Fabrication of flat stainless steel substrates with improved oxidation behavior for metal-supported solid oxide cells using aqueous tape casting

An aqueous tape casting procedure was developed and optimized to fabricate thick, flat tapes for use as porous stainless-steel substrates for metal-supported solid oxide cells (MS-SOCs). Curling tape is one of the main challenges when using aqueous based slurry formation. This work demonstrated that the sedimentation problem can be solved by increasing solid loading rather than adding excessive binder to raise viscosity. The effect of various casting surfaces on tape curling was also investigated. Materials that allow easy tape release resulted in flatter tapes once the water was evaporated. In addition, substrate oxidation resistance at high temperature was evaluated with respect to starting powder size, sintering extent, and pore former types. High sintering extent that removes or encloses the porosity between steel particles while retaining porosity left by pore formers can effectively prevent breakaway oxidation due to local chromium depletion. Carbon residue in the steel substrates from the slurry organic content can be decreased when formulating the slurry to prevent Cr-rich phase formation in the steel, which severely compromises the substrate oxidation resistance and ductility. By dwelling the substrate in high purity hydrogen, the sensitization can be reversed, but more detailed investigation of the reaction dynamics is needed. By combining the strategies described, this work produced crack-free, flat, 400–500 μm thick stainless steel substrates with 28.7 vol% porosity and improved oxidation resistance compared to previous substrates fabricated by dry pressing of fine powders.


Introduction
Solid oxide cells (SOCs) are promising devices for converting chemical energy into electrical energy.They offer high efficiency, tolerance of diverse fuel compositions, scalability, and modularity for various applications.SOCs can operate with electrical efficiency as high as 60% in the fuel cell (SOFC) mode [1], which is higher than other types of fuel cells.Oxide ions in SOFCs directly oxidize fuel at the anode, allowing the use of not only H 2 but also light hydrocarbon gases, carbon monoxide, and liquid fuels [2].Two-way operation of SOCs can be used with intermittent energy sources such as solar and wind to obtain stable output [3].The number of cells and cell stacks can be configured to make it possible for SOCs to fit a wide range of applications, from large stationary power plants to vehicular power [4]. 1 3 High fabrication cost, long fabrication time, and limited robustness are some of the primary obstacles that prevent SOCs from being widely commercialized.SOCs typically consist of three functional layers: air electrode, electrolyte, and fuel electrode, one of the cell layers typically needs to be bulky enough to provide mechanical strength, which adds to the material cost.Conventional fabrication methods for SOCs, such as screen printing and tape casting, involve suspending material particles in a pre-made solution, casting, drying, and firing to complete the process, which is timeconsuming and can take at least 2 days to complete the fabrication.In particular, each firing step takes approximately 1 day to complete [5].
Plasma spraying technology can offer a solution to the fabrication problems mentioned above by allowing the functional layers of SOCs to be deposited on a porous support structure made of stainless steel.Compared to traditional methods, plasma spraying is a rapid and well-established method to fabricate thin SOC functional layers, with a process time at the scale of minutes [6].Stainless steel is significantly cheaper than the ceramic materials commonly used in the functional layers and has a higher fracture toughness and a similar thermal expansion coefficient (TEC) to 8 mol% yttria stabilized zirconia (TEC YSZ ∼ 10.5 ppm K −1 and TEC 430L ∼ 11.4 ppm K −1 ) [7,8].Additionally, it has higher thermal shock resistance than ceramic materials.However, the microstructure of the metal support needs to be carefully engineered to prevent defects in the plasma-sprayed functional layers due to large surface pores and to minimize power losses from mass transfer impediments [9][10][11].One drawback of stainless steel is its susceptibility to oxidation, which can cause air electrode poisoning [11], structural degradation from oxide scale deposition [4], and rapid iron oxidation after local Cr depletion [12].
To fabricate SOFC stainless steel substrates, tape casting offers the advantage of producing material in a thin sheet format with the ability to control the thickness and final product size while maintaining high consistency of material properties at low cost [13].These advantages are even more prominent when compared to the die pressing method, which is another common fabrication method for SOC supports [14].The substrate fabrication is a distinct step from cell layer fabrication, and it adds additional time to the process.But overall, a single tape casting and firing step for the substrate with three additional plasma spraying steps takes less time than multiple tape casting or screen printing steps followed by additional firing steps to make all of the substrate and cell layers.
Tape casting starts by mixing powder with a solution of binder and other additives, such as defoamer, plasticizer, dispersant, and thickener.The binder and plasticizer provide the tape with sufficient strength for further shaping processes, while the dispersant prevents agglomerate formation, and the thickener increases viscosity to stabilize the slurry [15].The mixed slurry is typically de-gassed in a vacuum before the casting step to prevent pinholes in the cast tape.The tape is cast by passing the slurry under a blade, whose height controls the tape thickness.The casting surface supports the tape until the slurry dries.A common commercial option for the casting surface is Mylar film, which is often used on a conveyor, and is usually coated with silicone.Such an option is often used with organic slurry systems that have short drying times.After the solvent fully evaporates, the tape can be peeled off from the casting surface for its intended uses [16].
When fabricating substrates for SOCs, additional porosity is introduced by pre-mixing steel powder with other particles as pore formers in the slurry formation stage.The particles are later eliminated in subsequent thermal treatments and leave the tape in gas form, creating voids where they were located [17].The porosity volume fraction is crucial to SOC performance, as low gas permeability limits available reactant concentration at the electrodes, thus constraining the power output.To make the substrate compatible with the plasma spraying method, Metcalfe [18] reported that fine steel particles are required to prevent defects on the deposited layers, but this comes at the cost of fast substrate oxidation.
Among tape casting methods reported in the literature, organic solvent-based recipes are used more frequently for fast drying and better quality control than aqueous recipes.However, aqueous slip has the advantage of nonflammability, non-toxicity, no solvent recycling being required, and lower cost [19,20], although it is still necessary to investigate strategies to mitigate or overcome the disadvantages of the aqueous recipe.
The difficulty of low volatility is magnified when combined with thick casting, which is required to obtain sufficient mechanical strength as an SOC substrate.Chang et al. [21] reported thick slip casting with metal powder in a mold, while Mercadelli et al. [13] and Roosen et al. [22] provided organic solvent-based recipes for porous and dense substrates, respectively.Scheithauer et al. [23] reported a procedure for dense aqueous stainless steel tape casting, which is not intended for SOC application.
There are two criteria that determine if a continuous chromia layer can form on top of stainless steel.The first criterion is that the chromium concentration in the alloy must be high enough to prevent internal oxidation, caused by inward oxygen diffusion.The second criterion is that chromium diffusion from the bulk to the surface must be rapid enough to supply the consumed chromium [24].For porous substrates with small fine porosity, the high surface area increases the consumption rate of chromium, and the small thickness of solid between the small fine pores can prevent the chromium solid state diffusion within the material from being sufficient to replace the consumed chromium, as the reservoir of chromium supplying the nearby surfaces would be rapidly depleted.Sebastian Molin et al. [8] reported very different oxidation behaviors of porous 430 stainless steel sintered below 1100 °C from dense steel and they did not try to sinter the substrates to a denser state.
To prevent breakaway oxidation in SOC metal substrates, it is important to reduce the substrate surface to volume ratio.The traditional approach to achieve this goal is to use coarse metal powder as the starting material [9,18,21,25].However, this method still leaves some fine porosity between the coarser particles, which can be difficult to remove.The larger particles also present challenges in slurry development due to faster sedimentation.In this study, a new approach is proposed, which involves using sub-22 micron powder combined with large pore former particles to create uniform porosity with pore sizes matching that of the pore former particles.The gaps between steel particles in areas without pore former are sintered to near full density, so that the majority of the porosity is created by the pore former.The use of finer particles facilitates the dense sintering required to eliminate or close any fine porosity between unsintered steel particles.By doing so, the surface area to volume ratio is reduced to a lower range compared to substrates sintered to a lesser extent, significantly decreasing the chance for local Cr depletion and subsequent breakaway oxidation.
This study describes an aqueous tape casting procedure that produces porous, thick substrates from aqueous slurry despite the challenges described above.By applying the strategy stated above, we achieved properties that can alleviate the trade-off between spray deposition quality and substrate longevity.We also evaluated the effect of sintering conditions and tape composition on substrate oxidation resistance.

Starting materials and characterization methods
Three different stainless steel powders are used in this study: high chromium alloy 1C44Mo20 (Sandvik Osprey Ltd., Neath, UK) with sizes of − 22 μm and − 32 μm, plus medium chromium alloy 430L (Sandvik Osprey Ltd., Neath, UK) with a size of − 10 μm.The 1C44Mo20 is a specially designed composition for SOC interconnects and is a brand name from Sandvik.
For the fine particle size, 430L steel has a much lower cost than that of 1C44Mo20, yet sufficient oxidation resistance for solid oxide cell applications [26].The particle topographic images are taken using a scanning electron microscope (TM3000, Hitachi, Japan).Energy dispersive spectroscopy (EDS, 6610VL, JEOL, Japan) is used for elemental mapping of the particles and oxidized substrates.
Polyvinyl alcohol (98-99% hydrolyzed low molecular weight, Thermo Scientific, China) is the binder; the plasticizer is ethyl glycol (EG) (Thermo Scientific, USA), and antifoam 204 (Sigma Aldrich, Germany) is the defoamer.The pore formers used for this study include Poly(methyl-methacrylate) (PMMA) beads with sizes of 20 μm (Lamberti, Skedsmokorset, Norway), and food grade potato starch that was sieved to size 32-45 μm.Tapes with 55 vol% PMMA and 35 vol% starch were prepared and sintered, and their oxidation performance studied.Poly(ethyleneimine) (PEI) solution (M w ~ 2000, Sigma Aldrich, China) is the dispersant to stabilize the slurry.
The amount of dispersant is determined by measuring the viscosity (HAAKE Viscotester 7 plus, Thermo Scientific, Germany) of 15 vol% steel powder binder suspension while increasing dispersant to 3.2 wt% at a step size of 0.16 wt%.The slurry is diluted for easy removal of air bubbles could entrapped inside, yet the viscosity is high enough for the particle to stabled during measurement while being stirred.
The thermal properties of steel powder and slurry additives are studied by thermogravimetric analysis (SDT Q600, TA Instruments, USA) in air at 2 °C/min to determine the organic removal temperature profile.The sintered substrate structure is obtained by optical microscope (Axio Scope, Zeiss, Germany) and the porosity is analysed by Image J software.

Slurry preparation
To make the PVA binder solution, it is necessary to pre-mix it in hot water for complete dissolution.A stock binder solution is pre-made with a binder to solvent weight ratio of 10:43, which gives the desired binder content in the final tape.Steel and pore former particles are weighed, mixed at their target volume fraction, and then mixed with the binder solution and dispersant in a high-density polyethylene bottle with zirconia grinding media.The mixture is then mixed on a rolling mill (Model 784 VM, Stoneware, USA) for 18 h.The resulting slurry has a solid content of 33 vol%.Plasticizer can be added at this step, but for the purpose of making planar SOC substrates, no tape bending is needed, so it is not necessary.The milled slurry is then degassed in a vacuum chamber for 10 min.To prevent film forming on top of the slurry, an additional wet paper towel can be used to cover the plastic bottle during degassing.The finalized slurry is placed in a test tube to settle for 4 h, which is the time needed for tape to damp dry.If there is a thick layer of translucent fluid relative to the total slurry column height, it implies significant particle sedimentation, which can cause nonhomogeneous structure in the dry tape.Table 1 summarizes all the slurry formations used in this study as a weight percentage of the total slurry weight.Further description of the slurry formulation can be found in the corresponding sections of the text.

Tape casting
Once slurry is prepared, it is cast with a doctor blade with width of 100 mm on a flat surface at a height of 1 mm and a speed of 1 mm/s.The doctor blade is driven by a custom-made fixture, shown in Fig. 1, which utilizes a step motor connected to a threaded rod that goes through a brass block.As the rod rotates, the block moves along the rod, and it is connected to an aluminum plate that pushes the doctor blade to move at a fixed speed.Four different rigid plates, including boron glass, silicone rubber, polypropylene (PP) and polytetrafluoroethylene (PTFE) are used as casting surfaces to study their impact on tape flatness.All the plates have a width of 152 mm, and lengths vary from 152 to 304 mm.The as-cast tape is left on a bench to dry at room temperature without forced air flow.It takes up to 6 h for a 1 mm thick slurry to dry, and the dry green tape can be easily peeled off from the casting plate.

Thermal treatments
The green tapes were cut into 35 mm by 76 mm coupons, to fit into tube furnaces.The tapes first get organic content removed in an intermediate temperature tube furnace (Thermolyne 79325, Thermo Fisher).The organic removal step was carried out in air flowing at 500 SCCM up to 350 °C, with two 1-h dwelling steps at 180 °C and 300 °C.Air was then purged by a nitrogen and hydrogen mixture (96% N 2 and 4% H 2 ) before switching to pure hydrogen at 100 SCCM before continue ramping up temperature.The tape was pre-sintered at 850 °C for 2 h to provide sufficient mechanical strength for transfer to the final sintering tube furnace (1632 12HT, CM Furnaces Inc, Bloomfield, USA).The final sintering step was carried out in pure hydrogen (99.99999%) from an electrolyser (H2-800, Parker Balston) at 100 SCCM for various sintering times ranging from 2 to 6 h and temperatures from 1100 to 1280 °C.Alumina plates were used as the carrying supports during thermal treatments.Fig. 4 TGA result of a steel powders and b organic additives in air atmosphere up to 500 °C Fig. 5 Tape formulated from previous work [29].The binder and plasticizer content yields strong tape, but the resulting tape may warp after water is completely evaporated, with inconsistent morphology results

TGA profiles of organic additive and steel powder
TGA results of tape materials are important to determine the dividing point of oxidizing and reducing environments during the organic material removal processes.An oxidizing environment is required to effectively remove organic content, as residual carbon will significantly compromise the sintered substrate oxidation resistance [27].Meanwhile, the substrate cannot be significantly oxidized, so the flow gas needs to be switched to a reducing environment at higher temperature.Figure 4a reveals that the steel powder starts to have visible mass gain from 300 °C and it increases rapidly above 400 °C.This behaviour is more significant for the finer Cr16 powder.With a mass gain of 0.3 wt% and 0.2 wt%, this implies that 1.2 mol% and 0.6 mol% of Cr has been oxidized for the Cr16 and Cr22 powders, respectively, assuming Cr is the only oxidized element.Ethylene glycol is not included in the TGA analysis because it has a boiling point of 197 °C and is not normally added as a tape component.The TGA result from Fig. 4b for the organic additives shows that PMMA pore former has nearly complete burnout at 350 °C, while PVA and starch still have 47.5 wt% and 35.6 wt% residue left, respectively.This residue content decreased to 33.8 wt% and 21.9 wt% at 400 °C, which is still significant.Note that there is a bump in the PVA-TGA curve, possibly due to ignition.With the data listed above, the gas-switching point is determined to be 350 °C, which results in a low steel powder oxidation extent.400 °C is not chosen because despite the additional mass loss for both PVA and starch, the residual mass is still high until 450 °C at the continuous ramping condition.Therefore, a better strategy is to avoid using starch and to decrease the quantity of PVA to minimal amounts in the slurry formation stage.

Tape curling problem and observation
Bergner [28] reported that a 7 wt% PVA to steel powder ratio provides excellent tape flexibility.The slurry composition was further refined for porous stainless steel substrates using Cr22 powder with details described in previous work [29].The resulting slurry had a composition of 5 wt% (to stainless steel) PVA, 2.2 wt% (to binder solution) defoamer, 20 wt% (to PVA) ethylene glycol (EG) plasticizer and 33 vol% solid loading.However, this composition has a curling problem to be addressed, as shown in Fig. 5.This is the starting point of this work.
Fig. 8 Fixing warped 5 wt% binder tape (a) can be achieved by adding additional plasticizer to soften the tape and fix the shape using a weighted steel mesh while removing plasticizer at 120 °C for 2 h (b).The outgassed tape shows acceptable flatness (c) Fig. 9 Trimmed dry green Cr22 steel tape without pore former cast on a glass plate with 2.5 wt% binder (a) and 5 wt% binder (b).2.5 wt% binder yielded a flat tape, while with 5 wt% binder, the tape still has some warping issues, which increases the chance of crumbling after outgassing (c) Experimental observations suggest that the extent of tape curling is influenced by drying conditions, where shorter drying times result in more severe curling.During the drying process, a film forms on top of the slurry, and stress builds up at the edges of the tape, where water is lost more quickly.This stress causes deformation by compressing the wet and soft tape center.Uneven drying occurs when there is less lateral water diffusion within the tape plane.However, when the tape is small or thin, it can dry uniformly in both the lateral and longitudinal directions, resulting in a completely flat dried tape.Efforts were made to address the compression stress by trimming the dry edges while the center is still wet, covering the tape edges while keeping the center exposed to slow down the edge drying, and reducing overall air convection.While these methods can mitigate the curling problem, they do not fully solve it.In most cases, curled tapes are still acceptable for the next steps, but curvature increases the chance of the outgassed tape crumbling while transferring it to the final sintering furnace, since the curled green tape has little contact with the support plate.

Dispersant optimization
To determine the optimal amount of PEI required for full dispersion of the Cr16 powder, the first step was to analyze the rheological behavior of the slurry and identify a rotation speed range in which the viscosity is less dependent on rotation speed [10].The HAAKE Viscotester 7 plus was used.It employs four spindles with different geometries to calculate viscosity directly with respect to the given spindle rotation speed.The viscometer does not report shear rate, only rotation speed and spindle type used for testing, which is sufficient to find an optimized dispersant amount.Thus the viscosity plot is represented as a function of L2 spindle rotation speed instead of shear rate, as shown in Fig. 6a.The slurry demonstrated shear-thinning behavior, and viscosity became stable when the rotation speed was greater than 100 rpm.For subsequent viscosity measurements with different dispersant contents, the rotation speed was maintained at 200 rpm.Figure 6b demonstrates the viscosity change as the PEI content increases from 0 to 3.2 wt%.After the fourth data point at 0.47 wt%, the viscosity becomes stable regardless of increasing PEI content, indicating that the particles have achieved full dispersion.Therefore, 0.47 wt% is chosen as the dispersant content for the slurry.

Effect of plasticizer
To flatten the warped tapes, it is necessary to consider adding plasticizer to aid the additional shaping process required.
The effect of plasticizer on the tape warping issue is investigated by adding 100 wt% (with respect to binder) ethylene glycol to the 5 wt% (with respect to powder) binder slurry.The viscosity of the resulting slurry decreases significantly, and the tape becomes very soft and flexible.However, after outgassing, the tape still warps, as shown in Fig. 7.This result implies that plasticizer only temporarily counters the effect of the binder, but after it volatilizes, the effect of uneven drying still results in warpage.Nevertheless, this observation opens a possible solution.The green tape, as described in 3.3.1, is softened by spreading additional plasticizer to 100 wt% of binder on the surface of the cast tape and letting it sit for 30 min.Then the tape is sandwiched between two steel meshes with weight applied at the top and placed in a furnace at 120 °C for 2 h to evaporate the plasticizer.This steel mesh is intended to mechanically constrain the green tape while allowing plasticizer evaporation.This additional step is intended to solidify the tape in a relatively flat shape, and subsequent outgassing does not significantly change the tape flatness, as illustrated in Fig. 8.However, this method does not solve the warping issue from the root, but rather serves to modify the warped tape, and the treated tape still has small curved parts.

Effect of binder
To reduce the impact of film formation on thick tape, a new approach was taken by reducing the original 5 wt% binder content to 2.5 wt%.The result showed improvement in tape flatness, as the compressive stress within the top film and tape edges decreased with a lower binder content.While 5 wt% binder yielded a relatively flat tape, the remaining curling can cause outgassed tape to crumble where it has little to no contact with the support plate, as shown in Fig. 9. On the other hand, 2.5 wt% binder resulted in almost fully flat tape, although the improvement in flatness came at the cost of weaker tape, as the tape was found to crack when bent.However, for the application of flat SOC substrates, such strong flexibility is not required, and the binder is still sufficient to maintain tape integrity while cutting the tape to coupon sizes.
Fig. 13 Cr22 substrates sintered at temperatures from 1100 °C to 1200 °C for 6 h before (left) and after (right) oxidization at 900 °C for 24 h

Effect of casting surface
Besides reducing binder content, different casting surfaces were explored for the 2.5 wt% binder slurry.Silicone rubber, PTFE and poly propylene PP were tested in addition to boron glass.The results are shown in Fig. 10.The tapes dried on PP and PTFE have better flatness than those on glass and silicone.These results are likely due to the higher surface energy of the interface between the plastics and the binder compared to the silicone and glass.The high surface energy of the interface promotes the release of the tape edges from the casting surface, and thereby helps the tape centre to dry from below by allowing a pathway for the volatile phases to escape.

Effect of solid loading
The last factor investigated is the solid loading in suspension.Using 2.5 wt% binder and 10 wt% plasticizer to binder in a slurry with 33 vol% solid as a control, two additional slurry formulations, with steel powder loading decreased to half and increased by a factor of 1.5 while the binder solution remains the same were tested.The new corresponding solid loading and binder weight percents are 20 vol%, 43 vol%, 5 wt% and 1.7 wt%, respectively.Finer Cr16 powder with higher surface area is used for the test.Figure 11 shows the resulting tape with modified solid loading.Decreased solid loading leads to faster particle settling, and formation of a binder film is observed.The resulting tape has significant warping, which suggests that particle settling can also contribute to tape warping by leaving a thin translucent layer of only solvent-binder solution at the top of the slurry, resulting in a strong binder film at the top of the cast tape where there are fewer particles.This film then pulls the edges of the tape toward the centre as it shrinks during drying, to a greater extent than the shrinkage of the portion of slurry where the particle concentration remains high.Increased solid loading leads to a very flat tape, which can be a result of a combination of two factors: stronger interaction between particles raises fluid viscosity and hinders settling, and reduced volume shrinkage leads to less compression stress along the top plane of the tape that pulls the outside edges toward the centre of the tape.In addition, the increased solid loading compensates the viscosity lost from the reduced binder content, which is crucial for sterically stabilizing high density particles like steel, while previous work has preferred to raise binder content to raise viscosity [13,21].

Pore former addition
When adding pore former, a portion of the steel particles are simply replaced with pore former at the desired volume fraction, with no additional modification in the other organic additives.Three pore former fractions from 35 vol% to 55 vol% with 10 vol% increments were tested.The PMMA pore former results in a tape with similar properties as the full steel one regardless of the fraction, but the starch pore former altered the slurry properties, as starch absorbs water and increases the tape viscosity, making the slurry hard to manage.This effect became more pronounced as the starch volume fraction increased to 55 vol%.When tapes are made with 55 vol% of the steel particles replaced by 32-45 μm starch particles, the tape curls and shows cracks.This is possibly due to high viscosity making it difficult for the particles to disperse uniformly.The slurry composition is not optimized for a high starch volume fraction tape due to the sintering problem introduced by starch, as will be discussed in Sect.3.6.

Summary of slurry formation
In conclusion, the final optimized slurry for the full steel powder tape has the following composition: 43 vol% solid fraction, 1.65 wt% binder to steel powder, and 0.47 wt% dispersant to steel powder.The resulting tape is still on the flexible side despite the decreased binder and lack of plasticizer, but the excellent flatness allows direct cutting into intended sizes without breaking, which is the only shaping process needed for SOC substrate application.In addition, the slurry was placed in a test tube and allowed to settle for a few hours.After a 4-h slurry sedimentation test, there was only less than 1 mm clean fluid left by particle settling for a 46 mm deep sample of slurry (Fig. 12).This time is sufficient for slurry to completely dry out, and the 1 mm clean fluid is negligible compared to the whole 46 mm slurry column.

Effect of steel particle size on outgassed substrate handling
Three different powder sizes were tested to evaluate the effect of powder size on pre-sintered tape strength.Besides the Cr16 and Cr22 powders described above, an additional 1C44Mo20 powder with 90% volume particles smaller than 32 μm (Cr22-32) was used for comparison purposes.The Cr16 tape has higher mechanical strength compared to both the Cr22 and Cr22-32 tapes and can be picked up by hand.The Cr22-32 tape is the weakest and can crumble during outgassing without any handling.The Cr22 tape is also fragile and needs to be carefully handled when transferring to the sintering furnace.A weak tape during the transportation step between intermediate temperature furnace and sintering furnace increases the chance for the tape to mechanically fail.

Sintering strategy for breakaway oxidation prevention
Figure 13 shows the sintered microstructure evolution of Cr22 steel tape with 35 vol% stainless steel replaced by 45-75 μm starch pore former.The same composition is sintered at three temperatures, from 1100 to 1200 °C, with a 6-h dwell time.The porosity was higher, 43.7%, when sintered at 1100 °C for 6 h, compared to 28.8% when sintered at 1200 °C  for 6 h.The microstructure sintered at 1200 °C for 6 h achieved near-full density in the portions of the microstructure that began without pore former.Only this substrate survived the oxidation test without breakaway oxidation.For the Cr16 powder, a similar porosity can be achieved at 1200 °C with a 4-h dwell time, as shwon in Fig. 14.

Effect of organic residue in tape on substrate corrosion resistance and a solution to reverse carbide phase formation
When optimizing the sintering conditions of Cr22 powder tapes with different pore former fractions, a 2-h sintering time and a sintering temperature of 1280 °C were used.However, this led to the observation of Cr segregation and oversintering in the substrates.Cr segregation refers to the formation of regions with higher-than-normal Cr concentration in the steel, which can be seen in backscattered electron images as areas with lower brightness due to the lower atomic mass of Cr compared to Fe.This result was confirmed by EDS mapping, as shown in Fig. 15.
When sintering substrates with high a content of starch pore former, the substrate trends to be over sintered.The substrates have undergone excessive shrinkage, resulting in a complete loss of porosity.There is an obvious relationship between the extent of over-sintering and the starch pore former fraction, as shown in Fig. 16.On the other hand, for PMMA tapes, such over-sintering was not observed, only Cr segregation.
Referring again to the continuous-heating TGA results of PMMA and starch, the starch may have had a significant residue left at the point when the gas was switched from air to hydrogen during the tape outgassing process.The residual carbon can precipitate from austenite when the temperature is lowered, forming chromium carbide, and depleting the neighboring region of chromium at approximately 950 °C.This phenomenon is known as sensitization [1].PMMA has nearly complete burnout at the temperature where the gas is switched from oxidizing to reducing environment, but 5 wt% binder can still be sufficient for sensitization at the 1280 °C-2 h sintering condition.Cr22 substrates have Cr segregation with poor oxidation resistance, where the oxide layer fills most of the porosity and contains significant iron compared to less oxidation-resistant 430 steel, as shown in Fig. 17.When a longer sintering time of 6 h (at 1200 °C) is used, and hydrogen flows at 500 SCCM, no sensitization occurs.However, it does occur if the hydrogen flow is decreased to 100 SCCM.This result indicates that high purity hydrogen flow can help to remove extra carbon, possibly by the Sabatier reaction.Further experiments found that for Cr16 powder tape, 2.5 wt% binder with PMMA pore former can be sintered without sensitization at 1200 °C for 1 h at 100 SCCM hydrogen gas flow, which is an additional benefit of decreasing the binder content, in addition to preventing warpage.
To further understand the role of hydrogen in sensitization, both Cr22 and Cr16 substrates with Cr-rich phase presence were tested by dwelling at 1200 °C for 12 h with a hydrogen flow rate of 500 SCCM to remove any residual carbon remaining from the slurry organic content.Both substrates have porosity introduced by 35 vol% starch.The substrate microstructure before and after the carbon-removal test is shown in Fig. 18.In both cases, the high brightness carbide disappeared after hydrogen treatment.EDS analysis confirmed that the Cr content returned to the nominal value of 22 wt% and 16 wt% after treatment.Figure 19 shows elemental mapping of a Cr22 substrate after the treatment, confirming the elimination of chromium carbide.However, niobium carbide is still present, possibly due to stronger reactivity with of the niobium with carbon than of the hydrogen with carbon.

Conclusion
This work focuses on overcoming the challenges of a low-volatility aqueous thick tape casting process to obtain tape with high flatness, uniformity, and resistance to oxidation.Despite the compromise on tape flexibility with low binder and without plasticizer, the final tape can still be cropped into the desired size, which fits the demand of SOC substrate application.Raising solid loading proves to be an effective strategy to prevent slurry particle sedimentation rather than raising the viscosity with additional binder, which can cause stainless steel sensitization if not effectively removed.By applying extensive sintering with fine particles to preserve only porosity from the pore former particles, a microstructure with uniform porosity is created that can help to prevent breakaway oxidation by minimizing the surface area.The optimized slurry composition is 40 vol% solid loading with − 10 μm 430SS powder, with binder and dispersant loadings of 1.7 wt% and 0.47 wt% of the steel powder weight, respectively.PMMA as the pore former burns out to a sufficient extent at 350 °C without causing sensitization.29 vol% porosity can be obtained by adding 55 vol% PMMA and sintering at 1200 °C for 4 h.Sintering steel tape in a high flow rate of high-purity hydrogen for an extended time can be used as a strategy to remove carbon and to reverse sensitization if the carbon removal is not

Fig. 3
Fig. 3 Secondary electron image of as-received a Cr16 and b Cr22 stainless steel powder

Fig. 6 aFig. 7
Fig. 6 a Shear-thinning rheology behavior of 15 vol% Cr16 powder slurry with 2.5 wt% PVA at different L2 spindle rotation speeds.b Viscosity of the same slurry at 200 RPM while increasing PEI content

Fig. 12
Fig.12 Slurry with composition described in Table 1 in a 4 h-sedimentation-test. Roughly 1 mm clear layer fluid is observed at the top of a 46 mm deep slurry column, indicating good slurry stability

Fig. 14
Fig. 14 Cr16 tape with 55 vol% PMMA sintered at 1200 °C for 4 h has similar porosity to Cr22 tape with 35 vol% starch

Fig. 15 Fig. 16
Fig. 15 EDS elemental mapping of Cr22 steel sintered at 1280 °C for 2 h.The internal formation of Cr-rich phases and Nb-rich phases is also visible from backscattered electron images

Fig. 17
Fig.17SEM images of Cr22 (a) and Cr16 (b) substrates after 900 °C-24 h oxidation tests.In (a), the formation of a Cr-rich phase lowers the Cr concentration in the bulk ferrite phase to 10-11%, exposing the substrate to breakaway oxidation, as evidenced by the large quantities of oxide in the pores, as well as the high iron content in the oxides.In contrast, the Cr16 material in (b) does not experience breakaway oxidation, so the pores contain less oxide, and the oxide contains less Fe compared to the Cr22 material

Fig. 18
Fig.17SEM images of Cr22 (a) and Cr16 (b) substrates after 900 °C-24 h oxidation tests.In (a), the formation of a Cr-rich phase lowers the Cr concentration in the bulk ferrite phase to 10-11%, exposing the substrate to breakaway oxidation, as evidenced by the large quantities of oxide in the pores, as well as the high iron content in the oxides.In contrast, the Cr16 material in (b) does not experience breakaway oxidation, so the pores contain less oxide, and the oxide contains less Fe compared to the Cr22 material

Table 1
in a 4 h-sedimentation-test. Roughly 1 mm clear layer fluid is observed at the top of a 46 mm deep slurry column, indicating good slurry stability