No abstract

The essential oils of various commercial citrus species were studied by ATR/FT-IR as well as NIR-FT Raman spectroscopy. Some of the most relevant monoterpene components occurring in the oil, such as limonene and γ-terpinene, show characteristic key bands in the IR and Raman spectra, which are useful for making a quantitative distinction between different oil types. Applying principal component analysis (PCA) to the spectral data, the individual citrus products are located, in most cases, as separated clusters in a 3-dimensional factor space. Cross-validation statistics showed that a partial least-square (PLS) algorithm was able to produce high quality predictions of the relative oil composition as well as general quality parameters. Both complementary spectroscopic techniques described in this study have the potential to replace existing analytical procedures because they are capable of faster quality control measurements and they allow on-line analyses during distillation or extraction processes to be performed.

NIR reflection and ATR-IR spectroscopy methods are developed to determine the echinacoside content in roots of Echinacea angustifolia and Echinacea pallida. Based on the recorded spectra and the HPLC reference data, chemometrical analyses are performed using a partial least squares (PLS) algorithm. Generally, good calibration statistics are obtained for the prediction of the echinacoside content presenting comparatively high coefficients of determination (R(2)) and low root mean standard errors of cross validation (RMSECV). It is demonstrated that optimal predictions are possible when using a dispersive spectrometer covering the spectral range from 1,100 to 2,500 nm. In contrast to the time-consuming HPLC method, the described non-destructive measurements allow us to predict the echinacoside content already after an analysis time of approx. one minute. Both spectroscopic techniques presented in this paper are shown to be useful in agricultural practice as well as in the phytopharmaceutical industry.

The aim of this study was to investigate the accuracy and transferability of near-infrared (NIR) calibrations for estimating the content and composition of the volatile fraction in fennel fruits (Foeniculum vulgare Miller) as an example of medicinal and spice plants. A master calibration with spectra obtained on a scanning monochromator was generated using 345 samples from three different harvests (1997-1999). A subset of 70 samples from 1999 was also measured on a dispersive grating and a scanning diode array system to gain an insight into the influence of sample presentation and scanning techniques. For all instruments, calibrations with standard errors in the range of the reference method were achieved. Furthermore the influence of storage on NIR spectra and, additionally, the potential of transferring spectra between both scanning monochromators was studied.

No abstract