02.05.2014 Views

Nuclear Production of Hydrogen, Fourth Information Exchange ...

Nuclear Production of Hydrogen, Fourth Information Exchange ...

Nuclear Production of Hydrogen, Fourth Information Exchange ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

CAUSES OF DEGRADATION IN A SOLID OXIDE ELECTROLYSIS STACK<br />

Introduction<br />

<strong>Hydrogen</strong> is considered an alternative transportation fuel because its high gravimetric energy density<br />

and its reaction product is water, a clean alternative to carbon oxides from fossil fuels. Although<br />

hydrogen is the most abundant element, its diatomic gas cannot be mined or captured terrestrially.<br />

<strong>Hydrogen</strong> gas is derived mainly from reforming fossil fuels or by the electrolysis <strong>of</strong> water. A High<br />

Temperature Steam Electrolysis (HTSE) process based on Solid Oxide Fuel Cell (SOFC) technology<br />

(Baukal, 1976; Doenitz, 1982; Salzano, 1985; Hino, 2004) and utilising electricity and waste heat<br />

is presently being developed by the US Department <strong>of</strong> Energy’s <strong>Nuclear</strong> <strong>Hydrogen</strong> Initiative<br />

(Herring, 2007).<br />

In the HTSE process, multiple solid oxide cells are arranged in “stacks”, identical to stacks used in<br />

SOFC systems, but unlike a SOFC, heat and a positive potential are applied to the stack to reduce the<br />

steam to hydrogen; these being called Solid Oxide Electrolysis Cells (SOEC). Stacks <strong>of</strong> cells provide for<br />

a manageable voltage/current input as well as a manufacturable cell size. However, stacks necessarily<br />

have an increased cell area and a multiplicity <strong>of</strong> components that introduce a number <strong>of</strong> sources <strong>of</strong><br />

degradation that take place during operation. It has been observed that SOEC stacks show an<br />

unusually large degradation rate compared to a SOFC stack (Mawdsley, 2009).<br />

Argonne National Laboratory has analysed post-test, several cells and stack components run as<br />

SOEC to determine the major causes <strong>of</strong> degradation. Resistance and elemental maps <strong>of</strong> stack component<br />

surfaces where initially created using respectively, a four-point electrical probe and X-ray fluorescence<br />

measurements at Argonne’s Advanced Photon Source. In addition, Raman micro-spectroscopy and<br />

X-ray absorption spectroscopy were utilised to determine phase structure and oxidation state <strong>of</strong> the<br />

main constituents and elements <strong>of</strong> stack components. These maps were used to identify regions <strong>of</strong><br />

probable degradation. From this information, stack components were cross-sectioned in the areas<br />

<strong>of</strong> interest and analysed by scanning electron microscopy (SEM) and energy dispersive X-ray<br />

spectroscopy (EDS). Results <strong>of</strong> the SEM/EDS work are presented here.<br />

Experimental<br />

Cells and stacks were fabricated by Ceramatec, Inc. (Salt Lake City), and tested at Ceramatec and INL.<br />

After testing, the stacks were disassembled and several repeat units (cells, interconnects and gas flow<br />

fields) were sent to Argonne for evaluation. Stack components from the following tests were<br />

evaluated: i) a ten-cell stack operating for 200 h; ii) a 25-cell stack operating for 1 000 h, both stacks<br />

described by O’Brien (2007); iii) a half-ILS module consisting <strong>of</strong> two 60-cell stacks, operating for 2 040 h<br />

(Hartvigsen, 2007); iv) a ILS module containing four stacks <strong>of</strong> 60 cells each that operated for 420 h. This<br />

article focuses on SEM and EDS analysis <strong>of</strong> components from the ½ILS module, including some<br />

observations from other tests.<br />

A typical stack repeat unit consists <strong>of</strong> a ferritic stainless steel interconnect plate including edge<br />

rails, seals and corrugated gas flow fields attached to each side <strong>of</strong> the plate (O’Brien, 2007). Figure 1<br />

shows a composite SEM image <strong>of</strong> the other essential components <strong>of</strong> a stack repeat unit. Beginning at<br />

the top <strong>of</strong> the figure and moving down the cross-sections that were analysed are: i) the oxygen side <strong>of</strong><br />

the interconnect, including a native oxide layer at the interconnect interface; ii) the bond layer used to<br />

electrically connect the interconnect plate to the corrugated flow field (not shown); iii) another bond<br />

layer to attach the flow field to the oxygen electrode <strong>of</strong> the cell; iv) strontium-doped manganite<br />

oxygen electrode; v) Scandia-stabilised zirconia electrolyte; vi) nickel zirconia cermet steam/hydrogen<br />

electrode; vii) the bond layer on the steam/hydrogen electrode and the gas flow field, both which are<br />

not shown; viii) nickel bond layer connecting the flow field to the interconnect; ix) the steam/hydrogen<br />

side <strong>of</strong> the next interconnect plate. A stack repeat unit contains at least nine interfaces between<br />

dissimilar materials.<br />

Maps from electrical resistance and X-ray fluorescence (Mawdsley, 2009) measurements were<br />

used to identify areas which showed probable signs <strong>of</strong> degradation. With resistance measurements,<br />

we determined that the gas flow fields in the oxygen and steam-hydrogen compartments, as well as<br />

the bond layers connecting the bipolar plate in the steam-hydrogen compartment, were functioning<br />

well and maintained excellent electronic conductivity throughout all the tests. As a result, these were<br />

not considered further. The other stack components were cross-sectioned in regions where<br />

degradation was evident, then they were mounted, polished and examined using SEM and EDS.<br />

148 NUCLEAR PRODUCTION OF HYDROGEN – © OECD/NEA 2010

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!