Journal of Materials Science & Technology  2020 , 44 (0): 148-159 https://doi.org/10.1016/j.jmst.2020.01.027

Research Article

On the microstructural evolution pattern toward nano-scale of an AISI 304 stainless steel during high strain rate surface deformation

Hongwang Zhangab*, Yiming Zhaoab, Yuhui Wangab, Chunling Zhangb, Yan Penga

a National Engineering Research Center for Equipment and Technology of Cold Strip Rolling, College of Mechanical Engineering, Yanshan University, Qinhuangdao 066004, China
b National Key Laboratory of Metastable Materials Science and Technology, Yanshan University, Qinhuangdao 066004, China

Corresponding authors:   * National Engineering Research Center for Equipment and Technology of Cold Strip Rolling, College of Mechanical Engineering, Yanshan University, Qinhuangdao 066004, PR China. E-mail address: hwzhang@ysu.edu.cn (H. Zhang).

Received: 2019-06-30

Revised:  2019-09-26

Accepted:  2019-10-6

Online:  2020-05-01

Copyright:  2020 Editorial board of Journal of Materials Science & Technology Copyright reserved, Editorial board of Journal of Materials Science & Technology

More

Abstract

In the present investigation, an austenitic AISI 304 stainless steel was subjected to high strain rate surface deformation by Pipe Inner-Surface Grinding (PISG) technique. The depth-dependent deformation parameters (strain, strain rate and strain gradient) were evaluated and the microstructures were systematically characterized. Microstructural evolution from millimeter- to nano-scale was explored, with special attention paid to the localized deformation. Microstructural evolution begins with the formation of planar dislocation arrays and the twin-matrix lamellae, which is followed by the localized deformation characterized by the initiation and the development of shear bands. A twinning-dominated process that was supplemented with dislocation slip-dominated one governed the microstructural evolution inside shear bands. The twin-matrix lamellae transform into extended/lamellar structure and finally the nano-sized grains. Austenitic grains were substantially refined and martensitic transformation was effectively suppressed, of which the underlying mechanisms were analyzed.

Keywords: Microstructural evolution ; Plastic deformation ; AISI 304 stainless steel ; Pipe inner-surface grinding ; Shear band

0

PDF (6775KB) Metadata Metrics Related articles

Cite this article Export EndNote Ris Bibtex

Hongwang Zhang, Yiming Zhao, Yuhui Wang, Chunling Zhang, Yan Peng. On the microstructural evolution pattern toward nano-scale of an AISI 304 stainless steel during high strain rate surface deformation[J]. Journal of Materials Science & Technology, 2020, 44(0): 148-159 https://doi.org/10.1016/j.jmst.2020.01.027

1. Introduction

Great efforts have been taken to understand the evolution of deformation microstructure from millimeter to nanometer scale as plastic deformation has been an effective method to fabricate desirable strong nanostructured materials [[1], [2], [3], [4]]. The microstructural evolution is essentially related to the strain accommodation modes: dislocation slip and/or deformation twinning, depending upon the deformation conditions (strain, strain rate, deformation temperature) and material parameters (grain size, stacking fault energy (SFE), etc.) [[5], [6], [7], [8]]. Twinning plays a key role in the deformation of low SFE metals in particular at high strain rate and/or under low temperatures [7,8]. However, the deformation is more complicated compared to dislocation slip-governed one, and extensive investigation is needed to establish the microstructural evolution pattern in a wide range of length scale.

One important feature for twinning-dominated process is the strongly favored localized deformation as the coarse grains were subdivided into twin-matrix lamellae (T/M lamellae), which is characterized by the formation of high density of shear bands (SBs). For instance, in an AISI 310 stainless steel cold rolled to a thickness reduction of 90% (equivalent strain of 2.6), SBs can account for 86% [9] and in a cold rolled 70:30 brass they may cover 90% of the total volume at a strain of 3 [10]. The significance of the SBs on the microstructural evolution toward submicron- and/or nano-scale for low SFE metals has been emphasized. As demonstrated by the investigations on austenitic steels deformed by cold rolling at low strain rate [11] and Cu-alloys subjected to compression at high strain rate by dynamic plastic deformation (DPD) [12], twinning-dominated microstructural evolution was followed by the development of SBs toward finer scale. However, the microstructural evolution of SBs was governed by dislocation slip, whereas twinning is of less importance. For instance, in a low SFE Cu-Al alloy subjected to DPD up to an equivalent strain of 1.7, dislocation structure follows the disappearance of the original twin-matrix lamellae inside SBs and finally the formation of nanograins [12]. Identical dislocation slip dominated microstructural evolution inside SBs is also reported in copper alloys deformed by low strain rate equal-channel angular pressing (ECAP) to a strain of 4.6 [7]. Several fundamental questions are thus arisen for answer:

(1) Whether twinning-dominated process is involved in the development of SBs? Is there the transition between twinning- and dislocation slip-dominated strain accommodation mode?

(2) What governs the microstructural evolution and the size limit after twin-matrix lamellae were completely consumed by SBs?

(3) What’s the pattern of microstructural evolution of low SFE metals involving localized deformation in a wide range of of length scales?

As most of these previous investigations were performed in different samples subjected to bulk deformation, which is unable to capture the continuous evolution of microstructure in a wide range of length scale. Our recent investigation reported a surface processing technique that is able to generate monotonic deformation through shearing with ball tip, i.e. pipe inner surface treatment (PISG) [13]. A low SFE AISI 304 stainless steel pipe has been processed and gradient distribution of plastic deformation along depth has been generated. This offers a unique advantage in answering the above fundamental questions by systematic examining the microstructures at different depths of one sample. Here in the the present study, the general pattern of evolution of deformation microstructure toward nanometer regime was explored, paying more attention to the twinning-dominated microstructural evolution inside SBs.

2. Experimental

An AISI 304 stainless steel pipe with 80 mm in inner diameter and 108 mm in outer diameter was processed by pipe inner-surface grinding (PISG). Prior to PISG processing, the starting material was subjected to solution treatment, namely, 1080 °C holding for 1 h, followed by quenching into water, producing a single face centered cubic (fcc) austenitic structure with an average grain size of 50μm. PISG was performed under room temperature through shearing the sample surface by three carbide balls (8 mm in diameter) that were arranged on a tool bar at an interval of 120˚. After penetrating into the surface for certain depth, the balls rotate at a speed of V1 together with the tool bar while move at a speed of V2 along the pipe. The sample surface was thus sheared from one end to the other, undergoing 1-pass processing. The detailed processing can be found elsewhere [13]. Here, in the present investigation, AISI 304 stainless steel sample was subjected to 4-pass processing with V1 = 80 rpm, V2 = 50 mm/min and a penetration depth of 40μm.

The deformation microstructures were characterized by FEI-Scios scanning electron microscopy (SEM) with electron channel contrast (ECC) technique and FEI-Talos transmission electron microscopy (TEM, operated at 200 kV). Cross-sectional samples were prepared: i) electro-deposition of protecting Ni layer; ii) cutting 0.8 mm-thick slices and mechanical thinning down to 40μm; iii) final thinning by a double-jet electrolytic polishing (electrolyte: 10 vol.% HClO4 + 90 vol.% C2H5OH, 0 °C) supplemented with ion-milling (Leica 102) to generate large transparent areas.

3. Results

3.1. Deformation parameters

Following previous investigations [6,13,14], strain, strain rate and strain gradient can be estimated according to the deflection of an annealing twin boundary that is originally perpendicular to the treated surface. An exponential dependence of displacement (y) on depth (x) was commonly expressed as:

y(x)=ysexpkx(1)

ln(y)=ln(ys)+kx(2)

where ys and k are constants that can be determined by fitting the data using Eq. (2). However, the influence of surface curvature radius (R) on the data fitting should be considered. For pipe inner surface (Fig. 1(a)), the depth (x) for given displacement (y) in the X-O-Y coordination is actually smaller than that for flat surface by$\Delta x=R-\sqrt{R^{2}-y^{2}}$, and thus needs to be amended by x+x. The rod surface, on the contrary, has the measured depth amended by x-x. Besides, as seem from Fig. 1(b), even the influence of surface curvature was considered, ln(y) and x do not hold a strict linear relationship, implying the deviation from the exponential dependence expressed by Eq. (1). Here we tentatively replace ln(ys) in Eq. (2) by its Taylor expansion:

ln(y)=ln(ys)+k1x+k2x2+k3x3+o(x3) (3)

where k1, k2, k3 are the constants that can be determined by the experimental data. As depth data were amended (R = 5 mm for the surface mechanical grinding treatment (SMGT) Ni [6], R = 40 mm for the PISG 304 stainless steel [13]), Eq. (3) gives better fitting as shown in Fig. 3(b). As a result, the deformation parameters for the 4-pass PISG sample were estimated by the following equation:

y(x)=ysexp(k1x+k2x2+k3x3) (4)

Fig. 1.   (a) Schematic illustration of the influence of surface curvature on the calculation of deformation parameters according to the deflection of inner marker for pipe inner surface and rod outer surface, (b) Fitting the displacement-depth using Eqs. (2) and (3) of 8-pass SMGT Ni [6], 1-pass PISG 304 stainless steel [13] and the present 4-pass PISG 304 stainless steel; (c-f) Evaluation of the depth-dependent strain (e), strain gradient and strain rate (f) according to the displacement of twin boundary (d) measured in the SEM image (c).

According to the cross-sectional SEM image, ys = 298.9 μm, k1=-0.041, k2 = 1.77 × 10-4 and k3=-7.85 × 10-7, the shear strain (ε) and strain gradient (χ) can thus be calculated:

$ε(x)=- \frac{\partial y }{\partial x }=(12.3-0.11x+7.05x^{2}) \\ (0.041x+1.77 \times 10^{-4}x^{2}-7.85 \times 10^{-7}x^{3})$(5)

$x=\frac{\partial y }{\partial x }=(-0.61+14.21x-0.289x^{2}-2.496 \times 10^{-3}x^{3} \\ -1.66 \times 10^{-5}x^{4})exp(-0.041x+1.77 \times 10^{-4}x^{2}-7.85 \\ \times 10^{-7}x^{3})$(6)

As shown by Fig. 1(e and f), 4-pass PISG induces a shear strain of 9 on the topmost surface, namely, a comparable level for most heavy plastic deformation techniques [1,2,5], and a strain gradient of 0.4 μm-1 that is one order’s magnitude larger than that induced by high pressure torsion, namely, ~10-2 μm-1 [4]. A shear distance of ys was induced by the rotating balls at a velocity of V1 in a duration of t. Assuming that no slip occurs between ball tip and sample surface, the straining time can be: t = ys/(2πRV1). A rough estimation of the straining time is 10-3-10-4s, corresponding to a strain rate on the topmost surface of 104-105 s-1. Strain, strain rate and strain gradient decrease with depth to low level as a depth >100 μm. Note that it is a low bound estimation as twin boundary deflection was not detected at depths >100 μm, whereas low strain deformation microstructures like dislocations and deformation twins were actually generated.

3.2. Deformation microstructure

3.2.1. SEM observations

Three regions were observed along depth in the 4-pass PISG sample (Fig. 2), showing distinct deformation microstructures: i) Z1 (700-250 μm), a less deformed region characterized by dislocations and deformation twins. Dislocations slip dominates the deformation at deep Z1 close to the undeformed matrix, leading to the different contrasts inside the grain interiors owing to the induced orientation change (Fig. 2(c)). Twinning was activated at smaller depths, which manifests itself by the straight and parallel lines in the grain interiors. The twinning system increases at smaller depths and multiple twins were frequently activated inside one grain extending along different directions and intersecting each other. Such deformation microstructures are typical for AISI 304 stainless steel subjected to plastic deformation at relatively small strains, as detailed in previous literatures [5,9,11]. ii) Z2 (250-40 μm), a highly deformed region composed of fine scale twins and SBs. Here the grains commonly form multiple twins of two or three systems, owing to the enhanced plastic deformation. Note that the twinning system varies from grain to grain, reflecting the influence of crystallographic orientations on the twinning behaviors. Twins subdivide the coarse grains into T/M lamellae, of which the lamellar thickness decreases with an increase of twin density at smaller depths. SBs are other distinct structural features that are characterized by a bundle of parallel bands that cut through the T/M lamellae. SBs start to form in deep Z2, which bend the original T/M lamellae along the shear direction, as demonstrated by Fig. 2(b) that the straight twin boundaries change into S-shape (guided by the dotted line). At smaller depths, SBs increase their density and the T/M lamellae were finally destroyed. iii) Z3 (<40 μm), a most heavily deformed region with uniform contrast. Significantly refined microstructures are unable to be revealed by SEM, which will in the following be detailed by TEM.

Fig. 2.   Cross-sectional SEM-ECC images of the 4-pass PISG sample, showing the gradient deformation microstructures along depth (a) and the structural detail in the rectangle marked areas (b, c). Three distinct deformation regions (Z1, Z2 and Z3) were observed. The dotted line in (b) guides the tilted twin boundaries and the white dashed lines indicate the SB.

3.2.2. TEM observations

3.2.2.1. Z1 (700-250 μm)

Dislocations activated at relatively deeper layers in Z1 are planar dislocation arrays. It is a dislocation configuration where dislocations were arranged in the planar form on their respective {111} planes. Such configuration can be manifested vividly as the sample was tilted with [011] parallel with the electron beam, where the traces of the two (1$\bar{1}$1) and (1$\bar{1}$1) planes on the (011) plane can be determined by the inserted selected area diffraction pattern (SAED), being roughly intersected by ~57°, Fig. 3(a). Dislocations are observed with planar form along the traces of these two {111} planes.

Fig. 3.   Cross-sectional TEM observations of the dislocation structure (a) and deformation twins (b) in Z1 of the 4-pass PISG sample. TEM sample was tilted with [011] parallel with the electron beam, where diffraction spots from twin and matrix are mirror symmetric along the twin plane normal indicated by the dashed line in the SAED pattern in (b).

Deformation twins at smaller depth in Z1 have their crystal lattices symmetric with those of the adjacent matrix with respect to the twinning plane. The orientation relationship between twin and matrix manifest itself by the SAED pattern as the <011> of the matrix or twin was tilted parallel with the electron beam. As demonstrated by the inserted SAED pattern in Fig. 2(b), [011] zone axis for both matrix and twin are parallel, generating two overlapped single crystal diffraction patterns that have a rotation of 70.5° by [011] or are mirror symmetric with respect to the (1$\bar{1}$1) twinning plane normal (dashed line in the inserted diffraction pattern). Under such condition, twinning plane shows narrowest thickness, i.e. edge-on position, and can be indexed according to SAED pattern. Note that high densities of dislocations are presented in T/M lamellae, and these dislocations are apparently not the planar form, but exhibit complicated dislocation tangle that has been observed in previous investigations [8,9]. From a SEM-ECC image (Fig. 4(a)), multiple twins in one austenitic grain are intersected by 57° and subdivide the coarse grain into submicron-scale parallelogramic units bordered by twinning boundaries. The activated twin systems as well as the twin-twin intersection angle differ from grain to grain and hence the subdivision effect varies. Besides, another important observation is the formation of deformation induced martensite (indicated by α) that commonly appears at the twin-twin intersection sites. Multiple twins were further characterized by TEM under a condition that [110] is parallel with the electron beam (Fig. 4(b)). The inserted SAED pattern obtained from the circled area is overlapped by three single crystal diffraction patterns. Mirror symmetric twin-matrix [110] patterns evidence the multiple twins, while the rest one is the α-martensite with body centered cubic (bcc) lattice structure. Note that the [110] of γ-austenitic crystallographic direction is parallel with the [111] of α-martensite, and the (11$\bar{1}$) of γ-austenitic crystallographic plane is parallel with the (1$\bar{1}$0) of α-martensite, agreeing with the classical K-S relationship between martensite and the parent austenite: α<111 > //γ<110 > , and α{110}//γ{111} [15]. The dark-field image (Fig. 4(b)) operated by using the diffraction spot of α-martensite confirms the deformation-induced martensite transformation occurs at the twin-twin intersection sites, agreeing which the observations in previous literatures [8,16,17].

Fig. 4.   Cross-sectional SEM-ECC image (a) and dark filed TEM image (b) of the twin-twin intersection in Z1 of the 4-pass PISG sample. Note that deformation induced martensite (α) in the intersection sites, giving rise to an overlapped SAED pattern by α[111] and γ [110].

3.2.2.2. Z2 (250-40 μm)

As seen from the dark field TEM image together with the inserted SAED in Fig. 5(a), alternatively distributed bright and dark T/M lamellae show thickness of tens of nanometers. A dark band crosses the T/M lamellae, marking the local shearing that changes the crystallographic orientations for both twin and matrix. Note that such orientation change is so slight that the SAED pattern remains unchanged. The band shows wedge shape, reflecting the propagation of the local shearing from left to right. The original straight and parallel T/M lamellae in the narrow right part of the wedge-band gradually become bended in the right wide part (marked by the arrow in Fig. 5(b)). Some T/M lamellae become necking and finally invisible.

Fig. 5.   (a) Dark field cross-sectional TEM image and (b) the enlargement of the rectangle marked area (a) of the local shearing of deformation twins in Z2 of the 4-pass PISG sample. Dotted lines in (b) indicate the sheared region and an arrow indicates the bended twin boundaries.

At smaller depth in Z2, local shear is extensive and SBs were formed among the T/M lamellae. From Fig. 6(a), a SB (indicated by the dotted lines) of 200nm thick makes the original T/M lamellae bend toward shear direction and changes the inclination angle between the traces of twin boundaries and the shear direction (the dashed line) from original θ0 to θ. The maximum shear strain (ε) can be estimated following the method in [12]:

ε=cot(θ)-cot(θ0) (7)

Fig. 6.   (a) Bright and (b) dark field TEM images of the microstructure in Z2 of the 4-pass PISG sample, showing the SB formed in twin-matrix lamellae. (c) The local enlargement of (a) at a position where [011] of the SB is parallel with the electron beam. In (a), shear strain (ε) was calculated according to the angle between shearing direction and the original (θ0) and sheared (θ) twin boundaries. Note that SB contains twin substructure that is irrelevant to the original twin-matrix lamellae.

From Fig. 6(a), θ0 and θ are 60° and 40°, respectively, corresponding to a strain of ~0.61. The local shear of the T/M lamellae can be clearly resolved from the dark field image obtained by using the twin diffraction, Fig. 6(b), where the original twin-matrix lamellae experience bending, necking and fragmenting, leaving segments inside SB.

Layered structures were observed inside SB. According to the contrast difference in Fig. 6(b), these layered structures are oriented differently from the original or bended T/M lamellae. When the layers in “A” were tilted with the [011] parallel with the electron beam, typical twin-matrix diffraction pattern was obtained (the inserted SAED patterns in Fig. 6(c)), implying that these layers are T/M lamellae. Unlike the 40° inclination to the shear direction for the bended original T/M lamellae, these layers extend along the SB propagation direction or the shear direction. Their thickness is much smaller than that of the T/M lamellae outside the SB. These T/M lamellae are thus viewed as newly formed ones rather than the original T/M lamellae that were involved into SBs by shearing. The present evidenced T/M lamellae differ from previous investigations [7,9,[11], [12], [13]], where the layered structures were believed to be dislocation structures. These T/M lamellae will subject to deformation by local shearing and give rise to orientation changes, as demonstrated by the elongated diffraction spots from the SAED pattern obtained in area “B” in Fig. 6(c), implying the deviating from the standard twinning relationship by several degrees around [011].

SBs formed in multiple twins show identical microstructures. One 500 nm thick SB was shown in Fig. 7 to locally shear two sets of T/M lamellae (T1 and T2). The straight and parallel layers inside SB are observed along the shearing direction and with much finer scale compared to that of the outside T1 and T2 lamellae. Shear direction or the interior layer structure is inclined with the traces of T1 and T2 by 8° and 50°, respectively. When the [011] of the intersected T1 and T2 in the circled region “b1” was tilted parallel with electron beam, two overlapped twin-matrix diffraction patterns (Fig. b1, the dashed lines mark the twin plane normal of T1 and T2, respectively) was obtained. As the [011] of the SB was tilted parallel with electron beam, the layers were evidenced to be T/M lamellae by the inserted SAED pattern from “b2” (Fig. 7(b2)). Note that these T/M lamellae experienced orientation change, and a rotation up to 15° around [011] is noticed from the elongation of the diffraction spots. These T/M lamellae inside SB are also newly formed rather than the sheared T1 and T2.

Fig. 7.   (a) Cross-sectional TEM observation of SB formed in multiple twins (T1, T2) in Z2 of the 4-pass PISG sample. (b1) and (b2) the [011] SAED patterns from circled areas b1 and b2, respectively. Dashed lines in Fig. b1 indicate the twinning plane normal of T1 and T2, respectively.

At a smaller depth, local shear is extensive and the microstructure becomes very complicated. The microstructure is composed of deformed T/M lamellae and high density of SBs. As seen from the SEM-ECC image (Fig. 8(a)), T/M lamellae are curved and cut by many SBs that are roughly parallel with the shearing surface (vertical direction). These SBs subdivide the T/M lamellae into different isolated regions named as “twin islands”. With a decrease of depth, SBs increase their thickness and the original T/M lamellae gradually replace by layer structures that are parallel with the treating surface. TEM observation (Fig. 8(b)) shows that twin islands contain T/M lamellae of tens of nanometer in thickness, exhibiting a wavy appearance. The SAED pattern from circled area of “b1” (Fig. 8(b1)) shows typical twin-matrix diffraction as that in Fig. 5 but with elongated spots, implying the orientation change by a magnitude of several degrees. The surrounding layer structures of SBs have wavy boundaries with the spacing of tens to hundreds of nanometers, which differs significantly from those inside the twin island and also the T/M lamellae inside SBs in Fig. 6, Fig. 7. The SAED pattern from “b2” is composed of discontinuous diffraction circles, implying that these layers contain both high and low angle boundaries that differ from the twin boundaries in T/M lamellae in Fig. 6, Fig. 7.

Fig. 8.   (a) Cross-sectional SEM-ECC image and (b) TEM observation of extensive local shearing in Z2 of the 4-pass PISG sample. (b1) and (b2) the SAED patterns from circled areas in (b), respectively. Dashed lines in (b) indicate the boundaries between SB and the remaining twinned area.

3.2.2.3. Z3 (<40 μm)

The microstructure of Z3 is characterized by parallel and straight lamellar boundaries (Fig. 9(a)), and in between these lamellar boundaries high density of dislocations and occasionally deformation twins were observed. The spacing between lamellar boundaries is spanned from several to 250 nm with an average of 83.5 nm, Fig. 9(b). These extended structure yields a SAED pattern that is composed of continuous diffraction circles, implying the presence of many high angle boundaries. Such microstructure is the typical large strain deformation microstructure that is frequently observed in metals with high SFE, such as Ni [[4], [5], [6]]. Besides, the diffraction circles of the SAED pattern can be solely indexed as fcc austenite (γ) phase, whereas bcc structure that is commonly induced in highly deformed AISI 304 stainless steel was not detected.

Fig. 9.   (a) TEM observation and (b) the distribution of boundary spacing of extended microstructure in Z1 of the 4-pass PISG sample.

The lamellar structure experiences refinement under heavy plastic deformation at smaller depths. As demonstrated in Fig. 10(a), a lamella with a thickness of 100 nm was fragmented into fine-scale equiaxed regions by low angle boundaries that are indicated by dotted lines. From the HRTEM image of rectangle marked area in (a), Fig. 10(b), the low angle boundary is composed of an array of edge dislocations with an average distance of 3-4 nm. From the lattice image and the inserted FFT image, the (1$\bar{1}$1) planes of the crystals across the low angle boundary are deviated by 5.5°.

Fig. 10.   (a) Cross-sectional TEM observation of fragment of extended structure in Z1 of the 4-pass PISG sample. (b) The HRTEM image and the inserted IFFT image of the low angle boundary in the extended structure.

At the region close to the topmost surface, lamellar structure was replaced by equiaxed- or irregular-shaped grains (Fig. 11(a)). These grains are single austenitic grains with random orientations as reflected by the continuous diffraction circles in the inserted SAED pattern. The grain size was reduced below 60 nm with an average of 17.8 nm (Fig. 11(b)), which is very close to that for pure Ni subjected to SMGT at room temperature [6].

Fig. 11.   (a) TEM observation and (b) the boundary spacing distribution of the nanostructure in Z1 of the 4-pass PISG sample.

4. Discussion

The above systematic characterization of the deformation microstructures reveals the structural evolution of a low SFE AISI 304 stainless steel in a wide range of length scale from millimeter down to 20 nm. The general pattern can thus be built as schematically illustrated in Fig. 12, involving 4 stages which will in the following be discussed:

i) Generation of planar dislocation arrays;

ii) Formation of twin-matrix lamellae;

iii) Initiation and development of SBs;

iv) Formation of nano-sized austenitic grains

Fig. 12.   Schematic illustration of the microstructural evolution of AISI 304 stainless steel during plastic deformation.

4.1. Planar dislocation arrays and deformation twins

Plastic deformation of low SFE metals was accommodated through dislocation slip and twinning with an increase of deformation, yielding the first two processes. Dislocation slip was firstly activated but mainly restricted on their respectively slip planes owing to the low SFE that suppresses the dislocation cross-slip, forming planar dislocation arrays (see Fig. 3(a)). This dislocation configuration differs from the typical cell structure in high SFE metals, where dislocations are predominately presented in the cell boundaries but with few of them inside the equiaxed cell interiors [5,6]. In the present 4-pass PISG sample, planar dislocation arrays were formed in Z1 at a depth >400 μm, where both strain and stress are relatively low (see Fig. 1(e)). Deformation twinning occurs at smaller depth as strain and stress increase, and the grains were subdivided into the T/M lamellae by twin boundaries that can be viewed as a special high angle boundaries with misorientation of 60° around <111> [16]. When two or more sets of deformation twins were activated under large strain, structural refinement becomes more effective as martensitic transformation occurs at twin-twin intersection sites. Martensitic transformation was observed in the present case in Z1 with a depth of 300 μm (Fig. 4). This grain refinement mechanism has been proposed to operate downward nanometer scale for an AISI 304 stainless steel subjected to multidirectional plastic deformation via blasting the surface with high speed shots [8].

4.2. Shear bands

The deformation of twin-matrix lamellae gradually becomes localized and the plastic flow is concentrated in narrow regions, forming SBs that are composed of layered structures [3,7,9,11,12]. For low SFE metals, SBs were termed as brass-type as their development plays a decisive role in forming the final brass texture [18,19]. SBs are presented in the form of thin layers that are close to one of the maximum shear planes, for instance, ±35° to the rolling direction for cold rolling [9], perpendicular to the compression direction for DPD [12] and parallel to the shear plane for HPT and ECAP [3]. SBs in the present PISG sample were observed among both single and multiple twins with the extending direction roughly parallel to the surface, namely, the plane of the maximum shear.

The maximum shearing plane is commonly not coincident with the twinning plane or the closely-compacted plane. The original straight T/M lamellae were bended toward shear direction through cooperative deflection around the transversal direction that is parallel to the shear plane and perpendicular to the shear direction. This is supported by the observation that some bended twins and the original ones have their <110> paralleled with the transversal direction, i.e. the electron beam direction under zero-tilt position (Fig. 6). Such deflection can be resulted from the operation of two coplanar slip systems (111)[0$\bar{1}$1] and (111)[ $\bar{1}$01] when the shear plane is coincident with one of the octahedral planes, leading to rotation around <110> axis that is parallel with the transversal direction [20]. Besides, additional rotation away from the transversal direction is also observed, as when the crystals outside SB were tilted with [011] parallel the transversal direction, some bended T/M lamellae inside SB are apparently deviated from [011] (Fig. 6, Fig. 7). This rotation can be ascribed to the activation of only one of the two coplanar slip system (111)[0$\bar{1}$1] and (111)[ $\bar{1}$01], leading to the rotation about one of the <112> [21]. Note that {111}<112> twinning system can be viewed as reorientation of matrix toward the Goss position, and the twinning of a deflected original matrix lamella inside SB contributes to the same texture component as that of the deflected twin lamellae, the rotation of twin layers within SBs can be amplified by the newly formed deformation twinning [21]. The present formation of high density of twins inside SBs among either single or multiple twins may substantially enhances the rotation away from [011], Fig. 6, Fig. 7. Such incorporation of T/M lamellae into SBs has been one of the mechanisms responsible for the disappearance of twin relationship of the original T/M lamellae inside SBs, which was proposed to interpret the formation of brass texture {110}<112> in single crystal Cu and Cu-alloy with C {112}<111> during plane strain compression at 77 K [20]. Here in the present case, twinning inside SBs acts as a dominant mechanism to destroy the twin relationship, which cuts through the original T/M lamellae, leaving twin segments inside SBs (Fig. 6, Fig. 7). Besides, twin-slip interaction may also operate for the destroying of the twinning relationship, as a slip dislocation that can penetrate into an obstacle twin may in most cases dissociate into a slip dislocation in the twin and a partial dislocation at the twin boundaries [21]. This has been used to explain the twinning removal mechanism for a Cu-Al alloy subjected to high strain rate compression by DPD [12]. An issue remains uncertainty with respect to the formation of twins and dislocations inside SBs. These crystal defects were generated to accommodate the local shear strain. Resemble that in deep Z1 with low strain, single twinning system was activated insides SBs that are formed among both single and multiple twins, having one-directional T/M lamellae parallel with the local shearing. However, the density of T/M lamellae is much higher and the thickness is smaller, implying the local shear with low strain but high strain rate. The formation of single twins can follow the partial dislocation slip mechanisms in coarse-grained fcc metals, including pole [22], prismatic glide [23], faulted dipole [24] etc, which requires the partial dislocations slip continuously on consecutive {111} slip planes. With a reduction of structural scale, dislocation sources in the grain interior become lack and these mechanisms will cease to operate. Boundary related twinning mechanism should be activated, like coincidental overlapping of wide stacking fault ribbons inside nanosized grains [25], emission of partial dislocations from grain boundaries [26], grain boundary splitting and migration [27] and random activation of partials from grain boundaries [28]. The detailed process of formation of twins inside SBs is at present not clear and deep investigation is needed for the forthcoming paper.

The newly-formed twins inside SBs experience also the disappearance of twinning relationship. As shown by the inserted SAED patterns in Fig. 6, Fig. 7, the diffraction patterns from T/M lamellae are not perfectly symmetric along twin boundary normal but rotate about <011> by as large as 15°. Here twin nuclei were scarcely observed inside the newly formed T/M lamellae in SBs, the aforementioned disappearance of twinning through formation of twins seems of less importance. Note that the local shear inside SBs are roughly parallel with the {111} twin plane, namely the closely-compacted planes for dislocation slip, which may enhance the coplanar slip-induced rotation. This is supported by Fig. 6 that the SAED pattern of twin lamella rotates with respect to that of matrix around <011> axis by several degrees. Such rotation was also observed for SBs that were formed among the multiple twins, reaching as high as 15° (Fig. 7). The disappearance of twinning relationship was detailed in Fig. 13, where T/M lamellae inside SB were gradually replaced by layers with extended boundaries parallel the shear direction. The propagation of the extended structure along shear direction (indicated by the arrows) destroys the T/M lamellae (Fig. 13(b)). From the Fast Fourier Transformation (FFT) image (Fig. 13(c)) obtained from the squared area in Fig. 13(b), rotation away from the standard twin-relationship around <011> by 10° was underpinned. The Inverse Fast Fourier Transformation (IFFT) image (Fig. 13(d)) illustrates the lattice dislocations that are presented as series of additional half atomic planes close to the twin boundaries (TBs) with same sign, as marked by T. These dislocations arrange themselves with the distance of 1-2 nm, producing a misorientation angle (θ) of 7-15°, following $θ=\frac{b}{d}$, where b is the Burgers vectors for dislocation (0.254nm for austenite), d is the distance between dislocations.

Fig. 13.   Disappearance of twinning inside SBs of the 4-pass PISG sample. (a) TEM observation of the extended structure and twinned structure in SB; (b) HRTEM image of the propagation of extended structure toward twinned area in area “b” in a; (c) FFT and (d) IFFT image of the squared area “c” in b, indicating the disappearance of twinning by crystal rotation via dislocation slip.

Extended structure finally dominates the microstructure inside SBs. The extended structure is characterized by the parallel boundaries that are roughly aligned with the shear direction, in between the extended boundaries loose dislocations and interconnecting boundaries are presented. The extended structure has an average boundary spacing of around 50 nm that is one order of magnitude larger than that (5 nm) of the outside twin-matrix lamellae. All these features have been reported inside the SBs for low SFE metals [3,7,[9], [10], [11], [12]], showing great resemblance with the medium to large strain extended structure in high SFE metals deformed by monotonic deformation like cold rolling [29], torsion [30], HPT [5] etc. It is thus anticipated that the dislocation-slip mechanism governs the microstructural evolution inside SBs after T/M lamellae were destroyed. Dislocations were generated to accommodate the local shear with high strain and high strain gradient inside SBs, which on the one hand destroy the T/M lamellae, on the other hand, arrange themselves into dislocation boundaries as shown in Fig. 13. Dislocation slip induced boundaries have two types according to the formation mechanisms [31]. One is the incidental dislocation boundary (IDB) like cell boundaries and interconnecting dislocation boundaries formed via trapping of moving dislocations in a random manner, the other is the geometrical necessary boundary (GNB) for instance extended boundaries, lamellar boundaries, etc. owing to the different slip activities across the boundaries. The evolution of low angle boundaries into high angle ones can be originated from the different textural components besides boundaries and/or the continuously pumping dislocations into boundaries [32]. The general trend of the dislocation slip-induced structural evolution is the cell structure and extended structured at medium to low strains, followed by the lamellar structure at large strains [[29], [30], [31], [32]]. The well-developed SBs containing typical lamellar structure that is composed of long and parallel high angle lamellar boundaries was observed, Fig. 8, reflecting the applied large localized shear strains.

4.3. Formation of nano-sized austenitic grains

At smaller depths with increased deformation, lamellar structure and the equiaxed nano-sized austenitic grains were formed owing to the propagation and thickening of SBs. Previous investigations on the deformation of austenitic stainless steel [9,11] have reported the consumption of the T/M lamellae through the development of SBs, but the formation of nano-scale austenitic lamellar and equiaxed structure was not observed. PISG induces substantial structural refinement, while the deformation-induced martensitic transformation was effectively suppressed.

The structural scale can be reduced as small as 20 nm that is one order’s of magnitude smaller than the minimum value achieved in pure metals like Ni [2,5], pure iron [4] subjected to severe plastic deformation, where the structural evolution was governed by dislocation slip. The refinement is comparable with that for high SFE Ni and IF steel deformed by surface deformation via surface mechanical grinding treatment (SMGT) [6]. The microstructural evolution after the T/M lamellae were consumed was dominated by dislocation slip, of which the size limit can be viewed governed by the equilibrium between deformation-induced dislocation pumping and dynamic recovery-induced dislocation removal, as proposed in previous investigations on the deformation induced grain refinement of high SFE metals [2,5,6]. Superior structural refinement can be ascribed to the PISG with large strain, high strain rate and high strain gradient. As demonstrated by the Fig. 1, the surface shear strain reaches ~ 9 that is comparable to most heavy plastic deformation, however the strain rate (104-105s-1) and the strain gradient (0.4 μm-1) can be orders of magnitude higher. High strain rate facilitates dislocation generation and suppresses dislocation annihilation and the dislocation density can be much elevated. Previous investigation has demonstrated that an increase of strain rate from 10-1 to 105s-1 raises the dislocation density by a factor of two [33]. The reduction of structural size is more efficient, as evidenced by a pure Ni deformed by DPD at a strain rate of 102-103s-1 that a strain of 2.3 refines structure down to ~110 nm [34], which is comparable with that in conventionally deformed bulk Ni deformed to steady-state at much high strains >10 [35]. High strain gradient promote the dislocation storage, as high density of geometrically necessary dislocations were introduced to accommodate the strain gradient, following $ρ_{GND}=\frac{4X}{b}$ [36], where ρGND is the density of geometrically necessary dislocation, b is the Burgers vector (2.54 nm) and X is the strain gradient. A rough estimation of the ρGND for the present strain gradient of 0.4 μm-1 can be 6.3 × 1015 m-2 that is much higher than the steady-state dislocation density for pure metals (2 × 1015 m-2) during heavy plastic deformation [37]. High density of geometrically necessary dislocations are expected to influence the accumulation, mutual annihilation and trapping of statistically stored dislocations that is related to the strain amplitude, an additional contribution term was added to the classical Kocks model [38]:

$d_{ρs}=(k_{0ρG}+\frac{k_{1}}{b} \sqrt{ρs}-k_{2ρs})d \varepsilon$ (8)

where $ \frac{k_{1}}{b} \sqrt{ρs}$refers as to the dislocation storage/multiplication and k2ρs the dynamic recovery. As a result, high density of dislocations were pumped in and the lamellae were further fragmented by formation of dislocation boundaries.

The suppressed martensite transformation were evidenced by the formation of single phase austenite (Fig. 11), of which the underlying causes can be attributed the PISG-induced high shear stress and high temperature rising. Previous investigation on the high pressure torsion deformed pearlitic steel has proposed that very high shear stress may enhance the driving force for martensitic transformation and promote the reserve one [39]. PISG-induced shear stress τy can be roughly estimated according to the flow stress (σy) expressed by hardness (Vickers hardness, HV/3) following von Mises criterion for shearing [40]: $τ_{y}=\frac{1}{\sqrt{3}}\sigma_{y}=\frac{HV}{3\sqrt{3}}$. A hardness of 3-5 GPa may produce shear stress of 0.5-1 GPa that is able to induce a driving force for martensite transformation of 1.1-1.65 kJ/mol, comparable to that for martensitic transformation in steels, namely, 1.2 kJ/mol [41]. Temperature rise can be partially induced by high speed shearing during PISG, which may give rise to temperature rising around one hundred of degrees. Recent investigation on the dynamic plastic deformation of AISI 304 stainless steel has shown that martensite transformation can be fully suppressed as deformation was performed at 423 K [42]. Besides, martensite transformation was not observed in the nano-austenite that was subjected to further PISG processing, the suppression may also be related to the fine structural scale. The deformation induced martensitic transformation in duplex stainless steel also demonstrated that the fine austenitic scale may effectively retard the martensite transformation [43].

5. Conclusions

An austenitic AISI 304 stainless steel pipe was processed by PISG at room temperature. The microstructures along depth were systematically characterized and the following conclusions were reached:

(1) Gradient plastic deformation was induced by PISG. Surface plastic deformation with a strain of 9, a strain rate >104 -105s-1 and a strain gradient of 0.4 μm-1 was followed by decreased deformation with depth till undeformed matrix.

(2) The microstructural evolution of low SFE metals was built. Planar dislocation arrays and twin-matrix lamellae were followed shear banding. The development of shear bands gradually replaces T/M lamellae by extended/lamellar structure and eventually the nano-sized grains.

(3) Both twinning and dislocation slip were observed during the development of shear band. Twinning inside shear bands leads to the formation of T/M lamellae that were parallel with the shear direction, which plays an important role in destroying the original T/M lamellae. The T/M lamellae inside shear bands experience losing of twinning relationship through dislocation slip-dominated process, leading to the formation and the evolution of extended/lamellar structure toward small length scale.

(4) Substantial grain refinement was induced by PISG, while deformation-induced martensitic transformation was effectively suppressed. Single phase austenitic lamellae or equiaxed grains with an average size of 20 nm can be induced by PISG. The substantial structural refinement can be related to the large strain, high strain rate and high strain gradient, while the suppression of martensitic transformation has its cause related to the high shear stress and the temperature rising.

Acknowledgment

The authors gratefully acknowledge the financial support of the Hundred Outstanding Creative Talents Projects in Hebei University, China, the Project Program of Heavy Machinery Collaborative Innovation Center, China, the National Natural Foundation of Hebei Province, China (Grant No. E2018203312).


Reference

[1] R.Z. Valiev, R.K. Islamgaliev, E.V. Alexandrov, Prog. Mater. Sci. 45(2000) 103-189.

DOI      URL      [Cited within: 2]     

[2] R. Pippan, S. Scheriau, A. Taylor, M. Hafok, A. Hohenwarter, A. Bachmaier, Ann. Rev. Mater. Res. 40(2010) 319-343.

DOI      URL      [Cited within: 4]     

[3] X.H. An, S.D. Wu, Z.G. Wang, Z.F. Zhang, Prog. Mater. Sci. 101(2019) 1-45.

DOI      URL      [Cited within: 3]     

[4] A.P. Zhilyaev, T.G. Langdon, Prog. Mater. Sci. 53(2008) 893-979.

DOI      URL      [Cited within: 4]     

[5] H.W. Zhang, X.X. Huang, N. Hansen, Acta Mater. 56(2008) 5451-5465.

DOI      URL      [Cited within: 8]      Abstract

AbstractA quantitative analysis of microstructure and strength as a function of strain is presented for polycrystalline nickel (99.5%) deformed by high-pressure torsion in the strain range 1–300 (εVM, von Mises strain). Typical lamellar structures consisting of extended boundaries and short interconnecting boundaries have been found, with additional features at large strains which are equiaxed regions, small equiaxed subgrains and deformation twins. The evolution of microstructure and microstructural parameters falls in stages: (i) the first stage at εVM = 1–12; (ii) a transition stage at εVM = 12–34; and (iii) a saturation stage at εVM ? 34. A scaling analysis of spacing between boundaries shows a universal behavior up to εVM = 300, indicating that the predominant deformation mechanism is dislocation glide whereas twin formation is of minor importance. A clear link is observed between the evolution in structure and flow stress, which can guide the development of strong metals with a structural scale extending below 50–100 nm.]]>
[6] X.C. Liu, H.W. Zhang, K. Lu, Acta Mater. 96(2015) 24-36.

DOI      URL      [Cited within: 9]     

[7] S. Qu, X.H. An, H.J. Yang, C.X. Huang, G. Yang, Q.S. Zang, Z.G. Wang, S.D. Wu, Z.F. Zhang, Acta Mater. 57(2009) 1586-1601.

DOI      URL      [Cited within: 4]      Abstract

AbstractUltrafine-grained (UFG) or nanocrystalline (NC) Cu–Al alloys were prepared using equal-channel angular pressing (ECAP) to investigate the influence of stacking fault energy (SFE) on the microstructural evolution during deformation and the corresponding mechanical properties. The grain refinement mechanism was gradually transformed from dislocation subdivision to twin fragmentation by tailoring the SFE of alloys. Meanwhile, homogeneous microstructures and nanoscale grains were readily achieved in the low-SFE Cu–Al alloys and the equilibrium grain size was decreased by lowering the SFE. Moreover, in the Cu–Al alloy with extremely low SFE, shear fracture occurred during ECAP at strain levels higher than two due to the formation of macroscopic shear bands. In addition, the normalized deformation conditions at large strain were qualitatively discussed. More significantly, the strength and uniform elongation were simultaneously improved by lowering the SFE. This simultaneity results from the formation of profuse deformation twins and microscale shear bands, and their extensive intersections.]]>
[8] H.W. Zhang, Z.K. Hei, G. Liu, J. Lu, K. Lu, Acta Mater. 51(2003) 1871-1881.

DOI      URL      [Cited within: 5]      Abstract

AbstractA nanostructured surface layer was formed on an AISI 304 stainless steel with low stacking-fault energy by means of the surface mechanical attrition treatment (SMAT). The microstructure of the surface layer of the SMATed sample was characterized by using X-ray diffraction (XRD), scanning electron microscopy (SEM) and transmission electron microscopy (TEM), and cross-sectional TEM observation was performed to reveal the deformation-driven grain refinement mechanism for the f.c.c. materials with very low stacking-fault energy during SMAT. The grain refinement process in the surface layer involves formation of planar dislocation arrays and twins in deformed grains, twin–twin intersections leading to grain subdivision and a martensite transformation as well, and formation of randomly orientated refined crystallites. The formation of nanocrystallites in the top surface layer was ascribed to the much large strain and strain rate, as well as the multidirectional repetitive loading.]]>
[9] T. Morikawa, K. Higashida,

Recrystallization-fundamental aspects and relations to deformation microstructure

, in: N. Hansen, X. Huang, J. Jensen, E.M. Laurdisen, T. Leffers, W. Pantleon, T.J. Sabine, J.A. Wert (Eds.), Proceedings of the 21stRisoe International Symposium on Materials Science, Roskilde, Denmark: Riso. Nat. Lab., 2000, pp. 467-472.

[Cited within: 7]     

[10] M. Hatherly, A.S. Malin, Scripta Metall. 18(1984) 449-454.

DOI      URL      [Cited within: 2]     

[11] T. Morikawa, D. Senba, K. Higashida, R. Onodera, Japan, 1999, pp. 891.

[Cited within: 6]     

[12] C.S. Hong, N.R. Tao, X. Huang, K. Lu, Acta Mater. 58(2010) 3103-3116.

DOI      URL      PMID      [Cited within: 8]      Abstract

The electronic and vibrational spectra of 9-(Diethylamino)-benzo(a)phenoxazin-7-ium-5-N-methacrylamide (Nile Blue-5-N-methacrylamide) are measured, and the results are compared with the theoretical values obtained by quantum chemical calculations. The geometry, electronic transitions, charge distribution, and the IR normal modes of this new dye and of its precursor Nile Blue have been computed by using Density Functional Theory (DFT) method with the functional B3LYP and the 6-31G(d) Gaussian basis set. The molecular properties of the two dyes, predicted and observed, are very similar in the electronic ground state. In the excited state, however, the longer lifetime and larger fluorescence quantum yield of the Nile Blue-5-methacrylamide is ascribed to an inhibition of the twisted intramolecular charge transfer (TICT) process, when the NH2 is substituted by the methacrylamide in the 5-position of the aromatic extended ring of the dye. The change in charge density of the N atom in 5-position, as well as the difference in dipole moment and ionization potential of the two dyes molecules, explain the attenuation of TICT process. The vibration spectra of both dyes are simulated properly by using the DFT method.
[13] H.W. Zhang, Y.M. Zhang, Y.H. Wang, H.X. Yu, C.L. Zhang, J. Mater. Sci. Technol. 34(2018) 2125-2130.

DOI      URL      [Cited within: 6]     

[14] P. Heilmann, D.A. Rigney, Wear 72 (1981) 195-217.

DOI      URL      [Cited within: 1]     

[15] G. Kurdjumov, G. Sachs, Ztsch. Phy. 64(1930) 325-343.

[Cited within: 1]     

[16] J.A. Venables, Philos. Mag. 7(1962) 35-44.

DOI      URL      [Cited within: 2]     

[17] R. Lagneborg, Acta Metall. 12(1964) 823-843.

DOI      URL      [Cited within: 1]     

[18] B.J. Duggan, M. Hatherly, W.B. Hutchinson, P.T. Wakefield, Met. Sci. 12(1978) 343-351.

[Cited within: 1]     

[19] T. Leffers, R.K. Ray, Prog. Mater. Sci. 54(2009) 351-396.

DOI      URL      [Cited within: 1]     

[20] H. Paul, A. Morawiec, E. Bouzy, J.J. Fundenberger, A. Piatkowski, Metall. Mater. Trans. 35A(2004) 3775-3786.

[Cited within: 2]     

[21] K. Morii, H. Mecking, Y. Nakayama, Acta Metall. 33(1985) 379-386.

[Cited within: 3]     

[22] A. Ookawa, J. Phys. Soc. Jpn. 25(1957), 825-825.

[Cited within: 1]     

[23] J.A. Venables, Philos. Mag. 6A(1961) 379-396.

[Cited within: 1]     

[24] M. Niewczas, G. Saada, Philos. Mag. 82A(2002) 167-191.

[Cited within: 1]     

[25] V. Yamakov, D. Wolf, S.R. Phillpot, A.K. Mukherjee, H. Gleiter, Nat. Mater. 3(2004) 43-47.

DOI      URL      PMID      [Cited within: 1]      Abstract

Molecular-dynamics simulations have recently been used to elucidate the transition with decreasing grain size from a dislocation-based to a grain-boundary-based deformation mechanism in nanocrystalline f.c.c. metals. This transition in the deformation mechanism results in a maximum yield strength at a grain size (the 'strongest size') that depends strongly on the stacking-fault energy, the elastic properties of the metal, and the magnitude of the applied stress. Here, by exploring the role of the stacking-fault energy in this crossover, we elucidate how the size of the extended dislocations nucleated from the grain boundaries affects the mechanical behaviour. Building on the fundamental physics of deformation as exposed by these simulations, we propose a two-dimensional stress-grain size deformation-mechanism map for the mechanical behaviour of nanocrystalline f.c.c. metals at low temperature. The map captures this transition in both the deformation mechanism and the related mechanical behaviour with decreasing grain size, as well as its dependence on the stacking-fault energy, the elastic properties of the material, and the applied stress level.
[26] M.W. Chen, E. Ma, K.J. Hemker, H.W. Sheng, Y.M. Wang, X.M. Cheng, Science 300 (2003) 1275-1277.

DOI      URL      PMID      [Cited within: 1]      Abstract

We report transmission electron microscope observations that provide evidence of deformation twinning in plastically deformed nanocrystalline aluminum. The presence of these twins is directly related to the nanocrystalline structure, because they are not observed in coarse-grained pure aluminum. We propose a dislocation-based model to explain the preference for deformation twins and stacking faults in nanocrystalline materials. These results underscore a transition from deformation mechanisms controlled by normal slip to those controlled by partial dislocation activity when grain size decreases to tens of nanometers, and they have implications for interpreting the unusual mechanical behavior of nanocrystalline materials.
[27] J. Wang, H.C. Huang, Appl. Phys. Lett. 85 (2004) 5983-.

[Cited within: 1]     

[28] X.L. Wu, X.Z. Liao, S.G. Srinivasan, F. Zhou, E.J. Lavernia, R.Z. Valiev, Y.T. Zhu.Phys. Rev. Lett. 100(2008) 1-4, 095701.

[Cited within: 1]     

[29] D.A. Hughes, N. Hansen, Acta Mater. 48(2000) 2985-3004.

[Cited within: 2]     

[30] D.A. Hughes, N. Hansen, Mater. Sci. Technol. 7(1991) 544-553.

[Cited within: 2]     

[31] D. Kuhlmann-Wilsdorf, N. Hansen, Scripta Metall. Mater. 25(1991) 1557-1562.

[Cited within: 2]     

[32] D.A. Hughes, N. Hansen, Acta Mater. 45(1997) 3871-3886.

DOI      URL      PMID      [Cited within: 2]      Abstract

The rheological and fractal characteristics of raw (unconditioned) and conditioned water treatment residuals (WTRs) were investigated in this study. Variations in morphology, size, and image fractal dimensions of the flocs/aggregates in these WTR systems with increasing polymer doses were analyzed. The results showed that when the raw WTRs were conditioned with the polymer CZ8688, the optimum polymer dosage was observed at 24 kg/ton dry sludge. The average diameter of irregularly shaped flocs/aggregates in the WTR suspensions increased from 42.54 μm to several hundred micrometers with increasing polymer doses. Furthermore, the aggregates in the conditioned WTR system displayed boundary/surface and mass fractals. At the optimum polymer dosage, the aggregates formed had a volumetric average diameter of about 820.7 μm, with a one-dimensional fractal dimension of 1.01 and a mass fractal dimension of 2.74 on the basis of the image analysis. Rheological tests indicated that the conditioned WTRs at the optimum polymer dosage showed higher levels of shear-thinning behavior than the raw WTRs. Variations in the limiting viscosity (η(∞)) of conditioned WTRs with sludge content could be described by a linear equation, which were different from the often-observed empirical exponential relationship for most municipal sludge. With increasing temperature, the η(∞) of the raw WTRs decreased more rapidly than that of the raw WTRs. Good fitting results for the relationships between lgη(∞)∼T using the Arrhenius equation indicate that the WTRs had a much higher activation energy for viscosity of about 17.86-26.91 J/mol compared with that of anaerobic granular sludge (2.51 J/mol) (Mu and Yu, 2006). In addition, the Bingham plastic model adequately described the rheological behavior of the conditioned WTRs, whereas the rheology of the raw WTRs fit the Herschel-Bulkley model well at only certain sludge contents. Considering the good power-law relationships between the limiting viscosity and sludge content of the conditioned WTRs, their mass fractal dimensions were calculated through the models proposed by Shih et al. (1990), which were 2.48 for these conditioned WTR aggregates. The results demonstrate that conditioned WTRs behave like weak-link flocs/aggregates.
[33] B. Zhang, V.P.W. Shim, Acta Mater. 58(2010) 6810-6827.

[Cited within: 1]     

[34] Z.P. Luo, H.W. Zhang, N. Hansen, K. Lu, Acta Mater. 60(2012) 1322-1333.

[Cited within: 1]     

[35] H.W. Zhang, X. Huang, R. Pippan, N. Hansen, Acta Mater. 58(2010) 1698-1707.

[Cited within: 1]     

[36] M.F. Ashby, Philos. Mag. 21(1970) 399-424.

[Cited within: 1]     

[37] E. Schafler, M. Zehetbauer, I. Kopacz, T. Ungár, H. Amenitsch, S. Bernstorff, Phys Stat Sol 175 (1999) 501-511.

[Cited within: 1]     

[38] U.F. Kocks, H. Mecking, Prog. Mater. Sci. 48(2003) 171-273.

[Cited within: 1]     

[39] Y. Ivanisenko, I. MacLaren, X. Sauvage, R.Z. Valiev, H.J. Fecht, Acta Mater. 54(2006) 1659-1669.

[Cited within: 1]     

[40] M.F. Ashby, D.R.H. Jones, Engineering Materials. 1: An Introduction to Their Properties and Applications, 2nd ed., Butter worth Heinemann, Oxford, 1996.

[Cited within: 1]     

[41] Z. Nishiyama, Martensitic Transformations, Academic Press, New York, NY, 1978.

[Cited within: 1]     

[42] H.Y. Yi, F.K. Yan, N.R. Tao, K. Lu, Mater. Sci. Eng. 647A(2015) 152-156.

[Cited within: 1]     

[43] L. Xie, T.L. Huang, Y.H. Wang, G.L. Wu, N. Tsuji, X.X. Huang, Steel Research Int. 87(2017) 1-9, 1700169.

DOI      URL      [Cited within: 1]     

/