(648c) Electrochemical Deoxygenation of Bio-Oil from Fast Pyrolysis of Lignocellulosic Biomass: Energy Balance Estimated By Bond Dissociation Energies

Given the chemical structure of lignin polymers, aromatic rings interconnected by ether linkages, featuring hydroxyl and methoxyl side groups (Prasomsri et al. 2014), most processing of lignocellulosic biomass renders bio-oils rich in oxygen (up to 50 %), which is associated with high acidity, instability, high viscosity, and low heating values. So does the bio-oil generated by fast pyrolysis of lignocellulosic (woody) biomass residues. For its compatibility with respect to standard petroleum fuels, the bio-oil from pyrolysis needs thus upgrading by a deoxygenation process.

The state-of-the-art technology for deoxygenation, catalytic hydrodeoxygenation (HDO), uses H₂ gas in high stoichiometric excess with respect to the bio-oil (342 L to 669 L per L of bio-oil, Elliot et al. 2012), to effectively break C-O or C=O bonds on phenyl rings, with the aid of a catalyst (mostly supported metals such as Ru, Rh, Pd, Pt, Re, Cu, Ni, Fe, and their heterometallic alloys) at moderately high temperatures (150 °C to 400 °C) and pressure (70 bar to 200 bar) (Zacher et al. 2014). HDO is therefore quite energy consuming and requires external production of H₂, associated with increase of the total environmental impact.

The novel electrochemical deoxygenation process (EDOx), developed by Ceramatec, Inc. (Salt Lake City), uses an electrode catalytic membrane that is permeable to oxygen ions, for in situ generation of H₂ at the cathode site. Constructed membrane cells can be located in the fast pyrolysis unit itself, before the condenser, optimizing the overall energy efficiency. The electrical current through the membrane is caused by permeating oxygen ions, which are oxidized at the anode, resulting in a pure oxygen gas stream at this side of the membrane.

In order to benchmark the EDOx process against the state-of-the-art catalytic HDO processing using life cycle assessment methods, mass and energy balances are required. We used the bond dissociation energy (BDE) of organic compounds to estimate the theoretical energy demand of deoxygenation and validate those estimates with experimental data. Firstly, as a proof of concept, an energy balance was calculated for syringol and guaiacol, which are both considered model compounds for bio-oil.

Apart from EDOx experiments on bio-oil samples, preliminary tests were also performed using model compounds for bio-oil, such as syringol and guaiacol. The use of these model compounds allows gaining in-depth insights in the reaction pathways, and constructing a mass and energy balance for each experiment. We used the experimental data of Elangovan et al. (2015) as input for the development of mass and energy balances. In this paper, EDOx experimental results with a mixture of syringol, containing 31.1 wt% of bound oxygen, and steam were used as an example of the applied methodology. After the EDOx process (at 550 °C and 1.3 V), a liquid product was obtained containing 76 wt% of phenol, 22 wt% of o-cresol, 1 wt% of 2,6-xylenol and 1 wt% of 2-ethylphenol, on a dry basis, as measured by GC-MS. This liquid product has 16.4 wt% of bound oxygen. With a corresponding deoxygenation rate of 47.2 %, the performance of this syringol experiment was not yet optimal, but the fact that only four liquid products were formed allowed for understanding the chemical mechanisms and associated energy balance. Apart from the O₂ gas generated at the anode, at the cathodic (bio-oil) side of the unit a gas stream was obtained (after condensation of the liquid), containing methane (0.1 wt%), ethane (0.1 wt%), ethene (0.1 wt%), propane (1.4 wt%), CO (0.9 wt%), CO₂ (3.8 wt%), and H₂ (0.01 wt%), as measured by GC, with the remaining part N₂ carrier gas.

Although the mass flux was not determined for the product streams, a mass balance was obtained using the mass conservation law with respect to elemental carbon, while assuming that phenyl rings remained intact during the EDOx experiments. To determine the theoretical energy balance, Hess’ law was applied using BDEs taken from literature. Although these BDEs are not available for all bonds of syringol and guaiacol, to our knowledge, Prasomsri et al. (2014) showed that the effect of other substitute groups on phenyl rings on the BDEs are negligible. As a consequence, the BDEs of the methoxy substitutes of syringol and guaiacol can be estimated by the BDE of the carbon-methoxy bond in anisole (C₆H₅-OCH₃), which can be found in e.g. Blanksby & Ellison (2003).

The energetic assessment of the deoxygenation reactions in the syringol EDOx process was performed using three different stadia: (i) the stoichiometric reactions of syringol to the four aromatic liquid products with hydrogen gas, yielding methanol and water as side products, (ii) the reaction of methanol, with or without hydrogen gas, yielding the aforementioned seven gaseous compounds and water, (iii) the total hydrogen balance, including all the consumption and generation in all stoichiometric reactions of steps i and ii, is translated into an electric energy demand, together with the measured H₂ excess in the product gas stream.

For the syringol experiment, the reactions leading to the main liquid products, phenol and o-cresol, are exothermic, whilst the reactions of syringol with H₂ to 2,6-xylenol and 2-ethylphenol are endothermic. Overall, these reactions yield a heat generation of 17.5 kcal per mol of syringol. The reactions of the gas phase intermediate methanol to the gaseous product stream are also net exothermic, leading to another 2.8 kcal of heat generated per mol of syringol. The hydrogen balance shows that there is a net generation of H₂ in the chemical reactions, and together with the measured excess of H₂, this would correspond to a 2.3 kcal per mol of syringol reduction in the electricity demand.

Altogether, the chemical reactions thus yield an energy gain of 26 kcal per mol of syringol, corresponding to 0.71 MJ (196 Wh) per kg of syringol. By determining the oxygen content in the liquid and gaseous product streams, it was shown that 4.7 wt% of the syringol’s mass was effectively removed as pure O₂ gas at the anodic side of the membrane. Because the transport of oxygen ions through the membrane is responsible for the electric current, a theoretical minimum electric power consumption can be calculated to be 0.73 kJ (203 Wh) per kg of syringol that is deoxygenated in the experiments.

The global theoretical energy balance of the EDOx syringol experiments showed that the applied electric energy (203 Wh per kg of syringol), initiates the deoxygenation reactions, but meanwhile the exothermic nature of these reactions results in a heat generation of 150 Wh, which can be recovered, albeit with a lower exergy than the electric energy input. In addition, our analysis showed that hydrogen is generated in the process, not only by electrolysis of the steam, but also by the deoxygenation reactions themselves. Given the O₂ flow (of 4.7 wt% of the syringol input) resulting from electrolysis, our mass balance shows that the corresponding hydrogen amount is insufficient for the observed deoxygenation rate. We therefore for the first time experimentally showed a deoxygenation process that takes advantage of the reaction generated hydrogen.

The beneficial use of the reaction generated hydrogen, next to the electrolysis generated hydrogen, turns out to be one of the main advantages of the EDOx process, with respect to catalytic HDO. The overall low partial pressure of hydrogen gas, at atmospheric pressure, of the EDOx process, indicates the much lower activation energy of the occurring reaction mechanisms. Based on this observation, alternative reaction mechanisms can be proposed, involving not only catalytic site stabilization of intermediates, but also the active contribution of electrons from the cathode.

A similar strategy was applied to EDOx experiments with guaiacol, as a model compound for bio-oil, which yielded a more complex liquid product mix. As shown in this paper, the energetic assessment of these tests results in fundamental insights in the electrochemical deoxygenation reactions, and helps in the development of the process towards the production of high quality biofuels.

References

Prasomsri, T., Shetty, M., Murugappan, K., Román-Leshkov, Y. Energ. Env. Sci. 2014: 7, 2660-9

Zacher, A.H., Olarte, M.V., Santosa, D.M., Elliott, D.C., Jones, S.B. Green Chem. 2014: 16, 491-515

Elliott, D.C., Hart, T.R., Neuenschwander, G.G., Rotness, L.J., Olarte, M.V., Zacher, A.H., Solantausta, Y. Energy Fuels 2012: 26(6), 3891-6

Elangovan, S., Larsen, D., Hartvigsen, J., Mosby, J., Staley, J., Elwell, J., Karanjikar, M. Electrochemical upgrading of bio-oil. Accepted for publication in ECS Transactions, 2015

Blanksby, S.J., Ellison, G.B. Acc. Chem. Res. 2003: 36, 255-63