Rapid monitoring and quantification of unripe banana flour adulteration using visible-near infrared spectroscopy.
Loading...
Date
2021
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
A general lack of strict regulations in South Africa to monitor processed foodstuff increases
chances of unfair producers and traders to intentionally mislabel and adulterate high valued food
products with inferior lookalikes. Recently, unripe banana flour (UBF) has gained global
attention and has been identified as a replacement for cereals flours due to its gluten free traits
and resistant starch nutritional qualities, yet has no quality control standards. The objective of
this research was to develop rapid prediction models based on a visible to near infrared (Vis-
NIR) spectroscopy (Vis-NIRS) combined with multivariate analysis to classify, detect, and
quantify different adulteration levels of staple flours (i.e. wheat and maize flours) in unripe
banana flour. The other aim was to identify important biomarkers of unripe banana flour that
could be used to discriminate unripe banana flour adulteration at different concentration levels.
A critical evaluation of the portable Vis-NIR spectroscopy combined with chemometrics
analysis indicated that it was possible to discriminate between unripe banana flour with wheat
and maize flours and associated different adulteration levels. The partial least square (PLS)
regression (PLSR) analysis quantified individually maize and wheat flours, based on different
adulteration levels, showed that optimal PLSR detection models performances were obtained
using the first derivative Savitsky-Golay (7-point smoothing, 2nd order polynomial) and the
second derivative Savitsky-Golay (19-point smoothing, 2nd order polynomial). The study to
optimise and test the handheld Vis-NIR instruments’ feasibilty to simulteniously develop a
standard model for rapid solution to detect both maize and wheat flours adulteration indicated
high classification and prediction accuracies could be achived through principal component
analysis (PCA) and partial least squeres regression (PLSR). The study found that gluten could
be utilised as a biomarker to test for unwanted adulteration of unripe banana flour with wheat
flour, and showed good and reliable rapid spectroscopic PLSR model was achieved with high
precision. Near infrared spectroscopy showed great potential to detect the nutritional changes
of unripe banana flour during adulteration based on resistant starch content. The results of this
investigation indicated that wheat adulteration is a threat to unripe banana flour importnt
attribute as signification reduction of this parameter was observed with the increasing levels
wheat adultearation. Vis-NIR spectroscopy with multivariate analysis detected the varying
resistant starch concentration unripe banana flour samples successfully with high accuracy. The
results and stability of the models developed in this study demonstrated clearly that the Vis-
NIRS method has a potential of providing unripe banana flour processing industry with a rapid
and non-destructive technique to manage unripe banana flour quality as well as adulteration by
staple flours, therefore ensuring fair and safe trading of the product in retail markets of South
Africa.
Description
Doctoral Degree. University of KwaZulu-Natal, Pietermaritzburg.