Study on Water Electrolysis for Oxygen Production ...

47th International Conference on Environmental Systems ICES-2017-215 16-20 July 2017, Charleston, South Carolina

Study on Water Electrolysis for Oxygen Production

- Reduction of Water Circulation and Gas-Liquid

Separator - Masato Sakurai,1 Asuka Shima, Takuma Terao

Japan Aerospace Exploration Agency, Chofu, Tokyo, 182-8522, Japan

The Japan Aerospace Exploration Agency (JAXA) has been studying and developing

Environmental Control and Life Support System (ECLSS) technologies, particularly an

air regeneration system to support long-duration manned space missions beyond Earth

orbit. One of our goals is to develop a lightweight, compact and energy efficient air

regeneration system. The air regeneration system is comprised of a water electrolysis

subsystem for oxygen generation, a CO2 removal subsystem, and a CO2 reduction

subsystem. Previous research had reported on a water electrolyzer unit without water

circulation and an independent gas/liquid separator, and made the subsystems more

simple and lightweight. This report on the oxygen generation subsystem is updated and

scaled up from the previous report. JAXA has developed a new electrolyzer unit level of

100A (3-cell stacks). It also has heatsinks for electrolysis at high current density. The

results of this report showed that the characteristics of the new unit are almost the same

as those of previous systems, and that heatsinks are also effective in the new unit system.

Nomenclature

CO2 = Carbon dioxide

CH4 = Methane

ECLSS = Environmental Control and Life Support System

H2 = Hydrogen

H2O = Water

ISS = International Space Station

JAXA = Japan Aerospace Exploration Agency

LEO = Low Earth Orbit

MEA = Membrane Electrode Assemblies

SPE = Solid Polymer Electrolyte

O2 = Oxygen

I. Introduction

There are growing expectations for manned space activities on the Moon and Mars. Such activities require a life

support system with a long material recycling rate and operation period. Therefore, it is necessary to develop a

compact, lightweight, power-saving system. As has been adopted in the ISS, in the current technology, the air

regeneration system1-3 shown in Fig. 1 is a practical improvement to the system. This could be achieved by a

recirculating system. A water electrolysis system produces O2 and H2 from water. The crew members consume

O2 and exhale CO2, which is concentrated in an absorbent by a CO2 removal system using a pressure swing

adsorption process. The concentrated CO2 is then reduced by a CO2 reduction system using the Sabatier reaction

(CO2 + 4H2 → CH4 + 2H2O), producing H2O and CH4. The water is then electrolyzed and consumed by the

crew again. This proposed air regeneration system can thus significantly reduce the volume of water supply

needed and its only waste material is CH4.

1 Senior Researcher, Japan Aerospace Exploration Agency (JAXA), 7-44-1 Jindaiji-Higashi-machi, Chofu,

Tokyo, 182-8522, Japan

International Conference on Environmental Systems

2

Figure 1. Concept of air revitalization system using CO2 reduction

The ultimate goal of our group is to realize a compact, lightweight, low energy, highly stable air

regeneration system for manned space activities beyond the ISS orbit. The O2 generation subsystem requires

high power and mass, as shown in the actual example of the International Space Station (ISS). The supply of

crew O2 on the ISS is also produced mainly by electrolyzing water.4, 5 Producing O2 by water electrolysis entails

much lower cost than supplying O2 from the ground. Thus, water electrolysis for O2 generation has been an

important topic of study.6, 7 It is desirable to further reduce the supply of water required for O2 generation to

enable longer periods of human habitation in LEO and missions beyond it. The theory of the air regeneration

system has been well studied, but a system that operates well enough to be of practical use has yet to be

developed. Although realizing a closed-cycle air regeneration system for space applications at an early stage is

desirable, many technical problems must first be resolved. One of the main problems is the limited amount of

available energy in space. The maximum power-generating capacity of the ISS is about 120 kW, and the

available power on manned missions beyond LEO will typically be more constrained. A lot of electrical power

is needed to generate sufficient O2 for ten crew members by water electrolysis, most of which is required for

controlling systems and consumed by various energy losses. Another problem is that gas/liquid separation is

required in many parts of the air regeneration system; but it is difficult to design a system that operates well in

microgravity. This problem is particularly acute in the water electrolysis system, because the water required for

electrolysis and the generated gas are constantly in contact, and incomplete gas/liquid separation not only has

adverse effects on downstream subsystems but also leads to waste gas having a high H2O content.

The anode water feed method involves feeding water from the anode side. This system needs water/gas

separation on the anode side and creates the problem of generating water vapor with the cathode side as well.

Therefore, it will be necessary to extract the gases by drying the oxygen in order to make it breathable for

humans in space. The cathode water feed method (Fig. 2) also enables a simple system, but produces relatively

dry O2. In this method, water migrates to the anode side of the cell from the cathode side through the membrane.

As a result, although the generated hydrogen contains much water vapor, it is possible to obtain dry O2. Cathode

feed conversely has a lower rate of capability, thereby requiring a larger active area or number of cells.

To reduce the mass, size and power of the O2 generation system (Fig. 3), we attempted to build a system that

does not need water circulation or a gas/liquid separation device. In the previous system (Fig. 2), water

circulation was necessary as water was supplied to the electrolysis cell in order to eliminate the generated H2

and cool the cell. However, a previous study reported problems caused by water circulation.5 When the water

cycle is no longer needed, these problems could possibly be solved. Moreover, the size and mass of the system

could be significantly reduced. We created a simple system by developing an electrolysis unit that has a

gas/liquid separation function. Then by reducing the drops in pressure and maintaining thermal efficiency on

each part, the system is expected to save more power. In terms of life-span or the replacement of malfunctioning

units, the simple system may offer a favorable answer for future possible problems. Simplifying the subsystem


3

could save energy for other devices and improve both heat efficiency and thermal control response. It could also

lead to higher subsystem stability.

We had certified the feasibility of the system shown in Fig. 3 by using a single cell stack electrolyzer unit.8

From the result, we had no concerns about the basic characteristics of the unit, polarization curves, gas purity,

cross leaks, etc., in comparison with related data that included water circulation. We are concerned about the

thermal characteristics in case of no water circulation. The unit had not broken down, but we noted that the

center temperature of the unit became higher in comparison with using water circulation. As our next step, in

this paper, we tried to make a 3-cell stack electrolyzer unit with a gas/liquid separation function as scaled up.

Then we evaluated its basic characteristics. Furthermore, we mounted heatsinks on the unit for cases of

electrolysis at high current density, and confirmed its thermal characteristics.

This a preliminary study on a new type of water electrolysis. The level of technology readiness is still low,

as the water supply system has not matured and long-time operation for lifetime checks has not been carried out.

This water electrolysis cell works in different gravity orientations.

II. Objective

In this study in consideration of past research,8 we updated the single cell stack, and then tested and

evaluated the scaled up the 3-cell stack electrolysis unit. And according to the results of previous research, when

using the PEM developed by JAXA,9 power loss is smaller than that of the cells using Nafion; therefore, we had

not considered active cooling. However, as the system proposed in this study has no water circulation, we must

consider operation at high electrolytic current density. This is one option when aiming at further miniaturization

of the system in the future. We thus adopted an air cooling type radiator and examined the tendencies of its

thermal properties. The purpose of this research is summarized below.

1. To make potential issues with eliminating water circulation in a multiple cell stack evident. Focus on

evaluating the uniformity of water supply to each cell in electrolysis.

2. To confirm that gas/liquid separation function is adequate in a multiple stack unit. Each stack layer is

connected by a fluid line, so it can inhibit gas/liquid separation by water and gas movement.

3. To confirm the tendencies of thermal characteristics shown by the air cooling radiator. We evaluate

changes in temperature of the electrolysis unit due to the heat dissipation of natural convection.

Figure 3. Improved system using the electrolysis

unit with a gas/liquid separation function and

requiring no water circulation. Figure 2. Previous system using water circulation.


4

III. Experiment

Electrolytic cell unit without water circulation and with a gas/liquid separator (3-cell stack, 100A class)

The electrolysis unit used in this study has the PTFE hydrophobic membrane inside and an internal

gas/liquid separation function. The effective electrolysis area of the unit is 192 cm2 (Table 1), which is three

times that of the past electrolysis cell unit.8 And for the case of stacks with additional cells in the future, we

must understand the problems of parallelizing the fluid lines without water circulation. We thus designed the

unit to have a 3-cell stack structure as the first step. In order to save power in the O2 production system, we also

adopted air cooling type radiator cooling using heat pipes and heatsinks to enable high current density operation

and energy-saving cooling (Fig. 4). In considering heat transfer under microgravity, we plan to replace the heat

pipe with a loop heat pipe, etc. in the future. Then current density was varied from up to 0.4A/cm2, with current

voltage, the dew points of generated gas, and H2 concentration in O2 being measured as basic characteristics.

Moreover, the surface temperature of the electrolysis unit, internal temperature, and cooling fin temperature

were measured to provide data on thermal characteristics.

Table 1 Electrolysis unit specifications

Electrode area: 192 cm2 (64 cm2 × 3

cells)

Polymer electrolyte membrane: Nafion 212

Mass of electrolysis unit: 24.04 kg

Feed method: Cathode feed

Figure 4. Electrolysis unit requiring no water circulation.

Result and discussions

Figure 5 shows the current-voltage characteristics of the 3-cell stack with or without water circulation and

those of a single cell stack or 3-cell stack without water circulation. The result of the 3-cell stack is the average

value. The result of no-water circulation (Fig.5 (a)) and using water circulation (Fig.5 (b)) has a small difference

of about 0.02 V for electrolytic current densities up to 0.3 A/cm2. In the result of the single stack (Fig.5 (c)) and

3-cell stack (Fig. 5 (a)) with no-water circulation, the difference in voltage becomes slightly larger as the current

density became higher. The temperature control of units or supply water was difficult except in the case of the

3-cell stack with water circulation. Moreover, in case of Fig. 5 (a) and (c), the temperature of the unit tends to

change due to the heat generated by electrolysis. Therefore, it was inferred that differences in voltage in the

result shown in Fig. 5 were caused by the temperature of the electrolytic environment; moreover, the three data

graph shows that the 3-cell stack has almost the same characteristics.

Heatsink

H2O inlet

H2 outlet

O2 outlet

Inside

view

window


5

Figure 5. Current-voltage characteristics of new cathode feed electrolyzer and conventional reference cell.

○: (a) 3-cell stack electrolysis unit with gas/liquid separation function and no-water circulation. □: (b)

Same electrolysis unit as (a) with water circulation. ＋: (c) JAXA's previous single cell electrolysis unit

with gas/liquid separation function and no-water circulation.

Figure 6 shows the relationship between the dew point of H2 gas, the flow rate of moisture in the H2 gas

calculated from it, and the electrolytic current density. The H2 gas dew point was approximately room

temperature and the generated gas was almost the amount of saturated water vapor. Therefore, in proportion to

the generated gas flow rate, that is, the electrolysis current density, H2O is lost from the electrolytic cell unit

using H2 gas as a medium. Figure 7 shows the result in the case of continuing to send only the theoretical

amount of water consumed in water electrolysis to the electrolysis cell unit. Cell voltage in one of the three cells

suddenly became higher in 120 minutes. This indicates a shortage of water supply to the MEA. Until now, the

regularity at which a cell has a water shortage in the 3-cell stack has not been obtained. The amount of water

supplied is predicted to be very small, and the random inner state of gas and water of each unit has a large

influence on water supply to each cell. Although countermeasures are being considered against water loss

except for the electrolysis of water, stable electrolysis lasting at least 180 minutes was possible by sending the

corrected amount of estimated water supply from the results shown in Fig. 6 and Table 2 (see Fig. 8).

Figure 6. Dew points and H2O amount of hydrogen gas.

×: (a) Dew points of hydrogen; □: (b) H2O weight per minute in hydrogen.


6

Table 2 Amount of H2O permeating through the PTFE hydrophobic membrane.

0.3 (A/cm2) 0.4 (A/cm2)

Vapor amount in H2 gas (mg/min.) 7.24 9.97

H2O recovery amount using traps

(not vapor) (mg/min.) 4.42 7.93

H2O consumption by electrolysis

(theoretical value) (ml/min.) 0.322 0.430

Figure 7. Changes in cell voltages in case of lack of H2O supply (at 0.4 A/cm2).

Figure 8. Changes in cell voltages in case of corrected H2O supply amount (at 0.4 A/cm2).

· Water supply and dissipation

In the electrolytic cell based on the capillary feed system, a dead end water supply method was adopted to

supply water for water electrolysis. Water electrolysis consumes water at 0.322ml/min., but saturated steam was

found to be mixed in the generated gas. The calculation was made that 0.00609 g/min. of water dissipates with

H2. Water vapor dissipated with H2 gas was measured by using the silica gel absorption method (where the

weight of vapor absorbed by the silica gel was measured); moreover, the result was almost the same as the

calculation result of dissipation by saturated steam. In addition to water electrolysis, the capillary feed type

electrolysis cell must supply water for dissipation; moreover, maintaining the interface in the microgravity field

was also found to be necessary.

Figure 9 shows the flow rates of H2 and O2 with respect to electrolytic current density. For each electrolysis

current density, the flow rates of H2 and O2 gas almost equal to the theoretical values were obtained. Figure 10


7

shows the concentration of H2 in O2 with respect to electrolytic current density. When the electrolysis current

density is low, the percentage of hydrogen in oxygen appears to be high. As the electrolysis current density

increases, that is, as the O2 gas flow rate increases, the H2 concentration decreases in inverse proportion thereto.

When the electrolysis current density is medium, the percentage of hydrogen in oxygen appears adequate. This

indicates that the H2 cross leak amount of the MEA is almost constant independent of the electrolytic current

density, which is a general characteristic when water electrolysis is performed using the Nafion membrane. In

other words, it indicates that the purity of the generated gas does not change due to the gas/liquid separation

function of the present electrolytic unit with the PTFE hydrophobic membrane. These results indicate that the

gas/liquid separation function of the electrolytic cell unit works sufficiently.

On the other hand, pressure fluctuations in the H2O line of the electrolytic cell unit were observed (Fig. 11).

When the initial moisture content in the electrolytic cell unit was large (Fig. 11, gray lines) and when it was

small (Fig. 11, black lines), the range of fluctuating pressure was different. From this, it can be seen that as the

amount of water increases, pressure fluctuations tend to occur, and fluctuations in the interface between the

generated H2 gas, water, and PTFE hydrophobic membrane are largely involved in the gas/liquid separation

phenomenon, so that fluctuating pressure propagates between the cells. We predicted that the value oscillated. In

order to obtain a more stable gas/liquid separation function, it is thus necessary to keep the initial moisture

content low inside the electrolytic cell unit so as to smoothly release the hydrogen gas through the PTFE

hydrophobic membrane.

Figure 9. Current density and generated gas flow rate (standard).

●: Hydrogen; ■: Oxygen

Figure 10 Hydrogen concentrations in generated oxygen and flow rates of generated oxygen.

○: (a) hydrogen concentrations, +: (b) Flow rates of generated oxygen.


8

Figure 11. Changes in pressure of the H2O line.

When initial moisture content in the electrolytic cell unit was large (gray lines) and small (black lines)

· Exhaust heat evaluation of the electrolytic cell

In order to compare the effect of cooling fins on the electrolysis cell unit, electrolysis was conducted at

electrolytic current density of 0.3 A/cm2 and 0.4 A/cm2 with and without cooling fins, respectively, and

electrolysis was carried out for 180 minutes, with the surface and internal temperatures of the electrolytic cell

unit being measured.

Figure 12 shows the temporal change of the average value of internal temperature (positions c, d, e in Fig.

13) of each cell in the electrolytic cell unit when operating at an electrolysis current density of 0.3 A/cm2 for 180

minutes. The temperature rises with time, but in the case without heat radiation fins, the ultimate temperature is

about 4°C higher. Figure 14 shows the temperature distribution on the surface of the electrolytic cell unit after

operating for 180 minutes at an electrolysis current density of 0.3 A/cm2. When the cooling fins are removed,

heat is not transferred from the electrode to the whole cell, and thus the whole area is cooler (Fig. 14 (a)). When

the cooling fins are mounted, heat is transferred from the electrode to the whole cell, and thus the whole area is

hotter (Fig. 14 (b)). And given the temperature distribution in the direction perpendicular to the MEA, the

results for the surface shown in Fig. 13 and the measurement points inside it were obtained (Fig. 15). In both

cases of 0.3 A/cm2 and 0.4 A/cm2, the temperature was lower at all measurement points when the cooling fins

were mounted. In the center of the electrolytic cell unit, the temperature was about 25% lower when the cooling

fins were mounted. These results show that when there are no cooling fins, more energy is stored internally than

when the cooling fins were mounted. And it was confirmed that even with natural convection, there is a large

waste heat effect.

Figure 12. Changes in average temperatures

Gray line: without a radiator; Black line: using a radiator; Dashed line: room temperature.


9

Figure 13. Structure of the electrolysis unit and thermocouple positions of the surface and inside.

(a): without a radiator (b): using a radiator

Figure 14. Temperature distribution of the electrolysis unit surface (at current density of 0.3 A/cm2).

(a): 0.3 A/cm2 (b): 0.4 A/cm2

Figure 15. Temperature distribution in the electrolysis unit (at current density of 0.4 A/cm2).

Position indicates distances from the cell 1 side surface. ΔT indicates temperature differences from room

temperature. ×: Without fins;○: With fins.


10

IV. Conclusion

In this report, a scaled up electrolysis unit with a gas/liquid separation function and without water circulation

was evaluated regarding the “influence of a multi stack on stable water electrolysis at no-water circulation,”

“influence of a multi stack on gas/liquid separation performance” and the “effectiveness of the air cooling type

radiator for the electrolytic unit.” As a result, the following findings were obtained.

- From the results of the current-voltage characteristics, H2 cross leak amount, etc., there was almost no

difference in the basic characteristics between the 3-stack cell and single stack cell, and with or without

water circulation.

- H2O vapor contained in H2 gas and H2O leaking from the PTFE hydrophobic membrane were larger

than the amount of H2O consumed by electrolysis. Therefore, stable water electrolysis without water

circulation was possible by correcting the water supply amount, including H2O vapor and leaking loss.

- There was no effect of stacking on the gas purity of gas/liquid separation performance. However,

pressure fluctuations occurred in the H2O line, which were expected to be caused by the amount of gas

and water in the unit. Therefore, the interface between gas and water inside the electrolysis unit should

be researched for stable gas/liquid separation.

- As the result of attaching the heat pipe and heat sink radiator to the electrolysis unit, and only cooling

the heat sink by natural convection, temperature in the center of the unit was 25% lower than in the case

without a radiator. From the result, it was confirmed that air cooling using a radiator has a large exhaust

heat effect for the unit.

References

1Masato Sakurai, Mitsuo Oguchi, Shoichi Yoshihara and Mitsuru Ohnishi (2005) “Water Electrolysis Cells

Designed for Microgravity Conditions in Order to Establish Air Revitalization System.” Proc. International

Conference on Environmental Systems (ICES) July, Roma, 2005-01-2945. 2M. Sakurai, A. Shima, Y. Sone., M. Oguchi and M. Ohnishi (2013) “Air Revitalization Demonstration on

the JEM (KIBO) for Manned Space Exploration for Manned Space Exploration.” 43rd ICES, Vail, Colorado,

14-18 July 2013. 3M. Sakurai, A. Shima, Y. Sone, M. Ohnishi, S. Tachihara and T. Ito (2014) “Development of Oxygen

Generation Demonstration on JEM (KIBO) for Manned Space Exploration.” 44th ICES, Tucson, Arizona, 13-17

July 2014. 4E. A. Kurmazenko, N. M. Samsonov, L. I. Gavrilov, N. S. Farafonov, Ju. E. Belavencev, N. V. Pavlova, V.

Ju. Proshkin, S. Ju. Romanov, A. G. Geleznyakov, A. M. Ryabkin, G. A. Lyubimov and O. V. Kirushin (2005)

“Off-normal Situations Related to the Operation of the Electron-VM Oxygen Generation System aboard the

International Space Station.” Proc. International Conference on Environmental Systems (ICES), 2005-01-2803. 5Robert J. Erickson, John Howe Jr., Galen W. Kulp and Steven P. Van Keuren (2008) “International Space

Station United States Orbital Segment Oxygen Generation System On-orbit Operational Experience.” Proc.

International Conference on Environmental Systems (ICES), 2008-01-1962. 6Wolfram Knorr, Gijsbert Tan and Johannes Witt (2004) “The FAE Electrolyze Flight Experiment

FAVORITE: Current Development Status and Outlook.” Proc. International Conference on Environmental

Systems (ICES), 2004-01-2490. 7Wolfram Knorr, Gijsbert Tan, Johannes Witt and Berengaria Houdon (2005) “The FAE Electrolyze Flight

Experiment FAVORITE: Final Design and Pre-flight Ground Test Results.” Proc. International Conference on

Environmental Systems (ICES), 2005-01-2809. 8Masato Sakurai and Takuma Terao (2016) “Study of Water Electrolysis under Microgravity Condition for

Oxygen Generation - Applying to Ground Demonstration System and Development of New System -.” Proc.

International Conference on Environmental Systems (ICES) July, Washington, ICES-2016-204. 9Masato Sakurai, Asuka Shima, Yoshitsugu Sone, Takuma Terao and Mitsuru Ohnishi (2015) “Development

of Water Electrolysis System for Oxygen Production Aimed at Energy Saving and High Safety.” Proc.

International Conference on Environmental Systems (ICES) July, Washington, ICES-2015-273.

Study on Water Electrolysis for Oxygen Production ...

Documents

Transcript of Study on Water Electrolysis for Oxygen Production ...