical frequency and limit of detection were estimated for both viscous matrices (edible and lubricating oils) at