Final Report Summary - BEINGENERGY (Integrated low temperature methanol steam reforming and high temperature polymer electrolyte membrane fuel cell)
Executive Summary:
Hydrogen is foreseen to be the energy vector of the future but there are still significant technological barriers to store and transport it, especially in small portable applications. The success of a portable Fuel Cell system depends to a large extent on the fuel supply that should be accomplished in a cost-effective and comfortable manner.
SP1-JTI-FCH.2011.4.4 call frames the challenge as follows: Europe has a narrow technology base in portable fuel cell systems and related fuelling options, which limits opportunities to address a wide number of early market segments which have been shown elsewhere to be attractive for early commercial introduction, such as: construction site tool recharging, emergency and/or remote power, powering recreational applications (for camping, caravanning, boating etc.), personal portable power / powering consumer electronics, and educational devices.
Development of portable power technologies would potentially overlap into applications such as powering autonomous airborne and ground vehicles and small auxiliary power units for vehicles, particularly if the fuelling capability extended across the fuel spectrum to include conventional and renewable liquid fuels.
BeingEnergy project was selected as holding the potential for further developing portable power technologies. Having started in September 2012 and with a 42-month duration, BeingEnergy project aimed to develop a 500 We high-efficiency, cost-effective fuel cell power supply prototype, and prepare its deployment and exploitation. BeingEnergy proposes an integrated low temperature methanol steam reforming and a high temperature polymer electrolyte membrane fuel cell power supply.
The project had 4 objectives, each providing input for the next:
1 - Synthesizing, characterizing, and optimizing of catalysts for low temperature methanol steam reforming (LT-MSR, 180 °C) and the developing of strategies for industrial preparation of the selected catalysts;
2 - Development, characterization and optimization of a cell-reactor for the LT-MSR;
3 - Integration, characterization and optimization of the low temperature methanol steam reforming reactors with a high temperature polymer electrolyte membrane fuel cell (HT-PEMFC);
4 - Development, characterization and optimization of the LT-MSR/HT-PEMFC 500 We prototype.
The BeingEnergy project was able to achieve all the proposed objectives, having as final result a 500 We LT-MSR/HT-PEMFC prototype, fully tested and characterized. The consortium has developed an exploitation strategy, and is now moving towards the mass-production of this system.
Project Context and Objectives:
1-Context
Increasing costs of petroleum production and environmental concerns are driven research into new and clean sources of energy. H2-based energy has long been regarded as a promising alternative to fossil fuel sources. The use of H2 as a fuel from an energy efficiency and emissions viewpoint should be preferably used with fuel cells. H2-based fuel cells turn hydrogen gas into useful electric power with very high efficiencies.
Despite being the best fuel for polymer electrolyte membrane fuel cells (PEMFC) hydrogen is difficult to store and to transport, for this reason fuel processors that can convert liquid fuel into H2 have attracted great interest.
Among the different hydrogen feedstocks available, alcohols are very promising candidates because they are easily decomposed in the presence of water and generate a H2-rich mixture at a relatively lower temperature. Methanol Steam Reforming (MSR) has been extensively studied in recent years due to its several advantages. It has high H/C ratio and low propensity for soot formation and it exists as liquid at room temperature. Therefore, the refuelling system required is compatible with present gasoline distribution infrastructures. Furthermore, it can be produced from renewable sources and it is biodegradable. Methanol can be converted into H2 by different chemical reactions, among that the MSR reaction provides the highest amount of H2.
2-Objectives
Having started in September 2012 and with a 42-month duration, the BeingEnergy project aimed to develop a 500 We high-efficiency, cost-effective fuel cell power supply prototype, and prepare its deployment and exploitation. The system should be compact, easy to use, resulting in a power supply based on the synergetic integration of high temperature polymer electrolyte fuel cell (HT-PEMFC) stack and low temperature methanol steam reforming reactor using a very active new reforming catalyst.
The main objectives of BeingEnergy are hence to:
1 - Synthesizing, characterizing, and optimizing of catalysts for LT-MSR (180 °C) and the developing of strategies for industrial preparation of the selected catalysts;
2 - Development, characterization and optimization of a cell-reactor for the LT-MSR;
3 - Integration, characterization and optimization of the low temperature methanol steam reforming reactors with a high temperature polymer electrolyte membrane fuel cell (HT-PEMFC);
4 - Development, characterization and optimization of the LT-MSR/HT-PEMFC 500 We prototype
To this end the BeingEnergy project assembled a consortium of eight leading organisations in the energy and consultancy sectors with the technical expertise to develop an effective solution, the market knowledge to guarantee that the solution is commercially attractive, and the contact networks to promote it through the industry. The BeingEnergy consortium was a well-balanced structure where each partner has a specific main role: catalyst development (UPORTO, UPVLC-ITQ and Rhodia/Solvay), modelling of reformer reactors (VTT, ITM-CNR), integration of the catalyst and reformer (DLR), development the combined power supply prototype (SerEnergy), and project management and dissemination (INOVA+). Each partner was chosen based on the expertise required to fulfil its role, resulting a complementary of skills that guarantee the success of the BeingEnergy project.
Project Results:
1 - Synthesizing, characterizing, and optimizing of catalysts LT-MSR
1.1 - Context
Methanol Steam Reforming has several advantages and was used as the basis for the system to be developed in BeingEnergy. MSR is accompanied with side reactions (i.e. methanol decomposition), resulting in the production of CO as by-product. Although CO is produced in small amounts, it has a strong impact in the fuel cell performance. The anodic catalyst of fuel cells is extremely sensitive to poisoning with CO. Since CO is formed by the highly endothermic methanol decomposition, lower reaction temperatures would allow working under more favourable thermodynamic conditions and should reduce the yield of CO. While the LT-MSR reaction is an attractive source of nearly CO-free H2, the design of the catalyst is crucial in order to achieve large amounts of H2 with low CO formation.
The catalysts used for MSR can be classified in two main groups: Cu-based and group 8-10 metal-based catalysts (mainly Pd-based catalysts), each having advantages and disadvantages. Cu-based catalysts, which have been commercially produced and successfully used in industry for many years due to their high activity and selectivity, are sensitive to the process conditions. In comparison to copper catalysts, Pd-based catalysts have excellent thermal stability and are not pyrophoric but they are more expensive and less selective than their Cu-based counterparts. Thus, a highly active and selective catalyst for MSR is crucial to achieving a high efficiency compact fuel cell system.
1.2 - Catalyst development
The target of the BeingEnergy project was to perform the MSR reaction at temperatures below 200 °C (ideally 180 ˚C). This objective of low-temperature methanol conversion to hydrogen is of great economic impact, since it will extremely simplify the operation of HT-PEMFCs based on MSR and since the MSR is endothermic and the fuel cell is exothermal it will allow the energy integration of both reactors. The accomplishment of this important goal requires the design of efficient and robust LT-MSR catalysts. Ideally, the catalysts should be highly active in order to originate large amounts of H2, highly selective to H2 production, such that produced CO becomes negligible, and they should present long-term stability.
In the framework of the BeingEnergy project, two very performing type of catalysts for selective H2 production through LT-MSR have been developed: a Cu-based catalyst (CuZrAl) and a supported Pd catalyst (PdCu/Zr-m). The produced catalysts can be applied in different methanol reformers configurations. The CuZrAl catalyst can be directly applied as pellets in a conventional plug flow methanol reformer. The Pd-based catalyst would be more appropriate for a microreformer, where the catalyst could be coated directly onto the reactor walls. In the following, the most relevant findings from both catalytic systems are presented.
1.3 - Cu-based catalysts for LT-MSR
Series of CuZrAl catalysts (with a nominal composition of 65 % CuO and 25% ZrO2) were prepared by a carbonate coprecipitation method and studied for the MSR reaction at 180 ˚C. For the optimization of the CuZrAl catalysts various synthesis parameters, such as: pH, temperature, solution concentration, maturation time, reactants dosing rate and calcination conditions, were carefully studied. This detailed study resulted in the development of a highly performing CuZrAl catalyst. The resulting CuZrAl BeingEnergy catalyst proved to be significantly more active and selective than the benchmark (a CuZnAl catalyst with a nominal composition of 65% CuO and 25% ZnO, denoted as G66-MR and supplied by Süd Chemie) while showing comparable long-term stability. In fact, after 80 h on stream, the CuZrAl BeingEnergy catalyst has about 1.7 times higher activity (per gram of sample) and up to 3-times higher selectivity (meaning that the production of CO is significantly reduced) than that of the reference G66 MR catalyst. This means that the amount of CuZrAl BeingEnergy catalyst required to achieve the same methanol conversion is reduced by about 40 % compared to the G66-MR catalyst, which allows the design of a more compact methanol reformer. These results are very promising and comparable or better than that of the best state-of-the-art MSR catalyst. The physicochemical properties of the CuZrAl BeingEnergy catalyst were deeply investigated by several techniques and compared to those of the G66-MR reference sample.
The results of the detailed characterization study evidence marked differences in the physicochemical properties of both catalysts. It was observed that the CuZrAl BeingEnergy catalyst shows a large surface area, homogeneous particle sizes and high copper dispersion. As a result, the activation of the reactant molecules over both catalysts is significantly different, being easier and/or faster for CuZrAl BeingEnergy catalyst.
1.4 - Pd-based catalysts for LT-MSR
Monoclinic (ZrO2-m) and cubic (ZrO2-c) ZrO2 crystal phases, prepared by a hydrothermal method, were used as supports for monometallic (4 wt.% Pd) and bimetallic (4 wt.% Pd and 20 wt. % Cu) catalysts. The produced Pd-based and CuPd-based catalysts were studied for MSR reaction.
For bimetallic CuPd catalysts the use of ZrO2-m as support resulted in the improvement of both, conversion (an increase of 84 % vs ZrO2-c) and selectivity (the production of CO is notably reduced on CuPd/ZrO2-m, about 40 %). For CuPd/ZrO2-m an activity of 30.1 mmol H2·h-1·g-1 was obtained at 180 ˚C. This value is higher than that reported for supported Pd catalysts and Cu-based systems, and also than that of a commercial CuZnAl catalyst (G66-MR, Süd Chemie, with a nominal mass composition of 65 % copper oxide and 25 % zinc oxide.) tested under similar conditions (23.5 mmol H2·h-1·g-1 at 180 °C).
The new CuPd/ZrO2-m catalyst has activity and selectivity similar to those of G66-MR catalyst despite its lower metal loading, which makes it a good candidate to be used in a coated microreformer. The results from the characterization of this new catalyst suggest that Cu and Pd are strongly interacting when supported on ZrO2-m, resulting in a catalyst made up of highly dispersed CuPd clusters with different electronic properties with respect to those of each metal on their own. The particular “electronic structure” of these CuPd clusters may result in a different interaction with the reactants, leading to different intermediate species and thus different MSR reaction mechanisms. This would be similar to that reported for PdZn and PdGa alloys, where the formation of a specific Pd-alloy promotes the conversion and selectivity towards MRS, as observed for the BeingEnergy CuPd/ZrO2-m catalyst. The conversion enhancement of CuPd/ZrO2-m is attributed to the presence of small metallic particles that led to better dispersion, as evidenced by SEM, XPS and XRD. The low CO selectivity is likely due to the strong interaction between Cu and Pd in the so-called “CuPd-clusters”, (XPS, TPR and XRD), which would behave similarly to an alloy, affecting the MSR reaction mechanism and promoting the formation of CO2 and H2, as reported for Pd alloys catalysts.
2 - Development, characterization and optimization of a cell-reactor for the LT-MSR
2.1 – Context
An innovative Low Temperature Methanol Steam Reformer, palladium membranes for hydrogen separation and CO2 selective materials were studied towards their integration in a HT-PEMFC power supply system. The idea behind the low temperature reformer is taking advantage of the wasted heat by the HT-PEMFC if the operating temperatures of the reformer and fuel cell stack are the same, namely 180 °C. This way the system efficiency can be improved – energy integration. The target hydrogen production of the reformer is 6.5 l/min, corresponding to a 500 We fuel cell system.
2.2 - Kinetic characterization of the BASF catalyst
Commercial catalyst is being used by partner SerEnergy (as whole pellets) in their power supply units, thus it was decided to characterize this catalyst and use the information for the reformer design. The reaction kinetics of methanol steam reforming was therefore obtained within the temperature range of 453 K and 493 K.
The commercial catalyst is produced in pellets, but in order to improve the performance of the catalyst, smaller particle sizes were decided to be implement in the next reformer design. Particle sizes from 50 µm to 1.5 mm of diameter were tested at different temperatures (453 K, 473 K and 493 K). The best compromise between methanol conversion, pressure drop and mechanical strength was found for a range between 250 µm to 400 µm.
2.3 - Kinetic characterization of the BeingEnergy catalysts
Several in-house catalysts have been produced in the WP2 and for those with the highest performances the reaction rate equations were obtained. Two kinetic models were used, power law and mechanistic models. The mechanistic kinetic model was based in the work developed by Peppley et al. (Applied Catalysis A: General, 179 (1999) 31-49).
2.4 - Packed bed reformer
The development of the low temperature reformer started by characterizing commercial catalysts and the in-house developed catalyst in order to determine the conversion efficiency. The results were used in mathematical modelling. This way it was possible to screen different operating conditions to define the reformer performance, and also test different reformer geometries. It was clear that if the operating temperature is decreased notably, heat transfer would be the most important feature in the reformer structure.
Conventional methanol steam reformers typically have a tube-shell structure where thin tubular catalyst channels are within a large tube with heat transfer liquid. In tubular channels, however, heat transfer to the middle part of the catalyst bed is notably lower than to the edges. Regarding this limitation, the new reformer structure was designed to have thin layers of catalyst with heat transfer liquid layers in between. This way the heat transfer distance would remain small to the whole catalyst bed.
The first design for the low temperature steam reformer consisted of stainless steel plates that were stacked together so that they formed 2 mm gap for each catalyst channel and 1 mm gap for the heat transfer liquid channel. These two types of layers would alternate in the structure. Project partner Solvay produced the in-house catalyst. The obtained catalyst conversion efficiency was smaller than expected in the beginning of the project and therefore the reformer volume was set to 1 litre. The expected catalyst density was 600 g/l, and the received amount from Solvay was 700 g. The final amount fitted in the reformer was 575 g with catalyst particle size of 53-180 micrometres. Characterization measurements were performed and the reformer conversion corresponded quite well to the modelling results. Testing at 180 °C did not quite produce the required amount of hydrogen, but when increasing the temperature to 200 °C hydrogen amount was sufficient, namely 8.3 l/min.
The weight of the first design of the packed bed was very high due to the thick stainless steel plates which it was manufactured of. The next phase was to optimize the structure regarding the weight and heat transfer efficiency. The channel/layer type structure was preserved. Multi-physical modelling was utilized in this design work as well. The material was chosen to be aluminium and the core of the new design was formed by multiport extruded aluminium tubes which would act as catalyst channels. The size of the channel openings was 6 mm x 5 mm. Tubes were laser welded to end frames that created 2 mm gaps for the heat exchange liquid. As the catalyst, commercial BASF RP-60 was chosen because it would have been difficult to produce the in-house catalyst is such large amounts. Catalyst pellets were crushed to size of 250-400 µm, and the amount fitted in the reformer was 1.9 kg. Performance measurements were conducted. Full conversion was achieved at slightly lower methanol flow than with the stainless steel version but the heat transfer distance was also larger and the catalyst particles larger. A conventional tube-shell reformer from SerEnergy was also tested, and the aluminium tube reformer performed better than that, though it has to be stated that the SerEnergy reformer used catalyst pellets.
2.5 - Cellular reformer
As another line of development a cellular reformer concept was introduced in the project. Here the reformer would be constructed as cell plates in between the fuel cell stack plates. This way the heat from the fuel cell reaction could be utilized straight on the other side of the plate in the reformer channels. For this type of structure, new bipolar plate material was developed at VTT. This PPS composite could be used as both reformer and fuel cell manufacturing material.
PPS plates were moulded and machined, though the uniformity of the material was not perfect and cracks were formed in some plates during machining. There were also good quality plates which were assembled with crushed commercial BASF RP-60 catalyst into fuel cell end plates, however without other fuel cell components. The heating was electric. The performance of the reformer was measured, but the results were not reliable as the flows in this small plate reformer were notably lower than the packed bed reformer, and therefore the dimensioning of the tubes and flow meters was not adequate. SerEnergy was able to find a commercial producer for the PPS composite and the development work at VTT was stopped.
In the most sophisticated way the integrated device plates would consist of having fuel cell cathode on one side and reformer channels on one side, or having the fuel cell anode on one side and a lid for the reformer channels on the other side. This way the heat losses can be minimized as the heat from the fuel cell reaction on the cathode is directly transported via the common plate to the endothermic reaction taking place in the reformer. The key idea of the optimised (cellular) reformer is to use the high temperature waste heat of the fuel cell stack for the methanol steam reforming in the most efficient way. When reforming is taking place in plates that are positioned inside the stack, in between the fuel cell single cells, the heat transfer losses can be kept at minimum and fuel cell waste heat would have the highest possible temperature.
Different reformer geometries were experimentally tested and modelled using a three-dimension CFD approach. Regarding coupling a MSR reactor with a HT-PEMFC, the multi-serpentine reformer appears to be best approach, due to their simplicity, good flow distributions, good heat transfer and low pressure drop. Two bipolar plates, one of 25 cm2 made in PPS and other of 45 cm2 made of aluminium gold plated were designed, with fuel cell flow field in one side and a reformer flow field in the other.
2.6 - Palladium membranes
The activity at ITM-CNR has been focused on both the experimental and modelling utilization of unsupported and composite Pd-based membranes for application in the methanol steam reforming reaction to produce high grade hydrogen for high temperature PEM fuel cells supplying. Furthermore, silica based membranes have been used for their application in modelling activity in the same topic.
The modelling activities (followed by an experimental validation) have been realized by using 1-D and 2-D mathematical models, while only 1-D models for silica membrane reactors.
As not satisfactory results, from an experimental point of view, it was evidenced that the unsupported Pd-based membrane reactors cannot be useful for carrying out MSR reaction at low temperature (below 300 °C) without having drawbacks for the system (embrittlement phenomenon and CO poisoning of the membrane). Furthermore, the modelling of a silica membrane reactor for MSR reaction showed that the purity of produced hydrogen cannot be satisfactory for HT-PEMFC supplying.
Much attention has been reserved for the composite Pd-membranes for their experimental utilization in membrane reactors. In detail, supported Pd/Al2O3 membranes (with thin dense Pd-layers deposited onto the support) has been characterized to gas permeation tests in order to evaluate the H2/other gas ideal perm-selectivity. Successively, this typology of membranes has been allocated in membrane reactors and methanol steam reforming reaction tests have been performed with the intent of producing as much as possible high grade hydrogen for HT-PEMFC supplying. The main results were interesting in terms of conversion and hydrogen recovery with hydrogen purity higher than 90% and without CO content (less than 20 ppm) and, then, useful for HT-PEMFC supplying.
As a parallel research activity, polymeric membranes have been experimentally tested for their applications in hydrogen enrichment (and contemporary CO2 concentration) of reformed streams coming from a first stage of methanol steam reforming reaction testing.
3 - Integration, characterization and optimization of the low temperature methanol steam reforming reactors with a high temperature polymer electrolyte membrane fuel cell (HT-PEMFC);
3.1 - Experimental System Test
The synergetic integration of the developed low temperature reformer with the high temperature PEMFC was implemented and studied. A test bench has been developed for testing this integration; the test bench was designed in such a manner to allow control of all relevant fluids and also data acquisition of all important system conditions through numerous sensors. In particular, the feed to the reformer, the operating temperature of stack and reformer and the air supply to the fuel cell are controlled. Additionally, operating conditions for the HT-PEMFC like flow rates are also controlled allowing defined detection of I-V curves and utilization of fuel and air.
The 12-cell high temperature polymer electrolyte fuel cell stack was supplied by project partner SerEnergy. The liquid-cooled stack has an active area of 165 cm². For coolant circulation and temperature control a Huber temperature control system was utilized. Typically, HT-PEMFCs are operated at around 160 °C to 180 °C as with higher temperatures the degradation of the cell evolves exponentially. The cooling liquid is used to realize the heat exchange between the fuel cell and the methanol steam reforming reactor. First operation tests of the stack with pure hydrogen have shown a peak performance of 500 We at 0.41 A·cm-2 current density.
In the first run a tube shell reactor supplied by the project partner SerEnergy was used in the experimental system tests. This unit contains a mass of 2.7 kg of commercial BASF RP-60 CuZnO catalyst as pellets. Both components were originally designed to operate at temperatures above
200 °C. As the fuel cell stack maximum operation temperature is 180 °C the tube shell reformer reactor had to be operated at the same or slightly lower temperature level to allow heat transfer from the stack to the reformer. The maximum achievable hydrogen flow from the reactor was 5.5 lmin 1 at full methanol conversion. Therefore, the performance of the reformer was below the required 7.5 l/min to 8.0 l/min of hydrogen to reach 500 Wel gross power output.
The experiments with the tube shell reformer have shown that self-sustained operation of the coupled fuel cell / tube shell reformer system was difficult to achieve without optimizing the overall system design and operation strategy. On the stack side this was mainly due to high heat losses to the environment. To mitigate these losses different design and system control measures have been analysed. Different thermal insulation materials like glass wool and rubber based foams have been tested and compared identifying a concept which increased the amount of heat transferred to the cooling liquid by app. 21 %.
Besides the thermal insulation to the environment heat optimized control strategies of the fuel cell stack have been analysed as well. As fuel and air flow through the bipolar plates of the stack, the heat dissipation via these streams needs to be considered as well. Some HT-PEFC stacks are solely cooled with cathode air without any other liquid cooling. This shows that the heat loss through the cathode air flow is of great importance on the heat balance of the stack.
The heat transfer in the anode and cathode side of the stack is influenced by several factors like stoichiometry, current density, inlet temperature of the fluid and operation temperature of the stack. While the current density is determined by the power request to the system, the stoichiometry is a more flexible parameter. Respecting minimum stoichiometric airflow through the stack to ensure stable operation and minimal cell voltage variation the cathode stoichiometry can be varied in a wide range in HT-PEMFC systems. Tests have shown that by reducing the cathode stoichiometry and therefore the mass flow passing through the stack the amount of heat transferred into the cooling liquid, instead of the cathode air, can be increased by 11 %.
Architectural concepts on system level for increasing the available heat of the fuel cell have also been studied. Therefore, a gas/gas heat exchanger has been integrated in order to minimize the cooling effect of the cathode airflow through the stack by pre-heating the air supply utilizing the thermal energy of the cathode exhaust gas. Experiments have proven that the inlet air temperature can be significantly increased with this concept resulting in a noticeable benefit in available heat from the liquid cooling loop.
The second experimental reformer was designed within the scope of the project by partner VTT on the basis of a plate type heat exchanger. The reformer was built using complex extruded aluminium tubes creating a total volume of 1.8 l for the reforming catalyst. The reformer housing has been manufactured using laser welding allowing for minimal fastener use. Special focus during the design process of the reformer was on the heat transfer between the coolant media and the catalyst bed to ensure a homogenous temperature profile.
As the large-scale production of the BeingEnergy catalyst was still difficult, the consortium decided to equip the reformer with 1.9 kg of commercial RP-60 catalyst. The catalyst has been crushed to particle sizes between 250 and 400 micrometres as study performed by UPORTO has shown that the catalyst performance increases with reduced particle size. Tests with the new reformer design proved its superior performance compared to the tube shell reformer design. With app. 8.0 l/min of maximum hydrogen production at full methanol conversion the new designed reformer performed nearly 1.5 times better using only 70 % of the catalyst amount of the tube shell design.
In combination with the system control and architectural measures described above a self-sustained system operation of the heat-coupled fuel cell / reformer system was successfully demonstrated. Generating 427 Wel of power the fuel cell produced 572 Wth of heat of which nearly 87 % were captured in the cooling liquid leaving only app. 13 % of losses to the environment. A hydrogen production of 8.0 l/min was generated with the help of the fuel cell off heat.
3.2 - System Simulation
Decisions on experimental level were supported by system simulation. This includes one-dimensional models of the core components, which are the methanol steam reforming reactor and the HT-PEFC. Compared to non-dimensional models physical effects of varying gas compositions through the unit could be considered.
A variation of cathode stoichiometry is limited at the low end due to increasing kinetic losses. This means that there is an explicit disaccord between system efficiency and power density. The heat recovery from the cathode off gas by heating the air inlet is an effective measure for reducing the heat losses and additionally levelling out the temperature distribution within the stack. This reduces thermal stress within the bipolar plates, the membrane and the gaskets in the stack and consequently promotes operational lifetime. On the other hand, heat recuperation of the off air requires bulky devices due to low heat transfer coefficients of the fluids. An alternative approach with more compact hardware is the integration of the cathode preheating into the liquid cooling loop of the fuel cell.
Then the ideal order of units downstream the fuel cell is first the reactor requiring the highest temperature available, second the superheating of the fuel water mixture to enter the reformer, third the cathode preheat and finally the cooling radiator which is controlling the outlet temperature of the fuel cell stack. With the simulation it could be shown that the cooling power of the radiator can be fully compensated if the stack is operated at a cathode stoichiometry of 5.3 at 0.4 Acm-2 current density. This means that the radiator unit including the cooling fan with all its mass and power consumption can be fully eliminated. Then the air blower acts in a double role supplying the stack with cathode air and at the same time performing as temperature controller of the stack similar to an air cooled system. This is the basis for a compact system set-up, where the reforming unit as well as the superheating and the cathode air preheating are integrated into a compact design without external pipes.
From an operational point of view there is a theoretical limit of self-sustaining operation of the system. The fuel cell is more efficient in the lower load range; hence less heat is produced. At a current density of 0.14 Acm-2 there is less off heat available from the fuel cell than is required from the methanol steam reforming including the fuel evaporation. If the system is operated in the idle state, a minimum electrical power of 170 Wel has to be consumed. This can either be done by charging a coupled battery system in mobile applications or grid feed in stationary systems.
The system efficiency of the thermal loop coupled reformer fuel cell system is determined between 0.33 and 0.4 at 0.6 and 0.15 Acm-2 respectively with a rather conservative anode stoichiometry setting of 1.25. The system efficiency can be increased effectively by reducing the anode stoichiometry. A stable operation at 1.18 already has been performed experimentally. Additionally, a recovery of the hydrogen in the anode off gas by any means of gas separation could greatly improve the overall system efficiency. Different variations of separation have been implemented in the simulation. Basically there are the two options of H2 or CO2 separation. Both variants involve an additional water loop as sweep gas stream, which in the case of hydrogen separation requires cooling and condensation leading to more complex system architecture with less power density.
In order to enhance the power density of the system a compact fuel cell reformer design has been conceived. After each fuel cell a reforming cell is integrated into the fuel cell stack. With this approach the off heat from the fuel cell is directly used by the reforming cells. A multichannel design has been developed such that the performance of the compact reforming fuel cell unit could be represented by a one dimensional model representing the thermal interaction between the fuel cell stack, the gas channels and the methanol steam reforming channels. All four flow configurations have been investigated. The objective in this context is to find the solution with the most homogenous temperature distribution to avoid thermal stress inside the stack, fuel cell degradation and weak reforming performance. As the reforming process itself only represents about a sixth of the fuel cell off heat if operated at nominal power the heat management in this concept represents a challenge.
Best results were achieved in a counter flow configuration between the cathode and the reforming channel. The effect of anode flow direction was negligible. In a first approach the thermal control was performed by the cathode stoichiometry. Then the system was operated comparable to an air cooled system. To avoid excessive cooling, air preheating is required, such that the catalyst bed in the reforming channel is still correctly tempered to ensure full conversion of the fuel. This leads to high cathode air flows, which on the other hand requires increased channel diameters to reduce pressure drop and compression power reducing the power density of the system.
An alternative way of cooling the system would be the integration of the evaporation into the reforming fuel cell. Therefore, the entry zone of the reforming channel is filled with non-catalytic pellets of high heat conductivity. This would also require increased width of the reforming channels finally resulting in reduced power density.
4 - Development, characterization and optimization of the LT-MSR/HT-PEMFC 500 We prototype
SerEnergy took the various components, as well as the ones developed by themselves, and combined them to create a liquid cooled fuel cell system with liquid heated reformer (LT-MSR/HT-PEMFC) 500 We prototype.
The system was operated for 852 hours in the lab under realistic conditions. System had an average over the life efficiency of 38 % that was significantly above target. Failure to reach more operation hours was fouling of the evaporator and many unintended breakdowns that degraded the system. System lifetime under more optimal conditions is considered to be +1500 hours.
Volume and weight of the system did not reach the desired targets. As the system were to heavy and had to large volume. This was mainly due to the reasons of having a liquid cooled/heated system that demanded more components and a more complex system. And not having the optimized catalyst in large enough amounts. The cost targets of the system were within reach.
Below is a list of first the specifications in the initial proposal and secondly the measured parameters for the final system in operation:
-Nominal power: 350 We; 500 We
-Electrical efficiency: > 30%; 38%
-Operation lifetime: > 1000 h; +1500
-Number of start-stop cycles: > 100; 350
-Volume and weight densities: < 35 kg/kW and < 50 L/kW; 82 kg/kW and 251L/kW
-Mass production price: < 5,000 €/kW; 5656 €/kW @ 1000 pcs per year.
-Environmental operating conditions: -20 and +60 °C (estimated)
The original design of the 350 We generator had a start-up time of 45 minutes. With the current project development of a revised hardware and control software, the 500 W system that was developed in the project have been optimized have improved start-up performance.
The optimized setup consists of an improved catalytic burner design together with a novel design for a burner evaporator that makes it possible to evaporate sufficient fuel during the start-up phase.
The tested setup made it possible to achieve a time to stack operation temperature of 25 minutes. This was short from the desired 15 minutes, and primary caused by 3 factors:
1) A thermal mass of the system that was higher than anticipated. This was due to the new BeingEnergy catalyst had yet not been produced in large batches, so a commercial catalyst was used at 180 ºC and a larger mass of catalyst used, increasing the mass of catalyst.
2) Secondly but also very important was the difficult to circulate the heating fluid when the system was below 50 degrees. This was due to certain viscosity characteristics of the circulating fluid.
3) Third was the time lack for the initial heating of the catalytic monolith that in the current design takes 5 minutes before ignition phase.
During this work a new material for bipolar plates for the fuel cell stack was tested. Bipolar plates consist of a blend of carbon material and a polymer binder. The polymer binder properties are important for the performance and life time of the bipolar plates. Previous plate materials have had problems with corrosion of the polymer binder and hence with time led to porous plates and loss of mechanical strength.
In the project, a certain binder polymer called PPS was investigated, as it was thought to have superior properties to the existing material used. In the preparation phase of the project, commercial PPS composite materials were not commercially available. However, SerEnergy was able to find commercial vendors and order materials for bipolar plates during BeingEnergy, parallel with the development work done by VTT, and test them in their HT-PEMFC stack. Due to the good results, it was decided that the VTT development for the better machinable compound was stopped and efforts were moved to the long-term testing of the commercial PPS material as HT-PEMFC bipolar plate material.
The initial plate material that were used were compound consisting of 85% graphite and 15% polymer of the phenolic resin type.
There were two major reasons to look into alternative materials:
1) The phenolic based materials have a maximum service temperature of 180 ºC. Which is in the borderline with the current service temperature of the fuel cell and therefore insufficient in future perspective, when the membrane technology develops into the 200 º C regime for enhanced performance and better synergies with the Reformer temperature.
2) It was observed that the lifetime of the cells during operation in the stack, were significantly lower than for the single cell test case. Careful examinations revealed that the phenolic resin based materials were microporous, and that these micropores had an affinity to absorb phosphoric acid from the membrane due to capillary pressure forces. Hence leading to an ‘early death’ for the fuel cell.
It was during the project period confirmed that the PPS based composite material had superior qualities compared the existing state of the art material.
The use of the PPS based composite material, improved the lifetime of the fuel cell by a factor of at least 5, from 3000-3500 hours to +16000 hours and probably more since the tests continues after the end of the project period.
It was also concluded that the PPS composite material, was both mouldable into blank sheets and from there machinable, and is mouldable into final shape by the principle of compression moulding.
Another activity in the project period was the development of a simulation tool to simulate the system performance at various operation points and configurations. This was done to optimize the system before building it in real.
A phenomenological model was developed. The model incorporates the energy and mass balances, as well as performance parameters on the fuel cell in terms of a model based on empirical data. The model was used to simulate different load parameters and system efficiencies. And particular attention is put on the role of anode stoichiometry and system efficiencies, as well as the heat flows at different load conditions and system settings.
The external cooling needs as function of various system settings was carefully examined, as the heat balances and temperature levels were known to be of major importance to make the system work.
Some of the major conclusions was that system heat balance depends very much on fuel flows and cathode air flows. As substantial amounts of heat are transferred to the cathode air, and the high amounts of water in the fuel, requires much of the heat generated by the fuel cell to evaporate and be converted in the reformer.
The simulation tool together with the experiments in the lab, have helped in optimizing the system layout and will serve as design tool for future systems
Potential Impact:
Increasing costs of petroleum production and environmental concerns are driven research into new and clean sources of energy. H2-based energy has long been regarded as a promising alternative to fossil fuel sources. The use of H2 as a fuel from an energy efficiency and emissions viewpoint should be preferably used with fuel cells, as H2-based fuel cells turn hydrogen gas into useful electric power with very high efficiencies.
The BeingEnergy project developed a 500 W high-efficiency, cost-effective fuel cell power supply prototype, and is now working to commercialize it. This effort will improve the visibility and feasibility of H2-powered systems, and bring their advantages further into light, contributing to the move away from fossil fuels.
Through the project’s life the consortium performed various dissemination actions, building awareness to the project, and to the H2-based economy. These actions targeted the main stakeholders in the energy sector, as well as industrial and academic ones related to it. Some of these actions were:
-The publication of over a dozen of papers in scientific journals and conferences, based on the research work done in the project;
-Participation at various industrial fairs and other industrial and energy-related events, including the display of information about the project, and in later events, showing the fuel cell prototype. These events include the F-Cell conference, Hannover Messe Fair 2013, 2014, 2015, and others;
-Participation in the 11th and the 12th International Conference on Catalysis in Membrane Reactors (ICCMR) conferences, including a half-hour presentation to a large audience of the project results up to that point.
The dissemination effort the BeingEnergy project performed had a large impact in the building of awareness to the project and its achievements in H2-energy.
List of Websites:
http://www.beingenergy.eu/
Hydrogen is foreseen to be the energy vector of the future but there are still significant technological barriers to store and transport it, especially in small portable applications. The success of a portable Fuel Cell system depends to a large extent on the fuel supply that should be accomplished in a cost-effective and comfortable manner.
SP1-JTI-FCH.2011.4.4 call frames the challenge as follows: Europe has a narrow technology base in portable fuel cell systems and related fuelling options, which limits opportunities to address a wide number of early market segments which have been shown elsewhere to be attractive for early commercial introduction, such as: construction site tool recharging, emergency and/or remote power, powering recreational applications (for camping, caravanning, boating etc.), personal portable power / powering consumer electronics, and educational devices.
Development of portable power technologies would potentially overlap into applications such as powering autonomous airborne and ground vehicles and small auxiliary power units for vehicles, particularly if the fuelling capability extended across the fuel spectrum to include conventional and renewable liquid fuels.
BeingEnergy project was selected as holding the potential for further developing portable power technologies. Having started in September 2012 and with a 42-month duration, BeingEnergy project aimed to develop a 500 We high-efficiency, cost-effective fuel cell power supply prototype, and prepare its deployment and exploitation. BeingEnergy proposes an integrated low temperature methanol steam reforming and a high temperature polymer electrolyte membrane fuel cell power supply.
The project had 4 objectives, each providing input for the next:
1 - Synthesizing, characterizing, and optimizing of catalysts for low temperature methanol steam reforming (LT-MSR, 180 °C) and the developing of strategies for industrial preparation of the selected catalysts;
2 - Development, characterization and optimization of a cell-reactor for the LT-MSR;
3 - Integration, characterization and optimization of the low temperature methanol steam reforming reactors with a high temperature polymer electrolyte membrane fuel cell (HT-PEMFC);
4 - Development, characterization and optimization of the LT-MSR/HT-PEMFC 500 We prototype.
The BeingEnergy project was able to achieve all the proposed objectives, having as final result a 500 We LT-MSR/HT-PEMFC prototype, fully tested and characterized. The consortium has developed an exploitation strategy, and is now moving towards the mass-production of this system.
Project Context and Objectives:
1-Context
Increasing costs of petroleum production and environmental concerns are driven research into new and clean sources of energy. H2-based energy has long been regarded as a promising alternative to fossil fuel sources. The use of H2 as a fuel from an energy efficiency and emissions viewpoint should be preferably used with fuel cells. H2-based fuel cells turn hydrogen gas into useful electric power with very high efficiencies.
Despite being the best fuel for polymer electrolyte membrane fuel cells (PEMFC) hydrogen is difficult to store and to transport, for this reason fuel processors that can convert liquid fuel into H2 have attracted great interest.
Among the different hydrogen feedstocks available, alcohols are very promising candidates because they are easily decomposed in the presence of water and generate a H2-rich mixture at a relatively lower temperature. Methanol Steam Reforming (MSR) has been extensively studied in recent years due to its several advantages. It has high H/C ratio and low propensity for soot formation and it exists as liquid at room temperature. Therefore, the refuelling system required is compatible with present gasoline distribution infrastructures. Furthermore, it can be produced from renewable sources and it is biodegradable. Methanol can be converted into H2 by different chemical reactions, among that the MSR reaction provides the highest amount of H2.
2-Objectives
Having started in September 2012 and with a 42-month duration, the BeingEnergy project aimed to develop a 500 We high-efficiency, cost-effective fuel cell power supply prototype, and prepare its deployment and exploitation. The system should be compact, easy to use, resulting in a power supply based on the synergetic integration of high temperature polymer electrolyte fuel cell (HT-PEMFC) stack and low temperature methanol steam reforming reactor using a very active new reforming catalyst.
The main objectives of BeingEnergy are hence to:
1 - Synthesizing, characterizing, and optimizing of catalysts for LT-MSR (180 °C) and the developing of strategies for industrial preparation of the selected catalysts;
2 - Development, characterization and optimization of a cell-reactor for the LT-MSR;
3 - Integration, characterization and optimization of the low temperature methanol steam reforming reactors with a high temperature polymer electrolyte membrane fuel cell (HT-PEMFC);
4 - Development, characterization and optimization of the LT-MSR/HT-PEMFC 500 We prototype
To this end the BeingEnergy project assembled a consortium of eight leading organisations in the energy and consultancy sectors with the technical expertise to develop an effective solution, the market knowledge to guarantee that the solution is commercially attractive, and the contact networks to promote it through the industry. The BeingEnergy consortium was a well-balanced structure where each partner has a specific main role: catalyst development (UPORTO, UPVLC-ITQ and Rhodia/Solvay), modelling of reformer reactors (VTT, ITM-CNR), integration of the catalyst and reformer (DLR), development the combined power supply prototype (SerEnergy), and project management and dissemination (INOVA+). Each partner was chosen based on the expertise required to fulfil its role, resulting a complementary of skills that guarantee the success of the BeingEnergy project.
Project Results:
1 - Synthesizing, characterizing, and optimizing of catalysts LT-MSR
1.1 - Context
Methanol Steam Reforming has several advantages and was used as the basis for the system to be developed in BeingEnergy. MSR is accompanied with side reactions (i.e. methanol decomposition), resulting in the production of CO as by-product. Although CO is produced in small amounts, it has a strong impact in the fuel cell performance. The anodic catalyst of fuel cells is extremely sensitive to poisoning with CO. Since CO is formed by the highly endothermic methanol decomposition, lower reaction temperatures would allow working under more favourable thermodynamic conditions and should reduce the yield of CO. While the LT-MSR reaction is an attractive source of nearly CO-free H2, the design of the catalyst is crucial in order to achieve large amounts of H2 with low CO formation.
The catalysts used for MSR can be classified in two main groups: Cu-based and group 8-10 metal-based catalysts (mainly Pd-based catalysts), each having advantages and disadvantages. Cu-based catalysts, which have been commercially produced and successfully used in industry for many years due to their high activity and selectivity, are sensitive to the process conditions. In comparison to copper catalysts, Pd-based catalysts have excellent thermal stability and are not pyrophoric but they are more expensive and less selective than their Cu-based counterparts. Thus, a highly active and selective catalyst for MSR is crucial to achieving a high efficiency compact fuel cell system.
1.2 - Catalyst development
The target of the BeingEnergy project was to perform the MSR reaction at temperatures below 200 °C (ideally 180 ˚C). This objective of low-temperature methanol conversion to hydrogen is of great economic impact, since it will extremely simplify the operation of HT-PEMFCs based on MSR and since the MSR is endothermic and the fuel cell is exothermal it will allow the energy integration of both reactors. The accomplishment of this important goal requires the design of efficient and robust LT-MSR catalysts. Ideally, the catalysts should be highly active in order to originate large amounts of H2, highly selective to H2 production, such that produced CO becomes negligible, and they should present long-term stability.
In the framework of the BeingEnergy project, two very performing type of catalysts for selective H2 production through LT-MSR have been developed: a Cu-based catalyst (CuZrAl) and a supported Pd catalyst (PdCu/Zr-m). The produced catalysts can be applied in different methanol reformers configurations. The CuZrAl catalyst can be directly applied as pellets in a conventional plug flow methanol reformer. The Pd-based catalyst would be more appropriate for a microreformer, where the catalyst could be coated directly onto the reactor walls. In the following, the most relevant findings from both catalytic systems are presented.
1.3 - Cu-based catalysts for LT-MSR
Series of CuZrAl catalysts (with a nominal composition of 65 % CuO and 25% ZrO2) were prepared by a carbonate coprecipitation method and studied for the MSR reaction at 180 ˚C. For the optimization of the CuZrAl catalysts various synthesis parameters, such as: pH, temperature, solution concentration, maturation time, reactants dosing rate and calcination conditions, were carefully studied. This detailed study resulted in the development of a highly performing CuZrAl catalyst. The resulting CuZrAl BeingEnergy catalyst proved to be significantly more active and selective than the benchmark (a CuZnAl catalyst with a nominal composition of 65% CuO and 25% ZnO, denoted as G66-MR and supplied by Süd Chemie) while showing comparable long-term stability. In fact, after 80 h on stream, the CuZrAl BeingEnergy catalyst has about 1.7 times higher activity (per gram of sample) and up to 3-times higher selectivity (meaning that the production of CO is significantly reduced) than that of the reference G66 MR catalyst. This means that the amount of CuZrAl BeingEnergy catalyst required to achieve the same methanol conversion is reduced by about 40 % compared to the G66-MR catalyst, which allows the design of a more compact methanol reformer. These results are very promising and comparable or better than that of the best state-of-the-art MSR catalyst. The physicochemical properties of the CuZrAl BeingEnergy catalyst were deeply investigated by several techniques and compared to those of the G66-MR reference sample.
The results of the detailed characterization study evidence marked differences in the physicochemical properties of both catalysts. It was observed that the CuZrAl BeingEnergy catalyst shows a large surface area, homogeneous particle sizes and high copper dispersion. As a result, the activation of the reactant molecules over both catalysts is significantly different, being easier and/or faster for CuZrAl BeingEnergy catalyst.
1.4 - Pd-based catalysts for LT-MSR
Monoclinic (ZrO2-m) and cubic (ZrO2-c) ZrO2 crystal phases, prepared by a hydrothermal method, were used as supports for monometallic (4 wt.% Pd) and bimetallic (4 wt.% Pd and 20 wt. % Cu) catalysts. The produced Pd-based and CuPd-based catalysts were studied for MSR reaction.
For bimetallic CuPd catalysts the use of ZrO2-m as support resulted in the improvement of both, conversion (an increase of 84 % vs ZrO2-c) and selectivity (the production of CO is notably reduced on CuPd/ZrO2-m, about 40 %). For CuPd/ZrO2-m an activity of 30.1 mmol H2·h-1·g-1 was obtained at 180 ˚C. This value is higher than that reported for supported Pd catalysts and Cu-based systems, and also than that of a commercial CuZnAl catalyst (G66-MR, Süd Chemie, with a nominal mass composition of 65 % copper oxide and 25 % zinc oxide.) tested under similar conditions (23.5 mmol H2·h-1·g-1 at 180 °C).
The new CuPd/ZrO2-m catalyst has activity and selectivity similar to those of G66-MR catalyst despite its lower metal loading, which makes it a good candidate to be used in a coated microreformer. The results from the characterization of this new catalyst suggest that Cu and Pd are strongly interacting when supported on ZrO2-m, resulting in a catalyst made up of highly dispersed CuPd clusters with different electronic properties with respect to those of each metal on their own. The particular “electronic structure” of these CuPd clusters may result in a different interaction with the reactants, leading to different intermediate species and thus different MSR reaction mechanisms. This would be similar to that reported for PdZn and PdGa alloys, where the formation of a specific Pd-alloy promotes the conversion and selectivity towards MRS, as observed for the BeingEnergy CuPd/ZrO2-m catalyst. The conversion enhancement of CuPd/ZrO2-m is attributed to the presence of small metallic particles that led to better dispersion, as evidenced by SEM, XPS and XRD. The low CO selectivity is likely due to the strong interaction between Cu and Pd in the so-called “CuPd-clusters”, (XPS, TPR and XRD), which would behave similarly to an alloy, affecting the MSR reaction mechanism and promoting the formation of CO2 and H2, as reported for Pd alloys catalysts.
2 - Development, characterization and optimization of a cell-reactor for the LT-MSR
2.1 – Context
An innovative Low Temperature Methanol Steam Reformer, palladium membranes for hydrogen separation and CO2 selective materials were studied towards their integration in a HT-PEMFC power supply system. The idea behind the low temperature reformer is taking advantage of the wasted heat by the HT-PEMFC if the operating temperatures of the reformer and fuel cell stack are the same, namely 180 °C. This way the system efficiency can be improved – energy integration. The target hydrogen production of the reformer is 6.5 l/min, corresponding to a 500 We fuel cell system.
2.2 - Kinetic characterization of the BASF catalyst
Commercial catalyst is being used by partner SerEnergy (as whole pellets) in their power supply units, thus it was decided to characterize this catalyst and use the information for the reformer design. The reaction kinetics of methanol steam reforming was therefore obtained within the temperature range of 453 K and 493 K.
The commercial catalyst is produced in pellets, but in order to improve the performance of the catalyst, smaller particle sizes were decided to be implement in the next reformer design. Particle sizes from 50 µm to 1.5 mm of diameter were tested at different temperatures (453 K, 473 K and 493 K). The best compromise between methanol conversion, pressure drop and mechanical strength was found for a range between 250 µm to 400 µm.
2.3 - Kinetic characterization of the BeingEnergy catalysts
Several in-house catalysts have been produced in the WP2 and for those with the highest performances the reaction rate equations were obtained. Two kinetic models were used, power law and mechanistic models. The mechanistic kinetic model was based in the work developed by Peppley et al. (Applied Catalysis A: General, 179 (1999) 31-49).
2.4 - Packed bed reformer
The development of the low temperature reformer started by characterizing commercial catalysts and the in-house developed catalyst in order to determine the conversion efficiency. The results were used in mathematical modelling. This way it was possible to screen different operating conditions to define the reformer performance, and also test different reformer geometries. It was clear that if the operating temperature is decreased notably, heat transfer would be the most important feature in the reformer structure.
Conventional methanol steam reformers typically have a tube-shell structure where thin tubular catalyst channels are within a large tube with heat transfer liquid. In tubular channels, however, heat transfer to the middle part of the catalyst bed is notably lower than to the edges. Regarding this limitation, the new reformer structure was designed to have thin layers of catalyst with heat transfer liquid layers in between. This way the heat transfer distance would remain small to the whole catalyst bed.
The first design for the low temperature steam reformer consisted of stainless steel plates that were stacked together so that they formed 2 mm gap for each catalyst channel and 1 mm gap for the heat transfer liquid channel. These two types of layers would alternate in the structure. Project partner Solvay produced the in-house catalyst. The obtained catalyst conversion efficiency was smaller than expected in the beginning of the project and therefore the reformer volume was set to 1 litre. The expected catalyst density was 600 g/l, and the received amount from Solvay was 700 g. The final amount fitted in the reformer was 575 g with catalyst particle size of 53-180 micrometres. Characterization measurements were performed and the reformer conversion corresponded quite well to the modelling results. Testing at 180 °C did not quite produce the required amount of hydrogen, but when increasing the temperature to 200 °C hydrogen amount was sufficient, namely 8.3 l/min.
The weight of the first design of the packed bed was very high due to the thick stainless steel plates which it was manufactured of. The next phase was to optimize the structure regarding the weight and heat transfer efficiency. The channel/layer type structure was preserved. Multi-physical modelling was utilized in this design work as well. The material was chosen to be aluminium and the core of the new design was formed by multiport extruded aluminium tubes which would act as catalyst channels. The size of the channel openings was 6 mm x 5 mm. Tubes were laser welded to end frames that created 2 mm gaps for the heat exchange liquid. As the catalyst, commercial BASF RP-60 was chosen because it would have been difficult to produce the in-house catalyst is such large amounts. Catalyst pellets were crushed to size of 250-400 µm, and the amount fitted in the reformer was 1.9 kg. Performance measurements were conducted. Full conversion was achieved at slightly lower methanol flow than with the stainless steel version but the heat transfer distance was also larger and the catalyst particles larger. A conventional tube-shell reformer from SerEnergy was also tested, and the aluminium tube reformer performed better than that, though it has to be stated that the SerEnergy reformer used catalyst pellets.
2.5 - Cellular reformer
As another line of development a cellular reformer concept was introduced in the project. Here the reformer would be constructed as cell plates in between the fuel cell stack plates. This way the heat from the fuel cell reaction could be utilized straight on the other side of the plate in the reformer channels. For this type of structure, new bipolar plate material was developed at VTT. This PPS composite could be used as both reformer and fuel cell manufacturing material.
PPS plates were moulded and machined, though the uniformity of the material was not perfect and cracks were formed in some plates during machining. There were also good quality plates which were assembled with crushed commercial BASF RP-60 catalyst into fuel cell end plates, however without other fuel cell components. The heating was electric. The performance of the reformer was measured, but the results were not reliable as the flows in this small plate reformer were notably lower than the packed bed reformer, and therefore the dimensioning of the tubes and flow meters was not adequate. SerEnergy was able to find a commercial producer for the PPS composite and the development work at VTT was stopped.
In the most sophisticated way the integrated device plates would consist of having fuel cell cathode on one side and reformer channels on one side, or having the fuel cell anode on one side and a lid for the reformer channels on the other side. This way the heat losses can be minimized as the heat from the fuel cell reaction on the cathode is directly transported via the common plate to the endothermic reaction taking place in the reformer. The key idea of the optimised (cellular) reformer is to use the high temperature waste heat of the fuel cell stack for the methanol steam reforming in the most efficient way. When reforming is taking place in plates that are positioned inside the stack, in between the fuel cell single cells, the heat transfer losses can be kept at minimum and fuel cell waste heat would have the highest possible temperature.
Different reformer geometries were experimentally tested and modelled using a three-dimension CFD approach. Regarding coupling a MSR reactor with a HT-PEMFC, the multi-serpentine reformer appears to be best approach, due to their simplicity, good flow distributions, good heat transfer and low pressure drop. Two bipolar plates, one of 25 cm2 made in PPS and other of 45 cm2 made of aluminium gold plated were designed, with fuel cell flow field in one side and a reformer flow field in the other.
2.6 - Palladium membranes
The activity at ITM-CNR has been focused on both the experimental and modelling utilization of unsupported and composite Pd-based membranes for application in the methanol steam reforming reaction to produce high grade hydrogen for high temperature PEM fuel cells supplying. Furthermore, silica based membranes have been used for their application in modelling activity in the same topic.
The modelling activities (followed by an experimental validation) have been realized by using 1-D and 2-D mathematical models, while only 1-D models for silica membrane reactors.
As not satisfactory results, from an experimental point of view, it was evidenced that the unsupported Pd-based membrane reactors cannot be useful for carrying out MSR reaction at low temperature (below 300 °C) without having drawbacks for the system (embrittlement phenomenon and CO poisoning of the membrane). Furthermore, the modelling of a silica membrane reactor for MSR reaction showed that the purity of produced hydrogen cannot be satisfactory for HT-PEMFC supplying.
Much attention has been reserved for the composite Pd-membranes for their experimental utilization in membrane reactors. In detail, supported Pd/Al2O3 membranes (with thin dense Pd-layers deposited onto the support) has been characterized to gas permeation tests in order to evaluate the H2/other gas ideal perm-selectivity. Successively, this typology of membranes has been allocated in membrane reactors and methanol steam reforming reaction tests have been performed with the intent of producing as much as possible high grade hydrogen for HT-PEMFC supplying. The main results were interesting in terms of conversion and hydrogen recovery with hydrogen purity higher than 90% and without CO content (less than 20 ppm) and, then, useful for HT-PEMFC supplying.
As a parallel research activity, polymeric membranes have been experimentally tested for their applications in hydrogen enrichment (and contemporary CO2 concentration) of reformed streams coming from a first stage of methanol steam reforming reaction testing.
3 - Integration, characterization and optimization of the low temperature methanol steam reforming reactors with a high temperature polymer electrolyte membrane fuel cell (HT-PEMFC);
3.1 - Experimental System Test
The synergetic integration of the developed low temperature reformer with the high temperature PEMFC was implemented and studied. A test bench has been developed for testing this integration; the test bench was designed in such a manner to allow control of all relevant fluids and also data acquisition of all important system conditions through numerous sensors. In particular, the feed to the reformer, the operating temperature of stack and reformer and the air supply to the fuel cell are controlled. Additionally, operating conditions for the HT-PEMFC like flow rates are also controlled allowing defined detection of I-V curves and utilization of fuel and air.
The 12-cell high temperature polymer electrolyte fuel cell stack was supplied by project partner SerEnergy. The liquid-cooled stack has an active area of 165 cm². For coolant circulation and temperature control a Huber temperature control system was utilized. Typically, HT-PEMFCs are operated at around 160 °C to 180 °C as with higher temperatures the degradation of the cell evolves exponentially. The cooling liquid is used to realize the heat exchange between the fuel cell and the methanol steam reforming reactor. First operation tests of the stack with pure hydrogen have shown a peak performance of 500 We at 0.41 A·cm-2 current density.
In the first run a tube shell reactor supplied by the project partner SerEnergy was used in the experimental system tests. This unit contains a mass of 2.7 kg of commercial BASF RP-60 CuZnO catalyst as pellets. Both components were originally designed to operate at temperatures above
200 °C. As the fuel cell stack maximum operation temperature is 180 °C the tube shell reformer reactor had to be operated at the same or slightly lower temperature level to allow heat transfer from the stack to the reformer. The maximum achievable hydrogen flow from the reactor was 5.5 lmin 1 at full methanol conversion. Therefore, the performance of the reformer was below the required 7.5 l/min to 8.0 l/min of hydrogen to reach 500 Wel gross power output.
The experiments with the tube shell reformer have shown that self-sustained operation of the coupled fuel cell / tube shell reformer system was difficult to achieve without optimizing the overall system design and operation strategy. On the stack side this was mainly due to high heat losses to the environment. To mitigate these losses different design and system control measures have been analysed. Different thermal insulation materials like glass wool and rubber based foams have been tested and compared identifying a concept which increased the amount of heat transferred to the cooling liquid by app. 21 %.
Besides the thermal insulation to the environment heat optimized control strategies of the fuel cell stack have been analysed as well. As fuel and air flow through the bipolar plates of the stack, the heat dissipation via these streams needs to be considered as well. Some HT-PEFC stacks are solely cooled with cathode air without any other liquid cooling. This shows that the heat loss through the cathode air flow is of great importance on the heat balance of the stack.
The heat transfer in the anode and cathode side of the stack is influenced by several factors like stoichiometry, current density, inlet temperature of the fluid and operation temperature of the stack. While the current density is determined by the power request to the system, the stoichiometry is a more flexible parameter. Respecting minimum stoichiometric airflow through the stack to ensure stable operation and minimal cell voltage variation the cathode stoichiometry can be varied in a wide range in HT-PEMFC systems. Tests have shown that by reducing the cathode stoichiometry and therefore the mass flow passing through the stack the amount of heat transferred into the cooling liquid, instead of the cathode air, can be increased by 11 %.
Architectural concepts on system level for increasing the available heat of the fuel cell have also been studied. Therefore, a gas/gas heat exchanger has been integrated in order to minimize the cooling effect of the cathode airflow through the stack by pre-heating the air supply utilizing the thermal energy of the cathode exhaust gas. Experiments have proven that the inlet air temperature can be significantly increased with this concept resulting in a noticeable benefit in available heat from the liquid cooling loop.
The second experimental reformer was designed within the scope of the project by partner VTT on the basis of a plate type heat exchanger. The reformer was built using complex extruded aluminium tubes creating a total volume of 1.8 l for the reforming catalyst. The reformer housing has been manufactured using laser welding allowing for minimal fastener use. Special focus during the design process of the reformer was on the heat transfer between the coolant media and the catalyst bed to ensure a homogenous temperature profile.
As the large-scale production of the BeingEnergy catalyst was still difficult, the consortium decided to equip the reformer with 1.9 kg of commercial RP-60 catalyst. The catalyst has been crushed to particle sizes between 250 and 400 micrometres as study performed by UPORTO has shown that the catalyst performance increases with reduced particle size. Tests with the new reformer design proved its superior performance compared to the tube shell reformer design. With app. 8.0 l/min of maximum hydrogen production at full methanol conversion the new designed reformer performed nearly 1.5 times better using only 70 % of the catalyst amount of the tube shell design.
In combination with the system control and architectural measures described above a self-sustained system operation of the heat-coupled fuel cell / reformer system was successfully demonstrated. Generating 427 Wel of power the fuel cell produced 572 Wth of heat of which nearly 87 % were captured in the cooling liquid leaving only app. 13 % of losses to the environment. A hydrogen production of 8.0 l/min was generated with the help of the fuel cell off heat.
3.2 - System Simulation
Decisions on experimental level were supported by system simulation. This includes one-dimensional models of the core components, which are the methanol steam reforming reactor and the HT-PEFC. Compared to non-dimensional models physical effects of varying gas compositions through the unit could be considered.
A variation of cathode stoichiometry is limited at the low end due to increasing kinetic losses. This means that there is an explicit disaccord between system efficiency and power density. The heat recovery from the cathode off gas by heating the air inlet is an effective measure for reducing the heat losses and additionally levelling out the temperature distribution within the stack. This reduces thermal stress within the bipolar plates, the membrane and the gaskets in the stack and consequently promotes operational lifetime. On the other hand, heat recuperation of the off air requires bulky devices due to low heat transfer coefficients of the fluids. An alternative approach with more compact hardware is the integration of the cathode preheating into the liquid cooling loop of the fuel cell.
Then the ideal order of units downstream the fuel cell is first the reactor requiring the highest temperature available, second the superheating of the fuel water mixture to enter the reformer, third the cathode preheat and finally the cooling radiator which is controlling the outlet temperature of the fuel cell stack. With the simulation it could be shown that the cooling power of the radiator can be fully compensated if the stack is operated at a cathode stoichiometry of 5.3 at 0.4 Acm-2 current density. This means that the radiator unit including the cooling fan with all its mass and power consumption can be fully eliminated. Then the air blower acts in a double role supplying the stack with cathode air and at the same time performing as temperature controller of the stack similar to an air cooled system. This is the basis for a compact system set-up, where the reforming unit as well as the superheating and the cathode air preheating are integrated into a compact design without external pipes.
From an operational point of view there is a theoretical limit of self-sustaining operation of the system. The fuel cell is more efficient in the lower load range; hence less heat is produced. At a current density of 0.14 Acm-2 there is less off heat available from the fuel cell than is required from the methanol steam reforming including the fuel evaporation. If the system is operated in the idle state, a minimum electrical power of 170 Wel has to be consumed. This can either be done by charging a coupled battery system in mobile applications or grid feed in stationary systems.
The system efficiency of the thermal loop coupled reformer fuel cell system is determined between 0.33 and 0.4 at 0.6 and 0.15 Acm-2 respectively with a rather conservative anode stoichiometry setting of 1.25. The system efficiency can be increased effectively by reducing the anode stoichiometry. A stable operation at 1.18 already has been performed experimentally. Additionally, a recovery of the hydrogen in the anode off gas by any means of gas separation could greatly improve the overall system efficiency. Different variations of separation have been implemented in the simulation. Basically there are the two options of H2 or CO2 separation. Both variants involve an additional water loop as sweep gas stream, which in the case of hydrogen separation requires cooling and condensation leading to more complex system architecture with less power density.
In order to enhance the power density of the system a compact fuel cell reformer design has been conceived. After each fuel cell a reforming cell is integrated into the fuel cell stack. With this approach the off heat from the fuel cell is directly used by the reforming cells. A multichannel design has been developed such that the performance of the compact reforming fuel cell unit could be represented by a one dimensional model representing the thermal interaction between the fuel cell stack, the gas channels and the methanol steam reforming channels. All four flow configurations have been investigated. The objective in this context is to find the solution with the most homogenous temperature distribution to avoid thermal stress inside the stack, fuel cell degradation and weak reforming performance. As the reforming process itself only represents about a sixth of the fuel cell off heat if operated at nominal power the heat management in this concept represents a challenge.
Best results were achieved in a counter flow configuration between the cathode and the reforming channel. The effect of anode flow direction was negligible. In a first approach the thermal control was performed by the cathode stoichiometry. Then the system was operated comparable to an air cooled system. To avoid excessive cooling, air preheating is required, such that the catalyst bed in the reforming channel is still correctly tempered to ensure full conversion of the fuel. This leads to high cathode air flows, which on the other hand requires increased channel diameters to reduce pressure drop and compression power reducing the power density of the system.
An alternative way of cooling the system would be the integration of the evaporation into the reforming fuel cell. Therefore, the entry zone of the reforming channel is filled with non-catalytic pellets of high heat conductivity. This would also require increased width of the reforming channels finally resulting in reduced power density.
4 - Development, characterization and optimization of the LT-MSR/HT-PEMFC 500 We prototype
SerEnergy took the various components, as well as the ones developed by themselves, and combined them to create a liquid cooled fuel cell system with liquid heated reformer (LT-MSR/HT-PEMFC) 500 We prototype.
The system was operated for 852 hours in the lab under realistic conditions. System had an average over the life efficiency of 38 % that was significantly above target. Failure to reach more operation hours was fouling of the evaporator and many unintended breakdowns that degraded the system. System lifetime under more optimal conditions is considered to be +1500 hours.
Volume and weight of the system did not reach the desired targets. As the system were to heavy and had to large volume. This was mainly due to the reasons of having a liquid cooled/heated system that demanded more components and a more complex system. And not having the optimized catalyst in large enough amounts. The cost targets of the system were within reach.
Below is a list of first the specifications in the initial proposal and secondly the measured parameters for the final system in operation:
-Nominal power: 350 We; 500 We
-Electrical efficiency: > 30%; 38%
-Operation lifetime: > 1000 h; +1500
-Number of start-stop cycles: > 100; 350
-Volume and weight densities: < 35 kg/kW and < 50 L/kW; 82 kg/kW and 251L/kW
-Mass production price: < 5,000 €/kW; 5656 €/kW @ 1000 pcs per year.
-Environmental operating conditions: -20 and +60 °C (estimated)
The original design of the 350 We generator had a start-up time of 45 minutes. With the current project development of a revised hardware and control software, the 500 W system that was developed in the project have been optimized have improved start-up performance.
The optimized setup consists of an improved catalytic burner design together with a novel design for a burner evaporator that makes it possible to evaporate sufficient fuel during the start-up phase.
The tested setup made it possible to achieve a time to stack operation temperature of 25 minutes. This was short from the desired 15 minutes, and primary caused by 3 factors:
1) A thermal mass of the system that was higher than anticipated. This was due to the new BeingEnergy catalyst had yet not been produced in large batches, so a commercial catalyst was used at 180 ºC and a larger mass of catalyst used, increasing the mass of catalyst.
2) Secondly but also very important was the difficult to circulate the heating fluid when the system was below 50 degrees. This was due to certain viscosity characteristics of the circulating fluid.
3) Third was the time lack for the initial heating of the catalytic monolith that in the current design takes 5 minutes before ignition phase.
During this work a new material for bipolar plates for the fuel cell stack was tested. Bipolar plates consist of a blend of carbon material and a polymer binder. The polymer binder properties are important for the performance and life time of the bipolar plates. Previous plate materials have had problems with corrosion of the polymer binder and hence with time led to porous plates and loss of mechanical strength.
In the project, a certain binder polymer called PPS was investigated, as it was thought to have superior properties to the existing material used. In the preparation phase of the project, commercial PPS composite materials were not commercially available. However, SerEnergy was able to find commercial vendors and order materials for bipolar plates during BeingEnergy, parallel with the development work done by VTT, and test them in their HT-PEMFC stack. Due to the good results, it was decided that the VTT development for the better machinable compound was stopped and efforts were moved to the long-term testing of the commercial PPS material as HT-PEMFC bipolar plate material.
The initial plate material that were used were compound consisting of 85% graphite and 15% polymer of the phenolic resin type.
There were two major reasons to look into alternative materials:
1) The phenolic based materials have a maximum service temperature of 180 ºC. Which is in the borderline with the current service temperature of the fuel cell and therefore insufficient in future perspective, when the membrane technology develops into the 200 º C regime for enhanced performance and better synergies with the Reformer temperature.
2) It was observed that the lifetime of the cells during operation in the stack, were significantly lower than for the single cell test case. Careful examinations revealed that the phenolic resin based materials were microporous, and that these micropores had an affinity to absorb phosphoric acid from the membrane due to capillary pressure forces. Hence leading to an ‘early death’ for the fuel cell.
It was during the project period confirmed that the PPS based composite material had superior qualities compared the existing state of the art material.
The use of the PPS based composite material, improved the lifetime of the fuel cell by a factor of at least 5, from 3000-3500 hours to +16000 hours and probably more since the tests continues after the end of the project period.
It was also concluded that the PPS composite material, was both mouldable into blank sheets and from there machinable, and is mouldable into final shape by the principle of compression moulding.
Another activity in the project period was the development of a simulation tool to simulate the system performance at various operation points and configurations. This was done to optimize the system before building it in real.
A phenomenological model was developed. The model incorporates the energy and mass balances, as well as performance parameters on the fuel cell in terms of a model based on empirical data. The model was used to simulate different load parameters and system efficiencies. And particular attention is put on the role of anode stoichiometry and system efficiencies, as well as the heat flows at different load conditions and system settings.
The external cooling needs as function of various system settings was carefully examined, as the heat balances and temperature levels were known to be of major importance to make the system work.
Some of the major conclusions was that system heat balance depends very much on fuel flows and cathode air flows. As substantial amounts of heat are transferred to the cathode air, and the high amounts of water in the fuel, requires much of the heat generated by the fuel cell to evaporate and be converted in the reformer.
The simulation tool together with the experiments in the lab, have helped in optimizing the system layout and will serve as design tool for future systems
Potential Impact:
Increasing costs of petroleum production and environmental concerns are driven research into new and clean sources of energy. H2-based energy has long been regarded as a promising alternative to fossil fuel sources. The use of H2 as a fuel from an energy efficiency and emissions viewpoint should be preferably used with fuel cells, as H2-based fuel cells turn hydrogen gas into useful electric power with very high efficiencies.
The BeingEnergy project developed a 500 W high-efficiency, cost-effective fuel cell power supply prototype, and is now working to commercialize it. This effort will improve the visibility and feasibility of H2-powered systems, and bring their advantages further into light, contributing to the move away from fossil fuels.
Through the project’s life the consortium performed various dissemination actions, building awareness to the project, and to the H2-based economy. These actions targeted the main stakeholders in the energy sector, as well as industrial and academic ones related to it. Some of these actions were:
-The publication of over a dozen of papers in scientific journals and conferences, based on the research work done in the project;
-Participation at various industrial fairs and other industrial and energy-related events, including the display of information about the project, and in later events, showing the fuel cell prototype. These events include the F-Cell conference, Hannover Messe Fair 2013, 2014, 2015, and others;
-Participation in the 11th and the 12th International Conference on Catalysis in Membrane Reactors (ICCMR) conferences, including a half-hour presentation to a large audience of the project results up to that point.
The dissemination effort the BeingEnergy project performed had a large impact in the building of awareness to the project and its achievements in H2-energy.
List of Websites:
http://www.beingenergy.eu/
