quantifying heat-affected zone hydrogen-induced cracking ...266-s september 2013, vol. 92 welding...

Introduction

BlastAlloy 160 (BA-160) was devel-oped at Northwestern University based ona theoretical computational materials de-sign concept. It has high yield strength of1100 MPa (160 ksi) and impact toughnessof 176 J (130 ft-lb) at 25°C and was de-signed as a blast-resistant steel for use innaval surface ships (Ref. 1).

The chemical composition and mi-crostructure of BA-160 steel are shown inTable 1 and Fig. 1, respectively. Thestrengthening of BA-160 results from thecombined effects of a martensite/bainitematrix, M2C carbides (where M repre-sents Cr, Mo, and V), and copper precipi-tates. The optimal size of the M2C car-bides and copper precipitates for effectivestrengthening is designed to be in therange of 3–5 nm. High toughness resultsfrom the presence of finely dispersed, Ni-stabilized austenite. The austenite willtransform to martensite in the stress field

at the crack-tip, and can inhibit crackgrowth by delay of microvoid shear local-ization during ductile fracture leading tovery high impact toughness (Ref. 2).

In order to achieve the carefully de-signed strengthening and toughening forBA-160, multistep tempering treatment isemployed, as shown in Table 2. The sam-ple is solution treated at 900°C for 1 h fol-lowed by a water quench and liquid nitro-gen hold, in order to ensure completemartensite transformation with no re-tained austenite. The first temperingtreatment at 550°C for 30 min is used to

nucleate a fine, uniformly dispersed in-tralath austenite and suboptimal sizestrengthening precipitates. The secondtempering treatment at 450°C for 5 h isthen used to enrich the austenite with Ni,thus stabilizing the austenite. At the sametime, the precipitation of copper and M2Ccarbides are completed, therefore a peakaging condition is achieved (Ref. 3).

Some aspects of BA-160’s weldabilityhave previously been studied by Caron etal. (Refs. 4, 5). It was found that BA-160exhibits a moderate heat-affected zone(HAZ) liquation cracking susceptibility asdetermined by hot ductility testing. Inother testing, BA-160 was found to be re-sistant to reheat cracking in the tempera-ture range of 450° to 700°C. Susceptibilityof HAZ hydrogen-induced cracking(HIC), which is usually a big problemwhen welding high-strength steels (Refs.6–8) was not included in this study.

One of the main objectives of the de-sign of BA-160 was to develop a weldablehigh-strength, high-toughness steel.Therefore, in the design stage, the carboncontent of BA-160 is intentionally con-strained at 0.05 ± 0.01 wt-% to locate thesteel in Zone I of the Graville diagram, asillustrated in AWS D1.1/D1.1M:2010(Ref. 9), as shown in Fig. 2. Steels locatedin this zone are typically resistant to HICbased on carbon content and its effect onthe hardness of the martensite that forms.However, Caron et al. (Ref. 10) reportedthat high-hardness martensite can formover a wide range of cooling rates in thesimulated HAZs investigated using theGleeble™. Prior austenite grain size canexceed 100 μm under simulated high-heat-input welding conditions, which is consid-erably larger than that of HSLA-100 (~50μm) (Ref. 11). It is therefore logical to ex-pect that the formation of martensitic mi-crostructure of high hardness of 360 HV inthe coarse-grained (CG)HAZ (Ref. 10)might render BA-160 susceptible to HAZHIC.

265-sWELDING JOURNAL

WE

LD

ING

RE

SE

AR

CH

KEYWORDS

Implant TestHeat-Affected Zone (HAZ)Hydrogen-Induced CrackingCoarse-Grained HAZ

(CGHAZ) MicrostructureFracture BehaviorBA-160HY-100HSLA-100

ABSTRACT

The implant test was used to compare a recently developed blast-resistant steel(BA-160) to two existing naval steels, HY-100 and HSLA-100, in order to assess po-tential susceptibility to heat-affected zone (HAZ) hydrogen-induced cracking (HIC).Based on the implant test results of applied tensile stress vs. time to failure, the lowercritical stress (LCS), normalized critical stress ratio (NCSR), and embrittlement index(EI) were determined in order to make a quantitative comparison. In all cases, fail-ure during implant testing occurred in the coarse-grained HAZ (CGHAZ). The mi-crostructure of the CGHAZ was characterized using optical and transmission elec-tron microscopy. It was found that martensite is the predominant microstructuralfeature in the CGHAZ of BA-160 and HY-100, while a mixture of martensite and bai-nite are observed in HSLA-100. The fracture behavior was characterized and relatedto microstructure and cracking susceptibility. Based on both NCSR and EI values, itcan be concluded that BA-160 is more susceptible to HAZ HIC compared to HSLA-100, but more resistant than HY-100.

X. YUE ([email protected]), X. L. FENG,and J. C. LIPPOLD are with the Welding Engi-neering Program, The Ohio State University,Columbus, Ohio.

Quantifying Heat-Affected Zone Hydrogen-Induced Cracking in

High-Strength Naval Steels

The implant test was used to compare BA-160 with HY-100 and HSLA-100 with regard to the HAZ HIC resistance

BY X. YUE, X. L. FENG, J. C. LIPPOLD

Yue Supplement September 2013_Layout 1 8/6/13 10:57 AM Page 265

In the present investigation, the im-plant test is used to evaluate the inherentHAZ HIC susceptibility of BA-160. Forthe purpose of comparing BA-160’s per-formance with existing high-strengthnaval steels, implant testing was also con-ducted on HY-100 and HSLA-100 navalsteels. In order to relate HAZ HIC sus-ceptibility to microstructure, the CGHAZwas characterized using optical and trans-mission electron microscopy, and fracturebehavior was determined using the scan-ning electron microscope (SEM). The ob-jective is to qualify the alloy design for BA-160 with regard to HAZ HIC resistance.The research results are used to develop aweldability database for these steels,which can be used as a benchmark for

evaluating the performance of futurehigh-strength steels.

Materials and ExperimentalProcedures

Table 1 summarizes the chemical com-positions of the three steels evaluated inthis investigation. BA-160 was provided inthe form of 34.9-mm- (1.375-in.-) diame-ter barstock by QuesTek Innovations LLC,Evanston, Ill. It was heat-treated follow-ing the procedure in Table 2. HY-100 andHSLA-100 were provided in the form of12.7-mm (0.5-in.) rolled plate by the NavalSurface Warfare Center, Carderock Divi-sion, West Bethesda, Md.

The implant test, which was originally

developed by Henri Granjon (Ref. 12), isused in the present investigation to evalu-ate the HAZ HIC susceptibility. It is an ex-ternally loaded cracking test, in which theexternal load level can be varied inde-pendently from the welding parameters incontrast to self-restraint cold crackingtests such as Y-groove and controlledthermal severity tests, where the stresscannot be quantified and a “critical stress”below which cracking does not occur can-not be determined. The implant test hasbeen shown by many investigators (Refs.13–17) to be an effective means to quanti-tatively evaluate the HAZ HIC suscepti-bility of steels and provide consistent as-sessment. It has also been shown to be anexcellent tool for alloy development, sinceit uses very little material.

The details of the implant test areshown in Fig. 3A–D. The implant speci-mens were machined from the respectivebase metals along the forging or rolling di-rection, with dimensions shown in Table 3.One end of the specimen is 0.5 in. longwith a 10-32 UNF thread and is fitted intoa clearance hole in the center of the spec-imen plate, with the top of the 10-32 UNFthread section flush with the specimenplate surface. The other end of the im-plant specimen is 0.5 in. long with a 1⁄4-20UNC thread, and is threaded into a con-nection rod of The Ohio State UniversityModified Implant Testing System (OSU-MITS) so that the hydraulic system of theOSU-MITS can apply the tensile load onthe specimen after welding.

A weld bead is deposited using the gasmetal arc welding (GMAW) process with0.047-in. (1.2-mm) SuperArc® LA-100 wire(AWS: ER100S-G) on the surface of thespecimen plate directly over the 10-32 UNFthread and the hole. Welding parametersused are as follows: voltage, 30 V; current,215–225 A; and travel speed, 12 in./min (5.1mm/s). This corresponds to a heat input in

SEPTEMBER 2013, VOL. 92266-s

WE

LD

ING

RE

SE

AR

CH

Fig. 1 — Microstructure of the BA-160 base metal. Fig. 2 — Zone classification of BA-160, HY-100, and HSLA-100 accordingto AWS D1.1/D1.1M:2010. Note Graville diagram is with extended abscissa.

Table 1 — Measured Chemical Composition of the Three Steels

Element (wt-%) BA-160 HY-100 HSLA-100

C 0.059 0.18 0.051Mn 0.001 0.28 0.90Si 0.015 0.21 0.25P <0.005 0.008 0.008S <0.001 0.002 0.002

Cu 3.39 0.15 1.17Ni 6.80 2.32 1.58Cr 1.90 1.37 0.60Mo 0.61 0.26 0.37V <0.001 <0.01 <0.01

Nb <0.001 <0.01 0.017Ti 0.016 <0.01 <0.01

Table 2 — Heat Treatment Procedure for BA-160

Step Temperature/°C Duration Post step procedure

1. Austenitization 900 1 h Water quench2. Liquid nitrogen hold –196 30 min Air warm to room temp.3. Tempering 550 30 min Water quench4. Tempering 450 5 h Air cool to room temp.

Yue Supplement September 2013_Layout 1 8/6/13 2:32 PM 66

the range of 32.25 to 33.75 kJ/in. (1.27 to1.33 kJ/mm). Ar+15%H2 shielding gas at aflow rate of 45 ft3/h (21.2 L/min) was used tointroduce sufficient hydrogen into the weldpool to allow diffusible hydrogen in theHAZ to cause cracking. Using Ar+15%H2shielding gas to introduce hydrogen, insteadof the method of applying lubricant oil uti-lized in the earlier research (Ref. 18), aimsto ensure the testing method had better op-erability and repeatability. More impor-tantly, in future research, varying the dif-fusible hydrogen content can be easilyrealized by adjusting the H2 content in theAr+H2 mixture shielding gas.

A HAZ is created in the 10-32 UNFthread region of the implant specimen afterwelding. After 2 min from completion ofwelding, the implant sample is subject to astatic tensile load. The tensile stress is de-termined by the load divided by the cross-sectional area of the root diameter of the10-32 UNF thread. The implant sample isfree of bending, torsion, or shock loading asa result of the specific design of OSU-MITS. The stress concentration caused bythe 10-32 UNF thread forces cracking tooccur in the susceptible HAZ region insteadof the fusion zone. A computer equippedwith a data-acquisition system was con-nected to the OSU-MITS to monitor theload and measure the time to failure. Inorder to generate the implant test curve,multiple samples were run with the samewelding parameters and different loads inorder to generate a tensile stress vs. time tofailure relationship. The highest stress atwhich no failure occurred after 24-h loadingis defined as the Lower Critical Stress(LCS) (Refs.12, 19), which was taken as anindex to determine susceptibility to HIC.Test runs to determine the LCS were re-peated twice to verify the accuracy of the ex-perimental results.

Metallographic samples were sectionedperpendicular to the welding directionalong the axis of the implant specimens.Then they were mounted, ground, polished,and etched with 5% nital and examinedusing optical microscopy. The TEM sam-ples were evaluated in a Philips CM200TEM operated at 200 kV. The fracture sur-face of the implant samples was examinedunder a Philips XL30F ESEM. Vickershardness measurements were conducted onthe as-polished samples using a 1-kg load, inaccordance with ASTM E 384-10.

Results and Discussion

Weld Macrostructure

All welding was carried out on the im-plant specimen/specimen plate with thesame welding parameters. Themacrostructure of a representative weldperpendicular to the welding directionsectioned along the axis of the implant

specimen is shown in Fig. 4. The width ofthe HAZ in the implant specimen is ap-proximately 3.5 mm, which is much wider

than that in the adjacent specimen plate asa result of the difference in heat flow in theimplant specimen, which creates a shal-


WE

LD

ING

RE

SE

AR

CH

Fig. 3 — The implant test system (OSU-MITS) and specimen. A — Schematic drawing of the implanttest; B — full view of the OSU-MITS; C — close-up view showing an implant specimen under loadingand an unloaded one on the top-right corner; D — the implant specimen.

Table 3 — Specimen Plate/Implant Specimen Dimensions

Specimen Plate

Material A36 steelPlate thickness in. (mm) 0.5 (12.7)Plate width in. (mm) 2 (50.8)Plate length in. (mm) 4 (101.6)Length of test bead in. (mm) 3.5 (88.9)Hole diameter in. (mm) 0.201 (5.1)

Implant Specimen

Material BA-160, HY-100, and HSLA-100Total length of implant specimen in. (mm) 1 (25.4)Type of thread 10-32 UNFPitch in. (mm) 1/32 (0.79)Major diameter in. (mm) 0.1900 (4.83)Minor diameter in. (mm) 0.1517 (3.85)Thread length in. (mm) 0.5 (12.7)Thread angle 60 degThread root radius in. (mm) 0.004 (0.1)

A B

C D


lower temperature gradient. Porosity canbe observed in the fusion zone near the fu-sion boundary that results from the hy-drogen that is intentionally introducedinto the weld pool through the use ofAr+15%H2 shielding gas.

After welding, the implant specimen issubjected to tensile loading and held untilfracture occurs up to a maximum of 24hours hold time. A representative frac-tured weld sectioned to reveal the HAZHIC fracture path is shown in Fig. 5. It canbe seen that cracking initiates from theroot of the first unfused thread and prop-agates in the CGHAZ of the implant spec-imen. This is because of large prioraustenite grain size, and the presence ofhigh hardness in the CGHAZ provides themost susceptible microstructure. This lo-cation for fracture initiation and propaga-tion was consistent for all three steels.

Vickers Hardness Test Results

Vickers hardness measurements weretaken along the axis of the implant speci-mens of the three steels, starting in the fu-sion zone and running through the HAZto the unaffected base metal, as shown inFig. 6A–C. The hardness variation acrossthe HAZ is a direct indication of forma-tion of different microstructures. For thethree steels, the fusion zone hardness is inthe range of 250 to 280 HV. It is shown inFig. 6A that in the HAZ of the BA-160 im-plant specimen, the Vickers hardness in-creases from the CGHAZ to the fine-grained HAZ (FGHAZ) and into theintercritical HAZ (ICHAZ). TheCGHAZ hardness is in the range of 350 to376 HV, which is even lower than that ofthe base metal, representing a “softening”in the HAZ. This phenomenon was inves-

tigated by Yu et al. (Ref. 20), and it was at-tributed to different lath martensite mor-phologies and Cu precipitation behaviorsdue to different thermal cycles experi-enced in different HAZ regions. However,even though the CGHAZ has a lowerhardness than other HAZ regions, it is themost HIC-susceptible region, which is dueto the formation of the HIC-susceptiblemicrostructure of martensite with coarsegrain size (Ref. 6). It is shown in Fig. 6Bthat the hardness of HY-100 CGHAZ is inthe range of 440 to 464 HV, which is muchhigher than both the fusion zone and thebase metal. For HSLA-100, the CGHAZhardness is in the range of 293 to 329 HV,which is slightly higher than fusion zoneand base metal. It should be noted that thelocation of the CGHAZ, as shown in Fig.6A–C, is determined by metallographicobservation and the red dotted line is only


WE

LD

ING

RE

SE

AR

CH

Fig. 4 — Typical implant test specimen sectioned near the implant axis. BA-160 implant specimen. 5% nital etch.

Fig. 5 — Macrostructure of a typical fractured implant specimen after load-ing showing the fracture path. HY-100 implant specimen. 5% nital etch.

Table 4 — Implant Test Results

Steel CGHAZ Max CGHAZ Tensile Lower Critical Nominal Yield NCSR(b) Embrittlement CEAWS(d)

Hardness Strength(a) Stress Strength Index(c)

(HV) ksi(MPa) ksi(MPa) ksi(MPa)

BA-160 376 173 (1192) 91 (627) 160 (1102) 0.57 0.53 1.24HY-100 464 225 (1550) 55 (379) 100 (689) 0.55 0.24 0.75

HSLA-100 329 150 (1034) 102 (703) 100 (689) 1.02 0.68 0.62

(a) The CGHAZ tensile strength is converted from the CGHAZ max hardness using the ASTM hardness conversion chart.(b) NCSR stands for Normalized Critical Stress Ratio, which is the ratio of lower critical stress to nominal yield strength.(c) Embrittlement Index is the ratio of lower critical stress to the CGHAZ tensile strength.(d) AWS carbon equivalent: CEAWS = C + (Mn + Si)/6 + (Cr + Mo + V)/5 + (Ni + Cu)/15.

Fig. 6 — Vickers hardness measurements taken along the axis of the implant specimen. A — BA-160; B — HY-100; C — HSLA-100.

CA B

Yue Supplement September 2013_Layout 1 8/6/13 4:11 PM Page 268

the approximate boundary between theCGHAZ and the adjacent FGHAZ. Themaximum CGHAZ hardness is 376, 464,and 329 HV for BA-160, HY-100, andHSLA-100, respectively.

Microstructure Characterization ofCGHAZ of the Three Steels

Since the CGHAZ is the most HIC-susceptible region, the CGHAZ mi-crostructure of BA-160 was characterizedby means of optical microscopy and TEM,as shown in Fig. 7. It can be seen in Fig. 7Athat martensite is the predominant mi-crostructural feature. The microstructure isfurther investigated at higher magnificationunder TEM, as shown in Fig. 7B. Marten-site laths can be observed in Fig. 7B and thedark film between adjacent martensite lathsis retained austenite. The presence of re-tained austenite is a result of the high nickeladdition (6.8 wt-%) in BA-160, which is aneffective austenite stabilizer depressing theMs temperature. In addition, a heavy dislo-

cation network exists within the martensitelaths, which is a feature of low-carbon lathmartensite. The BA-160 CGHAZ marten-sitic microstructure has maximum Vickershardness of 376 HV.

For the purpose of comparing BA-160with HY-100 and HSLA-100, the CGHAZmicrostructures of the latter two steels havealso been studied, as shown in Figs. 8 and 9,respectively. From Fig. 8A, it is apparentthat the HY-100 CGHAZ microstructure isalso primarily martensitic. Packets ofmartensite laths with dislocation networkscan be observed in Fig. 8B. The HY-100CGHAZ martensitic microstructure hasmaximum Vickers hardness of 464 HV,which is much higher than that of BA-160.

The CGHAZ microstructure of HSLA-100 is shown in Fig. 9. Because of lowerhardenability than HY-100 (Ref. 11), themicrostructure of the HSLA-100 CGHAZis not fully martensitic. Light-etching nee-dle-like bainite can be observed in Fig. 9A,which nucleates from the prior austenitegrain boundaries and grows into the grain

interior. In the lower-left part of the TEMmicrostructure shown in Fig. 9B, a group ofparallel laths free of precipitates is ob-served, which are the martensite laths.While in the upper-right corner of Fig. 9B,intralath platelet-like cementite precipi-tates can be seen oriented at approximately55 deg from the primary lath growth direc-tion, confirming the presence of lower bai-nite. Therefore, the CGHAZ microstruc-ture of HSLA-100 is a mixture of martensiteand bainite, with maximum Vickers hard-ness of 329 HV, which is the lowest amongthe three steels.

Implant Test Results

The implant test results for the threesteels are shown in Fig. 10A–C. The de-layed nature of HIC can be clearly seenfrom the implant test curves. All threecurves show a general trend that a longerincubation time is required before frac-ture occurs under lower applied stress lev-els. The implant test results for the three


WE

LD

ING

RE

SE

AR

CH

Fig. 7 — CGHAZ microstructure of BA-160. A — Optical; B — bright-field TEM.

Fig.8 — CGHAZ microstructure of HY-100. A — Optical; B — bright-field TEM.

A B

A B


steels are summarized in Table 4. The LCS(from lowest to highest) is determined tobe 55 ksi (379 MPa), 91 ksi (627 MPa), and102 ksi (703 MPa) for HY-100, BA-160,and HSLA-100, respectively.

Two indexes are used in the presentstudy to compare BA-160 with HY-100and HSLA-100. The first is the normalizedcritical stress ratio (NCSR) (Ref. 17),which is the ratio of LCS to nominal yieldstrength of the test steel. Yield strength isan important index for evaluating a steel’smechanical properties, and in highly re-strained welds, the residual stress can evenreach the yield strength level. By takingthe ratio of LCS to nominal yield strength,the percent HAZ strength degradation(under the influence of diffusible hydro-gen) from the nominal yield strength canbe determined. The higher the NCSR, thebetter the performance of the steel. As thenominal yield strength is 160 ksi (1102MPa), 100 ksi (689 MPa), and 100 ksi (689MPa) for BA-160, HY-100, and HSLA-100, respectively, the NCSR is determinedaccordingly to be 0.57, 0.55, and 1.02 forBA-160, HY-100, and HSLA-100,respectively.

Another index used in this paper is theembrittlement index (EI), which is theratio of LCS to the CGHAZ tensilestrength. However, it is difficult to meas-ure the CGHAZ tensile strength directlyfrom the implant test. A convenient meansto obtain an approximation of theCGHAZ tensile strength is to do the con-version from the maximum CGHAZ Vick-ers hardness according to the ASTM hard-ness conversion chart. The EI is proposedherein aiming to evaluate the steel’sCGHAZ inherent susceptibility to HIC.That is, if EI is higher, the degradation ofthe CGHAZ strength because of dif-fusible hydrogen is lower, which means abetter CGHAZ inherent resistance toHIC. By doing the conversion, the ap-proximate CGHAZ tensile strength is 173ksi (1192 MPa), 225 ksi (1550 MPa), and150 ksi (1034 MPa) for BA-160, HY-100,and HSLA-100, respectively. The EI is de-termined accordingly to be 0.53, 0.24, and0.68 for BA-160, HY-100, and HSLA-100,respectively. The result shows that BA-160CGHAZ has less inherent HIC suscepti-bility than HY-100, but is more suscepti-ble to HIC compared with HSLA-100.

Fracture Behavior

Representative SEM fractographs ofthe implant specimens of the three steelsare presented in Figs. 11–13. Figure 11Ashows the general fracture appearance ofthe BA-160 implant specimen, and regionsin Fig. 11A with different features areshown in Fig. 11B–D at higher magnifica-tion. Cleavage type failure can be ob-served at the crack initiation site (Fig.11B), which is near the root of the first un-fused thread, and no large area of inter-granular (IG) mode can be observed. Asthe crack further propagates, the fracturemode is quasi-cleavage (QC), as shown inFig. 11C. Microvoid coalescence (MVC)is observed near the final failure region, asshown in Fig.11D.

For the HY-100, Fig. 12A shows a largearea of IG fracture, which is in sharp con-trast to BA-160. Cracking initiates fromthe stress concentration area and propa-gates intergranularly (Fig. 12B) for a cer-tain distance, and then transitions to QCand MVC, as shown in Fig. 12C and D, re-spectively.

Similar to HY-100, all the three frac-


WE

LD

ING

RE

SE

AR

CH

Fig. 9 — CGHAZ microstructure of HSLA-100. A — Optical; B — bright-field TEM.

Fig. 10 — The implant test curves of the following: A — BA-160; B — HY-100; C — HSLA-100.

C

A B

Yue Supplement September 2013_Layout 1 8/6/13 2:36 PM Page 270

ture modes, IG, QC, and MVC, can be ob-served on the HSLA-100 implant speci-men fracture surface, as shown in Fig.13A–D. However, the grain size observedin the IG region is smaller for HSLA-100compared with HY-100, and the area ofIG fracture is less than that of HY-100.Quasi-cleavage and MVC are observed insequence as the crack propagates.

The sequential occurrence of IG, QC,and MVC on the fracture surface was ex-plained using Beachem’s model (Ref. 21) inan earlier publication (Ref. 18). Much ofBeachem’s theory is based on the fact thefracture behavior is dependent on the inter-action of stress intensity and hydrogen level.For the present study, when the implantsample is subjected to tensile loading, acrack will not initiate immediately due tothe lack of sufficient diffusible hydrogen atthe stress concentration area. An incuba-tion period is thereby required so thatatomic hydrogen can continuously diffuseto the triaxially stressed region, and when itreaches the critical level, a crack will initi-ate. As the crack propagates, the cracklength increases and the stress intensity fac-tor will increase and the hydrogen level willdecrease. The fracture mode will thereforechange to QC and MVC.

As discussed in the previous section, theCGHAZ is the most susceptible to HICamong the HAZ regions. Therefore, thefracture behavior and CGHAZ mi-crostructure must be taken together to ex-plain the difference in HAZ HIC suscepti-bility of the three steels. The BA-160CGHAZ microstructure consists ofmartensite with some retained austenite,with maximum Vickers hardness of 376 HV.For HY-100, the CGHAZ microstructureis also martensite but with a much highermaximum hardness of 464 HV. TheCGHAZ grain size of BA-160 and HY-100are similar, in the range of 70–80 μm, whilethe CGHAZ microstructure of HSLA-100is a mixture of martensite and bainite,which has the lowest maximum hardness of329 HV among the three steels and also thelowest CGHAZ grain size, in the range of50–60 μm. It is known that a martensitic mi-crostructure with high hardness and coarseprior austenite grain size is detrimental toHIC resistance (Ref. 22). Therefore, HY-100 CGHAZ is the most susceptible to HICamong the three steels. In addition, it isfound that no large area of IG failure canbe observed on the fracture surface of BA-160 while a large area of IG exists on thatof HY-100. Since IG involves the least

amount of plastic deformation comparedto QC and MVC modes, occurrence of alarge area of IG on the fracture surface in-dicates a high potential for HIC. Charac-terization of the CGHAZ microstructureand fracture behavior supports the implanttest results. If taking both NCSR and EIinto consideration for HAZ HIC suscepti-bility evaluation, BA-160 shows a better re-sistance to HAZ HIC than HY-100 but ismore susceptible to HAZ HIC comparedto HSLA-100.

Conclusions

1. Under the welding conditions usedfor implant testing, the maximum Vickershardness of the CGHAZ is 376, 464, and329 HV for BA-160, HY-100, and HSLA-100, respectively.

2. The CGHAZ microstructure of BA-160 consists of martensite with retainedaustenite. For HY-100, the CGHAZ mi-crostructure is also martensite while forHSLA-100 a mixture of martensite andlower bainite forms in the CGHAZ.

3. Under the implant test conditionsused in this study, the Lower CriticalStress (LCS) was determined to be 55 ksi(379 MPa), 91 ksi (627 MPa), and 102 ksi


WE

LD

ING

RE

SE

AR

CH

C D

A B

Fig. 11 — Fracture morphology of BA-160 implant specimen under a stress of 118.9 ksi (819MPa) with failure after 13 min of loading. A — General fractureappearance (white arrow indicates the direction of crack growth); B — crack initiation; C — quasi-cleavage; D — microvoid coalescence. B, C, and D are ob-served in sequence along the crack propagation path.


(703 MPa) for HY-100, BA-160, andHSLA-100, respectively. The lower valuesof LCS indicate higher susceptibility toHIC.

4. The Normalized Critical StressRatio (NCSR) is determined by normaliz-ing the LCS using the nominal yieldstrength of the base metal. The NCSR wasdetermined to be 0.55, 0.57, and 1.02 forHY-100, BA-160, and HSLA-100, respectively.

5. The Embrittlement Index (EI) is de-termined by normalizing the LCS usingthe approximate tensile strength of theCGHAZ. The EI for these steels was 0.24,0.53, and 0.68 for HY-100, BA-160, andHSLA-100, respectively.

6. Fractographic analysis revealed thatIG, QC, and MVC fracture modes are ob-served on the fracture surface of HY-100and HSLA-100 implant specimens. Cracksinitiate in the CGHAZ stress concentra-tion region and grow intergranularly for acertain distance before transitioning toQC and MVC in sequence as the crackpropagates with an increase in stress in-tensity factor and decrease in hydrogenlevel.

7. No large area of IG fracture modescould be found on the fracture surface ofBA-160. Among the three steels, thelargest area of IG failure with the coarsestgrains was found on the fracture surface ofHY-100, indicating the most seriousdegradation due to effect of diffusiblehydrogen.

8. Taking both NCSR and EI into con-sideration for HAZ HIC susceptibilityevaluation, BA-160 showed better resist-ance to HAZ HIC than HY-100 but moresusceptibility to HAZ HIC compared toHSLA-100. This is because with similarCGHAZ grain size, the martensitic mi-crostructure in the BA-160 CGHAZ is ofa much lower hardness compared withHY-100. HSLA-100 shows the highest re-sistance to HAZ HIC than BA-160 andHY-100, as a result of formation of lower-hardness bainite and martensite mixturewith smaller grain size in its CGHAZ.

Acknowledgments

The authors gratefully acknowledgethe financial support of the Office ofNaval Research, Award No.

N000140811000. Its grant officers are Dr.Julie Christodoulou and Dr. WilliamMullins. The authors would like to thankJohnnie DeLoach, Matthew Sinfield, andJeffrey Farren with the Naval SurfaceWarfare Center Carderock Division, WestBethesda, Md., for providing the HY-100and HSLA-100 steels used in this studyand for valuable discussions regarding theweldability of these steels. Thanks are ex-tended to Prof. Gregory Olson’s researchgroup at Northwestern University for col-laboration on this research project andQuesTek Innovations LLC for providingthe BA-160 steel. Dejian Liu and GeoffreyTaber are acknowledged for their con-structive ideas and assistance with build-ing the implant testing system.

References

1. Saha, A. 2004. Systems design of trans-formation toughened blast-resistant naval hullsteels. PhD dissertation. Evanston, Ill.: North-western University.

2. Saha, A., and Olson, G. B. 2007. Computer-aided design of transformation toughened blast-resistant naval hull steels: Part I. Journal of Com-puter-Aided Materials Design 14: 177–200.


WE

LD

ING

RE

SE

AR

CH

Fig. 12 — Fracture morphology of HY-100 implant specimen under a stress of 107.9 ksi (743MPa) with failure after 9 min of loading. A — General fracture ap-pearance (white arrow indicates the direction of crack growth); B — coarse intergranular; C — quasi-cleavage; D — microvoid coalescence. B, C, and D areobserved in sequence along the crack propagation path.

C D

A B


3. Saha, A., Jung, J., and Olson, G. B. 2007.Prototype evaluation of transformation tough-ened blast-resistant naval hull steels: Part II.Journal of Computer-Aided Materials Design 14:201–233.

4. Caron, J. 2010. Weldability evaluation ofnaval steels. PhD dissertation. Columbus,Ohio: The Ohio State University.

5. Caron, J. L., Babu, S. S., and Lippold, J.C. 2011. Weldability evaluation of a Cu-bearinghigh-strength blast-resistant steel. Metallurgicaland Materials Transactions A 42A: 4032–4044.

6. Easterling, K. 1992. Introduction to thePhysical Metallurgy of Welding. pp: 216–233, Ox-ford, UK: Butterworth-Heinemann Ltd.

7. Devletian, J. H., and Fichtelberg, N. D.2001. Controlling hydrogen cracking in ship-building. Welding Journal 80(11): 46-s to 52-s.

8. DeLoach, J. J., Null, C., Flore, S., andKonkol, P. 1999. The right welding wire couldhelp the U.S. Navy save millions. Welding Jour-nal 78(6): 55–58.

9. AWS D1.1/D1.1M:2010, Structural Weld-ing Code — Steel. Miami, Fla.: American Weld-ing Society.

10. Caron, J. L., Babu, S. S., and Lippold, J.

C. 2011. Welding-induced microstructure evo-lution of a Cu-bearing high-strength blast-resis-tant steel. Metallurgical and Materials Transac-tions A 42A: 4015–4031.

11. Yue, X., Lippold, J. C., Alexandrov, B.T., and Babu, S. S. 2012. Continuous coolingtransformation behavior in the CGHAZ ofnaval steels. Welding Journal 91(3): 67-s to 75-s.

12. Granjon, H. 1969. The implants methodfor studying the weldability of high strengthsteels. Metal Construction and British WeldingJournal 1(11): 509–515.

13. Gedeon, S. A., and Eagar, T. W. 1990.Assessing hydrogen-assisted cracking fracturemodes in high-strength steel weldments. Weld-ing Journal 69(6): 213-s to 220-s.

14. Sawhill, J. M. Jr., Dix, A. W., and Savage,W. F. 1974. Modified implant test for studyingdelayed cracking. Welding Journal 53(12): 554-sto 560-s.

15. Vasudevan, R., Stout, R. D., and Pense,A. W. 1981. Hydrogen-assisted cracking inHSLA pipeline steels. Welding Journal 60(9):155-s to 168-s.

16. Signes, E. G., and Howe, P. 1988. Hy-drogen-assisted cracking in high-strength

pipeline steels. Welding Journal 67(8): 163-s to170-s.

17. Dickinson, D. W., and Ries, G. D. 1979.Implant testing of medium to high strengthsteel — A model for predicting delayed crack-ing susceptibility. Welding Journal 59(7): 205-sto 211-s.

18. Yue, X., and Lippold, J. C. 2013. Evalu-ation of heat-affected zone hydrogen-inducedcracking in Navy steels. Welding Journal 92(1):20-s to 28-s.

19. AWS B4.0:2007, Standard Methods forMechanical Testing of Welds. pp: 67–71. Miami,Fla.: American Welding Society.

20. Yu, X., Caron, J. L., Babu, S. S., Lippold,J. C., Isheim, D., and Seidman, D. N. 2010.Characterization of microstructural strength-ening in the heat-affected zone of a blast-resis-tant naval steel. Acta Materialia 58: 5596–5609.

21. Beachem, C. D. 1972. A new model forhydrogen-assisted cracking (hydrogen “embrit-tlement”). Metallurgical Transactions 3:437–451.

22. Kou, S. 2003. Welding Metallurgy. pp:410–417. Hoboken, N.J.: John Wiley & Sons,Inc.


WE

LD

ING

RE

SE

AR

CH

Fig. 13 — Fracture morphology of HSLA-100 implant specimen under stress of 113.5 ksi (782 MPa) with failure after 45 min of loading. A — General fractureappearance (white arrow indicates the direction of crack growth); B — intergranular; C — quasi-cleavage; D — microvoid coalescence. B, C, and D are ob-served in sequence along the crack propagation path.

C D

A B


quantifying heat-affected zone hydrogen-induced cracking ...266-s september 2013, vol. 92 welding...

Documents