A COMPARISON OF DAMAGE IN GLASS AND CERAMIC TARGETS

Brady Aydelotte, Phillip Jannotti, Mark Andrews and Brian Schuster

US Army Research Laboratory
321 Colleran Road RDRL-WML-H
Aberdeen Proving Ground, MD 21005

Keywords: Cone Crack, Impact, Indentation, Ceramic, Damage

Abstract

The high strength and low density of many ceramics and the transparency and low cost of glasses make them potentially useful candidates for many applications, including armor. Both ceramics and glasses are very brittle and they can go through a complex fracture process when impacted. Ballistic impacts on ceramics produce different types of damage including varying levels of comminution, cone cracking, and radial cracking. Sphere impacts on brittle targets are a useful way to study the evolution of ceramic damage. We performed sphere impact experiments on fused silica glass targets. This work is compared with X-ray computed tomography scans of recovered samples generated from previous work on boron carbide [B. Aydelotte and B. Schuster, in Dynamic Behavior of Materials, Volume 1, Springer International Publishing, 2016, pp. 19-23.]. The damage morphologies of the sphere impacted fused silica and boron carbide targets are compared. We found that cone cracks in boron carbide and fused silica have the same general shape in response to temporally and spatially changing loads and appear to exhibit cone rotation that is related to the component of the velocity tangential to the target surface. Cone angles in boron carbide were larger when measured in a plane which contains the shot-line vector and intersects the apex of the fracture conoid. Measurements of the fracture cone angle on a plane perpendicular to the plane containing the shot-line vector were consistently smaller for the same velocity. Measurements of cone angles in fused silica exhibited no such trends.

Introduction

Damage due to normal impact on ceramics and glasses has received considerable attention (see for example [1, 2, 3, 4, 5, 6, 7]). Earlier research has demonstrated that the formation of cone cracks tends to happen at characteristic cone angles and that increasing impact velocity tends to lead to decreases in the cone angle (see for example [1]).

Relatively less published literature exists on the oblique impact of ceramics. Salman et al. [8] studied the effect of oblique impacts on alumina particles, finding that oblique impacts resulting in mostly similar forms of damage with lower probability of failure, likely due to the reduced normal velocity component.. Much of the oblique impact literature is similar to that published by Sandanandan and Hethrington [9] and Hohler et al. [10]. Sadanandan and Hethrington[9] and Hohler et al. [10] focused primarily on studying various simplified armor packages impacted at different obliquities and their resistance to single impacts as measured by metrics like V50. There was little information on characterizing impact-induced damage. Fawaz et al. [11] presented results of modeling oblique and normal impacts on ceramic targets. They reported the ability to accurately model conoid formation, though no simulation images of fracture conoids from normal and oblique impacts were presented. Some sliding indentation studies on polycrystalline ceramic materials have been published [12, 13, 14].

Figure 1: Diagrams of our terminology for the various angles associated with cone cracking in normal and oblique impacts and how they are measured. (a) Schematic of a cone crack resulting from an oblique impact and the leading and trailing edge angles as viewed in a plane which contains the shot-line and intersects the apex of the cone crack. The terms leading and trailing edge angles are used when referring to a cone crack, resulting from an oblique impact, as viewed in the plane containing the shot-line as shown here. (b) Schematic of symmetric cone crack with left and right edge angles indicated. This schematic also represents a cone crack resulting from an oblique impact as viewed from a plane perpendicular to the one defined in (a).

Some studies of damage in oblique glass targets have also been published [15, 16, 17]. For the purpose of studying the phenomenology of damage in the glass targets, the work published by Chaudhri and Liangyi [15] and Forde et al. [17] are the most useful. Forde et al. [17] reported a series of normal and oblique impacts on borosilicate glass targets by mild steel rods and published some high speed camera images of damage formation in the borosilicate glass targets. They were able to measure cone angles for some of the impacts at normal incidence, but provided little quantitative information about oblique impacts.

Chaudhri and Liangyi [15] conducted sphere impact studies on glass targets at various obliquities and filmed the damage evolution using a high speed camera. They observed that cone cracks, which form as a result of oblique impacts, form such that the leading edge of the cone crack tilts away from the impact surface toward the bottom surface and the trailing edge of the cone crack tilts toward the impact surface as a result of the changing position of the center of pressure as the projectile translates down the target. Chaudhri and Liangyi [15] deemed the effect of friction on the cone crack orientation insignificant by visualizing the stress field with circularly polarized light. The resulting fringes are symmetric about the surface normal suggesting that minimal shear stress is transmitted across the interface. However, this is early in the impact when surface damage is very minimal.

Table 1: Selected mechanical properties of hot-pressed B4C and fused silica. The B4C values are taken from from Vargas-Gonzalez et al. [19]. The fused silica properties are drawn from various sources.

Aydelotte and Schuster [18] conducted normal and oblique impact experiments to study the damage morphology in polycrystalline ceramics. In this paper, they compared the cone cracking induced by normal and oblique impacts from tungsten carbide spheres on hot-pressed boron carbide (PAD B4C) targets. They observed that cone cracks which form in ceramics as a result of oblique impacts have concave down curvature on the leading edge of the cone crack and concave up curvature on the trailing edge; the same mechanisms at work as with Chaudhri and Liangyi [15].

Experimental Setup

In this paper, the terminology for the various cone crack related angles and how the angles will be measured are shown in Fig. 1a and 1b. The rotation of the cone crack is equal to one half of the difference between the edge angles. Where the top surface is not sufficiently intact to provide a measurement, the appropriate angle with the bottom surface is measured.

The experimental setup for the ceramic cylinders is described in some detail in [18] and it will be repeated here briefly. B4C cylinders 38.1 mm (1.5 in) diameter x 25.4 mm (1.0 in) in length were impacted with 6.35 mm diameter (0.25 in) tungsten carbide-6% cobalt (WC) spheres. Some selected properties of pressure assisted densification (PAD) formed boron carbide are shown in Table 1. Impact experiments were conducted at three different obliquities: 0°, 30°, and 60°. The spheres were fired out of a 0.30 caliber smooth-bore laboratory powder gun using discarding sabots at velocities between 200 and 500 m/s. The experimental setup is diagrammed in Fig.2.

In Aydelotte and Schuster [18], flash X-ray systems were used to view the cone cracks and make measurements. In some cases the flash X-rays were found to have some alignment issues so only measurements derived from XCT will be discussed here.

The experiments conducted on glass cylinders here are very similar to those conducted previously. Fused silica cylinders were procured from McMaster-Carr (Princeton, NJ). The cylinders were 38.1 mm (1.5 in) diameter x 19.05 mm (0.75 in) in length; some selected properties of fused silica are shown in Table 1. The cylinders were transparent on the ends and had a ground finish on the cylinder sides. The fused silica cylinders were impacted using either steel or borosilicate glass spheres. The steel spheres generally did so much damage to the cylinders that it was difficult to recover the samples for further analysis. Both the steel and glass spheres were 6.35 mm diameter (0.25 in). The steel spheres were bearing balls made from hardened 52100 steel. The steel spheres had an average mass of 1.044 ± 0.001 g. The borosilicate glass spheres were purchased from Winsted Precision Ball. Their average mass was 0.328 ± 0.001 g. The spheres were fired out of a .30 caliber smooth-bore laboratory powder gun using full-caliber plastic obturators at velocities between 200 and 500 m/s. The obturators generally yielded better accuracy than discarding sabots, but at the cost of having the plastic obturators strike the targets.

Figure 2: Schematic of the impact experiments showing the obliquity measurement.

XCT scans were performed on the recovered samples. The XCT machine used for scanning all...