Effect of replacement strike-face material on the ballistic performance of multi-ply soft armor targets

Abstract

In this study, the impact-face material of a multi-ply soft armor system was varied to different ratios and tested for the effects on the ballistic performance. It is known that the first few layers of multi-ply soft armor material typically fail inelastically near the system ballistic limit and can be replaced with a “sacrificial” material with other more desirable properties. Previous studies have determined that the ballistic performance of these hybrid systems is largely dependent on the amount of high-performance backing material. However, the extent to which the high-performance fabric can be replaced has yet to be fully quantified and examined. Materials of different properties, namely stainless steel mesh, Makrolon® polycarbonate sheets, and cotton, were used as replacement frontal material for 840 d Twaron® panels, and the hybrid panels were impacted by O1 tool steel right-circular cylinder projectiles fired using a single-stage smooth-bore gas gun. Results show that the ballistic performance is maintained up to a frontal material ratio of about 40%, and off-axis material properties play a role in energy dissipation.

Keywords

ballistic limit soft armor high-performance fabric shear strength strike face

High-performance ballistic fibers are used in a multitude of different fields due to their unparalleled strength-to-weight ratios. These fibers have been utilized in commercial and industrial applications such as tow ropes, fishing nets and lines, and turbine fragmentation containment systems in aircraft. More importantly, these fibers have proven to be extremely effective against ballistic threats, and a wide range of experimentation has been performed on these materials. However, the underlying principles by which these fibers are able to stop ballistic threats with such efficiency are still not well and fully understood. The ballistic limit is a metric by which these fabric armor systems are evaluated and compared, and is most commonly defined using the V50, the velocity at which a projectile has a statistical 50% chance of penetrating the target system.

At striking velocities below or near the ballistic limit, the major energy-dissipation mechanisms involve fiber axial strain energy and kinetic energy, as well as through-thickness kinetic energy when the fabric system is moved in the out-of-plane direction.^1–3 Cunniff,^3–5 in his various seminal studies on soft armor impact, has described this portion of energy absorption as “elastic.” The energy absorbed by the material per unit mass is carried away from the impact site at the speed of sound in the material. Past the ballistic limit, these elastic strain energy mechanisms start to become less significant, while localized (henceforth described as “inelastic”) failure modes start to take over. The energy does not get transferred rapidly enough away from the site of impact, and is further prevented from dissipating due to localized damage to the material.

The ballistic limit of a multi-ply fabric system can be predicted given the target system/projectile areal density ratio.^3–5 The relationship between the ballistic performance and the areal density ratio has been empirically verified for a broad range of ballistic materials to a rather high degree of accuracy, and thus provides an excellent design basis for soft armor structures. An analytical membrane model was developed by Phoenix and Porwal⁶ to explain this relationship. The areal density ratio parameter is given by

η = A_{d} A_{p} / m_{p}

(1)

where A_d is the target system areal density, A_p is the projectile’s presented area, and m_p is the projectile mass. In addition, the ballistic response of such a multi-ply target system is typically decoupled through the system’s layers.⁴ In other words, “the first few plies of a multi-ply armor system respond as if they were free-standing, not backed by the remaining plies of the system.”⁴ At the V50 velocity for the whole target system, the velocity is sufficiently high with respect to the frontal layers for them to behave inelastically. These initial layers may then be swapped with a different material to result in a similar or possibly even improved ballistic limit, along with other desirable characteristics, such as low cost or comfort. The caveat to using this “low-cost” frontal material is that the impacting projectile should not yield a sufficiently low ballistic limit for the rear high-performance material portion of the hybrid system. In such scenarios, for the same areal density ratio η, the hybrid system will perform worse than a full high-performance fabric system.

“Shear plug” failure mode

The formation of a shear plug typically occurs for continuum targets such as ductile metals, bulk polymers, or certain stiffened composite targets.^7,8 As its name suggests, an actual plug of material is sheared off during impact on a target plate at sufficiently high velocities. In high-performance ballistic fabrics, however, the term “shear plug” is likely to be a misnomer, as there has been little to no direct evidence of shear plugs forming, even at high impact velocities. Most of the analysis in the existing literature refers more to a localized inelastic failure mode than an actual plug. Regardless, various authors have suggested an analytical shear plug-type localized deformation mechanism that occurs as a significant failure mode,^5,9–12 through either explicit analysis in their respective literature or by inference from their equations of motion. Most recently, Nguyen et al.⁸ and Hudspeth¹² derived equations for this shear-plugging mechanism to describe the through-thickness energy absorption. Equations (2)–(4) below are a rearrangement of both authors’ formulations to provide clarity

E_{shear} = \int_{0}^{t} τ_{max} ({π κ D}_{p}) t d t = \frac{1}{2} τ_{max} {π κ D}_{p} t^{2}

(2)

{KE}_{plug} = \frac{1}{2} k^{2} A_{d} A_{p} V_{r}^{2}

(3)

E_{absorbed} = \frac{1}{2} m_{p} (V_{S}^{2} - V_{r}^{2}) = E_{shear} + {KE}_{plug}

(4)

In the equations, E_shear is the amount of energy required to shear a fabric plug of mass, t is the through-panel thickness, τ_max is the shear strength in the through-thickness direction, D_p is the projectile diameter, V_s is the striking velocity, V_r is the residual velocity, and κ is a non-dimensional scaling coefficient (given as 1.6 by Nguyen et al.⁸). It is seen from equation (2) and Hudspeth¹² that this inelastic energy is minimally dependent on the impact velocity, and consequently the energy absorption fraction via this mechanism decreases with increasing striking energy. In fact, at high striking velocities way past the ballistic limit of a certain system, the “shear energy” contribution is negligible and the Cunniff derivation for inelastic energy³ is recovered. Nonetheless, near the ballistic limit, a higher shear strength material may contribute more to the overall energy absorption than a weaker material.

A set of experiments performed by Alesi¹³ in 1957 investigated the synergistic effects of frontal layers on the armor system as a whole. Specifically, Alesi tested hybrid panels of window glass/nylon fabric, A-110AT titanium alloy/nylon fabric, and polymethyl methacrylate (PMMA)/polyvinyl butyral (PVB) under impact from 0.22-caliber fragment-simulating projectiles (FSPs). The three panels exhibited synergistic effects, meaning that the hybrid system performed similar to or even better than a target panel of purely nylon fabric. However, many factors appear to be involved, such as the ratio of frontal material areal density to the whole target system areal density and target/projectile areal density ratios, as well as the amount of high-performance backing material. As a conclusion, Alesi states that at lower velocities, the difference between the hybrid and fabric target diminishes quickly and, eventually, the nylon fabric system becomes superior.

A subsequent experiment⁴ was performed by Cunniff using 2-, 4-, and 16-grain right-circular cylinder (RCC) projectiles (and, consequently, different target/projectile areal density ratios) on a Pyrex glass/Kevlar® KM2 fabric hybrid panel target system to help narrow down these factors (Figure 1). In particular, the target areal density was kept constant while changing the projectile areal density. It was found that at lower areal density ratios, that is, larger projectile areal density for the same target system, the full KM2 fabric system will outperform the Pyrex/KM2 hybrid panel. At higher areal density ratios (smaller projectile mass), the Pyrex/KM2 hybrid outperforms the full KM2 fabric system – the transition areal density ratio seems to occur at η ≈ 0.294 based on linear fits. This has been attributed to the deformation of the smaller 2- and 4-grain projectiles during impact, whereas the same mass projectiles “do not typically deform during impact onto fabric armor at this areal density.”⁴ Deformation during impact increases the overall presented area, resulting in better ballistic performance of the target. However, further examination of the recovered projectiles was not performed. Regardless, these results show the influence of material strength when replacing these frontal layers.

Figure 1.

Hybrid Pyrex/KM2 and full KM2 fabric panel ballistic performance, as performed by Cunniff.⁴

While high-performance ballistic fabric systems show excellent resistance against ballistic impact, cut- and stab-resistance is another major requirement of these body armors,¹⁴ since these damage modes remain a viable threat to users of these armor systems. The feasibility of replacing the frontal material with a more stab-resistant material as well as knowing the fraction to which these high-performance fabrics can be replaced will no doubt be of great use in designing an efficient armor system. As mentioned, the amount of high-performance fabric material that can be replaced and the effects on the ballistic performance have yet to be fully quantified. The current work therefore focuses on (a) the systematic study of varying frontal material properties and (b) the amount of frontal material that can be replaced without affecting the overall ballistic performance of the hybrid armor system.

Experimental procedure

Gas gun setup

Projectiles were shot with a single-stage smooth-bore light-gas gun with an inner bore diameter of 9.80 mm and barrel length of 3.66 m. In order to improve accuracy and reduce trajectory instability due to drag on the RCC projectiles, the target is located approximately 0.4 m from the tip of the barrel. In Figure 2, alignment was performed to ensure perpendicularity of the target panel to the shot axis. A steel safety chamber was placed behind the target panel mount to retrieve any exiting projectiles using either terry cloth rags or 10% by weight porcine skin ballistic gelatin (temperatures permitting).

Figure 2.

Front and side views of the (a) target panel mount and (b) velocity-measurement device on the gas gun setup. A point laser was used to ensure accuracy of the right-circular cylinder projectile shot on the target panel.

Target panels were clamped on all four corners using L-brackets with inner neoprene rubber linings (50 A Durometer) of 25.4 mm (1”) width and 1/8-inch thickness to grip the target panel firmly, leaving an exposed surface area of 0.229 × 0.229 m² (9” × 9”). The L-brackets were secured using 12 flanged screws equally spaced on all corners, and then tightened using a torque wrench to a maximum torque of 2.8 N-m (25 in-lb). Velocity measurements were made using a magnetic chronograph (MagnetoSpeed® Sporter) due to space limitations. The magnetic chronograph measurements have been compared with existing instrumentation, such as a laser diode system and an optical chronograph, to ensure that measured velocities were accurate to within 3.05 m/s (10 ft/s).

Target materials

The base Twaron® 2040 fabric samples (840 denier yarns) were obtained from Barrday and manufactured in 2015, with an areal density of 0.215 kg/m² per ply and 4.732 kg/m² for 22 plies. The hybrid target panel materials were selected for flexibility and to have a variety of shear strengths, namely Makrolon® film, cotton, and 120 × 120 grid size 304 stainless steel mesh (single-ply A_d 0.1585, 0.2720, and 0.4736 kg/m², respectively). The shoot packs were designed and tested in various combinations, as listed in Table 1, with the system areal densities kept relatively similar to the 22 plies (A25-22T) and 14 plies (B33-14T) of Twaron®. The percentage deviation of each system areal density from the full fabric pack is also calculated. The number of backing Twaron® plies was chosen for inelastic energy analysis, as will be detailed in the next section.

Table 1.

Target panel materials and properties used in this study

	Panel material
Panel No.	Front matl.	Twaron®	Sys. A_d [kg/m²]	% Front matl	% Dev.
A25-22T	-	22	4.732	0	0.0
A18-2C-19T	2 Cotton	19	4.631	12	2.1
A32-2S-18T	2 Steel	18	4.819	20	1.8
A29-5M-18T	5 Makrolon	18	4.664	17	1.4
A34-6C-14T	6 Cotton	14	4.643	35	1.9
A27-4S-14T	4 Steel	14	4.906	39	3.7
A35-11M-14T	11 Makrolon	14	4.755	37	0.5
A36-10C-9T	10 Cotton	9	4.656	58	1.6
A30-6S-9T	6 Steel	9	4.778	59	1.0
A37-18M-9T	18 Makrolon	9	4.789	60	1.2
A31-8S-4T	8 Steel	4	4.649	81	1.8
B33-14T	-	14	3.011	0	0.0
B40-2S-9T	2 Steel	9	2.883	33	4.3
B39-7M-9T	7 Makrolon	9	3.045	36	1.1
B38-8C-4T	8 Cotton	4	3.036	72	0.8
B41-14M-4T	14 Makrolon	4	3.079	72	2.3

T: Twaron®; C: cotton; S: 304 stainless steel; M: Makrolon® polycarbonate.

The fabric materials were cut to sample sizes of 0.305 × 0.305 m² (12” × 12”), and subsequently edge-stitched three times together with a 25.4 mm (1-inch) margin from the edges for easier handling. Due to the loose stitching and stitch spacing, the edge-stitching does not significantly alter the ballistic performance. Since stitching is not possible for stronger target materials (e.g., Makrolon® and 304 stainless steel mesh), these non-fabric layers were first attached at the edges with tape. For consistency, all panels were then heat-shrunk in a polyvinyl chloride (PVC) bag with negligible areal density and thickness (0.051 mm). Panels were kept in storage in an air-tight container with clay desiccant packets for at least 12 hours prior to shooting to absorb any moisture that may be present due to transportation.

Projectiles

O1 tool steel RCC pieces were cut and faced from 9 mm rod stock to make the RCC projectiles. A nominal diameter of 9 mm was chosen as part of a larger study that was ongoing concurrently. Diameters and lengths of the RCC projectiles were measured to be within 9.01 ± 0.01 mm (0.1% deviation) and a nominal mass of 4.48 g. Due to the slightly over-sized bore diameter compared to the RCC projectiles, Lyman copper gas checks were lightly attached using a thin layer of petroleum jelly to the non-impact end of the RCC projectiles to form a better gas seal for higher gun efficiency and higher exit velocities. The gas checks weigh approximately 0.4 g, and this weight is taken into consideration for effective density calculations. Prior to shooting, the rear non-impact end of these RCC projectiles were marked using permanent ink to distinguish the impact end for post-mortem analysis.

The corners of the RCC projectiles were examined using scanning electron microscopy (FEI Nova NanoSEM 200) to verify the effect of target strength on any potential blunting of the corners that may affect the ballistic performance during perforation. The pre-shot cylinders have a corner radius of approximately 125 µm (Figure 3(a)). Pre-shot projectile surfaces appear to be relatively smooth with minimal damage except for microscopic residual grooves due to machining processes (Figure 3(b)).

Figure 3.

Micrographs of O1 tool steel right-circular cylinder projectiles prior to shooting, with measured corner radius of approximately 250 µm (a) and microscopic grooved surfaces due to machining (b).

Shooting procedure

A total of 12 shots were performed to determine the ballistic limit using the bracketing method, as detailed in NIJ-0101.0615. Shot locations were pre-determined and marked using a template such that the shots were located 25.4 mm (1 inch) from the panel stitching, at least 50.8 mm (2 inches) apart from each other, and (as much as possible) that the principal yarns do not overlap. For uniformity in testing, pre-and post-test temperatures and humidity levels were also recorded to ensure that testing conditions do not vary significantly.

As the test chamber and targets are not exactly the same as detailed in the NIJ-0101.06¹⁵ standard, the methodology was slightly modified. The first shot is fired at a desired velocity of 304.8 m/s (1000 ft/s) – this velocity was referenced and estimated with respect to the manufacturer’s datasheet for a 9 mm full metal jacket (FMJ) projectile impact. If the shot outcome is a partial penetration, a shot outcome of “0” is assigned to the shot number and the subsequent desired shot velocity is increased by 30.5 m/s (100 ft/s); if the shot outcome is a complete penetration, a shot outcome of “1” is assigned to the shot number and the subsequent desired shot velocity is decreased by 304.8 m/s (1000 ft/s). In the case of an unacceptable shot, for example an inaccurate shot location or large deviation of actual striking velocity from desired shot velocity of more than 3.05 m/s (10 ft/s), the shot is repeated.

The process is repeated until the first shot outcome “reversal,” that is, from partial to complete penetration or vice versa. At this point, the change in desired velocity is lowered to 22.9 m/s (75 ft/s). Similarly, if the shot outcome is a partial penetration, a shot outcome of “0” is assigned to the shot number and the subsequent desired shot velocity is increased by 22.9 m/s (75 ft/s); if the shot outcome is a complete penetration, a shot outcome of “1” is assigned to the shot number and the subsequent desired shot velocity is decreased by 22.9 m/s (75 ft/s).

Again, this process is repeated until the next shot outcome “reversal,” where the desired velocity step is further lowered to 15.2 m/s (50 ft/s). If the shot outcome is a partial penetration, a shot outcome of “0” is assigned to the shot number and the subsequent desired shot velocity is increased by 15.2 m/s (50 ft/s); if the shot outcome is a complete penetration, a shot outcome of “1” is assigned to the shot number and the subsequent desired shot velocity is decreased by 15.2 m/s (50 ft/s). The procedure is then repeated until a total of 12 acceptable shots are completed, up to a total of 16 possible shots per target panel if necessary in the case of unacceptable shots.

Results and discussion

Recorded pre- and post-testing temperatures and relative humidity levels were between 17.0℃ and 20.0℃ and 34% and 41%, respectively. Tests were typically completed within 5 hours of test commencement. For the 14-ply Twaron®-equivalent Series B targets, high-speed images were taken using a Shimadzu HPV-X2 (high-speed imaging capabilities were not available during the shooting phase for Series A panels). These images were used to ensure normal trajectory of the projectile without any significant yaw or pitch during flight (Figure 4).

Figure 4.

High-speed image sequence capture of the projectile impacting the fabric target at 339 m/s, with a frame rate of 400 kHz and 200 ns exposure. A brief flash occurs at the time and site of impact (t = 0). Note that image corrections of +20% brightness and +20% contrast were applied to improve image clarity.

The image sequences revealed a particularly interesting phenomenon whereby a brief flash of light occurred at the time of impact, and seemingly appears only where the projectile initially contacts the fabric target. The flash only occurs very briefly for a maximum duration (estimated from the frame rate and exposure) of about 5 µs. At t = 0, a small region around the right-hand side of the impact site of the RCC projectile lights up, and this corresponds with yarn movement in the fabric around the impact site. This is more apparent at t = 2.5 µs, where the engaged principal yarns running vertically are being strained – the flash looks to be the brightest in this area as well.

A similar phenomenon was previously observed for ultra-high molecular weight polyethylene (UHMWPE) by Chocron et al.¹⁶ when impacting Dyneema® HB80 laminates with a polyurethane matrix, and recently by Yang and Chen¹⁷ when impacting Dyneema® SB71 UD laminates. Their images obtained were from the back, and it is only inferred that the flash happens on impact, although there is no direct visual evidence. Chocron et al.¹⁶ attribute this to isentropic shock loading of the polyurethane matrix upon impact, and the flash is a result of an “autoignition effect” from the shock, resulting in localized melting of either the UHMWPE fibers or the polyurethane matrix, or both. As far as the authors are aware, there are currently no prior reports of similar phenomena occurring for aramid fibers, but it is possible that this flash is related to extremely localized deformation/damage.

Ballistic limit results

The outcome of each shot was assigned a value of “0” for non-perforated (or partial penetration) shots, and a value of “1” for perforated (or complete penetration) shots. Perforation of the panel was verified visually during the test, and via post-mortem for confirmation. Two different methods of estimating the V50 ballistic limit are given by the guidelines detailed in NIJ-0101.06¹⁵ and MIL-STD-662F,¹⁸ and these two methods were compared and averaged. The former uses a regression for a logistical S-curve to determine two different regression constants. The shot outcomes (either “1” or “0”) were plotted against their respective striking velocities.

The data points were curve-fitted using a Levenberg–Marquardt algorithm (the exact alogirthm for fitting was not specified) to obtain two curve-fit parameters, β₀ and β₁, as in Figure 4. The ballistic limit is then calculated, with the regression curve given as

π (V_{s}) = \frac{e^{β_{0}} + β_{1} V_{s}}{1 + e^{β_{0}} + β_{1} V_{s}}

(5)

V_{50} = - \frac{β_{0}}{β_{1}}

(6)

In equation (5), V_s is the striking velocity and π(V_s) is the probability of penetration at that particular striking velocity, as defined previously by either partial penetration or complete penetration. The V50 is then calculated using equation (6). Using Panel A27-4S-14T as an example, β₀ and β₁ are fitted as −287.5 and 0.8294, respectively, giving a calculated V50 via this method as 346.6 m/s.

In view of some of the possible uncertainties in calculating the V50 using just one method, the MIL-STD-662F¹⁸ method of calculating the V50 was also employed. This method uses the arithmetic mean of the same number of lowest complete penetration velocities and highest partial penetration velocities. Acceptable velocities are selected within a range of 125 ft/s, and at least three points from each group have to be selected. As an example, in Figure 5, the lowest complete penetration points used in calculation are highlighted in red, and the highest partial penetration points are highlighted in green. Using this method, the V50 calculated is 347.0 m/s. The V50s calculated using both methods are averaged and presented in Table 2. Although there are slight differences in areal densities of the target systems, this slight difference of up to 3.7% is not expected to contribute significantly to the differences in ballistic performance between panels. The results are all given in Table 2.

Table 2.

Experimental V50 results for Series A and B panels

	Panel material			V50 [m/s]
Panel	Front matl.	Twaron®	% Front matl.	NIJ	662F	Ave.	% Diff.
A25-22T	–	22	0	331.8	339.0	335.4	–
A18-2C-19T	2 Cotton	19	12	331.4	332.6	332.0	–1.0
A32-2S-18T	2 Steel	18	20	346.0	349.7	347.9	+2.8
A29-5M-18T	5 Makrolon	18	17	341.8	339.9	340.9	+0.7
A34-6C-14T	6 Cotton	14	35	339.9	332.0	336.0	–0.7
A27-4S-14T	4 Steel	14	39	346.6	347.0	346.8	+2.5
A35-11M-14T	11 Makrolon	14	37	315.5	314.9	315.2	-6.8
A36-10C-9T	10 Cotton	9	58	293.0	309.5	301.2	–11.0
A30-6S-9T	6 Steel	9	59	295.1	289.9	292.5	–13.6
A37-18M-9T	18 Makrolon	9	60	322.6	315.9	319.2	–5.7
A31-8S-4T	8 Steel	4	81	263.7	265.2	264.5	–21.8
B33-14T	–	14	0	300.6	298.7	299.7	–
B40-2S-9T	2 Steel	9	33	298.7	299.7	299.2	–0.2
B39-7M-9T	7 Makrolon	9	36	302.7	302.1	302.4	+0.9
B38-8C-4T	8 Cotton	4	72	280.5	279.0	279.7	–7.7
B41-14M-4T	14 Makrolon	4	72	286.3	282.6	284.5	–5.1

The V50 values were then plotted against the percentage of frontal material, as per Figure 5. A Weibull-type fitting was used with the equation

V_{50} = V_{50 f} exp (- f_{s}^{α})

(7)

where V₅₀ _f is the V50 ballistic limit of a full fabric target system of equivalent areal density, f_s is the sacrificial material percentage, and α is a regression coefficient for the curve fit. The coefficient α has a value of 5.176 for Series A shots and 8.472 for Series B shots. The darker dashed line in Figure 6 represents the Weibull curve-fit and the lighter dashed lines represent one standard deviation from the curve-fit.

Figure 5.

Typical plot of penetration probability against striking velocity for a target panel. Shown here are the shot outcomes for Panel A27-4S-14T.

Figure 6.

Plotted ballistic limits against frontal material ratio for the respective materials for (a) Series A and (b) Series B target panels.

The ballistic performance for all different frontal materials in Series A is not significantly altered up to 40%. Past this point, the ballistic limit begins to fall, even though the areal density is the same. The same effect is observed for Series B panels. After penetrating this frontal material, the residual velocity must exceed the V50 of the remaining Twaron® – the overall performance is therefore dependent on their elastic properties. This is where the excellent mechanical behavior of high-performance fibers comes into play. At lower frontal material ratios, the high-performance Twaron® fabric still absorbs an appreciable amount of impacting energy from the RCC projectiles. However, with higher frontal material ratios (at the same areal density), the percentage of high-performance material decreases. The V50 with respect to the same RCC projectile then becomes low enough that the overall performance suffers.

Interestingly, even though cotton has a much lower strength than either steel or Makrolon® polycarbonate, there appears to be little to no discernible difference in ballistic performance for a given frontal material percentage, even at large ratios. These results only serve to reinforce the efficacy of using ballistic fibers as protective gear, as they still achieve the best performance at this given areal density ratio. A superficial understanding of the results seems to suggest that the effect of target strength does not seem to play a huge role in performance. However, to imply that the ballistic performance is only dependent on the areal density of the target is not entirely true, as this also means that the layering order of the target material and fabric is not important. This has been shown not to be the case, as Cunniff⁵ demonstrated a distinct difference in performance by a factor of about two when changing the order of Kevlar and Spectra single-ply layers.

This drastic difference in ballistic performance was easily demonstrated by shooting a hybrid panel Twaron® fabric face first and steel mesh first, that is, a 4/14 steel mesh/Twaron® hybrid panel (Panel A-4S-14T) and a 14/4 Twaron®/steel mesh panel respectively, in this case (Figure 7). Both panels were shot at approximately 300 m/s. For Panel A27-4S-14T, the 304 stainless steel mesh frontal layers sheared off before the projectile was stopped by the Twaron® backing material without even punching through the first layer. On the other hand, for the 14/4 Twaron®/steel mesh hybrid panel (Panel 14T-4S, not listed in Tables 1 or 2), the front 14 layers of Twaron® were perforated, but the steel mesh backing material failed critically after just one shot (Figure 7).

Figure 7.

Post-impact images of 4/14 steel/Twaron® panel A27-4S-14T (a) and 14/4 Twaron®/steel hybrid panel 14T-4S (b).

For Panel A27, the frontal steel mesh obviously failed locally, resulting in a very well-defined shear plug. The residual velocity of the RCC projectile after punching out the steel shear plug was insufficient to cause localized damage in the Twaron® fabric, and so this absorbed energy gets somewhat dissipated around the impact site, as can be observed from the resulting tent shape.

Target post-mortem analysis

Target panels were analyzed post-mortem for different modes of damage and deformation near the impact site (Figure 8). For cotton, stainless steel mesh, and Makrolon® sheets, shear plugs were consistently formed for each shot regardless of the outcome. For Twaron®, formation of such a shear plug only occurred once out of all the impacted panels and their constituent plies, indicating the extreme unlikelihood of such an event. Instead, even though the initial few plies of Twaron® experience an impact velocity much higher than their individual ballistic limits, they often fail halfway along the circumference of the RCC projectile to form a semi-circular tab rather than punching out the material fully (Figure 8). The failure of the Twaron® fabric appears to be similarly localized, as observed from optical microscope images (Figure 9).

Figure 8.

(a) Recovered post-impact target “shear plugs”: Twaron® (A), 304 stainless steel mesh (B), cotton (C), and Makrolon® polycarbonate (D), with O1 tool steel right-circular cylinder projectile for reference. (b) Typical semi-circular tab formation of initial plies of Twaron®.

Figure 9.

Optical microscope images of Twaron® shear plug edges at 12.5× magnification. Fiber and yarn failure appear to be extremely localized.

The measured diameters of the recovered shear plugs are 8.60, 9.01, 8.11, and 8.94 mm for Twaron®, stainless steel mesh, Makrolon®, and cotton, respectively, reflecting the extremely localized inelastic failure of the initial plies via shearing/cutting or stress concentration-induced localized tensile failure. This indicates again that this is possibly where their respective failure strengths may become significant.

Energy analysis

The decoupled through-thickness ballistic response of a target armor system allows for a somewhat uncommon method of energy analysis to investigate the effects of the frontal material strength on the inelastic impact response. From the experimental results in Table 2 and Figure 6, it is observed that the V50 limit does not drop significantly up to a frontal material ratio of 40%, which includes all of the Series A-XX-14T panels (i.e. all panels with 14 plies of Twaron® as backing material). Series B comprises panels of areal densities equivalent to 14 plies of Twaron® fabric. In other words, at the V50 ballistic limit of the entire Series A-14T armor system, the residual velocity of the RCC projectile after penetrating the frontal layer is sufficiently high to defeat a system of equivalent areal density to Series B. We first partition the overall system into three constituent subsystems I, II, and III, each with a certain percentage by mass of the overall system areal density. In Figure 10, subsystems II and III, when combined, will form an equivalent Series B panel, as in Table 3. The percentage by mass of each subsystem is similar for all panels, and subsystem III for all panels consists of nine Twaron® plies.

Figure 10.

Schematic of decoupled response for Series A-14T panels into three subsystems and equivalent Series B panel on backing plies.

Table 3.

Series A-XX-14T and equivalent Series B panels for backing material behind subsystem I

		Subsystem
A Series panel		I		II		III		Equivalent B Series panel
No.	A_d [kg/m²]	Mat’l	% A_d	Mat’l	% A_d	Mat’l	% A_d	Equivalent B Series panel
A25-22T	4.732	8 T	36.4	5 T	22.7	9 T	40.9	B33-14T
A34-6C-14T	4.643	6 C	35.1	5 T	23.2	9 T	41.7	B33-14T
A27-4S-14T	4.900	4 S	38.7	5 T	21.8	9 T	39.5	B33-14T
A35-11M-14T	4.755	11 M	36.4	5 T	22.9	9 T	40.7	B33-14T
A30-6S-9T	4.778	4 S	39.6	2 S	19.9	9 T	40.5	B40-2S-9T
A37-18M-9T	4.789	11 M	36.2	7 M	23.4	9 T	40.4	B39-7M-9T

Panel A25-22T is included because this panel can be considered a 14-layer Twaron® fabric with a frontal eight layers of Twaron®. Since the V50 is obtained statistically, we make the necessary assumption that at this velocity, the RCC projectile fully penetrates the panel but exits with negligible residual velocity. Equation (4) gives the inelastic energy absorption as the difference between the projectile striking and residual kinetic energies – in this case, the residual kinetic energy of the initial subsystem I for Series A-14T panels is the V50 kinetic energy of the backing 14 plies of Twaron®, that is, V50 for B33-14T. This difference in projectile kinetic energy is given in Table 4 under the column ΔKE.

Table 4.

Series A-14T and Series B panels for 14-ply Twaron® equivalent backing material

A panel		Subsystem			Equivalent B panel		ΔKE [J]	KE_plug [J]	E_shear [J]
No.	V50 KE [J]	I	II	III	No.	V50 KE [J]	ΔKE [J]	KE_plug [J]	E_shear [J]
A25	279	8 T	5 T	9 T	B33	219	60.3	4.92	55.3
A34	275	6 C	5 T	9 T	B33	219	56.3	4.65	51.7
A27	293	4 S	5 T	9 T	B33	219	74.3	5.42	68.9
A35	242	11 M	5 T	9 T	B33	219	23.3	4.94	18.3
A30	209	4 S	2 S	9 T	B40	218	-9.67	5.39	–15.1
A37	249	11 M	7 M	9 T	B39	223	25.5	5.04	20.4

From the third term in equation (3), the kinetic energy of an assumed fabric “shear plug” of mass A_dA_p moving at the residual velocity after penetrating subsystem I (i.e., V50 velocity of the equivalent Series B panel) was also calculated under column KE_plug. Since recovered “shear plugs” were approximately the same diameter as the impacting RCC projectile (Figure 8), we assume that η = 1; a typical value of η for a Kevlar 29 fabric system is given as 1.3.⁶ At striking velocities much higher than the ballistic limit, the difference in kinetic energy ΔKE should be approximately equal to the plug kinetic energy, since E_shear in equation (3) remains constant. Again, it is stressed that for ballistic fabric material, this inelastic “shear energy” more than likely refers to localized failure around the impact site rather than a physical shear plug, as the latter has not been observed experimentally in the existing literature or in this study.

From the calculated energy values, we find that this shear plug kinetic energy far exceeds the change in kinetic energy of the projectile. Recall that the A-14T series panels have similar subsystem I areal densities, implying that the energy dissipated due to inelastic mass collision must be similar regardless of material. This unaccounted deficit must therefore be related to some strength property inherent to the frontal material in subsystem I. At this moment, the authors have yet to find a definitive correlation to the shear strength as per equation (2), mainly due to the difficulties in obtaining an effective shear strength and thickness value for meshes and fabrics.

Although included as examples of subsystem energy partition, panels A30 and A37 do not have 14 plies of Twaron® as backing material and are therefore excluded in the above analysis. It is noted that comparing panel A30 to A27 numerically gives the change in energy absorption if the middle five layers of Twaron® were replaced by two layers of steel mesh; the same can be said by comparing panel A37 to A35, which numerically gives us the change by replacing the five Twaron® layers with Makrolon®. This conclusion is incorrect, as the impacting velocity on these middle layers may not actually be sufficiently fast to result in inelastic impact, and the energy partitioning analysis breaks down.

Projectile post-mortem analysis

The post-impact projectiles were recovered, and their dimensions were measured using calipers. Although projectiles were captured post-impact to the best of the authors’ abilities, some projectiles unexpectedly yawed excessively after exiting the target and subsequently became damaged upon hitting the safety chamber – these projectile dimensions were not included. Measured projectiles were 9.01 ± 0.01 mm in diameter and 9.00 ± 0.02 mm in length, indicating negligible macro-scale deformation for O1 tool steel RCC projectile impact at these striking velocities.

This brings us back to the initial hypothesis that projectile deformation during impact results in a different ballistic limit despite having the same areal density for the whole system. A quick look at the formulation of the areal density ratio suggests that a system can deform the projectiles either via mass loss or via “mushrooming” of the projectile’s impact end, both of which would increase η and consequently the ballistic limit. For this study, the macro-scale deformation was negligible and so η did not change significantly regardless of the outcome of the shot, implying that any deformation must happen at the micro-scale. To investigate this micro-scale deformation, the projectiles were subsequently examined using a scanning electron microscope. Figures 11 –14 show the micrographs for a variety of frontal material strengths and ratios, and impact velocities. Some tilt when mounting the projectiles may result in a spuriously large radius, and attempts were made to minimize any sample tilt under the microscope.

Figure 11.

Micrographs of post-impact O1 steel right-circular cylinder projectile (a) corner and (b) impact end circumferential surface for Panel A32-2S-18T Shot 1, partial penetration at 311 m/s.

Figure 12.

Micrographs of post-impact O1 steel right-circular cylinder projectile (a) corner and (b) impact end circumferential surface for Panel A36-10C-9T Shot 2, partial penetration at 251 m/s.

Figure 13.

Micrographs of post-impact O1 steel right-circular cylinder projectile (a) corner and (b) impact end circumferential surface for Panel A36-10C-9T Shot 6, complete penetration at 313 m/s.

Figure 14.

Micrographs of post-impact O1 steel right-circular cylinder projectile (a) corner and (b) impact end circumferential surface for Panel A35-11M-14T Shot 10, partial penetration at 300 m/s.

Surface marring around the circumference of the cylinder impact end was apparent for all projectiles, presumably due to the high-performance fabric. Typical post-impact corner radii deformation for all impacted projectiles varied between approximately 150 and 250 µm from an original radius of about 125 µm, with no significant correlation between the deformation in radii and the impact velocity, frontal material strength, and frontal material ratio. On the other hand, there appears to be a slight correlation between the amount of Twaron® backing material and the change in corner sharpness. For example, for Panel A-10C-9T Shot 6 (nine Twaron® layers), the corner radius increased to 190 µm, while Shot 1 of Panel A32-2S-18T (18 Twaron® layers) increased its corner radius to 245 µm, even though they were fired at similar striking velocities. While the radius deformation is orders of magnitude smaller than the projectile radius, the constituent Twaron® fibers/yarns may be failing due to stress concentrations at the RCC projectile corner. Single fibers and yarns have been shown to fail at lower than theoretical critical velocities due to local variations in stress states, off-axis or otherwise.^19–25 Despite the multitude of studies on off-axis stress states during the transverse loading of yarns and fibers, these results have not been directly translated to a full fabric system due to its complexity. The micro-scale deformation of the RCC projectiles in this study suggest that the efficiency of high-performance fabric systems (or in fact any frontal material that may be used in place) is somewhat dependent on their ability to “round off” these corners, thereby reducing the effects of stress concentration and consequently any localized failure.

Comparison of results with previous works

As previously mentioned, Alesi¹³ shot 17 grain T37 FSPs on three different hybrid target panels: window glass/12-ply nylon fabric; A-110AT titanium alloy/8-ply nylon fabric; and PMMA/5-layer PVB, as well as nylon and PVB of equivalent areal densities (i.e., 100% high-performance material). In that study, the frontal material ratios and areal density ratios were not kept constant. Cunniff^3–5 shot 2-, 4-, and 16-grain RCC projectiles on 0.13-inch borosilicate glass/12-ply Kevlar® KM2 fabric as well as KM2 fabric panel of equivalent areal density, keeping the frontal material ratio constant but varying the areal density ratio. The frontal material ratios and areal density ratios are calculated in Table 5, and the comparison of V50 limits is given as well. The check mark signifies that the panel resulted in a higher V50.

Table 5.

Comparison of hybrid and 100% high-performance backing material V50s from different studies

	Materials				Higher V50
Study	Front	Backing	% Front	100 × η	Hybrid	100% Fabric
Alesi, 1957¹³	Glass	Nylon fabric	55	28.4	Similar
	A-110AT	Nylon fabric	66	25.0	✓
	PMMA	PVB	84	28.8	✓
Cunniff, 1999⁴	Pyrex	KM2 fabric	73	23.9		✓
	Pyrex	KM2 fabric	73	37.9	✓
	Pyrex	KM2 fabric	73	47.8	✓
Current study	Varied	Twaron fabric	Varied	6.17		✓
Current study	Varied	Twaron fabric	Varied	3.93		✓

PVB: polyvinyl butyral.

It appears from previous studies and the current dataset that the performance of a hybrid panel compared to a 100% high-performance backing material of equivalent areal density is determined largely by the areal density ratio η. To re-iterate Cunniff’s argument,^3–5 lower areal density RCC projectiles (2- and 4-grain) deformed when impacting the harder borosilicate glass frontal layer, whereas they do not deform during fabric armor impact. Similarly, since the larger 16-grain projectiles did not deform enough to affect the presented area significantly, the fabric armor behaves as expected, and the performance surpasses that of the hybrid panel. The transition seems to occur around a η value of 0.25–0.30.

Conclusions

In this study, the effects of replacing different amounts of high-performance Twaron® fabric with stainless steel mesh, Makrolon® polycarbonate, and cotton were studied as a possibility of replacing initial layers of high-performance material in an armor system with a lower cost alternative, or some material with more desirable characteristics. These target panels were designed to have about 15%, 33%, 60%, and 75% frontal “sacrificial” material for two different areal densities at 4.732 and 3.011 kg/m² (Series A and B, respectively), and were impacted with an O1 tool steel RCC projectile. It was found that the ballistic performance of the system was maintained up to approximately 40% when using a Weibull-type curve-fit. Non-high-performance frontal materials exhibited extremely localized failure and formed shear plugs consistently via a shearing/cutting mechanism, while the initial plies of Twaron® fabric exhibited a semi-circular tab at the impact site instead, although the failure mode appeared to be similarly localized or sheared. Since the ballistic response is decoupled, energy comparisons were made by partitioning the Series A armor system into three subsystems, and then comparing the striking energy at the ballistic limit with Series B panels. Deductions again showed the possibility of further energy absorption via inelastic “shear plug” formation related to transverse mechanical properties. Scanning electron microscopy of post-impact projectiles showed slight deformation of the RCC projectile corners, thereby reducing any stress concentration effects.

The design of a hybrid armor system therefore must meet two requirements: the high-performance material portion at any frontal material ratio must still have a ballistic limit high enough to maintain the same performance as a full fabric armor system,⁴ and any frontal material used should be able to reduce the effects of stress concentration sufficiently for any improvement in ballistic performance. The frontal materials used in this study may not deform the corners significantly compared to the Twaron® fabric, but this is most likely where target frontal material strengths come into play, although further studies are recommended.

Footnotes

Acknowledgements

The authors would like to thank the US Army, P.M. Soldier Protection and Individual Equipment, Technical Management Directorate for their support. The authors would also like to thank Stephanie Martinez-Morales for their help and contribution with sample preparation and technical help with running the experiments, and Joan Goetz of the Purdue Textile Science Lab for access and help with fabric panel stitching.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Roylance

Wilde

Tocci

. Ballistic impact of textile structures. Text Res J 1973; 43: 34–41.

Cunniff

. An analysis of the system effects in woven fabrics under ballistic impact. Text Res J 1992; 62: 495–509.

Cunniff PM. A design tool for the development of fragmentation protective body armor. In: 18th international symposium on ballistics, November 1999, pp. 1295–1302. IBS.

Cunniff PM. Decoupled response of textile body armor. In: 18th international symposium on ballistics, November 1999, pp. 814–821. IBS.

Cunniff

. A semiempirical model for the ballistic impact performance of textile-based personnel armor. Text Res J 1996; 66: 45–59.

Phoenix

Porwal

. A new membrane model for the ballistic impact response and V50 performance of multi-ply fibrous systems. Int J Solids Struct 2003; 40: 6723–6765.

Flanagan

Zikry

Wall

et al.

An experimental investigation of high velocity impact and penetration failure modes in textile composites. J Compos Mater 1999; 33: 1080–1103.

Nguyen

Ryan

Cimpoeru

et al.

The effect of target thickness on the ballistic performance of ultra high molecular weight polyethylene composite. Int J Impact Eng 2015; 75: 174–183.

Prosser A. Penetration of nylon ballistic panels by fragment-simulating projectiles Part II: mechanism of penetration. Text Res J 1988; 58: 161–165.

10.

Walker JD. Constitutive model for fabrics with explicit static solution and ballistic limit. In: 18th international symposium on ballistics, San Antonio, Texas, 1999, pp.1231–1238. IBS.

11.

Walker JD. Ballistic limit of fabrics with resin. In: 19th international symposium of ballistics, Interlaken, Switzerland, May 2001, pp.1409–1414. IBS.

12.

Hudspeth MC. Multi-axial Failure of High-Performance fiber During Transverse Impact. PhD Thesis, Purdue University West Lafayette, USA, 2016.

13.

Alesi AL. Composite personnel armor (No. TR-CP-5). Report, Quartermaster Research and Engineering Command, US Army Natick, USA, December 1957.

14.

NIJ Standard-0115.00:2000. Stab resistance of personal body armor.

15.

NIJ Standard–0101.06:2008. Ballistic resistance of personal body armor.

16.

Chocron

King

Bigger

et al.

Impacts and waves in Dyneema HB80 strips and laminates. J Appl Mech 2013; 80: 31806–31806.

17.

Yang

Chen

. Investigation of failure modes and influence on ballistic performance of Ultra-High Molecular Weight Polyethylene (UHMWPE) uni-directional laminate for hybrid design. Compos Struct 2017; 174: 233–243.

18.

MIL-STD-662F:1997. V50 ballistic test for armor.

19.

Walker

Chocron

. Why impacted yarns break at lower speed than classical theory predicts. J Appl Mech 2011; 78: 51021–51021.

20.

Hudspeth

Nie

Chen

. Dynamic failure of Dyneema SK76 single fibers under biaxial shear/tension. Polymer (Guildf) 2012; 53: 5568–5574.

21.

Mayo

Wetzel

. Cut resistance and failure of high-performance single fibers. Text Res J 2014; 84: 1233–1246.

22.

Hudspeth

Chen

Zheng

. Why the Smith theory over-predicts instant rupture velocities during fiber transverse impact. Text Res J 2016; 86: 743–754.

23.

Sun

Hudspeth

Chen

. Biaxial shear/tension failure behavior of Spectra single fibers. Compos Part A Appl Sci Manuf 2016; 88: 286–294.

24.

Hudspeth

Spatola

et al.

The effects of off-axis transverse deflection loading on the failure strain of various high-performance fibers. Text Res J 2016; 86: 897–910.

25.

Hudspeth

Chu

Jewell

et al.

Effect of projectile nose geometry on the critical velocity and failure of yarn subjected to transverse impact. Text Res J 2017; 87: 953–972.