Articles
Mid-infrared spectroscopy, a routine tool for improving the quality of processed tomato products
Article number
1445_7
Pages
43 – 50
Language
English
Abstract
The aim of this study is to use the mid-infrared spectroscopy (MIRS, spectrometer Bruker ALPHA II) in real conditions meaning routine analyses in a technical institute to control quality of raw tomatoes on arrival at the plant and monitor the quality changes during processing.
A diversity of processing tomatoes was selected in the South-East and South-West of France in 2022 (36 cultivars) and 2023 (38 cultivars), representing respectively 93 and 152 samples.
For each sample, a part of the fruit was characterized as raw material (15 fruits) for soluble solids content (°Brix), titratable acidity (mmol H+ 100 g‑1) and dry matter (%), while the other part was processed using either hot break (HB) or cold break (CB) conditions, and then characterized for their viscosity (Bostwick value). All samples were also systematically analyzed using MIRS (4000-600 cm‑1). For the raw material, the prediction models by partial least squares regression (PLSR) were good for soluble solids content (R2=0.90, RMSEP=3.9%) in external validation, for titratable acidity (R2=0.90, RMSEP=4.2%) and for dry matter (R2=0.83, RMSEP=4.2%). These results are very closed to those obtained in laboratory conditions.
Concerning viscosity, the best prediction was obtained (R2=0.89, RMSEP=11.6%) by the model built with all HB and CB purees.
These results point out the possibility of using MIRS as a professional tool to routinely control raw product quality, monitoring product processing and controlling output quality.
A diversity of processing tomatoes was selected in the South-East and South-West of France in 2022 (36 cultivars) and 2023 (38 cultivars), representing respectively 93 and 152 samples.
For each sample, a part of the fruit was characterized as raw material (15 fruits) for soluble solids content (°Brix), titratable acidity (mmol H+ 100 g‑1) and dry matter (%), while the other part was processed using either hot break (HB) or cold break (CB) conditions, and then characterized for their viscosity (Bostwick value). All samples were also systematically analyzed using MIRS (4000-600 cm‑1). For the raw material, the prediction models by partial least squares regression (PLSR) were good for soluble solids content (R2=0.90, RMSEP=3.9%) in external validation, for titratable acidity (R2=0.90, RMSEP=4.2%) and for dry matter (R2=0.83, RMSEP=4.2%). These results are very closed to those obtained in laboratory conditions.
Concerning viscosity, the best prediction was obtained (R2=0.89, RMSEP=11.6%) by the model built with all HB and CB purees.
These results point out the possibility of using MIRS as a professional tool to routinely control raw product quality, monitoring product processing and controlling output quality.
Authors
M. Clerc-Pithon, L. Lanoë, R. Giovinazzo, C. Boudot, D. Page, S. Bureau
Keywords
industrial tomato, processing, quality, MIRS, modeling
Online Articles (49)