# Published Data

### These pages provide an access point to data contained in CCFE published journal papers. By selecting a paper, and then a specific figure or table, you can request the related underlying data if it is available for release.

### Publication Figures

Publication Date:

2019-05-13

First Author:

D. R. Mason

Title:

Atomistic-Object Kinetic Monte Carlo simulations of irradiation damage in tungsten

Paper Identifier:

Search the list of Figures using keywords (case insensitive):

To drill down to information on Data within a published figure, select a table cell where the figure's number of Data Items > 0

Use the Download buttons to download information on each listed figure, e.g. the image file and how it was created

Figure Reference | Title | Description | Number of Figure Data Items | Identifier | Download Figure Details | ||
---|---|---|---|---|---|---|---|

3a | AOKMC_MSMSE_Mason_3 | Histograms of relative frequencies of migration barriers computed over 1000 kMC steps at 600 K, using the MNB potential in tungsten [54]. Top: vacancytype objects. A single monovacancy has a barriers at exactly 1.75 eV by construction, but a quad-vacancy explores a wider range. As the quad-vacancy itself is a low energy cluster, many are above the isolated vacancy barrier. Bottom: interstitial-type objects. A single crowdion shows barriers near 0 and 0.4 eV, corresponding to translation and rotation modes, and a few higher energy transitions to other single-interstitial formations, octahedral, ?100? dumbbell etc. The di-interstitial has a wider range of translation modes where one string pulls past the other, and a few rotations are found. The 7 interstitial cluster and 19-interstitial loop show mostly translation modes. | 0 | CF/20/17 | Download | ||

3 | AOKMC_MSMSE_Mason_3b | Histograms of relative frequencies of migration barriers computed over 1000 kMC steps at 600 K, using the MNB potential in tungsten [54]. Top: vacancytype objects. A single monovacancy has a barriers at exactly 1.75 eV by construction, but a quad-vacancy explores a wider range. As the quad-vacancy itself is a low energy cluster, many are above the isolated vacancy barrier. Bottom: interstitial-type objects. A single crowdion shows barriers near 0 and 0.4 eV, corresponding to translation and rotation modes, and a few higher energy transitions to other single-interstitial formations, octahedral, ?100? dumbbell etc. The di-interstitial has a wider range of translation modes where one string pulls past the other, and a few rotations are found. The 7 interstitial cluster and 19-interstitial loop show mostly translation modes. | 0 | CF/20/18 | Download | ||

4 | AOKMC_MSMSE_Mason_4 | Data from a single run computing the diffusion constant of a 13 interstitial cluster at 600 K in a 32?×?32?×?32 unit cell box. Left: for three minutes wall time (using a single core on a desktop PC), string pulling moves are considered and stored. After this they can just be recalled, and so the code accelerates. The solid black line is an indication of MD speed, assuming a good MD code can perform one million atom update steps per second (per core), and each timestep is 1fs. Right: the (x-) position of the centre of mass of the cluster, computed using equation (6), showing expected Brownian motion with no significant drift. This is a trivial exercise using standard okMC where the rules for translations are predetermined, but less so for an on-the-fly code where the rules for translation are computed | 0 | CF/20/19 | Download | ||

5 | AOKMC_MSMSE_Mason_5 | The diffusion constant (points) computed using our model, multiplied by the square root of the interstitial count. The dashed lines are the theoretical prediction from Swinburne et al [67], and Derlet et al [46], and solid symbols data points from MD simulations reported in those papers. We conclude that the model presented here has order of magnitude correct diffusion coefficients for isolated interstitial clusters. Note that non-Arrhenius behaviour can be seen in this plot, a consequence of the temperature-dependent rate prefactor (see table 1.) | 0 | CF/20/20 | Download | ||

10a | AOKMC_MSMSE_Mason_10a | interstitials are randomly placed in close proximity and relaxed, a wide range of different structures are found. The top figure shows a histogram of binding energies found for three-interstitial clusters, with some of the structures drawn. The bottom figure shows similar histograms for cluster sizes 2–13 interstitials. Inset: a comparison of the binding energy for the minimum energy structures computed with DFT [76] and different empirical potentials. The lowest energy structures found for each cluster size are mobile ?111?-type defect clusters, but the majority are sessile and difficult to simply categorise. We conclude that if defects produced in cascades collide, they are likely to form a sessile metastable configuration before transforming to a mobile configuration. | 0 | CF/20/21 | Download | ||

AOKMC_MSMSE_Mason_10b | If interstitials are randomly placed in close proximity and relaxed, a wide range of different structures are found. The top figure shows a histogram of binding energies found for three-interstitial clusters, with some of the structures drawn. The bottom figure shows similar histograms for cluster sizes 2–13 interstitials. Inset: a comparison of the binding energy for the minimum energy structures computed with DFT [76] and different empirical potentials. The lowest energy structures found for each cluster size are mobile ?111?-type defect clusters, but the majority are sessile and difficult to simply categorise. We conclude that if defects produced in cascades collide, they are likely to form a sessile metastable configuration before transforming to a mobile configuration. | 0 | CF/20/22 | Download | |||

6 | AOKMC_MSMSE_Mason_6 | Performance of the code, plotted as simulated versus wall time. The thick black line corresponds to a good MD code, running at 1 fs/1M atoms s?1. After the first nanosecond of simulated time the kMC code overtakes MD. Note that we are considering multiple independent simulations on a single processor, and that we acknowledge parallelising MD is significantly easier than parallelising kMC. | 0 | CF/20/23 | Download | ||

7 | AOKMC_MSMSE_Mason_7 | Evolution of the energy in the cascades considered as a function of simulated time. We can broadly split the evolution into the three parts, indicated by the vertical bands. The greatest reduction in energy comes in the first few ns of simulated time. After this the energy evolves through a number of descending plateaux. At the longest times simulated vacancy motion becomes possible, and further relaxation is possible. | 0 | CF/20/24 | Download | ||

Loading... please wait...