Mechanical Properties and Microstructural Evolutionof Simulated Heat-Affected Zones in Wrought EglinSteel

BRETT M. LEISTER, JOHN N. DUPONT, MASASHI WATANABE,and RACHEL A. ABRAHAMS

A comprehensive study was performed to correlate the mechanical properties and microstruc-tural evolution in the heat-affected zone of Eglin steel. A Gleeble 3500 thermo-mechanicalsimulator was used to simulate weld thermal cycles with different peak temperatures at a heatinput of 1500 J/mm. These samples underwent mechanical testing to determine strength andtoughness in the as-welded and post-weld heat-treated conditions. The inter-critical heat-af-fected zone (HAZ) had the lowest strength following thermal simulation, while the fine-grainand coarse-grain heat-affected zone exhibited increased strength when compared to theinter-critical HAZ. The toughness of the heat-affected zone in the as-simulated condition islower than that of the base metal in all regions of the HAZ. Post-weld heat treatments (PWHTs)increased the toughness of the HAZ, but at the expense of strength. In addition, certaincombinations of PWHTs within specific HAZ regions exhibited low toughness caused bytempered martensite embrittlement or intergranular failure. Synchrotron X-ray diffraction datahave shown that Eglin steel has retained austenite in the fine-grain HAZ in the as-simulatedcondition. In addition, alloy carbides (M23C6, M2C, M7C3) have been observed in thediffraction spectra for the fine-grain and coarse-grain HAZ following a PWHT of 973 K(700 �C)/4 hours.

DOI: 10.1007/s11661-015-3131-x� The Minerals, Metals & Materials Society and ASM International 2015

I. INTRODUCTION

EGLIN steel is an ultra-high strength steel alloy thatwas developed at Eglin Air Force Base in the early 2000sand has since been patented in 2009.[1] The steel hasstrength levels similar to AerMet100, AF1410, andHP9-4-30, but at a reduced cost due to a reduction orelimination of expensive alloying elements such as nickeland cobalt, which can both range from 10 to 14 wt pct inthe previously mentioned alloys.

Table I shows the alloy composition for Eglin steelused in this work. Silicon is added to enhance toughnessand stabilize austenite. Silicon is well known to makecementite precipitation difficult at tempering tempera-tures used for Eglin steel.[2–6] This is critical becausesimilar alloys can experience tempered martensiteembrittlement due to cementite formation at tempera-tures as low as 548 K (275 �C).[2] In order to ensure thatEglin steel does not form cementite while being tem-pered at 473 K (200 �C), the increased silicon content is

necessary. According to the patent document, chro-mium is added to increase strength and hardenability,while molybdenum is also added to increase harden-ability. Nickel is used to increase toughness, andtungsten is added to increase strength and wearresistance.[1]

Eglin steel typically has a quenched and temperedmicrostructure consisting of tempered martensite with avariety of carbide sizes and morphologies. Paules et al.[7]

have reported M3C, M6C, and MC carbides that formafter heat treatment with sizes of 180, 250, and 10 to 20nm, respectively. Two heat treatment schedules wereinvestigated. In the first heat treatment, the samples werenormalized at 1363K (1090 �C) for 1 hour, followed by asub-critical anneal at 923 K (650 �C) for 1.5 hours. Thesamples were then austenitized at 1173 K, 1223 K, and1283 K (900 �C, 950 �C, or 1010 �C) for 0.5 hours, oilquenched, and then tempered at 533 K (260 �C) for1 hour. The second heat treatment was the same as thefirst, but did not contain the normalization and sub-crit-ical anneal. M3C and MC carbides were found in allheat-treated samples, but the M6C carbides were onlypresent following the low-temperature austenitizationtreatment. Increasing the austenitization temperaturefrom 1173 K to 1283 K (900 �C to 1010 �C) caused thedissolution of the M6C carbides.Eglin steel will undergo a variety of fabrication

processes such as casting, rolling, forging, fusion weld-ing, and heat treating. Welding of Eglin steel will beespecially necessary, and a comprehensive study of

BRETT M. LEISTER, Associate, is with Exponent FailureAnalysis Associates, Menlo Park, CA, and also with the Departmentof Materials Science and Engineering, Lehigh University, Bethlehem,PA 18015. Contact e-mail: bleister@exponent.com JOHN N.DUPONT and MASASHI WATANABE, Professors, are with theDepartment of Materials Science and Engineering, Lehigh University.RACHEL ABRAHAMS, Research Scientist, is with Eglin Air ForceBase, Eglin AFB, FL 32542.

Manuscript submitted October 16, 2014.Article published online September 10, 2015

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, DECEMBER 2015—5727

heat-affected zone (HAZ) microstructure and propertieshas yet to be completed for this alloy. Thus, theobjective of this work is to measure the changes inmechanical properties in the HAZ of Eglin steel andcorrelate those changes to the evolution of microstruc-ture as a result of the weld thermal cycle. In addition, apost-weld heat treatment (PWHT) study has beenundertaken in an attempt to restore properties in theHAZ following weld thermal cycles.

II. EXPERIMENTAL PROCEDURE

The chemistry for wrought Eglin steel can be seen inTable I. The Eglin steel was received in the quenchedand tempered condition according to the following heattreatment: normalize at 1339 K (1066 �C) for 1 hour (air

cool), austenitize at 1283 K (1010 �C) for 1 hour (waterquench), and temper at 477 K (204 �C) for 4 hours (aircool).A Gleeble 3500 thermo-mechanical simulator was

used to perform the HAZ simulations. Thermal cyclesthat were used for the HAZ simulations were calculatedby Sandia’s SmartWeld (SOAR) program[8] using a heatinput of 1500 J/mm. Four peak temperatures of 948 K,1098 K, 1198 K, and 1623 K (675 �C, 825 �C, 925 �C,and 1350 �C) were used to be representative ofthe sub-critical HAZ (SCHAZ), inter-critical HAZ(ICHAZ), fine-grain HAZ (FGHAZ), and coarse-grainHAZ (CGHAZ), respectively. The peak temperatureswere chosen based upon dilatometry data collected forEglin steel at a range of heating rates. The Ac1temperature ranged from 1052 K to 1086 K (779 �C to813 �C), and the Ac3 temperature ranged from 1168 K

Table I. Composition (Weight Percent) of Eglin Steel

Alloy C Co Cr Fe Mn Mo Ni Si W

Eglin 0.26 — 2.70 Bal. 0.65 0.42 1.00 1.00 1.00

Fig. 1—Mechanical properties of as-simulated Eglin steel showing the effect of thermal cycle peak temperature and sample geometry on (a) yieldstrength, (b) tensile strength, (c) percent reduction in area, (d) charpy impact toughness.

5728—VOLUME 46A, DECEMBER 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 2—Light optical micrographs, scanning electron micrographs, and scanning electron fractographs of Eglin steel base metal and heat-affectedzones in the as-simulated condition. Arrows indicate as-quenched martensite at the prior austenite grain boundary in the ICHAZ sample. Thefracture surfaces shown are from the broken charpy specimens.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, DECEMBER 2015—5729

to 1184 K (895 �C to 911 �C). Samples used for HAZsimulations were 11 9 11 9 70 mm3 rectangular bars.

Post-weld heat treatments (PWHT) were conducted at473 K (200 �C/392 �F) and 973 K (700 �C/1292 �F) for1 and 4 hours. These temperatures were selected in orderto stay outside the temper embrittlement trough forEglin steel. Recent results by Abrahams et al. haveshown that Eglin steel exhibits a loss in impact tough-ness due to tempered martensite embrittlement betweentempering temperatures of 260 �C and approximately650 �C.[9] The PWHT temperatures of 473 K and 973 K(200 �C and 700 �C) were based on the results of thisstudy. The samples used for PWHT had previouslyundergone thermal cycling on the Gleeble 3500 using a1500 J/mm heat input. The samples were heated inambient atmosphere and allowed to air cool followingheat treatment.

Tensile testingwasperformed inaccordancewithASTME8[10] using a crosshead speed of 0.5 mm/min until yieldfollowed by 5 mm/min until fracture, using samples with agage diameter of 2.87 mm (0.113 in.) and a gage length of11.48 mm (0.452 in.). Samples were also tested with adouble-reduced gage section (DRS) and a gage length of3 mmand a gage diameter of 2.87 mm.TheDRSwas used

to ensure failure occurred within the uniform temperatureregion of the Gleeble samples. Charpy testing was per-formed at room temperature in accordance with ASTME23[11] using a standard size sample.Synchrotron X-ray diffraction characterization was

performed at the 12-BM-B beamline at the AdvancedPhoton Source of Argonne National Laboratory. Sam-ples used in the X-ray diffraction experiments were cutfrom the thermally simulated Gleeble samples with an11 9 11 mm2[2] cross section and were polished to a0.05 lm surface finish. The wavelength of the X-radia-tion used was 0.827 nm. When being examined forretained austenite, the samples were scanned from 25 to35 deg in 2h with a step size of 0.01 deg and a count timeof 3 seconds per step. The (200)a¢ and (200)c peaks arecontained within this 2h range. When being examinedfor the presence of carbides, the samples were scannedfrom 19 to 23 deg in 2h with a step size of 0.01 deg and acount time of 10 second per step.Microstructural characterization was performed on a

Reichert-Jung MeF3 inverted light optical microscope.Samples were prepared using standard metallographictechniques with a 0.05 lm colloidal silica final polish.The samples were etched first in 4 pct picral, cleaned

Fig. 3—Mechanical properties of Eglin Steel showing the effect of post-weld heat treatment on (a) yield strength, (b) tensile strength, (c) percentreduction in area, and (d) charpy impact toughness.

5730—VOLUME 46A, DECEMBER 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A

with cotton and ethanol, and further etched in 2 pctnital. Higher magnification scanning electron micro-scopy (SEM) was performed using a Hitachi 4300 FEG/SEM with a 10 kV accelerating voltage, and fractogra-phy was performed using a Phillips XL30 ESEM with a15 kV accelerating voltage.

Samples characterized by transmission electronmicroscopy (TEM) were prepared by cutting a thinsection from the thermally simulated bars using aStruers Accutom-50 high-speed abrasive cut-off sawwith an Al2O3 blade. The thin section was thenmechanically ground to a thickness of ~100 lm using600 grit SiC paper as a final step. TEM disks with 3-mmdiameter were then punched out from the sheet and

further ground to a thickness of ~50 lm to reducemagnetism effects in the microscope. The disk wasjetpolished at 19.5 V using a room-temperature solutionof 10 pct perchloric acid/90 pct acetic acid. Character-ization was performed using a JEOL 2000FX TEM witha LaB6 filament at a 200 kV accelerating voltage.

III. RESULTS

A. Mechanical Properties

The mechanical properties of the thermally simulatedheat-affected zones (HAZs) in Eglin steel are shown in

Fig. 4 Scanning electron fractographs of Eglin steel base metal in the (a) as-received condition and following PWHTs of (b) 200 �C/1 h, (c)200 �C/4 h, (d) 700 �C/1 h, and (e) 700 �C/4 h. Note: the fracture surfaces are from charpy specimens.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, DECEMBER 2015—5731

Figure 1. These results show that there is a decreasein both yield (Figure 1(a)) and tensile strength(Figure 1(b)), compared to the base metal, up to andincluding the ICHAZ, with a regain in strength in theFGHAZ and CGHAZ. There was also a change insample geometry from an ASTM standard geometrywith a 2.87-mm gage diameter to a custom double-re-duced gage section (DRS). The ductility (Figure 1(c)) ofthe SCHAZ, measured by percent reduction in area, isshown to be similar to that of the base metal, but there isa decrease in ductility in the FGHAZ and CGHAZ.There is also a uniform decrease in impact toughness

(Figure 1(d)) in all regions of the HAZ when comparedto the base metal. The base metal exhibits an impacttoughness of 28 J, and the HAZ exhibits an impacttoughness of 16–22 J at room temperature. It should benoted that all further tensile tests were conducted onsamples using the DRS geometry.

B. Optical Microscopy

Figure 2 shows the light optical micrographs, scan-ning electron micrographs, and scanning electron frac-tographs for the base metal and HAZ of the Eglin steel

Fig. 5—Scanning electron fractographs of the SCHAZ in Eglin steel in the (a) as-simulated condition, and following PWHTs of (b) 200 �C/1 h,(c) 200 �C/4 h, (d) 700 �C/1 h, and (e) 700 �C/4 h. Note: the fracture surfaces are from charpy specimens.

5732—VOLUME 46A, DECEMBER 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A

in the as-simulated condition. The fractographs weretaken from the surfaces of the fractured charpy speci-mens. The base metal exhibits a quenched and temperedmartensitic microstructure with microvoid coalescence(MVC) upon fracture. The SCHAZ exhibits a marten-sitic microstructure that has undergone further temper-ing, due to the elevated peak temperature during theweld thermal cycle, and now exhibits cleavage failure.The ICHAZ has a dual-phase microstructure, withtempered martensite present at the grain interior, andas-quenched martensite on the prior austenite grainboundary (labeled with arrows in Figure 2). The

as-quenched martensite at the prior austenite grainboundary had a hardness of 542 HV, and the temperedmartensite at the grain interior had a hardness of 450HV. The charpy sample also exhibits cleavage failure inthe as-simulated condition. The FGHAZ and CGHAZboth exhibit an as-quenched martensite microstructurewith a failure mode of MVC.

C. Post-weld Heat Treatment

Figure 3 displays the mechanical properties of thebase metal and HAZ of Eglin steel following post-weld

Fig. 6—Scanning electron fractographs of the ICHAZ in Eglin steel in the (a) as-simulated condition, and following PWHTs of (b) 200 �C/1 h,(c) 200 �C/4 h, (d) 700 �C/1 h, and (e) 700 �C/4 h. Note: the fracture surfaces are from charpy specimens.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, DECEMBER 2015—5733

heat treatments (PWHT). The PWHTs were performedon the base metal and HAZ samples that had previouslyundergone thermal simulation using a 1500 J/mm heatinput. Tensile tests (Figures 3(a) and (b)) show that aPWHT of 473 K (200 �C) for 1 or 4 hours has a minimaleffect on the strength of the HAZ. A PWHT of 973 K(700 �C) for 1 hour results in a decrease in strength inthe base metal, FGHAZ, and CGHAZ, but has no effecton the SCHAZ and ICHAZ. A PWHT of 973 K(700 �C) for 4 hours results in a significant decrease inthe strength of the base metal and all regions of theHAZ. The strength of Eglin steel following this PWHT

results in a yield strength of ~720 MPa and a tensilestrength of ~860 MPa. The ductility of Eglin steel(Figure 3(c)) generally increases with increasing timeand temperature of the PWHT. The impact toughness(Figure 3(d)) following each of the PWHTs has a variedeffect on the different regions of the HAZ. The 473 K(200 �C) PWHT has the greatest effect on the FGHAZand CGHAZ, increasing the toughness back to basemetal levels; however, it has no effect on the SCHAZand ICHAZ. The 973K (700 �C)/1 hour PWHT gener-ally has little effect on toughness, except in the case ofthe ICHAZ sample; however, it does have an effect on

Fig. 7—Scanning electron fractographs of the FGHAZ in Eglin steel in the (a) as-simulated condition, and following PWHTs of (b) 200 �C/1 h,(c) 200 �C/4 h, (d) 700 �C/1 h, and (e) 700 �C/4 h. Note: the fracture surfaces are from charpy specimens.

5734—VOLUME 46A, DECEMBER 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A

fracture mode which will be discussed later. The 973 K(700 �C)/4 hours PWHT had the greatest impact on thetoughness of the HAZ, leading to toughness in excess of60 J in the case of the ICHAZ.

D. Fractography

Figures 4 through 8 show the fractography results forthe base metal and HAZ regions of the Eglin steelsamples in both the as-simulated condition and follow-ing PWHT. Figure 4 shows the fractography from the

base metal charpy samples which exhibit MVC in theunaffected base metal and following both PWHTs at473 K (200 �C). Interlath cleavage is present followingPWHT at 973 K (700 �C)/1 hour, and intergranularfailure is seen in the base metal sample that was heattreated at 973 K (700 �C)/4 hours. Figure 5 shows thefracture surfaces of the SCHAZ which exhibits interlathcleavage failure in the as-simulated condition, followingboth PWHTs at 473 K (200 �C) and at 973 K (700 �C)/1 hour. The SCHAZ sample subjected to a PWHT of973 K (700 �C)/4 hours exhibited intergranular failure.

Fig. 8—Scanning electron fractographs of the CGHAZ in Eglin steel in the (a) as-simulated condition, and following PWHTs of (b) 200 �C/1 h,(c) 200 �C/4 h, (d) 700 �C/1 h, and (e) 700 �C/4 h. Note: the fracture surfaces are from charpy specimens.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, DECEMBER 2015—5735

The fractography of the ICHAZ is shown in Figure 6.Interlath cleavage is seen in the as-simulated conditionas well as the sample following the 473 K (200 �C)/1 hour PWHT. Following a PWHT of 473 K (200 �C)/4 hours, the ICHAZ exhibits MVC. Following a PWHTof 973 K (700 �C)/1 hour, it exhibits interlath cleavage.MVC with very fine dimples is seen in the ICHAZfollowing a PWHT of 973 K (700 �C)/4 hours. Figure 7shows the results of the fractography for the FGHAZ.MVC is present in the as-simulated condition and afterboth PWHTs at 473 K (200 �C). Interlath cleavage ispresent in the FGHAZ following a PWHT of 973 K(700 �C)/1 hour, and MVC with very fine dimples isexhibited by the FGHAZ after a PWHT of 973 K(700 �C)/4 hours. The fractography of the CGHAZ canbe seen in Figure 8. The fracture surfaces seen in theCGHAZ are the same as those in the FGHAZ, with theexception that intergranular failure is observed in theCGHAZ following a PWHT of 973 K (700 �C)/4 hours.

E. Synchrotron X-ray Diffraction

Using the previously described results, a subset ofsamples was selected for analysis using synchrotron

X-ray diffraction. The presence of retained austenite andsecond-phase carbides was examined during the scans.The base metal and the four regions of the HAZ in theas-simulated condition were chosen to gather baselinedata about Eglin steel. The ICHAZ following a 473 K(200 �C)/4 hours PWHT was chosen because it exhib-ited cleavage failure in the as-simulated condition butMVC following the PWHT. The base metal following a973 K (700 �C)/1 hour PWHT was chosen because thelight optical micrograph was similar to the SCHAZ inthe as-simulated condition, and both had lower impacttoughness compared to the base metal. Finally, both theFGHAZ and CGHAZ following a PWHT of 973 K(700 �C)/4 hours were chosen because they showed largedifferences in toughness and a difference in fracturemode (MVC and intergranular, respectively).Synchrotron X-ray diffraction spectra are shown in

Figure 9. For all synchrotron X-ray diffraction results, a¢is used to signify martensite and c is used to signifyaustenite. Figures 9(a) and (b) shows results from samplesin the as-simulated condition, and Figures 9(c) and (d)shows results from samples following PWHT.Figure 9(a)shows the spectra for the basemetal and each region of theHAZ. Each spectrum shows a (200)a¢ peak, but only the

Fig. 9—Synchrotron X-ray diffraction spectra for Eglin steel showing (a) presence of retained austenite in FGHAZ, (b) lack of carbides in sam-ples in the as-simulated condition, (c) presence of retained austenite and M23C6 carbides in PWHT samples, and (d) presence of M2C and M7C3

carbides in the FGHAZ and CGHAZ PWHT samples.

5736—VOLUME 46A, DECEMBER 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A

FGHAZexhibits a (200)cpeak.The presence of the (200)cpeak in the FGHAZ shows that retained austenite ispresent within the sample. Using the method specified inASTM E975,[12] there is approximately 4 vol pct retainedaustenite contained within the FGHAZ sample in theas-simulated condition. It should be noted that theminimum detectability limit for synchrotron X-raydiffraction is in the range of 0.1 to 2 vol pct dependingon the setup of the instrument.[13] Figure 9(b) shows thesame samples, but when scanned at a different 2h range.These spectra do not show the presence of second-phasecarbides within the samples in the as-simulated condition.The peaks that starts to appear on the far right of thespectra are the (110)a¢ peaks. The presence of retainedaustenite and carbides was also examined in PWHTsamples using the same 2h range as the as-simulatedsamples. Figure 9(c) shows that the ICHAZ following a473 K (200 �C)/4 hours PWHT exhibits a (200)c peak,

and the FGHAZ and CGHAZ following a 973 K(700 �C)/4 hours PWHT each exhibited a (511)M23C6

peak. In addition, Figure 9(d) shows that both theFGHAZ and CGHAZ samples also exhibit (101)M2C

peaks and (202)M7C3 peaks.

F. Transmission Electron Microscopy

The data gathered from the synchrotron X-raydiffraction testing allowed another subset of samplesto be selected for TEM analysis. The base metal waschosen to gather baseline data about the microstructureand presence of carbides in Eglin steel. Both theFGHAZ and CGHAZ samples in the as-simulatedcondition were selected because the fracture mode(MVC) was the same for both samples, but thesynchrotron diffraction results indicated that theFGHAZ should have retained austenite, but the

Fig. 10—Tempered martensite microstructure seen in Eglin steel base metal with (a) bright-field TEM micrograph, (b) dark-field TEM micro-graph with arrows indicating e-carbides, (c) theoretical SAED pattern for martensite and e-carbide using the Jack orientation relationship, and(d) the SAED pattern taken from the area shown in (a) with the arrow indicating the spot used to obtain the dark-field image in (b).

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, DECEMBER 2015—5737

CGHAZ should not. The FGHAZ, after a 973 K(700 �C)/1 hour PWHT, was selected because the lightoptical micrograph was similar to that seen in theSCHAZ and because there is a possibility for temperedmartensite embrittlement (TME) in both samples. TheICHAZ following a PWHT of 973 K (700 �C)/4 hourswas selected because it exhibited the highest toughnessof all the samples. The grain interior also underwent thehighest amount of tempering of all the samples tested.Finally, the CGHAZ following a PWHT of 973 K(700 �C)/4 hours was selected because it showed adecreased toughness compared to other samples giventhe same PWHT and also exhibited intergranularfailure. Synchrotron results also indicate that M2C andM7C3 carbides should be present.

In the TEM results, martensite is symbolized by a¢,austenite is symbolized by c, and cementite is symbol-ized by h. Figure 10 shows transmission electron micro-graphs and selected area electron diffraction (SAED)patterns for Eglin steel base metal. The bright-fieldTEM micrograph (Figure 10(a)) shows a tempered lathmartensite microstructure with second-phase particleswithin the laths. The dark-field TEM micrograph(Figure 10(b)) shows those second-phase particles aswhite spots (further indicated by arrows). The SAEDpattern for the area shown in Figure 10(a) is shown inFigure 10(d) with a theoretical SAED patterns for a¢and e shown in Figure 10(c). Using the theoreticaldiffraction patterns, simulated using WebEMAPScodes,[14,15] the secondary spots (i.e., those not createdby diffraction from the martensite matrix) were deter-mined to be from e-carbides. Thus, the second-phaseparticles in the lath martensite microstructure aree-carbides, which obey the Jack orientation relation-ship.[16] This is also confirmed with the dark-field TEM

micrograph, which was captured using the diffractionspot from e-carbide indicated with an arrow in theSAED pattern.TEM analysis of the FGHAZ and CGHAZ in the

as-simulated condition is shown in Figure 11 (FGHAZ),Figure 12, and Figure 13 (CGHAZ). The FGHAZexhibits an as-quenched lath martensite microstructureas seen in the bright-field TEM micrograph inFigure 11(a). The dark-field TEM micrograph, shownin Figure 11(b), from the same area as Figure 11(a)shows the presence of retained austenite at the lathboundaries of the martensite, and the arrow shown inthe SAED pattern in Figure 11(c) indicates the areafrom where the dark-field TEM micrograph wasobtained. There is no diffraction spot for the retainedaustenite because there is such a small amount ofaustenite present. However, a dark-field TEM image canstill be obtained by knowing where the diffraction spotfor austenite is located, even if it is not visible. TheCGHAZ in the as-simulated condition shows that thereis a degree of autotempering occurring during the weldthermal cycle. Figures 12(a) and (b) shows bright-fieldTEM micrographs of an autotempered lath of marten-site with second-phase particles within the lath. Com-parison of the theoretical and actual SAED patterns(Figures 12(c) and (d), respectively) confirms thatthe second-phase particles present in the CGHAZmicrostructure are e-carbides which obey the Jackorientation relationship. Further, Figure 13 shows aseparate area in the CGHAZ which exhibits a smallamount of interlath retained austenite. Figure 13(a)shows the bright-field TEM micrograph withFigure 13(b) showing the dark-field TEM micrographof the same region. The arrow in the SAED pattern(Figure 13(c)) shows the area from the diffraction

Fig. 11—Virgin martensite microstructure seen in the FGHAZ in the as-simulated condition with (a) bright-field TEM micrograph, (b) dark-fieldTEM micrograph with retained austenite showing as bright white films at the lath boundaries, and (c) the SAED pattern from the area in (a)and the arrow indicating the area used to obtain the dark-field image in (b).

5738—VOLUME 46A, DECEMBER 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A

pattern from where the dark-field TEM micrograph wasobtained.

Figure 14 shows the microstructure of the FGHAZfollowing a PWHT of 973 K (700 �C)/1 hour. Thebright-field TEM micrograph is shown in Figure 14(a)and shows a tempered martensite microstructure withsecond-phase particles present. Obtaining a SAEDpattern (Figure 14(d)) and comparing with a theoreticalSAED pattern (Figure 14(c)) using the Bagaryatskiorientation relationship[17] between martensite andcementite confirm the presence of cementite within thissample. The spots in the SAED pattern in Figure 14(d)that are not accounted for in the theoretical SAEDpattern are formed by the surrounding laths of themartensite matrix. Using the diffraction spot indicated

by an arrow in the SAED pattern, the dark-field TEMmicrograph (Figure 14(b)) was acquired, showing thepresence of cementite at the interlath region of thetempered martensite microstructure.Figure 15 shows the ICHAZ following a PWHT of

973 K (700 �C)/4 hours. The ICHAZ has a temperedmartensite microstructure with cuboidal second-phase particles, and the bright-field TEM micrograph(Figure 15(a)) shows that the lath nature of themicrostructure is starting to degrade. Figure 15(c) showsthe theoretical SAED pattern for martensite and M23C6

carbide using the Kurdjumov–Sachs orientation rela-tionship.[18] Comparing this theoretical SAED patternwith the actual SAED pattern obtained from the sampleconfirms that the cuboidal second-phase particles are

Fig. 12—Autotempered martensite microstructure seen in the CGHAZ in the as-simulated condition with (a) bright-field TEM micrographshowing autotempered martensite lath, (b) higher magnification image showing e-carbides within the lath, (c) theoretical SAED pattern formartensite and e-carbide using the Jack orientation relationship, and (d) the SAED pattern taken from the area shown in (a).

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, DECEMBER 2015—5739

M23C6 carbides. Further confirmation is gathered whena dark-field TEM micrograph is obtained (Figure 15(b))using the spot indicated by the arrow in Figure 15(d).

Analysis of the CGHAZ following a PWHT of 973 K(700 �C)/4 hours provided a variety of information.Figure 16 shows a bright-field TEM micrograph with aprior austenite grain boundary (PAGB) in the middle ofthe image running from the top to the bottom. There is asemi-continuous line of second-phase particles thatdecorate the boundary. In addition, the grains on eitherside of the boundary exhibit a highly tempered marten-sitic microstructure. A higher magnification image of theboundary is shown in Figures 17(a) and 18(a). There aretwo second-phase particles at the PAGB, each with adistinct morphology. Figure 17 provides analysis of theelongated particle on the left (indicated by the arrow inFigure 17(a)). Energy dispersive X-ray spectrometry(EDS) analysis was performed on this particle, and itis shown that the particle is enriched in W and Mo.Matrix contributions were reduced as much as possibledue to the formation of a hole around the particleduring jetpolishing. A SAED pattern was also obtained(Figure 17(d)) and compared to the theoretical SAEDpattern (Figure 17(c)) for martensite and cementiteusing the Bagaryatski orientation relationship. Combin-ing all of this information, it is shown that the elongatedparticle at the PAGB in the CGHAZ sample is cementitewhich is enriched in W and Mo. The cuboidal particle,on the right of the micrograph in Figure 18(a) (indicatedby an arrow), shows a different EDS spectrum(Figure 18(b)). The cuboidal particle is shown to beenriched in Cr with no appreciable enrichment in W orMo. The SAED pattern is slightly more complex for thisparticle, as it has contributions from both thegrain above [labeled with (1)] and the grain below

[labeled with (2)]. Comparing the theoretical SAEDpattern (Figure 18(c)) and the actual SAED pattern(Figure 18(d)), the cuboidal second-phase particle isshown to be M23C6 carbide.

IV. DISCUSSION

As previously stated, there was a change in tensilegeometry from an ASTM standard geometry to acustom DRS geometry. This change was prompted bymicrostructural investigation and the physical nature ofhow the Gleeble heats the specimens. When the brokentensile samples with the ASTM standard geometry wereexamined, it was found that all of the samples, includingthose that underwent thermal cycling representative ofthe FGHAZ and CGHAZ, broke within a microstruc-tural region similar to that of the SCHAZ and ICHAZ.Thus, a change in the geometry of the tensile specimenswas made to more accurately test the strength of thevarious regions of the HAZ. When samples are heated inthe Gleeble, a temperature gradient exists along thelongitudinal axis, which can result in a microstructuralgradient. As a result, areas exist outside of the center 2to 3 mm that have undergone thermal cycles similar tothe SCHAZ and ICHAZ. However, the center 2 to3 mm is within a relatively uniform ‘hot zone’ where thetemperature difference is approximately 60 �C. It shouldbe noted that all samples that underwent mechanicaltesting were visually examined to ensure that the failureoccurred within the proper microstructural region. Thefailures that occurred outside the proper microstructuralregion using the ASTM geometry are also the reason forthe lack of strength data on the plots in Figures 1(a)through (c) for the FGHAZ and CGHAZ.

Fig. 13—Autotempered martensite microstructure seen in the CGHAZ in the as-simulated condition with (a) bright-field TEM micrograph, (b)dark-field TEM micrograph with retained austenite showing as bright white spots at the lath boundaries, and (c) the SAED pattern from thearea in (a) and the arrow indicating the area used to obtain the dark-field image in (b).

5740—VOLUME 46A, DECEMBER 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A

The decrease in both yield and tensile strength(Figure 1) in the SCHAZ and ICHAZ is a result oftempering of the martensitic microstructure present inthe base metal. The regain in strength in the FGHAZand CGHAZ is due to the formation of virgin marten-site (Figure 2) throughout the entire microstructure.

The decrease in toughness observed in both theFGHAZ and CGHAZ is a result of the as-quenchedmartensite that forms following the thermal cycle. Bothsamples were heated above the Ac3 temperature, result-ing in full austenitization upon heating and the forma-tion of martensite upon cooling. The base metal heattreatment included a 4 hour temper at 473 K (200 �C),but no such temper occurred for each of these regions ofthe HAZ. The CGHAZ has a slightly higher toughness

than the FGHAZ due to a small amount of autotem-pering that occurred during the thermal cycle. This isevidenced by the presence of e-carbide within some ofthe martensitic laths in the CGHAZ (Figure 12). It hasbeen shown that the prior austenite grain size plays arole in the martensite start temperature of steel alloys.An increase in the austenite grain size leads to anincrease in the martensite start temperature.[19] Thus, inthe CGHAZ, martensite is formed at a higher temper-ature, leading to a higher degree of autotemperingcompared to the FGHAZ.The various PWHTs had different microstructural

effects on each region of the HAZ (Figure 3). When thesamples were tempered at 200 �C, there was no signif-icant impact on strength, as this tempering temperature

Fig. 14—Tempered martensite microstructure seen in the FGHAZ following a 700 �C/1 h heat treatment with (a) bright-field TEM micrographshowing tempered martensite laths with cementite, (b) dark-field TEM micrograph showing the cementite decorating the interlath regions, (c)theoretical SAED pattern for martensite and cementite using the Bagaryatski orientation relationship, and (d) the SAED pattern taken from thearea shown in (a) and the arrow indicating the spot used to obtain the dark-field micrograph shown in (b).

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, DECEMBER 2015—5741

was too low to alter the microstructure significantly. Ifthe matrix is as-quenched martensite, as is the case in theFGHAZ and CGHAZ, then tempering at 473 K(200 �C) will only cause the formation of e-carbide. Ifthe matrix has been tempered by the weld thermal cycle,as is the case for the SCHAZ and ICHAZ, and theformation of cementite has already occurred, then the473 K (200 �C) PWHT will have little effect on thematrix or the cementite present. The 473 K (200 �C)PWHT has the greatest benefit on the toughness of theFGHAZ and CGHAZ due to the tempering of theas-quenched martensite microstructure. The full thermalhistory of each of these HAZ regions is similar to that ofthe base metal, where an as-quenched martensiticmicrostructure was formed first, followed by a 473 K(200 �C)/4 hours temper that was employed. It can beseen that the toughness of the FGHAZ and CGHAZ,

following this PWHT, has returned to base metaltoughness levels.The PWHT of 973 K (700 �C)/1 hour results in

uniform strength (Figures 3(a) and (b)) in the basemetal and the four HAZ regions. The amount of time atthis elevated temperature is much longer (1 hour)compared to the thermal cycles that each of thesesamples had undergone previously (~2 minutes). Thus,the total time at elevated temperature for each of thesamples that have undergone this PWHT is essentiallyidentical. This leads to similar tempering of the marten-site matrix and uniform strength levels through theHAZ. The toughness of the FGHAZ following thisPWHT is still very low, and following the TEMinvestigation, it was observed that the FGHAZ hasundergone TME. This is evidenced by the decreasedtoughness, interlath cleavage failure, and the presence of

Fig. 15—Tempered martensite microstructure seen in the ICHAZ following a 700 �C/4 h heat treatment with (a) bright-field TEM micrographshowing tempered martensite laths with M23C6 carbides, (b) dark-field TEM micrograph showing the M23C6 evenly distributed throughout themicrostructure, (c) theoretical SAED pattern for martensite and M23C6 using the Kurdjumov–Sachs orientation relationship, and (d) the SAEDpattern taken from the area shown in (a) and the arrow indicating the spot used to obtain the dark-field micrograph shown in (b).

5742—VOLUME 46A, DECEMBER 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A

cementite at the interlath boundary (Figure 14).[2,20]

These results could most likely be extended to explainthe low toughness in the CGHAZ, as well as the basemetal and SCHAZ following the same heat treatment.All three of these regions exhibit low toughness andinterlath cleavage failure, which are the two keycomponents when identifying TME. While the peaktemperatures of these PWHT can be high enough tocause the formation of alloy carbides (those with Cr,Mo, and W), the time at temperature seems too low.Fe-rich carbides form much easier because C diffusesinterstitially, where Cr, Mo, and W (alloy carbideformers) diffuse substitutionally.[21]

The PWHT of 973 K (700 �C)/4 hours had thegreatest benefit to toughness in the HAZ, but thistoughness increase came at the expense of strength. Thepeak toughness of Eglin steel occurred in the ICHAZfollowing a PWHT of 973 K (700 �C)/4 hours. TheTEM investigation showed that there was a highlytempered martensite matrix with M23C6 present(Figure 15). Neither the high degree of tempering, northe presence of M23C6 are detrimental to toughness.However, the toughness of the CGHAZ following thesame PWHT is more than 50 pct lower than that seen inthe ICHAZ, and the fracture mode in the CGHAZ isprimarily intergranular failure, compared to MVCwhich is found in the ICHAZ. The formation ofcementite[22] and the possible rejection of impurityelements from the cementite to the PAGB[23] are themain causes of the decrease in toughness and change infracture mode in the CGHAZ following this PWHT(Figures 16 and 17). The presence of cementite at thePAGB leads to the formation of small cavities aroundthe precipitates. The eventual growth and linkup ofthese cavities lead to the growth of a crack along theboundary.[22] Cementite should be completely dissolved

during a PWHT of 973 K (700 �C)/4 hours; however, itis possible for the cementite to re-precipitate duringcooling if the cooling rate is sufficiently slow.[22,23] TheFGHAZ following the same PWHT does not exhibitdecreased toughness, nor does it exhibit intergranularfailure. Assuming that the same volume fraction ofcementite forms in both the FGHAZ and CGHAZduring slow cooling, the CGHAZ would allow for morecomplete grain boundary coverage because of thedecreased area of grain boundary present within thesample.[18,19] It is not solely the presence of cementite atthe PAGB that leads to IG failure. As stated previously,IG failure is typically caused by the segregation ofimpurities to the PAGB; however, the likelihood ofembrittlement is caused by the interaction between theseimpurities and the alloying elements, not necessarily theabsolute level of impurities within the materials.[24–26]

For instance, silicon (which is added to Eglin steel todrive cementite precipitation to higher temperatures)increases the tendency for embrittlement even though itis not enriched at the PAGB itself. This is because siliconreduces the bulk solubility of impurity elements withoutallowing them to precipitate. This results in theincreased segregation of impurities at the PAGB.[23] Itis likely that the base metal and SCHAZ, following thesame PWHT, have cementite present at the prioraustenite grain boundaries as well. This is supportedby the low toughness (similar to that of the CGHAZ)and the presence of IG failure.Although TEM analysis was not performed on the

SCHAZ and ICHAZ samples in the as-simulatedcondition, the reason for differences in toughnesscompared to the base metal can be inferred from otherevidence. It is highly likely that the e-carbide present inthe base metal transformed to cementite as a result ofthe elevated peak temperatures caused by the SCHAZand ICHAZ thermal cycles. As stated previously, thepeak temperature of the thermal cycle for the SCHAZand ICHAZ should be high enough to start to formalloy carbides (Cr, Mo, and W). However, the total timeabove 573 K (300 �C) for these thermal cycles is~30 seconds (Figure 1), which does not allow enoughtime for the formation of alloy carbides.[21] The decreasein toughness seen in both the SCHAZ and ICHAZ(Figure 1) is most likely due to tempered martensiteembrittlement (TME). The fracture mode of each ofthese areas shows interlath cleavage compared to theMVC fracture mode seen in the base metal. The elevatedpeak temperature of the thermal cycle most likely causedthe e-carbide present in the base metal to transformationto cementite at the interlath regions, resulting incleavage failure and decreased toughness, similar tothat seen by Horn and Ritchie.[2] In addition, thedual-phase microstructure observed in the ICHAZ(Figure 2) contributed to this lower toughness as well.The peak temperature of 1098 K (825 �C) fell betweenthe Ac1 and Ac3 temperature for Eglin steel, resulting inpartial austenitization at the PAGB and formation ofas-quenched martensite upon cooling. The TME presentat the grain interior and the as-quenched martensite atthe PAGB lead to the decreased toughness in theICHAZ.

Fig. 16—TEM micrograph of the CGHAZ following a 700 �C/4 hPWHT showing second-phase particles decorating the prior austenitegrain boundary.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, DECEMBER 2015—5743

Results of the synchrotron X-ray diffraction studyshow the presence of a c peak for the FGHAZ sample inthe as-simulated condition, but not for theCGHAZ in thesame condition (Figure 9(a)), even though the TEMresults indicate otherwise. The minimum detectabilitylimit of synchrotron X-ray diffraction is 0.1 vol pct, butcan be as high as 2 vol pct based upon the setup of theequipment.[13] The beamline setup would require furthertuning to resolve smaller volume fractions of c in theCGHAZ. The peak seen at a 2h value of 26.4 deg in theFGHAZ and CGHAZ spectra after a PWHT of 973 K(700 �C)/4 hours does not signify the presence of c as itdoes in the other spectra that show a peak in this sameposition (Figure 9(c)). The PWHT temperature of 973 K(700 �C) is high enough that it would completely elimi-nate any retained austenite in analloy of this composition.At elevated temperatures [as low as 473 K (200 �C)],carbon starts to diffuse from the austenite. The reductionin carbon contentwithin the austenite causes the austeniteto becomeunstable anddecompose.[2,3,20,27,28] The peak isactually from the (511)M23C6, and the presence ofM23C6 is

confirmed from the TEManalysis on theCGHAZ samplefollowing a PWHT of 973 K (700�C)/4 hours. Thespectra for both the FGHAZ and CGHAZ followingthe same PWHT (Figure 9(d)) exhibit peaks for both(101)M2C and (202)M7C3, but the results of the TEManalysis do not support this result. This is most likelycaused by a difference in sampling volume, as the volumeanalyzed by the X-ray beam was much larger than thevolume investigated during the TEM study.A PWHT of 973 K (700 �C) for 1 hour leads to a

decrease in strength in all regions of the HAZ and hasbeen shown to be detrimental to toughness in someregions because of tempered martensite embrittlement.A PWHT of 973 K (700 �C)/4hours had a large increasein toughness in the ICHAZ and FGHAZ, but resulted ina dramatic decrease in strength in all areas. In addition,intergranular fracture was seen in the base metal,SCHAZ, and CGHAZ, indicating that brittle fractureis occurring. Among the post-weld heat treatmentsconsidered, the treatment at 473 K (200 �C) for 4 hoursappears to present the best compromise between

Fig. 17—Tempered martensite microstructure seen in the CGHAZ following a 700 �C/4 h heat treatment with (a) bright-field TEM micrographshowing carbides at the prior austenite grain boundary, (b) EDS spectrum of the carbide labeled by an arrow in (a) showing a enrichment in Wand Mo, (c) theoretical SAED pattern for martensite and cementite using the Bagaryatski orientation relationship and (d) the SAED pattern ta-ken from the area shown in (a). The excess spots shown in (d) are due to the multiple orientations of the cementite particle.

5744—VOLUME 46A, DECEMBER 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A

strength and toughness in all regions of the HAZ. Thestrength of the FGHAZ and CGHAZ met or exceededthe strength of the base metal (within experimentalerror), and the toughness of the FGHAZ and CGHAZwas returned to base metal levels after undergoing thisPWHT. The yield strength of the SCHAZ (1186 MPa) isslightly below that of the base metal (1365 MPa), andthe yield strength of the ICHAZ (1062 MPa) is certainlybelow that of the base metal. However, the toughness ofthe ICHAZ is near that of the base metal. It should alsobe noted that, with the exception of the SCHAZ, allregions of the HAZ and the base metal exhibited MVCafter a 473 K (200 �C)/4 hour PWHT. Although it ispossible to consider other temperatures between 200 and700 �C, recent results presented by Abrahams et al.[9]

have shown that Eglin steel exhibits a loss in impacttoughness due to tempered martensite embrittlementbetween tempering temperatures of 260 �C and approx-imately 650 �C. Thus, use of PWHT temperatureswithin this range would be expected to have a detri-mental effect on the impact toughness to the base metal.

V. CONCLUSIONS

A comprehensive study has been conducted to corre-late the mechanical properties and microstructuralevolution in simulated heat-affected zones in Eglin steel.In addition, a post-weld heat treatment study has beenperformed in an attempt to restore properties to theHAZ following weld thermal cycles. The followingconclusions can be made:

(1) The strength decrease, compared to the basemetal, in the SCHAZ and ICHAZ is associatedwith tempering of the martensite matrix. Theregain of strength in the FGHAZ and CGHAZ isdue to the formation of as-quenched martensite.

(2) The toughness decrease in the SCHAZ andICHAZ is most likely due to the formation ofcementite at the lath boundaries caused by theelevated peak temperature of the thermal cycles.The interlath cleavage failure and decreasedtoughness lends support to the conclusion thatthese regions have undergone tempered marten-

Fig. 18—Tempered martensite microstructure seen in the CGHAZ following a 700 �C/4 h heat treatment with (a) bright-field TEM micrographshowing carbides at the prior austenite grain boundary, (b) EDS spectrum of the carbide labeled by an arrow in (a) showing a enrichment in Cr,(c) theoretical SAED pattern for martensite and M23C6 using the Kurdjumov–Sachs orientation relationship, and (d) the SAED pattern takenfrom the area shown in (a). Note: The two grains on each side of the carbide are labeled as (1) and (2) in (a) and (c).

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, DECEMBER 2015—5745

site embrittlement. The decrease in toughness inthe FGHAZ and CGHAZ is due to the formationof as-quenched martensite following the thermalcycle.

(3) Retained austenite is present in both the FGHAZand CGHAZ samples, although the amountpresent in the FGHAZ is higher than that in theCGHAZ. The 473 K (200 �C) PWHT had littleeffect on the strength of the base metal and HAZ,but increased the toughness in the FGHAZ andCGHAZ back to base metal levels.

(4) The 973 K (700 �C)/1 hour PWHT lead to auniform strength level in the base metal and allregions of the HAZ. The presence of cementite atthe lath boundaries and a fracture mode ofinterlath cleavage support the conclusion thattempered martensite embrittlement occurred inthe FGHAZ following this PWHT.

(5) The 973 K (700 �C)/4 hours PWHT lead to thehighest toughness, but at the expense of strength.The ICHAZ had a toughness of 64 J, with a failuremode of microvoid coalescence. This increasedtoughness was caused by a heavily temperedmatrixand the formation of M23C6 which is not detri-mental to toughness. The CGHAZ had a lowtoughness of 20 J, with a failure mode of inter-granular fracture. The decreased toughness wascaused by the formation of cementite at the prioraustenite grain boundary.

(6) Among the post-weld heat treatments considered,the treatment at 473 K (200 �C) for 4 hoursappears to present the best compromise betweenstrength and toughness in all regions of the HAZ.The strength of the FGHAZ and CGHAZ met orexceeded the strength of the base metal (withinexperimental error), and the toughness of theFGHAZ and CGHAZ was returned to base metallevels after undergoing this PWHT.

ACKNOWLEDGMENTS

This work was performed with funding from DOTCinitiative, DOTC-09-01-INIT120. Synchrotron X-raydiffraction work was performed using the AdvancedPhoton Source, an Office of Science User Facility

operated for the U.S. Department of Energy (DOE)Office of Science by Argonne National Laboratory,and was supported by the U.S. DOE under ContractNo. DE-AC02-06CH11357.

REFERENCES1. M. Dilmore and J.D. Ruhlman: United States Patent: 7537727.2. R.M. Horn and R.O. Ritchie: Metall. Mater. Trans., 1978,

vol. 9A, pp. 1039–53.3. G. Speich and W. Leslie: Metall. Mater. Trans. B, 1972, vol. 3,

pp. 1043–54.4. J. Youngblood and M. Raghavan: Metall. Mater. Trans. A, 1977,

vol. 8A, pp. 1439–48.5. E. Kozeschnik and H.K.D.H. Bhadeshia: Mater. Sci. Technol.,

2008, vol. 24, pp. 343–47.6. D. Delagnes, P. Lamesle, M.H. Mathon, N. Mebarki, and C.

Levaillant: Mater. Sci. Eng. A, 2005, vol. 394, pp. 435–44.7. J.R. Paules, M.F. Dilmore, and K.J. Handerhan: Materials Sci-

ence Technology 2005 Conference, vol. 2, Association for Iron andSteel Technology, AISTECH, 25–28 Sept. 2005, pp. 13–23.

8. Sandia National Laboratories: Adv. Mater. Process., 2004,vol. 162 (162), pp. 18–19.

9. R. Abrahams: Unpublished Research, Penn State University, StateCollege, 2008.

10. American Society for Testing and Materials: ASTM E8 ASTMBook Stand, ASTM, West Conshohocken, 2004.

11. American Society for Testing and Materials: ASTM E23 - 07ae1ASTM Book Stand, ASTM, West Conshohocken, 2007.

12. American Society for Testing and Materials: ASTM E975 - 13ASTM Book Stand, ASTM, West Conshohocken, 2013.

13. B.D. Cullity and S.R. Stock: Elements of X-Ray Diffraction,Prentice Hall, Upper Saddle River, 2001.

14. J.M. Zuo and J.C. Mabon: Microsc. Microanal., 2004, vol. 10.15. H.L. Yakel: Int. Met. Rev., 1985, vol. 30, pp. 17–44.16. K.H.J. Jack: Iron Steel Inst., 1951, vol. 169, pp. 26–36.17. Y.A. Bagaryatski: Dokl. Akad. Nauk SSR, 1950, vol. 73, p. 1161.18. G.V. Kurdjumov and G.Z. Sachs: Z. Phys., 1930, vol. 64, p. 325.19. H.-S. Yang and H.K.D.H. Bhadeshia: Scripta Mater., 2009,

vol. 60, pp. 493–95.20. G. Thomas: Metall. Mater. Trans. A, 1978, vol. 9, pp. 439–50.21. S. Bjorklund, L.F. Donaghey, and M. Hillert: Acta Metall., 1972,

vol. 20, pp. 867–74.22. C.L. Briant and S.K. Banerji: Int. Met. Rev., 1978, vol. 23,

pp. 164–99.23. I. Olefjord: Int. Met. Rev., 1978, vol. 23, pp. 149–63.24. G.C. Gould: Temper Embrittlement Steel 59–59–15, ASTM

International, 1968.25. M. Guttmann et al.: Metall. Trans., 1974, vol. 5, pp. 167–77.26. H.L. Marcus, L.H. Hackett, and P.W. Palmberg: Temper

Embrittlement Alloy Steels 90–90–14, ASTM International, 1972.27. G.N. Haidemenopoulos and A.N. Vasilakos: J. Alloys Compd.,

1997, vol. 247, pp. 128–33.28. M.J. Hebsur: Mater. Energy Syst., 1982, vol. 4, pp. 28–37.

5746—VOLUME 46A, DECEMBER 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A