Introduction either strengthen or weaken a material as

Introduction
to segregation engineering and grain boundary segregation engineering

Segregation
engineering refers to the self-organized microstructure manipulation whereby
the solutes, under thermodynamic driving force, are segregated to specific
lattice defects and consequently trigger microstructural changes at the site of
segregation. The underlying principle is that, upon heating into regimes good
enough for diffusion, the solute atoms show strong segregation at the
attractive trap sites i.e. lattice defects such as grain boundaries and
dislocations. The resulting local strain fields and alteration in chemical
composition at the lattice defects can induce structural transitions, phase
transformations and solute ordering. In order to realize segregation
engineering, certain thermodynamic and kinetic factors have to be taken into
account 1. Rather than considering such solute segregation just as an
undesirable phenomenon, segregation engineering aims to exploit it as a tool
for site-specific manipulation of interfaces or microstructures so as to obtain
desirable macroscopic material behavior.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

Grain
boundaries are inevitable defects in polycrystalline metallic materials. They
can either strengthen or weaken a material as they influence a range of
properties such as tensile strength, fracture toughness and corrosion. A lot of
grain boundary properties such as cohesion and strength are sensitive to solute
decoration. Thus, grain boundary segregation engineering (GBSE), which is the
manipulation of grain boundaries “via solute decoration or even confined
transformation”,
is an important branch of segregation engineering 2.

Adsorption isotherm

The amount of
segregation of a solute at a specific site such as grain boundary is inversely proportional
to the solubility of that solute in the bulk phase (matrix). Moreover, the
thermodynamics of grain boundary segregation are similar to monolayer gas
adsorption and hence can be modeled using the adsorption isotherm.

?i =
– (1/RT) x (d?/dln xi)T,V

Where,

?i = excess concentration of element at
grain boundary

xi = molar concentration of
element i in bulk

d?
= change in Gibbs energy upon segregation at constant temperature and volume

The equation
above represents the Gibbs adsorption isotherm. It shows the relation between
bulk concentration and solute segregation (excess concentration). However, due
to the challenge of experimentally measuring the interfacial energy as a
function of temperature and composition in the case of Gibbs adsorption isotherm,
it is more natural to use the Langmuir-McLean isotherm, which is approximates
to (system taken as dilute case):

?i
= ?iGB / ?iB = exp (-?GiGB
/ RT)

Where,

?i = segregation coefficient or grain
boundary enrichment factor

?iGB = molar grain boundary
occupation fraction of element i

?iB = molar concentration of
element i in bulk

?GiGB
= segregation free energy

The
Langmuir-McLean isotherm states that: grain boundary segregation occurs for ?GiGB
< 0 and that it decreases with increasing bulk solute content and temperature. The Langmuir-McLean isotherm model is usually more practical than its Gibbs counterpart as unlike the latter, the former does not require detailed knowledge of grain boundary energy 2. Characterization methods for evaluating grain boundary segregation Segregation engineering demands characterization methods with sufficient resolution for extraction of chemical, spatial and crystallographic information on the atomic scale. Such characterization is achieved with atom probe tomography (APT) in conjunction with another electron microscopy method such as scanning electron microscopy (SEM), scanning transmission electron microscopy (STEM), electron backscatter diffraction (EBSD) or transmission electron microscopy (TEM). Now days, almost all the electron microscopy methods can be applied directly to the atom probe tips, prepared usually by focused ion beam (FIB) cutting. Usually, the sequence of characterization is as follows: 1.       Preparation of atom probe tip using focused ion beam (FIB) cutting, 2.       Structural analysis of the segregation sites via imaging and diffraction of tip using electron microscopy, 3.       Local compositional analysis (down to ppm level precision) and evaporation of the tip using APT 1. Simulation methods for segregation engineering The progress in simulation methods presents opportunities for use in segregation engineering. Several studies have combined ab initio simulations with electron microscopy to analyze grain boundary segregation and its effects. For example, the origin of brittle fracture in Cu-Bi system was investigated by analyzing the electronic and geometric structure of the Cu grain boundary with and without Bi decoration. TEM was used to map the Bi distribution on Cu grain boundary, followed by density functional theory (DFT) calculations to find the induced charges in the electronic structure of Cu. It was revealed that at sites of Bi segregation, the Cu atoms assumed a Zn-like electronic structure, which reduced grain boundary cohesion. While DFT-based results are most precise, a drawback is that they are limited to high symmetry, low coincidence site lattice (CSL); this is because DFT calculations require periodic boundary conditions. For high CSL, atomic scale simulation methods such as molecular dynamics are appropriate. However, they provide lesser electronic detail as compared to DFT. Atomic scale simulation has been utilized to study segregation-driven grain size stabilization in nanocrystalline materials. At more mesoscopic scales, for example at single grain boundary level, phase field simulations are relevant. In fact, the phase field method has already been used to predict localized segregation-induced phase transformation at grain boundaries 2.       Segregation engineering to restore toughness via local phase transformation Mn segregation causing grain boundary embrittlement The less expensive Fe-Mn alloys have the potential to replace Fe-Ni alloys for applications demanding high strength. However, the prospect of this is hindered as it has been shown that quenched and tempered medium Mn steels (Mn 3-12 %) demonstrate strong embrittlement. It has also been observed that there is a shift in ductile to brittle transition temperature towards higher temperatures. It was observed that this temper embrittlement occurs after as little as 10 s of tempering, and is the direct consequence of Mn segregation at martensite grain boundaries, which severely reduces toughness by promoting intergranular fracture. Structural relaxation, solute-induced decohesion or local volume changed due to magnetic contributions are among the suggested mechanisms 3. Martensite-to-austenite transformation to consume segregated Mn Raabe et al. demonstrated that the ductility could be restored via consumption of Mn that segregates at the grain boundaries. Water quenched Fe-9 wt. % Mn alloy was tempered at 450 oC and 600 oC for a period between 10 s to 860 hours. The impact toughness decreased initially (as expected), but increased after holding times of 672 h and 30 seconds at 450 oC and 600 oC, respectively (Figure 1). As the holding time increased at 450 oC, the SEM images (Figure 2b and 2c) revealed a transition from a completely intergranular fracture (10 s; with prior austenite grain boundaries (PAGB) being weakest) to appearance of ductile dimples at intergranular facets (672 h). Whereas, after 3 min of tempering at 600 oC, the fracture mode is predominantly ductile (Figure 2d). Figure 1 Figure 2                       Atom probe tomography results confirmed that for tempering at 450 oC, the Mn concentration at the PAGB increased up to 18 at. % after 48 h (as expected), but reduced strongly to ~8 at. % after  336 h, whereby an Mn-rich phase with up to 32 at. % Mn appeared at the grain boundary (Figure 3). This Mn-rich phase was confirmed to be reversed austenite (?) by EBSD done after 672 h (Figure 4). Thus, the decrease in Mn concentration at the grain boundary is explained by the formation of austenite, which Figure 4 Figure 3                       consumes the Mn. This is because the nucleation of austenite requires ferrite and austenite to have equal chemical potentials, which implies oversaturation of up to 36 at. % Mn. The reduction in Mn at grain boundary reduced embrittlement and hence higher impact toughness. Besides "cleaning" the Mn, the presence of reversed austenite at grain boundary also helps to improve impact toughness via crack deflecting mechanism. A double-step heat treatment of tempering for 10 min at 600 oC followed by 10 min at 450 oC exploits the changing segregation behavior of Mn. The first step accelerates the nucleation of austenite, and then, the second step partitions the Mn into austenite, thus promoting its further growth 3. Thermodynamic analysis of martensite-to-austenite transformation In a separate study conducted by Raabe et al., thermodynamic analysis of the Fe-Mn system at 450 oC were carried out using Thermo-Calc® to estimate the Mn partitioning between newly formed austenite and martensite by using nominal alloy composition. Surprisingly, the results of this analysis matched well with that observed by APT analysis at the grain boundary where the localized martensite-to-austenite transformation had occurred at the same temperature. This shows that a thermodynamic force, which favors the transformation to austenite, exists locally in the grain boundary region. It was concluded that a low nuclear barrier for this transformation can be attributed to: 1.       A high chemical driving force owing to the high equilibrium segregation of Mn to the martensite boundary, 2.       Low free energy of the newly formed austenite-martensite interface, 3.       High elastic relaxation energy associated with the martensite. Furthermore, three criteria have been identified to select a segregating element for successful application of segregation engineering: (i) Solute should have high tendency to segregate at the defects, (ii) Solute should reduce the transformation temperature for the newly formed phase, and (iii) Should prefer local segregation over precipitation in the matrix 4. Optimization of mechanical properties via confined precipitation and dissolution of vessel phases Figure 5: reference 1 Springer et al. presented a new approach to achieve a compromise between strength and ductility in advanced structural steels. It is based on utilizing precipitates formed during ageing as temporary reservoirs of elements, the so-called vessel phases, which then partially dissolve at higher temperatures to form local chemical ingredients in the surrounding matrix. This was demonstrated on an 11.6 Cr-0.32 C (wt. %) steel, using M23C6 carbides as vessel phases, which were used to obtain locally enriched Cr and C for blending of metastable retained austenite (up to 14 vol. %) within a high strength martensitic matrix.                 Figure 6:     The processing sequence comprised of three main steps (Figure 5): 1.       Conditioning: Heat treatment (1150 oC for 2 h in argon atmosphere, followed by oil quenching to room temperature) to achieve thorough homogenization and supersaturated martensite. 2.       Accumulation: Heat treatment (750 oC for 47 h in argon atmosphere, followed by oil quenching) for nucleation and growth of precipitates i.e. the vessels. This step is critical in defining the composition and stability of the vessels. 3.       Dissolution: Heat treatment (850-1350 oC for 1-90 s, followed by hydrogen quenching to room temperature) for complete or incomplete dissolution of the vessels so that Cr and C can be locally enriched, thus lowering the martensite start temperature locally, which favors the formation of retained austenite. Figure 6 above shows the tensile test results of: (i) Complete vessel segregation engineering samples, (ii) Reference sample with conditioning and dissolution only, and (iii) Reference sample with conditioning and tempering only. It can be concluded from the results that sample (i) showed optimal combination of both strength (1.9 GPa) and ductility (~8% strain) as compared to (ii), which showed brittle fracture at low strain, and (iii), which had lower strength 5.   Linear complexions for nanostructuring of alloys Raabe et al. reported the formation of stable austenitic confined structure on edge dislocations in an otherwise martensitic matrix in Fe-9 at. % Mn. The formation of this localized austenite is believed to be a segregation-driven local transformation as a consequence of Mn solute composition and elastic strain levels at the defect being sufficient to stabilize a phase which is different from the matrix. It is suggested that the concept of these linear complexions extends the Gibbs isotherm to interface stabilized structural states. It was found that these linear complexions can be formed by annealing the quenched and cold-rolled (50% reduction) Fe-9 at. % Mn at 450 oC for 6 hours. Increasing the annealing treatment for longer holding time did not cause further growth or change in size or composition of the austenite, thereby confirming the confined nature of a complexion. Moreover, an excellent agreement between the thermodynamically predicted equilibrium Mn concentration in austenite and experimentally observed local composition was noted. With the discovery of linear complexions, the opportunities for nanostructuring alloys via targeted segregation 6. Challenges and outlook in segregation engineering The simultaneous characterization of both the structure and chemistry of grain boundary at atomic scale is one of the challenges in segregation engineering. The underlying difficulty there are multiple degrees of freedom associated with the description of a grain boundary, and all of them affect segregation. This means there is a huge number of the type of grain boundaries, and hence their correlation with segregation for effective manipulation is a tedious task. The simplification of grain boundary into three classes, namely high angle random, low angle and highly coherent, has various limitations. Another challenge lies in using appropriate simulation method. DFT is not adequate for chemical trend screening of grain boundary decoration. Therefore, it is suggested that phase field crystal method, with incorporation of atomic interactions obtained by DFT, is worth considering. There are several opportunities presented by segregation engineering that need to be investigated further. Applications where material performance may be improved include grain boundary oxidation and cohesion, and damage resistance. Although the focus in this review was on improving the mechanical properties of metallic alloys, the principles of segregation engineering can be explored for functional materials such as semiconductors 2.