AIAA-91-2444

Propellant Production on Mars: Single Cell Oxygen Production Test Bed

J. Colvin, P. Schallhorn, and K. Ramohalli UA/NASA Space Engineering Research Center University of Arizona Tucson, AZ

A I AA/SA E/AS M E/AS E E 27th Joint Propulsion Conference

June 24-26/1991 / Sacramento. CA For permission to copy or republish, contact the American lnsfiie of Aeronautics and Astronautics 370 L'Enfant Promenade, S.W., Washington, D.C. 20024

AIM-91-2444-CP

PROPELLANT PRODUCTION ON MARS: SINGLE CELL OXYGEN PRODUCTION TEST BED

James Colvin', Paul Schallhorn" and Kumar Ramohalli"' UA/NASA Space Engineering Research Center

Aerospace and Mechanical Engineering University of Arizona, Tucson, Arizona 85721

AHSTRACT

This paper describes one area of a project whose general aim is to produce oxygen from the indigcnous resowces on Mars. After discussing briefly the projcct's background, the case for indigenous material utilization, and the hcncfits in tcriiis of lower overall cost (through the reduction i n launch mass), specific cxpcrimental details of the clcctrolytic cclls are prcscntcd. At the heart of the oxygen production systcm is a bank of solid zirconia electrolytic cclls that will clcctrochemically separate oxygen from a high tempcraturc stream of carbon dioxide.

Expcrimcntal results arc discussed. The parameters varied include: the cell operating temperature, the carbon dioxide flow rate, and the voltage applicd across tlic cell. The results confirm our thcorctical cspcctalions. A novcl fcaturc of our work, bcsidcs bcing the first full-system production of oxygen from carbon dioxidc, i s the high-tech diagnostics through a n inlrarcd video imaging camera that monitors the overall system zinc1 alerts us to local hot spots and ~ii;~l~iinction.

w

BACKGROUND

A primary concern for any vcnture into space must be how much cnergy will the mission require. As we plan to visit local cntratcrrcstrial bodies niorc frequently, or a s we plan to venture fiirthcr i n t o space, an important consideration has to be, should we continue to bring all propellants with us from carth, or should we take advantage of the many resources that arc available to us once we reach our destination? With this idea in mind and wishing to expand on the success of the Martian Viking program, Ash, Dowlcr, and Varsi' in the Iatc seventies envisioncd an in-situ propellant plant which would make use of the Martian atmosphere to produce an oxygen and mcthanc propellant.

At the heart of this systcm would hc an array of zirconia solid clcctrolytc cells. Thcsc cclls liavc tlic ability to sclcctivcly transport oxygen ions thus allowing the production of pure oxygen. Tlic oxygen plant has undergone many changes sincc it was first envisioned by Ash et al. Lawon and Frisbee, at the Jet Propulsion Laboratory (JPL),

* Gr;iduiite Kescarcli Assistant, Student Member, AlAA

.. NASA/IJA Space Engineering Fellow, Student Member, AlAA

I . .

Professor, Associate Fellow, AIAA

Research funded by NASA as part of the UA SERC grant. Grant No. NAG-1332

Copyright (e) 1991 by James Colvin, Paul Schallhorn, and Kumar Ramohalli Reprinted by the American Institute of Aeronautics and Astronautics, with permission.

-."

have done extensive work on improving the overall s y s t e n ~ ~ ' ~ ~ ~ , reducing the total system mass and increasing the total system reliability.

Their improved system would have the Martian atmosphere drawn in through an electrostatic dust filter which is necessary ;is tlierc arc numerous long tcrm dust storms on the Mtirtian suifacc. The ;itinospliere, which consists of approximatcly 95'% carbon dioxide, will be drawn into the system by a

the atmosphere to be compressed from [lie ambient pressure of 6.8 nib to a pressure of 1 bar for delivery to tlie cathode side of tlic electrolyte. Before entering the zirconia array, the flow will pass through :I Iie;it exchanger which will raise the tcmpcr;itiire from the C 0 2 compressor's exit of GOO K to :tpproxini;rtely 1000 K. The source of the lieat I'm tlic lieat exchanger is tlie exhaust flow from tlic array. Oncc the llow has cntcrcd the zirconia ii iray, i t will l ie fiirtlicr heated to a temperature of 1273 K. This temperature is sufficient t o partially dissociate t l ic carbon dioxide into carbon monoxide and diiitoiiiic oxygen.

11 is the 0, that the cell will selectively transport to the anode side of the cell where the pressure of the 0, will be increased to 4 bar. The cell's exhaust \ v i consist of mostly CO, with some co also

ill. The 0, at the anode will now pass through ;I riidiator whcri it will be cooled from 1270 K to 250 K at 3.8 1x11, The 0, flow will next pass through an 0, adsorption compressor where the prcssiire will -be increased to 28 bar and its Icmperatore to 400 K. After passing through one more radiator, where the flow will be cooled to 230 K, the 0, will be finally cooled to 100 K by a niolcciilnr adsorption cryo-cooler refrigerator and stored for its final use. This use could initially be Llic oxidizer f o r the propellant necessary to return a Mart ian snmple to earth, and then eventually, life support lor :I manned Marti;in mission.

C02 adsorption compressor. Present pions 3 require

RIA STAUILIZED ZlRCONlA SOLID ELECI'ROLYTE

3 In L:iwton's work , he lists development risk factors {or components. In his option 111, thc oxygen cell is thc only component still listed as risk factor 4 meaning "there are still serious problems that must I)c xltlressed as well as some intensive development rcqiiiied." This is an area of current research being

W' conductcd at the NASA sponsored University of Arizona Space Engineering Rcsearch Center (NASAJUA SERC). Further details were worked out i n refcrence 5.

Since the Martian atmosphere is prcdomiiiiintly CO,, [lie remnindcr of this rcport will coiisidcr tlic atmosphere ;IS CO,. Tilc CO,, when i t ciitcis tlic cel l array, will be heated to 1273 IC. At this temperature, the CO, will bcgin to pnrti;illy dissociate into CO and OT The zirconia electrolyte is siindwiched between two porous p1:ititiiini

electrodes. This dissoci;ition will occur on tlic cathode side with the 0, entering tlic electrode i i i i d

moving towards the electrode-electrolyte inteifiicc.

Tlic driving potentiid for this movement is tlic par t ia l pressure gradient dcvclopcd by thc electrolyte removing oxygen from the interfacc aie;i. Oncc tlic O2 r~ci)cI~es tlic intcri;icc, i f is iu i l l icr reduced to monatomic oxygen. This oxygen rcccivcs two electrons from the negative electrode imd

beconies an oxygen ion and begins to migi;ttc through the electrolyte towards the anode. Upon reaching the anode, the ion will release its two

recombine with a n o h x oxygen atom to foiin llic 0, molccule. Qualitatively, this is how the oxygcn sepiiration process occiirs. Figure l4 shows this proccss schcmatically.

electrons to the positive electrode and tlieii W

I,".% -_

Richter, at JPL, performed the initial intcnsivc testing of this electrochemical process with thc aim of quantifying this physical procedure6. His w o ~ k

dcvelopcd the basic tlierniodynamic mid was performed using a tubular zirconia cell. I le v

2

v

electrochemical models for tlie reduction of C 0 2 and the subsequent production of O2 A few years later at JPL, Suitor continued the investigation? In his experiments, he used the disk geometry for his cells. Additionally, he investigated the use of different electrode materials.

I n tlic work being conducted at our ccntcr, we are using cells of the tubular geometry. Our aini is to invcstig;ite various cell cfliciciicies by varying several of the controlling p;irarneters. These prunctcrs arc: a ) the potential applied across the electrolyte; I)) tlic electrolyte operating temperature; and c)the incoming C02 flow rate. We would like IO know the 0, - production rate a s ;I function of tlicsc piiramctcrs. The various e[liciencies include: a ) the cell's over potential (Nernst efficiency), and b) t l ic system efficiency. The definitions of these efficiencies will be dcveloped in this report.

1

To simulate the M;irti;m atmosphere, wliicI1 contains ;ipproxiiii;itcly 95% CO, ; i d only 0.13% O,, Coleman grade CO, wits used to feed the test bLd. This grade contains less than 20 ppni O2 The flow was ni;iint;iincd :it slightly above local atmospheric pressure (13.7 psia) at a tempcrature 1)f

75 degrees F. Thc flow rates were varied between 37.5 sccni and 1475 sccm. Flow would enter the zirconia tube through a 1/8 inch alumina tube and then pass to the far end of tlie zirconia ch;imber (see figure 3).

I I . SING1.E ZIItCONIA CEI.1. TEST IIU

TIE single zirconia cell test bed currently i n place at tlic NASA/UA Space Engineering Center is shown schematically in figure 2. The test bed consists of the following:

I ) One tubular zirconia cell 2) One low voltage DC power supply 3) Two digital multimcters 4) Two clam shell type heaters surrounding thc cell 5) One Watlow lieater controller 6) I<;iowool type ceramic irisulation 7) Onc oxygen pressure transducer

(measuring in absolute pressure, 0-500 psia) S) One carbon dioxide prcssure transducer

(measuring in absolute pressure, 0-500 p ia ) 9) One oxygen flow meter (0-50 ISCCM) 10) One carbon dioxide flow meter (0-5oOa ISCCM) 11) Three K-type thermocouples 12) Anaerobic carbon dioxide 13) Oiie 386 based PC for t l i i t i i acquisition 14) Gas Cliromntograpli

--./

The clam shell heaters are centered about the middle 7 inches of the zirconia device. This means flow heating will begin in tlic aluminit tube. Exiting the alumina tube, the flow reverses direction, continues to be heated, and flows across the cathode of the electrolyte. The C02 begins dissociating :it the cathode. Tlic stream is now :I

mixture of CO,, - CO, and 02. O2 is then electrochemically pissct l throiigli the zirconia to the anode while tlic CO, ;mtl CO cxllaust pass out tlirougI1 t11c exit. n 1 C CO, Ilow rate is controlled by a metering valve in this exhaust flow. 0, flow passes through a volumetric flow meter and then can be directed to a water hubhling device, 21

sample cylinder, or directed to the gas

3

chromatograph for analyzing. The exhaust flow can also be directed to the gas chromatograph for analyzing. The zirconia, and its electrode, havc an upper temperature limit of 1200°C. A critical voltage limit of 2.23 VDC was assumed in accordance with Frisbee4. With these material liiiiils i n mind, a self imposed limit of 110o"C, and 2.0 VDC was used during a11 testing. The i isscmhlcd single cell test bed is shown i n figure 4.

v increments of 0.1 VDC. The C0, flow rate varied from 37.5 cc/niin to 1415 cc/min. The followiiig figures show the results of this testing. Figure 5 is a characteristic curve of the production rate of O2 vs the applied cell potential for a variety of COz flow ratcs at a temperature of 1000"C, while figure 6 displays 1111: production for four teniperatorcs ill a C0, flow rate of 137.5 cc/min. Note, on both figurcs 5 and 6, the sccoiid order dcpcndcncc of oxygcn production on the applied cell voltagc.

&tire 4: Sin& Cell Test Ilucl

111. SINGLE CELL TEST RESULTS

FUNDAMENTAL RESULTS

A series of tests were conducted to attempt to characterize the effects of tcmperature, cell potential, and carbon dioxide flow rate on the production rate of oxygen. During the testing tcmperatiire w a s varied from 800°C to 1100°C in iiicrenieiits of 25 degrees. Cell potential varied from the zero 0, production voltage to 2.0 Volts in

Fietire 5: Oxyeen Prochiction vs Cell Viiltiig (T = l(1011 "C)

Ficiire Oxycrn Prodiiction vs Cell Voltwe (co, Flow R ; l k = 137.5 SCCM)

4

v

Figure 7 demonstrates the dependence the oxygen production has on temperature of the cell, for an applied cell potential of 2.0 VDC. Notice the weak dependence the oxygen production has on the carbon dioxide flow rate. Figure 8 presents the oxygen production vs cell temperature at a carbon dioxide flow rate of 137.5 cc/min for various flow rates. Figure 8 clearly depicts the dependence of voltage on the oxygen production.

v Fimre 7: Omcen Production vs Temperature fCell Voltage = 2.0 VDCl

0, PRODUCTION YS TEUPEUTURE 7

Firrure 8: Oxvgen Production vs Temperature (co, Flow Rate = 137.5 SCCM)

Figure 9 presents the 0, production dependence on CO, flow rate, for an applied cell potential of 2.0 VDC. This is the most graphic illustration of the

..."

lack of dependence of the carbon dioxide flow rate on the oxygen production, especially for C 0 2 flow rates greater than 200 SCCM. It, like previous graphs, shows the temperature dependence on the 0, production.

Fiaure 9 Oxvaen Prnduction vs CO, Flow Rate

INTERPRETATION OF THE RESULTS

The four important results are: flow rate effects, cell voltage effects, cell temperature effects, and chemical effects. The latter refers to phenomena such as carbon formation, zirconia cell breakdown, local "hot-spot'' effects, etc

The flow rate effects are the easiest to infcrprcc. The basic schematic is shown in figure 10. It was initially found that thc oxygen yield r a k (production rate) increased as the C0, flow rate increascd, but only up to a certain point, after which increased flow rate actually resulted in the decrease of the yield of oxygen. This was observed while other parameters such as cell voltage and cell tcmpcraturc was held constant. It was suspected that the carbon dioxide may not have had sufficient residence time within the heated cell to achicvc the required temperature at the high flow rates To verify this hypothesis, a simple thermal difrusiviry analysis was performed, after confirming that thc flow within the tube is indeed laminar via simple Reynolds number calculation. The residence time is given by I/u, where 1 is the tube length and u is the mass averaged velocity. The characteristic timc for licat transfer within the tube is given by (d/2)'/a, where

5

9

n

li'ignre 13: Oween Conversion Rate

As tlic concern of this report is the complete cliai actcrization of the zirconia electrolytic cell, all mnvcrsion efficiencies must be investigated. Richtci, in his early work6 on the reduction of CO,, began with the basic Nernst relation:

v

Hcrc, F& is thc ncrnst voltage, T is the temperature (K), z is chargc/molc (2=4), R is the universal gas constant, and F is the Faraday constant. Atlditionally, P2 i s the pressure at the anode of the ccll wliilc Xo3 P, is the partial pressure of the oxygen in the incoming CO, flow stream at the cathodc. Richter then continued with his model and dcvclopcd an instantaneous Nernst potential:

In this equation, K is the normal equilibrium constant and n is the mole fraction of CO, reduced to CO. The value n can be calculated from this equation:

I n-C- . vco,

'The value C is a constant from Richter's work (C=6.969), I is the current produced across the cell, and Vco2 is the volumetric flow rate of the incoming C0, stream. The current (A) was

measured experimentally, as well ;IS the volumetric flow rate of the C 0 2 (SCCM). The cell temperature was measured by a k-type thermocouple positioned at the outside of the cell wall (K), and the pressure of the 0, collection side of the ccll was assumed to be one atmosphere. Using the proper values for the cquilibrium constant K*, there is now enough information availablc to calculate the instantaneous Ncrnst voltage. Richter cautions the use of this equation by stating it is valid only if the critical voltage over the cell is not reached. This can be defined as the potential which will be just sufficient to hcgin driving oxygen from the lattice of the zirconia. This process will be apparent from an "elbow" being ohscrvcd i n the current vs potential plot. Figure 5 shows this potential has not been reached in this work. It is evident that this potential is now a function of the operating temperature a i d the actual amount of 0, removed and not the partial pressurc of the 0, at the cathode. The Ncrnst potential is used in defining the Ncrnst efficiency.

This equation shows thc Ncrnsc efficiency is the ratio of the theoretical potential divided by the actual potential (Eact). What this indicates i s the instantaneous amount of cell overpotcntial. Figure 14 shows the results of this analysis.

Figure 1 4 Nernst Elficiencv vs Cell Vnitwe

It is apparent from this figure that the amount of ccll overpotential rises as [he actual cell potential rises. One sourcc of this overpotential could be the pressure drop through the porous electrode. It is important to remember that from figure 5, the rise in 0, production is almost directly proportional to the potential applied to the ccll. Ash9 ct ai, point oi i t that experiments have shown that as thc 0, production rate rises, there is an increase in the prcssurc drop across the negative electrode which can approach thc partial pressure of the oxygen in thc CO, stream resulting in an evcr increasing pumping power requirement. Another possible sourcc of the ovcrpotentiat is slow diffusion, adsorption, or dissociation processes near the electrode-electrolyte interfacc of the cell resulting in concentration ovcrpotential'O. They also state thc possibility of a transition overpotential due to slow clcctroclicmical reaction. Thc sourccs of ovcrpotcntial can not be cxpcrimcntally determined with this system as currently constructed, but any lulurc char;ictcrization of the zirconia clcctrolytc should dcfinitcly kccp in mind these processes.

With the Ncrnst efficiency established, it was ncxt desired to arrivc at a system efficiency. The control volume is taken around the entire cell system. The useful work inside the control volume is divided into two parts. The first being the rate of tlicrniodynamic encrby being used to dissociate the CO,. Tlic second being the work required to clcctrochcniically conduct the oxygen ions through the zirconia. The power passing into the control volume is also divided int two components. The first being the power delivered to the ceramic hcatcrs and the second being the power delivered to the clcctroly~e. The energy required to thermally dissociate the CO, is given by the equation:

Hcrc, O.,.,, i s tlic thermodynamic power input with thc calculation showing the normal enthalpy of reaction equation with the cnthalpy of formation and scnsiblc cnthalpics. The subscripts i and j rc.prcscnt the products and reactants respectively. The cocfficicnt N represents the molar flow rate of cacli respective constituent. This value and thc other power valiics give the following definition for the systcm efficiency:

In this equation, the subscript heat rcfcis lo obscrvcd conditions on the ceramic hcatcr a i i d [lie subscript ccll refers to obscrvcd conditions with tlic clectrolyte, E being the potential (V) and 1 bcing the currcnt (1). Figurc 15 displays the results of this analysis.

O w o h - & ' 1: ' , i ' ,!5 5T-Z ' 7 , r-. . -~ 2 ' )

CELL VOLTACE I W C )

Firure 15: Sinde Cell System Elliciency

N. FUTURE WOKK

In the future there arc plans to address issiics such as how the 0, and CO, back prcssurc may ;lCfccl the production, ctc. Concerns such as ccll degradation due to extended periods of opcriitioii require the study of the loug term opcratioo of tlic single cell test bed.

In addition to the single ccll test bcd, w o r k is beginning this sumnicr for the first scale-up of [ h i s work. This scale-up includcs the use or sixiccii cells, the use of tlic adsorption compressor, ;I

vacuum chenibcr (to simulate the Marli;in atmosphere, complete with dust), an cIcctrost i i~ic dust filter, and a device to allow for selective ci i r lx~n deposition (necessary due to problcms with CO, - tit high temperatures).

8

Y

V. SUMMARY

This paper has presented results from the first phase of a well planned multi-phase research program aimed at significantly reducing costs of futurc space missions; terrestrial applications are always kept in mind. This proof of concept study has realistically used full system hardware, realistic solid electrolytic cell material, realistic operating temperatures, full cell voltages, full scale flow rates, and most important, realistic long-duration operation. These features distinguish our cxperiments from the usually understood test-tube demonstrations, where the conditions are subst;intially different from the real life counterpart. The iinly feature that is not duplicated here is the

scale of oxygen production.

Important efficiencies arc defined and measured. The basic electrolytic efficiency refers to the efficiency of using electrical input in producing oxygen, while the overall system efficiency refers to the enthalpy difference achieved between the product stream and the reactant stream divided by - t h e overall energy input. The carbon dioxide flow ralc, cell voltage, cell temperature and the duration

A simple heat transfer theory explains the flow rate effects. The voltage effects arc in good agreement with the manufacturer's specifications.

'rhcsc thorough characterizations of the component performance parameters naturally lead us to the ncxt step of scale-up and full system demonstration with active controls. Creative solutions to these engineering designs indicate that h ture space missions could rcalizc substantial cost savings through lhc use of local (in-space) rcsourccs.

operation are all varied parametrically.

REFERENCES

1. Ash, R. L.; Dowler, W. L.; Varsi, G.: "Feasibility of Rocket Propellant Production on Mars." Acta Astronautica, Vol. 5, pp. 705-724. (1978)

4.

5.

6.

7.

8.

9.

10

Lawton, A. L.; Frisbee, R. H.: "A New Look at Oxygen Production on Mars ISPP." JPL D-2661, September 1985

Lawton, A. L.: "Risk Factors in the Development of Zirconia Cell Technology for the Production of Oxygen from the Martian Atmosphere." JPL D-3546, August 1986

Frisbee, R. H.: "Mass and Power Estimates for Martian In-Situ Propellant Production Systems." JPL D-3648, Octobcr 1986

Ramohalli, K.; Lawton, E.; Ash, R.: "Recent Concepts in Mission to Mars: Extraterrestrial Processes." AIAA Journal of Propulsion and Power, Vol. 5, No. 2, pp 181-187. Mar-Apr 1989

Richter, R.: "Basic Investigation into thc Production of Oxygen in a Solid Electrolyte." AIAA-81-1175, June 1981

Suitor, J.W.; Berdahl, C.M.; Fcrrall, J.F.; Marncr, W.J.; Schrodcr, J.E.; Shichta, P.J.: "Development of an Alternate Oxygen Production Source Using a Zirconia Solid Electrolyte Membrane." JPI. D-4320, May 1987

Wark, K.: Thermodvnamics. Third cd. p. 840. McCraw-Hill, Inc. 1977

Ash, R. L.; Richter R.; Dowlcr, J. A,; Hanson and Uphoff, C . W.: "Autonomous Oxygen for a Mars Return Vchiclc." IAF-82-210 1982

Etsell, T. H.; Flengas, S. N.: "Ovcrpotcntiel Behavior of Stabilized Zirconia Fuel Cells." Journal of the Electrochemical Society. Vol. 118, No. 12, pp 1890-1900. Dec. 1971

9