Effect of d-ferrite on impact properties ofsupermartensitic stainless steel

heat affected zonesD Carrouge H K D H Bhadeshia and P Woollin

The work presented in this paper focuses on the effect ofthe presence of non-equilibrium d-ferrite on the impactproperties of a supermartensitic stainless steel Togenerate homogeneous d-ferrite containing microstruc-tures the material was in the first part of the presentstudy subjected to a series of high temperature furnaceheat treatments Three microstructures possessingvariable ferrite content and different grain sizes weregenerated Charpy impact results indicate that thepresence of 14 d-ferrite in a martensitic matrix of60 mm prior austenite grain size raises the ductile tobrittle transition temperature (DBTT) by about 50uCcompared with a fully martensitic microstructure of80 mm For a similar grain size reducing the amount ofd-ferrite from 14 to 2 restored the DBTT to a levelcomparable to that of the tempered parent material AGleeble simulator was used to create the range ofmicrostructures found in a weld heat affected zone(HAZ) Testing of the simulated HAZ indicated thattoughness was not significantly affected by the presenceof d-ferrite The DBTT was comparable to that of theparent material and lower than that of the heat treatedspecimens containing 14 d-ferrite STWJ410

Keywords supermartensitic stainless steels weld heataffected zone delta-ferrite toughness

At the time this work was carried out Dr Carrougeand Professor Bhadeshia were in the Department ofMaterials Science and Metallurgy University ofCambridge Pembroke Street Cambridge CB2 3QZUK Dr Carrouge is now with Corus plc Dr Woollinis with TWI Ltd Granta Park Great AbingtonCambridge UK Manuscript received 3 December2002 accepted 27 May 2003

2004 Institute of Materials Minerals and Mining

Published by Maney on behalf of the Institute

INTRODUCTIONSupermartensitic stainless steels have recently been intro-duced in the oil and gas industry to substitute moreexpensive duplex (ferritic-austenitic) stainless steels foronshore and offshore tubing applications The optimisedmicrostructure of the material is free from d-ferrite andoffers good corrosion resistance in environments contain-ing CO2 and some H2S In addition to reducing carbideprecipitation the low carbon content (ie about 001 wt-)of supermartensitic steels provides good weldability withconventional arc welding techniques

Recent studies of the phase transformation occurring inthe HAZ of the welded material revealed the presence ofsmall amounts of d-ferrite in the high temperature regions

of the HAZ12 While it is well established that coarsed-ferrite grains strongly reduce the Charpy notch tough-ness of low carbon semi-ferritic and martensitic stainlesssteels3ndash5 there is as yet no evidence to support thecontention that the d-ferrite found in supermartensiticstainless steel heat affected zones has such a pronouncedeffect Consequently the aim of the present study was toassess whether the presence of d-ferrite in supermartensiticsteels significantly affects the impact properties of the HAZ

Charpy impact testing has been chosen to evaluate theimpact properties of a series of microstructures The first partof this work concentrates on microstructures generated usingfurnace heat treatments The other part focuses on HAZ thathave been simulated using a Gleeble weld-simulator

Material studiedSteel A whose chemical composition is reported in Table 1is a highly alloyed supermartensitic stainless steel stabilisedby the addition of titanium The stabilisation works byforming stable carbonitrides Ti(C N) thereby decreasingthe carbon or nitrogen in solid solution and preventing theprecipitation of chromium or molybdenum carbonitrideswhich may lead to a reduction in the corrosion resistance6

The minimum titanium content required to prevent theseeffects is four times the carbon plus nitrogen concentration(wt-) owing to the stoichiometry of Ti(C N) Steels ofcomposition similar to that of steel A have been reported tohave a martensite start temperature Ms of about 200uC thealloy transforms entirely to martensite during cooling fromthe solution temperature7

The material was supplied as a seamless pipe of2731 mm outer dia and 120 mm wall thickness A studyrevealed this steel to be prone to HAZ d-ferrite retention189

Phase balance and transformation temperaturesEquilibrium calculations using MT-DATA software havebeen conducted on steel A to determine the phase balanceand phase transformation temperatures MT-DATA usesexperimentally determined parameters to model thermo-dynamic variables For this reason the reliability of thecalculations it performs depends on the parameters presentin the databases Simpler systems are more accuratelypredicted by MT-DATA hence only the elements in major

Table 1 Chemical composition of steel A wt-

C 001Si 026Mn 046Ni 646Cr 1220Mo 248Cu 003Ti 009N 00070

DOI 101179136217104225021823 Science and Technology of Welding and Joining 2004 Vol 9 No 5 377

quantities are taken into account in calculations The latterwere made without silicon since MT-DATA has been foundto give silicon an excessively strong ferrite-forming ten-dency in the systems of concern The most importantelements phases common carbides and nitrides have beenallowed and are listed in Table 2

Calculations have been performed over the temperaturerange 400ndash1600uC at intervals of 25uC using the plus andsub-sgte databases Figure 1 shows the evolution of themass fractions of the equilibrium phases as a function oftemperature Steel A is predicted to have the capacity tobecome fully austenitic and then undergo completeferritisation before melting As a result of the presence oftitanium the precipitation of carbides and nitrides isrestricted to TiC and TiN particles Titanium carbides arepredicted to dissolve at temperatures close to 900uC

Table 3 reports the predicted equilibrium phase trans-formation temperatures Under equilibrium conditions theAe1 temperature is lower than 400uC owing to the relativelyhigh nickel content of the steel Ferrite starts forming at1237uC and the ferritisation interval is 159uC The latterinterval is relatively large mainly as a result of the strongferrite stabilising effect of molybdenum

IMPACT PROPERTIES OF HEAT TREATEDSPECIMENSThe impact properties of the parent material in its as-received condition were first studied Then to investigatethe effect of d-ferrite on the ductile to brittle temperature(DBTT) fully martensitic and ferritic-martensitic micro-structures were required The results of a preliminary studyrevealed that furnace heat treatments at 1100uC and 1350uCfor 15 min were satisfactory to produce the desired

microstructures9 However to facilitate the interpretationof Charpy transition curves a difference in prior austenitegrain size between specimens heat treated at differenttemperatures is not advisable To study the impact proper-ties at a given grain size the ferritic-martensitic micro-structure was re-austenitised to dissolve ferrite

Procedure and experimental techniquesForty specimens of 12 mm square section and 60 mm longwere machined from steel A in the longitudinal direction asshown in Fig 2 Ten specimens were kept in the as-receivedstate (quenched from 950uC and tempered at 640uC) Thirtyother samples were individually placed in a preheatedfurnace at the centre of the bottom surface and were left infor 15 min followed by water quenching Using thisprocedure 10 specimens were heat treated at 1100uC and20 others at 1350uC After cooling 10 of the latterspecimens were re-austenitised at 1100uC for an hour andthen water quenched Four different microstructures weretherefore produced

To reveal the microstructures specimens were ground onSiC paper to grit 1200 and polished with a 1 mm clothcoated with diamond paste The samples were etched usingone of the two etchants listed in Table 4 Digitalphotographs were taken using a Leica DMR digitalcamera

The volume fractions of d-ferrite in the specimens havebeen determined using SigmaScan Pro 50 software usingthe following procedure A series of digital micrographswere taken from representative areas of the specimens Theimages were first enhanced to use the full intensity spectrumand the total number of pixels in the image was determinedTo highlight d-ferrite the intensity threshold method wasemployed and the sum of pixels constituting the selectedfeatures was performed The ratio of the latter sum with theoriginal number of pixels led to the determination of thevolume fraction

To investigate the presence of any retained austenite (cr)in the different specimens X-ray diffraction tests were

Table 2 List of elements and phases used for equilibriumcalculation performed on steel A

Alloy Allowed elements Allowed phases

AFe C Mn Ni CrMo Ti N

M23C6 c d Fe3CM7C3 TiC TiN Liq

1 Result of MT-DATA equilibrium calculation performedon steel A

Table 3 Equilibrium phase transformation temperatures(uC) obtained after calculations performed onsteel A

Alloy Ae1 (5) Ae3 (95) Ae4 (5) Ae5 (95) Solidus (5)

A 400 705 1237 1396 1461

Ae1 Ae4 and solidus temperatures have been obtained by takingtemperature at which 5 of growing phase has formed SimilarlyAe3 and Ae5 temperatures have been obtained by takingtemperature at which 95 of growing phase has formed

2 Orientation of Charpy specimen within pipe

Table 4 Compositions and conditions of use of etchants

Name Composition Etching procedure Specificity

Sulphuric10 20 mL H2SO4 001 g NH4CNS80 mL H2O

Immersion of sample andelectrolytic etching at 4 V for 20 s

Reveals d-ferrite but not martensite

Stock11 1 g K2S2O5 20 mL HCl80 mL H2O

Immersion of sample untilspecimens becomes red

Reveals d-ferrite in blue andmartensite yellowbrown

378 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

carried out The latter were performed on a Phillipsdiffractometer using Cu Ka radiation operating at 40 kVand 40 mA Step scan mode was employed to cover theangular 2h range from 471 to 103u The 2h step size was005u with a dwell time of 30 s Samples were ground andpolished using the normal metallographic preparationprocedure with two cycles of polishing and etching beforethe X-ray diffraction test to remove any deformed layercaused by polishing The diffraction peaks were indexed bycomparing theoretical to experimental interplanar spacingvalues for ferrite and austenite planes

The volume fraction of austenite was estimated frommeasurements of the integrated intensities of martensiteand austenite peaks assuming they are the only phasespresent

As the volume fractions are measured from theintensities which are corrected for the structure and othersystematic factors the patterns obtained did not requirecalibration12 The ratio of the intensities of diffractionpeaks from two phases of a polycrystalline sample is givenby equation (1)12

Ic(hkl)

Ia0(hkl)~

Rc(hkl)

Ra0(hkl)|

Vc

Va0 (1)

where Ic(hkl) and Ia9(hkl) are the integrated intensity from agiven plane (hkl) from the austenite and martensite (orferrite) phase respectively Vc and Va9 are the volumefractions of austenite and martensite respectively andRc (hkl) and Ra9(hkl) are given for a specific peak by

R~1

v2frac12jF j2(p)(Lp)e2m (2)

where v5volume of unit cell F5structure factorp5multiplicity factor Lp5Lorentz-polarisation factorand e22m5temperature factor

Considering that all materials in reality exhibit crystal-lographic texture the average integrated intensity for atleast three specific reflections for austenite and martensite(Table 5) were taken into account (equations (3) and (4))

Ic

Rc~

1

3

X3

i~1

Ici

Rci

(3)

Ia0

Ra0~

1

3

X3

i~1

Ia0

i

Ra0i

(4)

The value of VcVa9 in equation (1) can be obtained fromthe measurement of IcIa9 and the calculation of RcRa9Once VcVa9 is found the value of Vc can be obtained fromthe additional relationship

VczVa0~1 (5)

The 2h values for three austenite peaks were used tocalculate the d spacings with Braggrsquos law and then thelattice parameter a These values were plotted against sin2hAn accurate value of ac was found by extrapolation tosin2h5112

To study the location of any retained austenite aJEOL 200 CX transmission electron microscope (TEM)

operating at 200 kV has been used Thin foils of 3 mm diaand approximately 50 mm thick were obtained from thespecimens employed for X-ray analysis The discs wereelectropolished in a solution of 5 perchloric acid 15glycerol and 80 ethanol at an applied voltage of 50 V atroom temperature Conventional selected-area diffractionwas used to distinguish austenite (fcc structure) from lowcarbon martensite (bcc)

After optical verification of the desired microstructuresthe heat treated specimens were machined to the standarddimension of 10 mm square section 6 55 mm with a 2 mmdeep V-notch positioned through-thickness at the middleof the specimens Charpy impact tests were finally carriedout in accordance with British Standard 10045-113 between20 and 2200uC to reveal any ductilendashbrittle transition

The fracture surfaces of Charpy specimens were observedusing a JEOL 5800 scanning electron microscope (SEM)Secondary electron mode was employed at an acceleratingvoltage of 15 kV To identify the nature of particles micro-analytical experiments were carried out using an energydispersive X-ray (EDX) analysis system ZAF (atomicnumber absorption fluorescence) corrections wereemployed the acquisition time was set to 100 s and thedead time was kept below 20

As-received microstructureThe initial microstructure of the material electrolyticallyetched is shown in Fig 3 It consisted of temperedmartensite with no evidence of d-ferrite and some apparenttitanium carbonitride particles

X-ray diffraction tests revealed that steel A contained asignificant amount of retained austenite (Fig 4) Thevolume percentage of the latter was 13 within anexperimental error of iexcl05 The error calculation wasbased on the integrated intensity since this error ischaracteristic of the specimen studied

This austenite remained stable after immersion in liquidnitrogen for 30 min but after re-austenitisation at 1000uCfor 30 min and water quenching no austenite could bedetected (Fig 5) Those results indicated that the presenceof retained austenite in the as-received condition was theconsequence of the manufacturerrsquos tempering treatment Infact tempering must have been carried out a few degreesabove the steel Ac1 temperature In this case the newaustenite is significantly enriched in carbon and nickel andremains stable after cooling7

A TEM examination of a thin foil prepared on steel Ashowed this retained austenite to be present as smallelongated grains located at martensite lath boundaries asillustrated in Fig 6

3 Optical micrograph of initial microstructure of steel Aelectrolytically etched

Table 5 Diffracting hkl planes used for martensite andaustenite

Phase i Diffracting plane (hkl)

Martensite 1 200Martensite 2 211Martensite 3 220Austenite 1 200Austenite 2 220Austenite 3 311

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 379

Science and Technology of Welding and Joining 2004 Vol 9 No 5

Characterisation of heat treated specimensTo facilitate the interpretation of the results heat treatedspecimens were characterised before impact testing in termsof phase presence volume fraction hardness and grain size

Figure 7 shows the microstructures of the fully martensi-tic martensiticferritic and annealed martensiticferriticspecimens observed on the actual-size Charpy specimensTypical cuboid Ti(CN) precipitates were the only observednon-metallic particles These particles were up to 10 mmlong and occurred in random dispersion or as small distinctclusters

In the first martensiticferritic microstructure d-ferritewas located both at grain boundaries and within thegrains Its volume percentage measured by image analysisat a magnification of 200 was 14iexcl2 Complete re-austenitisation of the latter microstructure gave a reductionof the d-ferrite content to 2iexcl05 and spheroidisation ofthe remaining ferrite The dissolution of d-ferrite appearedto be a slow process since 2 ferrite were still present afteran hour at 1100uC This is because significant partitioningof substitutional solutes occurred at 1350uC as indicated bythe EDX measurements reported in Table 6

There were no apparent differences in the microstructurein terms of Ti(CN) particles This was expected since theferritising and the re-austenitising operations were per-formed above the predicted temperature of TiC dissolution(about 900uC) and water quenching prevents precipitationduring cooling

The hardness and grain sizes measured on all materialsare reported in Table 7 Despite the presence of 14

5 X-ray diffraction pattern of steel A after indicatedthermal treatment (WQ water quench)

6 a Bright field image of martensite plus retained auste-nite region b dark field image of same region as a cselected area diffraction pattern of austenite corre-sponding to dark field image

Table 6 Summary of 15 EDX microanalyses (wt-) per-formed on ferrite and martensite

Composition

Phase Cr Ni Mo

d-ferrite 135iexcl09 48iexcl10 26iexcl04Matrix 117iexcl03 70iexcl04 17iexcl02

4 X-ray diffraction pattern of steel A in as-received condition

Table 7 Macrohardness (HV10) and grain size of differ-ent microstructures

Microstructure Hardness Grain size mm

Tempered martensitezcr 296iexcl3 25iexcl4Fresh martensite 317iexcl8 60iexcl9Fresh martensitez14 d-ferrite 311iexcl8 80iexcl12Fresh martensitez2 d-ferrite 304iexcl7 80iexcl12

380 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

d-ferrite there was little hardness change between thedifferent microstructures The latter was attributed to thevirtually carbon free martensite as a result of the presence oftitanium in concentrations greater than the stoichiometricamount As tensile strength is proportional to hardness allspecimens could be assumed to have a similar strength14

There was a small difference in grain size between the variousmicrostructures In fact the prior austenite grain size of thespecimens heat treated at 1350uC was about 20 mm largerthan the one heat treated at 1100uC As a result of the grainboundary pinning effect of d-ferrite there was no grain sizedifference between the microstructures consisting of freshmartensite and various amounts of ferrite

The presence of retained austenite in the furnace heattreated specimens has been checked using X-ray diffractionNo austenite could be detected in any of the specimens(Fig 8)

Results of Charpy impact testsTo facilitate the comparison and the determination of theductile to brittle transition temperature (DBTT) thetoughness results have been fitted by a non-linear regressiontechnique to the equation15

Impact energy (J)~A0zB0 tanhTT0

C0

(6)

where A0 B0 C0 and T0 are fitting constants and T isthe temperature in uC The DBTT was taken as thetemperature corresponding to half the height of thetransition curve Table 8 summarises the measured ductile

7 Microstructures obtained after isothermal treatment ata 1100uC and b 1350uC for 15 min and water quenchedc Microstructure obtained after treatment at 1350uC for15 min and water quenched followed by re-austenitisationat 1100uC for 60 min and water quenched

8 X-ray diffraction patterns of fresh martensite freshmartensitez14 d-ferrite and fresh martensitez2d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 381

Science and Technology of Welding and Joining 2004 Vol 9 No 5

to brittle transition temperatures for each tested micro-structure

Figure 9 shows the impact transition curves obtainedfrom the tempered martensitezcr (as-received condition)fresh martensite and fresh martensitez14 d-ferritemicrostructures The transition curves of the as-receivedand fresh martensite microstructures were similar and theirDBTT were identical and far below 0uC However the as-received microstructure showed slightly higher failureenergies at low temperatures This is because the retainedaustenite tends to transform to martensite during impactand thus absorbs some energy and raises the toughness asproposed by Bilmes et al16 At room temperature there wasmore scatter in the energy required to break the specimensbut there was apparently no significant difference betweenthe two microstructures

Similarly there were no apparent differences between theupper shelf failure energies of the fresh martensite andfresh martensitez14 d-ferrite microstructures (Fig 9)However the transition occurred over a shorter tempera-ture range in the biphased microstructure and a significantshift in ductile to brittle transition temperature wasobserved In fact the DBTT of the duplex microstructurewas raised by 52uC to about 246uC The origin of thisincrease could not be attributed directly to the presence ofd-ferrite since there was a 20 mm difference in prioraustenite grain size between the two microstructures

The observation of the transition curves of the freshmartensitez14 d-ferrite and fresh martensitez2 d-ferritemicrostructures of identical grain size removed this uncer-tainty (Fig 10) In fact specimens containing a reducedamount of d-ferrite had significantly better DBTT (286uCcompared to 246uC) As mentioned in the previous sectionafter re-austenitisation and quenching the only differencebetween the two ferrite containing microstructures was theactual level of ferrite Consequently it was clear that for aprior austenite grain size of 80 mm the presence of d-ferritesignificantly affected the transition temperature

Figure 11 compares the transition curves of the freshmartensitez2 d-ferrite (grain size 80 mm) with the temperedmartensitezcr microstructure (grain size 25 mm) The reduc-tion of the ferrite content clearly improves the notch toughnessto a level comparable to that of the as-received condition

FractographyObservation of the fracture surfaces in the SEM revealed avariety of fracture modes In particular ductile rupturecleavage and quasi-cleavage could be identified McDonaldet al summarise these fracture modes as follows1718

(i) Ductile rupture Dimples resulting from microvoidformation and coalescence are characteristic ofductile rupture Microvoids are generally initiatedat inclusions which can sometimes be observedwithin the dimples

(ii) Cleavage Stepped facets are caused by fracturethrough regions of different crystallographic orien-tation Cleavage occurs by the rapid propagation ofa crack along crystallographic planes and may beinitiated by inclusionmatrix decohesion This is avery low energy-absorbing mode of fracture

(iii) Quasi-cleavage A less well-defined fracture modeknown as quasi-cleavage is associated with tear ridgeswhich represent limited plastic deformation separatingcleavage facets Unlike pure cleavage energy isabsorbed as a result of the limited plastic deformation

Table 8 Ductile to brittle transition temperatures obtainedfor each microstructure studied

Microstructure DBTT (uC)

Tempered martensitezcr 298Fresh martensite 298Fresh martensitez14 d-ferrite 246Fresh martensitez2 d-ferrite 286

9 Charpy impact energy curves obtained on temperedmartensitezretained austenite (as-received condition)fresh martensite and fresh martensitez14 d-ferritemicrostructures

10 Charpy impact energy curves obtained on fresh mar-tensitez14 d-ferrite and fresh martensitez2 d-ferrite microstructures Note that both microstructureshave identical prior austenite grain size of 80 mm

11 Comparison of Charpy impact energy curves betweenas-received and fresh martensitez2 d-ferritemicrostructures

382 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

Examples of ductile rupture obtained from the fullymartensitic and martensiticferritic microstructures testedat 0uC are shown in Fig 12 Most cuboidal shaped titaniumcarbonitride particles were cracked

Specimens tested at low temperatures fractured in atransgranular manner (Fig 13) This indicated that thepresence of grain boundary d-ferrite did not embrittle thelatter In general the cleavage facet sizes appeared similarbut the fresh martensitez14 d-ferrite microstructureshowed more secondary cracks

River markings on the cleavage facets of the specimensshowed evidence of both cleavage (Fig 14) and quasi-cleavage fracture (Fig 15) initiated by Ti(CN) inclusionsThe capacity of these particles to reduce toughness is wellknown Brittle fracture induced by the cleavage of TiNinclusions has recently been investigated by Fairchildet al1920 Those particles are well bonded to the matrixand thus allow high stresses in a notch tip plastic zone toact on inclusions without debonding the interface Once aninclusion cleaves the strong bond allows for transfer of thecrack into the martensitic matrix

In addition to the SEM observation cross-sections per-pendicular to the fracture surfaces as shown in Fig 16 havebeen observed to reveal the fracture path Figures 17 and 18show example of such cross-sections for specimens of thefresh martensitez14 d-ferrite and fresh martensitez2d-ferrite microstructures which failed by cleavage

In the fresh martensitez14 d-ferrite specimens sec-ondary intragranular cracks propagated through marten-site and d-ferrite stringers whereas shorter cracks were seento initiate both in martensite and at the d-ferritemartensiteinterface in the fresh martensitez2 d-ferrite In the lattercase cracks could not easily link up when d-ferrite wasmore dispersed and possessed a spheroidised morphologythus explaining the better impact properties

IMPACT PROPERTIES OF GLEEBLESIMULATED SPECIMENSThe first part of this work established the effect of d-ferriteon the ductile to brittle transition temperature of homo-genous and biphased microstructures The microstructure

12 Dimpled fracture surfaces of a fresh martensite and b fresh martensitez14 d-ferrite microstructures tested at 0uC

13 Fracture surfaces of a fresh martensite and b fresh martensitez14 d-ferrite microstructures tested at 2150 and2100uC respectively

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 383

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of a weld HAZ is however far from being homogeneous andit is not known whether the presence of delta-ferrites affectthe HAZ impact properties

In supermartensitic stainless steels five HAZ regions canbe distinguished as illustrated in Fig 192

(i) partially melted region located beside the fusionboundary This is where incipient melting of thenewly formed d-ferrite occurs on heating

(ii) region which transforms completely to d-ferrite onheating with significant grain growth also calledthe coarse-grained HAZ (CG-HAZ)

(iii) high-temperature two-phase region characterisedby partial transformation of austenite to d-ferriteon heating

(iv) region which completely transforms to austeniteduring heating

(v) a low peak temperature two-phase region char-acterised by partial transformation of the tem-pered martensite to austenite during heating

The HAZ represents the area where all phase transforma-tions occur in the solid state and therefore the partiallymelted region should not be regarded as belonging to theHAZ The other regions can be artificially classified intotwo categories depending on the maximum temperaturethey experienced before cooling the high temperature andthe low temperature HAZ The high temperature HAZ

14 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedcleavage

15 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedquasi-cleavage

16 Location of transverse section in Charpy specimens

384 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

(HT-HAZ) includes the prior d-ferrite region and the hightemperature two-phase region whereas the low temperatureHAZ (LT-HAZ) refers to the remainder of the HAZ Anexample of a typical HAZ found in a single pass weldperformed on steel A is shown in Fig 20

As presented in another publication the high temperatureregion of a supermartensitic stainless steel HAZ is likely tocontain non-equilibrium d-ferrite1 High heat inputs singlepass bead on plates such as the one showed in Fig 20 yieldslow cooling rates in the CG-HAZ and favour the formationof Widmanstatten austenite during cooling from the fullyferritic region It is between the Widmanstatten austenitesubunits that ferrite is found in coarse-grained HAZ1 Indual-phase HAZ on the other hand ferrite is usually locatedat grain boundaries and as thin films within the grains Thevolume fraction of d-ferrite is dependent upon the actualwelding conditions but is usually smaller in coarse-grainedHAZ This is because fast cooling rates restrictWidmanstatten austenite formation in coarse-grained HAZbut prevent ferrite dissolution in the dual-phase HAZ9

Gleeble weld-simulatorBeing close to the weld metal sampling of all the HAZregions for Charpy impact testing is difficult To avoid this

problem a Gleeble weld-simulator has been employed TheGleeble 1500 is a thermomechanical testing device that isused in a wide range of applications including simulationand testing of weld heat affected zones A major advantageof the Gleeble is that it generates large volumes ofmicrostructure that simulate small hard to study regionsin actual weldments A range of specimen geometries can beused but by using bars Charpy specimens can directly bemachined after simulation

The machine (Fig 21) is interfaced to a computer that isreadily programmed to provide reference signals for closed-loop control of the applied thermal cycle Heating isaccomplished by the flow of low-frequency alternatingcurrent in the specimen The current distribution iscontrolled by the cross-section of the specimen and thejaws used to mount it The feedback signal necessary forclosed-loop control is obtained from a thermocouplewelded to the specimen surface21

ProcedureTen specimens of dimension 116116100 mm (takenaccording to Fig 2) were machined for HAZ simulationThe operation was performed at the maximum temperature

17 Cross-section of fracture surface of fresh martensitez14 d-ferrite microstructure tested at 2100uC Electrolyticallyetched to reveal d-ferrite

18 Cross-section of fracture surface of fresh martensitez2 d-ferrite microstructure tested at 2150uC Electrolyticallyetched to reveal d-ferrite

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 385

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of 1350uC using a cooling time between 800 and 500uC of10 s After simulation all specimens were polished andetched to reveal the underlying microstructure Standardsize Charpy specimens were finally machined with the notchlying across the centre of the simulated specimens

MicrostructureFigure 22 shows a representative macrograph of thesimulated specimens before grinding and polishing

After polishing and electro-etching the central region ofall simulated specimens showed a microstructural gradientThe latter indicated an inhomogeneous thermal experienceacross the sample sections and might have been caused byan improper contact between the samples and the grips in

the Gleeble machine As Fig 23a shows the central regionof the simulated specimens consisted of three distinct areasThe upper region had semi-circular shape and depth ofabout 25 mm and was identified as the coarse-grained region (Fig 23b) The latter was surrounded by

19 Schematic diagram illustrating structural changes that occur within heat affected zone of single-pass supermartensiticstainless steel weld2

20 Optical micrograph of single-pass bead on plate per-formed on steel A using tungsten inert gas (TIG) pro-cess with heat input of 1 kJ mm21 Plate thickness10 mm Etched electrolytically to reveal d-ferrite

21 Arrangement of equipment within Gleeble simulator

22 Macrograph of typical specimen after HAZ simula-tion using Gleeble simulator Thermocouple employedto record temperature was located at middle of centreline

386 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

the dual-phase region (Fig 23c) and in turn wassurrounded by the prior fully-austenitic area The existenceof this microstructural gradient was not expected but it wasdecided to study its impact properties since the generatedcentre line microstructure was representative of that of areal weld HAZ

The coarse-grained and dual-phase regions had grain sizeof about 160 mm and 30 mm respectively The d-ferritevolume percentage in the dual-phase region measured byimage analysis at a magnification of 200 (Fig 23c) was40iexcl5 This was larger than the volume percentagemeasured in heat treated specimens (Fig 7b and c)

Results of Charpy impact testThe transition curve obtained on the Gleeble simulatedspecimens is presented in Fig 24 For comparison the

results obtained on the as-received and martensiticferriticmicrostructures have been reported The Gleeble specimenshad the lowest upper shelf energy of the three curves Thelatter was probably because of the combined action of largegrain size and presence of d-ferrite in the tested micro-structure More importantly the simulated HAZ had aductile to brittle transition temperature comparable to thatof the as-received material (294 and 298uC respectively)although like the martensitic microstructure containing14 d-ferrite the transition occurred over a shortertemperature range The behaviour of the Gleeble simulatedspecimens at low temperatures was therefore significantlybetter than the ferriticmartensitic microstructure studied inthe first part

FractographyAll specimens tested at temperatures above 2100uC showedductile failure mode whereas those broken at temperaturesbelow 2100uC showed completely brittle behaviourInterestingly the Gleeble specimen tested at 2100uCshowed two fracture modes (Fig 25) Observation of across-section revealed the coarse-grained region to fail bycleavage whereas the dual-phase region showed evidence ofductility below the transition temperature (Fig 26) Theductility of the dual-phase HAZ indicates that the DBTT ofthis region is lower than 2100uC These results provideevidence to support the view that the presence of d-ferrite indual-phase HAZ up to 40 does not significantly affect theimpact properties of the whole high temperature HAZ Thespecimen transition temperature was more strongly influ-enced by the large grain size of the coarse-grained HAZ

SUMMARY AND CONCLUSIONThis work revealed that for a similar strength particlevolume-fraction and distribution and a given prioraustenite grain size the ductile to brittle transitiontemperature was very sensitive to the volume fraction ofd-ferrite in homogeneous microstructures On the otherhand tests carried out on simulated HAZ indicated that thedual-phase region did not play a major role in determiningthe DBTT even when d-ferrite was present up to 40 Thesmall grain size of the dual-phase region offers suitableobstacles to crack propagation Conversely the large grainsize of the coarse-grained region strongly influences theimpact properties of the high temperature HAZ Howeverwhen compared with the tempered parent material contain-ing austenite the transition temperature of the Gleeblesimulated specimens was raised by only a few degreesConsidering that the microstructure observed in simulated

23 Microstructure of Gleeble simulated specimens a lowmagnification of central region b close-up of coarse-grained region and c close-up of dual-phase regionAll specimens were electrolytically etched to reveald-ferrite

24 Charpy impact energy curve of Gleeble simulated spe-cimens and comparison with those of as-received andfresh martensitez14 d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 387

Science and Technology of Welding and Joining 2004 Vol 9 No 5

HAZ is always coarser than that found in a multipass girthweld and that the DBTT of the simulated specimens waswell below room temperature the presence of d-ferrite inHAZ of supermartensitic stainless steels is not likely toconstitute a threat to the impact properties

ACKNOWLEDGEMENTSThe authors are grateful to TWI and the EPSRC forsponsoring this research

REFERENCES1 D CARROUGE and P WOOLLIN in lsquoStainless steel world 2002rsquo

61ndash67 2002 Houston USA2 J ENERHAUG Oslash GRONG and U M STEINSMO Sci Technol Weld

Joining 2001 6 330ndash3383 T G GOOCH and B J GINN lsquoHeat affected zone toughness of

MMA welded 12Cr martensitic-ferritic steelrsquo TechnicalReport 373 Welding Institute Members Report 1988

4 C GROBLER lsquoWeldability studies on 12 and 14 chromiumsteelsrsquo PhD thesis University of Pretoria 1987

5 J J J ZAAYMAN and G T VAN ROOYEN Proc 1st Int ChromiumSteels and Alloys Congress Volume 2 SAIMM Cape TownJohannesburg 1992 137ndash142

6 K KONDO M UEDA K OGAWA H AMAYA H HIRATA andH TAKABE in lsquoSupermartensitic stainless steels 99rsquo 11ndash181999 Belgium

7 T G GOOCH P WOOLLIN and A G HAYNES in lsquoSupermartensiticstainless steels 99rsquo 188ndash195 1999 Belgium

8 P WOOLLIN and D CARROUGE in lsquoSupermartensitic stainlesssteels 2002rsquo 199ndash204 2002 Paper 027

9 D CARROUGE PhD thesis Department of MaterialsScience and Metallurgy University of Cambridge 2002Available from httpwwwmsmcamacukphasetrans2000phdhtmldominique

10 P WOOLLIN lsquoMetallurgical examination of martensitic stainlesssteel weldmentsrsquo Technical Report 620886195 WeldingInstitute Internal Report 1995

11 E BERAHA and B SHPIGLER lsquoColor metallographyrsquo 1977American Society for Metals

12 B D CULLITY lsquoElements of X-ray diffractionrsquo 1978Addison-Wesley

13 BS EN 10045-1 lsquoCharpy impact test on metallic materialsPart 1 Test methodrsquo British Standard 1990

14 D WILLIAM and J R CALLISTER lsquoMaterials science andengineering an introductionrsquo 1999 New York USA JohnWiley and Sons Inc

15 W OLDFIELD CSNI Specialists Meeting on lsquoInstrumentedprecracked Charpy testingrsquo Palo Alto CA 1981 5ndash59

25 Fracture surfaces of a coarse-grained and b dual-phase region of Gleeble simulated specimen tested at 2100uC

26 Cross-section of fracture surface shown in Fig 25 a coarse-grained and b dual-phase region This cross-section wastaken perpendicular to examination plane shown in Fig 16

388 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

quantities are taken into account in calculations The latterwere made without silicon since MT-DATA has been foundto give silicon an excessively strong ferrite-forming ten-dency in the systems of concern The most importantelements phases common carbides and nitrides have beenallowed and are listed in Table 2

Calculations have been performed over the temperaturerange 400ndash1600uC at intervals of 25uC using the plus andsub-sgte databases Figure 1 shows the evolution of themass fractions of the equilibrium phases as a function oftemperature Steel A is predicted to have the capacity tobecome fully austenitic and then undergo completeferritisation before melting As a result of the presence oftitanium the precipitation of carbides and nitrides isrestricted to TiC and TiN particles Titanium carbides arepredicted to dissolve at temperatures close to 900uC

Table 3 reports the predicted equilibrium phase trans-formation temperatures Under equilibrium conditions theAe1 temperature is lower than 400uC owing to the relativelyhigh nickel content of the steel Ferrite starts forming at1237uC and the ferritisation interval is 159uC The latterinterval is relatively large mainly as a result of the strongferrite stabilising effect of molybdenum

IMPACT PROPERTIES OF HEAT TREATEDSPECIMENSThe impact properties of the parent material in its as-received condition were first studied Then to investigatethe effect of d-ferrite on the ductile to brittle temperature(DBTT) fully martensitic and ferritic-martensitic micro-structures were required The results of a preliminary studyrevealed that furnace heat treatments at 1100uC and 1350uCfor 15 min were satisfactory to produce the desired

microstructures9 However to facilitate the interpretationof Charpy transition curves a difference in prior austenitegrain size between specimens heat treated at differenttemperatures is not advisable To study the impact proper-ties at a given grain size the ferritic-martensitic micro-structure was re-austenitised to dissolve ferrite

Procedure and experimental techniquesForty specimens of 12 mm square section and 60 mm longwere machined from steel A in the longitudinal direction asshown in Fig 2 Ten specimens were kept in the as-receivedstate (quenched from 950uC and tempered at 640uC) Thirtyother samples were individually placed in a preheatedfurnace at the centre of the bottom surface and were left infor 15 min followed by water quenching Using thisprocedure 10 specimens were heat treated at 1100uC and20 others at 1350uC After cooling 10 of the latterspecimens were re-austenitised at 1100uC for an hour andthen water quenched Four different microstructures weretherefore produced

To reveal the microstructures specimens were ground onSiC paper to grit 1200 and polished with a 1 mm clothcoated with diamond paste The samples were etched usingone of the two etchants listed in Table 4 Digitalphotographs were taken using a Leica DMR digitalcamera

The volume fractions of d-ferrite in the specimens havebeen determined using SigmaScan Pro 50 software usingthe following procedure A series of digital micrographswere taken from representative areas of the specimens Theimages were first enhanced to use the full intensity spectrumand the total number of pixels in the image was determinedTo highlight d-ferrite the intensity threshold method wasemployed and the sum of pixels constituting the selectedfeatures was performed The ratio of the latter sum with theoriginal number of pixels led to the determination of thevolume fraction

To investigate the presence of any retained austenite (cr)in the different specimens X-ray diffraction tests were

Table 2 List of elements and phases used for equilibriumcalculation performed on steel A

Alloy Allowed elements Allowed phases

AFe C Mn Ni CrMo Ti N

M23C6 c d Fe3CM7C3 TiC TiN Liq

1 Result of MT-DATA equilibrium calculation performedon steel A

Table 3 Equilibrium phase transformation temperatures(uC) obtained after calculations performed onsteel A

Alloy Ae1 (5) Ae3 (95) Ae4 (5) Ae5 (95) Solidus (5)

A 400 705 1237 1396 1461

Ae1 Ae4 and solidus temperatures have been obtained by takingtemperature at which 5 of growing phase has formed SimilarlyAe3 and Ae5 temperatures have been obtained by takingtemperature at which 95 of growing phase has formed

2 Orientation of Charpy specimen within pipe

Table 4 Compositions and conditions of use of etchants

Name Composition Etching procedure Specificity

Sulphuric10 20 mL H2SO4 001 g NH4CNS80 mL H2O

Immersion of sample andelectrolytic etching at 4 V for 20 s

Reveals d-ferrite but not martensite

Stock11 1 g K2S2O5 20 mL HCl80 mL H2O

Immersion of sample untilspecimens becomes red

Reveals d-ferrite in blue andmartensite yellowbrown

378 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

carried out The latter were performed on a Phillipsdiffractometer using Cu Ka radiation operating at 40 kVand 40 mA Step scan mode was employed to cover theangular 2h range from 471 to 103u The 2h step size was005u with a dwell time of 30 s Samples were ground andpolished using the normal metallographic preparationprocedure with two cycles of polishing and etching beforethe X-ray diffraction test to remove any deformed layercaused by polishing The diffraction peaks were indexed bycomparing theoretical to experimental interplanar spacingvalues for ferrite and austenite planes

The volume fraction of austenite was estimated frommeasurements of the integrated intensities of martensiteand austenite peaks assuming they are the only phasespresent

As the volume fractions are measured from theintensities which are corrected for the structure and othersystematic factors the patterns obtained did not requirecalibration12 The ratio of the intensities of diffractionpeaks from two phases of a polycrystalline sample is givenby equation (1)12

Ic(hkl)

Ia0(hkl)~

Rc(hkl)

Ra0(hkl)|

Vc

Va0 (1)

where Ic(hkl) and Ia9(hkl) are the integrated intensity from agiven plane (hkl) from the austenite and martensite (orferrite) phase respectively Vc and Va9 are the volumefractions of austenite and martensite respectively andRc (hkl) and Ra9(hkl) are given for a specific peak by

R~1

v2frac12jF j2(p)(Lp)e2m (2)

where v5volume of unit cell F5structure factorp5multiplicity factor Lp5Lorentz-polarisation factorand e22m5temperature factor

Considering that all materials in reality exhibit crystal-lographic texture the average integrated intensity for atleast three specific reflections for austenite and martensite(Table 5) were taken into account (equations (3) and (4))

Ic

Rc~

1

3

X3

i~1

Ici

Rci

(3)

Ia0

Ra0~

1

3

X3

i~1

Ia0

i

Ra0i

(4)

The value of VcVa9 in equation (1) can be obtained fromthe measurement of IcIa9 and the calculation of RcRa9Once VcVa9 is found the value of Vc can be obtained fromthe additional relationship

VczVa0~1 (5)

The 2h values for three austenite peaks were used tocalculate the d spacings with Braggrsquos law and then thelattice parameter a These values were plotted against sin2hAn accurate value of ac was found by extrapolation tosin2h5112

To study the location of any retained austenite aJEOL 200 CX transmission electron microscope (TEM)

operating at 200 kV has been used Thin foils of 3 mm diaand approximately 50 mm thick were obtained from thespecimens employed for X-ray analysis The discs wereelectropolished in a solution of 5 perchloric acid 15glycerol and 80 ethanol at an applied voltage of 50 V atroom temperature Conventional selected-area diffractionwas used to distinguish austenite (fcc structure) from lowcarbon martensite (bcc)

After optical verification of the desired microstructuresthe heat treated specimens were machined to the standarddimension of 10 mm square section 6 55 mm with a 2 mmdeep V-notch positioned through-thickness at the middleof the specimens Charpy impact tests were finally carriedout in accordance with British Standard 10045-113 between20 and 2200uC to reveal any ductilendashbrittle transition

The fracture surfaces of Charpy specimens were observedusing a JEOL 5800 scanning electron microscope (SEM)Secondary electron mode was employed at an acceleratingvoltage of 15 kV To identify the nature of particles micro-analytical experiments were carried out using an energydispersive X-ray (EDX) analysis system ZAF (atomicnumber absorption fluorescence) corrections wereemployed the acquisition time was set to 100 s and thedead time was kept below 20

As-received microstructureThe initial microstructure of the material electrolyticallyetched is shown in Fig 3 It consisted of temperedmartensite with no evidence of d-ferrite and some apparenttitanium carbonitride particles

X-ray diffraction tests revealed that steel A contained asignificant amount of retained austenite (Fig 4) Thevolume percentage of the latter was 13 within anexperimental error of iexcl05 The error calculation wasbased on the integrated intensity since this error ischaracteristic of the specimen studied

This austenite remained stable after immersion in liquidnitrogen for 30 min but after re-austenitisation at 1000uCfor 30 min and water quenching no austenite could bedetected (Fig 5) Those results indicated that the presenceof retained austenite in the as-received condition was theconsequence of the manufacturerrsquos tempering treatment Infact tempering must have been carried out a few degreesabove the steel Ac1 temperature In this case the newaustenite is significantly enriched in carbon and nickel andremains stable after cooling7

A TEM examination of a thin foil prepared on steel Ashowed this retained austenite to be present as smallelongated grains located at martensite lath boundaries asillustrated in Fig 6

3 Optical micrograph of initial microstructure of steel Aelectrolytically etched

Table 5 Diffracting hkl planes used for martensite andaustenite

Phase i Diffracting plane (hkl)

Martensite 1 200Martensite 2 211Martensite 3 220Austenite 1 200Austenite 2 220Austenite 3 311

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 379

Science and Technology of Welding and Joining 2004 Vol 9 No 5

Characterisation of heat treated specimensTo facilitate the interpretation of the results heat treatedspecimens were characterised before impact testing in termsof phase presence volume fraction hardness and grain size

Figure 7 shows the microstructures of the fully martensi-tic martensiticferritic and annealed martensiticferriticspecimens observed on the actual-size Charpy specimensTypical cuboid Ti(CN) precipitates were the only observednon-metallic particles These particles were up to 10 mmlong and occurred in random dispersion or as small distinctclusters

In the first martensiticferritic microstructure d-ferritewas located both at grain boundaries and within thegrains Its volume percentage measured by image analysisat a magnification of 200 was 14iexcl2 Complete re-austenitisation of the latter microstructure gave a reductionof the d-ferrite content to 2iexcl05 and spheroidisation ofthe remaining ferrite The dissolution of d-ferrite appearedto be a slow process since 2 ferrite were still present afteran hour at 1100uC This is because significant partitioningof substitutional solutes occurred at 1350uC as indicated bythe EDX measurements reported in Table 6

There were no apparent differences in the microstructurein terms of Ti(CN) particles This was expected since theferritising and the re-austenitising operations were per-formed above the predicted temperature of TiC dissolution(about 900uC) and water quenching prevents precipitationduring cooling

The hardness and grain sizes measured on all materialsare reported in Table 7 Despite the presence of 14

5 X-ray diffraction pattern of steel A after indicatedthermal treatment (WQ water quench)

6 a Bright field image of martensite plus retained auste-nite region b dark field image of same region as a cselected area diffraction pattern of austenite corre-sponding to dark field image

Table 6 Summary of 15 EDX microanalyses (wt-) per-formed on ferrite and martensite

Composition

Phase Cr Ni Mo

d-ferrite 135iexcl09 48iexcl10 26iexcl04Matrix 117iexcl03 70iexcl04 17iexcl02

4 X-ray diffraction pattern of steel A in as-received condition

Table 7 Macrohardness (HV10) and grain size of differ-ent microstructures

Microstructure Hardness Grain size mm

Tempered martensitezcr 296iexcl3 25iexcl4Fresh martensite 317iexcl8 60iexcl9Fresh martensitez14 d-ferrite 311iexcl8 80iexcl12Fresh martensitez2 d-ferrite 304iexcl7 80iexcl12

380 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

d-ferrite there was little hardness change between thedifferent microstructures The latter was attributed to thevirtually carbon free martensite as a result of the presence oftitanium in concentrations greater than the stoichiometricamount As tensile strength is proportional to hardness allspecimens could be assumed to have a similar strength14

There was a small difference in grain size between the variousmicrostructures In fact the prior austenite grain size of thespecimens heat treated at 1350uC was about 20 mm largerthan the one heat treated at 1100uC As a result of the grainboundary pinning effect of d-ferrite there was no grain sizedifference between the microstructures consisting of freshmartensite and various amounts of ferrite

The presence of retained austenite in the furnace heattreated specimens has been checked using X-ray diffractionNo austenite could be detected in any of the specimens(Fig 8)

Results of Charpy impact testsTo facilitate the comparison and the determination of theductile to brittle transition temperature (DBTT) thetoughness results have been fitted by a non-linear regressiontechnique to the equation15

Impact energy (J)~A0zB0 tanhTT0

C0

(6)

where A0 B0 C0 and T0 are fitting constants and T isthe temperature in uC The DBTT was taken as thetemperature corresponding to half the height of thetransition curve Table 8 summarises the measured ductile

7 Microstructures obtained after isothermal treatment ata 1100uC and b 1350uC for 15 min and water quenchedc Microstructure obtained after treatment at 1350uC for15 min and water quenched followed by re-austenitisationat 1100uC for 60 min and water quenched

8 X-ray diffraction patterns of fresh martensite freshmartensitez14 d-ferrite and fresh martensitez2d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 381

Science and Technology of Welding and Joining 2004 Vol 9 No 5

to brittle transition temperatures for each tested micro-structure

Figure 9 shows the impact transition curves obtainedfrom the tempered martensitezcr (as-received condition)fresh martensite and fresh martensitez14 d-ferritemicrostructures The transition curves of the as-receivedand fresh martensite microstructures were similar and theirDBTT were identical and far below 0uC However the as-received microstructure showed slightly higher failureenergies at low temperatures This is because the retainedaustenite tends to transform to martensite during impactand thus absorbs some energy and raises the toughness asproposed by Bilmes et al16 At room temperature there wasmore scatter in the energy required to break the specimensbut there was apparently no significant difference betweenthe two microstructures

Similarly there were no apparent differences between theupper shelf failure energies of the fresh martensite andfresh martensitez14 d-ferrite microstructures (Fig 9)However the transition occurred over a shorter tempera-ture range in the biphased microstructure and a significantshift in ductile to brittle transition temperature wasobserved In fact the DBTT of the duplex microstructurewas raised by 52uC to about 246uC The origin of thisincrease could not be attributed directly to the presence ofd-ferrite since there was a 20 mm difference in prioraustenite grain size between the two microstructures

The observation of the transition curves of the freshmartensitez14 d-ferrite and fresh martensitez2 d-ferritemicrostructures of identical grain size removed this uncer-tainty (Fig 10) In fact specimens containing a reducedamount of d-ferrite had significantly better DBTT (286uCcompared to 246uC) As mentioned in the previous sectionafter re-austenitisation and quenching the only differencebetween the two ferrite containing microstructures was theactual level of ferrite Consequently it was clear that for aprior austenite grain size of 80 mm the presence of d-ferritesignificantly affected the transition temperature

Figure 11 compares the transition curves of the freshmartensitez2 d-ferrite (grain size 80 mm) with the temperedmartensitezcr microstructure (grain size 25 mm) The reduc-tion of the ferrite content clearly improves the notch toughnessto a level comparable to that of the as-received condition

FractographyObservation of the fracture surfaces in the SEM revealed avariety of fracture modes In particular ductile rupturecleavage and quasi-cleavage could be identified McDonaldet al summarise these fracture modes as follows1718

(i) Ductile rupture Dimples resulting from microvoidformation and coalescence are characteristic ofductile rupture Microvoids are generally initiatedat inclusions which can sometimes be observedwithin the dimples

(ii) Cleavage Stepped facets are caused by fracturethrough regions of different crystallographic orien-tation Cleavage occurs by the rapid propagation ofa crack along crystallographic planes and may beinitiated by inclusionmatrix decohesion This is avery low energy-absorbing mode of fracture

(iii) Quasi-cleavage A less well-defined fracture modeknown as quasi-cleavage is associated with tear ridgeswhich represent limited plastic deformation separatingcleavage facets Unlike pure cleavage energy isabsorbed as a result of the limited plastic deformation

Table 8 Ductile to brittle transition temperatures obtainedfor each microstructure studied

Microstructure DBTT (uC)

Tempered martensitezcr 298Fresh martensite 298Fresh martensitez14 d-ferrite 246Fresh martensitez2 d-ferrite 286

9 Charpy impact energy curves obtained on temperedmartensitezretained austenite (as-received condition)fresh martensite and fresh martensitez14 d-ferritemicrostructures

10 Charpy impact energy curves obtained on fresh mar-tensitez14 d-ferrite and fresh martensitez2 d-ferrite microstructures Note that both microstructureshave identical prior austenite grain size of 80 mm

11 Comparison of Charpy impact energy curves betweenas-received and fresh martensitez2 d-ferritemicrostructures

382 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

Examples of ductile rupture obtained from the fullymartensitic and martensiticferritic microstructures testedat 0uC are shown in Fig 12 Most cuboidal shaped titaniumcarbonitride particles were cracked

Specimens tested at low temperatures fractured in atransgranular manner (Fig 13) This indicated that thepresence of grain boundary d-ferrite did not embrittle thelatter In general the cleavage facet sizes appeared similarbut the fresh martensitez14 d-ferrite microstructureshowed more secondary cracks

River markings on the cleavage facets of the specimensshowed evidence of both cleavage (Fig 14) and quasi-cleavage fracture (Fig 15) initiated by Ti(CN) inclusionsThe capacity of these particles to reduce toughness is wellknown Brittle fracture induced by the cleavage of TiNinclusions has recently been investigated by Fairchildet al1920 Those particles are well bonded to the matrixand thus allow high stresses in a notch tip plastic zone toact on inclusions without debonding the interface Once aninclusion cleaves the strong bond allows for transfer of thecrack into the martensitic matrix

In addition to the SEM observation cross-sections per-pendicular to the fracture surfaces as shown in Fig 16 havebeen observed to reveal the fracture path Figures 17 and 18show example of such cross-sections for specimens of thefresh martensitez14 d-ferrite and fresh martensitez2d-ferrite microstructures which failed by cleavage

In the fresh martensitez14 d-ferrite specimens sec-ondary intragranular cracks propagated through marten-site and d-ferrite stringers whereas shorter cracks were seento initiate both in martensite and at the d-ferritemartensiteinterface in the fresh martensitez2 d-ferrite In the lattercase cracks could not easily link up when d-ferrite wasmore dispersed and possessed a spheroidised morphologythus explaining the better impact properties

IMPACT PROPERTIES OF GLEEBLESIMULATED SPECIMENSThe first part of this work established the effect of d-ferriteon the ductile to brittle transition temperature of homo-genous and biphased microstructures The microstructure

12 Dimpled fracture surfaces of a fresh martensite and b fresh martensitez14 d-ferrite microstructures tested at 0uC

13 Fracture surfaces of a fresh martensite and b fresh martensitez14 d-ferrite microstructures tested at 2150 and2100uC respectively

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 383

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of a weld HAZ is however far from being homogeneous andit is not known whether the presence of delta-ferrites affectthe HAZ impact properties

In supermartensitic stainless steels five HAZ regions canbe distinguished as illustrated in Fig 192

(i) partially melted region located beside the fusionboundary This is where incipient melting of thenewly formed d-ferrite occurs on heating

(ii) region which transforms completely to d-ferrite onheating with significant grain growth also calledthe coarse-grained HAZ (CG-HAZ)

(iii) high-temperature two-phase region characterisedby partial transformation of austenite to d-ferriteon heating

(iv) region which completely transforms to austeniteduring heating

(v) a low peak temperature two-phase region char-acterised by partial transformation of the tem-pered martensite to austenite during heating

The HAZ represents the area where all phase transforma-tions occur in the solid state and therefore the partiallymelted region should not be regarded as belonging to theHAZ The other regions can be artificially classified intotwo categories depending on the maximum temperaturethey experienced before cooling the high temperature andthe low temperature HAZ The high temperature HAZ

14 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedcleavage

15 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedquasi-cleavage

16 Location of transverse section in Charpy specimens

384 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

(HT-HAZ) includes the prior d-ferrite region and the hightemperature two-phase region whereas the low temperatureHAZ (LT-HAZ) refers to the remainder of the HAZ Anexample of a typical HAZ found in a single pass weldperformed on steel A is shown in Fig 20

As presented in another publication the high temperatureregion of a supermartensitic stainless steel HAZ is likely tocontain non-equilibrium d-ferrite1 High heat inputs singlepass bead on plates such as the one showed in Fig 20 yieldslow cooling rates in the CG-HAZ and favour the formationof Widmanstatten austenite during cooling from the fullyferritic region It is between the Widmanstatten austenitesubunits that ferrite is found in coarse-grained HAZ1 Indual-phase HAZ on the other hand ferrite is usually locatedat grain boundaries and as thin films within the grains Thevolume fraction of d-ferrite is dependent upon the actualwelding conditions but is usually smaller in coarse-grainedHAZ This is because fast cooling rates restrictWidmanstatten austenite formation in coarse-grained HAZbut prevent ferrite dissolution in the dual-phase HAZ9

Gleeble weld-simulatorBeing close to the weld metal sampling of all the HAZregions for Charpy impact testing is difficult To avoid this

problem a Gleeble weld-simulator has been employed TheGleeble 1500 is a thermomechanical testing device that isused in a wide range of applications including simulationand testing of weld heat affected zones A major advantageof the Gleeble is that it generates large volumes ofmicrostructure that simulate small hard to study regionsin actual weldments A range of specimen geometries can beused but by using bars Charpy specimens can directly bemachined after simulation

The machine (Fig 21) is interfaced to a computer that isreadily programmed to provide reference signals for closed-loop control of the applied thermal cycle Heating isaccomplished by the flow of low-frequency alternatingcurrent in the specimen The current distribution iscontrolled by the cross-section of the specimen and thejaws used to mount it The feedback signal necessary forclosed-loop control is obtained from a thermocouplewelded to the specimen surface21

ProcedureTen specimens of dimension 116116100 mm (takenaccording to Fig 2) were machined for HAZ simulationThe operation was performed at the maximum temperature

17 Cross-section of fracture surface of fresh martensitez14 d-ferrite microstructure tested at 2100uC Electrolyticallyetched to reveal d-ferrite

18 Cross-section of fracture surface of fresh martensitez2 d-ferrite microstructure tested at 2150uC Electrolyticallyetched to reveal d-ferrite

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 385

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of 1350uC using a cooling time between 800 and 500uC of10 s After simulation all specimens were polished andetched to reveal the underlying microstructure Standardsize Charpy specimens were finally machined with the notchlying across the centre of the simulated specimens

MicrostructureFigure 22 shows a representative macrograph of thesimulated specimens before grinding and polishing

After polishing and electro-etching the central region ofall simulated specimens showed a microstructural gradientThe latter indicated an inhomogeneous thermal experienceacross the sample sections and might have been caused byan improper contact between the samples and the grips in

the Gleeble machine As Fig 23a shows the central regionof the simulated specimens consisted of three distinct areasThe upper region had semi-circular shape and depth ofabout 25 mm and was identified as the coarse-grained region (Fig 23b) The latter was surrounded by

19 Schematic diagram illustrating structural changes that occur within heat affected zone of single-pass supermartensiticstainless steel weld2

20 Optical micrograph of single-pass bead on plate per-formed on steel A using tungsten inert gas (TIG) pro-cess with heat input of 1 kJ mm21 Plate thickness10 mm Etched electrolytically to reveal d-ferrite

21 Arrangement of equipment within Gleeble simulator

22 Macrograph of typical specimen after HAZ simula-tion using Gleeble simulator Thermocouple employedto record temperature was located at middle of centreline

386 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

the dual-phase region (Fig 23c) and in turn wassurrounded by the prior fully-austenitic area The existenceof this microstructural gradient was not expected but it wasdecided to study its impact properties since the generatedcentre line microstructure was representative of that of areal weld HAZ

The coarse-grained and dual-phase regions had grain sizeof about 160 mm and 30 mm respectively The d-ferritevolume percentage in the dual-phase region measured byimage analysis at a magnification of 200 (Fig 23c) was40iexcl5 This was larger than the volume percentagemeasured in heat treated specimens (Fig 7b and c)

Results of Charpy impact testThe transition curve obtained on the Gleeble simulatedspecimens is presented in Fig 24 For comparison the

results obtained on the as-received and martensiticferriticmicrostructures have been reported The Gleeble specimenshad the lowest upper shelf energy of the three curves Thelatter was probably because of the combined action of largegrain size and presence of d-ferrite in the tested micro-structure More importantly the simulated HAZ had aductile to brittle transition temperature comparable to thatof the as-received material (294 and 298uC respectively)although like the martensitic microstructure containing14 d-ferrite the transition occurred over a shortertemperature range The behaviour of the Gleeble simulatedspecimens at low temperatures was therefore significantlybetter than the ferriticmartensitic microstructure studied inthe first part

FractographyAll specimens tested at temperatures above 2100uC showedductile failure mode whereas those broken at temperaturesbelow 2100uC showed completely brittle behaviourInterestingly the Gleeble specimen tested at 2100uCshowed two fracture modes (Fig 25) Observation of across-section revealed the coarse-grained region to fail bycleavage whereas the dual-phase region showed evidence ofductility below the transition temperature (Fig 26) Theductility of the dual-phase HAZ indicates that the DBTT ofthis region is lower than 2100uC These results provideevidence to support the view that the presence of d-ferrite indual-phase HAZ up to 40 does not significantly affect theimpact properties of the whole high temperature HAZ Thespecimen transition temperature was more strongly influ-enced by the large grain size of the coarse-grained HAZ

SUMMARY AND CONCLUSIONThis work revealed that for a similar strength particlevolume-fraction and distribution and a given prioraustenite grain size the ductile to brittle transitiontemperature was very sensitive to the volume fraction ofd-ferrite in homogeneous microstructures On the otherhand tests carried out on simulated HAZ indicated that thedual-phase region did not play a major role in determiningthe DBTT even when d-ferrite was present up to 40 Thesmall grain size of the dual-phase region offers suitableobstacles to crack propagation Conversely the large grainsize of the coarse-grained region strongly influences theimpact properties of the high temperature HAZ Howeverwhen compared with the tempered parent material contain-ing austenite the transition temperature of the Gleeblesimulated specimens was raised by only a few degreesConsidering that the microstructure observed in simulated

23 Microstructure of Gleeble simulated specimens a lowmagnification of central region b close-up of coarse-grained region and c close-up of dual-phase regionAll specimens were electrolytically etched to reveald-ferrite

24 Charpy impact energy curve of Gleeble simulated spe-cimens and comparison with those of as-received andfresh martensitez14 d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 387

Science and Technology of Welding and Joining 2004 Vol 9 No 5

HAZ is always coarser than that found in a multipass girthweld and that the DBTT of the simulated specimens waswell below room temperature the presence of d-ferrite inHAZ of supermartensitic stainless steels is not likely toconstitute a threat to the impact properties

ACKNOWLEDGEMENTSThe authors are grateful to TWI and the EPSRC forsponsoring this research

REFERENCES1 D CARROUGE and P WOOLLIN in lsquoStainless steel world 2002rsquo

61ndash67 2002 Houston USA2 J ENERHAUG Oslash GRONG and U M STEINSMO Sci Technol Weld

Joining 2001 6 330ndash3383 T G GOOCH and B J GINN lsquoHeat affected zone toughness of

MMA welded 12Cr martensitic-ferritic steelrsquo TechnicalReport 373 Welding Institute Members Report 1988

4 C GROBLER lsquoWeldability studies on 12 and 14 chromiumsteelsrsquo PhD thesis University of Pretoria 1987

5 J J J ZAAYMAN and G T VAN ROOYEN Proc 1st Int ChromiumSteels and Alloys Congress Volume 2 SAIMM Cape TownJohannesburg 1992 137ndash142

6 K KONDO M UEDA K OGAWA H AMAYA H HIRATA andH TAKABE in lsquoSupermartensitic stainless steels 99rsquo 11ndash181999 Belgium

7 T G GOOCH P WOOLLIN and A G HAYNES in lsquoSupermartensiticstainless steels 99rsquo 188ndash195 1999 Belgium

8 P WOOLLIN and D CARROUGE in lsquoSupermartensitic stainlesssteels 2002rsquo 199ndash204 2002 Paper 027

9 D CARROUGE PhD thesis Department of MaterialsScience and Metallurgy University of Cambridge 2002Available from httpwwwmsmcamacukphasetrans2000phdhtmldominique

10 P WOOLLIN lsquoMetallurgical examination of martensitic stainlesssteel weldmentsrsquo Technical Report 620886195 WeldingInstitute Internal Report 1995

11 E BERAHA and B SHPIGLER lsquoColor metallographyrsquo 1977American Society for Metals

12 B D CULLITY lsquoElements of X-ray diffractionrsquo 1978Addison-Wesley

13 BS EN 10045-1 lsquoCharpy impact test on metallic materialsPart 1 Test methodrsquo British Standard 1990

14 D WILLIAM and J R CALLISTER lsquoMaterials science andengineering an introductionrsquo 1999 New York USA JohnWiley and Sons Inc

15 W OLDFIELD CSNI Specialists Meeting on lsquoInstrumentedprecracked Charpy testingrsquo Palo Alto CA 1981 5ndash59

25 Fracture surfaces of a coarse-grained and b dual-phase region of Gleeble simulated specimen tested at 2100uC

26 Cross-section of fracture surface shown in Fig 25 a coarse-grained and b dual-phase region This cross-section wastaken perpendicular to examination plane shown in Fig 16

388 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

carried out The latter were performed on a Phillipsdiffractometer using Cu Ka radiation operating at 40 kVand 40 mA Step scan mode was employed to cover theangular 2h range from 471 to 103u The 2h step size was005u with a dwell time of 30 s Samples were ground andpolished using the normal metallographic preparationprocedure with two cycles of polishing and etching beforethe X-ray diffraction test to remove any deformed layercaused by polishing The diffraction peaks were indexed bycomparing theoretical to experimental interplanar spacingvalues for ferrite and austenite planes

The volume fraction of austenite was estimated frommeasurements of the integrated intensities of martensiteand austenite peaks assuming they are the only phasespresent

As the volume fractions are measured from theintensities which are corrected for the structure and othersystematic factors the patterns obtained did not requirecalibration12 The ratio of the intensities of diffractionpeaks from two phases of a polycrystalline sample is givenby equation (1)12

Ic(hkl)

Ia0(hkl)~

Rc(hkl)

Ra0(hkl)|

Vc

Va0 (1)

where Ic(hkl) and Ia9(hkl) are the integrated intensity from agiven plane (hkl) from the austenite and martensite (orferrite) phase respectively Vc and Va9 are the volumefractions of austenite and martensite respectively andRc (hkl) and Ra9(hkl) are given for a specific peak by

R~1

v2frac12jF j2(p)(Lp)e2m (2)

where v5volume of unit cell F5structure factorp5multiplicity factor Lp5Lorentz-polarisation factorand e22m5temperature factor

Considering that all materials in reality exhibit crystal-lographic texture the average integrated intensity for atleast three specific reflections for austenite and martensite(Table 5) were taken into account (equations (3) and (4))

Ic

Rc~

1

3

X3

i~1

Ici

Rci

(3)

Ia0

Ra0~

1

3

X3

i~1

Ia0

i

Ra0i

(4)

The value of VcVa9 in equation (1) can be obtained fromthe measurement of IcIa9 and the calculation of RcRa9Once VcVa9 is found the value of Vc can be obtained fromthe additional relationship

VczVa0~1 (5)

The 2h values for three austenite peaks were used tocalculate the d spacings with Braggrsquos law and then thelattice parameter a These values were plotted against sin2hAn accurate value of ac was found by extrapolation tosin2h5112

To study the location of any retained austenite aJEOL 200 CX transmission electron microscope (TEM)

operating at 200 kV has been used Thin foils of 3 mm diaand approximately 50 mm thick were obtained from thespecimens employed for X-ray analysis The discs wereelectropolished in a solution of 5 perchloric acid 15glycerol and 80 ethanol at an applied voltage of 50 V atroom temperature Conventional selected-area diffractionwas used to distinguish austenite (fcc structure) from lowcarbon martensite (bcc)

After optical verification of the desired microstructuresthe heat treated specimens were machined to the standarddimension of 10 mm square section 6 55 mm with a 2 mmdeep V-notch positioned through-thickness at the middleof the specimens Charpy impact tests were finally carriedout in accordance with British Standard 10045-113 between20 and 2200uC to reveal any ductilendashbrittle transition

The fracture surfaces of Charpy specimens were observedusing a JEOL 5800 scanning electron microscope (SEM)Secondary electron mode was employed at an acceleratingvoltage of 15 kV To identify the nature of particles micro-analytical experiments were carried out using an energydispersive X-ray (EDX) analysis system ZAF (atomicnumber absorption fluorescence) corrections wereemployed the acquisition time was set to 100 s and thedead time was kept below 20

As-received microstructureThe initial microstructure of the material electrolyticallyetched is shown in Fig 3 It consisted of temperedmartensite with no evidence of d-ferrite and some apparenttitanium carbonitride particles

X-ray diffraction tests revealed that steel A contained asignificant amount of retained austenite (Fig 4) Thevolume percentage of the latter was 13 within anexperimental error of iexcl05 The error calculation wasbased on the integrated intensity since this error ischaracteristic of the specimen studied

This austenite remained stable after immersion in liquidnitrogen for 30 min but after re-austenitisation at 1000uCfor 30 min and water quenching no austenite could bedetected (Fig 5) Those results indicated that the presenceof retained austenite in the as-received condition was theconsequence of the manufacturerrsquos tempering treatment Infact tempering must have been carried out a few degreesabove the steel Ac1 temperature In this case the newaustenite is significantly enriched in carbon and nickel andremains stable after cooling7

A TEM examination of a thin foil prepared on steel Ashowed this retained austenite to be present as smallelongated grains located at martensite lath boundaries asillustrated in Fig 6

3 Optical micrograph of initial microstructure of steel Aelectrolytically etched

Table 5 Diffracting hkl planes used for martensite andaustenite

Phase i Diffracting plane (hkl)

Martensite 1 200Martensite 2 211Martensite 3 220Austenite 1 200Austenite 2 220Austenite 3 311

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 379

Science and Technology of Welding and Joining 2004 Vol 9 No 5

Characterisation of heat treated specimensTo facilitate the interpretation of the results heat treatedspecimens were characterised before impact testing in termsof phase presence volume fraction hardness and grain size

Figure 7 shows the microstructures of the fully martensi-tic martensiticferritic and annealed martensiticferriticspecimens observed on the actual-size Charpy specimensTypical cuboid Ti(CN) precipitates were the only observednon-metallic particles These particles were up to 10 mmlong and occurred in random dispersion or as small distinctclusters

In the first martensiticferritic microstructure d-ferritewas located both at grain boundaries and within thegrains Its volume percentage measured by image analysisat a magnification of 200 was 14iexcl2 Complete re-austenitisation of the latter microstructure gave a reductionof the d-ferrite content to 2iexcl05 and spheroidisation ofthe remaining ferrite The dissolution of d-ferrite appearedto be a slow process since 2 ferrite were still present afteran hour at 1100uC This is because significant partitioningof substitutional solutes occurred at 1350uC as indicated bythe EDX measurements reported in Table 6

There were no apparent differences in the microstructurein terms of Ti(CN) particles This was expected since theferritising and the re-austenitising operations were per-formed above the predicted temperature of TiC dissolution(about 900uC) and water quenching prevents precipitationduring cooling

The hardness and grain sizes measured on all materialsare reported in Table 7 Despite the presence of 14

5 X-ray diffraction pattern of steel A after indicatedthermal treatment (WQ water quench)

6 a Bright field image of martensite plus retained auste-nite region b dark field image of same region as a cselected area diffraction pattern of austenite corre-sponding to dark field image

Table 6 Summary of 15 EDX microanalyses (wt-) per-formed on ferrite and martensite

Composition

Phase Cr Ni Mo

d-ferrite 135iexcl09 48iexcl10 26iexcl04Matrix 117iexcl03 70iexcl04 17iexcl02

4 X-ray diffraction pattern of steel A in as-received condition

Table 7 Macrohardness (HV10) and grain size of differ-ent microstructures

Microstructure Hardness Grain size mm

Tempered martensitezcr 296iexcl3 25iexcl4Fresh martensite 317iexcl8 60iexcl9Fresh martensitez14 d-ferrite 311iexcl8 80iexcl12Fresh martensitez2 d-ferrite 304iexcl7 80iexcl12

380 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

d-ferrite there was little hardness change between thedifferent microstructures The latter was attributed to thevirtually carbon free martensite as a result of the presence oftitanium in concentrations greater than the stoichiometricamount As tensile strength is proportional to hardness allspecimens could be assumed to have a similar strength14

There was a small difference in grain size between the variousmicrostructures In fact the prior austenite grain size of thespecimens heat treated at 1350uC was about 20 mm largerthan the one heat treated at 1100uC As a result of the grainboundary pinning effect of d-ferrite there was no grain sizedifference between the microstructures consisting of freshmartensite and various amounts of ferrite

The presence of retained austenite in the furnace heattreated specimens has been checked using X-ray diffractionNo austenite could be detected in any of the specimens(Fig 8)

Results of Charpy impact testsTo facilitate the comparison and the determination of theductile to brittle transition temperature (DBTT) thetoughness results have been fitted by a non-linear regressiontechnique to the equation15

Impact energy (J)~A0zB0 tanhTT0

C0

(6)

where A0 B0 C0 and T0 are fitting constants and T isthe temperature in uC The DBTT was taken as thetemperature corresponding to half the height of thetransition curve Table 8 summarises the measured ductile

7 Microstructures obtained after isothermal treatment ata 1100uC and b 1350uC for 15 min and water quenchedc Microstructure obtained after treatment at 1350uC for15 min and water quenched followed by re-austenitisationat 1100uC for 60 min and water quenched

8 X-ray diffraction patterns of fresh martensite freshmartensitez14 d-ferrite and fresh martensitez2d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 381

Science and Technology of Welding and Joining 2004 Vol 9 No 5

to brittle transition temperatures for each tested micro-structure

Figure 9 shows the impact transition curves obtainedfrom the tempered martensitezcr (as-received condition)fresh martensite and fresh martensitez14 d-ferritemicrostructures The transition curves of the as-receivedand fresh martensite microstructures were similar and theirDBTT were identical and far below 0uC However the as-received microstructure showed slightly higher failureenergies at low temperatures This is because the retainedaustenite tends to transform to martensite during impactand thus absorbs some energy and raises the toughness asproposed by Bilmes et al16 At room temperature there wasmore scatter in the energy required to break the specimensbut there was apparently no significant difference betweenthe two microstructures

Similarly there were no apparent differences between theupper shelf failure energies of the fresh martensite andfresh martensitez14 d-ferrite microstructures (Fig 9)However the transition occurred over a shorter tempera-ture range in the biphased microstructure and a significantshift in ductile to brittle transition temperature wasobserved In fact the DBTT of the duplex microstructurewas raised by 52uC to about 246uC The origin of thisincrease could not be attributed directly to the presence ofd-ferrite since there was a 20 mm difference in prioraustenite grain size between the two microstructures

The observation of the transition curves of the freshmartensitez14 d-ferrite and fresh martensitez2 d-ferritemicrostructures of identical grain size removed this uncer-tainty (Fig 10) In fact specimens containing a reducedamount of d-ferrite had significantly better DBTT (286uCcompared to 246uC) As mentioned in the previous sectionafter re-austenitisation and quenching the only differencebetween the two ferrite containing microstructures was theactual level of ferrite Consequently it was clear that for aprior austenite grain size of 80 mm the presence of d-ferritesignificantly affected the transition temperature

Figure 11 compares the transition curves of the freshmartensitez2 d-ferrite (grain size 80 mm) with the temperedmartensitezcr microstructure (grain size 25 mm) The reduc-tion of the ferrite content clearly improves the notch toughnessto a level comparable to that of the as-received condition

FractographyObservation of the fracture surfaces in the SEM revealed avariety of fracture modes In particular ductile rupturecleavage and quasi-cleavage could be identified McDonaldet al summarise these fracture modes as follows1718

(i) Ductile rupture Dimples resulting from microvoidformation and coalescence are characteristic ofductile rupture Microvoids are generally initiatedat inclusions which can sometimes be observedwithin the dimples

(ii) Cleavage Stepped facets are caused by fracturethrough regions of different crystallographic orien-tation Cleavage occurs by the rapid propagation ofa crack along crystallographic planes and may beinitiated by inclusionmatrix decohesion This is avery low energy-absorbing mode of fracture

(iii) Quasi-cleavage A less well-defined fracture modeknown as quasi-cleavage is associated with tear ridgeswhich represent limited plastic deformation separatingcleavage facets Unlike pure cleavage energy isabsorbed as a result of the limited plastic deformation

Table 8 Ductile to brittle transition temperatures obtainedfor each microstructure studied

Microstructure DBTT (uC)

Tempered martensitezcr 298Fresh martensite 298Fresh martensitez14 d-ferrite 246Fresh martensitez2 d-ferrite 286

9 Charpy impact energy curves obtained on temperedmartensitezretained austenite (as-received condition)fresh martensite and fresh martensitez14 d-ferritemicrostructures

10 Charpy impact energy curves obtained on fresh mar-tensitez14 d-ferrite and fresh martensitez2 d-ferrite microstructures Note that both microstructureshave identical prior austenite grain size of 80 mm

11 Comparison of Charpy impact energy curves betweenas-received and fresh martensitez2 d-ferritemicrostructures

382 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

Examples of ductile rupture obtained from the fullymartensitic and martensiticferritic microstructures testedat 0uC are shown in Fig 12 Most cuboidal shaped titaniumcarbonitride particles were cracked

Specimens tested at low temperatures fractured in atransgranular manner (Fig 13) This indicated that thepresence of grain boundary d-ferrite did not embrittle thelatter In general the cleavage facet sizes appeared similarbut the fresh martensitez14 d-ferrite microstructureshowed more secondary cracks

River markings on the cleavage facets of the specimensshowed evidence of both cleavage (Fig 14) and quasi-cleavage fracture (Fig 15) initiated by Ti(CN) inclusionsThe capacity of these particles to reduce toughness is wellknown Brittle fracture induced by the cleavage of TiNinclusions has recently been investigated by Fairchildet al1920 Those particles are well bonded to the matrixand thus allow high stresses in a notch tip plastic zone toact on inclusions without debonding the interface Once aninclusion cleaves the strong bond allows for transfer of thecrack into the martensitic matrix

In addition to the SEM observation cross-sections per-pendicular to the fracture surfaces as shown in Fig 16 havebeen observed to reveal the fracture path Figures 17 and 18show example of such cross-sections for specimens of thefresh martensitez14 d-ferrite and fresh martensitez2d-ferrite microstructures which failed by cleavage

In the fresh martensitez14 d-ferrite specimens sec-ondary intragranular cracks propagated through marten-site and d-ferrite stringers whereas shorter cracks were seento initiate both in martensite and at the d-ferritemartensiteinterface in the fresh martensitez2 d-ferrite In the lattercase cracks could not easily link up when d-ferrite wasmore dispersed and possessed a spheroidised morphologythus explaining the better impact properties

IMPACT PROPERTIES OF GLEEBLESIMULATED SPECIMENSThe first part of this work established the effect of d-ferriteon the ductile to brittle transition temperature of homo-genous and biphased microstructures The microstructure

12 Dimpled fracture surfaces of a fresh martensite and b fresh martensitez14 d-ferrite microstructures tested at 0uC

13 Fracture surfaces of a fresh martensite and b fresh martensitez14 d-ferrite microstructures tested at 2150 and2100uC respectively

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 383

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of a weld HAZ is however far from being homogeneous andit is not known whether the presence of delta-ferrites affectthe HAZ impact properties

In supermartensitic stainless steels five HAZ regions canbe distinguished as illustrated in Fig 192

(i) partially melted region located beside the fusionboundary This is where incipient melting of thenewly formed d-ferrite occurs on heating

(ii) region which transforms completely to d-ferrite onheating with significant grain growth also calledthe coarse-grained HAZ (CG-HAZ)

(iii) high-temperature two-phase region characterisedby partial transformation of austenite to d-ferriteon heating

(iv) region which completely transforms to austeniteduring heating

(v) a low peak temperature two-phase region char-acterised by partial transformation of the tem-pered martensite to austenite during heating

The HAZ represents the area where all phase transforma-tions occur in the solid state and therefore the partiallymelted region should not be regarded as belonging to theHAZ The other regions can be artificially classified intotwo categories depending on the maximum temperaturethey experienced before cooling the high temperature andthe low temperature HAZ The high temperature HAZ

14 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedcleavage

15 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedquasi-cleavage

16 Location of transverse section in Charpy specimens

384 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

(HT-HAZ) includes the prior d-ferrite region and the hightemperature two-phase region whereas the low temperatureHAZ (LT-HAZ) refers to the remainder of the HAZ Anexample of a typical HAZ found in a single pass weldperformed on steel A is shown in Fig 20

As presented in another publication the high temperatureregion of a supermartensitic stainless steel HAZ is likely tocontain non-equilibrium d-ferrite1 High heat inputs singlepass bead on plates such as the one showed in Fig 20 yieldslow cooling rates in the CG-HAZ and favour the formationof Widmanstatten austenite during cooling from the fullyferritic region It is between the Widmanstatten austenitesubunits that ferrite is found in coarse-grained HAZ1 Indual-phase HAZ on the other hand ferrite is usually locatedat grain boundaries and as thin films within the grains Thevolume fraction of d-ferrite is dependent upon the actualwelding conditions but is usually smaller in coarse-grainedHAZ This is because fast cooling rates restrictWidmanstatten austenite formation in coarse-grained HAZbut prevent ferrite dissolution in the dual-phase HAZ9

Gleeble weld-simulatorBeing close to the weld metal sampling of all the HAZregions for Charpy impact testing is difficult To avoid this

problem a Gleeble weld-simulator has been employed TheGleeble 1500 is a thermomechanical testing device that isused in a wide range of applications including simulationand testing of weld heat affected zones A major advantageof the Gleeble is that it generates large volumes ofmicrostructure that simulate small hard to study regionsin actual weldments A range of specimen geometries can beused but by using bars Charpy specimens can directly bemachined after simulation

The machine (Fig 21) is interfaced to a computer that isreadily programmed to provide reference signals for closed-loop control of the applied thermal cycle Heating isaccomplished by the flow of low-frequency alternatingcurrent in the specimen The current distribution iscontrolled by the cross-section of the specimen and thejaws used to mount it The feedback signal necessary forclosed-loop control is obtained from a thermocouplewelded to the specimen surface21

ProcedureTen specimens of dimension 116116100 mm (takenaccording to Fig 2) were machined for HAZ simulationThe operation was performed at the maximum temperature

17 Cross-section of fracture surface of fresh martensitez14 d-ferrite microstructure tested at 2100uC Electrolyticallyetched to reveal d-ferrite

18 Cross-section of fracture surface of fresh martensitez2 d-ferrite microstructure tested at 2150uC Electrolyticallyetched to reveal d-ferrite

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 385

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of 1350uC using a cooling time between 800 and 500uC of10 s After simulation all specimens were polished andetched to reveal the underlying microstructure Standardsize Charpy specimens were finally machined with the notchlying across the centre of the simulated specimens

MicrostructureFigure 22 shows a representative macrograph of thesimulated specimens before grinding and polishing

After polishing and electro-etching the central region ofall simulated specimens showed a microstructural gradientThe latter indicated an inhomogeneous thermal experienceacross the sample sections and might have been caused byan improper contact between the samples and the grips in

the Gleeble machine As Fig 23a shows the central regionof the simulated specimens consisted of three distinct areasThe upper region had semi-circular shape and depth ofabout 25 mm and was identified as the coarse-grained region (Fig 23b) The latter was surrounded by

19 Schematic diagram illustrating structural changes that occur within heat affected zone of single-pass supermartensiticstainless steel weld2

20 Optical micrograph of single-pass bead on plate per-formed on steel A using tungsten inert gas (TIG) pro-cess with heat input of 1 kJ mm21 Plate thickness10 mm Etched electrolytically to reveal d-ferrite

21 Arrangement of equipment within Gleeble simulator

22 Macrograph of typical specimen after HAZ simula-tion using Gleeble simulator Thermocouple employedto record temperature was located at middle of centreline

386 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

the dual-phase region (Fig 23c) and in turn wassurrounded by the prior fully-austenitic area The existenceof this microstructural gradient was not expected but it wasdecided to study its impact properties since the generatedcentre line microstructure was representative of that of areal weld HAZ

The coarse-grained and dual-phase regions had grain sizeof about 160 mm and 30 mm respectively The d-ferritevolume percentage in the dual-phase region measured byimage analysis at a magnification of 200 (Fig 23c) was40iexcl5 This was larger than the volume percentagemeasured in heat treated specimens (Fig 7b and c)

Results of Charpy impact testThe transition curve obtained on the Gleeble simulatedspecimens is presented in Fig 24 For comparison the

results obtained on the as-received and martensiticferriticmicrostructures have been reported The Gleeble specimenshad the lowest upper shelf energy of the three curves Thelatter was probably because of the combined action of largegrain size and presence of d-ferrite in the tested micro-structure More importantly the simulated HAZ had aductile to brittle transition temperature comparable to thatof the as-received material (294 and 298uC respectively)although like the martensitic microstructure containing14 d-ferrite the transition occurred over a shortertemperature range The behaviour of the Gleeble simulatedspecimens at low temperatures was therefore significantlybetter than the ferriticmartensitic microstructure studied inthe first part

FractographyAll specimens tested at temperatures above 2100uC showedductile failure mode whereas those broken at temperaturesbelow 2100uC showed completely brittle behaviourInterestingly the Gleeble specimen tested at 2100uCshowed two fracture modes (Fig 25) Observation of across-section revealed the coarse-grained region to fail bycleavage whereas the dual-phase region showed evidence ofductility below the transition temperature (Fig 26) Theductility of the dual-phase HAZ indicates that the DBTT ofthis region is lower than 2100uC These results provideevidence to support the view that the presence of d-ferrite indual-phase HAZ up to 40 does not significantly affect theimpact properties of the whole high temperature HAZ Thespecimen transition temperature was more strongly influ-enced by the large grain size of the coarse-grained HAZ

SUMMARY AND CONCLUSIONThis work revealed that for a similar strength particlevolume-fraction and distribution and a given prioraustenite grain size the ductile to brittle transitiontemperature was very sensitive to the volume fraction ofd-ferrite in homogeneous microstructures On the otherhand tests carried out on simulated HAZ indicated that thedual-phase region did not play a major role in determiningthe DBTT even when d-ferrite was present up to 40 Thesmall grain size of the dual-phase region offers suitableobstacles to crack propagation Conversely the large grainsize of the coarse-grained region strongly influences theimpact properties of the high temperature HAZ Howeverwhen compared with the tempered parent material contain-ing austenite the transition temperature of the Gleeblesimulated specimens was raised by only a few degreesConsidering that the microstructure observed in simulated

23 Microstructure of Gleeble simulated specimens a lowmagnification of central region b close-up of coarse-grained region and c close-up of dual-phase regionAll specimens were electrolytically etched to reveald-ferrite

24 Charpy impact energy curve of Gleeble simulated spe-cimens and comparison with those of as-received andfresh martensitez14 d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 387

Science and Technology of Welding and Joining 2004 Vol 9 No 5

HAZ is always coarser than that found in a multipass girthweld and that the DBTT of the simulated specimens waswell below room temperature the presence of d-ferrite inHAZ of supermartensitic stainless steels is not likely toconstitute a threat to the impact properties

ACKNOWLEDGEMENTSThe authors are grateful to TWI and the EPSRC forsponsoring this research

REFERENCES1 D CARROUGE and P WOOLLIN in lsquoStainless steel world 2002rsquo

61ndash67 2002 Houston USA2 J ENERHAUG Oslash GRONG and U M STEINSMO Sci Technol Weld

Joining 2001 6 330ndash3383 T G GOOCH and B J GINN lsquoHeat affected zone toughness of

MMA welded 12Cr martensitic-ferritic steelrsquo TechnicalReport 373 Welding Institute Members Report 1988

4 C GROBLER lsquoWeldability studies on 12 and 14 chromiumsteelsrsquo PhD thesis University of Pretoria 1987

5 J J J ZAAYMAN and G T VAN ROOYEN Proc 1st Int ChromiumSteels and Alloys Congress Volume 2 SAIMM Cape TownJohannesburg 1992 137ndash142

6 K KONDO M UEDA K OGAWA H AMAYA H HIRATA andH TAKABE in lsquoSupermartensitic stainless steels 99rsquo 11ndash181999 Belgium

7 T G GOOCH P WOOLLIN and A G HAYNES in lsquoSupermartensiticstainless steels 99rsquo 188ndash195 1999 Belgium

8 P WOOLLIN and D CARROUGE in lsquoSupermartensitic stainlesssteels 2002rsquo 199ndash204 2002 Paper 027

9 D CARROUGE PhD thesis Department of MaterialsScience and Metallurgy University of Cambridge 2002Available from httpwwwmsmcamacukphasetrans2000phdhtmldominique

10 P WOOLLIN lsquoMetallurgical examination of martensitic stainlesssteel weldmentsrsquo Technical Report 620886195 WeldingInstitute Internal Report 1995

11 E BERAHA and B SHPIGLER lsquoColor metallographyrsquo 1977American Society for Metals

12 B D CULLITY lsquoElements of X-ray diffractionrsquo 1978Addison-Wesley

13 BS EN 10045-1 lsquoCharpy impact test on metallic materialsPart 1 Test methodrsquo British Standard 1990

14 D WILLIAM and J R CALLISTER lsquoMaterials science andengineering an introductionrsquo 1999 New York USA JohnWiley and Sons Inc

15 W OLDFIELD CSNI Specialists Meeting on lsquoInstrumentedprecracked Charpy testingrsquo Palo Alto CA 1981 5ndash59

25 Fracture surfaces of a coarse-grained and b dual-phase region of Gleeble simulated specimen tested at 2100uC

26 Cross-section of fracture surface shown in Fig 25 a coarse-grained and b dual-phase region This cross-section wastaken perpendicular to examination plane shown in Fig 16

388 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

Characterisation of heat treated specimensTo facilitate the interpretation of the results heat treatedspecimens were characterised before impact testing in termsof phase presence volume fraction hardness and grain size

Figure 7 shows the microstructures of the fully martensi-tic martensiticferritic and annealed martensiticferriticspecimens observed on the actual-size Charpy specimensTypical cuboid Ti(CN) precipitates were the only observednon-metallic particles These particles were up to 10 mmlong and occurred in random dispersion or as small distinctclusters

In the first martensiticferritic microstructure d-ferritewas located both at grain boundaries and within thegrains Its volume percentage measured by image analysisat a magnification of 200 was 14iexcl2 Complete re-austenitisation of the latter microstructure gave a reductionof the d-ferrite content to 2iexcl05 and spheroidisation ofthe remaining ferrite The dissolution of d-ferrite appearedto be a slow process since 2 ferrite were still present afteran hour at 1100uC This is because significant partitioningof substitutional solutes occurred at 1350uC as indicated bythe EDX measurements reported in Table 6

There were no apparent differences in the microstructurein terms of Ti(CN) particles This was expected since theferritising and the re-austenitising operations were per-formed above the predicted temperature of TiC dissolution(about 900uC) and water quenching prevents precipitationduring cooling

The hardness and grain sizes measured on all materialsare reported in Table 7 Despite the presence of 14

5 X-ray diffraction pattern of steel A after indicatedthermal treatment (WQ water quench)

6 a Bright field image of martensite plus retained auste-nite region b dark field image of same region as a cselected area diffraction pattern of austenite corre-sponding to dark field image

Table 6 Summary of 15 EDX microanalyses (wt-) per-formed on ferrite and martensite

Composition

Phase Cr Ni Mo

d-ferrite 135iexcl09 48iexcl10 26iexcl04Matrix 117iexcl03 70iexcl04 17iexcl02

4 X-ray diffraction pattern of steel A in as-received condition

Table 7 Macrohardness (HV10) and grain size of differ-ent microstructures

Microstructure Hardness Grain size mm

Tempered martensitezcr 296iexcl3 25iexcl4Fresh martensite 317iexcl8 60iexcl9Fresh martensitez14 d-ferrite 311iexcl8 80iexcl12Fresh martensitez2 d-ferrite 304iexcl7 80iexcl12

380 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

d-ferrite there was little hardness change between thedifferent microstructures The latter was attributed to thevirtually carbon free martensite as a result of the presence oftitanium in concentrations greater than the stoichiometricamount As tensile strength is proportional to hardness allspecimens could be assumed to have a similar strength14

There was a small difference in grain size between the variousmicrostructures In fact the prior austenite grain size of thespecimens heat treated at 1350uC was about 20 mm largerthan the one heat treated at 1100uC As a result of the grainboundary pinning effect of d-ferrite there was no grain sizedifference between the microstructures consisting of freshmartensite and various amounts of ferrite

The presence of retained austenite in the furnace heattreated specimens has been checked using X-ray diffractionNo austenite could be detected in any of the specimens(Fig 8)

Results of Charpy impact testsTo facilitate the comparison and the determination of theductile to brittle transition temperature (DBTT) thetoughness results have been fitted by a non-linear regressiontechnique to the equation15

Impact energy (J)~A0zB0 tanhTT0

C0

(6)

where A0 B0 C0 and T0 are fitting constants and T isthe temperature in uC The DBTT was taken as thetemperature corresponding to half the height of thetransition curve Table 8 summarises the measured ductile

7 Microstructures obtained after isothermal treatment ata 1100uC and b 1350uC for 15 min and water quenchedc Microstructure obtained after treatment at 1350uC for15 min and water quenched followed by re-austenitisationat 1100uC for 60 min and water quenched

8 X-ray diffraction patterns of fresh martensite freshmartensitez14 d-ferrite and fresh martensitez2d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 381

Science and Technology of Welding and Joining 2004 Vol 9 No 5

to brittle transition temperatures for each tested micro-structure

Figure 9 shows the impact transition curves obtainedfrom the tempered martensitezcr (as-received condition)fresh martensite and fresh martensitez14 d-ferritemicrostructures The transition curves of the as-receivedand fresh martensite microstructures were similar and theirDBTT were identical and far below 0uC However the as-received microstructure showed slightly higher failureenergies at low temperatures This is because the retainedaustenite tends to transform to martensite during impactand thus absorbs some energy and raises the toughness asproposed by Bilmes et al16 At room temperature there wasmore scatter in the energy required to break the specimensbut there was apparently no significant difference betweenthe two microstructures

Similarly there were no apparent differences between theupper shelf failure energies of the fresh martensite andfresh martensitez14 d-ferrite microstructures (Fig 9)However the transition occurred over a shorter tempera-ture range in the biphased microstructure and a significantshift in ductile to brittle transition temperature wasobserved In fact the DBTT of the duplex microstructurewas raised by 52uC to about 246uC The origin of thisincrease could not be attributed directly to the presence ofd-ferrite since there was a 20 mm difference in prioraustenite grain size between the two microstructures

The observation of the transition curves of the freshmartensitez14 d-ferrite and fresh martensitez2 d-ferritemicrostructures of identical grain size removed this uncer-tainty (Fig 10) In fact specimens containing a reducedamount of d-ferrite had significantly better DBTT (286uCcompared to 246uC) As mentioned in the previous sectionafter re-austenitisation and quenching the only differencebetween the two ferrite containing microstructures was theactual level of ferrite Consequently it was clear that for aprior austenite grain size of 80 mm the presence of d-ferritesignificantly affected the transition temperature

Figure 11 compares the transition curves of the freshmartensitez2 d-ferrite (grain size 80 mm) with the temperedmartensitezcr microstructure (grain size 25 mm) The reduc-tion of the ferrite content clearly improves the notch toughnessto a level comparable to that of the as-received condition

FractographyObservation of the fracture surfaces in the SEM revealed avariety of fracture modes In particular ductile rupturecleavage and quasi-cleavage could be identified McDonaldet al summarise these fracture modes as follows1718

(i) Ductile rupture Dimples resulting from microvoidformation and coalescence are characteristic ofductile rupture Microvoids are generally initiatedat inclusions which can sometimes be observedwithin the dimples

(ii) Cleavage Stepped facets are caused by fracturethrough regions of different crystallographic orien-tation Cleavage occurs by the rapid propagation ofa crack along crystallographic planes and may beinitiated by inclusionmatrix decohesion This is avery low energy-absorbing mode of fracture

(iii) Quasi-cleavage A less well-defined fracture modeknown as quasi-cleavage is associated with tear ridgeswhich represent limited plastic deformation separatingcleavage facets Unlike pure cleavage energy isabsorbed as a result of the limited plastic deformation

Table 8 Ductile to brittle transition temperatures obtainedfor each microstructure studied

Microstructure DBTT (uC)

Tempered martensitezcr 298Fresh martensite 298Fresh martensitez14 d-ferrite 246Fresh martensitez2 d-ferrite 286

9 Charpy impact energy curves obtained on temperedmartensitezretained austenite (as-received condition)fresh martensite and fresh martensitez14 d-ferritemicrostructures

10 Charpy impact energy curves obtained on fresh mar-tensitez14 d-ferrite and fresh martensitez2 d-ferrite microstructures Note that both microstructureshave identical prior austenite grain size of 80 mm

11 Comparison of Charpy impact energy curves betweenas-received and fresh martensitez2 d-ferritemicrostructures

382 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

Examples of ductile rupture obtained from the fullymartensitic and martensiticferritic microstructures testedat 0uC are shown in Fig 12 Most cuboidal shaped titaniumcarbonitride particles were cracked

Specimens tested at low temperatures fractured in atransgranular manner (Fig 13) This indicated that thepresence of grain boundary d-ferrite did not embrittle thelatter In general the cleavage facet sizes appeared similarbut the fresh martensitez14 d-ferrite microstructureshowed more secondary cracks

River markings on the cleavage facets of the specimensshowed evidence of both cleavage (Fig 14) and quasi-cleavage fracture (Fig 15) initiated by Ti(CN) inclusionsThe capacity of these particles to reduce toughness is wellknown Brittle fracture induced by the cleavage of TiNinclusions has recently been investigated by Fairchildet al1920 Those particles are well bonded to the matrixand thus allow high stresses in a notch tip plastic zone toact on inclusions without debonding the interface Once aninclusion cleaves the strong bond allows for transfer of thecrack into the martensitic matrix

In addition to the SEM observation cross-sections per-pendicular to the fracture surfaces as shown in Fig 16 havebeen observed to reveal the fracture path Figures 17 and 18show example of such cross-sections for specimens of thefresh martensitez14 d-ferrite and fresh martensitez2d-ferrite microstructures which failed by cleavage

In the fresh martensitez14 d-ferrite specimens sec-ondary intragranular cracks propagated through marten-site and d-ferrite stringers whereas shorter cracks were seento initiate both in martensite and at the d-ferritemartensiteinterface in the fresh martensitez2 d-ferrite In the lattercase cracks could not easily link up when d-ferrite wasmore dispersed and possessed a spheroidised morphologythus explaining the better impact properties

IMPACT PROPERTIES OF GLEEBLESIMULATED SPECIMENSThe first part of this work established the effect of d-ferriteon the ductile to brittle transition temperature of homo-genous and biphased microstructures The microstructure

12 Dimpled fracture surfaces of a fresh martensite and b fresh martensitez14 d-ferrite microstructures tested at 0uC

13 Fracture surfaces of a fresh martensite and b fresh martensitez14 d-ferrite microstructures tested at 2150 and2100uC respectively

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 383

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of a weld HAZ is however far from being homogeneous andit is not known whether the presence of delta-ferrites affectthe HAZ impact properties

In supermartensitic stainless steels five HAZ regions canbe distinguished as illustrated in Fig 192

(i) partially melted region located beside the fusionboundary This is where incipient melting of thenewly formed d-ferrite occurs on heating

(ii) region which transforms completely to d-ferrite onheating with significant grain growth also calledthe coarse-grained HAZ (CG-HAZ)

(iii) high-temperature two-phase region characterisedby partial transformation of austenite to d-ferriteon heating

(iv) region which completely transforms to austeniteduring heating

(v) a low peak temperature two-phase region char-acterised by partial transformation of the tem-pered martensite to austenite during heating

The HAZ represents the area where all phase transforma-tions occur in the solid state and therefore the partiallymelted region should not be regarded as belonging to theHAZ The other regions can be artificially classified intotwo categories depending on the maximum temperaturethey experienced before cooling the high temperature andthe low temperature HAZ The high temperature HAZ

14 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedcleavage

15 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedquasi-cleavage

16 Location of transverse section in Charpy specimens

384 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

(HT-HAZ) includes the prior d-ferrite region and the hightemperature two-phase region whereas the low temperatureHAZ (LT-HAZ) refers to the remainder of the HAZ Anexample of a typical HAZ found in a single pass weldperformed on steel A is shown in Fig 20

As presented in another publication the high temperatureregion of a supermartensitic stainless steel HAZ is likely tocontain non-equilibrium d-ferrite1 High heat inputs singlepass bead on plates such as the one showed in Fig 20 yieldslow cooling rates in the CG-HAZ and favour the formationof Widmanstatten austenite during cooling from the fullyferritic region It is between the Widmanstatten austenitesubunits that ferrite is found in coarse-grained HAZ1 Indual-phase HAZ on the other hand ferrite is usually locatedat grain boundaries and as thin films within the grains Thevolume fraction of d-ferrite is dependent upon the actualwelding conditions but is usually smaller in coarse-grainedHAZ This is because fast cooling rates restrictWidmanstatten austenite formation in coarse-grained HAZbut prevent ferrite dissolution in the dual-phase HAZ9

Gleeble weld-simulatorBeing close to the weld metal sampling of all the HAZregions for Charpy impact testing is difficult To avoid this

problem a Gleeble weld-simulator has been employed TheGleeble 1500 is a thermomechanical testing device that isused in a wide range of applications including simulationand testing of weld heat affected zones A major advantageof the Gleeble is that it generates large volumes ofmicrostructure that simulate small hard to study regionsin actual weldments A range of specimen geometries can beused but by using bars Charpy specimens can directly bemachined after simulation

The machine (Fig 21) is interfaced to a computer that isreadily programmed to provide reference signals for closed-loop control of the applied thermal cycle Heating isaccomplished by the flow of low-frequency alternatingcurrent in the specimen The current distribution iscontrolled by the cross-section of the specimen and thejaws used to mount it The feedback signal necessary forclosed-loop control is obtained from a thermocouplewelded to the specimen surface21

ProcedureTen specimens of dimension 116116100 mm (takenaccording to Fig 2) were machined for HAZ simulationThe operation was performed at the maximum temperature

17 Cross-section of fracture surface of fresh martensitez14 d-ferrite microstructure tested at 2100uC Electrolyticallyetched to reveal d-ferrite

18 Cross-section of fracture surface of fresh martensitez2 d-ferrite microstructure tested at 2150uC Electrolyticallyetched to reveal d-ferrite

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 385

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of 1350uC using a cooling time between 800 and 500uC of10 s After simulation all specimens were polished andetched to reveal the underlying microstructure Standardsize Charpy specimens were finally machined with the notchlying across the centre of the simulated specimens

MicrostructureFigure 22 shows a representative macrograph of thesimulated specimens before grinding and polishing

After polishing and electro-etching the central region ofall simulated specimens showed a microstructural gradientThe latter indicated an inhomogeneous thermal experienceacross the sample sections and might have been caused byan improper contact between the samples and the grips in

the Gleeble machine As Fig 23a shows the central regionof the simulated specimens consisted of three distinct areasThe upper region had semi-circular shape and depth ofabout 25 mm and was identified as the coarse-grained region (Fig 23b) The latter was surrounded by

19 Schematic diagram illustrating structural changes that occur within heat affected zone of single-pass supermartensiticstainless steel weld2

20 Optical micrograph of single-pass bead on plate per-formed on steel A using tungsten inert gas (TIG) pro-cess with heat input of 1 kJ mm21 Plate thickness10 mm Etched electrolytically to reveal d-ferrite

21 Arrangement of equipment within Gleeble simulator

22 Macrograph of typical specimen after HAZ simula-tion using Gleeble simulator Thermocouple employedto record temperature was located at middle of centreline

386 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

the dual-phase region (Fig 23c) and in turn wassurrounded by the prior fully-austenitic area The existenceof this microstructural gradient was not expected but it wasdecided to study its impact properties since the generatedcentre line microstructure was representative of that of areal weld HAZ

The coarse-grained and dual-phase regions had grain sizeof about 160 mm and 30 mm respectively The d-ferritevolume percentage in the dual-phase region measured byimage analysis at a magnification of 200 (Fig 23c) was40iexcl5 This was larger than the volume percentagemeasured in heat treated specimens (Fig 7b and c)

Results of Charpy impact testThe transition curve obtained on the Gleeble simulatedspecimens is presented in Fig 24 For comparison the

results obtained on the as-received and martensiticferriticmicrostructures have been reported The Gleeble specimenshad the lowest upper shelf energy of the three curves Thelatter was probably because of the combined action of largegrain size and presence of d-ferrite in the tested micro-structure More importantly the simulated HAZ had aductile to brittle transition temperature comparable to thatof the as-received material (294 and 298uC respectively)although like the martensitic microstructure containing14 d-ferrite the transition occurred over a shortertemperature range The behaviour of the Gleeble simulatedspecimens at low temperatures was therefore significantlybetter than the ferriticmartensitic microstructure studied inthe first part

FractographyAll specimens tested at temperatures above 2100uC showedductile failure mode whereas those broken at temperaturesbelow 2100uC showed completely brittle behaviourInterestingly the Gleeble specimen tested at 2100uCshowed two fracture modes (Fig 25) Observation of across-section revealed the coarse-grained region to fail bycleavage whereas the dual-phase region showed evidence ofductility below the transition temperature (Fig 26) Theductility of the dual-phase HAZ indicates that the DBTT ofthis region is lower than 2100uC These results provideevidence to support the view that the presence of d-ferrite indual-phase HAZ up to 40 does not significantly affect theimpact properties of the whole high temperature HAZ Thespecimen transition temperature was more strongly influ-enced by the large grain size of the coarse-grained HAZ

SUMMARY AND CONCLUSIONThis work revealed that for a similar strength particlevolume-fraction and distribution and a given prioraustenite grain size the ductile to brittle transitiontemperature was very sensitive to the volume fraction ofd-ferrite in homogeneous microstructures On the otherhand tests carried out on simulated HAZ indicated that thedual-phase region did not play a major role in determiningthe DBTT even when d-ferrite was present up to 40 Thesmall grain size of the dual-phase region offers suitableobstacles to crack propagation Conversely the large grainsize of the coarse-grained region strongly influences theimpact properties of the high temperature HAZ Howeverwhen compared with the tempered parent material contain-ing austenite the transition temperature of the Gleeblesimulated specimens was raised by only a few degreesConsidering that the microstructure observed in simulated

23 Microstructure of Gleeble simulated specimens a lowmagnification of central region b close-up of coarse-grained region and c close-up of dual-phase regionAll specimens were electrolytically etched to reveald-ferrite

24 Charpy impact energy curve of Gleeble simulated spe-cimens and comparison with those of as-received andfresh martensitez14 d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 387

Science and Technology of Welding and Joining 2004 Vol 9 No 5

HAZ is always coarser than that found in a multipass girthweld and that the DBTT of the simulated specimens waswell below room temperature the presence of d-ferrite inHAZ of supermartensitic stainless steels is not likely toconstitute a threat to the impact properties

ACKNOWLEDGEMENTSThe authors are grateful to TWI and the EPSRC forsponsoring this research

REFERENCES1 D CARROUGE and P WOOLLIN in lsquoStainless steel world 2002rsquo

61ndash67 2002 Houston USA2 J ENERHAUG Oslash GRONG and U M STEINSMO Sci Technol Weld

Joining 2001 6 330ndash3383 T G GOOCH and B J GINN lsquoHeat affected zone toughness of

MMA welded 12Cr martensitic-ferritic steelrsquo TechnicalReport 373 Welding Institute Members Report 1988

4 C GROBLER lsquoWeldability studies on 12 and 14 chromiumsteelsrsquo PhD thesis University of Pretoria 1987

5 J J J ZAAYMAN and G T VAN ROOYEN Proc 1st Int ChromiumSteels and Alloys Congress Volume 2 SAIMM Cape TownJohannesburg 1992 137ndash142

6 K KONDO M UEDA K OGAWA H AMAYA H HIRATA andH TAKABE in lsquoSupermartensitic stainless steels 99rsquo 11ndash181999 Belgium

7 T G GOOCH P WOOLLIN and A G HAYNES in lsquoSupermartensiticstainless steels 99rsquo 188ndash195 1999 Belgium

8 P WOOLLIN and D CARROUGE in lsquoSupermartensitic stainlesssteels 2002rsquo 199ndash204 2002 Paper 027

9 D CARROUGE PhD thesis Department of MaterialsScience and Metallurgy University of Cambridge 2002Available from httpwwwmsmcamacukphasetrans2000phdhtmldominique

10 P WOOLLIN lsquoMetallurgical examination of martensitic stainlesssteel weldmentsrsquo Technical Report 620886195 WeldingInstitute Internal Report 1995

11 E BERAHA and B SHPIGLER lsquoColor metallographyrsquo 1977American Society for Metals

12 B D CULLITY lsquoElements of X-ray diffractionrsquo 1978Addison-Wesley

13 BS EN 10045-1 lsquoCharpy impact test on metallic materialsPart 1 Test methodrsquo British Standard 1990

14 D WILLIAM and J R CALLISTER lsquoMaterials science andengineering an introductionrsquo 1999 New York USA JohnWiley and Sons Inc

15 W OLDFIELD CSNI Specialists Meeting on lsquoInstrumentedprecracked Charpy testingrsquo Palo Alto CA 1981 5ndash59

25 Fracture surfaces of a coarse-grained and b dual-phase region of Gleeble simulated specimen tested at 2100uC

26 Cross-section of fracture surface shown in Fig 25 a coarse-grained and b dual-phase region This cross-section wastaken perpendicular to examination plane shown in Fig 16

388 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

d-ferrite there was little hardness change between thedifferent microstructures The latter was attributed to thevirtually carbon free martensite as a result of the presence oftitanium in concentrations greater than the stoichiometricamount As tensile strength is proportional to hardness allspecimens could be assumed to have a similar strength14

There was a small difference in grain size between the variousmicrostructures In fact the prior austenite grain size of thespecimens heat treated at 1350uC was about 20 mm largerthan the one heat treated at 1100uC As a result of the grainboundary pinning effect of d-ferrite there was no grain sizedifference between the microstructures consisting of freshmartensite and various amounts of ferrite

The presence of retained austenite in the furnace heattreated specimens has been checked using X-ray diffractionNo austenite could be detected in any of the specimens(Fig 8)

Results of Charpy impact testsTo facilitate the comparison and the determination of theductile to brittle transition temperature (DBTT) thetoughness results have been fitted by a non-linear regressiontechnique to the equation15

Impact energy (J)~A0zB0 tanhTT0

C0

(6)

where A0 B0 C0 and T0 are fitting constants and T isthe temperature in uC The DBTT was taken as thetemperature corresponding to half the height of thetransition curve Table 8 summarises the measured ductile

7 Microstructures obtained after isothermal treatment ata 1100uC and b 1350uC for 15 min and water quenchedc Microstructure obtained after treatment at 1350uC for15 min and water quenched followed by re-austenitisationat 1100uC for 60 min and water quenched

8 X-ray diffraction patterns of fresh martensite freshmartensitez14 d-ferrite and fresh martensitez2d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 381

Science and Technology of Welding and Joining 2004 Vol 9 No 5

to brittle transition temperatures for each tested micro-structure

Figure 9 shows the impact transition curves obtainedfrom the tempered martensitezcr (as-received condition)fresh martensite and fresh martensitez14 d-ferritemicrostructures The transition curves of the as-receivedand fresh martensite microstructures were similar and theirDBTT were identical and far below 0uC However the as-received microstructure showed slightly higher failureenergies at low temperatures This is because the retainedaustenite tends to transform to martensite during impactand thus absorbs some energy and raises the toughness asproposed by Bilmes et al16 At room temperature there wasmore scatter in the energy required to break the specimensbut there was apparently no significant difference betweenthe two microstructures

Similarly there were no apparent differences between theupper shelf failure energies of the fresh martensite andfresh martensitez14 d-ferrite microstructures (Fig 9)However the transition occurred over a shorter tempera-ture range in the biphased microstructure and a significantshift in ductile to brittle transition temperature wasobserved In fact the DBTT of the duplex microstructurewas raised by 52uC to about 246uC The origin of thisincrease could not be attributed directly to the presence ofd-ferrite since there was a 20 mm difference in prioraustenite grain size between the two microstructures

The observation of the transition curves of the freshmartensitez14 d-ferrite and fresh martensitez2 d-ferritemicrostructures of identical grain size removed this uncer-tainty (Fig 10) In fact specimens containing a reducedamount of d-ferrite had significantly better DBTT (286uCcompared to 246uC) As mentioned in the previous sectionafter re-austenitisation and quenching the only differencebetween the two ferrite containing microstructures was theactual level of ferrite Consequently it was clear that for aprior austenite grain size of 80 mm the presence of d-ferritesignificantly affected the transition temperature

Figure 11 compares the transition curves of the freshmartensitez2 d-ferrite (grain size 80 mm) with the temperedmartensitezcr microstructure (grain size 25 mm) The reduc-tion of the ferrite content clearly improves the notch toughnessto a level comparable to that of the as-received condition

FractographyObservation of the fracture surfaces in the SEM revealed avariety of fracture modes In particular ductile rupturecleavage and quasi-cleavage could be identified McDonaldet al summarise these fracture modes as follows1718

(i) Ductile rupture Dimples resulting from microvoidformation and coalescence are characteristic ofductile rupture Microvoids are generally initiatedat inclusions which can sometimes be observedwithin the dimples

(ii) Cleavage Stepped facets are caused by fracturethrough regions of different crystallographic orien-tation Cleavage occurs by the rapid propagation ofa crack along crystallographic planes and may beinitiated by inclusionmatrix decohesion This is avery low energy-absorbing mode of fracture

(iii) Quasi-cleavage A less well-defined fracture modeknown as quasi-cleavage is associated with tear ridgeswhich represent limited plastic deformation separatingcleavage facets Unlike pure cleavage energy isabsorbed as a result of the limited plastic deformation

Table 8 Ductile to brittle transition temperatures obtainedfor each microstructure studied

Microstructure DBTT (uC)

Tempered martensitezcr 298Fresh martensite 298Fresh martensitez14 d-ferrite 246Fresh martensitez2 d-ferrite 286

9 Charpy impact energy curves obtained on temperedmartensitezretained austenite (as-received condition)fresh martensite and fresh martensitez14 d-ferritemicrostructures

10 Charpy impact energy curves obtained on fresh mar-tensitez14 d-ferrite and fresh martensitez2 d-ferrite microstructures Note that both microstructureshave identical prior austenite grain size of 80 mm

11 Comparison of Charpy impact energy curves betweenas-received and fresh martensitez2 d-ferritemicrostructures

382 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

Examples of ductile rupture obtained from the fullymartensitic and martensiticferritic microstructures testedat 0uC are shown in Fig 12 Most cuboidal shaped titaniumcarbonitride particles were cracked

Specimens tested at low temperatures fractured in atransgranular manner (Fig 13) This indicated that thepresence of grain boundary d-ferrite did not embrittle thelatter In general the cleavage facet sizes appeared similarbut the fresh martensitez14 d-ferrite microstructureshowed more secondary cracks

River markings on the cleavage facets of the specimensshowed evidence of both cleavage (Fig 14) and quasi-cleavage fracture (Fig 15) initiated by Ti(CN) inclusionsThe capacity of these particles to reduce toughness is wellknown Brittle fracture induced by the cleavage of TiNinclusions has recently been investigated by Fairchildet al1920 Those particles are well bonded to the matrixand thus allow high stresses in a notch tip plastic zone toact on inclusions without debonding the interface Once aninclusion cleaves the strong bond allows for transfer of thecrack into the martensitic matrix

In addition to the SEM observation cross-sections per-pendicular to the fracture surfaces as shown in Fig 16 havebeen observed to reveal the fracture path Figures 17 and 18show example of such cross-sections for specimens of thefresh martensitez14 d-ferrite and fresh martensitez2d-ferrite microstructures which failed by cleavage

In the fresh martensitez14 d-ferrite specimens sec-ondary intragranular cracks propagated through marten-site and d-ferrite stringers whereas shorter cracks were seento initiate both in martensite and at the d-ferritemartensiteinterface in the fresh martensitez2 d-ferrite In the lattercase cracks could not easily link up when d-ferrite wasmore dispersed and possessed a spheroidised morphologythus explaining the better impact properties

IMPACT PROPERTIES OF GLEEBLESIMULATED SPECIMENSThe first part of this work established the effect of d-ferriteon the ductile to brittle transition temperature of homo-genous and biphased microstructures The microstructure

12 Dimpled fracture surfaces of a fresh martensite and b fresh martensitez14 d-ferrite microstructures tested at 0uC

13 Fracture surfaces of a fresh martensite and b fresh martensitez14 d-ferrite microstructures tested at 2150 and2100uC respectively

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 383

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of a weld HAZ is however far from being homogeneous andit is not known whether the presence of delta-ferrites affectthe HAZ impact properties

In supermartensitic stainless steels five HAZ regions canbe distinguished as illustrated in Fig 192

(i) partially melted region located beside the fusionboundary This is where incipient melting of thenewly formed d-ferrite occurs on heating

(ii) region which transforms completely to d-ferrite onheating with significant grain growth also calledthe coarse-grained HAZ (CG-HAZ)

(iii) high-temperature two-phase region characterisedby partial transformation of austenite to d-ferriteon heating

(iv) region which completely transforms to austeniteduring heating

(v) a low peak temperature two-phase region char-acterised by partial transformation of the tem-pered martensite to austenite during heating

The HAZ represents the area where all phase transforma-tions occur in the solid state and therefore the partiallymelted region should not be regarded as belonging to theHAZ The other regions can be artificially classified intotwo categories depending on the maximum temperaturethey experienced before cooling the high temperature andthe low temperature HAZ The high temperature HAZ

14 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedcleavage

15 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedquasi-cleavage

16 Location of transverse section in Charpy specimens

384 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

(HT-HAZ) includes the prior d-ferrite region and the hightemperature two-phase region whereas the low temperatureHAZ (LT-HAZ) refers to the remainder of the HAZ Anexample of a typical HAZ found in a single pass weldperformed on steel A is shown in Fig 20

As presented in another publication the high temperatureregion of a supermartensitic stainless steel HAZ is likely tocontain non-equilibrium d-ferrite1 High heat inputs singlepass bead on plates such as the one showed in Fig 20 yieldslow cooling rates in the CG-HAZ and favour the formationof Widmanstatten austenite during cooling from the fullyferritic region It is between the Widmanstatten austenitesubunits that ferrite is found in coarse-grained HAZ1 Indual-phase HAZ on the other hand ferrite is usually locatedat grain boundaries and as thin films within the grains Thevolume fraction of d-ferrite is dependent upon the actualwelding conditions but is usually smaller in coarse-grainedHAZ This is because fast cooling rates restrictWidmanstatten austenite formation in coarse-grained HAZbut prevent ferrite dissolution in the dual-phase HAZ9

Gleeble weld-simulatorBeing close to the weld metal sampling of all the HAZregions for Charpy impact testing is difficult To avoid this

problem a Gleeble weld-simulator has been employed TheGleeble 1500 is a thermomechanical testing device that isused in a wide range of applications including simulationand testing of weld heat affected zones A major advantageof the Gleeble is that it generates large volumes ofmicrostructure that simulate small hard to study regionsin actual weldments A range of specimen geometries can beused but by using bars Charpy specimens can directly bemachined after simulation

The machine (Fig 21) is interfaced to a computer that isreadily programmed to provide reference signals for closed-loop control of the applied thermal cycle Heating isaccomplished by the flow of low-frequency alternatingcurrent in the specimen The current distribution iscontrolled by the cross-section of the specimen and thejaws used to mount it The feedback signal necessary forclosed-loop control is obtained from a thermocouplewelded to the specimen surface21

ProcedureTen specimens of dimension 116116100 mm (takenaccording to Fig 2) were machined for HAZ simulationThe operation was performed at the maximum temperature

17 Cross-section of fracture surface of fresh martensitez14 d-ferrite microstructure tested at 2100uC Electrolyticallyetched to reveal d-ferrite

18 Cross-section of fracture surface of fresh martensitez2 d-ferrite microstructure tested at 2150uC Electrolyticallyetched to reveal d-ferrite

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 385

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of 1350uC using a cooling time between 800 and 500uC of10 s After simulation all specimens were polished andetched to reveal the underlying microstructure Standardsize Charpy specimens were finally machined with the notchlying across the centre of the simulated specimens

MicrostructureFigure 22 shows a representative macrograph of thesimulated specimens before grinding and polishing

After polishing and electro-etching the central region ofall simulated specimens showed a microstructural gradientThe latter indicated an inhomogeneous thermal experienceacross the sample sections and might have been caused byan improper contact between the samples and the grips in

the Gleeble machine As Fig 23a shows the central regionof the simulated specimens consisted of three distinct areasThe upper region had semi-circular shape and depth ofabout 25 mm and was identified as the coarse-grained region (Fig 23b) The latter was surrounded by

19 Schematic diagram illustrating structural changes that occur within heat affected zone of single-pass supermartensiticstainless steel weld2

20 Optical micrograph of single-pass bead on plate per-formed on steel A using tungsten inert gas (TIG) pro-cess with heat input of 1 kJ mm21 Plate thickness10 mm Etched electrolytically to reveal d-ferrite

21 Arrangement of equipment within Gleeble simulator

22 Macrograph of typical specimen after HAZ simula-tion using Gleeble simulator Thermocouple employedto record temperature was located at middle of centreline

386 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

the dual-phase region (Fig 23c) and in turn wassurrounded by the prior fully-austenitic area The existenceof this microstructural gradient was not expected but it wasdecided to study its impact properties since the generatedcentre line microstructure was representative of that of areal weld HAZ

The coarse-grained and dual-phase regions had grain sizeof about 160 mm and 30 mm respectively The d-ferritevolume percentage in the dual-phase region measured byimage analysis at a magnification of 200 (Fig 23c) was40iexcl5 This was larger than the volume percentagemeasured in heat treated specimens (Fig 7b and c)

Results of Charpy impact testThe transition curve obtained on the Gleeble simulatedspecimens is presented in Fig 24 For comparison the

results obtained on the as-received and martensiticferriticmicrostructures have been reported The Gleeble specimenshad the lowest upper shelf energy of the three curves Thelatter was probably because of the combined action of largegrain size and presence of d-ferrite in the tested micro-structure More importantly the simulated HAZ had aductile to brittle transition temperature comparable to thatof the as-received material (294 and 298uC respectively)although like the martensitic microstructure containing14 d-ferrite the transition occurred over a shortertemperature range The behaviour of the Gleeble simulatedspecimens at low temperatures was therefore significantlybetter than the ferriticmartensitic microstructure studied inthe first part

FractographyAll specimens tested at temperatures above 2100uC showedductile failure mode whereas those broken at temperaturesbelow 2100uC showed completely brittle behaviourInterestingly the Gleeble specimen tested at 2100uCshowed two fracture modes (Fig 25) Observation of across-section revealed the coarse-grained region to fail bycleavage whereas the dual-phase region showed evidence ofductility below the transition temperature (Fig 26) Theductility of the dual-phase HAZ indicates that the DBTT ofthis region is lower than 2100uC These results provideevidence to support the view that the presence of d-ferrite indual-phase HAZ up to 40 does not significantly affect theimpact properties of the whole high temperature HAZ Thespecimen transition temperature was more strongly influ-enced by the large grain size of the coarse-grained HAZ

SUMMARY AND CONCLUSIONThis work revealed that for a similar strength particlevolume-fraction and distribution and a given prioraustenite grain size the ductile to brittle transitiontemperature was very sensitive to the volume fraction ofd-ferrite in homogeneous microstructures On the otherhand tests carried out on simulated HAZ indicated that thedual-phase region did not play a major role in determiningthe DBTT even when d-ferrite was present up to 40 Thesmall grain size of the dual-phase region offers suitableobstacles to crack propagation Conversely the large grainsize of the coarse-grained region strongly influences theimpact properties of the high temperature HAZ Howeverwhen compared with the tempered parent material contain-ing austenite the transition temperature of the Gleeblesimulated specimens was raised by only a few degreesConsidering that the microstructure observed in simulated

23 Microstructure of Gleeble simulated specimens a lowmagnification of central region b close-up of coarse-grained region and c close-up of dual-phase regionAll specimens were electrolytically etched to reveald-ferrite

24 Charpy impact energy curve of Gleeble simulated spe-cimens and comparison with those of as-received andfresh martensitez14 d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 387

Science and Technology of Welding and Joining 2004 Vol 9 No 5

HAZ is always coarser than that found in a multipass girthweld and that the DBTT of the simulated specimens waswell below room temperature the presence of d-ferrite inHAZ of supermartensitic stainless steels is not likely toconstitute a threat to the impact properties

ACKNOWLEDGEMENTSThe authors are grateful to TWI and the EPSRC forsponsoring this research

REFERENCES1 D CARROUGE and P WOOLLIN in lsquoStainless steel world 2002rsquo

61ndash67 2002 Houston USA2 J ENERHAUG Oslash GRONG and U M STEINSMO Sci Technol Weld

Joining 2001 6 330ndash3383 T G GOOCH and B J GINN lsquoHeat affected zone toughness of

MMA welded 12Cr martensitic-ferritic steelrsquo TechnicalReport 373 Welding Institute Members Report 1988

4 C GROBLER lsquoWeldability studies on 12 and 14 chromiumsteelsrsquo PhD thesis University of Pretoria 1987

5 J J J ZAAYMAN and G T VAN ROOYEN Proc 1st Int ChromiumSteels and Alloys Congress Volume 2 SAIMM Cape TownJohannesburg 1992 137ndash142

6 K KONDO M UEDA K OGAWA H AMAYA H HIRATA andH TAKABE in lsquoSupermartensitic stainless steels 99rsquo 11ndash181999 Belgium

7 T G GOOCH P WOOLLIN and A G HAYNES in lsquoSupermartensiticstainless steels 99rsquo 188ndash195 1999 Belgium

8 P WOOLLIN and D CARROUGE in lsquoSupermartensitic stainlesssteels 2002rsquo 199ndash204 2002 Paper 027

9 D CARROUGE PhD thesis Department of MaterialsScience and Metallurgy University of Cambridge 2002Available from httpwwwmsmcamacukphasetrans2000phdhtmldominique

10 P WOOLLIN lsquoMetallurgical examination of martensitic stainlesssteel weldmentsrsquo Technical Report 620886195 WeldingInstitute Internal Report 1995

11 E BERAHA and B SHPIGLER lsquoColor metallographyrsquo 1977American Society for Metals

12 B D CULLITY lsquoElements of X-ray diffractionrsquo 1978Addison-Wesley

13 BS EN 10045-1 lsquoCharpy impact test on metallic materialsPart 1 Test methodrsquo British Standard 1990

14 D WILLIAM and J R CALLISTER lsquoMaterials science andengineering an introductionrsquo 1999 New York USA JohnWiley and Sons Inc

15 W OLDFIELD CSNI Specialists Meeting on lsquoInstrumentedprecracked Charpy testingrsquo Palo Alto CA 1981 5ndash59

25 Fracture surfaces of a coarse-grained and b dual-phase region of Gleeble simulated specimen tested at 2100uC

26 Cross-section of fracture surface shown in Fig 25 a coarse-grained and b dual-phase region This cross-section wastaken perpendicular to examination plane shown in Fig 16

388 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

to brittle transition temperatures for each tested micro-structure

Figure 9 shows the impact transition curves obtainedfrom the tempered martensitezcr (as-received condition)fresh martensite and fresh martensitez14 d-ferritemicrostructures The transition curves of the as-receivedand fresh martensite microstructures were similar and theirDBTT were identical and far below 0uC However the as-received microstructure showed slightly higher failureenergies at low temperatures This is because the retainedaustenite tends to transform to martensite during impactand thus absorbs some energy and raises the toughness asproposed by Bilmes et al16 At room temperature there wasmore scatter in the energy required to break the specimensbut there was apparently no significant difference betweenthe two microstructures

Similarly there were no apparent differences between theupper shelf failure energies of the fresh martensite andfresh martensitez14 d-ferrite microstructures (Fig 9)However the transition occurred over a shorter tempera-ture range in the biphased microstructure and a significantshift in ductile to brittle transition temperature wasobserved In fact the DBTT of the duplex microstructurewas raised by 52uC to about 246uC The origin of thisincrease could not be attributed directly to the presence ofd-ferrite since there was a 20 mm difference in prioraustenite grain size between the two microstructures

The observation of the transition curves of the freshmartensitez14 d-ferrite and fresh martensitez2 d-ferritemicrostructures of identical grain size removed this uncer-tainty (Fig 10) In fact specimens containing a reducedamount of d-ferrite had significantly better DBTT (286uCcompared to 246uC) As mentioned in the previous sectionafter re-austenitisation and quenching the only differencebetween the two ferrite containing microstructures was theactual level of ferrite Consequently it was clear that for aprior austenite grain size of 80 mm the presence of d-ferritesignificantly affected the transition temperature

Figure 11 compares the transition curves of the freshmartensitez2 d-ferrite (grain size 80 mm) with the temperedmartensitezcr microstructure (grain size 25 mm) The reduc-tion of the ferrite content clearly improves the notch toughnessto a level comparable to that of the as-received condition

FractographyObservation of the fracture surfaces in the SEM revealed avariety of fracture modes In particular ductile rupturecleavage and quasi-cleavage could be identified McDonaldet al summarise these fracture modes as follows1718

(i) Ductile rupture Dimples resulting from microvoidformation and coalescence are characteristic ofductile rupture Microvoids are generally initiatedat inclusions which can sometimes be observedwithin the dimples

(ii) Cleavage Stepped facets are caused by fracturethrough regions of different crystallographic orien-tation Cleavage occurs by the rapid propagation ofa crack along crystallographic planes and may beinitiated by inclusionmatrix decohesion This is avery low energy-absorbing mode of fracture

(iii) Quasi-cleavage A less well-defined fracture modeknown as quasi-cleavage is associated with tear ridgeswhich represent limited plastic deformation separatingcleavage facets Unlike pure cleavage energy isabsorbed as a result of the limited plastic deformation

Table 8 Ductile to brittle transition temperatures obtainedfor each microstructure studied

Microstructure DBTT (uC)

Tempered martensitezcr 298Fresh martensite 298Fresh martensitez14 d-ferrite 246Fresh martensitez2 d-ferrite 286

9 Charpy impact energy curves obtained on temperedmartensitezretained austenite (as-received condition)fresh martensite and fresh martensitez14 d-ferritemicrostructures

10 Charpy impact energy curves obtained on fresh mar-tensitez14 d-ferrite and fresh martensitez2 d-ferrite microstructures Note that both microstructureshave identical prior austenite grain size of 80 mm

11 Comparison of Charpy impact energy curves betweenas-received and fresh martensitez2 d-ferritemicrostructures

382 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

Examples of ductile rupture obtained from the fullymartensitic and martensiticferritic microstructures testedat 0uC are shown in Fig 12 Most cuboidal shaped titaniumcarbonitride particles were cracked

Specimens tested at low temperatures fractured in atransgranular manner (Fig 13) This indicated that thepresence of grain boundary d-ferrite did not embrittle thelatter In general the cleavage facet sizes appeared similarbut the fresh martensitez14 d-ferrite microstructureshowed more secondary cracks

River markings on the cleavage facets of the specimensshowed evidence of both cleavage (Fig 14) and quasi-cleavage fracture (Fig 15) initiated by Ti(CN) inclusionsThe capacity of these particles to reduce toughness is wellknown Brittle fracture induced by the cleavage of TiNinclusions has recently been investigated by Fairchildet al1920 Those particles are well bonded to the matrixand thus allow high stresses in a notch tip plastic zone toact on inclusions without debonding the interface Once aninclusion cleaves the strong bond allows for transfer of thecrack into the martensitic matrix

In addition to the SEM observation cross-sections per-pendicular to the fracture surfaces as shown in Fig 16 havebeen observed to reveal the fracture path Figures 17 and 18show example of such cross-sections for specimens of thefresh martensitez14 d-ferrite and fresh martensitez2d-ferrite microstructures which failed by cleavage

In the fresh martensitez14 d-ferrite specimens sec-ondary intragranular cracks propagated through marten-site and d-ferrite stringers whereas shorter cracks were seento initiate both in martensite and at the d-ferritemartensiteinterface in the fresh martensitez2 d-ferrite In the lattercase cracks could not easily link up when d-ferrite wasmore dispersed and possessed a spheroidised morphologythus explaining the better impact properties

IMPACT PROPERTIES OF GLEEBLESIMULATED SPECIMENSThe first part of this work established the effect of d-ferriteon the ductile to brittle transition temperature of homo-genous and biphased microstructures The microstructure

12 Dimpled fracture surfaces of a fresh martensite and b fresh martensitez14 d-ferrite microstructures tested at 0uC

13 Fracture surfaces of a fresh martensite and b fresh martensitez14 d-ferrite microstructures tested at 2150 and2100uC respectively

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 383

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of a weld HAZ is however far from being homogeneous andit is not known whether the presence of delta-ferrites affectthe HAZ impact properties

In supermartensitic stainless steels five HAZ regions canbe distinguished as illustrated in Fig 192

(i) partially melted region located beside the fusionboundary This is where incipient melting of thenewly formed d-ferrite occurs on heating

(ii) region which transforms completely to d-ferrite onheating with significant grain growth also calledthe coarse-grained HAZ (CG-HAZ)

(iii) high-temperature two-phase region characterisedby partial transformation of austenite to d-ferriteon heating

(iv) region which completely transforms to austeniteduring heating

(v) a low peak temperature two-phase region char-acterised by partial transformation of the tem-pered martensite to austenite during heating

The HAZ represents the area where all phase transforma-tions occur in the solid state and therefore the partiallymelted region should not be regarded as belonging to theHAZ The other regions can be artificially classified intotwo categories depending on the maximum temperaturethey experienced before cooling the high temperature andthe low temperature HAZ The high temperature HAZ

14 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedcleavage

15 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedquasi-cleavage

16 Location of transverse section in Charpy specimens

384 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

(HT-HAZ) includes the prior d-ferrite region and the hightemperature two-phase region whereas the low temperatureHAZ (LT-HAZ) refers to the remainder of the HAZ Anexample of a typical HAZ found in a single pass weldperformed on steel A is shown in Fig 20

As presented in another publication the high temperatureregion of a supermartensitic stainless steel HAZ is likely tocontain non-equilibrium d-ferrite1 High heat inputs singlepass bead on plates such as the one showed in Fig 20 yieldslow cooling rates in the CG-HAZ and favour the formationof Widmanstatten austenite during cooling from the fullyferritic region It is between the Widmanstatten austenitesubunits that ferrite is found in coarse-grained HAZ1 Indual-phase HAZ on the other hand ferrite is usually locatedat grain boundaries and as thin films within the grains Thevolume fraction of d-ferrite is dependent upon the actualwelding conditions but is usually smaller in coarse-grainedHAZ This is because fast cooling rates restrictWidmanstatten austenite formation in coarse-grained HAZbut prevent ferrite dissolution in the dual-phase HAZ9

Gleeble weld-simulatorBeing close to the weld metal sampling of all the HAZregions for Charpy impact testing is difficult To avoid this

problem a Gleeble weld-simulator has been employed TheGleeble 1500 is a thermomechanical testing device that isused in a wide range of applications including simulationand testing of weld heat affected zones A major advantageof the Gleeble is that it generates large volumes ofmicrostructure that simulate small hard to study regionsin actual weldments A range of specimen geometries can beused but by using bars Charpy specimens can directly bemachined after simulation

The machine (Fig 21) is interfaced to a computer that isreadily programmed to provide reference signals for closed-loop control of the applied thermal cycle Heating isaccomplished by the flow of low-frequency alternatingcurrent in the specimen The current distribution iscontrolled by the cross-section of the specimen and thejaws used to mount it The feedback signal necessary forclosed-loop control is obtained from a thermocouplewelded to the specimen surface21

ProcedureTen specimens of dimension 116116100 mm (takenaccording to Fig 2) were machined for HAZ simulationThe operation was performed at the maximum temperature

17 Cross-section of fracture surface of fresh martensitez14 d-ferrite microstructure tested at 2100uC Electrolyticallyetched to reveal d-ferrite

18 Cross-section of fracture surface of fresh martensitez2 d-ferrite microstructure tested at 2150uC Electrolyticallyetched to reveal d-ferrite

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 385

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of 1350uC using a cooling time between 800 and 500uC of10 s After simulation all specimens were polished andetched to reveal the underlying microstructure Standardsize Charpy specimens were finally machined with the notchlying across the centre of the simulated specimens

MicrostructureFigure 22 shows a representative macrograph of thesimulated specimens before grinding and polishing

After polishing and electro-etching the central region ofall simulated specimens showed a microstructural gradientThe latter indicated an inhomogeneous thermal experienceacross the sample sections and might have been caused byan improper contact between the samples and the grips in

the Gleeble machine As Fig 23a shows the central regionof the simulated specimens consisted of three distinct areasThe upper region had semi-circular shape and depth ofabout 25 mm and was identified as the coarse-grained region (Fig 23b) The latter was surrounded by

19 Schematic diagram illustrating structural changes that occur within heat affected zone of single-pass supermartensiticstainless steel weld2

20 Optical micrograph of single-pass bead on plate per-formed on steel A using tungsten inert gas (TIG) pro-cess with heat input of 1 kJ mm21 Plate thickness10 mm Etched electrolytically to reveal d-ferrite

21 Arrangement of equipment within Gleeble simulator

22 Macrograph of typical specimen after HAZ simula-tion using Gleeble simulator Thermocouple employedto record temperature was located at middle of centreline

386 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

the dual-phase region (Fig 23c) and in turn wassurrounded by the prior fully-austenitic area The existenceof this microstructural gradient was not expected but it wasdecided to study its impact properties since the generatedcentre line microstructure was representative of that of areal weld HAZ

The coarse-grained and dual-phase regions had grain sizeof about 160 mm and 30 mm respectively The d-ferritevolume percentage in the dual-phase region measured byimage analysis at a magnification of 200 (Fig 23c) was40iexcl5 This was larger than the volume percentagemeasured in heat treated specimens (Fig 7b and c)

Results of Charpy impact testThe transition curve obtained on the Gleeble simulatedspecimens is presented in Fig 24 For comparison the

results obtained on the as-received and martensiticferriticmicrostructures have been reported The Gleeble specimenshad the lowest upper shelf energy of the three curves Thelatter was probably because of the combined action of largegrain size and presence of d-ferrite in the tested micro-structure More importantly the simulated HAZ had aductile to brittle transition temperature comparable to thatof the as-received material (294 and 298uC respectively)although like the martensitic microstructure containing14 d-ferrite the transition occurred over a shortertemperature range The behaviour of the Gleeble simulatedspecimens at low temperatures was therefore significantlybetter than the ferriticmartensitic microstructure studied inthe first part

FractographyAll specimens tested at temperatures above 2100uC showedductile failure mode whereas those broken at temperaturesbelow 2100uC showed completely brittle behaviourInterestingly the Gleeble specimen tested at 2100uCshowed two fracture modes (Fig 25) Observation of across-section revealed the coarse-grained region to fail bycleavage whereas the dual-phase region showed evidence ofductility below the transition temperature (Fig 26) Theductility of the dual-phase HAZ indicates that the DBTT ofthis region is lower than 2100uC These results provideevidence to support the view that the presence of d-ferrite indual-phase HAZ up to 40 does not significantly affect theimpact properties of the whole high temperature HAZ Thespecimen transition temperature was more strongly influ-enced by the large grain size of the coarse-grained HAZ

SUMMARY AND CONCLUSIONThis work revealed that for a similar strength particlevolume-fraction and distribution and a given prioraustenite grain size the ductile to brittle transitiontemperature was very sensitive to the volume fraction ofd-ferrite in homogeneous microstructures On the otherhand tests carried out on simulated HAZ indicated that thedual-phase region did not play a major role in determiningthe DBTT even when d-ferrite was present up to 40 Thesmall grain size of the dual-phase region offers suitableobstacles to crack propagation Conversely the large grainsize of the coarse-grained region strongly influences theimpact properties of the high temperature HAZ Howeverwhen compared with the tempered parent material contain-ing austenite the transition temperature of the Gleeblesimulated specimens was raised by only a few degreesConsidering that the microstructure observed in simulated

23 Microstructure of Gleeble simulated specimens a lowmagnification of central region b close-up of coarse-grained region and c close-up of dual-phase regionAll specimens were electrolytically etched to reveald-ferrite

24 Charpy impact energy curve of Gleeble simulated spe-cimens and comparison with those of as-received andfresh martensitez14 d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 387

Science and Technology of Welding and Joining 2004 Vol 9 No 5

HAZ is always coarser than that found in a multipass girthweld and that the DBTT of the simulated specimens waswell below room temperature the presence of d-ferrite inHAZ of supermartensitic stainless steels is not likely toconstitute a threat to the impact properties

ACKNOWLEDGEMENTSThe authors are grateful to TWI and the EPSRC forsponsoring this research

REFERENCES1 D CARROUGE and P WOOLLIN in lsquoStainless steel world 2002rsquo

61ndash67 2002 Houston USA2 J ENERHAUG Oslash GRONG and U M STEINSMO Sci Technol Weld

Joining 2001 6 330ndash3383 T G GOOCH and B J GINN lsquoHeat affected zone toughness of

MMA welded 12Cr martensitic-ferritic steelrsquo TechnicalReport 373 Welding Institute Members Report 1988

4 C GROBLER lsquoWeldability studies on 12 and 14 chromiumsteelsrsquo PhD thesis University of Pretoria 1987

5 J J J ZAAYMAN and G T VAN ROOYEN Proc 1st Int ChromiumSteels and Alloys Congress Volume 2 SAIMM Cape TownJohannesburg 1992 137ndash142

6 K KONDO M UEDA K OGAWA H AMAYA H HIRATA andH TAKABE in lsquoSupermartensitic stainless steels 99rsquo 11ndash181999 Belgium

7 T G GOOCH P WOOLLIN and A G HAYNES in lsquoSupermartensiticstainless steels 99rsquo 188ndash195 1999 Belgium

8 P WOOLLIN and D CARROUGE in lsquoSupermartensitic stainlesssteels 2002rsquo 199ndash204 2002 Paper 027

9 D CARROUGE PhD thesis Department of MaterialsScience and Metallurgy University of Cambridge 2002Available from httpwwwmsmcamacukphasetrans2000phdhtmldominique

10 P WOOLLIN lsquoMetallurgical examination of martensitic stainlesssteel weldmentsrsquo Technical Report 620886195 WeldingInstitute Internal Report 1995

11 E BERAHA and B SHPIGLER lsquoColor metallographyrsquo 1977American Society for Metals

12 B D CULLITY lsquoElements of X-ray diffractionrsquo 1978Addison-Wesley

13 BS EN 10045-1 lsquoCharpy impact test on metallic materialsPart 1 Test methodrsquo British Standard 1990

14 D WILLIAM and J R CALLISTER lsquoMaterials science andengineering an introductionrsquo 1999 New York USA JohnWiley and Sons Inc

15 W OLDFIELD CSNI Specialists Meeting on lsquoInstrumentedprecracked Charpy testingrsquo Palo Alto CA 1981 5ndash59

25 Fracture surfaces of a coarse-grained and b dual-phase region of Gleeble simulated specimen tested at 2100uC

26 Cross-section of fracture surface shown in Fig 25 a coarse-grained and b dual-phase region This cross-section wastaken perpendicular to examination plane shown in Fig 16

388 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

Examples of ductile rupture obtained from the fullymartensitic and martensiticferritic microstructures testedat 0uC are shown in Fig 12 Most cuboidal shaped titaniumcarbonitride particles were cracked

Specimens tested at low temperatures fractured in atransgranular manner (Fig 13) This indicated that thepresence of grain boundary d-ferrite did not embrittle thelatter In general the cleavage facet sizes appeared similarbut the fresh martensitez14 d-ferrite microstructureshowed more secondary cracks

River markings on the cleavage facets of the specimensshowed evidence of both cleavage (Fig 14) and quasi-cleavage fracture (Fig 15) initiated by Ti(CN) inclusionsThe capacity of these particles to reduce toughness is wellknown Brittle fracture induced by the cleavage of TiNinclusions has recently been investigated by Fairchildet al1920 Those particles are well bonded to the matrixand thus allow high stresses in a notch tip plastic zone toact on inclusions without debonding the interface Once aninclusion cleaves the strong bond allows for transfer of thecrack into the martensitic matrix

In addition to the SEM observation cross-sections per-pendicular to the fracture surfaces as shown in Fig 16 havebeen observed to reveal the fracture path Figures 17 and 18show example of such cross-sections for specimens of thefresh martensitez14 d-ferrite and fresh martensitez2d-ferrite microstructures which failed by cleavage

In the fresh martensitez14 d-ferrite specimens sec-ondary intragranular cracks propagated through marten-site and d-ferrite stringers whereas shorter cracks were seento initiate both in martensite and at the d-ferritemartensiteinterface in the fresh martensitez2 d-ferrite In the lattercase cracks could not easily link up when d-ferrite wasmore dispersed and possessed a spheroidised morphologythus explaining the better impact properties

IMPACT PROPERTIES OF GLEEBLESIMULATED SPECIMENSThe first part of this work established the effect of d-ferriteon the ductile to brittle transition temperature of homo-genous and biphased microstructures The microstructure

12 Dimpled fracture surfaces of a fresh martensite and b fresh martensitez14 d-ferrite microstructures tested at 0uC

13 Fracture surfaces of a fresh martensite and b fresh martensitez14 d-ferrite microstructures tested at 2150 and2100uC respectively

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 383

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of a weld HAZ is however far from being homogeneous andit is not known whether the presence of delta-ferrites affectthe HAZ impact properties

In supermartensitic stainless steels five HAZ regions canbe distinguished as illustrated in Fig 192

(i) partially melted region located beside the fusionboundary This is where incipient melting of thenewly formed d-ferrite occurs on heating

(ii) region which transforms completely to d-ferrite onheating with significant grain growth also calledthe coarse-grained HAZ (CG-HAZ)

(iii) high-temperature two-phase region characterisedby partial transformation of austenite to d-ferriteon heating

(iv) region which completely transforms to austeniteduring heating

(v) a low peak temperature two-phase region char-acterised by partial transformation of the tem-pered martensite to austenite during heating

The HAZ represents the area where all phase transforma-tions occur in the solid state and therefore the partiallymelted region should not be regarded as belonging to theHAZ The other regions can be artificially classified intotwo categories depending on the maximum temperaturethey experienced before cooling the high temperature andthe low temperature HAZ The high temperature HAZ

14 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedcleavage

15 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedquasi-cleavage

16 Location of transverse section in Charpy specimens

384 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

(HT-HAZ) includes the prior d-ferrite region and the hightemperature two-phase region whereas the low temperatureHAZ (LT-HAZ) refers to the remainder of the HAZ Anexample of a typical HAZ found in a single pass weldperformed on steel A is shown in Fig 20

As presented in another publication the high temperatureregion of a supermartensitic stainless steel HAZ is likely tocontain non-equilibrium d-ferrite1 High heat inputs singlepass bead on plates such as the one showed in Fig 20 yieldslow cooling rates in the CG-HAZ and favour the formationof Widmanstatten austenite during cooling from the fullyferritic region It is between the Widmanstatten austenitesubunits that ferrite is found in coarse-grained HAZ1 Indual-phase HAZ on the other hand ferrite is usually locatedat grain boundaries and as thin films within the grains Thevolume fraction of d-ferrite is dependent upon the actualwelding conditions but is usually smaller in coarse-grainedHAZ This is because fast cooling rates restrictWidmanstatten austenite formation in coarse-grained HAZbut prevent ferrite dissolution in the dual-phase HAZ9

Gleeble weld-simulatorBeing close to the weld metal sampling of all the HAZregions for Charpy impact testing is difficult To avoid this

problem a Gleeble weld-simulator has been employed TheGleeble 1500 is a thermomechanical testing device that isused in a wide range of applications including simulationand testing of weld heat affected zones A major advantageof the Gleeble is that it generates large volumes ofmicrostructure that simulate small hard to study regionsin actual weldments A range of specimen geometries can beused but by using bars Charpy specimens can directly bemachined after simulation

The machine (Fig 21) is interfaced to a computer that isreadily programmed to provide reference signals for closed-loop control of the applied thermal cycle Heating isaccomplished by the flow of low-frequency alternatingcurrent in the specimen The current distribution iscontrolled by the cross-section of the specimen and thejaws used to mount it The feedback signal necessary forclosed-loop control is obtained from a thermocouplewelded to the specimen surface21

ProcedureTen specimens of dimension 116116100 mm (takenaccording to Fig 2) were machined for HAZ simulationThe operation was performed at the maximum temperature

17 Cross-section of fracture surface of fresh martensitez14 d-ferrite microstructure tested at 2100uC Electrolyticallyetched to reveal d-ferrite

18 Cross-section of fracture surface of fresh martensitez2 d-ferrite microstructure tested at 2150uC Electrolyticallyetched to reveal d-ferrite

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 385

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of 1350uC using a cooling time between 800 and 500uC of10 s After simulation all specimens were polished andetched to reveal the underlying microstructure Standardsize Charpy specimens were finally machined with the notchlying across the centre of the simulated specimens

MicrostructureFigure 22 shows a representative macrograph of thesimulated specimens before grinding and polishing

After polishing and electro-etching the central region ofall simulated specimens showed a microstructural gradientThe latter indicated an inhomogeneous thermal experienceacross the sample sections and might have been caused byan improper contact between the samples and the grips in

the Gleeble machine As Fig 23a shows the central regionof the simulated specimens consisted of three distinct areasThe upper region had semi-circular shape and depth ofabout 25 mm and was identified as the coarse-grained region (Fig 23b) The latter was surrounded by

19 Schematic diagram illustrating structural changes that occur within heat affected zone of single-pass supermartensiticstainless steel weld2

20 Optical micrograph of single-pass bead on plate per-formed on steel A using tungsten inert gas (TIG) pro-cess with heat input of 1 kJ mm21 Plate thickness10 mm Etched electrolytically to reveal d-ferrite

21 Arrangement of equipment within Gleeble simulator

22 Macrograph of typical specimen after HAZ simula-tion using Gleeble simulator Thermocouple employedto record temperature was located at middle of centreline

386 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

the dual-phase region (Fig 23c) and in turn wassurrounded by the prior fully-austenitic area The existenceof this microstructural gradient was not expected but it wasdecided to study its impact properties since the generatedcentre line microstructure was representative of that of areal weld HAZ

The coarse-grained and dual-phase regions had grain sizeof about 160 mm and 30 mm respectively The d-ferritevolume percentage in the dual-phase region measured byimage analysis at a magnification of 200 (Fig 23c) was40iexcl5 This was larger than the volume percentagemeasured in heat treated specimens (Fig 7b and c)

Results of Charpy impact testThe transition curve obtained on the Gleeble simulatedspecimens is presented in Fig 24 For comparison the

results obtained on the as-received and martensiticferriticmicrostructures have been reported The Gleeble specimenshad the lowest upper shelf energy of the three curves Thelatter was probably because of the combined action of largegrain size and presence of d-ferrite in the tested micro-structure More importantly the simulated HAZ had aductile to brittle transition temperature comparable to thatof the as-received material (294 and 298uC respectively)although like the martensitic microstructure containing14 d-ferrite the transition occurred over a shortertemperature range The behaviour of the Gleeble simulatedspecimens at low temperatures was therefore significantlybetter than the ferriticmartensitic microstructure studied inthe first part

FractographyAll specimens tested at temperatures above 2100uC showedductile failure mode whereas those broken at temperaturesbelow 2100uC showed completely brittle behaviourInterestingly the Gleeble specimen tested at 2100uCshowed two fracture modes (Fig 25) Observation of across-section revealed the coarse-grained region to fail bycleavage whereas the dual-phase region showed evidence ofductility below the transition temperature (Fig 26) Theductility of the dual-phase HAZ indicates that the DBTT ofthis region is lower than 2100uC These results provideevidence to support the view that the presence of d-ferrite indual-phase HAZ up to 40 does not significantly affect theimpact properties of the whole high temperature HAZ Thespecimen transition temperature was more strongly influ-enced by the large grain size of the coarse-grained HAZ

SUMMARY AND CONCLUSIONThis work revealed that for a similar strength particlevolume-fraction and distribution and a given prioraustenite grain size the ductile to brittle transitiontemperature was very sensitive to the volume fraction ofd-ferrite in homogeneous microstructures On the otherhand tests carried out on simulated HAZ indicated that thedual-phase region did not play a major role in determiningthe DBTT even when d-ferrite was present up to 40 Thesmall grain size of the dual-phase region offers suitableobstacles to crack propagation Conversely the large grainsize of the coarse-grained region strongly influences theimpact properties of the high temperature HAZ Howeverwhen compared with the tempered parent material contain-ing austenite the transition temperature of the Gleeblesimulated specimens was raised by only a few degreesConsidering that the microstructure observed in simulated

23 Microstructure of Gleeble simulated specimens a lowmagnification of central region b close-up of coarse-grained region and c close-up of dual-phase regionAll specimens were electrolytically etched to reveald-ferrite

24 Charpy impact energy curve of Gleeble simulated spe-cimens and comparison with those of as-received andfresh martensitez14 d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 387

Science and Technology of Welding and Joining 2004 Vol 9 No 5

HAZ is always coarser than that found in a multipass girthweld and that the DBTT of the simulated specimens waswell below room temperature the presence of d-ferrite inHAZ of supermartensitic stainless steels is not likely toconstitute a threat to the impact properties

ACKNOWLEDGEMENTSThe authors are grateful to TWI and the EPSRC forsponsoring this research

REFERENCES1 D CARROUGE and P WOOLLIN in lsquoStainless steel world 2002rsquo

61ndash67 2002 Houston USA2 J ENERHAUG Oslash GRONG and U M STEINSMO Sci Technol Weld

Joining 2001 6 330ndash3383 T G GOOCH and B J GINN lsquoHeat affected zone toughness of

MMA welded 12Cr martensitic-ferritic steelrsquo TechnicalReport 373 Welding Institute Members Report 1988

4 C GROBLER lsquoWeldability studies on 12 and 14 chromiumsteelsrsquo PhD thesis University of Pretoria 1987

5 J J J ZAAYMAN and G T VAN ROOYEN Proc 1st Int ChromiumSteels and Alloys Congress Volume 2 SAIMM Cape TownJohannesburg 1992 137ndash142

6 K KONDO M UEDA K OGAWA H AMAYA H HIRATA andH TAKABE in lsquoSupermartensitic stainless steels 99rsquo 11ndash181999 Belgium

7 T G GOOCH P WOOLLIN and A G HAYNES in lsquoSupermartensiticstainless steels 99rsquo 188ndash195 1999 Belgium

8 P WOOLLIN and D CARROUGE in lsquoSupermartensitic stainlesssteels 2002rsquo 199ndash204 2002 Paper 027

9 D CARROUGE PhD thesis Department of MaterialsScience and Metallurgy University of Cambridge 2002Available from httpwwwmsmcamacukphasetrans2000phdhtmldominique

10 P WOOLLIN lsquoMetallurgical examination of martensitic stainlesssteel weldmentsrsquo Technical Report 620886195 WeldingInstitute Internal Report 1995

11 E BERAHA and B SHPIGLER lsquoColor metallographyrsquo 1977American Society for Metals

12 B D CULLITY lsquoElements of X-ray diffractionrsquo 1978Addison-Wesley

13 BS EN 10045-1 lsquoCharpy impact test on metallic materialsPart 1 Test methodrsquo British Standard 1990

14 D WILLIAM and J R CALLISTER lsquoMaterials science andengineering an introductionrsquo 1999 New York USA JohnWiley and Sons Inc

15 W OLDFIELD CSNI Specialists Meeting on lsquoInstrumentedprecracked Charpy testingrsquo Palo Alto CA 1981 5ndash59

25 Fracture surfaces of a coarse-grained and b dual-phase region of Gleeble simulated specimen tested at 2100uC

26 Cross-section of fracture surface shown in Fig 25 a coarse-grained and b dual-phase region This cross-section wastaken perpendicular to examination plane shown in Fig 16

388 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of a weld HAZ is however far from being homogeneous andit is not known whether the presence of delta-ferrites affectthe HAZ impact properties

In supermartensitic stainless steels five HAZ regions canbe distinguished as illustrated in Fig 192

(i) partially melted region located beside the fusionboundary This is where incipient melting of thenewly formed d-ferrite occurs on heating

(ii) region which transforms completely to d-ferrite onheating with significant grain growth also calledthe coarse-grained HAZ (CG-HAZ)

(iii) high-temperature two-phase region characterisedby partial transformation of austenite to d-ferriteon heating

(iv) region which completely transforms to austeniteduring heating

(v) a low peak temperature two-phase region char-acterised by partial transformation of the tem-pered martensite to austenite during heating

The HAZ represents the area where all phase transforma-tions occur in the solid state and therefore the partiallymelted region should not be regarded as belonging to theHAZ The other regions can be artificially classified intotwo categories depending on the maximum temperaturethey experienced before cooling the high temperature andthe low temperature HAZ The high temperature HAZ

14 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedcleavage

15 Fracture surfaces of fresh martensitez14 d-ferrite microstructure tested at 2100uC showing Ti(CN) initiatedquasi-cleavage

16 Location of transverse section in Charpy specimens

384 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

(HT-HAZ) includes the prior d-ferrite region and the hightemperature two-phase region whereas the low temperatureHAZ (LT-HAZ) refers to the remainder of the HAZ Anexample of a typical HAZ found in a single pass weldperformed on steel A is shown in Fig 20

As presented in another publication the high temperatureregion of a supermartensitic stainless steel HAZ is likely tocontain non-equilibrium d-ferrite1 High heat inputs singlepass bead on plates such as the one showed in Fig 20 yieldslow cooling rates in the CG-HAZ and favour the formationof Widmanstatten austenite during cooling from the fullyferritic region It is between the Widmanstatten austenitesubunits that ferrite is found in coarse-grained HAZ1 Indual-phase HAZ on the other hand ferrite is usually locatedat grain boundaries and as thin films within the grains Thevolume fraction of d-ferrite is dependent upon the actualwelding conditions but is usually smaller in coarse-grainedHAZ This is because fast cooling rates restrictWidmanstatten austenite formation in coarse-grained HAZbut prevent ferrite dissolution in the dual-phase HAZ9

Gleeble weld-simulatorBeing close to the weld metal sampling of all the HAZregions for Charpy impact testing is difficult To avoid this

problem a Gleeble weld-simulator has been employed TheGleeble 1500 is a thermomechanical testing device that isused in a wide range of applications including simulationand testing of weld heat affected zones A major advantageof the Gleeble is that it generates large volumes ofmicrostructure that simulate small hard to study regionsin actual weldments A range of specimen geometries can beused but by using bars Charpy specimens can directly bemachined after simulation

The machine (Fig 21) is interfaced to a computer that isreadily programmed to provide reference signals for closed-loop control of the applied thermal cycle Heating isaccomplished by the flow of low-frequency alternatingcurrent in the specimen The current distribution iscontrolled by the cross-section of the specimen and thejaws used to mount it The feedback signal necessary forclosed-loop control is obtained from a thermocouplewelded to the specimen surface21

ProcedureTen specimens of dimension 116116100 mm (takenaccording to Fig 2) were machined for HAZ simulationThe operation was performed at the maximum temperature

17 Cross-section of fracture surface of fresh martensitez14 d-ferrite microstructure tested at 2100uC Electrolyticallyetched to reveal d-ferrite

18 Cross-section of fracture surface of fresh martensitez2 d-ferrite microstructure tested at 2150uC Electrolyticallyetched to reveal d-ferrite

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 385

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of 1350uC using a cooling time between 800 and 500uC of10 s After simulation all specimens were polished andetched to reveal the underlying microstructure Standardsize Charpy specimens were finally machined with the notchlying across the centre of the simulated specimens

MicrostructureFigure 22 shows a representative macrograph of thesimulated specimens before grinding and polishing

After polishing and electro-etching the central region ofall simulated specimens showed a microstructural gradientThe latter indicated an inhomogeneous thermal experienceacross the sample sections and might have been caused byan improper contact between the samples and the grips in

the Gleeble machine As Fig 23a shows the central regionof the simulated specimens consisted of three distinct areasThe upper region had semi-circular shape and depth ofabout 25 mm and was identified as the coarse-grained region (Fig 23b) The latter was surrounded by

19 Schematic diagram illustrating structural changes that occur within heat affected zone of single-pass supermartensiticstainless steel weld2

20 Optical micrograph of single-pass bead on plate per-formed on steel A using tungsten inert gas (TIG) pro-cess with heat input of 1 kJ mm21 Plate thickness10 mm Etched electrolytically to reveal d-ferrite

21 Arrangement of equipment within Gleeble simulator

22 Macrograph of typical specimen after HAZ simula-tion using Gleeble simulator Thermocouple employedto record temperature was located at middle of centreline

386 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

the dual-phase region (Fig 23c) and in turn wassurrounded by the prior fully-austenitic area The existenceof this microstructural gradient was not expected but it wasdecided to study its impact properties since the generatedcentre line microstructure was representative of that of areal weld HAZ

The coarse-grained and dual-phase regions had grain sizeof about 160 mm and 30 mm respectively The d-ferritevolume percentage in the dual-phase region measured byimage analysis at a magnification of 200 (Fig 23c) was40iexcl5 This was larger than the volume percentagemeasured in heat treated specimens (Fig 7b and c)

Results of Charpy impact testThe transition curve obtained on the Gleeble simulatedspecimens is presented in Fig 24 For comparison the

results obtained on the as-received and martensiticferriticmicrostructures have been reported The Gleeble specimenshad the lowest upper shelf energy of the three curves Thelatter was probably because of the combined action of largegrain size and presence of d-ferrite in the tested micro-structure More importantly the simulated HAZ had aductile to brittle transition temperature comparable to thatof the as-received material (294 and 298uC respectively)although like the martensitic microstructure containing14 d-ferrite the transition occurred over a shortertemperature range The behaviour of the Gleeble simulatedspecimens at low temperatures was therefore significantlybetter than the ferriticmartensitic microstructure studied inthe first part

FractographyAll specimens tested at temperatures above 2100uC showedductile failure mode whereas those broken at temperaturesbelow 2100uC showed completely brittle behaviourInterestingly the Gleeble specimen tested at 2100uCshowed two fracture modes (Fig 25) Observation of across-section revealed the coarse-grained region to fail bycleavage whereas the dual-phase region showed evidence ofductility below the transition temperature (Fig 26) Theductility of the dual-phase HAZ indicates that the DBTT ofthis region is lower than 2100uC These results provideevidence to support the view that the presence of d-ferrite indual-phase HAZ up to 40 does not significantly affect theimpact properties of the whole high temperature HAZ Thespecimen transition temperature was more strongly influ-enced by the large grain size of the coarse-grained HAZ

SUMMARY AND CONCLUSIONThis work revealed that for a similar strength particlevolume-fraction and distribution and a given prioraustenite grain size the ductile to brittle transitiontemperature was very sensitive to the volume fraction ofd-ferrite in homogeneous microstructures On the otherhand tests carried out on simulated HAZ indicated that thedual-phase region did not play a major role in determiningthe DBTT even when d-ferrite was present up to 40 Thesmall grain size of the dual-phase region offers suitableobstacles to crack propagation Conversely the large grainsize of the coarse-grained region strongly influences theimpact properties of the high temperature HAZ Howeverwhen compared with the tempered parent material contain-ing austenite the transition temperature of the Gleeblesimulated specimens was raised by only a few degreesConsidering that the microstructure observed in simulated

23 Microstructure of Gleeble simulated specimens a lowmagnification of central region b close-up of coarse-grained region and c close-up of dual-phase regionAll specimens were electrolytically etched to reveald-ferrite

24 Charpy impact energy curve of Gleeble simulated spe-cimens and comparison with those of as-received andfresh martensitez14 d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 387

Science and Technology of Welding and Joining 2004 Vol 9 No 5

HAZ is always coarser than that found in a multipass girthweld and that the DBTT of the simulated specimens waswell below room temperature the presence of d-ferrite inHAZ of supermartensitic stainless steels is not likely toconstitute a threat to the impact properties

ACKNOWLEDGEMENTSThe authors are grateful to TWI and the EPSRC forsponsoring this research

REFERENCES1 D CARROUGE and P WOOLLIN in lsquoStainless steel world 2002rsquo

61ndash67 2002 Houston USA2 J ENERHAUG Oslash GRONG and U M STEINSMO Sci Technol Weld

Joining 2001 6 330ndash3383 T G GOOCH and B J GINN lsquoHeat affected zone toughness of

MMA welded 12Cr martensitic-ferritic steelrsquo TechnicalReport 373 Welding Institute Members Report 1988

4 C GROBLER lsquoWeldability studies on 12 and 14 chromiumsteelsrsquo PhD thesis University of Pretoria 1987

5 J J J ZAAYMAN and G T VAN ROOYEN Proc 1st Int ChromiumSteels and Alloys Congress Volume 2 SAIMM Cape TownJohannesburg 1992 137ndash142

6 K KONDO M UEDA K OGAWA H AMAYA H HIRATA andH TAKABE in lsquoSupermartensitic stainless steels 99rsquo 11ndash181999 Belgium

7 T G GOOCH P WOOLLIN and A G HAYNES in lsquoSupermartensiticstainless steels 99rsquo 188ndash195 1999 Belgium

8 P WOOLLIN and D CARROUGE in lsquoSupermartensitic stainlesssteels 2002rsquo 199ndash204 2002 Paper 027

9 D CARROUGE PhD thesis Department of MaterialsScience and Metallurgy University of Cambridge 2002Available from httpwwwmsmcamacukphasetrans2000phdhtmldominique

10 P WOOLLIN lsquoMetallurgical examination of martensitic stainlesssteel weldmentsrsquo Technical Report 620886195 WeldingInstitute Internal Report 1995

11 E BERAHA and B SHPIGLER lsquoColor metallographyrsquo 1977American Society for Metals

12 B D CULLITY lsquoElements of X-ray diffractionrsquo 1978Addison-Wesley

13 BS EN 10045-1 lsquoCharpy impact test on metallic materialsPart 1 Test methodrsquo British Standard 1990

14 D WILLIAM and J R CALLISTER lsquoMaterials science andengineering an introductionrsquo 1999 New York USA JohnWiley and Sons Inc

15 W OLDFIELD CSNI Specialists Meeting on lsquoInstrumentedprecracked Charpy testingrsquo Palo Alto CA 1981 5ndash59

25 Fracture surfaces of a coarse-grained and b dual-phase region of Gleeble simulated specimen tested at 2100uC

26 Cross-section of fracture surface shown in Fig 25 a coarse-grained and b dual-phase region This cross-section wastaken perpendicular to examination plane shown in Fig 16

388 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

(HT-HAZ) includes the prior d-ferrite region and the hightemperature two-phase region whereas the low temperatureHAZ (LT-HAZ) refers to the remainder of the HAZ Anexample of a typical HAZ found in a single pass weldperformed on steel A is shown in Fig 20

As presented in another publication the high temperatureregion of a supermartensitic stainless steel HAZ is likely tocontain non-equilibrium d-ferrite1 High heat inputs singlepass bead on plates such as the one showed in Fig 20 yieldslow cooling rates in the CG-HAZ and favour the formationof Widmanstatten austenite during cooling from the fullyferritic region It is between the Widmanstatten austenitesubunits that ferrite is found in coarse-grained HAZ1 Indual-phase HAZ on the other hand ferrite is usually locatedat grain boundaries and as thin films within the grains Thevolume fraction of d-ferrite is dependent upon the actualwelding conditions but is usually smaller in coarse-grainedHAZ This is because fast cooling rates restrictWidmanstatten austenite formation in coarse-grained HAZbut prevent ferrite dissolution in the dual-phase HAZ9

Gleeble weld-simulatorBeing close to the weld metal sampling of all the HAZregions for Charpy impact testing is difficult To avoid this

problem a Gleeble weld-simulator has been employed TheGleeble 1500 is a thermomechanical testing device that isused in a wide range of applications including simulationand testing of weld heat affected zones A major advantageof the Gleeble is that it generates large volumes ofmicrostructure that simulate small hard to study regionsin actual weldments A range of specimen geometries can beused but by using bars Charpy specimens can directly bemachined after simulation

The machine (Fig 21) is interfaced to a computer that isreadily programmed to provide reference signals for closed-loop control of the applied thermal cycle Heating isaccomplished by the flow of low-frequency alternatingcurrent in the specimen The current distribution iscontrolled by the cross-section of the specimen and thejaws used to mount it The feedback signal necessary forclosed-loop control is obtained from a thermocouplewelded to the specimen surface21

ProcedureTen specimens of dimension 116116100 mm (takenaccording to Fig 2) were machined for HAZ simulationThe operation was performed at the maximum temperature

17 Cross-section of fracture surface of fresh martensitez14 d-ferrite microstructure tested at 2100uC Electrolyticallyetched to reveal d-ferrite

18 Cross-section of fracture surface of fresh martensitez2 d-ferrite microstructure tested at 2150uC Electrolyticallyetched to reveal d-ferrite

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 385

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of 1350uC using a cooling time between 800 and 500uC of10 s After simulation all specimens were polished andetched to reveal the underlying microstructure Standardsize Charpy specimens were finally machined with the notchlying across the centre of the simulated specimens

MicrostructureFigure 22 shows a representative macrograph of thesimulated specimens before grinding and polishing

After polishing and electro-etching the central region ofall simulated specimens showed a microstructural gradientThe latter indicated an inhomogeneous thermal experienceacross the sample sections and might have been caused byan improper contact between the samples and the grips in

the Gleeble machine As Fig 23a shows the central regionof the simulated specimens consisted of three distinct areasThe upper region had semi-circular shape and depth ofabout 25 mm and was identified as the coarse-grained region (Fig 23b) The latter was surrounded by

19 Schematic diagram illustrating structural changes that occur within heat affected zone of single-pass supermartensiticstainless steel weld2

20 Optical micrograph of single-pass bead on plate per-formed on steel A using tungsten inert gas (TIG) pro-cess with heat input of 1 kJ mm21 Plate thickness10 mm Etched electrolytically to reveal d-ferrite

21 Arrangement of equipment within Gleeble simulator

22 Macrograph of typical specimen after HAZ simula-tion using Gleeble simulator Thermocouple employedto record temperature was located at middle of centreline

386 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

the dual-phase region (Fig 23c) and in turn wassurrounded by the prior fully-austenitic area The existenceof this microstructural gradient was not expected but it wasdecided to study its impact properties since the generatedcentre line microstructure was representative of that of areal weld HAZ

The coarse-grained and dual-phase regions had grain sizeof about 160 mm and 30 mm respectively The d-ferritevolume percentage in the dual-phase region measured byimage analysis at a magnification of 200 (Fig 23c) was40iexcl5 This was larger than the volume percentagemeasured in heat treated specimens (Fig 7b and c)

Results of Charpy impact testThe transition curve obtained on the Gleeble simulatedspecimens is presented in Fig 24 For comparison the

results obtained on the as-received and martensiticferriticmicrostructures have been reported The Gleeble specimenshad the lowest upper shelf energy of the three curves Thelatter was probably because of the combined action of largegrain size and presence of d-ferrite in the tested micro-structure More importantly the simulated HAZ had aductile to brittle transition temperature comparable to thatof the as-received material (294 and 298uC respectively)although like the martensitic microstructure containing14 d-ferrite the transition occurred over a shortertemperature range The behaviour of the Gleeble simulatedspecimens at low temperatures was therefore significantlybetter than the ferriticmartensitic microstructure studied inthe first part

FractographyAll specimens tested at temperatures above 2100uC showedductile failure mode whereas those broken at temperaturesbelow 2100uC showed completely brittle behaviourInterestingly the Gleeble specimen tested at 2100uCshowed two fracture modes (Fig 25) Observation of across-section revealed the coarse-grained region to fail bycleavage whereas the dual-phase region showed evidence ofductility below the transition temperature (Fig 26) Theductility of the dual-phase HAZ indicates that the DBTT ofthis region is lower than 2100uC These results provideevidence to support the view that the presence of d-ferrite indual-phase HAZ up to 40 does not significantly affect theimpact properties of the whole high temperature HAZ Thespecimen transition temperature was more strongly influ-enced by the large grain size of the coarse-grained HAZ

SUMMARY AND CONCLUSIONThis work revealed that for a similar strength particlevolume-fraction and distribution and a given prioraustenite grain size the ductile to brittle transitiontemperature was very sensitive to the volume fraction ofd-ferrite in homogeneous microstructures On the otherhand tests carried out on simulated HAZ indicated that thedual-phase region did not play a major role in determiningthe DBTT even when d-ferrite was present up to 40 Thesmall grain size of the dual-phase region offers suitableobstacles to crack propagation Conversely the large grainsize of the coarse-grained region strongly influences theimpact properties of the high temperature HAZ Howeverwhen compared with the tempered parent material contain-ing austenite the transition temperature of the Gleeblesimulated specimens was raised by only a few degreesConsidering that the microstructure observed in simulated

23 Microstructure of Gleeble simulated specimens a lowmagnification of central region b close-up of coarse-grained region and c close-up of dual-phase regionAll specimens were electrolytically etched to reveald-ferrite

24 Charpy impact energy curve of Gleeble simulated spe-cimens and comparison with those of as-received andfresh martensitez14 d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 387

Science and Technology of Welding and Joining 2004 Vol 9 No 5

HAZ is always coarser than that found in a multipass girthweld and that the DBTT of the simulated specimens waswell below room temperature the presence of d-ferrite inHAZ of supermartensitic stainless steels is not likely toconstitute a threat to the impact properties

ACKNOWLEDGEMENTSThe authors are grateful to TWI and the EPSRC forsponsoring this research

REFERENCES1 D CARROUGE and P WOOLLIN in lsquoStainless steel world 2002rsquo

61ndash67 2002 Houston USA2 J ENERHAUG Oslash GRONG and U M STEINSMO Sci Technol Weld

Joining 2001 6 330ndash3383 T G GOOCH and B J GINN lsquoHeat affected zone toughness of

MMA welded 12Cr martensitic-ferritic steelrsquo TechnicalReport 373 Welding Institute Members Report 1988

4 C GROBLER lsquoWeldability studies on 12 and 14 chromiumsteelsrsquo PhD thesis University of Pretoria 1987

5 J J J ZAAYMAN and G T VAN ROOYEN Proc 1st Int ChromiumSteels and Alloys Congress Volume 2 SAIMM Cape TownJohannesburg 1992 137ndash142

6 K KONDO M UEDA K OGAWA H AMAYA H HIRATA andH TAKABE in lsquoSupermartensitic stainless steels 99rsquo 11ndash181999 Belgium

7 T G GOOCH P WOOLLIN and A G HAYNES in lsquoSupermartensiticstainless steels 99rsquo 188ndash195 1999 Belgium

8 P WOOLLIN and D CARROUGE in lsquoSupermartensitic stainlesssteels 2002rsquo 199ndash204 2002 Paper 027

9 D CARROUGE PhD thesis Department of MaterialsScience and Metallurgy University of Cambridge 2002Available from httpwwwmsmcamacukphasetrans2000phdhtmldominique

10 P WOOLLIN lsquoMetallurgical examination of martensitic stainlesssteel weldmentsrsquo Technical Report 620886195 WeldingInstitute Internal Report 1995

11 E BERAHA and B SHPIGLER lsquoColor metallographyrsquo 1977American Society for Metals

12 B D CULLITY lsquoElements of X-ray diffractionrsquo 1978Addison-Wesley

13 BS EN 10045-1 lsquoCharpy impact test on metallic materialsPart 1 Test methodrsquo British Standard 1990

14 D WILLIAM and J R CALLISTER lsquoMaterials science andengineering an introductionrsquo 1999 New York USA JohnWiley and Sons Inc

15 W OLDFIELD CSNI Specialists Meeting on lsquoInstrumentedprecracked Charpy testingrsquo Palo Alto CA 1981 5ndash59

25 Fracture surfaces of a coarse-grained and b dual-phase region of Gleeble simulated specimen tested at 2100uC

26 Cross-section of fracture surface shown in Fig 25 a coarse-grained and b dual-phase region This cross-section wastaken perpendicular to examination plane shown in Fig 16

388 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

of 1350uC using a cooling time between 800 and 500uC of10 s After simulation all specimens were polished andetched to reveal the underlying microstructure Standardsize Charpy specimens were finally machined with the notchlying across the centre of the simulated specimens

MicrostructureFigure 22 shows a representative macrograph of thesimulated specimens before grinding and polishing

After polishing and electro-etching the central region ofall simulated specimens showed a microstructural gradientThe latter indicated an inhomogeneous thermal experienceacross the sample sections and might have been caused byan improper contact between the samples and the grips in

the Gleeble machine As Fig 23a shows the central regionof the simulated specimens consisted of three distinct areasThe upper region had semi-circular shape and depth ofabout 25 mm and was identified as the coarse-grained region (Fig 23b) The latter was surrounded by

19 Schematic diagram illustrating structural changes that occur within heat affected zone of single-pass supermartensiticstainless steel weld2

20 Optical micrograph of single-pass bead on plate per-formed on steel A using tungsten inert gas (TIG) pro-cess with heat input of 1 kJ mm21 Plate thickness10 mm Etched electrolytically to reveal d-ferrite

21 Arrangement of equipment within Gleeble simulator

22 Macrograph of typical specimen after HAZ simula-tion using Gleeble simulator Thermocouple employedto record temperature was located at middle of centreline

386 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

the dual-phase region (Fig 23c) and in turn wassurrounded by the prior fully-austenitic area The existenceof this microstructural gradient was not expected but it wasdecided to study its impact properties since the generatedcentre line microstructure was representative of that of areal weld HAZ

The coarse-grained and dual-phase regions had grain sizeof about 160 mm and 30 mm respectively The d-ferritevolume percentage in the dual-phase region measured byimage analysis at a magnification of 200 (Fig 23c) was40iexcl5 This was larger than the volume percentagemeasured in heat treated specimens (Fig 7b and c)

Results of Charpy impact testThe transition curve obtained on the Gleeble simulatedspecimens is presented in Fig 24 For comparison the

results obtained on the as-received and martensiticferriticmicrostructures have been reported The Gleeble specimenshad the lowest upper shelf energy of the three curves Thelatter was probably because of the combined action of largegrain size and presence of d-ferrite in the tested micro-structure More importantly the simulated HAZ had aductile to brittle transition temperature comparable to thatof the as-received material (294 and 298uC respectively)although like the martensitic microstructure containing14 d-ferrite the transition occurred over a shortertemperature range The behaviour of the Gleeble simulatedspecimens at low temperatures was therefore significantlybetter than the ferriticmartensitic microstructure studied inthe first part

FractographyAll specimens tested at temperatures above 2100uC showedductile failure mode whereas those broken at temperaturesbelow 2100uC showed completely brittle behaviourInterestingly the Gleeble specimen tested at 2100uCshowed two fracture modes (Fig 25) Observation of across-section revealed the coarse-grained region to fail bycleavage whereas the dual-phase region showed evidence ofductility below the transition temperature (Fig 26) Theductility of the dual-phase HAZ indicates that the DBTT ofthis region is lower than 2100uC These results provideevidence to support the view that the presence of d-ferrite indual-phase HAZ up to 40 does not significantly affect theimpact properties of the whole high temperature HAZ Thespecimen transition temperature was more strongly influ-enced by the large grain size of the coarse-grained HAZ

SUMMARY AND CONCLUSIONThis work revealed that for a similar strength particlevolume-fraction and distribution and a given prioraustenite grain size the ductile to brittle transitiontemperature was very sensitive to the volume fraction ofd-ferrite in homogeneous microstructures On the otherhand tests carried out on simulated HAZ indicated that thedual-phase region did not play a major role in determiningthe DBTT even when d-ferrite was present up to 40 Thesmall grain size of the dual-phase region offers suitableobstacles to crack propagation Conversely the large grainsize of the coarse-grained region strongly influences theimpact properties of the high temperature HAZ Howeverwhen compared with the tempered parent material contain-ing austenite the transition temperature of the Gleeblesimulated specimens was raised by only a few degreesConsidering that the microstructure observed in simulated

23 Microstructure of Gleeble simulated specimens a lowmagnification of central region b close-up of coarse-grained region and c close-up of dual-phase regionAll specimens were electrolytically etched to reveald-ferrite

24 Charpy impact energy curve of Gleeble simulated spe-cimens and comparison with those of as-received andfresh martensitez14 d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 387

Science and Technology of Welding and Joining 2004 Vol 9 No 5

HAZ is always coarser than that found in a multipass girthweld and that the DBTT of the simulated specimens waswell below room temperature the presence of d-ferrite inHAZ of supermartensitic stainless steels is not likely toconstitute a threat to the impact properties

ACKNOWLEDGEMENTSThe authors are grateful to TWI and the EPSRC forsponsoring this research

REFERENCES1 D CARROUGE and P WOOLLIN in lsquoStainless steel world 2002rsquo

61ndash67 2002 Houston USA2 J ENERHAUG Oslash GRONG and U M STEINSMO Sci Technol Weld

Joining 2001 6 330ndash3383 T G GOOCH and B J GINN lsquoHeat affected zone toughness of

MMA welded 12Cr martensitic-ferritic steelrsquo TechnicalReport 373 Welding Institute Members Report 1988

4 C GROBLER lsquoWeldability studies on 12 and 14 chromiumsteelsrsquo PhD thesis University of Pretoria 1987

5 J J J ZAAYMAN and G T VAN ROOYEN Proc 1st Int ChromiumSteels and Alloys Congress Volume 2 SAIMM Cape TownJohannesburg 1992 137ndash142

6 K KONDO M UEDA K OGAWA H AMAYA H HIRATA andH TAKABE in lsquoSupermartensitic stainless steels 99rsquo 11ndash181999 Belgium

7 T G GOOCH P WOOLLIN and A G HAYNES in lsquoSupermartensiticstainless steels 99rsquo 188ndash195 1999 Belgium

8 P WOOLLIN and D CARROUGE in lsquoSupermartensitic stainlesssteels 2002rsquo 199ndash204 2002 Paper 027

9 D CARROUGE PhD thesis Department of MaterialsScience and Metallurgy University of Cambridge 2002Available from httpwwwmsmcamacukphasetrans2000phdhtmldominique

10 P WOOLLIN lsquoMetallurgical examination of martensitic stainlesssteel weldmentsrsquo Technical Report 620886195 WeldingInstitute Internal Report 1995

11 E BERAHA and B SHPIGLER lsquoColor metallographyrsquo 1977American Society for Metals

12 B D CULLITY lsquoElements of X-ray diffractionrsquo 1978Addison-Wesley

13 BS EN 10045-1 lsquoCharpy impact test on metallic materialsPart 1 Test methodrsquo British Standard 1990

14 D WILLIAM and J R CALLISTER lsquoMaterials science andengineering an introductionrsquo 1999 New York USA JohnWiley and Sons Inc

15 W OLDFIELD CSNI Specialists Meeting on lsquoInstrumentedprecracked Charpy testingrsquo Palo Alto CA 1981 5ndash59

25 Fracture surfaces of a coarse-grained and b dual-phase region of Gleeble simulated specimen tested at 2100uC

26 Cross-section of fracture surface shown in Fig 25 a coarse-grained and b dual-phase region This cross-section wastaken perpendicular to examination plane shown in Fig 16

388 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

the dual-phase region (Fig 23c) and in turn wassurrounded by the prior fully-austenitic area The existenceof this microstructural gradient was not expected but it wasdecided to study its impact properties since the generatedcentre line microstructure was representative of that of areal weld HAZ

The coarse-grained and dual-phase regions had grain sizeof about 160 mm and 30 mm respectively The d-ferritevolume percentage in the dual-phase region measured byimage analysis at a magnification of 200 (Fig 23c) was40iexcl5 This was larger than the volume percentagemeasured in heat treated specimens (Fig 7b and c)

Results of Charpy impact testThe transition curve obtained on the Gleeble simulatedspecimens is presented in Fig 24 For comparison the

results obtained on the as-received and martensiticferriticmicrostructures have been reported The Gleeble specimenshad the lowest upper shelf energy of the three curves Thelatter was probably because of the combined action of largegrain size and presence of d-ferrite in the tested micro-structure More importantly the simulated HAZ had aductile to brittle transition temperature comparable to thatof the as-received material (294 and 298uC respectively)although like the martensitic microstructure containing14 d-ferrite the transition occurred over a shortertemperature range The behaviour of the Gleeble simulatedspecimens at low temperatures was therefore significantlybetter than the ferriticmartensitic microstructure studied inthe first part

FractographyAll specimens tested at temperatures above 2100uC showedductile failure mode whereas those broken at temperaturesbelow 2100uC showed completely brittle behaviourInterestingly the Gleeble specimen tested at 2100uCshowed two fracture modes (Fig 25) Observation of across-section revealed the coarse-grained region to fail bycleavage whereas the dual-phase region showed evidence ofductility below the transition temperature (Fig 26) Theductility of the dual-phase HAZ indicates that the DBTT ofthis region is lower than 2100uC These results provideevidence to support the view that the presence of d-ferrite indual-phase HAZ up to 40 does not significantly affect theimpact properties of the whole high temperature HAZ Thespecimen transition temperature was more strongly influ-enced by the large grain size of the coarse-grained HAZ

SUMMARY AND CONCLUSIONThis work revealed that for a similar strength particlevolume-fraction and distribution and a given prioraustenite grain size the ductile to brittle transitiontemperature was very sensitive to the volume fraction ofd-ferrite in homogeneous microstructures On the otherhand tests carried out on simulated HAZ indicated that thedual-phase region did not play a major role in determiningthe DBTT even when d-ferrite was present up to 40 Thesmall grain size of the dual-phase region offers suitableobstacles to crack propagation Conversely the large grainsize of the coarse-grained region strongly influences theimpact properties of the high temperature HAZ Howeverwhen compared with the tempered parent material contain-ing austenite the transition temperature of the Gleeblesimulated specimens was raised by only a few degreesConsidering that the microstructure observed in simulated

23 Microstructure of Gleeble simulated specimens a lowmagnification of central region b close-up of coarse-grained region and c close-up of dual-phase regionAll specimens were electrolytically etched to reveald-ferrite

24 Charpy impact energy curve of Gleeble simulated spe-cimens and comparison with those of as-received andfresh martensitez14 d-ferrite microstructures

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 387

Science and Technology of Welding and Joining 2004 Vol 9 No 5

HAZ is always coarser than that found in a multipass girthweld and that the DBTT of the simulated specimens waswell below room temperature the presence of d-ferrite inHAZ of supermartensitic stainless steels is not likely toconstitute a threat to the impact properties

ACKNOWLEDGEMENTSThe authors are grateful to TWI and the EPSRC forsponsoring this research

REFERENCES1 D CARROUGE and P WOOLLIN in lsquoStainless steel world 2002rsquo

61ndash67 2002 Houston USA2 J ENERHAUG Oslash GRONG and U M STEINSMO Sci Technol Weld

Joining 2001 6 330ndash3383 T G GOOCH and B J GINN lsquoHeat affected zone toughness of

MMA welded 12Cr martensitic-ferritic steelrsquo TechnicalReport 373 Welding Institute Members Report 1988

4 C GROBLER lsquoWeldability studies on 12 and 14 chromiumsteelsrsquo PhD thesis University of Pretoria 1987

5 J J J ZAAYMAN and G T VAN ROOYEN Proc 1st Int ChromiumSteels and Alloys Congress Volume 2 SAIMM Cape TownJohannesburg 1992 137ndash142

6 K KONDO M UEDA K OGAWA H AMAYA H HIRATA andH TAKABE in lsquoSupermartensitic stainless steels 99rsquo 11ndash181999 Belgium

7 T G GOOCH P WOOLLIN and A G HAYNES in lsquoSupermartensiticstainless steels 99rsquo 188ndash195 1999 Belgium

8 P WOOLLIN and D CARROUGE in lsquoSupermartensitic stainlesssteels 2002rsquo 199ndash204 2002 Paper 027

9 D CARROUGE PhD thesis Department of MaterialsScience and Metallurgy University of Cambridge 2002Available from httpwwwmsmcamacukphasetrans2000phdhtmldominique

10 P WOOLLIN lsquoMetallurgical examination of martensitic stainlesssteel weldmentsrsquo Technical Report 620886195 WeldingInstitute Internal Report 1995

11 E BERAHA and B SHPIGLER lsquoColor metallographyrsquo 1977American Society for Metals

12 B D CULLITY lsquoElements of X-ray diffractionrsquo 1978Addison-Wesley

13 BS EN 10045-1 lsquoCharpy impact test on metallic materialsPart 1 Test methodrsquo British Standard 1990

14 D WILLIAM and J R CALLISTER lsquoMaterials science andengineering an introductionrsquo 1999 New York USA JohnWiley and Sons Inc

15 W OLDFIELD CSNI Specialists Meeting on lsquoInstrumentedprecracked Charpy testingrsquo Palo Alto CA 1981 5ndash59

25 Fracture surfaces of a coarse-grained and b dual-phase region of Gleeble simulated specimen tested at 2100uC

26 Cross-section of fracture surface shown in Fig 25 a coarse-grained and b dual-phase region This cross-section wastaken perpendicular to examination plane shown in Fig 16

388 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

HAZ is always coarser than that found in a multipass girthweld and that the DBTT of the simulated specimens waswell below room temperature the presence of d-ferrite inHAZ of supermartensitic stainless steels is not likely toconstitute a threat to the impact properties

ACKNOWLEDGEMENTSThe authors are grateful to TWI and the EPSRC forsponsoring this research

REFERENCES1 D CARROUGE and P WOOLLIN in lsquoStainless steel world 2002rsquo

61ndash67 2002 Houston USA2 J ENERHAUG Oslash GRONG and U M STEINSMO Sci Technol Weld

Joining 2001 6 330ndash3383 T G GOOCH and B J GINN lsquoHeat affected zone toughness of

MMA welded 12Cr martensitic-ferritic steelrsquo TechnicalReport 373 Welding Institute Members Report 1988

4 C GROBLER lsquoWeldability studies on 12 and 14 chromiumsteelsrsquo PhD thesis University of Pretoria 1987

5 J J J ZAAYMAN and G T VAN ROOYEN Proc 1st Int ChromiumSteels and Alloys Congress Volume 2 SAIMM Cape TownJohannesburg 1992 137ndash142

6 K KONDO M UEDA K OGAWA H AMAYA H HIRATA andH TAKABE in lsquoSupermartensitic stainless steels 99rsquo 11ndash181999 Belgium

7 T G GOOCH P WOOLLIN and A G HAYNES in lsquoSupermartensiticstainless steels 99rsquo 188ndash195 1999 Belgium

8 P WOOLLIN and D CARROUGE in lsquoSupermartensitic stainlesssteels 2002rsquo 199ndash204 2002 Paper 027

9 D CARROUGE PhD thesis Department of MaterialsScience and Metallurgy University of Cambridge 2002Available from httpwwwmsmcamacukphasetrans2000phdhtmldominique

10 P WOOLLIN lsquoMetallurgical examination of martensitic stainlesssteel weldmentsrsquo Technical Report 620886195 WeldingInstitute Internal Report 1995

11 E BERAHA and B SHPIGLER lsquoColor metallographyrsquo 1977American Society for Metals

12 B D CULLITY lsquoElements of X-ray diffractionrsquo 1978Addison-Wesley

13 BS EN 10045-1 lsquoCharpy impact test on metallic materialsPart 1 Test methodrsquo British Standard 1990

14 D WILLIAM and J R CALLISTER lsquoMaterials science andengineering an introductionrsquo 1999 New York USA JohnWiley and Sons Inc

15 W OLDFIELD CSNI Specialists Meeting on lsquoInstrumentedprecracked Charpy testingrsquo Palo Alto CA 1981 5ndash59

25 Fracture surfaces of a coarse-grained and b dual-phase region of Gleeble simulated specimen tested at 2100uC

26 Cross-section of fracture surface shown in Fig 25 a coarse-grained and b dual-phase region This cross-section wastaken perpendicular to examination plane shown in Fig 16

388 Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel

Science and Technology of Welding and Joining 2004 Vol 9 No 5

16 P D BILMES M SOLARI and C L LLORENTE MaterCharacterisation 2001 46 285ndash296

17 R D KNUTSEN and R HUTCHINGS Mater Sci Technol 1988 4127ndash135

18 M M MCDONALD and D C LUDWIGSON J Test Eval 1983 11165ndash173

19 D P FAIRCHILD D G HOWDEN and W A T CLARK MetallMater Trans A 2000 31A 641ndash652

20 D P FAIRCHILD D G HOWDEN and W A T CLARK MetallMater Trans A 2000 31A 653ndash667

21 R D CAMPBELL and D W WALSH ASM Metals Handbook 19716 603ndash613

Carrouge et al Effect of d-ferrite on impact properties of supermartensitic stainless steel 389

Science and Technology of Welding and Joining 2004 Vol 9 No 5