Effect of GMAW Process and Material Conditions on DP 780

WELDING RESEARCH

-s135WELDING JOURNAL

ABSTRACT. The drive to reduce vehicleweight and improve crash performancehas led automotive manufacturers to in-troduce higher-strength grades of ad-vanced high-strength steels (AHSS). Forthese materials to be used effectively, theinfluence of material and process condi-tions on gas metal arc (GMA) weld prop-erties must be understood. The objectiveof this work was to characterize the effectsof material prestrain, cooling rate condi-tions (welding heat input and fixture heatsink), filler metal selection, dilution, andpostbaking on the microstructure and me-chanical properties of GMA welds oncoated dual-phase (DP) and transforma-tion-induced plasticity (TRIP) steels. Theprimary materials studied were DP 780and TRIP 780; for comparison purposes alimited amount of work was conductedwith DP 980. The DP steels showed vary-ing degrees of heat-affected zone (HAZ)hardening and softening depending on thematerial grade, prestrain, and cooling ratecondition. The relatively high aluminumcontent of the TRIP 780 allowed retainedferrite to be present in all regions of theHAZ, along with a continuous region ofcoarse ferrite along the weld interface.This resulted in the TRIP 780 havinglower peak HAZ hardness than the DP780. Fusion zone microstructure and hard-ness were found to be affected by the basemetal chemistry, the cooling rate condi-tion, and the filler metal composition.Filler metal strength did not affect the sta-tic or dynamic tensile properties of eitherthe TRIP 780 lap or butt joint welds, or theDP 780 butt joint welds. All of the TRIP780 and DP 780 butt joints failed in thesoft HAZ. The results of the lap joint tests

showed a greater variation in strength thatis attributed to porosity at the root of theweld.

Introduction

The use of AHSS has increased as ameans of reducing vehicle weight throughthe use of thinner material gauges (Ref.1). Future applications will likely requirecoated DP and TRIP steels with thick-nesses of less than 2.0 mm (0.08 in.) andstrength levels greater than 700 MPa(101.5 ksi). Gas metal arc welding(GMAW) is often employed where partgeometry prevents the use of resistancespot welding (RSW) or when the designrequires additional joint strength and stiffness.

To use AHSS effectively, it is importantfor designers and manufacturing engi-neers to understand the factors that mayaffect the performance of GMA-weldedstructures. This paper primarily exploresthe effects of common manufacturingvariations on the microstructure and me-chanical properties of coated DP 780 andTRIP 780. The effects of the followingmanufacturing variations were evaluated:

• Material Prestrain. Various degreesof strain may be imparted in the sheetmetal during stamping or forming opera-

tions prior to welding. • Filler Metal Type. Mild steel sheet

applications typically employ commonelectrode compositions (e.g., ER70S) thatare designed to produce weld metal hav-ing an ultimate tensile strength of at least70 ksi (483 MPa). Higher-strength gradesmay be required to weld some AHSSgrades in order to match the strength ofthe base metal.

• Cooling Rate Conditions. Weld cool-ing rates can be affected by the weldingprocess parameters (i.e., welding heatinput) as well as heat sinking from toolingor other adjacent materials.

• Dilution. Depending on the joint de-sign and welding process parameters, theweld fusion zone alloy can consist of dif-ferent fractions of base metal and fillermetal.

• Postbaking. Components may bepainted and baked after welding.

This work allows the relative impor-tance of these factors on the weld mi-crostructure and properties of DP 780 andTRIP 780 to be assessed. A limitedamount of work was conducted withcoated DP 980; the results of which arelisted in table format for comparison purposes.

Experimental Approach

Mechanized GMA welds were pro-duced on coated steels under a range ofconditions, including:

• Material Grades. Table 1 is a list ofthe steel grades that were evaluated, alongwith the nominal gauge thickness andcoating type of each. Table 2 lists thechemical composition of each steel. Thechemical compositions of the DP 780 andTRIP 780 steels were determined by in-ductively coupled plasma (ICP) analysis,

SUPPLEMENT TO THE WELDING JOURNAL, JUNE 2008Sponsored by the American Welding Society and the Welding Research Council

KEYWORDS

AHSSCooling RateDP SteelsFiller MetalGMAWGas Metal ArcTRIP Steels

Effect of GMAW Process and MaterialConditions on DP 780 and TRIP 780 Welds

The effects of variables on microstructure and mechanical propertieswere explored for advanced high-strength steels

BY N. KAPUSTKA, C. CONRARDY, S. BABU, AND C. ALBRIGHT

N. KAPUSTKA, C. CONRARDY, and S. BABUare with Edison Welding Institute, Columbus,Ohio. C. ALBRIGHT is with The Ohio State Uni-versity, Columbus, Ohio.

Kapustka 6 08corr:Layout 1 5/5/08 3:01 PM Page 135

WELDING RESEARCH

JUNE 2008, VOL. 87-s136

while the composition of the DP 980 steelis that which was provided in the manu-facturer’s certification.

• Joint Types. Lap joint fillet weldswere produced for HAZ characterization.Both lap joint fillet welds and squaregroove butt joint welds were produced forfusion zone and mechanical propertycharacterization. As shown in Fig. 1, thelap joints had 25.4 mm (1.0 in.) of overlap.Lap joints were produced using 140- ×203-mm (5½- × 8-in.) coupons, while buttjoints used 127- × 203-mm (5- × 8-in.)coupons.

•Prestrain. Two conditions were used:as-received and prestrained via roll reduc-tion in thickness of approximately 8%.

• Filler Metal. Two filler metals wereused: a nominally “low”-strength elec-trode (ER70S-6) and a nominally “high”-strength electrode (ER100S-G).

• Heat Sink Effects. Welds were pro-duced at nominally high and low coolingrate conditions. High cooling rate wasachieved with short-circuit transfer and acopper backing bar. Low cooling rate wasachieved with spray transfer and no back-ing bar. Table 3 lists typical welding para-meters and heat sink conditions used toproduce welds with a nominally high orlow cooling rate. Parameters were devel-oped to achieve adequate fusion withoutexcessive penetration (i.e., no melt-through on lap joints). For butt jointwelds, the major acceptance criteria wasfull penetration, without excessive root re-inforcement. The fixture (Fig. 2) had a ½-in.-deep × 1-in.-wide channel along theweld centerline. A copper bar was insertedinto the channel for the high heat sinkwelds. For the no heat sink welds, the cop-per bar was removed. A computerized

data acquisition system was employed tomeasure the current, voltage, and time ofeach weld; travel speeds were determinedby dividing the weld length by the weldtime. The average theoretical heat inputof each weld was calculated to provide anindication of the actual heat input. Be-cause heat input calculations were deter-mined for comparative purposes, a heattransfer efficiency of 1.0 was used.

• Postbake. Samples were evaluated

Fig. 1 — Reduced cross-sectional geometry of lapjoint tensile specimens.

Table 2 — Base Metal Compositions

Element TRIP 780 DP 780 DP 980

Carbon 0.17 0.13 0.112Manganese 2.08 2.01 2.45

Silicon 0.023 <0.005 0.028Phosphorous 0.011 0.016 0.013

Sulfur <0.001 0.002 0.004Copper 0.034 0.026 0.02Nickel 0.021 0.015 0.01

Molybdenum 0.055 0.20 0.325Chromium 0.094 0.23 0.24Niobium <0.001 <0.001 0.004

Vanadium 0.002 0.003 0.001Titanium 0.010 0.002 0.003

Boron <0.001 <0.001 0.0001Aluminum 1.81 0.049 0.052Nitrogen 0.0060 0.0096 —Oxygen 0.0033 0.0078 —

Lead <0.001 <0.001 —Tungsten 0.002 <0.001 —

Zirconium 0.003 <0.001 —

Table 1 — Base Materials Investigated

Grade Coating Thickness(mm)

DP 780 HDGA 1.54DP 980 HDGA 1.60TRIP 780 HDGA 1.35

Note: HDGA = Hot-dipped galvanized and annealed.

Table 3 — Typical Welding and Heat Sink Conditions for Welds with Nominally High and LowCooling Rates

Nominal Average Average Travel HeatCooling Transfer Current Voltage Speed Heat Input

Grade Rate Mode (A) (V) (in./min) Sink (kJ/mm)

DP 780 High SC 48.7 17.2 11.9 Copper 0.18TRIP 780 High SC 45 19.8 13.2 Copper 0.16

DP 780 Low Spray 230 24.4 32 Air 0.40TRIP 780 Low Spray 237 24.8 45.1 Air 0.29

Table 4 — Test Matrix of Lap Welds for HAZ Characterization

Nominal Transfer Heat Heat InputGrade Cooling Rate Mode Sink (kJ/mm)

DP 780 High SC Copper 0.17DP 780 Low Spray Air 0.41

TRIP 780 High SC Copper 0.16TRIP 780 Low Spray Air 0.29

Note: Test matrix repeated for prestrained coupons.

Table 5 — Gleeble Test Conditions for HAZ Characterization

Peak Temperature Heating RateSample (°C) (°C/s)

1 500 252 1000 253 1000 0.16


WELDING RESEARCH


both with and without postbaking. Post-baking consisted of placing welded speci-mens in a preheated oven, allowing thespecimens to heat, holding the specimensat 170°C for 25 min., then removing fromthe oven and allowing to air cool.

The HAZ and fusion zone hardnessprofiles and microstructures of the GMAwelds were characterized. Static and dy-

namic tensile properties of DP 780 andTRIP 780 were also evaluated and relatedto the microstructural observations. Theweld evaluation approach is describedbelow in three parts.

HAZ Characterization

The effects of grade, prestrain, cooling

rate condition, and postbake on the HAZmicrostructure and hardness profile wereevaluated. Lap welds for HAZ evaluationwere produced using ER70S-6 wire. Table4 is the test matrix, which was applied forthe lap joints both with and without pre-strain. To ensure that steady-state condi-tions had been reached, metallographiccross sections were taken at least 3 in.from the start of the weld. One cross sec-tion was examined in the as-received con-dition, while another underwent a post-bake treatment prior to metallographicpreparation. Hardness traverses weretaken along the centerline of the top sheet.

Gleeble testing was performed to re-late the phase transformations that occurat different HAZ locations to the temper-ature at which the location was heated.Table 5 is the test matrix used during theGleeble trials. 5- × 75-mm (0.020- × 2.95-in.) samples were heated at the rates listedand then cooled to room temperature atapproximately 40°C/s. Each sample was

Fig. 2 — Fixture used to produce different heat sink conditions.

Fig. 4 — Hardness profiles of DP 780 and TRIP 780 lap welds producedwith the nominally low cooling rate: no prestrain or postbaking.

Fig. 3 — Hardness profiles of DP 780 and TRIP 780 lap welds produced withthe nominally high cooling rate: no prestrain or postbaking.

Fig. 5 — Hardness profiles of DP 780 and TRIP 780 lap welds produced bothwith and without prestrain for the high cooling rate condition.

Table 6 — Test Matrix of Lap Welds for Fusion Zone Characterization

Steel Filler Nominal Transfer Heat Heat InputGrade Metal Cooling Rate Mode Sink (kJ/mm)

DP 780 ER70S-6 High SC Copper 0.15DP 780 ER70S-6 Low Spray Air 0.43DP 780 ER100S-G High SC Copper 0.19DP 780 ER100S-G Low Spray Air 0.41

TRIP 780 ER70S-6 High SC Copper 0.16TRIP 780 ER70S-6 Low Spray Air 0.34TRIP 780 ER100S-G High SC Copper 0.19TRIP 780 ER100S-G Low Spray Air 0.29


WELDING RESEARCH

JUNE 2008, VOL. 87-s138

then cross sectioned along its center (thelocation where the temperature reachedthat listed in Table 5) and prepared formetallographic examination and hardnesstesting.

Fusion Zone Characterization

The effects of filler metal selection,cooling rate condition, and dilution on thefusion zone microstructure and hardnesswere evaluated. The test matrices of thelap and butt joints are shown in Tables 6

and 7, respectively. DP 780 and TRIP 780lap and butt joints were produced usingboth ER70S-6 and ER100S-G wire types.To ensure that steady-state conditions hadbeen reached, metallographic cross sec-tions were taken at least 3 in. from thestart of the weld. For each specimen dilu-tion of the filler metal by the base metalwas approximated from the macrographsusing imaging software. In addition, theweld metal chemistry of the high and lowcooling rate TRIP 780 lap welds producedwith the ER70S-6 wire was analyzed to

verify the dilution estimates.

Mechanical Property Characterization ofDP 780 and TRIP 780

The static and dynamic tensile strengthof lap and butt joints produced usingTRIP 780, and butt joints produced usingDP 780 were assessed. Table 8 is the weld-ing matrix used in producing the mechan-ical test specimens. To achieve completepenetration and avoid melt-through, thecalculated welding heat input of all buttjoint welds was similar. Thus, cooling ratewas affected primarily by the heat sinkcondition for the butt joint welds. For theTRIP 780 lap joints, a broader range ofcalculated welding heat input levels wereused in combination with heat sinking toaffect cooling rate.

For each welding condition listed inTable 8, three specimens underwent statictensile testing and three specimens under-went dynamic tensile testing. Weld metalreinforcement was machined from thesurface of the butt joint welds so that thefusion zone thickness was similar to that ofthe base material. Coolant was used dur-ing machining to minimize metallurgicaleffects. All specimens were laser cut intothe reduced cross-sectional geometryshown in Fig. 1. The butt and lap joint ten-sile specimens were similar; with the pri-mary difference being the location of theweld. For the butt joints, the weld was inthe middle of the specimen, while for thelap joints the overlapped region was in themiddle of the specimen. To avoid slippage,12.5-mm- (½-in.-) diameter holes werefirst placed 25 mm (1 in.) from the end ofeach specimen. These holes were then re-inforced with 12.5-mm-diameter washersthat were spot welded to the specimen. Tominimize bending on the lap joint tensile

Table 7 — Test Matrix of Butt Joint Welds for Fusion Zone Characterization

Steel Filler Nominal Transfer Heat Heat InputGrade Metal Cooling Rate Mode Sink (kJ/mm)

DP 780 ER70S-6 High SC Copper 0.20DP 780 ER70S-6 Low Spray Air 0.17DP 780 ER70S-6 High SC Copper 0.17DP 780 ER70S-6 Low Spray Air 0.18

TRIP 780 ER70S-6 High SC Copper 0.20TRIP 780 ER70S-6 Low Spray Air 0.18TRIP 780 ER100S-G High SC Copper 0.19TRIP 780 ER100S-G Low Spray Air 0.18

Table 8 — Test Matrix of Lap and Butt Joint Welds for Mechanical Property Characterization

Steel Filler Nominal Transfer Heat Joint HIGrade Metal Cooling Rate Mode Sink Geometry (kJ/mm)

DP 780 ER70S-6 High SC Copper Butt 0.20DP 780 ER70S-6 Low Spray Air Butt 0.17DP 780 ER100S-G High SC Copper Butt 0.17DP 780 ER100S-G Low Spray Air Butt 0.18

TRIP 780 ER70S-6 High SC Copper Lap 0.16TRIP 780 ER70S-6 Low Spray Air Lap 0.29TRIP 780 ER100S-G High SC Copper Lap 0.18TRIP 780 ER100S-G Low Spray Air Lap 0.29TRIP 780 ER70S-6 High SC Copper Butt 0.20TRIP 780 ER70S-6 Low Spray Air Butt 0.18TRIP 780 ER100S-G High SC Copper Butt 0.19TRIP 780 ER100S-G Low Spray Air Butt 0.18

Fig. 6 — Hardness profiles of DP 780 lap welds with and without postbak-ing for both prestrained and not prestrained sheet. (Welds were produced withthe nominally high cooling rate condition.)

Fig. 7 — Hardness profiles of TRIP 780 lap welds with and without postbakingfor both prestrained and not prestrained sheet. (Welds were produced with thenominally high cooling rate condition.)

Kapustka 6 08corr:Layout 1 5/6/08 9:28 AM Page 138

WELDING RESEARCH


specimens, 12.5-mm-thick plates wereplaced on the surface opposite the wash-ers during testing.

The static tensile specimens weretested at a rate of 50 mm/min (2.0 in./min).The dynamic tensile specimens weretested using two different weights and aconstant drop height of 10 ft. Initially, a58.5-lb weight was used. When it was real-ized that this weight was inadequate tobreak all specimens, a 106.7-lb weight wasused. Analysis of the peak load vs. strainat peak load, and the peak load vs. elon-gation at peak load curves indicated thatthe different weights did not have a signif-icant effect on dynamic tensile properties.For all tensile tests, peak load, energy atpeak load, and elongation at peak loadwere recorded.

Results

The results of the HAZ characteriza-tion, fusion zone characterization, andmechanical property characterization arepresented separately. The Discussion sec-tion of this paper then relates these results.

HAZ Characterization

Figures 3–6 show hardness profiles forwelds made on DP 780 and TRIP 780using ER70S-6 wire under a range of heatsink and prestrain conditions. The zero onthe X-axis represents the fusion boundarylocation with HAZ to the right and fusionzone to the left. Note that the HAZboundary was defined as the location inwhich the hardness in the HAZ reachesthe average hardness of the base metal.The data presented in these figures aresummarized in Table 9; data for DP 980

lap welds produced with ER70S-6 weldingwire are also listed. Referring to Table 9,ΔHPk and ΔHMin are the degrees of HAZhardening and softening, respectively.The figures show various degrees of HAZhardening and softening depending onmaterial grade and other conditions. Thehighest hardness occurs in the near HAZ(adjacent to the fusion boundary), whilethe softest point is in the far HAZ. Com-

pared to TRIP 780, DP 780 has a higherdegree of hardening and a slightly lowerdegree of softening. The following pro-vides additional description of the resultsin these figures:

• Cooling Rate. Figure 3 shows hard-ness profiles for welds produced with thenominally high cooling rate condition (i.e.,combination of low heat input and heatsink). Figure 4 shows hardness profiles for

Fig. 8 — Comparison of measured temperature cycle from DP 780 and TRIP780 steels during heating and cooling to 1000°C at 25°C/s. [It is important tonote during cooling cycle no forced cooling was used to track the heat of trans-formation from austenite to ferrite. The DP 780 steels showed significantchange in cooling rate as the transformation starts at ~600°C (solid arrow).No such change in cooling rate was observed in TRIP 780 steels.]

Fig. 9 — Measured dilation corresponding to the temperature cycle shown inFig. 8. [DP 780 steel shows classical behavior: on heating austenite formationabove 700°C and on cooling transformation from austenite to ferrite below600°C (marked by solid arrow). It is interesting to note that the TRIP 780 steelsshow a small change in dilation on heating above 500°C and then no classicalaustenite to ferrite transformation.]

Table 9 — HAZ Characterization Data for Lap Welds Produced on DP 780, TRIP 780, and DP980 Using ER70S-6 Welding Wire

Condition HPk ΔHPk HMin ΔHMin Width Grade Cooling Rate Prestrain Baking (Hv) (%) (Hv) (%) (mm)

DP 780 High No No 372 80 197 –5 3.0Low No No 261 26 189 –9 4.0High Yes No 362 52 211 –11 3.2Low Yes No 294 24 207 –13 4.5High No Yes 357 72 181 –13 3.0Low No Yes 280 35 187 –10 4.0High Yes Yes 370 55 218 –8 3.2Low Yes Yes 311 31 210 –12 4.8

TRIP 780 High No No 334 34 218 –8 3.8Low No No 277 11 211 –11 4.2High Yes No 327 5 253 –18 4.2Low Yes No 282 –9 261 –16 5.0High No Yes 388 55 235 –6 3.5Low No Yes 302 21 220 –12 4.2High Yes Yes 308 –1 252 –19 4.2Low Yes Yes 300 –3 259 –16 5.0

DP 980 High No No 392 28 266 –13 3.5Low No No 344 12 250 –18 5.2High Yes No 402 15 274 –22 NALow Yes No 381 9 252 –28 NAHigh No Yes 407 27 270 –16 3.2Low No Yes 362 17 258 –17 5.4High Yes Yes 400 14 291 –17 NALow Yes Yes 383 9 284 –19 NA

HPk = Peak HAZ hardness ΔHPk = [(HPk – Base Metal Hardness) – 1] × 100HMin = Minimum HAZ hardness ΔHMin = [(HMin – Base Metal Hardness) – 1] × 100


WELDING RESEARCH

JUNE 2008, VOL. 87-s140

the DP 780 and TRIP 780 welded with thenominally low cooling rate (i.e., combina-tion of high heat input and no heat sink).The plots show that cooling rate tends tohave the largest effect on the DP steel,with the TRIP steel being somewhat lessaffected. Lower cooling rates tend to pro-duce a wider HAZ (as identified from thehardness profile) with lower peak hard-ness. For example, the DP 780 peak hard-

ness was about 80% higher than the basemetal for the high cooling rate condition,and about 25% higher than the base metalfor the low cooling rate condition. Thelower cooling rate also tends to produceslightly more HAZ softening. The TRIP780 showed about 11% reduction in hard-ness compared to the base metal.

• Material Prestrain. Figure 5 showshardness profiles for DP 780 and TRIP

780 welds made both with and withoutprestrain for the high cooling rate condi-tion. Prestrain has the largest effect on theTRIP base material, increasing the basemetal hardness by about 25%. The hard-ness of the softest location of the TRIP780 HAZ is also increased by prestrain;with the degree of softening being slightlyincreased (about 3–10%). The softestpoint in the TRIP 780 HAZ remainsgreater than the hardness of the as-received base metal. Prestraining in-creased the DP 780 base metal hardnessby only about 10%. Prestraining did notaffect the peak HAZ hardness for eithermaterial.

• Postbake. Figure 6 shows hardnessprofiles of DP 780 welds with and withoutpostbaking for both prestrained and notprestrained sheet. Based on these data,postbaking did not appear to have a sig-nificant influence on the HAZ hardnessprofiles of the DP 780 material, regardlessof prestrain condition. Figure 7 shows a

Fig. 10 — Micrographs of the DP 780 HAZ and base metal for a lap weld produced with the nominally high cooling rate condition (2% Nital etch). A — 2ndindent in HAZ (370 HV, 1000×); B — 5th indent in HAZ (340 HV, 1000×); C — 10th indent in HAZ (195 HV, 1000×); D — base metal (207 HV, 500×).

Table 10 — Fusion Zone Microstructure, Hardness, and Dilution of DP 780 Welds Produced withER70S-6 Electrode

Heat NominalInput Cooling Heat Dilution Average

(kJ/mm) Rate Sink (%) Hardness Microstructure

Lap Welds

0.15 High Copper 21 269 Predominately AF with a small fraction of B0.43 Low Air 59 251 Predominately AF with a small fraction of B

Butt Joint Welds0.20 High Copper 46 309 Predominately AF with some B and M

(very fine microstructure)0.17 Low Air 52 266 Mixture of B and AF

A B

C D


WELDING RESEARCH


Fig. 11 — Micrographs of the TRIP 780 HAZ and base metal for a lap weld produced with the nominally high cooling rate condition (2% Nital etch). A —2nd indent in HAZ (335 HV, 1000×); B — 5th indent in HAZ (275 HV, 1000×); C — 10th indent in HAZ (224 HV, 1000×); D — base metal (254 HV, 500×).

similar result for the TRIP 780 material.Referring to Table 9, comparison of

these two steels with DP 980 indicates thatalthough DP 980 had the highest peakHAZ hardness, DP 780 had the highestdegree of HAZ hardening. Furthermore,the DP 980 generally had a higher degreeof HAZ softening compared to the DP780 and TRIP 780 materials.

Figures 8 and 9 show Gleeble test re-sults for the DP 780, and TRIP 780 steelheated to 1000°C at 25°C/s. Figure 8 is aplot of measured thermal cycles duringheating and cooling of these steels. It isnoteworthy that during these experi-ments, the heating rates were controlledprecisely and the cooling rate was allowedto be modified by heat of transformation.The plots clearly show an abrupt change inthe cooling curve slope (marked by arrow)for the DP 780 steels. Figure 9 shows thecorresponding measured dilatation for thesame samples. In DP 780, specimen ex-pansion is nearly linear with increased

temperature within the ranges of roomtemperature to 720°C and 860° to 1000°C.Upon heating through the temperaturerange of 720° to 860°C, the specimen widthdecreases and then begins to once againincrease. This, along with the fact that thelinear coefficient of thermal expansion(CTE) is lower in the range of 100° to720°C than from 860° to 1000°C, suggests

that the initial microstructure trans-formed to austenite between 720° and860°C. Similar heating curves were ob-served for the DP 980 specimens that un-derwent similar thermal cycles and will bepresented in future work. The measureddilation curve for TRIP 780 is significantlydifferent than that which was observed forthe DP materials. Upon heating to

Table 11 — Fusion Zone Microstructure, Hardness, and Dilution of DP 780 Welds Produced withER100S-G Electrode



Lap Welds

0.19 High Copper 38 326 Mixture of M and B0.41 Low Air 53 274 Predominately AF with a small fraction of B

Butt Joint Welds0.17 High Copper 46 317 Predominately AF, which is

probably interweaved with M. Some WF0.18 Low Air 54 307 Predominately AF with some areas of B

C D

BA


WELDING RESEARCH

JUNE 2008, VOL. 87-s142

1000°C, no large changes in the CTE wereobserved, which suggests that the TRIP780 does not fully transform to austenite.The difference in transformation strains isattributed to increased ferrite stability inthe TRIP 780. However, more work is nec-essary to understand these complex trans-formations at different heating and cool-ing rates.

To relate the hardness data to mi-crostructures, micrographs were taken atvarious HAZ locations of the DP 780 andTRIP 780 welds. Figure 10 shows regionsof the DP 780 HAZ and base metal of ahigh cooling rate weld. The top-left mi-crograph was taken near the location ofpeak HAZ hardness (370 HV). The bot-tom half of the image (where the hardnesswas measured) has a microstructure that

consists of predominately martensite withsome bainite. Moving farther away fromthe fusion boundary, the top-right micro-graph shows a slightly softer (340 HV) re-gion with a mixture of martensite and bai-nite. The bottom-left micrograph shownin Fig. 10 was taken near the minimumhardness (195 HV) location of the highcooling rate weld HAZ. As shown, the mi-crostructure is indistinguishable from thatof the base metal.

By comparison, the microstructure ofthe low cooling rate DP 780 weld (notshown) in the area of peak hardness (260HV) consists of predominately bainitewith some grain boundary ferrite. Thesedata suggest that the significant increasein peak hardness in the high cooling rateweld near HAZ is due to the higher vol-

ume fraction of martensite.Micrographs taken at various locations

of the TRIP 780 high cooling rate weldHAZ are shown in Fig. 11. The top-left mi-crograph was taken near the point of peakHAZ hardness (335 HV) and has a mi-crostructure that consists of a mixture ofmartensite and large ferrite grains. Thetop-right micrograph was taken fartherfrom the fusion boundary in a softer HAZregion (275 HV) and has a mixture oflarge ferrite grains and degeneratemartensite. Note that the term degeneratemartensite is used to describe regions thatappear to be martensite at optical mi-croscopy levels, but may contain temperedmartensite and/or bainite constituents(Ref. 2). Comparison of the top-left andtop-right micrographs shown in Fig. 11 in-

Fig. 12 — Equilibrium phase diagram of TRIP 780 calculated withThermo-Calc™ software, chemical composition determined by ICPanalysis.

Fig. 14 — Weld metal microstructure of the TRIP 780/ER70S-6 lap weldproduced at high cooling rate.

Fig. 15 — Static tensile test results of DP 780 butt joints.

Fig. 13 — Weld metal microstructure of the DP 780/ER70S-6 lap weld producedat high cooling rate.


WELDING RESEARCH


Fig. 16 — Dynamic tensile test results of DP 780 butt joints. Fig. 17 — Static tensile test results of TRIP 780 lap joints.

dicates that the fraction of ferrite in-creases with distance from the point ofpeak HAZ hardness. The bottom-left mi-crograph is in the softened region and ap-pears similar to that of the base metal,consisting of predominately ferrite with is-lands of degenerate martensite. Thedarker contrast of many of the degeneratemartensite islands suggests tempering hasoccurred during the weld thermal cycle.

By comparison, the microstructure ofthe low cooling rate TRIP 780 weld HAZ(not shown) at the location of peak hard-ness (275 HV) consists predominately offerrite grains separated by regions of a finedispersion of degenerate martensite.Again, this suggests that the lower marten-site content is responsible for the reducedpeak hardness.

It should be noted that the weld inter-face of the TRIP 780 welds consists of acontinuous band of ferrite grains. The fer-rite grains are larger along the weld inter-face of the high cooling rate weld thanalong the weld interface of the low coolingrate weld. It was observed that for bothcooling rate conditions the fraction of fer-rite decreased from the weld interface tothe location of peak HAZ hardness, andthen increased as the point of minimumHAZ hardness was approached. Betweenthe point of minimum HAZ hardness andthe HAZ boundary (unaffected basemetal), the area fraction of ferrite in themicrostructure appears constant.

Additional analysis was done to betterunderstand the TRIP 780 HAZ phasetransformations. The composition of theTRIP 780 was determined by ICP analysis.The aluminum content was found to be1.8%, which is about 1% higher than thenominal composition indicated by themanufacturer. Using this composition, theequilibrium phase diagram shown in Fig.12 was determined using Thermo-Calc™software (Ref. 3). As shown in Fig. 12, theTRIP 780 steel isn’t expected to transform

to 100% austenite on heating to highertemperature. In other words, ferrite isthermodynamically stable at all tempera-tures below the melting point with 1.8%aluminum. The phase diagram also showsthat we would expect 100% austenite for-mation if the nominal composition of0.8% aluminum was used.

Fusion Zone Characterization

Figures 3 and 4 show hardness profilesof DP 780 and TRIP 780 welded withER70S-6 for the high and low cooling rateconditions, respectively. Tables 10–13 listthe nominal cooling rate conditions, dilu-tion, average hardness, and microstruc-ture of DP 780 and TRIP 780 lap and buttjoint welds produced with both the

ER70S-6 and ER100S-G wires. For com-parison purposes, similar data for DP 980lap welds produced with ER70S-6 wire arelisted in Table 14.

Figure 13 shows the weld metal mi-crostructure of the DP 780/ER70S-6 lapweld produced with a high cooling rate,consisting predominately of acicular fer-rite with a small fraction of bainite. AsTable 10 indicates, lap joints produced atboth high and low cooling rates had a sim-ilar microstructure, although the low cool-ing rate was found to have a somewhatsofter and coarser microstructure. Thebutt joint welds produced with high cool-ing rate consisted of predominately acicu-lar ferrite with some bainite and marten-site. At low cooling rates the butt jointweld consisted of a mixture of acicular fer-

Table 12 — Fusion Zone Microstructure, Hardness, and Dilution of TRIP 780 Welds Producedwith ER70S-6 Electrode



Lap Welds0.16 High Copper 37 271 Predominately B with a small fraction of AF0.34 Low Air 59 256 Predominately B with a small fraction of AF,

(B has higher aspect ratio)Butt Joint Welds

0.20 High Copper 44 283 Predominately B, no AF0.18 Low Air 57 278 Predominately B with a small fraction of AF

Table 13 — Fusion Zone Microstructure, Hardness, and Dilution of TRIP 780 Welds Producedwith ER100S-G Electrode


(kJ/mm) Rate Sink (%) Hardness MicrostructureLap Welds

0.19 High Copper 31 369 Mixture of M and B0.30 Low Air 54 317 Predominately B with some AF

Butt Joint Welds0.19 High Copper 44 389 Most likely M0.18 Low Air 57 333 Predominately B with some M


WELDING RESEARCH

JUNE 2008, VOL. 87-s144

rite and bainite. Table 11 indicates thecooling rate conditions and relates themto the average hardness, dilution, and mi-crostructure of the DP 780 welds pro-duced with the ER100S-G wire. The mi-crostructure of the high cooling rate lapweld consisted of a mixture of martensiteand bainite, while the low cooling rate lapweld consisted of predominately acicularferrite with a small fraction of bainite. Buttjoint welds were predominantly acicular

ferrite with either martensite (for the highcooling rate welds) or bainite (for the lowcooling rate welds). Thus, the fusion zonemicrostructure of the DP 780 was affectedby filler metal type, joint design, dilution,and cooling rate.

Figure 14 shows the weld metal mi-crostructure of the TRIP 780/ER70S-6 lapweld produced at a high cooling rate, con-sisting predominately of bainite with asmall fraction of acicular ferrite. As Table

12 indicates, the fusion zone microstruc-ture and hardness obtained for the TRIP780 lap welds consists predominately ofbainite with a small fraction of acicularferrite. The microstructure of the buttjoint welds consisted primarily of bainite,with the low cooling rate weld having a

Fig. 18 — Fracture appearance of TRIP 780 low cooling rate lap joint produced with the ER100S-G wire. (Failure is along the weld interface of the top sheet.)

Fig. 19 — Fracture appearance of TRIP 780 high cooling rate lap joint produced with the ER100S-G wire. (Failure is along the weld interface of the top sheet,possibly initiating at porosity present at the root.)

A B

A B

Table 14 — Fusion Zone Microstructure, Hardness, and Dilution of DP 980 Lap Welds Producedwith ER70S-6 Electrode



0.15 High Copper 20 263 Predominately AF with a small fraction of B0.44 Low Air 59 253 Predominately AF with a small fraction of B

Table 15 — Carbon Equivalents, Aluminum,and Silicon Contents of Each Base Material

Aluminum SiliconMaterial CE (wt-%) (wt%)

TRIP 780 0.53 1.81 0.023DP 780 0.47 0.049 0.005DP 980 0.49 0.052 0.028

Note: CE = C + A(C) * [Si/24 + Mn/6 + Cu/15 + Ni/20 + (Cr + Mo + Nb + V)/5 + 5*B]A(C) = 0.75 + 0.25*tanh[20*(C – 0.12)]


WELDING RESEARCH


small fraction of acicular ferrite. Table 13lists the microstructure and hardness ofthe TRIP 780 welds produced with theER100S-G wire. The fusion zone mi-crostructure of the welds at high coolingrate consisted predominantly of marten-site. These welds had significantly higherhardness than those produced at low cool-ing rate, which were predominantly bai-nite with some martensite or acicular ferrite.

Table 14 lists the microstructure, aver-age hardness, and dilution of the DP 980lap welds produced with a high and lowcooling rate using ER70S-6 wire. The mi-crostructure of both welds is primarily aci-cular ferrite with a small fraction of bai-nite. In this case, cooling rate did not havea large influence on microstructure orhardness. Referring to the HAZ charac-terization data listed in Table 9, the DP980 weld fusion zone is softer than anypoint in the weld HAZ. This suggests thatan undermatched strength condition mayexist when DP 980 is welded with ER70S-6 wire. Depending on the loading condi-tion, failure may be expected to occurthrough the softer weld metal. Comparingthe fusion zone characterization data ofthe DP 780 and DP 980 lap joints (Table10 and Table 14, respectively) indicatesthat the microstructures and hardness val-ues obtained were very similar. Therefore,the base metal composition did not signif-icantly affect the fusion zone microstruc-ture for the DP steels.

Mechanical Property Characterization

The static tensile test results of the DP780 butt joints are shown in Fig. 15. Theresults are expressed in terms of joint effi-ciency (i.e., weld ultimate tensilestrength/measured base metal ultimatetensile strength) and the strain at peak

load. All of thesewelds failed in thesoftened region of thefar HAZ. As shown,the high cooling ratewelds had joint effi-ciencies in excess of90%; whereas the lowcooling rate weldshad joint efficienciesin the range of 85 to90%. The high cool-ing rate welds also ap-pear to have slightlygreater strains atpeak load. Fillermetal strength did nothave a distinguish-able effect on the sta-tic tensile properties.

The dynamic ten-sile test results of theDP 780 butt joints areshown in Fig. 16. The results are presentedin terms of ultimate tensile strength andthe strain at peak load. All of these speci-mens failed in the softened region of thefar HAZ. Ultimate tensile strengthsranged from 122 to 132 ksi (841 to 910MPa), and strain at peak load ranged from2.25% to less than 4.0%. Neither fillermetal strength nor cooling rate conditionhad a distinguishable effect on the dy-namic tensile properties of the DP 780butt joints.

Figure 17 shows TRIP 780 lap joint sta-tic tensile results for different filler metaland cooling rate conditions. The data fromFig. 17 indicate the following:

• Joint efficiencies ranged from about50% to about 98%. Strains at peak loadranged from less than 3% to nearly 8%.

• Specimen failure occurred either inthe far HAZ (i.e., near the point of greatestsoftening) or at the weld fusion boundary.

• Filler metal strength had no discern-able effect on the weld tensile properties.

• Welds produced under high coolingrate conditions had a significantly lowerjoint efficiency and strain than the weldsproduced under low cooling rate condi-tions. This difference was primarily attrib-uted to porosity in the welds producedwith the high cooling rate conditions(using short-circuit transfer).

The low cooling rate lap welds pro-duced with the ER70S-6 welding wirefailed in the softened region of the HAZ.Figure 18 shows micrographs of a lowcooling rate weld produced with theER100S-G wire, which failed along theweld interface of the top sheet. The rightside of the right micrograph reveals a verysmall amount of porosity to be present.However, two of the low cooling ratewelds produced with the ER100S-G wirefailed along the weld interface of the topsheet without observable porosity.

Fig. 20 — Dynamic tensile test results of TRIP 780 lap joints. Fig. 21 — Static tensile test results of TRIP 780 butt joints.

Fig. 22 — Dynamic tensile test results of TRIP 780 butt joints.


WELDING RESEARCH

JUNE 2008, VOL. 87-s146

Five of the six high cooling rate weldshave joint efficiencies in the range of 50 to78%, and strains at peak load between2.75 and 4%. Failure in these welds mayhave initiated at porosity present at theweld root, with failure along the fusionboundary of the top sheet. Figure 19shows micrographs of one of these welds.The outlying high cooling rate weld (withabout 85% joint efficiency and 5.75%strain at peak load) failed along the weldinterface of the bottom sheet.

Figure 20 shows the TRIP 780 lap jointdynamic tensile results for different fillermetal and cooling rate conditions. Asshown, ultimate tensile strengths rangedfrom 54 to 126 ksi (372 to 867 MPa) andstrain at peak load ranged from less than1% to over 5%. Referring to Fig. 20, thehigh cooling rate TRIP 780 lap joints hadlower strengths and strains at peak load.These welds failed along the weld interfacepresumably due to porosity present at theroot. All of the low cooling rate TRIP 780lap welds produced with the ER70S-6 wirefailed in far HAZ of the bottom sheet. Ofthe low cooling rate TRIP 780 lap jointsproduced with the ER100S-G wire, two dy-namic tensile specimens failed in the soft-ened region of the far HAZ, and one failedalong the weld interface of the top sheetwithout the presence of porosity at the weldroot. Analysis of Fig. 20 indicates that fillermetal strength did not have a distinguish-able effect on the dynamic tensile test re-sults of the TRIP 780 lap joints.

The static tensile test results of theTRIP 780 butt joints are shown in Fig. 21.All of these welds failed in the softened re-gion of the far HAZ with joint efficienciesin excess of 89%. On average, high coolingrate welds had higher strains at peak loadthan the low cooling rate welds. As was thecase with the DP 780 butt joint welds andthe TRIP 780 lap welds, filler metalstrength did not appear to influence thestatic tensile properties.

Figure 22 shows the dynamic tensiletest results of the TRIP 780 butt joints. Allof these specimens failed in the softenedregion of the far HAZ. The ultimate ten-sile strengths of the TRIP 780 butt jointsranged from 122 to 130 ksi (841 to 896MPa), and strain at peak load was gener-ally between 3 and 4%. Analysis of Fig. 21indicates that neither filler metal strengthnor cooling rate condition had a distin-guishable effect on the dynamic tensiletest results of the TRIP 780 butt joints. Itshould be noted that the low cooling rateTRIP 780 butt joints had dynamic tensileproperties comparable to the TRIP 780lap joints and the DP 780 butt joints. How-ever, on average, the TRIP 780 butt jointshad slightly higher strain at peak loadcompared to the DP 780 butt joints.

Discussion

The chemical compositions of DP 780and DP 980 are similar, with the latter hav-ing a higher percentage of carbon andother hardenability alloying additions.The base metal microstructure of the DPsteels consists of a ferrite matrix with de-generate martensite islands, with the DP980 having a higher proportion of marten-site. The results of this project indicatethat the process variables investigatedhave similar effects on the DP 780 and DP980 HAZ and weld metal microstructure.This is expected when one considers thatthe base metal microstructure and chem-istry of the two steels are similar.

The base metal microstructure of TRIPsteels consists of a ferrite matrix with islandsof martensite, bainite, and retained austen-ite. The TRIP 780 investigated in this studycontains a relatively high amount of alu-minum (1.8 wt-%). Aluminum is a ferritestabilizer and (when above approximately0.8 wt-%) can allow ferrite to remain stableat temperatures approaching the meltingpoint of the material (Refs. 1, 3). The resultsof this investigation verified that retainedferrite is present in all regions of the TRIP780 HAZ.

Because the microstructures of the DPand TRIP steels are substantially differ-ent, the results for these steels are dis-cussed separately.

DP 780 and DP 980 Steels

Hardness traverses indicate that theweld has regions of significant hardeningand softening depending on the basemetal grade, filler metal type, and coolingrate conditions (as determined by weldingheat input and heat sinking). The locationof greatest hardening is in the near HAZ(adjacent to the fusion boundary), wherethe far HAZ experienced softening.

For DP 780 welded with ER70S-6 wire,the softest location in the weld is the farHAZ. Gleeble test results indicate that theDP steels both fully transform to austeniteupon heating to 1000°C. The minimumhardness location corresponds to thepoint in the HAZ where the temperatureis near the minimum required to beginforming austenite (AC1 boundary). Thissoftening is probably attributable to tem-pering of the degenerate martensite. Me-chanical testing showed that this softenedregion limits the DP 780 weld strength,and that a higher-strength consumable didnot provide static or dynamic tensilestrength improvements. During static ten-sile testing, cooling rate had only a slighteffect on the DP 780 butt joint efficiency;for the high cooling rate condition joint ef-ficiencies exceeded 90%, while for the lowcooling rate condition joint efficiencies

ranged from 85 to 90%. During dynamictensile testing, cooling rate did not have adistinguishable effect on the strength ofthe DP 780 butt joint welds.

The microstructure and peak hardnessof the near HAZ were strongly influencedby both cooling rate and base metal com-position. A predominately martensitic mi-crostructure is present in the region ofpeak hardness (395 HV) in the DP 980high cooling rate weld, while the DP 780had a somewhat softer (375 HV) mixtureof martensite and bainite. The peak hard-ness of the low cooling rate DP 780 weld(260 HV) is significantly less than that ofthe high cooling rate DP 780 weld (375HV). This difference in hardness is prob-ably a result of autotempering of themartensite during the slower coolingcycle. In other words, the slower coolingrates allow martensite formed in the nearHAZ sufficient time at temperatures inwhich tempering can occur. The potentialimplications of the hardness increase inthe near HAZ region are not wellunderstood.

As previously mentioned, the mi-crostructure of both DP 780 and DP 980consists of a ferrite matrix with degeneratemartensite islands. During cold rolling ofthese materials, the stress and strain dis-tribution in the two structures is different(Refs. 4, 5). The strain tends to concen-trate in the soft ferrite (Ref. 4). BecauseDP 980 has a lower fraction of ferrite thanDP 780, it may have a higher dislocationdensity in the ferrite phase of its mi-crostructure. Therefore, prestrain pro-duces a larger increase in base metal hard-ness for the DP 980 steel than for the DP780 steel.

Welding on prestrained DP 780 andDP 980 resulted in a wider HAZ and theminimum HAZ hardness being slightlygreater than that of the welds without pre-strain. In addition, it was observed that foreach respective material the high coolingrate weldment had a slightly greater mini-mum HAZ hardness than the low coolingrate welds with nominally the sameamount of prestrain. The differences inthe HAZ of the welds produced with pre-strain to those produced on the as-received base metal are attributable to theprocess of recovery. For both DP grades,the minimum HAZ hardness of the pre-strained weld remains greater than that ofthe weld without prestrain, which is likelydue to the weld thermal cycle in the farHAZ being insufficient for completion ofthe recovery process. Compared to DP780, a greater reduction in base metalhardness occurs in the HAZ of prestainedDP 980. This may be due to the higher dis-location density present in the ferritephase of this material producing a largerdriving force for recovery.


WELDING RESEARCH


Based on the fusion zone data pre-sented in the tables, weld metal mi-crostructure and hardness appear to bepredominately influenced by the coolingrate, rather than the DP grade. Welds onDP 780 produced using both wires tendedto consist of predominately acicular fer-rite with varying degrees of bainite andmartensite. The welds on the DP 780 andDP 980 produced using the ER70S-6 wirehad similar microstructures and hardness,and no correlation could be made betweenbase metal dilution and microstructure.Analysis of the microstructure and the re-sulting fusion zone hardness indicates thatdilution of the filler metal by the basemetal does play a role in weld metal mi-crostructure evolution, although the effectof dilution was more pronounced for theTRIP 780, which is discussed later.

When the DP 980 was welded with theER70S-6 consumable, the lowest hardnessoccurred in the fusion zone. This suggeststhat a higher-strength electrode would berequired to avoid undermatched weldmetal strength and consequent failurethrough the fusion zone. Application of ahigher-strength electrode may be requiredto shift the failure location to the softenedregion of the HAZ.

TRIP 780 Steel

The TRIP 780 welds had a continuousstructure of ferrite grains along the fusionboundary. The presence of ferrite alongthe fusion boundary can be explained withthe equilibrium phase diagram shown inFig. 12. For TRIP 780, at temperaturesnear the liquidus, ferrite is the only ther-modynamically stable phase.

For all locations within the TRIP 780HAZ, the fraction of ferrite in the mi-crostructure is greater for the welds pro-duced with low cooling rate compared tohigh cooling rate. The microstructure ofthe low cooling rate weld at the point ofpeak HAZ hardness consists of predomi-nately ferrite grains separated by regionsof degenerate martensite. As previouslynoted, the term degenerate martensite isused to describe regions that appear to bemartensite at optical microscopy levels,but may contain tempered martensiteand/or bainite constituents. The mi-crostructure in the high cooling rate weldat the location of peak HAZ hardness con-sists of a mixture of martensite and largeferrite grains. In both cases, the fraction offerrite in the HAZ microstructure in-creases from the point of peak HAZ hard-ness to the location of minimum HAZhardness. The microstructure of the highcooling rate weld has greater hardness ateach HAZ location between the points ofminimum and peak hardness.

The microstructure evolution in the

HAZ of TRIP steels can be separated intotwo different regions. The regions thatheated above the AC1 temperature(700°C) and the regions heated below theAC1 temperature.

Referring to Fig. 12, the regions heatedabove the AC1 do not entirely transform toaustenite due to increased stability of fer-rite. As a result, the regions heated abovethe AC1 will remain in a two-phase (austen-ite + ferrite) region throughout the heatingcycle. Depending upon the peak tempera-ture, the ferrite fraction may increase fromthe original level. This often resulted in acontinuous necklace of ferrite along theweld interface of the welds produced on theTRIP 780 steel. Similar microstructural ob-servations have been noted for multipassself-shielded gas metal arc welds (Refs. 6,7). As the HAZ starts cooling, a small frac-tion of the austenite may be retained, butthe larger fraction decomposes into eitherbainite or martensite. The rate of coolingwill determine the nature of this mi-crostructure mixture. This is supported bythe HAZ microstructure of welds madewith both the nominally high and low cool-ing rate conditions.

The regions heated below the AC1 (farHAZ) and cooled are expected to un-dergo subtle microstructural changes. Aspreviously noted the base metal mi-crostructure of TRIP steel consists of aferrite matrix with islands of martensite,bainite, and retained austenite. Depend-ing upon the peak temperature, the re-tained austenite may transform to marten-site upon cooling; this can only bedetermined using transmission electronmicroscopy. At a given location of the farHAZ, the martensite formed from the re-tained austenite, as well as that present inthe as-received base material, may un-dergo tempering. The degree of marten-site tempering is dependent on the weldthermal cycle at the given location. Basedon the measured hardness of the weldsproduced with both cooling rate condi-tions, lesser degrees of martensite tem-pering is expected in the welds made withthe high cooling rate condition. The extentof softening appears to decrease with anincrease in distance from the AC1boundary.

Prestrain had a more pronounced ef-fect on the TRIP 780 than was observedwith the DP steels. During deformation ofTRIP steels at ambient temperatures theretained (quasi-stable) austenite progres-sively transforms into martensite as thematerial is strained (Refs. 4, 8). Therefore,the increased base metal hardness is dueto both the new martensite formed, and anincrease in the dislocation density of theferrite grains. For both cooling rate condi-tions, prestrain did not increase the peakHAZ hardness, but did increase the hard-

ness in the region between the AC1 bound-ary and the peak HAZ hardness location.In addition, the minimum HAZ hardnessof the welds made on the prestrained basemetal remained greater than the hardnessof the as-received base metal.

Welds produced on TRIP 780 with theER70S-6 wire consisted primarily of bai-nite with a small fraction of acicular fer-rite. Compared to the DP 780 welds, foreach condition of joint geometry and cool-ing rate these welds tended to have slightlyhigher hardness. When the ER100S-Gwire was used, the microstructure of theTRIP 780 welds tended to consist either ofmartensite and/or bainite. For this elec-trode/base material combination, acicularferrite was only observed in the mi-crostructure of the low cooling rate lapjoint weld. Compared to the DP 780 weldsproduced with the ER100S-G wire, foreach condition of joint geometry and cool-ing rate the TRIP 780 welds had signifi-cantly higher hardness.

The fusion zone hardness data suggestthat base metal dilution has a greater ef-fect on weld metal hardenability for TRIP780 than for DP 780 or DP 980. Of thethree materials, TRIP 780 appears to bethe most heavily alloyed (Table 2). Table15 lists the carbon equivalents of thesematerials, which were calculated using theformula proposed by Yurioka et al. (Ref.9). As listed, TRIP 780 has the highest car-bon equivalent. However, it is importantto note that the carbon equivalency for-mula does not consider aluminum. In thisregard, the major effect of dilution is at-tributed to higher carbon content in TRIPsteels. As Kou reported, increasing the al-loying content of weld metal increases itshardenability by pushing the nose of thecontinuous cooling transformation curvesto longer times (Ref. 10). Thus, harderconstituents such as bainite and marten-site are more likely to form in the weldsproduced on TRIP 780.

A wide range of joint efficiencies (50 to90%) and ultimate tensile strengths (54 to128 ksi) were obtained for the TRIP 780lap welds during static and dynamic test-ing, respectively. Failure was either alongthe weld interface or more commonly inthe softened region of the HAZ. Five ofthe six TRIP 780 lap welds produced withhigh cooling rate conditions failed alongthe weld interface during static testing(presumably due to porosity at the weldroot) at significantly lower strength andductility than the welds produced with lowcooling rate conditions. Notably, some ofthe welds failed along the weld interfacewithout any indication that porosity waspresent at the weld root. As mentionedpreviously, the weld interface of the TRIP780 welds contains a continuous region oflarge ferrite grains. This region most likely


WELDING RESEARCH

JUNE 2008, VOL. 87-s148

has lower strength than the surroundingweld metal and HAZ microstructure. Theweld root and any porosity present nearthe root serve as stress concentrations.Fracture can initiate at these stress con-centrations and propagate through thelower-strength ferrite grains along theweld interface. Butt joint welds that lackedsuch stress concentrations all failed in thesoftened region of the HAZ at joint effi-ciencies of over 90% (static testing) andultimate tensile strengths between 122and 130 ksi (dynamic testing). For the buttjoints that underwent static tensile testing,higher cooling rate conditions tended toincrease the joint efficiency slightly. How-ever, cooling rate conditions did not havea distinguishable effect on joint strengthduring dynamic tensile testing.

For each condition of cooling rate andjoint geometry, filler metal strength didnot have a distinguishable effect on eitherjoint strength or ductility. Filler metalstrength may play a larger role in staticstrength for prestrained TRIP 780. Pre-straining increased the hardness of thesoftest point in the HAZ, which may shiftthe failure to the fusion zone. It is possi-ble, however, that the continuous ferriteband along the fusion boundary may bethe point of failure on prestrained mater-ial, regardless of the filler metal strength.Additional work is needed to assess the ef-fect of prestrain and the continuous ferriteband on weld tensile properties.

Summary and Conclusions

This work primarily investigated the ef-fects of cooling rate (welding heat inputand fixture heat-sinking), prestrain, fillermetal selection, dilution, and postbakingon the microstructure and mechanicalproperties of GMA welds on coated DP780 and TRIP 780 sheet metal lap jointsand butt joints. A limited amount of test-ing was also conducted for DP 980 lapjoints. The most significant conclusionsare listed below.

1) The DP steels showed varying de-grees of HAZ hardening and softening de-pending on the material grade, prestrain,and cooling rate condition. The relativelyhigh aluminum content of the TRIP 780steel allowed retained ferrite to be presentin all regions of the HAZ, along with acontinuous region of coarse ferrite alongthe weld interface. This resulted in theTRIP 780 having lower peak HAZ hard-ness than the DP 780.

2) Fusion zone microstructure andhardness were found to be affected by thebase metal chemistry, the cooling ratecondition, and the filler metal composi-tion.

3) Filler metal strength did not affectthe static or dynamic tensile properties of

either the TRIP 780 lap or butt joint welds,or the DP 780 butt joint welds. All of theTRIP 780 and DP 780 butt joints failed inthe soft HAZ. The results of the lap jointtests showed a greater variation instrength that is primarily attributed toporosity at the root of the weld.

4) Additional work is needed to relateHAZ and fusion zone microstructures,hardness profiles, and geometric disconti-nuities to fatigue, bend, impact, and crushperformance. In particular, work isneeded to evaluate the effect that the fer-rite region along the fusion boundary hason the mechanical properties of TRIP 780lap joints.

References

1. Biro, E., and Lee, A. 2004. Welded prop-erties of various DP600 chemistries. Proc. SheetMetal Welding Conference XI, Sterling Heights,Mich.

2. Gould, J. 2007. Interview. Edison Weld-ing Institute. Columbus, Ohio.

3. Sundman, B., Jansson, B., and Anderson,J. O. 1985. The thermo-calc databank system.Calphad 9 (2): 1–153.

4. Hulka, K. 2005. Dual phase and TRIP steels.www.us.cbmm.com.br/english/sources/techlib/info/dualph/dualphas.htm.

5. Hsu, C., Soltis, P., Barton, D., and Oc-chialini, C. 2004. Weldability of dual-phasesteel with arc welding processes. Proc. SheetMetal Welding Conference XI, Sterling Heights,Mich.

6. Babu, S. S., David, S., and Quintana, M.2001. Modeling microstructure development inself-shielded flux cored arc welds. Welding Jour-nal 80 (4): 91-s to 97-s.

7. Quintana, M., Mclane, J., Babu, S. S., andDavid, S. 2001. Inclusion formation in self-shielded flux cored arc welds. Welding Journal80(4): 98-s to 105-s.

8. Senuma, T. 2001. Physical metallurgy ofmodern high strength steel sheets. ISIJ Interna-tional 41(6): 520–532.

9. Yurioka, N., Suzuki, H., Ohshita, S., andSaito, S. 1983. Determination of necessary pre-heating temperature in steel welding. WeldingJournal 62(6): 147–153.

10. Kou, S. 2003. Welding Metallurgy, SecondEdition. Hoboken, N.J.: John Wiley & Sons,Inc.

11. Farrar, R. A., and Harrison, P. L. 1987.Acicular ferrite in carbon-manganese weldmetals: An overview. Journal of Material Science22: 3812–3820.

12. Abson, D. J., and Pargeter, R. J. 1986.Factors influencing as-deposited strength, mi-crostructure, and toughness of manual metalarc welds suitable for C-Mn steel fabrications.International Metals Review 31(4): 141–194.

13. Babu, S. S., and Bhadeshia, H. K. D. H.1991. Mechanism of the transition from bainiteto acicular ferrite. Materials Transactions, JIM,32(8): 679–688.

14. Bode, R., Meurer, M., Schaumann, T.W., and Warnecke, W. 2004. Selection and useof coated advanced high-strength steels for au-tomotive applications. Proc. Galvatech ’04 Con-ference. Association for Iron & Steel Technol-ogy, Warrendale, Pa.

15. Shaw, J. R., and Zuidema, B. K. 2001.New high strength steels help automakers reachfuture goals for safety, affordability, fuel effi-ciency and environmental tesponsibility. Jour-nal of Materials and Manufacturing 5(110):976–983.

16. Lippold, J. 2004. Welding MetallurgyPrinciples. National Excellence in MaterialsJoining Education and Training. Columbus,Ohio.

17. Wagner, R. Ferrous Metallurgy. The OhioState University Materials Science & Engi-neering. Columbus, Ohio.

REPRINTS REPRINTS

To order custom reprints of 100 or more of articles in

the Welding Journal,call FosteReprints at

(219) 879-8366 or(800) 382-0808 or

Request for quotes can befaxed to (219) 874-2849.

You can e-mailFosteReprints at

[email protected].

An ImportantEvent

on Its Way?Send information on upcomingevents to the Welding JournalDept., 550 NW LeJeune Rd.,Miami, FL 33126. Items canalso be sent via FAX to (305)443-7404 or by e-mail to [email protected].


Documents

Effect of GMAW Process and Material Conditions on DP 780