49th International Conference on Environmental Systems ICES-2019-326 7-11 July 2019, Boston, Massachusetts

Mapping the Capabilities and Attributes of Solid Oxide Electrochemical Systems to Human Spaceflight Needs

John Graf, PhD1 National Aeronautics and Space Administration, Johnson Space Center, Houston, TX, 77058

NASA is sponsoring the development of a solid oxide electrochemical oxygen separation and compression system. This system is capable of extracting oxygen from a process stream of spacecraft cabin air, and compressing it to >27,000 kPa without mechanical compressors. Process temperatures are >650C, but the materials in contact with the oxygen do not burn. The system is called eCOG-C, Electrochemical Oxygen Generator and Compressor. A system capable of producing high pressure, high purity oxygen may have application as a method of recharging space suit oxygen tanks during human exploration missions. This paper describes the key performance parameters of a space suit oxygen tank recharge system, and places the key performance parameters of eCOG-c in the context of other methods of space suit oxygen tank recharge. This paper compares the eCOG-c configuration to an earlier electrochemical oxygen generator that has been prototyped and tested. Areas of emphasis for eCOG-C development are listed and described.

Nomenclature ABO = aviator’s breathing oxygen CTE = coefficient of thermal expansion eCOG-C = electrochemical oxygen generator and compressor EMU = extravehicular mobility unit EVA = extra vehicular activity GOX = gaseous oxygen HX = heat exchanger ISS = international space station kPa = kilopascals LOX = liquid oxygen lpm = liters per minute O2 = oxygen PEM = polymer electrolyte membrane PSA = pressure swing adsorption ppm = parts per million SEOS = solid electrolyte oxygen separation

I. Human Exploration – Life Support System – Oxygen Needs HERE are three different needs for oxygen in a life support system capable of supporting Human Exploration T beyond low earth orbit: 1) metabolic oxygen, 2) medical oxygen, and 3) space suit oxygen. Each of these uses has a unique set of requirements and key performance parameters. A well-planned life support system meets the metabolic, medical, and space suit oxygen needs in a coordinated way. Metabolic oxygen requires the greatest quantity of oxygen per year. Metabolic oxygen must be produced and delivered on a nearly continuous basis. Metabolic oxygen can be delivered at ambient pressure. Medical oxygen is needed only on a contingency/emergency basis. The key aspects of medical oxygen are integration with medical ventilator equipment and delivery rates. Peak oxygen delivery rates used in emergency medicine can be substantially higher than metabolic oxygen consumption rates, so sustained use of medical oxygen in a closed cabin environment can cause oxygen levels to rise above fire safety limits. Some medical oxygen system approaches address this issue ______1 Research Engineer, Crew and Thermal Systems Division, 2101 NASA Parkway / Mail Code EC3 by careful integration with ventilator hardware (coordinating oxygen delivery with ventilator cadence can reduce the rate of oxygen vented to the cabin). Additionally, some medical oxygen system designs use an oxygen concentrator, so medical oxygen can be administered continuously without raising the oxygen levels in the cabin. Space suit oxygen systems have three key needs: 1) pressure, 2) purity, and 3) system safety. Many space suit designs for future exploration missions involve two active oxygen tanks rather than using the EMU system architecture, which employs a main tank with a lower pressure, and a contingency tank with a higher pressure. Trade studies indicate that two active oxygen tanks initially filled to 20,000 – 25,000 kPa meet EVA oxygen system needs with favorable system size and system weight characteristics. Recharge tanks must hold oxygen at pressures higher than 20,000 kPa, and supply tanks with higher pressures can recharge more effectively. The design goal for the electrochemical oxygen generator and compressor (eCOG-C) prototype is 28,000 kPa. The purity requirement for oxygen in the EMU system is specified in the Aviator’s Breathing Oxygen (ABO) specification (Ref1). Operationally, the EMU system has used oxygen with much higher purity. The design goal for eCOG-C prototype is >99.99% oxygen. System safety for a space suit oxygen supply system for human exploration missions is critically important for three specific reasons: 1) oxygen is involved, 2) gas management system pressures will likely be higher than either space shuttle or ISS EMU systems, 3) launch constraints make it very likely that oxygen will be stored in the form of water; and mechanical, chemical, or electrochemical processes will be used to add energy to the oxygen in the cabin environment. This paper introduces a new approach to producing high pressure, high purity oxygen; a ceramic oxygen transport membrane. The first section of this paper places the eCOG-C method in the context of other methods of producing high pressure, high purity oxygen. The second section of the paper describes the principal technical development risk for eCOG-C; interconnecting seals. Finally, results of a feasibility assessment for interconnecting seals are shown. .

II. Comparing eCOG-C to Other Methods of Space Suit Oxygen Resupply

Table 1 summarizes some key attributes of seven methods of providing space suit oxygen. This table is not intended to be a comprehensive analysis, and it is not intended to recommend a specific system for human exploration. Instead, the intent of this table is to emphasize that providing space suit oxygen is hard, to stress that every system has issues and complexities, and to note that based on a mapping of system attributes to key performance parameters, there is a rationale for developing eCOG-C. In-flight operations are likely to be the simplest, safest, and most sustainable for a high pressure gas (GOX) based system. The oxygen is pressurized and purity is verified prior to launch. Energy is not added to the oxygen during the mission, so safety aspects are favorable. Gaseous systems are ideally suited to missions with few planned EVAs, because compared to other systems, gaseous systems tend to be large and heavy. If exploration mission plans involve many EVAs, it is likely that gaseous systems will be too heavy to meet launch constraints. Cryogenic liquid oxygen (LOX) systems share many of the purity and pressure attributes as gaseous oxygen systems: purity can be verified prior to launch and pressurization can be achieved without mechanical compressors. LOX systems are substantially smaller than comparable gaseous oxygen systems. If mission durations are short enough to accommodate system boil-off, LOX systems will likely be the best method of space suit oxygen supply. Many exploration mission plans have long mission durations, and system sustainability is an issue. Low pressure electrolysis is the reference method of metabolic oxygen supply for many mission plans. The low pressure electrolysis can be combined with a gas drying system and a mechanical compressor for space suit oxygen needs. This approach is the reference method for many exploration planning studies. Because the technology is relatively mature, the technical risk for developing this approach is relatively low – but the approach involves adding energy to low pressure oxygen. NASA does not have much operational experience adding energy to oxygen in flight. Slow moving compressors that emphasize safety tend to be large and heavy. All mechanical compression systems introduce a risk of contamination, so on-board purity verification will likely be required. This approach is relatively mature, and relatively low-risk; especially if mechanical compression can be done safely, and a reliable method of onboard purity verification can be developed. The electrolysis system can be uncoupled from the high pressure system if a pressure swing adsorption (PSA) oxygen concentrator is added to the system configuration. Product purity testing of a PSA system (described in reference 2) demonstrate purity concerns, especially with variable argon, and fast production rates.

2

International Conference on Environmental Systems

High Pressure O2 Gas Tanks Pressure: No Issues: Pressurization is done prior to launch

Purity: No Issues: Purity can be validated prior to launch, risk of contamination after launch is low

Safety: Favorable: No energy is added to the system during the mission

Summary: Great for missions with few planned EVAs. Concerns about size and weight for exploration

Cryogenic Liquid O2 (LOX)

Pressure: Pressurization occurs as heat is added to the system

Purity: No Issues: Purity can be validated prior to launch, risk of contamination after launch is low Safety: Lots of stored energy. Apollo 13 mishap involved LOX Summary: Best suited for short duration missions with many planned EVAs

Low Pressure Water Electrolysis + Drier + Mechanical Compressor Pressure: Demonstrated with technology demonstrator systems Purity: In-flight purity verification will likely be required; contamination risk is credible (compressor)

Safety: Adding energy to O2 during mission – compressors that keep O2 cool can be large

Summary: May trade well for exploration missions, especially if reliable purity verification is developed

Low Pressure Water Electrolysis + PSA O Concentrator + Mechanical Compressor 2 Pressure: Demonstrated with technology demonstrator systems Purity: In-flight purity verification will likely be required; severe concerns about PSA product purity Safety: Adding energy to O2 during mission – compressors that keep O2 cool can be large Summary: PSA product purity concerns need to be addressed before this can be evaluated High Pressure Water Electrolysis + O Drier 2 Pressure: Demonstrated with technology demonstrator systems (thick end caps) Purity: Must remove water, verification may be easier if only water needs to be measured Safety: Fundamental safety issue: stack has hydrogen, high pressure O2, and ignition source Summary: Safety issues cannot be designed out of the system

PEM Electrolyte Electrochemical O Compressor 2 Pressure: 14,000 kPa demonstrated with lab prototype Purity: Must remove water, verification may be easier if only water needs to be measured

Safety: Fundamental, but less severe: solid fuel (not H2), high pressure O2, and ignition source Summary: Kinetics needs to be demonstrated before this can be evaluated

Electrochemical Oxygen Generator & Compressor (eCOG-C) Pressure: 14,000 kPa demonstrated with lab prototype Purity: No in-flight purity verification needed

Safety: Adding energy to O2 during mission - high temp hazards, but fire triangle ok (no fuel)

Summary: Good potential for human exploration: solid state, high purity, no fuel near O2

Table 1: Pressure, Purity, and Safety Attributes for Seven Different Methods of Space Suit O2 Supply

High pressure electrolysis involves the fewest number of systems. Some system designs enable production of high pressure and low pressure oxygen, so metabolic oxygen needs can be served, and purity verification may be simpler if water vapor is the only credible contaminant. Oxygen system safety assessments of high pressure water electrolysis systems emphasize that high pressure electrolysis involves highly mobile and highly flammable hydrogen gas, high pressure oxygen, and a credible ignition source in close proximity to each other. This fire-triangle safety issue is intrinsic to high pressure electrolysis systems. There are variant electrochemical designs that try to balance the system so hydrogen stays within the membrane. These systems are referred to as PEM electrolyte electrochemical oxygen compressors. When these systems are balanced and working properly, there is no free gaseous hydrogen in the cell stack. Removing free hydrogen from the

3

International Conference on Environmental Systems

gas stream greatly improves the safety characteristics of the system, and operational temperatures are nearly ambient, but the cell stack contains solid phase fuel. There is a credible possibility for an oxygen fire. Research systems have been developed, and high pressure operation has been demonstrated, but these research systems did not have kinetics suitable for an operational system. When the technology has developed to a point where operationally relevant oxygen production rates have been demonstrated, this system can be evaluated for use in a human exploration system. eCOG-C electrochemically generates high pressure oxygen, using a solid state process. If crystalline interconnecting seals are developed, >99.99 purity should be inherent to the system design, and there would be no need for in-flight purity verification. Pressure integrity verifies product purity. As with every system that converts water to high pressure oxygen, energy is added to the oxygen in the cabin environment. A unique hazard to eCOG-C is the high operating temperature. Flammability hazards are addressed through material selection: the materials in contact with high pressure oxygen do not burn. eCOG-C would be better suited to supporting human exploration missions if operational temperatures were lower, but in the context of other systems, a solid state device, with the potential of achieving oxygen purity of >99.99% with such reliability that in-flight purity verification is not required, made from materials that do not burn, has promise for meeting human exploration needs.

III. A Comparison between SEOS and eCOG-C

As a general technology area, solid electrolyte oxygen separation relates to any electrochemically active oxide material that operates in a high temperature environment, has a dc potential placed across the material, and can electrochemically pump oxygen from one side of the oxide layer to the other. Generally, some solid electrolyte oxygen separation systems are planar and some systems are tubular. In the context of this paper, Solid Electrolyte Oxygen Separation (SEOS) refers to a specific configuration of a planar wafer consisting of several different layers that are co-sintered into a single element with matched coefficients of thermal expansion (CTE) so it can retain structural integrity across a wide range of temperatures. A stack of SEOS wafers is shown in figure 1. The size of each individual wafer is approximately 7 X 14 cm. The internal configuration of a SEOS cell stack is shown in figure 2. Figure 2 is not to scale, it is exaggerated to make each individual feature easier to identify. Structural integrity is provided by a strong, thick layer of dense material at the base. This dense material also seals the edges of the wafer. The top layer of the wafer is porous, and allows air to diffuse to the electrolyte layer. The electrolyte layer is relatively thin. This reduces the voltage needed to move oxygen ions from the air side to the oxygen side. The electrolyte is supported on both sides by porous layers, enabling the wafer to sustain an internal oxygen pressure of >1500 kPa without sustaining any damage or degradation (ref 4).

Figure 1 Photograph of a cell stack of SEOS wafers (left) and a quartered-section of a wafer stack (right) 4

International Conference on Environmental Systems

Individual wafers are held together using a glass interconnecting seal. The seal has an annular shape. It connects each of the oxygen layers into a single oxygen flow system, connecting a series of circular thru-holes into the center of each wafer.

Glass seal

Process air channel

Porous layer with O2

Electrolyte

Porous layer with air

Dense layer

Figure 2, cross section top view (above) and cross section side view (below) of the Solid Electrolyte Oxygen Separator (SEOS) cell stack configuration. Process air surrounds the wafers, and oxygen is pumped internally.

The SEOS cell stack was integrated into a thermal management system shown in figure 3, and tested by NASA. NASA tests confirmed measurements made by the manufacturer for power consumption, delivery rate, and delivery pressure. NASA tests had one unexpected result: manufacturer measurements of product purity were >99.999% O2, and the manufacturer did not detect any water in the oxygen product. NASA tests identified trace amounts of water in the oxygen product: 2-50 ppm depending on operating conditions. Further investigations identified the source of the water contamination: at process temperatures, the glass interconnecting seal is slightly permeable to water (ref 5). The manufacturer’s tests of product purity were conducted using a single wafer – with no interconnecting seal. Amounts of water contamination can be increased by: 1) increasing the humidity of the process air steam, which increases the permeability driving force, 2) increasing the operating temperature, which increases the permeability of the glass, or 3) decreasing the oxygen production rate, which effectively reduces the ratio of oxygen molecules to water molecules. The SEOS system has demonstrated the capacity of producing >1500 kPa oxygen with >99.99% product purity, but space suit recharge pressures will never be directly achieved with the SEOS wafer configuration. Internal oxygen pressure forces will cause the wafer to fail in tension.

5

International Conference on Environmental Systems

The eCOG-C project intends to modify the configuration of the SEOS wafer, put the wafer in compression, and produce oxygen with pressures >27,000 kPa. Figure 4 describes the configuration of the eCOG-C wafer. The SEOS wafer has ambient pressure process air flowing around the outside of the wafer, oxygen accumulating internally,

Air

inlet cell stack fan s

Air Outlet heater rr Figure 3. SEOS cross sectional schematic (left), photograph of prototype (right)

Crystalline seal

Process air channel

Process air direction

O in pressure vessel 2

Electrolyte

Porous layer with air

Dense layer

Pressure vessel

Figure 4: cross section view from top (above) and cross section view from side (below) of the eCOG-C cell stack configuration. Process air surrounds flows internally through the wafers, and oxygen is pumped externally.

6

International Conference on Environmental Systems

and a single port routing oxygen from the wafer to a delivery port. The eCOG-C wafer has two ports for routing ambient pressure process air through a set of internal flow channels. The channels are narrow, to withstand the large compressive load. The electrolyte layer pumps oxygen ions from the inside of the wafer to the outside. Because the eCOG-C cell stack is placed inside a pressure vessel, with a single oxygen port that is pressure regulated using a backpressure regulator external to the pressure vessel. This configuration allows the oxygen to be electrochemically pressurized to design pressure. The ambient pressure process air provides a source of fresh oxygen, and a mechanism to remove heat from the system. The eCOG-C requires crystalline interconnecting seals (ref 6). Glass seals are acceptable for the SEOS application because the pressures and mechanical forces are relatively low, but glass seals are too soft to maintain structural integrity across a pressure gradient greater than 27,000 kPa.

IV. Areas of Emphasis for eCOG-C Development

The next phase of eCOG-C development will emphasize four aspects of the system: 1) seals, 2) wafer manufacture, 3) thermal management, 4) pressure vessel design. ECOG-C interconnecting seals must be fully devitrified in order to withstand pressure forces exerted by the oxygen. Additionally, fully devitrified seals are capable of preventing water vapor permeation into the oxygen product stream. These newly developed devitrified seals will perform the best if the CTE of the seals exactly match the CTE of the wafers. CTE mismatches put the seal/wafer interface under mechanical stress when operating temperatures and storage temperatures are different from manufacturing temperatures. The SEOS glass seals have a slight CTE mismatch, but because the SEOS system operates at lower pressures, and because the glass is relatively soft and malleable, the CTE mismatch in the SEOS system is acceptable. Devitrified interconnecting seals with the attributes needed for eCOG-C have already been developed for solid oxide fuel cells. The solid oxide fuel cell seals are formed by starting with a precise mixture of a glassy powder, and subjecting the powdery mixture to a precise set of time/temperature/pressure conditions. Ceramic wafers with two ports and complex internal flow channels have been manufactured for heat exchanger applications – illustrated in figure 5. The first round of prototype wafers will be evaluated for geometric stability and compressive strength. Many elements contribute to the thermal management of the system: 1) process air flow rate 2) internal heat exchanger performance 3) flow distribution 4) current density 5) properties of insulation inside the pressure vessel 6) geometry of the cell stack and pressure vessel 7) properties of the insulation external to the pressure vessel Many of these elements can be developed by making small modifications to mature designs of existing systems that have been developed for different applications. Some of these elements need new designs and new tests to establish performance in unique operating conditions. The insulation inside the pressure vessel will be operating in a moderate temperature, high pressure operating environment. At the time of writing this manuscript, the eCOG-C development team has been unable to find published thermal properties for insulating materials that operate in a high pressure oxygen environment. Fundamental material characterization tests will be needed to establish insulation material properties in this unique, high pressure oxygen environment. Process temperature for the eCOG-C cell stack will nominally be 650C. High nickel monel alloys that will likely be used for the pressure vessel (for reasons of oxygen compatibility) maintain their structural properties in the 400- 450C temperature range. eCOG-C will be have structural elements in close proximity to electrochemical elements operating in the 600-700C temperature range. The eCOG-C development team will analyze and test structural systems across a broad range of temperatures to account for nominal and off-nominal operating conditions.

7

International Conference on Environmental Systems

Figure 5. Example of a ceramic wafer with internal flow channels.

V Feasibility Assessment of a Devitrified Sealant

Glass seals have been successfully used in the ceramic-ceramic joints that join each individual wafer in a cell stack with a maximum operating pressure of 300 psig, but this application has more demanding structural requirements. Glass, by its nature, is more forgiving of slight mismatches in coefficients of thermal expansion, but structurally weaker with respect to pressure loads. A feasibility study was conducted to determine if the constituents of the glass seal could be bonded under conditions of temperature and pressure that would enable the constituents to devitrify, and form a crystalline structure that bonds to the wafer.

Samples were prepared using a configuration described in Figure 6. Sealant materials were placed in direct contact with test coupons made from materials matching the top surface and bottom surface of the oxygen transport membrane wafer. The materials were placed under a relatively small, but constant compressive load (approximately 8.5 kPa. This configuration was heated and then held under different time/temperature conditions to determine the mechanisms of devitrification, and their rates at different temperatures.

Figure 6. Configuration for Devitrified Sealant Sample Preparation

8

International Conference on Environmental Systems

Differential scanning calorimetry was performed on the samples. The calorimetry test results (shown in Figure 7) clearly show the glass transition temperature at 743 degrees C, and a glass re-crystallization exotherm (i.e. devitrification) at 950 degrees C. The calorimetry tests offer a guide for determining the time/temperature/pressure conditions to enable the glassy material to flow and fill seal gaps, and then to crystallize and fully devitrify. The calorimetry tests suggest a four part sealant process: 1) fast heating to tempeatures that enable the glassy matrix to soften and fill small gaps in the wafer surface, 2) slow heating to a temperature above the main glass transition temperature, to enable the sealant to flow and fill small scale irregularities 3) a time/temperature/pressure hold at devitrification temperatures to allow for crystallization, 4) slow cooling maintain the bond between the sealant and the wafer. Figure 7 offers guidance about the temperature conditions needed for steps two and three of the process – for one leading sealant formulation.

Figure 7. Differential Scanning of Candidate Sealant Material

In addition to differential scanning calorimetry, temperature vs. viscosity trends were measured for a group of candidate sealant formulations. The results are shown in figure 8. Viscosity measurements are good indicators of a material’s ability to flow and fill small surface irregularities along the sealing surface (ref 7). Based on previous sealant development efforts, a strong seal can be achieved if viscosity is less than 106 Pa-s if special processing techniques are used. If the viscosity is less than 105 Pa-s, strong seals can be formed with no special processing steps.

Figure 8. Viscosity vs Temperature Trends for Candidate Sealants 9

International Conference on Environmental Systems

In addition to flowing, and filling, and forming a continuous bond, a seal should have a coefficient of thermal expansion that is similar to the coefficient of thermal expansion of the wafer. The wafer has a coefficient of thermal expansion of 12 ppm/C. The allowable mismatch between wafer CTE and sealant CTE depends on the maximum temperature, the seal thickness, maximum gas delivery pressure, and assumptions about safety factors and non- linearities. At the current stage of development, the target window for sealant CTE and flow characteristics are graphically shown in figure 9.

Figure 9. CTE and Viscosity Trends for Candidate Sealant Materials

One of the candidate seals has been manufactured, and subjected to preliminary testing. Coupons were manually fractured, and the fracture surface was visually evaluated by scanning electron microscopy. Examples of the images are shown in figure 10. Additionally elemental maps of the sealant were made using transmission electron microscopy. The elemental maps help determine if new (and potentially disruptive) crystal structure are formed. The analyses are preliminary, and the results are limited to a small number of fracture surfaces, but results to date are favorable: the bond surface is continuous, and there are no indications of unexpected crystal structures. It should be stressed that a substantial amount of additional testing is required before the sealing integrity can be fully verified.

Figure 10. Scanning Electron Microscopic images of the sealant/wafer boundary. 10

International Conference on Environmental Systems

VI Conclusions

The eCOG-C is a solid state, electrochemical method of producing space suit grade, high pressure, high purity oxygen from a process stream of ambient pressure, ambient temperature process air. The main rationale for pursuing ceramic oxygen transport membranes as a technology for supplying spacesuit grade oxygen is the solid state nature of the technology, the intrinsic purity attributes, and the oxygen compatibility attributes. The system involves high operating temperatures, but the materials in contact with the oxygen do not burn. The primary technology development risk for eCOG-C is interconnecting seals. To withstand structural loads caused by the pressure stress – at elevated temperatures – the seal needs to make a continuous bond with the wafer surface, and it needs to have a coefficient of thermal expansion that is almost identical to the surfaces of the wafer. A preliminary assessment indicates that candidate seal designs with these attributes are available. Future work will focus on developing, and fully characterizing these interconnecting seals.

References

1SAE Aerospace – A-10 Aircraft Oxygen Equipment Committee., “Aviator’s Breathing Oxygen Purity Standard,” SAE-AS-8010C, October 2012. 2Graf, John and Neumeyer, Derek. “A Cabin Air Separator for EVA Oxygen (CASEO)”. 41st International Conference on Environmental Systems, Portland OR, July 2011. 3Ankilander, William and Molter, Trent. “Oxygen Generator Cell Design for Future Submarines.” 25th International Conference on Environmental Systems, Monterey CA, July 1996. 4Hutchings, K.N., Bai, J., Cutler, R.A, Wilson, M.A. and Taylor, D.M. “Electrochemical Oxygen Separation and Compression Using Planar, Cosintered Ceramics,” Solid State Ionics, p. 442-450, 2008. 5Graf, John, Taylor, Dale, and Martinez, James. “Determining the Source of Water Vapor in a Cerium Oxide Electrochemical Oxygen Separator to Achieve Aviator Grade Oxygen,” Microscopy and Microanalysis 2014, Microscopy Society of America. 6Sohal, M.S., O’Brien, J.E., Stoots, V.I., Sharma, B., Yildiz, B., and Virkar, A., “Degradation Issues in Solid Oxide Cells During High Temperature Electrolysis,” J. Fuel Cell Sci. Technol., vol. 9, 2012. 7Chen, K. and Jiang, S.P., “Review – Materials Degradation of Solid Oxide Electrolysis Cells,” J. Electrochem. Soc., pp. F3070-F3083, 2016.

11

International Conference on Environmental Systems