a. van der steen- shock wave interaction with composite materials

8/3/2019 A. van der Steen- Shock Wave Interaction with Composite Materials

1/12

JOURNA L DE PHYSIQUE IVColloque C4, supp l6ment au Journal de Physique 111,Volume 5,mai 1995Shock Wave Interaction with Composite MaterialsA. van der SteenTNO Prins Maurits Laboratory, Lange Kleiweg 137, P.O . Box 45, 2280 AA Rijswijk, The Netherlands

Abstract:A large number of physical and chemical parameters determine the sensitivity ofexplosives. Special emphasis is placed on the influence of the crystal quality on the initiationprocess. First the crystal size is considered for m onomodal and bimodal distributions. It seems thatthe Lee-Tarver model is not capable to explain all the experimental observations. Other parame terslike the crystal shape, the smoothness of th e surface of the crystal, internal defects and dislocations(like voids), and the effect o f a binder in a PBX are also discussed. From the experimental resultsknown up to now it is clear that an optimised crystal quality can help to obtain less sensitiveexplosives. For this the study of the crystallisation process is a key item.

1. INTRODUCTIONThe sensitivity of explosives is like a balance. On one side the explosive has t o be as insensitive as possibleto enhance the safety and on th e other side the h nctio ning of explosives will be negatively effected if thesensitivity is too low. Fo r those reasons sensitivity has always been a topic in the detonation community.To describe the sensitivity of explosives is very complicated. First of all it is an intrinsic property of th eexplosive molecule itself. Enorm ous progress is being made in relating the chemical structure t o the sensi-tivity. While Karnlet related sensitivity to th e comp osition of organic explosives, later Delpuech and Odiotlinked the sensitivity t o the properties of th e excited electronic state of th e molecule [I]. In this way a rela-tively g ood estimate of the sensitivity of a new molecule can be obtained before it has ever been sensitised.On a larger scale the physical prope rties of the crystal lattice play a role in th e sensitivity. Coffey andArmstrong for example relate the m echanical properties o f the lattice to the impac t sensitivity 121. Theyassume that a pile-up o f phonons will occur on dislocations in the la ttice and will act a s ignition sites.Many investigators have related the crystal size to the sensitivity. Mou lard has shown the importance ofthe crystal size in relation to the shock sensitivity of PB X's [ 3 ] . Spear has investigated the shock sensitivityo f RDX as a function of the crystal size [4]. Amore theoretical approach is given by the Lee-Tarver modeltha t describes the ignition and pressure build-up during shock initiation assuming spherical particles withone specific diameter.How ever, crystals never behave ideally and m ost of the time th e imperfections in the crystals determinephysical prope rties such as th e shock sensitivity. Just as for liquid explosives that can be sensitised by airbubbles, the sensitivity of solid explosives depends strongly on the number and type of dislocations andimperfections in the crystal lattice. Mishra was one of the first who showed experimentally that the impactsensitivity ofRDX depends on the solvent used for crystallisation [5]. The different solvents resulted in

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jp4:1995409
http://www.edpsciences.org/http://dx.doi.org/10.1051/jp4:1995409http://dx.doi.org/10.1051/jp4:1995409http://www.edpsciences.org/


2/12

C4-108 JOURNAL DE PHYSIQUE IVdifferent crystal shapes. Also van der S teen et al. have shown the influence of the crystal shape on theshock sensitivity [6] .They have shown that a decrease o f the critical initiation pressure of more than 1 GPacan be reached if speroidized crystals with smoothed surfaces are used Borne has shown that theconcentration of voids in RDX crystals effects the sensitivity [ 7 ] . The small number of experimentalevidence is contrary to the large number of theoretical models describing the creation o f hot-spots.A last but certainly not least important contribution comes fiom the state in which the explosive is applied.Large single crystals are not used in w arheads o r boosters. All kinds of different processing techniques areused and processing aids are added to make th e explosive applicable: pressing, casting, extruding, mixing(like CompB ), addition of waxes o r polymers, plasticizers, etc .. Pressed explosives with a m inimum o fvoids between the binder and the ex plosive have a significantly higher sensitivity than cast-curedexplosives. The interaction between the binder and the explosive contributes also to the sensitivity.According to Swallone the mechanical properties of the po lymer influence the sensitivity of the PB X [S].This is also suggested by Sh edelbauer [9] .Others have investigated the influence of the impedancemismatch of the binder and the explosive to the sensitivity. Schrader et al. have shown that thedecomposition of the polymer followed with catalytic decom position of the exp losive can increase thesensitivity of AP-based rocket propellants [I 01.The above show s that there is a large range of interest from the fundamental level of understanding for theexplosive molecule until the a pplication of a PB X in a warhead or a s a booster. In all stages we have tounderstand the nature of the sensitivity of explosive. At T NO -PML w e are trying to understand andcontrol the sensitivity of RD X-based PB X's by controlling the quality of the R DX crystals. In that way wetry t o understand the role of the crystal size, crystal defects and im perfections, crystal interaction with thebinder and with other crystals, etc. However it seem s that each question being solved generates more newquestions. In the follow ing chapters a series of examples is given on different factors effecting thesensitivity of ex plosives.2. EXPLOSIVE AN D TEST METHODMo st experiments are carried out on cast-cured HTPB -based PB X's with RDX as a high explosive. Thesolid loading of the PBX varies depending on the use of m onomodal (65%) or bimodal ( 85%) RDX.Standard procedures and binders are used to cast-cure the PB X [6]

Pressure (GPa)Figure 1: Schem atic representation of the initiation distance of an explosive consisting of coarse o r finecrystals, as a function of th e initiation pressure.


3/12

The sh ock sensitivity of the PB X's is determined with a gap test [6]. It has a diameter of 50 mm, exocireas a donor explosive (an RDX-wax composition) and a PMMA gap. U sing an optical streak cam era thedistance between the PMMA -acceptor interface and the place the detonation wave emerges from the oute rsurface of the charge is m easured. This initiation distance is slightly different as measured on th e centralaxis of the set-up but this test method allows fast exp eriments. The initiation distance is measured as afunction of the shock pressure in the PMMA at the PM MA -explosive interface. The initiation distance isconsidered to be a go od d escription of the shock sensitivity of the cha rge.3.MONOMODALRDXGenerally it is accep ted that the critical initiation pressure becom es higher if the crystal size decreases.Schem atically this is illus trated in F igure I . It is assumed that at relatively low pressures the com positionwith the coarser crystals is more sensitive (gives shorter initiation d istances) because those crystals aremore easily ignited than the finer crystal. In o ther wo rds the ign ition process is the controlling step. Athigher pressures the grow th o f the reaction is controlling and the larger surface area of finer crystals causesa higher sensitivity. This crossing of the sensitivity as a function of the initiation pressure has been found byseveral investigators. Table I lists the particle sizes of four b atches o f RDX with different crystal sizes(two co arse and two fine batches).Table 1 Ave rage crystal size (50%), 10% and 9 0% limits of the crystal size distribution and specificsurface of two fine and two c oarse RDX batches.

50% 10%-90% Specific surfaceur n urn m2/cm3

In Figure 2 the initiation distances for four PBX 's with a solid loading of 65% R DX are given as a fincti onof the initiation pressure h o crossing of the sensitivities is observed at higher pressures rh e largestcrystals (C2) are not the most sens itive but the sensitivity lies in betw een the sensitivitv of the F2 and C1crystals.

3.5 4.0 4.5 5.0 5.5 6.0 6.5Pressure (GPa)

Figure 2 Initiation distances as a h n c t ~ o n f the initiation pressure for four PB X's with different RDXc r y s t a l s . s e e ~a b l e 1 = C l , D = C 2 , A = F 2 a n d + = F l


4/12

C4-110 JOURNALDE P H Y S I Q U E I V

Crossing of the sensitivities is observed when t he initiation distances are determined w ith a so-calledelectric gun. In that case the exp losives are initiated by the impact of a thin kapton flyer. For thin flyers therare-faction waves extinguish the shock waves and therefore much higher pressures are needed t o initiatethe explosive. Tw o different thicknesses of flyers (125 ym and 250 pm) are used and the thicker the flyerthe more the shock wave will resemble the initiation wave for th e gap test. Results are presentedschematically in Figu re 3. Contrary t o the gap test the coarse crystals are the most sensitive for a flyerthickness of 125 pm. The critical pressure for the fine is about 21 GPa and fo r the coarse 15 GPa. For aflyer of 250 pm these critical pressures lie much closer toge ther although the fine is still the most sensitive(10 and 12 GPa). As has been ob served already in Figure 2 the situation is reversed for the gap testexpe riments although the difference is still very small (about 4 and 6 GPa). M ore details can be found inreference [I 11.

Pressure (GPa)Figure 3: Schematic representation of the initiation distances as a function of the initiation pressure forPBX's with fine (----)and coarse (-) RDX for three different impulses. A: gap test, B: flyer 250 ym, C :flyer 125 pm .These results for mon omodalRDX indicate that in general the model of Figure 1 holds for the sensitivityof explosives. Deviations occur as is demonstrated in Figure 2 and can be found in [ l ]. Also the length ofthe p ressure pulse has to b e taken into accoun t. The influence of the solid loading, all presentedexperimen ts have a solid loading of 65%,on the shock initiation process has t o be clarified further. Itseems that the application of different thicknesses of flyers in the electric gun offers a very simple and fasttechnique to study the initiation of explosives as a function of the length (impulse) of the shock w ave.4. BIMODAL RDXBimodal mixtures of RD X are applied to increase the amount of high explosive in the PB X. T o explain thesensitivity for bimodal m ixtures becomes however more complicated. Important param eters are - again -the cry stal sizes and the ratio of the coarse and fine fraction applied.In Figure 4 the sensitivities of tw o P BX 's w ith bimodal RDX are comp ared. The coarse fraction has inboth cases the sam e diameter (370 pm). T he fine fraction is about 20 pm fo r one sample and 55 pm for theother. T he coarse fine ratio is 64/36. It shows that th e PBX with the smallest crystals has a critical pressurethat is about 0.5 GP a lower than for the larger crystals.


5/12

cD0C([I.-V ).--0s0.-.-([I.-C.-

Pressure (GPa)Figure 4 initiation distances as a function of the initiation pressure for a PBX with a fine fraction of 20 pm( + 1and a tine fraction of 55 ym (a).The coarse fraction is the same for both RDX's.The same is observed by varying the ratio coarselfine in the PBX. The results for three coarselfine ratio'szre shonn in Figure 5. In this case also the PBX with the largest amount of fine crystals is the mostsens~tive

40

r([I5 20-m

Pressure (GPa)Figure 3 . lnitiation distance as a function of initiation pressure for PBX's with different ratio's of coarse2nd fineRDX. 0 )C/F= 59/41,.) C/F = 64/36, +) C/F = 76/24.Both observations seem to be contradictory to the model shown in Figure 1. It assumes that (at relativelylow pressures, and that is the case in these examples) the smaller the crystals the less sensitive is theexplosive. The reversal in the bimodal samples is caused by the presence of the coarse crystals. If fine andcoarse crystals are both present in a PBX it is most likely that the coarse crystal will be ignited at thelowest (critical) pressures. The subsequent build-up to a detonation reaction will be determined by thespecific surface of the RDX. Since this is considerably higher for the fine crystals it is not unexpected that


6/12

C4-112 JOURNALDE PHYSIQUE IV

the increase o ft h e amount of fine crystals or the de crease of the crystal size of the fine fraction increasesthe sensitivity of the PBX.These results are only qualitatively and many questions remain. For example it is not yet clear how thedistance between th e coarse and the fine crystals effects the initiation process (how is the energy transfer).Also it is not evident how much o f a coarse fraction is needed to ignite the de composition reaction. Andlast but certainly not leas t: why are the coarse crystals the most sensitive?5. MODELLINGThere are many models available to simulate the shock initiation of explosives. The most well know n arethe Forest Fire and the Lee-T arver m odel. Th e Forest Fire model is very crude and empirical and cannot beapplied to param eter studies as are presented above. The m odel is more suitable for simulating the shockinitiation behaviour of different explosives.The Le e Tarver model is also referred to as the ignition and growth m odel and is founded on theknow ledge that the shock initiation in heterogeneous exp losives is based o n the form ation and subseque ntreaction of the hot spots that are created when the shock pulses compresses the solid. In the model oneignition term and two growth terms are applied. The ignition and grow th terms can help to understandqualitatively the shock initiation process. They could be an indication for the number of defect whereignition starts and how easily a reaction grow s to com pletion. A disadvantage of the model is that it canonly simulate one (sphe rical) particle size. Bimodal m ixtures (two crystal sizes) cannot be simulated Thetransfer of the ignition reaction from the coarse cry stals to the fine crystals is not yet understood.The simulation of monom odal explosives gives som e interesting result [12]. In these simulations theinitiation distance was calculated a s a lknction o f the initiation pressure . In Figures 6 the results are givenfor a variation of the amplitude of the first grov"~1;erm (GI) and the ignitiot: tern1 (I)

Pr e s sur e (G Pa ) Pr e s sur e (GPa !Figure 6: The simulated initiation distance as ~ L I I I C I I C I : of the initiation pressure for different value., :>i rncignition term (left) and the g row th term (right) 111 tilt Lee-Tarver model.


7/12

Figure 7: Scanning Electron Micrographs of a typical RDX crystal (top) and a crystal that has been post-treated (bottom) to obtain a better morphology.


8/12

C4-114 JOURNAL D E PHYSIQUE IVThese results show that the growth terms dominate the initiation process. A variation of 0.05 to 0.3 in thegrowth term results in a much larger variation in the initiation distances than a variation of 10 '~ n theignition term. This seems to be in accordance with the findings for the bimodal mixtures. For thosemixtures it is also found that small variations in the fine material (growth) causes large variations in thesensitivity. On the other hand, these simulations do not explain the results for the monomodal PBX's.Considering the simulations the ignition term plays a minor role and the crossing of the sensitivity curvescannot be explained.6. CRYSTALQUALITYAll results presented above are interpreted with respect to the crystal size and the crystal size distributionwhile the crystals are assumed to be spheres. Unfortunately crystals seldom have the shape of a sphere.This is illustrated in Figure 7 where two Scanning Electron Micrographs (SEM) are given of RDX crystals.Both have about the same diameter. One has the typical structure of an RDX crystal as obtained directlyfrom the recrystallization vessel. The faces of the crystal can be assigned to the different growth directionsof the crystal. The other crystal is more oval shaped and has a very smooth surface. This crystal has beenpost-treated with the solvent ethyl acetate. Its morphology is much more regular and can be defined withonly two parameters.The influence of the morphology of the crystals on the shock sensitivity is shown in Figure 8. PBX's withthe same fine fraction (47pm) but with an angular and not post-treated coarse fraction and a coarsefraction that has been post-treated are compared. The diameter of the coarse fraction is 480 pm before thepost-treatment and 450 pm afterwards. It shows that especially at the initiation threshold the PBX with thesmoother crystal is about 0.4 GPa less sensitive than the angular crystals. The differences are lesspronounced for the higher pressures. Unfortunately the sensitivity for monomodal PBX's has not (yet)been measured.

Pressure (G Pa)Figure 8: Initiation distances as a hnction of initiation pressure for a PBX with +) angular crystals anda)oval shaped crystals as the coarse fraction.These results show that not only the particle size plays a role in the sensitivity but also the morphology ofthe crystals. There are several possible explanations for the decreased sensitivity:


9/12

Figure 9:Light m icroscopy o f RDX crystals. Empty and filled voids are visible.

Figure 10: SEM picture of a cross section o f an RDX crystal. A void and different crystals orientations arevisible.


10/12

C4- 116 J O U RN A L DE PHYSIQUE IV- The tap density with the spheroidized particles is higher. The increased average distance between thecrystals could caus e the decreased sensitivity.- A better contact betw een the binder and the c rystal is possible for sm oother crystals, voids cou ld still bepresent at the interface for the angular crystals and cause ignition.- Angular crystals are more sensitive than speroidized crystals because ignition occurs at the corners of thecrystals.The m orphology tells something about the "outside" of the crystals. Figure 9 show s the "inside" of RDXcrystals as can be seen with a light microscope. Voids, som etimes filled with solvent, of all different sizesand shape s can be observed. In Figure 10a c ross section is given of on e crystal. An elongated void thatwas probably filled with the so lvent is clearly visible. All these im perfections are caused by thecrystallisation process. D epending on the typ e o f crystallisation (cooling, evaporation, e tc.) and thegrowing rate the internal and external quality of the crystal will be influenced. Bo rne has already indicatedthat crystals with a higher density, and therefore a smaller number of voids, have a lower initiationthreshold [7]. Also C han has shown that the sensitivity of CL20 can be less by improving th e crystallisationmethod [13]. It is assumed that these voids act as ignition sites. Void collapse, adiabatic heating, plasticflow, etc. are generally considered a s the processes to form hot-spots.Cutting crystals and etching the surfaces yields even more information as ca n be seen in Figure 11.Thedifferent grow th directions and the interfaces in between are clearly visible. It can not be excluded that atthese interfaces with a relatively high num ber o f dislocations a pile-up o f energy occurs as has bee ndescribed by Coffey and Armstrong [ 2 ] .Crystallisation studies and detailed techniques to follow theinitiation (or better the ignition and grow th) p rocesses have to reveal these questions in the future.

Figure 11 Cross section of an RDX crystal. The g rowth sectors and a seed crystal in the cen tre are clearlyvisible.7. CONCLUSIONS

Most of the time fundam ental studies on sensitivity refer to ideal system s as on e molecule or a singlecrystal. It is impossible to detonate a m olecule and from the studies of Dick on PETN [I41 w e know that itis very hard to detonate a cry stal. Recently at Lawrence L ivermore National Laboratory it has been shownthat initiation in single crystals only occurs after the shock w ave ha s reflected against an interface.


11/12

This paper h as tried to give some mo re insight in physical param eters affecting the sensitivity. The resultsare mainly qualitative and still many questions remain. The most important are:- How are (coarse) crystals ignited?- How does the energy transfer between crystals take place?- How far can we decrease the sensitivity of an RDX based PBX by optimising the crystal quality?It is this last question that is the most im portan t. Crystallisation of RD X is getting m ore attention in thedetonation community at the moment and h op ehl ly there will be a few answers in the coming years.AcknowledgementsThe research project was carried out under assignment A93KL447 for the R oyal Dutch Army. The authorwould like to thank all his colleagues who contributed to this work: W illem Duvalois, Aat Hordijk,Antoine van der Heijden, Wim Prinse, Murk van Rooijen, Benoit Coleau, Jos Mul, Rene Oostdam , RiesVerbeek en Ed de Jong.ReferencesI . A. Delpuech , Initiation des explosifs a l'echelle moleculaire, Ap proches Microscopique etMacroscop ique des Detonations (Les editions de Physique, Megeve, 1987), 353 (1987)A.E. Delpuech, The use of time-resolved spectrometries in the study of initiation of explosives atmolecular level, Proceedings Ninth Symposium (International) on D etonation, Portland, 172 (1989)S. Odiot, Mechanisms of detonation in molecular crystals: a review, Approches Microscopique etMacroscop ique des Detonations (Les editions de Physique, Megeve, 1987), 225 (1987)S. Odiot and M . Peyrard, Why does an explosive explode?, Approches M icroscopique et

Macroscop ique des Detonations (Les editions de Physique, Megeve, 1987), 393 (1987)2. C.S . Coffey, Energ y localization and the initiation and detonation of crystalline explos ives by shock orimpact, Proceedings W orkshop Desensitization of Explosives and Propellants, TN O Prins MauritsLaboratory, Rijswijk, Th e Netherlands, November 199 1R.W. Arm strong, Dislocation mechanics aspects of d esensitization to impact or shock d eformations,Proceedings Wo rkshop D esensitization of Ex plosives and Propellants, TNO Prins Maurits Laboratory,Rijswijk, The Netherlands, November 1991C.S. Coffey and R .W A rmstrong, This worksho p3. H. Mo ulard, J W. Kury and A D eclos, The effect of RDX particle size on the shock sensitivity of castPBX formulations, Proceedings Eight Symposium (International) on Detonation, Albuquerque, 9 02(1 985)H. Mou lard, Particular aspects o f the explosive particle size effect on shock sensitivity of cast PB Xformulations, Proceedings Ninth Symposium (International) on Detonation, Portland, 18 (1989)

4. R J. Spear and V. Nanut, Mechanism of and particle size effects on shock sensitivity of heterogeneouspressed explosives: preliminary assessment of binderless RDX in h z e trains, DOD Materials ResearchLaboratories, M elbourne Australia, MRL-R-1077 (1987)5. I.B . Mishra and L.J. Vande Kiefi, Novel approach to insensitive explosives, 19th International AnnualConference of ICT 1988, Karlsruhe, 25-1 (1988)6. A.C. van der S teen, H J. Verbeek and J .J. Meulenbmgg e, Influence of RDX crystal shape on the shocksensitivity of PBXes, P roceedings Ninth Symposium (International) on Detonation, P ortland, 83(1989)A.C van der Steen and E. Skjold, RDX particle shape and the sensitivity of PBX es, 1990 Joint

GovernmentJIndustry Symposium on Insensitive Munitions Technology (Addendum Proceedings),White Oak, 235 (1990)


12/12

C4-118 JOURNALDE PHYSIQUE IVA.C. van der Steen, E.G. de Jong, W.C. Prinse, A.C. Hordijk and W Duvalois, Crystal quality andshock sensitivity of RDX-based PBXes, Tenth Symposium (Intemational) on Detonation, Boston,(1993) to be published7. L. Borne, Influence of intragranular cavities ofRDX particle batches on the sensitivity of cast plasticbonded explosives, Proceedings Tenth Symposium (international) on Detonation, Boston, (1993) to bepublished8. G.M. Swallone and J. Field, Proc. R. Soc. Lond. A,,Vol. 379,389 (1982)9. F. Schedlbauer and A. Kretschmer, The influence of particle size and mechanical properties on thesensitivity of high explosive charges (PBX), Proceedings Tenth Symposium (Intemational) onDetonation, Boston, (1993) to be published10. M.A. Schrader, M.W. Leeuw and A.C. van der Steen, The heat sensitivity of solid propellants,AGARD Conference Proceedings No. 367, Hazard studies for solid rocket motors, Lisse, 19 (1984)11. W. Prinse, M. van Rooijen, B. Coleau, J. Mul, R. Verbeek and A. van der Steen, Shock sensitivitytesting of explosives with the gap test and thin flyer impact test, Europyro 1995, to be published12. R. Verbeek and A. van der Steen, The simulation of shock initiation of less sensitive explosives usingthe hydrocode autodyn, Proceedings Tenth Symposium (international) on Detonation, Boston, (1993)to be published13. M. Chan, NAWC, China Lake, Private communications14. J. Dick, This workshop

a. van der steen- shock wave interaction with composite materials

Documents