AD .

DEFENS;E DOCUMENTATION CENTERFOB

SCIENTIFIC AND TECHNICAL INFORMATION

CAHIEBON STATION, ALEXANDRIA, VIRGINIA

UNCLASSIFIED

RESEARCH DIVISION

EFFECT OF RADIAL DISTANICE FROM HIGH EXPLOSIVE ON

BREAKUP CHARAC TERIS TICS OF SOLID MATERIALS

IN SPHERICAL BOMBLETS

C>Contract DA-18-1O8-4O5-CML'829 I t

~ ~ Special Report 0395-04(16)SP I February 1963 /Copy 3

APR 7 1964

EROJET A

wGNR I

ABSTRACT

J This special report describes an experimental method fordetermining the effect of radial distance from an explosivecharge on the net size reduction of solid particles in aJ spherical CW bomblet. Data showing the effect of positionin bomblet, size of bomblet, agent-to-explosive mass ratioand shock amplification on final breakup are presented.

Methods for chemical analysis of both the inorganic andorganic materials used as agents (simulants) and tracersare presented.

1m |

4,

11r

t 0395-04(16)SP

Page iiiICONTENTS

1. INTRODUCTION ............................. 1

2. METHOD OF DETERMINING RADIAL EFFECTS ......... 2

Z. 1 Experimental Procedure ................... 22. 2 Preliminary Experimental Results With

Inorganic Salts .......................... 5

3. DETAILED EXPERIMENTAL RESULTS ............ 11

3. 1 Effect of Varying Tracer Position ................ 113. 2 Effect of Bomblet Casing Variation . ........... 173.3 Layer Stripping Effects . ......... .......... 263. 4 Shock Amplification Effects .. ............... 26

4. ANALYSIS OF RESULTS ....................... 41

5. CONCLUSIONS AND RECOMMENDATIONS ... o........ 43

REFERENCES ...................... f......44

APPENDIX A ....... .. ...................... .... 45

ACKNOWLEDGEMENT ... ...................... 50

Ia

1

I0395-04(16)sPPage ivI

ILLUSTRATIONS

f Figure Page

1. Spherical Explosive Dissemination Device . ....... 3

Z. Concentric Spherical Devices to Show RadialEffects 6...............................6

3. Variation of Resultant Gross MMD With TracerPositions ............................ 8

4. Plot of Tracer Breakup vs Radial DistanceFrom Charge ........ .......................... 9

5. Silicone Grease Impaction Sampler ................ 10

6. Particle Size ot Tracer vs Radial Distance FromExplosive Charge . ....................... 13

7. Particle-Size Distribution of Starting Material(Anthraquinone) ........ ......................... 15

8. Particle-Size Distribution of Starting Material(p-Aminobenzoic Acid) ....... ................... 16

9. Plot of Tracer Breakup vs Radial Distance FromCharge (Organic) ........ ........................ 19

10. *Variation of Resultant Gross MMD WithTracer Positions (Organic) ..... ................. .20

11. Breakup of Total Salt (Thick Wall) ..... ............ 22

1Z. Breakup of Total Salt (Thin Wall) ................. Z3

13. Tracer Breakup in Various Bomblet Types 24......4

14. Average Tracer Breakup in Various BombletTypes .......... ............................... 25 I

0395-04(16)SPPave v

ILLUSTRATIONSI

T n

15. Amount of Fine Tracer Particles in Thick-j Walled Bomblets .. ..................... 28

16. Particle Size of Salt vs Tracer Position(Series E-5b) ......................... 30

17. Comparison of Layer StTipping With VariableTracer Position (Particle Size of Tracer vsPosition) .......... ............................ 31

18. Shock-Amplification Device (1) ...... .............. 35

19. Shock-Amplification Device (2) ...... .............. 36

2 0. Tracer MMD of Radial-Effect Tests ..... .......... 37

21. Total Salt MMD of Radial-Effect Tests .............. 38

22. Smnall Particles Produced in Radial-Effect Tests 40

23. MMD Total Salt as a Function of Agent-to-Burster Ratio. ......................... 42

I

I

-5m

U

1 0395 -04(16)SPPage vi

I TABLES

f Page

1. Properties of Chrome and Potassium Alums ...... 4

2. Experimental Data for Concentric Sphere Series(Alum) (Group 1) ........................ ..... 7

3. Experimental Data for Concentric Sphere Series(Alum) (Group 2) ............ 0 ...... I ... 12

4. Properties of Anthraquinone and p-AminobenzoicAcid . .......... * a ....... # . . .. ....... . 14

5. Experimental Data for Concentric Sphere Series(Organic) .............................. 18

6. Experimental Data for Concentric Sphere Tests

(Casing Variation) ........................ 21

7. Samples Remaining Airborne at 27 Min ......... 27

8. Experimental Data for Layer Stripping -Concentric Sphere Tests ................... 29

9. Results of Shock-Amplification Experiments ....... 33

10. Small Particles Produced by Pressure-Amplification Tests ...................... 34

III

0395-04(16)SPPage 1

!NT1RODTCTIONj

The dissemination of chemical warfare agents by various explosivedevices was studied by numerous investigators (References I through 7).A spherical device containing the agent surrounding a central explosiveburster charge was chosen for this study because of its simplicity, anduniform geometry. Most investigations to date on the explosive dissem-ination of solid and liquid materials were designed around parametricstudies of initial states and final aerosol cloud measurements. Studiessuch as those conducted by Buck at CRDL (References 4 and 7) and by theStanford Research Institute (Reference 1) have produced infoimation onaerosol particle-size distribution as functions of agent-to-burster weightratios, length-to-diameter ratios (in cylindrical devices), gross size ofbomblet, etc. Some theoretical consideration was given to the mechan-isms involved in the production of aerosols by explosive devices (Ref-erences 5 and 8) but not enough experimental work was conducted to con-firm these theoretical approaches.

In order to pursue further work of either a theoretical or practicalnature in the area of explosive dissemination, experiments must bespecifically designed to study the intervening processes between initialand final states.

In the case of spherical bombs, there are three general regions ofinterest. The first, neglecting the structural members, is the regionclosest to the explosive charge. In this region, both the shock andthermal effects from the explosive detonation products should begreater. Proceeding outward is the intermediate region which issubjected mostly to shock compression, with less thermal effect,and finally, the outer region which is subjected to shock compressionand tensile effects caused by reflections from the outer boundary ofthe device and aerodynamic effects as the material moves out throughI the surrounding atmosphere. The difference in behavior of particu-late materials located in these various regions is the subject of thisinvestigation.

O396 -04(l0SPPage 2

3 2. METHOD OF DETERMINING RADIAL EFFECTS

Spherical devices were constructed (Figure 1) with concentric sphericallayers which contained inert potassium alum (Mallinckrodt No. 3216) or

"I Jtracer chrome alum (Mallinckrodt No. 4496). These materials wereselected because of similar physical properties and the ease of analysisof the chromium ion from the tracer. A relatively uniform size range of

*1 these materials was obtained by isolating a sieve fraction between 20 and32 mesh. The comparison of properties of these two materials is givenin Table 1.

All of the structural components of the bomblets were made of

0. 010-in.-thick extruded cellulose acetate sheet shaped on a vacuum-forming apparatus.

The devices were fired in the center of a 21. 5-ft-dia by 24-ft-highaerosol test chamber. The resultant aerosol clouds were sampled asfollows: Twenty-four trays, covering 11. 3% of the totasl chamber floorarea, were left in place for three hours after the firing, In most casesat least 95% of the disseminated material had fallen out after this timemaking the tray samples representative of the cloud material. Thematerial collected on these floor trays was weighed and sieved througha series of screens with openings ranging from 833 p. to 37 p.. The saltcollected on each sieve was weighed and analyzed for chromium (seeAppendix A). A particle-size-di stribution curve was plotted for bothtotal salt and chrome alum for each shot based on these weights.Millipore filter samples were taker, six feet from the chamber walland 12 ft above the floor at various time intervals during the course ofthe first half hour after detonation. The Millipore samples wereweighed and chemically analyzed for chromium. From this filter datamass-decay curves can be plotted from which particle-size distributionsmay be calculated.

Test results described herein are divided into two groups: (1) overallresults for the entire salt content of the bomb, and (2) results for thetracer material. Gross results for the entire salt charge were deter-mnined by weighing sieve cuts and Millipore samples. Tracer resultswere determined by analyzing all of the samples (i. e., sieve cuts andmillipore filters) chemically for the tracer and running the samecalculations on tracer weights that were run on the gross salt weights.Material balances were calculated by weighing the salt recovered onthe trays and multiplying this weight by 8. 85.

LI

VDETONATOR

I~ EI

I ZN2

, 1

I • CHAMBER

r WALL

1 4•H

ZNWZONE 3

S•-REFERENCE

U Figure 1. Spherical Explosive Dissemination. Device.

IIt...

STable 1. Properties of Chrome and Potassium Alums.

De signation Chrome Alum Potassium Alum

Formula K Cr(SO):12H 0 K Al1SO412H 04 2 AS 4) 2.

Specific Gravity 1. 83 1. 76

I Melting Point 890 C .92°C

Specific Heat 0.265 cal/gm (150C) 0. 324 cal/gm (154C)

I Formula Weight 499.4 474,.4

Refractive Index 1.4814 1,4564Crystal System Cubic Cubic

Color Purple White

Crystal Habit Octahedral Octahedral

Use in Munition Tracer Inert

IAI

j ~0395-04(106)P

S°a 5°;

Composition C-4 was the explosive used in these tests. This explosiveiis a mixture of 9 1%6 RDX (cyclotrimethylenetrinitramine), 2. 1 61 polyiso.butylene, 1, 6% motor oil, and 5. 3% di-(2-ethylhexyl) eebacate. It was

E-83 detonator.f

2.2 PRELIMINARY EXPERIMENTAL RESULTS WITHI INORGANIC SALTS

The first experiments were generally to determine whether any differencein behavior of the material was noted in the three general zones as illus-trated in Figure 2. The agent-to-burster weight ratio of these tests was

I on the order of 28:1. Results from these tests are shown in Table 2.

Tests were conducted in which the bomblets contained either all tracerI or all inert material. These tests were 10, 11, and 13 (Table 2). Theseshots were made to compare the similarity between tracer and inertagent, the assumption being that the tracer and agent would each duplicatethe breakup characteristics of the other. If this assumption were correctthe final breakup of the total salt content of the bomblets, as characterizedby mass median diameter (MMD), should be the same for all cases,regardless of the position of the tracer. Figure 3 indicates the slightdeviation from this ideal behavior for the average values of total saltcontent for Tests 2 through 7.

Plotting the data for tracer breakup vs radial distance from the explosivecharge surface gave the graph illustrated in Figure 4. The data recordedon Figure 4 indicate that the breakup of salt is greatest in the regionimmediately adjacent to the high explosive, diminishes with distance, andthen increases again at the outer surface of the device.

i To qualitatively eliminate the possibility of the increased breakup in theoutermost section of the bomblet being due to Impaction on the aerosolchamber walls, a standard concentric sphere device loaded with chromealum was fired. Petri dishes filled with Dow-Corning silicone grease,Type 4, smoothed on top, were taped to the walls of the chamber at thesame height as the bomb. Penetrations of the grease by particles of thechrome alum indicate the force with which the particles impact thechamber walls. Eight areas around the tank wall were covered with the

a grease dia.heR Typic•al dishes recovered after tests are shown inFigure 5. Of the particles striking the surface, -2%6 (by number)penetrated the grease; an insignificant fraction of these particles pene-

I trated to any appreciable depth. These results indicate that secondarybreakup by impaction on the walls of the chamber is probably of minor,if any, importance in these tests.

.• ,.11t

* 0395- Q�(]�)SP.?a.ge6

III -

II

'F Li

I A

tI 0

1 *� � 0

h

1I

I

I

.... -bar_

~3. . 395-04(16)SP

N4 u* C C O In. 00 CO 14 LM

r. 0% r4 uz 0 , t- I nL -LA ý C O D 00 00 I' I0O C

14 0n 00 10 44 Ld

4)0 fn '0 0 1- L

>O N

N N M

u$4 te;L C Ld

>3 - N4 1- N

0 3

$44

4.' 4)

0n N N N N

c- C;O

COn

03.95.04-( 6)SP

1 24

o u

d co

x 0

1.4

C-C,

(IW)(*IVLOI OW-IV I

O39B-Q4O.6)S��agj�91'* .7 18

1** I'

H4)21I. �. (�, I

I-

II I

-�I ".4

I0

Ij�4)'.4

1. II'.4I

'I.E0

4-I

In -I0I CIId

1 4)'.4

-'-4I0 4I

�1I - A

I I

� ia�fwNfl ONW �VNIi

I �..a

S 0395:04Ij1) P

Page 1

(Side View)

J1

( T o V i e w )

Fiue5 iioeGeaeipcinSmlr

3, DETAILED EXPERIMENTAL RESULTS

i•-i

1 ~3. 1. 1 Inorganic Salts

The dual breakup effect noted in Table 2 was confirmed by runningadditional experiments in which the zones previously loaded with only

I inert potassium alum were loaded with the tracer. In addition, twotests were conducted using 0. 060-in. -thick explosive chamber wallsinstead of the usual 0. 010-in. walls. These are Tests 8 and 9 inTable 3 and were tested to see whether additional thermal insulationof the explosive products would result: i.a change in final size dis-tribution due to a thermal effect of the hot detonation gases.. Resultsfrom all of these additional tests are shown in Table 3. The combineddata from all tests listed, in Tables 2 and 3 are illustrated in Figure 6.

3. 1. 2 Organic Compounds

I The apparently reproducible behavior of the breakup patterns of inor-ganic alums L-aises the question of whether the behavior of organicmaterials (which closely simulate the properties of CW agents) wouldbe similar. To study this behavior, a series of concentric sphereexperiments were conducted similar to those already describedexcept that organic materials rather than alums were used in theI devices. The use of ascorbic acid as the inert filler and calcium:ascotrbate as tracer was considered but rejected when it was foundexperimentally that these materials, after explosive dissemination,1 became hygroscopic. Anthraquinone was therefore chosen as theinert solid with p-aminobenzoic acid as the tracer. These materialshad somewhat similar properties, could be readily assayed, andwere in a size range typically resulting from large-scale organicproduction methods, such as those probably used with active agents.Properties of these materials are given in Table 4. Particle-sizedistributions of these materials as used were determined by a inivco-merograph and are given in Figures 7,and 8.

I

:1 0395-.04(1b)sp

4-b

.4- M. 0 v 0 -0)0

4414 00 0 6

54)J) >

@00

@0

144

E. 0)

%0 0 O1" -4 4 -

0395-04( 16)SP

I 0 0

-4

A IOMO81 OM VI

"rm

0 0.±S

ITable 4. Properties of Anthraquinone and p-Aminobenzoic Acid."

uDesignation Anthraqunone p-Aminobenzoic Acid

zk o H2N-*(

Specific Gravity 1.4 1.5

Melting Point Z86.C 1860C

jSpecific Heat 0. Z cal/g(15"C) 0. 287 cal/g(15"C)

Formula Weight 208.06 137.06

SColor Yellow Yellow

1 Use In Munition Inert Ttacer

I

ji

." 039-,Q4(16)AP

Ii 7age '.

II')00a ANTHRACUNONE

150I-50

140--

I

I013011 220 0 2 0 09 0 0

6 TKSDAMTRI MICRONG

I o" !Figure 7. Ptie-zeDistribution, of Starting Material (Anthraquinone).

I . . . . . ."P a g 1 6

Ii~W r00-

20

I

II

1-J|a'°

02 4 6 10 20 40 so so 100 200STONES DIMWETER IN MICRONS r

SFigure 8. Pairticle-Size Distribution of Starting Material (p-..Aminobenzoic Acid).

0390-04(16)SP

Page 17

The same experimental details of firing and sampling wesre used in thecase of organic materials andin the alum experiments. The sieve frac-tions and millipore filters were analyzed by ultraviolet spectrophoto-metry for p-aminobenzoic acid content (A rrnri 3c Al lniiltr fr'n, the

V'$ AA. &AlUAADCK ien in Liu~e o. ine tracer datafrom these tests are shown in Figure 9.

'1 In Figure 9, the resultant particle-size pattern again shows the apparentincrease in particle size with distance from the high explosive chargeup to a point. Then the particle size decreases again near the outersurface of the bomblet. Figure 10 illustrates the final size distributionof the total salt content of these bomblets.

1.3.2 EFFECT OF BOMBLET CASING VARIATION

if Since the increased breakup in the outer layers of the spherical bombletshad been duplicated in several experiments, it seemed possible that theshockwave interactions at the outer surface as well as aerodynamicprocesses were the major influence on this breakup. Replacing the thin(0. 010-in. ) cellulose acetate with a more massive confining materialchanges the manner of shock passage to some degree since more energyI will be required to breakup and move the heavier casing. Tenite-IIplastic, already available in 3. 0-in.-ID 9pheres with 0. 0625-in. -thickwalls, was the material chosen. The internal construction of thebomblets made from these thick-walled spheres was essentially thesame as for the first group bf experiments, except that the 3. 0-in.-dialimitation permitted only four tracer positions rather than the originalfive. The same firing and sampling procedures as previously describedwere employed in these tests. The inert material was again potassiumalum and the tracer, chrome alum. For direct comparison with thesethick casing tests, four bomblets having the same structure but with0. 010-in. -thick cellulose acetate outer walls were tiried. The datafrom these two sets of experiments are given in Table 6.

The final MMD of the total salt content vs average distance from highexplosive-to-tracer, for thick- and thin-walled spheres, ie. plotted inFigures 11 and 12 respectively. There is no significant differencebetween test results obtained from bombs with thick and thin casings.Again,some deviation from the ideal horizontal line is evident.Figure 13 is a plot of all of the data points from the original design andthe thick-and thin-walled 3. 0-in.-dia spheres. For clarity, theaverage values determined in these tests are plotted in Figure 14.

I'I4

0395- 04(16)SipPage 18

IE If; _o: _ _ _lo

mI 0 D 00 r- 0'

1 .4 r- 0ý v a 0 a

~0

Oo -4

1 44

;tJt

4--4

C'. -6 c 2 0 ý c

o Cd-".4

0d> 4

U _ _ l _ _ _ III _ _ _N_ _

"F.- Nc N0 N 0 '

I _t

- ~ ~~~~ - -7=7-- - - -" ý=77~

~to

II 0

be

-. .|Jn 0

P4

'1 oj •

i:i

00 ON N

.0;

[ 0395-04(16)sP-

ii°,~I!

I, !§I U

l o a

0395 -04(16)SP

Page 21

mEE4J 1ý

(d '0 Ln 0- %0 C- '0 A

0 tl 00 'Lt 0% 0 N0 en 0 Ln

cd co ~ ~ ~ dvU Ln 0% '0 L - '0 00 0' '0

7 Me QM L

0U (4b'0 c - C- N

(d oo L LA m~ N 0 0 rn L) N

(d k 4.) C; C; -; - 7

0) u

(U

I~~ N 4Jj- r ' ' '

4) )

IfIo) N N M

7 '7

-• 0-p -•04(i p'

O35CO( 16)SPPage . .

I I- . _ ____ ___

I P .

' E-4 .. l •,I2

0

U. to

1.A

w

lz

c 4

w

I

_ _ _ _z nt .1 ......

1 10

-E4;Ig

P-agJo

3wwI-i

10

SN

I0

1.0

W

,u(I

-0 . 0 ,__0.0-0

.0I •,0s

30395-'604(60Page 24

I - 0

III x

U. 4J

I: 0 0

0l 0 8 80Havu ON IV~

039I0(6S

Iit

V 5

11U, W FE

I- k

0 9

06 V3V -R 1VN

4r IImprovements in the analytical procedure for chromiumrn mad•. po a •1-I -analysis of the millipore filters in some of these tests.• Ta••b..7..

A:A jUA% A a -

* • indicate that the percentage of very small particles producedis lighest ýt* in the layers near the explosive charge and is lowest at the outer surface.

Evidence of this nature supports the theory that the very fine particlesi• ~~are a product of the evaporative effect of the hot detonation products,. i

13.3 LAYER STRIPPING EFFECTSa

Since the general trend of results in the alum experiments is evident when A

1 the tracer layer is placed at various positions in a constant size bomblet,it is expedient to examine the effect of making each of these tracer layersan outer layer to determine whether this outer layer effect is a functionof absolute size or of relative position. This was accomplished by run-nling a series of experiments in which the breakup of tracer in the outerlayers of successively smaller spheres was observed. The series ofspheres was produced by sequentially stripping the outer layers from the

. standard concentric sphere device previously described.

SData from these experiments are shown L Table 8. Figure 16 is a plot

-4. of the data for total salt content of the layer-stripping devices. A com-parison of the tracer data for the original concentric sphere series andj the layer stripping series is given in Figure 17.

1 3.4 SHOCK AMY• LIFI CATION EFFECTS

Tests 8 and 9 (Table .31 apparently were not affected by the reduction ofthe hot detonation gases obtained by confining the explosive charge in athick plastic chamber. Another series of experiments was conducted inwhich the explosive was confined in a metal chamber. A metal chamber

j should more effectively hold the hot detonation products and, in this

design, should also allow the expanding explosive chamber a relativelyfree run before impacting the first salt layer, thereby increasing the

i peak shock pressure. This "free run" was provided by surrounding theexplosive casing with an air layer surrounded by the alum layer.

1IU

0395-04(16)SPI Pa~ge 27

I 14- IL

0

00

N 414. LA c.,

00 0

A. 0, C

"-4-

V: 0~ LnC46 N o

C-d

r- N N N No

d)

* ~C;K.- *..4 L ) * 0

00 p.4 %

A N N CO

S!," O395-O4(16)5P

I 2. Page 28

2-

w1o

0 0.2 0. .6 0 1.0

AVG TRACER-CHARGE DISTANCE (IN.)

* Figure 15. Amount of Fine Tracer Particles in Thick-Walled Bomblets.

JI!iI!

- page 29

IN

-4

n co In 0n %

aL %0 In-q~f

to)1

4)

"A t- 63 4

40 %0 1 - o 0

0 395 -04(16) SP

21 Page 30

34

("ItmU '4W"I

0395-04( 16)SP

Page 31

0 0

0 -z

0

"M4

000

0~4

0

00

II

("XMIODVIIJ) OWN NI

-~- ----- - ----- - - --I -

.,-• 0395-04116)SP

Page 32

Eleven tests were run in this study, using three conditions. Results ofthese tests are provided in Table 9.

The explosive chamber employed in the lirst series tl.gure iol conubieuof a 0.250-in,-aluminum wall container with a 0. 125-in. air gap. Theremaining three spheres of the device were constructed of 0. 010-in.-thickcellulose acetate. Similar devices were used in the second and third

*1 series, except that the metal explosive chambers employed a 0. 063-in.-thick wall and four agent spheres (Figure 19), The explosive chamberemployed in the second test series was made of aluminum, in the thirdseries, copper. The explosive used in all three experiments was AerexL-1 (Aerojet Liquid Explosive), detonated with a du Pont No. 8 detonator.Standard sedimentation, sieving, and chemical assay techhiques wereused to sample and analyze the resulting aerosol.

Figure 20 illustrates results obtained from the tracer analyses and com-.1 pares these data with previously reported results of tests using standard

plastic spheres. It is apparent that there is a tendency toward increasedbreakup in the outer layers of the spheres. As shown in Table 9, materialbalances are generally good for both the alum and the tracer.

Plots of total salt recovered in the devices are shown in Figure 21. Ifall salt layers were identical, the illustrated curves would be horizontal;there is, however, a slight variation from this ideal slope. The pointson the curves fu- the thin-walled shots show little deviation (indicatinggood reproducibility). A slightly greater breakup was produced with thealuminum than with the copper.

Results of the thick-walled shots show a somewhat greater variation. Itwas noted that both types of aluminum chamber fractured into manyrelatively small pieces while the copper vessels generally produced afew large sheets (< 10). This result indicates that the copper stretchedmuch more than the aluminum and was thus more likely to impart a

7 uniform high-amplitude shock.IResults of tests using the two typ... of metal generally appear to be quitesimilar, with the exception of data from the layers nearest the plateimpact. If this difference is significant, it could be accounted for bydifferences in shock characteristics, or by increased effects of escapingihnt detonation uases in the aluminum liners.

Table 10 presents results from tracer assays of millipore filters and pro-vides weights of small particles.(>l2 t) produced by these devices.

t~j 0395 -04( 16)SP

Page 33

Q) t en N %0 a v4 co n 000c N

u to : Ln %.0 00 %.0 %D.0 00 %0 '.0

-W -:L %0 L A 0 00 '0% N 0% 0'. 00

"00 10

u m 10 1~-4 O n 0

m~ %0 N- fn0 LA 0n 0'. 00 0'00P- -4 - . 1 4 .4 -

0.C1

0IV

-d fs -- 0 N " - N LN '-~

00

co

U ,a~m 0 0- a, ~ 0. - N 0LA L 0".4~ 0n 0 -n tn4~ ~

IL

U 0395-04(16)SPPage 34

- 0* 14

4A 0 0 L A A 0 L

0- * d 4 d d

'I 0

-. 04

I ~ ~~ ~ ~ ~ -b n0LnL .4) Ln 4)00

:I (Dnu

rh

0395-04(16)SPPage 35

} /-DETONATOR

o/

/

i ~/,o

AIR GAP

I

�" E,-EXPLOSIVECHAMBER

1 "•-. "\ EXPLOSIVE

"•'•'--ZON E I

ZONE WALLS ZONE 20.010-IN. PLASTIC \-ZONE 3

i Figure 18. Shock-Amplification Device (1).

I

U

03954-4(16".I Page 36

I• DETONATOR

iIR GAP

I

EXPLOSIVECHAMBER

S-EXPLOSIVE

ZONE WALLS ZO-ZO I

0.010-IN.. PLASTIC ZONE 2

I ZONE Ai --.. NE 4

Figure 19. Shock-Amplification Device (2).

I

*1III

03i95-04(16)SPPing. 3?

HIt-.-%-.'

I a1 0/ N

�. / II

�. / I�TTY VSi / I i-I

I V

M

j / 2g

/ 0

/J ). �0.4

7. ox '-4

//// '4V

QJ .2 U-� w 00 g�

N 1N'

(A N

VIAI #1

I- (AIII2 00N

Io 0 '---0

Kg

Ii (M2�WNJ.) OlIN 'WNW

UU

I

0395-04(1.6)P` I Page 38

.4j

IIn

N

.I 0

I

SI6 I n,, o

• II0 "• ( "I V L ~ l l O ~ ie "IV I .,-

Page 39It

... ...

As shown in Figure 22, very few small particles are formed in thisprocess (<2%). Because the absolute amount of particles produced'is small, the variation between the experiments is not considered

the sam e am ount of sm all particles (Figure 15).,:1

T!I-

t

!-

II

I

I

I

II

I

ig 40,

-c o LAX ,-441

GOI

E-4

444 A

/u

III

0 200r0- - --

I /4"1>M•OII

i0

4Iux0. L

0w:

'4.

T PLw-

lei

W911 ld3V~iN

0395.-o4.(116)Sp - .

Page 41

4, ANALYSIS OF RESULTS

Results from the primary group of experiments in this series indicate_~ that, for inorganic materials such as alum, a reduction in particle

diameter is achieved when these materials are explosively disseminated.

The resulting MMD's of particles in these tests were reduced by two-thirds to three-fifths. Reducing the agent-to-burster ratio from 28:1 to0.91:1 results in a further breakup of the total salt to one-seventh of itsoriginal MMD as shown in Figure 23. This small increase in sizereduction with a large increase in explosive-to-agent weight ratiosindicates the low efficiency of breakup of relatively small particles inexplosive dissemination devices.

The tests conducted with organic materials (p-aminobenzoic acid andanthraquinone) resulted in no decrease in particle size and generallyshowed an increase. Itis probable that some fracture did occur inthese cases but the competing reagglomeration mechanism wassufficient to result in a net increase in average particle size.

The tracer breakup-vs-distance from explosive charge investigationproduced trends which were similar for both inorganic and organicmaterials. In both cases maximum resultant particle sizes wereobtained in the intermediate layers of the bomblets. In the case ofthe inorganic agents however, there was a net reduction in size ofparticles. In the case of the organic agents there was a net increasein size. This trend was obtained by using both the total salt and thetracer alone for organic materials. The organic material, whichhad an initial MMD of - 30 ý, is in the particle size region describedby Mac Lean (Reference 8) where inorganic materials (silica, glass)are very susceptible to agglomeration, and where some of theoriginal particles may be disseminated without fracture. If the sizeincrease is less at the explosive charge surface and at the outersurface of the device, as the experiments indicate (Figure 9), thefinal agglomeration is mainly produced in the central area of theagent section. The magnitude of the increased breakup (or decreasedagglomeration) noted in-the outer layers of agent and in thoseimmediately surrounding a charge is experimentally significant aRdemonstrated by duplication of runs (Figures 1, 2, and 3).

I

I~

0395-04(16)SP

150

1 aoage !

ii

800

00 510 I 20 25 30

AGENT TO BURSTER WEIGHT RATIO

Figure Z3. MMD Total Salt as a Function of Agent-to-Burster Ratio.

IIIII

0395-04(16)SPPage 43

"5. CONCLUSIONS AND RECOMMENDATIONS Hsystems, in the case of solid materials, should be designed to take

I advantage of their dissemination potential rather than their particlecomminution function. It should be noted that, due to the consequentagglomeration of very fine particles, it may be necessary to presizethe agent load several times smaller than the desired final effectivesize.

The mass fraction of particles in the small size ranges, i. e., those

physiologically effective (materials below 10 dia), was very low inall the tests, generally less than 10% and usually less than 5% of thetotal. These results along with the low fracture efficiency tends tosupport the conclusion that presizing of the solid agent to smaller thanthe desired size range and reducing agglomeration effects would be themost effective means of employing an explosive dissemination device.

Although the exact nature of the mechanisms responsible for the "radialbreakup profile" across a spherical bomblet has not been determined,the profile was demonstrated to exist and a technique for quantitativelydescribing it was developed. It is anticipated that this technique will beapplicable not only in determining breakup characteristics describedhere, but also, with methods such as infrared and ultraviolet absorptionspectrophotometry and gas-liquid chromatography, to decompositionstudies of agents and slmulants,

I

II|

L 0395 04 16)SPPage 44

REFERENCES

~ j 1. Excplosive Dissemination of Liquid Agents, Stanford Research Institute,-ýT n TA -I nTni TDA -18O88-

405-CML-261.

2. George Washington University, Research Laboratory, Final ProgressReport, DA-18-064-CML-163, January 1, 1959.

3. A Series of Static Tests of Single Bis and VX-Filled 8-in. Shell Equippedwith Agent-to-Burster Ratios of 2/I and 23/ 1, and Detonated on the Groundat Various Heights, CWL Technical Memorandum 33-16, 10 December 1958.

4. Buck, D., The Explosive Dissemination of Liquids of Low Volatility,CWLSpecial Publication 1-7, Status Report August 1959.

5. Research Study on the Dissemination of Solid and Liquid Agents, Aerojet-General Quarterly Progress Report 0395-04(01)QP, July - September 1962.

6. Research Study on the Dissemination of Solid and Liquid Agents, Aerojet-General Quarterly Progress Report 0395-04(03)QP, January - March 1963,p. 11.

7. Buck, D., Explosive Dissemination of Low-Volatility Liquids from Cylind-rical Devices, CRDLR 3038, January 1961.

8. Temperley, H. N. V., "Further Study of the Dispersal of Drops of LiquidBy an Explosive Charge". Final Report, Directorate of Chemical DefenseResearch and Development, November 10, 1955.

I..

I!0

!

0395-04(16)SP. .• • ~~~~~Page 45 .. .. ,.

- I, APPENDIX A i-U

ASSAY PROCEDURES

A l. DETERMINATION OF CHROMIUM IN AEROSOLIZED CHROME ALUM

The explosive dissemination of chrome alum described in this reportresulted in a series of samples ranging from Millipore filters holdingmicrogram quantities of chromium to sieve fractions with gramamounts.

The analytical procedure used consisted essentially of three steps:(1) Extraction of salt from the millipore filter or sieve sample,(2) oxidation of trivalent to hexavalent chromium, and (3) spectro-photometric determination of the hexavalent chromium in the formof the red-violet complex which it forms on reaction with 1, 5-diphenyl-carbohydrazide (s- diphenylcarbazide).

Procedure:

Weigh out sufficient sample to contain - 2 to 10 ý± gm of chromium permilliliter of sample solution. Disolve the sample in 5%o sulfuric acidand make up to 100. 0 ml with the acid. Heat a 5. 0 ml aliquot with10 ml of 2% ammonium persulfate on a boiling water bath for 20 min.Dilute the heated samples to - 40 ml and make neutral to methylorange with a 10%o solution of sodium hydroxide', using an outsideindicator. Add 3.3 ml of 1:5 sulfuric acid to the solution and transferto a 100 ml volumetric flask with distilled water. Add 1. 0 ml s-diphen-ylcarbazide solution (4. 0 gm phthalic anhydride and 0. 25 gm s-diphenyl-carbazide q. s. to 100 ml with 95%o ethanol) to the flask and dilute to themark (100 ml) with distilled water. Read the maximum color develop-ment of the solution in a Klett-Summerson photoelectric colorimeterwith a No. 54 filter (- 6 mim after the addition of the diphenylcarbazidesolution). Run reagent blanks with each set of determinations.

Run standard solutions in the same concentration range with each setof determinations and compare the unknown readings with the standard.A typical absorbne'ebmncpntration curve obtained with this method isshown in Figure Al.

!

i

. "0395.-04(16\SPPage 46

F - t

II

200 -_ _ _

160

so

0iM40-

"0 C 4 6 a 10

SCONCENTRATION (GAMMAS/M L)

SFigure Al. A Standard Cidorimetric Curve for Chromium.

!I!I

0395-04(16)SPPage 47

AZ. SPECTROPIOTOMETRIC DETERMINATION OF p-AMINOBENZOIC ACID

.. ... -* ...... ~-. •.ivv n? %-nr•nobenzoic acid In aqueous

solution shows a maximum at /tb mi. ins •ft -•auL•Lu.AA&A% .it in these tests has no significant contribution to this absorbance.

Procedure:

Weigh out sufficient sample to contain - 5 IL gm of p-aminobenzoic acid permilliliter of final solution. Disolve the sample in distilled water andread the absorbance of the solution directly at Z66 mp. using a distilled"water blank. Compare the readings with those of a standard in the sameconcentration range.

Typical absorbance and concentration curves obtained with this method areshown in Figures AZ and A3.

A3. INERT SPECIES

The general procedure for determining the amount of inert material in agiven sample is to chemically determine the amount of tracer as describedabove and subtract this weight from the total weight of the sample. Thedifference is taken as the weight of the inert species.

I

lIII

•4 f

0395-04(16)SP

S]! Page 48

0.50.

S oii

o.41

io 3z

040

0.3

0.20

0.5

0.2IftS 8o0 2 75 300 325WAVE LENGTH t.mý.i

Figurc AZ. Ultravinlpt Absorption Spectrum of p-Amino-benzoic Acid.

._ , 0395-04(16)SP4 Page 49

0.9A

3 0d

09 o.,_______/

"I• ___ __ _

0. ---------

01

0. _____ I__

6 to

CONCENTRATION (WAMMA•I/ML.)

I ~ Figure A3. Standard Absorbance Concentration Curve for p-.Aminobenzoic Acid.

.0/,

Page 50

ACKNOWLEDGEMENT

The assistance of Mr. K. L. Harris on thechemical analysis phase of this program isgratefully acknowledged.

1

I

IU