Palm sugar which is also named brown sugar is powdered sugar produced from palm
extract. Due to the high price of palm sugar, its contamination of materials that are cheap
or low quality is inevitable. Usually, adulteration detection is done by conventional
methods such as HPLC, TLC, or NMR which are time-consuming and require high-priced
equipment, thus impractical for routine and large sample analysis. The aim of this research
was to detect adulteration in palm sugar using Fourier Transform Infrared (FT-IR)
spectroscopy. The samples used in this study were palm sugar as the main ingredient and
coconut sugar as the adulterant. Two chemometric methods namely principal component
analysis (PCA) and partial least squares regression (PLSR) were used for analysis. The
absorbance data were taken at wavenumber 4000-650 cm-1
. Several concentrations of
coconut sugar as an adulterant ranging from 0 to 100% were added to palm sugar. A total
of 110 spectra of both pure and adulterated palm sugar samples were divided into two
groups, i.e. 73 samples for developing calibration model and 37 samples for developing
prediction model. The spectral obtained were pre-processed and analyzed using The
Unscrambler X version 10.4. a total of six pre-processing methods were used, i.e.,
Normalization, Standard Normal Variate (SNV), Multiplicative Scatter Correction (MSC),
and Baseline. Results showed that PCA was able to classify palm sugar based on
adulterant concentrations. PLSR calibration model with a coefficient of determination
(Rc2
) of 0.94 and root mean square error of calibration (RMSEC) of 8% was obtained by
applying the MSC method. The model was able to predict coconut sugar adulteration in
palm sugar with Rp2
of 0.89 and root mean square error of prediction (RMSEP) of
10.68%. The results confirmed the potential of FT-IR spectroscopy for detecting
adulteration in palm sugar.