CONDENSED PHASED:ECQUPQSITION OF NH4C104
3545
Species Resulting from Condensed Phase Decomposition of Ammon~umperchlorate^
. Ellsworth Hackman, 111, Thwkol Chemical Corporation, Elkton Division, Elkton, Maryland
21921
Henry H. Hesser, and Harold C. Beachell* Department of Chemistry, University of Delaware, Newark, Delaware
1971 1
(Received March 23,1973")
Publicalwn cosEs assisted by the A i r Force Ofice of Scientific Research
The thermal decomposition of current or potential propellant ingredients has been studied from a number of aspects A variety of mechanisms have been proposed. Further theory development and acceptance, however, is being restrained by the fact that some of the key species used in the theories have not been detected by chemical analysis. Ammonium perchlorate (AP) has been a favorite candidate for study because it Seems to be a ubiquitous oxidizer in current solid propellants. Detection of initial and transitory decomposition species is nearly impossible by wet methods, chromatography, and slower analytical techiiiques. Time-of-flight and, for the first time, higher resolution methods of mass spectrometry of ultrahigh purity A P and of its deuterated derivatwe were used in this study. Of particular importance to several proposed decomposition mechanisms is the sure detection of nitroxyl (HNO), the hydrides of nitrogen, and the oxides of chlorine. The detection met,hod must not cause confusion by adding species of its own, such as by cracking due to electron bombardment in the mass spectrometer. Furthermore, it is important to find what species might be present in the condensed phase-- as opposed to the gas phase-that would indicate strong exothermic reactions. Since it 1s ultimately desirable to relate the species found by mass spectrometry to what might be presented during combustion under higher pressures, the pertinence of the one to the other is discussed. The species of most concern, IIKO (mass 31), was found as a minor constituent by both high and low resolution measurement at temperaturer above 80". It was confirmed by the strong reduction of mass 31 when deuterated AP was decomposed. DNO occurs at mass 3%. It 1s considered highly improbable that HNO could have occurred by reactions other than surface thermal reactions. High resolution spectra were used to separate the atomic oxygen (or doubly charged Q 2 ) and NH2 peaks at mass 16, thereby confirming the presence of a key hydride of nitrogen. The other hydrides of nitrogen were also found: NH4, NHI, and NH. No condensed phase decomposition species were found at 80" or below. Perchloric acid, atomic chlorine, and all the simple chlorine oxides were found, but Clz, HC10, HC1Q2, N[C103, and Clod were not detected. Also not detected in the condensed phase were nitrogen, NzO, and NOz.The mass range monitored was from m/e 12 to 200, No parent NH&lO*+ ion was detected in any experiments. Evidence of enough strong exothermic reactions in the condensed phase was found. It supports theories stating that this is a major source of energy to help balance the endothermic requirements for vayorization to sustain combustion.
1. Imtroduetb For several decades there has been interest in clearly defining tbc irnpor tnnt energy absorption, generation,
merous. In the late 1950's the final reaction products for simple systems were readily calculated by computcr. By the mid-1960's reliable final product distributions and transfer steps during solid propellant c o m b u ~ t i o n . ~ were ~ ~ being calculated for complex propcllanx mixtures and were being confirmed as well as possiblc by chcmBetter understanding should lead t o better control. ical and spectral analyses. Early simplified theories stated that ignition took place The great experimental difficulty lies in capturing on a surfacc when 5, critical temperature was reached. and identifying the proper quantity of a species before The flame zone 1,hat was then established above the a loss of temperature or pressure, or brforc rcactions surface provided t h r sustaining energy by conduction with other species or with surfaces cause a change in and radiation for steady-state surface regression. For mole fraction. Gas temperatures range from 2000 t o solid propellants there are many strongly exothermic 3500°K; the zones being observed rangc from 100 to reactions that can be proposed t o take place in the lo3 pm. The surface being obscrvcd regrcsses 1inc:triy flame zone to support such theories. Oxidation of hydrocarbons to CO, and HzO and of aluminum to Al20$ are good examples. (1) This research was sponsored by a grant from the Air Force Office of Scientific Research to the University of Delaware. Thc experimental difficulties associated with de(2) (a) A. D. Crow and W. E. Grimshaw, Phil. Trans. Roy. SOC. tccting and measuring initial and midcourse reaction London, S e r . A , 230, 389 (1931); (b) 0. K. Rice a r i d It. Ginnel, J . proditcts, gas, and solid zone dimensions and temperaPhys. Chem., 54, 885 (1960). ture gradieuta for a burning solid propellant are nu(3) R . G. Parr and B. L. Crawford, Jr., ibid., 54, 929 (1950). The Journal of Physical Chemistry, Vol. 70, dl'o, 24, 2912
3N6
E. E. HACKMAN, 111, H. H. HESSEn, A N D H.
at rates in centimeters per second and the ambient pressure is usually between 20 and 200 atm. The high t,emperatures severely restrict the range of materials and processes usable for sampling and analysis. The minute dimensions being observed and the rapid movement of the reference vast,ly complicate the study of a given plane in the react,ing system, while the customary high pressures reduce thc dimensions and materials usable for viewing and spectrometry. Condensed Phase Reactions. To circumvent some of these difficulties, spectral analyses of individual decomposition zones have been made at much lower temperatures and pressures; most of the effort ha.. been concentrated on one material: ammonium perchlorate. This report deals with primary reactions occurring largely at the decomposing surface. Microcincmatography of burning ammonium perchlorate surfaces and scanning electron microscopy of quenched surfaces have shown the qualitative features of the reacting surfaceingood detail.' Those photographs have shown almost conclusivcly that import,ant reactions other than sublimation are taking place :it, and somewhat below, the surface. Figure 1 is an artist's conception, based on a number of photographs, showing what the structure of the condensed phase looks like under certain conditions of high prcssure combustion. We arc most, intcrested in 24, the porous or noncrystalline zone, that even shows cvidcnce of a liquidlikc phase. This cvidcnce of some condensed phase reactions mpports some radiat,ion heat transfer calculations. Thcy showed that the flnmc zone could not havc the requisite temperature and emissivity to transfer enough heat to the surface and thus satisfy the endothermic requirements of solids gasification or sublimation. This meant that conduction and heat generation in other zones must be important processes. Since there is strong convection or mass transport away from the surface, there would appear to be only limited molecular back-diffusion from hot gas to surface. This means that important exothermic reactions must be occurring in the condensed phase to a t least partially satisfy the endothermic r e quirements for steady-state gasification. Further interest in the condensed phase springs from the conviction that if overall combustion rates arc to be tailored or altered, the condensed phase reactions are the slowest and are also probably rate limiting, and therefore the best point for study and attack. Chemical analyses have been made of quenched propellant surfaces. They havc not been too rewarding since they indicate largely undecomposed products, as well as other products that, could havc been condensed from tlic gas phase during qurnching. But, of course, during quenching, severe changes of state arc taking place. Anothrr rrason for great interest in surface reactions is the dcsirc to get at thc very carlicst stages of decomThe Journal oJPhus*nl Chcmislru. Vol. 78. No. z4#1872
Figure 1.
c. BEACHELL
Crystalline oxidizer, single crystal eomhiistion model.
position. Key transient or intermediate specks might be prcscnt and detcct,ablc in the condensed phase, but may rapidly disappear and be lost to dctoction in the gas phase.
11. Theory Postulated Ezistace and Importance of H N O and N H , . A detailed theoretical analysis of AP decomposition by Jacobs and Pearsons calls for t.hc formation of nitroxyl, HNO, as NHI
+ O2
-+
HNO
+ OH;
AH = -6.2 kcal/mol
but NHz has not been dctcctrd xs :I decomposition product. Thcy point out that the method of production of the oxides of nitrogcn io AI' drcomposition wa.9 first theorized6 on the basis of HNO rawtioils 11s HNO
+ HNO
-+ HsO
+ N,O;
_ .. .
AH
=
-S%S kcal/mole
(1)
unimolecular to give nitric oxide HNO
-+
H
+ NO;
AH
= 49.6
kcal/mol
(2)
and oxidation to give nitrogrn dioxide HNO
+ Oz -+ NOI + OH;
AH
= -6.6
kcal/mol
(3) G u h o and Williams' propawl an 2dditional reaction for HNO leading t o nitric oxide
HCIO,
+ HNO
C108
+NO + H,O; AH
=
-11.79 kcal/mol
They state that the abovn reaction is essential to their gas phase kinetic scheme, i n the sense that the overall reaction rate would be much lowrr n.ithout the (4) T.L. Bog@, K . H. Kmeutle. iintl D. 1,;. Zuni,A I A A 3.. IO, 15 (1972). (5) P. W. M . Jacobs and G. S. I'eearson, Cmnbuat. Flame. 13, 419 (1969). (6) J. V. Davies. 1'. W. M. Jueobs, and A. Ilussell-Jones. Trona. Foradou &.. 63, 1737 (19G7). (7) C. Guirao and F. A. Williams. A I A A J . . 9, 1345 (1971).
CONDENSED P EASX DECOVPOSITION OF XH4C104 postulated step. They point out, however, that HNO has never been observed in AP decomposition or combustion experirnentE. Wilde8 has also pointed out the importance of HNO in the gaseous &-S 0 reaction. Selection of Bxpcrzmental Method. The method we feel holds the great e3t promise at this time for furthcr definition of overall reaction mechanisms and of surface decomposition products is mass spectrometry. Small samples can be heated t o temperatures near the surface temperatures obtained during combustion. Such temperatures can be supplied to a small sample near the inlet ?o a mass spectrometer by resistance heating of the sample container or by radiant healing of the surface. Under tlie ambient high vacuum conditions, any surface decomposal ion should be rapidly followed by vaporization. Unfortunately, in the case of ammonium perchlorate, rapid sublimation will also be taking place. Once i n the gas phase, under high vacuum, the species have little chince of collision reactions. They are ionized in thr electron stream and carried to the detector. HopefuIly, this occurs with a minimum of fragmentation or cracking. Pai Verneker and Maycock9#a and otbws have \Tamed that electron bombardment, of tlie sublimate can also produce many species. HPoivever, the ion:zmg voltage can be varied over a m ide range t o me if &astic changes in species take place for different t k c tron energies. Species requiring bimolecular reactions should not occur due to bombardment. Ako, < s a check on the mass spectrometer operation analyses caii b t run under test conditions in which other quantitative methods can be used. Check methods have revealed good correlation. However, the mass spectromelcr also identifies a number of highly reactive species that could not exist long enough to be detected by tl e s1ov:er methods. At any rate, time-of-figlit and high resolution mass Spectrometry hold the promise of providing data on primary specm that have long been postulated to occur, but have not been proved. Time-of-flight analyses have been wed in the past and have detected some of the primary specie^.^-^^ Pellett12 has shown that much the same kinds of products occur 1s hcther decomposition is caused by laser gash heating or by conduction heating of the sample in the spectrometer heater. To sirnplifj the bpecies identification problem, we have used ncat, high purity ammonium perchlorate (AF) and its deuterated derivative as the candidate propellants. AP has been found to burn as a monopropellant under a variety of physical conditions and thus makes art idcal test case.13 It is also a major component of most sold propellants. Wc and others have postulatrd the following primary products cap3ble of being formed in the condensed phase: Kj&, KPIy, C101, HC104, and the other hydrides and oxide, of nitrogen and oxides of chlorine.
3547 Yone of these occur as final products of combustion, either by chemical analyses or by thermochemical calculation. As stated above, Jacobs and Pearson and Guirao and Williams have postulated detailed mechanisms for the combustion process depending on the existence of HNO in addition. Time-of-flight spectrometry was chosen to scan all masslcharge values, while higher resolution methods were used to stJudy rcgions of interest, Prior work reported has not used high resolution spectra. In the future, further isotopic tagging is to be used to assure separation of species with similar mass/cbarge values. Future work will also be devoted to anionic and cationic changes to the candidate propellant material. Mack and Guill~ry"-'~have initiated some of this work, which will allow study of the reactivity of a variety of species to help strengthen the validity of the models proposed. Postulated Relationship of Lower T e m p e ~ a t u r e . Vacuum Conditions to Combustion Environments. The usual combustion testing environment for ammonium perchlorate as a single crystal monopropellant has been about 65 atm in a bomb pressurizcd with nitrogen at about 25". Under these conditions, the linear burning rate is about 1 cm/sec. The calculated equilibrium gas temperature is 1130", the most prominent gas species being 1320, 02, HCl, and Nz.DodeI7 and Levy18 have partially confirmed the products by chemical analyses. Attempts a t measurcment have indicated that this temperature might be as lox? as 1000", probably due to nonequilibrium conditions and mixing with surrounding cooler gases. The surface temperature is known from photographs to be well beyond the orthorhombrc-to-cubic phase transition temperature of 240" and is probably close t o the fast decomposition temperature recorded by clifferentia1 thermal analysis of 300 to about 450", depending strongly on the purity of the AP. Apparmtly, at temperatures greater than about 300", the sIc~ady-slate, self-sustaining decomposition reaction wc call combustion can take over. Below that temperature, only about 30% of the AP decomposes, and then reaction stops. MayeockQhas proposed a theory for this be(8) K. A. Wilde, Combust. Flame, 13, 173 (1969). (9) J . E.Maycock, 1'. R. Pai Verneker, and P. TV, M. Jacobs, J. Chem. Phys., 46, 2857 (1967). (10) V. R. Pai Verneker and J . N. Maycock, ibid., 47, 3618 (1967). 60, 1783 (11) G. A. Heath and J. R. Majer, Trans. Faraday SOC., (1964). (12) G. L. Pellett and A. R. Saunders, "Mass Spectrometer Pyrolysis of Ammonium Perchlorate at Low Pressure," CPlA Publ. No. 138, Vol. 1, John Roplcins Press, Baltimore, Md., 1967, pp 29-38. (13) E. E. Hackman, 111, and H. C. Beachell, A I A A J., 6 , 561 (1968). (14) J. L. iMack and G. B. Wilmot, J . Phys. Chem., 71, 2155 (1961). (15) W. A. Guillory and M.King, ibid., 73, 4367 (1969). (16) W. A . Guillory, J. L. Mack, and M . King, ibicl., 73, 4370 (1969). (17) M. Dode, Bull. SOC. Chim. Fr., 5, 170 (1938). (18) J. B. Levy, J . Phys. Chem., 6 6 , 109%(1962).
The Journal of Physical Chemistry, V o l . 76, X o . 24, 1972
E. E. HACKMAN, 111, H. H. HESSER, AXD R.6. BEACHELL
354s
havior. Our assumption in working at temperatures of 100 to 300” which is about the only practical range for examination b37 mass spectrometers, unless a flash heating method is used, is that the ma,jority of the nonelectronic reactions will be the same, at least initially, while the 30% decomposition is taking place. The higher combustion temperatures simply remove surface decomposition products fast enough by vaporization to allow steady-state and complete decomposition to take place. The pressure effect on combustion rates is very strong. I n fact, singlc AP crystals of millimeter dimensions will not burn at all below about 10 atm. Reasons for this were analyzed by Olfe and Penner.lg Even though some of the bases they used for calculations have been later found to be somewhat in error, their fundarnental conclusions seem valid. In addition to the first effect one would predict-that of increasing reaction rates with iiicreasing pressure due to greater reacting mass concentrations-the greatest effect is on the emissivity of the hot gases. Pressure would be expected to have little effect on the hot surface emissivity, but its value would be much larger than that of the gas cloud For smrili dimensions and low pressures. Olfe and Pcnner showed that radiation from the gas cloud to the surface for millimeter size dimensions increases tenfold when pressure is increased from 23 to 300 atm. Ehen more striking is the effect of increasing the simple geom ic (not necessarily the physical size with all its irregularities) size of the burning surface from a. square 3f 20 mm on a side to one of 20 cm on a side. This increases the radiant energy flux from gas to surface by about 30,000 times because thc larger gas cloud above the larger surface has a larger beam length, which- increams the gas cloud’s emissivity greatly. Thus, it can be s that where a surface temperature of greater than O is required, the hottest source available is the hot gas, and its effectiveness can be increased several hundred thousandfold by nonchemical changes; factors like pressure and geometry ar0 merely acticg as a hotter source to stimulate and sustain surface reaction. Therefore, wc make the assumption that, in mass spectrometry, although the pressures we use are very many times lower than those used in combustion, n-e are compensating for the lost thermal input from the gas phase by providing a direct input to the eample by the mass spectrometer heater. It is realized that dissociative sublimation will be very strong under the mass spectrometer conditions. Here we ho-oe that the high sensitivity of the devices being used wil enablr us to detect the surface decomposition products amid an anticipated flood of ammonia and perchloric acid, the first products of decomposition. 91 is interesting that most ammonium perchlorate containing propellants burn quite well, although very The Jozrnal of Ph:ysicul Chemistry, Vol. 76, N o .
$4, 197.9
slowly, at 1 atm. We believe this indicates that, in comparison with pure AP, condensed phase exothermic reactions are stronger and more controlling for the propellant (as opposed to gas phase radiant heating). The likely candidates are the oxidations C -+ COZ and AI -+Ale03 .
111. Equipment and Experimental Procedures The three experimental features of the study were the mass spectrometers used, the special sample probe designed to get surface decomposition products to the detector with minimum delay, and the high purity of the sample materials. Mass Spectrometers. Tzme-o&Flight (TOP). For screening studies of wide temperature ranges, probe experimental work, and the like, two time-of-flight mass spectrometers were used. This gavc an opportunity to compare background traces and t o determine the presence of any equipment artifacts in the scans. The instruments were Bendix Model 12-101 spectrometers operating with ionizing currents of 78 eV. This level of electron energy gave good reprodilcible spectra. We realize that this energy level may be causing some cracking of the gas species in addition to causing ionization sufficient for detection. Room temperature is associated with only about 0.025 el-. Expected surface temperatures of burning AP (400-900”) might be considered to have an energy level of approximately 0.10 eV. The energy associated with the highest gas combustion temperatures nould br lcss than 0.33 eV, so the electron beam energies used in any spectrometer have a far higher potential for causing decomposition than the thermal stress applied. The critical difference between the electron beam and thermal energy is that there is usually only one electron impact n ~ t ha species, but thermal radiation impinges on the species for a finite time. Other investigators have already recorded the chlorine oxides spectra produced by hot HC‘lOi. Guillory and King refer to the spectra as a cracking pattern quite similar to the species we find from AP. Other investigators have used very low- elecxtron beam energies in the region of 20 eV. This is further discusscd in section ii. The time-of-flight instr aments had a capture sensitivity of A. A Bendix Model 843 hot filament sample controller was used to control sample temperatures. A sample of approximately 10 mg was used. When the desired test temperature was reached and stabilized, a number of spectra w r c ruii until they were reproducible. Time-of-Flight (TOP) Sample Pmbc Sample orien(19) D. Olfe and 8.S. Penner, “Radiant Energy Emission from the Equilibrated Reaction Products of a Pure Ammonium Perchlorate Pellet,” Air Force Office of Scientific Research Technical S o t e No. 59-1094 (Contract N o . AF 49 (638)-412), Loekheed hIisdes and Space Division, Sunnyvale, Calif., Sept 1959.
CONDENSED PHASE DECOMPOSITION OF NH,ClO,
3549 cally large. The vacuum in the sample region was lo-' Torr and a t the detector it was lo-* Torr. High Resolution Sample Probe. As Figure 3 shows, the solid sample of AP to be tested is placed as near as possible to the electron stream. As the sample is depleted during the test,, it recedes somewhat from the ionization area, but in no case is it farther than about 20 mm. As shown, a Knudsen effusion cell is available, although it \vas not used during these t a t s . Maten'als. The ammonium perchlorate used was ultra high-purity grade prepared by American Potash Corp. This material had previously been used to prepare very pure single crystals. The deuterated vcrsion was prepared by triple recrystallization from D20. After vacuum drying, it gave no detectable proton signal on nmr analysis.
IV. Results Figure 2. Top ma48 Rpectmneter probe sketch.
tation and thermal condit,ioning wit,hin the sprctromrter are critical to reproducibility. Thr probe we used is shown in Figure 2. I n our studies we endeavored to assure that, the sample was heated uniformly and placed as near ns possible to the rlcctron beam. Thc limitation on nearness was to restrict the beam from being able to impingc directly on the solid samplc. Our goal was to know the temperature of the solid phase, to operate a t temperatures that would cause measurable decomposition of the condensed phase and production of gaseous specirs, and then to identify those species before any further decomposition or collisions could take place. The short period of timc see) and high vacuum (lO-'Torr) during the period from vaporization to detection would t,end to preserve the spccics from collision reactions. On the other hand, the high vacuum and the electron impact will tend to cause dccomposition that is an artifact of the analytical method. Our hope is that the artifact will be quantitative, rather than qualitative. That is, we are tacitly assuming a t this stage that the degree of decomposition we are measuring for a given temperature is really that for a somewhat higher temperatureif higher pressures and no electron beam were the ambient conditions. Thus, we are putting emphasis on detecting condensed phase decomposition rather than gas phase reactions. High Resolution Spectrometry. A CEC DuPont 211llOB instrument was used for these studies. A molccular leak of perfluorokcrosene 1va5 used in the background as a reference. With high resolution spectroscopy a number of runs are required to give assurance that peak heights at a given mass/chargc arc mcaningful, particularly with solid samplcs. A burst of d o composition occurring when the detector was measuring a given mass/charge \vould make that peak unrealisti-
Wit,h the m a s spectrometers used, resolution could be varied from differentiating between unit masses (TOF) to differentiating bctaccn massen differing only at the fourth decimal place. Background traces were a ncccssary adjunct to the understanding of samplc species and their amounts. Although care is taken to clean sample and ionization chambers and stabilizr the trace recording network, there is always the possibility of a peak being misintcrprctcd if sufficirnt background traces and rcpctitivc traces arc not availablc. The background trace i n 1"igurc 4 in typical for timcof-flight (TOF) mea.urcmcnts. Air impurities arc readily seen. Doubly charged nitrogen and oxygen arc seen a t m/c 14 and 1G. Hydroxyl and water arc nccn at 17 and IS. The mass 29 peak is probably C2Hj+ from a previous hydrocarbon sample. The peak at mass 40 is Ar+ as a minor impurity in air, as is the COX+ a t mass 44. Although attempts were made to kccp AI' samples pure and dry, there is thr ponnibility that some water and C02were absorbed and were introduced into the results. This tends to confuse thr undrrstanding of how much water is due to APdccompositioii and how much is due to absorbed water. When only
NO, c:
"/,""W
rr,,m u m p k aur,.c*
10 I" EIcClnm "I*l*"l.
UI 1 0
"7".'
Figure 3. Sample and probe orientation, high resolution spectrometer. The Journol o/PhlysicOl Chcmislry. Vol. 76, No. 24, 1972
E. E. HACKMAN, 111, H. H. HESSER, AND H. 6. BEACHELL
3550
28 32
OH 17
l _ _ l _ l
Ci03
HC104
MAWCHARGE
Figure 6. Time-of-flight mass spectrometer scan; NHiClOI at 95". Figure 4. Time-of-flight mass spectrometer scan; background a t 100". 28 32
h'iASS/CHA RG E
Figure 7. Time-of-flight mass spectrometer scan; NHaC104 at 115".
MASSiC H A RG E
Figure 5 , Time-of-flight mass spectmmeter scan ; I?dI'1&104 at 80'.
unit maw discrimination is possible, such as with TOF, the CO, prak also covers thr region at maw 44 whew S,O might be cletcctrcl. We are very much interested, h o w v r r , in thr completc absence of any prak at mass 31, ~ h r r HSO r is being qought in the sample. In our timc-of-flight scrcening studies, 80" was found to bc the hiqhryt tcmpcrature at mhich there was no dctcctablc dccamposition of AP (see Figure 5). Scans madr at 10 to 20" intervals from room tcmpcrature to SO" dctrctcd nolhmg but background. The strong prak.: at 28 :Lnd 32 arr due to a background of' air. It can br wcn that nothing is detectable at masses: 51 or 53 (35C10and 37ClC)); 67 or 69 ("ClOZ and 37C10L); 83 or 85 (35C103arid 37c103) ; or at 100 or 102 (H3%101 a d H jrCIO11, ,-nyof n hich \\-odd indicate perchlorate &.conpoiition. Btcfrrrmg to Figurr 6, it can bc seen that at 9.5" the appraraiicr of i h r chlorine oxide species has brgun. Thrrc 14 nlho rvidmcf for S O a t 30, HNO a t 31, and atomic chloruw at 3.5 and 37. HC1 at rnasws 36 and T h e Journal of L'hypical Chemistry, Val. 76, N o .
24, 1972
33 has not yet appeared, and molecular chlorine at 70 is definitely absent. At 115', Figure 7, HC1 at 36 and 35 suddenly becomes prominent. High resolution scan4 show that the strong increase in mass 44 is dur to Eurtbrr COL being desorbed from the sample. X30 mas detccted. HNO as mass 31 and NO at mass 30 and C1 at 35 and 37 have all reached a maximum, the examination for hydrides of nitrogen at lomcr m / c is discussed later. At 136' (Figure 8) HNO, ClO, and GIOL arc seen to be no more abundant than at 115". @lo3(83 and 85) and HC104 (100 and 102) now appear at a maximum. HC1 at mass 36 is nou- about as prominent as c103 at 85. It can be easily seen that there k no ClOS in any of these or in the following traces at masses 99 and 101. In Figure 9, at lbs", HC1 and C1O3 still vie €or position as the most abundant species. Tlicrc is still absolutely no indication of molecular chlorine. No SOL at mass 46 was detected in any of the scans; nor was HOC1 at masscs 46 and 48. Figure 10 is the decomposition pattern for deuterated AP at 140". DC104 is now prominent at ~nasses101
3551
CONDENSED PHASE DECOMPOSITION OF NH4C104 HDO
32
fJ7
I
YI
83
1Q0
z
2z 80
-I
2 GO a
M ASSKHARG E MAWCHARGE
Figure 8. Time-of-flight mass spectrometer scan; YH4C104 st 135". 28
212
31 HPIO
MASSICHARGE
Figure 9. Time-of-flight mass spectrometer scan ; KH46104 at 165'.
Figure 10. Time-of-flight mass spectrometer scan; lrTD4CIO4 at 140".
16, 17, and 18, respectively, are NH?, NH3, and SN,. WH4 is quite small and easily lost on lower sensitivity and low resolution traces. The presence of some SIL+ was confirmed by detection even when the electron beam of the mass spectrometer was turned off. The NH, peak at mass 16 is roughly 50% of the overall mass 16 signal due t o the combination of doubly charged oxygen and YH2. The oxygen, unlike nitrogen, is mostly due to AI' decomposition. The strong OH signal is almost all due to the sample; we assume part is due t o AP decomposition. The strong water signal is almost all due to the sample, and we assume some part of it is AP decomposition and the remainder is water and or an impurity in the sample. Alhougb not shown here, the higher mass scans show-ed C 0 2 to be present due to the AP, but no N2Qwas present at mass 44.
V. Data Evaluation and 103. 'The extent of proton impurity can be readily detected by the HC104 peaks at 100 and 102 amounting to about 10% of the total. This seems to indicate that delays in andysic; of deuterated AP, while subject to ambient air from time to time, can lead to absorption of water and significant reversion to NH4C104. The freshly prepared deuterated AP, after drying, gave no minimum signal on nmr analysis, indicating less than a few hundredths OF I% proton content. Most significant in this figure is the absence of mass 31, evidently due to D E 0 now occurring at mass 32 (along with [I2).k puzzling part of this trace is the nearly SOc% content of HC1 in the overall HC1-DC1 signal. There is no indication of IYD, at mass 22 in this lower sensitivily run, but ND is present at mass 16 and D20 is strong at mass 20. A typical higheo resolution (complete peak separation between 011 and SH3) scan of the lower mass numbers is thomi in Figure 11. The major peak at mass 14 is doubl-y. charged K2,but it is entirely due to the background. The major peak at mass 1 5 is NH. The smaller in s i e r but higher mass number peaks, at
In our studies, as compared with those conducted at atmospheric or higher pressures, we should be producing larger quantities of NH, and HC10, due to dissociative sublimation. If we did not detect these two as initial and major species as temperature was raised we would suspect that tlie low pressure in the mass spectrometer and the electron beam energy were cracking the ammonia and perchloric acid known to be formed and were producing the decomposition species. A review of the figures showing uncorrected relative species abundance from 80 to 165" shows that NHS and HC10, are among the first species formed, and they remain as major species at all temperatures tested. Maycock and Pai Vernekerlo h a w proposed il point defect mechanism t o explain the fact that at atmospheric pressure only 30% of ammonium perchlorate decomposes below 300". At temperatures above 350", decomposition is complete. The species we have found and the mechanisms t o account for them are in accord with their mechanism, which calls for production of species such as NH4, SEIs, H20,HClO,, and C103 in the condensed phase. The Journal of Physical Chemistry, Vol. 76, K O .24, 1972
E. E. HACKMAN, 111, H. H. HESSER, AND H. 6. BEACHELL
3552 OH
H2°
NOTE. N o peaks recorded below M/C = 14.
22
MadChargr
'
28
'
24
Figure 11. High. resolution mass spectra; ammonium perchlorate at 165", m/c 14-24.
I n this set of cxpcriments IT(: did not' find a temperat'ure a t which NH, and HClOl ww tho only products. In the first appearancc of HC10, a t 95", C103, the first decomposition product of HClO.), is already equally abundant, and C10, is possibly one-third as abundant. A more detailed stJudy of the temperature rangc SO95" should be z " e . Our spcctra scemcd to correlate quite closely with thn figures and descriptions given by Guillory and. Iiing.15 This includes the chlorine oxide specics " r c v e r d r s f f cct" they referred to. Guillory and Icing also gave the spect'rum for perchloric acid itself. Honcvrr, thcy stat'e that HClO, produccd as an cvaporatcd product remains relativcly stable up to 320". Therefore, t'hcy attribute chlorine oxido species in t'hc 200" region arid be lo^ to HClO, cracking by the el.ect,ronbeam. This may ncll be, but Levy'8 reports heterogeneous RClO, reactions bclow 300", the rate depcndirig on the nature of the fiurfacc with which it is in cont'act. It might bc clxpcctcd that tho surface of decomposing AP, containing NH4 and ISH3 species, and probably NH2 and NH specks, could provide an environment favoring HC101 decomposition. The abstraction of OH at the wealier 61-OH bond tvould start the forma~tionof thc three chlorine oxides. Apparently, we cannot yet say for certain what portion of t'hc HCI0, decomposition species (and by analogy the hydridcs of nitrogen spccics) is duc to thermal drcomposition and what portion is due t o electron beam cracking. WC do know that the intensities of the species grow in a striking fashion as the tcmperature is increased from 95 to 165". Also Pellett and Saunders12 have rcported not much difference in the relative abundance of the species when reducing clectron beam energy from 70 don-n to 20 eV. They also state that craclring products mid solid decomposition products can bc differentiated. It must bc realized, however, that %O-cV dcctrons are approximately several hundred tirnw m.ore powerful than the thermal environment for initiat'ing uncatalyzed decomposition. The factor t,hat sdd~icredence to thermal decomposition as the sourcc: of many species-as opposed t,o cracking-is the prescsicc of species such as HC1, HNO, The Journal of Piiysical Chemistry, V o l . 7 6 , No. 24, 1972
and NO. Prrchloric acid has not becn shown to produce HCl as a dccomposition product, H X O and NO almost certainly require a number of decomposition steps by NH3 and HC104, fo1lowc.d by oxidation of nitrogen hydrides. Such rca.;oning leads us to believe that many of the species detected actually were formed on the surfacc of thc thermally dccornposing NH4C101. Once such spccies arc formed on the surface, it requireq only that they be vaporized, usually as a radical, struck with an electron which strips off onc electron stnd forms the positive ion. Then, with almost no chance of any further reaction, the ion is acccleratcd to the idcntification sector of the mass spectrometer. Thus, surface reactions should be the last reactions occurrixig. Table I gives a series of reactions IW believe could occur on thc surface of thermally decomposing ammonium pcrchlorate a t temperatures as low as about 100" when under high vacuum. Although direct spectral evidence for mobile C104- has not been found, the detection of small quantities of N H p both with and without the electron beam turned on indicate that the perchlorate ion might havc been present. Reaction IC indicates pathways by which both NH, and NH4+ might havc becn formed. To malic the perchlorate ion a positive ion so that it can be detected rcquirc.s stripping off an electron pair
6104-
-3ClO*+
This is much less likely to occur than (3104-
( 3 1 0 4 . (radical)
and the radical would not bc detected. On the other hand, any NH4+leaving the surface in the high resolution apparatus would be inimediatcly accelerated to the dctector. Such was found with the electron beam turned off. Further work will be required to determine the proportions of NH4 and ISH4+being formed. The presence of NH, and absence of C104 might seem to militate against the interpretation that radicals left the surface because, in that case, one might assume a nearly equal probability of detecting either ammonium
CONDENSED PHASE ECOMPOSITION
OF
3553
NH4C104 Table I1 : Ammonium Perchlorate Monopropellant Combustion Equilibrium Calculntionsa
l---(bj
.--+ NH4++ Clod-- (minor)
OR Y
O
z
Combustion species, mo1/100 g
+ O2
c1 HCl C10 Clz OH Ha0
NO Na 0 2 a
,-------Pressure-------? 1 atm
34 a t m
68 atm
0.0054 0.8168 0.0001 0.0144 0.0002 1.2937 0.0011 0.4250 1.0546
0.0022 0.7301 0.0002 0.0593 0.0001 1,3371 0.0013 0.4249 1.0329
0.0019 0.7005 0.0002 0.0742 0.0001 1.3519 0.0013 0.4249 1.0254
Other products less than 0.00005: H, NH, Hg, NH3, N, and
0.
or perchlorate. However, ClOJ is highly unstable, decomposing unimolecinlarly to ClOz and 0 2 . The second rc.action producing OH and C103 has been widely proposed and is well supported by our findings. Our reaction steps suggest that all decomposition starts with this step. In II@104t5e syirimctry of the perchiorate ion is upset. The Cl-O bond, which was 1.408 A for each of the oxrgens, is lengthened to a C1-OH bond length of 1.630 A. Thk sets the stage for He104 decomposition t o @LO3 and OH radicals. The third and fourth reactions providing two pathways to ClOz formation tend to explain the presence of more ClOz and C10. Reaction 10 suggests one method of C10 formation. Reaction 5 shows how 6 1 0 2 could decompose without going to ClO. Reactions 6 through 9 show reasonable ways for forming the various hydrides that are found, plus NCl, HNO, and NO. Oxygen was found to substantiate its utiagc in several reactions. Thp reaction steps shown as occurring allow for the formation of most of the stable end products of equilibrium combustion except nitrogen. Tables I1 and IiI show the ealculated combustion temperatures, gas molecular weights, and species distribution for three pressures. It C % L ~ be wen that low pressures favor formation of E[G1 and 61 at the expense of Clz. Thus, where reactions arc carried out a t very low pressure, we mould not expect to see Clz. Table IV shons a series of nine reaction steps that :ire not justified bawd on the mass spectral analyses. The iormatior: of Clz (reaction 5) probably would take place a t higher pressure. The underlined species represent those species not detected. NOz, NzO, HOC1, Nz, and 612hssve been reported as decomposition products by other investigators. We believe this may be the differentiation hctween condensed phase or surface rractions and gas phase reactions. We know that nitrogen is a major m d product. However, we find no
Table I11: Ammonium Perchlorate Monopropellant Combustion Equilibrium Calculations
---..
H I = -70.69 koal/mol-------. ____-_ Pressure---------
Parameters
1 atm
34 a t m
68 a t m
Temp, "K Temp, OF Av mol wt
1375 2015 27.691
1397 2064 27,871
1403 2066 27.930
Table IV : Typical Nonjustified AP Coriderised Phase Ilecomposition Reaction Steps 1. (Condensed) NH4
+ C1O 3. OH + c1 4. HNO + 5. c1 + c1 6. NHz + E?( 7. NI3z + 8. N - + NO 9. NH2 $- HNO 10. N H + C1 2. NHI
0 2
0 2
L1
+- ClO4-
---+ NH4ClOa (gas phafie)" -~ + NW* + ___ HOC1 --+ IiOCl _ I _ .
> NO2 _- + OH * _._. c12 > NzO MzO --+ NH 4- _HOZ
+
_ _ I
--
NzQ
* N ~ O-t hrzo * .-M +- HC1 ~
Underlined species not detected.
evidence of rither N or Nz coming from the surface. We theorize that the oxidation of the nitrogen hydrides just does not proceed beyond N H at the surface.
Conclusions 1. I'lost of the species postulated as occurring in A P decomposition were found. The spectrometers recorded only positive ion species. Thercfore, the species detected were positive ions or were formed by electronimpact stripping of one or more elrctrons from the actual thermal decomposition specks. 2 . Thr species HNO and XH, were found as conTho Journal of Physical Chem.islry, Vol. '76, ;L'o. ,%* 1972
METZGER, et al.
3554 densed phase products providing support for two theories of decomposition mechanisms. 3. Sma!l quantities of NH4were found, but no Clod. NH4+leaving the condensed phase was detected with the electron beam turncd off. Clod-, on the other hand, could only have been detected by the loss of two electrons. C104 ~ o u l dhave been formed, but unimolccularly decomposed so rapidly as not to be detected. 4. No evidence for Clz was found. Atomic chlorine was formed among sll decomposition. species. This is probably due to the low pressure studied. This checks
with the results of Guillory and King, but is a t variance with the results of a number of other investigators. 5. Nitrogen, NzO, NOz, and HOC1 were riot detectable as condensed phase decomposition products. 6. There is evidence that the majority of all proposed reactions of AP decomposition take place ih the condensed phase at least to some extent. Thus, sufficient exothermic reactions can bc made available under proper pressure conditions to feed energy to the prime condensed phase endothermic reactions and to sustain combustion once the reaction chain is established.
et on the Fading Rate of Photochromic 3-Substituted
Benzothiazolinic Spiropyrans by A. Samat, J. Metzger,* Laborntoire de chimie organique A associe’ au C.N.R.S. (LA126), Universiti de Provence, 1S-Marseille (ZSO), France
F. Mentienne, F. Garnier, J. E. Dubois, Laoortztoire de chimie organique physique associe’ au C.N.R.S. (LASd), UniversitB de Paris V U , 75-Paris (Se), France
and R.Guglielmetti Laborntoire de synthhse organique, Universitd de Bretagne Occidentale, 29283 Brest-Ceder, France (Reeeiised March SO, 1972) Publication costs assisted by the Universite‘ de Rretagne Occidentale
Reversible transformations of spiropyrans into merocyanines are studied in a series of seventeen photochromic 3-substituted benzothiazolinic spiropyrans. The absorption spectra of the colored photomerocyanines and their first-order tjhermal fading kinetics are followed using a rapid scanning spectrometer coupled to a flash photolysis apparatus. Activation energies and entropies are calculated in toluene. The rate processes are shown to be very sensitive to 1-1bonding of the solvent and to the nature and position of substituents; a rate enhancement of lo5is observed between the substituents X = OCH3and X = i-C3H7. The data indicate that steric hindrance of substituents in the 3 position of a planar photomerocyanine has major importance in the rate of corwersion of the colored form back to the spiropyran.
Introduction Spiropyrans arc known to behave as photochromic compounds giving merocyanirie dyes when uv irradiated. The stabi1i.y of the photomerocyanine is related to the high degree of conjugation allowed by a nearly planar conformation. A great deal of data collected on the spiropyrans ha9 led to the structure and properties of this colorltm form S. On the other hand, thermodynamic and spectroscopic properties of the colored form and the cffeet of substituent X in the 3 position are not very ne11 known. In the work reported here, we arialyzc thc problem, oftcn mentioned in the literature but rarely discussed, of the conformation and stability The Journal of Physical Chemistry, Vol. 76, N o . $4, 197.9
5’
0 NO,
dZ’O s
+&;@16
CHJ
hduv)
~===.=2
7
OCH,
S
CH:,
M
