Direct Internal Reforming Molten Carbonate Research Articles

A new alkali-resistant Ni/Al2O3-MSU-1 core–shell catalyst for methane steam reforming in a direct internal reforming molten carbonate fuel cell

An alkali-resistant catalyst for direct internal reforming molten carbonate fuel cell (DIR-MCFC) is prepared by growing a thin shell of mesoporous MSU-1 membrane on Ni/Al2O3 catalyst beads. The MSU-1 shell is obtained by first depositing a monolayer of colloidal silicalite-1 (Sil-1) on the catalyst bead as linkers and then using NaF stored in the beads to catalyze the growth of the MSU-1 layer. The resulting core–shell catalysts display excellent alkali-resistance and deliver stable methane conversion and hydrogen yield in an out-of-cell test simulating the operating conditions of an operating DIR-MCFC. Higher conversion and yield (i.e., up to over 70%) are obtained from the new core–shell catalyst with MSU-1 shell compared to the catalyst with microporous Sil-1 shell. A mathematical model of the reaction and poisoning of the core–shell catalyst is used to predict the optimum shell thickness for its reliable use in a DIR-MCFC.

Thermodynamic analysis of application of organic Rankine cycle for heat recovery from an integrated DIR-MCFC with pre-reformer

This work deals with waste heat recovery from a proposed direct internal reforming molten carbonate fuel cell (DIR-MCFC), including an integrated pre-reformer. In this regard, some advantages are attainable over exhaust gas recycling. For instance, due to low temperature in the pre-reformer, carbon deposition and coke formation resulting from higher hydrocarbons can be eliminated. In this study, the cathode outlet provides the heat requirement for the pre-reforming process. After partial heat recovery from the cathode outlet, the stream still has a considerable energy and exergy (352.55°C and 83.687kW respectively). This study investigates waste heat recovery from the proposed DIR-MCFC, using an organic Rankine cycle (ORC) with two different configurations. In the first case, the cathode outlet provides the heat requirement for the pre-reforming process; then, it enters the heat recovery vapor generator of the organic Rankine cycle. In the second case, the cathode outlet is split into two streams for using in an ORC and supplying the pre-reforming process required heat. Several substances are selected as working fluids in order to compare their performance in the waste heat recovery system. The overall results at optimum conditions indicate that the energy and exergy efficiencies of the compound system are increased and its exergy loss is decreased with cathode splitting for all substances (1.1% average over all fluids). It is concluded that cathode splitting has a significant impact on the substances which have better performance in the first case. In both cases, toluene is the most efficient working fluid. In the second case, global energy and exergy efficiencies are found to be 60.45% and 57.75%, respectively, which are almost 2.67% higher compared to the first case for toluene.

Control of a direct internal reforming molten carbonate fuel cell system using wavelet network-based Hammerstein models

Inspired by a direct internal reforming molten carbonate fuel cell (DIR-MCFC) coupled with complicated nonlinear dynamics, the identification and control design of the Hammerstein model is presented. Through the sequential identification procedure, the static nonlinearity block is considered as the wavelet network which is trained and validated by the on-line learning algorithm, and the linear dynamic block is described by the state-space model in which parameters are estimated by the recursive least square algorithm. Using the numerical interpolation technique to approximate the implicit nonlinear function, we present a composite control framework consists of a nonlinear inversion and linear control. Through the closed-loop simulation tests, the nonlinear inversion design for the nonlinearity cancellation of a class of nonlinear systems is validated.

Identification and Control of a fuel cell system using wavelet network-based Hammerstein models

Inspired by a direct internal reforming molten carbonate fuel cell (DIR-MCFC) coupled with complicated nonlinear dynamics, the identification and control design of the Hammerstein model is presented. Through the sequential identification procedure, the static nonlinearity block is considered as the wavelet network which is trained and validated by the on-line learning algorithm, and the linear dynamic block is described by the state-space model in which parameters are estimated by the recursive least square algorithm. When a nonlinear inversion via the numerical interpolation technique is established, a composite control framework is used to ensure the good output regulation.

Direct internal reforming molten carbonate fuel cell with core–shell catalyst

A sort of core–shell catalyst as a novel anti-alkali-poisoning concept was prepared, tested and applied in the direct internal reforming molten carbonate fuel cell (DIR-MCFC). Results showed that the core–shell catalyst possessed good alkali-poisoning resistance capacity, which was explained well by the micropore model of the catalyst. And the cell performance could keep above 0.75V during 100 h test. When the steam carbon ratio was 2 (S/C = 2) and the current density was 150 mA cm −2, the cell potential varied from 0.826 to 0.751 V and the voltage fluctuant phenomenon was explained specifically. Furthermore, the short stack (three cells) was also assembled, and the maximum output power density of the short stack was 338.4 mW cm −2. The above results indicated that the core–shell catalyst could be applied into the DIR-MCFC successfully.

Preparation of core (Ni base)–shell (Silicalite-1) catalysts and their application for alkali resistance in direct internal reforming molten carbonate fuel cell

Alkali-resistant Ni/SiO 2–Sil-1 and Ni/Al 2O 3–Sil-1 core–shell catalysts were prepared for use in direct internal reforming molten carbonate fuel cell (DIR-MCFC). A thin zeolite shell was grown on the surface of catalyst beads to create a diffusion barrier against alkali poisons in the vapors generated from the electrolyte during DIR-MCFC operation. The synthesis of low defect zeolite shell was investigated and the effects of shell thickness on catalyst activity were examined. A mathematical model of the reaction and alkali-poisoning was developed and the optimum zeolite shell thickness was determined. The experimental and modeling results demonstrated that the core–shell catalyst is more resistant to alkali poisoning and a zeolite shell thickness of 3.5 μm can protect the catalyst for at least 100 h following a failure of the anode barrier in DIR-MCFC to give sufficient time for repair.

The effect of impurities on the performance of bioethanol-used internal reforming molten carbonate fuel cell

The effects of impurities in bioethanol, such as diethyl amine, acetic acid, methanol, and propanol, on the performance of a direct internal reforming molten carbonate fuel cell (MCFC) were investigated. A single cell with an Ni/MgO-coated anode was operated using bioethanol containing 1% impurities. The cell voltages of the single cells increased after the introduction of methanol and diethyl amine impurities. On the other hand, the single cell operated with propanol and acetic acid impurities showed somewhat worse performance than when it was operated without impurities. Activity tests were carried out to explain this phenomenon in detail. These tests showed that methanol as an impurity is reformed easily and increases the H 2 concentration in the reformed gas thereby increasing conversion and selectivity. Propanol, which is reformed incompletely, negatively affects the conversion and H 2 selectivity. Diethyl amine follows a basic reaction pathway and is reformed completely without coke formation, and acetic acid, which follows an acidic pathway, forms significant coke on the catalyst which deactivates the catalytic activity. Mixtures of these four impurities, in varying mole ratios were tested for catalytic activity, and it was found that of the mixtures tested, the one with a ratio 2:1:4:1 (methanol:propanol:diethyl amine:acetic acid) showed the best catalytic activity.

Electrolyte effect on the catalytic performance of Ni-based catalysts for direct internal reforming molten carbonate fuel cell

An active and tolerant Ni-based catalyst for methane steam reforming in direct internal reforming molten carbonate fuel cells (DIR-MCFCs) was developed. Deactivation of reforming catalysts by alkali metals from the electrolyte composed of Li 2CO 3 and K 2CO 3 is one of the major obstacles to be overcome in commercialization of DIR-MCFCs. Newly developed Ni/MgSiO 3 and Ni/Mg 2SiO 4 reforming catalysts show activities of ca. 80% methane conversion. Subsequent to electrolyte addition to the catalyst, however, the activity of Ni/Mg 2SiO 4 decreases to ca. 50% of its initial value, whereas Ni/MgSiO 3 catalyst retains its initial activity. Results obtained from temperature-programmed reduction and X-ray photoelectron spectroscopy identify unreduced Ni 3+ as a decisive factor in keeping catalytic activity from the electrolyte.

Characteristics of alkali-resistant Ni/MgAl 2O 4 catalyst for direct internal reforming molten carbonate fuel cell

Direct internal reforming – molten carbonate fuel cell (DIR–MCFC) has advantages of higher efficiency and smaller size. However, deactivation of the catalyst by alkali carbonate electrolytes poses a significant problem in MCFC. To solve this problem, Ni/MgO and Ni/MgAl 2O 4 catalysts were compulsively mixed with a eutectic mixture of Li 2CO 3 and Na 2CO 3 prior to a methane steam reforming activity test. Activity of Ni/MgO rapidly decreased, while that of Ni/MgAl 2O 4 remained steady due to good alkali resistance. To analyze the effects of alkali addition, N 2 adsorption-desorption, X-ray diffraction, temperature-programmed reduction and oxidation, scanning electron microscopy, and X-ray photoelectron spectroscopy experiments were carried out. Both Ni/MgO and Ni/MgAl 2O 4 showed sintering of Ni and blocking of pores, which reduced the catalytic activity. However, Ni/MgAl 2O 4 showed other positive effects such as stronger metal–support interaction and increased dissociative adsorption.

Optimum start-up strategies for direct internal reforming molten carbonate fuel cell systems

The study of start-up performance for a direct internal reforming molten carbonate fuel cell (DIR-MCFC) system is presented. Since a kW-class stack is assembled with an additional preheating design, the improvement of start-up behavior is conducted to find the proper operating strategy. For a cold start-up fuel cell system, both start-up delay and inverse response are strictly detected. When the optimum operating strategy is determined by solving the steady-state optimization algorithm subject to stack temperature constraint, the rapid system start-up as well as the maximum power output can be achieved simultaneously.

The catalytic performance of Ni/MgSiO 3 catalyst for methane steam reforming in operation of direct internal reforming MCFC

Active and tolerant Ni-based catalyst for methane steam reforming in direct internal reforming molten carbonate fuel cell (DIR-MCFC) was developed. Deactivation of reforming catalysts by alkali metals from electrolyte composed of Li 2CO 3 and K 2CO 3 is one of the major obstacles to be overcome in commercialization of DIR-MCFC. Newly developed Ni/MgSiO 3 reforming catalyst showed activities of ca. 82% methane conversion for 240 min in out-of-cell test. In duration test, the unit cell containing Ni foam impregnated with Ni/MgSiO 3 in anode gas channel did not give performance degradation for more than 2000 h, while the unit cell assembled with Ni/MgSiO 3-coated anode showed a significant performance loss after an operation of 1200 h. Results obtained from X-ray diffraction and Brunauer–Emmett–Teller technique revealed that Ni sintering and support deterioration were decisive factors in decreasing the catalytic activity.

Preparation of alkali-resistant, Sil-1 encapsulated nickel catalysts for direct internal reforming-molten carbonate fuel cell

Abstract A thin layer of silicalite-1 zeolite membrane was grown on the surface of Ni/SiO 2 and Ni/Al 2 O 3 catalyst beads after seeding and secondary regrowth to create core–shell catalysts that are resistant to alkali poisoning from direct internal reforming-molten carbonate fuel cell (DIR-MCFC). The zeolite shell thickness was optimized to prevent poisoning and minimize diffusion resistance. An out-of-cell test was designed to simulate the fuel cell operating conditions, which showed that the new core–shell catalysts maintained a high activity similar to the original fresh catalyst in spite of the exposure to alkali vapor at high temperature. The conventional Ni/SiO 2 and Ni/Al 2 O 3 catalysts suffered higher than 80% decrease in activity for steam reforming of methane reaction (SRM).

Structural Characteristics of (NiMgAl)Ox Prepared from a Layered Double Hydroxide Precursor and its Application in Direct Internal Reforming Molten Carbonate Fuel Cells

Abstract(NiMgAl)Ox materials prepared from layered double hydroxide (LDH) precursors show a typical mesoporous structure with an average pore size of about 6 nm, excellent activity for the methane steam reforming (MSR) reaction at 650 °C, and a strong resistance against Li poisoning, suggesting the possibility for future application in direct internal reforming molten carbonate fuel cells (DIR‐MCFC). X‐ray diffraction (XRD) experiments indicate that the molar ratio of Al/Mg exerts a strong effect on the properties of both the LDH precursors and final catalysts. Temperature‐programmed reduction (TPR) analysis reveals that the Al/Mg molar ratio influences the properties of Ni in the final catalyst. The catalytic performance of the catalysts prepared is greatly influenced by the molar ratio of Al/Mg for MSR. The activity gradually increases with an increase in the Al/Mg ratio (0.14–1.5). However, with further increases in the amount of Al added (Al/Mg = 2), the activity decreases. The activity is strongly related to the BET surface area and Ni dispersion. The Li‐poisoning test proves that Mg‐rich catalysts lose their activity quickly upon exposure to Li, while Al‐rich catalysts maintain virtually all of their original activity. NiMgAl (Al/Mg = 1.5) is found to be an excellent catalyst for the DIR‐MCFC, having both the highest activity and strongest resistance against Li poisoning.

Visualization of electrolyte volatile phenomenon in DIR-MCFC

Volatilization of molten salt is one of the factors that control the performance of molten carbonate fuel cells (MCFC). Volatilization of molten salt promotes cross-leakage and the corrosion of metallic components. Moreover, piping blockage is caused by the solidification of volatile matter. Because reforming catalysts filling the anode channel are polluted by molten salt volatile matter in direct internal reforming molten carbonate fuel cells (DIR-MCFC), the volatilization of molten salt is an especially serious subject. However, neither the behaviour nor the volatilization volume of molten salt volatile matter has heretofore been elucidated on. Because molten salt volatile matter that has strong alkalinity cannot be supplied directly to an analyzer, its volatilization volume is small, and analytical accuracy is poor. Therefore, an attempt has been made to elucidate about the electrolyte volatile phenomenon in an MCFC by using a non-contact image measurement technique. A 16 cm 2 MCFC single cell frame has an observation window and an irradiation window. The image of the volatile phenomenon is shown by irradiating a YAG laser light sheet 2 mm thick from an irradiation window into the anode channel, and taking measurements from an observation window with a high spatial resolution video camera (12 bit). As a result, though the volatile matter is not observed in an anode channel at OCV, the volatile matter flows in a belt-like manner from the inlet side near the electrode toward the outlet at a current density of 150 mA cm −2. In addition, volatile matter is difficult to observe with the conventional thickness of an anode electrode. Because the composition of these volatile matters is 15Li 2CO 3/85K 2CO 3 (the result of conversion into molten salt) by ion chromatography analysis, it is not an electrolyte (62Li 2CO 3/38K 2CO 3) but rather the volatile matter of potassium, such as KOH. Therefore, it is understood that the volatile matter K 2CO 3 is generated as KOH, comprising the water generated by the cellular reaction with an electrolyte, in a reaction with CO 2. Conversely, the volatile matter flows to the surface of the cathode electrode without regard to changes in current density. In addition, the volatile matter exists on the electrode surface, although it decreases more or less with the conventional thickness of the cathode electrode. Therefore, it is understood that the volatile matter, in order for it not to be related to the cellular reaction, does not comprise the dispersion of molten salt.

B214 MCFCの電解質揮発現象の可視化

Volatilizing of molten carbonate as electrolyte in Molten Carbonate Fuel Cell (MCFC) is one of degradation factors. The Volatilization of the molten carbonate brings the cross-leak and the corrosion of the metallic materials. Moreover, the piping blockage caused by solidification of the volatile chemical species in piping will be happened at lowering temperature. Especially, since the reforming catalysts filled to the anode channel in Direct Internal Reforming Molten Carbonate Fuel Cells (DIR-MCFC) are polluted by volatile species from the molten carbonate, volatilizing of the molten carbonate is an important issue. However, the behavior of molten carbonate volatile issue has been not elucidated, since the volatilization volume is too small amount to analyze them. Last symposium, we informed that at anode side, amount of volatilized particle was increasing with increasing current density. Although volatilization phenomenon was observed at cathode side, there was no relationship between amount of volatilized particle and current density. The volatile chemical species from anode and cathode were potassium rich composition.

Simulation of the performance for the direct internal reforming molten carbonate fuel cell. Part I: distributions of temperature, energy transfer and current density

This study examined the temperature distributions of the anode and cathode gases of the cell body as well as the current density distributions at each point of the direct internal reforming molten carbonate fuel cell (DIR-MCFC) using numerical modelling. The model was based on assumptions and experimental data from a 5 cm × 5 cm sized unit cell operation. The results showed there was an approximately 13°C temperature difference between the initial point (0, 0) and end point (1, 1) of the cell body and the temperature increased steadily along with the direction of the anode gas flow. The temperature distribution of the anode gases showed a similar trend to those of the cell body. The temperature of the anode gases was an average 11°C lower than that of the cell body. The temperature distributions of cathode gases were relatively higher than those of the anode gases and the cell body. The temperature distributions at each point of the cell body, including the anode and cathode gases, could be explained by the different rates of the electrochemical, methane steam reforming and water–gas shift reactions at each point in the cell body. The current density distribution at the entrance of the cell was the highest at 290 mA cm−2, and decreased steadily to 150 mA cm−2 at the exit. These results were also confirmed by the amount of hydrogen reacted in the electrochemical reaction (referred to Part II). Finally, modelling simulations showed a non-uniform distribution of the temperature and current density throughout the DIR-MCFC were observed. In addition, it was confirmed that the distributions of the reaction rates and gas compositions at each point of the cell also showed a great deal of difference throughout the DIR-MCFC. The non-uniformity of these temperature distributions can lead to deterioration in the cell performance. These might provide the necessary information for solving these problems. Copyright © 2005 John Wiley & Sons, Ltd.

Simulation of the performance for the direct internal reforming molten carbonate fuel cell. Part II: comparative distributions of reaction rates and gas compositions

This study examined the distributions of the three reaction rates and the compositions of the gases at each point of the unit cell in DIR-MCFC using a numerical simulation. The electrochemical reaction rates at the anode gas entering position were almost two times faster than those at the anode gas outlet position and most of the feeding CH4 reacted in the region from the position x=0 to the position x=0.3. In addition, the water–gas shift reaction became faster from near the half position of the unit cell to the gas outlet position. Therefore, in the rear position of the unit cell, the steam reforming reaction played an important role as a supplementary reaction for providing the H2 needed in the electrochemical reaction. The rates of the two catalytic reactions in the case without the electrochemical reaction were relatively slower than those in the DIR-MCFC. Unlike the distributions of temperature, the current density, gas compositions and the reactions rates at each point of the DIR-MCFC cell, the exit gas compositions from the simulation in particular could be comparative to those of the experimental results. Although there was an approximately 10% difference between both of them, the extent of the difference was considered to be reasonable for this simulation considering the experimental values that could be included in this simulation such as the lower conversion of the reactions, the lower current density and any other values. Copyright © 2005 John Wiley & Sons, Ltd.

Performance of a unit cell equipped with a modified catalytic reformer in direct internal reforming-molten carbonate fuel cell

The performance of a unit cell (50 mm × 50 mm) equipped with a modified direct internal reformer is investigated by comparison with that of a traditional reformer in which the reforming catalyst is loaded constantly throughout the cell. The channel depth ratio of the inlet and outlet in the modified reformer is 1:2, and the depth of the channel increases linearly to the outlet position. Half the amount of the reforming catalyst in the outlet position is loaded in the inlet position. The unit cell with a modified reformer shows good stability in terms of open-circuit voltage and its overall electrical resistance is 94% that of a unit cell equipped with a traditional reformer. In long-term cell performance tests with a cell utilization of 40%, the modified reformer unit cell shows greater durability; it provides 40 more hours of performance than the cell with a traditional reformer. Furthermore, the average temperature of the outlet gases from the modified reformer unit cell over the course of 200 h is lower than that of a traditional reformer unit cell by 8 °C. The potassium and lithium poisoning levels of catalysts in the modified reformer are respectively, 54 and 45% of those in a traditional reformer. The results might be attributed to the different loading amount of the reforming catalyst, leading to a more uniform dispersion of the reforming reaction throughout the cell. This dispersion of the reforming reaction may also lead to the uniform dispersion of other reactions throughout the cell, improving the overall cell performance.

Carbon deposition and alkali poisoning at each point of the reforming catalysts in DIR-MCFC

The extent of local carbon deposition and alkali poisoning of the reforming catalysts in a unit cell (5 × 5 cm) of a direct internal reforming molten carbonate fuel cell was examined. The catalysts in the catalyst bed were sampled from 25 points individually after the unit cell had been operated for both 24 and 100 h, and the amount of carbon, alkali poisoning were then analyzed. There was little difference in the amount of carbon deposition and alkali poisoning with respect to the direction of the cathode gas. They were only dependent on the direction of the anode gas. After 24 h, the amount of carbon deposited on the catalysts loaded in the front region, and the extent of alkali poisoning in the catalysts loaded in the rear region were relatively higher than those of any other regions. After 100 h, the amount of carbon deposition in the catalysts loaded in the front region was still the highest and the amount of alkali poisoning increased rapidly. Meanwhile, relatively low carbon deposition rate and the small increase in alkali poisoning were observed in the catalysts loaded in the rear region. Therefore, the rate of alkali poisoning of the catalyst in the rear region was much slower than those in any other region. These results for catalyst poisoning were explained by a simulation of the unit cell, which was performed to determine the temperature profiles and the extent of the reaction at each point of the unit cell. The simulation showed that the catalysts in the front region treated 90 mol of the initial methane flow, and the temperature was the lowest at this region. Therefore, the catalysts in this region had the highest level of carbon deposition and the rate of alkali poisoning was faster than that of the catalysts in the other regions. This simulation also explained the reasons for the relatively low level of carbon deposition in the rear region, as well as the relatively high rate of alkali poisoning from the start and the low rate of alkali poisoning in the rear region.

Evaluation of volatile behaviour and the volatilization volume of molten salt in DIR-MCFC by using the image measurement technique

The volatilization of molten salt is one of the factors that control the performance of molten carbonate fuel cells (MCFC). Volatilization of molten salt promotes the cross-leakage and corrosion of metallic components. Moreover, pipe blockage is caused by the solidification of volatile matter. Especially, because reforming catalysts filling the anode channel are polluted by molten salt volatile matter in direct internal reforming molten carbonate fuel cells (DIR-MCFC), volatilizing of the molten salt is a weighty subject. However, neither the behaviour nor the volatilization volume of molten salt volatile matter has been elucidated, because molten salt volatile matter that has strong alkalinity cannot be supplied directly to an analyzer, its volatilization volume is small, and the analytical accuracy is poor. Therefore, an attempt was made to elucidate the behaviour of vaporized alkali hydroxide by using a non-contact image measurement technique. The DIR-MCFC electrolyte is generally 62Li 2CO 3/38K 2CO 3. Consideration was given to the DIR-MCFC catalyst pollution mechanism as follows. Molten salt volatile matter is KOH generated as water generated in the cell reacts with the electrolyte. The generated KOH returns to K 2CO 3 again in high CO 2 concentration regions, and catalyst pollution is caused by the adherence of the K 2CO 3 to the catalyst. Moreover, the K 2CO 3 particles mutually cohere when the generated water assists bonding and blocks the piping. The present report experimentally evaluates the volatilization volume of KOH, the change from KOH to K 2CO 3, and the particulate growth of K 2CO 3, using the image measurement technique. In measuring the KOH volatilization volume, K 2CO 3 is generated as KOH volatilized by heating it in a crucible in an electric furnace reacts with CO 2, and is then injected into a reaction tube. The amount of K 2CO 3 is measured by measuring the image of the K 2CO 3 particle with a YAG laser and a CCD camera, thereby obtaining the KOH volatilization volume from the calculation of the stoichiometry of the amount of K 2CO 3. In order to study the change from KOH to K 2CO 3, the particulate growth of K 2CO 3 can be monitored by taking a picture of a K 2CO 3 particle generated with KOH and CO 2 for an extended period. As a result, a conglomerate is generated by the mutual adhesion small particles, causing piping blockage.