Technical Note:Theeffectofstrain rate onthemechanical ... · PDF fileAustempered ductile iron(ADI)isaductile ironwhichhashadits matrixtoughened byan austempering process. Thealloyexhibits

~,

mr!"f~

(;~'~',i" '

"'.'.'4' .

~~.

.."'

. -.. ''" :{::'J,':;:I/::.c.~;;;;:i."~C:.i:.

Synopsis

Austempered ductileiron (ADI) is a ductileiron which has had itsmatrix toughened by anaustempering process.The alloy exhibitsremarkable properties,which it owes to aunique microstructurecomprising graphitenodules embedded in amatrix qf adcular ferriteand carbon-enrichedaustenite. Theproperties include highstrength combined withhigh toughness andductility. Excellent wearresistance is alsoexhibited because,under strain, theaustenite is tranq'ormedlocallY to martensite.

Samples qf ductileiron were austemperedto produce grades II andIV AD!, and were testedat various strain rates.The grade IV ADIexhibited higherstrengths and lowerductilities, and thestrengths were found tobe more dependent onstrain rate. Transfor-mation to martensite

* Grqyton Cc, P.O. Box

12513, Benoryn 1504.

t School if Process andMaterials Engineering,University if theWitwatersrand, PrivateBag 3, Wits 2050.

@ The South iifricanInstitute if Mining andMetallurgy, 1995. SAISSN 0038-223X/3.00

+ 0.00. Technical notereceived Feb. 1995;revised technical notereceived Aug. 1995.

Technical Note: The effect of strain rateon the mechanical properties ofAustempered Ductile Ironby B. Eatwell * and L.A. Cornisht

Introduction

Austempered ductile iron (Am) is a ductile ironwhich has been subjected to an austemperingheat treatment to produce a matrix of acicularferrite in carbon-enriched austenite.

The initial composition of the base alloy isimportant to produce a good-quality ADI. Alloyingelements are usually considered in groups: thosewhich are carbide formers and segregate betweenthe nodules, and those which segregate prefer-entially to the nodule-matrix interface!.2. Thepresence of certain alloying elements affects thelocal duration of austenitization and the austem-pering reaction. The carbon equivalent isrecommended to lie in the range 4,3 to 4,6 percent by weight depending on the sectionthickness3. Thicker sections usually need alloyingelements, e.g. copper, nickel, and molybdenum,to promote hardenability. The high silicon contentsuppresses the bainite reaction (which wouldoccur in most steels), and an intermediate micro-structure consisting of ferrite and carbon-enrichedaustenite is formed instead. The structure ofacicular ferrite in carbon-enriched austenite iscommonly referred to as ausferrite.

The austempering heat treatment is affectedby the chemical composition, time, and tempera-ture. The components are austenitized at atemperature around 900°C to allow sufficienttime for the matrix to be completely transformedto austenite4. The austenitizing temperature iskept to a minimum to avoid reduction in theductility, which occurs as the temperatureapproaches 900°C. The material is then quenchedto the austempering temperature (between 250

and 400°C), and is held isothermally to producethe required microstructure. The holding time atthe austempering temperature is usually between1,5 and 4 hours5 and is dependent on theaustempering temperature. The austemperingtime is a minimum of 1 hour, and depends onthe alloy content. If the alloy contains elementsthat segregate to the nodule-matrix interface(e.g. copper), the austempering time should beincreased to allow a maximum amount of carbonto diffuse from the nodule back into the matrix.The austempering temperature determines thegrade of the Am, and is the key to the productionof the desired properties. In common with othercast irons, the grades are specified by mechanicalproperties rather than by simple composition.For specification purposes, the material mustexceed given minimum values for yield andultimate tensile stress (UTS), ductility, impactenergy, and hardness. These minimum valuesare given in Table I. To ascertain the grade for aparticular shape of component, an individualtrial must be undertaken, and data from othercomponents cannot be substituted.

During austempering, acicular ferrite isformed, rejecting carbon, which is absorbed bythe austenite, thereby stabilizing it. The amountof carbon in the enriched austenite can liebetween 0,8 and 2,0 per cent because austem-pering occurs in the austenite + carbide field,which allows higher carbon contents inaustenite. This renders the martensite transfor-mation more difficult, which gives the materialits toughness. At high austempering temper-atures' a coarser structure is produced, whichhas lower tensile strength, and usually greater

Table I

Specification for various grades of austempered ductile iron (ASTM 897-90)

MinimumUTSstressMPa

8501050120014001600

The Joumal of The South African Institute of Mining and Metallurgy

* Un-notched Charpy bars tested at 22 :!: 4"C

Minimumelongation

%

269 - 321302 - 363341-444388 - 477444 - 555

Typicalhardness

BHN

10741

n/a

NOVEMBER/DECEMBER 1995 357 ....

Mechanical properties of austempered ductile iron

was more prevalent atthe lower strain rates. Amechanism if increased

}low stress and workhardening in theaustenite if the matrixis proposed to explainthe higher strength ifthe samples at highstrain rates, wheresignjficantlY lessmartensite isformed inthe matrix.

Acknowledgements

The authors thank Graham

Clarke of Grayton CC and

Martin Reeves of Mineral-loy

for their help and financialsupport, and the Electron

Microscopy Unit at the

University of the

Witwatersrand for providing

the facilities.

toughness and fracture toughness. Conversely,lower austempering temperatures give a structurewhich is finer, and has more strength, hardness,and wear resistance6. The time at the austem-pering temperature is also important: too short atime results in unstable austenite which is readilytransformed into martensite and is usually morebrittle. A small amount of fine martensite can bebeneficial in improving both the wear and the.fatigue resistance. On the other hand, too long atthe austempering temperature means that theaustenite decomposes into bainite, Le. ferrite andcarbide, which results in embrittlement. Theprocess control during austempering must avoidthese extremes. The ideal treatment times areshorter at high austempering temperatures, andlonger at low temperatures3. At low-temperatureaustempering, carbides can form.

Very little work has been done on the effectof strain rate on these materials, but Kobayashiand Yamamot07 report that elongation decreaseswith increasing strain rate. This would beexpected if martensite was being formed. Theaim of the investigation described here was toshow the effect of the strain rate on selectedmechanical properties.

Table"Composition of the experimental ADI (inper cent by weight)

Composition

3,542,460,270,0380,0030,070,0620,530,04

TraceTrace

0,001

Experimental procedure

The samples were analysed to ensure that thechemical composition was suitable for theproduction of Am (Table II). The microstructurewas studied for its suitability, and was found tocomprise carbon nodules surrounded by ferritein a predominantly pearlitic matrix. The carbonequivalent of 4,36 per cent was found by use ofthe following formulas:

Carbon equivalent =% C + % Si/3. [1]

This carbon equivalent, together with the nodulecount of 200 nodules per square millimetre andthe high pearlite content (Figure 1), gave asuitable ADl. Ignoring the graphite, the propor-tions of pearlite and ferrite were 88 and 12 percent respectively, which is considered good forADl.

The samples were austenitized in molten saltat 880°C for 2 hours. Austempering wasconducted in a nitrite-nitrate bath for 1,5 hoursat both 310°C and 380°C. The austenitizing timecompensated for the presence of copper, whichsegregates at the nodules and is a barrier tocarbon diffusion. Hardness tests were undertakenon the initial heat-treated material to ensure thatthe alloys were within specification.

Mechanical testing comprised hardness andtensile tests at two different strain rates. Thetensile test specimens were machined from keelblocks to a gauge length of 30 mm and a gaugediameter of 6 mm. Precautions were taken toensure that the keel blocks were the same sizeso that the nodule counts and matrix cell sizeswould be similar. Four specimens were tested foreach condition. The different strain rates wereproduced by cross-head displacements of 0,5and 6,2 mm per minute. The average strain ratewas deduced per specimen, and was found tovary slightly but to be consistent within eachcondition. The hardness was measured after thestraining, the average of five indentations beingtaken. Microstructural examination by opticaland scanning electron microscopy (SEM) wasundertaken before and after the straining, andthe specimens were also examined underpolarized light to check for the presence ofmartensite. The latter was also checked by theuse of different etches, such as 2 per cent nitaland Vilella's reagent. The fracture surfaces werestudied by SEM.

X-ray diffraction was conducted to confirmthe phases present. Copper Ka radiation with awavelength of 0,15406 nm was used, and theincident angle was varied between 2,5 and 45degrees. As the spectra showed very highbackground-to-peak ratios, they gave onlyindications rather than precise values.

Figure 1-Microstructure of the ductile iron, showing graphite spheroids in a matrix ofpearlite, with ferrite surrounding the nodules: a classic 'bull's eye' structure Ix 100)

~ 358 NOVEMBER/DECEMBER 1995 The Joumal of The South African Institute of Mining and Metallurgy

Mechanical properties of austempered ductile iron

Table 11/

Mechanical properties of the different grades at different strainrates*

Figure 2-5EM micrograph in secondary mode: matrix of grade 11ADI, showing coarseferrite plates in a matrix of austenite. Bar represents 10 IIm

Figure 3-SEM micrograph in secondary mode: matrix of grade IV ADI, showing a featherylath network of ferrite, with seams of stabilized austenite and martensite. Bar represents10llm

The Journal of The South African Institute of Mining and Metallurgy

Results

The results of the initial hardness tests on theunstrained material (415 BHN and 331 BHNrespectively) gave a rough indication thataustenitizing at 310 and 380°C had producedgrades IV and II AD! respectively. This wasconfirmed by the results of the tensile test, asshown in Table Ill.

It can be seen that grade IVAD! austemperedat the lower temperature, when compared withgrade II AD!, exhibited higher strength (shownby the yield stress and UTS), and higher hardnessand lower ductility (shown by the elongationand area-reduction values). The UTS showedless sensitivity to strain rate than the yield stressin both grades. GradeII showed a 7 per centincrease in yield stress when tested at the higherstrain rate, whereas grade IV showed a 6 percent increase under similar conditions. Confidencetesting showed that the increases for each materialwere significant at 99,9 per cent confidence levels.However, the difference between the increases inthe two grades is not deemed significant.

The differences in UTS with strain rate wereless significant. The strain rate appeared to havelittle effect on the ductility, which is contrary toKobayashi7 but, as the tests were undertakenunder different conditions, the results are notdirectly comparable. Although an increase instrain rate increased the yield strength in theindividual grades, the effect on hardnessappeared to be the opposite, which is unusual.

Figures 2 and 3 show the microstructures ofthe unstrained grades II and IV respectively.Coarse ferrite plates in austenite can easily beseen in Figure 2, whereas the structure is muchmore angular and finer in Figure 3. The angularappearance was primarily due to the fine lathstructure of the ferrite, rather than to theintermediate bands of austenite and possiblymartensite. The sizes of the austenite grainswere estimated by identification of regions ofparallel ferrite plates or laths, and were verysimilar, with grade II slightly larger.

Figures 4 and 5 are secondary and back-scattered SEM images of unstrained grades IIand IV. The images show a contrast between thetopography (secondary electron mode on the lefthand) and the composition (backscatteredelectron mode on the right hand). Figure 4indicates two phases (ferrite and austenite) ,whereas in Figure 5 it can be seen that there arethree phases (ferrite, austenite, and martensite).The contrast is limited, because there are onlysmall compositional differences between thephases (up to 2 per cent by weight). The phasescan be identified in order of increasing carboncontent, the darkest being ferrite plates, thenmartensite and, finally, the lightest beingaustenite, which is not in lath form and appears

NOVEMBER/DECEMBER 1995 359 <l1li

Mechanical properties of austempered ductile iron

more equiaxed. The austenite would be expectedto have the highest carbon content since it iscarbon-enriched. Attempts to reveal themartensite more clearly by etching were notsuccessful, but observations with polarized lightagreed with the above.

Micrographs of the material after it had beentested indicated more martensite in the grade IVmaterial which had been subjected to the lowerstrain rate. The microstructures of grade Il weresimilar, and consisted mainly of two phases.

The fracture surfaces showed very littledifference with strain rate. Fracture had occurredin the matrix, and the graphite nodules werepulled out. The grade Il material fractured withmore ductile features at both strain rates.

Figure 4-SEM micrograph (secondary electron mode on the right and backscatteredelectron mode on the left), showing matrix of grade 11ADI with coarse and distinct ferriteplatelets in austenite. Bar represents 1 IIm

Figure 5-5EM micrograph (secondary electron mode on the right and backscatteredelectron mode on the left), showing matrix of grade IVADI with ferrite laths surrounded byaustenite, and (darker) martensite between. Bar represents 1 IIm

~ 360 NOVEMBER/DECEMBER 1995

Discussion

Ferrite plates, or laths in the finer morphology,were observed in the stabilized austenite. Theheat treatment was successful in producing therequired structure in the grade Il material, butthe presence of martensite in the unstrainedgrade IV material was an indication of aprematurely terminated austempering stage.

Mechanical testing indicated that the appliedstrain rates affected the yield stress, but not theUTS nor the ductility to any great extent.Although the mechanical effects were similar,the microstructures were different for the twogrades, and the same mechanism is unlikely tohave been responsible in both grades. In grade Ilthe ferrite was present in the form of coarseplates, whereas in grade IVthe laths were verymuch finer, and the grain boundaries wereclearer. The finer structure is expected from thedifferent austempering temperatures. Reduceddiffusion rates at lower temperatures wouldresult in a finer structure. The orientation of theferrite phase in both grades provided an indica-tion of the original austenite grain size, and thegrade Il specimen had a slightly larger austenitegrain size. This might be expected, since it wasaustempered at a higher temperature. Thefineness of the ferrite structure would increasethe strength of the material, as was shown bythe higher strength of grade IV. The apparentimproved ductility of grade Il could originatefrom the coarser ferrite structure, but theincrease was not significant. It is unlikely thatthe different austenite grain sizes were sufficientto promote any distinguishable effect.

The presence of martensite was at firstpuzzling, since it is supposed to form understrain. However, the grade IV specimenscontained marten site prior to straining, whichwas deduced to have originated from a prematuretermination of the austempering stage. Bothsamples were austempered for the same period,even though it is known that the best results areachieved by austempering for short times at hightemperature, or longer times at low temperature3.The period of 1,5 hours was chosen because itwas thought that this was long enough to allowtransformation at lower temperatures.Preparatory work had indicated, for short timesat least, that there was no apparent risk ofbainite formation when the material was heldtoo long at 310°C. However, the occurrence ofmartensite even before straining shows that theperiod was too short, and some austenite, whichwas not sufficiently enriched in carbon tobecome stable, was transformed to martensitewhen the specimens cooled3. The X-ray resultsindicated that the grade IV material alsocontained carbide, but this was not distinguishedmicrographically because it is formed in smallparticles at the ferrite-austenite interface. Itspresence could also explain the lower bainite-likeappearance. Carbide is more likely after treatmentat a lower temperature.

The Journal of The South African Institute of Mining and Metallurgy

Mechanical properties of austempered ductile iron

References

1. KOVACS,B,V.,KEOUGH,J.R., andPRAMSTALLER,D.M. Theaustempered ductile ironprocess. Final report toGas Research Institute,

1989.

2. KOVACS,B.V., andKEOUGH,J.R. AD! anengineering material,iron integrity casting.Proceedings worldMaterials Congress,1988.

3. VOIGHT,R.C. Austemperedductile iron processingand properties. CastMetals, vo!. 2, no. 2.1989. pp. 71-93.

4. KOVACS,B.V.Austempered ductileiron: fact and fiction.Modem castings,Mar. 1990. pp. 38-41.

5. RAcE, J., and STOTT,loPractical experience inthe austempering ofductile iron. Heattreatment if metals.1991. pp. 105-109.

6. GUNDLACH,R.B., andjANowAK, J.F.

Austempered ductile ironcombines strength withtoughness and ductility.Metal Progress,Ju!. 1985.

7. KOBAYASHI,T., andYAMAMOTO,H.Transformation inducedplasticity in austemperedlow alloyed ductile iron.Trans. japan.Foundrymens'Soc.,vo!. 8. 1989. pp. 30-34.

8. KEOUGH,J.R. Thedevelopment, processingand application ofaustempered ductile iron.World CorJferenceon ADI.A.S.M.E., 1991.pp. 640-641.

9. SIDjANIN, lo, SMALLMAN,R.E., and BouToRABl, S.M.Microstructure andfracture of aluminiumaustempered ductile ironinvestigated usingelectron microscopy.Mater. Sci. and Tech.,vo!. 10. 1994.p. 711-720.

10. MAYER, P., VETTERS,H.,and WALLA, J.Investigation on thestress induced martensiteformation inaustempered ductile iron.Proceedings2ndInternational CorJferenceon AustemperedDuctileIron. A.S.M.E., 1986.pp. 171-178.

Although the fracture surfaces showedcharacteristics of both ductile and brittle failure,the grade IV material exhibited a greater area ofductile failure in the material of low strain ratethan in that of higher strain rate. The amount ofmartensite was higher in the more slowly strainedspecimen, and the yield stress was lower, than inthe material of faster strain rate. Therefore, thefactor contributing to the higher proportion ofbrittle fracture in the faster -strained specimenwas not martensite. Sidjanin et al.9, in aninvestigation of an Am with 2,2 per centaluminium, by weight, found that the occurrenceof brittle fracture was associated with carbidetype E and particularly X,and a fine plate spacing.when carbides, particularly

X'precipitated on the

ferrite-austenite interfaces, deformation wasconstrained, and brittle fracture resulted. The X-ray-diffraction analyses indicated that the gradeIV material contained carbides, and this materialshowed the most variation in failure mode.However, the proportions of brittle failure weremuch less than in the alloy tested by Sidjanin etal.9, and the composition was also different.

Martensite was difficult to discern by SEM inthe grade II material. However, X-ray analysisand optical microscopy using polarized lightconfirmed the existence of martensite in thesamples of both low and high strain rate. Thegrade IV material contained marten site bothinitially and after being strained. As with thelower-grade Am, there was more martensiteafter straining at the slower rate. There was lessaustenite in the material subjected to the slowerstrain rate than in that subjected to the fasterrate, and less austenite in the faster-strainedmaterial than in the unstrained material. Theseresults suggest that, under strain, the austenitewas transformed to martensite, but that theeffect was more pronounced at low strain rates.The hardness results appeared to agree with this:material strained at the slower rate was harder inboth cases. The hardness values after testing atboth strain rates were higher than the unstrainedvalues, with the greater increase for the lowerstrain rate, which was associated with moremartensite in the microstructure.

The difference in marten site could beexplained by the austenite transformation. Oncesome martensite has formed, it places theaustenite in compression (since martensite isless dense), and hinders further transformation.Thus, the austenite is stabilized in two ways: bythe higher carbon content and by constraint, andso needs more time to be transformed. The moreslowly strained material contained more marten-site because there was sufficient time for it to betransformed, whereas the faster strain rate didnot allow significant transformation. Mayer etal. 10 recorded that martensite transformationdepends on the stability of the austenite, andreported that a minimum of 20 per cent retainedaustenite is needed. This value agrees with thosefound in the present investigation. The apparentanomalous relationship of decreasing hardnessand increasing yield stress with increased strainrates is explained by the constant (and muchfaster than in the tensile tests) loading rate inthe hardness tests. There would be much less (if

The Joumal of The South African Institute of Mining and Metallurgy

any) transformation to marten site during hard-ness tests than in tensile tests. Thus, hardnesstests were undertaken on a more stable micro-structure, while the tensile tests were carried outon a more dynamic microstructure.

The strengthening mechanism causing thehigher yield stresses in the material strained atthe higher rate was not directly related to eithermarten site or ferrite since the phase proportionswere nearly constant in the microstructures.Ferrite also contains very limited amounts ofcarbon, and there is no known enrichment effect,which makes ferrite strengthening unlikely. Thecarbides might have had a small effect on thegrade IV material, but no carbides were presentin the grade II material, which also exhibited asimilar relationship between yield stress andstrain rate.

Alternatively, the higher yield stress couldhave been due to some factor in the austenite,and a speculative explanation is offered. It isprobable that the inherent strength of theaustenite had increased, as a result of solutionstrengthening because of the higher carboncontent. Lower austempering temperature allowsless diffusion of carbon, and the austenite of thegrade IV material might be expected to have ahigher carbon content than that of the grade IImaterial. Solid-solution strengthening improvesas the proportion of the second element increases,usually to a limit. Thus, a higher carbon contentwould provide strengthening and, even withoutany workhardening, the carbon-enriched austenitewould be more difficult to deform, and hence thehigher yield stress and hardness. More carbonwould also increase the workhardening rate byinteracting with dislocations, which would alsoincrease the UTS. Both these factors wouldpromote high strength, but with a tendency tobrittleness, which was seen in the fracturesurfaces of the grade IV samples. It must beemphasized that this mechanism may not haveoccurred in the grade IV material which had beenaustempered longer, since martensite would nothave been present initially, and the austenitewould probably have been transformed moreeasily. The evaluation of this explanationrequires more tests, especially on material thathas been austempered for longer times.

Summary

Am grades II and IVwere tested at differentstrain rates, and it was found that there was anincrease in yield stress with strain rate. Therewas an increase in hardness compared with theunstrained values, although the effect decreasedwith increasing strain rate. However, the differ-ences were still within the Am specifications,and probably do not have any significance to theapplication of Am in practice. More martensitewas observed in the specimens subjected to slowerstrain rates, which suggests that the martensitetransformation was time-dependent. A speculativeexplanation regarding strengthening in theaustenite was proposed to explain the higheryield stresses and hardnesses which were asso-ciated with lower martensite content. .

NOVEMBER/DECEMBER 1995 361 ...

----