Life expectancy - graphite behaviour in nuclear reactors

Materials World magazine
3 Jul 2011
Neutron collisions create and/or move basal dislocations in graphite

A joint research project between six UK universities is assessing the durability of materials for long-term nuclear power generation. Dr Aidan Westwood and Dr Andrew Scott from Leeds University, UK, highlight studies underway to investigate the long-term behaviour of graphite used in nuclear reactors.

Graphite is the moderator and a major structural component in 90% of UK nuclear power generation capacity, and is proposed for use in future high temperature gas-cooled reactors (HTRs) that need to operate for 60-100 years. These reactors are seen internationally as an important source of power, such as for hydrogen production and in particular for direct process heat.

However, there are challenges. Under neutron irradiation, the graphite crystallites change their dimensions and shape, and polycrystalline blocks become more porous. Designs for these fourth generation reactors (Gen IV) that operate above 1,000K and for longer lifetimes, may be held up due to gaps in knowledge about how graphite properties are affected by irradiation. Lack of knowledge could also affect the dates at which it would no longer be possible to underwrite safe operation of the UK’s advanced gas-cooled reactor (AGR) fleet. This would have an impact on the UK’s commitments to carbon emission reduction to 2020 and beyond.

It is increasingly recognised that existing ‘mechanistic’ models, invoking atomic displacements by neutrons, do not fully explain the observed behaviour of graphite in these reactors, and such atomic-scale models do not extend to predictions of multi-scale behaviour. Furthermore, the UK is suffering from a declining expert knowledge base, with the ageing and retirement, of key proponents. Substantial gaps in understanding have become apparent. Mistakenly, regarded as a mature research field for which the fundamentals have long been solved, the effort put into the science of radiation damage has dwindled.

Collaborative research

To begin to redress this deficiency, a £1.3m grant from the Engineering and Physical Sciences Research Council (EPSRC) has recently been awarded to researchers from the Universities of Leeds, Manchester, Nottingham, Salford (The University of Huddersfield is now also involved, following relocation of key personnel from Salford) and Sussex to help improve understanding of irradiated graphite, using state-of-the-art techniques that were not available during the 1960s and 1970s. There is potential for this information to support an extension to the planned lifetimes of nuclear power reactors. In-depth microscopic understanding of neutron irradiation damage evolution in nuclear graphite will feed into a novel multi-scale modelling framework to predict behaviour at larger length scales and higher total neutron irradiation doses. This is necessary to assist industry in validating its models for use in safety and life extension cases, and for the design of Gen IV reactors.

The project will run for three-and-a-half years and involves around 25 academics, postdoctoral researchers and postgraduate students, along with key partners from industry and internationally invited experts.

Tracking damage

The project aims to measure the evolution of graphite damage and microstructure, identify the controlling atomic and mesoscopic processes, and deduce, for the first time, quantitative and consistent relationships between damage-induced physical property changes, and dose, temperature and time for crystallites and their interfaces.

With experimental measures of structure and porosity, these relationships will be pooled to inform a finite element (FE) model of graphite ‘bricks’ under likely operating conditions, demonstrating the feasibility of a component-level engineering design tool. The consortium will investigate irradiation-induced dimensional change, creep and other property changes in graphite as a function of irradiation dose, time and stress, using complementary techniques, from nanoscopic to macroscopic length scales.

At the nanoscale, microscopy will be undertaken on various fast neutron-damaged reactor graphites to reveal and understand structural changes. Complementary transmission electron microscopy observations of ion bombardment (see image, above) are planned at temperatures between 100-1,300K, and modelling will be used to scale these results to fast neutron damage. Day-long ion irradiation causes atom displacement damage equivalent to a full reactor lifetime.

Ab-initio calculations through density functional theory (DFT) will be
applied to model irradiation-induced defects (see image right) as well
as to simulate
spatially resolved electron energy loss (EEL) spectra from defects and
atomistic dynamical properties, such as the observed coherent inelastic
neutron scattering from polycrystals (polyCINS, see image, below)
intensities. These depend exclusively on the spatial dependence of the
vibrational modes. Thermal, mechanical and dimensional intracrystal
property changes can be elucidated from DFT defective supercells. 


Polycrystalline coherent inelastic neutron
scattering (polyCINS) from graphite, generated using ‘Scatter’ and GULP,
from a theoretical model. The main sinusoidal feature gives information
about the strength of the inter-layer forces and is expected to be
different for irradiated graphites

In detail

On the microstructural scale, neutron and X-ray diffraction, polyCINS, Raman spectroscopy and electron energy loss spectroscopy (EELS) will reveal the structural and vibrational characteristics of the crystallites (and the nanoscale C-C bond length distribution). Small angle neutron scattering (SANS) will be applied to study porosity and its variation with temperature, related to the filling of Mrozowski cracks by anisotropic expansion. The pore size distribution is also amenable to study using gas sorption, small angle X-ray scattering and X-ray tomography. The latter technique will also provide information on microstructure, like low-loss EELS and local sample density.

Dislocation theory will be applied to understand the stresses and strains of the evolving dislocation distribution, with time and irradiation, as a route to dimensional change and creep on the micrometre scale. Dislocation dynamics will be applied to reveal the interaction of dislocation movements with pores and Mrozowski cracks.

Crystallite-scale studies will focus on grain size determination using X-ray and neutron diffraction. C-C pair distributions revealed by total neutron scattering can also be correlated with EELS estimates of sp2/sp3 ratio. Since crystallites interact through boundaries that may be more or less diffuse, intercrystal properties estimated from grain boundary DFT modelling simulations will be used to quantify stress and strain transmission across boundaries (including possible slip), and to study intercrystalline porosity and homogenise properties of polycrystal elements for the thermal and mechanical properties of FE elements as a function of dose.

PolyCINS modelling of buckling, interstitial loops, ‘ruck and tuck’ defects (see main image, top), dislocations and (partial) stacking faults in graphite will be compared with experiment. Finite element modelling will be applied to determine heat flow, heat capacity, T(t) relations, volume change, stress change and likely positions of fracture for the macro elements.

Macroscopic (component) FE modelling will exploit a continuum damage function for the amount of damage induced over a specific history. This accumulation will depend upon stress history, porosity and irradiation level. The representational capacity of the basic damage function will be analysed and improved by comparison with available empirical results, which reveal relevant microstructural features, such as thermal analysis, porosity distributions and irradiation-produced strain fields from neutron strain scanning.

A ‘unified’, multi-scale, continuum damage mechanics-based material behaviour model will then be developed, in conjunction with a discrete crack approach, to represent the fracture behaviour of graphite. The prime feature will be the flow rule (strain rate) function. The damage functions include inter alia, the time dependent micro-structural variation, the porosity and the environment, including creep, oxidation and irradiation.

For the first time, multi-scale modelling will offer a foundation for understanding, from first principles calculations, through atomistic simulations, to mesoscopic and macroscopic FE models. The process of homogenisation of detailed atomic and mesoscopic structural elements into an effective, aggregated medium will be crucial and represent a substantial innovation in multiscale modelling, as a by-product of this work. In the absence of sufficient empirical data, this approach will be essential in understanding the structural integrity of graphite components as they are used in current and future nuclear reactors.

Further information

Dr Aidan Westwood, Advanced Carbon Materials Research Group, Institute for Materials Research, School of Process, Environmental & Materials Engineering, University of Leeds, Leeds, LS2 9JT, UK.
Tel: +44 (0)113 343 2555. Email:

Further details of the consortium project may be found at