DTHE DrIESIONAL STABILITY OF - DTIC · The present report suuisrises the work under-takcen by...
Transcript of DTHE DrIESIONAL STABILITY OF - DTIC · The present report suuisrises the work under-takcen by...
CID-
DTHE DrIESIONAL STABILITY OF
SELECTED ALLOY SYSTBMS
HARD COPY ' $-__ ________ ,__ MIROI'CHE $.C.J
C.). Final Report
June 1, 1963 to August 31, 1964
Prepared under:
Contract No. N140(131)75098-Naval Applied Science Laboratory
New York Naval Shipyard
Development, Engineering, Bro-klyn 1, New York
and Manufacturinghigh and ultra high ALLQXD.QENERAL CORPORATIONvacuum equipment M,,N OfFICES: 37 CAMBRIDGE PKY. CAMBRIDGE. MLAS O 2142 P•I4ONE , USM AREA COE 61,
electron beamn euipmentadvanced materials
research ,
mH I31$emu=
THE DIbW ICKAL BABILXT Of8Elf-UCTM ALLM BMIWTa
byl
SI,?J. Weihrmaxch)&.J. Mordoft
?imal Rleport
June Is 1963 to August 31., 19644
Prep.m under&t
Contract no. 3lL4U(l31)75Q98MNaval. P'Ppled Sclience Lakoratory
3fie York Noval SbLipyarciXrooklyn 1., Nov York
a ubmitted by I
Alloya Gawwrl Corprortion,ý7 Caaobvidge Parkway
c~ia~tiridge 4&0. nasaachusetts
Pan
1. . UCT oB .... . ... ION . . .............. . . . 1
s MATRUAI.& AND .P.P.TIO.. ........ .............. .
3o RXP~Xh4ZNTL PROCEDURE, * .* ** * *. ** . *. .oa* . 5
4. MS ADOZCU8B.sS * AC .. .. .. .. .. . ..... 77
5. ]DISTORTION MEASURa4NK ýACIarTY **ov*ewe*4S40*o*S 77
6. M UM WORK. ........................ ........ .... a3
7. *~ ~ ~ ................... . 84
x. choical. composi~tion ot specimen materials 3
it H eat Troatiaig and I4CMachini Scbedules 4
111. A~pplied Cvwe) Stresse 1
IV. SlAsthc Unmits and 2lastic Koduli ot. Tkhr"Alloys at 75m, 1.5U and ~icuJW 4d
V. Crsep Constanto for Squation 14 66
V1. Creep Constanits for lquatioin 3 7
Zan
Figure I - OcLeastic view of capacitance gauge andSpOcLAn.. ......... .. .... .............* 6
FiguXe i - Capacitance gauge test specin.......... 7
Figure, - Constant temperature chamber tar elasticI Imit measurements . . .. .. .. . . . ., . . .* .* a* 11
Figure 4 - Coautant temedrture and constantatmosphere chamber for creep measures•nts. 14
Figure 5 - Storage specimen for dimensional stabilLtytestj ................ ....*.. .... *. ....*o . 18
Figure 6 - Stress-total strain And stress plasticstrain curves of free cut Invar "•6u at75, 150 and kC¥..F .... . ..... . 4
Figure 7 - Stream-total strain 8nd stress plasticstraLn cur•ves of 356-T6 aluminum at 75,150 and •,o,... . , . . . . ... . .
Figure 8 - Stress-total strain and stress-plasticstrain curves of 31Q stainless steel at75, 150 and i0VF. .... ..... ** *.*..*... . .. .. .
Figure 9 - Stress-total stxain and stress-plasticstrain curves of 6U61 aluminum at 75,150 and 2 . 5
Figu•re 1-8tress-total strain and stress-plasticstrain curves of PZ9dA magnrAsusm t 75,150 and 2 0 .*7 .. .• • * * * .. *.. . .. * * * * * ..... * 6
Figu"* U-Square root of plastic strain vs. stressfor five alloys tested at 7528....... ....
Figure le-?otal caeep curves of Xnvar 136" at 5*YF.. 30
Av
Latt 01 PLUrMeS .- ContOWiA9
uAM
FLgue 13 - Total ;:reep curves of Invar "36" at 150F. 31
71$4r@ 1.4 - Total creep curves of lnvar "36" at ýUQ:k. 3
tLguxOe 15 - Total creep curves of 356-T6 aluminum at85°*r 4, .a a . . .. . . .. . . . .. .. *......#a4 6 0aa4a a0a06 33
Fagure 16 - Total c"ep curves of 356-T6 aluminumat . ..... ........ .. 4.. . 34
Figure 17 - Total creep curves of 356-T6 aluminumat •UOF. ......... .. 35
Figure A• - Total creep curves of 31 stalinesssteel at 85•* ..................... ..... 36
FLOgM" 19 - Total creep cur-e' of 310 stainlesssteel at 15 o0 * . . . . .. .... . .* a 0 * 0 * * 4 * * & .* 37
rIgure 2U. - Total cxewp curvs of 3Lu stainlesssteel at 20u*... ....... ...4..4 ...... -3s
FLqure 21 - Total caeep curves of 6,.,6L-T6 alumi.nuaat 8 0 . .. . . . .. . . . . . .. . . .. 39
Fig•i• i• - Total creep curves of 606i-T6 aluminumat 4u.. . . . . . . . ... . .. . . . .•
Figwm 23 - Total czeep curves of 606l-v6 aluminumat 2uU*F.................... ..... ......... 41
?J~gIxr i; -Total creep curves of ?59eA mgnesiumat 65F ............. . ............. 4k.
Figure 25 - total creep curves of Azr9A wigneIuMat 45°..... * ......... .. 43
Aigume 26 - Total creep curves of A29"A •agnesiumat 4O04.......... ... .
V
LAst of V*ulr - Co!tlnumd
iLgure 27 Total cr"ep curves of f•ree cut Invar at7 5 O* °51
figure 48 - Creep curves of free cut Ynvar at I3UdW 52
tigure 29 - Creep curves o9 fret cut Invar at -Oue. 53
Vigure 3U - Creep cuirves of s56'-T6 alumainum at 75 • ,* 544
FPIure 31 - Creep curves of 356-T6 aluminum at 150Ci . 55
Figure 3ý -Creep curves of 356-'•6 alu•ai•um at 2•QF* . 56
Figure 33 - Creep Gurves of 310 stainless PtOeOat 75 . .* * . .................. .. ...... 57
Fiure a - Creep curves of 31( stainless steelat 151U,*P .. .. . .. . . •. .. .........*oo 56
Figure 35 - Creep curves of 310 ataLnules stoeelat idou~p ..... *.,. .. .. .. .. . . *. . . . .. ... ... . . 59
r gure 36 creep curves of 6u61 a.1wminum at 75el1.... 6Q
Figure 37 - Creep curves ox 6U6L aluminum at L507P ... 61
Figure 38 - Creep curves of 6u61 aluminu at 2uur *..6
Figure 39 - Cxep curves of1 AkZ9ii - T6 magmeiumat 75of .... . 4.e....*.. ...... . .......... 63
FAgure 4U - Creep curves of A29.\ a0gsesium •It 15u7, 64
FIgure 414 - Creep curves of ,-Z92A nagnesium at iuOJr . 65
rigure 42 - DiensiaAoal change of fow cut Irivar 36stored at 75, 150 and O&1a0 '• .... ,,..... 71
FIgure 43, - DiznaLouaL chn"es oi 356-Tu6 Aumiuumstomd at 75, 150 and 2OfW .*............ 7*
vi
Linat o liaxes - Cati.4
Figure 44 - Dimensional Changes of 3.v staLLeOssteel Otore( at 75, 15u and i"6......... "73
FIqvre 45 - Dimonsional changes of 6•6• Aluainumstored at 75, 150 and 74*.............. "
Figure 46 - Dimenslonal change of r;.9P. magnesiumstored at 75o, 150.. and •jt....,......, ,.. 75
L'gure 47 - Distortion specimen ............ ..... • 7•
FiwMe 46 - Scematic view oi distortion measur~ugfacility. ............. ... *6#609460's 6* a *
vii
Slastic limits, elastic moduli, strems-total
strain and stress-plastic *train curves were dtermined
for tfa cuttnvg Invar 36, 356-T6 sliumins 310 stainless
steel 6Q61 aluminum and ?,Z9k, nagmasium at 715,, L5 and d'o.
Tas creep behavior of these aolaj & was studiLd at the same
temperatures of interest at stresses above, below and at
the elastic limit.
In all cases It was found that the precision
elastic limit did not define the sinism stress to cause
creep, in fact creep took place at stresses below tha3
elastic limit at a much reduced rat&. The data was
analysed in the light of current creep theory and micro-
yield theory. A dislocatIon Model was Used to deriv4
analytic eAprsiLons for creep nar the 10"6 elastic limit
stress. below the elastic limit creep was found to
follow an exponential time depwduc, while above tke
elastic limit a ta tine dependence was followed wLth
values of a from 1/) to 1/2.
The dimwnsLonal Oangias taking place in these
alloys under storage conditions at the three texperatuma
of interest is also presented.
A facility fox noasariw~ dintortios as a fuwtictn
of beat tz~tuq an mach mcih*Jw oplaatlow was dooLqued
wkd coustmuited and is deacrLbo.d
The present report suuisrises the work under-
takcen by Jklloyd Geoneral. CorpozatAoz duzizag the period
jume is, 19b6i to August 31, 1964 unader the sponsotshl.p
of the Naval. Appl~ied Science Lab~oratory,, New York~ Saval
Shipyard& to investigate the dinenajommi. sta~i)lity at
several alloy systems.
In the course of this program dimeanionAl
staoitJlty evaluationse were madoe ot five alloys,, free
cut in2var 36, .356-T6 aluoinum, 31o st~ainless steel 6u6).
aluminum and A4SVA magn~esium. An osperimental, program
was developed for dete rmining the dimension&l distortion
of castlags acconpnying metal. fabricating operations.,
and the necessary equipment was deaIgned and constructed.
Masurements, of distortions acconpauying heat treating an~d
muchining operations were made tar cast~ins ot type
i. .& 0. 25UZ M..A& .ZMM
e alloys tasted were Fzve cut Inver 036',
.56-•6 a3luminum, 31u st.in•l is stoeel,6u61 aluwzaiurm and
A29d. magriessam. The Invar and 6&06l alum~nuu were
received in the form of I 4nr.h diaaeter coCLu drawn
unanarasled rod. The 356 aluminum and 3L. otaine.as
stock was received as 1 A 1 x 6 iz41h cast blanks. The
magaesium was roeav± d in te form of e5 pound cast
Aaiots. ?as ;h*m1.c&l compositions ot these alloys are
Siven in Table 1.
k 2 Beet =- atA- anad .achgiAM Sckedu1L
•The as-receivad materIals Were givon heat
tratmes reomnded for maximum disionsionaL wtsabiity. (i)
The heat treating and uacbining schedules -s well as the
final hardn as of eaec alloy Are prosented in Table It.
Chmi~ C~aa~in atMaciin M&Aerials
invar 31
maz 1.0 -. .24
9L. 0.3I5 2.070u.6 <2
cc 23.0 U.2 -4
N±+Co 36. L, .0 do .10
mi ft 19 u - -
ire ~ a. bal..-a
-- bal. bal. .
- m1.LI bal.
Cua u d .o
36 a. -
heat 9rati*a and iac_€xinip schidu~le
xnvar " 1. Is8t treat 1541Vr in hydrogen -V1r ho", water quanih.B. lough machine to over also diamisoms.
bardanesi . Neat treat 12uJ in hydrogen - airUockwell "&,, 85 cooL.
4. Final macaine5, sabi W'r - '48 house, air cool.
A1U asta.Lea 1. oueQnc. Aabaa• . 195,.'a Ln hydrogen - "steel A/• hour, water quench. -'
2. sough machine to oves iz.e 4iunowione.hardness 3. Stress reollev 750"r Ln a.r - 1 hour,
ockWale "so 83 aiAr cool.4. Pinal. Uabhine5. gtabILla 20047 - 2u hours, air cool.
356-26 1. solution ana"ml- I-U•U" Ln air - d bra.aLuminum i. uenck% in boiling water"arCdn 3s. 3* h wa miain to oversL" d4Jnions.' 14lLwlSl "Vo, T7 4. Temper 3.Q*P - 4 'hours, air cool.
5. ftnal machine
6(6i 1. Solution aanaL - 970F ia air - k bira.alumaum 2. Water quenchhardness 3. Roshs mchin to oversize d&hm•inmsiPockwlL "am, 55 4. Aje Iirden - 3% i, 6 bra.., aIs cool
5. rual maChLaw
AM. I. Solution anneal - 750OF in tOkMagis AtmostWre - 2V hes.berdnme* 2. Mr blast cool3otell o" , 86 3. •ugh madbLve to overaise dimasions
4. Age harfn - 5UO', 44 brs., eaLr cool- -- A.
3.1I Cawaacitanc StriMM& g eamiain
All creep Vbavior as well as the elastic liait
and stress-strain deterl•e tiona ot tbe alloys were made
with a capacIta•ce Otrain gauge. This gauge (Figure 1)
consists of a two piece precision machined Invar housing
which Is mounted on the houldexs of the speclmen (Figure 4.),
A quarts disc plated with platinum on one face is mounted
in the upper section and acts as tkm positive electrode
of a parallel plate capacitor. The face of the lower
housing Is highly polished and acts as tre ground electrodf
of the capauitor. The height of the housUng can be varied
with the adjustment screwn and the gauge sensitivity can
be varied within broad liits by changing the capacitor
spacing. with this device, changes in specimen leagth
result in a cnaage of capacitor spacing which. in torn,
is translated into a change in capacitance.
'The elastic limit at a material is defuled as
the lowest stress VWNqured to produce a mmessrsbl amamt
of -r~-t- ww-Pw ý. I- -040if-f -ste - e in these detecaSatiOss
was the lood-u-load technique of Stalr and Averbacb. (2)
-5-"
SPECIMEN 5QUARTZ PLATE HOLDER
INSULATOR-SPECIMEN
PLATE LEAD------. CLAMPING SCREW
AIR GAP ADJ.PLATE
_____LOCKING SCREY
QUARTZPLATE SPECIMEN
; SPECIMENHOLDER
USPECIMEN
FIGURE 1 SCHEMATIC VIEW OF CAPACITANCE GAUGE
AND SPECIMEN
o z
00
0 Li.
(D LU
CD 00 L
00 C3
0 z D
-1-
CD LU
LOU
4;olj C0 q
0 CnLL
IL -. Q
Tests %ere conducted In an ixstron gpet-driven testing
machine Of 1.0,0U pounds capacity. Th 50Jý and 1W(M
pound scales Were used and loads vex* road to the noarest
5 mad 10 poundA, rspectLvely. The test epecluen, shown
Ln Figure i, was gripped against Its 45 shoulders by
matechd split nuts wb.cb weve threaded Lnto grips. the
grpS woewe connected to the frams and cross-head ot the
amchine by mild steel rafd hangers through spherical rod-
end bearings. Load" were applied in increments ot 4M) pei
in the egion of the antsicpatod elastic l.mit. The load-
ajg rate was approiLmately 2UIO(Q psL per Anaute and tbe
specimen was immdiately unloaded on reaching the desired
stress. Repeatable Osevog stiress condition vwre achieved
by al41owig the specime to UMag freely udemr tAe weLght
of the lower grip and hamger for unloaded strain readings.
AlU the ComjpnQts of the gripping system we precision
machined to maintan a &ahILty of loading.
upon re&acIng tke elastic limit the 3.oad-Unlod
seqmeco swas continued until the msacroecopic yield point
was zoached. lbo change In capacitanco was read at the
msximm load and u unloadi6ng, Zn this way the Stres-
total strain and stress-plastic strain curves were gemerated.
-8-
The chubgO in gaug capScitauce was msaued wfth a
lobertsbaw and V'alton Frow~niaty "~tog, geg O&Luvate
os car~reLd out for each Specimen by Loading the specim
to stesses well DeLoW the elastic limit and notixq the
motor 4f1ctLon. There are sAz mot scale factors Which
allow strain xmasuxements -o be mId in a range £fro
I x I-7 to 8O&x i-6 ionches/1wh. "N caljbration was
found to be linear over this range as evidenced by
comparison with electrical resistance strain-Sauge
measurementm. The ma.m strain sensitivity achieved vLth
the capacitance gouge was I x lu' 7 bnches/ibh., howaver,
dii iiculties in teaerstux* control caused theural *m~axwom
and contraction durjing the loa4d-unload cycle which, An some
cases$ dwarfed length changes due to plastic strain, As
a result te s&train onsItLvity aused was 5 x LO.-7 I&eW
inch.
TestMW wAs carri out at tbree temperatureso
75, 150 an 2(JO?. Th Moo. t*NP'ture tests (75V)
wore carvied out in asi without texperature coatrol.
to tIMO lave: thermav asas of the spscimn-capscitaace gauge
contiguratio, te erature f1uctuations took place voye
slowly so that expansion ot contraction duriig a•y Load-
-i9-.
unload cycle was lose than 5 x 10`7 ichee/incch. 'At the
higher tapme'atum-e (150zoo 6=4 fto~r totag Was.. Car!"ou
in a temperature conatrlLed air cbhaber. The apparatus
(8ee Liguqeo 3) consi~ted Of an annular container. The
inner cyLinder of the container was a alchroe woamd
resistance furnace and served as a constant temperaturs
enclosure for the spec • ien and grips. The temperature
was controlled to withli t 1V by a thermocouple controller.
1-l temperature was read independently with an iron-
colatantan tkaeruocouplo mounted on the spt-imen.
3.3. tl~asti.c Mdu~un M)e~surmnta
Sineo the PoALmity Moter does not give
absolute values ot stjran aA4 muot be calibrated for each
spec"Wn, it is natessa•y to measure the elastic modulus
indepndently at each temperature. To acco•lplih this,
etched foil electrical rsst-nze strain gauges wexe bonded
to tho gauge section of the specim"n with MIQ( epo•y
cement. To gauges mounted on opposite sides o f the
specim and w£ietd in series Aare used to balance ony
effect due to noaaoxality Q" loading and to produce a
eaImy of aveegro Strain. 'The gouges W03* umoistua prooted
vith a coat of stycast i651 hiGh teRPeature ePoy.
Sa~ -- - -
I HEflmOCOUPLELEAD UPP~ER ROD
FLLU) 7 ROUGH
GAS
I OUTLET
LIMIT MSSURCIMET
-LI -T
Fox tests at &"Lbent temperature a simsilarly prepared
ustressed specimen wasn mounted AeAr the test aPeiA.~Me to
act mus a tamprature compenbator. At l5c#* and eit-O the
unstressed spec imen was mounted to an air furnace and
An~opedSently maintained at temperature. both the test
OPe41110 OW~ compOUsator spec Len %emperatures Were
monitored with Iron-constantan thermocouples. For room
texperature tests no temperatuare control was necessary
since the ambient did not vary more than I O.,eF during
tL~e short testing tiames. The stress values were read
anb the Instron Chart to an accuracy of ±t lwa pal, and the
$train values to t d l- L~ nches/inch with a Baldwin
type N &train Indicator.
A stress was applied below t~he elastic limit
and the strain was memaurod, rive independent measure-
*auts weze sade for esob specimesn at each temperature and
time modulus was calculated. Thu five values were then
averaged.
3.44 ui.aa
Creep tests we"e conducted In a bank of six
dead-load creap seichimes using ain Interrupted lcoading
tocdaique. Yhe test specimen was encloed An a &cistant
temperature and Constalt atmosphere chamber bLeh was
mounted an supports connected to the creep tram ("ee
Figure 4). Whe temperature was maIntaIned by a tibUlar
nichroue would resistance furnace and controlled by a
tharAstor element to & 4.10r. Nitrogen was contiauously
fed into the chamber after passimng throug a dryang and
heating column.
I, creep stress was applied for a period og time
and then removedu the stress was mesusred with a otaian
gage load cell and the residual strain was measured by
trA capacitance Btra.An 1age. The speciAmAn Was thea
reloaded for A JouV4er per, od o£ time. the PXcedure was
continued for increasing perio4s of time until a total
oi 5W0 hours of creep was accumulated.
The strain was measured using the capacitance
gauge and the PecLmen of tIgure 1.
m" changes in gage capacity WreW measured with
an llectro-6cAOntif ic XndustrAes 707A capacitance brl*ýe
of variable sensitivity. The meter was calibrated by
loading the specimen with knoun %ieights and f ECUthe stwes-
strain curve, equatisg the strain to the meter defgdcaton.
dPI34&
L EAD UPPER
ROD
GASBAFFEL
GAS
GROUNDLEAD
C APAS ITANC EGLASZ -- , eGUAGE
SPECIMEN
GAS
RUBE ELOSAOF5FE TTO
LOASRODE
KaOtmM strmas resoluttiton of 1 • 107 inches per inch is
The specimen was gripped against its shoulders
by matched split nuts wich woez threaded into grips.
The grips were connected by threade rous to the upper
and lower yokes. Te lower yoke, ft&ic also served as
a load cell for messurixg stress, was connected by a
spherical thrust bearing to the frame of the machine.
The upper yoke wae pinned to the lever arm. The lever
arm was pivoted on a pin in a support block and could be
arranged to give a wmchanIcal advantage of ten or twenty
to one. Stress was applied to the creep specimen by mons
of cast iron weights stacked on a hanger attached to tIe
lever arm. The lhad was removed by means of a hydraulic
jack. For wro readings, the lever arm was held an the
unloaded posation by a support arm ond pin.
The creep stress was Measured by Oman of a
load cell in the lower yoke of the czeep maeAine. The
cell consisted of a .J35v Inch daaeter section of the yoke
machined into a ; In.ch tensile bLr. The tensile bar wes
instrumented with two etched foil strain gages located on
opposite sides of tke gage section perpendicular to the
-l5 -
lover azm and wired Am seties. the strain gages were
molituxeoprooted with a coatLaq of Stycast 251 eposy
potting compound. P Similar dmW bar was buwg next to
the load coll as a temperature Cospeastor. A Maldvin
type 3 strain indicator was used for strain readings
ii1ch were converted to stzess by the atrea-stxraLn curve
of the mild steel rod.
Table IIZ Lists the creep stresses and tomper-
atures investigated for each material dutring this program.
3.5 CON""1AU~ ~hnIs it XU aem ADD1±c LRAdIS
In order to study the dimensional J;ngeo under
storage conditiona, specimens Of the four alloys were
macined to the dtmnsLoas of rigure 5 following t:le same
heat treating and machLnan schedule previously outlined.
After the final bat treatmeat the speci• ss were given
* hand polih with 3/u grLt emey paper tfl.olecl wdy a
careful. polish with a fine micropolJshoang paper.
Toe specimns were then placed in air furnaces
maintained at T5,.15U and ?.uUr and their lengths
periodically mnasured. Tkb turnace teerature was
maintained at t u,.5*r of the designated teaperatures.
j j ... j 16
I1ree Cut50vQI48,ujItwZavar 36 3 i,3sw4sW
356ft1354.K) 15QU
6,3Q691ou 5 9 Wh
31.0 stainleos 3s5Q3, w>,5ksteol 25O ~o4 -3 0
16,,70 16,luuu~i c4,O(
175 5 a5w~ 15 s OW
Magnesium 5 i8u 5SU4,78o.3,70U 3,,53U 3 p34(ý
-1.700
K
-4
-j
A
4
'.1)r2 U 5
'-4 -' 0* -4
,�j I-.�00 C'K)
r-1
*�1
42w*��1
C)
CLoC) U
-�
oo '-4
C)'4-4
�3)-I
UC)U4U)
(!)
��1
04)U)
�J (j I'-1
4-.
(;,)
�1 C)w wcILl *�-4
a,. -0 �
(n 41) -4.4 U"
0 -4 -�4--. 0) I J
4-)
(2CL 0
- L
D.naLonaal cisagos were determxved fgom
precision 1amsurwmsnts of the 0- *rall length oa the
specimen after each storage period. All aeasuremeate
were made using a abeffield aechanical optical com.rstor
with an accuracy oL 1 5 mLcroinchss/inch. The comparator
indicated the dLfferwae in length betweem the test
specimen and a gage block of known length.
The nominal finished length oa the storage
spec imens was •4 inches but the length varied froem specieia
to specimen. Since the scale range of the comparator is
Sj.4 iwhes, a single gauge block (4.-ob inches in
length) was used. Whaen requirod, a small standard block
could be selected from a series of standards and placed
under the specimen so that the overall height of the
specimen plus block was within u. l Inches of 4.05U inches.
The storage specimse, upon removal from the
furnace, was placed on a large copper block for a period
in escess of 30 minutes. Because of its large thermal
mass$ tme copper block served as a comstane• temperature
plate. A reading of the specimen was taken on the
ft
The gage block to then road. rAsA procedure was epested
thasow tWwsa and aws&.gud forc w-ith spec 1n and gage
block. Tbe relatIve Leagth of the mpcAMn was caJlulated
£rom these values. The reoolation of tAe comparator was
5 A0 LO6 ±nhes. Voa a 4 iLwh peci•Ison ,ength, tLhs
reulted LA a smesitivity of 1.2 x 1- O inches per inch.
The speciamaw and gage blocks vex* handled wLth gloved
haads at a&L tims to reduce hesting and miasture changes
MaW handlung was kept to a ainisum.
-m;O0-
i • k3*IIW •tu DL___SJJ___JJttSU
4., -- atic 1lt mtaAgd aUX ctavild flhbvvhr
The stress-total strain and atresa-plastAc strain
curves for the five alloy systems are pesnted in Figuris
6 - 10. All determination* reported were made witu the
capacitanme strain gauge and the elastic limits are th
stresses at wich the first 5 x IU-7 strain was detected.
Th elastic limits and elastic nodulU are suarixed in
Table IV. The elastic •Lmits and the elastic moduli showed
a similar temperature diependene, but the elastic liAmLts
were somewhat move senuitiv to temperature. This has
been confirmed in other pure as( ad aluoys(4) aud
has Aeen explained by a thermally activated achaniza ot
,dislocation lultJ• L4plation and motion.
brown and iLakens( 5) and ?homss and Averbach( 6)
have oboesved and explained a parabolic relatiowsbip
between stress adn plastic stratn. Ine room temperstuwe
plastic strain curves of Figures 6-10 are replotted Ia
Figure 11 wit the stress plotted against the squave voot
of the plastic strakin A straight Line dopen4swue was
observed cOnfiusing the parabolic worX hardening law.
0
4-(0
wL 0 07U. 0 0
0 * rW I.- '
00 ~ 0
C- -4
4-)s, T,
0G4J
M
!--4
' ul
0 * 4J:000
CLU) U-i
00 000 qt
(Isdrl) s~i00
4.
0 r0
9-4a 02-1
U)- J,
L
U- U)
0r-4
00-41
-41
Ul f
ý4
(Isdol) s~u0
LL 0
0 10
UU)
000
, 14
044
(15dr~l SS.U)
0to
U..
0,00 0 o
0~-
-iCA.J
InJ
U)
0 i
--4
4-3
- I In.
NO 41
In
0 .j f
(4,- 0~)S 3 I
4'4~4-1
4J
14)
(jj)
4-1
0I
'Jc i
r14
21a"tic Limi~t~s and zlastic Modaul ofZhree allovz at 1%. I1%0 and aweV
Test slastic. Slastic
Inver 75 ýý6,4vUQ2150
356-'T6 75 8,iUU 11.3Alumi.num 150 7o3550 11.1
ko0 7slW1U.
310 stainl~ess 75 22,70i9.stool 15v 2(, 4u 9.J.
.. a:U ýýOUQU 8.7
6cl75 ie.40 o.8Pluminiam 150 1L,70Q 1045
AZ92A 75 5t280 6.5magnesium 15U 5,0u4u .
2W4j7& 6a.
4-4 ) -4
4 r-)
I ->
I I I - -q
.I4 4-1
U~"3-
F- 44'-
4..2 Crenfevol
The creep behavior uAftr load ot free cut lava:
6, 556-T6 alUminum, 3l1 stain'less .tIe1. 6061 a3msn"u
and AZ92F magneIAuSM at 85,15U and kuuOy is paesemted in
Figures The creep rate in all cases was Seen to
start at a hbgh value andi cdcgase continuouslLy until it
approached a steady state condition. The early creep
rate and creep strain are seen to increase witn te"perature
at ¢Omparable stresses.
At creep stresses Lalow the elastIc liSit the
.r-eep rate in all cases decreased continuously approachi•g
zero at the lower te"Ceratures in a swxt tim. %ear
the elastic limit stress the creep rate decreased
continuously approaching a steady state creep rate. At
still hiher temperatures the creep rate was hLghar and
decreasd more slowly.
In order to und"Stand the creep behavior ot
tbase alloys it is necessary to develop a dislocation
maecbanism of creep. Creerp at high load* generally consists
of tbree stages, the iArat of which is characterAsed by a
decreasing creep rate (transient creep). RventualLy the
i -'9-.
500
400
50,000 PSI'"
b3000
z4[
0
c. 200
--W
uo 34,000 PSIUa:
Jt-0
I--I00
22a,000 PSi
IIi100 200 300 400 500
TIME, HOURS
Figure 1.....Total Crccp Curves of fnvar " at;,-F.
-50-
500
400- 46,000 PSI
'"300
_z4 o
Ir
,00 22,000 PSI
too 200 300 400 500
TIME, HOURS
Figjure 1 - Total creep curve.; of inv'wr at 1.'--
. .. I I I I I
500
400
.O --- -- ....
C'300 - 35,000 PSI
'200-I.w 0C.,
-J 25,000 PSiI.-
1t 00
S. . ..... .... ... ... . .... . .. 15 ,0 0 0 P S I
100 200 300 400 500
TIME, HOURS
FLJurel. - Total Crccp curves of Invar "-" at F.
500
400
300 13.000 PSI
Ch 200Q
C9 9600 PSI
6300 PSI
00
100 200 300 400 500
TIME, HOURS
F'igure l) Totz,1 Croup curves of '"--T,' luinum a
500
400
~'300
00
~'200
0~
100
00
100 200 300 400 500
TIME, HOURS
Fiuurc i,.-Toti-l creep-I curves of 7 KT.Aluminum ast L . P
00 12,000 PSI
0
400 -
6 300z
,
/
G. 200- /0w
U /Z.,J / 0
1- 00 0,'O0 C
/o . 5800 PSI
t00 200 300 400 S00
TIME, HOURS
•.v•jur i' - TotaI c-•• cur-ves or .... P'- Aluminulm lat F.
500 3,0 S
400
*:300 o e
CC0
200
100 t00 300 400 500
TIME, HOURS
F'Ii qure -Total cree'p curves of -*I(" stainless steel.a tt r>F
31, 000 PSI
500
400-
400 -1
2~00
10 0
a-t
-37
S00 30,000 PSI
400
22,000 PSI
0
- 300
7
t-
S200
j 14,500 PSI
I-O1- 00
0I- 100
100 200 300 400 800
TIME, HOURS
Figure 7 - Tota•l creep curves of KlQ stainless steel
at '•< .
-38-
Lps
timeo. noursz
I'Ljurc2 Zi Total crecl- curves of ><~1-T' aluminum at 5.
Time. Hours
Ficju~rc '2,2' Total1 creep curvcsq of aluminum atLj(.
.........
time. hours
ri,.urc *-rpota! creep curves of I-( le~uninuin at Ž000ýOF.
U0
Fiyurc: 24-ý Total creep curves Of PZ-,;.A magnesiun, at IF
-- 0
-ps
timeo. hours
Figure z- -'rotai creep curves of t\Z WA magnlesiumn at 1500F.
(.°
/,p
C
. I.
"." .....-. ;I-. .- . I -..
creep rate diecreases to a GM ntant or steady state
condition (mecO1.6 stage) wni" is faiu±oJ.l.owd bY ani
increasitg creep rate (tertiary creep) until failure
occurs.
To explain creep behavior in the first or
transient stage two mechanisms have Leon advanced.
Considering the rate determining swchanism at low tem-
peratures as the motion of dislocations through a torest
ot dislocations b, te crea4Ion oi jogs Leads to the well
known logarlthmic creep law(T) given bya
A t higher temperatures where the dislouations can
acquire enough therml energy to climb and tkereby
avoid obsetaces, analysis gLves the Pudrode relation-
ship. (8)
which is Ulnown as the i-creip law.
Both these macuaniass assume a Large degriee
of cold work prior to creep in order to create the forest
of dislocations. nowever, it has been recently denmostrated
that although well detinmd defornation occucs 4n the micro-
yield region, the dislocatLos are not extanenLve enough
.W45-
to form ectensive forests. ( 9) Thus, at stress below aad
si~ghtly above thte elastic 1LAit the i and logarithni•
creeps laws are not expected to be strictly valid.
It has been dmsnstrated that in the region
of the •-b or 1V"7 elastic limLt, permanent plastic
deaorwition tages place by the multiplication and motion
of pinned dislo%ationn and that this mechanism in
tersailly activated.(-' 4 ) With tie applicatlon of a
stress in the region of the elastic limit, dis•lo8ations
bow out creating 4ew dislocations Vftich move until
oistacl~es are mat. The dislocation Wualtipliaticn
ceases whers the back st"ess rom the blocked diiuslotions
stops the operation oz US source. ThIS is the itate
%ben mcr:ocroep beg .s.
ince the average sources active at the constant
stress level have been eohausted further deformatio can
only be accoalished by activating move dislocation
soCoes or by re-activating the daikausted sou0ces. It
Is aseuamd that this occurs uba a fluctuation of the
internal stress aided by thermal activation takes place.
?he creep rate fax this procees s ±: "1. "P-
dt
I I I I I I I I I I I I I I ' ' ' , i i
where N Ls tol number of dis1l.catLon loops per sqisre
contia-eter, A Las te area of active slip planes In the
specizan, b tne burger's vector, .a the rate of malti-
plication of disLocations, and n is the number of fluc-
tuations necessary to build up to the mLnisum stress
to activate a dislocatioa source. it we assume that
random stress tluctuations take place with a mean square
distribution around the applied stress -it, then:
.n (4)
From the stress-plastic strain curves of Figure 11, at
stress levels above the elastic Limit and stress
fluctuation Is given by&
V 1=• (5)
Bowever, sinwe the creep curves represent total creep
rather than the plastic cawonut of total croep, the
strams iluctuatiCna may not be defined by the plastic
strain curves but should fall nearer the stress-total
strain curves. Vr~ttlAq the stress fluctuation as
t" " tie •(6 )
0047-
,fte re O<: a ,..1.,P
dt 2
wkhich Integrates to
€2la t) + (8)
or
S• K (. t) i,=+].(
For the parabolic stresB dependence, m 14: and
Eq. (9) reduces to )y5,7( t) 3 /". The difftronce
AjtweeI the cuie root ties dependence o 1 -crmp and
the parabolic time dependence of equation ' a"ies
trom tie value of the stress fluctuations wbirch in
the -3-c p derivatLon were assumed to ke directly
proportional to the strain.
In the theory of P-creep the rate of dis-
location multiplIcation is replaced by the rate of
dislocation climb. To evaLuate w is *iauat
tko probabilit~y of creatLsAg a mobile &tislocotion. Xt hasn
ba~t eauutated t1A~t th.Le process tAkal Place mhGRe an
unpinped length of dislOcatiob bOws Out to &A Unstable
mie v~eier its radius ot curvatura reacbem hail tbe
orig inal pinned dislocat~ion length. FurtherInOte, this
lpra:.ss Ls tbermAlly activa ted sad aided by tAe initernal
&tress. Denoting the activoti4c energy as 0, the
activastion proba ility tor dialacatIon .ultipLicatiosi
Ina the lorward dixection is given bys
%dam"* 0a is the critical Strsas necessary to octivate
tb* avexage dislocation source. Xn the reve-rse direction,
the SAPteSUa is$±5
Biwae the rate of dAslocation saltpl.ctiosa
As given by' ttie product 99 the number og attempts to
climb tUs energy barter and the probabiLity at Over-
caim4A the barrier,, adding equations 10 and 11 Sivems
the ase% rate of dislooation multiplieation ase
kT
whore n is the number of sources and v the vibration
frequency of the dislocation lirne. Combining equations
9 and 1 Fields a tinal expression tor the ceep strain
asove the 1U"6 elastic limit stress of
is ( +l exp- (enal)k? (3)lkT -
or
L* J-2 ' j t siub 33.*i (314)
In FVi~ure a4-A41 the creep curves are
replotted on a logaritbmic. scale. The constants of
equation 1~4 were calcuLated from the curves at striasses
above the elastic Limuit and are tabulated in Tdable V.
the values of 09e rnn to vary betwen 1/.3 and 11W.
The value of Ilk results fron a stxess fluctuation
defined by the stress-plastic strain curve and equation5
whihle the 1/3 value wouald come from equation 6 with a -,- 1.
-504
ý4)u1
0
-4J
ol -1E
4J
'44
141
UUl
*.-
"-44
,4)(U
Ci)
$14
UTL~,4s IODC
00
.4-
,)
I[ d aU
,.-.
(r~.. T UL?..X~3 ~zuz
3
F *14
4
.1
1
'-4 -4
4J'U
IC
0 0
0 (1)* -4;4�) H
I'410 -
(.1)
-4 (1)
r) I�I4 -
ci)
U
1-4
ci))
*1-4
1-4 4 1-4
-_ 1 UTc214� d�o.t� I
IL
"44
-,4
1-44
'U
""5
*-4)
-2.2-
~~4-1
4-3I
ý4.
u~~~ -I1,7 4s o .[Z
CLn
-4
-4 1(1
QH(U::3
U-um-4 (()I
01
ý4
ý4
U)l
u-, a Da
:1i
41
cicmj
F--4
-i'
04.
44 ;
u-rea4s dz~jZ
ILI
44
1ý4
$13
* -S4
FT I I
II
qI
GJ
'U fV
~I-4
U.) IU . ,>•••
'. . ' ' 'I I I I
04
'u
uluaqo
K L
.. 7
H
0 'U
o
-I..
-N-
l , 0. • ,.) ,Cl)
.~
_' |.H
t.., ...
nv 310 356-T6 6u61 AZ92A
n .3 -3 u.J45 0.5
k cn3/mole U-3 Q~i1.5 U.2
A hrl 1.6xl1,0* 5 Ixloo*4 2.5xl106 3.Lxlci"'7
Q cal/mol. ~48,oou 62,,000 21,IUOO 24,(XJ0
A.nalysis of the activation 5nlty, , Yields
reaults very close to reported values for the activat~ion
energy or selt d•i£union. DOtLusionaOI type Orep, howIvOr,
was not observed gor the ;A49,A mmgnesium *=Opt at the higher
teaperatures. This may be duo to the low toutiAg stresses
at •hich i•Ltualonal cJOOp may not predominate.
At satressas below th] elastic lit small stress
fluctuetLonA will not result in instantaneous creep defor-
mation since the stress-plastic strain curves have LnfinLto
sloyes in this regIon. aweosr, many dislocation sources
are active at stress*s below the elastic limit and ýbese
Will. *"Cate,, contribuating to tke creep stjrain untIl
exhausted. As these sources are exacusted the creep rate
uhouid dimJnih eventuaally going to sero when no more active
sotarces exist.
fOoloing Seneral reacti& rate theory and
Oasvse• w tha rate of Cuslocation geneaition is of first
order, the rate of exhaustion of active sources would be
proportLoaal to the concentration of active sources
dt
-67II
which upon integration yields
N - 104t (06)
Since the creep rate is given by
~L~ t ~(17)dt
where b is the urugers nector, p the rate at which the
dislocation sources operate and X the number of active
sources, the creep strain is given bys
SX02 (l -••)(8
wbere A is the limiting strain after infinite tim.pIt to expected that the linitiAq strain should be daxetly
proportional to the stress and the te erature, that Is
with increased temperature and at higher stresses the
'imiting strain increases. Analysis of the curves of
Fgwmes 27-41 at stremses below the elastic limit indicatoa
that the limiting trxain can be expremsed as
WhOre 'a4 Ls the applied stress sad I the absolute tempersatre.
-68-
CcmbLng equations 18 &ad 19 gives a final emp• esrion-v- balr• •...x jj6 ifm-jo.- .m!t Og
%F 40 r WW %.F
r A (1-.0"t) "'? (
te constants calculatad from figures ••4mr•l are tab-
ultated ln ealie V1.
In su:my, the reaults soWW that .isrocreep
can be expressed by two relationships depending on
%tither the applied stress Ls larger or smaller than the
1(i-"6 *mtjC limLt stress, co"
pirther analyst. will Sim at resoxving the
an5maous behavior of A292A aagniuim. afforts will be
mtds to evaluate the constants from a theoretical aodll.
2be fcti•onsl dependence o* te creep strain on
temperatuare and applied steos will be analyzed from
theoretical considerations.
4. n Diim% Uses Chn ma deg 21orUNA Cga itioaa
The results of tve sealuatLons of 1009;h
changes in the five alloys under storage conditions at
75j! 50) &ad 2(.w4 F are shown in figures 4~.Z-16. fegati ye
lenath changea were observed itn al.l the alloy after
50 hours With the e 0ptLon oi the 6U61 aluauinm.
-69-
3'
�
.46.4
.794I
4�% Qii'.4- I0.43 I
EQ ;.) 3.4
14'0 4� I
.4I
6.4
I94I
.4
I4.
U)
'44
'44I
'4tj
0
U)
.- 4.-
(1-4
WNi
01"J
E-4
fa C:
C I u
0
to)
"144
U4.
01' 7ouq le-~UD
*-72-
24 4J
ILI) 0 L
(D
00
"r-4
1440
00
r- ".
-4 t
0i
'44
OT IbeoTUTUUT
I 4-
--
u 'X
r44
-tio~ Teoiszult
'D3
C))
- 0
w
r0
C-
t;I 0
rI rq
a~ueo Teo~sutul
2The 356-6 aImWinm and 310 etalUiess steel heoved a
slight increase in length followed by a decr*seo In 1ength.
The 6061 Aluamium showed an initial negative change which
slowly increased to an overall positive length change.
Tho 356-T6 Pluminum showed the greatest diuensional
lintawility. In all cases the magnitude of the divlnsional
changes were insignificant in comparison to the creep
changes under load.
The measurements, although not complete, show
dimensional changes similar to that reported by lUment
and Averb.ach.(1o)
4, -76-
A facility fox the Getexr• Ation ot distortioM
LA CastIsAs and Welded St~rtntUreS after heat treOatin
and machining operations was designed a*4 trio vecessary
*qtpment was pusrchoed and coustrited. fte fa#•lltty
Coisists of the foLlowing Items.
1. silger-watts TA-. autocollimator
k. RiLge.-wattes goo prism
3. e optically flat silvred mirrors
4. flat PYWOA platO Ld Ac li A 1/Se inC11
5. i standarw blacks
6. precision 44MUC stainless steel balls k,.375
inch ia diameter in several sise ranges.
2%e complete asawbl~y is momtmad on a granite
surf•a* plate f•.i±c is IsoIated from the floor by shock
alweormer to adaJlmie building vibrtion and ambLguJity
Ln tS es•ureMents.
""M distortion spacimen was sough MachLned to
the dimmimoas of figure 47 on a maiinLg sactine . On
the eternal faces eb oles were xoagh drilled with a
tap dxLU at the L•ic;atod points.
-77-
Following ageing a slot was cut as shown Ln
Figure 47 with an Ilox electrical discharge machining
unit with a brass electrode. The surtaceswere then
milled with all faces parallel to within v.001 inches.
The conical holes were then finish machined in a Moore
No. 3 jig-borer to a depth Of Q.233 inch using an
electronic depth stop. Carbide toolinj was used to
minimize tool wear. The hole dimension allows a u.375
inch steel ball to rest at a depth of U.100 Inch in the
distortion specimen.
The testing configuration is shown in Figurs 48.
The specimen rests on balls on the pyrex plate. The
standard block is placed against the distortion block.
For mesurements in the horizontal position the standard
block is fitted with a single steel ball. The mirror
rests on two balls on the specimen and one on the standard.
In this coufiguration a measuremant in the change of
height can Le made, and by turning the autocollimmtor
cross-haLis 90 the change In pitch of the specimen
surface can be masured.
-79-
.>-4
14J--r4
(i
-4
S--4
I 1-4-)
00
.,- 0
• 4 '•3 U.-4 £ () U---4 ,L -4 .,4(')
4 J CU (I)n
C) (U-A-ir
, .o
JA"I ""L
In the vertical position the standard bloch is
fitted with two balls. fTe sirror rests on these two
balls and one in the specimen. Rwasurewze s of the
distortion across the slot are made by resting the mirror
on three balls placed in holes on both sides of th6 *lot.
A boam of light from the autocolllmator passes through
the prism and is reflected by the mirror back through
the prism and into the autocollinator. The angular
divergence of the light beam in its passage appears as
a shadow displaced from the cross-hairs of the izitnimant.
This displacemnt can be read and from simple geometry
converted to the angle @ that the mirror makes with the
distortion block. By measuring this angle 0 before and
after a machining and/or beat treating operation, it Is
possible to measure the distortion due to these operation*.
By rearranging the speclamw, linar distortion of the
specimen can be aeasred in all directions.
Zn the event that distortion causes cetlain of
tte measurements to fall outside the field of the auto-
collimator, balls of difterent *Laos can be calibrated
and used.
Wg.ininaxy -wasureatim of dLetortLoa in i56-"6
S LUm"UNm riAV* 96 Sadie **ach bave desmaotirated the
tebqa high resolution a•s umisfuJleu. DiLstortion
vwil be det*±U as a functLoR of mbchhnLat operation.,
age hiddULng and BtseZss relief beat tmatmests for each
of th give alloys evaluated Ln th•s pogram and for the
thre alloys to b ev wamited Ln the coming year.
-a1a-
6. .. in
This work will continue in the next year under
a reeWS1 contract. Dimensional stability evaluations
will be conducted for H).A-T6 magnesium-thorLum,
beryllium-copper and H-i1 ausforamng tool &tool.
The analysis of microyield and microcreep data
will continue in order to understand the maechanisms for
these phenomena.
The measurements of dimensional changes under
storage conitLons will be carried out using auto-
collimation techniques. This technique will eliminate
errors due to slight temperature variations and wear on
the ends of the specimens from the comparator probe
A testing program to evaluate distortion as a
function of beat treating and machining operations will
be LuntLated as outlined in section 5. The five alloys of
the previo•s program as well as those mtiood a8boeo
will be tested after fabrication procodureo sLmLlar to
those comonly used in structural appliati.oms of each
alloy.
-8)-
7. Wa
kJ1Xi1 JanuarY 1957.
ýk Mwip Hop Avexbaf..bs m.LO and CCoenp stop Trans@ )'N5MP
3. loaiemtialdp 2.1k.. and Avexbacbp N.L,, Acta. Not., .
4., ift~bx*UCAh P.P. and Averbaech, B.L., M.X.T. MastorsTbOeLS4, J~anuary (1964).
5. broim, No and Luk~ens, X,]?. Jr., Act&. Ket., 2 (1961)p. 106.
6. 2thoii, D.A. and Averbach, B.L.p Acta. Mt., Z (1959)p. 69.
7. r@lthow, P., Proc. Phiys. Soc., j~k (1956) y. 1173.
8. relthanj P. and iibkakt J.D., Acts. met., 2 (L959)p. 614.
9. suits, J.C. and Chalmas a, ., Act&. net., (1961)p. ilj6.
1.0. Leanto 3.5. and Averimcbp S.L.,, M.I.T. LaatximamstatIonLab.,, Report 2-95p December 1955.