High dislocation density–induced large ductility in deformed and partitioned steels

See allHide authors and affiliations

Science  08 Sep 2017:
Vol. 357, Issue 6355, pp. 1029-1032
DOI: 10.1126/science.aan0177

A ductile steel shows its strength

Many industrial applications require materials to have high strength while remaining pliable, or ductile. However, the microstructure that increases strength tends to reduce ductility. He et al. used a processing mechanism to create a “forest” of line defects in manganese steel. This deformed and partitioned steel was produced by cold-rolling and low-temperature annealing and contained a dislocation network that improved both strength and ductility.

Science, this issue p. 1029


A wide variety of industrial applications require materials with high strength and ductility. Unfortunately, the strategies for increasing material strength, such as processing to create line defects (dislocations), tend to decrease ductility. We developed a strategy to circumvent this in inexpensive, medium manganese steel. Cold rolling followed by low-temperature tempering developed steel with metastable austenite grains embedded in a highly dislocated martensite matrix. This deformed and partitioned (D and P) process produced dislocation hardening but retained high ductility, both through the glide of intensive mobile dislocations and by allowing us to control martensitic transformation. The D and P strategy should apply to any other alloy with deformation-induced martensitic transformation and provides a pathway for the development of high-strength, high-ductility materials.

Strength and ductility are key mechanical properties of metallic materials for developing energy-efficient and lightweight structural components in a wide variety of industries, including automotive and aerospace. Unfortunately, improving strength often results in the degradation of ductility, which is known as the strength-ductility trade-off (1). Cost is also an issue, as alloying elements that help improve both properties, such as cobalt and titanium, tend to be expensive. Previous efforts toward resolving this trade-off focused on engineering defects such as grain boundaries (2, 3) and coherent twin boundaries (4, 5). However, the strength may reach a limit when grain size or twin-boundary spacing is reduced to nanometers (68). Dislocations or line defects are a different pathway for engineering alloy properties. In contrast to the above planar defects, the strength of metallic materials monotonically increases with dislocation density according to the well-known Taylor hardening law (9). The problem is that ductility tends to decrease as the number of dislocations increases. However, improved choices in alloy composition and clever processing strategies may allow dislocation engineering to circumvent the strength-ductility trade-off. In particular, achieving high ductility in the presence of high dislocation density in an inexpensive material produced by facile processing routes is desirable for broad industrial applications at an economic cost.

Severe plastic deformation (SPD) and phase transformation both generate high dislocation density in metals. Cold drawing is one SPD that introduces a very high dislocation density in pearlitic steel wires (10). Cold drawing and other SPD methods also may lead to grain refinement (<100 nm) in metals because of the formation of high-angle grain boundaries by rearrangement of intensive dislocations (10, 11). By contrast, the first-order solid-state martensitic transformation produces a highly dislocated martensite microstructure without a lot of grain refinement (~500 nm) in steels (12). Either quenching or deformation can trigger martensitic transformation, depending on the alloying content of steels (13, 14). We show that rolling and low-temperature tempering produced a high dislocation density in steel, also enabling a large ductility. In addition to the high dislocation density, interstitial C atoms partition into and help stabilize the austenite phase during tempering (13). We define this thermomechanical treatment as the “deformed and partitioned (D and P)” process and the corresponding steel as D and P steel. The D and P process could be realized with conventional processing techniques that are compatible with existing industrial production lines.

We used a medium Mn steel (10% Mn, 0.47% C, 2% Al, and 0.7% V, by weight) for the D and P process. The Mn and C atoms are effective austenite stabilizers (fig. S1). The addition of 2% Al content is to suppress cementite precipitation during the tempering process. The addition of 0.7% V is to form intensive nanometer-sized V carbides, which provided enhanced resistance to delayed fracture induced by hydrogen embrittlement (15, 16). The D and P steel that we produced by multiple deformation and annealing steps (17) possesses a heterogeneous lamella dual-phase microstructure in which metastable austenite is embedded in a martensite matrix (Fig. 1A). We obtained the martensite matrix by cold rolling (fig. S2) followed by tempering. The tempered martensite matrix has heterogeneous grain morphologies and substructures. Large lenticular martensite grains, which constitute most of the matrix, are mainly decorated with dislocation cells (Fig. 1B). Nevertheless, some large lenticular martensite grains also possess dislocations as well as twins (Fig. 1C) and some small lath martensite grains with dislocations (Fig. 1D) in the martensite matrix. We estimated the average dislocation density of martensite matrix (~1.27 × 1016 m–2) with synchrotron x-ray diffraction (XRD) (fig. S3). Such a high dislocation density is due to the combination of plastic deformation and displacive shear transformation. The martensite we created differs from the thermally formed martensite obtained by quenching in that it inherits dislocations generated from warm deformation. The early transformed martensite is also subjected to cold rolling, which we confirmed by the formation of dislocation cells in martensite (Fig. 1B) (18). Consequently, the dislocation density in our martensite matrix is much larger than that in thermally formed martensite (~2.36 × 1014 m–2) (fig. S3). The austenite phase also has a heterogeneous microstructure, including coarse lamella grains (Fig. 1A), ultrafine lamella grains (Fig. 1D), and submicron granular grains (Fig. 1E). The large austenite grains are decorated with a high number of dislocations as well as stacking faults (Fig. 1F), whereas the small austenite grains mainly possess dislocations (Fig. 1E). Both martensite and austenite phases contain nanometer-sized vanadium carbides (Fig. 1, G and H) and solid solutes (C, Mn, and V atoms). Although some layered austenite grains have higher Mn content (fig. S4) from the banding of Mn in hot-rolled strip (fig. S1), we also observed austenite grains with low Mn content and martensite with high Mn content (fig. S4). This suggests that the Mn banding does not determine retention of the layered austenite grains in the D and P steel. Instead, it could be due to the mechanical stabilization induced by intensive defects in these layered austenite grains (Fig. 1F) (19).

Fig. 1 Microstructure of our D and P steel before the tensile test.

(A) Electron backscatter diffraction phase image and orientation image showing the lamella microstructure of austenite (γ) embedded in tempered martensite (α′) matrix. The austenite area fraction is 15%, and the martensite area fraction is 85%. RD indicates the rolling direction, ND indicates the normal direction, and TD indicates the transverse direction. (B) Dislocation structures in martensite. The inset is a selected-area diffraction pattern (SADP). (C) A typical lenticular martensite distributed with twins and dislocations. (D) The lath martensite and lamella austenite as observed in transmission electron microscopy (TEM) bright-field and dark-field images. The right image is a SADP showing a Kurdjumov-Sachs (K-S) relation between lamella austenite and lath martensite. (E) Dislocations in submicron granular austenite. (F) Dislocations and stacking faults in a large austenite grain captured by TEM bright-field and high-resolution images. (G) Distribution of nanosized vanadium carbide (V4C3) in tempered martensite matrix. (H) High-resolution TEM image of nanosized vanadium carbide in austenite.

The as-developed D and P steels possessed ultrahigh yield strengths of 2.21 and 2.05 GPa (Fig. 2). The high yield strength results from the high dislocation density (ρt) in the martensite matrix. For instance, the yield stress (σ) that results from dislocation forest strengthening is estimated to be 1.6 GPa according to the Taylor hardening law (9) Embedded Image where Taylor factor M is 2.9 (20), empirical constant α is 0.23 for cell-forming dislocation distribution (21), shear modulus μ is 85 GPa (22), and Burgers vector b is 0.25 nm (20). Therefore, the yield stress induced by the high dislocation density is on the same order of magnitude as the macroscopic yield strength (Fig. 2). The collective contributions from the other strengthening defects, such as solid solutes, nanoprecipitates, stacking faults, twin boundaries, and ultrafine grains (Fig. 1), elevated the yield strength beyond 2 GPa. Consequently, the D and P steel has a higher yield strength than steel with a bimodal coarse-grained microstructure (Fig. 2 and fig. S5). Intriguingly, the D and P steel also had a larger tensile uniform elongation (>15%) than its bimodal coarse-grained counterpart. Our D and P steel with high dislocation density has an inverse strength-ductility trade-off similar to what has been found in other materials (23).

Fig. 2 Engineering stress-strain curves of our steels at room temperature.

Curves a and b represent the D and P steels, processed by warm rolling, intercritical annealing, and cold rolling followed by tempering at 400oC for 6 and 15 min, respectively. Curve c represents the deformed steel, processed by warm rolling, intercritical annealing, and cold rolling without tempering at 400oC. Curve d represents the sample with bimodal coarse-grained microstructure (fig. S5) obtained by annealing at 900oC for 30 min. The uniform elongation is almost the same as the total elongation for all tensile curves. An increased yield strength of 180 MPa in curve a as compared to that in curve c is due to the segregation of C atoms at dislocations during the tempering at 400oC (24). The serrations at elongations between 10 and 14% in curves a and b manifest the Portevin-Le Châtelier effect (35).

The high dislocation density in our D and P steel results in an ultrahigh yield strength. The large uniform elongation is also due to the dislocation density. The D and P steel demonstrates discontinuous yielding followed by large Lüders strains (ε) (Fig. 2). The enhanced yield drop after the prolonged annealing is a typical feature of strain aging during which both mobile and immobile dislocations are segregated by solid solutes (24). Martensite has an inherently high ratio (~18%) of mobile screw dislocations from displacive shear transformations (25, 26). The mobile and immobile dislocations in martensite could rearrange to form dislocation cells (Fig. 1B) during cold rolling, while the characteristics of dislocations remain the same (18, 27). The dislocation cells consist of a cell interior with a low dislocation density and a cell wall with a high dislocation density. The catastrophic release and glide of the mobile screw dislocations from the cell wall lead to the collapse and break of original cell boundaries and the formation of elongated dislocation cell structure after Lüders strain, which we confirmed by comparing Fig. 1B and Fig. 3A. The elongated cell has an increased length of ~400 nm compared to that of the original cell structure, which indicates that the average glide distance of the mobile screw dislocations is ~400 nm. A general relationship between plastic strain and mobile dislocation density (ρm) can be written as (28) Embedded Image where L is the average glide distance of the mobile dislocations (~400 nm). We assumed the ratio of mobile dislocation density to total dislocation density to be 18% for deformation-induced martensite (26), given a mobile dislocation density of 2.3 × 1015 m–2 in the martensite matrix. Consequently, the plastic strain induced by the glide of these mobile dislocations in the martensite matrix is estimated to be 6.8%, which constitutes a large portion of measured Lüders strain (~7 to 9%) in the D and P steel (Fig. 2). The high mobile dislocation density in the martensite matrix accommodates a large plasticity upon yielding. The catastrophic release and glide of the mobile screw dislocations result in a low, but positive, strain-hardening rate during Lüders strain regime (fig. S6). Such a low but positive strain-hardening rate has been reported in irradiated metals (29) and ω-containing Ti alloys (30), in which the operation of dislocation channeling is the governing deformation mechanism.

Fig. 3 Microstructure of our D and P steel after the tensile test.

(A) The elongated dislocation cell structure in the tempered martensite matrix after tensile straining to 8%. The inset is the corresponding SADP. Arrows mark the break of cell boundaries. (B) The XRD profiles at different strains. The XRD measurement on the specimen of 5.9% strain is on the gauge part swept by Lüders band. Compared to the (220)γ peak, the substantially enhanced (211)α′ peak intensity at strains beyond 5.9% suggests that martensitic transformation is active in the large-strain regime. a.u. represents arbitrary units. (C) The formation of lenticular martensite in the coarse austenite grains after tensile straining to fracture. (D) The generation of deformation twins in the submicron austenite grains after tensile straining to fracture. The right image is the SADP showing a K-S relation between twinned lamella austenite and lath martensite.

In addition to the high mobile dislocation density, the D and P steel had a continuous transformation-induced plasticity (TRIP) effect at a large strain during the tensile test (Fig. 3B). The TRIP effect resulted from the formation of martensite in the coarse layered austenite grains. This TRIP effect applied compressive residual stress to effectively blunt localized plasticity during tensile straining (fig. S7) (31) and also provided dynamic strain partitioning between phases and improved strain hardening (fig. S6) (32). The formation of the large martensite grains suggests that coarse layered austenite grains are less stable than ultrafine austenite grains during uniaxial tensile deformation (Fig. 3C). Nevertheless, most of these coarse layered austenite grains only transform to martensite at strains beyond the Lüders strain (Fig. 3B), explaining their high mechanical stability in the D and P steel. We ascribed the enhancement in the mechanical stability to the high dislocation density in large austenite grains (Fig. 1F), where dislocations can act as barriers for glissile martensite interface and therefore stabilize austenite grains (19). Moreover, the hard martensite matrix (fig. S8) can shield austenite from deformation (33), allowing the austenite to transform at a large strain regime. The austenite grains in the D and P steel were further stabilized by C partitioning from martensite (fig. S9) (13) relative to the austenite grains in deformed steel (fig. S10). In addition to the effect of C partitioning and relatively higher Mn content (fig. S4), the high dislocation density in the D and P steel also controlled release of the TRIP effect, improving ductility.

The D and P steel also had a twinning-induced plasticity (TWIP) effect during the tensile test. We mostly observed the TWIP effect, induced by the formation of deformation twins, in ultrafine austenite grains (Fig. 3D). The initiation of nanotwins from phase boundary suggests a high stress level experienced by lamella austenite grain (Fig. 3D). Therefore, the TWIP effect also operated in the large strain regime (fig. S6). The minor austenite volume fraction (Fig. 1A) of the small grains means that we expect the TWIP effect to be much less important than the TRIP effect, even as deformation twins can accumulate deformations and improve ductility (4).

We compared the bulk properties of D and P steel to other high-strength metallic materials (Fig. 4). The D and P steel exhibits a yield strength that is 50% higher than that of nanobainite steel while maintaining a comparable uniform elongation (Fig. 4). Moreover, it exhibits a uniform elongation that is one order of magnitude larger than that of commercial maraging steels while maintaining an equivalent yield strength. Consequently, our D and P steel achieves excellent tensile properties and defines a new space in the strength-ductility map (Fig. 4). Despite its discontinuous yielding and Lüders strain (Fig. 2), the ultrahigh yield strength of our D and P steel makes it a desirable alloy for applications where yield strength is the main design criterion.

Fig. 4 Tensile properties of our steel compared with those of other existing high-strength metallic materials.

These include nanobainite steel (36), maraging steel (37, 38), dual-phase steel (39), martensitic steel (40), TRIP steel (41, 42), quenching and partitioning (Q and P) steel (43), nanotwinned steel (44, 45), maraging-TRIP steel (46), high–specific strength steel (47), high-entropy alloys (48, 49), nanostructured lamella Ti (3), Ti6Al4V (47), tungsten (50), Inconel 718 (51), nanostructured Mo alloy (52), and nanostructured Co alloy (53). Our D and P steel is clearly separated from the general trend and therefore defines a new space in the strength-ductility map. Note that the uniform elongation is selected here not only because it is a desirable property but also because it is less affected by the sample dimensions used in different studies (2).

The D and P overcomes the challenge of creating martensite, which is known to be an issue for compositionally similar steels that rely on a quenching and partitioning process (13) (fig. S5). The high dislocation density in the D and P steel not only increases the yield strength by dislocation forest hardening but also enables a large ductility by the glide of existing mobile dislocations and by the controlled release of the TRIP effect. The high dislocation density is the origin of the inverse strength-ductility trade-off. We expect that this strategy will be useful in other systems with similar deformation-induced martensitic transformation mechanisms such as titanium alloys (34). The D and P steel exhibits a low raw-materials cost as compared to the maraging steel while maintaining a comparable ultimate tensile strength (fig. S11). Therefore, by engineering dislocations, we simultaneously alleviate the economic concerns while achieving ultrahigh strength.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S14

References (5458)

References and Notes

  1. Materials and methods are available as supplementary materials.
  2. Acknowledgments: The authors thank K. Lu for his insightful and constructive comments on this paper. M.X.H. is grateful for financial support from the Research Grants Council of Hong Kong (grant nos. 712713, 17203014, and 17255016) and the National Natural Science Foundation of China (grant no. U1560204). H.W.L. is grateful for financial support from the National Natural Science Foundation of China (grant no. U1460203). H.W.Y. acknowledges the Ministry of Science and Technology of the Republic of China for providing financial support under contract MOST-104-2218-E-002-022-MY3 and thanks the Instrumentation Center at National Taiwan University for technical support during use of the JSM 7800F PRIME high-resolution scanning electron microscope. The authors also acknowledge the Shanghai Synchrotron Radiation Facility for providing the synchrotron XRD facility at beamline no. 14 B. The present work has pending patents with application numbers 201610455155.3 and PCT/CN2016/096509 and another awarded patent, number 201410669029.9. M.X.H. supervised the study. M.X.H., B.B.H., and H.W.L. designed the study. B.B.H. and B.H. prepared the thermomechanical treatment and the mechanical tests. H.W.Y., G.J.C., and B.B.H. conducted the microstructure characterization. B.B.H., M.X.H., H.W.L., and Z.K.W. analyzed the data. B.B.H. and M.X.H. wrote the paper. The authors declare no competing financial interests. Data are available in the manuscript and supplementary materials.
View Abstract


Navigate This Article