Hydrogen production by steam reforming of diluted bioethanol solutions

INTRODUCTION Bioethanol attracted growing interest as raw material for the production of hydrogen. The latter can be successfully used to feed fuel cells with the aim of combined heat and power (CHP) cogeneration. A demonstrative project has been carried out c/o the Dept. of chemistry of Università degli Studi di Milano 1 by running an integrated CHP unit with residential size, i.e. 5 kWel + 5 kWth. It was constituted by 6 integrated reactors, connected in series, fed with bioethanol and water, to produce reformate gas (prereformer and steam reformer) and to accomplish gas purification from CO (high- and low-temperature water gas shift reactors and two methanators). The purified reformate is suitable to feed a PEM fuel cell with the above mentioned power output. The aim of this work was to focus on process intensification, to achieve an economically sustainable solution. This was done at first by identifying the effect of bioethanol concentration on process output and proposing suitable means to achieve bioethanol purification. This is particularly straightforward because second generation bioethanol is nowadays proposed as fuel or as blending agent for gasoline, thus requiring deep dehydration. However, if used as feed for steam reforming, much lower concentration is needed, allowing to limit distillation/dehydration cost, which account for 50-80% of bioethanol production expenses 2,3. EXPERIMENTAL/THEORETICAL STUDY Process simulation has been carried out by using the Aspen Plus software and economic assessment of the solutions proposed has been carried out by using the Aspen ONE cost evaluation tool. The experimental apparatus was furnished by Helbio SA and was extensively described elsewhere 1. Activity resting with bioethanol solutions of different concentration and purity have been carried out by means of a bench scale continuous plant, at 300-750°C, 1 atm, H2O/CH3CH2OH = 3 (mol/mol). RESULTS AND DISCUSSION Process layout has been revised by comparing different possible schemes, compared as for heat and power duty (input) and output. Different solutions to account for energy input to the reformer have been also compared. In order to use diluted bioethanol solutions, the reformer was thermally sustained by catalytic combustion of part of the reformate. When using more and more diluted streams, the amount of reformate fed to the combustion line increased (to vaporize more water), so decreasing the electrical output of the fuel cell. However, the spent heat of the generated steam was recovered as thermal energy, so increasing the thermal efficiency. Additionally, different options for the purification of raw bioethanol beer have been compared, in order to meet the required specifications and to limit the cost of the reformer feed. Depending on the desired ethanol concentration, a flash drum was sufficient and economically feasible to reach intermediate bioethanol concentration (15-25 vol%), whereas for lower desired concentrations, the raw bioethanol flow was partly rectified to the azeotrope and then diluted with the untreated beer. This solution proved convenient to limit the heat input to the column reboiler, provided that no poisons for the reforming catalyst are present in the feed. To address this latter point, the effect of the purity of the resulting bioethanol stream on reformer performance has been also experimentally addressed. The effect of reactor temperature was considered, and this was set at the minimum level to guarantee optimal product yield, suitable catalyst durability and minimum heat input to the reactor. CONCLUSION The possibility to use diluted bioethanol streams as less expensive raw material for hydrogen production by steam reforming has been demonstrated and coupled with suitable bioethanol purification strategies. REFERENCES 1. I. Rossetti et al, Int. J. Hydrogen Energy, 37, 8499 (2012) 2. I. Rossetti et al, Chem Eng. J., 281, 1024 (2015) 3. I. Rossetti et al, Chem Eng. J., 281, 1036 (2015) ACKNOWLEDGMENTS The financial support of Linea Energia, Università degli Studi di Milano, Provincia di Lodi and Parco Tecnologico Padano is gratefully acknowledged.