Sealing glass for solid oxide electrolysis cell (SOEC) stacks

Abstract

Solid oxide electrolysis cell (SOEC) stack obtainable by a process comprising the use of a glass sealant with composition 50 to 70 wt % SiO2, 0 to 20 wt % Al2O3, 10 to 50 wt % CaO, 0 to 10 wt % MgO, 0 to 2 wt % (Na2o 1K2O), 0 to 10 wt % b2O3, and 0 to 5 wt % of functional elements selected from TiO2, ZrO2, ZrO2, F, P2O5, Mo03, FeO3, MnO 2, LaSrMnO perovskite (LSM) and combinations thereof. Preferably, the sealant is a sheet of E-glass fibers with a composition in wt % of 52-56 SiO2, 12-16AL2O3, 16-25 CaO, 0-6MgO, 0-2 Na2+K2O, 0-10 B2O3, 0-1.5 TiO2, O-1F.

Claims

1. Solid oxide electrolysis cell (SOEC) stack, comprising: a first cell stack assembly comprising at least one interconnector plate alternating with at least one cell unit, in which each cell unit comprises a first electrode, a second electrode and an electrolyte arranged between these electrodes, and a glass fibre sealant sheet disposed in between the interconnector plate and each cell unit, in which the glass sealant is an E-glass and is a barium oxide free composition which contains: 52-62 wt % SiO.sub.2, 10-15 wt % Al.sub.2O.sub.3, 18-25 wt % CaO, 0.5-4 wt % MgO, Na.sub.2O, K.sub.2O, both Na.sub.2O and K.sub.2O being present in a combined amount of 0.25-2 wt %, 3.5 wt % B.sub.2O.sub.3, and 0-5 wt % of functional elements selected from TiO.sub.2, ZrO.sub.2, F.sub.2, P.sub.2O.sub.5, MoO.sub.3, Fe.sub.2O.sub.3, MnO.sub.2, LaSr-M perovskite (LSM) and combinations thereof; said first cell stack assembly converted into a second assembly having a glass sealant of thickness between 5 and 100 m by heating said first assembly to a heating temperature of 500 C. or higher and subjecting the cell stack to a load pressure of 2-20 kg/cm.sup.2; and said second assembly converted into a final solid oxide electrolysis cell stack assembly by cooling the second assembly to a temperature below the heating temperature, the SOEC exhibiting a degradation rate of about 1%/1000 hours.

2. Solid oxide electrolysis cell stack according to claim 1 wherein the heating temperature is 800 C. or higher and the load pressure is 2-10 kg/cm.sup.2, and results in a glass sealant of thickness between 5 and about 50 m.

3. Solid oxide electrolysis cell stack according to claim 1, wherein the content of SiO.sub.2, Al.sub.2O.sub.3, CaO and MgO represents 85-95 wt % or 87-97 wt % of the glass sealant composition, the content of Na.sub.2O+K.sub.2O and B.sub.2O.sub.3 represents 3.75 to 5.5 wt % of the glass sealant composition and functional elements selected from TiO.sub.2, F.sub.2, ZrO.sub.2, P.sub.2O.sub.5, MoO.sub.3, Fe.sub.2O.sub.3, MnO.sub.2 and LaSrMnO perovskite (LSM) and combinations thereof represent 0-5 wt %.

4. Solid oxide electrolysis cell stack according to claim 1, wherein the glass sealant is loaded with filler material in the form of MgO, steel-powder, quartz, leucite and combinations thereof.

Description

(1) The FIGURE shows the average cell voltage during operation of a SOEC stack prepared according to Example 1.

EXAMPLE 1

(2) A 300 m thick anode supported cell with internal feeding and exhaust holes has demasked contact layers in the manifold areas in order to minimise the leakage through these porous structures. A metal gasket frame covered with equally shaped, punched E-glass fibre paper having a composition according to the invention (e.g. ASTM D578-05: 52 to 62 wt % SiO.sub.2, 12 to 16 wt % Al.sub.2O.sub.3, 16 to 25 wt % CaO, 0 to 5 wt % MgO, 0 to 2 wt % Na.sub.2O+K.sub.2O, 0 to 10 wt % B.sub.2O.sub.3, 0 to 1.5 wt % TiO.sub.2, Fe.sub.2O.sub.3 0.05 to 0.8 wt % and 0-1 wt % fluoride) on both sides is placed on both sides of the cell in such a way that air from the manifold holes is allowed to pass over one electrode (air side), and such that steam gas is allowed to pass over the other electrode (steam side) of the cell. Above and below the cell and gasket assemblage, an interconnect plate with manifold holes is placed. The E-glass paper contains fibres in an amount of 100 g/m.sup.2 towards the cell and 50 g/m.sup.2 towards the interconnect plate corresponding to a 40 and 20 m, respectively, thick dense layer after treatment according to the invention at temperatures of about 880 C. and a load pressure of about 6 kg/cm.sup.2. Building a stack with 5 cells, a cross-over leak between the anode and the cathode sides has been measured at RT to as low as 0.05 and 0.09% in two stacks after a full thermal cycle. With gas chromatography using steps of 2N.sub.2 concentration in oxygen on the air side and measuring the N.sub.2 mole concentration on the steam side during operation with the same gas pressure on the steam and oxygen/air side, we obtained a doubling of the N.sub.2 mole % in the anode of each step showing that the there is a leakage and that it is diffusion driven, presumably due to the diffusion through the porous structures of the cell (mainly the anode support). An increasing of the gas pressure on the oxygen side had no effect on the cross-over leak on the steam side.

(3) XRD-spectres of the E-glass show the presence of wollastonite, CaSiO.sub.3 (diopside, (Ca,Mg) SiO.sub.3 also fit the spectrum and its presence depends on the MgO-content of the glass) together with anorthite (CaAl.sub.2Si.sub.2O.sub.8, which may contain up to 10 mole % NaAlSi.sub.3O.sub.8) and cristobalite, (SiO.sub.2).

(4) The flat profile of the FIGURE shows that the SOEC does not degrade significantly during operation. In the electrolysis mode at 850 C., 0.5 A/cm.sup.2 45% H.sub.2O-45CO.sub.2-10% H.sub.2 the solid oxide cell stack has operated with a degradation as low as 1%/1000 hours between 30 to 800 hours. At 0.75 A/cm the overall voltage increase seems to level out before the stack test was stopped due to a system failure. The degradation rate is low compared to literature where degradation rates of 2%/1000 or more in high temperature operation of SOECs are normal. For instance, degradation rates of 2%/1000 hours at 850 C., p(H2O)/p(H2)=0.5/0.5 and 0.5 A/cm.sup.2 and 6%/1000 hours at 950 C. p(H2O)/p(H2)=0.1/0.9 and 1.0 A/cm.sup.2 have been reported in literature. Normally the degradation has been attributed to the delamination of the O.sub.2 electrode, Cr-contamination, as well as contamination of the H.sub.2-electrode's triple phase boundary by silica. The silica could also originate from the interconnect plate. In the present case the low degradation of 1%/1000 hours, compared to e.g. 2%/1000 hours or more of prior art SOECs indicates that the E-glass seal does not significantly contaminate the electrodes of the cell over 800 hours. Without being bound by any theory the reason for this appears to be that the E-glass seal was crystallized to a stable assemblage of MgCaSi.sub.2O.sub.6, CaSiO.sub.3, CaAl.sub.2Si.sub.2O.sub.8 and SiO.sub.2 (cristobalite) with a reduced area of exposed (SiO.sub.4).sup.4 units compared to the albite glass. Also a smaller exposed surface due to the design of the stack with very thin layers of sealing glass. Some preliminary results from another stack operating in electrolysis mode at 0.65 A/cm.sup.2 show no degradation so it is unknown to what extent the degradation of 1%/1000 hours is driven by Si or Cr-contamination.

(5) Therefore the invention enables to prepare by simple means (use of E-glass fibre paper as glass sealant precursor) a final cell stack in which the components of the stack including the sealant work well together without creating leakages during ordinary operation and thermal cycling. No deteriorating reactions occur between the oxide scale of the interconnect and the E-glass.