USE OF NEAR INFRARED SPECTROSCOPY (NIRS) FOR MONITORING AND PREDICTION OF BIOCHEMICAL METHANE POTENTIAL (BMP) FROM ARUNDO DONAX: PRELIMINARY RESULTS

Introduction The global energy demand is growing rapidly, while the traditional fossil energy sources are progressively decreasing. Anaerobic Digestion (AD) is a well-established technology to achieve energy from non-edible/low-input crop biomasses like Arundo donax. Biochemical Methane Potential (BMP), an assay to evaluate the AD energetic potential of a substrate, is time consuming and expensive, indicating the need of new fast analytical tools. Here, we were aimed to test weather Near Infrared Spectroscopy (NIR) can be used for fast prediction and monitoring of energetic potential of A. donax. Materials and Methods A. donax is a fast growing perennial cane belonging to Gramineae growing in damp soils. It has gained interest in the last decade as a potential biomass source because it is characterized by a high relative growing rate, high production yields, low input required and deep roots. We compared the BMP of four different biomasses: a) A. donax from a river-side wetland where it was growing spontaneously; b) A. donax cultivated in North Italy; c) maize silage, which is now frequently used as fermenting biomass; d) cheese whey from a milk processing in a dairy farm in North Italy. Maize silage and A. donax samples were chopped into pieces of 0.2 mm with a grinder. Digestate from the anaerobic digester of wastewater treatment plant was used as inoculum for the experimental trials. The inoculum sample was sieved through a 0.2 mm sieve and pre-incubated in anaerobic conditions in order to deplete the residual biodegradable organic content. This was done by storing the digestate for 5 days at 37+1°C. For the BMP tests, the AMPTS II system (Bioprocess Control - Sweden) was used. It is a multi-batch lab-scale equipment, which allows on-line measurements of ultra-low bio-methane flows produced during the AD process. The batch digesters (0.5 L working volume) are lodged in temperature-controlled chambers and are connected to a gas-volume measuring device. To prevent the formation of dry and inactive floating layer, all batch digesters were continuously stirred. All BMP values were expressed in NmL CH4 g-1 of fed Volatile Solids (VSfed). Trials were carried out at 37+1°C, until no methane production was detected, for a period of about 30 days. Mixtures of biomass and inoculum were prepared at 1:2 ratio, based on VS content. Control trial (inoculum only) was also carried out in triplicate. The biogas production of control batches was subtracted from the biogas production obtained in the substrate tests. Before the BMP tests, all substrates were analyzed for pH, total solids (TS), VS, total Kjeldahl nitrogen (TKN) and neutral detergent fiber (NDF). TS and VS were measured according to the Standard Methods procedures (APHA, 1998). TKN was determined by Kjeldahl instrument after total mineralization and NDF was evaluated by Van Soest methods (Van Soest et al., 1991). NIR spectra were collected on all samples taken at different times from each batch reactor during the BMP trials (126 scans, 4 time for each sample, every 4 cm-1 for a total of 1501 points; resolution 8 cm-1) in the whole NIR range (4000-10000 cm-1) by a NIR FT-NIR NIRFlex® N-500 (Buchi Italia, Assago, Milan) equipped by fiber optics. Spectra were then processed by using NIRCal v. 5.21 software (BUCHI Italia). The acquired spectra were averaged and converted to absorbance as spectra pre-treatment. Models for prediction of BMP were developed using PLS regression and validated by cross-validation. Prior to PLS regression, all data were normalized. PLS regression was carried out using The Unscrambler X software version 10 (Camo inc., Oslo, Norway). Results and Discussion The main results from this research are summarized as follows: a) evaluation of biochemical methane potential of A. donax, substrate characterized by great yield reached in low input cultivation conditions; b) evaluation of the possibility to generate a model to predict the bio-methane potential of the tested substrates through the correlation between the data obtained from chemical characterization, NIR spectroscopy and BMP trials; c) development of a system for on-line monitoring of fermentation processes through the use of NIR methodology. The quality of the obtained models, in term of model robustness, assessed through rigorous criteria such as the prediction error on independent validation set, and the reproducibility of results due to the confirmation on an adequate number of samples analyzed in parallel with the reference technique (BMP) are considered as indicators of results. Conclusion The suitability of this approach to obtain a real-time evaluation of how the anaerobic digestion process is going on and to predict the BMP of A. donax as good alternative as feedstock for anaerobic digestion process is evaluated in order to demonstrate the NIR applicability in this field. Novelty statement Few studies are available about the energetic yield of A. donax through AD and about the possibility to use NIR spectra for BMP prediction. Fast and simultaneous measurements through NIR can contribute to the evaluation of the process versus complex and time-consuming BMP test. Summary Anaerobic Digestion (AD) is a well-established technology to achieve energy from non-edible crop biomasses like A. donax and Near Infrared Spectroscopy (NIR) may be a valuable alternative to the time consuming and expensive Biochemical Methane Potential (BMP) test for the evaluation of the energetic yield of this substrate. Thus our research was aimed to evaluate the potential of A. donax as an energy source and to assess if NIR can be used for fast prediction and monitoring of its AD energetic potential.