Modeling Defects in Transparent Ceramics to Improve Military Armor

December 1, 2009

Army Research Laboratory, Adelphi, Maryland

The dominant materials solution used for ballistic transparency protection of armored tactical platforms in commercial and military applications is low-cost glass backed by polycarbonate. Development of next-generation ceramics is critical to offering enhanced protection capability and extended service performance for future armored windows to the soldier. Among the potential ceramic materials considered for armor — sapphire, edge-form-growth sapphire, magnesium aluminate spinel, aluminium oxynitride — one was selected for the current pursuit: magnesium aluminate spinel (MgAl₂O₄).

Simulations of Damage Propagation into spinel-urethane-polycarbonate stacks 11 μs after impact, withslug moving 975 m/s with (a) 5 defect samples and (b) 9 defect samples.

Finite element modeling has progressed substantially in the ability to predict failure of materials under extreme dynamic loading conditions. One of the limitations of predictive models is lack of a complete dynamic materials properties database, which is needed for materials models for each of the materials in the simulations. In order to compensate for parameters whose dynamic values were extrapolated from their static or quasistatic properties, baseline experiments are often used to recalibrate the models.

The objective of this work was to study the effect of various shape defects, located in the interior and on the surface of spinel, on the failure of the transparent material.

Coupons for ballistic testing consisted of laminated layers of spinel bonded using Huntsman 399 polyurethane adhesive to a Bayer polycarbonate. To reduce variables, the backing layer thickness was fixed at 12.7 cm of polycarbonate. The ceramic striking material for this investigation was 11 mm. The bonding layer is typically 1 mm. Experimental samples were evaluated only to attain penetration velocity to confirm the model parameters. However, the experimental results were also used to compare the actual cracking pattern with that produced from the simulation. In addition, square cuts of 1.5 × 5 mm, and cones of 4-mm diameter and 4-mm height, were introduced into the surface of the spinel. The density of the surface defects varied and represented a 2% and 4% mass loss of the solid spinel. The internal defects represented a 4% mass loss of the solid spinel.

The ballistic behavior of a model identical to the actual target geometry, which consisted of spinel, polyurethane (PU), and polycarbonate (PC), and impacted by a surrogate projectile, was simulated using the nonlinear ANSYS/AUTODYN commercial package. The material models used were obtained from the AUTODYN library. The 2D modeling laminated target consisted of panels of spinel, polyurethane, and polycarbonate of 900 cm² cross-sectional area. The defects were filled with air at one atmospheric pressure. Due to the lack of the strength and failure material models of the spinel, these were obtained by modifying the existing at the AUTODYN materials library alumina (Al₂O₃) strength and failure model by using existing experimental ballistic data.

Results were obtained by simulating projectiles impacting the targets at the experimental velocity of 975 m/s. The PC was modeled using a shock equation of state (EOS), piecewise Johnson-Cook (JC) strength model, and a plastic strain failure criterion. The urethane was modeled using a linear EOS and a principle stress failure criterion. The projectile steel was modeled using a shock EOS and a JC strength model. The spinel, however, due to lack of an existing material model, was modeled using a recalibrated form of the existing AUTODYN library Al2O3 material model, produced from existing experimental data and consequently validated many times by predicting within 3% results of new target design, which included a polynomial EOS and Johnson-Holmquist (JH2) strength and failure models. The air was modeled using ideal gas equation of state, with no strength and failure models.

The mass loss of the internal defects corresponds to 4%. The elliptical and rectangular internal defects were oriented having their long axis and larger dimension perpendicular and parallel, respectively, to the line of impact of the projectile. Analysis of the experimental tests and the simulations includes evaluation of extent of damage, residual velocity after impact, and extent of deflection.

During the impact event, failure modes appear consistent. Therefore, the 2D model shows an effective and rapid method of producing laminate constructions for interrogation of failure criteria. The baseline results allow confident investigation of the defect models. That prediction disagreement may be attributed to the fact that 2D defects, when expanded in 3D, is not a localized 3D defect, but rather a groove. Therefore, more material is removed in a 2D simulation than the 3D respective simulation.

The surface simulations indicated that the resistance of the ceramic hard face to the penetrating projectile depends mainly on the mass ahead of the projectile.

This work was done by C.G. Fountzoulas, J.M. Sands, G.A. Gilde, and P.J. Patel of the Army Research Laboratory. For more information, download the Technical Support Package (free white paper) at www.defensetechbriefs.com/tsp under the Materials category. ARL-0083