Dislocation Interactions with Irradiation Produced Defects: In Situ TEM Deformation Study
I. M. Robertson

Deformation experiments can be conducted in situ in the TEM but they have a complex stress state with cracks around the edges of the hole of the sample, and one cannot determine anything about the stress states of the materials. A new method is to deposit the test material onto Si and use lithographic processes to create the test specimen that is completely free standing. Then the sample can be loaded into a conventional TEM strain stage. The applied load and strain can be measured using this method. They are, however, difficult to fabricate with a low success rate and difficult to load in the TEM.
All three modes of dislocation deformation, twinning, partial dislocation and full dislocations can operate simultaneously, even at the grain boundary. A grain boundary source will eject all the same type of dislocation, but not necessarily at the same point. Using in situ TEM deformation of irradiated material shows the formation of dislocation channels just like the bulk irradiated material, giving confidence that in situ studies are going to reliably mimic the bulk. Preexisting dislocation are not a source of dislocation channels, however. The width of the channels is caused by the presence of dislocations of the same slip system on multiple planes.
Dislocation obstacle strength can be measured in situ and it shows that one value does not characterize obstacle strength, but rather a range of values are needed. Similar measurements can be made for dislocation interaction with He bubbles. MD simulations from B. Wirth suggest that dislocations actually shear the He bubbles. Screw dislocations are much more efficient than edge dislocations at annihilating defects. There are no obvious effects of elevated temperatures on the mechanism for dislocation movement and interaction.
Loop chasing shows that a dislocation loop moving can cause the movement of a dislocation line ahead of it, forcing it to a sink and setting it up for annihilation. The reason for this is that the line is attracted to the loop, allowing them to move together. There is no one mechanism for dislocation interaction, however, because as more dislocations are introduced into a system, the number and types of dislocation interactions become more complex.


Atomic-scale Plasticity in Presence of Frank Loops
David Rodney

The linking of plasticity on a multi-scale level is being done to model deformation of irradiated materials. On the atomic scale using MD, a systematic study of the interactions between edge/screw dislocation and Frank loops was simulated. The purpose was to get an “interaction matrix” to define the defect configuration. There is a strong analogy in interaction mechanisms between a Frank loop and a stacking fault tetrahedral (SFT). The interactions between dislocations and loops can be described in 3 ways: simple shear of a dislocation by the loop, the partial absorption of a loop by the dislocation, and the annihilation of the loop by the dislocation, which transforms into a helical structure. The latter mechanism results in the creation of a screw dislocation because the dislocation is strongly pinned after absorption of the loop. Screw dislocations can glide to create edge dislocations are unpin by Orowan mechanisms and leave behind large, jogged prismatic loops.
Work is just beginning on dislocation dynamics simulations where the MD dislocation interactions can be used to study the clear band formation of dislocation channels as a result of screw dislocations. The key reaction is the formation of a helical turn and unpinning a plane difference from the initial glide plane which appears to be sufficient to form the clear bands.


In-situ TEM Observation of ODS EUROFER97 and Its Relation to Mechanical Properties
Amuthan Ramar

The purpose of oxide-dispersion strengthened steels is to increase the upper operational temperature limit for ferritic-martenisitic steels and decrease radiation effects to mechanical properties. In the both the rolling and transverse direction, the material has a low DBTT below 0°C and a large drop in strength after 500°C. The microstructure has a grain size of 2 – 5 µm with an average yttria particle size of 8 nm. Heat treatments at 800°C show a large reduction in dislocation density with larger ferrite grains. Changes in the microstructure start to occur at 600°C and by 900°C the microstructure is very ‘clean’ and largely free of a dislocation network. However, the yttria particles are extremely stable at all temperatures tested. Differences in yield stress of the ODS steel can be seen only at temperatures above 500°C.
Activation energies are very large, above 3.6 eV at a temperature of 500°C and increasing with temperature. This large activation energy correlates with a dislocation climb mechanism, suggesting perhaps the climb of dislocations over the yttria particles that requires such a large activation energy.
The ODS steel has a higher strength but is also more brittle than the reference material. It has lower elongation at high temperatures. The microstructure is tempered martensite with very stable yttria particles that are at most semi-coherent in the matrix. There does appear to be an arrangement of the yttria particles, but it does not correspond to alignment with the lath boundaries and its arrangement is not really understood.


Atomic-Scale Modeling of Dislocation Dynamics in Environment of Radiation Defects
Yury Osetsky

In irradiated Cu, there is a high density of small stacking faults with a narrow size distribution. The goal of atomic-scale modeling is to understand the details of dislocation dynamics in a complicated environment. In particular modeling the strain rate, temperature and kinetic effects, among other phenomena, on dislocations needs to be understood. As a function of straining, with higher strain, as the shear stress increases, voids with increasing sizes are eliminated from the materials. By the time the largest void size is eliminated, softening of the material occurs. Dislocation interaction with a void also causes a drop in shear stress as there is partial dislocation elimination with the interaction with the void. In terms of stress-strain relationships, the behavior of small voids and small precipitates is very similar.
In modeling of critical resolved shear stress (CRSS), line modeling always predicted a higher stress than atomistic modeling. Continuum theory does well at modeling voids in Fe, however, it is poor at calculating voids or precipitates in Cu. Continuum modeling does well with strong obstacles based on Orowan mechanisms, but it does not describe well large particles with lower obstacle strength.


Radiation-Induced Segregation in Advanced Reactor Materials
Jeremy Busby

Radiation-induced segregation (RIS) has been studied extensively precisely because it has been correlated with cracking of stainless steels after irradiation. The loss of grain boundary Cr below the amount that makes the steel stainless means a loss of passive oxide formation that protects the steel in water environments. The creation of point defects from irradiation above thermal equilibrium levels means that the defects will migrate to sinks such as grain boundaries. The effect is highly dependent on temperature and dose rate, among other effects, but the range is broad and encompasses light water reactor operation. However, corrosion is not the only phenomenon which RIS can affect that is of concern for steels in advanced reactor concepts. Moreover, RIS is a concern in more than just Fe-Cr-Ni alloys.
RIS can lead to local concentrations that exceed solubility limits which leads to radiation-induced precipitation (RIP). The latter phenomenon is also highly dependent on temperature and dose rate.
Mo-Re alloys have been developed for high temperature applications due to the high melting temperature and high strength. At 800°C the alloy retains some ductility and fails under ductile fracture. However, by 1100°C the failure is brittle, intergranular fracture, and most ductility is lost. From 800 - 1100°C the failure transitions from ductile to brittle, and this transition can be correlated with RIS, which peaks in this temperature range. Additionally, transmutation complicates the phenomenon further. RIS and RIP in this system subsequently have a significant effect on mechanical properties.
In ferritic-martensitic steels, RIS could also have an impact on mechanical properties. Although difficult to measure, RIS seems to cause either the enrichment or depletion of Cr at the grain boundary, depending on the alloy with just small changes in the composition. This is very different than austenitic steels where Cr only depletes. Regardless of the trend, RIS can lead to precipitation of α’ which results in material embrittlement.
Rather than trying to experimentally measure a wide range of alloy systems, it would be preferable to develop models that can help to understand the RIS phenomenon. However, while ternary alloy codes do exist, to model minor solute additions requires interaction parameters that have to be measured experimentally and are difficult and time-consuming to do. New approaches do exist, however, by using first principles calculations such as ab initio to determine these interaction parameters rather than having to obtain them experimentally. With this method, interaction parameters can be determined, refined using MD simulations and fed into kinetic rate theory models to predict RIS. Then experimental measurements can be used as a feedback to benchmark the codes and further refine the multi-scale modeling approach.


Microstructural Development and Solutes Segregation during Electro-Irradiation around Cascade Damage Introduce by Ion-Irradiation
Heishichiro Takahashi

Irradiations have been performed on Fe-20Ni-15Cr-0.5Si alloy using Ni+ ion, He+ ion implantation and electron irradiation in a multi-beam HVEM. The Ni+ ions introduce cascade damage as opposed to electron irradiations which create only isolated point defect pairs. Cascade damage with further electron irradiation can transform into stacking fault tetrahedral (SFT) or be reduced to smaller, isolate cascades. The clusters of vacancies or interstitials can be transformed to the SFT or dislocation loops, respectively.
Looking at the SFT with HREM, the contrast of the defect can be imaged and correlated with the cascade damage region. After annealing, the cascades transform into these defects. Introducing cascade damage around a bubble, however, leaves the bubble intact and does not transform its nature. Focusing closer on the damage cascade, cluster composition around these cascades appears to change from the matrix material. Results suggest that in the cascade region, segregation behavior is manifested similar to the process that occurs at other sinks. Cr depletes and Ni enriches around these damage cascade clusters at substantial levels. This shows that the ion damage cascades act as sinks at which segregation can occur, making them similar to other sinks in the material. Moreover, radiation-induced precipitation has also been observed around the damage cascades. He bubbles, however, do not demonstrate any substantial segregation behavior.


Monte Carlo Simulation of Grain Boundary Segregation in Fe-Cr and Fe-Cr-Ni
Je-Wook Jang

Using experimental grain boundary segregation data, the purpose of this work was to develop interatomic potentials for Fe-Cr-Ni alloy systems. The goal is to try to reproduce the segregation behavior using Monte Carlo (MC). This work uses 2nd nearest neighbor Modified EAM potentials. While the MEAM is able to accurately calculate lattice parameter for bcc Fe-Cr systems, there are some discrepancies for the fcc Fe-Ni system. The conditions for the MC simulation are for high angle twist grain boundaries (GB) and low angle tilt GBs.
For the Fe-Cr system, MC simulations show that Cr enriches at the grain boundary, going from disordered to an ordering of Cr toward the boundary. For the bcc Fe-10Ni system, there is some segregation seen, but none shown for the fcc Fe-10Ni system.
For the Fe-18-Cr-13Ni, phase separation is seen on high angle twist boundaries, with large Cr depletion. However, no segregation is observed for tilt boundaries.
Grain boundary segregation is very different between bcc and fcc systems, where bcc system shows strong segregation for Fe-Cr but none for the Fe-Ni alloy. For Fe-Cr-Ni systems, only the twist GB for bcc alloy shows strong Cr segregation. No segregation is observed for the fcc system. Some errors probably exist with the Fe-Ni binary system due to an incomplete description of the composition dependence on the lattice parameter.


Ab initio-Based Thermokinetics of the Ni-Cr Binary
Julie Tucker (Dane Morgan)

The project goal of this work is to try to understand how diffusion constants in the Fe-Cr-Ni are affected by composition and temperature in order to simulate the grain boundary segregation process. This presentation focuses on just the Fe-Cr system, however.
To start calculations determine the activation barriers for migration, in particular for interstitial migration. The result was that Ni migration through interstitials was much larger, ~ 250 meV higher, than it was for Cr. Studying the effect of composition by clustering Cr around the interstitial by placing them around the nearest neighbor shells ends up showing that with increasing Cr concentration, both Ni and Cr migration energies decrease generally. The decrease is even more dramatic for Cr migration, however, until it reaches a ‘negative’ migration energy, which just means that the Cr atom is more stable as an interstitial at the saddle point rather than traditional interstitial lattice sites when the environment is Cr rich. The negative migration energy does not manifest any instability in the crystal structure around the interstitial, however.
The message here is that electronic interactions between the Cr interstitial and other Cr atoms or Ni atoms are needed to understand these migration energies. Simple strain effects do not explain the migration energy barriers as they change as a function of composition.
Using this barrier influence on diffusion in a dilute Cr system, diffusion constants can be calculated. Also correlation factors are determined using the five-frequency model. The Ni self-diffusion constants and Cr self-diffusion constants are compared to experimental data at high temperature, and the trends in diffusion data match experimental data well, certainly within the error of the experimental measurements. However, when translating this work to lower temperatures, the calculations do not correlate well with experimental data. While it might seem that low-temperature calculations are not accurate, in fact experimental measurements are made at high temperature and extrapolated at low temperature. Calculations, however, are made at 0K, so the fact that they match experimental measurements at high temperature suggests that they should be equally if not more realistic at low temperature.
The work here shows that there is significant local environment dependence. The Cr barrier is significantly lower than Ni, but the lower barrier does not necessarily result in such a huge difference in the Cr diffusion constant because while the exponential for the diffusion constant for Cr increases dramatically, the correlation factor between Cr and Ni significantly decreases, and results in only about an order of magnitude increase in the Cr diffusion constant. Although confusing, one can think that while the barrier is much lower, Cr diffusion will eventually become trapped in the matrix of Ni and slow the diffusion of Cr substantially, making the large barrier energy difference not as significant as it would be otherwise.
TextText