Characterization of Detonator Performance Using Photonic Doppler Velocimetry M. P. Maisey and M. D. Bowden1 Atomic Weapons Establishment, Aldermaston, Reading, Berkshire, RG7 4PR, United Kingdom ABSTRACT Detonators are used to convert electrical or other energy into an explosive output. This output can then be used to initiate further explosive charges. To aid in the development of explosive systems, it is important to characterize the output of detonators, in particularly the pressure produced. Recent advances over the last five years in high-speed digitizing oscilloscopes and high-bandwidth photodiodes, driven primarily by the telecommunications industry, have enabled the development of a new type of interferometer for measuring high velocities, such as those found in detonics experiments. The Photonic Doppler Velocimeter (PDV) can be visualized as a fiber-based Michelson interferometer. The light from a single-mode fiber laser at 1550 nm is passed through a circulator, which acts to separate bi-directional light. The beam is then reflected via free-space optics off the surface of interest, and then focused back into the same fiber. This reflected light is then mixed with an approximately equal amount of non-reflected light, and the resulting interference is recorded using a high-bandwidth photodiode and oscilloscope. In contrast to more traditional Velocimetry techniques such as VISAR, only a single data channel is required. We have used our PDV system to investigate the performance of optical and electrical detonators. The detonators examined are the commercially available RISI RP-80, and an AWE DOI (Direct Optical Initiation) detonator. The RP-80 is an exploding bridgewire (EBW) detonator, utilizing Pentaerythritol Tetranitrate as the initiating explosive and a RDX output pellet. The DOI detonator uses an aluminum flyer to initiate a Hexanitrostilbene (HNS) pellet. Both detonators are canned in aluminum and the velocity of the can was measured, and from this, the output pressure of the detonator has been determined. This is compared to calculated values.
1. INTRODUCTION A detonator is an explosive device used to convert a non-explosive energy source, such as electrical energy stored in a capacitor, into an explosive output. This explosive output is generally used to initiate further explosive charges. It is therefore desirable characterize the output of detonators, in particular output pressure. This allows for reliable detonation systems to be designed. There are several methods available for measuring the output pressure of an explosive charge. We have chosen to use velocity interferometry to track the motion of the metal can enclosing the explosive charge. By measuring the velocity the metal can is instantaneously accelerated to (jump-off velocity) we can calculate the pressure. By using an optical technique we can reduce the likelihood of spurious electrical noise compromising our measurements. There are several variants of velocity interferometry used for dynamic experiments. VISAR (Velocity Interferometer System for Any Reflector) and its higher time resolution sibling, ORVIS (Optically Recorded Velocity Interferometer System) are the most widely used methods. They typically use a single line, single mode, Nd:YAG laser operating at 532 nm, with several watts of power. A fiber optic is used to deliver the laser light, with a second fiber being used to collect the scattered light from the surface of interest. This return light is then split, one half delayed, and then recombined. If the target is moving, the difference in frequencies between the delayed and non-delayed light produces interference. From the frequency produced, the velocity of the surface can be calculated. As the return signal is interfered with a timeshifted version of itself, relatively low frequencies for a given velocity are produced. VISAR/ ORVIS cannot resolve multiple velocities. Due to their size, VISAR/ORVIS systems are not easily portable, and the free-space optics 1
[email protected],
[email protected]; phone +44 118 982 5067 Optical Technologies for Arming, Safing, Fuzing, and Firing IV, edited by Fred M. Dickey, Richard A. Beyer, Proc. of SPIE Vol. 7070, 70700P, (2008) 0277-786X/08/$18 · doi: 10.1117/12.796267 Proc. of SPIE Vol. 7070 70700P-1
2008 SPIE Digital Library -- Subscriber Archive Copy
employed can be vulnerable to misalignment. PDV systems offer both portability and highly accurate measurements in ultra high velocity regimes.
2. THE PHOTONIC DOPPLER VELOCIMETER We reported the development of our Photonic Doppler Velocity in a previous paper1, based upon new type of velocity interferometer reported by Strand et al [1] in 2004. Photonic Doppler Velocimetry is based upon two physical processes; these being the Doppler Effect and Optical Mixing, i.e. heterodyning. When a laser signal is reflected from the surface of a moving target the wavelength alters, the degree of change being dependent upon the velocity of the target. This is the Doppler Effect, and for a medium-independent wave, such as light, the change in wavelength is:
v2 = v1
c−u c+u
For the PDV system, the velocity is related to the measured beat frequency by:
v=
λlaser 2
f beat
Extraction of the beat frequency from the heterodyned signal is not a straightforward process as the returned signal is invariably non-analytic. Frequency extraction is generally done by use of a sliding fast Fourier transform (i.e. a spectrogram) coupled with a maximum component locator. This technique does bring an unavoidable degree of ambiguity to the process, as spectrograms are resolution-limited in either frequency or time. Figure 3 shows a fairly typical spectrogram. Figure 2 shows the Velocimeter concept. A fiber-coupled laser is connected to a circulator. This device has the property of separating bi-directional light, which allows the return light to be separated from the original light. Of course, a portion of original light is required to interfere with the return light, and this is obtained from the back-reflectance of the probe. This is then matched to the return light intensity from the experiment. The interference between original and return beams is measured by a high-bandwidth photodiode, connected to a digitizing oscilloscope.
Amplifier
Figure 1 - Interferometer Concept A fuller examination of the physics behind PDV can be found in a previous paper by the authors [2].
Proc. of SPIE Vol. 7070 70700P-2
Time/s
xlU
Figure 2 – Typical PDV Raw Signal for HNS Pellet Spectrogram
lug
6.66
4.6 4
r
(l
3.6
= = U-
2.6 2
1.6
2
Time / a
2.6
x 1u6
Figure 3 – Typical Time Frequency Representation of DOI Jump Off Signal for HNS Pellet Our Photonic Doppler Velocimeter is contained entirely within fiber, except where the beam is incident upon the surface of interest. This allows for an exceptionally portable and robust system. The entire system, including laser, oscilloscope and interferometer, is mounted within a commercially available compact, shock-mounted 19” rack enclosure. A Photonic Doppler Velocimeter has several advantages over more traditional methods of velocimetry. It is capable of resolving multiple velocities simultaneously. Further, as it is constructed from standard telecommunications optical fiber
Proc. of SPIE Vol. 7070 70700P-3
components, it is relatively inexpensive, with a cost per point of approximately $35,000 for a high-bandwidth system as presented here.
3. THE DETONATORS We examined the output performance of two detonators: the commercially-available RP-80 manufactured by Risi, and an AWE-designed DOI detonator. The AWE developed DOI detonator (Figure 4) uses a Q-Switched laser pulse to irradiate the interface between a transparent substrate and a metal coating. This forms a high density plasma the expansion of which drives a flyer plate across an air gap into an explosive pellet, which then undergoes Shock to Detonation (STD) transition. The DOI detonator was fired with a range of HNS (similar to type IV) charges pressed to a density of 1.6 g/cm3, for a total of 147 mg of explosive. As this detonator has a bare explosive output, an aluminum disk 125 µm thick was bonded to the output charge using a cyanoacrylate adhesive. The Risi RP80 Electric Bridge Wire (EBW) Detonator is a high power EBW type detonator, consisting of a gold bridgewire, a low density PETN fill and a high density RDX based fill. A high voltage-high current (typically of the order of 1000 Volts at 1000 amps) signal is transmitted through the bridgewire, causing it to explode. Traditionally it was considered that this explosion drove a shock wave into the low density PETN fill, causing a shock to detonation transition (SDT) which is then picked up and amplified by the high density output pellet. An alternate opinion that is currently gaining ground is that the EBW functions via a Deflagration to Detonation transition in the low density fill [3].
PDV Probe
Detonator Body
Fibre Connector HNS Pellet
Flyer, Substrate and Barrel
Figure 4 – AWE DOI Detonator
Proc. of SPIE Vol. 7070 70700P-4
2 1 34 65 '-.M '-'-,
——::::::
1fl ((fl
1. Plastic molded head 2. Brass sleeve 3. Bridgewire (Gold) 4. Initiating explosive: 80 mg PETN 5. Output explosive: 123mg RDX 6. Aluminum cup 0.007 "thick Figure 5 – Risi RP80 EBW Detonator (ref)
4. EXPERIMENTAL METHOD The detonators were mounted in a polycarbonate experimental fixture, with the PDV probes held at approximately 3-4 mm standoff from the output face of the detonator of interest. Collimating probes were used, producing a collimated beam of 0.5 mm. No alignment of the probe relative to the surface was performed, primarily due to safety concerns relating to high power laser impingement upon explosives. Experiment data was lost for some shots, and for higher cost experiments, alignment is likely to be necessary.
5. DATA ANALYSIS Once the data has been collected from the PDV it was transformed using a spectrogram (the parameters for which are in Table 1). A maximum component finder was then run on the spectrogram to locate the component with maximum power, above some selected noise floor. Window Type Hanning Window Size 512 Overlap 500 Fourier Transform Length 512 Table 1 – Spectrogram Parameters Once the maximum power components were identified we extracted the time-frequency data for that point and converted frequency into velocity, yielding Time-Velocity graphs such as those shown in Figure 6. Once a time-velocity representation was determined the jump off velocity was read off. A 1D hydrocode model was then produced using Sandia National Laboratories CTH code of the system, and the detonation pressure required to accelerate the flyer to the recorded velocity was determined. A range of equations of state were used for both the aluminium and the explosive pellets in order to ensure that material parameter effects did not significantly distort the pressure calculations. The HNS parameters were taken from Goveas et al [4], whilst the aluminium parameters were taken from the library of materials provided with the CTH hydrocode.
Proc. of SPIE Vol. 7070 70700P-5
The Time-Velocity graphs can also be integrated to provide Time-Distance and Distance-Velocity graphs, providing improved insight into margins and uncertainties, and facilitating future detonator designs 4000 3500 3000
Velocity / m/s
2500 2000 1500 1000 500 0 2.50E-06
3.00E-06
3.50E-06
4.00E-06
4.50E-06
5.00E-06
5.50E-06
6.00E-06
-500 Tim e / s
Figure 6 – Time vs. Velocity Plot for the Free Surface of a HNS Pellet
6. RESULTS AND DISCUSSION We analyzed the output of our DOI detonator with 5 differing HNS compositions, along with the commercially available RP-80 detonator. All of the shots showed the same predominant features of a very rapid acceleration to a “jump-off” velocity followed by shock reverberation induced acceleration to a steady velocity before the signal is lost. The results of these shots are summarized in Table 2, whilst Figure 7 shows a selection of Time-Velocity plots for the free surface of the aluminium flyer. The results of the HNS shots show pressures that are, in general slightly higher than those previously published in the open literature by Dobratz and Crawford5 (21.5 GPa); with the exception of the first shot we performed on a 10.76 cc/g Specific Surface Area (SSA) HNS. This showed a significantly higher value for the output pressure of the HNS than has been reported previously. Setting this shot aside for the moment this is not surprising that the pressure recorded do not match with Dobratz and Crawford as they report on the CJ pressure, a state that only truly develops in steady state detonation, i.e. cylinder tests and rate stick type experiments. Another explanation for the higher than historical pressure values calculated are the Equations of State used to model the system. The range of pressures over which the materials are characterized is often far short of those found in detonic applications, we are thus unavoidably extrapolating to some degree in applying the available material models to our system. Finally a systematic effect may be responsible for our higher pressures in relation to 5. Our system occasionally produces identical duplicate signals at a range of frequencies. If the lowest frequency signal is considered the true, or base, signal then these higher frequency signals occur at two and three times the base signals frequency, i.e. they appear to be second and third harmonic signals. This is due to additional Doppler shifting of the original laser signal used to diagnose the free surface we are interested in. In essence the signal reflected from the target, undergoes some degree of reflection from the probe, reflects from the target again, and is collected by the probe having been Doppler shifted twice. In order to get a third signal, or forth, or fifth, this merely need to keep occurring, however detector limitations and the reduction of power with each reflection limit us currently to three Doppler shifted signals. The analysis software is in general is not
Proc. of SPIE Vol. 7070 70700P-6
fooled by such multiply-doppler shifted signals as they are in general far lower power than the singly shifted signal however. There is no significant correlation of HNS jump off velocity with SSA, with or without the inclusion of the 30 GPa shot 6. The standard deviation of our calculated pressure outputs for the RP-80 detonator agree to within 5% with [5] if one assumes that the output pellet of the RP-80 detonator is composed of PBX-9407 (94% RDX, 6% Exon 46 bw, 1.6 g.cc-1). Calculations of the output pressure of pure RDX at 1.6 g.cc-1 produces a far lower value of 25 GPa, making it unlikely that the RP-80 has just a RDX output pellet. Another possible explanation is that the RDX in the RP-80 is contaminated with HMX, a common effect of some synthesis processes, which would also increase the output pressure. In order to improve the fidelity of the system and future measurements a number of developments are apparent. In no particular order these are; investigate the cause and possible methods for reducing or removing multiply-doppler shifted signals; extend equation of state data so it is applicable to the pressure regime of interest to us; and to develop a means of aligning the PDV without affecting an experiment.
Detonator
Shot Number
Specific Surface Area
Jump-off Velocity Km/s
Calculated Pressure GPa
9.9 8.3 8.3 4.5 4.5 10.76 10.76 13.9 13.9
1.76 1.94 1.88 1.76 1.76 2.00 1.82 1.82 1.93 1.85 0.09
21.0 25.0 23.9 21.0 21.0 30.0 23.1 23.1 25.6 23.74 2.91
3.00 2.70 2.79 3.03 Risi RP-80 EBW 2.60 Unknown Detonator 2.85 (RDX @ 1.6 g/cc) 2.91 2.63 2.66 2.79 Mean 2.79 Std Dev 0.15 Table 2 – Summary of Results for Detonator Firings
28.2 27.0 27.4 28.3 26.5 27.6 27.9 26.7 26.8 27.4 27.4 0.63
AWE HNS Based DOI Detonator (HNS @ 1.6 g/cc)
Mean Std Dev
Proc. of SPIE Vol. 7070 70700P-7
3500
S1 C1 S2 C1 S3 C1 S6 C1 S7 C1 S8 C1 S9 C1
3000
2500
Velocity / m/s
2000
1500
1000
500
0 1.50E-06
1.70E-06
1.90E-06
2.10E-06
2.30E-06
2.50E-06
-500 Time / s
Figure 7 –Selected Time-Velocity Profiles for HNS
7. CONCLUSION We have demonstrated that our PDV system is suitable for non-contact velocity measurements to measure the early time velocity of explosively driven thin foils, allowing us to determine material properties of the explosive donor, in this particular case pressure. The values calculated for the RP-80 detonator agree with published values for PBX-9407 to within 5%. A range of future work has been identified to improve the fidelity of the measurement system including improvement of equation of state data, a deeper understanding of the cause and mitigation techniques of multiply-doppler shifted signals, and experimental work to reduce the effect of alignment issues upon the performance of the system.
8. ACKNOWLEDGEMENTS The authors would like to thank Andrew Stoodley and Sarah Knowles for experimental assistance, Andrew Critchley, Ed Price and Martin Philpot for technical discussions and advice, along with Steven Clarke, Adrian Akinci (both Los Alamos National Laboratory, US) and Ted Strand (Lawrence Livermore National Laboratory, US) for technical discussions and advice.
9. REFERENCES 1. 2.
O. T. Strand, V. Berzins, D. R. Goosman, W. W. Kuhlow, P. D. Sargris, and T. L. Whitworth, Optical Technologies for Arming, Safing, Fuzing and Firing II Proc of SPIE Vol 6287 (2005). M. D. Bowden and M. P. Maisey, Optical Technologies for Arming, Safing, Fuzing and Firing II Proc of SPIE Vol 6287 (2007).
Proc. of SPIE Vol. 7070 70700P-8
3. 4. 5.
E. J. Welle, K. J. Flemming, and S. K. Marley, Optical Technologies for Arming, Safing, Fuzing and Firing II Proc of SPIE Vol 6287 (2006). S. G. Goveas, J. C. F. Millett, N. K. Bourne, and I. Knapp, (2006). B. M. Dobratz and P. C. Crawford, LLNL Explosives Handbook - Properties of Chemical Explosives and Explosive Simulants (Lawrence Livermore National Laboratory, 1985).
Proc. of SPIE Vol. 7070 70700P-9
