r. yazici and d. kalyon- microstrain and defect analysis of cl-20 crystals by novel x-ray methods

8/3/2019 R. Yazici and D. Kalyon- Microstrain and Defect Analysis of CL-20 Crystals by Novel X-Ray Methods

1/16

Energetic Materials, 23: 43-58, 2005Copyright Taylor & Francis Inc.ISSN: 0737-0652 printDOl: 10.1080/07370650590920287

WTaylor &Francis~ Taylor&FrandsGroup

Microstrain and Defect Analysis ofCL-20 Crystals by Novel X-Ray

MethodsR. YAZICID.KALYONHighly Filled-Materials Institute, Stevens Institute ofTechnology, Hoboken, NJ, USA

Microstrains and defects are introduced during synthesisand crystal-growth stages of energetic particles and increaseduring processing stages such as grinding, mixing, andextrusion. The detection and quantification of these micro-strains and defects in a given particle population is a diffi-cult task that requires highly sensitive techniques. In thisstudy a novel X-my diffraction technique (XAPS) basedon simultaneous rocking-curve analysis of individual parti-cles was s'uccessfully applied to CL-20 powders. The effectsof synthesis, grinding, and static loads on the extent ofmicrostrain and defect development in CL-20 particleswere quantitatively determined as frequency versus half-width of rocking curves. The greater half-width valuesobserved for the samples subjected to grinding and staticloads indicated greater microstraisi and defect densityin comparison to the as-received samples of CL-20. Itmay be possible to relate the findings of such analysis tocombustion calculations for energetic particles in generaland to CL-20 particles in particular.Keywords: CL-20, X-ray diffraction

Address correspondence to D. Kalyon, Highly Filled MaterialsInstitute, Stevens Institute of Technology, Castle Point, Hoboken,NJ 07030, USA. E-mail: [email protected]

4 3
mailto:[email protected]:[email protected]


2/16

4 4 R. Yazici and D. KalyonIntroductionCL-20 or hexanitrohexaazaisowurtzitane is a relatively newenergetic material that exhibits higher density and heat of for-mation than RDX and HMX, the conventional energetic nitra-mines. These superior properties are due to its unique cagedstructure (isowertzitane) with characteristics of high density,strained ring, and high branching, respectively. Superior pro-ducts already have been demonstrated by replacing the conven-tional nitramines with CL-20. These include solid rocketpropellants, gun propellants, and shaped charge explosives.

The combustive properties of energetic materials such asCL-20 depend on their composition and microstructural charac-teristics such as molecular and crystal structure and particlesize distribution [1]. Although some of these characteristicshave been fairly well understood, the effects of more subtle fea-tures, such as molecular and crystal defects on the combustiveproperties ofenergetics, have not been investigated. The presenceof microdefects such as misaligned domains and dislocations insingle crystallites (particles) is known to increase their chemicalactivity, by straining the crystal lattice and making more reactionsites available at the domain boundaries and dislocation sites[2, 3]. The strain energy of a dislocation is about 8eV for eachatom plane threaded by the dislocation, while the core energy isin the order of O.5eV per atom plane [4]. This large positivestain energy means that the free energy of a crystal is increasedby the introduction of a dislocation. Processing practices in crys-tal growth, such as solution stoichiometry, temperature, cata-lysts, impurities, and even mechanical vibrations, could inducedefective and partially amorphous structures that would not beapparent to most conventional characterization techniques.Highly sensitive and reliable techniques are required to determinequantitatively the extent ofthese defects in energetics to optimizetheir processing parameters and energetic properties.

The goal of this project was to investigate the applicability ofnovel X-ray methods to quantitative determination of crystalimperfections, that is, microstrains and defects in CL-20 particles.For this purpose a very sensitive X-ray diffraction technique


3/16

Defect Analysis of CL-20 4 5(XAPS), which isbased on the simultaneous rocking-curve analysisof individual particles [2,3,5,6]was applied to CL-20 powders.Experimental ProceduresMaterialsThe materials used in this study were two CL-20 powder sam-ples manufactured by Thiokol Corporation and provided byONR. The two powder samples received had average grainsize of 130 microns (coarse) and 6 microns (fine), respectively.

Some of the coarse (130 urn) CL-20 powder samples weresubject to deformation under static loads to simulate processingconditions. The applied static loads ranged from 600 to 2000 psiand were held for 10 minutes. The deformation experimentswere carried out at two temperatures of 25 and 90C.

In addition to the as-received powders, new CL-20 crystalswere also grown with different particle morphologies. Thesenewcrystals were grown and used, together with the as-receivedpowders, for crystal structure analysis. All CL-20 samples werekept in sealed cups with neutralizing water suspension, exceptduring testing. The sealed cups were placed in steel bomb enclo-sures during storage.

The Fourier transform infrared (FTIR) analysis of the pow-ders was carried out, and the purity of the epsilon, s, polymorphwas established. The presence of the doublet between 8.20 cm-1and 850cm-1 (and the lack of a single peak) establishes thepurity of the s polymorph. e-CL-20 is the most stable andenergetically favorable polymorph.

The differential scanning calorimeter (DSC) scans of CL-20powder were also carried out. The CL-20 crystals exhibit anexothermic transition (melting) at 168C. The heat of fusionis 3.46 call g.

Characterization Methods: X-ray Analyzerfor Particles (XAPS)In this study the X-ray Analyzer for Particles (XAPS)method, which is based on simultaneous rocking-curve analysis


4/16

4 6 R. Yazici and D. Kalyonof individual particles, was employed to determine quantita-tively the microstrain and defect distributions in CL-20 ener-getic crystals. XAPS is a suitable method for measuring thedegree of imperfection and lattice misalignment of individualcrystallite through X-ray rocking-curve half-width measure-ments [2,3,5,6].Briefly, in the XAPS rocking-curve technique the powdersample is irradiated with a crystal monochromatized parallelX-ray beam (Figure 1). Each reflecting crystalline particle actsindependently as the second crystal or test crystal of a doublecrystal diffractometer. Those particles, which are in reflectingpositions, give rise to individual diffraction spots along theDebye arc (Figure 1). Each diffracting spot emanating from

PSD

crystal Rotation(m in ute s o f o re )

Figure 1. Principle of operation of XAPS particle rocking-curve analyzer.


5/16

Defect Analysis of CL-20 47the powder sample is registered as a function of location andintensity by a linear position-sensitive detector (PSD), whichis aligned parallel to the Debye arc. After a short exposure theX-ray intensity spectra are collected by a multichannel analyzer(MeA) and stored by a computer (Figure 1). When the datatransmission is completed, the sample is rotated automaticallyby a chosen increment (1 minute of arc) before the next exposureis started. This stepwise rocking process is repeated until theentire angular range of the grain reflections are recorded (20minutes of arc). The rocking curve for each particle as wellas the rocking curves for the entire particle population arecomputed by analyzing the entire reflection matrix. The rock-ing-curve half-width (full width at half maximum intensity) pro-vides a measure of microstrain and angular lattice misalignmentinduced by dislocations within the diffracting grain (Figure 2).

The width at half the maximum intensity of the rockingcurve (half-width) or ( 1 3 ) provides a measure of the angular lat-tice misalignment of the crystallite. Thus, the half-width ( 1 3 )values provide information about the local densities of theaccumulated excess dislocations of one sign, that is, or sameorientation, with respect to lattice directions.

Some illustrations of the relationships between the distribu-tion of excess dislocations and the resultant rocking-curveprofiles are schematically shown in Figure 2. If the profile issmooth, a homogeneous distribution of excess dislocationswould, as marked in "deformed", exist. If, on the other hand,a multipeaked profile is produced, the excess dislocations areheterogeneously distributed, giving rise to distinct lattice tiltsas would be encountered, for example, in grains with subgrainboundaries, as marked "polygonized." The angle between thesubpeaks of such rocking curves then would represent therelative misorientation of adjacent lattice domains or subgrains,whereas the width ofwell-resolved subpeaks indicate the degreeof internal distortion in these domains. Annealed grains withlow excess dislocation densities give rise to smooth rockingcurves with small 1 3 values as marked in "annealed."

Assuming that the excess dislocations are randomlylocated in the grains and thus can be represented by a Gaussian


6/16

4 8 R. Yazici and D. Kalyon

a

b

Figure 2. Rocking-curve and particle defect structure charac-teristics (schematic). (a) Construction of rocking curves.(b) State of the related particle defect structures.

distribution, the excess dislocation density is determined byD= fj2 /9 b2, where b is the magnitude of the Burgers vector. Ifthe excess dislocations are aligned in thesubgrain boundary andthe tilt angle of adjacent subgrains, e ', is measured, the excessdislocation density in the subgrain boundary DSB = e ' /3bt,where t is the subgrain size [7]. If resolvable, the subgrain sizecan be measured from the images ofX-ray reflection topographs.

The microstrains, that is, the residual strains (and stresses)in individual crystallites or particles, induced by the excess dis-locations can also be calculated from the excess dislocation den-sity D values, provided that the exact natures and distributionsof the excess dislocations are known and the crystal structureand physical constants are available. The microstresses asso-ciated with a dislocation are biaxial in nature and vary inverselywith distance from the dislocation core. Although the stresses atthe core reach to infinite values, more reasonable stresses areinduced over the average body of the crystal [4].


7/16

Defect Analysis of CL-20A Picker four-circle goniometer was utilized for XAPS

experiments. CuKt radiation monochromatized with a Si (111)crystal at 35 kV and 20 rnA was used. Five to 15 minutes timedexposures were applied at each rocking step. Three hundred to1000 individual particles were evaluated from each sample.

Otber MethodsWide-Angle X-my Diffractometry. Wide-Angle X-ray Diffrac-tometry (WA-XRD) was utilized for determination of thedegree of crystallinity of the CL-20 powders. Computer-aideddeconvolution methods with Images software were applied todetermine the relative contributions of the amorphous and crys-talline portions of the samples from their diffraction patterns.The unit used was the Rigaku DXR-3000 diffractometersystem. Culv, radiation at 40kV and 20 rnA and a graphitemonochromat ion with focusing geometry were used for all WA-XRD runs.

Kappa Diffractometry. Kappa diffractometry was used tocarry out single-crystal analysis and to determine the crystalstructure parameters. This work was carried out because thecrystal structure parameters were required for dislocation den-sity calculations. A Nonius Difractis 586 X-ray unit andCAD4 and NRC software packages were used for single-crystalanalysis.

Scanning Electron Microscopy. A scanning electron micro-scope (SEM) was used to examine the particle morphology. AJEOL 840 SEM with EDX was utilized for this purpose.

ResultsCrystal Structure AnalysisThe crystal structure of E-CL-20 was determined using one ofthe crystals grown from solution. The results of these prelimin-ary studies indicated the following:


8/16

50 R. Yazici and D. KalyonCrystal systemUnit cell parameters:

Unit cell volume:Number of molecules

Monoclinica = .84 (2) Ab = 12.50 Ac=13.36 AtX =Y =900

f J = 106.8V = 1414.3 A _ 3z=4

per unit cell:X-ray density: d =2.04g/cc

from d =zMw/ ( V - : Y A ) with molecular weight equal to 438.24.XAPS CurvesTypical XAPS rocking-curve data, that is, intensity versusposition versus rotation, as stored in the computer memory isshown in Figure 3. The data shown are only a small portionof a single rocking-curve run, where 2/3rd of the rotations and2/3rd of the channels were omitted for simplification. Theanalysis program locates the intensity from a specific particleat a specific MCA channel(s) at a specific rotation step andfollows its variation at the same channel(s) in each rotation andobtains the rocking-curve distributing profile before calculatingthe width at half maximum. This procedure is carried outfor each reflecting grain. Twenty to 100 grains are analyzed ineach run.

Coarse versus Fine (Ground) CL-20PowdersThe results of the XAPS rocking-curve analysis of the as-received coarse (Figure 4) and fine CL-20 powder samples areshown in Figures 5 and 6, respectively. In these figures the sta-tistical distribution of the half-width values of the individualparticles is exhibited as the particle frequency observed foreach half-width value. In these graphs the larger half-widthvalues emanate from those particles, which constitute highmicrostrains and defects, and, conversely, the smaller half-width values emanate from those particles, which exhibit a


9/16

Defect Analysis of CL-20 51> C A P S R C X:K JN GC lR V E O F A S IN G .E C R ' Y S J ' A I . l . J ' J ECHON\EL.pam~ V IEW OF nE3 D Q !.\T A

Figure 3. Rocking-curve step scan patterns.

smaller number of defects and low microstrains. The rocking-curve half-width mean values of the fine powder (5.1 minutesof arc) were measurably larger than that of the coarse powder(4.2 minutes). Also, the maximum half-width value observedin the fine powder (38 minutes) was considerably higher thanthe coarse powder (18 minutes).

Assuming that the excess dislocations are randomly locatedin the grains, the excess dislocation density D = f 3 2 / 9 b 2 can be


10/16

52 R. Yazici and D. Kalyon

Figure 4. As-received coarse CL-20 powder at 300 x magni-fication.

035 -r-------------------------,0.3 Half-width

Mean: 4.2 minutesStandard deviation 2.2 minutes0.25';:!l'l5 0.2

cci0.15. .~. .E 0.1:::;z;o 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40

Half-width, minutes of arcFigure 5. Particle defect half-width distribution by XAPS foras-received coarse CL-20.


11/16

Defect Analysis of CL-20 530.45,-----------------------,0.4

Half-widthMean: 5.1 minutesStandard deviation: 5.2 minutes

3 5 7 9 11131517192123252729313335373941Half-width, minutes of arc

Figure 6. Particle defect half-width distribution by XAPS foras-received fine CL-20.

calculated for each case. Taking the Burgers vector b=8.84A.,then the excess dislocation density values can be determined.For the as-received coarse powder the excess dislocation densityis 2.1 x 107jcm2, and for the as-received fine (ground) powder itis 3.1 x 107jcm2.

A fine CL-20 batch was obtained by grinding of the as-grown particles. The higher half-width values of the fineCL-20 powder, therefore, could be taken as a measure of theeffect of the grinding process on the crystal defects. Thus,the applied grinding process appears to have increased the lat-tice imperfections and defects in the CL-20 particles. TheXAPS rocking-curve technique, therefore, can be utilized todetermine quantitatively the effects of grinding on crystaldefects.

The typical results of the WA-XRD scans are shown inFigures 7 and 8 for the coarse and fine powders, respectively.According to the deconvolution analysis carried out on thesepatterns, the degree of crystallinity of the CL-20 powders wasin the range of 90-93 %. The difference in the degree of crystal-linity of the coarse and fine powders was negligible.


12/16

5 4 R. Yazici and D. Kalyon

Figure 7. Wide-angle X-ray diffraction pattern of coarse CL-20powder.

CL-20 Powders Subjected to DeformationAs noted earlier, specimens of the coarse CL-20 powderwere subjected further to uniaxial compression (Figure 9),using a modified compression-molding machine to simulate the

Figure 8. Wide-angle X-ray diffraction pattern of fine CL-20powder.


13/16

Defect Analysis of CL-20 55

Figure 9. Coarse CL-20 particles fractured under static load.

thermomechanical conditions that they would experience duringfurther processing. During these processes high shear stressesare generated, and the particles generally undergo shear defor-mations. Invariably, it is the excess shear stresses that causethe microplastic deformations in the crystals. Under adequatelyhigh shear stresses, crystals tend to generate dislocations alongtheir slip-planes and shear deform by slip. The "characteristicslip-plane" of a crystal tends to be the lattice plane where theseparation between the atoms/molecules is maximum.

By applying normal loads to randomly oriented crystalliteparticles, "resolved shear stresses" are generated along the"characteristic slip-plane" of each particle. The magnitude ofthe shear stress would vary depending on the orientationangle of the characteristic slip-plane and the characteristic slipdirection of the given particle with respect to the direction ofthe applied normal stress. For the angle between the appliednormal stress direction and the characteristic slip-planenormal being c / J , and the angle between the slip direction and


14/16

56 R. Yazici and D. Kalyonthe slip-plane normal being A , the relationship betweenthe applied normal stress (J'app and the resolved shear stress'rrss (on the characteristic slip system of a single crystal) isgiven by [4J

This shear stress resolved on the characteristic slip-plane and inthe slip direction of a single crystal ranges from zero, for either< p =0 to )_= 0, to maximum, for < p = A = 45. The maximumresolved shear stress is

1'rss(max) =2 (Japp'

Therefore, in the static-load experiments carried out in thiswork, the effective or resolved shear stresses ('rrs.) on individualparticle slip systems ranged from 0 to 1,000 psi for the(J'app =2,000 psi experiment.

The typical results of the XAPS rocking-curve analysis ofthe coarse CL-20 powder (subjected to deformation under astatic load of 2000 psi for 10 minutes time at 90C) are shownin Figure 10. The mean half-width value exhibited by theCL-20 crystals increased upon being subjected to pressure (theincrease was from 4.2 minutes of arc of the as-received coarseCL-20 particles to 9.4 minutes of arc upon pressurization).Upon pressurization the breadth of the half-width valuesincreased substantially with the maximum half-width valuereaching 40 minutes of arc upon pressurization. The meanexcess dislocation density upon pressurization became1.1 x 1Q8jcm2, with a maximum value of 1.9 x 109/cm2.

The significant increase in the mean half-width value uponpressurization primarily arises from the generation of a rela-tively small number of particles (about 5% by number),which exhibit half-width values in the 20-40 minutes range. Ifthese relatively few particles that exhibit such high half-widthvalues are removed from the analysis, the magnitude of thehalf-width difference between the particles subjected to the2000 psi load versus those that were not reduces considerably.


15/16

Defect Analysis of CL-20 572000 psi at90 C

O . z r - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - .Half-widthMean: 9.4 minutesStandard deviation: 5.9 minutes

~ 0213'Ei I : JC.'I s 0.15co'Q 0.1

' " "jE 0.05Zo 2 4 6 8 10 12 14 16 1& 20 22 24 15 28 30 32 34 36 33 40

Half-width, minutes of arcFigure 10. Particle defect half-width distribution by XAPSfor CL-20 subjected to 2000 psi.The mean half-width of the CL-20 particles subjected to pres-surization becomes 6.44 when the particles with half-widthvalues in the 20-40 range are not included in the analysis.Thus, upon axial compression and the associated deformationthe increase in the defect density of the CL-20 powder is asso-ciated with the introduction of microstrains and defects to arelatively small number of particles, which presumably carrythe load during axial compression of the powder bed.

ConclusionsThe applicability of the novel X-ray method XAPS to the quan-titative determination of crystal defects and excess dislocationdensity of the CL-20 crystals has been demonstrated.

AcknowledgmentsThis project was funded by BMDO as managed by ONR underproject No. NOOO-14-96-1-1202.We are grateful for this support.We also thank Dr. Richard S. Miller for his contributions, input,and help during the course of this project.


16/16

58 R. Yaz'ici and D. KalyonReferences[1] Urbanski, T. 1990. Chemistry and Technology of Explosives, vol. 4,

New York: Pergamon Press.[2] Pangborn, R. N., R. Yazici, T. Tsakalakos, S.Weissmann, and 1. R.

Kramer. 1979. Determination of prefracture damage in fatiguedand stress corroded materials by X-ray double-crystal diffractome-try. NBS Special Publication 567, p. 433.

[3] Yazici, R., W. Mayo, T. Takemoto, and S. Weissmann. 1985.1. Appl. Crystallography, 16: 89.

[4] Dieter, G. E. 1986. Mechanical Metallurgy, New York: McGrawHill.

[5] Weissman, S., Z. H. Kalman, J. Chaudhuri, R. Yazici, and W.Mayo. 1982. Residual Stress and Stress Relaxation, ed. Eric Kulaand Volker Weiss, New York: Plenum Publications.

[6] Yazici, R., K. E. Bagnoli, and Y. Bae. 1988. In NondestructiveMonitoring of Materials Properties, F. Holbrook and J. Bussiere(eds.), City: MRS Publications, vol. 142, pp. 65-70.

[7] Hirsch, P. D. 1956. Metal Physics, 6: 282-283.

r. yazici and d. kalyon- microstrain and defect analysis of cl-20 crystals by novel x-ray methods

Documents