Optimization of pre-treatment process with alkaline hydrogen peroxide to improve the efficiency of methane production from greenhouse wastes

In this thesis; alkaline hydrogen peroxide pre-treatment process was examined and optimized to increase the amount of effective substrate quantitiy of lignocellulosic waste material originated from greenhouses, to enhance the anaerobic degradability, to increase the amount of produced methane and to shorten the digestion time. In this thesis; primarily, the most widespread cultivated five vegetables (tomato, cucumber, eggplant, green peppers and zucchini) in Antalya region were mixed by using their roots, stems, leaves and fruits considering the amount of production. The analyses of total solids, volatile solids, total Kjeldahl nitrogen, protein, total and dissolved chemical oxygen demand, dissolved reduced sugar, extractive matter including lipid, Van Soest Fraction (cellulose, hemicellulose, lignin, soluble matter) and elemental composition were determined. To optimize alkaline H2O2 pre-treatment process, Response Surface Methodology (RSM) used for analysis of engineering problems, modeling and optimization was performed to plan experimental set up and to determine the optimum conditions. Alkaline H2O2 pre-treatment process was examined through the effects of selected independent variables on dependent variables. Independent variables which considered to have effects on alkaline H2O2 pre-treatment process are reaction temperature, reaction time, solid concentration and hydrogen peroxide (H2O2) concentration, while selected dependent variables are dissolved COD increase, dissolved reducing sugars increase, the change of extractives-free lignin amount and biochemical methane potential (BMP). The central composite design (CCD) method which is present in statistical response surface methodology (RSM) Design Expert software was used to assess the effects of independent variables on the dependent variables. CCD pre-treatment experiments are planned with Design Expert® program, recommended CCD pre-treatment experiments were performed, models were developed for soluble COD, soluble reduced sugar, change of extractive-free lignin amount and BMP results and finally the validity of the models was evaluated by ANOVA test. The proposed models for soluble COD, soluble reduced sugar, BMP and lignin have low p-values and were found to be statistically significant. Regression coefficients for soluble COD, soluble reduced sugar, change of extractive-free lignin amount and BMP models were determined as 0.9682, 0.7740, 0.8376 and 0.5728, respectively. As a result of alkaline H2O2 pre-treatment CCD trials; it was found that soluble COD and soluble reduced sugar concentrations were increased and BMP was decreased by increasing the applied reaction temperature. According to the scope of this thesis which aims to increase BMP production, different solution proposals were evaluated in alkali H2O2 pre-treatment process optimization because of the different behaviors of dissolved COD, dissolved reduced sugar and BMP. Alkali H2O2 pre-treatment process optimization was made according to maximum methane production and minimum process cost criteria by using models developed for increase of soluble COD and soluble reduced sugar, extractive-free lignin amount and BMP. Two different optimization solutions considering the maximum methane production and process cost were evaluatedfor maximum process yield. In the optimization which targets the maximum methane production; reaction temprature was minimized and solid concentration, H2O2 concentration and the reaction time were remained in range. In the second optimization that targets the maximum methane production with taking process cost into consideration, reaction temprature, H2O2 concentration, reaction time were minimized and solid concentration was maximized. Optimum pre-treatment conditions has been found at 1% H2O2 concentration, 50°C reaction temperature, 6 hours of reaction time and 7% solid content at the maximum methane production optimization considering the costs. In these conditions, soluble COD and soluble reduced sugar concentrations were estimated 296.352 mg COD/gVS and 102.130 mgsugar/gVS, respectively by the proposed model. As a result of model validation experiments soluble COD and soluble reduced sugar concentration were measured as 290.3 mgCOD/gVS and 106.9 mgsugar/gVS, respectively. Since error between predicted and measured values were calculated 2.1% and 4.67%, respectively for the models of soluble COD and soluble reduced sugar, these models can be used safely in design space. Also, BMP value of the pretreated samples in these conditions was measured as 309 mLCH4/gVS. As a result of this pre-treatment optimization, 79.15% more BMP was produced from raw sample's BMP. Optimum pre-treatment conditions for maximum BMP production without considering cost have been found as 2.1% H2O2 concentration, 50°C reaction temperature, 13.5 hours of reaction time and 5.6% solids content.Soluble dissolved reduced sugar concentrations were estimated as 262.549 mgCOD/gVS and 48.558 mgsugar/gVS, respectively by the models. As a result of model validation experiments, soluble COD and soluble reduced sugar concentrations were measured as 256.9 mgCOD/gVS and 47.3 mgsugar/gVS, respectively. Since, error between predicted and measured values were calculated 2.21% and 2.60%, respectively for the models of soluble COD and soluble reduced sugars, these models can also be used safely in design space. Furthermore, BMP value of the pre-treated samples in these conditions was measured as 328 mLCH4/gVS. As a result of this pretreatment optimization, 89.85% more BMP was obtained compared to raw sample's BMP value. Lignocellulosic structure surface properties, characterization of molecular bonds and distribution of Van Soest fractions of alkaline H2O2 pretreated greenhouse wastes were investigated by scanning electron microscope, Fourier transform infrared spectroscopy and VanSoest analysis, respectively. SEM images revealed that the raw sample has a rigid, stable, and uniform continuous surface, while fibrils of preteated sample are separated from the main structure and exposed. Even though low level slidings of wavelength in the FTIR spectra were observed, peaks of hemicellulosic breakdown fragments and lignin components were detected. According to Van Soest results; decrement in the amount of hemicellulose was observed and significant changes were not detected in the amount of lignin. Application of alkaline H2O2 pretreatment to greenhouse wastes and the process optimization has been studied for the first time in literature. Since being the first study in the literature including the optimization of alkaline H2O2 pre-treatment process for greenhouse wastes, determination of BMP potential and investigation of pretreatment effects on surface characteristics, it has the potential of being a reference point.