James Marrow
My research is focussed on the degradation of structural materials and the role of microstructure. A significant proportion of this work is related to materials utilised in the nuclear industry. This work has been funded by organisations including EPSRC, Rolls- Royce, British Energy, EdF, the Health and Safety Executive (Nuclear Installations Inspectorate) and the Nuclear Decommissioning Authority. A key aspect is the investigation of fundamental mechanisms of damage accumulation using novel materials characterisation techniques. This has concentrated recently on computed X-ray tomography and strain mapping by digital image correlation.
Address: Department of Materials
University of Oxford
Address: Department of Materials
University of Oxford
less
Related Authors
InterestsView All (32)
Uploads
Papers by James Marrow
Gilsocarbon graphite. The test geometry adopted has a ‘five-point’ bending configuration, i.e. a cruciformshaped
specimen that creates a tensile biaxial stress on the surface. This allows the effect of changing the
biaxial ratio on the load-displacement and fracture characteristics to be considered. An acoustic emission
(AE) technique has been applied to monitor and identify the occurrence of acoustic events and their locations
in specimens loaded either monotonically to failure or via several progressively increasing load-unloading
cycles. It was found that the fracture path changes with biaxial ratio. AE events occurred from low load in
all loading modes, and increased progressively with the increase of applied load. The total number of AE
events and the sum of the cascade energy from the AE measurements were similar for specimens fractured
at the same load under a particular loading condition.
This work has highlighted areas where the development of image correlation methods that are optimised for analysis of discontinuities would be beneficial, for better detection of small cracks and the early development of damage against the background displacement field; improved precision in crack displacement field measurement by intelligent “masking’ or analysis algorithms and better integration with finite element software packages to make use of advanced tools for 2D and 3D deformation analysis.
This paper reviews some of this recent work on the analysis of 2D and 3D damage in engineering materials, and describes developments in quantitative analysis of defects by image correlation. The examples covered include brittle crack propagation in nuclear graphite, fatigue loading in magnesium alloys and indentation damage in brittle and ductile materials.
The activities of isotopes contained within nuclear graphite may be theoretically calculated from the trace elemental impurities present within virgin graphite material and the cross sectional areas of these elements. This combined with reactor operational conditions provides background to the isotopic inventory currently accepted. However, other isotopes may arise from impurities trapped in the porous graphite during reactor operation. These activated impurities will need to be accounted.
This paper presents microstructural and radiochemical techniques used to quantify the isotopic location and distribution within the graphite. These impurities have been characterised in terms of location and retention using high resolution techniques such as Scanning Electron Microscopy, Raman, micro X-ray Tomography and Energy Dispersive X-ray Spectroscopy.
This presentation summarises progress in work to observe deformation and fracture in nuclear graphite, using synchrotron X-ray tomography and digital volume correlation to measure three-dimensional strain fields. High precision synchrotron diffraction studies on strained samples and the fracture process zone of propagating cracks provide new insights into the inelastic deformation of graphite. Microcracked fracture process zones are common to quasi-brittle materials as diverse as high toughness monolithic ceramics, polymeric and natural biological composites, geological minerals and even volcanic structures. Experimental methods that support the study and modeling of damage development are thus important to a wide range of problems, beyond nuclear graphite.
The objective of this work is to better understand how the microstructure of a coarse grained polygranular graphite accommodates applied strain, and the effect of this applied strain on its mechanical properties. The relation between applied strain and residual inelastic deformation, and the difference in behaviour under tension and compression, are of particular interest. To study this, it is necessary to be able to observe, in situ, the relationship between the applied strains, the total strains in the material’s microstructure and the elastic strains in the crystals. The effect of radiolytic oxidation, which occurs progressively in the UK’s ageing nuclear reactors, on notch strength and its variability is an important element of methodologies to assess the probable development of cracking from stress concentrators such as keyway roots. At present, notch strength must be inferred from flexural tests on smooth specimens, with very limited data on irradiated graphite.
This presentation summarises progress in work to observe deformation and fracture in nuclear graphite, using synchrotron X-ray tomography and digital volume correlation to measure three-dimensional strain fields. High precision synchrotron diffraction studies on strained samples and the fracture process zone of propagating cracks provide new insights into the inelastic deformation of non-irradiated graphite, with implications for the behaviour of irradiated graphite and the effects of specimen size and stress gradients. Finally, novel modeling techniques are being developed to evaluate the sensitivity of small specimen fracture tests to microstructure.
As a surface characterization technique, indentation cannot provide information about the deformation processes and damage induced within the material. To gain an understanding about the deformation beneath the indenter, a combined approach of high-resolution synchrotron-based X-ray computed tomography (CT) and three-dimensional digital image correlation ('Digital Volume Correlation' or DVC) was employed. Firstly, the microstructure of the material is captured with CT before and after indentation, obtained in situ under load and revealing the microstructural inhomogenities of the material. Subsequently, this information is used to calculate the 3D displacement field and estimate the strain field beneath the indenter. The resulting strain data may then be used to validate models for deformation and fracture behaviorr.
This paper demonstrates the application of CT/ DVC to two different types of material: a ductile aluminium-silicon carbide composite (Al-SiC) and brittle alumina (Al2O3). In Al-SiC, the measured displacements for Hertzian indentation are in good agreement with elastic-plastic finite element simulations. In Al2O3, radial cracking is observed beneath a Vickers indentation and the crack opening displacements are measured with sub-voxel resolution.
Introduction
Compact tension (CT) and chevron notch (CN) specimens are common choices for measurement of the fracture resistance of brittle and quasi-brittle materials, as both geometries can exhibit stable crack propagation under displacement-controlled loading. The standard CT test originates in the testing of ductile materials, such as structural alloys [1] whereas methods using the short bar chevron notch specimen originated in the testing of brittle materials, such as rocks and minerals [2]. Consequently, there are differences between these approaches in the parameters measured and also the potential interpretation of results. In particular, the method for separating elastic and inelastic effects through loading and unloading cycles, which is common in elastic-plastic materials, can be misinterpreted in quasi-brittle materials; nonetheless, this method has gained a degree of popularity in research into the fracture behaviour of artificial graphite [3].
A review of similar tests on other quasi-brittle materials and an examination of the underlying theory, based on the Griffith’s energy approach, shows that intermediate unloading and reloading cycles in these materials are unnecessary for quantifying fracture resistance; complex analyses and interpretations based on measurements made during the intermediate cycles can therefore be misleading. To support this claim and to investigate the physical processes during the unloading and reloading cycles that cause apparent inelastic deformations in quasi-brittle materials, results are presented from a short bar chevron notch specimen fracture test of artificial graphite, performed at the Diamond Light Source facility and analysed by X-ray computed tomography and digital volume correlation.
Experiment
The specimen was loaded by driving in a steel wedge, opening the faces of the notch to initiate and propagate a stable crack. A series of high resolution X-ray computed tomography (XCT) three-dimensional images were obtained, in situ, at intervals as the specimen was loaded, unloaded and reloaded using the Diamond Light Source and also laboratory X-rays. Finally, the wake of the propagated crack was removed by electro-discharge machining and the cut specimen was tomographed again. The datasets were analysed by digital volume correlation (DVC) to map the internal displacements in three-dimensions. Consequently, the geometry and surface separations of the crack were measured with high precision. In particular, crack opening displacement profiles in the loaded and cut specimens were compared.
Analysis and Conclusions
These data shows there is an enhanced crack opening ahead of the physical crack tip. This confirms previous observations by lower resolution laboratory tomography [4], which also found that this zone remained measurable on removal of the wedge. The crack tip position observed directly by XCT is an underestimate due to limited resolution of crack openings in attenuation contrast. The observed fracture process zone might be interpreted as having a significant component of inelastic (i.e. plastic) deformation. However, the virtually complete relaxation of the fracture process zone after removal of the crack wake shows its deformation is essentially elastic; analysis of fracture experiments that assume energy-consuming processes due to plastic deformation in quasi-brittle materials may therefore be inappropriate.
The presentation concludes by providing some guidance for fracture testing of quasi-brittle materials, including precautions to be taken when selecting appropriate loading apparatus, designing test specimens, specifying the parameters than need to be measured and how the measurements should be analysed and interpreted.
References
[1] R6. Assessment of the Integrity of Structures Containing Defects: British Energy; 2001.
[2] Ouchterlony F. International society for rock mechanics commission on testing methods - Suggested methods for determining the fracture toughness of rock. INT J ROCK MECH MINING SCI. 1988;25:71-96.
[3] Sakai M, Urashima K, Inagaki M. Energy principle of elastic-plastic fracture and its application to the fracture mechanics of a polycrystalline graphite. J AM CERAM SOC. 1983;66:868-74.
[4] Mostafavi M, McDonald SA, Mummery PM, Marrow TJ. Observation and quantification of three-dimensional crack propagation in poly-granular graphite. ENG FRACT MECH. 2012; http://dx.doi.org/10.1016/j.engfracmech.2012.11.023.
Acknowledgements
Awards of beam time at the Diamond Light Source (experiment EE7119) and the Manchester X-ray Facility are gratefully acknowledged.
The foreseen operating conditions of the Generation IV concepts will place significant demands on their structural materials. These demands are far more stringent than those for existing nuclear plant, and there will be a requirement for design lives in excess of 60 years. Their material requirements have been well reviewed by numerous papers and candidate materials for critical component have been identified, drawing on experience of fast reactors prototypes operated in the latter part of the 20th century, some in the UK.
The aim of this paper is to summarise recent developments within Europe in support of the design of Generation IV plant, to highlight some issues concerning UK involvement in the structural integrity aspects of the European fast neutron reactor research programme, and lessons learned from the UK's 40 years of experience of fast reactor operation.
We could then grow an intergranular stress corrosion crack in the sample, and make in situ tomographic observations of how the crack interacted with the microstructure, revealing which boundary types tend to resist crack propagation."
Metallurgy at The University of Manchester explores solutions"
This paper briefly describes some of the research programmes being carried out by the NGRG at Manchester.
Cracking was initiated beneath saturated MgCl2 droplets in an atmospheric environment at 80°C and relative humidity of 30-33%. As-received and 10% cold rolled samples (with two orientations transverse and longitudinal to the loading direction) were subjected to an applied strain of 0.03 under displacement controlled tests. Regular optical observations were recorded of the droplets and their surrounding area. DIC analyses used the differentiation of the displacement fields to obtain the apparent surface strains used to detect crack initiation and propagation, and to measure crack opening displacements.
It was found that DIC was efficiently observed the strain developments and the displacements in observed surfaces outside of the droplets but it could not identify or quantify the initiation of the cracks inside the droplets because of the mobility of the salt film and the high amount of the corrosion products formed which obscure the vision under the droplets. In addition, results showed that early stage microcracks were initiated in α phase and α/γ interfaces and propagated preferentially in the ferrite phase. Also, SCC initiation and propagation was accelerated by cold rolling and the grains orientations were of major effects on the retardation of crack propagation which was more severe in the transverse rolling direction. Also, there was no relation established between the strain level and the density of pitting in either phase.
A generic mechanistic model for short fatigue crack propagation proposed by Navarro and Rios (N-R model) was implemented to assess its suitability for predicting the fatigue behaviour of specimens with various controlled surface conditions, obtained by machining. The surface/material properties required to implement this model were obtained by electron backscatter diffraction (EBSD), surface profilometry, hardness testing and X-ray diffraction residual stress measurement. The fatigue limits were determined using rotating-bending by means of the staircase method.
The fatigue limits predicted by the N-R fatigue model were compared with the results of the fatigue tests. There was no agreement between the prediction and observations, indicating that the original form of the N-R model is not appropriate for austenitic stainless steels.
In AISI 304L, the surface residual stresses are the dominant parameter, allowing prediction of the effects of machining on fatigue resistance while, the surface roughness developed by machining has no significant effect. In AISI 316L, the effect of surface roughness is found to be negligible, with a weaker effect of surface residual stress than has been observed for AISI 304L.
Crack nuclei in run-out (>107 cycles) fatigue tests were observed to arrest at twins and martensite packets, developed by fatigue in AISI 316L and AISI 304L, respectively. Good agreement with experiments was achieved by using a modification to the fatigue model, which takes account of the observed effect of the plastic deformation on the microstructure.
In this PhD thesis, the effects of time, applied stress and microstructure on populations of short crack nuclei have been investigated in sensitised type 304 austenitic stainless steel, tested under static load in an acidified potassium tetrathionate (K2S4O6) environment. Statistical evaluation, using the Gumbel extreme value distributions enables analysis of the growth rate of the population of short crack nuclei. This methodology has been developed, in order to quantitatively evaluate the influence of grain boundary control on crack development. These investigations showed an increase in the expected crack length with increasing time and grain size. Although the crack length tends to increase with stress, the effect is not strong. The grain boundary controlled microstructures exhibited significantly higher resistance to intergranular crack propagation. Direct observations of intergranular crack initiation and propagation, using digital image correlation (DIC) along with electron back scatter diffraction (EBSD), in various microstructures has been used to study the crack nucleation sites and crack interactions with grain boundaries of different characteristics. The effect of microstructural modification on crack growth kinetics has also been investigated. A significantly longer incubation period for crack initiation and lower crack growth rate were observed in thermo-mechanically treated microstructure.
New methods have been developed to assess the clustering characteristics of grain boundaries of particular properties. The network properties of boundaries classified by EBSD data have been compared with the network of corroded grain boundaries in electro-chemically tested samples. Image analyses (IA) was employed to evaluate the geometrical properties of susceptible boundaries clusters in a range of microstructures produced by sequential thermo-mechanical processing (TMP).
DL-EPR testing method of sensitisation assessment has been augmented by large area image analysis (IA) assessments of optical images to measure the dimensions and connectivity of the attacked grain boundary network. This approach determines the degree of sensitisation of the susceptible grain boundaries in the microstructure, and is used to explain IGSCC behaviour. A new method of degree of sensitisation determination is proposed, based on normalisation by a cluster parameter for the network of susceptible grain boundaries.
Six different surface machining operations on two different sample geometries, varied by milling cut depth and feed rate, were applied to cylindrical and rectangular-prismed 316Ti steel samples cut from plate. Two of the faces from rectangular samples were left ’as-received’, i.e. these surfaces corresponded to the upper and lower faces of the original plate the samples were cut from. One face was ground using a grinding wheel such that no roughness was visible to the naked eye. The remaining (end) faces were coarse-cut using a bandsaw. Sets of these samples (containing one of each of the milled profiles) were boiled in a magnesium chloride solution for two weeks, a similar (refer- ence) set of six samples were stress relieved by annealing in argon at 1100oC for 30 minutes. Subsequently, the resulting cracks and pits were examined by means of optical microscopy and scanning electron microscopy. Also, x-ray diffraction (XRD) residual stress characterisation work was performed on the samples, including electropolished ones (in order to characterise residual stress with depth). Finally, optical profilometry was employed to acquire single line and maps of the roughness.
The results indicate that machined samples retain significant and relatively high tensile stresses that are present over several 10’s of microns into the material from the surfaces. Beneath the tensile layer, a compressive sub-layer was found. SCC cracks were generally observed to grow into the tensile layer and were then deflected close to the compressive layer. However, many cracks were able to extend into the compressive layer, growing to over 100 microns in length. The dominant feature affecting cracking appeared to be the residual tensile layer; the surface profile had no clear affect.
Grain Boundary Engineering has been used to produce austenitic metals with a high increased fraction of ‘special’ boundaries that have increased resistance to SCC. Crack bridging at these ‘special’ boundaries has been shown in thermally sensitised stainless steels and modelling work predicts that crack bridging slows down the propagation of Intergranular Stress Corrosion Cracking. ‘Special’ boundaries have been defined as those with a low Sigma Coincident Site Lattice (CSL) structure but work has also shown that grain boundaries whose plane lies on a low index plane of the adjoining grains can also exhibit special properties.
An austenitic stainless steel was irradiated with protons to produce a microstructure similar to that obtained in nuclear reactors. Autoclave testing to initiate SCC was performed and grain boundary elemental composition measured. Evidence of crack bridging was observed in the autoclave sample and differences in segregation behaviour were observed. In particular it was noted that while segregation occurred at some random boundaries, no segregation was found at a random boundary which had a plane that lay on the {221} planes of the adjacent grains.
This PhD thesis investigates the effect of microstructure and stress on intergranular stress corrosion cracking in Type 302 / Type 304 austenitic stainless steels. High-resolution X-ray tomography has been successfully applied to examine, for the first time in 3-dimensions, in- situ, the interaction between microstructure and crack propagation. The development and subsequent failure of crack bridging ligaments has been observed and correlated with regions of ductile tearing persistent on the fracture surface. These ductile regions were consistent with the morphology of low-energy, twin-type grain boundaries, and are believed to possess the capability of shielding the crack tip.
Following this observation, a new grain bridging model has been developed, in order to quantify the effect of static stress and crack bridging on the maximum likely crack length. The model was compared and evaluated with in the literature available percolation-like models.
Intergranular stress corrosion tests in tetrathionate solutions have been designed and carried out to validate the new model. The assessment comprised,
(i) a thorough examination of the microstructure and analysis parameters employed,
(ii) the determination of the degree of sensitisation with subsequent crack path
investigations,
(iii) the identification of a suitable test system with associated grain boundary
susceptibility criteria,
(iv) the application of Grain Boundary Engineering (GBE) for microstructure control,
(v) statistical crack length assessments of calibrated IGSCC test specimens.
The results of these tests showed that the new model successfully predicts the magnitude of stress and the effect of grain boundary engineering on the maximum crack lengths.
This chapter provides an overview of the mechanical properties of materials and describes fracture mechanics principles, which are used widely to assess the severity of such defects in engineering structures.
The application of high-resolution synchrotron X-ray tomography to two in-situ experiments is described. Localised corrosion and intergranular cracking in sensitised 5083 aluminium alloy has been studied. These results show the progressive development, transition, and coalescence of two forms of damage within the bulk of the sample. In-situ observations of intergranular cracks in sensitised 302 stainless steel have also been obtained. These provide evidence for crack bridging ligaments, caused by the high resistance of special grain boundaries.
Further applications of high resolution X-ray tomography are described, such as in-situ studies of pitting and the transition from pitting to cracking in aluminium alloys and stainless steels, including the effects of near-surface residual stress.