Status Ofsofc Development At Siemens

W. Drenckhahn, L. Blum, H. Greiner Siemens AG, Power Generation Group Freyeslebenstr. 1,91058 Erlangen, Germany

The Siemens SOFC development programme reached an important milestone in June 1995. A stack operating with hydrogen and oxygen produced a peak power of 10.7 kW at a current density of 0.7 A/cm2 and was running for more than 1400 hours. The SOFC configuration is based on a flat metal separator plate using the multiple cell array design. Improved PENs, functional layer and joining technique were implemented. Based on this concept, a 100 kW plant was designed.

The SOFC development at Siemens has been started in 1990 after a two years preparation phase. The first period with the goal of the demonstration of a 1 kW SOFC stack operation ended in 1993. This important milestone was finally reached in the begin of 1994. The second project phase with the final milestone of a 20 kW module operation will terminate at the end of 1996. This result will form a basis for the next phase in which a 50 to 100 kW pilot plant will be built and tested.

The planar design of the Siemens high-temperature fuel cell combines metallic and ceramic materials and allows high power densities.

A fuel cell stack consists of two metallic end plates and several bipolar plates containing channels which direct the process gases to the electrochemically active elements. A main task has been to adapt the thermal expansion coefficient of the metallic plate to that of the electrolyte made of 8YSZ and at the same time attaining high corrosion resistance. These goals have been reached with a new metal alloy called CrFe5Y203l, which has been developed together with an Austrian partner, Metallwerke Plansee. This metallic bipolar plate has good electrical conductivity making it especially well-suited for high current densities, and produces a very uniform temperature distribution due to its good thermal conductivity (1). Its high mechanical strength at elevated temperatures permits the manufacture of large plates as well as large-volume stacks consisting of numerous individual cells.

A characteristic of the Siemens design is the multiple cell array concept (2) which depends on the manufacture of large plates as mentioned above. It allows the arrangement of several ceramic single cell elements (so-called PENs) in parallel in one layer. By this measure larger electrode areas can be realized in one stack. The actual bipolar plate has dimensions of260 x 260 mm2. This allows the placement of 16 PENs of the size 50 x 50 mm2 in parallel in one layer. The total electrode area per layer is 256 cm2.

On the anode side, a Ni-grid works as functional layer to provide good electrical contact. For this purpose LaCoOj is used on the cathode side. This perovskite is applied by wet spraying of a powder with a certain grain size by an air brush method. This results in a deformable layer compensating the thickness differences of the large number of parts which have to be assembled in one layer in parallel as well as in series.

Decreasing the internal cell resistance, mainly by reduction of the polarization losses at the electrode/electrolyte interfaces, has led to improved cell behaviour with a power output of 0.9 W/cm2 (current density of 1.2 A/cm2 at a cell voltage of 0.75 V, reacting hydrogen and air at 950 °C). Using a mixture of hydrogen with 50 % water vapour a power density of 0,6 W/cm2 was reached. This power output is reduced by about 45 %, if the operating temperature is lowered to 850 °C. The different current-voltage relations are illustrated in fig. 1.

current density (/(A/cm2)

Fig. 1:1/U Curves of Improved Cells with Different Gases, at 850 and 950 °C

current density (/(A/cm2)

Fig. 1:1/U Curves of Improved Cells with Different Gases, at 850 and 950 °C

Parallel to the improvement of current-voltage relations, also the long term behaviour has been investigated. Changes in electrode preparation and micro structure resulted in degradation rates of less than 2 % in 1000 hours, tested at 950 °C in ceramic housing in H2/H20-atmosphere of 1:1. At 850 °C no degradation could be observed over a period of 4500h. In a first long term test at 850 °C in a metallic housing, a degradation rate of 2 % per lOOOh was observed during 2000h of operation. First tests at 950 °C, using a protective layer against chromium evaporation showed similar degradation rates as tests in ceramic housings (3).

A very important step towards manufacturing of bigger stacks has been the transfer of the PEN manufacturing from laboratory scale to a pilot plant. In this plant, 30000 electrolytes of the size 50 x 50 mm2 or 10000 parts of the size 100 x 100 mm2 can be manufactured per year. The capacity of screen printing and sintering of electrodes is slightly higher. This means PENs for a module up to 100 kVV power (operating with air and 80% fuel utilization) can be manufactured per year. The production of PENs with the size 100 x 100 mm2 has been started in the beginning of 1996.

Based on the design described above, in June 1995 a stack with 80 cell layers has been assembled, each layer consisting of 16 parallel PENs, which means 1280 PENs with a total electrode area of about 2 m2. The stack had dimensions of260 x 260 x 260 mm3, including end plates. After brazing and anode reduction the stack performed well. The open circuit voltage of the stack was 104 V. The mean voltage of a single layer was 1.3 V, which indicates that all PENs have been sealed gas tight to the bipolar plates. The initial loading of the stack showed a power output of 10.7 kW (corresponding stack voltage 60 V, mean cell voltage 0,75 V ± 50 mV) at 950 °C and gas utilisation of 50 %. Further loading of the stack was limited by the electrical equipment of the test stand. Operating with hydrogen and air gave a power output of 5.4 kW. The power output at 850 °C was still 4.1 kW. The used PENs were of an older type than described above. The different current-voltage relations of the stack are plotted in fig. 2.

Afler a number of different I/U-measurements, the stack was operated at constant load of 200 -300 mA/cm2 for 1000 hours at 850 °C. During this time, the stack showed a relatively high degradation rate and compared to the initial performance a power loss of 19 % was observed at the designed load of 270 mA/cm2. Afterwards, a thermal cycle of 950 °C/RT/950 °C was run without serious damage to the stack.

Based on these results a 20 kW system was designed and built. It is suited for the operation with hydrogen and air (with the possibility to shift to oxygen).

It will be a self-sustaining system down to a power output of about 10 kW. Therefore, the heat loss to the environment was reduced to about 4 kW. The test facility is designed in a manner that 4 stacks can be operated in parallel. It is controlled by a stored programme computer. Fuel gas flow is adjusted proportional to the load and the air flow regulated so as to keep the max. stack temperature constant. The gases are preheated by the hot waste gases in specially designed recuperative heat exchangers. The plant will be put into operation at the end of August 96. Then the integration and test of the stacks will follow.

The design described above is used as a basis for the layout of bigger stacks and modules. The design and manufacturing of bipolar plates of the size 360 x 360 mm2 has been started, implementing 9 PENs of the size 100 x 100 mm2 with a total electrode area of 729 cm2 in one layer.

Based on this, it is planned to build up a 50 to 100 kW module consisting of 25 kW stacks till end of 1998. This results in stacks with a height of about 0.5 m.

Investigations on system behaviour and system calculations have led to a flow scheme of a lOOkW combined heat and power plant with an electrical efficiency of about 50% and a total efficiency above 90% (fig. 3). These values are based on the use of air and natural gas with internal reforming.


The test results described above proved the feasibility of manufacturing and operating larger stacks based on the multiple cell design using metallic bipolar plates.

The development of PENs with high power density shows the potential of this SOFC technology. The aim must be to reach these values as near as possible under real operating conditions and to further improve the long term stability.

High electrical and system efficiencies can be attained, even for plants with a power output of 100 kW. This represents a great advantage compared to existing technologies.


The European Community (research contract J01JE-CT-92-0105), the German Ministry of Education, Research and Development (BMBF) and the MWMT of Nordrhein-Westfalen are greatly acknowledged for financial support.


1. H. Greiner, T. GrSgler, W. KOck, R. Singer, Chromium Based Alloys for High Temperature SOFC Applications, 4th International Symposium on SOFC, 1995

2. H. Greiner, E. Ivers-Tiffee, W. Wersing, European Patent 0 425 939 B1

3. E. Fendler, R. Henne, R. Ruckdaschel, H. Schmidt, Protecting Layers for the Bipolar Plates of Planar SOFCs Produced by VPS, 2nd European SOFC Forum, 06-10 May 1996, Oslo/ Norway


Yasuo Miyake, Yukinori Akiyama, Takashi Yasuo, Shunsuke Taniguchi, Masataka Kadowaki and Koji Nishio

New Materials Research Center, SANYO Electric Co... Ltd. 1-1, Dainichi-higashimachi. Moriguchi City, Osaka 570, Japan Telephone: +81-6-900-3553 (Direct) Facsimile: +81-6-900-3556 E-Mail: [email protected]

Hirokazu Sasaki

New Energy and Industrial Technology Development Organization (NEDO)

0 0

Post a comment