Iftuti "MuLIR 69-92

EXPLOSIVE BEHAVIOR OF DINITROTOLUENE

~10 July 1969

UNITED STATES NAVAL ORDNANCE LABORATORY, WHITE OAK, MARYLAND

0.

0 OCT 1 71S9 m0 LTETOz A0IM 'ZON -

This document has been approved forpublic release and sale, its distributionis un-ibated.

NOLTI 69-92

EXPLOSIVE BEHAVIOR OF DINITROTOLUENE

D. Price, J. 0. Erkman, A. R. Clairmont, Jr.and D. J. Edwards

ABSTRACT: DNT shows Group 1 explosive behavior although it can be

dead-pressed moderately easily. Its infinite diameter detonation

velocity is approximated byDi (mm/ltsec) = 1.84 + 2.913 pc

Its detoanbildty and shock sensitivity are, respectively, near thoseof high bulk density nitroguanidine and DATB. A review of factorsaffecting detonability and sensitivity of a chemically hom(geneous

explosive is also presented.

Riblished 10 July 1969

Approved by: Carl Boyars, ChiefAdvanced Chemistry Division

CHEMISTRY RESEARCH DEPARTMENTU. S. NAVAL ORDNANCE LABORATORY

White Oak, Silver Spring, Maryland

NOLTR 69-92 10 July 1969

EXPLOSIVE BEHAVIOR OF DIUITROTOLUENE

This work was carried out under the tasks MAT 03L 000/RO11 01 01FR 59 and ORDTASK 033 102 F009 06 01. Its results are most usefulin contributi-r to the knowledge of explosive detenability and shock

sensitivity.

JOHN C. DOHERTYCaptain, USNCommander

By direction

ii

TABLE OF COMENTS~Page

EXPERIMENTAL ......................... 1

REVIEW OF RELEVANT PUBLICATIONS ............ 2

PRESENT RESULTS

Detonation Velocity .................. . . . 5Detonability............... . . . . . . . . . . 10

Shock Sensitivity .................. 15

General Considerations of Shock Sensitivity andDetonability ............. ...................... .. 22

SUMIARY ............................... 28

ACKNOWLEMMENT .................... . . . . . 31

APPENDIX, SUPPLIE ARY DATA. .... . . . . 3

REFERENCES ............................. . . 35

ILLUSTRATIONS

Figure Title PageI Diameter Effect on Detonation Velocity of DNT, X628 . . 7

2 Pattern of D vs p0 Curves for DNT at Two DifferentCharge Diameters .................... 9

3 vs d- Curve for DNT/RDX, 80/20, at 99.18% TMD . . . 12

4 Detonability Curve of 3-10p DNT, X628 . . . . ... . 14

5 Comparison of Shock Sensitivity of DNT With That ofSome Other Explosives. .............. 20

6 Detonability and Shock Sensitivity of TNT Charges . . . 27

7 Initiation Pressure vs Critical Diameter for FourPure, Homogeneous Explosives ....... . . . . . 30

TABLES

Table Title Page

1 Blinov's Values for Lead Cylinder Test of DNT 3

~iii

NOLTR 69-92

TABLE OF CONTEhTS (COT'ID)

Table Title Page

2 Detonation Velocity Measurements on Fine Dh4T, X628 6

3 D M.easurements by Probes on Coarse DNT, X587 . .... 8

4 Detonation Velocity of DNT/RDX, 80/20 .. ........ . 11

5 Failures Observed on Fine DNT, X628 ........... . 13

6 Comparison of d. of DNT, X628 with That of SimilarH.E . ........ .. . . .......... .16

7 Shock Sensitivity Tests on DNT, X587 . . . . . . . . . 17

8 Shock Sensitivity Tests on DNT, 11137 . ........ . 19

9 Source of Comparative Shock Sensitivity Curves of. 5 ........ ........ . . . . . 21

10 Shock Sensitivity and Detonability of TNT ...... .26

11 Measurements of d t Initiation Pressure (Pi) onHomogeneous Fxposives .............. ..... 2 9

A-I Ro-Tap Sieve Analyses of DIT (100g samples) ..... 33

A-2 Details of Measurements Summarized in Table 3 . . . . 34

iv

NOLTR L9-92

E.- .OSIVE BEHAVIOR OF DINITROTOLUENE

INTRODUCTION

2In earlier worh . we quoted a Russian review which stated that

Blinov3 had shown th.,: dinitrotoluene (DNT), dinitrophenol, and

other dinltro-compour-... of aromatic hydrocarbons exhibit Group 2

explosive behavior, i. , their critical diameter for detonation

increases with increasing loading density (po,). The objective of

the present work was to investigate the validity of that reported

behavior. The results if our studies of the detonation velocity (D)

as a function of po and diameter (d), of detonability, and of shock

sensitivity are reported here. They show clearly that the explosive

behavior of DNT is generally that of Group 1, not Group 2, materials.

The erroneous statement in the literature resulted, we believe, from

an improper interpretation of experimental observations.

EXPERI4ENTAL

The explosive, 2,4-dinitrotoluene CH3"C6H 3 "(NO2)2 has a crystal

density of 1.52 g/cc and melts at 70°C . By arbitrary decomposition

mechanisms a and b*, on a pe. gram basis, its heat of detonation is

89-82% that of TNT, and the volume of its gas products is the same.

It would be expected to resemble TNT rather closely in its general

behavior.

Two lots of 2,4-dinitrotoluene, tech., were obtained from

du Pont; they satisfied the du Pont sales specifications of 10/26/65

for this material (97.5% DNT or better). The two lots were given

the designations X587 and N137. They had average particle sizes of

150 and 350g respectively, and corresponding pour densities of 0.70

and 0.57 g/cc. Sieve analyses are given in the appendix. A portion

of lot X587 was used to prepare, by re-crystallization, about 30

pounds of fine (3 to 1L) DNT:. designated as X628.Charge preparation and handling were identical to those of

previous work.5 Charges were of various diameters, 20.32 cm long,a. Formation of H2 0, C02 in sequence; b. formation of CO, H20, CO2

in sequence.

1

NOLrT 69-92and boosters were 50/50 pentolite ( .o = I.56 g/cc) of the same

diameter and 5.08 cm long. The experimental assembly sand instru-

mentation for detonation vlocity measurements (70mm smear camerawith writing speed up to t!sec or ionization probes at variousstations) were also the same5. Record reduction was carried out as

in the previous work and, in addition, small corrections6 have been

made to the optically measured D value.

REVIEW OF RELEVANT PUBLICATIONS

As we stated above, the work of Blinov and his colleagues seemsto be the source of the statement that dinitro-compounds of aromatichydrocarbons exhibit Group 2 behavior. Although we still have not

seen Blinov's original :7ublication3 , we have obtained the translationof a later work7 whlch claims to show the same results. Also avail-

able now is the translation of a second, earlier paper 8 which is alsoa study of dinitro aromatics.

In both papers " 8 Blinov used the compression (change in height)

of a lead cyl! nder as a measure of "explosiveness." Moreover, in

1959, he states 8 , "...the explosiveness of dinitro compounds usuallydrops off with increase in bulk density..." The lead cylinder test

with no instrumentation, in particular without detonation velocity

measurements, is evidently the basis for assuming a trend of in-

creasing critical diameter with increasing density. This is in-

direct evidence very like that of a plate dent in the gap test.

Such evidence has already been shown unreliable for high bulk density

nitroguanidine, NQ-h9 .

Blinov further itated in 1959 that DNT packed In paper cannot

be detonated without use of a booster (at least 3g of pressed tetryl).

This is for a low density charge, one close to the pour-density of

the material. The work was carried out on technical grade materials

which passed a No. 30 screen. Similar material was also used for

the later work7 in which the lead cylinder values of Table 1 were

given for various corfined charges of DNT at 0.6 and 0.8 g/cc bulkdensity. Blinov states that they confirm. "that the critical

diameter of propagation of detonation in it grows with an increase in

density." We believe that the data of Table 1 suggest exactly thereverse of his conclusion. (See discussion below)

2

*|

NOLTR 0-9-92

0 cc0 430 fi04) ~ L- -

NLe (;N co cor' r. ~0C C~jc Cci cu m P4O-4_q

#a

4) .ko 04) E

Cu~C~4-4-4 0'j

0.1-)m 0 1-0 \0 . .-. CM U-

!-Z z~- Cij 0b~' - -0C 0. cc

*4 r4 '- -4 , 4 IAr

~r f-44 r r4 a) v0 tD4'; CO

0

E- .4- 34CD300 000ot a' 0

9-4 m4 CXZt:: 0- z )(V- r-4 N -4c

0- 4 0 _

c- .1 000000)F-4 V4~~ Coo0000CCD 0 4 ;4

0 1

I CL

ca -P-000 000) 0 ca .

04

.,40 4

iff 4-1 coS 054 0 0=~~V w C$ .0 0a

4)3

NOLTR 69-92

Our technical grade DNT (X587) had a pour-density of about 0.7

g/cc and the slave analysis shown in the appendix. Ln large diameters(6.3 W 7.6cm) a standard pentolite donor iritated reaction for un.-confined charges 0.7oj:p 1.33 g/cc, but the front velocity decreased

as it propagated and was not a detonation. At po = 0.7 g/c the front

velocity see.%t*d nearly constant and at Po = 1.5 g/ec no reaction was

initiated. in' the gap test confinement, charges at p0 = I.0 and po

1.5 glcc did uctonate. Ln order to obtain detonation of smaller

diameter unconfined charges, It was necessFxy to use a fine (-nIQA) DIll.It is quite easy to initiate non-steady state reactions in gre-

lar charges of coarse materials and such reactions, as we have shown

for NQ-h, are powerful enough to punch or dent a steel witness plate

(or flatten a lead cylinder). It is such reactions t.%at are probably

responsible for most of the lead cylinder results of Table I. Tiese

show that the outpat Increases with po for strong confinement, but

shows the reverbe trend for weak confinement. Only the four re ults

marked with a (c) seem to be possible detonations, and, if so, these

show that the higher density favors detonability or that the d de-

creases as p0 increases.10

Finally, in yet another paper , Blinov comments again that he

had previously established an increase in critical diameter with an

increase in loading density for the dinitro-compounds. %t he adds,"At the same time, upon lowering of density, failures are observed

with these methods of initiation that ensure explosion of charges of

large density." However, his results were confused by usntg es-

sentially point initiation (both boostered and unboostered) on lar-ge

dzameter, shock insensitive materials. Most of the lead cylinder

results on the weakly confincd charges in Table i seem best expla.ned

as caused by a sub-detonation, reactive shock, possibly in many cases

at effective diareters below the crit-cal diameter for detonation.

The opposite trends (with density) found for such a reactive shock ascompared to true detonation are analogous to those demonstrated for

NQ-h.

it

I

Detonvaation VelocItyBecause It waz lu 11,11be to- initilate j ataic the ~cn

rt ccars -Ma (1581 In =011rges smU tr to b-r-I'e convenietly

to use for ru-st- of tbe opltlcal vo-s. Mhe velocl-ty rnnrnete alt

lsted in Table 2. Pig. I sb-ous the extrapolat-Ion- of the D) vsddata to% inintei'rter for the Vsoadwi n i1tas 1.507 glee a:4

1301 s-/cc (two paint extrzapolatIon only). - is givesz thbe tUlwovalues shavwn n Mab' 2'. IT a linear D., vs p0 is also aewyed, these

data giveD1rz~sc .84 2.91T p, (1)

as the mt in' te diauter relat1.onship for DW. We re*alize that Eqft. (1)could be far better established w!th a larger nunater of deternir-atlona,but,. our supplyT of fine D)U was United - Eq. (1) is ad.-quate as aguaide to the D) D(p, I) pattern of mnW "M seIas to be the only D,vs P,, carw aailble- for this nt~erial. Koreover, It is Suv*crt*-edby sow maae~ts car-ried out on confM-La '.aItral boosteredcoarse DT, as will be tcxo~. M~g. 2sw tte D =D( p0 d) pattern

tO' .It Is a norsnal Gro-up I pattern.

Canfiurlemrent should !-cease the detoabllty. (decrease thecritca.L 6iL~eer) off coarse L.W. To assess this effect, prcobe

measuvrementa were nae on can-fined coarse nr. r587; they are givnin ?Able 3. Mhis, lo atD in the gap test'F canftriwn efldblts D)values enry clo-se to those of tm fin er detenantimg rctrntind 15m7.6-2 edirerchargses (see Sable 2); It iscr'nideotx.

*As noted above, how pr3 it:6 was .ery easy to, inie-tialte a vgrcreactlon with the d---or shock . For d = 7.46,0 =a" po. = 1,0, %A0 1.139gl=., reactIon fra-tn 0 o l decreas"UvC cte o bt23to 4.7 5J~ZC ere vtm4 At 1 .5 /c a~6 i. a -- a.

falure waMs rn.ch rma rAp-

NoLTR 69-92

:IN.AbC('J

skiU- A nAAA A.;t*AA

00

00f

I tC'cu (\ r %D c\ lU %DW~ CP\

0

4)I ~ tCJ 'ODI' eC'~Cj 0- NDv oC'o4 - 01).,r 0a~x %D..- C'i --t0 0 c

E' 0

oo0- N q ko8 (In . t%0

C LN 4) O11' f'0\ I-t0 0 4

.- 9- .4. r4, 9, H 9 9 0>

Io % 00cO' . O De co 0t.i CC) tJ43 t\cut- 61\\ - Lfl r-i IC 4)_t

0 Ul tn nnn inU'\0 n -\0

I NOLTR 69-92

0" U

C.V

0. z

I LUIU

II.-

'UA

a. u(D3Sri/WW)G~~ kL:03 .IVO3

7 ~

NOLTR 69-92

TABLE 3

D Measurements by Probes on Coarse DWI, X587*

Shot No. Do Dg/cc mm/bsec

1 1.00 3.788 0.0132 1.00 3.772 0.037

3 1.50 Failed1.50 5.908 0.051

*Charges confined in 12 in. length of standard gap test tubing boredto take probes. Detailed messurements given in Appendix.

8

NOLTR 69-942

7.0

r d (CM)

A 7.626.0- 058

5.0 5.0

0

I--z/

c4.0 -1

0.8 1.0 1.2 1.4

LOADING DENSITY p0 (G/CC)

FIG. 2 PATTERN OF D VS. p. CURVES FOR DNT AT TWO DIFFERENT CHARGE DIAMETERS

9

NOLTR 69-92

This increase in the effective diameter by corfinement is important7

because of the conditional relationship between d0 at Pg, i.e., only

for d Z d can detonation be achieved and hence a sensitivity

(P ) to shock-to-detonation transition measured. Because this ccn-9

dition is satisfied for coarse DIT in the range 1.0 to 1.5 g/cc and

in the gap test corfinement, mearingful shock sensitivity measure-ments (P ) can be made. (The particle size effect cn shock sensi-

tivity is small compared to Its effect on critical diameter.)Finally, the failure of shot 3 (Table 3) at po = 1.50 g/cc suggests

the possibility that dead pressing occurs in this material.

Another check of the detonation pattern of Fig. 2 and of the

ideal curve, Eq. (1), was made on an 80/20 mixture of DNT/RDX. TheDNT was coarse (X587); the PEDX, medium fine (Type B, Class A). Theresults, detonation velocity as a function of charge diameter, aregiven in Table 4 and plotted in Fig. 3 which also shows the extrapola-tion to the infinite diameter value D. of the mixture. Since the

chargee were of low porosity (99.2% T), the additivity rule for

H.E. mixtures should be applicable. It gives for pure DNT, D. - 6.25mnLsec at Po = 1.5075 ,/cc. At the same density Eq. (1) predicts

6.23 mu/psec which is an excellent check.

* Detonabilit'

* Table 2 contains the velocities easured on charges of fine DWwhich detonated. Table 5 presents the failures that were observedand summarizes the failure limit data obtained by combining the re-sults of both tables. Fig. 4 is a plot of the resultant detonability

curve. It lies somewhat above and to the right of the limit curve for

NQ-h 9 and far below and to the left of the limit curve of the coarse

DNT, Z587. Over most of the %TD range, the fine DNT has a greater

d. than does NQ-h, but the two curves cross at 87% TMD. At and above

87% TMD, therefore, the NQ-h has a greater dc than DNfT. It should be

noted, however, that the JNQ-h and DNT X587 are about comparable in

particle size (ca. I00-15CU) whereas, the fine DNT consists of 3-IQLparticles. It can be concluded that the propagation of detonation is

much more difficult in DNT than in any of the other H.E. studied

except NQ-h.

10

El NOLTR 69-92

TABLE 4

Detonation Velocity of DNT/RDX, 80/20

* d D

559 1560, 2.54 5.934 594 5.924565 3.455 6.5T3 6:552 6552

561 52 6.6755665 6.654562 1.5 7.62 6.T51 6.651

Di - 6.75 m/pepec at p. - 1.556 g/cc(99. 18%mh)

*Pv 1.569 g/cc

Coarse DNT, 1587Medium EDX, X597 (Type B, Class A)

11

NCLTR 69-92

0

100

-O

- M

12

NOLTR 69-92

TAME 5

Pailures Observed on Fine DK-, X628

Fail-ae atd 8

Shot No. cm cm

518 5.08 1. '01h 15.2

511 2.54 a.80Ch 57506 Za5:,512 1. H 6.6

5215141 <10. 2573 1.97 1 .5 -e 7

57.04,<1.94

Siwary of Detonability Data (Tables 2 and 5)

d+ d

1.00 7.62 5.081.16 5.08* 2.541.30 3.81r 2.54 'Record vakeats this1.. 0 5.08 2.54 is very near d .1.506 2.54 --1.514 -- 2.54

a. Preliminary laboratory sauple of I0-15p DUT used.

b. Booster only 2.54 cm long, by error.c. Shavlmgs from previous charges used.d. No trace on record; unveacted DYT recovered.

13

NOLIR 6'0- 92

75 * DE-TONATION0 FMLU2E

-4 9NEAR d

- I

0

1.0 1.1 1.2 1.3 1.4 1.5

LOWING DEN~snY o,,IGiCC)

FIG. A 0EIO NASIU.Y"CIRVE OF 3-10p DNT, X2

F1Gi

___-

NOLTH 69-92

NQ-h shows a definite reversal of its detonability curve athigher densities, a dead press phenomenon. The limit curve of Fig. 4only suggests this possibility for DNT by its curvature near thecrystal density. The two failuri-s observed at pO Z 1.507 g/00(Table 5) also indicate the possibility as does the one failure ofcoarse DNT at 1.5 g/cc (Table 3). Subsequent work in the gap testconfiguration confirmed the fact that DINT does exhibit dead pressingalthough not as readily as NQ-h.

The trend of Fig. 4 Is definitely that of a Group 1 material aswould be expected from the D(pod) pattern of Fig. 2. Nevertheless,it seems likely for the reasons given above, that DNT in someparticle size distribution and state of compaction will exhibit a

reversal in its detonability similar to that shown by Nh-h. On thehigh density branch of the limit curve, it would be expected to showa Group 2 (e.g., ammonium perchlorate) behavior pattern. It shouldbe ncted that the present fine DNT has a d. comparable to that of fineAP's in the 70 to 75 percent TD range. Table 6 compares the presentresults for fine DNT with NK-h, AP's, DATB, and TATB.

Shock Sensitivity

Shock sensitivity was studied on coarse DNT because the fine DNTwas available only in limited supply. However, the work on nitro-guanidine 9 showed that the particle size effect on shock sensitivityis small (particularly as compared to the effect on critical diameter)provided that the test material is detonating and not undergoing avigorous but subdetonation reaction. Moreover, the shock sensitivitycurves (P vs pc ) of the fine and coarse NQ approach each other as

gPo increases.

Table 7 contains the few gap test results on coarse DNT, lotX587. The very high P required to initiate the 98.9 percent -TD

charge indicates that dead pressing should occur at some %T1W >98.9%(that limit was the highest compaction of the coarse material that we

not be initiated to detonation in the same configuration (See Table 7)

strengthens the expectation that dead pressing will occur at some

high compaction.

15

NOLR 69-92

%,o VDA A

AA A

091

It\ tc\ A A

o o

100 q

0u jgI ; at0#4-' #tJnL1

lof I A

UN A

CD, 0.

0

0. ~

0

1 -I 1-p %p %0

16

ax a

1 - 4" ? 0 - 0a)- -

C%

5 Eo

E-C

0

"7

LiNOLTR 69-92

After lot X587 of the coarse DNT was exhausted, lot N137, a

somewhat coarser DNT was obtained. Gap test results on this material

are given in Table 8. Surprisingly, it was impossible to compact

this material to a dead-press density (in the gap test) or even to as

insensitive a state as was achieved with lot X587. However, the shots

made at smaller diameter (data at bottom of Table 8) show that if the

acceptor and its confinement were scaled to 0.7 the standard testdimensions, dead pressing occurred between 93 and 97 percent TMD.

In the region 79-98% TMD, the regular gap test results gave a

normal sensitivity curve for this material: P increases with in-

creasing %TD. The curve is shown in Fig. 5 where it is compared tosensitivity curves for NQ-h, TATU.. DATB, and TNT. (References forthese curves are given in Table 9.) All of these materials except

TNT give negative results on the impact test8,9"lI, i.e., they are

difficult to inite even in powder form. In the LSGT, there is in-

creasing difficulty of igniticn under shock in the order TNT, DATB,PNT, TATB, and NQ-h. The last is by far the most difficult to ignite

and exhibits dead pressing in the regular gap test (with consequentreversal of its detonability curves) in the higher range of % TMD.

We have not been able to demonstrate that any of the other Group Iexplosives will show this phenomenon in the regular gap test although

DNT, X587, and TATB seem to be headed toward dead-pressing at 100% TD.The two dashed curves of Fig. 5, for TATB and DATE, have been

derived from small scale gap test (SSGT) data and their correlation

with the LSGT data described in previous work12 . Testing the shock

* sensitivity of DNT to detonation in the SSGT is impossible. NQ-h

will not detonate in the SSGT9 " 12 , and DNT has a critical diameter

which is, over most of the % TMD range, even larger than that of NQ-h.Hence DNT is subcritical in the SSGT.

The dotted portion of the TATB curve in Fig. 5 is for the regionof % TMD > 95% where a good correlation does noq exist between the

LSGT and SSGT results. A sharp increase in gradient of the LSGT

shock sensitivity curve has been found only for NQ-h 9 , API 4, and AP14Imixtures 1 , and in each case it signals the occurrence of dead-

pressing. On the other hand, it occurs commonly in the SSGT shock

18

NOI.1Y 69-92

0

U)

4-')

4) ) 0

10040 4

W) n 4, 3

434

:04

L4.) 4)0co-r- 43 00

(4 4. esI0 $4

k ta

0) 4) kw r

n2 0

N4 4) EA4>to ;4 H

4) 0> 0.0 aH 0 x C LAA t% A0 'j14 v-I 002K 4L 0~0

'/3 40I 4)) -0

0 1% go V-1 -H

QJ41 V"4 4 LAL A$

.i 0 .2 0=0*C

k19443t

NOLTR 69 -92

110r LSGTDN

-CONVERTED SSGT VALUES 0N---------------BEYOND REGION OF ESTABUSHED CORLTINX8

NQ-h90-

TATB(X335)

70 -~

V~.

~50-

30- AM(33

0- TNT (X!412

65 75 85 95PERCENT THEORETICAL MAXIMUMA DENSITY (% TMD)

FIG. 5 COMPARISON OF SHOCK SENSITIVITY OF DNT WITH THAT OF SOME OTH4ER EXPLOSIVES

20

"our|e of Comparative Shock Sensitivity Curves of Fig. 5

Material Reference Coimnent

NQ-h 9 Average particle size up to lOC01LSDAT, X31 12 I50-1Lj, production lot

TNT, X412 12* TATE, X406 12*TATE, X335 13

DATE, X315 13 if 50-1O4L, production lot

CONVERSION OF SSGT VALUES TO LSOT VALUES 12

5 PointPo~g/c) %TM D33GSSGT, P LSGTs

p0 (/cc h- (kbar)_g (kbar) 9

-- a

DATE, X3151.233 67.12 6.94 31.2 25.01.455 79.20 7.38 35.8 28.5M.601 87.15 7.88 42.2 33.21.676 91.24 8.10 45.3 34.81.761 95.86 9.00 6o.8 43.2

TATE, T335*

1.519 78.38 7.92 43.2 33.71.645 84.88 8.56 52.5 39-3

1.762 90.92 9.63 74.1 48.81.840 94.94 11.10 117.5 62.51.887 97.37 13.47 162.2 83.0

TData for TAT, X406 of Reference 12 give almost coincident curve.

Particle size of these lots is unlknlown.

21

NOLTR 69-92

12sensitivity curves 1 2 The similarity suggests that many materials are

approaching a dead press limit near 100% TMD in the SSGT. Difference

in the location of such a limit in the two cases probably arises from

the difference in the value of the ratio of the effective test

diameter to the critical diameter (larger for the LSGT) and the lesser

insulation of hot spot areas in the smaller test. Ignition at low

porosity will, be equally difficult in both cases, but propagation of

either burning or detonation should be more difficult in the smaller

test.

Finally it should be mentioned that the chemical energy released

by decomposition must have a role in affecting both ease of ignition

and propagation although It alone is insufficient to determine shock

sensitivity order:Lng. The shock sensitivities shown in Fig. 5 are in

the same approximate order as the computed energy release in detona-

tion, but fine AP (with a lower chemical energy than any of them)

ranges in shock 3ensitivity from that of DATB to less than that of

NQ-h, depending on its % TMD.

GENERAL CONSIDERATIONS OF SHOCK SENSITIVITY AND DETONABILITYEvery study, such as that of the present report, contributes

something to our general knowledge of the shock sensitivity and

detonability of explosives. It is therefore appropriate to give here

a brief revised summary of our present view of the whole field.

The variable, % TMD, is related to the percent porosity or

(100 - % TMD). Hence it is a relative measure of internal surfaces

available for ignition and reaction. It is only relative because

a different particle size distribution will give different size voids

and a different sensitivity curve. But, as we have pointed out

before, the difference in P is not very large and two curves, cor-

responding to two different particle size distributions, tend to be-

come coincident at high compaction (high % TMD). Hence the trend of

the usual shock sensitivity curve P vs % TMD, shows that ignition byshock becomes more difficult as the amount of internal surface de-

creases. For any series of cold pressed charges, % TMD is a valuable

guide to both shock sensitivity (P ) and detonability (dc). There9 cseems to be some relationship between P and dc within any given

22

A

rI,

NOLTR 69-92

cold-pressed series, but It must be a very limited one. Change of

explosive particle size effects a very large change in de and only amoderate one in P . Moreover, the trends of P and dc with % IYDare opposite for Group 1 explosives and the same for Group 2 explo-

sives. Consequently, there is no evidence of a general relationshipbetween P and dc for pressed charges, and, as we shall show below,very little evidence of one for more homogeneous explosives.

Another way (besides compaction) of decreasing internal surfaceor changing its reactivity or both is by changing charge preparation.Thus for castable explosives, charges which are pressed, hot pressed,cast, and single crystal, respectively, range from relatively largeto near zero internal surface. In such a series of charges with ahigh (96-100) % 11, density or % TMD is w.o longer useful in pre-dicting either P or dc . Instead, the dominant factor seems to bethe extent to which the charge approximates perfect physicalhomogeneity (a perfect single crystal). Both P and d show largeincreases as true homogeneity is approached,

The two cases: (a) clearly heterogeneous, porous, granularcharges (%TMD dominant factor and particle size distribultion asecondary factor) and (b) an approximately homogeneous state (degreeof homogeneity dominant factor) can best be illustrated with TNTdata. For case (a), the P vs % TVD trend is that shown in Fig. 5with maximum Pg of 23 kbar at 99% TMD. But seven typical cast chargesof TNT ranged from 26-46 kbar in Pg and ore cast TNT, prepared by

very slow cooling of the mold, was off-scale because it could not beinitiated to detonation in the gap test. These cast charges were

all high density (96-98% TMD) and indistinguishable by this variable.The large effect of the cooling schedule indicates that here homogeneity,especially as affected by grain size, is the dominant factor in de-termining ignitability under shock (P g). Hence cast TNT falls undercase (b) above.*

* It seems reasonable to consider a major difference between pressed

and cast TNT (at the same density) as many and uniformly distributed

small voids compared to few and randomly distributed larger voids.

A second difference, when the cast material closely approximates

homogeneity, is small crystals in vhe pressed and large crystals inthe cast TNT.

j 23

NOLTR 69-92

The qualitative interpretation of these data seems straight-forward. For case (a) ignition is a surface phenomenon controlled

by the number of hot 3pots formed under shock. These in turn dependon the porosity (measured by % TM), number and shape of voids, amount

of internal surface and its reactivity. For case (b), at its ex-treme of a perfect crystal, there is no internal surface, and igni-tion (thermal explosion) must occur within homogeneous material as aresult of bulk heating caused by shock compression. This seems the

ony shock ignition m echanism for completely homogeneous exploslies,

and. the data on approximately homogeneous (cast) TMT suggest that itis the dominant factor in the approximately homogeneous region ofcase (b).

For completeness it should be noted that although the approxi-

mate homogeneity of case (b) has been obtained by the method of

charge preparation (e.g., a castng or forming a single crystalinstead of compacting small crystals), it is also obtained by ex-tending the range of compaction. Thus, both castable and non-

castable explosives can be made non-detonable (in given dimensions)by dynamic precompression to 105% or 110% Ta 1 5 . A major effect of

such preliminary1 shocking of compacted charges would be to squeeze

out all. internal voids and eliblinate internal surfaces where hot

spots could form. In other words, the precompressed material would

have to be ignited by bulk compression heating or not at all.

To cover both cases (a) and (b), we need an ignition Index

which includesi, properly weighted. all of the conditions that will

affect ignition under shock. In the region of the compacts such an

index would take account chiefly of available hot spot sites*; inthe approximately homogeneous region, it would indicate degree ofhomogeneity and compression (precompression will govern the heatgenerated by subsequent compression). In both regions, it would also

For pure materials such sites would be expected to be voids (high

stagnatlon temperatures from impacted spalls or Jets) or particle/

particle boundaries. But one material embedded in another produces

another type of heterogeneity where shock interactions can create

hot spots.

IlT 69-..

include, of course, the activation and reaction erergles, factors

that determine the Ignitabillty of the particular explosive. If such

an index could be used instead of % -TMD, all of the curves of FIg. 5should show a very sharp upward bend at the point where the material

approaches non-detonability under LSGT conditions.The explosive for which the most coplete data are available on

detonability and shock sensitiv1ty Is ThT. These data are assembled

in Table 10 and have been used to construct the qualitative curves of

Fig. 6 where shock sensitivity and critical diameter are shown as they

would be expected to vary with the ignItion I dex rTe lower range

of corresponds to the cold pressed charges; these TI., charges areporous in the literal sense of being permeable by liquids. The upper

range of % covers non-porous charges and .-uns from cast to sir _ecrystal and highly precom-pressed explosive. Porous charges can be

converted to non-porous only dyramically. Otherwise. the method of

preparation determ nes the permeability.Fig. 6 has been drawn to emphasize that Pg and dc vary i, the

same marner for a chemically homogeneous, approximately physically

homogeneous, non-porous TIT charge. For porous charges, P and d.vary in the same way for Group 2 materials, but in opposite directionsfor Group 1 materials as Fig. 6 illustrates with T141. It seems

reasonable to assume that the same factor is dominant in both

measurements (P9 and de) in region (b) and that that factor Is Ig-nlt~abillty under shock. For TNT, the sile crystal limit of region

(b) is a voidless solid, and both !gnItion and propagation must behomogeneous processes. In region (a), on the other hand, we expect

ignition to be a surface phenomenon, i.e., heterogeneous, whereas

propagation can conceivably be by surface or bulk reaction or both.

By the domina tion of the heterogeneous mechafl.sms in ignition and of

the howogeneous in propagation, it Is possible Vo have the opposite

trends of P_ and d shown in region (a) for T6 cEvery explo-Ive is expected to show a shock sensitIvity curve

like that -f Fig. 6. However, the division between regions (a) and

(b), which should occur at the same value of our ignition index A,

will occur at different % TMD and method of preparation. For example,

it occurs in DNT aust above 99% TD (cold press) and somewhat below

25

NN

- 0 CI

DeU 0-4 r-4 4

;7-C 0aa~os cm osco-o.

o. - - 41 a --.. I-- r" one

AI U. . oeo=I e -C, 00 0 AL.

W3 04 = 0I-

- 0 0 --- 0* CC s:L~ .c I CC% SC

u-i ;Z EV *i0 9-e4ov 0, *q -. Ctt 0 10 ~ l

CL IV -4 >~3 fl3.t 4 940 4 0 O 4 =i rd.

Ckc 00 oEl- = v,- fc>xS r C MS t-4.r3

x 9 C9 -cS 4a.ac w41r

C o cc-4<G 0 2 C C -4 434 >0a>

al fi Lc-o %; WU 4 1 0. C31 X r 0) 3e31 .o --c cci actS 000oC E- -- .- i .-0'..11eO EC3

o V4 a\ C- 4-P C Zr C -* QC.aN 0'. r- 0.s 04 0L a

Nt- 0 - ~ - t. 0 04~cLIN UN C -0 -- TD0 .VVC P.-4 0 OECc P4 C >-to 4c -o- -i4 3- 0 40*'C ri C O t

.0 cs C C -oo 0,i -Z -V -- * *

ao a2 0 0 ID

ts O-

NOLTP 69-92

b

IGNITION INDEX &

100[

a b

olIGNITION INDEX A

___, ____S - HOMOGENEOUS

I f-TEROGENOUSI,I t

PRESSED PRECOMPRESSED

SINGLECRYSTAL

FIG. 6 DETONABIUTY AND SHOCK SENSITIVITY OF TNT CHGARES

27

:OLTR 69-92

cast D.,T. The general detonail!1ty curve is U shaped as we have re-

ce:,tly shown9 . Again, its mr.i,mum d should occur at the same valuecof the Ignition index i%. It fact, it occurs at high % TMD for

Group 1 materials, at low % 77J) for Group 2. Whether it will be

possible to devise an igniltion index capable of describlng the be-

havior of all explosives reuains, of course, an open question.

Although P and dc show the same trends wltn \ 1:: region (b) for

a pure compound, there seems no obvious relationship between P anddc from compound to compound. This is illustrated by the data of

Table 11 for three liquids and one single crystal, forts in which it

is reasonable to expect all four explosives tc be in the (b) region.

FIG. 7 shows that all four of these materials require high initiation

pressures PI as would be expected then bulk compression ie the only

mechanism for heating. Thus PI ranges from 82 to 112 kbar. fut dcvaries "rom 2 to 68 rim and shows no obvious correlation with P1 . in

the more cor!ple=x case of voldless but chemically (and hence

physically) heterogeneous charges, e.g., aluminized, the trends in

V and dc with Al content can be the same (plastisol propellant) oropposite (TU'P). A great deal more work is needed to clarify factors

affecting both P and c

SMSIRY

1. DNTI shows Group 1 explosive behavior by both detonation

velocity pattern and detonability limit curve.

2. At 98-9W T-J), it appears to be approachin a dead press

limit in the LSGT. Hence a reversal in its detonability curve is

probable at high compaction.

3. The InfInite diameter detonation velocity is

D. (1.6sec) = 134 - 2.913%p

This Is nIased on too few expeirLts to be highly reliable, however.

4. The detonabiLity curve of ( 5-!0) D lies closest to tat

of, U- amoN~. explosiies wah1h have been studied.

5. The shock senstI r4 _Vv curve of PRT lies betwen that for

DATB (productlo.- qualIt:y) a:nd TAT?.

C -'

i*

r~wQI~p 69-92

TANS U

Measurements of d-, InItIatIon Pressure (P)onBzgncsSpnle

Material For.-(ks) ~ m

i, 20*C ~13f-

nflsin4le crystal .1371 85d~I

a. A. W. Caupell V- C. Davis, and 3. R. ?-ravis, Thys. -Iuts498(19061)

I,. A. W. Caampbe1, X. :,-; IMAUn.. and T- :;- Holad 3. ATpI. Phys.el, 963(1956)

o. Table 10

d. V. S. flynkln a-.- P. F. -PokM1A. Doki. Akad. Nwuk USSR 1260. 1719±!9611 thru RZ3iC-122

e. A. F. Belyaev and .. h. unfl. Ru I ihssIan 1. 17s. Ce.a2821t(1960)

f. T. F.. Eollaz4 A. Vi. Carph-fl, pa-m X. E. Malin, J. Appi. Th75.286. 12t'l7(97

29

NOLTR 69-92

PETN TNT-fF 0100-

0 NG NM0

ol

0 20 40 60 80

d (MM)C

FIG. 7 INITIATION PRESSURE VS. CRITICAL DIAMETER FOR FOUR PUREHOMOGENEOUS EYDLOSIVES

30

NOLTR 69-92

ACKNOWLEDGMENT

We wish to thank Dr. K. G. Shipp for re-precipitating the

commercial DNT to prepare batch X6S8 of the fine material.

NMTE ADDED IN PfRESS: Corrections made to the measured detonationvelocity are based on two assumptions: spherical expansion of the

detonation front with distance of travel and an increase in theradius of curvature of the front passing from the booster to an ac-

ceptor of lower detonation velocity (e.g., DNT).6 In very .vcentwork we have found a number of situations for which these assumptions

are incorrect. Hence we do not feel that the corrections made to themeasured D values of this report can be Justified without an addi-tional study of the detonation wave p-ofiles in DNT. Both the"corrected" and uncorrected data lead to the same conclusions exceptfor a small change in Eq. (1). This becomes

Di (mm/psec) = 1.96 + 2.913Po

when derived from the measured D values.

31

NOLTR 69-92

APPENDIX

Supplementary Data

Table Al contains the Ro-Tap Sieve Analyses of DIT X587 and

N137. Table A2 contains all the pin measurements which were sum-marized in Table 3 of the text.

32

NOLTR 69-92

TABLE Al

Ro-Tap Sieve Analyses of DNT(1og samples)

Lot X5§7Screen Ho. 60 100 140 200 230 270 Pan

" openingt, 250 149 105 74 62 53 --

Retained, Run],g 24.4 31.5 16.4 13.6 6.1 5.2 2.1

Retained, Run 2,g 23.4 30.1 16.9 13.6 6.8 5.5 3.6

By microscopic examination, this material consisted ofchunky cylinders with L/d of 4 to 7.

Screen No. 10 14 18 30 60 Pan

opening,g 2000 1410 1000 590 250 --

Retained, g 0.6 0.5 2.1 6.8 71.7 18.o

This material too was in the form of chunky cylinders withand 4/d of 2 the most comon.

33

L m• u~u.mln m

NoLTR 69-92

oj'4.0acoco'

.14 0 -- 0 I\ H tC4. C' 0 ' ', 0

0 00

0\ c 0-

00

I0 f

E- H4

o c o- c - 0 i0 a )( &Z 04.'

0~~ Z0- r

*l k % .0 0- t- , o >t-~O... 00.0 qz t-C %490o, % -% O\ %D t- OD t-- z* iD i0s

4 w 0.0 0 . . . .40 rq0 0

tu 0

10 t

I1 1 0 ts0-o :I -4P %D0 '..'0'0 0 - c 0z

t-~~~. t-COCO CO00 9-41 4

04

H 4430 0 \D L4. t.- k t- t- \~0 i3)44

NoLTR 69-92

REFERENCES

1. D. Price, "Contrasting Patterns in the Behavior of High Ex-plo-sives," Eleventh Symposium (International) on Combustion, TheCombustion Institute, Pittsblurgh (1967) pp 693-701.

2. K. K. Andreyev and A. F. Belyayev,, Theory of Explosives, StateScientific and Technical Publishing House (OBORONGIZ), Moscow(1960) through a partial translation by W. A. Erwin, Jr. Sec-tion of interest is "The Critical Diameter of the Charge andIts Dependen~ce on Different Factors."

3. 1. F. Blinov, Scientific Reports of Higher School, Chemistry andChemical Technology, No. 4, 1958, p 640.

4. W. C. VMcCrone, Analyt. Chem. 26 1848 (1954).

5. D. Price, A. R. Clairmont, Jr., and I. Jaffe, Combustion andFlame 11,, 415-25 (1967).

6. j. 0. Erkman, "1Velocit1Ly of Detonation from Streak Camera Records,"NoLTR 68-1L17 (19 Sept 1968).

7. 1. F. Blinov and L. M. Svetlova, "Influence of Shell on theDetonation Ability of Certain Disnitro Compounds of the BenzineSeries," in K. K. Andreyev, A. F. Belyayev, A. 1. Gol'binder, andA. Go Gorst, Theory of Explosives, Sbornik Stateye (OBORONGIZ),Moscow (196.3) through AD 605-706.

8. 1. F. Blinov, Khim. Prom. i pp, 419-25 through AE-tr-5528.

9. D. Price and A. R. Clairmont, Jr, "The Response of Nitroguanidineto a Strong Shook." NoLTR 67-169 (2 Feb 1968).

10. 1. F. Blinov and L. M. Svetlova, "On the Influence of the Dimen-sions of the Focus of Initiation of the Detonation Ability ofLowly-Sensitive Explosives." In collection given in ref (7).

11e N. L. Coleburn and B. E., Drimmer, "Explosive Properties of Amino-Substituted, Symmetrical Trinitrobenzenes," NOLTR 63-81 (May 1963)Confidential.

12. D. Price and T. P. Liddiard, "The Small Scale Gap Test: Calibra-tion and Comparison with the Large Scale Gap Test," NOLaTR 66-87(7 Jul 1966).

13. N. 0. Holland, editor, "Explosives Effects and Properties"NOLTR 65-218 (21 Feb 1967) Confidential.

35

nOLTR 69-92

REFERENWES (Cont'd)

14. D. Price, J. 0. Erkman, A. R. Clairmont, Jr., and D. J. Edwards,'Explosive Behavior of a Simple Composite Model," NOLTR 69-16,in process.

15. B. E. Dri mer and T. P. Liddiard, "Propagation Failure in SolidExplosives Under Dynamic Precompression, NOLTR 64-60 (22 Jul1964).

36

UNMlASIFIEDSectuity CLassificatiofi

AXCW4E NT CONTROL DATA - R&D

I 04IGINATING ACTVt-Y X Q40 A IST 1CURt~ty C LI:3SI7;CA?1t%

U. S. Naval a;M e laboratory I U~ASFEW-ite Oak, Silver Spring,, Md. G'-s

31 PECUYWrTILE

Explosive Behavior of Dinitrotoluene

4 OSSCMtPylve NOTES trvp. o rav &.d ft e .IV*

S AuTNOWS) (Le.,~ r~m& r nme, tas. if;

Price, D.; Erkman, J. 0.; Clairmont, Jr., A. R.; and Edwards, D. J.

6- O"*T PT I OTAL JWO- OF PA1591 no. or 0'me"

10 July 1969 147 is180- :;QdT~ACT 0 R .r tO S.-OMtpAT0*- RZPvOM? MUMIDE25)

03C6L 000/R0ll 01 01 NOTR699

TASK 033 102 F009 06 01 SOL O~~PORT f*O (Any in*.t ~mamb beft*a o"9

10 A VA IL AOILIY/4.IWI1A YVON UOT1C3

ftis dootmt bas been approved for p2blic release and sale, itsdlstribution ix, unilmitd.

11 SUPPLEMCENTARY NOTES 1Ia. SPOitSMcr WL;YAY ACTIVITY

INaval Ordnance Systems ComidIWashi~ton, D. C.

DNT shows Group 1 explosive behavior although it can be dead-pressed moderately easily. Its infinite diame~ter detonationvelocity is approximated by

Its detonability and shock sensitivity ere, respectively, nearthose of high bulk density nitroguanidine and DAM. A reviom offactors aff'ecting detonability and sensitivity of a chemically7homogeneous explosive is also presented.

DD IJA*18 1473 ~S FSecutity Classiflcaeft

MOL NYIL T P19 m

Deintlnetdtnto vlct

Infinite fi deoainvlct

Explosive classifimtIonI

1. O0!=NAfllG ACI MTV: Enter the nine and address Imposed by sccmiitp c!asificatio.. usiig standard tatesmetaof the Contractor. scN:9 actr. grantee, Oapnttment of Ia- such asfe*"e activiy orother ortveLzad"o (co.Ziorsto .jusrr) issuing I(I) rOatif td requesters my obtain copies of this

r'epmL report boo DI)."2a. REPORT SEC(TrY CLASSIFICATIOIt Enter the *wer- (2) -Foreign afmvceneat and dlsesioso Of this£1.1 security claassficetstm of the report. Indicate whether**Restuicted Data" s included. UKaung is to ba in accord Itpi yIJta n aeonce with a pmriame secwury regulations. (3 " U. L. Gavermot agencies COY otain copies of

Zb. ROU: Ato..ticdowadina. Srthis report directly fhas DDC. Other qnslified DVCnvcu* 52Do. 10 mAted Foce ladvaL-W inMM Enter -sMr.a rqes ~athe V ou aet- Also, when applicable. *bow that optionalmane hare been used foe Group 3 and Group 4 as author- (4) "L% SL military agencies may Obtain cqpies of this

lie. I report directly hon, DCC. Other qulified -eers3. REPORT TITLE: Enter the complete report tWte in allhlltqeutogcapial letters Tites in All cases should he mclaesIfied...-If a mraningful tile, canot be aectad without clssific-tion. show tidle classificolion in all capital* in paenthesis I (S l dsrti oth eptIacnold.Q -imeditely folowing the tile Iad DDC users shell roqaeer theug4. DESCRIPTIVE NOTES If appropriate. enter the type ofreport. *&. interim. proeg. 9-imar7 a"al or fial. If the epot has be furnshd to the Office of TachulcalGive the inclusive dates wbe aspecij~ficVeportin peis1 Sersies Dartmew of C*Aar* foe sal- to the p%&I, L6di.covered. Cato tise fact a" eOmer tMe price if know.S. AUTHMKR3): Eater the 2s~) of authov(s) an shows on IL SUPPLEKEMTARY NOTU Use for additiusi eapls-aor la the report. Eater lat Same. f"t samse. atide initial toyntsIf miita*ry, show rook end branch of service. The sam ofthe prnia aehor i*a ahvlz utiiu reureet 12. SPWV G MILITARY ACTIVITY: Efter the ftm of

the departmental proJect office or laboratory sponsotifg tAa?&. R13VT D)ATE Fatter the date of the repot as day. -)herearc nsd dodopest Include ad~esamonth. est or mnhyean If ame thtan -ve date appears g -

on the repeat. use date of potilcat;04. 13 ABSTRACT- Enter asabstractgivingabriefsd facftal7*. OTALN1J~R OFPAG~The otalage oast awa of the docent indicative of the report eve though

should follow normal pagination prncedwes. La.. ene the bea l.apa lcbesi h oyo h beha-~ ll~er o ~age cotainng Iforwton.port. If additional space is requited. a c.ontioustion sheet shl

7b. NUMBER OF REFERECCES Ente the total Ofub It -1 highly desisbe tha the aieanci or clssified -eprtreereci z cited iii the repaw1. Ibe unlassified- Eachb paragraph, of the aheteact shall as. withta& CONTRACT OR GRANIT Nt1~tt It appropriate. enter an iadacation of the military security classificatiof. the fa-the aplicable mambvt of the contract or gr a nder which Jfotmation in the paragrapk. represented as (7TS). (51. C). or (7)-the repoit was written. I Mhere, is so limitation on the length of the abstract mow-St. W. & Id. PROJECT NUSM Enter !he arpite Iever, the suggested lengths is from 150 to 22S wolds.subpli ta nupmber isstemlcato. thas purec e. 14- NZY WORDS: ter words ame technically meaningful tease

*~poc~ a~r, yaeu er. tskc~ir.etc Ior shr prses tha characterize a report sand may he used as94L ORIGIATORS REPORT NUERS) Eater the offt- index enti"a for cataloging the report. Key words nuast hecial report vanmber by which the douent will he identified aesetetd so *hAt 00 Mecurity classificatioft is rieuired. Ider'ti-mad controlled by the originatifig activity. 7%e ntx=be- most j ts, such as equqpment model designation. trade ease. militarybe sisque to this -eport.L project code name. ereogmpsphi tocaiioe. soy he used as keyAd9b- OTHER REPORT MNJMES): It the seo ha be words but will be followed by a idication of technical con-assigned toy other report numbers (erther br the "nabisorh isi~n flnk.rls n teats MsotinlMOr by the SpOnw*). alSO enter this number S).610. AVAfL.AB!LITY/(LIMTATION NOTICE& Entr any lift.ittios on further dikeemirntion of Obe report. Wwhe than Jh-e

Security Clssificetioe

aM

lik