MARCH 2014, VOL. 9398-s

WELD

ING

RESEA

RC

H

Introduction

Steels are currently being used as theprincipal structural material in naval shipconstruction, primarily because of theirrelatively low cost and good combinationof mechanical properties. For many years,the U.S. Navy has focused on developingstronger and tougher steels for the hulland deck applications (Refs. 1–3). Theprojected property requirements for high-strength, blast-resistant naval steelsneeded for future ship applications is im-pact fracture toughness above 115 Jat –64°C with a yield strength in the rangefrom 1030 to 1240 MPa. These steelsshould possess good formability and weld-ability, especially good resistance to heat-affected zone (HAZ) hydrogen-inducedcracking (HIC) (Ref. 4).

A new steel, BlastAlloy 160 (BA-160),was developed at Northwestern Univer-sity to meet these rigorous requirementsfor blast-resistant naval material applica-tions. It was designed based on a theoret-ical computational materials design con-cept, using a multiscale materialsmodeling method and detailed advancedmicrostructural characterization tech-niques. It has high yield strength of 1100MPa (160 ksi) and impact toughness of

176 J at 25°C. The chemical compositionof BA-160 is listed in Table 1. The steelheat treatment procedure, as outlined inTable 2, was used to achieve the combina-tion of high strength and toughness. Thehigh strength results from the combinedeffects of a martensite/bainite matrix,M2C carbides (where M=Cr, Mo, and V)and copper precipitates, which are in therange of 3–5 nm. High toughness resultsfrom the presence of finely dispersed, Ni-stabilized austenite, based on a dispersedphase transformation toughening mecha-nism. More information on the design andmicrostructure development of BA-160can be found in publications by Saha et al.(Refs. 5, 6).

As with most steels developed for ship-building applications, one of the primarydesign objectives for BA-160 is that itshould possess good weldability. Amongthe weldability issues associated withwelding high-strength steels, HAZ HIC isone of the biggest concerns, as validatedby numerous publications by previous re-searchers (Refs. 7–15). Heat-affectedzone HIC was also referred to as HAZcold cracking or delayed cracking in thewelding literature. Researchers have pro-posed a number of theories to describe thecracking mechanism. Some of the morepopular theories that have evolved overthe years include the surface adsorptiontheory by Petch (Ref. 16), the decohesiontheory by Troiano (Ref. 17), and the the-ory proposed by Beachem stating thatfracture behavior is dependent on thecombined effect of stress intensity at cracktip and hydrogen concentration (Ref. 18).Even though a unified mechanism for HICstill does not exist, it is generally agreedthat the occurrence of HIC in the HAZ ofwelds in high-strength steels requires thesimultaneous presence of a threshold levelof hydrogen, a susceptible microstructure,and tensile residual stress (Ref. 19).

In order for the new steel to be weldedwith good resistance to HAZ HIC, an un-derstanding of the influence of welding

Effect of Welding Parameters on theHeat-Affected Zone Hydrogen-Induced

Cracking Tendency of a Blast-Resistant SteelThe effect of heat input and preheat on the HAZ hydrogen-induced cracking

tendency of BA-160 was evaluated by the implant test

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

ABSTRACT

An implant test was used to investigate the effect of welding conditions on thesusceptibility of a blast-resistant steel, BlastAlloy-160 (BA-160), to heat-affectedzone (HAZ) hydrogen-induced cracking (HIC). The lower critical stress (LCS)was determined using the implant test with different heat input conditions andpreheat temperatures. Welding was performed using the gas metal arc process andhydrogen was introduced using an Ar-15%H2 shielding gas. The microstructureof the coarse-grained heat-affected zone (CGHAZ) of BA-160 was characterizedby means of both optical and transmission electron microscopy. The CGHAZ mi-crostructure of the as-welded, low-heat-input welds consisted of untemperedmartensite with some retained interlath austenite. Increased heat input leads tothe formation of lower bainite, decreasing the hardness of the microstructure. Useof preheat for low-heat-input welds also slightly reduced CGHAZ hardness, andthe microstructure is predominantly martensite with the possibility of some bai-nite. It was shown in the implant test results that both increasing heat input andusing preheat improved the HIC resistance of the HAZ. The fracture behaviorwas studied using scanning electron microscopy. It was shown that both weldingwith high-heat input and applying preheat resulted in an increase in fracture mor-phology dominated by microvoid coalescence with no intergranular fracture.

X. YUE (yuexinosu@gmail.com), X.-L. FENG,and J. C. LIPPOLD are with the Welding Engi-neering Program, The Ohio State University,Columbus, Ohio.

KEYWORDS

High-Strength SteelsHydrogen-Induced CrackingHeat InputPreheatImplant TestCGHAZ MicrostructureFracture BehaviorBA-160

Yue (201339) March 2014_Layout 1 2/13/14 9:58 AM Page 98

parameters on the cracking tendencyusing lab test is needed. In the presentstudy, the implant test was used to evalu-ate the HAZ HIC tendency of BA-160under different welding conditions. Theinfluence of heat input and effectivenessof preheat on reducing cracking suscepti-bility have been investigated using the im-plant test. The CGHAZ microstructurewas characterized using optical and trans-mission electron microscopy, and the frac-ture behavior was studied with scanningelectron microscopy.

Material and ExperimentalProcedures

BA-160 was provided in the form of 35-mm- (1.375-in.-) diameter bar stock byQuesTek Innovations LLC, Evanston, Ill.The composition of the material used inthis study is listed in Table 1. It was heat-treated following the procedure in Table 2,and the base metal microstructure isshown in Fig 1.

The implant test, which was originallydeveloped by Henri Granjon (Ref. 20),was used in the present investigation toevaluate the HAZ HIC susceptibility. Ithas been shown to be an effective HAZHIC test method, which provides a quan-titative measure of HIC susceptibility(Refs. 21, 22). The schematic of the im-plant test, testing system, and implantspecimen are shown in Fig. 2A–D. The im-plant specimens, as shown in Fig. 2D, weremachined from the BA-160 steel along therolling direction of the bar stock, with di-mensions shown in Table 3. One end of theimplant specimen was 0.5 in. (12.7 mm)long with a 10-32 UNF thread and was in-

serted into a clearance hole in the centerof the specimen plate, with the top of the10-32 UNF thread section flush with thespecimen plate top surface. The other endof the implant specimen was 0.5 in. (12.7mm) long with a 1/4-20 UNC thread andwas threaded into a connection rod of TheOhio State University Modified ImplantTesting System (OSU-MITS) so that aconstant tensile load could be appliedafter welding is completed.

A weld bead was deposited using thegas metal arc welding (GMAW) processwith 0.047-in.- (1.2-mm-) diameter Su-perArc® LA-100 wire (compositionshown in Table 1) on the surface of thespecimen plate directly over the 10-32

UNF thread and the hole. A low and ahigh heat input, which were 33 kJ/in. (1.3kJ/mm) (voltage, 30 V; current, 220 A;travel speed, 12 in./min); and 66 kJ/in(2.6kJ/mm) (voltage, 30 V; current, 220 A;travel speed: 6 in./min), respectively,were used to evaluate the effect of heatinput on the HAZ cracking tendency. Inaddition, for the low heat input, preheatof 60°, 100°, and 150°C were applied be-fore welding to evaluate the preheat ef-fect on reducing the HIC tendency. Be-fore welding, the specimen plate andimplant specimen were heated to 300°Cin a heating furnace, then were quicklymoved to the OSU-MITS and fixed onthe testing system. A Type K thermocou-

99-sWELDING JOURNAL

WEL

DIN

G R

ESEA

RC

H

Fig. 1 — Microstructure of the BA-160 base metal,which shows a tempered martensite/bainite matrixwith various precipitates dispersed on the matrix.

Table 1 — Chemical Composition of BA-160 Steel and SuperArc LA-100 (wt-%)

C Mn Si P S Cu Ni Cr Mo V Nb Ti Al Zr

BA-160 0.059 0.001 0.015 <0.005 <0.001 3.39 6.8 1.9 0.61 <0.001 <0.001 0.016 0 0SuperArc LA-100 0.05– 1.63– 0.46– 0.005– 0.002– 0.11– 1.88– 0.04– 0.43– ≤0.01 0 0.03– ≤0.01 ≤0.01

0.06 1.69 0.50 0.009 0.005 0.14 1.96 0.06 0.45 0.04

Fig. 2 — 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.

A

C

B

D

Yue (201339) March 2014_Layout 1 2/13/14 10:00 AM Page 99

ple was connected to the data-acquisition system of OSU-MITS andused to measure the temperature of thespecimen plate. Welding was started onthe preheated specimens once the tem-perature of the specimen plate droppedto the predetermined preheat tempera-ture. Argon + 15% H2 shielding gas at aflow rate of 45 ft3/h (21.2 L/min) was usedto introduce sufficient diffusible hydro-gen into the weld joint to promote HIC inthe HAZ.

A HAZ was created in the 10-32 UNFthread region of the implant specimenafter welding. Two minutes after comple-tion of welding, the implant sample wassubjected to a static tensile load. The ten-sile stress was determined by the load di-vided by the cross-sectional area of theroot diameter of the 10-32 thread. The im-plant sample was free of bending, torsion,or shock loading as a result of the specificdesign of OSU-MITS. The stress concen-tration caused by the 10-32 UNF threadforced cracking to occur in the susceptibleHAZ region rather than the fusion zone.

The data-acquisition system was used tomonitor the load and measure the time tofailure. To generate the implant test curve,multiple samples were welded with thesame welding parameters and subjected todifferent loads in order to generate a ten-sile stress vs. time to failure relationship.The highest stress at which no failure oc-curs after 24-h loading was defined as thelower critical stress (LCS) (Ref. 23), whichwas taken as an index to determine sus-ceptibility to HIC in the HAZ.

Metallographic samples were sectionedperpendicular to the welding directionalong the axis of the implant specimens. Fol-lowing mounting and polishing, they wereetched with 5% nital and examined usingoptical microscopy. Transmission electronmicroscope (TEM) samples were evaluatedin a Philips CM200 TEM operated at 200kV. The fracture surface of the implant sam-ples was examined under a Philips XL30FESEM. Vickers hardness measurementswere conducted along the axis of the im-plant samples using a 1-kg load, in accor-dance with ASTM E 384-10.

Results and Discussion

Weld Macrostructure

The weld macrostructures under dif-ferent welding conditions are shown inFig. 3A–C. It can be seen that a distinctHAZ was created on the 10-32 UNFthread region for all three weld condi-tions. When increasing heat input (Fig.3B) and using preheat (Fig. 3C), the areaof the fusion zone is larger as compared tothat welded with low heat input (Fig. 3A),and the width of the HAZ becomes largerwhen high heat input or preheat was used.Figure 3D shows a fractured implant spec-imen sectioned to reveal the HIC fracturepath. The fusion boundary separating thefusion zone and HAZ can be clearly seen.The crack initiates from the root of the un-fused thread and propagates through theCGHAZ approximately 100–300 micronsaway from the fusion boundary. In moststeels, the CGHAZ is the most susceptibleto HAZ HIC since it has the largest prioraustenite grain size and high hardness(Refs. 24–26).

Vickers Hardness Test Results

Hardness is a direct indication of a dif-ferent microstructure formed, and is animportant factor that influences the HAZHIC tendency. Therefore, Vickers hard-ness measurements were taken along theaxis of the implant specimens weldedunder different conditions, starting in thefusion zone and running through the HAZ

MARCH 2014, VOL. 93100-s

WELD

ING

RESEA

RC

H

Fig. 3 — Macrostructure of implant test specimens welded under dif-ferent conditions. A — Weld with low heat input; B — weld with highheat input; C — weld with low heat input, preheat at 150°C beforewelding; D — macrostructure of a fractured implant specimen afterloading showing the fracture path.

A B

C D

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 temp3. Tempering 550 30 min Water quench4. Tempering 450 5 h Air cool to room temp

Yue (201339) March 2014_Layout 1 2/13/14 10:01 AM Page 100

to the unaffected base metal, as shown inFig. 4A–C. The HAZ is indicated in thethree hardness plots as the region betweenthe two black solid lines. The red dottedline is the approximate boundary betweenthe CGHAZ and its adjacent fine-grainedHAZ (FGHAZ).

It can be seen in all three hardnesstraverses that the hardness of the fusionzone is lower compared to HAZ and basemetal. And when welded with high heatinput or using preheat, the fusion zonehardness decreases as compared to usinglow heat input without preheat, indicat-ing different cooling rates under differ-ent welding conditions lead to the forma-tion of different microstructures in thefusion zone.

For all three conditions, the hardnessof the CGHAZ is actually the lowestacross the HAZ. This is attributed to thedifference in lath martensite morphologyand Cu precipitation behavior in differentHAZ regions as a result of different ther-mal cycles experienced (Ref. 4). Eventhough the CGHAZ has the lowest HAZhardness, it is the most HIC-susceptibleregion, as shown in Fig. 3D. When usinglow heat input as shown in Fig. 4A, the av-erage CGHAZ hardness is 370 HV1.0, andit slightly decreases to an average of 358HV1.0 when welded with high heat input,and to an average of 363 HV1.0 when using

preheat at 150°C. The hardness results aresummarized in Table 4.

Weld CGHAZ Microstructure

Since cracking occurs in the CGHAZregion adjacent to the fusion boundary,the CGHAZ microstructure will influencethe cracking tendency. Therefore, in this

study, the CGHAZ microstructure of BA-160 under different welding conditionswas characterized.

It is known that cooling rate influencesthe phase transformation behavior. Usingan online weld modeling tool, EWI E-Weld Predictor™ (Ref. 27), the t8/5, whichrepresents the cooling time from 800° to500°C, is estimated to be 3.5, 7.3, and 4.8 s

101-sWELDING JOURNAL

WEL

DIN

G R

ESEA

RC

H

Table 3 — Specimen Plate/Implant Specimen Dimensions

Specimen plateMaterial 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 in. (mm) 3.5 (88.9)beadHole diameter in. (mm) 0.201 (5.1)

Implant specimen

Material BA-160Total length of implant 1 (25.4)specimen in. (mm)Type of thread 10-32 UNF Pitch 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 CB

Fig. 4 — Vickers hardness measurements taken along the axis of the implant specimen. A — Weld with low heat input; B — weld with high heat input; C — weld with low heat input, preheat at 150°C before welding.

BA C

Fig. 5 — Optical micrographs for the CGHAZ of BA-160. A — Weld with low heat input; B — weld with high heat input; C — weld with low heat input, pre-heat at 150°C before welding.

Yue (201339) March 2014_Layout 1 2/13/14 10:01 AM Page 101

for low heat input, high heat input, andusing preheat of 150°C, respectively. Notethat the calculated t8/5 is only an approxi-mate value; however, it can be clearly seenthat both welding with high heat input andapplying preheat decrease cooling rate ascompared to welding with low heat input.

It is shown in the optical micrographsin Fig. 5A–C that no diffusion-controlledtransformation products such as ferrite orpearlite can be observed, and martensite isthe predominant feature. The predomi-nant formation of martensite in theCGHAZ is primarily due to the high alloyaddition in BA-160, which results in highhardenability as indicated by the high car-bon equivalent (CEAWS = 1.24) (Ref. 28).The CGHAZ microstructure forming at

different welding conditions was furtherinvestigated under higher magnificationTEM, as shown in Fig. 6A–C. A packet ofmartensite laths can be observed in Fig.6A. Because of the low carbon content inBA-160 (0.059 wt-%), the martensiteformed in the CGHAZ is lath martensite,which can also be confirmed by the exis-tence of a dislocation network within themartensite laths. The dark region betweenmartensite laths is retained austenite.

The existence of retained austenite re-sults from the high-nickel addition (6.8 wt-%) in BA-160, which is an effectiveaustenite stabilizer depressing the Ms andMf temperatures. It was determined thatfor BA-160 CGHAZ, Ms is 355°C and Mfis 178°C (Ref. 29). During the welding

cooling process, austenite rich in Ni didnot transform to martensite. It was stabi-lized to the ambient temperature andtherefore results in incomplete austenitetransformation to martensite, even underfast cooling rates. At higher heat input,lower bainite is formed in the CGHAZunder slower cooling rates, which can beconfirmed by the formation of intralathplate-like cementite precipitates withinthe bainite laths, as shown in Fig. 6B. Thecementite precipitates are oriented at apreferential angle with the primary bainitelath growth direction, and this is the char-acteristic feature distinguishing lower bai-nite from upper bainite or lath martensite(Ref. 30). Note that lower bainite repre-sents only a small fraction of the CGHAZmicrostructure, and because of the thinlaths and limited amount, it cannot be re-solved in the optical microscope. Marten-site laths free of intralath cementite canalso be seen in Fig. 6B and C, and it wasfound that less retained austenite waspresent with increasing heat input or ap-plying preheat. The average Vickers hard-ness of the CGHAZ with low heat input is370 HV1.0 and decreases to 358 HV1.0 withthe increase in heat input and to 363 HV1.0with applying preheat of 150°C. Due to thehardness decrease when using preheat, itis postulated that a small quantity oflower-hardness bainite may form in theCGHAZ when welding with preheat of150°C because of the lower cooling rate ascompared to welding with low heat input.

MARCH 2014, VOL. 93102-s

WELD

ING

RESEA

RC

H

Table 4 — Summary of Microstructure, Hardness, and Implant Test Results for Different Welding Conditions for BA-160

Welding Conditions Low HI High HI Low HI Low HI Low HI (33 kJ/in.) (66 kJ/in.) (33 kJ/in.), preheat (33 kJ/in.), preheat (33 kJ/in.), preheat

at 60°C at 100°C at 150°C

CGHAZ microstructure M M+B M M M+(B)a

CGHAZ Hardness (Avg)b 370 358 368 366 363Implant test LCS 91 ksi 96 ksi 94 ksi 103 ksi 107 ksi

(627 MPa) (661 MPa) (648 MPa) (710 MPa) (737 MPa)

(a) M represents martensite; B represents bainite; (B) represents possible formation of bainite in the microstructure.(b) Avg means the average hardness of the CGHAZ, which is determined by taking the average of hardness data points in the CGHAZ region

together. The approximate boundary of CGHAZ is determined by metallographic observation.

Fig 6 — Bright-field TEM micrographs for the CGHAZ of BA-160. A — Weld with low heat input; B — weld with high heat input; C — weld with low heatinput, preheat at 150°C before welding.

A B C

BA

Fig. 7 — Implant test results for BA-160. A — Weld with low heat input; B — weld with high heat input.

Yue (201339) March 2014_Layout 1 2/13/14 10:02 AM Page 102

Effect of Heat Input on BA-160 HAZHydrogen-Induced Cracking Tendency

Heat input is a factor that influences theHIC tendency. Different heat inputs will re-sult in different cooling rates and, as shownin Figs. 5 and 6, lead to the formation of dif-ferent microstructures in the HAZ duringcooling after welding. As discussed in theprevious section, the low heat input resultedin a harder microstructure as compared tothe high heat input.

The implant test results for the twoheat inputs are shown in Fig. 7A and B.The lower critical stress (LCS) for eachwelding condition was determined. Lowercritical stress is defined as the maximumstress at which cracking does not occurafter 24-h loading under the influence ofdiffusible hydrogen. Therefore, the higherthe LCS, the better is the resistance toHIC. It can be seen in the implant test re-sults that an incubation time is observed

before the final failure occurs at eachstress level, and represents the time takenfor microcracks to form under the influ-ence of stress and hydrogen that then linkup together to lead to the final failure.When interpreting the implant test re-sults, it is considered that for an equivalentexternal stress applied to the implant spec-imen, a longer incubation time indicatesbetter resistance to cracking (Ref. 31).The LCS values were determined to be 91ksi (627 MPa) and 96 ksi (661 MPa) forlow and high heat input, respectively, aslisted in Table 4. The increase in LCS whenusing high heat input indicates that in-creasing heat input when welding BA-160steel reduces the tendency for HIC in theCGHAZ. The reduction in cracking sus-ceptibility when using high heat input canalso be seen by the comparison of the in-cubation time before failure. When theapplied stress is 107 ksi (737 MPa), the in-cubation time for the low heat input is 17min, while it increases to 139 min for thehigh heat input.

Clearly, the reduced HIC susceptibilityfor the CGHAZ of BA-160 when increas-ing heat input is the difference in mi-crostructure as a result of slower coolingrates. The lower hardness microstructureof the mixed martensite and lower bainiteformed at slow cooling rates has a betterresistance to cracking as compared to the

103-sWELDING JOURNAL

WEL

DIN

G R

ESEA

RC

H

B

C

A

Fig. 8 — Implant test results for BA-160, weld withlow heat input. A — Preheat at 60°C; B — preheatat 100°C; C — preheat at 150°C.

A

A

C D

B

B

C

Fig. 9 — Fracture morphology of implant specimen, weld with low heat input, failed after 17 min undertensile stress of 107.9 ksi. A — General fracture appearance; B — intergranular; C — quasi-cleavage;D — microvoid coalescence.

Fig. 10 — Fracture morphology of implant speci-men, weld with high heat input, failed after 139 minunder tensile stress of 107.9 ksi. A — General frac-ture appearance; B — quasi-cleavage; C — mi-crovoid coalescence.

Yue (201339) March 2014_Layout 1 2/13/14 10:04 AM Page 103

higher hardness martensitic microstruc-ture formed at low heat input.

However, it should be noted that in realwelding practice, reducing HAZ HIC ten-dency by using a very high heat input mayeither not be practical or lead to severegrain coarsening, which may either reducetoughness or cause softening. As a result,the determination of heat input shouldalso consider other factors.

Effect of Preheat on Reducing BA-160HAZ Hydrogen-Induced CrackingTendency

Preheat and/or interpass temperaturecontrol are used as effective means in ac-tual welding practice to alleviate HAZHIC in steels. First, preheating the steelbefore welding will lead to a slower cool-ing rate, which generally results in the for-mation of lower hardness microstructuressuch as ferrite, pearlite, and/or bainite inthe HAZ. These transformation productsare less susceptible to HIC as compared tohard and brittle martensite. Second, pre-heat can drive off any preexisting moistureand promote longer weld cooling timesboth of which reduce the diffusible hydro-gen content in the HAZ. The reduction ofdiffusible hydrogen will thereby reducethe cracking tendency. Since BA-160 is de-signed as a structural steel for naval shipapplications, it is likely that high-restraint,high-hydrogen conditions will be encoun-tered during fabrication. Therefore, it isnecessary to investigate using the lab testif preheat is effective to alleviate HAZHIC when welding BA-160.

In the present study, implant testingwas conducted with preheat temperaturesof 60°, 100°, and 150°C to evaluate the pre-

heat effect on reducing the cracking ten-dency. The implant test results are shownin Fig. 8A–C. It can be seen that with pre-heat before welding, all three curves showthat the incubation time prior to failure islonger than that without preheat at equiv-alent tensile stress levels, as compared toFig. 7A, indicating applying preheat re-duces HAZ cracking tendency. This is be-cause applying preheat can reduce the dif-fusible hydrogen content, and thereby alonger incubation time is required toreach the critical hydrogen level to causecracking to occur. This is one evidence in-dicating that preheat is effective to allevi-ate the HAZ HIC cracking tendency forBA-160.

The LCS with preheating at 60°C wasdetermined to be 94 ksi (648 MPa), whichis slightly higher as compared to 91 ksi(627 MPa) without preheat. Increasing thepreheat temperature to 100°C, the LCSwas increased to 103 ksi (710 MPa), andthis was also with a concomitant increasein incubation time. By further increasingthe preheat temperature to 150°C, theLCS was increased to 107 ksi (737 MPa).This clearly indicates that increasing pre-heat temperature is beneficial to reduceHAZ cracking tendency for BA-160. How-ever, increasing preheat temperature to150°C to reduce cracking was not so effec-tive compared with the 9-ksi LCS increasefrom preheat at 60° to 100°C. Based on theimplant test results, it is shown that apply-ing preheat and increasing preheat tem-perature are beneficial to reduce thecracking tendency for BA-160. One rea-son is that applying preheat results in alower cooling rate, which leads to the for-mation of a lower hardness microstructurein the CGHAZ as found in the Vickershardness test results. The other reason isthat preheat is effective to reduce the dif-fusible hydrogen level in the HAZ as dis-cussed previously.

Effect of Welding Parameters on FractureBehavior

The fracture surfaces of the implantspecimens welded under different condi-tions were studied using SEM to analyzethe effect of welding parameters on frac-ture behavior. The SEM fractographs ofthe BA-160 implant specimens are pre-sented in Figs. 9–11. For the three weldingconditions, the fracture surface studiedwas from the implant specimens failed atthe same tensile stress (107.9 ksi) for thepurpose of comparison.

Figure 9A shows the general fractureappearance of a BA-160 implant specimenwith low heat input, and regions with dif-ferent fracture modes are shown in Fig.9B–D at higher magnification. Coarse in-tergranular (IG) failure, as shown in Fig.9B, is only observed in the region close tothe root of the thread where the crack ini-tiates. A quasi-cleavage (QC) mode, asshown in Fig. 9C, constitutes the majorityof the fracture surface. Microvoid coales-cence (MVC) is observed in only a smallarea close to the final failure region, asshown in Fig. 9D.

The fracture morphology of a speci-men from a high-heat-input weld is shownin Fig. 10A–C. There is no faceted, coarseIG fracture observed on the fracture sur-face, in contrast to the fracture surfacewith low-heat input. Quasi-cleavage is alsothe predominant feature, as shown in Fig.10A and B. However, the region of MVCbecomes larger as compared to that of lowheat input.

Figure 11A–C shows the fracture mor-phology with low heat input and a preheatlevel of 100°C. Similar to Fig. 10A, no ob-vious IG failure can be observed at thecrack initiation site. The crack initiatesand propagates with QC mode for a longdistance, and then the fracture modechanges to MVC.

It is found that using both high heatinput and preheat eliminate the IG frac-ture mode, and increase the MVC area onthe fracture surface as compared to thelow-heat-input condition. This is an indi-cation that increasing heat input and usingpreheat improve HIC resistance. And thisis the result of formation of lower hard-ness microstructure in the CGHAZ withdecreasing cooling rate, and also becauseof the preheating effect to drive off hy-drogen from the weld pool.

Conclusions

1. Using low-heat-input welding condi-tions (33 kJ/in.), the CGHAZ microstruc-ture of BA-160 is lath martensite with re-tained austenite. With high heat input (66kJ/in.), lower bainite forms in theCGHAZ. The mixture of martensite andlower bainite has a lower hardness as com-

MARCH 2014, VOL. 93104-s

WELD

ING

RESEA

RC

H

A

C

B

Fig. 11 — Fracture morphology of implant speci-men weld with low heat input and preheat of100°C, failed after 80 min under tensile stress of107.9 ksi. A — General fracture appearance; B —quasi-cleavage; C — microvoid coalescence.

Yue (201339) March 2014_Layout 1 2/13/14 10:08 AM Page 104

pared to the martensitic microstructureformed with low heat input.

2. When preheat of 150°C is applied inconjunction with the low-heat-input weld-ing conditions, the hardness is decreasedand the microstructure is a mixture ofmartensite and a possibly small fraction ofbainite.

3. Implant test results show that atequivalent stress levels, the incubationtime prior to failure is longer when highheat input is used as compared to low heatinput. The lower critical stress (LCS) wasfound to increase from 91 ksi (627 MPa)with low heat input to 96 ksi (661 MPa)with high heat input. Both the longer in-cubation time present and increase in theLCS indicate that using higher heat inputreduces the HAZ cracking tendency forBA-160.

4. The LCS was determined to be 94,103, and 107 ksi (648, 710, and 737 MPa)with preheat levels of 60°, 100°, and 150°C,respectively. The LCS with preheat ishigher than that without preheat. And theincubation time at equivalent tensile stressis also longer than that without preheat.Both indicate preheating can effectivelyreduce the tendency for HIC in BA-160.

5. Intergranular, quasi-cleavage, andmicrovoid coalescence fracture modeswere observed on the fracture surface ofimplant samples from low-heat-inputwelds. Only quasi-cleavage and microvoidcoalescence were observed on the fracturesurface when using high heat input and ap-plying preheat. The elimination of IG andincrease in MVC area on the fracture sur-face correlated well with LCS values andindicated better resistance to cracking.

Acknowledgments

The authors gratefully acknowledgethe financial support of the Office ofNaval Research, Award No.N000140811000. Grant Officers: Dr. JulieChristodoulou and Dr. William Mullins.The authors would like to thank JohnnieDeLoach, Matthew Sinfield, and JeffreyFarren with the Naval Surface WarfareCenter Carderock Division, WestBethesda, Md., for valuable discussionsregarding the weldability of naval steels.Thanks are extended to Prof. GregoryOlson’s research group at NorthwesternUniversity for collaboration on this re-search project and QuesTek InnovationsLLC for providing the BA-160 steel. Dr.Yuping Yang with EWI is acknowledgedfor providing access to the E-Weld Predic-

tor™. Dejian Liu and Geoffrey Taber arethanked for their constructive ideas andassistance with building the implant test-ing system.

References

1. Czyryca, E. J. 1993. Advances in highstrength steel technology for naval hull con-struction. Key Engineering Materials 84-85:491–520.

2. Czyryca, E. J., Link, R. E., Wong, R. J.,Aylor, D. A., Montemarano, T. W., and Gudas,J. P. 1990. Development and certification ofHSLA-100 steel for naval ship construction.Naval Engineers Journal 102(3): 63–82.

3. Montemarano, T. W., Sack, B. P., Gudas,J. P., Vassilaros, M. G., and Vanderveldt, H. H.1986. High strength low alloy steels in navalconstruction. Journal of Ship Production 2(3):145–162.

4. 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.

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

6. 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.

7. Gedeon, S. A., and Eagar, T. W. 1990.Thermochemical analysis of hydrogen absorp-tion in welding. Welding Journal 69(7): 264-s to271-s.

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

9. Rowe, M. D., Nelson, T. W., and Lippold,J. C. 1999. Hydrogen-induced cracking alongthe fusion boundary of dissimilar metal welds.Welding Journal 78(2): 31-s to 37-s.

10. Park, Y. D., Maroef, I. S., Landau, A.,and Olson, D. L. 2002. Retained austenite as ahydrogen trap in steel welds. Welding Journal81(2): 27-s to 35-s.

11. Brauser, S., and Kannengiesser, Th.2010. Hydrogen absorption of different weldedduplex steels. International Journal of HydrogenEnergy 35: 4368–4374.

12. Maroef, I., Olson, D. L., Eberhart, M.,and Edwards, G. R. 2002. Hydrogen trapping inferritic steel weld metal. International MaterialsReviews 47(4): 191–223.

13. 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.

14. Cullison, A. 1994. Two paths, one goal:A consumable to weld HSLA 100. Welding Jour-nal 73(1): 51–53.

15. Moon, D. W., Fonda, R. W., and Spanos,G. 2000. Microhardness variations in HSLA-

100 welds fabricated with new ultralow-carbonweld consumables. Welding Journal 79(10): 278-s to 285-s.

16. Petch, N. J., and Stables, P. 1952. De-layed fracture of metals under static load. Na-ture 169(4307): 842–843.

17. Troiano, A. R. 1960. The role of hydro-gen and other interstitials in the mechanical be-havior of metals. Transactions of American So-ciety for Metals 52: 54–80.

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

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

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

21. 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.

22. 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.

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

24. Nawrocki, J. G., DuPont, J. N., Robino,C. V., and Marder, A. R. 2001. The postweldheat treatment response of simulated coarse-grained heat-affected zones in a new ferriticsteel. Metallurgical and Materials Transactions A32A: 2585–2594.

25. 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.

26. Spanos, G., Fonda, R. W., Vandermeer,R. A., and Matuszeski, A. 1995. Microstructuralchanges in HSLA-100 steel thermally cycled tosimulate the heat-affected zone during welding.Metallurgical and Materials Transactions A 26A:3277–3293.

27. Zhang, W., and Yang, Y. P. 2009. Devel-opment and application of online weld model-ing tool. Welding in the World 53(1/2): 1–9.

28. Yue, X., Feng, X. L., and Lippold, J. C.2013. Quantifying heat-affected zone hydro-gen-induced cracking in high-strength navalsteels. Welding Journal 92(9): 265-s to 273-s.

29. 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.

30. Bhadeshia, H. K. D. H. 2001. Bainite insteels: Transformations, microstructure andproperties. pp. 63–75. London, UK: IOM Com-munications Ltd.

31. 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.

105-sWELDING JOURNAL

WEL

DIN

G R

ESEA

RC

H

Yue (201339) March 2014_Layout 1 2/13/14 10:09 AM Page 105