Embed Size (px)
Citation preview
Nav~~~l CAmfd.Sr Ds AD- A285 773
* ot~tand 0011 Divi2of
cotoloc Center RDT&E ivso
sur333
.......... 3.. ..
3il
j. [) Stci
94- 324894
Ir
Technical Report 159231 June 1994
3 Structural Performance of CylindricalPressure Housings of DifferentCeramic Compositions Under
3 External Pressure LoadingPart III, Sintered Reaction Bonded Silicon
3 Nitride Ceramic
i R. R. KurkchubascheR. R JohnsonJ. D. Stachiw Accesion For
NTIS CRA&I
IDIC TAB fJUI,i:WouLced _--_
By ...........................
Distributionj
3A'- i!ablity Codes
Avail ai,djorDIMt Special -
IA-,
I
NAVAL COMMAND, CONTROL ANDOCEAN SURVEILLANCE CENTER
RDT&E DIVISIONSan Diego, California 92152-5001
K. E. EVANS, CAP1 USN " R. T. SHEARERCommanding Officer Executive Director I
IADMINISTRATIVE INFORMATION
This work was performed by the Marine Materials Technical Staff, RDT&E Division Iof the Naval Command, Control and Ocean Surveillance Center, for the Naval Sea Sys-tems Command, Washington, DC 20362. IReleased by Under authority of 3R. L. Wernli C. A. Keeney, HeadOcean Technology Branch Ocean Engineering
Division
ACKNOWLEDGMENT
A special thank you for the contributions of the following people from the Technical IInformation Division (TID): Eric Swenson of the Technical Writing Branch; Patty Gra-ham, Sally Lycke, and Tim Ruiz of the Computerized Production and Printing Branch;and the Photography and Computer Graphics Branches; and Dr. Andre Ezis andDr. Richard Policka of CERCOM, Inc.
EIIIIU
ES I
FEATURED RESEARCH
SUMMARY ness. Each cylinder was ultrasonically inspected
prior to testing and then fitted with epoxy-bondedTen 12-inch-outside-diameter (OD) by 1 8-inch-'ong titanium end-cap joint rings. After being Instrum-silicon nitride (Si3N4) cylinders were fabricated by ented with strain gages, the cylinders were pres-CERCOM, Inc., nondestructively inspected, and sure tested cyclically and to destruction. CERCOMassembled. They were subsequently Instrumented has demonstrated that it can successfully andand pressure tested at Southwest Research Insti- repeatably cast 12-inch-OD by 18-inch-long cylin-tite (SRI) under the supervision of the Naval Corn- ders from PSX. These cylinders are the largest-mand, Control and Ocean Surveillance Center known monolithic silicon-nitride components ever(NCCOSC) RDT&E Division (NRaD) under the fabricated. All ten cylinders passed proof testing toProgram for the Application of Ceramic to Large at least 10,000 psi. Cylinder failures can be attrib-Housings for Underwater Vehicles (reference 1). uted either to cyclic fatigue or to intentional pres-CERCOM's material, designated PSX. was chosen surization to critical collapse pressure. Failure offor Its high specific compressive strength, high the cylinders by buckling matched closely the pre-specific elastic modulus, and high fracture tough- dictions made by hand and computer calculations.
Il
FEATURED RESEARCH
CONTENTS
INTRODUCTION I
BACKGROUND 1
FABRICATION APPROACHES 1
SRBSN PROCESSING 2
OBJECTIVES 3
APPROACH 3
FABRICATION 3
PSX Fabrication Process 3
Powder Processing 3
Cold-Isotatic Pressing (Isopressing)_ 3
Pre-nitriding Green Cylinder 4
5 Green Machining_ 4
Nitriding 4
Sintering 5
QUALITY ASSURANCE 5
3 Dimensional Inspection 5
Dye Penetrant Inspection 5
Material Characterization_ 5
Pulse-Echo Ultrasonic Inspection 6
PRESSURE TESTING 6
Test Setup 6
Pressure Testing 7
U TEST OBSERVATIONS/DISCUSSION 8
MATERIAL PROPERTIES 8
U STRAINS 9
IvI
FEATURED RESEARCH I
STRUCTURAL PERFORMANCE 9
CYCLIC FATIGUE LIFE 9
CONCLUSIONS 10 URECOMMENDATIONS 11
REFERENCES 12
GLOSSARY 113
FIGURES 31. Process flow diagram for SRBSN and SSN fabrication 14
2. Fabrication process for PSX_15 33. Schematic of cold isostatic-pressing tooling set 16
4. Cold-walled, vacuum-nitriding iumace shown with nitrided cylinder 17 35 CERCOM's graphite-based (2200 degrees C) sintoring furnace used for densilying Si3N4 18
6. Sintered S13 N4 cylinder prior to grinding_ 19
7 Diamond grinding of fully densified Si3 N4 cylinder 20
8. Typical Vikers Indentation for measuring fracture toughness 21
9. EngIneering drawing for 12-inch-OD by 18-inch-long S13N4 ceramic cylinder 22
10. Pulse-echo C-scan of S13 N4 ceramic calibration standard used for inspection of
deliverable components 23
11. 12-inch cylinder test assembly, Type 1 configuration, Sheet 1 24
11. 12-inch cylinder test assemoly, Type 1 configuration, Sheet 2 25
11. 12-inch cylinder test assembly, Type 1 configuration, Sheet 3 26
12. 12-inch cylinder Mod 1, Type 2 end-cap joint ring_ 2_ 7
13. 12-inch cylinder spacer 28
14. 12-inch hemisphere 29 315. 12-Inch cylinder Mod 1 end-cap joint O-ring 30
16. 12-inch hemisphere clamp band 31 317. 12-inch hemisphere plug 32
18, 12-inch hemisphere washer 33
19. 12-inch hemisphere wooden plug 34
i
FEATURED RESEARCH
3 20. 12-inch cylinder test assembly, Type 2 configuration, Sheet 1 35
20. 12-inch cylinder test assembly, Type 2 configuration, Sheet 2 36
21. 12-Inch flat end plate 37
22. 12-Inch cylinder test assembly, Type 3 configuration, Sheet 1 38
22. 12-Inch cylinder test assembly, Type 3 configuration, Sheet 2 39
23. 12-inch cylinder Mod 1, Type 1 end-cap joint drig __40
24. 12-Inch fiat end-plate tie rod_ 41
25. 12-Inch fiat end-plate feed through 42
26, 12-Inch end-plate wooden plug 43
27. Photo of fully machined S13N4 ceramic cylinder prior to bonding of titanium end rings 44
28, Photo of fully machined Si3 N4 ceramic cylinder with titanium end rings bonded in place_ 45
29. Enc -cap joint ring removae fixture 46
30. Pressure vs. strain plot for cylinder 001 47
31. Circumferential cracking on the bearing surface of cylinder 001. The titaniumend-cap joint ring was removed after pressure testing to !nspect for cracking 48
32. Pulse-echo C-scan of cylinder 001, Sheet 1 49
32. Pulse-echo C-scan of cylinder 001, Sheet 2 50
S32. Pulse-echo C-scan of cylinder 001, Sheet 3 51
33. Pulse-echo C-scan of the other end of cylinder 001 52
S34. Presst-re vs. strain plot for cylinder 002 53
35. Pressure vs. strain plot for cylinder 003 5336. Pressure vs. strain plot for cylinder 004 54
37. Pressure vs. strain plot for cylinder 005 54
38. Pressure vs. strain plot for cylinder 006 55
39. Pressure vs. strain p~ot for cylinder 007 55
40. Pressure vs. strain plot for cylinder 006 56
41. Pr-sesure vs. strain plot for cylinder 009, pressurization No. 1 56
42. Pressure vs. strain plot for cylinder 009, pressurization No. 2 57
43. Pressure %s. strain plot for cylinder 010, pressurization No. 1 57
II 44. Pressure vs. strain plot for cylinder 010, pressurization No. 2 58
i45. WI vs. depth curves for AL-600 96-percent alumina ceramic and PSX ceramic 58
3 vii
Ii
PEATURED RESEARCH
TABLES U1. Typical properties for PSX commercial grade of silicon nitride 59
2. Actual dimensions of deliverable PSX cylinders 59
3. Messured material properties of deliverable PSX cylinders 60
4. Summary of findings made in PSX cylinders using pulse-echo ultrasonic inspection 61
5. Average hoop and axial strains measured on inner mid-bay diameter of PSX cylindersat 10.000-psi external hydroststic pressure 82
6. Summary of test plans and results for cylinders 001 through 010, Sheet 1 63
6. Summary of test plans snd results for cylinders 001 through 010. Sheet 2 64
7. Comparison of cyclic pressure testing performance of PSX Siiicon Nitride cylinders andof AL-60U ýo6-pu6f.iýnt dJumns.-'2temmic cyiitders 65 3
6. Comparison of tho material compositions evaluated for the application of ceranic,to large housings for UUVs 66
III
I
II
I
FEATURED RESEARCH
INTRODUCTION most extensive testing had been completed onS..94-percent alumina-ceramic housings only. Testing
showed limited cyclic fatigue life under repeatedUnmanned underwater vehicles (UUVs) require pressurization (references 5 and 6). The eventualpressure-resistant housirgs for containment of failure of components due to repeated pressuriza-their electronics and pocer supply. Currently, such tion is attributed to radial tensile stres. at thehousings are fabricated from metals such as alumi- ceramic-to-titaniuri metal-bearing interface (seenum, titanium, or steel. However, these materials figure 25 of refeience 1). This stress leads to inter-Iesult in very hea-y housings when vehicles are nal circumferential cracks which run through thedesigned to operate at depths as great as wall, eventually breaking off in shards, causing20.000 feet. leakage or catastrophic failure. It is believed thatCeramic materials, because of their high specific one way of increasing the cyclic fatigue life of
cor,1pressive strength and high specific elastic ceramic componeols is to use compositions having
modulus, are ideally suited for application to exter- higher fracture toughiness than alumina-ceramic
nal pressure-resistant housings tor underwater composition.
applications. A more detailed justification for using The program ior the application of ceramics toceramic niaterials ;n external pressi ire-resistant large housings for unaerwater vehicles gave NRaLhousings may be found in the outline for the pro- the opporturity to test some new ceramic somposi-aram under whi(uh this work. was performed by the tons exhiblting greater fracture toughness. TheseNaval Command, Control and Ocean Surveillance include silicon-nitride ceramic, ZTA ceramic, andCenter (NCCOSC), RDT&E Division (NRaD). See Lanxide's silicon carbide particulate-reinforced alu-reference 1. mina-ceramic composition (90-X-089). Silicei
One of the objectives of this program was to evalu- nitride ceramic was chosen for evaluation because
compositions for of its high fracture toughness, high zompressiveate various advanced ceramic Compositions s;ength and elastic modulus, and low specificuse in pressure -resistant housings. Composions gravity. Sintered reaction-bonded silicon nitrideavaluated include zirconia-toughened alumina - (~S)wscoe tcuei stecmoi
(ZTA) Ceramic (reference 2), silicon carbide partic- (SRBSN) was chosen btcause it is the composi-ulate-reinforoed alumina (SiC/AI2Oa/AI) ceramic tion CERCOM offered to make for NRsD under
comoetitive contract award. CERCOM was the(reference 3), and silcuon-nltriae (Si3 N4) ceramic. lowest bidder meeting the physical propertyThe common base of comparison for testing these requiremen g the pysical ---
materials was 96-percent alumina ceramic requirements chosen by NRaD.
manufactured by WESGO. Inc. (reference 4). FABRICATION APPROACHESThis report summarizes the fabrication, nonda.structive inspection, ana testing of ten 12-inch- There are two generic types (based or, fabricationoutside-diameter (OD) by 1 8-inch-long Si3N4 route of silicon nitride that can meet the generalceramic cylinders fabricated by CERCOM, Inc. material and performance requirements for pres-
sure housings used for deep ,ubmergenoe ser-BACKGROUND vice: sintered silicon nitride (SSN) produced by
Si3N4 powder, and sintered reaction-bonded siliconSbrnitride (SRBSN) produced from Si metal powderNRaD has been procurng and testing cylindrical (figure 1)1. CERCOM, with extensive expertiseand hemispherical components made from producing components using both fabricationceramic over the last decade. Most of the work, approaches, has selected the SRBSN process as
however, has focused on 94- and 96-percent the appro priatefor this pplcation.
alumina-ceramic compositions fabricated by Coors
Ceramics Company and WESGO. When the pro-gram for the application of ceramic to large hous-ings for underwater vehicles began, the 'Figures and tables are placed at the end of the text.I'
I!
FEATURED R~ESEARCH _____
The selection methodology was based not only on table 1 for typical property data). 1 he intergranularthe material mechanical property requirements, but phase (grain boundary) was engineered to be crys.also on the manufactured materal reliability, fab- talline so that the material would resist corrosionrication adaptation to the required component size and have good Impact resistance (optimel fractureand shape, cost comoetitiverness, and scale-up toughness).ability to larger diameter housings. In short, the As Illustrated, silicon nitride can be considered aSRBSN process offers the flexibility for material family of materials in the same manner as stain-and manufacturing design to produce high-quality less steel or brass. In material families certainceramic pressure housings. chiaracteristics are shared, but vary sufficiently toSRBSN PROCESSING permit the engiieering of propertie3 fir specific
applications. The family concept for silicon nitrideThe standard SRBSN process, as shown. in fig- originates with the ability to promote sintering withure 1, consists of (1) comminuting and blunding many different densification aids and with the abil-silicon metal powder and selected sintering aids, ity to fabricate bodies by a wide variety of available(2) cold forming the powder mixture into a net processing routes.shape, (3) nitriding tho green compact to convertthe r'licon metal to silicon nitride and concurrently Below its dissncistin temperature, the atomicrea(,tng the sintering aids to form second-phase mobility of Si3N4 is insufficient to produce the mass
compounds, and (4) thermrlly consolidating the transport required for full densification. Additives
reaction bonded preform to a Iliy dense state. that promote densification through the formation ofa liquid phase are required (alloying). Theoretically
CERCOM presently uses this basic fabrication pro- dense silicon nitride can be produced with thecess to manufacture eight different grades of sili- addition of one or moie oxides, such as MgO.con nitrilde. Each grade is engineered for a specihic Y20 3, A12 0 3, Si0 2 , ZrO2 , and Ced 2 . The type andapplication; i.e., properties are developeo through amount of aCditlve used in conjunction with the
grain boundary and microstructural engineering to selected densification process and heat treatmentmeet the specific needs of the application. For will ;argely determine the resultant properties of theexample, CERCOM manufactures one grade oW material. For example, certain additives or additiveSi3 N4 (CI SI3NA) designed specifically for use as combinatiors will produce an amorphous intergran-an Indexable cutting tool for the high-speed ular phase, whereas crystalline phases are pro.machining of cast iron. Another grade, known as duced by other additive selections and/or preciseNB Si3N4, is manufactured for the use of machir.- heat Ireatments. The nature of the intergranularIng nickel base alloys. In these examples, the phase wil; tron~ly influence the performance char-chemistry and nature of the grain boundary (crys- actoristics of the material.talline vs. amorphous) for each grade is developed Ato maximize wear resistance under the conditions As shown, with th many possibilities in additivefound at the interface between the metal work selection coupled with a wide variety st fabricationpiece and cutting tool. rutes, te f•mily of silicon nitride is large. From
th!s material family, CERCOM selected the PSX as
Another application example relates to the engi- the most appropriate for submersib!e applications.neering ot a suitable silicon nitride used for the PSX is produced using the SRBSN process; whereballs of ball-valve assemblies in pump mechia- densificaticn is accomplished using a proprietarynisms. These pumps ere used in 'down-hole" pressureless sinterng (PS) technique. The mate-applications for extracting oil where the compo- rial is characterized as isotropic with a high frac-nents are subjected to severe pressure changes ture toughns.t&--an important requirement forand a variety of corrosive environments including pressur,4 housings. The material also has excellentsalt water/brine. To meet these and other require- corrosion resistance because of the crystallized (X)ments, CERCOM developed a pressureless, sint- intergranular phase and has exce'ient strength withered grade of Si3N4 designated as PSX (see good reliability.
21
FEATURED RESEARCH
OBJECTIVES process flow apoears to be "straightforward," ar'd,generally, it is for standard shapea and sizes. How-ever, for the cylinder sizes required for this
The objectives of fabricating and testing CER- program, the required logistics for eac" stepCOM's PSX ceramic cylinders were: became a major technological challenge.
1. To evaluate CERCOM's ability to fabricate12-1nch-OD by 1 B-inch-long PSX cylinders. Powder ProcaulngThese parts are the larjest known monoli(hic Powd6rS of sil~con meta;, Y2 03, and A1203 arePSX parts ever to be fabricated. Cylinders sewd o aillig n me dia and awere to be nondestructNlvy inspected to sealed into a iar mill with grinding media and a
determine whether this composition could be nonaqueous La,'rier fluid. Comminution ainw
fabricated relatively free of defects. homogenization of the powder blend is achievedby rolling the jar mill for a specified speed and time
2. To determine tre structural performance of interval. To maintain process consistency, thePSX under external hydrostatic pressure load, slurry viscosity is monitored and adjusted duringlng. The structural performance parameters milling/honiogenization, as requ'ired, by carrier fluidinvestigated Included cyclic fatigue life and additions. Sili'orn nitride grinding media (PSX com-U critical collapse pressure. position) prevents contamination due to media
attrition. All powder han'ling and processing is per-3. To compare the performance of the SRBSN formed in a controlled environment to eliminate
S13N4 composition against the performance of air-bore contam~natlon which cAn form inclusionsthe baseline composition, WESGOs AL-600 in finished components. After achievinG satisfac-96-percent alumina ceramic, tory hnmogenizaton, the powder slurry is dried,
screened, and readied for green forming.IAPPROACHCold-1sostatlc Pressing (Isopressing)
The test plan for the ten cylinders included fabrics- A variety of processes have oean used for the con-tion, quality assuronce, and pressure testing. solidation of Si metal and Si3N4 green parts. The
green forming technique of choice for large cylin-FABRICATION ders is usually isopressing--the technique used
here. Isopresaing is a well-defined procedure forPSX Fabrication Process consolidating silicon-based powders and offers aI number of advantages:An overview of the fabrication process for PSX is
presented in figure 2. The overall process begins 1. A uniform, homogenous powder mixture canwith the selection of raw materials: silicon metal, be handled and pressed without exposure to3 and Y2 0 3 and AJ2 0 3 powders (densification aids). air and moisture.The raw materials require comminution, blending, 2. Isopressed cylinders can be fabricated with-and homogenizUor,. The processed powders are out binders, eliminating the need fc* bur-outshaped into cylinders by cold-isostatic pressing. normally required for injection molded f r slipScylinders are partially nitrided to allow for cast processed parts.rapid, conventionas green machining. Themachined net-shaped parts are nitrided to convert 3. Isopressed cylinders can be produced withthe remaining silicon to silicon nitride a• id then the controlled and uniform porosity requireddensifled using pressureless sintering technology, for pre-nitriding and nitnding,Cylinders are finished to final dimensions by dia-mond grinding, and all cormponents ar nspected 4. Isopresslng can be adapted to high-rate. full-
and characterized to ensure quality. Note that the scale produc(tion.
3
FEATURED RESEARCH 5
5. Isopressed cylinders are easily analyzed large cross section parts because of diminished Uusing standard Instrumental techniques gas paths (may lead to incomplete nitriding).and high-resolution computed tomographyto assure that critical size incluslons and The other technique for achieving acceptable han- Uagglomerates are not present. dMing strengths for green machining is termed pre-
nitriding, which was the procesb used for this6. lsopressing produces green-formed parts with program. Pre-nitriding, simply, uses a shortened
uniform cioss sections with relativnly high nitriding cycle to achieve the partial conversion ofdense/. silicon metal to silicon nitride, thereby developing
7. Cuntrolled uniform, high-density isopressed sufficient strength for subsequent green machiningparts result in low sioterng shrinkage. using conventional cutting tools and techniques.Optimum pre-nitriding conditions are those where
The prucedure requires a tooling set consisting ol about 10 percent of the normal nitrided welglit gain
a metal mandrel and a rubber bag wilt end raps is realized. Pre-nitrided cylinders with Insufficient(see figure 3 for build-up schematic). The mandrel weight gain•s do not have sufficient strength forforms the insid6 diameter (ID) of the isopressed green mactlining, whereas cylinders with excessivecylinder, the rubber bag defines the OD, and, weight gains become too "hard" and brittle fortogether with the end caps, they define the press- green machirning with conventional tooling.Ing or powder cavity. The rubber bag also trans-mits the hydrostatic pressure into the pressing Green Machining 3(powder) cavity during the pressing cycle. The OD and the cylinder ends are green machined
Thc metal maiidrel Is first placed into tne rubber on a lathe using carbide cutting tools without cool-sleeve. The blended powders are weighed and ants. The ID of the cylinder is determined by the Icharged into the pressing cavity under a controlled size of the isopress metal mandrel; therefore, theenvironment using vibiatiaoal energy. The cavity is ID does not require green machining.capped, sealed, and treaied with vacuum toensure de-airing. The mold then is placed into a Nltridinghydrostatic ch&mber, where pressure is isostatically appli6d k, the mold through a fluid. The Nitriding Is carried out in a cold-wall vacuum fur-prassed cylirider Is examined visually to ensure to nace (figure 4) where all inlterior construction mate- Iab3ence of cracks. rals are sele"ted for their overall inertness.
Heating elements, hearth, and heat shields arePur-nltrldlng Green Cylinder fabricated from molyboenum and high-ptrity alu-
mira is used as insulation. The green-formed cylin-When large silion metal cylinde,•s are isopressed ders are placed on Si3 N4 plates within the furnacethey nave insufficient strength for extensive han- iiot zone. The furnace vessel is evacuated to out-dling or green machining. There are two tech- gas the green ware and construction materials.niques that can be used to increase the strangth of Gas backfill techniques are used to asist this pro-green-formed silicon metal compacts: (1) argon cedure. Once the system integrity is confirmed bysinterting and (2) pre-nitrlding. a leak check, tha tempersture is ?SiSvi to 600 Idegrees C and held until a vacuum to 10-3 torr orArgon ciotering involves heating the isopressed lowfr Is established. This procedure assists theform in. an argon stmosphere at about 1150 oiuigasiny process and removes all chemicallydegrees C for several hoLrs, resulting in slight combined water and residual volatiles asseciatedfusion of silicon metal grains at cont-ct points with the system. The furnace then is backfilled to abetween particles in the compact. The fusicn pressure of 20.0 kPa with a gas containing 3 per-causes some shrinkage which can lead to cent H2 , 25 percent He, and 72 percent N2 for theincreased problems during nitriding, particularly for ramainder of the cycle.
4 1
FEATURED RESEARCH
The conversion process takes place ;n the temper- The sintered cylinder is machined with an OD/IDature range of 1100 to 1400 degrees C with two grinding machine (figure 7). Both diameters, asprinciple nitriding reactions: well as length, are diamond ground to print. Test
specimens also are machined from the "extra3 Si(s) + 2N2 (g) - Si3N4(s) (1) lengths" from all sintered cylinders.
3 Si(g) + 2N2 (g) - Si3 N4s) (2) QUALITY ASSURANCE
Successful nitridation is achieved by precise Quality assurance inludes dimensional and dye-control over the highly exothermic nature of reac- penetrant inspection, material characterization.tions (1) and (2). The kinetics at a given tempera- and ultrasonic inspection of each cylinder.ture produce an initial rapid nitriding rate whichslows until an essentially complete raaction asym- Dimensional Inspectiontote is reached (Arrhenius curve). Therefore, the Dimensional inspection of the ten silicon-nitrideacceptable technique tar 'exotherm" management cylinders shows that the actual cylinder dimnen-is to proceed increanentally, beginning at 1100 sions do not conform to the drawing dimensions.degrees C through the nltrindng temperature rai ge. CERCOM had dcfculty pressing parts which hadAt each temperature increment, the nitriding rate enoL, . stock to grind to the desired drawingis allowed to reach its asyMtote level before the dimensions. Dimensions were allowed to betemperature is increased again. Continuing this altered slightly, however, the requirement that theprocess to the maximum nitriding temperatumo of wall-thicknoss-to-diameter ratio remin constant1400 degrees C will result in full conversion of the had to be met. Table 2 is a summary of cylinder
silicon-to-silicon nitride. The time required a. a dimensions. Cylinders 006 and 007 were short-
given temperature for the reaction rate to reach the ened due to cracks which were found near the
asymtote level Is dependent upon many variables, ends of tho parts.
Some of the most critical variables are Si metal
purity, surface area and particle size, compact Dye Penetrant Inspectiongreen density, and cross-sectional thickness. Ifmanagement of the exctherm is not precise, the Dye penetrant inspectior of each cylinder was per-kinetics will cascade and the heat generated by the formed. Only cylinders without cracks were deliv-reaction will cause melting of the silicon. Once ered.
melted, the silicon coalesces and cannot be Material Characterizationnitrided.
The material characterization consists of elasticity,Slntering compression strength, flexural strength, and frac-S Sintering Is performed in an envicmnmrientally con- ture toughness measurements. All characteriza-tiotled graphite vacuum furnace, as bhown ir fig- tion, with the exception of elasticity, is conducted
ure 5. Thermal consolidation (densification) is on actual cylinder material. That is, all cylindersaccomplished th,'ough a liquid-phase sintering were fabricated with excess length. The excess
technique known as solution-recrystallization. Dur- material was removed during diamond grinding of
Ing sintering, the sintering additives form a liquld the cylinder and prepared into specimens appropri-
which promotes the alpha-to-beta phase trans- ate ror charactenfzation. The elasticity propertmes
formation in Si3 N4 . The sintering temperature is were measured from specimens fabricated from
selected to be high enough to form the liquid "like" processed matials.
phase, but low enough not to sublimate Si3 N4 . Young's Modulus, shear modulus, and Poisson'sSpecial sintering tooling is designed for dimen- Ratio are measured using the acoustic resonancesional control of the component during densifica- method, as de-scribed in ASTM C848-78. Thetoin. This technique allows for minimal distortion compression tests follow the proceudres outlinedduring sintering. Figure 6 shows the sintered part. in ASTM C773. Ths flexural strength, or Modulus
"5
IFEATURED RESEARCH 5
of Rupture, is measured using a four-point bending PRESSURE TESTING Itest; performed to MIL-STD-1 842 specificationusing the Type B testing configuration. The Weibull Test Setupmodel also is applied for statistical data analysis. IThat is, both the Weibull Modulus and the charac- Three types of assemblies were tested. They dif-
teristic strength are calculated using this modol. fered in the type of end closure used (hemispheri-
The Weibull Modulus is a relative indicator of cal or flat) and the type of titanium end-cap joint
material reliability. The higher the Welbull Modulus, ring used.the more reliable the date for design purposes. Hemispherical end closures were planned for useThe characteristic strength, always presented in a in cyclic tests, while fiat end plates were plannedWeibull analysis, is the stress condition where the for use in imp1osion tests (these being moreprobability of failure is 63 percent. rugged and, therefore, more capable of withstand
Fracture toughness is measured using the Vickers ing the force of the implosion at high pressure,Indentation Technique. The toughness measure- making them reusable). At program initiation, itmcnt is based on crack propagation under point was not planned to cycle any cylinders aboveloading. After diamond indenting the specimen, 12,000 psi. The machined hemispherical endsfour lateral cracks are generated from the comers ware calculated not to provide enough buckling Iof the indent (figure 8). The fracture tcughness is resistance above 12,000 psi. For this reason, all
calculated from the loading and the crack length, cyclic tests above 12,000 psi were run using flat-using the equation outlinod by Niihara pJ. Mat. Sci. end plates.Let. 1, 1982, pg. 13). Table 3 tabulates the data The difference in the two titanium end-cap jointgenerated for all fabricated cylinders, ring designs is the external seal. NRaD Mod 1,Pul-Echo Ultrsonic Inspection Type 2 end caps iave a lip on the OD. During cyl- I
inder a•ssembly, a silicon sealant is applied to thisEach cylinder was ultrasonically inspected by the lip. NlaD Mod 1, Type 1 end ceps do not have apulse-echo ultrasonic inspection technique. The lip. Type 1 end caps were originally intended forinspection was performed by Sonic Testing and use in proof and implosion tests; Type 2 end capsEngineering, Inc. (Southgate, CA). Sonic Testing were originally intended for cyclic tests becauseand Engineering used a 3/8-inch diameter, 3-inch they were believed to ensure a better seal underfocal length, 10-MHz transducer to perform the repeated pressurization. 3inspection. A small tile with fiat-bottom drilled holce(see figure 10) was used as a calibration standard. The three types of assemblies can be summarizedThis tile was made from the same composition as as follows:the cylinders. Details about the pulse-echo ultra- __
sonic inspection technique and other nondestruc- Assembly End Cap End Closuretive inspection techniques for ceramic components Type I Type 2 Hemisphericalcan be found in references 7 and 8. Table 4 lists Type 11 Type 2the results of the ultrasonic Inspection. It shows Type II Type 2 Flat Platethat some of the cylinders haa numerous large Type III Type 1 Flat Platedefects, while others had almost none. The first ""few deliverable cylinders were the "cleanest." Test assembly Type 1 is shown in figure 11. NRaDCERCOM attributes the dramatic difference in the Mod 1, Type 2 end caps (figure 12) are epoxynumber of defects to program changes made bonded to the ends of the ceramic cylinder usingbetween cylinders. Isostptbc pressing facilities were the procedure described in note 4 of figure 11. Achanged during the program. Green body quality 0.010-inch-thick manila paper gasket (figure 13)could have been affected by any number of vari- ensures a minimum 0.010-inch thickness of epoxysoles including weather, humidity, facility cleanli- on the bearing interface between the ceramic and Iness, and processing schedule. titanium.
I6 =; . ; i_ -_ 3
FEATURED RESEARCH
Cylinders 001 through 005 were bonded onto gated by a cylindrical wooden plug (figure 26). Fig-the smoothly machined ceramic surface finish. ure 27 shows a fully machlned cylinder. Figure 28However, test personnel noticed extrusion of epoxy is a fully machined cylinder with NRaD Mod 1,on the ID of some of the cylindrical test specimens Type 2 end caps epoxy-bonded to it.after pressure testing. It was decided that a betterbond b.:. ween the ttanium and ceramic could be Pressure Testingachieved if the ceramic surface to which the end- Pressure testing was performed in accordancecap joint rings would be bonded was grit blasted with the test-plan/result summary shown in table 6.prior to bonding. The bond surfaces on cylinders Strains were read at 1,000-psi intervals on the first006 through 010 were sandblasted by CERCOM. pressurization for each cylinder. When cylinders
In test assembly Type 1, the cylindrical assembly were to be purposely taken to failure pressure,is closest atseboth Tyends by ftahemiialaspheres strains also were read during the second cycle. Inis closed at both ends by titanium hemispheres some of the tests where the cylinder withstood all
(figure 14). The assembly is made watertight by a planned pressure testing, at least one end cap was
surface seal using a nitrile O-ring (figure 15) for removed us ing the end rmv fu swhich there is an O-ring gland machined into the in figure 29.
titanium end-cap joint ring. The titanium hemi-
sphere is joined to the titanium cylinder via a The method of end-cap removal involves heatingV-shaped steel clamp band (figure 16). the cylinder end to be removed, which breaks
down the epoxy, and then pulling the end cap offEach cylinder was instrumented with five the cylinder. Adequate force can be applied toCEA-06-250-UT-1 20 strain gage rosettes. These remove the end cap by using the mechanicalwere located at the center of the cylinder length advantage of turning nuts on the four 1/2-inch-and spaced 72 degrees apart. Each of the diameter tie rods. After the end caps were1/4-inch, 90-degree rosettes had one leg oriented removed, the cylinder was cleaned and taken toin the hoop direction and the other in the axial Sonic Testing and Engineering for pulse-echo ultra-direction. Electrical leads for the strain gages were sonic inspection to determine the presence andpassed through the pole of the upper hemisphere extent of internal circumferential cracking.via a plug (figure 17) held In place by a washer The following are brief summaries of the pressure(figure 18) and nut on the inside of the hemi- te for ach orielindes:sphere. The bottom hemisphere had a drain plug testing for each of the cylinders:which could be opened to determine whether there Cylinder 001 was proof tested to 16,000 psi andhad been any leakage during testing. A cylindrical inspected for damage. Strains were read atwooden plug (figure 19) was placed inside the 1,000-psi intervals and plotted (figure 30). After theassembly to mitigate the shock of implosion should proof test, the cylinder was subjected to pressurefailure occur. cycling at 16,000 psi to external pressure (this cor-
responds to an average maximum membraneFigure 20 shows test assembly Type 2. This test stress of 233,010 psi). Testing was terminatedassembly is identical to test assembly Type 1, after 581 cycles because the cylinder was leakingexcept that the cylinder ends are closed by flat at one end. The cylinder was shipped back tosteel bulkheads (figure 21) instead of titanium NRaD. and the end caps were removed. Figure 31hemispheres. shows the circumferential cracking on the bearing
Figure 22 shows test assembly Type 3 which uses surface of the cylinder, Figure 32 is a full-scaleNRaD Mod 1, Type 1 titanium end caps (figure 23) C-scan of this end. Cracking was extensive (up toinsteao of the Type 2 end caps used in test assem- 5.7 inches in length). The C-scan of the other end
bly Type 2. Test assemby Types 2 and 3 are held (figure 33) shows almost no cracking.together by four 1/2-inch tie rods (figure 24). Strain This cylinder demonstrates that extensive internalgage leads are passed through the feed through circumforential crackir'g can exist in a cylinder
shown in figure 25. The force of implosior is miti- without catastrophic failure. It is possible that the
7
FEATURED RESEARCH 5
leakage of water into the space between the Cylinder 007 was proof tested to 11,000-psi exter- 3ceramic and the titanium end-cap joint ring may nal hydrostatic pressure. Strains are plotted in fig-have accelerated crack growth, a phenomenon ure 39. This cylinder withstood 3,010 cycles toknown as stress-corrosion cracking. 11,000 psi without visible damage. Testing was
terminated because the test plan called forThe extrusion of the epoxy on this cylinder con- 3,000 cycles. This cylinder also was not ultrasoni-vinced NRaD that the smooth ceramic surface was cally inspected due to budgetary constraints.
not bonding strongly to the epoxy and the titanium. IiIt was decided that the ends of the remaining deliv- Cylinder 008 was proof tested to 11,000 psi. No
erable cylinders would be sandblasted to facilitate damage was noted. Strains are plotted in figure 40.
bonding. The first deliverable sandblasted cylinder The cylinder failed on the 1,665th pessurization to
was 006. 11,000 psi. UCylinder 0 proof tested to 6,000-psi exter- Cylinder 009 was proof tested to 10,000 psi.nalindesure. S2ais weroof restd and 16,00-psixted - Strains are plotted in figure 41. The cylinder wasnau pressure. Strains were read and plotted (fig- inspected for damage and then pressurized to fail-urmage 34) nod. The cylinder was ie thedand n d ure which occurred at 22,600 psi. Strains to failuredamage was noted. The cylinder was then cycled are plotned in figure 42. I
at 16,000-psi external hydrostatic pressure when it Iafailed on cycle number 337. Cylinder 010 was proof tested to 10,000 psi.
Strains are plotted in figure 43. The cylinder wasCylinder 003 was proof tested to 13,000-psi exter- inspected, and no damage was noted. It then wasnal pressure. The strains were read and plotted pressurized to failure which occurred at 21,600 psi I(figure 35). Subsequenty, the cylinder was cycled oxtemal hydrostatic pressure. Strains to failure areat 13,000 psi until failure on cycle number 1,379. plotted in figure 44.
Cylinder 004 was proof tested to 9,000-psi exter- A summary of test plan and test results is shown in Inal pressure. Strains were read and plotted (fig- table 6.ure 36). The test plan called for cycling the cylinder TEST OBSERVATIONS/DISCUSSION3,000 times to 9,000 psi. Testing was actually ter- Iminated after 3,008 cycles. The cylinder wasshipped back to NRaD where one end-cap joint MATERIAL PROPERTIESring was removed and that end of the cylinder was The measured material properties are shown insonically inspected. The pulse-ocho ultrasonic table 3. AJI material properties met or exceeded the Iinspection revealed no internal circumferential material property requirements shown on the cylin-cracking. der drawing (figure 9) with the exception of Weibull
Cylinder 005 was proof tested to 12,000-psi exter- Modulus which was measured to be 11. 1 with a
nal pressure. Strains were read and plotted (fig- low of 8.6 in cylinder 006 and a high of 20.4 in cyl-
ure 37). The cylinder was inspected for damage inder 010.
and then cycled to 12,000 psi until it failed on cycle The average measured specific gravity wasnumber 743. 3.27 gm/cc.
Cylinder 006 was the first cylinder with sand- The mean Young's Modulus was 43,960,000 psi
blasted ends to be tested. The cylinder was proof with a standard deviation of only 340,000 psi.tested to 12,000 psi external pressure. Strains are Young's Modulus ranged from 43,500,000 psi to
plotted in figure 38. The cylinder withstood 44,690,000 psi. The average Poisson's Ratio
2,241 cycles to 12,000 psi without failure or visible was 0.26.damage. Testing was terminated because the test The mean compressive strength was measured toplan only called for 2,000 cycles. This cylinder was be 486,600 psi with a standard deviation ofnot ultrasonically inspected due to budgetary 17,900 psi. Flexural strength was measured to beconstraints. 102,900 psi with a standard deviation of 5,300 psi.
8
FEATURED RESEARCH
The mean fracture toughne3s was measured to 1be 6.1 MPaem 1/2 w;th a standard deviation of E,, (o,," -0.2 MPaom1/2 , a minimum value of 5.8 MPaom 1 /2,and a maximum measured value of 6.2 MPa~m 1 2 . E= . (op - Y,a,)
where E = compressive (Young's) modulusSTRAINS and y = Poisson's Ratio
The average calculated Poisson's Ratio was foundAxial and hoop strains were measured and plotted to be 0.275. This value, with its standard deviationfor the proof cycle and in cylinders 009 and 010 for of 0.006, is not within range of the measured valuethe pressurization to failure. Strain plots show that of 0.252 with standard deviation of 0.006, Theboth the axial and the hoop strains remain within mean calculated Young's Modulus was 45.25 Msiclose range of each other circumferentially (stan- with a standard deviation of 1,11 Msi; this
dard deviation does not exceed 10 percent on axial exceeded the measured Young's Modulus ofstrains and 7 percent on hoop strains) and remain 43.96 MSi.
linear up to within about 1,500 psi of failure, at
which point the hoop strains diverge rapidly due to STRUCTURAL ,- ERFORMANCEthe buckling deformation of tha cylinder (see fig-ures 42 and 44). Table 5 summarizes strain read- The structural performance of the ten silicon nitride
ings at 10,000-psi external pressure. Strain cylinders tested was very good. None of the cylin-
measurements taken on the second pressurization ders failed on the first (proof) cycle. Proof pres-
to 10,000 psi In cylinders 009 and 010 show no sures ranged from 10,000 psi to 16,000 psi. Critical
significant difference In strain readings between collapse pressure exceeded 21,600 psi. The good
the first and second cycle as was found in the sili- structural performance is more impressive when
con-carbide particulate-reinforced alumina ceramic one considers the number of defects Oetected in
(Larxide composition 90-X-089). Therefore, silicon the cylinders during pre-service ultrasonic inspec-
nitride does not appear to exhibit a compaction tion.
effect. CYCLIC FATIGUE LIFE
Young's Modulus and Poisson's Ratio were calcu- Cyciic fatigue life data shows that the silicon-nitridelated for each cylinder using the simultaneous cylindeis had poor (low) fatigue life at a high stresssolution of the following equations: level (190,000 psi membrane stress). Testing
showed that fatigue life improved significantlyA thi&c-wall stress equation from reference 9 was when membrane stress levels were belowused to compute the expected stress at 10,000-psi 175,000 psi. This is demonstrated by cylindersextemal psi on the ID of the cylinder: 004, 006, 007, and 008.
--qaz The sandblasting of the cylinder ends seemed toa'-qZ have a dramatic effect on the cyclic performance
aIb2of the cylinders. It can be stated that cylinders 006
-q22 thnrough 008 performed significantly better thaniOb~p =a-.b-' cylinders 001 through 005. The difference is espe-
cially noteworthy oetween cylinders 005 and 006.
where q = extemel pressure Cylinder 006 had sand-blasted ends, while cylindera = outer radius 005 did not. When cy•.led at the same level ofb = inner radius stress (approximately 175,000 psi), cylinder 005
fa:.3d after 743 cycles while cylinder 006 exceededThen, the following two equations were solved 2,241 cycles. These two cylinders did not differsimultaneously for compressive modulus and Pois- significantly in the number of internal defects,son's Ratio: size, or in material properties (cylinder 006 had
II
FEATURED RESEARCH :lower flexural strength but higher fracture tough- tigated include critical collapse pressure and 3ness), cyclic fatigue life.
The very poor performance at a high stress level The structural performance of the silicon(233,000 psi) of cylinders 001 and 002 shows that nitride cylinders was excellent. None of thecyclic fatigue life degrades quickly with increased cylinders failed on proof test, even with thelevels of stress. However, the excellent perfor- numerous defects found in some of the cylin-
mance of cylinders such as cylinder 004, which ders. The critical collapse pressure exceeded 3withstood 3,008 cycles to 9,000 psi (131,0W psi) the design pressure (9,000 psi) by more then
and showed no signs of internal circumferential a factor of two. The structural performance of
cracking during pulse-echo ultrasonic inspection, silicon nitride can be considered predictable.
shows that the material performs extremely well at The cyclic performance of the silicon nitridethe intended design stress level of 131,000 psi at cylinders is not as predictable as would bean ocean depth of 20,000 feet. desired. However, the conclusion can be
drawn that the cyclic life can exceed 1,000Silicon nitride still presents some level of unpre- cycles at stress levels below 175,000 psi.
dictabillty in cyclic fatigue life as is evidenced by Cyclic performance degrades significantly atcylinders 007 and 008. These cylinders both had stress levels above 175,000 psi. Evidence Isandblasted ends and were tested to the same indicates that cylinders with ends sand-level of stress (160,000 psi), yet their perfor- blasted before end-cap joint rings are epoxy-mances differed dramatically, with cylinder 007 bonded to them perform better cyclically thanwithstanding 3,010 cycles and cylinder 008 failing those without sandblasted ends. The dataon cycle number 1,665. generated by cyclic pressurization of the cylin-
ders was not sufficient to draw any conclu-CONCLUSIONS sions about the relationship between any one
material property and cyclic performance. IThe objectives of this study were met: 3. Compare the performance of the SRBSN sili-
1. Evaluate CERCOM's ability to fabricate con nitride composition against the perfor- I12-Inch-OD by 18-inch-long Si3N4 cylinders. mance of the baseline composition,
Nondestructively inspect the cylinders to WESGO's AL-600 96-percent aluminadetermine whether this ceramic composition ceramic (reference 4).can be fabric.ated relatively free of defects. Table 7 compares the cyclic fatigue perfor-CERCOM was ab!e to fabricate the cylinders, mance of PSX silicon nitride cylinders andbut not without difficulty. The difficulties AL-600 cylinders subjected to external pres- 1appeared to be related to logistics, rather than sure cycles. This table indicates that thetechnical feasibility problems. CERCOM had structural performance of PSX silicon nitride isnumerous problems pressing quality parts, comparable to AL-600. AL-600 cylinders
leading to a high reject rate and eventually showed a clear trend of decreased fatigue liWeforcing CERCOM to make parts with modified as the peak external pressure of each cycleJimensions to meet NRaD's delivery require- was increased. The relationship betweenments. The fact that the first five cylinders cyclic life and peak external pressure for eachdelivered met dimensional requirements arnd cycle was not as defined for the PSX silicon
appeared to have the highest quality (in terms nitride cylinders tested for this report. Theof number and size of defects detected) dem- more predictable behavior of the AL-600 cylin-
onstrated that the difficulties were not of a ders can be attributed to the WESGO's sub-technical nature. stantial prior experience in fabricating large
hull shapes from their alumina-ceramic com-2. Determine the structural performance of Si3N4 positions. On the other hand, the silicon
under external hydrostatic pressure loading, nitride cylinders that were evaluated repre-The structural performance parameters inves- sented the largest parts fabricated by
I10 1
FEATURED RESEARCH
CERCOM using their PSX composition, and, 2. If CERCOM's PSX silicon nitride compositionconsequently, some manufacturing difficulties is to be used in an underwatei external pres-were encountered. sure application and 1,000 cycles to design
Table 8 contains the material properties of depth are expected, then the design should
CERCOM's PSX silicon nitride and WESGO's be such that maximum nominal membrane
AL-600 96-percent alumina ceramic. The stress does not exceed 160,000 psi. The
properties of WESGO's ZTA ceramic (refer- engineering properties to be used for engi.
ence 2) and Lanxide's 90-X-089 SiC/A120 3/AI neering calculations and to be called out oncomposition (reference 3) are included for the engineering drawing should be:comparison. As the table shows, Si3N4 has Compressive strength: 460,000 psibetter properties than AL-600 96-percent alu- Compressive Modulus: 43.9 Msimina ceramic with the exception of compres- Flexural Strength: 100,000 psisive modulus and Weibull Modulus. Ninety-six Fracture Toughness: 6.0 ksi (in1/2)percent alumina ceramic's higher elastic Specific Gravity: 3.275 g/ccmodulus turns out to be important as designs Poisson's Ratio: 0.26are often driven by buckling requirements. Standard finite-element analysis and bucklingThe difference in Weibull Modulus could analysis can be used to analyze pressure-account for the difference in the predictability housing designs using CERCOM's PSX sili-of the material during cyclic testing. con de.
It should be noted that Si3N4 has the highest com- Furthermore, the cylinder assembly designpressive strength, the highest flexural strength,shudio prtMd1,Te2en-aand the lowest specific gravity of the four composi- should incorporate Mod 1, Type 2 end-captions evaluated as part of the program for the joint rings bonded in accordance with note 4application of ceramic to underwater pressure of figure 9 to a ceramic surface which hashousings. Figure 45 shows W/D vs. depth curves been roughenea by sandblasting on the ODfor Si3N4 and 96-percent A120 3 . The example is for and ID circumferential surfaces of the part.a housing having a bength-to-diameter ratio of 1.5. 3. Prior to putting the cylindrical assemblyThe maximum design stress level for both materi- together, the ceramic cylinder should beals is chosen for a cyclic fatigue life of 1,000 cycles inspected ultrasonically. A pulse-ecno ultra-to design depth. Although the curves change for sonic inspection using a 3/8-inch diameter,different length-to-diameter ratios, the general 3-inch focal length. 1 0-MHz transducer is rec-trend remains the same. ommended. The circumferential scanning
interval should be 0.010 inch.RECOMMENDATIONS 4. Cylinders should be inspected by an ultra-
sonic method for internal circumferentialThe following recommendations are made based cracks every 100 dives to 75 percent or moreon the testing of the ten CERCOM PSX Si3N4 of design depth. A handheld thickness-detec-12-inch-OD by 18-inch-long by 0.412-inch-thick tor unit may be sufficient to perform thiscy!inders. Inspection. Detection of internal cracks
extending beyond the titanium end-cap joint1. CERCOM composition PSX silicon nitride is rings should signal the removal of thatrecommended for use in the construction of ceramic cylinder from service.
external pressure housings of up to
12-inch-OD by !8-inch-long measurements. 5. Because silicon nitride shows promise forUse of the composition for cylinders with future application to huoaings for underwaterdimensions greater than this is recommended service, it is recommended that the fabricationonly after a thorough test-and-evaluation pro- process be scaled up to 20-inch-OD bygram. 30-inch-long cylinder sizes.
"IP" =i•I'" ' • m ' ' I "5 + "r r'l I.I" +rill +' ,r l+'!I 'I I I"1
- • i
FEATURED RESEARCH -
REFERENCES Alumina Ceramic," NRaD TR 1590 (Aug) INCCOSC RDT&E Division, San Diego, CA.
1. Kurkchubasche, R. R., R. P. Johnson, 5. Stachiw, ,I. D. 1993. "Exploratory Evaluation ofJ. D. Stachiw. 1993. "Application of Ceramics Alumina Ceramic Housings for Deep Submer- Ito Large Housings for Underwater Vehicles: gence Service--Third Generation Housings,"Program Outline," NRaD 7,D 2585 (Oct) NRaD TR 1314 (Jun) NCCOSC RDT&E Divi-NCCOSC HDT&E Division, San Diego, CA. sibn, San Diego, CA.
2. Johnson, R. P., R. R. Kurkchubasche, 6. Stachiw, J. D. 1990. "Exploratory Evaluation ofJ. D. Stachiw. 1993. "Structural Performance Alumina Ceramic Cylindrical Housings forof Cylindrical Pressure Housings of Different Deep Submergence Service-Fourth Genera- ICeramic Compositions Under External Pres- tion Housings," NOSC TR 1355 (Jun) Navalsure Loading: Part ll-Zirconia Toughened Alu- Ocean Systems Center, San Diego, CA.mina Ceramic," NRaD TA 1593 (Dec)NCCOSC RDT&E Division, San Diego, CA. 7. Kurkchubasche, R. R., R. P. Johnson,
3. Kurkchubasche, A. R., A. P. Johnson, J. D. Stachiw. 1993. "Evaluation of ND TestingTechniques for Quality Control of Alumina
J. D. Stachiw. 1993. "Structural Performance Ceramic Housing Components," NRaDof Cylindrical Pressure Housings of Different TR 1588 (Sep) NCCOSC RLT&E Di-vision, ICeramic Compositions Under External Pres- San Diego, CA.sure Loading: Part IV-Silicon Carbide Particu-late Reinforced Alumina Ceramic," NRaD 8. Kurkchubasche, R. R. 1993. "Non-destructiveTR 1594 (Dec) NCCOSC RDT&E Division, Evaluation Techniques for Deep SubmergenceSan Diego, CA. Housing Components Fabricated from Alumina
4 Johnson, Ft. PR, .R. Kurkchubasche, Ceramic," Marine Technology Society Confer-
J. D. Slachiw. 1993. "Structural Performance ence Proceedings, 1993.
of Cylindrical Pressure Housings of Different 9. Young, Warren C. 1989. "Roark's Formulas forCeramic Compositions Urder External Pres- Stress and Strains, Sixth Edition," McGraw-Hillsure Loading: Part I--lIostatically Pressed Company.
i
_~I!I
I-I
-- _.12 3
FEATURED RESEARCH
I GLOSSARY OD outside diameterPS pressureless sir'.erng
alumina aluminum oxide PSX CERCOM's prbssureless,3 ASTM Amencan Society for Tasting sintered grade of Si3 N4of Materials SF safety factor (factor of safety)SSN sintered silicon nitrideFEA finite element adalysis specific strength strength-to-density ratio
,FEM finite element models SRBSN sintered reaction-bondedModulus of Rupture flexural strength sil ion -ndesilicon nitrideID insioe diameter t thicknessL Length t/OD tnickness-to-outer-diameterIJOD lei igth- to-outer-diameter ratio ratio
UUV unmanned underwater vehicleMOI modulus of rupture W/D weight tc displacementNDE nondestructive evaluation ZTA zirconia-toughened alumina
IUI
II
-I
I[ I I• ., ' , n r m .m m ! . ,. • .• • . , ! ( -_! , " •!" • . . . - '3
U"ATURED RESARCH .
SRBSN PROCESSING SSN PROCESSING I
SI Metal PowderI SI Powderand Additive(s) and Additive(s)
Comminution Commimnution
and andIHomogenization HomogenizationI
Gh Injecion Molding b
Sl'p Castng Green Shaping
F Exzruclon uum
lLtrildng 3--enlficaon DenshlflUon I
Figure 1. Proce3s flow diagram for SRBSN and SSN fabrication. 3
-1I IS
-'', ,,, ' '" "" " "! II '1• ; -- ; ; ; ; ; ; i rI
FEATURED RESEARCH
Ioc rwsn Comniinutse~ andI Homogepwnize I
r C l-I msticShape Forming
Irun
Machinng [- Net hape
I CompneMtS-to
*Wn D ED
3 Quatity Assmo rdIMG
3 ~Figure 2. Fabrication process for PSX.
I 15
IF
,IRUBBER END CAP
SEAL RING
METAL RUBBER BAGMANDREL
II
Figure 3. Schematic of cold Isostatic-ptssing tooling set.
I1
16I
I-. FEATURED RESEARCH
UIII
UIIIIIIIU3 Figuro 4. Gold-welled, vacuum-nitrtdlng furnace shown with nltnded cylinder.
I* 17
____________________________________ ________
II
I*o
E Ia
iiiI&:10
II III
IwC.)
.1&
III
18 U
Z ~Fl.ATURED RESEARCH
IIIIIIIIIIIII
I Figure 6. Slrntered Si 3N4 cylinder prior to grinding.
II 19
FEATURED RESEARCH3
0 0
CD
Cc,
2C,
CD5
20o
FEATURED RESEARCH
4A
a t
-V'4
-- -01at 1
3 .. x..
21'2
I IiL
z3D
(I,
free INil M'it; W T
it9
22I
FEATURED RESEARCH
8/64" dia. flat-bottomed hole
"5/64 dia. flat-bottomed hole
I-- 2/64" dia. flat-bottomecd hole
IStandard was scanned using a 3M* damter,
3" fo" length. 10 MHz tansducer.Scan sp~g: 0.010"
3 Figure 10. Pulse-echo C-scan of Si3 N4 ceramic calibration standard used for
inspection of deliverable components.
IIIIUIIUII
*I2
j I I I II I I I I I I I I,, ll l I
, I
N IN
I .
II
14
; i t! iIiII ,,i
!a ji 3
S-ii ,in ~
4 ~46
-. .I2 I
I a 4, 4
* I Cti'l I-'Iii �a'U nol
if
U,U i-I
a'I ZnII 'C
I. II V
a A* IU. S
*rU. a r
I, a.
III _______
S
* ____________ aU a 25 I
- I 4 1U
ii' I0I-
I. Smol U
- I* I
I4 .1
I13I
* el UU
III- 3
* U0
II
S.
IU
0
* I* I 26 0 4 I
I a 4, a
II �'
-u I fl1!�J'-a
I _____ II a
- II III II ' *'�I�'.7*' T �
I t iiI �'ei5 3 is.
fa�IjjI� 2b
___ .11.1I jjg'Iv I
I a iI!� �a �j ,...
5I� N hII a
I* .111
�1
I,'* illUII U
a,- 0 U 4
U U I; I
I Iii �i � UI
ii USI a U:! h'p�' �i, a
I
i.E11 iiIW� WI. U
A KU
Ua
IU
Ig �. U� Ni INb'! II
0 Ua U 28 a U
""I I u U
I-i I il1 '
I !I
ulp
S ,oI •
La
tL4
U2
..,. .- ,,
I ii* -
j6 I111'! Iii '~Ii
II I ' I I I I I I I I I I I I I , • , . . .
, , J~ill.I11
* 0 A
Si I3
II III -
"II .I' . " +. • IIl.7,
*. 1'1*s At1 1 1
- I' -- I
l!I,•- 4II
Ii U
a qlb
I a I - -
I -"j
I
i Ij4, . .
' 'I
It-V
"Il a I.
Iil
IlI
343
I I II,..~ v1=121.iiii it
iL 0
al a 1§11
-- - h I flly IMISI~IIf0 cm,
11 bC
0 ~~ UB I . K
14 Id 4 44 4
35
a I, a 1n -
01 ________ VII Iif__________________ TV
ii' II I
a I
Icol
II* w N
�I
I I'a
1�3Nwfl
*
2
U-P IN N
IU
U __ ___ S
a 0 a C I
4--.
IVItI
37S
""= -v' ": t r ij j
He I
fe ItIII
,it
f la If 9- 1
I,,._,__ ! t I ! !
l I I
1i s I I I I I i II I I I
I I d
1
0
4 A a I1 i f I I L h'I 'M ix l IN! t ai b h~3
1 I 4 a C
* I CI I-* S
aola
* a
I* ml
I ** Iwo,I II.*
&
U.S
.2'
* U.
I p.
p.
II _______
0
3I a 0 a C
J;FI -:T
U I ii. I40
Si i ,Ii" Jj
III -i l]
L E
II.
U11 I
I
Nol
I!
I 4 1I 41 4
I I I I I I I I I I •1 I ,I, I ' I I tI ,I , IV ,
I 5I' a
S' iii I* ii .. m u..-
I !'!LiT: I'' I
* 0I L'I• i ii' :! '
* 'lj i
INB
*
1 6
I a N
* it
art4
* a43
IFEATURED RESEARCH i
III
iP
+'IS8 I
S~I
I..0
I
| in iin i' ! i 1" ]• =E
II r • "Pm m t U
IFEATURED RESEARCH
IIUUIUIIIIUUIU Figure 28. Photo of �UI1y machined SI�N� �eram1c cVlinder with titanium end rings bonded in p�aoe.
IU* 45
WTURERMMI
77I-----.... ....
Figre 9.Endca jontrin rmovl ixtre
46I
wo Isc FEMID
I0
z
0 zO~
I30. PreSsure VS. stmn" P.lot tor cytinder 001.
FEATREDRESEARCH
Fiur 3.Cicufretilcrckn o teberigsutae tcyine 01.Te itnumen-apjIarin ws rmoedaftr resur tstig~o nspctfo crckng
48I
FEATURED RESEARCH
Uz
to ni
I
UJ
W:DIM
49
49
FEATURED RESEARCH I
Ii
8 I' I
Goo
a.
9ILD
CD
Z L)
MCkqI I
FEATURED RESEARCH
I c
n~EJ
I
IL
FEATURED RESEARCH
C,,Dco 1 (DI
wIcrU
Z U 0
z uv Ulqc z
14 < a
Nn IMb I)
zt) o O52
IFEATURED RESEARCHI o 4 -• .,,.., . . . . . .
04
-2%,000 Axial Strains
000 _______ *4
-4,00 -~
Test Configuration I Hoop StrainsI __ _ _ _ __ _ _ _ _ Stan0 4,000 8,000 12,000 16,000
External Pressure (psi)
3 Figure 34. Pres-ure vs. strain plot for cylinder 002.
0
3_Axial Strains
I NIi ~-4,000 L
Hoop Strains
Test Configuration I
-6,000 1, ._ _ _ _,
0 4,000 8,000 12,000 16,000Exterms: Pressure (pMl)
3Figure 35. Pressure vs, strain plot for cylinder 003.
* 53
INIFEATURED RESEARCH
Axial Strains
-2,000---_ _ _
C tioop Strains
-6,0001
04,000 8,000 112,000 16,0001
External Pressure (psi)I
Figure 36. Pressure VS. strain plot for cylinder 004.5
0
0 4,000 8,000 12,000 16,0003
External Pressure (psi)I
Figure 37. Pressure vs. strain plot for cylinder 005.
54I
FEATURED RESEARCH
0
I Axial Strains
-2,01)0t'*
I Hoop Strains
I _ _ _ _ _ _ _ __.
-4,0001--
(ps) 1,00
04,000 8,000 12,000 16,000
External Presulre (psil)
3 Figure 38. Pressure vs. strain plot for cylinder 006.
*IIIIb&Ng Axial Strains
0 4,000 8airu 12,000 16,000
External Preisuke (psi)
I Figure 39. Pressure vs. strain plot for cylinder 007.
SI 55
FEATURED RESEARCH
Hoop Strains
-2,0!n xilSran
IIAxial Strais~n
-4,000
0 4,00 ,000 12,000 16,000 3External Presure (psi)
Figure 40. Pressure vs. strain plot for cylinder 008. 3
AxiAl Strains
-2,O *" Hoop Strains
-S,000I
Prs.sutton Nom I
LL0 5,000 10,000 15:000 20,000 25,000
Extarnal Prpssure (psi) 3Figure 41. Pressure vs. stra!n plot for cy!inder 009, pressurization No. 1.
56 _
UFEATURED RESEARCH
3 oAxial Strains
I _ _60
U ~~-3,000--
Pressurization No. 2
"-7,00 -Hop Strains 111
0 5,00o 10,000 15.000 20,000 2,000
External Pressure (pal)
3 Figure 42. Pressure vs. strain plot tor cylinder 009, pressurization No. 2.
I j "• • "T•,' ,.Axial Strains
P5000l ji ooo -
3 Fressurlzetion No. 1
I-7,SO ... . . .... . ,0 5,000 10,000 115,00 20,000 25,000
External Pressur (psl)
IFigure 43. Pressure v. strein plot for cylinder 010, pressurization No. 1.
* 57
IFEATURED RESEARCH 3
0
Axial Stl.I i. I
S0 S,0o0 10,000 15,O000 20,oWo 25,00
ExMMSI PrOSUre (PSI)
Figure 44. Pressure vs. ctrain plot for cylinder 010, pressurization No. 2. i
Ill
10,00 . MMfuaky noorA I GOll W PSIJ
.0000
I
Figure 44. Pre) sur, s. setrainupot for cyl-6 nd6-erc00,pentauriz~na No.mi 2.3
\ %
N.~~98 N%3.Sly O*
40,0000,00ps
0.2 0.3 0.4 0.5 0.6 0.7 0.6
PSX ceramic.58
FEATURED RESEARCH
Table 1. Typical properties for PSX commercial grade of silicon nitride.
Bulk Density 3.29 u/cm'
3Flxural Strength (MOR) 703 MP*o RT, 4 point average 102 ksl
Characterist•lc Strength 744 MPrRT, 4 point 108 ksl
Welbull Modulus (W) 14
Elastic Modulus IE) 317 Gpa48 Mpsi
3 Polion's R2tlo Iv) 0.24
Hardness (Knoop I kg) 1480 kg/mmn
Fracture Toughness (Viokers) 8.0 MP&m14
3 Thermal Expansion (RT-10006C) 3.3 Io-/st.
Then-,ll Conductivity (RT) 30.0 W/lrOK
Thermal Shock Resistance (ATJ) 700 OC
U iTable 2. Actual dimensions of deliverable PSX cylinders.
NO.n der No r Ou.tr diamete Wall thickness Lenglh
Sist]4 001 12 0.412 18
"" S.gN4 00? 12 0.412 18
SiN 4 003 12 0.412 18
SSi3N, 0U4 12 0.412 18
SiN 4 005 12 0.412 18
3 Si5N4 006 12.077 0.418 13.0
S13N, 007 121053 0.413 16.0
SIIN, 008 12.070 0.416 17.938
SI2N4 009 12.058 0.415 18.0
SiN,. 010 12.041 0.406 18.0
U 59
FEATURED RESEARCH
-L 0 In i00
V- "It* *g *i q g ~
M 212 -d -
~~"It-s - - - -
40
t -: V- I0 ". K C;
8 8I
0600
0o ooO~~
-FEATURED RESEARCH
Table 4. Summary of findings made in PSX cylinders using pulse-echo ultrasonic inspection.
Cylinder Number of indications lndication~DepthNO. (return signal smplltud.'/nches belowU surface)
S13N4 002 4 57% @0.3; 59% 00.12; 69% @0. 12;
S13N4 003 4 43% @0.10;4C-% @0.2; 62% @0.3;______ _____________63% @0.1IS
S41N4 004 4 21% @0.3; 24% @0.2; 32% @0.2; 42%____ ___ __ ___ ___ ___ ___ _ _ ___ ___ @0.2USL 3N4 005 4 66% @0.12; 83% %30.1; 92% @0.13;
______ _____________4dB @0.13
SINo 006 9 47% @0.15; 50% @0. 15; 51 % @0.2;51% @0.3; 54% @0.2; 56% @0.2; 58%
______ ____________ 00.3; 62% @0.2; 80% @0.25
S6N4 007 13 40% 00.35; 48% @0.2; 49% @0.3; 53%@0.2;:55% 00.15; 59% 00.3; 67%@0.2; 79% @0.1; 80% @0.15; 80%@0.15; 90%@Q0.1; 60%@0&3; 93%
S13N4 008 17 47% @0035; 52% 00.3; 54% @0.2; 5"%@003; 61% @0.3; 62% 00.35,77%U @0-15; 78% @0.3; 812% @0.3; cluster
up to 85% @0.1-0.3; ldB@002; [email protected];,2dB @0.2; 3dB @0.3; 4dB
______________ @0.3;6dS @0.15; 8dB @0.3Sq6N 4 COOP 22 30% @0035; 45% @0.23; 65% @0.2;
65% @00.15; 70% @00.15: M0% @0.2;U75% @0025; 75% @0.2.76% 00.2; 80%@003, 80% @0.2; 80% @0035; 83%@0.15; 85% @0.3; 90% @0.20; 90%3 @0015; 90% @0.Z0, 95% @0M15; 95%
00.3; 25% @0.15S12N, 1 10 24% 00.3; 26% @0.3; 38% @0.2; 42%
@0.3; 56% @00,15; 62% @0015; 70%1 @0.35; 74% @0.25:7dB @0.2; 28%
flat-bottom drille to. Indtauons whch ame thwomticaly Wger thn 1W~ din are shown bold.
* 61
FEATURED RESEARCH
12 "m Ic--c•8• I
o -I
.08
--- - - -
+. +
0. 1p
Icc + .I ...L 0 L i
)CL > J? I0e 2F F ;e;e ;
62
FEATURED RESEARCH
Table 6. Summary of test plans and results for cylinders 001 through 010, Sheet 1.
CflIndet No. Test Test Plan/rTest ResultConfiguration
S6%001 II Proof laes cylinder to 16,000 psi, read strains, Cyclecylinder to 16,000 ps, mop lan afor 2.000 cycles.
i-proof cycle to 16,00 ps. Cylinder wtMtood Wcyces to 16,000 pol. Testing was terminated due to
________ __________leakage.
M81A 002 11 Proof test cylinder to 16,000 psi, mad strains. Cyclecylinder to 16,000 psi, Nap test!Ng after 2,090 cycles.
1 proof cycle to 16,010 psi, no damage noted. CyldiUa on cycle numwer 327 during presurlzatlcn.
51A 003 11 Proof tool cylinder to 13,000 psi, mad strains. Cyclecylinder to 13.000 psi. stp tesingafer 2.000 cycle._I 1-proof cycle to 13,00• pi, no damage noted. Cyeear3elld ore cycle "number 17 9 dn pressurizalon.
'0 4 004 Proof test cylinder to 9,000 psi, reW strains. CycleSn to 9,000 psi. stop tsin after 3,000 cycles.I1-proof cycle to 9,AW pal. no damage noted. Cylinderwwlo 3X ,cycles to ,900 psi w~how vilable
. damage.
oo4 005 11 Proof teas cylider to 12.000 psi, read strains. Cycleclne ID 12,000 pa. Mop !!L afe 3.000 cycles.
I •-iCW- eye to 1120110 psi, no damae noted. Cykvtofled on eyc number 743 during premsr, Iotn.
6i
I
I
I
IFEATURED RESEARCH 3
Table 6. Summary of test plans and results for cylinders 001 through 010, Sheet 2. 3
Configuration
SIN 006 II Proof test cylinder to 12,000 psi, read straino. Cyclesand-basted ends cylinder I* 12,00 psi,. sop testing after 3,000 cycles.
Iprof cycle to 12.000 psi, no damlae noted. Cylinderwithstood 2,241 yclas to 12.MO pal wOtW viable
0 danmpe.
SI^00711Proof lost r,ylnr to 11.000 pal. read strais. Cycle"san-blatsted ends cylinder to 11,000 psi, stop tesing after 3.000 cycles.
1- o cycle to 11.000 PSI, no damage noted. Cylinderwithsood 3,010 cycles to 11,= pal without visbledamage.
SiN, 008 II Proof test cylinder to 11,000 psi. Cycle cylinder to 11,000sani-blasted ends psi, sMop esting after 3,000 cycles.
1-proof cyele to 11,000 pal. No damage noted. CylinderMW on cycle nunmer 1.665.
SVN. 000 i11 Proof tNet cylinder to 10.000 psi, read strains. Thensand-blasted ends pressudze to failure while reading strains.
I-proof cycle to 10,000 psi, no damae noted. Cylnderwas p•rmIzed to faiure w occured at 22,600
%A 010 i11 Proo IeM cylinder to 10.000 psI, read strains. Thansaan.blested ends presaurze to failure whie reading strains.
1-proof cycle to 10,000 psi, no damage noted. Cylinderwas presunzed to bib" whdch occured at 21800
IIII
II
64 I
FEATURED RESEARCH
Table 7. Comparison of cyclic pressure testing performance of PSX Silicon Nitride cylin-ders and of AL-600 96-percent alumina-ceramic cylinders.
External Maxinmum nominal Number of pressurizationshydrostatic hoop stress CronPX Wso L0pressure (psi) (psI) Corcom PSX Wesgo AL600
SIN 4 96% A1203
cyinder cylinder9,000 131,067 withstood 3,008 withstood 3,000
10,000 145,631 "_i
11.000 160,194 withstood 3,000 withstood 1,380failure at 1,665 failure at 2,963
12,000 174,757 withstood 2,241 failure at 1,065failure at 743,
12,500 182,039
13,000 189,320 failure at 1,379 failure at 762
14,000 203,883 failure at 214
15,000 218,447 failure at 707
16,000 233,010 Leakage at 581failure at 337
I
I
III 8
' I II I I ' I i i i' :. . .
FEATUREUi RESEARCH ________________________
WUZo -70 CD
cv
.2 A
0
S 03
0w I
LU (D
SL U
S0
CD)
.0. Lo
06£
,Q mE
1 C
66U
IFEATURED RESEARCH
I THE AUTHORS
I RAMON R. ing from the University of California at Santa- KURKCHUBASCHE is Barbara in 1986, and an M.S. in Structural Engi-
a Research Engineer neering from the University of California, Sanfor the Ocean Engi- Diego, in 1991. He is a member of the Marineneering Division and has Technology Society and has published "Stressworked since November Analysis Considerations for Deep Submergence1990 in the field of deep Ceramic Pressure Housings," Intervention '92, andsubmergence pressure "Structural Design Criteria for Ahumina-Ceramichousings fabricated from Deep Submergence Pressure Housings," MTS '93ceramic materials. His Proceedings.education includes aB.S. in Structural DR. JERRY STACHIW isEngineering from the Staff Scientist for Marine
University of California at San Diego, 1989; and Materials in the Cceanan M.S. in Aeronautical/Astronautical Engineering Engineering Division.from Stanford University in 1990. His experi- He received his under-ence includes conceptual design, procurement, graduate engineeringassembly, testing, and documentation of ceramic degree from Oklahomahousings. Other experience includes buoyancy State University inconcepts utilizing ceramic, nondestructive 1955 and graduateevaluation of ceramic componerns. He is a mem- degree fromber of the Marine Technology Society, and has Pennsylvania State Uni-published "Elastic Staibility Considerations for Deep versity in 1961.Submergence Ceramic Pressure Housings," Inter-vention '92, and "Nondestructive Evaluatinn Tech- Since that time he has devoted his effc; Ls at vari-niques for Deep Submergence Housing ous U.S. Navy Laboratories to the solution of chal-Components Fabricated from Alumina Ceramic," lenges posed by exploration, exploitation, and
MTS '93 Proceedings. surveillance of hydrospace. The primary focus ofhis work has been the design ana fabrication of
A Ppressure resistant structural components of divingSRICHARD .r JOHNSON systems for the whole range of ocean depths.Ii is an Engineer for the Because of his numerous achievements in the fieldOcean Engineering of ocean engineering, he is considered to be theDivision. He has held leading expert in the structural application of plas-this position since 1987. tics and brittle materials to external pressure hous-Before that, he was a ings.Laboratury Technician
<7 for the Ocean Engi- Dr. Stachiw is the author of over 100 technicalIJneering LaboraTory, reports, articles, and papers on design and fabrica-University of California at tion of pressure resistant viewports of acrylic plas-i/-"Santa Barbara from tic, glass, germanium, and zinc sulphide, as wetll
/ 1985-1986, and Design as pressure housings made of wood, concrete,Engineer in the Energy glass, acrylic plabtic. and ceramics. His book on
Projects Division of SAIC from 1986-1987 His "Acrylic Plastic Viewports" is the standard refer-education includes a B.S. in Mechanical Engineer- ence on that subject.
II
FEATURED RESEARCH
For the contributions to the Navy's ocean engi- Technology Codes Outstanding Performance Cer-neering programs, the Navy honored him witn the tificate.Military Oceanographer Award and the NCCOSC'sRDT&E Division honored him with the Lauritsen- Dr. Stachiw is past-chairman of ASME OceanBennett Award. The American Society of Mechani- Engineering Division and ASME Comrnittee oncal Engineers recognized his contributions to the Safety Standards for Pressure Vessels for Humanengineering profession by election to the grade of Occupancy. He is a member of the Mar;neLife-Fellow, as well as the presentation of Centen- Technology Society, New York Academy of Sci-nial Medal, Dedicated Service Award and Pressure ence, Sigma Xi and Phi Kappa Honorary Society.
III
I;
I
F',,I'I
"ic, ropywli burcien w thi =wwol 01 Inaboatn is estimated to average I flour cew reycoomse. inctuc.flg the time for reviewinlg instructions. sacige-trgdt OC ahrn nfratsia'ito the data needed. and complteing and reviwing he colttection of sfli09That.O Send comr. writs regarding this burden' osttmatc or any oth itr asMc' of11 this colecio o9 ifllormition includingauggeebnnu k, reucng emis urde.'. to Washingto ileadquarteri Services. Directorate tor Infornatijor Operations and Reports. 11 of-o,03Hgwy ut 24 rigo.V22202-4302, end to the o~mee 01 Managemetad ugl Paperwork. Reduction Protect (0O0O-0l88). Washington. DC 20503
1.AGENCY USE ONLY (Loaam 0(111) 2 REPORT DATE 3 REVORT TYPE AND DATES COVERED
June 1994 Final
4 TITLE AND SUBTITLE 5 UNDING NUMBERSI STRUCTURAL PERFORMANCE OF CYLINDRICAL PRESSURE HOUSINGSOF DIFFEREN4T CERAMIC COMPOSITIONS UNDER EXTERNAL PRESSURE PE. 06037 13NLOADING PROJ: S0397Part III, Sintered Reaction Bonded Silicon Nitride Ceramic ACC- DN302232
6. AVTHlOP(S)
R. R. Kurkchubasche, R. P Johnson, J. D. Stachiw
7. PERFORMING ORGANIZATION NWtE(S) AND ADDRESS(ES) a PERFORMING ORGANIZATION
Naval Command, Control and Ocean Surveillance Center (NCC'OSC) REPORT NUMBER
RDT&E Divisien R19SnDiego, CA92152-5001
AGEN( Y REPORT NUMBER
Naval Sea Systems CommandWashington, DC 20362
1t SUPPLEMENTARY NOTES
1 2a. DISTRIBUTtONAVAJLAB9U1-TY STATEMENT 120) DISTRIBUTION CODE
Approved for public release; distribution is unlimited.
1ABTRAC (Ma2 00 words)b 18-inch-long reaction-bonded sintered silicon-nitride cyliniders were fabricated, aseembled, and
pressure tested to determine their suitability for use as exverna.. pressure- resist~ant housings for underwater applications.The material, designated PSX Si3N4 , is made by CERCOM, Inc. (Vista, CA) by a proprietary process. The primary advan-U tges of thsis material are high compressive strength, high elastic mcduius, high fracture toughness, and low specific graivity.Pressure test results are presen,,ed along with strain gage data and cyclic fatigue life u'ata. Conclusions regarding the stiit-ability of the material for application to pressure housings for under-water applic.ations are presented along with a compari-son to WESGO's AL-600 96-percent aluminA ceramic, which was chosen as the base of comparison for various ajvancedmaterials being evaluated under the same programs. Recommendations for design implementation, nondestruc-ive inspec-
tion, and further research are made.
lt. SUBJECT TERMS t5. NUMBER OF WA~3S
ceramnics 7
external pressure housing 16 PRICE CODE
ocean engineering17. SECURrTN CLASSIFICATION 18 SECUAITY CLASSIFIC.ATION 19. SECURITY CLSSIFICATtON 20 UMrTATtGjN OF ABSTA~cr
OF REPORT OF THIS PAGE OF ASST.,~
UNCLASSIFIED UNCLASSIFIED IUNCLASSIFIED SAME AS REPORT,
NSN 7640-1-2M05800 standard! fotm 296 (FRONT)
UNCLASSIFIED21& NAiME O REOINIMBLE INONQUAL I21b.TELEPHONE (urcieu. C0v) 21c OFFICE SYMBOLR. RI Kurkchubawche (619) 553-1949 Code 564
I
I __
II
QW-mw %xar So sn@ e MACIC)
IUNCLASSIFIED
I m I I " lr • I W "•l I j " ' '• " ; ...
