July 18, 2008

A Comparison of Mechanical Properties and Hydrogen Embrittlement Resistance of Austempered Vs Quenched and Tempered 4340 Steel

By Tartaglia, John M Lazzari, Kristen A; Hui, Grace P; Hayrynen, Kathy L

This study was conducted to compare the hydrogen embrittlement (HE) resistance of austempered 4340 steel with quenched and tempered (Q&T) 4340 steel with an identical yield strength (YS) of 1340 MPa (194 ksi). A baseline comparison showed that the austempered steel with a lower bainite microstructure exhibited higher hardness, tensile strengths, Charpy V-notch (CVN) impact toughness, and ductility at both low 233 K (-40 F) and ambient temperatures, as compared to the Q&T steel with a martensite microstructure. After machining and just prior to testing, subsized CVN specimens and notched bend specimens were immersed in hydrochloric acid-water baths. The HE resistance was higher for the austempered steel than the Q&T steel. No differences in room-temperature CVN energy resulted from hydrogen charging of the austempered and Q&T steels vs their unexposed counterparts. However, in the notched bend specimens, the hydrogen charging caused significant peak load decreases (40 pct) for the Q&T steel, while the austempered steel exhibited only small (6 pct) decreases in peak load. Intergranular (IG) fracture occurred solely in the charged Q&T bend samples, which is further evidence of their embrittlement. DOI: 10.1007/s11661-007- 9451-8

(c) The Minerals, Metals & Materials Society and ASM International 2008


A. Practical Significance and Study Purpose

Quenched and tempered (Q&T) processing of medium-carbon steels is the most common form of hardening and strengthening heat treatment. As a result, many researchers have extensively characterized the common properties and problems of Q&T steels, e.g., environmental embrittlement susceptibility under certain conditions.

The use of austempered steels is less common, but these steels offer advantages over Q&T steels in some applications that require limiting distortion and residual stress. Early work also showed that austempered steels with lower bainite microstructures offer improved Charpy impact toughness over their Q&T counterparts with predominantly martensitic microstructures. Figure 1, from the ASM Handbook, shows that austempered steels exhibit moderately (but significantly) lower Charpy V-notch (CVN) impact ductile-to-brittle transition temperatures (DBTTs) and higher upper shelf energies (USEs) at comparable strengths.[1]

Since austempered steel has greater toughness than Q&T steel in the absence of embrittlement, it is reasonable to investigate whether austempered steel also has greater toughness when subjected to environmental conditions that deleteriously affect Q&T steel. After conducting a comprehensive literature search, the present authors found some articles that suggested that austempered steels might have greater hydrogen embrittlement (HE) resistance than their Q&T counterparts, but the comparisons were limited, for various reasons. For example, Chuang et al.[2] reported that the fatigue crack growth rates of Q&T 4140 steels were greater than those in austempered samples; however, the experimental steel they used was exceptionally dirty (0.093 pct S) and they conducted their Q&T heat treatments in the blue-brittle range of 4140 steel, which caused their Q&T samples to be temper embrittled in the absence of hydrogen exposure. Reynolds and Hayes[3] reported that austempering exhibited less HE than Q&T heat treatment in the bend testing of spring steel and Lantsman et al.[4] reported similar results in a brief note.

The present study represents the first straightforward comparison, in a widely available and applicable open publication, of the HE resistance of a low-alloy steel that was heat treated by two different methods: austempering vs quench and tempering. A secondary benefit of this study is that other properties of austempered steel were determined and openly published for the first time.

B. Heat-Treatment Effects on Microconstituent Formation and Toughness

It is worthwhile to briefly define the microstructural constituents that are the subject of this article. Figure 2 shows[5] a continuous cooling transformation (CCT) diagram for an alloy steel. A fully martensitic microstructure forms when this steel cools faster than the critical cooling rate for martensite (CCR^sub M^) formation. Note that this cooling rate is the fastest cooling rate superimposed on Figure 2.

There are various bainite microstructures and, for this study, it is important to define the bainite formed during austempering. Bainite forms through the decomposition of austenite-to-acicular ferrite and carbides at a temperature above the martensite-start (M^sub s^) temperature. The two primary forms of bainite are upper bainite and lower bainite. In upper bainite, the carbides are typically located between the acicular ferrite grains; however, in lower bainite, the carbides tend to precipitate at an inclined angle to the major growth direction or longitudinal axis of the acicular ferrite grains.[6]

Figure 2 shows that a fully bainitic (B) micro structure can form by continuous cooling (slack quenching) at the cooling rate labeled 3. This type of bainite usually forms inadvertently during Q&T treatments intended to form all martensite, due to insufficient alloy content, oversized sections, or insufficient quench speed. Bainite formation can also occur when a slack quench is intentionally used to avoid quench cracking or residual stress formation. This Q&T process corresponds to the cooling rate labeled 2 and results in a micro structure containing bainite interspersed within a martensitic matrix. When bainite forms by continuous cooling, it is possible to form an acicular structure along with austenitic-martensitic regions. A final mixed microstructure results when bainite forms by continuous cooling or slack quenching.

Bainite transformation occurs by continuous cooling over a range of temperatures. As a result, bainite formed by continuous cooling is usually considered undesirable, because its microstructural refinement is inconsistent.

Austempering utilizes a quench temperature above the M^sub s^ temperature, with an isothermal transformation (IT) time sufficient to exceed the bainite-finish time. When bainite forms by IT or austempering (lines labeled 4 and 5 in Figure 2), the resulting microstructure is uniform and refined, as compared to bainite that forms during continuous cooling. Steels with high alloy contents may have excessively long bainite-finish times. In such instances, a small percentage of martensite may be present in the final microstructure. However, this microstructure is not like a slack quench microstructure, in which the predominant matrix microconstituent is martensite rather than bainite.

Tempered martensite has higher toughness than that of the upper bainite that forms during continuous cooling or slack quenching.[7] Figure 3 shows that low-and medium-carbon-alloy steels with tempered bainitic microstructures or tempered martensite-bainite mixtures exhibit significantly higher CVN DBTTs than tempered martensite, at all carbon contents. Although tempered martensite exhibits better transition toughness than tempered bainite, Figure 3 shows that the USEs of bainite and martensite are roughly equivalent, regardless of carbon content. However, lower bainite formed by austempering (and without further tempering) has higher toughness than does tempered martensite at high hardness levels, as shown in Figure 1 and by recent authors.[8] Another ASM Handbook volume[9] and a recent article[10] contain a more complete comparison and fundamental discussion of martensite and lower bainite structure-property relationships.

C. Hydrogen Embrittlement

The literature contains some indications that austempered steel has superior HE resistance,[2-4] but this has never been investigated in a widely applicable and straightforward comparison. Fully reviewing HE phenomenology is beyond the scope of this article, and a recent ASM Handbook[11] gives a complete assessment of recent theories. However, the following key components of HE are relevant to this work: intergranular (IG) cracking, quasi-cleavage, lower cohesive strength, and slow strain rate.


A. Work Scope

To properly compare HE resistance for austempered vs Q&T processing, the first requirement is that the same steel composition must be subjected to the two heat treatments. The present investigators chose SAE 4340 (UNS G43400) steel, because it is a common medium-carbon steel.

The second requirement is that the steel must also be heat treated to the hardness at which Q&T processing can result in HE susceptibility. The HE of Q&T steel occurs at various hardness levels and, as a result, specifications either cite different hardness levels to avoid or require preventative actions (such as baking heat treatments after pickling or plating). However, most metallurgists and manufacturing procedures state that steels with hardness levels above 40 HRC are susceptible to HE. To follow this guideline and generate useful data for commonly specified 4340 applications, the authors targeted 45 HRC for this study.

The present authors first verified that the properties of the austempered and Q&T samples typified the steel and respective heat treatments. This preliminary procedure had the auxiliary benefit that the present investigators obtained useful microstructural, monotonic mechanical property and fractographic engineering data for design and failure analysis. B. Materials and Their Characterizations

The investigators purchased UNS 43400 steel bar samples that were nominally 25 mm (1 in.) in diameter by 125 mm (5 in.) in length. The bars were obtained from a commercial heat of aircraft quality steel. The bars were supplied with a fine (ASTM No. 7) grain size, and were in a hot-finished and annealed condition (Brinell hardness of 208 HB), in accordance with SAE Aerospace Material Specification (AMS) 6415.

The bars were commercially heat treated by austenitizing at 1158 K (1625 F) and either quenching in oil at 344 K (160 F), followed by tempering at 491 K (425 F) for 90 minutes, or austempering at 585 K (594 F) to a desired Rockwell C hardness of nominally 45 HRC.

The elemental contents of an austempered and a Q&T specimen were determined using glow discharge-optical emission spectrometry (GD- OES),[12] except for carbon, sulfur, and nitrogen, which were determined using combustometric methods.[13]

Sections were cut from a Q&T and an austempered bar. Seven Rockwell hardness measurements were made[14] along a suitably prepared transverse cross section for each bar.

Transverse and longitudinal sections were cut from a Q&T bar and from an austempered bar. The sections were mounted and prepared using standard metallographic techniques for low-alloy steel.[15] Inclusion content cleanliness was determined in the as-polished condition. Optical micrographs were taken in the aspolished condition and after etching in 10 pct sodium metabisulfite. The etched transverse samples were then electropolished [with a 0.15-mm (0.006-in.) sample removal], and the retained austenite content was determined by X-ray diffraction.[16]

Inclusion content cleanliness was determined per ASTM Methods A and D, using automated image analysis.[17] In the ASTM Method A results, the worst observed field of view is located for each of four types of inclusions (in both thin and heavy form). In ASTM Method D, every observed field of view is rated in the same way. However, in Method D, the number of inclusion fields corresponding to each inclusion type is tabulated; this is in lieu of a statement of the worst severity rating, as with Method A.

C. Mechanical Testing

Three tensile specimens were machined from both the austempered and Q&T steels. The tensile specimens had threaded grip ends and gage diameters of 9 mm (0.35 in.) and lengths of 36 mm (1.4 in.). For each heat-treatment condition, duplicate tensile tests were conducted with an extensometer and strain gage at room temperature, to determine 0.2 pct offset yield strength (YS), ultimate tensile strength (UTS), and total elongation (El);[18] elastic modulus (E);[19] and monotonie strength coefficient (K) and strain- hardening exponent (n).[20] One tensile test was conducted at 233 K (-40 F) for each heat-treatment condition, using an extensometer to determine 0.2 pct offset YS, UTS, and El.[18]

For all tests, the samples were tested at strain rates of 0.3 and 10 pct/min in the elastic and plastic ranges, respectively. The percent elongation at fracture was determined using the El measured by the extensometer.

Multiple CVN specimens were machined from both austempered and Q&T steel, and were impact tested[21] at temperatures between 233 and 373 K (-40 and 212 F), to obtain an impact energy, lateral expansion, and fracture appearance (percent shear) transition curve. Duplicate (and a few triplicate) full-sized samples were tested at all temperatures, for each condition. Lateral expansion was determined using the device described in the ASTM standard[21] for that purpose, and fracture appearance was determined by manually measuring with a graduated ruler on the fractured Charpy specimen halves.

In addition, eight subsized CVN specimens 5-mm thick and 10-mm wide and having the V-notch geometry[21] were machined and tested at room temperature for each of the austempered and Q&T conditions. Duplicate samples were tested for both the austempered and Q&T heat treatments in the uncharged condition. Immediately after hydrogen charging, six samples were tested for each of the austempered and Q&T conditions.

Bend specimens were machined from the austempered and Q&T steels. Duplicate uncharged bend samples and duplicate hydrogen-charged bend samples were machined, for each of the austempered and the Q&T steels. The specimens were ground to a maximum surface roughness of 2 [mu]m (roughness average, Ra), with nominal dimensions of 13 mm (0.5 in.) in thickness, 16 mm (0.625 in.) in width, and 119 mm (4.68 in.) in length. The bend specimens were notched by grinding at the longitudinal centerline of the 13-mm (0.5-in.) face, which reduced the width dimension to 14 mm (0.55 in.); the notch had the same geometry and the 2-mm depth as the V-notches in Charpy samples.[21]

Duplicate bend tests were conducted in each of the following four key conditions: austempered vs Q&T and uncharged vs charged with hydrogen. Three-point bend tests were conducted at room temperature using a 100-mm (4-in.) span, and a slow crosshead velocity of 0.025 mm (0.001 in.) per minute on a screw-driven universal testing machine. The bend testing and analytical procedures in the present study were similar to those contained in an ASTM standard,[22] but the standard requires simultaneous hydrogen charging and static loading. However, the present specimens were charged prior to bend testing at a slow strain rate, because insufficient starting sample size and safety equipment made it impossible to test in full compliance with the ASTM standard.[22] The ultimate load, the displacement at ultimate load, the displacement at fracture, and the percent shear were determined quantitatively. The percent shear was determined using manual methods and low-magnification photographs.

To characterize the fracture mode of the Charpy and bend test specimens, their fracture surfaces were documented and examined with an optical stereo microscope and scanning electron microscope (SEM). The SEM was operated with 20-kV accelerating voltage and in secondary electron (SE) mode. For the three-point bend specimens, the pictures were taken at selected magnifications and depths below the notches, to facilitate fractographic comparison between the heat treatments and charging conditions.

D. Hydrogen Charging

To develop a procedure for hydrogen charging steel samples, preliminary experiments were conducted. The objective was to cause hydrogen absorption into the steel and induce embrittlement in the Q&T samples. Proof of embrittlement in the form of IG cracking was required before the authors considered the procedure successful.

Thin sections, approximately 1.6-mm (0.06-in.) thick, were cut from the Q&T steel bars, and a fine square notch was machined into them. The Q&T samples were charged with hydrogen by soaking them for approximately 6 hours in a bath of water containing 18 pet concentrated hydrochloric (HCl) acid. When the sections were removed from the bath, they were covered with black deposit. Within minutes of removal from the bath and after drying with no further preparation, the sections were mounted in a vise and struck with a hammer, to cause fracture. The SEM examination of the thin sections verified the presence of IG fracture. Therefore, after final machining and just prior to testing, selected subsized CVN and bend test specimens were charged similarly.


A. Composition and Microstructure

The results of the chemical analyses of Q&T and austempered samples are shown in Table I. Both samples conformed to the chemical specifications for SAE 4340 and UNS G43406 low-alloy steel contained in SAE AMS-6415. The authors' and the steel supplier's analyses were in excellent agreement.

After correcting for alloying element interference, large grain size, and preferred orientation, retained austenite contents of 0.5 and 0.4 pct were measured for the austempered and Q&T samples, respectively. These values are virtually identical and negligible, so retained austenite influenced no mechanical testing or fractographic results reported in this article. The lack of retained austenite is considered important because some authors[23] have noted that retained austenite content strongly influences the toughness of isothermally transformed bainitic 300 M steel. The composition of 300 M has significantly greater silicon content (1.6 wt pct) than 4340 steel (0.25 pct Si), and this probably results in a vastly different microstructure (higher retained austenite content and different carbides in the 300 M). For the purposes of this article, the entire matter of retained austenite influencing toughness can be ignored, because the 4340 steel of this study contained no retained austenite.

Figures 4 and 5 show optical micrographs of longitudinal views of the as-polished metallographic mounts of the Q&T and austempered samples, respectively. Table II shows the inclusion severity analysis results for ASTM Methods A ("worst fields") and D ("low inclusion content").[17]

This steel had numerous sulfide stringer inclusions and a few globular oxides, fragmented alumina inclusions, and elongated silicate inclusions, which represent only moderate cleanliness. The rather long (maximum severity level 3.0) sulfide inclusions are consistent with the sulfur content (Table I), and the inclusion- forming and some residual elements are present in concentrations that exceed typical commercial steelmaking practices.

Figures 6 and 7 are optical micrographs of the mounted specimens after they were etched in 10 pct sodium metabisulfite. All the micrographs were obtained in the interior of the bar at an original magnification (OM) of 500 times. Etching with sodium metabisulfite helps identify the difference between martensitic and bainitic microstructures, because martensite appears straw brown and bainite appears blue. Figure 6(a) shows that the Q&T samples typify a tempered martensite structure and Figure 6(b) shows that the austempered specimens are mostly bainitic with some martensite. Figure 7(a) shows that the samples (both Q&T and austempered), all of which were made from the same steel bar stock, are heavily banded in the longitudinal orientation. The light etching regions are probably enriched with carbon and manganese, and they contain most of the manganese sulfide inclusions. Figure 7(b), which is the same condition as Figure 6(a) but with a different sample, shows that the bands are actually tubular "packets" with highly alloyed boundaries.

It should be noted that the banding was parallel to the length dimension of all specimens, because this was bar stock. All the mechanical test specimens (tensile, Charpy, and bend) were machined with transverse fracture planes. Therefore, the fracture plane and cracking direction were perpendicular to the banding direction and the crack propagation path was not influenced by the banding.

B. Mechanical Properties

Rockwell C hardness values of 46.7 and 44.5 HRC were obtained as the average of seven measurements for the austempered and Q&T samples respectively, with respective standard deviations of 0.5 and 0.4 HRC. Both samples conformed to the desired hardness of approximately 45 HRC, with the austempered samples having slightly higher hardness. The hardness was relatively constant (small standard deviations in Table III) across the cross section, and mechanical test specimens were only obtained from the central 19 mm (0.75 in.) of the bar.

Table III shows the room- and low-temperature tensile test results for both heat-treatment conditions. Room-temperature stress- strain curves are plotted in Figure 8, for one sample in each heat- treated condition.

Both austempered and Q&T samples had the same tensile YS and modulus at room temperature. However, the austempered samples had significantly greater tensile strength, and somewhat higher ductility and monotonie strain hardening than the Q&T samples. Furthermore, the tensile strengths of the austempered and Q&T samples were virtually identical with the predictions based on hardness,[24] as shown in the footnote of Table III. Based on limited low-temperature tensile test data (one test each at 233 K (- 40 F)), the low-temperature yield and tensile strengths increased slightly for both conditions vs room temperature, with no significant decrease in ductility or change in strain-hardening behavior.

Table IV shows all the full-sized CVN results obtained for both heat-treatment conditions. Figures 9(a) through (c) show the impact energy, fracture appearance and lateral expansion transition data, and polynomial regression curves for both heat-treatment conditions. The impact energy vs temperature results from Figure 1 from the earlier[1] study are superimposed on Figure 9(a), for ease in comparing the energies achieved for the samples from the present study.

Figure 9(a) shows that the austempered samples generally absorbed significantly more energy than did the Q&T samples at all temperatures, due to the higher toughness associated with a lower bainite microstructure.

Figure 9(b) shows somewhat scattered fracture appearance data for the two heat-treatment conditions. Near the lower shelf and in the transition region, the Q&T samples exhibited somewhat larger shear lips than did the austempered samples. This contributed to a slightly lower fracture appearance transition temperature (FATT) result for the Q&T condition, as shown in Table V. Near the USE, the reverse trend was observed, wherein the austempered samples exhibited higher-percent shear values.

The austempered samples exhibited higher USEs than the Q&T samples; however, the lower shelf energies (LSEs) of the austempered samples were only slightly better than those of the Q&T samples. It is also important to note that the USEs and LSEs for both conditions probably would have increased and decreased significantly, if samples had been tested at temperatures below the minimum temperature of 234 K (-38 F) and above the maximum temperature tested in this study of 373 K (212 F). The percent shear data in Table V support this hypothesis, because approximately 15 and 19 pct shear values were obtained for the austempered and Q&T conditions, respectively; also, 25 and 40 percent brittle (flat and opposite to shear) fracture surface areas were obtained for the austempered and Q&T conditions, respectively, after testing at 373 K (212 F).

Although Table V shows that the DBTTs obtained in this study (based on the average energy criterion) were slightly lower for the Q&T condition, this is likely erroneous, because the USE and LSE were never reached with the test temperatures employed in this study. The energy transition curves from Reference 1, reproduced here in Figure 9(a), are probably more accurate, because true asymptotic behavior was found at both energy shelves. Table V also shows that the two conditions had identical DBTTs, when calculated with the data from Reference 1.

Figure 9(c) shows that the CVN austempered samples generally exhibited more lateral expansion than did the Q&T samples, but the data were highly scattered. The higher lateral expansion values obtained for the austempered samples also evince greater ductility, even when the austempered steel fractures in a brittle mode.

C. Fractography of Charpy Impact Samples

Figures 10 through 12 contain SE images obtained during SEM examination of selected full-sized Charpy impact samples fractured at 273 K (32 F), room temperature, and 373 K (212 F), respectively. The (a) and (b) figures show the austempered and Q&T samples, respectively.

At low magnification, the fracture surfaces of all samples (and at all temperatures) had a transition fracture appearance. The outer edges exhibited shear (angular) dimple rupture and the central region exhibited various forms of brittle (flat) fracture.

At 273 K and room temperature, the dominant fracture morphology in both the austempered and Q&T samples was a mixed mode of dimple rupture and partially formed transgranular (TG) cleaves, which is often called quasi-cleavage. However, the Q&T samples exhibited more quasi-cleavage at both those temperatures, and some isolated IG fracture was observed at 273 K (32 F). At 373 K, the dominant fracture morphology in both the austempered and Q&T samples was dimple rupture.

D. Fractography after Preliminary Hydrogen Charging Experiment

Figures 13(a) and (b) are SE images obtained from the thin and notched slivers of the austempered and Q&T samples, respectively, after the samples were charged with hydrogen by immersing the samples in the HCl bath and fracturing with a hammer. The fracture surfaces of the austempered samples exhibited a mixed mode of dimple rupture and quasi-cleavage, which is identical to that discussed with respect to the full-sized (and uncharged) Charpy samples in Figures 10(a) and 11(a). However, the fracture mode of the Q&T samples was mixed dimple rupture and isolated IG fracture. Figure 13(b) shows a localized area of one Q&T sample, in which IG facets were fully formed and not surrounded by dimple rupture.

The appearance of the IG fracture in the Q&T sliver sample validated that the charging procedure could produce HE. This result also gave some preliminary indication that the austempered samples were not embrittled under the same conditions, because the fracture morphology in the austempered slivers did not change from that of the uncharged Charpy samples.

E. Mechanical Testing of Charged vs Uncharged Specimens

Both uncharged and hydrogen-charged subsized austempered Charpy specimens had impact energies that ranged from 16 to 26 J (12 to 20 ft-lb) at room temperature. Both uncharged and hydrogen-charged subsized Q&T Charpy specimens had impact energies that ranged from 14 to 18 J (11 to 13 ft-lb) at room temperature. No significant difference in impact energy was observed between the two heat- treatment conditions or the charging conditions on the subsized samples; therefore, no fractography was performed on any subsized Charpy specimens (either uncharged or hydrogen charged).

Figure 14 shows the load cell force vs crosshead position curves obtained from one-half of the three-point bend tests performed on the Q&T and austempered samples in the uncharged and charged conditions. Maximum peak forces and positions are shown in Table VI, for all the tests.

The duplicate tests for the charged vs uncharged conditions exhibited excellent agreement. In all cases, there was a large (44 pct) drop in peak load (proportional to strength) as the Q&T samples were charged with hydrogen, whereas a much lower (6.1 pct) drop in peak load occurred when the austempered samples were charged with hydrogen. Both conditions exhibited lower drops in position at the peak load (proportional to sample displacement, strain, or ductility) when the samples were charged, but the drop in displacement of the austempered condition was roughly half that of the Q&T condition.

F. Fractography of Bend Samples

Figures 15 through 17 show selected SEM fractographs obtained at 1100 times OM during SEM examination of the three-point bending samples after testing. Figures 15 through 17 correspond to increasing distances from the notch tips. The (a), (b), (c), and (d) images of each figure correspond to the uncharged-austempered, charged-austempered, uncharged-Q&T, and charged-Q&T conditions, respectively.

It is noteworthy that all the slow bend specimens contained dimple rupture, TG cleavage, and quasicleavage, but the present authors only observed IG fracture for one condition. As shown in Figure 15(d), adjacent to (0.2 mm beyond) the notch in a charged Q&T specimen, the fracture mode was primarily IG fracture. This fracture mode was also observed in the companion specimen for this condition. The IG fracture was absent in all of the uncharged bend test specimens and the charged austempered specimens. However, the charged austempered samples exhibited slightly more quasi-cleavage than the uncharged austempered samples at a small depth below the notch, as shown in Figures 15(b) and (a), respectively. Dimple rupture and TG cleavage were observed across the remainder of the fracture surface corresponding to the charged Q&T condition. In the charged austempered samples, the uncharged Q&T samples, and the uncharged austempered samples, the fracture modes had varying mixtures of TG cleavage, dimple rupture, and quasi-cleavage throughout the entire fracture surfaces. The authors observed this same morphology, as depicted in Figure 11, for the full-sized Charpy samples that were impact tested at room temperature. Furthermore, Tables IV and VI show a similar trend in shear fracture in the Charpy and bend samples as a function of heat treatment, i.e., the Q&T samples exhibited more shear lip area than the corresponding austempered samples.


A. Quantitative Comparison of Bainite and Martensite Toughness and Resilience

Although the toughness of a material represents its ability to absorb energy in the plastic range, toughness is a commonly used concept that is otherwise difficult to define.[25] In this study, almost all quantitative measures of toughness show that lower bainite is tougher than tempered martensite. (The DBTT values in Table V may represent an exception to this trend, but it is unlikely, because the LSEs and USEs were never attained in the temperature range tested in this study, and Table V shows an identical DBTT for the austempered and Q&T steels from Figure 1.)

In this study, the austempered samples exhibited greater quantitative toughness and ductility than their Q&T counterparts, when they were compared at constant YS and even slightly higher hardness and tensile strength. The Charpy impact energy, tensile elongation, tensile reduction in area (RA), peak bending load, and bending displacements were all higher for the austempered samples than for the Q&T samples.

One measure of toughness is the area (Vj) underneath the (true) stress-strain curve, which corresponds to the ability to absorb energy in the plastic range or the work per unit volume that the material can sustain prior to rupture.[25] This important parameter illustrates that toughness is composed of both strength and ductility, and recently it has been proposed[26] for use in ranking materials based on toughness, with and without other parameters such as Charpy and plane strain fracture toughness properties.

The materials in this study necked significantly. This necking prevented the authors from determining the area under the tensile stress-strain curve, because the equations converting engineering stress-strain data to true stress-strain data become invalid once constancy of volume becomes violated at the maximum load at which necking occurs. In addition, the materials exhibited too much plasticity and gave unrealistic values when the authors attempted to correct the bending force-displacement data to axial stress-strain data using the elastic equations in an ASTM standard27 after subtracting the machine compliance with data from bend testing a large and stiff sample.

To approximate the area under the engineering stressstrain curve, the authors used two different relationships,[25,26] and Table VII shows that the austempered steel had significantly higher values using either formula. This means that the austempered steel has greater ability to withstand occasional stresses above the yield stress without fracturing, such as in freight-car couplings, gears, chains, and crane hooks25

Table VII also shows that the austempered and Q&T heat treatments of 4340 steel in this study effected identical values for the modulus of resilience (U^sub R^), which is the strain energy per unit volume and the ability to absorb energy when deformed elastically from zero stress to the yield stress and to return the energy when unloaded.[25] The two heat treatments produced identical U^sub R^ values, because they were compared at an identical YS and modulus (Table III), which are used to calculate U^sub R^. From a practical standpoint, the U^sub R^ value represents a material's resistance to energy loads in applications in which the material must not undergo permanent distortion due to deformation (such as a spring).

The superior toughness and ductility of the austempered samples with lower bainite than martensite in the Q&T samples was also found in earlier work.[1] However, it is noteworthy that this is only observed when lower bainite is formed by isothermal transformation.[8] Mixed microstructures containing upper bainite formed by continuous cooling and slack quenching have lower toughness than martensitic micro structures.[7,8,23]

B. Structure-Property Relationships

Earlier authors have noted similarities between martensite and lower bainite in fracture and strengthening behavior. Offset YS and DBTT obey Hall-Petch type relationships with the size (diameter) of well-defined packets in both microstrucures.[8,23] The boundaries of the packets have been compared to grain boundaries strengthening in ferrite.[23]

Figure 8 shows that the Q&T samples exhibited slightly discontinuous yielding, whereas the austempered samples exhibited completely continuous yielding behavior. Therefore, a different or an additional yielding mechanism may be operative in the austempered steels. Although discontinuous yielding is often associated with low- carbon sheet steels, it is now associated with solute-dislocation interaction or with precipitation along dislocations in multiple other systems in which Hall-Petch boundary strengthening is absent.12 In standard low-carbon steels, yield point elongation (YPE), representing barrier-free dislocation motion, is observed between discrete upper and lower yield points. However, discontinuous yielding without discrete yield points and YPE (like the Q&T condition evaluated here) is often observed in other materials such as normalized and aged steels; in these materials, the work-hardening theories are less widely understood. Discontinuous yielding in the latter materials must be associated with dislocation interactions with barriers vs the dislocation multiplication rate. The last stage of tempering in the Q&T heat treatment causes cementite and ferrite formation, which is the same micro structure as low-carbon sheet steels that yield discontinuously. However, there is more cementite formation in Q&T 4340, and that may be why YPE is absent in the Q&T steels. One author has recently speculated[10] that austempering forms lower bainite with a high concentration of carbon remaining trapped at defects within the bainitic ferrite with no precipitates forming. This would explain why continuous yielding was obtained in the present work with austempering, i.e., no initial particle barrier to dislocation motion is available with a lower bainite microstructure.

C. Fracture Appearance Considerations

The quasi-cleavage fracture morphology observed for the uncharged steels is similar to that observed by earlier investigators.8,23 Quasi-cleavage exhibits both cleavage and plastic deformation, but it is really just a form of cleavage.[15] The prefix quasi does not accurately describe the fracture, because it implies that the fracture resembles but is not really cleavage.[28]

The results in the last columns of Tables V and VI show that martensite has the qualitative appearance of greater toughness as compared to lower bainite, even though all quantitative measures show that the bainite is tougher than the martensite. Note that room temperature is in the mixed ductile-brittle region in the transition curve shown in Figure 9(b); therefore, both constraint and strain rate differences are hidden variables that can play greater roles in fracture appearance differences vs inherent toughness differences.

The low-magnification fracture appearance of the austempered bend test samples had less shear lip area as compared to Q&T bend samples, regardless of charging conditions. In addition, the uncharged Q&T samples exhibited dimple rupture at the notch root and in the last material to fracture. These fracture appearance differences probably evince less constraint (plane strain) and more plane stress in the Q&T samples, as the greater curvatures of the bend testing curves in Figure 14 suggest for the Q&T samples vs the austempered samples.

In the same austempered and uncharged bend test specimen, the microscopic fracture mode changed from full dimple rupture to greater portions of TG cleavage and quasi-cleavage as the crack depth increased from the notch tip (compare Figures 15(a), 16(a), and 17(a)). Furthermore, on several bend specimens, the authors noted two or three different zones of macroscopic fracture morphologies, in addition to the shear lip.

Since the bend samples were tested at constant crosshead speed, the greater portions of TG cleavage in the remaining material to fracture resulted from the increased strain rate that occurred at greater crack depths.

D. HE Considerations

Although other fracture modes can represent embrittlement, only IG fracture is confined to some type of embrittlement (in steels without case layers). In this study, the IG fracture found in the hydrogen-charged-Q&T samples clearly represents HE, because IG fracture was not found in any of the other conditions, i.e., uncharged-austempered, charged-austempered, and uncharged-Q&T conditions. As reported by McMahon,[29] IG fracture is usually obtained when the stress-controlled decohesion mechanism of HE is operative. Stress-controlled decohesion is initiated by blocked plastic flow on the microscale; this mechanism is of particular concern, because it can lead to failure at very low stress intensities.29 Many of the HE observations from this study were consistent with previous work. Small fractions of IG fracture were only found adjacent to the notch in the slow bend samples of this study, and this is as expected.11'25'29"32 Slow bend tests and notched specimens are usually required to observe HE in the transition region,32,33 as was found in the present study. Lastly, HE is usually only found at low strain rates and not in Charpy samples.32,33 Although the subsized Charpy samples from this study showed no impact energy reduction with hydrogen charging, the Q&T sliver samples (that were tested at an unknown strain rate) did exhibit IG fracture, whereas the austempered sliver samples did not.

The austempered samples did show some evidence of embrittlement, but the charging effect was minor. The charged austempered samples exhibited a greater fraction of quasi-cleavage near the notch tip, and the peak loads decreased somewhat. Other authors2 have noted quasi-cleavage in austempered samples in both the presence and absence of hydrogen.

Although the present study is a research study, mechanistic determinations are beyond the scope of this work. Based on the literature search, other authors[34] have developed models for the HE of martensitic microstructures. The present work does have some significant practical considerations. The hydrogen charging method of dipping samples in concentrated acid-water solutions is employed in industrial practice to remove mill oxide and prepare surfaces for plating. The load drop method is employed in a variety of HE test methods,35 especially Reference 22.

E. Commercial Comparisons of Austempering vs Q&T Processing

When a Q&T heat treatment is used, distortion issues and quench cracking are more probable than they are in austempering. When bainite is formed isothermally by austempering, the microstructure develops over many minutes or hours at a quench temperature above the M^sub s^, thus avoiding nonuniform transformation and resulting in the absence of quench cracks. This is especially true for high- alloy steels that have low M^sub s^ temperatures and correspondingly large austenite-to-martensite volume expansions.

The Q&T processing cost is less expensive than the austemper processing cost, but austempering is less expensive when the total cost of heat treatment is combined with the cost of distortion and cracking. The cost of cracking or distortion may include increased inspection, sorting scrap components, straightening, or increased stock removal during machining.

Since both Q&T and austempering processes involve precision heat treatments, they must be performed after welding; therefore, neither is more advantageous from that standpoint. However, after a part is welded, austempering might prove more advantageous than Q&T processing, because austempering promotes less part distortion, more fracture toughness, and less susceptibility to HE and liquid metal embrittlement after light metal plating, because of the absence of quench cracks.

F. Suggestions for Further Work

Many peripheral questions arose from the results of this study. It seems worthwhile to conduct the following testing.

(1) Low- and high-cycle fatigue comparisons of austempered and Q&T steels, to determine whether fatigue crack propagation and initiation resistance differ under strain and stress-controlled cycling, respectively.

(2) Charpy and bend sample data cannot be employed for design. Plane strain fracture toughness (K^sub Ic^) testing should be performed on larger samples with greater opportunities for constraint. The tests should be conducted with and without hydrogen charging for both austempered and Q&T steels, to obtain design quality data.

(3) Some minimal testing should be conducted under strain-rate control on larger samples, to try to separate the inherent toughness effects from the effects of constraint and strain rate.


The following mechanical properties and fracture characteristics were determined for Q&T 4340 steel and compared to those of austempered 4340 steel.

1. The microstructure of the Q&T 4340 steel was fully martensitic, with a hardness of 44.5 HRC. The microstructure of the austempered steel was predominately lower bainite with some martensite. The austempered steel was slightly higher in hardness (46.7 HRC), which resulted in a somewhat higher tensile strength with the same YS as the Q&T condition.

2. Although the austempered steel had slightly higher tensile strength, it had similar modulus, YS, and strain-hardening constants. Even with its higher tensile strength, the austempered steel had significantly higher tensile ductility at both room and low temperatures.

3. The Q&T and austempered steels exhibited discontinuous and continuous yielding, respectively.

4. As found in previous studies, the austempered steel had significantly greater impact energy and lateral expansion in CVN testing over a wide range of temperatures. However, the FATT for the Q&T steel was slightly lower than that of the austempered steel.

5. At most test temperatures, both the austempered and Q&T steels exhibited similarly mixed fracture morphologies of quasi-cleavage and dimple rupture, although the Q&T steel exhibited less dimple rupture.

6. As determined by other investigators, this study found that impact test specimens of both steels were not affected by hydrogen exposure, and a slow strain rate was required for HE to occur. Slow bend testing of notched specimens produced significantly lower strength and ductility in the Q&T steel after hydrogen exposure. The austempered steel exhibited only a small decrease in toughness with hydrogen exposure. The austempered steel had significantly better resistance to HE than did the Q&T steel.

7. Similar fractographic results were obtained after hydrogen charging and testing slivers and three-point bend specimens. At the base of notch in the hydrogen-charged Q&T specimens, the primary fracture morphology was IG fracture. Dimple rupture was the fracture mode on the remaining surface of these Q&T samples. Mixed fracture morphologies of dimple rupture, quasi-cleavage, and TG cleavage were obtained for the other three samples types: the uncharged Q&T specimens and both the hydrogen-charged and uncharged austempered specimens.


The authors thank all the test technicians at Stork Climax Research Services (CRS) and Terry Lusk at Applied Process, Inc., for meticulously conducting the experiments. The authors also acknowledge the guidance that Stork CRS metallurgical engineers Rick Gundlach and Art Griebel provided with the hydrogen charging and bend testing procedures, respectively.


1. ASM Handbook, 10th ed., vol. I, Properties and Selection: Irons, Steels, and High-Performance Alloys. Notch Toughness of Steels, ASM INTERNATIONAL, Materials Park, OH, 1990, p. 748, which cites the original article: R.F. Hehemann, VJ. Luhan, and A.R. Troiano: Trans. ASM, 1957, vol. 49, pp. 409-26.

2. J.H. Chuang, L.W. Tsay, and C. Chen: Int. J. Fatigue, 1998, vol. 20(7), pp. 531-36.

3. L.F. Reynolds and M.P. Hayes: Trans. lnst. Met. Finish., 1987, vol. 65 (2), pp. 50-57.

4. P.Sh. Lantsman, G.G. Vernovskaya, and V.N. Kudryavtsev: translated from Metallovedenie I Termiceskaya Obrabolka Metallov., 1973, No. 5, pp. 73-74.

5. L.J. Habraken and M. Econompoulos: Transformation and Hardenability in Steels, Climax Molybdenum Company of Michigan and University of Michigan Symp., Feb. 27-28, Climax Molybdenum Company of Michigan, Ann Arbor, MI, 1967, pp. 69-107.

6. F.B. Pickering: Transformation and Hardenability in Steels, Climax Molybdenum Company of Michigan and University of Michigan Symp., Feb. 27-28, Climax Molybdenum Company of Michigan, Ann Arbor, MI, 1967, pp. 109-29.

7. ASM Handbook. 10th ed., vol. 1, Properties and Selection: Irons, Steels, and High-Performance Alloys: Notch Toughness of Steels, ASM INTERNATIONAL, Materials Park, OH, 1990, p. 747.

8. D.R. Johnson and WT. Becker: J. Mater. Eng. Perf., 1993, vol. 2 (2), pp. 255-62.

9. ASM Handbook, 10th ed., vol. 4, Heat Treatment: Austempering of Steel, ASM INTERNATIONAL, Materials Park, OH, 1991, pp. 152-63.

10. H.K.D.H. Bhadeshia: Solid [arrow right] Solid Phase Transformations in Inorganic Materials, Vol. 1, TMS, Warrendale, PA, 2005, pp. 469-84.

11. ASM Handbook, vol. 13A, Corrosion: Fundamentals, Testing and Protection, ASM INTERNATIONAL, Materials Park, OH, 2003, pp. 367- 74.

12. ISO 14707:2000E First Edition (2000-18-15): Surface Chemical Analysis-Glow Discharge Optical Emission Spectrometry (GD-OES)- Introduction To Use, ISO 2000, ISO Copyright Office, Geneva, Switzerland, www.iso.ch.

13. ASTM E1019, 2003: Standard Test Methods for Determination of Carbon, Sulfur. Nitrogen, and Oxygen in Steel and in Iron, Nickel, and Cobalt Alloys, ASTM International, West Conshohocken, PA, www.astm.org.

14. ASTM E18, 2005e 1 : Standard Test Methods for Rockwell Hardness and Rockwell Superficial Hardness of Metallic Materials, ASTM International, West Conshohocken, PA, www.astm.org.

15. ASTM E3, 2001: Standard Practice for Preparation of Metallographic Specimens, ASTM International, West Conshohocken, PA, www.astm.org.

16. ASTM E975, 2003: Standard Practice for X-Ray Determination of Retained Austenite in Steel with Near Random Crystallographic Orientation, ASTM International, West Conshohocken, PA, www.astm.org.

17. ASTM E45, 2005el: Standard Test Methods for Determining the Inclusion Content of Steel, ASTM International, West Conshohocken, PA, www.astm.org.

18. ASTM ES, 2004: Standard Test Methods for Tension Testing of Metallic Materials, ASTM International, West Conshohocken, PA, www.asttn.org.

19. ASTM E111, 2004: Standard Test Method for Young's Modulus, Tangent Modulus, and Chord Modulus, ASTM International, West Conshohocken, PA, www.astm.org.

20. ASTM E646, 2000: Standard Test Method for Tensile StrainHardening Exponents (c-Values) of Metallic Sheet Materials, ASTM International, West Conshohocken, PA, www.astm.org. 21. ASTM E23, 2006: Standard Test Methods for Notched Bar Impact Testing of Metallic Materials, ASTM International, West Conshohocken, PA, www.astm.org.

22. ASTM G129, 2000 (2006): Standard Practice for Slow Strain Rate Testing to Evaluate the Susceptibility of Metallic Materials to Environmentally Assisted Cracking, ASTM International, West Conshohocken, PA, www.astm.org.

23. Y. Tomita and K. Okabayashi: Metall. Trans. A, 1986, vol. 17A, pp. 1203-09.

24. ASTM A370, 2006: Standard Test Methods and Definitions for Mechanical Testing of Steel Products, ASTM International, West Conshohocken, PA, www.astm.org.

25. G.E. Dieter: Mechanical Metallurgy, 3rd ed., McGraw-Hill Book Company, New York, NY, 1986, pp. 197-201 , 282-83, and 490-93.

26. P.M. Novotny: Adv. Mater. Processes, 2007, vol. 165 (5), pp. 29-31.

27. ASTM E855, 1990: Standard Test Methods for Bend Testing of Metallic Flat Materials for Spring Applications Involving Static Loading, ASTM International, West Conshohocken, PA, 2000, www.astm.org.

28. ASM Handbook: Fractography, 9th ed., ASM INTERNATIONAL, Metals Park, OH, 1987, vol. 12, pp. 12-31.

29. CJ. McMahon Jr.: Eng. Fract. Mech., 2001, vol. 68, pp. 773- 88.

30. A.R. Troiano: Trans, ASM, 1960, vol. 52, pp. 54-80.

31. G. Krauss: Steels: Processing, Structure, and Performance, ASM INTERNATIONAL, Materials Park, OH, 2005, pp. 406-10.

32. ASM Handbook: Failure Analysis and Prevention, vol. 11, ASM INTERNATIONAL, Metals Park, OH, 2002, pp. 99-101.

33. M. Wang, E. Akiyama, and K. Tsuzaki: Scripta Mater., 2005, vol. 53, pp. 713-18.

34. W.W. Gerberich, T. Livne, X.-F. Chen, and M. Kaczorowski: Metall. Trans. A, 1988, vol. 19A, pp. 1319-34.

35. Mechanical Hydrogen Embrittlement Methods for Evaluation and Control of Fasteners, ASTM, West Conshoshocken, PA, 2001, pp. 1-25.

JOHN M. TARTAGLIA, Engineering Manager and Senior Metallurgical Engineer, and KRISTEN A. LAZZARI and GRACE P. HUI, Engineering Associates, are with Stork Climax Research Services, Wixom, MI 48393. Contact e-mail: [email protected] KATHY L. HAYRYNEN, Technical Director, is with the Technologies Division, Applied Process, Inc., Livonia 48150, MI.

Manuscript submitted January 23, 2007.

Article published online February 1, 2008

Copyright Minerals, Metals & Materials Society Mar 2008

(c) 2008 Metallurgical and Materials Transactions; A; Physical Metallurgy and Materials Science. Provided by ProQuest Information and Learning. All rights Reserved.