A fundamental study of model fuel conversion reactions in sub and supercritical water

Abstract

Model reactants under hydrothermal conditions were examined to improve our understanding of chemical transformations in this high temperature and pressure environment. Results have a direct impact on present and future hydrothermal fuel conversion research for a range of fossil-based and bio-mass based feed stocks. Methane was chosen as a model compound and two different approaches were taken to examine its conversion in supercritical water. Catalytic reformation of methane was studied experimentally while partial oxidation of methane was studied through the application of a previously developed detailed chemical kinetic model that was analyzed and refined specifically for this study. Glucose and glycine were also chosen as model compounds to study related conversion pathways experimentally under hydrothermal conditions for biomass-based feed stocks. An experimental study of the catalytic reformation of methane in supercritical water (SCW) was completed that explored the use of carefully chosen catalysts under a variety of conditions and measured the conversion of methane and yields of various products.

(cont.) Eight metal catalysts were selected based on a review of previous catalysis experiments in hydrothermal conditions and those thought to be active for methane reforming. The range of conditions studied included 350 - 630⁰C, 150 - 400 bar, 0.01 - 2 wt% methane, 10 seconds to 72 minutes residence time, and with and without catalyst present. Four different experimental reactor designs were employed; a packed bed reactor, a continuous stirred tank reactor and two different batch reactor designs. A variety of techniques for reducing the metal catalysts and keeping them active in SCW were examined. Despite the range of conditions studied here, significant conversion of methane was never achieved. The most encouraging result was the relatively low yield of CO₂ (2.19% of the product gas volume) in the experiments employing 1% Ru/TiO₂ catalyst pellets. An analysis of each catalyst before and after exposure to SCW revealed significant degradation which helped to explain the low methane conversions. Based on this analysis and our experimental results, the most promising active metal identified was ruthenium, and the most promising support was titania (rutile) with some promise for zirconia and activated carbon.

(cont.) Although active for steam reforming and other hydrothermal catalyst applications, the nickel and platinum catalysts examined in this study showed signs of rapid degradation and deactivation and yielded little conversion of methane. In a previous study, researchers claimed to produce hydrogen from methane in SCW in the presence of alkali salts. Experiments with alkali salts in SCW were investigated here to further examine this claim. Our experiments with alkali salts revealed the importance of corrosion in the evolution of hydrogen from this media. Comparable amounts of hydrogen were produced from argon-alkali-SCW mixtures and from methane-argon-alkali-SCW mixtures suggesting that a significant amount of hydrogen in SCW reaction effluents can be attributed to water oxidizing the metal reactor material and not from hydrocarbon sources. Additional SCW alkali salt experiments in the same Hastelloy C-276 reactor eventually revealed an increasing catalytic conversion of methane, further emphasizing the likely importance of progressive metal corrosion.

(cont.) In the Hastelloy C-276 reactor, corrosion was confirmed by the presence of metal particulates and measurable amounts of dissolved nickel and chromium from the reactor metal alloy in the effluent. Comparable experiments in a gold-plated reactor still showed evidence of hydrogen generation from metal oxidation, but did not show evidence of corrosion. A detailed chemical kinetic model (DCKM) for single carbon species (C1) was refined and analyzed to support an examination of the effects of experimental conditions on methanol selectivity and methane conversion for the partial oxidation (POX) of methane in SCW. Although a formal sensitivity analysis was not performed on this model, a study of several key reactions and rates from literature resulted in good agreement of model predictions with reliable C1 SCW oxidation experimental data. SCW methane POX predictions from the refined model were then compared with POX experimental data. Disagreements between the model and the data were discussed along with a detailed critique of experimental issues associated with all previous SCW methane POX experimental studies.

(cont.) A reaction path analysis was developed from the DCKM which helped to elucidate the fate of methane and methanol in this environment and to identify a set of promising conditions to maximize methanol selectivity. Upon detailed analysis of both experimental and modeling results, the maximum methanol selectivity of about 80 % and maximum methane conversion of about 1% occurs at low temperatures ([approx.] 400⁰C), medium to high pressure (P > 300 bar), and high methane concentration ([CH₄]₀ > 50mM) with fuel-rich conditions at medium to high methane to oxygen ratios of [CH₄]₀/[0₂]₀ > 10. The experimental results may have achieved less than the maximum possible methanol selectivity due to issues such as inadequate mixing and wall effects. The modeling results may also be under-predicting methanol selectivity due to inadequate inclusion of non-ideal PVTN effects and solvent effects. However, the current model predictions and experimental results both substantiate our concern that SCW methane POX may fall short of the goal of greater than 70% methanol selectivity and 15% methane conversion. Nevertheless, other sets of experimental conditions that may show more promise have not been fully explored experimentally.

(cont.) In particular, the use of stable, selective catalysts, or inert wall material, or partial oxidation in the presence of hydrothermal flames have not been thoroughly analyzed here, and may improve the limited success discovered in this study. Glucose, glycine and glucose-glycine mixtures were studied as a model Maillard reaction system in a hydrothermal environment to explore a range of conditions that might alter the formation of undesired Maillard-type polymeric products. These polymeric products reduce the yield of biomass-derived fuels and complicate the separation and processing steps of biomass-to- fuel applications. Initial experiments were performed to study the individual hydrothermal degradation pathways of glycine and glucose and how those pathways change when the model compounds are mixed. Despite varying pH, time and temperature, we did not observe significant changes in the proposed Maillard mechanism, but product chromatograms did show possible development of alternate pathways particularly with furfural-type compounds.

(cont.) Glycine alone was found to be largely refractory (only 0 - 33% conversion) in our hydrothermal conditions from 50 to 300⁰C at 55 - 110 bar and 4 - 67 minutes residence time while glucose alone was quite reactive. In most conditions studied here, glucose conversion was greater than 85%, but moderate glucose conversions were achieved in a new, short residence time plug flow reactor (e.g., conversion of 35 % was measured after 7.3 seconds at 200⁰C and 55 bar). The degradation of glucose-glycine mixtures was studied at times of 7 seconds and 6 minutes at pH 2 and pH 5 and over a range of temperatures from 100 - 300⁰C. Near complete conversion of both reactants was observed in almost all conditions. Several liquid phase products were identified and analyzed, but total organic carbon (TOC) and carbon-hydrogen- nitrogen (CHN) analysis showed that significant reacted carbon is still unaccounted for.