American Institute of Aeronautics and Astronautics1

Low Temperature Low-Earth-Orbit Testing of MarsSurveyor Program Lander Lithium-Ion Battery

Concha Reid*

NASA Glenn Research Center, Cleveland, OH 44135

A flight-qualified, lithium-ion (Li-ion) battery fabricated for the Mars Surveyor Program2001 lander is undergoing life-testing at low temperature under a low-Earth-orbit (LEO)profile to assess its capability to provide long term energy storage for aerospace missions.Li-ion batteries are excellent candidates to provide power and energy storage for satellites inLEO due to their high specific energy, high energy density, and excellent low temperatureperformance. Although Li-ion batteries are increasingly being used for aerospace missionsin geosynchronous orbit, some challenges still remain before they can be deemed a suitablereplacement for their secondary alkaline battery counterparts in long cycle life LEOapplications. Life cycle testing of this battery is being conducted in the laboratory tocharacterize battery-level performance and to examine the dynamics of individual cellswithin the stack under aerospace conditions. Data generated in this work is critical toestablish confidence in the technology for its widespread use in manned and unmannedmissions. This paper discusses the performance of the 28 volt, 25 ampere-hour batterythrough 9000 LEO cycles, which corresponds to over 18 months on LEO orbit. Testing isconducted at 0 °C and 40% depth-of-discharge. Individual cell behaviors and their effect onthe performance of the battery are described. Capacity, impedance, energy efficiency, end-of-discharge voltages, and cell voltage dispersions are reported. Relationships between celltemperatures, cell impedance, and their relative position in the battery stack are discussed.

I. Introductionesting of lithium-ion (Li-ion) cells and batteries is being conducted by the National Aeronautics and SpaceAdministration (NASA) Glenn Research Center (GRC) and other partner organizations to assess their

capability to perform under a variety of conditions and temperatures in simulated low-Earth-orbit (LEO) regimes.Currently, nickel-hydrogen (Ni-H2) and nickel-cadmium (Ni-Cd) batteries provide energy storage for the

majority of satellites in LEO. Li-based battery technology initially broke ground in its utilization in the US spaceprogram due to its ability to enable missions operating in low temperature environments, such as Mars. In additionto its low temperature capability, Li-ion offers improvements in mass, volume, efficiency, and operating temperaturerange over secondary (rechargeable) alkaline battery chemistries.

At present, Li-ion liquid electrolyte batteries provide 2.5 times the specific energy and 4.5 times the energydensity of state-of-the-art Ni-H2 batteries when operating at 100% depth-of-discharge (DOD), and advancementscontinue to be made. Li-ion batteries have a high nominal cell level operating voltage compared to Ni-H2 batteries(3.6 volts (V) versus 1.2 V), so fewer Li-ion cells are required to meet the bus voltage. Hence, fewer intercellconnections are required in the battery assembly, which provides the system with greater simplicity and reliability.

Long cycle life is an essential performance metric for batteries in LEO. Since most satellites in LEO operate forfive years or more while delivering electrical energy during approximately 16 cycles per day, the energy storagesystem must be capable of providing upwards of 30,000 cycles. Since a Li-ion battery that has the equivalent massand voltage of a Ni-H2 battery can operate at a much lower DOD, the cycle life of the battery could be extended dueto the lower DOD requirement, which is a more benign operating condition. If Li-ion technology is able to meet thecycle life requirements of a mission while operating at a higher DOD, the resulting mass and volume savings in theenergy storage system could contribute to increased payload capability onboard the spacecraft.

Some practical issues that contribute to the increased use of Li-ion batteries in space are that Li-ion batteries areless expensive than Ni-H2 batteries and that Ni-H2 technology is in danger of becoming obsolete.†,1 Aerospace Li-

* Electrical Engineer, Electrochemistry Branch, Power and Electric Propulsion Division, 21000 Brookpark Road/MS309-1, non-member.

T

3rd International Energy Conversion Engineering Conference15 - 18 August 2005, San Francisco, California

AIAA 2005-5562

This material is declared a work of the U.S. Government and is not subject to copyright protection in the United States.

American Institute of Aeronautics and Astronautics2

ion cell and battery research and development efforts can benefit from ongoing advances in the thriving commercialLi-ion battery market, as well as from common NASA/military drivers for targeted improvements in the technology.Joint development efforts and the spinning-in of commercial technology for aerospace applications allow for certaindevelopment costs to be shared or reduced.2

Prior to routine use, real-time life testing of Li-ion batteries is required to assess multicell battery-levelperformance, establish cycle life, and otherwise validate the chemistry for its flight readiness for LEO and othertypes of aerospace missions. Results reported on in this paper will contribute valuable data to the ongoingassessment of the life performance of Li-ion aerospace battery technology and the low temperature operation of Li-ion batteries, and will help to establish confidence in the technology for its widespread use in manned and unmannedmissions.

II. BackgroundThe 28 V, 25 ampere-hour (Ah)

battery was designed and built for the2001 Mars Surveyor Program Lander,and was therefore designed to operateat the low temperatures that it wouldendure on the surface of Mars‡. Thebattery contains 8 prismatic Li-ioncells connected in series with nocharge balancing electronics. Thecells contain a liquid organicelectrolyte, a mesocarbon microbeads(MCMB) anode and a LiNiCoO2

cathode. Li-ion batteries of thisvintage typically deliver drasticallyreduced capacity at low temperatures.In addition, rate capability can alsosuffer. Due to the low temperaturerequirements of the original mission, aspecial electrolyte formulation wasused in the cells to enhanceconductivity, improve the redoxstability at the electrodes, and providebetter rate capability.7 Manufacturercharacteristics of the battery areshown in Table 0.

The 2001 Mars Surveyor Programwas the first major NASA missionthat baselined Li-ion battery technology. In addition to the battery being reported on in this paper, four other flightbatteries of the same design were built for the program. Following the cancellation of the mission due toprogrammatic issues, a coordinated plan was developed among several organizations to evaluate the performance ofall five batteries under a variety of conditions. In addition to NASA GRC, other participating organizations includethe NASA Jet Propulsion Laboratory (JPL), the Naval Research Laboratory (NRL), and the Air Force ResearchLaboratory (AFRL). The batteries have undergone testing in various LEO, geosynchronous orbit (GEO), andplanetary mission profile regimes. NASA GRC’s role in the test program is to perform low temperature LEO testingat specific orbital parameters. Reports on this work and other testing efforts with cells of the same design fromwhich these batteries were built are available in the literature.3-6 As a result of these and other related test programs,this battery design has been baselined for the 2007 Phoenix Mars Lander scout mission.

† Thomas Miller, NASA Ni-H2 battery expert, personal communications‡ For actual operation on Mars, the battery would have been housed in a warm electronics box passively heated by aradioisotope thermoelectric generator (RTG) to prevent it from being exposed to extremely cold temperatures. Theaverage recorded temperature on Mars is -63 °C.

Table 0. Manufacturer specifications of the Mars Surveyor ProgramLander battery.Cell/BatteryManufacturer

Yardney Technical Products, Inc.

Nameplate Capacity 25 AhRated Capacity 30 Ah at C/5 at 20 °C; 24 Ah at C/5 at -20 °CNominal Voltage 28.8 VOperating temperature -20 to 40 °CCell type PrismaticNumber of Cells 8Anode Mesocarbon microbeads (MCMB)Cathode LiNiCoO2

Electrolyte 1.0 M LiPF6 EC+DMC+DEC (1:1:1)

Design Mission

Storage: Up to two years, one year of which duringcruise to Mars at 0 °C to 30 °CEntry, descent and landing (EDL) on Mars: Several2C pulses, 100 milliseconds in duration at 0 °COperation on Mars: Minimum of 90 cycles, typicaldischarge of C/5 at a maximum of 50% DOD at -20°C to 40 °C

Special OperatingConsiderations

Can operate under any orientation and in zero-gravity

Source: Yardney Technical Products, JPL4

American Institute of Aeronautics and Astronautics3

III. StorageThe battery was stored at open circuit at 15 degrees Celsius (°C) and 50% state-of-charge (SOC) (roughly 3.6

V/cell) for almost two years prior to beginning characterization in preparation for LEO testing (referred to from hereon as pre-LEO characterization). The cells displayed extremely low self-discharge while in storage (less than 0.5%per month). The OCVs of the cells after storage is shown in the second column of Table 0. Due to the excellent celland battery voltages after storage and the fact that the battery delivered greater than nameplate capacity during itspre-LEO characterization, no wake-up procedures were required after the two year storage period.

IV. ExperimentalTesting involves pre-LEO characterization,

followed by life-testing under LEO conditions at40% DOD and 0 °C. Periodic characterizationtests are run at 1000 cycle intervals, and cellbalancing is performed as required. Parameters foreach of these test procedures are summarized inTable 0.

Pre-LEO characterization was performed afterthe battery was removed from storage. LEO testingwas initiated after pre-LEO characterization wascompleted.

Cell balancing is performed during a regularlyscheduled 1000 cycle characterization if the cellvoltage dispersion exceeds 100 millivolts (mV)on charge or 80 mV on discharge. The cellvoltage dispersion on charge exceeded 100 mV atcycle 6510, so cell balancing was performed atthe 7000 cycle characterization interval.

Since cell balancing was performed after7000 LEO cycles, there are noticeablediscontinuities in the data before and after cellbalancing. These two data series will beaddressed separately throughout this paper.

Testing is performed using an ArbinInstruments Model BT2000 battery cycler.Current, battery voltage, cell voltages, and celltemperatures (outside case) are continuallymonitored. The battery is tested under an inertatmosphere within an environmental chamberthat is held at a constant temperature of 0 °Cduring LEO testing and periodic characterization.During pre-LEO characterization, the chamber was set to the temperature required for each test and the battery wassoaked at that temperature for at least 24 hours before testing. Cell balancing is conducted at room temperaturebecause portions of this procedure must be done manually. After cell balancing, the battery is soaked at 0 °C for atleast 24 hours before characterization and testing continues. Figure 0 shows a picture of the battery in the testchamber. Cell 1 is located at the bottom of the stack. Cell 8 is located at the top of the stack. Thermocouples arelocated on the side of each cell near the top of the case (cells are oriented sideways in the battery to form the stack).A 30 ampere (A) fuse is connected in-line with the positive current cable that connects the battery cycler to thebattery to prevent accidental overcurrents.

V. Pre-LEO CharacterizationA pre-LEO characterization procedure was performed when the battery was removed from storage. A C/5 (5 A)

discharge was initiated to measure the residual capacity in the battery. During this discharge, an equipmentmalfunction caused Cell 3 to overdischarge to 1.55 V, well below the 2.5 V vendor-recommended safe operatinglimit of the cell. The test was terminated after 30 seconds, after which the open circuit voltage (OCV) of Cell 3 wasmonitored, but it did not recover on its own. The voltage on the cell was recovered by charging it at 1 A constant

Table 0. Cell OCVs before Cycling.Cell OCV after

Storage (V)Pre-LEO OCV (after

Equipment Malfunction) (V)1 3.55 3.502 3.55 3.473 3.34 1.554 3.54 3.355 3.48 3.256 3.46 3.347 3.42 3.378 3.53 3.51

Figure 0. Mars Surveyor Program Lander Battery testset-up.

American Institute of Aeronautics and Astronautics4

current while it was electrically disconnected from the seven other cells in the battery. Cell balancing was thenrequired to restore its voltage to the same nominal voltage of the other cells. After the cells were balanced, it wasobserved that their OCVs were within 20 mV of each other after a one hour rest period. Pre-LEO characterization

procedures were then resumed. There has been no indication of any reduced performance on Cell 3 duringsubsequent cycling. Table 0 shows the OCVs of each cell after storage, compared to their voltages after theequipment malfunction.

Before commencing LEO testing, the battery underwent several characterization tests. To determine the effectsof low temperature operation on the capacity of the battery, 100% DOD baseline capacity tests were performed at 20°C and 0 °C using identical test conditions. At each temperature, a constant current charge was performed at C/5 (5A) until the battery voltage reached 32 V or until any of the cell voltages reached 4.05 V. The battery voltage wasthen held constant and charging continued until the current tapered to C/50 (0.5 A). The battery was then discharged

Table 0: Battery Test Conditions and ProceduresTest Conditions Test Procedure

Pre-LEO

• Measure OCV.• Discharge at a constant current of C/5 (5 A) to 24 V or until any cell reaches 3 V inorder to measure residual capacity.• Perform baseline capacity measurements for several conditions and temperatures. If testtemperature will change, soak battery for at least 24 hours at new test temperature beforetest.

1. At 20 °C, charge at a constant current of C/5 (5 A) to 32 V or until any cellreaches 4.05 V; hold voltage constant and continue to charge until current tapersto C/50 (0.5 A); discharge at C/5 (5 A) to 24.0 V or until any cell reaches 2.5 V.

2. At 23 °C, charge at a constant current of C/2 (12.5 A) to 32.8 V or until any cellreaches 4.05 V; hold voltage constant and continue to charge until current tapersto C/125 (0.2 A); discharge at C/2 (12.5 A) to 24.0 V or until any cell reaches 3 V.

3. Repeat step 1 at 0 °C. 4. 100% DOD Capacity Test at 0 °C (as given below in this table)

LEO

• Charge at a constant current of C/2 (12.5 A) to 32 V or until any cell reaches 4.05 V.• Hold voltage constant and continue to charge for the remainder of the 55 minute chargeperiod.• Discharge at 0.7C (17.14 A) for 35 minutes (to 40% DOD).• Failure is defined when EODV is 24 V or 2.5 V on any cell.

100%DOD

CapacityTest

• Continue LEO discharge at 0.7C (17.14 A) to 24.0 V or until any cell reaches 2.5 V.• Charge at a constant current of C/2 (12.5 A) to 32 V or until any cell reaches 4.05 V.• Hold voltage constant and continue to charge until current tapers to C/50 (0.5 A).• Discharge at 0.7C (17.14 A) to 24.0 V or until any cell reaches 2.5 V.

1000

Cyc

leC

hara

cter

izat

ion

DCImpedance

1. Charge at a constant current of C/5 (5 A) to 32.4 V.2. Rest at OCV for 2 hours.3. Discharge at a constant current of 1C (25 A) for 10 seconds.4. Rest at OCV for 2 hours.5. Discharge at a constant current of C/10 (2.5 A) for 2 hours (remove 5 Ah).6. Repeat steps 2-5 four times (at each of 80, 60, 40 and 20% SOCs).

Cell Balancing

1. Discharge at a constant current of C/5 (5 A) to 24 V or until any cell reaches 3 V.2. Discharge at a constant current of C/50 (0.5 A) until any cell reaches 2.75 V. Allowvoltages to recover at open circuit.3. Individually drain discharge remaining 7 cells to 2.75 V using a 1 ohm (Ω) resistor.Allow voltages to recover at open circuit.4. Charge at a constant current of C/5 (5 A) to 32.8 V or until any cell reaches 4.1 V.5. Hold voltage constant and continue to charge until current tapers to C/125 (0.2 A).6. Repeat steps 1-3, then perform a drain discharge of individual cells to 2.75 V using 1.5Ω resistor.7. Repeat 100% DOD Capacity test.

American Institute of Aeronautics and Astronautics5

at C/5 (5 A) until its voltage reached 24.0 V or any of the cell voltages reached 2.5 V. These and other baselinecapacity test conditions are described in Table 0 under Pre-LEO characterization. The results of the C/5, 20 °C and0 °C are shown in Figure 0. 30.6 Ah were delivered at 20 °C, versus 27.4 Ah at 0 °C. The battery was charged at thedischarge temperature for each case. Because the battery was charged at the colder temperature for the 0 °C case,there was less capacity available in the battery prior to its discharge at that temperature than for the 20 °C case.

VI. LEO CyclingA LEO orbit period

varies slightly dependingupon the orbit altitude,from 86.48 to 105.12minutes at altitudes of100 and 1000 km,respectively.8 A nominal90 minute orbit periodwas selected for thistesting. During 55minutes of the orbit, thespacecraft is exposed tosunlight. Solar panelsprovide primary power tothe spacecraft andprovide energy to chargethe batteries.

For this testing, thebattery is charged usinga constant-current/constant-voltageprofile. The battery ischarged at C/2 (12.5 A)to 32 V or until the firstcell reaches 4.05 V.Next, the battery voltageis held constant whilethe current is allowed totaper for the remainderof the 55 minute chargeperiod. This two-stepcharge regime allowsfor additional capacityto be added to thebattery once it hasachieved its uppervoltage cut-off limit,without increasing itsvoltage beyond theacceptable limits.

For the remaining 35 minutes of the orbit, the sun is eclipsed by the Earth (with respect to the satellite), whichresults in the satellite being in shadow. During this time, the battery discharges to provide primary power to thespacecraft. A discharge rate of 17.14 A (about 0.7C) was required to discharge 40% of the capacity of the battery(10 Ah) during the 35 minute eclipse portion of the orbit. The 40% DOD was chosen for this testing in order to testthe outer envelope of possible DODs for Li-ion batteries for LEO applications. The test will continue until thebattery fails. Battery failure is defined as the point when the battery end-of-discharge-voltage (EODV) reaches 24 Vor when any cell reaches 2.5 V during a LEO discharge. To date, the battery has accumulated over 9000 cycles,

20

22

24

26

28

30

32

34

0 5 10 15 20 25 30Discharge Capacity, Ah

Vo

ltag

e,V

0 degrees C20 degrees C

Discharged at C/5 (5 A) from 32 to 24 VCapacity delivered at 20 °C = 30.6 AhCapacity delivered at 0 °C = 27.4 Ah

Figure 0. Pre-LEO capacity at 0 °C and 20 °C

-25

-15

-5

5

15

25

35

0 1 2 3 4 5 6 7 8 9 10 11 12 13Time, hours

Cu

rren

t,A

;V

olt

age,

V CurrentVoltage

Figure 1. LEO cycling on the Mars Surveyor Program Battery

American Institute of Aeronautics and Astronautics6

equivalent to more than 18 months in LEO.Figure 1 shows representative current and batteryvoltage over several cycles near the beginning ofthe LEO cycling.

Figure 1 shows the individual cell voltagesversus time during LEO cycling near thebeginning of life, just before cell balancing, andjust after cell balancing. At the beginning of life,the cell voltages are operating tightly together andaverage 3.5 V at the end of discharge. Before cellbalancing, at almost 7000 cycles, the cell voltagesare separating, with cells 6 and 7 never achievingfull charge before the battery reaches its cut-offvoltage. The EODV of the cells averages 3.42 Vat 6997 cycles and 3.44 V at 7100 cycles. Theincrease in the average EODV of the cells is dueto cell balancing.

VII. Periodic CharacterizationPeriodic characterization tests are conducted

to assess the overall state of health of the battery.Capacity and impedance checks are performedevery 1000 cycles.

A. 100 % DOD Capacity TestTo measure battery capacity, following 1000

LEO cycles, the battery is allowed to continuedischarging at its LEO rate until the batteryvoltage reaches 24 V or any cell voltage reaches2.5 V. This process allows us to determine thecapacity remaining in the battery. Full capacity isthen measured by conducting a 100% DODcapacity test. The battery is charged at a constantcurrent of C/2 (12.5 A) to 32 V or until any cellreaches 4.05 V. The battery voltage is then heldconstant until the current tapers to C/50 (0.5 A).The battery is then discharged at the LEO rate to24.0 V or until any cell reaches 2.5 V.

The full capacity that the battery can deliverdecreases as the battery cycles. Figure 1 showsthe 100% DOD capacity at each 1000 cycleinterval. Full capacity delivered decreased fromthe 25.7 Ahr delivered during the initial capacitytest to 24.1 Ahr at 7000 cycles. Because the cellvoltage dispersion was reduced significantly afterthe cells were balanced, the capacity deliveredimmediately after the cells were balancedincreased by 0.5 Ahr to 24.6 Ah, equivalent to thecapacity delivered at 3000 cycles. Capacitiesmeasured at all intervals are given in Table 1.Further details on cell balancing are discussedlater in the section on Cell Balancing.

3.35

3.45

3.55

3.65

3.75

3.85

3.95

4.05

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

Time, hours

Vo

ltag

e,V

Cell 1Cell 2Cell 3Cell 4Cell 5Cell 6Cell 7Cell 8

a)

3.35

3.45

3.55

3.65

3.75

3.85

3.95

4.05

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

Time, hours

Vo

ltag

e,V

Cell 1Cell 2Cell 3Cell 4Cell 5Cell 6Cell 7Cell 8

b)

3.35

3.45

3.55

3.65

3.75

3.85

3.95

4.05

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

Time, hours

Vo

ltag

e,V

Cell 1Cell 2Cell 3Cell 4Cell 5Cell 6Cell 7Cell 8

c)

Figure 1. Cell Voltages during LEO: a) Cycles 10-13; b)Cycles 6995-6997; and c) Cycles 7100-7102.

American Institute of Aeronautics and Astronautics7

B. Current-Interrupt Impedance TestCurrent-interrupt impedance tests are

performed in which the DC impedance ismeasured at different SOCs. The batteryimpedance at each SOC increases as the batterycycles, indicating that electrochemical aging isoccurring in the battery. Figure 2 shows the DCbattery impedance at 1000 cycle intervals to6000 cycles.

VIII. Cell BalancingCell balancing results in individual cell

voltages being brought tightly together so thatthey operate at similar SOCs. This procedureprovides for better overall uniform health of thebattery system. As discussed previously, beforeinitializing this life test, the eight cells wereuniformly balanced. As cycling continues, thecells tend to drift apart. If allowed to continuecycling in this manner, an individual cell mayreach its cut-off voltage before adequatecapacity is delivered to the load, causingbattery failure. A cell balancing procedure canbe performed to bring the cell voltages closertogether so the battery can deliver the requiredcapacity over a longer period of time.

Although the battery manufacturer specifiedthat cell balancing should occur when cellvoltage dispersion is 250 mV, we chose moreconservative values of 80 mV on discharge or100 mV on charge for this testing. Under theseconditions, cell balancing was required after7000 cycles, when the cell voltage separationon charge exceeded 100 mV. As shown inFigure 3, once the cells were balanced, the cellvoltage dispersion on both charge anddischarge was reduced from 102 mV to 39 mVand from 39 mV to 32 mV, respectively.

In actual orbit conditions, the cells should ideally be balancedas seldom as possible, since payload operations may be impactedduring battery maintenance. However, depending upon themission, performing this type of maintenance on orbit may not bean option. Other options to maintain cell balance are to allow thebattery to charge to a higher EOCV at predetermined conditions,i.e. when the battery EODV reaches a certain value, or to integratecharge balancing electronics with the battery that wouldcontinually balance the cells with each cycle.9,10 This particularbattery was not specifically designed to operate under LEOconditions, so it was not optimized to deliver thousands of cycleswithout the need for cell balancing.

Cell balancing was performed after 14 months of cycling. Fora 5 year mission, assuming a linear degradation, this indicates thatsuch maintenance would be required a minimum of 3-4 times overthe life of the mission. The time between balancing could beextended, depending upon whether or not more cell voltage

24

25

26

27

28

29

30

31

32

0 2 4 6 8 10 12 14 16 18 20 22 24 26

Discharge Capacity, Ah

Vo

ltag

e,V

Pre-LEOAfter 1000 CyclesAfter 2000 CyclesAfter 3000 CyclesAfter 4000 CyclesAfter 5000 CyclesAfter 6000 CyclesAfter 7000 Cycles

a)

24

25

26

27

28

29

30

31

32

0 2 4 6 8 10 12 14 16 18 20 22 24 26

Discharge Capacity, Ah

Vo

ltag

e,V

7000 Cycles

8000 Cycles

9000 Cycles

b)

Figure 1: Capacity delivered at 1000 cycle intervals after a fullcharge and a discharge to 100% DOD: a) before cellbalancing; and b) after cell balancing.

Table 1: 100% DOD Capacity ResultsPre-Cell Balancing Post-Cell Balancing

CycleInterval

Capacity(Ah)

CycleInterval

Capacity(Ah)

Pre-LEO 25.7 7000 24.6

1000 25.0 8000 24.5

2000 24.9 9000 24.2

3000 24.6

4000 24.5

5000 24.5

6000 24.2

7000 24.1

American Institute of Aeronautics and Astronautics8

dispersion would be tolerated. If theend-of-charge voltages of the cells wereallowed to separate more, this type ofmaintenance could be required fewertimes during the life of a 5 year LEOmission. This would be contingent uponwhether the battery and each of the cellsare able to continue to perform abovetheir recommended cut-off voltages andwhether the battery is still able todeliver the required capacity over asmany cycles.

IX. Battery PerformanceSummary

A. EfficiencyLi-ion battery chemistries typically

display excellent efficiencies. For thisbattery, coulombic efficiency isconsistently 100% throughout LEOcycling. The test battery demonstratesenergy efficiency between 90 and 93%during LEO cycling.

Figure 4 shows energy efficiency versuscycle number during LEO cycling. Theenergy efficiency increase between cycles4500 and cycles 5300, is attributed toequipment calibration errors. (Straymeasurement values are the result of testinterruptions for periodic characterizationor utility outages.)

B. EODV PerformanceDuring cycling, the battery loses

capacity and its impedance increases,resulting in a lower EODV with increasingnumber of cycles, as shown in Figure 4. Alinear trend projection of the performanceover the first 7000 cycles would predict abatter EODV of approximately 26.3 Voltsat 30,000 cycles (if the cells had not beenbalanced). This is well above the 24 Vminimum battery voltage level. Theelectrochemical processes that contribute tobattery aging have not been accuratelymodeled in this simple projection, so atsome point during the battery cycle life,there may be some discontinuities in theEODV curve that will not follow thecurrent trend. (Stray measurement valuesare the result of test interruptions forperiodic characterization or utility outages).

0

0.02

0.04

0.06

0.08

0.1

0.12

0 20 40 60 80 100State-of-Charge, %

Bat

tery

Imp

edan

ce,o

hm

s

Pre-LEO1000 Cycles2000 Cycles3000 Cycles4000 Cycles5000 Cycles6000 Cycles

Figure 2. Battery impedance at 1000 cycle intervals.

0

20

40

60

80

100

120

0 1000 2000 3000 4000 5000 6000 7000 8000 9000Cycle Number

Cel

lVo

ltag

eD

isp

ersi

on

,mV EOCV dispersion

EODV dispersion

Before Balancing After Balancing

Figure 3. Cell voltage dispersion on charge and discharge.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

0 1000 2000 3000 4000 5000 6000 7000 8000 9000

Cycle Number

En

erg

yE

ffic

ien

cy

Figure 4. Energy efficiency during LEO cycling.

American Institute of Aeronautics and Astronautics9

Some of the capacity loss appears to bereversible, and can be recovered throughreconditioning the battery. As seen in Figure 4, periodic characterization procedures appear tohave a “reconditioning effect” on cell balancing,as evidenced by the fact that the cell voltagedispersion on charge and discharge is reducedafter capacity and impedance testing have beenperformed. This reconditioning leads to a higherbattery EODV during LEO cycling. After cycle7000, the EODV was increased and the cellvoltage dispersion was decreased due to acombination of this reconditioning and the cellbalancing procedure.

C. Full Capacity after LEO CyclingAt 1000 cycle intervals, the battery is fully

discharged to 24 V at the LEO rate to measure thefull capacity in the battery during LEO cycling. Asseen in Figure 4, we observe that the batterydelivers fewer Ah as the number of cyclesincreases. Since the charge voltage cut-off limit iscontrolled on the battery-level (the sum of the seriescell voltages), certain cells do not reach their fullvoltage during the charge period. On discharge,these same cells will determine the capacitydelivered by the battery. The cell voltagedispersion tends to get larger over several thousandsof cycles, so the battery delivers less and lesscapacity with the increase in cell voltage dispersion.When the cells are brought closer together by cellbalancing, the battery can deliver more capacityunder the same test conditions. At 9000 cycles, the21.4 Ah delivered after LEO cycling is identical tothat delivered at 6000 cycles.

A lengthy taper charge period can also result inbringing cell voltages closer together, so morecapacity is delivered from a 100% DOD dischargethan from a full discharge after LEO cycling. Forexample, at 4000 cycles, the 100% DOD capacitydelivered was 24.5 Ah (see Table 1) whereas 21.9Ah was delivered from the full discharge after 4000LEO cycles (see Figure 4), a 2.6 Ah difference.Similar results are observed when comparing thecapacity delivered at 6000 and 9000 cycles.

D. Specific EnergySpecific energy for the battery is calculated

using the mass of the entire flight battery assembly,including cell stack, battery wiring, deck plate, andconnectors. In aerospace-design batteries, masspackaging factors can vary widely depending uponthe requirements and constraints of individualmissions, even for batteries of the same chemistry.As a result, battery-level specific energy can also vary significantly. In order to meet mission requirements, thelander battery was fortified with robust construction designed to tolerate the impact it would sustain during its

20

22

24

26

28

30

32

0 1000 2000 3000 4000 5000 6000 7000 8000 9000

Cycle Number

Vo

ltag

e,V

Figure 4. End-of-Discharge Voltage during LEO cycling.

0

0.02

0.04

0.06

0.08

0.1

0.12

1000 3000 5000 7000 9000

Cycle Number

Vo

ltag

eD

isp

ersi

on

,VBefore Full Discharge

After Full Discharge

a)

0

0.02

0.04

0.06

0.08

0.1

0.12

1000 3000 5000 7000 9000

Cycle Number

Vo

ltag

eD

isp

ersi

on

,V

Before Full Discharge

After Full Discharge

b)

Figure 4. Effect of characterization procedures on cell:a) EOCV dispersion; and b) EODV dispersion.

American Institute of Aeronautics and Astronautics10

landing on the surface of Mars. The specific energy values are indicative of this sturdy construction. The specificenergy averages 76 watt-hours per kilogram (Wh/kg) at 100% DOD and 32 Wh/kg during LEO cycling.

X. Cell Level PerformanceDuring LEO cycling, Cells 6 and 7

consistently have the lowest EODVs. The cellsare near the top of the stack and havetemperatures in the middle range of thetemperature distribution. The cycle performanceshown in Figure 1 shows cells 6 and 7 divergingfrom the rest of the battery. At ~cycle 7000,these cells are only at ~3.95 volts when the highrate charge is terminated based on the batterycutoff of 32.0 volts, while the other cells in thebattery all have voltages above 4.0 volts. Thelower EOCV indicates that these cells are at alower SOC than the rest of the battery. This isconsistent with the EODV behavior as well,where Cells 6 and 7 are approximately 20 mVlower that the other cells in the battery.

Figure 4 shows the OCV recovery for eachcell after a full discharge at the LEO dischargerate following the LEO charge at cycle 6000, andthe OCV recovery following the 100% DODcapacity check that followed. In both cases, Cells6 and 7 have the lowest EODV.

Cell impedance as a function of state of chargefor cycle 6000 is shown in Figure 5. Cell 7 has thelowest impedance at the highest states of charge.The impedance of Cell 6 is generally in the middlerange at most sates of charge. Figure 5 shows thechange in impedance between the pre LEO-cycling measurement and the measurement at6000 cycles as a function of SOC.

In future work, additional analysis will beperformed to further understand the relationshipsbetween the individual cell operating voltages, thecell impedances, the thermal characteristics ofeach cell relative to the other cells in the stack, andthe battery level interactions.

XI. ConclusionThe 2001 Mars Surveyor Program Lander

battery has achieved over 9000 LEO cycles at 40%DOD. The battery is performing well andcontinues to deliver required capacity at 0 °Cduring the LEO testing. If the currentperformance trends continue, it is projected thatthe battery can attain thousands of additional LEOcycles before it fails. The eight cells are operatinguniformly together and it has been demonstratedthat several thousand cycles can be achievedbefore cell voltage dispersion reaches 100 mV oncharge. By periodically rebalancing the cells,battery life can be extended and can possibly

0

2

4

6

8

10

12

14

16

18

20

22

4000 6000 9000

Cycle Number

Cap

acit

y,A

h

Capacity Discharged during LEO Residual Capacity

Figure 4. Capacity remaining in battery after LEO discharge.

3.40

3.45

3.50

3.55

3.60

3.65

0.0 0.5 1.0 1.5 2.0

Time, hour

Vo

ltag

e,V Cell 1 Cell 2

Cell 3 Cell 4

Cell 5 Cell 6

Cell 7 Cell 8

a)

3.20

3.25

3.30

3.35

3.40

3.45

0.0 0.5 1.0 1.5 2.0

Time, hours

Vo

ltag

e,V

Cell 1 Cell 2

Cell 3 Cell 4

Cell 5 Cell 6

Cell 7 Cell 8

b)

Figure 4: OCV recovery versus time at 6000 cycles for2 hours after full discharge at LEO rate: a) after LEOcycling; and b) after 100% DOD capacity check

American Institute of Aeronautics and Astronautics11

extend the life of the mission. Future analysis will focus on identifying the causes of cell imbalance and methodsfor prolonging battery life.

AcknowledgmentsThis work was performed in-house at the National Aeronautics and Space Administration (NASA) Glenn

Research Center (GRC). Funding for this work is provided by the legacy Energetics Program under the ExplorationSystems Mission Directorate of NASA. The author would like to acknowledge Michelle Manzo, Thomas Miller,Russel Gemeiner and the Lithium-ion Group at NASA GRC, Marshall Smart of NASA Jet Propulsion Laboratory,and Rob Gitzendanner at Yardney Technical Products for their technical insight and collaboration during thisproject. The author would also like to acknowledge Daniel Hauser, NASA summer intern and student at OhioUniversity, and Keith Johnson, NASA detailee and student at Cleveland State University, for assisting withcalculations, plotting of the data, and performing background research.

0

0.0005

0.001

0.0015

0.002

0.0025

0 20 40 60 80 100 120State-of-Charge, %

Ch

ang

ein

Imp

edan

ce,o

hm

s

Cell 1

Cell 2

Cell 3

Cell 4

Cell 5

Cell 6

Cell 7

Cell 8

Figure 5: Change in cell impedances from pre-LEOcharacterization to 6000 cycles

0

0.002

0.004

0.006

0.008

0.01

0.012

0.014

0 20 40 60 80 100 120

State-of-Charge, %

Imp

edan

ce,O

hm

sCell 1 Cell 2

Cell 3 Cell 4

Cell 5 Cell 6

Cell 7 Cell 8

Figure 5. Cell impedances at 6000 cycles at different SOCs.

American Institute of Aeronautics and Astronautics12

References1Dunlop, J.D., Brill, J.N., and Erisman, R., “Nickel-Hydrogen Batteries”, Handbook of Batteries 3rd ed., edited by D. Linden

and T.B. Reddy, McGraw-Hill Companies, Inc., New York, 2002, pp. 32.1-32.32.

2Marsh, R.A., Vukson, S., Surampudi, S., Ratnakumar, B.V., Smart, M.C., Manzo, M., and Dalton, P.J., “Lithium-IonBatteries for Aerospace Applications”, Journal of Power Sources, vol. 97-98, 2001, pp. 25-27.

3Reid, C., “Low Temperature Life-Cycle Testing of a Lithium-Ion Battery for Low-Earth-Orbiting Spacecraft”, 2004 NASAAerospace Battery Workshop, November 16-18, 2004.

4Smart, M.C., Ratnakumar, B.V., Whitcanack, L., Surampudi, S., Lowry, L., Gitzendanner, R., et al., “PerformanceCharacteristics of Yardney Lithium-Ion Cells for the Mars 2001 Lander Application”, AIAA-2000-2914, 2000.

5Smart, M.C., Ratnakumar, B.V., Whitcanack, L., Surampudi, S., Byers, J., and Marsh, R., “Performance Characteristics ofLithium-Ion Cells for NASA’s Mars 2001 Lander Application”, IEEE AES Systems Magazine, November, 1999, pp. 36-42.

6Smart, M.C., Ratnakumar, B.V., Whitcanack, L.D., Chin, K.B., Surampudi, S., Gitzendanner, R., et al., “Lithium-IonBatteries for Aerospace”, IEEE AESS Systems Magazine, January 2004, pp. 18-25.

7Salomon, M., Lin, H., Plichta, E.J., and Hendrickson, M., “Temperature Effects on Li-Ion Cell Performance”, Advances inLithium-Ion Batteries, edited by W.A. van Schalkwijk, and B. Scrosati, Kluwer Academic/Plenum Publishers, New York, 2002,pp. 309-344.

8Wertz, J.R., “Orbit and Constellation Design,” Space Mission Analysis and Design 2nd ed., edited by W. J. Larson, and J.R.Wertz, Microcosm, Inc and Kluwer Academic Publishers, Torrence, CA and Dordrecht, The Netherlands, 1992, pp. 165, 179,end pages.

9Reid, C., Hand, E., Button, R., Manzo, M., McKissock, B., Miller, T., Gemeiner, R., and Bennett, W., “Lithium-Ion ChargeControl Circuit Unit”, 2004 NASA Aerospace Battery Workshop, November 16-18, 2004.

10Isaacson, M.J., Hollandsworth, R.P., Giampaoli, P.J., Linkowsky, F.A., Salim, A., and Teoflio, V.L., “Advanced LithiumIon Battery Charger”, The Fifteenth Annual Battery Conference on Applications and Advances, January 11-14, 2000, pp. 193 –198.