Our research focuses on CO2 emission reduction of energy conversion systems in terms of exergy engineering.

Capture of CO2 from power plants and industrial emissions is commonly achieved by an amine solvent-based chemical method. However, small amounts of amine can be released to the environment during this process. These amines are converted by natural processes into nitrosamines which can pose a threat to the environment and human health even at the ppm order. Amine release can occur through three pathways: 1) as aerosol emission, 2) as liquid entrained by a carrier gas, and 3) as an evaporated volatile amine. Aerosol amine emissions can be prevented by using a wet electrostatic precipitator. Gas-entrained liquid amine can be captured by demisters. However, volatile amines cannot be sufficiently removed using existing methods. Current research focuses on utilizing a water wash column to remove volatile amines. In this research, the physical properties of 5 amines were acquired by bench-scale experiment with the aim of designing a washing tower for volatile amine removal. While the existing theoretical framework in the literature qualitatively captured the behavior of observed in empirical testing, a degree of quantitative error remained. By manipulating the highly contributing gas phase diffusion coefficients of amines, the resulting values were within the theoretical framework. Inclusion of this data increased the prediction accuracy such that the conventional theoretical framework can be used to design a packed tower for removal of vapor amines.

Thermochemical recuperation (TCR) is an attractive technology for the utilization of unused heat in excess of 1000 ˚C, as can be found in industries using high temperature furnaces. In the TCR process, high temperatures are used to upconvert fuel via steam reforming and dry CO2 reforming reactions. The higher value fuel reduces fossil fuel consumption and CO2 emissions. A prototypical example of TCR is the conversion of natural gas (CH4) into syngas (a mixture of H2 and CO). An issue with TCR is the deposition of elemental carbon on alumina catalyst surfaces, blocking the gas-catalyst reactions at the heart of TCR. To evaluate the potential for carbon deposition and preventative measures, the TCR reactions were modelled in CHEMKIN-II and CRESLAF. CHEMKIN-II can consider non-isothermal gas phase reactions, Al2O3 catalyst surface reactions, and mass transport. The 2D CRESLAF model operates in parallel to the CHEMKIN-II model. In this research, we aimed to improve the carbon deposition model.
An in-house, carbon deposition calculation to supplement the basic TCR reaction scheme provided qualitative approximation of carbon deposition but lacked the quantitative accuracy necessary for system design and implementation. However, this method only considered gas phase reactions on the assumption that CH4, CO2, and CO would not be adsorbed on the alumina catalyst surface. However, we confirmed experimentally the adsorption of CO and CO2 on the catalyst surface. To incorporate carbon deposition into the theoretical framework, a new gas phase reaction scheme including surface adsorption was created.
In the elementary carbon deposition reaction, the activation energy was gleaned from the literature and the frequency factor was determined based on fitting to experimental results to be 1.0 × 1022 s-1. Using the new reaction scheme and the fitted parameters, carbon deposition and low CO concentration TCR reactions (i.e., steam/CH4= 3~4) can be accurately simulated.

A spatiotemporal mismatch between unused heat exhausted from sugar mill and heat demands at factories exists on the island of Tanegashima. The unused heat comes from the burning of biomass that remains after sugar production (called, ‘bagasse’) in a boiler. A thermochemical energy storage and transport system utilizing a zeolite steam ad/desorption cycle is proposed to move the unused heat from bagasse combustion to heat demands at nearby factories. Zeolite releases heat upon adsorbing water and is ‘re-charged’ by simple drying. This study focused on the heat storage portion of the system. First, a unit to dry zeolite using flue gas from the bagasse boiler was designed; this unit is called the ‘Heat Charger’. Two types of gas-solid heat exchangers were considered for the Heat Charger: a counter-flow type and a cross-flow type. Each Heat Charger was designed using 1D and 2D simulations that coupled mass, energy, and steam conservation equations. Due to pressure loss through the Heat Charger, power must be supplied to a blower, which in turn increases operating costs. The design method of the Heat Charger aimed to minimize this blower power consumption. The analyses indicated a higher blower power consumption in the cross-flow design. Accordingly, only the counter-flow design optimized for minimal power consumption was adopted.
Based on the optimal Heat Charger design, a techno-economic analysis of the heat charging cost (JPY/kWh) was calculated on a levelized cost basis. The sensitivity of levelized cost to the gas temperature entering the Heat Charger and the dryness of zeolite exiting the Heat Charger was analyzed, accounting for differences in Heat Charger design needed to achieve these operational criteria. From this analysis, an optimal pinch temperature between the inlet gas and flue gas of 5.0 °C was determined. Additionally, optimal operations occur when zeolite exiting the Heat Charger achieves a dryness of 0.06 kg-H2O/kg-zeolite.
In addition to design of the Heat Charger, it was posited that the enthalpy of the flue gas could be increased by drying bagasse prior to combustion in the boiler. Process simulation of heat and material flow in the sugar mill found that reducing the moisture content of bagasse by 2% increased the flue gas temperature from 182 °C to 188 °C and increased the power generation by 31 kW. This increased power generation can be applied to the blower that provides hot air to the Heat Charger. Applying all the additional power to blower operation allows for decrease in operational heat charging costs and results in up to 2600 kg/h of zeolite to be re-charged. Since the reduction in blower power cost exceeds the increase in equipment cost which accompanies the bagasse drying process, if bagasse drying is applied to the optimized 2000-2200 kg-zeolite/h Heat Charger, it is expected that heat charging costs will be further reduced.

After the Great East Japan Earthquake, there is demand for transformation to a bi-directional energy transmission system with more distributed power sources. In this role, a fuel cell co-generation system with improved energy utilization is an attractive technology. Even in their current state of development, solid oxide fuel cells (SOFC) are a highly efficiency power generation system. However, it is necessary to reduce the raw material cost and increase the power density to accelerate penetration of SOFC. In SOFC, electrochemical reactions convert fuel directly into electricity. These reactions occur only at locations called triple phase boundaries (TPB). The TPB is where an uninterrupted pathway for electron flow, ion diffusion, and fuel gas flow exist. Increasing the volumetric density of TPB produces a SOFC with improved performance. This researched aimed to generate a multi-layered, cross-hatched microscale geometry to increase the TPB density. Each layer consists of a solid anode and solid electrolyte material which are manufactured in a method similar to weaving. Numerical calculations were used to determine the impact on SOFC performance of fiber diameter, degree of overlap between neighboring fibers, degree of overlap between fiber layers, and the number of layers. By decreasing the fiber diameter, anode overvoltage was reduced. When the fiber diameter was reduced to 3 μm or less, the anode overvoltage fell below that of a SOFC without a designed microstructure. Additionally, increasing overlap between neighboring fibers and fiber layers resulted in reduction of anode overvoltage; this effect was due to an increase in the TPB density. The highest TPB density was found to occur when neighboring fiber overlap is 8% of fiber diameter and fiber layer overlap is 30% of fiber diameter. The optimal structure to reduce anode overvoltage was determined to have a fiber diameter of 1 μm, the degrees of overlap mentioned above, and 14 fiber layers. Using this structure reduced anode overvoltage compared to an unstructured SOFC by 66.4%.

Carbon dioxide removal from the atmosphere at the scale of 10 GtCO2/y by the end of the century is likely necessary for a 50% chance of meeting the 2°C scenario set out by the Paris Agreement. We have previously developed a profitable pathway to use desalination brine to mineralize CO2 into stable magnesium carbonates while also producing fresh water, gypsum (CaSO4˙2H2O), halite (NaCl), sodium sulfate (Na2SO4 ˙10H2O), and aqueous hydrochloric acid. The process uses evaporative concentration of desalination brine to separately precipitate gypsum and halite. Subsequent cooling of the solution removes sodium sulfate. Hydrochloric acid is added to precipitate NaCl and then recovered by addition of magnesium chloride. The remaining magnesium chloride is thermally decomposed, forming hydrochloric acid and periclase (MgO). The MgO is then used to mineralize atmospheric CO2 into stable MgCO3˙3H2O.
The sub-process consuming the greatest amount of energy is the evaporative concentration of desalination brine due to the large latent heat of water vaporization. To reduce the energy consumption of evaporative concentration, a crossflow, thin-film evaporator was designed. The system takes advantage of low relative humidity air and the sensible heat of desalination brine to drive evaporation. The system, sized to treat 100,000 tonne/d of desalination brine, achieved brine evaporation at an electrical energy consumption of 23 kJ/kg-water while increasing the footprint of the desalination plant by only ~5%.
The other bottleneck in the process is the mineralization of gaseous CO2 by solid MgO, which is typically considered a slow reaction. In this year, experiments showed that the thermal decomposition of magnesium chloride in air can be realized at 350 ˚C, instead of 400 ˚C as is commonly reported in the literature. Performing thermal decomposition at this lower temperature not only reduces energy consumption, but also generates an amorphous MgO structure. Experiments showed that this amorphous MgO mineralizes gaseous CO2 many orders of magnitude faster than crystalline MgO. Mineralization was further accelerated by producing MgO as a micrometer thin sheet, resulting in complete CO2 mineralization within one day.

The iron and steel making industry (ISM) is one of the largest CO2 emitters in the world. In the ISM process, iron ore, calcium carbonate, and coking coal are reacted to form metallic iron. Concurrently, the process generates a molten calcium silicate melt (called ‘slag’) from impurities present in the starting materials. We propose reducing CO2 emissions in ISM by recovering CaO and FeO from molten slag via centrifuging. These materials would be recycled to the iron and steelmaking process, reducing CO2 emissions by ~13%. Lab-scale demonstration of molten slag separation by centrifuge for a number of slag compositions has been reported in the literature. However, no theoretical framework has been developed which can accurately predict the material composition and degree of separation of molten slag during centrifuging. This lack of theoretical basis prevents large-scale system design and optimization. Therefore, we simulated molten slag centrifuging experiments reported in the literature via numerical calculation with an aim to elucidate the relevant physical mechanisms controlling the separation phenomenon. In our numerical calculation model, the molten slag was simulated as an ensemble of microscale spherical aggregations of the fundamental compounds composing the molten slag. Particles did not directly interact but were also not allowed to share the same location. The particles moved under the balance of centrifugal, gravity, buoyancy, and drag forces based on the local molten slag properties. The molten slag properties of density, viscosity, liquidus temperature, and nucleation lag time were predicted using deep neural networks based on the local elemental composition and temperature. Initial simulations used monodisperse aggregations of simple oxides with some success. Greater accuracy was achieved by calculating aggregate sizes, accounting for the molar volumes, diffusivity, and statistical mechanics. Results from hard-sphere mixing theory and calculation by phase diagram suggested the presence of additional compounds for the given slag composition and experimental temperature. Inclusion of these compounds to the simulation yielded simulation results that matched quantitatively with experimental results. Based on the developed theoretical framework for molten slag separation, the increase of slag temperature will yield more rapid separation of purer materials through a lower slag viscosity and suppression of mixed compound nucleation.