Applicability of near infrared spectroscopy for real-time soil detection during automatic dishwashing

Abstract

The purpose of this study was to utilize NIR spectrometry to develop a novel method to detect and determine concentrations of different soils in dishwashing liquor during automatic dishwashing in real-time. If it is possible to differentiate between soils, this could be an opportunity to react specifically to them (e.g. by increasing the water temperature if fat components are not sufficiently emulsifying). The possibility of an automatic adaptation of the dishwashing process to different soils and soil levels could lead to a shorter, more environmentally friendly and cost-reducing process. In a first approach, an emulsion containing three soil types (oatmeal, egg-yolk and butterfat), water and detergent were used to develop NIR spectrometry prediction models. Transmittance spectra obtained with an Fourier transform near infrared (FT-NIR) spectrometer of testing standards of 76 automatic dishwashing cycles with seven samples per cycle were taken at various times during the main washing process for calibration (and validation) of the NIR spectrometry prediction models. The spectra were pretreated to develop NIR spectrometry prediction models for each type of soil using the partial least squares regression method with cross-validation. Overall, the coefficients of determination in cross-validation are R 2 > 0.92 for all NIR spectrometry prediction models developed. The results of the prediction models developed show that NIR spectrometry technology is a promising method to predict different levels of predefined soils in dishwashing liquor. The NIR spectrometry models were applied to an automatic dishwashing process with soiled dishes instead of emulsions containing soils to test their applicability. The resulting dishwashing process could be tracked in real-time by the dissolved soil concentrations, observed in the dishwashing liquor.

Developing and testing a new quantitative near infrared spectroscopy online tracking measuring system for soil detection during automatic dishwashing

“…The length of the groups of vertical dotted points also represents the change of homogenization during the dishwashing process caused by the circulation pump and temperature. 12…”

“…The new online tracking measurement system was used to develop a new calibration model from NIR spectra that was recorded with soiled dishes inside the dishwasher during the dishwashing process. The soil components BU (DFF Dairy Fine Food GmbH, Ratzeburg, Germany), OM (Peter K€ olln GmbH & Co. KGaA, Elmshorn Germany) and EY powder (Ovodan Eiprodukte GmbH & Co. KG, Zeven, Germany) were used, analogous to the work of Heidrich et al 12 The cleaning detergent used in the IEC 60,436 standard is "detergent D" (containing silicates, citrate, soda, oxygen-based bleach, nonionic surfactants, polycarboxylates, tetraacetylethylenediamine). 16 The soil components are more or less emulsified respectively dispersed in the dishwashing liquor together with water and the cleaning detergent.…”

“…The length of the groups of vertical dotted points also represents the change of homogenization during the dishwashing process caused by the circulation pump and temperature. 12 The dynamics at the beginning of the washing process are recorded exactly as they occur in the dishwasher. This is important because there is a large change in concentration at the beginning of the dishwashing process.…”

“…It has now become possible to detect soils in a quantitatively differentiated manner by means of near infrared (NIR) spectroscopy with subsequent chemometrics. Heidrich et al created the first NIR calibration model in the dishwasher by discontinuous sampling of the dishwashing liquor during a 65 C main wash cycle, 12 where, similar to the IEC standard, butterfat (BU), oatmeal (OM) and egg-yolk powder (EY) were chosen as representatives of fat-, carbohydrate-and protein-rich types of soiling, respectively, here called soil components. However, these measurements were previously carried out by taking samples every 10 min as "discontinuous tracking".…”