Biomass-derived carbon as a precursor for bipolar plate in vanadium redox flow battery

Abstract

In vanadium redox flow battery (VRFB), electrodes comprise of carbon/graphite felt and composite bipolar plate. The felt provides reaction sites for vanadium redox reactions and controls mass transport of electrolyte flows throughout VRFB electrodes, while the bipolar plate provides electrical conductivity and connects adjacent cells in VRFB. However, current studies have focused mainly on the carbon/graphite felt, giving few attention to bipolar plates. In chemistry, a heteroatom is an atom that is not carbon or hydrogen, such as oxygen, nitrogen and phosphorus. Bipolar plates lack heteroatom functional groups and porous structures, which have been shown to improve kinetics of vanadium redox reactions and energy efficiency. Both heteroatom functional groups and porous structures are prominent properties of biomass-derived carbon, which is renewable and abundant. Therefore, the aim of the PhD study is to examine the ability of applying biomass-derived carbon as a precursor for bipolar plate in VRFB, in order to simultaneously leverage the strengths of biomass-derived carbon in vanadium redox reactions and reduce the price of bipolar plates. To achieve this aim, walnut shell-derived carbon was obtained via heat treatment in various environments i.e. N2 and CO2 at 600-1000C, and hydrothermal carbonisation (HTC – 200-290C). Walnut shell-derived carbon was functionalised through a combination of chemical and thermochemical treatments e.g. NH3, HNO3, K2CO3 and H3PO4, and dielectric barrier discharge (DBD) cold plasma in NH3. Properties of walnut shell-derived carbon were characterised to understand the effects of operating conditions on functional groups, surface morphology and carbonaceous structures. DBD cold plasma is a novel method to incorporate nitrogen into walnut shell-derived carbon, in which DBD cold plasma power is a dominating factor controlling nitrogen surface contents and nitrogen surface species with surface nitrogen contents varying in 3.9-8.7 at%. Low cold plasma power 20 W favours the formation of amide-N groups (2.8-5.1 at%), but increasing cold plasma power to 40 and 80 W creates more amine-N groups (2.8-3.5 at%). The ratio of amine-N/amide-N group increases from ~0.4 to ~2.5 with increasing DBD cold plasma power from 20 W to 80 W. This is significantly different to heat treatment in NH3 at 800C with pyridinic-N, pyrrolic-N and quaternary-N being main nitrogen species. The total surface nitrogen contents of carbons obtained from NH3 heat treatment are 3.1-5.6 at% with the ratio of pyridinic-N and pyrrolic-N being ~2:1. Furthermore, DBD cold plasma has negligible influence on porous structure and carbonaceous structure of both biochar (obtained from pyrolysis) and hydrochar (obtained from HTC) with BET N2 surface area remaining < 10 m2 g -1 . It is recommended to use different analysis techniques such as temperature programmed desorption to examine nitrogen functional groups besides X-ray photoelectron spectroscopy due to overlapping binding energies of amine, amide, and pyridinic-N, pyrrolicN and quaternary-N. Regarding H3PO4 treatment, although phosphorus incorporation using H3PO4 did not significantly increase the phosphorus contents in walnut shell-derived carbon (1-2 wt%), it is proved that both heat treatment temperatures (in the range of 600-1000C) and concentration of H3PO4 solution (0.5-5 M) govern the relative contents of C-P groups (C-PO3/C2-PO2/C3- PO) and C-O groups (C-O-PO3/(C-O)3PO). At H3PO4 0.5 M, temperature  1000C is required to transform C-O-PO3/(C-O)3PO to C-PO3/C2PO2, but most phosphorus functional groups are shifted to C-PO3/C2PO2 in the whole temperature range of 600-1000C when using H3PO4 0.5 M. It has been found that micropores account up to 57-78% of total pore volumes in most walnut shell-derived carbon samples, and ultramicropores and micropores do not play any significant role in contributing to reaction kinetics of vanadium redox reactions, due to diffusion limitations. Although the total pore volumes vary in the range of 0.013-0.380 cm3 g -1 , the domination of ultramicropores and micropores in most samples suggested that the porosity and surface area do not involve in the conversions of vanadium redox reactions. The oxidation of VO2+ to VO2 + was affected by differences in molecular structures as characterized by Raman spectroscopy and PXRD between 600-800°C with ID/IG ration increasing from 0.69 to 0.94 and La increasing from 18.33 Å to 23.15 Å, respectively. With the significant increase of total oxygen content from 7.54 wt% to 23.84 wt% through HNO3 treatment, it has been found that the presence of oxygen functional groups improves the conversion of V 2+/V3+ and V 3+/VO2+. NH3 heat treatment at 800°C increased the nitrogen contents markedly leading to the improvement of both the kinetic transfer of V2+/V 3+ and V3+/VO2+ couples. Surface oxygen functional groups does not enhance the conversion of VO2+/VO2 + couple, but they contribute to the reversibility of the oxidation and reduction processes of VO2+ and VO2 + .