[Thesis defence] 11/12/2025 – Moustapha Dieng: «Real-time monitoring of apple purée quality attributes during processing: physicochemical mechanisms and statistical modelling» (UMR SQPOV)

Mr Ahmadou Moustapha DIENG will publicly defend his thesis entitled: «Real-time monitoring of the quality attributes of apple purée during processing: physicochemical mechanisms and statistical modelling», supervised by Ms Sylvie BUREAU, on Thursday 11 December 2025.

Laboratory

UMR_A 408 SQPOV - Safety and Quality of Plant Products

Composition of the jury

Mrs Sylvie BUREAU	Avignon University	Thesis supervisor
Mr Jean-Roch Mouret	Montpellier SPO Unit	Rapporteur
Mr Pierre PICOUET	USC GRAPPE Angers	Rapporteur
Mrs Nathalie DUPUY	IMBE Marseille	Examiner
Mr Alexandre LECA	SQPOV Unit Avignon	Thesis co-supervisor
Mr Raphaël PLASSON	SQPOV Unit Avignon	Thesis co-director

Summary

This thesis aimed to identify markers of apple cooking during its transformation into purée in order to understand and model the product's physicochemical changes in real time. Quality criteria, viscosity, particle size, VOCs, colour and biochemical parameters were characterised at four key stages of cooking, and VIS-NIR spectra were acquired continuously every 10 seconds. A complete two-factor, two-level factorial design (temperature and grinding speed) was used to study the effects of these parameters on the composition and structure of the purée made from Golden Delicious apples. Fifteen VOCs from the alcohol, aldehyde, ester and ketone families were identified and quantified. Marked changes occur between 45°C and 95°C, with a significant loss of compounds at 95°C. In terms of structure, particle size and viscosity are also relevant markers of product evolution. Two modelling approaches were used: a phenomenological approach, based on Henry's law and ideal gas laws, was applied to VOCs and showed that chemical reactions predominated over volatilisation. The predictive approach to quality criteria based on VIS-NIR spectral data, using the SO-PLS method, made it possible to predict four parameters: viscosity, particle size and acetaldehyde concentration. Then, using the data set organised into six blocks, the SO-PLS method made it possible to predict viscosity (RPD=4.00; R²=0.89). This thesis is a preliminary work that paves the way for online and continuous control of processing methods. The acquisition of additional data on apples and other plant products over several years will contribute to the development of new perspectives for real-time control of food quality.

Keywords Malus domestica x Golden Delicious, processing, organoleptic quality, online measurement, visible and near-infrared spectroscopy, multivariate analysis, phenomenological modelling, PLS-R and SO-PLS prediction