TWI901031B - Fuel cell interconnect with iron rich rib regions and method of making thereof - Google Patents

Abstract

An interconnect for a solid oxide fuel cell stack includes a chromium alloy interconnect having ribs separated by channels, and metal or metal oxide caps located directly on the top surfaces of the ribs on both air and fuel sides of the interconnect but not on bottom of the channels, where the metal or metal oxide caps comprise iron, manganese, cobalt, copper, a superalloy or an oxide thereof.

Description

Fuel cell interconnects having iron-rich rib regions and methods of making the same

The present invention relates to fuel cell stack assemblies, and more particularly, to interconnects and methods of preparing interconnects for use in fuel cell stacks.

A typical solid oxide fuel cell stack consists of multiple fuel cells separated by metal interconnects (ICs), which provide electrical connections between adjacent cells in the stack and channels for transporting and removing fuel and oxidant. The metal interconnects are typically composed of Cr-based alloys, such as those known as CrFe, with a composition of 95 wt% Cr-5 wt% Fe, or Cr—Fe—Y with a composition of 94 wt% Cr-5 wt% Fe-1 wt% Y. CrFe and CrFeY alloys maintain their strength and are dimensionally stable under typical solid oxide fuel cell (SOFC) operating conditions, such as 700-900°C in air and wet fuel atmospheres. However, during SOFC operation, chromium in the CrFe or CrFeY alloy reacts with oxygen and forms chromium oxide, leading to degradation of the SOFC stack.

_Two of the main degradation mechanisms affecting SOFC stacks are directly related to chromium oxide formation on metal interconnect components: i) higher stack ohmic resistance due to the formation of native chromium oxide (chromia, _Cr2O3 ) on the interconnects, and ii) chromium poisoning of the SOFC cathode.

Although _Cr₂O₃ is an electronic _conductor , its conductivity at SOFC operating temperatures (700-900°C) is extremely low, reaching a value of approximately 0.01 S/cm at 850°C (compared to 7.9× ^10₄ S/cm for Cr metal). The thickness of the chromium oxide layer on the interconnect surface grows over time, and thus the ohmic resistance of the interconnect, and therefore the ohmic resistance of the SOFC stack due to this oxide layer, increases over time.

A second degradation mechanism, related to the formation of chromium oxide in metal interconnects, is called cathode chromium poisoning. At SOFC operating temperatures, chromium vapor diffuses through cracks or pores in the coating, and chromium ions can diffuse through the crystal lattice of the interconnect coating material via solid-state diffusion into the SOFC cathode. Furthermore, during fuel cell operation, ambient air (humid air) flows through the air (cathode) side of the interconnect, and wet fuel flows through the fuel (anode) side of the interconnect. At SOFC operating temperatures and in the presence of humid air (cathode side), chromium on the surface of the _{Cr2O3 layer} on the interconnect reacts with water and evaporates as a gaseous species, chromium oxide hydroxide ( _CrO2 (OH) ₂ ). The chromium oxide _hydroxide species is transported from the interconnect surface in vapor form to the cathode electrode of the fuel cell, where it can deposit in solid form ( _{Cr2O3 )} . _Cr2O3 deposits on and in the SOFC cathode (e.g., via grain boundary diffusion) and/or reacts with the cathode (e.g., to form Cr—Mn spinel), leading to significant performance degradation of the cathode electrode. Typical SOFC cathode materials, such as calcium-titanium materials (e.g., LSM, LSC, LSCF, and LSF), are particularly susceptible to chromium oxide degradation.

One embodiment includes a method of preparing an interconnect for a solid oxide fuel cell stack, comprising providing an iron-rich material into a channel of a mold; providing a powder containing 4-6 wt% Fe, 0-1 wt% Y, and the balance Cr over the iron-rich material containing at least 25 wt% iron in the mold; pressing the iron-rich material containing at least 25 wt% iron and the powder containing 4-6 wt% Fe, 0-1 wt% Y, and the balance Cr in the mold to form an interconnect; and sintering the interconnect to form a sintered interconnect having iron-rich regions having an iron concentration greater than 10% in ribs of the interconnect.

Another embodiment includes a method of preparing an interconnect for a solid oxide fuel cell stack, comprising providing a chromium alloy interconnect having ribs separated by channels, and forming a metal or metal oxide coating directly on the top surfaces of the ribs rather than on the bottoms of the channels, wherein the metal or metal oxide coating comprises iron, manganese, cobalt, copper, superalloys, and oxides thereof.

Another embodiment includes an interconnect for a solid oxide fuel cell stack comprising a chromium alloy interconnect having ribs separated by channels; and a metal or metal oxide coating directly on top surfaces of the ribs rather than on the bottoms of the channels, wherein the metal or metal oxide coating comprises iron, manganese, cobalt, copper, a superalloy, or oxides thereof.

To limit the diffusion of chromium ions (e.g., Cr ³⁺ ) through the interconnect coating material to the SOFC cathode, materials with very few cation vacancies and, therefore, low chromium diffusion rates can be selected. A family of materials with low cation diffusion rates belongs to the calcium-titanium family, such as strontium oxides, such as La _{1 − x} Sr _x MnO ₃ (LSM), where 0.1 ≤ x ≤ 0.3, e.g., 0.1 ≤ x ≤ 0.2. These materials have been used as interconnect coating materials. In the case of LSM, the material has high electronic conductivity but low anion and cation diffusion.

A secondary function of the interconnect coating is to inhibit the formation of native oxide on the interconnect surface. Native oxide forms when oxygen reacts with chromium in the interconnect alloy to form a relatively high resistance layer _{of Cr₂O₃} . If the interconnect coating can inhibit the transport of oxygen and water vapor from the air to the interconnect surface, the kinetics of oxide growth can be reduced.

Similar to chromium, oxygen (e.g., ^{O2 -} ions) can be transported through the coating by solid-state diffusion or by gaseous transport through pores and cracks in the coating. This mechanism can also be used for airborne water vapor, a promoter for Cr evaporation and possible oxide growth. As discussed above, in a humid air environment, chromium evaporates from the surface of _Cr2O3 in the form of gaseous molecules _{CrO2 (} OH) ₂ , which can then diffuse through defects in the coating (such as pores and cracks). In the case of oxygen and water vapor, depending on the size of the defect or pore, the molecules diffuse through the defect by bulk diffusion or the Knudsen diffusion process.

If _CrO2 (OH) ₂ molecules contact the coating surface, they can react to form crystals and then re-evaporate to continue diffusing in the gas stream (in cracks or pores). Experiments have shown that _CrO2 (OH) ₂ reacts with the LSM interconnect coating 104 to form a spinel phase 101, such as manganese chromium oxide (Mn, Cr) _3O4 as shown in Figure _1. Although _CrO2 (OH) ₂ reacts with the LSM to form the spinel phase, the chromium species does not inhibit re-evaporation and further diffuses into cracks or defects. Chromium has been observed to be transported along the length of cracks in LSM IC coatings operated for long periods of time in fuel cells. Figure 2 shows chromium crystals 101 in cracks 103 in an LSM IC coating 104 that was operated in an SOFC stack under typical conditions of 800-850°C in ambient air on the cathode side. The formation of chromium-containing crystals is characteristic of crystals formed by a gas-to-solid phase transition. SEM and EDS analysis of the bulk LSM coating away from the cracks revealed no chromium. Therefore, it can be concluded that the majority of chromium transport from the CrFe interconnect through the LSM IC coating occurs via vapor-phase transport through micro- and macro-cracks, interparticle spaces, and pores in the LSM coating.

In the case of solid-state transport, materials with very few oxide ion vacancies and, therefore, low oxide ion conductivity are selected. For example, the unique feature of calcium-titanium LSM is that it exhibits low cation and anion conductivity, but high electronic conductivity, making it an excellent coating material. Other calcium-titanium ores (such as La _{1 − x} Sr _x FeO _{3 − d} , La _{1 − x} Sr _x CoO _{3 − d ,} and La _{1 − x} Sr _x Co _{1 − y} Fe _y O _{3 − d} ) all exhibit high electronic conductivity and low cation conductivity (low chromium diffusion rate). However, these particular materials also exhibit high oxide ion conductivity and are therefore less effective in protecting interconnects from oxidation (oxide growth).

A second family of materials that can be used for interconnect coatings is manganese cobalt oxide (MCO) spinel materials. In one embodiment, MCO spinels range from _Mn2CoO4 to _Co2MnO4 . That is, any spinel with _the composition _{Mn2 − xCo1 +xO4 (} 0≤x≤1 ₎ , or written as z( _Mn3O4 )+(1− _z )( _Co3O4 ) (where (⅓≤z≤⅔)), or written as (Mn _{, Co)3O4, such as Mn1.5Co1.5O4 , MnCo2O4 , or Mn2CoO4 , can be used} . Many spinels containing transition metals exhibit good electronic conductivity and relatively low anion and cation diffusion rates and are therefore suitable coating materials.

In one embodiment, spinel (e.g., (Mn, Co) ₃ O ₄ ) powder is doped with Cu to lower the spinel's melting temperature. The lowered melting temperature improves (increases) the coating density during deposition using coating methods such as air plasma spraying (APS) and increases the conductivity of the oxide in the reaction zone. Due to the lower melting temperature, the increase in coating density can occur during APS deposition and during extended operation at SOFC temperatures.

Adding Cu to the spinel layer has additional advantages. Cu-doped spinels, such as (Mn, Co) ₃ O ₄ , can lead to higher conductivity in the underlying spinel phase and in any reaction zone oxides formed between the spinel and the native Cr ₂ O ₃ oxide. Examples of conductivity for oxides from the (Mn, Co, Cu, Cr) ₃ O ₄ family include: CuCr ₂ O ₄ : 0.4 S/cm at 800°C, Cu _1.3 Mn _1.7 O ₄ : 225 S/cm at 750°C, and CuMn ₂ O ₄ : 40 S/cm at 800°C.

The spinel family of materials has the general formula AB ₂ O ₄ . Depending on the elements occupying the A and B sites, these materials can form octahedral or cubic crystal structures. Furthermore, depending on the doping conditions, copper atoms can occupy the A site, the B site, or a combination of A and B sites. Generally speaking, Cu prefers to enter the B site. When the A element is Mn, the B element is Co, and the spinel is doped with Cu, the spinel family can be described by the general formula (Mn, Co, Cu) ₃ O ₄ . More specifically, depending on the position of the Cu alloying element, the spinel family can be described by the following formula:

(1) If Cu enters the A site, then Mn _{2 − x − y} Co _1+x Cu _y O ₄ (0≤x≤1), (0≤y≤0.3)

(2) If Cu enters the B site, then Mn _{2 − x} Co _{1+x − y} Cu _y O ₄ (0≤x≤1), (0≤y≤0.3)

(3) If Cu enters sites A and B equally, then Mn _{2 − x − y} /2Co _{1+x − y} /2Cu _y O ₄ (0≤x≤1), (0≤y≤0.3)

Specific (Mn, Co, Cu) ₃ O ₄ compositions include, but are not limited to, Mn _1.5 Co _1.2 Cu _0.3 O ₄ , Mn _1.5 Co _1.4 Cu _0.1 O ₄ , Mn ₂ Co _0.8 Cu _0.2 O ₄ , and Co ₂ Mn _0.8 Cu _0.2 O ₄ . If Cu is incorporated into the B site, additional compositions include Mn ₂ Co _{1 − y} Cu _y O ₄ , where (0 ≤ y ≤ 0.3). These compositions can also be written as (Mn ₂ O ₃ ) + (1 − z)(CoO) + z(CuO), where (0 ≤ z ≤ 0.3). If Cu is incorporated into the B site, further compositions include Co ₂ Mn _{1 − y} Cu _y O ₄ , where (0 ≤ y ≤ 0.3). These compositions can also be written as (Co ₂ O ₃ )+(1−z)(MnO)+z(CuO), where (0≤z≤0.3). In a preferred Mn, Co spinel composition, the Mn/Co ratio is 1.5/1.5, for example, M _1.5 Co _1.5 O ₄ . When the B site is doped with Cu, preferred compositions include Mn _1.5 Co _{1.5 − y} Cu _y O ₄ , where (0≤y≤0.3).

In another embodiment, the (Mn, Co) ₃ O ₄ or (Mn, Co, Cu) ₃ O ₄ spinel family is doped with one or more monovalent species. That is, one or more species that have only one valence. Doping with monovalent species reduces cation transport at high temperatures and thus reduces the thickness of the intermediate oxide layer. The primary ion transport mechanism in spinels is cation diffusion through cation vacancies in the lattice structure. In spinels with multivalent species M ^2+/3+ (such as Mn ^3+/4+ and Co ^2+/3+ ), cation vacancies are generated when the M species oxidizes from a lower valence state to a higher valence state to maintain local charge neutrality. The introduction of monovalent species generally reduces the amount of cation vacancies and reduces the amount of interdiffusion between the spinel coating 102 and the native Cr ₂ O ₃ oxide or CrFe substrate 100. In this way, the amount of intermediate oxide layer formed is reduced. Examples of monovalent species that can be introduced into the spinel coating include Y ³⁺ , Al ³⁺ , Mg ²⁺ and/or Zn ²⁺ metals. In one embodiment, the spinel coating has a composition of (Mn, Co, M) ₃ O ₄ , where M = Y, Al, Mg, or Zn. For example, if M = Al is doped in the A site, the spinel composition may include Mn _{2 − y AlyCoO 4} (0≤y≤0.3) or (1−z)(Mn ₂ O ₃ )+z(Al ₂ O ₃ )+CoO, where (0≤z≤0.15).

In one embodiment, the interconnect coating is deposited onto a Cr-based alloy interconnect, such as an IC containing 93-97 wt% Cr and 3-7 wt% Fe, such as the Cr—Fe—Y or CrFe interconnects described above, using an air plasma spray (APS) process. The air plasma spray process is a thermal spray process in which powdered coating material is fed into a coating apparatus. Coating particles are introduced into a plasma jet, where they are melted and then accelerated toward the substrate. Upon reaching the substrate, the molten droplets are flattened and cooled, forming the coating. Plasma can be generated by direct current (DC plasma) or by induction (RF plasma). In addition, unlike controlled atmosphere plasma spraying (CAPS), which requires an inert gas or vacuum, air plasma spraying is performed in ambient air.

Cracks in the coating can appear at two different times: a) during deposition, and b) during operation under SOFC conditions. Cracks formed during deposition are influenced by the spray gun parameters and the material properties of the coating. Cracks formed during operation vary greatly with the material properties, more specifically, the density and sinterability of the material. Without being bound by a particular theory, it is believed that cracks that appear during operation are a result of continued sintering of the coating and, therefore, increasing densification of the coating over time. As the coating densifies, it shrinks laterally. However, the coating is constrained by the substrate and therefore forms cracks to relieve stress. Coatings applied at lower densities are more likely to densify further during operation, leading to crack formation. In contrast, coatings applied at higher densities are less likely to form cracks.

In a first embodiment, a sintering aid is added to the IC coating to reduce crack formation and, therefore, chromium evaporation. A sintering aid is a material that increases the density of the deposited coating and/or reduces densification after the coating is deposited. Because the sintering aid increases the deposited density of the coating material, it reduces crack formation after coating formation due to subsequent densification and/or operational stresses on relatively porous materials. Suitable sintering aids include materials that: a) lower the melting temperature of the bulk phase of the coating material, b) melt at a lower temperature than the bulk phase, resulting in liquid phase sintering, or c) form a secondary phase with a lower melting temperature. For the calcium-titanium family, including LSM, sintering aids include Fe, Co, Ni, and Cu. These transition metals are soluble in LSM and readily dope the B sites in the _ABO3 calcium-titanium phase. The melting temperatures of the oxides in the 3d transition metals tend to decrease in the order Fe > Co > Ni > Cu. Adding these elements to the B sites of the LSM will lower the melting temperature and increase the as-sprayed density. In one embodiment, one or more of Fe, Co, Ni, and Cu are added to the coating such that the coating comprises 0.5 wt% to 5 wt% (e.g., 1% to 4%, e.g., 2% to 3%) of these metals. In an alternative embodiment, the coating composition is expressed in atomic percentages and comprises La _{1 - x} Sr _x Mn _{1 - y} My _y O _{3 - d} , where (M = Fe, Co, Ni, and/or Cu), 0.1 ≤ x ≤ 0.3, 0.005 ≤ y ≤ 0.05, and 0 ≤ d ≤ 0.3. It should be noted that the atomic percentage ranges of Fe, Co, Ni, and Cu do not necessarily have to match the weight percentage ranges of these elements of the previous embodiments.

Other elements may also be added in combination with the above transition metals to maximize conductivity, stability, and sinterability. Such elements include, but are not limited to, Ba, Bi, B, Cu, or any combination thereof (e.g., a Cu+Ba combination), for example, in an amount of 5 wt% or less, such as 0.5-5 wt%. Furthermore, sintering aids, such as Y, may be added to achieve similar effects, particularly at the A _- site of the LSM. An example according to this embodiment is _{LayYxSr1 - x - yMnO3 ,} where _x = 0.05-0.5 and _{y =} 0.2-0.5, such as _{La0.4Y0.1Sr0.5MnO3 .} For coating materials other than LSM, copper can be used as a sintering aid in the MCO spinel materials described above.

In another embodiment, instead of introducing a transition metal powder into the air plasma jet during deposition, a metal oxide powder that readily reduces to its metallic state in an APS atmosphere is added to the plasma. Preferably, the metal of the metal oxide exhibits a melting temperature lower than that of the coating phase (calcite or spinel phase). For example, binary oxides such as cobalt oxide (e.g., CoO, _{Co₃O₄ ,} or _Co₂O₃ ), NiO, _In₂O₃ , SnO, _B₂O₃ , copper oxide ( _e.g. , CuO or _Cu₂O ), BaO, _{Bi₂O₃ ,} ZnO, or any combination thereof (e.g., (Cu, _Ba )O) can be added as a second phase to the coating powder (i.e., LSM powder or La+Sr+ _Mn powder or oxides thereof). This addition produces a two-phase powder mixture that is fed to the spray gun. The amount of the second phase can be less than or equal to 5 wt%, such as in the range of 0.1 wt% to 5 wt% of the total powder weight.

In the APS gun, as the LSM particles solidify on the surface of the IC, the metal oxides are reduced to their metallic phase, melting and promoting sintering of the melted LSM particles. The lower melting temperatures of the metal and binary oxides promote densification during deposition and solidification.

In another embodiment, a material that reacts with the coating material (such as LSM) and forms a secondary phase with a lower melting temperature is added to the coating feed during the APS process. The lower melting temperature secondary phase promotes densification. For example, silicate and/or calcium aluminate powder can react with the coating material powder in the hot plasma portion of the APS gun to form a glassy phase. In one embodiment, La from the LSM material reacts with Si—Ca—Al oxide (which may also include K or Na) to form a glassy phase, such as La—Ca—Si—Al oxide, which forms between the LSM particles. The coating may include less than or equal to 5 wt%, such as 0.5-5%, of a silicate, Ca-Al oxide, or Si-Ca-Al oxide.

In a second embodiment, the coating is post-treated in a manner that causes stress-free densification. This post-treatment can be carried out with or without the addition of the sintering aid of the first embodiment. In one example post-treatment according to the second embodiment, a "redox" cycle is carried out in an _N2 and _O2 atmosphere. In this cycle, the coating is alternately exposed to a neutral and an oxidizing atmosphere. For example, the coating can be treated in a neutral atmosphere comprising nitrogen or an inert gas (e.g., argon) and then treated in an oxidizing atmosphere comprising oxygen, water vapor, air, etc. If necessary, one or more cycles, such as 2, 3, 4 or more, can be carried out. If necessary, a reducing (e.g., hydrogen) atmosphere can be used instead of a neutral atmosphere or in addition to a neutral atmosphere. The redox cycle in the _N₂ and _O₂ atmospheres creates a cation vacancy concentration gradient, which increases the diffusion of cation vacancies and thereby effectively increases the sintering rate. This effect can be further enhanced by using a _low -Sr content LSM coating of _{La₁ − xSrxMnO₃ − d} (where x <= 0.1, e.g., 0.01 ≤ x ≤ 0.1, d ≤ 0.3), maximizing the non-stoichiometric amount of oxygen. Using this sintering step can enhance any or all of the sintering aid techniques described above.

In a third embodiment, the surface area of electrical interaction between the coating and the underlying Cr—Fe IC surface is increased. As the thickness of the chromium oxide layer increases, the chromium oxide layer that forms between the coating and the IC causes a millivolt drop over time. The total voltage drop depends on the area and thickness over which the voltage drop occurs. Increasing the area of oxide growth between the IC and the coating reduces the impact on voltage loss, thereby increasing the life of the stack. By adding a substance that penetrates deeply into the coating, this embodiment effectively increases the surface area of contact and thereby reduces the impact of the growing chromium oxide layer.

The method according to this third embodiment includes embedding a small amount of coating material into the IC. There are two alternative aspects of this embodiment. One aspect includes distributing the coating material, such as LSM or MCO, thoroughly and uniformly within the IC powder (e.g., Cr—Fe powder) prior to pressing to form the IC. The coating powder (e.g., LSM and/or MCO powder) may be included when the lubricant and Fe, Cr (or Cr—Fe alloy) powders are mixed together prior to pressing. Preferably, the powder mixture is capable of withstanding sintering temperatures and a reducing environment. A second aspect includes incorporating (e.g., embedding) a predetermined amount of coating powder only into the top surface of the Cr alloy IC. After the IC sintering step, the oxide regions embedded in the surface of the CrFe or CrFeY IC increase the surface roughness of the IC. After the pressing and sintering steps, the overall coating is deposited on the Cr alloy interconnect.

3A to 3C illustrate a method for embedding a coating material in the top surface of an interconnect. As shown in FIG3A , a lubricant and Cr/Fe powder 202 for forming the body of the IC are added to a mold cavity 200 using a first shoe (not shown) or another suitable method. As shown in FIG3B , prior to the pressing step, a coating material powder 204 (e.g., LSM or MCO) or a mixture of coating material powder 204 and lubricant/Cr/Fe powder 202 is provided to the mold cavity above the powder 202 in the mold cavity using a second shoe 206. 3C , a punch 208 is used to press the powders 204 , 202 to form an interconnect having the coating material embedded in its surface on the air side (i.e., if the air side of the IC is formed in the mold facing upward).

Alternatively, coating material powder 204 (e.g., LSM or MCO) (or a mixture of coating material powder 204 and lubricant/Cr/Fe powder 202) is first provided into mold cavity 200. If the air side of the IC is formed facing downward in the mold, lubricant/Cr/Fe powder 202 is provided to the mold cavity above powder 204 in mold cavity 200 before the pressing step. In this way, the coating material is incorporated into the IC primarily at the top of the air side surface of the IC.

4 , the coating powder 204 may be electrostatically attracted to the upper punch 208 of the press. The upper punch 208 then presses the coating powder 204 and lubricant/interconnect powder material 202 in the mold cavity 200 to form an IC with the coating material 204 embedded in the top portion of the air side.

Using the above method, after the pressing step, the coating powder can be uniformly incorporated into the surface of the air side of the IC. The pressing step is then followed by a sintering and coating step, such as an MCO and/or LSM coating step by APS or another method described herein.

The ratio of the coating powder to the Fe in the Cr-Fe alloy is preferably selected so that the top coating material has a coefficient of thermal expansion (CTE) similar to that of the sintered and oxidized interconnect. The CTE of the Cr-Fe alloy is a function of the alloy's composition and can be selected by selecting the ratio of Cr to Fe. The sintering process can be adjusted to keep the powder oxidized and stable. For example, sintering can be performed using wet hydrogen or in an inert atmosphere such as nitrogen, argon, or another inert gas. Wet hydrogen or inert gas atmospheres are oxidizing or neutral, respectively, and thereby prevent the oxide powder from reducing.

In a fourth embodiment, the coating is a multi-layer composite. FIG5 illustrates an example of a fourth embodiment of an IC having a composite coating. The composite coating consists of a spinel layer 102 and a calcium-titanium layer 104. The spinel layer 102 is first deposited on a Cr alloy (e.g., CrFe) interconnect 100. A calcium-titanium layer 104 (e.g., an LSM layer described above) is then deposited on top of the spinel layer 102. During deposition of layer 102 and/or during high temperature operation of a fuel cell stack containing the interconnect, an interfacial spinel layer 101 containing native chromium may form between the interconnect 100 and layer 102.

Preferably, lower spinel layer 102 comprises the Cu and/or Ni-containing MCO spinel described above. Layer 102 acts as a doping layer, increasing the conductivity of the underlying manganese chromium oxide (Mn, Cr) ₃ O ₄ or manganese cobalt chromium oxide (Mn, Co, Cr) ₃ O ₄ interface spinel layer 101. In other words, during and/or after forming layer 101, Cu and/or Ni from spinel layer 102 diffuse into interface spinel layer 101. This produces a Cu and/or Ni doped layer 101 (e.g., (Mn and Cr) _{3 − x − y} Co _x (Cu and/or Ni) _y O ₄ , where (0≤x≤1), (0≤y≤0.3)), which reduces the resistivity of the layer 101 .

Layer 102 may include the Cu-containing MCO layer and/or the Ni-containing MCO layer and/or the Ni-and-Cu MCO layer described above. In the MCO layer, when element A is Mn, element B is Co, and the spinel is doped with Cu and/or Ni, the spinel family can be described by the general formula (Mn, Co) _{3 − y} (Cu, Ni) _y O ₄ , where (0 < y ≤ 0.3). More specifically, depending on the position of the Cu and/or Ni alloying elements, the spinel family can be described by the following formula:

(1) If Cu and/or Ni enter the A site, then Mn _{2 − x − y} Co _1+x (Cu, Ni _y O ₄ (0≤x≤1), (0≤y≤0.3)

(2) If Cu and/or Ni enter the B site, then Mn _{2 − x} Co _{1+x − y} (Cu, Ni) _y O ₄ (0≤x≤1), (0≤y≤0.3)

(3) If Cu and/or Ni enter the A and B sites equally, then Mn _{2 − x − y/2} Co _{1+x − y/2} (Cu, Ni _y O ₄ (0≤x≤1), (0≤y≤0.3).

Although the spinel doped layer 102 containing Cu and/or Ni reduces the ASR of the interconnect, it is permeable to both oxygen and chromium. Therefore, in an embodiment of the present invention, a second calcium-titanium barrier layer 104 is formed above the doped layer 102. Preferably, layer 104 is a dense LSM layer that reduces or prevents the diffusion of Cr and oxygen. Layer 104 can be formed using the sintering aid described above to increase its density. The dense layer 104 reduces or prevents the growth of the interfacial spinel layer 101 by blocking the diffusion of air and oxygen from the cathode side of the fuel cell to the CrFe IC surface during stack operation. Layer 104 also reduces or prevents chromium poisoning of the cathodes of fuel cells in the stack by reducing or preventing chromium from diffusing from the IC to the cathode.

Thus, the composite coating 102/104 reduces or eliminates the area specific resistance (ASR) degradation contribution of interconnects and the stack, and reduces overall degradation of the fuel cell stack, by reducing or eliminating Cr poisoning of the fuel cell cathode. First, the spinel-doped layer 102 dopes the chromium-containing interfacial spinel layer 101 with elements (e.g., Ni and/or Cu) that reduce the electrical resistance of the spinel layer 101. Second, the spinel layer 102 prevents direct interaction between the calcium-titanium layer 104 and the chromium-containing interfacial spinel layer 101, which could lead to the formation of undesirable and resistive secondary phases. Third, the spinel layer 102 (e.g., a Mn-containing spinel with Co, Cu, and/or Ni) is less prone to cracking than the LSM layer 104, which enhances the integrity of the coating. Fourth, the top calcium-titanium layer 104 acts as a second barrier layer that reduces oxygen transfer to the interfacial oxide 101 on the interconnect surface. Thus, the top calcium-titanium layer 104 reduces the growth rate of the native oxide layer 101 and reduces the transfer of chromium from the doped layer 101 to the fuel cell cathode via the doped layer 102.

FIG6A shows the air side of an exemplary interconnect 100. The interconnect can be used in a stack, with a manifold for fuel inside and a manifold for air outside. The interconnect contains airflow passages (or channels) 8 between ribs 10 to allow air to flow from one side 13 of the interconnect to the opposite side 14. Annular (e.g., ring-shaped) seals 15 are located around fuel inlet opening 16A and outlet opening 16B (i.e., through-holes 16A, 16B in interconnect 100). Strip seals 19 are located on the side of interconnect 100.

FIG6B shows a close-up view of an exemplary seal 15, channel 8, and rib 10. Seal 15 can comprise any suitable sealing glass or glass-ceramic material, such as borosilicate glass. Alternatively, seal 15 can comprise the glass-ceramic material described in U.S. application Ser. No. 12/292,078, filed Nov. 12, 2008, which is incorporated herein by reference.

If desired, interconnect 100 may include a raised or bossed area below seal 15. Alternatively, as illustrated in FIG6B , seal 15 is preferably located in flat region 17 of interconnect 100. That is, seal 15 is located in a portion of the interconnect that does not include rib 10. If desired, interconnect 100 may be configured for use with a stack that incorporates manifolds for air and fuel. In this case, interconnect 100 and the corresponding fuel cell electrolyte would also include additional inlet and outlet openings (not shown).

FIG7 illustrates the fuel side of interconnect 100. Window seal 18 is located on the edge of interconnect 100. Also shown are fuel distribution plenums 17 and fuel flow channels 8 between ribs 10. It is worth noting that interconnect 100 shown in FIG7 has two types of fuel flow channels; however, this is not a limitation of the present invention. The fuel side of interconnect 100 can have fuel flow channels of equal depth and length, a combination of short and long channels, and/or a combination of deep and shallow channels.

In one embodiment, interconnect 100 is coated with Mn _1.5 Co _1.5 O ₄ (MCO) spinel at room temperature using an aerosol spray method and further treated with one or more thermal treatments. Generally, the MCO coating is omitted in the sealing areas (rings 15, stripes 19) by masking or removing the MCO deposited in these areas.

The MCO coating can be reduced by fuel in the riser bore and then react with the glass seal material at the annular seal 15. Therefore, in one embodiment, for the interconnect 100 shown in FIG6B , the MCO coating is removed from the flat areas 17 on the air side of the interconnect (e.g., by sandblasting) prior to stack assembly and testing. Alternatively, the flat areas 17 can be masked during aerosol deposition to prevent coating of the flat areas 17. Thus, the MCO coating is omitted in the area 17 below the annular seal 15 adjacent to the fuel inlet opening 16A and/or outlet opening 16B.

In another embodiment, the interconnect 100 is manufactured using a powder metallurgy process. The powder metallurgy process can produce a part that has connected pores within the body of the interconnect 100, which allow fuel to diffuse from the fuel side to the air side. This fuel transmitted through the pores can react with the MCO coating on the air side at the coating/interconnect interface. This reaction can cause seal failure and stack separation. In one embodiment, this failure can be mitigated by masking the seal 19 locations on the edges of the interconnect during MCO deposition, omitting the MCO coating under the strip seal 19, thereby eliminating the coating in these sealed areas and allowing the glass seal 19 to be directly bonded to the metal interconnect.

In another embodiment, the interconnect 100 forms a thin green _{Cr2O3 oxide} layer 25 on the fuel side of the interconnect 100. A cross-sectional micrograph of this fuel side oxide is shown in FIG8. _{The Cr2O3} oxide thickness was found to be between 0.5 and 2 microns. The three methods described below can be used to convert or remove this undesirable chromium oxide layer.

In a first embodiment of the method, the oxide layer is removed by any suitable method, such as sandblasting. This method is effective. However, it is time-consuming and increases processing costs.

Alternatively, _{the Cr2O3} oxide layer 25 can be left in place and converted into a composite layer. In this embodiment, a nickel mesh anode contact is deposited on the _Cr2O3 oxide layer 25 and allowed to diffuse into the chromium oxide layer. The nickel reacts with _{the Cr2O3 oxide layer} 25 and forms _a Ni-metal/ _Cr2O3 composite layer that reduces the ohmic resistance of layer 25. If necessary, the mesh can be heated after contacting layer 25 to accelerate composite formation.

In another embodiment, the oxide layer ₂₅ is reduced or completely eliminated by firing the MCO-coated interconnect in an environment with a low oxygen partial pressure. For example, based on thermodynamics, _Cr2O3 can be reduced to Cr metal at 900°C at a _pO2 (partial pressure) of 10-24 atm, while CoO can be reduced to Co-metal at 900°C at a _pO2 of 10-16 atm. By reducing the oxygen partial pressure of the firing atmosphere (i.e., lowering the dew point) to less than 10-24 atm at 900°C, the formation of _{Cr2O3 oxide} on the fuel side (uncoated side) is prevented, while allowing the MCO coating on the air side of the interconnect to be reduced to MnO (or Mn metal if _pO2 < 10-27 atm) and Co-metal, which benefits sintering. At _pO2 < 10-27 atm, MCO will be reduced to Mn-metal and Co-metal, which can lead to better sintering and a denser coating compared to MnO/Co-metal. Generally, the MCO-coated interconnect can be annealed at T>850°C (e.g., 900°C to 1200°C) at a pO ₂ of 10−24 atm (e.g., 10−25 atm to 10−30 atm, including 10−27 atm to 10−30 atm) for 30 minutes to 40 hours, such as 2-10 hours.

In another embodiment described below, the formation of a Cr ₂ O ₃ oxide layer 25 is reduced or avoided in the location between the nickel mesh anode contact compliant layer and the fuel side of the interconnect.

As described above, a compliant layer in the form of nickel is typically introduced in SOFC stacks to conform to surface topography changes and improve contact with the cell's anode electrode. At the beginning of its life, simple contact is sufficient to deliver the expected power from the active area. However, when the interconnects are primarily made of chromium, an oxide layer 25 can subsequently form on the fuel side during operation due to the water content in the fuel. In the following examples, the growth of this oxide layer 25 can be reduced or eliminated beneath the Ni mesh.

The inventors have observed that interconnects containing oxide layer 25 growth between the IC and the Ni mesh typically exhibit high voltage losses and degradation rates that are closely related to the area-specific resistance degradation (ASRD) rate (via Ohm's law and active region). The growth of low-conductivity oxides can generally lead to increased ASRD. Conversely, interconnects with minimal oxide layer 25 growth between the IC and the Ni mesh typically exhibit low ASRD. With the absence of oxide layer 25, the inventors have also observed the formation of a Cr—Fe—Ni alloy in the interconnect immediately beneath the Ni mesh. Without wishing to be bound by a particular theory, the inventors believe that the formation of this Cr—Fe—Ni alloy or Cr—Ni alloy results in lower ASRD and is more resistant to oxide growth than Cr—Fe interconnects. Therefore, it is advantageous to form this alloy at as many contact points as possible between the Ni mesh and the IC, especially in areas of high current density, such as in the middle of the interconnect. Furthermore, during stack operation, current must be forced through the resistive oxide layer 25 that grows later in the stack's life, or consolidate to higher conductivity points, thereby reducing the effective active area.

Furthermore, the inventors believe that the formation of this alloy is influenced by several factors, including the compression pressure between the Ni mesh and the interconnect, the percentage of undiffused Fe in the local interconnect beneath the Ni mesh, surface contamination between the interconnect and the Ni mesh, the bonding between the mesh and the interconnect, and/or the addition of nickel to the interconnect alloy. For example, if the percentage of undiffused Fe is low and contaminants are high, high pressure can be applied to overcome these obstacles. In contrast, if the percentage of undiffused Fe increases and the contaminant content decreases, less pressure can be applied to avoid an increase in ASRD.

In a first aspect of this embodiment, the compressive pressure between the Ni mesh and the interconnect in the fuel cell stack is increased to reduce the formation of the chromium oxide layer 25. One way to increase the pressure on the mesh in the stack is to make the interconnect thickness non-uniform to create a pressure field or gradient across the mesh. Preferably, the height of the interconnect ribs in the middle of the interconnect is slightly greater than the height of the ribs at the edges of the interconnect (i.e., the thickness of the middle portion of the interconnect is slightly greater than the thickness of the outer portion of the interconnect). This creates a pressure field in the middle of the interconnect (where most of the current is generated in adjacent fuel cells in the stack) and, after the mesh and interconnect are placed in the fuel cell stack, applies a higher pressure to the nickel mesh contacting the middle of the interconnect than to the edges of the interconnect. This, in turn, is believed to increase the formation of Cr—Fe—Ni alloy beneath the mesh and/or reduce ASRD.

In a second aspect of this embodiment, contamination between the fuel side of the interconnect and the web is reduced. This can be achieved by reducing the presence of contaminants during the stack manufacturing process and/or by cleaning the surface of the interconnect.

In a third aspect of this embodiment, a sufficiently high percentage of undiffused iron remains on at least the fuel side of the interconnect to form a Cr—Fe—Ni alloy. Undiffused iron includes regions of iron that have not yet been alloyed with the chromium matrix of the interconnect (e.g., in an interconnect having 4-6 wt% Fe with the balance being Cr and, optionally, 0-1 wt% yttrium oxide or yttrium). Achieving a high percentage of undiffused iron can be achieved by any suitable method, such as less sintered pressed powder interconnects and/or starting with larger iron particles. Less sintering involves partially sintering the interconnect after pressing the chromium and iron-containing powders into the interconnect at a lower temperature or for a shorter duration than that required to fully alloy the pressed iron and chromium powder particles. Larger iron particles are effective in achieving the desired percentage of undiffused iron for ASRD reduction purposes, but may require longer sintering times and/or higher sintering temperatures. Therefore, one method of achieving undiffused iron involves pressing a mixture of chromium powder having a first average particle size and iron powder having a second particle size that is larger than the first particle size (e.g., 30-200% larger in diameter, such as 50-100% larger) to form the interconnect, followed by sintering the interconnect.

In a fourth aspect of this embodiment, the nickel mesh is physically bonded to the fuel-side surface of the interconnect to prevent chromium oxide from forming between the mesh and the interconnect surface. For example, the nickel mesh can be heat-welded, welded, or brazed to the interconnect surface across the entire surface of the mesh, at least in the middle of the interconnect, and preferably in the middle and edges of the interconnect. By welding the mesh to the interconnect in multiple locations, particularly in the middle of the interconnect where low pressures are often found, the effective active area is increased and high conductivity is found in that active area. Alternatively, a pressed powder Cr—Fe interconnect can be placed in contact with a Ni mesh and then sintered, while in contact with a Ni mesh that is below the melting point of Ni (e.g., below 1450°C, such as at 1350-1425°C). This sintering temperature maintains the CTE of the part while thermally fusing the Ni mesh to the interconnect in all contact points as the sintering time increases.

In a fifth aspect of this embodiment, nickel is added to the Cr—Fe interconnect alloy to promote the formation of a Cr—Fe—Ni alloy at least on the fuel surface of the interconnect. Iron powder is added to the base Cr powder to increase the CTE of the interconnect to exceed that of chromium and match the CTE of the solid oxide fuel cell. In the case of Ni, which has approximately the same CTE as Fe, it is reasonable that Ni can replace Fe. Ni powder is doped into chromium powder or a mixture of chromium and iron (or chromium-iron alloy) powders in a powder metallurgy press/die, and the powders are subsequently pressed to produce a pressed interconnect portion containing Cr—Ni or Cr—Fe—Ni alloy throughout the portion. Furthermore, Co and Ni are slightly more compressible than Fe, so replacing Fe with these elements will only aid the pressing process. It is well known that adding Fe to a Cr matrix reduces oxidation, so maintaining a certain Fe content may still be beneficial. Thus, all or part of the iron in the interconnect may be substituted with nickel (e.g., 1-100%, such as 10-90%. For example, 30-70% of the iron in a Cr-Fe (4-6 wt%) alloy may be substituted with nickel to form a Cr-M (4-6 wt%) alloy, where M = 1-99% Ni and 99-1% Fe).

Powder composition adjustments are described in the above embodiments to assist the coating function on the air side of the interconnect by partially replacing Fe in the powder mixture with LSM, MCO, Co and/or Mn to promote the formation of a Mn—Co—Cr spinel layer on the cathode side (air side) of the interconnect.

In another aspect of this embodiment, alloying elements suitable for both the air side (e.g., Co and/or Mn) and the fuel side (e.g., Ni) are incorporated into the Cr—Fe interconnect. Thus, the powder composition placed in the press/die includes Cr, Fe, Ni, and at least one of Co and Mn. However, only Co and/or Mn are required on the air side, and only Ni is required on the fuel side. The inventors have observed that a certain amount of separation of the Fe and Cr powders occurs in the press/die, resulting in smaller Cr particles being sifted downward in the press cavity (i.e., in the die cavity), resulting in a more diluted Fe content on the fuel side and a more concentrated Fe content on the air side. This phenomenon can be exploited to layer the IC with the desired materials by mixing the powder composition and placing it into a press cavity, where the Ni particles are smaller than the Cr and Fe particles, which are of the same size. If particles containing Co and/or Mn are also used (e.g., Co and/or Mn metal particles and/or MCO and/or LSM metal oxide particles), they are larger than the Cr and Fe particles of the interconnect pressed air-side up. The punch and die can then be vibrated as the press cavity is filled to aid the separation process. This results in Ni settling on the bottom of the cavity on the fuel side of the interconnect, Co and/or Mn settling on the top of the cavity on the air side of the interconnect, and Fe and Cr remaining in the middle. For interconnects pressed with the fuel side facing up, the Ni particle size is larger than the Cr and Fe particles, the Cr and Fe particles are the same size, and the particles containing Co and/or Mn are smaller than the Cr and Fe particles. As used herein, particle size refers to the average particle size, and larger particles can have an average particle size that is 25-200% larger than the average particle size of Fe and Cr, and smaller particles can have an average particle size that is 25-200% smaller than the average particle size of Fe and Cr.

The following is a non-limiting example of the average particle size of this embodiment.

If Fe is not required for chromium oxide management, then:

2.5% Co, particle size about 100 µm

95% Cr, particle size approximately 50 µm

2.5% Ni, particle size approximately 25 µm

If Fe is required for chromium oxide management, then:

2% Co, particle size about 100 µm

95% Cr, particle size approximately 50 µm

1% Fe, particle size about 50 µm

2% Ni, particle size approximately 25 µm

If a Mn-based spinel is formed, then:

1% Co, particle size about 100 µm

1% Mn, particle size about 100 µm

95% Cr, particle size approximately 50 µm

1% Fe, particle size about 50 µm

2% Ni, particle size approximately 25 µm

In another aspect of this embodiment, nickel powder can be added only to the fuel side of the IC using the method described above with respect to Figures 3 and 4. If desired, nickel powder can be added to the fuel side, while Co, Mn, cobalt oxide, and/or manganese oxide powders can be added only to the air side of the interconnect.

Figures 3A through 3C illustrate a method for embedding an alloy material in the top surface of an interconnect. As shown in Figure 3A, a lubricant and Cr/Fe powder 202, which will form the body of the IC, are added to a mold cavity 200 using a first shoe (not shown) or another suitable method. As shown in Figure 3B, prior to the pressing step, a second shoe 206 is used to provide alloy material powder 204 (e.g., Ni) or a mixture of alloy material powder 204 and lubricant/Cr/Fe powder 202 into the mold cavity above the powder 202 in the mold cavity. Then, as shown in Figure 3C, a punch 208 is used to press the powders 204, 202 to form an interconnect with the alloy material (e.g., Ni) embedded in its surface on the fuel side (i.e., if the fuel side of the IC is formed in the mold facing upward). Before the alloy (e.g., Ni) material is formed on the top side of the interconnect, an air side coating material powder (e.g., LSM, MCO, Co and/or Mn) may be formed on the opposite bottom side of the interconnect as described above with respect to Figures 3A-3C.

Alternatively, alloy material powder 204 (or a mixture of alloy material powder 204 and lubricant/Cr/Fe powder 202) is first provided into mold cavity 200. If the fuel side of the IC is formed facing downward in the mold, lubricant/Cr/Fe powder 202 is then provided above powder 204 in mold cavity 200 before the pressing step. In this way, Ni is incorporated into the IC primarily at the top of the fuel side surface of the IC. As described above with respect to Figures 3A to 3C, air side coating material powder (e.g., LSM, MCO, Co, and/or Mn) can then be formed on the opposite top side of the interconnect.

4 , the alloy powder 204 may be electrostatically attracted to the upper punch 208 of the press. The upper punch 208 then presses the alloy powder 204 and lubricant/interconnector powder material 202 in the die cavity 200 to form an interconnect with the alloy material 204 embedded in the top portion of the fuel side.

Using the above method, the alloy powder can be uniformly incorporated into the surface of the fuel side of the interconnect after the pressing step. The pressing step is then followed by sintering and nickel mesh forming steps.

In a sixth aspect of this embodiment, as shown in Figure 13, a metal or metal oxide contact layer 27 is formed on at least the top portion of the rib 10 on the fuel side of the interconnect. The contact layer 27 contacts the nickel mesh 31, which in turn contacts the anode electrode 3 of the adjacent fuel cell in the stack.

Without wishing to be bound by a particular theory, it is believed that the contact layer 27 breaks up the continuous chromium oxide scale of the chromium oxide layer 25 and promotes the formation of Cr—Fe—Ni or Cr—Ni alloy in the “reaction zone” 29 in the interconnect rib 10 below the web 31, especially when the contact layer 27 comprises nickel or nickel oxide. Specifically, nickel can diffuse from the nickel or nickel oxide contact layer 27 into the chromium oxide layer 25, thereby creating a conductive path through the oxide (e.g., a nickel path formed by solid-state nickel diffusion) and into the "reaction zone" 29 of the interconnect rib, thereby increasing the formation of the Cr—Fe—Ni or Cr—Ni alloy and increasing the size (e.g., depth and/or width) of the "reaction zone" 29. Furthermore, the contact layer 27 increases the contact surface area between the nickel or nickel oxide material and the interconnect rib 10 (compared to the contact surface area between the conductive lines of the nickel mesh 31 and the rib 10). Thus, the contact layer 27 can facilitate good contact between the interconnect rib 10 and the mesh 31. Furthermore, it is believed that the contact layer 27 prevents or reduces the diffusion of impurities 33 from the mesh 31 into the chromium oxide layer 25, which could otherwise potentially cause an undesirable increase in the electrical resistance of the chromium oxide layer 25.

Contact layer 27 can be made of any suitable metal, such as nickel, platinum, or a platinum group metal (such as rhodium, palladium, or ruthenium), copper, iron, cobalt, silver, gold, tungsten, any other transition group metal, or any alloy of the foregoing. The metal can be applied in its metallic (reduced) phase (e.g., nickel metal) or as its oxide (e.g., nickel oxide). The oxide will reduce to a metallic, conductive phase during normal operation with a hydrogen-rich fuel stream.

The contact layer 27 in the form of a metal or metal oxide may be applied by any suitable method, such as screen printing, sputtering, electron beam deposition, evaporation, atomic layer deposition, electroplating, electrodeless plating, thermal spraying, brushing, dip coating, aerosol spraying, electrophoretic deposition, etc. Any of the above manufacturing processes may be optionally followed by a thermal process (e.g., annealing) to achieve bonding and interdiffusion, such as sintering, reduction, oxidation, diffusion bonding, or brazing.

For example, the contact layer 27 (such as a contact layer metal or metal oxide containing ink) can be screen printed on the ribs 10, on another layer (such as Ni mesh 31), or on the fuel cell anode electrode 3. The anode printing can be in a rib pattern aligned with the ribs 10, perpendicular to the ribs (or at any other angle), or as a continuous layer.

For interconnects containing a screen-printed contact layer 27 and mesh 31 soldered to ribs 10, it is preferred that the contact layer 27 not coat the entire top surface of all ribs to provide uncoated areas where the mesh 31 will be soldered to the ribs 10. Generally, screen-printed layers can have poor electrical conductivity. Therefore, the contact layer 27 pattern preferably has gaps 35 to accommodate the mesh solder points. For example, as shown in FIG14A, there may be four solder points 37A, 37B, 37C, and 37D for soldering, one in each corner of the interconnect. The contact layer 27 pattern should have gaps 35 that expose these points 37A to 37D so that they are not coated by the contact layer 27.

14A contains printed lines that do not extend all the way to the ends of the ribs to expose solder joints 37A-37D in stripe gaps 35. However, the areas covered with anode contact layer 27 will have better electrical contact with the web than the areas in gaps 35.

A more complex pattern is shown in FIG14B. This contact layer 27 pattern has more ribs coated with anode contact ink, thereby forming a better electrical contact. FIG14C illustrates the intermediate background between the patterns of FIG14A and 14B. The contact layer 27 pattern in FIG14C is cross-shaped, so that the gaps 35 are only located at the corners of the interconnect. Because the contact layer 27 ink has a finite thickness and compressibility, sufficient gap 35 space should be provided around the weld points 37A to 37D so that the mesh does not deform too much when pressed down on the bare rib 10 to be welded. The pattern shown in FIG. 14C has ample room to accommodate weld points 37A-37D and variations in the automated welding process, while still maintaining good contact with the ends of the ribs in the center of the cell, which is important for electrochemistry and recombination.

The patterns described above are not limited, and other contact layer 27 patterns may be used. For example, a hatched ("dashed") pattern may be used. Printing the contact ink in dashed (i.e., non-continuous) lines may be advantageous. This can increase local contact pressure and thus electrical contact. In another configuration, every printed rib, or every second rib (or every third rib, etc.), is printed with the contact layer ink.

The contact layer may have a thickness of 5 microns to 1000 microns (e.g. 25 microns). If a thicker print is required, multiple layers may be screen printed in successive steps. The anodic contact ink should be printable, sufficiently stable under ambient conditions and printed with suitable wear resistance. The powder may be a metal, a metal alloy or an oxide, such as nickel oxide which is reduced to Ni metal during operation. The ink may contain a solvent, such as water, ethanol, ethylene glycol, terpineol, isopropyl alcohol, toluene, hexane or acetone. The ink may also contain a dispersant, a binder and/or a plasticizer. Anti-wear components may also be added to the ink. Depending on the particle size of the powder and the relevant surface area, a solids loading of 50-90% can be used, for example, about 80%. After printing, the ink can be dried in a low-temperature process to make the printed layer more stable and wear-resistant. It can be dried at a temperature of 80°C-200°C (for example, about 120°C).

The contact layer 27 can be optionally formed on the top of the rib, on the top and sides of the rib, or applied over at least a portion of the entire surface of the interconnect 10 and channels 8. Preferably, if the contact layer 27 is printed, it is located only on the top of the rib 10. If ink pours down into the fuel channels 8, fuel flow may be affected, which in turn may affect fuel distribution in the hot box. In severe cases, it may be necessary to reduce hot box fuel utilization, which reduces system efficiency and/or power output. The printing can be carefully aligned and periodically inspected. Manual inspection or an automated vision system can be implemented to screen out incorrectly printed interconnects. The ink can be tinted with contrast additives to improve the accuracy of automated vision systems. Screening should be carefully matched to the "pitch" (the distance between the ribs). Therefore, interconnect manufacturing variations may require multiple screens with different rib pitches to ensure well-registered printing.

In summary, the formation of the chromium oxide layer is reduced or avoided by at least one of increasing the press pressure between the nickel screen and the interconnect, providing undiffused Fe in the interconnect below the nickel screen, reducing surface contamination between the interconnect and the nickel screen, bonding the nickel screen to the interconnect, adding nickel to the interconnect alloy, or coating a metal or metal oxide contact layer over the ribs on the fuel side of the interconnect, including combinations of any two, three, four, five, or all six of the above steps.

In another embodiment, to reduce the cost of the MCO coating process, the MCO coating can be annealed (e.g., sintered or fired) during the sintering step of the powder metallurgy (PM) interconnect. The sintering of the PM interconnect 100 and the MCO coating on the interconnect can be performed in the same step in a reducing environment, such as a hydrogen reduction furnace with a dew point between -20 and -30°C and a temperature between 1300 and 1400°C for a duration between 0.5 and 6 hours. At these temperatures and oxygen partial pressures, the MCO coating will be completely reduced to Co-metal and Mn-metal. However, the melting temperature of Mn is about 1245°C, the melting temperature of Co is about 1495°C, and the Co-Mn system has a low-pressure liquidus. Therefore, sintering at temperatures between 1300 and 1400°C may result in the formation of an undesirable liquid phase.

Possible solutions to avoid liquid formation include lowering the sintering temperature to below 1300°C, such as below 1245°C, for example, 1100°C to 1245°C; increasing the oxygen partial pressure to reduce the Mn (but not oxidized Cr) in the MCO to MnO (melting temperature 1650°C) as the Mn-metal counterpart; lowering the Mn:Co ratio in the MCO to increase the melting temperature of the Mn-Co metal system; adding dopants to the MCO (such as Cr) to increase the melting temperature of the Co-Mn-Cr metal system; and/or adding dopants (such as Fe, V, and or Ti) to the MCO coating to stabilize binary and ternary oxides (to prevent reduction to the metallic phase). For example, at a sintering temperature of 1400°C, MnO is reduced to Mn-metal at a pO ₂ of 10-17 atm, while Cr ₂ O ₃ is reduced to Cr-metal at a pO ₂ of ^10-15 atm, which provides a small range (pO ₂ between ^10-17 and ^10-15 atm) where Cr is reduced to metal but MnO remains as an oxide with a high melting point. Therefore, the interconnect and MCO coating can be sintered at ^1300-1400 °C at pO ₂ = ^10-15-10-17 atm.

In another embodiment, the IC sintering step can be performed first, and then the MCO coating is applied to the sintered IC. The IC and coating are then subjected to the reduction step described in the previous embodiment, which is more suitable for the MCO coating.

In another embodiment, interconnect manufacturing costs can be reduced by depositing the MCO layer as a mixture of reduced components (e.g., MnO, CoO, Mn metal, Co metal, or any combination of these components). The mixture is then sintered, preferably under low _pO2 conditions. However, such sintering may be easier or the starting materials may be more dense, thereby reducing sintering time. Additionally, these precursor particles may be significantly less expensive than MCO precursors, which require expensive synthesis methods to produce.

Additionally, before coating the interconnect with the MCO layer, a sandblasting step can be performed to remove native chromium oxide layers from both the air and fuel sides of the interconnect. To reduce costs, native oxide can be removed only from the air side of the interconnect before forming the MCO coating on the air side. The MCO coating is then deposited on the air side and the interconnect is annealed as described above. After the annealing is complete, oxide can then be removed from the fuel side, such as by sandblasting. In this way, the number of sandblasting steps is reduced because an additional sandblasting step is not required to remove oxide growth on the fuel side of the interconnect that occurs during the annealing of the MCO coating.

In other embodiments, the composition of the MCO coating is modified to enhance stability at SOFC operating temperatures (e.g., 800-1000°C). The MCO composition of some previous embodiments is Mn _1.5 Co _1.5 O ₄ . This material has high electrical conductivity. However, the MCO material can be reduced to the binary oxides, MnO and CoO, or to the binary oxides, MnO and Co-metal.

In some fuel cell geometries, the MCO coating is directly exposed to the fuel flow only at the riser openings 16A, 16B. This fuel/coating interface can be eliminated by not coating the flat areas 17 around the openings (Figure 6B). However, interconnects manufactured by powder metallurgy methods have portions with some connected (open) pores that allow fuel to diffuse through these portions to the air side. The fuel that diffuses through the pores can react with and reduce the MCO at the MCO/interconnect interface (shown in Figure 9), thereby producing a porous layer composed of MnO and Co-metal. As shown in Figure 9, the coating/IC interface can be damaged, leading to adhesion failure and separation of the cell from the interconnect during conventional handling.

When exposed to a fuel environment, it is desirable to have a coating material that is more stable and less likely to reduce. The embodiments described below optimize the composition and/or dope the MCO with other elements to stabilize the material in a reducing atmosphere.

Figures 11 and 12 illustrate the theory of electrolyte corrosion. In the prior art SOFC stack shown in Figures 11 and 12, the LSM coating 11 on the interconnect contacts an annular seal 15. Seal 15 contacts the cell electrolyte 5. Without wishing to be bound by a particular theory, it is believed that manganese and/or cobalt from the metal oxide (e.g., LSM of LSCo) layer 11 contains manganese and/or cobalt, which leach into and/or react with the glass seal 15 and is then transported from the glass to the electrolyte. Manganese and/or cobalt can be transferred from the glass to the electrolyte as manganese and/or cobalt atoms or ions, or as compounds containing manganese and/or cobalt (e.g., manganese and/or cobalt-rich silicate compounds). For example, it is believed that manganese and cobalt react with the glass to form a (Si, Ba)(Mn, Co)O _6+δ mobile phase, which is transferred from the glass seal to the electrolyte. Manganese and/or cobalt in or at the electrolyte 5 (e.g., as part of the mobile phase) tends to accumulate at the grain boundaries of zirconium oxide-based electrolytes. This results in intergranular corrosion and pitting that weakens the electrolyte grain boundaries, ultimately leading to cracks in the electrolyte 5 (e.g., the opening 16A to opening 16B crack). Without being bound by a particular theory, it is also possible that the fuel (e.g., natural gas, hydrogen, and/or carbon monoxide) passing through the fuel inlet riser 36 may also react with the metal oxide layer 11 and/or the glass seal 15 to produce a mobile phase and enhance the filtration of manganese and/or cobalt from the layer 11 into the seal 15, as shown in FIG. 11 .

As discussed above, in other embodiments, the composition of the MCO coating is modified to increase stability at SOFC operating _temperatures (e.g., 800-1000°C). Thus, the MCO composition can be optimized based on stability and conductivity. Example compositions include, but are not limited to _{, Mn2CoO4} , _{Mn1.75Co0.25O4 , Co1.75Mn0.25O4} , _{CoMnO4 , and Co2.5Mn0.5O4 .}

Based on the phase diagram ( FIG. 10 ) and from a stability perspective, it may be beneficial to have a multiphase composition rich in Mn, such as Mn _2.5 Co _0.5 O ₄ and Mn _2.75 Co _0.25 O ₄ (e.g., a Mn:Co atomic ratio of 5:1 or greater, such as 5:1 to 11:1). Higher Mn contents may also result in more stable compositions because the compositions are in a higher oxidation state than the two-phase spinel + binary oxide found at high Co contents. However, any composition within the (Mn, Co) ₃ O ₄ family between the final compositions of Co ₃ O ₄ and Mn ₃ O ₄ may be suitable.

In another embodiment, MCO is stabilized by adding an additional, less reducible dopant. For example, MCO is known to react with Cr in IC alloys to form (Cr, Co, Mn) ₃ O ₄ spinel. Intentionally adding Cr to an MCO coating at low levels, such as 0.1 atomic % to 10%, produces a spinel (Cr, Co, Mn) ₃ O ₄ that is more stable than MCO because Cr ³⁺ is extremely stable. Other transition metal elements soluble in the spinel structure that can enhance stability include Fe, V, and Ti. Example coating materials include spinel (Fe, Co, Mn) ₃ O ₄ with 1% to 50 at% Fe, (Ti, Co, Mn) ₃ O ₄ with 1% to 50% Ti, or a combination of (Fe, Ti, Co, Mn) ₃ O ₄ .

Adding Ti can lead to more stable secondary phases, including _Co2TiO4 , _Mn2TiO4 , or _Fe2TiO4 . These phases contribute to overall coating stability. Spinels with any combination of the dopants mentioned above may include (Fe, _Cr , Co, _Mn ) _3O4 , ( _Cr , Ti, _Co , Mn) _3O4 , _etc.

Spinels based on Mg, Ca, and Al are known to be extremely stable and impedance-reducing. However, these spinels have low electrical conductivity and are therefore not well-suited for interconnect coatings. In contrast, doping conductive spinels such as MCO with low levels of Ca, Mg, and/or Al increases the material's stability while only slightly reducing conductivity. Example spinels include (Ca, Co, Mn) ₃ O ₄ with 1% to 10 at% Ca, (Mg, Co, Mn) ₃ O ₄ with 1% to 10 at% Mg, (Al, Co, Mn) ₃ O ₄ with 1% to 10 at% Al, or combinations such as (Ca, Al, Mn, Co) ₃ O ₄ , where Ca, Al, and/or Mg are added at 1-10 at%. Si and Ce are other elements that can be used as dopants (1-10 at%) in MCO spinel.

In addition to the methods described above that fall under the general category of material-specific stabilization efforts, alternative embodiments are provided for designing changes that can be combined with or used as an alternative to the above embodiments to improve the stability of the coating. In a first alternative embodiment, a stable barrier layer can be added to the interconnect before the MCO coating is added. This barrier layer will preferably be made of an oxide that is more stable than MCO, will be conductive, and will be thin enough not to adversely affect the conductivity of the interconnect assembly. Furthermore, this barrier layer is preferably dense and hermetic. Example barrier layers include, but are not limited to, layers of Ti-doped oxides (e.g., _TiO2 ) or strontium manganate (LSM).

A second alternative embodiment involves adding a reactive barrier layer between the interconnect and the MCO coating, which includes any of the elements discussed above as possible dopants (e.g., Cr, V, Fe, Ti, Al, Mg, Si, Ce, and/or Ca). When the interconnect is heated to standard operating temperatures (800-1000°C), this layer diffuses these elements into the MCO coating, creating a gradient doping profile with a higher concentration of dopants at the interconnect interface where reduction occurs. In this way, the majority of the coating contains relatively few dopants, and thus the conductivity can be less affected by the uniform doping of the coating material. The reactive layer is a metal layer (e.g., Ti or a metal-containing compound that allows metal to diffuse outward at 800°C or higher).

A third embodiment involves designing the interconnect material to contain reactive dopants (e.g., Si, Ce, Mg, Ca, Ti, and/or Al for a Cr-4-6% Fe interconnect) that are diffused into the MCO coating in the same manner as just described. Thus, the interconnect will contain ≥90 wt% Cr, 4-6 wt% Fe, and 0.1-2 wt% Mg, Ti, Ca, and/or Al.

Additionally, any method of depositing or treating the IC to reduce or seal the porosity of the component, beyond standard oxidation methods, will help limit MCO coating reduction. For example, a Cr layer can be electroplated onto porous portions before the MCO annealing step to further reduce porosity. Alternatively, as described above, adding a reactive barrier layer, if dense and sealed, will also reduce or block hydrogen diffusion from surface pores.

To ensure good long-term electrical connection between the Ni mesh 31 and the Cr-Fe alloy rib 10, it is desirable to have Fe at the top of the rib 10 to ensure good long-term electrical connection between the Ni mesh 31 and the top of the Cr/Fe rib 10. However, in the case of interconnect 100 having an iron content of 4 to 6 wt%, the Fe particles are insufficient to ensure that all, or at least most, of the Ni is in contact with the Fe-containing Cr. By intentionally placing Fe at the top of the rib 10, preferably before sintering, the electrical contact life can be improved without unacceptably changing the thermal expansion coefficient of the entire interconnect 100.

The following non-limiting example methods of increasing the iron concentration at the top of the interconnect rib 10 compared to the base portion of the interconnect 100 between the channel 8 and the rib on the opposite side of the interconnect include (1) placing pure Fe particles, such as 99% pure, such as 99.9% pure, at the top of the rib 10, (2) using Fe foil and/or (3) placing Cr containing a higher concentration of Fe across the fuel side of the rib 10.

The particles can be applied to the top of the ribs 10 by spraying the particles over the fuel side of a pressing tool (e.g., a mold 201 containing a cavity 200) that contains ribs 310 separated by channels 308 in an opposite configuration to the interconnecting ribs 10 and channels 8. Gravity holds the particles in the bottom of the channels 308 of the pressing tool. The ribs 310 and channels 308 can be located in the lower punch of the pressing tool if the mold comprises a hollow cylinder with upper and lower punches, or in the bottom of the mold 201 if the bottom of the mold is fixed and the tool contains only one upper punch.

As shown in FIG15A , pure Fe powder 205P (e.g., 99.9 wt% pure Fe particles) is provided (e.g., sprayed) into the channel 308 of the mold cavity 200. As shown in FIG15B , a powder 202 comprising 4-6 wt% Fe, 0-1 wt% Y, and the balance Cr is provided on top of the pure Fe powder 205P in the channel 308 of the mold cavity 200. Then, as also illustrated in FIG15B , the powder 202 is pressed using a punch 208 (e.g., an upper punch). As illustrated in FIG15C , the result is an interconnect 100 having an iron-rich region 129 at the top of the rib 10 on the fuel side of the interconnect 100. Region 129 has greater than 10 wt% Fe (e.g., 15-99 wt% Fe, e.g., 25-75 wt% Fe), optionally 0-1 wt% Y, and the balance chromium. This method has the advantage of being inexpensive and resulting in lower high area ratio resistive degradation (ASRD). A feature of this embodiment is that a small amount of iron powder grains are placed in place using a spraying method, rather than a powder shoe method.

In another embodiment, as illustrated in FIG16 , an Fe foil strip 205S is placed in the channel 308 of the cavity 200 of the press tool 201. In one embodiment, the Fe foil strip 205S can be 20-80 microns thick, such as 30-70 microns thick. The foil strip 205S can be approximately one-half to one millimeter wide, such as 0.25 to 0.75 mm wide. In one embodiment, the foil strip 205S can be held in place at the bottom of the channel 308 using an adhesive 312 (such as an acrylic adhesive or another suitable adhesive). An advantage of this method is that the amount of Fe provided to the top of the rib 10 of the interconnect 100 is more controlled than with powder spraying methods. Similar to the method illustrated in FIG15B , a powder 202 comprising 4-6 wt% Fe, 0-1 wt% Y, and the balance Cr is provided atop an Fe foil strip 205S in a channel 308 of a cavity 200 of a mold 201. Powder 202 is then pressed using a punch 208. The result is an interconnect 100 having an Fe-rich region 129 at the top of the rib 10, similar to the interconnect 100 illustrated in FIG15C . Advantageously, when a shoe filled with Cr-Fe 4-6 wt% powder is used, the displacement of Fe particles associated with the above method is avoided.

17A , pure iron powder (e.g., 99% pure iron powder) or Fe—Cr powder 305 having a higher Fe concentration (e.g., 25-50 wt% Fe, 50-25 wt% Cr) than the base portion of the interconnect 100 is provided to the channel 308 of the cavity 200 of the pressing tool 201 before providing Fe—Cr powder 202 having a lower Fe concentration (e.g., 4-6 wt% Fe) to the cavity 200 of the pressing tool 201. In this embodiment, the first shoe 206 is used to place the higher Fe-containing powder 305 into the rib channel 308. After the second shoe 206 is filled with the remaining space in the cavity 200 with the powder 202 comprising 4-6 wt% Fe, 0-1 wt% Y, and the balance Cr, as illustrated in FIG17B , a punch 208 is used in a pressing step, similar to that illustrated in FIG15B , to reduce the height of the area containing the powder 305 by approximately half. The target value of at least approximately 25% Fe is sufficient to provide sufficient Fe particles at the top of the rib 10 of the interconnect 100 without unacceptably affecting the CTE of the interconnect. In this embodiment, the rib 10 of the interconnect 100 has an intentional gradient in the Fe concentration of the powder.

In another embodiment, the fuel side of the metal interconnect 100 contacts the nickel screen 31 in the anode chamber in the presence of a wet fuel atmosphere. Embodiments include materials and methods that modify the chemical properties of the top of the rib 10 and the microstructure of the rib 10 to reduce the electrical resistance of the layer formed between the nickel screen 31 and the top of the rib 10. It is believed that depositing certain materials onto the top of the rib 10 will reduce the ohmic resistance. Such materials include Fe, Mn, Co, Cu, and high-temperature superalloys (including, but not limited to, Inconel 625 and 718, Haynes 230, and various Hastelloy alloys) and their oxides. The chemical formulas of Inconel Nickel 625, 718, Haynes 230 and various Herschel alloys are provided below. Cr Mo Co Nb+Ta Al Ti C Fe Mn Si P S Ni Min 20 8 -- 3.15 -- -- -- -- -- -- -- -- The rest are Max twenty three 10 1 4.15 0.4 0.4 0.1 5 0.5 0.5 0.015 0.015 The rest are Inconel 718 (wt%) Ni Cr Fe Mo Nb & Ta Co Mn Cu Al Ti Si C S P B 50.0-55.0 17.0-21.0 The rest are 2.8-3.3 4.75-5.5 ≤1.0 ≤0.35 ≤0.3 0.2-0.8 0.65-1.15 ≤0.35 ≤0.08 ≤0.015 ≤0.015 ≤0.006 Haynes 230 C Mn Si P S Cr Co Fe Al Ti B Cu La W Mo Ni 0.05-0.15 0.30-1.0 0.25-0.75 0.03 0.015 20.0-24.0 5 3 0.20-0.50 0.1 0.015 0.5 0.005-0.05 13.0-15.0 1.0-3.0 Rem Hoechst combination alloy* C% Co% Cr% Mo% V% W% Ai% Cu% Nb % Ti% Fe% Ni% other% Herschel alloy 0.1 1.25 0.6 28 0.3 - - - - - 5.5 Remaining/Remainder is Mn 0.80; Si 0.70 Herschel alloy 0.02 1 1 26.0-30.0 - - - - - - 2 Remaining/Remainder is Mn 1.0, Si 0.10 Hoechst Alloy C 0.07 1.25 16 17 0.3 40 - - - - 5.75 Remaining/Remainder is Mn 1.0; Si 0.70 Hoechst Alloy C4 / Hoechst Alloy C-4 0.015 2 14.0-18.0 14.0-17.0 - - - - - 0..70 3 Remaining/Remainder is Mn 1.0 ; Si 0.08 Hoechst Alloy C276 / Hoechst Alloy C-276 0.02 2.5 14.0-16.5 15.0-17.0 0.35 3.0-4.5 - - - - 4.0-7.0 Remaining/Remainder is Mn 1.0; Si 0.05 Hoechst Alloy F 0.02 1.25 twenty two 6.5 - 0.5 - - 2.1 - twenty one Remaining/Remainder is Mn 1.50; Si 0.50 Herster Alloy G 0.05 2.5 21.0-23.5 5.5-7.5 - 1 - 1.5-2.5 1.7-2.5 - 18.0-21.0 Remaining/Remainder is Mn 1.02.0; P0.04; Si Herster Alloy G2 / Herster Alloy G-2 0.03 - 23.0-26.0 5.0-7.0 - - - 0.70-1.20 - 0.70-1.50 Remaining/Remainder is 47.0-52.0 Mn 1.0; Si 1.0 Herschel alloy N 0.06 0.25 7 16.5 - 0.2 - 0.1 - - 3 Remaining/Remainder is Mn 0.40; Si 0.25; B 0.01 Herster Alloy S 0.02 2 15.5 14.5 0.6 1 0.2 - - - 3 Remaining/Remainder is Mn 0.50; Si 0.40; B0.0009; La 0.02 Herster Alloy W 0.06 1.25 5 24.5 - - - - - - 5.5 rest/bal Mn 0.050; Si 0.50 Herster Alloy X 0.1 1.5 twenty two 9 - 0.6 - - - 18.5 - Remaining/Remainder is Mn 0.6; Si 0.60

For example, if an Fe capping layer is deposited and _{metallurgically} bonded to the rib top, _Cr₂O₃ does not form between the Fe capping layer and the Cr-Fe top of the rib 10. However, Cr can still diffuse from the interconnect rib through the Fe capping layer and deposit on the nickel wire mesh 31 via Cr evaporation and deposition. Therefore, the "new" interface capping layer will be between the Fe capping layer and the nickel mesh 31. If an oxide forms at this interface, it will be an oxide of an Fe-Ni-Cr alloy (Fe and Ni are in contact with the addition of Cr by evaporation/diffusion). This Fe-Ni-Cr alloy can be an austenitic phase and bond the nickel mesh 31 to the Fe capping layer in the absence of oxides. However, if an oxide forms, it will be a (Fe, Ni, Cr) ₃ O ₄ spinel phase, which is more conductive than Cr ₂ O _3. In the case of a Mn capping layer bonded to the top of the rib 10, the principle is similar, but the phase will be different. For example, in addition to the Mn-Ni-Cr alloy, a highly conductive (Mn, Ni, Cr) ₃ O ₄ oxide can also be formed. However, this also forms a lower resistance interface.

Another embodiment of reducing the ohmic resistance of the oxide on top of the rib 10 includes depositing a capping layer of inconel or other Ni-Cr superalloy on top of the Cr-Fe interconnect material. As is well known, the oxidation rate of Cr-Fe, and therefore the subsequent Cr ₂ O ₃ layer thickness, is much greater than that of Ni-Cr superalloy. By depositing a capping layer of dense superalloy and metallurgically bonding it to the top of the rib 10, the thickness of the Cr ₂ O ₃ layer formed under the nickel wire (on top of the capping layer) should be significantly thinner than that of the Cr-Fe alloy, such as less than half the thickness, such as between 10-50% the thickness. In addition, many superalloys form dual oxide coatings consisting of Cr ₂ O ₃ and (Mn, Cr) ₃ O ₄ spinel, which is more conductive than Cr ₂ O ₃ and has lower Cr-evaporation.

In the above-described embodiment, an additional benefit is that the metal capping layer 470 can fill the pores and increase the density of the top portion of the rib 10. There is some evidence that a porous and poorly formed rib top portion can lead to excessive oxide formation and a thicker oxide layer. In embodiments where the metal or metal oxide capping layer 470 is formed on the top of the rib 10 of the interconnect 100, the Ni contact layer 27 described with respect to FIG. 13 can be omitted. However, the Fe-rich top portion of the rib 10 of the embodiment of FIG. 15A to 17B can be included below the capping layer 470 or omitted.

To achieve a dense metal coating on top of the sintered interconnect 100 formed from powder metallurgy, a laser can be used to melt the metal powder that has been deposited on top of the rib 10. In a first embodiment, the interconnect 100 is placed in a bed of metal powder of the desired coating material. Additional powder is then spread over the interconnect 100 to fill the channels 8 and reach a given depth (e.g., 10-50 μm) on top of the rib 10. The laser then scans the top of the rib 10 to subsequently melt the metal powder and bond it to the top of the rib 10. In a second embodiment, a selected metal powder is mixed with one or more organic compounds and screen-printed onto the top of ribs 10 of the desired width and thickness. A laser is then scanned across the top of the ribs 10 to melt and bond the powder. In both cases, an inert atmosphere (argon, neon, helium, etc.) should be selected to prevent the metal powder from oxidizing. Other parameters (such as laser power, laser type, scan rate, and powder particle size) can be optimized for each material combination to achieve the desired bonding, thickness, and density of the metal coating.

18A , the metal or metal oxide capping layer 470 is formed only on the top surface of the rib 10, and not on the sidewalls of the rib 10 or the bottom of the channel 8. In an alternative embodiment shown in FIG18B , the metal or metal oxide capping layer 470 is formed on the top of the sidewalls of the rib 10 and at least a portion thereof.

The capping layer 470 can be advantageously formed at different stages of interconnect 100 fabrication. For example, the capping layer 470 can be added to the interconnect 100 after pressing or after oxidation of the interconnect 100, each of which has different advantages. For example, after pressing a powder comprising 4-6 wt% Fe, 0-1 wt% Y, and the balance Cr, adding the Fe capping layer 470 to the top of the rib 10 will result in a higher density of Fe-rich regions, while still containing enough Cr to prevent Fe oxidation. This ensures that sufficient Fe metal remains on the top of the rib 10 to perform its resistance-reducing function.

If Fe is added after oxidation of the pressed and sintered interconnect 100 (e.g., as described in U.S. Patent No. 8,652,691, which is hereby incorporated by reference in its entirety), subsequent exposure to water (e.g., wet fuel) in the SOFC stack during stack operation can result in some oxidation of the Fe capping layer 470. Some oxidation states of iron are electrically conductive as oxides. This has the advantage of producing even higher local concentrations of Fe without negatively impacting the interconnect's coefficient of thermal expansion.

In one embodiment of forming the Mn capping layer 470, due to the high vapor pressure of Mn, it is preferred to form the Mn on top of the ribs 10 after sintering and before oxidizing the interconnect 100. The subsequent pure oxidation step then forms a higher quality, more conductive manganese oxide compared to oxidation of the Mn in wet fuel as part of SOFC operating conditions.

In an alternative embodiment shown in FIG18C , a metal capping layer 570 is formed on top of the air-side ribs 10 of the interconnect 100. This metal capping layer 570 can also reduce ASRD at the air-side interface. As described in the previous embodiments, the air-side of the interconnect 100 can be coated with an LSM, MCO, or LSM/MCO composite material that reduces or prevents Cr-evaporation by atmospheric plasma spraying (APS). However, a resistive Cr ₂ O ₃ layer can still be formed beneath the coating. In one embodiment, the interconnect is sandblasted, the top of the air-side rib is capped with a metal capping layer 570 (e.g., Inconel nickel alloy), and then the entire air-side of the interconnect is coated with an APS LSM, MCO, or LSM/MCO layer. Inconel nickel (or other superalloys) preferably reduces _the thickness of _Cr2O3 formed on the top of the rib. In addition, because the top of the rib 10 carries a larger proportion of the current, _the reduced _Cr2O3 thickness reduces ASRD on the air-side. Similarly, the Fe or Mn capping layer promotes the formation of (Cr, Fe) ₃ O ₄ and (Cr, Fe) ₃ O ₄ spinel phases, which are more conductive than the Cr ₂ O ₃ native oxide skin. Capping layer 570 can be formed in addition to, or in place of, the metal or metal oxide capping layer 470 on the fuel side of the interconnect 100.

In another alternative embodiment shown in FIG. 18D , the first metal coating 470 described above with respect to FIG. 18A or 18B is formed on the fuel side of the rib 10 , while the second metal coating 570 described above with respect to FIG. 18C is formed on the top of the air side rib 10 of the interconnect 100 .

Thus, in an alternative embodiment, a metal or metal oxide coating (e.g., Fe, Mn, superalloy, Co, Cu, etc.) 570 is formed on top of the ribs 10 on the air side of the interconnect 100 beneath the LSM, MCO, or LSM/MCO coating 111, which covers the ribs 10 and channels 8 of the interconnect 100 on the air side of the coating.

The embodiments illustrated in Figures 15A to 17B include a method of preparing an interconnect for a solid oxide fuel cell stack, comprising providing an iron-rich material (205P, 205S, 305) containing at least 7 wt% (e.g., at least 10 wt%, such as at least 20 wt%, such as at least 25 wt% iron, for example, from about 10 wt% to about 99.9 wt%, such as from about 10 wt% to about 95 wt% iron, for example, from about 20 wt% to about 80 wt% iron, including from about 25 wt% to about 50 wt% iron) into a channel 308 of a mold 201; providing an iron-rich material (205P, 205S, 305) containing 4-6 wt% Fe, 0-1 wt% A powder 202 of Y and the remainder Cr is provided over the iron-rich material (205P, 205S, 305) in a mold 201; the iron-rich material and powder are pressed in the mold 201 (e.g., using a punch 208) to form an interconnect 100; and the interconnect 100 is sintered to form a sintered interconnect having an iron-rich region 129 having an iron concentration that is 10% greater than the iron concentration of the rib 10 of the interconnect 100, as shown in FIG. 15C .

In one embodiment, the method further includes connecting a Ni mesh 31 (shown in FIG13 ) to the iron-rich region 129 in the rib 10 on the fuel side of the interconnect 100. In one embodiment, the iron-rich region 129 comprises 10-99 wt% iron, such as 15-99 wt% iron, such as 20-80 wt% iron, such as 25-75 wt% iron. The interconnect 100 can be placed in a solid oxide fuel cell stack containing solid oxide fuel cells.

In one embodiment, the iron-rich regions 129 are located only in the top ends of the ribs 10 on the fuel side of the interconnect beneath the nickel screen 31 (shown in FIG. 13 ) that contacts the top ends of these ribs. In another embodiment, the iron-rich regions 129 are located only in the top ends of the ribs 10 on the air side of the interconnect 100. In yet another embodiment, the iron-rich regions 129 are located in the top ends of the ribs 10 on both the fuel and air sides of the interconnect.

In one embodiment shown in FIG. 15A , providing the iron-rich material containing at least 25 wt % iron into the channel of the mold comprises spraying iron powder 205P containing at least 99 wt % iron into the channel 308 of the mold 201 .

In another embodiment shown in Figure 16, the step of providing the iron-rich material into the channel of the mold includes providing the iron foil strip 205S into the channel 308 of the mold 201. Optionally, an adhesive 312 may be disposed between the iron foil strip 205S and the channel 308 of the mold 201.

In another embodiment shown in FIG17A , providing the iron-rich material into the channel of the mold includes providing a powder 305 comprising pure iron or chromium and at least 25 wt% iron into the channel 308 of the mold 201 using a first shoe 206. As shown in FIG17B , providing a powder comprising 4-6 wt% Fe, 0-1 wt% Y, and the balance Cr includes providing a powder 202 comprising 4-6 wt% Fe, 0-1 wt% Y, and the balance Cr into the mold 201 above the powder 305 comprising pure iron or chromium and at least 25 wt% iron using a second shoe 206.

18A-18C , a method of preparing an interconnect for a solid oxide fuel cell stack includes providing a chromium alloy interconnect 100 having ribs 10 separated by channels 8, and forming a metal or metal oxide coating (470, 570) directly on the top surfaces of the ribs 10 on the air and fuel sides of the interconnect, rather than on the bottom of the channels 8. The metal or metal oxide coating (470, 570) includes iron, manganese, cobalt, copper, a superalloy, or an oxide thereof.

In one embodiment shown in Figure 18A, the covering layer 470 is located only on the top surface of the rib 10 and not on the side surfaces of the rib 10. In another embodiment shown in Figure 18B, the covering layer 470 is located on the top surface of the rib 10 and on the side surfaces of the rib 10, but not on the bottom of the channel 8.

In one embodiment, the method further includes pressing a powder 202 comprising 4-6 wt% Fe, 0-1 wt% Y, and the balance Cr in a mold 201 to form the interconnect 100; and sintering the interconnect before forming the metal or metal oxide coating (470, 570); and placing the interconnect into a solid oxide fuel cell stack containing solid oxide fuel cells after forming the metal or metal oxide coating.

In one embodiment, the iron-rich region 129 described above and illustrated in FIG. 15C may be located beneath the metal or metal oxide blanket ( 470 , 570 ) on the fuel side and/or the air side of the interconnect 100 .

In one embodiment shown in FIG. 18C , the method further includes forming a strontium manganate layer, a manganese cobalt oxide layer, or a composite strontium manganate/manganese cobalt oxide layer 111 on the metal or metal oxide capping layer 570 and on the bottom of the channel 8 .

Although the foregoing relates to preferred embodiments, it should be understood that the present invention is not limited thereto. Those skilled in the art will recognize that various modifications may be made to the disclosed embodiments and that such modifications are intended to be within the scope of the present invention. All publications, patent applications, and patents cited herein are hereby incorporated by reference in their entirety.

3: Anode 5: Battery electrolyte 8: Airflow channel 10: Rib 11: LSM coating 13: One side 14: Opposite side 15: Ring seal 16A: Fuel inlet opening/through-hole 16B: Fuel outlet opening/through-hole 17: Flat area/fuel distribution plenum 18: Window seal 19: Strip seal 25: Oxide layer 27: Metal or metal oxide contact layer 29: Reaction zone 31: Nickel mesh 33: Impurities 35: Gap 36: Fuel inlet riser 37A: Welding point 37B: Welding point 37C: Welding point 37D: Welding point 100: Interconnect 101: Spinel phase/chromium crystals 102: Spinel layer 103: Cracks 104: LSM interconnect coating 111: Coating 129: Iron-rich region 200: Mold cavity 201: Mold 202: Lubricant and Cr/Fe powder 204: Coating material powder 205P: Pure Fe powder 205S: Fe foil 206: Second shoe 208: Punch 305: Fe-Cr powder 308: Channel 310: Rib 312: Adhesive 470: Metal cover 570: Metal cover

FIG1 is a micrograph showing the Mn-Cr spinel phase in the pores of the LSM-based cathode.

Figure 2 is a micrograph showing the Cr-containing phase in the cracks of the LSM interconnect coating deposited by air plasma spraying. The SOFC stack was operated at 850°C for 2000 hours.

3A-3C are schematic diagrams illustrating steps in a method of making an interconnect according to one embodiment.

FIG4 is another schematic diagram of a method for preparing an interconnect according to one embodiment.

FIG5 is a schematic side view of one embodiment of an interconnect having a double-layer composite coating.

Figures 6A to 6B and Figure 7 are schematic diagrams illustrating the following: (6A) the air side of the interconnect according to one embodiment, (6B) a close-up view of the sealing portion of the air side of the interconnect, and (7) the fuel side of the interconnect.

FIG8 is a micrograph illustrating chromium oxide on the fuel side (uncoated side) of the interconnect after the reduction sintering step.

FIG9 is a micrograph of a portion of a SOFC stack illustrating the reduction of the MCO coating at the coating/IC interface (in the ribbon seal area) due to diffusion of fuel through the porous IC.

Figure 10 is a phase diagram illustrating the Mn ₃ O ₄ —Co ₃ O ₄ system.

FIG11 is a schematic diagram of a fuel inlet riser in a conventional fuel cell stack.

FIG12 is a schematic diagram of a SOFC illustrating the electrolyte corrosion theory.

13 is a schematic side cross-sectional view of a portion of an interconnect according to one embodiment.

14A, 14B, and 14C are top views of interconnects according to alternative embodiments.

15A-15C are schematic diagrams of steps in a method of making an interconnect according to another embodiment.

FIG16 is a schematic diagram of a method for preparing an interconnect according to another embodiment.

17A-17B are schematic diagrams of steps in a method of making an interconnect according to another embodiment.

Figures 18A to 18D are schematic diagrams of interconnects prepared according to various embodiments.

8: Airflow channel

10: Ribs

100: Interconnectors

129: Iron-Rich Area

Claims (6)

A method of preparing an interconnect for a solid oxide fuel cell stack comprises: providing a chromium alloy interconnect having ribs separated by channels; and forming metal or metal oxide coatings on the air side and the fuel side of the interconnect, wherein the metal or metal oxide coatings comprise iron, manganese, cobalt, copper, a superalloy, or oxides thereof; wherein the metal or metal oxide coatings are located on the top surfaces of the ribs and on the sides of the ribs but not on the bottoms of the channels. A method for preparing an interconnect for a solid oxide fuel cell stack comprises: providing a chromium alloy interconnect having ribs separated by channels; forming metal or metal oxide coatings directly on the top surfaces of the ribs on the air and fuel sides of the interconnect, but not on the bottoms of the channels, wherein the metal or metal oxide coatings comprise iron, manganese, cobalt, copper, a superalloy, or an oxide thereof; and forming a strontium manganate (lanthanum strontium manganite) layer, a manganese cobalt oxide layer, or a composite strontium manganate/manganese cobalt oxide layer on the metal or metal oxide coatings and on the bottoms of the channels. The method of claim 2, wherein the step of forming a strontium manganate layer, a cobalt manganate oxide layer, or a composite strontium manganate/cobalt manganate oxide layer comprises forming the strontium manganate layer on the metal or metal oxide capping layers and on the bottoms of the channels. The method of claim 2, wherein the step of forming a strontium manganate layer, a cobalt manganate oxide layer, or a composite strontium manganate/cobalt manganate oxide layer comprises forming the cobalt manganate oxide layer on the metal or metal oxide capping layers and on the bottoms of the channels. The method of claim 2, wherein the step of forming a strontium manganate layer, a cobalt manganate oxide layer, or a composite strontium manganate/cobalt manganate oxide layer comprises forming the composite strontium manganate/cobalt manganate oxide layer on the metal or metal oxide capping layers and on the bottoms of the channels. An interconnect for a solid oxide fuel cell stack, comprising: a chromium alloy interconnect having ribs separated by channels; and metal or metal oxide coatings on the air and fuel sides of the interconnect, wherein the metal or metal oxide coatings comprise iron, manganese, cobalt, copper, a superalloy, or oxides thereof; wherein the metal or metal oxide coatings are on the top surfaces of the ribs and on the sides of the ribs but not on the bottoms of the channels.