AMMRC TR 71-59 1AD 1 .

"| OXIDATION BEHAVIOR OF A

COMPLEX DISILICIDE COATING-

STa-10W ALLOY SYSTEM AT

TEMPERATURES OF 1700 TO 2700 F

MILTON LEVY and JOSEPH J. FALCOMETALS DIVISION

December 1971

Approved for public release, disti ibution unlimited. L J L:1.NATIONAýLTECHNICALINFOPMATION SERVICE

ARMY MATERIALS AND MECHANICS RESEARCH CENTERWatertown, Massachusetts 02172

I ITI 1YRI •+ET I• s ,"

a BUFF 21001

T FIC........

k...... .......oE. . S...........

Div

3, ,

The findings in this report are not to be construed as•• an official Department of the Army position, unless soI designated by other authorized documents.

S' :' Mention of any trade names or manufacturers in this reportshall not be construed as advertising nor as an official

/.'• indorsement or approval of such products or companies bythe United States Government.

I.D SPOSITION INSTRIUCTIONS

Destroy this report when it is no longer aepded.SDo ,,ot return it to the oriettostor,

UNCLASSIFIEDS. cuntl Classification

DOCUMENT CONTROL DATA- R & D(Security cltosilic&.,• of ti.l, body .1 of .trar, ,nd inde.xi.9 anwoln.gi Musuat be e,,naed Whe,,th overall fqrpor, It 080e0011)

I ORIGINA'ING ACTIVITY (Corporate autho) &. REPONT •ACiUNITY CLASSIFICATION

Army Materials and Mechanics Research Center UnclassifiedWatertown, Massachusetts 02172 8. GROUP

S•E PORT T-TLV

OXIDATION BEHAVIOR OF A COMPLEX DISILICIDE COATING - Ta-IOW ALLOY SYSTEM ATTEMPERATURES OF 1700 TO 2700 F

4 OESCR#1PTIVOL NOTES (7rpe of-spopt and Inclusive do#**)

S. AU THOR($) (Fftrat nam, mIdlde Inl1•te, rae0 nowm)

Milton Levy and Joseph J. Falco

6 REPORT DATE 7t. TOTAL NO OF' PAG, ' 7b. NO Oft ••FS

December 1971 19 8,. CONT rRACT OR C11r-T NO S. On,,INATORS REPORT NUMalU,•sl

6. PROJCTNo. D/A IT062105A328 AMMRC TR /1-59

S.AMCMS Code 502E. 11.294 Sb. OTHER REPORT NOS, (Any o6,.,number.s ,. A.o be &*dlo-

'. Agency Accession No. DA 0477010 OISTRIOUTICIN STATEMENT

Approved for public retease; distribution unlimited.

II. FUPPLEMKNTA1Y NOTES '. SI'ONSORING MILITARY ACTIVITY

Details of illustratio;,s in U. S. Army Materiel Commandthis document may be better Washington, D. C. 20315

IS. ASS711ACT 810 ie un micW fu'i4I

The oxidation behavior of t complex disilicide coating - 90Ta-1OW alloy systemwas studied utilizing thermogravimetric, X-ray diffraction and electron microprobeanalyses. The coating afforded the alloy substrate complete protection againststatic oxidation for at least 200 hours at temperatures between 1700 to 2700 F. Theexcellert oxidation resistance of the coating system i3 due to the formation of aprotective layer whiý*h consists mainly of Si02 with some Ti, V, and W in solutionand a second dispersei phase of TiC2. While this protective layer is glassy at andabove 2100 F, the rea:tion products become increasingly crystalline with decreasingtemperature between 1700 and 2000 F. (Authors)

Pow 146PLACES F0l ' ORM 1474. 0 JAN 4,-1, 4C" I0

D DII...OUT.Y73 D,. UNCLASS!11'1 I

UNCLASSIFIEDSecurity Classifcdt1on

K~v wOROS .

ROLE W ROLE A foO. E

OxidationHigh-temperature researchProtective coatingsRefractory alloysTantalum allcysDisilicidesThermal analysisX-ray diffiactionElectron microprobe

AMMRC TR 71-59

OXIDATION BEHAVIOR OF A COMPLEX DISILICIDE COATING -Ta-1OW ALLOY SYSTEM AT TEMPERATURES OF 1700 TO 2700 1-

Technicai Report by

MIL TON LEVY and JOSEPH J. FALCO

December 1971

D/A Project 1T062105A328 -, "AMCMS Code 502E.11.294Metals Research for Army MaterielAgency Accesion No. DA OE,4770

Approved for public release, distribution unlimited,

METALS DIVISIONARMY MATERIALS AND MECHANICS RESEARCH CENTERWatertown, Mas:schutetts 02172 ,,,

ARMY MATERIALS AND MECHANICS RESEARCH CENTER

OXIDATION BEHAVIOR OF A COMPLEX DII ILICIDE COATING - Ta-lOWALLOY SYSTEM AT TEMPERATURES OF 1700 TO 2700 F

ABSTRACT

The oxidation behavior of a complex disilicide coating - 90Ta-IOW alloysystem was studied utilizing thermogravimetric, X-ray diff-action and electronmicroprobe analyses. The coating afforded the alloy substrate complete protec-tion against static oxidation for at 'east 200 hours at temperatures between1700 to 2700 F. The excellent oxidation resistance of the coating system is dueto the formation of a protective layer which colsists mainly of SiO2 with someTi, V, and W in solution and a second dispersed phase of TiO,. While this pro-tective layer is glassy at and above 2100 F, the reaction products become in-creasingly crystalline with decreasing temperature between 1700 and 2000 F.

CONlTENTS

Page

ABSTRACT

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . I

M, rERIALS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I

EXPERIMENTAL .. .. .. . . . . . . . . .. . . . . . . . 3

DISCUSSION OF RESULTS ............ ....... .......................... 3

1700 to 1800 F.... . .... ............. ......................... 31900 to 2000 F . . . . . . . . . . . . . . . . . . . . . .. . . 42100 F .......... ............. ............................... 62200 to k0O0 F ............ ......... .......................... 62500 to 2700 F ............ ......... ... ...................... 9

DIFFUSION OF COATING AiND SUBSTRATE ELEMENTS ....... ...... .......... 11

EFFECT OF HEATING RATE ON THE OXIDATION OF THE COATED AL',OY.... . 13

CONCLUSIONS .................................. . 15

FUTURE WORK ....................... ............................ IS

LITERATURE CITED ........... ....... ........................... .... 16

I

ItNTRODUCTION

The Army has propulsion systems waterials requirements for second generationgas turbine engines which include turbine blade, disk, and vane materials: corn-bustor liner, and duct materials. For the required work extraction (for every100 F rise in engine operating temperature there is a corresponding increase of12% in engine performance), turbines must operate at inlet temperatures between2300 and 3000 F.

Refractory alloys of tantalum and particularly columbitum are candidate mate-rials for such applications, provided coatings can be developed to protect themagainst high-temperature oxidation. One of our efforts is directed at character-izing -existing promising coating systems to provide data for further optimizationof these coatings.

Pure and alloyed sili1iide coatings offer some of the most promising meansof protecting refractory metals from high-temperature oxidation. However, inthe intermediate temperaturc range (1400 to 1600 F), they fail due to pesting.It has been proposed that pesting essentially involves a type of intercrystallineoxidation of a silicide in which each grain eventually becomes sarrounded byreaction products. Pest.ng is not only common in silicidps, but also in otherijitermetallic compoLmds 4-id leads to the rapid disintegration of the bulk material.

Ile have evolved a simple method for alleviating pesting in a modified sili-cide coating for a Ta-101 alloy, i.e., a preoxidation treatment at 1900 F. Ourstudy of the mechanisms of failure and protection of a modified silicide coatingfor a Ta-lOW1 concentrates on the oxidation behavior of the coated alloy abovethe pesting range, namely 1700 to 2700 F.

MATERIALS

A swaged and annealed Ta-101 alloy rod was machined into disks 0.5-inch in•diameter and C'.12S-inch thick. The disks were coated with a modified silicideby Solar,* employing the following two-step process:

1. A slurry consisting of 50W, 20Mo, 15V, and 1STi (in powder form) wassprayed onto the disks. The specimens were then sintered at 2780 F for 15 hoursunder a partial prcssure of 10-5 Torr.

2. The coated specimens were packed in pure-200-mesh silicon and were sili-cided for 15 hours at 2150 F in an argon atmosphere under a pressure of 800 Torr.The coating thickness resi1ting from the two-step process was 7 to 8 mils persurface.

The coating system is a duplex one, having a porous outermost layer (reser-voir layer) and a diffusion layer as shown in Figure '1. Electron microprobeanalysis (data which accompanies this photomicrograph) shows that the light

*Division of International Harvester Company, San Diego, California, U.S.A.

WW

areas of the reservoir layer contain mixed disilicides. At Points 1, 2, and 3,Mo Si 2 predominates, whereas If Si, is the main constituent 50 Microns from thesubstrate (Point 4). Although the outer layer is rei-rred to as a porous reser-voir layer, there are isolated areas that contain SiO, and TiO2 as shown inFigure 2, Points A and B (Si02 is present at Point A, while SiO2 and TiO2 arepresent at Point B). This is probably due to the introduction of oxygen as animpurity in the powders during the coating processing.

Wag. 200X

1 i Distance Iron,Point Identification SI Ta V i Mc Ti V Substrate

1_2 1 .Mo So 40.6 <1.1 0.1 44.5 &. 60 235JU2 AoS 42.3 <0.1 4.0 35.4- 7.4 10.4 155yJ3 h1o Sti 40.4 <D1 2.2: 37ý3 9.5 10.5 8 5A14 W Si2 30.7 <b.1 56 0 4:3 4.2 5. 50;5 (Ta. WI 'S' 25.3 48.1 19,3 1'S 28 1.8 25116 (Ta, WI S•2 22.3 58.1 16A4 0.3! 2.2 0,9 10l/

S7 10.1 81.5 8.0 <0.3 0.o, Co t-,,,t w - Interface

Distance from5:6 Coating78 Alloy 0.08 89.5 10,5 <0.3 <0.01 <I0.01 lOp

'fo 9 Alloy <0.05 89.5 10.5 <0.3 <0.01 <41.01 20P110 Alloy <0.05 89.5 10.5 <03 <0.01 '<0.0. 30)J

1 L

Figure 1. MICROPROBE A;\.ALYSIS OF AS-COATED REFRACTORY ALLOY19.066-140/AMC-70

Point Identification Si Ta W Mo Ti V Area

A SiO2 20-25 <0.1 1.8 <041 2.1 I .4 Coating Oxide"8A 1O2 1TO 20.25 <0.1 6 <0.8 jI) 214.1 < .1. Coating Oxide

Figure 2. MICROPROBE ANALYSIS OF TYPICAL OXIDES IN THE"*, .', , AS-COATED REFRACTORY ALLOY

"19-06. IllAMCy7O

2

The derse diffusion layer consists mainzy of tantalum and tungsten disilicide(Points S and 6). It is interesting to note that only small anounts of silicon(0.080) have diffused into th,, substrate. Not shown in this photomicrograph, butcharacteristic of silicide coatings, are microcracks which extend from the dh-fusion layer into the reservoir layer. Trhese microcracks are due to -he mismatchbetween the thermal expansion coefi'icients of the coating and the substrate.

EXPERIMENTAL

The Mettler thermogravimetric apparatus describee elsewhere1 was used forobtaining oxidation rates of the coated alloy as a function of time and tempera-ture. 1n',estigations were carried out at atmospheric pressure at an air flowrate of 50 cm/sec.

X-ray diffraction analyses were made with a Norc Ico X-ray diffractorieterusing a copper target operating at 35 kV and 20 mA. A two-theta (2-) szan wasmade from 15 to 1100 at a scanning speed of 10 nev mizaute, A 10 divergence slitwvith a nickel filter to eliminate the Ks radiation was used.

The diffracted radiation -;as recorded on a strip chart at a speed of 30 inch-es per hour. All variables of the unit wer" held constant so that a Semiquantita-tive analysis of the compounds present at various temperatures could be made f,'ortpeak heights and widths. Observed d-spacings were compared with known d-spael sz

from ASTM X-ray-diffraction index cards.

Electron microprobe analysis was performed wth an ANIR oelctron beam micro-analyzer. The accuracy of the concentrations obtained from the point analysis i,1generally of the order of ±5% of the amount reported. The analytical precisionvaries slightly depending upon the element studied, but is gonerally within -2',relitive for concentrations ranging from 10 to 100 wt %. The precision is t5"re'ative for concentrations ranging from I to 10 wt % and ±10c for conccntratiinranging from 0.1 to 1.0 wt %.

DISCUSSION OF RESULTS

For discussion purposes, results of oxidation-rate studies, X-ray diffractioni,and electron-microprobe analyses have been conveniently grouped according to thefollowirg temperature ranges: 1700 to 1800, 1900 to 2000, 2100, 2200 to 2400,and 2500 to 2700 F. Thus, identification of reaction products and their topologywere related to changes in oxidation kinetics.

1700 to '1800 F

At 1700 to 1800 F, Figure 3, oxidation proceeds parabolically for the firs•tfew hours after which it exhibits no further weight gain. This indicates thatthe protective film formed over the coating during the initial oxidation ismaintained throughout the entire run of 200 hours. Figure 4 is a photograph ofthe coated allo:.y oxidized at 1700 F for 20 minutes in air. A cracking patternis o'.bserved in the lighter areas. It has been amply demonstrated that largestresses are present in metal-oxide syste-s during and after oxidation which can

result in cracking. Many factors could be operative in the forg:ation of cracks,e.g., ratio of the volune of equivalent anounts of oxide and retai, variationsin growth rates of different layers in na.ltilayered scales, differences in ther-mal expansion of metal and oxide, etc. Additional expericents were carried cut-At 1700 F in which air was replac 4 by nitrogeo- in order to establish the roleef oxygen in the cracking process. Under these conditions, neither cracking nor.Vight gain were observed, suggesting that the formation of an oxide was neces-"-ary for cracking to occur. The mechanism is r.ost likely volumne expansion due to)Xidation rather than difference in thernal expanson or reaction with nitrogen"'ich is in agreement with the work of Hiaym'an and Restall. Regardless of the

.actors ,hich qnre operative, the presence of cracks does not necessarily lead tofailare of the coated specimen. The darker areas of the photograph sh,"ow thatself-healing of the zracks is occurring. Electron mficroprobe and X-ray diffrac-tion analyvis show in Figure 5 that this outer protective layer contains Sin-and TiO, fl(o~nts A and B) as well as an unknown com~pound having d-spacings 3.108,.. 735, and 1.93S .

The oxides at Point A appear to be denser than Point B (based on totalight percent of elements), suggesting that self-healing initiates at the outer

"--face cf the coating and progresses inward. Since it is difficult to forn a,tvctive oxide which remains absolutely sound under all conditions, the sc-If-iling capacity (ability to regener,1te protective oxide in case of surfaceicking) is importalt. Thiis capability depends on the preservation of an active

,:xservoir of disilicides in the coating which is identified as Wi, Mo, Ti, V).: at Points 1, 2, and 3. Points C and D are predominant!y SiO-, indicating

"%at Ti is diffusing to the outer layer of the coating where it is present asiiO2 . Scale areas of tile diffusion layer of the as-coated material [containing

ra, W) Sij have become oxidized as shown in the analysis at Points G and 1i.'his has occurred only at those sites in the diffusion layer wh-ere cracks which

we formed are broad enough to act as diffusion paths for oxygen (see Point E).dthough oxidation of the diffusion layer is undesirable, failure has not occur-red because the self-healing capability of the coating system prevents furtheroxygen di ffusion.

1900 to 2000 F

Figure 6 contains oxidation curves at 1900 and 2000 F where the initialsegments of both curves ar- parabolic followed by linear portions. At 1900 F,the linear portion represents no further weight gain; while at 2000 F, there isan increase in weight. In both cases, complete protection is afforded the sub-strate alloy for 200 hours.

Between 1700 and 2000 F, the initial parabolic oxidation rate increases withincreasing temperature. flowever, if one compares the secondary linear portion ofthe curves only, the weight gain at 1900 F is less than at 1800 F. This anomajymay be explained with the aid of Figure 5. The photomicrograph (coated alloy oxi-dized at 1800 P) is not characteristic of the entire coating but is representativeof the few sites which exhibited diffusion-layer oxidation. In this case, thecrack is much broader than the microcracks observed in the as-coated condition.A volume increase occurs as a result of the localized oxidation of the diffusionlayer at the base of the crack which creates internal stresses and a rarid broad-ening of the crack. However, furthrr oxidation at these sites is totally arrested

a 1800 F

o w170DF OF!

F'pre 3 OXIDATION CURVES OF AE C-O-----------------4 OATED REFRACTORY ALLOY IrN THE.5 TEM'PERATURE RANGE OF 1700 Ir TO

^100 F-

20 40 60 -40 80 120 160 200S-Minutes - Hours -

Time

Figure 4, COATED ALLOY OXIDIZED AT 1700 F FORT 2 O MINUTES M14g. 5X.•. =-,•19-066-929/tAMC.69

Point Idcrit.ficatpon Sd Ta J T ov D a. 4*Ai I w r~.iVSf: 402 04 293. aO 641]74 214

S2 (W Mo,T, V\I5, 375 0414611120 54 t 1703 ,W.Mo.T, VIS,' 271 0(6 415 15.2 48 108 100

' 4 Z")" Allay '.o-.o 83 5 97 (03 3 5 - to.S 5 Alloy ýo0 05 88 O ioo C3 07 10 .o :161• Aily 0( 06 R9 10 1040006 30.7 ,o j o 31oo ,0 o o1• 1 '4

Se4 AS. O 40 3I1O 01' OfIt t

OO

0: 'Ov C ?0 Aj 60 1 1 1 1D ,i0: 311.401 02 i 1 Iv 3 ' 4 7,-E Vo16 I G 01. 41 j 1 00t 03 01 -F S0: 2 0 .5 021 -10, 01 U ,G Mi xed ,.Jc 14 0 1225[?12 1

1 i 38H 01x~. d R ' 1) 37

*D istj ' npa .' -ri ft. r" coat,, .1 a i .h.'r ftam fItowi.

1 1t • Eb ,7 Fiqure,5 MICROPROBE ANWLYSIS OF COATED REvRACIUr:Y' . - V -,, -'J ALLOY AFTER 200 HOURS ') Al,1 OXIDATION AT I 80F-

1 .-a 1.75X19-0664^ 42,11,%C,70

within a few hours (as shown by the leveling off of the oxidation curves at1800 F), because a thin layer of oxides has fored at the coating surface. Theoxide layer seals the crac,. opening and behaves as an oxygen diffusion barrier.The factor which controls coating spallation is largely dependent unon tile amountof oxidation that has occurred at various sites throughout the diffulion layerbefore self-healing is complete,. Regardless of the amount of oxidation undergone!, this diffusioii layer at 1800 F, the coating would likely be susceptible toi)oth mechanical and thermal shock under thermal cycling conditions."

At 1900 F, oxidation of the diffusion layer and hence broadening of cracksliu noL occur which probably accounts for the lower weight gain observed. At1900 and 2000 F, the outer protective layer still consisted of SiO , TiO , andthe unknoin compound present at 1700 and 1800 F. There was also some diminutionof the reservoir of mixed disilicides.

2100 F

Figure 7 shows that oxidation at 2100 F follows the parabolic growth law,nd ,omplete protection i.; afforded the substrate alloy by the coating. Someolatilization of oxidation products may be occurring at this temperature sinceSo init ai oxidation rate is less than observed at 2000 F. At this temperature."00 1" the outer protective layer first takes on a glazed appearance. Fitzer:",eport, ,that disi licide coatings for molybdenum take on this glassy appearanceýt 2400 F. ihe lower temperature we report may be due to the complex compositionof our coating system. In addition to Mo and Si, the coating contains Ti, V, and

". which may lower the melting point of the glassy-like compounds formed and haie'he heneficial effect of allowing self-healing to take place at lower temperatures.igure 8 which is a photomicrograph of the coated alloy oxidized at 2100 F along

::ith electron-microprobe analysis shows that the outer glassy layer consistslainly of SiO- containing <0.1 wt % Ta, 7.4, wt ' W, <0,1 wt % Mo, 1.6 wt % rTi,

mud 1.9 wt % V in solution (Point A), and small amounts of TiO2 as a separatephase. Thiis is more readily seen in Figure 9 which magnifies the oxide layer bya factor of 5. Pure SiOý (cristobalite) could not be present as a glassy ormolten phase at 2100 1 since its melting point is 3120 F. X-ray diffraction scansshow tVe presence of an unknown compound which electron microprobe analysis iden-tified as an oxide containing mainly silicon along with the aforementioned ele-ments in much smaller amounts. Note in Figure 8 that the outer protective layerhas bridged a crack that has formed in the adjacent porous layer of mixed disili-cides (reservoir). Tlhere appears to be no significant changes in the reservoirand diffusion layers other than some slight diffusion of tantalum into the reser-voir layer (Points 3 and 4). As noted by Points 11, 12, and 13, no diffusion ofthe coating elements into the substrate Ta-101, ,lloy was found.

2200 to 2400 F

In this temperature range, oxidation rates increase with increasing tempera-ture as shown in Figure 10. At each temperature, after one hour of oxidation,the oxLdation rates decrease with time, particularly at 2300 F probably due tovolatilization. Spectrographic analysis of the effluent gases, which were passedthrough a scrubber containing 10' Na 011, showed that silicon was present infairly large amounts and Mo, IV, and V in lesser amounts and Ti in a trace amount.The electron microprobe analysis for the coated allcy oxidized at 2300 F shotw

0 2000 F

'O 1900 F

c4E2L

EFigure 6. OXIDATION CURVESOF A COATED REFRACTORY

C =ALLOY IN THE TEMPERATURERANGE OF 1900 F TO 2000 F."19.066.1023/AMC.71

I I I I

20 40 60 40 80 120 160 200S--Minutes ,-, Hours

Time

3

2-"Eo

C Figure 7. OXIDATION CURVE(3 OF A COATED REFRACTORY. ALLOY AT 2100 F.

"19-066.1022/AMC071

ij• , I I I

20 40 60 40 80 120 160 200

-. m------ Minutes - - HoursTime

P I..dentific on So' T ..I . .T V Distance ;j

I w.o.,,.V)S,2 435 - 0.,I26 1456. 91. 240390 (0~• o131 ISltS144 ,5 1, 1,853 25,1 011569 94132 s5.5, 150

425.3 05f555 99 3.4 521 75S. . 2;.2 69!457 126 38 _.91 50

6faWS2 268 ~48.5151 3.9 34 2.9( 40S25.1 61.5,3 (03 1.8 24 308 22.9 67.3 9A1 (03 08 (t) 209 i1_7.5 7

18109 (03102 1(O1 toi.8 k012' ' 8:3:, 8j•.~~~o o~c,.,'•,,10 { 8.3 of 8 J (03(01(,f~ '0 2 lCoat-nq!Substr:le16 t iO fri

It Aitoy 1 00 8,811321(2 1(101012 o0 5 88 12• n140.31 W01 o , 0.1' 20"

13 (058821 119! Og 3 (OI 0 01 30,-AL A,, S 120 25i 02 7V 0 1" 16 291 300

'Distance%$ measured from Colatig, 4II others irom SubstrateFigure 8. MICROPROBE ANALYSIS OF COATED REFPACTORY ALLOY AFTER 200 HOURS OFAIR OXIDATION AT 2100 F. Mag. 200X. 19.066,140/AMC.71

Figjure 9. MICROPROBE ANALYSIS OF COATEDReservoir REFRACTORY ALLOY AFTER 200 HOURS OF

Layer AIR OXIDATION AT 2100-F. Mag. 1000X.1 19,066.941 AMC-71

Substrave Identi-

Oxide Point fication Si Ta W Mo Ti VLayer "r B TiO 2 4.5 0.3 8.7 (0.1 35.40 0.4

. 20.25 (0.1 9.0 (0.1 1.2 (0.1

D SiC 2 40.50 0.1 0.9 (0.1 1.4 0.8

o 2400 F3~ a 2300p F K.

o 2200 F

~ 1 REFRACTORY ALLOY IN THE TEMPERATURE RANGE# OF 2200 F TO 2400 F.

19.066.1021 /AMC,71

20 4 60 0 80120 160 •.00SMinutes -4---- dours ---- o

Time

-Eli

49

Fiue1.OIDTO UVS FACAE

in Figure 12- when compared to the analysis at 2100 F (Figure 8) does indeed sub-stantiate a volatilization reaction. In general, the elements Si, Mo, Ti, and Vdecrease in amount as the temperature is raised from 2100 to 2300 -degrees. Theamount of tungsten on the other hand, in reases at 2300 F probably because tung-sten diffusion from below the porous or eservoir area occurs more rapidly thandoes volatilization. The amount of oxid in the reservoir or porous area alsoappears to be increasing at the higher t mperature which, to some extent, dimin-ishes the reservoir layer or amount of d silicide available for self-healing.There is a significant change in the diffusion layer at 2300 F, namely, conver-sion of (Ta, W) Si 2 to (Ta, W)5 Si 3 and (Ta, W)2 Si. It is likely tha'ý the lowersilicides will exhibit poorer oxidation resistance but would eventually exhibitprotective behavior sincc larger periods of oxidation would be needed to reachthe protective stage, the lower the silicon content. 6 Perkins 7 and Berkowitz-Mattuck and Di]s 8 have reported such behavior for molybdenum silicides, namely,oxidation resistances decrease as MO Si 2 > Mo05 Si 3 > MO3 Si. Figure 12 which mag-nifies the outer protective oxide layer shown in Figure 11 brings out a point ofinterest rega'-ding this layer. Points A and C both consist primarily of SiO2with smal:l r'mbunts of Ta, WV, Mo, Ti, and V in solution. However, Point A, whichis in the outermost area of the layer appears to have been in the molten state at2300 F whereas Point C, further into the layer, does not. The elemental analysisappears to indicate that the higher concentration of elements at Point A hasfacilitated melting in that area.

2500 to 2700 F

Figure 13 shows that the parabolic rate law still holds in the range 2500to 2700 F, and oxidation rates increase with increasing temperature. The markedincrease in oxidation rate at 2700 F after the first hour indicates that theprotective limit of the coating is being approached. However, complete protec-tion was afforded the substrate alloy for the entire run. Some volatilizationhas probably occurred but is considered to be minor since a net weight gain of12 mg/cm2 at 2700 F is noted. This relatively small change in weight againattests to the excellent protective properties of the coating system.

Photomicrographs with accompanying electron microprobe anaiyses, represent-ing 2700 F oxidation of the coated alloy, are shown in Figures 14 and 15. Fig-ure 14 concentrates on the outermost glassy layer of the coating system which isshown filling and sealing a crack that has formed. This glassy layer is increas-ing in thickness at this temperature and consi'sts mainly of SiO2 with smallamounts of WV and V and a larger amount of Ti in solution (Points B, C, and D)which was also the case at temperatures as low as 2100 F. Again, TiO2 (Point A)is present as a separate phase dispersed within this lay-r. The reservoir layer(Points 1 and 2) still contains mixed disilicides (IV, Mo) Si 2 with very little',f any porosity remaining, because the pores have become filled with oxides.1itanium which is also present in the reservoir layer has diffused to the outerprotective layer (based on a comparison of the elements present in both thereservoir and outer glassy layers at temperatures of 2300 and 2700 F). Thereservoir layer has further diminished in thickness at 2700 F. The diffusionlayer shown in Figure IS has broadened as a result of the diffusion of silicontoward the substrate where it has formed lower silicides with Ta, W, and V.Also some vanadium has diffused from the reservoir layer toward the substratealloy, Figure 16 shows how the thickness of the diffusion layer and the combined

9

*Lm S. -Identification SW c a ~ iý TA V Distance H_ __L - -- - ] T ,, -

. ." ,I (W.MoV,TiS, 25.8 0 55.00 7.8 t3.9 7.0 195"24.1 0. 5301423 .9 5.71 155

"" 3 124.6 0.651.01134 3.6 6.55 1404. 22.3 6.5,37.6125.4, 3.0 541 55

'5 216 16.5•42 It 10- 6 i 36 5 --(TWS1 895 756.3-223 3.8 26 36G 30

0T:~ : 0.5 2. 1 4ý 20

8, (TaW)S 7,9 79 9.o0.51 i 0 '6 0

9 - ,)s 7.6 83.71 ~8.4 0. 0101 (0.11 Interface

~ 1 Alloy (0.05189.710.3 103 (0.01 ý0 20'V 1 _ 1 03(0141 19 U 12 Alloy (0.05J89o110 o03 001 O 01 30"

" .•. . *Distances measured from coating, all others from substrale

19066-939/AMC.71 Figure 11. MICROPROBE ANALYSIS OF COATED REFRACTORY ALLOYAFTER 200 HOURS OF AIR OXIDATION AT 2300 F. Mag. 200X.

Figure 12. MICROPROBE ANALYSIS OF COATEDREFRACTORY ALLOY AFTER 200 HOURS OF AIROXIDATION AT 2300 F. Mag. 50OX.

Reservoir Layer 19.066.942/AMC.71

Stibstrale "Oxide Layer

Point fication S!i Ta W Mo Ti V tance p--

A Si0 2 20.25 (0.1 3.1 (0.1 2.6 0.4 270B TiO2 (1.0 0.3 4.3 0.3 55.60 0.3 265C Si0 2 35.40 (0.1 ). ,10 1 29 (0.1 260

i U2600 F

A. 2700F F/

! /

- :iqure 13. OXIDATION CURVES OF A COATEDS/ ,R�EFRACTORY ALLOY IN THE TEMPERATURE

. •. - RANGE OF 2500 F TO 2700 F.19.066,175/AMC.70

-J420 40 60 40 -0 120 1o0 20

hMm -- - HTime

10

D stow. frCrnl

Point Idenotfication Si 7a "I o T. V sub.:A TO 02 17 5 5837 0.49 -381

B SiO: 26.7 0,1 - .I 46 A ntC SiO 30,5 <0.1I <.0 1 4.1 0.9 -10(lD SiO: 2G.5 0.1 ' O -01 3.8 09 2t,E SiO2 + TiO2 14.5 0.8 1 0.3 19.3 3V 24!I (W, .1o0)Si' 25.3 0.9 57 .j 115.0 03 1.2 225p.2 (W, Mo)S 2 24.2 2. 1 555 15.6. 0.3, 23 1*,k

ft Figure 14. MICFOPROBE ANALYSIS OF COATED REFRAC'wH-ALLOY AFTER :30 HOURS OF AIR OXIDATION AT 2700 P.mA Mag. 20OX19*066.139/AMC*70

DittanCe fromPoint Identification Si Ta WN Mo Ti V Substrate~

3 (Ta,W, V)5S13 10.0 58.3 1 6.2 2.2 38 95 90pu*4 (Ta, W)$Sitj 8.4 82.2 8,8 -10.31 0.2 0.3 GOP1

4 *5 (Ta. W)Z&S 6.2 80.5 13.3 <&3' <f0V <0.02 1 5)1* * Dstance front

Coatinq_~.... 6 Alloy 1.1 886 10.1 e0. 3 <DO01<0D.02 5)

7Alloy 100 898 102[<, <01<.2 0)

~ ,.~ ~Figure 15. MICROPROBE ANALYSIS OF COATED REFRACTORY~J ALLOY AFTER 200 HOURS OF AIR OXIDATION AT 2700 F.

Mag. 20OX

reservoir (porous) and outer '?rotective layers vary with temperature after 200hours of oxidation. Both coating layers generally increase in thickness withincreasing temperatures, particularly above 2300 F.

DIFFUSION OF COATING AND SUBSTRATE ELEMENTS

Diffusion of coating and substrate elements was examined by electron micro-probe analysis for the as-coated condition and after 200 hours of oxidation at2100 and 2700 F. This was accomplished by traicing element concentrations acrossthe coating thickness and into the su~bitrate for a short distance. It should benoted that the electron microprobe data is not intended to be taken as a strictpoint to point analysis because of the inhomogeneity of the coating system. Re-sults showing element concentration as a function of distance from the substrateare plotted in Figures 17 through 22.

11

376

.300-

f0

i200°F- Figure 16. CHANGE OF

Reservoir plus Ox~de Layer (a) COATING THICKNESS WITHTEMPERATURE.

100- 1066-943/AMC*71

____________ _ . Diffusion Layer (b)

500 1000 1500 2000 2500Temperature (deg F)

I ubstrafeReservoir Layer 50 substrate

'Diffusion Layer Resevoi Lae40 fI ODffusion Layer

40

20 30!

As Cooled

10

L0!er 2:As Coatod

Ubstrate Reservoir Lyr 10

o30- S• Substrate Resefvoit Ljy~r

0iO

20- , 0 20 Diffusion Layer2100 F Oxirdizedfor 200 hours

I02100FOxidizod for

..L .. L00 ho s

30 -1SubstrateLouser 20 - - Reservoir Layer20 - ,,L . • Diffusion I

Layer /'-

20 2YOt uoiione

[0I 211t0 F 0.1d.We 111 o2700 Fý OIzfor 200 hourfor2C0us

0 25 12b 175 225 0 25 50 75 IW~ 125 150175 200 22S

Dlstms from Substrate • P Distance lorn Substrate A

Figure 17. DIFFUSION OF SILICON Figure 18. DIFFUSION OF MOLYBDENUM

IN THE COATING SYSTEM IN THE COATING SYSTEM19.066-1020/AMC-71 19-066.1019/AMC.71

12

S• • - -_-• -- •=-- : •--• i-i.- . • -= -.

Figure 17, the plot for silicon, shows that the silicon concentration de-creases in both reservoir layer and diffusion layer and at the interface betweensubstrate and diffusion layer as the coa "ed alloy is oxidized at 2100 and 2700 F.Some silicon has penetrated iito the sub trate alloy at 2700 F.

The molybdenum plot, Figure 18, she 3 that molybdenum concentrations as !Ighas 45 wt % were present in the as-coated zondition in the res:ervoir layer. f.croxidation, the molybdenum concentration ?creased marked]: in the reservoir .1:There was little effect in the diffusion layer wnere -he 'iolyhdenun conc.ntrat Cwas low, and none was present at the int rface of substrate said diffusion ia~Ž,.

Titanium also decreases significanitly in the reservoir X.iyer as a result foxidation (see Figure 19). There was az slight increase of titanium in the d ffusion layer and only very small amounts were present at the substrate-diffiut-i,•layer interface. At 2700 F, the titanium clusters at the inttcrface between th.diffusion and reservoir layers. Figure 20 shows that vanaditmi behaved similnr~yto titanium.

In the as-coated condi" ion, tungsten is concentrated in the reservoir layerclose to the diffusion layer, also in the diffusion layer and, of course, itconstitutes 10% of the substrate alloy (see Figure 2i7. Oxidation causts thetungsten to be more uniformly dispersed in the reservoir layer (55%), while tbeamount of tungsten in the diffusion layer decreases. Therefore, tungsten is dif-fusing from the diffusion layer into and throughout the reservoir layer.

Figure 22 shows that tantalum, in the as-coated and under oxidizing condi-tions clusters mainly at the substrate/diffusion layer interface, and in thediffusion layer. At 2700 F, there is some diffusion of tantalum (0.9 wt %) upto 225)1 from the substrate into the reservoir layer. Diffusior of tantalum fromthe substrate into the diffusion layer contributes to the broadening of thediffusion layer noted at 2700 F.

EFFECT 01' HEATING RATE ON THE OXIDATION OF THE COATED ALLOY

Since we reportedI that the complex disilicide coatings are susceptible topesting in the temperature range 1400 to 1600 F it seemed advisable to establishthe minimum heating rate at which the coating could safely traverse the pestingrange without deleterious effects. Accordingly, oxidation runs were ,wde atheating rates of 1, 6, 10, and 25 C rise per minute from ambient to 2G00 F. Theresults were plotted in Figure 23. Failure of the coating occurred only at theslowest heating rate of 1 C rise/minute. In the temperature range 800 to 1300 Fwhere the slope of the curves is greatest, low temperature oxides are forning,whereas high temperature oxides are forming between 1500 and 2100 F. Sin:e theexposure time is longer at a I C rise/minute heating rate at a given temperature,the opportunity for the formation of th.., lower temperature oxides is optimized,but failure occurs due to pestlng. Therefore, higher temperature oxides must bepresent to alleviate pesting. Although high temperature oxides have formed tosome extent during this run, it appears that damage to the intercrystallinestructure of the coating during its exposure in the pesting range has progressedbeyond the point where self-healing by the high temperature oxides formed canrepair the damage.

Is

A, t ,

i$~ Wo I i I Im I-

_______________________________Ia II)tv~ v4 20010C 0O.4.'4 f.

2100 F *,*@ P

ol7J~.oA- Um

0 20 ' 0 5 I00 12$ I00 I70 ZOO 700 02000S 15 I00I2010 15I$20022

Figure 19. DIFFUSION OF TITANIUM Figure 20. DIFFUSION OF VANADIUM IN THEIN THE COATING SYSTEM COATING SYSTEM

19-066.108/AMC-71 19-66.1-1 7/AMC-71

I\

I40 I

I Al CTIed

s0 20

Figure 219. DIFFUSION OF TITANIEN Figure 20. DIFFUSION OF VANTADIUM IHIN THE COATING SYSTEM I H COATING SYSTEM

19.066,1016/AM C.71 19-066.1015/AMC.'/1

$00 L0600

000

0 4-I", . U0

401-

21 0 t o70 ' 0s0.

to01

10 00 $ 0 1$ 0 9 0 2007 5 2 T 0

O.',A9 0.09,191 U 0590,6 90,,at-4I1 It

Figure ~ ~ ~ ~ ~ ~ ~ ~ ~~t 21.oo DIFSOUFTNSE iue 2wIFSO FTNAUIN THE OATINGSYSTEM N THEsoATN.SSE

1906&-106/AM707 210 F.6*0 5/AMC'

601.noF4

01C ds, pit ~e4!25

5' 10 /S6AG // Figure 23. EFFECT OF HEATING RATES ON THE

3., .•."-- = OXIDATION OF A COATED REFRACTO'Y ALLOY

- 7.'• -/ FROM ROOM TEMPERATURE TO 2800'F.19-066-176/AMC-70

High temperature oxides are completely formed at 2100 F, suggesting thatthis would be the most advantageous temperature to preoxidize the coating systembefore subjecting it to pesting temperatures. A coated specimen which was pro-oxidized at 2100 F for 5 hours and then subjected to pesting conditions affordedexcellent protection for 200 hours. After 200 hours of oxidation, the coatinghad a uniform glazed appearance which is preferred over the specimen preoxidationtreatment at 1900 F reported previously by the authors.

CONCLUSIONS

The modified complex silicide coating afforded the 9OTa-lO1 alloy completeprotection against static oxidation for at least 200 hours at temperatures be-tween 1700 and 2700 F. The excellent oxidation resistance of the coat-., systemis due to the formation of a protective layer which consists mainly of S1O2 withsome Ti, V, and W in solution and a second dispersed phase of TiO2 . While tne ox-ide layer at and above 2100 F is glassy, the reaction products become increasinglycrystalline with decreasing temperature between 1700 and 2000 F. After the glassylayer is formed, silicon and titanium are preferentially oxidized. The oxidationrate curves suggest that the parabolic growth law is applicable, indicating thatoxide growth is controlled by solid state diffusion through the scale. Duringoxidation, the disilicido is partially consumed both by oxidation and an inwarddiffusion of silicon. At 2300 F the (Ta, 11) Si 2 in the diffusion layer thus be-comes con-erted to (Ta, 1)5 Si 3 and (Ta, W)2 Si. The lower silicides in turnwould not ixhibit as good oxidation resistance as the disilicides, and performancecould be reduced. Since the glassy protective layer was formed at 2100 F, it ap-pears that the preoxidation treatment to alleviate pesting would more advanta-geously be carried out at 2100 F rather than the 1900 F previously reported.Without a preoxidation treatment, pesting failures can occur at heating rites be-low a 10 C rise per minute as the temperature is increased from ambient to 2800 F.

FUTURE WORK

We plan to carry out similar studies with SU31 and FS85 columbium alloys -disilicide coating systems and to investigate the use of barrier layers andelement additives to the disilicides to reduce the thermodynamic activities ofdiffusing species.

is

LITERATURE CITED

1. FALCO, J. J., and LEVY, M. Alleviation of the Silicide Pest in a Coatingfor the Protection of Refractory Metals Against High-Tenperature Oxidation.Army Materials and Mechanics Research Center, A4MRC TR 69-30, December 1969;also J. Less-Common Metals, v. 20, 1970, p. 291-297.

2. STEPHENS, J. R., and KLOOP, IV. D. Exploratory Study of Suicide, Aluminum,and Boride Coatings for Nitration/Oxidation Protection of Chromium Alloys.Transactions of the Metallurgical Society of AIME, v. 245, 1969, p. 1975-1981.

3. IHAYMAN, C., and RESTALL, J. E. Improved Siuicide Coatings for the Protectionof Molybdenum - 0.5% Titanium Alloy. J. Less-Common Metals, v. 19, 1969,p. 9-21.

4. RESTALL, J. E. The Development and Testing of Protective Coat",ngs on F48and SV16 Niobium Alloqs. J. Less-Commor. Metals, v. 16, 1968, p. 11-36.

S. FITZER, E. In Passivierende Film und Deckschichten. If. Fischer, K. Hauffe,and W. Wiederholt, eds., Springer, 1956, p. 43.

6. KOFSTAD, P. High Temperature Oxidation of Metals. John Wiley & Sons, Inc.,New York, 1966, Ch. IX.

7. PERKINS R. A. Materials Science and Technology for Advanced Applications.v. II, 1964 Golden Gate Metals Conference, San Francisco, February 1964.

8. BERIKOWITZ-MATTUCK, J. B., and DILS, R. R. High-Temperature Oxidation.11. Molybdenum Disilicides. J. Electrochem. Soc., v. 112, 1965, p. 583.

5 16