Commercially pure nickel (Ni) was thermomechanically processed to promote an increase in Σ3 special grain boundaries. Engineering the character and chemistry of Σ3 grain boundaries in polycrystalline materials can help in improving physical, chemical, and mechanical properties leading to improved performance. Type-specific grain boundaries (special and random) were characterized using electron backscatter diffraction and the segregation behavior of elements such as Si, Al, C, O, P, Cr, Mg, Mn, B, and Fe, at the atomic level, was studied as a function of grain boundary character using atom probe tomography. These results showed that the random grain boundaries were enriched with impurities to include metal oxides, while Σ3 special grain boundaries showed little to no impurities at the grain boundaries. In addition, the influence of annealing time on the concentration of segregants on random grain boundaries was analyzed and showed clear evidence of increased concentration of segregants as annealing time was increased.
Nickel is the major alloying element in some of the most advanced alloys used today in applications ranging from nuclear power plant components to high temperature aircraft engine components [
Segregation, a local enrichment at the GB, can alter the mechanical and physical properties including tensile strength, ductility, and fracture toughness, electrical properties, and so forth [
Grain boundaries can be classified as low angle (angle of misorientation <10–15°), high angle, and coincident site lattice (CSL) [
Altering the GB character distribution in a microstructure is known as “GB Engineering” (GBE) [
The above discussion suggests that it is possible to thermomechanically process Ni-based alloys to increase the SGBs which should result in improved segregation resistance to impurities. Most GB studies have focused on models and simulations because experimental studies on segregation are difficult to conduct due to the difficulties in the randomness of sample preparation and experimental procedures, identifying the type-specific GB in such a small volume of material, isolating it for extraction, preparing the sample, and studying it atomically [
A study was designed by the authors to thermomechanically process Ni to promote the formation of SGBs and then atomically investigate the concentration of segregants in SGBs and RGBs. In addition, the effect of increased annealing time on 2 RGBs and their segregation behavior was studied. We utilized electron backscatter diffraction (EBSD) to characterize the grain boundaries and a focused ion beam (FIB) for site-specific lift-out procedure [
Commercially pure Ni, 3.12 mm thick sheet, was thermomechanically processed to drive segregation [
Specimen parameters.
Specimen | GB type | Strain (cold-rolled) | Annealing temperature (°C) | Annealing time (hours) |
---|---|---|---|---|
1 | RGB | 3% | 1000 | 30 |
1 | SGB | 3% | 1000 | 30 |
2 | RGB | 3% | 1000 | 1 |
EBSD was conducted on JEOL JSM-7000F Field Emission Scanning Electron Microscope (FESEM). The EBSD GB maps were obtained with a step size of 2 microns at 30 kV followed by analyses using the HKL Channel 5 software program where Brandon criterion was applied [
(a) Specimen 1 EBSD GB maps identifying the SGB (in red) extracted for analysis (location of extraction highlighted in yellow). (b-c) Specimen 1 SGB SEM images of the FIB preparation of the APT tip. (d) Specimen 1 SGB 3D atom-by-atom LEAP tomography reconstruction showing the cylindrical ROI for analysis (ions shown are Si only; 100% of ions are shown). (e) Specimen 1 EBSD GB maps identifying the RGB (in blue) extracted for analysis (location of extraction highlighted in yellow). (f-g) Specimen 1 RGB SEM images of the FIB preparation of the APT tip. (h) Specimen 1 RGB 3D atom-by-atom LEAP tomography reconstruction showing the cylindrical ROI for analysis (ions shown are Si only; 100% of ions are shown).
(a) Specimen 1 EBSD GB maps identifying the RGB (in blue) extracted for analysis (location of extraction highlighted in yellow). (b-c) Specimen 1 RGB SEM images of the FIB preparation of the APT tip. (d) Specimen 1 RGB 3D atom-by-atom LEAP tomography reconstruction showing the cylindrical ROI for analysis (ions shown are Si only; 100% of ions shown). (e) Specimen 2 EBSD GB maps identifying the RGB (in blue) extracted for analysis (location of extraction highlighted in yellow). (f-g) Specimen 2 RGB SEM images of the FIB preparation of the APT tip. (h) Specimen 2 RGB 3D atom-by-atom LEAP tomography reconstruction showing the cylindrical ROI for analysis (ions shown are B only; 100% of ions shown).
FEI 3D Quanta was used to lift out the grain boundaries using a typical beam current ranging from 2.3 nA to 7 nA and beam energies of 30 kV and 5 kV for milling and cleanup, respectively. Each GB was approximately 20
APT studies were conducted on an IMAGO (Cameca) LEAP 3000X-Si. LEAP provides the required sensitivity and resolution (spatial resolution of 0.1 nm–0.5 nm) to study intergranular segregation at the atomic level [
ROIs encompassing the GB and the matrix on either side were analyzed for this study as well. A rectangular prism ROI was divided into 1 nm slices and analyzed across the array parallel to the GB. Resultant 1D compositional profile graphs were developed from this data.
Figure
(a) 3D atom-by-atom LEAP reconstructions of specimen 1 SGB; point size = 1.5 and the percentage of ions shown is 100% except for Ni (0.1%) and Al (50%). (b) 3D reconstructions of specimen 1 RGB; point size = 1.5 and the percentage of ions shown is 100% except for Ni at 0.10%, Al at 50%, and Si at 50%.
A quantitative comparison of the segregation of all impurities between the RGB and SGB was evaluated by identifying a region of interest (ROI) demarcated by the cylindrical volumes that pass perpendicularly through each GB within the 3D atom-by-atom LEAP tomography reconstructions shown in Figure
RGB and SGB ROI bulk compositions of specimen 1.
RGB | SGB | ||||
---|---|---|---|---|---|
Ion type | Count (atomic) | Ranged | Ion type | Count (atomic) | Ranged |
Al | 4513.0 | 0.2273% | Al | 4649.0 | 0.2277% |
Si | 3684.0 | 0.1855% | Si | 3590.0 | 0.1758% |
Ni | 1975573.0 | 99.4936% | Ni | 2032126.0 | 99.5377% |
O | 65.0 | 0.0033% | O | 126.0 | 0.0062% |
C | 449.0 | 0.0226% | C | 120.0 | 0.0059% |
P | 232.0 | 0.0117% | |||
AlO | 86.0 | 0.0043% | |||
Cr | 51.0 | 0.0026% | |||
Mg | 122.0 | 0.0061% |
(a) 1D concentration profile graphs for the special GB of specimen 1. (b) 1D concentration profile graphs for Al, Si, C, and P for the random GB of specimen 1. (c) 1D concentration profile graphs for O, AlO, Cr, and Mg for the random GB of specimen 1.
Figure
This significant segregation of elements at the random GB and the lack of segregation, or reduced segregation in the case of Si, at the special GB provide evidence at the atomic level that special grain boundaries, specifically Σ3, do inhibit segregation of elements. The exception is silicon, although the concentration is half of that found in the RGB, which correlates with studies on segregation enthalpies of various elements [
This lower tendency for segregation at the SGB is due to a lower grain boundary energy which translates to lower diffusivity for Σ3 boundary. GB energy for Σ3 is on the order of <0.1 J/m2 compared to a high angle random grain boundary of >0.4 J/m2 [
First, the large number of coinciding lattice points translates into a more ordered structure and low free volume since one out of every three sites between the two grains coincides [
Second, segregation resistance in the SGB is caused by the inhibited GB migration due to the large fraction of Σ3 boundaries [
To further investigate the silicon enrichment at the SGB, another ROI (rectangular prism) of 30 nm × 30 nm × 50 nm was extracted from this specimen. This procedure extracts fewer ions and therefore identifies elements that are present in very low concentration. These would not necessarily show up in the entire spectrum because of various background issues.
This ROI encompassed the GB and the matrix on either side as shown in Figure
(a) The atom-by-atom 3D reconstruction with a representation of the rectangular prism ROI in yellow. (b) The IVAS image of the one-nanometer slices of the rectangular prism. (c) The concentration in atomic counts of the segregants (Al and Si) of the ROI for the SGB. (d) The concentration in atomic counts of the segregants (O and C) of the ROI for the SGB.
(a) The atom-by-atom 3D reconstruction with a representation of the rectangular prism ROI in yellow. (b) The IVAS image of the one-nanometer slices of the rectangular prism. (c) The concentration in atomic counts of the segregants (Al, Si, C, and P) of the ROI for the RGB. (d) The concentration in atomic counts of the segregants (AlO, Mg, SiO, and NiO) of the ROI for the RGB. (e) The concentration in atomic counts of the segregants (O, Cr, and Mn) of the ROI for the RGB.
This same analysis on the RGB was done for further quantification and identification of other possible segregants. Figures
The random grain boundaries from specimens 1 (annealed for 30 hours) and 2 (annealed for one hour) were studied by evaluating the effect of increased annealing times at 1000°C on segregation levels at random grain boundaries.
Figure
Atom probe tomography reconstructions of the RGB in specimen 2 (annealed for one hour); point size = 1.5 and the percentage of ions shown is 100% except for Ni (0.03%), Si (15%), Cr (50%), B (50%), and Al (7%).
A quantitative comparison of the segregation of all impurities was evaluated by identifying a region of interest (ROI) demarcated by the cylindrical volumes that pass perpendicularly through each GB within the 3D atom-by-atom LEAP tomography reconstructions shown in Figures
RGB ROI bulk compositions (atomic) for specimens 1 and 2, respectively.
Specimen 1 |
Specimen 2 |
||||
---|---|---|---|---|---|
Ion type | Count (atomic) | Ranged | Ion type | Count (atomic) | Ranged |
Al | 4513 | 0.2273% | Al | 5263.0 | 0.2238% |
Si | 3684 | 0.1855% | Si | 3824.0 | 0.1626% |
Ni | 1975573 | 99.4936% | Ni | 2340822.0 | 99.5411% |
O | 65 | 0.0033% | O | 30.0 | 0.0013% |
C | 449 | 0.0226% | C | 403.0 | 0.0171% |
P | 232 | 0.0117% | P | 49.0 | 0.0021% |
AlO | 86 | 0.0043% | B | 1030.0 | 0.0438% |
Cr | 51 | 0.0026% | Cr | 89.0 | 0.0038% |
Mg | 122 | 0.0061% | Mg | 103.0 | 0.0044% |
RGB of specimen 2 1D concentration profile graphs: (a) Al and Si, (b) C and B, (c) P, (d) Cr, (e) Mg, and (f) O.
As shown in Figure
Figures
Although this phenomenon is more likely due to complex grain boundary migration processes, this enrichment could also be due to physical site competition, which warrants some discussion. The atomic radii of P, B, and C are small as seen in Table
Atomic radii of all Ni elements and impurities.
Element | Atomic radius (pm) |
---|---|
Ni | 149 |
C | 67 |
Cu | 145 |
Fe | 156 |
Mn | 161 |
S | 88 |
Si | 111 |
Al | 118 |
B | 87 |
Cr | 166 |
Co | 152 |
Mg | 145 |
P | 98 |
V | 171 |
However, as mentioned above, the differences in segregation of these elements within the same specimen are more likely due to complex high temperature grain boundary migration processes and interactions. Annealing increases grain boundary migration, which in turn affects segregation behavior and diffusion processes [
Figure
To further investigate the lack of segregation of Si, Mg, Cr, and Al in the RGB of specimen 2, another ROI (rectangular prism) of 30 nm × 30 nm × 50 nm was exported from this specimen in IVAS which encompassed the GB and bulk matrix on either side as shown in Figure
Atom-by-atom 3D reconstruction with a representation of the rectangular prism ROI in yellow (b) and the IVAS image of the one-nanometer slices of the rectangular prism (a) analyzed (atoms shown in image are B only).
First, segregants not previously identified in the initial analyses shown in the paper thus far were found in this ROI such as B, NiO2, and Fe. Second, Figure
The concentration in atomic counts against the distance across the ROI (rectangular prism) of the segregants in the RGB of specimen 2 for (a) Al, Si, C, and B; (b) P, AlO, and NiO2; (c) Mg, Cr, and Fe.
In this study, increased annealing time does show an increase in the amount of segregation of some elements at the random GB as would be expected since an increase in annealing time at 1000°C would provide significant driving force for segregation. Si and Al, the main impurity elements in Ni, exhibit a 74% and 43% increase, respectively, in their concentration at the RGB from one-hour annealing to 30-hour annealing. The standard deviation is shown in Table
Standard deviations (in units of percentage since the data is in units of percentage).
Specimen | At.% Si | At.% Al |
---|---|---|
Specimen 1 | 0.059 | 0.042 |
Specimen 2 | 0.034 | 0.029 |
This study successfully used EBSD GB maps to identify and lift out type-specific grain boundaries to study segregation behavior of Ni at the atomic level using APT. The significant findings and observations are listed as follows: Σ3 SGBs do inhibit segregation of impurities at the GB compared to RGBs in Ni under the current research conditions. The evidence in this research suggests that increasing annealing time from 1 hour to 30 hours for RGBs does increase the concentration of segregants at the GB.
The 1000°C annealing for 30 hours should have been a driving force for segregation but the SGB still showed very little quantified data providing evidence of segregation. Some Ni-based superalloys in some gas turbine components can experience temperatures in the range of 650°C to over 1000°C [
The role of segregants at grain boundaries is not well understood and certainly not well quantified. There are questions concerning segregants and their interactions with one another and their interactions with type-specific grain boundaries [
This insight will feed future efforts in better understanding segregation behavior as well as the structure of type-specific grain boundaries during segregation in materials for performance improvement in countless applications.
Future work in this area of study could incorporate texture analysis to include the identification of the dominant crystal orientations of each type-specific grain boundary to more completely describe and understand the boundary conditions. A better understanding of the diffusion process with respect to this texture analysis could provide additional insights into the segregation behavior at the atomic level. As grain boundary migration is a complex topic, more work is needed in understanding the mechanisms at play when discussing the driving forces that would cause some impurities/elements to segregate more than others within the same sample.
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.
The authors declare that there is no conflict of interests regarding the publication of this paper.
This research is supported by the National Science Foundation under Grant no. DMR-1151109. The authors would like to thank Johnny Goodwin at the Central Analytical Facility, University of Alabama, Tuscaloosa, Alabama, for his expertise in the Electron Backscattered Diffraction and would also like to thank Chad Hornbuckle, University of Alabama, for his expertise in the local electrode atom probe. Professor Gregory B. Thompson and Dr. Monica Kapoor thank ARO-W911NF-13-1-0436 for support of this research.