The use of lignocellulosic biomass for the production of value added chemicals and fuel molecules is a promising possibility to meet future society needs for renewable resources. Crucial steps within a lignocellulosic biomass refinery are reactions from hexoses or pentoses, such as glucose or xylose, to the corresponding platform chemicals. Thereby, the utilization of biomass should be sustainable with regard to the principles of green chemistry. In literature, the use of toxic chromium salts , corrosive halides , toxic ionic liquids  or sulfonated solvents often improves the reaction towards the specific platform chemical, but these reaction systems might be ecologically questionable.
In this work, we focus on the optimization of an environmentally favorable reaction system. The examined model reaction system from xylose to furfural exhibits high economic and environmental potential due to the use of low priced iron catalysts. The reaction system contains none of the problematic aforementioned substances. The consecutive reaction to humins is the substantial challenge.
We use a holistic methodology to investigate and to optimize the model reaction system with regard to reaction conditions and reactor design that lead to high selectivity. Therefore, we combine the use of PAT (Process analytical technology) tools, kinetic modelling and further reaction engineering.
The key task of the xylose dehydration is to minimize or avoid the consecutive reaction to humins. Thus, the objective of this work is to perform real-time monitoring of the model reaction from xylose to furfural with ATR-UV/Vis, ATR-mIR and Raman spectroscopy to gain insights into the reaction kinetics. A special autoclave is used that allows spectroscopic investigations of multiphase systems at elevated pressures and temperatures. The findings hereof allow the kinetic modeling and a design of a suitable reactor. To scale-up the investigated reaction system, a continuous plant is designed, building on the results of the batch experiments. DoE (Design of Experiments) together with PAT tools are used to investigate the optimal reaction conditions of the continuous system.
The batch reaction was performed in a two-phase system containing water with dissolved iron (II) and iron (III) sulfate as catalyst and 2‑methyltetrahydrofuran as organic extraction phase (see figure 1). Both reaction phases could be analyzed simultaneously with two fiber optical probes under reaction conditions. ATR mIR and Raman spectroscopy were used in both phases to gain kinetic data at reaction temperatures between 100 °C and 160 °C.
Figure 1: Scheme of the fiber optical probes and the investigated batch reaction from xylose to furfural with simultaneous extraction of furfural.
The continuous xylose dehydration was performed in a liquid-liquid Taylor flow in a capillary, containing water with dissolved iron (III) sulfate as catalyst and 2‑methyltetrahydrofuran as extraction phase. The reaction conditions were varied along a statistical plan. Thereby, the reaction was performed at different temperatures, residence times, concentrations of the substrate xylose and of the catalyst.
The influence of temperature on the yield of furfural can be examined by the in-situ spectroscopic methods. The initial reaction rate of the xylose dehydration is determined at different temperatures with the help of chemometrics. Hereby the activation energy of this reaction is modelled with an Arrhenius plot. During the dehydration, humins are mostly formed in the water phase. Initial reaction rates and an activation energy that refers to the humin formation are modelled. Furthermore the time span is identified after which the consecutive reaction to humins becomes dominant. These investigations are shown for both, iron (II) and iron (III) sulfate, which gives a hint for the actual active catalytic species.
The activation energy of the xylose dehydration is higher than the activation energy referring to the humin formation. Regarding both investigated iron sulfates, both determined activation energies of the xylose dehydration are equal. However, the use of iron (III) sulfate exhibits a higher activity at lower temperatures. Hence, the continuous experiments are performed in a biphasic system with iron (III) sulfate as catalyst tending to relatively high temperatures and a short residence time. The batch experiments also showed the need for an optimization of the extraction. Therefore, the continuous reactor is designed as liquid-liquid Taylor flow, which leads to high mass transfer rates.
The biphasic reaction solutions of the different reaction conditions referring to the statistical plan are analyzed by ATR UV/Vis, ATR mIR and Raman spectroscopy. The results of the PAT analytics are compared with classical offline analytics, such as HPLC and GC. The reaction conditions are optimized with a DpE central composite design, which lead to superior conditions regarding the selectivity towards furfural.
We show that the comprehensive reaction monitoring by the use of ATR-UV/Vis, ATR-mIR and Raman spectroscopy in combination with kinetic modelling and further reaction engineering are powerful tools for the research on the utilization of a lignocellulosic feedstock.
The modelled kinetics are crucial for minimizing the humin formation and thereby increasing the selectivity of the reaction towards furfural. With that gained knowledge, a continuous reactor design with specific reaction parameters is proposed to intensify the yield of the platform chemical. The optimal reaction conditions of the continuous reactor design lead to an intensified yield of furfural and a more sustainable use of the renewable lignocellulosic biomass.
This work was performed as part of the Cluster of Excellence "Tailor-Made Fuels from Biomass", which is funded by the Excellence Initiative by the German federal and state governments to promote science and research at German universities.
 J. Janssen, F. W. Kremer, J.
H. Baron, M. Muether, S. Pischinger and J. Klankermayer, Energy Fuels 2011, 25, 4734-4744.
 P. T. Anastas and J. C. Warner, Green Chemistry: Theory and Practice, Oxford University Press: New York, 1998, p.30.
 J. B. Binder, J. J. Blank, A. V. Cefali and R. T. Raines, ChemSusChem 2010, 3, 1268-1272.
 K. R. Enslow and A. T. Bell, ChemCatChem 2015, 7, 479-489.
 R. J. van Putten, J. C. van der Waal, E. de Jong, C. B. Rasrendra, H. J. Heeres and J. G. de Vries, Chemical Reviews 2013, 113, 1499-1597.
 T. vom Stein, P. M. Grande, W. Leitner and P. Domínguez de María, ChemSusChem 2011, 4, 1592-1594.
 S. Hardy, I. M. de Wispelaere, W. Leitner and M. A. Liauw, Analyst 2013, 138, 819-824.