AD-A133 976 APR SPECTROSCOPY OF THERMALLY INITIATED FREE RADICALS 1/1IN ROT U) NAVAL RESEARCH LAR WASHINGTON DCM DPACEET AL 270OCT 83NRL-MR-5212

UNCLASSIFIEO FG 9/1 N

128 *1o 25

L'.

13211 12 l2Ll 6

'"'U- li, 112.0-

MICROCOPY RESOLUTION TEST CHARTNAIIONAL BLIRI AUL 01 SlANILARLLS ]%3 A

I.o*

SECURITY C6ASSIFICATION OF T.IS PAGE 'fWhe, Dae 1E1.060

READ INSTEUCIOffsREPORT DOCUMENTATION PAGE strog 9MPLEMG911 omsI. REPORT MUMm1 2. GOVT ACCESSION NO. 1. NEC2llIEN.TS CAMALOG iUMBE

4. TITEI (Wai sbIIg) S. TYPE OF REPORT a PlimOO cOVElmao

EPR SPECTROSCOPY OF THERMALLY INITIATED Progres report on a contnui-gNRL problem. ;FREE RADICALS. IN RDX .NLpolm6. PE[R0FORMING ORG. REPORT MIUM9E9

7. AuTwONf(&. S. CONTNACY O GRANT MUERmseld)

M.D. Pace, D.R. Fmarm, A.D. Britt and W.B. Moniz

9. PERFORMIGN G ORGANIZATION NAM9 ANO AORESS 10. PROGRAM L9tE9&IE P* Pt9OCT, TASK

NVRA Re ch Laborry , IT U SWavnlton, DC 20875 61153N-13; RRO13-05-42;

61-0086-0.3

I . CONLTRLIOG' OFFICE NA[MC AMC ADDRESS is. REPORT OATE

October 27, 1983I. N3.UMER OF PAGES

1214. MONITORING AGENCY NAME & AODRESSO(S II.W bom CaieWloAp Offtce) IS. SECURITY CI.ASS. (tel npeff)

UNCLASSIFIEDIS&. O"C AS 11CATION/ DOWNGRADING

16. DISTRIGUTION STATEMENT ( a io Reat)

Approed for public relems; distribution unlimited.

17. OISTRI|UTION STATEMENT (of loe abset snobed In' leak 20. I Ai flne lom fA pot)

IS. SUPPLEMENTARY NOTES

it KEY IFOROS (Continue on.. roe *dd* If meedw~ mod ldbutir by block ,uee)

EPRRDXThermal itiation

20. A@STRACT (Caes onl -evrne dlds' It aeeeeeiW ad Id"f.ly by Meek MAN"

-. ! The exploleo RDX exhibits EM siipsas when heated to temperatures ranging from 120'C to 2004C.NMat RDX snoles give intes EPR signals at 197'C. RDX dissolved into sulfolane gives intense EPRSignals at 17'0 and weaker !Ss detected W- low a 120fC. A 0.1M solution of RDX in sulfolanegives an EPR spectrum at 1T -C which is attributed to a nitroxide radical [H2-(NNO2)2(CH2NOJA 1M solution of RDX in sulfolane at 17"C and a neat sample of RDX at 197'C give an identical -

(Continua)

0D, i, 1473 EDTOOFINV55BSET

II

t AN 7r3

SECRIT COMICTO 010 ?"IS4 ,1) Dat

ENON

sccuo*lv CL&SSIVIChYIOk Of ?$dos V&05 (111hes. ... Lra.

2. ABSTRACT (Conflnue)

*petrm of a Scond nmda.4!'bi Spectrum is simulateod by the following combination ofhyperfine complinp: (1) Wz-1/2, A-9.98G; (2) WI-1, A-8.21G; (3) 4xi-1/2, A-6.66G;(4) Wi-1/2, A3.40; (6) Wz-112, A-1.26G.

56Uem'Y CLASUICAtION OF THIS PAGCe~h"e Doslas)

CONTENTS

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

EXPERIMENTAL RESULTS ................................ 1

DISCUSSION .............................................. 6

REFERENCES ............................................ 9

2

DTlCELECTEOCT2 11 1. .. '

B Availajbil!v Codes:, ,I d/ r j

Dist Spec ad

LA

EPR SPECTROSCOPY OF THERMALLY INTIATED FREE RADICALS IN RDX

Free-radicals in RDX [CHsNs0sj hexahydro-,3,5-trinitro-5-triazinepcyclotiauthylene-trinitramines cyclonites hezogen] have been observed byboth thermal and potocheical initiation. 1- Termal initiation of solid-state explosives is accamanied by an increase in temperature (described asthe "exothermw).' An exotherm is not unusual since many v- reactionsproceed through exothermic pathways. For example the oxidation of sugar byburning is exothermic. The unusual characteristic of explosives is thevery rapid rate of iate heat evolution during thermal initiation.

We are investigating the production of free radicals in s om nexplosives by using ER and hW spectroscopy. The purpose of this articleis to demostrate the cw ER detection of free radicals in the explosiveRDX at temeratures below the normal" decomposition temperature of neatRDX (205-206'C). This was accmplished by dissolving RK into sulfolane[C H,0,Ss Totrahydrothiophene 1,l-dioxide] and heating the solution. EPRof neat RDX during heating was also recorded. Same of our previous work onREX includes photocheically produced free radicals. This article reportsfree radicals produced strictly by thermal energy. Future studies areplanned to look for transient free radicals (free radicals with millisecondlifetimes ).

EXPERII1L RLTS

The EPR spectra were recorded with an IBM Ut 200 x-band spectrometer.The modulation frequency was 100 KHz, the modulation amplitude was 0.63 G,and the microwave power was 8 milliwatts. Samples were heated in thecavity by flowing comressed air through a resistive heating coil. AVarian 4540 controller was used to regulate current to the heating coil.The temperature controller was calibrated with a chrml-alunelthermocouple. A quartz insert dewar was used to minimize heating of thecavity. Samples of neat MK were heated in a standard 3mn, i.d. quartztube. Samples of RDX plus sulfolane were heated in a lam i.d. sampletube. (Sulfolane contributes to loss in cavity Q which necessitates mallsample tubes.) The sample tubes were open to the air at 1 atm. Nobackground EPR signals were observed from neat samples of sulfolane or MO(used as a solvent in some experiments) when heated to 2009C. The

measurement of g-values was acoeplished with an internal Nt9+ standard.

Same facts about sulfolane: At roan temperature sulfolane is a solid,but above its melting point (.p. 27.4-27.8C) it becomes liquid and ismiscible with water, acetone and toluene. RDX dissolves in sulfolane andform a 1:1 omplex at room tumperature. 7 Above 704C the complexdissociates into one RDC molecule and one sulfolane molecule, so attmperatures above 704C omplexing is not a problem. Five i of sulfolaneat 700C dissolves greater than 100 mg of RDX without saturating. In theDR experients UK was dissolved in sulfolane to form solutions ranging inManusaipt apptovd Septembe 1, 1983.

..

oncmntration from O.M to 1M. Because the boiling point of sulfolane(b.p. 285'C) is much greater that the decomposition temperature of T= itis possible to decmose the RDX while it is still in solution. When thesolutions were heated, spectra were observed at a temperature as low as1208C. The optimn temperature where the EPR signals were strongest was1709C. At 1906C outgassing of the sample tended to detune the microwavebridge.

In the first experiment an 0.1M solution of REDX in sulfolane washeated to 1700C and the spectrum shown in Fig. la was recorded. This EPRspectru has 29 lines which are visible above the background noise. Usinghigh gain weaker signals were observed at the field extream of thespectrum. This spectrum is assigned to three inequivalent hyperfinecouplings. The largest coupling is from a single I 'N (I-l) coupling of15.8 G. The 'IN contributes 3 lines with an intensity ratio of 1:1:1. Thesecond coupling is from four equivalent protons (I-1/,) of 7.7 G. Theintensity ratio of this quintet is 1:4:6:4:1. The third coupling is fromtwo equivalent 'IN nuclei of 1.8 G. The intensity ratio of this quintet is1:2:3:2:1. These couplings were used to compute a simulation of thespectrum which is shown in Fig. lb. The ER spectrum in Fig. la-b isattributed to a nitroxide free radical having the molecular structure I.This free radical is formed after net loss of ND from the RDX parentmolecule. The same free radical has been observed by photo-chemicalgeneration in an EX- acetoitrile solution.$ We have also observed similarradicals produced by pbotolysis of RFMXK- solutions.I After prolongedheating (about 1 hour) the spectrua in fig. la changed to a differentspectrum (not shown) which is attributed to a different free radical.

In a second EER experiment a 1 solution of 1MZ in sulfolane washeated to 1706C and the spectrun shown in Fig. 2a was recorded. This EPRspectrum has 36 lines. The 4 highest field lines and the 4 lowest fieldlines are obscured by the noise. The spectrum is symetric, but there isno EPR line at center field. There is an apparent background signal whichhas a broad linewidth (approximately 25 Gauss). No attempt was made toanalyze this background signal. Ccputer simulation of the spectrum inFig. 2a was difficult but a reasonably accurate 'fitm to the spectrum isshown in Fig. 2b. The simulation required 5 inequivalent couplings. Thelargest coupling is assigned to a nuclear spin '/1 coupling of 9.98 G andcontributes a doublet with each line of equal intensity. The secondcoupling is assigned to a spin 1 coupling of 8.21 G which contributes atriplet of equal intensity. The third coupling of 6.66 G is assigned to 4equivalent spin /, nuclei which contribute a quintet of lines with anintensity ratio of 1:4:6:4:1. The fourth coupling is assigned to a spinI/ coupling of 3.4 G which contributes a doublet. The fifth coupling isassigned to another spin V. coupling of 1.25 G which contributes a doubletof lines. (These assignments are listed more conveniently in Fig. 2b.) Asin Figure 1 there is a broadline signal with approximately 25 G linewidthwhich is included in the sizulation shown in Fig. 2b. In many cases it ispossible to rely upon comquter simulations to unambiguously assign aradical structure. Such is not the case here because the parameters whichare used to calculate the simulation are not unique. There are fourdifferent paraeters which can be simlztaneously varied: 1) the number ofinequivalent couplinge 2) the nuclear spin number (I-1 or '/,)o 3) thenumer of equivalent spinl and 4) the value of the hyperfine coupling. Asthe naber of inequivalent couplings increases (in this case 5 inequivalent

2

- -- V*..''0'

0.I

H\ N./H

0 2N"N /C\' NO02EXPERIMENTAL HH

10GSIMULATION

HYPERFINE INTENSITY. EQUIVALENT SPINS COUPLINGS RATIOS

I 15.8 GAUSS 1:1:1

1/2 4 7.7 1:4:6:4:1

1 2 1.8 1:2:3:2:1

Fig. 1. (a) This first derivative EPR spectrum was recorded with an O.1Msample of RDX in sulfolane at 1700C. (b) The spectrm wascomputer simulated. The simulation required three "IN couplingsand a coupling from four equivalent protons.

3

• '-,,4 ±,.i - ,

V EXPERIMENTAL

10OG

I I SIMULATION

Fig. 2. (a) This first derivative EPR spectrum was recorded fromi a 1M4sample of RDX in sulfolane, at 1700C.

(b) The spectrumn was comiputer simiulated by using the followingcouplings:

1) lxI=L/, A-9.98 G2) W-1m A-8.21 G3) 4xl-'/, A-6.66 G4) lxI-1/, A=3.4 G5) Ixl-'/1 A-1.25 G

4

couplings are used) the probability of a unique "f it" decreases because thetotal number of variables increases. For example substituting 2 equivalentspin 1 couplings for the 4 equivalent spin V/, couplings in this assigqentresults in an almost identical spectrum (not showns the total numer oflines reains the sam, but subtle differences in relative peak intensitiesoccur). Further attempts to substitute different equivalent combinationsof hyperfine couplings into Fig. 2b were not conclusive. From theseresults: 1) It is possible to simulate the experimentally observedspectrum (Fig. 2a) as a single radical (neglecting the broadline backgroundsignal). 2) The hyperfine couplings shown in Fig. 2b may be used indeducing a structure for the radical but a different set of equivalentcouplings would lead to a different interpretation of the radical'sstructure.

In a third experiment 100 mg of neat RDO was heated to 1970C. The EPRspectrum which was recorded (Fig. 3) is identical to the spectrum in Fig.2a-b. (e magnetic field sweep width is different in Fig. 2a and Fig. 3.)This indicates that the same free radical is detected in a IM solution ofRDX dissolved in sulfolane at 1700C and in neat RF= at 1979C. At 1970C theRIX is apparently molten and forms a solution phase environment containingthe free radicals. Prolonged heating of the neat RDX at 1970C results indecomposition and gas evolution. Much of the sample eventually passes as agas from the tube.

lOG

Fig. 3. This first derivative ERR spectrum was recorded frum 100 mg ofneat RDX at 197'C. The spectrum is the same as in Fig. 2a. (Themagnetic field calibration between Fig. 2a and Fig. 3 is differentand the signals in Fig. 3 are weaker that those in Fig. 2a. Thebackground signal in Fig. 3 is not as intense as in Fig. 2a.)

5

DISCYSSIOH

ydifferent pathways for the thermal decomposition of rWK have beensuggested. Two categories for describing initial deomposition stepsinclude: 1) decomposition with the molecular ring intact and 2)descmpition involving cleavage of the molecular ring. The results inthis article support an initial decozosition step which leaves the ringintact, but subsequent ring cleavage is suggested to rapidly follow. Theevidence supporting an intact molecular ring is the occurrence of thenitroxide spectrum in Fig. 1. The cyclic nitroxide (I) can be derived fromthe parent RDX molecule by a net loss of ND. There are at least twomechanism for this: 1) hanolytic cleavage of an N-ND, bond followed byreaction with molecular oxygen as shown in (I1 or 2) recombination andrearrangemnt resulting in the loss of NO as in [2].

The conditions for detecting radical I are dependent on the RDXconcentration in sulfolane and the time of heating. We found that a -low-concentration (0.1m1) was required to observe I and that after heating thesample for aproximately 30 minutes the spectrum slowly changed into thespectrum shown in Fig. 2. The 1M sample never gave a spectrum as in Fig.1, but always appeared as in Fig. 2. This suggests that radical I is anintermediate radical which is converting into a second free radical (asindicated in [3]). We observed no gas evolution from the samples of RDXplus sulfolane at 1706C. When the temperature was increased gases wereevolved but no nitric oxide was detected by using the wiron ringo test.'This suggests that the nitrogen dioxide and nitric oxide may be reactingwith the solvent molecules. If so, then a solvent which is unreactiveshould depress the formation of radical I. A test with RDX (0.1M) inpropylene carbonate at 1700C gave no EPR signals. This is weak butfavorable evidence for equation (2]. Another possibility is that thenitroxide radicals (I) are reacting with each other to give a secondradical. Disproportionation of nitroxides with alpha protons according to(4] is not unumon (although neither product is predicted to be a freeradical).

The radical structural assignment for the EPR spectrum in Fig. 2 and 3is uncertain because of the complexity of the spectrum. The couplingassignment in Fig. 2b does offer a basis for suggesting a structure. The9.98 G coupling may be assigned to 1 proton. The 8.21 G coupling isassigned to 1 nitrogen-14 coupling. The 6.66 G coupling is assigned to 4equivalent protons. The 3.4 G coupling is assigned to 1 proton and the1.25 G coupling is assigned to a different proton. A plausible structure(this is a tentative assignment) is indicated as II in Table 1. It is alsopossible that a bi-radical is responsible for the spectrum in Fig. 2 and3. Such a structure would resemble III in Table I. Attempts to deduce anunambiguous structure have been unsuccessful but further experiments usingisotopic substitution are planned.

6

N 0 2 N 0 2 N0 2N-22 NO-N--N 02 K NJ N NN9+0

ill (,. N-N -, + -No~t -. o! I

N0 2 NO 2 N0 2

NO 2 N 0 N 022 I

[2] _No2 *_ K2 + .o, - N---

N INO02 NO02

NO02

NO 2

N--a + Not

NO

12

NO 2 NO2N I

N 0 2 N 0 2[3] N- N 0 2 -\ M-N 5 I Radical [

NO 2 NO 2

NOH NON NO

41 2HX N-0 N- + +H'( N-OHH NN H N -

NOH NO? H NO? H

71

Table 1

Free Radical and Molecular Structure Assignments Based Upon EPR Results

Designation Structure

0*

I. H1CL%,1

II. (tentative) H

Where R and R'= aliphatic groups

III. (tentat ive) 6-NC N-6

8

FJUPERES

1. H. H. Miles, K. L. DeVries, A. D. Britt, and W. B. Moniz, Propellants,Explos. Pyrotech. ZL 100 (1982).

2. C. Dorey, J. S. Wilkes, L. P. Davis, H. L. Pugh, W. R. Carper, A. G.Turner, and K. E. Siegentbaler Frank J. Seilar Research Lab, ReportFJSRL-T-80-0026, 1980.

3. M. D. Pace and B. S. Holmes, J. Magn. Rescn. U& 143 (1983).

4. M. D. Pace and W. B. Mniz, J. Magn. Reson. A7. 510 (1982).

5. C. Darnez and J. Paviot, Int. J. Radiat. Phys. Chem. .4. 11 (1972).

6. A. Stolovy, E. C. Jones, J. B. Aviles, A. I. Nanenson, and W. A. Fraser,J. Chem. Phys. 2L. 229 (1983).

7. C. Michaud, H. Merx, C. R. Acad. Sci., Paris, Ser. C 2&7.L 652 (1968).

8. J. W. Bodnar and C. F. Rowell, J. Chem. Phys. M6. 707 (1972).

9. H. Remy, *Treatise on Inorganic Chemistry, Volume In, ElsevierPublishing CaqwV, 1956, P. 592.

9

I