Effects of Microstructure on Inverse Fracture Occurringduring Drop-Weight Tear Testing of High-Toughness X70Pipeline Steels

BYOUNGCHUL HWANG, YANG GON KIM, SUNGHAK LEE, NACK J. KIM,and JANG YONG YOO

The effects of microstructure on inverse fracture occurring in the hammer-impacted region were ana-lyzed after conducting a drop-weight tear test (DWTT) on high-toughness pipeline steels. Three kindsof steels were fabricated by varying the alloying elements, and their microstructures were varied bythe rolling conditions. The pressed-notch (PN) or chevron-notch (CN) DWTT and Charpy V-notch(CVN) impact tests were conducted on the rolled steel specimens, and the results were discussed incomparison with the data obtained from CVN tests of prestrained specimens. In the hammer-impactedregion of the DWTT specimens, abnormal inverse fracture having a cleavage fracture mode appeared,and the inverse fracture area correlated well with the upper-shelf energy (USE) obtained from theCVN test and with the grain size. The steel specimens having a higher USE or having coarse poly-gonal ferrite tended to have a larger inverse fracture area than those having a lower USE or havingfine acicular ferrite. This was because steels having a higher impact absorption energy required higherenergy for fracture initiation and propagation during the DWTT. These results were confirmed bythe CVN data of prestrained steel specimens.

I. INTRODUCTION

RECENTLY, rapid technological advancements of thethermomechanical control process (TMCP), composed ofcontrolled low-temperature hot rolling and accelerated cool-ing, have considerably improved the toughness of pipelinesteels.[1–5] In these high-toughness pipeline steels, however,an abnormal fracture appearance frequently occurring dur-ing the drop-weight tear test (DWTT) has caused manydifficulties in evaluating DWTT properties.[6–18] The abnor-mal fracture appearance can roughly be divided into twotypes. First, the test is invalid, in accordance with the APIRP 5L3 specification,[15] as shear fracture initiates at thenotch of the DWTT specimen. In the second case, the so-called “inverse fracture” of a cleavage-fracture mode occursin the hammer-impacted region of the DWTT specimen,after cleavage fracture initiated at the notch propagates in ashear-fracture mode. Thanks to intensive studies, the formerproblem was partially solved by inserting a static precrack,a fatigue precrack, an electron-beam weld, a chevron notch,etc., into the DWTT specimen, instead of the conventionalstandard pressed notch.[7,8] However, the latter problem stillwaits for clear solutions or alternatives to prevent or reduceits occurrence. Empirically speaking, the size of this inversefracture tends to become more serious, as the DWTT speci-

men thickness increases and the test temperature decreases.[6–9]

It is also known that the inverse fracture is associated withthe high fracture-initiation energy of the recently developedhigh-toughness pipeline steels.[11,12] Nevertheless, in the caseof the chevron-notch (CN) DWTT specimen, in which thefracture-initiation energy is reduced by inserting the sharpnotch, the inverse fracture also occurs in the hammer-impacted region, leaving the problem unsolved.[12,13] Fewworks have been done on how inverse fracture occurs in therecently developed high-toughness pipeline steels and onhow microstructures and other properties, together withDWTT type and test conditions, affect the inverse fracture.

Iwasaki et al.[6] investigated the plastic deformation andfracture behavior of DWTT specimens and reported that ahigh compressive prestrain was introduced inside the DWTTspecimen as both the three-point bending and the hammerimpact were working in the hammer-impacted region.Recently, Hwang et al.[12] concluded that the work harden-ing due to such a compressive prestrain was correlated withthe size of the inverse fracture, based on the results of hard-ness and indentation tests of the hammer-impacted region.They pointed out the occurrence of more serious inversefracture in the case of pressed-notch (PN) DWTT specimensbecause of the high fracture-initiation energy at the notch.Mannucci and Harris[11] observed that the thickness of thehammer-impacted region significantly increased during theDWTT and suggested that this thickening further reducedthe toughness by increasing the stress triaxiality in the ham-mer-impacted region.

In the present study, the size, shape, and characteristics ofthe inverse fracture occurring in the hammer-impacted regionof high-toughness X70 pipeline steels were evaluated to inves-tigate the formation mechanisms of the inverse fracture andits effects. In order to suggest methods to prevent the inversefracture formation, the size of the inverse fracture wasmeasured from the fracture surface of the fractured DWTT

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 36A, FEBRUARY 2005—371

BYOUNGCHUL HWANG, Postdoctoral Research Associate, and YANGGON KIM, Research Assistant, are with the Center for Advanced Aero-space Materials, Pohang University of Science and Technology, Pohang790-784, Korea. SUNGHAK LEE and NACK J. KIM, Professors, Centerfor Advanced Aerospace Materials, Pohang University of Science and Tech-nology, are also with the Materials Science and Engineering Department,Pohang University of Science and Technology. Contact e-mail: shlee@postech.ac.kr JANG YONG YOO, Principal Researcher, is with the PlateResearch Group, Technical Research Laboratories, POSCO, Pohang 790-785,Korea.

Manuscript submitted April 1, 2004.

372—VOLUME 36A, FEBRUARY 2005 METALLURGICAL AND MATERIALS TRANSACTIONS A

specimen, and the measured data were analyzed in relationto the microstructures, mechanical properties obtained fromtensile and Charpy V-notch (CVN) impact tests, work hard-ening, and hardness of the DWTT hammer-impacted region.

II. EXPERIMENTAL

Three kinds of X70-grade pipeline steels, having a yield-strength level above 483 MPa (70 ksi), were used in this study,and their chemical compositions are shown in Table I. Steel Ahad the basic composition of the X70 steel. In steel B, theamounts of carbon, copper, and molybdenum, which increasethe hardenability, were slightly reduced from those of steelA. Steel C had a lower amount of carbon than steel B. Varioushot-rolled specimens were processed from these three steelsby varying rolling conditions, as shown in Table II. Rollingwas started at the two temperatures of 980 °C and 910 °Cand was finished at the two temperatures of 830 °C to 860 °Cand 720 °C to 750 °C, which were the temperatures of theaustenite single-phase region above Ar3 and the temperatureof the (austenite � ferrite) two-phase region below Ar3, respec-tively. After the finish rolling, the specimens rolled in the sin-gle-phase region began to be cooled at 780 °C to 810 °C,while the specimens rolled in the two-phase region began tobe cooled at 690 °C to 710 °C. The finish cooling temperature(FCT) was varied in the range from 100 °C to 600 °C, andthe cooling rate was about 10 °C/s. For convenience, the spec-imens which were rolled in the single-phase region and cooledto different FCTs are referred to as the A1 to A6 specimens,while those rolled in the two-phase region and cooled to dif-ferent FCTs are referred to as the A7 to A12 specimens. Therolling conditions of steels B and C were similar to those ofsteel A, as shown in Table II.

The longitudinal-transverse plane of the rolled specimenswas polished, etched by a nital solution, and observed byan optical microscope. In addition, secondary phases suchas the martensite-austenite (MA) constituent and retainedaustenite (RA) were identified using a two-step etchingmethod,[19] and their volume fraction was measured by animage analyzer. In this two-step etching, specimens wereetched first in an etching solution of 96 pct ethanol, 4 pctpicric acid, and a few drops of hydrochloric acid and werethen etched in a solution of sodium hydroxide andmetabisulfide.

Tensile round bars with a gage diameter of 6 mm anda gage length of 30 mm were prepared in the transversedirection and were tested at room temperature at a crossheadspeed of 5 mm/min using a 10-ton Instron machine. Charpyimpact tests were performed on subsized CVN bars of 7.5 �10 � 55 mm in size and with a transverse-longitudinal(T-L) orientation, in the temperature range from �196 °C to20 °C, using a Tinius Olsen impact tester of 500 J capa-city, in accordance with the ASTM Standard E23-02.[20] Inorder to reduce errors in data interpretation, regression analy-sis for absorbed impact energy vs test temperature was doneusing a hyperbolic tangent curve-fitting method.[21] Basedon the data from regression analysis, the energy transitiontemperature (ETT), which corresponds to the average valueof upper-shelf energy (USE) and lower-shelf energy, wasdetermined.

The DWTT specimens had a size of 76.2 � 305 � 20 mm(full-plate thickness) in the T-L direction, in accordance withthe API RP 5L3 specification.[11,15] These specimens weretested at �15 °C using a DWTT machine with a maximumenergy capacity of 20,000 J. Both the PN DWTT and CNDWTT were conducted on the steel A specimens, while onlythe PN DWTT was done on the steel B and C specimens.

Table I. Chemical Compositions of the X70 Steels used in this Study (Weight Percent)

Steel C Si Mn Cu � Ni � Mo Nb � V � Ti N

A 0.075 0.26 1.54 0.63 0.095 to 0.105 �0.005B 0.070 0.25 1.55 0.48 0.095 to 0.105 �0.005C 0.046 0.26 1.51 0.55 0.095 to 0.105 �0.005

Table II. Rolling Conditions and Microstructures of the X70 Steel Specimens

Start Cooling Finish Cooling

Specimen Temperature Temperature (° C) Cooling Rate (°C/s) Microstructure*

Number (°C) A Steel B Steel C Steel A Steel B Steel C Steel A Steel B Steel C Steel

1 780 to 810 100 — 135 16.0 — 12.3 AF � QPF,BF — AF � QPF2 200 — — 19.0 — — AF � BF,QPF — —3 340 320 — 16.3 10.0 — AF � BF,QPF AF � QPF —4 430 — 390 9.4 — 10.5 AF � QPF — AF � QPF5 480 530 540 14.1 17.0 10.0 AF � QPF AF � QPF QPF,PF � AF6 580 590 615 12.4 11.1 8.8 AF � QPF AF � QPF QPF,PF � AF7 690 to 710 100 — — 12.5 — — PF � M — —8 240 — 220 14.5 — 8.2 PF � M,AF — PF � M,AF9 330 320 265 11.0 7.3 9.0 PF � AF,M PF � AF,M PF � AF,M

10 420 430 — 9.3 7.0 — PF � AF,BF PF � AF,BF —11 505 — 495 12.0 — 7.8 PF � AF — PF � AF12 600 590 590 10.8 9.2 8.3 PF � WF PF � AF PF � AF

*Designation: major matrix � minor phases.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 36A, FEBRUARY 2005—373

After the DWTT, inverse fracture was observed by a scan-ning electron microscope (SEM), and the percent shear area(pct SA) and percent inverse fracture area (pct IFA) weremeasured by an image analyzer. Here, inverse fracture wasincluded in the calculation of the pct SA of the DWTT spec-imen. Figures 1(a) and (b) illustrate the shape and detaileddimensions of the DWTT specimen and its notch.

In order to evaluate the work hardening of specimens aris-ing from the hammer impact after the DWTT, the steel A speci-mens were sectioned perpendicular to the fracture surface ofthe hammer-impacted region, and the hardness of the ham-mer-impacted region was measured under a 300-g load usinga Vickers microhardness tester. On the A1 and A7 specimensshowing the lowest USE and on the A6 and A10 specimensshowing the highest USE, the hardness test was performedunder a 300-g load from the hammer-impacted region towardthe notch up to 10 mm in 2 mm intervals. In order to learnthe effects of work hardening on Charpy impact properties,prestrain was introduced inside the A1, A6, A7, and A10 speci-mens by reducing the thickness of the specimens by 2 and5 pct through cold rolling. Charpy impact tests were performedon them, and the ETT and USE were evaluated.

In addition, the effect of work hardening on the occur-rence of the inverse fracture during DWTT was studiedfrom the tensile test results. The work-hardening exponentwas calculated using two methods: the Hollomon equation(� � K�p

n),[22,23] and the Swift equation (� � a (b ��p)N).[23,24] The work-hardening exponent (n) of theHollomon equation was obtained by measuring the slopeof the straight line in the ln (�)-ln (�p) plot. In the Swiftequation having the correction coefficient marked as b, thework-hardening exponent (N) was obtained from the ln(�)-ln (b � �p) plot.

III. RESULTS

A. Microstructure

Microstructures of the X70 pipeline steel are very com-plicated, but can be roughly divided into polygonal ferrite(PF), quasipolygonal ferrite (QPF), Widmanstätten ferrite(WF), acicular ferrite (AF), granular bainitic ferrite (GB),bainitic ferrite (BF), and martensite (M), in the order ofdecreasing transformation temperature.[25,26,27] The PF is anequiaxed microstructure, and QPF has an irregular grainboundary. The WF is a platetype microstructure initiated atand growing from austenite grain boundaries, while AF isan acicular microstructure formed inside austenite grains andhas an irregular grain boundary. The GB contains equiaxedMA constituents and has well-developed substructures inside.The BF has well-developed laths formed from austenite grainboundaries. Microstructures of the rolled specimens in thisstudy were analyzed in terms of these categories.

Figures 2(a) through (d) are optical micrographs of thesteel A specimens rolled at the single-phase region (speci-mens A1 through A6), and their detailed microstructures aresummarized in Table II. The matrix of these specimens iscomposed of AF overall, and the fraction of BF increasesas the FCT decreases below 430 °C. In specimen A6, fab-ricated at the highest FCT, the fraction of QPF increases(Figure 2(d)). The MA remains less than 1 pct, and the frac-tion of RA is about 2 pct. At a FCT below 430 °C, thefraction of MA increases to about 3 pct, but the fraction ofRA stays at about 2 pct, irrespective of the FCT (Figure 2(a)).Figures 2(e) through (h) are optical micrographs of the steelA specimens rolled in the two-phase region (specimens A7through A12). Unlike the specimens rolled in the single-phase region, whose microstructure does not show muchvariation with the FCT change, PF is present at about 68 pctin the specimens rolled in the two-phase region, and phasestransformed from retained austenite during or after finishcooling, except for PF, vary with FCT.[28,29,30] In specimenA12, fabricated at the FCT of 600 °C, WF is observed to besurrounded by PF transformed prior to cooling (Figure 2(h)).When the FCT decreases to about 400 °C, AF is primarilyobserved instead of WF (Figure 2(g)), and the fraction ofmartensite abruptly increases at the FCT below 330 °C (Fig-ures 2(e) and (f)). Black phases observed in specimen A12(Figure 2(h)) are cementite precipitated inside MA con-stituents. At the FCT below 420 °C, cementite is rarelyobserved, but MA or RA phases are mostly found.

Optical micrographs of the steel B specimens are shownin Figures 3(a) through (f). The steel B specimens rolled inthe single-phase region (specimens B1 through B6) consistmainly of AF and QPF (Figures 3(a) through (c) and Table II).The volume fraction of QPF decreases with decreasing FCT,whereas the fraction of AF increases. In comparison withmicrostructures of the steel A specimens, the RA fractionis considerably reduced, whereas the MA fraction tends toslightly increase. At the FCT of 590 °C, the fraction ofcementite is relatively high at 3 pct, because cementite isprecipitated at interfaces between the PF or QPF phases (Fig-ure 3(c)). The steel B specimens rolled in the two-phaseregion are also composed of 74 pct PF formed prior to fin-ish rolling, fine PF transformed during slow cooling, andsomewhat coarser AF (Figures 3(d) through (f)). As the FCTdecreases, the fraction of PF formed during slow cooling

(a)

Fig. 1—Shape and dimensions of (a) the DWTT specimen and (b) its notch.

(b)

374—VOLUME 36A, FEBRUARY 2005 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 2—Optical micrographs of the (a) A1, (b) A4, (c) A5, (d) A6, (e) A7, ( f ) A9, (g) A10, and (h) A12 steel specimens. Nital etched.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 36A, FEBRUARY 2005—375

Fig. 3—Optical micrographs of the (a) B3, (b) B5, (c) B6, (d) B9, (e) B10, and ( f ) B12 steel specimens. Nital etched.

decreases, the fraction of AF and martensite graduallyincreases, and the grain size tends to decrease.

Figures 4(a) through (h) are optical micrographs of thesteel C specimens. Since the carbon content in the steel C

specimens is lower than that in the steel A or B specimensby about 0.03 wt pct, hardenability is reduced, and thevolume fraction of high-temperature transformed phasestends to increase. In all the steel C specimens rolled in the

376—VOLUME 36A, FEBRUARY 2005 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 4—Optical micrographs of the (a) C1, (b) C4, (c) C5, (d) C6, (e) C8, ( f ) C9, (g) C11, and (h) C12 steel specimens. Nital etched.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 36A, FEBRUARY 2005—377

single-phase region (specimens C1 through C6), the frac-tion of QPF is high, and the microstructure is composedmostly of QPF and PF at the FCT above 390 °C (Fig-ures 4(b) through (d) and Table II). Due to the reduced car-bon content, the fraction of total secondary phases is lowerthan in the steel A or B specimens. Secondary phases aremostly observed around prior-austenite grain boundaries,and coarse MA is formed at the FCT of 135 °C and 615 °C(Figures 4(a) and (d)). The steel C specimens rolled in thetwo-phase region (specimens C7 through C12) consist of88 pct PF transformed before cooling, and AF takes up mostof the rest (Figures 4(e) through (h)). At the FCT of 265 °C,martensites are present at about 4 pct, and their fractionsignificantly increases to 8 pct at the low FCT of 220 °C.At the FCT of 495 °C, cementites together with MA con-stituents are also formed at AF boundaries or sub-boundaries(Figure 4(g)).

B. Tensile Properties

Variations in tensile strength, yield strength, and elongationas a function of FCT of the specimens rolled in the single-phase region and two-phase region are shown in Fig-ures 5(a) and (b), respectively. Since the steel specimensfabricated in this study show a yield strength of over483 MPa, all satisfy the strength requirement of X70-gradepipeline steels. Compared with the specimens rolled in thetwo-phase region, those rolled in the single-phase regionshow less variation in tensile and yield strengths, becausethe matrix stays almost the same and the volume fraction ofphases varies within 10 pct. In the specimens rolled in the

(a) (b)

Fig. 5—Tensile strength, yield strength, and elongation vs finish cooling temperature for the steel specimens rolled at the (a) single- and (b) two-phase regions.

two-phase region, the tensile and yield strengths show a largevariance because about 15 to 30 vol pct of RA formed priorto cooling varies with FCT. The yield strength is at a maxi-mum at the FCT of about 500 °C in the specimens rolledin the single-phase region, while it is at about 400 °C inthose rolled in the two-phase region. The yield strengthtends to increase with decreasing FCT because of a grain-refinement effect and increased volume fraction of phasestransformed at low temperatures. However, in the specimensrolled in the single-phase region, the yield strength dropsas the FCT decreases below 300 °C. This is because MAgenerates lots of dislocations at boundaries of adjacent duc-tile phases, when MA is present in ductile phases such asPF and AF formed in the specimens rolled in the single-phase region. As the higher volume fraction of hard phasessuch as MA leads to an increase in the tensile strength,MA plays a role in decreasing the yield ratio (yieldstrength/tensile strength) of the steel specimens.[31,32]

C. Charpy Impact Properties

Figures 6(a) through (d) show the results of the USE andETT tests as a function of FCT. The specimens rolled in thesingle-phase region show an overall higher USE than thespecimens rolled in the two-phase region. The USE of thespecimens rolled in the single-phase region is almost onthe same level in the steel A and B specimens, and it ishigher in the steel C specimens than in other steel speci-mens at the FCT above 400 °C. The variation of USE in thespecimens rolled in the two-phase region shows almost thesame relationship to the FCT in all the three steels, and the

378—VOLUME 36A, FEBRUARY 2005 METALLURGICAL AND MATERIALS TRANSACTIONS A

the 2 and 5 pct prestrained specimens is lower than that ofthe 0 pct prestrained ones.

D. The DWTT Properties

Figures 8(a) through (f) show the overall fracture surfacesof the DWTT specimens. In the hammer-impacted regionof most DWTT specimens, inverse fracture of a cleavagetype is found, as marked by arrows. In many modern pipelinesteels, inverse fracture is frequently observed in DWTT speci-mens and is particularly expected to become more seriousas the DWTT temperature decreases and the toughness andthickness of pipeline steels increase.[6–12] In the steel B spec-imens, separations on DWTT fractured surfaces occur moreoften than in the other specimens (Figures 8(c) and (d)) and,thus, reduce the USE.[1,2,8] All the specimens were valid inaccordance with the API RP 5L3 specification since cleav-age fracture was initiated from the notch, and the pct SAwas calculated including the inverse fracture.

Observation of the fracture surface of the DWTT speci-men reveals that the cleavage fracture initiated from the notchis transformed to a shear fracture after propagating over acertain distance and then returns into a cleavage fracture on

(a) (b)

(c) (d)

Fig. 6—(a) and (b) USE and (c) and (d) ETT obtained from Charpy impact test vs finish cooling temperature.

USE decreases in the order of steels C, A, and B. The USEsof all the specimens rolled in the two-phase region reach thehighest levels at the FCT of 450 °C to 550 °C and thendecrease. The ETT in the specimens rolled in the single-phase region is extremely low (about �130 °C to �140 °C)in the steel C specimens at the FCT of 400 °C to 550 °Cand is almost on the same level in the steel A and B speci-mens. The steel A and B specimens rolled in the two-phaseregion show the lowest ETT at the FCT of 350 °C to 450 °C,and the steel C specimens show a peculiar tendency that theETT decreases with decreasing FCT.

The Charpy impact test data of the 2 and 5 pct prestrainedspecimens are shown in Figures 7(a) through (d). Table Ilists the USE, ETT, and increase in ETT values obtainedfrom these figures. The ETT rises with increasing prestrain,while the USE tends to decrease. The ETT of the 2 pct pre-strained specimen rolled in the single-phase region is slightlyhigher by 5 °C to 10 °C than that of the 0 pct prestrainedspecimens, while the ETT of the 2 pct prestrained specimenrolled in the two-phase region is considerably higher by 30 °Cto 40 °C. The ETT of the 5 pct prestrained specimens, eitherrolled in the single- or two-phase regions, is slightly higherthan that of the 2 pct prestrained specimens. The USE of

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 36A, FEBRUARY 2005—379

the inverse fracture surface occurring in the hammer-impactedregion. Figure 9(a) provides a schematic diagram of a typicalDWTT fracture surface, and Figure 9(b) is a low-magnificationSEM fractograph of the marked rectangular area in Figure 9(a).The upper part of Figure 9(a) shows the shear-fractured duc-tile fracture area. Cleavage fracture initiates at the arrow-marked area below the ductile fracture area and propagatesin a radial manner. This indicates that the arrow-marked areais where the inverse fracture of the cleavage-fracture modeinitiates. A typical SEM fractograph of the inverse fractureis shown in Figure 9(c). These fractographs clearly show thatthe inverse fracture is formed when the cleavage fracture ini-tiated at the notch propagates in a shear-fracture mode andthen transforms again into cleavage fracture near the hammer-impacted region.

The variations in pct SA and pct IFA as a function ofFCT are shown in Figures 10(a) through (d). The pct SAof the specimens rolled in the single-phase region is aboutthe same in the three steels for the FCT below 400 °C anddecreases considerably for the FCT above 500 °C in thesteel A and B specimens (Figure 10(a)). In the specimensrolled in the two-phase region, the pct SA of the steel Band C specimens is slightly lower than that of the steel A

specimens for the FCT above 400 °C, but all show excel-lent DWTT properties (over 85 pct). The overall DWTTproperties are excellent for the FCT below 400 °C, irres-pective of the starting rolling temperature, and they maydeteriorate for the FCT above 400 °C because of the largeformation of inverse fracture. The pct IFAs of the speci-mens rolled in the single-phase region are low and exhibitsimilar trends in the three steels for the FCT below 500 °C(Figure 10(c)). For the FCT of 600 °C, the pct IFAs of steelsA and B increase abruptly, while those of steel C remainnearly constant. The pct IFAs of the specimens rolled in thetwo-phase region are generally low for the FCT below300 °C and increase for the FCT above 400 °C (Figure 10(d)).Those of steels B and C are somewhat higher than those ofsteel A.

Figures 11(a) through (d) provide the Vickers hardnessdata as a function of distance from the hammer-impactedregion of the DWTT specimen of the steel A specimen. Thehardness in the hammer-impacted region shows little dif-ference with the notch-type or microstructure and decreasesas the distance from the impacted location increases. How-ever, the hardness increase of 20 to 50 VHN in comparisonto that of the base metal far from the impacted region

(a) (b)

(c) (d)

Fig. 7—Effect of prestrain on Charpy absorbed energy vs test temperature for the (a) A1, (b) A6, (c) A7, and (d) A10 specimens.

380—VOLUME 36A, FEBRUARY 2005 METALLURGICAL AND MATERIALS TRANSACTIONS A

is observed in all the specimens. This confirms the work-hardening effect caused by the hammer impact.[12,16]

IV. DISCUSSION

The X70-grade steel used in the present study is a high-toughness pipeline steel having a high USE and low ETT.

Inverse fracture heavily occurred in the hammer-impactedregion of most of the DWTT specimens. The reduced tough-ness in the hammer-impacted region will lead to a larger for-mation of cleavage-type inverse fracture, which makes thesatisfaction of the 85 pct shear-appearance transition temper-ature quite difficult even in the case of inherently toughpipeline steels. The size of the inverse fracture is not feasi-ble for quantitative approach, because scatter is significant

Fig. 8—Overall fracture surfaces appeared in DWTT specimens of the (a) and (b) A steel specimens, (c) and (d) B steel specimens, and (e) and ( f ) C steelspecimens. (a), (c) and (e) DWTT specimens of the steel specimens rolled at the single-phase region, and (b), (d), and (f) those of the steel specimens rolledat the two-phase region.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 36A, FEBRUARY 2005—381

even in an identical specimen because of the variance inmicrostructure or the intrinsic characteristic of the impact test-ing method.[33] Consequently, it is desirable to qualitativelyevaluate the overall tendency through comparative analysis,instead of quantitatively evaluating the size and presence ofthe inverse fracture. In this study, the effects of prestrain onthe toughness of high-toughness pipeline steels were analyzed,and the effects of the reduced toughness due to prestrain onthe inverse fracture of PN and CN DWTT specimens wereinvestigated. This study further discussed the correlationbetween mechanical properties, which were obtained fromtensile and Charpy impact tests, and the inverse fracture.

A. Effects of Notch Type on Pct IFA

Figures 12(a) and (b) present the USE and pct IFA of thesteel A specimens rolled in the single- and two-phase regionsvs finish cooling temperature. The pct IFA stays almost steadyup to an FCT of 500 °C in the specimen rolled in the single-phase region, while it increases up to an FCT of 400 °C andthen drops in the specimen rolled in the two-phase region.In all the specimens, the effects of the FCT change are simi-lar, irrespective of the kind of notches. Comparing the pctIFA of the steel A specimens with the USE, specimens rolledin the single- and two-phase regions show the highest pct

IFA at the FCT having the highest USE, and variations ofthe USE and pct IFA with the FCT are almost the same. Thisindicates that the impact absorption energy is partially relatedto the formation and size of the inverse fracture.

According to Figures 12(a) and (b), showing the effects ofnotch type on pct IFA, the pct IFA of the steel A specimensrolled in the single-phase region is higher in the PN DWTTthan in the CN DWTT at all the FCTs, but not so in the speci-mens rolled in the two-phase region. The specimens rolledin the single-phase region require a higher energy for the frac-ture owing to the high absorption energy of the specimen itselfand to the high fracture-initiation energy of the PN type. Thus,much higher compressive prestrain is caused in the hammer-impacted region of the PN DWTT specimen, and, conse-quently, much more inverse fracture occurs.[6,11] However,in the specimens rolled in the two-phase region, this phe-nomenon is unlikely to be observed because the absorptionenergy is considerably lower than in the specimens rolled inthe single-phase region. Although the current API RP 5L3specification recommends the use of CN DWTT specimensduring the DWTT for high-toughness steels and PN DWTTspecimens for low-toughness steels,[15] there are no clear stan-dards on the notch type to be used depending on the tough-ness. Therefore, if the suppliers and users of pipeline steelsuse different notches for the DWTT, problems may arise in

Fig. 9—(a) Schematic diagram of an overall fracture surface typically observed in DWTT specimens and (b) and (c) SEM fractographs of typical inversefracture occurring in the hammer-impacted region. (b) was taken from the marked rectangular area in (a).

382—VOLUME 36A, FEBRUARY 2005 METALLURGICAL AND MATERIALS TRANSACTIONS A

specifying the supply conditions, as the pct SFA can beaffected by the inverse fracture, depending on the notch type.

B. Correlation between Impact Absorption Energyand Pct IFA

The relations between pct IFA and USE are presented inFigures 13(a) and (b) and 14(a) and (b). In the specimensrolled in the two-phase region, those having the USE below200 J show a low pct IFA, but some specimens having anUSE above 200 J show a high pct IFA (above 5 pct) (Fig-ures 13(b) and 14(b)). In the specimens rolled in the single-phase region, those having a high USE show a higher pctIFA, but they all generally have low pct IFA values below3 pct (Figures 13(a) and 14(a)). Figures 13 and 14 indicatethat the pct IFA can increase abruptly above a certain USElevel in the specimens rolled in the single- and two-phaseregions, even though both specimens show some variationsdepending on the notch type or the steel. The reason behindthe higher pct IFA of the PN DWTT than the CN DWTTin the steel A specimens rolled in the single-phase region,which have relatively higher absorption energy, can also

apply to the greater formation of inverse fracture in the spec-imens having higher absorption energy. In other words, thespecimens having a higher Charpy impact absorption energy,most of which is related to fracture initiation, require greaterenergy for the fracture initiation during the DWTT. In thiscase, much higher compressive prestrain occurs in the ham-mer-impacted region during the fracture process, the tough-ness decreases, and the tendency for cleavage fractureincreases. Consequently, pipeline steels having an extremelyhigh USE need considerable care during the evaluation ofthe DWTT results due to the presence or absence of inversefracture and its size. If the inverse fracture is heavily formedduring the DWTT of pipeline steels and the large inversefracture is included in the cleavage fracture when measuringthe pct SFA under the current API RP 5L3 specification,the pct SFA decreases and, thus, the transition temperaturecan be artificially high.

C. Effects of Prestrain on Toughness and Pct IFA

The effects of prestrain on the Charpy impact properties ofthe specimens rolled in the single- and two-phase regions were

(a) (b)

(c) (d)

Fig. 10—(a) and (b) Percent shear area (pct SA) and (c) and (d) percent inverse fracture area of the fractured PN DWTT specimens of steel specimensrolled at the single-phase region and two-phase region vs finish cooling temperature.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 36A, FEBRUARY 2005—383

examined so as to elucidate the correlation between the for-mation mechanism of inverse fracture and the toughness vari-ation with prestrain (Figures 7(a) through (d) and Table III).

In the case of the DWTT specimen, compressive prestrain isintroduced into only some parts of the specimen, whereas inthe prestrained Charpy impact specimen, compressive prestrain

(a) (b)

(c) (d )

Fig. 11—Vickers hardness vs distance from the hammer-impacted side of the PN DWTT and CN DWTT specimens for the (a) A1, (b) A6, (c) A7, and(d) A10 specimens.

(a) (b)

Fig. 12—USE and pct IFA of the A steel specimens rolled at the (a) single- and (b) two-phase regions vs finish cooling temperature.

384—VOLUME 36A, FEBRUARY 2005 METALLURGICAL AND MATERIALS TRANSACTIONS A

is artificially introduced by cold rolling of the whole specimen.Although the amount and distribution of prestrain presentinside the prestrained specimens may vary, it can be useful

(a) (b)

Fig. 13—Relationship between pct IFA and USE of the A steel specimens rolled at the (a) single- and (b) two-phase regions.

(a) (b)

Fig. 14—Relationship between pct IFA and USE of all the specimens rolled at the (a) single- and (b) two-phase regions.

Table III. Charpy Impact Test Results of the SteelSpecimens Prestrained by Cold Rolling

Prestrain Increase inSpecimen (Pct) USE (J) ETT (°C) ETT

A1 0 185.3 �92.2 —2 152.2 �81.5 10.75 165.6 �78.7 13.5

A6 0 238.5 �87.2 —2 186.4 �82.9 4.35 211.3 �80.0 7.2

A7 0 109.8 �107.1 —2 97.4 �75.9 31.25 108.8 �71.3 35.8

A10 0 214.4 �120.8 —2 160.5 �79.9 40.95 183.1 �78.1 42.7

for understanding the effect of prestrain on fracture proper-ties. According to the variation in the ETT vs prestrain asshown in Table I, the increase in the ETT of the specimensrolled in the single-phase region is not much larger than thatof those rolled in the two-phase region. This can be attributedto microstructural differences and indicates that the primaryphase of AF in the specimens rolled in the single-phase regionhas less toughness reduction than the primary phase of PF inthe specimens rolled in the two-phase region.

Pussegoda et al.[34] investigated the variation in transitiontemperature through a Charpy impact test and crack-tipopening displacement test[35] after introducing tensile andcompressive prestrains into steels having different microstruc-tures. They reported that steels having finer ferrite grainsdue to accelerated cooling showed a smaller increase in tran-sition temperature due to prestrain than the typical hot-rolledsteels having coarse ferrite grains. This is consistent withthe research results of this study and indicates that it isnecessary to consider the formation of subgrains or dislo-cation structures inside the microstructure as well as ferritegrain size. The AF of the specimens rolled at the single-phase

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 36A, FEBRUARY 2005—385

region can be considered as fine ferrite having a high dis-location density and subgrains, while the PF, the primarymicrostructure of the specimens rolled at the two-phaseregion, can be considered as coarse ferrite having a lowerdislocation density than AF, as it has already been formedin the two-phase region before accelerated cooling. Theincrease in transition temperature generally means a decreasein toughness due to prestrain and can partially explain thecorrelation between the formation of inverse fracture andthe variation in absorption energy.

Comparison of the pct IFA of the three steels after the PNDWTT (Figures 10(c) and (d)) shows, in general, a higher pctIFA in the specimens rolled in the two-phase region than inthose rolled in the single-phase region, except for a few speci-mens rolled in the single-phase region, which had serious inversefracture at the FCT of 600 °C. Also, according to the prestrainedCharpy impact data (Table III), the specimens rolled in the two-phase region show a larger reduction in toughness than thoserolled in the single-phase region. Relating this with the inversefracture formation, it is likely that the specimens rolled in thetwo-phase region form more inverse fracture than those rolledin the single-phase region, as the former might experience agreater reduction in toughness in the hammer-impacted regionthan the latter, even if the amount of compressive prestrain intro-duced inside the DWTT specimen is the same in both.

D. Correlation between Work Hardening and Pct IFA

Figure 15 shows the correlation between the hardness andpct IFA measured in the hammer-impacted region of steel A.It does not show any difference due to notch type or a directcorrelation between hardness and pct IFA, but the maximumvalue of the pct IFA slightly increases with increasing hard-ness in the hammer-impacted region. The correlation betweenhardness and pct IFA in the hammer-impacted region isnot so obvious as the relation between USE and pct IFA,as detailed in Figures 13 and 14.

Table IV provides the factors related to work harden-ing obtained from the Hollomon and Swift equations, andFigures 16(a) through (d) show the correlation between

these factors and pct IFA. Two work-hardening exponents,n and N, obtained from the Hollomon and Swift equations,respectively, show some variations with specimens, buthave a similar trend. According to Figures 16(a) and (c),the work-hardening exponents and pct IFA do not showany particular relationship, as the test results are mostlycrowded in one area. The material constants, b and b/N,calculated from the Swift equation, tend to be proportionalwith the pct IFA, although the proportionality might bemore or less ambiguous (Figures 16(b) and (d)). The mate-rial constants of the A6, A10, A11, and A12 specimens,showing heavy formation of inverse fracture, are 0.009,0.008, 0.011, and 0.008, respectively, which are all over0.005.

The present results indicate that the formation of inversefracture is affected by the microstructure of specimens andtheir intrinsic properties. However, as these factors aresubsidiary ones showing up in developing high-strength, high-toughness pipeline steels, it is not desirable to controlcompositions or to change fabricating methods for the solepurpose of reducing inverse fracture. Instead, we have to comeup with some fundamental solutions for this problem. Althoughthe DWTT developed some 40 years ago is still widely usedas a test to evaluate the fracture-propagation transition tem-perature (FPTT) and toughness of pipeline steels, it poses prob-lems for testing the properties of currently developed high-strength pipeline steels such as API X100 and API X120,[18,30]

because they are not included in the API RP 5L3 specifica-tion. The present DWTT method may be used after someadjustments, but the problem of reliability remains. Althoughthe reliability of the DWTT can be verified by conducting afull-scale burst test[11,17,36] simultaneously with the DWTT, thisis expensive. Furthermore, even if the reliability of the DWTTfor currently developed high-strength pipeline steels is veri-fied, it cannot apply likewise to new pipeline steels having dif-ferent microstructures and properties that will come in thefuture, and it may also cause other problems because of thetesting method itself. Thanks to continuous advancement infracture mechanics after the 1960s and 1970s,[37] many stud-ies are under way to evaluate the resistance to unstable or run-ning ductile-fracture propagation. For example, a parameter

Table IV. Work-Hardening Exponent and MaterialsConstants Measured from Tensile Tests of the

Steel Specimens

Hollomon Equation Swift Equation

Specimen n* b** N* b/N

A1 0.153 0.0035 0.1431 0.0245A2 0.145 0.0035 0.1363 0.0257A3 0.134 0.0030 0.1289 0.0233A4 0.126 0.0030 0.1249 0.0240A5 0.101 0.0050 0.0999 0.0500A6 0.141 0.0090 0.1669 0.0539A7 0.147 0.0035 0.1450 0.0241A8 0.158 0.0030 0.1377 0.0218A9 0.151 0.0040 0.1508 0.0265A10 0.112 0.0080 0.1269 0.0630A11 0.118 0.0110 0.1655 0.0665A12 0.137 0.0080 0.1686 0.0474

*Work-hardening exponent.**Materials constant in the swift equation.[24]

Fig. 15—Relationship between pct IFA and hardness of the hammer-impacted region of all the specimens rolled at the single-phase region.

386—VOLUME 36A, FEBRUARY 2005 METALLURGICAL AND MATERIALS TRANSACTIONS A

based on fracture mechanics called crack-tip opening angle,instead of the conventional Charpy impact test or DWTT,has been introduced as a method to investigate characteristicsof ductile-fracture propagation and arrest, but has not yetachieved widespread attention.[38–41] Since the current DWTTentails many abnormal fracture appearances during testing andit is a qualitative testing method, not based on fracture mechan-ics, it does not seem to be a reliable guideline to guarantee thestructural stability of pipeline steels. In order to procure fun-damental fracture data and precisely predict the FPTT of actualpipeline steels, therefore, new testing methods based on frac-ture mechanics are required in steel or structural designing ofvarious pipeline steels.

V. CONCLUSIONS

In the present study, the inverse fracture occurring in thehammer-impacted region was analyzed after conducting theDWTT on high-toughness X70 pipeline steels.

1. Observation of the inverse fracture revealed that the inversefracture of a cleavage-fracture mode was not initiated from

the hammer-impacted region, but that the cleavage frac-ture initiated at the notch was propagated in a shear-fracturemode and then transformed to cleavage fracture again toform inverse fracture near the hammer-impacted region.

2. The size of the inverse fracture occurring in the hammer-impacted region was found to be related to the USEobtained from the Charpy impact test. Since steels hav-ing a higher Charpy impact absorption energy requiredhigher energy for fracture initiation and propagation dur-ing the DWTT, a higher compressive prestrain was intro-duced in the impacted region. Thus, the toughnessdecreased, and the tendency of cleavage fracture increased.

3. Since steels containing polygonal ferrite, which wascoarser than acicular ferrite, experienced more reduc-tion in toughness due to compressive prestrain, the largerinverse fracture occurred in the hammer-impacted regionduring the DWTT.

ACKNOWLEDGMENTS

This work was financially supported by the NationalResearch Laboratory Program of the Ministry of Science and

(a) (b)

(c) (d)

Fig. 16—Relationship between pct IFA and (a) work-hardening exponent (n), obtained from the Hollomon equation, and (b) materials constant (b), (c) work-hardening exponent (N), and (d) b/N value obtained from the Swift equation for the A steel specimens.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 36A, FEBRUARY 2005—387

Technology of Korea and by POSCO under Contract No.PL-03909. The authors thank Drs. Wung Yong Choo, ChongSoo Woo, Dong Han Suh, and Seong Soo Ahn, POSCO, andMessrs. Young Min Kim and Han Sang Lee, POSTECH, fortheir help of the testing and data analyses of DWTT.

REFERENCES1. I. Tamura, H. Sekine, T. Tanaka, and C. Ouchi: Thermomechanical

Processing of High-Strength Low-Alloy Steels, Butterworth & Co., Ltd,London, 1988.

2. R. Denys: Pipeline Technology, Proc. 3rd Int. Pipeline TechnologyConf., Elsevier, Brugge, Belgium, 2000, vol. I & II.

3. J.R. Ellwood: Proc. 2000 Int. Pipeline Conf., ASME, Calgary, AB,Canada, 2000.

4. X.W. Zhao, J.H. Luo, M. Zheng, M.X. Lu, and H.L. Li: Met. Mater.Int., 2002, vol. 8, pp. 479-85.

5. J.S. Chung and S. Prinsenberg: Proc. 13th Int. Offshore and PolarEngineering Conf. and The ISOPE Symp. on High-Performance Mater-ials in Offshore Industry (HMOI), Honolulu, HI, 2003.

6. N. Iwasaki, T. Yamaguchi, and T. Taira: Mech. Work Steel Process,1975, vol. 13, pp. 294-314.

7. H. Kashimura, M. Ogasawara, and H. Mimura: Metal Prog., 1976,Nov., pp. 58-62.

8. N. Nozaki, K. Bessyo, Y. Sumitomo, I. Takeuchi, and A. Yamashita:Sumitomo Search, 1981, vol. 26, pp. 76-90.

9. K. Seifert: Mater. Testing, 1984, vol. 26, pp. 277-80.10. L. Pussegoda, L. Malik, A. Dinovitzer, B.A. Graville, and A.B. Rothwell:

in Proc. 2000 Int. Pipeline Conf., J.R. Ellwood, ASME, New York,NY, 2000, vol. 1, pp. 239-45.

11. G. Mannucci and D. Harris: “Fracture Properties of API X100Gas Pipeline Steels,” Final Report, European Commission, Brussels,Belgium, 2002.

12. B. Hwang, S. Lee, Y.M. Kim, N.J. Kim, J.Y. Yoo, and C.S. Woo:Mater. Sci. Eng. A, 2003, vol. A368, pp. 18-27.

13. S. Lee, B. Hwang, Y.M. Kim, N.J. Kim, and J.Y. Yoo: Proc. 17thConf. on Mechanical Behaviors of Materials, J.H. Hong and Y.S. Kim,eds., Pohang, Korea, 2003, pp. 40-49.

14. B. Hwang, Y.M. Kim, S. Lee, N.J. Kim, and S.S. Ahn: Metall. Mater.Trans. A, in press.

15. API Recommended Practice 5L3, API, Washington, D.C., 1996.16. E. Heier: “Drop Weight Tear Testing of High Toughness Pipeline

Material,” Technical Report, Det Norske Veritas, Norway, 2003.17. G.M. McClure, A.R. Duffy, and R.J. Eiber: J. Eng. Industry, 1965,

vol. 4, pp. 265-78.

18. J.Y. Koo, M.J. Luton, N.V. Bangaru, R.A. Petkovic, D.P. Fairchild,C.W. Petersen, H. Asahi, T. Hara, Y. Terada, M. Sugiyama, H. Tamehiro,Y. Komizo, S. Okaguchi, M. Hamada, A. Yamamoto, and I. Takeuchi:Proc. 13th Int. Offshore and Polar Engineering Conf., J.S. Chung andS. Prinsenberg, eds., Honolulu, HI, 2003, pp. 10-18.

19. A.K. De, J.G. Speer, and D.K. Matlock: Adv. Mater. Process, 2003,vol. 161, pp. 27-30.

20. ASTM Standard E23-02, ASTM, Philadelphia, PA, 2002.21. W. Oldfield: ASTM Standardizations News, 1975, pp. 24-29.22. J.H. Hollomon and J.D. Lubahn: Phy. Rev., 1946, vol. 70, p. 77523. S.K. Kim: Ph.D. Dissertation, POSTECH, Pohang, Korea, 2002.24. H.W. Swift: Met. Ind., 1940, vol. 56, pp. 127-3025. Atlas for Bainitic Microstructures, ISIJ, Tokyo, 1992, vol. 1.26. T. Hayashi, F. Kawabata, and K. Amano: Conf. Proc. from Materials Solu-

tion ’97 on Accelerated Cooling/Direct Quenching of Steels, R. Asfahani,ed., ASM INTERNATIONAL, Materials Park, OH, 1997, pp. 93-99.

27. G. Krauss and S.W. Thompson: Iron Steel Inst. Jpn. Int., 1995, vol. 35,pp. 937-45.

28. N.J. Kim, A.J. Yang, and G. Thomas: Metall. Trans. A, 1985, vol. 16A,pp. 471-74.

29. J.Y. Yoo and J.S. Woo: Proc. Int. Pipe Dreamer’s Conf., M. Toyodaand R. Denys, eds., Scientific Surveys, Ltd., Yokohama, Japan, 2002,pp. 441-56.

30. S. Okaguchi, H. Makino, M. Hamada, A. Yamamoto, T. Ikeda,I. Takeuchi, D.P. Fairchild, M.L. Macia, S.D. Papka, J.H. Stevens, C.W.Perersen, J.Y. Koo, N.V. Bangaru, and M.J. Luton: Proc. 13th Int. Off-shore and Polar Engineering Conf., J.S. Chung and S. Prinsenberg,eds., Honolulu, HI, 2003, pp. 36-42.

31. B.C. Kim, S. Lee, N.J. Kim, and D.Y. Lee: Metall. Trans. A, 1991,vol. 22A, pp. 139-49.

32. S. Kim, S. Lee, Y.R. Im, H.C. Lee, S.J. Kim, and J.H. Hong: J. Kor.Inst. Met. Mater., 2002, vol. 42, pp. 939-48.

33. W. Späth: Impact Testing of Materials, Gordon and Breach, New York,NY, 1961, ch. IX.

34. L.N. Pussegoda, L. Malik, and W.R. Tyson: Can. Metall. Q., 1997,vol. 36, pp. 39-47.

35. G.E. Dieter: Mechanical Metallurgy, McGraw-Hill, London, 1988.36. S.D. Papka, J.H. Stevens, M.L. Macia, D.P. Fairchild, and C.W.

Petersen: Int. J. Offshore and Polar Eng., 2004, vol. 14, pp. 42-51.37. T.L. Anderson: Fracture Mechanics—Fundamentals and Applications,

CRC Press, Boca Raton, FL, 1995.38. P.E. O’Donoghue, M.F. Kanninen, C.P. Leung, G. Demofonti, and

S. Venzi: Int. J. Pressure Vessel Piping, 1997, vol. 70, pp. 11-25.39. C.G. Lee, S.-J. Kim, B.-H. Song, and S. Lee: Met. Mater. Int., 2002,

vol. 8, pp. 435-41.40. C.-M. Kim, W.-S. Kim, and Y.-T. Kho: Met. Mater. Int., 2002, vol. 8,

pp. 197-202.41. D.J. Horsley: Eng. Fract. Mech., 2003, vol. 70, pp. 547-52.