Volume 57, Number 1, 2003 APPLIED SPECTROSCOPY 23 0003-7028 / 03 / 5701-0023$2.00 / 0 q 2003 Society for Applied Spectroscopy New Method for Pollen Identication by FT-IR Spectroscopy C. S. PAPPAS, P. A. TARANTILIS, P. C. HARIZANIS, and M. G. POLISSIOU * Laboratory of Chemistry, Department of Science (C.S.P., P.A.T., M.G.P.) and Laboratory of Sericulture and Apiculture, Department of Crop Science (P.C.H.), Agricultural University of Athens, 75 Iera Odos Street, 11855 Athens, Greece A new methodology for identication of pollen was developed based on FT-IR spectroscopy. Pollen samples of twenty different plant species were collected and the diffuse reectance infrared Fourier transform (DRIFTS) and KBr pellet spectra were recorded. Li- braries of spectra were created. Spectra of unknown plant origin pollen were recorded and compared with those of the corresponding pollen library and the match value was measured automatically using the appropriate software (OMINC ver. 3.1). From the same pollen samples, microscopic slides were prepared and the photo- graphs of the pollen grains were used as a second comparison meth- od. Using light microscopy, the pollen identication is usually lim- ited to the family or generic name, while FT-IR spectroscopy can distinguish species belonging to the same genus. This method is sim- ple and fast, and when the DRIFTS technique is used the sample is not destroyed. Index Headings: Pollen; Microscopic; FT-IR; Identication. INTRODUCTION Palynology is the study of pollen grains and spores produced by plants. Examination of pollen grains is of great importance for many scientic studies such as plant taxonomy, honey studies (melissopalynology), genetics and evolution, forensic science, allergy studies, tracing vegetation history, and paleontology. A lot of knowledge has been gained about pollen grains through morpholog- ical and chemical studies. 1–7 It has been demonstrated that pollen grains can be clas- sied by automating pollen counting through images ob- tained from slides. A standard data set of pollen images is required, and in the prepared slide the pollen grains must be well separated. Also, the depth of focus of the microscope dramatically affects the focal plane of the camera that captures the image. 8 Using light microscopy, the identication of pollen grains is sometimes limited to the family or generic level. 9 Pollen can be electrodynamically trapped directly from the atmosphere and analyzed by Raman spectroscopy. The Raman spectra provide chemical rather than mor- phological information and can be used to distinguish different pollens in the 1600–400 cm 21 region. 10 Many papers have been published using Fourier transform infrared (FT-IR) spectroscopy to study plant materials, but not pollen grains. FT-IR spectroscopy is a method suitable for monitoring chemical changes in plants. 11–14 This paper describes the development of a method for pollen identication from different plants in a simple and fast manner without destroying the Received 2 June 2002; accepted 4 September 2002. * Author to whom correspondence should be sent. sample. It is also able to distinguish different plant spe- cies of the same genus. MATERIALS AND METHODS Pollen Samples. Three pollen samples were collect- ed from each of the following twenty honey plants: lemon ( Citrus limon ), sour orange ( Citrus aurantium ), tangerine (Citrus deliciosa ), orange ( Citrus sinensis ), saffron ( Crocus sativus ), pine ( Pinus halepensis ), chamomile ( Chamomilla recutita ), poppy ( Papaver rhoeas ), sage-leaved ( Cistus salvifolius ), pyracantha ( Pyracantha coccinea ), heather ( Erica manipuliora ), eucalyptus ( Eucalyptus sp. ), thyme ( Corydothymus capitatus ), pink cistus ( Cistus incanus ), yellow sweet clover (Melilotus ofcinalis ), crepis ( Crepis foetida ), chestnut ( Castanea sativa ), cotton ( Gossypium hirsu- tum ), corn (Zea mays ), and sunower ( Helianthus an- nuus ). The pollen samples were collected from differ- ent regions of Greece. They were dried at 40 8C in an oven for 8 h and kept at 220 8C until use. One sample from each of the above plants was used as a standard. The other two were titled as unknown. Microscopic Analysis. From the standard pollen samples, pollen grains of four Citrus species ( C. au- rantium, C. deliciosa, C. limon, and C. sinensis ) were xed on glass slides for microscopic analysis. Several anthers were placed on a clean glass slide and the pol- len grains were separated from the anthers by adding 2–3 drops of diethyl ethyl and pressing them with a glass rod. After evaporation of the ether (1–2 min) and removal of the empty anthers, 1–2 drops of 10% iso- glucose (high fructose corn syrup) solution was added; then the sample was dried on a hot plate at a temper- ature of 50–60 8C for about 5 min. The pollen samples were xed by adding 2–3 drops of Entellen t and then replacing the cover slip. Photographs were taken with a light microscope at 3330. FT-IR Spectroscopy. FT-IR spectra were obtained in the diffuse reectance infrared Fourier transform spectroscopy (DRIFTS) mode using the micro sampling cup of a Spectra- Tech diffuse reectance accessory (2 mg of dried pollen) and in a pellet pressed with KBr (2 mg dried pollen and 148 mg dried KBr) against a KBr background on a Nicolet Magna 750 FT-IR spectrometer (DTGS detector, Nichrome source; beamsplitter: KBr), with a total of 100 scans (res- olution, 4 cm 21 ). Spectra were collected and manipulated using the OMNIC (ver. 3.1) software supplied from the manufacturer of the spectrometer. Spectra of each sample were collected in triplicate. All spectra were smoothed using the ‘‘automatic smooth’’ function of the above software, which uses the