JOURNAL DE PHYSIQUE Colloque C6, supplCment

JOURNAL DE PHYSIQUE Colloque C6, supplCment a u n O l l , Tome 48, novembre 1987

A FIM/ATOM-PROBE STUDY OF PHASE TRANSFORMATIONS IN MOLYBDENUM STEELS

I. Stark and G.D.W. Smith

Department of Metallurgy, Oxford University, Parks Road, Oxford OX1 3PH, U.K. AlSXRMPFIM/AP,T!%,~ X-ray analysis ard dilatonetry irdicate that both 3 C precipitation and a solutedreglike effect are associated with reacticn retaz&tim i n F d H b allay. %ere is no e v i h for molybdennn or carbcn/wlybdenm clustering in awtenite.

INIIICDUCPIZH The inhibiting e f f e c t s of mlybdmm rn the t r a m f o m i c n of Fe++b alloys are d i a m m [1-6], but t h e i r origin d m mlear. What is clear is that l d of mlyb&mu of only a.few tenths of an atomic per cent significantly retard the rate of deccmposition of a t s t e d t e t o ferrite 4% carbids at a l l temperatures. This e f f e c t is particularly marked a t i n t e n d l a t e teapetahlres ( 50 C 703 C ) resulting in the deep lsys in the lllkwws of mny mlyWenrm steels. Similar effects are observed with other solute additions tut n b t o the sare degree a s with molylxbum. The molybdenum effect has been mriously attributed t o (1) Mo/C clustering in the azsrenite pinning the interface [4],(2)solutedrag at the transPonnaticn ineerfaoe [3,6,7] a d (3)Merphsse interface precipitation [8,9]. There is no strong preferential evidence for any one of these hypotheses, principally because the m , l y ~ c a r b a d i n t e r f a c einteractiau appear t o be on an a t i c d e and are ~ c e s s i b to t the experhmtal t d i n t q applied ~ ~ ~ previcusly. In the present iinrestigathn the uee of a field-ion microscope/atom-probe has pruvided the f i r s t analysis on a suEficiadly fine scale of the atomic processes associated with mtardathn both above and helm the b y region of the 'ITCarrves. A fieldanissim FlCFM has also k e n used to investigate grain-hdary -tion.

-

-

Fe+.20wt%C-0.52wt;ZMo and Fe43.32wt%C-2.05wt% alloys ere EhOSen on the h i s of their mnrkemy different TlT amw [ref.l p.119.,ref.2,p.106]. Transfonmtion t h ndmtes in fomer,ohtt can t&e hcurs or days in the latter. In the 2& all9 it is p a r t i q l y slow betwem 650 C ad 550 C and a d e p bayformsinthe~curve.Specimeus~haPDgenisedat12MCfor3daysad~qrarhed~obrine. Following reaustenltisation a t 1050 C they were quen&ed h t o a salt bath or transferred to -her furnace. Specimens were transformed as bars or discs of thfckwss 0.5-fmn i n order t o &eve wid cooling. No significant dacattmisaticn was observed. OFIM/AP analysis was carried out using a VG Scientific FIMlUl [lo] with the tfq hpxature set at 90 K. Molybdenum atoms image very brightly in necn and the disMbutim of mlybdenm c d d be deduoedfromtbFIM~. X-ray analyses *re obtained using the VG HD501 fiel-ssim SIB4 at the ZKAFA HaweU laboratories by Dr. J.M.TLtchm~9h. The TIT auve for a F e 0 . 3 ~ ~ - 1 . alloy 9 ~ was determind using dilatanetry. lhts is expected t o alloy as the tmperature rerge of each reacticn regime is a d i f f e r s l i g h t l y from that f o r the 2.0satsitive f m c t i m of mlylxbum content [2]

9

.

To t e s t whether clustering is assodated with the deep b y in the 2.05wtXEb allcy, the distribution of molybdenum i n austenite under these conditions was determined. In a l l instances neither FIM observatim nor atomprole analysis sIowed any -sth that ~ l y b & menl u s t e e had tdtaz p k . Ihe results obtained fmn a 20,0a)-icn sanple after 23 hours at 570 C are s b m in table 1 below. Ihe laJ X' value supports the lull hypothesis that the distribution of mlyWanm is rwdan. Analysis of the spct* between successive molybdenum ataos also -tsd a rzadaP rllstributim. S q l e axxi spa* amlysis also s l m d no widence of carbon/m,lybdenrm asmdation.

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1987673

JOURNAL DE PHYSIQUE

C6-448

Table 1 : % c c n p o s i t h of b a t a n sarples (20-atan step). Mo per 40-atan w l e Observed W t e d i f randan 0 1 2

633 282 67 18

h

622 297 69 11

%-squared

=

prolnblllty

5.5

>

11%

(2) pol/AP ad 2PN anal+ af ferrite 0 Optical microscopy revealed four t m f o n n a t i c n regimes between the u p p r nose a t 730 C ad M , cad1 of which was slwwn by FIM/AP analysis t o be associated with a particular type of k / C interactfa,. llle aralyses are smmarised in table 2 and the mLcrmtnrctures i n figs. 1-3 d 6. Dilatometry slwwed t h a t the lower transformation region, usually represented by a sirgle set of TIT-curves, i n f a c t comprises two closely-overlapping regimes ( f i g . 4) with distinctly rLCEfere~u mlcrc6tructures. Ihe four g n e r a l reactial reginw wre: 0 (a)Above 590 C: The i n i t i a l decomposition product ws ferrite which f i r s t F o n d alorg t l e prior austenite grain-boundaries (fig. lb). Throughout t h i s range ahnost a l l of tleilulylxhrun kcwas incorporated i n t o carbides. I n i t i a l l y , these a r e approxhwtely stoichiaetrlc YC du?re sole of tlle molybdenum s i t e s a r e occupied by ircn uterus. In the latter s ~ a g r s ,wlcre ~ r l l t fonlr;, e boLl~Elf tuxl molybdeny-enriched E C are 0@3erved i n the FlM sp~ji~m. arecLpltdLe wra ~ n i ~ drailers y (1(~200& X 3 0 0 Aa t 730 C and spheraida ( l ~ 2 . 5 d ) a t 670 C- 90 C (fig.la). Mos~of UE! rallets were similarly oriented alttlere a F w occepticns The size iud s l q of ~ t k railers was rile sse after both ooe how: a d ane day at 730 C. A n example of the mlcroetructure f o d a t 595 C and associared with slow r e t i a t kLncLLcs is s l m i n the electron micrograph (fig. 6). 'Iho density of dislocations in tlle ferrite w a particularly ldgh a d carbides.

-

9)5C

vre

( b ) ~ t570'~: grain-boundary dotri-rp118, needles am3 blocky f e r r i t e =re all olselwd (fig.2b). The range of f e r r i t e morp1~Logiesm rdlected i n an e q d l y diverse r a g e of p r d p i t a t e s . Figure 2a shows a case h e r e rcda and elagated plates precipitated in two 100 directlcns. In otirer instawes elle rods were aligned a l q a single dfrecticn or were present as an array of rcugt~splrralds of raditts MR. Because of the s i z e and of the carbides, the analyses w r e certain! to cuztah s a w o x i t r i h ~ t i m from the f e r r i t e , but the observed Mo:C r a t i o of just less than 2:1 ws identical t o that at higher temperatures. 0 (c)510-555 C: blocky needles -re the doninant rroq>lmlogy (fig.3b). 'lltese c o n t a W sralL (1-10fi3), b r i g l 1 t l p i m g l ~preclpitatea a c h appeared i n clusters t-acller t l m a t rardan i n tJw fcrrlte. nSe =re present i n very high concentratiaw a t the l m a r g l e b a u d a r l s ktferrLte d t s (Eig.3a). 111eLr dlstLnct electron d i f f r a c t i o n pattern indlooted tl& tlry w r e preclpitatcs rat1:lrcr tlun cluskrx. rile average ratio of mlybdennn t o carbon i n these precipitates ws al>proxhmtely 2:l. Ihe dawity of single bright spots i n the f e r r i t e irdicated that i t retained a substantial amunt of oolybchm arl Llds ws c o d i d by the analyses. TP1 slmwed that d y of the ferrite units containerl cementite wlrreas F C m s aodprsent. n~ cementite shape and distrihltim d t e r 8 mlmtes m t e d tW r e p l m a s k of C by FSC I d IX* y e t become significant. I h e i m d i c a t i m was that m s t of t l e KC observed 1 4 fon2d1u-LNI austenite decomposition: The single cementite particle analysed o x i t a i d t l e bulk mlybbmm level

-

T%.76~0.24

~pc~.98~0.0?.3)1~ 0 (d)Below 500 C: packets of fine, characteristically b a l d t i c wdles fonnrl. llle a d y precipitate present i n these d e s vas cementite.

I n t h i s f l l l o y , t h g transformation was complete witllin tllc f i r s t ftw dmtm a t a l l teqeratures between 510 C and 725 C. t$C precipitation was observed thraghoat the rar\ge lut f e r d t e 1 ~ 1 y L d 3 m u 1-1s bere generally close t o the W c level (fig.8b). The AP analysts are s l m i~n table 3. 0 0 0 (a)*620 C: Sphcmldal t$.C f o m d a t 620 C a d 725 C. (he of the specittransfonued for 103 s x o d s a t 620 C cartalned a highlydemrated I ~ Q h - ~ fl ecr d t e t f e r r i t e bouodary. llle docorarim \mi restricted t o the intedece i n the form of a sub-mmolayer covering rather t l m an array of Interface precipitates. Evaporation resulted i n the interface mving a s a smooth curve towards t l r tip apex with the me e z e p t i m slnm in figure Sa, whem i t was kinked taranls a Y C precipitate. Analysis w a w r i e d out

when the boundary was close t o the apex and the zone of h t e r s e c t i m of the interface with the t i p surface was wider than the probe hole. I n t h i s way the boundary contribution was mdmised. 'lhe mlybdenrn and carbon 1 4 =re 15-20at% and 10-15arX respectively. Zhe ccnpoaitim profile (figme 5b) shlws that 5-lOaa of 8ulpiur and sane phosphorus were a@ present. Identical decoration WE observed m two occasicns in this alloy aFter tramformation a t 525 C. Ccnsideraticn of the awtedte grain size and the mmber of specimens dried w s t e d that it was d k e l y that a U of tkse were prlor austenite grain Lumdari83. (b)595'~:

TIN s t w d

-

M2C precipitation i n dl ferrite grains f n m

the emliest

steges.

(c)52?j0c: The low-angle f e r r i t e / f e r r i t e boundaries =re uuch less hmvily decorated thm thase formed i n the 2.05wt%Mo a l l o y under tlle sam caditions. Zhe diffeWE M yw r t h w the relative d i f f e r a r e Cn bulk mlyWeram content.

m x-ray d p s i s of tran9formtial i n t e r f ~ ~ e s Analyses obtained from transformagion i n t e r f a c e s in the 0.52wtXMD end 2.0at2ays aEter 15 seconds and 23 hrxlrs respectively a t 595 C are slum with the associated SIPI higea of precipitatefree interfaca (ffg. 7). W y s i s of th sl-t g r a d q interface seetiara in the 0.5d a y , which are t h e l i k e l i e s t s i t e s of segregation, is extreoely difficult due t o their highly c-luted nature. 'Ihe smooth interface s l m i n the ,SIlPl ndcrcgraph is s d t a b l e for analysis, but m y nct be represmtative of a l l interEaces. Molybdenm a d phospllorus enrichnent is observed in the -2. allqr but nct in the 0.52wtDb alloy.

The analysis shown i n t a b l e 1 was obtained w h e r e ~ 2 0atoms were detected per 1x of material evaporated. Additional analyses wre oMraIned over a range of &table q l e and step sizes and in no instance vas tllere any evidare of clustering. 'Ihe analysis of the betwsm irrllvidual w1yMenm atoms shows no clustering m on atunic scale. 'Ihe s w i t i v f t y of the spadrrg analysis is reduoed ly the 5CB: inst-t detection efficiency and the 2-3nn diameter area f n m where auressively-Qtected icns my originate. Consequently the AP analysis does nct totally disc& the possibility that sam 2-3 atom clusters fom, altkmgh i t makes it very fnprobable. Because carbon diffuses when mrterrsite forms on ,-q the observed distributicn will nct be the same a s i n the austenite. Therefore the observed absence of a carbm/mQ%&nm aasodatfm is of q u e s t i d l e significam. Precipitation end Segregation The m s t obvious result of the imrestigoticn is that there is saue C predpitaticn i n both alloys throughout the intermediate transformation temperature range (i.e%st a h and kdcu the by). Dilatometry has a l s o sllown t h a t MZC precipitation just below the deepest p a r t of thelTl!bey is associated with a decomposition process that differs distinctly fraa the w e d U t e umphology observed a t lmr tenperaturn. FIM/M analyses shows that tlle rate of reactim a d the mcmt of t$C predpitaticn are inversely related. That t h e cxtent of precipitation is not s h p l y related t o the differanes in c e t i m is conEinned by tables 2&3a c h glve the mlyb&mmI c m e n t of ferrite end s b m that precipitate &mity depends on more than the bulk molyLxhm level. It i s the effects which came precipitation t o teccme extensive that a r e the centre of debate, t h e p r i d p a l . questicn bekg "Is prectpitatim a cause or effect of the d q e s i n i n t e r f e velodty a d transformation rate?". at a Solute-drag tlieory predicts that mlybdenm is segregated a t the interface, which then r a t e a t which t h i s enhanced molybdenum l e v e l can be raintained. Prectpitatim a t these slcu mwlrrg boundaries i s tllils a secondary process. It is almost certain that interfthat are so intemely decorated nuve rrore slowly than tlwse tlmt are solute-free. Zhe relatimship tetwsm p d p i t a t e h i t y and reaction r a t e is m r e ccnsistent, althougti the precipitate density observed a t a f d t e / a u s t e n i t e i n t e r f a c e seems t o be too low t o N l y a m m t for i t s haw velocity and m X-ray analysis slum the precipitate-free regions i n the 2.05wt.%Pb alloy a t 595 C t o be highly auiched in mlybcknm. At the some temperature i n t h e 0.52wtX Mo alloy, the f e r r i t e / a u s t e n i t e interface i s not obvicxlsly mlybdemwemiched. It seems l i k e l y t h a t a cmbination of the tw processes aocamts for the o b d behavlau. Zhe r e l a t i v e importance of these & q e 6 with canposition, precipitation irrreaaf&y daninant at higher molylxknnu and c a r h levels. Because of the attractim of w1yWenm f o r carbcn a d because of a

-

C6-450

JOURNAL DE PHYSIQUE

s i z e effect, a higher wlybdenum l e v e l is expected at the i n t e r f e i n all of the alloys, h r t t k evidence s u g g e s t s t h a t transformation is strongly inhibited c d y &ere there is significant precipitation. The pinn&tg e f f e c t s of C a r e suggested by the high density of dislocatians in the ferrite a t 1 9 a d~525 ~ C in ~t k Z.OS*rdl: dlq the interface kink d a t e d wffh c e of the fat precipitates formed i n the 0.52wtXMo alloy a t 620 C. A t higher taperatures, *re the energetically expensive s t e r a t i o n of lattice Qfects is not possible, different p r m s e s may M b i t tnursfomticn. F e r r i t e folmafipl my require ledges that are high emugh to overgrow interphase 5 C or, &re 5 C fonus a t the growth f r m t i n a eutectoid -r, tnursformtian proceed at a rate gwemed bj intedaoe wlyMenm diffusial.

9

axmmms (1) No evidence for Mo o r Mo/C clustering hm been ohenred in amtenfte in FeC* alloys u d e r caditione t h m t r a m f o w a t h is stralgly irhibited. (2) Some 3 C formation accaqades trarrrfolmaticn both above a d below the bay in both alloys. The extent of this precipitation is not sinply related to the the hulk of velslybderun a d carcarbon. (3) The ' 'molybdenum effect" is probably a combination of the e f f e c t s of segregation and precipitation. The r e l a t i v e importof these can charge ctring the aurse of transformtian as the austenlte becomes progressively carbon-enriched. Significant effects are associated with extemive irmrporaticm af mlybdenm into 3C.

-

(1) AW,Atlas d Iaothermnl wd Cmlim Trwsformation M w m s . (2) 'Ihelning,K-E,Steel & its Heat lhstnents.httenwths, 2nd ed. (3) 8-xwLl e t al, M.Preps. and Ph&e Trws. i n Mat., &t.Soc. 3Lpe,(1986) ,p.445. (4) Sharma,R, z d rndy,G.R., Met-Trws., 4, (1974), p.939. (5) Ehkshia,H.K.D.H., J.Wt.Sci.. 3 (1983), p.1473. (6) Aarmsc*1 e t al, ihid, 3 (1985), p.4232. (7) Bumto,M., and Aammm,H.I. @t.Trans. (1986), p.1385. (8) Purdy,G.R., Acta Met. (1978), p.477 (9) Purdy,G.R., ibid, p.487. (lO)Ceteeo,A. et al, J.de PIysw6,artppl.t o 3(1984),329. (11)CTabkqH.J. e t al, Surf.& Int.Analysts,~p.2(12,(1987)

a.

,a

~abl. 2'

o~

Time at T ( C) (minutes) 730 60 725 1200 420 670 1680 620' 590

Ferrite atXMotC 0.05 0.03 0.20 0.08 0.07 0.05 0 0.01 0.01 0.13 0.04 0 0.13 0.08 0 0.06 0.06 0

;;:: 1140 34 16 8

470 430

4 0.5

Atom-probe analyees of re-0.33wtXC-2.OSutXHo atXC 0 0

6600

550 540 533 526'

8

*o

0.06 0.03 0.04 0.02

0.09 0.03 0.10 0.02

0.13 0.70 0.33 0.8 1.22 0.91 1.02

0.04 0.13 0.05 0.16 0.15 0.05 0.06

1.18 0.09 1.30 0.06

atXC f a 35 1.0 33 1.6 40 2.5 33 1.7 25 1.4 43 8 44 2.7 381.5 343.4 32 10 34 6 25 6 24 2.0 31 4.3 25 1.2

Carblde atXM0tG 56 1.1 63 2.2 60 3.1 65 2.4 5.7 0.7 57 9 56 3.0 622.0 66 4.6 68 15 66 8 75 9 1.0 0.4 0.6 0.6 1.2 0.3

I r Atorprobe analyses of ~e-0.2OwtXC-0.52wtXMo

Table 3

8

Time at T (minutes) 725 16 625 1.7 6001 30 526 8

Ferrite atXC f G atXMofG 0.09. 0.02 0.20 0.03 0.13 0;02 0.31 0.03 0.07 0.02 0.38 0.05 0 0.19 0.04

513

0.04 0.03

( C)

*I 1

1.1

0.36 0.07

etXC t G 33 15

--

29 22 43 23

7.5 1.2 11 2.3

Carbide atXMotG 44 11

--

71 1.1 57 1.2

12 0.3 14 0.5

In the mjority of caaes tlie analyses given summarlse those obtalned traeveral apecimena. The carbides ere too m a l l for the Fe content to be eccuretely determined. lko dlstinct forms of ti C have been enalysed.In one I4 is principally Mo. whilst i n thn other i t fsbmainly Fe.

Pigs. 1-3

ELM observation of the cllanges in &C precipitation and overall ferrite 0 526 C in a Fe-0.32C-2.05Ho w t X alloy.

Bi .

*

I *

r

5-

Fig. 2b

~ i g . 2a 5 7 0 " ~ , 95 hours

Fig. 3a

0 526 C, 8 minutes

70@l I

~7 I

10%

\

23 hours

Fig. 3b 5400C, 16 minutes

*

_./ .................... ----.--' z:

570'~.

Figure 4 Dilatometrically determined TTTcurve for Fe-0.33wtXC-1.96wtXMo. The 10% transformation curve is shown.

JOURNAL DE PHYSIQUE

C6-452

F i g u r e 5a

F i g u r e 5b

Fig. 5 I n t e r f e r e n c e of a 3 C p r e c i p i t a t e w i t h a h e a v i l y d e c o r a t e d f e r r t e / f e r r i t e boundary and a n a l y s i s a c r o s s t h i s boundary. Fe-0.20wtXC-0.52wtXM0, 620 C, 100s.

d

Figure 6 Micrograph of Fe-0.32wt;bC-2.05wtMo showing a transformatdon i n t e r f a c e a f t e r 23 hours a t 595 C. Note t h e h i g h dislocation density in f e r r i t e , p a r t i c u l a r l y around M2C.

0.4pm Fig. 7 Fig. 7a

STEM X-ray a n a l y s e s of f e r r i t e / a u s t e n i t e i n t e r f a c e s .

Fe-0.32wtXC-2.OawtXMo 2 3 hours 595 C.

ferrite

Position i n t e r f a c e (1) " (2) " (3) lOnm i n t o f e r r i t e ( 4 ) 50nm i n t o a u s t e n i t e ( 5 )

wt XMo 4.0

3.9 1.8 1.O, 1.3

wtXP

1.1 0.7 0.6 0.6, 0.5

Position ferrite t I t t

austenite I t I t 11

i n t e r £a c e II I I

I t

*

= Less t h a n 2 s t a n d a r d d e v i a t i o n s .

1.o 0.4 0.3 0.2 0.2 0.6, 0.3, 0.5, 0.5

NoPin any of t h e analyses.