materials and manufacturing - nasa€¦ · materials and manufacturing ... in addition, there were...

Materials andManufacturing

IntroductionGail Chapline

Nondestructive Testing InnovationsWillard Castner

Patricia HowellJames Walker

Friction Stir Welding Advancements

Robert DingJim Butler

Characterization of Materials in the Hydrogen EnvironmentJon Frandsen

Jonathan BurkholderGregory Swanson

Space Environment: It’s More Than a VacuumLubert Leger

Steven Koontz

Chemical FingerprintingMichael KillpackEnvironmental Assurance

Anne Meinhold

Unprecedented Accomplishments in the Use of Aluminum-Lithium AlloyPreston McGill

Jim ButlerMyron Pessin

Orbiter Payload Bay Door

Lubert LegerIvan Spiker

Engineering Innovations200

To build a spacecraft, we must begin with materials. Sometimes the

material choice is the solution. Other times, the design must

accommodate the limitations of materials properties. The design of the

Space Shuttle systems encountered many material challenges, such as

weight savings, reusability, and operating in the space environment.

NASA also faced manufacturing challenges, such as evolving federal

regulations, the limited production of the systems, and maintaining

flight certification. These constraints drove many innovative materials

solutions. Innovations such as large composite payload bay doors,

nondestructive materials evaluation, the super lightweight tank, and

the understanding of hydrogen effects on materials were pathfinders

used in today’s industry. In addition, there were materials innovations

in engineering testing, flight analysis, and manufacturing processes.

In many areas, materials innovations overcame launch, landing, and

low-Earth orbit operational challenges as well as environmental

challenges, both in space and on Earth.

201Engineering Innovations

Nondestructive Testing Innovations

Have you ever selected a piece of fruitbased on its appearance or squeezed it for that certain feel? Of course youhave. We all have. In a sense, youperformed a nondestructive test.Actually, we perform nondestructivetesting every day. We visually examineor evaluate the things we use and buy to see whether they are suitable for theirpurpose. In most cases, we give the item just a cursory glance or squeeze;however, in some cases, we give it a conscious and detailed examination.We don’t think of these routineexaminations as nondestructive tests,but they are, and they give us a sense ofwhat nondestructive testing is about.

Nondestructive testing is defined as theinspection or examination of materials,parts, and structures to determine theirintegrity and future usefulness withoutcompromising or affecting theirusefulness. The most fundamentalnondestructive test of all is visual

inspection. In the industrial world,visual examination can be quite formal,with complex visual aids, pass/failcriteria, training requirements, andwritten procedures.

Nondestructive testing depends onincident or input energy that interactswith the material or part being examined.The incident or input energy can bemodified by reflection from interactionwithin or transmission through thematerial or part. The process ofdetection and interpretation of themodified energy is how nondestructivetesting provides knowledge about thematerial or part. Tests range from thesimple detection and interpretation ofreflected visible light by the human eye(visual examination) to the complexelectronic detection and mathematicalreconstruction of through-transmittedx-radiation (computerized axialtomography [CAT] scan). From anondestructive testing perspective, thesimilarity between the simple visualexamination and the complex CAT scanis the input energy (visible light vs.x-rays) and the modified energy

(detected by the human eye vs. anelectronic x-ray detector).

Nondestructive testing is a routine part of a spacecraft’s life cycle. For thereusable shuttle, nondestructive testingbegan during the manufacturing andtest phases and was applied throughoutits service life. NASA performed many such nondestructive tests on theshuttle vehicles and developed mostnondestructive testing innovations inresponse to shuttle problems.

Quantitative NondestructiveTesting of Fatigue Cracks

One of the most significantnondestructive testing innovations was quantifying the flaw sizes thatconventional nondestructive testingmethods could reliably detect. NASAused artificially induced fatigue cracksto make the determination because such flaws were relatively easy to grow and control, hard to detect, andtended to bound the population of flaws of interest. The need to quantifythe reliably detectable crack sizes was

Two examples of the most basic nondestructive testing:Left, a gardener checks ripening vegetables. Right, Astronaut Eileen Collins, STS-114 (2005) mission commander, looks closely at a reinforcedcarbon-carbon panel on one of the wings of the Space Shuttle Atlantis in the Orbiter Processing Facility at Kennedy Space Center (KSC). Collins andthe other crew members were at KSC to take part in hands-on equipment and Orbiter familiarization.


mandated by a fracture control interestin having confidence in the startingcrack size that could be used in fracture and life calculations. Althoughthere was no innovation of any specific nondestructive testing method,quantifying—in a statistical way—thereliably detectable crack sizes associatedwith the conventional nondestructiveevaluation methods was innovative andled the way to the adoption of similarquantitative nondestructive evaluationpractices in other industries.

The quantification of nondestructivetesting methods is commonly referredto today as probability of detection.The Space Shuttle Program developedsome of the earliest data for thepenetrant, x-ray, ultrasonic, and eddycurrent nondestructive testingmethods—the principal nondestructivetesting methods used to inspect shuttlecomponents during manufacturing.Data showed that inspectors certified toaerospace inspection standards could,on average, perform to a certainprobability of detection level definedas standard nondestructive evaluation.

Beyond standard nondestructiveevaluation, NASA introduced a specialnondestructive evaluation level ofprobability of detection wherein thedetection of cracks smaller than thestandard sizes had to be demonstratedby test. Engineers fabricated fatigue-cracked specimens that were used overmany years to certify and recertify, bytest, the inspectors and theirnondestructive evaluation processes tothe smaller, special nondestructiveevaluation crack size. The size of thefatigue cracks in the specimens wastargeted to be a surface-breakingsemicircular crack 0.127 cm (0.050 in.)long by 0.063 cm (0.025 in.) deep, asize that was significantly smaller thanthe standard nondestructive evaluationcrack size of 0.381 cm (0.150 in.) longby 0.19 cm (0.075 in.) deep.

The special probability of detectionspecimen sets typically consisted of 29 randomly distributed cracks ofapproximately the same size. Bydetecting all 29 cracks, the inspectorand the specific nondestructiveevaluation process were consideredcapable of detecting the crack size to a 90% probability of detection with95% confidence.

Nondestructive Testing ofThermal Protection System Tiles

The development of Thermal ProtectionSystem tiles was one of the most unique and difficult developments ofthe program. Because of this material’s“unknowns,” the tile attachmentscheme, and their extremely fragile

nature, NASA examined a number ofnondestructive testing methods.

Acoustic Emission Monitoring

Late in the development of the shuttleThermal Protection System and just before the first shuttle launch,NASA encountered a major problemwith the attachment of the tiles to theOrbiter’s exterior skin. The bondstrength of the tile system was lowerthan the already-low strength of the tile material, and this was notaccounted for in the design. The lowbond strength was due to stressconcentrations at the tile-to-strainisolation pad bond line interface. A Nomex® felt strain isolation pad was bonded between each tile and the Orbiter skin to minimize the

Quantitative Nondestructive Testing

Fatigue-cracked Panel

X-ray Inspection

Penetrant Inspection

Ultrasonic Inspection

Eddy Current Inspection

Probability of Detection Curve

Pro

bab

ility

of D

etec

tion

(%)

Flaw Size Increasing

100

90

80

70

60

50

40

30

20

10

0

X-ray Inspection

Penetrant Inspection

Ultrasonic Inspection

Eddy Current Inspection

P


lateral strain input to the tile from the aluminum skin. These stressconcentrations led to early andprogressive failures of the tile material at the tile-to-strain isolationpad bond line interface when the tile was loaded.

To determine whether low bondstrengths existed, engineers resorted toproof testing for each tile. This requiredthousands of individual tile proof testsprior to first flight. Space ShuttleColumbia (Space TransportationSystem [STS]-1) was at Kennedy SpaceCenter being readied for first flightwhen NASA decided that proof testingwas necessary. Since proof testing was not necessarily nondestructive andtiles could be damaged by the test,NASA sought a means of monitoringpotential damage; acoustic emissionnondestructive testing was an obviouschoice. The acoustic signatures of a low bond strength tile or a tile damagedduring proof test were determinedthrough laboratory proof testing offull-size tile arrays.

To say that the development andimplementation of acoustic emissionmonitoring during tile proof testing was done on a crash basis would be anunderstatement. The fast pace wasdictated by a program that was alreadybehind schedule, and the tile bondstrength problem threatened significantadditional delay. At the height of theeffort, 18 acoustic emission systemswith fully trained three-person crewswere in operation 24 hours a day, 7 days a week. The effort was thelargest single concentration of acousticemission equipment at a single job site. As often happens with suchproblems, where one solution can beovertaken and replaced by another, a tile densification design fix for thelow-strength bond was found andimplemented prior to first flight, thusobviating the need for continued

acoustic emission monitoring. By thetime the acoustic emission monitoringwas phased out, NASA had performed20,000 acoustic emission monitoredproof tests.

Sonic Velocity Testing

Another early shuttle nondestructivetesting innovation was the use of anultrasonic test technique to ensure that the Thermal Protection System tiles were structurally sound prior to installation. Evaluation of pulse or sonic velocity tests showed a velocity relationship with respect to

both tile density and strength. Thesemeasurements could be used as aquality-control tool to screen tiles forlow density and low strength and couldalso determine the orientation of the tile.

The sonic velocity technique input ashort-duration mechanical impulse intothe tile. A transmitting transducer and areceiving transducer, placed on oppositesides of the tile, measured the pulse’stransit time through the tile. For theLockheed-provided tile material, LI-900 (with bulk density of 144 kg/m3

[9 pounds/ft3]), the average through-the-thickness sonic velocity was on the

Coating

Tile Body

Silicon Rubber(RTV 560)

Silicon Rubber(RTV 560)

Nomex® StrainIsolation Pad

Primer

AluminumSubstrate

T

Acoustic Emission Monitoring of Tiles During Proof Test

Tile Body

AcousticEmissionSensor

To Acoustic EmissionSignal Processing

T

Felted Nomex® Pad Showing Stress Concentration in Tile Bond Line Caused by Needling

Aluminum Substrate

Acoustic EmissionsFrom Local Tile Failure

ProofLoad

Stiff Spot / Stress Concentration

Tile Proof Test

Tile Attachment Scheme


order of 640 m/sec (2,100 ft/sec), andthe through-the-thickness flat-wisetensile strength was on the order of 1.69 kg/cm2 (24 pounds/in2). The LI-900acceptance criterion for sonic velocitywas set at 518 m/sec (1,700 ft/sec),which corresponded to a minimumstrength of 0.91 kg/cm2 (13 pounds/in2).Sonic velocity testing was phased out in the early 1990s.

Post-Columbia AccidentNondestructive Testing of External Tank

A consequence of the Columbia(STS-107) accident in 2003 was thedevelopment of several nondestructiveinnovations, including terahertz imagingand backscatter radiography of ExternalTank foam and thermography of thereinforced carbon-carbon—both onorbit and on the ground—during vehicleturnaround. The loss of foam, reinforcedcarbon-carbon impact damage, andon-orbit inspection of ThermalProtection System damage were allproblems that could be mitigated tosome extent through the application ofnondestructive testing methods.

Nondestructive Testing of ExternalTank Spray-on Foam Insulation

Prior to the Columbia accident, nonondestructive testing methods wereavailable for External Tank foaminspection, although NASA pursueddevelopment efforts from the early1980s until the early 1990s. The foamwas effectively a collection of smallair-filled bubbles with thin polyurethanemembranes, making the foam a thermaland electrical insulator with very highacoustic attenuation. Due to theseproperties, it was not feasible to inspectthe foam with conventional methodssuch as eddy current, ultrasonics, orthermography. In addition, since thefoam was considered nonstructural,problems of delaminations occurringduring foam application and foampopping off (“popcorning”) duringascent were considered manageablethrough process control.

After the Columbia accident, NASAfocused on developing nondestructivetesting methods for finding voids and delaminations in the thick,hand-sprayed foam applications around protuberances and closeout

areas. The loss of foam applied to thelarge areas of the tank was not as muchof concern because the automatedacreage spray-on process was bettercontrolled, making it more unlikely tocome off. In the event it did come off,the pieces would likely be smallbecause acreage foam was relativelythin. NASA’s intense focus resulted inthe development and implementationof two methods for foam inspection—terahertz imaging and backscatterradiography—that represented new andunique application of nondestructiveinspection methods.

Terahertz Imaging

Terahertz imaging is a method thatoperates in the terahertz region of theelectromagnetic spectrum betweenmicrowave frequencies and far-infraredfrequencies. Low-density hydrocarbonmaterials like External Tank foam wererelatively transparent to terahertzradiation. Terahertz imaging used apulser to transmit energy into astructure and a receiver to record theenergy reflected off the substrate orinternal defects. As the signal traveledthrough the structure, its basic wave

Pulse Velocity Measurement Unit

TimeTransmitter

Sound Waves

Receiver

Tile

Tile

Sonic Velocity Testing of Tiles at Kennedy Space Center Thermal Protection System FacilityThe speed of sound through the tile is related to density and strength.


properties were altered by theattenuation of the material and anyinternal defects. An image was made byscanning the pulser/receivercombination over the foam surface anddisplaying the received signal.

Probability of detection studies ofinserted artificial voids showed around90% detection of the larger voids insimple geometries, but less than 90%detection in the more-complicatedgeometries of voids around protrusions.Further refinements showed thatdelaminations were particularly difficultto detect. The detection threshold for a2.54-cm- (1-in.)-diameter laminar defectwas found to be a height of 0.508 cm(0.2 in.), essentially meaningdelaminations could not be detected.The terahertz inspection method wasused for engineering evaluation, andany defects found were dealt with by anengineering review process.

Backscatter Radiography

Backscatter radiography uses aconventional industrial x-ray tube togenerate a collimated beam of x-rays

that is scanned over the test object. Thebackscattering of x-rays results fromthe Compton effect—or scattering—in which absorption of the incident or primary x-rays by the atoms of the

Transmitted Pulse

Transmitter ReceiverTransmitted Pulse

Time

Time

Air GapInsulatingFoam

Metal-Foam-Metal Re�ection

Aluminum Substrate

Foam-Air Re�ection

Air-Metal Re�ection

Air Gap

Foam-Air Re�ection indicates an air gap or delamination.

3

3

3

2

2

2

1

1

1

1

X-ray Tube

Detector

X-rays

InsulatingFoam

BackscatterX-rays

Aluminum Substrate

An irradiated column of foam that has voids produces less backscattered x-rays than a void-free column of foam.

Collimator

CollimatedX-ray Beam

I

Backscatter X-ray Imaging System

Terahertz Imaging SystemThis system uses high-frequency electromagnetic pulses.

Insulating foam covers theExternal Tank.


test material are reradiated at a lowerenergy as secondary x-rays in alldirections. The reradiated orbackscattered x-rays were collected in collimated radiation detectorsmounted around the x-ray source. Voids or defects in the test materialwere imaged in backscatter radiographyin the same manner as they were inconventional through-transmissionradiography. Imaging of voids ordefects depended on less absorbingmaterial and less backscattered x-raysfrom the void.

Since only the backscattered x-rayswere collected, the technique wassingle sided and suited for foaminspection. The foam was well suitedfor backscatter radiography sinceCompton scattering is greater from lowatomic number materials. Thetechnique was more sensitive to nearsurface voids but was unable to detectdelaminations. Like terahertz imaging,backscatter radiography was used forengineering evaluation, and defectsfound were dealt with by anengineering review process.

Nondestructive Testing ofReinforced Carbon-Carbon SystemComponents

A recommendation of the ColumbiaAccident Investigation Board stated:“Develop and implement acomprehensive inspection plan todetermine the structural integrity of allReinforced Carbon-Carbon (RCC)system components. This inspectionplan should take advantage of advancednon-destructive inspection technology.”To comply with this recommendation,NASA investigated advancedinspection technology for inspection ofthe reinforced carbon-carbon leadingedge panels during ground turnaroundsand while on orbit.

Ground Turnaround Thermography

NASA selected infrared flashthermography as the method todetermine the structural integrity of thereinforced carbon-carbon components.Thermography was a fast,noncontacting, one-sided applicationthat was easy to implement in theOrbiter’s servicing environment.

The Thermographic Inspection System was an active infrared flashthermogaphy system. Thermographicinspection examined and recorded the surface temperature transients of the test article after application of a short-duration heat pulse. The rate of heat transfer away from the testarticle surface depended on the thermaldiffusivity of the material and theuniformity and integrity of the testmaterial. Defects in the material wouldretard the heat flow away from thesurface, thus producing surfacetemperature differentials that werereflective of the uniformity of thematerial and its defect content. Adefect-free material would uniformly

transfer heat into the underlyingmaterial, and the surface temperaturewould appear the same over the entiretest surface; however, a delaminationwould prevent or significantly retardheat flow across the gap created by thedelamination, resulting in more-localheat retention and higher surfacetemperature in comparison to thematerial surrounding the delamination.Temperature differences were detectedby the infrared camera, which providedvisual images of the defects. Electronicsignals were processed and enhancedfor easier interpretation. The heat pulsewas provided by flashing xenon lampsin a hooded arrangement that excludedambient light. The infrared camera wastransported along a floor-mounted railsystem in the Orbiter Processing Facilityfor the leading edge panel inspections,allowing full and secure access to all ofthe leading edge surfaces. After thetransport cart was positioned, thecamera was positioned manually via agrid system that allowed the same areasto be compared from flight to flight.

The thermography system wasvalidated on specimens containing flatbottom holes of different diameters and depths. Validation testing confirmedthe ability of the flash thermographysystem to detect the size holes thatneeded to be detected.

After the first Return to Flightmission—STS-114 (2005)—thepostflight thermography inspectiondiscovered a suspicious indication in the joggle area of a panel.Subsequent investigation showed that the indication was a delamination.This discovery set in motion an intensefocus on joggle-area delaminations and their characterization andconsequence. Many months of furthertests, development, and refinement of the thermography methodology

Infrared thermography inspection of the Orbiter nose captured at the instant of the xenon lamp flash. Kennedy Space Center OrbiterProcessing Facility.


determined that critical delaminationswould be detected and sized by flashthermography and provided the basisfor flightworthiness.

On-orbit Thermography

The success of infrared thermographyfor ground-based turnaround inspectionof the wing leading edge panels and theextensive use of thermography duringReturn to Flight impact testing made it the choice for on-orbit inspection ofthe leading edge reinforcedcarbon-carbon material. A thermalgradient through the material must existto detect subsurface reinforcedcarbon-carbon damage with infraredthermography. A series of ground testsdemonstrated that sunlight or solarheating and shadowing could be used togenerate the necessary thermalgradient, which significantly simplifiedthe camera development task.

With the feasibility of on-orbitthermography demonstrated and with the spaceflight limitations onweight and power taken into account,NASA selected a commercialoff-the-shelf microbolometer camerafor modification and development intoa space-qualified infrared camera forinspecting the reinforced carbon-carbonfor impact damage while on orbit.

The extravehicular activity infraredcamera operated successfully on itsthree flights. Two reinforced carbon-carbon test panels with simulateddamage were flown and inspected onSTS-121 (2006). The intentional impactdamage in one panel and the flat bottomholes in the other panel were clearlyimaged. Engineers also performed asimilar on-orbit test on two otherintentionally damaged reinforcedcarbon-carbon test panels during a spacestation extravehicular activity with the

Extravehicularactivity infraredflight camera.

Processedinfrared images

of reinforcedcarbon-carbon

test panels.

Astronaut Thomas Reitermounting pre-damaged

reinforced carbon-carbon test panels on the

International Space Stationduring STS-121 (2006).

On-orbit Thermography


same result of clearly imaging thedamage. The end result of these effortswas a mature nondestructive inspectiontechnique that was transitioned anddemonstrated as an on-orbitnondestructive inspection technique.

Additional NondestructiveTesting

Most nondestructive testinginnovations resulted from problemsthat the shuttle encountered over the years, where nondestructive testingprovided all or part of the solution.Other solutions worth mentioninginclude: ultrasonic extensometermeasurements of critical shuttle bolttensioning; terahertz imaging ofcorrosion under tiles; phased arrayultrasonic testing of the External Tank friction stir welds and the shuttle crawler-transporter shoes;thermographic leak detection of themain engine nozzle; digitalradiography of Columbia debris;surface replication of flow liner cracks;and the on-board wing leading edgehealth monitoring impact system.

In the mid 1990s, NASA pursued the implementation of friction stir welding

technology—a process developed by The Welding Institute of Cambridge, England—

to improve External Tank welds. This effort led to the invention of an auto-adjustable

welding pin tool adopted by the Space Shuttle Program, the Ares Program (NASA-

developed heavy launch vehicles), and industry.

Standard fusion-welding techniques rely on torch-generated heat to melt and join the

metal. Friction stir welding does not melt the metal. Instead, it uses a rotating pin and

“shoulder” to generate friction, stir the metal together, and forge a bond. This process

results in welds with mechanical properties superior to fusion welds.

Standard friction stir welding technology has drawbacks, however; namely, a

non-adjustable pin tool that leaves a “keyhole” at the end of a circular weld and the

inability to automatically adjust the pin length for materials of varying thickness. NASA’s

implementation of friction stir welding for the External Tank resulted in the invention

and patenting of an auto-adjustable pin tool that automatically retracts and extends in

and out of the shoulder. This feature provides the capability to make 360-degree welds

without leaving a keyhole, and to weld varying thicknesses.

During 2002-2003, NASA and the External Tank prime contractor, Lockheed Martin,

implemented auto-adjustable pin tool friction stir welding for liquid hydrogen and liquid

oxygen tank longitudinal welds. Since that time, these friction stir welds have been

virtually defect-free. NASA’s invention was being used to weld Ares upper-stage

cryogenic hardware. It has also been adopted by industry and is being used in the

manufacturing of aerospace and aircraft frames.

Friction Stir Welding AdvancementsNASA invents welding fixture.

Friction stir welding units, featuring auto-adjustable pin tools, welded External Tank barrelsections at NASA’s Michoud Assembly Facility in New Orleans, Louisiana. The units measured8.4 m (27.5 ft) in diameter and approximately 7.6 m (25 ft) tall to accommodate the largestbarrel sections.


Characterization ofMaterials in theHydrogen Environment

From the humid, corrosion-friendlyatmosphere of Kennedy Space Center,to the extreme heat of ascent, to thecold vacuum of space, the SpaceShuttle faced one hostile environmentafter another. One of those harshenvironments—the hydrogenenvironment—existed within theshuttle itself. Liquid hydrogen was the fuel that powered the shuttle’scomplex, powerful, and reusable mainengine. Hydrogen provided the highspecific impulse—the bang per poundof fuel needed to perform the shuttle’sheavy-lifting duties. Hydrogen,however, was also a potential threat to the very metal of the propulsionsystem that used it.

The diffusion of hydrogen atoms into a metal can make it more brittle andprone to cracking—a process calledhydrogen embrittlement. This effect can reduce the toughness of carefullyselected and prepared materials. A concern that exposure to hydrogenmight encourage crack growth waspresent from the beginning of the SpaceShuttle Program, but the rationale forusing hydrogen was compelling.

The Challenge of the HydrogenEnvironment

Hydrogen embrittlement posed morethan a single engineering problem for the Space Shuttle. This was partlybecause hydrogen embrittlement canoccur in three different ways. The most common mode occurs whenhydrogen is absorbed by a material that is relatively unstressed, such as the components of the shuttle’s main

engines before they experienced the extreme loads of liftoff and flight; this is called internal hydrogenembrittlement. Under the rightconditions, internal hydrogenembrittlement has the potential torender materials too weak and brittle to survive high stresses applied later.

Alternatively, embrittlement can affect a material that is immersed inhydrogen while the material is beingstressed and deformed. Thisphenomenon is called hydrogenenvironment embrittlement, which canoccur in pressurized hydrogen storagevessels. These vessels are constantlystressed while in contact withhydrogen. Hydrogen environmentembrittlement can potentially reduceductility over time and enablecracking, or hydrogen may simplyreduce the strength of a vessel until itis too weak to bear its own pressure.

Finally, hydrogen can react chemicallywith elements that are present in ametal, forming inclusions that candegrade the properties of that metal oreven cause blisters on the metal’ssurface. This effect is called hydrogenreaction embrittlement. In the shuttle’smain engine components, the reactionbetween hydrogen and the titaniumalloys occurred to internally formbrittle titanium hydrides, which wasmost likely to occur at locations wherethere were high tensile stresses in thepart. Hydrogen reaction embrittlementcan affect steels when hydrogen atoms combine with the carbon atomsdissolved in the metal. Hydrogenreaction embrittlement can also blistercopper when hydrogen reacts with theinternal oxygen in a solid copper piece,thereby forming steam blisters.

Insights on HydrogenEnvironment Embrittlement

NASA studied the effects of hydrogenembrittlement in the 1960s. In the early1970s, the scope of NASA-sponsoredresearch broadened to include hydrogenenvironment embrittlement effects onfracture and fatigue. Engineersimmersed specimens in hydrogen andperformed a battery of tests. Theyapplied repeated load cycles tospecimens until they fatigued and brokeapart; measured crack growth rates incyclic loading and under a constantstatic load; and tested materials inhigh-heat and high-pressure hydrogenenvironments. Always, results werecompared for each material to itsperformance in room-temperature air.

During the early years of the SpaceShuttle Program, NASA and contractorengineers made a number of keydiscoveries regarding hydrogenenvironment embrittlement. First,cracks were shown to grow faster whenloaded in a hydrogen environment. This finding would have significantimplications for the shuttle design, asfracture assessments of the propulsionsystem would have to account foraccelerated cracking. Second, scientistsobserved that hydrogen environmentembrittlement could result in crackgrowth under a constant static load.This behavior was unusual for metals.Ductile materials such as metals tend to crack in alternating stress fields, not in fixed ones, unless a chemical or an environmental cause is present.Again, the design of the shuttle wouldhave to account for this effect. Finally,hydrogen environment embrittlementwas shown to have more severe effects at higher pressures. Intriguingly,degradation of tensile properties wasfound to be proportional to the squareroot of pressure.


The overall approach to hydrogenenvironment embrittlement researchwas straightforward. As a matter ofcommon practice, NASA characterizedthe strength and fracture behavior of itsalloys. To determine how these alloyswould tolerate hydrogen, engineerssimply adapted their tests to include ahigh-pressure hydrogen environment.After learning that high pressureexacerbates hydrogen environmentembrittlement, they further adapted thetests to include a hydrogen pressure of703 kg/cm2 (10,000 psi). Later in theprogram, materials being considered for use in the main engine were testedat a reduced pressure of 492 kg/cm(7,000 psi) to be more consistent withoperation conditions. The differencebetween room-temperature air materialproperty data and these new results wasa measurable effect of hydrogenenvironment embrittlement. Now thatthese effects could be quantified, thenext step was to safeguard the shuttle.

Making Parts Resistant to Hydrogen EnvironmentEmbrittlement

One way to protect the main enginesfrom hydrogen environmentembrittlement was through materialsselection. NASA chose naturallyresistant materials when possible. Therewere, however, often a multitude ofconflicting demands on these materials:they had to be lightweight, strong,tough, well suited for themanufacturing processes that shapedthem, weldable, and able to bearsignificant temperature swings. Theadditional constraint of imperviousnessto hydrogen environment embrittlementwas not always realistic, so engineers

experimented with coatings and platingprocesses. The concept was to shieldvulnerable metal from any contact withhydrogen. A thin layer of hydrogenenvironment embrittlement-resistantmetal would form a barrier thatseparated at-risk material fromhydrogen fuel.

Engineers concentrated their researchon coatings that had low solubility and low-diffusion rates for hydrogen at room temperature. Testing haddemonstrated that hydrogenenvironment embrittlement is worst at near-room temperature, so NASAselected coatings based on theireffectiveness in that range. The mostefficient barrier to hydrogen, engineersfound, was gold plating; however, thecost of developing gold platingprocesses was a significant factor.Engineers observed that copper plating provided as much protection as gold, as long as a thicker andheavier layer was applied.

Protecting weld surfaces was oftenmore challenging. The weld surfacesexposed to hydrogen fuel during flightwere typically not accessible to platingafter the weld was complete.Overcoming this problem required amore time-consuming and costlyapproach. Engineers developed weldoverlays, processes in which hydrogenenvironment embrittlement-resistantfiller metals were added during a finalwelding pass. These protective fillerssealed over the weld joints and providedthe necessary barrier from hydrogen.NASA used overlays in combinationwith plating of accessible regions toprevent hydrogen environmentembrittlement in engine welds.

These approaches—a combination oftwo or more hydrogen environmentembrittlement prevention methods—were the practical solution for many ofthe embrittlement-vulnerable parts ofthe engines. For example, the mostheavily used alloy in the engines wasInconel® 718, an alloy known to beaffected by hydrogen environmentembrittlement. Engineers identified analternative heat treatment, differentfrom the one typically used, whichlimited embrittlement. But this alonewas insufficient. In the most criticallocations, the alternative heat treatmentwas combined with copper plating andweld overlays.

A unique processing approach was alsoused to prevent embrittlement in theengine’s main combustion chamber.This chamber was made with a highlyconductive copper alloy. Its wallscontained cooling channels thatcirculated cold liquid hydrogen andkept the chamber from melting in theextreme heat of combustion. But thehydrogen-filled channels became prone to hydrogen environmentembrittlement. These liquid hydrogenchannels were made by machining slotsin the copper and then plated withnickel, which closed out the open slotand formed a coolant channel. Thenickel plate cracked in the hydrogenenvironment and reduced the pressurecapability of the channels. Engineersdevised a two-part solution. First, theydeveloped an alternative heat treatmentto optimize nickel’s performance inhydrogen. Next, they coated the nickelwith a layer of copper to isolate it fromthe liquid hydrogen. This two-prongedstrategy worked, and liquid hydrogencould be safely used as the combustionchamber coolant.


Addressing Internal Hydrogen Embrittlement

Whereas hydrogen environmentembrittlement was of great concern atNASA in the 1960s, internal hydrogenembrittlement was largely dismissedeven through the early years of theSpace Shuttle Program. Internalhydrogen embrittlement had never been a significant problem for the typesof materials used in spaceflighthardware. The superalloys andparticular stainless steels selected byNASA were thought to be resistant tointernal hydrogen embrittlement.Engineers thought the face-centered,cubic, close-packed crystal structurewould leave too little room forhydrogen to permeate and diffuse.

Recall that internal hydrogenembrittlement occurs when hydrogen isabsorbed before high operationalstresses. Hydrogen enters into the metaland remains there, making it morebrittle and likely to crack when extremeservice loads are applied later. It is theaccumulation of absorbed hydrogen,rather than the immediate exposure atthe moment of high stress, thatcompromises an internal hydrogenembrittlement-affected material. WhenNASA initially designed the mainengine, engineers accounted forhydrogen absorbed duringmanufacturing. Engineers, however,thought that the materials that wereformed and processed withoutcollecting a significant amount ofhydrogen were not in danger ofabsorbing considerable amounts later.

This notion about internal hydrogenembrittlement was challenged duringthe preparation of an engine failureanalysis document in 1988. The engine

was repeatedly exposed to hydrogen in flight and after flight, at hightemperatures and extreme pressure. The report suggested that in theseexceptional heat and pressure conditionssome engine materials might, in fact,gather small amounts of hydrogen witheach flight. Gradually, over time, thesematerials could accumulate enoughhydrogen to undermine ductility.

Engineers developed a special testregimen to screen materials forhigh-temperature, high-pressurehydrogen accumulation. Test specimenswere “charged” with hydrogen at649°C (1,200°F) and 351.6 kg/cm2

(5,000 psi). They were then quicklycooled and tested for strength andductility under normal conditions.Surprisingly, embrittlement by internal hydrogen embrittlement wasobserved to be as severe as byhydrogen environment embrittlement.As a subsequent string of fatigue testsconfirmed this comparison, NASA had to reevaluate its approach topreventing hydrogen embrittlement.The agency’s focus on hydrogenenvironment embrittlement had been anear-total focus. Now, a new awarenessof internal hydrogen embrittlementwould drive a reexamination.

Fortunately, the process for calculatingdesign properties from test data hadbeen conservative. The margins ofsafety were wide enough to bound thecombined effects of internal hydrogenembrittlement and hydrogenenvironment embrittlement. The wealthof experience gained in studyinghydrogen environment embrittlementand mitigating its effects also worked inNASA’s favor. Some of the samemethodologies could now be applied to

internal hydrogen embrittlement. Forinstance, protective plating wouldoperate on the same principle—thecreation of a barrier between hydrogenand a vulnerable alloy—whetherhydrogen environment embrittlement orinternal hydrogen embrittlement wasthe chief worry. Continued testing of“charged” specimens would allowquantification of internal hydrogenembrittlement damage, just as hydrogenimmersion testing had enabledmeasurement of hydrogen environmentembrittlement effects.

Taking strategies generated to avoidhydrogen environment embrittlementand refitting them to prevent internalhydrogen embrittlement, however, oftenrequired additional analysis. Forexample, from the beginning of theSpace Shuttle Program NASA usedcoatings to separate at-risk metals fromhydrogen. The agency intentionallychose these coatings for theirperformance at near-room temperature,when hydrogen environmentembrittlement is most aggressive. Testsshowed the coatings were less effectivein the high heat that promotes internalhydrogen embrittlement. New researchand experimentation was required toprove that these protective coatingswere adequate—that, although theydidn’t completely prevent the absorptionof hydrogen when temperatures andpressures were extreme, they did reduceit to safe levels.

Special Cases: High-Pressure Fuel Turbopump Housing

NASA encountered a unique hydrogen embrittlement issue duringdevelopment testing of the main engine high-pressure fuel turbopump.


High-Pressure Fuel and Oxidizer Turbopump Turbine Blade Cracks

After observing cracks on polycrystalline turbine blades,

NASA redesigned the blades as single-crystal parts.

When tested in hydrogen, cracks were detected.

Scientists used a Brazilian disc test to create the tensile

and shear stresses that had caused growth. NASA

resolved cracking in the airfoil with changes that

eliminated stress concentrations and smoothed the flow

of molten metal during casting. To assess cracking at

damper contacts, scientists extracted test specimens

from single crystal bars, machined contact pins from the

damper material, and loaded two specimens. This

contact fixture was supported

in a test rig that allowed the

temperature, loads, and load

cycle rate to be varied.

Specimens were pre-charged

with hydrogen, tested at

elevated temperatures, and

cycled at high frequency to

actual operating conditions.

Disc-shaped SpecimensClamped in Place

Contact Pin

Schematic of Test Rig

Clamp

Solid Core

Normal Force

DynamicDisplacement

DynamicDisplacement

Normal Force

First Stage Blade 42 Trailing Edge Root


A leak developed during the test; this leak was traced to cracks in themounting flange of the turbopump’shousing. The housing was made fromembrittlement-prone nickel-chromiumalloy Inconel® 718, and the cracks werefound to originate in small regions ofhighly concentrated stress. So, engineerschanged the material to a more-hydrogen-tolerant alloy, Inconel® 100,and they redesigned the housing toreduce stress concentrations. Thisinitially appeared to solve the problem.Then, cracks were discovered in otherparts of the housing. Structural andthermal analysis could not explain thiscracking. The locations and size of thecracks did not fit with existing fatigueand crack-growth data.

To resolve this inconsistency, engineersconsidered the service conditions of the housing. The operating environmentof the cracked regions was a mixture ofhigh-pressure hydrogen and steam at149°C to 260°C (300°F to 500°F).Generally, hydrogen environmentembrittlement occurs near roomtemperature and would not be asignificant concern at that level of heat;however, because of the unexplainedcracking, a decision was made to testInconel® 100 at elevated temperatures inhydrogen and hydrogen mixed withsteam. Again, the results wereunexpected. Engineers observed apronounced reduction in strength andductility in these environments atelevated temperatures. Crack growthoccurred at highly accelerated rates—as high as two orders of magnitudeabove room-temperature air when thecrack was heavily loaded to 30 ksi √in

—

(33 MPa √m—

) and held for normalengine operating time. Moreover, crackgrowth was driven by both the numberof load cycles and the duration of eachload cycle. Crack growth is typicallysensitive to the number and magnitudeof load cycles but not to the length oftime for each cycle.

Clearly, the combination of thehydrogen and steam mixture and the uncommonly high stressconcentrations was promotinghydrogen environment embrittlementin Inconel® 100 at high temperatures.Resolving this issue required threemodifications. First, detailed changesto the shape of the housing were made,further reducing stress concentrations.Second, gold plating was added toshield the Inconel® 100 from the hothydrogen and steam mixture. Finally, a manufacturing process called “shotpeening” was used to fortify thesurface of the housing against tensilestresses by impacting it with shot,determined to be promoting fracture,and therefore eliminated.

Summary

The material characterization done inthe design phase of the main engine,and the subsequent anomaly resolutionduring its development phase,expanded both the material propertiesdatabase and the understanding ofhydrogen embrittlement. The range of hydrogen embrittlement data hasbeen broadened from essentiallyencompassing only steels to nowincluding superalloys. It was alsoextended from including primarilytensile properties to includingextensive low-cycle fatigue andfracture-mechanics testing inconditions favorable to internalhydrogen embrittlement or hydrogenenvironment embrittlement. Theresultant material properties database,now approaching 50 years of maturity,is valuable not only because thesematerials are still being used, but alsobecause it serves as a foundation forpredicting how other materials willperform under similar conditions—andin the space programs of the future.

Space Environment: It’s More Than a Vacuum

We know that materials behavedifferently in different environments on Earth. For example, aluminum does not change on a pantry shelf foryears yet rapidly corrodes or degradesin salt water.

One would think that such materialdegradation effects would be eliminatedby going to the near-perfect vacuum of space in low-Earth orbit. In fact,many of these effects are eliminated.However, Orbiter systems produced gas,particles, and light when engines,overboard dumps, and other systemsoperated, thereby creating an inducedenvironment in the immediate vicinityof the spacecraft. In addition, movementof the shuttle through the tenuous upper reaches of Earth’s atmosphere(low-Earth orbit) at orbital velocityproduced additional contributions to the induced environment in the form of spacecraft glow and atomic oxygeneffects on certain materials. Theinteractions of spacecraft materials with space environment factors likesolar ultraviolet (UV) light, atomicoxygen, ionizing radiation, andextremes of temperature can actually be detrimental to the life of materialsused in spacecraft systems.

For the Orbiter to perform certainfunctions and serve as a platform forscientific measurements, the effects of natural and Orbiter-inducedenvironments had to be evaluated andcontrolled. Payload sensitivities to theseenvironmental effects varied, dependingon payload characteristics. Earth-basedobservatories and other instruments areaffected by the Earth’s atmosphere interms of producing unwanted lightbackground and other contaminationeffects. Therefore, NASA developed


essential analytical tools forenvironment prediction as well asmeasurement systems for environmentdefinition and performance verification,thus enabling a greater understanding of natural and induced environmenteffects for space exploration.

Induced EnvironmentCharacterization

NASA developed mathematical modelsto assess and predict the inducedenvironment in the Orbiter cargo bayduring the design and developmentphase of the Space Shuttle Program.Models contained the vehicle geometry,vehicle flight attitude, gas and vaporemission source characteristics, andused low-pressure gas transport physicsto calculate local gas densities, columndensities (number of molecular speciesseen along a line of sight), as well ascontaminant deposition effects onfunctional surfaces. Gas transportcalculations were based on low-pressuremolecular flow physics and includedscattering from Orbiter surfaces and thenatural low-Earth orbit environment.

The Induced EnvironmentContamination Monitor measured the induced environment on threemissions—Space TransportationSystem (STS)-2 (1981), STS-3 (1982),and STS-4 (1982)—and was capable of being moved using the ShuttleRobotic Arm to various locations forspecific measurements. Mostmeasurements were made during theon-orbit phase. This measurementpackage was flown on the threemissions to assess shuttle systemperformance. Instruments included ahumidity monitor, an air sampler forgas collection and analysis after return, a cascade impactor forparticulate measurement, passivesamples for optical degradation of

surfaces, quartz-crystal microbalancesfor deposited mass measurement, a camera/photometer pair for particlemeasurement in the field of view, and a mass spectrometer. Additionalflight measurements made on STS-52 (1992) and many payloadsprovided more data.

Before the induced environmentmeasurements could be properlyinterpreted, several on-orbit operationalaspects needed to be understood.Because of the size of the vehicle andits payloads, desorption of adsorbedgases such as water, oxygen, andnitrogen (adsorbed on Earth) took afairly long time, the inducedenvironment on the first day of amission was affected more than on

subsequent days. Shuttle flight attituderequirements could affect the cargo baygaseous environment via solar heatingeffects as well as the gases produced byengine firings. These gases could reachthe payload bay by direct or scatteredflow. Frequently, specific payload orshuttle system attitude or thermalcontrol requirements conflicted with the quiescent induced environmentrequired by some payloads.

With the above operationalcharacteristics, data collected with themonitor and subsequent shuttleoperations showed that, in general, themeasured data either met or were closeto the requirements of sensitivepayloads during quiescent periods. A large qualification to this statement

The Atlantic Ocean southeast of the Bahamas is in the background as Columbia’s Shuttle Robotic Armand end effector grasp a multi-instrument monitor for detecting contaminants. The experiment, calledthe Induced Environment Contaminant Monitor, was flown on STS-4 (1982). The tail of the Orbiter canbe seen below.

had to be made based on a newunderstanding of the interaction of the natural environment with vehiclesurfaces. This interaction resulted insignificantly more light emissions andmaterial surface effects than originallyexpected. Data also identified anadditional problem of recontact ofparticles released from the shuttleduring water dumps with surfaces in thepayload bay. The induced environmentcontrol program instituted for the SpaceShuttle Program marked a giant stepfrom the control of small free-flyinginstrument packages to the control of alarge and complex space vehicle with amixed complement of payloads. Thisapproach helped develop a system withgood performance, defined the vehicleassociated environment, and facilitatedeffective communication between theprogram and users.

The induced environment program also showed that some attachedpayloads were not compatible with the shuttle system and its associated

payloads because of the release ofwater over long periods of time. Other contamination-sensitive payloadssuch as Hubble Space Telescope,however, were not only successfullydelivered to space but were alsorepaired in the payload bay.

Unique Features Made It Possible

The Orbiter was the first crewedvehicle to provide protection ofinstrumentation and sensitive surfacesin the payload bay during ascent and re-entry and allow exposure to the low-Earth orbit environment.Effects were observed without beingmodified by flight heating or grosscontamination. Also, as part of theinduced environment control program,the entire payload bay was examinedimmediately on return. Because ofthese unique aspects, NASA was ableto discover and quantify unexpectedinteractions between the environmentof low-Earth and the vehicle.

Discovery of Effects of Oxygen Atoms

After STS-1 (1981) returned to Earth,researchers visually examined thematerial surfaces in the payload bay for signs of contamination effects. Most surfaces appeared pristine, except for the exterior of the televisioncamera thermal blankets and somepainted surfaces. The outside surface of the blankets consisted of an organic(polyimide) film that, before flight,appeared gold colored and had a glossy finish. After flight, most filmswere altered to a yellow color and nolonger had a glossy finish but, rather,appeared carpet-like under highmagnification. Only the surfaces oforganic materials were affected; bulkproperties remained unchanged.

Patterns on modified surfaces indicateddirectional effects and, surprisingly, the flight-exposed surfaces were foundto have receded rather than havingdeposited contaminants. The patternson the surfaces were related to the

Engineering Innovations 215

a) Scanning electron microscope image of a typical Kapton® polyimide plastic sheet. The various specs and bumps are from the inorganicfiller used in plastic sheet manufacture.

b) Scanning electron microscope image of a typical Kapton® polyimide plastic sheet after exposure to surface bombardment by atomicoxygen in low-Earth orbit. The rough surface is typical of atomic oxygen attack on plastics in low-Earth orbit and is the result of the strongdependence of chemical reaction on atom-surface collision energy. Note how some of the inorganic filler particles are standing onpedestals because they protect the underlying plastic from atomic oxygen attack.

c) Scanning electron microscope image of a microelectron fabrication etching target also flown on STS-46 and exposed to low-Earth orbitatomic oxygen. The highly directional attack of low-Earth orbit atomic oxygen produced a clean, high-resolution removal of the unprotectedplastic around the pattern of protective inorganic surface coatings. High-speed neutral atomic oxygen beams in ground-based productionfacilities may be a useful adjunct to microelectronic production as described in US Patent 5,271,800.

a b c

Atomic Oxygen Effects on Polymers and Plastics in low-Earth Orbit as Seen With the Scanning Electron Microscope; STS-46 (1992)

vehicle velocity vector. Whencombining these data with theatmospheric composition and densities,the material surface recession wascaused by the high-velocity collision of oxygen atoms with forward-facingOrbiter surfaces leading to surfacedegradation by oxidation reactions.Oxygen atoms are a major constituent ofthe natural low-Earth orbit environmentthrough which the shuttle flew at anorbital velocity of nearly 8 km/sec(17,895 mph). The collision energy ofoxygen atoms striking forward-facingshuttle surfaces in low-Earth orbit wasextremely high—on the order of 5electron volts (eV)—100 times greaterthan the energy of atoms in typicallow-pressure laboratory oxygen atomgenerators. The high collision energy ofoxygen atoms in low-Earth orbit playsan important role in surface reactivityand surface recession rates.

Material recession rates are determinedby normalizing the change in samplemass to the number of oxygen atomsreaching the surface over the exposuretime (atoms/cm2, fluence). Atomdensity is obtained from the standardatmospheric density models used byNASA and the Department of Defense.Since oxygen atoms travel muchslower than the Orbiter, they impactedthe surfaces in question only whenfacing toward the vehicle velocityvector and had to be integrated overtime and vehicle orientation. STS-1recession data were approximatebecause they had to be integrated over changing vehicle attitude; hadlimited atom flux, uncontrolled surface temperatures and solar UVexposure; and predicted atom densities.Recession rates determined frommaterial samples exposed during theSTS-5 (1982) mission and InducedEnvironmental Contamination Monitor

flights had the same limitations but supported the STS-1 data.Extrapolation of these preliminaryrecession data to longer-term missionsshowed the potential for significantperformance degradation of criticalhardware, so specific flightexperiments were carried out toquantify the recession characteristicsand rates for materials of interest.

On-orbit Materials Behavior

Fifteen organizations participated in aflight experiment on STS-8 (1983) tounderstand materials behavior in thelow-Earth orbit environment. Theobjective was to control some of theparameters to obtain more-accuraterecession rates. The mission had adedicated exposure to direct atomimpact (payload bay pointing in thevelocity direction) of 41.7 hours at analtitude of 225 km (121 nautical miles)resulting in the largest fluence of theearly missions (3.5 x 1020 atoms/cm2).Temperature control at two set pointswas provided as well as instruments tocontrol UV and exposure to electricallycharged ionospheric plasma species.

The STS-8 experiment providedsignificant insight into low-Earth orbitenvironment interactions withmaterials. Researchers establishedquantitative reaction rates for morethan 50 materials, and were in therange of 2-3 x 10-24 cm3/atom forhydrocarbon-based materials.Perfluorinated organic materials werebasically nonreactive andsilicone-based materials stoppedreacting after formation of a protectivesilicon oxide surface coating. Materialreaction rates, as a first approximation,were found to be independent oftemperature, material morphology, andexposure to solar radiation orelectrically charged ionspheric species.

Researchers also evaluated coatingsthat could be used to protect surfacesfrom interaction with the environment.

Reaction rates were based on atomicoxygen densities determined fromlong-term atmospheric density models,potentially introducing errors inshort-term experiment data. In addition,researchers obtained very little insightinto the reaction mechanism(s).

An additional flight experiment—Evaluation of Oxygen Interaction withMaterials III—addressing both of thesequestions was flown on STS-46 (1992).The primary objective was to producebenchmark atomic oxygen reactivitydata by measuring the atom flux during material surface exposure.Secondary experiment objectivesincluded: characterizing the inducedenvironment near several surfaces;acquiring basic chemistry data relatedto reaction mechanism; determining the effects of temperature, mechanicalstress, atom fluence, and solar UVradiation on material reactivity; and characterizing the induced andcontamination environments in theshuttle payload bay. This experimentwas a team effort involving NASAcenters, US Air Force, NASA SpaceStation Freedom team, AerospaceCorporation, University of Alabama in Huntsville, National Space Agencyof Japan, European Space Agency, andthe Canadian Space Agency.

STS-46 provided an opportunity tomake density measurements at severalaltitudes: 427, 296, and 230 km (231,160, and 124 nautical miles). However,the vehicle flew for 42 hours at 230 km(124 nautical miles) with the payloadbay surfaces pointed into the velocityvector during the main portion of the mission to obtain high fluence. The mass spectrometer provided by the

216 Engineering Innovations

US Air Force was the key component of the experiment and was capable ofsampling both the direct atomic oxygenflux as well as the local neutralenvironment created by interaction of atomic oxygen with surfaces placedin a carousel. Five carousel sectionswere each coated with a differentmaterial to determine the materialeffects on released gases. Materialsamples trays, which providedtemperature control plus instruments to control other exposure conditions,were placed on each side of the massspectrometer/carousel.

NASA achieved all of the Evaluation of Oxygen Interaction with Materials III objectives during STS-46.A well-characterized, short-term,high-fluence atomic oxygen exposurewas provided for a large number ofmaterials, many of which had neverbeen exposed to a known low-Earthorbit atomic oxygen environment. Thedata provided a benchmark reaction ratedatabase, which has been used by theInternational Space Station, Hubble, andothers to select materials and coatings toensure long-term durability.

Reaction rate data for many of thematerials from earlier experiments wereconfirmed, as was the generally weakdependence of these reaction rates ontemperature, solar UV exposure,oxygen atom flux, and exposure tocharged ionospheric species. The roleof surface collision energy on oxygenatom reactivity was quantified bycomparing flight reaction rates of keyEvaluation of Oxygen Interaction withMaterials III experiment materials with reactivity measurements made inwell-characterized laboratory oxygenatom systems with lower surfacecollision energies. This evaluation also provided an important benchmarkpoint for understanding the role of

solar extreme UV radiation damage in increasing the generally low surface reactivity of perfluorinatedorganic materials. The massspectrometer/carousel experimentproduced over 46,000 mass spectraproviding detailed characterization of both the natural and the inducedenvironment. The mass spectrometerdatabase provided a valuable resourcefor the verification of various models of rarified gas and ionospheric plasmaflow around spacecraft.

Intelsat Satellite

Knowledge gained from atomic oxygen reactivity studies played a key role in the STS-49 (1992) rescue of the communications satellite

Intelsat 603 that was used to maintaincommunications from a geosynchronousorbit. Failure of the Titan-3 upper stageleft Intelsat 603 marooned in anunacceptable low-Earth orbit andsubject to the effects of atomic oxygendegradation of its solar panels, whichcould have rendered the satellite useless.NASA quickly advised the InternationalTelecommunications SatelliteOrganization (Intelsat) Consortium ofthe atomic oxygen risk to Intelsat 603,leading to the decision to place thesatellite in a configuration that wasexpected to minimize atomic oxygendamage to the silver interconnects onthe solar panels. This was accomplishedby raising the satellite altitude andchanging its flight attitude so thatatomic oxygen fluence was minimized.


Evaluation of Oxygen Interaction with Materials III flight experiment in the Orbiter payload bay ofSTS-46 (1992). Material exposure samples are located on both sides of the mass spectrometer gasevolution measurement assembly in the center.

To provide facts needed for a finaldecision about a rescue flight, NASAdesigned and executed the IntelsatSolar Array Coupon flight experimenton STS-41 (1990). The experimentresults, in combination withground-based testing, supported thedecision to conduct the STS-49 satelliterescue mission. On this mission,Intelsat 603 was captured and equippedwith a solid re-boost motor to carry it to successful geosynchronous orbit.

NASA Discovers Light Emissions

On the early shuttle flights, NASAobserved another effect caused by the interaction between spacecraftsurfaces and the low-Earth orbitenvironment. Photographs obtained by using intensified cameras andconducted from the Orbiter cabinwindows showed light emissions(glow) from the Orbiter surfaces whenin forward-facing conditions.

The shuttle provided an excellentopportunity to further study thisphenomenon. On STS-41D (1984),astronauts photographed variousmaterial samples using a special glowspectrometer to obtain additional dataand determine if the glow wasdependent on surface composition.These measurements, along with thematerial recession effects and dataobtained on subsequent flights, led to a definition of the glow mechanism.

Spacecraft glow is caused by theinteraction of high-velocity oxygenatoms with nitrous oxide absorbed onthe surfaces, which produces nitrogendioxide in an electronically excitedstate. The excited nitrogen dioxide isreleased from the surfaces and emitslight as it moves away and decays from its excited state. Some nitrousoxide on the surface and some of thereleased nitrogen dioxide result fromthe natural environment. The lightemission occurs on any spacecraftoperating in low-Earth orbit; however, the glow could be enhancedby operation of the shuttle attitudecontrol engines, which producednitrous oxide and nitrogen dioxide as reaction products. These findingsled to a better understanding of thebehavior of spacecraft operating inlow-Earth orbit and improved accuracyof instrument measurements.


STS-62 (1994) orbits Earth duringa “night” pass, documenting theglow phenomenon surroundingthe vertical stabilizer and the

Orbital Maneuvering System podsof the spacecraft.

The Intelsat Solar Array Coupon flight experimentshown mounted on the Shuttle Robotic Armlower arm boom and exposed to spaceenvironment conditions during STS-41 (1990).

ChemicalFingerprinting

Comprehensive ElectronicSystem for Greater Flight Safety

A critical concern for all complexmanufacturing operations is thatcontaminants and material changes overtime can creep into the productionenvironment and threaten productquality. This was the challenge for thesolid rocket motors, which were inproduction for 30 years.

It is possible that vendor-supplied rawmaterials appear to meet specifications

from lot to lot and that supplier process changes or even contaminatedmaterial can appear to be “in spec” but actually contain subtle, criticaldifferences. This situation has thepotential to cause significant problemswith hardware performance.

NASA needed a system to readily detectthose subtle yet potentially detrimentalmaterial variances to ensure thepredictability of material properties andthe reliability of shuttle reusable solidrocket motors. The envisioned solutionwas to pioneer consistent and repeatableanalytical methods tailored to specific,critical materials that would yieldaccurate assessments of material

integrity over time. Central to thesolution was both a foolproof analysisprocess and an electronic data repositoryfor benchmarking and monitoring.

A Chemical “Fingerprint”

Just as fingerprints are a precisemethod to confirm an individual’sidentity, the solid rocket motor projectemployed chemical “fingerprints” toverify the quality of an incoming rawmaterial. These fingerprints compriseda detailed spectrum of a givenmaterial’s chemical signature, whichcould be captured digitally and verifiedusing a combination of sophisticatedlaboratory equipment and customanalytical methods.

The challenge was to accuratelyestablish a baseline chemical fingerprintof each material and developreproducible analytical test methods tomonitor lot-to-lot material variability. A further objective was to gain agreater understanding of criticalreusable solid rocket motor materials,such as insulation and liner ingredients,many of which were the samematerials used since the Space ShuttleProgram’s inception. New analyticaltechniques such as the atomic forcemicroscope were used to assessmaterials at fundamental chemical,molecular, and mechanical levels.These new techniques provided thehigh level of detail sought. Because ofunique attributes inherent in eachmaterial, a one-size-fits-all analysismethod was not feasible.

To facilitate documentation and datasharing, the project team envisioned acomprehensive electronic database toprovide ready access to all relevant data. The targeted level of backgrounddetail included everything from whereand how a material was properly used to details of chemical composition.


During the Space Shuttle Program’s operation, issues arose regarding the use of

substances that did not meet emerging environmental regulations and current

industry standards. NASA worked to develop chemicals, technologies, and processes

that met regulatory requirements, and the agency strove to identify, qualify, and

replace materials that were becoming obsolete as a result of environmental issues.

The stringent demands of human spaceflight required extensive testing and

qualification of these replacement materials.

EnvironmentalAssurance

Reuseable Solid Rocket MotorTCA* Reduction History

* 1,1,1 trichloroethane

The ideal system would enable aqualified chemist to immediatelyexamine original chemical analysis datafor the subtle yet significant differencesbetween the latest lot of material andprevious good or bad samples.

To develop such a system, commerciallyavailable hardware and software wereused to the greatest extent possible.Since an electronic framework to tie the data together did not exist, one wasdesigned in-house.

The Fingerprinting Process

The chemical fingerprinting program,which began in 1998 with a prioritizedlist of 14 critical materials, employed a team approach to quantify anddocument each material. Theinterdisciplinary team included designengineering, materials and processesengineering, procurement qualityengineering, and analytical chemistry.Each discipline group proposed testplans that included the types of testingto be developed. Following approval,

researchers acquired test samples(usually three to five lots of materials)and developed reliable test methods.Because of the unique nature of eachmaterial, test methods were tailored toeach of the 14 materials.

A “material” site in the projectdatabase was designed to ensure alldata were properly logged and criticalreports were written and filed. Oncethe team agreed sufficient data hadbeen generated, a formal report wasdrafted and test methods were selectedto develop new standard acceptanceprocedures that would ultimately beused by quality control technicians tocertify vendor materials.

The framework developed to packagethe wide-ranging data was termed the Fingerprinting Viewer. Programdata were presented through a series of cascading menu pages, each withincreasing levels of detail.

The Outcomes

Beyond meeting the primary programobjectives, a number of resultingbenefits were noted. First, throughincreased data sharing, employeescommunicated more effectively, bothinternally and with subtier suppliers.The powerful analytical methodsemployed also added to the suppliers’materials knowledge base. Subtlematerials changes that possibly resulted from process drift or changesat subtier suppliers were detectable.Eight subtier suppliers subsequentlyimplemented their own in-housechemical fingerprinting programs toimprove product consistency, recertifymaterial after production changes, or even help develop key steps in the manufacturing process to ensurerepeatable quality levels.

Additionally, engineers could nowaccurately establish shelf-lifeextensions and storage requirements


Image 3-D Plot

The atomic force microscope affords a visual evaluation of surface preparation processesto improve understanding of their effects on bonding. The top panel represents topographyof a grit blast surface for comparison to a highly polished one. The atomic force microscopeuses an extremely fine probe to measure minute interactions with surface features evendown to an atomic scale. The maps at left are scaled from black at the bottom of valleys towhite at the tops of peaks within the scanned area. The 3-D projections at right are on acommon height scale. The grit blast surface clearly offers greatly increased surface areaand mechanical interlocking for enhanced bonding. Beyond simple topography, the probeinteractions with atomic forces can also measure and map properties such as microscopichardness or elastic modulus on various particles and/or phase transitions in a compositematerial, which in turn can be correlated with chemical and physical properties.

Grit Blasted

Polished1 µm

1 µm

Tools for Materials Evaluation Atomic Force Microscope Images of Metal Surface

© A

TK. A

ll rig

hts

rese

rved

.

for stockpiled materials. The ability tostore greater amounts of materials overlonger periods of time was valuable incases where new materials needed to becertified to replace existing materialsthat had become obsolete.

Finally, investigators were able to solveproduction issues with greaterefficiency. Comprehensive databasefeatures, including standardized testmethods and the extensive onlinereference database, provided resourcesneeded to resolve production issues in amatter of days or even hours—issuesthat otherwise would have requiredmajor investigations. In some cases,fingerprinting was also used to indicatethat a suspect material was actuallywithin required specifications. Thesematerials may have been rejected inprevious cases but, by using thefingerprinting database to assess the

material, the team could look deeper tofind the true root cause and implementproper corrective actions.

From Fingerprints to Flight Safety

The overarching value of the chemicalfingerprinting program was that itprovided greater assurance of the safetyand reliability of critical shuttle flighthardware. The fundamentalunderstanding of critical reusable solidrocket motor materials and improvedcommunications with vendors reducedthe occurrence of raw materials issues.NASA will implement chemicalfingerprinting methods into theacceptance testing of raw materialsused in future human space explorationendeavors. The full benefits of theprogram will continue to be realized inyears to come.

UnprecedentedAccomplishments in the Use ofAluminum-LithiumAlloy

NASA was the first to use weldedaluminum-lithium alloy Al 2195 at cryogenic temperatures,incorporating it into the External Tank under circumstances thatdemanded innovation.

From the beginning of the SpaceShuttle Program’s launch phase, NASAsought to reduce the weight of theoriginal tank, thereby increasingpayload capacity. Since the tank wascarried nearly to orbit, close to 100% ofthe weight trimmed could be applied tothe payload. NASA succeeded inimplementing numerous weight-savingmeasures, but the biggest challenge wasto incorporate a lightweight aluminumalloy—aluminum-lithium Al 2195—into the tank structure. This alloy hadnever been used in welded cryogenicenvironments prior to NASA’sinitiative. Several challenges needed tobe overcome, including manufacturingthe aluminum-lithium tank components,welding the alloy, and repairing thewelds. NASA and the External Tankprime contractor broke new ground inthe use of aluminum-lithium to producethe “super lightweight tank.”

The original tank weighed 34.500metric tons (76,000 pounds) dry. By the sixth shuttle mission, the tank’sweight had been reduced to 29.900metric tons (66,000 pounds). Thisconfiguration was referred to as the“lightweight tank.”

The real challenge, however, was stillto come. In 1993, the InternationalSpace Station Program decided tochange the station’s orbital inclination


This high-performance liquid chromatography/mass spectometry is employed to document minutedetails of a material’s chemical and molecular composition. Through the chemical fingerprintingsystem, seemingly minuscule discrepancies raise red flags that trigger investigations and precludedefective materials from reaching the production floor. Dr. Ping Li shown here at ATK in Utah.

© A

TK. A

ll rig

hts

rese

rved

.

to 57 degrees (a “steeper” launchinclination), allowing Russian vehiclesto fly directly to the station. Thatchange cost the shuttle 6,123 kg (13,500 pounds) of payload capacity.The External Tank project officeproposed to reduce the dry weight ofthe tank by 3,402 kg (7,500 pounds).

The Space Shuttle Program sought to incorporate lightweightaluminum-lithium Al 2195 into themajority of the tank structure, replacingthe original aluminum-copper alloy Al 2219; however, NASA first needed to establish requirements formanufacturing, welding, and repairingaluminum-lithium weld defects.

NASA started the super lightweighttank program in 1994. During the early phase, advice was sought fromwelding experts throughout the UnitedStates and the United Kingdom. The consensus: it was virtuallyimpossible to perform repairs onwelded aluminum-lithium.

The aluminum-lithium base metal also presented challenges. LockheedMartin worked with ReynoldsAluminum to produce the aluminum-lithium base metal. One early problemwas related to aluminum-lithiummaterial’s fracture toughness—ameasure of the ability of material with a defect to carry loads. Althoughmaterial was screened, flight hardwarerequirements dictated that structuresmust have the ability to function in the event a defect was missed by thescreening process. The specificdifficulty with the aluminum-lithiumwas that the cryogenic fracturetoughness of the material showed little improvement over theroom-temperature fracture toughness.

Since the two propellant tanks wereproof tested at room temperature andflown cryogenically, this fracturetoughness ratio was a crucial factor.

A simulated service test requirementwas imposed as part of lot acceptancefor all aluminum-lithium material used on the tank. The test consisted ofapplying room temperature andcryogenic load cycles to a crackedsample to evaluate the ability of thematerial to meet the fracture toughnessrequirements. Failure resulted in theplate being remelted and reprocessed.

Implementation of simulated servicetesting as a lot acceptance requirementwas unique to the aluminum-lithiummaterial. Testing consisted of croppingtwo specimens from the end of eachplate. Electrical discharge machining (a process that removes metal bydischarging a spark between the tooland the test sample) was used tointroduce a fine groove in each sample.The samples were then cyclicallyloaded at low stresses to generate asharp fatigue crack that simulated a defect in the material.

The first sample was stressed to failure;the second sample was stressed to nearfailure and then subjected to cyclicloading representative of load cyclesthe tank would see on the launch padduring tanking and during flight.

In the second sample, initial loadingwas conducted at room temperature.This simulated the proof test done onthe tank. Next, the sample was stressed 13 times (maximum tankingrequirement) to the level expectedduring loading of propellants atcryogenic temperatures and, finally,stressed to maximum expected flight

stress at cryogenic temperature. This cycle was repeated three moretimes to meet a four-mission-lifeprogram requirement with the exceptionthat, on the fourth cycle, the sample was stressed to failure and had toexceed a predetermined percent of the flight stress. Given the size of thebarrel plates for the liquid hydrogen and liquid oxygen tanks, only one barrelplate could be made from each lot ofmaterial. As a result, this process wasadopted for every tank barrel plate—32 in each liquid hydrogen tank andfour in each liquid oxygen tank—andimplemented for the life of the program.

Another challenge was related to thealuminum-lithium weld repair processon compound curvature parts. Theeffect of weld shrinkage in the repairscaused a flat spot, or even a reversecurvature, in the vicinity of the repairsand contributed to significant levels ofresidual stress in the repair. Multipleweld repairs, in proximity, showed thepropensity for severe cracking. Afterexamination of the repaired area, it wasfound that welding aluminum-lithiumresulted in a zone of brittle materialsurrounding the weld. Repeated repairscaused this zone to grow until theresidual stress from the weld shrinkageexceeded the strength of the weldrepair, causing it to crack.

The technique developed to repairthese cracks was awarded a US Patent.The repair approach consisted ofalternating front-side and back-sidegrinds as needed to remove damagedmicrostructure. It was also found thataluminum-lithium could not tolerate as much heating as the previousaluminum-copper alloy. This requiredincreased torch speeds and decreased


fill volumes to limit the heat to whichthe aluminum-lithium was subjected.

Additional challenges in implementingeffective weld repairs caused NASA toreevaluate the criteria for measuring thestrength of the welds. In general, weldrepair strengths can be evaluated byexcising a section of the repairedmaterial and performing a tensile test.The strength behavior of the repairedmaterial is compared to the strengthbehavior of the original weld material.In the case of the aluminum-copperalloy Al 2219, the strengths were

comparable; however, in the case of the aluminum-lithium alloy repair, thestrengths were lower.

Past experience and conventionalthinking was that in the real hardware,where the repair is embedded in a long initial weld, the repaired weld will yield and the load will beredistributed to the original weld,resulting in higher capability. Todemonstrate this assumption, a tensiletest was conducted on a 43-cm- (17-in.)-wide aluminum-lithium panelthat was fabricated by welding two

aluminum-lithium panels together and simulating a weld repair in thecenter of the original weld. The panelwas then loaded to failure. The test that was supposed to indicate betterstrength behavior than the excisedrepair material actually failed at alower stress level.

To understand this condition, anextensive test program was initiated to evaluate the behavior of repairs on a number of aluminum-copper alloy (Al 2219) and aluminum-lithium alloy (Al 2195) panels.


The use of aluminum-lithium AI 2195 in manufacturing major External Tank components, such as the liquid hydrogen tank structure shown above, allowed NASA to reduce the overall weight of the External Tank by 3,402 kg (7,500 pounds). The liquid hydrogen tank measured 8.4 m (27.5 ft) in diameterand 29.4 m (96.6 ft) in length. Photo taken at NASA’s Michoud Assembly Facility in New Orleans, Louisiana.


With any space vehicle, minimum weight is of critical

importance. Initial trade studies indicated that using a

graphite/epoxy structure in place of the baselined aluminum

structure provided significant weight savings of about 408 kg

(900 pounds [4,000 newtons]), given the large size and excellent

thermal-structural stability. Two graphite/epoxy composite

materials and four structural concepts—full-depth honeycomb

sandwich, frame-stiffened thin sandwich, stiffened skin with

frames and stringers, and stiffened skin with frames only—

were considered for weight savings and manufacturing

producibility efficiency. These studies resulted in the selection

of the frame-stiffened thin sandwich configuration, and

component tests of small specimens finalized the graphite

fiber layup, matrix material, and honeycomb materials.

Graphite/epoxy properties at elevated temperatures are

dependent on moisture content and were taken into account

in developing mechanical property design allowables.

Additionally, NASA tracked the moisture content through all

phases of flight to predict the appropriate properties during

re-entry when the payload bay doors encountered maximum

temperatures of 177°C (350°F).

Payload bay doors were manufactured in 4.57-m (15-ft)

sections, resulting in two 3 x 18.3 m (10 x 60 ft) doors.

The panel face sheets consisted of a ± 45-degree fabric ply

imbedded between two 0-degree tape plies directed normal to

the frames and were pre-cured prior to bonding to the Nomex®

honeycomb core. A lightweight-aluminum wire mesh bonded

to the outside of face sheets provided lightning-strike

protection. Frames consisted primarily of fabric plies with the

interspersions of 0-degree plies dictated by strength and/or

stiffness. Mechanical fasteners were used for connection

of major subassemblies as well as final assembly of the doors.

All five Orbiter vehicles used graphite/epoxy doors, one of the

largest aerospace composite applications at the time, and

performance was excellent throughout all flights. Not only was

the expected weight saving achieved and thermal-structural

stability was acceptable, NASA later discovered that the

graphite/epoxy material showed an advantage in ease of repair.

Ground handling damage occurred on one section of a door,

resulting in penetration of the outer skin of the honeycomb core.

The door damage was repaired in 2 weeks, thereby avoiding

significant schedule delay.

Orbiter Payload Bay Door One of the largest aerospace composite applications of its time.

Test panels were covered with aphoto-stress coating that, underpolarized light, revealed the strain pattern in the weld repair. The Al 2219 panel behaved asexpected: the repair yielded, the loadsredistributed, and the panel pulled wellover the minimum allowable value. In aluminum-lithium panels, however,the strains remained concentrated in the repair. Instead of the 221 MPa(32,000 pounds/in2) failure stressobtained in the initial welds, the welds were failing around 172 MPa(18,000 pounds/in2). These lowerfailure stress values were problematicdue to a number of flight parts that had already been sized andmachined for the higher 221 MPa(32,000 pounds/in2) value.

Based on this testing, it was determinedthat weld shrinkage associated with therepair resulted in residual stresses in the joint, reducing the joint capability.To improve weld repair strengths,engineers developed an approach toplanish (lightly hammer) the weld bead,forcing it back into the joint andspreading the joint to redistribute andreduce the residual stresses due toshrinkage. This required scribing andmeasuring the joint before every repair,making the repair, and then planishingthe bead to restore the weld to itsprevious dimensions. Wide panel testresults and photo-stress evaluation ofplanished repairs revealed that thenewly devised repair procedure waseffective at restoring repair strengths toacceptable levels.

Testing also revealed that planishing ofweld beads is hard to control precisely,resulting in the process frequentlyforming other cracks, thus leading toadditional weld repairs. Because of the

difficulty in making and planishingmultiple repairs, a verification ground rule was established that every“first repair of its kind” had to bereplicated on three wide tensile panels,which were then tested either at room temperature or in a cryogenicenvironment, depending on thein-flight service condition expected for that part of the tank.

All these measures combinedaccomplished the first-ever use ofwelded aluminum-lithium at cryogenictemperatures, meeting the strictdemands of human spaceflight. Thesuper lightweight tank incorporated 20 aluminum-lithium ogive gores (the curved surfaces at the forward end of the liquid oxygen tank), fourliquid oxygen barrel panels, 32 liquidhydrogen barrel panels, 12 liquidoxygen tank aft dome gores, 12 liquidhydrogen tank forward dome gores, and 11 liquid hydrogen aft dome gores.

Through this complex and innovativeprogram, NASA reduced the 29,937-kg(66,000-pound) lightweight tank byanother 3,401.9 kg (7,500 pounds). The 26,560-kg (58,500-pound) superlightweight tank was first flown onSpace Transportation System (STS)-91(1998), opening the door for the shuttle to deliver the heaviercomponents needed for construction of the International Space Station.


materials and manufacturing - nasa€¦ · materials and manufacturing ... in addition, there were...

Documents