NASA Technical Reports Server (NTRS) 19930017507: Full system engineering design and operation of an oxygen plant

18 Pages·0.61 MB·English

by  NASA Technical Reports Server (NTRS)

N93-26696
Full System Engineering Design and Operation of an -
James Colvin, Paul Schallhom, and Kumar Ramohalli
University of Arizona/NASA Space Engineering Research Center
Abstract
This paper describes one area of a project whose general aim is to produce oxygen from the
Indigenous resources on Mars. After discussing briefly the project's background and the
experimental systemdesign, specificexperimental resultsofthe electrolytic cellare presented. At
the heart of the oxygen production system is a tubular solid zirconia electrolyte cell that will
electrochemically separate oxygen from a high-temperature stream of Coleman grade carbon
dioxide. Experimental results are discussed and certain system efflclencles are defined. The
parameters varied include 1)the celloperatingtemperature; 2) the carbon dioxide flow rate; and
3) the voltage applied acrossthe cell.The resultsconfirmour theoretical expectations.
IV-33

Introduction
Backaround
A primary concem for any mission to Mars must be how much energy will be necessary to
complete the mission. An important consideration has to be: Should we continue to bring all
propellants withusfrom Earth,orshouldwe takeadvantage ofthe many known resources that are
availableto uson Mars?Withthisidea inmindandwishingtoexpand onthe success ofthe Martian
Viking program, Ash, Dowler, and Varsi_inthe late seventiesenvisionedan in-situpropellant plant
which would make useofthe Martianatmosphere to produce an oxygen and methane propellant.
The heart ofthissystemwould be anarrayofyttriastabilizedzirconia solidelectrolyte cells.These
cells have the ability to selectively conduct oxygen ions, thus allowing the production of pure
oxygen. Theoxygen planthasundergone many changessinceitwasfirst envisioned byAsh1etal.
FrisbeeandLawtonhavedone extensiveworkon improvingtheoverallsystem byreducingthetotal
system massand increasingthe total systemreliability.2-4
Their improved systemwould havethe Martianatmosphere drawn inthrough an electrostaticdust
filter, which isnecessaryasthereare numerouslong-term dust stormson the Martian surface. The
atmosphere, which consistsofapproximately 95%carbondioxide,willbedrawn intothe systemby
a CO2adsorption compressor. Presentplans3requirethe atmosphere to be compressed from the
ambient pressure of 6.8 mb to a pressure of 1 bar for delivery to the cathode of the electrolyte.
Before enteringthe zlrconla array,the flowwillpass through a heat exchanger which will raisethe
temperature from the CO2compressor's exit temperature of 600 K to approximately 1000 K.The
source of the heat for the heat exchanger isthe exhaustflow from the array. Once the flow has
entered the array, itwill be further heatedto atemperature of 1273 K.
This temperature is sufficientto partially dissociatethe carbon dioxide into carbon monoxide and
diatomic oxygen. The 02 will be increased to 4 bar.The cell'sexhaust will consist of mostly CO2
with some CO resultingfrom the removal of02. The 02 produced at the anode will pass through
a radiator where it will be cooled from 1270 to 250 Kwitha pressure of 3.8 bar. The 02 will next
pass through an 02 adsorption compressor where its pressurewillbe Increased to 28 bar and its
temperature to 400 K.After passingthrough aradiator,theflowwill becooled to 230 K.The02 will
be finally cooled to 100 Kbya molecularadsorptioncryo-cooler refrigerator and storedfor itsfinal
use.The usecould initiallybethe oxidizerfor the propellant necessaryto returnaMartian sample
IV-34
to earth, and then eventually,lifesupportfor a manned Martian mission.
Yttd_-Stabilized Zlrconla Solid Electrolyte
InLawton's work,3he listsdevelopment dskfactors for components. In hisoption III, the oxygen
cell isthe onlycomponent stilllistedas riskfactor 4, meaning"thereare stillseriousproblems that
must be addressed aswell as some Intensivedevelopment required."This Is an area of current
research. Further details were worked out in Reference 5. Since the Martian atmosphere is
predominantly CO2,the remainder ofthisreportwillrefertotheatmosphere asCO2.TheCO2,when
itenters the cellarray, willbe heatedto 1273 K.At thistemperature, the CO2will beginto partially
dissociate into CO and 02. The zirconla electrolyte is sandwiched between two porous platinum
electrodes. Thedissociationwilloccuratthe cathodewiththe 02 enteringthe electrodeand moving
towards the electrode-electrolyte interface. The driving force for this movement is the partial
pressure gradient developed by theelectrolyteremoving oxygen from the interface area. Once the
O2reaches the interface, it is further reduced to monatomic oxygen. The oxygen received two
electronsfrom the negative electrode and becomes an oxygen ion and beginsto migrate through
the zirconia electrolytetowards the anode. Upon reaching the anode, the ion will release its two
electrons to the positive electrode andthen recombine withanother oxygen atom to reform the O2
molecule. Qualitatively,thisdescribestheoxygen separationprocess. Figure 14showsthisprocess
schematically.
Richter performed the initial intensivetesting of this electrochemical process with the aim of
quantifying this physical procedure._ His work was performed using a tubular zlrconia cell. He
developed the basicthermodynamic and electrochemical models for the reduction ofCO2 andthe
subsequent production of 02. A few years later, Suitor continued the investigation/ In his
experiment, heusedthediskgeometry for hiscells.Additionailylhe investigatedthe useofdifferent
electrode materials.Although hissupplygas wasair, many ofhisresultscan also be applied when
CO2isthe supply gas.
Inthisstudy, zirconia cellswiththe tubular geometry were used. The aim was to develop various
cellefflcienciesbyvaryingseveralsystemcontrolparameters. These parameters were: the potential
applied across the electrolyte, the electrolyte operating temperature, and the incoming CO2 flow
rate. We wouldliketo knowhowthe CO2production ratevariesasafunction ofthese parameters.
The various efflciencles include the cell's Nemst efficiency (a measure of the theoretical energy
required against the actualenergy putintothe system). Thedefinitionsof these efficiencleswill be
IV-35
COz--> CO + Oz(Ps) ......
N
+
_'%\\'%\\\\\\\\\\'_, _\\\',._
o_ (P2)
Fig. 1 Oxygen migration through the ZrO2 cell.
02 Purge ÷
Iv
)Flow Control
120 VAC
C02 * CO
Heater. J
Conti-oller I
Bu le Sample
Indicator
(_) Thermocouple @ Pressure Transducer
Q (_) (_ Flow/Volt/Current Meter
Fig. 2 Single cell test bed (schematic).
IV-36
developed later inthis report. The singlezirconia cell test bed currently being studied is shown
schematically inFigure 2.The testbedconsistsofthefollowing: 1)onetubularzirconia cell;2) one
voltage DC power supply; 3) two digital multimeters; 4) two ceramic clam shell heaters; 5) one
Watlow heatercontroller;6)Kaowooltype ceramicinsulation;7)one 02 pressuretransducer (0-500
psia);8)one CO2pressuretransducer (0-500psia);9) one02 flow meter (0-50 sccm); 10)one CO2
flowmeter (0-5000 sccm); 11)threektypethermocouples; (12)Coleman grade CO2for supplygas;
13)one PC/386 processor for data acquisition;and 14)Varian model 3700 gas chromatograph.
To simulate the Martian atmosphere, which containsapproximately 95% CO2and only 0.13% 02,
Coleman grade CO2wasusedtosupplythetestbed.Thisgrade containslessthan 20 ppm 02. The
flow was maintained at slightlyabove local atmospheric pressure (13.7 psia) at a temperature of
75°F. The flow rateswere varied between 38and 1475 sccm. Flow would enter the zlrconla tube
through a 1/8 in.aluminatube andthen passtothe far endofthe zirconiachamber (seeFigure 3).
The clam shell heaters are centered about the middle 7 in. of the zirconia device. This means
hearing of the flow will begin inthe alumlna tube. Exiting the alumina tube, the flow reverses
direction while continuing to be heated andflows across the cathode ofthe electrolyte. The CO2,
now heated, beginsdissociation and tsdrawn to the cathode. Thefree stream Is now amixture of
CO2,CO, and 02. O2isdissociatedto monatomic oxygen atthe electrode-electrolyte interface and
electrochemically conducted through thezirconiatothe anode whiletheCO2and CO exhaust pass
out through the exit ofthe tube. The supplyCO2flow rateis controlled by a metering valve inthe
exhaustflow.The 02 produced (at approximately 13.7psia)flowsthrough the massflow meterand
then can be directed to a water bubbling device, a sample cylinder, or directed to the gas
chromatograph for analyzing. The zirconia electrolyte and its platinum electrode have an upper
temperature limitof 1150°C wherea phasechange inthezirconia will take place. A criticalvoltage
limit of 2.23 VDC was assumed in accordance with the work of Frisbee.4 Keeping below this
potential will prevent the oxygen from being driven from the zirconia lattice structure causing
permanentdamage tothe cell.Withthesemateriallimitsinmind,aself-imposed limitof 1100°C, and
2.0 VDC was usedduring alltesting.
Single-CeU Test Results ==
F_nd_m_ntal Results
A series of tests were conducted to attempt to characterize the effects of temperature, cell
potential, and carbon dioxide flow rate on the production rate of oxygen. During testing, the
temperature was varied from 800 to 1100°C in increments of 25o. Cell potential varied from z
approximately 0.6 to 2.0 VDC In incrementsof 0.1 volts.The CO2flow ratevaried from 38to 1475 =
IV-37
eO I l i i I i ! i i i ! i ! i i ! !
C.)
5.0
.<
_4.0
Z
0 8
F-3.0
a
o2.0
r_ C02 FLOWRATE
z_zxzxzx 1000 SCCM
o 1.0 DDnDD 175 SCCM
ooooo I00 SCCM
A 0
¢r_ 38 SCCM
0 _"
A
0.0 ' I _r i , i • I ' I ' I ' I ' I '
0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
GURE4 CELL POTENTIAL (VDC)
_I_GI,E Z,,g&C_ _IIDLL_CJ_11C
ALUMINA ELECTROLYTE
, o, = o, =t, N
c _ co,. co I
= ........ ---- ....... "---'- 0 I -
_////_/_////////S//A
INCONEL
SURE3
SEALS
IV-38
sccm. Thefollowing figuresshowthe resultsofthistesting.Figure 4 showsthe dependence of the
production rateof 02 on the applied cell potentialfor avadety ofCO2 flow ratesat a temperature
of 1000°C, while Figure 5displaysthe production at fourdifferenttemperatures ata CO2flow rate
of 138 sccm. Note, on both Figs.4and5, the second-order dependence of oxygen production on
the applied cell potential.
Figure 6 demonstrates the dependence the oxygen production hason the temperature of the cell
with anapplied cellpotential of2.0 VDC. Notice theweak dependence the oxygen production has
on the carbon dioxide flowrate.Figure 7 presentsthe oxygen production vs celltemperature at a
carbon dioxide flow rate of 138sccm. Figure 7alsoclearly depicts the dependence on voltage of
the oxygen production rate.
Figure 8 indicatesthe 02 production rate dependence on the CO2 supplyflow rate for an applied
cell potential of 2.0 VDC. This isthe most graphic Illustrationof the lack of dependence on the
carbon dioxide flow ratefor the oxygen production rate,especiallyfor CO2flow rates greater than
about 200 sccm. Figure 8, likethe previousfigures, showsthe dependence on the temperature for
the O2production rate.
IntQrDretatlonof the Results
The resultsof the extensivesystemanalysiscan be discussedinfive areas: 1)flow rate effects, 2)
massflow ratio,3)oxygen conversionefficiency,4)Nernst efficiency,and 5)the systemefficiency.
Each of these areas will be addressed inorder.
The flow rate effects were the easiest to interpret.The basic schematic isshown inFigure 2. Itwas
initiallyfound that the oxygen yield rate (production rate) increased asthe CO2flow rate increased,
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
temperature were held constant. It was suspected that the carbon dioxide may not have had
sufficientresidence time within the heated cell to achieve the required temperature at the high flow
rates. To verify this hypothesis, asimplethermal diffusivity analysis was performed, after confirming
that the flow within the tube Is Indeed lamlnar via simple Reynolds number calculation. The
residence time isgiven by I/u, where/is the tube length and u isthe mass averaged velocity. The
characteristic time for heat transfer within the tube isgiven by (d/2)2/o:, where d isthe cell inside
diameter and = is the thermal diffusivity (k/pc_) of the carbon dioxide at high temperatures. As can
IV-39
10
! i
9
8-
7-
0
0
$-
(/I
V
Z
5"
0
F-
o
4-
o
0
n_
o
0.
COz FLOWRATE _r
3-
0CW ¢rA"A-A_IO00 SCCM
_A_ 175 SCCM
DOODO 100 SCCM
ooooo 38 SCCM
' I ' I '
0.0007 0.0008 0.0009 0.00 0
IGURE6 INVERSE CELL TEMPERATURE (K-')
7.0
i i i i i I I I I i i ! ! _ i
6.0 z_
_E CELL TEMPERATURE
o5.0
(..) _n_1100 C []
(/') _ 1000 C A []
DDODD 900 C
z4.0 cx:x:x:x_800 C 1_ _ []
0
o
I,--
(..) []
_3.0
0
o
0
o
z_ o
o-2.0 [] o
A o
¢_ o
,r,,l o
0 A o
,A, [] 0
1.0
cram 0
z_ [] 0
o o
.0 ' I ' I i I i t i I ' I -r I t I i
0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
:IGURE5
CELL POTENTIAL (VDC)
IV-40
be seen inFigure 9,the peak occursalmostpreciselywhere thetwo times become equal. Itisthus
clear that up to the point where the residencetime issufficientto bdng the carbon dioxide to the
requisite temperature, the production rate ofthe oxygen Increaseswith the flow rate, but beyond
thispointthe production ratefalls off.Itwouldbe instructiveto measurethe actualgastemperature,
and not merelythe cellwalltemperature as isdone Inthe experiments reported here.
The mass flow ratio (the mass production rate of 02 divided bythe CO2supply massflow rate) is
of much interest.This isdue to sizing constraintswithin scaled-up systems such as the primary
carbon dioxide compressor, theheat exchanger,andthe radiators.Forthese reasons, Figs. 10and
11 plotthe massflow ratiovsCO2 flowrate andtemperature, respectively.
Inpreviousstudiesofthe propellantproductionplant,theemphasiswas placed on sizingofthetotal
production system.2"4Forthese studies,oxygen conversionefficienclesof25-30% were usedwhen
considering the estimated surfacearea needed to produce a required amount of oxygen per day.
The use of these efflclencles was necessary to minimize oxygen plant mass. However, when
consideringthe operation ofthezlmonia cellalone,Figure 12showsthe complete range of oxygen
conversion efflciencles.
As the concem of this report isthe complete characterization of the zirconla electrolytic cell, all
conversion efflciencles must be investigated. Richter, in hiseady workBon the reduction of CO2
began withthe basicNemst relation:
Here, E_Nisthe Nemst voltage, Tisthetemperature (10,z ischarge/mole (z = 4), R isthe universal
gas constant, andF isthe Faraday constant.Additionally,P_2isthe pressure atthe anode while Xo2
P1isthe partial pressureofthe oxygen inthe supplyCO2flow stream atthe cathode. Richterthen
continued with hismodel and developed an instantaneousNernst potential
]
E_i" RTIn [ K(l-n)
zF ( )_.
n
IV-41