Shock initiation sensitivity and Hugoniot-based equation of state of Composition B obtained using in situ electromagnetic gauging

A series of gas gun-driven plate impact experiments were performed on vacuum melt-cast Composition B to obtain new Hugoniot states and shock sensitivity (run-distance-to-detonation) information. The Comp B (ρ0 = 1.713 g/cm3) consisted of 59.5% RDX, 39.5% TNT, and 1% wax, with ~ 6.5% HMX in the RDX. The measured Hugoniot states were found to be consistent with earlier reports, with the compressibility on the shock adiabat softer than that of a 63% RDX material reported by Marsh.[4] The shock sensitivity was found to be more sensitive (shorter run distance to detonation at a given shock input condition) than earlier reports for Comp B-3 and a lower density (1.68-1.69 g/cm3) Comp B formulation. The reactive flow during the shock-to-detonation transition was marked by heterogeneous, hot spot-driven growth both in and behind the leading shock front.


Introduction
Composition B is a widely-used high explosive consisting of an approximately 60:40 ratio of RDX and TNT by weight, and is part of a larger class of melt castable explosives based on TNT. "Comp B" can be prepared by a variety of methods including vacuum melt, open melt or hot pressing, and with or without 1% wax desensitizer [1]. Despite its prevalence as a military explosive, there are only a few reports of unreacted equation of state (Hugoniot) data or measurements of shock initiation sensitivities of various Comp B formulations [2][3][4][5][6][7][8][9][10]. Traditional Comp B consists of 59.5 wt% RDX, 39.5 wt% TNT, and 1 wt% wax desensitizer, and is melt-cast either under vacuum melt or open melt conditions to obtain bulk charges with densities of 1.713-1.715 g/cm 3 or 1.680-1.690 g/cm 3 , respectively. Historically, high quality melt castings of Composition B often resulted in a "richening" of the formulation with up to 4% RDX; for example "LASL Comp B" contained 63 wt% RDX, 36 wt% TNT, 1 wt% wax [4]. Composition B-3 is similar to Composition B, but contains no wax densensitizer, and can similarly, be prepared by (vacuum or open) melt casting, or hot pressing.
In the present work, gas gun-driven plate impact experiments were employed to: 1) obtain new unreacted shock (Hugoniot) states, and determine the shock sensitivity (Pop-plot) of vacuum melt Comp B, with comparisons with previous reports, and 2) apply in situ electromagnetic gauging to investigate the reactive flow during and subsequent to the shock-to-detonation transition. Photonic 1 To whom any correspondence should be addressed. Doppler velocimetry (PDV) was also applied, for the first time, to measure the spatiotemporal characteristics of the detonation profile and chemical reaction zone following turnover to detonation.  figure 1A.  Microsecond-duration, sustained shock waves were introduced into Comp B samples by gas gundriven plate impact using a 50 mm bore launch tube two-stage light gas gun, described previously [11][12][13], to obtain projectile velocities of up to 2.034 km/s. The principal diagnostic in the experiments was electromagnetic gauging [13], which allows for measurement of particle velocity wave profiles at up to 10 Lagrangian positions in the sample, including at the impact interface. Figure 1B shows a photograph of an assembled Comp B target showing a stirrup gauge on the impact face, and gauge inserted on 30 o angle. Figure 2A shows the details of the embedded gauge package which consists of 5 µm patterned Al foil gauge elements, sandwiched between two 25 µm films of FEP-PTFE, forming a membrane ~60 µm thick. In the gauge package, there are 9 "particle velocity trackers," which are the horizontal elements indicated in the figure. The particle velocity trackers, in concert with a stirrup gauge element at the impact interface, figure 2B, allow for measurement of particle velocity wave profiles at up to 10 positions in-material. In addition, there are 3 "shock trackers," providing closelyspaced gauge elements for higher temporal resolution, redundant measurements of the shock velocity. The final target geometry is shown in figure 2C. The embedded gauge membrane is inserted into the wedge sample at a 30 o angle, which spaces the gauges ~ 0.8 mm apart from the impact face. At the back interface of the target (approximately 22-23 mm from the impact face), an 8 kÅ Al mirror was coated to a PMMA window and affixed to the Comp B for measurement of a short run-distance detonation profile using PDV. [14] Figure 2. A) Photograph of the LANL embedded electromagnetic gauge package. B) A single "stirrup gauge" element is affixed at the impact face of the target to obtain the shock input condition. C) In the full target assembly, the embedded gauge package is inserted on a 30 o angle. A PMMA window was mounted at the rear surface of the Comp B, 22-23 mm from the impact face, to obtain a short rundistance detonation profile using PDV.

Experimental
A Lexan projectile fitted with a Kel-F 81 polymer impactor was impacted into the Comp B targets at velocities ranging from 1.321-2.034 km/s using the LANL two stage light gas gun [12]. A single experiment was performed using a z-cut sapphire impactor launched by a 72 mm bore single stage light gas gun [15]. The error in measured projectile velocity is ~0.1%, and is reported for each experiment in table 2.

Results and Discussion
A total of six plate impact experiments were performed on Comp B, with shock input stresses ranging from 1.0 to 8.4 GPa. In all of the two-stage experiments, Comp B shock initiated to detonation, and features of the reactive flow along the shock-to-detonation transition were measured in situ. Comp B initiates via a heterogeneous initiation mechanism derived from hot spot-driven reactive growth, as evidenced by an increase in particle velocity, both in and behind the shock front, similar to heterogeneous plastic-bonded explosives [11]. In contrast to explosives based on HMX, TATB, and TNT studied by these authors, electrical noise was measured on the embedded gauges up to a condition near the turnover to detonation, which is believed to be due to piezoelectric effects in RDX. Example particle velocity wave profiles from shot 2s-685 are shown in figure 3A. The response of the right shock tracker is shown in figure 3B, which illustrates the disappearance of electrical noise near the turnover to detonation. In shot 2s-685, Comp B was shocked to 6.6 GPa, and turnover to detonation is observed between gauges 5 and 6, or t* = 1.03 ± 0.07 µs and x* = 4.8 ± 0.1 mm.

Unreacted Hugoniot
The unreacted Hugoniot data, table 2, are plotted in the pressure-particle velocity (P-u p ) plane in figure  5A, along with experimental data reported by Marsh [4] and Lemar [3], and linear Rankine-Hugoniot relationships (U s = c 0 + su p ) by Coleburn and Liddiard [2], and Urtiew et al. [7][8]. A linear Rankine-Hugoniot fit to the new Hugoniot data in the U s -u p plane to 8.4 GPa is U s = 2.41 (± 0.27) + 2.30 (± 0.36)u p . As seen in the figure, there is qualitative agreement of the new unreacted Hugoniot states with those reported by references 2 and 7, but softer behavior than in Ref. 4 due to a lesser RDX content in the material studied here. An increase in scatter in the experimental data is observed as the input shock pressure is increased into the initiation regime, consistent with other explosives. In a single experiment (1s-1563), a quasi-elastic limit was measured, u p,el = 0.029 mm/µs, U s,el = 3.185 mm/µs, P HEL = 0.16 GPa, and a low pressure Hugoniot state below u p = 0.5 mm/µs.

Initiation mechanism and Pop-plot
Comp B initiates via a heterogenous initiation mechanism in which there is a continual increase in the shock and particle velocities in the front leading to detonation. There are marked similarities between Comp B, and plastic-bonded explosives, such as PBX 9501, in which the peak particle velocity associated with reactive burn lags the initial shock wave in space and time [11]. Also notable is the large increase in particle velocity from the unreacted shock input condition to the von Neumann condition at the front of the detonation wave. Following turnover, the detonation is nearly steady, similar to other heterogeneous explosives, and the measured detonation velocities are reasonably consistent with the reported steady detonation velocity D s = 7.92 km/s [15]. From the embedded gauge records, the run distance and time to detonation were determined to < 5%. The Pop-plot for vacuum melt Comp B is overlaid with that of Comp B-3 from Ramsay [5][6][7][8][9], and Urtiew [7][8] for a lower density formulation in figure 4B. The vacuum-melt Comp B was found to be more sensitive than previous reports, with the run distance being shorter by >5 mm at some input conditions. The detonation profile, measured at the PMMA window, showed a peak interface particle velocity u p ~ 3.

Conclusions
A series of shock initiation experiments have been performed on vacuum melt Composition B with an initial density of 1.713 g/cm 3 . From multiple in situ embedded electromagnetic gauges, new unreacted Hugoniot and run-distance (time)-to-detonation data are presented. Comp B initiates by a hot spotdriven "heterogeneous" initiation mechanism, with reactive growth close both in and behind shock front. The shock sensitivity of the vacuum melt formulation is greater than that reported by Ramsay (Comp B-3) [5][6][7][8][9] and Urtiew et al. [7][8]. Future work will compare the shock sensitivity and chemical reaction zone of vacuum melt Comp B with pressed Comp B-3 and related formulations.