Bullet proof - ceramic body armour impact prediction
Proving that a material can withstand the force of a bullet is one thing. Predicting it can is quite another. Jon Binner of Loughborough University, Richard Todd of the University of Oxford, and Peter Brown of Dstl, all UK, explain how a collaborative research project is helping to increase understanding of ceramic materials for armour.
It is often said that where armour is concerned, a material’s performance can’t be predicted, only demonstrated. This is certainly the case for ceramic armours, where development of new materials for personnel and vehicle armour is very slow and expensive, owing to the tremendous amount of research needed with little certainty of success until actual ballistic test results are known. Challenges include the often fragmented nature of the specimens left behind after impact, and the need to consider the whole system in the dynamic failure process hinders the comparison of different material specimens from different laboratories.
To address the various challenges facing the development of ceramic armour in the UK, the EPSRC and the Defence, Science and Technology Laboratory (Dstl) have jointly funded the project Understanding and Improving Ceramic Armour Materials (UNICAM). Now nearing its conclusion, the project takes a holistic approach to developing an understanding of the high strain-rate performance of ceramics.
Ceramics in armour
For many years, the high hardness and low density of engineering ceramics such as alumina and silicon carbide have seen them widely used in personnel and vehicle armour. However, the brittle nature of these materials reduces their ability to withstand multiple ballistic hits. This commonly results in long radial and ring cracks, along with microcracks and comminution beneath the impact region. Resistance to such cracking and failure in ceramics depends on the mechanical and physical properties of the material in question, as well as specific microstructural features, but the details of this are far from well established – particularly at the extremely high strain-rates caused by ballistic impacts. This begs the question, what are the material characteristics of the ideal armour ceramic?
Modelling is playing a dual role in addressing this question, enabling:
• the design and interpretation of experiments aimed at understanding a material’s behaviour on a macroscopic scale
• insight into the role of micromechanisms in determining ballistic performance
Performing high strain-rate testing on poorly characterised, non-reproducible materials yields data that is of little use. The research team is instead determining the properties of a material prior to testing, and then inputting this data into the model. The model is developed and validated by comparing predicted and measured results – those of instrumented low and high strain-rate laboratory tests and actual ballistic tests, with the results of careful post-mortem characterisation.
The approach has been used on a number of different ceramics with a range of characteristics, yielding comparative information on the microstructural and mechanical characteristics of ceramics deemed successful in terms of ballistic performance. This method is giving scientists a greater understanding of high strain-rate performance in these materials, allowing the design of enhanced ceramic armour systems. One study of particular note is the dynamic response of a variety of ceramics. The approach is also helping the research team to pinpoint the precise failure mechanisms and material properties that are directly relevant to the development of a verifiable, multi-scale ceramic material model. The team has so far shown that it is possible to reliably produce a range of ceramic materials with controlled and quantified microstructures and properties. This is a crucial step forward and one that few researchers have previously been able to address at this level of detail.
Re-writing the book
The work has also shown that much of the current literature data on the high strain-rate compressive properties of ceramics obtained by the Split Hopkinson Pressure Bar (SHPB) test – a commonly used technique for testing the dynamic stress-strain response of materials – is most likely suspect, and difficult to interpret at best. One reason for this thinking has come through the project’s use of highspeed photography, which has shown that anvils used in several previous studies probably failed before the specimen – as such, the results would mainly have reflected the properties of the anvil material rather than the ceramic. But when using transparent specimens, the researchers found that the failure of the ceramic usually spreads from a localised region, meaning the shape of the nominal stress-strain curve cannot be interpreted in terms of uniform behaviour.
Furthermore, the load–displacement results are strongly influenced by the experimental conditions, such as bar and specimen dimensions, and the shape of the pressure pulse produced by the test equipment. An example is linear elastic loading, which under certain conditions cannot be established in the initial stages, hence there is no simple interpretation of the results. Often in the past, the importance of such factors to the validity and meaning of test results has not been appreciated and, therefore, has gone unreported. But through careful experimental design, the UNICAM team has developed a much more reliable method of testing ceramics in compression at high strain rates that yields high quality data, which can later be interpreted with the aid of the models developed.
Not so brittle
The research findings, summarised opposite, challenge the perceived wisdom that armour ceramics mainly behave in an elastic/brittle manner. When the team performed detailed microstructural and fractographic analysis of ceramics at different strain rates, using small-scale impact tests that involved firing tungsten carbide projectiles at small specimens using a gas gun, the results showed that plasticity is a far more prevalent deformation and failure mechanism than was previously envisaged. These principles have been borne out in postmortem examination of ceramics used in full-scale ballistic tests and have strongly influenced the evolution of the ceramic material model.
A new multi-scale constitutive model for ceramics is being developed in an attempt to simulate the deformation and strain localisation that leads to quasi-brittle fracture. This relies on observations on the microstructural-scale of the various ceramics being assessed, while ensuring the conservation of energy and momentum essential to understanding the thermo-mechanical response of ceramic materials to dynamic loading. At the smallest length scales, the model is validated against micromechanical tests at the microstructural scale, while the dynamic indentation tests described above and their static equivalents provide experimental comparisons for the longer length scales.
Raman spectroscopy was used to measure the fraction of zirconia phase transformation in the nano zirconia toughened alumina (nZTA) samples after testing, and the residual stress and dislocation density in the alumina grains were quantitatively measured using Cr3+/Al2O3 fluorescence spectroscopy. The results indicated that the occurrence of the zirconia phase transformation reduced the residual stress and dislocation densities in the nZTA samples, resulting in less damage, less plastic deformation and shorter crack propagation.
From a UK defence perspective, it is of significant strategic importance to have a greater understanding of, and an ability to address, the issues surrounding armour materials. The interdisciplinary nature of the UNICAM team is equally important, and has allowed the engineering relevance of the materials knowledge generated during the project to be captured far more effectively than has traditionally been the case.
Some of the most important outcomes of the project are the development of a multi-scale ceramic material model that can be implemented via widely used, commercial finite-element software, as well as the use of micro-cantilever beam testing of armour ceramics to evaluate, refine and validate this model at higher strain rates. Detailed high strain-rate experimentation conducted on transparent and opaque armour ceramics is of equal note, and has revealed both the limitations of data previously thought reliable and new methodologies for generating high-quality data suitable for modelling.
These and other UNICAM achievements are providing the UK with a new ceramic armour testing and modelling capability that looks set to make a meaningful contribution to the design and performance of future protective systems.
This is one of the less expensive ceramic armour materials and is known to display a range of performance based on purity and grain size. Different alumina ceramics were studied with purities varying from 96–99% and mean grain sizes from 1.8µm– 8µm. A detailed analysis of the fragments generated from around impact sites in high-purity alumina revealed three different types of fracture surface, each closely linked to fragment size. Plastically deformed surfaces and transgranular fracture were dominant for fragments smaller than around 1,250µm, while larger fragments displayed intergranular fracture. Dislocations were observed to be the dominant mechanism of plastic deformation at high projectile velocities (around 800ms-1) while twins dominated at lower velocities (around 100ms-1).
Three different grades were studied alongside a detailed fragment analysis. Hexoloy SA yielded a high fraction of <50µm fragments that were created through transgranular fracture when subjected to ballistic impact at 820ms-1. The flaw population is a key input for numerical modelling of ceramics under dynamic compression–loading conditions. Direct measurement of surface flaw sizes for two grades of SiC with uniform and bimodal microstructures (SiC-D and SiC-E respectively) indicated that the SiC-E had a broader range of flaws, the largest measuring around 40µm), while the SiC-D had uniform flaws of less than 20µm. An acoustic emission study showed that fewer, but larger, cracks formed on loading the SiC-D, although these began forming at lower loads in the SiC-E. Based on this, better impact resistance was expected from the SiC-E, a result that was corroborated by high strain-rate gas gun impact tests. Zirconia toughened a lumina (Z TA)
The presence of zirconia can increase the toughness of alumina by a factor of up to four, for only a relatively small penalty of increased density and higher material costs. Previous studies have shown that the benefits are particularly notable at higher impact velocities, reducing the depth of penetration compared to alumina. In this project, the focus was on additions of nanostructured zirconia, and considerable effort was devoted to learning how to produce genuinely nanostructured composites. Proportions of 10%, 15% and 20% nano ZTA with 1.5mol yttria stabiliser were produced, with densities of around 99% of theoretical and a range of grain sizes. The nano zirconia additions appeared to influence the fracture behaviour of the material. While pure alumina undergoes extensive plastic deformation, the alumina grains in the ZTAs experienced insignificant plastic deformation.
The alumina samples also generated a much broader fragment size distribution on impact, 10–1,200µm compared to the ZTAs, which measured less than 500µm. In terms of the depth of penetration and size of the indents formed from gas gun impacts using tungsten carbide penetrators at 130ms-1, the results clearly showed these were smaller in ZTA compared those of pure alumina. Along with the ring cracks at the impact site, a large population of microcracks was also observed. The maximum projectile speed that the nano ZTA specimen could tolerate without fracture was also higher than for alumina.
For more information, contact Jon Binner, J.Binner@lboro.ac.uk
Authors: Jon Binner, Santonu Ghosh, Shuo Huang, Bala Vaidhyanathan and Houzheng Wu – Loughborough University, UK, Peter Brown – Dstl, UK, Claire Dancer, Simone Falco, Robert Gerlach, Emilio Lopez-Lopez, Nik Petrinic and Richard Todd – University of Oxford, UK.
The authors are grateful to the EPSRC and Dstl for funding for the UNICAM project.