Technical Problems To Be Solved Before The Solid Oxide Fuel Cell Will Be Commercialized

C. Bagger0, N. Christiansen2', P.V. Hendriksen", E.J. Jensen2', S.S. Larsen2', M. Mogensen"

"Materials Department, Ris0 National Laboratory, P.O.BOX 49, DK-4000 Roskilde, Denmark

2' Haldor Tops0e A/S, Nym0lIevej 55, DK-2800 Lyngby, Denmark

1. INTRODUCTION

The problems which must be solved before SOFC-systems are competitive with todays power production technology are of both technical and economical nature. The cost of SOFC stacks at the 25 kW level of today is about 30,000 ECU/kW IV and it is bound to come down to about 500 ECU/kW. The allowable cost of a SOFC system is anticipated to be around 1500 ECU/kW.

As part of the Danish SOFC program (DK-SOFC) a 0.5 kW stack was built and tested during the second half of 1995 121. Based upon the experience gained, an economic analysis has been made. The tools required to approach an economically acceptable solution are outlined below.

2. ECONOMIC ANALYSIS

The Danish 0.5 kW SOFC stack was built according to the Bipolar Flat Plate SOFC (BFP-SOFC) concept. 50 cells, each with an active area of 50 cm2, produced the targeted 0.5 kW with an overall area specific internal resistance (ASR) of 1 Si'cm2 at 1000°C. Cheap ceramic techniques with a potential for upscaling had been selected for the technological development. The 8x8 cm electrolyte was tape cast from high purity zirconia (8 mole% yttria) to a sintered thickness of 160p.m. 5-10pm composite electrode layers with La(Sr)MnOj(LSM) /3/ and NiO /4/, respectively, were deposited by spray painting of ceramic slurries and sintered. Current collecting layers were made from LSM by tape casting and from NiO/YSZ by spray painting. Ceramic interconnects were fabricated from La(Sr)Cr(V)03(LSCV) 151 by uniaxial pressing and sintering at high temperature. The plates were machined with diamonds to give suitable surfaces and cross flow gas channels. The stack was assembled from quality controlled 10-celI substacks.

Based upon the fabrication experience of the first stack, costs were calculated to 3,000-10,000 ECU/kW, depending on the area specific internal stack resistance. The lower limit of the fabrication cost corresponds to 0.4 Qcm2. The materials cost, including actual losses during the various fabrication processes and components rejected by quality controls, constitutes -50% (Fig. 1). The remaining 50% are the cost of capital and manpower in a pilot plant with a total production capacity of 2 MW SOFC stacks per year. In spite of the low-cost ceramic techniques chosen, the stack cost requires a reduction of more than 10 times before it is of commercial interest (500 ECU/kW). A number of tools are available for reduction:

1. Redesign of identified high-cost elements

2. Cheaper materials where acceptable

3. Reduction of the number of components in the repetitive unit of the stack

4. Reduction of losses during fabrication

5. Automation of the production

The interconnect is by far the most expensive individual stack element. The requirement of gas tightness implicates low materials porosity and the presence of gas channels makes the stack ele-

CapitalCosI 24,1%

CapitalCosI 24,1%

Fig. 1. Distribution of present stack fabrication cost

ment the most voluminous one, altogether maximizing the consumption of the relatively costly lanthanum chromite. The thickness and the layout with channels in cross flow configuration limit the range of applicable ceramic shaping techniques. Increasing demands to pla-narity minimizes in-plane resistance losses in the electrodes but introduces expensive mechanical grinding after sintering. Reduction of the interconnect element cost clearly requires reconsideration of the design, leading to a lower complexity of the element.

Materials of high purity are usually used for the development of new materials and components in laboratory scale. Commercialization, however, necessitates determination of the importance of purity to enable substitution of frequently used materials with cheaper ones. The cost of pure lanthanum compounds is high. The relatively high content in a stack (Fig. 2) makes the use of much cheaper lanthanide mixtures desirable, and a substitution seems to be acceptable in current collecting elements 161. The cost of yttrium stabilized zirconia is high, too, when the content of silica, which is detrimental to ion conductivity, is to be kept at a low level. Additions to YSZ which may neutralize the effect of Si in cheaper zirconia are desired.

Generally, losses were high during the laboratory scale fabrication of the stack elements. Mass production and automation will of course reduce the waste fraction somewhat, but losses are to a significant extent inherent to the individual ceramic technique (cut-away green tape, slurry remnants). It is therefore important to establish recycling of waste in the fabrication processes.

A stack cost reduction factor of 3 may be realized as a result of identified possible improvements of the technology. The relative cost distribution in Fig. 3 shows that materials cost is reduced below 25% of the total, while capital investment cost increases to above 40%.

Fig. 2. Distribution of present materials cost
0 0

Post a comment