Journal on Processing and Energy in Agriculture 21 (2017) 1 17 Biblid: 1821-4487 (2017) 21; 1; p 17-22 Original Scientific Paper UDK: 615.32 Originalni naučni rad USE OF NIR SPECTROSCOPY AND 3D PRINCIPAL COMPONENT ANALYSIS FOR PARTICLE SIZE CONTROL OF DRIED MEDICINAL PLANTS UPOTREBA BLISKE INFRACRVENE SPEKTROSKOPIJE I 3D ANALIZE GLAVNIH KOMPONENTI ZA KONTROLU VELIČINE ČESTICA SUŠENOG LEKOVITOG BILJA Davor VALINGER * , Maja BENKOVIĆ * , Tamara JURINA * , Ana JURINJAK TUŠEK * , Ana BELŠČAK-CVITANOVIĆ ** , Jasenka GAJDOŠ KLJUSURIĆ * , Ingrid BAUMAN * * University of Zagreb, Department of Process Engineering, Faculty of Food Technology and Biotechnology, Croatia ** University of Zagreb, Department of Food Engineering, Faculty of Food Technology and Biotechnology, Croatia e-mail: mbenkovic@pbf.hr ABSTRACT The regular or in-line measurement of particle size is an essential requirement for quality control throughout food industry to ensure a consistent product output. Non-invasive spectroscopy in the near infrared range from 904 - 1699 nm (NIR) was applied to investigate the possibility of differentiation of five different medicinal plants (chamomile, dandelion, nettle, broadleaf plantain and yarrow). Dry plants were milled and after sieving four fractions were obtained with following particle sizes: <100 μm, 100-280 μm, 280-450 μm and >450 μm. Raw spectra and first derivative of raw spectra were recorded and principle component analysis (PCA) of first three factors was applied (3D PCA). 3D PCA showed better differentiation for first derivative compared to the raw spectra for all five medicinal plants and all analyzed particle sizes. Particle size fractions were photographed using microscope in order to explain analysis results for fractions 280-450 μm and >450 μm, where the presence of smaller particles was visible for all five medicinal plants. Key words: near infrared spectroscopy, principle component analysis, particle size, medicinal plants. REZIME Redovno ili linijsko merenje veličine čestica je ključni uslov za kontrolu kvaliteta u prehrambenoj industriji kako bi se osigurao konzistentan proizvod. Jedna od nedestruktivnih i najbržih analitičkih tehnika koja pruža uvid u hemijska i fizička svojstva ispitivanih materijala je bliska infracrvena spektroskopija. Neinvazivna spektroskopija u bliskom infracrvenom opsegu od 904 - 1699 nm (NIR) je primenjena da se ispita mogućnost diferencijacije pet različitih lekovitih biljaka (kamilica, maslačak, kopriva, stolisnik i trputac). Osušene biljke su samlevene i prosejane na odgovarajućem slogu sita pri čemu su dobijene 4 frakcije čestica sledećih veličina: <100 μm, 100-280 μm, 280-450 μm i >450 μm. Dobijene frakcije su snimane NIR instrumentom bez prethodnog tretmana. Na snimljene spektre i njihove prve derivacije primenjena je analiza glavnih komponenta (PCA) za prva tri faktora (3D PCA) korišćenjem programskih paketa StatSoft Statistica v.10 i Wolfram Research Mathematica v.10. 3D PCA je pokazala bolju diferencijaciju za prve derivacije nego za početne spektre za svih pet lekovitih biljaka i za sve analizirane veličine čestica. Različite frakcije čestica su fotografisane pomoću mikroskopa (Bresser LCD-35) kako bi objasnili rezultate analize za frakcije 280-450 μm i > 450 um, gde je bilo vidljivo prisustvo manjih čestica za svih pet lekovitih biljaka. Ključne reči: bliska infracrvena spektroskopija, analiza glavnih komponenata, veličina čestica, lekovito bilje. INTRODUCTION A significant number of literature reports have emerged over the past 20–30 years regarding the use of near infrared spectroscopy (NIRS) as both a research tool for formulation development and as a quality control technology for monitoring unit operations for product manufacturing (Ely et al., 2008). The demand for consistent and improved quality of food products requires development of instrumental methods for rapid determination of quality parameters. Currently, a great deal of interest is focused on the application of spectroscopic techniques such as nearinfrared (NIR) that enable non-destructive at-/on- line monitoring in the food and pharmaceutical industries. NIR spectroscopy measures molecular vibrations that, in contrast to midinfrared (MIR) radiation, can be transmitted through quartz- based optical fibers, enabling simple and flexible sampling on the process line. Previous studies have shown that NIR spectroscopy has great potential as an analytical technique for the determination of homogeneity of powder blends (Scotter, 1990; Wargo and Drennen, 1996; Muzzio, Robinson, Wightman and Brone, 1997; Luner, Majuru, Seyer and Kemper, 2000; El- Hagrasy, Morris, D’Amico, Lodder and Drennen, 2003) and quantification of lactose content (Kaffa, Norris, Kulscár and Draskovits, 1982; Baer, Frank, Loewenstein and Birth, 1983; Buckton, Yonemochi, Hammond and Moffat, 1998; Laporte and Paquin, 1999; Lane and Buckton, 2000; Tarkosova and