Study on hydrogen assisted cracking susceptibility of HSLA steel by implant test

Gop CHAKRABORTY,R.REJEESH,S.K.ALBERT

aMaterials Technology Division,Indira Gandhi Centre for Atomic Research,Kalpakkam,603102,India

bNational Institute of Technology,Surathkal,India

Study on hydrogen assisted cracking susceptibility of HSLA steel by implant test

Gopa CHAKRABORTYa,*,R.REJEESHb,S.K.ALBERTa

aMaterials Technology Division,Indira Gandhi Centre for Atomic Research,Kalpakkam,603102,India

bNational Institute of Technology,Surathkal,India

DMR-249A is an indigenously developed high strength low alloy steel for Indian ship building industry for making ship-hull and is extensively used in the construction of war ships and submarines.Welding electrodes conforming to SFA 5.5AWS E8018 C1 has been indigenously developed for welding of this steel using shielded metal arc welding process.In the present study,susceptibility to hydrogen assisted cracking of DMR-249A steel welds made using this electrode has been assessed using implant test.Implant tests were conducted using this electrode at two different levels of diffusible hydrogen,measured using gas chromatography technique.It is observed that both the steel and the welding consumable are not susceptible to hydrogen assisted cracking even with a high diffusible hydrogen level of 9 mL/100g of weld metal.In implant tests,specimen did not fracture even after loading to stress levels higher than the yield strength of the base metal.The good resistance of this steel and the welding consumable,even with high levels of diffusible hydrogen,is attributed to absence of a susceptible microstructure in both the weld metal and heat affected zone.Hence,this study shows that,in the absence of a susceptible microstructure,hydrogen assisted cracking is unlikely to occur even if hydrogen level is high.It also confirms that in welding of DMR-249A with indigenously developed E8018 C1 electrode,hydrogen assisted cracking is not a concern and no preheating is required to avoid it during welding.

HSLA steel;Hydrogen assisted cracking;Diffusible hydrogen;Implant test;Lower critical stress

1.Introduction

Microalloyed high strength low alloy(HSLA)steels containing low carbon and small additions of Nb,V,Ti exhibit an outstanding combination of high strength,resistance to brittle fracture and good weldability[1].DMR-249A is a low carbon HSLA steel with micro additions of Nb,V andTi,indigenously developed for Indian ship-building industries and is being used in the construction of war ships and submarines[2].Shielded metal arc welding(SMAW)is one of the major welding processes employed by shipping industries.Complex dynamic loading,extreme temperature conditions during service together with residual stresses generated in the weld due to fit up and fabrication can make these weld joints susceptible to brittle fracture in service[3].The presence of undetected cracks caused by hydrogen assisted cracking(HAC)during fabrication can further assist brittle fracture in service.Hence,there is a need to assess the susceptibility of these welds to HAC and ensure that there is no risk of HAC during fabrication of naval structures using this steel and consumables[4].

Earlier studies[5]have shown that the conditions for HAC to occur in steel welds are:presence of diffusible hydrogen(HD), residual stress and susceptible microstructure in the weld and temperature in the range of ambient to 200°C.In this regard, martensiticmicrostructurewithhighhardnessismostsusceptible and ferritic microstructure with low hardness is least susceptible. Hence,duringwelding,effortsaremadetoreduceriskofHACby avoiding development of a susceptible microstructure and minimizing the hydrogen levels in welding.The probability of havingasusceptiblemicrostructureintheHAZorweldisassessed fromthecompositionofthebasemetalandweldmetal,heatinput and preheating(which will reduce the cooling rate of the weld) chosen for welding[6].In order to reduce hydrogen level,low hydrogen welding consumables,proper baking of the consumables to remove moisture content in the consumables and appropriate preheating or preheating+post heating conditionsthat would provide more time for hydrogen to diffuse out at high temperature are chosen.For HSLA steels like DMR-249A,the hardenability is very low and the as-welded microstructure is ferritic and hence susceptibility to HAC is expected to be low. However,susceptibility of a weld to HAC can be quantified from implanttestintermsoflowercriticalstress(LCS),thestressbelow whichthewelddoesnotfractureduringthetest,andinthepresent study this is attempted for welds of DMR-249A steel made with indigenously developed consumables.

In order to achieve two different levels of hydrogen in the weld metals,welds for implant tests were made both with baked electrodes and unbaked electrodes.Baking brings down the moisture content in the flux coating which in turn reduces the diffusible hydrogen content in the welds.Hence,during fabrication using low hydrogen welding consumables,they are baked prior to use as per the instruction provided by manufactures.In the present study,implant tests were conducted using electrodes with and without baking to produce welds with different levels of diffusible hydrogen during implant testing.Diffusible hydrogen levels in electrodes were measured using thermal conductivity based gas chromatography.

2.Experimental

DMR-249A is a low carbon(C:0.09,Mn:1.14,Si:0.18, Ni:0.62,Al:0.026,Nb:0.039,V:0.02,Ti:0.02,S:0.006,P:0.14, N:56 ppm)HSLA steel with minimum yield strength and tensile strength of 390 MPa and 510 Mpa,respectively.AWS E8018 C1 is a basic coated low hydrogen electrode.Nominal chemical composition of the electrode(Ys:483 MPa,UTS: 552 MPa)is given in Table 1.

2.1.Measurement of diffusible hydrogen content in the weld

For diffusible hydrogen measurement,DMR-249A steel samplesarefabricatedasperISO3690specification.Thespecimen of size 30 mm×15 mm×10 mm is fixed in a copper jig with run-on and run-off pieces each of size 40 mm×15 mm×10 mm. Beadonplateweldingwascarriedoutwith3.15 mmdiametergrade SFA5.5AWSE8018C1electrode.Weldingparametersusedforthe weldingare:current–110A,voltage–23V,weldingtime–～30 s, weldinglength–～70 mm.Approximateheatinputcorresponding to the welding parameters is～1100 J/mm.Diffusible hydrogen measurementswerecarriedoutforthreedifferentconditionsofthe electrodes:(1)baking the electrode at 450°C for 4 hr and maintaining the electrode temperature at 150°C after baking prior to welding;(2)baking the electrode at 150°C for 4 hr prior to welding;and(3)without any baking.It is to be noted here that the normalpracticeofweldingistobaketheelectrodepriortowelding, which gives a minimum level of diffusible hydrogen.However,in the present study diffusible hydrogen in the electrodes in the as-receivedcondition(nobaking)aswellasintheelectrodesbaked at lower temperature was also measured.The objective was to prepare implant test specimens with different levels of diffusible hydrogen in the welds using these electrodes.Immediately after completion of the welding,the specimens for diffusible hydrogen measurement were immersed in ice cold water for 5 s and kept inside liquid nitrogen to cool to subzero temperature until they are taken out for hydrogen extraction and measurement.

Fig.1.Schematic diagram of implant test specimen.

The HE_GCTCD set up used for diffusible hydrogen measurement consists of a diffusible hydrogen collection chamber,a heater to heat the chamber and a gas chromatograph (GC).The detail of HE_GCTCD set up and diffusible hydrogen measurement technique is provided elsewhere[7].The specimen iskeptinsidethechamberat400°Cfor30minutesforextraction ofdiffusiblehydrogenfromthetestspecimen.Hydrogencollected inthechamberistransportedtoaGCwithathermalconductivity detectorusingArascarriergasandthesignalisrecorded.Priorto measurement,GCiscalibratedusingknownvolumesofhydrogen injected into the GC and from this,the volume of hydrogen evolved from the weld specimen and collected in the chamber is estimated.Usingtheweightofthedepositedmetalintheweld,the volumeofdiffusiblehydrogeniscalculatedinmillilitersper100g of deposited weld metal.For each condition,three tests were performed and average of the data is reported.

2.2.Implant test

Fig.1 shows the schematic diagram of implant specimen and base plate,prepared as per Doc.IIW-802 guidelines[8].The implant testing machine is a computer controlled and mechanically operated machine along with a load-cell attached to it to display the load and time duration during the testing[8].The specimen assembly consists of a base plate with a hole,into which implant specimen is inserted in such a way that the top surface of the implant specimen and base plate are at the same level.Single pass bead on plate welding was made on this specimen assembly using the test electrode and employing the same welding parameters used for making the specimens for diffusible hydrogen measurement in such a way that the weld bead passes over the implant specimen fusing its top surfacecompletely with the base plate.Two separate sets of test were conducted using the specimens prepared with baked(at 450°C for 4 hr)and unbaked electrodes.A thermocouple was attached to the base plate to monitor the temperature and loading was done when the assembly cools down to 100°C.A series of tests with first specimen loaded at 1000 kg were conducted.Subsequently,loading was increased to 2000 kg in steps of 200 kg and later on in steps of 100 kg until failure of the sample.Two tests were repeated for each loading condition.After implant test,selected samples were sectioned to observe for micro cracks.The samples were sliced,metallographically polished and etched with 2%nital solution to study under optical microscope.Microhardness measurements were performed on base metal,HAZ and weld metal at 500 g load.

Table 1 Chemical composition(Wt%)of weld metal.

Fig.2.Schematic diagram of(a)notch tensile test and(b)tensile test specimen.

2.3.Notch tensile test and impact tests of simulated HAZ specimens

Implant test specimens contain notch in the HAZ produced by the weld.Hence,in order to compare the LCS with notch tensile strengthoftheHAZ,notchtensiletestsoftheHAZwereconducted. For this purpose,HAZ was simulated on a plate of size 100 mm×150 mm×10 mm by heating up to 1080°C and then coolinginair,basedonHAZmicrostructureobservedfromimplant samples and available literature reference for this steel[9].Notch tensilesampleswerefabricated(Fig.2(a))fromthesimulatedHAZ and tested in tensile testing machine under 10-4/s strain rate.For comparisonpurpose,tensiletestingofthebasemetal(Fig.2(b))was also done at similar strain rate.Standard Charpy“V”notch impact testing samples were fabricated from the simulated HAZ material andimpacttestingwasdoneatroomtemperature.Fracturesurface ofthefailedsamplesaftertensileandCharpytestingwasobserved under scanning electron microscope(SEM).

3.Results

The HE_GCTCD data indicate that diffusible hydrogen(HD) content in the welding consumable,after baking at 450°C/4h, is 3.1 mL/100 g of weld metal.This is certainly a low value of diffusible hydrogen,and as per IIW and AWS classifications, this electrode comes under very low hydrogen category of electrode[10–12].Results of the HDmeasurement for the electrode without any baking is 9.6 mL/100 g of weld metal and the same for baking at 150°C is 8.3 mL/100 g of weld metal.Thus,by altering the baking conditions of the electrodes,one can get different levels of HDcontents for the same batch of electrodes. This enables carrying out implant tests using electrodes differing only in their HDcontent and determining LCS at these levels of HDcontents for the same consumables and base materials.

As mentioned earlier,implant test was carried out for two sets of welding prepared with baked(450°C)and unbaked electrode. The tests were conducted in the range of 1000 kg(equivalent stress:205 MPa)to 2000 kg(410 MPa)load in steps of 200 kg and above 2000 kg;load was increased up to 2400 kg(490 MPa) in a step of 100 kg.For samples welded with baked electrode,no failureoccurredwithin24 h(Fig.3(a))uptoaloadingof2400 kg (490 MPa).No further test was carried out above 2400 kg (490 MPa)sinceitisabovetheyieldstrengthofthebasemetal[9]. To estimate the stress level at which specimen fracture,for one specimenloadwasincreaseduntilfractureandthisoccurredatthe basemetal(Fig.3(b))farawayfromtheHAZandthecorresponding fracture stress is 554 MPa(load=2700 kg),which is nearly equivalenttothetensilestrengthoftheweldjoint[9].AsLCScould not be determined for the properly baked electrode,implant tests were conducted for specimens prepared using electrodes without baking(HDlevels=9.6 mL/100 gofweldmetal).Thesespecimensalso did not fracture even after loading up to 2400 kg(490 MPa). Theimplantspecimenstestedatthehigheststresslevelsweresliced along the length,polished,etched and then examined for microcracks under optical microscope.No cracks were found as shown in Fig.4.Thus,results clearly confirm that the steel is not susceptible for HAC even at high levels of diffusible hydrogen.

Fig.3.Implant sample tested at(a)2400 kg and(b)2700 kg loads.

Fig.4.Micrograph of implant sample welded with(a)baked and(b)unbaked electrodes.

The optical micrographs of DMR-249A steel base metal,weld metalandHAZareshownin Fig.5(a)-(c).Thebasemetalconsists offinegrainedequiaxedferriteandsomepercentageofpearliteasa banded structure(Fig.5(a)).Micrograph of weld metal shows fine bainitic structure along with acicular ferrite(Fig.5(b)).The HAZ microstructure consists of acicular ferrite with some polygonal ferrite(Fig.5(c)).No martensitic phase could be identified in the HAZorintheweldmetal.HardnessoftheHAZ(275VHN)isfound tobemarginallyhigherthanthebasemetal(235VHN).Weldmetal hardness(315VHN)is higher than that of HAZ(Fig.6(a)).

Fig.5.Optical micrographs of(a)base metal,(b)weld metal and(c)HAZ.

Fig.6.(a)Hardness profile of the weld joint and(b)stress–strain diagram for base metal and HAZ(notch-tensile specimen).

Mechanical properties of the simulated HAZ are given in Table 2 and Fig.6(b).Since notch effect is experienced by the HAZ of the implant sample due to presence of the helical notch, the notch tensile strength of the simulated HAZ was determined and the value obtained for the same is 660 MPa,which is higher than the tensile strength of the base metal(575 MPa),revealing the strengthening effect of the notch.Toughness of the HAZ simulated structure is appreciably high(170J),although much lesser than base metal(350J)[13].Fractographs(Fig.7(a)and (b))also show ductile cup-cone fracture for tensile tested samples of both base metal and HAZ.Dimple size in case of base metal is much smaller as compared to HAZ,indicating high ductilityof thematerial.Fibrous ductile fracture(Fig.7(c)) is also noted for the impact tested specimen of simulated HAZ.

4.Discussion

In the present investigation,the results indicate that HSLA steel of grade DMR-249A is not susceptible to HAC irrespective of HDcontent of the welding consumable.The reason for the same is investigated in the following section.

It is clear from the results presented above that without baking, HDcontentinthefluxoftheelectrodeisquitehigh,andbybakingat 150°Conlythemoistureabsorbedbythefluxcoatingisdrivenoff, whereas by baking at a temperature recommended by the manufacturer(450°C),chemically bonded water in some of the flux constituents is also removed.During arc welding processes, hydrogen gets introduced into the weld from the moisture of the atmosphereaswellasflux,andfromhydrogenousmaterialssuchas oil,grease,paint,etc.[14].Themoisturecanberemovedbydrying at moderately higher temperature whereas elevated temperature (above400°C)isrequiredtodrivethechemicallyassociatedwater of the flux[14].

It can be further seen from the results that LCS could not be determined for weldments produced from either properly baked or unbaked electrode by implant tests as the LCS values areabove the yield strength of the material in both the conditions. It can be assumed that fully ferritic microstructure of the HAZ, similar to base metal,and predominantly ferritic structure of the weld metal are the major reasons for good resistance of the weld joint to HAC.It has also been seen for other grades of HSLA steel that ferritic microstructures are resistant to HAC irrespective of HDcontent of the weld[15].Another factor contributing to HAC resistance could be the presence of carbides,especially TiC in the weld metal,which are known to be strong traps for hydrogen[16].Though Rishi et al.[13]have characterized in detail the inclusion content of SMAW weld metal of DMR-249A steel with similar electrode,detailed study on effectiveness of various precipitates and inclusion as hydrogen traps has not been attempted.Such a study may reveal more information about high resistance to HAC of these welds.

Table 2 Mechanical test results of DMR-249A steel.

The tendency of HAC is much higher in weldments with a stringent variation in hardness from base metal to weld or HAZ [8].However,in this case variation in hardness between HAZ (275 VHN)and base metal(235 VHN)is only marginal due to the presence of predominantly ferritic microstructure.Weld metal hardness(315 VHN)is slightly higher probably because of the presence of bainite in the weld metal.However,adverse effect of high diffusible hydrogen content is much nullified by the almost uniform hardness distribution across the weldment [7].Mechanical properties of the simulated HAZ also indicate that fine acicular ferritic structure of HAZ contributes to its good mechanical properties.From the notch tensile strength of the HAZ and fractographs of the same,it can be concluded that a ferritic microstructure of low hardness and high toughness is resistant to HAC,irrespective of the hydrogen level in the weld.

In this context,susceptibility to HAC of DMR-249A can be compared to that of modified 9Cr-1Mo steel.For modified 9Cr-1Mo steel,the LCS is reported to be 185 MPa corresponding to HDcontent of 3.7 mL/100 g of weld metal,which is considerably lower than the yield strength of the material (1000 MPa)[8].With preheat to a temperature of 250°C before welding,HDcontent comes down to 1.8 mL/100 g of weld metal;however,LCS increases only to 267 Mpa,which is still much below the yield strength of the steel[8].In contrast to this,in spite of high hydrogen levels LCS could not be determined for DMR-249A grade of steel because of the good resistance of the steel to HAC.Results emphasize the point thatin the absence of a susceptible microstructure,high hydrogen content may not cause HAC[15].This comparison clearly indicates that even very low hydrogen level can cause HAC in a martensitic microstructure as in case of modified 9Cr-1Mo steel;but if it is a ferritic microstructure,even high hydrogen level does not cause HAC.

Fig.7.Fractograph of(a)impact tested base material,(b)tensile tested base material and(c)notch tensile tested HAZ.

5.Conclusions

The following conclusions can be made from the present study:

1)HSLA steel grade DMR-249A is not susceptible to HAC.

2)Thefineacicularferritic/bainiticmicrostructureproducedin HAZ and weld can adequately resist HAC in this steel.

3)In absence of a susceptible microstructure,HAC is unlikely to occur even if hydrogen levels in the welding consumables are high.

[1]Show BK,Veerababu R,Balamuralikrishnan R,Malakondaiah G.Effect of vanadium and niobium modification on the microstructure and mechanical properties of a microalloyed HSLA steel.Mater Sci Eng A 2010;527:1595–604.

[2]RodriguesPCM,PerelomaEV,SantosDB.MaterSciEng A 2000;283:136–43.

[3]Yue X,Lippold JC.Evaluation of heat affected zone hydrogen induced cracking in navy steels.Welding Res Suppl 2013;92:20s–28s.

[4]Devletian JH,Fichtelberg ND.Controlling hydrogen cracking in ship building.Welding J 2011;80:46s–52s.

[5]Dickinsons DW,Ries GD.Implant testing of medium to high strength steel-a model for predicting delayed cracking susceptibility.Welding Res Suppl 1979;205s–211s.

[6]Padhy GK,Komizo Y.Diffusible hydrogen in steel weldments-a status review.Trans JWRI 2013;42:39–62.

[7]Padhy GK,Ramasubbu V,Albert SK,Murugesan N,Ramesh C.Hot extraction of diffusible hydrogen and its measurement using a hydrogen sensor.Welding World 2012;56:7–8.

[8]Albert SK,Ramasubbu V,Sunder Raj SI,Bhaduri AK.Hydrogen assisted cracking susceptibility of modified 9Cr-1Mo steel and its weld metal. Welding World 2011;7–8.

[9]Pamnani R,Vasudevan M,Jayakumar T,Vasantharaja P,Ganesh KC. Numerical simulation and experimental validation of arc welding of DMR-249A steel.Def Technol 2016;12:305–15.

[10]ISO 2560:2009,Welding consumables-covered electrodes for manual arc welding og non-alloy and fine grain steels-Classification.

[11]Specification for carbon steel electrodes for shielded metal arc welding, American National Standard,AWS A5.1/5.1M:2004,approved by American National Standards Institute;2003.

[12]Specifications for low alloy steel electrodes for shielded metal arc welding,American National Standard,AWSA5.5/A5.5M:2006,approved by American National Standards Institute;2006.

[13]Pamnani R,Jayakumar T,Vasudevan M,Sakthivel T.Investigation on the impact toughness of HSLA steel arc welded joints.J Manuf Process 2016;21:75–86.

[14]Magudeeswaranan G,Balasubramaniana V,Madhusudhan Reddy G. Hydrogen induce cold cracking studies on armour grade high strength, quenched and tempered steel weldments.Int J Hydrogen Energy 2008;33:1897–908.

[15]Madhusudhan Reddy G,Mohandas T,Sarma DS.Cold cracking studies on low alloy steel weldments:effect of filler metal composition.Sci Technol Welding Joining 2003;8:407–14.

[16]Depover T,Monbaliu O,Wallaert E,Verbeken K.Effect ofTi,Mo and Cr based precipitates on the hydrogen trapping and embrittlement of Fe-C-X Q&T alloys.Int J Hydrogen Energy 2015;40:47.

Received 30 August 2016;revised 21 September 2016;accepted 22 September 2016 Available online 7 October 2016

Peer review under responsibility of China Ordnance Society.

*Corresponding author.Tel.:+914424780500.

E-mail address:gopa_mjs@igcar.gov.in(G.CHAKRABORTY).

http://dx.doi.org/10.1016/j.dt.2016.09.003

2214-9147/?2016 The Authors.Production and hosting by Elsevier B.V.on behalf of China Ordnance Society.This is an open access article under the CC BY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).

?2016 The Authors.Production and hosting by Elsevier B.V.on behalf of China Ordnance Society.This is an open access article under the CC BY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).